GAS GLANDS OF SOME TELEOSTEAN FISUES. 215 



Coggi 1886 (25), Serranits, Perca, Ballsfes, Corvina, Julis, Motella 

 and other types, ami 1889 (23) Esox lucius ; Corning 1888 (28), 

 Esox, Perca and other types; Vincent &, Barnes 1896 (75), 

 Zeus, Gaelics and other types; de 8eabra 1897 (31), Esox and others ; 

 Jaeger 1903 (44), Sci(ena, Lacioperea; Deineka 1904 (29), Perca, 

 Esox and other types; Bykowski & Nusbaum 1904 (24), 

 Fierasfer, Ophidium; Reis '& Nusbaum 1905, 1906 (62, 63), 

 Poll/acanthus {Macropodus), Flerasfeyr, Ophidium, Gharax, Perca-, 

 Reis 1906(64, 65), Sarcjus, Gharax, Pagellus, Gorvina, Umhrina, 

 Chri/sophr>/s and others described in a general way. 



Summary of Part I. and General Gonclusions resjjecting 

 ^ Gas Gland Structure. 



From the foregoing it is now manifest that all teleostean " led 

 boilies," *". e. the vasculo-glandular gas-producing organs situated 

 in the bladder wall, are composite structures (consisting of two 

 distinct and sometimes separate parts — the glandular epithelium 

 or gas gland* and the rete mirabilef. The mutual relationships 

 of these two parts are quite constant, and though the gas gland 

 must be described as the essential part, yet the rete, on account 

 of its constant J presence and undoubtedly highly important 

 function, is probably of almost equal importance. The rete 

 mirabile bipolare geminum (Miiller, 53), i. e. the arterial and 

 venous rete already described, is always formed in connection 

 with the artery and vein supplying the gas gland, and may either 

 be anatomically quite distinct from the glandular epithelium (Eel 

 type), the arterial and venous capillaries reuniting to a varying 

 extent on the gland side of the rete to form large arteries and 

 veins before supplying the glandvilar epithelium, or the rete 

 mirabile may be contiguous with the glandular epithelium, the 

 capillaries of the rete undergoing little (e. g. Si/7ignatkus) or no 

 (e, g. Ophidium) reunion before supplying this. 



The diffuse pancreas, when present, is often intermingled with 

 the arteries and veins situated at the extreme anterior end of the 

 rete, and so actuall}^ becomes situated inside the bladder wall ; it 

 was at first, on account of several j)eculiarities, mistaken by me 

 for a new gland with a special function coimected with the gas 

 gland, but this view was quite a mistaken one, the penetration 

 of the pancreas into the bladder wall being of no physiological 

 significance. 



The gross anatomical difterences, often visible to the naked 

 eye, which are to be found in " red bodies " taken from different 

 fishes, are determined by the disposition and degree of subdivision 

 of the rete mirabile — since the distribution of the glandular 



* The " driisige siiuine " of Miiller .and the " corpo epitheli.ile " of Coggi. 



t The " vaso-gaiiglioH " of Owen and " organo vascolare " of Emery. 



X That is to sa}', using the term " gas gland " in the usual sense as equivalent to 

 "oxygen gland" (see Part II.), the rete is constantly associated with it; the rete, 

 however, is absent in the cases of those bladders the lining-epithelia of which only, 

 or principallv, " secrete " nitrogen and carbon dioxide. 



15* 



