ME. P. H. CAEPENTER ON THE GENUS ACTINOIVIETEA. 75 



areas on the ventral surface of tlie centrodorsal piece of Ant. rosacea (PI. IV. fig. 15, r.ar), 

 in each of which hcs the convex dorsal surface of a single first radial. 



What, theu, are the five radiating punctated grooves which Agassiz aud Roemer 

 regarded as the articular surfaces for the attachment of the five arms of Glenotremltes ? 

 I believe them to be the representatives of the basal grooves on the ventral surface of 

 the centrodorsal piece of Actinometra (PI. V. figs. 2, 7, li, PI. VI. figs. 3, 8, 10, 15, 17, 

 21, h.g). They are sometimes slightly developed in Ant. rosacea, one lying between every 

 two of the de2:)ressions (</) mentioned above (PI. IV. fig. 15), in precisely the same manner 

 as the grooves and " genital openings" alternate on the ventral surface of Glenotremltes. 



We thus see that the peculiarities of the ventral surface of Glenotremltes may be 

 readily explained by what we know of the corresponding parts in Autedon and Actino- 

 metra. The apertures in the centre of the dorsal surface admit of an equally simple 

 explanation. 



The quinquelocular organ forming the dorsal termination of the axial prolonga- 

 tion of the adult Comatula consists of five chambers, arrann;ed around a central axis 

 which contains numerous vessels. In Pentacriniis there is no centrodorsal piece, but the 

 quinquelocular organ is contained in a cavity, the sides of which are formed by the first 

 radials above and by the basals below. Its five chambers are not closed below, but 

 narrow considerably, and are continued down the stem as five long vessels arranged 

 symmetrically around a central axis. The same appears to be the case in the stem of 

 the Pentacrinoid larva of Comatula; for, as already mentioned. Dr. Carpenter has 

 described a minute five-rayed perforation occupying the central portion of the dorsal 

 surface of the recently detached Autedon. I regard this perforation as homologous with 

 the five small apertures arranged around a single central one on the dorsal su.rface of 

 the centrodorsal piece of Glenotremites, and with the similar openings on the underside 

 of the calyx of the other stalked Crinoids — for example, of Cupressocrlnus. The fact 

 that the young Antedon rosacea has only three rows of dorsal cirrhi when liberated 

 from its stem, while there are four or six rows on the dorsal surface of Glenotremites, 

 does not at all tell against this view. Indeed Sars^ has shown that the pentacrinoid stage 

 persists in Antedon Sarsii very much longer than in Ant. rosacea, and he has found 

 specimens with nearly thirty cirrhi still in a pedunculate condition, the cirrhi being 

 placed in such close proximity to one another that any trace of a distinct order in their 

 arrangement was entirely obliterated. The exterior of the centrodorsal piece of Ant. 

 Sarsii, therefore, immediately after its liberation from the stem, would present (its size, 

 of course, excepted) a very similar appearance to the convex dorsal surface of Gleno- 

 tremites, viz. a central five-rayed opening, or possibly even a single opening with five 

 others round it, the rest of the surface being covered with sockets for the i3rticulation of 

 the dorsal cirrhi. 



(iv.) The Pentagonal Base of the Calyx. 

 (§ 51) In aU the Actinometra; with which I am acquainted the Pentagonal Base of the 

 calyx formed by tlie close mutual adhesion of the five first radials, together with the 



' ' CrinoVdcs vivants,' p. 57. 

 SECOND SERIES. — ZOOLOGY, VOL. II. 11 



