have known what the long-winged ancestors of the group looked like. The experimentally proved bisexual generation of erinacei is an insect with fully developed wings (figs. 1–2), and other characters typical of long-winged Cynips. One can hardly agree with the literature in which these two generations have been placed in different genera. Both of them must represent true Cynips. Here is evidence that long-winged and short-winged species may be nearer than closely related species, for they may be alternate generations of one species.
In several of the cases cited above, we were delayed in our recognition of relationships by the peculiar hypopygial spines which, we have already mentioned, are found nowhere but among short-winged gall wasps. In connection with a generic re-arrangement which we are undertaking for the whole family Cynipidae, we have found that the hypopygial spine is one of the most constant of generic characters among the long-winged insects. This is confirmed by preliminary studies of more than five hundred species of Cynipidae, and the drawings for such long-winged subgenera as Cynips, Besbicus, and Atrusca in the present paper will illustrate our point. For this reason, we were at first reluctant to believe that shortwinged Acraspis or short-winged Philonix, with distinctive spines, were not genera of nearly as ancient standing as previous classifications had indicated. On the other hand, there is one case of an insect, apache, which is obviously a short-winged Acraspis altho it, as far as we can determine from inadequate material, has the same spine (fig. 402) as the long-winged Cynips acraspiformis (fig. 400). Final evidence, however, is to be drawn from such a long-winged, bisexual generation as we have just described for Cynips erinacei where the short-winged agamic form has the peculiar spine (fig. 420) of a short-winged species of Acraspis, and the long-winged bisexual form has the typical spine (fig. 406) of a long-winged species of Cynips. One must conclude that the form of the spine, while thoroly diagnostic among long-winged Cynipidae, is liable to modification among short-winged forms.
In verification of these conclusions, an inverse application of our procedure was called for in the case of Cynips fulvicollis whose small, round, downy galls (fig. 228) are common everywhere in the northeastern sector of the United States.
