10 THE VOYAGE OF H.M.S. CHALLENGER. 



they do not appear within the middle portion of the centro-dorsal in the specimen of 

 Anteclon eschrichti figured in PI. I. fig. Sd, though they are comparatively large in other 

 forms of this type, as I have noticed elsewhere. 1 



The peripheral part of the ventral surface of the centro-dorsal is divided by ridges or 

 grooves into the five trapezoidal areas in which the radial plates are lodged, and they 

 are occasionally marked by more or less definite pits which receive the ends of the radial 

 axial canals, as already explained (PI. IV. figs. Id, 2d). In most Comatulse every two 

 fossse are separated by one of the five basal grooves which lodged the rays of the basal 

 star, to be described subsequently. They are sometimes comparatively insignificant, as 

 in Antedon antarctica (PI. I. fig. 6d), while in the Pacific species they are usually very 

 strongly marked (PI. II. figs. 1-56). On the other hand, if no basal star is present, the 

 radial fossse on the centro-dorsal are usually separated by tolerably sharp ridges as in 

 Antedon eschrichti (PI. I. fig. 8d), Antedon quinduplicava (PI. IV. fig. id), and Antedon 

 disciformis (PI. IV. fig. 2d). The last-mentioned species, however, has indications of 

 basal grooves at the proximal ends of these ridges. The grooves are fairly distinct in 

 both the species of Promachocrinus which I have examined, but though the radials are 

 ten in number, there are only five fossse on the centro-dorsal, the ventral surface of 

 which is distinctly pentagonal in outline, with its angles interradial, just as in Antedon 

 (PI. I. figs. \c, Id, 5). 



In fact, I know of no Comatula in which the general shape of the centro-dorsal is 

 not more or less distinctly pentagonal with its ventral ridges and angles interradial. 

 Wachsmuth and Springer regard this fact as indicating the probable presence of radially 

 situated under-basals in the Co matula -larva. Their extensive and important investiga- 

 tions into the structure of the calyx in the Palseocrinoids have led them to formulate the 

 following rule : 2 — " In species with under-basals, whenever the column is pentangular, 

 its longitudinal angles are directed interradially, the sides and columnar cirrhi radially." 



They proceed to state 3 that the centro-dorsal of Comatulse is interradial " and rests, 

 as in the Apiocrinidse, against the outer face of the basals, not within the basal ring " ; 

 while they continue — " upon this mainly we base the opinion that perhaps also the 

 Comatula? in their early larva had rudimentary under-basals. That these plates, if 

 present, were not observed, is not surprising, as they may have been very minute and 

 been covered entirely by the column." 



Whether this be the case or not, the statement that the centro-dorsal of Comatulse 

 rests against the outer face of the basals is a somewhat misleading one. The " outer 

 face " can only mean that which appears on the outside of the calyx ; and this, from its 

 very nature, cannot rest against the centro-dorsal, for it would then be internal and con- 

 cealed. 



1 Quart. Journ. Geol. Soc, 1880, vol. xxxvi. p. 47. 



4 Revision of the Palaeocrinoidea, pt. iii. sect. 1, p. 7 (229); Proc. Acad. Nat. Sci. Philad., 1885. 



3 Ibid., sect. 2, 1886, p. 298 (222). 



