156 University of California Publications in Zoology. [^ou 4 



mentioned have fulfilled the first two requirements, but not the 

 third one. 



Method of training the Starfish. — It occurred to me that the 

 difficulty might perhaps be met by interfering directly and con- 

 tinuously in the course of the reaction and preventing the star- 

 fish from succeeding save by one method. For this the righting 

 reaction was selected. The starfish was placed on its back, then, 

 by holding the rays with the hands or forceps, releasing them 

 from their attachment, etc., etc., the starfish was prevented from 

 turning save on a certain pair of rays. The starfish was, as it 

 were, trained to turn on a certain pair of rays. 



It needs to be pointed out that this "training" does not 

 imply anything different in principle from what is found in 

 nature; there is nothing "artificial" about its results. The 

 starfish merely finds in the course of its reaction that it can right 

 itself in only one way, so that we are able to test whether this 

 will result in habit-formation. It is clear that the same results 

 would be attained if this effect were produced in any other way 

 than by the use of the human hand, and conditions might readily 

 arise in nature that would have the same action. 



Experiments in Training the Starfish. 



A thorough, extended course of training was given to but 

 two individuals, while three others were worked with for a 

 shorter period and in a less thorough way. The effect of a 

 brief series of "lessons" was, however, tested on many other 

 individuals. The small number of individuals employed in 

 extended and thorough work was due to the very great amount 

 of time required in the experiments, and to the fact that the 

 successful method of work was not hit upon till near the end 

 of the season. A given starfish may require two to seven minutes 

 for its reaction, and the careful description to be written requires 

 still more time. Then there is the period of rest between suc- 

 cessive lessons. Altogether, by the time five specimens have 

 been given ten lessons, at intervals, the part of the day left for 

 other work is small. Moreover, the work is extremely tedious to 

 the experimenter, so that it is almost necessary to vary it with 

 work of a different character,- — especially in view of the fact 



