J. HYM. RES. 2(1), 1993 pp.227-304 Phylogeny of Aculeata: Chrysidoidea and Vespoidea (Hymenoptera) DENIS J. BROTHERS AND JAMES M. CARPENTER (DJB) Department of Zoology and Entomology, University of Natal, P.O. Box 375, Pietermaritzburg, 3200 South Africa; (JMC) Department of Entomology, American Museum of Natural History, Central Park West at 79th Street, New York, NY 10024, U.S.A. Abstract.-The development of ideas on the phylogeny of the aculeate Hymenoptera, especially Vespoidea and Chrysidoidea, since Brothers's 1975 and Carpenter's 1986 studies is reviewed. The results oftheir detailed analyses of aculeate higher taxa are re-evaluated in the light of new information and/or reinterpretations by subsequent workers. Almost all of their earlier results, including the relationships within the Chrysidoidea, the holophyly of Vespoidea (including Pompilidae), the sister-group relationship ofScoliidae (including Proscoliinae) and Vespidae, that ofSapygidae and Mutillidae (including Myrmosinae), and the composition ofBradynobaenidae are confirmed. The final preferred cladogram, using 219 variables and based on ground plans for all families ofChrysidoidea and Vespoidea and three taxa of Apoidea, indicates the following relationships (components of the superfamilies included within curly brackets): {Plumariidae + (Scolebythidae + «Bethylidae + Chrysididae) + (Sclerogibbidae + (Dryinidae + Embolemidae»)))} + ({Heterogynaidae + (Sphecidae s.l. + Apidae s.l.)} + {Sierolomorphidae + (Tiphiidae + (Pompilidae + (Sapygidae + Mutillidae») + (Rhopalosomatidae + (Bradynobaenidae + (Formicidae + (Scoliidae + Vespidae»»)}). INTRODUCTION ofthe new information. The final result is the best­ supported cladogram available for the superfami­ Current ideas on the phylogeny of the aculeate lies of Aculeata and for the components of the Hymenopteradatefrom the publicationofBrothers's Chrysidoidea and Vespoidea. We do not analyze (1975) paper, which was the first attempt to apply the Apoidea in any detail since it is clearly a cladistic principles in an analysis of the entire holophyletic group and analyses of some of its group. Since the initial purpose of that study was components are presented by Alexander (1990, merely the elucidation of the relationships of the 1992); that superfamily is in any case the group components of the Mutillidae s.L, the paper had dealt with in least detail in the 1975 analysis. limitations in that the component taxa were dealt Limits and names ofthe various taxa included in with in differing detail (analysing tribes in some the Aculeata are sometimes problematic. Thus, taxa but lumping families presumed to comprise Brothers's Bethyloidea and Sphecoidea should be holophyletic groups elsewhere) and the sample of Chrysidoidea and Apoidea respectively in terms of exemplars used to derive taxon ground plans was nomenclaturalpriority (Day 1977, Michener 1986), probably inadequate for some taxa. Although not and the correct names are used below for these taxa all of the conclusions of that study have been even though other names may have been used by accepted, it has fulfilled one of its major functions the authors of the papers under discussion. The in stimulating further investigations ofthe relation­ abbreviations 's.L' (sensu lato) and 's.s' (sensu ships among the vmious higher taxa of aculeates, stricto) are used to indicate more and less inclusive Carpenter's (1986) analysis of the families of concepts where confusion could result, e.g., Chrysidoidea being particularly significant. The Vespoidea s.L is more or less the concept ofBroth­ present paper aims to survey the relevant literature ers (1975), whereas Vespoidea s.s. comprises only which has appeared on the topic since 1974, to Vespidae s.L (Masaridae + Eumenidae + Vespidae analyze new characters and interpretations pre­ s.s.). sented therein, and to modify and amplify the data base ofBrothers (1975) and re-analyzeitin the light 228 JOURNAL OF HYMENOPTERA RESEARCH PREVIOUS STUDIES Over three years, Saini & Dhillon investigated various modifications ofthe metatibial spurs (1978), In the following survey of papers on this topic, mouthparts (1979a, b) and metathorax (1980) in 22 we generally deal with them in chronological order, varied families of Hymenoptera. Single and often starting in 1975 with Brothers's study which exam­ relatively derived representatives were apparently ined 25 taxa and 92 characters of Aculeata. The used for each family, so that the studies were very cladogram he obtained is reproduced here (Fig. 1) limited, providing no information on intrafamilial in the format generated by CLADOS (Nixon 1992) variation. Furthermore, there was no differentia­ from Hennig86 version 1.5 (Farris 1988), the com­ tion between plesiomorphies and apomorphies, in­ puter programs used for our new analyses, for validating their conclusions. On the basis ofnum­ easier comparison with them. The distribution of ber and development of the metatibial spurs, they derived character states on the various internodes linked Mutillidae and Formicidae (including their of the 1975 cladogram is also provided (Appendix Dorylidae) in one line, and Chrysididae, Scoliidae, IA) to remedy a lack in the original paper; this is Sphecidae, Vespidae S.S., Eumenidae, Pompilidae similar to the listing given by Wahl (1990) but with and Apoidea s.s. in another. Looking at the mouth­ a few corrections. (Note that the distribution and parts, they identified two lines of modification (in numbering ofvariables shown in Fig. 1results from the maxillae involving therelative sizes ofthe galea one of our new analyses (see below) and is not the and lacinia andinthe labiathe relative development same as that used in the 1975 paper.) Major of glossa and paraglossa), the first leading from conclusions from the 1975 analysis were the estab­ Ichneumonoidea to Chrysididae, Mutillidae and lishment of the holophyletic nature of the Formicidae, and the other from Chalcidoidea to Chrysidoidea, with Plumariidae as the sister taxon Scoliidae, Sphecidae, Vespidae S.S., Pompilidae, of the remaining chrysidoid families (exemplified Eumenidae and Apoidea s. s. Theiraccountofmodi­ by Scolebythidae and an amalgam of other taxa); fications of the metapleuron and metapostnotum the recognition of the polyphyletic nature of the could be interpreted to indicate close relationships traditional superfamily Scolioidea, with placement between Chrysididae, Scoliidae and Sphecidae, a ofthe Scoliidae as the sister taxon ofthe Vespidae lineage including Vespidae S.S., Eumenidae, s.l. rather remote from the Tiphiidae; the accep­ Formicidae and Apoidea S.S., and distinctness of tance of only three superfamilies (Chrysidoidea, the Pompilidae. They disagreed with Brothers's Vespoidea and Apoidea) instead of the traditional (1975, 1976) interpretation of the origin of the seven; inclusion ofPompilidae in Vespoidearather 'propodeal triangle' (as an expanded metapost­ than close to Apoidea; inclusion ofMyrmosinae in notum) in Apoidea s.l. Mutillidae rather than Tiphiidae; and inclusion of Konigsmann (1978), in that part ofhis survey of Typhoctinae, Chyphotinae, Apterogyninae and hymenopteran phylogeny covering the Aculeata, Bradynobaeninae in a single newly constituted based his treatment to a great extent on Brothers's family (Bradynobaenidae) rather than in Mutillidae (1975) analysis but indicated large areas ofuncer­ and Tiphiidae. Brothers (1976) further investi­ tainty (Fig. 2), usually where he felt that the char­ gated the structure of the metapostnotum and sec­ acters given by Brothers in support of particular ond and third phragmata in various aculeates, find­ internodes were weak or homoplastic. He placed ing corroboration for his earlier conclusions. Sclerogibbidae as the sister group of all other In 1977 Rasnitsyn described a new subfamily of aculeates (on the basis of the multisegmented an­ Sco1iidae based on a monotypic genus, Proscolia tennae and apparent lack of synapomorphies with Rasnitsyn, which he considered indicated that "the any particular aculeate group), excluding it from ancestor of the family was at least as primitive as the Chrysidoidea, but otherwise accepted the divi­ the Anthoboscinae (Tiphiidae)", and thus probably sion ofthe aculeates into three holophyletic groups. most closely related to that taxon. Such a conclu­ He analyzed the remaining taxa within the sion does not necessarily follow, however, since it Chrysidoidea in greater detail than Brothers had, is based on shared plesiomorphies. and suggested a sister-group relationship between VOLUME 2, NUMBER 1, 1993 229 Plumariidae and Scolebythidae (based on the com­ coding ofcharacters, so that his information cannot mon reduction of the pronotal collar), between easily be included in any new cladistic analysis. Embolemidae and Dryinidae (based on the 10­ In his analysis of the evolution of the Hy­ segmented antennae and single mesotibial spur in menoptera, Rasnitsyn (1980) reinterpreted some of both) and betweenChrysididae and Cleptidae (based Brothers's (1975) characters and added a few new on integumental sculpture, wing venation, form of characters. In the Aculeata (his Vespomorpha; ig­ the ovipositor and possibly the lack of articulation noring taxa known only from fossils, which were and sensillar fields between the first and second not considered by Brothers), he recognized the metasomal segments) but could not resolve the Chrysidoidea as a holophyletic group (including relationships among thesepairs oftaxaorBethylidae Sclerogibbidae), but split