THE AURORA BOREALIS 217 



regions of the atmosphere, that is, that its proportion, 

 relatively to that of the more abundant gases oxygen and 

 nitrogen, may increase as the altitude grows greater ? To 

 this, I think, an affirmative answer may be given. Let us 

 consider the grounds for the supposition. 



When a gas is compressed it turns warm, as every one 

 knows who has used a bicycle-pump. Conversely, when 

 it escapes from compression it cools itself. But all gases 

 do not heat or cool equally for equal amounts of compres- 

 sion or expansion ; for some gases are raised to a higher 

 temperature than others by absorbing the same quantity 

 of heat; and the same quantity of heat is, practically, 

 generated by the same degree of compression, or absorbed 

 by the same degree of expansion; for work is quanti- 

 tatively equivalent to heat, as was shown by Joule half a 

 century ago. Now 40 grams of argon should, if the 

 specific heat of that gas were the same as that of oxygen, 

 require the same amount of heat to raise its temperature 

 through 1 degree ; or put in another way, if each of these 

 gases were expanded to the same amount, they would be 

 equally cooled, if equal amounts of heat were requisite to 

 raise their temperatures through an equal number of 

 degrees. But this is not the case. Argon requires less 

 heat to raise its temperature than oxygen, in the ratio of 

 3 to 5 if it is not allowed to expand, or in the ratio 

 of 5 to 7 if expansion is possible under constant pressure. 

 If allowed to expand under circumstances in which volume 

 increases while pressure falls, some ratio intermediate 

 between these would show the difference in cooling ; the 

 exact amount depending on the degree of expansion in 

 question. Broadly stated, argon, in expanding will cool 

 itself considerably more than oxygen or nitrogen ; while its 

 congeners, helium, neon, krypton, and xenon, will exhibit 

 a degree of cooling practically identical with that of argon. 



The next point to be considered is that gases diffuse 





