252 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



central canal that is still left in the upper segments (at least) of the stem. This 

 perforation, however, is yctv soon closed up by an extension of the calcareous net- 

 work, so that no trace of it remains visible externally. Around the stellate aper- 

 ture is seen a circular series of five sockets for the articulation of the dorsal cirri, 

 each of them having a pore in its center which is usually at the summit of a minute 

 elevation. This pt)re gives passage to a sarcodic thread which proceeds from the 

 sarcodic axis contained within the cavity of the basin, and runs along the central 

 canal of the cirrus to its termination. A second series of sockets, alternating in 

 position with the first, is seen nearer the upper margin of the basin. This margin, 

 when viewed from above, is somewhat pentagonal; but the opening left by the 

 inversion of the lip is nearly circular. Throughout the whole period of growth the 

 increase of the centrodorsal takes place at a greater rate than that of any other 

 part of the skeleton, so that it soon comes to pass beyond the circlet of basals and 

 to abut on the proximal edge of the radials; instead of stopping here it continues 

 to increase in diameter until it conceals the whole inferior surface of the radials, 

 sometimes encroachuig on the first primibrachs. The increase in size from a diam- 

 eter of 0.05 inch to 0.16 inch, vnth a corresponding augmentation of its central 

 cavity is brought about by a continuous deposit of new material on the external 

 surface and a continual removal of old material from the internal surface. With 

 this general augmentation in size there is an increase both in the number of sockets 

 for the articulation of the doi-sal cirri and in the size of the individual sockets, and 

 there is also a marked change in their disposition. I [Cai-penter] have not been 

 able to satisfy myself that after the development of the fii-st two whorls, each con- 

 sisting of five cirri, any similar regularity is obsei-vable in their subsequent multi- 

 phcation; but since the real oi-igin of each cirrus is in a peduncle of sarcodic sub- 

 stance put forth from the central axis in the cavity of the centrodorsal basin, and 

 since the arrangement of the whole aggregate of such peduncles is distinctly verticil- 

 late, the want of a definite plan in the grouping of the cirri on the external surface 

 of that plate seems attributable to their very close apposition. The new cirri 

 always make their appearance between those previously formed and the base of 

 the calyTi:, so that their sockets are close to the margin of the basin. The increase 

 of the cirri in diameter is by no means proportional to the increase in di- 

 ameter of the centrodoreal, so that not only is space made on its surface for the 

 augmentation in the number of their sockets from 10 to between 30 and 40, 

 but a vacancy gradually comes to be left in the central part of the exterior of the 

 basin which extends with its growth and finally comes to bear a considerable 

 proportion to its diameter. This vacancj" can not be accounted for solely by 

 the widemng out of the innermost circle of sockets by the general growth of the 

 basin; it is principally due to a progressive loss of the first-formed cirri from within 

 outward, and the fiUing up of their sockets with new deposit, concurrently with the 

 formation of new cirri about the margin. Thus it appears that the total number 

 of cirri developed during the fife of any individual Antedon considerably exceeds 

 that %vilh which we meet at any one epoch." 



In the Ohgophreata the cirri are tenacious, and are seldom to any cxtenl lost 

 by the process of capture, no matter how rough the treatment accorded them may 



