POWERS OP COMMUNICATION. 167 



the larva? ; but only one passed over the little bridge to 

 the larva?, while 1 5 went over the bridge L to I. On 

 repeating this experiment with another marked ant, she 

 herself made 40 journeys, during which 19 other ants 

 found their way to the point of junction. Only 2 went 

 over the little bridge to the larvae, 8 went over L to I, 

 and the remainder on to D. 



In another similar experiment the marked ant made 

 16 journeys; and during the same time 13 other ants 

 came to the point of junction. Of these 13, 6 went 

 on to D, 7 crossed over L to I, and not one found the 

 larvae. Altogether, out of 92 ants, 30 went on to D, 

 51 crossed over in the wrong direction to I, and only 

 1 1 found their way to the larvae. 



From January 2 to January 24 (1875) I made a 

 series of similar observations; and during this time 56 

 ants came in all. Of these, 20 went straight on to 

 D, 26 across the paper to I, and only pj g> 4 

 10 to the larvae. 



This, I think, gives strong reason to 

 conclude that, under such circumstances, 

 ants track one another by scent. 



I then slightly altered the arrange- 

 ment of the papers as shown in the 

 accompanying diagram (fig. 4). A, as 

 before, is the nest, o being the door. 

 B is the board ; h is a glass on which 

 are placed the larvae ; m is a similar glass, but empty ; 

 n a strip of paper : to the end of n are pinned two 



