SUPPOSED USE OF A PASSWOED. 129 



strangers two were taken into the nest and fifty were 

 thrown into the water. I think it most probable that 

 even these two were subsequently brought out and 

 treated like the rest. 



It is clear, therefore, that in these species, and I 

 believe in most, if not all others, the ants of a com- 

 munity all recognise one another. The whole question 

 is full of difficulty. It occurred to me, however, that 

 experiments with pupge might throw some light on 

 the subject. Although all the communities are deadly 

 enemies, still if larvae or pupte from one nest are trans- 

 ferred to another, they are tended with apparently as 

 much care as if they really belonged to the nest. In 

 ant-warfare, though sex is no protection, the young are 

 spared, at least when they belong to the same species. 

 Moreover, though the habits of ants are greatly changed 

 if they are taken away from their nest and kept with 

 only a few friends, still, under such circumstances, the}^ 

 will carefully tend any young who may be confided to 

 them. Now if the recognition were individual — if the 

 ants knew any one of their comrades, as we know our 

 friends, not only from strangers, but from one another 

 — then young ants taken from the nest as pupse and 

 restored after they had come to maturity would not 

 be recognised as friends. On the other hand, if the 

 recognition were effected by means of some signal or 

 password, then the pupse which were intrusted to ants 

 from another nest would have the password, if any, of 

 that nest ; and not of their own. Hence in this case 



K 



