1945] Observations on Rhachiocrema 113 
primitive forms in the genus. Mr. Santschi in his recent study 
on the subgenera of Crematogaster has excluded from it certain 
species which show a furrow or a more or less marked impres- 
sion on the postpetiole, on which he founds his subgenus Neo- 
crema. This division which comprises neotropical and malaga- 
sian species does not seem homogeneous to me; this is why I 
have fused it with Orthocrema.’’ 
If Neocrema is heterogeneous neither it nor Orthocrema 
would become more homogeneous when fused. It is unlikely, 
therefore, that Emery’s arrangement of Orthocrema was de- 
signed to secure structural uniformity. On the contrary he 
seems to have deliberately constructed a heterogeneous assem- 
blage for the sake of having all the species which he considered 
primitive in the same subgenus. However desirable this may be 
from a phyletic standpoint it is not sound taxonomy to found a 
group on inconstant characters, which was what Emery did. 
He apparently regarded as primitive the rectangular petiole 
and the entire, globose postpetiole which many of the species in 
Orthocrema possess. But neither of these features holds for all 
the species which Emery included in Orthocrema. In the Aus- 
tralian species frivola the postpetiole is as clearly bilobed as in 
any species of Acroccelia and the petiole is not markedly rec- 
tangular. The shape of the petiole of paradoxa is certainly very 
far from rectangular. Yet both these extreme conditions can be 
reached through species in which the conditions are intermediate 
so that from a phyletic viewpoint the assemblage can be de- 
fended. I believe that the phyletic gain which accrues to Em- 
ery’s arrangement is more than offset by the taxonomic dis- 
abilities which it entails. In the form which Emery gave it the 
subgenus Orthocrema has to be characterized along such gen- 
erous lines that no satisfactory delimitation of the group is 
possible. The recognition of Neocrema and Rhachiocrema as 
valid subgenera relieves Orthocrema of its most incongruous 
species and permits a much better demarkation of all three sub- 
genera. That the three subgenera tend to intergrade is not a 
matter for concern for other subgenera of Crematogaster also 
possess integrading species. 
I propose to treat Rhachiocrema as a valid subgenus even 
though the features which separate it are not as distinct as was 
originally supposed. Reference has already been made to the 
fact that antennal structure cannot be used as a basis for sepa- 
