144 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



a jirimitivo motainoric division, but for tho ju-imarily dorsal structures tho inter- 

 ainhilacral areas each represent a jirimitivo metanieric division. 



A very strong reason for considering the five dorsal motamoric units of the 

 ochinoderms to bo the five interradial areas plus one-lialf of the radial areas on either 

 side is that the connection between the internal structures and the exterior is 

 always hiterradial; the stone canals, the madroporites, and tho madreporic poras, as 

 well as the genital openings (except in the crinoids, in which the genital system is 

 scarcely comparable in a broad morphological way to that in the other echinoderms) 

 are always interradial, exactly as the connection between the internal structures and 

 the exterior, tho nei)hiidial, genital, or tracheal jjores, in hisects and crustaceans are 

 always in the middle of a metamere and never on the border line between two 

 metameres. 



Moreover, in the original ring of 10 coronal plates the interradial plates (basals 

 and genitals) are always much larger than the radial (infrabasals and oculars). 

 This in itself would suggest that these interradial plates indicate areas of phylo- 

 genetically greater significance. 



Furthermore, the teeth in the echiuoids, each of which moves out and back 

 like tho mandibles of the bilateral mvcrtebratcs, and the orals of the crinoids, which 

 have the same motion, are interradial, each undoubtedly occupying the center of a 

 somite just as do the mandibles of crustaceans and insects. 



But the most conclusive proof of tho extraordinary alternation between the 

 metamcric divisions of the dorsal and of the ventral portions of the body Ues in the 

 fact that the primordial tentacles and the coelomic chambers, ventral structures, 

 are developed in the center of the ambulacral areas, while tho primary nerves arising 

 from the dorsal nervous center lie in the center of the interambulacral areas. 



The unit of the pentamerous symmetry in the echinoderms, therefore, so far 

 as the calcareous structures and the nerves are concerned, can not be considered 

 as a single ambulacral system plus one-half of each of the adjacent interambulacral 

 systems, but must be regarded as a smgle interradius plus one-half of the ambulacral 

 systems on either side. Ventrally, however, the unit of the pentamerous symmetrj' 

 is tho radial ambulacral extensions of the variotis circunioral systems, all of which 

 are single. Thus, in the echinoderms, while the pentamerous symmetry of the 

 calcareous structures and dorsal nerves is strictly interradial in its arrangement, 

 that of all the other ambulacral structures is strictly radial, and we find two differ- 

 ent phases of the same type of symmetry in the same animal. But though more 

 organs are involved in the ventral radial pentamerous sj^inmetry than in the dorsal 

 interradial pentamerous symmetrj^ the latter is of far greater phylogenetical 

 significance; it has resulted from a fundamental readjustment of one of the most 

 significant systems of the ecliinodermal organization, accompanied by a profoimd 

 change in a system recognized as possibly the most diagnostic in comparative 

 morpholog}', while the former merely is the result of the extraordinary development of 

 five radial buds on each of the circumoral rings, made possible by the existence 

 of the latter. 



Now according to the former interpretation the five crinoid arms represent 

 five individual structures each complete m itself and each commencing with one of 



