BEHAVIOR OF ANTS 415 



she again moves off, but in the same direction which she took 

 on leaving the nest. She may again halt and explore, and 

 the movement away from the nest in the original direction in a 

 straight or sinuous trail, with occasional exploratory excursions, 

 may be repeated several times. If food is found or if the ant's 

 efforts are fruitless, she returns to the nest, adopting the re- 

 verse direction and omitting the exploratory movements, but 

 never passing along the outgoing trail even w^hen this is en- 

 countered. As a result of this behavior the ant finally lands 

 at one side of the nest entrance. Then she begins to turn around 

 till she recognizes the vicinity of the nest and finally enters it. 

 This turning, which is very characteristic, is called by Cornetz 

 " le tournoiement de Turner," and is distinguished from the 

 other exploratory movements ("tournoiements de recherche 

 de provende"). The straight or sinuous portions of the trail 

 following the original direction from the nest and reversing 

 it on the return are called "trajets de Pieron." Cornetz, there- 

 fore, rejects the conclusions of Bethe and others to the effect that 

 the individual ant depends on her odor trail, and agrees with 

 Pieron in believing that on her return trip she depends on her 

 kinesthetic or muscular movements; but he maintains that 

 these can give only a rough indication of the amount of terri- 

 tory covered on the outward trip. The maintenance of a given 

 direction, even when interrupted by exploratory excursions, 

 both on the going and returning trails, must be ascribed to an 

 inner sense. The ant behaves, in short, "as if she contained 

 a compass." The kinesthetic sense acts merely like a roughly 

 constructed podometer, giving the insect a vague notion of the 

 distance she has traveled away from the nest. That the homing 

 ant follows a direction and not a trail is shown by her main- 

 taining this direction even when she is lifted from her path 

 by the wind or by the experimentor, and deposited some dis- 

 tance away, even if this be on the opposite side of the nest 

 opening (Pieron's experiment). Cornetz does not, of course^, 

 exclude the ant's recognition of landmarks by means of sight, 

 smell, or touch, or her resorting to the odor of a well-established 

 trail under certain conditions, especially during the "tour- 

 noiements de Turner," but these are not the determining factors 

 in the first developmental stages of the trail. The territory 

 covered by ants of the superior genus {Myrmecocystus) in 



