A WORD ON CHALCIDOID CLASSIFICATION. John S. Noyes. Dept. of Entomology, The Natural History Museum (NHM), Cromwell Road, London, 5W7 5BD, England.

Arnold Menke's jibe at us chalcidologists for not getting our house in order (see CF #12) has spurred me into action. I feel more than a little responsible because it was the classification used by me in The Hymenoptera that initiated his comments.

Firstly, it would seem that the problem with the Chalcidoidea is not really with the number of families generally regarded as valid but with what we understand of the phylogeny of the group. Even if we decided on some arbitrary number (from 1 to 30) of families, we would still have the problem of what to do with lower ranked taxa. We really need to get some good character evidence to enable us to decide on the phylogenetic relationships of the group as a whole before we can even start to sort out the validity of higher taxa. To my way of thinking this is a very long way off, and may in fact be impossible given the undoubted complexity of the group.

However, just to throw my oar in I thought it would be interesting if we could come up with some sort of phylogeny which reflects all our views of the possible relationship of the higher taxa regarded as monophyletic entities in the Chalcidoidea. My current thinking is shown opposite. This phylogeny is almost entirely intuitive and is not based on any good evidence. If I was asked to do another tomorrow it would certainly look entirely different. Perhaps readers of CF have their own ideas or would like to comment on this phylogeny. Perhaps eventually we could arrive at some sort of intuitive "concensus" tree. Any takers?

Although the phylogeny given opposite is basically intuitive it may be worth commenting on some of the groupings. My "eulophid" lineage (Trichogrammatidae, Eulophidae, Elasmidae, Aphelinidae, Signiphoridae) is based on reductional characters which are not regarded as good ones, eg. reduction in antennal segments, tarsal segments (what about the mesofurca?). The "pteromalid" lineage consists of all other taxa, except of course the Mymaridac, Rotoitidac and Tetracampidae, which I think are basic. Could the "pteromalid" lineage have started with parasitoids of the eggs or larvae of xylophagous or other beetles? Within this lineage I think that the Pteromalidae, as it is presently considered must be polyphyletic and this I have indicated by listing some of the major subfamilies (underlined). The monophyletic relationship of the Eucharitidae and Perilampidae is generally accepted without question, but I am not sure what to do with the Echthrodapinae (are they disguised Torymidae?), Chrysolampinae (which seems to link the Miscogasterinae and Perilampidae), and Philomidinae (any comment Chris/John?). I am pretty sure that the Ormyridae-Ormocerinae group is a good one although I cannot come up with any good synapomorphies except for a possible phytophagous relationship. The Spalangiinae Diparinae group seems to be a good one although I base it mainly on similarity of general habitus and wing venation and the association with the immature stages on non-xylophagous Coleoptera and synanthropic Diptera. I have no doubt that the Cleonyminae-Encyrtidae lineage is good and reflects the development from an ancestral group which was probably parasitic in the immature stages of xylophagous beetles through to a group of secondary phytophagous habit (Tanaostigmatidae) or parasitic on the immature stages of holometabolous insects and auchenorrhynchous Homoptera (Encyrtidae).

A WORD ON CHALCIDOID CLASSIFICATION. Gary Gibson.

I am glad that John has raised the issue of chalcidoid classification because it gives us all the chance to exchange gut feelings and personal opinions which we would never dare put into print for fear of being ridiculed, as well as exchanging bits and pieces of information or questions that may, with discussion, help form the 'big picture'. It is about time that we stopped crying about the chaos in classification and started a rapid exchange of ideas and information to help solve problems, which is why CF was started in the first place. I hope that chalcidoid taxonomists throughout the world will take part in this discussion. My own initial two cents worth is given below. You will note that there are generalizations and unfounded statements and opinions, but this is the place for them.

I agree totally with John that the problem in chalcidoid familial classification is not in the number of families, but in our understanding (or the lack thereof) of the monophyly and relationships of the morphotaxa we variously classify as families, subfamiies, tribes, etc. As he indicated, what we really need is some "good character evidence" to enable us to decide on relationships. For the most part I think that we lack accurate knowledge of structure, and particularly accurate knowledge of character state distribution throughout the superfamily. This is a severe problem for cladistic analyses and determining relationships, but it is not unique to the Chalcidoidea. Many workers decry the amount of homoplasy found in their groups because it results in 32 billion equally parsimonious trees. I suspect that much of this homoplasy is 'false' homoplasy resulting from incorrect information on structure and its distribution and leading to incorrect hypotheses of homology or polarity. If I have found anything from my morphological work it is that the literature on Hymenoptera is repleat with inaccuracies. At times I've wondered if this has resulted partly from a different way of 'looking at things' by pre-cladistic and post-cladistic systematists. To generalize, pre-cladistic taxonomy emphasized differences between taxa for classification whereas post-cladistic taxonomy emphasizes 'sameness' (synapomorphies). The concept of homology is of critical importance to post-cladistic taxonomy and because of this 'sameness' is examined rigorously (at least in theory if not always in practice). In the pre-cladistic era, was 'sameness' (either superficial similarity, or superficial 'non-sameness' resulting from character state transformation) examined less critically because one was looking for differences? Because of this philosophical difference, do we now have a greater knowledge of what apparently differs among taxa (and a hodgepodge of higher level taxa and classifications) rather than what is truly shared among taxa? Regardless, the advent of cladistics should have resulted in a renaissance in morphology because of character state analysis. For the most part I do not think that this has happened. Most institutions have always seemed to be quite excited about supporting studies that use computers to massage data, but much less willing to fund the basic morphological studies that discover the data to be massaged (including the National Science Foundation judging by some proposals that I have been asked to review and were not funded (I hope that there is no correlation)). Even more insidious at present is an apparently universal institutional concept that "molecular techniques" and "science" are synonymous (the only True science is Molecular Science -praise be Its name). It has been said that there are only two sure things in life - death and taxes. A third may be the universal addiction of funding agencies, and thus managers of organizations, for supporting the latest, new and improved, 'cutting edge' of science.

Whether the above opinions are valid or not, I do not think we are going to get very far in resolving higher level chalcidoid relationships until we have a much more accurate understanding of structure and its distribution. Much of this work will have to be done during M.Sc. and Ph.D. degrees because most graduates will find employment with organizations that, at best, are willing to tolerate morphological studies. Any professorial type turning out chalcid taxonomists should demand as part of the degree program that the candidate study, and publish on, one character system throughout at least the Chalcidoidea and preferably the Hymenoptera (see Darling's paper on the labrum as an example of such a study). Is this too much to ask?

Some of you may be beginning to wonder if I am ever going to say anything about chalcidoid classification and John's phylogeny, but first a few more general comments. One of the questions that has always intrigued me is why are chalcidoids structurally so diverse compared with other groups of parasitic Hymenoptera (if you've seen one ichneumonid you've seen them all), and why do there always seem to be intermediates. This pattern would indicate that Chalcidoidea is a relatively young (terminal) and radiating lineage that has not had time for extinctions to parcel out nice distinct groupings. However, some features seem to indicate Chalcidoidea as a very old (basal) lineage in Apocrita. For example, chalcidoids are one of the very few parasitic Hymenoptera to retain a free prepectus (retained also in Mymarommatidae (?), Stephanidae, Monomachidae, Austroniidae and Roproniidae) and a tergo-trochanteral muscle (retained in Mymarommatidae, Stephanidae, Diapriidae, Megalyridae and Ceraphronoidea). Consequently, both (plesiomorphic) features are otherwise known only in Stephanidac, and possibly Mymarommatidae. If Mymarommatidae is the sister group of Chalcidoidea (by no means definitively established), then we know that the common ancestor of Chalcidoidea is at least lOOmy old (early Cretaceous) because mymarommatids are knowo from Lebanese amber. Yet several groups of Proctotrupoidea s. 1. that appear to be more derived in structure than are Chalcidoidea are known from the Jurassic. I would think that the common ancestor of Mymarommatidae + Chalcidoidea, if not the Chalcidoidea itself, must have been present during the Jurassic. Could the pattern of diversity we see in extant Chalcidoidea be a consequence of diversification over an extremely long time period combined with a more recent (middle to late Cretaceous) secondary explosive radiation of the group? Does anyone want to comment on this?

To finally get to John's comments about relationships within the Chalcidoidea, not surprisingly there are parts I agree with and parts I don't. I certainly agree with the Eurytomidae + (Chalcididae + Leucospidae) branch. There does not seem to be much of a problem with the monophyly of Leucospidae, but does anyone know of characters that indicate the monophyly of Chalcididae exclusive of Leucospidae, i.e., is Chalcididae paraphyletic relative to Leucospidae? The same question for Eurytomidae re: the Chalcididae.

I have slightly different views on the "eulophid" and "tetracampid" lineages. Some reductions do not seem to be relevant. For example, Zdenek showed in his Australasian work that the groundplan number of antennal segments in Eulophidac probably was 12. Based on non-chalcidoid parasitic Hymenoptera I think that the common ancestor of chalcidoids must have had 5-segmented tarsi and most probably a curved, bifurcate fore tibial spur. I restrict the "eulophid" lineage to Eulophidae, Elasmidae and Trichogrammatidae, because they share 4-segmented tarsi (secondarily reduced to 3 segments in Trichos.) and a short, simple fore tibial spur. Other groups that have four segmented tarsi, such as some aphelinids, encyrtids and mymarids retain the primitive structure of the fore tibial spur, and these I think probably are independent reductions in number of tarsal segments. Consequently, I would put aphelinids and signiphorids into the "pteromalid" lineage. On the other hand, rotoitids (4-segmented tarsi) and tetracampids (4-segmented in tetracampine males) have a short and relatively straight, but still bifurcate fore tibial spur [for tetracampid structure see Figures 5-7 in Boucek (1958): Revision der europaischen Tetracampidac (Hym. Chalcidoidea) mit einem Katalog der Arten der Welt. Acta Ent. Mus. Natl. Pragae 32: 41-90]. Could this be an intermediate structure, which would indicate that Rotoitidae and/or Tetracampidae are the sister group of the Eulophidae + Elasmidae + Trichogrammatidae? As always there is a fly in the ointment with this character system. Although only one extant genus of Mymarommatidac is considered valid at present, there actually are two good genera, one of which has a simple fore tibial spur and the other a conspicuously bifurcate spur (both types of spurs relatively long). So even here we have a problem with (outgroup?) comparison. I would really like to see someone make a comprehensive, detailed SEM study of the fore tibial spur and fore tarsal structure throughout Chalcidoidea so that we had some 'facts' to base hypotheses on.

I also have many more doubts about a strong "Cleonyminae-Encyrtidae" lineage than John because I don't buy the monophyly of Eupelmidae. All the characters by which the family has been defined, or associated with encyrtids and tanaostigmatids, either are primitive features or apparently were derived independently several times (see Gibson (1989): Phylogeny and classification of the Eupelmidae, ... Mem. Ent. Soc. Canada, No. 149]. I am convinced that there is a close (possibly sister group) relationship between Cleonyminae (Cleonymini) and Calosotinae, but Eupelminae may be more closely related to some other cleonymine group (perhaps the sister group of Chalcedectini), and I have no idea of the true relationships of the Metapelmatinae. Interestingly, Eusandalum Ratzeburg (Calosotinae) and Metapelma Westwood (Metapelmatinae) have 9 metasomal (10 abdominal) terga, ie. abdominal terga 9 and 10 are not fused into a composite tergum. Cleonymines often have a very distinct suture between the two presumptive terga, but I do not know of any that have the terga separate (primitive condition for Hymenoptera). Ibe epipygium of torymids and some agaonids certainly is not homologous with the 10th tergum of the eupelmids, but can anyone tell me of any pteromalid that has 10 abdominal (9 metasomal) terga? Could this be the groundplan structure for Chalcidoidea, or must I assume that the apical tergum in the above two eupelmid genera is secondarily divided?

I certainly agree with the Tanaostigmatidae + Encyrtidae lineage, with Cynipencyrtus Ishii probably the sister group of the Encyrtidae (based on transverse axillae), but possibly of the tanaos. + encyrtids (second choice), or of the tanaos. alone (third choice). How about that for taking a stand. The metapelmatines may also be involved in this grouping, and for some reason I feel that the aphelinids somehow are related to tanaos. + encyrtids, but I have no hard characters to base this on. I should note that Mike Sharkey recently caught a specimen of Cynipencyrtus in Japan, which I was able to CPD and mount. This specimen shows that I was incorrect in stating in my 1989 paper that a completely enlarged acropleuron is a synapomorphy for Tanaostigmatidae + Encyrtidae (p. 21, and Table I character 3). The specimen clearly shows that the acropleuron in Cynipencyrtus is not quite completely enlarged ventrally to the mesocoxa so that a slender, but distinct, lower mesepimeron remains (the ventral part of the mesopleural suture is distinct). This structure is more primitive than that known for tanaos. and encyrtids (is this statement correct for encyrtids ?), but I donÍt know how much significance to place on it. Certainly, the pronotal-prepectal structure of Cynipencyrtus appears primitive relative to both the tanao. and encyrtid structure, both of which can be derived from a Cynipencyrtus-like structure. Could choice 2 be correct? If so, could Cynipencyrtus be some sort of intermediate dade between aphelinids and tanaos. + encyrtids? [The specimen does have fore tibial apical pegs, which I did not know for the 1989 paper.]

My knowledge of pteromalids is limited, to say the least, so I can't comment on the other lineages John proposed. However, to me, one of the most interesting questions in Chalcidoidea is the relationships between torymids and agaonids (sensu whatever). I can't wait for some resolution on this.

What I would at least like to see from other chalcidoid specialists are submissions that discuss the monophyly of their group of expertise. If you believe your group is monophyletic, what are the characters that indicate this, and do you know of any derived features that link your group with other chalcidoids? I also can't wait for the next issue of CF.