Bulletin of the British Museum (Natural History). British Museum (Natural History) Converted as part of the ABLE project by Dauvit King London : BM(NH) Continued as: Bulletin of the Natural History Museum. Entomology series Vol.1 (1950) - vol.61 (1992) 2 4 This document has been converted to TEI XML as part of the ABLE project to make it more widely available to biodiversity researchers in a useful format. eng text No corrections have been made in the text. The original source has not been regularized or normalized. Quotation marks have not been processed. They are as in the original DjVu XML document. Hyphens, including end-of-line hyphens, have not been processed. They are as in the original DjVu XML document. The text has been segmented based purely on layout based on page breaks. No language level segmetation, such as sentences, tone-units or graphemic, has been applied. Additional mark up using taXMLit has been applied to the TEI XML based on analysis of the original source through the uBio and OpenCalais web services. (Add comment for fuzzy matching once this has been brought into the final workflow too.) M BRITISH MUSEUM(NATURAL HiSi 26 JU PRESENTEDGENERAL UC.-lARY Bulletin of the British Museum (Natural History) The ichneumon-fly genus Banchus(Hymenoptera) in the Old World M. G. Fitton Entomology series Vol51 Nol 25 July 1985 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in fourscientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology,and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique andever-growing collections of the Museum, both by the scientific staff of the Museum and byspecialists from elsewhere who make use of the Museum's resources. Many of the papers areworks of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself,available separately, and individually priced. Volumes contain about 300 pages and severalvolumes may appear within a calendar year. Subscriptions may be placed for one or more ofthe series on either an Annual or Per Volume basis. Prices vary according to the contents ofthe individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History),Cromwell Road, London SW75BD,England. 'World List abbreviation: Bull. Br. Mus. nat. Hist. (Ent.) Trustees of the British Museum (Natural History), 1985 The Entomology series is produced under the general editorship of the Keeper of Entomology: Laurence A. Mound Assistant Editor: W. Gerald Tremewan ISBN 565 06011 2 ISSN 0524-6431 Entomology series Vol 51 No 1 pp 1-60British Museum (Natural History)Cromwell RoadLondon SW7 5BD Issued 25 July 1985 BRITISH MU: The ichneumon-fly genus Banchus (Hymenoptera) inthe Old World -^1 M. G. Fitton Department of Entomology, British Museum (Natural History), Cromwell Road, LondonSW7 5BD Contents Synopsis 1 Introduction 1 Banchus Fabricius 3 Taxonomy 3 Biology and hosts 5 Systematic list of hosts 6 Nomenclatural summary 7 Format, material examined and depositories 8 Keys to species 11 Males 11 Females 12 Systematic section 14 Species excluded from Banchus 43 Acknowledgements 45 References 45 Index 59 Synopsis This paper provides the first comprehensive account of the genus Banchus (Ichneumonidae, Banchinae) inthe Old World. Twenty-three species are recognized, described and keyed. Seven of the species are new.Geographical distributions and data on biology and hosts are summarized. Known hosts are Lepidoptera(mainly of the family Noctuidae) and Banchus species are known to play an important role in the control ofsome pest species in Europe and North America. The taxonomic treatment includes 11 new synonymies,one provisional synonymy, three new combinations, one revised combination, two revised statuses, threelectotype designations and four nomina dubia. Introduction Banchus has had a chequered history. Taxonomists found it difficult to assign to one of the fivetraditional subfamilies of the Ichneumonidae, sometimes placing it in the Ophioninae andsometimes in the Pimplinae. As originally defined by Fabricius (1798; 1804) the genus includedmembers of the families Braconidae, Cephidae, Pompilidae, Proctotrupidae and Ibaliidae, aswell as the Ichneumonidae. This exotic assortment was reduced by Gravenhorst (1829) to agrouping resembling the genus as understood at present. The main error in his treatment was thetransposition of the sexes. Developments in the following 150 years were mainly concerned withdescription of new species, splitting the genus and faunistic studies. Much of this work was ofpoor quality; no one made a comprehensive study and those describing and keying species failedto appreciate all but the most superficial characters. A real advance in our understanding of thetaxonomy of Banchus came only with publication of the revision of the Nearctic species byTownes & Townes in 1978. Bull. Br. Mus. nat. Hist. (Ent.) 51 (1): 1-60 Issued 25 July 1985 M. G. FITTON I8 E ICHNEUMON-FLY GENUS BANCHUS 3 BANCHUSFabridus Banchus Fabricius, 1798: 209, 233. Type-species: Banchus pictus Fabricius, by subsequent designation(Curtis, 1836: 588). Corynephanus Wesmael, 1849: 631. Type-species: Banchus monileatus Gravenhorst, by monotypy. Cidaphurus Foerster, 1869: 159. Type-species: Ichneumon volutatorius Linnaeus, by subsequent mono-typy (Woldstedt, 1877: 439). Nawaia Ashmead, 1906: 184. Type-species: Nawaia japoni ca Ashmead, by monotypy. DIAGNOSIS. Moderately-built ichneumonids (Fig. 1), fore wing length 7-7-14-3 mm. Frons simple. Marginof clypeus with a median notch. Genal carina joining hypostomal carina above base of mandible. Uppertooth of mandible much wider than lower tooth and longer, its apex truncate and weakly subdivided.Lower tooth pointed. Maxillary palp with segment 4 more or less widened and flattened distally, especiallyin males. Labium not elongate. Antennal flagellum long, tapered to a slender apex. In males, flagellumwith poorly- to well-developed specialized setae (flag setae), 2 to 4 per segment, often arising from apoorly- to well-marked, shallow trough on the dorsal surface. Epicnemial carina absent. Scutellumproduced posteriorly into a spine, although it is sometimes reduced to a very weak tubercle or even absent.Propodeum short, its posterior transverse carina strong laterally, weak or obsolete medially. Pleural carinapresent, at least in part. Tarsal claws pectinate. Laterotergites of segments 2 and 3 of gaster about 0-7 timesas deep as long. Gaster more or less compressed, relatively strongly so in most females. Ovipositor veryshort, its sheath about 0-1 times as long as hind tibia. REMARKS. The genus comprises 47 valid, described species. One is Holarctic in distribution, afurther 24 occur in the Nearctic (Townes & Townes, 1978) and another 22 are found in thePalaearctic and the mountains of the Oriental region. Banchus is the largest of a group of five genera united by the structure of the mandible. Themandible is of an unusual, almost tridentate form unlikely to have arisen more than once withinthe Banchinae. A mandible similar in structure is found only in the Diplazontinae (in whichsubfamily it occurs in all species). The other genera in the Banchus-group are: Rhynchoban-chus (with 3, possibly 4, Palaearctic species); Banchopsis (1 Palaearctic and 1 Oriental species);Ceratogastra (1 Palaearctic and 1 Nearctic species) and; Philogalleria (7 Australian species and 1species in Argentina which possibly deserves a separate genus). Within this group the mostimportant discriminant characters are the form of the claws, the presence or absence of theepicnemial carina, the extent of the propodeal carinae, the form of the frons, the position ofcross- vein 2m-cu in the fore wing, and the relative sizes of the coxae. The tribe Banchini, as currently recognized (Townes, 1970: 45; Townes & Townes, 1978:378), includes two disparate elements- the Banchus-group of genera and Exetastes and its allies.Cushman (1937), Perkins (in Beirne, 1941) and Beirne (1941) all supported separation of thetwo groups on the basis of adult and larval characters. Cushman and Perkins associated theExetastes-gioup more closely with the tribe Lissonotini. Townes (1944) on the other hand linkedthe two groups as the Banchini, separate from the Lissonotini, mainly on the basis of a ratherweak character in the hind wing venation. Townes has consistently maintained this position.Viktorov (1967) supported Cushman's (1937) conclusion, citing several characters of Exetasteswhich distinguish it from the Banchus-group. Aubert (1978) has gone further, placing Exetastesin a tribe of its own, Exetastini. As with many ichneumonid subfamilies the generic and tribalclassification of the Banchinae, although workable, is in need of closer study. Townes' (1970)treatment and key fail to take account of the structure of the mandible in Philogalleria (of whichTownes may have then seen only one specimen). Taxonomy Despite its long period of gestation this study, based on the examination of 3421 specimens ofOld World Banchus, cannot be regarded as definitive. It has raised as many taxonomic questionsas it has answered. The lack of specimens from some problem areas (taxonomic as well asgeographic), coupled with an inadequate knowledge of most aspects of the biology of the genus,has effectively prevented discrimination of all of the species. A conservative approach wasadopted and some of the morphospecies recognized undoubtedly comprise complexes of two or 4 M. G. FITTON more 'biological' species. With the material and resources available further splitting could not bejustified on scientific grounds. Apart from obvious species-pairs or very small groups it is difficult to deduce anything of thephylogenetic relationships of the species within the genus. As noted by Townes & Townes(1978) in their revision of North American Banchus, the characters which distinguish species donot correlate in ways which indicate clear-cut species-groups. Specialization by reduction (anobvious example is in the length of the scutellar spine) is probably one of the complicatingfactors. The best that can be done is to suggest trends of specialization for individual characters. Many species show considerable sexual dimorphism (including colour patterns). The mostobvious morphological differences are in the shape of the gaster , the antennae and the maxillarypalps. The specializations in the female seem to be associated with oviposition and in the malewith mate-finding and courtship. In some species the overall colour pattern is the same in maleand female, in others it is completely different, for example, in B. falcatorius and B. volutatoriusthe females are largely black with reddish legs while the males are conspicuously marked withyellow. This female colour pattern, although still aposematic, probably renders her lessconspicuous while searching for hosts (which, in B. falcatorius, probably occurs very low in thevegetation at soil level). As a general rule males have a yellowish face with a median black stripeand females a black face, sometimes with yellowish orbital marks, and there are often alsodifferences in the colour of the antennae. The characters which have been used in this study are discussed below. The formaldescriptions of species in the systematic section of the paper are restricted largely to thecharacters used in the keys. When comparing specimens with figures allowance should be madefor some variation. Head. The overall shape of the head is important in distinguishing species. The charactersmost easily used are the relative width of the face and the size of the malar space. These aremeasured as indicated in Figs 2 and 3. The shape of the head behind the eyes varies betweenspecies but is much more difficult to use as a distinguishing character. The maxillary palps arespecialized in many species (Figs 11-32, 103-123). The relative lengths of the two terminalsegments are important and in males the way in which the fourth segment is widened andflattened is characteristic. The antennae vary in length and in the shape of the terminalsegments, both characters being difficult to use in keys. In the male the distal half of the antenna bears specialized setae ('flag setae') on its dorsalsurface (Figs 124-129). The development of these varies greatly between species. In the mostprimitive cases the setae are not easily distinguished from others on the antenna. In the mostspecialized forms the setae (two, three or four per segment) are erect, widened and flattened andarise from a shallow polished trough. All segments do not bear these setae, there is a zone (whichvaries in position between species) in which they are best developed and proximally and distallyto this the setae become progressively less modified. The descriptions given for each speciesapply only to the zone bearing the most modified setae. The shape and number per segment aresubject to some variation, but together with the form of the male maxillary palps these setae givethe best 'key' characters for a species. Both are probably important in courtship. Thorax and propodeum. The scutellum bears a spine on its posterior apex (Figs 33-58). Theprimitive condition, it is supposed, is to have a well-developed spine. The length of the spine isimpossible to measure accurately because of the lack of reference points, and there is variationwithin species. Nonetheless it is still useful taxonomically. The lower postero-lateral corners ofthe mesothorax in some species are formed into weak tubercles which, although relativelyconstant in their development, are difficult to use comparatively. In the descriptions the'posterior part of the propodeum' refers to the posterior transverse carina and the area behindand enclosed by it. The posterior transverse carina itself is often obsolete or absent medially. The measured proportions of the hind femur (as shown in Fig. 5) are used to represent thegeneral proportions of the legs as a whole. The length of the fore wing (measured from the tip ofthe tegula to the tip of the wing) is used as a measure of overall size. Gaster. The form of the gaster in the female offers a number of very useful characters.However, their practical value is much reduced by the very variable amounts of distortion found ICHNEUMON-FLY GENUS BANCHUS 5 in dry specimens. Allowance must be made for this in comparing specimens with descriptionsand figures, particularly those of the apex of the gaster (Figs 82-102). The distortion most affectsthe amount of compression of the posterior half of the gaster, the telescoping of segments andthe compression and profile of sternite 6 (the subgenital plate) which in some species has aweaker, less sclerotised portion on its posterior margin. The form of tergite 1 is not subject todistortion and it offers useful characters in males and females (Figs 59-81), although subject tosome variation (compare Figs 65 and 66 of male and female gudrunae). The proportions oftergite 1 are measured as indicated in Fig. 4. Sculpture and colour. The development of microsculpture and puncturation on the thorax,coxae and gaster varies between species but variation within species and difficulties of descrip-tion limit its value. Colour patterns are subject to variation but are easily observed. Care wastaken to ensure that such variation was taken into account in construction of those parts of thekeys relying on colour. Biology and hosts Townes & Townes (1978) state that Banchus is a genus of open shrubby country and that femalesfly rather low and males higher and faster. However, there are few published observations on thehabits and habitats of individual species. Adults have been collected from flowers. The black andyellow and/or reddish colour patterns may be aposematic (Townes & Townes note thatspecimens give off a strong pungent odour when captured). B. falcatorius has been recorded asthe prey of an asilid fly. Most species probably are univoltine and adults are on the wing for some period during latespring or early summer. A few species (e.g. B. dilatatorius) occur in early spring. These probablypass the winter as adults or pharate adults within the cocoon. In Europe three species (B.zonatus, B. gudrunae and B. moppiti) have been taken in late autumn, winter and early spring.Of these three, B. zonatus and B. gudrunae are Mediterranean in distribution whilst the fewknown specimens of B. moppiti come from localities as scattered as Great Britain, Switzerlandand Spain. There appears to be a similar pattern in the species in the eastern Palaearctic andmountains of the Oriental region, with some specimens of the more southern species occurringin the period from October to May. However, some of these species might not be univoltine. Courtship behaviour in B. hastator is described by van Veen (1982): when the maleencounters a female he orientates himself face to face and starts fanning with his wings, probablycreating an airstream along the female and himself. The gaster of the male is also raised and atintervals these actions are interrupted by an attempt to mount the female. Using the tips of theantennae the male gently strokes the legs and sides of the thorax of the female. If the female isreceptive copulation takes place. The hosts of Banchus are Lepidoptera. All of the reliable observations relate to species ofNoctuidae, though species of seven other lepidopterous families are also recorded. Whether ornot any or all of the non-noctuid records should be rejected is difficult to assess. In the Nearcticspecies B. flavescens it has been shown that females oviposited readily in the larvae of its usualhost Mamestra configurata Walker and also in Scotogramma trifolii (Rottemburg) and Tricho-plusia ni (Hiibner) , but had to be induced to attack A utographa calif ornica (Speyer) , A . falcifera(Kirby) and Heliothis ononis (Denis & Schiffermuller) by an immediately preceding exposure toM. configurata larvae (Arthur & Ewen, 1975; Ewen & Arthur, 1976). All these hosts arenoctuids and in addition to the reluctance of the females to oviposit in some species the parasitefailed to develop in any except M. configurata, because of successful host defence reactions. VanVeen (1982), working with B. hastator in the Netherlands, found 'that an antennal contact of afew strokes with the cuticle of the host's skin was sufficient to reject a strange host by walkingaway or to a susceptible instar of Pfanolis] flammea [the usual host] immediately by an acutesting reaction'. He unfortunately does not identify the 'strange host'. All these results could beused to support a contention that Banchus species are probably monophagous on species ofNoctuidae, but for one species, B. volutatorius , there are reliable rearings from at least threenoctuid species (Anarta myrtilli (Linnaeus), Lacanobia oleracea (Linnaeus) and Xestia xanth- 6 M. G. FITTON ographa (Denis & Schiffermiiller)), so judgement must be reserved. The host records from twoHymenoptera (Maneval, 1935) can probably be dismissed out of hand. The larvae normally develop as solitary internal parasites and it is usually suggested thatoviposition is into an early instar host larva. In experiments, van Veen (1982) has demonstrateda preference for early instar larvae in B. hastator, although females will attack all larval instarsexcept the last (the fifth) (van Veen, 1982; Bledowski & Krainska, 1926). The larval develop-ment of B. hastator was studied in detail by Bledowski & Krainska (1926). The head sclerites ofthe final larval instar of four species are illustrated by Short (1978). Short's work should be usedwith caution: his figures of Banchus show why. The illustrations of B. femoralis and B. hastatorshow differences which might be thought significant; however, both specimens are of the samespecies (B. hastator, of which femoralis is a synonym) and one figure shows the view from outsidethe head capsule while the other shows the view from within! Other figures are given by Beirne(1941) and Barron (1976). The full-grown Banchus larva kills and leaves the host larva after it has left its foodplant andentered the soil to pupate (this has been reported in several species). Cocoons of Banchus areelongate-ovoid, the silk dense and almost black. The adult emerges through a hole cut next toone end. Barron (1976) records a species of Banchus (from Maryland) parasitized by Euceros medialisCresson (Hymenoptera: Ichneumonidae). Systematic list of hosts The list below includes all the hosts of Banchus detailed elsewhere in this paper. Those detailscan be accessed via the index. LEPIDOPTERA ZYGAENIDAE Zygaena ephialtes (Linnaeus)NYMPHALIDAE Melitaea didyma (Esper)THYATIRIDAE Achlyaflavicornis (Linnaeus) Habrosyne pyritoides (Hufnagel)GEOMETRIDAE Opisthograptis luteolata (Linnaeus)SPHINGIDAE Deilephila elpenor (Linnaeus) Deilephila porcellus (Linnaeus) Hyloicus pinastri (Linnaeus) Smerinthus ocellata (Linnaeus)NOTODONTIDAE Phalera bucephala (Linnaeus)LYMANTRIIDAE Dasychira fascelina (Linnaeus) Euproctis similis (Fuessly) Leucoma salicis (Linnaeus)NOCTUIDAE Acronicta megacephala (Denis & Schiffermiiller) Agrochola circellaris (Hufnagel) Agrochola helvola (Linnaeus) Agrotis exclamationis (Linnaeus) Agrotis segetum (Denis & Schiffermiiller) Anarta myrtilli (Linnaeus) Aporophyla lutulenta (Denis & Schiffermiiller) Aporophyla lutulenta luenebergensis (Freyer) Atethmia ambusta (Denis & Schiffermiiller) Bena prasinana (Linnaeus) Blepharita adusta (Esper) ICHNEUMON-FLY GENUS BANCHUS Ceramica pisi (Linnaeus) Euxoa nigricans (Linnaeus) Hadena compta (Denis & Schiffermuller) Hadena rivularis (Fabricius) Heliothis viriplaca (Hufnagel) Lacanobia contigua (Denis & Schiffermuller) Lacanobia oleracea (Linnaeus) Lacanobia suasa (Denis & Schiffermuller) Lycophotia porphyrea (Denis & Schiffermuller) Mamestra brassicae (Linnaeus) Panolis flammea (Denis & Schiffermuller) Phlogophora meticulosa (Linnaeus) Xestia xanthographa (Denis & Schiffermuller) HYMENOPTERA DIPRIONIDAE Diprion pini (Linnaeus)ICHNEUMONIDAEOphion luteus (Linnaeus) Nomenclatural summary agathae sp. n. altaiensis Meyer, 1927 (nomen dubium)cerinus Chandra & Gupta, 1977cre/e/densisUlbricht, 1916 croaticusHensch, 1928dilatatorius (Thunberg, 1822) stat. rev. acuminator (Fabricius, 1787) (homonym) syn. n. compressus (Fabricius, 1787) (homonym) sibiricus Meyer, 1927 syn. n.fa/catorius (Fabricius, 1775) variegator (Fabricius, 1775) intersectus (Geoffrey, 1785) aries (Christ, 1791) labiatus (Schrank, 1802) histrio (Schrank, 1802) (homonym) tricolor (Schrank, 1802) falcat or Fabricius, 1804 sachalinensis (Matsumura, 1911) luteofasciatus Ulbricht, 1911 nobilitatorMorley, 1915 sanguinator Meyer, 1922 lavrovi Meyer, 1927 nigromarginatus Constantineanu & Pisica, 1960 propitius Kuslitzky, 1979 (provisional synonymy)flavomaculat us (Cameron, 1904)gudrunae sp. n.hasta tor (Fabricius, 1793) pungitor (Thunberg, 1822) reticulator (Thunberg, 1822) syn. n. femoralis Thomson, 1897 kolosovi Meyer, 1925 syn. n.insulanus Roman, 1937japonicus(Ashmead, 1906)mauricettae sp. n.moppitisp. n.noxMorley, 1913pa/pa/JsRuthe, 1859 stat. rev. spinosus Cresson, 1865 syn. n. 8 M. G. FITTON formidabilis Provancher, 1874 syn. n. groenlandicus Aurivillius, 1890 syn. n. alticola (Ashmead, 1901) syn. n.picf us Fabricius, 1798 cultratus (Gmelin, 1790) (homonym) mutillatus (Christ, 1791) (homonym) zagoriensis Hensch, 1928 bipunctatus Hensch, 1928 russiator Aubert, 1981 syn. n.poppitisp. n.punkettaisp. n.sanjozanus Uchida, 1929tholussp. n. tumidus Chandra & Gupta, 1977furcaf or Aubert, 1981volutatorius (Linnaeus, 1758) venator (Linnaeus, 1758) umbellatarum (Schrank, 1786) syn. n. certator (Thunberg, 1822) (homonym) monileatus Gravenhorst, 1829 syn. n. farrani Curtis, 1836 calcaratus Szepligeti, 1910 alticola Schmiedeknecht, 1910 (homonym) obscurus Meyer, 1926zonafusRudow, 1883 algericus Schmiedeknecht, 1910 syn. n. Species excluded from Banchus Andricus villosulus (Gravenhorst) (nomen dubium) comb. n.Banchus rofeusfusRudow, 1883 (nomen dubium)Cephas pygmeus (Linnaeus, 1767) spinipes (Panzer, 1800) vim/atar (Fabricius, 1804)Earinuse/af or (Fabricius, 1804) comb. n.Exetastes fornicator (Fabricius, 1781)Exetastes tomentosus (Gravenhorst, 1829)Ibalia leucospoides (Hochenwarth, 1785) cultellator (Fabricius, 1793) Ichneumon vigilatorius Panzer, 1804 (nomen dubium) comb. rev.Lissonota histrio (Fabricius, 1798)Megarhyssa quadrator (Schellenberg, 1802) comb. n.Phytodietus armillatus (Morley, 1913)Pompilus (Episyron) annulatus (Fabricius, 1793)Proctotrupes gravidator (Linnaeus, 1758)Theronia atalantae (Poda, 1761) varius (Fabricius, 1793) Format, material examined and depositories In the systematic section the species are arranged in alphabetical order. For each (exceptaltaiensis) information is under the following headings. Synonymy. When types have not been available for study (because of loss or destruction orbecause they could not be obtained on loan) the inclusion of a nominal species in a synonymy isbased on consideration of the original description and/or it is substantiated by reference to apublished treatment or is considered more fully under 'Nomenclature'. Nomenclature. This includes discussion of synonymies, type-restrictions, etc. It takes up aconsiderable amount of space (and it took up a lot of time) because the nomenclatural problems ICHNEUMON-FLY GENUS BANCHUS 9 associated with work on European species are related to human historical and sociologicalfactors rather than biological ones. Description. The descriptions for females and males are separate. The characters included aremainly those used in the keys, the ones which apply equally to both sexes are given for the femaleonly (except where the female is not known). The characters, including details of measure-ments, are discussed in the 'Taxonomy' section above. Remarks. These include brief notes on recognition and relationships of the species anddiscussion of its taxonomy. Biology and hosts. This is a brief summary based on data associated with material examinedand information gleaned from the literature. Names of hosts have been up-dated to conformwith current usage. Distribution. As given this is based entirely on data taken from specimens examined, as arethe distribution maps. Map 1 shows the distribution of the genus (that is, the combineddistributions of all the species) in the Old World for comparison with the distributions ofindividual species (Maps 2-11). The records in the literature are not reliable and only a fewreferences are made to them (Aubert (1978) should be consulted for more details of these). Material examined. Except for new species and those known from only a few specimens, thedata under 'material examined' are restricted to totals of specimens, the names of countries (and Map 1 Distribution of Banchus in the Old World. 10 M. G. FITTON larger islands) from which they come and the depositories in which they are to be found. Allspecimens on which this revision is based bear my determination labels. Some detail of collectinglocalities is presented in the distribution maps for each species. However, a feature of muchEuropean material is inadequate locality data on labels and there are no points on the maps forperhaps a third of all the specimens examined. Dates of collection are also often lacking; forinstance, in an attempt to relate morphological to seasonal variation in Banchus pictus only 90 ofthe 184 specimens to hand could be utilized because the rest had no date of collection.The names of depositories are abbreviated as in the list below. AC Collection of J. Aubert, Paris, France ANS Academy of Natural Sciences, Philadelphia, U.S.A. BC Collection of R. Bauer, Grossschwarzenlohe, B.R.D. BMNH British Museum (Natural History) BRI Biosystematics Research Institute, Ottawa, Canada CM Castle Museum, Norwich, England EIHU Entomological Institute, Hokkaido University, Sapporo, Japan FSA Faculte des Sciences Agronomique de 1'Etat, Gembloux, Belgium GC Collection of V . Gupta , University of Florida , Gainsville , U . S . A . HC Collection of R. Hinz, Einbeck, B.R.D. IBMPP Ail-Union Institute of Biological Methods of Plant Protection, Kishinev, U.S.S.R. IEAU Istituto di Entomologia Agraria dell'Universita, Sassari, Italy IEE Instituto Espanol de Entomologia, Madrid, Spain IEUB Istituto di Entomologia, Universita degli studi di Bologna, Bologna, Italy IP Institut fur Pflanzenschutzforschung, Eberswalde, D.D.R. IRSNB Institut Royal des Sciences Naturelles de Belgique, Brussels, Belgium ITZ Institut voor Taxonomische Zoologie, Zoologisch Museum, Amsterdam, Netherlands IZPAN Instytut Zoologiczny, Polska Akademia Nauk, Warsaw, Poland JC Collection of R. Jussila, Paattinen, Finland JKC Collection of J . Kolarov , Sadovo , Bulgaria JPM Jena Phyletisches Museum , Jena , D . D . R . KC Collection of K. Kusigemati, Kagoshima, Japan KHC Collection of K. Horstmann, Wiirzburg, B.R.D. LELW Laboratorium voor Entomologie van de Landbouwhogeschool, Wageningen, Netherlands LSL Linnean Society, London, England MCSN Museo Civico di Storia Naturale, Genoa, Italy MHN Museum d'Histoire Naturelle, Geneva, Switzerland MIZS Museo ed Istituto di Zoologia Sistematica, Turin, Italy MLSU Zoological Museum, Moscow Lomonosov State University, Moscow, U.S.S.R. MNHN Museum National d'Histoire Naturelle, Paris, France MNHU Museum fur Naturkunde der Humboldt-Universitat , Berlin , D.D.R. MUM Manchester University Museum, Manchester, England NC Collection of A. Nakanishi, Fukuoka, Japan NM Naturhistorisches Museum, Vienna, Austria NMB Naturhistorisches Museum, Basle, Switzerland NMV National Museum of Victoria, Melbourne, Australia NR Naturhistoriska Riksmuseet, Stockholm, Sweden PC Collection of C. Pisica, lasi, Rumania RNH Rijksmuseum van Natuurlijke Historic, Leiden, Netherlands RSM Royal Scottish Museum, Edinburgh, Scotland SC Collection of H. Schnee, Markkleeberg, D.D.R. SMT Staatliches Museum fur Tierkunde , Dresden , D . D . R . TAU Department of Zoology, Tel- Aviv University, Tel- Aviv, Israel TC Collection of H. & M. Townes, American Entomological Institute, Ann Arbor, U.S.A. TM Termeszettudomanyi Muzeum, Budapest, Hungary UL Universite Laval, Quebec, Canada UM University Museum, Oxford, England USNM U.S. National Museum of Natural History, Washington, U.S.A. UU Department of Entomology, University of Uppsala, Uppsala, Sweden UZI Universitetets Zoologiska Institutionen, Lund, Sweden ICHNEUMON-FLY GENUS BANCHUS 11 UZM Universitetets Zoologisk Museum, Copenhagen, Denmark VRC Collection of G. van Rossem, Ede, Netherlands ZC Collection of K. Zwart, Wageningen, Netherlands ZI Zoological Institute, Wroclaw, Poland ZIL Zoological Institute, Leningrad, U.S.S.R. ZIM Zoological Institute and Museum, Sofia, Bulgaria ZMU Zoological Museum of the University, Helsinki, Finland ZPZ Zavod Za Poljoprivednu Zoologiju, Zagreb, Yugoslavia ZSBS Zoologische Sammlung des Bayerischen Staates, Munich, B.R.D. Keys to species The keys to females and males are separate. Sexing specimens of Banchus is not alwaysstraightforward. The short, broad ovipositor sheaths are sometimes mistaken for the malegenitalia. The most reliable way to sex specimens is to examine the last visible tergites andsternites of the gaster: in females they are relatively long (Figs 6-10) and in males they are short(Fig. 1). Males (the males of cerinus and insulanus are unknown) 1 Maxillary palp (Figs 104, 114) with segment 5 less than 0-2 as long as segment 4; segment 4 long and cylindrical , with only its extreme apex sharply expanded and flattened 2 - Maxillary palp (Figs 103, 105-113, 115-123) with segment 5 more than 0-4 as long as segment 4; segment 4 gradually and evenly widened and flattened from base to apex or more sharplyexpanded from near base 3 2 Hind femur less than 6-2 times as long as deep; entirely reddish yellow in colour. Gaster usually entirely black, the segments sometimes brownish or yellowish on their posterior margins,very rarely as conspicuous yellow bands on tergites 1, 2 and 3. Scutellar spine usually long(more than 0-8 as long as scutellum) (Fig. 47) palpalis Ruthe (p. 30) - Hind femur more than 6-5 times as long as deep; black, yellow proximally and distally and sometimes dorsally. Gaster with each segment black anteriorly and yellow posteriorly.Scutellar spine short (about 0-3 as long as scutellum) (Fig. 34) crefeldensis Ulbricht (p. 15) 3 Antennal flag setae well developed (Figs 126-129): upright, flattened, 2 or 3 (rarely 4) per segment, arising from a more or less well-defined and polished trough 4 - Antennal flag setae not or only poorly differentiated (as in Fig. 124, or at most as in Fig. 125 but shorter): at 60 degrees or less to the segment surface, not flattened, 2 per segment, neverarising from a trough 17 4 Antennal flag setae 3 (occasionally 4) per segment (Fig. 129) - Antennal flag setae 2 per segment (Figs 126-128) (very rarely a few segments with 3 setae) 5 Scutellar spine long (about 0-8 as long as scutellum) (Fig. 44) mauricettae sp. n. (p. 28) - Scutellar spine short (about 0-3 as long as scutellum) to absent (Figs 33, 37, 38) 6 6 Hind femur less than 5-3 times as long as deep, usually entirely reddish yellow in colour but can be darkened medially. Malar space more than 0-6 times basal width of mandible falcatorius (Fabricius) (p. 19) - Hind femur about 6 1 times as long as deep , yellow in colour with the median part black . Malar space about 0-5 times basal width of mandible agathae sp. n. (p. 14) 7 Tergite 1 of gaster at least 2-1 times as long as broad Tergite 1 of gaster at most 2-0 times as long as broad 11 8 Hind femur at least 7-0 times as long as deep. Tergite 3 of gaster with a crease separating laterotergite along its whole length gudrunaesp. n. (p. 23) Hind femur less than 6-5 times as long as deep. Tergite 3 of gaster with a crease separatinglaterotergite on only about its anterior 0-5 9 Meoscutum shining , with only weak microsculpture between punctures . Scutellar spine at most about 0-5 as long as scutellum (Fig. 39) ffavomaculatus (Cameron) (p. 23) - Mesoscutum not shining, with well-developed microsculpture between punctures. Scutellar spine at least about 0-7 as long as scutellum (Figs 46, 52) 10 Hind femur about 6-0 times as long as deep. Antennal flagellum black with underside ofsegment 1 and distal section reddish yellow. Segments of gaster each black anteriorly andyellow posteriorly , with j unction of the two areas reddish nox Morley (p . 30) 12 M. G. FITTON Hind femur at most 5-6 times as long as deep. Antennal flagellum reddish, darkened dorsally(especially first few segments and distal third). Segments of gaster each reddish with someblackish anteriorly, especially on tergites 1 and 2 sanjozanus Uchida (p. 37) 11 Maxillary palp (Figs 110, 116, 122) with segment 5 about 0-5 as long as segment 4 12 Maxillary palp (Figs 105, 113, 121 , 123) with segment 5 more than 0-7 as long as segment 4 14 12 Antennal flagellum black ventrally. Hind femur more than 6-2 times as long as deep. Gaster with tergite 1 at most 1-65 times as long as broad; tergite 3 with a crease separating the laterotergite on at least its anterior 0-8 japonicus (Ashmead) (p. 26) Antennal flagellum yellowish ventrally. Hind femur at most 6-1 times as long as deep. Gasterwith tergite 1 at least 1-60 times as long as broad; tergite 3 with a crease separating thelaterotergite on its anterior 0-5 or less 13 13 Hind femur reddish yellow, usually entirely, rarely darkened ventrally; at most 5-4 times as long as deep. Hind tibia and segment 1 of hind tarsus reddish yellow, the tibia and sometimes the tarsal segment distally blackish volutatorius (Linnaeus) (p. 40) Hind femur yellow, black-marked ventrally and internally; at least 5-6 times as long as deep.Hind tibia and segment 1 of hind tarsus blackish, the tibia partly reddish yellow ventrally poppiti sp. n. (p. 35) 14 Malar space at most 0-6 times basal width of mandible. Antennal flagellum with its proximal 0-6 orange turcator Aubert (p. 39) - Malar space more than 0-6 times basal width of mandible. Antennal flagellum with at least its dorsal surface entirely brownish or blackish 15 15 Antennal flagellum yellowish ventrally (and posterior margins of gastral tergites 1, 2 and 3 yellow) or, if antennal flagellum is black ventrally, the posterior margins of gastral tergites 1 ,2 and 3 yellowish brown or reddish brown. Hind coxa entirely black dilatatorius (Thunberg) (p. 17) - Antennal flagellum blackish ventrally. Posterior margins of gastral tergites 1, 2 and 3 yellow. Hind coxa black, usually with at least a dorsal yellow spot 16 16 Tergite 1 of gaster at most 1-6 times as long as broad. Hind femur at most 5-5 times as long as deep. Fore wing length at most 10-0 mm moppitisp. n. (p. 28) - Tergite 1 of gaster more than 1 6 times as long as broad . Hind femur more than 5 3 times as long as deep. Fore wing length more than 10-5 mm zonatus Rudow (p. 41) 17 Tergite 1 of gaster with a strong median swelling in front of level of spiracles, the anterior face of swelling at almost 90 degrees to its dorsal surface (Figs 77, 78). Scutellar spine long (at least 0-7 as long as scutellum) (Figs 53, 54) 18 - Tergite 1 of gaster at most weakly swollen at level of spiracles (Figs 64, 73, 75). Scutellar spine moderately long (about 0-6 as long as scutellum) to very small (Figs 41 , 49, 51) 19 18 Segment 4 of maxillary palp relatively weakly expanded (Fig. 119); uniformly reddish yellow in colour tholussp. n. (p. 37) Segment 4 of maxillary palp considerably expanded from the base (Fig. 120); bi-coloured tumidus Chandra & Gupta (p. 38) 19 Maxillary palp (Fig. 117) with segment 5 about 0-6 as long as segment 4; segment 4 considerably widened and flattened. Hind femur about 6-0 times as long as deep punkettai sp. n. (p. 36) Maxillary palp (Figs 109, 115) with segment 5 about 0-8 as long as segment 4; segment 4 slightlyor moderately widened and flattened. Hind femur less than 5-7 times as long as deep 20 20 Posterior part of propodeum entirely black. Hind coxa almost always entirely black. Segments of gaster each black, usually with the posterior margin brownish yellow. Segment 4 ofmaxillary palp relatively slightly widened (Fig. 109). Antennal flag setae poorly differenti-ated and very small (Fig. 124) hastator (Fabricius) (p. 24) Posterior part of propodeum almost always with yellow marks. Hind coxa black, often with ayellow dorsal patch. Segments of gaster each black anteriorly, broadly yellow posteriorly.Segment 4 of maxillary palp relatively more widened (Fig. 115). Antennal flag setae at about60 degrees to the segment surface, small (as in Fig. 125 but setae shorter) pictus Fabricius (p. 33) Females (the female of sanjozanus is not included in the key) 1 Maxillary palp (Figs 13, 24) with segment 5 at most about 0-5 as long as segment 4; segment 4 relatively very long and slender 2 Maxillary palp (Figs 11, 12, 14-23, 25-32) with segment 5 at least 0-7 as long as segment 4;segment 4 variable in proportions, but much less slender 3 ICHNEUMON-FLY GENUS BANCHUS 13 2 Hind femur less than 6-0 times as long as deep; reddish yellow in colour, sometimes dark- ened ventrally. Gaster usually entirely black, posterior edges of some segments may bebrownish. Scutellar spine long (usually more than 0-8 as long as scutellum) (Fig. 47) I ml pa Us Ruthe (p. 30) - Hind femur more than 6-3 times as long as deep; black and yellow (proximally and distally and sometimes dorsally). Gaster with each segment black anteriorly and yellow posteriorly.Scutellar spine short (about 0-3 as long as scutellum) (Fig. 34) crefeldensis Ulbricht (p. 15) 3 Posterior part of propodeum entirely black 4 - Posterior part of propodeum with at least some yellow or reddish marks 9 4 Malar space at least 0-8 times basal width of mandible. Width of lower face at least 1-1 times vertical length of eye. Hind femur mainly black dilatator ius (Thunberg) (p. 17) - Malar space at most 0-7 times basal width of mandible. Width of lower face at most 1-0 times vertical length of eye. Hind femur entirely reddish yellow or with more or less extensiveblack areas 5 5 Tergite 7 of gaster (Fig. 85) elongate, subacute, its upper surface rounded posteriorly falcatorius (Fabricius) (p. 19) - Tergite 7 of gaster (Figs 88, 89, 95, 101) of normal length, subtruncate, its upper surface not markedly rounded posteriorly 6 6 Hind femur more than 6-0 times as long as deep. Tergite 3 of gaster with a crease separating laterotergite along its whole length insulanus Roman (p. 26) - Hind femur at most 5-5 times as long as deep. Tergite 3 of gaster with a crease separating laterotergite on less than its anterior 0-5 7 7 Tergite 1 of gaster at least 1-8 times as long as broad. Hind femur entirely reddish yellow volutatorius (Linnaeus) (p. 40) - Tergite 1 of gaster at most 1-7 times as long as broad. Hind femur reddish yellow with more or less extensive black areas 8 8 Tergites of gaster each black anteriorly, broadly yellow posteriorly, the boundary between the two areas sharply defined pictus Fabricius (p. 33) - Tergites of gaster each black, grading posteriorly, to a greater or lesser extent, into a brownish and sometimes yellowish marginal area hastator (Fabricius) (p. 24) 9 Tergite 7 of gaster (Figs 84, 92, 102) relatively elongate, subacute, its upper surface rounded posteriorly AND width of lower face at least 1-0 times vertical length of eye 10 - Tergite 7 of gaster (Figs 6, 82, 86, 87, 90, 91, 93, 95-100) of normal length, subtruncate, its upper surface not markedly rounded posteriorly AND/OR width of lower face at most 0-9 timesvertical length of eye 12 10 Antennal flagellum yellowish orange in colour, with the base of segment 1 and the distal segmentsdark dilatatorius (Thunberg) (p. 17) - Antennal flagellum black, brownish distally 11 11 Tergite 7 of gaster (Fig. 102) more elongate. Hind femur at least 5-3 times as long as deep. Fore wing length at least 11-0 mm zonatus Rudow (p. 41) - Tergite 7 of gaster (Fig. 92) less elongate. Hind femur at most 5-2 times as long as deep. Fore wing length at most 10-8 mm moppiti sp. n. (p. 28) 12 Tergite 1 of gaster with a strong median swelling in front of level of spiracles, the anterior face of the swelling at almost 90 degrees to its dorsal surface (Figs 77, 78) tholussp. n. (p. 37) and tumidus (Chandra & Gupta) (p. 38) (It has proved impossible to separate reliably the females of these two species. See notesunder tumidus.} - Tergite 1 of gaster at most weakly swollen in front of level of the spiracles (Figs 6 , 59 , 63 , 65 , 66 , 68,69,71,73,74,75,79) 13 13 Antennal flagellum entirely black. Hind femur largely black (yellow or reddish proximally and distally) or yellow with a black area medially (which extends onto its dorsal surface) 14 - Antennal flagellum reddish, orange, or yellow, at least on its proximal 0-5, or dark dorsally (black or brownish) and paler ventrally (yellow or reddish). If antennal flagellum is dark(blackish or brownish dorsally and dark reddish ventrally) then hind femur reddish with ablack ventral mark, which does not reach to dorsal surface 16 14 Tergite 3 of gaster with a crease separating laterotergite along its whole length. Hind femur more than 6-6 times as long as deep. Hind tibia yellow, with its distal 0-2 blackish gudrunae sp. n. (p. 23) - Tergite 3 of gaster with a crease separating laterotergite on only its anterior 0-5. Hind femur at 14 M. G. FITTON most 6-5 times as long as deep. Hind tibia yellow with its distal 0-2 blackish or black with amedian yellowish band 15 15 Hind tibia yellow with its distal 0-2 black. Hind femur about 5-4 times as long as deep agathae sp. n. (p. 14)Hind tibia black with a median yellowish band. Hind femur more than 5-5 times as long as deep japonicus (Ashmead) (p. 26) 16 Hind femur entirely yellow or entirely reddish 17 Hind femur black and yellow or black and reddish , at least reddish with a ventral black mark .... 18 17 Hind femur more than 6-0 times as long as deep. Tergite 1 of gaster more than 2-5 times as long as broad cerinus Chandra & Gupta (p. 15) Hind femur less than 5-5 times as long as deep. Tergite 1 of gaster less than 2-5 times as long asbroad mauricettae sp. n. (p. 28) 18 Tergites of gaster each black anteriorly, broadly to narrowly yellow posteriorly, the boundary between the two areas sharply defined 19 - Tergites of gaster (at least 1, 2 and 3) each black anteriorly, reddish or yellow posteriorly, if yellow with an intermediate reddish band grading into the black and into the yellow 21 19 Hind tibia and tarsus mainly blackish, tibia and segment 1 of tarsus slightly lighter proximally. Tergites of gaster each only narrowly yellow posteriorly poppiti sp. n. (p. 35) - Hind tibia, on at least its proximal 0-5, and tarsus yellow. Tergites of gaster each broadly to relatively narrowly yellow posteriorly 20 20 Gaster (about as in Figs 7,8) relatively less compressed posteriorly ; tergite 7 and sternite 6 (Fig . 100) subtruncate. Antennal flagellum entirely yellowish orange (at least its proximal 0-7),sometimes with a very small dark mark dorsally on segment 1 turcator Aubert (p . 39) - Gaster relatively more compressed posteriorly; tergite 7 and sternite 6 (Fig. 95) more acute. Antennal flagellum yellowish orange, sometimes with proximal section dark dorsally andalmost always with at least proximal 0-5 of segment 1 blackish dorsally . . . pictus Fabricius (p. 33) 21 Tergites of gaster each black anteriorly (at least 1, 2 and 3), reddish posteriorly. Tergite 1 of gaster at least 2-25 times as long as broad 22 Tergites of gaster each black anteriorly, yellow posteriorly and with an intermediate reddishband. Tergite 1 of gaster at most 2 -25 times as long as broad punkettai sp. n. (p. 36) 22 Posterior part of propodeum entirely reddish. Gaster with tergites 1 to 5 each black anteriorly ffavomaculatus (Cameron) (p. 23) - Posterior part of propodeum black, with an irregular reddish yellow band along posterior transverse carina. Gaster with only tergites 1 , 2 and 3 black anteriorly no* Morley (p. 30) Systematic section Hunch us agathae sp. n. (Figs 11, 33, 59, 82, 103)[Banchus pictus Fabricius; Bischoff, 1930: 225. Misidentification (and wrong sex).] DESCRIPTION. Female. Width of lower face 0-90 times vertical length of eye. Malar space 0-60 times basalwidth of mandible. Maxillary palp as in Fig. 11. Antenna with apical segments as long as broad. Scutellarspine (Fig. 33) about 0-3 as long as scutellum. Mesopleuron and mesoscutum shining, only very weaklysculptured, with strong punctures, on the mesopleuron separated by much less than their diameter. Hindfemur 5-40 times as long as deep. Fore wing length 10-1 mm. Tergite 1 of gaster 2-00 times as long as broad,its dorsal profile as in Fig. 59. Gaster compressed from the posterior of segment 3, reaching just beyond thetips of the fore wings (when folded back). Tergite 3 with a crease along its anterior 0-5 separatinglaterotergite. Tergite 7 and sternite 6 as in Fig. 82. Colour: black and yellow. Face yellow with a median black stripe. Antenna black with scape and pedicelyellow ventrally. Maxillary palp yellowish with dark marks, including distal 0-7 of segment 4 and all ofsegment 5. Posterior part of propodeum yellow, narrowly black posteriorly. Hind coxa black with a yellowdorsal patch and distal margin narrowly yellow. Hind femur black, yellow proximally and distally.Segments of gaster each black anteriorly, very narrowly yellow posteriorly. Male. Width of lower face 0-90 times vertical length of eye. Malar space 0-45 times basal width ofmandible. Maxillary palp as in Fig. 103; segment 4 flattened and considerably widened; segment 5 about0-7 as long as 4. Antennal flag setae (similar to Fig. 129, but the setae only about half as long) upright, short,flattened, 3 per segment, arising from a polished trough. Hind femur 6-10 times as long as deep. Fore wing ICHNEUMON-FLY GENUS BANCHUS 15 length 8-5 mm. Tergite 1 of gaster 1-80 times as long as broad. Tergite 3 with a crease along its anterior 0-5separating the laterotergite. Colour: black and yellow. Antenna black with scape, pedicel and proximal and distal parts of flagellumyellow ventrally. Maxillary palp with segments 1 yellowish, 2 and 3 reddish, 4 and 5 blackish. Posterior partof propodeum yellow, narrowly black posteriorly. Hind coxa black with a yellow dorsal spot and narrowpostero-ventral margin. Hind femur yellow, black medially. Segments of gaster each black anteriorly,yellow posteriorly. REMARKS. The male should be readily recognized by the form of the antennal flag setae. The female iskeyed to a large extent on colour characters but the form of the posterior part of the gaster is alsoimportant. BIOLOGY AND HOSTS. Unknown. The specimens were taken in July and August.DISTRIBUTION (Map 2, p. 16). Known from only two specimens collected in the Pamir. MATERIAL EXAMINED Holotype cf , U.S.S.R.: Pamir, Umss-Tugai, 25.vii.1928 (Rickmers) (ZSBS).Paratype $, U.S.S.R.: Pamir, Maz, 3580 m, 15.viii.1928 (Reinig) (MNHU). Banchus altaiensis Meyer nomen dubiumBanchus altaiensis Meyer, 19276: 295. Type(s) cf , U.S.S.R. (destroyed). This species was described from material, probably only one specimen, collected in the Altai mountains.The description does not fit adequately any of the species from Central Asia known to me and I hesitate toidentify it with any species not proven to occur there. Its identification will have to wait until more extensivecollections have been made in the area. Banchus cerinus Chandra & Gupta (Figs 6, 12)Banchus cerinus Chandra & Gupta, 1977: 185. Holotype $, INDIA (GC) [examined]. DESCRIPTION. Female. Width of lower face 0-85 times vertical length of eye. Maxillary palp as in Fig. 12.Antenna with apical segments longer than broad. Scutellar spine of moderate length. Mesopleuron andmesoscutum weakly coriaceous, with dense, fairly small punctures. Hind femur 6-25 times as long as deep.Fore wing length 14-3 mm. Tergite 1 of gaster 2-75 times as long as broad, its dorsal profile as in Fig. 6.Gaster subcylindrical, only weakly compressed posteriorly, not reaching to tips of fore wings (when foldedback). Tergite 3 with a crease along its anterior 0-6 separating laterotergite. Tergite 7 and sternite 6 as inFig. 6. Colour: mainly yellow, with black marks. Face yellow with a black mark between and just belowantennal sockets. Antenna yellow with distal half of flagellum and an externo-lateral mark on pedicelblackish. Maxillary palp yellowish. Posterior part of propodeum yellow. Hind coxa yellow, blackishdistally. Hind femur yellow. Segments of gaster yellow, with basal third of tergites 1,2 and 3 and of sternite2 black. Male. Unknown. REMARKS. A pale-coloured species with relatively long appendages. Its closest relatives seem to be tholusand tumidus. DISTRIBUTION (Map 4, p. 20), BIOLOGY AND HOSTS. The species is known only from the holotype, which wascollected amongst mixed bushes in coniferous forest in the NW. Himalayas (Gupta, 1975: (appendix) 50). MATERIAL EXAMINED (1 $)India: 1 $ (holotype), Himachal Pradesh, Ahla, 2286 m, 18.vii.1971 (Gulati) (GC). Banchus crefeldensis Ulbricht(Figs 4, 13,34,60,83,104, 125) Banchus crefeldensis Ulbricht, 1916: 12. Holotype cf, WEST GERMANY (A. Ulbricht coll., Krefeld) [examined].Banchus croaticus Hensch, 1928: 99. Lectotype $, YUGOSLAVIA (ZPZ), designated by Horstmann, 1982a: 82 [examined]. 16 M. G. FITTON Map 2 Distribution of Banchus agathae, B. japonicus, B. mauricettae, B. nox, B. sanjozanus, B. tholusand B. tumidus. DESCRIPTION. Female. Width of lower face 0-90-1-00 times vertical length of eye. Malar space 0-60-0-75times basal width of mandible. Maxillary palp as in Fig. 13. Antenna with apical segments longer thanbroad. Scutellar spine (Fig. 34) about 0-3 as long as scutellum. Mesopleuron and mesoscutum stronglycoriaceous, with moderately strong punctures, on mesopleuron separated by about their diameter. Hindfemur 6-35-7-50 times as long as deep. Fore wing length 9-9-10-7 mm. Tergite 1 of gaster 1-80-2-00 times aslong as broad, its dorsal profile as in Fig. 60 (male). Gaster compressed from segment 4, reaching about totips of fore wings (when folded back). Tergite 3 with a crease along its entire length separatinglaterotergite. Tergite 7 and sternite 6 as in Fig. 83. ICHNEUMON-FLY GENUS BANCHUS 17 Colour: black and yellow. Face black with broad yellow orbital stripes and a vertical yellow mark beloweach antennal socket. Antenna entirely black except for a yellow patch on underside of scape, and rarelyalso on pedicel. Maxillary palp blackish with part or all of segment 2 and proximal 0-7 of segments 3 and 4reddish yellow. Posterior part of propodeum yellow anteriorly, black posteriorly. Hind coxa black with ayellow dorsal patch. Hind femur black, yellow proximally and distally and sometimes dorsally. Segments ofgaster each black anteriorly and broadly yellow posteriorly. Male. Width of lower face 0-90-1-00 times vertical length of eye. Malar space 0-55-0-70 times basalwidth of mandible. Maxillary palp as in Fig. 104; segment 4 narrow and cylindrical, flattened and widenedonly at its extreme distal apex; segment 5 less than 0-2 as long as 4. Antennal flag setae (Fig. 125) at about 50degrees, not flattened, 2 or 3 per segment, not arising from a trough. Hind femur 6-60-7-65 times as long asdeep. Fore wing length 9-1-10-8 mm. Tergite 1 of gaster 1-80-2-10 times as long as broad (Fig. 4). Tergite 3with a crease along its entire length separating laterotergite. Colour: black and yellow. Antenna black, with scape and pedicel yellow ventrally. Maxillary palpblackish with segments 2 wholly and 3 and 4 proximally reddish yellow. Posterior part of propodeum blackwith a variable amount of yellow anteriorly, ranging from a very broad band along posterior transversecarina to a small median spot to absent entirely. Hind coxa black, usually with a yellow dorsal patch. Hindfemur black, yellow proximally and distally and sometimes dorsally. Segments of gaster each blackanteriorly, yellow posteriorly (on at least tergites 2 and 3 relatively broadly so). REMARKS. Like the related B. palpalis, this is a very distinctive species; it is readily recognized by thestructure of the maxillary palps and its coloration. BIOLOGY AND HOSTS. Dates of collection range from mid-February (in Spain and Portugal) to mid-May (inScotland and Ireland). Stelfox (1936: 63) records males flying in numbers round ivy (Hedera helix) andgorse (Ulex europaeus) in Ireland. Specimen labels indicate capture of males on Betula and around Buxusand females from pine. I have seen 11 reared specimens, 3 without a satisfactory host identification, the remaining 8 (includingboth sexes) fromAporophyla lutulenta (Denis & Schiffermuller) or A. lutulenta 'subspecies' lueneburgensis(Freyer) (Noctuidae). The named hosts were collected at Bussum, Netherlands; Aviemore, Great Britain;and Bremen and near Hamburg, West Germany. DISTRIBUTION (Map 3, p. 18). Widely distributed in Europe but not found in Scandinavia or the U.S.S.R.In Great Britain it is restricted to the Scottish Highlands. MATERIAL EXAMINED (40 $ , 79 cf ) Austria, Belgium, France, Great Britain, Hungary, Ireland, Italy, Netherlands, Portugal, Spain,Switzerland, Tunisia, Turkey, West Germany, Yugoslavia (BMNH, FSA, IEE, IRSNB, LELW, MHN,MNHN, NM, NMB, NR, RNH, UM, USNM, ZC, ZSBS). Banchus dilatatorius (Thunberg) stat. rev.(Figs 2, 14, 35, 36, 61, 84, 105) Ichneumon acuminator Fabricius, 1787: 268. Type(s) cf , EAST GERMANY (lost). [Junior primary homonym of Ichneumon acuminator Miiller, 1776: 157.] Syn. n.Ichneumon compressus Fabricius, 1787: 381. Holotype $, SWEDEN (UZM) [examined]. [Junior primary homonym of Ichneumon compressus Sulzer, 1776: 190.] Ichneumon dilatatorius Thunberg, 1822: 279; 1824: 360. Holotype $, GERMANY (UU) [not examined].Banchus sibiricus Meyer, 19276: 294. Syntypes $, cf , U.S.S.R. (destroyed). Syn. n. NOMENCLATURE. In most of the literature this species is referred to by the invalid name compressus. Although Ichneumon acuminator has been placed previously in Banchus it has never been identified.The description fits certain males of dilatatorius if one allows that Fabricius failed to observe the pale colourof the underside of the antennae. The description of Banchus sibiricus fits well specimens from Kamchatka (which include those identifiedby Roman (1931: 29) as lavrovi var.) which are considered conspecific with dilatatorius but are discussedunder 'Remarks' below. The identity of the female from Rumania determined as sibiricus by Constan-tineanu & Pisica (1959: 190) is not known. DESCRIPTION. Female. Width of lower face 1-15-1-25 times vertical length of eye. Malar space 0-80-1-00times basal width of mandible (Fig. 2). Maxillary palp as in Fig. 14. Antenna with apical segments about asbroad as long. Scutellar spine (Figs 35, 36) very small or virtually absent. Mesopleuron and mesoscutumcoriaceous, with strong punctures, on mesopleuron separated by less than diameter. Hind femur 4-35-4-85 M. G. FITTON o * crefeldensis Map 3 Distribution of Banchus crefeldensis. times as long as deep. Fore wing length 8-7-11-5 mm. Tergite 1 of gaster 1-30-1-40 times as long as broad,its dorsal profile as in Fig. 61. Gaster strongly compressed from posterior of tergite 3, reaching beyond tipsof fore wings (when folded back). Tergite 3 with a crease along its anterior 0-3 separating laterotergite.Tergite 7 and sternite 6 as in Fig. 84. Colour: black and yellow (the yellow often rather creamy). Face black with yellow orbital marks.Antenna with scape and pedicel black, scape, and usually pedicel, yellowish or reddish ventrally.Flagellum yellowish orange, with base of segment 1 and distal segments dark. Maxillary palp blackish,entirely or with segments 2 and 3 brownish. Posterior part of propodeum black, entirely or, more usually,with a yellow band along posterior transverse carina, the band usually not extensive and narrowly absentmedially. Hind coxa black. Hind femur black, narrowly yellowish proximally and usually widely yellowishdistally. Segments of gaster each black anteriorly, yellow posteriorly, the yellow bands usually less than 0-3of length of segments. Male. Width of lower face 1-10-1-25 times vertical length of eye. Malar space 0-85-0-95 times basalwidth of mandible. Maxillary palp as in Fig. 105; segment 4 only slightly flattened and widened distally;segment 5 about 0-9 as long as 4. Antennal flag setae (similar to Fig. 126) upright, flattened, relativelynarrow, 2 per segment, arising from a polished trough. Hind femur 4-35-5-00 times as long as deep. Forewing length 8-6-10-7 mm. Tergite 1 of gaster 1-25-1-55 times as long as broad. Tergite 3 with a crease alongits anterior 0-3 separating laterotergite. Colour: black and yellow. Antenna blackish or brownish dorsally, yellow or yellowish ventrally (exceptfor apex of flagellum which is dark). Maxillary palp blackish or brownish. Posterior part of propodeumblack, entirely or with a yellow band (sometimes broken medially) along posterior transverse carina. Hindcoxa black. Hind femur yellow with a black mark on interne-lateral and ventral surface, and sometimesalso dorsally near base. Segments of gaster each black anteriorly, yellow posteriorly, the yellow bandsusually less than 0-3 of length of segment. ICHNEUMON-FLY GENUS BANCHUS 19 REMARKS. This species is distinctive and easily identified. The very wide face is a conspicuous feature.However, the poor characters used in earlier keys led to confusion with other species in several of thecollections examined. The details of colour given in the description above do not apply fully to all of the material examined.One female from Iran (AC) has the yellow coloration much more extensive than in any other specimensseen. All the specimens from Kamchatka (1 $ (MLSU) and 4 $ , 2 0" (NR)) have yellow absent but withreddish in place of the yellow on the legs and on the face of the male, and with the margins of some or all oftergites 1, 2 and 3 of the gaster narrowly to broadly reddish brown. Other than in these colour charactersthe Iranian and Kamchatka specimens do not differ much from 'typical' dilatatorius . It is impossible todecide whether or not they represent extremes of colour variation or distinct species. Unfortunately thecharacter which might enable a decision to be reached, the male flag setae, cannot be investigated. Of thetwo male specimens one lacks the head completely and the other lacks both antennae. The Kamchatkaspecimens match the description of B. sibiricus which was described from Irkutsk (see 'Nomenclature'above). The material from Mongolia identified by Momoi (1973: 242-243) as lavrovi might also be this darkform of dilatatorius. I could not examine these Mongolian specimens because Momoi has so far failed toreturn them to Budapest (Zombori, pers. comm.). The six specimens in Stockholm (NR) had beenidentified by Roman (1931: 29) as lavrovi var. BIOLOGY AND HOSTS. This is an early spring species with most records relating to the period from late Marchto late April. The earliest dates of collection (7 and 13 March) come from sand dune areas in theNetherlands. The label on one female records that it was found on catkins of Salix caprea. I have seen noreared specimens of this species, although the following hosts are recorded in the literature: Blepharitaadusta (Esper) (Noctuidae) (Bajari, 1960: 260); Euxoa nigricans (Linnaeus) (Noctuidae) (Schmiedek-necht, 1910: 1926; Meyer, 1934: 228); Phlogophora meticulosa (Linnaeus) (Noctuidae) (Gyorfi, 1944:106); Euproctis similis (Fuessly) (Lymantriidae) (de Gaulle, 1907: 19); and Leucoma salicis (Linnaeus)(Lymantriidae) (Leonardi, 1928: 83). DISTRIBUTION (Map 4, p. 20). Western and central Europe, mainly north of 50 degrees N. (but with veryfew records from the area bordering the Atlantic (France, the British Isles and Norway) and only a fewfrom Denmark and Sweden); the European Alps; and with some scattered records in the westernU.S.S.R. , north-west Iran and the Soviet Far East (see 'Remarks' above). MATERIAL EXAMINED (86 9 , 96 cf) Austria, Belgium, Czechoslovakia, Denmark, East Germany, Finland, France, Great Britain, Hungary,Iran, Italy, Netherlands, Sweden, Switzerland, U.S.S.R., West Germany (AC, BMNH, BRI, HC, IP,IRSNB, ITZ, LELW, MNHN, MHN, MIZS, NMB, NR, RNH, SMT, USNM, UZI, UZM, ZC, ZIL,ZMU, ZSBS). Banchus falcatorius (Fabricius)(Figs 5, 9, 10, 15, 37, 38, 62, 85, 106, 129) Ichneumon falcatorius Fabricius, 1775: 332. Holotype d", DENMARK (UZM) [examined]. Ichneumon variegator Fabricius, 1775: 339. LECTOTYPE d", SWEDEN (UZM), here designated [examined] . Ichneumon intersectus Geoffroy, 1785: 414. Type(s) cf , FRANCE (lost) (Horstmann, 19826: 243).Ichneumon aries Christ, 1791: 339. Type(s) $, no type-locality (lost). Ichneumon labiatus Schrank, 1802: 264. Syntypes $ [not cf as stated by Schrank], WEST GERMANY (lost).Ichneumon histrio Schrank, 1802: 265. Syntypes cf , WEST GERMANY and FRANCE (lost) (Horstmann, 19826: 243). [Junior primary homonym of Ichneumon histrio Christ, 1791: 356.]Ichneumon tricolor Schrank, 1802: 286. Syntypes cf , WEST GERMANY (lost). Banchus falcator Fabricius, 1804: 128. [Unjustified emendation of Ichneumon falcatorius Fabricius, 1775.]CorynephanessachalinensisMatsumura, 1911: 92. Lectotype d" [not $ as stated by Matsumura] , U.S.S.R. (EIHU), designated by Townes, Momoi & Townes, 1965: 236 [examined].Banchus falcatorius var. luteofasciatus Ulbricht, 1911: 151. Type(s) , HUNGARY (?lost).Banchus nobilitator Morley, 1915: 138. Holotype $, U.S.S.R. (BMNH) [examined].Banchus falcatorius var. sanguinator Meyer, 1922: 139. Holotype d", U.S.S.R. (destroyed).Banchus lavrovi Meyer, 19276: 294. Syntypes $, cf, U.S.S.R. (destroyed).Banchus falcatorius var. nigromarginatus Constantineanu & Pisica, 1960: 710. Syntypes 3 $, RUMANIA (?Constantineanu coll.) [not examined].Banchus propitius Kuslitzky, 1979: 351. Holotype d", MONGOLIA (ZIL) [not examined]. [Provisional synonymy.] M. G. FITTON Map 4 Distribution of Banchus cerinus, B. dilatatorius and B. turcator. NOMENCLATURE. Ichneumon labiatus was listed by Dalla Torre (1901: 63) in the synonymy of compressus(= dilatatorius). However, as noted by Gravenhorst (1829: 390) the description applies to B. falcatorius.Schrank misdetermined the sex of his species. Ichneumon tricolor has been synonymised usually with compressus (= dilatatorius) (e.g. Aubert, 1978:153). The description and the figure to which Schrank refers (Schaeffer, 1768: pi. 116, fig. 5), differ fromthat species in the coloration of the gaster. However, they fit the male of B. falcatorius and there can belittle doubt about the identity of the two species. This synonymy was first proposed by Gravenhorst (1829:396). The synonymy of sanguinator might be questioned. I have seen no male of B. falcatorius as red asdescribed by Meyer; the most reasonable explanation of the colour is that the specimen had suffered fromoverlong exposure to cyanide in a killing jar. The description of lavrovi fits specimens from the northern part of the eastern Palaearctic and hereincluded in falcatorius . One such specimen determined as lavrovi by Meyer himself has been examined.The oldest name applying to this form of falcatorius is nobilitator Morley. In terms of falcatorius as recognized in this revision (see 'Remarks' below) I have thought it bestprovisionally to synonymize the species, propitius, recently described by Kuslitzky (1979). DESCRIPTION. Female. Width of lower face 0-95-1-00 times vertical length of eye. Malar space 0-60-0-70times basal width of mandible. Maxillary palp as in Fig. 15. Antenna with apical segments just longer thanbroad. Scutellar spine (Fig. 37) minute or absent (but see 'Remarks'). Mesopleuron and mesoscutum ICHNEUMON-FLY GENUS BANCHUS 21 coriaceous, with moderately strong punctures, on mesopleuron separated by less than their diameter. Hindfemur 4-25^4-70 times as long as deep. Fore wing length 10-1-11-2 mm. Tergite 1 of gaster 1-45-1-70 timesas long as broad, its dorsal profile as in Fig. 62. Gaster (Figs 9, 10) strongly compressed, reaching beyondtips of fore wings (when folded back). Tergite 3 with a crease along its anterior 0-3 separating laterotergite.Tergite 7 and sternite 6 as in Fig. 85. Colour: mainly black, with legs largely reddish and usually some reddish and very rarely yellow on thegaster. Face black, often with very small reddish or yellowish orbital spots. Antenna dark brown orblackish dorsally, brown or reddish ventrally. Maxillary palp reddish with segments 4 apically and 5 whollydark. Posterior part of propodeum entirely black. Hind coxa black, very rarely with a postero-ventralreddish spot. Hind femur reddish yellow, entirely or, in a few specimens, blackish medially. Segments ofgaster black, rarely entirely so, usually with variable reddish marks on segments 1,2 and 3 and more rarelyon other segments. Very rarely reddish areas grade to yellow posteriorly. Male. Width of lower face 0-95-1-10 times vertical length of eye. Malar space 0-65-0-75 times basalwidth of mandible. Maxillary palp as in Fig. 106; segment 4 considerably widened and flattened; segment 5about 0-5 as long as 4. Antennal flag setae (Fig. 129) upright, long, flattened and considerably widened, 3(sometimes 4) per segment, arising from a polished trough. Hind femur (Fig. 5) 4-45-5-25 times as long asdeep. Fore wing length 9-2-11-7 mm. Tergite 1 of gaster 1-40-1-70 times as long as broad. Tergite 3 with acrease along its anterior 0-3 separating the laterotergite. Colour: black, yellow and reddish. Antenna black dorsally (and ventrally at distal apex of flagellum),yellow ventrally. Maxillary palp with segment 1 yellow, segments 2 and 3 wholly and 4 proximally reddishyellow, segment 4 distally and 5 wholly brown or blackish. Posterior part of propodeum yellow anteriorly,black posteriorly, the yellow sometimes reduced to a pair of latero-median spots or entirely absent. Hindcoxa black, usually with a yellowish or reddish patch postero-ventrally and sometimes also dorsally. Hindfemur reddish yellow, almost always entirely but in a few specimens blackish medially. Gaster with tergite 1black anteriorly, reddish and then yellow posteriorly; tergites 2 and 3 each reddish anteriorly, yellowposteriorly (with some blackish laterally); remaining tergites black, 4, 5 and 6 each with a postero-medianyellow spot. REMARKS. B. falcatorius has several specialized features and it is the species in which sexual dimorphism ismost marked. There is also some geographical variation in characters; for instance, specimens from theeastern part of the range have the hind femur darkened medially and a few specimens from Turkey have awell-developed spine on the scutellum. More than one species might be included in the taxon hererecognized, but in the limited material from the eastern Palaearctic available for study no clear-cutsegregates could be recognized. Although the sex association in this species is not in doubt, interesting confirmation comes from thediscovery of four gynandromorph specimens in the extensive material examined. These were the onlygynandromorphs found (or at least recognized) during the present study. Details of the specimens are asfollows. Italy: Piemonte, Susa, 24. vi. 1872 (Gribodo) (MCSN). Head male (unfortunately the antennal flagellaare missing), thorax and abdomen female. Italy: Cadore, Valle del Boite (IEUB). Head female, thoraxand propodeum apparently mainly female but with yellow marks on the left of the mesoscutum andscutellum (male characters), gaster apparently male (including the genitalia) but with some bilateralasymmetry in colour anteriorly. ?locality [illegible] (Giraud coll.) (MNHN). Head (including antennaeand palps) male on the left and female on the right, prothorax and anterior of mesothorax mainly femalebut partly male on left, posterior of mesothorax and remainder of body female. Interestingly this specimenhad been labelled (anonymously) as a hybrid between falcatorius and monileatus [= palpalis}. Denmark:Ordrup, 19.viii. 1877 (Drewsen) (UZM). Head (including appendages) female, remainder male except thatthe dark colour of the left fore and mid coxae suggests some female influence. BIOLOGY AND HOSTS. This species is on the wing from early June to mid August (with a few records for lateMay and late August). The adults have often been collected from umbellifer flowers. One female islabelled as being the prey of Dasypogon teutonics (Diptera; Asilidae). Despite the abundance of specimens in collections there are very few reared examples, and not all ofthose have an attached host name or host remains. The most reliable host data relate to Agrotis segetum(Denis & Schiffermiiller) (Noctuidae) and this is certainly an important host species. It is (or was) a majorpest of root crops (particularly various beets and carrots) and its association with B. falcatorius was thesubject of research in the U.S.S.R. in the 1920s and 30s (Kosobutzkii, 1928; Meyer, 1927a, 1928; Pospelov,1924; Samoilova, 1936). The long, knife-like gaster of the female is undoubtedly adapted to enable it toreach the host larvae, which feed at soil level. Other hosts recorded on specimen labels are: Deilephilaporcellus (Linnaeus) and D. elpenor (Linnaeus) (Sphingidae), Dasychira fascelina (Linnaeus) (Lyman- 22 M. G. FITTON triidae), Acronicta megacephala (Denis & Schiffermiiller) (Noctuidae) and Melitaea didyma (Esper)(Nymphalidae). The last named host record seems unlikely to be accurate, but it is difficult to judge theothers (all the specimens are without locality or date !). Other hosts recorded in the literature are: Agrotisexclamationis (Linnaeus) (Noctuidae) (Meyer, 1927a: 81) and, erroneously, two Hymenoptera (Diprionpini (Linnaeus) (Diprionidae) and Ophion luteus (Linnaeus) (Ichneumonidae) (Maneval, 1935: 74)). DISTRIBUTION (Map 5, p. 22). Most of the Palaearctic, but few specimens seen from large areas of theU.S.S.R. and further east than the Caucasus none from south of 50 degrees N. In the more southern partsof its range in the western Palaearctic it is apparently restricted to higher altitudes. MATERIAL EXAMINED (652 $ , 806 cf ) Austria, Belgium, Bulgaria, Corsica, Czechoslovakia, Denmark, East Germany, Finland, France, GreatBritain, Greece, Italy, Netherlands, Norway, Poland, Rumania, Sakhalin, Spain, Sweden, Switzerland,Turkey, U.S.S.R., West Germany, Yugoslavia (AC, BC, BMNH, BRI, CM, EIHU, FSA, HC, IEE,IEUB, IP, IRSNB, ITZ, IZPAN, JC, JKC, KHC, LELW, MCSN, MHN, MNHN, MNHU, MUM, NMB,NR, PC, RNH, SC, SMT, UM, USNM, UZI, UZM, VRC, ZIL, ZIM, ZMU, ZC, ZSBS). falcatorius T flavomaculatus Map 5 Distribution of Banchus falcatorius and B. flavomaculatus. ICHNEUMON-FLY GENUS BANCHUS 23 Banchus ffavomaculatus (Cameron)(Figs 16, 39, 63, 86, 107) Cidaphurusflavomaculatus Cameron, 1904: 346. Lectotype $ [not cf as stated by Cameron and by Morley,1913], INDIA (BMNH), fixed by Morley, 1913: 255. NOMENCLATURE. The lectotype and one paralectotype female (BMNH) are the only known syntypespecimens. Both bear labels 'Cidaphurus flavomaculatus Cam. Simla' in Cameron's handwriting. Thelectotype has, in addition, Nurse's typewritten data label 'Simla 5.97.'. It is unfortunate that Cameron'soriginal description applies better to the paralectotype because it is not conspecific with the lectotype and isB. punkettai. DESCRIPTION. Female. Width of lower face 0-75-0-80 times vertical length of eye. Malar space 0-50-0-55times basal width of mandible. Maxillary palp as in Fig. 16. Antenna with apical segments about as broad aslong. Scutellar spine (Fig. 39) varying from very short to about 0-5 as long as scutellum. Mesopleuron andmesoscutum shining, weakly coriaceous, with moderate punctures, on mesopleuron separated by about ora little more than their diameter. Hind femur 5-60-5-65 times as long as deep. Fore wing length 8-8-9-2mm. Tergite 1 of gaster 2-25-2-35 times as long as broad, its dorsal profile as in Fig. 63. Gaster stronglycompressed from posterior of segment 3, reaching to about tips of fore wings (when folded back). Tergite 3with a crease along its anterior 0-4 separating laterotergite. Tergite 7 and sternite 6 as in Fig. 86. Colour: black and reddish, with some yellow marks. Face reddish, yellowish laterally and with a blackmedian stripe. Antenna blackish dorsally, scape and pedicel yellowish ventrally, flagellum brownventrally. Maxillary palp reddish yellow, segments 4 distally and 5 wholly dark. Posterior part ofpropodeum entirely reddish. Hind coxa black with large dorsal and postero-ventral reddish patches. Hindfemur reddish with a black stripe ventrally. Segments of gaster black anteriorly, very broadly reddishposteriorly. Male. Width of lower face 0-80 times vertical length of eye. Malar space 0-60 times basal width ofmandible. Maxillary palp as in Fig. 107; segment 4 considerably widened and flattened; segment 5 about0-5 as long as 4. Antennal flag setae (similar to Fig. 126, but setae slightly longer) upright, long, flattened, 2per segment, arising from a polished trough. Hind femur 5-25 times as long as deep. Fore wing length 8-3mm. Tergite 1 of gaster 2-30 times as long as broad. Tergite 3 with a crease along its anterior 0-5 separatinglaterotergite. Colour: black and yellow, with some reddish, especially on legs. Antenna black dorsally, scape andpedicel yellow ventrally, underside of flagellum reddish yellow proximally and distally, black medially.Maxillary palp with segments 1 brownish, 2 reddish yellow, 3 brownish and 4 and 5 blackish. Posterior partof propodeum entirely yellow. Hind coxa black, with a yellow dorsal spot contiguous interno-laterally witha postero-ventral spot. Hind femur reddish yellow with a black mark ventrally. Segments of gaster eachblack anteriorly, yellow posteriorly, the yellow tending to reddish, especially on sternites. REMARKS. See 'Remarks' under nox. BIOLOGY AND HOSTS. Virtually unknown. Adults have been collected in May and October. The specimenscollected at Simla by Gupta and Joseph were from coniferous and deodar forest (Gupta, 1971: (appendix)2, 3, 54). DISTRIBUTION (Map 5, p. 22). Along the southern flank of the Himalayas at about 2000 m. MATERIAL EXAMINED (11 $, 2 cT). India: 1 $, Himachal Pradesh, Simla, v.1897 (Nurse) (BMNH) (lectotype); 4 $, 1 cf, HimachalPradesh, Simla, 2133 m, 1 & lO.x.1962, 3.X.1966 (Gupta, Joseph) (GC); 1 $, Uttar Pradesh, Khurpatal,ll.x.1978 (Gupta) (GC); 1 $, 1 cT, Uttar Pradesh, Mussoorie, 2100 m, l.x.1962, 21.x. 1972 (Gupta,Khana) (GC); 3 $ , Uttar Pradesh, Nainital, 1938 m, 2-10 & 12.X.1978 (Gupta) (GC). Nepal: 1 $ , 3-2 kmSE. of Sikha, 2100-2400 m, 23.V.1954 (Quinlan) (BMNH). Banchus gudrunae sp. n. (Figs 17, 40, 65, 66, 87, 108) DESCRIPTION. Female. Width of lower face 0-75-0-85 times vertical length of eye. Malar space 0-45-0-55times basal width of mandible. Maxillary palp as in Fig. 17. Antenna with apical segments longer thanbroad. Scutellar spine (Fig. 40) about 0-7 as long as scutellum. Mesopleuron and mesoscutum coriaceous,with moderately strong punctures, on mesopleuron separated by about their diameter. Hind femur6-75-7-20 times as long as deep. Fore wing length 9-7-10-7 mm. Tergite 1 of gaster 2-00-2-50 times as long 24 M. G. FITTON as broad, its dorsal profile as in Figs 65 (male), 66 (female). Gaster compressed from tergite 4, reachingabout to tips of fore wings (when folded back). Tergite 3 with a crease along its whole length separatinglaterotergite. Tergite 7 and sternite 6 as in Fig. 87. Colour: black and yellow. Face yellow with a black median stripe. Antenna black (sometimes slightlybrownish distally), scape, and sometimes pedicel, with a yellow patch beneath. Maxillary palp reddishyellow with segment 1 and distal parts of 4 and 5 blackish. Posterior part of propodeum yellow, narrowlyblack posteriorly. Hind coxa black with a yellow dorsal patch. Hind femur yellowish with proximal halflargely black. Segments of gaster each black anteriorly, yellow posteriorly, the yellow bands relativelywide. Male. Width of lower face 0-80-0-90 times vertical length of eye. Malar space 0-45-0-65 times basalwidth of mandible. Maxillary palp as in Fig. 108; segment 4 flattened and widened distally; segment 5 about0-5 as long as 4. Antennal flag setae (similar to Fig. 127, but with setae slightly shorter and very muchbroader) upright, relatively short, flattened and very broad, 2 (occasionally 3) per segment, arising from apolished trough. Hind femur 7-00-7-45 times as long as deep. Fore wing length 8-8-10-0 mm. Tergite 1 ofgaster 2-10-2-40 times as long as broad. Tergite 3 with a crease along its whole length separatinglaterotergite. Colour: black and yellow. Antenna black dorsally, scape and pedicel yellow ventrally, flagellumbrownish ventrally. Maxillary palp reddish yellow, segment 1 dorsally narrowly black, segment 4 distallyand 5 wholly blackish. Posterior part of propodeum yellow, very narrowly black posteriorly. Hind coxablack with a large yellow dorsal patch. Hind femur yellow, blackish on its proximal 0-3, especially ventrallyand laterally. Segments of gaster each black anteriorly, very broadly yellow posteriorly. REMARKS. This species has a number of characters which make relating it to any others in the Palaearcticdifficult. The form of the male palp suggests a relationship withpalpalis and crefeldensis but the flag setaeare of a very much more specialized kind and the female palp is not specialized in the same way. BIOLOGY AND HOSTS. Unknown. Dates of collection of adults range from the middle of November to themiddle of January. DISTRIBUTION (Map 6, p. 27). Known only from the island of Cyprus. MATERIAL EXAMINED Holotype cf , Cyprus: Polemedia Hills, 14.xii.1948 (Mavromoustakis) (BMNH). Paratypes 10 $ , 9 cf . Cyprus: 6 $ , 3 cf , Polemedia Hills, 14 & 20.xii.1948 (Mavromoustakis) (BMNH);3 $, Palodkia, 5 & 12. i. 1949 (Mavromoustakis} (BMNH); 1 $, Limassol, 4.U940 (Mavromoustakis)(USNM); 4 cf, Zakaki Marshes, 29.xi.1946 (Mavromoustakis) (TC); 1 cf, near Limassol, 18.xi.1946(Mavromoustakis) (TC); 1 cf , Paphos district, near Panayia, 900 m, 29.xi.1946 (Mavromoustakis) (TC). Banchus hastator (Fahricius ) (Figs 18, 41, 64, 88, 109, 124) Ichneumon hastator Fabricius, 1793: 167. Holotype $, EUROPE (UZM) [examined]. Ichneumon pungitor Thunberg, 1822: 265; 1824: 320. [Replacement name for Ichneumon hastator Fabricius.] Ichneumon reticulatorThunberg, 1822: 265; 1824: 321. Type(s) cf , SWEDEN (lost). Syn. n.Banchus femoralis Thomson, 1897: 2411. Lectotype 9, SWEDEN (UZI), designated by Townes, Momoi & Townes, 1965: 237 [examined].Banchus kolosovi Meyer, 1925: 10. Syntypes3 <J>, U.S.S.R. (destroyed). Syn. n. NOMENCLATURE. This species until recently has been referred to by the junior synonym femoralis. It is dealtwith in a relatively large number of non-taxonomic papers so the name change is particularly unfortunate.A further source of confusion is the incorrect use of the name hastator by Townes & Townes (1978) (seenomenclatural notes under palpalis). Ichneumon pungitor is a junior objective synonym of hastator. It was proposed apparently because of thesecondary homonymy in Ichneumon, in Thunberg's work, with Foenus hastator Fabricius, 1804 [which,incidentally, was miss-spelled as vastator by Thunberg (1822: 262; 1824: 315)]. The status of pungitor as areplacement name has not been recognized previously and it had been placed incorrectly in the synonymyof B. volutatorius (Linnaeus) (e.g. Aubert, 1978: 158). Ichneumon reticulator has not been identified since its description. Roman (1912: 277) reported thatthere were no specimens in Thunberg's collection and suggested that the species was a Banchus. Thedescription fits some males of hastator, with which it is here synonymised. After its description, except for an entry in the Zoological Record, B. kolosovi was not referred to (even ICHNEUMON-FLY GENUS BANCHUS 25 by Meyer) until Aubert (1978: 166) noted its existence. Aubert miss-spelled the name as kozlovi. Thedescription fits particularly dark specimens ofhastator, with which it is here synonymised. NotwithstandingMeyer's comments on kolosovi andfemoralis, it seems that he realised his mistake because in his treatmentof the genus (1934: 229) he mentions, under femoralis, material from Sverdlovsk (the type-locality ofkolosovi) collected by Kolosov. This is presumably the type-material of kolosovi. No material collected byKolosov is noted under other Banchus species by Meyer. DESCRIPTION. Female. Width of lower face 0-85-0-90 times vertical length of eye. Malar space 0-50-0-65times basal width of mandible. Maxillary palp as in Fig. 18. Antenna with apical segments longer thanbroad. Scutellar spine (Fig. 41) about 0-2 as long as scutellum. Mesopleuron and mesoscutum coriaceous,usually fairly weakly so, with moderate punctures, on mesopleuron separated by about their diameter.Hind femur 4-50-5-15 times as long as deep. Fore wing length 8-8-10-9 mm. Tergite 1 of gaster 1-50-1-70times as long as broad, its dorsal profile as in Fig. 64. Gaster weakly compressed, reaching about to tips offore wings (when folded back). Tergite 3 with a crease along its anterior 0-3 separating laterotergite.Tergite 7 and sternite 6 as in Fig. 88. Colour: mainly black, with yellow and reddish, particularly on legs. Face black with yellow orbitalmarks, which are usually small but are sometimes large and with a pair of very small reddish marks in centreof face. Antenna with scape and pedicel black, scape, and usually pedicel, yellow ventrally. Flagellumbrown dorsally, reddish ventrally, darker at extreme base and apex. Maxillary palp reddish with segment 1black and segments 2 proximally, 4 distally and 5 wholly darker. Posterior part of propodeum entirelyblack. Hind coxa black. Hind femur reddish or yellowish proximally and distally, medially varying fromreddish dorsally and blackish ventrally to all blackish. Segments of gaster black, sometimes almost entirelyexcept for brownish posterior margin of sternites 1 and 6 and tergite 7, but usually with posterior margins offurther segments narrowly yellowish and with the yellow grading into a wider brownish band. Male. Width of lower face 0-90-1-00 times vertical length of eye. Malar space 0-50-0-65 times basalwidth of mandible. Maxillary palp as in Fig. 109; segment 4 only slightly widened and flattened; segment 5almost as long as 4. Antennal flag setae (Fig. 124) relatively poorly differentiated, small, not flattened, 2per segment, not arising from a trough. Hind femur 4-60-5-50 times as long as deep. Fore wing length7-2-10-7 mm. Tergite 1 of gaster 1-45-1-80 times as long as broad. Tergite 3 with a crease along its anterior0-3 separating laterotergite. Colour: mainly black, with some reddish and yellow, particularly on legs. Antenna black dorsally, scapeand pedicel yellow ventrally, flagellum reddish ventrally. Maxillary palp with segment 1 blackish orblackish and yellow, 2 and 3 reddish, 4 and 5 blackish or brownish. Posterior part of propodeum entirelyblack. Hind coxa black (in a single specimen with a small yellowish dorsal spot and the postero- ventralmargin reddish). Hind femur reddish or yellowish proximally and distally, medially varying from reddishdorsally and blackish ventrally to all blackish. Gaster sometimes almost entirely black except for brownishposterior margins of some segments, but usually with posterior margins of segments narrowly yellowishwith adjacent brownish bands. REMARKS. This is an isolated species, with few obvious specializations. The primitive condition of theantennal flag setae in the male is particularly notable. BIOLOGY AND HOSTS. The species is an important parasite of Panolis flammea (Denis & Schiffermiiller)(Noctuidae), which is sometimes a serious pest in pine forests. B. hastator is univoltine and adults occur inMay, June and early July. Reared specimens have emergence dates in March and April also. Some detailsof the biology and ecology of the species have been investigated as a result of the association with aneconomically important host. Most of the work has been connected with its effect as a natural control agent(Friederichs, 1936; Habermehl, 1922, 1924; Pfeffer, 1933; Scheidter, 1934; Schwerdtfeger, 1952; Smits vanBurgst, 1927) but has included studies of embryonic and larval development and morphology (Beirne,1941; Biedowski & Krainska, 1926; Shevyrev, 1913), courtship, host identification and host-stagepreferences, egg-laying and the reproductive system (van Veen, 1982). There are many specimens reared from P. flammea in collections. Apart from that species, the followingspecies are recorded as hosts on specimen labels: Blepharita adusta (Esper) (Noctuidae) (1 $, 2 cf,Germany (Smits van Burgst coll.) LELW); Deilephila porcellus (Linnaeus) (Sphingidae) (1 cf , no locality(Adkin) BMNH); Hyloicus pinastri (Linnaeus) (Sphingidae) (1 $, Poland (Mazur) IZPAN) and ?Achlyaflavicornis (Linnaeus) (Thyatiridae) (1 $, Great Britain (Lyle) BMNH). While one might cast doubt onthese records it is difficult to investigate them critically. Kolubajiv (1934: 114, 116) records Lymantriadispar (Linnaeus) (Lymantriidae) as a host. DISTRIBUTION (Map 6, p. 27). Widely distributed in northern Europe as far east as Leningrad, perhapsextending further east to the Urals (if my synonymy of kolosovi is correct (see 'Nomenclature' above)). Italso occurs in the European Alps and I have seen single individuals collected in Corsica and Yugoslavia. 26 M. G. FITTON MATERIAL EXAMINED (113 $, 151 cf) Belgium, Corsica, Denmark, East Germany, Finland, France, Great Britain, Ireland, Netherlands,Poland, Sweden, Switzerland, U.S.S.R., West Germany, Yugoslavia (AC, BC, BMNH, BRI, HC, IP,IRSNB, ITZ, IZPAN, LELW, KC, MNHN, MUM, NMB, NR, RNH, RSM, SC, SMT, UM, USNM,UZI, UZM, VRC, ZC, ZIL, ZMU, ZSBS). Banchus insulanus Roman (Figs 19, 42, 67, 89)Banchus insulanus Roman, 1937: 18. Holotype $, MADEIRA (NR) [examined]. DESCRIPTION. Female. Width of lower face 0-85 times vertical length of eye. Malar space 0-55 times basalwidth of mandible. Maxillary palp as in Fig. 19. Antenna with apical segments longer than broad. Scutellarspine (Fig. 42) indistinct. Mesopleuron and mesoscutum coriaceous, with moderate punctures, onmesopleuron separated by more than their diameter. Hind femur 6-55 times as long as deep. Fore winglength 9-3-9-5 mm. Tergite 1 of gaster 1-65 times as long as broad, its dorsal profile as in Fig. 67. Gasteronly weakly compressed, not reaching to tips of fore wings (when folded back). Tergite 3 with a creasealong its entire length separating laterotergite. Tergite 7 and sternite 6 as in Fig. 89. Colour: mainly black, with yellow and reddish marks. Face black with irregular, yellow orbital marks.Antenna black. Maxillary palp blackish. Posterior part of propodeum black. Hind coxa black with a smallyellow dorsal patch. Hind femur yellowish, blackish ventrally and posteriorly on its proximal 0-5. Segmentsof gaster blackish, in one specimen with only posterior 0-5 of tergite 3 and posterior 0-3 of tergite 7 reddish;in the other specimen with posterior 0-7 of tergite 2, almost all of tergite 3, posterior 0-5 of tergite 4, andtergites 5 and 6 laterally, reddish. Male. Unknown (but see 'Remarks'). REMARKS. The specimen described as the male by Hellen (1949: 13) proved, on examination, to be afemale. The material collected in 1959 and supposedly including both sexes (Hellen, 1961: 37) cannot befound in Hellen's collection (A. Albrecht, pers. comm.).The relationships of the species are not at all obvious from the characters of the female. BIOLOGY AND HOSTS. Unknown. Adults have been collected in May and July-August.DISTRIBUTION (Map 6, p. 27). Known only from the island of Madeira. MATERIAL EXAMINED (2 9) Madeira: 1 $, 25 Fontes, Rabacal, 17.vrM.viii.1935 (Lunblad) (NR) (holotype); 1 $, Ribeiro Frio,3.v.l938(Fre>0(ZMU). Banchus japonicus (Ashmead) (Figs 20, 43, 68, 90, 110)Nawoia japonica Ashmead, 1906: 185. LECTOTYPE 9, JAPAN (USNM), here designated [examined]. NOMENCLATURE. Ashmead described this species from two specimens, also stating 'Type. - Cat. No. 7259,U.S.N.M.', thus giving the impression that one was the holotype. However, both specimens are labelledType No. 7259 U.S.N.M.' and must be considered as syntypes. Other labels include 'Nawaia japonicaAshm' on one specimen and 'Nawaia japonica Ash. Paratype' on the other, the latter on a similar label butin a different hand and obviously more modern than the former. The specimen labelled as paratype alsobears a label '47' and is in much better condition than the other. I have labelled and hereby designate it aslectotype. DESCRIPTION. Female. Width of lower face 0-80-0-90 times vertical length of eye. Malar space 0-35-0-45times basal width of mandible. Maxillary palp as in Fig. 20. Antenna with apical segments about as long asbroad. Scutellar spine (Fig. 43) about 0-3 as long as scutellum. Mesopleuron and mesoscutum shining andonly weakly coriaceous, with moderate punctures, on mesopleuron separated by about their diameter.Hind femur 5-55-6-50 times as long as deep. Fore wing length 9-2-12-1 mm. Tergite 1 of gaster 1-45-1-85times as long as broad, its dorsal profile as in Fig. 68. Gaster only weakly compressed, reaching to tips offore wings (when folded back). Tergite 3 with a crease along its anterior 0-5 separating laterotergite.Tergite 7 and sternite 6 as in Fig. 90. Colour: black and yellow. Face yellow with a broad median black stripe and sometimes a black markbelow each antennal socket. Antenna black, with a small yellow mark on scape. Maxillary palp blackish,with a longitudinal yellow stripe ventrally on segments 1 and 2. Posterior part of propodeum yellow, ICHNEUMON-FLY GENUS BANCHUS 27 Map 6 ^ gudrunihastatorV insulanus Map 6 Distribution of Banchus gudrunae, B. hastator and B. insulanus. entirely or with a black area postero-medially. Hind coxa black with a yellow dorsal patch. Hind femurblackish, yellow proximally and distally. Segments of gaster black with posterior margins broadly yellowand with a pair of yellow spots on tergite 2 and usually also on 1 and 3. Male. Width of lower face 0-90 times vertical length of eye. Malar space 0-35-0-45 times basal width ofmandible. Maxillary palp as in Fig. 110; segment 4 flattened and considerably widened; segment 5 about0-5 as long as 4. Antennal flag setae (similar in appearance to Fig. 126, but with setae longer) upright, long,flattened, 2 per segment, arising from trough which is not polished. Hind femur 6-25-6-75 times as long asdeep. Fore wing length 9-3-10-3 mm. Tergite 1 of gaster 1-50-1-65 times as long as broad. Tergite 3 with acrease along almost its entire length separating laterotergite. Colour: black and yellow. Antenna black with scape, pedicel and proximal end of flagellar segment 1yellow ventrally. Maxillary palp with segment 1 yellow, 2 and 3 brownish yellow to blackish, 4 and 5 black.Posterior part of propodeum yellow, entirely or with small black marks posteriorly. Hind coxa black andyellow. Hind femur yellow with a black mark ventrally and interno-laterally, usually extending also todorsal surface. Segments of gaster black with posterior margins broadly yellow and with a pair of yellowspots on tergite 2 and usually also on 1 and 3. REMARKS. Although japonicus has been confused withpoppiti the two species are easily separated.BIOLOGY AND HOSTS. Unknown. Adults have been collected in May and June. DISTRIBUTION (Map 2, p. 16). Japan. Recorded also from Korea (Kim, 1955: 493), from which country theonly species of Banchus examined is B. palpalis. MATERIAL EXAMINED (6 $ , 6 c?) Japan: 1 $, Gifu, Fujishiro (Nawa) (USNM) (lectotype); 1 $, Gifu, Gifu-yama (Nawa) (USNM)(paralectotype); 1 $, Hokkaido (Uchidd) (TC); 1 $, 3 cf, Hokkaido, Mt Soranuma, 26.vi.1965 and 28 M. G. FITTON 15.vi.1968 (Kusigemati) (KC); 1 cf, Hokkaido, Toya-ko, 14.vi.1967 (Miyazaki) (KC); 1 $, Honshu,Tokyo, Komaba, 9.V.1916 (Hirayama) (EIHU); 1 cf , Honshu, Hyogo Pref., Ryuzoji, Sasayama, 5. v. 1965(Nakanishi) (NC); 1 $, Honshu, Tottori Pref., Mt Daisen, 12.vi.1963 (Nakanishi) (NC); 1 cf , Honshu,Yatsugatake, 2.vi.l967 (Kocha) (KC). Banchus mauricettae sp. n. (Figs 21, 44, 69, 91, 111) DESCRIPTION. Female. Width of lower face 0-90-0-95 times vertical length of eye. Malar space 0-55 timesbasal width of mandible. Maxillary palp as in Fig. 21. Antenna with apical segments longer than broad.Scutellar spine (Fig. 44) about as long as scutellum. Mesopleuron and mesoscutum coriaceous, withmoderate to strong punctures, on mesopleuron separated by a little more than their diameter. Hind femur5-20-5-30 times as long as deep. Fore wing length 10-5-10-6 mm. Tergite 1 of gaster 1-90-2-35 times as longas broad, its dorsal profile as in Fig. 69 (male). Gaster moderately compressed, not quite reaching tips offore wings (when folded back). Tergite 3 with a crease along its anterior 0-6 separating laterotergite.Tergite 7 and sternite 6 as in Fig. 91. Colour: yellowish red with black and yellow marks. Face yellowish red with orbits yellow and sometimeswith a blackish median stripe. Antenna entirely reddish yellow except for extreme apex of flagellum andscape and pedicel dorsally, which are blackish. Maxillary palp reddish yellow with segments 4 distally and 5wholly dark. Posterior part of propodeum reddish, entirely or with black marks posteriorly. Hind coxavarying from reddish with black patches to almost entirely black. Hind femur entirely reddish. Segments ofgaster entirely reddish or with some black anteriorly. Male. Width of lower face 0-85-0-90 times vertical length of eye. Malar space 0-60-0-65 times basalwidth of mandible. Maxillary palp as in Fig. Ill; segment 4 flattened and widened; segment 5 about 0-8 aslong as 4. Antennal flag setae (similar in appearance to Fig. 129, but with setae only two-thirds as long)upright, fairly short and broad, 3 (occasionally 4) per segment, arising from a polished trough. Hind femur5 95-6 20 times as long as deep . Fore wing length 11-2-12-3 mm . Tergite 1 of gaster 2 20-2 35 times as longas broad. Tergite 3 with a crease along its anterior 0-6 separating laterotergite. Colour: yellow, reddish and black. Antenna blackish dorsally, scape and pedicel yellow ventrally,flagellum reddish yellow ventrally. Maxillary palp reddish yellow with segments 4 distally and 5 whollyblackish. Posterior part of propodeum entirely yellow. Hind coxa largely reddish yellow, blackish orbrownish antero-ventrally and dorso-posteriorly. Hind femur reddish, sometimes with a blackish markventrally. Segments of gaster each black, reddish and yellow (from anterior to posterior), the area of blackreduced or absent on posterior segments. REMARKS. This is a distinctive species; its relationships are difficult to discern.BIOLOGY AND HOSTS. Unknown. Dates of collection are in June and July. DISTRIBUTION (Map 2, p. 16). The Szechuen province of China. Altitudes given on data labels range from300-4500 m. MATERIAL EXAMINED Holotype cf , China: Szechuen, Yao-Gi, 1200-2400 m, 3.vii.l929 (Graham) (USNM). Paratypes 4 $, 10 cf. China: 1 9, 2 cf, Szechuen, Yao-Gi, 1200-2400 m, 3 & 16.vii.1929 (Graham)(USNM); 2 cf, Szechuen, near Mupin, 600-2400 m, 28.vi & 22.vii.1929 (Graham) (USNM); 1 cf,Szechuen, Ningyuenfu, 1800-3240 m, 24-26. vii. 1928 (Graham) (USNM); 1 cf , Szechuen, Mt Omei, 840m, vi.1937 (Graham) (USNM); 1 cf , Szechuen, Suifu, 300-450 m, l-21.vi.1928 (Graham) (USNM); 2 cf ,Szechuen, 14-5 km SW. of Tatsienlu, 2550-3900 m, 25-27.vi.1923 (Graham) (USNM); 1 cf , Szechuen, ULong Kong, near Tatsienlu, 3000-4500 m, 25-30.vi.1923 (Graham) (USNM); 2 9, Szechuen, Yachow,v-vi.1928 (Graham) (USNM); 1 $, Szechuen, Yachow to Mupin, 600-1500 m, 23-27.vi.1929 (Graham)(USNM). Banchus moppitisp. n. (Figs 22, 45, 70, 92, 113, 127) DESCRIPTION. Female. Width of lower face 1-10-1-20 times vertical length of eye. Malar space 0-85-1-00times basal width of mandible. Maxillary palp as in Fig. 22. Antenna with apical segments about as broad aslong. Scutellar spine (Fig. 45) about 0-6 as long as scutellum, usually downcurved at the tip. Mesopleuronand mesoscutum coriaceous, usually strongly so, with strong punctures, on mesopleuron separated bymuch less than their diameter. Hind femur 4-75-5-20 times as long as deep. Fore wing length 8-2-10-8 mm.Tergite 1 of gaster 1-60-1-80 times as long as broad, its dorsal profile as in Fig. 70. Gaster compressed from ICHNEUMON-FLY GENUS BANCHUS 29 segment 4, reaching just beyond tips of fore wings (when folded back). Tergite 3 with a crease along itsanterior 0-3 separating laterotergite. Tergite 7 and sternite 6 as in Fig. 92. Colour: black and yellow. Face black with yellow orbital stripes, usually broad but varying considerablyin width. Antenna black, usually with a small yellow patch on underside of scape. Maxillary palp blackish.Posterior part of propodeum black with a broad yellow stripe along posterior transverse carina. Hind coxablack, usually with a small yellow dorsal patch. Hind femur black, yellow proximally and distally andusually dorsally. Segments of gaster each black anteriorly, broadly yellow posteriorly. Male. Width of lower face 1-10-1 -20 times vertical length of eye. Malar space 0-75-0-95 times basalwidth of mandible. Maxillary palp as in Fig. 113; segment 4 flattened and widened; segment 5 not quite aslong as 4. Antennal flag setae (Fig. 127) upright, relatively short, flattened and widened, 2 per segment,arising from a polished trough. Hind femur 4-80-5-50 times as long as deep. Fore wing length 8-5-10-9 mm.Tergite 1 of gaster 1-45-1-60 times as long as broad. Tergite 3 with a crease along its anterior 0-3 separatinglaterotergite. Colour: black and yellow. Antenna black with scape and pedicel yellow ventrally. Maxillary palpblackish. Posterior part of propodeum yellow anteriorly, black posteriorly. Hind coxa black, usually with ayellow dorsal spot and sometimes also a postero-lateral one. Hind femur yellow with an extensive blackmark, extending ventrally and laterally. Segments of gaster each black anteriorly, yellow posteriorly. REMARKS. This species seems closely related to zonatus. Its generally much smaller size and relativelyshorter appendages (and gaster in females) readily differentiate it. BIOLOGY AND HOSTS. Unknown. The specimens were collected early in the year (in February, March andApril) with the exception of a female taken in October. DISTRIBUTION (Map 7, p. 29). The few known specimens come from widely scattered localities in westernEurope. Map? % moppiti Q moppiti. located only to country Map 7 Distribution of Banchus moppiti. 30 M. G. FITTON MATERIAL EXAMINED Holotype $, Spain: Madrid, Cercedilla, 22.X.1978 (Noyes) (BMNH). Paratypes 6 $, 4 cf . France: 1 cf , Vincennes, 20.ii.1885 (de Gaulle coll.) (MNHN); 1 cf , Boulogne,22.iii (Giraud coll.) (MNHN). Great Britain: 1 $ (Desvignes coll.) (BMNH). Spain: 4 $, 2 cf , Teruel,15.iii.82 (Hiendlmayr coll.) [note - only 1 $ of these 6 specimens has the locality and date but all areidentically mounted and set] (ZSBS). Switzerland: 1 $ , 7.iv.l861 (Sichel coll.) (MNHN). Banchus nox Morley(Figs 23, 46, 71, 93, 112) Banchus nox Morley, 1913: 255. Holotype cf , INDIA (BMNH) [examined].NOMENCLATURE. The species misidentified as nox by Chandra & Gupta (1977: 182) is B. punkettai. DESCRIPTION, female. Width of lower face 0-80 times vertical length of eye. Malar space 0-55 times basalwidth of mandible. Maxillary palp as in Fig. 23. Antenna with apical segments longer than broad. Scutellarspine (Fig. 46) about 0-7 as long as scutellum. Mesopleuron and mesoscutum coriaceous, with moderatepunctures, on mesopleuron separated by a little more than their diameter. Hind femur 6-40 times as long asdeep. Fore wing length 11-2 mm. Tergite 1 of gaster 2-70 times as long as broad, its dorsal profile as in Fig.71 (male). Gaster strongly compressed from posterior of segment 3, not reaching to tips of fore wings(when folded back). Tergite 3 with a crease along its anterior 0-4 separating laterotergite. Tergite 7 andsternite 6 as in Fig. 93. Colour: black and reddish, with some yellow marks. Face black with very broad yellow orbital markswhich also extend beneath antennal sockets. Antenna dark brown, blackish dorsally, scape and pedicelyellowish ventralty. Maxillary palp reddish yellow, with segment 1 dorsally and segment 4 distally blackish.Posterior part of propodeum blackish with an irregular reddish yellow band along posterior transversecarina. Hind coxa black with a large reddish dorsal patch. Hind femur reddish with a black stripe ventrally.Segments of gaster reddish, tergites 1, 2 and 3 black anteriorly. Male. Width of lower face 0-85 times vertical length of eye. Malar space 0-50 times basal width ofmandible. Maxillary palp as in Fig. 112; segment 4 flattened and considerably widened; segment 5 about0-6 as long as 4. Antennal flag setae (similar to Fig. 126, but setae slightly longer) upright, long, flattened, 2per segment, arising from a polished trough. Hind femur 6-05 times as long as deep. Fore wing length 9-7mm. Tergite 1 of gaster 2-15 times as long as broad. Tergite 3 with a crease along its anterior 0-5 separatinglaterotergite. Colour: black and yellow, with some reddish, especially on legs. Antenna black dorsally, scape andpedicel yellow ventrally, underside of flagellum with segment 1 and distal part reddish yellow, remainderblack. Maxillary palp with segment 1 yellowish, 2 and 3 wholly and 4 proximally reddish yellow, 4 distallyand 5 wholly brownish. Posterior part of propodeum yellow, narrowly black postero-medially. Hind coxablack, with a yellow dorsal spot contiguous interno-laterally with a postero- ventral spot. Hind femurreddish with a black mark ventrally. Segments of gaster each black anteriorly, yellow posteriorly, withjunction of the two areas reddish. REMARKS. Closely related toflavomaculatus. The aggregate differences between these taxa warrant theirseparation, but this will need to be reassessed when more material becomes available for study. BIOLOGY AND HOSTS. The only two known specimens were collected in October and November, the male onflowers of Spiraea (Morley, 1913: 255). DISTRIBUTION (Map 2, p. 16). Assam, between 1800 and 2400 m. MATERIAL EXAMINED (1 $, 1 cf) India: 1 cf, Assam, Shillong, Khasi Hills, 1800 m, x.1903 (Turner) (BMNH) (holotype); 1 $, Assam,Mishmi Hills, Delai Valley, Cha Che, 2200-2400 m, 21.xi.1936 (Steele) (BMNH). Banchus palpalis Ruthe (Figs 1,24, 47, 72, 94, 114) [Banchus monileatus Gravenhorst, 1829: 393; in part. Misidentification.] Banchus palpalis Ruthe, 1859: 377. Syntypes 2 cf , ICELAND (lost). Banchus spinosus Cresson, 1865: 274. Holotype 9 [not cf as stated by Cresson], U.S.A. (ANS) [not examined]. Syn. n.Banchus formidabilis Provancher, 1874: 61. Holotype $, CANADA (UL) [not examined]. Syn. n. ICHNEUMON-FLY GENUS BANCHUS 31 Banchus (Corynephanus) groenlandicus Aurivillius, 1890: 30. Lectotype cf , GREENLAND (NR), designated by Townes, 1961: 104 [examined]. Syn. n.Cidaphurus alticola Ashmead, 1901 : 148. Holotype $ [not cf as stated by Ashmead], U.S. A. (USNM) [not examined]. Syn. n. NOMENCLATURE. For almost 150 years the name monileatus was applied consistently to this species.However, two species were mixed in Gravenhorst's original material and a series of unfortunate eventsculminated in Townes & Townes (1978: 532) designating as lectotype of monileatus a female which wasvolutatorius , rather than the species as it had been identified previously. In addition, as a result ofmisinformation about my examination of the holotype of hastator, Townes & Townes incorrectlysynonymized monileatus with that species. Female volutatorius and palpalis superficially are strikinglysimilar, so much so that Aubert (1978: 156), without realising that he was examining volutatorius,commented that Townes' lectotype differed from volutatorius only in certain details! When he originallyexamined Gravenhorst's syntypes (in 1964) Townes probably checked the identity of the species and thenmade the mistake of selecting the specimen in best condition for designation as lectotype withoutre-checking it. Under the International Code the lectotype designation has priority over the previousrestriction of the taxonomic species. Thus monileatus becomes a junior synonym of volutatorius while the species previously known asmonileatus must take the next available name -palpalis. The name hastator correctly applies to the speciespreviously known by the junior synonym femoralis. The identity of palpalis was established by Roman (1928: 24; 1930: 285) and Fitton (1978a: 76). Banchus spinosus, B. formidabilis and Cidaphurus alticola are included in the synonymy of palpalis onthe basis of their treatment in Townes & Townes (1978). DESCRIPTION. Female. Width of lower face 0-80-0-90 times vertical length of eye. Malar space 0-65-0-75times basal width of mandible. Maxillary palp as in Fig. 24. Antenna with apical segments longer thanbroad. Scutellar spine (Fig. 47) long, more than 0-8 as long as scutellum (in a very few specimens only 0-3 aslong as scutellum). Mesopleuron and mesoscutum coriaceous, with fine to moderate punctures, onmesopleuron separated by a little more than their diameter. Hind femur 5-00-5-95 times as long as deep.Fore wing length 9-8-11-5 mm. Tergite 1 of gaster 1-65-2-15 times as long as broad, its dorsal profile as inFig. 72 (male). Gaster compressed from posterior of segment 3, not quite reaching to tips of fore wings(when folded back). Tergite 3 with a crease along its anterior 0-4 separating laterotergite. Tergite 7 andsternite 6 as in Fig. 94. Colour: almost entirely black except for mainly reddish legs and a few small yellowish marks. Face blackwith yellow orbital marks, usually very small and sometimes absent. Antenna black, entirely or withyellowish marks ventrally on scape and pedicel. Maxillary palp yellowish or brownish with segments 1 and 5wholly and 3 and 4 distally blackish. Posterior part of propodeum entirely black. Hind coxa black. Hindfemur reddish yellow, usually entirely, rarely darkened ventrally. Segments of gaster black, sternites andposterior edges of tergites 6 and 7 sometimes brownish. Male (Fig. 1). Width of lower face 0-90-0-95 times vertical length of eye. Malar space 0-60-0-70 timesbasal width of mandible. Maxillary palp as in Fig. 114; segment 4 narrow and cylindrical, flattened andwidened only at its extreme distal apex; segment 5 less than 0-2 as long as 4. Antennal flag setae (similar toFig. 125, but setae slightly longer and showing signs of flattening) at about 50 degrees, very weaklyflattened, 3 per segment, not arising from a trough. Hind femur 5-35-6- 10 times as long as deep. Fore winglength 9-8-11-4 mm. Tergite 1 of gaster 1-90-2-25 times as long as broad. Tergite 3 with a crease along itsanterior 0-5 separating laterotergite. Colour: black, with some yellow and with legs mainly reddish yellow. Antenna black, scape and pedicelyellow ventrally, proximal and distal parts of flagellum often brownish yellow ventrally. Maxillary palpwith segments 1 and 2 wholly and 3 and 4 proximally reddish yellow, 3 and 4 distally and 5 wholly blackish.Posterior part of propodeum usually entirely black, sometimes with a transverse yellow mark (or marks)immediately behind posterior transverse carina. Hind coxa sometimes entirely black, usually with anexterno-lateral reddish yellow mark and sometimes also dorsal and/or ventral yellowish spots. Hind femurentirely reddish yellow. Segments of gaster black, sometimes brownish or yellowish on their posteriormargins, very rarely with conspicuous yellow bands on tergites 1, 2 and 3. REMARKS. Easily identified from the structure of the maxillary palps. Nonetheless, it is confused withfemale volutatorius by incompetent or overconfident identifiers. Its presumed sister-species is crefeldensis.The males with well-developed yellow marking and conspicuous yellow bands on the gaster comemainly, but not exclusively, from the southern parts of the range. In North America, where thegeographical variation in colour may be more consistent and also applies to females, this form has beenrecognized as a separate 'subspecies' (Townes & Townes, 1978: 533). 32 M. G. FITTON BIOLOGY AND HOSTS. Adults normally occur from mid-June to mid-August, with one record fromSwitzerland as early as 22 May and one from Yugoslavia as late as September. In Europe the usual hostseems to be Blepharita adusta (Esper) (Noctuidae), from which I have seen 39 reared specimens. Of these,34, possibly 35, come from a single mass rearing (with the only data: Germany (Smits van Burgs t coll.)(LELW)). In Great Britain Blepharita adusta is more common in the north (Bretherton, Goater &Lorimer, 1983) and the records ofpalpalis show a similar pattern. Hosts recorded in the literature are: Panolis flammea (Denis & Schiffermuller) (Noctuidae) (Smits vanBurgst, 1927: 239), Lacanobia oleracea (Linnaeus) (Noctuidae) (Meyer, 1934: 231) and Deilephilaporcellus (Linnaeus) (Sphingidae) (Leonardi, 1928: 87). The record from L. oleracea, at least, canprobably be dismissed as having resulted from the misidentification of female volutatorius. DISTRIBUTION (Map 8, p. 32). The only Holarctic species of Banchus. Very widely distributed; in the moresouthern parts of its range it occurs at higher altitudes (for example, altitudes of 1800 to 2100 m arerecorded for the specimens from Austria, Yugoslavia, Bulgaria, Greece and Turkey). Map 8 tt. palpal is Map 8 Distribution of Banchus palpalis (the North American records are taken from Townes & Townes ,1978). ICHNEUMON-FLY GENUS BANCHUS 33 MATERIAL EXAMINED (103 $ , 88 cf ) Belgium, Bulgaria, Canada, Denmark, East Germany, Finland, France, Great Britain, Greece, Iceland,Ireland, Italy, Japan, Korea, Netherlands, Poland, Sweden, Switzerland, Turkey, West Germany, U.S.A.,U.S.S.R., Yugoslavia (BC, BMNH, BRI, EIHU, FSA, HC, IEUB, IRSNB, IZPAN, JC, JKC, KCLELW, MHN, MNHN, MNHU, MUM, NC, NR, RNH, USNM, UZM, VRC, ZC, ZI, ZMU, ZSBS). Banchus pictus Fab rid us(Figs 25, 49, 50, 73, 95, 115) Ichneumon cultratus Gmelin, 1790: 2708. Type(s) [?sex], EUROPE (destroyed). [Junior (by first reviserchoice of Gravenhorst, 1829: 382, 1006) primary homonym of Ichneumon cultratus Gmelin, 1790: 2699.] Ichneumon mutillatus Christ, 1791: 358. Type(s) cf , no type-locality (lost). [Junior primary homonym ofIchneumon mutillatus Gmelin, 1790: 2716.] Banchus pictus Fabricius, 1798: 234. Lectotype $, GERMANY (UZM), designated by Townes, Momoi &Townes, 1965: 238 [examined]. Banchus zagoriensis Hensch, 1928: 100. ?Syntypes 2 <J>, 1 cf , YUGOSLAVIA (ZSBS) [examined]. Banchus bipunctatus Hensch, 1928: 101. Holotype cf , YUGOSLAVIA (ZPZ) [examined]. Banchus russiator Aubert, 1978: 157. [Unavailable name published conditionally (Article 15 of the Code).] Banchus russiator Aubert, 1981: 18. Holotype $, U.S.S.R. (AC) [examined]. Syn. n. NOMENCLATURE. The synonymy of/, cultratus with pictus was queried by Aubert (1978: 157) but he gave noreasons for so doing and I can find none. The description of Ichneumon mutillatus Christ fits well the male of pictus and the synonymy, proposedby Gravenhorst (1829: 383) but queried by Aubert (1978: 157), is accepted. The Christ name is unavailablebecause of its homonymy with /. mutillatus Gmelin, which is a replacement name for /. mutillariusFabricius, 1787: 271 (a junior homonym of/, multillarius Fabricius, 1775: 342). I did not succeed in obtaining on loan material from the Hensch collection. However, Dr K. Horstmannwas able to visit Zagreb in 1980 and in the course of his work on the species of Ichneumonidae described byHensch he made the types of the Banchus species available to me. Unfortunately the specimen tentativelyselected for designation as lectotype of B. zagoriensis had the date of collection 28 August 1928 and couldnot therefore have been a syntype as the description was published on 1 September 1928 (Horstmann,1982a: 82). Three specimens labelled as cotypes (= syntypes) of B. zagoriensis are present in the Bauercollection (ZSBS). Because they bear no dates of collection and because the cotype labels were almostcertainly added by Bauer and not Hensch (Horstmann, pers. comm.) they are only tentatively regarded assyntypes. Selection and designation of a lectotype for B. zagoriensis is best deferred until the other materialin the Hensch collection can be examined. The identity of the species is not in doubt. DESCRIPTION. Female. Width of lower face 0-80-0-95 times vertical length of eye. Malar space 0-40-0-50times basal width of mandible. Maxillary palp as in Fig. 25. Antenna with apical segments slightly longerthan broad. Scutellar spine (Figs 49, 50) ranging from very small to about 0-6 as long as scutellum.Mesopleuron and mesoscutum coriaceous, with strong punctures, on mesopleuron separated by less thantheir diameter. Hind femur 5-00-5-50 times as long as deep. Fore wing length 8-1-10-6 mm. Tergite 1 ofgaster 1-45-1-65 times as long as broad, its dorsal profile as in Fig. 73. Gaster compressed, reaching justbeyond tips of fore wings (when folded back). Tergite 3 with a crease along its anterior 0-2 to 0-3 separatinglater otergite. Tergite 7 and sternite 6 as in Fig. 95. Colour: black and yellow. Face yellow, with a black median stripe (rarely rather wide). Antenna withscape and pedicel black dorsally, yellow ventrally. Flagellum orange, sometimes slightly darker dorsally,almost always with segment 1 proximally blackish and distal segments dark. Maxillary palp reddish yellow,with segment 1 black and sometimes segments 2 proximally and 4 distally darkened. Posterior part ofpropodeum yellow anteriorly, black posteriorly; the yellow area varying considerably in extent, fromcovering almost all of the area to being divided by the black area medially, reduced to a pair of spots, or(very rarely) entirely absent. Hind coxa black, entirely or with a yellow dorsal patch. Hind femur mediallyblack (sometimes brownish red, especially dorsally), yellow proximally and distally. Segments of gastereach black anteriorly, broadly yellow posteriorly. Male. Width of lower face 0-85-1-00 times vertical length of eye. Malar space 0-45-0-55 times basalwidth of mandible. Maxillary palp as in Fig. 115; segment 4 moderately widened and flattened; segment 5about 0-9 as long as 4. Antennal flag setae (similar to Fig. 125, but setae shorter) at about 60 degrees, short,not flattened, 2 per segment, not arising from a trough. Hind femur 5-05-5-65 times as long as deep. Forewing length 8-1-10-3 mm. Tergite 1 of gaster 1-45-1-70 times as long as broad. Tergite 3 with a crease alongits anterior 0-3 separating laterotergite. 34 M. G. FITTON Colour: black and yellow. Antenna with scape and pedicel black dorsally, yellow ventrally. Flagellumreddish orange, dorsally darker (blackish at base and apex). Maxillary palp blackish, sometimes entirely,but usually with segments 2 and 3 wholly and 4 and 5 proximally yellowish brown. Posterior part ofpropodeum black, entirely or with a pair of antero-lateral yellow patches (rarely the yellow moreextensive, very rarely posterior part of propodeum entirely yellow). Hind coxa black, often with a dorsalyellow patch. Hind femur medially black (sometimes brownish red, especially dorsally), yellow proximallyand distally. Segments of gaster each black anteriorly, broadly yellow posteriorly. REMARKS. Specimens of turcator have been confused with pictus. The two species can be distinguishedusing the characters given in the key. The morphospecies/7/ctaj, as here defined, has variations in time of occurrence of adults, length of thescutellar spine, and sex ratios which need further investigation and explanation. There are two distinct periods of occurrence of adults. When this was noted 184 specimens were to hand.Of these 90 had a date of collection, 48 were captured between 10 April and 3 June and 42 between 1 Julyand 2 October (most of the latter batch after 10 August). The length of the scutellar spine varies from verysmall to about 0-6 as long as the scutellum. Although there is continuous variation in the length of the spinemost of the specimens with a short spine were collected in the first period and most of those with a longerspine in the second. The sex ratio of the 48 specimens collected in the first period was 1 18 $ : 1 cf and of the42 in the second period 4-25 $ : Icf . Three possible biological explanations of these observations are that (1) pictus is bivoltine, with thegenerations varying slightly in morphology and more distinctly in sex ratio; (2) the early and late summergroups represent separate species; or (3) that one species has a partial second generation and a secondspecies (possibly parthenogenetic) occurs in late summer. If two species were involved diligent study should have revealed variation in some morphologicalcharacters, colour, geographical distribution, or host or other biological data correlated with date ofcapture or length of scutellar spine. None was found. Assignment of the 94 specimens without a date ofcapture to two subjective classes - with short or long scutellar spine - gave groups which varied in sex ratioin the same way as the early and late summer groups (short spine, 58 specimens, 1-55 $ : 1 cf ; long spine, 36specimens, 5 $: 1 cf). The only other relevant evidence comes from two reared specimens: two larvae ofthe noctuid Agrochola helvola (Linnaeus) collected near Sheffield, Great Britain (Ford) (UM) on 13 June1969 proved to be parasitized. The parasite larvae emerged in August 1969 and spun cocoons. The resultingadults (1 $ Id") did not emerge until early May 1970. None of this evidence gives unambiguous support to any of the three suggested explanations of theobservations. Even the one year life-cycle of the specimens reared in Sheffield could be explained bypostulating univoltinism in the northern part of the range of the species and bivoltinism in the south(although no evidence of this was found). However, the balance is perhaps in favour of the involvement ofmore than one species, if only because there is no undisputed evidence of bivoltinism in any other species ofBanchus. The poor quality and quantity of the available data and the possible complexity of the situationmean that the problem remains unresolved - a fruitful field for future study. BIOLOGY AND HOSTS. For information on phenology see 'Remarks' above. I have seen only three rearedspecimens of pictus, from Agrochola helvola (Linnaeus) (Noctuidae) (all from Great Britain (Lyle andFord) BMNH and UM). There are several hosts noted in the literature, but as black-and-yellow Banchusare frequently misidentified as pictus they should be treated with caution. The recorded hosts are:Agrochola circellaris (Hufnagel) (Noctuidae) (Habermehl, 1922: 269), Agrotis segetum (Denis & Schiffer-muller) (Noctuidae) (Bajari, 1960: 261), Atethmia ambusta (Denis & Schifferrmiller) (Noctuidae)(Schmiedeknecht, 1910: 1928), Lycophotia porphyrea (Denis & Schiffermiiller) (Noctuidae) (de Gaulle,1907: 119), Hadena rivularis (Fabricius) (Noctuidae) (Wagner, 1929: 11), Phalera bucephala (Linnaeus)(Notodontidae) (Leonardi, 1928: 83) and Smerinthus ocellata (Linnaeus) (Sphingidae) (Meyer, 1934: 228). DISTRIBUTION (Map 9, p. 35). Widely distributed in the western Palaearctic as far north as 56N in the west,and extending south-east as far as the Pamirs. MATERIAL EXAMINED (162 $ , 75 cf ). Austria, Belgium, Czechoslovakia, Denmark, East Germany, France, Great Britain, Greece, Hungary,Italy, Morocco, Netherlands, Poland, Portugal, Rumania, Sardinia, Spain, Sweden, Switzerland, Turkey,U.S.S.R., West Germany, Yugoslavia (AC, BMNH, BRI, CM, FSA, HC, IBMPP, IEAU, IEE, IP,IRSNB, KHC, LELW, MCSN, MHN, MIZS, MLSU, MNHN, MUM, NM, NMB, NR, PC, RNH, RSM,TM, UM, USNM, UZI, UZM, VRC, ZC, ZIL, ZMU, ZSBS). ICHNEUMON-FLY GENUS BANCHUS Map 9 Distribution of Banchus pictus, B. poppiti and B. punkettai. Banchuspoppitisp. n. (Figs 26, 48, 74, 96, 116) DESCRIPTION. Female. Width of lower face 0-90 times vertical length of eye. Malar space 0-55-0-65 timesbasal width of mandible. Maxillary palp as in Fig. 26. Antenna with apical segments longer than broad.Scutellar spine (Fig. 48) about 0-4 as long as scutellum. Mesopleuron and mesoscutum coriaceous, withmoderate punctures, on mesopleuron separated by about their diameter. Hind femur 5-25-5-65 times aslong as deep. Fore wing length 10-3-11-1 mm. Tergite 1 of gaster 1-65 times as long as broad, its dorsalprofile as in Fig. 74 (male). Gaster very weakly compressed, not reaching to tips of fore wings (when foldedback). Tergite 3 with a crease along its anterior 0-4 separating laterotergite. Tergite 7 and sternite 6 as inFig. 96. Colour: black and yellow, the margins of some yellow areas slightly reddish. Face black with yelloworbital stripes. Antenna with scape and pedicel black dorsally, yellow ventrally. Flagellum orangeventrally, dark dorsally. Maxillary palp yellowish with segments 1 dorsally, 4 distally, and 5 wholly dark.Posterior part of propodeum yellow, entirely or with posterior margin black. Hind coxa black with a largeyellow dorsal patch. Hind femur black, narrowly reddish yellow proximally and distally. Segments ofgaster black, narrowly yellow posteriorly. Male. Width of lower face 0-95 times vertical length of eye. Malar space 0-50-0-65 times basal width ofmandible. Maxillary palp as in Fig. 116; segment 4 flattened and widened; segment 5 about 0-5 as long as 4. 36 M. G. FITTON Antennal flag setae (similar to Fig. 127, but with setae narrower) upright, flattened, 2 per segment, arisingfrom a polished trough. Hind femur 5-60-6-10 times as long as deep. Fore wing length 11-1-11-2 mm.Tergite 1 of gaster 1-65-1 -90 times as long as broad. Tergite 3 with a crease along its anterior 0-4 separatinglaterotergite. Colour: black and yellow, the margins of some yellow areas slightly reddish. Antenna black dorsally,yellow ventrally. Maxillary palp yellow with segments 4 distally and 5 wholly blackish. Posterior part ofpropodeum yellow, entirely or narrowly black postero-medially. Hind coxa black with very large dorsaland ventral yellow patches. Hind femur yellow, black-marked ventrally and internally, tending to reddishdistally. Segments of gaster each black anteriorly, black posteriorly. REMARKS. This species has not previously been differentiated fromjaponicus.BIOLOGY AND HOSTS. Unknown. Adults have been collected in May and June.DISTRIBUTION (Map 9, p. 35). Japan. MATERIAL EXAMINED Holotype cf , Japan: Hakodate, 12.vi.1926 (Malaise) (NR). Paratypes 2 $, 1 cf. Japan: 1 $, 1 cf, Hakodate, 12.vi.1926 (Malaise) (NR); 1 , Mt Mino, 6.V.1929(Teranishi) (TC). Banchus punkettai sp. n. (Figs 27, 51, 75, 97, 117)[Banchus nox Morley; Chandra & Gupta, 1977: 182. Misidentification.] DESCRIPTION. Female. Width of lower face 0-75-0-80 times vertical length of eye. Malar space 0-45-0-60times basal width of mandible. Maxillary palp as in Fig. 27. Antenna with apical segments just broader thanlong. Scutellar spine (Fig. 51) about 0-5 as long as scutellum. Mesopleuron and mesoscutum shining, onlyvery weakly sculptured, with strong punctures, on mesopleuron separated by less than their diameter.Hind femur 5-60-6-10 times as long as deep. Fore wing length 8-6-9-7 mm. Tergite 1 of gaster 1-80-2-25times as long as broad, its dorsal profile as in Fig. 75. Gaster strongly compressed from posterior of segment3, reaching to tips of fore wings (when folded back). Tergite 3 with a crease along its anterior 0-3 separatinglaterotergite. Tergite 7 and sternite 6 as in Fig. 97. Colour: black and yellow, with some reddish, especially on legs. Face yellow with a black median stripe,the stripe sometimes rather wide but with reddish yellow patches remaining below antennal sockets.Antenna brown, often rather darker dorsally and distally, with scape and pedicel blackish dorsally andyellowish ventrally. Maxillary palp reddish yellow with distal parts of segments 4 and 5 dark. Posterior partof propodeum yellowish with a black patch medio-posteriorly, the yellow area sometimes reduced to aband along posterior transverse carina. Hind coxa black with a yellow dorsal patch and a red postero-ventral patch. Hind femur reddish with a blackish ventral stripe. Segments of gaster each black anteriorly,yellow posteriorly, with junction between two colours reddish. Tergite 1 also with posterior margin blackand yellow area reduced to two lateral and/or one median spot(s). Male. Width of lower face 0-90 times vertical length of eye. Malar space 0-65 times basal width ofmandible. Maxillary palp as in Fig. 117; segment 4 flattened and considerably widened; segment 5 about0-6 as long as 4. Antennal flag setae (similar to Fig. 124) relatively poorly differentiated, at about 50degrees, very short, not flattened, 2 per segment, without a polished trough. Hind femur 6-05 times as longas deep. Fore wing length 8-6 mm. Tergite 1 of gaster 1-90 times as long as broad. Tergite 3 with a creasealong its anterior 0-5 separating laterotergite. Colour: black and yellow, with some reddish, especially on legs. Antenna black dorsally, scape, pediceland proximal flagellar segments yellow ventrally, remainder of flagellum brownish yellow ventrally.Maxillary palp with segment 1 yellow, 2 and 3 reddish yellow, 4 and 5 blackish. Posterior part ofpropodeum yellow, black postero-medially. Hind coxa black, with a dorsal yellow spot which is contiguousinterno-laterally with a postero-ventral spot. Hind femur reddish, black ventrally and with a separatedorsal blackish area. Segments of gaster each black anteriorly, yellow posteriorly, the junction between thetwo colours very narrowly reddish. REMARKS. The colour pattern of this species is very similar to two other species, flavomaculatus and nox,found in the same area. The males are easily identified using the flag setae but the females are more difficultto separate. ICHNEUMON-FLY GENUS BANCHUS 37 BIOLOGY AND HOSTS. Virtually unknown. Adults have been collected in April and May. The specimenscollected by Kamath and Gupta came from mixed vegetation in coniferous forest (Gupta, 1975: (appendix)20,22). DISTRIBUTION (Map 9, p. 35). Along the southern flank of the Himalayas and south-east into Burma,between about 2000 and 3000 m. MATERIAL EXAMINED Holotype cf , Nepal: 2756'N, 85WE, 3030 m, 23-29.V.1967 (Can. Nepal Exped.) (BRI). Paratypes 13 $ . Burma: 5 $ , NE. , Kambaiti, 2000 m, 4, 6 & 7.iv. 1934 (Malaise) (NR); 1 $ , Mt Victoria,2800 m, 27.iv.1938 (Heinrich) (TC). India: 1 $, Himachal Pradesh, Dalhousie, 2132 m, 29.iv.1971 (Ram)(GC); 1 $, Himachal Pradesh, Kalatop, 2438 m, S.v.1971 (Kamath) (GC); 1 $, Himachal Pradesh, Simla(Nurse) (BMNH) (paralectotype of Cidaphurus flavomaculatus Cameron). Nepal: 2 $, 2756'N, 8500'E,3030 m, 23-29.V.1967 (Can. Nepal Exped.) (BRI); 2 $, 2800'N, 8500'E, 21-23.V.1967 (Can. NepalExped.)(BRI). Banchus sanjozanus Uchida (Figs 52, 76, 118)Banchus volutatorius var. sanjozanus Uchida, 1929: 184. Holotype cf , JAPAN (EIHU) [examined]. DESCRIPTION. Female. See 'Remarks'. Male. Width of lower face 0-86 times vertical length of eye. Malar space 0-50-0-60 times basal width ofmandible. Maxillary palp as in Fig. 118; segment 4 flattened and considerably widened; segment 5 about0-6 as long as 4. Antennal flag setae (similar to Fig. 128) upright, long, flattened, 2 per segment, arisingfrom a polished trough. Scutellar spine (Fig. 52) about as long as scutellum. Mesopleuron and mesoscutumcoriaceous, with moderately fine punctures, on mesopleuron separated by a little more than theirdiameter. Hind femur 5-15-5-60 times as long as deep. Fore wing length 10-2-10-4 mm. Tergite 1 of gaster2-10-2-22 times as long as broad, its dorsal profile as in Fig. 76. Tergite 3 with a crease along its anterior 0-4separating laterotergite. Colour: black and reddish, with some yellow. Face entirely reddish yellow or yellow with a narrowmedian black stripe. Antenna with scape and pedicel blackish dorsally, yellow or reddish yellow ventrally.Flagellum reddish, darkened dorsally, especially first few segments and distal 0-3. Maxillary palp withsegments 1 , 2 and 3 reddish yellow , 4 and 5 blackish . Posterior part of propodeum reddish , entirely or withposterior margin narrowly black. Hind coxa blackish with a large (but not sharply defined) reddish dorsalpatch and at least some reddish ventrally. Hind femur reddish with a blackish ventral mark. Segments ofgaster each reddish, with some blackish anteriorly, especially on tergites 1 and 2. REMARKS. Closely related to volutatorius. I have seen no females which I can associate with the males,although females were recorded by Uchida (1931: 52). BIOLOGY AND HOSTS. Unknown. Adults have been collected in August.DISTRIBUTION (Map 2, p. 16). Japan. MATERIAL EXAMINED (2 cf) Japan: 1 cf , Sanjodake, Yamato, 9.viii.l913 (Isshiki) (EIHU) (holotype); 1 cf , Hokkaido, Mt Yubari,ll.viii.1966 (Kusigemati) (KC). Banchus tholussp. n. (Figs 28, 53, 77, 98, 119) DESCRIPTION. Female. Width of lower face 0-70-0-80 times vertical length of eye. Malar space 0-65-0-85times basal width of mandible. Maxillary palp as in Fig. 28. Antenna with apical segments longer thanbroad. Scutellar spine (Fig. 53) about as long as scutellum. Mesopleuron shining, very weakly sculptured,with moderate punctures, separated by more than their diameter. Hind femur 6-60-7-05 times as long asdeep. Fore wing length 10-8-12-2 mm. Tergite 1 of gaster 1-90-2-20 times as long as broad, its dorsal profileas in Fig. 77. Gaster subcylindrical, only weakly compressed apically, not reaching to tips of fore wings(when folded back). Tergite 3 with a crease along its whole length separating laterotergite. Tergite 7 andsternite 6 as in Fig. 98. Colour: variable, from brown and yellowish cream to largely reddish orange (see 'Remarks' below).Face yellowish to reddish orange with orbits yellow. Antenna blackish, with scape, pedicel, and sometimesbase of flagellum, yellowish or reddish ventrally. Maxillary palp entirely reddish yellow. Posterior part of 38 M. G. FITTON propodeum pale yellow cream with a postero-median brown area or entirely reddish. Hind coxa yellowishcream and brown or entirely reddish. Hind femur entirely reddish. Segments of gaster reddish withsegment 1 pale cream or yellowish anteriorly, and sometimes with segments 1 to 4 each with a darkbrownish transverse median band. Male. Width of lower face 0-80 times vertical length of eye. Malar space 0-70 times basal width ofmandible. Maxillary palp as in Fig. 119; segment 4 moderately widened and flattened; segment 5 about aslong as 4. Antennal flag setae not differentiated (antennae in poor condition and proper observationdifficult). Hind femur 7-15 times as long as deep. Fore wing length 11-5 mm. Tergite 1 of gaster 1-60 timesas long as broad. Tergite 3 with a crease along its whole length separating laterotergite. Colour: brown and yellowish cream, with some reddish. Antenna brownish dorsally, brownish yellowventrally. Maxillary palp entirely reddish yellow. Posterior part of propodeum reddish yellow, narrowlydark brownish postero medially. Hind coxa yellowish cream and brown. Hind femur entirely reddish.Segments of gaster reddish, with segment 1 pale creamy yellow anteriorly and 1 and 2 more or less brownishmedially. REMARKS. The females show a considerable range of variation in colour, the specimens from Burmaresembling those of tumidus, while those from the Philippines are almost entirely reddish orange. Thespecimens from Sumatra and Java have an intermediate coloration. Unfortunately only a single male (fromBurma) is known. The females might represent more than one species. See also 'Remarks' under B.tumidus. BIOLOGY AND HOSTS. Virtually unknown. Adults have been collected in December, March, April, May andJune. The specimen from Mount Data was collected in oak forest. DISTRIBUTION (Map 2, p. 16). On mountains in Burma, Sumatra, Java and the Philippines. MATERIAL EXAMINED Holotype cf , Burma: NE., Kambaiti, 1800 m, ll.vi.1934 (Malaise) (NR). Paratypes 10 <J>. Burma: 1 $, Mt Victoria, 1400 m, iii.1938 (Heinrich) (GC) (paratype of Banchustumidus Chandra & Gupta); 2 $,Maymyo, 800m, xii. 1937 (HeinricK) (TC) (paratypes of Banchus tumidusChandra & Gupta). Sumatra: 1 $, Sungei Kumbang, Korinchi, 1370 m, iv.1914 (Robinson & Klass)(BMNH). Java: 1 $, Gedeh, Tjibodas, 1700 m, xii.1935 (Lieftinck) (TC). Philippines: 4 $, Mindoro,Hong, Mt Halcon, 1370 m, 9, 10 & ll.v.1954 (Townes) (TC); 1 $, Mt Data, 2380 m, 31.xii.1952 (Townes)(TC). Banchus tumidus Chandra & Gupta (Figs 29, 54, 78, 99, 120)Banchus tumidus Chandra & Gupta, 1977: 183. Holotype $, INDIA (GC) [examined]. DESCRIPTION. Female. Width of lower face 0-75-0-80 times vertical length of eye. Malar space 0-65-0-75times basal width of mandible. Maxillary palp as in Fig. 29. Antenna with apical segments longer thanbroad. Scutellar spine (Fig. 54) about as long as scutellum. Mesopleuron shining, only very weaklysculptured, with moderate punctures, separated by more than their diameter. Hind femur 6-80-7-05 timesas long as deep. Fore wing length 10-5-11-9 mm. Tergite 1 of gaster 1-85-1-95 times as long as broad, itsdorsal profile as in Fig. 78. Gaster subcylindrical, only weakly compressed posteriorly, not reaching to tipsof fore wings (when folded back). Tergite 3 with a crease along its whole length separating laterotergite.Tergite 7 and sternite 6 as in Fig. 99. Colour: yellowish cream, reddish and brown. Face yellowish. Antenna dark brown, with scape, pediceland base of flagellum yellowish ventrally. Maxillary palp entirely reddish yellow. Posterior part ofpropodeum yellowish cream with a postero-median brown area. Hind coxa yellowish cream and brown.Hind femur entirely reddish. Segments of gaster reddish yellow, with segment 1 pale anteriorly andsegments 1 to 4 each with a brownish transverse median band. Male. Width of lower face 0-80-0-85 times vertical length of eye. Malar space 0-60-0-70 times basalwidth of mandible. Maxillary palp as in Fig. 120; segment 4 very considerably widened and flattened;segment 5 about 0-76 as long as 4. Antennal flag setae not differentiated. Hind femur 6-70-6-90 times aslong as deep. Fore wing length 10-8 mm. Tergite 1 of gaster 1-70-1-75 times as long as broad. Tergite 3 witha crease along its whole length separating laterotergite. Colour: creamy yellow and brown. Antenna dark brown dorsally, yellow ventrally. Maxillary palpreddish yellow with segment 4 partly blackish (Fig. 120). Posterior part of propodeum reddish yellow,brownish postero-medially. Hind coxa creamy yellow and brown. Hind femur entirely reddish. Segmentsof gaster reddish yellow, with segment 1 pale cream anteriorly and 1 and 2 brownish medially. ICHNEUMON-FLY GENUS BANCHUS 39 REMARKS. This species is very closely related to tholus. The males are easily distinguished on the form of themaxillary palp, but not otherwise. The females are impossible to separate on morphological characters;those placed here as tumidus have segment 4 of the maxillary palp very weakly bicoloured (it is stronglybicoloured in males) . The females of tholus show a much wider range of variation in colour and the divisionbetween the two species was made largely on the basis of geography. See also 'Remarks' under tholus. BIOLOGY AND HOSTS. Virtually unknown. Adults have been collected in April, May, June, September andNovember. DISTRIBUTION (Map 2, p. 16). On the southern flank of the western Himalayas between 600 and 2000 m. MATERIAL EXAMINED (9 9 , 2 cf ) India: 1 $ , Uttar Pradesh, Kumaon Himalaya, Jeolikote, 1219 m, 12.ix.1965 (Tikar) (GC) (holotype); 1$ , Himachal Pradesh, Khajjiar, 1920 m, 24. vi. 1965 (Joseph) (GC); 1 $ , Himachal Pradesh, Khajjiar, 1828m, 30.iv. 1971 (Ram) (GC); 2 $ , Himachal Pradesh, Manali, 1828 m, 17 & 20. v. 1970 (Ram & Gulati) (GC);1 $, Uttar Pradesh, 1949 (Bianchi} (TC); 2 $, Uttar Pradesh, Dehra Dun, 600 m, 27.xi.1965 (Gupta}(GC); 1 cT, Uttar Pradesh, Kumaon Hills, Bhowali, 1700 m, 5-8.vi.1968 (Gupta) (GC); 1 cf, UttarPradesh, Dehra Dun, 7.iv.l967 (Tikar) (GC); 1 $, Uttar Pradesh, Gharwal Himalaya, Phata, 1524 m,l2.v.l961(Kamath)(GC). Banchus turcator Aubert(Figs. 30, 55, 79, 100, 121) Banchus turcator Aubert, 1978: 157. [Unavailable name published conditionally (Article 15 of the Code).]Banchus turcator Aubert, 1981: 18. Holotype cf , TURKEY (AC) [examined]. DESCRIPTION. Female. Width of lower face 0-90-1-00 times vertical length of eye. Malar space 0-50-0-55times basal width of mandible. Maxillary palp as in Fig. 30. Antenna with apical segments longer thanbroad. Scutellar spine (Fig. 55) usually very small, sometimes absent or reasonably well developed.Mesopleuron and mesoscutum strongly coriaceous, with moderate to strong punctures, on mesopleuronseparated by slightly less than their diameter. Hind femur 4-70-5-60 times as long as deep. Fore wing length8-0-10-7 mm. Tergite 1 of gaster 1-45-1-90 times as long as broad, its dorsal profile as in Fig. 79. Gasterrelatively weakly compressed, reaching about to tips of fore wings (when folded back). Tergite 3 with acrease along its anterior 0-3 separating laterotergite. Tergite 7 and sternite 6 as in Fig. 100. Colour: mainly black and yellow, the yellow areas on the appendages tending to reddish. Face varyingfrom yellow with a black median stripe to black with yellow orbital stripes. Antenna with scape and pedicelblack dorsally, yellow ventrally. Flagellum entirely orange except that it is dark at extreme apex. Maxillarypalp reddish yellow, with segments 1 and 5 wholly, 2 proximally and 4 distally dark. Posterior part ofpropodeum black with a yellow band (sometimes interrupted medially) along posterior transverse carina.Hind coxa black, entirely or with a yellow dorsal patch. Hind femur black, yellow or reddish yellowproximally and distally, the yellowish area sometimes extending along most of dorsal surface. Segments ofgaster each black anteriorly, yellow (sometimes broadly so) posteriorly. Male. Width of lower face 0-90-1-05 times vertical length of eye. Malar space 0-45-0-55 times basalwidth of mandible. Maxillary palp as in Fig. 121 ; segment 4 flattened and considerably widened; segment 5about as long as 4. Antennal flag setae (similar to Fig. 127, but setae more widened toward their apices)upright, flattened and widened, 2 per segment, arising from a polished trough. Hind femur 4-60-5-30 timesas long as deep. Fore wing length 8-7-10-0 mm. Tergite 1 of gaster 1-55-1-65 times as long as broad. Tergite3 with a crease along its anterior 0-3 separating laterotergite. Colour: black and yellow, with some reddish. Antenna with scape and pedicel black dorsally, yellowventrally. Flagellum orange, darkened distally. Maxillary palp with segment 1 yellowish (sometimes partlydarkened), 2 and 3 wholly and 4 proximally reddish yellow, 4 distally and 5 wholly blackish or brownish.Posterior part of propodeum black with a yellow band along posterior transverse carina, the bandsometimes broken medially, reduced to two lateral spots or even absent. Hind coxa black, usually with aventral and sometimes also a dorsal yellow spot. Hind femur yellowish red with a large medial black area.Segments of gaster each black anteriorly, yellow posteriorly. REMARKS. This species has been confused with pictus in collections. It is, however, quite unrelated asshown, for instance, by the rather different flag setae. BIOLOGY AND HOSTS. Unknown. Adults have been collected in April, May and June.DISTRIBUTION (Map 4, p. 20). Turkey and in the mountains of Tadzhikistan. 40 M. G. FITTON MATERIAL EXAMINED (5 9 , 9 cf ) Turkey: 1 of, Guriin, 12-15.vi.1976 (Heinrich) (AC) (holotype); 1 $, Alem-Dag, 600 m, 26-30. vi.,(Demelt) (FSA); 1 cf , Antalya, Termessus, 9.V.1968 (Halliri) (NR); 1 $ , Erzurum, 20 km tspir to Ikizdererd, 1700 m, 2.vi.l962 (Guichard & Harvey) (BMNH); 1 cf , Gumus,ane, near Maden, 1800 m, 29.V.1962(Guichard & Harvey) (BMNH); 1 $, Isik-Dag, 1200 m, vi.1966 (Real) (FSA); 1 $, Istanbul, 2Q.iv.(deGaullecoll.) (MNHN); 1 cf , Kizilcahamam, 1000m, 26-28.V.1964 (FSA); 1 $, 1 cf , Nevehir, Urgiip, 4 &6.vi.l978 (Schwarz) (AC); 1 cf, Zara Taiger (UZM). U.S.S.R.: 2 cf , Gissarskiy Khrebet, K-K chanch-ovon Aliche, 25.iv.1960 (Malyavin) (MLSU); 1 cf , Khorog region, Shugnansk Khrebet, 2600 m, 5.vi.l956(Zhelokhovtsev) (MLSU). Banchus volutatorius (Linnaeus)(Figs 3, 7, 8, 31, 56, 80, 101, 122, 128) Ichneumon volutatorius Linnaeus, 1758: 562. Lectotype cf , EUROPE (LSL), fixed by Roman, 1932: 14 [examined] . Ichneumon venator Linnaeus, 1758: 564. Type(s) $, EUROPE (lost).Ichneumon umbellatarum Schrank, 1786: 261. Type(s) $ [not cf as stated by Schrank], WEST GERMANY (lost). Syn. n. Ichneumon certatorThunberg, 1822: 266; 1824: 322. Holotype $, SWEDEN (UU) [not examined].Banchus monileatus Gravenhorst, 1829: 393. Lectotype $, POLAND (ZI), designated by Townes & Townes, 1978: 532 [examined]. Syn. n.Banchus farrani Curtis, 1836: 588. Lectotype cf , IRELAND (NMV), designated by Fitton, 1976: 322 [examined]. Banchus calcaratus Szepligeti, 1910: 186. Holotype $, HUNGARY (TM) [examined].Banchus volutatorius var. alticola Schmiedeknecht, 1910: 1931 . Syntypes 9 , cf , EAST GERMANY (7MNHU) [not examined]. [Junior secondary homonym of Cidaphurus alticola Ashmead, 1901: 148.]Banchus obscurus Meyer, 1926: 263. Type(s) $ [not cf as stated by Meyer], U.S.S.R. (destroyed). NOMENCLATURE. The identity of venator was established by Fitton (19786: 375). Gravenhorst (1829: 389) synonymised /. umbellatarum with B. falcatorius but subsequently it has beenincluded in the synonymy of B. compressus (= dilatatorius) (e.g. Aubert, 1978: 152 [with the date givenincorrectly as 1802]) without, however, being used as the valid name for that species, despite its seniority.The description does not fit dilatatorius; it could apply to some females of falcatorius or volutatorius. Iconsider that it best fits certain females of volutatorius, with which it is here synonymised. Although the types of var. alticola were not located a male determined by Schmiedeknecht, and from thetype-locality, was examined (ZSBS). Meyer undoubtedly mistook the sex of the type-material (probably a single specimen) of obscurusbecause the face coloration cannot apply to any male Banchus. The description fits perfectly the female ofvolutatorius and the host recorded by Meyer (Lacanobia oleracea) is one of those known for this species. DESCRIPTION. Female. Width of lower face 0-85-0-90 times vertical length of eye (Fig. 3). Malar space0-55-0-60 times basal width of mandible. Maxillary palp as in Fig. 31. Antenna with apical segments longerthan broad. Scutellar spine (Fig. 56) distinct, often almost 0-5 as long as scutellum. Mesopleuron andmesoscutum coriaceous, with moderate punctures, on mesopleuron separated by about their diameter.Hind femur 4-85-5-35 times as long as deep. Fore wing length 8-1-9-8 mm. Tergite 1 of gaster 1-80-2-05times as long as broad, its dorsal profile as in Fig. 80. Gaster (Figs 7, 8) moderately compressed, notreaching to tips of fore wings (when folded back). Tergite 3 with a crease along its anterior 0-4 separatinglaterotergite. Tergite 7 and sternite 6 as in Fig. 101. Colour: predominantly black, with legs largely reddish. Face black, occasionally with brownish orbitalmarks. Antenna black, with scape, pedicel and proximal flagellar segments reddish brown ventrally.Maxillary palp black, with segments 2 and 3 wholly and 4 proximally reddish or brownish. Posterior part ofpropodeum black, often entirely, sometimes with reddish areas (of varying extent) on segments 1, 2 and 3(in extreme cases with posterior margins of these segments narrowly yellowish). Male. Width of lower face 0-90-1-00 times vertical length of eye. Malar space 0-45-0-60 times basalwidth of mandible. Maxillary palp as in Fig. 122; segment 4 considerably widened and flattened; segment 5about 0-5 as long as 4. Antennal flag setae (Fig. 128) upright, very long, flattened and widened, 2 persegment, arising from a polished trough. Hind femur 4-85-5-40 times as long as deep. Fore wing length8-0-9-9 mm. Tergite 1 of gaster 1-65-2-00 times as long as broad. Tergite 3 with a crease along its anterior0-4 separating laterotergite. Colour: black and yellow, with some reddish. Antenna black with flagellum yellowish ventrally except at ICHNEUMON-FLY GENUS BANCHUS 41 its distal apex. Maxillary palp with segment 1 yellow, 2 and 3 wholly and 4 proximally reddish yellow, 4distally and 5 wholly blackish. Posterior part of propodeum black or with a yellow mark or marks along(usually just anterior to) posterior transverse carina. Hind coxa black, entirely or with small yellow marksexterno-laterally and/or postero-ventrally. Hind femur reddish yellow, rarely with a blackish ventral mark.Segments of gaster each black anteriorly, yellow posteriorly, tergites 1, 2 and 3 very broadly yellow andwith anterior area often partly (sometimes largely) reddish. REMARKS. The male is similar, superficially, to falcatorius but can be separated readily by the number offlag setae and the lack (usually) in falcatorius of a distinct spine on the scutellum. The female has a gasterwhich is quite unlike that of falcatorius and would not be confused with it. The female is, however,frequently confused with palpalis, although the maxillary palp is different in form and volutatoriusgenerally has a much shorter scutellar spine. BIOLOGY AND HOSTS. Adults have been collected mainly in the period from mid- June to early August, butthere are a few records as early as mid- April and as late as September. However, there is no evidence of twogenerations in the distribution of the records and information from rearings also indicates that volutatoriusis univoltine. I have seen 58 reared specimens, many with detailed and reliable host data. The hosts (indecreasing order of number of rearings and reared specimens) are: Anarta myrtilli (Linnaeus) (Noctuidae),12 $, 11 cf from 12 rearings, Great Britain, Germany, the Netherlands and Switzerland; Lacanobiaoleracea (Linnaeus) (Noctuidae), 14 $ , 1 cf from 5 rearings, Great Britain; Mamestra brassicae (Linnaeus)(Noctuidae), 1 $, 1 cf from 2 rearings, Great Britain; Hadena compta (Denis & Schiffermuller)(Noctuidae), 3 $, 1 cf from 1 rearing, Denmark; Ceramica pisi (Linnaeus) (Noctuidae), 1 $, 1 cf from 1rearing, Great Britain; Heliothis viriplaca (Hufnagel) (Noctuidae), 2 cf from 1 rearing, Germany; Xestiaxanthographa (Denis & Schiffermuller) (Noctuidae), 1 cf from 1 rearing, Great Britain; Habrosynepyritoides (Hufnagel) (Thyatiridae), 1 $ from 1 rearing, Germany; Opisthograptis luteolata (Linnaeus)(Geometridae), 1 $ from 1 rearing, Germany; and Zygaena ephialtes (Linnaeus) (Zygaenidae), 1 cf from1 rearing, Germany. The literature additionally records these noctuid hosts: Agrotis segetum (Denis & Schiffermuller)(Meyer, 1927a: 81), Lycophotia porphyrea (Denis & Schiffermuller) (Leonardi, 1928: 83; Meyer, 1934:231); Lacanobia suasa (Denis & Schiffermuller) and L. contigua (Denis & Schiffermuller) (Ljungdhal,1918: 82; Meyer, 1934: 231); and Bena prasinana (Linnaeus) (Hedwig, 1939: 22). Zorin & Zorina (1929)give some biological information on the association with Lacanobia oleracea. DISTRIBUTION (Map 10, p. 42). Widespread in northern and central Europe, extending into Turkey, and inthe U.S.S.R. occurring as far south as Alma-Ata and as far east as the Chitinskaya Oblast. MATERIAL EXAMINED (486 $, 301 cf) Austria, Belgium, Denmark, East Germany, Finland, France, Great Britain, Hungary, Ireland, Nether-lands, Norway, Poland, Rumania, Sweden, Switzerland, Turkey, U.S.S.R., West Germany (AC, BC,BMNH, BRI, CM, FSA, HC, IBMPP, IEUB, IP, ITZ, IZPAN, JC, KHC, LELW, MHN, MLSU,MNHN, MUM, NMB, NR, PC, RNH, RSM, TM, UM, USNM, UZI, UZM, VRC, ZC, ZI, ZIL, ZIM,ZMU, ZSBS). Banchus zonatus Rudow (Figs 32, 57, 58, 81, 102, 123, 126) Banchus zonatus Rudow, 1883o: 57. Type(s) $, EUROPE 'Sudeuropa' (?JPM) [not examined].Banchus algericus Schmiedeknecht, 1910: 1927. Holotype $, ALGERIA (MNHU) [examined]. Syn. n. NOMENCLATURE. Apart from some exaggeration of size, the description of zonatus fits perfectly, and only,the female of this species. The synonymy is further confirmed by the type-locality 'Vaterland Sudeuropa'(the species was not described from Germany as stated by Aubert, 1978: 166). It has not been possible toobtain on loan material from the Rudow collection. Rudow's methods and the poor condition of hiscollection were criticized during his lifetime. The collection, still neglected, was seen recently byHorstmann (pers. comm.), who made notes on its contents, including the fact that three specimens stand asB. zonatus. However, nothing is known of the status or identity of these specimens. DESCRIPTION. Female. Width of lower face 1-00-1 -15 times vertical length of eye. Malar space 0-75-0-95times basal width of mandible. Maxillary palp as in Fig. 32. Antenna with apical segments broader thanlong. Scutellar spine (Figs 57, 58) about 0-5 as long as scutellum or almost absent (see 'Remarks')-Mesopleuron and mesoscutum coriaceous with moderately strong punctures, on mesopleuron separatedby less than their diameter. Hind femur 5-30-6-15 times as long as deep. Fore wing length 11-1-13-5 mm. M. G. FITTON Map 10 Distribution of Banchus volutatorius. Tergite 1 of gaster 1-35-1-65 times as long as broad, its dorsal profile as in Fig. 81. Gaster stronglycompressed from posterior part of segment 3, reaching beyond tips of fore wings (when folded back).Tergite 3 with a crease along its anterior 0-4 separating laterotergite. Tergite 7 and sternite 6 as in Fig. 102. Colour: black and yellow. Face yellow with a black median stripe. Antenna black (sometimes brownishdistally), scape with a yellow patch beneath. Maxillary palp dark brown or black. Posterior part ofpropodeum black with a yellow stripe along posterior transverse carina. Hind coxa black, usually with ayellow dorsal patch. Hind femur black, yellow at extreme base, distally, and often dorsally. Segments ofgaster each black anteriorly, yellow posteriorly. Male. Width of lower face 1-00-1-20 times vertical length of eye. Malar space 0-60-0-90 times basalwidth of mandible. Maxillary palp as in Fig. 123; segment 4 flattened and slightly widened; segment 5 about0-8 as long as 4. Antennal flag setae (Fig. 126) upright, moderately long, flattened, 2 per segment, arisingfrom a polished trough. Hind femur 5-35-6-65 times as long as deep. Fore wing length 10-5-14-0 mm.Tergite 1 of gaster 1-65-1-90 times as long as broad. Tergite 3 with a crease along its anterior 0-4 separatinglaterotergite. Colour: back and yellow. Antenna black, scape and pedicel (and sometimes flagellum segment 1proximally) yellow ventrally, flagellum sometimes slightly brownish ventrally. Maxillary palp blackish orbrownish. Posterior part of propodeum yellow anteriorly, black posteriorly, the yellow varying from astripe along posterior transverse carina to covering almost entire area. Hind coxa black with a dorsal andoften a postero- ventral yellow patch. Hind femur yellow, blackish laterally and ventrally towards base. ICHNEUMON-FLY GENUS BANCHUS Map 1 1 Distribution of Banchus zonatus. Segments of gaster each black anteriorly, yellow posteriorly, the yellow less variable in extent than infemales, occupying about 0-5 of each tergite. REMARKS. This is the largest species of the genus and the appearance of the females in particular is verystriking. B. zonatus seems to be most closely related to dilatatorius and moppiti. The specimens from the eastern Mediterranean (Cyprus and Israel) consistently have the scutellar spinevery small or virtually absent and it is possible that they represent a separate species. However, this isdifficult to investigate without additional material. There are no specimens from peninsular Italy, Greeceor Turkey in collections. BIOLOGY AND HOSTS. Unknown. Dates of collection range from early September to mid-December. Thereis a single male with a date in March, unfortunately without an intelligible locality but with the intriguingdata'Salixcaprea'. DISTRIBUTION (Map 11, p. 43). A 'Mediterranean' species. Meyer (19276: 291) records the species fromthe Caucasus but I have seen no specimens to substantiate this north-eastward extension of its distribution. MATERIAL EXAMINED (28 $ , 21 C?) Algeria, Cyprus, France, Israel, Italy, Morocco, Spain, Tunisia (AC, BMNH, FSA, IEE, MHN,MNHN, MNHU, NR, TAU, TC, UZM). Species excluded from Banchus The following Old World species were described incorrectly in Banchus. Banchus armitfafiisMorley, 1913: 254. Holotype cf, NICOBAR ISLANDS (BMNH) [examined].Identity. Phytodietus armillatus (Morley). In Kaur & Jonathan's (1979) treatment of Oriental Phyto- 44 M. G. FITTON dietus the holotype runs to P. alasuffuscus Kaur & Jonathan. However, it differs in some details,particularly in the degree of constriction of tergite 1 of the gaster. Banchus elator Fabricius, 1804: 128. LECTOTYPE cf, AUSTRIA (UZM), here designated (selected byG. E. J. Nixon) [examined].Identity. Earinus elator (Fabricius) comb. n. (Braconidae) (see Nixon, 1986, in press). Banchus histrio Fabricius, 1798: 234. Lectotype cf, EAST GERMANY (UZM), designated by Horstmann,19826: 243 [not examined].Identity. Lissonota histrio (Fabricius) (Horstmann, 19826: 243). Itaiic/iifsqiiadratorSchellenberg, 1802: 21. Type(s) $, SWITZERLAND (lost). Identity. Megarhyssa quadrator (Schellenberg) comb. n. Schellenberg's description and figures areobviously of a rhyssine (Pimplinae). The species concerned corresponds to that named as M. citraria(Olivier) in the BMNH collection. The two recent catalogues (Oehlke, 1967; Aubert, 1969) coveringEuropean Megarhyssa differ in so many details that it is not clear what name is the valid one for this species.Whatever it is, quadrator will be a junior synonym. Gravenhorst's invitation (1829: 959) to comparequadrator with M. superba (Schrank) has been overlooked by all who have dealt subsequently with theEuropean Rhyssini. Banchus robustusRudow, 18836: 246. Type(s) $, EAST GERMANY (?JPM) [not examined]. Identity. Unknown, it remains a nomen dubium. It is excluded from Banchus because of the length of theovipositor: 'Legestachel fast halb so lang als Hinterlieb'. Banchus spinipes Panzer, 1800: 17. Type(s) $, GERMANY (lost).Identity. Junior synonym of Cephus pygmeus (Linnaeus) (Cephidae) (Muche, 1981: 283). Banchus tomentosusGravenhorst, 1829: 376, Holotype $, EAST GERMANY (ZI) [examined].Identity. Exetostes tomentosus (Gravenhorst) (Townes, Momoi & Townes, 1965: 235). Banchus viUosulus Gravenhorst, 1807: 267. Syntypes [?number] [?$], no type-locality (lost). Identity. The description of this species fits a cynipoid rather than an ichneumonid. Gravenhorst himselfcompared it to Ibalia and did not refer to it in any of his subsequent works on ichneumonids. It is unlikelythat any type-material will be found. The general nature of the description makes application of the nameto a particular species difficult, but it is desirable to formally transfer the name to the Cynipoidea so thatworkers on that group can consider its identity. The species is here placed as a nomen dubium in Andricus(comb, n.) (Cynipidae), to the agamic females of which the description seems best to apply. Banchus viridator Fabricius, 1804: 127. Syntypes $, AUSTRIA (UZM) [one of four putative syntypesexamined]. Fabricius cited Banchus spinipes Panzer as a synonym of his new species viridator. I can see no reason forhim not to have used Panzer's name for the species. Whatever the reasons, it complicates consideration ofthe name. It could be treated as an independent species, as published in synonymy, or as a replacementname (in which case the type-specimens would be those of B. spinipes rather than those cited above). Identity. Junior synonym of Cephus pygmeus (Linnaeus) (Cephidae) (Muche, 1981: 283). The following species have been placed incorrectly in Banchus at some time. Ichneumon annulatus Fabricius, 1793 is a pompilid, Pompilus (Episyron) annulatus (Fabliau's) (Schulz,1912: 73). Ichneumon cultellator Fabricius, 1793 is a junior synonym of Ibalia leucospoides (Hochenwarth) (Iba-liidae) (Kerrich, 1973: 73). Ichneumon fornicator Fabricius, 1781 is a species oiExetastes (Townes, Momoi & Townes, 1965: 229). Ichneumon gravidator Linnaeus, 1758 is a species of Proctotrupes (Proctotrupidae) (Fitton, 19786: 378;Townes & Townes, 1981: 179). Ichneumon varius Fabricius, 1793 is a junior synonym of Theronia atalantae Poda (Aubert, 1978: 167). Ichneumon vigilatorius Panzer, 1804 was synonymised with Banchus falcatorius by Gravenhorst (1829:390). This was later rejected in favour of synonymy with compressus (Aubert, 1978: 153) but was reinstatedby Horstmann (19826: 238). However, Schaeffer's illustration (1767: pi. 61, fig. 6), to which Panzer'sdescription and name refer, clearly shows an ichneumonid with segment 1 of the gaster strongly petiolateand which probably belongs in the Ichneumoninae. Although I cannot find an ichneumonine with all ICHNEUMON-FLY GENUS BANCHUS 45 details of the colour pattern correct, I think it best to place the species as a nomen dubium in Ichneumon(comb. rev.). Acknowledgements I am grateful to the following for the loan of material and provision of information relating to collections intheir care: A. Albrecht, A. A. Allen, A. Athanasov, J. F. Aubert, C. Baroni Urbani, J. R. Barron, R.Bauer, C. Besuchet, A. Brindle, R. W. Carlson, R. Danielsson, G. Delrio, P. Dessart, E. Diller, M. Dorn,R. Eck, W. A. Ely, P. P. d'Entreves, M. Fischer, D. Gerling, V. K. Gupta, R. Hinz, K. Horstmann, A. G.Irwin, R. Jussila, D. R. Kasparyan, S. Kelner-Pillault, J. Kolarov, E. Konigsmann, K. Kusigemati, W.Kuslitzky, E. Mellini, E. Mingo, A. Nakanishi, A. Neboiss, J. Oehlke, P. Oosterbroek, C. O'Toole, J.Papp, P. I. Persson, B. Petersen, C. Pisica, R. Poggi, W. J. Pulawski, G. van Rossem, J. Sawoniewicz, H.Schnee, M. R. Shaw, M. Sorg, M. Suwa, C. Thirion, H. K. Townes, M. C. Townes, L. Zimina, L.Zombori, and K. W. R. Zwart. I am indebted to my colleagues in the BMNH for discussion of variouspoints; help with translations; and technical assistance. References Arthur, A. P. & Ewen, A. B. 1975. Cuticular encystment: a unique and effective defense reaction by cabbage looper larvae against parasitism by Banchus flavescens (Hymenoptera: Ichneumonidae). Annals of the Entomological Society of America 68: 1091-1094.Ashmead, W. H. 1901. Some insects of the Hudsonian zone in New Mexico. II. Hymenoptera. Psyche, Cambridge 9: 147-148.1906. Descriptions of new Hymenoptera from Japan. Proceedings of the United States National Museum 30: 169-201.Aubert, J. F. 1969. Les Ichneumonides ouest-palearctiques et leurs holes. 1 (Pimplinae, Xoridinae, Acaenitinae). 302pp. Paris.1978. Les Ichneumonides ouest-palearctiques et leurs holes. 2 (Banchinae, et suppl. aux Pimplinae). 318 pp. Paris. 1981. Syllogismes, illogismes et innovations chez les Ichneumonides. Bulletin de la Societe Entomolo- gique de Mulhouse 1981: 17-22.Aurivillius, C. 1890. Gronlands Insektfauna. I. Lepidoptera, Hymenoptera. Kungliga Svenska Vetenskap- sakademiens Handlingar 15 (IV) (1): 1-33. Bajari. E. 1960. Furk6szdarazs-Alkatuak I. Ichneumonoidea I. Fauna Hungariae 54, 266 pp.Barron, J. R. 1976. Systematics of Nearctic Euceros (Hymenoptera: Ichneumonidae: Eucerotinae). Le Naturaliste Canadien 103: 285-375.Beirne, B. P. 1941 . A consideration of the cephalic structures and spiracles of the final instar larvae of the Ichneumonidae (Hym.). Transactions of the Society for British Entomology 7: 123-190.Bischoff, H. 1930. Entomologische Ergebnisse der Deutsch-Russischen Alai-Pamir-Expedition 1928 (I). 3. Hymenoptera I (Sphecidae, Vespidae, Scoliidae, Mutillidae, Chrysididae, Ichneumonidae, Evaniidae). Mitteilungen aus dem Zoologischen Museum in Berlin 16: 215-225.Bledowski, R. & Krairiska, M. K. 1926. Die Entwicklung von Banchus femoralis Thorns. (Hymenoptera, Ichneumonidae). Bibliotheca Universitatis Liberae Polonae (A) 16: 3-50.Bretherton, R. F., Goater, B. & Lorimer, R. I. 1983. Noctuidae [part]. In Heath, J. & Emmet, A. M. (Eds), The moths and butterflies of Great Britain and Ireland 10: 36-413.Cameron, P. 1904. Descriptions of new genera and species of Ichneumonidae from India (Hym.). (Continued). Zeitschrift fur systematische Hymenopterologie und Dipterologie 4: 337-347.Chandra, G. & Gupta, V. K. 1977. Ichneumonologia Orientalis. Part VII. The tribes Lissonotini and Banchini (Hymenoptera: Ichneumonidae: Banchinae). Oriental Insects Monographs 7, 290 pp. Delhi.Christ, J. L. 1791. Naturgeschichte, Klassification und Nomenclatur der Insekten vom Bienen, Wespen und Ameisengeschlecht. 535 pp. Frankfurt am Main.Constantineanu, M. I. & Pisica, C. 1959. Specii de Pimpline (Pimplinae Cresson) noi pentru fauna Republicii Populare Romine. In Omagiu lui Traian Sdvulescu cuprilejul tmplinirii a 70 de ani: 187-192.1960. Ichneumonide din Valea Bistrijrei dintre Poiana Teiului sj Vaduri (subfamiliile: Pimplinae, Lissonotinae, Xoridinae, Acoenitinae, Colyriinae i Banchinae). Analele Stiintifice de Universitdti "Al. I. Cuza" din lasl. (New series) (II. Stiinte naturale) 6: 701-712.Cresson, E. T. 1865. Catalogue of Hymenoptera in the collection of the Entomological Society of 46 M. G. FITTON Philadelphia, from Colorado Territory. Proceedings of the Entomological Society of Philadelphia 4: 242-313. Curtis, J. 1836. British Entomology 13: 578-625. London.Cushman, R. A. 1937. Revision of the North American species of ichneumon-flies of the genus Exetastes Gravenhorst. Proceedings of the United States National Museum 84: 243-312.Dalla Torre, C. G. de. 1901. Catalogus Hymenopterorum 3 (1): 1-544.Ewen, A. B. & Arthur, A. P. 1976. Cuticular encystment in three noctuid species (Lepidoptera): induction by acid gland secretion from an ichneumonid parasite (Banchusflavescens). Annals of the Entomological Society of America 69: 1087-1090.Fabricius, J. C. 1775. Systema Entomologiae. 832 pp. Flensburgi et Lipsiae. 1781. Species Insectorum 1, 552 pp. Hamburgii et Kilonii. 1787. Mantissa Insectorum 1, 348 pp. Hafniae. 1793. Entomologia Systematica 2, 519 pp. Hafniae. 1798. Supplementum Entomologiae Systematicae. 572 pp. Hafniae. 1804. Systema Piezatorum. 439 pp. Brunsvigae. Fitton, M. G. 1976. The western Palaearctic Ichneumonidae (Hymenoptera) of British authors. Bulletin of the British Museum (Natural History) (Entomology) 32: 301-373.19780. The Ichneumonidae (Hymenoptera) described by J. F. Ruthe. Zeitschrift der Arbeitsgemein- schaft Osterreichischer Entomologen 30: 75-79. 19786. The species of 'Ichneumon' (Hymenoptera) described by Linnaeus. Biological Journal of the Linnean Society 10: 361-383.Foerster, A. 1869. Synopsis der Familien und Gattungen der Ichneumonen. Verhandlungen des Naturhis- torischen Vereines der Preussischen Rheinlande und Westphalens 25: 135-221.Friederichs, K. 1936. Zur Phanologie einiger Parasiten der Forleule (Panolisflammea). Entomologische Zeitschrift 50: 89-91. Geoffrey, E. L. 1785. In Fourcroy, A. F. de (ed.), Entomologia Parisiensis 2: 233-544. Parish's.Gmelin, J. F. 1790. In Linnaeus, C., Systema Naturae 13th edn, 1 (5): 2225-3020. Lipsiae.Gravenhorst, J. L. C. 1807. Vergleichende Uebersicht des Linneischen und einiger neueren zoologischen Systeme. 476 pp. Gottingen. 1829. Ichneumonologia Europaea 3, 1097 pp. Vratislaviae. Gupta, V. K. 1971. Ichneumon hunting in India. 109 + 80 pp. Delhi. 1975. Ichneumonological explorations in India. 62 + 64 pp. Delhi. Gyorfi, J. 1944. Beitrage zur Kenntnis der Ichneumoniden Ungarns. III. Fragmenta Faunistica Hungarica 7: 103-107.Habermehl, H. 1922. Beitrage zur Kenntnis der palaearktischen Ichneumonidenfauna. Konowia 1: 266-282.1924. Beitrag zur Kenntnis der primaren und sekundaren Schmarotzerwespen der Kieferneule (Panolis flammea Schiff. = P. griseovariegata Goeze) (Hym.). Deutsche Entomologische Zeitschrift 1924: 183-184.Hedwig, K. 1939. Verzeichnis der bisher in Schlesien aufgefundenen Hymenopteren. Zeitschrift filr Entomologie 18 (3): 12-28.Hellen, W. 1949. Zur Kenntnis der Ichneumonidenfauna der Atlantischen Inseln. Commentationes BiologicaeS(lT):l-23. 1961. Ichneumonidenfunde aus Madeira. Notulae Entomologicae 41: 35-38. Hensch, A. 1928. Beitrag zur Kenntnis der jugoslawischen Ichneumonidenfauna. Konowia 7: 99-112.Horstmann, K. 1982a. Typenrevision einiger von A. Hensch beschriebener Ichneumoniden-Arten (Hymenoptera). Zeitschrift der Arbeitsgemeinschaft Osterreichischer Entomologen 33: 81-88.19826. Revision der von Panzer beschriebenen Ichneumoniden-Arten (Hymenoptera). Spixiana 5: 231-246.Kaur, R. & Jonathan, J. K. 1979. Ichneumonologia Orientalis. Part VIII. The tribe Phytodietini from India (Hymenoptera: Ichneumonidae). Oriental Insects Monographs 9: 276 pp. Delhi.Kerrich, G. J. 1973. On the taxonomy of some forms of Ibalia Latreille (Hymenoptera: Cynipoidea) associated with conifers. Zoological] ournal of the Linnean Society 53: 65-79.Kim, C. W. 1955. A study on the ichneumon-flies in Korea. [In Chinese and German.] Commemoration Theses, Fiftieth Anniversary , Korea University: 423-497.Kolubajiv, S. 1934. Die Ergebnisse der Ziichtung von parasitischen Insekarten aus ihren Wirten in der staatlichen Versuchsanstalt in Prag in den J. J. 1929-1933. Casopis Ceskoslovenske Spolecnosti Entomologicke3l: 59-68, 113-120, 155-163.Kosobutzkii, M. I. 1928. [The noctuid attacking winter crops (Euxoa segetum Schiff.) in the Votsk ICHNEUMON-FLY GENUS BANCHUS 47 Autonomous Region (Biology, Ecology and Measures of Control), 1926-1928. (Attempt at a mono-graphic investigation).] 192 pp. [In Russian.] [Original not seen. Information taken from Review ofApplied Entomology (A) 21: 217.] Kuslitzky, W. 1979. Banchini (Hymenoptera, Ichneumonidae) of the Mongolian People's Republic. [InRussian.] Insects of Mongolia 6: 343-353. Leonard!, G. 1928. Elenco delle species di insetti dannosi e loro parassiti ricordati in Italia fino all' anno 1911.Partell. 592pp. Portici. Linnaeus, C. 1758. Systema Naturae, 10th edn, 1, 824 pp. Holmiae. Ljungdhal, D. 1918. Lepidoperologiske anteckningar. Entomologisk Tidskrift 39: 82-91. Maneval, H. 1935. Observations sur des Hymemopteres de la faune franchise et description d'une especenouvelle. Revue Frangaise d' Entomologie 2: 65-76. Matsumura, S. 1911. Erster Beitrag zur Insekten-Fauna von Sachalin. Journal of the College of Agricul-ture, Tohoku Imperial University, Sapporo 4: 1-144. Meyer, N. F. 1922. [Towards the ichneumonid fauna of the Astrakhan province.] [In Russian andGerman.] Izvestiya Otdela Prikladnoi Entomologii 2: 125-140. 1925. [A new ichneumonid species from the Urals.] (Hymenoptera, Ichneumonidae). [In Russian and German.] Izvestiya Entomologicheskogo i fitopatlogicheskogo byuro Ural'skogo obshchestva lyubiteleiestestvoznaniya 4: 10. 1926. Einige neue Ichneumoniden und Cynipiden. Russkoe Entomologicheskoe Obozrenie 20: 260-264. 1927a. Schlufwespen, die in Russland in den Jahren 1881-1926 aus Schadlingen gezogen sind. [In Russian.] Izvestiya Otdela Prikladnoi Entomologii 3: 75-91. 1927ft. Zur Kenntnis der Tribus Banchini (Familie Ichneumonidae) und einiger neuer Schlupfwespen aus Russland. Konowia 6: 291-311. 1928. Zur Biologic der Parasiten von Feltia segetum Schiff . [In Russian.] Izvestiya Otdela Prikladnoi Entomologii 3: 201-218. 1934. Tables systematiques des Hymenopteres parasites (fam. Ichneumonidae) de 1'URSS et des pays limitrophes. 3. [In Russian.] Opredeliteli po Faune SSSR, isdawaemyje soologitscheskim Institutom Akademii Nauk 15: 271 pp.Momoi, S. 1973. Ergebnisse der zoologischen Forschungen von Dr. Z. Kaszab in der Mongolei. 332. Einige mongolischen Arten der Tribus Banchini (Hymenoptera: Ichneumonidae). Folia Entomologica Hungarica (Series Nova) 26 (Suppl.): 241-250.Morley, C. 1913. Ichneumonidae: I. Ichneumones Deltoidei. Fauna of British India. Hymenoptera 3: 1-531.1915 . A revision of the Ichneumonidae based on the collection in the British Museum (Natural History) . 4, 167 pp. London.Muche, H. 1981. Die Cephidae der Erde (Hym., Cephidae). Deutsche Entomologische Zeitschrift (N.F.) 28: 239-295. Miiller, O. F. 1776. Zoologiae Danicae prodromus. 274pp. Havniae.Nixon, G. E. J. 1986. A revision of the European Agathidinae (Hymenoptera: Braconidae). Bulletin of the British Museum (Natural History) (Entomology) [in press] .Oehlke, J. 1967. Westpalaarktische Ichneumonidae I: Ephialtinae. Hymenopterorum Catalogus (nova editio) 2, 49 pp. 's-Gravenhage. Panzer, G. W. F. 1800. Faunae Insectorum Germanicae 73, 24 pp. Nurnberg.1804. D. Jacobi Christiani Schaefferi iconum insectorum circa Ratisbonam indigenorum enumeratio systematica. 260pp. Erlangae.Pfeffer, A. 1933. [Invasion de Panolis flammea en Slovaquie occidentale. Biologic. Moyens repressifs, etc] Recueil des Travawc de L'Institut Recherche Agronomique de Tschecoslovaquie 116 (2): 3-45. [Original not seen. Information taken from Review of Applied Entomology (A) 22: 382-385.]Pospelov, V. P. 1924. [Annual report of the department of applied entomology.] [In Russian.] [Annals of the state institute of experimental agronomy] 2: 243-252. [Original not seen. Information taken from Review of Applied Entomology (A) 14: 604.]Provancher, L. 1874. Les ichneumonides de Quebec avec description de plusieurs especes nouvelles. [part] Le Naturaliste Canadien 6: 55-63. Roman, A. 1912. Die Ichneumonidentypen C. P. Thunbergs. Zoologiska Bidragfrdn Uppsala 1: 229-293.1928. Liste der aus Island bekannten Ichneumoniden. In Lindroth, C. H., Zur Land-Everte- bratenfauna Islands. I. Goteborgs Kungl. Vetenskaps och Vitterhets Samhalles Handlingar (B) 1 (6): 15-24.1930. Ichneumonidae collected by Major R. W. G. Hingston on the Oxford University Expedition to 48 M. G. FITTON Greenland, 1928. Annals and Magazine of Natural History (10) 5: 281-288.1931. Entomologische Ergebnisse der scwedischen Kamtchatka-Expedition 1920-1922. 33. Ichneumonidae, Subfamilien Pimplinae und Tryphoninae. Arkivfor Zoologi 23 A (6): 1-32. 1932. The Linnean types of Ichneumon flies. Entomologisk Tidskrift 53: 1-16. 1937. Die Arthropoden fauna von Madeira nach den Ergebnissender Reise von Prof. Dr. O. Lundblad Juli-August 1935. II. Hymenoptera: Ichneumonidae. Arkiv for Zoologi 30A (1): 1-26. Rudow, F. 1883o. Einige neue Hymenoptera. Entomologische Nachrichten 9: 57-64. 18836. Neue Ichneumoniden. Entomologische Nachrichten 9: 232-247. Ruthe, J. F. 1859. Verzeichnis der von Dr. Staudinger im Jahre 1856 auf Island gesammelten Hymenop-teren. Stettiner Entomologische Zeitung 20: 305-322, 362-379. Samoilova, Z. 1. 1936 [Influence of parasitism on the quantity of food consumed by infected or non-infectedlarvae.] [In Russian.] [Summary of scientific research work of the Institute of Plant Protection, 1935]:316-317. [Original not seen. Information taken from Review of Applied Entomology (A) 25: 153.] Schaeffer, J. C. 1767. Icones insectorum circa Ratisbonam indigenorum 1 (2): LI-C. Regensburg. 1768. Icones insectorum circa Ratisbonam indigenorum 2 (1): CI-CL. Regensburg. Scheidter, F. 1934. Forstentomologische Beitrage. 33. Sekundarparasiten aus Banchus femoralis Thorns.Zeitschriftfur Pflanzenkrankheiten 44: 507-508. Schellenberg, J. R. 1802. Entomologische Beytrage. 1, 24pp. Winterthur. Schmiedeknecht, 0. 1910. Opuscula Ichneumonologica 4 (fasc. 25): 1921-2000. Blankenburg. Schrank, F. P. 1786. Baiersche Reise. 276 pp. Munchen. 1802. Fauna Boica 2 (2), 412 pp. Ingolstadt. Schulz, W. A. 1912. Aelteste und alte Hymenopteren skandinavischer Autoren. Berliner EntomologischeZeitschrift 57: 52-102. Schwerdtfeger, F. 1952. Untersuchungen iiber den 'Eisernen Bestand' von Kiefernspanner (Bupaluspiniarius L.), Forleule (Panolis flammea Schiff.) und Kiefernschwarmer (Hyloicus pinastri L.). Zeit-schriftfur angewandte Entomologie 34: 216-283. Shevyrev, I. Y. 1913. Parasiti i swerchparasiti is mira nasekomych. [In Russian.] EntomologicheskilVestnik 1:117-222. Short, J. R. T. 1978. The final larval instars of the Ichneumonidae. Memoirs of the American Entomo-logical Institute 25, 508 pp. Smits van Burgst, C. A. L. 1927. Lijst van de namen der in Midden- en West-Europa waargenomenparasieten en hyperparasieten van de Gestreepte Dennenrups (Panolis griseovariegata Goeze). Ento-mologische Berichten 7: 237-240. Stelfox, A. W. 1936. The Irish ichneumon-flies of the genus Banchus. Irish Naturalist's Journal 6:63-64. Sulzer, J. H. 1776. Abgekiirzte Geschichte der Insekten nach dem Linnaeischen System. 1: 274 pp.Winterthur. Szepligeti, G. 1910. Ubersicht einiger Gattungen der Banchoinae. Archivum Zoologicum 1: 183-186. Thomson, C. G. 1897. Opuscula entomologica 22: 2407-2452. Lund. Thunberg, C. P. 1822. Ichneumonidea, Insecta Hymenoptera, illustrata. Pars I. Memoires de I'AcademieImperiale des Sciences de St. Petersbourg 8: 249-281. 1824. Ichneumonidea, Insecta Hymenoptera, illustrata. Pars II. Memoires de I'Academie Imperiale des Sciences de St. Petersbourg 9: 285-368. Townes, H. K. 1944. A catalogue and reclassification of the Nearctic Ichneumonidae. Memoirs of theAmerican Entomological Society 11 (1): 1-477. Townes, H. 1970. The genera of Ichneumonidae. Part 3. Memoirs of the American Entomological Institute13: 307 pp. Townes, H., Momoi, S. & Townes, M. 1965. A catalogue and reclassification of the eastern PalearcticIchneumonidae. Memoirs of the American Entomological Institutes, 661 pp. Townes, H. & Townes, M. 1978. Ichneumon-flies of America north of Mexico: 7. Subfamily Banchinae,tribes Lissonotini and Banchini. Memoirs of the American Entomological Institute 26: 614 pp. 1981. A revision of the Serphidae (Hymenoptera). Memoirs of the American Entomological Institute 32, 541 pp. Townes, H., Townes, M. & Gupta, V. K. 1961. A catalogue and reclassification of the Indo- AustralianIchneumonidae. Memoirs of the American Entomological Institute 1, 522 pp. Uchida, T. 1929. Drei neue Gattungen, neunzehn neue Arten und fuenf neue Varietaten der Ichneumo-niden aus Japan, Korea und Formosa (Hym). Insecta Matsumurana 3: 168-187. 1931. Beitraege zur Kenntnis der Tribus Banchini aus Japan. Insecta Matsumurana 6: 45-54. Ulbricht, A. 1911. Ichneumonidenstudien. Archiv fur Naturgeschichte 77 (1(2)): 144-152. ICHNEUMON-FLY GENUS BANCHUS 49 1916. Niederrheinische Ichneumoniden. 3. Nachtrag. Abhandlungen des Vereins fur Naturwissen- schaftliche Erforschung des Niederrheins 2: 1-21.Veen, J. C. van 1982. Notes on the biology of Banchus femoralis Thomson (Hym., Ichneumonidae) an endoparasitoid oiPanolisflammea (D. & S.) (Lep. , Noctuidae). Zeitschriftfur angewandte Entomologie 94:300-311.Viktorov, G. A. 1967. On the systematical position of ichneumonids of the genera Odinophora Foerster and Exetastes Gravenhorst and their new species (Hymenoptera, Ichneumonidae). [In Russian.] Entomologicheskoe Obozrenie 46: 864-868.Wagner, A. C. W. 1929. Schlupfwespen und ihre Wirte. Verhandlungen des Vereins fur Naturwissenschaft- liche Unterhaltung zu Hamburg 20: 1-17.Wesmael, C. 1849. Sur les ichneumonides de Belgique, appartenant aux genres Metopius, Banchus et Coleocentrus. Bulletin de I' Academic Royale des Sciences, des Lettres et des Beaux-Arts de Belgique 16: 621-634.Woldstedt, F. W. 1877. Beitrag zur Kenntnis der um St. Petersburg vorkommenden Ichneumoniden. Bulletin de I' Academic Imperiale des Sciences de St. Petersbourg 23: 432-460.Zorin, P. V. & Zorina, L. M. 1929. K biologii ogordnoj ssowki, Polia oleracea L. [In Russian.] Zashchita RasteniiS: 475-486. [Abstract in Review of Applied Entomology (A) 17: 588-589.] 50 M. G. FITTON 10 Figs 2-10 2-5, measurements used in descriptions, of (2) malar space, dilatatorius, female; (3) face,volutatorius, female; (4) gaster tergite 1, crefeldensis , male; (5) hind femur, falcatorius, male. 6-10,female gaster of (6) cerinus, lateral; (7) volutatorius, lateral; (8) volutatorius, dorsal; (9) falcatorius,lateral; (10) falcatorius , dorsal. Scale lines represent 1 mm. ICHNEUMON-FLY GENUS BANCHUS 51 13 15 16 \l 17 18 20 22 23 25 26 27 28 29 30 31 Figs 11-40 11-32, segments 4 and 5 of female maxillary palp, right, of (11) agathae; (12) cerinus; (13)crefeldensis; (14) dilatator ius; (15) falcatorius; (16) flavomaculatus; (17) gudrunae; (18) hastator; (19)insulanus; (20) japonicus; (21) mauricettae; (22) moppiti; (23) nox; (24) palpalis; (25) p/cft; (26)poppiti; (27) punkettai; (28) tfio/ws; (29) tumidus; (30) turcator; (31) volutatorius; (32) zonatus. 33-40,outline of scutellum, left lateral, of (33) agathae; (34) crefeldensis; (35, 36) dilatatorius; (37, 38)falcatorius; (39) flavomaculatus; (40) gudrunae. Scale lines represent 1 mm. 52 M. G. FITTON 57 58 Figs 41-58 Outline of scutellum, left lateral, of (41) hastator; (42) insulanus; (43) japonicus; (44)mauricettae; (45) moppiti; (46) nox; (47) palpalis; (48) poppiti; (49, 50) pictus; (51) punkettai; (52)sanjozanus; (53) tholus; (54) tumidus; (55) turcator; (56) volutatorius; (57, 58) zonatus. Scale linerepresents 1 mm. ICHNEUMON-FLY GENUS BANCHUS 53 Figs 59-73 Gaster tergite 1 , left lateral, of (59) agathae $ ; (60) crefeldensis cf ; (61) dilatatorius $ ; (62)falcatorius $ ; (63) flavomaculatus $ ; (64) hastator $ ; (65) gudrunae cf ; (66) gudrunae $ ; (67) insulanus$; (68) japonicus $; (69) mauricettae cf; (70) moppiti <j>; (71) no* cf; (72) palpalis cf; (13) pictus $.Scale line represents 1 mm. 54 M. G. FITTON 74 75 82 Figs 74-84 74-81, gaster tergite 1, left lateral, of (74) poppiti d"; (75) punkettai 9 ; (76) sanjozanus d";(77) tholus $ ; (78) tumidus <j> ; (79) turcator $ ; (80) volutatorius $ ; (81) zonatus $ . 82-84, apex of gasterof female, (a) dorsal, (b) left lateral, of (82) agathae; (83) crefeldensis; (84) dilatatorius. Scale linerepresents 1 mm. ICHNEUMON-FLY GENUS BANCHUS 55 Figs 85-93 Apex of gaster of female, (a) dorsal, (b) left lateral, of (85) falcatorius; (86)flavomaculatus;(87) gudrunae; (88) hastator; (89) insulanus; (90) japonicus; (91) mauricettae; (92) moppiti; (93) nox.Scale line represents 1 mm. 56 M. G. FITTON 97 Figs 94-102 Apex of gaster of female, (a) dorsal, (b) left lateral, of (94)palpalis; (95)pictus; (96) poppiti;(97) punkettai; (98) tholus; (99) tumidus; (100) turcator; (101) volutatorius; (102) zonatus. Scale linerepresents 1 mm. ICHNEUMON-FLY GENUS BANCHUS 57 '103 \ 104 105 111 118 119 123 Figs 103-123 Segments 3, 4 and 5 of male maxillary palp, right, of (103) agathae; (104) crefeldensis; (105)dilatatorius; (106) falcatorius; (107) flavomaculatus; (108) gudrunae; (109) hastator; (110) japonicus;(111) mauricettae; (112) nox; (113) moppiti; (114) palpalis; (115)pictus; (116) poppiti; (117) punkettai;(118) sanjozanus; (119) //ZO/MJ; (120) tumidus; (121) turcator; (122) volutatorius; (123) zonatus. Scaleline represents 1 mm. 58 M. G. FITTON Figs 124-129 Segments of male antennal flagellum, showing flag setae, right antenna, interno-lateral, of(124) hastator; (125) crefeldensis; (126) zonatus; (127) moppiti; (128) volutatorius; (129) falcatorius . ICHNEUMON-FLY GENUS BANCHUS Index 59 The index includes host names. Principal entries for the valid Banchus species are in bold. Achlya 6, 25 Acronicta 6, 22 acuminator 7, 17 adusta 6, 19,25,32 agathae7, 11, 14 Agrochola 6, 34 Agrotis6,21,22,34,41 alasuffuscus 44 algericus 8, 41 altaiensis 7, 15 alticola (Ashmead) 8,31 alticola (Schmiedeknecht) 8, 40 ambusta 6, 34 AnartaS, 6,41 Andricus 8, 44 annulatus 8, 44 Aporophyla 6, 17 aries 7, 19 armillatus 8, 43 atalantae 8, 44 Atethmia 6, 34 Autographa 5 Banchopsis 3Banchus 3Bena 6, 41bipunctatus 8, 33Blepharita 6, 19, 25, 32brassicae 7, 41bucephala 6, 34 calcaratus 8, 40 californica 5 Cephus 8, 44 Ceramica 7, 41 Ceratogastra 3 cerinus 7, 14, 15 certator 8, 40 Cidaphurus3,23,31 circellaris 6, 34 citraria 44 compressus 7, 17, 44 compta 7, 41 configurata 5 contigua 7, 41 Corynephanes 19 Corynephanus 3,31 crefeldensis 7, 11, 13, 15, 24, 31 croaticus 7, 15 cultellator 8, 44 cultratus 8, 33 Dasychira 6, 21 Deilephila6,21,25,32 didyma 6, 22 dilatatorius 5, 7, 12, 13, 17, 40, 43 Diprion 7, 22 Diprionidae 7 dispar 25 Earinus 8, 44 elator8,44elpenor 6, 21ephialtes 6, 41Episyron 8, 44Euceros 6Euproctis 6, 19Euxoa 7, 19exclamationis 6, 22Exetastes 3, 8, 44 falcator 7, 19 falcatorius4, 5, 7, 11, 13, 19, 41, 44 falcifera 5 farrani 8, 40 fascelina 6, 21 femoralis 6, 7, 24 flam me a 5, 7, 25, 32 flavescens 5 flavicornis 6, 25 flavomaculatus 7, 11, 14, 23, 30, 36 formidabilis 8, 30 fornicator 8, 44 Geometridae 6groenlandicus 8, 31gravidator 8, 44gudrunae 5, 7, 11,13,23 Habrosyne 6, 41 Hadena7,34,41 hastator5,6,7, 12, 13,24,31 Heliothis5,7,41 helvola 6, 34 histrio (Fabricius), 8, 44 histrio(Schrank),7, 19 Hyloicus 6, 25 Ibalia 8, 44 Ichneumon 8, 17, 19, 24, 33, 40, 45 insulanus7, 13,26 intersectus 7, 19 japonicus 7, 12, 14, 26, 36 kolosovi 7, 24kozlovi 25 labiatus 7, 19Lacanobia 5, 7, 32, 41lavrovi 7, 17, 19Leucoma 6, 19leucospoides 8, 44Lissonota 8, 44luenebergensis 6, 17luteofasciatus 7, 19luteolata 6, 41luteus 7, 22lutulenta 6, 17Lycophotia 7, 34, 41Lymantria 25Lymantriidae 6 Mamestra5,7, 41mauricettae 7, 11, 14, 28medialis 6megacephala 6, 22Megarhyssa 8, 44Melitaea 6, 22meticulosa 7, 19monileatus 8, 30, 40moppiti 5, 7, 12, 13, 28, 43mutillarius 33mutillatus 8, 33myrtilli 5, 6, 41 Nawaia3,26ni5 nigricans 7, 19nigromarginatus 7, 19nobilitator 7, 19Noctuidae 6Notodontidae 6nox 7, 11,14,23,30,36Nymphalidae 6 obscurus 8, 40ocellata 6, 34oleracea5,7,32, 41ononis 5Ophion 7, 22Opisthograptis 6, 41 palpalis 7, 11, 13, 17, 24, 27, 30 Panolis5,7,25,32 Phalera 6, 34 Philogalleria 3 Phlogophora 7, 19 Phytodietus 8, 43 pictus 8, 12, 13, 14, 33, 39 pinastri 6, 25 pini 7, 22 pisi 7, 41 Pompilus 8, 44 poppiti 8, 12, 14, 27, 35 porcellus6,21,25,32 porphyrea 7, 34, 41 prasinana 6, 41 Proctotrupes 8, 44 propitius 7, 19 pungitor7,24 punkettai 8, 12, 14, 23, 36 pygmeus 8, 44 pyritoides 6, 41 quadrator 8, 44 reticulator7,24Rhynchobanchus 3rivularis 7, 34robustus 8, 44russiator 8, 33 sachalinensis 7, 19salicis 6, 19 60 M. G. FITTON sanguinator 7, 19 tholus 8, 12, 13, 15, 37 venator 8, 40 sanjozanus 8, 12, 37 Thyatiridae 6 vigilatorius 8, 44 Scotogramma 5 tomentosus 8, 44 villosulus 8, 44 segetum 6, 21 , 34, 41 Trichoplusia 5 viridator 8, 44 sibiricus 7, 17 tricolor 7, 19 viriplaca 7, 41 similis 6, 19 trifolii 5 volutatorius 4, 5, 8, 12, 13, 31, 37, 40 Smerinthus 6, 34 tumidus 8, 12, 13, 15, 38 Sphingidae 6 turcator 8, 12, 14, 34, 39 xanthographa 5 , 7, 41 spinipesS, 44 Xestia5,7,41 spinosus 7, 30 umbellatarum 8, 40 suasa 7, 41 zagoriensis 8, 33 superba 44 variegator 7, 19 zonatus 5, 8, 12, 13, 29, 41 varius 8, 44 Zygaena 6, 41 Theronia 8, 44 vastator 24 Zygaenidae 6 British Museum (Natural History) An introduction to the Ichneumonidae of Australia /. D. Gauld In the important field of biological and integrated control of pests the parasitic Hymenopteraare of particular significance, and this work considers one of the largest families of Parasitica,the Ichneumonidae. The group has received little attention in Australia - though it has alreadybeen utilized successfully in curtailing the ravages caused by accidentally introduced pests. Forselective control programmes to be effective, however, a sound knowledge of the biology ofboth the pest and its parasites is essential - and a sound taxonomic base is vital for thedevelopment of such knowledge. Ironically, considering the group's economic importance, the parasitic Hymenoptera isamongst the least studied of any group of living organisms, and taxonomic difficulties havepresented major problems to many entomologists working with the Parasitica. AnIntroduction to the Ichneumonidae of Australia will go a long way towards rectifying thissituation, being a taxonomic treatment, by genus, of the Australian ichneumonids, acomprehensive illustrated identification guide, and a summary of all available information onthe group. It will also serve as an introduction to the biology and distribution of Australianichneumonids, and provide a check-list of the described species and an index to their knownhosts. It provides an important revision of ichneumonid nomenclature in order to bring thegroup into line with the generally accepted principles of zoological nomenclature. 1984, 413pp, 3 maps, 580 figs. Paperback. 565 00896 X 40.00 Titles to be published in Volume 51 The ichneumon-fly genus Banchus Banchus (Hymenoptera) in the Old World By M. G. Fitton The phylogeny, classification and evolution of parasitic wasps of the subfamily Ophioninae(Ichneumonidae) .By I. D.Gauld A cladistic analysis and classification of trichodectid mammal lice (Phthiraptera: Ischnocera).By C. H. C. Lyal The British and some other European Eriococcidae (Homoptera: Coccoidea). ByD. J.Williams Photoset by Rowland Phototypesetting Ltd, Bury St Edmunds, SuffolkPrinted in Great Britain by Henry Ling Ltd, Dorchester Bulletin of the British Museum (Natural I BRITISH MUbtUM .(NATURAL HISTORY) 30AU61985 PRESENTEDGENERAL LIBRARY istory) The phylogeny, classification and evolution of parasitic wasps of the subfamily Ophioninae (Ichneumonidae) lanD. Gauld Entomology series VolSl No 2 29 August 1985 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in fourscientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology,and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique andever-growing collections of the Museum, both by the scientific staff of the Museum and byspecialists from elsewhere who make use of the Museum's resources. Many of the papers areworks of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself,available separately, and individually priced. Volumes contain about 300 pages and severalvolumes may appear within a calendar year. Subscriptions may be placed for one or more ofthe series on either an Annual or Per Volume basis. Prices vary according to the contents ofthe individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History),Cromwell Road, London SW75BD,England. World List abbreviation: Bull. Br. Mus. nat. Hist. (Ent.) Trustees of the British Museum (Natural History), 1985 The Entomology series is produced under the general editorship of the Keeper of Entomology: Laurence A. Mound Assistant Editor: W. Gerald Tremewan ISBN 565 06012ISSN 0524-6431 British Museum (Natural History)Cromwell RoadLondon SW7 5BD Entomology seriesVol51No2pp61-185 Issued 29 August 1985 The phytogeny, classification and evolution ofparasitic wasps of the subfamily Ophioninae(Ichneumonidae) Ian D. Gauld Department of Entomology, British Museum (Natural History), Cromwell Road, LondonSW7 5BD Contents Synopsis 62 Introduction 62 The research potential of Ophioninae 63 The aim of the present work 63 Terminology 63 Material examined 63 Discussion of methodology 63 Choice of parsimony method utilized 64 The advantages and disadvantages of the parsimony method 65 Compatibility methods utilized 66 Value of LEQU . B AS for overcoming some problems of compatibility analysis 67 The operational procedure adopted in this study 70 An apologia for subjectivity in cladistic analysis 70 The holophyly of the subfamily 70 The systematic position of the subfamily 71 Polarity determination and character coding 73 Characters used in the study 74 The characters and their polarity 74 Preliminary remarks on the relative values of the characters 79 The phylogenetic analysis 80 Preliminary analysis of data 80 Analysis using Le Quesne test program, LEQU. B AS 80 Analysis using FOURS program 85 Phylogenetic analyses of groups of ophionine taxa 87 The Euryophion group 87 The Thyreodon group 88 The Dictyonotus/Ophionopsis group 89 The Thyreodon + Euryophion + Dictyonotus complex 90 The Stauropoctonus group 95 The Enicospilus + Stauropoctonus complex 95 The major groups and their relationship with the unassigned taxa 107 The inter-relationship of the two major generic complexes 107 The position of the unassigned taxa 107 Discussion of suggested phylogeny 116 A classification of the Ophioninae 118 Zoogeographic discussion 119 The subfamily Ophioninae 122 The Ophion genus-group 122 The Sicophion genus-group 128 The Eremotylus genus-group 130 The Thyreodon genus-group 133 BRITISH MUSEl{NATURAL HISTORY 30AUGI9& PRESENTEDGENERAL LIBRA Bull. Br. Mus. not. Hist. (Ent.) 51 (2): 61-185 Issued 29 August 1985 62 I. D. GAULD The Enicospilus genus-group 141 The Orientospilus subgroup 142 The Ophiogastrella subgroup 144 The Stauropoctonus subgroup 145 The Leptophion subgroup 149 The Enicospilus subgroup 154 Acknowledgements 162 References 163 Appendices 168 Index to generic names 185 Synopsis An attempt is made to reconstruct the phylogeny of the genera of the ichneumonid subfamily Ophioninaeusing both parsimony and compatibility methods of analysis. The problems of phylogenetic analysis ofhighly homoplastic data are discussed and the strengths and weaknesses of the various methods evaluated.It was concluded that there is no wholly adequate objective method of analysing highly homoplastic data,but it is suggested that analysis may be undertaken using subjective evaluation of characters supportingconflicting patterns. Within the Ophioninae five major evolutionary lineages are recognized, the Ophion,Sicophion, Eremotylus, Thyreodon and Enicospilus genus-groups. The largest, the Enicospilus genus-group, is subdivided into five subgroups, the Orientospilus, Ophiogastrella, Stauropoctonus, Leptophionand Enicospilus subgroups. A scenario for the possible evolution of the subfamily is suggested. Thirty-twogenera are recognized, one of which, Janzophion, is described as new. Rictophion is treated as a synonymof Euryophion, Ophionopsis is treated as a synonym of Dictyonotus, and Aulophion treated as a synonymof Stauropoctonus. The majority of genera are shown to be holophyletic, but Ophion and Enicospilus arethought to be paraphyletic, though for purposes of classificatory convenience they are retained as validgenera. Diagnostic descriptions of the genera are given and the inter-relationships of their componentspecies are discussed. Introduction Ophionines are mostly quite large, slender, orange-brown ichneumonids that can frequently beobserved flying around lights at night in virtually any part of the world. In temperate regionsthere are rather few species, but in the tropics there are very large numbers of taxa, andophionines form a conspicuous component of the ichneumonid fauna. Almost all members ofthe subfamily are nocturnal or crepuscular, though in drier areas a few are diurnally active.Males of many species fly at dusk, but most females are not active until an hour or two after dark. Ophionines are solitary protelean endoparasites of holometabolous insect larvae. The hostsof the vast majority of species are not known, but what records there are usually refer to speciesparasitizing lepidopterous larvae, most often exposed, solitary larvae, and in particular speciesof the families Noctuidae, Geometridae and Lymantriidae. Larvae of a wide range of otherfamilies are also attacked, but there are extremely few records of ophionines parasitizingMicrolepidoptera or Rhopalocera. One Nearctic species is exceptional in that it is known toparasitize coleopterous larvae (Townes, 1971). Oviposition is usually into the host larva. The parasitoid egg is apparently free in the host'shaemocoel where it hatches to produce a caudate first instar larva (Moutia & Curtois, 1952).Species attacking mature larvae undergo rapid development, but species (e.g. Enicospilusamericanus) that oviposit in very young larvae have a protracted first larval instar (Price, 1975).Several species are apparently host specific (Janzen, pers. comm.), whilst others (e.g.Thyreodon atriventris) parasitize a variety of taxonomically related hosts. Some temperatespecies seem to attack a number of different hosts in a similar niche, whilst a few species arebivoltine, with different generations attacking different hosts (Brock, 1982). A few species seemto attack a wide range of hosts (Gauld & Mitchell, 1981). The parasitoid larva completesdevelopment just prior to host-pupation, often after the host has constructed a cocoon. Theichneumonid larva spins a characteristic fibrous, ovoid cocoon which is generally dark brownwith a pale equatorial band. Species may remain as mature larvae or even adults in this cocoonfor the greater part of the year in seasonal habitats. PHYLOGENY OF THE OPHIONINAE 63 The research potential of Ophioninae The ease with which ophionines may be collected (using m.v. light-traps) makes them particu-larly suitable for zoogeographic and ecological study. Large samples may be collected in terrainwhere sweep netting and Malaise trapping yield poor results, or, as in the case of rain forestcanopy, where collecting can only be achieved by cumbersome, expensive and (for fast-flyinginsects) unproven techniques. Unlike Lepidoptera, which are similarly easy to collect, mosttropical ophionines are readily separable without recourse having to be made to time-consuminggenitalic preparation. Illustrated keys are available to facilitate identification of most Old Worldtropical species (Gauld, 1977; Gauld & Mitchell, 1978; 1981). The aim of the present work If the Ophioninae is to be used as a serious vehicle for zoogeographic study then it is necessary tohave an understanding of the phylogenetic inter-relationships of the genera, as many authorscontend that biogeographic speculations are valid only when related to the evolutionary historyof a group (Mackerras, 1962; Nelson & Platnick, 1981). As little has been published on thephylogeny of Ophioninae, a cladistic study of the group is necessary before much of thedistributional data available can be interpreted in an evolutionary manner. The aim of thepresent study is to investigate the phylogeny of the Ophioninae using a variety of moderncladistic techniques, and to relate the results to what is known about the distribution and biologyof the group in general. Terminology The morphological terminology in this work follows that proposed by Richards (1956) andinterpreted for the Ophioninae by Gauld & Mitchell (1978; 1981). Specialist terms, relevant toonly some ophionines, are defined in these works. Family-group names used conform with thedirectives of the International Code of Zoological Nomenclature and with various Opinions ofthe International Commission. Some recent authors (e.g. Townes, 1969; 1971) have chosen todisregard certain of these opinions and have not followed the Code when forming family-groupnames (see Fitton & Gauld, 1976). Material examined The majority of specimens examined are contained in the collections of the British Museum(Natural History) (BMNH), but valuable additional material was furnished by the AustralianNational Insect Collection (ANIC), Canberra, the Bernice P. Bishop Museum (BPBM),Hawaii, the Canadian National Collection (CNC), Ottawa, the Gupta Collection (GC), theMusee Royal de 1'Afrique Centrale (MRAC), Tervuren, the Museum National d'HistoireNaturelle (MNHN), Paris, the Taiwan Agricultural Research Institute (TARI), Taichung, theTownes Collection (TC), Ann Arbor, the United States National Collection (USNM),Washington and the Zoological Institute (ZI), Leningrad. The types of virtually all Old Worldspecies have been examined and exhaustive lists of material examined are contained in recentrevisionary studies (Gauld, 1977; Gauld & Mitchell, 1978; 1981; Gauld & Carter, 1983). Discussion of methodology The present work is an attempt to elucidate the phylogeny of the Ophioninae and therefore onlyphylogenetic methods of data analysis have been used. These methods are based on ideasinitially expounded by Hennig (1966). Although Hennig made a major contribution to system-atic philosophy, he greatly underestimated the difficulty that would occur in phylogeneticanalyses due to morphologically undetectable evolutionary parallelism and character-statereversal (i.e. homoplasy). Such events result in there being incompatible character sets intaxonomic data. An estimation of the extent of incompatibility can be made using the simple test 64 I. D. GAULD outlined by Le Quesne (1969) (see also Gauld & Mound, 1982; Underwood & Gauld, in prep.),and such incompatibilities are far from uncommon in most real data. In the past decade two main approaches have been adopted by cladists in an attempt to resolveconflicting character sets - parsimony methods, which seek to minimize the number ofcharacter-state transformations, and compatibility methods, which seek to find a series ofnesting groups supported by the largest number of compatible characters (Felsenstein, 1982).For simple data sets where there is a low incidence of homoplasy, these methods yield verysimilar results (e.g. Gauld, 1983), but as the frequency of homoplasy increases so the methodsusually give increasingly different results. Parsimony methods have gained wide acceptance in the literature, with the shortest rootedWagner tree often being uncritically presented as the best phylogenetic hypothesis. Simul-taneously, many authors have criticized compatibility analysis either for producing numerousapparent best solutions (Kluge, 1976), or as being less useful at producing congruent phy-logenies from different developmental stages (e.g. Mickevich, 1978; Rohlf & Sokal, 1980) (butsee also Rohlf etal., 1983). However, Felsenstein (1981; 1982) stated that both methods can bejustified as maximum likelihood methods, but under somewhat different circumstances. Ifhomoplasy is expected to be scattered at random throughout all characters then a parsimonymethod is favoured, but if homoplasy is expected to be concentrated in certain characters thencompatibility is supported. Felsenstein (op. cit.) continues by noting that both methods requireboth homoplasy to be rare and characters to have a low rate of change. In the case ofOphioninae, homoplasy demonstrably is not rare, suggesting results obtained by either methodshould be viewed circumspectly. The additional assumption necessary for parsimony methods toapproach maximum likelihood, that is that homoplasy is randomly scattered across all charac-ters, contradicts the consensus of opinion of most practising entomological taxonomists. Bothparsimony and compatibility methods were used, with caution, in this study. Choice of parsimony method utilized Of the parsimony methods available at the start of this study, the program selected was PHYLIP(package for inferring phytogenies) written by Dr J. Felsenstein. This program offers fourroutines for dealing with non-polymorphic, discrete state data: (a) the Camin-Sokal parsimonymethod; (b) the Dollo parsimony method; (c) the Wagner parsimony method and (d) a mixedmethod allowing a, b or c to be specified for each character. The Camin-Sokal and Dollomethods were not used in the present study. Neither gives as short a tree as that obtained by theWagner method as both place additional constraints on the 'tree-growing' method (Felsenstein,1982), the former by not allowing reversal, the latter by not permitting forward parallelism. In the insects being studied there is virtually no evidence at all for either of the additional apriori assumptions these methods necessitate. Indeed there is some biological evidence tosuggest that these extra assumptions are unwarranted for ichneumonids. Dollo parsimony mayjustifiably be invoked for treating complex structures (such as the vertebrate eye), but virtuallyall characters used in the present study involve small changes in simple structures. Very often thederived state involves reduction of a structure, and observation of other ichneumonid taxasuggests that certain of these apomorphies (such as reduction of extent of occipital carina) havebeen developed in parallel in numerous different evolutionary lineages. Even when theapomorphic state is the development of a novel structure (such as an alar sclerite), the evidencestrongly suggests (as the feature occurs in a few otherwise specialized species of several differentgenera) parallel development in closely related lineages. Reversal is more difficult to demonstrate, but clearly it does occur. For example, the alarsclerites apparently have been lost by some Enicospilus species on oceanic islands (Gauld &Carter, 1983) and the posterior transverse carina of the propodeum, which is absent in mostEnicospilus species, has been redeveloped in some members of the otherwise highly specializedE. signativentris species-complex (Gauld & Mitchell, 1981). Because of the objections to the Camin-Sokal and Dollo methods the option employed in thisstudy was that which necessitates no extra a priori assumptions, the Wagner method. Clado- PHYLOGENY OF THE OPHIONINAE 65 grams were rooted using a hypothetical all zero ancestor, as use of any outgroup taxon is liable tointroduce more incidences of homoplasy. The advantages and disadvantages of the parsimony method The major advantage of the parsimony method seemed to be that, because of its 'averagingprocedure' (see below), generally all species of a particular genus were associated. As ophioninegenera are polythetic, compatibility methods invariably exclude certain taxa (see Davies &Boratyriski, 1979). Despite its widespread acceptance there are serious flaws in parsimony methodology whenapplied to complex data sets. (There are also philosophical objections (see Pratt, 1972;Felsenstein, 1981; Friday, 1982; Panchen, 1982), but these are beyond the discussion of thepresent work.) One, and perhaps the most obvious methodological flaw, is that because themethod considers each character as an identical piece of binary information, and is attemptingonly to minimize transformation, a large set of coincidental 'bad' characters will be favoured atthe expense of even a very slightly smaller set of 'good' characters. This is best illustrated in thecase of some Hawaiian genera of ophionines (taxa 924-6 in the following study) and Ophio-gastrella (taxa 910-11). In both cases, the sets of characters responsible for positioning these taxain the Wagner analysis are composed of characters that are likely to be highly homoplastic inother taxa. Although a traditional taxonomist intuitively recognizes these as poor characters(and taxonomist's intuitive judgement may be very good, see Davies, 1981) and has noconfidence in them as indicators of phylogenetic affinity, it was not found to be possible totranslate this subjective bias into an objective taxonomic weighting scheme. Without anobjective weighting scheme the averaging procedure adopted by the Wagner method has theeffect of reducing the number of transformation steps of many 'weak' characters at the expenseof 'moderately good' ones (Strauch, 1984). Although some hybrid parsimony/compatibility trees produced in this study were longer thanthe shortest Wagner tree (mainly because certain characters appear to have been 'written-off,i.e. allowed to have a large number of transformations) it is noteworthy that these cladogramswere supported by more characters with minimal or nearly minimal homoplasy. This can best beillustrated by reference to two alternative cladograms produced during this study, the mostparsimonious 235-step one and the favoured 239-step one (Gauld, unpubl. PhD). The numbersof characters with various numbers of transformation steps are shown in Fig. 1. A second disadvantage of the parsimony method is that a multiplicity of almost equallyparsimonious cladograms may be generated from the same data set by the same procedure(Strauch, 1984), but altered merely by re-ordering the input order of the taxa (e.g. Figs 21-24).In this study it soon became apparent, when rearrangements are practised, that certain taxa areprone to 'hopping', i.e. moving from one lineage to another in different reconstructions. Taxon906 was found to be such an OTU - its final position in a cladogram depended solely upon theposition in which it was entered into the analysis. This was presumed to operate in the followingmanner: taxon 906 is almost equally associated (see Appendix 3) with three separate lineages,Simophion, the Thyreodon complex and the Enicospilus complex. This equality of association isso pronounced that once the taxon has linked to one or other of the alternatives (the firstencountered) it does not share enough derived features with any other taxon to disassociateitself. Thus if the first three taxa entered are 905 (Simophion), 906 and 940 (a more distantlyrelated taxon), 906 remains associated with 905, whilst if the first three taxa entered are 935 (oneof the Thyreodon complex), 906 and 940, 906 remains associated with the Thyreodon complex.Some programs attempt to circumvent this problem by computing an 'advancement index' forordering the taxon input, but I fail to see the intellectual justification for this. A third disadvantage of the parsimony method is that it is not possible to predict, from anygiven data set, the actual minimum tree length (Felsenstein, 1982; Day, 1983). This means that,for large data sets, in practice one is guessing that the minimum length tree obtained is theshortest possible. This study shows that for highly homoplastic data, a number of equally short,quite different cladograms may be obtained. It appears to be merely an act of faith that a slightly 66 I. D. GAULD Number ofcharacterswith Ntransform-ationsteps OQ _ 25- 20- 15- 10- 5- 239 step cladogram 235 step cladogram / A \ ;\ I \ is . \ i \ \\\ \ ' \ \ \ + + 4- 12345678 9 number of transformation steps, N Fig. 1 A comparison of the incidences of transformation required to fit all characters to the mostparsimonious 242-step cladogram and a favoured 249-step cladogram (From Gauld, unpubl. PhDthesis). It is noteworthy that the latter, although 7 steps longer, is supported by more minimallyhomoplastic characters. shorter and yet totally different arrangement does not exist, a rather disconcerting observation ifone accepts the principle of parsimony as paramount in phylogenetic reconstruction. Compatibility methods utilized Felsenstein's package, PHYLIP, provides an option for producing compatible character sets.This option, CLIQUE, was used to find the largest cliques for various sets of taxa considered inthis study. A second program (written by Dr G. Underwood) was developed by Underwood andGauld during the course of this study. This program was developed from the work of Le Quesne(1969; 1972) and involves calculating the probability of incompatibility between two characterson a null hypothesis of random distribution of states of both. This value was then used tocalculate for each character a coefficient of character-state randomness (herein called the O/Evalue). Typically the output of this program, LEQU.BAS, consists of a list of characters withtheir observed incompatibilities, calculated expected number of incompatibilities, and O/E PHYLOGENY OF THE OPHIONINAE 67 values tabulated; the characters are ranked by O/E value as a postscript (e.g. Tables 2, 6). Afacility has been incorporated so that multistate characters which have been coded in binaryfashion and numbered as decimal increments (e.g. 1-1, 1-2), and cannot logically fail, are notcompared. Included also in LEQU.BAS is a labelling feature. This is based on an idea given by Guise etal. (1982) for counting frequencies of each of the four possible character-state combinations(0,0; 0,1; 1,0; 1,1) for each character pair. If a pair of characters fail the test (because all fourcombinations are found in the data) on account of a single occurrence of one combination, thetaxon having this combination is labelled. The results of this operation are printed in the form ofa table (e.g. Appendix 4). The figure in brackets adjacent to the taxon is the total number oflabels the taxon has received. The taxa most frequently labelled will be the ones with the mostdiscordant character sets. A particularly high label score for a species for any one characterstrongly suggests homoplasy for that character with respect to other taxa in the set. Some care isneeded in the interpretation of results as characters with very unequal character-state distri-bution (either only two or two 1 scores) often have high label values. These are easilyrecognized as equally high scores are given for the two taxa with the minority states. Even thesevalues can sometimes be seen to be unusually high (i.e. suggestive of homoplasy) if the characterlabel values are compared with the label values of similarly uninformative characters. Value of LEQU.BAS for overcoming some problems of compatibility analysis Two major disadvantages with compatibility analysis are: (1) a maximum character set is oftenfar too small to allow full resolution of data (Felsenstein, 1982) and (2) the data set may yield anumber of more or less equally large cliques (Kluge, 1976). LEQU.BAS is of some value inovercoming both these problems. As the largest compatible character set is generally very small, it is useful to have a way ofordering characters. Homoplastic characters that would be eliminated from any compatible setare not all equally bad; some are far more discordant than others. For example, allowance of justone incidence of parallelism or reversal may be enough to make certain characters compatiblewith a cladogram, whereas others will necessitate the invocation of multiple incidences ofhomoplasy to achieve congruence. It can be seen from the hypothetical data set 1 (Tables 1-3)that LEQU.BAS offers a way of grading characters from slightly to extremely homoplastic.Eight taxa, A-H, exhibit 18 binary characters (Table 1). Characters 1-12 support the nested set(((AB)C)((DE)(F(GH)))). Characters 13-18 are differentially homoplastic, 13 and 14 requiringone extra transformation to fit the specified cladogram, 15 and 16 requiring two extratransformations and 17 and 18 needing three extra. Analysing these data with LEQU.BAS gavethe results presented in Table 2. It can be seen that the six homoplastic characters are ranked inorder of the extra number of transformations required to make them fit the specified cladogram.Several taxa are highly labelled for certain characters, suggesting homoplasy. For example,character 15, which is present in the apomorphic state in B and C, but is plesiomorphic in A, ishighly labelled for A, suggesting a reversal may have occurred. Table 1 Character state matrix for hypothetical data set 1. A-H represent taxa, 1-18 their independentcharacters. = plesiomorphic, 1 = apomorphic. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 A 111100000000110110 B 111100000000111101 C 001100000000101001 D 000011110000010010 E 000011110000010001 F 000011001100001111 G 000011001111101010 H 000011001111101101 68 I. D. GAULD Table 2 Result of LeQuesne test on hypothetical data set 1 . The number in italics is the character number ,the second column is the number of observed failures, the third column the number of expected failuresand the final column in each case the ratio of observed over expected failures. Characters with a low O/Evalue are considered to be good characters, and are ranked in the final line higher than those with highO/E scores. Character number: failures observed expected O/E ratio Ranking ratios7 8 9 10 1 2 11 12 3 4 5 6 13 14 15 16 17 18 Although there is a clear relationship between the number of extra transformations requiredto make a character 'fit' a favoured cladogram and that character's O/E value, the O/E value alsovaries with information value of the character, i.e. the number of 1 scores in relation to scores.For example, in data set 1, character 14 may be modified so it has the following distribution of 1states in taxa A-H, 11100000, 11001000 or 10001001. These require one, two and threetransformations, respectively, and have O/E values of 041, 0-90 and 1-24. However, if thenumber of transformations necessary is kept constant (at say two) and the number of derivedstates varied, the following relationship may be observed: 10001000 (O/E = 1-41), 11001000(O/E = 0-90), 11011000 (O/E = 0-61), 11100010 (O/E = 0-53), 11100011 (O/E = 0-50),11100111 (O/E = 0-23). Similar variations of character 16 (requiring three transformations)produced values with the range 0-74-1-24 (the largest being for 10010001), whilst a range of1-16-1-35 was found for a character requiring four transformations. Thus a homoplasticcharacter with a high proportion of derived states supporting various subgroups in the definitivecladogram may have a lower O/E value than a less homoplastic, but less informative character. Kluge (1976) pointed out that compatibility analysis may produce a large number of equallylarge cliques. From his data set of 139 binary characters he obtained six almost equally largecliques of 80 or 81 characters. Kluge remarked that not only is there no reason to prefer one toany other of the sets, but also pointed out that the most primitive species in each of the cliquesdid not correspond to the most primitive as assessed on the basis of the best documentedcharacters. To consider Kluge's objections necessitates re-examining his data. Of his six largestcliques, it is apparent that no less than 76 characters are common to all sets. The six dendrogramssupported by these cliques are essentially similar, differing mostly in the relative order of themost primitive taxa (im-ti in Kluge's fig. 7a) so a great deal of information is common to all, and Table 3 Labels matrix derived from hypothetical data set 1 . The column to the right of the taxon letter isthe total number of times that taxon is solely responsible for a character failing the LeQuesne test. Forfurther details see text. A (21) 333333 B ( 6} 3 3 C ( 8) - - D ( 5) - - E ( 7) - - F ( 5) - - G (12) - - H ( 8) - - 9 10 11 12 13 14 15 16 17 18- - - - 1 - PHYLOGENY OF THE OPHIONINAE 69 this agrees with Kluge's preferred phylogenetic arrangement. Most of the differences betweencliques (and between the dendrograms supported by these and the preferred arrangement)involve relatively uninformative characters. It is mathematically demonstrable that the prob-ability of chance compatibility of a particular character with any other set of characters isinversely proportional to the information value of the character. Therefore, one criterion ofwhich clique to select is to favour that with the highest information value. However, in practice one frequently finds that the subsets which are not common to all thesimilar-sized cliques comprise equally poorly informative characters. In such cases the averageO/E value for a clique is a useful indicator of the degree to which the clique is consistent withslightly homoplastic (and hence excluded) characters which are not normally considered in acompatibility analysis. For example, consider the hypothetical data set 2 (Tables 4, 5). There aretwo equally large, equally informative cliques: A (10) [1, 2, 3, 4, 5, 6, 7, 8, 9, 10] Average O/E value = 0-308B (10) [1, 2, 3, 4, 5, 6, 7, 8, 9, 14] Average O/E value = 0-323A DB = (1,2, 3, 4, 5, 6, 7, 8, 9} Table 4 Character state matrix for hypothetical data set 2. Conventions follow table 1. Table 5 Results of LeQuesne test on hypothetical data set 2. Conventions follow table 2. Character number: failures observed expected O/E ratio _3 : 3 9.86 0.3 6^ : 3 10.63 0.28 2 : 2 6.89 0.29 12 : 12 10.63 1.13 Ranking ratios 10 6 5 9 1 2 3 4 7 8 14 11 12 13 These sets differ in one character, possession of either 10 or 14. Character 10 supports J + Kwhereas 14 supports J + M. No other character that is compatible offers support for one or otherof these alternatives but some homoplastic characters do support J + K (plus other taxa also)reinforcing the J -I- K grouping suggested by character 10. No homoplastic character reinforcesthe J + M suggestion. This reinforcement can be detected by differences in the values of the O/Eratio, as character 10 scores 0-15 whereas character 14 scores 0-58. This difference in O/E valueis responsible for the difference in average O/E values of the two cliques A and B, and it thusseems reasonable to prefer the clique with the lowest average O/E value. This method offers asimple objective procedure for selecting one of a number of large cliques as 'most favourable' , asit will be the one reinforced by most secondary characters. In the present study compatibility analysis was found to give, up to a point, results that weresimilar to conventional taxonomic assessment. Frequently, characters considered important ingeneric classification by traditional taxonomists were found to be the intersect of the largestcliques. 70 I. D. GAULD The operational procedure adopted in this study This study was commenced by tentative recognition of groups of taxa from the shared derivedcharacter matrix (Appendix 3). The rigidity of these groups was tested using Underwood'sFOURS program (Underwood, 1982) and then further analyses using the compatibility andparsimony methods outlined above. Resulting cladograms were compared and differencessubjectively evaluated by appraisal of characters supporting different options. The variousgroups were gradually related to each other until the majority of taxa had been placed. Theremaining unplaced taxa all exhibited confusing affinities. To facilitate their placement anumber of hypothetical taxonomic units were proposed to represent the various genera alreadyplaced in the analysis. The data set comprising hitherto unplaced taxa and HTUs was analysedusing parsimony and compatibility techniques, and the resulting cladograms subjectivelyevaluated. An apologia for subjectivity in cladistic analysis In recent years there has been a great deal of intellectual activity devoted to making taxonomymore 'scientifically respectable' by attempting to remove subjectivity. Phenetic methods werethe first to be claimed as objective taxonomic procedures (Sokal & Sneath, 1963), but thesemethods were essentially a movement away from evolutionary taxonomy (Nelson & Platnick,1981) and have largely been eclipsed by the development of numerical cladistics. Despite theseeming objectivity of such numerical methods, they are really only objective ways of analysinglargely subjective data, as character selection is a highly subjective procedure (Pratt, 1972).Furthermore, the claim by many cladists that their favoured numerical methods (parsimonyanalyses) are scientific (in the Popperian sense) is flawed by the assumption that parsimony perse is a criterion for formulation of a scientific hypothesis (i.e. the least falsified hypothesis isaccepted - Gaffney, 1979), as is eloquently shown by Panchen (1982). A phylogenetichypothesis derived from a cladistic analysis rests on the supposition (or hypothesis) that aparticular derived character-state is a genuine synapomorphy which thus indicates commonalityof descent of two or more taxa. In a highly homoplastic data set, for every character that isaccepted as a genuine synapomorphy, others are rejected as showing false patterns. Farris(1969) defends parsimony in such cases, on the grounds that hierarchic correlations are morelikely to occur among cladistically reliable characters than among cladistically unreliablecharacters (i.e. random variables). This is true only if cladistically unreliable characters varyrandomly and is not a corollary of the generally accepted statement that random variables arecladistically unreliable characters. Clearly it is untrue; most taxonomists can cite examples ofsuites of characters being size related, habitat related, associated with a particular biologicalstrategy and so forth. How then is a taxonomist supposed to choose between competingcharacter sets if not by 'biggest is best' criteria? I think this can only be done at present,subjectively. Competing character sets are examined in the light of a taxonomist's knowledge ofthe variability of characters over a much wider group of organisms. Highly variable features(such as the number of ribs or vertebrae in some birds - Strauch, 1984 - or the form ofmammalian teeth - Butler, 1982) are rejected as unlikely to be indicative of phylogeneticaffinity. Other sets clearly comprise adaptive features associated with certain habitats and thesemay likewise be rejected. It is noteworthy that very many cladistic analyses seem to have beenundertaken on data sets that have initially been subjectively edited (with no explanation) bydiscarding characters which are presumably thought to have no phylogenetic value. In this workI have tried to explain my reasons for rejecting certain character sets in preference to others. The holophyly of the subfamily A prerequisite to meaningful cladistic analysis is the establishment of the holophyly of the groupbeing studied. The holophyly of the Ophioninae is supported by the following apomorphies. (a) Possession of a spurious vein extending from the vannal notch to the tornus of the fore wing.This feature, first recognized by Perkins (1959), is an autapomorphy of the subfamily. PHYLOGENY OF THE OPHIONINAE 71 (b) Possession of numerous setae on the labial sclerite of the final instar larva (Short, 1978).This feature is also an autapomorphy of the subfamily. (c) Presence of a single radio-medial cross vein (? 3r-ra) far distal to 2m-cu. This unusualcondition is elsewhere only found in isolated genera of the Campopleginae, Tersilochinae andAnomaloninae (Townes, 1970; 1971; Gauld, 1976). In none of these taxa is the cross vein as fardistal to 2m-cu as is the case in ophionines. (d) Possession of a sinuous pectinal comb on the tarsal claws. Many ichneumonids have a fewscattered pectinal teeth on the tarsal claws but very few have a strongly developed sinuous comblike that found in virtually all ophionines. (e) Total loss of glymmae. These structures are present in at least some members of virtually allother subfamilies. No trace has ever been observed in ophionines. (f) Enlargement of ocelli and related nocturnal specializations (Gauld & Huddleston, 1976).The pale colour, enlarged ocelli and eyes, long antennae and slender legs are features found inmany nocturnal ichneumonoids, but this suite of characters is apparently an apomorphic featureof the Ophioninae. The systematic position of the subfamily The subfamily Ophioninae belongs to the Ichneumonidae, a holophyletic group of apocriteHymenoptera (Konigsmann, 1978). Virtually nothing has been published about the phylo-genetic relationships of ichneumonid taxa and, if a reasonable assessment is to be made ofcharacter polarity in Ophioninae, it is necessary to attempt to place the group in relation to someother subfamilies. From study of Cretaceous fossil ichneumonoids (Townes, 19730 & b; Rasnitsyn, 1983),comparison with putative ancestral groups (such as siricids and cephids) and examination ofprimitive extant Apocrita (e.g. Megalyridae, Stephanidae, some Braconidae), it can be deducedthat primitive ichneumonpid features are likely to include: possession of an un-notched(possibly long) ovipositor; possession of a gaster that is broadly attached to the propodeum andhas a free first sternite; possession of an areolated propodeum; possession of notaular grooves;possession of simple or basally lobate claws. Study of the biological evolution of parasitoids (e.g.Shaw, 1983) suggests that the ectoparasitic habit is primitive with respect to endoparasitism, andas both types occur in the Ichneumonidae (Clausen, 1940; Gauld, 19846), one expects that themost primitive taxa will be found amongst ectoparasitoids. Some authors (e.g. Cushman, 1926;Telenga, 1969; Achterberg, 1976; 1984) suggest that primitive ichneumonoids are likely to havebeen parasites of xylophagous coleopterous larvae, as are archaic extant parasitoids such asorussids (Quinlan & Gauld, 1981; Middlekauf, 1983). These data suggest that the most primitiveextant ichneumonids are to be found amongst the Pimplinae, Labeninae and perhaps alsoTryphoninae. This hypothesis is supported by studies of larval morphology (Short, 1978) and isconcomitant with the informal higher classificatory scheme outlined by Townes (1969). Compared with the Pimplinae and Labeninae, the Ophioninae appears to be a ratherspecialized subfamily. It seems to belong to a large, holophyletic group of taxa that includes thesubfamilies Ophioninae, Campopleginae, Cremastinae, Tersilochinae, Banchinae and Cteno-pelmatinae. The holophyly of this group is suggested by the following apomorphies: (a) possession of a dorsal subapical notch on the ovipositor; (b) possession of similar female reproductive tract (Pampel, 1913); (c) having endoparasitic larvae which lack a labral sclerite and possess a Y-shaped prelabialsclerite (see also Fig. 2). Two of the included taxa, the Ctenopelmatinae and Banchinae, usually possess a plesio-morphic first gastral segment which is broadly attached to the propodeum, has a more or lesscentrally positioned pair of spiracles and a free sternite. In the Ophioninae, Campopleginae,Cremastinae and Tersilochinae this segment is specialized. It is more slender, has the tergite andsternite intimately fused and is lengthened anteriorly so that the spiracles are nearer to theposterior end. Furthermore, all these taxa possess apomorphic, laterally compressed gasters,suggesting they constitute a holophyletic clade. m m 14 L5 Fig. 2 Cladogram showing putative phylogenetic inter-relationship of the ophionoid group of ichneu-monid subfamilies. The apomorphic features supporting this arrangement are: 1 , possession of a fringedclypeus ; 2 , fusion of Rs with M in central part of fore wing ; 3 , development of sclerotized bridge betweentibial spurs; 4, possession of striae on tergite 2 of gaster; 5, loss of distal abscissae of veins in hind wing; 6,fusion of areae superomedia and petiolaris; 7, development of spurious vein in fore wing; 8, loss of 2r-min fore wing ; 9 , enlargement of pterostigma ; 10 , reduction of length of Rs in hind wing ; 1 1 , developmentof sinuous apex to ovipositor; 12, possession of expanded larval labial sclerite; 13, development oftubular petiole with reduced glymmae (note this is developed in parallel in some tersilochines); 14,gaster laterally compressed; 15, petiolar spiracle near hind end of tergite; 16, tergite and sternite 1intimately associated; 17, possession of a dorsal subapical notch on ovipositor; 18, possession of asimilarly modified female reproductive tract (see Pampel, 1913); 19, endoparasitic larva with a Y-shapedprelabial sclerite. PHYLOGENY OF THE OPHIONINAE 73 Within this clade two apparent groups are recognizable. One comprises the Tersilochinae +Cremastinae and is supported by two apomorphies, possession of an enlarged pterostigma andstrongly shortened vein Rs in the hind wing. In addition both these taxa share a number ofdevelopmental trends (underlying synapomorphies of Saether, 1979), most notably, the de-velopment of a sinuous ovipositor apex and the presence of dorsally convergent eyes in males.The second group, Ophioninae + Campopleginae, is supported by a larval specialization, thepossession of an expanded labial sclerite. The petioles of these two taxa are similarly modified,although this development appears to have been paralleled in some tersilochines (primitivetersilochines have a more 'cremastine-like' petiole). Most ophionines and campopleginespossess an apomorphic short, straight ovipositor; even when it is long it is very similar in the twogroups, being robust and up-curved. For the purposes of this study the Ophioninae is considered to be the sister-group of theCampopleginae, and the Campopleginae + Ophioninae is treated as the sister-group of theCremastinae + Tersilochinae (Fig. 2). Polarity determination and character coding The polarity of the majority of characters has largely been determined by the method ofout-group comparison (Watrous & Wheeler, 1981) using as out-groups the taxa mentionedabove. Unfortunately this method does not work for all characters, usually because both statesoccur in both the group under study and the out-groups. In these cases the polarity assignment isbased on unsatisfactory criteria such as common equals primitive, but this is stated in thediscussion. In many cases in the Ichneumonidae, character-states are progressive steps in the loss ofancestral features such as the occipital carina, propodeal carinae or segments of the palp. Suchcharacters can be arranged in simple transformation series, from plesiomorphic to the mostderived state. For example, -for maxillary palps - 5-segments to 4-segments to 3-segments. Suchtransformation series may simply be scored in binary form as two characters - 8-1 8-2 maxillary palp 5-segmented maxillary palp 4-segmented 1 maxillary palp 3-segmented 1 1 where = plesiomorphic and 1 = apomorphic. Hence 8-1 represents reduction from 5 to 4segments and 8-2 further reduction to 3. Such characters are obviously not independent for it isimpossible to have a 0,1 coding as, if extreme reduction is observed, it is assumed that anintermediate stage has been passed through. In other cases a particular structure may have been modified in one of several ways. Forexample, in the Ichneumonidae generally the mandible is bidentate equally, so in the Ophio-ninae teeth of the same length must be considered a plesiomorphic feature. Although the lowertooth is fairly constant in size, the upper may either be lengthened or shortened. Such a bifurcateseries can be scored as two binary characters - 4-1 4-2 mandible equally bidentate upper tooth enlarged 1 upper tooth reduced 1 Thus character 4-1 represents enlargement of the upper tooth, whilst 4-2 represents reduction.Clearly a 1,1 coding cannot exist for such character pairs. All transformation series in thecharacters dealt with below have been treated in similar fashion. From the very large number of characters exhibited by the Ophioninae a number have beenselected which show the greatest range of variation between the different genera. Included areall characters previously considered to be diagnostic of genera, and a number of additionalfeatures that have been found to be useful in delineating species-groups (e.g. in Gauld & 74 I. D. GAULD Mitchell, 1978). Many other characters, features in which a single species differs from others in aspecies-complex, were excluded from this analysis. These characters, such as the shape of thealar sclerites, form of the microsculpture of the alitrunk, density of hairs on the wing surface,general colour pattern, relative length of the mid tibial spurs etc. , show a considerable range ofvariation within the species of any single genus and, in most cases, this range of variation isrepeated in many genera. Such characters were considered to be unlikely to make significantcontribution to resolving the phylogeny of the genera. Characters used in the study In the following section considerable space has been allocated to explaining the rationaleinvolved in character scoring and polarity determination. Although this practice is not usual inall cladistic studies (characters are frequently relegated to an appendix) it is considered to beimportant in the present work. It is upon these scores that the results of any numerical analysisdepend; consequently, this section is seminal to the entire work. The characters and their polarity 1-1, 1-2 Occipital carina. A complete occipital carina is plesiomorphic (0,0). It may be centrallyinterrupted (1,0) or absent (1,1). 2 Mandibular axis. The plesiomorphic condition is for the axis of articulation of the mandible to be at 90 to the longest axis of the head. In a few ophionines the head is narrowed and the mandibular axis tiltedalmost into the vertical plane (1). 3 Mandibular swelling. A flat outer mandibular surface is plesiomorphic for ichneumonids; the swelling present near the mandibular base of some ophionines is apparently a derived feature (1). 4-1, 4-2 Mandibular teeth. The relative lengths of the teeth vary considerably in ichneumonids, but equallybidentate appears to be the plesiomorphic state (0). In ophionines the alternative specializations are -upper tooth elongate (1,0) and upper tooth reduced (0,1). 5-1,5-2 Torsion of the mandibles. The plesiomorphic condition for ichneumonids is to have the teeth of themandible aligned in the same plane as the mandibular axis (Figs 38, 39). Many ophionines have themandible twisted from 5-50 so that the lower tooth is directed forwards (Fig. 40). In a few species themandible is exceptional in being twisted more than 70 so that, when closed, the lower tooth occludesthe internal, upper one. This torsion is here considered a serial development, from untwisted (0,0)through slightly twisted (1,0) to exceptionally twisted (1,1). The selective advantage of this torsion isnot clearly understood, but it is suggested that it is important in facilitating egress from certain types ofcocoon. It is noteworthy that not only do all species of the huge and successful genus Enicospilus havesuch mandibles, but the totally unrelated tryphonine genus Netelia (which is also nocturnal and has asimilar host range) also has twisted mandibles. 6-1, 6-2 Shape of the clypeal margin in anterior aspect. The shape of the clypeus is often characteristic of agroup of ichneumonids. For example, in the pimplines it is usually bilobate, in the anomaloninesconvex with a central tooth, whilst in the ichneumonines it is usually flat and truncated. The weaklyconvex condition found in many ophionines, campoplegines and cremas tines is considered plesio-morphic. In the Ophioninae the clypeus may be modified one of two ways - either being stronglyconcave (1,0), or centrally produced and pointed (0,1). 7-1, 7-2 Clypeal profile. It is difficult to assign polarity to this character, but the widespread condition in thesubfamily and in the closely related out-groups was assumed to be plesiomorphic. This is thepossession of a virtually flat clypeus (0,0) (Fig. 39). The clypeus may be modified in one of two ways -either flared outwards (1 ,0) (Fig. 44) or with a groove present parallel to the margin, so that the actualmargin is sharp (0,1) (Fig. 38). 8-1, 8-2 Number of maxillary palp segments. The plesiomorphic condition for the Hymenoptera is6-segmented palps, but in the Ichneumonidae the number is reduced to 5. Fusion may occur betweenthe distal two or three segments leading to a reduction in apparent segment number to 4 or 3. Thischaracter is scored as a transformation series, 5-segmented being 0,0, 4 being 1,0 and 3 being 1,1. 9 Shape of central segments of maxillary palp. In most primitive ichneumonids, virtually all members of the out-groups and many ophionines, the maxillary palp segments are slender and elongate. This isconsidered to be plesiomorphic. In a few ophionines the central segments are specialized, beingglobose. 10 Maxillae. The maxillae of most ichneumonids resemble those of the more primitive mandibulate PHYLOGENY OF THE OPHIONINAE 75 hexapods (Richards, 1956). A few ophionines have the maxillae specialized, unusually elongated sothat the galea projects below the mandibles. This apomorphic condition is found in species inhabitingdry areas and is presumed to be an adaptation to feeding from certain flowers. 11 Labium. Like the maxillae, the labium is remarkably unspecialized in most ichneumonids. A few eremic ophionines possess greatly elongated glossae, an apomorphy that presumably has a similarfunction to character 10. 12 Ocelli. The majority of ophionines exhibit a set of features, the so-called ophionoid facies (Gauld & Huddleston, 1976), common to nocturnal Hymenoptera. These features are here regarded asapomorphies of the entire Ophioninae. They include possession of greatly enlarged ocelli. Themajority of species in the out-groups, being diurnally active, lack this specialization. Consequently thepossession of enlarged ocelli must, within the Ophioninae, be regarded as a plesiomorphy. In a fewareas where competition from other ichneumonids is low, e.g., remote islands, deserts, tops of highmountains, several ophionines have adopted a diurnal habit. This is presumed to be a secondaryfeature as several species retain some nocturnal features. However, these diurnal species have smallocelli, and in these cases small ocelli are considered apomorphic. Polarity determination of thischaracter can be questioned as it is based on an a posteriori inference. 13 Frontal grooves. Most ichneumonids, including almost all members of the out-groups, have the lateral part of the frons flat. In some ophionines a groove is present on either side, parallel to the inner orbit.Possession of this groove is regarded as a specialization. 14-1, 14-2 Flagellum length. In most ichneumonids the flagellum is setaceous and consists of a largenumber of more or less identical segments (flagellomeres). In the majority of ophionines there arebetween 45 and 65 such segments, and the flagellum is about as long as the fore wing. This isconsidered, because of its common occurrence within the group, to be the plesiomorphic condition.The flagellum appears to be modified in one of two ways. Either the segments may be very short andtransverse (scored 1 ,0) or the number may be greatly increased (to between 75 and 95) so the flagellumis very much longer than the fore wing (0,1). A short flagellum is found in many species inhabiting hot,dry areas, and is perhaps an attempt at reducing an evaporative surface. A very long flagellum is foundin species inhabiting humid areas, particularly rain forest canopies. Similarly shortened or elongatedantennae can be observed in other groups, e.g. the Anomaloninae (Gauld, 1976). 15-1, 15-2 Spiracular sclerite. In most ichneumonids, including the majority of species in the closely relatedout-groups, the spiracular sclerite is exposed and is clearly visible near the upper hind corner of thepronotum (Fig. 42). This is considered to be plesiomorphic. A large number of ophionines have theupper corner of the pronotum somewhat broadened and notched to partially occlude the spiracularsclerite (1,0), whilst in a few species this flap completely covers the sclerite (1,1) (Fig. 43). 16-1, 16-2 Notauli. The notauli are grooves in the mesoscutum that appear to extend backwards from theanterior margin, and at their most extreme, reach the scuto-scutellar groove. The presence of thesegrooves seems to be a plesiomorphic feature (0,0), as they are well developed in many sawflies and arevisible in many fossil apocritans. However, notauli are not present in all members of the out-groups sosome doubt remains about the correct polarity of this character. The majority of ophionines only havevestigial notauli impressed at the extreme anterior margin of the mesoscutum (1,0), whilst in severalthey are absent entirely (1,1). 17 Pronotal crest. The plesiomorphic condition for ichneumonids is possession of a more or less flatpronotum dorsally. In several ophionines a crest is present to protect the neck region, possibly againstattacks by asilids which habitually kill Hymenoptera by piercing the cervical region. The presence ofthis crest is regarded as an apomorphy. 18-1, 18-2 Mesopleural furrow. The plesiomorphic condition of the ichneumonid mesopleuron is for it tobe relatively flat with a small pit (the episternal scrobe) near to the middle of the mesopleural suture(Townes, 1969). Amongst ophionines two specializations have apparently arisen - the presence of adiagonal groove from the pit to just below the subalar prominence (1 ,0) or possession of a groove fromthe pit to the upper end of the epicnemial carina (0,1). 19 Epicnemial carina. It is a plesiomorphic feature for ichneumonids to have this carina complete,extending laterally from the medioventral line to near the subalar prominence. In a number of taxa thelateral portion of this carina is lost (1). 20-1, 20-2 Scutellar carinae. These carinae are of sporadic occurrence throughout the Ichneumonidae butin many of the more primitive groups such as Pimplinae they are not developed. In the majority ofspecies in the out-groups these carinae are absent, suggesting that their absence in ophionines may bea plesiomorphic feature. When present these carinae may be short, reaching to or not quite to thecentre (1,0), or they may be virtually complete, reaching 0-8 or more of the length of the scutellum(1,1). 76 I. D. GAULD 27-7, 27-2 Metanotal protuberances. The hind rim of the metanotum of most ophionines, as well as mostcampoplegines and cremastines, is unspecialized. In a few ophionines a small lateral tooth isdiscernible (1,0), whilst in some taxa this tooth is apparently enlarged to form a protuberance thatextends back almost to the propodeal spiracle (1,1). 22-7, 22-2 Propodeal anterior area. The anterior part of the propodeum (Fig. 50), immediately behind themetanotum, is, in ichneumonids, characteristically depressed to form a transverse groove which isoften somewhat broadened medially behind the postscutellum. This presumably plesiomorphic stateis found widely in Ophioninae and in virtually all members of the out-groups. Two modifications fromthis pattern have been observed. In some species the groove is broadened (lengthened) and striate,and thus forms a broad shallow concavity, almost as if the insect had been stretched (Fig. 51) (1,0).The other adaptation is for the groove to be much deeper and present as a U-shaped furrow (Fig. 52)(0,1). 23 Propodeal spiracle. In the majority of ophionines and members of the out-groups the propodealspiracle is oval or even subcircular, with the longest axis 4 or less times as long as the shortest. A fewophionines have very large, very elongate spiracles which are 8 or more times as long as broad. This isconsidered to be an apomorphic development. 24-1, 24-2 Anterior transverse carina of the propodeum. The presence of this carina, like the otherpropodeal carinae, is a plesiomorphic feature of ichneumonids as they are complete in fossils(Townes, I913b). The reduction of this carina is a progressive feature and has been coded as follows -complete (0,0), present only centrally (1,0), absent (1,1). 25-7, 25-2 Posterior transverse carina of the propodeum. Coded as for character 24 for the same reasons. 26 Longitudinal propodeal impression. The plesiomorphic condition for ichneumonids is to have thedorsum of the propodeum more or less flat. In a few ophionines a deep longitudinal impression ispresent and this is considered to be an apomorphic feature. 27-7, 27-2, 27-3 Posterior transverse carina of the mesosternum. This carina is usually complete incampoplegines and cremastines and thus a complete carina is here considered plesiomorphic for theOphioninae. The carina is usually lost at two points on either side of the midline, before the mid coxae,so that central and lateral vestiges remain (1,0,0). Sometimes the central vestige is completely lost soonly lateral traces remain (1,1,0). In a few species reduction seems to have occurred by loss of only thecentral part so two broad lateral portions remain (0,0,1). 28-1, 28-2 Lateromedian longitudinal carina of the propodeum. Coded as for character 24 for the samereasons. 29-7, 29-2 The thyridium. This is a moderately large indentation found on the second gastral tergite. In itsplesiomorphic condition it is close to the anterior margin of the tergite (0,0), but in a number ofophionines it can be seen to have been displaced posteriorly, leaving a scar between itself and thetergal margin (0,1). In a few taxa the thyridium is absent (1,0). 30 Epipleuron of tergite 2. This is a difficult feature to determine the polarity of as within the out-groups both conditions occur widely. In most ophionines it is upturned, and in a few otherwise specializedtaxa it is pendant. The latter condition is tentatively considered to be apomorphic. 31 Profile of tergite 2. Tergite 2 is laterally compressed in ophionines and members of the closer out-groups (Fig. 2). In profile it is much longer than posteriorly deep. In a few ophionines it is quadrate and this isconsidered to be an apomorphic development. 32 Position of spiracles on tergite 1. The plesiomorphic condition for ophionines and members of the out-groups is for the spiracles to be at or behind the level of the margin of the sternite. In a few taxa thespiracles are situated before the sternite margin, a presumed apomorphic condition (1). 33 Presence of an umbo on tergite 2. The umbo is a convex area on the midline at the anterior margin of tergite 2. It is typically present in many ophionines and members of the out-groups. The apomorphicstate (1) is where this structure has been lost. 34 Ovipositor length. The length of the ovipositor varies a great deal between ichneumonid taxa. Probably the ancestral condition for the family was long, but almost every evolutionary lineage showsreduction. Virtually all ophionines have short straight ovipositors resembling those of many campo-plegines (Townes, 1970), and thus this condition is considered to be plesiomorphic. The longovipositors found in a few taxa are considered to be apomorphic features (1). 35 Ovipositor sheath. The ovipositor sheaths (valvulae 3) of ichneumonids are almost always slender, just wide enough to enclose the ovipositor. In a few ophionines the sheaths are very stout. This isconsidered to be an apomorphic development (1). 36-7, 36-2, 36-3 Position oflm-cu in relation to Cu la in the fore wing. In most ichneumonids these two veinsare quite widely separated and are often at least as far apart as 0-75 of the length of Cu lb . With somereservation this is considered to be the plesiomorphic state (0,0,0). In many ophionines these veins are PHYLOGENY OF THE OPHIONINAE 77 separated by about 0-5 times the length of C lb (1,0,0), in a number of species by about 0-25 times(1,1,0), whilst in a very few instances the veins are practically contiguous (1,1,1). 37 Length of second discal cell. The plesiomorphic condition for ichneumonids generally, including ophionines, is for the length of the second discal cell (measured along C la ) to exceed 1-10 times thelength of the first subdiscal cell (measured along Cui). In a number of ophionines the second discal cellis unusually short, being less than the length of the first subdiscal cell. This is considered to be anapomorphic feature (1). 38 Presence of a ramellus. The ramellus on Im-cu in the fore wing is apparently a remnant of the vein that divided the first discal from the submarginal cell (the two are confluent in all extant ichneumonidsthough present in fossil forms (Townes, 19736)). The retention of this stub is tentatively considered tobe a plesiomorphic feature, but I have some doubts about the polarity assignment for similar reasonsto those stated in 47-1 below. 39 Base o>/Rs+2r in fore wing. In the majority of ophionines and members of the out-groups this vein is emitted from the pterostigma at about 30 to the fore margin of the wing; the base of the vein is straight(Fig. 3). In some ophionines the angle of emission is greater (40+) and the vein is rather sharply bentbasally (Fig. 4). This is considered to be an apomorphic feature (1). 40 Shaft o/Rs+2r in fore wing. The plesiomorphic condition for Hymenoptera generally is for this vein to be centrally straight; this condition is found widely in out-groups and many ophionines (Fig. 3). Insome ophionines the vein is markedly sinuous just before the centre and this (Fig. 6) is considered tobe an apomorphic development (1). 41 Position of confluence o/Rs and R! in fore wing. In virtually all ophionines and many members of the out-groups these veins meet at the extreme distal apex of the wing so the marginal cell is very long. In afew ophionines the confluence of these veins is more proximal on the fore margin of the wing, awayfrom the tip, so the marginal cell is correspondingly shorter (Fig. 4). This is considered to be anapomorphic feature (1). 42-1, 42-2 Shape of pterostigma. In most ichneumonids the pterostigma is quite broadly triangular and thisplesiomorphic condition prevails in most species in the out-groups. In some ophionines the ptero-stigma is still broad (0,0) but in many it is not abruptly narrowed distally but evenly tapered (1 ,0) (Figs4, 6). In a few the pterostigma is much more slender and elongately tapered (1,1) (Fig. 5). 43-1, 43-2 Position of distal abscissa of Cu } in hind wing. It is very difficult to assign polarity to thischaracter as all stages from Ci close to M to Cui close to L4 can be found in members of theout-groups. Furthermore, in most Campopleginae this vein is lost. However, in some of the moreprimitive ichneumonids the usual condition is for this vein to be closer to M than to L4 . This istentatively assumed to be the plesiomorphic condition (0,0). The derived states, treated as atransformation series are: intermediate between M and \A (1,0) and closer to IA than to M (1,1). 44-1, 44-2 Shape o/Rs in hind wing. The plesiomorphic condition of this vein in ichneumonids generally,including many ophionines, is for it to be straight (0,0). In many ophionines it is quite distinctly Figs 3-6 Fore wings: 3, Ophion; 4, Simophion; 5, Stauropoctonus; 6, Enicospilus. 78 I. D. GAULD concave whilst in a few taxa it is exceptionally curved so that the distal portion is parallel to the foremargin of the wing. These two derivations are treated as a transformation series and coded 1,0 and 1,1respectively. 45-1, 45-2 Glabrous area in fore wing. All primitive ichneumonids, most taxa in the out-groups, andvirtually all lower Hymenoptera have the membrane of the fore wing uniformly hirsute close to veinRs+2r. This is treated as the plesiomorphic condition (0,0). In most ophionines a small glabrous areais present in the discosubmarginal cell near where the pterostigma emits Rs+2r. This is consideredapomorphic (1,0). A further specialization is for this area to be very extensive, reaching at least 0-3 ofthe way along Rs+2r (1,1). 46 Alarsclerites. The possession of alar sclerites is an unusual apomorphic feature of certain ophionines.Other than in this subfamily these sclerites are only found in two species of ichneumonid (and then indifferent positions). Elsewhere in the Hymenoptera alar sclerites are apparently only found in somepepsine pompilids (M. C. Day, pers. comm.). 47-1, 47-2 Shape of 7m-cu in fore wing. In the lower Hymenoptera this vein (which in fact is almostcertainly not just \m-cu but a composite of this and other vein parts) is angled centrally so theantero-proximal side of the 2nd discal cell is angulate at about 90. In many ophionines this angulationis present (Fig. 3) and often accompanied by a ramellus (see character 38). This is here tentativelyconsidered to be the plesiomorphic condition (0,0). In many species this vein is evenly curved (1,0),whilst in some it is further modified by being sinuous (1,1). I have certain reservations aboutthe polarity of 47-1 as the evenly curved condition is the widespread condition amongst the out-groups. 48-1, 48-2 Length of3r-m in the fore wing. In the majority of more primitive ichneumonids 3r-m is longerthan the abscissa of M between 2m-cu and 3r-m. In many ophionines this is also the case and it is hereconsidered to be the plesiomorphic state (0,0). In some ophionines 3r-m is shorter, 0-75-0-50 times aslong as M (1,0), whilst in a very few it is extremely reduced, 0-35 or less times the length of M (1,1). 49 Position of abscissa ofCuj between 7m-cu and Cu la . This vein is positioned at about 90 to the axis of the fore wing in most ichneumonids and virtually all lower Hymenoptera. In some ophionines it isspecialized in being turned so that it is almost parallel to the wing axis (1). 50 Position of cu-a in fore wing. The plesiomorphic cdndition of this vein in ichneumonids is for it to subtend an angle of about 90 to Cu^. In a few ophionines it is strongly oblique, subtending an angle of 50-60 (1).51-1, 51-2 Flange on fore tibial spur. The plesiomorphic state for ichneumonids is the possession of a membranous flange on almost the entire length of the fore tibial spur, immediately behind the microtrichial 'comb' (Fig. 48) (0,0). In some ophionines this is reduced to about 0-3 of the length of the spur (1,0), whilst in others it is entirely lost (Fig. 49) (1,1).52-7, 52-2 Form of hind trochanter. The possession of a simple margin to the trochanter is plesiomorphic for ichneumonids (0,0). In a few ophionines a small marginal tubercle is present (1,0), whilst in some cases this may be long, curved and pointed (1,1). 53 Dorsal margin of hind trochanter. This is also unspecialized in most ichneumonids, but some ophionines are unusual in having a specialization - the margin extended as a flange over the articulation of thetrochantellus (1). 54 Hind tibial spurs. In many ophionines and members of the out-groups these spurs are flattened internally and bear a long fringe of close hairs. In some ophionines they are specialized in beingcylindrical and bearing only scattered hairs (1). 55 Penultimate distal hamulus. The unspecialized condition in Hymenoptera is for the hamuli to be of similar size and shape. In some ophionines the penultimate hamulus is specialized in being very muchlonger and more coiled than its fellows (1). 56 Grouping of distal hamuli. The plesiomorphic condition for Hymenoptera is for the row of hamuli to be fairly evenly spaced. In a few ophionines they are arranged in two groups, an apparently specializedcondition (1). 57 Number of hamuli distally. Most primitive Hymenoptera have from 9 to about 15 distal hamuli and, as this condition is widespread in primitive ichneumonids and many ophionines, it is consideredplesiomorphic in context of this study. Some ophionines have only 4 or 5 distal hamuli and thisreduction is considered to be an apomorphic condition (1). 58-1, 58-2 Shape of hind tarsal claw. In most members of the out-groups and many primitive ichneumonidsthe tarsal claws are fairly evenly curved (Fig. 45) and this condition, where it occurs in ophionines isconsidered to be plesiomorphic (0,0). In Ophioninae the claw may be modified in one of two ways -either by being longer and straighter (1,0) or by being almost geniculate (0,1). 59 Colour of inter ocellar area. In most Hymenoptera this area is concolorous with the vertex. In a few PHYLOGENY OF THE OPHIONINAE 79 ophionines it is sharply chromatically contrasted, and this is tentatively considered an apomorphic feature (1).60 Distal pectinal tooth of hind tarsal claw. The claws of ophionines possess a sinuous row of pectinal teeth or pectinae (Gauld & Mitchell, 1978) forming a comb-like internal surface. This is apparently an apomorphy of the subfamily, so in the present context such a claw may be considered a plesiomorphy. In some ophionines the claw is specialized in having the distal pectinal tooth on the outer surface of the main tooth (Fig. 46) (1).61-1, 61-2 Taper of mandibular teeth. In most primitive ichneumonids the mandible is weakly tapered so that the distal apex is more than 0-5 times as broad as the base. This is considered to be the plesiomorphic condition where it occurs in ophionines (0,0). Some species have the mandible more strongly tapered so the apex is 0-4-0-5 times as wide as the base (1,0) whilst a few are exceptional in having the apex <0-3 times as wide as the base (1,1). 62 Thickness of base of Rs+2r. The plesiomorphic condition for Ichneumonidae is for this vein to be slender at its junction with the pterostigma. In some ophionines it is specialized in being broadened sothat close to the pterostigma it is more than twice its central thickness (1). 63 Presence of laterotergite 1 of gaster. In primitive ichneumonids a distinct laterotergite is present associated with tergite 1 ; vestiges of this may be found by partial dissection in many ophionines. Someare specialized in having lost this structure (1). 64 Presence of a pectinal comb on both sides of male claw. The plesiomorphic condition for ophionines is for a single sinuous pectinal comb to be present on the tarsal claws; some males are specialized inhaving this comb present as a double row of teeth, surrounding a flattened central area (1) (Fig. 47). Preliminary remarks on the relative values of the characters For analytical purposes all characters have been treated initially as having equal numericalweight. However, a practising taxonomist does not value characters equally. Intuitively he ismore likely to consider some to be more important than others. 'Good' characters are likely toinclude many of the more bizarre morphological adaptations (Gauld & Mound, 1982). At theother extreme, characters may be considered to have very little value as indicators of phylo-genetic affinity. This is often for one of three reasons. (a) The apomorphic state of the character may involve the disappearance of a structure and, asit is obvious that certain structures are frequently lost in parallel in many distantly relatedlineages (e.g. the occipital carina has been lost in some taxa independently in virtually allichneumonid subfamilies), one suspects that an absence in two closely related taxa could also bethe result of parallel reduction. (b) There may be some doubt concerning polarity. This is particularly true in cases where onepostulates that a particular state is an apomorphy for the subfamily, but advocates reversal in afew exceptional taxa (e.g. character 12, ocellar size). (c) Even if the apomorphic condition involves the development of a particular structure, thecharacter may be judged to be unstable, as examination of out-groups suggests it has beenderived independently in numerous evolutionary lineages, or is correlated with a particular sizerange or extreme habitat occupied by a particular species. For example, very large ichneu-monids often have similarly coarse reticulate propodeal sculpture, whilst eremic species haveshort, stout antennae. The third of these reservations is discussed for individual characters in a subsequent section,but it is appropriate to mention a and b here. Those of the characters used in this study whichinvolve an apomorphic 'loss' are listed below with an asterisk(*), whilst those in which thepolarity assignment is suspect are denoted by a prime mark('). This convention is usedextensively in the cladograms presented in the following sections. For comparative purposes alist is given below. 1-1*, 1-2*, 2, 3, 4-1, 4-2*, 5-1, 5-2, 6-1, 6-2, 7-1', 7-2', 8-1*, 8-2*, 9, 10, 11, 12*, 13, 144', 14-2',15-1, 15-2, 16-1*', 16-2*', 17, 18-1, 18-2, 19*, 20-1, 20-2, 21-1, 21-2, 22-1, 22-2, 23, 24-1*, 24-2*,25-1*, 25-2*, 26, 27-1*, 27-2*, 27-3*, 28-1*, 28-2*, 29-1, 29-2*, 30', 31, 32, 33*, 34', 35, 36-1',36-2', 36-3, 37, 38*', 39, 40, 41, 42-1, 42-2, 43-1', 43-2', 44-1, 44-2, 45-1, 45-2, 46, 47-1', 47-2, 80 I. D. GAULD 48-1, 48-2, 49, 50, 51-1*, 51-2*, 52-1, 52-2, 53, 54, 55, 56, 57*, 58-1, 58-2, 59', 60, 61-1, 61-2, 62,63*, 64. The phylogenetic analysis An initial attempt was made to construct a Wagner tree, but with repeated runs it becameobvious that there exists a large number of equally or almost equally (less than 1% difference inoverall length) short trees, many of which exhibit quite different topologies. As there is no basison which to prefer one of these, and there is no way of knowing whether all the shortest treeshave been computed, Wagner analysis was initially rejected in favour of a more detailed study ofcharacters and taxa using the programs LEQU.BAS and FOURS. From these it was possible togain some indication of the 'shape' of the primary data set. Tentative groups of taxa recognizedinitially were reanalysed using Wagner and compatibility methods. Preliminary analysis of data Analysis using LeQuesne test program, LEQU.BAS Using the program LEQU.BAS the primary data matrix (Appendix 2) of 51 selected ophionine taxa was analysed (see Table 6). Of the 95 characters used, six, 8-2, 11, 28-1, 34, 49 and 53, Table 6 Results of LeQuesne test on the primary data matrix (Appendix 2). Conventions follow table 2.Note that a character with a single incidence of one or other state logically cannot fail the test and isrepresented by . Character number: incompatibilities observed expected 0/E ratio Grand total- 2598 observed. 3007.1 expected. Overall 0/E ratio = 0.86Ranking ratio of scoring characters 64 8.1 21.2 27.3 16.1 6.2 25.1 55 2 3 40 60 52.2 52.1 46 7.213 42.2 21.1 22.2 22.1 15.2 14.2 45.2 9 36.3 5.1 7.1 51.2 48.2 58.1 14.1 24.1 51.1 48.1 15.1 63 45.1 30 12 17 42.1 57 6.1 18 28.2 24.2 1.2 37 25.2 27.1 59 18.2 47.2 43.2 58.2 27.2 36.2 2929.1 43.1 16.2 31 44.1 41 44.2 56 50 36.1 10 4.2 5.2 47.1 35 54 2320.2 26 38 33 61.1 1 4.1 61.2 62 32 2 20.1 1.1 19 39 PHYLOGENY OF THE OPHIONINAE 81 showed only a single discordant state. In the case of 28-1 and 49 the majority state was thederived condition, whilst in the remainder it was the plesiomorphic condition that predomi-nated. As a minimum of two discordant character-states is logically required for a failure, itfollows none of these eight can fail; they are shown as . The remaining 89 characters have two or more discordant states and therefore can potentiallyfail the test. Each character was paired with every other one (except for the alternatives n. 1 with.2) and of the 7,855 comparisons made 2,598 resulted in failures. A total of 3,007 failures wasexpected, so although slightly better than random (O/E value = 0-86), the results suggest a veryhigh incidence of homoplasy. For individual characters a considerable variation in O/E ratio wasobserved. Characters 6-2, 84, 16-1, 21-2, 25-1, 27-3 and 64 all scored better than 0-55, whilst atthe other extreme eight characters, 4-2, 10, 36-1, 41, 44-1, 44-2, 50 and 56 all scored worse than1-00. Examination of the labels matrix (Appendix 4) showed that two taxa, 909 and 922, had beenlabelled exceptionally frequently (scoring 89 and 85 labels respectively). Six further species, 906,908, 921, 923, 942 and 951, were labelled between 45 and 67 times, whilst the rest were labelledless than 42 times. Sixteen taxa were labelled less than ten times and three, 912, 913 and 933 werenot labelled at all. As would be expected, the elimination of bad characters (i.e. those with an exceptionally highO/E score) produced, in general, a progressive improvement in the O/E ratios of the survivingcharacters. However, the average ratio improved very slowly. With 20 informative charactersremaining it was still 0-34. This clearly shows that not all the failures of the 'better' characterswere due to clashes with the worst characters; the majority of characters seem to be somewhathomoplastic! Complete elimination of homoplasy (as advocated by LeQuesne, 1972) results inthe removal of virtually all characters. The largest cliques contain only ten compatiblecharacters. Several of these have only a very low information value (i.e. have only two or threederived states), but five more informative characters (2, 6-2, 15-1, 15-2 and 22-2) were found tobe universal to all the larger cliques. All of these are features that have been used previously inichneumonid higher classification (Cushman, 1947; Townes, 1971). These characters supportthe cladogram shown in Fig. 7. Only the group supported by character 2 corresponds with anypreviously recognized grouping of taxa, the Thyreodon group of Cushman (1947). 905 909910-911916-918923 948938 939941 906 927 935 937 901 904 907 908 921 912 915 919 920 928 934 940 945 " 922 924 926 932 934 942 944 946 947 15.2 15.1 22.2 Fig. 7 Cladogram supported by largest clique of compatible characters. 82 I. D. GAULD The rate at which the O/E values of individual characters improved (after removal of badcharacters) varied tremendously. For example, after removal of the eight worst characters(Table 7), character 29-1, initially the seventy-eighth placed (and the twelfth one scheduled forelimination) progressively improved (with the elimination of each worse character) to sixty-seventh. Character 59, initially placed as sixty-sixth, steadily worsened its position to eighty-third ; it would in fact be the tenth character eliminated . If the ten worst characters were removeden bloc they would have been 4-2, 10, 36-1, 50, 56, 44-2, 41, 44-1, 31 and 16-2. By stepwiseremoval of the worst character they would be 4-2, 10, 50, 36-1, 44-1, 41, 39 and 59. All further'cleaning-up' of data was done by stepwise removal of characters. Table 7 Results of a LeQuesne test on the primary data matrix after the progressive removal of thehighest scoring characters (4-2, 10, 50, 36-1, 44-2, 56, 44-1, 41). Conventions follow table 2; non-informative characters omitted. Character number: incompatibilities observed expected O/E ratio Grand total- 2179 observed, 2613.3 expected. Overall O/E ratio = 0.83Ranking ratios of scoring characters 64 8.1 21.2 27.3 16.1 6.2 2 25.1 55 3 60 40 52.1 52.2 54 42.2 5.2 47.1 23 46 7.235 22.2 13 21.1 14.2 15.2 22.1 45.2 51.2 7.1 5.1 36.3 9 58.1 48.2 20.2 51.1 4.1 266.1 33 14.1 61.1 15.1 38 12 45.1 42.1 48.1 63 30 61.2 24.1 57 1.2 62 32 17 18.2 29.1 24.1 18.1 1.1 37 25.2 28.2 20.1 19 27.1 43.1 47.2 43.2 27.2 29.2 36.2 59 31 58.2 39 16.2 It is apparent from the generally high O/E values that homoplasy is a common feature in theOphioninae. Any attempt to produce a character set without homoplasy (a clique) necessitatesexcluding the great majority of characters, including some of the most informative ones. This isunacceptable for two reasons. (1) A large amount of information would have to be discarded leaving many unresolvablesituations. As 2N t 2 apomorphies are necessary to fully resolve N t taxa, a minimum of 100compatible characters would be needed to fully resolve the ophionine data. Consequently,without homoplasy, even the full set of 95 characters is insufficient for full resolution. Excludingautapomorphies for definition of terminal taxa, a minimum of 49 characters is necessary for PHYLOGENY OF THE OPHIONINAE 83 resolution, but in the present case it can be seen that even discarding half the characters, theresultant data set would still be extremely homoplastic. (2) It is important to realize that just because a character is homoplastic and has been derivedindependently in two (or more) lineages, this does not mean that it is not of phylogenetic value.Such a character can still be a genuine synapomorphy for species in two different groups of taxa,and many minimally homoplastic characters are still highly regarded by traditional taxonomists.For example, Mason (1981) states that one of the major synapomorphies of the Braconidae (theprobable sister-group of the Ichneumonidae) is the fusion of the second and third gastraltergites. Approximately six of the 60,000 species of Ichneumonidae also have tergites 2 and 3fused. This parallelism in a few specialized ichneumonids does not eliminate the character as auseful apomorphy for defining Braconidae in a phylogenetic sense, it merely reduces the value ofthe character as an infallible means of identifying braconids. If it were possible to consider allspecies of the two taxa Ichneumonidae and Braconidae, it is likely that the O/E value of thischaracter would be very low, but if the analysis were of all Parasitica, and a few taxa from eachfamily were selected as examples of the range of morphological variation, it is quite possible thatone of the six anomalous ichneumonids would be included, giving the character a much largerO/E value, and consequently risking its exclusion. In the present study certain species ofEnicospilus, Leptophion and Laticoleus were deliberately selected to show the range ofmorphological variation in the genus, so it is highly likely that these will contribute to the highlevel of homoplasy in the way outlined above. In the LeQuesne analysis eight characters were eliminated stepwise until none of the remainderscored worse than 1-00. The particularly high O/E value of these characters strongly suggeststhey are particularly homoplastic, and consequently of no real value in phylogenetic reconstruc-tion. Considered from a biological point of view it certainly seems these features are homo-plastic, but even so several seem to be of use in defining possible holophyletic groups.Considered in order of 'worst' first these characters are as follows. 4-2*: reduction of upper tooth of mandible. The occurrence of the derived state of this feature islimited in the matrix to three species, Rhynchophion flamipennis (908), Sicophion pleuralis(922) and Enicospilus unidens (946). Amongst the Ophioninae in general, this feature is found inthe apomorphic state in a group of Madagascan and southern Indian species of Enicospilus(Gauld & Mitchell, 1978; 1981) that seem to be related to E. unidens, an apparently unrelatedMelanesian species (E. interruptus Szepligeti) and in an Indian Ophion species. Thus althoughthe character has phylogenetic value in some cases (the definition of the unidens group) in othersit has none. Presumably the apomorphic condition found in three taxa in the matrix is the resultof parallel derivation, a conclusion that is strongly suggested by the very high label values(Appendix 4). Amongst out-groups this character is of similar sporadic occurrence (cf. Townes,1970), sometimes defining an apparent clade and other times being characteristic of a singlespecies. Structurally it is a simple adaptation for biting through fibrous cocoons and as such it isnot surprising that it has apparently arisen independently in many different evolutionarylineages. 10: elongation of maxillae. This too is a low information value character as the apomorphiccondition occurs in just two taxa in the matrix, Rhynchophion flamipennis (908) and Sicophionpleuralis (922). Amongst other ophionines it is limited in occurrence to some Neotropicalspecies of Eremotylus. All of these taxa are restricted to the drier areas of Central and SouthAmerica and the structural development seems to be an adaptation to drinking from certainflowers. Similar structural modification can be observed in many other ichneumonids in similarhabitats, including many Tersilochinae, Cremastinae and Anomaloninae. In arid areas ofsouth-western Australia species of three closely related cremastine genera, Trathala, Temeluchaand Pristomerus, all have identically modified maxillae. Only familiarity with the worldcremastine fauna enables one to recognize this as evolutionary convergence rather thanconsidering it indicative of a holophyletic group. These observations strongly suggest thatparticularly intense selection pressures in a particular area may elicit the development in 84 I. D. GAULD parallel, of apparently identical apomorphic features in closely related lineages. This mayexplain the similarity between the taxa in the matrix. 50: oblique position of cu-a in fore wing. Like the preceding two characters this is of lowinformation content as the derived condition occurs in just two taxa in the matrix, Lepiscelusdistans (909) and Barytatocephalus mocsaryi (927). Amongst other ophionines it is limited tooccurring in one or two species of Ophion. In the sister-group to the Ophioninae, theCampopleginae, the derived condition apparently is useful for defining some genera (e.g.Cymodusa, see Townes, 1970) but the character has little apparent phylogenetic significance inthe Ophioninae. 36-1': vein Im-cu separated from Cu la by less than 0-75, but more than 0-5 of the length ofCu lb .The majority of taxa have this character present in the presumed derived state. Only Lepiscelusdistans (909), Stauropoctonus occipitalis (917), Riekophion emandibulator (920), Ophionopsisnigrocyaneus (921), Sicophion pleuralis (922), Rictophion nebulifer (937) and Enicospilusnephele (944) are plesiomorphic for this feature. It is not found in the plesiomorphic state in anyother Stauropoctonus species, in most other Enicospilus species nor any other Riekophionspecies. If the state were a genuine symplesiomorphic feature of these taxa the 1 state musthave been derived in parallel in at least species of Riekophion, Enicospilus and Stauropoctonus(assuming they are natural genera) and therefore the 1 state is not synapomorphic for all speciesso scored. Alternatively, if the 1 state is a genuine synapomorphy for all 'other taxa', the statemust represent independent reversals in some taxa. Either way the character is homoplastic.The same argument can be applied even if the polarity is reversed, unless one were to accept theotherwise unsupported clade of 909+917+920+921+922+937+944. 44-2: strongly curved Rs in hind wing. In the matrix the apomorphic state of this character isfound in Simophion calvus (905), Lepiscelus distans (909), Laticoleus curvatus (915), Abancho-gastra hawaiiensis (925), Leptophion maculipennis (928), Euryophion adustus (936), Rictophionnebulifer (937) and Ophion luteus (939). Elsewhere in the subfamily all species of Ophion (ca 150species) have Rs strongly curved. The genus Abanchogastra is monotypic, but only some speciesof Leptophion, Enicospilus, Euryophion and Laticoleus possess the apomorphic state as dosome species of Eremotylus and Ophiogastrella. In both Leptophion and Laticoleus this featureseems to define holophyletic species-groups, but there is no evidence to support the idea that itcould be a synapomorphy of all the taxa listed above. Clearly it must have been derivedindependently in many ophionine lineages, but although a homoplastic feature, it is interestingto note that it still is quite characteristic of certain species-groups and even genera, suggestingthat once the apomorphic state has been derived, it is unlikely to undergo reversal. 56: separation of hamuli into two groups. This feature occurs in the matrix only in twoAfrotropical species, Lepiscelus distans (909) and Laticoleus curvatus (915). The latter is theonly species in its genus with this particular modification and Lepiscelus is monobasic (Gauld &Mitchell, 1978). The apomorphic condition is also found in some central Asian Ophion species.The only feature these taxa seem to have in common is that they inhabit rather dry areas. Thecharacter appears to have little phylogenetic value. 41: fore wing with R^ not reaching to tip. The apomorphic condition is found in Thyreodonatricolor (901), Simophion calvus (905), Orientospilus melasma (906), Ophionopsis nigro-cyaneus (921) and Sicophion pleuralis (922). It is not present in many other Thyreodon species soit is unlikely that the derived condition in 901 is synapomorphic with that of other species. Theexclusion of 901 from the analysis actually resulted in a worsening of the O/E value of thischaracter from 1-02 to 1-12, suggesting it is homoplastic in other taxa. The apomorphic state ofthis character is found in all species of the genera Simophion, Orientospilus and Ophionopsis,suggesting this feature, once derived, is not likely to undergo reversal. 44-1: bowed vein Rs in hind wing. The apomorphic state of this character is found in all taxalisted under 44-2 above and additionally in 14 other taxa, all from different genera. Elsewherethe feature is of sporadic occurrence in Ophiogastrella, Aulophion, Leptophion, Laticoleus and PHYLOGENY OF THE OPHIONINAE 85 Enicospilus, as well as being present in all species of Dictyonotus, Hellwigiella, Sclerophion,Eremotylus, Xylophion and Rhopalophion. This suggests parallel development in many dif-ferent evolutionary lines. The high O/E values of several of these characters were surprising as almost all representstructural specializations (not 'loss' apomorphies or dubious polarity assignments). Several (4-2,41 and 44-2) do apparently have considerable phylogenetic value at certain hierarchical levelsand really only two (36-1 and 44-1) are likely to be considered by a traditional taxonomist ascategory 3 'bad' characters. Character 10 almost certainly would intuitively be regarded as a'good' character (vide Cushman, 1947). It is interesting to note that two taxa (909 and 922) werecited for at least half of the eight worst characters, as having contributed to the high level ofhomoplasy . These taxa, also recognizable by their high label scores, were subsequently shown tobe amongst the more difficult to place, and small differences in their position in the order of taxasubmitted to the WAGNER program, produced large changes in the resulting cladogram. Another estimate of incidences of homoplasy in the primary data may be obtained fromperusal of the labels matrix (Appendix 4). Notably high scores, indicative of homoplasticderivation of the apomorphic state in two or more taxa, can be observed for character 1-2 (taxa909, 923 and 949), 4-1 (906, 909), 4-2 (908, 922, 946), 10 (908, 922), 26 (906, 923), 29-1 (922,937), 50 (909, 927), 56 (909, 915), 58-2 (942, 951) and 60 (928, 942). The shared plesiomorphiccondition for character 38 in taxa 941 and 943 was also highly labelled, suggesting one or otherhas undergone reversal. Of these 11 characters, six (4-2, 10, 50, 56, 58-2 and 60) are highlylabelled for all apomorphic states (two or three in each case), suggesting these are not usefulcharacters for phylogenetic reconstruction. Rather suprisingly all are 'gain' apomorphies. Four(4-2, 10, 50 and 56) are amongst the eight characters eliminated in cleaning up the LeQuesnematrix; one other (58-2) has a higher than average O/E score, but surprisingly one (60) has areasonably low O/E value. Analysis using FOURS program Using the FOURS program a shared derived character matrix (Appendix 3) was computed. Thisshows the number of derived characters common to each pair of taxa and has values in the rangeof 3-37. Multiple linkage clusters drawn for taxa with 29+ shared derived characters are shownin Fig. 8. Four groups are apparent. The largest includes 17 species (912-5, 919, 924, 928-30,932-4, 942-3, 945-7), the last eight of which form a particularly closely knit group. Taxon 925and the reciprocal nearest neighbours 911 and 910 are less strongly associated with this group.This complex contains most of the species of Enicospilus and Leptophion, all Dicamptus,Laticoleus and Pamophion, together with Ophiogastrella and representatives of two of the threeendemic Hawaiian genera, Pycnophion and Abanchogastra. It is noteworthy that Enicospilusnephele (944) and Leptophion tetus (931) are not associated. Each of these taxa share lessderived characters with their supposed congeners than their congeners share with other taxa,although in each case the species' nearest neighbours are its congeners (Table 13). Taxa 916, 917 and 918 (the Stauropoctonus group) form a cluster with 30-34 derivedcharacters in common. This group seems to be associated with the Enicospilus group quiteclosely as 918 shares 28 derived characters with 947, 943, 942 and 919. Taxa 935, 936 and 937 (the Euryophion group) cluster with 30-32 shared derived characters. Taxa 901-4 (the Thyreodon group) form a cluster sharing 28-32 derived characters. Alsoassociated with this cluster are the reciprocal nearest neighbours 907 and 921 (Dictyonotus IOphionopsis). When linkages of more than 22 shared derived characters are considered (Fig. 9), the majorityserve to consolidate the clusters already formed. The Stauropoctonus group and Ophiogastrellaare repeatedly linked to the Enicospilus cluster (both having in excess of 60% of possiblelinkages to the latter) as does taxon 931. Taxa 926 and 944 are less strongly but uniquelyassociated with this group and taxon 949 also links to this group. The Thyreodon, Euryophionand Dictyonotus/ Ophionopsis groups, together with taxon 908, form a second robust clusterwith only Euryophion showing any linkage to non-group taxa. Taxon 909 is somewhat 86 I. D. GAULD o I <0TD O O0\(S u<u 60 J I BC E PHYLOGENY OF THE OPHIONINAE 87 Fig. 9 Multiple linkage clusters drawn for all taxa with 22 or more derived characters in common. Asvirtually complete linkages occur between taxa enclosed within the boxes, these linkages are not shownso as to avoid confusion. intermediate linking to both the Enicospilus and Euryophion groups but shares most derivedfeatures with taxa in the Stauropoctonus group. Other taxa occupy a more ambivalent position. Taxon 906 is intermediate between Enico-spilus and Euryophion, whilst 905 appears to be fairly closely associated with Euryophionalthough its nearest neighbour is taxon 906. Taxon 927 is also intermediate. The remainingspecies, 920, 922, 923, 938-41, 948 and 950-51, are not closely associated with other taxa. Tosummarize therefore, two major groups are discernible, the Enicospilus/ Stauropoctonus com-plex containing taxa 909-919, 924-926, 928-934, 942-947 and 949, and the ThyreodonlEuryophion complex with taxa 901-904, 907-908, 921 and 935-937. Taxa 920, 922, 923,938^941, 948 and 950-951 are not associated at all whilst 905, 906 and 927 are intermediatesbetween the major clusters. It is interesting to note that, excluding the last three taxa, theremainder could be classified phenetically (if one counted shared plesiomorphies of which 920,922, 938-941, 948 and 950-951 have a considerable number in common) into three groupsapproximately corresponding to the Ophion (plesiomorphic group) , Thyreodon and Enicospilusgroups of Cushman (1947). Phylogenetic analyses of groups of ophionine taxa The Euryophion group (taxa 935, 936, 937) It can be seen from Appendix 3 that these taxa form a group with 30+ shared derived characters,and no more than 26 derived characters in common with any other taxon. Using these three asthe fixed taxa (3F option) of FOURS, and trying them against all other taxa, nothing broke themup. The three taxa were found to share 29 apomorphies (2, 6-1, 7-1, 16-1, 16-2, 22-2, 24-1, 24-2,25-1, 25-2, 27-1, 28-1, 28-2, 31, 32, 38, 39, 42-1, 42-2, 43-1, 44-1, 454, 47-1, 49, 51-1, 51-2, 58-1,61-1 and 62). Five of these apomorphies (2, 7-1, 22-2, 42-2 and 58-1) have average or lower O/Evalues (0-60, 0-84, 0-80, 0-79 and 0-87 respectively) and are shared by relatively few other taxa inthe matrix. Of these five apomorphies only one (7-1) has questionable polarity. No character is a 88 I. D. GAULD unique apomorphy of the group, but the four apomorphies accepted above suggest this group ismonophyletic. There are three possible arrangements of these taxa (Fig. 10). Taxa 935 and 936 have beenconsidered to be congeneric (in Euryophion) whilst 937 is usually placed in a separate genus,Rictophion (Townes, 1971; Gauld & Mitchell, 1978). Rictophion has been separated fromEuryophion by two unique apomorphies, loss of thyridia and possession of 3-segmented palps.Cladogram A supporting this arrangement is the weakest of the three as it involves only a singleinformative character, 36-1 which has both a dubious polarity assignment and a very high O/Escore. Both B and C also involve some poor scoring characters (19 and 44-2 both have O/Evalues of more than 0-95). Arrangement B is only supported by 'loss' apomorphies, though one(8-1) has a low O/E value (0-39). However, it has a rather high label score. The characterssupporting C are all fairly high scoring (O/E values 0-76+) but do not involve either apomorphicloss or dubious polarity assignment. On balance therefore, C seems to be the preferablearrangement, suggesting that Euryophion, as currently recognized, is paraphyletic with respectto Rictophion. This conclusion is not surprising as originally Rictophion was separated fromEuryophion on the basis of two autapomorphies. Rictophion ikuthana (937), the only species inthe genus, can be regarded as a specialized species of Euryophion. These species, and all theremainder in the genus, are further analysed below (p. 136). 935 936 937 935 937 936 936 937 935 8.1* 44. 2 54 A B Fig. 10 The three possible dichotomous arrangements of Euryophion latipennis (935), E. adustus (936)and Rictophion ikuthana (937). The Thyreodon group (taxa 901, 902, 903, 904) These four taxa, all species currently placed in the genus Thyreodon (Townes & Townes, 1966),form a closely knit group. Using the FOURS 3F option any combination of these speciesremained as a group when tested against any fourth taxon. The Thyreodon species share 27apomorphies (2, 6-2, 7-1, 14-1, 17, 22-2, 23, 24-1, 24-2, 25-1, 25-2, 27-1, 28-1, 28-2, 29-2, 30, 33,36-1, 38, 424, 42-2, 47-1, 49, 51-1, 51-2, 54 and 61-1). Seven of these apomorphies (2, 6-2, 7-1,22-2, 23, 42-2 and 54) have both lower than average O/E scores (0-60, 0-47, 0-84, 0-80, 0-78, 0-79and 76 respectively) and are shared by relatively few other taxa. Only one, 7-1, involvesdubious polarity assignment; the remainder are gain apomorphies suggesting the group is PHYLOGENY OF THE OPHIONINAE 89 monophyletic. Considering the possible phylogenetic relationships of these taxa to each other,the following result was obtained using the FOURS program (Underwood, 1982). TAXA NOS.- 901XX'Y XYZ'XY'ZX'Y'Z' 902 The most favoured arrangement X,XYZ' is shown in Fig. 11. This involves three forwardparallelisms 21-2 (in 901 and 903), 37 (in 901/2 and 904) and 36-2 (in 903 and 904) and onereversal (9 in 901). If the dubious polarity of 36-2 were to be reversed this character would fit thecladogram. It is noteworthy that some authors (e.g. Ashmead, 1900; Cushman, 1947) haveplaced taxon 904 and its relatives in a separate genus, Athyreodon. 904 12*, 13 26, 58.1 Fig. 11 The arrangement of Thyreodon species as supported by the largest number of compatiblecharacters. The DictyonotuslOphionopsis group (taxa 907, 921) These two taxa are reciprocal nearest neighbours. They share 30 apomorphies (2, 6-2, 7-1, 12,13, 14-1, 16-2, 18-2, 21-1, 21-2, 22-2, 23, 24-1, 24-2, 25-1, 25-2, 28-1, 28-2, 30, 33, 38, 424, 42-2,44-1, 47-1, 49, 51-1, 51-2, 54 and 58-1). Ten of these (2, 6-2, 7-1, 21-1, 21-2, 22-2, 23, 42-2, 54 and58-1) have both average or lower O/E values (0-60, 047, 0-84, 0-79, 0-39, 0-80, 0-78, 0-79, 0-76and 0-87 respectively) and are shared by relatively few other taxa in the matrix. One character, 90 I. D. GAULD 21 '2, is a unique apomorphy of this group. The primary label matrix (Appendix 4) suggestscharacters 16-1, 36-1 and 61-1 (0 scores in taxon 921) may have undergone reversal, whilstcharacter 41-2(1 score in 921) may have been derived in parallel in 921 and other taxa. The Thyreodon + Euryophion + Dictyonotus complex (the above three groups)These three groups of taxa appear to be associated (see shared derived character matrix,Appendix 3). Phylogenetically they also seem to form a distinctive clade, sharing 16 apomorphicfeatures (2, 7-1, 22-2, 24-1, 24-2, 25-1, 25-2, 28-1, 28-2, 38, 42-1, 42-2, 47-1, 49, 51-1 and 51-2).Four of these apomorphies (2, 7-1, 22-2 and 42-2) have both lower than average O/E values(0-60, 0-84, 0-80 and 0-79 respectively) and are shared by very few other taxa. Character 2 iselsewhere only found in the apomorphic state in taxon 908, character 7-1 is apomorphic for taxa908, 940 and 946, character 22-2 is apomorphic for taxa 906, 908, 927 and 950, whilst character42-2 is otherwise only apomorphic for taxa 908, 909, 916, 917 and 918. Examination of the labelsmatrix (Appendix 4) shows that for character 7-1, species 940 and 946 are labelled more thantwice as frequently as any other taxon. Similarly for character 22-2, species 906 and 950 are fairlyhighly labelled. These data suggest parallel derivation of the apomorphic condition of thesecharacters in taxa 906, 940, 946 and 950. Three other characters are found in the apomorphic condition in all except one taxon in thegroup . Of these ,61-1 has a high label score for taxon 921 , suggesting the apparent plesiomorphiccondition in this taxon may be a reversal. Character 54 is present in the apomorphic condition inall taxa in this group except 935; elsewhere it is only found, in the data matrix, in the derivedstate in three taxa, 908, 941 and 950. The higher values of these last two species suggest theapomorphic state may have been derived in parallel in them. The third of the three characters,58-1, is usually found in the derived state in Thyreodon, so may be presumed to have undergonereversal in 904. Elsewhere it is only found in the apomorphic state in taxa 908, 909, 927, 941 and950. Taxon 909 has a high label score for this character, suggesting homoplasy. It seemsreasonable to treat tentatively all these three characters as apomorphies of the group. Thisinterpretation is certainly the most parsimonious, for in each case the alternative to singlepostulated reversals must be the advocacy of multiple forward parallelisms. One other taxon, 908, seems to be consistently associated with this group. Its nearestneighbours all lie within the group (Table 8), and in cluster analysis it readily associated with thegroup. It shares 18 of the 19 apomorphies of the group and is discordant only in having character Table 8 The Thyreodonl Euryophion complex showing nearest neighbours with numbers of sharedderived characters. The rows give the five (or more if several are equal fifth) nearest neighbours of eachtaxon listed in the left-hand column. The columns show the number of times a species is cited as a nearneighbour. It can be seen that taxon 908 occupies an intermediate position between Thyreodon (901-4)and Euryophion (935-7), having nearest neighbours in both. The Thyreodon and Euryophion groupshave only one near neighbour in common (904-936). Only species in the latter group share largenumbers of derived characters in common with non-group taxa (the 'others' column). It is interesting tonote that taxon 907, the least autapomorphic in this group (see p. 89), is cited as a near neighbour by allother taxa. 901902903904907921908935936937 PHYLOGENY OF THE OPHIONINAE 91 28-2 plesiomorphic. However, this character is highly labelled for this taxon suggesting areversal. In FOURS analyses taxon 908 repeatedly broke into the group. For example - TAXANOS.-901 921 936 908 33 33 In both these cases exclusion of taxon 908 from the group, the XYZ' option, is not the mostparsimonious solution. Accordingly, taxon 908, Rhynchophion flammipennis , is included foranalysis in this group. Considering only the ten taxa of this group, 35 characters are informative for examiningintra-group relationships. A LeQuesne test on this subset of data yielded the results shown inTable 9. Progressive removal of the five worst characters (59, 36-1, 27-2, 9 and 39) markedlyaltered the ranking of the 'best' scoring characters (i.e. those with an O/E value of 0-6+) ; theoverall O/E ratio improved from 0-69 to 0-52 (Table 10). The data matrix was re-ordered withcharacters arranged in increasing value of O/E ratio (Tables 11, 12) before and after removal ofthe worst characters. Considering only the characters with an O/E value greater than 0-55 showsthat there are two competing patterns. Characters 16-2, 44-1, 31 and 32 unite the Euryophiongroup (935-7) with the Dictyonotus group (907, 921) and 908. This arrangement is contradictedby characters 14-1, 23, 33 and 6-2, which all favour uniting the Dictyonotus group with Table 9 Results of a LeQuesne test on the Thyreodonl Euryophion data set. Conventions as in Table 2. Character number: incompatibilities observed expected O/E ratio Grand total- 280 observed, 406.3 expected. Overall O/E ratio = 0.69 Ranking ratio of scoring characters 16.2 26 8.1 17 44.1 62 33 45.1 14.1 23 6.1 6.2 31 32 27.1 21.2 44.2 39 12 20.121.1 16.1 43.1 19 29.2 30 41 18.2 37 13 36.2 9 27.2 36.1 59 92 I. D. GAULD Table 10 Results of a LeQuesne test on the ThyreodonlEuryophion data set after progressive removal offive highest scoring characters (59, 36-1, 27-2, 9, 39). Conventions as in Table 2. Character number: incompatibilities observed expected 0/E ratio 8.1: 3 13.9 0.22 \2_ : 14 24.1 0.58 16.1: 13 22.3 0.58 16.2: 7 22.3 0.31 ^9 : 13 20.3 0.64 20.1: 10 13.9 0.72 ^3_ : 6 20.3 0.30 ^6_ : 6 20.3 0.30 30 : 16 20.3 0.79 31_ : 11 24.1 0.46 36.2: 20 20.3 0.98 39 : 12 23.2 0.52 44.1: 8 22.7 0.35 44.2: 6 13.4 0.45 6.1: 6 19.6 0.31 ^3 : 23 24.1 0.96 l]^ : 8 23.2 0.34 21.1: 13 22.7 0.57 27.1: 7 13.9 0.50 32 : 11 24.1 0.46 41^ : 11 13.9 0.79 45.1: 6 20.3 0.30 6.2: 6 19.6 0.31 14.1: 6 20.3 0.30 18.2: 22 23.2 0.95 21.2: 7 13.4 0.52 29.2: 19 23.2 0.82 23 : 6 20.3 0.30 43.1: 14 23.2 0.60 62 : 6 20.3 0.30Grand total- 158 observed, 303.4 expected. Overall 0/E ratio = 0.52Ranking ratio of scoring characters 8.1 45.1 62 23 26 33 14.1 6.1 6.2 16.2 17 44.1 44.2 31 32 27.1 39 21.2 21.1 1216.1 43.1 19 20.1 30 41 29.2 18.2 13 36.2 Thyreodon (901^). Characters 6-1, 8-1, 17, 26, 45-1 and 62 clash with neither of thesecombinations and so, in this context, are uninformative. The very best characters in the original35 character set (Table 11) favour the former grouping whilst in the 'cleaned-up' 30 character set(Table 12) the best characters favour Thyreodon + Dictyonotus. Of the eight characters in thetwo contradictory groups, two, 16-2 and 33, are 'loss' apomorphies and perhaps therefore theapomorphic state is particularly likely to be subject to parallel development. Even disregardingthese one is still left with a tie with three characters supporting each arrangement, so it isnecessary to consider further characters. Of the characters having an initial O/E value of averageor better (16-2-16-1 in Table ll)two, 12 and 21-1, suggest uniting Thyreodon with Dictyonotus ifone postulates minimal reversal in Thyreodon. Character 27-1 suggests uniting Thyreodon withEuryophion and taxon 908, whilst 21-2 is an autapomorphy of taxa 907 and 921. Character 39suggests linking a single Thyreodon species with Euryophion. Character 20-1 is apomorphic inonly two taxa, 901 and 907. Considering all characters the most parsimonious arrangement appears to be to treatEuryophion as the first branch. By postulating minimal homoplasy (i.e. single reversals orparallelisms) 12 characters can be made to support this arrangement. Only three, 30, 43-1 and44-2, have an O/E value below average, and only one, 33, is a 'loss' apomorphy. Thisarrangement involves the following homoplasy (and these are considered to be predictable fromthe Thyreodon labels matrix (Appendix 5) if, for a particular character, the score obtained bythe presumed discordant species is higher than that obtained by other species in the group) - 12,reversal in 903/4 (not predictable); 30, reversal in 908, parallelism in 935 (predictable); 39, Table 11 The Thyreodon/ Euryophion complex with all 35 informative characters ranked according totheir initial O/E value (lowest first). All characters to the left of 27-1 have an O/E value of 0-55 or less. 16.2 8.1 44.1 33 14.1 6.1 31 27.1 44.2 12 21.1 43.1 29.2 41 37 36.2 27.2 59 26 17 62 45.1 23 6.2 32 21.2 39 20.1 16.1 19 30 18.2 13 9 36.1 901902903904908907921937935936 010101010101000 1 , 11 10111 00011000111 1 1 1 1 1 1101 1001 1111 0101 1 1000 11101 11111 1 1 11100 1001 000 1 1 1 1 1 1 PHYLOGENY OF THE OPHIONINAE 93 Table 12 The Thyreodonl Euryophion complex with the 30 'best' informative characters ranked by O/Evalue computed after progressive removal of five highest scoring characters (see Table 10). Allcharacters to the left of 21- 1 have an O/E value of 0-55 or less. 901902903904908907921937935936 2623 1 111 27.1 1 C 1 1 1 1 1 13.1 C C C C 1 1 C C 1 1 20.1 1 1 1111 1111 00111 2 1318.2 1 1 11 1 parallelism in 904 (predictable); 43-1, parallelism in 907 (predictable); 44-2, parallelism in 921and reversal in 935 (not predictable). In attempting to resolve the sister-lineage of Euryophion, the Thyreodonl DictyonotuslRhynchophion branch, it is apparent that no character is a unique apomorphy of eitherRhynchophion + Dictyonotus, Rhynchophion + Thyreodon or Dictyonotus + Thyreodon.Character 30 suggesting the latter is paralleled in taxon 935, 16-2 and 44-1 suggesting Rhyn-chophion + Dictyonotus are paralleled in Euryophion as is character 27-1 which favours unitingRhynchophion + Thyreodon. In any compatibility cladogram (e.g. Fig. 12) these three taxamust remain as an unresolved trichotomy. 935/7 17 2, 7.1, 22.2, 24.1*, 24.2*. 25.1*, 25.2*. 28.1*.38*. 42.1, 42.2, 47.1, 49, 51.1*. 51.2* Fig. 12 Cladogram showing inter-relationships of taxa in the Thyreodon/ Euryophion complex assupported by the largest compatible clique. 94 I. D. GAULD To resolve these data further it was necessary to use the WAGNER option of PHYLIP. Theshortest rooted tree obtained involved 83 transformation steps for the 45 characters showingboth and 1 states for this group of taxa (that is the 35 characters shown in the LeQuesne matrix(Table 11) plus 1-1,4-2, 8-2, 10, 18-1, 29-1, 36-3, 43-2, 47-2 and 57 which are autapomorphies ofvarious individual taxa). To obtain this tree (Fig. 13) 21 characters (1-1, 4-2, 6-1, 6-2, 8-2, 10,14-1, 17, 184, 21-2, 23, 26, 29-1, 33, 36-3, 43-2, 44-2, 45-1, 47-2, 57 and 62) are presumed to havebeen uniquely derived, 11 characters (8-1, 12, 16-2, 18-2, 20-1, 27-1, 39, 41, 43-1, 44-1 and 59)have undergone two transformations (i.e. have either been derived in parallel or have 903 904 936 937 907 921 36.2 16.1, 30, 36.1 Fig. 13 Most parsimonious cladogram for taxa in the Thyreodonl Euryophion complex. Black squaresrepresent uniquely derived characters; circles, apomorphic features derived independently in twolineages; diamonds, in three lineages; squares, in four lineages, or involving four transformations. PHYLOGENY OF THE OPHIONINAE 95 undergone reversal), 12 characters (13, 16-1, 19, 21-1, 27-2, 29-2, 30, 31, 32, 36-1, 36-2 and 37)have undergone three transformations, whilst one (9) underwent four. It is noteworthy that thecladogram shows only one of a number of competing cladograms with the same arrangement oftaxa and the same number of transformation steps, but different positioning of the homoplasticcharacters. For example, character 30 is shown as a group apomorphy and is postulated asundergoing reversal in the stem 936/7, and in taxon 908, but it would be equally parsimonious tohave suggested that 30 was derived in parallel in stems 901/4, 907/21 and taxon 935. Biologicallyneither is implausible. The Stauropoctonus group (Taxa 916, 917, 918) These three taxa share 29 apomorphies (1-1, 5-1, 5-2, 14-2, 16-1, 16-2, 18-2, 21-1, 22-1, 25-1,28-1, 28-2, 29-2, 33, 37, 38, 39, 42-1, 42-2, 45-1, 47-1, 49, 51-1, 51-2, 52-1, 52-2, 61-1, 61-2 and62). Six of these apomorphies (5-1, 5-2, 14-2, 21-1, 42-2 and 52-2) have both lower than averageO/E values (0-84, 0-75, 0-83, 0-79, 0-79 and 0-71 respectively) and are shared by relatively fewother taxa in the matrix. Within this group all three possible arrangements of taxa are supported: 916 + 917 by 25-2*and 47-2; 916 + 918 by 1-2*, 36-1', 36-2', 43-1' and 63; 917 + 918 by 57*. All of these eightinformative characters have high LeQuesne test failure rates (0-9+) and most are either losscharacters or have dubiously assigned polarity, making the choice between arrangementsdifficult. The most parsimonious would obviously be 916 + 918. Currently 916 and 917 areplaced together in Stauropoctonus whilst 918 is placed in a separate taxon, Aulophion (Cush-man, 1947; Townes, 1971). Aulophion has traditionally been separated from Stauropoctonus bythe possession of two autapomorphies, loss of the posterior transverse carina of the meso-sternum and absence of the epicnemial carina. This suggests Stauropoctonus may be aparaphyletic assemblage. For the present these taxa are treated as an unresolved trichotomy,but their inter-relationship is discussed further below (p. 146). It is sufficient now to state thatthese three taxa seem to form a closely knit and apparently holophyletic group. The Enicospilus + Stauropoctonus complex (Taxa 909-919, 924-926, 928-934, 942-947)This, the largest apparent group of ophionine taxa in the matrix, contains 28 species. (N.B.Taxon 949 which associates with this group in the cluster analysis is exceptional in having theplesiomorphic state of a number of characters usually found in the apomorphic condition in thisgroup (e.g. 25-2, 28-2, 42-1, 51-1, 51-2 and 63). For the present this taxon is excluded from thecomplex and its position discussed later in the work.) Their nearest neighbours are shown inTable 13. These taxa share four apomorphic features (16-1, 25-1, 28-1 and 51-1). A further 13characters (22-1, 25-2, 28-2, 33, 36-1, 38, 42-1, 43-1, 45-1, 47-1, 49, 51-2 and 63) are present in theapomorphic state in all except one, two or three taxa. Most of these characters are labelled,often highly, for the species in this group that show the plesiomorphic condition (e.g. 38 fortaxon 943 and 47-1 for taxa 944 and 945), suggesting a reversal may have occurred. Theexceptions are 22-1 and 33 which are plesiomorphic for taxa 910 and 911. Both are unlabelled inthe primary label matrix (Appendix 4). Of the 17 characters listed above as possible apomor-phies of this complex, all except 22-1, 33 and 63 occur in the apomorphic condition in a largenumber of other ophionine taxa in the matrix. Individually therefore, they are not goodcharacters for defining the group. Character 49 is plesiomorphic only in the solitary taxon 926,suggesting 49 may be an apomorphy of all ophionines, and 926, an otherwise specialized species,may have undergone a reversal. Character 16-1 is apomorphic for all taxa except 901-3 and 921;character 25-1 for all except 948 and 951 ; character 25-2 for all except 918, 920, 922, 923, 938-41,945, 948, 949 and 951; character 28-1 for all except 939; character 28-2 for all except 908, 920,923, 928, 938-41, 949 and 951; character 36-1 for all except 909, 917, 920, 921, 922, 937 and 944;character 38 for all except 938, 939, 941, 943, 948 and 951; character 42-1 for all except 905, 922,925, 938-41 and 948-50; character 43-1 for all except 901^, 908, 909, 917, 921-22, 925 and 941;character 45-1 for all except 901-4, 907-8, 921, 923, 925-26, 931, 941 and 948; character 474 forall except 938, 939, 943, 944, 948 and 951; character 51-1 for all except 920, 922, 939-41 and949-51; and character 51-2 for all except 920, 922, 932, 938-41 and 948-51. Ten of thesecharacters (25-1, 25-2, 28-1, 28-2, 38, 42-1, 474, 49, 514 and 51-2) have been stated above (p. 96 I. D. GAULD Table 13 The EnicospiluslStauropoctonus complex showing nearest neighbours with numbers of sharedderived characters. Conventions as in Table 8. Particularly striking are taxa 944 (an Enicospilus species)and 931 (a Leptophion species). Both cite their congeners as nearest neighbours, but neither is cited asnear neighbours by its congeners. 926 925 924 947 946 945 944 943 942 934 933 932 919 915 914 913 912 931 930 929 928 918 917 916 909 910 911 Others 926 + -- 25 25 27 -- -- 24 27 - ,- 925 + 27 28 30 28 -- -- 24 24 924 25 27 + 27 29 29 947 + 35 31 -- 31 34 30 30 30 946 35 + -- -- 37 32 32 -- 32 945 31 31 + -- 28 31 28 28 28 944 24 26 24 + 26 24 943 31 30 28 + 31 28 942 34 37 31 31 + 31 934 30 32 31 + -- 30 -- 30 933 29 29 29 29 + -- 29 932 28 32 30 30 - + -- -- 29 - 919 29 29 29 29 29 + 32 33 30 29 30 915 32 + 36 33 34 31 914 32 33 36 + 33 33 913 29 29 33 33 + 31 912 - 30 34 33 31 + 29 931 - 27 27 27 + -- 27 27 930 28 29 28 29 -- 28 -- + 29 30 929 30 29 -- 29 29 + 29 928 30 31 29 30 29 + 918 28 - 28 28 - 28 - - + 30 34 917 24 24 24 -- 30 + 31 25 -- 916 27 27 34 31 + 27 -- -- 909 24 24 24 26 25 27 + -- -- 24-936 910 26 - -- 26 28 28 26 26 -- 26 + 30 911 -- 28 - 28 30 27 30 + 90) to be apomorphies of the ThyreodonlEuryophion complex, suggesting they are charactersthat unite the EnicospiluslStauropoctonus and ThyreodonlEuryophion generic complexes.Character 16- 1 could also be considered an apomorphy of both these major groups if one were topostulate reversal in 901-3 and 921. Character 22-1 is found in the apomorphic state outside the EnicospiluslStauropoctonuscomplex only in taxa 920, 940, 949 and 951; character 33 in taxa 901^, 907-8, 920-22 and 949and character 63 in 905, 906 and 923. Character 33 is a loss apomorphy and has been cited aboveas an apomorphy of the ThyreodonlDictyonotus lineage. It is perhaps not unreasonable topostulate parallel loss in the EnicospiluslStauropoctonus, the ThyreodonlDictyonotus and the920, 922 and 949 lineages. However, this feature, the loss of the umbo, is not apomorphic fortaxa 910 and 911 included above in the Enicospilus group. It is biologically rather unlikely thatthe umbo would be lost then redeveloped, so this feature favours placing Ophiogastrella(910-11) primitive with respect to the rest of the group. Such a position is also favoured bycharacter 22-1. The apomorphic condition, a lengthened anterior part of the propodeum, israther unlikely to have undergone reversal. It is interesting to note that character 22 has two alternative derived states, 22-1 or 22-2. Thederived state 22-2 is an apomorphic feature of the ThyreodonlEuryophion complex and is onlyfound elsewhere in taxa 906 and 927. The alternative derived state characteristic of theEnicospiluslStauropoctonus group (less Ophiogastrella) is also found in very few other taxa (seeabove). The plesiomorphic condition (0,0) is found in taxa 905, 910, 911, 922, 923, 938, 939, 941and 948. The character seems to be important in defining two major sister-lineages of thesubfamily. PHYLOGENY OF THE OPHIONINAE 97 Using the CLIQUE option of PHYLIP eight cliques were found with 12 or more informativecharacters. These were A (12)B(12)C(12)D(12)E(13)F(12)G(12)H(12)ADB 18-1, 18-2, 21-1, 22-1, 27-3, 33, 35, 36-3, 42-2,61-1, 61-2, 64]15-1, 15-2, 18-1, 18-2, 21-1, 22-1, 27-3, 33, 35, 36-3, 42-2, 64]17, 18-1, 18-2,21-1, 22-1, 27-3, 31, 33, 39, 42-2, 60, 64]15-1, 17, 18-1, 18-2, 21-1, 224, 27-3,31, 33, 42-2, 60, 64]12, 14-2, 17, 18-2, 21-1, 22-1, 27-3, 31, 33, 39, 42-2, 60, 64]12, 15-1, 17, 18-2, 21-1, 22-1, 27-3, 31, 33,42-2, 55, 64]13, 18-1, 18-2, 21-1, 22-1, 33, 35, 36-3, 42-2, 61-1, 61-2, 64]13, 15-1, 15-2, 18-1, 18-2, 21-1, 22-1, 33, 35, 36-3, 42-2, 64]H = (18-2, 21-1, 22-1, 33, 42-2, 64} Av. O/E = 0-76Av. O/E = 0-75Av. O/E =0-78Av. O/E = 0-77Av. O/E = 0-78Av. O/E = 0-76Av. O/E = 0-79Av. O/E = 0-78 The best clique in the sense of the largest is E but this has a lower average O/E value (0-78) thanseveral others. B has the lowest average O/E value with 0-75, closely followed by A and F with0-76. The remainder have higher values. The cladograms based on the largest clique (E) and thebest O/E scoring clique (B) are shown in Figs 14, 15. The cladogram produced from clique Bassociated the various taxa fairly well with their congeners. Species of Enicospilus (942-47),Dicamptus (932-34) and the Hawaiian genera (924-26) (regarded by Townes, 1971, as derivedEnicospilus) cluster together, except for E. cionobius (945) which is excluded by having theplesiomorphic condition of 15-2 (possibly a reversal). Laticoleus (912-15) and Leptophion(928-30) species form a separate cluster as do the Stauropoctonus group (916-18) plusLepiscelus (909). The cladogram derived from clique E has more confusing groupings, unitingsome species of Enicospilus and Laticoleus (character 17) or defining a clade containingDicamptus neavei, Leptophion tetus and Abanchogastra hawaiiensis (character 12). The heter-ogeneous collection of species united by character 14-2 includes representatives of Dicamptus,Enicospilus, Laticoleus and Leptophion; other species of these genera are excluded. The largestclique, E, is therefore not considered particularly useful in this case; clique B appears to be abetter indicator of relationship. It is noteworthy that clique B is more informative (s = 0-681)than clique E (s = 0-632). Fig. 14 Cladogram for taxa in the Enicospilus/ Stauropoctonus complex based on the largest clique (E). 98 I. D. GAULD Fig. 15 Cladogram for taxa in the EnicospiluslStauropoctonus complex based on the favoured clique (B) . It is notable that the intersect of these cliques includes several characters previouslyconsidered to be good indicators of phylogenetic relationship (Cushman, 1947; Townes, 1971).Three (18-2, 21-1 and 42-2) support the group 909 + 916-18, whilst two others (22-1 and 33)place Ophiogastrella (910-11), primitive to the other taxa. Character 64 is an autapomorphy ofOphiogastrella. A LeQuesne test was undertaken on the EnicospiluslStauropoctonus data set and the O/Evalues are given in Table 14. Stepwise elimination of all characters scoring worse than 1-00 (atotal of 15) resulted in considerable cleaning up. The overall ratio improved from 0-89 to 0-74.Amongst the scoring characters the most striking changes in rank occurred to 1-2 which rosefrom thirty-fifth to eleventh position, and 62 which dropped from thirty-first place to positionforty-three. Amongst the best scoring characters 42-2 and 18-2 rose from fourteenth andsixteenth positions to seventh and eighth positions (Table 15). Using the S option of FOURS thedata set was reorganized with characters ranked as per O/E value after removal of the 15 mostdiscordant characters (Table 16). It can be seen that the characters with the lowest O/E values,22-1 and 33, exclude Ophiogastrella (910-11) whilst 64 is an autapomorphy of this genus. Anumber of other characters may be postulated as apomorphies of Ophiogastrella, though allapparently have been derived in parallel elsewhere. Characters 18-2 and 42-2 support theStauropoctonus group -I- 909, whilst 21-1 supports just the former. Character 1-2 is incompatiblewith 21-1. Initially it had a poor O/E score (0-93 compared with 0-54 for 21-1) but progressiveelimination of the poorest characters produced rapid 'clean-up' until with 15 characterseliminated it scored 0-57 compared with 0-34 for 21-1. With only the 18 best characters left in thematrix both taxa scored equally 0-13. A considerable number of slightly homoplastic characters(e.g. 5-1, 5-2, 37, 52-1) support 21-1 but no other character supports 1-2 and it is for this reasonthat one would prefer the arrangement supported by the former character. Several other characters support the Stauropoctonus + 909 group but necessitate postulatingparallel derivation in other places . These include 1 1 (parallelism in 925 and 945) , 39 (parallelismin 930) ,61-1 (parallelism in 924-26 and 942-47) . Many of these parallelisms are also suggested inthe reduced labels matrix (Appendix 6), e.g. taxon 930 for character 39 acquires almost half of itstotal number of labels and taxa 925 and 945 are quite highly labelled for character 1-1. PHYLOGENY OF THE OPHIONINAE 99 Table 14 Results of a LeQuesne test on the EnicospiluslStauropoctonus data set. Conventions as in Table2. Character number: incompatibilities observed expected 0/E ratio 16.2 44.2 25.2 19 6.1 56 Table 15 Results of a LeQuesne test on the EnicospiluslStauropoctonus data set after progressiveremoval of 15 highest scoring characters (56, 19, 6-1, 25-2, 44-2, 16-2, 24-2, 44-1, 30, 57, 24-1, 43-2, 36-1,45-1, 36-2). Conventions as in Table 2. Character number: incompatibilities observed expected 0/E ratio 31 60 52.1 59 52.2 62 43.1 37 4.1 47.2 100 I. D. GAULD Table 16 The EnicospiluslStauropoctonus complex data set reordered according to O/E values of thecharacters given in Table 15. Characters to the left of 17 have an O/E value of average (0-74) or less. 22.1 64 27.3 42.2 36.3 1.2 61.1 47.1 7.2 15.1 55 29.2 1.1 48.2 20.2 27.2 5.1 20.1 5.2 60 59 62 37 47.2 Character 15-1 is compatible with those characters that suggest treating 910 + 911 and 909 +916-18 as the first two branches. This unites all the other taxa to form a residual group thatincludes Enicospilus (942-47), Dicamptus (932-34), Leptophion (928-31), Laticoleus (912-15)and Pamophion (919). This is rather a difficult group to resolve. Two of the best characters, 13and 27-3, are incompatible. Character 35 suggests uniting Laticoleus (912-15) with Leptophion(928-3 1 ) , and this is partially supported by character 3 . No character unites all the taxa that showthe plesiomorphic condition of 35. Three characters, 5-1, 61-1 and 61-2, suggest uniting theHawaiian genera (924-26) with Enicospilus (942-47). Only one taxon, 944, is at variance for asingle character, 61-2, and this has a high label value suggesting that a reversal to theplesiomorphic state may have occurred. In several places in the data set it is apparent that similarcases of a single taxon being at variance with its congeners is observable. In the case of 944, it isone of only about ten Enicospilus species out of nearly a thousand to have the plesiomorphiccondition of character 61-2. Similarly 945 has the plesiomorphic state of 15-2. Both of these taxado not appear to be primitive in any other features and as they are thus unlikely to be ancestral toall other taxa in the genus it is probable that the apparently unspecialized condition is a reversal. Character 15-2 favours uniting the Hawaiian genera, Enicospilus and Dicamptus (932-34) andthis is partially supported by character 40 although it is necessary to postulate reversal in theHawaiian genera. Character 46 is similar. Character 7-2 is contradictory, suggesting a groupcomprising 912-15, 919 and 928-34. No characters satisfactorily separate Laticoleus (912-15)from Leptophion (928-31) though 27-1, 27-2 and 55 suggest some separation. Character 3suggests that Pamophion belongs to the Laticoleus I Leptophion lineage. PHYLOGENY OF THE OPHIONINAE 101 Using the CLIQUE option of PHYLIP on the data set with the basal taxa (910-11, 909,916-18) removed revealed the existence of four cliques with eight or more informativecharacters. These are A (8)B(9)C(8)D(9) 4-1, 5-1, 18-1, 27-3, 35, 36-3, 61-1, 61-2]4-1, 5-1, 27-1, 27-2, 27-3, 35, 36-3, 61-1, 61-2]4-1, 54, 13, 18-1, 35, 36-3, 61-1, 61-2]4-1, 5-1, 13, 27-1, 27-2, 35, 36-3, 61-1, 61-2] Av. O/E = 0-81Av. O/E = 0-83Av. O/E = 0-86Av. O/E = 0-87 A n B C D = (4-1, 5-1, 35, 36-3, 61-1, 61-2} These sets support very similar cladograms; the favoured one, with the lowest average O/E, isshown in Fig. 16, but all are similar in showing two large groups (912-15 + 928-31 and 924-26 +942-47) and leaving taxa 919 and 932-34 unresolved. 18.1 61.2 I 5.1 61.1 Fig. 16 Cladogram for taxa in the EnicospiluslStauropoctonus complex less taxa 909-911 and 916-8,based on the favoured clique (A). Using the WAGNER option of PHYLIP several attempts were made to construct the shortestrooted tree possible. The minimum length tree (Fig. 17) necessitated 242 transformation steps. Both parsimony and compatibility analyses have certain similarities. Both taxa 909 + 916-18formed a group primitive to most other taxa whilst Enicospilus (942-47) and Dicamptus(932-34) are amongst the most specialized taxa and separate from Laticoleus (912-15) andLeptophion (928-31). However, the Wagner parsimony method usually separated the Hawaiiantaxa (924-26) as a relatively primitive, discrete group but placed Ophiogastrella (910-11) well inthe Enicospilus/ Dicamptus/ Leptophion/ Laticoleus complex. Compatibility methods suggestedthe reverse. To assess the relative merits of these alternative arrangements it is necessary toevaluate the characters upon which they are based. As mentioned above, the exclusion of taxa910 and 911 is based on characters 15-1, 22-1 and 33 which are present in the apomorphiccondition in all other taxa. The former two are striking structural modifications and whilst thelatter is a loss apomorphy , it is an unusual reduction (when considered for the family as a whole).There is no apparent functional reason why these characters should be linked, so theircongruence can be viewed as strong evidence for excluding 910-11 as the first branch. However,910 and 911 do have a number of apomorphies in common with many other taxa in the group, 102 I. D. GAULD IO CO Fig. 17 Minimum length cladogram produced by Wagner parsimony analysis of the EnicospiluslStauropoctonus complex. This tree requires 242 transformation steps. especially the Laticoleusl Leptophion complex (e.g. characters 7-2, 27-1, 27-2, 20-1) which itmust be assumed has been derived in parallel if Ophiogastrella is indeed the most primitivebranch. The Wagner analysis united all taxa excluding the Stauropoctonus/909 complex on the basis ofcharacters 15-1, 36-2 and 43-1. The former has been postulated as having undergone reversal inthe stem 910 + 911 so cannot be considered a group apomorphy, whilst 36-2 is highlyhomoplastic in any favoured arrangement, and in the minimum length tree necessitates sixtransformations that include being developed, lost and subsequently redeveloped. In fact, thischaracter is also present in the apomorphic condition in some Stauropoctonus (916, 918), and itis not stretching credibility to consider it an apomorphy of the entire group here being analysed,and postulate reversal in taxa 909 and 917. This hypothesis involves only a single additionaltransformation step, and would seem biologically more feasible than the gain-loss-gain schemefavoured by the Wagner analysis. The character itself, relative position of two wing veins, hasnot been used in higher classification, though it is of considerable use in separating species(Gauld & Mitchell, 1978), and has doubtfully assigned polarity. It is also amongst the very worstcharacters suggested by the original LeQuesne test (see Table 6). The third character, 43-1, isfound widely in the apomorphic condition throughout the Ophioninae. Either it has been PHYLOGENY OF THE OPHIONINAE 103 derived in parallel in the Euryophion group, in most Ophion and related taxa and in certainmembers of the present group, or it is an apomorphy of the Ophioninae and the apparentlyplesiomorphic state of most Stauropoctonus group taxa represents a further apomorphy.Initially the polarity of this character was considered to be tentatively assigned (p. 77).Elsewhere in the Ichneumonidae, although reliance is often placed on character 43-1 in generickeys, it is variable in most higher taxa (cf. Pimplinae in Townes, 1969), suggesting it is a highlyhomoplastic feature. Other characters involved in placing Ophiogastrella high up the tree are 7-2, 20-2, 27-1, 27-2,37, 43-2, 45-2, 48-1 and 59. Of these, 7-2 and 59 are highly homoplastic, gain-loss-gain characterswhich seem biologically implausible, and they also have dubious polarity assignment. Character45-2 is postulated as having undergone reversal in the stem 910 + 911 so there remains fivereasonably robust apomorphies uniting Ophiogastrella with the most specialized taxa (20-2,27-1, 27-2, 43-2 and 48-1). Considered individually, all of these characters are rather 'weak'. Theremarks made above about 43-1 can also be applied to 43-2; the loss apomorphies 27-1 and 27-2have identical state distributions in this group and should perhaps be considered as a singlefeature, the loss of the posterior mesosternal transverse carina. This carina has been lost,presumably independently, in other ophionine evolutionary lines (e.g. Ophion, some Lepto-phion, some of the Stauropoctonus group) and has commonly been lost in many ichneumonidevolutionary lineages outside the subfamily under consideration. As a consequence it is not acharacter which seems to unite convincingly Laticoleus (912-5) with Ophiogastrella. It isinteresting to note that there is at least some degree of correlation between the presence orabsence of this carina and the type of habitat occupied. In groups where it is usually absent (e.g.the Phygadeuontinae), species inhabiting very wet areas have the carina complete. In thesubfamily mentioned this includes common species of Paraphylax, Amauromorpha andApsilops found in Old World rice padi. In the Anomaloninae the two closely related generaTherion and Heteropelma differ in this feature; the former is characteristically found on dry openareas, whilst most Heteropelma species occur in more humid woodlands (Gauld, 1976). In theOphioninae the carina is always present in those species found in humid rain forests (e.g. mostEnicospilus, Leptophion and Dicamptus) but is incomplete in species favouring drier, moreexposed habitats (e.g. some Australian Leptophion, most Ophion). Both Ophiogastrella speciesand Laticoleus seem to favour drier forest habitats than Leptophion. Character 20-2, the development of scutellar carinae, has probably been derived indepen-dently in several evolutionary lineages (e.g. some species of Ophion and Leptophion (Gauld &Mitchell, 1981)), but it also appears to have undergone reversal in other groups (e.g. in taxon944 and also Enicospilus arduus (Gauld & Mitchell, 1978)). It would be only slightly lessparsimonious to suggest that the presence of complete scutellar carinae is an apomorphy of theentire Enicospilus I Stauropoctonus complex, and to suggest that it has been lost in a few lineages.Certainly the shared presence of these carinae is not a convincing character for unitingOphiogastrella with other genera. The remaining character, 48-1, is a difference in relative lengths of wing veins and itsapomorphic state is found scattered throughout the subfamily. To summarize therefore, the decision on where to position Ophiogastrella hinges on twocharacter sets, one of three unusual and fairly convincing apomorphies (15 1 , 22- 1 and 33) versusfive widely distributed, and thus unconvincing, apomorphies (20-2, 27-1, 27-2, 43-2 and 48-1).The Wagner method, in attempting to minimize tree length, opts for including Ophiogastrellaamongst the most derived species, whilst the compatibility method favours treating it as the mostprimitive taxon, thus eliminating homoplasy in three characters (22-1, 33 and 15-1) as the othercharacters are mutually incompatible. The position of the Hawaiian genera (924-6) also presents a problem. In the parsimonydendrogram (Fig. 17) the five characters defining the stem 924 + 925 + 926 (5-1, 15-2, 36-3, 61-1and 61-2) also define at least part of the Enicospilus + Dicamptus branch (932-4 + 942-7)suggesting the Hawaiian genera could be placed on the larger branch, as in the compatibilitydendrogram. The parsimony method has excluded the Hawaiian genera on the basis ofcharacters 7-2, 20-2, 40, 43-2, 45-2, 46, 48-1 and 59. Of these 7-2 and 59 are the biologically 104 I. D. GAULD implausible gain/loss/gain characters with dubiously assigned polarity, so the real choice has tobe made between the two character sets 5-1, 15-2, 36-3, 61-1 and 61-2 versus 20-2, 40, 43-2, 45-2,46 and 48-1. Considering the first set first, characters 5-1, 61-1 and 61-2 are all mandible characters. Theapomorphic state of all three is elsewhere only found in taxa 916-8 and 942-7 (excluding 944).Although this character set has almost certainly been derived independently in StauropoctonuslAulophion and Enicospilus it is not found elsewhere in the Ophioninae. It is of very rareoccurrence in other parts of the family, and where it does occur it is usually characteristic of agroup of genera e.g. the Orthocentrinae (Townes, 1971) and the XanthopimplalEchthromorphagroup of Pimplini (Townes, 1969). Character 15-2, a completely concealed spiracular sclerite, isfound in virtually all species of Enicospilus and Dicamptus. Elsewhere it is found in the derivedcondition only in the Hawaiian genera, in a few species of Leptophion and in the aberrantmonotypic genus Sicophion (922). The final character in this set, 36-3, an extreme reduction inthe length of a wing vein, is only found in a very few taxa, notably (in the matrix) the Hawaiiangenera, some Enicospilus, Euryophion and Xylophion. Elsewhere in the Ophioninae it onlyoccurs in some Dicamptus and one aberrant Laticoleus. The second and alternative character set (20-2, 40, 43-2, 45-2, 46 and 48-1) has partially beenconsidered above where characters 20-2, 43-2 and 48-1 have been suggested to be rather poorindicators of phylogenetic affinity. Two of the Hawaiian genera show the derived state of 48-1,so it would be equally parsimonious to postulate that a reversal had occurred in one Hawaiiangenus, as opposed to parallel derivation in the subgroup of two Hawaiian taxa and the major linein the larger group. The remaining characters all refer to the fore wing, in particular theantero-distal part of the discosubmarginal cell and the adjacent vein, Rs+2r. The apomorphicconditions of characters 40, 45-2 and 46 are the most characteristic features of Enicospilus/Dicamptus, though the same combination is also found in a very few other taxa (e.g. Riekophionspecies). Several species of Enicospilus may have one or more characters in the plesiomorphicstate (e.g. the E. senescens species-group (Gauld & Mitchell, 1978)), but the trace featuresfound in one or two species suggests this condition represents a reversal. This suite of charactersdoes seem therefore to be good evidence for excluding the Hawaiian taxa from the Enicospiluslineage. However, the plesiomorphic condition of all three of these characters occasionally isfound in some Enicospilus species inhabiting oceanic islands, and in these cases it can be seenthat a reversal has occurred as related taxa still have the apomorphic states. For example, thethree species E. vidus, E. ditor and E. donor form a closely interrelated group on the GalapagosIslands (Gauld & Carter, 1983) . They belong to the E. capensis species-group and probably haveevolved from a migrant South American species, all of which have the apomorphic condition forcharacters 40, 45-2 and 46, as has E. vidus. In E. donor they all are plesiomorphic, whilst in E.ditor intermediate conditions exist. This strongly suggests reversal has occurred. A similar casecan be demonstrated for Hawaiian Enicospilus (see Cushman, 1944). The majority of Enicospi-lus species are both synchronous and sympatric with numerous congeners. The three charactersare developed in such a way as to form a distinctive, species-specific pattern (especially the exactshape of the sclerite, character 46) (see figs 384-641 in Gauld & Mitchell, 1981). In areas wherefew species occur, such as deserts, there is considerable variation within a species in the exactexpression of these characters. Possibly their reversal on oceanic islands is facilitated by areduction in selection pressure that favours uniformity in areas of high possible interspecificinteraction. It is possible that the apparent plesiomorphic condition of these characters in theHawaiian genera is also a reversal, though in this case no definite close relatives are known. The remaining character in the second set, 48-1, refers to the relative lengths of veins 3r-mand M . The apomorphic condition has previously been used only to characterize species andboth states can be found in most genera, suggesting it is a rather homoplastic feature. To sum up, once again the Wagner method positioned a group on the basis of the larger setand again biological evidence suggests that perhaps this is incorrect and the position indicated bythe compatibility method is more plausible. The parsimony analysis united all species of Enicospilus (942-47) into a single clade butDicamptus was paraphyletic with respect to this. This is not surprising as traditionally Enico- PHYLOGENY OF THE OPHIONINAE 105 spilus is separated from Dicamptus on account of its specialized mandibles; Dicamptus isrecognized only by the specialized features it has in common with Enicospilus plus thepossession of unspecialized mandibles, a plesiomorphic feature (Townes, 1971). The compati-bility method proved to be less successful at producing groups corresponding with recognizedgenera, probably because Enicospilus and Leptophion are polythetic taxa (Gauld & Mound,1982). The relative positions of Leptophion, Laticoleus and Pamophion are less clear. The Wagnermethod suggested Leptophion and Laticoleus form separate groups, though both are charac-terized by the apomorphic condition of 35, and with one exception, 37. The compatibilitymethod suggests Laticoleus + Leptophion form a group defined by character 35 and these twotaxa, plus Pamophion, form a group defined by character 3. To conclude this section, on balance, secondary evidence suggests Ophiogastrella and theHawaiian genera are better placed by the compatibility method whilst Enicospilus is mostsuccessfully aggregated by parsimony analysis. A 'hybrid' compromise cladogram was post-ulated that required 249 transformation steps (Fig. 18). This compromise arrangement groupsvarious congeneric species together, thus to some extent corroborating the initial assumptionthat ophionine genera are not polyphyletic groups. The relationship of the component genera of the Enicospilus/ Stauropoctonus complex can beillustrated most clearly if the highly homoplastic characters are omitted, the relative positions of "O X5 o ^O ^Q ^^ ^QNO N 5^ 00 00 K) GO Fig. 18 Hybrid cladogram derived from subjective evaluation of parsimony and compatibility analyses ofEnicospilus/ Stauropoctonus data set. This arrangement requires 249 transformation steps. 106 I. D. GAULD 16-1, 25-1. 25-2, 28-1. 361, 38,42-1,43-1, 45-1, 471, 49, 51-1,51-2,63' Fig. 19 Cladogram showing putative phylogenetic relationships between the genera of the Enico-spiluslStauropoctonus complex as supported by the least homoplastic characters. PHYLOGENY OF THE OPHIONINAE 107 species within genera disregarded (and the object of this part of the analysis is to establish thephylogenetic relationships of the genera) and single exceptional character scores, that arealmost certainly incidences of homoplasy (e.g. 15-2 in taxon 945; 61-2 in taxon 944), ignored.This simplified cladogram together with supporting characters is illustrated in Fig. 19. Essenti-ally it is the same arrangement as Fig. 18. This cladogram suggests that Stauropoctonus is paraphyletic with respect to Aulophion,Leptophion is paraphyletic with respect to Laticoleus and Enicospilus is paraphyletic withrespect to the PycnophionlAbanchogastralBanchogastra lineage. The major groups and their relationship with the unassigned taxa The inter-relationship of the two major generic complexes The Enicospilus/ Stauropoctonus complex has been defined above as a holophyletic group on thebasis of 17 apomorphies (16-1, 224, 25-1, 25-2, 28-1, 28-2, 33, 36-1, 38, 42-1, 43-1, 45-1, 47-1, 49,51-1, 51-2 and 63) (though Ophiogastrella is primitive in respect of two, 22-1 and 33) and theThyreodon/Euryophion complex has similarly been defined by a possible 19 apomorphies (2,7-1, 22-2, 24-1, 24-2, 25-1, 25-2, 28-1, 28-2, 38, 42-1, 42-2, 47-1, 49, 51-1, 51-2, 54, 58-1 and 614).Ten of these apomorphic features (25-1, 25-2, 28-1, 28-2, 38, 42-1, 47-1, 49, 51-1 and 51-2) arecommon to both sets, suggesting a sister group relationship. Of the apomorphies remaining fordefining the Thyreodon/Euryophion lineage, three (24-1, 24-2 and 61-1) occur extensively invarious species-groups and species in the Enicospilus/ Stauropoctonus complex, thus weakeningtheir credibility as characters defining the former lineage. Character 42-2 has apparently beenderived in parallel in the Stauropoctonus line, but the remaining five features (2, 7-1, 22-2, 54and 58-1) are robust characters that strongly support the monophyly of the ThyreodonlEuryophion lineage. Of the six remaining apomorphies defining the Enicospilus/ Stauropoctonuslineage (16-1, 22-1, 33, 36-1, 45-1 and 63) two (22-1 and 33) exclude Ophiogastrella; theapomorphic state of 33 has also been derived in parallel in the Thyreodon line. The derivedstates of characters 16-1 and 36-1 occur in most species of the Thyreodon/Euryophion complex,suggesting they are group apomorphies, whilst the derived state of 45-1 also occurs, in parallel,in Euryophion. Character 63 is the single convincing apomorphy supporting the monophyly ofthe Enicospilus/ Stauropoctonus complex. A cladogram showing the putative phylogeny of thesetaxa is shown in Fig. 20. The position of the unassigned taxa The position of the remaining taxa (905, 906, 920, 922, 923, 927, 938-41 and 948-51) can now beconsidered in relation to the robust arrangement of the majority of ophionine taxa presentedabove. To reduce the data matrix to more manageable proportions a number of hypotheticaltaxonomic units (HTUs) were constructed to represent genera reasonably placed in the earlierpart of the work. Scores were assigned to these HTUs by marking each character with thecondition found in the majority of species of the genus (including all taxa examined in earliertaxonomic studies (Gauld, 1977; Gauld & Mitchell, 1978, 1981) but not included in the primarydata matrix). This method of scoring, rather than deducing a hypothetical ancestor, was adoptedbecause of the high incidence of homoplasy. This idiosyncracy is unlikely to alter any cladisticarrangement significantly, whereas the assumption that the group-ancestor has a particularcharacter present in the plesiomorphic condition, based on occasional incidence of the occur-rence of the plesiomorphic (and possibly reversed) condition in some members of the group,could result in considerable alteration to the position assigned to the group. The HTUs utilizedare 960 (Laticoleus), 961 (Ophiogastrella), 962 (Leptophion), 963 (Stauropoctonus), 964(Thyreodon/ Dictyonotus) , 965 (Euryophion), 967 (Dicamptus), 968 (Enicospilus). [N.B. Thereis no 966.] The character scores of these HTUs are given in Appendix 7. The results of a LeQuesne test on taxa 905, 906, 920, 922, 923, 927, 938-41, 948-51 and 960-68are given in Table 17. The overall O/E value is strikingly high (0-90), indicating the extremelyhomoplastic nature of this data set. Perusal of the labels matrix (Appendix 8) shows that a large 108 I. D. GAULD 16.1*', 25.1*. 25.2*. 28.1*. 28.2*, 36. 1'^38*', 42.1, 47.1', 49, 51.1*. 51.2*B Fig. 20 Putative phylogenetic inter-relationships between groups of taxa in the Enicospilus/Stauropocto-nus and ThyreodonlEuryophion complexes. number of characters have been labelled highly for certain taxa; this will be discussed below. Wagner analysis of this data set yielded a number of equally short cladograms with quitedifferent topologies (Figs 21-24). A large number of other trees that were only slightly lessparsimonious were also generated (Gauld, unpublished data). Compatibility analysis yielded asingle largest clique of 11 informative characters (2, 3, 7-1, 17, 25-1, 26, 38, 40, 41, 47-1 and 52-1)which supports the cladogram shown in Fig. 25. Although differing in many topological details,certain features are common to the dendrograms. In all of the Wagner trees the two majorcomplexes emerged as the most derived groups. Five taxa (938, 939, 941, 948 and 951) occupythe most primitive position in three of the dendrograms and are amongst the eight most primitivetaxa in the other two, suggesting these species are amongst the most primitive of all ophionines.This group of species comprises examples of the genera Xylophion, Ophion, Agathophiona,Rhopalophion and Sclerophion, Gauld (1979) suggested these might represent a holophyleticgroup, the Ophion subgroup, characterized by having a slender Rs+2r, which is also straightbasally, having an incomplete posterior transverse carina of the mesosternum, having asubapically impressed clypeus, having Im-cu usually with a ramellus, and always angled, andhaving the anterior area of the propodeum 'occluded'. However, the holophyly of the group isopen to question. The shape of Rs+2r (corresponding to characters 39, 40 and 62), of Im-cu (38 PHYLOGENY OF THE OPHIONINAE 109 Table 17 Results of LeQuesne test on data set comprising HTUs and unplaced taxa (905, 906, 920, 922, ^ : 26 38.2 (J.68 7.2: 47 61.0 0.77 14.2: 46 54.1 0.85 IT_ : 24 25.6 0.94 20.1: 60 60.0 1.00 23 : 43 51.1 0.84 25.2: 58 61.2 0.95 28.2: 57 59.6 0.96 31^ : 28 25.6 1.09 36.1: 25 24.9 1.00 38 : 34 51.1 0.67 42.1: 61 60.2 1.01 44.1: 59 61.2 0.96 46 : 33 45.9 0.72 48.2: 32 37.5 0.85 54 : 39 45.9 0.85 j>9 : 45 59.6 0.76 63 : 52 60.9 0.85 Grand total- 1528 observed, 1705.7 expected. Overall 0/E ratio = 0.90 Ranking ratios of scoring characters 47.1 25.1 40 2 38 3 46 22.1 36.3 59 7.2 55 6.1 19 45.2 15.2 33 58.1 23 54 14.2 48.2 63 52.1 61.1 37 24.2 43.2 29.2 24.1 18.1 15.1 47.2 20.2 27.1 51.2 44.2 61.25.1 9 27.2 18.2 7.1 17 26 57 12 25.2 28.2 41 44.1 1.1 16.2 62 51.1 22.2 32 29.142.2 20.1 36.1 1.2 42.1 39 35 48.1 36.2 45.1 31 30 14.1 43.1 and 47) and the form of the propodeum (22) are symplesiomorphies and therefore notadmissable for defining a holophyletic group. Both the incomplete posterior transverse carina ofthe mesosternum (equivalent to 1 scores for 27-1 and 27-2) and the impressed clypeus(corresponding to a 1 score for character 7-2) are apomorphic features found in many other taxa.In many of the Wagner trees generated the five taxa formed a clade defined by characters 18-1and 44-1 in addition to 7-2 and 27-2 (27-1 is most parsimoniously considered an apomorphy ofthe subfamily) . All of these characters are highly homoplastic so it is not surprising that thecompatibility cladogram leaves these taxa as an almost unresolved paraphyletic assemblage. Four further taxa, 920 (Riekophion emandibulator) , 922 (Sicophion pleuralis) , 940 (Eremo-tylus boguschi) and 950 (Hellwigiella nigripennis) have been placed in the same group as theOphion subgroup (Townes, 1971; Gauld, 1979). The undescribed genus (949) would, onTownes' (1971) criteria, also belong to this group. There is no evidence in any analysis thatsuggests that these taxa and the Ophion subgroup constitute a holophyletic clade. The definitionof the group as adopted by previous authors rests on a plesiomorphic feature, possession of amembranous flange on the fore tibial spur, suggesting this 'group' is in fact a paraphyleticassemblage. In Wagner analyses 920, 922, 940, 949 and 950 may be positioned primitively withrespect to most other taxa (e.g. Figs 22, 24) but some members may occasionally be united withother taxa. Taxa 949 and 950 are sometimes placed in much more derived positions, the formerwithin the Enicospilusl Stauropoctonus complex (e.g. Fig. 23) and the latter as the sister-lineageto the ThyreodoniEuryophion complex (e.g. Fig. 21). Hellwigiella (950) is placed as thesister-group to the ThyreodoniEuryophion complex on the basis of a number of shared, ratherstriking, derived characters. These characters (which may be shared by several or all species inthe complex) are adaptations to a diurnal eremic existence, and include possession of small ocelli(12), a shorter stouter flagellum (14-1), pendant epipleuron 2 (30), cylindrical hind tibial spurs(54) and long weakly curved claws (58-1). Many or all of these features are also found in the 110 I. D. GAULD 3> ' i t ( ) O Fig. 21 283 transformation step cladogram derived from Wagner analysis of HTUs plus unplaced taxa. apomorphic condition in other diurnal eremic ophionines including an undescribed species ofOphion from Australia (in ANIC) and Agathophiona species from Mexico. In another group ofichneumonids, the Anomaloninae, similar differences occur between closely related eremic andnon-eremic organisms (e.g. Gravenhorstia (Erigorgus) and Gravenhorstia (Gravenhorstia)species (Gauld, 1976)). The occurrence of so many apomorphic states of these characterstogether in day-flying species strongly suggests the characters should be considered as acharacter suite and perhaps accorded less taxonomic weight. Unlike members of the ThyreodonlEuryophion complex, Hellwigiella shows the plesiomorphic state of important characters usedfor defining the group (2, 7-1, 44-2). Furthermore, a position near to the base of the common PHYLOGENY OF THE OPHIONINAE 111 -~ -< Fig. 22 283 transformation step cladogram derived from Wagner analysis of HTUs plus unplaced taxa. stem of the two major complexes is suggested by the fact that Hellwigiella shows the plesiomor-phic states of characters 16-1, 42-1, 52-1 and 52-2. The position of the new genus (949) is rather perplexing as the taxon exhibits a number ofderived features shared with the more specialized taxa, especially those in the EnicospiluslStauropoctonus complex (e.g. 15-1, 16-1, 22-1, 33, 45-1) yet at the same time lacks one majorapomorphy of the group (63) . Taxon 949 also shows the plesiomorphic state of many of the stemcharacters, including 25-2, 28-2, 42-1, 51-1 and 52-2. It does, however, share a small number ofderived features (including 1 1 , 15 1 , 16- 1 , 33 and 45-1) with another enigmatic taxon , Sicophionpleuralis (922). Sicophion, like taxon 949, is primitive in a surprising array of other features andit seems plausible that these two taxa have a sister-group relationship and represent a primitivegroup that evolutionarily converged with the Enicospilus group, possibly as a result of similarselection pressures. The striking phenetic resemblance between 949 and some species ofLeptophion is perhaps more understandable when one considers these taxa have apparentlyevolved in mid altitude tropical cloud forests - 949 in the Neotropics and Leptophion in SouthEast Asia. Taxon 920, Riekophion emandibulator , is often placed with 967 and 968 (Dicamptus and 112 I. D. GAULD co m GO Fig. 23 284 transformation step cladogram derived from Wagner analysis of HTUs plus unplaced taxa. Enicospilus) in compatibility analyses on the basis of sharing the derived states of characters 40and 46 (e.g. Fig. 25). In other respects Riekophion does not appear to be at all closely related toeither Dicamptus or Enicospilus (Gauld, 1977) and the high label scores obtained by the speciesfor the derived states of these characters (Appendix 8) strongly suggests parallel derivation intwo lineages. PHYLOGENY OF THE OPHIONINAE 113 C/> i i i i Fig. 24 284 transformation step cladogram derived from Wagner analysis of HTUs plus unplaced taxa. Accepting that the taxa of the Ophion 'genus group' (920, 922, 938-41, 948-51) are the mostprimitive ophionines, then the most parsimonious arrangement of these taxa with respect to allother groups is that presented in Fig. 26. All other ophionine taxa, 922 + 949 + 920 and 950 +940 remain as an unresolved trichotomy. The Ophion genus-group (i.e. the Ophionini ofTownes, 1971) is therefore a paraphyletic grade, an assemblage of less-specialized ophionines.The 'group' possibly comprises three apparently holophyletic clades, the Ophion-group (=Ophion subgroup of Gauld, 1979), the Eremotylus group (940, 950) and the Sicophion group(920, 922, 949). The clade, 'all other ophionine taxa', comprises the two groups discussed above(the Thyreodon/Euryophion complex and the Enicospilusl Stauropoctonus complex) and fourunplaced taxa, 905, 906, 923 and 927. The position of these four taxa can now be considered inrelation to the fairly rigid structure derived above. Taxa 905 and 906 (Simophion calvus and Orientospilus melasmd) share 27 derived characters(Appendix 3); 906 also has a large number in common with other taxa, particularly 947 (28), 946 114 I. D. GAULD O O O "O -O O O "O O O -OE:ao*<>KJK)p!OpOi O CO ' O K3 NJ C*> Oi *_zt "O ^O O O *O Oa ^ O O K) |v(Ji O SI 00 40 t 47.1' 8g ^ 25.1* Fig. 25 Cladogram based on largest compatible set of characters obtained from HTUs plus unplaced taxa. (27), 942 (26), 918, 915, 914, 911 and 910 (25). Taxon 905 is less closely associated with othertaxa but shows some affinity with the Euryophion group (937 (25), 936 (24) and 935 (23)). Taxa905 and 906 share several unusual apomorphic features including 6-1 (elsewhere only found inthe derived state in 909, 935-37 and 946), 19 (elsewhere only found in 908, 918, 935 and 937), 23(elsewhere only found in the DictyonotuslThyreodon lineage) and 41 (elsewhere only found in922 and 923). Taxa 905 and 906 also share all the stem characters of the EnicospiluslStauropoctonus + Thy reodonl Euryophion lineage, except that 905 is plesiomorphic for 42-1.This suggests that they may belong near the other groups, possibly as a sister-species pair. Of thefeatures defining the two major complexes these taxa share only one, an apomorphy of 63,suggesting placement near the base of the Enicospilusl Stauropoctonus lineage. It is noteworthythat in earlier Wagner analyses (e.g. Figs 22-24) 905 and 906 generally were split up, 905 oftengrouping with 923 , 964 or 965 , whilst 906 usually was united with 960 or 961 . If the two taxa wereentered together first in the data file, then they were not separated , but remained as a distinctclade(Fig.21). Taxon 927 (Barytatocephalus mocsaryi) shares 24 derived characters with taxon 936, 23 with914 and 946 and 22 with 913. It shares all the stem features with the major complexes andadditionally has the apomorphic condition for 22-2 and 58- 1 , suggesting it belongs near the baseof the Thy reodonl Euryophion lineage. It does not have the major apomorphies of theEnicospilusl Stauropoctonus branch (that is 22-1, 33 and 63). Taxon 923 (Prethophion latus) is an enigmatic Neotropical taxon that has relatively fewcharacters in common with any other species. The largest number, 22, are shared with 918 and ithas 21 in common with 906. It does have all the 'gain' stem apomorphies in common with otherhigher taxa, but is plesiomorphic for 25-2, 28-2 and 45. It does not have any of the apomorphiesdefining the Thyreodonl Euryophion complex, but is apomorphic for character 63, suggesting itbelongs near the base of the Enicospilusl Stauropoctonus stem. It is most parsimonious to treatPrethophion latus as the sister species of the 905 -I- 906 lineage (on the basis of the derived stateof characters 27-1 and 27-2), but this association must be regarded as very tenuous. PHYLOGENY OF THE OPHIONINAE 115 941 951 939 948 938 950 940 922 949 920 o 51.1 Fig. 26 Most parsimonious arrangement of 'Ophion genus-group' and all other taxa. Conventions as inFig. 13. The cladogram showing the preferred arrangement of taxa as described above is shown in Fig.27. It is interesting to note that this user defined cladogram is one step more parsimonious thanany obtained by Wagner analysis. 116 I. D. GAULD o . . Fig. 27 Cladogram derived from subjective evaluation of character complexes involved in grouping taxain cladograms presented in Figs 21-25. It is noteworthy that this, at 242 steps, is a more parsimoniousarrangement than the best obtained by Wagner analysis, though this is a fortuitous event. Discussion of suggested phytogeny A putative phylogeny of the Ophioninae is shown in Fig. 27. Parts of this cladogram seem to befairly robust, but other groups are only supported by weak homoplastic characters. The overallarrangement, placing the Ophion and Eremotylus groups in primitive positions and having abifurcated evolutionary line, does seem reasonably sound. It corresponds with increasingmorphological complexity, particularly in the form of the fore tibial spur (character 51), theregion of the spiracular sclerite (character 15), the modification of the propodeum (characters22, 25) and the loss of umbo (character 33). What little is known about the structure of the finalinstar larvae supports this arrangement. The cephalic capsules of the larvae of Enicospilus,Dicamptus, Euryophion and Thyreodon are clearly more specialized than those of Ophion inhaving a modified hypostoma (Short, 1978). The larvae of Euryophion and Thyreodon aresimilar in having numerous setae (11+) on each part of the maxillary lobe median to thehypostomal spur and adjoining the maxillary palp. Dicamptus and Enicospilus have eight or lesssetae in this position. Certain venational characters also support the phylogeny proposed. Theseinclude modification of Im-cu and broadening of the base of Rs+2r (characters 38, 47 and 62). PHYLOGENY OF THE OPHIONINAE 117 However, in almost all adult characters it is necessary to postulate parallelism. Reduction of thetibial spur membrane has occurred independently in Xylophion and the main evolutionary line,and specialization of the propodeum has occurred in both the main lineage and Eremotylus. It isquite disconcerting to see the degree of evolutionary convergence that has occurred in separatelineages (such as the development of alar sclerites and a sinuous Rs+2r in Riekophion andDicamptuslEnicospilus and the general similarity of structure between the undescribed CostaRican genus and species of Leptophiori) . No amount of rearrangement of lineages could possiblyremove any but a small proportion of such homoplasy. Perhaps the development of suchsimilarities is due to some underlying adaptation of the genotype not visually manifested (theunderlying synapomorphies of Saether, 1979). It is difficult to explain otherwise (unless onepostulates reticulate evolution - 'ancient hybrids' giving introgression) how such an unusualfeature as alar sclerites has appeared, apparently independently in at least six evolutionarylineages of Ophioninae (Afrophion, Sclerophion, Leptophion, Laticoleus, Riekophion andDicamptuslEnicospilus). Similar examples of such unusual parallelisms can be found in themodification of the penultimate distal hamulus, the projecting pecten of the hind tarsal claw andthe development of a spine on the hind trochantelli. The Ophion group of genera (Ophion, Alophophion, Sclerophion, Afrophion, Rhopa-lophion, Agathophiona, Xylophion) is defined by the possession of the apomorphic states offour rather homoplastic characters, the loss of the posterior transverse carina of the meso-sternum (27-2), the possession of a diagonal mesopleural furrow (18-1), possession of animpressed clypeal margin (7-2) and having Rs in the hind wing at least slightly curved (44-1). Notall species of these genera necessarily possess all these apomorphies; Afrophion species areexceptional in having both a blunt clypeal margin and a straight Rs whilst some Rhopalophionspecies have the mesopleural furrow obsolescent. Although this genus-group is amongst themost primitive of ophionines, its position at the base of the phylogenetic tree rests on theassumptions that characters 38 and 47-1 (shape of Im-cu and presence of a ramellus) are cor-rectly polarized. As mentioned above (pp. 77, 78) the polarity of both features is questionableand if it could be demonstrated that the assigned polarity is incorrect then these two featureswould become convincing apomorphies supporting the holophyly of the Ophion group. Thischange in polarity would necessitate placing the Ophion group a little higher up the evolutionarytree, possibly as the sister lineage to the Eremotylus group; these two groups share theapomorphic states of characters 18-1 and 44-1. The holophyly of the Eremotylus group is supported by the angulate base of Rs+2r (39),which is somewhat thickened (62), and by the form of the clypeus. In most species it is slightlyflared outwards in profile (7-1) and often slightly concave. The mandibles are a subtly differentshape to those of virtually all other ophionines but this feature is very difficult to define. The Sicophion group is a rather tenuous association of primitive taxa. The holophyly of thegroup is supported by the apomorphic states of characters 33 and 45-2. Apart from possession ofcertain primitive features such as the plesiomorphic state of character 51-2, and more extensivepropodeal carination, all of these taxa appear to be highly specialized, sharing a variety ofapomorphic features with other taxa placed higher up the phylogenetic tree. More evidence ofthe holophyly of this group needs to be obtained before it can confidently be accepted. The holophyly of the Thyreodon group is supported by the specialized form of the anteriorpart of the propodeum (character 22-2) and by the form of the claw (character 58-1). Theinclusion of Barytatocephalus within this clade does seem reasonable. In addition to thecharacters used in the analysis, two other features suggest Barytatocephalus belongs here. Vein3r-m in the fore wing of species of this genus forms a more obtuse angle with Rs+2r than is thecase with most other ophionine taxa. Most other members of this clade have a similarspecialization. At the base of the hind wing, in the anal cell, a vestige of a vein (? 2A) is oftendiscernible. In most ophionines this trace, when observable, can be seen to be fairly near to andparallel with the hind margin of the wing; in both Barytatocephalus and other members of theThyreodon complex (but not Euryophion) this vein trace is remote from the wing margin andclose to and parallel with \A. Townes (1971) considered Barytatocephalus to be a derivation ofEnicospilus, but there is little evidence for this supposition. The exposed spiracular sclerite, the 118 I. D. GAULD general head shape and form of the propodeum suggest Barytatocephalus should not be includedin the Enicospilus lineage. The Enicospilus group is the largest complex and one of the most difficult to define. Itsholophyly is supported by a single apomorphy, loss of vestigial first laterotergites (character 64),but the included taxa do resemble each other in possessing a large number of derived features incommon. Within this group five subordinate lineages are recognizable, the Orientospilus,Ophiogastrella, Stauropoctonus, Leptophion and Enicospilus subgroups. Each of these isapparently a holophyletic group. These are the various groupings recognized in the classificationproposed below. A classification of the Ophioninae Hennig (1966) pointed out the existence of two sorts of monophyletic taxa - holophyletic andparaphyletic groups. Traditional taxonomists may accord either group the status of a super-generic rank, though cladists do not recognize-paraphyletic taxa (Farris, 1979; Carpenter, 1982).Whilst there are many cases of small paraphyletic taxa that can be satisfactorily amalgamatedwith a holophyletic taxon to form a slightly larger holophyletic taxon (e.g. Aulophion +Stauropoctonus), there are other cases where dogmatic adherence to cladistic tenets producesimpractical results. For example, if the Hawaiian genera were incorporated into Enicospilus togive a single holophyletic taxon, the resultant genus would contain an extraordinary range ofmorphological and probably biological diversity. I concur with Martin (1981) in believing that aclassification, as opposed to a phylogeny, is a compromise between known or hypothesizedinter-relationships and nomenclatural convenience. A good classification has at least twoimportant facets: (a) it facilitates generalizations to be made about the component taxa, and (b)it is predictive. Farris (1979) and Mickevich (1978) both contend that a purely phylogeneticclassification is an optimum as it is both more predictive and more stable than one that is notentirely phylogenetic. Whilst I agree with these authors that this is often the case, I do not think itis always so, especially where there are very unequal rates of evolution in different lineages.Consider, for example, the ichneumonid subfamily Mesochorinae. Five genera, Cidaphus,Astiphromma, Mesochorus, Stictopisthus and Plectochorus each show progressive morphologi-cal complexity. All are quite large genera and are functionally useful, but only Plectochorusappears to be holophyletic. Each of the other four taxa is apparently paraphyletic with respect tothe genera that are more specialized than it (Townes, 1971; Gauld, 19846). Although not a strictphylogenetic classification, this arrangement of the species (which is based on adult morpho-logical features) was found to have predictive value for larvae (see Short, 1978) which seem tocorroborate the idea that these taxa are a nesting paraphyletic series. I personally see no way ofestablishing a strict phylogenetic classification for such a group, other than by lumping all thegenera, an action that would result in a much less informative and less useful classification thanthe present one. The existence of such nesting paraphyletic groups would seem to be a corollary of thepunctuational model of evolution. Stanley (1979) argues convincingly that many major adaptivechanges have occurred very rapidly in isolated populations - thus a new organizational level(higher taxon) may have arisen from a small population of a pre-existing (and, if the rate ofphyletic evolution is small, virtually unchanging) species placed in a different higher taxon. To return to the present example. If the Hawaiian genera have resulted from rapid radiationfrom an Enicospilus ancestor into niches hitherto, for enicospilines, unexploited, it is fairlyunlikely that any prediction made from study of the biology of continental Enicospilus specieswill apply to Pycnophion. Similarly, any generalizations made about Enicospilus are likely tohave to be prefaced by 'except in some Hawaiian species'. Consequently the purposes ofprediction and generalization would both be better served by recognizing the Hawaiian taxa asdistinct genera, even though Enicospilus thus becomes a paraphyletic taxon. A second area of contention relevant to the present work concerns the reality and recognitionof polythetic superspecific taxa. The currently recognized ophionine genera have been acceptedinitially as monophyletic groups, though in several instances it has been mentioned that a PHYLOGENY OF THE OPHIONINAE 119 number are polythetic (Cushman, 1947). Some authors (e.g. L0vetrup, 1973) tend to dismisspolythetic taxa as the product of unsatisfactory discrimination. Whilst I accept that this is anexplanation, I dispute that it is the most obvious one. Most cladistic studies reveal thathomoplasy is a common phenomenon. If a data set is highly homoplastic, then logically it followsthat some phylogenetically 'real' clades will only be definable in a disjunct (sensu Hull, 1965)way. Imposing a monothetic classification on such groups would result in the erection of amultiplicity of new genera, many of which would be monobasic. The resultant classification willhave little predictive value as a high proportion of new taxa are likely to necessitate new generafor their accommodation (Gauld & Mound, 1982). The following classification is suggested for the subfamily Ophioninae. It is not strictlycladistic, as it recognizes some paraphyletic taxa; however it is a fair approximation of thepresumed phylogeny. Formal tribes are not recognized, but these could be used for the variousgenus-groups. It should be noted that some very small genera were not incorporated in thecladistic analysis (these are denoted by an asterisk), but they were closely studied in an earlierwork (Gauld, 1979). This has facilitated their placement in the present system. Subfamily OPHIONINAE Ophionidae Shuckard, 1840Enicospilini Townes, 1971 OPHION genus-group Op/iion Fabricius, 1798Alophophion Cushman, 1947*Sclerophion Gauld, 1979 rfitCAfrophion Gauld, 1979*AgathophionaWestwood, 1882Rhopalophion Seyrig, 1935Xylophion Gauld, 0-979 / ? y SICOPHION genus-groupRiekophion Gauld, 1977Sicophion Gauld, 1979Janzophion gen. n. EREMOTYLUS genus-groupEremoty/usFoerster, 1869Trophophion Cushman, 1947*Hel/wig/e/faSzepligeti, 1905 THYREODON genus-group BarytatocephaIusSchu\z, 1911Euryophion Cameron, 1906 Rictophion Townes, 1971 Syn. n.Dictyonotus Kriechbaumer, 1894Ophionopsis Tosquinet, 1903 Syn. n. IWiync/iop/iJonEnderlein, 1912T/rjreodonBrulle, 1846 ENICOSPILUS genus-group Orientospilus subgroupPrethophion Townes, 1971Simophion Cushman, 1947Orientospilus Morley, 1912 Ophiogastrella subgroupOphiogastrellaBrues, 1912 Stauropoctonus subgroupLepiscelus Townes, 1971Stauropoctonus Br nuns, 1889Aulophion Cushman, 1947 Syn. n. Leptophion subgroupPamophion Gauld, 1977Laticoleus Townes, in Townes & Townes, 1973Leptophion Cameron, 1901 Enicospilus subgroup P/campfusSzepligeti, 1905Enicospilus Stephens, 1835Pycnophion Ashmead, 1900Banchogastra Ashmead, 1900Abanchogastra Perkins, 1902 Zoogeographic discussionPresent day distribution The approximate numbers of species per area are given in Table 18. For the purposes of thisdiscussion Melanesia, Australia and New Zealand are considered as a single region, theAustralo-Pacific, whilst for purposes of comparison Hawaii, with three endemic and onecosmopolitan genera, is ignored. Altogether 20 of the 32 ophionine genera (62-5%) are restricted to a single zoogeographicregion. This is a surprisingly high proportion compared with many other ichneumonid subfami-lies. For example, only about 30% of pimpline genera are endemic to a single region (Townes, 120 I. D. GAULD Table 18 Approximate numbers of species of various ophionine genera in each zoogeographic area. PRETHOPHION ORIENTOSPILUS OPHIOGASTRELLA ABANCHOG ASTRABANCHOGASTRA M O O H > M ALOPHOPHION SCLEROPHION AFROPHION AGATHOPHIONA RHOPALOPHION XYLOPHION RIEKOPHION SICOPHION ~ ~ " 2 TiMyflPHTDN 1 EREMOTYLUS 8 ~ 1 5 20 TROPHOPHION HELLWIGIELLA 1 BARYTATOCEPHALUS 2 EURYOPHION ~ 7 1 DICTYONOTUS 1 2 2 RHYNCHOPHION ~ ~ 3 1 40 3 LEPISCELUS ~ 1 STAUROPOCTONUS 1 1 2 1 1 4 PAMOPHION -- -- -- 1 LATICOLEUS 10 LEPTOPHION 11 17 7 DICAMPTUS 13 12 1 5 ENICOSPILUS 50 150 180 150 50 2 20 100 PYrNflPHTflN 3 1969) and a similar proportion of genera of other subfamilies are likewise restricted. Themajority of the restricted ophionine genera are small taxa with three or fewer species; onlyAlophophion, Ophiogastrella and Laticoleus are larger than this and they have ten or morespecies each. Six of the more widely distributed genera occur in only two regions: Rhynchophion andThyreodon (Neotropical/Nearctic), Euryophion and Orientospilus (Afrotropical/Oriental),Leptophion (Oriental/ Australo-Pacific) and Simophion (Palaearctic/Nearctic). Two generaoccur in just three regions, Dictyonotus (Palaearctic/Oriental/Afrotropical) and Dicamptus(Afrotropical/Oriental/Australo-Pacific), whilst Eremotylus occurs in four regions (Nearctic/Palaearctic/Oriental/Neotropic). Ophion occurs in all regions except the Afrotropical andStauropoctonus in all but the Nearctic (though it is absent from the African mainland). OnlyEnicospilus is truly cosmopolitan with quite large numbers of species in all areas. PHYLOGENY OF THE OPHIONINAE 121 Considered from a geographical standpoint the Neotropical region contains both the mostgenera (12) and has the highest degree of generic endemicity (50%). The Afrotropical andOriental regions both contain ten genera, but the former has a far higher degree of endemicity(40% compared with 10%). The Australo-Pacific region contains eight genera, three of which(37%) are endemic whilst the Palaearctic, with a similar number of genera, has only twoendemics (25%). The Nearctic is the most impoverished with seven genera, one of which (14%)is endemic. The faunal similarity (at generic level) between the regions is shown in Table 19. Table 19 The generic faunal affinities between major zoogeographic regions (Hawaii is excluded). Thevalues at the intersects of rows and columns V may be defined as: , _ number of genera common to areas X and Ytotal number of genera present in areas X and Y It is noteworthy that contiguous or recently contiguous areas have a significantly higher value for V thando areas that are not contiguous or that have only been contiguous in the distant past (more than 40million years ago), suggesting that dispersal may be a more important event than vicariance in thebiogeographical history of the subfamily. o r 1 o f r 1 PALAEARCTIC 0.36 0.25 0.23 0.38 0.20 AFROTROPICAL 0.06 0.10 0.20 0.43 ORIENTAL 0.21 0.22 0.38 AUSTRALO-PACIFIC 0.15 0.18NEOTROPICAL 0.36 Hypothesized evolutionary history of the group: a scenario Primitive ichneumonids are known from the Upper Cretaceous (80-90mya) (Townes, 1973ft)and the family may have originated at the beginning of the Cretaceous (Rodendorf & Rasnitsyn,1980), possibly from a protoichneumonoid ancestor such as the Praeichneumonidae (Rasnitsyn,1983). The age of the subfamily Ophioninae is not known but it is presumed that it post-dates theprimitive Upper Cretaceous groups. A fossil Ophion species is known from from the LowerOligocene (35-40mya) in France and Rodendorf (1962) states that the Ophioninae (?sensu lato,i.e. including the Campopleginae) dates back to the Palaeocene (55-60mya). Possibly the groupradiated around the beginning of the Tertiary some 65-70 million years ago. At this time the continents were still in close proximity to one another ( Audley-Charles et al. ,1981; Owen, 1981). Australia was connected to Antarctica and an archipelagic connectionprobably existed between this landmass and South America. Europe and North America werecontiguous and South America was not widely separated from Africa. The most primitivelineage of the Ophioninae (the Ophion genus-group) may have been widespread at this time, aswould have been its sister-group, the stem group of the remaining Ophioninae. The Sicophionlineage became isolated in the southern continent of Australia/ Antarctica/South America at anearly stage whilst its sister-lineage, the ancestor of the Eremotylus, Enicospilus and Thyreodonlineages, remained in Laurasia or Africa. The ancestor of the Eremotylus lineage probablydifferentiated in Laurasia whilst the ancestor of the Enicospilus I Thyreodon radiated in Africa. Isuggest that early offshoots of this line must have spread to South America, either by flying 122 I. D. GAULD across the widening South Atlantic, or by dispersal through Laurasia and across a water gap toSouth America. Some kind of barrier to the spread of organisms to and from South America issuggested by the high degree of generic endemicity and the fact that the Neotropical species ofEnicospilus seem to belong to very few species-groups compared with those of other tropicalareas. I suggest that the less vagile Thyreodon lineage reached South America much later,possibly from Asia via the Bering Straits. Probably the radiation of the Enicospilus lineageresulted in widespread extinction of members of the Ophion group, thus leaving the isolatedspecialized groups extant today. Ophion almost certainly has undergone secondary radiationsgiving rise to species complexes in South America (derived from the Nearctic) and Australia(derived initially from the Palaearctic fauna spreading through the Oriental region, where it hassubsequently become isolated on mountain tops) (Gauld, 19840). Except for Xylophion, aremnant of the early widespread Ophion group-distribution, and Riekophion, a remnant of thesouthern radiation of the Sicophion lineage, the ophionine fauna of the Australo-Pacific haslargely been derived from the Oriental region (as has been observed for other groups, e.g.Wilson, 1959; Gupta, 1962), though New Guinea has served as the epicentre of a considerableradiation, particularly in the case of Enicospilus , over 100 endemic species of which occur on theisland. Although the above scenario is largely speculative it is the most parsimonious interpretationof the present distribution in relation to the postulated phylogenetic history of the group. Otherscenarios would involve postulation of more widespread extinctions, more transoceanic disper-sive events or necessitate postulating a much earlier origin for the subfamily. The subfamily Ophioninae In the following section the various genera and new synonymies are discussed in some detail.Keys have not been given here to genera as any attempt to produce a key to world genera wouldinvolve the usage of rather difficult couplets to allow for evolutionary parallelism betweendifferent species-groups in similar habitats in different zoogeographic regions. Practical keys togenera on a regional basis already exist. The Afrotropical region is covered by Gauld & Mitchell(1978) and the Indo- Australian region by Gauld & Mitchell (1981). This latter key will alsosuffice for the Palaearctic region. A key to the Neotropical genera is currently in preparation andthis will suffice for the Nearctic. Townes (1971) also offers reasonable generic keys, but these arenow rather dated, particularly his key to the 'Ophionini' (see Gauld, 1979). The OPHION genus-group This group contains seven genera, Afrophion, Agathophiona, Alophophion, Xylophion, Scler-ophion, Rhopalophion and Ophion. The first six are holophyletic groups but Ophion is probablyparaphyletic with respect to them (see p. 125). The group is characterized by the absence of a posterior mesosternal transverse carina,usually by possession of a diagonal mesopleural furrow, generally by possession of an impressedclypeal apex and by having Rs in the hind wing usually at least slightly curved and often verystrongly bowed. Virtually all species in this genus-group have Im-cu in the fore wing centrallyangled somewhat, and usually have a distinct ramellus present at this point. As Ophion is apparently the paraphyletic stem-group from within which all other genera inthis group have arisen, and the genus is primarily a Holarctic taxon, it seems probable that thisgroup originated in the temperate north. It is probable that at some period it was present in mostregions and has gradually disappeared from equatorial regions leaving isolated relicts in SouthAfrica (Afrophion), Australia (Xylophion), Madagascar (Rhopalophion} and Patagonia(Alophophion). The possibility that there has been repeated expansion into and extinctionwithin the tropics is suggested by the presence of some groups of Ophion species on isolatedmountains in South East Asia, New Guinea and South America, and by the occurrence ofdistinctive Ophion species-complexes in Australia and New Zealand (Gauld, 19840). PHYLOGENY OF THE OPHIONINAE 123 AFROPHION Gauld Afrophion Gauld, 1979: 79. Type-species: Ophion nubilicarpus Tosquinet, by original designation. Mandibles stout, not twisted, subequally bidentate, barely narrowed apically; outer mandibular surfaceconvex, with strong proximal concavity. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus inprofile weakly convex with apical margin sharp but not clearly impressed; clypeus in anterior aspect weaklyconvex. Ocelli large, the posterior ones very close to the eyes; frontal carina absent; occipital carinacomplete, ventrally joining hypostomal carina. Antennae moderately long, 1-3 times length of fore wing.Pronotum unspecialized; spiracular sclerite exposed; notauli vestigial; epicnemial carina complete;mesopleural furrow strong, diagonal. Scutellum moderately convex, quite narrow and not carinatelaterally; posterior transverse carina of mesosternum absent except laterally. Propodeum with anteriorarea either occluded or extremly short, the remainder of the propodeum rather abruptly declivous, withirregular vestiges of carinae and rather coarsely microreticulate. Fore wing with pterostigma broad;marginal cell elongate ; Rs+2r very broad and slightly angled before joining pterostigma; anterior corner ofdiscosubmarginal cell glabrous, in one species extensively so and with a detached alar sclerite (Gauld &Mitchell, 1978); Im-cu with small ramellus, centrally angled. Hind wing with Rs from straight to weaklycurved. Fore tibial spur with a membranous flange behind macrotrichial comb; mid and hind trochantelliunspecialized; inner hind tibial spur flattened, with a margin of close fine hair; hind tarsal clawsunspecialized, those of male slightly more closely pectinate than those of the female. Gaster slender;tergite 2 elongate, with thyridia very weak, close to anterior margin; umbo distinct; epipleuron up- turned.Ovipositor sheath slender; male with gonosquamae very large, ploughshare-like. Afrophion is restricted to the extreme south of Africa. The genus contains two species, the type-speciesand A. hynnis (Gauld & Mitchell). It is distinguished from Ophion by the form of the propodeum, theunique structure of the male genitalia and by the basally incrassate Rs+2r in the fore wing. AGATHOPHIONA WestwoodAgathophiona Westwood, 1882: 19. Type-species: Agathophiona fulvicornis Westwood, by monotypy. Mandibles twisted 5-10, barely tapered, with lower tooth slightly the longer; outer mandibular surfaceslightly convex, with a strong proximal concavity. Maxillary palp 5-segmented, labial palp 4-segmented;mouthparts exceptionally specialized in that the distal part of the maxilla projects below the clypeus andthe labial glossae project by a distance almost equal to length of hind wing; clypeus in profile rather flat,margin blunt; clypeus in anterior aspect broad, very weakly convex. Ocelli quite small, the posterior onesseparated from eye by about their own minimum diameter; frontal carina absent or present but weak;occipital carina dorsally complete or narrowly obsolescent centrally, ventrally obsolescent, not joining thevery weak hypostomal carina. Antennae rather short and quite stout, distal segments quadrate. Pronotumunspecialized; spiracular sclerite exposed; notauli short but strongly impressed near anterior margin;epicnemial carina more or less complete; mesopleural furrow virtually absent. Scutellum moderatelystrongly convex, not laterally carinate; posterior transverse carina of mesosternum present only laterally asvestiges. Propodeum with anterior area occluded, rather deeply impressed; propodeal carinae vestigial;posterior area striate or coriaceous. Fore wing with pterostigma moderately broad; marginal cell slender;Rs+2r more or less straight, not thickened before joining pterostigma; discosubmarginal cell without aglabrous anterior area; Im-cu generally with only a trace of a ramellus, rather evenly but quite stronglycurved. Hind wing with Rs quite strongly curved. Fore leg with tibial spur bearing a membranous flangebehind the macrotrichial comb; mid and hind trochantelli unspecialized; inner hind tibial spur cylindrical,with scattered hairs marginally; hind tarsal claws long and weakly curved. Gaster rather stout; tergite 2 inprofile only slightly longer than posteriorly deep; thyridia small and close to anterior margin; epipleuronup-turned. Female with subgenital plate enlarged, as long as tergite 2 and medially notched; ovipositorsheath unspecialized. Agathophiona is a monobasic genus; the type-species occurs in Mexico. A. fulvicornis is diurnally active(Townes, 1971). Structurally this is one of the most specialized of all ophionine genera. The remarkablemouthparts are presumably an adptation to feeding from certain types of flowers but the function of therather unusually modified female terminalia is not known. 124 I. D. GAULD ALOPHOPHION CushmanAlophophion Cushman, 1947: 439. Type-species: Ophion chilensis Spinola, by original designation. Mandibles not twisted, weakly narrowed apically, subequally bidentate; outer mandibular surface flat,usually punctate and hirsute. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profileweakly convex, margin often impressed, sometimes very narrowly so; clypeus in anterior aspect weaklyconvex. Ocelli generally large, the posterior ones close to the eyes; frontal carina absent; occipital carinaabsent. Antennae moderately to very long, at least 1-3 times length of fore wing. Pronotum unspecializedor mediodorsally somewhat flattened and quite long; spiracular sclerite exposed; notauli present onanterior part of mesoscutum; epicnemial carina generally strong; mesopleural furrow distinct, diagonal,extending from episternal scrobe to near subalar prominence. Scutellum very weakly convex, usuallynarrow and not carinate laterally; posterior transverse carina of mesosternum absent except for lateralvestiges. Propodeum with anterior area occluded, transverse and often lateromedian longitudinal carinaediscernible, often almost complete; posterior area smooth or rugulose. Fore wing with pterostigma broad;marginal cell long; Rs+2r slender, curved near proximal 0-3 before joining pterostigma near centre;dicosubmarginal cell with glabrous area anterior; Ira-cu generally centrally angled, sometimes with a shortramellus which is directed more anteriorly than that of Ophion. Hind wing with Rs curved. Fore tibial spurwith a membranous flange behind macrotrichial comb; mid and hind trochantelli unspecialized; inner hindtibial spur flattened, with a margin of long close hairs; hind tarsal claws unspecialized. Gaster moderatelyslender; tergite 2 in profile elongate, thyridia oval, separated from anterior margin of tergite by its ownlength or less, umbo distinct; epipleuron up-turned. Ovipositor sheath narrow. Alophophion is a moderately large genus that is restricted to South America. The majority of speciesoccur in southern Chile and Patagonia. Cushman, when describing Alophophion as a distinct genus,commented that it barely warranted generic distinction from Ophion. Whilst it is undoubtedly very close toOphion, the combination of characters exhibited by the group clearly separate it as a holophyletic lineage.This lineage is characterized by the following apomorphies: occipital carina entirely absent; Rs+2r joiningpterostigma near centre; first subdiscal cell stouter than normal; ramellus, when present, directed moreanteriorly than that of other ophionines. OPH/OJVFabricius Ophion Fabricius, 1798: 210, 235. Type-species: Ichneumon luteus L., by subsequent designation, Curtis, 1836: 600. Paniscus Schrank, 1802: 316. Type-species: Ichneumon luteus L., by monotypy.Psylonychia Szepligeti, 1905: 21. [Nomen nudum.]Stenophthalmus Szepligeti, 1905: 23. Type-species: Stenophthalmus algiricus Szepligeti, by subsequent designation, Viereck, 1914: 137. [Homonym of Stenophthalmus Becker, 1903.]Pachyprotoma Kohl, 1906: 223. Type-species: Ophion (Pachyprotoma) capitatus Kohl, by monotypy.Australophion Morley, 1912: 4, 30. Type-species: Ophion peregrinus Smith, by monotypy.Neophion Morley, 1912: 4, 30. Type-species: Neophion crassus Morley, by subsequent designation, Viereck, 1914: 100. Apatophion Shestakov, 1926: 262. Type-species: Apatophion mirsa Shestakov, by original designation.Platophion Hellen, 1926: 13. Type-species: Platophion areolaris Brauns, by subsequent designation, Cushman, 1947: 475. Potophion Cushman, 1947: 476. Type-species: Potophion caudatus Cushman, by original designation.Psylonychia Cushman, 1947: 476. [Unavailable name, proposed in synonymy.]Apomesus Townes, 1971: 54. Type-species: Apomesus longiceps Townes, by original designation.Mecetron Townes, 1971: 60. Type-species: Stenophthalmus choaspese Uchida, by original designation. Mandibles not twisted, from very weakly to moderately narrowed distally, generally subequally bidentateor with upper tooth slightly the longer; outer mandibular surface more or less flat, except for basalconcavity, moderately punctate. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profileweakly to moderately convex, margin impressed, acute; clypeus in anterior aspect weakly convex, rarelytruncate or even slightly concave. Ocelli usually large, the posterior ones separated from eyes by less thantheir own diameters; frontal carina absent; occipital carina usually complete dorsally, ventrally notreaching the hypostomal carina, rarely with occipital carina mediodorsally obsolescent, or in a few specieswith it entirely absent. Antennae generally of moderate length, in a few deserticolous species the flagellumshort and with central segments quadrate. Pronotum unspecialized; spiracular sclerite usually completelyexposed; notauli weak but discernible on anterior 0-2 of mesoscutum; epicnemial carina generally strongand well developed on mesopleuron; mesopleural furrow strongly impressed, extending diagonally from PHYLOGENY OF THE OPHIONINAE 125 episternal scrobe towards subalar prominence. Scutellum weakly to moderately convex, usually carinateonly on anterior 0-2 or less, rarely with lateral longitudinal carinae complete to posterior margin; posteriortransverse carina of mesosternum usually present only laterally as vestiges. Propodeum with anterior areaoccluded except for a small semicircular depression centrally; propodeal carina variously developed, atmost with both transverse carinae and the lateromedian longitudinal carinae complete, enclosing an areasuperomedia, in the most exceptional cases with all carinae only vestigial; posterior area usually rathersmooth. Fore wing with pterostigma moderately to very stout; marginal cell very long and slender; Rs+2rusually virtually straight, slender, barely broadened before joining pterostigma, in some species evenlybroadened to join pterostigma; discosubmarginal cell with glabrous area in anterior corner; Im-cu usuallywith a well-developed ramellus, rarely with stub-like indication of this vein present; Im-cu centrallystrongly geniculate. Hind wing with Rs from weakly to very strongly curved. Fore tibial spur withmembranous flange behind the macrotrichial comb; mid and hind trochantelli unspecialized; inner hindtibial spur flattened, with a margin of close long hairs; hind tarsal claws unspecialized, those of malesusually more closely pectinate than those of the female. Gaster moderately slender; tergite 2 in profilemoderately to strongly elongate, usually with thyridia large, close to anterior margin; umbo welldeveloped; epipleuron up-turned. Ovipositor sheath usually slender, in one species-group exceptionallystout. Ophion is a very large genus that is widely distributed throughout the world except for the Afrotropicalregion. In tropical South America and South East Asia few species occur and these are restricted to highaltitudes (Gauld & Mitchell, 1981). Ophion as currently defined is almost certainly a paraphyleticassemblage; Agathophiona, Rhopalophion, Sclerophion, Xylophion, Alophophion and Afrophion arephenetically highly divergent (Gauld, 1979), holophyletic species-groups which have probably arisen fromwithin Ophion. A number of other phenetically less divergent species-groups were included by Gauld(1979) within Ophion, but as the phylogenetic relationship of the taxa becomes better understood it will benecessary to erect a number of additional genera in order to establish a system of holophyletic taxa.Previous authors (e.g. Morley, 1912; Cushman, 1947; Townes, 1971) have attempted to remove otherspecies-groups from Ophion and treat these as separate genera, but in each case the group removed has notbeen holophyletic as comparatively few of the subtropical and eremic Ophion species are known. Recentcollecting has greatly increased our knowledge of these insects but more work needs to be undertakenbefore it is possible to fully resolve this genus. The major species-groups of Ophion are outlined below; allexcept the unsatisfactorily resolved luteus-group are holophyletic but I have avoided treating these groupsas distinct genera until the luteus anathema can be resolved. As each of these groups appears to be definedby a series of unique apomorphies it is not possible to suggest their phylogenetic inter-relationship. Ophion peregrinus species-group. This species-group contains all the described New Zealand Ophionspecies (Townes et al, 1961). Its monophyly is suggested by the possession of a mediodorsallylengthened pronotum, having Rs+2r joining the pterostigma distal to the pterostigma base and havingthe longitudinal lateral carina of the propodeum with a raised ridge diverging towards the propodealspiracle. In O. peregrinus this is only weakly developed but in other taxa this ridge is frequently verystrong and continuous with the carina, and the front part of the carina is absent so the apparent carinacurves anteriorly towards the spiracle where it terminates. O. peregrinus is apparently the mostprimitive taxon as it retains most of the propodeal carinae and has a strongly geniculate Im-cu; theother taxa have only a vestige of the anterior carina (the lateromedian and posterior transverse carinaare lacking) and frequently have a reduced ramellus and more evenly curved Im-cu. Most also have a- fairly straight Rs in the hind wing. O. bicarinatus species-group. This complex contains O. bicarinatus Cameron, O. facetious Gauld &Mitchell, O. gerdius Gauld & Mitchell, O. horus Gauld & Mitchell and all of the described Australianspecies (Gauld, 1977). The group is characterized by possession of a proximally broadened Rs+2rwhich is slightly curved before reaching the pterostigma. The European species, O. minutus, maybelong to this group. The five Australian species form a monophyletic subgroup characterized by theblack interocellar area. In the Oriental region the species of the bicarinatus group are restricted tomountains in Sri Lanka, India, Burma, Malaysia, Taiwan and Sumatra. Their present fragmenteddistribution and their paraphyletic nature with respect to the Australian species-group suggests thatthis group was more widespread throughout the Indo- Australian region in the past, probably when theclimate was cooler (Gauld, 1984a). O. cronus species-group. This monobasic group is restricted to montane New Guinea (Gauld & Mitchell,1981). The group is characterized by having a very sparsely hirsute discosubmarginal cell and a veryshort 3r-m. It does not appear to be closely related to any other Indo- Australian species. 126 I. D. GAULD O. caudatus species-group. This group is characterized by the elongate head shape; the eyes are more ovalthan normal and the labium is specialized in having the prementum extended far beyond the insertionof the labial palps (Gauld & Mitchell, 1981). This group contains O. caudatus (Cushman), O. silusGauld & Mitchell, O. longiceps (Townes), O. ascus Gauld & Mitchell, O. sumptions Gauld &Mitchell, O. mastrus Gauld & Mitchell and an undescribed species from Sulawesi (BMNH). All thesespecies occur in mountains in the Oriental region. An undescribed species from high altitude in Peru(TC) has a similarly modified head and like other members of the O. caudatus group it has large wings,elongate trochantelli and virtually obscured mesopleural punctures. This Neotropical species clearlyseems to belong in this species-group, but it is possible that the apomorphies defining the group are asuite of characters facilitating existence at high altitude and that the Neotropical and Oriental lineageshave undergone morphological convergence. O. areo/ar/s species-group. This group is characterized by the loss of occipital carina, possession of aquadrate scutellum and a somewhat broadened ovipositor sheath, and by having a characteristicpattern of propodeal carina in which the area superomedia is more or less discernible, the posteriortransverse carina is often complete and the anterior transverse carina is absent except centrally. Thisgroup contains O. areolaris Brauns, O. ocellaris Ulbricht and O. fuscomaculatus Cameron. The firsttwo are western Palaearctic species whilst O. fuscomaculatus has an eastern Palaearctic range thatextends into the higher mountains of the Oriental region. O. similis species-group. This group contains the Palaearctic species O. similis (Szepligeti), O. mirsa(Shestakov), O. buchariensis Meyer and two undescribed Nearctic species. All are stout insects withshort, compact gasters, rather convex, irregularly sculptured propodea and somewhat shorterantennae than is usual for species of this genus. The ocelli and eyes are moderately small so theorbital-ocellar distance and the malar space are broader than is normal for Ophion species. All areapparently diurnally active and most are associated with dry areas. O. dentatus species-group. This group contains O. dentatus Smith, O. turcomanicus Szepligeti and O. virusGauld & Mitchell. The group is characterized by possession of unusually long, fairly slendermandibles, having long, weakly curved claws and possessing numerous spines on hind tarsal segments1-3. Species of this group are widely distributed throughout the eastern Palaearctic region, particu-larly Central Asia. Several also extend into the drier parts of the Indian subcontinent. O. luteus species-group. This large group contains all of the other described species of the genus from thePalaearctic, Nearctic and Neotropical regions. It is apparently a paraphyletic group and I can onlydefine it in terms of plesiomorphic features (i.e. absence of apomorphic characters exhibited by otherspecies-groups). With more study it may be possible to subdivide this group into several holophyleticspecies-complexes but little is yet known about either the eastern Palaearctic or the Nearctic species. RHOPALOPHION Seyrig Rhopalophion Seyrig, 1935: 49. Type-species: Rhopalophion curvus Seyrig (= Ophion discinervusMorley), by original designation. Mandibles not twisted, weakly narrowed apically, more or less equally bidentate; outer mandibular surfaceslightly convex with distinct proximal concavity. Maxillary palp 5-segmented, labial palp 4-segmented;clypeus in profile weakly convex, margin slightly impressed and reflexed or sometimes almost blunt;clypeus in anterior aspect weakly convex or truncate. Ocelli moderately small to large, the posterior onesseparated from the eyes by less than 0-8 times their own maximum diameter; frontal carina absent;occipital carina complete, joining hypostomal carina close to base of mandible. Antennae of moderatelength, at least 1-2 times as long as fore wing. Pronotum unspecialized; spiracular sclerite exposed; notaulivirtually absent; epicnemial carina present, extending onto pleuron; mesopleural furrow absent or veryindistinct. Scutellum weakly to moderately convex, laterally carinate for its entire length; posteriortransverse carina of mesosternum absent except for lateral vestiges. Propodeum with anterior areaoccluded; anterior and posterior transverse carinae usually complete, rarely with the former vestigial, veryrarely all carinae absent; longitudinal carinae not developed; posterior area finely alutaceous to smooth.Fore wing with pterostigma moderately broad; marginal cell long; Rs+2r almost straight, not or only veryslightly expanded before joining pterostigma; discosubmarginal cell anteriorly broadly glabrous from baseof Rs+2r to ramellus; Im-cu very angulate, with an extremely long ramellus that reaches at least 0-6 ofdistance to Rs&M. Hind wing with Rs virtually straight or very weakly curved. Fore tibial spur withmembranous flange behind macrotrichial comb reaching 0-5-0-7 of length of spur; mid and hindtrochantelli unspecialized; inner hind tibial spur moderately flattened, with a fringe of long close hairs; PHYLOGENY OF THE OPHIONINAE 127 hind tarsal claws unspecialized, those of male far more closely and finely pectinate than those of the female.Gaster slender; tergite 2 in profile moderately long with thyridia oval and close to anterior margin; umboquite well developed; epipleuron up-turned. Ovipositor sheath slender. Rhopalophion is a small genus with three described species, discinervus which is widespread throughoutthe Afrotropical mainland, and divergens and parallelus which are restricted to Madagascar (Delobel,1975). R. divergens and R. parallelus are apparently sister-species and constitute a clade defined by thefollowing apomorphic features, possession of small ocelli, possession of subquadrate lower face andpossession of a relatively long malar space. This clade is the sister- lineage to R. discinervus which is definedby possession of a very angulate Im-cu and a distally swollen ramellus (Gauld & Mitchell, 1978). Thefemales of discinervus are unusual amongst ophionines in that the distal tarsal segment of the female has alateral projection. This is similar to one found in a species of Ophiogastrella (Cushman, 1947). Rhopa-lophion species are easily distinguished from other taxa in the Ophion genus-group on account of theirunique venation, total lack of propodeal longitudinal carinae and long fore tibial spur. SCLEROPHION Gauld Sclerophion Gauld, 1979: 77. Type-species: Pleuroneurophion longicornis Uchida, by original desig-nation. Mandibles stout, not twisted, barely narrowed apically; outer mandibular surface weakly convex.Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profile weakly convex, margin notimpressed; clypeus in anterior aspect weakly convex. Ocelli large, the posterior ones separated from eyesby less than 0-2 of their minimum diameter; frontal carina absent; occipital carina complete, ventrallyreaching hypostomal carina. Antennae very long and slender, more than 1-7 times length of fore wing.Pronotum unspecialized; spiracular sclerite exposed; notauli weak but discernible; epicnemial carinacomplete, strong; mesopleural furrow weak, diagonal, with upper end rather broad and shallow. Scutellummoderately convex, carinate laterally on at least its anterior 0-4; posterior transverse carina of meso-sternum absent except as lateral vestige. Propodeum with anterior area occluded except medially where itis semicircular; anterior and posterior transverse carinae usually complete, other carinae weak; posteriorarea usually rather smooth. Fore wing with pterostigma moderately slender; marginal cell elongate; Rs+2rthickened and evenly curved before joining pterostigma; discosubmarginal cell with an extensive glabrousarea anteriorly, the distal margin of this area with a thickened corneous mark; Im-cu with long ramellus,centrally geniculate. Hind wing with Rs very weakly curved. Fore tibial spur with a membranous flangebehind macro trichial comb; mid and hind trochantelli unspecialized; inner hind tibial spur flattened, withmargin of close, long hairs; hind tarsal claws unspecialized. Gaster slender; tergite 2 very elongate, withthyridia elliptical, close to anterior margin; umbo well developed; epipleuron up-turned. Ovipositorsheath slender. Sclerophion species may be recognized by their characteristic fore wing venation, particularly thethickened Rs+2r and the corneous mark in the discosubmarginal cell. The genus contains two species, thetype-species and S. uchidai Gauld & Mitchell. These occur in the mountains at the eastern Palaearctic/Oriental interface (Gauld & Mitchell, 1981). XYLOPHION GauldXylophion Gauld, 1979: 77. Type-species: Ophion xylus Gauld, by original designation. Mandibles not twisted, weakly to moderately tapered, subequally bi dentate; outer mandibular surface flat,punctate, with proximal concavity. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profileconvex, margin impressed, acute; clypeus in anterior aspect weakly convex. Ocelli large, the posterior onesseparated from eyes by their own minimum diameter or less; frontal carina absent; occipital carina strong,ventrally often reaching hypostomal carina, sometimes ventrally obsolescent. Antennae of moderatelength or long, usually at least 1-5 times length of fore wing. Pronotum more or less unspecialized, thoughone species has trace of a median transverse crest; spiracular sclerite exposed; notauli quite weak butusually distinct on anterior 0-2 of segment, rarely absent; epicnemial carina strong, reaching onto pleuronabove level of lower corner of pronotum ; mesopleural furrow weak , diagonal . Scutellum weakly convex , atmost carinate laterally on anterior 0-2; posterior transverse carina of mesosternum present only laterally asvestiges. Propodeum with anterior area more or less occluded except centrally where it is discernible as asemicircular area; anterior transverse carina usually more or less complete, posterior transverse carinafrom present laterally as vestiges to almost complete; longitudinal carinae vestigial; posterior area rathersmooth. Fore wing with pterostigma large and triangular; marginal cell long; Rs+2r evenly broadened andcurved before joining pterostigma; discosubmarginal cell with a large glabrous area in anterior corner; 128 I. D. GAULD lra-cu with a short ramellus, centrally angulate. Hind wing with Rs strongly curved. Fore tibia with amembranous flange extending from 0-1-0-3 of its length behind macrotrichial comb; mid and hindtrochantelli unspecialized; inner hind tibial spur flattened, with a margin of long, close hairs; hind tarsalclaws of female unspecialized, of male unique in having a central gap between an inner and outer row ofpectinal teeth. Gaster slender; tergite 2 in profile very elongate with thyridia elliptical, separated fromanterior margin by its own length or less; umbo strongly developed; epipleuron up-turned. Ovipositorsheath slender. Xylophion is a small genus restricted to the Australian tectonic plate. Three species, one of which isundescribed, occur in south-eastern Australia. One of these, X. xylus, also occurs more widely throughoutAustralia and has been found in the New Guinea highlands (Gauld, 1984a). Xylophion is characterized by the venation, particularly the possession of a very short 3r-w, and thereduced flange on the fore tibial spur. A particularly striking, and hitherto unrecognized, autapomorphy ofthe genus is the form of the male tarsal claw. Instead of having a sinuous row of pectinal teeth present (as isnormal for ophionines), Xylophion males have a marked discontinuity in the centre of the row between theteeth on the inner and outer edges of the claw. In its least derived form only the central gap exists(Xylophion species 1, BMNH) but the other two species are specialized further in different ways. X. ketushas the inner and outer marginal rows extended laterally so that there is a central overlap between the twoparallel ends of the rows of teeth. In X. xylus there is a less pronounced overlap but the distal portion of theclaw is flattened and the terminal tooth reduced so the pectinal row forms a 'fence' around the distal end ofthe claw. The SICOPHION genus-group This group comprises three genera, Sicophion and Janzophion from montane tropical SouthAmerica and Riekophion from Australia. The species in this complex exhibit an unusualcombination of plesiomorphic and apomorphic features. All possess a well-developed membra-nous flange on the fore tibial spur though none has an umbo on tergite 2 and often the uppercorner of the pronotum is expanded to partially occlude the spiracular sclerite. It is suggestedhere that these three genera comprise a distinct lineage which is derived with respect to theOphion group, and that the Sicophion group represents a separate southern radiation of theophionines. However, as mentioned above (see p. 117), if the polarity of characters 38 and 47-1has been misinterpreted then the Ophion lineage may represent a more derived group than theSicophion lineage. If this were the case then it is possible that the Sicophion group is a collectionof relict genera whose present southern distribution can be explained by Darlington's (1965)'glove hypothesis'. JANZOPHION gen. n. Type-species: Janzophion nebosus sp. n. Mandibles twisted perhaps 5, slightly tapered distally, with upper tooth slightly the longer; outermandibular surface with a diagonal line of hair extending from upper proximal corner to near centre.Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profile convex, margin impressed, inanterior aspect with margin weakly convex. Ocelli very large, the hind ones more or less contiguous witheyes; frontal carina absent; occipital carina absent. Antennae very long and slender, about 1-5 times lengthof fore wing. Pronotum unspecialized; spiracular sclerite partially concealed; notauli vestigial; epicnemialcarina present laterally and ventrally; mesopleural furrow indistinct. Scutellum weakly convex, laterallycarinate only on anterior 0-1-0-3; posterior transverse carina of mesosternum complete. Propodeum withanterior area long; anterior transverse carina and usually posterior transverse carina complete, often with amedian longitudinal carina present; propodeum otherwise rather smooth. Fore wing with pterostigmamoderately broad; marginal cell very long; Rs+2r bowed, thickened before joining pterostigma; discosub-marginal cell with glabrous area near anterior corner, but anterior to this is narrow hirsute region; \m-cuwithout a ramellus, proximally fairly straight, distally strongly bowed. Hind wing with Rs weakly curved;marginal cell proximally glabrous; penultimate distal hamulus longer than its fellows, the distal one slightlyshorter but longer than the proximal ones. Fore tibial spur with a membranous flange behind themacrotrichial comb; mid and hind trochantelli unspecialized; inner hind tibial spur flattened, with marginof close long hairs; hind tarsal claw unspecialized. Gaster slender; tergite 2 in profile very elongate, withthyridia remote from anterior margin; umbo vestigial; epipleuron up-turned. Ovipositor sheath slender. PHYLOGENY OF THE OPHIONINAE 129 Janzophion is a distinctive genus which bears a very strong phenetic similarity to the Old World genusLeptophion. Unlike Leptophion, Janzophion species lack the occipital carina, and have a well-developedmembranous flange present on the fore tibial spur, behind the macrotrichial comb. Janzophion nebosus sp. n. Fore wing length 14-16 mm. Head slightly more elongate than normal for ophionines; lower face polished, 1-4-1 -5 times as long asbroad; malar space 0-50-0-55 times basal mandibular width; head strongly narrowed behind eyes, occiputmediodorsally slightly concave. Flagellum with 66-68 segments, the tenth segment about 2-0 times as longas wide. Mesoscutum with margin slightly out-turned; scutellum finely shagreened; mesopleuron withupper part highly polished, finely and sparsely punctate, ventrally slightly more coriaceous; metapleuronsimilar. Propodeum in profile evenly declivous; propodeal spiracle joined to pleural carina by weak ridge;most of alitrunk bearing fine pale pubescence. Fore wing with AI = 1-20-1-25; CI = 0-43-0-47; ICI =0-61-0-70; SDI = 1-14-1-17; cu-a proximal to base of Rs&M by about 0-3 times its own length; outer hindcorner of 2nd discal cell about 90. Hind wing with about 8 distal hamuli; NI = 3-70-4-60. Legsunspecialized; hind tarsal claws of male with slightly finer and denser pectinate comb than that of female.Gaster slender, male with subgenital plate bearing long fine pubescence; gonosquamae quite long, dorsallysomewhat membranous. Pale yellowish species, with interocellar area, part of mesoscutum, much of mesopleuron, metapleuronand part of propodeum blackish; gaster with distal part of tergite 5 and tergites 6+ infuscate. Pterostigmaand Rs+2r blackish, other veins flavous; wing hyaline, proximal angle of marginal cell infumate. REMARKS. This species has been taken at light in cloud forests between 1500 and 2350 m in Costa Rica.Nothing is known of its biology. MATERIAL EXAMINED Holotype cf , Costa Rica: Alajuela Prov; Volcan Poas N. P., xii.1982 (Janzen & Hallwachs) (BMNH). Paratypes. Costa Rica: 1 cf , same data as holotype (BMNH); 1 cf , same locality as holotype, xii.1981(Janzen & Hallwachs} (BMNH); 1 $, Monte Verde Reserve, 1500 m, ii.1980 (Mason) (TC). RIEKOPHION GauldRiekophion Gauld, 1977: 21. Type-species: Allocamptus emandibulator Morley, by original designation. Mandibles not twisted, weakly evenly tapered or distally parallel-sided, subequally bidentate; outermandibular surface flat, punctate. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profileconvex, margin usually blunt, in one species, impressed; clypeus in anterior aspect from weakly convex toslightly concave. Ocelli from large, the hind ones contiguous with eyes, to rather small, the hind onesseparated from eyes by 0-5 times their minimum diameter; frontal carina absent, or weakly present in onespecies; occipital carina usually complete, in one species ventrally obsolescent, not reaching hypostomalcarina. Antennae moderately long and relatively slender, at least 1-3 times length of fore wing. Pronotumeither unspecialized, or in two species mediodorsally long and flat; spiracular sclerite exposed, or in onespecies with slight expansion of dorsal lobe of pronotum; notauli vestigial or absent; epicnemial carinacomplete, generally curved to meet anterior margin of pleuron; mesopleural furrow very weak to quitedistinct diagonally. Scutellum weakly convex, very characteristic in being fairly narrow and barely tapered,and medially longitudinally higher than laterally; carinae if present only on anterior 0-3 or less; posteriortransverse carina of mesosternum complete. Propodeum with anterior area moderately to very long, oftenirregularly striate; propodeum usually with anterior and posterior transverse carinae complete, latero-median ones weaker but often complete behind anterior transverse carina and usually enclosing a longnarrow area superomedia. Fore wing with pterostigma moderately to very slender; marginal cell usuallylong; Rs+2r proximally sinuous, only weakly broadened and not curved abruptly to join the pterostigma,unusual in joining stigma distal to proximal end; discosubmarginal cell anteriorly glabrous, often withgroup of isolated hairs near base of Rs+2r, always bearing a detached corneous or sclerotized patch inmembrane. Im-cu weakly sinuous or fairly evenly curved. Hind wing with Rs straight or very weaklycurved. Fore tibial spur with a membranous flange behind macrotrichial comb; mid and hind trochantelliunspecialized; inner hind tibial spur flattened, the margin with close long hairs; hind tarsal claws elongate,with thyridia small, widely separated from anterior margin to which it is joined by a groove; umbo absent;epipleuron up-turned. Ovipositor sheath slender; unusual in having male subgenital plate ornamentedwith projections.Riekophion is a very distinctive genus, characterized by the quadrate scutellum, unusual venation and 130 I. D. GAULD ornamented subgenital plate (Gauld, 1977). It contains three species which occur in southern and westernAustralia. The phylogenetic inter-relationships of these species are suggested in Fig. 28. Nothing is knownof the biology of these insects. SICOPHION GauldSicophion Gauld, 1979: 71. Type-species: Sicophion pleuralis Gauld, by original designation. Mandibles stout, twisted about 25-35 and with lower tooth slightly the longer; outer mandibular surfacemore or less flat, sparsely punctate. Maxillary palp 5-segmented, labial palp 4-segmented; maxillaeelongate; clypeus in profile convex, apical margin not impressed, in anterior aspect truncate. Ocelli verylarge, the posterior ones close to margin of eye; frontal carina absent; occipital carina mediodorsallyincomplete, ventrally obsolescent, not reaching hypostomal carina. Antennae very long and slender, atleast 1-4 times length of fore wing. Pronotum unspecialized; spiracular sclerite virtually completelyoccluded by enlarged flap formed from hind corner of pronotum; notauli absent; epicnemial carina strong,extending onto mesopleural furrow; mesopleural furrow strong, extending from episternal scrobe forward.Scutellum quite convex, narrow, carinate only on anterior 0-2-0-3; posterior transverse carina ofmesosternum present only laterally as vestiges. Propodeum with anterior area occluded; anterior trans-verse carina present centrally, laterally obsolescent; posterior transverse carina present as lateral vestiges;propodeum otherwise rather smooth, posterior area not differently sculptured from spiracular area. Forewing with pterostigma broad, distally abruptly narrowed; marginal cell moderately long and slender,unique amongst ophionines in being broadest distad of centre at sinuation in distal abscissa of Rs; Rs+2rnot angled near junction with pterostigma, but somewhat broadened, somewhat angled slightly proximalto centre; discosubmarginal cell very extensively glabrous anteriorly, the glabrous area bearing indistinctcorneous areas; Im-cu without a ramellus, rather abruptly curved. Hind wing with Rs unique in theOphioninae in being slightly convex. Fore tibial spur with membranous flange behind macro trichial comb;mid and hind trochantelli very elongate, but unspecialized; inner hind tibial spur strongly flattened with afringe of long, close hairs; hind tarsal claws unspecialized, markedly sexually dimorphic, those of malebeing far more closely pectinate than those of the female. Gaster exceptionally slender; tergite 2 elongate,thyridia absent; umbo absent; epipleuron very narrow, pendant. Ovipositor sheath slender, ovipositorunique in being proximally angled and without a subapical notch. Sicophion is a small Neotropical genus with a single described Bolivian species (Gauld, 1979). In theBMNH is a short series of a putative second species collected in Costa Rica by Janzen & Hallwachs. Thesediffer from S. pleuralis in having a more extensive glabrous area in the discosubmarginal cell and a lessobviously 'bent' Rs+2r. There are subtle differences in colour and head shape also. Sicophion is one of the most distinctive of ophionine genera. The sinuous Rs in the fore wing, the slightlyconvex Rs in the hind wing and the basally angulate, acutely pointed ovipositor with no subapical notch are,amongst ophionines, unique autapomorphies of the genus. Their very slender form with extensivesemi-matt black coloration is a typical feature of ophionines from higher elevations (e.g. Enicospilusruwenzorius Gauld & Mitchell) and the rather elongate head shape is also found in some upper montanespecies (e.g. Ophion longiceps (Townes), Gauld & Mitchell, 1981). The presumed loss of the dorsal notchis a feature that occurs in several genera of ichneumonids with more slender ovipositors (e.g. Parania in theAnomaloninae, Gauld, 1976). Sicophion species are only known to occur at mid and high elevation in the Neotropical region.Specimens have been taken between 1600 and 3000 m. Nothing is known of their host ranges. The EREMOTYLUS genus-group This group contains three genera, Eremotylus, Trophophion and Hellwigiella. The last two aremonobasic and probably really represent single phenetically highly divergent species of Ere-motylus. The majority of species of all genera are eremic organisms. The Eremotylus genus-group is characterized by the fore wing venation; Rs+2r is stronglygeniculate and thickened before joining the pterostigma whilst \m-cu is usually fairly evenlyarcuate. The clypeus is usually flat or out-flared and the margin is not subapically impressed. EREMOTYLUS Foerster Eremotylus Foerster, 1869: 150. Type-species: Ophion marginatus Gravenhorst (= Anomalon margina-tum Jurine), by subsequent monotypy, Thomson, 1888: 1193. PHYLOGENY OF THE OPHIONINAE 131 Fig. 28 Suggested phylogenetic inter- relationships of species of Riekophion. The apomorphic characterssupporting this arrangement are: 1, mandible slender; 2, ocelli small; 3, occipital carina mediodorsallybroadened; 4, body extensively black; 5, pronotal lobe expanded; 6, pronotum medio-dorsally flat andlengthened; 7, presence of hair patch in fenestra; 8, cubital index small; 9, distal sclerite lost; 10, ocelligrossly enlarged; 11, scutellum quadrate; 12, possession of alar sclerites; 13, metapleuron inflated; 14,Rs+2r sinuous; 15, Rs+2r joining pterostigma near centre; 16, male subgenital plate ornamented. Camptoneura Kriechbaumer, 1901: 23. Type-species: Ophion marginatus Gravenhorst (= Anomalonmarginatum Jurine), by subsequent designation, Viereck, 1914: 27. [Junior homonym of CamptoneuraAgassiz, 1846.] Genophion Felt, 1904: 123. Type-species: Genophion gilletti Felt (= Ophion costale Cresson), by originaldesignation. Camptoneuroides Strand, 1928: 52. [Replacement name for Camptoneura Kriechbaumer.] Clistorapha Cushman, 1947: 450. Type-species: Ophion subfuliginosus Ashmead, by original designation. Boethoneura Cushman, 1947: 450. Type-species: Boethoneura arida Cushman, by original designation. Chilophion Cushman, 1947: 450. Type-species: Ophion abnormum Felt, by original designation. Chlorophion Townes, 1971: 55. Type-species: Chlorophion vitripennis Townes, by original designation. 132 I. D. GAULD Mandibles not twisted, usually quite long, moderately strongly narrowed, usually subequally bidentate orwith upper tooth slightly the longer; outer mandibular surface flat or slightly convex, punctate. Maxillarypalp 5-segmented, labial palp 4-segmented; clypeus in profile more or less flat or even slightly out-flared,margin never impressed; clypeus in anterior aspect truncate or slightly concave. Ocelli usually moderatelylarge to large, the posterior ones separated from the eyes by less than 0-5 their own maximum diameter, orin a few species with the ocelli small and the posterior ones separated from the eyes by more than their ownmaximum diameter; frontal carina absent or rarely very faintly indicated; occipital carina complete, rarelyventrally somewhat obsolescent. Antennae moderately long, at least 1-2 times length of the fore wing.Pronotum unspecialized; spiracular sclerite exposed; notauli vestigial or absent; epicnemial carina usuallycomplete; mesopleural furrow usually rather weak but discernible, faintly indicated between episternalscrobe and subalar prominence. Scutellum weakly convex, usually carinate 0-3 to 0-8 of its length, rarelywithout carinae; posterior transverse carina of mesosternum from complete to present only laterally asvestiges. Propodeum generally with anterior area short but not occluded, in some species moderately long;anterior transverse carina usually distinct, at least centrally, sometimes complete, rarely absent; posteriortransverse carina present laterally as vestiges or absent, other carinae generally absent; posterior area fromsmooth and polished to rugulose. Fore wing with pterostigma moderately slender; marginal cell long andslender, rarely rather short; Rs+2r abruptly geniculate and thickened near pterostigma; discosubmarginalcell with a small glabrous area in anterior corner; Im-cu without a ramellus, arcuate or very weaklysinuous. Hind wing with Rs from almost straight to strongly curved. Fore tibial spur with a membranousflange behind macrotrichial comb; mid and hind trochantelli unspecialized; inner hind tibial spur generallyflattened with a marginal fringe of close hairs; hind tarsal claws usually unspecialized, sometimes slightlylonger and less evenly curved than normal. Gaster slender; tergite 2 in profile elongate with thyridiaseparated from fore margin by its own length or a little more, characteristically at the posterior end of aweak trough; umbo usually distinct, rarely weak; epipleuron up-turned. Ovipositor sheath slender. Eremotylus is a moderately large genus containing about 35 species, most of which occur in the drierregions around the Mediterranean, the Middle East, Central Asia, the south-western United States andnorthern Mexico. Isolated species have a wider distribution; E. subfuliginosus occurs in the north-easternpart of the U.S.A., whilst the morphologically very specialized species E. marginatus is not uncommonthroughout much of western Europe. E. vitripennis occurs in the drier parts of southern South America,whilst E. perdix occurs in the Indian subcontinent (Gauld & Mitchell, 1981). The Palaearctic species weremonographed recently by Horstmann (1981) who recognized eight species. The New World species areextremely poorly known. Virtually nothing is known about the host ranges of species of Eremotylus. Theonly reliable host record to hand is of one European species (E. curvinervis Kriechbaumer) which has beenreared from a species oiDryobota Lederer (Lepidoptera: Noctuidae) (Seyrig, 1926). Some authors (e.g. Cushman, 1947; Townes, 1971) divided the species of Eremotylus between a numberof small genera characterized mostly by differences in the development of carinae, particularly thetransverse mesosternal carina. Gauld (1979) pointed out that despite these differences, all species share alarge number of features and seem to comprise a natural group. Horstmann (1981) accepted this treatment.The present study has reinforced the author's opinion that this is a natural group; it is definable on the basisof several apomorphies including the characteristically modified Rs+2r, the arcuate or slightly sinuousIm-cu and the unoccluded anterior propodeal area. The characteristic clypeus and rather slender, longishmandibles are useful confirmatory characters. The present disjunct distribution of the genus suggests that at one time the range of the genus must havebeen wider. Further discussion of zoogeography is best left until the New World species are better known. HELLWIGIELLA Szepligeti Hellwigiella Szepligeti, 1905: 23. Type-species: Hellwigiella nigripennis Szepligeti, by subsequent desig-nation, Viereck, 1914: 67. Mandible twisted about 15, not appreciably narrowed, with upper tooth slightly shorter than the lower andbearing a pronounced very sharp ventral flange; outer mandibular surface punctate, with a strong proximalconcavity. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profile quite convex centrally,with margin reflexed, quite thin; clypeus in anterior aspect almost truncate. Ocelli small, the posterior onesseparated from eye margin by more than their own maximum diameter; frontal carina absent; occipitalcarina strong, ventrally joining hypostomal carina. Antennae short and stout, central segments transverse,the flagellum barely longer than the fore wing. Pronotum mediodorsally quite long, otherwise unspecial-ized; spiracular sclerite exposed; notauli absent; epicnemial carina strong, complete; mesopleural furrowvestigial. Scutellum quite convex, laterally carinate about 0-8 of its length; posterior transverse carina ofmesosternum complete or obsolescent centrally. Propodeum with anterior area quite short, impressed as a PHYLOGENY OF THE OPHIONINAE 133 deep U-shaped groove; anterior transverse carina more or less complete, other carinae indistinct, posteriorarea reticulate. Fore wing with pterostigma moderately stout; marginal cell quite long and slender; Rs+2rabruptly geniculate and slightly broadened before joining pterostigma; discosubmarginal cell with a verysmall glabrous area anteriorly; Im-cu evenly curved, usually without a ramellus. Hind wing with Rs evenlybowed. Fore tibial spur with a membranous flange behind macrotrichial comb; mid and hind trochantelliunspecialized; inner hind tibial spur almost cylindrical, with a fringe of sparse hairs; hind tarsal claw quitelong and weakly curved. Gaster moderately slender, tergite 2 about twice as long as deep with thyridiaseparated from anterior margin by about their own length; umbo distinct; epipleuron membranous,pendant in most individuals. Ovipositor sheath slender. Hellwigiella contains what is, I believe, a single Mediterranean species, though it is sculpturally andchromatically rather variable and has been divided into two (Szepligeti, 1905). Hellwigiella is morphologi-cally extremely distinctive on account of the characteristic clypeus and mandible, and the form of thepropodeum. The majority of apomorphic features characterizing the genus are obviously adaptations to adiurnal existence and most are paralleled in other genera with similar habits. It is possible that Hellwigiellais merely a specialized diurnal offshoot of Eremotylus, but if this were so then it is necessary to postulatethat the specialized invaginated condition of the anterior area of the propodeum (character 22-2) has beenderived from the partially elongated condition of Eremotylus (character 22-1). In the preceding phylo-genetic analysis it has been suggested that the apomorphic states of these characters represent alternativespecializations, so it is possible that Eremotylus and Hellwigiella have a sister-group relationship. Nothingis known of the biology of Hellwigiella. TROPHOPHION Cushman Trophophion Cushman, 1947: 447. Type-species: Trophophion tenuiceps Cushman, by original desig-nation. Mandibles barely twisted, quite stout, weakly narrowed, subequally bidentate; outer mandibular surfaceweakly convex. Maxillary palp 5-segmented, labial palp 4-segmented; maxilla and labium elongate,projecting below mandibles; clypeus in profile flat, in anterior aspect almost truncate. Ocelli small, theposterior ones separated from the eye by more than their own maximum diameter; frontal carina absent;occipital carina complete dorsally, ventrally incomplete. Antennae short, slightly clavate, barely longerthan fore wing. Pronotum unspecialized; spiracular sclerite exposed; notauli absent; epicnemial carinacomplete; mesopleural furrow indistinct. Scutellum weakly convex, without lateral carinae; posteriortransverse carina of mesosternum present only laterally as vestiges. Propodeum with anterior area shortbut not occluded; anterior transverse carina present medially, posterior one present laterally as vestiges,other carinae obsolescent; posterior area polished, punctate. Fore wing with pterostigma moderatelyslender; marginal cell moderately short; Rs+2r thickened and curved to join pterostigma; discosubmar-ginal cell with glabrous area anteriorly; Im-cu evenly curved, without a ramellus. Hind wing with Rsbowed. Fore tibial spur with a membranous flange behind macrotrichial comb; mid and hind trochantelliunspecialized; inner hind tibial spur weakly flattened, fringed with long fine hairs; tarsal claws rather longand weakly curved. Gaster moderately stout and rather more cylindrical than that of most ophionines;tergite 2 in profile less than twice as long as posteriorly deep, with thyridia separated from anterior marginby about its own length; umbo rather weak; epipleuron up-turned. Female with tergites 3-7 shallowlynotched medially, with subgenital plate large, strongly sclerotized and centrally notched; ovipositor andsheath rather stout. Only a single species of Trophophion is known. This occurs in the drier south-west of the U.S. A. Its hostis unknown. Trophophion is recognizable on account of the specializations of the head and mouthparts andthe female gaster. In other respects it is very similar to some Eremotylus species, and it is almost certainly aspecialized offshoot of this genus. However, I have hesitated to synonymize Trophophion until somedetails of its biology and host relationship are known. If these are very different from those of Eremotylusthen I suggest if be left as a separate genus even though this leaves Eremotylus as a paraphyletic taxon.Additional material from the dry parts of the U.S. A. and northern Mexico might help in elucidating thephylogenetic position of this bizarre species. The THYREODON genus-group This group contains five genera, Barytatocephalus, Euryophion, Thyreodon, RhynchophionandDictyonotus. The first of these is included in the complex on the basis of three apomorphies, thepossession of an impressed anterior propodeal area (22-2), long, weakly curved tarsal claws, and 134 I. D. GAULD absence of any transverse carinae on the propodeum. Barytatocephalus lacks the tilted mandibu-lar axis, slender pterostigma and cylindrical hind tibial spurs of other genera, and thus seems toconstitute the most primitive branch of the lineage. The remaining genera form a holophyleticgroup, and this may be divided into two holophyletic subgroups, one comprising Thyreodon,Dictyonotus and Rhynchophion, the other containing just Euryophion. The former subgroup isdefined by the possession of a pointed clypeus, a very short and stout flagellum and anexceptionally elongate propodeal spiracle. A further possible apomorphy of this lineage is thepossession of an evenly hirsute anterior part of the discosubmarginal cell (assuming thatpresence of a small glabrous area is an apomorphy of the subfamily). Euryophion has a slightlyconcave clypeus and a somewhat thickened and usually proximally curved Rs+2r. These twolineages seem to be biologically distinct; species of theThyreodon subgroup have only beenrecorded as parasites of Sphingidae (Gauld & Mitchell, 1978; 1981; Carlson, 1979), whilstEuryophion species attack Eupterotidae and Saturniidae. The close relationship between Dictyonotus, an Old World genus, and Rhynchophion andThyreodon, primarily Neotropical genera, suggests they may have had a common ancestor thatwas widely distributed throughout the Nearctic and eastern Palaearctic regions. D. purpuras-cens, which could well be rather similar to the group ancestor, is currently widely distributed inthe eastern Palaearctic, occurring as far north as 50 (Townes et al. , 1965) , so a slight extension ofits range northwards would have permitted migration across the Bering Strait. Movement fromAsia to America is postulated on the belief the group has had an Old World origin, a suggestionfavoured by the present distribution of the most primitive member of the group (Barytato-cephalus) and the sister-lineage of the Thyreodon subgroup (Euryophion). The exact relationship between the genera Thyreodon, Rhynchophion and Dictyonotus isunclear, as slight evidence in the form of a very few highly homoplastic characters can bemarshalled for placing Thyreodon as the sister-group of either Rhynchophion or Dictyonotus, or S $ w. 10 16 Fig. 29 Suggested phylogenetic inter-relationships of Dictyonotus species. The length of the lines isproportional to the number of apomorphies characterizing it. The apomorphic characters supportingthis arrangement are: 1, presence of a metapleural tubercle; 2, striation present on gena; 3, hind tarsusflattened; 4, Im-cu sinuous; 5, body densely pubescent; 6, tergite 2 posteriorly deeper than long; 7,malar space longer than basal mandibular width; 8, occipital carina mediodorsally broadened; 9, petioledepressed; 10, tergite 2 posteriorly at least as deep as long; 11, petiolar spiracles anterior to margin ofsternite; 12, presence of metanotal swelling; 13, reduction in number of hamuli; 14, development ofstrongly punctate scutellum; 15, thyridia remote from anterior margin of tergite; 16, complete posteriortransverse carina of mesosternum (? a reversal); 17, presence of tubercle on metanotal margin; 18,epipleuron 2 pendant. PHYLOGENY OF THE OPHIONINAE 135 even Rhynchophion + Dictyonotus (see p. 93). What is clear is that both Rhynchophion and,more particularly, Thyreodon, are characterized by a string of apomorphies (as is nigrocyaneus),whilst purpurascens may well have changed very little from the form of the group ancestor (Fig.29). One wonders what adaptation acquired by the Thyreodon lineage has allowed it to radiateso markedly in the Neotropics and give rise to a complex of 30 or more quite closely relatedspecies attacking sphingids, whilst in Asia, Dictyonotus, although presented with a very similardiversity of potential hosts (Rothschild & Jordan, 1903), has apparently failed to radiateappreciably. BARYTATOCEPHALUS Schulz Barycephalus Brauns, 1895: 43. Type-species: Barycephalus mocsaryi Brauns, by subsequent designation, Viereck, 1914: 19. [Homonym of Barycephalus Guenther, I860.]Barytatocephalus Schulz, 1911: 23. [Replacement name for Barycephalus Brauns.] Mandible not twisted, very weakly narrowed, with upper tooth a little stouter but no longer than the lower;outer mandibular surface with a strong proximal concavity, and with a weak trace of a diagonal groove.Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profile convex, margin impressed, acute;clypeus in anterior aspect weakly convex. Ocelli small, the posterior ones remote from the eyes; frontalcarina absent, though a small trace may be discernible in some specimens; occipital carina complete,ventrally reaching to hypostomal carina. Antennae quite short, about 1-3 times as long as fore wing, centralsegments slightly elongate. Pronotum unspecialized; spiracular sclerite exposed; notauli absent; epic-nemial carina complete, present on mesopleuron; mesopleural furrow weak but discernible as a shortdiagonal impression extending from episternal scrobe towards subalar prominence. Scutellum convex,carinate laterally only on anterior 0-4; posterior transverse carina of mesosternum complete. Propodeumconvex, with anterior area impressed as a deep groove; propodeal carinae absent; posterior area coarselypunctate. Fore wing with pterostigma moderately slender; marginal cell moderately long; Rs+2r slender,proximally almost straight; discosubmarginal cell with a small glabrous area at anterior corner, remainderof cell sparsely hirsute; \m-cu somewhat irregularly convex, without a ramellus. Hind wing with Rsvirtually straight; genus unusual in having distal abscissa of Cu\ very weak and the first abscissa of Cu\ andcu-a forming an almost straight line, oblique, so anterior corner of first submarginal cell is about 50. Foretibial spur without membranous flange behind macrotrichial comb; mid and hind trochantelli simple; innerhind tibial spur somewhat flattened (though less so than in most taxa in this subfamily), with a fringe ofmoderately long hairs; hind tarsal claws long, weakly curved. Gaster moderately slender; tergite 2 inprofile with thyridia close to anterior margin, elongate, umbo vestigial; epipleuron up-turned. Ovipositorslender. Barytatocephalus is a small genus containing a single species that is widely distributed throughout theeastern Mediterranean region and the southern U.S.S.R. Nothing is known of its biology. DICTYONOTUS Kriechbaumer Dictyonotus Kriechbaumer, 1894a: 198. Type-species: Dictyonotus melanarius Kriechbaumer (= Thyreodon purpurascens Smith), by monotypy. Aglaophion Cameron, 1903: 131. Type-species: Aglaophion flavinervis Cameron, by monotypy.Ophionopsis Tosquinet, 1903: 389. Type-species: Ophionopsis fulvipes Tosquinet (= Ophionopsis nigro- cyaneus Tosquinet), by subsequent designation, Viereck, 1914: 106. Syn. n.Hybopleurax Enderlein, 1912: 624. Type-species: Hybopleurax sumatranum Enderlein (= Thyreodon purpurascens Smith), by monotypy.Hypselogastrina Enderlein, 1918: 217. Type-species: Hypselogastrina saliina Enderlein (= Ophionopsis nigrocyaneus Tosquinet), by original designation. Syn. n.Coracophion Shestakov, 1926: 260. Type-species: Coracophion manganicolor Shestakov (= Thyreodon purpurascens Smith), by monotypy. Mandibles stout, not twisted, barely narrowed distally, equally bidentate or with lower tooth slightly thelonger; outer mandibular surface more or less flat, with scattered hairs. Maxillary palp 5-segmented, labialpalp 4-segmented; clypeus in profile flat, with margin often slightly out-flared, in anterior aspect convexwith a median obtuse point. Ocelli small, the hind ones separated from the eye margins by at least their ownminimum diameter; frontal carina present; occipital carina complete, ventrally usually not joining thehypostomal carina. Pronotum unspecialized; spiracular sclerite exposed; notauli absent; epicnemial carinamore or less complete, always present on mesopleuron; mesopleural furrow weakly impressed, horizontal, 136 I. D. GAULD extending from episternal scrobe to near upper end of epicnemial carina. Scutellum moderately convex,with longitudinal lateral carinae present only on anterior end; posterior transverse carina of mesosternumcomplete; metanotum exceptional in being produced into a blunt prominence above upper end of spiracle.Propodeum with anterior area short, impressed as a deep groove; propodeal carinae absent or with vestigesof lateromedian ones discernible; posterior area convex, rugose-reticulate. Fore wing with pterostigmaslender; marginal cell of moderate length; Rs+2r very slightly bowed proximally, not appreciablybroadened; Im-cu evenly curved to somewhat sinuous, without a ramellus; discosubmarginal cell evenlyhirsute anteriorly. Hind wing with Rs weakly curved or almost straight. Fore tibial spur without amembranous flange behind macro trichial comb; mid and hind trochantelli unspecialized; inner hind tibialspur cylindrical, without a row of long marginal hairs; hind tarsal claws long and weakly curved. Casterstout to very stout; tergite 2 in profile from about 1-5 times as long as posteriorly deep to less than 1-0 timesas long as deep, with thyridia oval, separated from anterior margin by about its own diameter or more;umbo absent; epipleuron pendant. Ovipositor sheath slender. This genus contains four species, D. purpurascens (Smith), D. flavinervis (Cameron) from the Orientaland eastern Palaearctic regions and D. nigrocyaneus (Tosquinet) comb. n. and D. setus (Gauld & Mitchell)comb. n. from the Afrotropical region. The last two were formerly placed in Ophionopsis, a genusoriginally proposed to accommodate the morphologically aberrant nigrocyaneus. D. setus is much moresimilar to the main Dictyonotus line than is nigrocyaneus, suggesting the two genera should be united(Gauld & Mitchell, 1978). The Dictyonotus lineage is distinguished only by one weak apomorphy (Fig. 29),and purpurascens in turn by a further one, suggesting this species may resemble the ancestor of this genus,and also the ancestors of Rhynchophion and Thyreodon. The hosts of Dictyonotus are apparently the larvae of Sphingidae (Townes et al. , 1965 ; Gauld & Mitchell ,1978; 1981). EURYOPHION Cameron Euryophion Cameron, 1906: 83. Type-species: Euryophion nigripennis Cameron, by monotypy.Eurycamptus Morley, 1912: 27. Type-species: Ophion latipenne Kirby, by subsequent designation, Viereck, 1914: 57.Thoracophion Roman, 1943: 22. Type-species: Thoracophion ventrator Roman (= Ophion latipenne Kirby), by monotypy. Primophion Townes, 1971: 65. Type-species: Primophion adustus Townes, by original designation.Rictophion Townes, 1971: 66. Type-species: Euryophion nebulifer Morley (= Cymatoneura ikuthana Kriechbaumer), by original designation. Syn. n. Mandibles large, not twisted, weakly narrowed apically, subequally bidentate. Maxillary palp 3-5segmented, labial palp 3-4 segmented; clypeus in profile flat or with apex out-turned, in anterior aspectusually with margin slightly concave. Ocelli small to large; frontal carina present or absent; occipital carinausually complete, rarely dorsally incomplete. Antennae rather stout, not longer than fore wing. Pronotumunspecialized, spiracular sclerite exposed; notauli weak or vestigial; epicnemial carina present ventrally,sometimes laterally absent; mesopleural furrow usually vestigial. Scutellum without lateral carinae; hindmargin of metanotum at the most only weakly swollen before propodeal spiracle; posterior transversecarina of mesosternum absent except for lateral and rarely central vestiges. Propodeum with anterior areashort, impressed as a deep trough; carinae virtually absent though their former position may be indicatedby rugosities; posterior area from punctate to finely wrinkled to coriaceous. Fore wing with pterostigmaslender; marginal cell moderately long; Rs+2r abruptly curved and thickened basally; discosubmarginalcell without a distinct fenestra, but usually with a small glabrous area in anterior corner; Im-cu fairlyevenly curved, without a distinct ramellus. Hind wing with Rs from weakly to strongly curved. Fore tibialspur without a membranous flange behind macro trichial comb; mid and hind trochantelli simple; innerhind tibial spur subcylindrical or slightly flattened, often with a reduced fringe of hairs; hind tarsal clawslong and weakly curved, rarely with a small apical flange. Gaster stout, tergite 2 in profile short and deep,generally less than 3 times as long as deep posteriorly; thyridia present or absent; umbo vestigial;epipleuron pendant or up- turned. Ovipositor sheath slender, often concealed. Although no character is a unique autapomorphy of this genus, the species share a large number ofapomorphic features (p. 87). They resemble each other greatly in venation, shape of the gaster and formof the head. No other ophionines have reduced palpar segments nor do any have the claw flange found insome more specialized Euryophion. The initial analysis suggested that Euryophion is paraphyletic withrespect to Rictophion. In order to resolve this matter further a more detailed analysis was undertaken usingall species in the two genera, and this confirmed the paraphyletic nature of Euryophion. The sole PHYLOGENY OF THE OPHIONINAE 137 representative of Rictophion, R. ikuthana, can be considered a specialized species of Euryophion, and it ison this basis that Rictophion is here treated as a synonym. The more detailed analysis involved eight species, E. latipennis (201), E. adustus (202), E. ikuthana(203), E. nigripennis (204), E. meridionalis (205), E. variegatus (206), E. vexatious (207) and E. pisinnus(208). Eighteen characters from the original set (1-1, 8-1, 8-2, 9, 13, 18-2, 19, 27-2, 29-1, 29-2, 30, 36-1,36-2, 37, 43-2, 54, 57 and 59) were used together with the following eight characters. 70 Labrum shape. In most ichneumonids the labrum is either semicircular or triangular, but broader basally than medially long; some Euryophion are specialized in having the labrum longer than broad (1). 71 Hind tarsal claw. In some Euryophion the hind claw, is specialized in having a small flange near the apex (1). 72 Rs in hind wing. The moderately curved condition is considered plesiomorphic for Euryophion; in some taxa it is strongly curved (1). 73 Wing patterning. The plesiomorphic condition for ophionines (and members of out-groups) is unpatterned wings. In some Euryophion species the fore wings are distinctly patterned (1). 74 Wing ground colour. Transparent, virtually colourless wings are plesiomorphic for ophionines. The strongly infumate condition found in some Euryophion species is considered to be a derived feature (1). 75 Rs in fore wing. The plesiomorphic condition is for this vein to be simply arcuate. The sinuous condition of a few species of Euryophion is considered to be derived (1). 76 Rugosity of propodeum. The plesiomorphic condition for Euryophion species and related genera appears to be possession of at least some rugae close to the position of the vestigial carinae; thesmooth, punctate propodeum of one Euryophion species is considered to be derived (1). 77 Hairiness of ovipositor sheath. The plesiomorphic condition for most ophionines, and other ichneumonids, is for the sheath to bear close, moderately long pubescence. The short sparsepubescence of one species is considered to be a derived characteristic (1). The primary data matrix obtained (Table 20) was analysed using the methods outlined in the genericanalysis. Characters 8-1, 18-2, 29-1, 29-2, 76 and 77 each have only a single derived state and thereforemake no contribution to resolving phylogenetic relationships. A LeQuesne test on the remaining data set(Table 21) showed three characters (9, 59, 74) to be particularly homoplastic. Removal of these caused animprovement in the overall O/E ratio from 0-62 to 0-45. The high values evident in the labels matrix Table 20 Primary data matrix for genus Euryophion. The taxa corresponding to the numbers are given inthe text (p. 137). 1.1 8.2 13 19 29.1 30 36.2 43.2 57 70 72 74 76 8.1 9 18.2 27.2 29.2 36.1 37 54 59 71 73 75 77 Table 21 Results of LeQuesne test on Euryophion data matrix. Conventions as in Table 2. Character number: Failures observed expected O/E ratio 1.1: 2 9.3 0.22 8.1: 8 13.2 0.60 9 : 10 9.3 1.08 _13_ : 9 13.2 0.68 ^9_ : 8 9.3 0.86 27.2: 10 14.1 0.71 30 : 6 13.2 0.45 36.2: 10 13.2 0.76 37_ : 5 14.1 0.35 43.2: 12 13.2 0.91 54 : 6 13.2 0.45 57 : 8 13.2 0.60 59 : 13 13.2 0.98 70 : 2 9.3 0.22 7l_ : 5 13.2 0.38 72 : 8 13.2 0.60 .73 : 5 13.2 0.38 M. : 12 9 - 3 K29 Zi : 5 13 - 2 - 38 Grand totals- failures observed expected O/E ratio72 116.7 0.62 Ranking ratios 1.1 70 37 71 73 75 30 54 8.1 57 72 13 27.2 36.2 19 43.2 59 9 74 138 I. D. GAULD (Appendix 9) strongly suggest parallelism for the derived state of character 9 (in taxa 202 and 203), ofcharacter 59 (in taxon 208 with respect to other taxa) and character 74 (in taxa 202 and 204). The singlelargest compatible clique comprises nine informative characters, 1-1, 8-2, 30, 37, 54, 70, 71, 73 and 75.These support a cladogram that resolves all taxa except 206-8 which remain as a trichotomy. Parsimonyanalysis yields two equally long minimum length trees which differ in their arrangement of taxa 206-8.There is little biological justification for preferring one or other of these two arrangements (Figs 30, 31) butthe larger number of reversals involved in Fig. 31 mitigate marginally in favour of treating taxa 206 and 207as sister-species. 203 202 208 207 206 205 201 204 8.2 18.2 29.1 29.2 8.1, 9127.236.1 13 74 43.2 36.274 43.2 71 7375 27.2 37 36.25772 136.1 PHYLOGENY OF THE OPHIONINAE 139 203 202 208 207 206 205 201 204 8.218.2 29.1 27.236. Ir 36.2 Figs 30, 31 Alternative cladograms showing most parsimonious arrangement of species of Euryophion.Squares indicate autapomorphic developments; circles that an apomorphic feature has been derived inparallel in two separate lineages; diamonds that a feature has undergone three transformations. 140 I. D. GAULD All methods of analysis yielded similar results in suggesting that E. latipennis and E. nigripennis are veryclosely related, with the former possibly the ancestor of the latter. E. meridionalis was always placed on thesister-species to the clade latipennis + nigripennis, and these three taxa formed a rather distinctspecies-group. The three taxa whose interrelationship is difficult to resolve, E. variegatus, E. vexatious andE. pisinnus, form a very distinct clade whilst E. ikuthana and E. adustus are less closely related. Clearly itwould be unsatisfactory to place ikuthana in a separate genus without at least creating separate genera forthe variegatus species-group and the latipennis species-group. E. adustus could be incorporated into theformer or also treated as a separate genus. As these insects are essentially similar animals that form adistinct group with respect to other ophionines, it is suggested that they be placed in a single genus,Euryophion, but to represent the phylogeny of the group four species-groups may be recognized. Theformal classification may thus be summarized as follows. EUR YOPHION Cameron Rictophion Towneslatipennis species-group latipennis (Kirby) nigripennis Cameron meridionalis (Morley)ikuthana species-group ikuthana (Kriechbaumer)adustus species-group adustus (Townes)variegatus species-group variegatus Gauld & Mitchell pisinnus Gauld & Mitchell vexatious Gauld & Mitchell Euryophion is primarily an Afro tropical genus with one species, E. vexatious, inhabiting southern India(Gauld & Mitchell, 1978; 1981). This taxon is one of the most derived in the genus, and it is very closelyrelated to two African species. This suggests that Euryophion may only have recently become establishedin the Oriental region. The data strongly suggest an African origin for the genus. RHYNCHOPHION Enderlein Rhynchophion Enderlein, 1912: 630. Type-species: Rhynchophion odontandroplax Enderlein, by originaldesignation. Mandibles stout, not twisted, barely narrowed distally, with lower tooth somewhat longer than the upper;outer mandibular surface more or less flat, with scattered hairs. Maxillary palp 5-segmented, labial palp4-segmented; maxilla and labium elongate, projecting below apex of clypeus by about a distance equal tohalf the length of the head; clypeus in profile flat, with margin often slightly out-flared, in anterior aspectwith median obtuse point. Ocelli small, the hind ones separated from the eye by at least their owndiameter; frontal carina absent or present but weak; occipital carina complete, ventrally not joininghypostomal carina. Antennae short and stout, not as long as fore wing, and centrally with segmentstransverse. Pronotum unspecialized; spiracular sclerite exposed; notauli obsolescent; epicnemial carinapresent only ventrally, not extending onto mesopleuron; mesopleural furrow weakly impressed, hori-zontal. Scutellum convex, laterally carinate only at extreme anterior end; posterior transverse carina ofmesosternum present only laterally as vestiges. Propodeum with anterior area short, present as a deeptransverse groove; propodeal carinae absent or with lateral longitudinal ones present as vestiges posterior-ly; posterior area convex, rugose-reticulate grading to punctate. Fore wing with pterostigma very slender;marginal cell quite short; Rs+2r slightly bowed proximally, barely widened near pterostigma; discosub-marginal cell uniformly hirsute; \m-cu fairly evenly curved, without a ramellus. Hind wing with Rs veryweakly bowed. Fore tibial spur without a membranous flange behind macrotrichial comb; mid and hindtrochantelli simple; inner hind tibial spur cylindrical, without a fringe of long close hairs; hind tarsal clawlong, weakly curved. Gaster stout; tergite 2 in profile about as deep posteriorly as long, sometimes deeper,with thyridia close to anterior margin; umbo absent; epipleuron up- turned. Ovipositor sheath slender. Rhynchophion is a small genus restricted to the southern part of the U.S. A. and the Neotropical region.Three species have been described, but these may be variants of a single species. They differ principally inthe colour of the wings and antennae. PHYLOGENY OF THE OPHIONINAE 141 THYREODON Bm\\6 Thyreodon Brulle, 1846: 150. Type-species: Thyreodon cyaneus Brulle, by subsequent designation, Hooker, 1912: 107.Athyreodon Ashmead, 1900: 87. Type-species: Athyreodon thoracicus Ashmead (= Ophion atriventris Cresson), by original designation.Tipulophion Kriechbaumer, 19016: 75. Type-species: Tipulophion gigas Kriechbaumer (= Ophion atriventris Cresson), by monotypy.Macrophion Szepligeti, 1905: 32. Type-species: Macrophion ornatus Szepligeti (= Ophion atriventris Cresson), by subsequent designation, Viereck, 1912: 640.Oleter Shestakov, 1926: 259. Type-species: Oleter selenaction Shestakov (= Thyreodon laticinctus Cresson), by original designation. Mandibles not twisted, massive, weakly to moderately narrowed distally, usually fairly evenly bidentate,or with lower tooth slightly the longer; outer mandibular surface flat except for a deep proximal concavity.Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profile from flat to slightly out-flaredventrally, generally with margin sharp; clypeus in anterior aspect with a median apical tooth. Ocelli usuallysmall, the posterior ones widely separated from the eyes, or in a few species with ocelli very large, almostcontiguous with eyes; frontal carina present, often very strong; frons medially unusual in many species inbeing quite strongly raised between the antennal bases; occipital carina present, usually complete,sometimes not reaching hypostomal carina ventrally. Antennae short and stout, central flagellar segmentstransverse. Pronotum mediodorsally specialized in having anterior and posterior margins raised to formflanges or transverse crests; spiracular sclerite exposed; notauli often strong, generally extending the entirelength of the mesoscutum, frequently with crests across anterior ends; epicnemial carina complete, presenton mesopleuron; mesopleural furrow absent. Scutellum rather small, moderately convex, either withoutlateral carinae, or, at most, with carinae present on anterior 0-4; posterior transverse carina of meso-sternum present as lateral and central vestiges. Propodeum highly modified, with central part stronglyswollen so anterior part is in a very deep groove, as is the spiracle, and with the metapleuron appearingunusually small ; propodeal carinae absent, but usually with postero-dorsal surface of propodeum bearing adeep longitudinal concavity; posterior area otherwise from almost smooth to very strongly reticulate. Forewing with pterostigma slender; marginal cell from short to moderately long; Rs+2r usually only slightlycurved and weakly broadened before joining pterostigma; discosubmarginal cell without a glabrous areaanteriorly; Im-cu usually evenly curved, without a ramellus. Hind wing with Rs more or less straight. Foretibial spur without a membranous flange behind the macrotrichial comb; mid and hind trochantelli simple;inner hind tibial spur cylindrical, without a pronounced marginal fringe of long close hairs; hind tarsal clawslong and weakly curved. Gaster from moderately stout to long and slender; tergite 2 in profile from slightlyto very elongate, with thyridia widely separated from anterior margin; umbo absent; epipleuron pendant,but often creased anteriorly. Ovipositor sheath slender. Thyreodon is a New World genus with about 25 described species. The majority are diurnally active, andfrequently seen feeding from flowers. These species are generally black or brightly coloured, and havesmall ocelli whereas a few nocturnal species are predominantly brown and have large ocelli. It is perhaps aninteresting behavioural 'hangover' from their nocturnal ancestry, that predominantly diurnal species arequite frequently nocturnally active as well and are taken at light. The centre of diversity of the genus appears to be in northern South America. One species extends as farnorth as Canada. The ENICOSPILUS genus-group This is the largest genus-group in the subfamily, containing both the greatest number of generaand species; it also contains some of the most morphologically specialized of all ophionines. The14 genera in this group can be divided into five subgroups (the Orientospilus, Ophiogastrella,Stauropoctonus, Leptophion and Enicospilus subgroups) which correspond with the mainevolutionary lineages apparent in the group. The holophyly of this group is supported by a singleapomorphy, the loss of the vestigial first laterotergites, which is not a particularly convincingreason for uniting the component genera. However, all share a number of other features,including the complete atrophy of the flange on the fore tibial spur, though this feature is sharedwith the Thyreodon genus-group. It is possible that the Enicospilus group may be paraphyleticwith respect to the Thyreodon group but this relationship could not be resolved further with the 142 I. D. GAULD characters and material at hand. The majority of genera in this group (that is the Stauropocto-nus, Leptophion and Enicospilus subgroups) clearly constitute a holophyletic group (see Fig.19); the Ophiogastrella subgroup would seem to be the sister-lineage to these. The position ofthe Orientospilus subgroup is less clear, but it seems reasonably placed as the most primitiveextant lineage of the group. This genus-group is primarily a Pan-tropical complex of genera. Only isolated species ofSimophion, Stauropoctonus, Dicamptus and Enicospilus occur in the Holarctic region. Thehosts of the majority of species appear to be tree- and shrub-feeding lepidopterous larvae,though only a small percentage of all the species has been reared. The ORIENTOSPILUS subgroup This group contains three genera, Orientospilus, Simophion and Prethophion. The systematic position ofthe latter genus is far from clear as it is united with Orientospilus and Simophion solely on the basis of aweak character - an incomplete posterior transverse carina of the mesosternum - and its present position isreally only the most parsimonious possible. Prethophion does share certain features with Thyreodon - ithas a slightly twisted mandibular axis, though not as pronounced as that of species in the Thyreodonlineage, and a similarly stout gaster; like species of the Thyreodon lineage the anterior margin of the tegulais not reflexed (possibly a derived feature). However, the plesiomorphic form of the anterior part of thepropodeum and tarsal claws seem to exclude Prethophion from the Thyreodon group as here defined.Furthermore, the first laterotergite is lost and the mandibles are specialized in a quite different way, andmuch more closely resemble the derived form found in Orientospilus. Simophion and Orientospilus are more convincingly related; both have a similarly modified clypeus,have a basally angulate Rs+2r in the fore wing, have lost the epicnemial carina laterally and have a shortmarginal cell. ORIENTOSPILUS Morley Orientospilus Morley, 1912: 6. Type-species: Orientospilus individuus Morley, by subsequent designation,Morley, 1913: 378. Mandibles not twisted, long, strongly and evenly tapered, with upper tooth much longer than the lower;outer mandibular surface flat, sparsely hirsute. Maxillary palp 5-segmented, labial palp 4-segmented;clypeus in profile flat with margin out-flared; in anterior aspect concave. Ocelli moderately large, theposterior ones separated from the eyes by 0-3-1-3 times their minimum diameter; frontal carina absent;occipital carina complete, though ventrally not reaching hypostomal carina. Antennae moderately stout,the distal segments quadrate, the flagellum barely longer than the fore wing. Pronotum mediodorsally withanterior and posterior margins reflexed, raised as keels; spiracular sclerite exposed; notauli absent;epicnemial carina present ventrally, not reaching above level of lower corner of pronotum; mesopleuralfurrow vestigial or absent. Scutellum quadrate, weakly to moderately convex, with lateral longitudinalcarinae complete; posterior transverse carina of mesosternum present only laterally as vestiges. Prop-odeum with anterior area short, deeply impressed as groove; anterior transverse carina from complete toobsolescent, other carinae absent; posterior area coarsely reticulate, usually medially concave. Fore wingwith pterostigma of moderate breadth, evenly narrowed distally; marginal cell short; Rs+2r curved andbroadened before joining pterostigma; discosubmarginal cell with a glabrous area anteriorly; Im-cuevenly arcuate without a ramellus. Hind wing with Rs from virtually straight to curved abruptly proximally.Fore tibial spur without a membranous flange behind macro trichial comb; mid and hind trochantelliunspecialized; hind inner tibial spur more or less flattened, with margin of long close hairs; hind tarsal clawunspecialized. Gaster slender; tergite 2 in profile elongate, with thyridia obsolescent, or if discernible,small and quite close to anterior margin; umbo present; epipleuron usually up-turned, rarely, in someindividuals, the posterior part is pendant. Ovipositor sheath slender. This small genus contains three described species, capitatus Gauld & Mitchell from southern and westAfrica, melasma Townes from Madagascar and individuus Morley from eastern peninsular India.Structurally these are very similar insects. The Madagascan species is the least specialized; both the Indianand African species have lost the anterior transverse carina of the propodeum and have a shorter, stouterand apically more setaceous flagellum and have a broader malar space suggesting they are sister-species(Fig. 32). The Madagascan species may well be the ancestor of the other two. PHYLOGENY OF THE OPHIONINAE 143 Fig. 32 Putative phylogenetic arrangement of species of Orientospilus . This cladogram is supported bythe following apomorphic features: 1 , malar space very wide; 2, ocelli widely separated from eyes; 3, Rsin hind wing curved proximally; 4, flagellum with proximal segments almost quadrate; 5, anteriortransverse carina of propodeum absent; 6, malar space moderately wide, at least 0-4 times basalmandibular width. PRETHOPHION TownesPrethophion Townes, 1971: 74. Type-species: Prethophion latus Townes, by original designation. Mandibles very slightly twisted, very strongly narrowed, subequally bidentate; outer mandibular surfaceslightly convex, with a proximal concavity. Maxillary palp 5-segmented, labial palp 4-segmented; clypeusin profile almost flat, margin centrally blunt, clypeus in anterior aspect virtually almost truncate. Ocellilarge, the posterior ones more or less touching eyes; frontal carina absent; occipital carina entirely absent.Antennae moderately slender, barely longer than fore wing but with central segments clearly elongate.Pronotum unspecialized; spiracular sclerite exposed; notauli vestigial; epicnemial carina present onmesosternum, present but very weak on mesopleuron; mesopleural furrow very distinct, oblique,extending from episternal scrobe to subalar prominence, the speculum posterodorsal to this carina stronglyinflated. Scutellum convex, laterally carinate only on extreme anterior end; posterior transverse carina ofmesosternum absent except for lateral vestiges. Propodeum short and abruptly declivous posteriorly;anterior area unspecialized; anterior transverse carina complete; blunt tubercles (which are possiblyvestiges of posterior transverse carina) discernible; posterior area concave, very finely alutaceous. Forewing with pterostigma moderately slender; marginal cell slender; Rs+2r proximally not broadened, almoststraight; discosubmarginal cell without a glabrous area anteriorly; Im-cu fairly evenly curved, without aramellus. Hind wing with Rs strongly curved. Fore tibial spur without a membranous flange behindmacrotrichial comb; mid and hind trochantelli not specialized; inner hind tibial spur flattened, with amargin of long close hairs; hind tarsal claw unspecialized. Gaster quite stout, centrally almost cylindrical; 144 I. D. GAULD tergite 2 in profile elongate, with thyridia close to anterior margin; umbo low but distinct; epipleuronup-turned. Ovipositor sheath stout. This enigmatic genus contains a single species which occurs at low altitudes in Peru and Bolivia. Nothingis known of its biology. SIMOPHION CushmanSimophion Cushman, 1947: 446. Type-species: Simophion excarinatus Cushman, by original designation. Mandibles fairly evenly tapered, not or barely twisted, subequally bidentate or with lower tooth slightly theshorter; outer mandibular surface punctate, proximally concave. Maxillary palp 5-segmented, labial palp4-segmented; clypeus in profile flat, apically sometimes slightly out-flared, in anterior aspect concave.Ocelli moderately large, the posterior ones separated from eye by 0-2-0-7 times their maximum diameter;frontal carina absent; occipital carina complete, though generally not reaching hypostomal carinaventrally. Antennae moderately long and slender, about 1-4-1-6 times length of fore wing. Pronotumunspecialized; spiracular sclerite exposed; notauli vestigial; epicnemial carina present ventrally, notreaching onto mesopleuron laterally; mesopleural furrow weak, diagonal, extending from episternalscrobe to near subalar prominence. Scutellum moderately convex, narrow, without lateral carinae;posterior transverse carina of mesosternum present only laterally as vestiges. Propodeum with anteriorarea short, striate slightly centrally; anterior transverse carina absent as are other propodeal carinae, thepropodeum being narrowed distally; marginal cell short; Rs+2r very abruptly angled before joiningpterostigma, its extreme end abruptly broadened; discosubmarginal cell very sparsely hirsute, with aglabrous area anteriorly; Im-cu curved or sinuous, without a ramellus. Hind wing with Rs very stronglybowed. Fore tibial spur without a membranous flange behind macrotrichial comb; mid and hindtrochantelli unspecialized; inner hind tibial spur flattened, with a margin of long hairs; hind tarsal clawsunspecialized. Gaster slender; tergite 2 in profile elongate, with thyridia obsolescent, or if present quitelarge and relatively close to anterior margin of tergite; umbo distinct; epipleuron up-turned. Ovipositorsheath slender. This moderately small genus is restricted to the 'Mediterranean' type biomes of the northern hemi-sphere. Two species occur in the Middle East and Central Asia (Horstmann, 1981), whilst five or six occurin the south-west of the United States (Townes, 1971). Similar disjunct distributions occur in severalgenera of lower Aculeata (M. C. Day, pers. comm.) and in the myrmicine ant genus Messor (Bolton,1982). Amongst ophionines Eremotylus has a similar, though slightly more extensive distribution. The OPHIOGASTRELLA subgroup This group comprises a single Neotropical genus, Ophiogastrella, which was treated as a member of theOphion genus-group by Cushman (1947), but as a member of the Enicospilini by Townes (1971). Townes &Townes (1973) considered it to be closely related to Laticoleus. Structurally Ophiogastrella shows a remarkable combination of primitive and derived features. Thespecialized fore tibial spur and first laterotergite suggest it belongs close to the base of the Enicospiluslineage, but the well-developed umbo, exposed spiracular sclerite and unspecialized anterior part of thepropodeum suggest it is more primitive than many other genera. These features, together with the lack ofan impressed clypeal apex, externally flat mandible and slender ovipositor sheath suggest it is not at allclosely related to Laticoleus. The peculiar male claws are an autapomorphy of this taxon. Ophiogastrella is confined to the southern part of Central America and northern South America. Thereis no evidence to suggest the group has ever occurred elsewhere. OPHIOGASTRELLA Brues Ophiogastrella Brues, 1912: 201. Type-species: Ophiogastrella maculithorax Brues, by original desig-nation. Brachyscenia Enderlein, 1921: 36. Type-species: Brachyscenia nigriventris Enderlein, by original desig-nation. Mandibles not twisted, evenly but only moderately narrowed apically, subequally bidentate; outermandibular surface flat except for small proximal concavity, centrally with scattered hairs. Maxillary palp5-segmented; labial palp 4-segmented; clypeus in profile flat or weakly convex, margin blunt or sharp,never impressed; clypeus in anterior aspect weakly convex, straight or very slightly concave. Ocelli large,the posterior ones generally very close to the eyes; frontal carina absent; occipital carina complete,ventrally (as genal carina) usually sharply angled and complete to hypostomal carina. Antennae of PHYLOGENY OF THE OPHIONINAE 145 moderate length, 1-2-1-5 times the length of the fore wing, rarely slightly longer. Pronotum unspecialized;spiracular sclerite exposed; notauli absent; epicnemial carina complete; mesopleural furrow absent orweakly impressed, diagonal, extending from episternal scrobe to near subalar prominence. Scutellumweakly to moderately convex, with lateral longitudinal carina extending about 0-8 of its length; posteriortransverse carina of the mesosternum vestigial. Propodeum with anterior area short, unspecialized;anterior transverse carina from complete to absent, other carinae at most present only as vestiges;posterior part of propodeum generally rather smooth. Fore wing with pterostigma quite large and broad;marginal cell long; Rs+2r curved or slightly angled proximally, from slender to moderately broadened;discosubmarginal cell with a small glabrous area anteriorly; Im-cu very strongly and evenly curved,without a ramellus. Hind wing with Rs from almost straight to strongly bowed. Fore tibial spur without amembranous flange behind the macrotrichial comb; mid and hind trochantelli simple or with a blunt toothnear distal margin; inner hind tibial spur flattened, with a fringe of close hairs; hind tarsal claws of femaleunspecialized, of male flattened with pectinal comb present on both sides (Fig. 47). Gaster moderatelyslender; tergite 2 in profile elongate, with thyridia remote from anterior margin; umbo quite welldeveloped; epipleuron generally upturned. Ovipositor sheath slender. Ophiogastrella is a small genus containing about 10 species, of which only three are described (Townes &Townes, 1966). They are restricted to the Neotropical region from Costa Rica to about 10S in Brazil, andseem to be associated with areas which have a pronounced dry season. The species are rather similar to each other, and best separated by differences in the length of 3r-w,shape of Rs in the hind wing and sculpture of the propodeum. There are also more subtle differences in theshape of the clypeus, and the head. The STAUROPOCTONUS subgroup This subgroup contains two genera, Stauropoctonus (including as a synonym Aulophiori) and Lepiscelus.The exposed spiracular sclerites of species in this complex suggest it is one of the more primitive ofenicospiline lineages. It is also one of the taxonomically most distinctive as its members are characterizedby an incomplete occipital carina, a transverse mesopleural furrow, slender pterostigma, basally thickenedand bent Rs+2r and a very sinuous Im-cu. Most species have Im-cu and Cw la basally widely separated,and all have the mid and hind trochantelli specialized. The group contains few species. The most primitive Stauropoctonus species and the solitary Lepiscelusoccur in the Old World, suggesting an Old World origin for the group. LEPISCELUS Townes Lepiscelus Townes, 1971: 73. Type-species: Lepiscelus gracile Townes (= Eremotylus distorts Seyrig), byoriginal designation. Mandibles twisted about 5, proximally strongly narrowed, distally parallel-sided with upper tooth abouttwice as long as the lower; outer mandibular surface more or less flat. Maxillary palp 5-segmented, labialpalp 4-segmented; clypeus in profile weakly convex, margin narrowly acute, laterally blunted, short,exposing labrum; clypeus in anterior aspect truncate or even slightly concave. Ocelli large, the hind onesvery close to the eyes; frontal carina absent; occipital carina entirely absent. Antennae very long andslender, about 1-7 times length of the fore wing. Pronotum mediodorsally rather long, flat; spiracularsclerite exposed; notauli very short but often strongly impressed in part; epicnemial carina weak, laterallybecoming obsolescent at level of lower corner of pronotum; mesopleural furrow very weakly impressed,transverse. Scutellum almost rectangular, carinate laterally only at extreme anterior end; posteriortransverse carina of mesosternum complete. Propodeum with anterior area moderately long, dorsallywithout carinae, with posterior area finely wrinkled to rugulose. Fore wing with pterostigma slender;marginal cell long; Rs+2r abruptly curved and broadened before joining pterostigma; discosubmarginalcell anteriorly glabrous; Im-cu very strongly sinuate. Hind wing with Rs strongly curved; hamuli arrangedin two groups, the proximal group comprising two which are longer and flatter than those in the distal groupof three. Fore tibial spur without a membranous flange behind the macrotrichial comb; mid and hindtrochantelli extended apically as a broad flange over the proximal end of the femur; inner hind tibial spurflattened, with a margin of long, close hairs; hind tarsal claws quite long, those of male more closelypectinate than those of the female. Gaster slender; tergite 2 in profile elongate, with thyridia remote fromanterior end; umbo vestigial; epipleuron up-turned. Ovipositor sheath slender; male with gonosquamaunusually long. Lepiscelus contains a single species which is widespread, but apparently rather uncommon, throughoutthe Afrotropical region between latitude 10N and 20S. The most distinctive feature of this genus is the 146 I. D. GAULD flanged mid and hind trochantelli which are, amongst the Ichneumonidae, a unique feature of Lepiscelusdistans. Nothing is known of the biology of this insect, but the form of the mandibles and head shape arereminiscent of some species of Enicospilus that either inhabit arid areas or are active in the dry season. STAUROPOCTONUS Brauns Stauropoctonus Brauns, 1889: 75. Type-species: Ophion bombycivorus Gravenhorst, by monotypy. Stauropodoctonus Morley, 1913: 375. [Unjustified emendation.] Nipponophion Uchida, 1928: 201. Type-species: Nipponophion variegatus Uchida (= Ophion bombyci-vorus Gravenhorst), by monotypy. Aulophion Cushman, 1947: 458. Type-species: Aulophion bicarinatus Cushman, by original designation.Syn. n. Mandibles twisted about 85, evenly, but quite strongly narrowed apically, subequally bidentate. Maxillarypalp 5-segmented, labial palp 4-segmented; clypeus in profile weakly convex, in anterior aspect withmargin slightly convex. Ocelli large, the posterior ones close to eye margins; frontal carina absent; occipitalcarina usually absent, partially present in one species. Antennae very long and slender, sometimes up to 2times length of fore wing. Pronotum unspecialized, spiracular sclerite exposed; notauli vestigial or absent;epicnemial carina from present and complete to absent; mesopleural furrow strong, extending fromepisternal scrobe to upper end of prepectal carina (or the corresponding position if this carina is absent), insome species impressed as a deep groove. Scutellum moderately convex to convex, either without lateralcarinae or with carina incomplete; posterior transverse carina of mesosternum from complete to reduced tolateral vestiges. Propodeum with anterior area long; anterior transverse carina usually complete, theposterior one sometimes discernible; posterior area from weakly rugulose to reticulate. Fore wing withpterostigma moderately slender; marginal cell long; Rs+2r abruptly curved and slightly thickened basally;discosubmarginal cell with a glabrous area anteriorly, which may be expanded along Rs+2r; Im-cuwithout a ramellus, either evenly bowed or sinuous. Hind wing with Rs from almost straight to weaklybowed. Fore tibial spur without a membranous flange behind macrotrichial comb; mid and hindtrochantelli specialized in having the outer distal margin produced into a strongly decurved sharp spine;inner hind tibial spur flattened, with a marginal fringe of long hairs; hind tarsal claws unspecialized. Gasterslender; tergite 2 in profile very elongate, with the thyridia remote from anterior margin; umbo vestigial;epipleuron pendant or up-turned. Ovipositor sheath slender. In the analysis of the genera Stauropoctonus appeared to be paraphyletic with respect to Aulophion. Thelatter genus has been separated from Stauropoctonus by possession of two apomorphic features (Cushman,1947; Townes, 1971). No apomorphies have ever been suggested for Stauropoctonus, though the cladeStauropoctonus + Aulophion is one of the most clearly defined in the subfamily. It is characterized by thefollowing combination of apomorphies (though none is actually unique to the clade) - at least partial loss ofoccipital carina; presence of projections on mid and hind trochantelli; presence of impressed transversefurrow on mesopleuron; presence of strongly twisted mandibles; having Rs+2r in the fore wing basallyangled. In addition to these features the head and mesoscutal profile of species of this clade arecharacteristic, though these 'apomorphies' are so subtle I have not been able to code them. To furtherresolve the relationship between Aulophion and Stauropoctonus more data were analysed. The following taxa were included - Aulophion sp. 1 (? bicarinatus) (301), Aulophion sp. 2 (302),Stauropoctonus bombycivorus (303), 5. torresi (304), 5. townesorum (305) and S. occipitalis (306). This isall the species in the group except for one (or possibly two) Aulophion species which are very closelyrelated to taxon 302. To represent the range of interspecific variation in the group the following charactersfrom the primary set were utilized- 1-2, 19, 20-1, 24-1, 25-2, 27, 30, 36, 43-1, 43-2, 44-1, 45-2, 47-2, 57 and59. (It is noteworthy that 27 and 36 are composites of 27-1 and 27-2 and 36-1 and 36-2 respectively; thesecharacters showed identical state distribution over the data and therefore were treated as single charactersto avoid excess weighting.) In addition to these fifteen, three further characters were used. 80 Median carina of propodeum. This apomorphic feature (1) is found only in a few isolated Neotropical ophionines. The plesiomorphic condition, no median carina, is found in virtually all ophionines andmembers of the various out-groups. 81 Metapleural sculpture. In most out-group taxa, most ophionines and several species of this group the metapleuron is smooth and finely punctate. Some Stauropoctonus species are specialized in havingthis region coarsely rugose (1). 82 Position of cu-a with respect to Rs&M in fore wing. The plesiomorphic condition for ophionines apparently is for cu-a to be proximal to the base of Rs&M. A few Stauropoctonus are specialized inhaving these veins opposite (1). PHYLOGENY OF THE OPHIONINAE 147 Table 22 Primary data matrix for genus Stauropoctonus. The taxa corresponding to the numbers aregiven in the text (p. 146). 1.2 20.1 25.2 30 43.1 44.1 47.2 59 81 19 24.1 27 36 43.2 45.2 57 80 82 301302303304305306 .47.2 36r45.2 47.257 I- 1 80 81, 82 119, 27.1, 30143.1, 44.1 k45.2 1.2, 43.2 25.2 Fig. 33 Favoured cladogram showing putative phylogenetic arrangement of species of Stauropoctonus.This is based on the largest compatible set often informative characters and requires 25 transformationsteps. 148 I. D. GAULD The primary data matrix for taxa 301-6 is shown in Table 22. Parsimony and compatibility analyses,using techniques outlined above, yielded two 'best' alternative hypotheses of phylogenetic relationship(Figs 33, 34). The cladograms have a large number of features in common. Firstly, both suggestStauropoctonus is paraphyletic with respect to Aulophion, and this in turn suggests that Aulophion can betreated as a synonym of Stauropoctonus , as has been formally proposed above. Secondly, both suggest that i 19, 27.1, 30i 43.1, 44.1 45.2 |47.2r, 57 1.2, 43.2 25.2 , 47.2 Fig. 34 Alternative cladogram for species of Stauropoctonus. This is the most parsimonious arrangementof taxa discovered and requires 24 transformation steps. (Symbols as for Figs 30-31.) PHYLOGENY OF THE OPHIONINAE 149 Aulophion is both a holophyletic clade and the most derived lineage in the group. Thirdly, both place 5.ocdpitalis as the most primitive taxon in the group. The two cladograms differ in their arrangement of 5.bombycivorus, S. torresi and 5. townesorum and these different arrangements depend on the alternativecompatible sets of characters 36 and 45-2 versus 59, 81 and 82. Character 36 has been demonstrated to beboth unreliable and of dubious polarity, whilst 45-2 can be regarded as a loss apomorphy (loss of hairs onwing membrane). Character 59 (colour of interocellar area) is also far from a convincing apomorphy, but81 and 82 strongly suggest that torresi and townesorum are sister-species. S. bombycivorus is rather difficultto place, but could be either the ancestor of, or a close relative of the ancestor of both the 'Aulophion'species-group and torresi/townesorum. The more primitive species of this genus (ocdpitalis and bombycivorus) are restricted to the Old World,the former to the mountains of Madagascar and the latter to the southern part of the Palaearctic, except inAsia where it extends north into Kamchatka and the Kurile Islands (Townes, Momoi & Townes, 1965).Possibly the ancestor of these taxa was widely distributed in the Old World at one time, and theMadagascan survivor is a relict of this. There are clear indications in the Madagascan fauna of an ancientPalaearctic affinity amongst the Hymenoptera. The symphytan taxon Cephidae is represented on theisland and in the Holarctic only (Benson, 1935), and a number of ichneumonid genera have rather similardistributions, e.g. Neliopisthus, Euceros (Townes, 1969: Barren, 1978). As the widespread Palaearcticspecies bombycivorus may be almost directly ancestral to the two other species-groups, it may at one timehave also occurred in the Nearctic region. The torresi/townesorum group is Malesian/Melanesian, with theformer species constituting part of the characteristically intrusive, Indo-Papuan faunal element in northernAustralia (Gauld, 19840). The distribution of torresi, in transcending both Wallace's and Weber's lines,suggests its current range is the result of a dispersive rather than a vicariance event. The Neotropicalspecies form a rather distinctive clade. The most northerly extent of their distribution seems to be CostaRica where one species occurs in lower montane (1000-1500 m) forest, whilst none has been found south of25S. Perhaps the ancestor of this group reached South America from the north during the Miocene whenan archipelagic connection existed between the two continents (Rich & Rich, 1983). This evolutionary'scenario' is based on the supposition that a Stauropoctonus occurred in the United States. At present thereis no evidence for this; the scenario presented is merely the most parsimonious interpretation of data.Other scenarios would necessitate advocating much more widespread extinction. The LEPTOPHION subgroup This complex of genera is characterized by the form of the mandibles, which have a more or less discernibleimpressed groove extending diagonally, the convex clypeus with an impressed acute margin, the very longslender antennae and, for most species, a somewhat broadened ovipositor sheath. The venation of allspecies is rather similar in having Rs+2r basally broadened, and often angulate before joining thepterostigma; Im-cu is either sinuate or evenly curved. This subgroup contains three genera, Leptophion, Laticoleus and Pamophion which are difficult toresolve as the holophyly of Leptophion cannot convincingly be demonstrated. Pamophion, a monobasictaxon, is holophyletic and is apparently the sister-lineage of Leptophion + Laticoleus. The latter is aholophyletic group, but it may have arisen from within the former. The only apomorphy that can bepostulated for Leptophion is the presence of a specialized penultimate hamulus. (The apomorphies shownin Fig. 19 are subject to much exception and parallelism in some species of Laticoleus, and reference to theoverall classification suggests that presence of a complete posterior transverse carina of the mesosternum isplesiomorphic for this complex - though some taxa have partially lost this carina.) The assumption that thespecialized hamulus is a group apomorphy necessitates speculating that reversal has occurred in severaltaxa, but there is some evidence that this has happened. For example, L. tetus has a slight indication ofsome enlargement of the hamulus and its sister-species, L. yampus, has it clearly specialized, though not aslarge as that of L. iochus, the sister-species of yampus + tetus. L. ankylosus and L. eithos are highlyspecialized species (without a modified hamulus) which are closely related to L. vernalis which has a longpenultimate hamulus, and Leptophion species 1 (BMNH) which has it only slightly modified. These fourspecies apparently belong to the maculipennis lineage, the remainder of which all have a highly specializedhamulus. A tentative phylogeny for this subgroup is proposed in Fig. 35. Laticoleus is retained as a separategenus, but its validity needs to be reassessed as more material becomes available for study and thephylogeny of the species of Leptophion is better understood. Judging from the present distribution of this group (Fig. 36) it is most parsimonious to postulate an OldWorld origin for this complex of genera. The more primitive taxa are restricted to Melanesia or Australiawhich suggests an Australo-Melanesian origin for this group. However, if Laticoleus is truly primitive withrespect to Leptophion, then an alternative hypothesis is that this group was once widely spread in the Old LEPTOPHION MACULIPENNIS GROUP A Fig. 35 Cladogram showing putative phylogeny of groups of species in the Leptophion subgroup. Blacksquares represent apomorphic features, white plesiomorphic; diagonally divided squares indicate amixture of the two states in the component species of a particular lineage. The characters that supportthis arrangement are: 1, loss of specialized hamulus; 2, Rs+2r emitted from centre of pterostigma; 3,Im-cu very sinuous; 4, specialized hind tarsal claws; 5, loss of posterior transverse carina of mesoster-num; 6, presence of a 'hair brush' on mandible; 7, genal carina evanescent before joining hypostomalcarina; 8, lengthened penultimate distal hamulus; 9, exceptional broad ovipositor sheath that in profile isalmost quadrate; 10, long malar space; 11, ovipositor sheath at least moderately broad, in profile morethan 0-4 times as deep as long; 12, marginal cell of hind wing at least partially glabrous; 13, anterior areaof propodeum elongate; 14, epipleuron 2 pendant; 15, hind trochantellus marginally produced into blunttooth; 16, posterior transverse carina of mesosternum present. PHYLOGENY OF THE OPHIONINAE 151 Fig. 36 The geographical distribution of taxa of the Leptophion subgroup. World tropics and has suffered considerable extinction, leaving only a rather specialized primitive lineagein the Afrotropical region and a few relicts east of Wallace's line. Gauld (1984a) suggested that the aniciand iochus groups of Leptophion, in Australia, have had a northern origin, and that once their ancestorshad adapted sufficiently to cross the rain forest/savannah interface (Taylor, 1972) then they had ampleopportunity to give rise to radiations in species-poor Australia (Gauld, 1984a). The most specialized groupof Leptophion, the maculipennis group, is best represented in the more easterly parts of Indonesia andMelanesia. It is probable that this species-group arose in this area and a few species have spread west to theAsian mainland. LATICOLEUS Townes Laticoleus Townes, in Townes & Townes, 1973: 358. Type-species: Coiloneura unicolor Szepligeti, byoriginal designation. Mandibles not twisted, weakly narrowed apically, generally subequally bidentate; outer mandibularsurface with subbasal swelling weak or well developed, with a diagonal groove extending from the upperproximal corner to between bases of teeth. Maxillary palp 5-segmented, labial palp 4-segmented; clypeusin profile convex, margin impressed, in anterior aspect with margin weakly convex. Ocelli generallymoderately large, the posterior ones separated from the eyes by 0-2-1-0 times their minimum diameter;frontal carina absent; occipital carina complete, the lower part (genal carina) reaching to hypostomalcarina. Antennae very long and slender, at least 1-5 times length of fore wing. Pronotum usuallyunspecialized, or in one Madagascan species with anterior margin mediodorsally expanded and curvedback; spiracular sclerite partially concealed; notauli vestigial or absent; epicnemial carina complete, weakor even absent on mesopleuron; mesopleural furrow undeveloped. Scutellum weakly convex, with laterallongitudinal carinae present, more or less complete; posterior transverse carina of the mesosternumincomplete, present as lateral or rarely central vestiges. Propodeum with anterior area moderately long,striate or smooth; anterior transverse carina present or absent, other carinae usually absent; posterior areafrom rugose to virtually smooth. Fore wing with pterostigma moderately stout; marginal cell long; Rs+2r 152 I. D. GAULD conspicuously thickened near pterostigma, often abruptly curved in proximal part; discosubmarginal cellwith a glabrous area in anterior corner, this area generally quite large, and in one species, bearing a weaksclerite; Im-cu without a ramellus, usually fairly evenly bowed, in a few species slightly sinuous. Hind wingwith Rs from almost straight to strongly curved, with marginal cell proximally, at least narrowly glabrous;distal hamuli unspecialized, or with proximal three enlarged. Fore tibial spur without a membranous flangebehind macrotrichial comb; mid and hind trochantelli unspecialized or with a weak tooth projectingdistally; inner hind tibial spur flattened with a margin of close long hairs; hind tarsal claws usuallyunspecialized, rarely with very coarse pectinae. Gaster slender; tergite 2 in profile very elongate, withthyridia well removed from anterior margin; umbo absent; epipleuron up-turned. Ovipositor sheathexceptionally broad. Laticoleus is a moderately small Afrotropical genus containing 11 described species (Gauld & Mitchell,1978) . The majority occur in Madagascar and east Africa. Nothing is known of the biology of species of thisgenus. LEPTOPHION Cameron Leptophion Cameron, 1901: 227. Type-species: Leptophion longiventris Cameron, by monotypy.Spilophion Cameron, 1905: 124. Type-species: Spilophion maculipennis Cameron, by monotypy.Coiloneura Szepligeti, 1905: 35. Type-species: Coiloneura melanostigma Szepligeti (= Leptophion lon-giventris Cameron), by subsequent designation, Viereck, 1914: 35. Mandibles not twisted, weakly narrowed apically, generally subequally bidentate; outer mandibularsurface often with a basal swelling, and with a diagonal groove extending from near upper corner tobetween bases of teeth, sometimes with this groove bearing a brush of long hairs, other times with thegroove very weak. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profile convex, marginimpressed acute, in anterior aspect usually with margin weakly convex, rarely slightly concave. Ocelligenerally large, the posterior ones close to the margins of the eyes, or rarely in some Australian examples,with ocelli smaller; frontal carina absent; occipital carina either complete or obsolescent at extreme ventralend so the carina fails to join the hypostomal carina. Antennae in most species very long and slender, atleast 1-5 times the length of the fore wing, in some Australian species shorter, barely longer than the forewing. Pronotum unspecialized; spiracular sclerite partially concealed, in a few Malesian taxa almostcompletely covered; notauli vestigial or absent; epicnemial carina complete; mesopleural furrow usuallyundeveloped. Scutellum moderately convex, from without lateral carinae to with these carinae complete;posterior transverse carina of mesosternum complete, interrupted before mid coxae, or rarely absent.Propodeum with anterior area long and usually striate; anterior transverse carina usually complete, rarelyabsent, the other carinae usually not discernible; posterior area from strongly rugose to smooth andshining. Fore wing with pterostigma moderately broad; marginal cell long; Rs+2r conspicuously thickenednear pterostigma, usually abruptly curved or angled near base; discosubmarginal cell usually with aglabrous area in anterior corner, this area usually large and extending 0-2 of the way down Rs+2r, rarelywith a weak alar sclerite, sometimes secondarily with micro trichia obscuring part of glabrous area; Im-cuusually without a ramellus, either evenly bowed or, more usually, sinuous. Hind wing with Rs from almoststraight to very strongly bowed, with marginal cell adjacent to this vein at least narrowly glabrous; distalhamuli unspecialized or with penultimate hamulus long and coiled. Fore tibial spur without a membranousflange behind macrotrichial comb; mid and hind trochantelli unspecialized; inner hind tibial spur flattened,with a fringe of long close hairs; hind tarsal claws unspecialized or with distal pectinal tooth projectingbeyond the apex of the claw. Gaster usually slender, tergite 2 in profile from moderately to very elongate,with thyridia remote from anterior margin; umbo absent; epipleuron up-turned. Ovipositor sheathmoderate, broad. Leptophion is a moderately large Indo-Pacific genus containing 30 described species (Gauld, 1977;Gauld & Mitchell, 1981). The majority are restricted to lower montane forests on the islands to the east ofthe Sunda Shelf. Many are endemic to a particular island and recent collecting has yielded an undescribedspecies from Sulawesi (BMNH). Most species are only known from relatively few individuals. The genus is divisible into four species-groups, the largest of which, the maculipennis species-group, maybe subdivided into three species-complexes. Key to species-groups and complexes of Leptophion 1 Anterior transverse carina of mesosternum incomplete; propodeum often with anterior trans-verse carina absent anici species-group PHYLOGENY OF THE OPHIONINAE 153 - Anterior transverse carina of mesosternum complete; propodeum almost always with anterior transverse carina complete 2 2 Hind tarsal claw simple ; genal carina usually joining hypostomal carina 3 - Hind tarsal claw with distal pectinal tooth projecting apically; genal carina usually not reaching hypostomal carina (maculipennis species-group) 4 3 Mandible with a dense brush of long stout hair on outer surface iochus species-group - Mandible with scattered hairs on outer surface longicornis grade-group 4 Fore wing with Rs+2r emitted from near centre of pterostigma; anterior corner of discosubmar- ginal cell partially hirsute eithos species-complex - Fore wing with Rs+2r emitted from proximal end of pterostigma; anterior corner of discosub- marginal cell glabrous 5 5 Penultimate distal hamulus unspecialized; fore wing with Im-cu from evenly curved to weakly sinuous radiatus species-complex - Penultimate distal hamulus long and coiled; fore wing with Im-cu moderately to strongly sinuous maculipennis species-complex longicornis grade-group This paraphyletic assemblage contains two species, L. longicornis (Szepligeti) and L. bakeri (Cheesman).The group can only be characterized by symplesiomorphies. It is restricted to New Guinea and the NewHebrides. anici species-group This holophyletic group contains three species, L. anici Gauld, L. antennatus (Morley) and L. unical-caratus Gauld. The group is characterized by the loss of the posterior transverse carina of the mesoster-num; species tend to have less propodeal sculpture than most Leptophion species, and usually thetransverse carina is lost. The anici species-group is restricted to Australia and New Caledonia wherespecies occur in drier habitats than mos't Leptophion. iochus species-group This holophyletic group contains three Australian species, L. iochus Gauld, L. yampus Gauld and theaberrant L. tetus Gauld. The group is characterized by possession of a brush of hair on the outer surface ofthe mandible; they are generally more robust species with more densely pubescent wings than mostLeptophion species. maculipennis species-group This holophyletic group is characterized by the presence of a modified hind tarsal claw; the majority ofspecies have infumate marks in the proximal corner of the marginal cells and in most the genal carina doesnot join the hypostomal carina. This is the largest species-group and may be subdivided into threeapparently holophyletic species-complexes. eithos species-complex. This group contains four species, L. eithos Gauld & Mitchell, L. ankylosusGauld & Mitchell, L. vernalis Gauld & Mitchell and Leptophion species 1 (BMNH). It is characterized bythe highly modified fore wing venation in which Rs+2r is emitted near the centre of the pterostigma. Thiscomplex is restricted to Western New Guinea, Sulawesi and the intervening islands. radiatus species-complex. This group contains six species, L. radiatus (Uchida), L. pterospilus Gauld &Mitchell, L, vechti Gauld & Mitchell, L. lavellai Gauld & Mitchell, L. cheesmanae Gauld & Mitchell andL. illustrious Gauld & Mitchell. It is characterized by the possession of unspecialized hamuli; most specieshave a rather evenly curved Im-cu and very few have Rs in the hind wing appreciably curved. This group iswidely distributed from the Solomon Islands to the Continental Asian mainland. maculipennis species-complex. This group contains 13 species, L. maculipennis (Cameron), L. pubes-cens Gauld & Mitchell, L. gobius Gauld & Mitchell, L. kus Gauld & Mitchell, L. juxtus Gauld & Mitchell,L. magus Gauld & Mitchell, L. townesi Gauld & Mitchell, L. alleni Gauld & Mitchell, L. nodus Gauld &Mitchell, L. samari Gauld & Mitchell, L. samuelsoni Gauld & Mitchell, L. quorus Gauld & Mitchell andL. longiventris Cameron. These are the characteristic Leptophion species of Cushman (1947) and Townes(1971) in that they possess both a specialized hind tarsal claw and modified penultimate hamulus. Mostspecies have Im-cu very strongly sinuous and Rs in the hind wing strongly bowed. This is the mostwidespread of all groups with species throughout the Indo-Pacific region. One species has colonizedtropical Australia (Gauld, 1984a). 154 I. D. GAULD PAMOPHION GauldPamophion Gauld, 1977: 28. Type-species: Pamophion sorus Gauld, by original designation. Mandibles not twisted, weakly narrowed apically, subequally bidentate; outer mandibular surface withsubbasal swelling, and with a diagonal groove. Maxillary palp 5-segmented, labial palp 4-segmented;clypeus in profile convex, margin impressed, acute, in anterior aspect weakly convex. Ocelli large, theposterior ones close to the eye margins; frontal carina absent; occipital carina complete except at extremeventral end. Antennae long and slender. Pronotum unspecialized; spiracular sclerite partially concealed;notauli vestigial; epicnemial carina quite strong, present on mesopleuron; mesopleural furrow unde-veloped. Scutellum weakly convex, carinate laterally most of its length; posterior transverse carina ofmesosternum incomplete, usually discernible as a discontinuous ridge. Propodeum with anterior arearather short, striate; anterior transcarina complete, other carinae absent, posterior area rugose-reticulate.Fore wing with pterostigma moderately broad; marginal cell long; Rs+2r conspicuously thickened nearpterostigma, proximally curved; discosubmarginal cell broadly glabrous anteriorly; Im-cu without aramellus, moderately sinuous. Hind wing with Rs straight; marginal cell evenly hirsute; hamuli unspecial-ized. Fore tibial spur without a membranous flange behind the macro-trichial comb; mid and hindtrochantelli with a distinct blunt distal tooth; inner hind tibial spur flattened, with a fringe of long, closehairs; hind tarsal claws unspecialized. Gaster slender; tergite 2 in profile very elongate, with thyridiamoderately close to anterior margin; umbo absent; epipleuron pendant, narrow. Ovipositor sheathslender, unspecialized. A single species, P. sorus, is known to occur in Queensland. It is apparently the most primitive species inthis genus-group. Nothing is known of its biology. The ENICOSPILUS subgroup This group contains five genera, Dicamptus, Enicospilus, Pycnophion, Banchogastra and Abanchogastra.The last three are endemic Hawaiian taxa and probably constitute a monophyletic clade which almostcertainly arose from within Enicospilus. They are so phenetically divergent, however, that it would bequite impractical to include them within Enicospilus, an otherwise structurally uniform genus of over 700species. The relationship of Enicospilus to Dicamptus is not clear; the latter may be paraphyletic with respect tothe former though subtle differences in venation, alar sclerite form, sculpture and body shape suggest thatDicamptus is actually holophyletic. The Enicospilus group is characterized by a number of apomorphic features including having thespiracular sclerite concealed, having an elongate anterior propodeal area, having a rather sinuous Rs+2rand generally having alar sclerites. Individual species may be exceptional in one or two of these features. Dicamptus, clearly the least specialized genus in this complex, is restricted to the Old World, andEnicospilus is apparently most diverse in the Old World tropics, suggesting a palaeotropical origin for thegroup. Preliminary study of Neotropical species of Enicospilus suggests that very few large species-groupsoccur in South America; several of these also occur in the Nearctic region. The Australian Dicamptus andEnicospilus have apparently been derived from immigration from South East Asia (Gauld, 1984o). The Hawaiian genera seem to be a holophyletic group. This is attested by the total lack of alar sclerites,possession of a straight, rather slender Rs+2r and loss of the last 0-3 or so of the lateral scutellar carinae.There are a number of other similarities in sculpture and exact position of wing veins that further supportsthis clade. The inter-relationship of these three genera may be misrepresented in the cladogram (Fig. 37).Considering only the characters in the primary data matrix five (16-2, 27-3, 37, 48-1 and 49) supportPycnophion + Abanchogastra whilst three (13, 42-2, 43-1) support the group Pycnophion + Banchogastra.None supports the third combination. Of the first five characters 16-2 has almost certainly undergonereversal in Banchogastra as the character is apomorphic in the greater part of the presumed ancestrallineage of the species. The plesiomorphic state is otherwise generally only found in primitive ophioninesand (again as a reversal) in a few diurnal specialized species (Banchogastra has not been taken at light andthus may well be diurnally active). Character 27-3, a centrally interrupted posterior mesosternal carina, isonly found in the primary data set in Pycnophion and Abanchogastra, but this carina is weak centrally inBanchogastra and may even be absent narrowly in some specimens. A medioventrally evanescentmesosternal carina can therefore be considered an apomorphy of the Hawaiian genera. Character 49,position of Cu\ in the fore wing, is plesiomorphic only for Banchogastra. The derived condition musttherefore be considered to be an apomorphy of the subfamily and the anomalous inclination of this vein inBanchogastra is presumably a further specialization, perhaps resulting from the very close proximity of the PHYLOGENY OF THE OPHIONINAE 155 Fig. 37 Putative phylogenetic arrangement of Hawaiian genera in relation to Enicospilus moea Chees-man. This cladogram is supported by the following apomorphic features: 1, petiolar spiracles anterior tomargin of sternite 1; 2, tergite 2 depressed; 3, epipleuron 2 pendant; 4, further reduction in ocellar size;5, possession of an elongate ovipositor; 6, loss of hair in discosubmarginal cell centrally; 7, developmentof a stout gaster; 8, possession of large thyridia close to anterior margin of tergite; 9, metapleuroninflated; 10, epicnemial carina medioventrally incomplete; 11, possession of a short, rounded prop-odeum; 12, presence of incipient frontal carinae; 13, possession of inflated hind trochanters; 14, loss ofoccipital carina mediodorsally; 15, upper tooth of mandible compressed; 16, increase in torsion ofmandible; 17, reduction in size of ocelli; 18, loss of fenestra; 19, development of fine, granulate thoracicsculpture; 20, loss of alar sclerites; 21, reduction in length of second abscissa of Cui in fore wing; 22,medially evanescent posterior transverse carina of mesosternum ; 23 , straight and slender Rs +2r; 24, lossof posterior part of lateral carina of scutellum. bases of \m-cu and Cw la . In some individuals this abscissa of Cui is occluded. The remaining twoapomorphies (37 and 48-1) are venational features which are usually rather variable. The three apomor-phies uniting Pycnophion with Banchogastra are slightly more convincing, particularly 42-2, the ptero-stigma shape. Furthermore, Pycnophion and Banchogastra resemble each other in many other specializedfeatures. Both have smaller ocelli than normal, have a medioventrally interrupted epicnemial carina, have 156 I. D. GAULD a strongly inflated metapleuron and a rather short anterior propodeal area and possess rather inflatedtrochanters. The thyridia of tergite 2 are large and close to the anterior margin (? a reversal) and the gasteris stouter than most other ophionines. The first segment of the gaster is very much broader and shorter thanthe corresponding segment in other enicospilines. In view of these marked similarities it would seem thatPycnophion + Banchogastra constitute a distinct clade. A Marquesan species, Enicospilus moea Cheesman, shares a number of apomorphies with the Hawaiiangenera, including possession of a straight, rather slender Rs+2r, having a medioventrally obsolescentposterior mesosternal carina and having only the anterior 0-7 of the scutellum carinate. Like the Hawaiiangenera the abscissa of Cui between Im-cu and Cw la is very short and cu-a is well proximal to the base ofRs&M. This species does, however, possess fenestra but this, rather than any of the extant Hawaiianspecies, may be closest to the base of the Pycnophion/ Banchogastra/ Abanchogastra evolutionary line (Fig.37). ABANCHOGASTRA PerkinsAbanchogastra Perkins, 1902: 141. Type-species: Abanchogastra debilis Perkins, by monotypy. Mandibles twisted about 45, strongly and evenly narrowed with teeth subequal, slightly depressed; outermandibular surface almost flat. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profileweakly convex, margin blunt; clypeus in anterior aspect weakly convex. Ocelli large, the posterior onesseparated from eye by 0-1-0-2 times their own diameter; frontal carina absent; occipital carina present,mediodorsally interrupted, ventrally weak but joining hypostomal carina. Antennae incomplete. Pro-notum unspecialized; spiracular sclerite concealed by pronotal flange; notauli absent; epicnemial carinaweak but present on mesopleuron; mesopleural furrow absent. Scutellum weakly convex, carinate laterallyabout 0-7 of its length; posterior transverse carina of mesosternum centrally obsolescent. Propodeum withanterior area long, striate, carinae absent, posterior area coriaceous. Fore wing with pterostigma ofmoderate width; marginal cell long; Rs+2r straight and slender; discosubmarginal cell uniformly hirsute;Im-cu evenly curved, without a ramellus. Hind wing with Rs weakly curved. Fore tibial spur without amembranous flange behind macrotrichial comb; mid and hind trochantelli unspecialized; inner hind tibialspur flattened, hind tarsal^claws unspecialized. Gaster slender; tergite 2 very elongate, thyridia elliptical,remote from anterior margin, umbo absent; epipleuron up-turned. Ovipositor sheath slender. This genus contains a single Hawaiian species which differs strikingly from any Enicospilus in the forewing venation. Nothing is known of its biology. BANCHOGASTRA AshmeadBanchogastra Ashmead, 1900: 87. Type-species: Banchogastra nigra Ashmead, by original designation. Mandibles twisted about 25, strongly and evenly narrowed apically with upper tooth a little longer than thelower; outer mandibular surface with strong proximal concavity, remainder of surface virtually flat but withdiagonal tract of dense hair. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profileweakly convex, margin sharp, not impressed at all; clypeus in anterior aspect with margin very weaklyconvex. Ocelli small, the posterior ones separated from eyes by more than their own minimum diameter;frontal carina very weak but discernible; occipital carina complete, ventrally not quite reaching hypo-stomal carina. Antennae moderately slender, about 1-2 times length of fore wing. Pronotum unspecialized;spiracular sclerite completely occluded by pronotal flange; notauli weak; epicnemial carina strong, presenton mesopleuron but medioventrally interrupted; mesopleural groove virtually absent. Scutellum convex,carinate laterally about 0-8 of its length; posterior transverse carina of mesosternum complete, or slightlyweak medioventrally. Propodeum with anterior area moderately short but clearly discernible; anteriortransverse carina complete; posterior transverse carina vestigial, others absent; posterior area rugose.Fore wing with pterostigma moderately slender; marginal cell long; discosubmarginal cell evenly hirsute;Rs+2r straight, expanded slightly immediately before joining pterostigma; Im-cu evenly arcuate, ramellusabsent; unusual in having Im-cu and Cw la basally contiguous or united. Hind wing with Rs almost straight.Fore tibial spur without a membranous flange behind macrotrichial comb; mid and hind trochantelli shortand stout, otherwise unspecialized; inner hind tibial spur slightly flattened, with a fringe of fine hairs; hindtarsal claws long but closely pectinate. Gaster stout; tergite 2 depressed, barely longer than deep, notlonger than broad; thyridia large, oval, close to anterior margin; umbo absent; laterotergite narrow,pendant. Ovipositor sheath short, slender. Banchogastra is a small genus containing two Hawaiian species (Townes, 1971). Nothing is known oftheir biology. PHYLOGENY OF THE OPHIONINAE 157 Figs 38-^t3 Stereoscan photographs of ophionines: 38-41, mandibles of (38) Ophion sp.; (39) Ophio-gastrella sp.; (40) Enicospilus sp.; (41) Leptophion alleni Gauld & Mitchell. 42, 43, hind corner ofpronotum and spiracular sclerite of (42) Ophion sp. ; (43) Enicospilus sp. 158 I. D. GAULD Figs 44-49 Stereoscan photographs of Ophioninae: 44, mandibles of Thyreodon sp. 45-47, hind tarsalclaws of (45) Ophion sp. ; (46) Leptophion sp. $ ; (47) Ophiogastrella sp. cf . 48, 49, fore tibial spurs of(48) Ophion sp.; (49) Enicospilus sp. PHYLOGENY OF THE OPHIONINAE 159 Figs 50-52 Stereoscan photographs of dorsal region of posterior part of alitrunk of: 50, Ophion sp. ; 51,Enicospilus sp.; 52, Thyreodon sp. DICAMPTUS SzepligetiDicamptus Szepligeti, 1905: 21. Type-species: Dicamptus giganteus Szepligeti, by monotypy. Mandibles not twisted, generally very weakly narrowed, almost equally bidentate; outer mandibularsurface weakly convex with proximal concavity, sometimes with pronounced proximal swelling anddiagonal hirsute groove. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profile from flatto convex or even pyramidal, usually with margin impressed and acute; clypeus in anterior aspect truncateor weakly convex. Ocelli generally large, the posterior ones usually virtually contiguous with eyes except ina few Afro tropical species; frontal carina absent; occipital carina usually complete, ventrally joininghypostomal carina. Antennae moderately to extremely elongate, from slightly longer than fore wing to 2-0times fore wing length and with upwards of 90 flagellar segments. Pronotum unspecialized; spiracularsclerite virtually completely concealed by pronotal lobe; notauli vestigial or absent; epicnemial carinausually complete, often strong; mesopleural furrow generally absent. Scutellum weakly to moderatelyconvex, with lateral longitudinal carinae strong, usually complete; posterior transverse carina of meso-sternum complete. Propodeum with anterior area long, striate; anterior transverse carina usually completeand other carinae obsolescent or absent; posterior area usually coarsely reticulate. Fore wing withpterostigma fairly slender; marginal cell very long; Rs+2r broadened and variously sinuate before joiningpterostigma; discosubmarginal cell with a large glabrous fenestra near anterior corner, this fenestrabearing at least one sclerite (proximal one); Im-cu from arcuate to sinuous, usually without a ramellus.Hind wing with Rs virtually straight. Fore tibial spur without a membranous flange behind macrotrichialcomb; mid and hind trochantelli unspecialized; inner hind tibial spur flattened, with a margin of long closehairs; hind tarsal claws unspecialized. Gaster slender; tergite 2 very elongate, with thyridia remote fromanterior margin; umbo absent; epipleuron up-turned. Ovipositor sheath slender. Dicamptus is a moderate-sized genus with 27 described species and two undescribed species (ANIC;BMNH) in the Palaeotropical region. A few species extend into temperate areas, D. fuscicornis (Erichson)reaches Tasmania and D. nigropictus (Matsumura) occurs northwards as far as Japan and Korea. Two quite distinct species-groups are recognizable. The pulchellus group comprises gracile species with avery small cubital index (0-55 or less), a short vein 3r-m (less than 0-5 of the length of M between 2m-cuand 3r-ra) and a fairly well-developed mandibular groove. This group contains D. collessi Gauld, D.uptoni Gauld and Dicamptus species 1 (ANIC) which are Australian endemics, D. indicus Nikam and D.fuscicornis which are widespread Indo- Australian species, D. isshikii, an Oriental species and theAfrotropical species D. crassellus (Morley), D. xhosa Delobel, D. betsileo Delobel, D. pellucidus(Kriechbaumer), D. seyrigi Delobel, D. townesi Delobel and D. pulchellus (Morley). The remaining species (the giganteus group) are in general very much larger insects with a larger cubitalindex, a longer 3r-m and stouter mandibles. This may well be a paraphyletic grade-group, as almost 160 I. D. GAULD certainly the characters defining it are plesiomorphic for the genus. The majority of species in this groupoccur in areas with a pronounced dry season , particularly parts of Africa and India. Some of the very largestspecies (reticulatus (Cameron), nigropictus (Matsumura) and giganteus Szepligeti) occur in montaneforest, a very wet habitat. None of the species in the giganteus group occur east of Weber's line. In NewGuinea they seem to have been replaced by some exceptionally large species of Enicospilus (E. enormousGauld & Mitchell species complex). ENICOSPILUS Stephens Enicospilus Stephens, 1835: 126. Type-species: Ophion merdarius Gravenhorst sensu Stephens (=Ichneumon ramidulus L.), by subsequent monotypy, Stephens, 1845. Henicospilus Agassiz, 1846: 138. [Unjustified emendation.] Allocamptus Foerster, 1869: 150. Type-species: Ophion undulatus Gravenhorst, by subsequent desig-nation, Thomson, 1888: 1189. Dispilus Kriechbaumer, 18946: 309. Type-species: Ophion (Dispilus) natalensis Kriechbaumer, bymonotypy. Pleuroneurophion Ashmead, 1900: 86. Type-species: Pleuroneurophion hawaiiensis Ashmead, by originaldesignation. Cymatoneura Kriechbaumer, 1901a: 22. Type-species: Ophion undulatus Gravenhorst, by subsequentdesignation, Viereck, 1914: 8. Pterospilus Kriechbaumer, 1901c: 156. Type-species: Ophion (Enicospilus) dubius Tosquinet, by subse-quent designation, Viereck, 1914: 126. [Junior homonym of Pterospilus Rondani, 1856.] Trispilus Kriechbaumer, 1901c: 156. Type-species: Ophion (Enicospilus) trimaculatus Tosquinet (=Henicospilus seminiger Szepligeti), by monotypy. Metophion Szepligeti, 1905: 28. Type-species: Metophion bicolor Szepligeti, by subsequent designation,Viereck, 1914: 94. Ceratospilus Szepligeti, 1905: 28. Type-species: Ceratospilus biroi Szepligeti, by monotypy. Atoponeura Szepligeti, 1905: 34. Type-species: Atoponeura concolor Szepligeti (= Enicospilus ato-poneurus Cushman), by monotypy. Ophiomorpha Szepligeti, 1905: 34. Type-species: Ophion curvinervis Cameron (= Enicospilus cameroniiDalla Torre), by subsequent designation, Hooker, 1912: 134. [Junior homonym of OphiomorphaNilsson, 1836.] Cryptocamptus Brethes, 1909: 230. [Unnecessary replacement name for Allocamptus Foerster.] Eremotyloides Perkins, 1915: 530. Type-species: Eremotyloides orbitalis Ashmead, by monotypy. Amesospilus Enderlein, 1918: 222. Type-species: Ophion unicallosus Snellen, by original designation. Schizospilus Seyrig, 1935: 79. Type-species: Schizospilus divisus Seyrig, by original designation. Mandibles twisted from 10 to 90, weakly to very strongly narrowed, from equally bidentate to with uppertooth conspicuously the longer, rarely with lower tooth the longer; outer mandibular surface usually with aproximal concavity, the remainder either almost flat or with a diagonal hirsute groove. Maxillary palp5-segmented, labial palp 4-segmented; clypeus in profile from flat to nasute, with margin blunt or acute oracute and subapically impressed; clypeus in anterior aspect usually weakly convex, rarely truncate orjoncave, in one Neotropical species with an indication of a median tooth. Ocelli usually large, generallywith the posterior ones very close to or contiguous with the eyes, in a few species with the ocelli smaller;frontal carina absent; occipital carina usually complete, sometimes mediodorsally obsolescent or inter-rupted. Antennae usually more than 1-5 times as long as the fore wing. Pronotum usually unspecialized, insome Madagascan species with median transverse crests; spiracular sclerite almost always completelyoccluded by flange of pronotum; notauli vestigial or absent; epicnemial carina usually well developed,present on mesopleuron, usually reaching to level of lower corner of pronotum; mesopleural furrowusually absent. Scutellum from almost flat to strongly convex, almost always with strong lateral carinaevirtually entire, rarely with these carinae short; posterior transverse carina of mesosternum usuallycomplete. Propodeum with anterior area moderately long to long, generally striate; anterior transversecarina of propodeum usually present, almost always complete; posterior transverse carina usually absent,rarely present laterally; posterior area usually reticulate or rugose, sometimes finely irregularly wrinkled,in a few species concentrically striate, very rarely the posterior area almost smooth or punctate. Fore wingwith pterostigma moderately broad to quite slender; marginal cell long; Rs+2r almost always broadenedand variously sinuate before joining the pterostigma; discosubmarginal cell usually with a glabrous fenestraadjacent to the vein Rs+2r, this fenestra often bearing one or more detached sclerites; Im-cu usuallyarcuate or sinuous, rarely somewhat angulate; ramellus usually absent, rarely in one or two specieswell-developed. Hind wing with Rs straight or weakly bowed. Fore tibial spur without a membranous PHYLOGENY OF THE OPHIONINAE 161 flange behind macrotrichial comb; mid and hind trochantelli usually simple, rarely with a marginal tooth;inner hind tibial spur flattened, with margin of close long hairs; hind tarsal claws various, most usuallyunspecialized and slightly sexually dimorphic, sometimes extremely sexually dimorphic, the male having avery fine pectinal comb, in a few taxa with the claws strongly geniculate, occasionally with the clawsasymmetrical, in other taxa with the distal pectina projecting beyond the apex of the true claw, and veryrarely, in a few eremic species, with the pectina reduced. Gaster from moderately stout to very slender;tergite 2 usually very long and slender, rarely posteriorly almost as deep as long, almost always withthyridia remote from anterior margin; umbo vestigial or absent; epipleuron usually up- turned, pendant in afew (or one) Neotropical species. Ovipositor sheath slender; ovipositor usually short, straight, in isolatedtaxa it may be upcurved or decurved, and very rarely it may be straight but project well beyond the apex ofthe gaster. Enicospilus is an extremely large genus, most species of which occur in the tropics. The major centres ofradiation appear to be New Guinea, where there are about 200 endemic species, and Madagascar, whichhas about half that number. The Philippines, Hawaii and Zaire river basin are other areas of pronouncedendemicity. In the tropics, the greatest diversity of species seems to occur at mid-elevation in the cloudforests (moss forests) between 1200 and 2000 m. A number of species are restricted to areas that have apronounced dry season (e.g. Enicospilus capensis), and a few species seem to" inhabit deserts (e.g. E.psammus). Many species are capable of sustained flight over great distances (e.g. across the Tasman Seafrom Australia to New Zealand) and several occur on almost all South Pacific archipelagos as far east as theTuamotus. Most small oceanic islands apparently have several species of Enicospilus present, and on theMicronesian islands six of the 33 recorded ichneumonids are Enicospilus (Townes, 1958). Outside thetropics there are notably fewer species. Scaramozzino (1983) records only ten from Italy whilst Viktorov(1957) knew of only 15 species from the Soviet Union. Five nominal species are recorded from Britain, butprobably only four species occur there (Fitton et al., 1978). Certain generalized distribution patterns can be observed in the genus. New Guinea and Madagascarhave large numbers of endemic species in endemic species-groups; for example eight of the 24 species-groups recognized as occurring in the Afrotropical region are restricted to Madagascar, and these eightgroups contain 35 species (Gauld & Mitchell, 1978). The Philippines, Hawaii and South America eachcontain very few species-groups, but these may be very large. Relatively few of the numerous otherspecies-groups are endemic to a single zoogeographic area. Several of the species-groups that arewidespread in South East Asia have their greatest diversity in New Guinea (e.g. the xanthocephalus andtremulus species-groups, Gauld & Mitchell, 1981). Virtually no widespread South East Asian groups havea localized centre of diversity outside Melanesia; most species seem to be widespread with isolated localendemics (e.g. theflavicaput species-group, Gauld & Mitchell, 1981). Many widespread South East Asiaagroups are also represented in mainland Africa (e.g. the capensis and antefurcalis species-groups). Anumber of species-groups are either endemic to mainland Africa (e.g. the babaultii, biimpressus andrubens species-groups) or are most diverse in Africa (e.g. the dolosus species-group). Most speciesoccurring in the Palaearctic region belong to species-groups that are well-represented in the Nearcticregion, but neither region appears to have any endemic species-groups, nor is any species-group endemicto the Holarctic region. The Enicospilus species of Australia are virtually all either widespread South EastAsian species or are endemics derived from South East Asia (Gauld, 1984a). New Zealand has no endemicspecies, but shares two with Australia. A distinct faunal region is apparent which comprises part of theMediterranean basin, most of the Middle East and extends eastwards into Central Asia and south-eastwards into north-west India. Although the fauna of this area is poorly known, it seems that at least onespecies-group may be endemic to the region (the przewalskii species-group). Gauld & Mitchell (1978; 1981) outline a very large number of species-groups and several of these arerefined by Gauld (1984a). Large numbers of other species are currently unplaced; these may belong to lessclearly definable groups, or they may be aberrant members of existing groups, or they may representnumerous monobasic species-groups. The phylogenetic inter- relationships of most of these species-groupsare very difficult to assess as most are definable on the basis of a number of autapomorphies. Few shareobvious specializations with other species-groups. Repeated attempts have been made to subdivide Enicospilus into a number of smaller genera(Kriechbaumer, 1901c; Szepligeti, 1905; Seyrig, 1935) but none of these subdivisions has endured, largelyfor two reasons - the authors proposing the separation have very limited experience of the range ofmorphological diversity afforded by the genus, and the characters used to effect separation are superficialdifferences, usually in the number of alar sclerites. The clearest demonstration of the great variability ofalar sclerite form and number can be seen between the closely related species of the unidens species-group,a group definable by several autapomorphies. The proximal sclerite is present in all species but only E.unidens has a central sclerite; the distal sclerite is present in E. unidens, E. akainus and E. mirax but absent 162 I. D. GAULD in E. gonidius and E. amygdalis. The most important features for recognizing apparently 'natural'groupings seem to be the form of the tarsal claws (including the degree of sexual dimorphism), the structureof the mandibles and clypeus, the presence or absence of the proximal sclerite and the structure of the malegenitalia. Little is known of the biology of species of Enicospilus. The majority of host records refer to speciesparasitizing larvae of Noctuidae, Geometridae, Lymantriidae or Saturniidae. A few species attackpyralids. Many of the larger species that attack saturniid larvae spin their cocoon within the host cocoon,but other species spin a cocoon that is not enclosed by that of the host. Most species seem to attack larvaethat are free-living, tree-leaf-feeding caterpillars, but the few with longer ovipositors seem to attack larvaemining stems (e.g. E. terebrus). It must be stressed that the hosts of the majority of tropical species are notknown. PYCNOPHION AshmeadPycnophion Ashmead, 1900: 87. Type-species: Pycnophion molokaiensis Ashmead, by monotypy. Mandibles twisted about 20, evenly tapered, with upper tooth broader but of about equal length to thelower tooth; outer maridibular surface slightly concave, sparsely pubescent. Maxillary palp 5-segmented,labial palp 4-segmented; clypeus in profile moderately convex, margin blunt; clypeus in anterior aspectweakly convex. Ocelli moderately large, the posterior ones separated from the eye by about 0-5 times theirown diameter; frontal carina very weak; occipital carina complete, mediodorsally with a depression.Antennae moderately long, about 1-3 times the length of the fore wing. Pronotum slightly flattenedmediodorsally; spiracular sclerite covered by pronotal flange; notauli vestigial; epicnemial carina strong,curved to nearly reach the anterior margin of the pleuron above the level of the lower corner of thepronotum; mesopleural furrow absent. Scutellum convex, laterally carinate at least 0-6 of its length;posterior transverse carina of the mesosternum centrally obsolescent. Propodeum with anterior areamoderately short but clearly discernible; anterior transverse carina complete, at least centrally, theposterior one vestigial, the other absent; posterior area finely coriaceous. Pterostigma moderately slender;marginal cell long; Rs+2r virtually straight, evenly but abruptly widened before joining pterostigma;discosubmarginal cell with an ill-defined glabrous area anteriorly, the entire cell very sparsely hirsute;Im-cu fairly evenly arcuate, ramellus absent. Hind wing with Rs straight. Fore tibial spur without amembranous flange behind macro trichial comb; mid and hind trochantelli unspecialized; inner hind tibialspur somewhat flattened, with a fringe of long hairs; tarsal claws unspecialized. Gaster moderately stout;tergite 2 in profile, slightly longer than posteriorly deep, with thyridia large, oval and close to anteriormargin; umbo vestigial; epipleuron up-turned. Ovipositor exceptional in being very long and up-curved,reaching beyond apex of gaster by at least length of tergites 2-5. Pycnophion is a small genus containing three Hawaiian species (Townes etal., 1961). It is one of the mostatypical of ophionine genera and females superficially resemble campoplegines or phygadeuontines. Onespecies, P. fuscipennis Perkins, has been reared as fuparasite of the larvae of Hyposmocoma chilonella(Lepidoptera: Cosmopterygidae) (Swezey, 1931) which are borers in the pith of stems of Rubus, Acaciaand a variety of other woody plants. Acknowledgements Much of this work was undertaken as part of a PhD thesis under the supervision of Dr Garth Underwood. Iam extremely grateful for his advice at all stages of the study. I would like to thank Dr J. Felsenstein forproviding a copy of his program PHYLIP, and the staff of the computing centre at the City of LondonPolytechnic for their assistance with computational problems. I am grateful to many of my colleagues at theBritish Museum (Natural History) for their opinions on various problems and comments on various drafts,though the conclusions expressed do not necessarily reflect their points of view; in particular I would like tothank Dr A. D. Austin, Mr B. Bolton, Mr M. C. Day, Dr M. G. Fitton, Dr J. D. Holloway, Dr L. A.Mound, Dr G. S. Robinson, Mr R. I. Vane- Wright and Dr P. Whalley. I am grateful to the following forlending me particularly interesting specimens from the collections in their care: Dr D. Kasparayan, Dr H.Townes and Dr L. Zombori, and I am particularly grateful to all those entomologists, too numerous tomention, who have sent me ophionines from all over the world. Finally I would like to thank Mr J. Carterfor taking the stereoscan pictures and Ms P. A. Mitchell for typing the manuscript. PHYLOGENY OF THE OPHIONINAE 163 References Achterberg, C. van 1976. A preliminary key to the subfamilies of the Braconidae (Hymenoptera).Tijdschrift voor Entomologie 119: 33-78. 1984. Essay on the phylogeny of Braconidae (Hymenoptera: Ichneumonoidea). Entomologisk Tidskrift 105: 41-58. Agassiz, J. L. R. 1846. Nomendator zoologicus, Index Universalis. 1135 pp. Soloduri. Ashmead, W. H. 1900. Classification of ichneumon-flies or the superfamily Ichneumonoidea. Proceedings of the United States National Museum 23: 1-220.Audley-Charles, M. G., Hurley, A. M. & Smith, A. G. 1981. Continental movements in the Mesozoic and Cenozoic. In Whitmore, T. C. (Ed.), Wallace's Line and Plate Tectonics. 91 pp. Oxford.Barron, J. R. 1978. Systematics of the World Eucerotinae (Hymenoptera, Ichneumonidae) Part 2. Non-Nearctic species. Naturaliste Canadien 105: 327-374.Benson, R. B. 1935. On the genera of Cephidae, and the erection of a new subfamily Syntexidae. Annals & Magazine of Natural History (10) 16: 535-553.Bolton, B. 1982. Afrotropical species of the myrmicine ant genera Cardiocondyla, Leptothorax, Melisso- tarsus, Messor and Cataulacus (Formicidae). Bulletin of the British Museum (Natural History) (Ento-mology) 45: 307-370.Brauns, S. 1889. Die Ophionoiden. Archiv des Vereins der Freunde der Naturgeschichte in Mecklenburg 43: 58-100. 1895. Descriptions specierum novarum Ichneumonidarum e fauna Hungarica. Termeszetrajzi Fuzetek 18: 42-49. Brethes, J. 1909. Hymenoptera Paraguayensis. Anales del Museo National de Historia Natural de Buenos Aires 19: 225-256.Brock, J. P. 1982. A systematic study of the genus Ophion in Britain. Tijdschrift voor Entomologie 125: 57-97.Brues, C. T. 1912. Brazilian Ichneumonidae and Braconidae obtained by the Stanford expedition. Annals of the Entomological Society of America 5: 193-228.Brulle, M. A. 1846. In Lepeletier de Saint-Fargeau, A., Histoire naturelle des insectes 4. Hymenopteres. vii+680pp. Paris.Butler, P. M. 1982. Directions of evolution in the mammalian dentition. In Joysey, K. A. & Friday, A. E. (Eds), Problems of Phylogenetic Reconstruction, xi+442 pp. London.Cameron, P. 1901. On Hymenoptera collected in New Britain by Dr Arthur Willey. Proceedings of the Zoological Society of London 1901: 224-248.1903. Descriptions of new genera and species of Hymenoptera taken by Mr Robert Shelford at Sarawak, Borneo. Journal of the Straits Branch of the Royal Asiatic Society 39: 89-181.1905. On the phytophagous and parasitic Hymenoptera collected by Mr Ernest Green in Ceylon. Spolia Zeylanica 3: 67-144. 1906. Descriptions of new species of parasitic Hymenoptera chiefly in the collection of the South African Museum, Cape Town. Annals of the South African Museum 5: 17-186.Carlson, R. W. 1979. Family Ichneumonidae. In Krombein, K. V., Hurd, P. D., Smith, D. R. & Burks, B. D. , Catalog of Hymenoptera in America North of Mexico 1: 1-1198.Carpenter, J. M. 1982. The phylogenetic relationships and natural classification of the Vespoidea (Hymenoptera). Systematic Entomology 7: 11-38.Clausen, C. P. 1940. Entomophagous Insects. 688 pp. New York.Curtis, J. 1836. British Entomology 13: plates 578-625. London.Cushman, R. A. 1926. Some types of parasitism among Ichneumonidae. Proceedings of the Entomological Society of Washington 28: 25-51.1944. The Hawaiian species of Enicospilus and Abanchogastra (Hymenoptera, Ichneumonidae). Proceedings of the Hawaiian Entomological Society 12: 39-56. 1947. A generic revision of the ichneumon-flies of the tribe Ophionini. Proceedings of the United States National Museum 96: 417-482.Darlington, P. J. 1965. The biogeography of the southern end of the world, vii+236 pp. Cambridge, Massachusetts.Davies, R. G. 1981. Information theory and character selection in the numerical taxonomy of some male Diaspididae (Hemiptera: Coccoidea). Systematic Entomology 6: 149-178.Davies, R. G. & Boratyhski, K. L. 1979. Character selection in relation to the numerical taxonomy of some male Diaspididae (Homoptera: Coccoidea). Biological Journal of the Linnean Society 12: 95-165. 164 I. D. GAULD Day, W. H. E. 1983. Computationally difficult problems in phylogenetic systematics. Journal of Theoretical Biology 103: 429-438.Delobel, A. 1975. Deux nouveaux Rhopalophion Seyrig appartenant a la faune malgache. Bulletin de la Societe entomologique de France 80: 43-47. Enderlein, G. 1912. Beitrage zur Kenntnis aussereuropaischer Ichneumoniden II. Ophionoiden. DerGattung Thyreodon und ihre Verwandten. Zoologischer Anzeiger 39: 624-632.- 1918. Ichneumoniden. In Michaelson, W., Beitrage zur Kenntnis der Land und SusswasserfaunaDeutsch-Sudwestafrikas 2(4). 22 pp. Jena. 1921. Beitrage zur Kenntnis aussereuropaischer Ichneumoniden. Stettiner Entomologische Zeitung 82: 1-45.Fabricius, J. C. 1798. Entomologia systematica emendata et aucta . . . adjectis synonymis, locis obser- vationibus, descriptionibus. Supplementum. 572pp. Halfniae.Farris, J. S. 1969. A successive approximation approach to character weighting. Systematic Zoology 18: 374-385. 1979. The information content of the phylogenetic system. Systematic Zoology 28: 483-519. Felsenstein, J. 1981 . A likelihood approach to character weighting and what it tells us about parsimony and compatibility. Biological Journal of the Linnean Society 16: 183-196. 1982. Numerical methods for inferring evolutionary trees. Quarterly Review of Biology 57: 379-404.Felt, E. P. 1904. Nineteenth report of the state entomologist, 1903. Beneficial insects. Bulletin of the New York State Museum 76: 97-125.Fitton, M. G. & Gauld, I. D. 1976. The family-group names of the Ichneumonidae (excluding the Ichneumoninae). Systematic Entomology 1: 247-258.Fitton, M. G., Graham, M. W. R. de V., Boucek, Z. R. J., Fergusson, N. D. M., Huddleston, I., Quinlan, J. & Richards, O. W. 1978. Kloet and Hincks, a check list of British Insects. Part 4: Hymenoptera. Handbooks for the Identification of British Insects 11(4): 1-159.Foerster, A. 1869. Synopsis der Familien und Gattungen der Ichneumoniden. Verhandlungen des Naturhistorischen Vereins der Preussischen Rheinlande und Westfalens 25: 135-221.Friday, A. E. 1982. Parsimony, simplicity and what actually happened. Zoological Journal of the Linnean Society 74: 329-335.Gaffney, E. S. 1979. An introduction to the logic of phylogeny reconstruction. In Cracraft, J. & Eldredge, N. (Eds), Phylogenetic Analysis and Palaeontology, ix+233 pp. New York.Gauld, I. D. 1976. The classification of the Anomaloninae (Hymenoptera: Ichneumonidae). Bulletin of the British Museum (Natural History) (Entomology) 33: 1-135.1977. A revision of the Ophioninae (Hymenoptera: Ichneumonidae) of Australia. Australian Journal of Zoology, Supplementary Series 49: 1-112. 1979. An analysis of the classification of the Ophion genus-group (Ichneumonidae). Systematic Entomology 5: 59-82.- 1983. The classification, evolution and distribution of the Labeninae, an ancient southern group of Ichneumonidae (Hymenoptera). Systematic Entomology 8: 167-178. 1984a. The Australian Ophioninae (Insecta; Hymenoptera): a historical biogeographic study. Journal of Biogeography 11: 269-288. 19846. An introduction to the Ichneumonidae of Australia. 413 pp. London. Gauld, I. D. & Carter, J. M. 1983. The Ophioninae of the Galapagos Islands (Hymenoptera: Ichneumoni-dae). Journal of Natural History 17: 145-155.Gauld, I. D. & Huddleston, T. 1976. The nocturnal Ichneumonoidea of the British Isles, including a key to genera. Entomologist's Gazette 27: 35-49.Gauld, I. D. & Mitchell, P. A. 1978. The taxonomy, distribution and host preferences of African parasitic wasps of the subfamily Ophioninae. 287 pp. Slough. 1981. The taxonomy, distribution and host preferences of Indo-Papuan parasitic wasps of the subfamily Ophioninae. 611 pp. Slough.Gauld, I. D. & Mound, L. A. 1982. Homoplasy and the delineation of holophyletic genera in some insect groups. Systematic Entomology 7: 73-86.Guise, A., Peacock, D. & Gleaves, T. A. 1982. A method for identification of parallelism in discrete character sets. Zoological Journal of the Linnean Society 74: 293-303.Gupta, V. K. 1962. Taxonomy, zoogeography and evolution of Indo- Australian Theronia (Hymenoptera: Ichneumonidae). Pacific Insects Monograph 1: 1-142.Hellen, W. 1926. Beitrage zur Kenntnis der Ichneumoniden Finlands. 2 Subfam. Ophioninae und Anomaloninae. Acta Societatis pro Fauna et Flora Fennica 56: 3-27.Hennig, W. 1966. Phylogenetic Systematics. xiii+263 pp. Urbana, Illinois. PHYLOGENY OF THE OPHIONINAE 165 Hooker, C. W. 1912. The ichneumon-flies of America belonging to the tribe Ophionini. Transactions of the American Entomological Society 38: 1-176.Horstmann, K. 1981. Die Palaarktischen Arten der Gattungen Eremotylus Forster und Simophion Cushman (Hymenoptera, Ichneumonidae). Entomofauna 2: 415^32.Hull, D. L. 1965. The effect of essentialism on taxonomy: two thousand years of stasis 1 . British Journal for the Philosophy of Science 15: 314-326.Kluge, A. G. 1976. Phylogenetic relationships in the lizard family Pygopodidae: an evaluation of theory, methods and data. Miscellaneous Publications, Museum of Zoology, University of Michigan 152: 1-72.Kohl, F. F. 1906. Ichneumonidae. In Penther, A. & Zederbauer, E., Ergebnisse einer naturwissenschaft- lichen Reise zur Erdschias-Dagh (Kleinasien). Annalen des Naturhistorischen (Hof) Museums Wien 20: 220-246.Konigsmann, E. 1978. Das Phylogenetische System der Hymenopteren 3. Terebrantes (Unterordnung Apocrita). Deutsche Entomologische Zeitschrift (N.F.) 25: 1-55.Kriechbaumer, J. 1894a. In Sickmann, F., Beitrage zur Kenntnis der Hymenopteren-Fauna des nordis- chen China. ZoologischeJahrbiicherSystematikS: 195-236. 18946. Hymenoptera Ichneumonidae. Berliner Entomologische Zeitschrift 39: 297-318. 1901a. Bemerkungen iiber Ophioniden. Zeitschrift fur Systematische Hymenopterologie und Dipter- ologiel: 18-24.1901ft. Bemerkungen iiber Ophioniden. Zeitschrift fur Systematische Hymenopterologie und Dipter- ologie 1:73-79. 1901c. Ueber die Gattungen der von Tosquinet in seiner Ichneumonides d'Afrique beschrieben Ophionarten. Zeitschrift fur Systematische Hymenopterologie und Dipterologie 1: 155-156.LeQuesne, W. J. 1969. A method of selection of characters in numerical taxonomy. Systematic Zoology 18: 201-205. 1972. Further studies on the uniquely derived character concept. Systematic Zoology 21: 281-288. L0vetrup, S. 1973. Classification, convention and logic. Zoologica Scripta 2: 49-61. Mackerras, I. M. 1962. Speciation in Australian Tabanidae. In Leeper, G. W. (Ed.), The Evolution of Living Organisms. 459 pp. Melbourne.Martin, R. 1981. Phylogenetic reconstruction versus classification: the case for clear demarcation. Biologist 28: 127-132. Mickevich, M. F. 1978. Taxonomic congruence. Systematic Zoology 27: 143-158.Middlekauf, W. W. 1983. A revision of the sawfly family Orussidae for North and Central America (Hymenoptera; Symphyta, Orussidae). University of California Publications in Entomology 101: 1-46.Morley, C. 1912. A revision of the Ichneumonidae ... 1. ix+88 pp. London.1913. The fauna of British India, including Ceylon and Burma. Hymenoptera 3, Ichneumonidae, 1 Ichneumones Deltoidei. 532 pp. London.Moutia, L. A. & Curtois, C. M. 1952. Parasites of the moth borers of sugarcane in Mauritius. Bulletin of Entomological Research 43: 325-359.Nelson, G. & Platnick, N. 1981. Systematics and biogeography. Cladistics and vicariance. 567 pp. New York.Owen, H. G. 1981. Constant dimensions or an expanding Earth? In Cocks, L. R. M. (Ed.), The Evolving Earth. 264 pp. London & Cambridge.Pampel, W. 1913. Die weiblichen Geschlechtsorgane der Ichneumoniden. Zeitschrift fur Wissenschaftliche Zoologie 108: 290-357.Panchen, A. L. 1982. The use of parsimony in testing phylogenetic hypotheses. Zoological Journal of the Linnean Society 74: 305-328.Perkins, J. F. 1959. Ichneumonidae, key to subfamilies and Ichneumoninae 1. Handbook for the Identification of British Insects 7(2ai): 1-116.Perkins, R. C. L. 1902. Four new species and a new genus of Parasitic Hymenoptera (Ichneumonidae; Ophioninae) from the Hawaiian Islands. Transactions of the Entomological Society of London 1902: 141-143.1915. On Hawaiian Ophioninae (Hymenoptera, Ichneumonidae). Transactions of the Entomological Society of London 1914: 521-535. Pratt, V. 1972. Numerical taxonomy - a critique. Journal of Theoretical Biology 36: 581-592.Price, P. W. 1975. The evolutionary strategies of parasitic insects and mites. 244 pp. New York.Quinlan, J. & Gauld, I. D. 1981. Symphyta (except Tenthredinidae). Hymenoptera, new edition. Handbooks for the Identification of British Insects 6(2a): 1-67.Rasnitsyn, A. P. 1983. Ichneumonoidea (Hymenoptera) from the Lower Cretaceous of Mongolia. Contributions of the American Entomological Institute 20: 259-265. 166 I. D. GAULD Rich, P. V. & Rich, T. H. 1983. The Central American disperal route: biotic history and palaeogeography. In Janzen, D. H. (Ed.), Costa Rican Natural History . 816 pp. Chicago & London.Richards, O. W. 1956. Hymenoptera: introduction and key to families. Handbooks for the Identification of British Insects 6(1): 1-94.Rodendorf , B. B. 1962. Fossil Insects (English translation by H. Vaitactis. Unpublished - copy in BMNH). 358 pp.Rodendorf, B. B. & Rasnitsyn, A. P. 1980. [Historical development of the insect class.] [In Russian.] Trudy Palaeontologicheskogo Instituta 175: 1-269.Rohlf, F. J., Colless, D. H. & Hart, G. 1983. Taxonomic congruence re-examined. Systematic Zoology 32: 144-158. Rohlf, F. J. & Sokal, R. R. 1980. Comments on taxonomic congruence. Systematic Zoology 29: 97-101.Roman, A. 1943. Neue Schlupfwespen aus Ostafrika. Folium Entomologicum Festschrift zum 60- Geburtstage von Felix Bryk: 20-23.Rothschild, W. & Jordan, K. 1903. A revision of the lepidopterous family Sphingidae. cxxxv+813 pp. Tring. Saether, O. A. 1979. Underlying synapomorphies and anagenetic analysis. Zoologica Scripta 8: 305-312.Scaramozzino, P. L. 1983. II genere Enicospilus Steph. in Italia. Nota preliminare. Atti XIII Congresso Nazionale Italiano di Entomologia 1983: 113-117.Schrank, F. P. 1802. Fauna Boica 2(2): 977 pp. Ingolstadt. Schulz, W. A. 1911. Zweihundert alte Hymenopteren. Zoologischen Annalen (Wurzburg) 4: 1-220.Seyrig, A. 1926. Etudes sur les Ichneumonides (Hymenoptera) 1. Eos 2: 115-133. 1935. Hymenoptera 2, Ichneumonidae. Mission Scientifique de I'Omo 3(18): 1-103. Shaw, M. R. 1983. On evolution of endoparasitism: the biology of some genera of Rogadinae (Braconi- dae). Contributions of the American Entomological Institute 20: 307-328.Shestakov, A. 1926. Tabula diagnostica et species palaearcticae. Konowia 5: 256-263.Short, J. R. T. 1978. The final larval instars of the Ichneumonidae. Memoirs of the American Entomo-logical Institute 25: 1-508.Shuckard, W. E. 1840. In Swainson, W. & Shuckard, W. E. (Eds), The cabinet cyclopedia: on the history and natural arrangement of insects. iv+406pp. London. Sokal, R. R. & Sneath, P. H. A. 1963. Principles of numerical taxonomy, xv -1-573 pp. San Francisco.Stanley, S. M. 1979. Mac? revolution. Patterns and Process. 332 pp. San Francisco.Stephens, J. L. 1835. Illustrations of British Entomology Mandibulata 7. 306 pp. London. 1845. Illustrations of British Entomology Mandibulata 7. Index & list of plates. 6 pp. London. Strand, E. 1928. Miscellanea nomenclatorica zoologica et palaeontologica. Archiv fur Naturgeschichte 92A: 30-75.Strauch, J. R. 1984. Use of homoplastic characters in compatibility analysis. Systematic Zoology 33: 167-177.Swezey, O. H. 1931. Some observations on the insect faunas of native forest trees in the Olinda forest on Maui. Proceedings of the Hawaiian Entomological Society 7: 493-504. Szlpligeti, G. V. 1905. Hymenoptera, Ichneumonidae. In Wytsman, P., Genera Insectorum 34: 1-68.Taylor, R. W. 1972. Biogeography of insects of New Guinea and Cape York Peninsula. In Walker, D. (Ed.), Bridge and barrier: the natural and cultural history of Torres Strait. 437 pp. Canberra.Telenga, N. A. 1969. Origin and evolution of parasitism in Hymenoptera Parasitica and development of their fauna in the USSR (English translation). 112 pp. Jerusalem.Thomson, C. G. 1888. Ofversigt af de i Sverige funna arter af Ophion och Paniscus. Opuscula Entomo- logica 12: 1185-1201.Tosquinet, J. 1903. Ichneumonidae nouveaux. Memoires de la Societe Royale Entomologique de Belgique 10: 1^02.Townes, H. 1958. Hymenoptera: Ichneumonidae, Stephanidae and Evaniidae. Insects of Micronesia 19: 35-87. 1969. Genera of Ichneumonidae 1. Memoirs of the American Entomological Institute 11: 1-300. 1970. Genera of Ichneumonidae 3. Memoirs of the American Entomological Institute 13: 1-307. 1971. Genera of Ichneumonidae 4. Memoirs of the American Entomological Institute 17: 1-372. 1973a. Two ichneumonids (Hymenoptera) from the early Cretaceous. Proceedings of the Entomo-logical Society of Washington 75: 216-219. 19736. Three tryphonine ichneumonids from Cretaceous amber (Hymenoptera). Proceedings of the Entomological Society of Washington 75: 282-287.Townes, H., Momoi, S. & Townes, M. 1965. A catalogue and reclassification of Eastern PalaearcticIchneumonidae. Memoirs of the American Entomological Institute 5: 1-661. PHYLOGENY OF THE OPHIONINAE 167 Townes, H. & Townes, M. 1966. A catalogue and reclassification of Neotropic Ichneumonidae. Memoirs of the American Entomological Institute 8: 1-367.1973. A catalogue and reclassification of Ethiopian Ichneumonidae. Memoirs of the American Entomological Institute 19: 1-416.Townes, H., Townes, M. & Gupta, V. K. 1961. A catalogue and reclassification of Indo- Australian Ichneumonidae. Memoirs of the American Entomological Institute 1: 1-522.Uchida, T. 1928. Zweiter Beitrage zur Ichneumoniden Fauna Japans. Journal of the Faculty of Agriculture of Hokkaido Imperial University 21: 177-297.Underwood, G. 1982. Parallel evolution in the context of character analysis. Zoological Journal of the Linnean Society 74: 245-266.Viereck, H. L. 1912. Contributions to our knowledge of bees and ichneumon-flies including descriptions of twenty-one new genera and fifty-seven new species of ichneumon-flies. Proceedings of the United States National Museum 42: 613-648.1914. Type-species of the genera of ichneumon-flies. Bulletin of the United States National Museum 31: 1-186.Viktorov, G. A. 1957. Species of the genus Enicospilus Stephens in USSR. Entomologicheskoe Obozrenie 36: 179-210.Watrous, L. E. & Wheeler, Q. D. 1981. The out-group comparison method of character analysis. Systematic Zoology 30: 1-11.Westwood, J. O. 1882. Descriptions of new or imperfectly known species of Ichneumones Adsciti. Tijdschrift voor Entomologie 25: 2-48.Wilson, E. 0. 1959. Adaptive shift and dispersal in a tropical ant fauna. Evolution 13: 122-144. 168 I. D. GAULD Appendix 1 The ophionine taxa used in the cladistic study. 901 Thyreodon atricolor (Olivier) 902 Thyreodon laticinctus Cresson 903 Thyreodon fulvescens Cresson 904 Thyreodonflamminiger (Morley) 905 Simophion calvus Viktorov 906 Orientospilus melasma Townes 907 Dictyonotus purpurascens (Smith) 908 Rhynchophion flammipennis (Ashmead) 909 Lepiscelus distans (Seyrig) 910 Ophiogastrella sp. 1 (BMNH) 91 1 Ophiogastrella sp. 2 (BMNH) 912 Laticoleus unicolor (Szepligeti) 913 Laticoleus pronotalis Gauld & Mitchell 914 Laticoleus spilus Gauld & Mitchell 915 Laticoleus curvatus Delobel 916 Stauropoctonus bombycivorus (Gravenhorst) 917 Stauropoctonus occipitalis Gauld & Mitchell 918 Aulophion sp. 1 (BMNH) 919 Pamophion sorus Gauld 920 Riekophion emandibulator (Morley) 921 Ophionopsis nigrocyaneus Tosquinet 922 Sicophion pleuralis Gauld 923 Prethophion latus Townes 924 Pycnophion molokaiensis Ashmead 925 Abanchogastra hawaiiensis (Ashmead) 926 Banchogastra nigra Ashmead 927 Barytatocephalus mocsaryi (Brauns) 928 Leptophion maculipenhis (Cameron) 929 Leptophion anici Gauld 930 Leptophion pterospilus Gauld & Mitchell 931 Leptophion tetus Gauld 932 Dicamptus neavei Gauld & Mitchell 933 Dicamptus giganteus Szepligeti 934 Dicamptus fuscicornis (Erichson) 935 Euryophion latipennis (Kirby) 936 Euryophion adustus (Townes) 937 Rictophion ikuthana (Kriechbaumer) 938 Xylophion xylus (Gauld) 939 Ophion luteus (L.) 940 Eremotylus boguschi (Meyer) 941 Agathophionafulvicornis Westwood 942 Enicospilus tremulus Gauld & Mitchell 943 Enicospilus spathius Gauld & Mitchell 944 Enicospilus nephele Gauld & Mitchell 945 Enicospilus cionobius Gauld & Mitchell 946 Enicospilus unidens Seyrig 947 Enicospilus mahalonius Gauld & Mitchell 948 Rhopalophion discinervis (Morley) 949 Janzophion nebosus sp. n. 950 Hellwigiella nigripennis Szepligeti 951 Sclerophion uchidai Gauld & Mitchell Hypothetical taxa 960 Laticoleus ancestor 961 Ophiogastrella ancestor 962 Le-'ophion ancestor 963 Stauropoctonus ancestor Nearctic Neotropical Neotropical Neotropical Palaearctic Afrotropical Oriental/Eastern Palaearctic Neotropical Afrotropical Neotropical Neotropical Afrotropical Afrotropical (Madagascar) Afrotropical Afrotropical Palaearctic Afrotropical (Madagascar) Neotropical Australian Australian Afrotropical Neotropical Neotropical Hawaiian Hawaiian Hawaiian Palaearctic Oriental Australian/New Caledonian Oriental Australian Afrotropical Oriental Oriental/ Australian Afrotropical Afrotropical Afrotropical Australian/Papuan Palaearctic Palaearctic Neotropical Oriental Oriental Oriental Oriental Afrotropical Afrotropical (Madagascar) Afrotropical Neotropical Palaearctic Oriental/Eastern Palaearctic 964 Thyreodonl Dictyonotus ancestor 965 Euryophion ancestor 966 Dicamptus ancestor 967 Enicospilus ancestor PHYLOGENY OF THE OPHIONINAE 169 Appendix 2 Primary data matrix of 51 selected ophionines and 95 characters. indicates a presumedplesiomorphic condition, 1 a presumed apomorphic state. 4.1 5.1 6.1 7.1 8.1 9 11 13 14.2 15.2 3 4.2 5.2 6.2 7.2 8.2 10 12 14.1 15.1 16.1 170 I. D. GAULD Appendix 2 - cont. 16.2 18.1 19 20.2 21.2 22.2 24.1 25.1 26 27.2 28.1 29.1 17 18.2 20.1 21.1 22.1 23 24.2 25.2 27.1 27.3 28.2 29.2 PHYLOGENY OF THE OPHIONINAE 171Appendix 2 - cont. 30 32 34 36.1 36.3 38 40 42.1 43.1 44.1 45.1 46 47.2 31 33 35 36.2 37 39 41 42.2 43.2 44.2 45.2 47.1 172 I. D. GAULD Appendix 2 - cont. PHYLOGENY OF THE OPHIONINAE 173 Appendix 3 Shared derived character matrix for the 51 selected ophionines. 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 951 7 7 8 9 14 17 8 8 12 19 19 21 20 22 22 14 12 19 19 13 4 11 12 950 18 16 16 17 20 22 21 18 14 17 17 17 18 19 19 15 13 16 18 10 19 15 13 949 10 9 10 11 15 17 15 12 21 18 17 22 20 22 24 22 20 24 23 17 10 18 14 948 9 9 10 10 15 15 11 10 9 19 18 18 20 20 20 11 9 14 16 8 7 11 12 947 17 16 18 18 20 28 19 16 24 25 25 27 29 30 30 27 24 28 29 20 14 19 17 946 19 18 19 20 22 27 21 18 24 26 28 28 29 32 29 26 23 27 29 20 15 20 17 945 14 13 13 14 17 24 17 13 22 22 22 26 24 28 27 24 23 26 27 19 12 18 17 944 12 12 12 13 16 18 15 12 17 17 17 21 19 21 21 20 18 21 21 16 11 15 13 943 13 12 13 14 16 22 16 12 21 21 21 23 22 26 25 25 23 28 27 15 10 16 15 942 16 15 15 17 19 26 18 14 21 25 25 28 27 29 28 26 24 28 29 18 12 19 17 941 13 14 12 12 13 10 13 17 9 11 12 11 13 13 13 7 6 10 7 8 13 8 12 940 10 10 10 12 16 16 12 10 13 13 15 14 15 16 15 14 13 15 14 12 7 12 12 939 4 5 6 6 13 11 6 7 7 13 12 11 12 12 13 7 5 11 9 7 3 7 10 938 9 10 11 11 18 18 11 12 12 20 20 19 19 21 21 14 11 19 17 13 7 13 15 937 20 21 22 22 25 24 24 27 23 18 19 18 18 20 21 19 18 23 16 13 24 16 18 936 24 25 24 26 24 23 26 25 24 21 22 19 21 22 22 22 20 24 20 13 24 14 19 935 21 21 22 22 23 22 24 25 23 18 19 18 19 20 20 21 18 23 18 13 22 15 18 934 14 13 15 15 18 23 17 13 18 25 25 28 26 30 28 22 18 22 29 18 12 15 15 933 14 13 14 14 15 20 16 12 21 22 23 27 25 28 28 22 20 22 29 18 12 16 13 932 18 17 16 18 19 22 20 17 18 23 25 25 26 29 26 21 18 22 26 18 16 16 14 931 15 15 15 15 18 19 19 15 20 22 21 25 24 27 27 22 19 21 27 13 15 13 18 930 15 14 14 17 18 21 17 13 20 22 22 28 27 29 28 24 22 25 29 14 12 16 16 929 16 16 16 17 21 23 18 16 20 26 25 29 28 29 30 23 21 27 28 14 13 17 20 928 14 13 13 15 18 20 17 14 21 23 22 28 27 29 31 23 21 24 30 14 13 15 17 927 19 19 17 18 19 20 21 18 18 20 21- 20 22 23 21 18 17 18 21 11 16 13 15 926 15 14 13 14 15 21 17 15 17 18 19 21 20 22 21 20 18 21 22 11 14 14 16 925 17 16 16 18 20 20 19 16 22 21 21 20 23 23 23 23 22 24 21 12 14 17 16 924 16 15 14 16 18 22 18 14 18 21 20 23 24 24 23 23 21 23 23 13 13 16 16 923 13 15 16 15 19 21 14 16 16 17 16 17 18 18 19 19 14 22 15 10 12 12 922 13 11 12 12 16 17 14 15 16 13 14 18 18 18 19 16 18 21 17 12 12 921 27 26 25 23 15 15 30 27 18 12 12 12 14 15 15 16 16 16 14 9 920 11 11 12 12 12 13 12 11 15 14 15 16 16 19 18 15 13 16 16 919 16 15 16 17 17 22 18 13 21 26 26 30 28 33 32 27 24 28 .918 18 17 19 29 22 25 21 19 26 24 23 25 24 27 27 34 30 917 17 15 15 17 15 19 19 15 25 18 18 20 19 22 22 31 916 17 16 18 19 19 21 20 16 27 21 21 22 22 25 24 915 17 16 17 17 22 25 18 17 24 28 28 34 33 36 914 17 16 18 18 22 25 18 17 21 28 30 33 33 913 18 17 19 19 21 24 18 17 20 26 27 31 912 15 14 15 15 19 24 16 15 19 26 26 911 17 16 18 18 21 25 17 15 19 30 910 15 14 16 16 20 25 17 15 17 909 17 17 18 18 20 20 20 18 908 25 25 25 24 19 19 28 907 30 28 27 26 18 19 906 20 18 20 20 27 905 16 16 17 18 904 28 29 29 903 30 30 902 32 174 Appendix 3 - cont. I. D. GAULD 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 15 14 18 15 13 16 19 18 8 22 19 17 25 22 23 1 15 13 9 14 17 16 8 34 31 24 31 35 19 10 37 30 26 31 7 31 28 24 5 26 23 6 31 PHYLOGENY OF THE OPHIONINAE 175 Appendix 4 Primary labels matrix derived from analysis of 5 1 selected ophionines . The figure in bracketsto the immediate right of the taxon number indicates the number of times that taxon alone is responsiblefor the failure of a character in the LeQuesne test. Notably high values in any column suggest that aparticular character state shared by the respective characters is likely to be the result of paralleldevelopment or reversal. For example, character 1-2 has high values for taxa 909 and 923; as both taxashare an apomorphic 1 state, the labels matrix suggests this condition has been independently derived inthe two taxa, and thus should not be considered a synapomorphy uniting 909 + 923. iUNLABELLED TAXA: 912, 913, 933 176 I. D. GAULD Appendix 4 - cont. 17 18.2 20.1 21.1 22.1 23 24.2 25.2 27.1 27.3 29.1 30 18.1 19 20.2 21.2 22.2 24.1 25.1 26 27.2 28.2 29.2 31 901 1 - - 902 6 1 903 - - 904 - - -- 905 -- 4 - 2 1 3 - 1 906 2 - 1 2 1 5 6 6 1 11 1 1 8 1 907 -- 1 11 - - 1 1 908 ....*-. 7 2 3 3 1 1 8 1 2 2 909 1 1 4 - 3 2 2 1 -- 1 3 1 1 2 910 1 1 911 - - 1 1 - - 914 - - 1 1 -- - 1 915 4 1 1 1 1 1 1 1 916 - - -- 917 - 1 - -- -- 918 6 4 6 1 1 - 3 919 2 920 1 -- - 1 .... 2 921 1 - 2 1 4 11 1 1 2 1 2 3 -- 5 922 1 2 1 3 1 2 2 5 1 1 1 17 4 1 923 5 3 1 l 2 1 1 1 -- 3 10 4 1 -- 924 - -1 1 12 1 925 1 2 12 1 -- 926 - 2 1 1 1 - 5 927 1 1 1 2 3 - 1 928 -- - -- 2 929 - 1 - - 1 1 930 -- -- 931 -- 1 - - - 932 - 1 - - 6 934 - - - 935 - i 1 _ l .. l .. 2 2 936 - -- - - - 1 - 937 2 2 1 10 1 1 938 1 .... l - - - i - l l 939 1 - - 940 1 1 - - 941 i ._ 1 3 e 6 942 - l 1 1 1 l 2 1 -- 943 .. i ._ _ i ._ i i l .. 944 .. .. 2 -- - - 945 -- - 1 - - - - 1 -- -- -- -- -- - 946 l 2 1 -- - - 1 1 1 947 5 -- -- 5 - 2 - 948 - - - l .. - i - 15 - - - 1 - - -- -- 949 .. 3 - i __ _. ._ _ 950 - - 2 4 3 2 1 -- -- -- - 1 -- 951 2 -- l - 1 15 1 1 1 2 -- 1 - PHYLOGENY OF THE OPHIONINAE 177 Appendix 4 - cont. 32 35 36.2 37. 39 41 42.2 43.2 44.2 45.2 47.1 48.1 33 36.1 36.3 38 40 42.1 43.1 44.1 45.1 46 47.2 178 I. D. GAULD Appendix 4 - cont. 48.2 51.1 52.1 54 56 58.1 59 61.1 62 6450 51.2 52.2 55 57 58.2 60 61.2 63 901 -- - 902 - 903 1 7 -- -- 904 - 1 3 1 -- 905 - -- - -- - 906 6 3 5 - 1 907 -- - - - - - 1 -- 908 1 1 2 2 2 - 2 909 31 1 30 7 1 - 1 3 1 2 910 1 - 2 5 911 1 - 1 5 914 - - 6 -- - 915 i .... 3 30 - 1 1 1 916 i i 4 - 917 __ __ __ __ 3 4 __ __ __ __ __ ._ i __ 4 __ 918 - 2 2 1 919 920 - -- 1 1 1 921 1 11 - - -- 922 - 6 5 1 - 1 1 1 1 923 1 - . 1 4 - 1 -- 924 1 - 925 6 - 1 1 926 -- - 2 - 927 31 1 2 1 1 1 2 2 928 1 - 2 1 19 1 1 929 - 4 - 930 - 931 -- 1 932 2 - 1 - 1 934 i .. __ __ __ 1 _ 935 __ ._ ._ 2 - 3 - 936 - -- 2 -- 3 -- -- -- 937 - 3 2 1 - 938 7 - 1 939 - 1 1 940 2 1 - - - 941 4 1 4 1 3 - 1 942 2 -- 1 1 1 27 1 19 2 2 -- 1 943 -.-... i e 7 - -- -- 2 -- 1 944 - 1 - 945 -- -- - 5 2 946 2 -- - -- 1 - - 2 2 -- 947 1 - 1 - 948 -- 1 1 -- - 1 949 2 2 - 7 - 1 1 2 950 - 3 2 - 3 3 - - - -- 951 1 -- 2 1 27 1 1 1 1 1 1 -- PHYLOGENY OF THE OPHIONINAE 179 CO LT) CO i i ro I CM i-H CO LD CO i * CO CM CO t t CM CM I CM IO CM I CO I I-H <> CO LO CO -H I CM .-H IT) I-H CM CO I 3 CM CXl T-H r^ *" *VI CO CM CM CM i i co in co CT> I CT> I I CM I ^H I CO CM CO CM . < ro CM I-H i CM ro i ro I-H , i i cr> I-H i ,-H i CM fi ro ID co CM CM ro ^4 00 CM I-H co in co CM .-H ro LO co u o I 353 oo j^j i< OO i < i < I O I o o (U t-H f-1 C\J 00 00 r-l I-H ^H < i <i ro TD _X rt ^- r-l l-t U I -J I I I I PHYLOGENY OF THE OPHIONINAE 181 Appendix 6 - cont. 182 I. D. GAULD CM i-HOi-H.-HOOi-lr-l o o o CM O O O o o o o CM co i-H O O O o o o r~. r-l O O O i-H O i-H I-H O o o o o o o i-l O o o o o o o o o o o O r-l ,-H O .-H r-H -4 O O o o o o o o o o o OOOOO.-HOO OOr-HOOi-HOO o oo o -H O o o o o o o o o o o o o o o o o o o o o o o o o o o o o o o O 1-1 o o o o I-H I-H I-H o o O O O O .-H O O O o o o o o o o o o o o o o o o ,-H O .-I o o o o o o o o o O <-i o o -I O 1-H O o o o o o o O r* O i-H ^H O ^H rt o o o o o o o o o o o o o o o o o o r-l O O o o o I-H O O O ^-1 O ^-i O O o o o 1-H O O o o o O O ^H o o o o o o o o o ^H O O o o o o o o tr oooooooo OOOOOOi-H-HCO OOOt-HOi tOO tx oooooooo CM IO CO Or-Hl I I lOOr-Hi I Or-HOOOOOO 10 ,-H ,-1 ^H VO O O O O ^H 1-H ^H 10 O O 1-H 11 1-H If) o oo o o oo o CO i-HOt-HOOOOO CO IT) O O O O O O OO O O O O <-< I-H O 1-H O ^H o o r-< O o o o o oooooooo r i lOi li lt-HOi-H> I IT) i li-Hi Ir-lOOOr-l IQ OOOOOi-HOO to OOOOOOOO uo ,OOOOOOOO 10 OOi-HOOOOO a- OOOOOOOO 10 OOOOr-H^HOO Oi-i.-H^H.-!O.-i.-i 10 OOOOOOOO CM CMOOOOOOOO tO OOOi-HOOOO CM 1-Hi-Hi li li It-Hi-Hi-H IO Ol lOl-HOOOO CM OOOOOOOO 10 ^H^HrtrH^H^Hrt^H o i-Hi-HOOOOOO to OOOOOOOO en i-H i-H O CO r-4 i-H O CO ^f i-H i-H i-H t I i-H f I rH i-H CM oo OOOOOOOO <* Oi-HOOOOOO co 1-H 1-H 1-H I-H i-H 1-H 1-H 1-H ^" 1 * 1 I O CO CO O 1 I r 1 CM T--i-Hi-Hi-Hi-Hi-Hi-Hi-H^H 3" OOOi-HOOOO r- OOOOl-Hl-HOO ^- i-Hi-Hi-Hi-Hi-Hi-Hi-l!-! IO -HOOOi-Hi-lOO ^J" OOOOOOi-Hi-H CM to OOOOr-HOOO S" i-HOOOOOi-Hi-i 1-H tn OOOOi i i i O O ^ i if-Hi-Hi-HOi i t i i i Oi iCMco^rtor^ooto to to to to to to tocncncncncncncncn i a OITi ON ON" O i-l t-l i-H I-H ur> -H CM r-H Vf) ^H ^H ,_( 00 s <N ON <N ON i CM in <-i t-i CO OJ CM i-H CM VD CM r-l t-H IT) i-l T3 8"S. OJ "cl c <u X 5 03 CM i-H i-l 184 I. D. GAULD C\J I CNJ CvJ vo c\j co in co CM .-H i-H ,-H CM CO I-H evj c>o t-i f i n cvi oo I-H ro co CM ,-H r-t i-H I-H I-H CM i I CM i I i I .2 o i CO CTl O i I PHYLOGENY OF THE OPHIONINAEAppendix 9 Labels matrix derived from the Euryophion primary data matrix (Table 20). 1.1 19 30 37 54 59 71 73 185 75 Index to generic names Synonyms are in italics. British Museum (Natural History) An introduction to the Ichneumonidae of Australia /. D. Gauld In the important field of biological and integrated control of pests the parasitic Hymenopteraare of particular significance, and this work considers one of the largest families of Parasitica,the Ichneumonidae. The group has received little attention in Australia - though it has alreadybeen utilized successfully in curtailing the ravages caused by accidentally introduced pests. Forselective control programmes to be effective, however, a sound knowledge of the biology ofboth the pest and its parasites is essential - and a sound taxonomic base is vital for thedevelopment of such knowledge. Ironically, considering the group's economic importance, the parasitic Hymenoptera isamongst the least studied of any group of living organisms, and taxonomic difficulties havepresented major problems to many entomologists working with the Parasitica. AnIntroduction to the Ichneumonidae of Australia will go a long way towards rectifying thissituation, being a taxonomic treatment, by genus, of the Australian ichneumonids, acomprehensive illustrated identification guide, and a summary of all available information onthe group. It will also serve as an introduction to the biology and distribution of Australianichneumonids, and provide a check-list of the described species and an index to their knownhosts. It provides an important revision of ichneumonid nomenclature in order to bring thegroup into line with the generally accepted principles of zoological nomenclature. 1984, 413pp, 3 maps, 580 figs. Paperback. 565 00896 X 40.00 Titles to be published in Volume 51 The ichneumon-fly genus Banchus (Hymenoptera) in the Old WorldBy M. G. Fitton The phylogeny, classification and evolution of parasitic wasps of the subfamily Ophioninae(Ichneumonidae) By I. D. Gauld A cladistic analysis and classification of trichodectid mammal lice (Phthiraptera: Ischnocera)By C. H. C. Lyal The British and some other European Eriococcidae (Homoptera: Coccoidea)By D. J. Williams Photoset by Rowland Phototypesetting Ltd, Bury St Edmunds, SuffolkPrinted in Great Britain by Henry Ling Ltd, Dorchester BRITISH MUSEUMNATURAL HISTORY) Bulletin of the 26AUGI987 PRESENTED British Museum (Natural History) A cladistic analysis and classification oftrichodectid mammal lice(Phthiraptera: Ischnocera) C. H. C. Lyal Entomology series VolSl No 3 31 October 1985 The Bulletin of the British Museum (Natural History), instituted in 1949, is issued in fourscientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology,and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique andever-growing collections of the Museum, both by the scientific staff of the Museum and byspecialists from elsewhere who make use of the Museum's resources. Many of the papers areworks of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself,available separately, and individually priced. Volumes contain about 300 pages and severalvolumes may appear within a calendar year. Subscriptions may be placed for one or more ofthe series on either an Annual or Per Volume basis. Prices vary according to the contents ofthe individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History),Cromwell Road, London SW75BD,England. World List abbreviation: Bull. Br. Mus. nat. Hist. (Ent.) Trustees of the British Museum (Natural History), 1985 The Entomology series is produced under the general editorship of the Keeper of Entomology: Laurence A. Mound Assistant Editor: W. Gerald Tremewan ISBN 565 06013 9 ISSN 0524-6431 Entomology series Vol 51 No 3 pp 187-346British Museum (Natural History)Cromwell RoadLondon SW7 5BD Issued 31 October 1985 A cladistic analysis and classification of trichodectidmammal lice (Phthiraptera: Ischnocera) 2vC. H. C. Lyal Department of Entomology, British Museum (Natural History), Cromwell Road, LondonSW7 5BD Contents Synopsis 187 Introduction 187 Acknowledgements 188 Method of systematic analysis 188 Relationship of systematic analysis and classification 189 Morphology 189 Character survey 203 Character analysis 213 Identification of apomorphic states 213 Cladistic analysis 223 Taxonomic history of Trichodectidae 234 Proposed classification 241 Descriptions of genera and subgenera 247 Bovicolinae Keler 247 Eutrichophilinae Keler 265 Dasyonyginae Keler 267 Trichodectinae Kellogg 286 Neotrichodectinae subfam. n 321 Keys to Trichodectidae 334 Key to subfamilies 335 Key to genera and subgenera 335 References 340 Index 344 Synopsis Observations on phthirapteran morphology are analysed with particular reference to the Trichodectidae.Problems of structure, homology and nomenclature associated with major morphological features arebriefly reviewed and solutions suggested where possible; where appropriate, the morphological terminolo-gy is clarified. A cladistic analysis of the 351 species and subspecies of Trichodectidae is carried out usingthe states of 187 characters, and the taxa are re-classified in accordance with the results and the principles ofphyletic sequencing. Twenty genera are accepted in five subfamilies. Ten of the genera are divided into 27subgenera. This necessitates the description of one new subfamily, three new genera and four newsubgenera. Three genera are newly synonymised, eight genera and subgenera are recalled fromsynonomy, and four genera are reduced to subgenera. The generic placements of 106 species andsubspecies are changed. Keys to the subfamilies, genera and subgenera are provided, and the genera andsubgenera are described. Introduction The Trichodectidae is a family of ischnoceran chewing lice parasitic on mammals. There are 351described species and subspecies grouped into between 13 and 39 genera, various workershaving widely differing views on generic limits. The classification of the Trichodectidae at thegeneric level is perhaps more confused than that of any other group of lice (Hopkins, 1949, 1960;Emerson & Price, 1981) and no workable keys to genera are available. The confusion and Bull. Br. Mus. nat. Hist. (Ent.) 51 (3): 187-346 Issued 31 October 1985 188 C. H. C. LYAL disagreement surrounding the classification of the family persists despite a sound basis oftaxonomic knowledge at the species level, derived largely from the works of Werneck (1948,1950), although this author did not attempt to produce any keys. The purpose of this study is to present a classification of the species of Trichodectidae and todescribe and provide a key to the recognised genera and subgenera. The foundation of theclassification proposed below is a cladistic analysis, and for the purpose of determining characterpolarity in that analysis a suitable out-group had to be identified first. Ideally this out-groupshould be the sister-group of the taxon studied, and in this case would be expected to be a groupin the same suborder, the Ischnocera. However, the holophyly of the Ischnocera has not beendemonstrated (Lyal, 1985), although there is evidence that indicates the holophyly of a groupcomprising most of the Ischnocera (including the Trichodectidae). The classification of theIschnocera is unsettled, and the sister-group of the Trichodectidae has not been identified. Asingle teneral male specimen of a possible sister-group has been seen, purportedly collectedfrom a corvid in Austria. The specimen has the single tarsal claws of a mammal parasite, but noother apomorphies of the Trichodectidae, although it has the same general facies as members ofthe family. As out-group, therefore, the Ischnocera s.l. is used, reference being made whereappropriate to the other three suborders of Phthiraptera (Lyal, 1985). The term 'holophyletic' is used below to refer to groups of taxa comprising a single ancestralspecies (known or inferred) and all of its descendants. The more familiar term 'monophyletic' isavoided because ambiguities in its use have caused misunderstandings in the resolution ofsystematic problems. Scale lines equivalent to 50 /am are given in most figures. This study formed part of the results of a research project submitted to the University ofLondon for the degree of Doctor of Philosophy. Acknowledgements I am very grateful to all those who have given advice and read parts of this paper whilst in preparation,particularly Dr R. G. Davies, University of London; Dr L. A. Mound, Dr T. Clay, Mr W. R. Dolling, MrD. Hollis and Mrs J. M. Palmer, British Museum (Natural History); and Dr B. Heming, University ofAlberta. I also thank the following people who were of great assistance during part of this study which wascarried out in the United States: Dr K. C. Emerson, Florida; Dr K. C. Kim, Pennsylvania State University;Dr R. D. Price, University of Minnesota; Dr O. Flint, United States National Museum; Dr R. Traub andMiss H. Starke, University of Maryland School of Medicine; and Dr W. A. Drew and Dr D. Peters,Oklahoma State University. Method of systematic analysis Systematic analysis is carried out in order to determine 'relationships' between different taxa,the relationships discovered being employed to study some aspect of the biology or evolution ofthe taxa and/or to provide a framework for their classification. Taxa may be deemed to be'related' according to a number of different criteria, so the type of analysis performed is dictatedby the requirements of other studies to be undertaken. In this case the systematic analysis ispreliminary to an examination of the phylogenetic linkages between Trichodectidae and theirhosts (Lyal, in prep. , a). The relationships required from the analysis are therefore phylogeneticin nature. Although it has been suggested that phylogenetic relationships can be determined byphenetic analysis (Sneath & Sokal, 1973), this is generally believed not to be the case, and aphyletic analysis is required (Wiley, 1981). In this study a manual cladistic analysis was deemedmost suitable as computerized methods of sufficient power and capacity were unavailable whenthe study was initiated (Lyal, 1983). The first stage in the analysis was a character survey of all species of the family, with the aim ofdiscovering suitable apomorphies. As analysis proceeded some characters initially selected wererejected, if the polarity of their states could not be determined, or if their apomorphic stateswere restricted to single species. The observations were recorded in a data matrix, but it rapidly became apparent that this TRICHODECTID MAMMAL LICE 189 matrix was too large to be manipulated conveniently. Preliminary inspection of the taxa had ledto the recognition of sufficient apomorphies to divide the family into five or six plausibleholophyletic groups, so secondary data matrices were compiled for each of these. The smallernumber of species and apomorphies then necessary on each of these matrices enabled manualanalysis to proceed. These matrices were inclusive rather than exclusive, and taxa with lowprobabilities of membership were included; as a consequence some taxa were included in morethan one matrix. The matrices were examined and taxa or holophyletic groups sharing thegreatest number of synapomorphies were united as sister-groups. This process was continueduntil the cladistic relationships within each matrix were resolved as fully as possible. The fullmatrix was then compiled using the largest holophyletic groups identified in the analysis of thesecondary matrices. This matrix was then analysed using the same technique and the fullcladogram drawn. The cladogram, which gave the full distribution of apomorphies, wasexamined to ensure that the most parsimonious distribution of apomorphies had been achieved.Character states were weighted by a simple 'gain' or 'loss' criterion. 'Loss' states were onlyused in the analysis where 'gain' states were not available, and after 'gain' states had been used toresolve the cladogram as fully as possible. It follows that, whereas the distribution of 'gain' statesshould be maximally parsimonious over the cladogram, this is unlikely to be the case for 'loss'states. Relationship of systematic analysis and classification Biological classifications afe designed not only to enable the taxa classified to be located in theliterature and in collections, but also to store information about those taxa. In many cases thenature of this information is not clear from the classification, and taxa are grouped together onthe basis of some ill-defined combination of phenetic and phyletic relationships. The type ofinformation stored in the classification (i.e. the type of relationship used in its construction)may, however, be indicated precisely, and the classification used as an efficient data-retrievalsystem. This can only be done by utilisation of cladistic information alone; attempts toincorporate phenetic or anagenetic information lower the retrieval facility (Cracraft, 1974). The method whereby the classification was constructed from the results of the cladisticanalysis in this study was by using both rank and sequence in which taxa of equal rank are listed(G. Nelson, 1972, 1974; Cracraft, 1974; Wiley, 1979, 1981). This process has been termedphyletic sequencing (Cracraft, 1974). The convention employed is that within a classificationholophyletic taxa of equal rank are listed ('sequenced') so that each taxon is the sister-group ofall those taxa of the same rank (and within the same taxon of immediately higher rank) listedbelow it in the classification (modified from Cracraft, 1974). Use of this criterion allows recoveryof a dichotomously-branching cladogram from the classification. However, holophyletic groupswith a trichotomous or polychotomous interrelationship must also be accommodated. Wiley(1979) recommends that these be placed in the classification with equal rank and be noted as'sedis mutabilis' to indicate that their order is unknown or interchangeable. It must beremembered that this convention is for a formal classification, not for regular use in discussion. All hierarchical groups may be sequenced, but it is probably more efficient to employ informalgroups below the subgenus level, especially with larger genera. Morphology The discussion below is intended to clarify morphological terms used in this study, to point outstructures of taxonomic or systematic interest, and to describe briefly the morphology of theinsects. Head Figures 1 and 2 depict the dorsal and ventral aspects of a generalised trichodectid head with theregions and major features labelled with the terms used in this study.The internal carinae of the head are strongly developed in many species, and differences in the 190 C. H. C. LYAL dorsal preantennalsulcus median broadening ofclypeal marginal carina clypeo-f rental sulcus pulvinus clypeal marginal carina ? frontal sulcus f lagellum eye occipital ring osculum clypeus clypeus sub-genal area genal area temple f rons gula Figs 1, 2 Terminology applied to trichodectid head, modified slightly from Symmons (1952). Mandible,maxilla, hypopharynx and labium omitted. 1, structures of head ($); 2, regions of head ($). degree of sclerotisation and position are useful taxonomically, particularly at the species level.The postoccipital sulcus is marked by a greatly-developed internal carina ('occipital ring')projecting posteriorly into the thorax; this structure is present in many other lice, but rarely sowell developed. The postero-lateral margins of the head ('temple margins') are more or less rounded and maybe produced at the postero-lateral angles in some species. The lengths of the setae on thepostero-lateral angles may provide characters of taxonomic value. The eyes are positionedadjacent to the antennae or more posteriorly; in some species (e.g. those in the genusCebidicold) the eyes occupy lateral projections. Anterior to the antennal socket the margin of the head is produced into a variable sclerotisedconus (Fig. 1), but the trabecula (Clay, 1946) is absent. A sclerotised conical projection from thedorsal nodus of the clypeo-frontal sulcus may be present projecting parallel to the margin of theantennal socket. This projection is frequently present in only one sex of a species. TRICHODECTID MAMMAL LICE 191 The anterior margin of the head may, in plan, form an unbroken arc ('circumfasciate head'),but in most Trichodectidae it is interrupted medially by the pulvinus or the osculum (Symmons,1952). The pulvinus is a thick bilobed pad of unsclerotised tissue developed from the clypeo-labral suture, and is found in all Ischnocera (Fig. 1). The osculum is a median indentation of theanterior margin of the head (Fig. 2). When the insect is at rest the pulvinus and osculum havebeen observed to be applied to the hair of the host (see discussion of mandibular morphologybelow). The width of the osculum is therefore probably influenced by the diameter of the hair inthe region of the host's body inhabited by the louse. In cases where the pulvinus fails to reach theanterior margin of the head a small ventral sulcus ('ventral preantennal sulcus' of Clay, 1951)may connect it to the margin or osculum. Many species have a small dorsal preantennal sulcus(Clay, 1951) also interrupting the anterior margin medially. Between the antennae and the osculum the margin of the head may, in plan, be shallowly orbroadly convex, sinuate or straight (Figs 3-6); sclerotisation along the margin (forming the'clypeal marginal carina') may be minimal or heavy, but is usually pronounced in the mediandorsal region (Fig. 1). Much of the variation is possibly influenced by the density and texture ofthe hair of the host. Only two structures of the mouthparts need be mentioned, the mandibles and the cibarialsclerite. The mandibles of Trichodectidae, like those of many other Amblycera and Ischnocera, areasymmetric. There are three apical teeth on the right mandible and two on the left; on the rightthe centre tooth is generally longest, whilst on the left the posterior tooth is normally the betterdeveloped (Figs 7-10). Mandibular asymmetry is considered by Snodgrass (1935) to be a featureof insects that masticate their food, and the lack of asymmetry in the mandibles of some speciesoiRidnus (Amblycera) is associated by B. Nelson (1972) with blood-feeding in those species.Trichodectes canis, the only trichodectid known to take blood meals (Bouvier, 1945), hasdimorphic mandibles, as do all other Trichodectidae. In addition to the collection and preparation of food the mandibles of Trichodectidae areemployed in anchoring the insect. When at rest, Trichodectes canis and T. metis have beenobserved to enclose a hair in the pulvinus with the mandibles and whilst in this position maycompletely release the grip of the tarsal claws and straighten the legs laterally from the body.The insect is then held on only by the mandibles. In addition to the species above, specimens ofFelicola sp. collected from dried museum skins have been found in this position, and specimensof most genera have been found preserved in alcohol clinging to single hairs by their mandiblesalone. The selective pressures on the mandibles of Trichodectidae are different, therefore, fromthose operating on lice that do not have a mandibular anchoring mechanism, and possiblyprecludes the assumption of monomorphy in blood-feeding species. The 'interior' face of theright or of both mandibles may be ridged, so that when the mandibles are folded closed, theridges on one mandible are not covered by the other, and all or most of the exposed mandibulararea is ridged. The ridges are therefore not positioned in such a way that they can act against oneanother, and it is suggested that they are developed to prevent the mandibles slipping on a hairwhen clasping it (Figs 8-10). When the mandibles are closed, the right is always interior to theleft, which may have a concavity to receive it (Figs 7, 9, 10); the left mandible is not then kept bythe right from contact with the hair, and the greatest possible mandibular area is utilised. Insome species the right mandible has a basal notch on the anterior margin, which receives the tipof the left mandible (Figs 7, 10); this feature may act in concert with the restraint of the rightmandible to 'lock' the mandibles closed about a hair. Although, as said above, none of theTrichodectidae have monomorphic mandibles, the mandibles of Damalinia (T.) conectens (Fig.7) are very slender and pointed, and the centre tooth of the right mandible of Dasyonyx spp. andEurytrichodectes spp. (Fig. 9) is long and pointed, suggesting in each case a piercing function andthus possible haemophagy . In contrast, the mandibles of Damalinia (T.) indica (Fig. 8) are bluntand broad, suggesting an adaptation to grinding and chewing. The sitophore sclerite (Figs 11, 12) varies considerably in the Psocodea (Cummings, 1913;Haub, 1967, 1972, 1973, 1977). The form of the sclerite, particularly of its posterior border, isobscured in slide-mounted specimens, and differential inclination of the specimens may change 192 C. H. C. LYAL Figs 3-10 Variation in trichodectid head and mouthpart structure. 3-6, $ head, dorsal aspect, of (3)Bovicola caprae; (4) Damalinia lineata; (5) Felicola subrostratus; (6) Trichodectes zorillae. 7-10,mandibles, dorsal aspect, of (7) Damalinia conectens; (8) Damalinia indica; (9) Eurytrichodectesparadoxus, showing mandibles interlocking; (10) Trichodectes canis, with detail of postero-dorsalmargin of left mandible. TRICHODECTID MAMMAL LICE 193 14 13 12 11 Figs 11-16 Trichodectid morphology. 1 1 , Damalinia neotheileri, sitophore sclerite. 12, Bovicola hemitra-gi, sitophore sclerite. 13, Damalinia indica, right antenna of d", dorsal, arrow indicating prolongation ofantero-ventral pedicel margin. 14-16, lateral aspect of metatarsal claws of (14) Dasyonyx (D.)smallwoodae; (15) Dasyonyx (N.) diacanthus; (16) Eurytrichodectesparadoxus. the apparent dimensions (Haub, 1977). Trichodectidae differ from other Psocodea in having anopen posterior border to the sclerite, a character state identified by Haub (1973) as plesiomor-phic, but which is almost certainly apomorphic. Within the family the sclerite is present in allspecies and the form is fairly constant (Fig. 11), although in two species groups the posteriorarms are extended laterally (Fig. 12) and the sclerite is difficult to see. The minor variation in theform of the sclerite at the species level is so far unexploited in taxonomic study. The sitophore sclerite is cup-shaped, and is opposed on the dorsal surface of the sitophore by aclosely-fitting projection, the epipharyngeal crest. These two structures have been identified as a'mortar and pestle' for grinding food (Weber, 1936) and as components of a salivary pump 194 C. H. C. LYAL (Buckup, 1959; Keler, 1966). Rudolph (19820, 19826, 1983) demonstrates that it is a pump,sucking water absorbed from the atmosphere onto the lingual sclerites through the Y-shapedduct linking these and the sitophore sclerite. The antennae are made up of a scape, pedicel and flagellum. In common with otherThysanura-Pterygota the pedicel and flagellum are annulations formed by subdivision of a singlesegment (Imms, 1939; Matsuda, 1965). Consequently, the only intrinsic muscles of the antennaeare in the scape and insert on the pedicel, moving the pedicel and flagellum as a unit; the scapeitself is moved by muscles inserted on its basal margin and arising on the anterior arms of thetentorium. The positions of the muscle insertions on the scape and pedicel are such that the an-tenna can be moved through 360 degrees, but there is no muscular mechanism for flexing theflagellum. Whilst the prime function of the antenna is sensory, secondary modification in themale has taken place in some Ischnocera (including almost all Trichodectidae) and someAnoplura, the antennae being used to clasp the female round the abdomen during copulation(Keler, 19380; Sikora & Eichler, 1941; Lyal, in prep., b). This development has led to anincrease in length and degree of sclerotisation of the pedicel and flagellomeres and an increase instrength of the antennal muscles. The intrinsic muscles of the scape are enlarged and the scapeexpanded; the extrinsic muscles are also enlarged and the head concomitantly broader in themale than in the female. To facilitate free movement of the flagellum the joint between the scapeand the pedicel is broad and membranous (other than at the two articulatory points). However,because there is no direct muscular control of the flagellum, the degree of free movementrelative to the pedicel must be limited, so that control may be effected by the intrinsic muscles ofthe scape. Observations made in this study reveal that the apex of the pedicel is angled relative tothe long axis of the annulation, the longest margin of the pedicel being the antero-ventral (Fig.13). There is very little membrane between the pedicel and the flagellum on the antero-ventralmargin, but more on the postero-dorsal, so some flexibility between the pedicel and flagellum ispossible, although limited in the anterior direction by the prolongation of the pedicel. Duringcopulation the male clasps the female around the top of the abdomen from underneath(Werneck, 1936; Sikora & Eichler, 1941), the pedicel and flagellum being curved to match thecurvature of the abdomen. The antennae are raised above the head of the male and the intrinsicadductor muscles contracted. The pedicel and flagellum are thus brought down over theabdomen of the female, the flagellum being constrained by the form of its junction with thepedicel. The form of the pedicel-flagellum joint is such that, should the haemolymph pressure bereduced, the joint membrane would collapse and further contract the flagellum against thefemale. However, lice are not known to have control over haemolymph pressure in theirantennae, although larvae of Lepidoptera are known to control their antennal movementpartially by this means (Matsuda, 1965). The mechanical strength and degree of possible controlof a system involving joints as described above is likely to be inversely proportional to thenumber of joints in succession, with a single joint being the most efficient. The degree ofcurvature attainable using three segments is sufficient to grasp the female. For these reasonsonly the first flagellomere is required to take on a clasping function; the apical two flagellomeresmay be retained in a sensory capacity (most Ischnocera) or lost (Trichodectidae). InTrichodectidae the last two flagellomeres have contracted and fused to the first flagellomere inall males (although a small semicircular sclerotisation, probably representing the apical flagello-mere, is present in the male of Eurytrichodectes paradoxus) and in females of the Neo-trichodectinae, Trichodectinae, Bovicolinae and most Eutrichophilinae. The sensilla of the twoapical flagellomeres are retained on the remaining flagellomere. The firmness with which the female is held may be increased by projections in the form ofspikes or denticles on the antennae, particularly on the flagellum. Trichodectidae are character-ised by the possession of two modified setae apically on the male flagellum, which take the formof sharp, stout teeth (Fig. 13). Trichodectidae have two sensilla coeloconica and three sensilla placodea on the terminalflagellomere, sometimes closely associated (Clay, 1970; Kim & Ludwig, 1978). In Lorisicolamalaysianus and L. mjoebergi the sensilla are in pits with tongue-like projections around them. TRICHODECTID MAMMAL LICE 195 Thorax Figures 17 and 18 depict a generalised trichodectid thorax, labelled with the terms used in thisstudy. In all Trichodectidae the rhombic sclerite anterior to the pronotum is present; thepronotum itself generally bears two lateral sclerites, sometimes narrowly joined medially. Themesonotum and metanotum are fused, with prominent pleural ridges extending onto them;there may, as with the pronotum, be a medial zone of desclerotisation. Contrary to the assertionof Mayer (1954), the lateral cervical sclerite is always present and bears two anterior setae. Thepronotum is fused to the propleuron, which in turn is fused to the prosternum. The prosternummay extend anteriorly between the fore-coxae and be unsclerotised medially. The mesosternite,fused to the prosternite, may also be medially divided. The mesepisternum is difficult to delimit,but rarely extends unbroken between the metasternite and the metanotum. The metasternite, ifpresent, is only rarely fused to the mesosternite, and is never sclerotised medially. Posterior to rhombic sclerite prothorax pterothorax abdominalpleurum I prosternum lateral cervicalsclerite mesothoracicspiracle profurcal pit mesofurcal pit mesosternummetasternum postcoxale met epimeron Figs 17, 18 Diagrammatic representations of generalised trichodectid thorax, with terms used in thisstudy. 17, dorsal. 18, ventral. 196 C. H. C. LYAL the metacoxa may be a semicircular sclerite ('postcoxale' of Matsuda, 1970). In Procavicola(Meganarionoides) this is very heavily sclerotised and fused to abdominal pleurum II; the twopostcoxales may also fuse medially. One, two or three setae are always present medially on the anterior margin of the thoraxposterior to the temple margin of the head. The lateral margins of the thorax also bear setae,which may be more or less abundant. A row of setae is generally present across the posteriordorsal margin of the prothorax and pterothorax, sometimes interrupted medially ('mediangap'). Setae are sometimes present on the dorsal disc of the prothorax and pterothorax('anterior setae'), and medially on the mesosternum. The legs of Trichodectidae, in common with those of most other mammal lice, terminate in asingle tarsal claw. In most species this claw is smooth ventrally, but it is toothed ventrally inDasyonyx (Figs 14, 15) and ridged in Eurytrichodectes (Fig. 16). In many species a small hyalineprojection, which may be pointed or blunt (Figs 14, 15), is present at the base of the claw. Mayer(1954), in her study of Bovicola caprae, terms this structure a 'pulvillus' but, as a true pulvillus isgenerally paired (Richards & Davies, 1977; Chapman, 1982), this term is inappropriate. Theprojection is more probably an empodium or arolium, or even a simple basal tooth of no widerhomology. Kim & Ludwig (1978, 1982) maintain that the pulvillus and empodium do not occurin the Phthiraptera, but Clay (1969) demonstrates the presence of an empodium in Menoponi-dae (Amblycera), and Clay (1970) figures structures in Boopia (Amblycera) that almostcertainly are pulvilli, although she follows Keler (no reference given) in terming them euplantu-lae (plantulae) of the second tarsomere. Abdomen In this study the 'true' segment number is referred to by roman numerals to distinguish it fromthe apparent number. The terms 'sternite', 'pleurite' and 'tergite' are used for the sclerites of thesternum, pleurum and tergum respectively of each segment. Segment I is represented in Trichodectidae by the reduced tergum I only. In this respect thefamily is similar to the Trichophilopteridae, but differs from all other Ischnocera, in whichsegment I is absent. Segments III-VIII bear the spiracles, if these are present; spiracles are never present onsegments I and II. Within the Trichodectidae many species have fewer than the plesiomorphicnumber of six pairs of abdominal spiracles (Table 1). Loss has apparently taken placesequentially from the posterior, so that if any spiracles are present one pair is on segment III, andif more than one pair is present there are no intercalating segments which lack spiracles betweenthose segments with spiracles. There is, however, no evidence that spiracles have necessarilybeen lost one pair at a time. Whilst in Procaviphilus (Meganarionoides) angolensis, P. (M.)colobi and P. (M.) baculatus the posterior pair of spiracles only is very small and apparently inprocess of being lost, in Lorisicola (L.) hercynianus and L. (L.) siamensis the posterior two pairsare extremely small, probably non-functional, and apparently in process of being lost. In anumber of clades, sister-groups exhibit multiple discontinuities in spiracle number. Thesister-species Lorisicola (L.) mjoebergi and L. (L.) malaysianus have six and zero pairsrespectively; Felicola viverriculae and an undescribed sister-species have three and zero pairsrespectively; the Lorisicola (P.) lenicornis - wernecki clade and the sister L. (P.) acuticeps -neoafricanus clade have four and zero pairs respectively. Variation within species can occur,although it is generally erratic. Felicola subrostratus normally has three pairs of abdominalspiracles; the species is widespread and found on many hosts, but on Madagascar, where the hostis Eupleres goudoti, there may be three or two pairs, and specimens exhibiting asymmetry arepresent in the British Museum (Natural History) collection. Asymmetry has also been noted inTrichodectes emeryi, one paratype of which has six spiracles on one side of the abdomen and fiveon the other. Trichodectes (S.) potus is unusual in that the female has three pairs of abdominalspiracles and the male only two, the only known example in the lice of sexual dimorphism inspiracle number. Most Trichodectidae have either six, three or no pairs of abdominal spiracles, other numbers TRICHODECTID MAMMAL LICE 197 Table 1 Distribution of number of pairs of abdominal spiracles in the Trichodectidae. being less common (Table 1). From the cladogram, reduction to five, four and one pair can beseen to have occurred once, reduction to two and three pairs twice, and reduction to none eighttimes. The loss of abdominal spiracles plainly cannot be used a priori as a taxonomic characterdefining (holophyletic) genera, as proposed by Ewing (1936), but neither is it as variable assuggested by Keler (1938) and Hopkins (1941), who treat it as a character of specific value only.The selective advantage of this reduction is not known, but it may be an adaptation to excludedust from the tracheal system, or to reduce water loss. A posterior commissure joining the two main abdominal tracheal trunks is present inAnoplura, Rhyncophthirina, Boopiidae and Trimenoponidae (Amblycera), Philopteridae(Ischnocera) and 'some Trichodectidae' (Harrison, 1915; Ferris, 1931). All species ofTrichodectidae examined in this study possess a posterior commissure, which is consequentlyassumed to have a universal distribution throughout the family. The presence of the posteriorcommissure is assumed to be plesiomorphic for the Phthiraptera (Clay, 1970). In all lice segment IX is the genital segment, and the male genitalia open to its posterior andthe female genitalia to its anterior. Modifications of this segment are discussed in detail below. In both sexes of Trichodectidae segments X and XI are fused. In females segment XI liescaudally, posterior to X, but in males the modifications of segment IX have led to thedisplacement of both X and XI onto the dorsal surface of the abdomen. Most male Trichodecti-dae have the genital opening posterior or postero-dorsal. The postgenital segments are reducedto form a small anal cone arising from the dorsal (anterior) wall of the genital chamber (Keler,1957) (Fig. 19). Conical projections are present either dorsoposteriorly or ventroposteriorly on the first three 198 C. H. C. LYAL tVllp sVII tergum VII tVllla sternum VII tergum VIII tVlllp gc segment IXp-gs' Fig. 19 Diagrammatic representation of terminal segments of trichodectid male abdomen, with termsused in this study, t - tergite; s - sternite; p - pleurite; ac - anal cone; gc - genital chamber; p-gs -post-genital sclerite; sgpr- lateral rods of subgenital plate. pleura (II, III and IV) in many Trichodectidae. The distribution of the pleural projections issummarised in Table 2. Male Trichodectidae, alone among Phthiraptera, possess 'lateral abdominal flecks', firstnoted in Werneckiella by Moreby (1978). The flecks are small pits positioned on the antero-dorsal angles of pleura III-VII, occasionally on II and VIII, and anterolaterally on terga III-V inmany species (Fig. 218). In Werneckiella there are small sclerites situated anteriorly andposteriorly of the pit; in Trichodectes canis there are no sclerites bordering the pit but a smallsclerite is present at its base; in Neotrichodectes there are no sclerites associated with the pit atall. The function of these structures is unknown. In some species of Phthiraptera terga II and/or III of the males are ornamented with largesetae or projections. Such developments are numerous in the Trichodectidae. Many species ofFelicola have a pair of long setae medially on male tergum II (Figs 180, 187, 188), whilstGeomydoecus (Thomomydoecus) spp., Trichodectes ovalis, T. ugandensis and an undescribedsubspecies of T. galictidis have paired 'combs' of long setae on male terga II and III (Fig. 161),and Bovicola multispinosa and B. hemitragi have paired semicircular 'brushes' of setae on maletergum II (Fig. 41). Damalinia ornata has sclerotised blunt projections on male terga II and III.These specialised setae and projections may in some way assist the male to hold the femaleduring copulation, although in no case have the setae or projections been observed to bedamaged, as they might be expected to be should they operate against the female abdomen, andthere is no observational evidence to support the hypothesis. Males of Neotrichodectes species have a pair of small median setae on terga II-VII , sometimesseparated by a seta of normal length (Fig. 229). This feature, not found in females or males ofany other group, is of unknown function. Female genitalia The female subgenital plate of Trichodectidae comprises either sternite VIII or sternites VIIIand VII fused. The posterior margin of sternum VIII forms the ventral margin of the vulva. Inmany species the margin is expanded posteriad and slightly laterad (Fig. 94), and in others the TRICHODECTID MAMMAL LICE 199 Table 2 Distribution of abdominal pleural projections in the Trichodectidae. For pleura II, III and IV anindication is given whether a dorsal ('D') or ventral ('V') projection is present, and whether thoseprojections are sclerotised ('S'). In each case the presence state is indicated by ' + '. Very lightsclerotisations are indicated by '?'. centre of the margin is greatly expanded into a flat lobe termed the 'subgenital lobe' (Figs 149,153, 175). The subgenital lobe appears to have evolved at least three times in the Trichodecti-dae, and its form is of taxonomic and systematic value. Its function, however, is not known. To each side of the vulva is a projection from sternum VIII, the gonapophysis of segment VIII(Lyal, in prep., b). The structure of this is variable. In all cases a basal internal apophysis isproduced, presumably as a muscle attachment (Keler, 1938a). The gonapophysis lies longitudi-nally, parallel to the abdomen; it may curve onto the dorsal surface apically. It may be long,slender and apically acute (Fig. 235), long and apically rectangular (Fig. 154), broad andmembranous (Fig. 243), produced into a rounded, toothed or rectangular medial lobe with anapical 'spur' (Figs 175, 233, 236), with the spur reduced (Fig. 237) or absent (Fig. 211),sometimes with serrations laterally (Fig. 211). Setae may be present on the gonapophyses andthe ventral vulval margin in various configurations, occasionally arising from sclerotisedtubercles. In some cases a sclerotised band links the gonapophyses and the vulval margin, but inmost cases this is not present and the connection is membranous. The curvature of the gonapophyses about both longitudinal and lateral axes produces acomplex three-dimensional structure difficult to interpret on slide-mounted specimens. Theorientation of the gonapophysis may be altered by pressure of the coverslip during mount-ing, especially if it has a membranous base. In addition, differences in orientation, either 200 C. H. C. LYAL natural or caused by the mounting process, may give the impression of very different shapes. Observations by Ferris (1951) and Murray (1957a, 19576) indicate that the gonapophyses areused, at least in some species, to trap the hair on which the egg is to be laid, guide the egg onto it,and mould the attaching cement. For Bovicola ovis the diameter of the hair is of importance indetermining whether the egg will be laid (Murray, 19576) and this is almost certainly detected bythe gonapophysis. There is thus selection pressure on the form of the gonapophysis relating tothe structure of the host hair. Posterior to the vulva and sometimes covered by the subgenital lobe is the 'sub-vulval area',which sometimes bears characteristic spines or scales. Also posterior to the vulva is a single sclerite which, as it cannot be homologised with eitherthe gonangulum or sternite IX + X (Lyal, in prep., 6), is termed the 'post-vulval sclerite'(Moreby, 1978). This sclerite may be single, medially divided, narrow and strip-like, broad andtriangular, fused to the postgenital pleurite, or absent (Figs 154, 175, 236, 237, 243). The female genital chamber, opening at the vulva, is oval, dorso-ventrally compressed, andlightly-sclerotised. The dorsal wall of the chamber may be heavily sculptured and bearsclerotised spicules, ridges or spines; this sculpturing may extend on to the postgenital sterna.The ventral wall may be similarly sculptured, although usually to a lesser extent; where thesubgenital lobe is present the sculpturing of the genital chamber may extend onto its dorsalsurface . The sculpturing of the walls of the chamber probably provides a highly f rictional surfaceagainst which the spicular surface of the endophallus acts to provide a firm union duringcopulation. The nature of the sculpturing and the distributions of spicules or spines may betaxonomically useful at the specific or subspecific level. The common oviduct opens into theanterior end of the genital chamber and curves sharply posteriad to lie dorsally to the chamber.Dorsal to the vulva the common oviduct curves sharply anteriad and divides into the two pairedoviducts (Fig. 20). The genital chamber, although assuming a more or less circular cross-sectionduring the passage of an egg and perhaps during copulation, is, when at rest, a dorso-ventrally gonapophysis VIII paired oviducts common oviduct sternum VII sternum VIM genital chamber Fig. 20 Diagrammatic three-dimensional representation of trichodectid female genital chamber andoviducts (internal), with terms used in this study. TRICHODECTID MAMMAL LICE 201 flattened, fairly rigid structure, the minimum width of which is governed by the diameter of theegg. The oviduct, however, is an elastic, folded membranous tube, compressed and folded toreduce volume, and expanding only to allow passage of the egg. The common oviduct must, atthe junction with the genital chamber, be the same diameter as the chamber. At its division intothe paired oviducts, however, it is narrow and greatly folded when at rest. Between these twopoints it narrows more or less abruptly, and folds may be apparent in its walls. Price & Emerson(1971) interpret these folds in Geomydoecus as striations of the dorsal wall of the genitalchamber, and term this apparently membranous structure the 'genital sac'. The true dorsal wallis interpreted as the ventral wall and the true ventral wall is apparently not observed. The formof the folds of the 'genital sac' (common oviduct), as well as its length and the width of itsjunction with the genital chamber, have been utilised by Price and his co-workers as specific andsubspecific characters in their re visionary work on Geomydoecus (e.g. Price, 1974; Price &Emerson, 1971; Price & Hellenthal, 1976; Timm & Price, 1980). The apparent constancy of thedimensions of the oviduct is related to the size of the egg and perhaps to the restrictions invariability of size of the genital chamber imposed by selection. The apparent constancy of thefolds is more puzzling although it is possible that tubes of identical length and diameter have anoptimum folding pattern if compressed in the same way. The wall of the common oviductmay also be lightly sclerotised, at least near to the junction with the genital chamber. Thefolds of the common oviduct have not been utilised as a taxonomic character elsewhere in thePhthiraptera. Blagoveshtchenski (1956) examines several species of Trichodectidae but fails to findevidence of a spermatheca in the family. In this study examination of slide-mounted specimensof most species has failed to reveal a spermatheca in any genus apart from problematically inDasyonyx, where a lightly-sclerotised sac, differing in form between species, is developed fromthe wall of the common oviduct. The relatively anterior position of this sac in comparison to thespermatheca of other Phthiraptera suggests that the two structures are not homologous. Nohistological or anatomical studies have been carried out on the sac. The male abdomen and copulation The opening of the male genital chamber ('genital opening') is always posterior to sternum IX,and primitively in the lice is ventrally positioned and distant from the anus, which is terminal.During copulation the male and female are usually oriented the same way, with male ventral tofemale. The tip of the male's abdomen is curved dorsally and anteriorly so that the genitalopenings of the male and female meet. The male genitalia consequently enter the female genitalchamber 'upside-down' , with the ventral side of the former coming into contact with the dorsalwall of the latter. Should sclerotised tergites be absent (as in Neotrichodectes and Geomydoecus)or greatly reduced (as in Trichodectes galictidis and Felicold) flexion of the male abdomen duringcopulation is evenly distributed along the membranous dorsal surface. If sclerotised tergites arepresent and fully-developed, however, flexion must be about the sclerite-membrane-scleritejoints of the dorsal surface, these thus functioning as 'hinges'. At each such hinge, there will beconsiderable deformation of the internal structures during flexion, whilst between the hingesthere will be no deformation. Increasing the number of hinges on the dorsal surface permitssmoother curvature of the abdomen and reduces the internal deformation at each hinge. Manymale Trichodectidae have tergal sclerites divided into an anterior and a posterior plate on atleast some segments (Figs 103, 105, 136), possibly for this reason. The degree of internaldeformation may be further reduced by modification of the shape of the plates. The form of platewhich, when placed in series, causes maximum internal deformation and requires maximummembrane area: sclerite area ratio for flexion is a simple rectangle, with anterior and posteriormargins at right angles to the long axis of the abdomen. These factors can be diminished byintroducing curvature in the anterior and posterior margins of the plates, either with bothmargins parallel or opposite, producing a series of alternating biconvex and biconcave plates(Lyal, 1983). Both these patterns are approached in the Trichodectidae (Figs 68, 103, 105). Theflexibility required of the ventral surface of the abdomen is developed, should the sterna be 202 C. H. C. LYAL sclerotised, by increasing the lengths of the sclerites and allowing them to overlap when theabdomen is at rest (Figs 103, 105). Despite the increased flexibility of the abdomen achievable by modifications of the tergal andsternal plates, the position taken by the male during copulation is compatible with a ventralgenital opening only if the abdomen is long and slender. A short, broad abdomen will notdeform sufficiently to turn back on itself and bring the male genital opening into contact with thevulva. Such a short, broad abdominal form has been developed several times in the Phthirap-tera, however, including at least once in the Trichodectidae. In most cases the limitationimposed on the degree of curvature of the abdomen has been met by a posterior or even dorsalmigration of the genital opening, thus reducing the degree of curvature necessary. Thisdisplacement has been effected by an increase in length of sterna VII, VIII and IX and aconcomitant shortening of the corresponding terga. In Trichodectidae this may lead to thedivision of tergite VIII longitudinally by segment IX (Fig. 212). The re-positioning of the genitalopening has increased its proximity to the anus, and in Trichodectidae the anus is containedwithin the genital chamber with the recuced segments X and XI projecting from the genitalopening (Fig. 19). The curvature required of the abdomen has been further limited in someTrichodectidae by two distinct adaptations. In some species of Felicola segment IX is developedinto a long, slender posterior process and the genital opening is dorsal and apical (Fig. 187). Thebasal apodeme and parameres are elongate and slender (Fig. 205). Most of the curvaturerequired is probably developed at the base of segment IX and the junction of the parameres andthe basal apodeme. Constriction of the endophallus at the latter fold is prevented by thepresence of a reduced, circular mesomeral arch, lacking an extension, which, being fused to theendophallus, prevents the structure from being pinched shut. In some Lorisicola speciessegment IX projects dorsally from segment VIII, the genital opening being dorsal (Fig. 114).This positioning limits the degree of curvature required of the abdomen by increasing theproximity of the genital opening to the vulva on minimum curvature. Male terminalia In Trichodectidae the subgenital plate is fundamentally uniform in construction, although greatsuperficial differences may exist between species. The plate is here considered to comprise eightdiscrete elements: sternites VII, VIII, IX (and, occasionally, VI); a pair of rods (referred to as'subgenital plate rods' or 's.g.p.r.' in this study) which lie laterally to the sternites and sometimesfuse them together (Figs 19, 165, 212); the two pleurites of segment IX, and the post-genitalsclerite, which is of uncertain homology (Fig. 19). Any of these elements may be present,reduced or absent, or fused to adjacent sclerites. The sternites and the post-genital sclerite maybe whole or medially divided. In cases of extreme reduction only the lateral rods may be left(Fig. 163) or all the sclerites absent (Fig. 187). In the most complete form (Fig. 68), all thesclerites are fused, forming a squared plate; usually there are membranous areas surroundingthe sternal setae ('perisetal gaps'), but these may be absent (Fig. 115). The subgenital plate rods on sternum VIII are always connected to the ventral wall of thegenital chamber in Trichodectidae, a feature also observed in some Anoplura. The function ofthis attachment is not known, although in species with ventrally-positioned parameres and amedian posterior extension to the mesomeral arch, the difference in lengths between the dorsalwall of the genital chamber and the membrane between the basal apodeme and the subgenitalplate might cause divergence of the apices of the parameres and mesomeral arch extensionduring extrusion of the genitalia to allow evertion of the endophallus. In some Trichodectidae the posterolateral angles of the subgenital plate are greatly extendedinto setose lobes termed 'styli' by Eichler (1963) (Figs 46, 68). Abdominal sternal sclerites areprobably derived from a fusion of the sternal plate and the coxal elements of the pairedabdominal appendages (coxopodites) (Matsuda, 1976; Richards & Davies, 1977). In thesubgenital plate of the Acercaria the paired nature of the gonocoxopodite component may beexpressed as a concavity of the posterior margin of the plate and concomitant projection of theposterolateral angles (Matsuda, 1976). True abdominal styli are derived from the paired TRICHODECTID MAMMAL LICE 203 abdominal appendages and are serially homologous with either the shaft of the thoracic legs orthe coxal styli (spurs of Matsuda, 1976) of the thoracic legs as present in the Machilidae(Thysanura) (Matsuda, 1976; Richards & Davies, 1977). Styli are not, therefore, homologouswith the coxopodites, but arise from them, being separated by a clear sulcus. Although true styliare present in some Psocoptera (Matsuda, 1976), the posterior extensions of the subgenital plateof Trichodectidae are not demarcated by a sulcus and thus are not homologous with true styli.The so-called styli of the Trichodectidae are an indication of the gonocoxopodite component ofthe subgenital plate, and are here termed 'pseudostyli'. Taxonomic use may be made of the formof the pseudostyli, which is very variable in the Trichodectidae, but the difficulty of assigningpolarity to transformation series limits the applicability of this character within phylogeneticstudies. The distribution of the pseudostyli through the Trichodectidae, however, is utilised inthe construction of the phylogeny of the family (see below). Male genitalia The structure of the male genitalia of lice and the homologies of the major parts are discussed byLyal (1983 and in prep., b). The genitalia consist of a more or less sclerotised basal apodemesupporting caudally a pair of parameres which may be fused ventrally and a pair of mesomereswhich may be fused dorsally; fused to the interior faces of the parameres and mesomeres is thepermanently-everted portion of the eversible endophallus (Lyal, 1983 and in prep., b}. In mostTrichodectidae both parameres and mesomeres are present, the latter usually being fusedapically (Figs 170, 198), the fused portion generally being extended posteriorly (Fig. 224). Theparameres may also be fused, forming a median ventral plate (Fig. 250). The parameres andmesomeres may meet the basal apodeme together (Fig. 170) or separately (Fig. 250). The fullrange of variation of the male genitalia is discussed in the character analysis and taxonomysections below. During copulation the endophallus everts into the female genital chamber. As describedabove, the interior of the femlae genital chamber is roughened and lined with scales. Theendophallus is likewise roughened, being covered with small chitinous spicules or largersclerites. The probable function of this adaptation is to maintain a firm connection between themale and female genitalia during copulation. The form of the endophallus and the distribution ofthe spicules and sclerites are species-specific and very variable, possibly functioning as pre-zygotic isolating barriers. Character survey As noted above, the characters and character states listed below were selected from a muchlarger set of characters and states. In some cases reasons for rejection are discussed below. Thefull list of characters and the complete data matrix are presented in Lyal (1983) and aredeposited in the library of the British Museum (Natural History). 1 Posterior of basal apodeme symmetric in vertical planelateral struts ('b.a.l.s.') asymmetric in vertical plane (Figs 193, 195) 1 g 2 Anterior end of basal flat, convex, concave apodeme (Figs 91, 200, 250) deeply concave (Fig. 52) 1 g acuminate (Fig. 173) 1' g 3 Posterior bifurcation of absent b.a.l.s. present (Fig. 173) 1 g 4 Posterior of b.a.l.s. lacking lateral extension with lateral extension to mesomeres (Fig. 239) 1 g 5 Anteposterior spur of absent b.a.l.s. present (Fig. 83) 1 g 2046 Shape of basal apodeme Posterior of b.a.l.s. 8 Posterior of b.a.l.s. Posterior of b.a.l.s. 10 Basiparameral sclerites 11 Paramere fusion 12 Paramere fusion 13 Paramere fusion 14 Parameral apices 15 Paramere shape 16 Paramere shape 17 Parameral plate shape 18 Parameral orientation 19 Paramere shape 20 Paramere shape 21 Paramere shape C. H. C. LYAL not long and 'waisted'very long, with median'waist' (Fig. 119) not modified as belowbroad and obtuse in meetingparameres (Fig. 198) not modified as belowsharply inturned and convex(Fig. 200) not modified as belowincurved to parameres (Figs144, 145) absent present, fused to parameres present, free not fused to mesomeral archfused to mesomeral arch inpart (Figs 60, 75)completely fused tomesomeres (Fig. 74) not fused to basal apodemefused exteriorly to basalapodeme (Fig. 75)fused medially or completelyto basal apodeme (Figs 145,172, 174) not fused together or closely associated fused together not fused but closely associated, with line of division apparent (Figs 223, 224) unfusedfused not as described belowfused to shield-shaped plate(Figs 239, 242) not as described belowfused as described incharacter 15, with antero-median projection (Fig.238) not as described belowproduced apically intoincurving points (Fig. 171) similar at right-angles to each other (Fig. 170) more or less broad, thickvery thin, deflectedasymmetrically (Fig. 174) more or less broadnarrow rods (Figs 196, 197) not as described below r o i r 1 1 TRICHODECTID MAMMAL LICE 205 22 Median internal projectionof parameres 23 Paramere size 24 Paramere shape 25 Paramere shape 26 Paramere shape 27 Base of parameres 28 Base of parameres 29 Paramere and mesomereshape 30 Reduction of parameres andmesomeres 31 Mesomeres 32 Mesomeral position 33 Mesomere fusion 34 Lateral desclerotisationsof mesomeral arch 35 Lateral flexions ofmesomeral arch 36 Mesomeral arch 37 Mesomeres basally, betweenb.a.l.s. 38 Mesomeral arch mesally 39 Diameter of mesomeral arch very broad, lanceolate, scoop-shaped (Fig. 172) 1 g absent present 1 g large or moderate small discs (Fig. 84) 1 1 not as described below cylindrical (Figs 169, 205) 1 g not as described below basally very narrow, medially broad (Fig. 198) 1 g not as described below with characterically- differentiated base and blade (Fig. 51) 1 g not as described below broad, club-like (Fig. 192) 1 g block-like (Fig. 201) 1' g cuboid (Fig. 52) 1" g lacking flange with flange (Fig. 118) 1 g not as described below of characteristic asymmetric form (Fig. 249) 1 g not reduced as below characteristically reduced (Fig. 52) 1 1 characteristically greatly reduced (Fig. 53) 2 1 present absent 1 1 reach or fail to reach b.a.l.s. extend mesad of b.a.l.s. 1 g apically fused to form arch not apically fused 1 1 absent present (Fig. 118) 1 g absent present (Figs 107, 225) 1 g not as described below modified into tripartite arch (Fig. 93) 1 g not modified as below sharply directed posteriad (Fig. 224) 1 g smooth or with projection with two nipples (Fig. 221) 1 g less than half the length of the permanently-everted endophallus more than half the length of the permanently-everted endophallus 1 g 206 40 41 42 4344 45 46 47484950515253 54 55 56 5758 Shape of mesomeral arch Shape of mesomeral archShape of mesomeral arch Shape of mesomeral archShape of mesomeral arch Shape of mesomeral arch Shape of mesomeres Median extension ofmesomeral arch Apex of mesomeral archextension Shape of mesomeral archextension Tongue-like sclerite betweenparameres Everted part of endophallusEndophallus spiculationEndophallus spiculation Endophallus spiculation Endophallus spiculation Endophallus spiculationGonapophysis setaeGonapophysis setae C. H. C. LYAL as great as the length of the permanently-everted endophallus 2 not as described below with median anterior deflection (Figs 191, 192) 1 circular or elliptical rectangular (Fig. 198) 1 not as described below extension lost, arch 'looped' (Fig. 247) 1 not as described below widely circular (Fig. 55) 1 not as described below pentagonal, convex distally (Fig. 62) 1 smoothly curved sharply inturned to parameres (Fig. 81) 1 extending anteriad to posterior end of basal apodeme and sharply recurved (Figs 82, 219) not as described belowvery slender, string-like(Fig. 52) presentabsent not bifurcatebifurcate not as described belowbroadly expanded, lanceolate absentpresent not sclerotisedsclerotised not as described belownumerous 'V'-shaped rods not as described below comprising large hook-like spines (Fig. 107) 1 not concentrated about gonopore concentrated about gonopore 1 with dense 'V'-shaped patch about gonopore 2 not as described below including median row of hook-like scales 1 not as described below dense and refrigent in part 1 present absent 1 lacking sclerotised tubercles TRICHODECTID MAMMAL LICE 59 60 61 62 63 6465 66 67 6869 70 71 7273 Gonapophysis setaltubercles Gonapophysis setaltubercles Gonapophysis shapeGonapophysis shape Gonapophysis shape Gonapophysis lobeApex of gonapophysis Gonapophysis lobeGonapophysis lobe Gonapophysis lobeGonapophysis lobe Gonapophysis lobe Gonapophysis junctionwith ventral vulval margin Ventral vulval marginVentral vulval margin some having sclerotised basaltubercles (Fig. 150) absent, or present and notcharacteristically fusedpresent and characteristicallyfused (Fig. Ill) absent, or present and incharacteristic pattern (Fig. 153)present and modified fromcharacteristic pattern byloss of apical non-tuberculateseta (Fig. 146) not as described belowspoon-shaped (Fig. 156) not as described belowhook-shaped (Fig. 66)hook-shaped with spur (Fig.67) not explanate ventrally(other than as discrete lobe)thinly explanate ventrally(other than as discrete lobe) absentpresent acute or obtuse not projecting beyond lobe (if present); apparently absent not very thickvery thick absent, or not as described below broad, formed of fused tubercles (Fig. 233) absent, or not as described below narrow, rectangular, formed from 2 or 3 fused tubercles (Fig. 177) negative for character 67 or, if positive, flattened positive for 67 and folded anteriorly positive for 67 and greatly folded absent, or not as described below produced into spines posteriorly (Fig. 211) acute smoothly continuous unsclerotisedsclerotised no shorter than length ofgonapophyses 207g r 208 74 7576 777879 80 C. H. C. LYAL 81 8283 8485 86 87888990 9192 Ventral vulval margin Ventral vulval marginVentral vulval margin Ventral vulval marginSubgenital lobeSubgenital lobe margin Subgenital lobe surface Subgenital lobe, basally Subgenital lobe processes(ch. 81) Subgenital lobe processes(ch. 81) Subgenital lobe Subgenital lobe bifurcations(ch. 84) Subgenital lobe Sub-vulval areaDorsal vulval facePost-vulval scleritesPost-vulval sclerites Common oviductFemale genital chamber shorter than length of gonapophyses (Fig. 102) 1 lacking setal tubercles with setal tubercles (Fig. 156) 1 not expanded expanded (Figs 94, 126) 1 not as described below greatly produced and rounded (Fig. 113) 1 not as described below with median narrow projection 1 absent present 1 not serrate serrate 1 very serrate 2 smooth dorsally and ventrally with overlapping pointed scales (Fig. 210) 1 with small spines (Fig. 208) 2 with many overlapping spines (Fig. 207) 3 lacking lateral processes with lateral processes 1 not as described below thinly sclerotised and directed posteriad 1 not as described below membranous and serrate (Figs 155, 157) 1 not bifurcate (Fig. 210) bifurcate (Fig. 208) 1 not as described below rectangular, widely separate (Fig. 176) 1 lacking submarginal setal patch with submarginal setal patch 1 lacking pointed scales with pointed scales 1 not spinous spinous 1 present absent 1 broad, triangular long, narrow, oriented parallel to longitudinal axis 1 not as described below with folded 'collar' at branching-point 1 lined with overlapping scales TRICHODECTID MAMMAL LICE 93 Female genital chamber 94 Female genital chamber 95 Female genital chamber 96 Female genital chamber 97 Female sternum VII 98 Reproduction 99 Male tergite VIII (ifpresent) 100 Posterior margins of maletergum IX 101 Male segment IX 102 Pseudostyli 103 Pseudostyli (if present) 104 Pseudostyli (if present) 105 Male segment IX 106 Male sternite VIII 107 Male sternite VIII(if present) 108 Male perisetal gap 109 Male sternum VI lined with sclerotised nodules,some fused not as described below with median dorsal longitudinal fold not as described below with median dorsal area clear of scales not as described belowwith median anterior dorsalarea clear, thinlysclerotised not as described belowwith spines on dorsal face lacking processeswith two long spikes onposterior margin (Fig. 73) sexualparthenogenetic with posterior element absentor, if present, not fused totergite IX with posterior element fusedto tergite IX not as described belowgreatly expanded (Fig. 79) without two longitudinalstrengthening scleriteswith two longitudinalsclerites absentpresent not as described belowbroad, rounded, long (Fig.68) rounded apicallyangular, pointed apically(Fig. 78) lacking single apical projection with single apical projection (?fused pseudo- styli) not as described belowcharacteristically enlarged(Fig. 159) not convex posteriorlycharacteristically convexposteriorly (Fig. 218) presentabsent not as described belowwith anterior and posteriorsclerites 210 110 Female flagellomeres 111 Male flagellomeres 112 Male scape C. H. C. LYAL 113 115 Male scape setae 114 Male scape setal row (if present) Male flagellum 116 Male flagellar 'teeth' 117 Male flagellar 'teeth' 118 Male flagellum 119 Male flagellum 120 Inner face of maleflagellum 121 Male scape 122 Female pedicel 123 Male flagellum 124 Flagellar sensilla placodeaand coeloconica 125 Sitophore sclerite 126 Posterior temple angles 127 Posterior temple margins unfused (three flagellomeres) fused to form two flagellomeres 1 fused to form one flagellomere 2 unfused fused 1 greatly or slightly expanded not expanded 1 randomly scattered on posterior face in longitudinal row on posterior face 1 numbering at least three setae on posterior face reduced to two setae on posterior face lacking apical 'teeth'with single apical 'tooth'with two apical 'teeth'with three apical 'teeth'with four apical 'teeth'with eight apical 'teeth' not on protuberanceon protuberancearticulated basallyfused to flagellumlacking basal projectionwith simple basal projectionwith basal projection oflinked 'teeth'with broad, rough, basalprojection not as described belowwith simple median and basalprojections only not serrateserrate lacking apical projectionwith apical projection lacking projections with membranous projection not very longvery long not in pit in pit with marginal tongue-like processes not as described below with posterior arms extended (Fig. 12) 1 lacking projections with projections (Fig. 129) 1 with very long projections(Fig. 138) not as described below produced and convex 1 TRICHODECTID MAMMAL LICE 128 Osculum 129 Pretarsus 130 Pretarsal claws 131 Post-coxaleoflegUI 132 Sternum II 133 Atrium of thoracic spiracle 134 Abdominal spiracles 135 Abdominal spiracle VIII 136 Abdominal spiracle VII 137 Abdominal spiracle VI 138 Abdominal spiracle III 139 Setae of posterior setal rowof pleurum III 140 Pleurum VII 141 Abdominal tergal setae 142 Male tergum II 143 Male tergum III 144 Male terga II-IV not deep deep and with characteristicanterior marginal convexity(Fig. 69) bearing two clawsbearing one claw lacking ventral spineswith blunt ventral spineswith sharp ventral spines not fused to abdominal pleurum II fused to abdominal pleurum II, at least in female not as described belowwith sclerotised apophysisarticulated to pleurum II sphericaltubular all of similar size spiracle VIII of male very small spiracles VII and VIII very small in both sexes presentabsent presentabsent presentabsent presentabsent not stoutstout lacking tuft of very long setae with tuft of very long setae (Figs 88, 90) short or of medium length(Figs 105, 188) generally very long, obscuringp.l.s. (if present) (Fig. 158) lacking specialised setae asdescribed belowwith long stout setae notfound on other terga (Figs180, 188) lacking specialised setae asdescribed belowwith isolated pair of longmedian setae, longer thanthose of tergum IV (Fig. 187) without specialised setae asdescribed beloweach with single pair oflong median setae (Fig. 182) 1 1 1r 211 212 145 Female terga I-VIII 146 Female terga I-IV 147 Male terga II and III C. H. C. LYAL 148 149 150 151 152 153154155 156 157 158 159160 Female terga I-VIII Male terga II-III Male tergum II Postero-lateral setae Male terga II-VI Setae of abdominal sternaII-IV Female abdominal setaeSetae of male tergum II Abdominal setal basesAbdominal pleurum II Dorsal projection ofpleurum II Dorsal projection ofpleurum IV Ventral projection of without specialised setae as described below each with single pair of long stout median setae 1 with median setae lacking median setae 1 without specialised setae as described below with group of characteristically specialised long setae (Fig. 178) 1 without specialised setae as described below each with single pair of median setae 1 without specialised setae as described below with long setae arranged in straight rows of four or more (Figs 161, 244) 1 without specialised setae as described below with long setae arranged in curved rows (Fig. 41) 1 absent or possibly present but not clearly distinguished from lateral setal group clearly present on terga II-VI 1 with two median setae the same size as other setae of median group with two median setae appreciably smaller than other setae of median group 1 not as described below short, stout (Fig. 136) 1 not as described below very long, fine 1 not arising from modifiedsclerite arising from very long,medially-divided sclerite 1 not enlarged enlarged 1 not as described below extending narrowly onto sternum II 1 extending broadly onto sternum II 1 ' absent present, unsclerotised 1 present, sclerotised 2 absent present 1 absent TRICHODECTID MAMMAL LICE 213 pleurum IV present 1 g 161 Projections of pleurum not as described below IV (if present) very long (Fig. 136) 1 g 162 Maleterga VI-VIII (or VI, not as described belowif VII-VIII without with anterior sclerite sclerites) longitudinally divided medially 1 g 163 Male terga VI-VII not as described below with posterior scleritelongitudinally dividedmedially 1 g 164 Abdominal lateral flecks absent present 1 g Character analysis The character analysis is performed in two interconnected parts, the determination of characterpolarity and the construction of the cladogram. These are linked through the process of'reciprocal illumination' (Hennig, 1966), and thus, although they are considered individually inthe two sections below, there is some interaction between the two processes which will bemanifested in discussion. The following two sections are intended to explain the reasons for thepolarity assigned to the characters listed above, and for those groups developed in the cladogramwhere characters are apparently arranged non-parsimoniously. To increase clarity and conciseness the distributions of characters and character statesdiscussed below are related to taxa and clades developed in the analysis. Clades are referred toby the names of the taxa (species or genera) on the extreme left (top) and right (bottom) of theclade as depicted on the cladogram (Figs 23 to 34), reading from left (top) to right (bottom). Identification of apomorphic states In many cases an apomorphic character state is identified as such by its distinct complexity andvery limited distribution and, to avoid pointless repetition in the following discussion, suchinstances are not examined individually. The characters are examined under the following headings: Male genitalia (characters 1-56); Female genitalia (characters 57-97); Reproduction (character98); Male terminal abdominal segments (characters 99-109); Antennae (characters 110-124);Head (characters 125-128); Legs (characters 129-130); Postcoxale (character 131); Spiracles(characters 133-138); Abdominal setae (characters 139-156); Abdominal pleural projectionsand modifications (characters 157-161); Abdominal sclerae (characters 132, 162-164). Male genitalia (characters 1-56) Most species of Psocodea have symmetric male genitalia and this state is consequently assumedto be plesiomorphic for the superorder. In a few species of Trichodectidae the genitalia areasymmetric, but differences in the form of the asymmetry in different species (Figs 170, 174, 249)suggest that several independent autapomorphies have been developed. In some cases acharacteristic asymmetry is limited to a single species and is therefore of no relevance to phyleticanalysis, but the asymmetries described in characters 1, 18, 19 and 29 are more widelydistributed and are all employed. The distribution of other apomorphies indicates that character1, the vertical asymmetric deflection of the lateral struts of the basal apodeme ('b.a.l.s.'), hasbeen developed twice, once in Felicola (S.) bedfordi (Fig. 195), and once in the commonancestor of F. (F.) cynictis and F. (F.) setosus (Fig. 193). In most Psocodea the basal apodeme is not fused to the parameres, but in a few Trichodecti-dae this fusion, considered to be apomorphic, has taken place (character 12). In the Damalinia(D.) theileri-harrisoni clade the posterior ends of the b.a.l.s. are broad and fused exteriorly tothe parameres (Fig. 75); this fusion (character 12:1) is unlike that found elsewhere in the family 214 C. H. C. LYAL and is believed on that account to be autapomorphic. In the Trichodectes (S.) retusus-mustelaeclade, the T. (S.)fallax-potus clade, Neolutridia lutrae and Lutridia exilis the parameres are alsofused to the b.a.l.s. (character 12:1'), but there are no features in the fusion pattern to indicatewhether the apomorphy is homologous or convergent in the four clades. There are a number ofpossible sequences of gains and reversals. The fusion may have taken place three or four timeswith no reversals; once, to be lost at least four times; or twice, to be lost at least twice. Thegenitalia of Lutridia spp. differ from those found in the sister-group (the Trichodectes-Neolutridia clade), being more similar to those of Protelicola. Comparison of L. exilis and L.matschiei (Figs 144, 145) indicates fusion of the parameres and b.a.l.s. in the former species to beassociated with the virtual detachment of the basal fused portion of the parameres, a uniquefeature. For this reason the fusion in L. exilis is deemed to be autapomorphic. The Trichodectes(S.) retusus-mustelae clade and the T. (S.) fallax-potus clade are placed by character 144 in atrichotomy with T. (S.) emery i, a species in which the parameres are not fused to the b.a.l.s. Ifthe fusion is homologous in Trichodectes (Stachiella) and Neolutridia, then it must have been lostin Werneckodectes, Trichodectes (Trichodectes) and T. (Paratrichodectes), and T. (S.) emeryi.Parsimony suggests that fusion was developed independently in Neolutridia and the commonancestor of the two clades in Trichodectes (Stachiella) that possess the character, these latterbeing united as sister-groups. Fusion of the parameres to the b.a.l.s. is thus postulated to havetaken place four times in the Trichodectidae, three of those times in the Trichodectes-Lutridiaclade. In most Psocodea the parameres are not fused together, but such fusion is present, probablyapomorphically, in a number of species of Trichodectidae (character 13:1). In some speciesparameral fusion is difficult to observe, as the portion of the permanently-everted endophalluslying between the parameres is faintly sclerotised, giving the impression that the parameres arefused together; fusion has probably developed in some cases through sclerotisation of theendophallus. The distribution of other apomorphies suggests that parameral fusion exhibitsmore homoplasy than any other apomorphy in the analysis, being derived 24 times and lost once.In the Lorisicola (P.) bengalensis-juccii clade the parameres are closely associated with eachother but are not fused (Figs 223, 224), although fusion has sometimes been assumed (e.g.Werneck, 1948). This proximity is believed to be autapomorphic for the clade (character 13:1').The form of the fused parameres (parameral plate) may be apomorphic for groups of species(characters 15, 17). In a few Trichodectidae the parameres and mesomeres are fused, a probable apomorphy(character 11). The distribution of other apomorphies and slight differences in the fusion pattern(Figs 60, 74, 75, 81, 225) indicate some homoplasy in the character. As described above, the mesomeres are frequently fused apically in the Psocodea, andconsequently this fusion, when found in the Trichodectidae, is deemed to be plesiomorphic.Loss effusion (i.e. reduction to two unfused mesomeres) is therefore believed to be apomorphicwithin the Trichodectidae (character 33), and distribution relative to other apomorphiesindicates that it has occurred several times in the family (see cladogram). In most species of theProcaviphilus-Eurytrichodectes clade there is a lateral desclerotisation on each side of themesomeral arch (Figs 118, 122, 135). This character state is not found elsewhere and this, itsstructure, and the distribution of other apomorphies, all indicate its apomorphic status(character 34). The mesomeral arch may also have lateral flexions (Figs 107, 108, 225), whichcan give the arch the appearance of being broken (see Werneck, 1948). This modification isfound in species of the Lorisicola (P.) lenicornis-neoafricanus clade and of Procavicola( Condylocephalus) , which on the basis of other apomorphies are widely separated on thecladogram; the lateral flexion of the mesomeral arch (character 35) is consequently believed tobe a convergent apomorphy in the two clades named. In most species of the genus Eutrichophi-lus the mesomeral arch is divided into three parts by total desclerotisations laterally (Figs 91,93). This feature is unique within the Phthiraptera and therefore considered apomorphic(character 36). In many Trichodectidae a rod-like sclerite terminating posteriorly in a Y-shape or a broadplate is present longitudinally between the b.a.l.s. (Figs 55, 82, 148). Although this sclerite (the TRICHODECTID MAMMAL LICE 215 'central sclerotisation') is very clear in some species, it is poorly sclerotised in others, and may beeither absent or obscured by the sclerotisations of the endophallus in mounted specimens. Theuncertainty attached to the observation of this structure has precluded its use in analysis,although it may be of value taxonomically. The central sclerotisation is probably a derivative ofthe basal apodeme, developed for muscle attachment. A sclerite or pair of sclerites is present anteriorly to the parameres in some Trichodectidae,and are referred to here as the 'basiparameral sclerites' (Figs 107, 225) (character 10: 1+2). Theymay be fused to the parameres (Fig. 225) or to each other (Fig. 93). Their presence is believed tobe apomorphic and their derivation is probably from the basal ventral flanges of the parameresfound apomorphically (character 28) in a number of Trichodectidae (Figs 118, 122). If they areformed (as suggested here) by detachment of the flange from the main body of the paramere(Fig. 21), fusion of the sclerites and the parameres is a stage in the transformation series to thedevelopment of free sclerites, but fusion of the basiparameral sclerites to each other is a'terminal' apomorphic state (although it is not used in the cladistic analysis because of itsextremely limited distribution). The anterior end of the basal apodeme may be heavily or lightly sclerotised, or apparently notsclerotised at all. This degree of sclerotisation is very susceptible to modification duringpreparation of the specimen, and thus is difficult to assess accurately. The character is not used inanalysis. The most prominent features of the basal apodeme are the lateral struts (b.a.l.s.),which are generally fairly heavily sclerotised. These struts may approach the anterior of thebasal apodeme in parallel, convergently or divergently; this character is not used in cladisticanalysis because of the difficulty in assigning polarity to the different forms, but is usefultaxonomically, and can help in the determination of the form of the anterior margin of the basalapodeme. This anterior margin may be straight or broadly convex, shallowly concave, verydeeply concave (Fig. 52) or acuminate (Fig. 173) (character 2). Of these forms the last two arealmost certainly apomorphic within the Trichodectidae; the elongation of the apodeme and theconcomitant parallel-sided concavity is found only in the Bovicola alpinus-tibialis clade(character 2:1), and the acuminate form is found only in three species of Trichodectes , althoughin fact the latter apomorphy imparts little useful information for the construction of thecladogram. The polarity of the character for the other three states is difficult to assess, and theyare therefore not used in phyletic analysis. In some species of Damalinia (Tricholipeurus) the b.a.l.s. develop a lateral spur before thejunction with the parameres, probably at the point at which the dorsal and ventral layers of thebasal apodeme separate; this feature (the 'anteposterior spur', Fig. 83) is not found elsewhere inthe Trichodectidae and is considered apomorphic (character 5). The posterior ends of theb.a.l.s. are most frequently not, or only slightly, expanded laterally, but in some Trichodectidaethey are greatly broadened and scoop-shaped. This very broad form, whilst believed to beapomorphic, is not used as an apomorphy in phyletic analysis because of difficulties in delimitingthe state. The posterior forking of the b.a.l.s. (Fig. 173) is also considered apomorphic(character 3), but was probably developed twice in Trichodectes (see cladogram). The basalapodeme probably extends anteriorly as far as segment VI in the plesiomorphic state, but in afew Trichodectidae it extends up to segment II; in some cases this lengthening has beenaccompanied by a width restriction or 'waisting' medially, and this is believed to be apomorphic(character 6). Other features of the basal apodeme are, by virtue of their restricted distributionand concordance with the distributions of other apomorphies, believed to be apomorphic(characters 4, 7, 8, 9). Whilst it is not possible to be certain of the plesiomorphic form of the parameres in theTrichodectidae, it is assumed that this is fairly unspecialised, and that the forms of the parameresfound in groups of species that are also linked by other apomorphies are apomorphic (characters20-27). As noted above, apical fusion of the mesomeres is plesiomorphic for the Trichodectidae, soloss of apical fusion, reduction in size and complete loss of the mesomeres are all considered tobe apomorphic within the family (characters 30, 31, 33). The presence of a median longitudinalextension to the mesomeral arch is also considered plesiomorphic, as it is present in a number of 216 C. H. C. LYAL taxa outside the Trichodectidae. Loss of this extension, or modification of its form from 'simplelanceolate' (Fig. 118), are considered apomorphic within the family (characters 47-49). In theplesiomorphic state the mesomeres articulate basally with the basal apodeme; articulation of thebasal apodeme with any other part of the mesomeres is considered apomorphic. Mesadextension of the mesomeres between the b.a.l.s. (character 32) has apparently arisen twice, oncein the Dasyonyginae, and once in the Lorisicola, and in each case providing an autapomorphyfor the clade named. In Lorisicola (P.) bengalensis andphilippinensis the parts of the mesomeresbetween the b.a.l.s. are apomorphically deflected posteriad (Fig. 224) (character 37). A similarrecurving of the mesomeres occurs elsewhere in the Trichodectidae (character 45), but in thiscase the mesomeres are exterior to the b.a.l.s., and their recurved portions lie ventrally to theb.a.l.s. This apomorphy is seen as a transformation series of states in Damalinia (Tricho-lipeurus) (character 45:1-45:2), the most apomorphic of which (45:2) is also exhibited byLorisicola mjoebergi, although in this case the recurved parts of the mesomeres are very difficultto see (Fig. 219). Other modifications to the mesomeres believed to be apomorphic are presentin restricted groups within the Trichodectidae (characters 38-46). The endophallus may be sclerotised in a number of apomorphic ways within the Trichodecti-dae (characters 51-56). Female genitalia (characters 57-97) The gonapophyses of most Trichodectidae and many other Psocodea bear at least some setae,which arise directly from the structure and not from tubercles. Absence of setae (character 57)and development of sclerotised setal tubercles (character 58) are therefore both believed to beapomorphic within the Trichodectidae. Setal tubercles are found in Protelicola, Procaviphilusand the Trichodectes-Neolutridia clade (T-N), but as a sister-group relationship between thelatter two is not supported by other apomorphies, and the form of the tubercles differs betweenthe two clades (Figs 111, 157), the character is probably convergent. The relationship betweenProtelicola and the T-N clade is discussed below. The characteristic pattern taken by thetubercles in each clade is modified by loss (character 60) or fusion (character 59); in both casesthese are believed to be apomorphic modifications because of their concordance with otherapomorphies. Tuberculate setae are also found on the ventral vulval margin of most species inTrichodectes (Fig. 157); concordance with other apomorphies suggests the apomorphy of thischaracter (character 74). The plesiomorphic form of the gonapophyses is not certain, but some forms, because of theirvery restricted distribution, are believed to be apomorphic (characters 61, 62, 63); someconvergence in character 62 is indicated by the distribution of other apomorphies. Thedevelopment of a lobe on the ventral margin of the gonapophysis is restricted to the Trichodecti-dae and, for this reason, is believed to be apomorphic within the clade (character 64).Distribution of other apomorphies suggests that the lobe developed independently in severaldifferent clades, sometimes taking only one form in a clade (character 70), sometimes beingapomorphically modified (characters 66-69). The reduction of the 'spur' - the portion of thegonapophysis distal to the lobe - is considered apomorphic, as it is confined to two small groupsof taxa within the family (character 65). Where present, the gonapophyses in most Phthirapterameet the ventral vulval margin at an angle (Figs 94, 175), but in some Trichodectidae they meetin a smooth curve (Fig. 154), which may be sclerotised (characters 71, 72). The ventral vulvalmargin may extend in a more or less smooth curve between the gonapophyses, as is mostfrequently the case in lice with gonapophyses, or it may be produced in some manner (characters75-78). Each of these projections is considered apomorphic, although the distribution of otherapomorphies suggests that some are homoplastically developed in different clades of the family.The distinction between the subgenital lobe (character 78) and the expansion of the ventralvulval margin (character 75, and its apomorphic derivative, character 76), may not be im-mediately clear, but whilst the former term is applied to structures that arise abruptly from themargin, the latter is a more extensive posterior production of the whole of the margin. Both ofthese apomorphies occur more than once in the Trichodectidae. The form of the subgenital lobeis variable, though frequently it is marginally serrate, sometimes with the serrations greatly TRICHODECTID MAMMAL LICE 217 developed (character 79:2). Several other apomorphies, of restricted distribution within theTrichodectidae, are found in the form of the subgenital lobe (characters 80-83, 85, 86). Twoprobable apomorphies, the presence of an internal sclerite in the subgenital lobe and thepresence of lateral setal patch, are not used in the cladistic analysis. The internal sclerite is notreadily observable, but this is probably due to the sclerite being rendered undetectable duringpreparation of the specimens (especially the smaller species), and this likelihood precludes itsuse. The distribution of the sclerite, where detected, suggests it to be plesiomorphic within theTrichodectinae, possibly linked to the development of the subgenital lobe in this clade. Thelateral marginal setae appear to share part of the distribution of the lateral processes of the lobe(character 81), being absent in a few species only, and the two characters are probably closelyassociated; for this reason the setal character is not used. The plesiomorphic form of the genital chamber in the Trichodectidae is not known, butobservations on other Phthiraptera suggest light sclerotisation with a few internal spicules. Thedevelopment, in some restricted groups of Trichodectidae, of particular patterns of spicules,scales, spines and broad sclerotised areas (characters 92-96) is considered apomorphic. Reproduction (character 98) Parthenogenetic reproduction (character 98) occurs in a few Trichodectidae, mostly in theBovicolinae, but also in the species Geomydoecusscleritus. As all other Phthiraptera reproducebisexually, the character is taken as apomorphic. Parthenogenesis appears to have developed atleast four times in the Trichodectidae. Male terminal abdominal segments (characters 99-109) The sclerites of the terminal segments of the male trichodectid abdomen are very variable inpresence or absence states, extent, and degree of subdivision. This variability is not, in manycases, readily associated with transformation series of other characters to which polarity hasbeen applied, and the plesiomorphic state (and hence apomorphic states) of the characters ofthese sclerites is, in most cases, not known. In a few instances the sclerites are distinctly modifiedin a restricted group of species, and thus polarity can be assigned (characters 99, 101, 106, 107,108, 109). Segment IX and the genital opening are apomorphically positioned more or less dorsally inmany species of Trichodectidae, as discussed above. Distortion due to preparative processesobscures the position in many of the specimens examined, however, and characters associatedwith this positioning cannot be used with any confidence, and are excluded from analysis. In some species of Trichodectidae the posterior margin of tergum IX is greatly expanded toproduce a double convex lobe (character 100); this development is believed to be apomorphic,although the distribution of other apomorphies indicates that it developed twice. The presence of pseudostyli, discussed in detail above, is believed to be apomorphic for theTrichodectidae (character 102). The plesiomorphic form of the pseudostyli is not known, as theextant forms are very variable and cannot, in most cases, be resolved into transformation series.In two cases (characters 103, 104), the pseudostyli are of very distinct form and restricted togroups of species believed to be holophyletic on other grounds; these character states arebelieved to be apomorphic. The presence of a single projection posteriorly on segment IX insome species of Trichodectidae, believed to be formed of fused pseudostyli, is also consideredapomorphic (character 105). Although found in only three species, the distribution of otherapomorphies suggests two independent developments of this character state. Antennae (characters 110-124) In all Trichodectidae the male flagellomeres are fused together (character 111), a state foundelsewhere only in the anopluran families Echinophthiriidae and Hamophthiriidae and thereforeconsidered apomorphic for the Trichodectidae. Fusion of the flagellomeres has also occurred insome female Trichodectidae (character 110), but the distribution of other apomorphies suggestsfour homoplastic derivations of this apomorphy in the family. The expansion of the scape in themale to house the enlarged musculature is probably plesiomorphic for the Trichodectidae, as 218 C. H. C. LYAL similar expansion is found in many Ischnocera and Anoplura. Reduction of this expansion is,however, apomorphic within the family (character 112), and is believed to have taken placethree times. The flagellum of most male Trichodectidae bears a number of setae modified intobroad pointed 'teeth' (Figs 13, 231), a feature not found in the same form in any otherPhthiraptera, and therefore considered apomorphic for the family. The plesiomorphic numberof 'teeth' is almost certainly two, as this number is the most common in all groups ofTrichodectidae; any variation from this number (to zero, one, three, four or eight) is believed,therefore, to be apomorphic (character 115). The loss of the basal articulation of the 'teeth'(character 117) and the development of a supporting protuberance (character 116) are bothapomorphic. In order that the male antennae should clasp the female with maximum efficiencythe 'inner' (posterior) surface of some or all of the antennal segments may be roughened or bearprojections; such developments are considered apomorphic in each form (characters 118-121).The presence of a membranous projection on the female antenna (character 122) is also believedto be apomorphic. In most Phthiraptera the setae of the scape are scattered over its surface in nocoherent pattern, and this is true of some Trichodectidae (Fig. 70); in most Trichodectidae,however, the setae of the dorsoposterior surface of the scape are apomorphically arranged in aline along the segment (Fig. 13) (character 113). In some cases where a row of setae might beexpected from the construction of the cladogram, the number of the setae involved in theputative row is only two, and this is believed to represent an apomorphic reduction in number(character 118). The sensilla of the antennae in Trichodectidae and other Phthiraptera have been discussedabove; the presence of a fringed pit surrounding the sensilla of the flagellum in two species ofTrichodectidae is unique and believed to be apomorphic (character 124). Head (characters 125-128) Although the sitophore sclerite is variable in most Phthiraptera (Haub, 1973), it is comparative-ly uniform in the Trichodectidae. The form found in most Trichodectidae (Fig. 11) is believed onthis account to be plesiomorphic, and is departed from in Bovicola (Spinibovicola), Dasyonyxand Eurytrichodectes , where the posterior arms are extended (Fig. 12) (character 125). Thedistribution of other apomorphies indicates that the modification is convergent in Bovicola(Spinibovicola) and Dasyonyx plus Eurytrichodectes. The posterior margins of the temple are generally broadly rounded in Trichodectidae, but inspecies of the genus Eutrichophilus the convexity is much greater than in the rest of the family(Fig. 87); this development is believed to be apomorphic (character 127). In the three species ofthe genus Eurytrichodectes (only two of which are described) and the four of Procavicola(Condylocephalus) the posterior temple angles are developed into pointed projections (charac-ter 126), these being very long in the former genus (character 126: 2). This modification is notfound elsewhere in the Trichodectidae, although small rounded projections are found in someDasyonyx spp. and some Damalinia spp. The presence of pointed projections is believed, on thebasis of the distribution of other characters, to be homoplastic in the two genera mentioned. The form of the osculum has been largely excluded from consideration in the cladistic analysisbecause of the direct influence of the hair of the host (see above). However, in the Damalinia(D.) theileri-baxi clade it is quite different from other species of Trichodectidae (Fig. 69), and ishere suggested to be apomorphic (character 128). Legs (characters 129, 130) The loss of one tarsal claw on each leg (character 129) is an apomorphy associated withectoparasitism on mammals. This character is proposed as an autapomorphy of the Trichodecti-dae, although it may be autapomorphic for a postulated holophyletic group comprising theAnoplura, Rhyncophthirina and Trichodectidae (Lyal, 1985). A number of Psocodea have teeth on the 'inside' face of the tarsal claws, and manyTrichodectidae have what appears to be a single basal tooth (Fig. 15). The occurrence of teeth allalong the 'inside' face of the tarsal claws (Figs 14, 15) is restricted in the Trichodectidae toDasyonyx, .and is believed to be autapomorphic for the genus (character 130). The two TRICHODECTID MAMMAL LICE 219 subgenera of Dasyonyx have tarsal claw teeth of different forms: D. (Dasyonyx) have sharpslender teeth (Fig. 14), whilst D. (Neodasyonyx) have blunter, broader teeth (Fig. 15). Thesetwo forms may be co-apomorphies, indicating that the subgenera are sister-groups, or one maybe the plesiomorphic state. No other characters have been found within Dasyonyx that indicatereliable sister-group relationships within the genus. In this study the two tooth forms areaccepted as co-apomorphies and the subgenera retained, but further work on the genus maycause this hypothesis to be challenged. Postcoxale (character 131) In most species of Trichodectidae the metathoracic postcoxale is either not sclerotised orsclerotised and small, but the polarity of the transformation series with the extreme states'sclerotised' and 'not sclerotised' is not known. In Procaviphilus (Meganarionoides) and somespecies of Dasyonyx (Dasyonyx) the postcoxale is greatly enlarged and heavily sclerotised, anapomorphic state not included in the analysis for reasons given below. A further apomorphiccondition, the fusion of the postcoxales, is also found in some members of the same subgenera.The presence of the sclerotised postcoxale is difficult to determine in some of the smaller speciesof Dasyonyx, but in any case the distribution of other apomorphies indicates convergence of thepostcoxale characters in the two subgenera. Neither apomorphy is used in the analysis. In P. (Meganarionoides), uniquely, the postcoxale is fused to abdominal pleurum II (charac-ter 131), and this apomorphy is used in the analysis. It is interesting that the great development of the postcoxales, restricted to lice parasitic onhyraxes, is morphologically convergent on the development of the apophysis of abdominalsternum II (character 132), which is found in another group of hyrax lice. Spiracles (characters 133-138) For the Phthiraptera (and the Trichodectidae) the plesiomorphic number of spiracles is a singlethoracic pair and six abdominal pairs; further reduction in the number of abdominal spiracles isapomorphic. The numbers of abdominal spiracles in the different species of Trichodectidae aresummarised in Table 1. Each reduction is considered to be an apomorphy (characters 135-138),though some homoplasy has occurred. Inspection of the distribution of other apomorphiesindicates that reduction to five, four and one pair of spiracles has occurred once, reduction tothree and two pairs twice, and reduction to none eight times. Because of the sequential patternof spiracle loss, apomorphy 136 is always associated on the cladogram with apomorphy 135, 137with 136 and 135, and 138 with 137, 136 and 135. In most species of lice all of the abdominal spiracles have atria of roughly the same size; as hasbeen pointed out in earlier discussion, however, some Trichodectidae have posterior spiracleswith atria much smaller than those more anterior on the abdomen (character 134). Thisdifference in size is believed to be an apomorphic reduction. In most species of lice the atrium of the thoracic spiracle is as broad or broader than long;species of the genus Cebidicola, however, have a tubular atrium associated with the thoracicspiracle (character 133). This modification of form is believed to be apomorphic. Abdominal setae (characters 140-156) Study of abdominal setal patterns throughout the Psocodea suggests that the plesiomorphicpattern is a row of setae running around the abdomen on each of segments I to VIII. Ontrichodectid abdominal pleura II-VII this row of setae (referred to here as the posterior setalrow or 'p.s.r.') is generally clear, and is absent in only a few species. The distribution of otherapomorphies indicates this absence to be apomorphic, although limited to a very few, distantly-related, species. In some Trichodectidae, the p.s.r. of pleura II, III and IV comprises setae thatare much stouter than those of other pleura, and the distribution of other apomorphies suggeststhe apomorphic status of each of these, although each exhibits some homoplasy. Preliminaryanalysis indicated that specialisation of the p.s.r. on pleura II and IV conveys little phylogeneticinformation, and only character 139 (specialisation of the p.s.r. on pleurum III) is used in thefinal cladistic analysis. The setae on pleura VIII and IX are frequently longer than the setae of 220 C. H. C. LYAL the p.s.r. on anterior pleura; the extreme length of these setae in Eutrichophilus is, however,recognised as apomorphic (character 140). The setal row on sterna II and III is usually similar to the row on other sterna, but in the twospecies of Eurytrichodectes the setae of these two rows are short, stout and conical (Fig. 136).This unique feature is believed to be apomorphic (character 153). The tergal setal row of many Trichodectidae, especially males, is clearly composed of fourdiscrete groups - two lateral and two median - each separated by a gap (Fig. 22) . The positioningof the groups and the number of setae in them are useful taxonomic characters, and may beutilised as landmarks for the identification of particular setae. The groups are, however, difficultto use in phyletic analysis because of the difficulty of assigning polarity to any transformationseries. In some Trichodectidae (and in no other Phthiraptera) a seta- termed here the 'posterolateralseta' or 'p.l.s.' - is present posterolaterally on each side of terga II-VI (Fig. 22). The restrictionof distribution of this seta within the Phthiraptera suggests that its presence is apomorphic(character 151). In some cases there is more than one p.l.s. on each side of the tergum (Fig. 201);this is believed to be an apomorphy but its sporadic occurrence (in terms of clades indicated byother apomorphies) has led to its omission from the cladistic analysis. The presence of the p.l.s.is difficult to assess in some species, either because the lateral group may be reduced in numberor because the lateral group is composed of very long setae. In the former case, a single seta inthe position of the p.l.s. may be this seta (the lateral group being absent), or it may be the soleremaining seta of the lateral group (the p.l.s. being absent) (Fig. 159). In the latter case (mostTrichodectes species and the Neotrichodectinae), the most lateral seta of the lateral groupfrequently lies slightly posterior to the rest of the row (Fig. 158) and a more differentiated p.l.s.is absent. In both these cases the p.l.s. is postulated to be present, though modified. The anterior setae posterior setalrow postero-lateralseta lateralgap mediangap median setalgroup lateral setalgroup Figs 21, 22 Trichodectid morphology. 21, postulated evolution of basi-parameral sclerite by detachmentof basal flange of paramere. 22, Abdominal setal arrangement, illustrated by anterior terga and pleura ofmale. TRICHODECTID MAMMAL LICE 221 distribution of other characters suggests, however, that secondary loss of the p.l.s. has occurredwithin some taxa. Some setae of the median tergal group, particularly in males, may be specialised. In males ofNeotrichodectes spp. the two central setae of the united median groups, or two setae very nearthe centre (perhaps separated by one or two unmodified setae) are very much smaller than theother setae of the row (Fig. 229). These 'tergocentral microsetae' (character 152) are foundnowhere else, and this and the concordance of their distribution with that of apomorphy 118:l'+2' suggests the apomorphy of the character state. In Felicola (S.) pygidialis and F. (S.)macrurus the median group of tergum III is modified in a distinctive manner in the males (Fig.178), and this modification is assumed to be apomorphic for the two species (character 147).Some or all of the setae of the median group on terga II and III of male Trichodectidae may beenlarged relative to the other tergal setae. This enlargement occurs sporadically both within andoutside the family, and each case is believed to be autapomorphic. Within the Trichodectidae, inmales of Geomydoecus (Thomomydoecus), G. (G.) copei, Trichodectes (Paratrichodectes)ovalis and ugandensis, and the undescribed sister-subspecies of Trichodectes (T.) galictidis, thesetae of the median row on both terga II and III are enlarged but remain in a straight row (Figs161 , 244) ; this arrangement is believed to be apomorphic, but probably convergent in each of thefour groups (character 149). In Bovicola (Spinibovicola) hemitragi and multispinosa a similarenlargement is confined to some of the setae of tergum II, and the lines are curved (Fig. 41)(character 150). In males of Felicola, the holophyly of which is supported by several apomor-phies, the median setal group is reduced to a single, greatly enlarged seta (Fig. 188). Theapomorphic status of this character (character 142) is indicated by its restricted distribution andcorrelation with other apomorphies. It is notable that the setae are single, but of normal size(very small) in the Felicola (F.) rahmi-viverriculae clade (Fig. 183), and secondarily increased innumber to six in Felicola (S.) bedfordi and F. (F.) setosus (Figs 180, 189). On the basis of otherapomorphies, the former is believed to be a single autapomorphic reversal, whilst the latter isbelieved to be a convergent gain. The sclerite from which the pair of setae arises may be long and of characteristic shape (Fig.181); this feature is found only in conjunction with the enlarged setae (character 142) and ispostulated to be apomorphic (character 155). In many males of Felicola the median setal group on terga III-VII is also reduced to a singleseta, although in most cases this does not approach the size of the seta of tergum II. In theFelicola (S.) cooleyi-quadraticeps clade this reduction has taken place on terga III-VII, but thesetae are similar in length to those of tergum II, the latter being reduced relative to those of otherspecies of Felicola and the former enlarged (Fig. 186) (character 143). This apomorphy isconvergent on the apomorphic setal pattern of the males of Trichodectes (Stachiella) (character144), although in this case the setae are all generally long and stout. The median setal group ofthe female tergum may also be reduced to a single seta (character 145) or lost (character 146).The distribution of these female apomorphies is as follows. The reduction of the median groupto a single seta is found only in the Trichodectes (S.) fallax-octomaculatus clade; the sister-species, T. (S.) potus, and the sister-group to this clade, the T. (S.) retusus-mustelae clade, lackthe female median group entirely. The sister-species to the whole T. (S.) retusus-potus clade, T.(S.) emeryi, has the median group unreduced, numbering three setae, on terga I and II, reducedto one seta or absent on tergum III, and absent on terga IV- VIII. It is not certain whether setalloss in the female has taken place only once, the setae being regained in the fallax-octomaculatusclade, or has taken place independently three times (in emeryi, the retusus-mustelae clade, andpotus). The length of the abdominal setae is difficult to employ in phyletic analysis because of theproblem of establishing the polarity of the transformation series 'very short - medium - verylong'. The restricted distributions of the two extremes of the series (concordance with otherapomorphies) indicate their apomorphic status, however. The very short, sparse setae (Fig. 183)are found in no other Phthiraptera but the Felicola-Lorisicola clade and some Trichodectes spp. ,and are probably apomorphic but convergent in the two groups. The very long refringent setaeof some Trichodectidae (Fig. 158) are considered apomorphic for a similar reason (character 222 C. H. C. LYAL 141), though in this case similar setae are found in some Philopteridae. These long setaeprobably evolved twice in the Trichodectidae: once in Trichodectes and once in the Neo-trichodectinae. The fine, long setae of the females of the Felicola (S.) cooleyi-quadraticeps cladeare found in no other group and are considered apomorphic (character 154). The setal bases - the circular 'pits' of the setal articulations - are of fairly constant size relativeto the setae in most Trichodectidae. However, in the Bovicola (B.) alpinus-tibialis clade thebases are noticeably large in relation to the setae, and seem to have a double margin. Thisfeature needs to be examined using the scanning electron microscope to elucidate its truestructure, but examination using the light microscope is sufficient to detect its presence. Thisfeature is here considered as apomorphic (character 156). Abdominal pleural projections and modifications (characters 157-161) In many Trichodectidae the dorsoposterior and/or the ventroposterior pleural angles project onpleura II, III or IV (see discussion above, and Table 2). Projections on these segments of thetype found here do not occur elsewhere in the Phthiraptera, and are therefore considered asapomorphic (characters 158-160). Preliminary analysis reveals that the projections on pleurumIII contribute no useful phyletic information, so the apomorphy is omitted from the finalcladistic analysis. Variation in the degree of development of the dorsal and ventral lobes of theprojection on pleurum IV is omitted for the same reason, except for the extreme development inEurytrichodectes (character 161). Both the presence of a dorsal and a ventral projection onpleurum IV (characters 159 and 160 respectively) are included in the analysis, though the latterapomorphy is reversed in some clades. The presence of a projection on pleurum II (character158) provides a synapomorphy for Geomydoecus spp., which are also united as a holophyleticgroup on other grounds; the projection on this pleurum is found convergently in Trichodectes(Paratrichodectes) zorillae. The sclerotisation of the dorsal projection on pleurum II (character158:2) is an autapomorphy of Geomydoecus (Thomomydoecus) (and some species in Geomy-doecus s.str. - see discussion below -) and T. zorillae, but sclerotisation of the projections on theother pleura is very variable, and is not used in cladistic analysis. In Damalinia (Damalinia) pleurum II extends on to sternum II, and the pleurite is expandedat the expense of the sternite (character 157). This extension may be broad (character 157:1') ornarrow (character 157:1), but the more plesiomorphic state of these two (should they not beco-apomorphies) is not known. The species with a broad ventral extension of pleurite II alsopossess a more or less extensive dorsal extension, but this is not found in species with a narrowventral extension. In this treatment the two forms of the ventral extension are used tocharacterise each of two sister-groups, but this hypothesis is open to challenge, as the groupindicated by character 157:1 has no other supporting apomorphy. Abdominal sclerae (characters 132, 162-164) As noted above, in species of the trichodectid genus Procavicola sternite II is greatly developedas a heavily sclerotised internal apophysis, articulated to pleurum II (character 132). Thepresence of this unique structure is considered apomorphic. The presence of the lateral flecks and their associated small sclerite is considered apomorphic,as the structure occurs in no Phthiraptera other than the Trichodectidae (character 164). The pleura, sterna and terga of the trichodectid abdomen may be sclerotised or not; althoughsome groups (identified on the basis of other apomorphies) may, in general, be more or lesssclerotised, the polarity of the transformation series 'sclerotised - not sclerotised' for eachsegment cannot be determined, and these characters are not used in the cladistic analysis. Themale abdomen may have a characteristic sclerotisation dorsally, in that the terga may haveanterior and posterior sclerites; this feature is present in some Anoplura, but is probablyconvergent in this suborder. The presence of doubled tergal sclerites in male Trichodectidae isbelieved to be apomorphic for the family, but is not used in cladistic analysis because of the largenumber of reversals. The tergal sclerites are not further modified in most male Trichodectidae,but in some there is longitudinal division of the anterior or posterior sclerites (characters 162 and163 respectively), this division being accepted as apomorphic. TRICHODECTID MAMMAL LICE 223 D. (Trlchollpeurus) Eutrichophllus 00) Dasyonyginae i o CD) Trichodectini Felicola _ - -* -C* -* CT) CD -r^ j^wojoj t/i oj <_n ** Neotrichodectinae Fig. 23 Cladogram of Trichodectidae. Clades numbered 1-12 are resolved in Figs 24-34. The Eu-trichophilinae (Genus Eutrichophllus only) is not resolved further. For explanation of numberedapomorphies see text. Cladistic analysis The holophyly of both monotypic and polytypic species is accepted without the need forjustification, so species-level autapomorphies have not been indicated unless they are homoplas-tic with character states elsewhere on the cladogram. Omission of the autapomorphies of speciessaves both space in the data matrix and time taken for analysis, and for the same reasons manysister-species pairs are justified on the cladogram with fewer autapomorphies than are available.Of the 187 apomorphic character states used in the analysis, 86 are postulated to have beendeveloped more than once or to have been secondarily lost, 363 such homoplasies beingproposed. When, in the analysis, a choice is available between postulating one reversal or a pairof homoplastic gains (i.e. three clades in a holophyletic group are involved and the topographyof the tree is not affected whichever the choice), the latter is chosen (e.g. character 13:1 in theDamalinia theileri-appendiculata clade). This choice is made so that the distribution ofapomorphic character states can more easily be discerned on the cladogram. The number of 224 C. H. C. LYAL Blsonlcolg s. seaeclmdecembrll Fig. 24 Cladogram of Bovicolinae (part) (clades 1-4 of Fig. 23: genera Bovicola, Bisonicola, Tragulicolaand Werneckiella). For explanation of numbered apomorphies see text. TRICHODECTID MAMMAL LICE 225 D. (D.) oppendlculata Fig. 25 Cladogram of Damalinia (Damalinia) (clade 5 of Fig. 23). For explanation of numberedapomorphies see text. homoplasies could be slightly reduced without affecting the topology of the tree because, asexplained below, the less parsimonious presentation is sometimes chosen to make the clado-gram more informative and less potentially misleading. In a number of places in the discussion ofapomorphic state identification above, reference is made to single apomorphic character statesarising more than once. This is superficially contradictory, especially if the terms apomorphyand homology are equated. Such situations have been detected during one 'round' of reciprocalillumination. Should all such convergent apomorphic states be receded as separate apomorphies(which they are believed to be, even if they cannot be differentiated morphologically) thecladogram would appear more parsimonious. The loss of the median extension of the mesomeral arch (character 47) is placed on thecladogram 17 times, frequently in combination with the loss of apical fusion of the mesomeres(character 33). These apomorphies are not arranged in the most parsimonious manner on thecladogram, as can be seen by inspection of the Bovicolinae. As presented, the cladogram depictsthe loss of the extension 11 times in this subfamily. A more parsimonious arrangement of theapomorphies is achieved by postulating characters 33 and 47 as synapomorphic for Damalinia(Damalinia) and Damalinia (Cervicola), and character 47 as synapomorphic for two clades:Werneckiella plus Tragulicola and Bisonicola, and Bovicola (Bovicola) plus B. (Lepikentron)and B. (Spinibovicola) . This arrangement reduces the number of proposed homoplasies ofcharacter 47 to six within the Bovicolinae, and reduces the number of polychotomies on thecladogram. Alternatively, the loss of the extension might be postulated to have occurred onlyonce, in the common ancestor of the Bovicolinae, and regained six times (B. crassipes, D. 226 C. H. C. LYAL Fig. 26 Cladogram of Damalinia (Cervicola) (clade 6 of Fig. 23). For explanation of numberedapomorphies see text. r-CH -+9- )la extroriusC, semlarmatus Procovicola (P.) 16 speciesP. (Condylocephalus) 1 indf leldiP. (C.) hopkinsl Fig. 28 Cladogram of Dasyonyginae (clade 8 of Fig. 23, part: genera Cebidicola and Procavicold) . Clade13 is resolved in Fig. 29. For explanation of numbered apomorphies see text. TRICHODECTID MAMMAL LICE 227 co en - co ' o-9-m- mallnla (Trlcholipeurus) aepycer Fig. 27 Cladogram of Damalinia (Tricholipeurus) (clade 7 of Fig. 23). For explanation of numberedapomorphies see text. elongata, D. moschatus, D. clayi and the D. albimarginata-indica clade). The most parsimo-nious hypothesis is that the structure was lost in the ancestor of Bovicolinae as suggested above,regained twice (B. crassipes and the D. (T.) albimarginata-elongata clade), and secondarily losttwice (D. (T.) lineata-victoriae and D. (T.) pakenhami-bedfordi). This last hypothesis, althoughmore parsimonious than the distribution on the tree presented, does not change the topology ofthe tree. The distribution of character 47 is not as apparent from inspection of the tree in its mostparsimonious distribution as it is in the tree presented, as the more scattered distribution of thelosses and reversals obscures the alternative possible distributions and implies a spuriousconfidence in the tree as supported by them. The distribution of character 33 (the loss of mesomeral fusion) in Werneckiella is notpresented in the most parsimonious manner. There is great difficulty in the observation of thischaracter state in Werneckiella, and the morphological difference between 'loss of fusion' and'fusion' is very slight. A detailed examination of the species of this genus for other characters to 228 C. H. C. LYAL , 00 00 CO -^ O Fig. 29 Cladogram of Dasyonyginae (part; clade 13 of Fig. 28: genera Procaviphilus , Dasyonyx andEurytrichodectes). For explanation of numbered apomorphies see text. TRICHODECTID MAMMAL LICE 229 --m- Trlchodectes CO O) - 01 co y; Fig. 30 Cladogram of Trichodectini (clade 9 of Fig. 23, part: genera Protelicola, Lutridia, Neolutridia andWerneckodectes). Clade 14 is resolved in Fig. 31. For explanation of numbered apomorphies see text. complete a full analysis was not made, character 33 only being noted because it occurs elsewhereon the cladogram. It is possible but not likely that the distribution of character 33 as observed issupported by other apomorphies, but the proposal of holophyletic groups within the genus onthe basis of the observations made of this single character would be unwise. It is notable thatWerneckiella fulva and W. neglecta, which differ in the state of character 33, are otherwise verysimilar, the females apparently being indistinguishable (Emerson & Price, 1979), and it is verylikely that they are sister-species. The arrangement of Protelicola, Lutridia and the Trichodectes-Neolutridia clade (T-N) onthe cladogram (Figs 30, 31) does not accord with the most parsimonious distribution of theapomorphies. The cladogram contains four convergences for 'gain' apomorphies: 12:1' (fusionof parameres and b.a.l.s.) is postulated as homoplastic in Lutridia, Neolutridia and Trichodectes(Stachiella); 13:1 (fusion of parameres to each other) is postulated as homoplastic in Protelicolaand Lutridia; 20 (development of rod-shaped parameres) is postulated as homoplastic inProtelicola and Lutridia; and 58 (development of tubercles for the gonapophysis setae) ispostulated as homoplastic in Protelicola and T-N. Apomorphy 12: 1 ' has been discussed in detailabove, and the distribution suggested in the cladogram is believed consistent with the morpholo-gical evidence. Apomorphy 58 could be considered in two ways other than that presented: as anautapomorphy supporting the sister-group relationship of Protelicola and T-N, or as anautapomorphy of the Trichodectini (the Trichodectes-Protelicola clade), reversed in Lutridia.The first alternative is not supported by the distribution of any other apomorphies, whereas thetwo alternative arrangements are each indicated by more than one apomorphy (see below); thesister-group relationship of Protelicola and T-N is therefore rejected. The plesiomorphicarrangement of the gonapophysis tubercles in T-N is clearly distinct from the arrangement inProtelicola. If the tubercles are postulated to be homologous in the two clades two furtherapomorphies (the form of the tubercles in each clade) would have to be proposed, as neitherform appears to be plesiomorphic with respect to the other. This manipulation does not affectthe topology of the cladogram (whatever the position of Lutridia), and does not clarify therelationships of the clades involved, so the hypothesis of convergence of character 58 inProtelicola and T-N is retained. The other two apomorphies may now be considered together asthey both suggest the sister-group relationship of Protelicola and Lutridia. The alternativehypothesis (of the cladogram as presented) is supported by apomorphies 72 (development of asclerotisation along the ventral vulval margin) and 47 (loss of the median extension of themesomeral arch). 'Loss' characters are given much less weight than 'gain' characters in thisanalysis, so character 47 should be left out of consideration. The sister-group relationship ofProtelicola and Lutridia is therefore supported by two apomorphies and the relationshipproposed on the cladogram supported by one. As noted in the generic descriptions below,however, an undescribed species of Protelicola has been seen which does not share apomorphy20. The cladistic position of this species with respect to the other two species in the genus has not 230 C. H. C. LYAL I O-OD+B CH -si CO CO L oo .0. r\j .0. - Co r\> en - -0-0 1 OO-O-B-B-H- no -*-*CJ en uij. oo to oo A -I to en ** O> * to f CTJ u. - A CO <0 - - O CO CO CO - t CO Fig. 31 Cladogram of Trichodectes (clade 14 of Fig. 30). For explanation of numbered apomorphies seetext. TRICHODECTID MAMMAL LICE 231 Fig. 32 Cladogram of Felicola (clade 10 of Fig. 23) . For explanation of numbered apomorphies see text. 232 Ul ~%l CO 00 O3 GO L. (L.) spenceri L. (L.) sudamerlcanus L. (L.) hercvnlanus L. (L.) Siem L. (Pnradoxuroecus) bengalensls L. (P.) philipplnensls L. (P.) luce 11 L. (P.) nspldorhvnchus Fig. 33 Cladogram of Lorisicola (clade 11 of Fig. 23). For explanation of numbered apomorphies see text. TRICHODECTID MAMMAL LICE 233 a Geomydoecus (G.) unresolved species G. (G.) copelG. (G.) thonoiwus G. (G.) duchesnensls G. (G_. ) dakotensls - G. (Thoniomydoecus) wardl G. (I. ) a - z clade Neotrlchodectes (N.) thoraclcus N. (N.) mlnutus N. (N.) osbornl N. (N.) mephltldls N. (N.) wolffhuegell r o N. (Trigonodectes) bgrbarge N. (Nasulcola) pallldus N. (Lakshmlnarayanella) gastrodes N. (L. ) cunmilngsl N. (ConeeatliQlg) chilensls N. (C.) interruptofosclatus N . ( C . ) semlstrlotus N. (C.) arlzonae Fig. 34 Cladogram of Neotrichodectinae (clade 12 of Fig. 23: genera Neotrichodectes and Geomydoecus). For explanation of numbered apomorphies see text. been determined because of the poor state of preservation of the specimens, but its existenceraises the possibility that character 20 is an apomorphy not of Protelicola but of only two specieswithin the genus (the alternative being a reversal in the undescribed species). Character 20 isalso homoplastically developed in Felicola and Bovicola (Lepikentron) . If this character isdisregarded, apomorphies 72 and 13:1 must be compared for their comparative likelihood ofhomoplasy. Apomorphy 13:1 is homoplastically developed at 22 other points on the cladogramwhilst 72 is found elsewhere only in Bovicola (Lepikentron) . Apomorphy 72 should clearly begiven much more weight than 13:1 in construction of the cladogram, and 20 is considered ofuncertain value in view of the undescribed species of Protelicola. For these reasons thecladogram is retained as proposed, even though it is not maximally parsimonious. The dorsal projection of pleurum IV (character 159) is lost in the Trichodectini, but postulated as secondarily regained in Werneckodectes and Trichodectes (Paratrichodectes) zorillae. The form of the projection is different in the two species, however, which indicates the independent development of the structure. Whilst the genus Geomydoecus s.l. is almost certainly holophyletic, this probably does not 234 C. H. C. LYAL apply to either of the two included subgenera (Fig. 34). The question of holophyly should beaddressed first in the smaller subgenus Geomydoecus (Thomomydoecus). All but one of theincluded species (G. (T.) war.di) have characteristically asymmetric male genitalia (character29), and are proposed on this basis to be a holophyletic group (the asymmetricus-zacatecae cladeor 'a-z clade'). G. (T.) wardi and the a-z clade share the following apomorphies: posterolateraltemple margin with single stout seta and associated shorter, finer setae (a character not includedin the data matrix); male parameral plate apically pointed (character 14); gonapophysissmoothly continuous with ventral vulval margin (character 71); male abdominal terga II and HIwith median setal group comprising exceptionally long, stout setae (character 149); and pleuralprojections sclerotised, especially in females (character 158:2). The possession of a single stouttemple seta is unique to these species, but may be a reduction from the two stout setae found inthis position in some Geomydoecus (Geomydoecus). Apomorphies 14, 71 and 149 are alsoshared by G. (G.) copei, and this species has the mesomeral arch and parameral plate veryslender, approaching the shape of the genitalia of the a-z clade more closely than does G. (T.)wardi; the posterolateral temple margin lacks any specially-modified setae, but this may be dueto secondary loss. Apomorphies 14, 71 and 158:2 are shared by the G. (G.) thomomyus-dakotensis clade, but the male genitalia are considerably broader than those of G. (T.) wardi,and the mesomeral arch lacks a median extension (an autapomorphy of the clade). This cladehas a further autapomorphy in the form of the posterolateral setae of the temple margin, whichcomprise a single long fine seta and associated shorter fine setae. As with G. (G.) copei, theplesiomorphic form of the temple setae is unknown, and could have been the form found in G.(Thomomydoecus) . Other species of G. (Geomydoecus) have a single apex to the parameralplate (character 14), but do not share any of the other apomorphies mentioned. G. (T.) wardi,the a-z clade, and the G. (G.) thomomyus-dakotensis clade are all parasitic on Thomomys spp. ,whilst G. (G.) copei is a parasite of Orthogeomys hispidus; both host genera are parasitised byother members of Geomydoecus (Geomydoecus) . The apomorphies listed above plainly do not support unequivocally any of the three possiblesister-group relationships of the a-z clade without invoking homoplasy to an unjustifiableextent. It is apparent, however, that G. (Geomydoecus) is paraphyletic with respect to G.(Thomomydoecus) and that the latter subgenus is possibly polyphyletic. A full phylogeneticanalysis of the 102 species and subspecies of Geomydoecus, which would have been necessary toresolve the problem, was not attempted. The hosts of the genus are all geomyid rodents, thesystematic and taxonomic understanding of which is of questionable accuracy (Price, pers.comm.). For the purposes of this study the subgeneric concepts proposed by Price & Emerson(1972) are retained. Taxonomic history of Trichodectidae Burmeister (1838) divided the Mallophaga into two families, Liotheidae and Philopteridae, thelatter comprising the two genera Philopterus and Trichodectes . Kellogg (1896) proposed thesuborders Amblycera and Ischnocera for Liotheidae and Philopteridae (sensu Burmeister)respectively, and erected the family Trichodectidae for the genus Trichodectes. Mjoberg (1910) described Damalinia and Eutrichophilus , the second and third genera ofTrichodectidae, and Stobbe (1913a) described a fourth genus, Eurytrichodectes. Stobbe (1913b)revised the family for the first time. Ewing (1929) described four further genera and provided akey to all eight, although Ferris (1929) regarded E wing's new genera as of 'most dubious value'.Bedford (1929, 19320, 19326, 1936) described a further 10 genera, two of which were juniorsynonyms of genera proposed by Ewing (1929), thus bringing the total to 16; Ewing (1936)provided a key to 14 of these. Keler (1938a) recognised 24 genera, 10 of them new (although oneof these had been published previously by Keler, 1934 as a nomen nudum). The two generaomitted by Keler (1938a) were the same two previously omitted by Ewing (Cebidicola Bedford,1936 and Lorisicola Bedford, 1936); the three species included in these genera were placed byKeler with two others, also from primates, in his new genus Meganarion (although with theproviso 'without, of course, intending to establish the congeneric status of these species'). Keler TRICHODECTID MAMMAL LICE 235 (1938fl) provided a key to most of the genera described in his paper (with the sole exception ofMeganarion) and to many of the species. Werneck (1941) introduced the subgenus concept tothe taxonomy of Trichodectidae , describing three new subgenera in two of the four genera of thefamily parasitic on hyraxes. During the decade following Keler's (1938a) review of the family anumber of new genera were described, bringing the total number of available names in thegenus-group to 43 by the end of 1948. Werneck (1948, 1950) reviewed all the Trichodectidae,and recognised 20 genera (though one of these doubtfully) and three subgenera. No genera andonly one subgenus have been described since 1948. The most recent name to be added(Lakshminarayanella Eichler, 1982) was published as a replacement name for a juniorhomonym, and brings the total of available names to 45. There have been no revisionary worksof the family since those of Werneck (1948, 1950), although Hopkins & Clay (1952), whencataloguing the 'Mallophaga', accepted 13 genera, some of these doubtfully, and Eichler (1963)recognised 38 genera (with no subgenera). The problem of establishing criteria by which taxa can be distinguished at the generic levelreceived early attention. An attempt to identify morphological characters for this purpose wasinitiated by Bedford (1929). Bedford (1932) provided a more thorough discussion, andconcluded that whilst the shape of the head, the presence or absence of abdominal sclerites andthe form of the female gonapophyses were of value, the form of the male genitalia and thenumber of abdominal spiracles provided no useful guide. Ewing (1936) came to a quite differentconclusion, regarding abdominal spiracle number as 'the most important generic character'.Bedford (1939) realised the unworkability of any system involving a priori assessment ofmorphological characters for generic discrimination, although he still felt that abdominalspiracle number was not of value at the generic level, and noted that (morphological) genericcharacters 'may not be very striking'. To supplement or replace morphological charactersBedford (1939) made use of host data, the possibility of which was first discussed by Kellogg(1913, 1914) and Harrison (19166). Bedford (1939) wrote: 'Before placing a species in a newgenus one should ask oneself: would it be possible to say from what kind of host the parasite wastaken off had it not been recorded? If it is impossible to answer the question, then one should bejustified in placing it in a new genus.' Hopkins (1941) used this principle to a certain extent in hisdiscussion of Felicola. He also discussed the morphological characters used for the discrimina-tion of the genus from others, and pointed out that 'the singling out of one character [on which tobase genera] . . . only tends to obscure natural relationships'. Werneck (1936) perceived andtreated the problem of generic discrimination in a rather different way from those describedabove. He noted that whilst the type-species (and sometimes a few species similar to the type) ofeach of the described genera were quite distinct, other species showed intermediate characters.The existence of these 'transitional forms' convinced Werneck that there was no validity in theseparate genera, so he synonymised them all (with the exception of some genera not found inSouth America, which were outside the scope of the paper). Bedford (1939) regarded this actionas 'unwarrantable' and reinstated all the genera. Hopkins (1942, 1943) reviewed the charactersused to separate the genera of Trichodectidae parasitic on carnivores and antelopes respect-ively. In each case he found annectent species as described by Werneck (1936), and took similaraction, though modified by the belief that louse genera should somehow reflect host taxa.Hopkins (1942) therefore accepted three genera of Trichodectidae parasitic on carnivores andlater (1943) accepted one genus parasitic on antelopes. Werneck (1948, 1950), although lessinfluenced by the host data, recognised more genera than had Hopkins. He accepted themorphologically 'distinct' species and species groups as genera, and placed annectent species inthe genus which they most closely resembled. Hopkins (1949) 'conceded subgeneric status tomany groups which seem likely to be accepted by systematists whose views . . . differ from mine'and recognised 14 genera and 20 subgenera. Hopkins & Clay (1952) synonymised some of thesubgenera accepted by Hopkins (1949), but raised others to generic status. Ledger (1980) heldviews similar to Hopkins & Clay (1952), although in some cases followed the views of Hopkins(1949); the resultant generic arrangement still involved fewer genera than accepted by Werneck(1948, 1950), and many more subgenera. Emerson & Price (1981), however, could 'find no basisfor rejecting the classification of Trichodectidae given by Werneck (1948, 1950)', and suggested 236 C. H. C. LYAL that the 'question of genera vs subgenera will perhaps continue until Mallophaga have beendescribed from all likely hosts'. A number of other workers also follow these views. A thirdgroup of taxonomists (e.g. Eichler, 1963; Zlotorzycka, 1972) have accepted not only all thegenera recognised by Werneck (1948, 1950), but also a number of genera that Werneckconsidered as junior synonyms; the subgenus category is not used, however. The present genericplacement of most of the Trichodectidae is thus a matter of some contention and a review of thevariations in status of some genera and subgenera is presented in Tables 3-5. Table 3 Generic concepts in the Bovicolinae. The genera included in the table are represented bynumbers 1-11; '2' indicates that the genus (Bovicola) is given full generic status, '1(2)' indicates that thegenus (Bovicola) is considered a subgenus (of Damalinia) and ' = 1' indicates that the genus is consideredas a junior synonym (of Damalinia). The generic name Bovidoecus Bedford, 1929 is omitted, as it wassynonymised with Bovicola by Bedford (1932a) and has not since been used. There is no reference in thetable to Werneck (1936), who treated all genera of Trichodectidae as synonyms of Trichodectes . Table 4 Generic concepts in the Trichodectini (plus Neotrichodectes and Trigonodectes) . Coding as forTable 3. Protelicola is included in Table 5. The generic name Grisonia Keler, 1938a is omitted and itsreplacement name, Galictobius Keler, 1938ft, used throughout. TRICHODECTID MAMMAL LICE 237 Table 5 Generic concepts in the Felicolini (plus Protelicold). Coding as for Table 3. The generic namesBedfordia and Felicinia (a junior homonym and an absolute synonym respectively) have not beenincluded as their status has not varied; they are discussed in the comments following the description ofFelicola s. str. below. Hopkins (1941) presented an explanation for the presence of so many annectent species in theTrichodectidae and a justification for synonymising many of the genera, writing: 'I believe theexplanation to be that the Trichodectidae are in the process of dividing up into genera; in somecases the divergence has proceeded far enough for us to recognise the segregates as genericallydistinct, but in a much greater number of cases extreme members of a group may have becomestrikingly distinct whilst the others remain as connecting-links which entirely undo our attemptsto find characters peculiar to the group.' Hopkins' statement implies that the species in a genusare somehow evolving as a unit, and is linked to the typological approach to taxonomy.Relationships between three or more taxa, if assessed by a simple count of character states (i.e.not distinguishing between plesiomorphies, apomorphies and homoplasies) are frequentlyreticulate in aspect (Simpson, 1961; Hennig, 1966; Mayr, 1969). The greater the number ofhomoplasies the more complex the reticulum is likely to be, and the more difficult it is tocombine the taxa into groups. If genera are constructed on this principle some morphologicallydistinct species and species-groups will be distinguished, 'linked' by annectent species, with theconcomitant absence of 'gaps' between genera - precisely the problem with traditional group-ings of the Trichodectidae. The 'problem' of annectent species is therefore engendered by thetypological approach; the difficulties in distinguishing supra-specific groups are also a result ofthis, but combined in the Trichodectidae with a high degree of homoplasy. The difficulties discussed above have discouraged authors from attempting to produce keys togenera of the Trichodectidae. Since the key to genera published by Keler (1938a), very few havebeen published, and none included all the genera. Keler (1944) produced a key to some genera,slightly emended from Keler (1938), but a promised second half to the paper containing the restof the key was never published. Werneck (1948, 1950), despite describing all of the genera, didnot attempt to produce a key. A few keys have since been published in faunistic works, forexample Toulechkoff (1955) produced a key (in Bulgarian) to the genera found in Bulgaria, andZlotorzycka (1972) published a rather inaccurate key (in Polish) to the genera found in Poland. Although most authors follow Kellogg (1896) in their conception of the Trichodectidae, andretain familial rank for the group (e.g. Hopkins & Clay, 1952; Hopkins, 1960; Ledger, 1980;Emerson & Price, 1983), the rank of the group has been raised by others. Keler (19380) raisedthe Trichodectidae to superfamily level and included three families: Trichodectidae, Bovicoli-dae and Dasyonygidae, the latter two being described as new. In the Trichodectidae he placedfour subfamilies, all of which he indicated to be new: Trichodectinae, Felicolinae, Eury-trichodectinae and, dubiously placed in this family, Eutrichophilinae (Fig. 35). Eichler (1940) 238 C. H. C. LYAL MSIOHIX EURYTRICHODECTES EUTRICHOPHILVS PROCAVICOLA PROCAVIPHILUS BEDFORDIA FELICOLA PROTELICOLA SURICATOECUS GEOMYDOECUS GRISONIA LUTRIDIA NEOTRICHODECTES STACHIELLA TRICHODECTES URSODECTES BOVICOLA CERVICOLA DAWLINIA HOLAKARTIKOS LEPIKEKTRON RHABDOPEDILON TRICHOLIPEURUS DASYONYGIDAE EURYTRICHODF.CTINAF EUTRICHOPHILINAE FELICOLINAE TRICHODECTIDAE TRICHODECTINAE BOVICOLIDAE Incertae sedis MEGANARION __ Fig. 35 Classification of Trichodectoidea' according to Keler (1938a). described two further subfamilies in the Trichodectidae (sensu Keler, 19380): Lymeoninae andCebidicolinae. Eichler (1941) described the new subfamily Damaliniinae in the Bovicolidae,attributing the nominate subfamily to himself. He also transferred the Eurytrichodectinae andthe Eutrichophilinae to the Dasyonygidae, again attributing the newly-defined nominatesubfamily to himself. Eichler (1941) considered the rank of the whole group to be notsuperfamily but 'family group' and termed it the Trichodectiformia (attributed to Keler, 19380);the subfamily Trichodectinae is also attributed to Keler (19380) but the nominate family isattributed to Burmeister (1838). The classification proposed by Eichler (1941) is depicted in Fig.36. Keler (1944) retained the Eurytrichodectinae in Trichodectidae and moved the Eutrichophi-linae to Bovicolidae; the subfamilies described by Eichler (1940, 1941) were not mentioned (Fig.37). The Trichodectidae were attributed by Keler (1944) to Kellogg (1896) but the nominatesubfamily to Keler (19380). The superfamily rank was retained and the 'family group' notmentioned. Hopkins (1949) regarded the families proposed by Keler (19380) as subfamilies andthe subfamilies as at most tribes. Hopkins also, following his synonymy of Bovicola withDamalinia, considered that 'Bovicolinae must be known as Damaliniinae'. Eichler (1963)retained the higher ranks and included 'interfamilia Trichodectiformia' within the superfamilyTrichodectoidea. He moved the Lymeoninae to Dasyonygidae (wherein he retained Euryt-richodectinae), but accepted the move of Eutrichophilinae to Bovicolidae proposed by Keler(1944). Eichler (1963) also proposed the division of the Trichodect(oidea) into tribes, indicatingtheir existence and composition by variations in typography in the list of genera presented(Eichler, 1963: 159, lines 31-37). Eichler (1963) did not publish any of the tribal names, butLakshminarayana (1976) listed all of them. Neither Eichler (1963) nor Lakshminarayana(1976), however, gave any statement that purported 'to give characters differentiating thetax(a); or ... a definite bibliographic reference to such a statement' as is required by theInternational Code of Zoological Nomenclature for any name published after 1930 (Article 13). TRICHODECTID MAMMAL LICE 239 DASYOtlYX EURYTRICHODECTES EUTRICHOPHILUS PROC AVI COLA PROCAVIPHILUS BEDFORDIA FELICOLA PROTELICOLA SURICATOECUS GALICTOBIUS GEOMYDOECUS LUTRIDIA NEOTRICHODECTES STACHIELLA TRICHODECTES URSODECTES LYMEOH CEBIDICOLA LORISICOLA MEGANARIOHOIDES CERVICOLA DAMALINIA TRICHOLIPEURVS BOVICOLA HOLAKARTIKOS LEPlKENTROtI RHABDOPEDILOtl UERNECKIELLA TRICHOPHI LOPTEPUS DASYONYGINAEEURYTRICHODECTINAE EUTRICHOPHILIHAE FELICOLINAE TRICHODECTINAE LYMEONINAE CEBIDICOLINAE DAMALINIINAE BOVICOUNAE DASYONYGIDAE TRICHODECTIDAF, BOVTCOLTDAE TRIC1IOP11ILOPTERIDAE Fig. 36 Classification of Trichodectiformia' according to Eichler (1941). None of the (11) names, therefore, are available for taxonomic use. Eichler (1963) attributedTrichodectoidea, Trichodectiformia, Trichodectidae and Trichodectinae to Burmeister (1838)but, whilst (correctly) attributing Bovicolidae and Dasyonygidae to Keler (19380) he attributedthe nominate subfamilies of both to Eichler (1941). The classification proposed by Eichler(1963) is depicted in Fig. 38. Keler (1969) proposed a classification similar to that proposed byKeler (1944), but omitting a number of genera (Fig. 39). Article 36 of the International Code of Zoological Nomenclature (1984) states that allcategories in the family-group (tribe, subfamily, family, superfamily and any supplementarycategories, according to Article 35a) are co-ordinate, and a name established for any categorywithin the group is available with its original date and author for a taxon with the same typegenus in each of the categories. The Trichodectinae, Trichodectidae, Trichodectiformia andTrichodectoidea should therefore all have the same date and author. The first use of afamily-group name based on the type genus Trichodectes was by Kellogg (1896), who ranked'the Nitzschian families as suborders, the Nitzschian genera as families, and the Nitzschiansubgenera, the genera of present-day writers, as genera.' Kellogg (1908) attributed theTrichodectinae (the only subfamily included in the Trichodectidae, which bore no attribution)to 'Burmeister (?)'. As explained above, subsequent attribution has frequently been toBurmeister (1838) and, in the case of some co-ordinate names, to Keler (19380). Burmeister(1838) did not mention any taxon in the family-group with the type genus Trichodectes. All 240 EURWRICHODECTESFASTIGATOSCULUM 1 SURICATOECUS MEGANARIONOIDES GALICTOBIUS GEOMXDOECUS LUTRIDIA NEOTRICHODECTES STACHIELLA TRICHODECTES TRIGONODECTES URSODECTES EUTRICHOPHILUS PROCAVICOLA 1 ,2 1,2 PROCAVIPHILUS BOVICOLA 1 CERVICOLA 1 DAMALINIA 1 HOLAKARTIKOS 1 LEPIKENTRON 1 RHABDOPEDILON 1 TRICHOLIPEURUS 1 TRICHOPHILOPTERUS C. H. C. LYAL EURYTRICHODECTINAE FELICOLINAE DASYONY TRICHODECTINAE EUTRICHOPIIILINAE BOVICOLINAE TRICIIODECTIDAF. BOVICOLinAE TRICHOPMILOPTERIDAE Fig. 37 Classification of Trichodectoidea' according to Keler (1944). Position of genus inferred fromKeler (1938a). Position of genus inferred from key in Keler (1944). family-group names with the type genus Trichodectes should therefore be attributed to Kellogg(1896). Trichodectidae Kellogg, 1896 has been placed on the Official List of Family-groupNames in Zoology (Opinion 627, Bull. zool. Nom. 19: 91-96, (1962)). The names Bovicolinae,Bovicolidae, Dasyonyginae and Dasyonygidae should all be attributed to Keler (1938a), notEichler (1941). The action of Hopkins (1949) is synonymising the senior family-group (Bovicoli-nae) with the junior (Damaliniinae) was taken because he believed Bovicola and Damalinia (thetype-genera) to be synonyms, and Damalinia is senior to Bovicola. This action is incorrect underArticle 40 of the Code, however, which states that, at least after 1961, in the case of type-genussynonymy the senior family-group name is to be used for the family-group taxon that containsboth senior and junior synonyms. This Rule can be set aside for such an action if taken before1961 , if the name has 'won general acceptance' (Article 40b). The subfamily Damaliniinae sensuHopkins (1949) has rarely if ever been used since, whilst the name Bovicolidae (= Damaliniinaesensu Hopkins) has been employed by Eichler (1963) and Keler (1969). The action of Hopkins(1949) is therefore rejected. Keler (1944) included the Trichophilopteridae - a family containing a single genus, parasiticon Lemurs - within the Trichodectoidea, although Keler (1969) referred this family to thePhilopteroidea. Eichler (1963) retained the Trichophilopteridae in the Trichodectoidea, butdistinguished it as 'interfamily Trichophilopteriformia' as opposed to 'interfamily Trichodecti-formia'. Stobbe (19130), Ferris (1933) and Werneck (1948) all considered the affinities of TRICHODECTID MAMMAL LICE "Tribes" 241 DASYONYX NEODASYOKYX PROCAVIPHILUS EVRYTRICHODECTES PROC AVI COLA CONDYLOCEPHALUS MEGAHARIOHOIDES LYHEON FELICOLA SURICATOECVS EICHLERELLA FASTIGATOSCULUM FELICOMORPHA PROTELICOLA NEOFELICOLA PARAFELICOLA PARADOXUROECUS TRICHODECTES OALICTOBIUS POTVSDIA TRIGONODECTES URSODECTES WERtiECKODECTES STACHIELLA NEOTRICHODECTES LUTRIDIA GEOMYDOECVS CEBIDICOLA LORISICOLA BOVICOLA HOLAKARTIKOS LEPIKENTRON RHABDOPEDILOH WERKECKIELLA DAMALINIA CERVICOLA TRICHOLIPEURUS EUTRICHOPHILUS ZlZl =1 Zl DASYONYGINAE F.URYTRICHODECTINAE ID LYMEONINAE FELICOUNAE Zl ZJ Zl Zl TRICHODECTTNAE Zl n CEBIDICOLTNAF BOVICOLINAE DAMALINIINAE EUTRICHOPHILTNAE DASYONYGIDAE TRICHODECTIDAE Fig. 38 Classification of Trichodectiformia' according to Eichler (1963). The 'tribes' are indicated bysquare brackets in the appropriate column, the first genus in each 'tribe' being intended by Eichler (1963)as the type-genus. See text. Trichophilopterus to lie with the Thilopteridae' rather than with the Trichodectidae. In thisstudy no apomorphies were found to indicate a sister-group relationship between Trichophilop-terus and all or part of the Trichodectidae. Proposed classification The proposed classification is derived from the results of a cladistic analysis of the Trichodecti-dae (Trichodectiformia sensu Eichler, 1963) at the species level (Figs 23-34). The species aregrouped on the four criteria discussed below, and ranked according to the principles of phyleticsequencing. Holophyly. The classification includes, as far as possible, only holophyletic groups. Some generaand subgenera, however, may be found not to follow this criterion (see discussion of Damalinias. str., Dasyonyx and Geomydoecus below). Utility. Genera are ideally of 'moderate' size and relative morphological uniformity. If a genus islarge and diverse, recognition is difficult and useful discussion on many aspects of biology ordistribution prohibited; if genera are too small, identification is time-consuming and discussion 242 C. H. C. LYAL DASYONYX NEODASmim PROCAVIPHILVS EURtTRICHODECTES PROCAVICOLA PKOTELICOLA FELICOLA SVRICATOECUS NEOFELICOLA CEBIDICOLA LORISICOLA GEOMYDOECUS STACHIELLA TRICHODECTES TRIGONODECTES URSODECTES EVTRICHOPHILUS BOVICOLA CERVICOLA DAHALINIA LEPIKENTRON TRICHOLIPEURUS DASYONYGINAE EURYTRICHODECTIHAE FELICOLINAE TRICHODECTINAE EUTRICHOPHILINAE BOVICOLINAE DASYONYGIDAE TRICHODECTIDAF. BOVICOLIDAE 1-/ERNECK1ELLA Fig. 39 Classification of Trichodectoidea' according to Keler (1969). again impeded. No 'absolute' size can be recommended, however, as the most satisfactory sizewill depend on a number of properties of the species, and must (in this study) conform to suchlimitations as are imposed by the criterion of holophyly. Subfamilies are chosen in this study toaid discussion by providing names for holophyletic groups of genera, and to fulfil the logic of thephyletic sequencing convention. Stability. To ensure that the classification has maximal stability the generic concepts accepted inthis study conflict as little as possible with established usage. Distinctness. To facilitate identification, taxa in the genus-group should be as distinct from oneanother as possible. The requirements of Mayr (1969) that genera must be separated by adecided gap, and that the 'size' of the gap should be inversely proportional to the size of thetaxon, are not necessarily compatible with the criterion of holophyly followed here, however,and the problem of annectent species (that 'fill' any such gap) has been discussed above. Despitethe apparent drawback of adherence to holophyletic groups at the expense of inter-generic'gaps', it has been possible in this study to produce a key to the genera of Trichodectidae (seediscussion of keys to genera of Trichodectidae above). The formation of the genera of Trichodectidae is discussed below, to give an indication of therationale behind each decision. The genera are discussed by subfamily, and the division intosubfamilies is discussed last. Bovicolinae (Figs 24-27, 40) This clade was not resolved fully in the analysis, and a primary pentachotomy was obtained. Theclade has been treated as a single genus (see Table 4), but the diversity of morphology and ofhosts indicates that this concept is too broad to be of great value. Subdivision of the clade intosmaller holophyletic groups increases the value of the classification for information retrieval,and leads to the acceptance of genera that approach the concepts of Werneck (1950). To obtain a TRICHODECTID MAMMAL LICE 243 measure of conformity each of the five branches of the clade has been accorded generic status.The monobasic genera Bisonicola and Tragulicola require no comment, and the genusWerneckiella was revised by Moreby (1978). Of the two remaining clades, one correspondsapproximately to a restricted concept of Bovicola, the other to Damalinia plus Tricholipeurus(sensu Werneck, 1950). The Bovicola clade (genus Bovicola) has a primary tetrachotomy, withmost of the species belonging to only one of the four resultant clades (Fig. 24). The species in thislarge clade are morphologically more similar to one another than they are to any of the species inthe other three clades. To recognise this morphological divergence (and thus facilitate identifica-tion), and to demonstrate in the classification the extent of the phylogenetic knowledge, the fourbranches are each accorded subgeneric status. The Damalinia plus Tricholipeurus clade (genusDamalinia} has a primary trichotomy (Fig. 23), and it is clear that discussion of the genus will befacilitated by the recognition of each of these branches as a subgenus. Damalinia s. str.comprises two major clades, each characterised by the form of an apomorphic development ofabdominal pleurum II onto the sternum. As discussed above, these two forms may beco-apomorphies, or may represent two states in a transformation series. If the latter interpreta-tion is correct, one of the clades is probably paraphyletic with respect to the other. Eutrichophilinae Only the single genus Eutrichophilus is included, with no change in generic concept. Dasyonyginae (Figs 28, 29) The previously-accepted generic concepts in this subfamily remain essentially unchanged at thesubgenus level. The only change is the transfer of the subgenus Meganarionoides fromProcavicola to Procaviphilus , and the inclusion of Procaviphilus sclerotis and P. serraticus in P.(Meganarionoides). Subgenera are used (as in Werneck, 1941, 1950; Ledger, 1980) as noadvantage accrues from regarding each of the clades so recognised as a full genus, andapplication of the principles of phyletic sequencing allows retention of all the currently-usedgeneric and subgeneric names with no higher taxa required, whereas recognition of all these asgenera would require the description of a number of intercalating family-group taxa. It is notable that one of the two subgenera of Dasyonyx may be paraphyletic with respect tothe other, as the subgenera are characterised by apomorphic developments of the teeth of thetarsal claws. These may be co-apomorphies or two states in a transformation series (see above).If the latter interpretation is correct, one of the subgenera is probably paraphyletic with respectto the other. Trichodectinae (Figs 23, 30-33, 40) The first dichotomy in this clade splits it roughly into Felicola (sensu Ledger, 1980, but withoutProtelicola and with Lorisicola) on one side and Trichodectes (sensu Ledger, 1980, but withoutNeotrichodectes and Trigonodectes , and with Protelicola) on the other. The diversity ofmorphology of the lice, and the variety of hosts infested, indicates that the very broad genericconcepts endorsed by Ledger (1980) are too inclusive to be of great value in data-retrieval anddiscussion. For this reason the genera proposed here are smaller than those of Ledger (1980)and, in some cases, approach the concepts held by Werneck (1948). Most of the species in the Trichodectes side of the initial dichotomy arise from the threebranches of an apical trichotomy (Figs 30, 31). The branch of this trichotomy comprising thepinguis-galictidis clade corresponds roughly to the concept of Trichodectes held by Werneck(1948), whilst the other two branches (the ovalis-zorillae clade and the emeryi-potus clade)correspond roughly to Stachiella sensu Werneck (1948) (though fallax, octomaculatus andpotuswere placed in Trichodectes by Werneck, 1948). However, placing two of the three clades of thetrichotomy in a taxon Stachiella and excluding the third results in a group that is notholophyletic. Recognising each of the three branches of the trichotomy as a separate genus isundesirable, as the three intergrade phenetically. The course followed here is to recognise the 244 C. H. C. LYAL genus Trichodectes comprising all three branches, each of these being considered a subgenus(Fig. 40). Using the principle of phyletic sequencing the sister-group of Trichodectes is alsoconsidered a genus, for which the name Werneckodectes is available. Likewise the next threebranches of this clade are also considered genera. This process necessitates dividing Werneck'sgenus Lutridia into two genera, but retention of the genus as it stood calls for recognition of aparaphyletic group in the classification, and, although the species in the two clades comprisingLutridia (sensu Werneck, 1948) are superficially similar, some of these similarities may behomoplastic. The other branch of the initial dichotomy of the Trichodectinae clade comprises, as notedabove, most of the species consigned to Felicola by Ledger (1980) plus the single species of thegenus Lorisicola (sensu Werneck, 1950). The two branches of this clade (Fig. 23) are eachconsidered as genera which, taking the most senior available names, are known as Felicola andLorisicola. For reasons of utility, each genus is divided into two holophyletic subgenera. None ofthe genera or subgenera coincides with any previous generic concept, as such concepts reliedheavily on head shape and abdominal spiracle number, both of which characters have proved tobe subject to a considerable degree of homoplasy. In order to maintain the logic of phyletic sequencing, if the Felicola-Lorisicola clade is to beconsidered as comprising two genera, the rank of this clade and of the Trichodectes-Protelicolaclade must be equal and formally recognised. Use of the tribal category permits this, and thefamily-group names Trichodectini and Felicolini are available (see full classification below). Itmust be stressed that these tribes are inserted to maintain the formal structure of theclassification, and are not intended (or believed) to have any other significance. Neotrichodectinae (Fig. 34) The first dichotomy in this subfamily divides the clade into those species previously assigned tothe genus Geomydoecus on one side, and species from Neotrichodectes, Lakshminarayanellaand Trichodectes (sensu Werneck, 1948) on the other. The two branches will be discussedseparately. The genus Geomydoecus as previously recognised is fairly uniform in morphology, distribu-tion and host species, and may be identified readily. To divide this genus into others wouldinhibit rather than encourage discussion, and the genus is retained in its present form. The twosubgenera as proposed by Price & Emerson (1972) are also retained though, as indicated above,neither are holophyletic groups. The other branch of the primary dichotomy comprises the 10 species previously assigned tothe genus Neotrichodectes (considered a subgenus of Trichodectes by Hopkins, 1949 and Ledger,1980), the two species previously assigned to the genus Lakshminarayanella (formerly Ly-meon), and a single species formerly placed in Trichodectes by most authors (T. barbarae}. Theclade is plainly close to the established concept of Neotrichodectes, and it is preferable that thisname is applied to as much of the group as possible. The 10 species of Neotrichodectes auctt. donot form a holophyletic group, however, though morphologically they are quite uniform.Inclusion of T. barbarae is unlikely to create problems, but Lakshminarayanella (as Lymeon)has been placed by some authors in a subfamily of its own (Eichler, 1940, 1963), and consideredclose to the hyrax lice (Keler, 1944; Hopkins, 1949; Eichler, 1963). If Lakshminarayanella issynonymised with Neotrichodectes and given no formal recognition it is likely to be raised fromsynonomy by future workers because of its distinctive morphology, leaving Neotrichodectesparaphyletic. The course taken here is to recognise Lakshminarayanella as a subgenus ofNeotrichodectes, which necessitates recognition of four other (holophyletic) subgenera, namesalready being available for two of these. Application of the principles of phyletic sequencingpermits equal ranking of the subgenera within the genus. Subfamilies (Fig. 40) To divide the family into 'manageable' holophyletic groups for the purposes of discussion and tomaintain the logic of phyletic sequencing, supra-generic groupings had to be employed. Use of .o " TRICHODECTID MAMMAL LICE Bovlcola (Bovlcola) B. (Holakartlkos) B. (Leplkentron) B. (spinlpovlcola) Blsonlcola Tragullcola Hernecklella iianiallnlo (DnmallDJa)- D. (Cervlcola) D. (Trlchollpeurus) Eutrlchophllus Procavlcola (Procavlcola) P. (Condylocephalus) Procavlphllus (Procavlphllus) P. (MegQngrj.QQQides) Dasyonyx (Dasyonyx) D. (Neodasyonyx) F-un'trlchodectes - Lutrldla Neolutrldla Werneckodectes Trlchodectes (Trlchodectes) I. (Paratrichodectes) I. (Stachlella) Felicolc (Felicola) F. (Surlcatoecus^ Lorlsicold (LorisjCQlg) - L. (Paradoxuroecus) Neotrlchodectes (Neotrlchodectes) N. (Trigonodectes) ft. (Nasulcola) N. (Lakshmlnarcyonella) N. (Conepotlcola) Geomydoecus (Geomydoecus) G. (Thomomydoecus) 245 > BOVICOLINAE EUTRICHOPHILINAE > DASYONYGINAE *\ > TRICHODECTINAF. > NEOTRICHODECTINAE Fig. 40 Cladogram of the genera and subgenera of Trichodectidae, with subfamily assignments. the principles of phyletic sequencing permitted the use of the subfamily category throughout(with the addition of the tribes mentioned above). The limits of the subfamilies were chosen formaximum utility, modified by the dictates 01 the sequencing convention. It would be surprising,given the high degree of homoplasy of structures in the Trichodectidae, 4f the subfamiliesfulfilled the criterion 'distinctness' described above and were readily distinguishable. A key to 246 C. H. C. LYAL subfamilies is provided, however, largely to satisfy the requirements of the International Code ofZoological Nomenclature (Article 13) for a description to accompany any new name for,although names were available for most of the subfamilies, a single new name is required. A complete classification of the Trichodectidae to generic level is set out below in phyleticsequence (as recommended by Wiley, 1979, 1981). Sequenced classification of the Trichodectidae Family TRICHODECTIDAE Kellogg, 1896Subfamily BOVICOLINAE Keler, 1938 (all genera sedis mutabilis)Genus BOVICOLA Ewing, 1929 (all subgenera sedis mutabilis) Subgenus BOVICOLA Ewing, 1929 Subgenus HOLAKARTIKOS Keler, 1938 Subgenus LEPIKENTRONKeler, 1938 Subgenus SPINIBOVICOLA subgen. n.Genus BISONICOLA gen. n.Genus WERNECKIELLA Eichler, 1940Genus TRAGULICOLA gen. n.Genus DAMALINIA Mjoberg, 1910 (all subgenera sedis mutabilis) Subgenus DAMALINIA Mjoberg, 1910 Subgenus CERVICOLA Keler, 1938 Subgenus TRICHOLIPEURUS Bedford, 1929Subfamily EUTRICHOPHILINAE Keler, 1938 Genus EUTRICHOPHILUS Mjoberg, 1910Subfamily DASYONYGINAE Keler, 1938Genus CEBIDICOLA Bedford, 1936Genus PROCAVICOLA Bedford, 1932 Subgenus PROCAVICOLA Bedford, 1932 Subgenus CONDYLOCEPHALUSWemeck, 1941Genus PROVCAVIPHILUS Bedford, 1932 Subgenus PROCAVIPHILUS Bedford, 1932 Subgenus MEGANARIONOIDES Eichler, 1940Genus DASYONYX Bedford, 1932 Subgenus DASYONYX Bedford, 1932 Subgenus NEODASYONYXWerneck, 1941Genus EURYTRICHODECTESStobbe, 1913Subfamily TRICHODECTINAE Kellogg, 1896Tribe TRICHODECTINI Kellogg, 1896Genus PROTELICOLA Bedford, 1932Genus LUTRIDIA Keler, 1938Genus NEOLUTRIDIA gen. n.Genus WERNECKODECTESCond, 1946Genus TRICHODECTESNitzsch, 1818 (all subgenera sedis mutabilis) Subgenus TRICHODECTESNitzsch, 1818 Subgenus PA RATRICHODECTES subgen. n. Subgenus STACHIELLA Keler, 1938Tribe FELICOLINI Keler, 1938Genus FELICOLA Ewing, 1929 Subgenus FELICOLA Ewing, 1929 Subgenus SURICATOECUS Bedford, 1932Genus LORISICOLA Bedford, 1936 Subgenus LORISICOLA Bedford, 1936 Subgenus PARADOXUROECUSConci, 1942Subfamily NEOTRICHODECTINAE subfam. n.Genus NEOTRICHODECTES Ewing, 1929 Subgenus NEOTRICHODECTES Ewing, 1929 Subgenus TRIGONODECTESKeler, 1944 Subgenus NASUICOLA subgen. n. Subgenus LAKSHMINARAYANELLA Eichler, 1982 TRICHODECTID MAMMAL LICE 247 Subgenus CONEPATICOLA subgen. n.Genus GEOMYDOECUSEwing, 1929Subgenus GEOMYDOECUSEv/ing, 1929 (paraphyletic)Subgenus THOMOMYDOECUS Price & Emerson, 1972 (polyphyletic?) Descriptions of genera and subgenera The generic and subgeneric descriptions below are arranged by subfamily in the order of thesequenced classification of the Trichodectidae (see above). Descriptions are set out in the following order: paragraph one - head, both sexes, with detailsof sexually-dimorphic features of antennae, if present; paragraph two - thorax, both sexes,omitting mention of the anterior setae (on the post-temporal margin) which are present in allspecies; paragraph three - abdomen, both sexes, with details of sexually-dimorphic features ofthe setae, sclerites or shape, if present; fourth paragraph - female terminalia and genitalia; fifthparagraph - male subgenital plate, terminalia and genitalia. Descriptions are given of eachgenus as a whole, even where subgenera are present. The descriptions of subgenera (if any arepresent) follow that of the genus in which they are placed, and give only subgeneric characters,so that some of the paragraphs listed above may be omitted. Characters that vary betweensubgenera, if mentioned in the generic description, are indicated by an asterisk (**). Each description is followed by an indication of the host group or groups parasitised, and byany pertinent comments on the taxonomy, morphology or biology of some or all of the includedspecies. A check-list of all species included in each genus or subgenus is also given, the namesbeing placed in alphabetical order. Following each species name in the check-lists is anindication of the number of specimens of each sex examined in the study. Two species have not been placed, and are considered incertae sedis. Trichodectes baculusSchommer, 1913; type-host: Caprahircus Linnaeus. Trichodectes tigris Ponton, 1870; type-host:Felis tigris Linnaeus. These species are discussed by Werneck (1950). The subfamilies, genera and subgenera are keyed (p. 335). BOVICOLINAE Keler Genus BOVICOLA EwingThe genus Bovicola comprises four subgenera. DESCRIPTION. Anterior of head with osculum absent or, if present, broad and shallow* ; pulvinus of normallength or short and not attaining anterior margin of head; dorsal preantennal sulcus present or absent*;clypeal marginal carina not or only slightly broadened medially, or broadened to variable degree into barwith posterior and anterior margins roughly parallel, bar either straight and at right angles to long axis ofhead or curved and parallel to anterior margin of head* ; anterolateral margin of head smoothly rounded;preantennal portion of head short, outline broadly rounded or trapezoid*. Temple margin smoothlyconvex or with posterior projection*, sometimes convexly produced posteriad*. Male scape expanded ornot expanded* , with setal row apparently present or setae randomly scattered; flagellomeres fused in malesand females; male flagellum with two basally-articulated 'teeth' and interior face not 'roughened'. Dorsumof head with more or less abundant setae , short , long or of moderate length * . Sitophore sclerite unmodifiedor with posterior arms extended*. Thorax with more or less abundant setae, short, long, or of moderate length, frequently longest onpostero-lateral margin of pterothorax*. Abdomen oval or elongate, frequently tapering posteriorly more in male than in female*. Abdominalspiracles present on segments HI- VIII. Abdominal setae variable*; anterior setae always present onpleura, sometimes on sterna and terga; postero-lateral setae absent. Abdominal pleural projectionsabsent. Sclerites present at least on sterna III-VII (males) and III- VIII (females), terga II-VII (males) andIII-IX (females) and pleura II- VIII; male terga with posterior sclerites present or absent*. Gonapophyses with marginal setae; ventral lobe present, though sometimes not pronounced*. Gona-pophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval margin notsclerotised, or sclerotised only medially; subgenital lobe absent, though small median membranousprojection may be present (Fig. 42)*. Genital chamber sometimes with median antero-dorsal area lackingscales or spicules*. 248 C. H. C. LYAL 41 45 43 Figs 41-45 Bovicola species. 41, fi. (Spinibovicola) hemitragi, cf abdominal terga I and II. 42, fi. (5.jellisoni, $ terminalia, ventral. 43, B. (Lepikentron) breviceps, 9 terminalia, ventral. 44, B. (Holakartikos) crassipes, $ gonapophysis, ventral. 45, B. (B.) caprae, 9 gonapophysis, ventral. TRICHODECTID MAMMAL LICE 249 250 57 Figs 51-57 Bovicola species, cf genitalia. 51, B. (B.) bovis. 52, B. (B.) caprae. 53, B. (B.) concavifrons.54, B. (B.) concavifrons, detail of right paramere (p) and mesomere (m). 55, B. (Lepikentron) breviceps.56, B. (Spinibovicola) hemitragi. 57, B. (Holakartikos) crassipes. Male subgenital plate variable*. Pseudostyli present or absent*. Male genital opening dorsal orpostero-dorsal. Male genitalia variable*. HOSTS. Bovidae, Cervidae and Camelidae (Artiodactyla). COMMENTS. Some species of Bovicola are parthenogenetic, males being rare or unknown. A summary of the varying taxonomic treatments of Bovicola, its subgenera and synonyms, is presenteein Table 3. TRICHODECTID MAMMAL LICE 251 Subgenus BOVICOLA Ewing (Figs 3, 42, 45, 46, 48, 49, 51-54) Bovicola Ewing, 1929: 193. Type-species: Trichodectes caprae Gurlt, by original designation. Bo vidoecus Bedford, 1929: 518. Type-species: Pediculus bovis Linnaeus, by original designation. [Synony-my by Bedford, 1932a: 356.] Rhabdopedilon Keler, 1938a: 453. Type-species: Trichodectes longicornis Nitzsch, by original designation.[Synonymy by Werneck, 1950: 59.] DESCRIPTION. Clypeal marginal carina not broadened medially, or more or less broadened into bar withposterior margin straight or matching curvature of osculum. Temple margin smoothly convex, lackingprojection on postero-lateral angle, not convexly produced posteriad to great extent. Male scape notexpanded or only slightly expanded. Dorsum of head with setae short or of moderate length, of greaterabundance anteriorly than posteriorly. Sitophore sclerite unmodified. Thorax with lateral and dorsal setae long and of moderate length, sometimes abundant and numerous ondisc of prothorax and pterothorax, otherwise less abundant and sparsely scattered on disc of pterothoraxwith only two setae present on disc of prothorax; setae present along lateral margins and posteriorly(dorsally) on prothorax and pterothorax; posterior setal row of prothorax marginal, with median gappresent or absent; posterior setal row of pterothorax submarginal, with median gap absent, rowincorporating two very long setae between postero-lateral and postero-median angles or, if setae generallyabundant on thorax, postero-lateral setae of pterothorax longer than others. Abdominal setae short, long or of medium length; setal bases, at least of setae of posterior setal row onsterna and terga, enlarged, clearly with doubled margins. Pregenital sclerites present on sterna and terga(where present) of all segments, except sometimes tergum I and (independently) tergum VIII of males;terga of males, at least of segments IV- VI, with both anterior and posterior sclerites. Gonapophyses with lobe rectangular, acute, rounded or not pronounced; marginal setae confined tolobe, long. Ventral vulval margin not sclerotised; convex, biconvex with median indentation, or convexwith small median membranous projection (Fig. 42); margin smooth or spinose. Postgenital area lackingspinose patch. Genital chamber with antero-median dorsal area lacking spicules, scales or other decora-tion , either very narrow and strongly-defined or wide and ill-defined , or with very narrow longitudinal fold . Male subgenital plate variable; stqrnites VII and IX present, fused to s.g.p.r., sternite VIII absent or, ifpresent, fused or not fused to s.g.p.r. (Figs 46, 48). Pseudostyli absent (Fig. 49) or, if present, setose andlobulate (Figs 46, 48). Basal apodeme very concave anteriorly, the sides of the concavity frequently beingparallel, though sometimes obscure. Parameres with broad basal flange or block; sometimes very reduced.Basiparameral sclerites present and fused, or absent. Mesomeres, if fused apically, forming very narrowarch lacking median extension; otherwise mesomeres not fused, sometimes very reduced and obscure.Male genitalia depicted in Figs 51-54. HOSTS. Bovidae and Cervidae (Artiodactyla). COMMENTS. Some of the species in this subgenus are parthenogenetic. SPECIES INCLUDEDalpinus Keler, 1942 (5 cf , 3 $)bovis (Linnaeus, 1758) (7 cf, 137 9)caprae (Gurlt, 1843) (c.50 Cf , c.50 $) concavifrons (Hopkins, 1960) [Recalled from synonymy with longicornis (Nitzsch).] (2 cf , 98 9)jellisoni Emerson, 1962 (10 cf , 10 $)7/mbafus(Gervais, 1844) (c.SOcT, c.65 9)longicornis (Nitzsch, 1818) (44 $)oreamnidis (Hopkins, 1960) (holotype cf)ov/s(Schrank, 1781) (59 cf, 64 9)tarandi (Mjoberg, 1910) (2 $ , 7 nymphs)t/Wa/is(Piaget, 1880) (c.100 9) Subgenus HOLAKARTIKOS Keler gen. rev., stat. n. (Figs 44, 47, 57) Holakartikos Keler, 1938a: 461. Type-species: Trichodectes pilosus Piaget (nee Giebel) [= Trichodectescrassipes Rudow], by original designation. DESCRIPTION. Anterior of head with osculum absent; pulvinus very short, not attaining anterior margin of 252 C. H. C. LYAL head; dorsal preantennal sulcus absent, though ventral preantennal sulcus sometimes present; clypealmarginal carina not always pronounced and not, or only slightly, broadened medially; preantennal portionof head very short, outline smoothly and shallowly rounded. Temple margin smoothly convex, lackingprojection on postero-lateral angle, convexly produced posteriad. Male scape very slightly expanded, withsetae randomly scattered. Dorsum of head with abundant setae of moderate length; temple with longpostero-lateral marginal setae. Sitophore sclerite unmodified. Thorax with abundant setae, long and of moderate length, present on margins and disc of prothorax andpterothorax; setae longest on the rounded postero-lateral angles of prothorax and ptero thorax. Abdomen with long setae of posterior setal row, and shorter anterior setae, present on sterna, terga andpleura (where present) of all segments (Fig. 47). Pre-genital sclerites sometimes very faint, present onsterna and terga (where present) of all segments except tergum I and sternum II; male terga lackingposterior sclerites. Gonapophyses with broadly rounded lobe smoothly continuous with ventral margin; marginal setaelong, present all along ventral margin, including lobe. Ventral vulval margin not sclerotised; produced intothree weakly-developed lobes. Postgenital pleural area with patch of short, spine-like setae. Genitalchamber lacking dorsal non-ornamented area or fold. Male subgenital plate with s.g.p.r. not joining sternites VII and VIII, and sometimes failing to contacteither or both; sternites sometimes very faintly sclerotised, obscure; s.g.p.r. with broad lateral flange onVIII and IX (Fig. 47). Pseudostyli absent. Male genital opening dorsal. Basal apodeme long, not concaveanteriorly. Parameres long, slender, with basal block and flange. Basiparameral sclerites absent. Meso-meres fused apically, with median extension present (see comments below). Male genitalia depicted in Fig.57. HOSTS. Bovidae (Artiodactyla). COMMENTS. The only included species is not known to be parthenogenetic. Werneck (1950) failed to recognise the median extension of the mesomeral arch, and considered itabsent. Holakartikos was considered a synonym of Bovicola by Werneck (1950) and Emerson & Price (1981); amore extensive history of the variations in status of this subgenus is presented in Table 3. SPECIES INCLUDEDcrassipes(Rudow, 1866) (24 cf, 31 $) Subgenus LEPIKENTRONKeler gen. rev., stat. n. (Figs 43, 55)Lepikentron Keler, 1938a: 452. Type-species: Trichodectes breviceps Rudow, by original designation. DESCRIPTION. Anterior of head with osculum absent; pulvinus very short, not attaining anterior of marginof head; dorsal preantennal sulcus absent; clypeal marginal carina not pronounced, not broadenedmedially; preantennal portion of head shorter in male than female, outline broadly and smoothly rounded.Temple margin smoothly convex, lacking projection on postero-lateral angle, not convexly producedposteriad to great extent. Male scape expanded, with setal row apparently present, though setae may bescattered randomly. Dorsum of head with setae of moderate length, slender; setae sparsely distributed,more abundant anteriorly than posteriorly. Sitophore sclerite unmodified. Thorax with lateral and dorsal setae slender, long and of moderate length; setae present postero-laterally and posteriorly on prothorax and along lateral margins and posteriorly (dorsally) on pterothorax;posterior setal row of prothorax submarginal, sparse, with large median gap; posterior setal row ofpterothorax submarginal, with small median gap, setae shorter medially than laterally, with two long setaelaterally; pair of setae, widely spaced, present on disc of prothorax dorsally; setae not present on disc ofpterothorax. Abdominal setae of moderate length, slender; anterior setae never present on sterna and terga.Pre-genital sclerites present on terga II-VII (males) and terga III-IX (females) and sterna III-VII (males)and III- VIII (females); male terga III-VII with both anterior and posterior sclerites, though the posteriorelements may be very faintly sclerotised and difficult to see. Gonapophyses with small, pronounced lobe and broad tapering spur (Fig. 43); marginal setae confinedto lobe. Ventral vulval margin sclerotised medially; shallowly convex or biconvex (Fig. 43). Postgenitalpleural area lacking spinous patch. Genital chamber lacking median non-ornamented area or fold. Male subgenital plate with sternite VII fused to s.g.p.r. and sternites VIII and IX absent; s.g.p.r. broad.Pseudostyli absent. Male genital opening poste/o-dorsal. Basal apodeme slightly longer than parameres, TRICHODECTID MAMMAL LICE 253 convex anteriorly. Parameres long, slender. Basiparameral sclerites absent. Mesomeres not apically fused;each with median desclerotisation, and appearing as two rods (Fig. 55). HOSTS. Camelidae (Artiodactyla). COMMENTS. Only one male of the single included species is known and the species may be parthenogenetic.The subgenus was treated as a synonym of Bovicola by Werneck (1950) and Emerson & Price (1981); amore extensive history of the variations in status of this subgenus is presented in Table 3. SPECIES INCLUDED6rev/ceps(Rudow, 1866) (1 cf , 24 $) Subgenus SPINIBOVICOLA subgen. n. (Figs 12, 41, 50, 56)Type-species: Trichodectes hemitragi Cummings. DESCRIPTION. Anterior of head with osculum absent; pulvinus very short, but attaining anterior margin ofhead; dorsal preantennal sulcus absent; clypeal marginal carina broadened medially into straight bar withposterior margin slightly irregular; anterolateral margin of head smoothly rounded; preantennal portionshort, with outline rounded, almost straight anteriorly. Temple margin convex laterally, straight posterior-ly, with short posteriorly-directed projection on postero-lateral angle bearing two setae. Male scapeslightly expanded, with setae randomly scattered. Dorsum of head with abundant setae of moderatelength. Sitophore sclerite with posterior arms extended (Fig. 12). Thorax with dorsal and lateral setae abundant, long or of moderate length, present marginally and ondisc of prothorax and pterothorax; longest setae present postero-laterally on pterothorax. Abdomen tapering more acutely in male than in female. Abdomen with posterior setal row comprisinglong setae on sterna, terga and pleura, anterior setae shorter; anterior setae present on sterna, terga andpleura (where present) of all abdominal segments except sometimes tergum I; male tergum II with curvedrow of 3-4 long stout setae on each side, modified from posterior setal row, these setae being linked by acurved sclerite (modified tergite) (Fig. 41). Pre-genital sclerites present on terga II-VII or VIII (males) andII-IX (females) and sterna II-VII (males) and II- VIII (females); male terga lacking posterior sclerites. Gonapophyses with broadly rounded lobe smoothly continuous with ventral margin; marginal setaelong, confined to lobe. Ventral vulval margin not sclerotised; convex. Postgenital pleural area lackingspinose patch. Genital chamber lacking dorsal non-ornamented area or fold. Male segment IX produced posteriorly into narrow, sclerotised extension; subgenital plate taperingcharacteristically, comprising sternites VI and VII linked by s.g.p.r. (Fig. 50). Pseudostyli absent. Malegenital opening dorsal. Basal apodeme not as long as parameres, not concave anteriorly. Parameres fusedbasally, long and tapering to pointed apices. Basiparameral sclerites absent. Mesomeres absent, orrepresented by very short sclerites, not apically fused. Male genitalia depicted in Fig. 56. HOSTS. Bovidae (Artiodactyla). COMMENTS. Neither of the two included species is known to be parthenogenetic. SPECIES INCLUDED hemitragi (Cummings, 1916) (20 cf , 26 $)multispinosus Emerson & Price, 1979 (8 C?, 11 $) Genus BISONICOLA gen. n. (Figs 58-60)Type-species: Bovicola sededmdecembrii Eichler. DESCRIPTION. Anterior of head with osculum absent, though pulvinus attaining margin; margin anteriorlyto pulvinus membranous, hyaline; dorsal preantennal sulcus present; clypeal marginal carina slightlybroadened medially; anterolateral margin of head smoothly convex; preantennal portion of head short orlonger, but not as long as postantennal portion, outline broadly rounded, though slightly truncateanteromedially. Temple margin broadly and smoothly convex. Male scape expanded, with setae randomlyscattered; flagellomeres fused in males and females; male flagellum with two or three basally-articulated'teeth'; male flagellum not 'roughened' on interior face. Dorsum of head with numerous setae of mediumlength. Sitophore sclerite unmodified.Thorax with lateral and dorsal setae long and of moderate length; setae present along lateral margins and 254 C. H. C. LYAL 58 60 Figs 58-60 Bisonicola sedecimdecembrii. 58, $ terminalia, ventral. 59, O" subgenital plate, setae omitted.60, cf genitalia. posteriorly (dorsally) on prothorax and pterothorax; posterior setal row on prothorax marginal, withmedian gap; posterior setal row of pterothorax submarginal, with median gap absent; pterothorax withposterior setal row incorporating two very long setae between postero-lateral and postero-median angles;pair of setae, widely spaced, present on disc of prothorax dorsally; scattered setae sometimes presentposteriorly on disc of pterothorax dorsally. Atria of thoracic spiracles very large. Abdomen oval. Abdominal spiracles present on segments III- VIII. Abdomen with setae short and ofmedium length; anterior setae present on sterna, terga and pleura; postero-lateral setae absent. Abdomin-al pleural projections absent. Sclerites present on sterna, terga and pleura (where present) of all pre-genitalabdominal segments except tergum I; male terga V, VI and VII with anterior and posterior sclerites. Gonapophyses broad, truncate; setae present along postero-median margin; ventral lobe absent (Fig.58). Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulvalmargin not sclerotised; very short, more or less straight; subgenital lobe absent. Male subgenital plate with sternites VII and IX fused to s.g.p.r., sternite VIII present but not fused tos.g.p.r.; s.g.p.r. heavily sclerotised, widest on sternum VIII (Fig. 59). Pseudostyli present, large, broadlytriangular (Fig. 59). Male genital opening dorsal. Parameres separate, rod-like, fused to mesomeral arch.Basiparameral sclerites absent. Mesomeres fused apically; median extension absent. Male genitaliadepicted in Fig. 60. HOSTS. Bovidae (Artiodactyla). SPECIES INCLUDED sedecimdecembrii sedecimdecembrii (Eichler , 1946) comb. n. from Bovicola (5 d", 4 9)sedecimdecembrii bison (Blagoveshtchenski, 1967) comb. n. from Bovicola (2 C?, 1 $) TRICHODECTID MAMMAL LICE 255 Genus WERNECKIELLA Eichler gen. rev. (Figs 61, 62)Werneckiella Eichler, 1940: 160. Type-species: Trichodectes equi Denny, by original designation. DESCRIPTION. Anterior of head with osculum absent; pulvinus not attaining margin; dorsal preantennalsulcus present; clypeal marginal carina slightly broader medially than laterally, or not broadened;anterolateral margin of head smoothly convex; preantennal portion of head not long, outline broadlyrounded, sometimes slightly flattened anteriorly. Temple margin convex or rectangular. Male scapeexpanded, with setae randomly scattered; flagellomeres fused in males and females; male flagellum withtwo basally-articulated 'teeth'; male flagellum not 'roughened' on interior face. Dorsum of head withabundant short setae. Sitophore sclerite unmodified. Thorax with lateral and dorsal setae short and of medium length; setae present along lateral margins andposteriorly (dorsally) on prothorax and pterothorax; posterior setal row on prothorax marginal, withmedian gap; posterior setal row on pterothorax marginal or submarginal, with no median gap; posteriorsetal row of pterothorax incorporating two very long setae with intervening shorter setae betweenpostero-lateral and postero-median angles; prothorax with seta or setae on disc laterally (dorsally). Abdomen elongate-oval. Abdominal spiracles present on segments III-VIII. Abdomen with setae shortand of moderate length; anterior setae frequently present on sterna and terga, always present on pleura;postero-lateral setae absent. Abdominal pleural projections absent. Sclerites present on abdominal pleuraII- VII and sterna and terga of at least abdominal segments II- VII; male terga with single sclerites only. Gonapophyses broad, truncate, with median faces almost parallel to one another; marginal setae ofmoderate length; ventral lobe absent. Gonapophyses meet ventral vulval margin acutely, not linked bysclerotised band. Ventral vulval margin not sclerotised, very short, straight; subgenital lobe absent. Male subgenital plate with sternite VII present and fused to s.g.p.r., sternites VIII and IX absent;s.g.p.r. sinuate and broadest on sternum VIII (Fig. 61). Pseudostyli present, small, simple setose lobes.Male genital opening postero-dorsal. Parameres long, straight or flared and twisted medially, sometimesfused basally. Basiparameral sclerites absent. Mesomeres fused or almost fused into pentagonal mesomer-al arch with median extension absent; mesomeres broadest basally (external to b.a.l.s.) and more or lessbroad distally; mesomeres projecting basally between b.a.l.s. to contact parameres. Male genitaliadepicted in Fig. 62. HOSTS. Equidae (Perissodactyla) and Bovidae (Artiodactyla). COMMENTS. Some species are parthenogenetic, the males being unknown. Werneckiella was considered a subgenus of Damalinia by Hopkins (1949) and a synonym of Bovicola byWerneck (1950); it is here raised from synonomy with Bovicola. A more extensive history of the variationsin status of this genus is presented in Table 3. The genus was revised by Moreby (1978). SPECIES INCLUDED aspilopyga (Werneck, 1956) comb. n. from Bovicola (9 cf , 11 $)equi (Denny, 1842) comb. n. from Bovicola (3 cT, c. 100 <j>)fulva (Emerson & Price, 1979) comb. n. from Bovicola (4 cf , ?1 $)neglecta (Keler, 1942) comb. n. from Bovicola (5 cf , 6 $)ocellata (Piaget, 1880) comb. n. from Bovicola (17 $)zebrae Moreby, 1978 comb. rev. from Bovicola (1 cf , 2 <j>)zuluensis (Werneck, 1950) comb. n. from Bovicola (13 (?, 13 <j>) Genus TRAGULICOLA gen. n. (Figs 63, 64)Type-species: Damalinia traguli Werneck. DESCRIPTION. Anterior of head with osculum present, broad; dorsal preantennal sulcus present; clypealmarginal carina broadened medially into less heavily sclerotised dorsal sclerite, which is broad, posteriorlyconvex and with median posterior projection (Fig. 63); anterolateral margin of head straight or slightlysinuate; preantennal portion of head as long as its maximum width, outline trapezoid. Temple marginconvex or rectangular. Male scape expanded, with setal row present and comprising at least four setae;flagellomeres fused in males and females; male flagellum with two basally-articulated 'teeth'. Dorsum ofhead with setae short posteriorly and of moderate length anteriorly. Sitophore sclerite unmodified.Thorax with dorsal setae short or of moderate length; prothorax and pterothorax with marginal or 256 C. H. C. LYAL I 64 62 Figs 61-64 Bovicolinae species. 61, 62, Werneckiella equi C?, (61) terminalia; (62) genitalia. 63, 64,Tragulicola traguli, (63) $ head, dorsal; (64) cf genitalia. submarginal posterior setal row, the longest setae being posterolaterally except in the male, which has apair of long setae medially on the posterior row of the pterothorax; male with setal patch centrally on disc ofpronotum, but no other setae present on disc of either sex. Abdomen elongate, with male tapering to more acute posterior angle than female. Abdominal spiraclespresent on segments III- VIII; atria oblate-spheroids, very large. Abdomen with setae of moderate length,the longest being those comprising the pleural posterior setal row, particularly of the posterior pleura;anterior setae present on all pleura, but not sterna or terga; postero-lateral setae absent. Abdominalpleural projections absent. Sclerites present on sterna, terga and pleura (where present) of all abdominalsegments except tergum I, which is reduced and obscure; male terga HI-VI with anterior and posteriorsclerites. Gonapophyses broad medially, tapering smoothly distally; ventral margin with long, abundant setae;ventral lobe absent. Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band.Ventral vulval margin not sclerotised: convex; subgenital lobe absent; marginal spines present, thoughdifficult to see. Female genital chamber with dorsal wall lacking spicules over narrowly triangular areaanteromedially. Male subgenital plate with sternites IX and VIII linked by s.g.p.r., but sternite VII not attache< TRICHODECTID MAMMAL LICE 257 Pseudostyli present, short, conical; median ventro-posterior projection also present, longer than pseudo-styli. Male genital opening postero-dorsal. Basal apodeme acuminate apically. Parameres broad, triangu-lar, poorly-sclerotised, asymmetrically deflected (may be artifact of preparation, though deflected thesame way in all specimens seen). Basiparameral sclerites absent. Mesomeres fused apically, medianextension absent; mesomeral arch fused to b.a.l.s. about one-third length of basal apodeme anteriad fromposterior end. Male genitalia depicted in Fig. 64. HOSTS. Tragulidae (Artiodactyla). SPECIES INCLUDED fragu/j (Werneck, 1950) comb. n. from Damalinia (70 cf , 75 $) Genus DAMALINIA MjobergThe genus Damalinia comprises three subgenera. DESCRIPTION. Anterior of head with osculum present, narrow or broad, deep or shallow, or osculumabsent, in which case head as described below for D. (T.) conectens*; dorsal preantennal sulcus present orabsent*; clypeal marginal carina more or less broadened medially and of variable form*; anterolateralmargin of head straight, slightly concave, slightly sinuate or convex; preantennal portion of head ofvariable length, outline triangular, trapezoid, rectangular or rounded*. Temple margin smoothly convex,sometimes convexly produced posteriad* , with posterolateral angle sometimes developed laterally or withsmall posterior rounded projection*. Male scape expanded, with setal row present or setae randomlyscattered*; flagellomeres fused in males and females; male flagellum with two or three basally-articulated'teeth'* and interior face serrate or 'roughened'. Dorsum of head with setae sparse or more or lessabundant, short or of moderate length, frequently longer along the anterolateral margins and across theclypeus than elsewhere. Sitophore sclerite unmodified. Thorax with dorsal and marginal setae short, long or of moderate length, frequently longest onposterolateral margin of pterothorax. Prothorax with setae sparse or absent on anterolateral margin;posterior setal row marginal, though directed onto disc medially and median setal pair sometimes isolated,row more or less sparse, with median gap between setae (other than isolated median pair) present,sometimes wide*; single seta frequently present on dorsal disc anterolaterally. Pterothorax with posteriorsetal row marginal or submarginal, sometimes irregular or 'doubled'*, median gap present or absent,posterior setal row incorporating two very long setae with intervening shorter setae between postero-lateral and postero-median angles; setae absent from disc. Abdomen oval, elongate, or very elongate and narrow*. Abdominal spiracles present on segmentsIII- VIII. Abdominal setae short or of moderate length, frequently longer on pleura than on sterna and 65 Figs 65-67 meyeri, Damalinia species. 65, D. (Tricholipeurus) elongata, $ terminalia, ventral. 66, D. (Cervicola)gonapophysis, ventral. 67, D. (C.) hendrickxi, $ gonapophysis, ventral. 258 C. H. C. LYAL 68 Figs 68-70 Damalinia species. 68, D. (D.) crenelata, C? abdomen. 69, D. (D.) baxi, $ head, dorsal. 70,D. (Cervicola) martinaglia, cf scape. terga; anterior setae present on all pleura except, occasionally, pleurum II, rarely on sterna and terga;postero-lateral setae absent. Abdominal pleural projections absent. Pregenital sclerites present on sterna,terga and pleura (where present) of all segments except tergum I , sometimes absent or very small on pleura(Fig. 80)*; male terga with or without posterior sclerites*. Gonapophyses variable; ventral margin without rounded lobe, but sometimes with hook-shapedprojection*; marginal setae present. Gonapophyses meet ventral vulval margin acutely, not joined bysclerotised band. Ventral vulval margin sclerotised or not sclerotised; subgenital lobe present or absent*.Dorsal margin of vulva and post-vulval area with or without pointed scales*. Common oviduct, atbranching point, with or without collar (see subgenus Cervicola)*. Male subgenital plate variable, sternites VII and VIII always being present and fused to s.g.p.r. , sterniteIX sometimes modified. Pseudostyli absent or, if present, of variable form*. Posterior margins of tergum TRICHODECTID MAMMAL LICE 259 76 Figs 71-76 Damalinia (D.) species. 71, D. theileri, cf terminalia. 72, D. appendiculata, cf terminalia. 73,D. theileri, $ terminalia. 74-76, cf genitalia of (74) D. orientalis; (75) D. neotheileri; (76) D. crenelata. 260 C. H. C. LYAL 77 78 79 80 Figs 77-80 Damalinia species, d" terminalia. 77, D. (Cervicola) natalensis, ventral, setae omitted. 78, D.(C.) martinaglia, ventral, setae omitted. 79, D. (Tricholipeurus) indica, ventral, setae omitted. 80, D.(T.) aepycerus. IX sometimes greatly expanded * . Male genital opening postero-dorsal or dorsal . Male genitalia variable * .HOSTS. Bovidae and Cervidae (Artiodactyla). Subgenus DAMALINIA Mjoberg (Figs 11, 68, 71-76)Damalinia Mjoberg, 1910: 69. Type-species: Trichodectes crenelatus Piaget, by monotypy. DESCRIPTION. Anterior of head with osculum present, narrow or broad, deep or shallow; dorsal preanten-nal sulcus present; clypeal marginal carina broadened medially either into more or less developed simple TRICHODECTID MAMMAL LICE 261 81 Figs 81-86 Damalinia species, cf genitalia. 81, D. (Tricholipeurus) victoriae. 82, D. (T.) indica. 83, D.(T.) aepycerus. 84, D. (Cervicola) hopkinsi. 85, D. (C.) meyeri. 86, D. (C.) reduncae. bar with posterior margin straight or concave, or into more or less broad U-shaped sclerite, or into broad,heavily-sclerotised margin of deep osculum; anterolateral margin of head convex or slightly sinuate, in thelatter case slightly concave at junction of margin and clypeof rental sulcus and convex anteriorly;preantennal portion longer or shorter than posterior portion, outline triangular, trapezoid or rounded,sometimes with slight protuberances on either side of osculum. Temple margin smoothly convex, slightlyproduced posteriad, sometimes with postero-lateral angle developed laterally, or with small posteriorrounded projection. Male scape with setae randomly scattered; male flagellum with two basally-articulated'teeth'. Dorsum of head with setae more or less abundant. 262 C. H. C. LYAL Pterothorax with posterior setal row sometimes irregular or 'doubled'. Abdomen oval or elongate, sometimes very narrowly elongate. Abdominal setae present anteriorly onall pleura, occasionally on terga and sterna, but only laterally and as irregularity or 'doubling' of posteriorsetal row. Pleurum II with sclerite extending broadly or narrowly onto sternum II and sometimes tergumII, frequently 'crowding' sternite or tergite II (Fig. 68); pleurites not reduced in size or absent; tergum Ilacking sclerite; male terga V and VI (at least) with both anterior and posterior sclerites. Gonapophyses variable, sometimes hook-shaped, though lacking distal spur, more frequently obtuse,sometimes with ventral (median) margin concave or convex; ventral margin with setae long or of moderatelength, abundant, setae sometimes also present on anterior margin. Ventral vulval margin not sclerotised,sometimes short, straight or convex; subgenital lobe absent, though posterior margin of sternum VIIsometimes developed into two spikes (Fig. 73). Dorsal margin of vulva and postvulval area usually withoutpointed scales. Common oviduct without 'collar'. Male subgenital plate with sternites VII, VIII and IX fused to s.g.p.r. , sternite IX and postgenital scleritesometimes fused; s.g.p.r. more heavily sclerotised than sternites (Fig. 68). Pseudostyli absent or, ifpresent, long and broad, parallel-sided or with basal constriction (Figs 68, 71), or long and narrow (Fig.72). Posterior margin of tergum IX not greatly expanded. Parameres more or less broad, sometimes fusedtogether. Basiparameral sclerites absent. Mesomeres unfused apically, sometimes fused to b.a.l.s. incharacteristic manner (Fig. 75) or to parameres, basally or in entirety, in the latter case apparently absent.Endophallus lacking spicular patch (cf. subgenus Tricholipeurus). Male genitalia depicted in Figs 7476. HOSTS. Bovidae (Artiodactyla). COMMENTS. Emerson & Price (1982) distinguish their new species orientalis (described in Bovicold) fromthe very similar species thompsoni Bedford on the following grounds: The female of B. orientalis is smallerthan that of B. thompsoni and the lateral margins of the forehead are even [sic] so slightly indented for B.orientalis and always even for B. thompsoni; the median plates on tergites II- VIII are of different shapesfor the two species; the chaetotaxy of terminal abdominal segments is different, with each gonapophysishaving at least 20 median and anterior setae for B. thompsoni [orientalis having, according to the precedingdescription, 12-17 setae]; and the posterior margin of the temple of B. thompsoni has small projections thatare not present for B. orientalis. The male of B. thompsoni is unknown.' Five male and seven femaleparatypes of B. orientalis were examined in this study, together with a further eight males and eight femalesfrom the same host (Capricornis crispus swinhoei) not examined by Emerson & Price when they preparedtheir description of orientalis, and three females of thompsoni, including the holotype. Treating thesupposed distinguishing characters in order, the three specimens of thompsoni are larger than any of thefemales from C. c. swinhoei; whilst none of the specimens of thompsoni has an indentation on the forehead(at the junction of the clypeo-frontal sulcus with the margin), not all of the specimens from Capricornis c.swinhoei have either; the shapes of abdominal tergites II- VIII are not significantly different in the twospecies; the chaetotaxy of the terminal abdominal segments is not different, and no specimen of thompsonihas more than 17 setae on the median and anterior margins of the gonapophyses; the small projections ofthe posterior temple margins are present in all specimens of orientalis. It seems, therefore, that thedifferences between the two species are twofold: host (orientalis being described from Capricornis crispusswinhoei and thompsoni being known only from Capricornis sumatrensis sumartrensis) , and size. Thebiological significance of the latter character is not clear, and the two species may be found to differ in othercharacters not so far discovered. Until a larger sample can be examined, collected from more localities, notaxonomic action is taken to reduce the rank or synonymise orientalis, though on the basis of theinformation so far available the species probably should not stand. SPECIES INCLUDED adenota (Bedford, 1936) comb. n. from Bovicola (39 cf , 35 9) appendiculata (Piaget, 1880) (19 cf , 25 9) fcaxiHopkins, 1947 (16 cf , 24 9) chorleyi (Hopkins, 1941) (21 cf , 19 9) crenelata (Piaget, 1880) (27 cf , 20 9) dimorpha (Bedford, 1939) comb. n. from Bovicola (syntype 9) /a/ireii/io/zi'(Eichler, 1949) comb. n. from Tricholipeurus harrisoni (Cummings, 1916) (3 cf , 3 9) hiltt (Bedford, 1934) comb. n. from Bovicola (42 cf , 27 9) neotheileri Emerson & Price, 1971 (1 cf , 6 9) orientalis (Emerson & Price, 1982) comb. n. from Bovicola (108 cf , 135 9) ornate Werneck, 1957 (Holotype cf ) pelea (Bedford, 1934) comb. n. from Bovicola TRICHODECTID MAMMAL LICE 263 semitheileri Emerson & Price, 1971 (holotype cf , allotype $) theileri Bedford, 1928 (2 cf . 4 $) thompsoni (Bedford, 1936) comb. n. from Bovicola (3 $) Subgenus CERVICOLA Keler gen. rev., stat. n. (Figs 66, 67, 70, 77, 78, 84-86) Cervicola Keler, 1934: 263. [Nomen nudum.] Cervicola Keler, 1938a: 460. Type-species: Trichodectes tibialis Keler (nee Piaget) [= Trichodectes meyeriTaschenberg], by original designation. DESCRIPTION. Anterior of head with osculum present, narrow or broad, deep or shallow; dorsal preanten-nal sulcus present; clypeal marginal carina broadened medially, either into simple bar with posteriormargin straight or convex but occasionally with median posterior projection, or into longer posteriorly-developed sclerite with posterolateral angles more or less convex and more or less pronounced medianposterior projection; anterolateral margin of head straight or slightly sinuate, in the latter case slightlyconcave at junction of margin and clypeof rental sulcus and convex anteriorly; preantennal portion of headas long as or shorter than postantennal portion, outline triangular, trapezoid or rounded. Temple marginsmoothly convex, slightly produced posteriad. Male scape with setae randomly scattered; male flagellumwith two basally-articulated 'teeth'. Dorsum of head with setae more or less abundant, sometimes less soposteriorly than anteriorly. Pterothorax with posterior setal row single. Abdomen oval or elongate. Abdominal setae present anteriorly on all pleura except, occasionally,pleurum II, but may be very short, fine and difficult to see; anterior setae never on sterna and terga.Pleurum II never with sclerite extending onto sternum II; pleurites not reduced in size or absent; male tergaV and VI (at least) with both anterior and posterior sclerites. Gonapophyses hook-shaped, apex of curved portion acute or rounded, sometimes with distal (dorsal)spur (Figs 66, 67). Gonapophyses with setae long or of moderate length on posterior margin and sometimeson apex of 'hook', smaller setae sometimes present on anterior margin of 'hook'. Ventral vulval margin notsclerotised. Subgenital lobe absent. Dorsal margin of vulva and post-vulval area with pointed scales.Common oviduct at branching-point with folded and more or less apparent 'collar', sometimes partiallysclerotised and refracting light when viewed in phase-contrast or bright field transmitted light. Male subgenital plate with sternites VII, VIII and IX fused to s.g.p.r. , though sternite IX sometimes notcomplete, s.g.p.r. sometimes not attaining posterior margin of segment IX, perisetal gaps sometimesabsent; s.g.p.r. more heavily sclerotised than sternites. Pseudostyli absent or, if present, apically pointedand more or less broad (Fig. 78) or apically rounded and very narrow (Fig. 77). Posterior margins of maletergum IX not greatly expanded. Parameres broad or narrow, larger or smaller than mesomeres, may bereduced to small discs, in which case mesomeres absent; parameres fused or unfused; apices sometimeswidely divergent. Basiparameral sclerites present or absent. Mesomeres absent or, if present, not fusedapically, nor fused to parameres or b.a.l.s. Endophallus lacking spicular patch (cf. subgenus Tricho-lipeurus).^ Male genitalia depicted in Figs 84-86. HOSTS. Bovidae and Cervidae (Artiodactyla). COMMENTS. Cervicola was treated as a synonym of Damalinia by Werneck (1950), and of both Damaliniaand Bovicola by Hopkins & Clay (1952, pp. 102 and 67 respectively). The history of the variations in statusof Cervicola is presented in Table 3. SPECIES INCLUDED annectens Hopkins, 1943 comb. rev. from Tricholipeurus (21 cf , 25 $)forficula (Piaget, 1880) (5 cf , 8 $)hendrickxi Hopkins, 1947 (4 cf , 6 $)hopkinsi Bedford, 1936 (15 cf , 28 $)lerouxi (Bedford, 1930) comb. n. from Tricholipeurus (9 $)maa/ Emerson & Price, 1973 (holotype cf , allotype $)martinaglia (Bedford, 1936) (34 cf , 27 $)meyeri meyeri (Taschenberg, 1882) (5 cf , 54 $)meyeri hydropotis (Dobroruka, 1975) comb. n. from Cervicolameyeri sika (Dobroruka, 1975) comb. n. from Cervicolamuntiacus (Seguy, 1948) (12 cf , 12 $)natalensis Emerson, 1964 (2 cf , 2 $) 264 C. H. C. LYAL reduncae (Bedford, 1929) stat. n., comb. n. from Tricholipeurus [Raised from subspecies of D. trabeculae.} (2% <3, 29%) trabeculae (Bedford, 1929) comb. n. from Tricholipeurus (10 cf , 10 $)Uganda? (Werneck, 1950) stat. n., comb. n. from Tricholipeurus [Raised from subspecies of D. trabeculae.] (14 cf , 13 $)Unless otherwise stated, all species were previously considered as placed in Damalinia s. str. Subgenus TRICHOLIPEURUS Bedford stat. n. (Figs 4, 7, 8, 13,65,79-83) Tricholipeurus Bedford, 1929: 514. Type-species: Tricholipeurus aepycerus Bedford, by original designa-tion. DESCRIPTION. Anterior of head variable, one of two types. (a) Osculum absent; pulvinus short, not attaining anterior margin of head; dorsal preantennal sulcusabsent; clypeal marginal carina insignificant, not, or only slightly, broadened medially; anterolateralmargin of head straight posteriorly, convex anteriorly; preantennal portion of head as long as postantennalportion, outline rounded anteriorly (D. (T.) conectens only). (b) Osculum present, rarely deep; dorsal preantennal sulcus present; clypeal marginal carina broadenedmedially either into simple bar with posterior margin straight or convex, or into longer posteriorly-developed U- or W-shaped sclerite, with posterolateral angles more or less acutely convex, and frequentlywith more or less pronounced median posterior projection; anterolateral margin straight, slightly sinuate,or slightly concave; preantennal portion of head longer or shorter than postantennal portion but not short,outline trapezoid or rectangular. Temple margin smoothly convex, more or less convexly producedposteriorly. Male scape with setal row frequently present, comprising four or five setae, though setaesometimes more or less randomly scattered; male flagellum with two or three basally- articulated 'teeth'.Dorsum of head with setae sometimes sparse. Pro thorax with median gap of posterior setal row sometimes almost the width of the posterior margin.Pterothorax with posterior setal row single. Abdomen usually very elongate, narrow. Abdominal setae present anteriorly on all pleura except,occasionally, pleurum II, but may be very short, fine and difficult to see; anterior setae never on sterna andrarely on terga. Pleurum II never with sclerite extending onto sternum II; pleural sclerites frequentlyreduced to small anterior plate or absent; tergum I sometimes lacking sclerite; male terga lacking posteriorsclerites, or with both anterior and posterior sclerites present on at least segments V and VI (though may bepresent on any segments up to II- VIII). Gonapophyses variable, sometimes hook-shaped with distal spur and marginal setae on posterior(dorsal) margin of 'hook', or long with ventral margin convex, straight or sinuate (though not with lobe),apically pointed or rounded; marginal setae of moderate length. Ventral vulval margin sometimessclerotised; subgenital lobe present, variable, not marginally serrate, ventrally smooth or scaled. Dorsalmargin of vulva and post-vulval area without pointed scales. Common oviduct without 'collar'. Male subgenital plate with sternites VII, VIII and IX fused to s.g.p.r.; perisetal gaps sometimes large,occasionally absent; s.g.p.r. heavily sclerotised and sternites sometimes very lightly scerlotised; s.g.p.r.sometimes curved or sinuate (Fig. 80). Pseudostyli absent or, if present, variable, short and rounded orapically angular, posteriorly or medially directed, narrow or broad, sometimes fused to form single caudalprojection. Posterior margins of male tergum IX frequently greatly expanded (Fig. 79). Lateral struts ofbasal apodeme sometimes with anteposterior spur (Fig. 83). Parameres unfused or, if fused, plate apicallypointed or bifurcate. Basiparameral sclerites present or absent. Mesomeres unfused apically or, if fused,symmetric or asymmetric, median extension absent or present; mesomeral arch frequently recurvedabruptly at base to contact parameres, sometimes extended between b.a.l.s.; mesomeres not fused toparameres or b.a.l.s. Endophallus with patch of regularly-arranged and numerous spicules sometimes veryapparent. Male genitalia depicted in Figs 81-83. HOSTS. Bovidae and Cervidae (Artiodactyla). COMMENTS. Damalinia (T.) longiceps is included following the statement of Clay & Hopkins (1955) that itresembles D. (T.) spinifer Hopkins 'most closely among known species'. Tricholipeurus has been treated as a genus, synonym (Hopkins, 1943) and subgenus (Hopkins, 1949) ofDamalinia; the history of the variations in status of Tricholipeurus is presented in Table 3. SPECIES INCLUDEDaepycerus (Bedford, 1929) comb. n. from Tricholipeurus (1 cf , 1 $) TRICHODECTID MAMMAL LICE 265 albimarginata (Werneck, 1936) comb. n. from Tricholipeurus (9 cf , 13 $) antidorcus (Bedford, 1931) comb. n. from Tricholipeurus (11 cf, 18 9) bvdfordi (Hill , 1922) comb. n. from Tricholipeurus (2 cf , 2 $ ) clayi( Werneck, 1938) comb. n. from Tricholipeurus (14 cf , 12 $) conectens Hopkins, 1943 comb. rev. from Tricholipeurus (18 cf , 16 $) cornuta cornuta (Gervais, 1844) comb. n. from Tricholipeurus (20 cf , 27 $) cornuta ourebiae Hopkins, 1943 comb. rev. from Tricholipeurus (18 cf , 17 $) dorcephali( Werneck, 1936) comb. n. from Tricholipeurus (2 cf , 2 $) elongata (Bedford, 1934) comb. n. from Tricholipeurus (10 cf , 10 $) indica (Werneck, 1950) comb. n. from Tricholipeurus (65 cf , 60 $) //jieafa (Bedford, 1920) comb. n. from Tricholipeurus (75 cf , 83 $) //peuro/des(Megnin, 1884) comb. n. from Tricholipeurus (106 cf , 91 $) 70iigiceps(Rudbw, 1866) moschatus (Emerson & Price, 1971) comb. n. from Tricholipeurus (holotype cf , disassociated cf head) pakenhami (Werneck, 1947) comb. n. from Tricholipeurus (21 cf , 20 $) parallels (Osborn, 1896) comb. n. from Tricholipeurus (50 cf , 106 $) parkeri (Hopkins, 1941) comb. n. from Tricholipeurus (10 cf , 10 $) spinifer Hopkins, 1943 comb. rev. from Tricholipeurus (17 cf , 16 $) victoriae Hopkins, 1943 comb. rev. from Tricholipeurus (33 Cf , 35 $) EUTRICHOPHILINAE Keler Genus EUTRICHOPHILUS Mjoberg (Figs 87-93) Eutrichophilus Mjoberg, 1910: 71. Type-species: Eutrichophilus cercolabes Mjoberg, by subsequentdesignation (Harrison, 1916a: 21). DESCRIPTION. Anterior of head with osculum present or absent, but pulvinus always attaining margin;dorsal preantennal sulcus absent; clypeal marginal carina with median expansion absent or slight, orpresent as broad or narrow parallel-sided bar with transverse margins convex, straight, or concave;anterolateral margin of head straight or convex; preantennal portion of head long or short; outlinetriangular, rounded or broadly trapezoid. Temple margin convex or with posterolateral angle apparent;temples greatly expanded posteriad (Fig. 87). Male scape expanded, with longitudinal setal row presentand comprising two setae; male flagellomeres fused; female flagellomeres fused or flagellum comprisingtwo flagellomeres; male flagellum very long, with two basally-articulated 'teeth'. Dorsum of head withsetae short or long, sometimes longer anteriorly than posteriorly. Sitophore sclerite unmodified. Thorax with dorsal setae short or of moderate length marginally or submarginally on posterior ofprothorax and pterothorax, absent from disc of both; one or two long setae on posterolateral margins ofpterothorax. Abdomen oval and elongate. Abdominal spiracles present on segments III- VIII. Abdominal setae shortor of moderate length, with tufts of long setae on at least pleurum VIII, sometimes also pleurum VII(males) or IX (females) (Fig. 88); anterior setae present on all pleura but absent from sterna and terga;postero-lateral setae present. Abdominal pleura lacking projections dorsally or ventrally. Sclerites presenton sterna, terga and pleura (where present) of all pre-genital abdominal segments except, sometimes,tergum I, which may be very small; male terga, at least of abdominal segments V and VI, with anterior andposterior sclerites. Gonapophyses frequently large, broadly triangular or rounded, ventral margin lacking lobe but withmore or less dense marginal setae which are long or of moderate length. Gonapophyses meet ventral vulvalmargin acutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; smoothly convex,with or without median indentation or setose projection; subgenital lobe absent. Female terminaliadepicted in Fig. 90. Male subgenital plate with sternites VII and VIII present and fused to s.g.p.r. , IX absent or, if present,fused to s.g.p.r. (Fig. 88). Pseudostyli absent. Male genital opening dorsal, male segment IX posterior.Parameres long or short, narrow or broad; with basiparameral sclerite or flange sometimes present andfused medially, thus linking parameres, but otherwise unfused. Mesomeres present, fused apically to formarch with no median extension; arch smoothly rounded, or with lateral desclerotisations, in which casemedian portion is straight and at right angles to lateral portions, very poorly sclerotised and thin, or absent.Male genitalia depicted in Figs 89, 91, 92. 266 C. H. C. LYAL Figs 87-89 Eutrichophilus species. 87, E. minor, $ head, dorsal. 88, E. setosus, cf abdomen. 89,setosus, cf genitalia. HOSTS. Erethizontidae (Rodentia). SPECIES INCLUDED cerco/aftesMjoberg, 1910 (29 cf , 25 9)comitans Werneck, 1950 (6 cf , 2 $)cordicepsMjoberg, 1910 (23 cf , 27 $)exiguus Werneck, 1950 (holotype cf , allotype $)guyanensis Werneck, 1950 (8 Cf , 7 $)/o&afusEwing, 1936 (5 cf , 8 $)maximus Bedford, 1939 (11 cf , 11 $)mexicamis(Rudow, 1866) (50 cf , 50 $)minor Mjoberg, 1910 (34 cf , 27 $)moojeni Werneck, 1945 (3 cf , 3 })setosus (Giebel, 1874) (102 cf , 102 $) TRICHODECTID MAMMAL LICE 267 91 Figs 90-93 Eutrichophilus species. 90, E. maximus, 9 terminalia. 91, E. guyanensis, cf genitalia. 92, E.guyanensis, cf genitalia, detail. 93, E. moojeni, cf genitalia. DASYONYGINAE Keler Genus CEBIDICOLA Bedford(Figs 94-99) Cebidicola Bedford, 1936: 52. Type-species: Trichodectes armatus Neumann, by original designation.Meganarion Keler, 1938a: 465. Type-species: Trichodectes armatus Neumann, by original designation.[Synonymy by Eichler, 1941.] 268 C. H. C. LYAL Fig. 94 Cebidicola armatus, 9 terminalia. ,/ Fig. 95 Cebidicola armatus, cf abdomen. TRICHODECTID MAMMAL LICE 98 Figs 96-99 Cebidicola species. 96, C. semistriatus , $ head, dorsal. 97, C. armatus, 9 head, dorsal. 98, C.armatus, cf genitalia. 99, C. extrarius, C? genitalia. DESCRIPTION. Anterior of head with osculum present, deep; dorsal preantennal sulcus present or absent;clypeal marginal carina broadened medially into dorsal, posteriorly convex, sclerite; anterolateral marginstraight, slightly convex, concave or sinuate anteriorly, more or less abruptly concave at junction withclypeof rental sulcus, with or without anterior sclerotised projection on either side of osculum (Figs 96, 97);preantennal outline broadly triangular. Temple margin convex or slightly acute and angular posterolateral-ly, with eyes more or less prominent (Figs 96, 97). Male scape expanded, with setal row present andcomprising two or more setae; male flagellomeres fused, with two basally-articulated 'teeth'; femaleflagellomeres fused or unfused. Dorsum of head with setae short or of moderate length, sparse. Sitophoresclerite unmodified. Tarsal claws lacking ventral spines or teeth. Postcoxale absent or present, not greatly developed.Thoracic setae present dorsally only along posterior and posterolateral margins of pterothorax; setae shortmedially, longer laterally. Atrium of thoracic spiracle tubular or conical. Abdomen oval, sometimes tapering posteriorly more in male than female. Abdominal spiracles presenton segments III- VIII. Abdominal setae short or of moderate length, longest on pleura VI- VIII; anteriorsetae sometimes present on pleura and laterally on sterna and terga; postero-lateral setae present,sometimes numbering more than one per site. Pleural projection present ventrally on abdominal pleurumIV, large, sclerotised. Sclerites present on all abdominal pleura, on at least abdominal terga II to VIII andat least abdominal sterna V to VII; male terga, at least on segments V-VII, with anterior and posteriorsclerites (Fig. 95). Gonapophyses broad, especially medially, though lobe absent; marginal setae long, densely crowded.Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval marginnot sclerotised; sometimes expanded, otherwise smoothly convex (Fig. 94); bilobed median spinoseprojection may be present, but subgenital lobe absent. Median longitudinal sclerite sometimes present onfemale sternum VIII (Fig. 94). Male subgenital plate with sternite VII fused to s.g.p.r. , VIII fused or not fused to s.g.p.r. , and IX absent 270 C. H. C. LYAL or, if present, not fused to s.g.p.r. Pseudostyli absent or, if present, small, slender, incurved (Fig. 95). Malegenital opening postero-dorsal or dorsal. Parameres fused or unfused; basiparameral sclerites present,fused. Mesomeres present, fused or unfused and, if fused, median extension absent; mesomeral archproduced basally between b.a.l.s. to contact parameres, which do not meet b.a.l.s. Male genitalia depictedin Figs 98, 99. HOSTS. Cebidae (Primates). SPECIES INCLUDED armatus (Neumann, 1913) (3 cf , 4 $)exfrariusWerneck, 1950 (21 cf , 13 $)semiarmatus (Neumann, 1913) (12 (J, 12 <J>) Genus PROCA VICOLA BedfordThe genus Procavicola comprises two subgenera. DESCRIPTION. Anterior of head with osculum present, semicircular; dorsal preantennal sulcus presentclypeal marginal carina broadened medially into dorsal, posteriorly convex, sclerite (Fig. 106); conuslarge, as long as female scape; anterolateral margin of head straight or convex anteriorly, more or lessabruptly concave at junction with clypeofrontal sulcus; preantennal outline broadly triangular. Templemargin smoothly convex, sometimes produced posteriad, or with postero-lateral angle developed intoposteriorly-projecting triangular or rounded process*. Male scape expanded, with setal row represente(by two setae only; male flagellomeres fused, with two basally-articulated 'teeth'; female flagellomeresunfused and closely associated, fused to two closely associated annulations, or completely fused*. Dorsumof head with setae short, sparse; anterior margin of head with setae longer than on disc. Sitophore scleriteunmodified. Tarsal claws lacking ventral spines or teeth. Postcoxale of metathoracic leg absent or present; if presentmay be well developed, but not to the same degree as described for Procaviphilus (Meganarionoides) amnot fused to abdominal pleurite II. Thoracic setae present dorsally only along posterior and lateroposterior margins of prothorax and pterothorax; setae short, except for laterally on pterothorax, where omoderate length. Abdomen oval, more or less elongate (Figs 103, 105). Abdominal spiracles present on segmentsIII- VIII, all approximately the same size. Abdominal setae short or of moderate length; anterior setaepresent on pleura only; postero-lateral setae present, sometimes numbering more than one per site*Pleural projections present dorsally and ventrally on abdominal pleurum IV, sclerotised. Sclerites presenon sterna, terga and pleura of all abdominal segments except, occasionally, tergum I* ; male terga, at leasof segments IV- VI, with anterior and posterior sclerites; second abdominal sternum with broadheavily-sclerotised apophysis underlying sternite, articulating with median extensions of abdominapleurite II (Figs 103, 105). Gonapophyses broad, lacking lobe; marginal setae lacking tubercles, occasionally on small conicaprotuberances (Figs 101, 102). Gonapophyses meet ventral vulval margin acutely, not linked by sclerotisecband. Ventral vulval margin not sclerotised; expanded, sometimes W-shaped medially (Fig. 100),sometimes broadened posteriorly (Fig. 101), sometimes contracted, shorter than length of gonapophyses(Fig. 102)*. Male subgenital plate with sternites VII and VIII present and fused to s.g.p.r., sternite IX absent or, ifpresent, fused to s.g.p.r. and perisetal gaps small*; s.g.p.r. not always attaining posterior margin ofsegment IX*. Pseudostyli absent (Figs 103-105). Male genital opening postero-dorsal. Male genitalia veryvariable* ; parameres not fused, basiparameral sclerites present or absent* , mesomeres fused or unfused*. HOSTS. Procaviidae (Hyracoidea). Subgenus PROCA VICOLA Bedford (Figs 100, 104-106, 109)Procavicola Bedford, 1932: 711. Type-species: Trichodectes sternatus Bedford, by original designation. DESCRIPTION. Temple margin smoothly convex, sometimes projecting posteriorly, but never with posteriormembranous or lightly-sclerotised process. Postero-lateral setae present, single at each site. Sclerites present on sterna, terga and pleura of allabdominal segments except segment I, where tergal sclerite absent. TRICHODECTID MAMMAL LICE 271 100 Figs 100-102 Procavicola species, $ terminalia, ventral. 100, P. (P.) natalensis. 101, P. (Condylocepha-lus) lindfieldi. 102, P. (C.) dissimilis. Ventral vulval margin expanded as described, much broader than length of gonapophyses (Fig. 100). Male subgenital plate with sternites VII and VIII present and fused to s.g.p.r., sternite IX absent;s.g.p.r. may be very slender, may not attain posterior margin of segment IX (Figs 104, 105). Parameresunfused, more or less narrow, rod-like, sometimes asymmetrically curved. Basiparameral scleritesfrequently present, fused or separate. Mesomeres not fused, short. Endophallus lacking large, hook-likesclerites. Male genitalia depicted in Fig. 109. HOSTS. Procaviidae (Hyracoidea). SPECIES INCLUDEDaffijMsWerneck, 1941 (10 C?, 13 $) 272 C. H. C. LYAL 106 Figs 103-106 Procavicola species. 103, P. (Condylocephalus) dissimilis , C? abdomen. 104, P. (P.)vicinus, C? terminalia. 105, P. (P.) eichleri, cf abdomen. 106, P. (P.) natalensis, $ head, dorsal. TRICHODECTID MAMMAL LICE 273 107 108 109 Figs 107-109 Procavicola species, cf genitalia. 107, P. (Condylocephalus) dissimilis. 108, -P. (C.)dissimilis, detail of right paramere and mesomere at junction with basal apodeme. 109, P. (P.)pretoriensis. brucei Werneck, 1941 (43 cf , 48 $)eichleri Werneck, 1941 (45 cf , 56 $)emarginatus (Bedford, 1928) (16 5, 19 $>)furca Bedford, 1939 (1 cf)heterohyracis Bedford, 1932 (3 cf , 1 $)lopesi lopesi Bedford, 1939 (32 cf , 40 $)lopesi vicinus Werneck, 1941 (8 cf , 10 9)mokeetsi Bedford, 1939 (20 cf , 15 $)natalensis Bedford, 1932 (23 cf , 12 $)parvus Bedford, 1932 (8 Cf , 6 $)pretoriensis Bedford, 1932 (23 Cf , 21 $)sAoanusMaltbaek, 1937sternal us (Bedford, 1928) (9 cf, 8 ?)subparvus Bedford, 1932 (6 cf , 9 $)thorntoni Hopkins, 1942 (18 cf , 22 $)ugandensis Werneck, 1941 (17 cf , 18 $) 274 C. H. C. LYAL Subgenus CONDYLOCEPHALUS Werneck (Figs 101-103, 107, 108) Condylocephalus Werneck, 1941: 497 [as subgenus of Procavicola Bedford]. Type-species: Procavicola(Condylocephalus) bedfordi Werneck, by original designation. DESCRIPTION. Temple margin convex, with posteriorly-projecting membranous or lightly-sclerotisedprocess, more apparent in male than female, triangular or as small rounded bump (linfieldi females).Female flagellomeres unfused, but closely associated. Abdomen with postero-lateral setae present, frequently doubled, trebled or numerous at each site.Sclerites present on sterna, terga and pleura of all abdominal segments. Ventral vulval margin expanded, sometimes broadened posteriorly and broader than length of gona-pophyses (Fig; 101), otherwise narrower, width less than length of gonapophyses (Fig. 102). Male subgenital plate with sternites VII, VIII and IX present and fused to s.g.p.r., with perisetal gapssmall (Fig. 103). Parameres unfused, curved, with anterolateral projections, not asymmetric. Basipar-ameral sclerites present, fused or unfused. Mesomeres fused apically; mesomeral arch with medianextension and lateral double flexion (Figs 107, 108). Endophallus ornamented with large, hook-likesclerites (Fig. 107). HOSTS. Procaviidae (Hyracoidea). COMMENTS. Though Condylocephalus has been treated by most authors as a subgenus of Procavicola,Eichler (1963) considered it to have full generic status. SPECIES INCLUDEDbedfordi Werneck, 1941 (2 cf , 1 $)dissimilis Werneck, 1941 (64 cf , 58 $)hopkinsi Werneck, 1941 (24 cT, 29 $)JbMlfieW(Hill, 1922) (77 cT, 65 $)univirgatus (Neumann, 1913) (33 cf , 32 $) Genus PROCAVIPHILUS BedfordThe genus Procaviphilus comprises two subgenera. DESCRIPTION. Surface of head, thorax and abdomen frequently covered with clearly- visible scales orsclerotised nodules.Anterior of head variable, one of two types: 'procaviphilus' or 'procavicola'*. (a) 'procaviphilus' type. Osculum absent or, if present, slightly concave only; dorsal preantennal sulcusabsent; clypeal marginal carina broadened medially into straight or slightly curved bar (Fig. 112); conussmall, not as long as female scape; anterolateral margin of head straight or convex anteriorly, no abruptconcavity at junction with clypeofrontal sulcus; preantennal outline trapezoid (Fig. 112). (b) 'procavicola' type. Osculum present, semicircular; dorsal preantennal sulcus present or absent; clypealmarginal carina broadened medially into dorsal, posteriorly convex, sclerite (Fig. 106); conus large, as longas female scape; anterolateral margin of head convex anteriorly, more or less abruptly concave at junctionwith clypeofrontal sulcus; preantennal outline broadly triangular. Temple margin smoothly convex, more or less projecting posteriorly. Male scape expanded, with setalrow represented by two setae only; male flagellomeres fused, with two basally-articulated 'teeth'; femaleflagellomeres unfused , though sometimes very closely associated . Dorsum of head with setae short , sparse .Sitophore sclerite unmodified. Tarsal claws lacking ventral spines or teeth. Postcoxale of leg III absent or, if present, frequentlyenlarged, heavily sclerotised, displaced posteriad to occupy abdominal sternum II, and fused to sclerite ofabdominal pleurum II, in which case gap between postcoxales sometimes obscured by sternite II*.Thoracic setae present dorsally only along posterior and lateroposterior margins of prothorax andpterothorax; setae short, except for laterally on pterothorax, where of moderate length. Abdomen oval-elongate. Abdominal spiracles present on segments III to VIII, though sometimes verysmall and possibly non-functional on VIII*. Abdominal setae short or of moderate length; anterior setaepresent on pleura only, sparse; postero-lateral seta present. Pleural projections present ventrally anddorsally on abdominal pleurum IV, sclerotised. Sclerites, frequently faint, present on sterna, terga andpleura of all abdominal segments except I; male terga, especially tergum VI, frequently with anterior andposterior sclerites. Gonapophyses with setae non-tuberculate and, frequently, tuberculate; setal tubercles, if present, TRICHODECTID MAMMAL LICE 275 1 10 112 1 13 Figs 110-113 Procaviphilus species. 110, P. (Meganarionoides) n. baculatus, $ terminalia, ventral. Ill,P. (M.) n. baculatus, $ gonapophysis, ventral. 112, P. (P.)f. granuloides, $ head, dorsal. 113, P. (M.)scutifer, $ terminalia, ventral. sometimes fused characteristically*; lobe absent or, if apparently present, formed of fused tubercles andthick, with submarginal setae (Fig. 111). Gonapophyses meet ventral vulval margin acutely, not linked bysclerotised band. Ventral vulval margin not sclerotised; expanded, frequently with posterior broadening(Fig. 110), sometimes lengthened (Fig. 113) or with median lobulate process (Fig. 110)*. Male subgenital plate with sternites VII, VIII and IX present and fused to s.g.p.r., perisetal gaps verysmall or absent, rarely large (Figs 114, 115). Pseudostyli absent. Male genital opening posterodorsal. Malegenitalia very variable: basal apodeme short or long with median constriction*; parameres unfused,frequently with basal flange, or fused, faintly or clearly, and parameral plate apically bifurcate. Basipar-ameral sclerites absent. Mesomeres fused, basally extending between b.a.l.s. to contact parameres, which 276 C. H. C. LYAL TRICHODECTID MAMMAL LICE 277 116 120 Figs 116-122 Procaviphilus species, cf genitalia. 116, P. (Meganarionoides) angolensis. 117, P. (P.) f.granuloides. 118, P. (P.) dubius. 119, P. (M.) n. neumanni. 120, P. (M.) jordani. 121, P. (M.)jordani,detail of junction of mesomere, paramere and basal apodeme. 122, P. (M.) serraticus. 278 C. H. C. LYAL may also contact b.a.l.s. (Figs 116, 117, 119-121); mesomeral arch with median extension more or lessbroad, lateral desclerotisations sometimes apparent (Figs 118, 122)*. HOSTS. Procaviidae (Hyracoidea) and Cercopithecidae (Primates). Subgenus PROCAVIPHILUS Bedford (Figs 112, 117, 118)Procaviphilus Bedford, 1932: 725. Type-species: Procaviphilus ferrisi Bedford, by original designation. DESCRIPTION. Anterior of head of 'procaviphilus' type. Postcoxale absent or, if present, not greatly developed and not fused to abdominal pleurite II. Abdominal spiracles all the same size. Gonapophyses with setal tubercles present but not fused; gonapophyses not apically truncate. Ventralvulval margin expanded, sometimes W-shaped medially, not broadened posteriorly. Male genitalia with basal apodeme short, attaining abdominal segment VII or VI, lacking medianconstriction; mesomeral arch with lateral desclerotisations. HOSTS. Procaviidae (Hyracoidea). SPECIES INCLUDED duWus Werneck, 1941 (10 cf , 10 $)ferrisi ferrisi Bedford, 1932 (12 cf , 5 $)ferrisi granuloides Bedford, 1939 (26 C? , 19 9)ferrisi Wndei Werneck, 1946 (3 cf , 4 $)granulates (Ferris, 1930) (13 cf , 17 $)Aarrisi Werneck, 1946 (15 cf , 19 $)robertsi (Bedford, 1928) (16 cf , 18 $) Subgenus MEGANARIONOIDESEichler (Figs 110, 111, 113-116, 119-122) Meganarionoides Eichler, 1940: 159. Type-species: Trichodectes colobi Kellogg, by original designation.Acondylocephalus Werneck, 1941: 478 [as subgenus of Procavicola Bedford]. Type-species: Trichodectescongoensis Ferris, by original designation. [Synonymy by Werneck, 1946: 85.] DESCRIPTION. Anterior of head of 'procaviphilus' type or, more frequently, of 'procavicola' type. Postcoxale of leg III enlarged as described in description of Procaviphilus s.l., and fused to abdominalpleurite II, at least in female. Gonapophyses with setal tubercles absent (in which case gonapophyses characteristically broad andvulval margin produced posteriad as in Fig. 113), or present basally, marginally and submarginally, andfused characteristically to form basal process (Fig. Ill); gonapophyses more or less truncate. Ventralvulval margin expanded, as described for Procaviphilus s. str. or, more frequently, broadened posteriorly,sometimes produced posteriad (Fig. 113) or with median lobulate process (Fig. 110). Male genitalia with basal apodeme attaining abdominal segment VII or VI or, more frequently, long,attaining segment III or II, with median constriction (Figs 116, 119); mesomeral arch with or withoutlateral desclerotisation. HOSTS. Procaviidae (Hyracoidea) and Cercopithecidae (Primates). COMMENTS. There has been some disagreement in the literature over the correct host of one species in thissubgenus. Most species included in P. (Meganarionoides) are parasites of Procaviidae, as are all otherspecies in the subfamily Dasyonyginae (other than the three species of Cebidicola, which are included inthe subfamily for the first time in this study). One species, however, P. (M.) colobi (Kellogg, 1910), wasdescribed from the monkey Colobus guereza caudatus Thomas. Keler (1938a) included this species withthe others described from Primates in his genus Meganarion (an objective synonym of Cebidicola},although realising that the species were not truly congeneric. Eichler (1940) described the new genusMeganarionoides for colobi, and placed it with Cebidicola and Lorisicola in the new subfamily Cebidicoli-nae. Werneck (1946) recognised the identity of colobi with the hyrax lice, and synonymised Meganar-ionoides with Procavicola (Acondylocephalus) Werneck, 1941, the subgenus thus taking the nameProcavicola (Meganarionoides). Werneck (1946) also suggested that Colobus was not the true host of P.colobi, but that the louse was probably a parasite of Dendrohydrax validus subsp. He suggested that thehost record of the type-specimens was erroneous and due to mislabelling (the collection having included TRICHODECTID MAMMAL LICE 279 both Colobus and Dendrohyrax), and that a second record was due to contamination (other hyrax-licehaving been associated with the specimens of colobi). Hopkins (1949) reported having examined 25 skinsof Colobus polykomos, which he identified as the 'supposed host', without having found any Trichodecti-dae, and agreed with Werneck (1946) that Dendrohyrax validus subsp. was the correct host. Hopkins &Clay (1952) also identified D. validus subsp. as the host, the record from 'Colobus caudatus' being termedan 'error'. Eichler (1963) agreed, and removed Meganarionoides from the Cebidicolinae and placed it inthe Dasyonygidae with Procavicola (Fig. 38). Emerson & Price (1981) include P. colobi, withoutcomment, as a parasite of Dendrohyrax validus validus, although the association with the nominatesubspecies of this animal has not appeared elsewhere in the literature. Kuhn & Ludwig (1964), however,reported a specimen of Colobus guereza with 'hundreds of eggs and adult and larval Procavicola on it, allclasping the hairs tightly; most of them on the back and on the throat', and were able to state that themonkey had not been in contact with a Dendrohyrax or any other Procaviidae after its death. Theyconcluded There is no doubt . . . that Colobus guereza is a natural host of Procavicola (Meganarionoides)colobi.' In view of the fact that there are now three records of the species from Colobus guereza and nonefrom any member of the Procaviidae, this conclusion seems fully justified. Meganarionoides was, as described above, treated as a subgenus of Procavicola by Werneck (1946). Inthis he has been followed by most authors, although Eichler (1963) considered it to be a full genus. Beforethe present study Meganarionoides had not been placed as a subgenus of Procaviphilus. SPECIES INCLUDED a/rica/ius( Werneck, 1941) comb. n. from Procavicola (5 cf , 5 $)angolensis (Bedford, 1936) comb. n. from Procavicola (8 cf , 11 9)colobi (Kellogg, 1910) comb. n. from Procavicola (1 d", 1 $)congoensis (Ferris, 1930) comb. n. from Procavicola (22 cf , 25 $)jordani (Bedford, 1936) comb. n. from Procavicola (2 cf , 2 $)muesebecki (Emerson & Price, 1969) comb. n. from Procavicola (10 cf , 10 $)/it'i/niann/neuma/in/(Stobbe, 1913) comb. n. from Procavicola (2 cf , 1 $)neumanni baculatus (Ferris, 1930) comb. n. from Procavicola (13 cf , 14 $)sclerotis sclerotis Bedford, 1932 [treated as Procaviphilus s. str. by previous authors] (10 cf , 17 $)sclerotis major Maltbaek, 1937 [treated as Procaviphilus s. str. by previous authors]scutifer (Werneck, 1941) comb. n. from Procavicola (14 cf , 19 $)serraticus (Hill, 1922) [treated as Procaviphilus s. str. by previous authors] (50 cf , 70 $)tendeiroi (Emerson, 1965) comb. n. from Procavicola (3 Cf , 2 $) Genus DASYONYX BedfordThe genus Dasyonyx. comprises two subgenera. DESCRIPTION. Anterior of head with osculum present, variable in degree of excavation* ; dorsal preantennalsulcus absent; clypeal marginal carina broadened medially, with posterior curvature of broadened portionsimilar to curvature of osculum* (Fig. 123); conus not large; anterolateral margin of head straight, convexor concave, though not very concave at junction with clypeofrontal suture; preantennal portion of headshort or long, outline broadly triangular, trapezoidal or rounded*. Temple margin shallowly convex,sometimes with small rounded projection postero-laterally*. Male scape expanded, with setal row present;male flagellomeres fused, with two basally-articulated 'teeth'; female flagellomeres unfused. Dorsum ofhead with setae of moderate length. Sitophore sclerite with posterior arms extended (cf. Fig. 12), thoughsclerite difficult to see. Tarsal claws with ventral spines or teeth (Figs 14, 15)*. Post-coxale of metathoracic leg absent or, ifpresent* enlarged, though not as described for Procaviphilus (Meganarionoides) and not fused toabdominal pleurite II. Thorax with dorsal setae present only posteriorly on prothoracic margin andposteriorly and posterolaterally on pterothoracic margin; setae short anteriorly, longer posteriorly with thelongest setae on the posterolateral margins of the pterothorax; setae generally sparse. Abdomen broadly oval, with male segment IX not projecting greatly (Fig. 129). Abdominal spiraclespresent on segments III to VIII, all approximately the same size, frequently inconspicuous. Abdominalsetae of moderate length; anterior setae present on pleura only; postero-lateral setae present. Pleuralprojections present dorsally and ventrally on abdominal pleurum IV, sclerotised. Sclerites present onsterna, terga and pleura (where present) of all abdominal segments except I; male terga, at least ofabdominal segment VI, with anterior and posterior sclerites. Gonapophyses with sparse marginal setae and variably-developed lobe ventrally, the lobe bearing twoapical or subapical setae and frequently being serrate along dorsal (posterior) margin (Fig. 126). 280 C. H. C. LYAL TRICHODECTID MAMMAL LICE 281 282 C. H. C. LYAL Fig. 129 Dasyonyx (D.) validus, C? abdomen. Gonapophyses meet ventral vulval margin acutely, not linked by sclerotised band. Ventral vulval marginnot sclerotised; greatly expanded, sometimes with postero-lateral angular projections (Fig. 126); subgenit-al lobe absent. Male subgenital plate with sternites VII, VIII and IX present and fused to s.g.p.r. , but variably modified(Figs 124, 125, 127-129), frequently lacking perisetal gap. Pseudostyli absent. Male genital openingpostero-dorsal or dorsal. Parameres fused or, if unfused, then with basal flanges (Figs 130, 132, 134, 135).Basiparameral sclerites absent. Mesomeres fused; mesomeral arch with median extension and lateraldesclerotisations; mesomeres more or less produced basally between b.a.l.s. to meet parameres, whichsometimes do not contact b.a.l.s. (Figs 130, 132-135). HOSTS. Procaviidae (Hyracoidea). Subgenus DASYONYX Bedford(Figs 14, 123-127, 129-132, 134, 135)Dasyonyx Bedford, 19326: 720. Type-species: Dasyonyx validus Bedford, by original designation. TRICHODECTID MAMMAL LICE 283 Figs 130-135 Dasyonyx species, C? genitalia. 130, D. (D.) validus. 131, D. (D.) validus, endophallus,sclerites shown by dashed outlines. 132, D. (D.) ovalis, 133, D. (Neodasyonyx) ruficeps. 134, D. (D.)guineensis. 135, D. (D.) minor. DESCRIPTION. Osculum deeply concave; preantennal outline of male head subtriangular or subtrapezoidal.Temple margin frequently with small rounded projection postero-laterally (Fig. 123). Tarsal claws with ventral spines slender and sharp (Fig. 14). Postcoxale of metathoracic leg generallypresent and enlarged, though not as described for Procaviphilus (Meganarionoides). HOSTS. Procaviidae (Hyracoidea). 284 C. H. C. LYAL SPECIES INCLUDED bedfordi Werneck, 1945 (11 cT, 16 $)dendrohyracis (Ferris, 1930) (12 cf , 15 $)guineensis Werneck, 1941 (5 cf , 3 $)hopkinsi Werneck, 1941 (33 cf , 32 $)minor Bedford, 1939 (3 cf , 1 $)oculatus Bedford, 1928ovalis Bedford, 1932 (36 cf , 35 $)smallwoodae Emerson & Price, 1969 (7 cf , 3 $)validus validus Bedford, 1932 (22 cf , 22 $)validus ugandensis Werneck, 1941 (29 cf , 26 <j>) Subgenus NEODASYONYX Werneck(Figs 15, 128, 133) Neodosyonyx Werneck, 1941: 543 [as subgenus of Dosyonyx Bedford]. Type-species: Dosyonyx trans-vaalensis Bedford, by original designation. DESCRIPTION. Osculum shallowly concave; preantennal region of male head short, outline rounded.Temple margin lacking postero-lateral projection.Tarsal claws with broad ventral teeth (Fig. 15). Postcoxale of metathoracic leg absent. HOSTS. Procaviidae (Hyracoidea). SPECIES INCLUDED capensis Emerson, 1965 (holotype cf , allotype $)d/acantfii/s(Ehrenberg, 1828) (9 cf, 8 $)nairobiensis Bedford, 1936 (89 cf , 92 $)ruficeps Emerson, 1964 (15 cf , 13 $)transvaalensis Bedford, 1932 (24 cf , 31 $)waterburgensis Bedford, 1932 (5 cf , 6 $) Genus EURYTRICHODECTES Stobbe (Figs 9, 16, 136-138)Eurytrichodectes Stobbe, 1913o: 111. Type-species: Eurytrichodectes paradoxus Stobbe, by monotypy. DESCRIPTION. Anterior of head with osculum absent or, if present, very shallowly concave; dorsalpreantennal sulcus absent; clypeal marginal carina slender, not greatly developed medially or, if de-veloped, in the form of a median posteriorly-directed narrow-based triangle; anterolateral margin of headslightly sinuate; preantennal portion of head very short, outline broadly triangular or trapezoid. Templemargin produced posteriorly into broad triangular spike, almost as long as prothorax or, if shorter,attaining front of pronotum (Fig. 138). Male scape expanded, with longitudinal setal row comprising onlytwo setae; male flagellomeres fused, though semicircular sclerite at apex may be vestige of terminal'flagellomere; two basally-fused 'teeth' present on male flagellum; female flagellomeres unfused; membra-nous projection present on female pedicel (Fig. 138). Dorsum of head with setae short, sparse. Sitophoresclerite with posterior arms extended (cf. Fig. 12), though sclerite difficult to see. Tarsal claws ridged ventrally, lacking teeth or spines (Fig. 16). Postcoxale of metathoracic leg absent.Thorax with dorsal setae present on posterior margin of prothorax and on posterior margin of pterothorax;setae short, sparse. Abdomen broad and oval, sometimes with male terminal segments tapering and projecting slightly.Abdominal spiracles present on segments III to VIII, all approximately of the same size. Abdominal setaeshort on sterna IV to IX, terga and pleura, some tergal setae very short; sterna II and III with stout, conicalsetae (Fig. 136); anterior setae present, sparse on pleura; setal row on terga and sterna may be irregular;postero-lateral setae present. Pleural projections present dorsally and ventrally on abdominal pleurum IV,sclerotised, very long, reaching or almost reaching posterior margin of pleurum V. Sclerites present onsterna, terga and pleura (where present) of all abdominal segments except, sometimes, tergum I; maleterga, at least of abdominal segments II to VI, with anterior and posterior sclerites. Gonapophyses with ventral marginal setae present, each with a small, conical tubercle; ventral lobeabsent. Gonapophyses meet ventral vulval margin smoothly, not linked by sclerotised band. Ventral vulvalmargin not sclerotised; medially expanded and trapezoid; subgenital lobe absent. TRICHODECTID MAMMAL LICE 285 ig. 136 Eurytrichodectes paradoxus , cf abdomen. Male subgenital plate with sternites VII, VIII and IX present and fused to s.g.p.r. , with setal gaps verysmall or absent (Fig. 136). Pseudostyli absent. Male genital opening dorsal. Parameres not fused,sometimes apically bifurcate. Basiparameral sclerites absent. Mesomeres fused; mesomeral arch withmedian extension (see comments below); mesomeral arch lacking lateral desclerotisations; mesomeressometimes produced basally between b.a.l.s. to contact parameres. Male genitalia depicted in Fig. 137. HOSTS. Procaviidae (Hyracoidea). COMMENTS. The illustration of the the mesomeral arch of E. paradoxus is reproduced upside-down inWerneck (1941: 452). SPECIES INCLUDED machadoi Werneck, 1958 (holotype cf , allotype $)paradoxus Stobbe, 1913 (29 cf , 28 $) 286 C. H. C. LYAL 138 137 Figs 137, 138 Eurytrichodectesparadoxus. 137, cf genitalia. 138, $ head, dorsal. TRICHODECTINAE Kellogg Genus PROTELICOLA Bedford gen. rev. (Figs 139, 140)Protelicola Bedford, 1932a: 355. Type-species: Protelicola intermedius Bedford, by monotypy. DESCRIPTION. Anterior of head with osculum present; clypeal marginal carina broadened just laterally toosculum, tapering medially and interrupted by dorsal preantennal sulcus; anterolateral margin of headstraight or convex; preantennal outline of head triangular or rounded. Temple margin convex. Male scapeslightly expanded, with longitudinal setal row present and comprising three or four setae; flagellomeresfused in males and females; male flagellum with two basally-articulated 'teeth'. Dorsal setae of head shortor of moderate length, longest and most abundant anteriorly. Sitophore sclerite unmodified. Thorax with setae long or of moderate length laterally and dorsally, with setae on postero-lateral anglesof pterothorax shorter, more spine-like. Prothorax with setae present sparsely on lateral and posteriormargins; median gap present, wide; single seta present antero-laterally on disc. Pterothorax with setaepresent postero-laterally and submarginally posteriorly; median gap present, wide; no setae present ondisc. Abdomen rounded, similar in shape in males and females. Abdominal spiracles present on segmentsIII- VIII. Abdominal setae of moderate length; anterior setae present on pleura only; postero-lateral setaeabsent. Abdominal pleura lacking projections. Abdominal segments with tergal sclerites absent excepttergite IX in female; pleural sclerites present on pleura II and III; sternites IV- VII present in male andV-VII present in female, very slender and difficult to see in both sexes. Gonapophyses with basal setae and rectangular lobe on ventral margin, lobe formed from more or lessfused setal tubercles. Gonapophyses meet ventral vulval margin smoothly, not linked by sclerotised band.Ventral vulval margin not sclerotised. Subgenital lobe present, not marginally serrate, but sometimes withapical papillae. Female genital chamber with small lapped scales on walls, dorsal wall with medianlongitudinal anterior 'slit' where scales are absent. Male subgenital plate comprising very slender sternite VII and s.g.p.r. only, s.g.p.r. not reachingposterior of sternum IX. Male genital opening dorsal or postero-dorsal. Pseudostyli absent. Basalapodeme with b.a.l.s. widely divergent anteriorly. Parameres long, slender, fused basally, projectinganteriorly between b.a.l.s. (but see second paragraph of 'comments' below). Basiparameral sclerites TRICHODECTID MAMMAL LICE 287 139 Figs 139, 140 Protelicola species, C? genitalia. 139, P. hyaenae. 140, undescribed species. absent. Mesomeres fused apically to form arch, with median projection present, broad (see second andthird paragraphs of 'comments' below). Male genitalia depicted in Fig. 139.Alimentary canal with numerous small spines in crop. HOSTS. Hyaenidae and Protelidae (Carnivora). COMMENTS. Protelicola was treated as a subgenus and a junior synonym of Felicola by Hopkins (1949) andWerneck (1948) respectively; its most recent placement (Emerson & Price, 1981) was as a synonym ofFelicola. A more detailed history of the variations in status of Protelicola is presented in Table 5. In the British Museum (Natural History) collection there is a slide bearing 1 nymphal, 3 female and 2male (one of which is teneral) lice from Proteles cristatus termes. Hopkins has identified the lice asProtelicola intermedius. The females are indistinguishable from females of P. intermedius s. str., but themale genitalia are very distinct, with the parameres completely fused to form a broad plate, the mesomeralarch wide, and the b.a.l.s. with a small postero-lateral projection contacting the mesomeres (Fig. 140). Inall other respects the males resemble P. intermedius s. str. It seems that the males represent a new species,differing from Protelicola intermedius by the structure of the genitalia, but the identity of the females isdoubtful. In view of the limited number of specimens available, the species is not formally described. Bedford (19320) described P. intermedius from Proteles cristatus, the first louse known from a hyaena.Hopkins (1960) described P. intermedius hyaenae (in Felicola) from Hyaena brunnea, distinguishing itfrom the nominate subspecies by the size, the outline of the preantennal portion of the head, and the male 288 C. H. C. LYAL genitalia. Ledger (1980) raised P. i. hyaenae to specific rank although Emerson & Price (1981) retained itssubspecific status. Hopkins (1960) indicated three features of the male genitalia in which the two taxadiffer: the greater anterior divergence of the b.a.l.s. in P. i. intermedius , the shape of the parameres (whichhe presumed to be completely fused in P. i. intermedius, and are only basally fused in P. i. hyaenae), andthe absence of the mesomeral arch extension in P. i. intermedius. The type-series of P. i. intermedius hasnot been seen in the present study, but a series of specimens from Proteles cristatus has been examined;these specimens agree with Bedford's and Hopkins' descriptions of P. i. intermedius in all but details of themale genitalia. The male genitalia of this series were found to be very similar to those of P. hyaenae, withthe b.a.l.s. variably divergent anteriorly, the parameres fused only basally and the mesomeral arch with amedian broad extension (although this is very thinly sclerotised and difficult to see in both taxa). The malegenitalia of P. hyaenae are depicted in Fig. 139. The two species may be distinguished by the smaller sizeand shorter preantennal region of P. hyaenae (see photographs in Hopkins, 1960). SPECIES INCLUDED hyaenae (Hopkins, 1960) stat. rev., comb. n. from Felicola (holotype cf , allotype $)intermedius Bedford, 1932 comb. rev. from Felicola (19 cf, 48 9) Genus LUTRIDIA Keler (Figs 141-145)Lutridia Keler, 1938a: 433. Type-species: Trichodectes exilis Giebel, by original designation. DESCRIPTION. Anterior of head with osculum absent; dorsal preantennal sulcus absent; clypeal marginalcarina broadened medially to form dorsal sclerite with three posteriorly-directed projections (Fig. 142);preantennal portion of head short, outline smoothly rounded. Temple margin convex or rectangular. Malescape not expanded, longitudinal setal row comprising two setae positioned distally on segment; flagello-meres fused in males and females; male flagellum lacking 'teeth'. Dorsum of head with setae short or ofmoderate length, sparse. Sitophore sclerite unmodified. Thorax with dorsal setae long or of moderate length, limited to posterior and postero-lateral margins ofprothorax and pterothorax. Abdomen oval or slightly elongate-oval, with male segment IX projecting slightly posteriorly (Fig. 141).Abdominal spiracles present on segments III- VIII. Abdominal setae as follows: pleurum II with setaesparse, stout and short, anterior setae and p.s.r. present; pleurum III with setae short and of moderatelength, stout, very sparse (exilis) or with p.s.r. present (matschiei); pleura IV- VII or VIII lacking setae,those on VIII, if present, very small and posteriorly positioned; sternal setae short and stout or longer(about two-thirds length of segment); sternum II with median gap of posterior setal row small or absent;sterna III-IV and VII- VIII (matschiei) or III- VIII (exilis) with setae very sparse, setae absent from sternaV-VI of L. matschiei; terga I-IV (males) or I-III (females) with median setal group including one or twosetae as long as segment, other terga with median group absent or comprising shorter setae; terga withlateral seta or setae generally present, of moderate length on terga II-III, shorter on more posteriorsegments (these setae may represent either lateral setal group or postero-lateral seta); anterior setaepresent only on pleurum II. Abdominal pleura lacking projections. Sclerites absent from abdominalpleura, present on terga IV- VIII (males) or terga IV-IX (females) and sterna III- VIII (males) or sternaIV- VIII (females); sclerites frequently very faint, may not be seen; male terga lacking posterior sclerites. Gonapophyses with non-tuberculate setae on ventral margin; ventral lobe not present. Gonapophysesmeet ventral vulval margin smoothly, linked by sclerotised band. Ventral vulval margin sclerotised, withchord at 90 degrees to long axis of abdomen, submarginal non-tuberculate setae present. Subgenital lobepresent, small, rectangular, sometimes serrate along posterior margin (Fig. 143). Male subgenital plate comprising s.g.p.r. only or with s.g.p.r. linked by broad sternite VIII (and possiblyVII); in either case s.g.p.r. not attaining segment IX. Pseudostyli absent. Male genital opening dorsal.Basal apodeme slender, long, attaining at least abdominal segment III. Parameres long, slender, basallyfused; basal fused portion may be partially detached from rest of parameres; parameres sometimes fused tob.a.l.s. Mesomeres absent. Male genitalia depicted in Figs 144, 145. HOSTS. Lutrinae (Carnivora: Mustelidae). COMMENTS. Lutridia has been treated as a synonym and a subgenus of Trichodectes (Hopkins, 1942 andHopkins, 1949 respectively), though the most recent treatment (Emerson & Price, 1981) consideredLutridia as, a full genus. The history of the variations in status of Lutridia is presented in Table 4. TRICHODECTID MAMMAL LICE 289 141 143 142 144 145 Figs 141-145 Lutridia species. 141, L. exilis, cf abdomen. 142, L. matschiei, $ head, dorsal. 143, L.matschiei, $ terminalia. 144, L. matschiei, cf genitalia. 145, L. exilis, cf genitalia. SPECIES INCLUDEDexi//s(Nitzsch, 1861) (3 cf , 10 $)matec/iJd(Stobbe, 1913) (36 cf , 38 $) Genus NEOLUTRIDIA gen. n. (Figs 146, 147)Type-species: Trichodectes lutrae Werneck. DESCRIPTION. Anterior of head with osculum absent, though pulvinus attaining margin; dorsal preantennalsulcus absent; clypeal marginal carina slightly broadened medially at junction with pulvinus; preantennaloutline broadly and smoothly rounded. Temple margin rectangular. Male scape not expanded; longitudin- 290 C. H. C. LYAL al setal row present, comprising four setae; flagellomeres fused in males and females; male flagellumlacking 'teeth'. Dorsum of head with setae short or of moderate length, sparse; temple margin with two orthree longer setae. Sitophore sclerite unmodified. Prothorax with two setae of medium length on posterior margin; pterothorax with one or two short,spine-like setae anterolaterally and six to ten long setae dorsally on posterior margin. Abdomen oval, with male segment IX projecting posteriorly. Abdominal spiracles present on segmentsIII-VIII. Abdominal setae as follows: pleurum III with short, stout setae anteriorly and posteriorly; pleuraIII-VIII lacking setae; terga I- VI (males) or I-IV (females) with central seta of median groups as long assegment, setae otherwise short; terga VII-VIII (males) or V-IX (females) with short setae, sparse; sternawith stout, short setae, sparse; anterior setae present only on pleurum II; postero-lateral setae absent.Abdominal pleura lacking projections. Sclerites absent from abdominal pleura and sterna, but present,slender, on at least terga III- VII (males) or V-IX (females), though may be very faint and not seen; maleterga lacking posterior sclerites. Gonapophyses with small ventral lobes formed from fused setal tubercles. Gonapophyses meet ventralvulval margin smoothly, linked by sclerotised band. Ventral vulval margin sclerotised, with chord at 90degrees to long axis of abdomen; submarginal non-tuberculate setae present. Subgenital lobe present,large, with lateral spine-like projections and associated setae present basally (Fig. 146). Male subgenital plate represented by sternite VIII with lateral arms extending anteriad. Pseudostyliabsent. Male genital opening dorsal. Basal apodeme attaining abdominal segment III, not slender.Parameres broad, scoop-shaped, not fused together, but fused to b.a.l.s. Mesomeres absent. Malegenitalia depicted in Fig. 147. HOSTS. Lutrinae (Carnivora: Mustelidae). SPECIES INCLUDEDlutrae (Werneck, 1937) comb. n. from Lutridia (1 cf , 1 $) Genus WERNECKODECTES Conci gen. rev. (Fig. 148)Werneckodectes Conci, 1946: 59. Type-species: TrichodectesferrisfWerneck, by original designation. DESCRIPTION. Osculum absent, though pulvinus attains anterior margin of head; dorsal preantennal sulcuspresent; clypeal marginal carina broadened slightly medially; preantennal portion of head short, outline 146 147 148 Figs 146-148 Trichodectini species. 146, Neolutridia lutrae, $ terminalia, ventral. 147, N. lutrae, cfgenitalia (after Werneck). 148, Werneckodectes ferrisi, cf genitalia (after Werneck). TRICHODECTID MAMMAL LICE 291 smoothly and broadly rounded. Temple margin convex. Male scape expanded; flagellomeres fused inmales and females; male flagellum with two basally-articulated 'teeth' and basal projection. Dorsum ofhead with setae of moderate length. Thorax with dorsal setae longest postero-laterally on pterothorax and along posterior margin onpterothorax; shorter setae present submarginally along posterior of pterothorax and on postero-lateralangles of prothorax, where there is a small setal patch; disc and posterior margin of pronotum each withpair of small setae. Abdomen oval, tapering posteriorly rather more in male than female. Abdominal spiracles present onsegments III- VIII. Abdominal setae numerous, anterior setae being present on sterna, terga and possiblypleura of all segments; anterior setae smaller than setae of posterior setal rows on each segment;postero-lateral setae, if present, obscured by large number of other setae. Abdominal pleurum IV withventral projection in male, possibly with dorsal projection in female; pleurum III possibly with dorsalprojection in female. Sclerites present on all abdominal pleura, on terga V-IX (males) or terga VII-IX(females) and on sterna III-VI (males only - sclerites absent on female sterna); male terga withoutposterior tergites, but anterior tergites, where present, with median longitudinal division. Gonapophyses with non-tuberculate setae on ventral margin; ventral lobe present, small; setae on lobestout, short, whilst setae distal to lobe longer, more slender. Gonapophyses meet ventral vulval marginsmoothly, linked by broad sclerotised band. Ventral vulval margin sclerotised, with chord at 90 degrees tolong axis of abdomen; marginal non-tuberculate setae present, stout, short. Subgenital lobe present,broad, with lateral rounded projections and associated setae present basally. Male subgenital plate with sternite VII fused to s.g.p.r., sternites VIII and IX absent; s.g.p.r. withsinuate margins. Pseudostyli absent. Male genital opening dorsal. Parameres not as long as basal apodeme,not fused together. Basiparameral sclerites absent. Mesomeres not apically fused, abutting parameres andb.a.l.s. basally. Male genitalia depicted in Fig. 148. HOSTS. Ursidae (Carnivora). COMMENTS. Werneckodectes has been treated as a synonym and a subgenus of Trichodectes (by Hopkins,1942 and Hopkins, 1949 respectively) ; its most recent placement was as a synonym of Trichodectes. A morecomprehensive history of the variations in status of Werneckodectes is given in Table 4. SPECIES INCLUDEDferrisi(Werneck, 1944) comb. rev. from Trichodectes. Genus TRICHODECTES NitzschThe genus Trichodectes comprises three subgenera. DESCRIPTION. Anterior of head with osculum present or absent* , but always with pulvinus attaining margin;dorsal preantennal sulcus present or absent; clypeal marginal carina broadened to variable extent mediallyto form simple bar with posterior margin straight or concave, or carina broadened into dorsal sclerite whichis heavily-sclerotised laterally (dorsal to margin of clypeus and pulvinus) and lightly or very lightly-sclerotised medially (posterior to the osculum) , more or less convex posteriorly or U-shaped with medianposterior process* ; antero-lateral margin of head straight, convex or sinuate* ; preantennal portion of headlong or short*, outline broadly rounded, broadly triangular, trapezoid or only slightly produced anteriadbetween coni*. Temple margin convex, rectangular or produced laterally*. Male scape expanded or notexpanded* ; longitudinal setal row present, comprising at least four setae; flagellomeres fused in males andfemales; male flagellum with one, two or four basally-articulated 'teeth' or 'teeth' absent*. Dorsum of headwith setae short, of moderate length or long, longest setae generally present along posterior templemargin; setae sometimes sparse. Sitophore sclerite unmodified. Thorax with prothoracic dorsal setae sparse, short or of moderate length* posteriorly and postero-laterally, absent from disc; pterothorax with setae on postero-lateral angles short and spine-like or ofmoderate length, dorsal setae otherwise present on posterior margin only, long or short, numerous, sparseor absent*. Abdomen oval, male segment IX sometimes slightly projecting posteriad, but usually positioneddorsally on the abdomen (Figs 156, 158, 160, 164, 165). Abdominal spiracles present on segments HI-IV,III-V, III- VII or III- VIII*; spiracle on segment VIII, if present, sometimes much smaller than those onsegments III- VII*. Abdomen with at least some tergal and sternal setae as long as segment, or setae veryshort, sparse and absent from pleura V and VI*; terga with lateral and median groups of setae frequentlydistinct; tergal setae numerous, or median group reduced to a single seta or absent*; male terga II and IIIsometimes with median group comprising exceptionally long, stout setae (Fig. 161)*; anterior setae 292 C. H. C. LYAL 151 Figs 149-152 Trichodectes (T.) species, $ terminalia. 149, T. (T.) emersoni. 150, T. (T.) canis. 151, T.(T.)galictidis. 152, T. (T.) p. pinguis, ventral. TRICHODECTID MAMMAL LICE 293 I I CH- o s TfIT) 294 C. H. C. LYAL 157 Figs 155-157 Trichodectes (Stachiella) species, $ terminalia. 155, T. (S.) octomaculatus . 156, T. (S.)erminiae. 157, T. (S.) potus. present only on pleurum II; postero-lateral setae presumed absent, or presence obscured by numerouslong setae or reduction (or absence) of lateral setal group. Abdominal pleura lacking projections, orprojections present dorsally on pleura II, III and IV* (Fig. 160). Abdominal sclerites variable, present orabsent*; male terga with or without anterior and posterior sclerites*. Gonapophyses with separate tuberculate setae and single apical non-tuberculate seta on ventral margin(Figs 149, 151, 153, 156, 159), or tubercles more closely associated (Fig. 155) or all setae non-tuberculate(Figs 152, 154)*; ventral lobe absent. Gonapophyses meet ventral vulval margin smoothly, linked bysclerotised band, or band absent. Ventral vulval margin sclerotised or, rarely, not sclerotised; with chord at TRICHODECTID MAMMAL LICE 295 00 in 296 C. H. C. LYAL TRICHODECTID MAMMAL LICE 297 B ^H I ^2 K' 3 h^ ^ s il 298 C. H. C. LYAL 167 166 170 Figs 166-171 Trichodectes species, C? genitalia. 166, T. (Paratrichodectes) ovalis. 167, T. (P.) zorillae.168, T. (P.) ugandensis. 169, T. (T.) canis. 170, T. (T.) galictidis. 171, T. (T.) emersoni. 90 degrees to long axis of abdomen; marginal setae present, tuberculate or non-tuberculate. Subgenitallobe present, usually with marginal serrations and lateral basal projections, the latter sometimes withassociated setae. Male subgenital plate absent (Figs 158, 160-162, 164), represented only by s.g.p.r. (Fig. 163) or enlargedsternite VIII (Fig. 159), with sternites VII, VIII and IX present and fused to s.g.p.r. (Fig. 165), or of thelatter form but with sternite VIII divided medially. Pseudostyli absent. Male genital opening dorsal orpostero-dorsal. Parameres fused or separate, fused to b.a.l.s. or free*. Basiparameral sclerites absent.Mesomeres absent or present, fused or unfused; if mesomeres fused apically, median extension absent. TRICHODECTID MAMMAL LICE 299 174 Figs 172-174 Trichodectes (Stachiella) species, c? genitalia. 172, T. (S.) octomaculatus. 173, T. (S.)emeryi. 174, T. (S.) erminiae. HOSTS. Canidae, Mustelidae, Procyonidae, Ursidae and Viverridae (Carnivora). COMMENTS. The different concepts of the extent of the genus Trichodectes held by various workers aresummarised in Table 4. Subgenus TRICHODECTES Nitzsch(Figs 10, 149-152, 158, 159, 162, 163, 169-171) Trichodectes Nitzsch, 1818: 294. Type-species: 'Trichodectes canis DeGeer (syn. T. latus N.)', by subsequent designation [Johnston & Harrison, 1911: 326].Ursodectes Keler, 1938a: 435. Type-species: Trichodectes pinguis Burmeister, by original designation. [Synonymy by Hopkins, 1942: 444.] Grisonia Keler, 1938a: 464. [No type-species designated] [Homonym of Grisonia Gray, 1843: 68].Galictobius Keler, 1938fc: 228. [Replacement name for Grisonia Keler.] Type-species: Trichodectes galictidis Werneck, by original designation. [Synonymy by Hopkins, 1942: 444.] DESCRIPTION. Male scape expanded; male flagellum with two or four basally-articulated 'teeth'. Abdominal spiracles present on segments III- VIII, spiracle on segment VIII not smaller than those onanterior segments. Abdomen with at least some tergal and sternal setae as long as segment, or setae veryshort, sparse and absent from pleura V and VI (kuntzi and emersoni); male terga II and III sometimes with 300 C. H. C. LYAL median group comprising exceptionally long, stout setae (undescribed sister-species to T. galictidis), tergalsetae never with median group reduced to single seta or absent (except sometimes on tergum I).Abdominal pleura lacking projections. Abdominal sternal sclerites absent, or present on sterna V-VIIIonly; abdominal tergal sclerites present or, more usually, absent; if tergites present on male, then not withanterior and posterior sclerites on each segment. Female genital chamber with ventral wall frequently obscure, dorsal wall bearing sclerotised nodules,sometimes fused together. Male subgenital plate absent, represented by s.g.p.r. only, or by enlarged sternite VIII only. Parameresfused to form plate or unfused; symmetric or asymmetric; not fused to b.a.l.s. Faintly-sclerotisedtongue-like sclerite of uncertain homology sometimes present dorsally between parameres if mesomeresabsent (Fig. 169). Mesomeres absent or, if present, fused or unfused. Male genitalia depicted in Figs169-171. HOSTS. Canidae, Mustelidae, Ursidae and Viverridae (Carnivora). COMMENTS. Trichodectes Nitzsch, 1818 was placed on the Official List of Generic Names in Zoology, withthe type-species Trichodectes canis DeGeer, by Opinion 627 of the International Commission of Zoologic-al Nomenclature (1962, Bulletin of Zoological Nomenclature 19: 91). SPECIES INCLUDED canis (DeGeer, 1778) (c.100 cf, c.100 $)emersoni Hopkins, 1960 (15 cf , 12 $) ga/icfMsWerneck, 1934 (15 cf , 16 $; also 2 cf, 1 $ of an undescribed sister-species)kuntzi Emerson, 1964 (15 cf , 14 $)me/is (Fabricius, 1805) (c.60 cf , c.60 $)pinguispinguisBurmeister, 1838 (1 cf , 5 9)pinguis euarctidos Hopkins, 1954 (20 cf , 20 $)vosse/er/Stobbe, 1913 (2 cf , 7 $) Subgenus PARATRICHODECTESsubgen. n. (Figs 6, 154, 160, 161, 166-168)Type-species: Trichodectes ovalis Bedford. DESCRIPTION. Anterior of head with osculum present; clypeal marginal carina broadened medially intodorsal sclerite which is heavily-sclerotised laterally and lightly-sclerotised medially, more or less convexposteriorly or U-shaped with median posterior process; antero-lateral margin of head convex or sinuate;preantennal portion of head not long, outline rounded or subtriangular. Temple margin convex orrectangular. Male scape not greatly expanded; male flagellum with two basally-articulated 'teeth'. Thorax with prothoracic dorsal setae sparse, of moderate length posteriorly and postero-laterally,absent from disc; pterothorax with setae on postero-lateral angle long dorsally, short and spine-likeventrally; dorsal posterior pterothoracic setae submarginal, long, comprising two pairs with wide median gap. Abdominal spiracles present on segments III- VII. Abdomen with at least some tergal and sternal setaeas long as segment; setae present on all pleura; tergal setae numerous, median group not reduced to a singleseta or absent except sometimes on tergum I or on posterior terga of males only , if male terga II and III withmedian setal group comprising exceptionally long, stout setae; postero-lateral setae presumed absent,though may be present as the most lateral seta of lateral group, which is frequently situated moreposteriorly than other setae. Abdominal sterna and terga with or without sclerites; male terga, if scleritespresent, with anterior sclerites only. Female genital chamber with dorsal wall not bearing sclerotised nodules. Male subgenital plate unsclerotised. Parameres separate or thinly fused to each other; symmetric orasymmetric; not fused to b.a.l.s. Tongue-like sclerite not present. Mesomeres absent. Male genitaliadepicted in Figs 166-168. HOSTS. Mustelinae (Carnivora: Mustelidae). SPECIES INCLUDED ovalis Bedford, 1928 comb. rev. from Stachiella (15 cf , 12 $)ugandensis Bedford, 1936 comb. rev. from Stachiella (44 cf , 47 $)zor/tfaeStobbe, 1913, comb. rev. from Stachiella (17 cf , 26 $) TRICHODECTID MAMMAL LICE 301 Subgenus STACHIELLA Keler stat. n. (Figs 153, 155-157, 164, 165, 172-174) Stachiella Keler, 1938a: 428. Type-species: Trichodectes pusillus Nitzsch [= Pediculus mustelae Schrank], by original designation.Potusdia Conci, 1942: 141. Type-species: Trichodectes potus Werneck, by original designation. [Synony- mised with Trichodectes by Werneck, 1948: 110; syn. n. of Stachiella.] DESCRIPTION. Clypeal marginal carina broadened medially into dorsal sclerite which is heavily-sclerotisedlaterally and lightly or very lightly-sclerotised medially, more or less convex posteriorly or U-shaped withmedian posterior process; antero-lateral margin of head smoothly convex; preantennal portion of headlong or short, outline broadly rounded, sometimes only slightly produced anteriad between coni. Malescape not, or only slightly, expanded; male flagellum with one or two basally-articulated 'teeth' or 'teeth'absent. Thorax with prothoracic dorsal setae sparse, of moderate length posteriorly and postero-laterally,absent from disc; ptero thorax with setae on postero-lateral angles short and spine-like or of moderatelength, dorsal posterior setae submarginal, long, comprising one or more pairs with wide median gap, orabsent. Abdominal spiracles present on segments III-IV, III-V, or III-VIII; spiracle on segment VIII, ifpresent, sometimes much smaller than those on segments III-VII. Abdomen with at least some tergal andsternal setae as long as segment; setae present on all pleura; terga with median setal group of male reducedto a single seta on most segments, lateral group small; female tergal setae of similar arrangement or withmedian group absent (see discussion on p. 221 above); male terga II and III never with median groupcomprising exceptionally long, stout setae; postero-lateral setae presumed absent, though perhaps presentas the most lateral seta of lateral setal group, which is frequently situated more posteriorly than other setaeof the group. Abdominal pleura lacking projections. Abdominal sterna variably sclerotised, with sternites,if present, most commonly on posterior segments; abdominal terga with sclerites present on segmentsIII-VIII or III-IX, sometimes on I and II; male terga frequently with both anterior and posterior sclerites,though posterior sclerites may be faintly-sclerotised or absent; abdominal pleurum II frequently sclero-tised, otherwise pleura unsclerotised. Gonapophyses with separate or closely-associated tuberculate setae and single apical non-tuberculateseta on ventral margin. Gonapophyses meet ventral vulval margin smoothly, linked by sclerotised band.Ventral vulval margin sclerotised. Female genital chamber with dorsal wall not bearing sclerotisednodules. Male subgenital plate absent, or with sternites VII, VIII and IX present and fused to s.g.p.r., or of thisform but with sternite VIII divided medially. Parameres separate or thinly fused to one another;symmetric, asymmetric or asymmetrically deflected; fused or not fused to b.a.l.s. Tongue-like sclerite notpresent. Mesomeres absent, present, small and unfused, or present and fused. Male genitalia depicted inFigs 172-174. HOSTS. Mustelidae: Mustelinae and Procyonidae (Carnivora). COMMENTS. Stachiella has been considered a synonym and a subgenus of Trichodectes (Hopkins, 1942 andHopkins, 1949 respectively), although its most recent treatment (Emerson & Price, 1981) was as a validgenus. Potusdia has also been considered as a valid genus, subgenus of Trichodectes and synonym ofTrichodectes but has not, before this study, been considered a synonym of Stachiella. A more comprehen-sive history of the variations in status of Stachiella and Potusdia is given in Table 4. SPECIES INCLUDED divaricat us Harrison, 1915 comb. rev. from Stachiellaemeryi Emerson & Price, 1974 [treated by Emerson & Price (1974, 1981) as Trichodectes s. str.] (4 cT, 17 ?) erminiae (Hopkins, 1941) comb. n. from Stachiella (92 d" , 100 $)fallax Werneck, 1948 [treated by Werneck (1948) and all subsequent authors as Trichodectes s. str.] (2cf,2$) jacobi (Eichler, 1941) comb. n. from Stachiellaking! McGregor, 1917 comb. rev. from Stachiella (14 cf , 25 <j>)larseni (Emerson, 1962) comb. n. from Stachiella (58 cf , 57 $)mustelae (Schrank, 1903) comb. n. from Stachiella (18 cf , 23 $)octomaculat us Paine, 1912 [treated by all authors as Trichodectes s. str.] (48 cf , 58 $) 302 C. H. C. LYAL potus Werneck, 1934 [treated by Werneck (1948) and Emerson & Price (1981) as Trichodectes s. str.] (19cf,25$) refiisusrefususBurmeister, 1838 comb. rev. from Stachiella (1 cf , 1 $)retusus martis ( Werneck, 1948) comb. n. from Stachiella (1 cf)retusussalfii (Conci, 1940) comb. n. from Stachiella Genus FELICOLA EwingThe genus Felicola comprises two subgenera. DESCRIPTION. Anterior of head with osculum present or absent; dorsal preantennal sulcus present orabsent; clypeal marginal carina, if osculum absent, with very slight median broadening, or, if osculumpresent," carina broadened medially into dorsal sclerite which is heavily-sclerotised laterally (dorsal tomargin of clypeus and pulvinus) and very lightly-sclerotised medially (posterior to osculum); anterolateralmargin of head straight, slightly sinuate or convex; preantennal portion of head long or short, outlinetriangular or broadly rounded*. Temple margin rectangular or convex. Male scape variably expanded ornot expanded* , with longitudinal setal row present and comprising at least three setae; flagellomeres fusedin males and females; male flagellum with one or three basally-articulated 'teeth', or variable number of'teeth' fused to flagellum, or 'teeth' absent*. Dorsum of head with setae short, of moderate length or long,but in any case frequently longer than abdominal tergal setae; setae sparse, though most numerousanteriorly and along lateral margins. Sitophore sclerite unmodified. Thorax with dorsal setae long or of moderate length though frequently short and spine-like onpostero-lateral angles of pterothorax, not present on disc of prothorax or pterothorax. Abdomen oval or elongate, frequently terminating in more or less acute projection of segment IX in themale (Figs 179, 181, 183-189). Abdominal spiracles absent, or present on segment III, segments III-IV, orsegments III-V*. Abdominal setae short, of moderate length, or occasionally long and fine* (Figs 182, 186,188); male tergum II frequently with 2-6 very long setae medially (Figs 179, 180, 187, 189); abdominalpleurum III frequently with posterior setal row comprising stout, conical setae (Figs 181, 182, 188);anterior setae, if present, only on pleurum II; postero-lateral setae present or absent. Abdominal pleurumIII with projections absent, or, if present, dorsal or ventral, sclerotised or unsclerotised; abdominalpleurum IV with projections absent or, if present, dorsal and occasionally ventral, sclerotised orunsclerotised. Abdominal sclerotisation variable; sternal, tergal and pleural sclerites, if present, generallyon anterior segments, becoming less clear on posterior segments; male terga sometimes with anterior andposterior sclerites, at least on tergum VI*. Gonapophyses with non-tuberculate setae and rounded or rectangular lobe present on ventral margin;apical spur present or absent* . Gonapophyses meet ventral vulval margin acutely, not linked by sclerotisedband. Ventral vulval margin not sclerotised; straight or concave, with chord less than 90 degrees to longaxis of abdomen; subgenital lobe present, apically single or bifurcate, with margins serrate, at leastposteriorly (Fig. 175) (see comment below). Male subgenital plate not present, though sternite VIII sometimes with postero-lateral projectionsprobably homologous with lateral rods of subgenital plate (Fig. 181). Pseudostyli absent. Male genitalopening postero-dorsal or dorsal; segment IX frequently produced posteriad. Parameres generally longand slender (Figs 191, 192, 197, 201), occasionally broader (Fig. 198); frequently fused completely or,more usually, basally only. Basiparameral sclerites absent. Mesomeres present or absent; if present,unfused (Fig. 200) or, if fused, median extension absent*. Male genitalia depicted in Figs 190-206. HOSTS. Herpestidae, Viverridae, Felidae and Canidae (Carnivora). COMMENTS. Emerson & Price (1980) distinguish the females of their new species Suricatoecus occidentals(transferred to Felicola in this study) from other species in the 'helogale Group' (equivalent to thecongoensis-occidentalis clade) by the presumed absence of the subgenital lobe in occidentalis . Examina-tion of the type-series of this species, however, reveals that the subgenital lobe, although very fine, ispresent in all the females. A second species of this clade, close to F. helogaloidis, has been taken from skinsof Crossarchus obscurus (the host of F. occidentalis) and specimens are in the collection of the BritishMuseum (Natural History). Subgenus FELICOLA Ewing(Figs 5, 175, 179-185, 190-193, 196-198, 202-204) FelicolaEwing, l.vi.1929: 121, 122, 192. Type-species: Trichodectes subrostratus Burmeister [attributed toNitzsch], by original designation. TRICHODECTID MAMMAL LICE 303 304 C. H. C. LYAL S >*; TRICHODECTID MAMMAL LICE 305 CO a 306 C. H. C. LYAL 183 Figs 183-185 Felicola (F.) species. 183, F. (F.) calogaleus, cf abdomen. 184, F. (F.) helogale, C?terminalia. 185, F. (F.) hopkinsi, cf terminalia. Felicinia Bedford, -.x.1929: 519. Type-species: Trichodectes subrostratus Burmeister [attributed to Nitzsch], by original designation. [Synonymised by Bedford, 1932a: 536.]Bedfordia Keler, 1938a: 463. Type-species: Felicola helogale Bedford, by original designation. [Homonym oi Bedfordia Fahrenholz, 1936: 55.] [Synonymised with Felicola by Hopkins, 1941: 36.]Fastigatosculum Keler, 1939: 11. [Replacement name for Bedfordia Keler.] [Synonymised with Surica- toecus by Werneck, 1948.] DESCRIPTION. Preantennal portion of head with outline narrowly or broadly triangular or rounded. Maleflagellum with 'teeth' absent or, if present, numbering one, two, three or four and fused to flagellum, notbasally articulated. Abdominal spiracles absent, or present on segments III-IV or III-V. Abdominal setae very short or ofmoderate length. Gonapophyses with rounded lobe on ventral margin; spur present or absent. TRICHODECTID MAMMAL LICE 307 CTJ b 308 C. H. C. LYAL O)CO o oo S V. 8 TRICHODECTID MAMMAL LICE 309 193 Figs 190-194 Felicola species, C? genitalia. 190, F. (F.) calogaleus. 191, F. (F.) inaequalis, endophalluspartially everted. 192, F. (F.) inaequalis, endophallus not everted. 193, F. (F.) setosus; 194, F.(Suricatoecus) fahrenholzi. Everted portion of endophallus frequently thinly sclerotised (Figs 190, 196, 204). Mesomeres presentand fused or absent. HOSTS. Felidae, Herpestidae and Viverridae (Carnivora). COMMENTS. F. genettae (Fresca) is included on the basis of the figures and description of Fresca (1924)which, although poor, suggest an affinity with the calogaleus-viverriculae clade. If this is a correctplacement, the host record of Genetta genetta rhodanica is anomalous. The various treatments of the junior 310 C. H. C. LYAL 195 196 199 Figs 195-199 Felicola species, O" genitalia. 195, F. (Suricatoecus) bedfordi. 196, F. (F.) robertsi. 197, F.(F.) hopkinsi. 198, F. (F.) congoensis. 199, F. (F.) helogale. synonyms of Felicola are summarised in Table 5; omitted from the table is Conci (1946) who, like Eichler(1941, 1963) regarded Fastigatosculum as a valid genus. SPECIES INCLUDED calogaleus (Bedford, 1928) (14 cf , 24 $) congoensis (Emerson & Price, 1967) comb. n. from Suricatoecus (23 cf , 31 $)cynictis (Bedford, 1938) (13 C?, 19 9)genettae (Fresca, 1924) TRICHODECTID MAMMAL LICE 311 200 205 Figs 200-206 Felicola species, cf genitalia. 200, F. (Suricatoecus) acutirostris. 201, F. (S.) vulpis. 202, F.(F.) liberiae. 203, F. (F.) minimus, parameres displaced slightly apart. 204, F. (F.) subrostratus. 205, F.(S.) decipiens. 206, F. (S.) decipiens, detail of paramere. 312 C. H. C. LYAL helogale Bedford, 1932, comb. rev. from Suricatoecus (3 cf , 7 $) Ae/oga7oid/s( Werneck, 1948) comb. n. from Suricatoecus hopkinsi Bedford, 1936 comb. rev. from Suricatoecus (2 cf , 1 $) inaegua/is(Piaget, 1880) (10 cf , 19 $) /ifreriae Emerson & Price, 1972 (7 cf , 3 $) minimus Werneck, 1948 (15 cf , 28 $) occidental!* (Emerson & Price, 1981) comb. n. from Suricatoecus (4 cf , 5 9) ra/i/ni Emerson & Stojanovitch, 1966 (9 cf , 16 $) robertsi Hopkins, 1944 (8 cf , 15 $) rohani Werneck, 1956 (68 cf , 62 $) setosus Bedford, 1932 (14 Cf, 18 ?) suferosfrafus(Burmeister, 1838) (103 cf, 120 $) viverriculae (Stobbe, 1913) comb. n. from Parafelicola (21 cf, 26 9; 6 cf, 6 $ of undescribed sister-species)zeylonicus Bedford, 1936 (6 cf , 7 $) Subgenus SURICATOECUS Bedford stat. n.(Figs 176-178, 186-189, 194, 195, 200, 201, 205, 206) Suricatoecus Bedford, 1932a: 354. Type-species: Trichodectes cooleyi Bedford, by monotypy.EichlerellaCond, 1942: 140. Type-species: Trichodectes vulpis Denny, by original designation. [Synonymyby Werneck, 1948: 172.] DESCRIPTION. Preantennal portion of head with outline narrowly triangular or broadly rounded. Malescape not, or very slightly, expanded; male flagellum with 'teeth' absent or, if present, numbering one orthree and basally articulated. Abdominal spiracles absent, or present on segment III or segments III-V. Abdominal setae very short,of moderate length, or long and fine. Tergal and sternal sclerites generally present on abdomen, thoughless clear on posterior segments; male terga never with posterior sclerites. Gonapophyses with rounded or rectangular lobe on ventral margin; spur present. Everted portion of endophallus never sclerotised. HOSTS. Canidae and Herpestidae (Carnivora). COMMENTS. Suricatoecus has been treated not only as a genus, but also as a synonym and a subgenus ofFelicola (Bedford, 1936 and Hopkins, 1949 respectively); a more complete history of the variations instatus of Suricatoecus and of Eichlerella is provided in Table 5. Eichler (1963) included the manuscriptname Felicomorpha in his catalogue, without providing any further details, the name therefore being anomen nudum. In an earlier, unpublished work, Eichler had attributed this name to Keler m.s. , and notedthe type-species, which is a junior synonym of T. vulpis Denny. Felicomorpha is, however, not an availablename. SPECIES INCLUDED acutirostris (Stobbe, 1913) [treated as Felicola s. str. by previous authors] (2 cf , 2 $)bedfordi Hopkins, 1942 [treated as Felicola s. str. by previous authors] (4 cf , 11 9)cooleyi (Bedford, 1929) comb. n. from Suricatoecus (30 cf , 28 $)decipiens Hopkins, 1941 comb. rev. from Suricatoecus (9 cf , 9 $)fahrenholzi( Werneck, 1948) comb. n. from Suricatoecus (16 cf , 16 $)fennecus (Emerson & Price, 1981) comb. n. from Suricatoecusguinlei( Werneck, 1948) comb. n. from Suricatoecus (4 cf , 10 $)macrurus Werneck, 1948 [treated as Felicola s. str. by previous authors] (23 cf , 19 $)pygidialis Werneck, 1948 [treated as Felicola s. str. by previous authors] (36 cf , 41 $)quadraticeps (Chapman, 1897) comb. n. from Suricatoecus (5 cf , 9 $)vulpis (Denny, 1842) comb. n. from Suricatoecus (18 Cf , 25 $) Genus LORISICOLA BedfordThe genus Lorisicola comprises two subgenera. DESCRIPTION. Anterior of head with osculum present or absent*; dorsal preantennal sulcus present orabsent; clypeal marginal carina broadened medially, median sclerite variable*; anterolateral margin of TRICHODECTID MAMMAL LICE 313 210 211 Figs 207-211 Lorisicola species. 207, L. (L.) mjoebergi, 9 abdomen. 208, L. (Paradoxuroecus)bengalensis, 9 subgenital lobe, ventral. 209, L. (P.) africanus, $ head, dorsal. 210, L. (L.) spenceri, $terminalia, ventral, setae omitted apart from on gonapophyses. 211, L. (L.) felis, $ gonapophysis,ventral. 314 C. H. C. LYAL 212 213 Figs 212, 213 Lorisicola (L.) species, C? terminalia. 212, L. (L.) similis. 213, L. (L.) spenceri. head sinuate or convex; preantennal portion of head of variable length, outline more or less broadlytriangular or rounded, sometimes, if osculum absent, convexly produced anteriorly*. Temple marginconvex, rectangular, or slightly produced laterally*. Male scape expanded or only slightly expanded, withlongitudinal setal row present and comprising at least two setae*; male and female flagellomeres fused;male flagellum with one or two basally- articulated teeth* . Dorsum of head with setae short or of moderatelength, sparse. Sitophore sclerite unmodified. Thorax with dorsal setae short or of moderate length* though frequently short and stout on postero-lateral angles of pterothorax, not present on disc or medially posteriorly on prothorax or pterothorax. Abdomen oval or elongate, male segment IX not produced greatly. Abdominal spiracles absent, orpresent on segments III- VI or III- VIII ; posterior two pairs of spiracles , if six pairs present , sometimes verysmall and possibly non-functional*. Abdominal setae short or very short, frequently sparse dorsally;abdominal pleurum III frequently with posterior setal row comprising stout, conical setae (Fig. 218);anterior setae absent except on pleurum II; postero-lateral setae present or absent, sometimes numberingmore than one per site* (Fig. 207). Pleural projections present dorsally and sometimes ventrally onpleurum IV, sclerotised or unsclerotised. Abdominal sclerotisation variable; sterna with sclerites absent TRICHODECTID MAMMAL LICE 315 } 214 215 Figs 214, 215 Lorisicola species, cf terminalia. 214, L. (L.) mjoebergi. 215, L. (Paradoxuroecus) laticeps. except for subgenital plate (in male), or present on more posterior segments (VII, VI + VII, V-VII,IV-VII or III-VII); terga with sclerites on I-VIII, II- VIII, III-VIII or IV-VIII; pleura with sclerites on atleast II, sometimes also on III and IV; male terga with anterior and posterior sclerites present on at leastterga IV-VII, or posterior sclerites not present*. Gonapophyses with ventral marginal non-tuberculate setae; rounded or rectangular lobe present orabsent on ventral margin*; apical spur present or absent*. Gonapophyses meet ventral vulval marginacutely, not linked by sclerotised band. Ventral vulval margin not sclerotised; straight or concave, withchord less than 90 degrees to long axis of abdomen; subgenital lobe present, ventral surface more or lesscovered in overlapping scales*. Male subgenital plate with sternites VII, VIII and IX present and fused to s.g.p.r. , VII and VIII fused tos.g.p.r. and IX absent, or VII fused to s.g.p.r. and VIII and IX absent or present but not fused to s.g.p.r. *.Pseudostyli absent. Male genital opening postero-dorsal or dorsal; male segment IX lying dorsally onabdomen. Parameres short, broad, sometimes fused. Basiparameral sclerites present or absent*. Meso-meres present, fused; median extension present or absent* ; mesomeres extending basally between b.a.l.s.to contact parameres, or terminating exteriorly to b.a.l.s.*. Male genitalia depicted in Figs 219-226. HOSTS. Felidae, Herpestidae and Viverridae (Carnivora) and Lorisidae (Primates). 316 C. H. C. LYAL Fig. 216 Lorisicola (Paradoxuroecus) acuticeps, C? abdomen. Subgenus LORISICOLA Bedford (Figs 207, 210-214, 219, 220)Lorisicola Bedford, 1936: 51. Type-species: Trichodectes mjoebergi Stobbe, by original designation. DESCRIPTION. Anterior of head with osculum present; clypeal marginal carina broadened medially to formrectangular or W-shaped sclerite, or broadened slightly to either side of osculum, very lightly sclerotisedposterior to osculum. Temple margin rectangular or slightly produced laterally. Male scape expanded oronly slightly expanded, with setal row comprising two setae; male flagellum with basally-articulated 'teeth'on projection. Abdominal spiracles absent, or present on segments III-VIII; posterior two pairs of spiracles, if six pairspresent, sometimes very small and possibly non-functional. Males with posterior tergal sclerites absent. Gonapophyses with rounded lobe, or lobe absent (Fig. 207); apical spur present or absent (Fig. 211).Subgenital lobe covered in overlapping pointed scales (Fig. 210) or spines (Fig. 207). Male subgenital plate with sternites VII, VIII and IX present and fused to s.g.p.r. (Fig. 214), or withsternite VII fused to s.g.p.r. and VIII and IX present but not fused to s.g.p.r. (Fig. 212), or of the latterform but lacking sternite IX (Fig. 213), or lacking VIII and IX but VII very broad. Basiparameral scleritesabsent. Mesomeres fused, with median extension. Mesomeres extending basally between b.a.l.s. to TRICHODECTID MAMMAL LICE 317 Fig. 217 Lorisicola (Paradoxuroecus) bengalensis, cf abdomen. contact parameres (Fig. 222), terminating exteriorly to b.a.l.s. (Fig. 220), or extending anteriad toposterior end of b.a.l.s. and abruptly reversing, lying ventrally to b.a.l.s., though reversed portion isdifficult to see (Fig. 219). HOSTS. Felidae and Viverridae (Carnivora) and Lorisidae (Primates). COMMENTS. The original spelling mjobergi is emended here to mjoebergi in accordance with Article 32(d)(i)of the International Code of Zoological Nomenclature (1984). SPECIES INCLUDED americanus (Emerson & Price, 1983) comb. n. from Felicolabraziliensis (Emerson & Price, 1983) comb. n. from Felicolacaffra (Bedford, 1919) comb. n. from Felicola (1 cf , 1 $)/e7is(Werneck, 1934) comb. n. from Felicola (1 cf , 1 ?)hercynianus (Keler, 1957) comb. n. from Felicola (6 cf , 6 $)ma/a^s/aiius(Werneck, 1948) comb. n. from Trichodectes (4 cf , 5 9)nyoeftergi(Stobbe, 1913) (c.100 cf , c.100 $)neofelis (Emerson & Price, 1983) comb. n. from Felicolasiamensis (Emerson, 1964) comb. n. from Felicola (5 cf , 2 <j>)similis (Emerson & Price, 1983) comb. n. from Felicola (1 (5", 1 9)spenceri (Hopkins, 1960) comb. n. from Felicola (8 Cf , 5 $)sudamericanus (Emerson & Price, 1983) comb. n. from Felicola 318 C. H. C. LYAL Fig. 218 Lorisicola (Paradoxuroecus) juccii, C? abdomen. Subgenus PAlMDOXl/tfOECI/S Conci gen. rev., stat. n. (Figs 208, 209, 215-218, 221-226) Paradoxuroecus Conci, 1942: 141. Type-species: Paradoxuroecus juccii Conci, by original designation. Parafelicola Werneck, 1948: 226. Type-species: Trichodectes acuticeps Neumann, by original designation.Syn. n. Neofelicola Werneck, 1948: 235. Type-species: Neofelicola aspidorhynchus Werneck, by original designa-tion. Syn. n. DESCRIPTION. Anterior of head with osculum present, in which case clypeal marginal carina broadenedslightly to either side of osculum and very lightly sclerotised posterior to osculum, or osculum absent, inwhich case clypeal marginal carina broadened medially to form posteriorly-convex or straight heavilysclerotised bar (Fig. 209); outline of preantennal portion of head more or less broadly rounded ortriangular, sometimes, if osculum absent, convexly produced anteriorly (Fig. 209). Temple margin convexor rectangular. Male scape expanded or only slightly expanded, with longitudinal setal row comprising atleast three setae; male flagellum with two basally-articulated 'teeth', only rarely on projection. Thorax with dorsal setae short. Abdominal spiracles absent, or present on segments III- VI. Postero-lateral setae present or absent, butnever numbering more than one per site if present. Males with posterior tergal sclerites absent, or anteriorand posterior sclerites present on at least terga IV- VII. Gonapophyses with lobe present, rounded or rectangular; apical spur present. Subgenital lobe bilobate,with scales modified into short spines in some cases, though spines may be sparse (Fig. 208). TRICHODECTID MAMMAL LICE 319 320 C. H. C. LYAL 224 223 225 226 Figs 223-226 Lorisicola (Paradoxuroecus) species, C? genitalia. 223, L. (P.) juccii. 224, L. (P.)bengalensis. 225, L. (P.) acuticeps. 226, L. (P.) acuticeps, endophallus everted. Male subgenital plate with sternites VII, VIII and IX present and fused to s.g.p.r. (Fig. 216), VII and IXpresent and fused to s.g.p.r. but VIII absent (Fig. 215), or VII and VIII present and fused to s.g.p.r. but IXabsent (Figs 217, 218). Parameres short, broad, not fused together (see second paragraph under'comments' below), but may be very closely associated (Figs 223, 224). Basiparameral sclerites present(Fig. 225) or absent. Mesomeres fused apically with median extension present or, if absent, arch with twoapical nipples (Fig. 221); mesomeral arch sometimes with lateral double flexion (Fig. 225); mesomeresextending basally between b.a.l.s. to contact parameres, sometimes sharply recurved posteriorly betweenb.a.l.s. (Fig. 224). Endophallus frequently with spicular collar, sometimes V-shaped, around gonopore(Fig. 225). HOSTS. Herpestidae and Viverridae (Carnivora). COMMENTS. Paradoxuroecus has been considered by most authors, following Werneck (1948), to be asynonym of Felicola; it is here recalled from synonymy and placed as a subgenus of Lorisicola. Neofelicolaand Parafelicola were both considered by Hopkins (1949) to be subgenera of Felicola. A more completehistory of the variations in the status of Paradoxuroecus, Neofelicola and Parafelicola is presented in Table5. Werneck (1948) figured the parameres of aspidorhynchus, sumatrensis and juccii as fused together;examination of the type-material of the first two species and of numerous specimens of the last has revealedthat this is not the case, although in L. juccii the parameres are very closely associated with each other. TRICHODECTID MAMMAL LICE 321 Lorisicola (Paradoxuroecus) bengalensis (Werneck, 1948) was described in Neofelicola from threefemales, the male being unknown. These females were taken from a museum skin of Paradoxurushermaphrodytus canus Miller, which was itself collected in Thailand. Female lice subsequently collectedfrom P. hermaphrodytus subspp. in Thailand agree with Werneck's (1948) description of N. bengalensis.Emerson (1965) describes the male of bengalensis, collected from the type host in Thailand, anddistinguishes a new species, N. philippinensis , collected from Paradoxurus philippinensis. Numerous specimens of lice from a number of subspecies of Paradoxurus hermaphrodytus have beenexamined during the course of this study, and it has become apparent that lice of two clades are present: theFelicola (F.) zeylonicus-viverriculae clade and the Lorisicola (P.) philippinensis-juccii clade. These clades may be distinguished by the following characters. F. (F.) zeylonicus-viverriculae clade - Three pairs of abdominal spiracles; male flagellum with 'teeth' notbasally articulated; parameres long, slender, extending anteriorly between b.a.l.s.; mesomeres notbasally extending between b.a.l.s. ; mesomeral arch lacking median extension; female subgenital lobebroad, smooth ventrally, with long, flattened marginal spines. L. (P.) philippinensis-juccii clade - Four pairs of abdominal spiracles; male flagellum with 'teeth' basallyarticulated; parameres short, broad, closely-associated with one another but not fused, not extendingbetween b.a.l.s., if reaching them; mesomeres basally extending between b.a.l.s. to contact para-meres; mesomeral arch with median extension; female subgenital lobe narrow, apically bilobate,covered ventrally with small pointed scales, lacking marginal spines. The female described as N. bengalensis by Werneck (1948) is, by the structure of the subgenital lobe andthe number of abdominal spiracles, a member of the L. (P.) philippinensis-juccii clade, as are both maleand female of N. philippinensis as described by Emerson (1965). The male described by Emerson (1965) asN. bengalensis is, however, a member of the F. (F.) zeylonicus-viverriculae clade and therefore not the truemale of N. bengalensis. Female lice of the latter clade are now known from Paradoxurus hermaphrodytussubspp. , as are males of the former. The male of L. bengalensis has genitalia indistinguishable from those of L. philippinensis. Emerson(1965) distinguishes L. philippinensis from L. bengalensis by the male genitalia, the greater number ofsternal and tergal setae in both sexes of the former, and the greater number of vulval setae in the former.The characteristics of the male genitalia, as stated above, are the same in the two species. Study of the largesample of specimens now available indicates that the vulval setal number of L. philippinensis is not outsidethe range of L. bengalensis. The tergal and sternal setae in the males are the same, as are the sternal setae inthe females. The tergal setae of the female paratype of L. philippinensis in the British Museum (NaturalHistory) collection are the same as those of L. bengalensis, but do not agree with the figure in Emerson(1965), where far more setae are depicted. The host species Paradoxurus philippinensis is at best asubspecies of P. hermaphrodytus, and all other subspecies appear to harbour L. bengalensis. L. philip-pinensis (Emerson) is consequently provisionally synonymised with L. bengalensis (Werneck), subject toexamination of the female allotype of philippinensis. SPECIES INCLUDED acuticeps (Neumann, 1902) comb. n. from Parafelicola (28 cf , 34 <J?)africanus (Emerson & Price, 1966) comb. n. from Parafelicola (15 cf , 7 $)aspidorhynchus ( Werneck, 1948) comb. n. from Neofelicola (6 cf , 7 9)bengalensis ( Werneck, 1948) comb. n. from Neofelicola (26 cf , 59 $)juccU Conci, 1942 comb. n. from Felicola (82 cf , 81 ?)laticeps( Werneck, 1942) comb. n. from Suricatoecus (10 cf , 13 $)/en/cora/s ( Werneck, 1948) comb. n. from Parafelicola (9 cf , 22 <J>)mungos (Stobbe, 1913) comb. n. from Suricatoecus (1 cf , 3 $)neoafricanus (Emerson & Price, 1968) comb. n. from Parafelicola (holotype cf , 4 $)paralaticeps ( Werneck, 1948) comb. n. from Suricatoecus (1 cf , 4 $)philippinensis (Emerson, 1965) comb. n. from Neofelicola (5 cf , 5 $)sumatrensis ( Werneck, 1948) comb. n. from Neofelicola (4 cf , 4 $)wernecW (Hopkins, 1941) comb. n. from Parafelicola (7 cf , 11 ?) NEOTRICHODECTINAE subfam. n. Type-genus: Neotrichodectes Ewing Genus NEOTRICHODECTES EwingThe genus Neotrichodectes comprises five subgenera. 322 227 228 229 231 Figs 227-231 Neotrichodectes species. 227, N. barbarae, cf anterior abdominal segments, dorsal. 228, N.mephitidis, 9 abdominal pleura II and III, dorsal. 229, N. mephitidis, cf terminalia. 230, N. barbarae, cftemple margin. 231, N. semistriatus, cf flagellum. DESCRIPTION. Anterior of head with osculum present, sometimes very shallow* ; dorsal preantennal sulcuspresent, not always clear; clypeal marginal carina broadened medially into dorsal sclerite of variable form*which is always more heavily sclerotised laterally (dorsal to margin of clypeus and pulvinus) than medially;antero-lateral margin of head convex, straight or sinuate*; preantennal portion of head short or longer,sometimes as long as postantennal portion*, outline broadly rounded or triangular*. Temple marginbroadly convex. Male scape expanded, with longitudinal setal row present and comprising at least foursetae; flagellomeres fused in males and females; male flagellum with two or more basally-articulated 'teeth' TRICHODECTID MAMMAL LICE 323 233 235 Figs 232-235 Neotrichodectes species, $ terminalia. 232, N. mephitidis, ventral. 233, N. chilensis,showing setal distribution. 234, N. chilensis, showing sclerites. 235, N. barbarae. distally, and with basal toothed or rough projection sometimes present* (Fig. 231); female pedicel lackingmembranous projection. Dorsum of head with most setae short or of moderate length, more or less sparse,sometimes with long seta on posterior margin of temple* (Fig. 230). Sitophore sclerite unmodified. Thorax with dorsal setae long or of moderate length* though frequently short and spine-like onpostero-lateral angles of prothorax and pterothorax; posterior margin of pronotum with four setae andwide median gap, posterior of pterothorax dorsally with setae more abundant, marginal or submarginal,median gap present or absent. Abdomen broadly rounded, not greatly projecting posteriad in male (Fig. 229). Abdominal spiraclesabsent. Abdominal setae grenerally abundant, as long as segment or sparse, shorter, with long setaepresent only on posterior pleura* ; terga, especially of males, with median and lateral setal groups distinct,though median groups generally united; male terga II- VI with median two setae much smaller than othersetae in the row, sometimes separated by one or more longer setae (Fig. 229); anterior setae present on 324 236 237 Figs 236, 237 Neotrichodectes species, $ terminalia. 236, N. gastrodes. 237, N. pallidus. pleura II and HI only; postero-lateral setae presumed absent, though possibly present as the most lateralseta of lateral group, which is frequently situated more posteriorly than other setae. Abdominal pleuralacking projections, except for small sclerotised or unsclerotised projection dorsally on pleurum III offemale N. mephltidis. Abdominal terga and sterna lacking sclerites, except for lateral rods of subgenitalplate in male and, sometimes, tergite IX in female; abdominal pleura usually unsclerotised, sometimeswith sclerites on pleura II, III, IV and, in females, VIII*. Gonapophyses with or without lobe on ventral margin, very variable*. Gonapophyses meet ventralvulval margin smoothly or acutely, but not linked by sclerotised band. Ventral vulval margin notsclerotised; generally convex medially*, subgenital lobe present or absent*. Common oviduct not notablystriate (cf. Geomydoecus}. Male subgenital plate with only s.g.p.r. present (Fig. 229). Pseudostyli absent. Male genital opening TRICHODECTID MAMMAL LICE 325 240 Figs 238-242 Neotrichodectes species, C? genitalia. 238, N. (Neotrichodectes) mephitidis. 239, N.(Nasuicola) pallidus. 240, N. (Trigonodectes) barbarae. 241, N. (Lakshminarayanella) gastrodes. 242,N. (Conepaticola) chilensis. dorsal. Parameres fused to form single plate with apex pointed or bifid*. Basiparameral sclerites absent.Mesomeres fused apically; mesomeral arch with median extension*; mesomeres basally abut postero-lateral projections of b.a.l.s. but do not contact parameral plate (Fig. 242), or contact neither b.a.l.s. norparameral plate (Fig. 240). HOSTS. Mustelidae and Procyonidae (Carnivora) and Bradypodidae (Edentata). COMMENTS. Neotrichodectes has been considered a synonym and a subgenus of Trichodectes (by Hopkins,1942 and Hopkins, 1949 respectively); a more complete history of the variations in status of Neo-trichodectes is presented in Table 4. 326 C. H. C. LYAL Subgenus NEOTRICHODECTES Ewing (Figs 228, 229, 232, 238)Neotrichodectes Ewing, 1929: 194. Type-species: Goniodes mephitidis Packard, by original designation. DESCRIPTION. Anterior of head with clypeal marginal carina broadened into dorsal sclerite which is more orless convex posteriorly, sometimes almost circular; preantennal portion of head sometimes as long aspostantennal portion. Male flagellum with two basally-articulated 'teeth' distally, and with toothedprojection sometimes present basally (Fig. 231). Temple with long seta present on posterior margin. Thoracic and abdominal setae long, abundant. Abdominal pleura lacking projections, except for smallsclerotised or unsclerotised projection dorsally on pleurum III of female N. mephitidis (Fig. 228).Abdomen lacking sternal, tergal and pleural sclerites, except for lateral rods of male subgenital plate. Gonapophyses broad, membranous, with ventral marginal setae absent or, if present, basal only (Fig.232); ventral lobe absent. Ventral vulval margin with lobe present though difficult to see; lobe serrate, atleast along posterior margin (Fig. 232). Female genital chamber with dorsal wall bearing slanting scaleslaterally, lightly sclerotised and lacking scales or other decoration medially (anteriorly). Parameral plate slender, apically bifid, with median basal extension reaching anteriorly between b.a.l.s.Mesomeral arch with median extension pointed; mesomeres basally abut postero-lateral extensions ofb.a.l.s. Male genitalia depicted in Fig. 238. HOSTS. Mustelidae and Procyonidae (Carnivora). COMMENTS. Neotrichodectes s. str. is most readily distinguished from other subgenera of Neotrichodectes bycharacters of the female terminalia: the retention of the plesiomorphic form of the gonapophyses (foundalso in Geomydoecus) and the apomorphic development of a membranous subgenital lobe (cf . the verydifferent structure in N. (Trigonodectes)). Assignment of male insects to the subgenus relies on absence ofthe basal flagellar projection (a character-state reversal not undergone by the type-species of thesubgenus), the presence of an anterior development of the parameral plate between the b.a.l.s., and theslenderness of that parameral plate in relation to its length. Neotrichodectes wolffhuegeli (Werneck) isknown only from the male, although Werneck (1948) predicted that the female would be very similar tothat of N. chilensis (placed in N. (Conepaticola) in this study), and so must be assigned to subgenus on thebasis of male characters. N. wolffhuegeli does have a toothed projection on the base of the male flagellum,although, as indicated above, cannot be eliminated from Neotrichodectes s. str. on that basis. The form ofthe parameral plate of N. wolffhuegeli is much the same as members of Neotrichodectes s. str. and on thisevidence the species is placed in the subgenus. Emerson (pers. comm.), however, suggests that the malegenitalia of N. wolffhuegeli lie within the limits of permissible variation of N. chilensis, which he thereforeconsiders as a junior synonym; N. wolffhuegeli and N. chilensis are found on the same host, although N.chilensis is also found on several other species of the host genus. All specimens identified as N. wolffhuegeli(by Werneck and in the present study) have a much narrower parameral plate than those identified as N.chilensis. However, the parameral plate, whilst not extending anteriad between the b.a.l.s. in mostspecimens ofN. chilensis, does do so in some. In those specimens of TV. chilensis where there is no anteriorextension of the parameral plate the endophallus lacks large, heavily-sclerotised teeth, or such teeth arefew in number; in those where the extension is present the teeth are correspondingly more developed. Thedegree of development of the endophallus teeth seems to be proportional to the degree of development ofthe anterior margin of the parameral plate in N. chilensis and such teeth are present and well developed inN. wolffhuegeli. The sample of specimens of both species was too small to permit any correlations ofgenitalia type with host species or geographical distribution, though within N. chilensis specimensexhibiting both extremes were found from the same host in the same area. For the purposes of this study,the two species are treated as separate, N. wolffhuegeli is assigned to Neotrichodectes s. str., and N.chilensis is assigned to Neotrichodectes (Conepaticola). This conclusion is regarded as the most satisfactoryfor the data presently available, but further collecting from species of the host genus (Conepatus) is neededto clarify the situation. The original spelling of the species name wolffhugeli is emended here to wolffhuegeli in accordance withArticle 32 (d)(i) of the International Code of Zoological Nomenclature (1984). SPECIES INCLUDED mephitidis (Packard, 1873) (c.50 cf , c.100 $)minutus (Paine, 1912) (64 cf , 76 $)osftorn/Keler, 1944 (16 cf , 21 $)f/ioraciciis(Osborn, 1902) (11 cf, 9 $)wolffhuegeli ( Werneck, 1936) (1 cf) TRICHODECTID MAMMAL LICE 327 Subgenus TRIGONODECTES Keler gen. rev., stat. n. (Figs 227, 230, 235, 240) Trigonodectes Keler, 1944: 179, 185. Type-species: Trichodectes barbarae Neumann, by original designa-tion. DESCRIPTION. Anterior of head with osculum very shallow; clypeal marginal carina broadened mediallyinto dorsal U-shaped sclerite with median posterior process; anterolateral margin of head convex;preantennal portion of head with outline rounded. Male flagellum with two basally-articulated 'teeth'distally, and with toothed projection present basally. Temple with long seta present on posterior margin. Thoracic and abdominal setae long, abundant. Abdominal pleura lacking projections. Abdomen lackingsternal, tergal and pleural sclerites, except for lateral rods of male subgenital plate. Gonapophyses slender, apically acute, sclerotised, with strong setae present along ventral margin;ventral lobe absent. Gonapophyses meet ventral vulval margin acutely. Ventral vulval margin with medianlobulate projection, with margin not serrate, and submarginal setal row present (Fig. 235). Female genitalchamber lacking lateral slanting scales and anterior sclerotised area on dorsal wall, but both ventral anddorsal walls bearing numerous scales with posterior spinules. Parameral plate triangular or shield-shaped, pointed apically. Mesomeral arch with median extensionrounded, covered in small tubercles; mesomeres extend basally anterior to ends of b.a.l.s., and do notcontact b.a.l.s. or parameral plate. Basal apodeme lacking postero-lateral projections on b.a.l.s. Malegenitalia depicted in Fig. 240. HOSTS. Mustelinae (Carnivora: Mustelidae). COMMENTS. Trigonodectes has been treated as a synonym and a subgenus of Trichodectes (by Werneck,1948 and Hopkins, 1949 respectively); in this study it is raised from synonymy with Trichodectes andbarbarae is placed for the first time in Neotrichodectes. A more complete history of the varying statusaccorded to Trigonodectes is presented in Table 4. SPECIES INCLUDEDbarbarae (Neumann, 1913) comb. n. from Trichodectes (13 C?, 13 $) Subgenus NASUICOLA subgen. n. (Figs 237, 239)Type-species: Trichodectes pallidus Piaget. DESCRIPTION. Anterior of head with osculum shallow; clypeal marginal carina broadened into dorsalrectangular sclerite; antero-lateral margin of head convex; preantennal portion of head not as long aspostantennal portion, outline broadly rounded. Male flagellum with two basally-articulated 'teeth' distally,and with toothed projection present basally. Temple with long seta present on posterior margin. Thoracic and abdominal setae long, abundant. Abdominal pleura lacking projections. Abdomen lackingsternal, tergal and pleural sclerites, except for lateral rods of male subgenital plate. Gonapophyses with rounded ventral lobe with submarginal setae; spur distal to lobe very short.Gonapophyses meet ventral vulval margin acutely. Ventral vulval margin convex, but subgenital lobe orlobulate process not present. Female terminalia depicted in Fig. 237. Female genital chamber with dorsalwall bearing slanting scales laterally, spines medially (spines most apparent anteriorly, though may beobscured). Parameral plate apically bifid, lacking median basal extension reaching anteriorly between b.a.l.s.Mesomeral arch with median extension pointed; mesomeres basally abut postero-lateral extensions ofb.a.l.s. Male genitalia depicted in Fig. 239. HOSTS. Procyonidae (Carnivora). SPECIES INCLUDEDpallidus (Piaget, 1880) [treated by previous authors as Neotrichodectes s. str.] (65 cf , 75 $) Subgenus LAKSHMINARAYANELLA Eichler stat. n.(Figs 236, 241) Lymeon Eichler, 1940: 158. Type-species: Trichodectes gostrodesCummings, by monotypy. [Homonym of Lymeon Foerster, 1868: 176.]Lakshminarayanella Eichler, 1982: 83. [Replacement name for Lymeon Eichler.] 328 C. H. C. LYAL DESCRIPTION. Anterior of head with osculum present, deep; clypeal marginal carina broadened mediallyinto posteriorly convex bar, parallel to curvature of osculum; antero-lateral margin of head straight orslightly sinuate; preantennal portion of head short in male, longer in female; outline broadly triangular.Male flagellum with seven basally-articulated 'teeth' distally, and with roughened projection presentbasally. Temple with no long setae present on posterior margin. Thorax with dorsal setae of moderate length, longest on postero-lateral angle of pterothorax; ptero-thorax with setae sparse along posterior dorsal margin. Abdominal setae of moderate length, not as long as segment except on posterior pleura; terga withlateral and median setal groups not clearly distinct, median gap sometimes pronounced. Abdominal pleuralacking projections. Abdomen lacking tergal and sternal sclerites except for lateral rods of male subgenitalplate and tergite IX of female; abdominal pleura II, III, IV and, in female, VIII, with sclerites, though thatof IV sometimes very small. Gonapophyses broad, very thick; ventral lobe present, thick, with setae along posterior margin; spurdistal to lobe not short, rounded apically (Fig. 236). Gonapophyses meet ventral vulval margin acutely.Ventral vulval margin convex, but subgenital lobe or lobulate process not present. Female genital chamberwith dorsal wall bearing slanting scales laterally, spines medially (spines most apparent anteriorly). Parameral plate apically bifid, sometimes projecting slightly anteriad between b.a.l.s. Mesomeral archwith median extension pointed, broad basally; mesomeres basally abut postero-lateral extensions olb.a.l.s. Male genitalia depicted in Fig. 241. HOSTS. Bradypodidae (Edentata). SPECIES INCLUDED cummingsi (Eichler, 1943) comb. n. from Lakshminarayanella gastrodes (Cummings, 1916) comb. n. from Lakshminarayanella (4 cf, 6 $ , 4 nymphs) Subgenus CONEPATICOLA subgen. n. (Figs 231, 233, 234, 242)Type-species: Neotrichodectes semistriatus Emerson & Price. DESCRIPTION. Anterior of head with clypeal marginal carina broadened medially into dorsal U-shapecsclerite with median posterior process; antero-lateral margin of head convex or sinuate; preantennaportion of head not as long as postantennal portion, outline broadly rounded. Male flagellum with twobasally-articulated 'teeth' distally, and with toothed projection basally. Temple with long seta present onposterior margin. Thoracic and abdominal setae long, abundant. Abdominal pleura lacking projections. Abdomen lackingsternal and tergal sclerites, except for sternal rods of male subgenital plate and tergite IX of female;abdominal pleura II, III and VIII in female sometimes with sclerites, though that of III may be very small. Gonapophyses with ventral lobe present, large, apparently comprising fused setal tubercles, with setaealong posterior margin and anterior margin, the latter frequently directed posteriad; spur distal to lobepresent, not short, frequently obtuse apically. Gonapophyses meet ventral vulval margin acutely. Ventralvulval margin convex, but subgenital lobe or lobulate process not present. Female terminalia depicted inFigs 233, 234. Female genital chamber with dorsal wall bearing slanting scales laterally, spines medially,sometimes lightly sclerotised and lacking scales, spines or other decoration antero-medially. Parameral plate apically bifid, sometimes projecting slightly anteriad between b.a.l.s. Mesomeral archwith median extension pointed; mesomeres basally abut postero-lateral extension of b.a.l.s. Male genitaliadepicted in Fig. 242. HOSTS. Mustelinae and Mephitinae (Carnivora: Mustelidae). SPECIES INCLUDED arizonaeWerneck, 1948 (18 C?, 39 $) cMens/sWerneck, 1948 (c.50 C?, c.50 $) interruptofasciatus (Kellogg & Ferris, 1915) (27 cf, 46 $) semjsfriatus Emerson & Price, 1976 (5 cf , 5 $)All the above species have been treated previously as Neotrichodectes s. str. Genus GEOMYDOECUSEwing The genus Geomydoecus comprises two subgenera. TRICHODECTID MAMMAL LICE 329 DESCRIPTION. Anterior of head with osculum present; dorsal preantennal sulcus present; clypeal marginalcarina broadened medially into posteriorly convex bar; antero-lateral margin of head convex or sinuate;preantennal portion of head not long, outline broadly triangular or rounded. Temple margin broadlyconvex. Male scape expanded; longitudinal setal row present and comprising at least three setae; malescape sometimes with median posterior projection; flagellomeres fused in males and females; maleflagellum with two basally-articulated 'teeth'; female pedicel with membranous postero-ventral projection(Fig. 245), sometimes obscure. Dorsum of head with setae short or of moderate length, more or less sparse;temple margin sometimes with specialised long, fine or short and stout latero-posterior setae*. Sitophoresclerite unmodified. 244 245 Figs 243-245 Geomydoecus species. 243, G. (G. ) calif ornicus , $ terminalia. 244, G. (Thomomydoecus)asymmetricus , cf anterior of abdomen, dorsal. 245, G. (T.)pattoni, $ antenna. 330 C. H. C. LYAL Fig. 246 Geornydoecus (Thomomydoecus) minor, 9 abdomen. Hatched areas indicate damage tospecimen. Thorax with dorsal setae long or of moderate length; posterior margin of pronotum with four setae andwide median gap, posterior margin of pterothorax dorsally with varying number of marginal or submargin-al setae. Abdomen broadly rounded or more elongate and tapered, particularly in male. Abdominal spiraclesabsent. Abdomen with at least some setae as long as segment; setae generally abundant; terga, especiallyof males, with median and lateral setal groups distinct, though median groups generally united; male tergaII- VI without median setae shorter than others; male terga II and III sometimes with median groupcomprising exceptionally long, stout setae* (Fig. 244); anterior setae present on pleura II and III only;postero-lateral setae sometimes clearly present (Fig. 244), otherwise obscured, though may be present asmost lateral seta of lateral group, which is frequently situated more posteriorly than other setae.Abdominal pleura with projections dorsally on pleura II, III, IV and, at least in female, ventrally on IV,sclerotised (Fig. 246) or unsclerotised*; projections generally more apparent in females than males.Abdominal terga and sterna lacking sclerites, except for lateral rods of male subgenital plate and,sometimes, terga II-IV of male*; abdominal pleura II and sometimes III and IV sclerotised, at least infemale; other pleura unsclerotised. TRICHODECTID MAMMAL LICE 331 249 248 Figs 247-250 Geomydoecus species, cf genitalia. 247, G. (G. ) thomomyus. 248, G. (G.) actuosi. 249, G.(Thomomydoecus) minor. 250, G. (T.) wardi. 332 C. H. C. LYAL Gonapophyses broad, membranous, with ventral marginal setae, if present, generally basal only; ventrallobe absent (Figs 243, 246). Gonapophyses meet ventral vulval margin smoothly or acutely, but not linkedby sclerotised band. Ventral vulval margin not sclerotised; generally convex or very convex medially;subgenital lobe not present. Female genital chamber with dorsal wall bearing slanting scales laterally,lightly sclerotised and lacking scales or other decoration antero-medially. Common oviduct generally withdistinct striae. Male subgenital plate with only s.g.p.r. present. Pseudostyli absent. Male genital opening dorsal.Parameres fused to form single plate with apex pointed or bifid*. Basiparameral sclerites absent.Mesomeres fused apically; mesomeral arch with or without median extension; mesomeres basally abutb.a.l.s., which sometimes have postero-lateral extensions; mesomeres do not contact parameral plate.Male genitalia depicted in Figs 247-250. HOSTS. Geomyidae (Rodentia). COMMENTS. A few of the species are parthenogenetic. No more than 25 specimens of most species were examined during the course of this study, though inmany cases large numbers were available. Detailed descriptions of all species of Geomydoecus s.l. and a phenetic treatment of the genus may befound in Hellenthal & Price (1976, 1980), Price (1974, 1975), Price & Emerson (1971, 1972), Price &Hellenthal (1975a, 19756, 1976, 1979, 1980a, 19806, 1980c, 1981a, 19816), Price & Timm (1979), Timm &Price (1979, 1980). Subgenus GEOMYDOECUS Ewing (Figs 243, 247, 248)Geomydoecus Ewing, 1929: 193. Type-species: Trichodectes geomydis Osborn, by original designation. DESCRIPTION. Temple margin sometimes with two short, stout setae latero-posteriorly, or single long, fineseta latero-posteriorly. Male abdominal terga II and III only rarely with median setal group comprising exceptionally long, stoutsetae (G. copei). Pleural projections rarely sclerotised. Male terga II-IV not sclerotised. Male genitalia not asymmetric. Parameral plate apically pointed or bifid. Male genitalia depicted in Figs247, 248. HOSTS. Geomyidae (Rodentia). SPECIES INCLUDED actuosi Price & Hellenthal, 1981 (25 cf , 25 9)albati Price & Hellenthal, 1981 (25 cf , 25 9)alcorni Price & Emerson, 1971 (6 cf , 6 $)aUeni Price & Emerson, 1971 (2 cf , 6 $)angularis Price & Hellenthal, 1981 (25 cf , 25 9)aureiaurei Price & Hellenthal, 1981 (25 cf , 25 $)aurei grahamensis Price & Hellenthal, 1981 (25 (?, 25 9)bajaiensis Price & Hellenthal, 1981 (25 cf , 25 9)bulleri Price & Emerson, 1971 (25 cf , 25 $)californicus (Chapman, 1897) (25 cf , 25 $)centra/is Price & Hellenthal, 1981 (25 cf , 25 9)chapini Werneck, 1945 (12 cf , 16 $)cherriei Price, 1974 chiapensis Price & Emerson, 1971 (18 cf , 20 $)chihuahuae chihuahuae Price & Hellenthal, 1979 (25 cf , 25 9)chihuahuae emersoni Price & Hellenthal, 1979 (25 cf , 17 9)clausonae Price & Hellenthal, 1981 (25 cf , 25 9)dittooi Price & Hellenthal, 1981 (25 cf , 25 9)copei Werneck, 1945 (25 cf , 25 9)corojiado/ Barrera, 1961 (25 cf , 25 9)costaricensis Price & Emerson, 1971 (7 cf, 7 9)crovelloi Price & Hellenthal, 1981 (25 cf , 25 9)dakotensis Price & Emerson, 1971 (25 cf , 25 9)dalgleishiTimm & Price, 1979 (2 cf , 2 9) TRICHODECTID MAMMAL LICE 333 dariensis Price & Emerson, 1971 (10 cf , 14 9)duchesnensis Price & Emerson, 1971 (5 cf , 8 9)ewingi Price & Emerson, 1971 (25 cf , 25 9)expansus(Duges, 1902) (25 cf , 25 $)art/in/ Price & Hellenthal, 1981 (26 cf , 25 9)fulvescens Price & Emerson, 1971 (25 cf , 25 <j>)fu/vi Price & Hellenthal, 1979 (25 cf , 25 9)geomydis(Osborn, 1891) (25 cf, 25 $)guadalupensis Hellenthal & Price, 1980 (25 cf , 25 9)AeanejiTimm & Price, 1980 (25 cf , 25 ?)hoffmanni Price & Hellenthal, 1976 (25 cf , 25 9)/Hieji Price & Hellenthal, 1980 (25 cf , 25 9)idahoensis Price & Emerson, 1971 (25 cf , 25 9)illinoensis Price & Emerson, 1971 (25 cf , 25 9)jaliscoensis Price & Hellenthal, 1981 (25 cf , 25 9)joaesi Price & Emerson, 1971 (4 cf , 5 9)Hmitaris limitaris Price & Hellenthal, 1981 (25 cf , 25 9)limitaris bakeri Price & Hellenthal, 1981 (25 cf , 25 9)limitaris halli Price & Hellenthal, 1981 (25 cf , 25 9)limitaris tolteci Price & Hellenthal, 1981 (25 cf , 25 9)mart/ni Price & Hellenthal, 1975 (25 cf , 25 9)mcgregori Price & Emerson, 1971 (25 cf , 25 9)merriami Price & Emerson, 1971 (25 cf , 25 9)mexicanus Price & Emerson, 1971 (25 cf , 25 $)mobilensisPrice, 1975 (25 9)jmiscufi Price & Hellenthal, 1981 (25 cf , 25 9)nayaritensis Price & Hellenthal, 1981 (16 cf , 25 9)nebrathkensisTimm & Price, 1980 (25 cf , 25 9)oklahomensis Price & Emerson, 1971 (25 cf , 25 9)oregonus Price & Emerson, 1971 (25 cf , 25 9)panamensis Price & Emerson, 1971 (23 cf , 22 9)pattoni Price & Hellenthal, 1979 (12 cf , 8 9)perotensis perotensis Price & Emerson, 1971 (25 cf , 25 9)perotensis irolonis Price & Emerson, 1971 (25 cf , 25 9)polydentatus Price & Emerson, 1971 (25 cf , 25 9)quadridentatus Price & Emerson, 1971 (25 cf , 25 9)scleritus (McGregor, 1917) (3 cf , 25 9)setzeri Price, 1974 (6 cf , 8 9)shastensis Price & Hellenthal, 1980 (25 cf , 25 9)sinaloae Price & Hellenthal, 1981 (25 cf , 25 9)sp/dta/Timm & Price, 1980 (25 cf , 25 9)subcalifornicus Price & Emerson, 1971 (25 cf , 25 9)subgeomydis Price & Emerson, 1971 (25 cf , 25 9)subnuhili Price & Hellenthal, 1975 (25 cf , 25 9)tamaulipensis Price & Hellenthal, 1975 (3 cf , 25 9)texanustexanusEv/ing, 1936 (25 cf , 25 9)texanus tropicalis Price & Hellenthal, 1975 (25 cf , 25 9)thomomyus (McGregor, 1917) (25 cf , 25 9)tolucae Price & Emerson, 1971 (25 cf , 25 9)traubi Price & Emerson, 1971 (25 cf , 25 9)trichopi Price & Emerson, 1971 (25 cf , 25 9)truncatus Werneck, 1950 (25 cf , 25 9)umbrini Price & Emerson, 1971 (25 cf , 25 9)ustulati ustulatiPrice & Hellenthal, 1975 (25 cf , 25 9)ustulati clarkii Price & Hellenthal, 1975 (25 cf , 25 9)veracruzensis Price & Emerson, 1971 (25 cf , 25 9)warmanae Price & Hellenthal, 1981 (25 cf , 25 $)welleri w?77eri Price & Hellenthal, 1981 (25 cf , 25 9)wellerimultilineat us Price & Hellenthal, 1981 (25 cf , 25 9) 334 C. H. C. LYAL werneckl Price & Emerson, 1971 (25 cf , 25 $)yucatanensis Price & Emerson, 1971 (25 cf , 25 $) Subgenus THOMOMYDOECUS Price & Emerson(Figs 244-246, 249, 250) Thomomydoecus Price & Emerson, 1972: 464 [as subgenus of Geomydoecus Ewing]. Type-species:Geomydoecus (Thomomydoecus) jamesbeeri Price & Emerson, by original designation. DESCRIPTION. Temple margin with single stout seta and finer, shorter adjacent setae latero-posteriorly. Male abdominal terga II and III with median setal group comprising exceptionally long, stout setae.Pleural projections sclerotised, at least in female (Fig. 246). Male terga II-IV sometimes with sclerites. Gonapophyses meet vulval margin smoothly. Male genitalia symmetric (Fig. 250) or asymmetric (Fig. 249). Parameral plate apically pointed. HOSTS. Thomomyusspp. (Rodentia: Geomyidae). COMMENTS. As discussed on p. 232, the subgenus Geomydoecus is probably paraphyletic and Thomomy-doecus paraphyletic or even polyphyletic. Geomydoecus (Thomomydoecus) was raised to generic status byHellenthal & Price (1984), on the basis that there are 'sufficient' morphological differences between it andGeomydoecus s. str. This action, reversed here, is consistent with the purely phenetic approach employedby the authors, but cannot be reconciled with the cladistic methods used in this study. The division ofGeomydoecus into subgenera is retained, even though neither of the two subgenera is 'natural' (i.e.holophyletic), because insufficient work has been done to resolve the relationships properly, and theindependent unit of the classification, the genus Geomydoecus, is holophyletic. Raising Thomomydoecusto generic status, however, would produce two independent units in the classification (Geomydoecus andThomomydoecus) that differ from all the others in not being holophyletic. Geomydoecus (T.) byersi (Hellenthal & Price) was described after the completion of this manuscript andhas not been included in the cladistic analysis. SPECIES INCLUDED asymmetricus Price & Hellenthal, 1980 (25 cf , 25 $)birneyi Price & Hellenthal, 1980 (25 cf , 25 $)byersi (Hellenthal & Price, 1984) comb. n. from Thomomydoecusdickermani Price & Emerson, 1972 (25 cf , 25 $)genowaysi Price & Emerson, 1972 (25 cf , 25 $)green Price & Hellenthal, 1980 (2 cf )jamesbeeri Price & Emerson, 1972 (8 cT, 10 $)johnhafneri Price & Hellenthal, 1980 (25 cf , 25 $)markhafneri Price & Hellenthal, 1980 (25 cf , 25 $)minor Werneck, 1950 (25 cf , 25 ?)neocopei Price & Emerson, 1971 (2 cf , 1 9)orizabae Price & Hellenthal, 1980 (10 cf , 26 $)peregrin! Price & Hellenthal, 1980 (4 cf , 4 $)potteri Price & Hellenthal, 1980 (16 cf , 25 $)timmi Price & Hellenthal, 1980 (25 cf , 25 $)wardi Price & Emerson, 1971 (25 cf , 25 $)williamsiPrice & Hellenthal, 1980 (8 cf , 16 $)zacatecae Price & Hellenthal, 1980 (25 cf , 25 $) Keys to Trichodectidae Two keys are provided: a key to subfamilies and a key to genera and subgenera. The latter keycontains all genera and subgenera, and it is not necessary to use the subfamily key as anintroduction to it. The subfamily key is included because of the formal requirement that any newly describedtaxon must be accompanied by some form of description in order to make the name available.The key to the five subfamilies of Trichodectidae therefore serves to distinguish Neotrichodecti-nae subfam. n. from all others. TRICHODECTID MAMMAL LICE 335 Key to subfamilies 1 No abdominal spiracles present; majority of tergal and sternal setae at least two-thirds length of segment or, if not, median setal group on tergum II comprising at least three setae (and,frequently, median groups running together). New World NEOTRICHODECTINAE subfam. n.(p. 321) - At least one pair of abdominal spiracles present or, if not, majority of abdominal sternal and tergal setae less than two-thirds length of segment or median setal group on tergum IIcomprising only one seta. Old and New World 2 2 Female subgenital lobe present, frequently with serrate margin, at least posteriorly; if margin of subgenital lobe smooth, gonapophyses meet vulval margin smoothly (Fig. 153); femaleflagellomeres fused; abdominal spiracles numbering six or fewer pairs. Parasitic on Carni-vora and Primates. Old and New World TRICHODECTINAE Kellogg, 1896 (p. 286) - Female subgenital lobe absent or, if present, not marginally serrate and gonapophyses meet vulval margin acutely; female flagellomeres fused or unfused; abdominal spiracles numberingsix pairs, though spiracles on segment VIII may be very small and inconspicuous, possiblynon-functional (some species of Procaviphilus (Meganarionoides)). Not parasitic on Carni-vora. Old and New World 3 3 Dorsal or ventral projection present on abdominal pleurum IV; mesomeral arch generally produced basally between b.a.l.s.; female antennal flagellomeres generally not fused, oronly partially fused; parasitic on hyraxes and primates. Old and New World DASYONYGINAE Keler, 1938 (p. 267) - Pleurum IV lacking any projection; mesomeral arch rarely produced basally between b.a.l.s. ; female flagellomeres generally fused or, if not, then female with long setal tufts onabdominal pleura VIII and IX; not parasitic on hyraxes or primates 4 4 Posterior margin of temple generally produced, very convex; (Fig. 87) ; very long setae present on at least pleurum VIII, sometimes also on pleura VII (male) or IX (female); basiparameralsclerites present; mesomeral arch lacking extension if complete, otherwise tripartite, medianpart sometimes obscure (Figs 91, 93); if mesomeral arch entire, male genitalia as in Fig. 89,temple margins not greatly produced, and female with two flagellomeres; parasitic on NewWorld porcupines (Erethizontidae) EUTRICHOPHILINAE Keler, 1938 (p. 265) - Temples not so developed; setae on pleurum VIII not exceptionally long; basiparameral sclerite present or absent; mesomeral arch, if present, with or without extension, but archnever tripartite; female flagellomeres fused; pseudostyli frequently present; parasitic onArtiodactyla and Perissodactyla. Old and New World BOVICOLINAE Keler, 1938 (p. 247) Key to genera and subgenera 1 No abdominal spiracles present ; maj ority of tergal and sternal setae at least two-thirds length of segment or, if not, median setal group on tergum II comprising at least three setae (and,frequently, median groups running together) . New World 2 - At least one pair of abdominal spiracles present or, if not, majority of abdominal sternal and tergal setae less than two-thirds length of segment or median setal group on tergum IIcomprising only one seta. Old and New World 8 2 Abdominal pleura II-IV with dorsal projections (Fig. 246), though most apparent in females and sometimes very inconspicuous; male lacking tergocentral microsetae; latero-posteriorcorner of temple margin frequently with single long fine seta or one or two shorter, stoutsetae; female pedicel with dorsal membranous projection (Fig. 245) (sometimes obscure).[Geomvidae](GOAfyDQEO75s.l.,p.328) 3 - Abdominal pleura lacking dorsal projections, or single membranous projection present on pleurum IV only (Fig. 228); male with tergocentral microsetae on abdominal terga II-VI(Fig. 227): long seta frequently present on temple margin but shorter stout seta notdeveloped; female pedicel lacking any projection. (NEOTRICHODECTESs.l. , p. 321) 4 3 Pleural projections on pleurum II sclerotised; temple margin with single stout seta and associated smaller finer setae; male abdominal terga II and III with rows of enlarged setae(Fig. 244); parameral plate with single apical point; male genitalia symmetric or asymmetric GEOMYDOECUS (THOMOMYDOECUS)(p. 334)Pleural projection on pleurum II unsclerotised or, if sclerotised, posterolateral temple margin 336 C. H. C. LYAL with single long fine seta and associated smaller setae; temple margin with or withoutspecialised setae but not with single stout seta; male abdominal terga II and III rarely withrows of specialised setae (G. copei only); parameral plate with single apical point or apicallybifid; male genitalia symmetric GEOMYDOECUS (GEOMYDOECUS)(p. 332) 4 Female subgenital lobe present, with serrate margins; female genital chamber with clear, flat dorsal region but lacking single scattered spines; gonapophyses broad, membranous, lackinglobe; parameral plate slender, with basal projection between b.a.l.s. (Fig. 238). [Mustelidae and Procyonidae] NEOTRICHODECTES (NEOTRICHODECTES) (p. 326) Female subgenital lobe absent or, if present, lobe with smooth margins and longitudinal setalrows (Fig. 235); female genital chamber, if with clear flat dorsal area, then with single spinesscattered over it; gonapophyses not broad and membranous, frequently with lobe; paramer-al plate broad, with very limited projection between b.a.l.s 5 5 Ventral vulval margin with lobulate process with smooth margins and longitudinal rows of setae (Fig. 235) ; gonapophyses slender, sclerotised, lacking lobe ; parameral plate with singleapical point; mesomeral arch extension broad, clubbed (Fig. 240). [Mustelidae] NEOTRICHODECTES (TRIGONODECTES)(p. 327) - Ventral vulval margin convex, but not produced; gonapophyses not slender, lobe present; parameral plate with apex bifid ; mesomeral arch with pointed extension 6 6 Large species, over 2-75 mm long; male flagellum with 7 articulated 'teeth'; female gonapo- physes thick, with lobe and spur (Fig. 236); female pleurum VIII sclerotised; abdominalsetae relatively small, not attaining following setal row. [Bradypodidae] NEOTRICHODECTES (LAKSHMINARAYANELLA)(p. 327) Smaller species, under 2-25 mm long; male flagellum with 2 articulated 'teeth'; femalegonapophyses otherwise; female pleurum VIII not sclerotised; abdominal setae long,attaining or nearly attaining setal bases of following setal row. [Carnivora] 7 7 Gonapophyses with flat lobe and small spur (Fig. 237); male mesomeral arch extension attaining end of parameral plate (Fig. 239) [Procyonidae] NEOTRICHODECTES (NASUICOLA)(p. 327) - Gonapophyses with lobe comprising fused setal tubercles, long spur present (Fig. 233); male mesomeral arch extension reaching beyond apex of parameral plate (Fig. 242). [Mustelidae] NEOTRICHODECTES (CONEPATICOLA)(p. 328) 8 Five pairs of abdominal spiracles present; vulval margin sclerotised, with or without setal tubercles, and meeting gonapophyses smoothly; subgenital lobe present; parameres notfused to b.a.l.s.; mesomeres absent; postcoxale absent; abdominal segments II-V withmedian setal group present, comprising at least three setae. [Mustelidae] TRICHODECTES (PARATRICHODECTES)(p. 300) Other than five pairs of abdominal spiracles present, though spiracles on segment VIII may bevery small, inconspicuous and possibly non-functional (some species of Procaviphilus(Meganarionoides) as described in key couplet 26, and some species of Trichodectes(Stachiella) , as described in key couplet 14) 9 9 Abdominal pleura V-VI (at least) lacking setae 10 Abdominal pleura III-VIII (at least) with posterior setal row and , sometimes , anterior setae ... 12 10 Abdominal tergal setae on segments I-VI less than half length of segment, shorter than postero-lateral setae; pleura V-VI lacking setae (Fig. 159); male flagellum with twobasally-articulated 'teeth'; mesomeres present, unfused; parameres fused, with distinctinturned apices arising from plate (Fig. 171); subgenital lobe bifurcate, with long basallateral processes (Fig. 149). [Mustelidae] TRICHODECTES (TRICHODECTES) (in part) (p. 299) At least some setae on abdominal terga I-III as long or longer than segment and postero-lateralseta, and postero-lateral setae sometimes absent; pleura IV- VII (at least) lacking setae (Fig.141); male flagellum lacking 'teeth'; mesomeres absent; parameres unfused or united at baseonly; subgenital lobe not bifurcate or only slightly so; basal processes of subgenital lobeabsent or, if present, not long (Fig. 146). [Lutrinae] 11 11 Posterior setal row present on pleurum III; parameres slender, rod-like, fused basally (Figs 144, 145); subgenital lobe lacking basal lateral processes; gonapophyses lacking setal tubercles (Fig. 143) LUTRIDIA(p. 288) Posterior setal row not present on pleurum III; parameres broad, not fused to each other (Fig. TRICHODECTID MAMMAL LICE 337 147); subgenital lobe with basal lateral processes; gonapophyses with setal tubercles (Fig. 146) NEOLUTRIDIA(p. 289) Ventral vulval margin meets gonapophyses smoothly, joined by sclerotised band; subgenitallobe present, frequently with basal lateral processes; sternal setae on at least segmentsIII-VI attaining or nearly attaining base of following setal row; dorsum of head with setaesparse; male scape expanded or, if not, parameres fused to b.a.l.s 13 Ventral vulval margin meets gonapophyses acutely or, if meeting smoothly, not joined bysclerotised band; subgenital lobe present or absent, but, if present, never with basal lateralprocesses (except Damalinia (Tricholipeurus) elongata; see Fig. 65); sternal setae onsegments III-VI not attaining base of following setal row, usually less than three-quarterslength of segment or, if longer, either female genitalia not as described and dorsum of headwith dense setal covering (Bovicola (Holakartikos) and B. (Spinibovicola)) or male scapenot expanded and parameres not fused to b.a.l.s 15 Pleurum IV with dorsal projection; anterior setae present on abdominal terga and sterna. [Ursidae] WERNECKODECTES(p.29Q) Pleurum IV without dorsal projection; anterior setae not present on abdominal terga and sterna 14 Male abdominal terga II-IV (at least) with median setal group reduced to one seta (Fig. 164); parameres fused to b.a.l.s. (Figs 172, 174) or characteristically asymmetric (Fig. 173); female abdominal terga III- VII (at least) with median setal group reduced to one seta or absent. [Mustelidae and Procyonidae] TRICHODECTES (STACHIELLA) (p. 301) Tergal setae of both sexes more abundant, with at least two setae in median setal group; parameres not fused to-b.a.l.s. [Canidae, Viverridae, Ursidae and Mustelidae] TRICHODECTES (TRICHODECTES) (in part) (p. 299) Posterior setal row of pleurum III with setae stouter than those of p.s.r. of pleurum V (Figs 182,188) or , if not , species with four pairs of abdominal spiracles ; otherwise species with 0,1,2,3,4 or 6 pairs of abdominal spiracles ; subgenital lobe present ; gonapophyses with lobe present 1 6 Posterior setal row of pleurum III with setae not stouter than those of p.s.r. of pleurum V; sixpairs of abdominal spiracles present , if gonapophyses with lobe , then subgenital lobe absent 19 Abdominal spiracles numbering 6, 4 or pairs; if no abdominal spiracles present, then femalewith gonopore surrounded by spicular refringent patch, or gonapophysis lobe comprisingtwo fused tubercles, or antennal sensilla in pit with peripheral tongue-like projections;female subgenital lobe frequently with overlapping scales or spines; male mesomeral archalways present, with median extension or two apical nipples; mesomeres produced basallybetween b.a.l.s. or, if not, antennal sensilla as described above; parameres usually broad,contacting mesomeres only, not b.a.l.s.; male abdominal tergum II lacking specialised setaeof median group. (LORISICOLAs.L, p. 312) 17 Abdominal spiracles numbering 3, 2, 1 or pairs; if no abdominal spiracles present, thenfemale gonopore not surrounded by spicular refringent patch, gonapophysis lobe notcomprising two fused setal tubercles; antennal sensilla of male and female never in pit withperipheral tongue-like projections; female subgenital lobe never with overlapping scales orspines; male mesomeres fused, unfused or absent; if mesomeres fused, mesomeral archnever with median extension or apical nipples; parameres frequently narrow, rod-like,contacting mesomeres, b.a.l.s. or both; male abdominal tergum II frequently with long,specialised setae (Figs 179, 180, 187-189). (FELICOLAs.L, p. 302) 18 Male antennal flagellum with 'teeth' on projection; six pairs of abdominal spiracles present orabdominal spiracles absent, in which case antennal sensilla in pit with peripheral tongue-likeprojections. [Felidae, Viverridae and Lorisidae] LORISICOLA (LORISICOLA) (p. 316) Male antennal flagellum with 'teeth' not on projection, or, if projection present, mesomeralarch lacking extension; four pairs of abdominal spiracles present or abdominal spiraclesabsent, in which case male gonopore surrounded by spicular patch (Fig. 225). [ViverridaeandHerpestidae] LORISICOLA (PARADOXUROECUS)(p. 318) Male antennal flagellum with one or three basally-articulated 'teeth', or 'teeth' absent, in whichcase male abdominal tergum III with median setal group reduced to one seta of similar size tothose on tergum II, which are not greatly enlarged, and parameres not fused; femalegonapophysis with lobe and spur present, lobe rounded or rectangular and formed of fusedtubercles ; subgenital lobe bifid or not ; if subgenital lobe bifid , lobes pointed , rounded or with 338 C. H. C. LYAL rectangularly obtuse posterior margins (Fig. 176); everted portion of male endophallusnever sclerotised; abdominal spiracles numbering 0, 1 or 3 pairs. [Herpestidae and Canidae] FELICOLA (SURICATOECUS)(p. 312) Male antennal flagellum with one, two, three or four nonarticulated 'teeth', or 'teeth' absent, inwhich case male abdominal tergum III with median setal group reduced to one seta muchsmaller than those on tergum II, which are greatly enlarged, and parameres fused, at leastbasally; female with gonapophysis lobe rounded, with or without spur, but lobe neverrectangular; subgenital lobe bifid or not; if subgenital lobe bifid, lobes of various shapes, butnever with rectangularly obtuse posterior margins; everted portion of male endophallusfrequently thinly sclerotised (Figs 190, 196); abdominal spiracles numbering 0, 2 or 3 pairs.[Herpestidae, Viverridae and Felidae] FELICOLA (FELICOLA)(p. 302) 19 Dorsal x>r ventral projection present on abdominal pleurum IV (Fig. 105); mesomeral arch generally produced basally between b.a.l.s.; female antennal flagellomeres generally not fused, or only partially fused. [Procaviidae and Primates] 20 Pleurum IV lacking any projection; mesomeral arch rarely produced basally between b.a.l.s. ;female flagellomeres generally fused or, if not, then female with long setal tufts onabdominal pleura VIII and IX (see couplet 27) ; not parasitic on hyraxes or primates 27 20 Sitophore sclerite modified , with posterior arms extended (Fig. 12) (sclerite difficult to see) 21 - Sitophore sclerite unmodified (Fig. 1 1) (not, generally, difficult to see) 23 21 Tarsal claws with ventral teeth or spines; temple margin with or without small rounded projection; pleural projection on abdominal pleurum IV not elongate. [Procaviidae] (DASYONYXs.L,p.279) 22 Tarsal claws lacking ventral teeth or spines; temple margin with long, broad, triangularprojection (Fig. 138); pleural projection on abdominal pleurum IV long (Fig. 136). [Pro-caviidae] EURYTRICHODECTES(p.284) 22 Tarsal claws with sharp, fine spines (Fig. 14) DASYONYX (DASYONYX) (p. 282) Tarsal claws with broad, saw-like teeth (Fig. 15) DASYONYX (NEODASYONYX)(p. 284) 23 Abdominal sternum II with broad, heavily-sclerotised band articulated with abdominal pleurum II (Fig. 105); setal row of male scape comprising only two setae; basiparameral sclerites present. [Procaviidae] (PROCAVICOLAs.L, p. 270) 24 Abdominal sternum II lacking sclerotised band or, if sclerotised band present, this is fused toabdominal pleurum II or medially broken; setal row of male scape numbering more than twosetae; basiparameral sclerites absent or, if present, thoracic spiracle with tubular atrium andfemale flagellomeres fused 25 24 Posterior angle of temple with small projection ; mesomeral arch with lateral double flexion and median extension; endophallus with large hook-like spines (Fig. 107) PROCAVICOLA (CONDYLOCEPHALUS)(p. 274) Posterior angle of temple lacking projection; mesomeres unfused and lacking lateral doubleflexion and median extension; endophallus lacking large hook-like spines (Fig. 109) PROCAVICOLA (PROCAVICOLA)(p. 270) 25 Atrium of thoracic spiracle spherical; mesomeral arch with median extension and lateral desclerotisations; gonapophyses with setal tubercles or, if not, postcoxale greatly developedand fused to abdominal pleurum II. [Procaviidae and Cercopithecidae] (PROCAVIPHILUS s.l.,p.274) 26 Atrium of thoracic spiracle tubular; mesomeral arch lacking median extension and notdesclerotised laterally; gonapophyses lacking setal tubercles; postcoxale not greatly de-veloped and fused to abdominal pleurum II. [Cebidae] CEBIDICOLA (p. 267) 26 Parameres with basal flange, sometimes fused faintly; perisetal gap of male subgenital plate absent; postcoxale not fused to abdominal pleurum II; setal tubercles of gonapophyses not fused characteristically. [Procaviidae] PROCA VIPHILUS (PROCA VIPHILUS)(p. 278) Parameres usually lacking basal flange; perisetal gap of male subgenital plate present or, ifabsent, parameres fused together and articulated with mesomeral arch as in Fig. 120, andmesomeral arch produced basally along b.a.l.s. (Fig. 121); postcoxale fused to abdominalpleurum II, at least in females; setal tubercles of gonapophyses fused characteristically (Fig.Ill) or, if not, ventral vulval margin as in Fig. 113. [Procaviidae and Cercopithecidae] PROCAVIPHILUS (MEGANARIONOIDES)(p. 278) TRICHODECTID MAMMAL LICE 339 27 Posterior margins of temple generally produced, very convex (Fig. 87) ; very long setae present on at least pleurum VIII (Figs 88, 90), sometimes also on pleurum VII (males) or IX(female); basiparameral sclerites present; mesomeral arch lacking extension if complete,otherwise tripartite, median part sometimes obscure (Figs 101, 102); if mesomeral archentire, male genitalia as in Fig. 100, temples not greatly produced, and female with twoflagellomeres, otherwise female flagellomeres fused. [Erethizontidae] EUTRICHOPHILUS(p. 265) - Temples not so developed; setae on pleurum VIII not exceptionally long; basiparameral sclerites present or absent; mesomeral arch, if present, with or without extension, but nevertripartite; female flagellomeres fused 28 28 Parameres narrow , rod-like and fused basally ; mesomeral arch with broad lobe-like extension ; b.a.l.s. widely divergent anteriorly (Fig. 139); gonapophyses with setal tubercles; pseudo- styli absent. [Protelidae and Hyaenidae] PROTELICOLA(p. 286) - Parameres not narrow and fused basally or, if so, then b.a.l.s. not widely divergent anteriorly; mesomeral arch without broad lobulate extension; gonapophyses lacking setal tubercles;pseudostyli frequently present 29 29 Subgenital lobe present ; endophallus with dense patch of regularly-arranged spicules or , if not , parameral plate with single apex (Fig. 81); mesomeral arch entire, with abrupt bend toenable bases to meet parameres (Figs 81, 82), or mesomeres unfused and b.a.l.s. withanteposterior spur (Fig. 83); interior face of male flagellum serrate (Fig. 13); abdominalsterna never with anterior setae; long, slender species. [Bovidae and Cervidae] DAMALINIA (TRICHOLIPEURUS)(p. 264) - Subgenital lobe absent or, if present, as small flap (Fig. 42) and species with anterior setae on abdominal sterna (Bovicola jellisoni); sternum VII sometimes developed posteriorly intotwo projecting spikes (Damalinia theileri, neotheileri and semitheileri, Fig. 73); endophalluslacking spicular patch; parameres with apices free; mesomeres apically fused, unfused orabsent, but abrupt bend not present; b.a.l.s. lacking anteposterior spur; interior face of maleflagellum without serrations ; broader species 30 30 Dorsal face of vulva with pointed scales; gonapophyses hook-shaped (Fig. 66); common oviduct, at branching point, with folded and more or less apparent collar, sometimespartially sclerotised and refracting transmitted light; mesomeres unfused; abdominalpleurum never extending ventrally onto abdominal sternum II; interior face of male flagellum serrate. [Bovidae and Cervidae] DAMALINIA (CERVICOLA)(p. 263) Dorsal face of vulva lacking pointed scales; gonapophyses not hook-shaped or, if they are, thenabdominal pleurum II extending onto sternum II (Fig. 68); common oviduct lacking 'collar'as described above; mesomeres fused, unfused or absent; interior face of male flagellum withor without serrations 31 31 Abdominal pleurum II with sclerite extending onto sternum II and occasionally tergum II, sometimes at the expense of sternite or tergite (Fig. 68); mesomeres unfused, may be fusedto parameres and apparently absent; pseudostyli absent or, if present, broad or narrow (Figs68, 71, 72); interior face of male flagellum serrate. [Bovidae]DAMALINIA (DAMALINIA) (p. 260) - Abdominal pleurum II not extending onto sternum II; mesomeres fused, unfused or absent; pseudostyli , if present , not as figured above ; interior face of male flagellum lacking serrations 32 32 Atria of abdominal spiracles large, clear; mesomeral arch fused to b.a.l.s. ; parameres broad, asymmetrically deflected (Fig. 64) ; thorax with setae sparse dorsally in female, but male withmedian patch of setae on pro thorax; head elongate, trapezoid, with deep osculum present(Fig. 63) [Tragulidae] TRAGULICOLA(p. 255) - Atria of abdominal spiracles not large; mesomeral arch not fused to b.a.l.s. or, if it is, parameres and mesomeres also fused (Fig. 60); parameres not broad or asymmetricallydeflected; thoracic setae less sparse but male thorax lacking central setal patch; head notelongate but rounded , osculum absent or, if present, not deep (Fig. 3) 33 33 Parameres fused to mesomeres (Fig. 60); pseudostyli present, apically angular (Fig. 59); gonapophyses broad, truncate (Fig. 58); osculum absent, but anterior margin of head slightlyflattened or concave medially, with hyaline border where pulvinus attains margin. [Bovidae] BISONICOLA(p. 253) - Parameres not fused to mesomeres; pseudostyli, if present, apically rounded; gonapophyses 340 C. H. C. LYAL not broad, and with lobe variably apparent, or, if gonapophyses broad and truncate(Werneckiella), then pulvinus not attaining anterior margin of head, which is smoothlyrounded and lacks a median hyaline border 34 34 Gonapophyses broad, truncate; mesomeres of characteristic pentagonal form (Fig. 62). [EquidaeandBovidae] WERNECKIELLA (p. 255) Gonapophyses with more or less discrete lobe (Figs 42, 43); mesomeres, if present, notpentagonal; base of parameres frequently heavily block-like. [Bovidae, Cervidae andCamelidae](JJOV7COLAs.l.,p.247) 35 35 Species with more or less dense covering of long setae; anterior setae present on abdominal terga, sterna and pleura, slightly shorter than setae of posterior setal rows on these elements(Fig. 47) ; gonapophyses with very limited lobe formation (Fig. 44) 36 - Species with shorter setae or, if setae long, then sparsely distributed and anterior setae not present on abdominal terga and sterna; gonapophyses generally with more developed lobe(Figs43,45) 37 36 Sitophore sclerite with posterior arms extended (Fig. 12); male with specialised setae on abdominal tergum II (Fig. 41); mesomeres absent (Fig. 56); female lacks spinose patch onpostgenitalpleuralarea BOVICOLA (SPINIBOVICOLA)(p. 253) - Sitophore sclerite with posterior arms not extended; male lacking specialised setae on abdominal tergum II; mesomeres present, fused apically, with median extension (Fig. 57);female with spinose setal patch on postgenital pleural area BOVICOLA (HOLAKARTIKOS)(p. 251) 37 Setae on head long, fine; osculum absent; preantennal sulcus absent; gonapophyses with very distinct lobe of characteristic form (Fig. 43); male genitalia with mesomeres not fused,bipartite (Fig. 55) BOVICOLA (LEPIKENTRON)(p. 252) - Setae on head not long and fine; osculum present or absent; preantennal sulcus present or absent ; gonapophyses with less distinct lobe , of different form (Figs 42 , 45) ; male genitalia ofdifferent form, mesomeres never bipartite BOVICOLA (BOVICOLA) (p. 251) References Bedford, G. A. H. 1929. Anoplura (Siphunculata and Mallophaga) from South African Hosts. AnnualReport of the Director of Veterinary Services, Union of South Africa 15: 501-549. 1932o. Trichodectidae (Mallophaga) found on African Carnivora. Parasitology 24: 350-364. 19326. Trichodectidae (Mallophaga) parasitic on Procaviidae. Proceedings of the Zoological Society of London 3: 709-730. 1936. Notes on species of Trichodectidae with descriptions of new genera and species. Onderstepoort Journal of Veterinary Science and Animal Industry 7: 33-58.Blagoveshtchenski, D. I. 1956. Structure and taxonomic significance of the genitalia of Mallophaga. [In Russian] Parazitologjcheskiy Sbornik Zoologicheskogo Instituta Akademiya Nauk. S.S.S.R. 16: 1-88.Bouvier, G. 1945. De 1'hemophagie de quelques Mallophages des animaux domestiques. Schweizer Archiv fur Tierheilkunde 87: 429-434.Buckup, L. 1959. Der Kopf von Myrsidea cornicis (De Geer) (Mallophaga-Amblycera). Zoologische Jahrbiicher. (Anatomic) 77: 241-288. Burmeister, H. 1838. Handbuch der entomologie. 2 (2), Gymnognatha: 418-443.Chapman, R. F. 1982. The insects. Structure and function, xiv + 919 pp. London, Sydney, Auckland, Toronto.Clay, T. 1946. Mallophaga miscellany. - No. 3. The Trabecula. Annals and Magazine of Natural History (11) 13: 355-359.1951. An introduction to a classification of the avian Ischnocera (Mallophaga): Part I. Transactions of the Royal Entomological Society of London 102: 171-194.1969. A key to the genera of Menoponidae (Amblycera: Mallophaga: Insecta). Bulletin of the British Museum (Natural History) (Entomology) 24: 1-76. 1970. The Amblycera (Phthiraptera: Insecta). Bulletin of the British Museum (Natural History) (Entomology) 25: 75-98.Clay, T. & Hopkins, G. H. E. 1955. Notes on the Rudow Collection of Mallophaga at Hamburg. Mitteilungen aus den Hamburgischen Zoologischen Museum und Institut 53: 49-73.Conci, C. 1942. Diagnosi preliminari di tre nuovi generi e di una nuova specie di Trichodectinae. Bolletino della Societa Entomologica Italiana 74: 140-142. TRICHODECTID MAMMAL LICE 341 1946. Un nuovo genere di Trichodectidae Sudamericano. Bolletino delta Societa Entomologica Italiana 76: 59. Cracraft, J. 1974. Phylogenetic models and classification. Systematic Zoology 23: 71-90.Cummings, B. 1913. On some points of the anatomy of the mouth-parts of the Mallophaga. Proceedings of the Zoological Society of London 1913: 128-141.Eichler, W. 1940. Notulae Mallophagologicae. I. Neue Gattungen und Subfamilien von Haarlingen. Zoologischer Anzeiger 129: 158-162.1941. Zur Klassifikation der Lauskerfe (Phthiraptera Haeckel: Rhynchophthirina, Mallophaga und Anoplura). Archivfilr Naturgeschichte (N.F.) 10: 345-398.1963. Phthiraptera. 1. Mallophaga. Bronns, Klassen und Ordnungen des Tierreichs 5 (3) 7: vii + 291 pp. Leipzig. 1982. Notulae Mallophagologicae. XIII. Goliathipon nov. gen. und weitere neue Taxa der Gattungs- stufe. Deutsche Entomologische Zeitschrift 29: 81-88.Emerson, K. C. 1965. A new species of Mallophaga from the Philippine Islands. Journal of the Kansas Entomological Society 38: 68-69.Emerson, K. C. & Price, R. D. 1979. Two new species ofBovicola (Mallophaga: Trichodectidae). Journal of the Kansas Entomological Society 52: 747-750.1980. A new species of Suricatoecus (Mallophaga: Trichodectidae) from the western cusimanse, Crossarchus obscurus (Carnivora: Viverridae). Florida Entomologist 63: 505-508.1981. A host-parasite list of the Mallophaga on mammals. Miscellaneous Publications of the Entomological Society of America 12: 1-72.1982. A new species of Bovicola (Mallophaga: Trichodectidae) from the Formosan Serow, Capricor- nis crispus swinhoei (Artiodactyla: Bovidae). Pacific Insects 24: 186-188. 1983. A review of the Felicola felis complex (Mallophaga: Trichodectidae) found on New World cats (Carnivora: Felidae). Proceedings of the Entomological Society of Washington 85: 1-9.Ewing, H. E. 1929. A manual of external parasites, xiv + 225 pp. Springfield & Baltimore.1936. The taxonomy of the mallophagan family Trichodectidae, with special reference to the New World fauna. Journal of Par asitology 22: 233-246. Fahrenholz, H. 1936. Zur Systematik der Anopluren. Zeitschrift fur Parasitenkunde 9: 50-57.Ferris, G. F. 1929. Review: a manual of external parasites, by Henry Ellsworth Ewing. Entomologist's News 40: 337-341.1931. The louse of elephants. Haematomyzus elephantis Piaget (Mallophaga: Haematomyzidae). Parasitology 23: 112-127. 1933. The mallophagan genus Trichophilopterus. Parasitology 25: 468-471. 1951. The sucking lice. Memoirs of the Pacific Coast Entomological Society 1: 320 pp. Foerster, A. 1868. Synopsis der Familien und Gattungen der Ichneumonen. Verhandlungen des Naturhis- torischen Vereins der Preussischen Rheinlande und Westfalens 25: 135-221.Fresca, A. G. 1924. Mal6fagos del Museo de Madrid. III. Un Trichodectido nuevo. Boletin de la Real Sociedad Espanola de Historia Natural 24: 75-78.Gray, J. E. 1843. List of specimens of Mammalia in the collection of the British Museum, xxviii + 216 pp. London.Harrison, L. 1915. The respiratory system of Mallophaga. Parasitology 8: 101-127. 1916a. The genera and species of Mallophaga. Parasitology 9: 1-156. 19166. The relation of the phylogeny of the parasite to that of the host. Report of the British Association for the Advancement of Science 85: 476-477.Haub, F. 1967. Der Kopf von Pseudomenopon pilosum (Scopoli) (Mallophaga-Amblycera). Zoologische Jahrbucher. (Anatomic) 84: 493-558.1972. Das Cibarialsklerit der Mallophaga-Amblycera und der Mallophaga-Ischnocera (Insecta). Zeitschrift filr Morphologic und Okologie der Tiere 73: 249-261.1973. Das Cibarium der Mallophagen/Untersuchungen zur morphologischen Differenzierung. Zoologische Jahrbucher. (Anatomic) 90: 483-525. 1977. Das Cibarialsklerit von Arten der Familie Colpocephalidae (Burmeister 1838) Mallophaga/ Amblycera. Untersuchungen zur Grossenund Formkonstanz funktionsbezogener Strukturen in derUberordnung Psocodea. Zoologische Jahrbucher. (Anatomic) 97: 294-322. Hellenthal, R. A. & Price, R. D. 1976. Louse-host associations of Geomydoecus (Mallophaga: Trichodecti-dae) with the yellow-faced pocket gopher, Pappageomys castanops (Rodentia: Geomyidae). Journal ofMedical Entomology 13: 331-336. 1980. A review of the Geomydoecus subcalifornicus complex (Mallophaga: Trichodectidae) from Thomomys pocket gophers (Rodentia: Geomyidae), with a discussion of qualitative techniques and 342 C. H. C. LYAL automated taxonomic procedures. Annals of the Entomological Society of America 73: 495-503. 1984. A new species of Thomomydoecus (Mallophaga: Trichodectidae) from Thomomys bottae pocket gophers (Rodentia: Geomyidae). Journal of the Kansas Entomological Society 57: 231-236.Hennig, W. 1966. Phylogenetic systematics. Translated by Davis, D. D. & Zangerl, R. 263 pp. Urbana, Chicago and London.Hopkins, G. H. E. 1941. New African Mallophaga. Journal of the Entomological Society of South Africa 4: 32-47. 1942. Notes on Trichodectidae. Revista Brasileira de Biologia 2: 439453. 1943. Notes on Trichodectidae (Mallophaga) 1. Revista Brasileira de Biologia 3: 11-28. 1949. The host-associations of the lice of mammals. Proceedings of the Zoological Society of London 119: 387-604. 1960. Notes on some Mallophaga from mammals. Bulletin of the British Museum (Natural History) (Entomology) 10: 75-95. Hopkins, G. H. E. & Clay, T. 1952. A check list of the genera and species of Mallophaga. 362 pp. London.Imms, A. D. 1939. On the antennal musculature in insects and other Arthropods. Quarterly Journal of Microscopical Science 81: 273-320.Johnston, T. H. & Harrison, L. 1911. Notes on some mallophagan generic names. Proceedings of the Linnean Society of New South Wales 36: 321-328.Keler, S. von 1934. Mallophaga von Polen. Die Familie Trichodectidae. Bulletin del' Academic Polonaise des Sciences et des Lettres (B) 2: 259-267.1938a. Baustoffe zu einer Monographic der Mallophagen. I Teil: Uberfamilie der Trichodectoidea. Nova Acta Leopoldina 5: 395-467.19386. Uber einige Mallophagen aus Paraguay und Kamerun. Arbeiten iiber Morphologische u. Taxonomische Entomologie 5: 228-241.1939. Baustoffe zu einer Monographic der Mallophagen. II Teil: Uberfam. Nirmoidea (1). Nova Acta Leopoldina (N.F.) 8: 1-254.1944. Bestimmungstabelle der Uberfamilie Trichodectoidea (Mallophaga). Stettiner Entomolog- ischer Zeitung 105: 167-191.1957. Der Haarling der Wildkatze (Felicola hercynianus n. sp.). Mallophaga: Trichodectidae. Deutsche Entomologische Zeitschrift 4: 172-178. 1966. Zur Mechanik der Nahrunsaufname bei Corrodentien. Zeitschrift fur Parasitenkunde 27: 64-79. 1969. Mallophaga (Federlinge und Haarlinge). Handbuch derZoologie. Berlin. 4 (2) 11/17: 1-72. Kellogg, V. L. 1896. New Mallophaga, I, - with special reference to a collection made from maritime birdsof the bay of Monterey, California. Proceedings of the California Academy of (Natural) Sciences (2) 6:31-168. 1908. Mallophaga. In Wytsman, P., Genera Insectorum 66: 1-87. 1910. Corrodentia: Mallophaga. Wissenschaftliche Ergebnisse der Schwedischen Zoologischen Ex-pedition dem Kilimandjaro, dem Meru und dem Umgebenden Massaisteppen Deutsch-Ostafrikas 1905-1906 15(4): 43-56. 1913. Distribution and species-forming of ecto-parasites. American Naturalist 47: 129-158. 1914. Ectoparasites of mammals. American Naturalist 48: 257-279. Kim, K. C. & Ludwig, H. W. 1978. Phylogenetic relationships of parasitic Psocodea and taxonomicposition of the Anoplura. Annals of the Entomological Society of America 71: 910-922. 1982. Parallel evolution, cladistics, and classification of parasitic Psocodea. Annals of the Entomol-ogical Society of America 75: 537-548. Kuhn, H.-J. & Ludwig, H. W. 1964. Mallophaga on catarrhine monkeys: Colobus guereza, a natural hostof Procavicola colobi. Nature 203: 424425. Lakshminarayana, K. V. 1976. Nomenclatural changes in Phthiraptera - some suggestions. AngewandteParasitologie 17: 160-167. Ledger, J. A. 1980. The arthropod parasites of vertebrates in Africa south of the Sahara. Volume IV.Phthiraptera (Insecta). Publications of the South African Institute for Medical Research no. 56: 1-327. Lyal, C. H. C. 1983. Taxonomy, phylogeny and host relationships of the Trichodectidae (Phthiraptera:Ischnocera). Ph.D. Thesis, University of London. 602 pp. 1985. Phylogeny and classification of the Psocodea, with particular reference to the lice (Insecta: Phthiraptera). Systematic Entomology 10: 145-165. in prep. [a]. Host relationships of the Trichodectidae (Phthiraptera: Ischnocera).in prep. [b]. The external genitalia of Psocodea, with particular reference to the lice. Matsuda, R. 1965. Morphology and evolution of the insect head. Memoirs. American EntomologicalInstitute 4: 1-334. TRICHODECTID MAMMAL LICE 343 1970. Morphology and evolution of the insect thorax. Memoirs of the Entomological Society ofCanada 76: 1-431. 1976. Morphology and evolution of the insect abdomen - with special reference to developmental patterns and their bearings upon systematics. viii + 532 pp. Oxford & New York.Mayer, C. 1954. Vergleich Untersuchungen am Skeletl-Muskelsystem des Thorax der Mallophagen unter Berucksichtigung des Nervensy stems. Zoologische Jahrbiicher. (Anatomic) 74: 1-188.Mayr, E. 1969. Principles of systematic zoology, x + 428 pp. Bombay. Mjoberg, E. 1910. Studien iiber Mallophagen und Anopluren. Arkivfor Zoologi 6 (13): 1-296.Moreby, C. 1978. The biting louse genus Werneckiella (Phthiraptera: Trichodectidae) ectoparasitic on the horse family Equidae (Mammalia: Perissodactyla). Journal of Natural History 12: 395-412.Murray, M. D. 1957a. The distribution of the eggs of mammalian lice on their hosts. I. Description of the oviposition behaviour. Australian Journal of Zoology 5: 13-18.1957 b. The distribution of the eggs of mammalian lice on their hosts. II. Analysis of the oviposition behaviour of Damalinia ovis (L.). Australian Journal of Zoology 5: 19-29.Nelson, B. C. 1972. A revision of the New World species of Ricinus (Mallophaga) occurring on Passeriformes (Aves). University of California Publications in Entomology 68: 1-175.Nelson, G. 1972. Phylogenetic relationship and classification. Systematic Zoology 21: 227-231. 1974. Classification as an expression of phylogenetic relationship. Systematic Zoology 22: 344-359. Nitzsch, C. L. 1818. Die Familien und Gattungen der Thierinsecten (Insecta Epizoica). Magazin der Entomologie (Germar) 3: 261-316.Price, R. D. 1974. Two new species of Geomydoecus from Costa Rica pocket gophers (Mallophaga: Trichodectidae). Proceedings of the Entomological Society of Washington 76: 41-44.1975. The Geomydoecus (Mallophaga: Trichodectidae) of the southeastern U.S.A. pocket gophers (Rodentia: Geomyidae). Proceedings of the Entomological Society of Washington 77: 61-65.Price, R. D. & Emerson, K. C. 1971. A revision of the genus Geomydoecus (Mallophaga: Trichodectidae) of the New World pocket gophers (Rodentia: Geomyidae). Journal of Medical Entomology 8: 228- 257.1972. A new subgenus and three new species of Geomydoecus (Mallophaga: Trichodectidae) from Thomomys (Rodentia: Geomyidae). Journal of Medical Entomology 9: 463-467. Price, R. D. & Hellenthal, R. A. 1975o. A reconsideration of Geomydoecus expansus (Duges) (Mallo-phaga: Trichodectidae) from the yellow-faced pocket gopher (Rodentia: Geomyidae). Journal of the Kansas Entomological Society 48: 3342.19756. A review of the Geomydoecus texanus complex (Mallophaga: Trichodectidae) from Geomys and Pappageomys (Rodentia: Geomyidae). Journal of Medical Entomology 12: 401^08.1976. The Geomydoecus (Mallophaga: Trichodectidae) from the hispid pocket gopher (Rodentia: Geomyidae). Journal of Medical Entomology 12: 695-700.1979. A review of the Geomydoecus tolucae complex (Mallophaga: Trichodectidae) from Thomomys (Rodentia: Geomyidae) based on qualitative and quantitative characters. Journal of Medical Ento-mology 16: 265-274.1980a. The Geomydoecus oregonus complex (Mallophaga: Trichodectidae) of the western United States pocket gophers (Rodentia: Geomyidae). Proceedings of the Entomological Society ofWashington 82: 25-38.19806. The Geomydoecus neocopei complex (Mallophaga: Trichodectidae) of the Thomomys umbrinus pocket gophers (Rodentia: Geomyidae) of Mexico. Journal of the Kansas Entomological Society 53: 567-580. 1980c. A review of the Geomydoecus minor complex (Mallophaga: Trichodectidae) from Thomomys (Rodentia: Geomyidae). Journal of Medical Entomology 17: 298-313. 1981o. Taxonomy of the Geomydoecus umbrini complex (Mallophaga: Trichodectidae) from Thomomys umbrinus (Rodentia: Geomyidae) in Mexico. Annals of the Entomological Society ofAmerica 74: 37-47. 19816. A review of the Geomydoecus californicus complex (Mallophaga: Trichodectidae) from Thomomys (Rodentia: Geomyidae). Journal of Medical Entomology 18: 1-23. Price, R. D. & Timm, R. M. 1979. Description of the male of Geomydoecus scleritus (Mallophaga: Trichodectidae) from the southeastern pocket gopher. Journal of the Georgia Entomological Society 14: 162-165.Richards, O. W. & Davies, R. G. 1977. Imms' general textbook of entomology 10th Edition. 2 volumes, vii + 1354 pp. London.Rudolph, D. 1982a. Occurrence, properties and biological implications of the active uptake of water vapour from the atmosphere in Psocoptera. Journal of Insect Physiology 28: 111-121. 344 C. H. C. LYAL 1982ft. Site, process and mechanism of active uptake of water vapour from the atmosphere in the Psocoptera. Journal of Insect Physiology 28: 205-212. 1983. The water-vapour uptake system of the Phthiraptera. Journal of Insect Physiology 29: 15-25. Sikora, H. & Kichler, W. 1941. Uber Kopulationseigentumlichkeiten der Mallophagen. Zeitschrift fur Morphologic und Okologie der Tiere 38: 80-84. Simpson, G. G. 1961. Principles of animal taxonomy . 247 pp. New York.Sneath, P. H. A. & Sokal, R. R. 1973. Numerical taxonomy - The principles and practice of numerical classification, xc + 573 pp. San Francisco. Snodgrass, R. E. 1935. Principles of insect morphology . 667 pp. New York.Stobbe, R. 1913a. Mallophagen. 1. Beitrag: Neue Formen von Saugetieren (Trichophilopterus und Eurytrichodectes nn. gg.). Entomologische Rundschau 30: 105-106, 111-112.1913ft. Mallophagen. 3. Beitrag: Die Trichodectiden des Berliner Museums fur Naturkunde. Sitzungsberichten der Gesellschaft naturforschen der Freunde, Berlin 8: 365-383.Symmons, S. 1952. Comparative anatomy of the mallophagan head. Transactions of the Zoological Society of London 27: 349-436.Timm, R. M. & Price, R. D. 1979. A new species of Geomydoecus (Mallophaga: Trichodectidae) from the Texas pocket gopher, Geomys personatus (Rodentia: Geomyidae) . Journal of the Kansas Entomological Society 52: 264-268.1980. The taxonomy of Geomydoecus (Mallophaga: Trichodectidae) from the Geomys bursarius complex (Rodentia: Geomyidae). Journal of Medical Entomology 17: 126-145.Toulechkoff, K. 1955. The Trichodectidae (Trichodectoidea, Mallophaga) ectoparasitic on mammals of Bulgaria. [In Bulgarian.] Izvestiya na Zoologicheskiya Institut. Bulgarska Akademiya na Naukite. Sofiya 4-5: 423-434. Weber, H. 1936. Copeognatha. In Schulz, P. (Ed.), Biologic der Tiere Deutschlands 39: 1-50.Werneck, F. L. 1936. Contribuigao ao conhecimento dos Mallophagos encontrados nos mammifero sul-americanos. Memorias do Instituto Oswaldo Cruz 31: 391-589. 1941. Os Mal6fagos dos Procaviideos. Memorias do Instituto Oswaldo Cruz 36: 445-576. 1946. Sobre alguns ma!6fagos de procaviideos. Revista Brasileira de Biologia 6: 85-97. 1948. Os Maldfagos de Mamiferos. Parte I: Amblycera e Ischnocera (Philopteridae e parte de Trichodectidae) Edicao da Revista Brasileira de Biologia, 243 pp. Rio de Janeiro. 1950. Os Maldfagos de Mamiferos. Parte II: Ischnocera (continuacao de Trichodectidae) e Rhyn- cophthirina. Edi^ao de Memorias do Instituto Oswaldo Cruz. 207 pp.Wiley, E. O. 1979. An annotated Linnaean hierarchy, with comments on natural taxa and competing systems. Systematic Zoology 28: 308-337. 1981. Phylogenetics. The theory and practice ofphylogeneticsystematics. xv + 439 pp. New York. Zlotorzycka, J. 1972. Mallophaga. Goniodoidea i Trichodectoidea. Klucze do Oznaczania Owadow Polski 15 (3): 1-48. Index Synonyms are in italics; original generic combinations are indicated for identical specific epithets. Acondylocephalus 278 angularis 332 baxi 262 actuosi 332 annectens 263 bedfordi, Dasyonyx 284 acuticeps321 antidorcus 265 bedfordi, Felicola 312 acutirostris 312 appendiculata 262 bedfordi, Procavicola 274 adenota 262 arizonae 328 bedfordi, Trichodectes 265 aepycerus 264 armatus 270 Bedfordia 306 affinis 271 aspidorhynchus 321 bengalensis 321 africanus, Parafelicola 321 aspilopyga 255 birneyi 334 africanus, Procavicola 279 asymmetricus 334 bison 254 albati 332 aurei 332 Bisonicola 253 albimarginata 265 Bovicola 247 alcorni332 baculatus279 Bovidoecus 251 alleni 332 baculus 247 bovis 251 alpinus 251 bajaiensis 332 braziliensis 317 americanus 317 bakeri 333 breviceps 253 angolensis 279 barbarae 327 brucei 273 bulleri 332byersi 334 caffra 317 californicus 332 calogaleus 310 canis 300 capensis 284 caprae 251 Cebidicola 267 centralis 332 cercolabes 266 Cervicola 263 chapini 332 cherriei 332 chiapensis 332 chihuahuae 332 chilensis 328 chorleyi 262 clarkii 333 clausonae 332 clayi 265 cliftoni 332 colobi 279 comitans 266 concavifrons 251 Condylocephalus 274 conectens 265 Conepaticola 328 congoensis, Suricatoecus 310 congoensis, Trichodectes 279 cooleyi 312 copei 332 cordiceps 266 cornuta 265 coronadoi 332 costaricensis 332 crassipes 252 crenelata 262 crovelloi 332 cummingsi 328 cynictis 310 dakotensis 332dalgleishi 332Damalinia 257dariensis 333Dasyonyx 279decipiens 312dendrohyracis 284diacanthus 284dickermani 334dimorpha 262dissimilis 274divaricatus 301dorcephali 265dubius 278duchesnensis 333 Eichlerella 312 eichleri 273 elongata 265 emarginatus 273 emersoni, Geomydoecus 332 emersoni, Trichodectes 300 emeryi 301 INDEX equi 255cr mi niac 301euarctidos 300Eurytrichodectes 284Eutrichophilus 265ewingi 333exiguus 266exilis 289expansus 333extimi 333extrarius 270 fahrenholzi, Suricatoecus 312 fahrehholzi, Tricholipeurus 262 fallax 301 Fastigatosculwn 306 Felicinia 306 Felicola 302 felis 317 fennecus 312 ferrisi, Procaviphilus 278 ferrisi, Trichodectes 291 forficula 263 fulva255 fulvescens 333 fulvi 333 furca 273 galictidis 300Gatictobius 299gastrodes 328genettae 310genowaysi 334geomydis 333Geomydoecus 328grahamensis 332granulatus 278granuloides 278greeri 334Grisonia 299guadalupensis 333guineensis 284guinlei 312guyanensis 266 halli 333harrisi 278harrisoni 262heaneyi 333helogale 312helogaloidis 312hemitragi 253hendrickxi 263hercynianus 317heterohyracis 273hilli 262hindei 278hoffmanni 333Holakartikos 251hopkinsi, Damalinia 263hopkinsi, Dasyonyx 284hopkinsi, Felicola 312hopkinsi, Procavicola 274hueyi 333hyaenae 288hydropotis 263 345 idahoensis 333illinoensis 333inaequalis 312indica 265intermedius 288interruptofasciatus 328irolonis 333 jacobi 301jaliscoensis 333jamesbeeri 334jellisoni251johnhafneri 334jonesi 333jordani 279juccii 321 kingi 301kuntzi 300 Lakshminarayanella 327larseni 301laticeps 321lenicornis 321Lepikentron 252lerouxi 263liberiae 312Iimbatus251limitaris 333lindfieldi 274lineata 265lipeuroides 265lobatus 266longiceps 265Iongicornis251lopesi 273Lorisicola 312lutrae 290Lutridia 288Lymeon 327 maai 263 machadoi 285 macrurus 312 major 279 malaysianus 317 markhafneri 334 martinaglia 263 martini 333 martis 302 matschiei 289 maximus 266 mcgregori 333 Meganarion 267 Meganarionoides 278 melis 300 mephitidis 326 merriami 333 mexicanus, Eutrichophilus 266 mexicanus, Geomydoecus 333 meyeri 263 minimus 3 12 minor, Dasyonyx 284 minor, Eutrichophilus 266 minor, Geomydoecus 334 minutus 326 346 C. H. C. LYAL mjoebergi 317mobilensis 333mokeetsi 273moojeni 266moschatus 265muesebecki 279multilineatus 333multispinosus 253mungos 321muntiacus 263musculi 333mustelae 301 nairobiensis 284Nasuicola 327natalensis, Damalinia 263natalensis, Procavicola 273nayaritensis 333nebrathkensis 333neglecta 255neoafricanus 321neocopei 334Neodasyonyx 284Neofelicola 318neofelis 317Neolutridia 289neotheileri 262Neotrichodectes 321neumanni 279 occidentalis 312ocellata, 255octomaculatus 301oculatus 284oklahomensis 333oreamnidis251oregonus 333orientalis 262orizabae 334ornata 262osborni 326ourebiae 265ovalis, Dasyonyx 284ovalis, Trichodectes 300ovis 251 pakenhami 265pallidus 327panamensis 333Paradoxuroecus 318paradoxus 285Parafelicola 318paralaticeps 321parallela 265Paratrichodectes 300parkeri 265parvus 273pattoni 333pelea 262peregrini 334 perotensis 333philippinensis 321pinguis 300polydentatus 333potteri 334potus 302Potusdia 301pretoriensis 273Procavicola 270Procaviphilus 274Protelicola 286pygidialis 312 quadra ticeps 312quadridentatus 333 rahmi 312reduncae 264retusus 302Rhabdopedilon 251robertsi, Felicola 312robertsi, Trichodectes 278rohani 312ruficeps 284 salfii 302scleritus 333sclerotis 279scutifer 279sedecimdecembrii 254semiarmatus 270semistriatus 328semitheileri 263serraticus 279setosus, Felicola 312setosus, Trichodectes 266setzeri 333shastensis 333shoanus 273siamensis 317sika 263similis 317sinaloae 333smallwoodae 284spenceri 317spickai 333Spinibovicola 253spinifer 265Stachiella 301sternatus 273subcalifornicus 333subgeomydis 333subnubili 333subparvus 273subrostratus 312sudamericanus 317sumatrensis 321Suricatoecus 312 tamaulipensis 333tarandi 251 tendeiroi 279texanus 333theileri 263Thomomydoecus 334thomomyus 333thompsoni 263thoracicus 326thorntoni 273tibialis251tigris 247timmi 334tolteci 333tolucae 333trabeculae 264traguli 257Tragulicola 255transvaalensis 284traubi 333Trichodectes 291Tricholipeurus 264trichopi 333Trigonodectes 327tropicalis 333truncatus 333 ugandae 264 ugandensis, Dasyonyx 284ugandensis, Procavicola 273ugandensis, Trichodectes 300umbrini 333univirgatus 274Ursodectes 299ustulati 333 validus 284veracruzensis 333vicinus 273victoriae 265viverriculae 312vosseleri 300vulpis 312 wardi 334warmanae 333waterburgensis 284welleri 333 wernecki, Felicola 321wernecki, Geomydoecus 334Werneckiella 255Werneckodectes 290williamsi 334wolffhuegeli 326 yucatanensis 334 zacatecae 334zebrae 255zeylonicus 312zorillae 300zuluensis 255 British Museum (Natural History) An introduction to the Ichneumonidae of Australia /. D. Gauld In the important field of biological and integrated control of pests the parasitic Hymenopteraare of particular significance, and this work considers one of the largest families of Parasitica,the Ichneumonidae. The group has received little attention in Australia - though it has alreadybeen utilized successfully in curtailing the ravages caused by accidentally introduced pests. Forselective control programmes to be effective, however, a sound knowledge of the biology ofboth the pest and its parasites is essential - and a sound taxonomic base is vital for thedevelopment of such knowledge. Ironically, considering the group's economic importance, the parasitic Hymenoptera isamongst the least studied of any group of living organisms, and taxonomic difficulties havepresented major problems to many entomologists working with the Parasitica. AnIntroduction to the Ichneumonidae of Australia will go a long way towards rectifying thissituation, being a taxonomic treatment, by genus, of the Australian ichneumonids, acomprehensive illustrated identification guide, and a summary of all available information onthe group. It will also serve as an introduction to the biology and distribution of Australianichneumonids, and provide a check-list of the described species and an index to their knownhosts. It provides an important revision of ichneumonid nomenclature in order to bring thegroup into line with the generally accepted principles of zoological nomenclature. 1984, 413pp, 3 maps, 580 figs. Paperback. 565 00896 X 40.00 Titles to be published in Volume 51 The ichneumon-fly genus Banchus (Hymenoptera) in the Old World By M. G. Fitton The phytogeny, classification and evolution of parasitic wasps of the subfamily Ophioninae(Ichneumonidae) By I. D.Gauld A cladistic analysis and classification of trichodectid mammal lice (Phthiraptera: Ischnocera)By C. H. C. Lyal The British and some other European Eriococcidae (Homoptera: Coccoidea)By D. J. Williams Photoset by Rowland Phototypesetting Ltd, Bury St Edmunds, SuffolkPrinted in Great Britain by Henry Ling Ltd, Dorchester Bulletin of the British Museum (Natural History) x %/ s PRESENTED The British and some other EuropeanEriococcidae (Homoptera: Coccoidea) D. J. Williams Entomology seriesVol51 No 4 28 November 1985 The Bulletin of the British Museum (Natural History) , instituted in 1949, is issued in fourscientific series, Botany, Entomology, Geology (incorporating Mineralogy) and Zoology,and an Historical series. Papers in the Bulletin are primarily the results of research carried out on the unique andever-growing collections of the Museum, both by the scientific staff of the Museum and byspecialists from elsewhere who make use of the Museum's resources. Many of the papers areworks of reference that will remain indispensable for years to come. Parts are published at irregular intervals as they become ready, each is complete in itself,available separately, and individually priced. Volumes contain about 300 pages and severalvolumes may appear within a calendar year. Subscriptions may be placed for one or more ofthe series on either an Annual or Per Volume basis. Prices vary according to the contents ofthe individual parts. Orders and enquiries should be sent to: Publications Sales, British Museum (Natural History),Cromwell Road, London SW75BD,England. World List abbreviation: Bull. Br. Mus. nat. Hist. (Ent.) Trustees of the British Museum (Natural History), 1985 The Entomology series is produced under the general editorship of the Keeper of Entomology: Laurence A. Mound Assistant Editor: W. Gerald Tremewan ISBN 565 06014 4 ISSN 0524-6431 Entomology series Vol51No4pp347-393 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 28 November 1985 ^v-,-", ,r, ...6^ (. !* -v^V ' 2 a MOV h The British and some other European Eriococcidae(Homoptera: Coccoidea) D. J. Williams Commonwealth Institute of Entomology, c/o British Museum (Natural History), CromwellRoad, London SW7 5BD Contents Synopsis 347 Introduction 347 Acknowledgements and depositories 348 Morphology 348 Biology and economic importance 350 Eriococcidae Cockerell 351 Key to genera of British Eriococcidae 352 Cryptococcus Douglas 352 Eriococcus Targioni Tozzetti 356 Kuwanina Cockerell 384 Noteococcus Hoy 384 Ovaticoccus Kloet 384 Pseudochermes Nitsche 385 References 389 Index 393 Synopsis Eleven of the 12 indigenous British Eriococcidae, Cryptococcus fagisuga Lindinger, Eriococcus devonien-sis (Green), E. glyceriae Green, E. greeni Newstead, E. inermis Green, E. insignis Newstead, E. munroi(Boratynski), E. placidus Green, E. pseudinsignis Green, E. spurius (Modeer) and Pseudochermes fraxini(Kaltenbach), are redescribed; the twelfth species, E. cantium, is described as new. Four introducedspecies, E. lagerstroemiae Kuwana, Kuwanina parva (Maskell), Noteococcus hoheriae (Maskell) andOvaticoccus agavium (Douglas), are discussed. Coccus buxi Fonscolombe, the type-species of EriococcusTargioni Tozzetti, is redescribed, as are Coccus spurius Modeer, Acanthococcus aceris Signoret andRhizococcus gnidii Signoret, the type-species of Gossyparia Signoret, Acanthococcus Signoret andRhizococcus Signoret respectively. In agreement with some earlier works ? these three genera remain synonyms of Eriococcus but GreeniscaBorchsenius and Kaweckia Koteja & Zak-Ogaza are newly synonymised with Eriococcus. R. gnidii isnewly synonymised with E. thy mi (Schrank). An unnamed species of Eriococcus near E. buxi, occurring inthe U.S.S.R. and Turkey, is also discussed and lectotypes of the British species are designated, whereappropriate. There are sections also on morphology, biology and economic importance. Introduction The family Eriococcidae has been formerly associated with the Pseudococcidae, or mealybugs,but the two are now considered to be distinct, and Afifi (1968) has given good characters toseparate the adult males of both families. Ferris (1957a, 1957ft) discussed numerous genera,based on the female, that might be included in the Eriococcidae, and defined the family, mainlyon negative characters or without certain characters normally found in other families of scaleinsects. Many genera that have been included in the family have never been studied critically andtheir affinities may lie elsewhere. In order to establish the true relationships and possible phylogeny of the Eriococcidae, Dr J.M. Cox, British Museum (Natural History), and the present writer are currently studying thegroup on a world basis, and it is hoped to publish on these aspects later. In the meantime this Bull. Br. Mus. Nat. Hist. (Ent . ) 51 (4) : 347-393 Issued 28 November 1985 348 D. J. WILLIAMS paper has been written to facilitate identification of the British species currently assigned to thefamily, and to discuss some European genera, the type-species of which are not British althoughsuch genera have been recorded from Britain. Williams (1984) discussed briefly the distribution of the family and commented on its possibleorigins in Gondwanaland. It is poorly represented in the tropics and much speciation has takenplace in the more temperate areas. The North American fauna was described by Ferris (1955)(under the name Dactylopiidae) and by Miller & McKenzie (1967). Much of the Palaearcticfauna has been discussed in recent years by Borchsenius (1949), Danzig (1962, 1980), Dziedicka& Koteja (1971) and Tereznikova (1981). Some interesting species from South America havebeen described by Miller & Gonzalez (1975) and a thorough study of the 75 New Zealand specieswas made by Hoy (1962). The British species were listed by Boratynski & Williams (1964). Ofthe 481 species included by Hoy (1963) in his catalogue of world species, almost 150 are knownfrom Australia. It is doubtful, however, if more than one or two Australian species can beidentified from the present literature and, because the Australian Region probably includessome of the most interesting and bizarre species, it is clear that a definition of the family and itslimits can only be made after the Australian species have been studied adequately. Species fromthe Oriental Region, although at present rather few, are virtually unrecognisable from theliterature and need revising. Acknowledgements and depositories The writer is much indebted to Dr A. Kaltenbach, Naturhistorisches Museum, Vienna, who haskindly supplied for study specimens from V. Signoret's collection mentioned in the text. Mrs D.Matile-Ferrero, Museum National d'Histoire Naturelle, Paris, has sent for study valuablematerial of Rhizococcus gnidii and Eriococcus thy mi collected in France, for which the writerextends his sincere thanks. Dr J. M. Cox, British Museum (Natural History), has given muchhelp in discussions during the preparation of the manuscript. Most of the material studied is deposited in the British Museum (Natural History) (BMNH)but other depositories of original material mentioned are the Imperial Agricultural ExperimentStation, Tokyo (IAES) and the New Zealand Arthropod Collection, Auckland (NZAC). Morphology Many of the characters of the North American species were discussed by Ferris (1955) and byMiller & McKenzie (1967), resulting in definitions of the family for this area. The morphologyhas also been discussed by Danzig (1980) and Hoy (1962) and a study of the labium was made byKoteja (19740). Present studies of the British and other genera have revealed characters thathave not normally been used to separate species or genera and these characters are discussedbriefly here. There is no generally accepted common name for the family but the names 'felted scaleinsects' or 'felted coccids' have gained some acceptance and the former name could easily beused. The name refers to the ovisac produced by most species from the dorsum and the ventralmargins. This ovisac, secreted from ducts, has a felted texture and completely covers the top ofthe insect except for a hole at the posterior end, allowing first instars to escape. The felted saccombines with mealy secretion on the venter, but on the dorsum rods of wax that seem to beproduced from the dorsal setae are also present. Many slide preparations show these rodsattached to the enlarged setae. Body shape and segmentation. The shape of the body varies from globular to broadly-oval andelongate, and although species may be found on leaves, stems and roots, the elongate species areusually found on grasses. In many species the body is strongly nodulose and when this takes upstain, it often masks minute characters. Numbering of the segments follows that given by Miller(1984) for Eriococcidae and by Williams (19850) for Pseudococcidae. In this system the vulva issituated between the 7th and 8th abdominal segments, so that the first abdominal segment on theventer is represented by areas lateral to the 3rd coxae and the 2nd abdominal segment is THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 349 complete after the metathorax. On the dorsum the 1st abdominal segment is complete, butbecause the segment containing the anal lobes is probably complex, it is here regarded as the 8thor anal lobe segment. Anal area. In many genera of Eriococcidae the posterior end of the body is produced into twoprominent anal lobes, each usually sclerotised and often conical with a long apical seta. Dorsallythere are frequently enlarged setae, varying in number, but in most species there are two suchsetae on the inner edge and one on the outer edge, although this may be reduced to a muchsmaller seta. Ventrally there are flagellate setae, often numbering 1-3, and near the inner cornerof each lobe there is a seta that Hoy (1962) named the suranal seta. This is usually slender but insome genera it is flat or spatulate. In all the species discussed here the suranal setae are slender.Sometimes the inner edges of the lobes are so strongly nodulose and sclerotised that they appearto be toothed or serrate. When there are no definite anal lobes present the suranal setae arerepresented by a pair just anterior to the anal ring on the venter. Species with anal lobes oftenhave a small dorsal lobe or plate projecting between the base of the lobes. It varies from beingonly a slight prominence to triangular, well developed and nodulose. In some species it is heavilysclerotised in the form of a definite plate, as in Eriococcus phyllanthi Ferris, and this led Ferris(19576) to refer to it as a cauda. It is not certain whether this plate is homologous with the caudaof the Aphidoidea and it is here called a dorsal plate no matter what the extent of itsdevelopment is. The anal ring, in species with well-developed anal lobes, is normally situatedvertically to the body between the dorsal plate and the apical margin of the venter between thesuranal setae. When prepared on microscope slides it often becomes distorted or is pushedinwards, lying neither on the dorsum nor venter. Although the shape varies considerably, andthis is mentioned in the text when appropriate, a normal anal ring usually has 8 setae and a singlerow of pores, except laterally, where there are a few pores forming a double row. Antennae. These vary from being normal with 6 or 7 segments, to small stubs with 1 or 2segments. A normal antenna appears to taper, with the terminal segment usually one of theshortest and not wider than the previous segments. Frontal lobes and frontal tubercles. Often just antero-medially to each basal antennal segmentthere is a lobe-like structure that varies in size and length. They are much more pronounced indistorted specimens when they can be observed to stick out from the surface. Normally they aremembranous but sometimes, as in Eriococcus buxi and its relative herein described, they tend tobe sclerotised. Their function is unknown. Signoret (1875) mentioned these lobes whendescribing Acanthococcus aceris. Boratynski described them as eversible membranous tuberclesin A. munroi but there is no evidence that they are eversible. They were mentioned also byDziedicka & Koteja (1971) when describing Rhizococcus palustris. These structures seem tohave been ignored in most other works but Dr J. M. Cox and the present writer have observedthem in numerous species from all the zoogeographical regions. In at least one species fromAustralia these lobes extend around and posterior to the basal antennal segment; for this reasona more appropriate term would be frontal lobes and this term is used throughout this work. Frontal tubercles are minute raised spots situated antero-medially to each antennal base.They are difficult to see but once their normal position is located their presence or absence canbe noted. Again, their function is unknown but they seem to be homologous with the 'conicaldisc pores' mentioned by Ben-Dov et al. (1975) in Protopulvinaria mangiferae (Green), and tothe 'preantennal spots' discussed by Ben-Dov (1979) in species of Kilifia. The species in thesegenera belong to the family Coccidae and the presence of these tubercles may be much morecommon throughout the group. They are certainly present in many species of Eriococcidae andmay have some taxonomic significance. In all species studied so far, frontal lobes and frontaltubercles never occur together and one may have been derived from the other. Pseudochermesfraxini, herein described, possesses frontal tubercles in all instars and this may be normal, butsome species have neither frontal lobes nor frontal tubercles. Legs. Unlike species of Pseudococcidae, most of which have the tibia longer than the tarsus, thetarsus in the Eriococcidae is often longer than the tibia. Also the trochanter pores in the 350 D. J. WILLIAMS Eriococcidae form a line from the anterior lateral corner to about half-way along the posterioredge, more or less dividing the trochanter into two equal parts. In the Pseudococcidae thetrochanter pores are situated towards the distal end. Labium. The basal segment often has 2 pairs of setae but in some genera these are reduced to asingle pair. In Cryptococcusfagisuga, herein described, they are absent except in the first instarwhen a single pair is present. Setae. On the dorsum and ventral edges, there are often enlarged setae that are spine-like,conical, pointed, blunt or truncate, but sometimes they are cylindrical, lanceolate or evenbulbous. Sometimes they are interspersed by quite slender setae. In some species, enlargedsetae are confined to the anal lobes or they are absent entirely, but usually the dorsal setae,which may be minute, show some signs of being stiff and not flagellate. When the dorsal setaetend to be small and slender in the adult, there are usually thicker setae in at least the first instar.On the median part of the venter, what are here referred to as normal setae are flagellate setaefound in many groups of the Coccoidea. Laterally there are often other setae which, althoughslender, are nevertheless stiff. Macroducts. These have been discussed by Ferris (1955), Miller & McKenzie (1967), Miller(1984) and by many other workers. They normally have the inner end reflexed in the form of acup and, although found in some other families, they represent one of the most importantcharacters of the Eriococcidae, although sometimes they are absent entirely. Microducts. The significance of these minute ducts has been little understood but their structuremay indicate affinities and evolutionary paths. They have been discussed in some detail by Goux(1948). Normally the inner end is bulbous, and the term ampulla is here adopted for it. It usuallybears a minute filament arising from the inner end. Between the ampulla and the external orificethere is a tube varying from filamentous and slender to short and bulbous at the inner end, andsituated internally just next to the orifice there is a structure often wing-nut-shaped in profile butwhich is here called a collar. Sometimes the orifice is bifid, as shown already by Miller &Gonzalez (1975) in Eriococcus araucariae Maskell. The bifid orifice is external and may be seenon microducts at the edge of the body where the bifid part protrudes from the surface of theintegument. They are not internal as shown by Hoy (1962) for E. araucariae. It is still not clearwhether the shape and type of the microduct has generic significance. Cryptococcus fagisugaherein described possess microducts in the form of a double tube. Enlarged ducts. These are illustrated here in E. buxi and E. sp. near buxi. They are usually largerthan the macroducts but differ in having a flat or slightly rounded inner end, instead of beingreflexed into a cup. Because Eriococcus possesses these ducts, Borchsenius (1949) separated thisgenus from all others. The significance of these special ducts is still not clear because Miller &Gonzalez (1975) discussed them in Exallococcus laureliae Miller & Gonzalez. Furthermore,some Australian species currently assigned to Eriococcus also have them. Pores. Although the disc pores are usually quinquelocular, there are often pores with up to 9loculi. They are usually on the ventral surface but when present on the dorsum they havesometimes been given generic significance. Cruciform pores are usually oval with a central slit inthe form of a cross. When present they are usually found on the venter but their presence on thedorsum has been given generic importance by Koteja & Zag-Ogaza (1981) when describingKaweckia. Biology and economic importance Little work has been done on the biology of the British Eriococcidae, but Newstead (1903) hasgiven useful field notes on some species. Gullan (1984) included the specialised gall-forminggenus Apiomorpha in the family and has given important information on gall formation.Normally British species are univoltine but Patel (1971), discussing Eriococcus coriaceusMaskell from South Australia, has stated that it has five generations a year there. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 351 British species are oviparous but whether this is true for all Eriococcidae is not clear from theliterature. So far as is known, there are always three instars in the female and five in the malewhen present. Koteja (1983) has discussed an additional moult in Acanthococcus greeni(Newstead) after the adult female had formed, but this was probably caused by parasitism. InPseudococcidae there are normally four female instars. The ovisac in the indigenous British species is felted but, apart from the gall-forming species,it is not clear how the covering of many is formed. In Ovaticoccus agavium it is described byBoratynski (1958) as loosely felted. The ovisac of Eriococcus spurius in mature^ adults is presentonly around the margins, but this is because there are no macroducts in the median areas. Theinsect is, therefore, easily recognisable and is specific to elms. Most of the British species feed on grasses but it is not known if they cause damage.Eriococcus devoniensis is always found on Erica tetralix in Britain, causing the stems to turn acomplete circle enclosing the insect and ovisac in the bend. Large areas of the plant can be killedand this led Warburton (1937) to recommend burning to destroy the insect. Pseudochermesfraxini is found in enormous numbers in the crevices of bark of ash trees and often causesconcern, but so far there are no records of mortality of trees. By far the most important insect isCryptococcusfagisuga, found in the crevices of bark of beech trees. It is particularly prevalent inolder trees (Wainhouse, 1980, states that trees around 25 years of age are particularlysusceptible) and the species is probably found throughout Britain wherever beech is grown.Parker (1975) has stated that beech planted in southern England 1951-60 will be susceptibleduring the 1980s. The insect is probably the only parthenogenetic species of Eriococcidae inBritain and is often present in'enormous numbers producing what has been termed fluffy whitewool, so that trunks appear to be completely white. Despite the large numbers, the insectsthemselves apparently do not cause damage. But, associated with the insect is an ascomycetefungus, Nectria sp., known as 'beech bark disease', a name coined by Ehrlich (1934) in acomprehensive account of the disease and insect in Canada and U.S.A. The fungus causesconsiderable damage resulting in mortality of the tree, and is dependent on the insect forproviding an entry wound at the point of feeding. Parker (1975) showed that Nectria infected thebark only following wounding, but as Lonsdale (1980) has shown, heavy infestations of theinsect induce a lowering of resistance to fungal invasion within bark tissue. In the Old World the insect is distributed throughout Europe, with incursions into Turkey andIran. Covassi (1975) stated that Sicily is the southern limit, and the Commonwealth Institute ofEntomology (1979) has produced a map showing the world distribution. Thomsen et al. (1949)gave a full and important account of the insect and fungus in Denmark. In Britain the insect hasbeen made the subject of a Forestry Commission Leaflet (Hussey, 1956). Wainhouse (1979)showed that the larvae are dispersed passively by wind and a small percentage are carried abovethe canopy. Some trees are more susceptible than others (Wainhouse & Howell, 1983) and,although there is possible intraspecific variation within populations of the beech scale, there isno evidence of adaptation of the scale to particular trees. It was reported by Elliott (1933) that, in Suffolk, this species was devoured by a small beetle,Enicmus [now Lathridius] minutus (L.), but this was unlikely because the beetle is normally afungus-feeder. ERIOCOCCIDAE Cockerell Acanthococcites Signoret, 1875: 16. Type-genus: Acanthococcus Signoret. [Placed on Official List ofFamily-Group Names in Zoology (Name Number 535) with an endorsement that it is not to be givenpriority over Eriococcini Cockerell, 1899 whenever both names are held to denote a single taxon(Melville, 1982: 96).] Acanthococcidae Signoret; Maskell, 1887: 47. Eriococcini Cockerell, 18990: 389. Type-genus: Eriococcus Targioni Tozzetti. [Placed on Official List ofFamily-Group Names in Zoology (Name Number 534) with an endorsement that it is to be givenprecedence over Acanthococcidae Signoret, 1875 whenever both names are held to denote a singletaxon (Melville, 1982: 96).] Eriococcidae Cockerell; Brues & Melander, 1932: 134. 352 D. J. WILLIAMS Cryptococcidae Kosztarab, 1968ft: 12. Type-genus: Cryptococcus Douglas. [Synonymised with Eriococci-dae by Danzig, 1980: 58.] Until all genera are studied, including the peculiar gall-producing forms, it is impossible todefine the family on a world basis. The following definition is based only on species that havebeen studied in some detail. ADULT FEMALES. Body elongate, broadly oval or globular, usually membranous but often stronglynodulose. Segmentation usually distinct in the oval and elongate forms, obscure in the globular forms.Anal lobes often well developed, conical and sclerotised; when present usually with a dorsal plate betweenand with suranal setae at inner ventral bases. Anal ring often with 8 setae and pores, sometimes reduced,without pores. Antennae, when normal, with 6 or 7 segments, the apical segment usually narrower thanpreceding segments, sometimes segments reduced to one or two. Frontal lobes or frontal tuberclessometimes present. Legs present or absent; when absent sometimes third pair represented by minute flaps;when present and normal, tibia frequently longer than tarsus, trochanter pores in a line almost dividingtrochanter into two equal parts. Labium usually with 1 or 2 pairs of setae on basal segment, rarely absententirely. Dorsal setae often enlarged, conical, bulbous or cylindrical, if small, remaining spine-like or stiff.Disc pores, when present, usually quinquelocular, but sometimes with up to 9 loculi, frequently on venterand sometimes on dorsum. Macroducts normally present, at least on dorsum, sometimes absent.Microducts usually present. Cruciform pores often present on venter, rarely on dorsum. Key to genera of British EriococcidaeAdult females 1 Legs absent except for third pair replaced by small leg flaps Legs present 2 Macroducts present on dorsum , CRYPTOCOCCUS Douglas (p. 352) - Macroducts absent on dorsum, replaced by large invaginated quinquelocular pores KUWANINA Cockerell(p. 384) 3 Anal lobes present, sclerotised 4 - Anal lobes absent 5 4 Anal lobes with a series of about 12 enlarged setae on dorsum. Segmentation absent between trochanter and femur and between tibia and tarsus NOTEOCOCCUS Hoy (p. 384) Anal lobes with at most 4 enlarged setae on dorsum. Segmentation distinct between trochanterand femur and between tibia and tarsus ERIOCOCCUS Targioni Tozzetti (p. 356) 5 Anal ring crescentic, containing pores and setae. Cruciform pores absent PSEUDOCHERMES Nitsche(p. 385)Anal ring not as above, without pores. Cruciform pores present OVATICOCCUS Kloet (p. 384) CRYPTOCOCCUS Douglas Cryptococcus Douglas, 1890: 155. Type-species: Coccus fagi Baerensprung [= Cryptococcus fagisugaLindinger], by original designation. The type-species is now known to occur on Fagus spp. throughout Europe and the eastern part of theU.S.A. C. aceris Borchsenius is known from Georgia and Azerbaizhan in U.S.S.R. and from Germany onAcer. Another species on Acer was described as C. williamsi Kosztarab from eastern U.S.A., and C.integricornis Danzig is known from the Primorsky region in the far east of U.S.S.R. on Tilia amurensis.Kosztarab (1968a) gave a key to the species known at the time. All four species are without legs in the adult female, but the hind pair are replaced by small flaps. In thetype-species these are simple, each with at most a single seta, but in the other species the flaps are eitherreticulate or they possess minute pores. When viewed at the edge of the body these structures are flap-likeand protrude . They are not plates on the surface of the body or pore plates , as has been suggested ; becausethey are always in the positions of the third legs and probably emit pheromones they are here referred to asleg flaps. C. integricornis possesses also spine-like remains of the first and second pairs of legs. All fourspecies have 2-5 segmented antennae that are stub-like, and macroducts each with a cup-shaped inner endon the dorsum and at least the ventral margins. Microducts are present. In the type-species they are in theform of double tubes, the ampulla of each being expanded, with a filament between. The microducts of theother species need more critical study. Quinquelocular pores are present and the anal ring is simple, squareor oval with 4-6 setae and a few pores. In C. fagisuga the basal segment of the labium is without setae but THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 353 there is a single pair present in the first instar. Specimens of C. fagisuga, C. aceris and C. williamsiexamined possess neither frontal lobes nor frontal tubercles. Hoy (1962) accepted C. nudatus Brittin in the genus. The species was described from New Zealand onHoheria spp., but because it lacks macroducts it may belong to another genus, although it is very close.Kuwanina parva (Maskell), described from Japan and recorded from Britain on Prunus, has at one timebeen assigned to Cryptococcus, but the genus Kuwanina is here accepted as distinct because it lacksmacroducts. These are replaced by large invaginated quinquelocular pores. Cryptococcus is represented by C. fagisuga only in Britain, where it is sometimes injurious to beech. Cryptococcus fagisuga Lindinger(Figs 1,2) Cocotf/ag/ Baerensprung, 1849: 174. Syntypes $, GERMANY: Berlin (probably lost). [Junior homonym of Coccus /og/Sulzer, 1776.]Coccus fagi Walker, 1852: 1086. LECTOTYPE $, ENGLAND (BMNH), here designated [examined]. [Junior homonym of Coccus fagi Sulzer, 1776.]Pseudococcus fagi (Baerensprung) Douglas, 1886: 152.Cryptococcus fagi (Baerensprung) Douglas, 1890: 155; Newstead, 1903: 215; Schmutterer, 1952: 417; Ferris, 1955: 83; Tereznikova, 1981: 46.Cryptococcus fagisuga Lindinger, 1936: 444; Borchsenius, 1949: 37; Hoy, 1963: 54; Danzig, 1964: 634. [Replacement name for Coccus fagi Baerensprung.] DESCRIPTION. Adult female (Fig. 1) in life, globular, bright to lemon yellow, completely covered in a whiteovisac. On the slide, rotund, membranous, nodulose, largest about 0-6 /u,m in diameter, often wider thanlong, anal lobes not developed, their positions removed to ventral surface of body. Antennae reduced,with 1-2 segments. Legs absent but position of each third leg represented by a small flap often containing asingle seta. Labium 48 /um long, often wider than long, shorter than clypeolabral shield, basal segmentwithout setae. Spiracles often triangular, sclerotised. Anal ring at apex of venter, sclerotised, almostquadrate, with 4 short setae 12 /um long occupying corners, a few minute pores on posterior margin ormorphologically anterior margin. Dorsal surface with small setae that are pointed and stiff, 5 /im long, in more or less single rows across thesegments. Macroducts in single rows across the segments, each about 8 ^tm long, with the cup wider than adorsal setal base. Microducts in moderate numbers, each about 6 /u,m long, in the form of a double tubetapering towards double orifice, each tube with inner end swollen into an ampulla with a filament between. Ventral surface with setae about same size as dorsal setae except on anal lobe segment where the suranalsetae, and one or two near margins, are larger and conical, about 10 /u,m long and usually a pair of slightlylonger setae near anterolateral corners of anal ring, probably the anal lobe setae. Macroducts of two types.A large type, similar to those on dorsum, few, around margins and sometimes in submarginal areas ofthorax and anterior abdominal segments. A small type, not much larger than a microduct, arranged singlyin submedian areas of third to fifth abdominal segments and on metathorax anterior to leg-flaps, each ductwith a shallow cup. Microducts, same as on dorsum, sparse, present in median areas as well as margins.Quinquelocular pores, about same diameter as the cup of a large macroduct, in more or less single rows onposterior abdominal segments, there being also one or two near each spiracular opening and one or twobetween clypeolabral shield and labium. Second instar female (Fig. 2B) broadly oval, smaller than adult female. Antennae near margins,2-segmented, reduced. Legs absent but position of each third leg represented by a minute sclerotised point.Labium 36 /urn long, smaller than clypeolabral shield, basal segment without setae. Anal ring on venter atapex of abdomen , similar to that of adult but smaller. Dorsal surface with small pointed but stiff setae, each4 fjLm long, many not much longer than a setal base. Macroducts smaller than in adult female, present insingle transverse rows on most segments. Microducts, as in adult, bitubular, represented by one or twoonly. Ventral surface with setae as on dorsum except for larger conical setae lateral to anal ring, a pair ofwhich longer than others and probably apical setae. Suranal setae small. Macroducts same as on dorsum,occasionally on margins only. Microducts represented by one or two on margins only. Quinquelocularpores arranged singly next to spiracular openings only. First instar (Fig. 2A), oval, the smallest 300 [im long, 180 urn wide. Anal ring displaced to apex of venter,as in later instars but smaller. Antennae 65 /xm long, with 5 segments. Legs normal, hind trochanter +femur 40 /um long, hind tibia 16 /urn long, hind tarsus 16 ^tm long, claw slender, without a denticle, 16 /xmlong. Labium 32 /um long, basal segment with a pair of minute setae. Dorsal surface with short pointed andstiff setae, 4-5 pm long, but some on head longer and flagellate. Macroducts absent. Microducts 354 D. J. WILLIAMS D.J.W. \ Fig. 1 Cryptococcus fagisuga Lindinger. England, Buckinghamshire, Burnham Beeches, on Fagussylvatica. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 355 Fig. 2 Cryptococcus fagisuga Lindinger. (A) First instar. (B) Second instar. 356 D. J. WILLIAMS represented at most by one or two varying in position. Quinquelocular disc pores sparse, rarely more than 2on any segment. Ventral surface with short setae similar to dorsal setae except for apical setae each 40 yxmlong, larger conical setae lateral to anal lobes, and longer flagellate setae on head. Suranal setae small.Macroducts and microducts absent. Quinquelocular pores few in submarginal areas of abdomen and headand singly near spiracular openings. One or two pores medially to first and second coxa often trilocular, butthis condition not constant. A minute indefinite scar present lateral to third coxae. MATERIAL EXAMINED (all on Fagus sylvatica unless stated otherwise) England: Berkshire, Silwood Park, 5.V.1948, 17.iii.1950, 29.V.1959 (K. L. Boratynski), Cookham,4.iv.l973, ll.viii.1983 (D. J. Williams); Buckinghamshire, Burnham Beeches, 18.iii.1973, 1, 4, 15,23.iv.1973 (D. J. Williams); Cheshire, nr Chester, 1890 (R. Newstead), Ince, xi.1895 (R. Newstead);Hampshire, Bolderwood, 29.iii.1975 (J. H. Martin); Kent, Mailing, ix.1895 (E. E. Green), Bearsted,27.vii.1913 (E. E. Green), nr Canterbury, 19.V.1973 (L. M. Hanford); London, Buckingham PalaceGardens, S.x.1963 (K. L. Boratynski); Northumberland, Riding Mill, S.x.1950 (D. J. Williams); Shrop-shire, l.xi.1933 (E. E. Edwards); Suffolk, xi.1970; Surrey, Camberley, 25.V.1920, 10.iv.1930 (E. E.Green), Wisley, 4.vii.l954 (D. J. Williams); Yorkshire, Bingley, lO.x.1933 (G. Fox-Wilson), HardcastleCrags, 22.viii.1961 (D. J. Williams). Scotland: Aberdeen, iv.1924 (G. D. Morison). Luxembourg:Beaufort, 29.vi.1961 (K. L. Boratynski). Hungary: Nagymilis Laszlotanya, 12.viii.1980 (F. Kozdr & M.Kosztarab). U.S.A.: Vermont, on F. grandifolia, 8.V.1975. DISCUSSION. When discussing the homonymy of Coccus fagi, Lindinger (1936) referred to Roemer (1789)who gave an illustration. The first 32 plates of this work are a re-issue of those in Sulzer (1776) where onPlate XI, fig. 11 there is an illustration in colour of C. fagi. This seems to be Eulecanium tiliae (L.) aspresent understood. C. fagisuga is always parthenogenetic insofar as males have never been found. Although Walker made numerous slide preparations of aphids, there is no evidence that he made anyslide preparations of the scale insects that he described. A batch of dry material mounted on a card labelledCoccus fagi in Walker's handwriting is present in BMNH. This is here regarded as Walker's originalmaterial and among 9 specimens prepared on separate microscope slides one has been selected aslectotype. The other 8 are labelled paralectotypes. ERIOCOCCUS Targioni Tozzetti Eriococcus Targioni Tozzetti, 1868: 726. Type-species: Coccus buxi Fonscolombe, by subsequent designa-tion (Signoret, 1872: 429) and by Opinion 1203 (Melville, 1982: 95).Gossyparia Signoret, 1875: 20. Type-species: Coccus ulmi Linnaeus sensu Linnaeus, 1766 [= Coccus spurius Modeer, 1778.], by original designation. [Synonymised by Ferris, 1955: 94.]Acanthococcus Signoret, 1875: 35. Type-species: Acanthococcus aceris Signoret, by monotypy and by Opinion 1203 (Melville, 1982: 95). [Synonymised by Ferris, 1955: 94.]Rhizococcus Signoret, 1875: 36. Type-species: Rhizococcus gnidii Signoret, by monotypy. [Synonymised by Ferris, 1955:94.]Greenisca Borchsenius, 1948: 502. Type-species: Eriococcus inermis Green, by original designation. Syn. n.Anophococcus Balachowsky, 1954: 61. Type-species: Eriococcus inermis Green, by original designation. [Synonymised with Acanthococcus Signoret by Danzig, 1980: 205.]Kaweckia Koteja & Zak-Ogaza, 1981: 501. Type-species: Eriococcus glyceriae Green, by original designation. Syn. n. Eriococcus Targioni Tozzetti; Melville, 1982: 95. [Addition to Official List, Name Number 2153.]Acanthococcus Signoret; Melville, 1982: 95. [Addition to Official List, Name Number 2154.] Although in time it may be necessary to recognise some of the genera Synonymised above, they are hereregarded as components of Eriococcus pending further research on the family and until the characters ofmicroducts, frontal lobes and tubercles can be assessed on a world basis. As Ferris (1955) has mentioned,division of the group seems undesirable (except for extreme forms) until comprehensive studies have beenmade, especially of the Australian fauna. Some reasons for the synonymy need explanation, however. Eriococcus. This genus was separated by Borchsenius (1948, 1949) because the type-species possessesenlarged ducts. As already stated, some Australian species possess these ducts and their true significance isnot yet known. The type-species also has well-developed frontal lobes, slender microducts with bifid orificeand only a single ventral seta on the anal lobes. Borchsenius based his interpretation of the type-species onRussian specimens which do not appear to be identical with specimens from the type-locality in France. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 357 Acanthococcus. The most important characters separating the type-species from Eriococcus are theabsence of enlarged ducts and the presence of 2 ventral setae instead of one on the anal lobes and 2 pairs ofsetae on the basal segment of the labium instead of a single pair. Frontal lobes are well developed and themicroducts are slender with bifid orifice. At present there seems to be no reason for recognising the genus. Gossyparia. Based on the type-species, this is almost a replica of Acanthococcus but lacks dorsalmacroducts in the median areas of the dorsum, regarded here as nothing more than a specific character.There are well-developed frontal lobes and the microducts are slender with bifid orifice. Rhizococcus. The genus was separated from Acanthococcus originally because, in the type-species, theantennae had 7 segments instead of 6 and because the body was presumed to be naked. It is now known thata normal ovisac is produced and the antennae may have either 6 or 7 segments. Other characters presentare frontal tubercles instead of frontal lobes and slender microducts with bifid orifice. With our presentknowledge the genus is not distinctive enough from Eriococcus. Greenisca. There are nomenclatural problems concerning the type-species. According to Danzig (1962)the species on which Borchsenius (1948) based Greenisca was not E. inermis Green but another species,which Danzig (1980) stated, was later described as Anophococcus gouxi Balachowsky; the type-species,therefore, of Greenisca should be A. gouxi Balachowsky, 1954 = E. inermis Green sensu Borchsenius1948. Article 70 of the International Code of Zoological Nomenclature states that 'if a zoologist considersthat a type-species designated for a new genus has been misidentified, then that person should refer thecase to the Commission to designate as the type-species whichever species will in its judgement best servestability and uniformity of nomenclature . . .' Furthermore, the anal lobes of G. gouxi illustrated byDanzig (1980) have three enlarged setae but the same species illustrated by Balachowsky (1954) has theouter seta much reduced in size. E. inermis Green has a minute outer seta on the anal lobes, frontaltubercles and microducts without bifid orifice. Based on this type-species the genus Greenisca is hereregarded as a component of Eriococcus. If Anophococcus gouxi is accepted as the type-species, a specieswith dorsal disc pores, there may be some justification for recognising the genus but until the significance ofdorsal disc pores is better understood, the species is here regarded as belonging to Eriococcus. Kaweckia. The type-species has numerous cruciform pores and disc pores on the dorsum and the numbersof dorsal enlarged setae are restricted to the margins of the posterior abdominal segments. Othercharacters are frontal tubercles and microducts without a bifid orifice. At present there seems to be littlejustification for accepting this genus and it is here regarded as a component of Eriococcus. Although the type-species of Eriococcus, Acanthococcus and Rhizococcus are not British species theyare described here to help facilitate identification and to assess the characters. The following key ispresented to separate the British indigenous and introduced species. Key to British species of Eriococcus 1 Enlarged setae on abdomen confined to inner edges of dorsum of anal lobes only. Outer dorsal setae on anal lobes much smaller 2 - Enlarged setae on abdomen on at least a few segments anterior to anal lobes. Outer dorsal setae on anal lobes same type as inner setae 3 2 Disc pores of quinquel ocular type present on dorsum of body. Enlarged setae present on head margins placidus Green (p. 376) - Disc pores on dorsum absent . Enlarged setae absent on head margins inermis Green (p . 370) 3 Enlarged setae in bands across head , thoracic and most abdominal segments 4 - Enlarged setae on margins only except for a few occasionally on head 9 4 Anal lobes with 4 enlarged setae cantium sp. n.(p. 363) Anal lobes with 3 enlarged setae 5 5 Enlarged setae in median areas of posterior abdominal segments much smaller than others on dorsum munroi (Boratynski) (p. 374) Enlarged setae in median areas of posterior abdominal segments same size as others on dorsum 6 6 Dorsal macroducts in bands across segments Dorsal macroducts absent in mid-regions of thorax and abdomen spur/us (Modeer) (p . 380) 7 With at least 2 setae in mid-dorsal area of anal lobe segment lagerstroemiae Kuwana (p. 374) Without setae in mid-dorsal area of anal lobe segment 8 8 Body elongate-oval. Dorsal setae sharply pointed. Frontal tubercles present greeniNewstead(p. 367) - Body broadly oval. Dorsal setae truncate. Frontal lobes present devoniensis (Green) (p. 365) 358 D. J. WILLIAMS 9 Dorsal enlarged setae truncate, with almost parallel sides, on margins only of 7th and two orthree preceding segments, in addition to anal lobe setae. Dorsal disc pores present. Dorsalcruciform pores present glyceriae Green(p. 367) - Dorsal enlarged setae conical, either truncate or bluntly pointed, in a continuous row around margins. Dorsal disc pores absent. Dorsal cruciform pores absent 10 10 Margins of 7th abdominal segment each with 4 enlarged marginal setae . All dorsal setae except marginal, minute and about same length isignisNewstead(p. 372) - Margins of 7th abdominal segment each with 3 enlarged marginal setae. Dorsal setae, excluding marginal, noticeably longer on head and thorax than on posterior abdominalsegments pseudinsignisGreen(p. 378) Eriococcus aceris (Signoret)(Fig. 3) Acanthococcus aceris Signoret, 1875: 35; Borchsenius, 1949: 347; Danzig, 1964: 632; Tereznikova, 1981: 15. Syntypes $ , SAVOIE and AUSTRIA: on Acer campestre (probably lost).Eriococcus aceris (Signoret) Cockerell, 1896: 323; Schmutterer, 1952: 406.Nidularia aceris (Signoret) Lindinger, 1933: 108. DESCRIPTION. Adult female on slide broadly oval, largest specimens 3-75 mm long, 2-2 mm wide. Surface ofbody strongly nodulose. Anal lobes protruding, about twice as long as wide, apically rounded, moderatelysclerotised with inner margins conspicuously nodulose, each lobe with apical seta 200 ftm long, two innerdorsal enlarged setae and one outer enlarged seta situated towards mid-dorsum; ventral setae flagellate,one subapical, one at outer base and suranal seta longest but shorter than anal ring setae. Dorsal platetriangular, lightly sclerotised, strongly nodulose. Antennae 6- or 7-segmented, 250-330 /im long. Frontallobes developed but smaller than basal antennal segment. Legs normal, well developed. Trochanter +femur 180-220 /xm long, tibia 110-140 /xm long, tarsus about 150 /im long, the tibia + tarsus always longerthan trochanter + femur, claw 35 pm long, stout and curved with denticle near apex. Coxa withouttranslucent pores but with spicules. Labium 160-170 pm long, shorter than clypeolabral shield, basalsegment with 2 pairs of setae. Anal ring with 8 setae, each about 150 /im long. Dorsal surface with numerous enlarged setae of two main sizes, but all with more or less straight sidestapering gradually to a truncate tip, in profile curved and bluntly thorn-like. The largest setae 50 /u,m longon margins of posterior segments, the shortest, which are the most numerous, about 28 /im long.Macrotubular ducts of one size, 30 /xm long, tapering to orifice, the cup wider than base of a dorsal seta,evenly distributed. Microducts about 10 /im long, elongate, each with orifice bifid, evenly distributed. Ventral surface with normal flagellate setae in median areas, short, slender but stiff in lateral areas.Macrotubular ducts of three sizes. A large type similar to dorsal ducts on margins of abdomen. A smallertype about half width of large type, not numerous, in more or less single rows on abdominal segments andaround submargins to head. A small type about half as wide again as the medium-sized ducts in smallnumbers across abdominal segments. Microducts absent. Disc pores predominantly quinquelocular butoccasional pores present that have 6 or 7 loculi; in bands across the abdominal segments, around margins tohead and on mid-venter. Cruciform pores present in a wide submarginal band from head to about fourthabdominal segment. MATERIAL EXAMINED Austria: Weidling, on Acer campestre. Germany: Baden, on A. campestre, 24.V.1916 (H. Wtinn);Munich, on A. platanoides, 26.V.1951 (H. Schmutterer), Erlangen, 6.vi.l949 (H. Schmutterer). Switzer-land: Valais, on A campestre, 20.viii.1906. DISCUSSION. No original material has been traced. The specimens from Austria, one of the type-localities,are from the Naturhistorisches Museum, Vienna, identified by F. Low, and the specimens from Switzer-land are from the collection of P. Marchal. There seems to be no doubt about the identity of the species nowrecognised as such by many modern workers. Eriococcus buxi (Fonscolombe)(Fig. 4) Coccus buxi Fonscolombe, 1834: 218. Syntypes $, FRANCE: Aix[-en-Provence] (probably lost).Eriococcus buxi (Fonscolombe) Targioni Tozzetti, 1868: 726; Fernald, 1903: 72.Nidularia buxi (Fonscolombe) Lindinger, 1933: 108. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 359 Fig. 3 Eriococcus aceris Signoret. Germany, Erlangen. 360 D. J. WILLIAMS Fig. 4 Eriococcus buxi (Fonscolombe). France, Orange, on Buxus sempervirens . THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 361 DESCRIPTION. Adult female broadly oval, largest specimens 1-9 mm long, 1-35 mm wide, not nodulose.Anal lobes about twice as long as wide, conical, each lobe with an apical seta 140 fjun long. Dorsally with anouter enlarged conical seta situated submarginally rather than on edge, and 2 inner enlarged setae.Ventrally with a single slender seta and a suranal seta shorter than anal ring setae. Dorsal plate only lightlysclerotised, rounded. Anal ring with 8 setae, each about 110 /u,m long. Antennae 160-175 /urn long, with 6segments. Frontal lobes well developed, about as long as width of basal antennal segment. Legs robust,hind trochanter + femur 170-190 pm long, hind tibia 80-90 /x,m long, hind tarsus 80-90 fim long, the tibia +tarsus either same length as trochanter + femur or slightly shorter, claw 25 /xm long, straight, with a minutedenticle near apex. Hind coxa with a few small translucent pores. Labium 80 /um long, shorter thanclypeolabral shield, basal segment with a single pair of setae. Dorsal surface densely covered with enlarged setae, lanceolate, 32-52 /am long, pointed, the sidesconcave but curved in profile and thorn-like. Macroducts of one size, evenly distributed, each about 20 pmlong, the cup about two-thirds width of diameter of setal base of a dorsal seta, the inner end of filamentsmall and barely perceptible. Enlarged ducts present, each 35 /u-m long, the sides almost parallel, with innerend flat and rim of orifice sclerotised, but in profile thicker towards one side; present either singly orsometimes in pairs on head margin where there are 2-4 present at most. Microducts in a regulararrangement, each about 12 ^tm long, very slender, the sides parallel, with orifice widely bifid. Ventral surface with normal slender setae in median areas. In submarginal areas the setae are shorter,stiff and blunt. Enlarged setae, same as on dorsum, around margins only. Macroducts of two sizes. A largertype, same as dorsal macroducts, numerous in a wide zone around margins and submargins, on thoraxreaching to spiracles, and on head occupying area between clypeolabral shield and antennal bases. Anarrower type on abdomen only, in transverse rows or bands across median areas of segments. Microducts,same as on dorsum, around margins only. Cruciform pores absent. Disc pores of quinquelocular type,numerous across abdominal segments almost to margins, occasional pores present in median areas ofthorax and head and around spiracular openings. MATERIAL EXAMINED (all on Buxus sempervirens) France: 'south' (det. Signoret); Lyons, 20.vi.1932 (L. Goux); Orange, 15.iv.1978 (D. Matile-Ferrero &D. J. Williams). Switzerland: Rolle, ix.1906 (P. Marchal), others labelled ex coll. P. Marchal. DISCUSSION. The most important characters are the dorsal setae 32-52 /im long, the dorsal macroductsabout half width of setal base of dorsal seta, the enlarged ducts on head only and the lateral enlarged seta onanal lobes submarginal. There seems to be no doubt about the identity of this species, even though the original material cannot betraced. The species was described from Aix-en-Provence and the specimens at hand collected in Orange,and from Lyons not far from the type-locality, are considered by French workers to be this species. Afurther specimen is available from material collected in the south of France and this was identified by V.Signoret as E. buxi. Eriococcussp. near buxi (Fonscolombe)(Fig- 5) DESCRIPTION. Body of adult female on slide, broadly oval, membranous, not nodulose, largest specimens2-3 mm long, 1-3 mm wide. Anal lobes sclerotised, conical, pointed, about twice as long as wide at base,each lobe with an apical seta 180 ^tm long; dorsally with 1 outer enlarged seta on groove on edge and 2lateral enlarged setae; ventrally with 1 slender seta and a suranal seta shorter than anal ring setae. Dorsalplate projecting slightly, rounded. Anal ring with 8 setae, each about 185 /nm long. Antennae 150-210 /u,mlong, 6-segmented. Frontal lobes lightly sclerotised at tip, elongate, sometimes twice as long as width ofbasal antennal segment. Legs well developed, hind trochanter + femur 160-190 pm long, hind tibia 80-95/Ltm long, hind tarsus 80-95 /xm long, the tibia + tarsus either same length as trochanter + femur, or slightlylonger, claw straight, 25 /xm long, with a minute denticle, near apex. Hind coxa with a few translucentpores. Labium 80-95 /u,m long, shorter than clypeolabral shield, basal segment with 1 pair of setae. Dorsal surface with numerous enlarged setae, evenly distributed, each seta broadly lanceolate, pointed,25-45 /xm long, curved in profile. Groups of these setae usually present submarginally on mid-head region,behind eyes, on prothorax, 1st and 7th abdominal segments, each group associated with a single enlargedduct or sometimes 2, each duct 40 /xm long with inner end flat and with sclerotised rim raised fromsurrounding integument. Other enlarged ducts sometimes present on submargins without the groups ofsetae, up to 17 marginal ducts present in some specimens and others present on midline, varying in numberbut there is usually one on meso thorax and another on 1st abdominal segment. Macroducts of one size,numerous, evenly distributed each duct about 25 /urn long, with a clear circular rim surrounding orifice, the 362 D. J. WILLIAMS .- ' -V^i 1 * .%' -^ NK'r^v.'v^ V.: // f x Fig. 5 Eriococcus sp. near fcux/ (Fonscolombe). U.S.S.R. , Crimea, on Buxus sp. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 363 cup about half width of setal base of a dorsal seta, the inner end of filament resembling a bunch of grapes.Microducts numerous, each about 12 pm long, slender, with parallel sides and with orifice widely bifid. Ventral surface with normal slender setae in median areas and laterally towards margins. Enlarged setaesame as on dorsum, around margins. Macroducts of two sizes. A large type, same size as dorsal ducts,present around margins and submargins, reaching to spiracles. A narrower type, not numerous in medianareas of abdomen only, each duct with a distinct rim around orifice. Microducts on margins, same shape ason dorsum. Quinquelocular pores numerous on abdomen where they almost reach margins; others presentin median areas of head and thorax and around spiracular openings, those next to first spiracle extending tomargin. Cruciform pores absent. MATERIAL EXAMINED U.S.S.R.: Crimea, Massandra, on Buxus sp., 24.vii.1951 (T. Bustchik). Turkey: Gerede, on B.sempervirens , 13.vii.1970. DISCUSSION. The large number of submarginal enlarged ducts, many surrounded by groups of setae, andother ducts on the midline, easily distinguish this species from E. buxi. Specimens from Crimea have thedorsal setae 25-40 /xm long and noticeably shorter than those in E. buxi; in Turkish specimens, although28-45 /xm long, these setae are still shorter than in E. buxi. The position of the lateral enlarged seta on eachanal lobe is always marginal, whereas in E. buxi this seta is displaced slightly to dorsum. Another maindifference is the size of a dorsal macroduct, which in this species has the cup about half the width of a setalbase, but in E. buxi it is about two-thirds the width. Furthermore, the dorsal ducts have a conspicuous rimaround the orifice. There is every indication that this species is new but for the moment it is left un-named pending furtherresearch by workers in U.S.S.R. It seems certain that it is the same as that described by Borchsenius (1949)and Tereznikova (1981) as E. buxi. Eriococcus cantium sp. n. (Fig. 6) DESCRIPTION. Adult female, when prepared on microscope slides, elongate-oval with almost parallel sides,2-6 mm long and 1-00 mm wide, nodulose. Anal lobes about twice as long as wide, moderately sclerotised.Each lobe with an apical seta 300 /im long, dorsally with 1 inner and 3 outer enlarged setae and ventrallywith 2 slender setae and a slender suranal seta shorter than anal ring setae. Dorsal plate not prominent,rounded, nodulose. Antennae 7-segmented, 300 j*m long. Frontal tubercle present, just anterior to basalantennal segment. Legs well developed, hind trochanter + femur 250 /u,m long, hind tibia 150 /im long,hind tarsus 160 /xm long, claw curved 40 /urn long, with a denticle near apex. Hind coxa with minutetranslucent pores on outer half and hind femur with a small group on mid-anterior margin. Labium 100 /xmlong, slightly shorter than clypeolabral shield and basal segment with 2 pairs of setae. Anal ring with 8setae. Dorsal surface with enlarged conical setae, pointed, 35-75 /im long, in bands across the segments. On7th segment medially a group of 4-5 present and on 6th segment a similar group present but this mergingwith lateral setae. Macroducts of one size, about 25 /xm long, the cup narrower than setal bases of largestsetae, fairly evenly distributed across segments. Microducts not numerous, each about 4 ^tm long, withshort collar, ampulla and swollen inner end to tube. Ventral surface with normal slender setae in median areas. Enlarged setae, same as on dorsum, onmargins of head, thorax and anterior abdominal segments, but submarginally a few setae that are moreslender but stiff. Macroducts of two sizes; a larger type, same size as on dorsum, around margins and anarrower type in transverse rows on abdominal segments, and in median to submarginal areas of thoraxand head. Microducts in small numbers on margins. Disc pores of quinquelocular type numerous acrossabdominal segments, in median areas of thorax and head and near spiracles. Cruciform pores present insmall numbers in a narrow submarginal zone from about 5th abdominal segment forward to head. Holotype $, England: Kent, Bearsted, on Brachypodium sylvaticum (Poaceae), 30.vii.1925 (E. E.Green) (BMNH). Paratype. England: 1 $ , same data as holotype (BMNH). DISCUSSION. This species is close to E. greeni in having a similar distribution of enlarged dorsal setae. Itdiffers, however, in possessing groups of 4-6 setae in the median areas of the seventh and sixth segmentswhereas in E. greeni these groups usually possess only 2 or at most 3 setae. The median group of setae onthe sixth segment of E. cantium is not so differentiated and is almost continuous with the lateral setae.Furthermore there are three lateral setae on each anal lobe whereas in E. greeni there are only two.The name is the Latin word for the county of Kent, and is used as a noun in apposition. 364 D. J. WILLIAMS Fig. 6 Eriococcus cantium sp. n. England, Kent, Bearsted, on Brachypodium sylvaticum. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 365 Eriococcus devoniensis (Green) comb. rev.(Fig. 7) Rhizococcus devoniensis Green, 1896: 260. LECTOTYPE $>, ENGLAND (BMNH), here designated [examined] . Eriococcus devoniensis (Green) Cockerell, 1897: 589; Newstead, 1903: 201; Schmutterer, 1952: 413.Acanthococcus devoniensis (Green) Borchsenius, 1949: 337. DESCRIPTION. A broadly oval species, attaining a length of 2-0 mm and a width of 1-3 mm, body nodulose.Anal lobes sclerotised, rather wide and pointed, triangular, about as wide as long at base in olderspecimens, but conical and rounded in young specimens. Each lobe with an apical seta 180 /u,m long anddorsally 2 inner submarginal enlarged setae and 1 outer marginal seta. Ventrally each lobe with 2 slendersetae and a suranal seta shorter than anal ring setae. Dorsal plate poorly developed, rounded. Anal ringwith 8 setae, each 100 /x,m long. Antennae 190-210 /im long, 7-segmented. Frontal lobes conspicuous andlarger than a basal antennal segment. Legs well developed. Hind trochanter + femur 180-200 pm long,hind tibia 100-115 /um long, hind tarsus 110-120 /xm long, the tibia + tarsus longer than trochanter +femur, claw stout and curved, 30 /u,m long, with a denticle near apex. Hind coxa with a few noticeabletranslucent pores and hind femur with a few on anterior edge. Labium 130-150 /im long, usually slightlyshorter than clypeolabral shield, basal segment with 2 pairs of setae. Dorsal surface with an even distribution of enlarged setae of various sizes, 28-40 /im long, each truncate,the sides straight or only slightly concave, and tapering gradually. Macroducts of one size, about 20 /nmlong, the cup about same size or smaller than diameter of setal base of largest setae, fairly numerous.Microducts each about 4 /u,m long, with ampulla, inner end of tube swollen, but without internal collar,opening to exterior with minute orifice, in moderate numbers over entire surface. Ventral surface with normal slender setae in median areas and enlarged setae similar to those on dorsum,in a marginal zone, on thorax, reaching almost to spiracles. Macroducts of two sizes. A large type, same asthose on dorsum, in a marginal zone, interspersed with the enlarged setae. A narrower type present acrossmedian areas of abdominal segments and in groups in median areas of thorax and head. Microducts few, onmargins only. Disc pores predominantly quinquelocular, numerous on abdomen and present in medianareas of thorax and around spiracles. Cruciform pores few, in submargins of head and thorax. MATERIAL EXAMINED (all on Erica tetralix unless stated otherwise) England: Devon, Budleigh Salterton, ix.1896 (lectotype), 20.ix.1896 (E. E. Green), Sherbrook,18.viii.1896; Surrey, Camberley, ll.iv.1922, vii.1914, ix.1914, xi.1929 (E. E. Green), 10.vi.1959 (D. J.Williams); Durham, Waldridge Fell, vi.1950, 16.viii.1960 (D. J. Williams); Yorkshire, Hawkesworth,12.viii.1960, Levisham, 27.viii.1959, Goathland, 25.viii.1959 (all D. J. Williams), Ickornshaw Moor,29.viii.1948 (A. Smith), Shipley Glen, 25.viii.1984 (N. B. Pungerl). Ireland: Kerry, Mt Mangerton, insphagnum, viii.1943 (R. S. Bagnalf). Austria: Lunz, on E. cornea. Germany: Ratzeburg. Corsica:Vizzavone Forest, on E. arborea, l.xi.1930 (A. Balachowsky). DISCUSSION. Green's original slide contains three specimens, and the lectotype selected is clearly marked.The two other specimens are here designated paralectotypes. Although Green described this insect originally on Erica cinerea he (Green, 1917) corrected themisidentification of the host-plant to E. tetralix and the insect seems to be restricted to this plant in Britain,even when other species of Erica are growing next to it. It distorts the stems at the point of feeding, so muchso that stems may curl in a circle enclosing the insect in the middle as illustrated in colour by Newstead(1903). Specimens are at hand from Corsica collected on E. arborea and the record led Green (1931) tocorrect an early impression that E. tetralix was the only host-plant. There are unconfirmed records on E.cinerea and Calluna by Goux (1934) and on C. vulgaris by Balachowsky (1937). Apart from the British records already listed, Harrison (1916a, 1916ft, 1918, 1948, 1949) has recorded itfrom Northumberland, Durham and Yorkshire, and from western Scotland including the Isle of Rhum andBenbecula, Ronay and Grimsay in the Outer Hebrides. Other records include those of Green (1917) fromCheshire and Green (1923) from Aberdeenshire, while Killington (1936) recorded it from Hampshire. The species is found throughout Europe as far north as Sweden and may be present throughout thePalaearctic region wherever E. tetralix is found. 366 D. J. WILLIAMS . . * ^ fi ^* , Fig. 7 Eriococcus devoniensis (Green). England, Durham, Waldridge Fell, on c tetralix. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 367 Eriococcus glyceriae Green comb. rev.(Fig. 8) Eriococcus glyceriae Green, 1921: 146; Schmutterer, 1952: 410. LECTOTYPE $, ENGLAND (BMNH), here designated [examined].Nidularia glyceriae (Green) Lindinger, 1933: 116.Greenisca glyceriae (Green) Borchsenius, 1949: 368; Danzig, 1964: 634; Danzig, 1980: 228; Tereznikova, 1981: 52.Kaweckia glyceriae (Green) Koteja & Zak-Ogaza, 1981: 506. DESCRIPTION. Body elongate-oval, largest specimens 3-4 mm long, 1-7 mm wide, membranous throughout,not nodulose. Anal lobes either membranous or faintly sclerotised, each lobe about as long as wide, with anapical seta 140 //,m long and dorsally 2 inner and 1 outer truncate setae and one ventral seta that is flagellateand situated towards outer margin. Suranal setae slender and flagellate, shorter than anal ring setae.Antennae 7-segmented, 220-230 tim long. Frontal tubercle present just anterior to each basal antennalsegment. Legs well developed, hind trochanter + femur 180 /im long, hind tibia 110 /u,m long, hind tarsus110-120 /Am long, the tibia + tarsus always longer than trochanter + femur, claw 35 /AHI long, curved withminute denticle near apex. Hind coxa with proximal edge indistinct, translucent pores numerous, a fewextending on to surrounding integument at base. Labium 85-100 /Am long, shorter than clypeolabral shield,basal segment with 2 pairs of setae. Spiracles heavily sclerotised, this sclerotisation completely surroundingatrium. Dorsal plate membranous in form of a narrow rounded lobe. Dorsal surface with enlarged setae on 5th and posterior segments only except for an occasional seta on4th segment, few, on margins, each seta with almost parallel sides, truncate, 15-25 /Am long. Other dorsalsetae slender, often curved but stiff, with blunt tip 12-20 ttm long. Macroducts of two sizes, the largestabout 16-20 tim long with the cup about same diameter as an enlarged setal base, in bands across thesegments except on anal lobe segment where they are replaced by a narrower type. Microducts minute,about 6 /AHI long, with ampulla, swollen inner end to tube and an inner collar, few, across the posteriorabdominal segments and around the margins to head. Disc pores, usually with 7 loculi, in more or lesssingle rows at anterior edges of segments. Cruciform pores present in wide bands across median areas ofsecond abdominal segment and forward to mesothorax; a few others present in lateral areas of thesesegments, on third abdominal segment and on prothorax. Ventral surface with normal setae in median areas, replaced by shorter stiff setae, similar to those ondorsum, around margins of thorax and anterior abdominal segments. Macroducts, of the large type, similarto those on dorsum, on margins only. A narrower type in moderate numbers across abdominal segments, inmedian areas of thorax and head and reaching to submargins of thorax. Microducts as on dorsum, few, onmargins. Cruciform pores restricted to a few only on margins of prothorax and head. Disc pores, usuallywith 7 loculi, but occasionally with more or fewer loculi at anterior and posterior edges of 5th and posteriorsegments, at anterior edges of anterior abdominal segments, in median area of thorax and in a zone aroundthoracic margins to head. MATERIAL EXAMINED England: Norfolk, Blakeney Point, on Glyceria maritima (now Puccinellia maritimd), vii.1920 (E. E.Green) (lectotype), ix.1920 (N. E. Brenchley). Hungary: Zicsiujfalu, on Agropyron sp., 8.ix.l981 (F.Kozar). U.S.S.R.: Odessa, on A. repens (A. Kiritshenko) . DISCUSSION. The lectotype is here designated from three specimens on the same slide and clearly indicated.The two other specimens are labelled paralectotypes. Although this species has concentrations of cruciform pores on the mid-dorsum and rather heavilysclerotised spiracles, these characters are probably of specific significance only and the species seems to becongeneric with the type-species. Specimens available from eastern Europe tend to have the enlarged setaewith slightly concave sides. Eriococcus green/ Newstead comb. rev.(Fig. 9) Eriococcus greeni Newstead, 1898: 96; Newstead, 1903: 200. LECTOTYPE $, ENGLAND (BMNH), here designated [examined] .Acanthococcus greeni (Newstead) Borchsenius, 1949: 340; Danzig, 1975: 71; 1980: 212; Tereznikova, 1981:27. DESCRIPTION. An elongate-oval species, attaining a length of 3-0 mm and a width of 1-6 mm, body 368 D. J. WILLIAMS Fig. 8 Eriococcus glyceriae Green. England, Norfolk, Blakeney Point, on Pucclnellia maritima. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 369 Fig. 9 Eriococcus greeni Newstead. England, Surrey, Camberley, on grass. 370 D. J. WILLIAMS nodulose. Anal lobes about twice as long as wide, conical, tending to be pointed, moderately sclerotised.Each lobe with an apical seta 380 /xm long and dorsally with 1 outer and 2 inner enlarged setae and ventrallywith 2 slender setae and a slender suranal seta much shorter than anal ring setae. Dorsal plate sclerotised,moderately developed, rounded but nodulose. Anal ring with 8 setae each about 140 /u,m long. Antennae280-300 /xm long, with 6 or 7 segments. Frontal tubercle present just anterior to basal segment. Legs welldeveloped. Hind trochanter + femur 240-270 /xm long, hind tibia 150-170 /u,m long, hind tarsus 150-170/im long, tibia + tarsus always longer than trochanter + femur. Claw curved, 35 /u,m long, with a minutedenticle near apex. Hind coxa with translucent pores on outer half and hind femur with a small group onmid-anterior edge. Labium 110-120 /am long, shorter than clypeolabral shield. Dorsal surface with pointed conical setae 30-70 jam long, the sizes in a haphazard arrangement but thereare always some of the longest in the marginal groups. On the mid-dorsum of 7th abdominal segment, thereis usually a group of 2 such setae, rarely 3, and a similar group is present on the mid-dorsum of 6thabdominal segment. Macroducts in a regular distribution, of one type, about 25 jum long, the cup smallerthan diameter of setal base of largest setae , tapering gradually , the orifice usually surrounded by a pale ovalarea. Microducts about 4 jam long, each with ampulla, swollen inner end to tube and an internal collar,evenly distributed. Ventral surface with normal slender setae in median areas, marginally with a few enlarged setae similarto dorsal setae, submarginally with more slender but stiff setae. Macroducts of two sizes. A large type,same as on dorsum, in small numbers around margins. A narrower type in bands across abdominalsegments, in median areas of thorax and head, extending in submargins almost to marginal macroducts.Microducts not numerous, in a marginal zone from anterior abdominal segments to head. Quinquelocularpores numerous on abdomen, in median areas of thorax and around spiracles. Cruciform pores few, insubmarginal areas only of head, thorax and first few abdominal segments. MATERIAL EXAMINED England: Devon, Budleigh Salterton, on grass, 20. ix. 1896 (E. E. Green) (lectotype); Somerset,Minehead, on grass, ix.1920 (E. E. Green), Cheddar, vii.1926 (E. E. Green); Surrey, Camberley,31.viii.1914, ix.1914, 25. ix. 1934, Hog's Back, 21.vii.1922, on grass (all E. E. Green); Kent, Thurnham, ongrass, 15. ix. 1926 (E. E. Green); Berkshire, Silwood Park, on Festuca sp., Deschampsia caespitosa,27.x. 1948, on grass 19.viii.1948 (all K. L. Boratynski);YorkshiTe, Hawkesworth, on grass, 20.viii.1961 (D.J. Williams). Scotland: E. Lothian, Gullane, on grass, viii.1925 (E. E. Green). DISCUSSION. The lectotype designated is the only original specimen seen. In all specimens there is always agroup of 2 large setae on the mid-dorsum of the 7th abdominal segment or rarely with a small seta forming agroup of 3. Furthermore, there is always a median group of setae on the 6th segment. These charactersseparate it from R. cantium herein described as new, a species that has a group of 4 on the mid-dorsum ofthe 7th segment and with the setae on the mid-dorsum of 6th segment not forming a distinct group. Green (1923, 1926, 1928) recorded this species also from Sea View, Isle of Wight; Bearsted, Kent;Yately, Hampshire; and from Monument Hill in East Lothian (Haddington). Eriococcus incrmis Green comb. rev.(Fig. 10) Eriococcus inermis Green, 1915: 176. LECTOTYPE $ , ENGLAND (BMNH), here designated [examined]. Nidularia inermis (Green) Lindinger, 1933: 116. Greenisca inermis (Green) Borchsenius, 1948: 502; 1949: 367. Anophococcus inermis (Green) Balachowsky, 1954: 61. Rhizococcus inermis (Green) Danzig, 1962: 854; 1964: 634; Dziedicka & Koteja, 1971: 576. Acanthococcus inermis (Green) Danzig, 1975: 64; 1980: 226. DESCRIPTION. Body of adult female not nodulose, elongate-oval, anterior end rounded, sides at timessubparallel, largest specimens 2-5 mm long, 1-5 mm wide. Anal lobes about twice as long as wide, pointed.Each lobe with an apical seta 180 /urn long, dorsally with 2 inner enlarged conical setae, the anterior usuallythe larger, about 28 /u,m long, the posterior about 24 /im long, and a minute outer seta near base; ventrallywith 2 slender setae and a suranal seta shorter than anal ring setae. Dorsal plate rounded, lightlysclerotised. Anal ring with 8 setae each 130 /urn long. Antennae 150-210 |u,m long, 6-segmented. Frontaltubercle present just anterior to basal antennal segment. Legs well developed, hind trochanter + femur140-170 fj.m long, hind tibia 90-110 /im long, hind tarsus 110-130 pm long, the tibia + tarsus conspicuouslylonger than trochanter + femur, claw curved, 30 /nm long, with a denticle near apex. Hind coxa with a fewlarge translucent pores on outer half and hind femur sometimes with one or two on anterior edge. Labium95-110 fim long, shorter than clypeolabral shield, basal segment with 2 pairs of setae. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 371 Fig. 10 Eriococcus inermis Green. England, Surrey, Camberley, on grass. 372 D. J. WILLIAMS Dorsal surface with minute bluntly pointed setae, most scarcely longer than diameter of setal base andsome appreciably shorter, in moderate numbers across segments. Macroducts evenly distributed, eachabout 20 /um long with cup nearly twice diameter of a dorsal setal base. Microducts each about 4 /um long,with ampulla, tube with inner end swollen, and an internal collar, in moderate numbers over surface. Ventral surface with slender setae in median areas and minute setae, similar to those on dorsum, aroundmargins and submargins. Macroducts of two sizes. A larger type, the same as dorsal ducts, on margins only.A narrower type in median areas as far forward as head margins. Disc pores varying in size, numerous onabdomen where, on the posterior segments, many have 7 or occasionally more loculi but anteriorly on theabdomen, the median area of thorax and around the spiracles, they are quinquelocular. Cruciform porespresent in a submedian zone from head to about 5th abdominal segment. MATERIAL EXAMINED England: Surrey; Camberley, on grass, ix.1914 (E. E. Green) (lectotype, paralectotypes), 15.iv.1922,l.x.1925, on Festuca ovina, ix.1929 (E. E. Green); Somerset, Cheddar, viii.1926 on F. ovina (E. E. Green);Berkshire, Silwood Park, on F. ovina, Deschampsia flexuosa 1948-1956 (various dates) (K. L. Bora-tynski); Yorkshire, Steeton Moor, 29.viii.1948 (A. Smith), Hawkesworth, on grass, 20.viii.1961 (D. J.Williams); Durham, Waldridge Fell, on grass, 17.vii.1960 (D. J. Williams). DISCUSSION. The original material comprises 12 specimens, six on each of two slides, one labelled 'type'and the other labelled 'co-type'. The lectotype has been selected from the slide labelled 'type' and is clearlymarked, the remaining 11 specimens are labelled paralectotypes. The distribution at present also includes France, western U.S.S.R. and the far eastern area of U.S.S.R.It will probably be found throughout the Palaearctic Region. Eriococcus ins/gn/s Newstead comb. rev.(Fig. 11) Eriococcus insignis Newstead, 1891: 164; 1903: 198. LECTOTYPE $, ENGLAND (BMNH), here desig-nated [examined]. Nidularia insignis (Newstead) Lindinger, 1933: 116. Rhizococcus insignis (Newstead) Borchsenius, 1949: 357; Danzig, 1962: 841; 1964: 633; Dziedicka &Koteja, 1971: 561. Acanthococcus insignis (Newstead) Danzig, 1975: 64; Tereznikova, 1981: 29. DESCRIPTION. Body of adult female elongate-oval, the sides often subparallel, nodulose, largest specimens2-9 mm long and 1-2 mm wide. Anal lobes conical, pointed, about twice as long as wide, moderatelysclerotised. Each lobe with an apical seta 280-320 /um long, on the dorsum 1 outer and 2 inner enlargedsetae, on the venter 2 slender setae and a suranal setae shorter than anal ring setae. Dorsal plate rounded,narrow, sometimes nodulose and lightly sclerotised. Anal ring with 8 setae, each 135 /um long. Antennae200-270 /um long with 7 segments. Legs well developed, hind trochanter + femur 200-250 /um long, hindtibia 120-160 /um long, hind tarsus 130-150 /um long, the tibia + tarsus always longer than trochanter +femur, claw curved, 35 /um long, with a denticle near apex. Hind coxa with conspicuous translucent poreson outer half. Labium 100 /um long, shorter than clypeolabral shield, basal segment with 2 pairs of setae. Dorsal surface with a marginal row of enlarged setae 40-65 /um long, each with slightly concave sides. Onsegments 5-7 these setae are truncate, but anteriorly they are bluntly pointed and on head they extend tosub marginal area near midline. The setae on the 7th abdominal segment always number 4. Elsewhere onthe dorsum there are small setae in moderate numbers, 6-8 /urn long. These vary in thickness on differentspecimens, even from the same batch, sometimes being quite slender but usually they are slightly conicaland often they are thicker on the head and thorax. Despite these differences they vary little in length.Macroducts of one size, evenly distributed, about 25 /um long, cup narrower than a setal base of a marginalseta. Microducts in an even distribution, each about 4 /um long with ampulla, tube with swollen inner end,and an internal collar. Ventral surface with normal slender setae in median areas, a few enlarged setae on head margin and stiffsetae in a marginal zone from head to anterior abdominal segments, these setae usually larger than thesmall dorsal setae but smaller than dorsal marginal setae. Macroducts of two sizes. A larger type, same ason dorsum around margins, and a narrower type in median areas, extending laterally almost to marginalmacroducts. Microducts few, around margins only. Disc pores usually with 7 loculi, fairly numerous acrossabdominal segments, but not reaching margins, also present in median areas of thorax and head andaround spiracular openings. Cruciform pores in a narrow submarginal zone around entire body, notnumerous. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 373 . -A Fig. 11 Eriococcus insignis Newstead. England, Surrey, Camberley, on grass. 374 D. J. WILLIAMS MATERIAL EXAMINED England: Cheshire, Ince, on Agrostis sp., 1890 (lectotype, paralectotypes) ; Surrey, Camberley, 1914-1934 (various dates), on Hypericum sp., ix.1923, Wisley, ix.1915, Box Hill, 26. ix. 1921; Sussex, Arundel,ix.1914; Kent, Thuraham, 15. ix. 1926; Somerset, Cheddar, viii.1926, on Brachypodium sylvaticum,viii.1926, Minehead, ix.1920; Warwickshire, Morton Paddocks, 14.viii.1920 (R. Newstead); Devon,Budleigh Salterton, 10, 20.ix.1896; Yorkshire, Goathland, 25.viii.1959 (D. J. Williams). Scotland: E.Lothian, Gullane, vii.1925; Argyllshire, Skye, Uig, on grass, 8.ix.l973 (D. J. Williams). Channel Is:Guernsey, Houmet Homtolle, 9.ix.24; Herm, 13. ix. 1924, all on grasses (all E. E. Green). DISCUSSION. The lectotype has been selected and clearly marked, from four specimens on a single slidelabelled 'Cotype $ $' by Newstead. The three other specimens are labelled paralectotypes. A discussion of the similarities between this species and E. pseudinsignis is given under the latter. Atpresent E. insignis may be identified by the presence of 4 marginal setae on each side of the seventhabdominal segment instead of 3 in E. pseudinsignis. Literature records for Britain include those of Green (1915, 1925ft, 1926, 1928) for Camberley, theChannel Is, Wales (Breconshire) and Cheddar. Eriococcus lagerstroemiae Kuwana comb. rev. Eriococcus lagerstroemiae Kuwana, 1907: 182; Boratynski & Williams, 1964: 91. Syntypes $, JAPAN: Ichijiku and Sarusuberi (IAES). Nidularia lagerstroemiae (Kuwana) Lindinger, 1933: 116.Acanthococcus lagerstroemiae (Kuwana) Borchsenius, 1960: 214. This species was recorded from England by Green (1915) from specimens collected in a nursery garden atSt. Albans on Lagerstroemia sp., growing in the open. It is not clear if the insects were destroyed at thetime. The species has not been recorded from Britain since, but it has been found also in India and China. No authentic material has been studied for this work, although some specimens are at hand from Japan,Tokyo, on L. indica sent by Kuwana to Green. The species is in great need of further study. Material shows that the species possesses frontal lobes, numerous slender microducts with bifid orifice,and a small group of setae on mid-dorsum of the anal lobe segment. It comes close to E. aceris but differs inhaving the marginal setae, especially those on the posterior segments, not appreciably longer than those onthe mid-dorsum. In E. aceris the marginal setae are about twice the length of the mid-dorsal setae. Eriococcus munroi (Boratynski) comb. n.(Fig. 12) Acanthococcus munroi Boratynski, 1962: 56; Danzig, 1975: 71; 1980: 218; Tereznikova, 1981: 32.Holotype $ , ENGLAND (BMNH) [examined] . DESCRIPTION. Body of adult female nodulose, elongate-oval, widest at mesothorax, largest specimens 2-1mm long and 1-5 mm wide. Anal lobes about twice as long as wide, almost conical, sclerotised, each withapical setae up to 300 /u,m long. Dorsal surface of each lobe with 2 inner enlarged conical setae, thesubapical longer than inner and with an outer enlarged seta, the longest, situated slightly on dorsumtowards base. Dorsal plate sclerotised, narrow and rounded, the posterior edge with 'teeth' because of thenodulosity. Anal ring with 8 setae each about 125 /urn long. Antennae usually 7-segmented, 250-300 /imlong. Frontal lobes present just anterior to basal segment. Legs normal, well developed, hind trochanter +femur 220-230 /urn long, hind tibia 130-150 /urn long, hind tarsus 140-150 jiim long, the tibia + tarsus alwayslonger than trochanter + femur, claw curved, 30 /urn long with denticle near apex. Hind coxa with a fewconspicuous translucent pores. Labium 120 /u,m long, shorter than clypeolabral shield, basal segment with 2pairs of setae. Dorsal surface with enlarged setae, each conical with almost straight sides and with blunt tip, arranged intransverse bands across the segments but absent in the submarginal areas. These setae of various sizes,30-60 /um long but always with one or two of the large type present in any marginal group. On the medianareas of 4th and 5th abdominal segments the conical setae tend to be shorter but the normal conical setaeare replaced on the median areas of 6th and 7th abdominal segments by minute conical setae about 6 /u,mlong and rarely longer than the diameter of a setal base, there being usually 4 across the middle of eachsegment. Macroducts of one size, each about 20 /urn long with almost straight sides and with cup narrowerthan a setal base of an enlarged seta, fairly numerous in wide bands across the segments. Microducts eachabout 4 fjim long with inner end of tube swollen, ampulla and a saucer-shaped inner collar, evenlydistributed. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 375 Fig. 12 Eriococcus munroi (Boratynski). England, Berkshire, Silwood Park, on Veronica chamaedrys. 376 D. J. WILLIAMS Ventral surface with normal slender setae in median areas and a few enlarged conical setae on marginsand submargins. Macroducts of two sizes. A larger type, similar to dorsal ducts, around margins and anarrower duct in transverse rows on abdominal segments, in median and submarginal areas of thorax andhead. Microducts few, around margins only. Cruciform pores not numerous, in submarginal areas of 6thabdominal segment and forward to head. Disc pores, usually quinquelocular, numerous in bands acrossabdominal segments but not reaching margins, present also in median areas of thorax and around thespiracles. MATERIAL EXAMINED England: Berkshire, Silwood Park, on Achillea millefolium (Asteraceae), 7.ix.l948 (holotype) andsubsequent specimens collected on the same host plant, on Veronica chamaedrys (Scrophulariaceae),1948, 1949 (various dates), on Deschampsia caespitosa (Poaceae), 1949, Chrysanthemum leucanthemum,21. vi. 1949 and on 'thistle' (Asteraceae), 7.ix.l949 (all K. L. Boratynski); Surrey, Camberley, on grass,ix.!914(E. E. Green); Somerset, Cheddar, on Teucrium scorodoniae (Lamiaceae); Isle of Man, Port Erin,27.ix.1918. DISCUSSION. Boratynski (1962) has given an extended description and an account of the biology of thisspecies. He has also discussed some specimens available labelled ex coll. Manchester Museum, withoutlocality data. These have the enlarged setae wider, shorter and more rounded or almost truncate at theposterior end of body. The single specimen available on Teucrium scorodonia has the minute setae on thesixth segment replaced by enlarged setae, and the posterior marginal setae are conspicuously rounded. Theextent of variation is still not clear. Since its description the species has been found throughout Europe andthefareastofU.S.S.R. Eriococcusplacidus Green comb. rev. (Fig. 13) Eriococcus placidus Green, 1915: 148. LECTOTYPE $, ENGLAND (BMNH), here designated [ex-amined]. Nidularia placida (Green) Lindinger, 1933: 116.Greenisca placida (Green) Rasina, 1955: 69; Danzig, 1964: 634. DESCRIPTION. Body of adult female elongate-oval, not nodulose, largest specimens 3-5 mm long, 1-7 mmwide. Anal lobes conical, pointed, about twice as long as wide, moderately sclerotised, each lobe with anapical seta 220-300 /um long, the dorsum with 2 inner enlarged conical setae, the anterior 36-50 /nm long,the posterior usually shorter, 32-40 /urn long, and an outer minute seta situated, in most specimens,towards centre. Ventral surface of lobe with 2 slender setae and a slender suranal seta. Dorsal platemoderately developed, lightly sclerotised, outer edge rounded, often nodulose. Anal ring with 8 setae each145 /im long. Antennae 280-300 /urn long, with 7 segments. Frontal tubercles present, minute, just anteriorto each basal antennal segment. Labium 100-120 /urn long, shorter than clypeolabral shield, basal segmentwith 2 pairs of setae. Legs well developed, slender, hind trochanter + femur 230-240 /xm long, hind tibia140-150 /im long, hind tarsus 160-170 /xm long, claw slightly curved, slender, 40 /xm long, with a minutedenticle near apex. Hind coxa with conspicuous translucent pores on outer half, hind femur with a smallgroup at mid-anterior margin. Dorsal surface with two types of setae. Apart from the enlarged setae on anal lobes there are otherspresent about 30 /xm long, on head margin in varying numbers but there are usually one or two present.Occasionally these are replaced by slender setae. Elsewhere the dorsum is beset with minute pointed setaescarcely more than 20 pm long, but often shorter, rarely more than twice as long as diameter of a setal base.Macroducts fairly evenly distributed, each about 25 /urn long, the cup about 2-3 times as wide as diameterof setal base of a small seta. Microducts in a regular arrangement, each about 6 /xm long, with ampulla, tubewith inner end swollen, and inner collar. Quinquelocular pores present in single to double rows mainly atanterior and posterior edges of segments, each pore with wide sclerotised rim and about half width of adiameter of cup of macroduct. Ventral surface with normal slender setae of various sizes. Short stiff setae present, often longer thansmall dorsal setae, on margins of head, thorax and anterior abdominal segments. Macroducts of two sizes.A larger type, same as on dorsum, around margins. A narrower type present in median to submarginalareas. Microducts apparently absent. Disc pores always smaller than dorsal pores, fairly numerous onabdomen, where on the posterior segments they have usually 7 loculi, but on the anterior abdominalsegments, thorax and head they are quinquelocular. Cruciform pores not numerous, in a narrowsubmarginal zone from about 5th abdominal segment forward to head. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 377 Fig. 13 Eriococcus placidus Green, England, Kent, Thurnham, on Avenaflavescens. 378 D. J. WILLIAMS MATERIAL EXAMINED England: Kent, Thurnham, on IFestuca sp., 8.ix.l920 (E. E. Green) (lectotype, paralectotypes), onAvena flavescens, Brachypodiwn sylvaticum, 13.vii.1921 (E. E. Green), Bearsted, on B. sylvaticum,30.vii.1925 (E. E. Green); Isle of Wight, on B. sylvaticum, 3.ix.l921 (E. E. Green); Berkshire, Wytham,on B. pinnatum, 24.viii.1966, 30.viii.1968 (G. Varley), x.1969 (D. Manawadu). Germany: Ebernburg, onB. pinnatum, 16.ix.1928 (H. Wiinn). DISCUSSION. The original slide contains four specimens and the specimen selected as lectotype is clearlymarked, the other three specimens are labelled paralectotypes. This species is easily recognisable by the dorsal quinquelocular pores that have wide rims and are alwayslarger than the ventral disc pores. Green (1923) recorded the species from the Isle of Wight and from Box Hill in Surrey. Eriococcus pseudinsignis Green comb. rev.(Fig. 14) Eriococcus pseudinsignis Green, 1921: 149; Schmutterer, 1952: 407. LECTOTYPE $, ENGLAND (BMNH), here designated [examined].Nidularia pseudinsignis (Green) Lindinger, 1933: 116.Rhizococcus pseudinsignis (Green) Borchsenius, 1949: 354; Danzig, 1962: 845; 1964: 633; Dziedicka & Koteja, 1971.Acanthococcus pseudinsignis (Green) Tereznikova, 1981: 35. DESCRIPTION. Slide-mounted specimens, elongate-oval, the largest 2-8 mm long and 1-4 mm wide,nodulose. Anal lobes conical, about twice as long as wide, sclerotised. Each lobe with an apical seta 280 ju,mlong; dorsally 1 outer and 2 inner enlarged setae and ventrally 2 slender setae and a suranal seta shorterthan anal ring setae. Dorsal plate lightly sclerotised, narrow, rounded and slightly nodulose. Anal ring with8 setae each about 140 /im long. Antennae 250-300 /AHI long with 7 segments. Frontal tubercle present justanterior to basal segment. Legs well developed, slender, hind trochanter -I- femur 230-260 /urn long, hindtibia 130-150 /im long, hind tarsus 150-160 /im long, claw 35 /u,m long, curved, with a small denticle nearapex. Coxa with a few translucent pores on outer half. Labium 100-120 pm long, shorter than clypeolabralshield, basal segment with 2 pairs of setae. Dorsal surface with a single row of enlarged marginal setae, 28-60 /u,m long, those on head extending tosubmarginal areas. Each seta almost conical but inner edges sometimes slightly concave or convex. On theposterior abdominal segments these setae are bluntly pointed, tending to be truncate, but anteriorly theyare more sharply pointed. Margins of 7th abdominal segment always with 3 setae varying little in length.Elsewhere on the dorsum the setae on the head and thorax are conical, often curved, sometimes onlyslightly smaller than the enlarged setae on head, but usually 12-15 /am long. On the posterior abdominalsegments the setae become shorter and more slender, 6-8 /xm long. Macroducts of one size in moderatenumbers, fairly evenly distributed, about 25 /im long, the cup only a little narrower than diameter of a setalbase of marginal seta. Microducts not numerous, about 4 /urn long, with inner end of tube swollen, ampullaand internal collar. Ventral setae normal in median areas but laterally towards margins they are conical and thick but not solarge as dorsal marginal setae. Macroducts of two sizes. A larger type, similar to dorsal ducts, aroundmargins only. A narrower type in bands across abdominal segments in median areas of head and thorax,extending to submargins. Microducts around margins only. Cruciform pores in a narrow submarginal zoneand in area between clypeolabral shield and antennae. Disc pores, usually with 7 loculi, numerous onabdominal segments but not reaching margins, present also in median areas of thorax and around spiracles. MATERIAL EXAMINED England: Kent, Thurnham, on Festuca sp., 9.ix.l920 (E. E. Green) (lectotype, paralectotypes); Surrey,Hog's Back, on Brachypodium sylvaticum, 21.vii.1922 (E. E. Green); Isle of Wight, Seaview, on B.sylvaticum, 3.ix.l921, Alum Bay, on grass, vii.1927 (E. E. Green); Berkshire, onAchillea millefolium,Veronica chamaedrys, Alopecurus pratensis, Agrostis sp., Holcus mollis, Deschampsia caespitosa, 1948(various dates) (K. Boratynski); Yorkshire, Ickornshaw Moor, x.1958 (A. Smith), Steeton Moor,29.viii.l948(A.Smith). DISCUSSION. Green's original slide contains three specimens, one of which is selected as lectotype andclearly marked, the other two are labelled paralectotypes. It is sometimes difficult to distinguish this species from R. insignis. At present the main differences arethe 3 marginal setae on the 7th abdominal segment in R. pseudinsignis compared with 4 in R. insignis. The THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 379 Fig. 14 Eriococcus pseudinsignis Green. England, Yorkshire, Ickornshaw Moor, on grass. 380 D. J. WILLIAMS dorsal setae on the head and thorax of/?, pseudinsignis are quite robust and are much longer than the setaeon the posterior abdominal setae, whereas all the dorsal setae in R. insignis are short and usually slender,although those on the head and thorax are often wider than the abdominal setae. At present both species are here regarded as distinct but intermediates may be found to warrant furtherresearch. Eriococcus spur/us (Modeer) (Fig. 15) Coccus ulmi Linnaeus, 1758: 455 (in part). [Coccus ulmi Linnaeus sensu Linnaeus, 1766: 740. Misidentification.]Coccus spurius Modeer, 1778: 43. Syntypes 9, EUROPE (probably lost).[Gossyparia ulmi (Linnaeus) sensu Linnaeus, 1766; Signoret, 1875: 21; Schmutterer, 1952: 416. Misiden- tifications.]Gossyparia spuria (Modeer) Cockerell, 1899ft: 268; Fernald, 1903: 68; Borchsenius, 1949: 330; Danzig, 1964: 632; Tereznikova, 1981: 50.Nidularia spuria (Modeer) Lindinger, 1933: 108.Eriococcus spurius (Modeer) Ferris, 1955: 164. DESCRIPTION. Adult female, when prepared on slides, broadly oval, strongly nodulose, 3-0 mm long and 1-8mm wide. Anal lobes well developed, protruding, about twice as long as wide, rounded at apex,moderately sclerotised, strongly nodulose with sclerotised teeth, especially on inner margins. Each lobewith an apical seta 250 /-im long, dorsally with 2 inner and 1 outer enlarged setae, ventrally with 2 slendersetae and a pointed suranal seta. Antennae 230-300 u,m long, with 7 segments. Frontal lobes present, justanterior to basal segment. Legs normal, well developed. Hind trochanter + femur 160-190 ^im long, hindtibia 100-110 /-im long, hind tarsus 130-150 //,m long, the tibia + tarsus always longer than trochanter +femur, claw 40 /*,m long, curved, with a denticle near apex. Hind coxa without translucent pores but withnumerous spicules. Labium 150 /xm long, shorter than clypeolabral shield, basal segment with 2 pairs ofsetae. Dorsal plate well developed, triangular, lightly sclerotised and strongly nodulose. Anal ring with 8setae, each about 150 fim long. Dorsal surface with numerous setae, in a regular distribution except on the intersegmental areas ofthorax where they are absent, but present on mid-line. Setae long and slender with almost straight sides,tapering gradually to a blunt point, the largest 60 /x,m long around posterior margins, the smallest about 25/Am long. Macrotubular ducts, each about 20 pm long, tapering slightly to orifice, cup wider than a setalbase, present around margins and submargins, rarely extending to submedian areas, absent entirely onmedian area. Microtubular ducts slender, about 10 /x,m long, with the orifice widely bifid, in an evendistribution over entire dorsum. Ventral surface with normal slender setae in median areas, a few enlarged setae the same as on dorsum,around margins, and a few short setae that are slender and stiff on submargins. Macroducts of two mainsizes. A large duct, same size as on dorsum, on margins, and a narrower duct in transverse rows onabdomen and in submedian areas of thorax and head. Microtubular ducts absent. Quinquelocular discpores present across median areas of abdominal segments, on metathorax and near spiracles and in an areabetween antennae and clypeolabral shield. Cruciform pores in submedian zone between first spiracles andanterior abdominal segments. MATERIAL EXAMINED England: Surrey, Camberley, on Ulmus angustifolia, iv-vii.1916, 18. iv. 1922 (E. E. Green); Farnham, onU. angustifolia, x.1915 (C. J. F. Fryer); Woking, on Ulmus sp., vi.1918 (G. C. Gough). France: (ex coll. P.Marchal). Germany: Hanover, on U. procera (U. campestris)^vii.l893>. Austria: 'Donau-Auen', on U.procera. Czechoslovakia: 'Bohemia', on Ulmus sp. (ex coll. K. Sulc). Turkey: Istanbul, on U. 'campestrisvar. pyramidalis.' In addition numerous specimens have been examined from U.S.A. DISCUSSION. Linnaeus (1758) cited Reaumur (1738) but unfortunately referred to two distinct species thatReaumur had figured and described. The first refers to pi. 5, figs 5, 6 and this is undoubtedly an armouredscale insect discussed on p. 78 as found on a branch of elm. This species is accepted at present asLepidosaphes ulmi (L.). The second species referred to by Linnaeus is illustrated on pi. 7, figs 1-10 anddiscussed by Reaumur on p. 119 as found on a branch of elm. There is no doubt from Reaumur'sillustrations that this is the species under discussion here. Linnaeus (1766) realising his mistake in 1758,listed C. ulmi but referred to Linnaeus (1761) where only the name is mentioned and to Geoffrey (1762) (awork not consistently binominal but nevertheless vital) who, apart from giving a short description, refers to THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 381 Fig. 15 Eriococcus spurius (Modeer). England, Surrey, Camberley, on f//mi angustifolia. 382 D. J. WILLIAMS the second of Reaumur's species mentioned by Linnaeus (1758). Clearly the listing by Linnaeus (1766) isbased on a misidentification of the species listed by Linnaeus (1758). All subsequent references to 'ulmi'the eriococcid, from Fabricius to the present day must refer to 'sensu Linnaeus, 1766'. Modeer (1778)described Coccus spurius and referred to the excellent illustrations of Reaumur (1738: pi. 7, figs 1-10).Modeer's name is accepted by all modern workers. Records from England, listed above, were discussed by Green (1916, 1917) when there were outbreaksin Surrey. The insect has not been found since and it is not clear if it has become established but its externalappearance is so distinctive that it should be easily recognised. Since it was introduced to the U.S.A.,probably in the last century, it has caused concern by its frequent large aggregations on elms. The speciesseems to be confined to the genus Ulmus. Macropterous and brachypterous males have been recorded and both forms have been described by Afifi(1968). Eriococcus f/yim (Schrank) comb. rev.(Fig. 16) Coccus thymi Schrank, 1801: 146. Syntypes $ , GERMANY: Kehlheim (probably lost). Eriococcus thymi (Schrank) Signoret, 1875: 32; Fernald, 1903: 79. Rhizococcus gnidii Signoret, 1875: 37; Fernald, 1903: 66. Syntypes $, FRANCE: Estrelle Mts, nr Cannes (probably lost). Syn. n. Nidularia gnidii (Signoret) Lindinger, 1933: 116.Nidularia thymi (Schrank) Lindinger, 1957: 548. DESCRIPTION. Body of adult female broadly oval, nodulose, largest specimens 1-9 mm long, 1-35 mm wide.Anal lobes well developed, conical, about twice as long as wide at base, heavily sclerotised on venter,lightly sclerotised on dorsum. Each lobe with apical seta 280 /u,m long, dorsal surface with 3 enlarged setae,ventral surface with 2 slender seta and a suranal seta much shorter than anal ring setae. Dorsal platemembranous, rounded, narrow, margin irregular, nodulose. Anal ring with 8 setae each 120 /um long.Antennae 170-250 /urn long, with 6 or 7 segments, the 3rd segment often divided. Frontal tubercle present,seta-like, just anterior to each basal antennal segment. Legs well developed. Hind trochanter + femur140-190 /u,m long, hind tibia 70-100 /urn long, hind tarsus 110-130 /u,m long, claw 30 /urn long, stout, curvedwith a denticle near apex. Ratio of lengths of hind tibia -I- tarsus to hind trochanter + femur 1-18-1-30.Hind coxa with a few translucent pores. Labium 130-160 /um long, slightly shorter than clypeolabral shield,basal segment with 2 pairs of setae. Spiracles heavily sclerotised, the sclerotisation continuing around theatrium in varying degrees. Dorsal surface with enlarged setae of different sizes. The largest around margins 28-55 /u,m long, eachseta blunt, with sides slightly concave but curved in profile, these setae tending to be longest on posteriorabdominal segments. Remaining dorsal setae shorter, 12-28 /urn long on the head, thorax and anteriorabdominal segments, sometimes approaching in length the shorter marginal setae, becoming short, 8 /u,mlong on the 6th and 7th segments, each seta blunt and curved in profile. Macroducts of one size, each about25 /am long, the cup narrower than a setal base of an enlarged seta. Usually a single pair of smaller ducts onanal lobe segment. Microducts in moderate numbers, each about 8 /am long, with inner sclerotisedampulla, sclerotised tube and collar and with bifid orifice. Ventral surface with normal flagellate setae in median areas. Enlarged setae, similar to dorsal setae onmargins and submargins, varying in size but not as long as dorsal marginal setae. Slender but stiff setae alsopresent in a submarginal zone. Macroducts of three sizes. A large type, similar to dorsal ducts, presentaround margins only. An intermediate type, only slightly narrower than the large type, in transverse rowson abdominal segments and in median areas of thorax and head. A small type sparse, present singly insubmedian areas of most abdominal segments. Microducts few, in marginal and submarginal areas.Quinquelocular disc pores in transverse bands on abdominal segments but not reaching margins, in medianareas of thorax and head and around spiracles. Cruciform pores few, in submarginal areas of anteriorabdominal segments and thorax, and in median areas of thorax and head. MATERIAL EXAMINED France: Montpellier, on Thymus vulgaris (Lamiaceae) (det. as E. thymi by V. Signoret), Cannes, onDaphne gnidium (Thymelaeaceae) (det. as R. gnidii by V. Signoret), Var, Agay, on D. gnidium, 4.V.1908(P. Marchal), Bouc-Bel-Air, on T. vulgaris, v.1932 (L. Goux). DISCUSSION. Although there is no original material available, there seems to be no doubt that thespecimens examined and identified by Signoret and others identified by Marchal and Goux represent thespecies described by Schrank on Thymus as C. thymi and on Daphne gnidium as R. gnidii. The characters THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 383 Fig. 16 Eriococcus thymi (Schrank). France, Montpellier, on Thymus vulgaris. 384 D. J. WILLIAMS on some specimens of D. gnidium tend to be larger but the specimens themselves are larger and allrepresent the same species. In well-stained specimens the microducts clearly show the bifid orifice but thischaracter is difficult to see in some specimens. The illustration is based on specimens from France onThymus vulgaris, identified as E. thy mi by Signoret. Further collecting may show a much wider host-plantrange. A record by Hardy (1876) from Watch Law in the north of Northumberland, England, on the roots ofThymus serpyllum (probably T. drucei), has never been substantiated. KUWANINA Cockerell Kuwanina Cockerell, in Fernald, 1903: 121. Type-species: Sphaerococcus parvus Maskell, by originaldesignation. Cockerell intimated that this genus differed from Antonina by the larva having 3 or 4 jointed antennae.Antonina belongs to the family Pseudococcidae but Kuwanina clearly belongs to the family Eriococcidaewhere it was accepted by Hoy (1963). Another species, K. hilli Laing, was shown by Williams (1985a) to beidentical with Chaetococcus australis Froggatt, a mealybug in the family Pseudococcidae. The type-species is in need of critical study, outside the scope of this work because the species is notnative to Britain, having been found only once on imported cherry trees from Japan. The genus, as represented by specimens at hand from Japan and Britain, seems to be related toCryptococcus in lacking legs and having the third pair replaced by pore-bearing leg-flaps. It differs,however, in lacking macroducts. These are replaced by quite large invaginated quinquelocular pores on thedorsum and ventral margins. In addition there are smaller quinquelocular pores of two distinct sizes on theventer. Microducts are present. Kosztarab (19686) placed Kuwanina in the Cryptococcidae and Koteja (1974ft), although agreeing thatthe genus was closely related to Cryptococcus, believed that the 'Kuwanina group' may form a distinctfamily. The genus is here regarded as a normal component of the Eriococcidae. Kuwanina parva (Maskell) Sphaerococcus parvus Maskell, 1897: 244. Syn types $, JAPAN: on cherry (NZAC).Kuwanina parva (Maskell) Cockerell, in Fernald, 1903: 121; Hoy, 1963: 165. MATERIAL EXAMINEDEngland: Hertfordshire, St Albans, on dwarf Prunus sp. , imported from Japan, 1914 (C. J. F. Fryer). DISCUSSION. Green stated that examples were found on gnarled branches in a nursery garden where theywere able to thrive and exist. It is doubtful if the species now exists in Britain. NOTEOCOCCUS Hoy Noteococcus Hoy, 1962: 164. Type-species: Eriococcus hoheriae Maskell, by original designation. Hoy erected this genus to accommodate E. hoheriae from New Zealand, a species with large sclerotisedanal lobes, each with 12 small enlarged setae on the dorsum in addition to the normal pair on the innermargins, and with an irregular series of teeth on the venter. Furthermore, apart from the strongly noduloseposterior abdominal segments, there is a large rugose plate anterior to the anal lobes on the dorsum. Thelegs are much reduced with the trochanter + femur and the tibia + tarsus fused. Noteococcus hoheriae (Maskell) Eriococcus hoheriae Maskell, 1880: 298. Lectotype $, NEW ZEALAND: Lyttelton (NZAC), designated by Deitz & Tocker (1980: 47).Noteococcus hoheriae (Maskell) Hoy, 1962: 164. Specimens are at hand recorded by Green (1925a) from Cornwall, Isles of Stilly, Tresco, on Hoheriapopulnea, iii. 1924 (/. C. F. Fryer), living in the open. Williams (1985ft), reporting on scale insects collectedat Tresco, stated that the species has not been found since, but in common with other exotic species thathave been introduced to Tresco, on plant material, the insect may still occur there. OVATICOCCUS Kloet Gymnococcus Douglas, 1888: 150. Type-species: Coccus agavium Douglas, by original designation (see below). [Homonym of Gymnococcus Zopf, 1887: 126.]Ovaticoccus Kloet, 1944: 86. [Replacement name for Gymnococcus Douglas.] THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 385 This genus has been revised by Miller & McKenzie (1967) who included nine North American species, butat least two others described from elsewhere have been listed by Boratynski (1958) and accepted by Hoy(1963). In a description of the genus, Miller & McKenzie (1967) stated that the anal lobes are absent andthat the anal ring is without pores, usually in the form of a modified circle with usually 3 pairs of setae.Enlarged dorsal setae are present in some species, as are macroducts, microducts and cruciform pores. The type-species has been redescribed and illustrated by Boratynski (1958) and by Miller & McKenzie(1967). An examination of numerous specimens shows that well-developed frontal tubercles are present.The genus is closely related to Pseudochermes but differs in possessing an anal ring without pores and inpossessing cruciform pores. Although O. agavium is not a native British species, it is included here because it was described from theRoyal Botanic Gardens, Kew (under glass) in 1888 on Agave, and Boratynski (1958) has indicated that hehad no trouble finding it again in 1956 and 1957. This species is easily separated from Pseudochermes fraxiniby the presence of numerous enlarged setae in addition to slender lanceolate setae on the dorsum. Douglas (1888) proposed Gymnococcus conditionally when describing C. agavium and according to theInternational Code of Zoological Nomenclature the name is available from that date. Ovaticoccus agavium (Douglas) Coccus agavium Douglas, 1888: 150. Syntypes 9, ENGLAND: Kew (probably lost, but see below).Gymnococcus agavium (Douglas) Cockerell, 1893: 1049; Newstead, 1903: 204; Fernald, 1903: 79; Borchsenius, 1949: 370; Tereznikova, 1981: 57.Ripersia agavium (Douglas) Newstead, 1897: 12.Ovaticoccus agavium (Douglas) Kloet, 1944: 86; Boratynski, 1958: 175; Danzig, 1964: 634; Miller & McKenzie, 1967: 509. MATERIAL EXAMINED England: Surrey, Royal Botanic Gardens (under glass), ii.1888, on leaves of Agave sp. (originalmaterial), 1956, 1957 (various dates) on A. decipiens, A. francescini, A. parryi (K. L. Boratynski), on A.francescini, 26. v. 1964 (5. A. Afifi). DISCUSSION. Boratynski (1958) has indicated that he designated a lectotype from a slide containingtype-material and that this is in the British Museum (Natural History). This slide is not present in thecollections and it is doubtful if Boratynski could have done so. There are numerous unmounted specimensfrom Douglas' collection on cards numbered 1052 and 1056. These numbers agree with the information inhis collection data books and three of these specimens were prepared on a slide by F. Laing. Douglas alsosent specimens to Newstead who labelled four slides Type lot ex coll. J. W. Douglas'. None of these slidesrepresents the original specimens used by Douglas in his description, although they are authentic enough.There is no doubt about the identity of the species.Afifi (1968) has described and illustrated the alate male. PSEUDOCHERMES Nitsche Pseudochermes Nitsche, in Judeich & Nitsche, 1895: 1249. Type-species: Chermes ? fraxini Kaltenbach, by original designation.Apterococcus Newstead, 1898: 97 [as subgenus of Ripersia Signoret]. Type-species: Eriococcus fraxini Newstead, by original designation [cited as Ripersia fraxini]. [Synonymised by Cockerell, 18996: 264 (see below).] This genus is now widely recognised, even though for many years most workers accepted the synonymy, byCockerell (18996), with Fonscolombia Lichtenstein. Cockerell synonymised both Pseudochermes andApterococcus with Fonscolombia but Fonscolombia is now regarded as an unrecognisable genus in thePseudococcidae, even so the synonymy of Apterococcus with Pseudochermes still stands. Pseudochermeshas remained monotypic. DESCRIPTION. Adult female oval to almost circular, posterior end rounded without recognisable anal lobesbut with normal apical setae. Anal ring crescentic with 6 setae and pores. Suranal setae normal. Antennae6-segmented. Frontal tubercles present. Legs normal, claw with denticle. Labium short, basal segmentwith a single pair of setae. Dorsal setae short, narrow but spine-like. Macroducts present, each with innerend cupped, on dorsum and venter. Microducts slender, elongate, with simple orifice. Disc poresquinquelocular, on dorsum and venter. Cruciform pores absent. First and second instars with normal legs. Antennae 6-segmented, basal segment with a pair of setae.Frontal tubercles present. Anal ring as in adult female but smaller. 386 D. J. WILLIAMS DISCUSSION. The genus has affinities with Ovaticoccus but differs in possessing an anal ring containing acrescentic plate with 6 setae and pores. In Ovaticoccus the anal ring is always without pores. Furthermorecruciform pores are present in Ovaticoccus, usually in large numbers, but are absent completely inPseudochermes . Koteja (1974a) allied this genus to Cryptococcus and included both in the family Cryptococcidae. Thereseems to be no reason for accepting this action. Pseudochermes fraxini (Kaltenbach)(Figs 17, 18) Chermes (?) fraxini Kaltenbach, 1860: 259; 1974: 433. Syntypes <j>, GERMANY (probably lost). Eriococcus fraxini Newstead, 1891: 165. LECTOTYPE $, ENGLAND (BMNH), here designated [ex-amined]. [Synonymised by Judeich & Nitsche, 1895: 1249.] Ripersia fraxini (Newstead) Newstead, 1892: 147. Coccus fraxini (Kaltenbach) Judeich & Nitsche, 1895: 1247. Pseudochermes fraxini (Kaltenbach) Judeich & Nitsche, 1895: 1249; Borchsenius, 1949: 365; Danzig, 1964:634; Tereznikova, 1981: 59. Fonscolombia fraxini (Kaltenbach) Cockerell, 1899ft: 264; Fernald, 1903: 114; Schmutterer, 1952: 418. Ripersia (Apterococcus) fraxini (Newstead) Newstead, 1898: 97. Apterococcus fraxini (Newstead) Newstead, 1903: 210. DESCRIPTION. Young adult female (Fig. 17) on microscope slide oval, 1-15 mm long, 0-9 mm wide, butmature adult female sometimes almost circular, membranous, posterior end rounded, anal lobes notdeveloped but position of each lobe with an apical seta 60 /urn long. Anal ring in form of crescenticsclerotised plate 30 pm wide, with a few pores and 6 setae, each about 30 ju,m long, blunt and stiff, situatedon ventral surface even in young adult females. Suranal setae present just anterior to anal ring, shorter thananal ring setae. Antennae 100-110 /im long, with 6 segments. Frontal tubercles present just anterior tobasal antennal segment. Legs small but well developed. Hind trochanter + femur 70-75 /itn long, hind tibia3540 /u,m long, hind tarsus 30-35 /-im long, claw 16 /n,m long, curved, with a denticle near apex. Ratio oflengths of hind tibia + tarsus to hind trochanter + femur 0-90-1-05. Hind coxa noticeably larger thananterior coxae, with a few large translucent pores. Labium 50-60 /-im long, much smaller than clypeolabralshield; basal segment with a single pair of minute setae. Dorsal surface with narrow conical setae on 7th and anal lobe segments, each about 20 /xm long andusually shorter than anal ring setae. Anteriorly the setae are shorter and more slender, but stiff, 8 /AHI long.Macroducts fairly numerous, in an even distribution, each about 20 /im long, tapering to narrow orifice, thecup wider than a setal base of largest setae. Microducts present, not numerous, tending to be in rows acrossthe segments, each duct filiform, about 4 /im long with a small sclerotised ampulla. Quinquelocular discpores present in more or less single rows at posterior edges of segments. Ventral surface with slender setae only slightly longer than dorsal setae, except on head where they aremuch longer. Macroducts of two sizes. A larger type, same as on dorsum, around margins only except onthorax near spiracles where they also occupy the submarginal areas. A narrower duct each with sclerotisedcup and filiform tube, tapering to a minute orifice, sparse, on abdominal segments only. Microducts, sameas on dorsum, few, on margins. Quinquelocular disc pores not numerous on head and thorax but morenumerous on abdomen. Second instar female (Fig. 18B), oval, similar in shape to young adult female but smaller. Anal lobesonly slightly produced, each with an apical seta 40 fjun long, and suranal setae minute. Anal ring 20 /^mwide, crescentic, as in adult female situated at apex on venter, with 6 setae. Antennae 6-segmented, 75 /Amlong. Frontal tubercles present. Legs similar in shape to those of adult female but all coxae approximatelysame size. Hind trochanter + femur 48 /nm long, hind tibia 25 /mi long, hind tarsus 25 /urn long, claw 12 /AHIlong. Labium 45 /xm long, basal segment with a single pair of setae. Dorsal surface with thick truncatesetae, 12 /4m long, on posterior segments, these longer than anal ring setae. Anteriorly the setae becomeshorter and more slender, but remain stiff. Macroducts in more or less single rows across the segments,each duct similar in shape to those of adult female but the cup narrower. Microducts as in adult, notnumerous. Quinquelocular pores smaller than in adult, in single rows at posterior edges of segments. Ventral surface with slender setae, usually shorter than on dorsum except for a few long setae on head.Macroducts of one size as on dorsum, on head and thoracic margins and extending almost to spiracles.Microducts few on margins only. Quinquelocular disc pores not numerous, in single rows on abdominalsegments and present on head and thorax but absent in median area of thorax. First instar (Fig. 18A) oval, 300 /urn long, 180 /tm wide, abdomen tapering. Anal lobes poorly developed,each with an apical seta 35 /im long. Anal ring as in second instar and adult female but 16 /xm wide, with 6 THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 387 Fig. 17 Pseudochermesfraxini (Kaltenbach). Berkshire, Cookham, on Fraxinus excelsior. 388 D. J. WILLIAMS V i- \1- t V B D.J.W. Fig. 18 Pseudochermes fraxini (Kaltenbach). (A) First instar. (B) Second instar female. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 389 setae. Antennae 65 ^tm long, with 6 segments. Frontal tubercles present. Legs normal, hind trochanter +femur 35 /xm long, hind tibia 35 /u,m long, hind tarsus 20 /ion long. Claw curved, 10 /urn long, with a minute[denticle near apex. Labium 30 /urn long, basal segment with a pair of setae. Dorsal surface with shorttruncate setae, tending to be cylindrical, tapering only slightly, those on abdomen the largest and thickest,about 8 fjLm long, becoming shorter anteriorly to head where they are 5 fjun long, all setae thicker than anal[ring setae. Microducts sparse, as in later instars. Ventral surface with slender setae, shorter than dorsalsetae except for a few long setae on head. Ventral ducts absent. Quinquelocular pores smaller than in laterinstars, represented usually by a pair on each of 4th to 7th segments and singly next to openings of spiracles. MATERIAL EXAMINED (all on Fraxinus excelsior unless stated otherwise) England: Berkshire, Silwood Park, F. excelsior and Populus tremula (Salicaceae), 29. vi. 1966 (K. L.\Boratynski), Cookham, 5.viii.l972, 8.ix.l972, ll.iii.1973, ll.viii.1983 (D. J. Williams); Cheshire, Ince,[viii.1890 (R. Newstead) (lectotype, paralectotypes), Chester, 1895, ll.xi.1895 (R. Newstead); Devon,Sidmouth, 26.vi.1963; Gloucestershire, Cheltenham, viii.1973 (D. J. Williams); Hampshire, Bentley,Il8.iv.1969 (C. I. Carter), Exbury, viii.1974; Kent, Bearsted, vii.1913 (E. E. Green); Leicestershire,[Loughborough, 8.xi.l938 (A. Roebuck); London, Buckingham Palace Gardens, 11.x. 1963 (K. L. Bora-tynski); Northumberland, Riding Mill, S.x.1950 (D. J. Williams); Oxfordshire, Goring, 25.ii.1914 (E. E.Green); Somerset, Bath, Ashrey Gullywood, 24.ii.1978; Surrey, Frimley, 23.V.1921 (E. E. Green),Wisley, on Syringa sp., 1958, 6.vii.l983 (D. J. Williams), Woking, 1914, vii.1915 (E. E. Green); Yorkshire,Adel, 2.vii.l963 (D. J. Williams). Wales (N.): 1976. Belgium: intercepted in England, 31.U967 (5. H.Blore). Hungary: Galgamocsa, 26.iv.1981 (Nagy). DISCUSSION. Newstead's single original slide of Eriococcus fraxini contains nine whole specimens and a few| pieces. The lectotype selected is clearly marked and the remaining eight whole specimens are labelledparalectotypes. Afifi (1968) has described and illustrated the adult male of this species. It is always minute, wingless anddegenerate. Apart from incursions into Iran, this species is European where its favourite host-plant is Fraxinus\excelsior, but it is known on other species of Fraxinus. It is often found in enormous numbers in the crevicesof the bark. The record on Syringa is not surprising because this genus and Fraxinus both belong to theI family Oleaceae. Furthermore, Syringa has been mentioned from Germany by Lindinger (1938) who also(listed Sorbus sp. S. aucuparia (as Pirus aucuparia) was previously listed by Lindinger (1912). The record onPopulus tremula is based on specimens collected at the same time as others on F. excelsior. There seems tobe every reason that the insects were feeding on Populus and that P. fraxini may yet be found on other(host-plants. The species is probably widespread in Britain on ash. Harrison (1944) recorded it from Inverarish in the(Isle of Raasay, Scotland. References Afifi, S. A. 1968. Morphology and taxonomy of the adult males of the families Pseudococcidae andEriococcidae (Homoptera: Coccoidea). Bulletin of the British Museum (Natural History) (Entomology)supplement 13: 1-120. Baerensprung, F. von 1849. Beobachtungen uber einige einheimische Arten aus der Familie der Coccinen.III. Coccus. ZeitungfiirZoologie, Zootomie und Palaeozoologie 1: 173-176. Balachowsky, A. 1937. Les cochenilles de Seine-eHDise (1) (Contribution a 1'etude des Coccides de France23me note). Bulletin de la Societe des Sciences Naturelles et Medicales de Seine-et-Oise (3) 5: 1-6.- 1954. Sur une nouvelle espece d'Eriococcini de la foret de Fontainebleau avec creation d'un genrenouveau. Bulletin de la Societe entomologique de France 59: 61-64. Ben-Dov, Y. 1979. A taxonomic study of the soft-scale genus Kilifia (Coccidae). Systematic Entomology 4:311-324. I Ben-Dov, Y., Williams, M. L. & Ray, C. H. 1975. Taxonomy of the Mango Shield Scale Protopulvinariamangiferae (Green) (Homoptera: Coccidae). Israel Journal of Entomology 10: 1-17. Boratynski, K. L. 1958. A note on Ovaticoccus agavium (Douglas) (Homoptera: Coccoidea: Eriococci-dae) and on the genus Ovaticoccus generally. Proceedings of the Royal Entomological Society of London(B) 27: 173-182. 1962. A new species of Acanthococcus Signoret, 1875 (Eriococcidae, Coccoidea, Homoptera) from Britain. Proceedings of the Royal Entomological Society of London (B) 31: 55-60.Boratynski, K. L. & Williams, D. J. 1964. Coccoidea, pp. 87-94. In Kloet, G. S. & Hincks, W. D., A 390 D. J. WILLIAMS check-list of British insects. Second Edition (Revised). Part 1: Small Orders and Hemiptera. Handbook for the Identification of British Insects 11(1): xv, 119 pp.Borchsenius, N. S. 1948. On the revision of the genus Eriococcus Sign. (Insecta, Homoptera, Coccoidea) Doklady Akademii Nauk SSSR 60: 501-503.1949. Sucking Insects Vol. 7. Suborder mealybugs and scale insects (Coccoidea), Family mealybugs (Pseudococcidae). Fauna SSSR. Novaya Seriya 38. 383 pp. 1960. Contribution to the Coccid fauna of China 4. Hard and soft scales, harmful to fruit and grape culture in northeast and east China. Acta Entomologica Sinica 10: 214-218.Brues, C. T. & Melander, A. L. 1932. Classification of insects. A key to the known families of insects am other terrestrial arthropods. Bulletin of the Museum of Comparative Zoology 73: 672 pp.Cockerell, T. D. A. 1893. Notes on the cochineal insect. The American Naturalist 27: 1041-1049.1896. A check-list of the Coccidae. Bulletin of the Illinois State Laboratory of Natural History 4: 318-339. 1897. Contributions to coccidology. - II. The American Naturalist 31: 588-592. 1899o. First supplement to the check-list of the Coccidae. Bulletin of the Illinois State Laboratory of Natural History 5: 389-398. 18996. Some notes on Coccidae. Proceedings of the Academy of Natural Sciences of Philadelphia 1899: 259-275.Commonwealth Institute of Entomology 1979. Cryptococcus fagisuga Lndgr. Distribution Maps of Insect Pests no. 393.Covassi, M. 1975. Nuovi reperti sulla corologia del Cryptococcus fagisuga Lndgr. in Italia e in Corsica (Homoptera, Crypt ococcidae). Redia 56: 555-564.Danzig, E. M. 1962. Revision of the genus Rhizococcus Signoret (Homoptera, Coccoidea) of the fauna of USSR. Entomologicheskoe Obozrenie41: 839-860.1964. Suborder Coccinea - Coccids or mealybugs and scale insects, pp. 616-654. In Bei-Bienko, G. Ya. (Ed.), Keys to the insects of the European part of USSR Vol. 1. Apterygota, Palaeoptera, Hemimetabola. Opredeliteli po Faune SSSR 84: 935 pp.1975. New species of mealybugs of the genus Acanthococcus Sign. (Homoptera, Coccoidea, Eriococcidae) from the Far East. Entomologicheskoe Obozrenie 54: 62-81. 1980. Coccids of the Far East of the USSR (Homoptera, Coccinea) with an analysis of the phylogeny of coccids of the world fauna. Opredeliteli po Faune SSSR 124: 367 pp.Deitz, L. L. & Tocker, M. F. 1980. W. M. Maskell's Homoptera: species-group names and type-material, i Information Series. Department of Scientific and Industrial Research. New Zealand no. 146: 76pp.Douglas, J. W. 1886. Note on some British Coccidae (No. 5). The Entomologist's Monthly Magazine 23: 150-155.1888. Notes on some British & exotic Coccidae (No. 12). The Entomologist's Monthly Magazine 25: 150-153. 1890. Notes on some British & exotic Coccidae (No. 16). The Entomologist's Monthly Magazine 26: 153-155.Dziedicka, A. & Koteja, J. 1971. A revision of the species of the genus Rhizococcus Signoret (Homoptera, Coccoidea) occurring in Poland. Acta Zoologica Cracoviensia 16: 557-579.Ehrlich, J. 1934. The Beech Bark Disease. A Nectria disease of Fagus, following Cryptococcus fagi (Baer.). Canadian Journal of Research 10: 593-692.Elliott, E. A. 1933. The Hemiptera of Suffolk. First Supplement. Transactions of the Suffolk Naturalists' Society 2: 134-156.Fernald, M. E. 1903. A catalogue of the Coccidae of the world. Bulletin of the Hatch Agricultural Experiment Station 88: 360 pp.Ferris, G. F. 1955. Atlas of the scale insects of North America. The families Aclerdidae, Asterolecaniidae, Conchaspididae, Dactylopiidae and Lacciferidae. 7: 1-233 pp. Stanford, California.1957a. Notes on some little known genera of the Coccoidea (Homoptera). Microentomology 22: 59-79. 1957 'b. A review of the family Eriococcidae (Insecta: Coccoidea). Microentomology 22: 81-89. Fonscolombe, L. J. H. Boyer de 1834. Description des kermes qu'on trouve aux environs d' Aix. Annales de la Societ^ Entomologique de France 3: 201-218.Geoffroy, E. L. 1762. Histoire abregee des insectes qui se trouvent aux environs de Paris; dans laquelle ces animaux son ranges suivant un ordre methodique. 1: xxvii, 523 pp. Paris.Goux, L. 1934. Notes sur les Coccides [Hem.] de la France (8e note). Description d'une Ripersia nouvelle et remarques sur quelques autres especes. Bulletin de la Societe Entomologique de France 39: 27-31.1948. Les caracteres specifiques chez les Eriococcus et la signification des microfilieres tubulaires. THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 391 Notes sur les Coccides (Hem. Coccoidea) de la France (39me note). Bulletin de la Societe Linneenne deProvence 16: 66-69. Green, E. E. 1896. A new British Coccid. The Entomologist's Record and Journal of Variation 8: 260-261. - 1915. Observations on British Coccidae in 1914, with descriptions of new species. The Entomologist'sMonthly Magazine 51: 175-185. 1916. On two new British Coccidae, with notes on some other British species. The Entomologist'sMonthly Magazine 52: 23-31. 1917. Observations on British Coccidae; with descriptions of new species. No. IV. The Entomologist'sMonthly Magazine 53: 260-264. 1921. Observations on British Coccidae, with descriptions of new species. The Entomologist'sMonthly Magazine 57: 146-152. 1923. Observations on British Coccidae, VIII. The Entomologist's Monthly Magazine 59: 211-218. 1925a. Observations on British Coccidae, IX. The Entomologist's Monthly Magazine 61: 34-44. 1925b. Notes on the Coccidae of Guernsey (Channel Islands), with descriptions of some new species.The Annals and Magazine of Natural History (9) 16: 516-527. 1926. Observations on British Coccidae. X. The Entomologist's Monthly Magazine 62: 172-183. 1928. Observations on British Coccidae. XI. With descriptions of new species. The Entomologist'sMonthly Magazine 64: 20-31. 1931. Observations on British Coccidae. XIII. The Entomologist's Monthly Magazine 67: 99-106. Gullan, P. J. 1984. A revision of the gall-forming Coccoid genus Apiomorpha Rubsaamen (Homoptera: Eriococcidae: Apiomorphinae). Australian Journal of Zoology (Supplementary Series) 97: 203 pp.Hardy, J. 1876. Contributions to the entomology of the Cheviot Hills. No. V. History of the Berwickshire Naturalists' Club (1873-1875), pp. 328-334. Harrison, G. H. 1944. Coccidae from the Western Isles of Scotland. The Entomologist 77: 110-111.Harrison, J. W. H. 1916a. Coccidae and Aleyrodidae in Northumberland, Durham, and north-eastYorkshire. The Entomologist 49: 172-174. 19166. Notes and records. Coccidae. Scale Insects. The Vasculum 2: 93. 1918. Additional localities for Eriococcus devoniensis Green. The Entomologist's Monthly Magazine54: 17. 1948. Zoocecidia in the Scottish Western Isles. The Entomologist's Monthly Magazine 84: 113-116.1949. The scale insect, Eriococcus devoniensis Green, at Ross Links. The Vasculum 34: 24. Hoy, J. M. 1962. Eriococcidae (Homoptera: Coccoidea) of New Zealand. Bulletin, New Zealand Department of Scientific and Industrial Research 146: 219 pp.1963. A catalogue of the Eriococcidae (Homoptera: Coccoidea) of the world. Bulletin. New Zealand Department of Scientific and Industrial Research 150: 260 pp. Hussey, N. W. 1956. Felted Beech Coccus. Leaflet of the Forestry Commission no. 15: 8 pp.Judeich, J. F. & Nitsche, H. 1895. Lehrbuch der Mitteleuropaischen Forstinsektenkunde Part IV. pp. 937-1421. Vienna.Kaltenbach, I. H. 1860. Die deutschen Phytophagen aus der Klasse der Insekten. Verhandlungen des naturhistorischen Vereines der preussischen Rheinlande und Westphalens 17: 203-260.1874. Die Pflanzenfeinde aus der Klasse der Insekten. 848 pp. Stuttgart. Killington, F. P. 1936. Further records of Coccidae (Hemipt.) from Hants. Journal of the Society for British Entomology 1: 119.Kloet, G. S. 1944. A new generic name in the Coccidae (Hemiptera). The Entomologist's Monthly Magazine 80: 86.Kosztarab, M. 1968a. A new Cryptococcus species from North America, with a key to the species of the genus (Homoptera: Coccoidea). The Virginia Journal of Science 19: 7-11. 19686. Cryptococcidae, a new family of the Coccoidea (Homoptera). The Virginia Journal of Science 19: 12. Koteja, J. 1974a. Comparative studies on the labium in the Coccinea (Homoptera). Zeszyty NaukoweAkademii Rolniczej w Krakowie 27: 1-162, 78 figs. j 19746. On the phylogeny and classification of the scale insects (Homoptera, Coccinea) (discussion based on the morphology of the mouthparts). Acta Zoologica Cracoviensia 19: 267-325. 1983. An additional moult in adult female Acanthococcus greeni (Newstead), (Homoptera, Coc- cinea). ActaBiologica Cracoviensia (Zoologia) 25: 59-62.Koteja, J. & Zak-Ogaza, B. 1981. Kaweckia gen. n. in the Eriococcidae (Homoptera, Coccoidea) and notes on related genera. Acta Zoologica Cracoviensia 25: 501-517.Kuwana, S. I. 1907. Coccidae of Japan, I. A synoptical list of Coccidae of Japan with descriptions of 392 D. J. WILLIAMS thirteen new species. The Bulletin of the Imperial Central Agricultural Experiment Station Japan II 177-207.Lindinger, L. 1912. Die Schildlause (Coccidae) Europas, Nordafrikas und Vorderasiens, einschliesslich dei\ Azoren, der Kanaren und Madeiras. 388 pp. Stuttgart.1933. Beitrage zur Kenntnis der Schildlause (Hemipt. - Homopt., Coccid.). Entomologischefi Anzeigerl3: 107-108, 116-117. 1936. Neue Beitrage zur Kenntnis der Schildlause (Coccidae). Entomologische Zeitschrift 49: 444. 1938. Verseichnis der aus Nordwest-Deutschland, insbesondere aus Gross-Hamburg, gemeldeterj Schildlause (Homoptera-Coccoidea). Verhandlungen des Vereins fur naturwissenschaftliche Heimfors- chung zu Hamburg 26 (1937): 1-15. 1957. Bin Weiterer Beitrag zur Synonymic der Cocciden. Beitrage zur Entomologie 7: 543-553. Linnaeus, C. 1758. Systema Naturae, &c. Editio decima, reformata. Tomus I. Regnum Animale. 824 pp. Holmiae. 1761. Fauna Svecica, &c. Editio altera, auctior. 578 pp. Stockholmiae. 1766. Systema Naturae, &c. Editio duodecima reformata. Tom. I. Pars II. Regnum Animale. pp. 533-1327. Holmiae.Lonsdale, D. 1980. Nectria infection of beech bark in relation to infestation by Cryptococcus fagisuga Lindinger. European Journal of Forest Pathology 10: 161-168.Maskell, W. M. 1880. Further notes on New Zealand Coccidae. Transactions and Proceedings of the New Zealand Institute 12 (1879): 291-301.1887. Further notes on New Zealand Coccidae. Transactions and Proceedings of the New Zealand Institute 19 (1886): 45-49. 1897. On a collection of Coccidae, principally from China and Japan. The Entomologist's Monthly Magazine 33: 239-244.Melville, R. V. 1982. Opinion 1203. Eriococcidae Cockerell, 1899 conserved: Type species designated for Eriococcus Targioni-Tozzetti, 1868 (Insecta, Homoptera). The Bulletin of Zoological Nomenclature 39: 95-98. Miller, D. R. 1984. Terminology. The Scale 10: 47-49.Miller, D. R. & Gonzalez, R. H. 1975. A taxonomic analysis of the Eriococcidae of Chile. Revista Chilena de Entomologia 9: 131-163.Miller, D. R. & McKenzie, H. L. 1967. A systematic study of Ovaticoccus Kloet and its relatives, with a key, to North American genera of Eriococcidae (Homoptera: Coccoidea: Eriococcidae). Hilgardia 38: 471-539.Modeer, A. 1778. Om Fastflyet, Coccus. Handlingar det Gotheborgska Wetenskaps och Witterhets Samhallets 1: 11-50.Newstead, R. 1891. On some new or little known Coccidae found in England. The Entomologist's Monthly Magazine 26: 164-166.1892. On new or little known Coccidae, chiefly English (No. 2). The Entomologist's Monthly Magazine 28: 141-147. 1897. On Coccus agavium, Douglas. The Entomologist's Monthly Magazine 33: 12-13. 1898. Observations on Coccidae (No. 17). The Entomologist's Monthly Magazine 34: 92-99. 1903. Monograph of the Coccidae of the British Isles 2: 270 pp. London. Parker, E. J. 1975. Some investigations with beech bark disease Nectria in southern England. European Journal of Forest Pathology 5: 118-124.Rasina, B. 1955. Materiali Latvijas PSR Coccoidea fauna. Latvijas PSR Zinatnu Akademijas Vestis 94: 67-75. Reaumur, R. A. F. de 1738. Memoirespourserviral'histoire des insectes. 4: xxvi, 636 pp. Paris.Roemer, J. J. 1789. Genera Insectorum Linnaei et Fabricii iconibus illustrata. viii, 86 pp. Vitoduri Helve torum.Schmutterer, H. 1952. Die Okologie der Cocciden (Homoptera, Coccoidea) Frankens 1. Abschnitt. Zeitschrift fur Angewandte Entomologie 33: 369-420.Schrank, F. von P. 1801. Fauna Boica. 2: viii, 412 pp. Ingolstadt.Signoret, V. 1872. Essai sur les cochenilles ou gallinsectes (Homopteres-Coccides), 8e partie. Annales dela Societe Entomologique de France (5) 1: 421-434.1875. Essai sur les cochenilles ou gallinsectes (Homopteres-Coccides), 14e partie. Annales de la Societe Entomologique de France (5) 5: 15-40.Sulzer, J. H. 1776. Abgekiirzte Geschichte der Insecten nach dem Linaeischen System. Part 1 , xxviii, 274 pp. Part 2, 71 pp. Winterthur.Targioni Tozzetti, A. 1868. Introduzione alia seconda memoria per gli studj sulle Cocciniglie, e catalogo THE BRITISH AND SOME OTHER EUROPEAN ERIOCOCCIDAE 393 del generi e delle specie della famiglie del Coccidi. Atti della Societa Italiana di Scienze Naturali II' 694-738.Tereznikova, E. M. 1981. Coccids, Issue 19, Eriococcidae, Kermesidae, Asterolecaniidae and Coccidae. Fauna Ukraini2Q (19): 215 pp.Thomsen, M., Buchwald, N. F. & Hauberg, P. A. 1949. Angreb af Cryptococcus fagi, Nectria galligena og andre Parasiter paa B0g i Danmark 1939-43. Forstlige Fors0ksvcesen i Danmark 18: 93-326.Wainhouse, D. 1979. Dispersal of the beech scale (Cryptococcus fagi Baer.) in relation to the development of beech bark disease. Mitteilungen der Schweizerischen Entomologischen Gesellschaft 52: 181-183.Wainhouse, D. & Howell, R. S. 1983. Intraspecific variation in beech scale populations and in susceptibility of their host Fagus sylvatica. Ecological Entomology 8: 351-359.Walker, F. 1852. List of the specimens of Homopterous insects in the collection of the British Museum. Part IV. pp. ix, 909-1188. London.Warburton, C. 1937. Annual report for 1937 of the zoologist. The Journal of the Royal Agricultural Society of England 98: 560-564.Williams, D. J. 1984. Some aspects of the zoogeography of scale insects (Homoptera: Coccoidea). Verhandlungen des X International Symposiums iiber Entomofaunistik Mitteleuropas pp. 331-333. j 1985a. Australian Mealybugs. 431 pp. London. t 19856. Scale insects (Homoptera: Coccoidea) of Tresco, Isles of Stilly. Entomologist's Gazette 36: I 135-144. Zopf, W. 1887. DiePilzthiereoderSchleimpilze,pp. 1-174. In Schenk, A. (Ed.),HandbuchderBotanik3: 716 pp. Breslau. Index canthococcus 356.ceris 358gavium 385{nophococcus 356, 357[pterococcus 385 luxi 358, 361 antium 357, 363Cryptococcus 352 levoniensis 357, 365Eriococcus 352, 356 to 353 agisuga 353raxini 386 Synonyms are in italics. glyceriae 358, 367gnidii 382Gossyparia 356gouxi 357greeni 357, 367Greenisca 356Gymnococcus 384 hoheriae 384 inermis 357, 370insignis 358, 372 Kaweckia 356Kuwanina 352, 384 lagerstroemiae 357, 374 munroi 357, 374Noteococcus 352, 384Ovaticoccus 352, 384 parva 384placidus 357, 376pseudinsignis 358, 378Pseudochermes 352, 385 Rhizococcus 356spurius 357, 380thymi 382ulmi380 British Museum (Natural History)Australian Mealybugs D. J. Williams Mealybugs, or Pseudococcoidea, represent one of about 20 families currently recognized asscale insects of the superfamily Coccoidea, and are responsible for a significant proportion ofinsect-related crop damage in Australia. Although they have natural predators, a lack ofauthoritative systematic literature on the group has made identification a haphazard business,and selection of the appropriate natural enemy difficult. Biological methods of controllingthese pests, therefore, have not been as fully exploited as they might have been. This work should help this situation by providing an authoritative identification guide andcomprehensive review of current knowledge on the biology of the 196 species of Mealybugscovered. It will fill the existing gap on the classification of these insects, and in enablingentomologists and agriculturalists to correctly identify and classify specimens, will promote theuse of biological control. 1985 viii + 43 Ipp, 177 line illustrations 0565009532 40 Titles to be published in Volume 51 The ichneumon-fly genus Banchus (Hymenoptera) in the Old World By M. G. Fitton The phytogeny, classification and evolution of parasitic wasps of the subfamily Ophioninae (Ichneumonidae) By I. D. Gauld A cladistic analysis and classification of trichodectid mammal lice (Phthiraptera: Ischnocera)By C. H. C. Lyal The British and some other European Eriococcidae (Homoptera: Coccoidea) ByD. J.Williams Photoset by Rowland Phototypesetting Ltd, Bury St Edmunds, SuffolkPrinted in Great Britain by Henry Ling Ltd, Dorchester