MUSEUM OF COMPARATIVE ZOOLOGY. 21 



the water of the river Seine contains oxygen and nitrogen in the pro- 

 portions determined by Humboldt et Provengal ('09), but that, owing 

 to the defective means of extracting the carbon dioxide employed by 

 them, they secured only one fortieth of the volume of that gas which 

 was actually present (Grehant, '69, p. 379). 



Such being the case, it is evident that reliance cannot be placed on 

 Baumert's conclusions, for his opinion — that there was no perceptible 

 absorption of gases under the conditions of experimentation previously 

 detailed — rested ultimately upon his experiment with intestinal respi- 

 ration in distilled water, and his ability subsequently to extract the 

 whole of the gas absorbed by the water during the experimentation. 



Being unwilling to sacrifice the limited number of young fishes in my 

 possession for the purpose of securing the contents of the air-bladder, 

 and not having the opportunity of getting the gas from the air-bladder 

 of adults, I determined to employ a method which seemed likely to 

 furnish positive evidence of the presence of carbon dioxide, if any were 

 really eliminated, even though it would not give a rigid quantitative test. 



The plan was, to place the fishes in a vessel of water having a limited 

 air-chamber, which could be rapidly swept of its contents by a con- 

 tinuous flow of pure air through it, and to lead the air, thus drawn 

 from the experimental jar, through baryta water, which would indicate 

 the presence of even a small amount of carbon dioxide by the formation 

 of a precipitate. 



Accordingly, a glass jar (Figure 2, j) of about 20 litres' capacity, pro- 

 vided with a thick ground-glass cover having a central neck and orifice 

 for a stopper, was selected. The cover was larded t© secure a close 

 joint, and in addition the edge was sealed with melted paraffine, and 

 the jar nearly filled with recently boiled distilled water. This was 

 siphoned into the jar with as much precaution as possible in order to 

 prevent exposure to the atmosphere. To make the water habitable for 

 the fishes, it was of course necessary to aerate it, and it was at the same 

 time important not to introduce in this process any carbon dioxide. To 

 accomplish this successfully was found to be a task much more labori- 

 ous than was at first anticipated. The apparatus shown in Figure 2 

 was that which finally proved satisfactory. 



A Bunsen pump (p) attached to the hydrant provided the suction 

 required, and the system of tubes and chambers through which the air 

 was drawn was arranged as follows. Four glass combustion tubes (k) 

 each about 2 meters long and 2 cm. in diameter were loosely filled with 

 fragments of potassic hydrate, and joined by short pieces of rubber 



