476 BULLETIN 82, UNITED STATES NATIONAL MUSEUM. 



also there is no nuclear division, but the original nucleus moves outvcard from the 

 base into the middle of the filament. 



Before the formation of these trabeculre begins the rudiments of the first 

 15 hydroccele tentacles increase considerably in size and appear within the vesti- 

 bule as small elevations. Scarcely have they reached an appreciable height when 

 the 10 interradial tentacles appear at the base of each of the five original evagi- 

 nations. There thus arise from each primary evagination five tentacles, three 

 central, larger and older, and two lateral, one on either side, as yet rudimentary. 

 The central tentacle in each group lies in the midradial plane. While the exten- 

 sions of the hydroccele forming the 15 older tentacles are parallel to the major 

 axis of the body, those giving rise to the 10 interradial tentacles lean inward 

 toward that axis. 



All the tentacles are tubes composed of a single layer of epithelium ensheathed 

 in the ectoderm of the vestibule. The lumen is very narrow, narrower than the 

 diameter of the walls, and in the outer ends usually is not demonstrable on account 

 of the two walls being pressed together. 



The borders of the cells in the tentacle walls are not always visible. When 

 they can be seen the cells in a cross section of the tentacle appear as a truncated 

 pyramid with the base outward, in a longitudinal section as a prism or cube. 



In the older free-swimming larv» the commencement of the stone canal can 

 be recognized as a pointed process on the right (originally left) end of the hydro- 

 ccele ring. After the assumption by the latter of a position at right angles to 

 the major axis it still occupies the same place, between radii I and V, arising 

 from the wall of the ring toward the oral ccelome. It has become a short S-sliaped 

 canal showing little individual variation, which lies entirely in the compass of the 

 oral ccelome, between its two side branches. From its point of origin on the 

 liydroccele ring it turns toward the left and anteriorly, then ventrally and toward 

 the right. For a long time its end remains closed, though thrust against the left 

 inner wall of the parietal canal. The opening of the blind end into the parietal 

 canal usually takes place on the fourth day. 



The walls of the stone canal are composed throughout of relatively high cells 

 with centrally situated nuclei, exhibiting the histological features which in pre- 

 vious stages had characterized the walls of the hydroccele ring itself, and which 

 recur in the budding tentacles. As in these, the lumen is in most cases of con- 

 siderably less diameter than the walls. 



In the part lying near the hydroccele ring a broadening of the lumen anteriorly 

 and toward the left usually occurs, which directly before the entrance into the 

 parietal canal may expand into an ampullalike swelling. Here Seeliger found 

 in some cases that the border of the lumen was cross striped, though he was unable 

 to determine the presence of cilia. 



PtBIBTAL CANAL. 



The changes in the parietal canal during the early part of the attached exist- 

 ence are closely connected with those in the vestibule and hydroccele. 



