the gases began to cool down. The period of combination would thus be prolonged just as the cyanogen flame is prolonged by inert nitrogen.

In the explosion of cyanogen with oxygen, the high temperature of the wave-front would prevent the direct union of the carbonic oxide formed there with the free oxygen; behind the wave-front, the carbonic oxide mixed with oxygen would fall in temperature below the dissociation point, and combination would begin. I think this hypothesis explains in a satisfactory way the lower intensity of the cyanogen flame when completely burnt, and the prolonged period during which the gases remain luminous behind the wave-front. I may add that both Bunsen and Deville held that the burning of carbonic oxide was retarded by dissociation.

It has been shown that carbonic acid is formed in the passage of the electric spark through the dried gases. The spark also causes dissociation in dried carbonic acid. It may be supposed in the latter case that the carbonic oxide and oxygen liberated by the heat of the spark partially combine on cooling, while the rest, rapidly mixed with inert gas by diffusion, escapes recombination. When the spark is passed in carbonic oxide and oxygen, the gases in its path, at first too hot to combine, are able to do so as they cool. In a mixture of the three gases, an equilibrium is reached when the rate of combination is equal to the rate of dissociation.

Again, the ready union of carbonic oxide and oxygen without flame on the surface of heated platinum may be explained on this hypothesis if we suppose that the platinum acts as a conductor to carry off the heat.*

But though the dissociation of carbonic acid explains some of the anomalies that have been discovered, there are still difficulties to be accounted for. In Smithells' experiment of the divided cyanogen flame, it is not easy to see why the carbonic oxide formed in the inner flame should burn in the outer when the flames are near together, but be extinguished on separating the two cones, unless we grant that freshly formed molecules have a power of combination which they lack when they have had time to "settle down." Possibly the heated gases formed by the first flame act like a metallic conductor, but in that case one would expect the flame of burning carbon bisulphide or cyanogen, when mixed with the carbonic oxide, would act in a similar way.

Brereton Baker (Trans., 1894, 617), who has carried out most patient and skilful experiments on the effect of moisture on chemical combination, is of opinion that there is a resistance to the direct union of elements (except in the nascent state), and that moisture * An idea suggested by Professor Osborne Reynolds.

facilitates chemical union by allowing the transfer of the opposite electric charges of the molecules. In view of this opinion, and of the experiments lately made by J. J. Thomson, it appeared to me of interest to submit the dry mixture of carbonic oxide and oxygen to the Röntgen rays. Professor Schuster was kind enough to submit my tube to a very active discharge. The gases were contained in a thin soda-lime glass tube fitted with platinum wires. The Röntgen bulb was placed about 10 mm. from the eudiometer. Sparks having been passed through the dried gases without inflaming them, the Röntgen rays were turned on, and from time to time, during a quarter of an hour, a spark was passed through the gases, but no inflammation occurred. To prove the nature of the mixture, a small bubble of hydrogen (less than 0.4 per cent. of the total gas) was admitted to the dried gases. After a few minutes had been allowed for diffusion, a spark was passed, causing a violent explosion. The Röntgen rays, therefore, if they affect these gases at all, are not possessed of marked activity.

While I think the dissociation of carbonic acid at very high temperatures may be regarded as one of the limiting factors in the reaction between carbonic oxide and oxygen, it would seem that there is some other cause limiting the direct combination of these gases at lower temperatures.

Owens College,

Manchester.

XLVII.—On the Detonation of Chlorine Peroxide. By Professor H. B. DIXON, M.A., F.R.S., and J. A. HARKER, D.Sc. M. BERTHELOT (Sur la Force des Matières Explosives, 1, 109) discovered the fact that certain endothermic gaseous compounds, e.g., acetylene, cyanogen, and nitric oxide, could be decomposed by the detonation of a small charge of fulminate. The sudden shock arising from the fulminate is communicated, according to M. Berthelot, to the gas immediately round it, breaking down the chemical compound with liberation of heat; the heated molecules so produced in turn act on their neighbours, and thus the decomposition is propagated from layer to layer with the violence and the velocity characteristic of the explosion-wave.

In 1889 Dr. Thorpe described to the Chemical Society the explosion of carbon bisulphide vapour, produced by means of a detonator. He adopted M. Berthelot's view, that the decomposition was of the nature of the explosion-wave.

According to this view, it would seem that the explosion, once initiated, should travel as far as the gas extends, and that the rate of the explosion should conform to the laws governing the propagation of the explosion-wave in gaseous mixtures. It appeared of interest, therefore, to determine whether the explosion, set up by the shock of the fulminate, was propagated along a tube filled with one of these endothermic gases; and, if so, to measure the rate at which the flame was propagated.

In 1891 we made experiments on acetylene and carbon bisulphide. In neither case could we succeed in setting up a true explosion-wave in a tube filled with these gases. Our explosion apparatus consisted of a steel bomb, holding 400 c.c., connected with a strong glass or a steel tube of 15 mm. internal diameter. A copper detonator holding 1 gram of fulminate was fired in the middle of the bomb. With acetylene, although the gas was completely decomposed in the bomb, the decomposition never extended more than 15 cm. down the tube. With the vapour of carbon bisulphide, on the other hand, the flame was more intense, and it travelled from 2 to 24 metres along the tube. The deposit of carbon and sulphur along the tube became gradually thinner until the explosion died out. A true explosion-wave was not set up (Manchester Memoirs, 1892). Last year M. Maquenne published in the Comptes Rendus some similar results with nitrous oxide and acetylene.

With chlorine peroxide, however, even when diluted with other gases, we have found that the explosion is propagated through a long

tube.

The chlorine peroxide mixed with oxygen (and a trace of chlorine) was evolved by warming sulphuric acid and chlorate of potassium in a glass flask by means of a water bath; the gas was washed in a small bulb, and, passing through a sampling vessel, was conveyed by flexible glass tubing to the lower end of a long and stout glass tube, AB, 9.9 metres long,' inclined at an angle of 30°. The two ends of

the long tube (made for us by Messrs. Molyneux and Webbe) were fitted with steel flanges, so that they could be readily clamped up to the two "bridge pieces." These bridge pieces, EF, CD, were steel tubes with taps at one end and flanges at the other. In both, an insulated bridge of silver foil was stretched across the tube close to the flange, by which an electric current was carried to the electromagnetic stylus on the chronograph. The silver bridge was coated, for protection, with paraffin. The lower bridge piece CD was bent; it was a metre in length, and was fitted with firing wires near the tap.

The chlorine peroxide passed into the tube without coming into contact with the metal flange, or any material other than glass. The explosion-tube was inclined, because it was found that the heavy peroxide flowed along the bottom of the tube without pushing the air before it when the tube was horizontal. When the long tube was filled, the flow of gas was diverted. The upper bridge piece was first clamped on, and then the lower; both being previously filled with electrolytic gas. On passing a spark through the electrolytic gas, the explosion-wave was set up, and this was communicated in turn to the chlorine peroxide in the glass tube. The flame was seen to suddenly illuminate the whole glass tube, the passage of the flame being too rapid to be followed by the eye.

We had many premature explosions of the gas, usually set up in the long tube. Only once did the explosion appear to start in the generating flask. At first, we attempted to prepare pure chlorine peroxide, but found it was impossible for us to fill the long tube and clamp on the bridge pieces without firing it; even with the diluted mixtures we used, we made seven experiments running, all of which failed at the last moment. The analysis of the sample showed considerable variations in the proportion of chlorine peroxide present, and we cannot be sure that the samples were of exactly the same composition as the gas in the tube.

Two experiments, we think, gave fairly trustworthy results:

The rate of explosion of chlorine peroxide is thus of the same order as the rate of explosion in gaseous mixtures, in hydrogen and chlorine, for example. Unfortunately, the heat of formation of chlorine peroxide has not been determined, so that we cannot com

pare the observed rate with the theoretical velocity calculated according to Berthelot's formula, or according to the "sound wave" hypothesis. Working backward from the latter, we should predict a heat of formation for CIO, of -15,800 cals. from the first experiment, and of 16,800 cals. from the second. It would be of interest to check this result by direct experiment.

From the density of the mixture and the velocity of the explosion, the pressure produced in the wave may be calculated by Riemann's equation.* According to this calculation, the faster explosion generated a pressure of 31 atmospheres in the wave.

Owens College,

Manchester.

XLVIII.-Morin. Part I.

By HERMANN BABLICH, Ph.D., and A. G. PERKIN, F.R.S.E.

THE yellow colouring matter, morin, exists, as has been known for some time, in old fustic, the wood of Morus or Maclura tinctoria, and has lately been shown by one of us and F. Cope (Trans., 1895, 67, 937) to be also present in the Indian dyestuff" jackwood" (Artocarpus integrifolia). When pure it appears as a glistening mass of almost colourless needles, soluble in alkalis with a yellow coloration, and yielding, with aluminium mordants, very beautiful bright yellow shades.

Hlasiwetz and Pfaundler (Annalen, 1863, 127, 352) assigned to morin the formula C12H,O,,H2O, and described various crystalline monobasic salts to support this formula. Löwe (Zeit. anal. Chem., 1875, 14, 119), however, considered the formula C1H10O, + 2H2O as more probable, and he obtained a lead salt, 2PbO,C1sH10O;; but although the percentage composition of the latter corresponded with this formula the molecular weight of morin was still open to doubt, it being considered by some as isomeric with maclurin, C3H10O6, which, on analysis gives closely similar numbers. By means, however, of the compounds of morin with mineral acids, it was recently established by one of us and L. Pate (Trans., 1895, 67, 644) that its true formula is C15H10O7, it being thus isomeric with quercetin. These acid compounds of morin are very similar to those of quercetin and fisetin, and differ from them only in the fact that during the formation of the sulphuric acid compound 1 mol. of water is eliminated. By gentle reduction with sodium amalgam in acid solution, morin

*See Prof. Schuster's note, Phil. Trans., 1893, 184, 152.