GAS GLAXDS OF SOME TELEOSTEAX FTSllES. 



187 



by tlie division of the initial artery and vein. The subdivision 

 of the arteries and veins in each mass proceeds rapidly, until 

 there is forined by this process ari ovoid mass of extremely fine 

 parallel arterial and venous capillaries, closely intermingled Avith 

 each other, but never intercommunicating and indistinguishable 

 structurally as arteries and veins under the highest powers of 

 the microscope (fig. 4). These two ovoid masses of parallel 

 arterial and venous capillaries lying on the dorsal and ventral 

 sides of the ductus pneumaticus just a.nterior to its junction 

 with the bladder are the retia mirabilia of the eel-bladder (text- 

 fig. 53, R.M.). 



The finest arterial and venous capillaries of each rete mirabile 

 run parallel to each other for a short distance, but soon they 

 commence to unite, arterial capillaries with arterial capillaries 

 and venous with venous, in order to re-form larger vessels similar 

 to those which pi-oduced the capillaries by subdivision. This 

 process of reunion proceeds so far as to form two masses of 

 lai'ge intermingled arteries and veins similar to those depicted 



Text-fig. 54. 



External aspect of a rete mirabile (after Quekett). The anastomoses of the 

 vessels shown in this figure do not occur in actiiality. 



in figure 3 (c/. figs. 3 & 5). When reunion has proceeded thus 

 far, each of these masses of re-foi-med arteries and veins once more 

 breaks up into fine capillaries in connection with the vascular 

 supply of the glandular folded epithelium which lines the bladder 

 internally. I reproduce here as faithfully as I am able (text- 

 fig. 54) Quekett's figure (61) of the construction of the Eel's rete 

 mirabile, which illustrates in a general way that which I have 

 just described. Text-fig. 54 is vm satisfactory in that it does not 

 indicate the fact that the vessels and capillaries forming the 

 rete are of two kinds — arterial and venous — closely intermingled 

 with each other, although never intercommunicating ; also anas- 

 tomoses never occur between adjacent large vessels in the manner 

 shown in the figure. A more satisfactory diagram of the 

 construction of the Eel's rete mii'abile is the one I have 



