222 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



Ontogenetically it has been conclusively proved in all the comatulids of wliich 

 the young are known that the centrodorsal is a single columnal, and is never formed 

 through a fusion of two or more, and such evidojice as we have ])ouits definitely to 

 the conclusion that it is phylogenetically also a single columnal, homologous with a 

 single nodal columnal, and at the same time with all the nodal colunmals collectively, 

 of the ])cntacrinitos. 



The centrodorsal of the comatulids is the exact equivalent of the so-called 

 proximale or "centrodorsal" found in very many of the Flexibilia. In these forms 

 at some undetermuied period in the ontogeny the infrabasals fuse with the topmost 

 columnal, which eidarges and, together with it, form a structure rcraaniing always 

 in permanent union with the calyx, the new columnals, if any be subsequently 

 added, being formed either directly beneath it, or by intercalation between the 

 columiaals already existmg beneath it. 



The centrodorsal of the comatulids is formed in exactly the same way, and 

 mamtains exactly the same relationship with the infrabasals and with the other 

 plates of the calyx. 



In such families as the Bourgueticrinidse and Apiocrinidae (both of which 

 include recent species) some forms possess a primitive proximale while others do 

 not, and we find an exactly parallel condition in the pentacrinite-thiollericrmite- 

 comatulid group, which collectively forms a precise equivalent to either of these 

 families. 



In Thiolliericrinus, which represents in all essentials the basic type from which 

 both the pentacrinites and the comatulids have been derived, through specialization 

 in exactly the opposite direction, there is a proximale which is the exact equivalent 

 of that in such genera as Bourgueticrinus and MUlencnnus, the only difference beiiag 

 that it is cirriferous instead of noncirriferous, a difference of no particular morpho- 

 logical consequence. In the comatulids this proximale has usurped the functions 

 of the entire stem which, having become useless, is now discarded before the adult 

 stage is reached. In the pentacrinites the topmost columnal, though enlarged, 

 never succeeds in forming an attachment with the infrabasals; this incipient ])roxi- 

 male formation, resulting only in the enlargement of the proximal columnal, con- 

 tinues throughout the life of the individual; each columnal formed just under the 

 calyx is an incipient proximale, but never becomes fused with the infrabasals ; pushed 

 outward from the calyx by the formation of another nodal columnal between it and 

 the calyx, it later becomes separated from the colunmal which preceded it by a 

 series of intercalated mternodals so that in the stem of the adult pentacrmito we find 

 a series of incipient proximales or nodals, cirriferous as in the comatulids and in 

 Thiolliericrinus, separated by a series of unspecializcd columnals or internodals. 



Phrynocrimis alone of the recent stalked crmoids appears to possess a jiroximale 

 of the i)rimitive type, and in this genus the columnals are all uniform in structure, 

 just as in the larval comatulids. But in all the other genera (or at least in nearly all 

 of them) inci])ient proximales occur as modified columnals throughout the stem, 

 with increasing frequency towai'd the calyx, each representing an attempt to form 

 a proximale. 



