In the air of the various countries of the earth, at different longitudes and at different altitudes above its surface, on the ocean or on the dry land—in a word, in the air of most diverse localities of the earth—the oxygen and nitrogen are found everywhere to be in a constant ratio. This is, moreover, self-evident from the fact that the air constantly diffuses (intermixes by virtue of the internal motion of the gaseous particles) and is also put into motion and intermixed by the wind, by which processes it is equalised in its composition over the entire surface of the earth. In those localities where the air is subject to change, and is in a more or less enclosed space, or, at any rate, in an unventilated space, it may alter very considerably in its composition. For this reason the air in 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, which abounds in the lower orders of plant life evolving oxygen, contains an excess of this gas.23 The constant composition of air over the whole surface of the earth has been proved by a number of most careful researches.24
The analysis of air is effected by converting the oxygen into a non-gaseous compound, so as to separate it from the air. The original volume of the air is first measured, and then the volume of the remaining nitrogen. The quantity of oxygen is calculated either from the difference between these volumes or by the weight of the oxygen compound formed. All the volumetric measurements have to be corrected for pressure, temperature, and moisture (Chapters I. and II.) The medium employed for converting the oxygen into a non-gaseous substance should enable it to be taken up from the nitrogen to the very end without evolving any gaseous substance. So, for instance,25 a mixture of pyrogallol, C6H6O3, with a solution of a caustic alkali absorbs oxygen with great ease at the ordinary temperature (the solution turns black), but it is unsuited for accurate analysis because it requires an aqueous solution of an alkali, and it alters the composition of the air by acting on it as a solvent.26 However, for approximate determinations this simple method gives results which are entirely satisfactory.
The determinations in a eudiometer (Chapter III.) give more exact results, if all the necessary corrections for changes of pressure, temperature, and moisture be taken into account. This determination is carried out essentially as follows:—A certain amount of air is introduced into the eudiometer, and its volume is determined. About an equal volume of dry hydrogen is then passed into the eudiometer, and the volume again determined. The mixture is then exploded, in the way described for the determination of the composition of water. The remaining volume of the gaseous mixture is again measured; it will be less than the second of the previously measured volumes. Out of three volumes which have disappeared, one belonged to the oxygen and two to the hydrogen, consequently one-third of the loss of volume indicates the amount of oxygen contained in the air.27
The most complete method for the analysis of air, and one which is accompanied by the least amount of error, consists in the direct weighing, as far as is possible, of the oxygen, nitrogen, water, and carbonic anhydride contained in it. For this purpose the air is first passed through an apparatus for retaining the moisture and carbonic anhydride (which will be considered presently), and is then led through a tube which contains shavings of metallic copper and has been previously weighed. A long layer of such copper heated to redness absorbs all the oxygen from the air, and leaves pure nitrogen, whose weight must be determined. This is done by collecting it in a weighed and exhausted globe, while the amount by weight of oxygen is shown by the increase in weight of the tube with the copper after the experiment.