a property discovered by Dr. Priestley, but first applied to the purpose of its analysis by Sir H. Davy, in the course of his researches into the compounds of nitrogen.* In the experiments of that philosopher, the results, approaching most nearly to precision, were obtained by detonating nitrous oxide with rather more than an equal volume of hydrogen, viz. 39 measures of the former to 40 of the latter. Both gases were in this case decomposed; water was produced; no nitrous acid was formed; and a volume of nitrogen remained, which always a little exceeded that of the nitrous oxide decomposed, viz. in the proportion of about 41 to 39.

F

In the repetitions which I have frequently made of this experiment, a similar excess of the accruing nitrogen, over the volume of the nitrous oxide employed, has always been observed, and generally in about the same proportion. But according to the law regulating the combination of gaseous bodies with each other, which has been deduced by M. Gay Lussac from a great variety of examples,† all tending to show, that gases unite in proportions as to volume, which are either. equal, or simple multiples of each other, nitrous oxide ought to be constituted of exactly one volume of nitrogen and half a volume of oxygen condensed into the space of one volume; and those products should result from every careful decomposition of the gas in question. That they are not correctly obtained by the method which I have just alluded to, appears to be owing to sources of inaccuracy, necessarily connected with that mode of analysis. I was induced, therefore, to try various other processes, among which there is one that may deserve to be made known, since it exhibits, in a very summary way, and by a single operation, the quantities of nitrogen and oxygen that enter into the constitution of nitrous oxide, with as much precision as, I believe, is attainable in the present state of gaseous analysis. This method consists in firing, by the electric spark, a mixture of nitrous oxide and carbonic oxide in due proportions. The nitrous oxide, which I employed, was obtained by the careful decomposition of nitrate of ammonia, and did not contain in 100 parts more than 3 parts of gas unabsorbable by well boiled water. The carbonic oxide was gene rated from recently ignited chalk and iron filings, and after having been washed with caustic potash, appeared, from the results of its combustion with oxygen, to be contaminated with not more than 3 per cent. of foreign gas, having the properties of nitrogen. Some nicety was found to be necessary in adjusting the proportions of the gases to each other, in order to obtain a perfect decomposition. When an excess of nitrous oxide

* Davy's Researches, London 1800, p. 286. itt Memoires de la Soc. d'Arcueil, ii. 207.

was used, some free oxygen was always detected in the residues and yet a slight redundancy of nitrous oxide appeared to be essential to the perfect combustion of the carbonic oxide. After firing the mixed gases, and removing the carbonic acid by liquid potash, I next determined the proportion of oxygen in the residue by commonly known methods, and considered the remainder as nitrogen gas. An example, taken from an experiment made with great care, will best illustrate the nature of the process.

Carbonic oxide 25 measures
Nitrous oxide
26 do.

51

24.25 pure +0.75 azote.

25.25 do. +0.75 do.

52 after combustion.

28 after potash; found to consist of 0.85 oxygen +27.15 nitrogen.

In this case, the carbonic acid, from 24-25 real carbonic oxide was 24 measures.

The nitrogen was by experiment

By calculation it ought to have been,

27.15

1

b From the nitrous oxide 26 ad Do. carbonic oxide

.75

Difference..

26.75

0.40

No person, I believe, who is much experienced in processes of this nature, will look for a nearer approach to accuracy, than in the results of the experiment which has been just described; for the carbonic acid falls short by only of the theoretical proportion; while the nitrogen exceeds that proportion by only

and the oxygen by. The experiment was several times repeated, with approximations fully as near as the above to those which the law of volumes would require.

When the object in view is solely or chiefly to determine the quantity of oxygen in nitrous oxide, there can be no source of fallacy in the use of a slight excess of carbonic oxide. For this purpose I made several experiments, which agreed so closely in their results, that it may be sufficient to particularize one of them as a specimen of the rest.

In this case, the carbonic acid was 40 215 18-5, and therefore exceeded in volume the theoretical proportion (18.4) by only, a deviation much within the limits of possible errors, arising either from the difficulty of measuring small quan tities, or of ascertaining the purity of nitrous oxide. We may, also, from this mode of operating, deduce the quantity of nitrogen, which exists as an element of nitrous oxide; for since 1 volume of carbonic acid results from the combustion of 1 volume of carbonic oxide, the residuary 21-5 measures must have contained 21 18.5 2.5 of carbonic oxide and nitrogen introduced by that gas, +19 measures of nitrogen disengaged from the nitrous oxide.

=

The results of the experiments with an excess of carbonic oxide suggested to me a ready and correct method of testing the nitrous oxide, which had hitherto been a desideratum. The only test, before applicable to this purpose, was the amount, to which the gas is absorbed, when agitated with well boiledwater. But besides the uncertainty whether all the nitrous oxide be in this case condensed, the proportion of the unab sorbed residuum is subject to variation, from the quantity of other gases extricated from the water itself. Reduced to the form of a rule, the new method may be stated as follows: Let a given volume of nitrous oxide be exploded with a slight excess of carbonic oxide of known purity; for example, 110 or 115 measures of the latter to 100 of the former. Now as each volume of real nitrous oxide gives, under these circumstances, an equal volume of carbonic acid, we may impute whatever carbonic acid is deficient of that proportion to the mixture of so much nitrogen with the nitrous oxide. If, for example, using an excess of carbonic oxide, there should result, from 100 measures of nitrous oxide, only 95 of carbonic acid, we' may safely consider the nitrous oxide to be contaminated with 5 per cent. of nitrogen gas. A proportion of nitrous gas may, I am aware, be occasionally mixed with the nitrous oxide, but this may be easily discovered, and previously separated, by solution of green sulphate of iron.

Having determined the application of carbonic oxide to the analysis of nitrous oxide to be so easy and satisfactory, I had hoped that the same agent might have been employed in the analysis of nitrous gas, which, as is well known, does not form a combustible mixture with simple hydrogen gas. But on trial, I could not, by any variation which I made in the proportions of the two gases, obtain a mixture combustible by electricity. I had recourse, therefore, to olefiant gas; but had nearly abandoned this method also as impracticable, on finding that the mixture could not be set on fire by a spark from the prime

* Davy's Researches, p. 136,

conductor of an electrical machine. The discharge, however, of a small Leyden jar, through a mixture of nitrous gas and olefiant gas, occasioned a vivid combustion, and both gases were entirely decomposed. The following experiment may be taken as an example:

Olefiant gas 6.5

Nitrous gas 465 451 pure + 1-4 nitrogen.

53.0

40 fired.

27 washed with potash.

In this case 40 27 13 measures of carbonic acid were formed, which are just double the volume of the olefiant gas. In the residuary 27 measures, I found 27 measures of free oxygen. But 65 measures of olefiant gas require for saturation 19.5 of oxygen, to which, adding the residuary 2·7, we have 22-2 measures of oxygen by experiment in 451 nitrous gas; while theory would require 22.55 or about 4 more than was actually obtained. Again, the residuary nitrogen was 27 - 2·7 243; while from theory it should have been half the volume of the pure nitrous gas, viz. 22:55 + the impurity of the latter 1.4 23.95. The actual proportion of nitrogen, therefore, exceeds the estimated by only 4th.

It may be stated, then, in general terms, as the results of analyzing nitrous oxide and nitrous gas by the methods which have been described in this paper;

1stly. That I volume of nitrous oxide is decomposed by 1 volume of carbonic oxide; and the products are 1 volume of carbonic acid and 1 volume of nitrogen. But to convert 1 volume of carbonic oxide into an equal volume of carbonic acid, half a volume of oxygen is required. Therefore 1 volume of nitrous oxide must be constituted of 1 volume of nitrogen + half a volume of oxygen in the space of 1 volume.

2dly. That 6 volumes of nitrous gas require for perfect decomposition 1 volume of olefiant gas, and the gaseous products are 2 volumes of carbonic acid and 3 volumes of nitrogen. But to form 2 volumes of carbonic acid by the combustion of carbon, 2 volumes of oxygen are necessary; and 1 volume of oxygen is required to saturate the 2 volumes of hydrogen existing in 1 volume of olefiant gas. The results of this experiment, therefore, confirm the analysis both of nitrous gas and olefiant gas by other methods; for the former gas must consist of equal volumes of nitrogen and oxygen gases not condensed in bulk; and 1 volume of olefiant gas must be constituted of 2 volumes of hydrogen + carbon sufficient for forming 2 volumes of carbonic acid.

ARTICLE XVII.

SCIENTIFIC NOTICES.

CHEMISTRY.

1. Ignition supported by Hydrophosphoric Gas, &c. (Extract of a Letter from M. J. B. Von Mons to M. Planche.)

I have lately observed, on kindling phosphuretted hydrogen not spontaneously inflammable, that the bubbles which are slowly generated maintain the ignition of a lighted match, without inflaming it, and are themselves inflamed by the incandescent flameless body: this has some relation to Doebereiner's lamp. You have undoubtedly already seen that hydrogen, after having burnt for some minutes in the philosophical candle, heats the end of the tube sufficiently to cause the gas to be relighted immediately after it is blown out. The hydrogen in this candle inflames spontaneously if the mixture of the sulphuric acid and the water be made in the bottle itself.-(Journal de Pharmacie.)

2. Effect of Prussic Acid on Vegetation.

C. I. Th. Becker (Dissertatio de Acidi Hydrocyanici Vi perniciosâ in Plantas. Jena, 1823. 4to.) has made many experiments, from which it follows that prussic acid prepared by Vauquelin's method destroys vegetables nearly in the same manner as it acts on animals. Seeds steeped in this acid either die or lose the power of germinating. The more delicate vegetables perish under its influence sooner than the more robust.(Journal de Pharmacie.)

3. To preserve the Colour of Red Cabbage.

Digest the leaves of the cabbage in warm alcohol, and when the whole of the colouring matter is extracted distil off a portion of the spirit, and evaporate the remainder, at a very gentle heat, to the consistence of a syrup. This extract may be preserved unimpaired for years, if kept in closely stopped phials. In order to use it, it is only necessary to add a small portion of it to water, in which it is readily soluble, when the addition of an acid or an alkali will produce its peculiar effect. When we wish to employ this test to discover small quantities of carbonic acid, it is necessary to render it slightly green by the addition of a diluted alkali. The carbonic acid will then restore the bluecolour, by saturating the alkali. Test papers may also be pre pared by means of the alcoholic tincture of the cabbage, which, when rendered green by immersion in a diluted alkaline solution, may be used in all those cases in which litmus papers are com monly employed.-(American Journal of Science.)