198 PROCEEDINGS OF THE ACADEMY OF [1881. 



In some of these species the two structures are found combined, 

 in others so closely associated that it is next to impossible to 

 separate them ; and in still others the extremes are separately 

 represented. Generic divisions based upon these variations, unless 

 accompanied by other distinctive characters, seem to be unneces- 

 sary and inexpedient. It would require the creation of a large 

 number of new genera, which would probably have to be sub- 

 divided to accommodate other transitional forms, and so on with- 

 out end. 



The double series of joints resulted from the increasing width 

 and outward growth of the arm ; hence arm joints which originally 

 Avere simple and cuneate did not in the mature animal extend 

 througli the full width of the arm, but gradually interlocked by 

 their shai'p angles, so that joints on each side, which at first were 

 separated b}^ another joint, came by degrees to be partly in con- 

 tact and to rest upon each other. Therefore, in double-jointed 

 arms every joint at each side bears a pinnule, while in those with 

 single joints the pinnules are found only on alternate sides. 



The arms in the Actinocrinidaj and Plat^crinidte divide rarely 

 af:er they become free, the branching as a rule taking place in the 

 body or in the free rays, MegUtocrinus^ Amphorac7-inus, and 

 Periechocrinus form the only exceptions. In the Rhodocrinidte, 

 on the contrary', the arms branch as a rule beyond the calyx. 



The ventral grooves in the arms of this family are less deep, 

 but comparatively wider than those of the Cyathocrinidae. They 

 are bordered on each side by a row of long pinnules, which cover 

 them perfectly. Whether the grooves were closed b}" marginal 

 plates seems to us doubtful, although Prof. Wetherby states that 

 he has observed such plates in Glyptocrinus. they probabl}^ occur 

 below the bifurcation where the arms should be regarded as free 

 ra^^s, or perhaps they are restricted to oral arms, such as P. Herbert 

 Carpenter describes in Actinometra 



The pinnules throughout this family are long, closely crowded 

 together laterally, and the two rows with which each arm is pro- 

 vided fit together so neatl}'^, and cover the arm furrow so perfectly, 

 that additional plates were scarcel}^ needed. Each pinnule is com- 

 posed of a number of joints, which differ in form in different genera. 

 In some they are of equal width .and height, outwardly convex ; 

 in others higher than wide, with the outer surface flat ; in some 

 they are entirely smooth, and in others provided with a peculiar 



