380 p. HERBERT CARPENTER. 



Returning now to Marsupites and to the consideration of 

 the nature of the central plate of its apical system (fig. iii, 

 1), we are at once met by the difficulty that we do not know, 

 and are never likely to know, the history of its development. 

 If it passed through a pedunculate stage as Comatula does, 

 the central disc of its calyx cannot be homologous with the 

 subanal plate (fig. ii, 1) of the Echinids as Loven supposes, 

 since the homologue of the latter would, on the view ad- 

 vanced above, have been the terminal plate at the base of 

 the stem. The central disc of Marsupites might then be 

 fairly regarded as homologous with the centrodorsal plate 

 (= uppermost stem segment) of Comatula (fig. i, cd). 

 Except in Glenotremites , this piece is imperforate as in 

 Marsupites, its primitive annular character becoming obli- 

 terated by a secondary extension of its original calcareous 

 reticulation. In most Comatula, however, it is marked by 

 sockets for the attachment of numerous cirrhi which serve 

 to fix the young animal after it is detached from its stem, 

 and there is no trace of these in Marsupites . 



Until quite recently I should have considered this as a 

 serious difficulty in the way of our regarding the central 

 disc of Marsupites as a centrodorsal piece homologous with 

 that of Comatula. But in the examination of the collection 

 of ComatulcR brought home by the *' Challenger," I have 

 come across many specimens in which there are not only no 

 cirrhi, but no sign of their having once existed. The centro- 

 dorsal piece is nearly flat outside, and is not marked with the 

 slightest trace of sockets. This is, however, merely the 

 result of age, the obliteration of the cirrhus socket by a 

 calcareous deposit having been carried to an extreme extent. 

 For in the same as well as in different species, I have met 

 with all kinds of intermediate stages between this and the 

 ordinary condition of two or three rows of cirrhus-sockets 

 which prevails in the younger, and in the commoner 

 Comatul<2, 



It is not altogether clear to me how these older cirrhusless 

 Comatula. can fix themselves in any one spot, as these 

 animals ordinarily do, and the same difficulty of course 

 presents itself in the case of Marsupites, the central disc of 

 which corresponds in many respects with the centrodorsal 

 plate of Comatula. On the whole, however, I am inclined 

 in this case to adopt Loven's views, and believe that Mar- 

 supites was an altogether stemless Crinoid like Holopus, 

 but that the primitive character of its central abactinal plate 

 never became obscured as it does in Holopus by an extraneous 

 calcareous dsposit. The chief reason which has led me to 



