APPARATUS FOR THE ANALYSIS OF GASSES. 89 



Marti. The residuary oxigen being deducted from the 

 quantity at the outset of the experiment shows how much 

 qxigen has been expended in the combustion of the inflam- 

 mable gas. It is scarcely necessary to observe, that the 

 gasses are carefully reduced, at each stage of the operation, 

 to a mean temperature and pressure, (60° of the thermome- 

 ter, and 30 inches of the barometer)*. 



The process of combustion, as thus stated in general 

 terms, appears sufficiently simple. It is often, however, ren- Source of w 

 dered complicated by the imperfect combustion of the in- rour * 

 flammable gas, a part of which escapes through the orifice 

 of the burner, either wholly unaltered, or only partially 

 burned. As this portion- is not absorbed by sulphuret of 

 lime, it gives a fallacious appearance of an actual addition 

 of nitrogen to the oxigen gas remaining in the receiver b. 

 I .am unacquainted with any method of entirely obviating 

 this difficulty ; but its amount may be diminished by an Precautions 

 attention to certain precautions. With this view, the pres- a £ ainst ll * 

 sure upon the gas, contained in the receiver o o, should, on 

 first opening the cocks q and r, be no more than is sufficient 

 for its gentle expulsion through the tube s s. When, how- 

 ever, the stream is once kindled, the larger the flame, and 

 the more active the combustion, within certain limits, the 

 more completely is the gas consumed. It is necessary, also, 

 to stop the combustion, before it is rendered languid by the 

 admixture of carbonic acid with the gas in the receiver b, 

 and by the diminished purity of the oxigen gas. If this be 

 not attended to, a large proportion of the inflammable gas, 

 toward the close of the process, makes its escape unaltered 

 into the receiver o. In general I have found, that, setting The more com- 

 out with oxigen gas of equal purity, the more combus- bustiblethegas 

 tible the inflammable gas submitted to experiment, the p i ete the " de- 

 more complete is its decomposition by slow combustion.' The composition. 

 apparatus, therefore, is better adapted to the analysis of ole- 

 fiant gas, of carburetted hidrogen gas, or of mixtures of 

 these two, than of carbonic oxide, or any gas of which this 

 oxide forms a large proportion. 



* The rules observed in these calculations are stated in my Epitome of 

 Chemistry, $th edition, p. 441. 



The 



