444 EDWIN S. GOODRICH. 



stoma, whilst in others this is not the case. These different 

 conditions may be dependent upon the state of expansion of 

 the thin-walled underlying nephridial sac. 



The structure of the nephridium can be very well studied in 

 this species. The nephridial canal is divided into several 

 branches which spread over the surface of the organ ; the 

 branches all converge towards and finally open into the canal 

 leading to the exterior, and their lumen is provided with 

 powerful long cilia (fig. 24). 



Coming off from the inner surface of these branches are 

 numerous secondary canals, which branch repeatedly, and 

 form a network throughout the substance of the nephridium 

 leading from one chamber to another (figs. 24, 29, 32, and fig. 

 15 ofGl. siphonostoma). The secondary canals appear not 

 only to branch, but to anastomose ; the canals coming from one 

 main branch, however, do not seem to open into those coming 

 from another. The system, except for the anastomosis, may 

 perhaps be compared to a river, tributaries of which are 

 separated by watersheds. Here and there the secondary 

 canals lead up to the very numerous chambers into which 

 open the tube-bearing cells. 



In Gl. unicornis the solenocytes are generally distributed 

 in pairs, never in groups, and are intermediate in structure 

 between those of Nephthys and those of the species of 

 Glycera described above. Instead of being entirely sup- 

 ported by the tubes, as in the latter, the cells are attached 

 at their base to the wall of the nephridium (figs. 9 and 32) by 

 a short stalk. The cell is more elongated, and a neck of 

 considerable length bends round from the body of the cell to 

 the top of the tube. Although in some specimens this neck is 

 quite long, yet it is always much shorter than in Nephthys 

 (fig. 10). The nucleus is large and oval. 



Such is the structure of a typical tube-bearing cell in Gl. 

 unicornis ; but in many cases these cells seem to resemble 

 those of Gl. siphonostoma (fig. 29). These apparent 

 exceptions may possibly be due to the worms having been 

 wrongly identified ; I am inclined to believe, however, that 



