494 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



the esophagus facing the posterior end of the gut. As in the central section it is 

 here also enveloped by a fold of the coelome wall, which in the outermost part, 

 however, is less closely connected with it, so that the axial organ is immediately 

 surrounded by a small space, which can be traced back to the primitive body 

 cavity, and has no connection with the true ccelome. On the oral side this cavity 

 ends under the ectodermal oral disk, to the inner wall of which the fold of the 

 cnplomic wall enveloping the axial organ is attached, after broadening out like a 

 funnel. Aborally it disappears, since the axial organ entirely fills the cavity. 



Toward the aboral end the axial organ, as already described, passes into the 

 vertical mesentery of the aboral coelome, traverses the central space which is 

 inclosed by the chambered organ, and finally loses itself between the cordlike 

 processes of the latter. How far it extends into the stem has not been determined. 



Sometimes in the oldest larvae of this period the aboral part of the axial organ 

 almost from the entrance into the chambered organ appears merely as a cell 

 cord without a central lumen. 



In its short course in the vertical mesentery it possesses in some cases a 

 special sheath which appears as a fine pavement epithelium. 



In the portion of the axial organ lying at the outermost end of the ccelome 

 the conditions are as they were in the oral portion at its first appearance. 



HTDROCCELB. 



In this stage the hydroccele ring usually becomes entirely closed by the 

 growing together of the two blind ends in the interradius I-V. The septum com- 

 posed of a double thickness of pavement epithelium representing the apposition 

 of the two blind ends ruptures in various places and finally disintegrates into a 

 number of trabeculse. But among larvae of five weeks individuals may still be 

 found in which the two layered septum still persists. 



The lumen is rather extensive. Its shape in cross section varies according 

 to the state of contraction, but apparently it is usually broader in the direction 

 parallel to the major axis than in the transverse plane. 



The wall is everywhere composed of a single layer of high cells, which become 

 larger in a zone running under the nerve ring. Along this zone ring fibrillae 

 appear which surround the whole hydroccele ring and form a sort of sphincter. 

 They lie on the outer side of the hydroccele wall and in cross sections are evident 

 as strongly refractive elongate or peg-shaped bodies. In longitudinal view the 

 fibrillae are seen to be cross-striped, and the nuclei appear in general to be 

 arranged in longitudinal rows. 



In some of the oldest larvae the ring fibrillae occur not only on the region 

 lying on the nerve ring, but also on the entire inner wall and on a part of the 

 aboral wall, where, however, they are very delicate and not entirely regular, not 

 accompanying the entire canal in an unbroken layer, but forming a more or less 

 complete investment of longer or shorter ring fibers. 



These muscular fibrillse arise from the hydrocrele wall, and not from mesen- 

 chyme cells. 



