48 DISCOPHOR.^. Paet III. 



on every marginal segment ; the oculiferous lobes are as broad as they are long ; 

 the plicated concentric folds cover fully one half of the diameter of the disk, and 

 extend very closely to the borders of the digestive cavity; the space occupied 

 by the branches of the forked radiating canals extends as far along the circular 

 canal as the length of the branches themselves, and therefore has the form of 

 a triangle with a broad base ; the pouch beneath the sexual organs is propor- 

 tionately twice as broad as in the last stage, and projects much farther upwards 

 and toward the centre of the disk ; and, finally, the lips of the proboscis are very 

 deeply fringed, and the farrows in the prolongations of the corners are very deep. 

 Such, in general terms, are the characteristics of this phase, which Ave now 

 proceed to describe in detail. The lips of the proboscis (PI. XI". Fig. 5 «') are 

 not only very thin and flexible, but they also begin to show the tortuous plications 

 so characteristic in the adult. The fringes (^Figs. 5 a and 6 a) stretch in one 

 unbroken line, gradually diminishing in length from the ends of the prolongations 

 half way to the base of the same ; they are mere digitate diverticles of the outer 

 wall, and have the form of hollow tubes (PI. XP. Fig. 19) ; and in the wall, 

 especially at the tip, are imbedded numerous groups of lasso-cells (« I c). The 

 inner surface of the proboscis is lined by exceedingly minute vibratile cilia (PL 

 XP. Fig. G c), which are very difficult to detect even with a power of five hundred 

 diameters; they are as long as the thickness of the two walls {Fig. Q a l>) which 

 underlie them. No new branches have been added to the forking canals; but the 

 three branches on each side of the main channel have begun to anastomoze among 

 themselves, and a few anastomozing channels have developed at the base of the 

 oculiferous lobes, and extend a short distance toward the centre of the disk. These 

 have more the character of lacunae (PI. XP. Fig. 29 c^) than canals, and are so 

 intimately interwoven with the marginal canal, or that part of it (PI. XP. Fig. 5 c*) 

 which diverts into the oculiferous lobes, that they may be said to form, at least 

 in a certain degree, a part of the circular chymiferous system. Beside these, we 

 may see also, between the forks and the simple canals (PI. XP. Fig. 5 e), several 

 centripetal jJi'olongations from the marginal canal (mc), which, in time connecting 

 with the forked tubes, will increase their peripheric branches. In the specimen used 

 for this illustration there was an abnormal development of a canal (e^), which ran 

 from near the outer end of the simple canal (e), obliquely outward and across to 

 the exterior branch of the next forked radiating tube. That the tentacles are not 

 always strictly in a single row, may be clearly shown by an illustration of one 

 of their phases of development. The view (PI. XP. Fig. 13) which we present for 

 this purpose is partly sectional ; that is, the upper edge of the interlobular sockets 

 is left out, and the bases of the tentacles are exposed, in order to exhibit the 

 connection of the walls of one of them with those of the others. Nearest the eye 



