78 DISCOVERY REPORTS 



be to supply 'high tension' gas-nuclei around which the bubbles could form. However, due to 

 surface-tension, the gas-pressure within a bubble is inversely related to the diameter of the bubble, 

 which means that considerable pressures will be required for its growth. But in the foam found in 

 lung-alveoli, Pattle (1958) discovered that the surface-tension of the bubbles was much reduced by 

 a surrounding layer of insoluble protein. Perhaps the foam covering the gas-gland has somewhat 

 similar properties. 



In conclusion, it must be acknowledged that little is known of gas-secretion by the swimbladder, 

 particularly in deep-sea fishes. As Scholander (1956, p. 523) has written: ' In spite of all the information 

 available regarding the function of the gas-gland in fishes, one may safely say that none of the three 

 cardinal feats of the gland can as yet be explained: namely, the production of 100-200 atmospheres 

 of oxygen, of 10-20 atmospheres of nitrogen and of 0-1-0-2 atmospheres of argon.' In the endeavour 

 to solve these problems the giant cell gas-gland of Vinciguerria should provide excellent experimental 

 material, particularly for histochemical studies. 



The resorbent part of the swimbladder 



We have already seen that the swimbladder volume of a marine teleost must be kept near or equal to 

 some 5 per cent of the body volume. There is evidence that teleost fishes have a certain latitude of 

 movement above a level of neutral buoyancy (Scholander, Claff, Teng and Walters, 1951 ; Jones, 

 1952), but beyond this range the fish will tend to be 'ballooned' out of control as the volume of its 

 swimbladder (and hence its own volume) increases. Apart from the dangers of injury, an upward 

 migration involving a threefold increase in the volume of the swimbladder means that the fish must 

 exert (by compensatory movements) an upward force equal to about 10 per cent of its weight in air 

 in order to make 'downwards headway' (Denton and Marshall, 1958). 



During a migration towards the surface, a teleost with a closed swimbladder must reduce the 

 volume of the sac to a manageable, just buoyant, level by loss of the contained gases to the blood. This 

 diffusion takes place through special resorbent surfaces with a rich supply of capillaries, some account 

 of which has been given in the earlier descriptive section (pp. 7-50). These findings can now be 

 summarized and discussed. Other aspects will also be considered in a later section on vertical 

 migrations. 



Order Isospondyli, Suborder Stomiatoidea 



In this group there is a close association between the resorbent system and the gas-gland, for the 

 venous return from the capillaries is generally by way of veins running from the gland to the rete 

 mirabile. These veins may return from the gland {Astronesthes niger, Vinciguerria and Maurolicus), or 

 run along the inner edge of the gland {Pollichthys). Argyropelecus also has periglandular veins but these 

 may also pass through the gland. (With the material available it was not possible to make a close 

 study of these venous channels.) 



The arterial capillaries of the resorbent surface are formed from a branch of the retial artery which 

 leaves this vessel just before it flows into the rete, to run forward alongside the latter. On reaching 

 the resorbent area, it breaks up into arterioles and capillaries, which together with the venous elements, 

 form the resorbent complex. 



Since the venous blood from this complex eventually flows through the rete mirabile, a bipolar 

 retial system would appear to be essential. To supply the venous part of a resorbent area a unipolar 

 rete would have to give off hundreds of separate venous capillaries, which might then have to run for 

 considerable distances before meeting their arterial counterparts (see, for instance, Vinciguerria, 

 Maurolicus and Astronesthes niger). Such an arrangement might well be functionally unbalanced 



