200 DR. W N F. WOODLAND OK THE 



pancreas into the bladclei^ wall cannot be supposed to be of any- 

 more significance than its penetration into the liver. I failed at 

 first to recognize this gland as the pancreas, both because of its 

 peculiar position inside the bladder wall and because of the 

 peculiar modification of the pancreatic acini in juxtaposition with 

 the veins, neither of which peculiarities have been described by 

 writers on the teleost pancreas. Another small peculiarity existing 

 in several teleost genera (Gobius niger, e. g.) which I studied more 

 than others was the syncytial character of the wall of the pancreatic 

 duct (PI. V.fig. 33 and Pi". IV. fig. 34), a feature not present in most 

 teleosts. The modification of the pancreatic acini in connection ' 

 with the veins (veins from the rete mirabile as well as hepa.tic- 

 portal veins from the gut wall) just referred to consists of the cells - 

 of each acinus next the thin vein wall being drawn out, the cells 

 on the side remote from the- vein being quite short, as shown in 

 text-fig. 56. This figure also indicates the similarity between the 

 zymogen granules and the erythrocyte globules, both globules and 

 zymogen granules being stained a bright emerald-green in my 

 preparations. As stated in Part II., I now have good reason to 

 believe that the erythrocyte granular matter is abstracted from 

 the blood in part by the liver and in part by special cell- 

 masses situated near the kidneys. Laguesse (4-10), K-ennie 

 (17), and others have shown that the teleost pancreas resembles in 

 all essentials the pancreas of higher Vertebrates — in the characters 

 of the acini, ducts, zymogen granules, and the presence of centro- 

 acinal cells and areas of Langerha,ns. With respect to these last, 

 I may mention incidentally that I can fully confirm the results of 

 Rennie (17) and there can be but little doubt now that these 

 areas constitute an organ physiologically and in many teleosts 

 anatomically distinct from the pancreatic acini. This view of 

 the physiological independence of the islets of Langerhans is sup- 

 ported by the researches of Diamare (3), Flint (3 a), Opie (14, 15), 

 de Witt (3), Lane (11) and others in opposition to the view that the 

 islets are patches of exhausted pancreas, supported by Lewaschevv 

 (12), Laguesse (6, 10), and Dale (1); and, as I have already re- 

 marked, researches on the teleost pancreas indicate quite plainly 

 that the former is the correct view. 



SiPHONOSTOMA TYPHLE (rondeletii Delar.), Nerophis 



^QUORIUS. 



The " red bodies " of these two genera are, like that of S'i/7i- 

 gnathus, small red patches situated at the extreme anterior end of 

 the bladder and identical in their plan of construction. The only 

 particular in which the " red body " of SipJiono stoma differs from 

 that of Syngnathus is that in the fonner the cells of the gas gland 

 are much larger and the intervening ducts much narrower ; 

 indeed, in the posterior portion of the gland the lumina are so 

 narrow as not to be seen ver}'' easily. The large capillaries are 

 also more circular in transverse section in Siphonostoma than in 

 Syngnatlnts. Gas bubbles and intracellular capillaries are in my 

 preparations to be found occasionally in the substance of the 



