1945] 
Observations on Rhachiocrema 
111 
cation it further stated that Mann had elevated the “ paradoxa 
group” into the subgenus Rhachiocrema. As has already been 
shown Mann made no such proposal, nor could he have done so 
for the description of Rhachiocrema was published two years 
before the “ paradoxa group” first appeared in print. It is easy 
to dismiss Emery’s mistake as a slip resulting from last minute 
alterations. It is not so easy to dismiss the consequences of the 
error. Regardless of what Emery believed and whether he in- 
tended to do so or not he originated the concept that all the 
members of his “ paradoxa group” belong to the subgenus Rha- 
chiocrema. There is clear proof that one eminent myrmecologist 
has accepted this view. When Menozzi published a key to the 
Malayan and Papuan species of Orthocrema in 1935 (5) he 
omitted the species in the “ paradoxa group.” The lack of dis- 
sident opinion in the literature indicates that the acceptance has 
been general. Yet the heterogeneity of the “ paradoxa group” is 
evident from Emery’s own work. During the course of his 
studies on New Guinea ants he described or redescribed each of 
the four species which he later built into the “ paradoxa group.” 
Three of them he figured as well. It is scarcely conceivable that 
Emery was unaware of their marked dissimilarities. These are 
even more evident when specimens are available for comparison. 
In the material sent from New Guinea by Mr. Levy were speci- 
mens of paradoxa , polita and irritabilis var. le-guilloui. The re- 
maining species in the group, emeryi , was not represented. The 
structural contrast between paradoxa and the species irritabilis 
and polita is striking. For the purpose of this study only two 
characters will be considered. In paradoxa the antennal scapes 
project beyond the occipital margin by at least one-quarter of 
their length and the funiculus is provided with a distinct two- 
jointed club. The huge epinotal spines are more than half as 
long as the thorax. The base of each spine is stout and cylin- 
drical and, although they arise at the angle between the basal 
and the declivious faces, their bases are so large that the angle 
itself is virtually obliterated. There is no infraspinal area in 
the usual sense of the term because the bases of the two spines 
involve the entire width of the thorax. In irritabilis the antennal 
scapes barely reach the occipital border; in polita they exceed 
it by an amount less than the greatest thickness of the scape. 
In both these species the funicular club is not distinctly two- 
jointed. The antepenultimate joint is enlarged so that, as Emery 
