No. 418.] NESTS OF AMERICAN ANTS. 797 
internally regulated variations. The premises to these conclu- 
sions I cannot accept, since they do not appear to me to be 
indefectibly established. They are based on a rather limited 
study of a few species of highly specialized European ants, and 
cannot, therefore, lay claim to great generality. My own 
observations, still incomplete, to be sure, on several Texan 
ants representing both the most primitive and the most special- 
ized subfamilies (Ponerinze and Camponotinz) convince me 
that worker ants not only very frequently lay eggs in consider- 
able numbers, but that these produce perfectly normal offspring. 
When workers are properly fed in the artificial nests, they 
seem to have no more desire to devour their own eggs than 
to devour those which are deposited in their keeping by the 
queens.. During February last a carefully isolated lot of 
workers and soldiers of a handsome Camponotus (C. texanus 
n.sp.)laid dozens of eggs in my Fielde nests. By the first 
week in June many of the larvae were mature, and a few of 
these had spun their cocoons before I was compelled to leave 
my laboratory for the summer. These cocoons were found to 
contain perfectly normal male pupz, thus adding fresh evidence 
for the generally accepted belief that the parthenogenetic off- 
spring of worker ants are males. Similar observations were 
made on workers of Camponotus marginatus n. Var. and Pachy- 
condyla, except that in these cases I did not follow the larva 
quite to the pupal stage. I am, moreover, convinced that 
numerous eggs are laid (probably by the soldiers) and reared 
in the frequently queenless nests of a gall-inhabiting Colo- 
bopsis (C. etiolata n. sp.) from Texas. In fact, for aught we 
know to the contrary, every well-developed ant colony may 
contain one or more fertile workers. Where the worker caste 
is dimorphic the soldiers probably have the greatest tendency 
to lay eggs. Judging by analogy with other Hymenoptera, 
like Polistes among the wasps, it is also probable that the older 
and more vigorous the ant colony, the greater the tendency for 
workers to take on the reproductive powers of the gecm 
That these conditions clearly imply the possibility of the inher- 
itance of worker characters through the male offspring goes 
Without saying. The comparatively frequent development of 
