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



Wyville made no further reference to them except to say that they sometimes remained 

 permanently in the adult Antedon, usually in groups of three or five. These last, 

 however, like those already noticed in Pentacrinus (PI. XIII. fig. 1), are merely periso- 

 matic plates developed in the tissue uniting the second radials. Good figures of them 

 were given by Dr. Carpenter ;* and it is by no means certain that they are a further 

 develoj^ment of the primitively single plates which appear between the orals and basals, 

 for the direct continuity of the two structures has never been definitely traced. 



I am inclined to believe that where they do appear in ordinary Comatulse these 

 primary calyx-interradials eventually undergo resorption like the orals and the anal 

 plate. But they are permanent in Thaumatocrinus (PI. LVI. figs 1-4), as are also 

 the orals (fig. 5). This remarkable genus has five calyx-interradials which rest on 

 basals and separate the radials just as in certain Ojahiurids and in some types of the 

 Palaeozoic Ehodocrinidse. It is much to be regretted that this extraordinarily interesting 

 form is only represented by a single individual. For the study of the distribution of 

 the axial cords within the calyx would have been of some importance. 



If Thaumatocrinus resembles the ordinary Comatulse and Pentacrinidse, the circular 

 commissure (PI. XXIV. fig. 9, c.co., i.co.) is formed by both interradial and intra- 

 radial commissures which connect the paired branches of the five jjrimary interradial 

 cords (PL XXIV. figs. 7-9 ; PI. LVIII. figs. 1-3 — ar) ; and the interradial commissures 

 must traverse, or at any rate, lie upon the inner surfaces of the interradial plates. But 

 the general embryonic characters of Thaumatocrinus lead me to think that the arrange- 

 ment of its axial cords must be more like that which occurs in BathycHnus. In this 

 genus the primary cords do not fork within the basal ring, but pass upwards through it 

 and enter more or less complete canals which are formed by the apposition of two 

 grooves, one on each of the contiguous lateral faces of adjacent radials (PL VII. fig. 6a). 

 When they reach about half the height of the radials they fork, and the resulting 

 branches themselves form the interradial commissures, entering the radials by the 

 apertures in their lateral faces. 



I cannot help suspecting that the same condition may occur in Thaumatocrinus, i.e., 

 that the primary interradial cords run upwards through the basals into the interradials, 

 and there fork, one branch entering each of the two radials which are separated by the 

 interradial lodging the primary cord. This simple condition would correspond very 

 well with the general embryological characters of the type, as revealed in other ways. 

 But owing to the want of material there is unfortunately but little chance of the above 

 hypothesis being verified or disproved. 



While resembling the Rhodocrinites in having five large plates separating the radials, 

 Thaunmtocrinus differs from most Pateocrinoids, with the exception of the Platycrinida?, 

 in the absence of any higher series of interradial plates. Except on the anal side these 



' Phil. Trans., 1866, pi. xxxiii. fig. 7 a, b. 



