128 University of California Publications i)i Zoology. [Vol.4 



the normal position is reached. (A model of the starfish, in 

 paper or cloth, will make clear the necessity of such movements 

 when three of the rays remain attached.) Five variants of this 

 method are evidently possible, since any three adjacent rays may 

 be used. 



4. Four of the rays take hold, two extending to the right, two 

 to the left. Then the fifth ray (M^hich we may call the posterior 

 one) is lifted straight up and swings directly over till its ventral 

 surface reaches the bottom, while the anterior attached pair walks 

 backward beneath the posterior attached pair, the latter walking 

 forward over the surface of the latter. Nearly this type of turn- 

 ing is shown in progress in figs. 16 and 17. Five variants of this 

 method are possible, depending on what arm is left free. 



4. All of the rays attach themselves (figs. 18, 19). Now the 

 turning can be accomplished only by the release of certain rays, 

 when the method passes to one of the types already described. 



5. An unusual method is that in which but one ray twists and 

 attaches itself, and by its unaided efl^orts turns the starfish about 

 an axis passing through this ray. This may be induced experi- 

 mentally by preventing the other rays from taking hold ; I have 

 seen it occur spontaneously in a number of cases. Of course 

 five variants of this type are possible, depending on which ray 

 is used. 



6. A still more unusual type is seen in the performance of the 

 righting action without attachment of the tube feet of any of the 

 rays. Preyer (1886) and Romanes (1885) have given accounts 

 of certain ways in which this is sometimes accomplished. The 

 typical method appears to be for the starfish to raise its disk high, 

 standing on the tips of all the five rays, then to swing one or more 

 rays over or one or more under, or both, until the body topples 

 over, ventral side down. In my own observations the righting 

 without attaching the tube feet was seen only when these were 

 experimentally prevented from taking hold. The starfish then 

 writhed and squirmed irregularly, taking various bizarre forms, 

 until it had succeeded in getting its ventral side down, when the 

 squirming ceased. No definable method was observable, — save 

 that of making all sorts of movements till the end was reached. 



As is apparent, each of these six methods admits of great 



