o38 PROCEEDINGS OF THE ACADEMY OF [Oct., 



taken from their own E nest, and upon the m and b runs earth 

 newly taken from the nest of aliens, making the earth from the 

 two nests to meet in the middle of the passages, 1, 2, 3, 4. Into 

 c-d I put washed earth. Immediately after such distribution of 

 the earth I put many ants from the E nest into the maze at T 

 upon the pupte pile, and recorded the number of journeys made 

 through each run. Fully half the journeys were made upon the 

 earth from the alien nest. 



I then closed the a and the n runs and sent many ants of colony 

 C over the m and b runs. While the trails were yet fresh, I 

 removed the C ants and their nest, and gave the maze over to the 

 E colony. The E ants in no wise avoided the m and b runs that 

 had just been used by the C ants, but they traversed them as often 

 as they traversed the runs through which no aliens had passed. 

 Variations and repetitions of this experiment gave results always 

 similar. I therefore think that the odor of the ant is discernible 

 to other ants only when it is either perceived upon or is immedi- 

 ately disseminated from the living body of the ant. This \new is 

 sustained by the fact that alien pupre placed in the nest just before 

 they emerge from the pupa-stage are at first accepted by the ants, 

 and are nevertheless often killed as soon as they cease to be inert. 



I have found that the ant's power of perceiving this odor lies 

 in the eleventh segments of her antennre. The contact of these 

 segments with any part of the body of another ant is followed 

 by reflexes denoting either satisfaction or repugnance. When the 

 ant is deprived of these segments by a cut across the tenth seg- 

 ment, she no longer discriminates between friend and foe. De- 

 prived of these segments, marked ants of two or of five colonies 

 lived peaceably together or fought one another with absolute 

 impartiality. Forel discovered that ants of alien colonies ceased 

 from hostile demons1;ration when deprived of the antenn?e; but in 

 my experiments this effect is as complete when no more than the 

 two distal segments are removed. The removal of the twelfth 

 segment alone is not effective, and as the segments telescope each 

 into its proximal neighbor, the destruction of the tenth is necessary 

 for^the complete removal of the eleventh. 



A healthy ant, with or without antenna), will fight a dead ant, 

 kindred or alien, if the dead ant be made to simulate an attack 

 upon the live ant; and an ant will sometimes continue a battle 



