112 University of California Puhlications in Zoology. ["^o^. 4 



All went directly away from the lighted side, in the direction of 

 the light rays. Replacing them in the iron vessel, each went 

 as before (with one exception) toward that side of the vessel that 

 was nearest it. It appears most decidedly as if the animals were 

 able to perceive the dark sides from a short distance, and to move 

 toward them, even though these sides cast no shadow. 



Now the following experiment was tried. The starfish was 

 placed in the center of a large fiat surface, in the sunlight, far 

 from any object. Then a black plate 12 cm. wide and 16 cm. 

 long was placed at one side of the animal, in such a way that its 

 shadow did not fall on the animal, the plate being about 2 cm. 

 from the tips of the nearest rays. Thereupon these nearest rays 

 were raised and began feeling about, and soon the starfish began 

 to move toward the black plate. The plate and starfish were now 

 removed, then replaced on the flat surface as before, save that 

 the plate now stood in a different direction from the starfish. 

 The animal moved toward it, as before. This was done a third 

 time, with the same results. In another case a starfish was in- 

 duced to move repeatedly toward a black ruler that was placed 

 upright about two cm. from the tip of a ray. 



It will naturally suggest itself to the reader that the proper 

 method of experimentation is as follows : The plate is placed to 

 the right of the starfish. As soon as the animal begins to move 

 toward it, the plate is transferred to the opposite side, in order to 

 determine whether the starfish will reverse its movement and fol- 

 low the plate. While this experiment succeeds in a few cases, it 

 is as a rule doomed to failure from beforehand, owing to a 

 marked characteristic of starfish behavior. When the animal gets 

 started in a certain direction, it tends to persist in that direction 

 without regard to changed conditions,— a matter which we shall 

 discuss later (p. 115). Thus, if the plate is transferred to a new 

 position, the starfish continues to go in the direction in which it 

 has started, until it has gone too far to be affected by the plate. 



It may be suggested that the movement toward the plate was 

 due merely to the disturbance in the water induced by the intro- 

 duction of the plate. There is no doubt but that the starfish does 

 sometimes move toward the seat of such a disturbance. But 

 when the plate was introduced first and the starfish after, the 



