the Salmon in Fresh Water. 15 



of several secreting tubes into a common duct occurs very frequently in 

 the trout, and I think that in the salmon it is still more frequent. 



2. Before the branching, of the glands begins, the characteristic upper 

 end (Oppel's, Oberende) of the cdl becomes less distinct and soon ceases 

 to be evident, and the whole of tlie cell body becomes granular and 

 undifferentiated. This I call the intermediate epithelium (Fig. 9). In 

 the glands of the cardiac or anterior part of the stomach this part of 

 the gland is very short, but in the pyloric region it occupies a large 

 part of the gland. 



3. The zyrnin-secreting epithelial cells do not exhibit any transi- 

 tion from the intermediate ones, but are sharply distinguished from 

 them (Fig. 9). They are cubical, have a large rounded nucleus poor 

 in chromatin, and a cell body with a very evident spongioplasm. The 

 .surface next the lumen of the gland is not differentiated in any way. 

 In the cell body are scattered numerous granules which stain with 

 eosin, though not very strongly, and which can be brought out best by 

 M. Heidenliain's iron-hsematoxylin, with which they stain an intense 

 black. I have 110 doubt, from the analogy of the granules in the 

 pancreas, that these are zymogen granules. They vary slightly in size, 

 and are always most numerous towards the free end of the cell. Their 

 actual number, or at least the ease with which they can be demon- 

 strated, varies very much in different stomachs, and even in different 

 parts of the same stomach, and probably would be found to depend upon 

 whether the organ had been recently called on to digest at the time of 

 death or not. 



The amount of this variety of epithelium present in the glands varies 

 greatly with the part of the stomach examined. About one-half or 

 one-third of the vertical extent of the cardiac glands is occupied by this 

 epithelium (Figs. 1, 4, 7), while at the pylorus there may be only a few 

 cells of this sort at the bottom of a tube formed in about equal propor- 

 tions of the superficial and intermediate epithelia. Perhaps one-twelfth 

 or less of the whole extent in the pyloric glands is occupied by this 

 epithelium. The glands intermediate in position between these two 

 regions are intermediate ako in this respect. 



The above description of the normal stomach is drawn up from 

 examination of the stomachs of seven fish caught at Berwick in March 

 1896. Portions of their stomachs were fixed in sublimate solution as 

 soon as the fish were killed. 



STOMACHS OF SALMON FROM UPPEK WATERS OF RIVERS. 



The stomachs of fish killed in the upper reaches of rivers present 

 very striking differences from the normal type, and these seem to be caused 

 by a desquamative catarrh of the mucous membrane. The muscular 

 and submucous layers show no change, the muscularis mucosse is 

 unaltered, but the connective tissue of the mucous membrane looks 

 swollen and hyaline. This appearance is best seen in the stratum 

 compactum, which is thicker than in the normal stomach, and often 

 much more folded, as though its total volume were increased. The 

 number of eosinophile leucocytes in the meshes of the connective tissue 

 is also usually increased. But the principal change is in the epithelium of 

 the glands. In the extreme cases this is almost entirely desquamated, 

 even from the fundi (Figs. 2 and 3); but usually more or less of the 

 zymin-secreting epithelium is preserved in situ, though in a degenerated 

 state. The superficial and intermediate epithelia are the first to dis- 

 appear, and usually cannot be recognised at all. The fundal epithelium 

 is never normal. The protoplasm of the cells loses its granular appear- 



