922 
MR. P. H. CARPENTER ON A NEW 
separate the radials (r) from the centrodorsal (cd), which is, as yet, but slightly 
differentiated from the stemjoints below it. In all cases, however, the basals become 
concealed very soon after the conclusion of the Pentacrinoid stage, if not before. I 
can find no trace of them in various abyssal Comatulce, which are no larger than 
Thaumatocrinus renovatus, and suspect therefore that in this type they persist 
through life as they do in Atelecrinus. Were they really only larval basals, and 
destined to have been eventually transformed into a £ ‘ rosette,” Thaumatocrinus 
would present a still more curious combination of characters than it actually does. 
Both the persistence of the basals and the considerable development of the orals 
are characters which, either singly or combined, would cause the type to be regarded 
as one of no little interest; but they are altogether cast into the shade by the other 
peculiarities of the calyx, viz., the complete separation of the radials by relatively 
large interradial plates, and the presence of the anal appendage. 
It might perhaps be suggested that the ten-rayed Promachocrinus affords some 
explanation of the first of these points. In this genus * the basals only exhibit a 
pentamerous symmetry, for the rosette into which they become transformed is 
connected with a basal star of five rays only, just as in any other Comatula. These 
five basal rays are attached to the dorsal surfaces of five out of the ten radials so as 
to partially separate them from the centrodorsal. These radials, therefore, are really 
interradial in position, and so correspond to the five interradial plates of Thauma¬ 
tocrinus. But here the resemblance ceases ; for the “ interradial radials” of Promacho¬ 
crinus precisely resemble the five true radials with which they alternate, and the 
arms borne by the two sets of plates are indistinguishable. I cannot, therefore, re¬ 
gard Promachocrinus as anything but a very regular variation of the usual pentamerous 
symmetry of the Crinoids. 
Failing Promachocrinus, there is no other Neocrinoid which presents anything like 
the remarkable morphological condition of Thaumatocrinus. To find a corresponding 
developmental stage we must go back to a very early period in the ontogeny of a 
Crinoid, i.e., one but little later than the appearance of the rudiments of the lowest 
arm-plates. The radials first appear as isolated plates in the spaces <£ where the 
upturned angles of two oral plates are opposed to the bevelled off upper angles of two 
adjacent basals.”! They gradually increase in size, and ere long come to form a nearly 
complete circle, two of them being separated for a time by the anal plate. This is 
eventually lifted out from between them, but the radials of Ant. rosacea do not come 
into complete lateral contact until after the appearance of the first whorl of cirri. In 
the larva represented in Plate 71, fig. 6, however, the cirri do not appear until the 
radials have met laterally, and the arm-bases are well developed. The radials of a 
mature Comatula, therefore, form a closed ring of five plates ; and any interradials 
which may subsequently appear are limited to the angles between adjacent second 
* Proc. R. S., No. 194, 1879, p. 385; see also Journ. Linn. Soc. Zool., vol. xv., p. 214, pi. 12, fig. 2S. 
f C. Wyville Thomson, “ On tlie Embiyogeny of Antedon rosaceus ,” Phil. Trans., 1865, p. 528. 
