148 SISTEE ANTS EECOGNISED 



1879, divided it into halves, which I will call A and B, 

 so that there were approximately the same number of 

 ants with a queen in each division. At this season, 

 of course, the nest contained neither young nor even 

 eggs. During April both queens began to lay eggu. 

 On July 20 I took a number of pupse from each division 

 and placed each lot in a separate glass, with two ants 

 from the same division. On August 30 I took four 

 ants from the pupse bred in B, and one from those in A 

 (which were not quite 80 forward), and after marking 

 them as usual with paint, put the B ants into nest 



A, and the A ant into nest B. They were received 

 amicably and soon cleaned. Two, indeed, were once 

 attacked for a few moments, but soon released. On 

 the other hand, I put two strangers into nest A, but 

 they were at once driven out. For facility of observa- 

 tion I placed each nest in a closed box. On the 31st 

 I carefully examined the nests and also the boxes in 

 which I placed them. I could only distinguish one 

 of the marked ants, but there were no dead ants either 

 in the nests or boxes. 



I carefully examined the box in the same way for 

 several successive mornings, but there was no dead 

 ant. If there had been I must have found the body, 

 and I am sure, therefore, that these ants were n«it 

 attacked. 



Again, on August 31 I put two more of the anta 

 which had emerged from the pupse taken out of nest 



B, and nursed by ants from that nest, into nest A at 



