40 P. H. CARPENTER ON COMATULiE FROM 



Goldfuss (Taf. li. fig. 1) it is decagonal, almpst what Schliiter would 

 call lobate ; in the other (Taf. xlix. fig. 9c) it is only pentagonal ; 

 so that in fossil as well as in recent Comatulce the shape of the axial 

 opening was not constant. I do not think, therefore, that any cor- 

 relation can he established for A. lenticularis, any more than for 

 A, italica, between the shapes of the two openings, dorsal and 

 ventral, in the axis of the centrodorsal piece. 



What I take to be the second part of Schliiter's argument seems 

 even more unsatisfactory than the first; for although in recent 

 Comatulce a lobate axial opening generally does have radial pits 

 around its border, yet this is not always the case, and the pits may 

 occur round a pentagonal opening, as in the fossil A. paradoxa^ while 

 Schliiter himself records that they may be sometimes present and 

 sometimes absent in Solanocrinus scrohiculatus, just as I have 

 found to be the case in A. rosacea and A. celtica. The occurrence 

 of such variations as these tells strongly against Schliiter's proposed 

 classification of the fossil Comatulw. He divides them into two 

 groups according as they have (A) radial pits and a 5-lobed axial 

 opening, or (B) no radial pits and an undivided opening. Were 

 recent Comatulce classified thus, A. rosacea and A. celtica would 

 appear in both groups, while some individuals with lobate or decago- 

 nal openings but no pits, would find a place in neither ! 



Although not constant in their occurrence, the radial pits of these 

 two species are peculiar to them among all the recent Comatvlce I 

 have examined. In no other species 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 

 ofi' from these surfaces. They have a precisely similar appearance 

 in Ant. rotunda (PI. Y. fig. 5 a), except that they are a trifle deeper 

 and more distinct thau in A. rosacea ; but they have nothing hke 

 the relative size of the radial pits in most of Schliiter's species. 

 Besides A. rosacea and A. celtica, I only know of one other recent 

 Comatula in which radial pits occur ; but their shape is very peculiar. 

 The axial opening is bluntly lobate, and the pits are situate at five 

 points on its projecting lip, or, rather, their central ends are ; for 

 they bifurcate and extend outwards so as to occupy the best part of 

 each radial area, ending a little inside its margin. This feature is 

 of some interest ; for in one of Goldfuss's two figures of A. ^aradoxa 

 (Taf. xlix. fig. 9 c) all the radial pits show a tendency to bifurca- 

 tion at their outer ends, two of them especially so. The same is 

 the case in our English example of this species (PI. Y. fig. 1) and 

 in Act. ahiormis (PL Y. fig. 8a). 



I have shown elsewhere f that, omitting the little-known Comaster 

 of Goldfuss J, the recent Comatulce may be referred to three well- 

 marked generic types, viz. Antedon, Actinometra, and FromacJio- 

 crinus. Loven's Plianogenia is, I believe, only an aberrant Actino- 



* '^Actinometra" pi. 4. fig. 15, q. 



t " Preliminary Eeport upon the CoTuatulcB of the ' Challenger ' Expedition," 

 Proc. Eoy. Soc. No. 194, 1879, p. 385. 



I Journ. Linn. Soc, Zoology, vol. xiii. pp. 454-456. 



