MONOGRAPH OF THE EXISTING CEINOIDS. 605 



Such individuals altered the direction in which they traveled simply by changing 

 the groups of arms which were to be extended. 



In adult Comatula purpurea, in which the anterior arms are very much longer 

 than the dwarfed posterior, observations on the reef, at the live car, and at the 

 laboratory showed that there is more or less definite orientation. Under ordinary 

 conditions the long arms are extended in locomotion, and hence are anterior, while 

 the short arms do the pushing, and hence are posterior. The mouth is at the base 

 of the longer arms, and hence is at the anterior side of the disk. 



If locomotion in a given direction were prevented the comatulid never re- 

 versed its movement, using the short arms to pull and the long ones to push, but 

 gradually swung itself around until ultimately the long arms were in advance. A 

 considerable number of experiments and observations showed that the mechanism 

 for orientation was not by any means perfectly adjusted, for an individual would 

 often go for some distance with the longest arms at one side, or only partially in 

 front. Nevertheless it was clear that long arms in front, short arms behind, gave 

 the normal method of progression. 



Doctor Clark remarks it may be a fair inference that the asymmetrical condi- 

 tion of the arms is the result of selection, since it is possible that long arms would 

 be better for pulling and short stout arms for pushing ; but it is no doubt true that 

 there are very diverse interpretations of the matter possible. 



Locomotion in all the comatulids observed at Mae'r, whether swimming or 

 creeping, invariably takes place with the oral surface uppermost. If dropped in 

 deep water with the mouth down they commonly righted themselves before reach- 

 ing the bottom. If not, or if placed on the bottom with the mouth down, they 

 slowly but surely turned over before attempting to either swim or creep. The 

 process of turning over is essentially the same as in a starfish ; that is, a group of 

 arms begins the task and continues gradually lifting up one side until the arms 

 of that side can be bent back far enough over and beyond the aboral surface to 

 secure a hold. Then by their pulling, and the pushing of the arms opposite to 

 them, the disk is righted. The process seems a tedious one and is probably not 

 often necessary in the normal life of comatulids on the reef. 



Particularly interesting is the question of how much these comatulids move 

 about under the natural conditions of their life, but no satisfactory evidence 

 on the point was secured. 



On one occasion a dozen healthy comatulids of four species were " planted " 

 on a part of the reef where no crinoids were living. Twenty-four hours later 

 they had all disappeared, and no trace of them could be found in the neighbor- 

 hood. But as there had been a strong wind and much surf during the night 

 they may have been simply swept away. 



On another occasion 10 large comatulids were carefully located around the 

 margin of a rock fragment about a meter in diameter. Circumstances, including 

 unfavorable tides, made it impossible to visit the spot again for a week. Then 

 it was found that more or less sand had washed in around the rock, and the 

 comatulids were apparently gone, but again there was doubt what part the surf 

 and strong tidal currents may have played. 



