132 BULLETIN OF THE BUREAU OF FISHERIES. 



My histological observations, which will be given in detail in a later paper, show that 

 the pancreatic cells are not all compactly arranged in acini, as is usually the case, but 

 that they are more or less scattered. Every cross section of the pancreas reveals the fact 

 that a large amount of adipose tissue is present in association with the pancreatic tissue. 

 The cross section of a pancreatic lobe is somewhat triangular in shape. The central 

 portion of the triangle in an adult organ always contains one or two blood vessels of 

 relatively large size, together with a considerable amount of fatty tissue. The pan- 

 creatic cells are arranged around the surface of the gland and in the interstices among 

 the fat cells and blood vessels. There are many acini where definite arrangement of 

 pancreatic cells around a central lumen can be shown. This lumen rarely forms a space 

 as large as one sees in the corresponding region of a mammalian pancreas. In these 

 groups, in favorable preparations, that portion of the pancreatic cell bordering on the 

 lumen is highly granular and the granules stain in a way characteristic of zymogen 

 granules. The pancreatic cells around the surface of the gland and in many portions of 

 the deeper region have a rather scattered arrangement in which it is extremely difficult 

 to show definite relation to ductules. Such relation is questionable in a large proportion 

 of the gland. The diminutive size of the pancreatic ducts, the presence of the large 

 amount of fat in the gland, the rich vascular arrangements of the gland, and the diffuse 

 arrangement of the gland cells of the salmon pancreas have led me to the conclusion that 

 the secretion of this gland is largely internal, i. e., that the pancreatic secretion is largely 

 thrown into the lymph and blood stream of the organ. In this statement it is not meant 

 to minimize the importance of the gland in the production of the pancreatic juice. 

 Rather, the intention is, to emphasize the importance of the internal lipase secreting 

 function. 



When digestion stops and the animal begins its fast, the pancreatic gland undoubt- 

 edly continues to function. This is shown especially by the histological appearance at 

 different stations of the migration journey. There is obviously no digestive function 

 to be accomplished by the secretion during the fast. Therefore, during this phase of the 

 life cycle the internal secretive function becomes the main, in fact, the only one. The 

 vascularity of the pancreas and of the pyloric cseca does not change in proportion to 

 the amount of retrogressive change shown in the caeca and in the rest of the alimentary 

 tract. Putting these facts and deductions together they may be summarized thus: 



1. The pancreatic gland is abundantly active during the migration fast. 



2. The activity consists in the production of an internal secretion which is not 

 essentialh' different in character from the normal secretion produced at an earlier stage 

 in the life cycle. It is therefore rich in lipases. 



3. The pancreatic lipase produced during the fasting period is chiefly, if not wholly, 

 discharged into the tissue spaces, from whence it reaches the blood stream. 



4. The circulation carries the lipase to the tissues of the body, which includes both 

 the fat-storing tissues and the active fat-using muscles now under especial consideration. 



Lipase from the granule cell layer. — There is a second tissue filled with zymogen 

 granules which I believe to be an active source of lipase, namely, the granule cell layer 

 of the alimentary tract. I have already called attention to the presence of a layer of 

 granule cells in the mucous coat of the alimentary tract. This layer is especially richly 

 developed in the pyloric caeca. In every section of the retrogressing caecum, one is 

 strongly impressed with the continued large size of this granule cell layer. Even in 



