CLARK: PLATES IN CRINOIDS 311 



interradius and its migration upward and toward the left leave 

 no room for doubt that the so-called anal of the pentracrinoid 

 larvae is nothing more nor less than the radianal of the fossil forms. 



Since the radianal is represented in the larvse of the comatulids 

 we should expect also to find a representative of the plate known as 

 anal x. Now in forms in which the radianal is present anal x 

 lies directly over the posterior basal, always to the left of the radi- 

 anal and always maintaining a closer relation with the radial to 

 the left of the posterior interradial area than with that to the 

 right, with which the radianal is associated. Whereas the radianal 

 is always a single plate, anal x commonly forms the base of a short 

 series of more or less similar plates. 



In the so-called Thaumatocrinus renovatus the posterior inter- 

 radial area is occupied by a large interradial plate bearing upon 

 its distal edge a conical process composed of a series of calcareous 

 rings; this process lies to the left of the base of the anal tube, and 

 therefore presumably to the left of the recently vanished radianal. 



The free arms of the crinoids are composed of an extension of 

 the boundary between the primarily skeleton forming dorsal 

 surface and the perisomic ventral surface. Although phyloge- 

 netically very complicated, ontogenetically they arise as a linear 

 (or double) series of ossicles, each new ossicle being added at the 

 extremity of the series; none of the phylogenetic processes by 

 which they originated are recapitulated. In short a crinoid arm 

 in all the forms we know is nothing but a double or single series of 

 ossicles supporting an extension of the ventral perisome — a series 

 of simple braces of long forgotten origin. 



The fixity of the crinoid arm as an individual structure entirely 

 distinct from the crinoid arm as a phylogenetic complex — the 

 conception of the crinoid arm as a structure with an identity of 

 its own and with an ontogenetically completely obliterated phy- 

 logenetic origin — must constantly be kept in mind; for when we 

 are able to grasp this idea we see at once that any series of ossicles 

 arising on the border between the dorsal skeleton forming and 

 ventral perisomic surfaces, and being composed in equal parts of 

 each, will assume the structure common to all the processes arising 

 in the same region, and will take on from the beginning the struc- 



