76 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.41. 



short colunmals. The fifth cohimnal is much higher, and is of a rather 

 peculiar shape. It expands from below upward, and contracts 

 slightly in its immediate proximal portion. The shape of this colum- 

 nal is unique among the other elements of the stem, and may be 

 recognized in the succeeding two figures. Such being the case it 

 may well serve as a datum plane for identifying the columnals. 



In the stage represented by figure 9 there has apparently been 

 a fusion of the columnals. The cirri are borne on a much longer 

 columnal, which apparently represents the proximal columnal plus 

 the next succeeding ossicle. Tliis fusion can not be demonstrated, 

 but it seems probable, inasmuch as there are but two columnals shown 

 between the cirriferous segment and the characteristic "fifth" ossi- 

 cle. This does not prove anything, for one may argue that the num- 

 ber of columnals between the proximale and "fifth" ossicle varies, as, 

 for all that we know, it may. However, an extremely close union is 

 indicated between the cirriferous segment and the next lower ossicle. 

 As here given the suture between the two elements it drawn as well 

 marked, whereas in Carpenter it is little more than indicated. The 

 next lower ossicle is somewhat smaller in diameter. 



In the next stage (fig. 10) there has been a complete fusion between, 

 the cirriferous element of the former figure and the next lower 

 columnal. The ''fourth" columnal has enlarged and is of practically 

 the same diameter as the centro-dorsal above it. It would appear 

 that a very close union exists between this ossicle and the centro- 

 dorsal, and that the two elements are weU on the way toward fusion. 

 Upon detachment it appears that the point of disruption would logi- 

 cally be between the "fifth" ossicle and the practically consolidated 

 centro-dorsal. The union between the two seems to be of a compara- 

 tively loose nature. 



From the facts as shown by these specimens we may reasonably 

 assume, I believe, that there has been a fusion of columnals with the 

 proximale to form the centro-dorsal in tliis species. Indeed, I tliink 

 we may be reasonably sure that three columnals in addition to the 

 proximale go to make up the centro-dorsal. By the direct evidence 

 as apparently afforded by a hving species we may conclude then that 

 the centro-dorsal may at times be composed of several fused colum- 

 nals. The same condition seems to obtain in the case of the fossil 

 Comatulse, but in regard to them the evidence is even less conclusive. 



Whether the fusion of several columnals to form the centro-dorsal 

 indicates independent derivation from that had by forms in which the 

 centro-dorsal is composed merely of the proximale such as is doubtless 

 the case in TMolliericrinus, is a question. The columnals of the 

 stalked Antedon, as shown in figures 11 and 12, Plate 7, have a decid- 

 edly Bourgueticrinus type of structure. The evidence seems to point 

 strongly to the conclusion, therefore, that we have to look among 



