226 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



length of the first, whoii it gives rise to a second nodal; but this second stem is 

 slii'litly diflVrent from the first; the columnals are slightly shorter, and their articu- 

 lar faces are very slightly modified. This process is repeated, each subsequent 

 rei)etiti()n of the original stem, mainly from mechanical reasons incident to mcreas- 

 wc size, taking on more and more of the adult character, untU at last the perfect 

 pentacrinitc stem is developed, in which each internode is homologous with the 

 entire larA-al column of the comatulid. 



The basals of the pentacrinitc, though modified by increasing size, remain at 

 phylogcnetically the same stage as the basals of the comatulids at the point where 

 the comatulids and jx-ntacrinites begin to diverge in their stem characters — the 

 stage of tlic development of the first whorl of cirri; otherwise the pentacrmite 

 crowns and the comatulids develop along exactly parallel lines as evidenced, for one 

 thing, by their peculiar, but exactly similar, types of arm division and of arm 

 structure. 



It is evident, then, that the centrodorsals of the comatulids both ontogenet- 

 ically and phylogcnetically are the representatives of, and are therefore homologous 

 with, the nodals of the pentacrinites mdividually, as well as collectively, as sup- 

 posed by Thomson; whereas in the comatulids the single nodal is enormously 

 enlarged and modified in various ways and permanently attached to the crown, in 

 the pentacrinites each nodal merely marks a stage in the development of a long 

 and continuously growing stem. Thomson's conception of the centrodorsal as a 

 coalesced series of nodals probably was suggested by the Very numerous cirri com- 

 monly present on the centrodorsal of such genera as Antedon, and their arrange- 

 ment in more or less regular rows, each row being correctly considered as the equiv- 

 alent of a pentacrinitc nodal. 



The increase in the number of cirri in the comatulids over the primitive five 

 may be easily accounted for. Ordinarily the crinoid stem, both in its calcified and 

 in its uncalcificd structures, undergoes continuous growth until the death of the 

 animal, continually forming new columnals just beneath the calyx. The abrupt 

 cessation of the development of new columnals m the comatulids has not been cor- 

 related with a similar cessation in regard to the uncalcificd constituents of the stem, 

 which, unable to develop normally along the usual Imes of crinoid growth, have 

 become repressed within the centrodorsal and have found relief from this repression 

 in the formation of cirri whenever the ontogenetical development of the repressed 

 stem constituents calls for the formation of a ciiTiferous nodal. We thus have a 

 very curious condition ; for, although the centrodorsal itself is strictly homologous 

 with a single pentacrmite nodal, as well as with all the nodals collectively, the 

 soft structures within it arc not, for they arc homologous with the entire penta- 

 crinite stem, and are, in effect, an entire pentacrinitc stem prevented from acquiring 

 the normal elongate form. The pentacrinitc stem in its develo])ment eontinuoush' 

 produces nodals at regular intervals; the comatulid <-entrodorsaI contmuousl}- ])ro- 

 duces new cirri between the most proximal row of cirri and the proximal edge of 

 the centrodorsal in just the same way, and the progressive development of the 

 cirri on succeedmg nodals ui the pentacrmite is exactly duplicated in the comatu- 



