some Fe2O3, H2O placed in the bend, and the tube again filled and re-weighed with dry hydrogen. Another similar U-tube, charged with powdered calcic chloride and weighed full of hydrogen, was connected with the outlet of the first U-tube. Sulphuretted hydrogen (containing 15 to 20 per cent. hydrogen) washed and dried through two drying cylinders charged with powdered calcic chloride, was now passed through the two U-tubes, the first, containing the Fe2O3, H2O, being immersed in a beaker of cold water to moderate the action. When the action appeared to be complete, that is to say, when the whole of the material in the first U-tube was quite black, the cold water in the beaker was changed for boiling water, and the current of hydrogen sulphide replaced by dry hydrogen to expel excess of hydrogen sulphide. The hot water in the beaker was changed for cold, and when sufficient time had elapsed to enable the first U-tube to become cold, the stopcocks were closed and the tubes weighed separately. The increase in weight of the two tubes was taken as the quantity of hydrogen sulphide which had entered into action. The increase in weight of the calcium chloride tube was taken as the 3 mols. of water formed by the reaction, and the 1 mol. expelled from the hydrated Fe2O,H2O which had entered into action, according to the equations(A.) Fe2O3,H2O Fe,S,+ 4H2O.

(B.) Fe2O,H,O= 2FeS + S + 4H2O.

The results I obtained are embodied in the following table which contains the following particulars:

(1.) The weight of Fe2O3,H2O operated upon.

(2.) The weight of above after treatment with hydrogen sulphide, and drying in a current of hydrogen at about 100°.

(3.) Difference between (1) and (2).

(4.) The water resulting from the reaction.

(5.) The weight of hydrogen sulphide entering in action obtained by adding (3) and (4).

(6.) The weight of hydrogen sulphide with which the Fe2O3, H2O was theoretically capable of reacting according to equations (A) and (B).

(7.) The percentage of Fe,O,H2O entering into action calculated from (5) and (6). 3H2S

(8.) The ratio

increase in weight of both tubes

or 4H2O' increase in weight of calcium chloride tube'

(9.) Sulphur calculated from (5).

(10.) Sulphur by analysis. This was not ascertained in all cases, for the products of the reaction were used for other inquiries.

(11.) Sulphur found in the free state by washing the product of reaction with carbon bisulphide in the manner usual in such cases, but without access of air.

I expected, or was justified in expecting, for the greatest possible

3H2S

care was observed in all the experiments. The average ratio, 4H2O' found, 1-438 is slightly higher than theory 1.417. In six experiments it is higher, in three experiments lower. This might be perhaps because the sulphides retained a little water, but I do not quite understand why this should be so in some cases and not in others.

Assuming that the free sulphur found was produced according to the equation (B), then it would appear that from 17 per cent. (Experiment 6) to 30 per cent. (Experiment 8) of the Fe2O,,H2O entering into action went to form ferrous sulphide, FeS, and that the remaining portion formed ferric sulphide, Fe2S., according to equation (A). The sulphides formed in the dry way by the action of H2S on Fe2O,,H2O, as in my experiments, as well as the sulphides prepared in the wet way by precipitation of neutral ferrous salts with hydrogen sulphide, or by adding a neutral ferric salt slowly to a solution of ammonium sulphydrate, are all soluble in hot potassium cyanide solution, with production of potassium ferrocyanide. They are completely decomposed by neutral cupric sulphate or silver nitrate. My experiments in this direction are at present too incomplete to be used in any way for the purpose of explaining the composition of the sulphides formed by means of H2S and Fe2O,H2O; but I think that the amount of water affords strong evidence that the reaction proceeds according to either one or both of the equations (A) and (B),

and that the amount of free sulphur indicates that both occur, the former (A) predominating, but they do not bear any definite relation to each other in amount, but vary in proportion to each other with slightly varying circumstances, which I have not determined, and upon which I can throw no light. However, I think it possible that further inquiry by means of the decompositions of the sulphides with cupric sulphate and silver nitrate may definitely prove the composition of the sulphide formed in the reaction between Fe2O3,H2O, and H.S.

The mixture of ferrous and ferric sulphides, as obtained in my experiments, is intensely pyrophoric, firing with the very first contact of air. On this account extraordinary precautions had to be observed, whilst the material was in the dry state; after it had been moistened with water, however, the oxidation was by no means so rapid.

The action of potassium cyanide on these iron sulphides is a useful means of distinguishing between the forms of ferric oxide active or inactive with sulphide, for any ferric oxide unacted upon by hydrogen sulphide or dilute potassium sulphide is left undissolved, and can be thus separated and estimated. To effect the solution of the ferrous and ferric sulphides it is necessary to boil them for about ten minutes with an excess of potassium cyanide. If, however, they be boiled with an insufficient amount of potassium cyanide, a curious action takes place. On filtering off the liquor from the undissolved sulphide, the filtrate runs through of a clear yellow colour; but on attempting to wash the undissolved sulphide with water, it immediately passes through the filter-paper in a similar condition to the "colloidal " ferric hydrate before spoken of. The filtrate is then thick and muddy, and quite black and opaque. On standing for some time, black flocks are deposited, and the solution partially clears. I have noticed the same appearance when cupric sulphide has been treated with an amount of potassium cyanide insufficient to effect its solution.

XXIII.-Note on some Derivatives of Fluorene, C13H10

By W. R. HODGKINSON and F. E. MATTHEWS.

THE hydrocarbon, isolated from coal-oil by Berthelot (Ann. Chim. Phys. [4], 12, 222), and termed fluorene from its supposed fluorescent property, has been investigated by several chemists, and its constitution as o-diphenylenemethane determined by its production from diphenylmethane by passing through a red-hot tube (Graebe, Annalen,

VOL. XLIII.

N

174, 194), and the reduction of diphenyleneketone by zinc-dust or phosphonium iodide (Fittig, Ber., 6, 187; Graebe, Ber., 7, 1625). Its constitution would therefore be expressed by—

CH2

More recently, however, Carnelley (C. J., 37, 708) has discussed the constitution of coal-tar fluorene, and the fluorene from diphenyleneketone, and notices that of the six possible isomerides of this body several occur in coal-oil, and several are also formed when the vapours of benzene and toluene are passed through a heated tube.

The object we had in view in commencing this investigation was in the first place to obtain the phenol or hydroxy-derivative, fluorol, C13H10O, as a starting point for further work.

As diphenylene ketone is somewhat difficult to obtain in quantity, and the yield of the hydrocarbon from it is by no means a quantitative one, we have employed coal-tar fluorene obtained from Kahlbaum in a state of approximate purity. For further purification it was distilled, and the fraction between 300-310° crystallised five or six times from alcohol, reserving only the crystals falling out between 25° and 30°. Thus purified, it melts constantly at 113°, and the slight fluorescence may be almost entirely removed by one or two crystallisations from glacial acetic acid, or by sublimation over potassium carbonate.

Glacial acetic acid is by far the best solvent from which to crystallise fluorene. By means of this solvent, it is possible to separate from fluorene boiling between 300-310°, hydrocarbons melting at 112113°; 124°; 118°; and, in small quantity only, at about 200°.

The one melting at 118°, and occurring in pretty large quantity in coal-tar fluorene, is most probably the 7-methylene-diphenylene of Carnelley. It certainly gave another body in addition to diphenyleneketone on oxidation; we have not, however, examined these substances further, as we desired only the ortho-modification.

The product we have employed oxidises in glacial acetic solution completely with chromic acid to diphenylene ketone, no quinone being produced.

On analysis 0.1106 gram gave 0.3796 gram CO, and 0.0645 gram OH2.

The a-dibromofluorene, C13H,Bra, was made by dissolving the hydrocarbon in chloroform, adding a slight excess of bromine, and leaving the chloroform to evaporate spontaneously. The compound is almost

insoluble in cold alcohol, but dissolves easily in boiling alcohol, from which solution it separates in almost colourless tabular crystals. From hot glacial acetic solution it forms perfectly colourless crystals melting at 165°. Light causes the crystals to become slightly yellow. A bromine determination by heating with Iceland spar gave

C13

49.71 per cent. Br. Theory H, = Br 49.38.

Br2

We did not notice any formation of the B- and -dibromofluorenes as noticed by Lehmann and Azruni; also by Fittig and Schmitz. Their formation may depend on the solvent employed.

Monobromofluorene, C1H,Br.

This substance is formed when bromine is very carefully added to a solution of the hydrocarbon in chloroform, care being taken to keep the temperature as low as possible. The product, even when the fluorene is maintained in excess, still contains the dibromo-compound, from which the mono-derivative may be separated by repeated solution in, and crystallisation from, about 90 per cent. alcohol. It forms colourless needles melting at 101-102°.

Analysis gave 32-65 per cent. bromine; the formula C,,H,Br requires 32.54 per cent.

This substance is exceedingly soluble in cold chloroform.

Both the mono- and di-bromofluorenes yield on oxidation a substituted diphenylene ketone. The product appears to be the same whether chromic acid in glacial acetic or permanganate is used.

Dibromodiphenylene ketone, C13HBг2O.

This is best obtained by oxidising the (165°)-a-dibromfluorene with chromic acid in acetic acid solution. It forms bright yellow microscopic crystals from acetic acid, melting at 198°.

Analysis gave 47.34 per cent. bromine; the formula requires 47.39.

The monobromfluorene, when carefully oxidised with the theoretical amount of chromic anhydride, yields bromdiphenylene ketone, C13H-BRO, in dark yellow needles melting at 104°.

The production of diphenylene ketone alone, by oxidation of the fluorene, and the formation of the dibromo-compound melting at 165°, was sufficiently conclusive that the hydrocarbon we are dealing with had the constitution assumed as that of ortho- or a-methylenediphenylene.