Even if Hennig86 or NONA are used, most relevant clades remain exactly the same in all trees. The classification of the Stephanidae, as derived from the interpretation of the above cladogram, is briefly described below.
Schlettererius and Stephanus, although with only a few species, are very distinct from other stephanids, and I propose that their taxonomic status be maintained, but interpreted in sensu stricto.
The genus Megischus is highly polyphyletic, but changing its taxonomic status would result in dramatic nomenclatural changes (mostly the proposition of many new genera), which seems hardly nescessary or adequate.
I was also glad to observe that the genus Hemistephanus was indeed recovered as monophyletic, as I proposed earlier (Aguiar 1998).
Other distinctly monophyletic taxa are the genus Parastephanellus and the subfamily Foenatopinae (longest branch in the tree above), which is comprised mostly of Oriental species, described in the literature as Foenatopus, Diastephanus or Neostephanus. The latter two genera are distinctly not monophyletic, and I propose that they be synonymized with Foenatopus.
In contrast, some "good" characters show such dimorphic states that they cannot be objectively ordered, making coding a true guessing game. To leave them as "unordered" is often to assume wild evolutionary possibilities, that is, the possibility of character state "X" to change into a dramatically different "Y" in a single step.
The selection of the characters used in the analysis was a very complicate issue. The problem was that the more taxa I included, the more "indefinite" the caracter states became, to the point of practically blending in a continuum. When that happened, there was no way of objectively defining discrete states, and the character had to be abandonned. I had to reject 43 "promising" characters because of this problem, and work very hard to convince my "skeptical side" about the other 175.
In other words, most stephanids look pretty much the same, but when examined closely, an extreme morphological variability soon becomes obvious. The variation in cuticular sculpture, for example, can be easily used to recognize individual specimens. This situation is reflected, I believe, in the fact that over 95% of the described species of Stephanidae are based on single specimens. Not surprisingly, this morphological variability makes subjective interpretations of genera and species very difficult and often contradictory.
Due to the first problem, i.e., the high morphological uniformity, it is very difficult to visualize "groups" within Stephanidae. For example, by the end of my Master's degree, I was convinced that two North American species were somehow related to the Australian species of Parastephanellus, since they shared nearly all important characters I knew by then. It was only after an arduous work with the Australian species that I could finaly detect the apomorphies for the genus Parastephanellus, and securely isolate it from the North American species. Also, it was only after this study that I was able to notice that the Australian species of Parastephanellus form a "unit" within this genus, which is different, for example, from the New Guinean species and the New Caledonian species of the same genus. Thus, the decision to use representatives from these three "subgroups" of Parastephanellus could only be made after a lengthy and difficult revision.
To simply include "as many species as possible", or "all that look somewhat different", would not be a good idea, because the high variability of subtle characters makes nearly every stephanid specimen an unique morph. It seems more accurate to work with species from distinct biogeographical regions first. They had to be carefully examined, and well understood, and only then a truly representative set of species could be selected.
Outgroup Analysis
There are no obvious sister groups for Stephanidae, and defining the phylogenetic position of this family among the Hymenoptera is not an easy task. I was initially considering representatives of the following taxa as outgroups:
This work, however, proved much more time consuming than initially planned, particularly because of the great difficulty in finding or establishing homologies between Stephanidae and the diverse outgroup taxa mentioned above. I did evaluate the relationships of stephanids with other Hymenoptera, but based on a combination and reanalysis of literature data.
This page last updated: February 14, 2000.