266 
the great variations in the results which they afford. 
In a tat>le, which presents the results of seven differ- 
ent experiments, conducted in a similar manner, the 
consumption of oxygen to the carbonic acid produced, 
is represented, in one instance, as bearing the pro- 
portion of 1OO to 91 ; in another, the proportion is 
as 100 to 50; and in a third experiment, as 100 to 
20 * ; so that the same animals appear, at one time, 
to have consumed only one-tenth ; at another time, 
one- half ; and, at a third time, four-fifths of the oxy- 
gen employed ; contrarieties which sufficiently de- 
monstrate the existence of some undiscovered sources 
of error. These experiments, therefore, can be con- 
sidered as proving only the general fact of the con- 
version of oxygen gas into carbonic acid by the re- 
spiration of fishes ; but they do not enable us to de- 
termine the actual extent to which this conversion 
proceeds. As, however, in the experiment that has 
been detailed (552.), nine-tenths of the oxygen, 
which disappeared, were obtained in the form of car- 
bonic gas, it is, surely, more reasonable to suppose 
that the remaining one-tenth was converted into the 
same gas, than that it entered into, and was retained 
in the animal system. On this point we may be per- 
mitted to call in the aid of analogy ; for as, in the 
respiration of every other class of animals, the bulk 
of acid produced is proved to be nearly or exactly- 
equal to that of oxygen consumed, it may be fairly 
inferred that a similar equality obtains in the respira- 
tion of fishes, to which, indeed, the volumes of air iu 
* Mem, cTArcueil, torn. ii. p. 378, 
