8 
William Mouton Wheeler. 
nisms can not be readily trapped and overwhelmed by liquid resins. 
But this kind of selective action, kas little bearing on our problem, 
unless we suppose what is not impossible, tkat large and powerful 
Ponerince and Camponotince — insects comparable to the Australian 
species of Myrmecia, the Brazilian species of Paraponera and Dino - 
ponera or the Malayan Camponotus gigas — may have lived in the 
amber forests. The differences of habit, however, are certainly more 
important. The existing Ponerince are n early all wary, terrestrial or 
even hypogseic ants, which rarely or never climb trees, but seek their 
insect prey on or under the surface of the g round, and there is every 
reason to believe that the early Tertiary species had the same habits. 
As Emery maintains 1 ), this would readily account for the small number 
of individuals of this subfamily in the material examined. The Myr- 
micincBy too, are largely terrestrial, although several genera, such as 
Sima , Monomorium, Leptothorax and Crematogaster are very largely 
or entirely arboreal. Sima and Leptothorax are, in fact, represented 
by a fair number of species in the amber, though the number of indi- 
viduals is scarcely as great as we should expect. Crematogaster is 
entirely wanting, though from its present cosmopolitan distribution 
we should certainly expect it to be present. The same is true of 
Pheidole. It would seem, therefore, that the absence of these and many 
other common Myrmicine genera, which in all probability existed as 
far back as the Lower Oligocene, must be due to their never having 
invaded the Baltic region rather than to the selective action of the 
liquid resin. The terrestrial habits of many other genera, such as 
Erebomyrma ) Stenamma , Aphcenogaster,Myrmica etc., sufficiently account 
for their small individual representation. Undoubtedly the prepon- 
derance of the Dolichoderince and Camponotince , which together constitute 
nearly 97 °/o of all the specimens, is due to the highly arboreal habits 
of these ants. The singulär disproportion between the individual re- 
presentations of these two subfamilies is brought about by two species 
of Dolichoderince , Iridomyrmex goepperti and I. geinitzi , the former 
represented by 5428, the latter by 1289 specimens. If we subtract 
the sum of these (6717) from the total number of Dolichoderince (7508) 
we have left only 791 individuals, which is certainly much nearer the 
modern ratio of Dolichoderince to Camponotince in a tropical forest. 
The absence of one whole subfamily of ants, the Dorylince, from the 
Baltic amber, is still to be accounted for, since we can hardly suppose 
that this group, which is nearly as primitive as the Ponerince , was 
h Le Formiche dell’ Ambra Siciliana, etc. loco citato p. 586. 
