p. H. CAEPENTEE ON CEETACEOUS COMATUL^. 555 



I append a key to the various species of Antedon from the English 

 Chalk which have been described in this and in my preceding 

 paper. 



Antedons from the English ChaXk. 



I. Cirrhus- sockets with a keyhole-shaped pit. 



A. Centrodorsal cavity has ribbed walls 1. A. paradoxa. 



B. Walls of centrodorsal cavity plain % A, rugosa. 



II. Oirrhus-sockets have an oval articular surface slightly 



elevated at the ends. 



C. Ten vertical rows of cirrhus-sockets 3. A. Lundgreni. 



D. Sockets numerous in alternating horizontal 



rows 4. A. striata. 



III. Cirrhus-sockets have a transverse articular ridge. 



E. Sockets numerous in irregular rows. Six aper- 



tures at dorsal pole h. A. perforata. 



F. Few sockets in one incomplete row. Dorsal pole 



imperforate 6. A. laticirra. 



In conclusion, I wish to say a few words on a morphological 

 subject which was alluded to in my previous paper (pp. 37, 38), 

 and has recently gained in interest from the writings of Messrs. 

 "Wachsmuth and Springer. 



In recent Comatulce, as in Pentacrinus, the lowest portions of the 

 coslom of the arms are lodged in the median ventral furrows of their 

 skeleton, which terminate in those of the rays. The latter are 

 continued down the inner faces of the first radials, where they are 

 more or less completely converted into canals that end bhndly 

 against the upper surface of the centrodorsal. They are some- 

 times received into five radial pits, which are very deep in 

 certain fossil Comatulce. This is the case in Ant. Betzii., in which 

 they remain outside and separate from the median cavity of the 

 centrodorsal that lodged the chambered organ, being shut off from 

 it by vertical partitions. In A. semiglohosa the central cavity is 

 united (perhaps by accident) with the five others round it and has 

 a stellate appearance, while there is a similar stellate pit on the 

 under surface of the centrodorsal. Stellate dorsal pits slightly 

 smaller in size occur in other fossil Comatulce ; these are far larger 

 than was necessary merely to transmit the vascular axis of the 

 larval stem, continued downwards from the chambered organ. One 

 is led to suspect, therefore, that they gave passage, not only to this 

 central axis, but also to five radial extensions of the body-cavity 

 around it, the radial pits of other Comatulce being here represented 

 by real perforations through the substance of the centrodorsal. 



The condition of Ant. perforata tends to support this view. On 

 its ventral surface (PL XXIII. fig. 2, a) the radial pits are more or less 

 confluent [with the central cavity, but dorsally (fig. 2, h) there are 

 six separate openings, a central one and five others around it cor- 

 responding to the radial extensions of the axial cavity above. 



This question is of interest in reference to the hypothesis recently 

 put forward by Messrs. "Wachsmuth and Springer *, " that the 



* "Eevision of the Palseocrinoidea. — Part i,," Proceed. Acad. Nat. Sci. 

 Philad. 1879, p. 15 (separate copy). 



