THE SWIMBLADDER AS A HYDROSTATIC ORGAN 71 



general terms, one may state the problem thus: How can a steep concentration gradient be maintained 

 across a barrier in spite of the fact that liquid is continuously streaming through it? The answer lies in the 

 arrangement of the vascular channels. 



Clearly, the retia will be most efficient in gaseous exchange if the capillary elements fit together so 

 as to obtain the greatest possible surface of contact. Scholander (1954, 1958) had some interesting 

 observations on this aspect. Transverse sections through the retia of a deep-sea eel (Synaphobranchus), 

 a rose-fish (Sebastes) and a rat-tail (Coryphaenoides) showed that the arterial and venous capillaries 

 fit together to form either a chequer-board 1 or a hexagonal star pattern. The eel has the first arrange- 

 ment, one giving the maximum gaseous diffusion between the afferent and efferent capillaries, and 

 Scholander (1954) went on to say '. . .it is remarkable that we find it in our deepest fish. The only 

 other solution to the topological problem of making four polygons (black or white) meet at one point 

 in such a way that black always borders white is realized in the hexagonal star pattern found in the 

 rete of the rosefish ' (and in that of the rat-tail). 



While looking through serial transverse sections of the swimbladder of various stomiatoids, 

 Argyropelecus aculeatus, Vinciguerria nimbaria and Polyipnus, I found both types of pattern in the 

 same rete. In Polyipnus, for instance, there is a transition from the hexagonal to the mosaic arrange- 

 ment in passing from the proximal to the distal end of the rete (see PL I, fig. 4). At the beginning 

 of the rete the arterial capillaries are occluded and form an hexagonal pattern round the larger 

 (6-8//) venous capillaries. About half-way down the rete, the arrangement is much the same, except 

 that the arterial capillaries are partly open. Lastly, over the distal third of the rete, the two sets of 

 capillaries form a mosaic pattern and are equal in size (6-7/1). In the specimen of Argyropelecus 

 aculeatus, except for a middle area at the proximal end, the rete showed a hexagonal star pattern 

 throughout. 



In a specimen of Vinciguerria attenuata, however, the rete of which was well expanded, the 

 capillaries formed a mosaic pattern at all levels (see Text-fig. 32a). The seemingly curious mixture 

 found in other stomiatoid retia is likely to be no more than a reflection of the unique arrangement 

 whereby venous blood from the resorbent capillary bed returns through the rete (p. 79). During the 

 secretory phase the rete will be fully expanded and both sets of capillaries will then form a mosaic 

 pattern. When gases are being lost from the swimbladder the arterial capillaries must be closed, while 

 the venous elements will be fully expanded. In this way a hexagonal pattern would be formed. The 

 mixture of patterns found in some retia would thus be due to the physiological state of the fish when 

 it died in the net. Preservation and fixation might also play some part, but evidently not in the rete of 

 Vinciguerria attenuata mentioned above. 



In the Myctophidae, the vascular system of the oval does not involve the retia, other than that the 

 arterial blood may come from a branch of the retial artery. In two species, Myctophum punctatum 

 and Diaphus dofleini, transverse sections through the retia showed the capillaries to be rounded rather 

 than polygonal, the appearance being more like a system of condenser tubes than a mosaic. In the 

 living fish, their shape may be otherwise. 



Apart from the intimacy of their association, the total surface of contact between the capillaries 

 will obviously be directly proportional to their length. And an increase in length will not only lead to 

 increased exchange of gases, but also slow down the rate of blood-flow and so further enhance the 

 efficiency of the exchange. But before considering this aspect in bathypelagic fishes, some mention 

 must be made of Scholander's (1954, 1958) concept of the rete as a device for the building up of 

 high pressures. 



Assuming that gases are liberated from the blood, Scholander derived an equation, showing that 



1 'Mosaic' would be a more apt descriptive term. 



