104 The Tracking Instinct in a Tortugas Am. 



dead flies, for if a fly be wounded and moving the ani usually be- 

 comes much excited and proceeds to bite it, and it is remarkable 

 how eflficacious the bites are in quieting the fly. In any event 

 this "finder ant" soon leaves the flies without carrying off any 

 piece of them, but instead of moving off in the erratic and tortuous 

 path it was pursuing before it found the flies, it now goes in a fairly 

 straight path toward some crevice in the floor, out of which there 

 soon pours an excited swarm of its nest-mates, who proceed toward 

 the flies in a fairly straight path, but which is not necessarily identical 

 with that taken by the ''finder ant" in returning from the flies to 

 the nest. Often the path of the return swarm is not quite right in 

 direction, and thus the ants would pass to one side or the other of 

 the flies; but, curiously, when the right distance has been made and 

 the ants are about to pass the flies, the swarm suddenly breaks up 

 into individuals coursing in random fashion in all directions (the 

 "Turner's curves" of authors). It is very dramatic to see the 

 straight path of the once orderly file of ants suddenly break into 

 this random wandering, but so accurately do they gage the distance 

 that I have never seen them miss it by more than 2 inches in a 

 journey of 8 feet, while often the direction may be so much in error 

 that in going 8 feet the ants may tend to pass as much as 4 inches 

 to one side or the other of the flies. Of course, in cases v/hen the 

 path of the main swarm does not happen to "strike" the flies, but 

 passes to one side and then breaks up, many of the ants will not 

 succeed in finding the flies, but must wander in erratic curves over 

 the floor. A considerable number, however, do find the flies, and 

 within a few minutes a fairly straight swarm-path is established 

 between the nest and the flies. 



Cornetz, observing ants in Algiers, finds that when an ant returns 

 to the nest it pursues a fsLirly straight path which is more or less 

 right in direction, but in any event, when the ant has gone the 

 correct distance, it begins to wander in more or less tortuous courses 

 until it finds the nest. 



EXPERIMENTS. 



In studying the behavior of the Tortugas ants I would usually 

 place a few recently killed house-flies upon the floor in order to 

 draw out the ants, so that many of them would be constantly moving 

 over the floor in all directions and the entrance to the nest would 

 be well defined by crow^ds of moving ants in its neighborhood. 

 Often paper was pinned down to the floor in order that the paths 

 of the ants might be drawn accurately in pencil. 



I. When an ant has discovered a dead fly and is engaged in 

 "inspecting" it, we maj^ draw a circular line of a solution of cor- 

 rosive sublimate in 35 per cent alcohol about a foot in diameter 



