MONOGRAPH OF THE EXISTING CRINOIDS. 337 



forms it is shoved outward beyond the radials into the primarily unplated portion 

 of the intermediate area, where it of necessity disappears. 



Anal x in the fossils develops between the two posterior radials, but probably 

 appears at a much later ontogcnetical stage. It thus develops along exactly the 

 same lines as the radials, giving forth, like the latter, a linear series of ossicles which 

 collectively represent the division series; but, handicapped by its late ontogenetical 

 origin, it lags far behind the radials in development, so that the ossicles following 

 it never reach the dorsoventral border line, and it remains as a partially developed 

 radial, followed by a series of interambulacrals which may be regularly arranged, 

 but which are never segregated and fused into pairs as are the ambulacrals arising 

 from the radials. 



The interradials of Promachocrinus and of Thaumatocrinus arise very early 

 in life and are from the first equal in height to the radials. They are probably in 

 these genera best interpreted as a sort of lateral budding from, or a delayed re- 

 duplication of, the radial to the left, and they are from the first equal in height to 

 the radials which they separate. As the radials move farther and farther apart 

 they continue to broaden, and their development in all ways is proportionate to 

 their breadth as compared with the breadth of the normal primary radials. 



Developing within the radial circlet, which they entirely span dorsoventrally, 

 their growth is in every detail parallel to that of the radials themselves, the differ- 

 ence in development between the two being at all stages proportionate to the 

 difference in breadth. 



The dorsoventral dimensions of the interradials are from the first equal to the 

 dorsoventral dimensions of the radials; therefore, as would be expected, the dorso- 

 ventral dimensions of the following ossicles are from the first equal to those of the 

 corresponding ossicles following the radials at the time of their formation. Devel- 

 oping under identical conditions, these plates develop in exactly the same way. 

 Reaching the border between the dorsal and ventral surface of the animal at exactly 

 the developmental stage at which this is reached by the ossicles arising from the 

 radials, thanks to the interradial sagging of this border line, the development of the 

 arms from the third brachial outward follows exactly the same lines as it does in 

 the arms of the primary radial series. 



It occasionally happens in Thaumatocrinus (and probably also in Promachocrinus, 

 though no instance has as yet been reported in that genus) that interradials occur 

 from which no arms arise, but which exist as broad single plates interpolated in the 

 radial circlet. These probably represent interradials delayed in development so 

 that they did not reach the dorsoventral border line, and therefore could not give rise 

 to the equivalents of postradial series. 



I have examined pentacrinoids of Promachocrinus kerguelcnsis in which both 

 the radianal and anal x are present, the former dwindling, the latter increasing in 

 size. They are situated side by side between the two posterior radials. 



Except for the large infrabasals and the position of the radianal farther to the 

 right and within the lower left-hand corner of the right posterior radial, the very 

 young of Promachocrinus kerguelensis does not differ in any essential particular 

 from the very young Antedon. The increase hi the number of arms is brought about 



