160 University of California Publications in Zoology. [Vol. 4 



than h -\- c, owing to the lack of coordination in the twisting of 

 the latter. 



In experiments 5 to 8, & took hold readily, its original dis- 

 inclination seeming to be quite conquered; it was usually at- 

 tached and began pulling before c had done so. Indeed, h was 

 almost the first of the rays to bend back, twist, and endeavor to 

 attach itself. The ray e, however, continued to make active 

 efforts to attach itself, as it did at first. 



Now, after these eighteen training lessons (ten the first day, 

 eight the second), which had been directed mainly toward induc- 

 ing h to play the part of attaching and pulling, the starfish was 

 allowed to right itself five times in succession without any inter- 

 ference or assistance (there being but the usual two minutes 

 between the last of the ''lessons" and the series of free reac- 

 tions). In every case the ray b at once began to bend down, 

 twist over, and attach itself ; it was the first ray to take hold in 

 most of the five free experiments. In the first of the free trials 

 this ray b turned the starfish practically unaided, though when 

 the turning was nearly finished the ray e took hold and helped 

 a little. In the succeeding four free trials the work was done 

 by b and e in common. In one case e became attached first, but 

 b persisted, took hold also, and played an equal part in accom- 

 plishing the reaction. 



Thus in this starfish A the method of turning had been com- 

 pletely changed by a course of eighteen lessons of training. 

 Before the training b was never employed in attaching and pull- 

 ing, and in the firet lessons it was almost impossible to compel 

 it to take this role. But at the end it spontaneously took this 

 role before any of the other rays, and played the main part in 

 the turning. It is interesting to observe that the ray e, which 

 was at first the ray most strongly inclined to play this part, still 

 retains this tendency; the turning is accomplished by the co- 

 operation of this originally active ray, with the one that has been 

 made active by training. 



It is evident, however, that we have not yet succeeded in 

 our original undertaking, which was to cause the rays b and c 

 to cooperate in the turning. Our training thus far has been 

 directed mainly toward b alone ; it is now necessary to work on 

 the two together. 



