302 SoiifJicrii Cross. 



iviiiiiiiis bill IVoin t'onr to six lart,'C'r cavities. Horizontal sections, 

 tlirnii'Ji specimens which exemplify tlie next stage of development 

 of the brood chambers, taken through the animal just below the zone 

 of invagination, show at that level several cavities (usually four in 

 IT. sulcata, six in U. carlgrcni), filled with compact solid masses, and 

 occupying the same relative position as the cavities in the earlier 

 stage. A'ertical sections passing through the invagination zone show 

 these cavities at this stage to be more or less closed to the exterior. 

 Figs. 12 and 13 are illustrations of such sections, the former of 

 U. carlrjreni, the latter of U. sulcata. In the former there are 

 curious interlacing outgrowths of the mesoglcea carrying the inva- 

 ginated ectoderm with them, and in this way a complete closing of 

 the brood chamber to the outside world is effected. In the latter 

 case, which is taken from the specimen figured (fig. 1), although the 

 mouth is closed by the juxtaposition of the invaginated ectoderm, 

 there is ouly very slight indication, at the lower end of what was the 

 opening to the chamber, of the interlacing outgrowths of mesoglcea 

 shown as being present in the U. carlcjreni example. It is probable 

 that this latter condition is a slightly earlier stage in development. 

 Of the two examples figured the brood chamber in U. carhjrcni is 

 11 mm. deep, while that of U. sulcata is about 6 mm., measured 

 from the invagination constriction. The walls are fairly stout, the 

 layers being of average strength except at the innermost parts 

 where the mesoglcea and endoderm appear somewhat thinned out. 

 The invaginated ectoderm is formed of tall columnar cells for the 

 greater part, but they become somewhat lower towards the bottom 

 of the chamber. 



A series of horizontal sections througli the brood chambers illus- 

 trates their relation at this stage to the adjacent organs of the body. 

 Two of these sections are represented in figs. 14 and 15, which are 

 taken through the chamber, the upper part of which is represented 

 in vertical section in Fig. 12. Here it is seen that the chamber 

 completely severs each mesentery into two parts, passing between 

 them on the outer side of the longitudinal muscle. It is interesting 

 to notice that although the invaginated endoderm investing the 

 brood chamber has already become continuous with the corresponding 

 layers of the divided mesenteries, the mesogloeal layers have 

 not yet .so joined, and on the inner side of the brood chamber the 

 me-soglcca of each mesentery is turned back upon itself, away 

 from the mesogkea investing the brood chamber (fig. 14), Fig. 15 

 represents a section of the horizontal series near the bottom of the 

 chamber, and shows the same mesenteries ; but each is now entire 



