294 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



those found in the other genera of the Thalassometridae, there is, curiously enough, 

 never the slightest trace of any modification of the proximal pinnules toward the 

 type found in the other genera of that group. 



Dr. W. B. Carpenter has noticed in the growing young of Antedon bifida that, 

 as in other species, after the formation of the first two whorls of cirri no special 

 regularity can be traced in the manner of development; the young cirri normally 

 appear between those previously formed and the radial pentagon, so that their 

 sockets are close to the margin of the centrodorsal; but as the centrodorsal grows 

 and new cirri appear around its margin, the older cirri which are attached close to 

 the dorsal pole drop away and their sockets become gradually obliterated by cal- 

 careous deposit. The result is that the dorsal surface of the centrodorsal is usually 

 left comparatively smooth, but in some species the deposit of new material con- 

 tinues after the cirrus sockets are obliterated and causes the dorsal pole to become 

 rough and irregular. On the other hand, the lower surface of the centrodorsal in 

 most species of the Comasteridse is almost flat and extremely smooth. This is 

 owing to the very extensive and uniform manner in which the new material is 

 laid down. 



Dr. P. H. Carpenter noticed that the primary trunks which leave the chambered 

 organ, subsequently dividing and passing to the cirri in the corresponding radial 

 areas, usually undergo their division within the cavity of the centrodorsal. It 

 sometimes happens, however, that more or less of this division takes place within 

 the substance of the centrodorsal, so that interiorly there may be only one radial 

 opening visible, whereas outwardly there may be found the apertures of half a dozen 

 cirrus canals. 



In regenerating cirri the basal segments are the longest, and the following 

 decrease rapidly in diameter, so that the whole cirrus tapers considerably from its 

 base to its point. This condition gradually becomes less and less marked as the 

 segments increase in size and their apposed faces become beveled off toward the 

 dorsal side, so that the cirrus ultimately acquires all the characters of maturity. 



In the comatulids*only the first few rows of cirri are developed, as described by 

 W. B. Carpenter. The cirri which appear subsequently gradually assume certain 

 of the developmental features of regenerating cirri, so that at the adult stage, and 

 usually some time before that stage is reached, the cirri which are constantly pro- 

 duced about the ventral margin of the centrodorsal arise exactly as if they were 

 formed at an old socket from which the original cirrus had been lost. 



In very old specimens of certain species a peculiar condition is found among 

 these last formed marginal cirri, which was first noticed in Florometra mageUanica. 

 The cirri are formed just as regenerating cirri, but with increasing age the ontogeny 

 of regenerated parts becomes gradually retarded, so that in old examples the last 

 formed cirri never assume mature characters, but remain slender and tapering. 



As the assumption of a definite number of segments and the cessation of further 

 addition after the full number is reached is a true and definite growth character and 

 therefore dependent, like all other growth characters, upon the virility of the animal, 

 incipient senescence affects this likewise, and the marginal cirri of very old specimens 

 therefore possesses the number of segments characteristic of the adult, plus an 



