320 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



of the interradial ridges at the upper end of the centrodorsal. * * The pen- 

 tagonal shape of the basals is still traceable in slightly older specimens * * * 

 but in still older ones * * * the amount of the radials which is visible on the 

 exterior of the calyx becomes relatively less and less, and the same is the case 

 with the basals. These are best described as triangular, with their lower angles 

 extended so as just to meet those of their fellows and separate the radials from the 

 centrodorsal by what is practically little more than a line, only visible at all under 

 specially favorable conditions of light." Carpenter believed that even this is 

 absent in part of some of the specimens, so that the radials actually come into 

 partial contact with the centrodorsal. This has been found to be the case in speci- 

 mens recently collected, in some of which the basals are only to be seen in the 

 angles of the calyx, where they are scarcely so prominent as are the basal rays in 

 many forms. 



The basals of Atelecrinus were said to be comparable to those of the penta- 

 crinites ; the comparison may be made still closer if the pentacrinite genus Endoxo- 

 crinus is suggested, for in Endoxocrinus there are no infrabasals, and the basal ring, 

 therefore, is quite free interiorly. 



In all the recent comatulids excepting Atelecrinus the basals in the adult 

 become metamorphosed into a peculiar plate, aptly termed by W. B. Carpenter the 

 rosette. In the words of Carpenter, the rosette of Antedon lifida "may be described 

 as consisting of a disk perforated in the center, with ten rays proceeding from it, 

 five of these rays being triangular in form and nearly flat whilst each of the other 

 five that alternate with these has parallel margins inflected on its ventral aspect in 

 such a manner as to form a groove, whilst the ray curves to its dorsal aspect in such 

 a manner as to bring this groove to the periphery of the rosette, and then terminates 

 abruptly as if truncated. Around the central perforation we sometimes find on the 

 ventral surface an irregular raised collar, obviously corresponding to the central 

 passage of the annulus of the pentagonal base, but more commonly this is replaced 

 by a number of vertical processes irregularly disposed. Its diameter in a full-grown 

 specimen is about 0.045 inch. When we look at this rosette in position we find that 

 the five triangular rays are directed to the sutures between the five radials, their 

 apices joining the contiguous pairs of these just between their two adjacent aper- 

 tures leading to the radial canals, whilst each of the five spoutlike rays join the 

 inflected margins of the former, being applied to the borders of the vertical furrow 

 of the latter in such a manner that the two grooves are united into a complete canal." 

 Notwithstanding the apparent continuity between the calcareous reticulation of 

 the rosette and that of the pentagonal base at the extremity of each ray of the 

 former, Carpenter was "disposed to think the continuity not real, since, after 

 boiling in a solution of potash, the rosette separates itself from the radials without 

 any positive fracture at these points. A real continuity, however, would seem to 

 exist between the central prolongations of the radials and the discoidal portion of 

 the rosette, these prolongations attaching themselves to it either separately or 

 after coalescing with each other either to a slight extent or so completely as to 

 form the collar just described, and this junction being so complete that its sepa- 

 ration can only be effected by fracture." 



