14 
Psyche 
[March 
namely the Myrmeciinae, Pseudomyrmecinae, Dolichoderinae (in- 
cluding Aneuretini) and Formicinae. (See the arrangement of ant 
subfamilies into the “myrmecioid” and “poneroid” complexes by 
Brown, 1954). Where does this leave the poneroid subfamilies in 
the scheme of things ? Among the poneroids, including the Ponerinae, 
Dorylinae, Leptanillinae, and Myrmicinae, the most primitive group 
is certainly the Ponerinae, and most particularly the ponerine tribe 
Amblyoponini. It is disconcerting to find that the ponerines are 
hardly closer to Sphecomyrma than they are to the primitive myrme- 
cioids. There is no trace of the ponerine gastric constriction in the 
latter two groups. More significantly, the Amblyoponini have an 
incompletely constricted petiole ( that is, the petiole is broadly attached 
posteriorly to the gaster), and they also have the constriction be- 
tween the first and second gastric segments. Sphecomyrma lacks the 
gastric constriction, but has the petiole strongly constricted behind. 
We have long considered the amblyoponine petiolar form as primitive 
among the ants, and it certainly resembles that of some tiphioids as 
much as or more than it does that of most other ants. 
One character that has never received any particular attention is 
the form of the amblyoponine male mandibles (Brown, i960: figs. 8, 
26) which is in fact now seen to be quite primitive. In the species 
of Amblyoponini so far described, the mandibles are narrow and 
wasp-like, sometimes bidentate, and sometimes tapering to a single 
acute point (the latter condition is evidently derived) ; they close 
tightly against the convex free clypeal margin, as do those of most 
wasps. (The more elaborate worker-female mandibles of Amblyo- 
ponini are not too difficult to imagine as derived from the bidentate 
wasp-like form.) When considered together, the petiole and the 
male mandibles of Amblyoponini certainly strengthen the general 
impression that this tribe is very primitive, although it shows tenden- 
cies toward specialization for life in cryptic habtats. 
If the amblyoponine petiole is truly more primitive than the 
Sphecomyrma petiole, then the split between myrmecioids and pone- 
roids must have come at a time when ants were still very wasp-like, 
and perhaps we should even consider the possibility that the diver- 
gence occurred before these groups had fully acquired their sociality. 
In the light of this last possibility, the metapleural glands assume a 
particular importance in our phylogenetic speculations. The fact 
that they are such complicated organs makes it unlikely that they 
were evolved independently in different ant lineages. If their func- 
tion — at present unknown — is eventually shown to be primarily 
