316 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



from each other by a sejjtiim within which runs the canal containing the 

 genital rachis. 



At the base of the arm somewhat different conditions are found. A section 

 through the first brachial shows the ventral canal without a median septum, while 

 the septum between it and the dorsal canal is much thickened and is traversed by a 

 large number of intercommunicating cavities, of which one contains the genital 

 cord. Passing outward along the arm these last mentioned cavities decrease 

 gradually in number until finally only the genital canal is left. 



The dorsal canal in Antedon mediterranea is much broadened in places, while in 

 the vicinity of the interbrachial musculature it expands dorsally so as to frequently 

 come, furrow-like, within a short distance of the dorsal nerve cord. Similar swell- 

 ings are also found in Ileliometra glacialis, but in this species are more complicated 

 in structure, the cells of the dorsal furrows being thickened and bearing long cilia. 

 Septa often appear, which cut off this ciliated furrow from the remaining portion 

 of the brachial extension of the body cavity, and as often vanish, so that all tiie 

 cavities in the arms (with the exception of the water vessels and the small subneural 

 schizocsele canal) are intercommunicating. 



In lieliometra glacidlis the epithelium lining all these cavities presents a very 

 varying aspect. It reaches its greatest development in the ciliated grooves. In 

 the section lying dorsally its cells are smaller and more flattened, while in the inter- 

 communicating canals, between the dorsal and ventral canals, it is composed of 

 flattened cells without cilia. 



It is from the sections of the dorsal canal furnished with ciliated grooves that 

 the dorsal canals of the pinnules, which bear the ciliated sacks, arise. 



ALIMENTARY CANAL. 



P. H. Carpenter observed that both in Antedon and in Heliometra the gullet 

 runs downward and backward, trending slightly to the left. When it has readied 

 the bottom of the visceral mass the intestine turns to the right and coils around 

 its anterior side until it has reached the hinder part of the disk behind the com- 

 mencement of the first coil where it turns upward and slightly forward, to end 

 in the anal tube. The descending portion of the fore-gut is comparatively short, 

 and the coiled intestine into which it passes lies spread out on the upper surface 

 of the radials, so that there is no general dilatation of the gut at the bottom of 

 the cup as in the pentacrinoid. 



He found that in the majority of the Comasteridse the digestive tube proceeds 

 directly downward to a point somewhat behind and to the left of the center of 

 the disk, and then commences to wind. Its direction, as in the endocyclic species, 

 follows the hands of the clock when seen from the ventral side, but there are four 

 coils instead of one. The first coil occupied the extreme edge of the lowest ])art 

 of the disk, and consequently passes in front of the mouth so as to appear beneath 

 it in longitudinal section. The second coil passes immediately behind it and is 

 followed by two more in an ever-narowing but ascending spiral which terminates 

 in the more or less central anal tube, which is often some little distance in front 



