8o 
Psyche 
[March 
in Discothyrea ) stemming from a condition in which the integument 
was more coarsely sculptured, with larger foveae, each fovea bearing 
a hair on a central tubercle. The Old World ( Indo-Melanesian) 
Gnamptogenys stocks include species (e.g., G. menadensis) that meet 
these specifications and are known to be epigaeic, even arboreal, for- 
agers (personal observations), but Gnamptogenys has reduced palpal 
segmentation and other characters that make it more likely a con- 
vergent than an ancestral stock to Proceratium. 
I believe that Acanthoponera is nearer to the ancestral line of 
Proceratium because of the higher basic palpal segment number and 
the traces in some Proceratium species of a median carina on the 
frontal area and vertex of the head, characters of Acanthoponera 
(Brown, 1958:188). But the main question we are led to consider 
is: Has P. avium preserved some version of the archetypal Procera- 
tium sculpture-pilosity pattern, or is its present condition a secondary 
reversion from the fine-scale pattern characteristic of other Procera- 
tium around the world ? 
Possibly we shall never have a clearly definitive answer to this 
question, but we do have one clue pointing toward the reversion 
hypothesis. This clue is the rather unusual eyes of P . avium — a 
rather large, glassy-looking orb on each side, backed by a perceptible 
amount of dark pigment, and relatively larger than the character- 
istically single-facetted “compound” eyes of other Proceratium species. 
I think we have to suppose that the ancestors of P. avium, and also 
of all other Proceratium, had already specialized for a cryptic ex- 
istence to the point where minute single-facetted eyes, probably barely 
enough to sense the difference between light and dark, served ade- 
quately the lifeways of these animals. My guess is that such lifeways 
also had forced selection for the fine sculpture-pilosity pattern of Pro- 
ceratium in the ancestral lineage of this genus. If my reasoning is 
correct, then P. avium is an interesting example of a “character- 
released” species on a remote oceanic island with a depauperate 
endemic ant fauna. 
The known endemic ant fauna of Mauritius numbers only about 
7 species, if we take Donisthorpe’s (1946, 1949) count as a base and 
deduct obviously introduced species, including synonyms, and add my 
collections. The endemics now appear to be restricted to the small 
areas of upland native forest; the cane fields and other culture areas 
are saturated with introduced ant species. In a sense, then, the moun- 
tain forests represent the “real” island (s) of Mauritius as far as the 
endemics are concerned. 
