380 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



Among those comastorids in whirh the centnxlorsal is reduced to a pentagonal 

 or stellate plate countei-sunk to the level of the radials a curiously specialized con- 

 dition obtains. 'Hie chambered organ and accessor}' structures primitively occupy 

 a position within the circlet of infrabasals, as it does in Isocrinus and in Metacrinus; 

 with the degeneration of those plates, as exemplified by Endoxocrinus in the pen- 

 tacrinitos, the chambered organ becomes more ventral in its location, and occupies 

 a position m the center of the basal circlet, corresponding to the position it formerly 

 occupied in the circlet of infrabasals; in most comatulids it is contained within 

 tine central cavity of the centrodorsal, and is bounded ventrally by the rosette, 

 wliich lies at the level of the dorsal surface of the radial pentagon (fig. 66, p. 93); 

 but in the comasterid species with stellate centrodorsals, it has again moved 

 ventrally, has l)eon entirely extruded from the centrodorsal, and has taken a posi- 

 tion within the radial circlet, exactly con-esponding to the position it formerly 

 held within the basal circlet, and before that within the infrabasal circlet (fig. 68, 

 p. 93). 



In two genera of comatulids, PromacJiocrinus and Thaumatocrinus, both known 

 only from the recent seas, each of the five radials has morphologically undergone 

 longitudmal twiiuiing or division which has resulted in the formation of two radials 

 (making 10 in all) each of which, so far as can be seen, is exactly like all the others. 



These two genera both belong to the Macrophreata, but to entirely different 

 families, Thaumatocrinus falling in the Pentametrociinidie near Pentametrocrinus, 

 and Promachocrinus falhng in the Antedonidse and in the subfamily HeUometrinse, 

 being very closely related to Solanometra, Anthometra and Florometra and, rather 

 less closely, -to Heliometra. 



Although Promachocrinus possesses 10 radials all exactly ahke, it possesses 

 the usual type of rosette and only five basal rays, each of which is situated directly 

 under the center of a radial. There are thus five radial and five interradial radials. 

 Although structurally and morphologically each interradial radial is the exact 

 counterpart and twin of a radial radial, its origin is altogether different. In the 

 early larva only radial radials occur, the interradial radials appearing at a con- 

 siderably later period as narrowly hnear interradial plates which rapidly increase 

 in size, give rise to a process on their distal edge, and finally become quite indis- 

 tinguishable from the original five radials, bearing post-radial series which also 

 are quite indistinguishable from those borne on the five original radials. 



In the genus Thaumatocrinus a young specimen of one sjiecies, T. renovatus, 

 has been studied, and the relationships of the radials of each of the five pairs arc 

 seen to be exactly as in Promachocrinus; in this specimen all five interradials have 

 reached a size not greatly inferior to that of the five original radials, though they 

 an; still much l(!ss convex dorsally, and one of them, the posterior, has given rise to 

 the rudiment of one of the five supernumerary arms. 



There are no basal rays in the species of Thaumatocrinus, but pseiulo-basal 

 rays are present. These are five m number, and are situated between alternate 

 radials so that the radials are grouped in five pairs, each pair lying in a depression 

 between two pseudobasal rays. Viewed tlorsallj' each of these pairs of radials 

 consists of the original radial to the left and the secondary (mterradial) radial to 



