But though Dalton's method is correct, as far as the analysis of atmospherical air is concerned, there.. can be little doubt that the proportions which he has assigned as the limits in which oxygen gas and nitrous gas combine are incorrect. We have many trials to ascertain these limits, but never could ob tain the proportions given by Mr Dalton. Mr Gay Lussac has lately turned his attention to this subject, and has given a very simple and satisfactory account of the proportions in which the two gases combine,. (ilIeJnoires D' Armed, tom. ii. 233.) In a paper which he published on the combination of gaseous, bodies, he showed, that in all cases they unite ei ther in equal bulks, or one part in bulk of one with two or three parts of another, and in no intermediate proportions. This opinion was obviously founded on a very ingenious hypothesis of Mr Dalton, relative to the way in which substances combine. This led him to examine the combination of oxygen gas and nitrous gas. The result was, that 100 parts of oxy gen gas unite either. with 200 or with 300 parts of. nitrous gas. The first compound constitutes nitric acid, the second nitrous acid. His method of ana lysing air, founded upon this discovery, is to let up 100 measures of air into a wide vessel, and then to add 100 measures of nitrous gas. In about a minute the absorption is completed. No agitation is to be employed. The fourth part of the diminution gives the oxygen. Suppose the diminution to amount to 84, the fourth of that number, or 21, represents the oxygen in 100 parts of air. On repeating this expe riment, we found it pretty exact, provided the diame ter of the glass vessel employed be not less than three or four inches. In narrow vessels it is very inaccu rate.
The second kind of eudiometer was first propo sed by Volta, and hence is usually known by the name of Volta's eudiometer. It consists in mixing 100 parts of the air to be analysed with 100 parts of hydrogen gas in a graduated tube, and passing an electric spark through the mixture. A detonation takes place ; the whole of the oxygen, and part of the hydrogen, being converted into water. This me thod is very easy, and is susceptible of great preci sion. From numerous and decisive experiments, it follows, that one part of oxygen combines with two parts of hydrogen, when the experiment is made in this way. It appears, too, that the whole oxygen disappears, provided the quantity of hydrogen pre sent be sufficient. We have only to mix 100 parts of air and 100 parts of hydrogen gas together, deto nate the mixture, and observe the diminution of bulk. The third part of that diminution indicates the quantity of oxygen present. Suppose, with the pre ceding proportions, that the diminution of bulk amounts to 63, the third part of that number, or 21, indicates the quantity of oxygen gas in 100 parts of air. He must be a careless experimenter, that, with this eudiometer, commits an error of 1 per cent. We therefore consider it as one of the best means of de termining the proportion of oxygen gas present in any gaseous mixture. It does not answer quite so well when we use it to ascertain the purity of oxy gen gas, or of a gas composed chiefly of oxygen ; be cause, in that case, the diminution of bulk in the ga seous mixture is so sudden and so great, that a va cuum is formed, and the water over which the expe riment is made lets go a portion of air, which mixes with the residue, and makes it appear greater than it otherwise would be, and, of course, diminishes the proportion of oxygen which the gas really contained.
This error is diminished if the water has been recent ly boiled. We cannot state the amount of this er ror; though we have convinced ourselves that it of ten exceeds 2 per cent.
Another method of analysing air, is to expose 100 measures of it to a solution of sulphuret of lime, or of sulphate of iron saturated with nitrous gas. The first of these liquids was recommended by De Marti, the second by Mr Davy. They both answer very well ; they gradually absorb the whole of the oxygen, and leave the azote: Hence the diminution of bulk gives the oxygen in the air examined. The absorption may be made in a graduated tube, or in a eudiometrical instrument contrived for the purpose by Mr Pepys, and described in the Philosophical Trans actions for 1 SOS.
The last method of analysing air which we shall mention, is that of Berthollet. It consists in pla cing 100 measures of air in contact with a cylin der of phosphorus. This method succeeds very well in warm weather ; but it does not answer at all in winter. We have kept a cylinder of phosphorus for a fortnight in contact with air, at a temperature a little above the freezing point, and the air only lost about 3 per cent. of its bulk ; but if you bring it near the fire, the phosphorus soon absorbs the whole oxygen. When the thermometer stands at 70°, the absorption is completed in a few hours. You know the completion by the phosphorus ceasing to smoke. The oxygen, by this process, is removed ; but as time azote dissolves a portion of the phosphorus, its bulk is a little greater than it ought to be. According to Berthollet, to obtain the true bulk of the residua ry azote, you must diminish it by part. This method is not so convenient as the preceding, at least in this country, because it is so tedious ; but it is sufficiently accurate.
The third constituent of the atmosphere is car bonic acid gas. Its presence in the atmosphere was recognised as soon as Dr Black had ascertained the cause of the difference between mild and caustic alkalies : For it was known, that a caustic alkali soon becomes mild by exposure to the air. Dr Black ascertained, that the mildness is owing to the absorption of carbonic acid. From the observations of Saussure we learn, that this gas exists in the at mosphere on the summit of Mount Blanc, which is nearly 16,000 feet above the level of the sea ; for lime water soon deposited its lime in the state of carbon ate, when exposed upon the summit of that mountain, (Saussure's Voyages, iv. 199.) Humboldt found it in a quantity of air. brought down by Garneriu from a height of 42S0 feet, to which he had ascend ed in an air balloon, (Jour. de Phys. xlvii. 202.) It appears, therefore, to constitute a part of every portion of the atmosphere to which we have access.