REPORT ON THE CRINOIDEA. 9 



radial pentagon, one at the inner end of each radial (PI. III. fig. 6?> ; PI. IV. fig. 2c). 

 These are the dorsal ends of the radial axial canals, which do not become obliterated as is 

 usually the case ; and in Antedon disciformis there is a small pit on the upper surface of 

 the centro-dorsal corresponding to each of these canals which terminate blindly in this 

 position (PL IV. figs. 2c, 2c/). Among recent Comatulse, however, the most striking 

 development in this respect is presented by Antedon quinduplicava ; for the radial axial 

 canals which pass over from the ventral to the inner faces of the radials turn outwards 

 ao-ain at the bottom of the calyx, and expand into relatively large bilobate cavities which 

 are formed by excavation in the apposed surfaces of the radials and the centro-dorsal 

 respectively, as is well seen in PI. IV. figs, lc, id. 



Among the fossil Cornatulte there are several species in which the ventral surface of 

 the centro-dorsal is marked by five small radial pits of this kind, that receive the 

 ends of the radial axial canals. But in Antedon retzii they appear as actual perforations 

 in the ventral surface of the centro-dorsal which reach downward to the bottom of its 

 internal cavity, being in fact only separated from it by a narrow septum, and this is 

 occasionally absent, so that the centro-dorsal cavity which is naturally decagonal or 

 pentagonal in outline becomes stellate. This condition is very common in the stem-joints 

 of some Palseocrinoidea, such for example as Cupressocrinus, and I think there can be 

 no doubt that the radial openings or the extensions of the central canal in all such cases 

 served for the passage of canals containing water in communication with that in the 

 coelom above. 



Messrs. Wachsmuth and Springer ' suggested long since that the complex stem of 

 many Palseocrinoids might have been " subservient to respiration " ; and the facts 

 mentioned above respecting the Bourgueticrinidse and the Comatulae certainly go far 

 towards supporting this view. 



The ventral surface of the centro-dorsal is usually flat or slightly hollowed, rarely 

 very convex, except in species like Actinometra paucicirra, Actinometra typica, &c, 

 in which the greater part of the centro-dorsal is enclosed within the radial pentagon, as 

 will be explained shortly. The internal openings of the canals leading to the cirrus- 

 sockets are frequently visible on the floor of its cavity, as is well shown in Promachocrinus 

 kerguelensis and in Antedon antarctica (PI. I. figs, id, 6d). In both these species and 

 also in others the walls of the centro-dorsal cavity are marked by strong ribs, the lower 

 ends of which are more or less distinctly visible through the axial opening, projecting 

 beneath its bp, which their upper ends help to support. Five of them, those at the 

 interradial angles, are often considerably larger than the rest, and may be the only ones 

 visible. In other cases, however, both these and numerous smaller intermediate ribs are 

 visible through the axial opening, as is seen in PI. I. figs. Id, 6d. These ribs are much 

 more distinct in some individuals than in others of the same species. Thus, for example, 



1 Revision of the Palseocrinoidea, part i. p 15, Proc. Acad. Nat. Sci. Philad., 1879. 



(ZOOL. CHALL. EXP. — PART LX. — 1888.) *-* 00 2 



