84 SMITHSONIAN INSTITUTION 



which is a feature not yet discovered in any Recent crinoid. With that excep- 

 tion the tegmen of this comatulid larva might almost be taken to be of the 

 same genus as those figured on Plates LXVII and LXVIII. 



The radianal, after being left far behind by the rectum when carried up- 

 ward with the anal tube, is resorbed along with the interradial plates, from 

 which in the later stages it scarcely differs, as is shown by figure 4 of Plate C, 

 a dorsal view of a specimen which has cast off the stem (the cirri beyond their 

 stumps being omitted in the drawing). 



It will be observed in this figure that the basals are no longer visible, hav- 

 ing been fused into the "rosette"; the orals in the stage represented by this 

 specimen have also vanished. Thus it results that the two rings of plates which 

 at the beginning of the series composed practically the entire calyx have now 

 completely disappeared; and instead of the globose cup which enclosed the 

 viscera in the youngest larval stage, there remains only a flat basin, serving as 

 a mere support for the relatively greatly enlarged visceral mass. 



To show the final outcome of the changes above described, I have figured 

 the tegmen of an adult specimen of the same species (PI. C, fig. 5). All that 

 is left of the several ventral structures is the granular skin, subcentral anal 

 tube somewhat flattened by shrinking, and the exocyclic mouth with the asym- 

 metric ambulacra, both permanently displaced by the pressure of the rectum. 



The origin and development of the radianal as shown in the Comactiuia 

 series is paralleled in a series of Promachocrinus kerguelensis, loaned me by 

 Mr. Clark out of a set of larval forms studied by him from the collection made 

 by the German South Polar Expedition near Gaussberg, in the Antarctic, and 

 now in his hands for investigation. I have figured a few of the specimens, on 

 account of the decisive way in which they reinforce the foregoing observations 

 and conclusions (PL A, figs. 10-14). The series begins with the pre-brachial 

 stage like our Comactiuia (figs. 1 and 2), in which the radials have just begun 

 to form as small dots, but in this case all the radials have appeared, that of the 

 right posterior ray being smaller than the others and the radianal much larger, 

 having already passed to the left of it. From the great difference in size of 

 these incipient plates it seems probable that the radianal, as already suggested 

 by Mr. Clark, actually developed before any of the radials were formed. In the 

 several stages which follow, the radianal does not increase to the relative size 

 which it attains in Comactiuia; but it runs a similar course until it is lifted out 

 from the ring of radials by the growth of the anal tube, and the interradial 

 spaces have become occupied by the interradial radials characteristic of this 

 genus. It must be remembered when examining the figures here given that 

 they do not represent the true relative sizes of the specimens ; for convenience 

 they are drawn of about the same general dimensions, but for the actual rela- 



