MUSEUM OF COMPARATIVE ZOOLOGY. 15 



Close attention to the movements of young fishes kept in confine- 

 ment led me at first to doubt the accuracy of this conclusion. The 

 result claimed by Moreau ('63 and 'G3 a , p. 820) — that, after the arti- 

 ficial removal of gas from the air-bladder of fishes generally, the gas 

 is regularly renewed, even when they are restrained from coming to the 

 surface — also served to confirm my doubts upon this point. Usually 

 it was only the tip of the bill which protruded from the water, and it 

 seemed incredible that such a movement could be sufficient for the 

 acquisition of even a limited volume of air. In addition, it has often 

 occurred that several times in succession a fish has been observed to 

 come near the surface and emit bubbles of gas without any portion of the 

 head region breaking the surface of the water. 



This seemed conclusive upon at least one point: coming near the 

 surface could not be solely for the purpose of acquiring air. I have 

 given elsewhere (pp. 12, 13) what appeared to me to be the probable ex- 

 planation of the upward movement of the fish previous to emitting gas. 



The possibility that fresh air was not always — perhaps not generally 

 — acquired, led me to reflect further about the possible functions of 

 the air-bladder. Starting with the assumption, that no organ arises 

 absolutely de novo, and that the air-bladder is the result of» a differen- 

 tiation in the alimentary tube, what, in a phylogenetic sense, was likely 

 to have been its first function 1 Did it arise as a hydrostatic apparatus 

 to be ultimately diverted to the service of respiration for the higher 

 vertebrates, or was the hydrostatic function already a superimposed 

 one'? Might it not be that the original function was purely one of 

 excretion ? Perhaps in so primitive a fish as Lepidosteus this original 

 function was still the dominant one. 



These reflections led me to think it more than ever desirable to 

 subject the gas emitted by the gar-pike to chemical analysis, — really 

 the only method of arriving at definite and satisfactory conclusions. 

 I therefore determined to undertake an analysis as soon as the fishes 

 had attained a sufficient size to make the amount of gas given off in the 

 course of a few hours voluminous enough for easy experimentation. 



Accordingly, in December, 1883, when the fishes I'emaining were six 

 or eight inches in length, some preliminary experiments were begun. 

 The analysis was attempted by using a bent tube of about 10 mm. 

 calibre, roughly graduated to 5 mm. ; but the result showed that satis- 

 factory conclusions were attainable only by employing more suitable 

 apparatus. Through the kindness of Professor Cooke the mercurial 

 bath and other apparatus for gas analysis at the Chemical Laboratory 



