besides the primary reaction in which nitrogen and chromium oxide are mainly formed, there is a secondary reaction resulting in the formation of one or more of the oxides of nitrogen, and of an oxide of chromium of a composition not expressed by the formula Cr203. I give the result of an actual experiment.

Volume of nitrogen at 0° C. and 760 mm. pressure, 111.946 c.c. Volume of nitrogen calculated at 0° C. and 760 mm. pressure (if Cr=52), 114.740 c.c.

Volume of nitrogen calculated at 0° C. and 760 mm. pressure (if Cr=52·5) 114·292 c.c.

I assumed that the atomic weight of O= 15·96, of N = 14·02, and that at standard temperature and pressure a litre of nitrogen weighs 1.25700 grams.

Method B. Reduction of Ammonium Dichromate to Chromium Chloride.

In this series of experiments, as previously mentioned, a known weight of the dichromate was reduced with great care to chromium chloride, precipitated with the least possible amount of ammonia, and then ignited and the oxide weighed. It is clear, therefore, that besides ammonium dichromate four other compounds, namely, water, hydrochloric acid, alcohol, and ammonia, took part in the reaction. These substances all readily lend themselves to purification, and the method employed for this purpose in each case I will briefly indicate. A platinum dish of 200 c.c. capacity was taken and tared against another dish of the same size and almost the same weight, the difference in the weight between the two dishes being made up by means of a small piece of platinum foil. These two dishes were treated in all respects alike, when one was placed on the water-bath, so also was the other, and for the same time; they were ignited in a similar manner, and they were placed in desiccators for the same length of time, &c. In this manner the various corrections, at best uncertain, which would have been necessary in computing the actual weight of the dish, were obviated. They were both covered with platinum foil, but their covers were not weighed with them.

The dishes having been carefully tared, a known volume, viz., 100 c.c. of pure trebly distilled water was then evaporated to dryness, the dishes ignited, and again weighed. I was unable to detect any difference in their weight, and the experiment having been repeated once or twice with the same result, it was assumed that the water was pure. The hydrochloric acid was bought as the purest, but it was again distilled most carefully, and 100 c.c. evaporated as before, there being in this case also either no residue or one which was absolutely

inappreciable. With the ammonia, which was also redistilled, there was no residue, though in this case I only evaporated a much smaller quantity, namely 20 c.c. The alcohol, however, gave me at first a residue of 0.003 gram per 20 c.c. of alcohol taken, but after being again very carefully distilled, was found to be pure.

Having thus purified the materials and determined the errors, if any, introduced into the estimation by the reagents employed, it was now possible to make some determinations of the required atomic weight. As the same plan was adopted throughout, a short description of one experiment will suffice for all the others.

The two platinum dishes having been carefully cleaned were ignited, cooled in desiccators, and weighed against one another, the piece of platinum foil being adjusted until the difference between the two was reduced to an amount less than 1 mgrm., the exact difference being carefully noted. The finely ground ammonium dichromate was then placed in one dish, weighed, and the temperature and pressure noted; 10 c.c. of water were then added, and when the salt had dissolved, 10 c.c. of hydrochloric acid; then with great care, and in small quantities at a time, 10 c.c. alcohol were added, and the whole evaporated to complete dryness on the water-bath, the platinum dish serving as tare being also placed on the water-bath. The above treatment was again repeated, so that there should be no doubt as to the complete reduction of the salt. The residue was taken up with 10 c.c. of water, and then 2 c.c. of the pure ammonia solution were added, and the whole stirred with a platinum wire bent once at right angles. Another 10 c.c. of water were added, and 3 c.c. of ammonia, and the mass again thoroughly stirred with the wire, which was washed with 10 c.c. of water. The mass was then evaporated to complete dryness, placed in an air-bath, and heated at about 140° for five hours. The two dishes, with their loosely-fitting platinum lids, were next placed in a gas-muffle, the floor of which was covered with asbestos card, the dishes themselves resting on pipe-clay triangles. The temperature was slowly raised to redness, and maintained at that point for an hour; the evolution of ammonium chloride was perfectly under control, and I do not believe that a trace of the oxide was mechanically carried away. In the volatilisation of the ammonium chloride, it was, however, possible that a certain amount of decomposition might have taken place, and that a trace of chromium chloride might have been re-formed, for it was found that, if a mixture of the hydrate and ammonium chloride were heated in a long glass tube, closed at one end and contracted at the other, a considerable amount of the chloride was produced. It is true that the conditions obtaining in the two cases were very different, still as there was the chance that an error might be introduced, it was

necessary to provide against it. The ignited residue was, therefore, treated with water, and then a little ammonia added, and the whole again evaporated to dryness and ignited. The dishes were then removed from the muffle by means of a fork with two prongs made of brass wire and coated with platinum foil, and placed at once in desiccators, allowed to cool in the balance-room, and weighed, the temperature and pressure being noted. They were again heated for half an hour and weighed, and also a third time; if the last two weighings agreed, the result was accepted. In this way, throughout the whole process, there was no transference at any time of any of the material from one dish to another, no filtering, no handling of any of the apparatus employed, and no burning of filter-papers, the whole operation being of the simplest kind. The chromium oxide was always of a beautiful green colour, and its purity was tested by treating it with water and filtering. The filtrate was clear and colourless, and on evaporation gave no residue. There were in all eight estimations made, but the first one was spoiled by using too much ammonia for the precipitation, and the second through incompletely drying the residue in the air-oven before igniting in the muffle. The remaining six observations gave the subjoined results.

:

The assumptions made in the calculations were these:

grams.

1 litre of air at 0° C. and 760 mm. pressure weighs 1.293
Correction for 1 gram (NH,),Cr2O, at standard

I give the experimental results and the value for the atomic weight of chromium deduced from each of them.

From this it will be seen that my experiments give results corresponding to the lower limit for the atomic weight of this element, and further, make the value approximate very closely to a whole number. I do not know of any constant source of error in my experiments. The only point in which there might be, perhaps, a risk of loss, is in the ignition of the mixture of chromium hydrate and ammonium chloride, with the loss of a little of the chromium oxide as the ammonium chloride volatilises.

I have never had, however, any ground whatever for supposing that such was the case, and further, the platinum covers loosely fitted to the dishes would, I think, practically remove even the chance of any such source of loss. The appearance of the residue, also, did not lend itself to any supposition of this kind.

XXVIII.—The Decomposition of Carbon Disulphide by Shock. (A Lecture Experiment.)

By T. E. THORPE, F.R.S.

It is usually stated in the text-books that only one compound of carbon and sulphur is known, viz., carbon disulphide, but the investigations of Hermann, Guignet, Löw, Sidot, and Raab have rendered it certain that several sulphides of carbon exist. The accounts which have been given of the nature and composition of these substances are, however, somewhat conflicting, and this fact induced me, some little time since, to begin some experiments with a view of obtaining further evidence on the subject.

Löw obtained a sesquisulphide, CS3, by the action of sodiumamalgam on carbon disulphide, and Raab prepared a compound, CS2, which he termed pentacarbon sulphide, by the action of sodium alone on the disulphide. One difficulty attending the use of the amalgam consists in the necessity of removing the mercury, and this, in Löw's process, involves the employment of sulphuretted hydrogen; in the other method the sodium becomes gradually coated with a thin crust of the product, and the action ceases after a time. It occurred to me that possibly better results might be obtained by the use of the fluid alloy of potassium and sodium, which is very mobile, and which, it was hoped, could be readily detached from the incrusting mass by shaking, and thus cause fresh surfaces of the mixed metals to be exposed to the action of the disulphide. Accordingly, as a prelimi

nary experiment, I treated a small quantity of rectified carbon disulphide, dehydrated by phosphorus pentoxide, with a few grams of the fluid alloy made by squeezing the two metals together through a sodium-press. After standing for a few hours, considerable quantities of a yellowish-brown powder were seen to have incrusted the globules of the alkaline alloy; but on shaking the bottle to detach the crust, the contents exploded with a loud report, and my hand was coated with a black deposit apparently consisting of finely divided carbon.

Further experiments on the yellowish-brown powder showed, in fact, that it is highly explosive; on simply pressing a few particles of it with a glass rod, it detonates with even more violence than diamine diiodide (iodide of nitrogen).

On reflecting on the circumstances of the explosion, it seemed to me that the deposit of carbon on my hand was far larger than could possibly have come from the decomposition of the yellowish-brown powder itself. The only other substance present which could have furnished such a deposit was, of course, the carbon disulphide. Now, carbon disulphide is an endothermic compound: when formed from its elements, and as liquid, it absorbs, according to Thomsen, 19,610 calories. Hence it was not improbable that it would be found to behave like acetylene, cyanogen, nitric and nitrous oxides, the oxides of chlorine, &c., and experience decomposition by sudden and violent shock. On this supposition, the carbon disulphide had in all probability been resolved into its elements by the violent explosion of the relatively small quantity of the yellowish-brown powder, just as Berthelot has shown that acetylene and cyanogen may be so resolved by the explosion of mercuric fulminate. A very few trials proved that this was actually the case, and as the experiment forms an exceedingly easy and perfectly safe method of demonstrating the resolution of an endothermic compound by shock, as a class illustration, I venture to bring it under the notice of the Society.

The apparatus required consists simply of a thick glass tube about 600 mm. long and 15 mm. wide, fitted at one end with a caoutchouc cork through which pass two stout wires or thin rods. On the end of one wire is fixed a small brass or iron cup like that of an ordinary deflagrating spoon, and the other is so bent as to nearly touch (to within 2 or 3 mm.) the bottom of the cup in which is placed about 0-05 gram of mercuric fulminate. The cork is now fixed tightly into the tube, and a piece of paper, slightly longer than the tube, is moistened with carbon disulphide and placed within the tube, which is supported at an angle of about 45° by means of a clamp and retort stand. After a minute or so, the tube will be practically filled with the vapour of carbon bisulphide, and the paper may be withdrawn.

VOL. LV.

R