150 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



metamere became dissociated from each other, so that the pentaradiate echino- 

 dcmi body consists really of ten radial divisions, five "radial," representing the 

 ventral portion of the five original metameres (oriented most obviously by the five 

 primordial tentacles), alternating with five "interradial," representing the dorsal 

 portion of the five original metameres (oriented by the five "dorsal" nerves). 



The dorsoventral axis remains as it was originally; the anteroposterior axis 

 has become resolved into a circle; each of the planes originally passing through the 

 center of each metamere and crossing the anteroposterior axis at right angles has 

 become divided into a dorsal and a ventral portion, and the resulting ten planes 

 have become radially arranged with their iimer edges coinciding with the dorso- 

 ventral axis. 



The digestive tube, originally lying along the anteroposterior axis, has been 

 forced out of this position through the rearrangement of Ihe five half metameres 

 in the form of a closed circle, and either comes to coincide with the dorsoventral 

 axis (echinoids and holothurians) or to occupy a position at the ventral pole 

 (crinoids). 



In this connection a very extraordinary feature of crinoid morphology, which 

 has hitherto passed unnoticed, should be considered. In the bilaterally symmet- 

 rical animals development begins at the head and gradually works backward along 

 the anteroposterior axis of the body toward the tail. Thus when wo pass from 

 the tail of an animal (embryo or adult) towanl the head we pass over segments 

 (or groups of segments) of progressively increasing specialization and perfection, 

 the most highly specialized and the most perfect being found at the extreme anterior 

 end. In the crinoids the head, or what remains of the head, occupies an apical 

 position at the focus of the five radial divisions which represent the neural portions 

 of the five (originally thoracic) half somites. But the remnant of the head stiU 

 retains its influence as the center and, as it were, the originator of morphological 

 specialization and perfection. This progressive morphological speciahzation and 

 perfection makes itself felt not along the original axis (now reduced to a circle from 

 which the head is entirely detached), but along the five radial divisions which rep- 

 resent the axes of the neural portion of the five half somites of which the echino- 

 demi body is composed, as well as along the axis of the column; in short, along each 

 and every line which departs from the central nerve mass, no matter what direc- 

 tion it takes. Thus it is that, as the new brachials and new pinnules are added 

 (list ally, each successive brachial and pinnule is less perfect than its predecessor, 

 for it is developed at a greater distance from the morphological center of perfec- 

 tion; and as the columnals and the cirri receive accessions to their number only 

 between those already formed and the central nerve mass, each new columnal and 

 each new whorl of cirri is more perfect than those preceding. On account of the 

 apical situation in the echinoderms of what represents the head in the bilaterally 

 s\Timietrical invertebrates, each of the five dorsal radial divisions of the bo<ly, and 

 in the Pelmatozoa also the column, have come to assume to a certain extent the 

 developmental features normal to the neural portion of the body of a bilaterally 

 symmetrical invertebrate. This idea ma}- be roughly indicated by comparing the 

 crinoid body to a cluster composed of the neural portion of six primitive crustacean 



