THE SWIMBLADDER AS A HYDROSTATIC ORGAN 69 



depth (Parr, 1937; Kanwisher and Ebeling, 1957), but in absolute terms this amount may be quite 

 small (see p. 84). However, many of the bathypelagic fishes with swimbladders undertake daily 

 vertical migrations. These will be considered in a later section (pp. 85-95). Here the main concern is 

 with the structures involved in maintaining the swimbladder as a hydrostatic organ. These are the 

 gas-secreting complex (rete mirabile and gas-gland) and those allowing of the loss of gases from the 

 swimbladder (the resorbent capillary complex). 



The gas-producing complex 

 The swimbladder produces gas by means of a capillary system supplying a glandular area. The gland is 

 formed by localized modification of the epithelial cells that line the sac. 



In physoclists and some physostomes (e.g. cyprinids, pike and eels) the capillaries form retia 

 mirabilia. These consist of regular and intimate intercalations of arterial and venous capillaries, which 

 follow parallel courses and carry blood to and from the gas-gland. 1 



The retia mirabilia 

 Some account of the form and size of the retial system has already been given in the descriptive part 

 of this report (pp. 7-50). In this section the emphasis will be on the fine structure. 



Woodland (19110, b) was the first to appreciate the essential features of retia mirabilia. He 

 divided them into two types, unipolar and bipolar. In both kinds, an artery and vein subdivide to 

 form the close and regular association of capillaries, (which may number many thousands), that supply 

 the gas-gland. Before entering the gland, the retial capillaries of the bipolar type recombine to 

 form arteries and veins, these then breaking up within the gland to form the capillary circulation. 

 In the unipolar type the retial capillaries merely continue into the gland. 



Both kinds of retia are found in the swimbladders of deep-sea fishes. All stomiatoids have a single, 

 bipolar rete mirabile : the myctophids have three unipolar retia. Now some species of both groups 

 live under similar hydrostatic pressures and undertake daily vertical migrations, and on this account, 

 there would appear to be ' nothing to choose ' between the efficiency of both types of system. 



Apart from this consideration, it is interesting that the two most diverse groups of pelagic, deep-sea 

 fishes, the stomiatoids and myctophids, differ in this particular way. However, it will be as well to 

 remember Pantin's (1951) observation: that given certain standard parts, the number of structural 

 solutions to a physiological requirement is limited by the nature and number of these parts. In this 

 instance, given an artery and a vein that form a retial system, there would appear to be only two ways 

 this system could feed a gas gland. The retial capillaries can either continue into the gas-gland (uni- 

 polar retia) or combine to form larger vessels (bipolar retia) before doing so. 



The nature of the swimbladder as a whole must also be considered. Woodland (191 1 a) and Fange 

 (1953) have shown that the retia mirabilia of teleosts with a euphysoclistous swimbladder are unipolar 

 in type. This survey has provided further evidence for this generalization. Besides the Myctophidae, 

 the other bathypelagic euphysoclists are the Miripinnati, Anoplogastroidea, Stvlophorus and Chias- 

 modon. All these fishes have unipolar retia. At least some of the deep-sea salmonoids are euphyso- 

 clists and their micro-retia run straight to the gas-gland. This is hardly surprising. 



Apart from the stomiatoids, the only other fishes known to have bipolar retia are the eels (Apodes). 

 The development of a bipolar retia system in the first group would appear to be linked to the venous 

 part of the resorbent area, which drains into vessels that also supply the gas-gland. Such a circulation 

 would hardly be feasible with the unipolar type of rete (see also p. 78). However, in both euphyso- 

 clists and eels, the secretory and resorbent parts have their own circulatory systems (Fange, 1953). 



1 In view of this association in parallel, the noun rete is quite inappropriate, but the adjective is justified, for these capillary 

 systems are wonderful instances of biological engineering in miniature. A better name would be fastis mirabilis. 



