Chemical Analysis of the Atmosphere. 281 



the eudioraetric methods show no greater discrepancy than the errors 

 which may be regarded as inseparable from experiment. The chan- 

 ges however which we observe in the atmosphere, — the varying phe- 

 nomena of meteorology, would naturaHy lead us to think that these 

 proportions must undergo some changes ; and although our experi- 

 ments do not justify such a supposition, it is desirable to ascertain with 

 certainty the variations, if any do occur, however small they may be. 



The two other substances, which are generally regarded as acci- 

 dental constituents of the air, water and carbonic acid, have equally 

 engaged the attention of chemists. The estimation of the water con- 

 stitutes the special branch of Physics called hygrometry, on which 

 subject we have a great number of very valuable researches. The 

 valuation of the carbonic acid has been less attended to. We are in- 

 debted to Theodore de Saussure for an interesting memoir on this 

 subject. 



Although the recent works on these latter points of physics leave 

 little to be desired in relation to the exactness of the method em- 

 ployed, I think it may be useful to possess various methods, based on 

 different principle's, and which may on that account serve as checks 

 to each other, and eventually lead to that exactness which this deli- 

 cate part of chemistry appears to require. It is with this view that 

 I have undertaken a series of experiments for obtaining an easy and 

 sure method of determining in a direct manner the quantity of wa- 

 ter and carbonic acid contained in a given volume of atmospheric air. 



I. Determination of the watery vapor contained in the atmosphere. 



A (Fig.l.) is a cylindrical vessel of glass or metal, — tin for exam- 

 ple, — having two openings a and b, the latter furnished with a stop 

 cock. It contains about thirty quarts of water. The upper orifice 

 a being connected with a horizontal tube of glass, it is evident that 

 the water which flows through the cock b is replaced, in the vessel, 

 by an equal volume of air passing through the tube. By filling the 

 latter with a hygrometric substance, capable of retaining the watery 

 vapor which this air contains, we determine by the increased weight 

 which the tube thus acquires, the proportion of water contained in a 

 volume of air equal to that of the discharged water. 



To use this very simple apparatus with perfect certainty, several 

 conditions are requisite. 



1. The vessel A must contain at least thirty quarts in order to act 

 upon a considerable volume of air. Although in common experi- 



VoL. XXIII.— No. 2. 36 



