604 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



the connection is. One may perhaps hazard the guess that they keep the ciliated 

 grooves from becoming clogged up. 



If the coral is broken away or the rock fragment turned over so that the 

 comatulid is exposed, the animal begins at once to move. In the species with well- 

 developed cirri, notably in Tro-piometra afra, the position is maintained by means 

 of them, but they apparently do not play any important part in locomotion. One 

 might supjjose that they would be of some use either in pulling or joushing, but 

 there never was the least indication of movement on the part of the cirri in any 

 of the species observed. In the species in which cirri are feebly developed or quite 

 wanting the position is maintained by the use of some of the arms. It seemed to 

 be usually the shorter arms that were so used, but under laboratory conditions none 

 of the anns remained still for a long period, and while it would seem quite likely 

 that the shorter arms are constantly used as anchors, evidence on the point is not 

 satisfactory. 



Creeping is a rather complex movement, a combined pulling and pushing. 

 Certain arms, usually three or four, but sometimes as many as seven, are stretched 

 out to their full extent and the pinnules of the terminal portion catch hold of any 

 available projections. This attachment is not only by the use of the minute hooks 

 which are found at the tips of the pinnules, but by the remarkably viscid nature 

 of the secretion from the glands in their epithelium. The degree of viscidity 

 varies in the different species, and more or less in individuals. It is least notice- 

 able in Trofiometra afra, and hardly more so in Comatula purpurea. It is most 

 marked in Comatella stelJigera and in some individuals of Comanthus annulata. 

 Thanks to this viscidity locomotion is possible even on very smooth surfaces, such 

 as that of an enamel-ware basin. It is, however, difficult on loose sand. Rock 

 surfaces and tightly packed sand are the most satisfactory for permitting the 

 pinnules to function successfully. When the pinnules of the extended arms have 

 taken a hold, or perhaps while they are taking hold, three or more arms on the 

 opposite side of the body are drawn up in a strongly arched position and their 

 widespread tips are closely appressed to the bottom very near the disk. Then 

 when the extended arms, having secured their hold, begin to contract and pull the 

 animal toward the place of attachment, the arched arms begin to straighten out 

 and push the animal in the same direction. The result is a movement more or less 

 marked according to the character of the bottom and the size of the comatulid. 

 The total movement may be from one-eighth to one-quarter of the arm length. 

 The process is continually repeated, and, with very slight resting periods, may be 

 kept up for hours. 



Doctor Clark found thatj on the whole, Comatella maculata and the adults of 

 Comatula purpurea were the most active, and were rarely at rest in aquaria. A 

 specimen of ComatelJa maculata with arms about 75 mm. long traveled approxi- 

 mately 85 mm. a minute and continiied to go around and around a large basin for 

 several hours. It probably traveled about 40 meters an hour, but on a natural 

 surface, such as the reef would afford, it would be able to do better than that. 



In comatulids with approximately equal arms there was no evidence of any 

 orientation, one group of arms being quite as likely to be in advance as another. 



