230 DR. W. N. F. WOODLAND ON THE 



similarly derived. It is true that the amount of nitrogen 

 dissolved in the blood is very minute as compared with the 

 amount of oxygen, but, on the other hand, nitrogen does not, like 

 oxygen, require to be produced at a rapid rate ; and since the 

 supposition of Nusbaum & Reis, that free nitrogen is generated 

 by the decomposition of cell-substance, is quite inadmissible for 

 the reasons supplied above, the nitrogen dissolved in the blood 

 can be the only source of that gas. It is well known that human 

 blood, when suddenly released from gTeat pressure, develops 

 bubbles of nitrogen owing to the inability of the blood to re- 

 dissolve the gas immediately, and doubtless the cells of the gas 

 gland also so act upon (though certainly not by a diminvition of 

 pressure) the blood plasma bathing their substance as to comjoel 

 the nitrogen to assiime a gaseous form. Once abstracted from, 

 the blood stream, the nitrogen, in the case of those bladders 

 containing nitrogen at high pressure, is pumped in the same 

 way as, and in many cases with, the oxygen into the bladder 

 lumen *. 



I now propose to oflfer some additional suggestions concerning 

 the physiology of the gas gland which have occviri'ed to me 

 during the course of my work. It is curious that not one of the 

 investigators mentioned in the foregoing pages has attempted to 

 explain in a satisfactory manner the striking conformation of the 

 rete mirabile. A year ago I published in a short note (78) 

 (based, as I have already explained, upon a mistaken conception 

 of the teleost pancreas) a sketch of a new theory concerning 

 the use of the rete which I have again outlined on a preceding 

 page of this paper, but a,part from this I have met with no 

 suggestion concerning the phj^siological significance of the rete 

 mirabile bipolare geminum (or more simply rete mirabile duplex) t. 

 I now propose to restate more fully this theory of mine, but 

 before doing so I will mention the only previous hypothesis of 

 which I am aware. Johannes Miiller (53) stated that in his 

 opinion the retia mirabilia associated with the gas glands of 

 teleost fishes (and Miiller was one of the first, if not the first, to 

 distinguish the gas glands — " luftdriisen" — from the retia mira- 

 bilia and to state their proper function as gas-producers) possessed 

 the same utility as the various other kinds of retia mirabilia de- 

 scribed by him, viz. to cause the blood stream to flow more slowly 

 for some physiological purpose, and this opinion has been adopted, 

 so far as I know, by all subsequent observers. Presumably 

 the slowness of the blood stream in connection with the gas glands 

 is supposed to be for the purpose of allowing the cells of the 

 gas gland time in which to absti-act the gases present in the blood, 

 and that this is one function of the rete- mirabile I myself do 



* Haldane (36) comes to the same conclusion. 



t The following suggestion concerning the use of the duplex rete associated with 

 the gas glands may throw some light upon the phj'siology of the similar duplex 

 retia mirabilia which are stated to occur in connection with other structures, 

 e. cf. the " choroid gland " associated with the eye of many teleosts and the rete 

 in connection with the liver of the Tunny. 



