In studying the phenomena of nature the conclusion is arrived at that the universally reigning state of mobile equilibrium forms the chief reason for that harmonious order which impresses all observers. It not unfrequently happens that we do not see the causes regulating the order and harmony; in the particular instance of carbonic anhydride, it is a striking circumstance that in the first instance a search was made for an harmonious and strict uniformity, and in incidental (insufficiently accurate and fragmentary) observations conditions were even found for concluding it to be absent. When, later, the rule of this uniformity was confirmed, then the causes regulating such order were also discovered. The researches of Schloesing were of this character. Deville's idea of the dissociation of the acid carbonates of sea-water is suggested in them. In many other cases also, a correct interpretation can only follow from a detailed investigation.
The difference of the weight of a litre of dry air (free from carbonic anhydride) at 0° and 760 mm., at different longitudes and altitudes, depends on the fact that the force of gravity varies under these conditions, and with it the pressure of the barometrical column also varies. This is treated in detail in my works On the Elasticity of Gases and On Barometric Levellings, and ‘The Publications of the Weights and Measures Department’ (Journal of the Russian Physico-Chemical Society, 1894). In reality the weight is not measured in absolute units of weight (in pressure—refer to works on mechanics and physics), but in relative units (grams, scale weights) whose mass is invariable, and therefore the variation of the weight of the weights itself with the change of gravity must not be here taken into account, for we are here dealing with weights proportional to masses, since with a change of locality the weight of the weights varies as the weight of a given volume of air does. In other words: the mass of a substance always remains constant, but the pressure produced by it varies with the acceleration of gravity: the gram, pound, and other units of weight are really units of mass.
For rapid calculation the weight of a litre of air (in a room) in St. Petersburg, may under these conditions (H, t, and f) be obtained by the formula e = 1·20671 + 0·0016 [H1 - 755 + 2·6(18° - t°)] where H1 = H - 0·38f. In determining the weight of small and heavy objects (crucibles, &c. in analysis, and in determining the specific gravities of liquids, &c.) a correction may be introduced for the loss of weight in the air of the room, by taking the weight of a litre of air displaced as 1·2 gram, and consequently 0·0012 gram for every cubic centimetre. But if gases or, in general, large vessels are weighed, and the weighings require to be accurate, it is necessary to take into account all the data for the determination of the density of the air (t, H, and f), because sensitive balances can determine the possible variations of the weight of air, as in the case of a litre the weight of air varies in centigrams, even at a constant temperature, with variations of H and f. Some time ago (1859) I proposed the following method and applied it for this purpose. A large light and closed vessel is taken, and its volume and weight in a vacuum are accurately determined, and verified from time to time.