146 University of California Publications in Zoologij. [Vol. 4 



on these rays will now cause the starfish to turn directly toward 

 the light. Yet it continues to turn on these rays a - b. The 

 direction of the light having once determined the impulse, a 

 change in its direction has no effect. 



This persistence is seen even if the starfish is turned around 

 before any of the tube feet or rays have attached themselves. 

 It is only necessary that the unified impulse toward turning in 

 a certain way shall have appeared, in the inclination of the tube 

 feet. Again, if after the starfish has partly attached itself and 

 has begun to turn in a certain direction, we forcibly release 

 it from its attachment and place it under quite new relations to 

 the outer conditions, it resumes, when again placed on the bot- 

 tom, its original way of righting, turning on that pair of rays 

 on which it had begun to turn. 



This persistence of the impulse is very marked in experiments 

 in which an attempt is made to change the impulse so as to cause 

 the formation of new habits, as described in the next section. 

 Thus, after the starfish has taken hold with a certain pair of 

 rays, if we forcibly release these from their attachment, the 

 animal frequently simply attaches them again, and continues 

 to turn as before. If not too great violence is used, sometimes 

 a given pair of rays may be detached half a dozen times in suc- 

 cession without causing other rays to take hold, or any change in 

 the direction of the impulse. In certain cases the starfish had 

 taken an impulse to turn on a certain pair of rays, a and e. 

 Releasing both of these, the animal continued its efforts to turn 

 in that direction. The two rays were forcibly prevented from 

 becoming attached, but the animal refused to take hold with 

 the other rays and made repeated efforts to bring a and e in 

 contact again. Finally, in spite of my efforts to prevent, it 

 managed to take hold with a, though I managed to keep e from 

 getting hold. But now the starfish turned just as it would have 

 done if both were attached, the unaided efforts of a pulling it 

 over in the usual way. 



Changing an Established Impulse. — Yet of course the im- 

 pulse is sometimes changed after it is once established. Different 

 starfish differ greatly in the ease with which the direction of 

 the impulse can be changed. Some specimens are extremely 



