If^OT] Jennings. — Behavior of the Starfish. 157 



that, in these first series of experiments, it was uncertain whether 

 the results mi^ht not be thoroughly negative, and the work prac- 

 tically thrown away. 



Our problem is, to train the starfish when righting itself to 

 use a certain pair of rays with which to grasp the bottom and 

 pull itself over. The rays thus "used" are those employed like 

 the pair to the right in fig. 13. (Of course all the rays are used 

 in one way or another; we shall confine attention to those used 

 for attaching and pulling over.) 



It is first necessary, of course, to determine whether the 

 animal has any predilection for the use of certain rays before 

 training is begun. This, as we have already seen (p. 144), is 

 actually the case. Of 95 experiments with a number of speci- 

 mens, the ray e was used 89 times, the ray b but fourteen times. 

 Out of the 76 cases in which a definite pair of adjacent rays 

 were employed, a pair containing e was used 64 times, a -{- e, 37 

 times; (/. -\- e, 27 times. On the other hand, the pair b -{- c was 

 not used once in the entire 95 experiments. 



Thus in endeavoring to train the starfish to turn on a certain 

 pair of rays, it will be necessary to avoid a -{- e and e -{- d, and to 

 select some pair on which the animal has no previous tendency 

 to turn. Selection of the pair b -\- c will evidently present the 

 most difficult test. 



1 shall give a rather full account of the methods and results 

 in the ease of the two starfish that were under training for a 

 long period, as this will bring out many essential points. 



Starfish "A." — The first individual with which I worked I 

 called starfish "A." Thirty test experiments with this indivi- 

 dual showed, as we have seen on page 143, a strong tendency to 

 use the rays e and a in the active work of turning, e being em- 

 ployed in 25 cases, a in 18 cases. The precise combination a -\- e 

 did the active work in 15 out of the thirty cases. On the other 

 hand, there was a strong disinclination to use the rays b and e ; 

 b was employed but twice and c only 5 times in the thirty ex- 

 periments, while the pair b -\- c was never used in combination. 



Will it be possible to train the starfish to employ by pref- 

 erence this combination b -\- c, Avhich it never uses before train- 

 ing? Experiments with this in view were begun July 21, 1906. 



