92 DISCOVERY REPORTS 



If the activities of these fishes are largely confined to night time, when they are swimming in the 

 well-oxygenated surface waters, it is difficult to see how enough oxygen can be obtained to fill their 

 swimbladders. Use of the gas-gland as an oxygen store would seem to be essential for such fishes. 

 The problem (a most intriguing one), remains. 



Gas resorption 



During its daily climb towards the surface, a deep-sea fish must lose gas from its swimbladder as the 

 hydrostatic pressure is reduced. While the physical problem is simply the converse of that faced 

 during descent, it is a more critical one. Being overweight is presumably no handicap to a diving 

 fish, but during the rise the swimbladder must be kept from exceeding the volume necessary for 

 neutral buoyancy. If this is not done, the fish will lose control and start to ' balloon ' upwards to the 

 surface at an ever-increasing rate. Even predatory fishes without a swimbladder are not entirely free 

 from this problem. Gunther (1887) relates how Johnson found a gulper-eel (Saccopharynx ampul- 

 laceus) floating at the surface. The fish had swallowed a nine-inch, deep-sea gadoid (Halargyreus) 

 ' . . . the stomach of which was forced up into the mouth by the distended air bladder, showing how 

 rapidly both fishes must have ascended to the surface '. 



As we have seen, most of the migrating fishes with swimbladders weigh between 0-5 and 10 g. 

 Considering once more a 5-g. fish with a swimbladder volume of 0-25 ml., a migration from a depth 

 of 400-50 m. will cause the sac to expand to about 2-0 ml. if no gas is lost. So during the ascent, about 

 1-75 ml. gas, the greater part of which will be oxygen, must disappear into the blood. 



Such problems have been considered by Jones (1951, 1952) and are based on experiments with 

 perch (Perca fluviatilis), which, like the deep-sea fishes with swimbladders is a physoclist. Jones (1952) 

 has calculated that a migration from 500 to 100 m. by a physoclist should extend over 34 hr. if the 

 swimbladder is to be kept at the volume of neutral buoyancy. He suggests the difficulty could be 

 overcome if the fish had a small swimbladder (the extra space being filled with fat), or thick, pressure- 

 resisting walls bounding the sac, or if the rate of resorption is 20-30 times that supposed in the perch. 



Considering these suggestions, the deep-sea fishes have a reduced swimbladder (compared to the 

 perch) in that they live in a denser medium and so need less buoyant support. In freshwater fishes, 

 the volume of the swimbladder is 8 per cent of the body volume (see Jones and Marshall, 1953). 

 Moreover, the walls of their swimbladders are quite thin (see pp. 60-65). The rate of removal of gas 

 from the sac must thus be high. 



The quantity of gas diffusing into a capillary bed will directly depend on: (1) the area of the bed; 

 (2) the rate of flow of blood; (3) the concentration gradient (the difference in tension between the 

 swimbladder gases and those in the blood; (4) temperature. The rate of diffusion will also be inversely 

 related to a ' frictional ' component, the thickness and nature of the tissue through which the gases 

 move, and on the size of the gas molecules. 1 



The Lilliputian fauna of bathypelagic fishes have decidedly large expanses of capillaries relative to 

 the volumes of their swimbladders (see the descriptive section, pp. 7-50). In Table 6 this is expressed 

 as a ratio of surface of resorbent area : volume of swimbladder. The swimbladder volume is taken to 

 be 5 per cent of the body volume. The surface-area of an ' oval ' is readily calculated (but is a minimal 

 value owing to some collapse of the swimbladder) since it is circular. In the stomiatoid fishes the 

 area of the capillary bed was got by dividing it into convenient sub-areas. This will also be a minimal 

 value. 



1 Fick's law for diffusion processes is expressed thus: y- = — ^> a -=- , where dn is the amount of substance diffusing 



across an area a, in time dt, dcjdx is the concentration gradient, R is the gas constant, T the absolute temperature, / the 

 'frictional resistance' and N the Avogadro number. 



