ANT COMMUNITIES 



tree. There the column made a detour of about six 

 inches. But an abandoned path continuous with the 

 main road, which apparently had been used at a recent 

 date, was traced for a considerable distance farther. 



Road No. 3 was the longest, being more than a hun- 

 dred feet long. It extended for nearly twenty feet in a 

 straight line, at which point it touched an old stump 

 that deflected it at a slight angle. Thence it was 

 continued in a nearly direct course as far as a beaten 

 footpath through the wood. Here the ant trail was oblit- 

 erated by passing human feet, although the ants still 

 thronged the pathway, there much broadened by the con- 

 tinual interference and loss caused by foot-passengers. 

 The trail, however, was resumed at a point nearly oppo- 

 site that at which it touched the path, and was con- 

 tinued again in a straight line about twenty feet farther 

 to the tree where it ended. 



When the entire trail (No. 3) was staked off, it was 

 found that its terminus deviated less than three feet from 

 the straight line drawn from the point of departure at the 

 ant-hill. The greater deviation in this case was doubt- 

 less caused by the peculiar difficulties in the track. 

 The three roads so radiated from the parent nest that 

 they were included within about one quadrant of a 

 circle, of which the two shorter trails might represent 

 radial boundaries of the quadrant, while the longer 

 trail was midway between the two. 



Looking simply at the results of these observations, it 

 is manifest that these rufous ants showed an accurate 

 sense of direction in marking out and following their 

 approaches to the trees. It would not be reasonable to 

 attribute such mathematical accuracy as above shown 

 to mere accident. The roads in point of directness were 



60 



