74 Chapter II. 



action according to their perception of this deficiency. 

 It is true, the smaller number of slaves in more popu- 

 lous san guinea nests may partly be due to another 

 circumstance, viz., that in the latter a greater per- 

 centage of robbed pupae is eaten than in smaller nests. 

 Nevertheless, this circumstance is far from explaining, 

 why in the weakest colonies of this marauding ant the 

 number of slaves even exceeds that of the masters. 

 The only way of accounting for this fact is to assume 

 that these colonies try to strengthen their forces by 

 the greatest possible number of auxiliaries. 



To perceive this necessity of increasing their 

 numbers does not go beyond the limits of the 

 instinctive powers of ants. A very interesting case of 

 this kind was observed by me both in the summer and 

 the fall of 1898 in my artificial nest of F. sanguined, 

 already repeatedly referred to. During my absence, 

 extending over several weeks of July and August,, the 

 nest had been badly cared for and had repeatedly dried 

 up; consequently many workers of sanguinea and the 

 greater number of the old slaves had perished. Mean- 

 while new auxiliary ants (F. runbarbis) had been 

 reared from cocoons which I had given to the 

 sanguineas. Now, I observed in the course of Sep- 

 tember, that new clusters of eggs, laid by the two 

 queens, made their appearance in the nest, and that a 

 number of larvae were reared, some of which were 

 conspicuous for their rapid development. This is an 

 exceptional case, generally not occurring in sanguinea 

 colonies during autumn. In October and the first part 

 of November I was absent again. On my return I 

 found, to my great surprise, that in place of the two 



