93 University of California Publications in Zoology. [Vol. 4 



under a unified impulse is well seen in the photograph shown in 

 fig. 13, though in this case the impulse was directed toward 

 righting an inverted starfish, rather than toward locomotion. 

 We shall have occasion to study thoroughly this unified impulse, 

 later, in examining the righting reaction. 



Method of Locomotion in the Starfish; Erroneous Current 

 View. — There is a further mistaken conception expressed in the 

 accounts commonly given of the locomotion of the starfish, and 

 this doubtless forms the foundation on which the erroneous 

 theory mentioned in the preceding paragraphs was based. The 

 common mistake lies in attributing too great importance to the 

 pulling of the tube feet as they shorten up after the suckers have 

 attached themselves. The starfish is supposed to haul itself along, 

 as one might do by attaching ropes in front, then shortening 

 them. This is certainly a mistaken idea, so far as the usual 

 locomotion is concerned. 



My attention was directed to this fact when, holding the usual 

 view, I endeavored to prevent the active locomotion of the star- 

 fish by placing it on loose sand or on greased surfaces. To my 

 surprise the starfish was able to walk approximately as well 

 under these conditions as when placed on stone or clean glass. 

 It is obvious that when the starfish is placed on loose sand it 

 cannot pull itself along by attaching the suckers to a firm object 

 and then pulling upon this as a fixed point. If even the lightest 

 pull is exercised, the sand grain to which the sucker is attached 

 is at once lifted. And on a smooth greased surface the attach- 

 ment of the suckers is prevented. 



Careful study of the locomotion when on loose sand, and com- 

 parison with the usual locomotion on a solid body, showed what 

 the real method of action is. It is the same on the sand as else- 

 where, — save when the starfish is climbing a steep surface or 

 hanging inverted from beneath a surface. 



The typical method of walking is clearly as follows : Each 

 tube foot is extended forward, in the direction in which the 

 animal is moving (thus at difl^erent angles to the ray axis in dif- 

 ferent rays, as shown above). Then the tip of the foot is placed 

 against the surface, the sucker aiding to give it firm attachment, 

 and the whole tube foot is given an impulse to swing back (with 



