354 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



tically functionless, and consequently greatly reduced; the transverse commissure 

 in the axillary is the representative apparently of the intrabasal commissure, the 

 commissure of both branches having beccnie superposed and merged into one. 



Now we know that the axillary is a double ossicle, arising from the fusion of 

 two ossicles interiorly with the result of forming the complicated chiasma; or, in 

 other words, the axillary represents a retarded phase in the transition from the bi- 

 serial to the uniserial type of arm. The exactly comparable structure, shown by 

 the nerve cords within the calyx, is just as evidently the result of the drawing apart 

 of the two derivatives from the primary interradial cord as the result of the fusion 

 of two ossicles exteriorly, an intermediate stage being seen in Encrinus. 



Viewed in this light the nervous system of the crinoid is seen to be after all 

 quite similar to that of the other higher invertebrates, especially to that of the 

 arthropods, instead of being unique as has commonly been supposed. 



In such fossil forms as have biserial arms it is to be remarked that at the arm 

 bases the brachials become uniserial; this is not to be interpreted as indicating 

 that the arms were originally uniserial, but quite otherwise; mechanical consider- 

 ations have forced the amalgamation of the two primitive radials into one, and simi- 

 larly have forced the uniserial arrangement "of the first two, and partially of the 

 third and fourth, brachials. The first four brachials, as will be shown later, are 

 intermediate in their character between the radials proximal to and the brachials 

 succeeding them; thus their relationship to each in the biserial arms is especially 

 instructive. 



Thus we have good evidence that the radials were primarily double ossicles 

 arranged in pairs, each pair superposed upon a single basal, just as the brachials 

 beyond them are primarily arranged in a double series, or else were primarily single 

 ossicles each superposed directly upon a single basal, each later dividing into two; 

 the five radials as we see them now resulted from the fusion of the primitive radials 

 into pairs exteriorly; that is, the two on each basal joined, not interiorly with each* 

 other, but exteriorly with those on adjacent basals. 



We know of no crinoids in which the radials are ten in. number arranged in 

 pairs over the five basals, each of the ten being the equivalent of half of a radial in 

 the forms in which the radials are five in number. Promachocrinus and TJiau- 

 matocrinus have ten radials, but each of these ten is the equivalent of one of the 

 five radials in allied forms or of one of the hypothetical original pairs, being, 

 though developed later, a perfect twin of the one lying at the side of it. 



Thus the dorsal portion of the ambulacral system of the crinoids (and of the 

 other echinoderms as well) is entirely a double system formed by the lateral union 

 exteriorly of ten interradial processes, though it supports ventrally single structures 

 arising from the prolongation along its ventral surface of various of the circular 

 circumoral systems. 



A consideration of the mechanical conditions affecting the structure of the cri- 

 uoids shows at once why ten single radials superposed upon the five basals are 

 never found. The echinoderms are divided into three or five radial divisions 

 because of the fact that the divisions are by lines of weakness and therefore must 

 be of some uneven number, for if the number were even the animal would be sub- 



