LOCOMOTIONS OF SURFACE-FEEDING CATERPILLARS. 147 



be beneath the upper portion of the incline and elevate the an- 

 terior portion of its body to make the exploring movements which 

 invariably accompany near contact with a change in its environ- 

 ment, it would be almost certain to touch the incline, and its head 

 would inevitably be pointed towards the upper edge of the in- 

 cline. Should the caterpillar approach the incline where one of 

 its edges touches the maze and make the usual exploring move- 

 ments what would happen? If it moved its uplifted body in- 

 wards it would come in contact with the incline and its head 

 would be directed towards the upper edge of the incline. Should 

 it move its body outwards, since the incline descends abruptly, 

 the caterpillar would come in contact with nothing and would 

 then wave its body inward. Should the sweep be sufficiently 

 long, the caterpillar would surely come in contact with the in- 

 cline and its head would be pointed towards its upper edge. 

 This suggests that it is not necessary to predicate a negative 

 geotropism to account for the upward movement of the cater- 

 pillar on its first contact with the maze. It simply moves off 

 in the direction that its head happens to be pointing when it 

 encounters the incline. 



To test the soundness of this conclusion, the following simple 

 experiment was devised. A caterpillar was placed in a long nar- 

 row test-tube; which was sufficiently wide to permit freedom of 

 longitudinal locomotion, but which was so narrow as to make it 

 inconvenienient for the insect to turn around or make exploring 

 movements. This device made it possible to induce the cater- 

 pillar to come in contact with the incline with its head pointing 

 in any desired direction. It was only necessary to wait until 

 the caterpillar was moving freely along; and then, after first 

 pointing the mouth of the test-tube in the desired direction , to 

 bring it in contact with the upper surface of the incline. In more 

 than 99 per cent, of the cases, the insect moved off in the direc- 

 tion its head was pointing. 1 By means of this device hundreds of 

 caterpillars were induced to move downward, upon their first 

 contact with the incline. 



This also explains why a caterpillar approaching the vertical 



1 In the few exceptions, I had waited too long and the caterpillar had curved 

 its head around the rim of the test-tube before I had brought it into contact with 

 the incline. 



