222 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



convex on the outer surface, just as do the plates often seen in a similar position 

 in many of the fossil species. 



This plating is always, so far as known, found in Comaster belli, almost 

 always in Comanthina scMegelii, but only occasionally in the two species of 

 Comaster, where, however, it reaches its highest development. 



I previously regarded these plates as the sum of all the calcareous particles 

 within the portion of the perisome affected, which have become cemented together 

 into a uniform granular calcareous sheet, which, with the consequent increase in 

 rigidity, always maintained sutures along definite lines determined by mechanical 

 considerations dependent upon the surface movements of the interradial and the 

 interbrachial integument. There is no doubt that these plates are very frequently, 

 if not usually, formed in this way, for all stages can be traced in such species 

 as Comanthina schlegelii; but since the discovery of the large perisomic plates 

 in the very young of Comacthiia meridionalis (part 1, fig. 412, p. 317) and their 

 persistance until comparatively late in life in the lateral interradial perisome it 

 has seemed possible to consider these plates in Comanthina and Comaster as the 

 interradial plates of the very young, which, instead of becoming resorbed, have 

 been retained and have increased in size correlatively with the general increase 

 in size of the animal, although the plates on the ventral surface of the disk have 

 long since disappeared. 



These plates usually have a very definite arrangement. At the proximal end 

 of the series, between the distal ends of, or just above, the radials, there is a single 

 more or less rhombic plate which is followed by a double column of alternating 

 subhexagonal plates; rarely there are some additional, usually small, plates 

 inserted between these two columns. 



In the Atlantic species of the genus Antedon (figs. 1018, pi. 7, and 1145, 1146, 

 pi. 23), as well as in ComatUia iridometrlformis and a few other forms (fig. 

 1143, pi. 23), there are found close down upon the radials in the interradial 

 portion of the perisome several small plates which may be so closely united 

 as to appear as a single interradial plate. Thomson supposed that in Antedon 

 hifida these are derived from the five minute interradial plates sometimes observed 

 in the very young, representing the last vestiges of the complete juvenile peri- 

 somic plating just described in Comactinia meridionalis. He did not trace out the 

 process of derivation, however, and P. H. Carpenter has suggested that they are 

 in reality secondary perisomic plates originating in later life. 



The perisomic plates on the ventral surface of the disk are ordinarily dis- 

 tinguishable from those in the lateral interradial areas by the common presence 

 of perforating water pores, but Carpenter found that these pores are occasionally 

 present in the plates occupying the interradial areas. 



P. H. Carpenter first noticed that in those comatulids which have short and 

 rounded gonads and plated ambulacra the enlarged part of the pinnule is pro- 

 tected by a very strongly developed anambulacral plating (figs. 1055, 1057, 1058, 

 1062, pi. 14), which is much more regular and closely set than that of the disk 

 and arm bases. Resting upon the four or five middle segments of the short 



