ME. P. H. CAEPENTEE ON THE GENTJS ACTINOMETEA. 67 



tlie large Act. robusta (PI. V. fig. 14), in which three of the basal furrows are widest at 

 their central ends, and consequently triangular, while the others are somewhat irregular 

 in shape. 



In both these specimens numerous small openings are visible on the floor of the 

 central cavity of the centrodorsal piece, but the principal ones leading to the marginal 

 cirrhus-sockets are concealed under its projecting lip. In the small centrodorsal piece of 

 Act. pectinata, however, these last are very distinct (PI. V. fig. 7, u) and correspond 

 in number to the eleven external cirrhus-sockets (PI. V. fig. 6, TJ), so that the five 

 principal cirrhus-vessels leaving the quinquelocular organ Avould seem to divide at once 

 within the cavity of the centrodorsal piece, and not within the substance of its walls, as 

 is the case in Act. polijmorpha. The ventral surface of the centrodorsal plate of Act. 

 pectinata (PI. V. fig. 7) is by no means so flattened as in the closely allied Act. Solaris 

 (fig. 2), but rises considerably from the circumference towards the centre, and the inter- 

 radial ridges are well marked. The basal grooves {b.g) are narrow and parallel-sided, and 

 terminate within the margins of the plate, beyond which the interradial ridges are not 

 continu.ed, so that there are none of the small processes extending outwards from the 

 angles as in some forms of Act. polymorpha. The median line of each of the radial areas 

 is occupied by a deep depression, which is particularly distinct at its central end. A 

 similar depression, though developed to a less extent, exists also in Act. polymorpha, N^a. 2 

 (PI. VI. fig. 17, r.ar). 



(§ 46) Nearly all the observers who have studied Comatula have regarded the " Knopf," 

 or centrodorsal piece, as of essentially the same nature as the stem of the stalked Crinoids. 

 The first author who put forward this opinion was Schweigger ^; and Miller's views-, pub- 

 lished two years later, were fundamentally the same, though somewhat modified in 

 form ; for the centrodorsal piece was regarded by Miller as composed of two separate 

 pieces, one forming the floor of the cavity and the other its sides and rim. The former 

 was described by him as a pentagonal unperf orated plate, " analogous in situation to the 

 first columnar joint of the Crinoidea; but as it is not required to transmit the passage 

 to the alimentary canal ^ (no prolongation of the column existing in this animal), it is 

 without central perforation." 



The other or ventral half of the centrodorsal piece was regarded by Miller as an 

 annular or basin-shaped plate, representing the " pelvis " or basal circlet of Fentacrinus, 

 though he described it as marked externally by numerous sockets for the attachment of 

 the cirrhi, which in Pentacrimis are borne by the stem-segments only, and never by 

 the basals. 



Goldfuss, who in most points followed Miller's views, difi'ered from him considerably 

 with regard to the nature of the centrodorsal piece of Comatula mediterranean which 

 they had both studied ; and his conclusions, though not absolutely correct, are much 

 nearer the truth than those of Miller. Pinding most specimens to bear three rows of 



' Op. cit. p. G4. - Oj}. cii. pp. 129, 130. 



'•' It must bo remembered that the canal which occupies the centre of the Crinoidal stem was originally supposed 

 to be a continuation of the alimentarj- canal, and not, as we now know it to be, of the general perivisceral cavity 

 or coelom. 



SECOND SERIES. — ZOOLOGY, VOL. II. 10 



