224 BULLETIX 82, UNITED STATES NATIONAL MUSEUM. 



In the very young stems of the peiitucrmites the eolumnals are longer than 

 broad, as in the stems of the hirval comatiilids, and they arc bound together by 

 artieulations of the bourgueticrinoid typo exactly resembluig those in the larval 

 comatulid stem (fig. 1 43, p. 205) ; but after each nodal the columnals become shorter 

 (those of each iiiteriiodc being always alike) and the articular faces become slightly 

 modified from the original type so that after five or six internodes an approxima- 

 tion to the true peiitacrinite type is reached. The first internode and the terminal 

 stem plate have never been observcil; but from the striking similarity, even in the 

 number of the component parts, I believe that we are justified in considering the 

 lowest internode in a pontacrinite stem which has been observed (the first post- 

 nodal to the second nodal columnals, both inclusive) as strictly homologous with 

 the entire larval comatulid stem, exclusive of the termmal stem plate, and plus a 

 very slight modification m the direction of the adult pcntacrinite internodal char- 

 acters. The following hiternodes progressively become and more differentiated in 

 the direction of the adult; hence we may confidently assume that the preceding 

 mternodes entirely lacked the very slight specialization which we find in the first 

 observed ; in other words, that they exactly resembled the stems of the larval 

 comatulids. 



Now a young pcntacrinite possessmg but a smgle mternode, the cirriferous 

 nodal bemg the last columnal under the calyx, would be in all its characters prac- 

 tically identical with a larval comatulid at the time of the development of the first 

 whorl of cirri, at which time the basals have only just begun to undergo their meta- 

 morphosis into the rosette. 



In the comatulids no further development of the stem as a whole occurs, but 

 the centrodorsal — the nodal of the pentacruiite — is enormously enlarged and gives 

 rise to usually one or more additional whorls of cutI, and fracture takes place 

 between this enlarged topmost columnal and that just beneath, largely as a result 

 of the great proportionate decrease m the area by which this enlarged topmost 

 columnal is attached to the followmg columnal, assisted by a modification from the 

 primitive bourgueticrinoid type of the articulation unitmg the two in the direction 

 of the so-called stem syzygv' (just as the articulation between the nodals and the 

 infranodals in the jM'ntacrinites is modified) and a consequent weakening of the 

 union. The metamorphosis of the basals into the rosette, it should be noticed, 

 docs not occur until after the development of the first whorl of cirri; that is, until 

 after the last possible common stage of development between the comatulid and the 

 pcntacrinite. 



In the pcntacrinite, on the other hand, the nodal (the centrodorsal of the 

 comatulids) docs not enlarge; a smgle whorl of cii-ri is developed, and the union 

 between the nodal and the infranodal is transformed into a syzygy as in the coma- 

 tulids although, because of the absence of any enlargement of the nodal or of any 

 other growth change, this does not weaken it, or at least does not weaken it enough 

 to induce fracture. The pcntacrinite, instead of enlargmg the first nodal as do the 

 comatulids, proceeds to form another stem in which thefiret nodal occupies a posi- 

 tion analogous to the terminal stem plate in the original stem of both the pcn- 

 tacrinitcs and the comatulids, and this stem grows to exactly the morphological 



