1901.] NATURAL SCIENCES OF PHILADELPIIIA. 523 



I placed some hundreds of pupre at the T corner, and taking 

 marked and unmarked ants from the nest I, introduced them to 

 the maze through the tube, T. In no case, except when mani- 

 festly lost, did any ant carry^ in a pupa without having first made 

 the journey without a burden. Sometimes an ant traversed more 

 than one run before she began the labor of carrying in, and if she 

 afterward used more than one of the runs she never used other 

 than those she had previously traversed, no matter how many of 

 her fellow burden-bearers were travelling the other routes. If the 

 glass floor was clean she apparently had greater difficulty in estab- 

 lishing her trail than when the floor was covered by Avood or by 

 earth. In the first case, she would go a short distance, as from 

 T to a^ or n'^, and return, and then would prolong her next jour- 

 ney as far as a* or n^, and she might make many excursions from 

 T over the same path before she discovered the entrance to the nest 

 at I. When the floor was covered with earth a single trip often 

 sufficed, and she began to carry in when on her second excursion 

 from T. The structure of the ant's feet adapts them to the partial 

 clasping of particles of earth, and it may be that pressure assists 

 in the deposit of the scent. 



Having begun to carry in, an ant usually continued her work 

 incessantly, making from thirty-seven to seventy round journeys in 

 an hour. No burden-bearing ant ever made a loop in her own 

 trail. The shortest run, c-d, was not oftenest taken, but no ant 

 ever carried in by a route longer than a single run, o, m, c-d, 

 n or h. 



The c-d run virtually counted as but one, for when any ant 

 had made the c or the d passage, she afterward used either side of 

 the central block c with little hesitation. 



There was no evidence that any ant was influenced by the fact 

 that the c-d run was shortest, and that there were certain ad- 

 vantages in following a straight line; but we should not therefore 

 hastily declare that ants have no reasoning power. The Chinese, 

 for reasons ^vell known to themselves, generally make their roads 

 crooked. 



The ingoings and returns of several marked ants during one 



*Tlie ants withdraw their youuo; from currents of air even more quickly 

 than from Hglit. If the ants liuddled upon the pupif, making their own 

 bodies a screen from the light, tbej were impelled by my gently blowing 

 into the tube 2' to search for a tranquil refuge iu the nest. 



