17s University of Calif omia Publications. [zoology. 



outer ends. It is important to observe that the nuclei never get 

 quite to the extreme inner ends of the cells and that the smaller 

 inner portions of the cells remain distinctly protoplasmic; and 

 further that the nuclei do not become flattened out and crecentic 

 from pressure as so frequently happens in loaded gland cells, but 

 on the contrary that they remain spherical throughout. The 

 significance of this will be seen as we proceed. 



The glandular nature of the gastric epithelium does not 

 pertain in equal degree to the entire stomach wall but reaches 

 its most pronounced expression on the two sides of the organ. 

 Dorsally and ventrally the cells though usually having something 

 of the same character, have it in but a relatively slight degree. 

 That these cells are secretory and that the blastocoel mass is 

 the result of their activity scarcely admits of doubt. Naturally, 

 of course, one asks for the direct evidence on the point that 

 would be afforded by observing the discharge of the secretion. 

 While it is true that in a majority of the specimens examined 

 there is an absence of such evidence, in a few instances we have, 

 with little doubt, found it. Such evidence is furnished by the 

 section shown in PI. XIX, Fig. 13. 



Here one finds a thin homogeneous layer on the outer surface 

 of the gastric epithelium, from which there extend off into the 

 adjacent coelomic mass more or less w-ell defined tracts or bands 

 in which the substance shows a fibrillar structure. These fibrils 

 are almost certainly continuous with the homogeneous layer on 

 the one hand, and with the coelomic mass on the other. They 

 are the result of post mortem changes of the secretion in all 

 probability but their relations show, according to our interpret- 

 ation, the origin and destination of this secretion. In this 

 particular specimen as in one or two others presenting a similar 

 condition, the secretion is considerably more deeply stained than 

 we have ever found it in specimens in which there is an absence 

 of indications that the secretory process is going on. 



The gastric cells appear to perform this secretory function 

 up to the time of the initial changes toward metamorphosis; that 

 is, to the first intimation of gill pockets. At least in all the 

 stages studied precedent to the beginning of the pouches, we find 

 the cells in the secretory regions either loaded with the secretion 



