278 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



increases in tliickness so that the inner core is bronght to an apex a considerable 

 distance from the ))omt of the terminal ohiw as a wliole. The relationsliip between 

 the central core and the cortical layer is analogous to that between the dentine and 

 the enamel in pointed mammahan teeth. 



In ontogenetically young develojiing cirri the terminal claw at first differs in no 

 way from the preceding segments in shape or size, except that it is rounded off at 

 the tip; during growth, however, it gradually becomes curved, slender, and pointed, 

 and commonly elongates with slightly greater ra])idity than the other distal seg- 

 ments, in certain cases, as in the genus Crinometra, involving the ])enultimate 

 segment in this elongation. In older or in regenerating cirri its growth is relatively 

 far more rajjid, and it becomes very long, slightly curved, and pointed, wliile the 

 following ossicles are as yet merely short cylinders, one-third or even one-fourth of 

 its length (see two enlargements in the lower center of fig. 382, p. 301) ; similarly its 

 growth ceases and it attains its ])erfect form long before the following segments 

 reach their full size. 



Probably the origin of the differentiation and of the specialization of the 

 terminal claw may be explained as the result of pure mechanics. The action of 

 gripping the soft but more or less resistant bodies of other organisms into wliich the 

 cirrus tip tended to penetrate to a greater or lesser degree has resulted in the ])aring 

 away or molding of the sides of the originally bluntly conical terminal segment, at 

 the same time causing a condensation of the cortical layer of stereom, and finally 

 resulting in the formation of a sharpened terminal spine, curved in the same degree 

 as the distal part of the cirrus as a whole. Tliis process would very quickly cause 

 the formation of a pronounced and perfected terminal claw, so that now we find that 

 character a very important feature in both the recent groups, the comatuhds and 

 the pentacrinites, which hve attached to the bottom or to other organisms b}' the 

 cirri. 



The penultimate segment in rare cases resembles the preceding segments (figs. 

 316, p. 273, and 356, p. 293), but it is usually modified more or less, tending to assume 

 certain of the characters of the terminal claw (figs. 314, 315, 317, 318, p. 273). 

 It is commonly somewhat tapering and of a lesser diameter than the segments 

 preceding, so that it appears smaller but proportionately more elongate, most 

 frequently about as long as broad, in contrast to very short preceding cirrals, 

 though in certain cases where the chstal cirrals are long the penultimate segment 

 may be somewhat shorter than those proximal to it, being intermediate in its pro- 

 portionate length between the terminal claw and the preceding cirrals (figs. 369, 

 370, p. 299). Its distal edge usually inclines inward (doi-salward) at a much larger 

 angle than the distal edges of the other cirrals (winch are nearly parallel to tlieir 

 proximal edges) , and therefore in lateral view the penultimate segment is roughly 

 trapezoidal, the base of the trapezoid being ventral. The dorsal surface is broadly 

 rounded and is never carijiate as is frequently the case on the preceding segments. 



There is less variation in the size and in the shape of the terminal claw and 

 penultimate segment than in any of the other elements of wliich the cirri tyre com- 

 posed, even than in the short basal segments. The sha])o and proportionate size of 

 the terminal claw is fairly constant when compared with the very variable shapes 



