l8 SMITHSONIAN MISCELLANEOUS COLLECTIONS VOL. 76 



exterior margin of the cup, and also extends deeply inward, where 

 it interlocks with the tube plate by means of the socket. Among 

 fragmentary specimens I have a niunber of separated tube plates, 

 including several of the lower ones, all of which have the curved 

 socket as above described (pi. 2, fig. 17) ; and in one of them (fig. 

 18) the anal plate remains firmly attached by the socket in its original 

 position; in others (figs. 14, 16) the relation of the tube plate with 

 its socket to the raised process at the margin of the cup is shown as 

 it appears in the absence of the anal plate and arms. 



This is the structure of the parts at the posterior side in C. tennes- 

 seeae; and in C. ivachsmuthi I am able, after some careful prepara- 

 tion, to show the tube itself in position, with its interlocking key- 

 plate wedged into the socket precisely as it was in life — an admirable 

 device for anchoring such a ponderous appendage (pi. 3, figs. 4, 5). 

 The two other closely associated species, C. twbinatiis and C. shu- 

 mardi, have the same elevated process for supporting the anal plate, 

 and their structure in relation to the tube is undoubtedly the same; — 

 thus including in this class all species up to the earlier Keokuk. 



But in the two succeeding species with which the series terminated, 

 C. hradleyi of the later Keokuk and C. car pent eri of the Chester, a 

 curious change occurs — a striking example of the infinite variety 

 which characterizes the processes of nature. The raised platform 

 disappears, and with it the anal plate; instead of them a rather wide 

 plate rests directly upon the upper face of the radial, at the same level 

 as that of the other radials, filling its entire width except space enough 

 at the right side for one arm, and bending inward between the arms. 

 This is entirely different from the separate , anal plate above de- 

 scribed, but is itself the first plate of the tube. The connection is 

 beautifully shown in the two specimens of C. bradleyi (pi. 3, figs. 14, 

 15), in which the long tube is exposed from its origin on both right 

 and left sides. It is also well shown in a specimen of C. carpenteri, 

 on which I removed the arms sufficiently to expose the tube in a side 

 view (pi. 4, fig. 18). Other specimens of both species, figured on the 

 same plates show the form and proportions of the first tube plate seen 

 directly from the posterior. 



The course of the four antecedent species, having the anal struc- 

 ture first above described, was marked by a steady decrease in size. 

 C hradleyi, in which the change was introduced, took on renewed 

 vigor, with increase in size and relative number of arms. But the 

 change was not permanently advantageous, and the next species, C. 

 earpentcri, after undergoing still another modification in the arrange- 



