A MONOGRAPH OF THE EXISTING CRINOIDS 51 



recent forms. Gisle'n said it has been supposed by the author that from the consider- 

 ably larger size of the centrodorsal cavity in the young we might conclude that a small 

 centrodorsal cavity represents a more specialized stage. This feature according to 

 Gislen has just as little phyletical significance as the longer segments in the young, 

 and must be attributed to an as yet unfinished calcareous deposition in the walls 

 around the chambered organ. Small and weakly calcified species always have, in 

 comparison with the usual type in their family, an unusually large centrodorsal 

 cavity. 



The stronger development of the basals is a more primitive feature, noticeable 

 both during the phylogeny and during the ontogeny of the comatulids. 



The articular faces of the radials seem originally to have been moderately de- 

 veloped and to have had an inwardly broader muscular fossa of about the same size 

 as the interarticular ligament. Later this muscular fossa was reduced to a low broad 

 band which may even disappear entirely, as in the Comasterida or Mariametrida, or 

 may become enormously enlarged in comparison with the interarticular ligament, as 

 in the Macrophreata. This expresses itself in the ontogeny by the young of the group 

 first mentioned having relatively larger muscular attachments, while the young in the 

 latter group have relatively smaller ones. A small radial cavity is found only in the 

 later comatulids, from the Upper Cretaceous onward. A radial cavity filled with 

 spongy calcareous tissue is never found in the earlier cornatulids, but occurs in the 

 recent Comasterida and Mariametrida. 



So far as we are acquainted with the fossil comatulids, syzygies appear in the 

 older forms, but synarthries not until the younger ones. In the great majority of the 

 older forms the syzygies seem to have been few in number, and their septa few. 

 The arms were usually 10 in number, sometimes 5; if greater, the number of the 

 component parts in the division series was very variable in older tunes. 



The evolution of the comasterid type. The oldest comatulids we know are found 

 in the lowest Jurassic and belong to the comasterid type. Even then this type had a 

 highly developed centrodorsal with small, closely set, alternating cirrus sockets, 

 usually indistinctly or not at all sculptured, and rather strongly reduced basals. It 

 may therefore be assumed that the comasterid type even at that remote period had 

 a fairly long developmental course behind it. The primary characteristics of the 

 older forms were the extremely small centrodorsal cavity and the inwardly broader 

 fairly large muscular fossae on the radial articular faces. The characteristics which 

 even today distinguish the comasterids were developed early. The first form without 

 cirri is met with in the Lower Cretaceous. The centrodorsal is always more or less 

 flattened. The radial fossae are from the very first vertical, and the radial cavity 

 large, its central depression broad and deep, in recent forms filled with spongy calcare- 

 ous tissue. The muscular fossae in the younger forms are low and broad. The arms 

 in fossil forms are very little known, and so far as we are acquainted with them they 

 are 5 or 10 in number and uniserial. Two unique features which in recent times 

 especially distinguish this group are the exocyclic mouth and the pear-shaped organs 

 instead of sacculi. 



As primitive features which are retained in forms now living may be reckoned 

 the clumsiness, the often undeveloped synarthries (as in Comatula and Comaster), the 

 usually variable number of division series and of their component ossicles, the oc- 



