COLE. — IMAGE-FORMING POWERS OF VARIOUS TYPES OF EYES. 353 



appeared very reluctant to leave the runway, often drawing back com- 

 pletely after the anterior end had been extended a centimeter or two. 

 It was only by stimulating such individuals at the posterior end that 

 they could be induced to continue crawling. 



As it seemed possible that the factor of thigmotaxis might have 

 vitiated in large measure the results, the second and third sets of 

 experiments {B, Cf) were made under somewhat different conditions. 

 The ground-glass plate, instead of being laid upon the table, was sus- 

 pended about 2 cm. above it by means of strings leading upward to a 

 common point, and thence to the ceiling in much the manner devised 

 by Parker and Arkin (:oi, p. 1 ol). In this way the horizontal position 



Figure 4. Apparatus used in later experiments witli eart/nconn, and with 

 the land planarian, menhcorm, snail, and slug, d, d', diaphragms or screens of 

 heavy pasteboard ; P, ground-glass plate ; v, worm. 



of the plate was maintained, but it could be moved in any direction in 

 a practically horizontal plane, and could be rotated freely about its 

 vertical axis. The worm could now be allowed to crawl freely on the 

 glass plate. By means of the opaque screens d and d' (Figure 4), it was 

 kept out of the iufl lence of the light until it had straightened out and 

 was moving well, whereupon, by revolving the plate if necessary, and 

 moving it horizontally in the right direction, the worm {u-) could be 

 brought into the normal position, with its anterior end projecting about 

 a centimeter beyond the directive axis, i. e., into the lighted area. In 

 this way it crawled out into the lighted area at right angles to the 

 direction of the rays of light, and without the disturbing thigmotactic 

 influence of the sides of the glass runway existing in the previous 

 experiments. "When once in position, the glass plate was held station- 

 voL. xm. — 23 



