70 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.41. 



ture has been described by various authors and some figures of A. 

 nohilis are here given as copied from Carpenter. In younger speci- 

 mens of this species the centro-dorsal is normally cirriferous. Subse- 

 quently, however, the cirri are lost and the centro-dorsal becomes 

 largely resorbed (PL 6, figs. 9-11). Still another variant is to be 

 observed in the case of such a form as Antedon laticirra (PL 8, figp .1,2). 

 But few cirri are borne by the centro-dorsal and these are of relatively 

 enormous size. Again, in many of the species of Comatulse, both 

 recent and fossU, the centro-dorsal bears large numbers of cirri either 

 regularly or irregularly arranged. The shape of the centro-dorsal 

 seems, therefore, to vary largely both as to shape, size, and number 

 of cirri borne. This variation seems to obtain not only among differ- 

 ent genera and species, but within the same species, dependent upon 

 the degree of maturity of the individual, 



i Definition of the centro-dorsal. — Before discussing even in a most 

 general way the interrelationships of the forms referred to the Coma- 

 tulae, or to endeavor to establish a connection between any of these 

 types and their stalked ancestors it will be necessary to enter rather 

 deeply into a discussion of the evolution of the centro-dorsal, the 

 organ that is so characteristic of these eleutherozoic forms. It may 

 perhaps be best to limit and define the term centro-dorsal. The 

 centro-dorsal may be held to consist in its simplest form of the 

 proximal columnal to which the infrabasals have been fused. Fur- 

 thermore, centro-dorsal as here used may only apply to a cirrifer- 

 ous organ. It is evident that the terms proximale and centro-dorsal 

 cover an essentially similar structure. Indeed, as here defined, the 

 only essential difference lies in the fact that one bears cirri and the 

 other does not. It is obvious that this distinction is a most arbitrary 

 one. It is nevertheless a matter of some convenience to draw such a 

 line. It is quite within the bounds of possibility, as will subse- 

 quently be shown, that the term "centro-dorsal" may apply to two 

 quite different structures. As above defined and as immediately dis- 

 cussed hereafter, the organ is held as having the same position gen- 

 erally assigned to it. After having indicated in a general way the 

 evolution of this type of centro-dorsal, the evidence for another and 

 divergent type will be given. 



Evolution of tJie centrals. — The question of the evolution of the 

 centro-dorsal involves necessarily the question of the evolution of the 

 centrale. This in tiu-n involves questions which are quite without the 

 province of the present paper. It is necessary, however, to state, 

 even though briefly, an opinion relative to these topics. The pres- 

 ence of a persistent proximal columnal has been used by Wachsmuth 

 and Springer and by subsequent authors as a character of prime 

 importance in the classification of the Crinoidea. It is on this struc- 

 ture that the Order Flexibilia was established. I do not believe that 



