284 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



In a few cases, as for instance in Antedon, Alastigometra, and in the genera of 

 the Comactiniinse, provision is made for this flexibility (which, however, is only 

 moderatolj' developed in these forms) by the beveling off or catting away through 

 resoqition of the dorsal distal ends of the segments below (dorsal to) the transverse 

 fulcral ridge (figs. 312, 313, p. 271). Usually no such adaptation is found, or if 

 present it is so slight as to be inadequate to servo the purpose; instead, the motion 

 of one segment iipon another and consequent mtermittent compression of the dis- 

 tal edge of the latter, working in connection with the progressive difference in the 

 size of the tlorsal and ventral ligament bundles, has resulted in the swelling or in 

 the eversion of this distal edge which rises obliquely upward as a broad thickened 

 rim or as a crescentic serrate transverse ridge. 



In a few species with comparatively primitive stout cirri, such as those belong- 

 ing to the genera Catoptometra or Tropiomefra (fig. 356, p. 293), or to the genera 

 included in the family Charitometridae, no further development is found; the ])lay 

 of the distal segments upon each other is made possible by a turning outward of 

 the <lorsal distal edge of each ; but in most cases such a condition is found only in 

 the more proximal of the segments bearing dorsal processes; as the amount of 

 ])ossiblo intersegmental motion gradually increases distall}^, we find that the pro- 

 duced distal dorsal edge of the segments gradually becomes more prominent, 

 increasing in height and becoming more and more erect, at the same time, on 

 account of the progressive dorsal carination of the segments, becommg progressively 

 narrower and moving mward from the ends of the segments to a subterminal or 

 even median jiosition, so that the dorsal processes have, on the subterminal seg- 

 ments, become sharp spines situated in the subterminal or median portion of the 

 dorsal side. 



The dorsal spines commonly are of a slightly more dense composition than the 

 remainder of the segments which bear them; though in some species they may 

 for a greater or lesser distance inward from the end of the cirrus be tipped with 

 vitreous condensed stereom, the amount of this tipping rapidly decreases prox- 

 unally on succeeding spines. The progressive distal mcrease ui height and erect- 

 ness, and the progressive attainment of a position further and further removed 

 from the extreme distal edge, are to be explained by the correlation in the develop- 

 ment of these structures and the progressive difference in size between the dorsal 

 and the ventral ligament bundles by which the cLrrals are articulated; where tliis 

 difference is greatest, the dorsal processes were first formed, and as the dorsal 

 processes developed here are the oldest, the}' have become the most i)erfected. The 

 transformation of the original transverse ridge into a sj)ino may bo simply a normal 

 growth change, or its origin may be mechanical along the lines suggested for ex- 

 plaining the original shai-pening of the termmal claw. 



In species ha\'ing the cirri unusually broad, as m the species comi)t)sing the 

 genera of the Colobomctrida? (figs. 345-348, p. 289, 349-352, p. 291, and 353-355, 

 p. 293), the i)rimitive transverse ridge does not simi)ly become more and more acute 

 and soon resolve itself into a spine as is commonly the case, but the cirri become 

 flattened below, and the originally crescentic transverse ridge resolves itself into a 

 sharp flattened serrate ritlgc (as in Oligomeira and in Prometra), bi- or tricuspid 



