AND IT.. .TO EECENT COMATTTL^l. 193 



of the basals, these four all differ in the height of the outer 

 surfaces of the radials. In none of them does it reach the same 

 relative proportion that it does in Goldfuss's figure (PI. IX. 

 fig. la). There is also a considerable amount of variation in the 

 proportions between the height and width of the articular faces, in 

 the shapes of their muscle-plates, and in the relations between 

 the diameter of the central funnel and that of the entire calyx 

 Iu Groldfuss's figure the upper ends of the muscle-plates are 

 bluntly pointed, their superior margins sloping sharply downwards 

 towards the intermuscular notch ; but in three of the specimens 

 before me they are more squared and nearly horizontal. Further, 

 while the total diameter of my figure of the top of the calyx (PL 

 IX. fig. 2 b) and the corresponding one of Goldfuss (PI. IX. fig. 1 b) 

 is the same, the diameters of the central funnel are very different 

 in the two cases, being 16 millims. in fig. 2 b, but only 12 millims. 

 in Groldfuss's figure (PI. IX. fig. 1 b). This appears to be due 

 to the distal faces of the radials of the Cambridge specimen 

 having a rather less inward slope than those of Goldfuss's 

 specimen. 



There is yet another difference between the type and most of 

 the specimens of Antedon costata which I have examined. In 

 the former the cirrhus-sockets are regularly arranged in ten 

 vertical rows. In the latter they are larger and much less regular, 

 very much as in the specimen represented in Quenstedt's ' En- 

 criniden ' (pi. 96. fig. 32), which has squarish muscle-plates. On 

 the other hand, his fig. 33 represents a specimen with a more 

 regular centrodorsal and pointed muscle-plates ; but it differs from 

 the type in having no interradial notches between the muscle- 

 plates of contiguous radials. The specimen figured by Quenstedt 

 in ' Der Jura,' tab. 88. fig. 10 (reproduced here on PI. IX. fig. 4) 

 is also different from the type, as the outer dorsal surface of the 

 radials is greatly reduced. In this respect it is just at the oppo- 

 site extremity of the series to Goldfuss's specimen, in which the 

 exterior of the radials is unusually large (PI. IX. fig. 1 a). 



These differences are slight exaggerations of the kind of varia- 

 tion that one finds in recent Comatulce ; but they are insufficienc 

 to form the basis of specific distinctions. Curiously enough, one 

 of the Cambridge specimens shows how variation may occur even 

 in individual cases. It is slightly smaller than the others, with a 

 more regularly ribbed centrodorsal (as in Goldfuss's specimen), 

 and a relative width of the central funnel which is intermediate 



