370 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



ft^rmation of a column, should occur, conditions woukl be quite different. Assuming 

 that ail the plates abut by their inner borders upon the periproctal area, it is evident 

 that the greatest width of the large plates is bej'ond the distal border of the small 

 plates. Thus a contraction of the circlet would gradually force the small ])lates be- 

 tween them inward so that the large plates would come into mutual contact beyond 

 them, forming a closed circlet with the closed circlet of small plates within it. 



A circlet of large plates in mutual contact with a similar circlet of small plates 

 within it is what we find in the crinoids in the circlet of basals enclosing the circlet 

 of infrabasals. 



But if the larger plates, interradially situated, came into contact with each 

 other beyond the small plates, radially situated, the latter would be entirely cut 

 off from the series of ambulacrals of which they formed the base. These would 

 therefore cease further growth and increase in numbers. 



Precisely this has happened in the crinoids; the development of the ambula- 

 crals comparable to those of the echinoid abruptly ceased in the phylogenetically 

 far distant past. 



Therefore the true homologies of the apical systems of the urchins and of the 

 crinoids seem to be that the large genitals of the former are the equivalent of the 

 large basals of the latter, and the small oculars of the former are the equivalent of 

 the small infrabasals of the latter. The oculars are extruded from the original 

 circlet of 10 altematuig large and small plates by a simple process of growth; the 

 infrabasals have moved mward from this circlet as a result of a contraction which 

 became necessary in order to meet the mechanical exigencies arising from the 

 development of a column. 



The individual radials in the comatulids are in close lateral apposition, usually 

 for nearly or quite their entire lateral length, so that the articular faces of atljacent 

 radials from the transverse ridge onward are barely separated from each other by a 

 narrow more or less shallow groove (figs. 431, p. 349, 441, p. 351). This groove 

 between the articular faces as a rule is broader and tieeper in the Macrophreata 

 than in the Oligophreata (reaching its maximum in the fanuly Pentamctrocrinidae) ; 

 in the young of certain macrophreate forms the radials may be entirely, and in the 

 young of certain oligoplireate forms partially, separated b}- intercalated interradials. 

 In the smaller species of the Oligoplu'cata the conditions resemble those found in 

 the Macrophreata; but usually in this group the intcrradial groove is reduced to a 

 minimum, both of width and depth. There are, however, some curious exceptions; 

 in the genus Pontiometra (fig. 432, p. 349) the radial faces are widely separated, 

 while in the Calometridse and in Comatilia not only are the radial faces widely 

 separated, but the radials extend upward in the angles of the cal_>-x, entirely and 

 more or less widely separatmg the bases of the first primibrachs, in several species 

 of the former and in the only known species of the latter terminatmg m broad 

 spatulate processes, each of these processes being composed of the anterior intcrra- 

 dial extensions of two adjacent radials. 



The dorsolateral edges of each radial are not sharp, but are more or less rounded 

 off, so that on the dorsal surface of the radial pentagon there are e^^dent five more 



