Al2S results. I cannot find any corroboration of this

statement.

Mr. Niewerth's process for reducing aluminium, in which he either uses A12S3 or else makes it as an intermediate product, will be found in full on p. 185, it being too long to repeat here. I cannot find any outside testimony as to the possibility of his schemes.

Reichel has probably proven the possibility of reducing Al'S by a metal having more affinity for sulphur. From a chemical standpoint its reduction by copper, iron, etc., should be under the proper conditions a very easy operation. These conclusions follow from the relative affinity of sulphur for the metals, which is set forth in the following investigation :—

"A. Orlowsky* has studied the affinity of sulphur for the metals. From his researches it was found that it usually possesses the greatest affinity for the alkaline metals, with which it forms polysulphides. Among the other metals, copper possesses the greatest affinity for sulphur, then follow in order mercury, silver, iron, lead, and after these platinum, chromium, aluminium, and magnesium, whose affinities for sulphur are quite insignificant."

EXPERIMENTS ON AL2S.

Taking the data given in the foregoing papers, I made a series of experiments on first making APS3 and then on reducing it.

Experiment I.

Took pure, white alumina, made by calcining pure sul

*Jahresb. der Chemie, 1881, p. 24.

phate of alumina, put it in porcelain boats in a hard glass tube, and passed vapor of carbon bisulphide, CS2, over it at bright redness for forty-five minutes. The product was cooled out of contact with the air. The result was a grayish-black powder, not sintered together in the least. On analyzing the product by Fremy's method, it showed 12.65 per cent. of A12S.

Experiment II.

Took equal parts of alumina, sulphur, and charcoal, ground intimately together in a mortar, and served as in Experiment I, prolonging the action of CS2 to an hour and a half. The product was a grayish-black powder, similar in appearance to the former product. It contained 38.51 per cent. AlS3.

Experiment III.

Repeated Experiment I, but used a porcelain tube, thus allowing a higher heat than the glass tube would stand. The treatment lasted an hour and a half. The product was of similar appearance to the previous ones, and contained 39.54 per cent. AS3.

Experiment IV.

I placed some ordinary aluminium sulphate, Al2(SO1)3.18H2O, in the porcelain tube, and heated it gradually up to bright redness with the tube open at both ends, calcining it thus for two hours. The result was that the tube was filled with very porous alumina. CS2 was then passed over it for two hours, the whole being kept at redness. The product was dirty-white, but lemon-yellow in places, and at the yellow parts sintered together. Analyzing an

average specimen, it showed 31.16 per cent, AlS3. It is probable that if a yellow piece had been singled out it would have shown much more A12S3 than this average sample.

Experiment V.

I placed some pure alumina in small hollows cut in pieces of charcoal, and placed these in the tube instead of the porcelain boats. The tube was then placed in an assay furnace and heated almost to whiteness for an hour and half, CS being passed through. The product was small, black, fused buttons melted down into the bottoms of the cavities in the charcoal. These lumps were black outside, brittle, compact fracture, and the broken surfaces mottled, dirty-white, and yellow. They had a strong smell of hydrogen sulphide, and when dropped into water this gas was evolved so actively as to make quite a buzz, resembling the action of a piece of zinc dropped into acid. one or two minutes the button was resolved into a black powder. This product contained 40.43 per cent. Al'S3.

Experiment VI.

In

Repeated Experiment V, but used porcelain boats. The product was still dark, and contained 38.80 per cent. Al2S3.

Experiment VII.

Wishing to make a quantity of the substance, I filled the tube with alumina, put it in a hot fire, and passed CS2 over it three hours. The product was grayish-black, with here and there touches of yellow, with lumps of considerable size sintered together. An average sample of it contained 32.32 per cent. Al'S.

Tabulating these results we have—

I.

Experiment
II. III. IV. V. VI. VII.
A12S3 (p. ct.) 12.65 38.51 39.54 31.16 40.43 38.80 32.32

First I would notice that, as remarked by Fremy, the A12S3 formed incloses the particles of alumina and prevents further action. It seems highly probable that a stirring apparatus to keep the alumina agitated would greatly improve the product. Experiment I gave poor results because the heat was not sufficient; Experiment II was done at a higher heat, with addition of carbon, and Experiment III at a still higher heat, without carbon. It appears from this that the presence of carbon had very little influence on the amount of Al2S3 produced. Experiment V, giving the best results, was worked, I think, at a higher heat than any of the others; but Experiment VI was conducted under as nearly as possible the same conditions; however, we may consider the products as being nearly enough alike, the carbon does not appear to have made a marked difference in the product.

To establish such a process on a practical scale, a wrought-iron or fire-clay retort would be necessary, with arrangements to heat it almost to whiteness. Boats of charcoal, holding ample charges of alumina, are made to fit in the retort. Some sort of stirring apparatus to agitate the alumina from time to time should be provided. The CS2 could be brought in superheated by waste heat from the furnace and passed out into a condenser. Or, to economize still further, the retort might be lengthened, its forepart made a producer of CS2, by passing sulphur vapor over carbon, and the rear part be filled with the alumina to utilize this CS2. Many other devices will occur to the

practical chemist in running such a process, the above being mere suggestions.

REDUCING THE AL'S3.

Experiment VIII.

I took about half a gramme of product of VII, and wrapping it tightly in lead-foil placed it on a cupel and heated in a muffle. Air was kept from the metal by a close-fitting porcelain cover. On removing the lid after a few minutes, there appeared a button of lead with some powder on its surface. I then cupelled the lead at as low a temperature as possible. The metal cupelled away entirely, leaving no aluminium. On repeating with every precaution the result was the same.

Experiment IX.

About one gramme of product VII was wrapped in copper foil, put in a porcelain crucible, and covered with NaCl and a little charcoal. A close cover was put on, the whole placed in the middle of a Hessian crucible, the latter filled up with fine charcoal, and a cover luted on. On heating this an hour at bright redness, hardly whiteness, there resulted a large button of copper. However, its specific gravity was that of pure copper, and a qualitative test showed no trace of aluminium. It occurs to me now that probably the NaCl reacted on the Al2S3, forming aluminium chloride and sodium sulphide, preventing the action of the copper.

Experiment X.

Repeated Experiment IX with tinfoil, and heating only twenty minutes. The tin resulting showed some alumin