No. 418.] NESTS OF AMERICAN ANTS. 799 
would be more than outweighed by the difficulty such a colony 
would experience in obtaining slaves! And even if the dulotic 
habit had manifested itself repeatedly as a chance variation, 
and had proved useful, there would still be lacking the heredi- 
tary basis for the instinct, since this is exhibited only by the 
workers. Wasmann states the difficulty in the following words 
(91 p. 236): ‘ Before the inclination to rear slaves had proved 
itself permanently beneficial during many generations, it could 
not have been established by natural selection as an hereditary 
Anlage; but the possibility of inheriting the Anlage must 
exist before the incipient inclination could be transmitted from 
one colony to another — ergo natural selection lacks the very 
point of departure for the development of an hereditary slave- 
making instinct from the accidental forms of mixed colonies. 
We must leave natural selection like Baron von Münchhausen 
to drag itself out of the morass by its own hair." 
Surely natural selection, however numerous its shortcom- 
ings, deserves better treatment at our hands. A careful 
perusal, however, of the above-quoted passage from the 
“Origin” and a consideration of the facts brought to light 
since its publication leave little cause for anxiety. Wasmann 
assumes that sanguinea robs the pupæ of other ants for the sake 
of rearing slaves. This is scarcely borne out by the facts. 
The young of the auxiliary species are sought for and appro- 
priated to serve as food, in obedience to an ancient and wide- 
spread formicid instinct that emerges to view very clearly in 
many often distantly related species of ants. Thus Adlerz 
has shown (96a) that on exceptional occasions even Lasius 
niger robs the larve and pupz of Z. flavus, and these may 
hatch and function as slaves in the nest of the dark-colored 
species. Wasmann, too (99a), has observed a colony of L fuli- 
ginosus appropriating the larvæ and pupæ of a neighboring 
colony of Myrmica levinodis. I have repeatedly observed the 
same instinct in Mexican and Texan Ecitons ('01; E. crassi- 
corne, schmitti, pilosum)? It is also probable that many cases 
pæ in their nests for some 
ut even if some of them 
