214 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



enormous swelling by external accretion of the calyx plates, which has also affected 

 the lower brachials and, together with the proximale, the columnals immediately 

 below it in rapidly decreasing degree. As enlargement is the chief factor involved in 

 the differentiation of the proximale from the other columnals, it naturally follows that 

 all columnals enlarged will take on the characteristics of proximales in proportion 

 to the amount of their enlargement. Thus in Apiocrinus we find not a single 

 proximale at the summit of the column, but a series of them of diminishing size, 

 distally grading more or less gradually into columnals of the usual type. 



The series of short discoidal columnals at the summit of the stem in Monachocri- 

 nus is the cone-shaped structure seen in Apiocrinus in an atrophied and obsolete con- 

 dition; it represents a group of imperfect proximales which occurs in these genera in 

 addition to the imperfect proximales found at rapidly increasing intervals toward the 

 distal end of the column. 



The pentacrinites also form a continuous series of proximales (called in this case 

 nodals) immediately beneath the calyx in exactly the same way; but in this group 

 stem growth is so exaggerated that intercalation of columnals at once begins and 

 progresses so rapidly that by the tune the proximales (nodals) are fully developed 

 they are separated from each other by from one or two to as many as 40 or more 

 columnals of the ordinary type in the group. 



Intercalation of columnals also occurs in Apiocrinus, but in this genus it is so 

 feebly evident as to be quite negligible as a factor hi column building. In Proiso- 

 crinus (fig. 128, p. 199), while the lower part of the column resembles that of Apio- 

 crinus, the proximal half has taken on the characteristics of the column found in the 

 pentacrinites. 



The repetition of the proximale throughout the length of the column in Monacho- 

 crinus and allied genera with decreasing frequency toward the distal end, and its 

 repetition in the pentacrinites at perfectly regular intervals, is singularly similar 

 to the conditions which we find in the arms. 



In the arms the axillaries (figs. 81, p. 134, and 164, p. 227) are all primarily redupli- 

 cated radials, and the radials themselves, like the proximales, are secondarily, not pri- 

 marily, calyx plates ; each one of the axillaries forms the base of what is essentially 

 an entirely new series of brachials, in exactly the same manner that the radial forms 

 the base of the pos1>radial series as a whole, and the proximales form the end of a 

 completed column. 



In extraneous division of the type occurring in Metacrinus the axillaries occur 

 with decreasing frequency toward the tips of the arms, just as the reduplications of 

 the proximale occur with decreasing frequency toward the distal end of the column in 

 Honachocrinus and its allies ; furthermore, with increasing distance from the calyx the 

 less perfect do the reduplications, both of the radial and of the proximale, become. 



In interpolated division as we see it in the comatulids and in all the pentacrinites 

 excepting Metacrinus (as well as in many other diverse types) the repetition of the 

 radial (forming the axillaries) occurs at regular intervals, just as the repetition of 

 the proximale occurs at perfectly regular intervals in the column of the pentacrinites; 

 moreover, the reduplications both of the radial and of the proximale are all exactly, 

 or very nearly exactly, alike, all being singularly perfect. 



