MOKOGKAPH OF THE EXISTING CRINOIDS. 



195 



arms, of which the outer is more highly developed than the inner. At the time 

 this occurs the larva is already a perfect comatulid, and therefore the two new 

 arms are of the perfected comatulid type. The origin of Pj appears to be identical 

 with that of the short outer arms of Comatula etheridgei, with this difference, that 

 at the time the oral pinnule appears tlie pentacrinoid arms are still in the Flexibilia 

 stage. 



The very slender distal pinnules are ^composed of segments which, except 

 for the first two, are greatly elongated, and they bear no resemblance whatever 

 to the distal portion of the arms. They are true comatulid pinnules, correlated 

 with the true comatulid type of arm. At the same time they are obviously of 

 exactly the same nature as the oral pinnules, and it is noticeable that their first 

 two segments always maintain the same size relationship with the brachials as 

 do those of the latter, no matter how slender they may become. 



The genital pinnules are transitional between the oral and the distal pinnules. 

 They appear after the oral pinnules and are at first greatly reduced replicas of 

 them, but soon the distal segments elongate, and this modification gradually 

 extends itself basally, so that they merge imperceptibly into the pinnules of the 

 distal type. 



Thus while the enormously elongated comatulid arm has developed along 

 lines characteristic of and more or less peculiar to the comatulids, the arm branches 

 have remained permanently in the state of supplementary arms of the Flexibilia 

 type, from which they differ only in their greater slenderness. This explains the 

 curious uniformity in the number of pinnule segments throughout the comatulids 

 as a whole. At the Flexibilia stage the arms of the pentacrinoid larvae consist 

 of about 1.5 brachials, and most pinnules have between 1.5 and 20 segments. 



The view that pinnules are dwarfed arm branched or ramules is quite correct, 

 but in the comatulids it must be borne in mind that these dwarfed arm branches 

 belong to the ancestral type of arm and not to the perfected comatulid arm type. 



That the pinnules are in reality modified arms is shoM'n by the more or less 

 frequent replacement of a pinnule by an arm, or the reverse, and the occasional 

 bifurcation of a pinnule on the second segment, which then becomes axillary. 



In the ontogeny of the comatulids pinnules first appear at the tip of tlie larval 

 arm after the first 12 to 15 brachials have been formed. Pi does not appear until 

 from one to half a dozen or more pairs of distal pinnules are present, and the inter- 

 mediate pinnules appear from the most proximal of the distal pinnules one by one 

 downward to the oral pinnules after the arm has assumed more or less the true 

 comatulid character, and after P, is well grown. 



In certain types, as ComntiUa irhlomrtriforrnis (fig. 183, p. 98) and " Antedon " 

 impinvafa, the pinnules between the most proximal distal and the oral never appear 

 at aU, while in many others the last of these. P^, is never found except upon regen- 

 erated arms. In Metaciinux (fig. 287, p. 215), Hypalocrimis, and C omastrocrinus 

 (part 1, fig. 126, p. 195) the pinnules toward the arm tip suddenly become greatly 

 reduced and finally disappear altogether, so that the terminal portion of the arm 

 is strictly comparable to arms of the ancestral Flexibilia type. 



