MONOGRAPH OF THE EXISTING CRINOIDS. 203 



They are, especially the right posterior, sometimes separated from the basals 

 by infraradials. 



In the Articiilata as a whole and in certain sections of other groups the size of 

 the radials is correlated with that of the basal portion of the postradial series and 

 bears no relation whatever to that of the basals and inf rabasals ; this is well illus- 

 trated in the comatulids and pentacrinites. 



The arms tend to arise as slender appendages from the edge of the disk, no 

 matter whether the plate situated there is a radial or one of the plates of a subse- 

 quent division series ; unless they happen to lie at the edge of the disk the size of 

 the radials agrees with that of the following postradial ossicles, and not with that 

 of the preceding plates. 



From this it is clear that the only invariable attribute of the radials is their 

 connection with the postradial series of ossicles. 



The origin of the radials as we Imow it in the comatulids indicates no relation- 

 ship whatever with the orals, basals, and infrabasals, but a most intimate connec- 

 tion with the arms. 



The radials do not appear until the orals, basals, and infrabasals, which arise 

 simultaneously at a very early period, have attained their perfected form and 

 proportions. 



The first traces of the radials are found when the swellings indicating the 

 budding arms first appear and tbe ventral ambulacral structures have begun to 

 extend toward them. 



Carpenter noticed that the radials, like the basals and orals, commence as 

 expanded cribriform films, while the endogenous additions by which they are sub- 

 sequently thickened are cribriform like those of the basals, and not fasciculated 

 like those of the primibrachs, which originate as imperfect rings, which soon become 

 filled up with lengthening fasciculated tissue, just as is the case with the columnals 

 and later brachials. 



This is probably correlated with the fact that the radials are broader than the 

 succeeding plates and without further significance. 



Structurally the radials are always related to the following ossicle in exactly 

 the same way that all the axillaries are related to the ossicles immediately following, 

 and in certain types, as, for example, Arachnocj-inus hnihosus (part 1, pi. 16, fig. 

 595), there is an unmisiakable agreement in form between the radials and the 

 axillaries. 



The orientation of the five primary nerve trunks, which are interradial, the 

 position of the water pores, which are interradial. and the location of the anus, 

 which always lies in the posterior interradius, show that in the larva the five seg- 

 ments of the body are the regions covered by the orals and the basals, and that the 

 radials are intersegmental plates. 



Radianal. — In the pentacrinoid young of the comatulids the radianal plate 

 is always found, is invariably well developed, and is always closely associated 

 with the right posterior radial, usually at first occupying a position beneath its 

 left lower angle from which it moves upward and outward, passing around the 

 entire left side of the radial and out onto the disk. 



