230 PRINCIPLES OF CHEMISTRY 



earth. In those localities where the air is subject to change, being in a 

 more or less enclosed space, or, at least, an unventilated space, it may 

 alter very considerably in its composition. For this reason the air of 

 dwellings, cellars, and wells, in which there are substances absorbing 

 oxygen, contains less of this gas, whilst the air on the surface of 

 standing water, abounding in the lower orders of plant life evolving 

 oxygen, holds an excess of this gas. 17 The constant composition of 

 air over the whole surface of the earth has been proved by a number 

 of most careful researches. 18 



17 As a proof of the fact that certain circumstances may actually change the composi- 

 tion of air, it will be enough to point out that the air contained in the cavities of glaciers 

 (of permanent mountain ice) contains only up to 10 p.c. of oxygen. This depends on the 

 fact that at low temperatures oxygen is much more soluble in snow-water and snow than 

 nitrogen. When shaken up with water the composition of air should change, because 

 the water dissolves an unequal quantity of oxygen and nitrogen. We have already seen 

 (Chapter I.) that the air boiled off from water saturated at about contains about thirty- 

 five volumes of oxygen and sixty-five volumes of nitrogen, and we have considered the 

 reason of this. It is remarkable that the solubility of oxygen and nitrogen in water de- 

 creases so uniformly with the temperature that the proportion of oxygen and nitrogen held 

 in an aqueous solution remains almost constant at the most varied temperatures. This ex- 

 plains the circumstance that the air over the sea (especially arctic) is, as certain observers 

 have found, poorer in oxygen than on dry land the water dissolves more oxygen than 

 nitrogen. The difference does not, however, exceed 0'3 p.c., and sometimes does not exist. 



18 The analysis of air by weight conducted by Dumas an'1 Boussingault in Paris, and 

 which they repeated many times between April 27 and September 22, 1841, under various, 

 conditions of weather, showed that the amount by weight of oxygen only varies between 

 22'89 p.c. and 23*08 p.c., the average amount being 23'07 p.c. Brunner, at Bern in Switzer- 

 land, and Bravais, at Faulhorn in the Bernese Alps, at a height of two kilometres above 

 the level of the sea, made analyses of the air at the same season of the year as Dumas,, 

 and found that the composition of the air at these places did not exceed the limits deter- 

 mined for Paris. Marignac at Geneva, Lewy at Copenhagen, and Stas at Brussels, 

 confirmed this. The analyses of air taken from different parts of the world, at the 

 surface of the ocean and at different heights above the level of the sea, lead to the con- 

 clusion that air everywhere contains an equal amount of oxygen, or that if it docs vary 

 it does so within very inconsiderable limits. 



As there is some basis (which will be mentioned shortly) for considering that the com- 

 position of the air at great altitudes is different from that at attainable heights namely, 

 that it is richer in nitrogen several fragmentary observations made at Munich and in 

 America gave reason for thinking that in the upward currents (that is in the region of 

 minimum barometric pressure or at the centres of meteorological cyclones) the air is 

 richer in oxygen than in the descending currents of air (in the regions of anticyclones 

 or of barometric maxima) ; but more carefully conducted observations showed this pro- 

 position to be incorrect. Improved methods for the analysis of air have shown that cer- 

 tain slight variations in the composition of air do actually occur, but in the first place 

 they depend on incidental local influences (on the passage of the air over mountains and 

 large surfaces of water, regions of forest and vegetation, and the like), and in the second! 

 place are limited by quantities which are scarcely distinguishable from possible errors in 

 the analyses. 



The considerations which make one think that the atmosphere at great altitudes con- 

 tains less oxygen than at the surface of the earth are based more particularly on the law 

 of partial pressures (page 81). It obliges one to consider that the equilibrium of the 

 oxygen in the strata of the atmosphere is not dependent on the equilibrium of the nitro- 



