72 DISCOVERY REPORTS 



the equilibrium pressure is directly proportional to the diffusion across the retial capillaries and the 

 amount of oxygen dissociated from a unit volume of blood. The inverse terms in the equation are the 

 blood solubility coefficient for oxygen and the rate of blood-flow in the retial capillaries. Again, the 

 longer the retial capillaries the greater the diffusion and the less the blood-flow, both of which will 

 enhance the build-up of equilibrium pressures. Theoretically, a small difference in gas-tension in the 

 arterial and venous capillaries at the beginning of the rete could lead to pressures of several thousand 

 atmospheres at the end near the gas-gland. However, active secretion of oxygen by the gas-gland 

 accords better with present knowledge than a liberation of this gas from the blood (Scholander, 1954; 

 Sundnes, Enns and Scholander, 1958). On the other hand, the gas-gland cells might not accept 

 oxygen from the blood unless the tension was higher than that in the swimbladder. This aspect must 

 be left for future experiments : here we may simply remark on the significance of long capillaries for 

 the efficient exchange of gases in the retia, particularly in those of deep-sea fishes. Regardless of the 

 attainment of high pressures, a deep-sea fish must have well-developed retia in order to prevent the 

 loss of gases from the swimbladder. Using his equations and reasonable values for the various con- 

 stants, Scholander (1958, p. 9) calculated ' . . . that the exchange through the rete in a deep-sea eel is so 

 great that if the blood in swimbladder has an oxygen tension of two hundred atmospheres it will leave 

 the bladder with an oxygen tension only a few millimetres higher than in the arterial blood '. 



In proportion to the size of the swimbladder, bathypelagic teleosts have large retia mirabilia, the 

 relative development of these systems being greater than those of shallow water species (Marshall, 

 19150, 1954; Jones and Marshall, 1953). This difference may be given rough quantitative expression 

 by obtaining values of the ratio / x bjr for species from the two types of environment (/ and b being the 

 lengths of the major and minor axes of the sac (which approaches an ellipsoid in form) and r the 

 length of the retial capillaries). 



Values of this ratio are given in Table 4 (p. 74) for various species of deep- and shallow-water 

 fishes. Drawings of the swimbladders of some of the latter species may be found in Text-fig. 36. 



The first six species in the shallow-sea group are epipelagic fishes. It will be seen that the ratios 

 in these species are far higher than the figures obtained for the deep-sea species. But as regards a 

 counter-current exchange-system it is the absolute rather than the relative length of the retial 

 capillaries that is significant. However, apart from Pollichthys, Vinciguerria, Argyropelecus and Astro- 

 nesthes among the deep-sea group and Hyporhamphns from the epipelagic species, the retia of the 

 bathypelagic species are longer (some much longer) than those in the surface-swimming species. And 

 there is another factor to be considered, the diameter of the capillaries, for the smaller this is the 

 greater the efficiency of gaseous exchange. Two of the exceptions among the deep-sea group have 

 relatively small capillaries, which measured 7-8 ft in diameter in Argyropelecus and Vinciguerria. (There 

 are no data for the other two.) In the epipelagic species the capillaries are 10// or more in diameter. 



Table 4 also reveals that the ratios and retial dimensions of Gadus minutus and Capros aper are close 

 to those of the bathypelagic species : this is not surprising in view of the depth-range of these two fishes. 



Considering only the deep-sea fishes, it would be reasonable to expect that the deeper the living 

 space the longer would be the retia. In Table 4 it will be seen that the retia of Stephanoberyx monae 

 and Melamphaes megalops are much longer than those of the other deep-sea fishes. Now the first 

 species may well be demersal rather than pelagic in habit and it has a depth-range extending down to 

 2295 m. (Grey, 1956). Norman's (1929, 1930) data for Melamphaes megalops suggests that this fish 

 tends to be concentrated well below the 500-m. level. The populations of the other species tend to be 

 centred above this depth. Even more striking instances of this correlation between the retial span 

 and depth can be found if abyssal fishes are also considered. But this will best be left until the final 

 section of this report. 



