NO. 1846. OU CERTAIN ELEUTHEROZOIC PELMATOZOA—EIRK. 69 



Figure 19, according to the explanation of the plate, is an adult 

 specimen, and figiu-e 18 certainly does not represent a very young 

 stage. Wachsmuth and Springer apparently overlook the fact that 

 Carpenter described the centro-dorsal in the later stages of develop- 

 ment as having interradial angles and conclusively proved that this 

 was purely a result of secondary growth. Indeed, as quoted above, 

 Carpenter states that in larvae in which only one or two cirri have 

 appeared, the centro-dorsal in the noncirriferous portion is radial, 

 while in the remaining portion the angles of the plate are interradial, 

 as a result of secondary change of form induced by the presence of 

 the cirri. 



Bather (1898, p. 425) again emphasizes this fact: 



That such secondary formation of stereom does take place is no hypothesis; it has 

 been described in Antedon by W. B. Carpenter, H. Biu-y, and others. In fact, the 

 odd thing about that genus is that the very features on which Wachsmuth and 

 Springer relied in their famous prediction that it would be proved dicyclic are of 

 purely secondary nature. ; 



Thus we find that the early form of the centro-dorsal in Antedon 

 is entirely dependent upon the nature of the adjacent basals and is 

 largely governed in subsequent growth by the addition of cirri and 

 the concomitant increase in size of the plate itself. 



It would have proved much more effective had Wachsmuth and 

 Springer applied their law to the primary verticil of cirri, instead of 

 attempting to make the centro-doreal conform to their law ; for these 

 cirri are radial in position, as they should be in a dicylic form. The 

 later cirri are given off apparently without regard to position. 



It is to be noted in the case of Antedon that the proximal columnal 

 fuses with the infrabasals at an extremely early stage in the ontogeny 

 of the animal. Indeed such fusion apparently takes place before thfe 

 constituent plates of the thee a have assumed definite positions and 

 have become apposed. Such a condition of affairs shows a high 

 degree of acceleration and clearly indicates the comparatively ancient 

 acquisition of a detached existence by this line. As a result of this 

 early fusion of the proximal columnal with the infrabasals the centro- 

 dorsal assumes a position withm the basal circlet as soon as the plates 

 acquire a definite form. This condition is quite different from that 

 to be noted in earlier and less specialized types, as will be indicated 

 in the discussion of the phylogeny of the centro-dorsal. 



In its adult expression among the Comatulse the centro-dorsal 

 assumes widely different forms. Primitively low and bearing but 

 few cirri, in some forms it reaches an extraordinary development. 

 In Antedon (Solanocrinus) scrohiculatus (PI. 8, fig. 3), for example, it 

 will be noted that the centro-dorsal is of great length. As opposed 

 to this condition of affairs a number of species of Actinometra have 

 the organ greatly reduced and noncirriferous. This type of struc- 



