73 



tained by Vauquelin, as to render it difficult to con* 

 ceive, but for some occasional error, that they should 

 not entirely coincide. Of the probable existence of 

 such error, the great difference in the analysis, made 

 by the combustion of phosphorus, and by the test of 

 nitrous gas, furnishes strong suspicion: and it is 

 known that this latter method requires many precau- 

 tions to insure complete accuracy. Even if all these 

 had been observed, more attention to the actual sur- 

 rounding temperature was required, than seems, in 

 these experiments, to have been bestowed ; for, in 

 a small graduated tube, air is itself the most sensible 

 thermometer, and readily suffers a change of bulk 

 from the slightest variations of temperature, and, in 

 the measurement of small quantities, demands, there- 

 fore, the utmost attention and correctness. In his 

 analysis of the air which had been respired by snails, 

 M. Vauquelin could not well be deceived by the non- 

 combustion of phosphorus ; and, although this fact 

 does not disprove the assertion of Spallanzani, yet, 

 as it proves more, it is entitled to the preference ; 

 for, where the same degree of credit is attached, a posi- 

 tive must ever outweigh a negative proof. The snail 

 too, in this experiment, might have died from some 

 accidental cause, before the consumption of the oxy- 

 gen gas was complete, while, in other instances, the 

 same species of animal would entirely consume it : 

 and we find, accordingly, that Spallanzani himself, 

 in several parts of his work, asserts that snails, by 

 their respiration, consume all the oxygen gas of the 



* Memoirs on frc-spiration, p. 153. 165. 172. 213. 222. 2*9. 340. 



