CORBITELLA AND HETEROTELLA. 15 



of the floricomes exists between this and the foregoing species 

 in the nature of the discohexaster. Whereas in C. sjieciosa this 

 kind of rosette occurs predominantly in the ordinary hexaster 

 form and only exceptionally in the modified hexactinose form, 

 the relative proportion of these two forms as they occur in 61 

 elegans is just the reverse. A further point of important difference 

 lies in the fact that the hexactinose discohexaster of the latter 

 species is usually nearly twice or more than twice as large as any 

 discohexaster in the former (PL, fig. 13). 



The hexactinose discohexaster of G. elegans was figured and 

 remarked upon by Marshall. In all my preparations it occurs 

 in very great abundance and in most places in crowded disposition. 

 Axial length 220-264 1*. The six rays arranged as in a regular 

 hexactin are usually straight or nearly so. Near the central node 

 about 7 h <>■ thick, they narrow somewhat outwardly but to thicken 

 again near their junction with the convex terminal disc. Breadth 

 at the middle about 6 /a The terminal disc measures up to 19 l*- 

 in diameter. Its recurved marginal prongs are strong ; in number 

 5-8, usually G, to each disc. Altogether the disc is shaped exactly 

 like that of the discohexaster of C. speeiosa. 



In the figure of this characteristic spicule given by Mar- 

 shall ('75, fig. 66 b) one important point requires a correction; 

 viz., the axial filament is represented by him as reaching right 

 up to the terminal disc. This is by no means the fact. Examina- 

 tion of the spicule in glycerine reveals that the axial cross is 

 confined to the central node, in a manner known to me to be 

 the case in all hexactinose rosettes derived from hexasters (Ij., 

 '97? PP- 44-45). I do not know how to account for another 

 figure of Marshall's (7. c, fig. 66 d), in which a spicular ray 

 with axial filament is shown as being forked into two disc-bearing 



