238 A. R. MOORE. 



dorsal side upward, and as rapidly as the twisting is accomplished 

 the tube feet secure a hold on the bottom. Next, if not simul- 

 taneously, the arm adjacent to the ventral side of the arm which 

 is obtaining a hold, twists, so that the ventral surfaces of the 

 two arms face each other, and secures itself in the same manner 

 as the first. If A and B have in this way attached themselves 

 to the bottom, inhibiting impulses are sent to C and D. The 

 latter release themselves if they have already seized the bottom, 

 rise ventrally, dragging E which either remains passive or bends 

 dorsally, even catching the bottom with its tube feet in some 

 cases. The righting is completed by C, D and E passing over 

 A and B and attaching. 



This simple and useful method of righting may be modified 

 by (i) inequalities in the length of the arms, (2) injuries to certain 

 of the arms, (3) any initial twist an arm may have due to its 

 position before the animal was laid upon its dorsal side. 



As to (i) short arms are more sluggish than ones of normal 

 length, (2) injury to an arm inhibits the active twisting and 

 seizing of the surface with the tube feet of this arm, (3) if an 

 arm is partially twisted its tube feet reach the bottom more quickly 

 than they otherwise would. As a result we have the following 

 modifications of the normal method of righting. 



1. If four arms are injured, their activity is inhibited and the 

 righting is accomplished by the one uninjured arm. It may 

 force an adjacent arm to coordinate weakly. 



2. If A and C (Fig. i) twist so as to face each other with their 

 ventral surfaces, B receives two impulses, from opposite direc- 

 tions, to coordinate, and therefore does not twist either way but 

 bends under dorsally, allowing A and C to accomplish what A 

 and B did in the normal case. The same result may be brought 

 about by injuring B, D and E. 



3. Sometimes inhibitions are weak and A, B, C and D may all 

 remain attached, C and. B facing ventrally toward A and D. 

 E alone is inhibited and the righting is accomplished by A and B 

 walking backward under C and D. 



I found, as Jennings noted, that in a few cases a starfish per- 

 sistently refused to use a certain arm for initiating the righting 

 movements. In most cases this was clearly due to an injury or 



