M BRITISH MUSEUM (NATURAL HiSi 26 JU PRESENTED GENERAL 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 four scientific 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 and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works 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 several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the 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 0 565 06011 2 ISSN 0524-6431 Entomology series Vol 51 No 1 pp 1-60 British Museum (Natural History) Cromwell Road London SW7 5BD Issued 25 July 1985 BRITISH MU: The ichneumon-fly genus Banchus (Hymenoptera) in the Old World -^1 M. G. Fitton Department of Entomology, British Museum (Natural History), Cromwell Road, London SW7 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) in the 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 of some 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, three lectotype designations and four nomina dubia. Introduction Banchus has had a chequered history. Taxonomists found it difficult to assign to one of the five traditional subfamilies of the Ichneumonidae, sometimes placing it in the Ophioninae and sometimes in the Pimplinae. As originally defined by Fabricius (1798; 1804) the genus included members of the families Braconidae, Cephidae, Pompilidae, Proctotrupidae and Ibaliidae, as well as the Ichneumonidae. This exotic assortment was reduced by Gravenhorst (1829) to a grouping resembling the genus as understood at present. The main error in his treatment was the transposition of the sexes. Developments in the following 150 years were mainly concerned with description of new species, splitting the genus and faunistic studies. Much of this work was of poor quality; no one made a comprehensive study and those describing and keying species failed to appreciate all but the most superficial characters. A real advance in our understanding of the taxonomy of Banchus came only with publication of the revision of the Nearctic species by Townes & Townes in 1978. Bull. Br. Mus. nat. Hist. (Ent.) 51 (1): 1-60 Issued 25 July 1985 M. G. FITTON I •8 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. Margin of clypeus with a median notch. Genal carina joining hypostomal carina above base of mandible. Upper tooth 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, especially in males. Labium not elongate. Antennal flagellum long, tapered to a slender apex. In males, flagellum with poorly- to well-developed specialized setae (flag setae), 2 to 4 per segment, often arising from a poorly- to well-marked, shallow trough on the dorsal surface. Epicnemial carina absent. Scutellum produced 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 carina present, at least in part. Tarsal claws pectinate. Laterotergites of segments 2 and 3 of gaster about 0-7 times as deep as long. Gaster more or less compressed, relatively strongly so in most females. Ovipositor very short, its sheath about 0-1 times as long as hind tibia. REMARKS. The genus comprises 47 valid, described species. One is Holarctic in distribution, a further 24 occur in the Nearctic (Townes & Townes, 1978) and another 22 are found in the Palaearctic and the mountains of the Oriental region. Banchus is the largest of a group of five genera united by the structure of the mandible. The mandible is of an unusual, almost tridentate form unlikely to have arisen more than once within the Banchinae. A mandible similar in structure is found only in the Diplazontinae (in which subfamily 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 1 species in Argentina which possibly deserves a separate genus). Within this group the most important discriminant characters are the form of the claws, the presence or absence of the epicnemial carina, the extent of the propodeal carinae, the form of the frons, the position of cross- 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 the two groups on the basis of adult and larval characters. Cushman and Perkins associated the Exetastes-gioup more closely with the tribe Lissonotini. Townes (1944) on the other hand linked the two groups as the Banchini, separate from the Lissonotini, mainly on the basis of a rather weak character in the hind wing venation. Townes has consistently maintained this position. Viktorov (1967) supported Cushman's (1937) conclusion, citing several characters of Exetastes which distinguish it from the Banchus-group. Aubert (1978) has gone further, placing Exetastes in a tribe of its own, Exetastini. As with many ichneumonid subfamilies the generic and tribal classification 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 which Townes may have then seen only one specimen). Taxonomy Despite its long period of gestation this study, based on the examination of 3421 specimens of Old World Banchus, cannot be regarded as definitive. It has raised as many taxonomic questions as it has answered. The lack of specimens from some problem areas (taxonomic as well as geographic), 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 was adopted 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 be justified on scientific grounds. Apart from obvious species-pairs or very small groups it is difficult to deduce anything of the phylogenetic 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 do not correlate in ways which indicate clear-cut species-groups. Specialization by reduction (an obvious example is in the length of the scutellar spine) is probably one of the complicating factors. 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 most obvious morphological differences are in the shape of the gaster , the antennae and the maxillary palps. The specializations in the female seem to be associated with oviposition and in the male with mate-finding and courtship. In some species the overall colour pattern is the same in male and female, in others it is completely different, for example, in B. falcatorius and B. volutatorius the females are largely black with reddish legs while the males are conspicuously marked with yellow. This female colour pattern, although still aposematic, probably renders her less conspicuous while searching for hosts (which, in B. falcatorius, probably occurs very low in the vegetation at soil level). As a general rule males have a yellowish face with a median black stripe and females a black face, sometimes with yellowish orbital marks, and there are often also differences in the colour of the antennae. The characters which have been used in this study are discussed below. The formal descriptions of species in the systematic section of the paper are restricted largely to the characters used in the keys. When comparing specimens with figures allowance should be made for some variation. Head. The overall shape of the head is important in distinguishing species. The characters most easily used are the relative width of the face and the size of the malar space. These are measured as indicated in Figs 2 and 3. The shape of the head behind the eyes varies between species but is much more difficult to use as a distinguishing character. The maxillary palps are specialized in many species (Figs 11-32, 103-123). The relative lengths of the two terminal segments are important and in males the way in which the fourth segment is widened and flattened is characteristic. The antennae vary in length and in the shape of the terminal segments, both characters being difficult to use in keys. In the male the distal half of the antenna bears specialized setae ('flag setae') on its dorsal surface (Figs 124-129). The development of these varies greatly between species. In the most primitive cases the setae are not easily distinguished from others on the antenna. In the most specialized forms the setae (two, three or four per segment) are erect, widened and flattened and arise from a shallow polished trough. All segments do not bear these setae, there is a zone (which varies in position between species) in which they are best developed and proximally and distally to this the setae become progressively less modified. The descriptions given for each species apply only to the zone bearing the most modified setae. The shape and number per segment are subject to some variation, but together with the form of the male maxillary palps these setae give the 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). The primitive condition, it is supposed, is to have a well-developed spine. The length of the spine is impossible to measure accurately because of the lack of reference points, and there is variation within species. Nonetheless it is still useful taxonomically. The lower postero-lateral corners of the mesothorax in some species are formed into weak tubercles which, although relatively constant 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 behind and 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 the general proportions of the legs as a whole. The length of the fore wing (measured from the tip of the 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 descriptions and figures, particularly those of the apex of the gaster (Figs 82-102). The distortion most affects the amount of compression of the posterior half of the gaster, the telescoping of segments and the compression and profile of sternite 6 (the subgenital plate) which in some species has a weaker, less sclerotised portion on its posterior margin. The form of tergite 1 is not subject to distortion and it offers useful characters in males and females (Figs 59-81), although subject to some variation (compare Figs 65 and 66 of male and female gudrunae). The proportions of tergite 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 was taken to ensure that such variation was taken into account in construction of those parts of the keys relying on colour. Biology and hosts Townes & Townes (1978) state that Banchus is a genus of open shrubby country and that females fly rather low and males higher and faster. However, there are few published observations on the habits and habitats of individual species. Adults have been collected from flowers. The black and yellow and/or reddish colour patterns may be aposematic (Townes & Townes note that specimens give off a strong pungent odour when captured). B. falcatorius has been recorded as the prey of an asilid fly. Most species probably are univoltine and adults are on the wing for some period during late spring or early summer. A few species (e.g. B. dilatatorius) occur in early spring. These probably pass 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 few known specimens of B. moppiti come from localities as scattered as Great Britain, Switzerland and Spain. There appears to be a similar pattern in the species in the eastern Palaearctic and mountains of the Oriental region, with some specimens of the more southern species occurring in 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 male encounters a female he orientates himself face to face and starts fanning with his wings, probably creating an airstream along the female and himself. The gaster of the male is also raised and at intervals these actions are interrupted by an attempt to mount the female. Using the tips of the antennae the male gently strokes the legs and sides of the thorax of the female. If the female is receptive copulation takes place. The hosts of Banchus are Lepidoptera. All of the reliable observations relate to species of Noctuidae, though species of seven other lepidopterous families are also recorded. Whether or not any or all of the non-noctuid records should be rejected is difficult to assess. In the Nearctic species B. flavescens it has been shown that females oviposited readily in the larvae of its usual host 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 to M. configurata larvae (Arthur & Ewen, 1975; Ewen & Arthur, 1976). All these hosts are noctuids and in addition to the reluctance of the females to oviposit in some species the parasite failed to develop in any except M. configurata, because of successful host defence reactions. Van Veen (1982), working with B. hastator in the Netherlands, found 'that an antennal contact of a few strokes with the cuticle of the host's skin was sufficient to reject a strange host by walking away or to a susceptible instar of Pfanolis] flammea [the usual host] immediately by an acute sting reaction'. He unfortunately does not identify the 'strange host'. All these results could be used to support a contention that Banchus species are probably monophagous on species of Noctuidae, but for one species, B. volutatorius , there are reliable rearings from at least three noctuid 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 two Hymenoptera (Maneval, 1935) can probably be dismissed out of hand. The larvae normally develop as solitary internal parasites and it is usually suggested that oviposition is into an early instar host larva. In experiments, van Veen (1982) has demonstrated a preference for early instar larvae in B. hastator, although females will attack all larval instars except 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 of the final larval instar of four species are illustrated by Short (1978). Short's work should be used with caution: his figures of Banchus show why. The illustrations of B. femoralis and B. hastator show differences which might be thought significant; however, both specimens are of the same species (B. hastator, of which femoralis is a synonym) and one figure shows the view from outside the 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 and entered the soil to pupate (this has been reported in several species). Cocoons of Banchus are elongate-ovoid, the silk dense and almost black. The adult emerges through a hole cut next to one end. Barron (1976) records a species of Banchus (from Maryland) parasitized by Euceros medialis Cresson (Hymenoptera: Ichneumonidae). Systematic list of hosts The list below includes all the hosts of Banchus detailed elsewhere in this paper. Those details can 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) ICHNEUMONIDAE Ophion luteus (Linnaeus) Nomenclatural summary agathae sp. n. altaiensis Meyer, 1927 (nomen dubium) cerinus Chandra & Gupta, 1977 cre/e/densisUlbricht, 1916 croaticusHensch, 1928 dilatatorius (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, 1937 japonicus(Ashmead, 1906) mauricettae sp. n. moppitisp. n. noxMorley, 1913 pa/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, 1929 tholussp. n. tumidus Chandra & Gupta, 1977 furcaf or Aubert, 1981 volutatorius (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, 1926 zonafusRudow, 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 (except altaiensis) information is under the following headings. Synonymy. When types have not been available for study (because of loss or destruction or because they could not be obtained on loan) the inclusion of a nominal species in a synonymy is based on consideration of the original description and/or it is substantiated by reference to a published treatment or is considered more fully under 'Nomenclature'. Nomenclature. This includes discussion of synonymies, type-restrictions, etc. It takes up a considerable 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 sociological factors rather than biological ones. Description. The descriptions for females and males are separate. The characters included are mainly those used in the keys, the ones which apply equally to both sexes are given for the female only (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 and discussion of its taxonomy. Biology and hosts. This is a brief summary based on data associated with material examined and information gleaned from the literature. Names of hosts have been up-dated to conform with current usage. Distribution. As given this is based entirely on data taken from specimens examined, as are the distribution maps. Map 1 shows the distribution of the genus (that is, the combined distributions of all the species) in the Old World for comparison with the distributions of individual species (Maps 2-11). The records in the literature are not reliable and only a few references 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, the data 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. All specimens on which this revision is based bear my determination labels. Some detail of collecting localities is presented in the distribution maps for each species. However, a feature of much European material is inadequate locality data on labels and there are no points on the maps for perhaps a third of all the specimens examined. Dates of collection are also often lacking; for instance, in an attempt to relate morphological to seasonal variation in Banchus pictus only 90 of the 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 always straightforward. The short, broad ovipositor sheaths are sometimes mistaken for the male genitalia. The most reliable way to sex specimens is to examine the last visible tergites and sternites 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 sharply expanded 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, never arising 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 separating laterotergite 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 of segment 1 and distal section reddish yellow. Segments of gaster each black anteriorly and yellow 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 some blackish 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. Gaster with tergite 1 at least 1-60 times as long as broad; tergite 3 with a crease separating the laterotergite 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 slightly or 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 of maxillary 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 a yellow 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 about 60 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 be brownish. 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 extensive black 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 times vertical 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 notes under 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 a black 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 a median 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 as broad 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 and almost 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 reddish band. 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 basal width of mandible. Maxillary palp as in Fig. 11. Antenna with apical segments as long as broad. Scutellar spine (Fig. 33) about 0-3 as long as scutellum. Mesopleuron and mesoscutum shining, only very weakly sculptured, with strong punctures, on the mesopleuron separated by much less than their diameter. Hind femur 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 the tips of the fore wings (when folded back). Tergite 3 with a crease along its anterior 0-5 separating laterotergite. 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 pedicel yellow ventrally. Maxillary palp yellowish with dark marks, including distal 0-7 of segment 4 and all of segment 5. Posterior part of propodeum yellow, narrowly black posteriorly. Hind coxa black with a yellow dorsal 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 of mandible. Maxillary palp as in Fig. 103; segment 4 flattened and considerably widened; segment 5 about 0-7 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 0-5 separating the laterotergite. Colour: black and yellow. Antenna black with scape, pedicel and proximal and distal parts of flagellum yellow ventrally. Maxillary palp with segments 1 yellowish, 2 and 3 reddish, 4 and 5 blackish. Posterior part of propodeum yellow, narrowly black posteriorly. Hind coxa black with a yellow dorsal spot and narrow postero-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 is keyed to a large extent on colour characters but the form of the posterior part of the gaster is also important. 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 dubium Banchus 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 to identify it with any species not proven to occur there. Its identification will have to wait until more extensive collections 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 and mesoscutum 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 folded back). Tergite 3 with a crease along its anterior 0-6 separating laterotergite. Tergite 7 and sternite 6 as in Fig. 6. Colour: mainly yellow, with black marks. Face yellow with a black mark between and just below antennal sockets. Antenna yellow with distal half of flagellum and an externo-lateral mark on pedicel blackish. Maxillary palp yellowish. Posterior part of propodeum yellow. Hind coxa yellow, blackish distally. Hind femur yellow. Segments of gaster yellow, with basal third of tergites 1,2 and 3 and of sternite 2 black. Male. Unknown. REMARKS. A pale-coloured species with relatively long appendages. Its closest relatives seem to be tholus and tumidus. DISTRIBUTION (Map 4, p. 20), BIOLOGY AND HOSTS. The species is known only from the holotype, which was collected 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. tholus and B. tumidus. DESCRIPTION. Female. Width of lower face 0-90-1-00 times vertical length of eye. Malar space 0-60-0-75 times basal width of mandible. Maxillary palp as in Fig. 13. Antenna with apical segments longer than broad. Scutellar spine (Fig. 34) about 0-3 as long as scutellum. Mesopleuron and mesoscutum strongly coriaceous, with moderately strong punctures, on mesopleuron separated by about their diameter. Hind femur 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 as long as broad, its dorsal profile as in Fig. 60 (male). Gaster compressed from segment 4, reaching about to tips of fore wings (when folded back). Tergite 3 with a crease along its entire length separating laterotergite. 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 below each antennal socket. Antenna entirely black except for a yellow patch on underside of scape, and rarely also on pedicel. Maxillary palp blackish with part or all of segment 2 and proximal 0-7 of segments 3 and 4 reddish yellow. Posterior part of propodeum yellow anteriorly, black posteriorly. Hind coxa black with a yellow dorsal patch. Hind femur black, yellow proximally and distally and sometimes dorsally. Segments of gaster 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 basal width of mandible. Maxillary palp as in Fig. 104; segment 4 narrow and cylindrical, flattened and widened only at its extreme distal apex; segment 5 less than 0-2 as long as 4. Antennal flag setae (Fig. 125) at about 50 degrees, not flattened, 2 or 3 per segment, not arising from a trough. Hind femur 6-60-7-65 times as long as deep. Fore wing length 9-1-10-8 mm. Tergite 1 of gaster 1-80-2-10 times as long as broad (Fig. 4). Tergite 3 with a crease along its entire length separating laterotergite. Colour: black and yellow. Antenna black, with scape and pedicel yellow ventrally. Maxillary palp blackish with segments 2 wholly and 3 and 4 proximally reddish yellow. Posterior part of propodeum black with a variable amount of yellow anteriorly, ranging from a very broad band along posterior transverse carina to a small median spot to absent entirely. Hind coxa black, usually with a yellow dorsal patch. Hind femur black, yellow proximally and distally and sometimes dorsally. Segments of gaster each black anteriorly, 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 the structure of the maxillary palps and its coloration. BIOLOGY AND HOSTS. Dates of collection range from mid-February (in Spain and Portugal) to mid-May (in Scotland and Ireland). Stelfox (1936: 63) records males flying in numbers round ivy (Hedera helix) and gorse (Ulex europaeus) in Ireland. Specimen labels indicate capture of males on Betula and around Buxus and females from pine. I have seen 11 reared specimens, 3 without a satisfactory host identification, the remaining 8 (including both 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 colour of the underside of the antennae. The description of Banchus sibiricus fits well specimens from Kamchatka (which include those identified by Roman (1931: 29) as lavrovi var.) which are considered conspecific with dilatatorius but are discussed under '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 0-80-1-00 times basal width of mandible (Fig. 2). Maxillary palp as in Fig. 14. Antenna with apical segments about as broad as long. Scutellar spine (Figs 35, 36) very small or virtually absent. Mesopleuron and mesoscutum coriaceous, 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 tips 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. 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 absent medially. Hind coxa black. Hind femur black, narrowly yellowish proximally and usually widely yellowish distally. Segments of gaster each black anteriorly, yellow posteriorly, the yellow bands usually less than 0-3 of length of segments. Male. Width of lower face 1-10-1-25 times vertical length of eye. Malar space 0-85-0-95 times basal width 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, relatively narrow, 2 per segment, arising from a polished trough. Hind femur 4-35-5-00 times as long as deep. Fore wing length 8-6-10-7 mm. Tergite 1 of gaster 1-25-1-55 times as long as broad. Tergite 3 with a crease along its anterior 0-3 separating laterotergite. Colour: black and yellow. Antenna blackish or brownish dorsally, yellow or yellowish ventrally (except for apex of flagellum which is dark). Maxillary palp blackish or brownish. Posterior part of propodeum black, entirely or with a yellow band (sometimes broken medially) along posterior transverse carina. Hind coxa black. Hind femur yellow with a black mark on interne-lateral and ventral surface, and sometimes also dorsally near base. Segments of gaster each black anteriorly, yellow posteriorly, the yellow bands usually 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 the collections 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 specimens seen. All the specimens from Kamchatka (1 $ (MLSU) and 4 $ , 2 0" (NR)) have yellow absent but with reddish in place of the yellow on the legs and on the face of the male, and with the margins of some or all of tergites 1, 2 and 3 of the gaster narrowly to broadly reddish brown. Other than in these colour characters the Iranian and Kamchatka specimens do not differ much from 'typical' dilatatorius . It is impossible to decide whether or not they represent extremes of colour variation or distinct species. Unfortunately the character which might enable a decision to be reached, the male flag setae, cannot be investigated. Of the two male specimens one lacks the head completely and the other lacks both antennae. The Kamchatka specimens 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 dark form of dilatatorius. I could not examine these Mongolian specimens because Momoi has so far failed to return them to Budapest (Zombori, pers. comm.). The six specimens in Stockholm (NR) had been identified 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 March to late April. The earliest dates of collection (7 and 13 March) come from sand dune areas in the Netherlands. The label on one female records that it was found on catkins of Salix caprea. I have seen no reared specimens of this species, although the following hosts are recorded in the literature: Blepharita adusta (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 very few records from the area bordering the Atlantic (France, the British Isles and Norway) and only a few from Denmark and Sweden); the European Alps; and with some scattered records in the western U.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 from that species in the coloration of the gaster. However, they fit the male of B. falcatorius and there can be little 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 as described by Meyer; the most reasonable explanation of the colour is that the specimen had suffered from overlong exposure to cyanide in a killing jar. The description of lavrovi fits specimens from the northern part of the eastern Palaearctic and here included 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 best provisionally 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 0-60-0-70 times basal width of mandible. Maxillary palp as in Fig. 15. Antenna with apical segments just longer than broad. 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. Hind femur 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 times as long as broad, its dorsal profile as in Fig. 62. Gaster (Figs 9, 10) strongly compressed, reaching beyond tips 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 the gaster. Face black, often with very small reddish or yellowish orbital spots. Antenna dark brown or blackish dorsally, brown or reddish ventrally. Maxillary palp reddish with segments 4 apically and 5 wholly dark. Posterior part of propodeum entirely black. Hind coxa black, very rarely with a postero-ventral reddish spot. Hind femur reddish yellow, entirely or, in a few specimens, blackish medially. Segments of gaster black, rarely entirely so, usually with variable reddish marks on segments 1,2 and 3 and more rarely on 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 basal width of mandible. Maxillary palp as in Fig. 106; segment 4 considerably widened and flattened; segment 5 about 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 as deep. Fore wing length 9-2-11-7 mm. Tergite 1 of gaster 1-40-1-70 times as long as broad. Tergite 3 with a crease 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 reddish yellow, 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. Hind coxa black, usually with a yellowish or reddish patch postero-ventrally and sometimes also dorsally. Hind femur reddish yellow, almost always entirely but in a few specimens blackish medially. Gaster with tergite 1 black anteriorly, reddish and then yellow posteriorly; tergites 2 and 3 each reddish anteriorly, yellow posteriorly (with some blackish laterally); remaining tergites black, 4, 5 and 6 each with a postero-median yellow spot. REMARKS. B. falcatorius has several specialized features and it is the species in which sexual dimorphism is most marked. There is also some geographical variation in characters; for instance, specimens from the eastern part of the range have the hind femur darkened medially and a few specimens from Turkey have a well-developed spine on the scutellum. More than one species might be included in the taxon here recognized, but in the limited material from the eastern Palaearctic available for study no clear-cut segregates could be recognized. Although the sex association in this species is not in doubt, interesting confirmation comes from the discovery of four gynandromorph specimens in the extensive material examined. These were the only gynandromorphs found (or at least recognized) during the present study. Details of the specimens are as follows. Italy: Piemonte, Susa, 24. vi. 1872 (Gribodo) (MCSN). Head male (unfortunately the antennal flagella are missing), thorax and abdomen female. Italy: Cadore, Valle del Boite (IEUB). Head female, thorax and propodeum apparently mainly female but with yellow marks on the left of the mesoscutum and scutellum (male characters), gaster apparently male (including the genitalia) but with some bilateral asymmetry in colour anteriorly. ?locality [illegible] (Giraud coll.) (MNHN). Head (including antennae and palps) male on the left and female on the right, prothorax and anterior of mesothorax mainly female but partly male on left, posterior of mesothorax and remainder of body female. Interestingly this specimen had 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 that the 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 late May and late August). The adults have often been collected from umbellifer flowers. One female is labelled 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 of those 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 major pest of root crops (particularly various beets and carrots) and its association with B. falcatorius was the subject 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 to reach the host larvae, which feed at soil level. Other hosts recorded on specimen labels are: Deilephila porcellus (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 the others (all the specimens are without locality or date !). Other hosts recorded in the literature are: Agrotis exclamationis (Linnaeus) (Noctuidae) (Meyer, 1927a: 81) and, erroneously, two Hymenoptera (Diprion pini (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 the U.S.S.R. and further east than the Caucasus none from south of 50 degrees N. In the more southern parts of 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, Great Britain, 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 syntype specimens. Both bear labels 'Cidaphurus flavomaculatus Cam. Simla' in Cameron's handwriting. The lectotype has, in addition, Nurse's typewritten data label 'Simla 5.97.'. It is unfortunate that Cameron's original description applies better to the paralectotype because it is not conspecific with the lectotype and is B. punkettai. DESCRIPTION. Female. Width of lower face 0-75-0-80 times vertical length of eye. Malar space 0-50-0-55 times basal width of mandible. Maxillary palp as in Fig. 16. Antenna with apical segments about as broad as long. Scutellar spine (Fig. 39) varying from very short to about 0-5 as long as scutellum. Mesopleuron and mesoscutum shining, weakly coriaceous, with moderate punctures, on mesopleuron separated by about or a little more than their diameter. Hind femur 5-60-5-65 times as long as deep. Fore wing length 8-8-9-2 mm. Tergite 1 of gaster 2-25-2-35 times as long as broad, its dorsal profile as in Fig. 63. Gaster strongly compressed from posterior of segment 3, reaching to about 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. 86. Colour: black and reddish, with some yellow marks. Face reddish, yellowish laterally and with a black median stripe. Antenna blackish dorsally, scape and pedicel yellowish ventrally, flagellum brown ventrally. Maxillary palp reddish yellow, segments 4 distally and 5 wholly dark. Posterior part of propodeum entirely reddish. Hind coxa black with large dorsal and postero-ventral reddish patches. Hind femur reddish with a black stripe ventrally. Segments of gaster black anteriorly, very broadly reddish posteriorly. Male. Width of lower face 0-80 times vertical length of eye. Malar space 0-60 times basal width of mandible. Maxillary palp as in Fig. 107; segment 4 considerably widened and flattened; segment 5 about 0-5 as long as 4. Antennal flag setae (similar to Fig. 126, but setae slightly longer) upright, long, flattened, 2 per segment, arising from a polished trough. Hind femur 5-25 times as long as deep. Fore wing length 8-3 mm. Tergite 1 of gaster 2-30 times as long as broad. Tergite 3 with a crease along its anterior 0-5 separating laterotergite. Colour: black and yellow, with some reddish, especially on legs. Antenna black dorsally, scape and pedicel 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 part of propodeum entirely yellow. Hind coxa black, with a yellow dorsal spot contiguous interno-laterally with a postero-ventral spot. Hind femur reddish yellow with a black mark ventrally. Segments of gaster each black 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 specimens collected 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, Himachal Pradesh, 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 km SE. 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 0-45-0-55 times basal width of mandible. Maxillary palp as in Fig. 17. Antenna with apical segments longer than broad. 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 femur 6-75-7-20 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, reaching about to tips of fore wings (when folded back). Tergite 3 with a crease along its whole length separating laterotergite. Tergite 7 and sternite 6 as in Fig. 87. Colour: black and yellow. Face yellow with a black median stripe. Antenna black (sometimes slightly brownish distally), scape, and sometimes pedicel, with a yellow patch beneath. Maxillary palp reddish yellow with segment 1 and distal parts of 4 and 5 blackish. Posterior part of propodeum yellow, narrowly black posteriorly. Hind coxa black with a yellow dorsal patch. Hind femur yellowish with proximal half largely black. Segments of gaster each black anteriorly, yellow posteriorly, the yellow bands relatively wide. Male. Width of lower face 0-80-0-90 times vertical length of eye. Malar space 0-45-0-65 times basal width of mandible. Maxillary palp as in Fig. 108; segment 4 flattened and widened distally; segment 5 about 0-5 as long as 4. Antennal flag setae (similar to Fig. 127, but with setae slightly shorter and very much broader) upright, relatively short, flattened and very broad, 2 (occasionally 3) per segment, arising from a polished trough. Hind femur 7-00-7-45 times as long as deep. Fore wing length 8-8-10-0 mm. Tergite 1 of gaster 2-10-2-40 times as long as broad. Tergite 3 with a crease along its whole length separating laterotergite. Colour: black and yellow. Antenna black dorsally, scape and pedicel yellow ventrally, flagellum brownish ventrally. Maxillary palp reddish yellow, segment 1 dorsally narrowly black, segment 4 distally and 5 wholly blackish. Posterior part of propodeum yellow, very narrowly black posteriorly. Hind coxa black with a large yellow dorsal patch. Hind femur yellow, blackish on its proximal 0-3, especially ventrally and 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 Palaearctic difficult. The form of the male palp suggests a relationship withpalpalis and crefeldensis but the flag setae are 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 the middle 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 dealt with 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) (see nomenclatural notes under palpalis). Ichneumon pungitor is a junior objective synonym of hastator. It was proposed apparently because of the secondary 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 a replacement name has not been recognized previously and it had been placed incorrectly in the synonymy of B. volutatorius (Linnaeus) (e.g. Aubert, 1978: 158). Ichneumon reticulator has not been identified since its description. Roman (1912: 277) reported that there were no specimens in Thunberg's collection and suggested that the species was a Banchus. The description 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. The description fits particularly dark specimens ofhastator, with which it is here synonymised. Notwithstanding Meyer's comments on kolosovi andfemoralis, it seems that he realised his mistake because in his treatment of the genus (1934: 229) he mentions, under femoralis, material from Sverdlovsk (the type-locality of kolosovi) collected by Kolosov. This is presumably the type-material of kolosovi. No material collected by Kolosov 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 0-50-0-65 times basal width of mandible. Maxillary palp as in Fig. 18. Antenna with apical segments longer than broad. 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 1-50-1-70 times as long as broad, its dorsal profile as in Fig. 64. Gaster weakly compressed, reaching about to tips 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. 88. Colour: mainly black, with yellow and reddish, particularly on legs. Face black with yellow orbital marks, which are usually small but are sometimes large and with a pair of very small reddish marks in centre of face. Antenna with scape and pedicel black, scape, and usually pedicel, yellow ventrally. Flagellum brown dorsally, reddish ventrally, darker at extreme base and apex. Maxillary palp reddish with segment 1 black and segments 2 proximally, 4 distally and 5 wholly darker. Posterior part of propodeum entirely black. Hind coxa black. Hind femur reddish or yellowish proximally and distally, medially varying from reddish dorsally and blackish ventrally to all blackish. Segments of gaster black, sometimes almost entirely except for brownish posterior margin of sternites 1 and 6 and tergite 7, but usually with posterior margins of further 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 basal width of mandible. Maxillary palp as in Fig. 109; segment 4 only slightly widened and flattened; segment 5 almost as long as 4. Antennal flag setae (Fig. 124) relatively poorly differentiated, small, not flattened, 2 per segment, not arising from a trough. Hind femur 4-60-5-50 times as long as deep. Fore wing length 7-2-10-7 mm. Tergite 1 of gaster 1-45-1-80 times as long as broad. Tergite 3 with a crease along its anterior 0-3 separating laterotergite. Colour: mainly black, with some reddish and yellow, particularly on legs. Antenna black dorsally, scape and pedicel yellow ventrally, flagellum reddish ventrally. Maxillary palp with segment 1 blackish or blackish and yellow, 2 and 3 reddish, 4 and 5 blackish or brownish. Posterior part of propodeum entirely black. Hind coxa black (in a single specimen with a small yellowish dorsal spot and the postero- ventral margin reddish). Hind femur reddish or yellowish proximally and distally, medially varying from reddish dorsally and blackish ventrally to all blackish. Gaster sometimes almost entirely black except for brownish posterior margins of some segments, but usually with posterior margins of segments narrowly yellowish with adjacent brownish bands. REMARKS. This is an isolated species, with few obvious specializations. The primitive condition of the antennal 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 in May, June and early July. Reared specimens have emergence dates in March and April also. Some details of the biology and ecology of the species have been investigated as a result of the association with an economically 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 van Burgst, 1927) but has included studies of embryonic and larval development and morphology (Beirne, 1941; Biedowski & Krainska, 1926; Shevyrev, 1913), courtship, host identification and host-stage preferences, egg-laying and the reproductive system (van Veen, 1982). There are many specimens reared from P. flammea in collections. Apart from that species, the following species 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 ?Achlya flavicornis (Linnaeus) (Thyatiridae) (1 $, Great Britain (Lyle) BMNH). While one might cast doubt on these records it is difficult to investigate them critically. Kolubajiv (1934: 114, 116) records Lymantria dispar (Linnaeus) (Lymantriidae) as a host. DISTRIBUTION (Map 6, p. 27). Widely distributed in northern Europe as far east as Leningrad, perhaps extending further east to the Urals (if my synonymy of kolosovi is correct (see 'Nomenclature' above)). It also 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 basal width of mandible. Maxillary palp as in Fig. 19. Antenna with apical segments longer than broad. Scutellar spine (Fig. 42) indistinct. Mesopleuron and mesoscutum coriaceous, with moderate punctures, on mesopleuron separated by more than their diameter. Hind femur 6-55 times as long as deep. Fore wing length 9-3-9-5 mm. Tergite 1 of gaster 1-65 times as long as broad, its dorsal profile as in Fig. 67. Gaster only weakly compressed, not reaching to tips of fore wings (when folded back). Tergite 3 with a crease along 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 small yellow dorsal patch. Hind femur yellowish, blackish ventrally and posteriorly on its proximal 0-5. Segments of 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, and tergites 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 a female. The material collected in 1959 and supposedly including both sexes (Hellen, 1961: 37) cannot be found 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 labelled Type No. 7259 U.S.N.M.' and must be considered as syntypes. Other labels include 'Nawaia japonica Ashm' on one specimen and 'Nawaia japonica Ash. Paratype' on the other, the latter on a similar label but in a different hand and obviously more modern than the former. The specimen labelled as paratype also bears a label '47' and is in much better condition than the other. I have labelled and hereby designate it as lectotype. DESCRIPTION. Female. Width of lower face 0-80-0-90 times vertical length of eye. Malar space 0-35-0-45 times basal width of mandible. Maxillary palp as in Fig. 20. Antenna with apical segments about as long as broad. Scutellar spine (Fig. 43) about 0-3 as long as scutellum. Mesopleuron and mesoscutum shining and only 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 1-45-1-85 times as long as broad, its dorsal profile as in Fig. 68. Gaster only weakly compressed, reaching to tips of fore 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 mark below 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 ^ gudruni • hastator V 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 femur blackish, yellow proximally and distally. Segments of gaster black with posterior margins broadly yellow and 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 of mandible. Maxillary palp as in Fig. 110; segment 4 flattened and considerably widened; segment 5 about 0-5 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 as deep. Fore wing length 9-3-10-3 mm. Tergite 1 of gaster 1-50-1-65 times as long as broad. Tergite 3 with a crease along almost its entire length separating laterotergite. Colour: black and yellow. Antenna black with scape, pedicel and proximal end of flagellar segment 1 yellow 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 and yellow. Hind femur yellow with a black mark ventrally and interno-laterally, usually extending also to dorsal surface. Segments of gaster black with posterior margins broadly yellow and with a pair of yellow spots 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 the only 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 times basal 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, with moderate to strong punctures, on mesopleuron separated by a little more than their diameter. Hind femur 5-20-5-30 times as long as deep. Fore wing length 10-5-10-6 mm. Tergite 1 of gaster 1-90-2-35 times as long as broad, its dorsal profile as in Fig. 69 (male). Gaster moderately compressed, not quite reaching tips of fore 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 sometimes with a blackish median stripe. Antenna entirely reddish yellow except for extreme apex of flagellum and scape and pedicel dorsally, which are blackish. Maxillary palp reddish yellow with segments 4 distally and 5 wholly dark. Posterior part of propodeum reddish, entirely or with black marks posteriorly. Hind coxa varying from reddish with black patches to almost entirely black. Hind femur entirely reddish. Segments of gaster 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 basal width of mandible. Maxillary palp as in Fig. Ill; segment 4 flattened and widened; segment 5 about 0-8 as long 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 femur 5 • 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 long as 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 wholly blackish. Posterior part of propodeum entirely yellow. Hind coxa largely reddish yellow, blackish or brownish antero-ventrally and dorso-posteriorly. Hind femur reddish, sometimes with a blackish mark ventrally. Segments of gaster each black, reddish and yellow (from anterior to posterior), the area of black reduced 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 from 300-4500 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, 840 m, 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, U Long 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 0-85-1-00 times basal width of mandible. Maxillary palp as in Fig. 22. Antenna with apical segments about as broad as long. Scutellar spine (Fig. 45) about 0-6 as long as scutellum, usually downcurved at the tip. Mesopleuron and mesoscutum coriaceous, usually strongly so, with strong punctures, on mesopleuron separated by much 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 its anterior 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 considerably in 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 coxa black, usually with a small yellow dorsal patch. Hind femur black, yellow proximally and distally and usually 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 basal width of mandible. Maxillary palp as in Fig. 113; segment 4 flattened and widened; segment 5 not quite as long 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 separating laterotergite. Colour: black and yellow. Antenna black with scape and pedicel yellow ventrally. Maxillary palp blackish. Posterior part of propodeum yellow anteriorly, black posteriorly. Hind coxa black, usually with a yellow dorsal spot and sometimes also a postero-lateral one. Hind femur yellow with an extensive black mark, 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 relatively shorter appendages (and gaster in females) readily differentiate it. BIOLOGY AND HOSTS. Unknown. The specimens were collected early in the year (in February, March and April) with the exception of a female taken in October. DISTRIBUTION (Map 7, p. 29). The few known specimens come from widely scattered localities in western Europe. 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 are identically 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 basal width of mandible. Maxillary palp as in Fig. 23. Antenna with apical segments longer than broad. Scutellar spine (Fig. 46) about 0-7 as long as scutellum. Mesopleuron and mesoscutum coriaceous, with moderate punctures, on mesopleuron separated by a little more than their diameter. Hind femur 6-40 times as long as deep. 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 and sternite 6 as in Fig. 93. Colour: black and reddish, with some yellow marks. Face black with very broad yellow orbital marks which also extend beneath antennal sockets. Antenna dark brown, blackish dorsally, scape and pedicel yellowish 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 transverse carina. 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 of mandible. Maxillary palp as in Fig. 112; segment 4 flattened and considerably widened; segment 5 about 0-6 as long as 4. Antennal flag setae (similar to Fig. 126, but setae slightly longer) upright, long, flattened, 2 per segment, arising from a polished trough. Hind femur 6-05 times as long as deep. Fore wing length 9-7 mm. Tergite 1 of gaster 2-15 times as long as broad. Tergite 3 with a crease along its anterior 0-5 separating laterotergite. Colour: black and yellow, with some reddish, especially on legs. Antenna black dorsally, scape and pedicel yellow ventrally, underside of flagellum with segment 1 and distal part reddish yellow, remainder black. Maxillary palp with segment 1 yellowish, 2 and 3 wholly and 4 proximally reddish yellow, 4 distally and 5 wholly brownish. Posterior part of propodeum yellow, narrowly black postero-medially. Hind coxa black, with a yellow dorsal spot contiguous interno-laterally with a postero- ventral spot. Hind femur reddish with a black mark ventrally. Segments of gaster each black anteriorly, yellow posteriorly, with junction of the two areas reddish. REMARKS. Closely related toflavomaculatus. The aggregate differences between these taxa warrant their separation, 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 on flowers 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 events culminated in Townes & Townes (1978: 532) designating as lectotype of monileatus a female which was volutatorius , rather than the species as it had been identified previously. In addition, as a result of misinformation about my examination of the holotype of hastator, Townes & Townes incorrectly synonymized monileatus with that species. Female volutatorius and palpalis superficially are strikingly similar, 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 originally examined Gravenhorst's syntypes (in 1964) Townes probably checked the identity of the species and then made the mistake of selecting the specimen in best condition for designation as lectotype without re-checking it. Under the International Code the lectotype designation has priority over the previous restriction of the taxonomic species. Thus monileatus becomes a junior synonym of volutatorius while the species previously known as monileatus must take the next available name -palpalis. The name hastator correctly applies to the species previously 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 on the 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 0-65-0-75 times basal width of mandible. Maxillary palp as in Fig. 24. Antenna with apical segments longer than broad. Scutellar spine (Fig. 47) long, more than 0-8 as long as scutellum (in a very few specimens only 0-3 as long as scutellum). Mesopleuron and mesoscutum coriaceous, with fine to moderate punctures, on mesopleuron 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 in Fig. 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 and sternite 6 as in Fig. 94. Colour: almost entirely black except for mainly reddish legs and a few small yellowish marks. Face black with yellow orbital marks, usually very small and sometimes absent. Antenna black, entirely or with yellowish marks ventrally on scape and pedicel. Maxillary palp yellowish or brownish with segments 1 and 5 wholly and 3 and 4 distally blackish. Posterior part of propodeum entirely black. Hind coxa black. Hind femur reddish yellow, usually entirely, rarely darkened ventrally. Segments of gaster black, sternites and posterior 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 times basal width of mandible. Maxillary palp as in Fig. 114; segment 4 narrow and cylindrical, flattened and widened only at its extreme distal apex; segment 5 less than 0-2 as long as 4. Antennal flag setae (similar to Fig. 125, but setae slightly longer and showing signs of flattening) at about 50 degrees, very weakly flattened, 3 per segment, not arising from a trough. Hind femur 5-35-6- 10 times as long as deep. Fore wing length 9-8-11-4 mm. Tergite 1 of gaster 1-90-2-25 times as long as broad. Tergite 3 with a crease along its anterior 0-5 separating laterotergite. Colour: black, with some yellow and with legs mainly reddish yellow. Antenna black, scape and pedicel yellow ventrally, proximal and distal parts of flagellum often brownish yellow ventrally. Maxillary palp with 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 an externo-lateral reddish yellow mark and sometimes also dorsal and/or ventral yellowish spots. Hind femur entirely reddish yellow. Segments of gaster black, sometimes brownish or yellowish on their posterior margins, 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 with female 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 come mainly, but not exclusively, from the southern parts of the range. In North America, where the geographical variation in colour may be more consistent and also applies to females, this form has been recognized 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 from Switzerland as early as 22 May and one from Yugoslavia as late as September. In Europe the usual host seems 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 van Burgst, 1927: 239), Lacanobia oleracea (Linnaeus) (Noctuidae) (Meyer, 1934: 231) and Deilephila porcellus (Linnaeus) (Sphingidae) (Leonardi, 1928: 87). The record from L. oleracea, at least, can probably 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 more southern parts of its range it occurs at higher altitudes (for example, altitudes of 1800 to 2100 m are recorded 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, KC LELW, 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 reviser choice 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 of Ichneumon 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 no reasons for so doing and I can find none. The description of Ichneumon mutillatus Christ fits well the male of pictus and the synonymy, proposed by Gravenhorst (1829: 383) but queried by Aubert (1978: 157), is accepted. The Christ name is unavailable because of its homonymy with /. mutillatus Gmelin, which is a replacement name for /. mutillarius Fabricius, 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. Horstmann was able to visit Zagreb in 1980 and in the course of his work on the species of Ichneumonidae described by Hensch he made the types of the Banchus species available to me. Unfortunately the specimen tentatively selected for designation as lectotype of B. zagoriensis had the date of collection 28 August 1928 and could not 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 Bauer collection (ZSBS). Because they bear no dates of collection and because the cotype labels were almost certainly added by Bauer and not Hensch (Horstmann, pers. comm.) they are only tentatively regarded as syntypes. Selection and designation of a lectotype for B. zagoriensis is best deferred until the other material in 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 0-40-0-50 times basal width of mandible. Maxillary palp as in Fig. 25. Antenna with apical segments slightly longer than 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 than their diameter. Hind femur 5-00-5-50 times as long as deep. Fore wing length 8-1-10-6 mm. Tergite 1 of gaster 1-45-1-65 times as long as broad, its dorsal profile as in Fig. 73. Gaster compressed, reaching just beyond tips of fore wings (when folded back). Tergite 3 with a crease along its anterior 0-2 to 0-3 separating later 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 with scape 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 of propodeum yellow anteriorly, black posteriorly; the yellow area varying considerably in extent, from covering 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 medially black (sometimes brownish red, especially dorsally), yellow proximally and distally. Segments of gaster each 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 basal width of mandible. Maxillary palp as in Fig. 115; segment 4 moderately widened and flattened; segment 5 about 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. Fore wing length 8-1-10-3 mm. Tergite 1 of gaster 1-45-1-70 times as long as broad. Tergite 3 with a crease along its anterior 0-3 separating laterotergite. 34 M. G. FITTON Colour: black and yellow. Antenna with scape and pedicel black dorsally, yellow ventrally. Flagellum reddish 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 of propodeum black, entirely or with a pair of antero-lateral yellow patches (rarely the yellow more extensive, very rarely posterior part of propodeum entirely yellow). Hind coxa black, often with a dorsal yellow patch. Hind femur medially black (sometimes brownish red, especially dorsally), yellow proximally and distally. Segments of gaster each black anteriorly, broadly yellow posteriorly. REMARKS. Specimens of turcator have been confused with pictus. The two species can be distinguished using the characters given in the key. The morphospecies/7/ctaj, as here defined, has variations in time of occurrence of adults, length of the scutellar 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 July and 2 October (most of the latter batch after 10 August). The length of the scutellar spine varies from very small to about 0-6 as long as the scutellum. Although there is continuous variation in the length of the spine most of the specimens with a short spine were collected in the first period and most of those with a longer spine in the second. The sex ratio of the 48 specimens collected in the first period was 1 • 18 $ : 1 cf and of the 42 in the second period 4-25 $ : Icf . Three possible biological explanations of these observations are that (1) pictus is bivoltine, with the generations varying slightly in morphology and more distinctly in sex ratio; (2) the early and late summer groups represent separate species; or (3) that one species has a partial second generation and a second species (possibly parthenogenetic) occurs in late summer. If two species were involved diligent study should have revealed variation in some morphological characters, colour, geographical distribution, or host or other biological data correlated with date of capture or length of scutellar spine. None was found. Assignment of the 94 specimens without a date of capture to two subjective classes - with short or long scutellar spine - gave groups which varied in sex ratio in the same way as the early and late summer groups (short spine, 58 specimens, 1-55 $ : 1 cf ; long spine, 36 specimens, 5 $: 1 cf). The only other relevant evidence comes from two reared specimens: two larvae of the noctuid Agrochola helvola (Linnaeus) collected near Sheffield, Great Britain (Ford) (UM) on 13 June 1969 proved to be parasitized. The parasite larvae emerged in August 1969 and spun cocoons. The resulting adults (1 $ Id") did not emerge until early May 1970. None of this evidence gives unambiguous support to any of the three suggested explanations of the observations. Even the one year life-cycle of the specimens reared in Sheffield could be explained by postulating 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 of more than one species, if only because there is no undisputed evidence of bivoltinism in any other species of Banchus. The poor quality and quantity of the available data and the possible complexity of the situation mean 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 reared specimens of pictus, from Agrochola helvola (Linnaeus) (Noctuidae) (all from Great Britain (Lyle and Ford) BMNH and UM). There are several hosts noted in the literature, but as black-and-yellow Banchus are 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 56°N 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 times basal 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, with moderate punctures, on mesopleuron separated by about their diameter. Hind femur 5-25-5-65 times as long as deep. Fore wing length 10-3-11-1 mm. Tergite 1 of gaster 1-65 times as long as broad, its dorsal profile as in Fig. 74 (male). Gaster very weakly compressed, not reaching to 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. 96. Colour: black and yellow, the margins of some yellow areas slightly reddish. Face black with yellow orbital stripes. Antenna with scape and pedicel black dorsally, yellow ventrally. Flagellum orange ventrally, 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 large yellow dorsal patch. Hind femur black, narrowly reddish yellow proximally and distally. Segments of gaster 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 of mandible. 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, arising from 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 separating laterotergite. 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 of propodeum yellow, entirely or narrowly black postero-medially. Hind coxa black with very large dorsal and ventral yellow patches. Hind femur yellow, black-marked ventrally and internally, tending to reddish distally. 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 0-45-0-60 times basal width of mandible. Maxillary palp as in Fig. 27. Antenna with apical segments just broader than long. Scutellar spine (Fig. 51) about 0-5 as long as scutellum. Mesopleuron and mesoscutum shining, only very 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 1-80-2-25 times as long as broad, its dorsal profile as in Fig. 75. Gaster strongly compressed from posterior of segment 3, reaching to tips 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. 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 and yellowish ventrally. Maxillary palp reddish yellow with distal parts of segments 4 and 5 dark. Posterior part of propodeum yellowish with a black patch medio-posteriorly, the yellow area sometimes reduced to a band 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 black and 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 of mandible. Maxillary palp as in Fig. 117; segment 4 flattened and considerably widened; segment 5 about 0-6 as long as 4. Antennal flag setae (similar to Fig. 124) relatively poorly differentiated, at about 50 degrees, very short, not flattened, 2 per segment, without a polished trough. Hind femur 6-05 times as long as deep. Fore wing length 8-6 mm. Tergite 1 of gaster 1-90 times as long as broad. Tergite 3 with a crease along its anterior 0-5 separating laterotergite. Colour: black and yellow, with some reddish, especially on legs. Antenna black dorsally, scape, pedicel and 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 of propodeum yellow, black postero-medially. Hind coxa black, with a dorsal yellow spot which is contiguous interno-laterally with a postero-ventral spot. Hind femur reddish, black ventrally and with a separate dorsal blackish area. Segments of gaster each black anteriorly, yellow posteriorly, the junction between the two 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 difficult to separate. ICHNEUMON-FLY GENUS BANCHUS 37 BIOLOGY AND HOSTS. Virtually unknown. Adults have been collected in April and May. The specimens collected 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: 27°56'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 $, 27°56'N, 85°00'E, 3030 m, 23-29.V.1967 (Can. Nepal Exped.) (BRI); 2 $, 28°00'N, 85°00'E, 21-23.V.1967 (Can. Nepal Exped.)(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 of mandible. Maxillary palp as in Fig. 118; segment 4 flattened and considerably widened; segment 5 about 0-6 as long as 4. Antennal flag setae (similar to Fig. 128) upright, long, flattened, 2 per segment, arising from a polished trough. Scutellar spine (Fig. 52) about as long as scutellum. Mesopleuron and mesoscutum coriaceous, with moderately fine punctures, on mesopleuron separated by a little more than their diameter. Hind femur 5-15-5-60 times as long as deep. Fore wing length 10-2-10-4 mm. Tergite 1 of gaster 2-10-2-22 times as long as broad, its dorsal profile as in Fig. 76. Tergite 3 with a crease along its anterior 0-4 separating laterotergite. Colour: black and reddish, with some yellow. Face entirely reddish yellow or yellow with a narrow median 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 with segments 1 , 2 and 3 reddish yellow , 4 and 5 blackish . Posterior part of propodeum reddish , entirely or with posterior margin narrowly black. Hind coxa blackish with a large (but not sharply defined) reddish dorsal patch and at least some reddish ventrally. Hind femur reddish with a blackish ventral mark. Segments of gaster 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 0-65-0-85 times basal width of mandible. Maxillary palp as in Fig. 28. Antenna with apical segments longer than broad. 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 as deep. Fore wing length 10-8-12-2 mm. Tergite 1 of gaster 1-90-2-20 times as long as broad, its dorsal profile as 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 and sternite 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 sometimes base 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 yellowish cream and brown or entirely reddish. Hind femur entirely reddish. Segments of gaster reddish with segment 1 pale cream or yellowish anteriorly, and sometimes with segments 1 to 4 each with a dark brownish transverse median band. Male. Width of lower face 0-80 times vertical length of eye. Malar space 0-70 times basal width of mandible. Maxillary palp as in Fig. 119; segment 4 moderately widened and flattened; segment 5 about as long as 4. Antennal flag setae not differentiated (antennae in poor condition and proper observation difficult). Hind femur 7-15 times as long as deep. Fore wing length 11-5 mm. Tergite 1 of gaster 1-60 times as 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 yellow ventrally. Maxillary palp entirely reddish yellow. Posterior part of propodeum reddish yellow, narrowly dark 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 brownish medially. REMARKS. The females show a considerable range of variation in colour, the specimens from Burma resembling those of tumidus, while those from the Philippines are almost entirely reddish orange. The specimens from Sumatra and Java have an intermediate coloration. Unfortunately only a single male (from Burma) 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 and June. 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 Banchus tumidus Chandra & Gupta); 2 $,Maymyo, 800m, xii. 1937 (HeinricK) (TC) (paratypes of Banchus tumidus Chandra & 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 0-65-0-75 times basal width of mandible. Maxillary palp as in Fig. 29. Antenna with apical segments longer than broad. Scutellar spine (Fig. 54) about as long as scutellum. Mesopleuron shining, only very weakly sculptured, with moderate punctures, separated by more than their diameter. Hind femur 6-80-7-05 times as long as deep. Fore wing length 10-5-11-9 mm. Tergite 1 of gaster 1-85-1-95 times as long as broad, its dorsal profile as in Fig. 78. Gaster subcylindrical, only weakly compressed posteriorly, not reaching to tips of 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, pedicel and base of flagellum yellowish ventrally. Maxillary palp entirely reddish yellow. Posterior part of propodeum 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 and segments 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 basal width 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 as long as deep. Fore wing length 10-8 mm. Tergite 1 of gaster 1-70-1-75 times as long as broad. Tergite 3 with a crease along its whole length separating laterotergite. Colour: creamy yellow and brown. Antenna dark brown dorsally, yellow ventrally. Maxillary palp reddish 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. Segments of 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 the maxillary 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 strongly bicoloured in males) . The females of tholus show a much wider range of variation in colour and the division between 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 and November. 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, 1828 m, 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, Uttar Pradesh, 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 0-50-0-55 times basal width of mandible. Maxillary palp as in Fig. 30. Antenna with apical segments longer than broad. Scutellar spine (Fig. 55) usually very small, sometimes absent or reasonably well developed. Mesopleuron and mesoscutum strongly coriaceous, with moderate to strong punctures, on mesopleuron separated by slightly less than their diameter. Hind femur 4-70-5-60 times as long as deep. Fore wing length 8-0-10-7 mm. Tergite 1 of gaster 1-45-1-90 times as long as broad, its dorsal profile as in Fig. 79. Gaster relatively weakly compressed, reaching about to tips 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. 100. Colour: mainly black and yellow, the yellow areas on the appendages tending to reddish. Face varying from yellow with a black median stripe to black with yellow orbital stripes. Antenna with scape and pedicel black dorsally, yellow ventrally. Flagellum entirely orange except that it is dark at extreme apex. Maxillary palp reddish yellow, with segments 1 and 5 wholly, 2 proximally and 4 distally dark. Posterior part of propodeum 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 yellow proximally and distally, the yellowish area sometimes extending along most of dorsal surface. Segments of gaster 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 basal width of mandible. Maxillary palp as in Fig. 121 ; segment 4 flattened and considerably widened; segment 5 about 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 times as long as deep. Fore wing length 8-7-10-0 mm. Tergite 1 of gaster 1-55-1-65 times as long as broad. Tergite 3 with a crease along its anterior 0-3 separating laterotergite. Colour: black and yellow, with some reddish. Antenna with scape and pedicel black dorsally, yellow ventrally. Flagellum orange, darkened distally. Maxillary palp with segment 1 yellowish (sometimes partly darkened), 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 band sometimes broken medially, reduced to two lateral spots or even absent. Hind coxa black, usually with a ventral 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 as shown, 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 Ikizdere rd, 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.(de Gaullecoll.) (MNHN); 1 cf , Kizilcahamam, 1000m, 26-28.V.1964 (FSA); 1 $, 1 cf , Nev§ehir, Urgiip, 4 & 6.vi.l978 (Schwarz) (AC); 1 cf, Zara Taiger (UZM). U.S.S.R.: 2 cf , Gissarskiy Khrebet, K-K chanch-ov on 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 been included in the synonymy of B. compressus (= dilatatorius) (e.g. Aubert, 1978: 152 [with the date given incorrectly 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. I consider 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 the type-locality, was examined (ZSBS). Meyer undoubtedly mistook the sex of the type-material (probably a single specimen) of obscurus because the face coloration cannot apply to any male Banchus. The description fits perfectly the female of volutatorius 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 space 0-55-0-60 times basal width of mandible. Maxillary palp as in Fig. 31. Antenna with apical segments longer than broad. Scutellar spine (Fig. 56) distinct, often almost 0-5 as long as scutellum. Mesopleuron and mesoscutum 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 1-80-2-05 times as long as broad, its dorsal profile as in Fig. 80. Gaster (Figs 7, 8) moderately compressed, not reaching to 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. 101. Colour: predominantly black, with legs largely reddish. Face black, occasionally with brownish orbital marks. 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 of propodeum 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 basal width of mandible. Maxillary palp as in Fig. 122; segment 4 considerably widened and flattened; segment 5 about 0-5 as long as 4. Antennal flag setae (Fig. 128) upright, very long, flattened and widened, 2 per segment, arising from a polished trough. Hind femur 4-85-5-40 times as long as deep. Fore wing length 8-0-9-9 mm. Tergite 1 of gaster 1-65-2-00 times as long as broad. Tergite 3 with a crease along its anterior 0-4 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, 4 distally 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 marks externo-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 and with anterior area often partly (sometimes largely) reddish. REMARKS. The male is similar, superficially, to falcatorius but can be separated readily by the number of flag setae and the lack (usually) in falcatorius of a distinct spine on the scutellum. The female has a gaster which 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 volutatorius generally has a much shorter scutellar spine. BIOLOGY AND HOSTS. Adults have been collected mainly in the period from mid- June to early August, but there are a few records as early as mid- April and as late as September. However, there is no evidence of two generations in the distribution of the records and information from rearings also indicates that volutatorius is univoltine. I have seen 58 reared specimens, many with detailed and reliable host data. The hosts (in decreasing 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; Lacanobia oleracea (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 1 rearing, Great Britain; Heliothis viriplaca (Hufnagel) (Noctuidae), 2 cf from 1 rearing, Germany; Xestia xanthographa (Denis & Schiffermuller) (Noctuidae), 1 cf from 1 rearing, Great Britain; Habrosyne pyritoides (Hufnagel) (Thyatiridae), 1 $ from 1 rearing, Germany; Opisthograptis luteolata (Linnaeus) (Geometridae), 1 $ from 1 rearing, Germany; and Zygaena ephialtes (Linnaeus) (Zygaenidae), 1 cf from 1 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 in the 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 to obtain on loan material from the Rudow collection. Rudow's methods and the poor condition of his collection were criticized during his lifetime. The collection, still neglected, was seen recently by Horstmann (pers. comm.), who made notes on its contents, including the fact that three specimens stand as B. 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 0-75-0-95 times basal width of mandible. Maxillary palp as in Fig. 32. Antenna with apical segments broader than long. 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 separated by 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 strongly compressed 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 brownish distally), scape with a yellow patch beneath. Maxillary palp dark brown or black. Posterior part of propodeum black with a yellow stripe along posterior transverse carina. Hind coxa black, usually with a yellow dorsal patch. Hind femur black, yellow at extreme base, distally, and often dorsally. Segments of gaster 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 basal width of mandible. Maxillary palp as in Fig. 123; segment 4 flattened and slightly widened; segment 5 about 0-8 as long as 4. Antennal flag setae (Fig. 126) upright, moderately long, flattened, 2 per segment, arising from 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 separating laterotergite. Colour: back and yellow. Antenna black, scape and pedicel (and sometimes flagellum segment 1 proximally) yellow ventrally, flagellum sometimes slightly brownish ventrally. Maxillary palp blackish or brownish. Posterior part of propodeum yellow anteriorly, black posteriorly, the yellow varying from a stripe along posterior transverse carina to covering almost entire area. Hind coxa black with a dorsal and often 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 in females, 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 very striking. 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 spine very small or virtually absent and it is possible that they represent a separate species. However, this is difficult to investigate without additional material. There are no specimens from peninsular Italy, Greece or Turkey in collections. BIOLOGY AND HOSTS. Unknown. Dates of collection range from early September to mid-December. There is a single male with a date in March, unfortunately without an intelligible locality but with the intriguing data'Salixcaprea'. DISTRIBUTION (Map 11, p. 43). A 'Mediterranean' species. Meyer (19276: 291) records the species from the 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 by G. 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 are obviously 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) covering European 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 compare quadrator with M. superba (Schrank) has been overlooked by all who have dealt subsequently with the European 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 the ovipositor: '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 himself compared it to Ibalia and did not refer to it in any of his subsequent works on ichneumonids. It is unlikely that any type-material will be found. The general nature of the description makes application of the name to a particular species difficult, but it is desirable to formally transfer the name to the Cynipoidea so that workers 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 syntypes examined]. Fabricius cited Banchus spinipes Panzer as a synonym of his new species viridator. I can see no reason for him not to have used Panzer's name for the species. Whatever the reasons, it complicates consideration of the name. It could be treated as an independent species, as published in synonymy, or as a replacement name (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 reinstated by Horstmann (19826: 238). However, Schaeffer's illustration (1767: pi. 61, fig. 6), to which Panzer's description and name refer, clearly shows an ichneumonid with segment 1 of the gaster strongly petiolate and 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 in their 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 various points; 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 of Applied Entomology (A) 21: 217.] Kuslitzky, W. 1979. Banchini (Hymenoptera, Ichneumonidae) of the Mongolian People's Republic. [In Russian.] 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 espece nouvelle. 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 and German.] 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 lyubitelei estestvoznaniya 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-infected larvae.] [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 Entomologische Zeitschrift 57: 52-102. Schwerdtfeger, F. 1952. Untersuchungen iiber den 'Eisernen Bestand' von Kiefernspanner (Bupalus piniarius 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.] Entomologicheskil Vestnik 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 waargenomen parasieten 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'Academie Imperiale 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 the American Entomological Society 11 (1): 1-477. Townes, H. 1970. The genera of Ichneumonidae. Part 3. Memoirs of the American Entomological Institute 13: 307 pp. Townes, H., Momoi, S. & Townes, M. 1965. A catalogue and reclassification of the eastern Palearctic Ichneumonidae. 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- Australian Ichneumonidae. 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 line represents 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 gaster of female, (a) dorsal, (b) left lateral, of (82) agathae; (83) crefeldensis; (84) dilatatorius. Scale line represents 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 line represents 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. Scale line 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 3 Banchus 3 Bena 6, 41 bipunctatus 8, 33 Blepharita 6, 19, 25, 32 brassicae 7, 41 bucephala 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 elator8,44 elpenor 6, 21 ephialtes 6, 41 Episyron 8, 44 Euceros 6 Euproctis 6, 19 Euxoa 7, 19 exclamationis 6, 22 Exetastes 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 6 groenlandicus 8, 31 gravidator 8, 44 gudrunae 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, 24 kozlovi 25 labiatus 7, 19 Lacanobia 5, 7, 32, 41 lavrovi 7, 17, 19 Leucoma 6, 19 leucospoides 8, 44 Lissonota 8, 44 luenebergensis 6, 17 luteofasciatus 7, 19 luteolata 6, 41 luteus 7, 22 lutulenta 6, 17 Lycophotia 7, 34, 41 Lymantria 25 Lymantriidae 6 Mamestra5,7, 41 mauricettae 7, 11, 14, 28 medialis 6 megacephala 6, 22 Megarhyssa 8, 44 Melitaea 6, 22 meticulosa 7, 19 monileatus 8, 30, 40 moppiti 5, 7, 12, 13, 28, 43 mutillarius 33 mutillatus 8, 33 myrtilli 5, 6, 41 Nawaia3,26 ni5 nigricans 7, 19 nigromarginatus 7, 19 nobilitator 7, 19 Noctuidae 6 Notodontidae 6 nox 7, 11,14,23,30,36 Nymphalidae 6 obscurus 8, 40 ocellata 6, 34 oleracea5,7,32, 41 ononis 5 Ophion 7, 22 Opisthograptis 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 reticulator7,24 Rhynchobanchus 3 rivularis 7, 34 robustus 8, 44 russiator 8, 33 sachalinensis 7, 19 salicis 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 Hymenoptera are 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 already been utilized successfully in curtailing the ravages caused by accidentally introduced pests. For selective control programmes to be effective, however, a sound knowledge of the biology of both the pest and its parasites is essential - and a sound taxonomic base is vital for the development of such knowledge. Ironically, considering the group's economic importance, the parasitic Hymenoptera is amongst the least studied of any group of living organisms, and taxonomic difficulties have presented major problems to many entomologists working with the Parasitica. An Introduction to the Ichneumonidae of Australia will go a long way towards rectifying this situation, being a taxonomic treatment, by genus, of the Australian ichneumonids, a comprehensive illustrated identification guide, and a summary of all available information on the group. It will also serve as an introduction to the biology and distribution of Australian ichneumonids, and provide a check-list of the described species and an index to their known hosts. It provides an important revision of ichneumonid nomenclature in order to bring the group into line with the generally accepted principles of zoological nomenclature. 1984, 413pp, 3 maps, 580 figs. Paperback. 0 565 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, Suffolk Printed in Great Britain by Henry Ling Ltd, Dorchester Bulletin of the British Museum (Natural I BRITISH MUbtUM .(NATURAL HISTORY) 30AU61985 PRESENTED GENERAL 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 four scientific 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 and ever-growing collections of the Museum, both by the scientific staff of the Museum and by specialists from elsewhere who make use of the Museum's resources. Many of the papers are works 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 several volumes may appear within a calendar year. Subscriptions may be placed for one or more of the series on either an Annual or Per Volume basis. Prices vary according to the contents of the 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 0 565 06012 0 ISSN 0524-6431 British Museum (Natural History) Cromwell Road London SW7 5BD Entomology series Vol51No2pp61-185 Issued 29 August 1985 The phytogeny, classification and evolution of parasitic wasps of the subfamily Ophioninae (Ichneumonidae) Ian D. Gauld Department of Entomology, British Museum (Natural History), Cromwell Road, London SW7 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& PRESENTED GENERAL 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 Ophioninae using both parsimony and compatibility methods of analysis. The problems of phylogenetic analysis of highly 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 supporting conflicting 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, Leptophion and Enicospilus subgroups. A scenario for the possible evolution of the subfamily is suggested. Thirty-two genera are recognized, one of which, Janzophion, is described as new. Rictophion is treated as a synonym of Euryophion, Ophionopsis is treated as a synonym of Dictyonotus, and Aulophion treated as a synonym of Stauropoctonus. The majority of genera are shown to be holophyletic, but Ophion and Enicospilus are thought to be paraphyletic, though for purposes of classificatory convenience they are retained as valid genera. Diagnostic descriptions of the genera are given and the inter-relationships of their component species are discussed. Introduction Ophionines are mostly quite large, slender, orange-brown ichneumonids that can frequently be observed flying around lights at night in virtually any part of the world. In temperate regions there are rather few species, but in the tropics there are very large numbers of taxa, and ophionines form a conspicuous component of the ichneumonid fauna. Almost all members of the 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 hosts of the vast majority of species are not known, but what records there are usually refer to species parasitizing lepidopterous larvae, most often exposed, solitary larvae, and in particular species of the families Noctuidae, Geometridae and Lymantriidae. Larvae of a wide range of other families are also attacked, but there are extremely few records of ophionines parasitizing Microlepidoptera or Rhopalocera. One Nearctic species is exceptional in that it is known to parasitize coleopterous larvae (Townes, 1971). Oviposition is usually into the host larva. The parasitoid egg is apparently free in the host's haemocoel where it hatches to produce a caudate first instar larva (Moutia & Curtois, 1952). Species attacking mature larvae undergo rapid development, but species (e.g. Enicospilus americanus) 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 temperate species seem to attack a number of different hosts in a similar niche, whilst a few species are bivoltine, with different generations attacking different hosts (Brock, 1982). A few species seem to attack a wide range of hosts (Gauld & Mitchell, 1981). The parasitoid larva completes development just prior to host-pupation, often after the host has constructed a cocoon. The ichneumonid larva spins a characteristic fibrous, ovoid cocoon which is generally dark brown with a pale equatorial band. Species may remain as mature larvae or even adults in this cocoon for 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 terrain where sweep netting and Malaise trapping yield poor results, or, as in the case of rain forest canopy, where collecting can only be achieved by cumbersome, expensive and (for fast-flying insects) unproven techniques. Unlike Lepidoptera, which are similarly easy to collect, most tropical ophionines are readily separable without recourse having to be made to time-consuming genitalic preparation. Illustrated keys are available to facilitate identification of most Old World tropical 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 to have an understanding of the phylogenetic inter-relationships of the genera, as many authors contend that biogeographic speculations are valid only when related to the evolutionary history of a group (Mackerras, 1962; Nelson & Platnick, 1981). As little has been published on the phylogeny of Ophioninae, a cladistic study of the group is necessary before much of the distributional data available can be interpreted in an evolutionary manner. The aim of the present study is to investigate the phylogeny of the Ophioninae using a variety of modern cladistic techniques, and to relate the results to what is known about the distribution and biology of the group in general. Terminology The morphological terminology in this work follows that proposed by Richards (1956) and interpreted for the Ophioninae by Gauld & Mitchell (1978; 1981). Specialist terms, relevant to only some ophionines, are defined in these works. Family-group names used conform with the directives of the International Code of Zoological Nomenclature and with various Opinions of the International Commission. Some recent authors (e.g. Townes, 1969; 1971) have chosen to disregard certain of these opinions and have not followed the Code when forming family-group names (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 Australian National Insect Collection (ANIC), Canberra, the Bernice P. Bishop Museum (BPBM), Hawaii, the Canadian National Collection (CNC), Ottawa, the Gupta Collection (GC), the Musee Royal de 1'Afrique Centrale (MRAC), Tervuren, the Museum National d'Histoire Naturelle (MNHN), Paris, the Taiwan Agricultural Research Institute (TARI), Taichung, the Townes Collection (TC), Ann Arbor, the United States National Collection (USNM), Washington and the Zoological Institute (ZI), Leningrad. The types of virtually all Old World species have been examined and exhaustive lists of material examined are contained in recent revisionary 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 only phylogenetic methods of data analysis have been used. These methods are based on ideas initially expounded by Hennig (1966). Although Hennig made a major contribution to system- atic philosophy, he greatly underestimated the difficulty that would occur in phylogenetic analyses due to morphologically undetectable evolutionary parallelism and character-state reversal (i.e. homoplasy). Such events result in there being incompatible character sets in taxonomic 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 resolve conflicting character sets - parsimony methods, which seek to minimize the number of character-state transformations, and compatibility methods, which seek to find a series of nesting 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 very similar results (e.g. Gauld, 1983), but as the frequency of homoplasy increases so the methods usually give increasingly different results. Parsimony methods have gained wide acceptance in the literature, with the shortest rooted Wagner tree often being uncritically presented as the best phylogenetic hypothesis. Simul- taneously, many authors have criticized compatibility analysis either for producing numerous apparent 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) (but see also Rohlf etal., 1983). However, Felsenstein (1981; 1982) stated that both methods can be justified as maximum likelihood methods, but under somewhat different circumstances. If homoplasy is expected to be scattered at random throughout all characters then a parsimony method is favoured, but if homoplasy is expected to be concentrated in certain characters then compatibility is supported. Felsenstein (op. cit.) continues by noting that both methods require both homoplasy to be rare and characters to have a low rate of change. In the case of Ophioninae, homoplasy demonstrably is not rare, suggesting results obtained by either method should be viewed circumspectly. The additional assumption necessary for parsimony methods to approach maximum likelihood, that is that homoplasy is randomly scattered across all charac- ters, contradicts the consensus of opinion of most practising entomological taxonomists. Both parsimony 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 four routines for dealing with non-polymorphic, discrete state data: (a) the Camin-Sokal parsimony method; (b) the Dollo parsimony method; (c) the Wagner parsimony method and (d) a mixed method allowing a, b or c to be specified for each character. The Camin-Sokal and Dollo methods were not used in the present study. Neither gives as short a tree as that obtained by the Wagner 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 a priori assumptions these methods necessitate. Indeed there is some biological evidence to suggest that these extra assumptions are unwarranted for ichneumonids. Dollo parsimony may justifiably be invoked for treating complex structures (such as the vertebrate eye), but virtually all characters used in the present study involve small changes in simple structures. Very often the derived state involves reduction of a structure, and observation of other ichneumonid taxa suggests that certain of these apomorphies (such as reduction of extent of occipital carina) have been developed in parallel in numerous different evolutionary lineages. Even when the apomorphic state is the development of a novel structure (such as an alar sclerite), the evidence strongly suggests (as the feature occurs in a few otherwise specialized species of several different genera) parallel development in closely related lineages. Reversal is more difficult to demonstrate, but clearly it does occur. For example, the alar sclerites 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 most Enicospilus species, has been redeveloped in some members of the otherwise highly specialized E. signativentris species-complex (Gauld & Mitchell, 1981). Because of the objections to the Camin-Sokal and Dollo methods the option employed in this study 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 to introduce 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 'averaging procedure' (see below), generally all species of a particular genus were associated. As ophionine genera are polythetic, compatibility methods invariably exclude certain taxa (see Davies & Boratyriski, 1979). Despite its widespread acceptance there are serious flaws in parsimony methodology when applied 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 the present work.) One, and perhaps the most obvious methodological flaw, is that because the method considers each character as an identical piece of binary information, and is attempting only to minimize transformation, a large set of coincidental 'bad' characters will be favoured at the expense of even a very slightly smaller set of 'good' characters. This is best illustrated in the case 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 taxa in the Wagner analysis are composed of characters that are likely to be highly homoplastic in other 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 no confidence in them as indicators of phylogenetic affinity, it was not found to be possible to translate this subjective bias into an objective taxonomic weighting scheme. Without an objective weighting scheme the averaging procedure adopted by the Wagner method has the effect of reducing the number of transformation steps of many 'weak' characters at the expense of 'moderately good' ones (Strauch, 1984). Although some hybrid parsimony/compatibility trees produced in this study were longer than the 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 cladograms were supported by more characters with minimal or nearly minimal homoplasy. This can best be illustrated by reference to two alternative cladograms produced during this study, the most parsimonious 235-step one and the favoured 239-step one (Gauld, unpubl. PhD). The numbers of 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 equally parsimonious 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 are prone to 'hopping', i.e. moving from one lineage to another in different reconstructions. Taxon 906 was found to be such an OTU - its final position in a cladogram depended solely upon the position in which it was entered into the analysis. This was presumed to operate in the following manner: 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 is so pronounced that once the taxon has linked to one or other of the alternatives (the first encountered) it does not share enough derived features with any other taxon to disassociate itself. Thus if the first three taxa entered are 905 (Simophion), 906 and 940 (a more distantly related taxon), 906 remains associated with 905, whilst if the first three taxa entered are 935 (one of 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' for ordering 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 any given 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 the shortest 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 of characters with N transform- ation steps 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 most parsimonious 242-step cladogram and a favoured 249-step cladogram (From Gauld, unpubl. PhD thesis). It is noteworthy that the latter, although 7 steps longer, is supported by more minimally homoplastic characters. shorter and yet totally different arrangement does not exist, a rather disconcerting observation if one 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 in this study. A second program (written by Dr G. Underwood) was developed by Underwood and Gauld 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 characters on a null hypothesis of random distribution of states of both. This value was then used to calculate for each character a coefficient of character-state randomness (herein called the O/E value). Typically the output of this program, LEQU.BAS, consists of a list of characters with their 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). A facility has been incorporated so that multistate characters which have been coded in binary fashion and numbered as decimal increments (e.g. 1-1, 1-2), and cannot logically fail, are not compared. Included also in LEQU.BAS is a labelling feature. This is based on an idea given by Guise et al. (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 four combinations are found in the data) on account of a single occurrence of one combination, the taxon having this combination is labelled. The results of this operation are printed in the form of a table (e.g. Appendix 4). The figure in brackets adjacent to the taxon is the total number of labels the taxon has received. The taxa most frequently labelled will be the ones with the most discordant character sets. A particularly high label score for a species for any one character strongly suggests homoplasy for that character with respect to other taxa in the set. Some care is needed in the interpretation of results as characters with very unequal character-state distri- bution (either only two 0 or two 1 scores) often have high label values. These are easily recognized as equally high scores are given for the two taxa with the minority states. Even these values can sometimes be seen to be unusually high (i.e. suggestive of homoplasy) if the character label 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 often far too small to allow full resolution of data (Felsenstein, 1982) and (2) the data set may yield a number of more or less equally large cliques (Kluge, 1976). LEQU.BAS is of some value in overcoming both these problems. As the largest compatible character set is generally very small, it is useful to have a way of ordering characters. Homoplastic characters that would be eliminated from any compatible set are not all equally bad; some are far more discordant than others. For example, allowance of just one incidence of parallelism or reversal may be enough to make certain characters compatible with a cladogram, whereas others will necessitate the invocation of multiple incidences of homoplasy 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 requiring one extra transformation to fit the specified cladogram, 15 and 16 requiring two extra transformations and 17 and 18 needing three extra. Analysing these data with LEQU.BAS gave the results presented in Table 2. It can be seen that the six homoplastic characters are ranked in order 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, is highly 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 independent characters. 0 = 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 failures and the final column in each case the ratio of observed over expected failures. Characters with a low O/E value are considered to be good characters, and are ranked in the final line higher than those with high O/E scores. Character number: failures observed expected O/E ratio 1 3 8.68 0.35 .2 3 8.68 0 35 .3 5 12.21 0.41 4 5 12.21 0.41 1 5 12.21 0 41 £ 5 12.21 0.41 7 2 8.68 0.23 a 2 8.68 0 23 2 4 12.21 0.33 12 4 12.21 0.33 jj_ 3 8.68 0 35 I?. 3 8.68 0.35 J_3 7 12.21 0.57 i* 8 13.09 0 61 J_5 10 12.21 0.82 Ii 13 13.09 0.99 \j_ 16 13.09 1 22 IS 16 12.21 1.31 Grand totals- f n lures observed expected O/E ratio 57 100.63 0.57 Ranking ratios 7 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 required to make a character 'fit' a favoured cladogram and that character's O/E value, the O/E value also varies with information value of the character, i.e. the number of 1 scores in relation to 0 scores. For example, in data set 1, character 14 may be modified so it has the following distribution of 1 states in taxa A-H, 11100000, 11001000 or 10001001. These require one, two and three transformations, respectively, and have O/E values of 041, 0-90 and 1-24. However, if the number of transformations necessary is kept constant (at say two) and the number of derived states 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 of 1-16-1-35 was found for a character requiring four transformations. Thus a homoplastic character with a high proportion of derived states supporting various subgroups in the definitive cladogram 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 equally large cliques. From his data set of 139 binary characters he obtained six almost equally large cliques of 80 or 81 characters. Kluge remarked that not only is there no reason to prefer one to any other of the sets, but also pointed out that the most primitive species in each of the cliques did not correspond to the most primitive as assessed on the basis of the best documented characters. To consider Kluge's objections necessitates re-examining his data. Of his six largest cliques, it is apparent that no less than 76 characters are common to all sets. The six dendrograms supported by these cliques are essentially similar, differing mostly in the relative order of the most 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 is the total number of times that taxon is solely responsible for a character failing the LeQuesne test. For further 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 - 2 222-------4 c 4 < . A - _ _ -22----1- - - 2 .' J - _ _ _ 2 2 - - - - 1 2 4 3 - - - - - - 1 1 - - 5 1 1 1 - - - -----2233-1 1 4 2 6 . -----1133-- 2 4 2 PHYLOGENY OF THE OPHIONINAE 69 this agrees with Kluge's preferred phylogenetic arrangement. Most of the differences between cliques (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 is inversely proportional to the information value of the character. Therefore, one criterion of which 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 the similar-sized cliques comprise equally poorly informative characters. In such cases the average O/E value for a clique is a useful indicator of the degree to which the clique is consistent with slightly homoplastic (and hence excluded) characters which are not normally considered in a compatibility analysis. For example, consider the hypothetical data set 2 (Tables 4, 5). There are two equally large, equally informative cliques: A (10) [1, 2, 3, 4, 5, 6, 7, 8, 9, 10] Average O/E value = 0-308 B (10) [1, 2, 3, 4, 5, 6, 7, 8, 9, 14] Average O/E value = 0-323 A DB = (1,2, 3, 4, 5, 6, 7, 8, 9} Table 4 Character state matrix for hypothetical data set 2. Conventions follow table 1. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 J 1 1 0 0 1 1 0 0 0 1 1 1 1 1 K 1 1 0 0 1 1 0 0 0 1 1 1 1 0 L 1 1 0 0 1 1 0 0 0 0 0 1 1 0 M 1 1 0 0 1 1 0 0 0 0 0 0 0 1 N 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 0 0 1 1 0 0 1 0 0 0 P 0 0 1 1 0 0 1 1 1 0 0 1 0 0 Q 0 0 1 1 0 0 1 1 1 0 0 0 1 0 Table 5 Results of LeQuesne test on hypothetical data set 2. Conventions follow table 2. Character number: failures observed expected O/E ratio li 3 9, ,86 0, ,30 2 : 3 9, ,86 0, .30 4 : 3 9, ,86 0, ,30 _5 : 3 10 ,63 0, ,28 I : 3 9, ,86 0, ,30 1 ; 3 9, ,86 0, ,30 12 : 1 6, ,89 0, ,15 ii : 11 9. ,86 1, ,12 13 : 12 10, ,63 1, 13 14 : 4 6, ,89 0, ,58 _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 + K whereas 14 supports J + M. No other character that is compatible offers support for one or other of 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 reinforces the J + M suggestion. This reinforcement can be detected by differences in the values of the O/E ratio, as character 10 scores 0-15 whereas character 14 scores 0-58. This difference in O/E value is responsible for the difference in average O/E values of the two cliques A and B, and it thus seems reasonable to prefer the clique with the lowest average O/E value. This method offers a simple objective procedure for selecting one of a number of large cliques as 'most favourable' , as it 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 were similar to conventional taxonomic assessment. Frequently, characters considered important in generic classification by traditional taxonomists were found to be the intersect of the largest cliques. 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 derived character matrix (Appendix 3). The rigidity of these groups was tested using Underwood's FOURS program (Underwood, 1982) and then further analyses using the compatibility and parsimony methods outlined above. Resulting cladograms were compared and differences subjectively evaluated by appraisal of characters supporting different options. The various groups were gradually related to each other until the majority of taxa had been placed. The remaining unplaced taxa all exhibited confusing affinities. To facilitate their placement a number of hypothetical taxonomic units were proposed to represent the various genera already placed in the analysis. The data set comprising hitherto unplaced taxa and HTUs was analysed using parsimony and compatibility techniques, and the resulting cladograms subjectively evaluated. An apologia for subjectivity in cladistic analysis In recent years there has been a great deal of intellectual activity devoted to making taxonomy more 'scientifically respectable' by attempting to remove subjectivity. Phenetic methods were the first to be claimed as objective taxonomic procedures (Sokal & Sneath, 1963), but these methods were essentially a movement away from evolutionary taxonomy (Nelson & Platnick, 1981) and have largely been eclipsed by the development of numerical cladistics. Despite the seeming objectivity of such numerical methods, they are really only objective ways of analysing largely subjective data, as character selection is a highly subjective procedure (Pratt, 1972). Furthermore, the claim by many cladists that their favoured numerical methods (parsimony analyses) are scientific (in the Popperian sense) is flawed by the assumption that parsimony per se is a criterion for formulation of a scientific hypothesis (i.e. the least falsified hypothesis is accepted - Gaffney, 1979), as is eloquently shown by Panchen (1982). A phylogenetic hypothesis derived from a cladistic analysis rests on the supposition (or hypothesis) that a particular derived character-state is a genuine synapomorphy which thus indicates commonality of descent of two or more taxa. In a highly homoplastic data set, for every character that is accepted 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 more likely to occur among cladistically reliable characters than among cladistically unreliable characters (i.e. random variables). This is true only if cladistically unreliable characters vary randomly and is not a corollary of the generally accepted statement that random variables are cladistically unreliable characters. Clearly it is untrue; most taxonomists can cite examples of suites of characters being size related, habitat related, associated with a particular biological strategy and so forth. How then is a taxonomist supposed to choose between competing character 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 of the 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 of mammalian teeth - Butler, 1982) are rejected as unlikely to be indicative of phylogenetic affinity. Other sets clearly comprise adaptive features associated with certain habitats and these may likewise be rejected. It is noteworthy that very many cladistic analyses seem to have been undertaken on data sets that have initially been subjectively edited (with no explanation) by discarding characters which are presumably thought to have no phylogenetic value. In this work I 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 group being 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 unusual condition is elsewhere only found in isolated genera of the Campopleginae, Tersilochinae and Anomaloninae (Townes, 1970; 1971; Gauld, 1976). In none of these taxa is the cross vein as far distal to 2m-cu as is the case in ophionines. (d) Possession of a sinuous pectinal comb on the tarsal claws. Many ichneumonids have a few scattered pectinal teeth on the tarsal claws but very few have a strongly developed sinuous comb like that found in virtually all ophionines. (e) Total loss of glymmae. These structures are present in at least some members of virtually all other 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 in many nocturnal ichneumonoids, but this suite of characters is apparently an apomorphic feature of the Ophioninae. The systematic position of the subfamily The subfamily Ophioninae belongs to the Ichneumonidae, a holophyletic group of apocrite Hymenoptera (Konigsmann, 1978). Virtually nothing has been published about the phylo- genetic relationships of ichneumonid taxa and, if a reasonable assessment is to be made of character polarity in Ophioninae, it is necessary to attempt to place the group in relation to some other subfamilies. From study of Cretaceous fossil ichneumonoids (Townes, 19730 & b; Rasnitsyn, 1983), comparison with putative ancestral groups (such as siricids and cephids) and examination of primitive extant Apocrita (e.g. Megalyridae, Stephanidae, some Braconidae), it can be deduced that 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 and has 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, and as both types occur in the Ichneumonidae (Clausen, 1940; Gauld, 19846), one expects that the most 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 have been parasites of xylophagous coleopterous larvae, as are archaic extant parasitoids such as orussids (Quinlan & Gauld, 1981; Middlekauf, 1983). These data suggest that the most primitive extant ichneumonids are to be found amongst the Pimplinae, Labeninae and perhaps also Tryphoninae. This hypothesis is supported by studies of larval morphology (Short, 1978) and is concomitant with the informal higher classificatory scheme outlined by Townes (1969). Compared with the Pimplinae and Labeninae, the Ophioninae appears to be a rather specialized subfamily. It seems to belong to a large, holophyletic group of taxa that includes the subfamilies 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 prelabial sclerite (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 less centrally 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 and sternite intimately fused and is lengthened anteriorly so that the spiracles are nearer to the posterior 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 fringed clypeus ; 2 , fusion of Rs with M in central part of fore wing ; 3 , development of sclerotized bridge between tibial 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-m in fore wing ; 9 , enlargement of pterostigma ; 10 , reduction of length of Rs in hind wing ; 1 1 , development of sinuous apex to ovipositor; 12, possession of expanded larval labial sclerite; 13, development of tubular 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 1 intimately associated; 17, possession of a dorsal subapical notch on ovipositor; 18, possession of a similarly modified female reproductive tract (see Pampel, 1913); 19, endoparasitic larva with a Y-shaped prelabial 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 and strongly shortened vein Rs in the hind wing. In addition both these taxa share a number of developmental 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, the possession 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 (primitive tersilochines have a more 'cremastine-like' petiole). Most ophionines and campoplegines possess an apomorphic short, straight ovipositor; even when it is long it is very similar in the two groups, being robust and up-curved. For the purposes of this study the Ophioninae is considered to be the sister-group of the Campopleginae, and the Campopleginae + Ophioninae is treated as the sister-group of the Cremastinae + Tersilochinae (Fig. 2). Polarity determination and character coding The polarity of the majority of characters has largely been determined by the method of out-group comparison (Watrous & Wheeler, 1981) using as out-groups the taxa mentioned above. Unfortunately this method does not work for all characters, usually because both states occur in both the group under study and the out-groups. In these cases the polarity assignment is based on unsatisfactory criteria such as common equals primitive, but this is stated in the discussion. In many cases in the Ichneumonidae, character-states are progressive steps in the loss of ancestral features such as the occipital carina, propodeal carinae or segments of the palp. Such characters can be arranged in simple transformation series, from plesiomorphic to the most derived state. For example, -for maxillary palps - 5-segments to 4-segments to 3-segments. Such transformation series may simply be scored in binary form as two characters - 8-1 8-2 maxillary palp 5-segmented 0 0 maxillary palp 4-segmented 1 0 maxillary palp 3-segmented 1 1 where 0 = plesiomorphic and 1 = apomorphic. Hence 8-1 represents reduction from 5 to 4 segments and 8-2 further reduction to 3. Such characters are obviously not independent for it is impossible to have a 0,1 coding as, if extreme reduction is observed, it is assumed that an intermediate stage has been passed through. In other cases a particular structure may have been modified in one of several ways. For example, 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 lower tooth is fairly constant in size, the upper may either be lengthened or shortened. Such a bifurcate series can be scored as two binary characters - 4-1 4-2 mandible equally bidentate 0 0 upper tooth enlarged 1 0 upper tooth reduced 0 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 the characters dealt with below have been treated in similar fashion. From the very large number of characters exhibited by the Ophioninae a number have been selected which show the greatest range of variation between the different genera. Included are all characters previously considered to be diagnostic of genera, and a number of additional features 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 a species-complex, were excluded from this analysis. These characters, such as the shape of the alar 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 of variation within the species of any single genus and, in most cases, this range of variation is repeated in many genera. Such characters were considered to be unlikely to make significant contribution to resolving the phylogeny of the genera. Characters used in the study In the following section considerable space has been allocated to explaining the rationale involved in character scoring and polarity determination. Although this practice is not usual in all cladistic studies (characters are frequently relegated to an appendix) it is considered to be important in the present work. It is upon these scores that the results of any numerical analysis depend; 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 centrally interrupted (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 tilted almost 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 equally bidentate 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 the mandible aligned in the same plane as the mandibular axis (Figs 38, 39). Many ophionines have the mandible twisted from 5-50° so that the lower tooth is directed forwards (Fig. 40). In a few species the mandible is exceptional in being twisted more than 70° so that, when closed, the lower tooth occludes the 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 is not clearly understood, but it is suggested that it is important in facilitating egress from certain types of cocoon. It is noteworthy that not only do all species of the huge and successful genus Enicospilus have such mandibles, but the totally unrelated tryphonine genus Netelia (which is also nocturnal and has a similar 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 a group of ichneumonids. For example, in the pimplines it is usually bilobate, in the anomalonines convex with a central tooth, whilst in the ichneumonines it is usually flat and truncated. The weakly convex 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 strongly concave (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 the subfamily and in the closely related out-groups was assumed to be plesiomorphic. This is the possession 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 actual margin is sharp (0,1) (Fig. 38). 8-1, 8-2 Number of maxillary palp segments. The plesiomorphic condition for the Hymenoptera is 6-segmented palps, but in the Ichneumonidae the number is reduced to 5. Fusion may occur between the distal two or three segments leading to a reduction in apparent segment number to 4 or 3. This character 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 is considered to be plesiomorphic. In a few ophionines the central segments are specialized, being globose. 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 so that the galea projects below the mandibles. This apomorphic condition is found in species inhabiting dry 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 similar function 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 as apomorphies of the entire Ophioninae. They include possession of greatly enlarged ocelli. The majority of species in the out-groups, being diurnally active, lack this specialization. Consequently the possession of enlarged ocelli must, within the Ophioninae, be regarded as a plesiomorphy. In a few areas where competition from other ichneumonids is low, e.g., remote islands, deserts, tops of high mountains, several ophionines have adopted a diurnal habit. This is presumed to be a secondary feature as several species retain some nocturnal features. However, these diurnal species have small ocelli, and in these cases small ocelli are considered apomorphic. Polarity determination of this character 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 large number of more or less identical segments (flagellomeres). In the majority of ophionines there are between 45 and 65 such segments, and the flagellum is about as long as the fore wing. This is considered, 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 and transverse (scored 1 ,0) or the number may be greatly increased (to between 75 and 95) so the flagellum is 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 found in species inhabiting humid areas, particularly rain forest canopies. Similarly shortened or elongated antennae 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 related out-groups, the spiracular sclerite is exposed and is clearly visible near the upper hind corner of the pronotum (Fig. 42). This is considered to be plesiomorphic. A large number of ophionines have the upper corner of the pronotum somewhat broadened and notched to partially occlude the spiracular sclerite (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 the anterior margin, and at their most extreme, reach the scuto-scutellar groove. The presence of these grooves seems to be a plesiomorphic feature (0,0), as they are well developed in many sawflies and are visible in many fossil apocritans. However, notauli are not present in all members of the out-groups so some doubt remains about the correct polarity of this character. The majority of ophionines only have vestigial notauli impressed at the extreme anterior margin of the mesoscutum (1,0), whilst in several they are absent entirely (1,1). 17 Pronotal crest. The plesiomorphic condition for ichneumonids is possession of a more or less flat pronotum dorsally. In several ophionines a crest is present to protect the neck region, possibly against attacks by asilids which habitually kill Hymenoptera by piercing the cervical region. The presence of this crest is regarded as an apomorphy. 18-1, 18-2 Mesopleural furrow. The plesiomorphic condition of the ichneumonid mesopleuron is for it to be 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 a diagonal groove from the pit to just below the subalar prominence (1 ,0) or possession of a groove from the 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 the lateral portion of this carina is lost (1). 20-1, 20-2 Scutellar carinae. These carinae are of sporadic occurrence throughout the Ichneumonidae but in many of the more primitive groups such as Pimplinae they are not developed. In the majority of species in the out-groups these carinae are absent, suggesting that their absence in ophionines may be a plesiomorphic feature. When present these carinae may be short, reaching to or not quite to the centre (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 most campoplegines and cremastines, is unspecialized. In a few ophionines a small lateral tooth is discernible (1,0), whilst in some taxa this tooth is apparently enlarged to form a protuberance that extends 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 the metanotum, is, in ichneumonids, characteristically depressed to form a transverse groove which is often somewhat broadened medially behind the postscutellum. This presumably plesiomorphic state is found widely in Ophioninae and in virtually all members of the out-groups. Two modifications from this 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 propodeal spiracle is oval or even subcircular, with the longest axis 4 or less times as long as the shortest. A few ophionines have very large, very elongate spiracles which are 8 or more times as long as broad. This is considered to be an apomorphic development. 24-1, 24-2 Anterior transverse carina of the propodeum. The presence of this carina, like the other propodeal 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 the dorsum of the propodeum more or less flat. In a few ophionines a deep longitudinal impression is present 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 in campoplegines and cremastines and thus a complete carina is here considered plesiomorphic for the Ophioninae. 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 so only lateral traces remain (1,1,0). In a few species reduction seems to have occurred by loss of only the central 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 same reasons. 29-7, 29-2 The thyridium. This is a moderately large indentation found on the second gastral tergite. In its plesiomorphic condition it is close to the anterior margin of the tergite (0,0), but in a number of ophionines it can be seen to have been displaced posteriorly, leaving a scar between itself and the tergal 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 specialized taxa 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 is considered 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 the spiracles 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 apomorphic state (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 shows reduction. 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 long ovipositors 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 is considered to be an apomorphic development (1). 36-7, 36-2, 36-3 Position oflm-cu in relation to Cula in the fore wing. In most ichneumonids these two veins are quite widely separated and are often at least as far apart as 0-75 of the length of Culb. With some reservation 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 the length of the first subdiscal cell (measured along Cui). In a number of ophionines the second discal cell is unusually short, being less than the length of the first subdiscal cell. This is considered to be an apomorphic 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 ichneumonids though present in fossil forms (Townes, 19736)). The retention of this stub is tentatively considered to be a plesiomorphic feature, but I have some doubts about the polarity assignment for similar reasons to 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 bent basally (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). In some ophionines the vein is markedly sinuous just before the centre and this (Fig. 6) is considered to be 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 a few ophionines the confluence of these veins is more proximal on the fore margin of the wing, away from the tip, so the marginal cell is correspondingly shorter (Fig. 4). This is considered to be an apomorphic feature (1). 42-1, 42-2 Shape of pterostigma. In most ichneumonids the pterostigma is quite broadly triangular and this plesiomorphic 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) (Figs 4, 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 this character as all stages from C«i close to M to Cui close to L4 can be found in members of the out-groups. Furthermore, in most Campopleginae this vein is lost. However, in some of the more primitive ichneumonids the usual condition is for this vein to be closer to M than to L4 . This is tentatively assumed to be the plesiomorphic condition (0,0). The derived states, treated as a transformation 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 fore margin of the wing. These two derivations are treated as a transformation series and coded 1,0 and 1,1 respectively. 45-1, 45-2 Glabrous area in fore wing. All primitive ichneumonids, most taxa in the out-groups, and virtually all lower Hymenoptera have the membrane of the fore wing uniformly hirsute close to vein Rs+2r. This is treated as the plesiomorphic condition (0,0). In most ophionines a small glabrous area is present in the discosubmarginal cell near where the pterostigma emits Rs+2r. This is considered apomorphic (1,0). A further specialization is for this area to be very extensive, reaching at least 0-3 of the 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 in different positions). Elsewhere in the Hymenoptera alar sclerites are apparently only found in some pepsine 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 almost certainly not just \m-cu but a composite of this and other vein parts) is angled centrally so the antero-proximal side of the 2nd discal cell is angulate at about 90°. In many ophionines this angulation is present (Fig. 3) and often accompanied by a ramellus (see character 38). This is here tentatively considered 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 about the 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 longer than the abscissa of M between 2m-cu and 3r-m. In many ophionines this is also the case and it is here considered to be the plesiomorphic state (0,0). In some ophionines 3r-m is shorter, 0-75-0-50 times as long 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 Cula. 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 is specialized 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 the trochantellus (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 being cylindrical 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 much longer 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 specialized condition (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 considered plesiomorphic in context of this study. Some ophionines have only 4 or 5 distal hamuli and this reduction 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 ichneumonids the tarsal claws are fairly evenly curved (Fig. 45) and this condition, where it occurs in ophionines is considered 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 so that 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. Some are 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 in having 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 numerical weight. However, a practising taxonomist does not value characters equally. Intuitively he is more likely to consider some to be more important than others. 'Good' characters are likely to include many of the more bizarre morphological adaptations (Gauld & Mound, 1982). At the other 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, as it is obvious that certain structures are frequently lost in parallel in many distantly related lineages (e.g. the occipital carina has been lost in some taxa independently in virtually all ichneumonid subfamilies), one suspects that an absence in two closely related taxa could also be the result of parallel reduction. (b) There may be some doubt concerning polarity. This is particularly true in cases where one postulates that a particular state is an apomorphy for the subfamily, but advocates reversal in a few exceptional taxa (e.g. character 12, ocellar size). (c) Even if the apomorphic condition involves the development of a particular structure, the character may be judged to be unstable, as examination of out-groups suggests it has been derived independently in numerous evolutionary lineages, or is correlated with a particular size range or extreme habitat occupied by a particular species. For example, very large ichneu- monids often have similarly coarse reticulate propodeal sculpture, whilst eremic species have short, 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 which involve an apomorphic 'loss' are listed below with an asterisk(*), whilst those in which the polarity assignment is suspect are denoted by a prime mark('). This convention is used extensively in the cladograms presented in the following sections. For comparative purposes a list 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 became obvious that there exists a large number of equally or almost equally (less than 1% difference in overall length) short trees, many of which exhibit quite different topologies. As there is no basis on which to prefer one of these, and there is no way of knowing whether all the shortest trees have been computed, Wagner analysis was initially rejected in favour of a more detailed study of characters and taxa using the programs LEQU.BAS and FOURS. From these it was possible to gain some indication of the 'shape' of the primary data set. Tentative groups of taxa recognized initially 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 is represented by 0 — . Character number: incompatibilities observed expected 0/E ratio 1.1 71 72.6 0.98 1.2 54 57.1 0.95 2_ 44 73.2 0.60 1 44 71.3 0.62 4.1 53 57.5 0.92 4.2 53 42.4 1.25 5.1 64 76.5 0.84 5.2 46 61.6 0.75 6.1 60 65.4 0.92 6.2 31 65.4 0.47 7.1 63 75.2 0.84 7.2 60 80.7 0.74 8.1 11 28.5 0.39 8.2 0 - - 9_ 55 66.1 0.83 1° 34 28.5 1.19 _U 0 - - li 68 74.8 0.91 11 52 66.1 0.79 14.1 65 73.8 0.88 14.2 66 79.8 0.83 15.1 73 81.2 0.90 15.2 62 75.2 0.82 16.1 73 50.8 0.45 16.2 77 77.5 0.99 17 65 71.3 0.91 18.1 69 75.2 0.92 18.2 69 72.3 0.95 11 65 66.1 0.98 20.1 78 79.8 0.98 20.2 68 80.5 0.84 21.1 52 65.9 0.79 21.2 11 28.2 0.39 22.1 65 80.9 0.80 22.2 61 76.3 0.80 il 57 73.2 0.78 24.1 72 81.1 0.89 24.2 76 80.5 0.94 25.1 15 28.1 0.53 25.2 72 75.8 0.95 26 49 57.8 0.85 27.1 77 80.7 0.95 27.2 77 80.0 0.96 27.3 12 27.5 0.44 28.1 0 - - 28.2 69 73.2 0.94 29.1 50 50.7 0.99 29.2 79 81.1 0.97 30 68 74.8 0.91 3i 69 69.0 1.00 li 67 71.3 0.94 33 69 79.7 0.87 34 0 - - 35_ 55 73.2 0.75 36.1 71 64.4 1.10 36.2 76 78.8 0.96 36.3 58 69.6 0.83 37 78 82.2 0.95 38 53 62.5 0.85 39 78 79.1 0.99 40 47 71.3 0.66 11 64 62.5 1.02 42.1 66 72.3 0.91 42.2 61 77.3 0.79 43.1 73 73.8 0.99 43.2 78 81.2 0.96 44.1 82 81.2 1.01 44.2 70 68.0 1.03 45.1 69 76.3 0.90 45.2 67 80.9 0.83 46 56 76.2 0.74 47.1 46 61.6 0.75 47.2 74 77.4 0.96 48.1 73 81.2 0.90 48.2 61 72.2 0.84 49 0 - 50 31 28.5 1.09 51.1 61 68.1 0.90 51.2 62 74.0 0.84 52.1 51 70.6 0.72 54 57 74.8 0.76 55_ 33 57.8 0.57 56 30 28.5 1.05 1Z. 74 80.8 0.92 58.1 67 76.9 0.87 58.2 27 28.1 0.96 59 78 81.7 0.95 60 19 28.5 0.67 61.1 70 80.5 0.87 61.2 72 77.3 0.93 62 75 80.3 0.93 63 74 81.9 0.90 64 5 28.5 0.18 Grand total- 2598 observed. 3007.1 expected. Overall 0/E ratio = 0.86 Ranking 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 7.2 13 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 29 29.1 43.1 16.2 31 44.1 41 44.2 56 50 36.1 10 4.2 5.2 47.1 35 54 23 20.2 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 the derived 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, it follows none of these eight can fail; they are shown as 0 — . The remaining 89 characters have two or more discordant states and therefore can potentially fail 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 was expected, so although slightly better than random (O/E value = 0-86), the results suggest a very high incidence of homoplasy. For individual characters a considerable variation in O/E ratio was observed. Characters 6-2, 84, 16-1, 21-2, 25-1, 27-3 and 64 all scored better than 0-55, whilst at the other extreme eight characters, 4-2, 10, 36-1, 41, 44-1, 44-2, 50 and 56 all scored worse than 1-00. Examination of the labels matrix (Appendix 4) showed that two taxa, 909 and 922, had been labelled 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 labelled less than 42 times. Sixteen taxa were labelled less than ten times and three, 912, 913 and 933 were not labelled at all. As would be expected, the elimination of bad characters (i.e. those with an exceptionally high O/E score) produced, in general, a progressive improvement in the O/E ratios of the surviving characters. However, the average ratio improved very slowly. With 20 informative characters remaining it was still 0-34. This clearly shows that not all the failures of the 'better' characters were due to clashes with the worst characters; the majority of characters seem to be somewhat homoplastic! Complete elimination of homoplasy (as advocated by LeQuesne, 1972) results in the removal of virtually all characters. The largest cliques contain only ten compatible characters. Several of these have only a very low information value (i.e. have only two or three derived states), but five more informative characters (2, 6-2, 15-1, 15-2 and 22-2) were found to be universal to all the larger cliques. All of these are features that have been used previously in ichneumonid higher classification (Cushman, 1947; Townes, 1971). These characters support the cladogram shown in Fig. 7. Only the group supported by character 2 corresponds with any previously recognized grouping of taxa, the Thyreodon group of Cushman (1947). 905 909 910-911 916-918 923 948 938 939 941 906 927 935 937 901 904 907 908 921 912 915 919 920 928 934 940 945 " 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 bad characters) 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 for elimination) 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 removed en bloc they would have been 4-2, 10, 36-1, 50, 56, 44-2, 41, 44-1, 31 and 16-2. By stepwise removal 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 the highest 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 1.1 63 67.7 0.93 1.2 49 53.8 0.91 2 38 68.4 0.56 I 40 66.7 0.60 4.1 47 54.4 0.86 5.1 58 71.1 0.82 5.2 43 57.9 0.74 6.1 53 61.3 0.86 6.2 26 6U3 0.42 7.1 57 69.9 0.82 7.2 56 74.6 0.75 8.1 9 26.9 0.33 9_ 51 62.0 0.82 li 62 69.8 0.89 11 48 62.0 0.77 14.1 60 68.8 0.87 14.2 59 73.8 0.80 15.1 66 75.0 0.88 15.2 56 69.9 0.80 16.1 20 47.9 0.42 16.2 71 72.0 0.99 \]_ 61 66.7 0.91 18.1 65 70.0 0.93 18.2 62 67.4 0.92 li 59 62.0 0.95 20.1 70 73.8 0.95 20.2 63 74.4 0.85 21.1 49 61.8 0.79 21.2 9 26.7 0.34 22.1 60 74.7 0.80 22.2 54 70.9 0.76 23 51 68.4 0.75 24.1 68 74.9 0.91 24.2 69 74.4 0.94 25.1 15 26.5 0.57 25.2 66 70.6 0.94 26 47 54.4 0.86 27.1 71 74.5 0.95 27.2 71 73.9 0.96 27.3 10 25.9 0.39 28.2 64 68.4 0.94 29.1 44 47.7 0.92 29.2 72 74.9 0.96 30 63 70.0 0.90 21 63 64.6 0.98 32 61 66.7 0.91 11 64 73.9 0.87 li 52 68.4 0.76 36.2 71 73.8 0.96 36.3 54 65.7 0.82 37 71 75.9 0.94 38 52 58.7 0.89 39 72 73.4 0.98 40 44 66.7 0.66 42.1 60 67.4 0.89 42.2 53 71.7 0.74 43.1 66 68.8 0.96 43.2 72 74.9 0.96 45.1 63 70.9 0.89 45.2 60 74.7 0.80 46 53 70.9 0.75 47.1 43 57.9 0.74 47.2 69 71.8 0.96 48.1 67 75.0 0.89 48.2 57 67.4 0.85 51.1 55 63.7 0.86 51.2 56 68.9 0.81 52.1 47 66.0 0.71 52.2 42 58.0 0.72 54 51 69.8 0.73 55 31 54.4 0.57 57 68 74.8 0.91 58.1 60 71.5 0.84 58.2 26 26.5 0.98 59 73 75.6 0.97 60 17 26.9 0.63 61.1 65 74.4 0.87 61.2 65 71.7 0.91 62 68 74.4 0.91 63 68 75.7 0.90 64 4 26.9 0.15 Grand total- 2179 observed, 2613.3 expected. Overall O/E ratio = 0.83 Ranking 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 7.2 35 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 26 6.1 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 the Ophioninae. Any attempt to produce a character set without homoplasy (a clique) necessitates excluding the great majority of characters, including some of the most informative ones. This is unacceptable for two reasons. (1) A large amount of information would have to be discarded leaving many unresolvable situations. As 2Nt— 2 apomorphies are necessary to fully resolve Nt taxa, a minimum of 100 compatible 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. Excluding autapomorphies 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, the resultant data set would still be extremely homoplastic. (2) It is important to realize that just because a character is homoplastic and has been derived independently 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 (the probable sister-group of the Ichneumonidae) is the fusion of the second and third gastral tergites. Approximately six of the 60,000 species of Ichneumonidae also have tergites 2 and 3 fused. This parallelism in a few specialized ichneumonids does not eliminate the character as a useful apomorphy for defining Braconidae in a phylogenetic sense, it merely reduces the value of the character as an infallible means of identifying braconids. If it were possible to consider all species of the two taxa Ichneumonidae and Braconidae, it is likely that the O/E value of this character would be very low, but if the analysis were of all Parasitica, and a few taxa from each family were selected as examples of the range of morphological variation, it is quite possible that one of the six anomalous ichneumonids would be included, giving the character a much larger O/E value, and consequently risking its exclusion. In the present study certain species of Enicospilus, Leptophion and Laticoleus were deliberately selected to show the range of morphological variation in the genus, so it is highly likely that these will contribute to the high level of homoplasy in the way outlined above. In the LeQuesne analysis eight characters were eliminated stepwise until none of the remainder scored worse than 1-00. The particularly high O/E value of these characters strongly suggests they 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 is limited 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 in the 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 unrelated Melanesian species (E. interruptus Szepligeti) and in an Indian Ophion species. Thus although the character has phylogenetic value in some cases (the definition of the unidens group) in others it has none. Presumably the apomorphic condition found in three taxa in the matrix is the result of 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 single species. Structurally it is a simple adaptation for biting through fibrous cocoons and as such it is not surprising that it has apparently arisen independently in many different evolutionary lineages. 10: elongation of maxillae. This too is a low information value character as the apomorphic condition occurs in just two taxa in the matrix, Rhynchophion flamipennis (908) and Sicophion pleuralis (922). Amongst other ophionines it is limited in occurrence to some Neotropical species of Eremotylus. All of these taxa are restricted to the drier areas of Central and South America and the structural development seems to be an adaptation to drinking from certain flowers. Similar structural modification can be observed in many other ichneumonids in similar habitats, including many Tersilochinae, Cremastinae and Anomaloninae. In arid areas of south-western Australia species of three closely related cremastine genera, Trathala, Temelucha and Pristomerus, all have identically modified maxillae. Only familiarity with the world cremastine fauna enables one to recognize this as evolutionary convergence rather than considering it indicative of a holophyletic group. These observations strongly suggest that particularly 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 may explain 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 low information content as the derived condition occurs in just two taxa in the matrix, Lepiscelus distans (909) and Barytatocephalus mocsaryi (927). Amongst other ophionines it is limited to occurring in one or two species of Ophion. In the sister-group to the Ophioninae, the Campopleginae, the derived condition apparently is useful for defining some genera (e.g. Cymodusa, see Townes, 1970) but the character has little apparent phylogenetic significance in the Ophioninae. 36-1': vein Im-cu separated from Cula by less than 0-75, but more than 0-5 of the length ofCulb. The majority of taxa have this character present in the presumed derived state. Only Lepiscelus distans (909), Stauropoctonus occipitalis (917), Riekophion emandibulator (920), Ophionopsis nigrocyaneus (921), Sicophion pleuralis (922), Rictophion nebulifer (937) and Enicospilus nephele (944) are plesiomorphic for this feature. It is not found in the plesiomorphic state in any other Stauropoctonus species, in most other Enicospilus species nor any other Riekophion species. If the 0 state were a genuine symplesiomorphic feature of these taxa the 1 state must have 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 species so scored. Alternatively, if the 1 state is a genuine synapomorphy for all 'other taxa', the 0 state must 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 the otherwise 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 is found in Simophion calvus (905), Lepiscelus distans (909), Laticoleus curvatus (915), Abancho- gastra hawaiiensis (925), Leptophion maculipennis (928), Euryophion adustus (936), Rictophion nebulifer (937) and Ophion luteus (939). Elsewhere in the subfamily all species of Ophion (ca 150 species) have Rs strongly curved. The genus Abanchogastra is monotypic, but only some species of Leptophion, Enicospilus, Euryophion and Laticoleus possess the apomorphic state as do some species of Eremotylus and Ophiogastrella. In both Leptophion and Laticoleus this feature seems to define holophyletic species-groups, but there is no evidence to support the idea that it could be a synapomorphy of all the taxa listed above. Clearly it must have been derived independently in many ophionine lineages, but although a homoplastic feature, it is interesting to note that it still is quite characteristic of certain species-groups and even genera, suggesting that 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 two Afrotropical species, Lepiscelus distans (909) and Laticoleus curvatus (915). The latter is the only 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. The character appears to have little phylogenetic value. 41: fore wing with R^ not reaching to tip. The apomorphic condition is found in Thyreodon atricolor (901), Simophion calvus (905), Orientospilus melasma (906), Ophionopsis nigro- cyaneus (921) and Sicophion pleuralis (922). It is not present in many other Thyreodon species so it is unlikely that the derived condition in 901 is synapomorphic with that of other species. The exclusion of 901 from the analysis actually resulted in a worsening of the O/E value of this character from 1-02 to 1-12, suggesting it is homoplastic in other taxa. The apomorphic state of this 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 taxa listed under 44-2 above and additionally in 14 other taxa, all from different genera. Elsewhere the 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 represent structural specializations (not 'loss' apomorphies or dubious polarity assignments). Several (4-2, 41 and 44-2) do apparently have considerable phylogenetic value at certain hierarchical levels and really only two (36-1 and 44-1) are likely to be considered by a traditional taxonomist as category 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) were cited for at least half of the eight worst characters, as having contributed to the high level of homoplasy . These taxa, also recognizable by their high label scores, were subsequently shown to be amongst the more difficult to place, and small differences in their position in the order of taxa submitted to the WAGNER program, produced large changes in the resulting cladogram. Another estimate of incidences of homoplasy in the primary data may be obtained from perusal of the labels matrix (Appendix 4). Notably high scores, indicative of homoplastic derivation of the apomorphic state in two or more taxa, can be observed for character 1-2 (taxa 909, 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 plesiomorphic condition for character 38 in taxa 941 and 943 was also highly labelled, suggesting one or other has undergone reversal. Of these 11 characters, six (4-2, 10, 50, 56, 58-2 and 60) are highly labelled for all apomorphic states (two or three in each case), suggesting these are not useful characters 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 LeQuesne matrix; one other (58-2) has a higher than average O/E score, but surprisingly one (60) has a reasonably low O/E value. Analysis using FOURS program Using the FOURS program a shared derived character matrix (Appendix 3) was computed. This shows the number of derived characters common to each pair of taxa and has values in the range of 3-37. Multiple linkage clusters drawn for taxa with 29+ shared derived characters are shown in 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 925 and 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 three endemic Hawaiian genera, Pycnophion and Abanchogastra. It is noteworthy that Enicospilus nephele (944) and Leptophion tetus (931) are not associated. Each of these taxa share less derived 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 derived characters in common. This group seems to be associated with the Enicospilus group quite closely 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. Also associated with this cluster are the reciprocal nearest neighbours 907 and 921 (Dictyonotus I Ophionopsis). When linkages of more than 22 shared derived characters are considered (Fig. 9), the majority serve to consolidate the clusters already formed. The Stauropoctonus group and Ophiogastrella are repeatedly linked to the Enicospilus cluster (both having in excess of 60% of possible linkages to the latter) as does taxon 931. Taxa 926 and 944 are less strongly but uniquely associated with this group and taxon 949 also links to this group. The Thyreodon, Euryophion and Dictyonotus/ Ophionopsis groups, together with taxon 908, form a second robust cluster with only Euryophion showing any linkage to non-group taxa. Taxon 909 is somewhat 86 I. D. GAULD •o I <0 TD O — O 0\ (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. As virtually complete linkages occur between taxa enclosed within the boxes, these linkages are not shown so as to avoid confusion. intermediate linking to both the Enicospilus and Euryophion groups but shares most derived features 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 Euryophion although its nearest neighbour is taxon 906. Taxon 927 is also intermediate. The remaining species, 920, 922, 923, 938-41, 948 and 950-51, are not closely associated with other taxa. To summarize 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 Thyreodonl Euryophion 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 intermediates between the major clusters. It is interesting to note that, excluding the last three taxa, the remainder 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 groups approximately corresponding to the Ophion (plesiomorphic group) , Thyreodon and Enicospilus groups 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 as the fixed taxa (3F option) of FOURS, and trying them against all other taxa, nothing broke them up. 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/E values (0-60, 0-84, 0-80, 0-79 and 0-87 respectively) and are shared by relatively few other taxa in the 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 is monophyletic. There are three possible arrangements of these taxa (Fig. 10). Taxa 935 and 936 have been considered to be congeneric (in Euryophion) whilst 937 is usually placed in a separate genus, Rictophion (Townes, 1971; Gauld & Mitchell, 1978). Rictophion has been separated from Euryophion 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 single informative character, 36-1 which has both a dubious polarity assignment and a very high O/E score. Both B and C also involve some poor scoring characters (19 and 44-2 both have O/E values 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 characters supporting C are all fairly high scoring (O/E values 0-76+) but do not involve either apomorphic loss or dubious polarity assignment. On balance therefore, C seems to be the preferable arrangement, suggesting that Euryophion, as currently recognized, is paraphyletic with respect to Rictophion. This conclusion is not surprising as originally Rictophion was separated from Euryophion on the basis of two autapomorphies. Rictophion ikuthana (937), the only species in the genus, can be regarded as a specialized species of Euryophion. These species, and all the remainder 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 species remained as a group when tested against any fourth taxon. The Thyreodon species share 27 apomorphies (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, 0-79 and 0 76 respectively) and are shared by relatively few other taxa. Only one, 7-1, involves dubious 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.- 901 X X' Y XYZ' XY'Z X'Y'Z' 902 903 904 1100: 12* 13 0011: 36- 2 ' 1010: 21- 1 1110: 26 58-1 1101: 37 0111: 9 The most favoured arrangement X,XYZ' is shown in Fig. 11. This involves three forward parallelisms 21-2 (in 901 and 903), 37 (in 901/2 and 904) and 36-2 (in 903 and 904) and one reversal (9 in 901). If the dubious polarity of 36-2 were to be reversed this character would fit the cladogram. It is noteworthy that some authors (e.g. Ashmead, 1900; Cushman, 1947) have placed 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 compatible characters. 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 and 58-1) have both average or lower O/E values (0-60, 047, 0-84, 0-79, 0-39, 0-80, 0-78, 0-79, 0-76 and 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) suggests characters 16-1, 36-1 and 61-1 (0 scores in taxon 921) may have undergone reversal, whilst character 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 apomorphic features (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 is elsewhere only found in the apomorphic state in taxon 908, character 7-1 is apomorphic for taxa 908, 940 and 946, character 22-2 is apomorphic for taxa 906, 908, 927 and 950, whilst character 42-2 is otherwise only apomorphic for taxa 908, 909, 916, 917 and 918. Examination of the labels matrix (Appendix 4) shows that for character 7-1, species 940 and 946 are labelled more than twice as frequently as any other taxon. Similarly for character 22-2, species 906 and 950 are fairly highly labelled. These data suggest parallel derivation of the apomorphic condition of these characters in taxa 906, 940, 946 and 950. Three other characters are found in the apomorphic condition in all except one taxon in the group . Of these ,61-1 has a high label score for taxon 921 , suggesting the apparent plesiomorphic condition in this taxon may be a reversal. Character 54 is present in the apomorphic condition in all taxa in this group except 935; elsewhere it is only found, in the data matrix, in the derived state in three taxa, 908, 941 and 950. The higher values of these last two species suggest the apomorphic 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 undergone reversal in 904. Elsewhere it is only found in the apomorphic state in taxa 908, 909, 927, 941 and 950. Taxon 909 has a high label score for this character, suggesting homoplasy. It seems reasonable to treat tentatively all these three characters as apomorphies of the group. This interpretation is certainly the most parsimonious, for in each case the alternative to single postulated reversals must be the advocacy of multiple forward parallelisms. One other taxon, 908, seems to be consistently associated with this group. Its nearest neighbours all lie within the group (Table 8), and in cluster analysis it readily associated with the group. 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 shared derived characters. The rows give the five (or more if several are equal fifth) nearest neighbours of each taxon listed in the left-hand column. The columns show the number of times a species is cited as a near neighbour. 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 groups have only one near neighbour in common (904-936). Only species in the latter group share large numbers of derived characters in common with non-group taxa (the 'others' column). It is interesting to note that taxon 907, the least autapomorphic in this group (see p. 89), is cited as a near neighbour by all other taxa. 901 902 903 904 907 921 908 935 936 937 901 902 903 904 907 921 908 935 936 937 Others + 32 30 28 30 27 — — — — 32 + 30 29 28 26 — — — — 30 30 + 29 27 25 25 ~ — — 28 29 29 + 26 — — — 26 — 30 28 27 — + 30 28 — — — 27 26 25 — 30 + 27 — — — 25 25 25 — 28 27 + 25 25 27 — — — — 24 — 25 + 30 31 (23- 905, 906, 918) — — — 26 26 — — 30 + 32 (26- 946) — — — — 24 24 27 31 32 + (25- 905; 24- 906) PHYLOGENY OF THE OPHIONINAE 91 28-2 plesiomorphic. However, this character is highly labelled for this taxon suggesting a reversal. In FOURS analyses taxon 908 repeatedly broke into the group. For example - TAXANOS.-901 921 936 908 X 1100: 21-1 30 41 X' 0011: 16-1 Y' 0101: 18-2 Y 1010: 29-2 37 Z' 0110: 9 XYZ' 1110: 13 28-2 XY'Z 1101: 6-2 12 14-1 23 X'YZ 1011: 27-1 36- 1 61-1 X'Y'Z' 0111: 16-2 31 32 44-1 TAXA NOS. - 901 907 936 908 X 1100: 204 21- 1 30 X' 0011: 31 32 Y 1010: 37 Y' 0101: 18-2 Z' 0110: 43-1 XYZ' 1110: 13 28-2 29-2 XY'Z 1101: 6-2 12 14-1 23 X'YZ 1011: 27-1 X'Y'Z' 0111: 16-1 16- 2 44-1 33 33 In both these cases exclusion of taxon 908 from the group, the XYZ' option, is not the most parsimonious solution. Accordingly, taxon 908, Rhynchophion flammipennis , is included for analysis in this group. Considering only the ten taxa of this group, 35 characters are informative for examining intra-group relationships. A LeQuesne test on this subset of data yielded the results shown in Table 9. Progressive removal of the five worst characters (59, 36-1, 27-2, 9 and 39) markedly altered the ranking of the 'best' scoring characters (i.e. those with an O/E value of 0-6+) ; the overall O/E ratio improved from 0-69 to 0-52 (Table 10). The data matrix was re-ordered with characters arranged in increasing value of O/E ratio (Tables 11, 12) before and after removal of the worst characters. Considering only the characters with an O/E value greater than 0-55 shows that there are two competing patterns. Characters 16-2, 44-1, 31 and 32 unite the Euryophion group (935-7) with the Dictyonotus group (907, 921) and 908. This arrangement is contradicted by 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 6.1: 11 22.9 0.48 6.2: 11 22.9 0.48 8.1: 7 16.2 0.42 1 : 30 26.9 1.11 \2_ : 18 27.9 0.65 11 : 26 27.9 0.93 14.1: 11 23.6 0.47 16.1: 18 26.0 0.69 16.2: 11 26.0 0.42 17_ : 12 26.9 0.45 18.2: 24 26.9 0.89 11 '• 17 23.6 0.72 20.1: 11 16.2 0.68 21.1: 18 26.4 0.68 21.2: 9 15.6 0.58 23 : 11 23.6 0.47 26 : 10 23.6 0.42 27.1: 9 15.7 0.57 27.2: 28 23.1 1.21 29.2: 22 26.9 0.82 30 : 20 23.6 0.85 li : 15 27.9 0.54 li : 15 27.9 0.54 JJ3 : 11 23.6 0.47 36.1: 21 15.7 1.34 36.2: 23 23.1 1.00 1Z. : 25 26.9 0.93 li : 17 26.9 0.63 41_ : 14 16.2 0.87 43.1: 19 26.9 0.71 44.1: 12 26.4 0.45 44.2: 9 15.6 0.58 45.1: 11 23.6 0.47 59 : 23 16.2 1.42 62 : 11 23.6 0.47 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 20.1 21.1 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 of five 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 0.30 Grand total- 158 observed, 303.4 expected. Overall 0/E ratio = 0.52 Ranking 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 12 16.1 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 these combinations and so, in this context, are uninformative. The very best characters in the original 35 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 the two contradictory groups, two, 16-2 and 33, are 'loss' apomorphies and perhaps therefore the apomorphic state is particularly likely to be subject to parallel development. Even disregarding these one is still left with a tie with three characters supporting each arrangement, so it is necessary to consider further characters. Of the characters having an initial O/E value of average or better (16-2-16-1 in Table ll)two, 12 and 21-1, suggest uniting Thyreodon with Dictyonotus if one postulates minimal reversal in Thyreodon. Character 27-1 suggests uniting Thyreodon with Euryophion and taxon 908, whilst 21-2 is an autapomorphy of taxa 907 and 921. Character 39 suggests linking a single Thyreodon species with Euryophion. Character 20-1 is apomorphic in only two taxa, 901 and 907. Considering all characters the most parsimonious arrangement appears to be to treat Euryophion as the first branch. By postulating minimal homoplasy (i.e. single reversals or parallelisms) 12 characters can be made to support this arrangement. Only three, 30, 43-1 and 44-2, have an O/E value below average, and only one, 33, is a 'loss' apomorphy. This arrangement involves the following homoplasy (and these are considered to be predictable from the Thyreodon labels matrix (Appendix 5) if, for a particular character, the score obtained by the 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 to their 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 36.1 901 902 903 904 908 907 921 937 935 936 0101 0101 0101 000 1 0 0 0 0 0 , 0 1 0 1 0 0 0 1 011010 0 0 1 011010 0 0 1 011010 0 0 1 011010 1 0 1 011011 1 0 1 011010 1 0 1 011011 1 1 0 100101 10111 0 0 0 0 00011 00011 1 0 0 1 1 1 0 0 1 1 0 1101 0 0 0 0 0 1001 1111 0101 0 1 1 0 000 11101 11111 1 1 11100 1001 000 1 1 1 1 1 1 0 PHYLOGENY OF THE OPHIONINAE 93 Table 12 The Thyreodonl Euryophion complex with the 30 'best' informative characters ranked by O/E value computed after progressive removal of five highest scoring characters (see Table 10). All characters to the left of 21- 1 have an O/E value of 0-55 or less. 901 902 903 904 908 907 921 937 935 936 26 23 1 1 1 1 27.1 1 C 1 1 1 1 0 0 1 13.1 0 C 0 C 0 C 0 C 0 1 1 C 0 C 1 1 20.1 1 1 1111 1111 00111 2 13 18.2 0 1 0 1 0 0 0 0 1 0 1 1 parallelism in 904 (predictable); 43-1, parallelism in 907 (predictable); 44-2, parallelism in 921 and reversal in 935 (not predictable). In attempting to resolve the sister-lineage of Euryophion, the Thyreodonl Dictyonotusl Rhynchophion branch, it is apparent that no character is a unique apomorphy of either Rhynchophion + 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 uniting Rhynchophion + Thyreodon. In any compatibility cladogram (e.g. Fig. 12) these three taxa must 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 as supported by the largest compatible clique. 94 I. D. GAULD To resolve these data further it was necessary to use the WAGNER option of PHYLIP. The shortest rooted tree obtained involved 83 transformation steps for the 45 characters showing both 0 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 of various 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 have been 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 squares represent uniquely derived characters; circles, apomorphic features derived independently in two lineages; 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 the cladogram shows only one of a number of competing cladograms with the same arrangement of taxa and the same number of transformation steps, but different positioning of the homoplastic characters. For example, character 30 is shown as a group apomorphy and is postulated as undergoing reversal in the stem 936/7, and in taxon 908, but it would be equally parsimonious to have suggested that 30 was derived in parallel in stems 901/4, 907/21 and taxon 935. Biologically neither 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 and 62). Six of these apomorphies (5-1, 5-2, 14-2, 21-1, 42-2 and 52-2) have both lower than average O/E values (0-84, 0-75, 0-83, 0-79, 0-79 and 0-71 respectively) and are shared by relatively few other 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 eight informative characters have high LeQuesne test failure rates (0-9+) and most are either loss characters or have dubiously assigned polarity, making the choice between arrangements difficult. The most parsimonious would obviously be 916 + 918. Currently 916 and 917 are placed together in Stauropoctonus whilst 918 is placed in a separate taxon, Aulophion (Cush- man, 1947; Townes, 1971). Aulophion has traditionally been separated from Stauropoctonus by the 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 a paraphyletic 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 that these 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 the plesiomorphic state of a number of characters usually found in the apomorphic condition in this group (e.g. 25-2, 28-2, 42-1, 51-1, 51-2 and 63). For the present this taxon is excluded from the complex and its position discussed later in the work.) Their nearest neighbours are shown in Table 13. These taxa share four apomorphic features (16-1, 25-1, 28-1 and 51-1). A further 13 characters (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 the apomorphic 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 for taxon 943 and 47-1 for taxa 944 and 945), suggesting a reversal may have occurred. The exceptions are 22-1 and 33 which are plesiomorphic for taxa 910 and 911. Both are unlabelled in the 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 large number of other ophionine taxa in the matrix. Individually therefore, they are not good characters 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 for all except 938, 939, 943, 944, 948 and 951; character 51-1 for all except 920, 922, 939-41 and 949-51; and character 51-2 for all except 920, 922, 932, 938-41 and 948-51. Ten of these characters (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 shared derived 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 as near 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 characters that unite the EnicospiluslStauropoctonus and ThyreodonlEuryophion generic complexes. Character 16- 1 could also be considered an apomorphy of both these major groups if one were to postulate reversal in 901-3 and 921. Character 22-1 is found in the apomorphic state outside the EnicospiluslStauropoctonus complex only in taxa 920, 940, 949 and 951; character 33 in taxa 901^, 907-8, 920-22 and 949 and character 63 in 905, 906 and 923. Character 33 is a loss apomorphy and has been cited above as an apomorphy of the ThyreodonlDictyonotus lineage. It is perhaps not unreasonable to postulate parallel loss in the EnicospiluslStauropoctonus, the ThyreodonlDictyonotus and the 920, 922 and 949 lineages. However, this feature, the loss of the umbo, is not apomorphic for taxa 910 and 911 included above in the Enicospilus group. It is biologically rather unlikely that the 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 by character 22-1. The apomorphic condition, a lengthened anterior part of the propodeum, is rather unlikely to have undergone reversal. It is interesting to note that character 22 has two alternative derived states, 22-1 or 22-2. The derived state 22-2 is an apomorphic feature of the ThyreodonlEuryophion complex and is only found elsewhere in taxa 906 and 927. The alternative derived state characteristic of the EnicospiluslStauropoctonus group (less Ophiogastrella) is also found in very few other taxa (see above). The plesiomorphic condition (0,0) is found in taxa 905, 910, 911, 922, 923, 938, 939, 941 and 948. The character seems to be important in defining two major sister-lineages of the subfamily. PHYLOGENY OF THE OPHIONINAE 97 Using the CLIQUE option of PHYLIP eight cliques were found with 12 or more informative characters. 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 = 0-76 Av. O/E = 0-75 Av. O/E =0-78 Av. O/E = 0-77 Av. O/E = 0-78 Av. O/E = 0-76 Av. O/E = 0-79 Av. 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) than several others. B has the lowest average O/E value with 0-75, closely followed by A and F with 0-76. The remainder have higher values. The cladograms based on the largest clique (E) and the best O/E scoring clique (B) are shown in Figs 14, 15. The cladogram produced from clique B associated 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 derived Enicospilus) cluster together, except for E. cionobius (945) which is excluded by having the plesiomorphic 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) plus Lepiscelus (909). The cladogram derived from clique E has more confusing groupings, uniting some species of Enicospilus and Laticoleus (character 17) or defining a clade containing Dicamptus 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 largest clique, E, is therefore not considered particularly useful in this case; clique B appears to be a better 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 previously considered 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 of Ophiogastrella. A LeQuesne test was undertaken on the EnicospiluslStauropoctonus data set and the O/E values are given in Table 14. Stepwise elimination of all characters scoring worse than 1-00 (a total 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 rose from thirty-fifth to eleventh position, and 62 which dropped from thirty-first place to position forty-three. Amongst the best scoring characters 42-2 and 18-2 rose from fourteenth and sixteenth positions to seventh and eighth positions (Table 15). Using the S option of FOURS the data set was reorganized with characters ranked as per O/E value after removal of the 15 most discordant 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. A number of other characters may be postulated as apomorphies of Ophiogastrella, though all apparently have been derived in parallel elsewhere. Characters 18-2 and 42-2 support the Stauropoctonus group -I- 909, whilst 21-1 supports just the former. Character 1-2 is incompatible with 21-1. Initially it had a poor O/E score (0-93 compared with 0-54 for 21-1) but progressive elimination of the poorest characters produced rapid 'clean-up' until with 15 characters eliminated it scored 0-57 compared with 0-34 for 21-1. With only the 18 best characters left in the matrix 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 reason that one would prefer the arrangement supported by the former character. Several other characters support the Stauropoctonus + 909 group but necessitate postulating parallel derivation in other places . These include 1 • 1 (parallelism in 925 and 945) , 39 (parallelism in 930) ,61-1 (parallelism in 924-26 and 942-47) . Many of these parallelisms are also suggested in the reduced labels matrix (Appendix 6), e.g. taxon 930 for character 39 acquires almost half of its total 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 Table 2. Character number: incompatibilities observed expected 0/E ratio 1.1 40 42.7 0.94 1.2 27 29.2 0.93 1 33 47.6 0.69 4.1: 39 35.8 1.09 5.1 45 50.5 0.89 5.2 41 42.3 0.97 6.1 28 19.9 1.40 7.2: 39 51.5 0.76 li 27 29.7 0.91 11 10 19.9 0.50 14.2 43 51.5 0.84 15.1: 37 42.3 0.88 15.2 39 50.0 0.78 16.2 48 40.1 1.20 17 25 29.7 0.84 18.1: 11 19.7 0.56 18.2 27 35.5 0.76 11 27 19.9 1.35 20.1 38 34.9 1.09 20.2: 46 50.1 0.92 21.1 16 29.7 0.54 22.1 5 19.9 0.25 24.1 49 48.1 1.02 24.2: 50 44.7 1.12 25.2 26 19.9 1.30 27.1 41 46.2 0.89 27.2 41 46.2 0.89 27.3: 11 19.1 0.58 29.2 34 43.2 0.79 30 22 19.9 1.10 li 19 19.9 0.95 33 : 5 19.9 0.25 35 28 47.6 0.59 36.1 33 28.6 1.15 36.2 44 44.3 0.99 36.3: 25 38.8 0.64 11 50 49.0 1.02 39 36 40.1 0.90 40 34 47.6 0.71 42.2: 27 35.8 0.75 43.1 32 29.0 1.10 43.2 50 48.3 1.03 44.1 55 48.2 1.14 44.2: 43 35.0 1.23 45.1 30 29.0 1.04 45.2 45 49.4 0.91 46 42 50.2 0.84 47.1: 15 19.5 0.77 47.2 52 50.5- 1.03 48.1 47 50.5 0.93 48.2 39 46.6 0.84 52.1: 45 48.1 0.94 52.2 40 42.3 0.95 li 25 35.8 0.70 56 28 19.9 1.40 5_7 : 35 35.8 0.98 59 48 51.7 0.93 60 17 19.9 0.85 61.1 35 50.7 0.69 61.2: 40 50.5 0.79 62 36 40.1 0.90 64 5 19.9 0.25 Grand total- 1035 observed, 1161.2 e xpected. Overall 0/E ratio = 0.89 Ranking ratio of scoring characters 22.1 33 64 13 21.1 18.1 27.3 35 36 .3 61.1 3 55 40 42.2 7.2 18. 2 47.1 15.2 29. 2 61.2 14.2 46 48.2 17 60 15.1 27.1 27.2 5.1 39 62 12 45.2 20.2 1.2 59 48.1 52. 1 1.1 52.2 31 5.2 57 36.2 24.1 37 47 .2 43.2 45.1 20.1 4.1 43.1 30 24.2 44.1 36.1 16.2 44.2 25.2 19 6.1 56 Table 15 Results of a LeQuesne test on the EnicospiluslStauropoctonus data set after progressive removal 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 1.1 26 33.2 0.78 1.2 13 23.0 0.57 3 23 36.9 0.62 4.1 27 28.3 0.96 5.1 33 38.7 0.85 5.2 29 32.8 0.89 7.2 27 39.7 0.68 li 20 23.5 0.85 11 7 15.8 0.44 14.2 32 39.7 0.81 15.1 23 32.8 0.70 15.2 27 38.3 0.70 17 18 23.5 0.76 18.1 9 15.5 0.58 18.2 13 28.0 0.46 20.1 24 27.4 0.88 20.2 32 38.4 0.83 21.1 8 23.5 0.34 22.1 1 15.8 0.06 27.1 30 35.5 0.85 27.2 30 35.5 0.85 27.3 6 14.9 0.40 29.2 25 33.7 0.74 li 14 15.8 0.89 33 1 15.8 0.06 35 20 36.9 0.54 36.3 17 31.4 0.54 37 36 37.9 0.95 39 22 31.4 0.70 40 25 36.9 0.68 42.2 13 28.3 0.46 43.1 22 23.5 0.93 45.2 34 38.7 0.88 46 32 38.7 0.83 47.1 10 15.3 0.65 47.2 38 38.8 0.98 48.1 33 38.7 0.85 48.2 29 35.9 0.81 52.1 33 37.0 0.89 52.2 30 32.8 0.91 55 20 28.3 0.71 5! 36 39.9 0.90 60 14 15.8 0.89 61.1 24 38.9 0.62 61.2 28 38.7 0.72 62 29 31.4 0.92 64 1 15.8 0.06 Grand total - 522 observed 708.4 expected Ove rail 0/E rat io = 0.74 Ranking ratio of scoring characters 22.1 33 64 21 .1 27.3 13 42.2 18.2 36.3 35 1.2 18.1 61.1 3 47.1 40 7.2 39 15 1 15.2 55 61.2 29.2 17 1.1 14.2 48.2 46 20.2 27.1 27.2 12 5.1 48 1 20.1 45.2 5 2 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 the characters 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 33 21.1 13 18.2 35 18.1 3 40 39 15.2 61.2 11 14.2 46 27 ,1 12 48.1 4b ,1 31 52.1 52 .2 43.1 4. 1 924 11001 1 0 0 1 0 0 0 1 0 1 0 0 0 1 1010 0 0 0 0 000 0 0 1 1 1 0 e 0 0 0 0 o 0 1 1 0 0 926 11001 0 0 0 1 0 0 0 1 0 1 0 0 0 1 1011 0 1 0 1 000 0 0 I 1 1 0 1 0 0 0 0 e e 0 1 0 0 926 11000 1 0 0 1 0 0 0 1 0 1 0 0 0 1 1010 0 0 0 0 0 1 0 0 1 \ 0 1 0 0 1 0 0 0 e 1 1 0 0 0 947 11000 0 0 0 0 0 0 0 1 0 1 0 0 0 1 1011 1 0 1 1 1 1 0 0 0 1 1 1 1 1 0 0 0 1 0 0 1 0 1 1 946 11000 0 0 0 1 0 0 0 1 0 1 1 0 0 1 1011 0 0 0 1 1 1 0 0 0 1 1 1 1 0 0 0 0 1 • 1 1 1 0 0 945 11000 0 0 0 0 0 0 0 1 0 1 1 1 0 1 0111 0 1 0 0 1 1 0 0 0 1 1 1 1 0 0 0 0 0 0 I 1 0 1 i 944 11000 0 0 0 0 0 0 0 1 0 0 1 0 0 1 1001 0 0 0 0 000 0 0 1 1 0 1 0 0 0 0 0 e 1 1 0 0 0 943 1 1 00-0 0 0 0 0 0 0 0 1 0 0 1 0 0 1 1011 0 0 1 0 1 1 0 0 0 1 0 1 1 1 0 0 1 0 1 1 1 0 1 e 942 11000 0 0 0 1 0 0 0 1 0 1 1 0 0 1 1011 0 0 1 1 1 1 0 0 0 1 1 1 1 0 0 1 0 0 0 1 1 1 0 0 934 11000 0 0 0 0 0 0 0 0 1 1 1 1 0 1 1001 0 0 0 1 1 1 0 0 0 0 1 1 1 0 0 0 0 1 0 1 1 0 e 0 933 11000 0 0 0 0 0 0 0 0 0 1 1 1 0 1 1001 0 0 1 0 1 1 0 0 0 0 0 1 1 0 0 0 0 1 0 1 1 0 0 1 932 11000 0 0 0 0 0 0 0 0 0 1 1 1 0 1 1001 0 0 0 1 1 0 0 0 1 0 1 1 1 0 1 0 0 0 0 1 1 1 0 0 919 11000 0 0 0 0 0 0 0 0 1 1 0 1 0 1 0001 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 0 1 1 1 1 1 1 0 1 931 11000 0 0 0 0 1 0 1 0 1 1 0 1 0 1 0001 0 0 0 0 000 0 1 0 0 1 0 0 0 0 0 1 0 1 1 0 0 1 930 11000 0 0 0 0 1 0 0 0 1 1 0 1 1 1 0101 0 0 1 0 1 0 0 0 0 0 0 1 1 0 0 0 0 0 0 1 1 1 0 0 929 11000 0 0 0 0 1 0 1 0 0 1 0 1 0 1 0101 0 0 1 0 0 0 1 1 0 0 0 1 0 0 0 0 0 0 0 1 1 1 0 0 9?8 11000 0 0 0 0 1 0 0 0 1 1 0 1 0 1 0101 0 0 1 0 0 1 0 0 0 0 0 1 1 0 0 1 0 0 0 1 1 1 0 1 915 11000 0 0 0 0 1 0 0 0 1 1 0 1 0 1 0000 1 0 1 0 0 1 1 1 0 0 1 1 1 0 0 0 1 1 0 1 1 1 0 1 914 11000 0 0 0 0 1 0 0 0 1 1 0 1 0 1 0000 0 0 0 0 1 1 1 1 0 0 1 1 1 0 0 0 1 1 1 1 1 1 0 1 913 11000 0 0 0 0 1 0 0 0 0 1 0 1 0 1 0000 1 0 1 0 0 1 1 1 0 0 1 1 1 0 0 0 0 1 0 0 1 1 0 0 912 11000 0 0 0 0 1 0 0 0 1 1 0 1 0 1 0000 0 0 1 0 0 1 1 1 0 0 1 1 1 0 0 0 0 1 0 1 1 1 0 0 909 11000 0 1 1 0 0 1 0 1 0 1 0 0 1 0 0 0 "1 1 0 1 1 0 000 0 0 0 0 0 0 0 0 0 0 1 0 1 0 0 1 1 918 11010 0 1 1 0 0 1 0 1 0 1 0 0 1 0 0011 0 1 1 0 0 0 1 1 0 1 0 0 1 1 0 0 1 0 1 1 1 1 0 0 917 11010 0 1 1 0 0 0 0 1 0 1 0 0 1 0 0011 0 1 1 0 000 0 0 1 0 1 0 1 0 0 1 0 1 1 0 1 0 1 916 11010 0 1 1 0 0 1 0 1 0 1 0 0 1 0 0011 0 1 1 0 000 0 0 1 0 0 0 1 0 0 1 1 1 1 1 1 0 1 911 00100 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0001 0 0 0 1 0 1 1 1 0 0 1 1 0 0 0 0 1 1 0 1 1 1 0 0 910 00100 0 0 0 0 0 0 0 0 0 1 0 1 0 0 0001 0 0 0 1 0 1 1 1 0 0 1 1 0 0 0 0 1 1 0 0 1 1 0 0 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 that includes 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, 13 and 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 show the plesiomorphic condition of 35. Three characters, 5-1, 61-1 and 61-2, suggest uniting the Hawaiian genera (924-26) with Enicospilus (942-47). Only one taxon, 944, is at variance for a single character, 61-2, and this has a high label value suggesting that a reversal to the plesiomorphic state may have occurred. In several places in the data set it is apparent that similar cases of a single taxon being at variance with its congeners is observable. In the case of 944, it is one of only about ten Enicospilus species out of nearly a thousand to have the plesiomorphic condition of character 61-2. Similarly 945 has the plesiomorphic state of 15-2. Both of these taxa do not appear to be primitive in any other features and as they are thus unlikely to be ancestral to all 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) and this is partially supported by character 40 although it is necessary to postulate reversal in the Hawaiian genera. Character 46 is similar. Character 7-2 is contradictory, suggesting a group comprising 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 3 suggests 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 informative characters. 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 = 0-81 Av. O/E = 0-83 Av. O/E = 0-86 Av. O/E = 0-87 A n B 0 C 0 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, is shown 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 shortest rooted tree possible. The minimum length tree (Fig. 17) necessitated 242 transformation steps. Both parsimony and compatibility analyses have certain similarities. Both taxa 909 + 916-18 formed 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) and Leptophion (928-31). However, the Wagner parsimony method usually separated the Hawaiian taxa (924-26) as a relatively primitive, discrete group but placed Ophiogastrella (910-11) well in the Enicospilus/ Dicamptus/ Leptophion/ Laticoleus complex. Compatibility methods suggested the reverse. To assess the relative merits of these alternative arrangements it is necessary to evaluate the characters upon which they are based. As mentioned above, the exclusion of taxa 910 and 911 is based on characters 15-1, 22-1 and 33 which are present in the apomorphic condition in all other taxa. The former two are striking structural modifications and whilst the latter 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 their congruence 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 Enicospilusl Stauropoctonus complex. This tree requires 242 transformation steps. especially the Laticoleusl Leptophion complex (e.g. characters 7-2, 27-1, 27-2, 20-1) which it must be assumed has been derived in parallel if Ophiogastrella is indeed the most primitive branch. The Wagner analysis united all taxa excluding the Stauropoctonus/909 complex on the basis of characters 15-1, 36-2 and 43-1. The former has been postulated as having undergone reversal in the stem 910 + 911 so cannot be considered a group apomorphy, whilst 36-2 is highly homoplastic in any favoured arrangement, and in the minimum length tree necessitates six transformations that include being developed, lost and subsequently redeveloped. In fact, this character is also present in the apomorphic condition in some Stauropoctonus (916, 918), and it is 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 additional transformation step, and would seem biologically more feasible than the gain-loss-gain scheme favoured by the Wagner analysis. The character itself, relative position of two wing veins, has not 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 worst characters suggested by the original LeQuesne test (see Table 6). The third character, 43-1, is found 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 certain members of the present group, or it is an apomorphy of the Ophioninae and the apparently plesiomorphic 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 generic keys, it is variable in most higher taxa (cf. Pimplinae in Townes, 1969), suggesting it is a highly homoplastic 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 characters which seem biologically implausible, and they also have dubious polarity assignment. Character 45-2 is postulated as having undergone reversal in the stem 910 + 911 so there remains five reasonably 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'. The remarks made above about 43-1 can also be applied to 43-2; the loss apomorphies 27-1 and 27-2 have identical state distributions in this group and should perhaps be considered as a single feature, 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 ichneumonid evolutionary lineages outside the subfamily under consideration. As a consequence it is not a character which seems to unite convincingly Laticoleus (912-5) with Ophiogastrella. It is interesting to note that there is at least some degree of correlation between the presence or absence 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 the subfamily mentioned this includes common species of Paraphylax, Amauromorpha and Apsilops found in Old World rice padi. In the Anomaloninae the two closely related genera Therion and Heteropelma differ in this feature; the former is characteristically found on dry open areas, whilst most Heteropelma species occur in more humid woodlands (Gauld, 1976). In the Ophioninae the carina is always present in those species found in humid rain forests (e.g. most Enicospilus, Leptophion and Dicamptus) but is incomplete in species favouring drier, more exposed habitats (e.g. some Australian Leptophion, most Ophion). Both Ophiogastrella species and 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 taxon 944 and also Enicospilus arduus (Gauld & Mitchell, 1978)). It would be only slightly less parsimonious to suggest that the presence of complete scutellar carinae is an apomorphy of the entire 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 uniting Ophiogastrella with other genera. The remaining character, 48-1, is a difference in relative lengths of wing veins and its apomorphic state is found scattered throughout the subfamily. To summarize therefore, the decision on where to position Ophiogastrella hinges on two character sets, one of three unusual and fairly convincing apomorphies (15 • 1 , 22- 1 and 33) versus five 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 Ophiogastrella amongst the most derived species, whilst the compatibility method favours treating it as the most primitive taxon, thus eliminating homoplasy in three characters (22-1, 33 and 15-1) as the other characters are mutually incompatible. The position of the Hawaiian genera (924-6) also presents a problem. In the parsimony dendrogram (Fig. 17) the five characters defining the stem 924 + 925 + 926 (5-1, 15-2, 36-3, 61-1 and 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 compatibility dendrogram. The parsimony method has excluded the Hawaiian genera on the basis of characters 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 to be 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. The apomorphic 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 Stauropoctonusl Aulophion and Enicospilus it is not found elsewhere in the Ophioninae. It is of very rare occurrence in other parts of the family, and where it does occur it is usually characteristic of a group of genera e.g. the Orthocentrinae (Townes, 1971) and the XanthopimplalEchthromorpha group of Pimplini (Townes, 1969). Character 15-2, a completely concealed spiracular sclerite, is found in virtually all species of Enicospilus and Dicamptus. Elsewhere it is found in the derived condition only in the Hawaiian genera, in a few species of Leptophion and in the aberrant monotypic genus Sicophion (922). The final character in this set, 36-3, an extreme reduction in the length of a wing vein, is only found in a very few taxa, notably (in the matrix) the Hawaiian genera, some Enicospilus, Euryophion and Xylophion. Elsewhere in the Ophioninae it only occurs 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 been considered above where characters 20-2, 43-2 and 48-1 have been suggested to be rather poor indicators 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 Hawaiian genus, as opposed to parallel derivation in the subgroup of two Hawaiian taxa and the major line in the larger group. The remaining characters all refer to the fore wing, in particular the antero-distal part of the discosubmarginal cell and the adjacent vein, Rs+2r. The apomorphic conditions 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. Riekophion species). Several species of Enicospilus may have one or more characters in the plesiomorphic state (e.g. the E. senescens species-group (Gauld & Mitchell, 1978)), but the trace features found in one or two species suggests this condition represents a reversal. This suite of characters does seem therefore to be good evidence for excluding the Hawaiian taxa from the Enicospilus lineage. However, the plesiomorphic condition of all three of these characters occasionally is found in some Enicospilus species inhabiting oceanic islands, and in these cases it can be seen that a reversal has occurred as related taxa still have the apomorphic states. For example, the three species E. vidus, E. ditor and E. donor form a closely interrelated group on the Galapagos Islands (Gauld & Carter, 1983) . They belong to the E. capensis species-group and probably have evolved from a migrant South American species, all of which have the apomorphic condition for characters 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 case can be demonstrated for Hawaiian Enicospilus (see Cushman, 1944). The majority of Enicospi- lus species are both synchronous and sympatric with numerous congeners. The three characters are developed in such a way as to form a distinctive, species-specific pattern (especially the exact shape of the sclerite, character 46) (see figs 384-641 in Gauld & Mitchell, 1981). In areas where few species occur, such as deserts, there is considerable variation within a species in the exact expression of these characters. Possibly their reversal on oceanic islands is facilitated by a reduction in selection pressure that favours uniformity in areas of high possible interspecific interaction. It is possible that the apparent plesiomorphic condition of these characters in the Hawaiian 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-m and M . The apomorphic condition has previously been used only to characterize species and both 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 set and again biological evidence suggests that perhaps this is incorrect and the position indicated by the compatibility method is more plausible. The parsimony analysis united all species of Enicospilus (942-47) into a single clade but Dicamptus 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 is recognized only by the specialized features it has in common with Enicospilus plus the possession of unspecialized mandibles, a plesiomorphic feature (Townes, 1971). The compati- bility method proved to be less successful at producing groups corresponding with recognized genera, probably because Enicospilus and Leptophion are polythetic taxa (Gauld & Mound, 1982). The relative positions of Leptophion, Laticoleus and Pamophion are less clear. The Wagner method suggested Leptophion and Laticoleus form separate groups, though both are charac- terized by the apomorphic condition of 35, and with one exception, 37. The compatibility method suggests Laticoleus + Leptophion form a group defined by character 35 and these two taxa, plus Pamophion, form a group defined by character 3. To conclude this section, on balance, secondary evidence suggests Ophiogastrella and the Hawaiian genera are better placed by the compatibility method whilst Enicospilus is most successfully aggregated by parsimony analysis. A 'hybrid' compromise cladogram was post- ulated that required 249 transformation steps (Fig. 18). This compromise arrangement groups various congeneric species together, thus to some extent corroborating the initial assumption that ophionine genera are not polyphyletic groups. The relationship of the component genera of the Enicospilus/ Stauropoctonus complex can be illustrated most clearly if the highly homoplastic characters are omitted, the relative positions of "O X5 o — ^O ^Q ^^ ^Q NO N 5^ 00 00 K) GO Fig. 18 Hybrid cladogram derived from subjective evaluation of parsimony and compatibility analyses of Enicospilus/ 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 the phylogenetic relationships of the genera) and single exceptional character scores, that are almost 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 with respect 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 the basis 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 the Thyreodon/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) are common to both sets, suggesting a sister group relationship. Of the apomorphies remaining for defining the Thyreodon/Euryophion lineage, three (24-1, 24-2 and 61-1) occur extensively in various species-groups and species in the Enicospilus/ Stauropoctonus complex, thus weakening their credibility as characters defining the former lineage. Character 42-2 has apparently been derived in parallel in the Stauropoctonus line, but the remaining five features (2, 7-1, 22-2, 54 and 58-1) are robust characters that strongly support the monophyly of the Thyreodonl Euryophion lineage. Of the six remaining apomorphies defining the Enicospilus/ Stauropoctonus lineage (16-1, 22-1, 33, 36-1, 45-1 and 63) two (22-1 and 33) exclude Ophiogastrella; the apomorphic state of 33 has also been derived in parallel in the Thyreodon line. The derived states 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 of the Enicospilus/ Stauropoctonus complex. A cladogram showing the putative phylogeny of these taxa 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 be considered in relation to the robust arrangement of the majority of ophionine taxa presented above. To reduce the data matrix to more manageable proportions a number of hypothetical taxonomic units (HTUs) were constructed to represent genera reasonably placed in the earlier part of the work. Scores were assigned to these HTUs by marking each character with the condition found in the majority of species of the genus (including all taxa examined in earlier taxonomic studies (Gauld, 1977; Gauld & Mitchell, 1978, 1981) but not included in the primary data matrix). This method of scoring, rather than deducing a hypothetical ancestor, was adopted because of the high incidence of homoplasy. This idiosyncracy is unlikely to alter any cladistic arrangement significantly, whereas the assumption that the group-ancestor has a particular character 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 utilized are 960 (Laticoleus), 961 (Ophiogastrella), 962 (Leptophion), 963 (Stauropoctonus), 964 (Thyreodon/ Dictyonotus) , 965 (Euryophion), 967 (Dicamptus), 968 (Enicospilus). [N.B. There is 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 960-68 are given in Table 17. The overall O/E value is strikingly high (0-90), indicating the extremely homoplastic 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 quite different topologies (Figs 21-24). A large number of other trees that were only slightly less parsimonious were also generated (Gauld, unpublished data). Compatibility analysis yielded a single 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 major complexes emerged as the most derived groups. Five taxa (938, 939, 941, 948 and 951) occupy the most primitive position in three of the dendrograms and are amongst the eight most primitive taxa 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 holophyletic group, the Ophion subgroup, characterized by having a slender Rs+2r, which is also straight basally, having an incomplete posterior transverse carina of the mesosternum, having a subapically impressed clypeus, having Im-cu usually with a ramellus, and always angled, and having the anterior area of the propodeum 'occluded'. However, the holophyly of the group is open 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, Character number : incompatibilities observ ed expected 0/E ratio 1.1: 44 45.4 0.97 1.2: 38 37.7 1.01 2_ : 15 25.6 0.59 5.1: 35 38.2 0.92 6.1: 30 38.2 0.79 7.1: 35 37.4 0.94 1 : 35 38.2 0.92 \2_ : 36 38.2 0.94 14.1: 50 45.1 1.11 15.1: 53 58.9 0.90 15.2: 30 37.5 0.80 16.2: 56 57.6 0.97 18.1: 54 60.2 0.90 18.2: 35 37.4 0.94 11 : 30 38.2 0.79 20.2: 53 58.6 0.90 22.1: 45 60.0 0.75 22.2: 49 50.1 0.98 24.1: 54 60.8 0.89 24.2: 44 50.1 0.88 25.1: 14 25.1 0.56 26 : 24 25.6 0.94 27.1: 53 58.6 0.90 27.2: 57 61.0 0.93 29.1: 37 37.5 0.99 29.2: 52 58.9 0.88 30 : 42 38.2 1.10 Ji : 45 45.9 0.98 33_ : 50 60.9 0.82 35 : 47 45.9 1.02 36.2: 59 56.5 1.05 36.3: 28 37.3 0.75 37 : 53 60.9 0.87 39 : 61 59.6 1.02 40 : 22 38.2 0.58 11 : 44 45.9 0.96 42.2: 37 37.4 0.99 43.1: 44 37.4 1.18 43.2: 53 60.2 0.88 44.2: 34 37.4 0.91 45.1: 48 45.0 1.07 45.2: 47 58.7 0.80 47.1: 22 45.4 0.48 47.2: 34 37.7 0.90 48.1: 59 56.9 1.04 51.1: 57 58.6 0.97 51.2: 55 60.7 0.91 52.1: 22 25.6 0.86 55 : 20 25.6 0.78 1Z. = 56 59.6 0.94 58.1: 42 51.1 0.82 61.1: 51 58.7 0.87 61.2: 41 45.0 0.91 62 : 56 57.6 0.97 ^ : 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 61.2 5.1 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 29.1 42.2 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 not admissable for defining a holophyletic group. Both the incomplete posterior transverse carina of the 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 18-1 and 44-1 in addition to 7-2 and 27-2 (27-1 is most parsimoniously considered an apomorphy of the subfamily) . All of these characters are highly homoplastic so it is not surprising that the compatibility 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 the Ophion subgroup (Townes, 1971; Gauld, 1979). The undescribed genus (949) would, on Townes' (1971) criteria, also belong to this group. There is no evidence in any analysis that suggests that these taxa and the Ophion subgroup constitute a holophyletic clade. The definition of the group as adopted by previous authors rests on a plesiomorphic feature, possession of a membranous flange on the fore tibial spur, suggesting this 'group' is in fact a paraphyletic assemblage. In Wagner analyses 920, 922, 940, 949 and 950 may be positioned primitively with respect to most other taxa (e.g. Figs 22, 24) but some members may occasionally be united with other taxa. Taxa 949 and 950 are sometimes placed in much more derived positions, the former within the Enicospilusl Stauropoctonus complex (e.g. Fig. 23) and the latter as the sister-lineage to the ThyreodoniEuryophion complex (e.g. Fig. 21). Hellwigiella (950) is placed as the sister-group to the ThyreodoniEuryophion complex on the basis of a number of shared, rather striking, derived characters. These characters (which may be shared by several or all species in the 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 of Ophion from Australia (in ANIC) and Agathophiona species from Mexico. In another group of ichneumonids, the Anomaloninae, similar differences occur between closely related eremic and non-eremic organisms (e.g. Gravenhorstia (Erigorgus) and Gravenhorstia (Gravenhorstia) species (Gauld, 1976)). The occurrence of so many apomorphic states of these characters together in day-flying species strongly suggests the characters should be considered as a character suite and perhaps accorded less taxonomic weight. Unlike members of the Thyreodonl Euryophion complex, Hellwigiella shows the plesiomorphic state of important characters used for 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 of derived features shared with the more specialized taxa, especially those in the Enicospilusl Stauropoctonus complex (e.g. 15-1, 16-1, 22-1, 33, 45-1) yet at the same time lacks one major apomorphy of the group (63) . Taxon 949 also shows the plesiomorphic state of many of the stem characters, including 25-2, 28-2, 42-1, 51-1 and 52-2. It does, however, share a small number of derived features (including 1 • 1 , 15 • 1 , 16- 1 , 33 and 45-1) with another enigmatic taxon , Sicophion pleuralis (922). Sicophion, like taxon 949, is primitive in a surprising array of other features and it seems plausible that these two taxa have a sister-group relationship and represent a primitive group that evolutionarily converged with the Enicospilus group, possibly as a result of similar selection pressures. The striking phenetic resemblance between 949 and some species of Leptophion is perhaps more understandable when one considers these taxa have apparently evolved in mid altitude tropical cloud forests - 949 in the Neotropics and Leptophion in South East 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 40 and 46 (e.g. Fig. 25). In other respects Riekophion does not appear to be at all closely related to either Dicamptus or Enicospilus (Gauld, 1977) and the high label scores obtained by the species for the derived states of these characters (Appendix 8) strongly suggests parallel derivation in two 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 most primitive ophionines, then the most parsimonious arrangement of these taxa with respect to all other 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 of Townes, 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 four unplaced taxa, 905, 906, 923 and 927. The position of these four taxa can now be considered in relation 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 -O E:ao*£<>KJK)p!OpOi «O CO — ' O K3 NJ C*> Oi *_zt "O ^O «O O *O O a ^ 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 other taxa but shows some affinity with the Euryophion group (937 (25), 936 (24) and 935 (23)). Taxa 905 and 906 share several unusual apomorphic features including 6-1 (elsewhere only found in the 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 in 922 and 923). Taxa 905 and 906 also share all the stem characters of the Enicospilusl Stauropoctonus + 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 the features 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 noteworthy that in earlier Wagner analyses (e.g. Figs 22-24) 905 and 906 generally were split up, 905 often grouping with 923 , 964 or 965 , whilst 906 usually was united with 960 or 961 . If the two taxa were entered together first in the data file, then they were not separated , but remained as a distinct clade(Fig.21). Taxon 927 (Barytatocephalus mocsaryi) shares 24 derived characters with taxon 936, 23 with 914 and 946 and 22 with 913. It shares all the stem features with the major complexes and additionally has the apomorphic condition for 22-2 and 58- 1 , suggesting it belongs near the base of the Thy reodonl Euryophion lineage. It does not have the major apomorphies of the Enicospilusl Stauropoctonus branch (that is 22-1, 33 and 63). Taxon 923 (Prethophion latus) is an enigmatic Neotropical taxon that has relatively few characters in common with any other species. The largest number, 22, are shared with 918 and it has 21 in common with 906. It does have all the 'gain' stem apomorphies in common with other higher taxa, but is plesiomorphic for 25-2, 28-2 and 45. It does not have any of the apomorphies defining the Thyreodonl Euryophion complex, but is apomorphic for character 63, suggesting it belongs near the base of the Enicospilusl Stauropoctonus stem. It is most parsimonious to treat Prethophion latus as the sister species of the 905 -I- 906 lineage (on the basis of the derived state of 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 in Fig. 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 than any obtained by Wagner analysis. 116 I. D. GAULD o — . •— . Fig. 27 Cladogram derived from subjective evaluation of character complexes involved in grouping taxa in cladograms presented in Figs 21-25. It is noteworthy that this, at 242 steps, is a more parsimonious arrangement 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 be fairly robust, but other groups are only supported by weak homoplastic characters. The overall arrangement, placing the Ophion and Eremotylus groups in primitive positions and having a bifurcated evolutionary line, does seem reasonably sound. It corresponds with increasing morphological complexity, particularly in the form of the fore tibial spur (character 51), the region of the spiracular sclerite (character 15), the modification of the propodeum (characters 22, 25) and the loss of umbo (character 33). What little is known about the structure of the final instar larvae supports this arrangement. The cephalic capsules of the larvae of Enicospilus, Dicamptus, Euryophion and Thyreodon are clearly more specialized than those of Ophion in having a modified hypostoma (Short, 1978). The larvae of Euryophion and Thyreodon are similar in having numerous setae (11+) on each part of the maxillary lobe median to the hypostomal spur and adjoining the maxillary palp. Dicamptus and Enicospilus have eight or less setae in this position. Certain venational characters also support the phylogeny proposed. These include 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 the tibial 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 is quite disconcerting to see the degree of evolutionary convergence that has occurred in separate lineages (such as the development of alar sclerites and a sinuous Rs+2r in Riekophion and DicamptuslEnicospilus and the general similarity of structure between the undescribed Costa Rican genus and species of Leptophiori) . No amount of rearrangement of lineages could possibly remove any but a small proportion of such homoplasy. Perhaps the development of such similarities is due to some underlying adaptation of the genotype not visually manifested (the underlying synapomorphies of Saether, 1979). It is difficult to explain otherwise (unless one postulates reticulate evolution - 'ancient hybrids' giving introgression) how such an unusual feature as alar sclerites has appeared, apparently independently in at least six evolutionary lineages of Ophioninae (Afrophion, Sclerophion, Leptophion, Laticoleus, Riekophion and DicamptuslEnicospilus). Similar examples of such unusual parallelisms can be found in the modification of the penultimate distal hamulus, the projecting pecten of the hind tarsal claw and the 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 of four 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 an impressed clypeal margin (7-2) and having Rs in the hind wing at least slightly curved (44-1). Not all species of these genera necessarily possess all these apomorphies; Afrophion species are exceptional in having both a blunt clypeal margin and a straight Rs whilst some Rhopalophion species have the mesopleural furrow obsolescent. Although this genus-group is amongst the most primitive of ophionines, its position at the base of the phylogenetic tree rests on the assumptions 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 questionable and if it could be demonstrated that the assigned polarity is incorrect then these two features would become convincing apomorphies supporting the holophyly of the Ophion group. This change in polarity would necessitate placing the Ophion group a little higher up the evolutionary tree, possibly as the sister lineage to the Eremotylus group; these two groups share the apomorphic 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 slightly flared outwards in profile (7-1) and often slightly concave. The mandibles are a subtly different shape 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 the group is supported by the apomorphic states of characters 33 and 45-2. Apart from possession of certain primitive features such as the plesiomorphic state of character 51-2, and more extensive propodeal carination, all of these taxa appear to be highly specialized, sharing a variety of apomorphic features with other taxa placed higher up the phylogenetic tree. More evidence of the 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 anterior part of the propodeum (character 22-2) and by the form of the claw (character 58-1). The inclusion of Barytatocephalus within this clade does seem reasonable. In addition to the characters used in the analysis, two other features suggest Barytatocephalus belongs here. Vein 3r-m in the fore wing of species of this genus forms a more obtuse angle with Rs+2r than is the case with most other ophionine taxa. Most other members of this clade have a similar specialization. At the base of the hind wing, in the anal cell, a vestige of a vein (? 2A) is often discernible. In most ophionines this trace, when observable, can be seen to be fairly near to and parallel with the hind margin of the wing; in both Barytatocephalus and other members of the Thyreodon complex (but not Euryophion) this vein trace is remote from the wing margin and close to and parallel with \A. Townes (1971) considered Barytatocephalus to be a derivation of Enicospilus, 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 included in the Enicospilus lineage. The Enicospilus group is the largest complex and one of the most difficult to define. Its holophyly 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 in common. Within this group five subordinate lineages are recognizable, the Orientospilus, Ophiogastrella, Stauropoctonus, Leptophion and Enicospilus subgroups. Each of these is apparently a holophyletic group. These are the various groupings recognized in the classification proposed below. A classification of the Ophioninae Hennig (1966) pointed out the existence of two sorts of monophyletic taxa - holophyletic and paraphyletic 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 amalgamated with a holophyletic taxon to form a slightly larger holophyletic taxon (e.g. Aulophion + Stauropoctonus), there are other cases where dogmatic adherence to cladistic tenets produces impractical results. For example, if the Hawaiian genera were incorporated into Enicospilus to give a single holophyletic taxon, the resultant genus would contain an extraordinary range of morphological and probably biological diversity. I concur with Martin (1981) in believing that a classification, as opposed to a phylogeny, is a compromise between known or hypothesized inter-relationships and nomenclatural convenience. A good classification has at least two important 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 phylogenetic classification is an optimum as it is both more predictive and more stable than one that is not entirely phylogenetic. Whilst I agree with these authors that this is often the case, I do not think it is 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 Plectochorus appears to be holophyletic. Each of the other four taxa is apparently paraphyletic with respect to the genera that are more specialized than it (Townes, 1971; Gauld, 19846). Although not a strict phylogenetic 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 to corroborate the idea that these taxa are a nesting paraphyletic series. I personally see no way of establishing a strict phylogenetic classification for such a group, other than by lumping all the genera, an action that would result in a much less informative and less useful classification than the present one. The existence of such nesting paraphyletic groups would seem to be a corollary of the punctuational model of evolution. Stanley (1979) argues convincingly that many major adaptive changes 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 of phyletic 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 radiation from an Enicospilus ancestor into niches hitherto, for enicospilines, unexploited, it is fairly unlikely that any prediction made from study of the biology of continental Enicospilus species will apply to Pycnophion. Similarly, any generalizations made about Enicospilus are likely to have to be prefaced by 'except in some Hawaiian species'. Consequently the purposes of prediction and generalization would both be better served by recognizing the Hawaiian taxa as distinct genera, even though Enicospilus thus becomes a paraphyletic taxon. A second area of contention relevant to the present work concerns the reality and recognition of polythetic superspecific taxa. The currently recognized ophionine genera have been accepted initially 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 dismiss polythetic taxa as the product of unsatisfactory discrimination. Whilst I accept that this is an explanation, I dispute that it is the most obvious one. Most cladistic studies reveal that homoplasy is a common phenomenon. If a data set is highly homoplastic, then logically it follows that 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 a multiplicity of new genera, many of which would be monobasic. The resultant classification will have little predictive value as a high proportion of new taxa are likely to necessitate new genera for their accommodation (Gauld & Mound, 1982). The following classification is suggested for the subfamily Ophioninae. It is not strictly cladistic, as it recognizes some paraphyletic taxa; however it is a fair approximation of the presumed phylogeny. Formal tribes are not recognized, but these could be used for the various genus-groups. It should be noted that some very small genera were not incorporated in the cladistic analysis (these are denoted by an asterisk), but they were closely studied in an earlier work (Gauld, 1979). This has facilitated their placement in the present system. Subfamily OPHIONINAE Ophionidae Shuckard, 1840 Enicospilini Townes, 1971 OPHION genus-group Op/iion Fabricius, 1798 Alophophion Cushman, 1947* Sclerophion Gauld, 1979 rfitC Afrophion Gauld, 1979* AgathophionaWestwood, 1882 Rhopalophion Seyrig, 1935 Xylophion Gauld, 0-979 • / «? y 0 SICOPHION genus-group Riekophion Gauld, 1977 Sicophion Gauld, 1979 Janzophion gen. n. EREMOTYLUS genus-group Eremoty/usFoerster, 1869 Trophophion Cushman, 1947* Hel/wig/e/faSzepligeti, 1905 THYREODON genus-group BarytatocephaIusSchu\z, 1911 Euryophion Cameron, 1906 Rictophion Townes, 1971 Syn. n. Dictyonotus Kriechbaumer, 1894 Ophionopsis Tosquinet, 1903 Syn. n. IWiync/iop/iJonEnderlein, 1912 T/rjreodonBrulle, 1846 ENICOSPILUS genus-group Orientospilus subgroup Prethophion Townes, 1971 Simophion Cushman, 1947 Orientospilus Morley, 1912 Ophiogastrella subgroup OphiogastrellaBrues, 1912 Stauropoctonus subgroup Lepiscelus Townes, 1971 Stauropoctonus Br nuns, 1889 Aulophion Cushman, 1947 Syn. n. Leptophion subgroup Pamophion Gauld, 1977 Laticoleus Townes, in Townes & Townes, 1973 Leptophion Cameron, 1901 Enicospilus subgroup P/campfusSzepligeti, 1905 Enicospilus Stephens, 1835 Pycnophion Ashmead, 1900 Banchogastra Ashmead, 1900 Abanchogastra Perkins, 1902 Zoogeographic discussion Present day distribution The approximate numbers of species per area are given in Table 18. For the purposes of this discussion Melanesia, Australia and New Zealand are considered as a single region, the Australo-Pacific, whilst for purposes of comparison Hawaii, with three endemic and one cosmopolitan genera, is ignored. Altogether 20 of the 32 ophionine genera (62-5%) are restricted to a single zoogeographic region. 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 ASTRA BANCHOGASTRA M O O H > M 50 -- 25 1 5 10 — 30 50 - - 25 - - 1 i 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. The majority of the restricted ophionine genera are small taxa with three or fewer species; only Alophophion, Ophiogastrella and Laticoleus are larger than this and they have ten or more species each. Six of the more widely distributed genera occur in only two regions: Rhynchophion and Thyreodon (Neotropical/Nearctic), Euryophion and Orientospilus (Afrotropical/Oriental), Leptophion (Oriental/ Australo-Pacific) and Simophion (Palaearctic/Nearctic). Two genera occur 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 and Stauropoctonus in all but the Nearctic (though it is absent from the African mainland). Only Enicospilus 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 most genera (12) and has the highest degree of generic endemicity (50%). The Afrotropical and Oriental 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 two endemics (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). The values at the intersects of rows and columns V may be defined as: , _ number of genera common to areas X and Y total 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 than do areas that are not contiguous or that have only been contiguous in the distant past (more than 40 million years ago), suggesting that dispersal may be a more important event than vicariance in the biogeographical history of the subfamily. o r1 o f r1 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 0.18 NEOTROPICAL 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 the primitive Upper Cretaceous groups. A fossil Ophion species is known from from the Lower Oligocene (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 group radiated 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 connection probably existed between this landmass and South America. Europe and North America were contiguous and South America was not widely separated from Africa. The most primitive lineage of the Ophioninae (the Ophion genus-group) may have been widespread at this time, as would have been its sister-group, the stem group of the remaining Ophioninae. The Sicophion lineage became isolated in the southern continent of Australia/ Antarctica/South America at an early stage whilst its sister-lineage, the ancestor of the Eremotylus, Enicospilus and Thyreodon lineages, remained in Laurasia or Africa. The ancestor of the Eremotylus lineage probably differentiated in Laurasia whilst the ancestor of the Enicospilus I Thyreodon radiated in Africa. I suggest 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 to South America. Some kind of barrier to the spread of organisms to and from South America is suggested by the high degree of generic endemicity and the fact that the Neotropical species of Enicospilus seem to belong to very few species-groups compared with those of other tropical areas. 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 lineage resulted in widespread extinction of members of the Ophion group, thus leaving the isolated specialized groups extant today. Ophion almost certainly has undergone secondary radiations giving 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 has subsequently become isolated on mountain tops) (Gauld, 19840). Except for Xylophion, a remnant of the early widespread Ophion group-distribution, and Riekophion, a remnant of the southern radiation of the Sicophion lineage, the ophionine fauna of the Australo-Pacific has largely 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 considerable radiation, particularly in the case of Enicospilus , over 100 endemic species of which occur on the island. Although the above scenario is largely speculative it is the most parsimonious interpretation of the present distribution in relation to the postulated phylogenetic history of the group. Other scenarios 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 would involve the usage of rather difficult couplets to allow for evolutionary parallelism between different species-groups in similar habitats in different zoogeographic regions. Practical keys to genera 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 also suffice for the Palaearctic region. A key to the Neotropical genera is currently in preparation and this will suffice for the Nearctic. Townes (1971) also offers reasonable generic keys, but these are now 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 probably paraphyletic 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 impressed clypeal apex and by having Rs in the hind wing usually at least slightly curved and often very strongly bowed. Virtually all species in this genus-group have Im-cu in the fore wing centrally angled 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 in this group have arisen, and the genus is primarily a Holarctic taxon, it seems probable that this group originated in the temperate north. It is probable that at some period it was present in most regions and has gradually disappeared from equatorial regions leaving isolated relicts in South Africa (Afrophion), Australia (Xylophion), Madagascar (Rhopalophion} and Patagonia (Alophophion). The possibility that there has been repeated expansion into and extinction within the tropics is suggested by the presence of some groups of Ophion species on isolated mountains in South East Asia, New Guinea and South America, and by the occurrence of distinctive 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 surface convex, with strong proximal concavity. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profile weakly convex with apical margin sharp but not clearly impressed; clypeus in anterior aspect weakly convex. Ocelli large, the posterior ones very close to the eyes; frontal carina absent; occipital carina complete, 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 carinate laterally; posterior transverse carina of mesosternum absent except laterally. Propodeum with anterior area either occluded or extremly short, the remainder of the propodeum rather abruptly declivous, with irregular 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 of discosubmarginal 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 weakly curved. Fore tibial spur with a membranous flange behind macrotrichial comb; mid and hind trochantelli unspecialized; inner hind tibial spur flattened, with a margin of close fine hair; hind tarsal claws unspecialized, 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-species and A. hynnis (Gauld & Mitchell). It is distinguished from Ophion by the form of the propodeum, the unique structure of the male genitalia and by the basally incrassate Rs+2r in the fore wing. AGATHOPHIONA Westwood Agathophiona Westwood, 1882: 19. Type-species: Agathophiona fulvicornis Westwood, by monotypy. Mandibles twisted 5-10°, barely tapered, with lower tooth slightly the longer; outer mandibular surface slightly 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 and the 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 ones separated 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 the very weak hypostomal carina. Antennae rather short and quite stout, distal segments quadrate. Pronotum unspecialized; spiracular sclerite exposed; notauli short but strongly impressed near anterior margin; epicnemial carina more or less complete; mesopleural furrow virtually absent. Scutellum moderately strongly convex, not laterally carinate; posterior transverse carina of mesosternum present only laterally as vestiges. 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 a glabrous anterior area; Im-cu generally with only a trace of a ramellus, rather evenly but quite strongly curved. Hind wing with Rs quite strongly curved. Fore leg with tibial spur bearing a membranous flange behind 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 in profile only slightly longer than posteriorly deep; thyridia small and close to anterior margin; epipleuron up-turned. Female with subgenital plate enlarged, as long as tergite 2 and medially notched; ovipositor sheath 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 remarkable mouthparts are presumably an adptation to feeding from certain types of flowers but the function of the rather unusually modified female terminalia is not known. 124 I. D. GAULD ALOPHOPHION Cushman Alophophion 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 profile weakly convex, margin often impressed, sometimes very narrowly so; clypeus in anterior aspect weakly convex. Ocelli generally large, the posterior ones close to the eyes; frontal carina absent; occipital carina absent. Antennae moderately to very long, at least 1-3 times length of fore wing. Pronotum unspecialized or mediodorsally somewhat flattened and quite long; spiracular sclerite exposed; notauli present on anterior part of mesoscutum; epicnemial carina generally strong; mesopleural furrow distinct, diagonal, extending from episternal scrobe to near subalar prominence. Scutellum very weakly convex, usually narrow and not carinate laterally; posterior transverse carina of mesosternum absent except for lateral vestiges. Propodeum with anterior area occluded, transverse and often lateromedian longitudinal carinae discernible, 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 short ramellus which is directed more anteriorly than that of Ophion. Hind wing with Rs curved. Fore tibial spur with a membranous flange behind macrotrichial comb; mid and hind trochantelli unspecialized; inner hind tibial spur flattened, with a margin of long close hairs; hind tarsal claws unspecialized. Gaster moderately slender; tergite 2 in profile elongate, thyridia oval, separated from anterior margin of tergite by its own length 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 species occur 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 to Ophion, 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 joining pterostigma near centre; first subdiscal cell stouter than normal; ramellus, when present, directed more anteriorly 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 bidentate or with upper tooth slightly the longer; outer mandibular surface more or less flat, except for basal concavity, moderately punctate. Maxillary palp 5-segmented, labial palp 4-segmented; clypeus in profile weakly to moderately convex, margin impressed, acute; clypeus in anterior aspect weakly convex, rarely truncate or even slightly concave. Ocelli usually large, the posterior ones separated from eyes by less than their own diameters; frontal carina absent; occipital carina usually complete dorsally, ventrally not reaching the hypostomal carina, rarely with occipital carina mediodorsally obsolescent, or in a few species with it entirely absent. Antennae generally of moderate length, in a few deserticolous species the flagellum short and with central segments quadrate. Pronotum unspecialized; spiracular sclerite usually completely exposed; notauli weak but discernible on anterior 0-2 of mesoscutum; epicnemial carina generally strong and 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 carinate only on anterior 0-2 or less, rarely with lateral longitudinal carinae complete to posterior margin; posterior transverse carina of mesosternum usually present only laterally as vestiges. Propodeum with anterior area occluded except for a small semicircular depression centrally; propodeal carina variously developed, at most with both transverse carinae and the lateromedian longitudinal carinae complete, enclosing an area superomedia, in the most exceptional cases with all carinae only vestigial; posterior area usually rather smooth. Fore wing with pterostigma moderately to very stout; marginal cell very long and slender; Rs+2r usually virtually straight, slender, barely broadened before joining pterostigma, in some species evenly broadened to join pterostigma; discosubmarginal cell with glabrous area in anterior corner; Im-cu usually with a well-developed ramellus, rarely with stub-like indication of this vein present; Im-cu centrally strongly geniculate. Hind wing with Rs from weakly to very strongly curved. Fore tibial spur with membranous flange behind the macrotrichial comb; mid and hind trochantelli unspecialized; inner hind tibial spur flattened, with a margin of close long hairs; hind tarsal claws unspecialized, those of males usually more closely pectinate than those of the female. Gaster moderately slender; tergite 2 in profile moderately to strongly elongate, usually with thyridia large, close to anterior margin; umbo well developed; epipleuron up-turned. Ovipositor sheath usually slender, in one species-group exceptionally stout. Ophion is a very large genus that is widely distributed throughout the world except for the Afrotropical region. In tropical South America and South East Asia few species occur and these are restricted to high altitudes (Gauld & Mitchell, 1981). Ophion as currently defined is almost certainly a paraphyletic assemblage; Agathophiona, Rhopalophion, Sclerophion, Xylophion, Alophophion and Afrophion are phenetically highly divergent (Gauld, 1979), holophyletic species-groups which have probably arisen from within 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 be necessary 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 other species-groups from Ophion and treat these as separate genera, but in each case the group removed has not been holophyletic as comparatively few of the subtropical and eremic Ophion species are known. Recent collecting has greatly increased our knowledge of these insects but more work needs to be undertaken before it is possible to fully resolve this genus. The major species-groups of Ophion are outlined below; all except the unsatisfactorily resolved luteus-group are holophyletic but I have avoided treating these groups as distinct genera until the luteus anathema can be resolved. As each of these groups appears to be defined by 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 Ophion species (Townes et al, 1961). Its monophyly is suggested by the possession of a mediodorsally lengthened pronotum, having Rs+2r joining the pterostigma distal to the pterostigma base and having the longitudinal lateral carina of the propodeum with a raised ridge diverging towards the propodeal spiracle. In O. peregrinus this is only weakly developed but in other taxa this ridge is frequently very strong and continuous with the carina, and the front part of the carina is absent so the apparent carina curves anteriorly towards the spiracle where it terminates. O. peregrinus is apparently the most primitive taxon as it retains most of the propodeal carinae and has a strongly geniculate Im-cu; the other taxa have only a vestige of the anterior carina (the lateromedian and posterior transverse carina are 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 Australian species (Gauld, 1977). The group is characterized by possession of a proximally broadened Rs+2r which is slightly curved before reaching the pterostigma. The European species, O. minutus, may belong to this group. The five Australian species form a monophyletic subgroup characterized by the black interocellar area. In the Oriental region the species of the bicarinatus group are restricted to mountains in Sri Lanka, India, Burma, Malaysia, Taiwan and Sumatra. Their present fragmented distribution and their paraphyletic nature with respect to the Australian species-group suggests that this group was more widespread throughout the Indo- Australian region in the past, probably when the climate 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 very short 3r-m. It does not appear to be closely related to any other Indo- Australian species. 126 I. D. GAULD O.