and the Observed Velocity of Sound in Air and Gases. 29 



chemical action, but few have ventured to regard the atmosphere 

 in any other light than a mechanical mixture. But if such an 

 action exists among the particles of dissimilar gases during 

 intermixture, it is a legitimate question whether there are any 

 indications of the manifestation of feeble chemical forces not 

 hitherto recognized by chemists. Let us see what bearing the 

 physical properties of the atmosphere have on this question. 



It is obvious that, the densities of the nitrogen and oxygen in 

 relation to air being known, the percentage of these gases can be 

 computed, provided the volume of the mixture is equal to the sum 

 of the volumes of the two constituents, which is the assumption 

 made by the theory of mechanical mixture. The validity of this 

 assumption can be tested by comparing the results of calcula- 

 tion with those of the best analyses. The experiments of Reg- 

 nault furnish the numbers in relation to density*. 



Let V = volume of oxygen, 

 V'= volume of nitrogen, 



d = density of oxygen (air being unity) =1*105612, 

 d' = density of nitrogen (air being unity) = 0*971346, 

 we have 



^xV + d'xV' = 100xl. 

 Also 



V + V' = 100. 

 Hence 



-. 100-100 x<* 2-8654 01 _._ 



V= d^dT- = 0134266 = 21 ' 341 P er Cent 



Similarly, 



V'= =78-659 „ 



100000 

 The proportion by weight would be 



Oxygen . . . . = 23-595 

 Nitrogen .... = 76-405 



100000 



The results of the volumetric method of analysis afford the 

 best test of this question, not only because they are probably 

 more accurate, but likewise because the analytical processes 

 assume the same condition of things in relation to the consti- 



* Memoires de VAcademie des Sciences, vol. xxi. pp. 138,142, 145. The 

 numbers expressing the densities of oxygen and nitrogen are slightly dif- 

 ferent from those given by Regnault in his memoir. This arises principally 

 from an error which is made in the reduction of his first experiment 

 (p. 136) for determining the exact weight of a given volume of air. Of 

 course this error pervades all the densities which are related to it. I have 

 carefully revised all Regnault's computations for the reductions, and 

 have obtained the numbers employed in the text. 



