378 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



as in the species of CoinasteridiE in which cirri are absent (figs. 151-159, p. 221; 

 162, p. 223, 164, p. 227, and 165-170, p. 229); but when thecentrodorsal is large, as 

 in the species of Grinometra, and in certain species of Catoptometra, Pachylometra, 

 Eeliomftra,Solanometra, Comatula, ComatuMla, ConioifZto, and a fewother genera (figs. 

 77, p. 130, 80, p. 133, 81 , p. 134, 99, p. 160, 100, p. 162, and 101 , 102, p. 163), the radials 

 may be so far withdrawn that part or even all of the first priinibrachs are concealed. 

 Quite independently of the increase in size of the centrodorsal, the radials may be 

 more or less reduced, as is seen in various comasterids; this of course assists con- 

 siderably in their concealment. There appears to be not the shghtest correlation 

 between these two processes. The outline of the centrodorsal, when large and show- 

 ing no trace of radial resorption is approximately cii'cular, whereas the periphery 

 of the radial circlet is pentagonal; moreover the outer surface of the individual 

 radials where not in mutual apposition or attached to the centrodorsal is convex; 

 hence, though the radials may be entirely concealed m the median line, thej' are 

 almost invariably to be seen in the inteiradial angles as a more or less prominent 

 triangle gabeling over the ends of the basal rays (fig. 95, p. 157). These interradial 

 triangles commonly appear as structures having an entity of their own, and have 

 frequently been mistaken for basals, but a close examination will reveal a very 

 close sutural line dropping perpendicularly from the apex toward the base, and in 

 the middle of the base the tubercular elevations marking the external ends of the 

 basal rays. 



P. H. Carpenter considered the radials of the pentacrinites comparable to 

 those of the stalked larva of Antedon, because they appear above the basals on the 

 exterior of the calyx as relatively large convex plates. This similarity is, however, 

 purely superficial; it is true that the external appearance of the radials in the two 

 bears a close resemblance but, while those of the pentacruiites are nearly horizontal, 

 the greater part of their external thickness extendmg horizontally inward over the 

 ventral surface of the basals to the center of the caljrx, those of the stalked young 

 of Antedon are more slanting, not yet having begun to undergo the change to the 

 nearly horizontal attitude of those of the adult. The radials of the pentacrinites 

 can only be compared with the radials of the adults of such macrophreate species 

 as show a comparatively large portion of their radials on the exterior of the calyx, 

 as do the species of Atelecrinus (figs. 123, p. 192, 124, 125, p. 193, 414, p. 319, and 

 573, 1)1. S) or Bathymetra; the radials of the verj' young comatulids are comparable 

 to the radials of such genera as Proisocrinus (fig. 128, p. 199), but by no means 

 comparable to the radials of the true pentacrinites. 



The radials of the comatulids are in a phylogenetically more advanced conchtion 

 than those of the pentacrinites; that is, they have become more recumbent and the 

 outer (now dorsal) side has become shorter so that they have withdrawn more or 

 less (often entirely) within the area covered by the centrodorsal. The radials of 

 Atelecrinus, like the basals of the same genus, have imdergone the least change, 

 and are essentially like the corresponding structures m the pentacrinites, in par- 

 ticular in the genus Enxioxocrinus. In the genus Bathymetra of the Antcdonidie 

 also the radials are essentially as in the pentacrinites, though here the basals have 

 disappeared entirely from external view. 



