34-0 C. H. TURNER. 



thing. Such was not the case ; for although there were about 

 two hundred workers within the nest, never more than six were 

 seen covering the crack at any one time. Sometimes it took the 

 few ants employed several days to cover the cracks in a satisfac- 

 tory manner ; yet in no case did a larger number participate. 



To test the second assumption two different kinds of experi- 

 ments were performed. I shall first describe an experiment of 

 the first kind. When all the ants were resting quietly within the 

 nest the trash was gently removed from crack e (Fig. 2). The 

 ants within the nest immediately became restless. Two of them 

 mounted the ceiling and examined the crack carefully. About 

 fifteen minutes later an ant was observed moving back and forth 

 on top of the nest. On coming in contact with crack e it paused 

 momentarily, examined the crack carefully and then passed on. 

 After it had roamed about the island for a few minutes, I im- 

 prisoned it. About five minutes later another ant from the nest 

 walked across the top of the nest to crack e. After examining 

 the crack carefully with its antennae, the ant began to cover the 

 crack with trash. In a few minutes this ant was joined by a third. 

 This experiment shows that all ants that cross the crack are not 

 stimulated to do the same thing. 



The second type of experiment was quite unlike this. As be- 

 fore, the trash was gently removed Irom crack e producing the 

 same restlessness within. The cover was then removed from com- 

 partment B (Fig. 1) and several worker ants transferred to a 

 beaker. The cover was then replaced on B. 



One of the captured ants, after having been marked with a 

 characteristic color was placed near the middle of the top of com- 

 partment £7 and covered with a transparent white glass cone eight 

 centimeters in diameter. At first the ant would make vigorous 

 efforts to escape from the cone. It would meander over the top of 

 the nest, go repeatedly round and round the circumference of the 

 base of the cone and sometimes mount its sides. In these random 

 movements the ant would of necessity cross that crack many 

 times. After more or less of this fruitless activity, the ant would 

 become quieter. It would then move more leisurely about, ex- 

 amining the crack at frequent intervals or else it would rest over 

 the crack or a short distance therefrom and quietly preen its 



