1956] 
Creighton — Ephebomyrmex 
55 
so feebly rugose as to be almost smooth and is marked 
only by scattered patches of punctures. In this same 
species the dorsum of the first gastric segment was said 
to be smooth and shining and without basal punctures. 
In townsendi, on the other hand, both the postpetiole 
and the dorsum of the first gastric segment were de- 
scribed as completely covered with fine, dense punctures 
and opaque. Wheeler also claimed that in general town- 
sendi is more opaque than imberbiculus. Olsen accepted 
townsendi as a separate species in 1934 (2) but the 
writer in 1950 (3) treated it as a subspecies of imber- 
biculus. At that time there was little material of town- 
sendi available but, since it was still the only member 
of the subgenus known from Mexico, it could be regarded 
as a southern race of imberbiculus. It is now clear that 
such a treatment is erroneous. As material began to 
accumulate it became apparent that imberbiculus occurs 
much further south in Mexico than does townsendi and 
that the latter form occurs at random, principally in the 
northwestern portion of the range of imberbiculus. The 
writer Pas taxen colonies wmcn agree wen witn tne type 
of townsendi in northern Sonora and at three stations 
in southern Arizona. In these specimens both the post- 
petiole and the basal two-thirds of the dorsum of the 
first gastric segment are densely and evenly covered with 
small, close-set punctures which render the surface opaque 
or nearly so. These specimens can, without any diffculty, 
be assigned to townsendi, but this is not true of others 
which are intermediate in character. One nest from 
northwestern Chihuahua has the gastric punctuation 
limited to the anterior third of the segment. Another 
large colony from southeastern Arizona shows some in- 
dividuals without gastric sculpture and others with only 
a small area of sculpture immediately adjacent to the 
postpetiole. In both these colonies the sculpture of the 
postpetiole is like that of townsendi. After it was ap- 
preciated that the supposedly definitive features of town- 
sendi vary, a reexamination was made of all specimens 
previously assigned to imberbiculus. This established the 
