528 PROCEEDINGS OP THE ACADEMY OF [Oct., 



the ant the direction toward or away from the nest. He caused 

 ants to make a path across a board, a section of which could be 

 made to revolve 180°. When the section was reversed, an ant 

 reaching the section from either direction w'as unable to directly 

 proceed. If the section was not reversed until the ant was upon 

 it, the ant AYould continue its way across the section, but on 

 coming to the end of the section would stop and give evidence of 

 having lost its trail. 



Perhaps by reversing the section the continuity of the indi- 

 vidual trail was broken, the ant being therefore unable to directly 

 proceed. My experiments show that the theory of polarization of 

 the scent is untenable. Dot Five, after establishing her trail, 

 made, in thirty minutes, twenty-eight round trips through a. 

 During her absence in the nest, at the end of every ingoing journey, 

 I changed the relative position of at least three centimeters of 

 earth forming her path, and I gradually extended the displace- 

 ments as far as from a^ to a^, being careful to make the road 

 level and quiet before her return. Not once did she hesitate or 

 wander during all these journeys, although in the twenty-eight 

 stirrings every millimeter of her trail had been displaced. Other 

 ants were equally able to follow their trail over displaced portions 

 of their path. 



I also caused many marked ants to make trails over strips of 

 wood covering the floor of the runs, and after the trails were 

 established I turned the strips end for end, being careful to 

 replace them so exactly that the continuity of the individual trail 

 should not be broken. In no case did an ant give sign of having 

 lost her trail. The advance of the ant over the reversed trail was 

 unhesitating even when the reversed strip was so much as eight 

 centimeters in length. 



A layer of washed earth* a millimeter or two in depth may be 

 sprinkled upon the trail without destroying it, as does a thicker 

 layer. The thin layer is doubtless pervious to the scent. The 

 trail may also be moistened for a distance of several centimeters 

 without destroying its continuity. Bui, as Bethe has pointed out, 

 the ant can be thrown off its trail by wiping the floor on which the 

 scent is laid, or by covering it with a strip of wood or of paper 



^ Earth taken from any Stenamma fulvum pieeum nest, well washed in 

 running water and then dried. 



