402 The American Naturalist. [May, 



opaque object and crawl towards the lighted end, but as soon as they 

 emerge from the obscured region they turn about with the head 

 towards the window and remain thus at the boundary of obscured and 

 illuminated parts of the tube — not scattering through the latter part. 



By leaving a small area uncovered along the upper side of the 

 obscured part of the tube when it is turned to the window, the cater- 

 pillars continue on into this region as if it were not covered at all, 

 though passing towards the window from a more generally illuminated 

 part of the tube into one only illuminated from a narrow line. They 

 move along the upper side with ventral aspects towards the light. 



All the experiments were conducted in diffuse daylight but succeed 

 the same, though more rapidly, in direct sunlight. One other experi- 

 ment upon the relation of movement to direction of light is this: the 

 test tube with animals at one end, a, is put on the table with this end 

 away from the window but in a strong beam of sunlight coming 

 obliquely from the window to strike the test tube nearly at right 

 angles. Now the animals are in the strong sunlight, but, directed as it 

 seems, by the diffuse light coming from the other end of the tube, b, 

 nearer the window, they all wander toward that end, b. Modifying 

 this so that the end, b, is in the sunlight while the end, a, is away 

 from the window and in diffuse light the animals still go toward the 

 window though now leaving the lesser illumination for the more 

 intense one. 



To determine the action of light of various refrangibility the tube is 

 covered by colored glass. Thus if blue glass cover the entire tube the 

 animals act as if white light were used, but under red glass they react 

 very slowly indeed. The same results are obtained with colored solu- 

 tions instead of glass. 



Covering one end of the tube, (either end) with blue glass produces 

 no visible result other than that observed when white light was 

 present all along the tube, but if red glass be used the animals move 

 just as if the red region were the obscured, not illuminated region, oi 

 the first experiments. When the window end is covered by red and the 

 room end by blue, the animals collect at the boundary between t e 

 two when the tube is perpendicular to the window, but when parallel 

 to the window they distribute themselves all along the blue region. 



It is thus the more strongly refrangible rays which are effective, t e 

 less refrangible rays having little influence. 



A certain intensity of light is necessary in these expel 

 in the evening the light gradually ceases to produce ni< 

 artificial light may be intense enough to act like sunlight. Moreover 



