BUCKINGHAM. — DIVISION OF L.\BOR AMONG ANTS. 



441 



nish. I have never succeeded in making this method work well, since 

 the paint often falls off or is removed by the ant or its companions. 

 Especially is this likely to happen in the course of experiments extend- 

 ing over considerable time, as mine have done. I have had no better 

 success with trying to aflix other substances to the ants. The following 



FlGUKE 5. 



method, however, works well with species of large ants : Pieces of col- 

 ored sewing-silk were split into their component strands. One or, if 

 different colors were to be combined, two or more of these were tied 

 with a single knot into a loop. The legs of the insect were then 

 grasped by the thumb and forefinger of the left hand, while with the 

 right hand the loop was slipped over the abdomen, adjusted imme- 

 diately back of the last pair of legs, and fastened by completing a 

 sc^uare knot, care being taken not to make the loop around the ant too 

 tight. A small pair of forceps is indispensable in adjusting and tying 

 the loop. Provided that the thread is sufficiently slender, I cannot 



