THE SWIMBLADDER AS A HYDROSTATIC ORGAN 79 



with the arterial system, particularly in fishes that make upward vertical migrations. To keep the 

 swimbladder from over-inflation the rate of gas-removal should be high : hence the need to remove 

 venous blood from the capillary system as quickly as possible (see also pp. 92-94). 



The resorbent circulation of the stomiatoid swimbladder would appear to be unique among deep- 

 sea fishes (perhaps also among teleosts in general). It presumably functions as follows: During gas- 

 resorption, contraction of the arterioles forming the retial capillaries will close down the flow of 

 arterial blood through the rete (Fange (1953) has shown how such contraction can lead to capillary 

 closure). As the retial artery is still open blood can only flow down the arterial branch supplying the 

 resorbent area. Blood returning from the system will be free to flow through the rete, since the 

 venous capillaries will be fully open (presumably they have extra room for expansion owing to the 

 virtual closing of their arterial counterparts). During gas-secretion both sets of retial capillaries will 

 be open and the gas-gland fully expanded. Arterial blood will now flow from the retial artery into 

 the dilated arterioles (and so through the rete) rather than down the by-pass vessel into the resorbent 

 system. Furthermore, when the gas-gland is expanded, the resorbent system is contracted (Fange, 

 1953) and this may well lead to a stoppage of the resorbent circulation. 



Just how this antagonistic action of the secretory and resorbent mucosa is brought about in the 

 stomiatoids has yet to be determined. In euphysoclist teleosts, the action is perfectly clear (see 

 p. 81 of this section). However, the smooth muscles of the swimbladder wall are likely to be 

 involved. (In Argyropelecus and Polyipnus the muscle layer is close to the inner epithelium.) The 

 disposition of the fibres may be such that those controlling the expansion of the capillary network are 

 contracted during resorption, while those around the gas-gland are relaxed. The opposite would occur 

 during gas-secretion. An extreme instance of this antagonism may be seen in Text-fig. 32 of the 

 swimbladder of Vincignerria. 



Dissections made of two specimens of Argyropelecus aculeahis show the swimbladder as it probably 

 appears during the resorptive and secretory phases. Considering only the former condition, the 

 relatively large expanse of the capillary layer and the bunching up of the gas-gland is particularly 

 striking (Text-fig. 38). 



As already indicated, the capillary circulation of stomiatoids is unique in that the venous return- 

 flow is through the rete mirabile. In the groups now to be considered the secretory part of the swim- 

 bladder is perfectly distinct from the resorbent part, this having a separate venous system (which 

 usually runs into the hepatic or cardinal veins). Teleosts with such a swimbladder are called euphyso- 

 clists and are to be distinguished from paraphysoclists, those in which the two parts are not so sharply 

 delimited (Rauther, 1922). 



Salmonoidea (deep-sea). In a recent revision of the argentinine fishes (genera Argentina and 

 Glossanodon), Cohen (1958) has shown that the swimbladder has two distinct parts, comprising an 

 anterior, thick-walled, cylindrical chamber and a posterior, thin-walled diverticulum. From Fange's 

 (1958) observations and my own, it is clear that the anterior chamber contains the gas-gland and the 

 micro-retia. Cohen described the posterior diverticulum as having a vessel from the distal end that 

 seemed to drain into the renal portal system (or is this the dorsal aorta?), while a vein from the front 

 of the down-turned part leads to the hepatic portal vein. In a specimen of Argentina sphyraena 

 (standard length 170 mm.), I found a well-marked posterior diverticulum in which the down-turned 

 part, that runs forward beneath the thick-walled chamber, measured 13 mm. in length. The inner 

 epithelium of this part was thrown into folds and contained a capillary network. There is thus good 

 indication that the diverticulum is the resorbent part of the swimbladder. I also found a thin-walled 

 posterior chamber in Microstoma microstoma, but this had no down-turned antrorse section. Nansenia 

 may also have such a structure (the individual I examined was not well-preserved). 



