312 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



crim/s are easily identified because of the 10 rays; those of the species of Thnu- 

 matocrinus because of the deeply stellate shape and 10 undivided ambulacral 

 grooves; those of the species of Penfametrocrinus because of the deeply stellate 

 shape and the five undivided ambulacral grooves; those of the species of Com- 

 asterida? because of the absence of sacculi; those of the species of Himerometridae 

 because of the incised borders and numerous irregularly-branched ambulacral 

 grooves; those of the Zygometridse (other than the species of Eudiocrinus) because 

 of the character of the plating, combined with the many -branched ambulacral 

 grooves; and those of the species of Tropiometridre because of the irregular division 

 of the five primary amliulacral grooves. 



There is a limited amount of correlation between the extent of plating on the 

 disk and the development of side and covering plates along the brachial and pinnide 

 ambulacra. Thus in the Calometridie both the plating of the disk and of the 

 ambulacra reach the highest degree of perfection, and the next most perfect stage 

 of each is found in the Thlassometridte and Charitometridse. On the other hand, 

 covering plates are highly developed in the genera Nemaster and CoviatUia in the 

 CapillasterinfB ( Com aster idse), and both side and covering plates in certain species 

 in the Heliometrina? (Antedonidse), in all of whicli the disks are naked; and the 

 disk is more or less hcavilj^ plated in the Zygometridfe, V oUjbometra diadcma^ and 

 Pentametrocrinus sem-peri and P. atlanticus, in which no side or covering plates can 

 be detected without minute examination. 



In the very young postpentacrinoid stages of Thaumafocrinus renovaius (part 

 1, fig. 117, p. 183) and of Comactinia vieridionalis the disk between the margin and 

 the outer borders of the orals is completeh' covered hj a pavement of small plates, 

 which later disappear. These do not appear to have any connection with the plates 

 just described, but seem to be the last vestige of a complete plating of the disk 

 which occurred in the ancestral types. It is curious that such a plating has not been 

 observed in the developing Antedon, but the stage at wliich it occurs has not as yet 

 received much attention. 



In certain of the larger species in the Comasteridae, most conspicuously de- 

 veloped in Comaster multifida, C. helli, C. novceguinem, and Comanthina schlegeJii, 

 in Erythrometra^ in Masfigoinetra., invariably in Antedon diibenii and -1. moroc- 

 cana, and frequently in A. bifida, interradials occur. In the comasterids these are 

 usually in the form of two, sometimes with a more or less developed third, colimms 

 of large polygonal plates resting upon a single plate in the interradial angle. 

 They may be supplemented by similar plates between the branches of the division 

 series. In Comanthina schlegelii they are almost invariably present, and large and 

 solid; in the other species they are sometimes extraordinarily developed, eacli 

 individual plate having a convex outer surface, and very frequently quite absent. 

 In the developing Comactini.a it has been noticed that the early complete plating 

 of the integument of the disk persists longest in the lateral interradial areas, and 

 it is quite possible that these interradials as seen in the large comasterids arise from 

 the retention and enlargement of these larval structures. In Erythrometra (figs. 

 666, 667, p. 329) the interradials take the form of several columns of small isolated 



