366 LEON J. COLE. 



Nevertheless, at the end of 56 minutes the animal had succeeded 

 in righting itself, though arm b was still crossed over arm a. 

 At the end of one hour, however, from the time it was placed on 

 its back, it had straightened out completely. The turning was 

 accomplished with arms e and a, although c swung over between 

 them and really seemed most instrumental in pulling the animal 

 over. 



In both of these experiments the lack of coordination was very 

 apparent, as mentioned by both Romanes and Loeb. Neither 

 could it be determined that in either experiment a definite 

 impulse toward turning in a given direction was established, as 

 is normally the case. In fact, especially in the second experi- 

 ment, it was noticed several times that as two or more arms 

 happened at the moment to be working in concert, the righting 

 could have been effected easily if another arm had but ceased its 

 efforts in an opposite direction, to say nothing of helping in the 

 same direction. The final turning appeared to be nothing but a 

 chance occurrence when enough arms happened to be pulling 

 in one direction to turn the specimen in spite of its other arms. 

 Thus, although there resulted complete lack of coordination, as 

 Loeb states is the case when the nerve connections are severed, 

 the end result was nevertheless finally accomplished, apparently 

 by accident. The pull of one arm on another does not seem, 

 therefore, to be a sufficient stimulus to induce a definite co- 

 ordination of the tube feet in the second arm. 



The second specimen did not crawl after righting itself. It 

 would be interesting to test whether by exerting a steady pull 

 in one direction on such an "uncoordinated" specimen, a more 

 or less unified impulse to crawl in that direction might not be 

 induced, comparable to the crawling of the severed earthworm. 1 

 This experiment was not tried, but the lack of coordination shown 

 in the righting would make it seem that a negative result might 

 in all probability be expected. 



1 These experiments might be made even more comparable by severing the rays 

 of the starfish completely from the disc and then sewing them in position again 

 with thread. 



