ME. P. H. CARPENTER OX THE GENUS ACTINOMETRA. 85 



cular opening. So far as I know, these openings are tolerably constant on the dorsal 

 surface of the radial pentagon of Ant. rosacea ; but the five depressions corresponding to 

 them on the ventral surface of the centrodorsal piece are very variable in the distinct- 

 ness of their development ; and Dr. Carpenter has found that in some cases they may be 

 absent altogether \ 



This last condition, in which there are no radial depressions (q) on the ventral surface 

 of the centrodorsal piece, appears to be the normal one in Ant. celtica, in which I have 

 rarely found any traces of such depressions (PI. IV. fig. 2). The margin of the central 

 opening is usually almost circular, though sometimes bluntly stellate as in Ant. rosa 

 (fig. 15) ; at the same time, the five openings ( Q) upon the dorsal surface of the radial 

 pentagon are but little developed or even entirely absent. In PI. IV. figs. 3, 7, they are 

 represented as present in the small variety and absent in the large one ; but I have 

 sometimes found exactly the reverse to be the case. 



"We shall find the same variability in the presence or absence of these openings in 

 Actinometra, not only in different individuals of the same species, but in the same indi- 

 vidual. This fact shows that too much reliance must not be placed on the presence, 

 absence, or difference in size of similar openings in the calyx of the fossil Crinoids 

 (the interradial " Li'icken " in Cupressocrinus, for example) as characters of any 

 systematic value. The absence or slight development of these openings in Ant. celtica is 

 principally due to the fact that the inner margin of the dorsal surface of the first radials 

 is not notched but straight, the radial axial furrow not being continued so far towards 

 the dorsal surface as in Ant. rosacea ; and also that processes grow inwards from the two 

 sides of the dorsal end of each of the five spout-like rays of the rosette, so that the 

 lumen of the canal it encloses becomes much diminished ; while in some cases similar 

 processes are put forward from the margin of the first radial, which unite with the 

 others so completely as eutirely to obliterate the lumen of the radial axial canal, and 

 thus form its dorsal boundary. 



(§ 58) The dorsal aspect of the pentagonal base of the calyx of Actinometra is by no 



1 Schliitcr (op. cit. p. 37) has proposed a division of the (fossil) Comatulce into two groups, characterized as fol- 

 lows : — (1) Centrodorsal with no radial pits and a round " Nahrungscanal " (central opening) ; and (2) Centrodorsal 

 with radial pits and a quinquepartite opening. 



These characters, however, are far too uncertain to be of any systematic value. For example, Schliiter himself 

 notes the absence of the " Eadialgruben " in his own specimen of Solanocrinus scrobiculatus, while they were present 

 in one examined by Quenstedt (Echinodermen, p. 179). I have some specimens of Ant. celtica answering to the 

 first, and others to the second of the above definitions: and although most specimens at Ant. rosacea would be classed 

 in group (2), yet individuals with a pentagonal or even quinquepartite opening, ?»tZ no radial jiits, are not uncommon. 

 This last condition is very common among the ' Challenger ' Antedons. In fact, the radial pits of Ant. rosacea and 

 Ant. celtica arc peculiar to these species, and not always present even in them. In no other recent Comatula have 

 I found any thing exactly like them. They are not parts of the generally concave surface of each radial area, but 

 have distinct peripheral borders marking them off from these surfaces, and corresponding to the openings of the radial 

 axial canals enclosed within the spout-like processes of the rosette. In Act, pectinata, Act. polymorpTia, var. 2, and 

 in a new ' Ch.illeii' ' \edon there is a distinct pit at the central end of each radial area, which is merely the 



deepened termination of a depression occupying its median line. Its nature (sects. 45 & (!1 ) is essentially the same as 

 that of the radial pits of Ant. rosacea, but its appearance is very different. Hence I cannot corroborate Sehluter's 

 statement that " Slanche Arten" possess '• Eadialgruben." He ouly describes them in ~> out of his 11 fossil species, 

 i besides the Solaru crinns scrooieidatus examined by Quenstedt. 



12* 



