GAS GLANDS OF SOME TKLEOSTEAN FISUES. 229 



rapidly proLlu^es oxygen in considerable quantity in order to 

 counteract the compression of the bladder due to the increased 

 pressure and so to maintain the equality of the specific gravity of 

 the fish with its medium ; if the hsh rises, either the absorption- 

 organ known as the " oval " comes into action and returns the 

 surplus oxygen in the swollen bladder to the blood, or, if the fish 

 be a physostome, the pneumatic duct allows the surplus gas to 

 escape to the exterior. Between the two extreme classes of 

 bladder — the oxygen-filled bladder with " red bodies " and the 

 nitrogen-filled bladder devoid of "red bodies" — there exist many 

 transitional kinds containing a relatively small percentage of 

 oxygen and feebly- developed "red bodies"*. As regards the 

 exact mode of abstraction by the gas gland of the gases contained 

 in the blood and their subsequent expulsion into the bladder, 

 papers already published afford little or no information on the 

 subject. In the case of those gases present in the bladder in 

 minute quantities, it is probable that they leave the blood by 

 simple diffusion: e.g. Traube-Mengai-ini (73, 74) and Priefer (60) 

 showed that this was the case when hydrogen was dissolved in the 

 water t. Ifc is evident, therefore, that the problem of gas- 

 abstraction becomes of importance only when the partial pressure 

 of any particular gas in the bladder exceeds that in the blood. 

 That the gas gland cells act as a pump is certain J ; it is also 

 certain that the gases first enter the cells in a dissolved condition 

 and that, just as a scleroblast converts dissolved calcareous salts 

 into a solid spicule,lso the gas gland cells cause these dissolved 

 gases to appear in a gaseous form as bubbles in the cytoplasm §. 

 In addition to this, however, the gland cells subject these gas 

 bubbles to a considerable pressure, greater than that existing in 

 the bladder, so that when the gas bubbles are expelled into the 

 bladder lumen and experience a diminution of pressure they 

 burst. As regards the varying composition of the bladder gas, it 

 can only be concluded that the cells of the gas gland, like the cells 

 of the kidney, exercise a selective power. That the oxygen 

 pumped into the bladder is derived from the blood no one doubts, 

 but Nusbaum & Reis, Thilo, and some others find a great 

 difficulty in supposing that nitrogen (and carbon dioxide) is 



* E.g. Fercafluviatilis contains on an average 15 percent, of oxygen, 83 of nitrogen, 

 and 2 of carbon dioxide (Hiifner) ; Lota vulgaris 65 per cent, oxygen, 30 nitrogen, 

 5 carbon dioxide (Hiifner), &c., &c. For analyses of the bladder gases in many 

 iish see the works of Biot (19), Contigliachi (27), Delaroche (30), Humboldt & 

 Provencal (39,40), Hiifner (38), Richard (67) among others. The percentage 

 composition of the bladder gas exhibits, as might be expected, considerable variation 

 not only in different hidividuals of the same species of iish but in the same individual 

 at different times. 



t It must be mentioned, however, that Humboldt & Proven9al (39) in 1809 per- 

 formed this experiment of impregnating water with hydrogen and failed to detect a 

 trace of hydrogen in the bladder, but probably their methods of analysis were too 

 ciude. 



X In this connection Hiifner ventures to compare the gas gland cells with the 

 contractile gland cells described by Drasch (32) in the neck-skin of the Frog, both 

 being similarly supplied with nerves. 



§ Of. the gas vacuoles in Arcella and other Thecanioehida. 



Piioc. ZooL. Soc— 1911, No. XYI. 16 



