The Effect of Steam on the Burning of Cyanogen to Carbonic acid. Considering the effect produced by the presence of steam on the explosion of a mixture of carbonic oxide and oxygen, it might be conjectured that when cyanogen is burnt to carbonic acid the presence of steam might materially affect the rate of combination, and consequently the length of the flame. But the flame produced by a well dried mixture of cyanogen and oxygen does not appear to differ from the flame given by the same gases mixed with 1.5 per cent. of steam. This result, which was contrary to our expectations, was shown by a series of photographs of the flames produced by the explosion of the dried and wet mixtures under similar conditions. As before, two tubes, furnished with similar windows, were fixed side by side. After thoroughly drying one tube, we filled it with the mixture which had been standing some days over phosphoric oxide. The second tube was filled with gas nearly saturated with steam. The mixtures were fired at the same time, and the flames on reaching the windows in the two tubes were photographed on the moving film. Both with slow and high velocities of the film, the images were drawn out equally. We can only conclude from this result that, in a mixture of freshly formed carbonic oxide and oxygen, these gases unite as the mass cools down without the intervention of steam; and where steam is present it does not appreciably alter the time during which the change continues.

This result agrees with Professor Smithells' observation that when a dried mixture of cyanogen and air is burnt in his " separator," the carbonic oxide formed in the inner cone will burn in dry air, provided the outer cone is kept close to the inner; but when the cones are widely separated, so as to allow time for the carbonic oxide formed below to "settle down," the outer flame is extinguished on bringing it into dry air.

The Effect of Steam on the Flame of Carbonic Oxide and Oxygen.

Experiments similar to the above, made with the tubes filled with dried and moist carbonic oxide and oxygen respectively, show a marked difference (Fig. 9). In the presence of steam, the flame is much shorter. In the dried mixture, the gases were passed through several tubes containing boiled sulphuric acid, and through a column of phosphoric oxide. Care had to be taken to allow the flame a long run before reaching the window, as in the dried gases the explosion-wave is not set up for several feet. The photographs of the carbonic oxide flame thus seem to bear out the hypothesis, founded on the rates of explosion, that "at the extreme temperatures

of the explosion-wave, as well as in ordinary combustion, carbonic oxide is oxidised by the steam" (Phil. Trans., 1893, 184, 112).

A considerable body of evidence has now been accumulated concerning the mode of burning of cyanogen. The principal points established are:

(i) Cyanogen burns and explodes with oxygen in the absence of moisture.

(ii) Cyanogen in explosions forms carbonic oxide with great rapidity, and carbonic acid with less rapidity.

(iii) Cyanogen, burning in air, appears to undergo oxidation in two stages.

(iv) The presence of moisture does not affect the formation of carbonic acid from freshly formed carbonic oxide.

If we have rightly interpreted the facts before us, it would appear that cyanogen, in explosions, burns first to carbonic oxide, and then, if oxygen is present, this carbonic oxide proceeds to unite with it to form carbonic acid. This formation of carbonic acid is slow com. pared with the formation of the carbonic oxide; nevertheless, the reaction is rapid enough to influence the velocity of explosion of cyanogen with a large excess of oxygen.

Owens College,

Manchester.

XLVI.-The Mode of Formation of Carbonic acid in the burning of Carbon Compounds.

By Professor HAROLD BAILY DIXON, M.A., F.R.S.

THE usual statements concerning the burning of carbon compounds, namely, that the carbon unites directly with oxygen to form carbonie acid, and that this is reduced in presence of an excess of the carbon to carbonic oxide, appear to have little or no experimental foundation. All the direct evidence points the other way.

But with regard to the mode in which carbonic acid is formed, and the part played by moisture in its formation, the evidence is curiously puzzling. Granting that carbon in any compound may burn with oxygen to carbonic oxide, we have to deal with the following facts:

1. An explosion is not propagated by a spark in a dried mixture of carbonic oxide and oxygen at ordinary temperature and pressure (Dixon, Brit. Assoc. Report, 1880, 503).

2. When a long tube full of the dried gases has moisture introduced at one end and the gases are ignited there, the flame traverses

the moist gases, but dies out on reaching the dry gases (Dixon, Chem. News, 1882, 46, 151). The flame of dry carbonic oxide burning in air is extinguished by plunging it into dry air (M. Traube, Ber., 1885, b, 1890).

3. The addition of steam increases the rapidity of inflammation of carbonic oxide and oxygen, both in the ordinary flame and in the explosion-wave (Dixon, Phil. Trans., 1884, Pt. 2, 640, and 1893, A, 111).

4. Carbonic oxide and oxygen appear to be inert towards one another in the incomplete combustion of hydrogen and carbonic oxide; at all events, they act as if they were incapable of rapidly uniting (Dixon, Phil. Trans., 1884, Pt. 2, 671).

5. Dried carbon does not glow when heated in dried oxygen; it burns, however, to carbonic oxide. In the presence of finely divided platinum, carbonic acid is formed (Baker, Phil. Trans., 1888, A, 571).

6. Dried carbonic oxide and oxygen will unite in the path of the electric spark (Lothar Meyer, Ber., 1886, 19, 1099; Beketoff, Bull. Acad. St. Pétersbourg, [N.S.], 2, 175; Dixon). The amount of combination depends upon the nature of the spark; a limit is reached when the rate of combination is equal to the rate of decomposition of the carbonic acid formed (Dixon and Lowe, Trans., 1885, 571).

7. The dried gases completely combine, without producing flame, in contact with a heated platinum wire (Dixon and Lowe, Trans., 1885, 576).

8. Dry carbonic oxide proceeding from a half-burnt cyanogen flame will burn in dried air, provided the two flames are close together (Smithells and Dent, Trans., 1894, 603).

9. An explosion of cyanogen and oxygen will cause the union of dry carbonic oxide and oxygen with which they are mixed (Beketoff, Bull. Acad. St. Pétersbourg, [N.S.], 2, 175).

10. In the explosion of cyanogen with an excess of oxygen, the formation of carbonic acid is complete, and is not affected by the presence of moisture (Dixon, Trans., 1886, 384). The complete combustion of cyanogen, however, does not take place so rapidly or with such violence as the incomplete combustion to carbonic oxide (Dixon, Strange, and Graham, Trans., 1896).

11. Moisture is active in the oxidation of other substances, e.g., sulphur and phosphorus (Baker, Phil. Trans., 1888, A, 581); it also affects the combination of other substances (Cowper,* Trans., 1883, 153; Baker, Trans., 1894, 611; Pringsheim, Ann. Phys. u. Chem.,

The effect of moisture in determining the combination of chlorine and sodium, the discovery of which is usually attributed to Wanklyn, is not mentioned by Wanklyn in his paper (Chem. News, 20, 271).

[N.F.], 32, 384), and facilitates the passage of the electric discharge in gases (J. J. Thomson, Phil. Mag., [5], 36, 313; Brit. Assoc. Report, 1894, 482).

Several explanations have been put forward to account for the influence of moisture in promoting the combination of carbonic oxide and oxygen; some of these have special reference to this reaction only, whilst others are of more general application. Most of the hypotheses advanced have, on examination, suggested experiments by which their probability might be tested. It will, I think, be useful to examine these hypotheses in the light of the evidence which has been accumulated. Even if it should appear, as I think, that no satisfactory explanation has been arrived at, a discussion of the results may focus attention on the essential points of the problem, and may serve to suggest future experiments.

In my papers on the subject I have made no attempt to suggest a reason why oxygen should not explode with carbonic oxide. I have limited myself to the conclusions which seemed to accord most closely with the observations: (a) that carbon burns primarily to carbonic oxide; (b) that carbonic oxide is inert, under ordinary conditions, towards oxygen; and (c) that carbonic oxide at a high temperature decomposes steam to form carbonic acid and hydrogen, so that, in the ordinary explosion of moist carbonic oxide and oxygen, the steam acts catalytically as a carrier of oxygen.

It is evident, from the results recorded in the previous paper on "The Explosion of Cyanogen," that the second conclusion (b) must be modified. Under certain conditions, carbonic oxide and oxygen unite readily. We are confronted, therefore, with the apparent anomaly two gases which are not exploded by the spark or application of a flame, which are inert, or do not uuite rapidly, as other gases do, in explosions, nevertheless combine in presence of heated platinum, and in presence of other combining gases, or if one of them has just been formed by a chemical reaction.

Professor H. E. Armstrong,* taking a wide survey of the conditions of chemical combination, considers the function of the steam in the explosion of carbonic oxide and oxygen to be that of the liquid in a galvanic cell. The two gases, inert in the pure state, require "the formation of a conducting system in which electrolysis can occur." Comparing the system of carbonic oxide, steam, and oxygen to a Grove gas battery, Armstrong regards the oxidation of the carbonic oxide by the oxygen of the water as dependent upon the simultaneous oxidation of the hydrogen of the water by the free oxygen.

"Presidential Address to Chemical Section, British Association, 1885," Proc. Roy. Soc., 1886, 40, 287; Trans., 1886, 49, 112. "Presidential Address to Chemical Society," Trans., 1895.

The general theory advocated by Armstrong is that interaction does not occur between two pure substances, and that water will condition the interaction between them only if it be capable of forming a composite electrolyte with one of them. I am not sure whether Armstrong considers that water forms a composite electrolyte with carbonic oxide, or whether water itself is the electrolyte; but, in either case, it is not easy to explain all the facts. The theory, as I understand it, supposes that the three substances, carbonic oxide, steam, and oxygen must be in contact while the change proceeds, and any two of these gases will be inert on meeting. If, therefore, we accept the molecular theory of gases, the change will only occur on the simultaneous collision of molecules of the three substances; or, taking Armstrong's equation, on the simultaneous collision of five particular molecules, two of carbonic oxide, two of steam, and one of oxygen. What the chances are in favour of such collisions occurring might, I suppose, be calculated; it is obvious that their number must be exceedingly small, compared with the collisions between any two molecules. Supposing this quintuple collision occurred, the heat evolved by the chemical change would be divided among the products of interaction; the heated steam and carbonic acid molecules so formed would fly in all directions, and inevitably some steam molecules would come into violent collision with unburnt carbonic oxide molecules. What would happen? Experiment shows that if we heat carbonic oxide and steam together, they react, forming carbonic acid and hydrogen. Do they do so in the case supposed above? If they do, carbonic acid is formed in a manner independent of electrolytic conduction. If they do not, how are we to explain the action of heat on carbonic oxide and steam?

Again, the electrolyte theory has to explain the oxidation of carbonic oxide in the burning of cyanogen. As we have seen, moisture does not appear to be necessary for this reaction or to appreciably affect the result if present. I have tried an experiment to see whether some oxide of nitrogen, which might be formed in the initial combustion of the cyanogen, can act the part of “catalyst" between carbonic oxide and oxygen. Peroxide of nitrogen I found to have no effect; nitrous oxide I had previously tried with the same result.

But apart from these difficulties, we must bear in mind that an explosion in moist carbonic oxide and oxygen is propagated at a