of the terminal groups an explanation must be offered which is somewhat different from that which accounts for the production of Düll's compound.

Action of Silver Nitrate.-It has been already stated that the bromine in the original substance is quantitatively separated by precipitation with silver nitrate in alcoholic solution; in order to examine the nature of this solution, a weighed quantity of the substance was dissolved in dilute (50 per cent.) alcohol, and mixed with slightly more than the calculated amount of silver nitrate dissolved in the same solvent. After filtration, the liquid was treated with a few drops of barium chloride solution to remove excess of silver, and, being strongly acid, it was then shaken with excess of barium carbonate and again filtered. After being kept in a vacuum desiccator for a day or two, yellow drops began to separate, and collected as a heavy oil at the bottom of the vessel. When most of the supernatant liquid had evaporated, the residue was taken up with absolute alcohol, filtered from barium salts, and the alcohol again evaporated. The oil or syrup thus obtained was moderately soluble in cold water, and dissolved easily in alcohol or ether; the solutions readily reduce Fehling's solution and ammoniacal silver nitrate, and with aniline acetate paper give a bright orange to brick-red coloration. With thymol and a few drops of moderately strong sulphuric acid, a fine pink colour is obtained, and phloroglucinol, with acids, gives an intense brownish-red. An ethereal solution of hydrogen bromide gives, immediately, an intense purple colour.

On heating, the oil begins to show signs of boiling at about 80°, but the temperature continues to rise, and no fixed boiling point is indicated. A small quantity of liquid distils over at about 200°, and behaves generally like the unaltered substance.

If the aqueous solution is treated with the calculated quantity (1 mol.) of phenylhydrazine as acetate, the liquid at once turns milky, and a red oil slowly separates, but this shows no signs of solidification, even after being kept for some days in a vacuum desiccator, or after purification by dissolving in hot dilute alcohol, from which it separates on cooling.

On oxidising this substance with silver oxide in the manner already described, an acid is obtained which melts at 162-163°, and is evidently identical with that obtained from the original substance.

From these observations, it is evident that this product bears a close resemblance to Düll's hydroxymethylfurfuraldehyde, but differs from it in some important particulars-for example, in the formation of a liquid hydrazone, and the character of the hydroxymethylpyromucic acid obtained from it on oxidation.

Action of Water.-A solution of the substance in dilute alcohol is

neutral to litmus, but on boiling it with water it slowly dissolves and the solution becomes strongly acid. If it is boiled with water, and excess of barium carbonate added to the hot solution, the filtered liquid, on extraction with ether, gives a product which appears to be identical in every way with that obtained by the action of alcoholic silver nitrate.

It appears, therefore, that the substance very readily undergoes hydrolysis, giving hydrogen bromide and hydroxymethylfurfuraldehyde. This fact, and the ease with which the original substance reacts with silver nitrate, further support the view, expressed above, that the bromine occupies a position in the side chain rather than in the nucleus.

Action of Sulphurous Acid.—When powdered crystals of the purified substance are covered with water, and sulphur dioxide is passed into the mixture, the gas is rapidly absorbed for a considerable time, and on allowing the mixture to remain for some hours, with further saturation by the gas if necessary, the whole of the solid dissolves; this solution, when treated with excess of caustic alkali, gives a magnificent blue or plum colour. Ether, chloroform, or benzene extract this colour-giving substance from the solution, and leave it as a crystalline solid on evaporation.

For the isolation of this substance, the mixture, after saturation with sulphur dioxide, was allowed to stand for 24 hours or more, and then neutralised with barium carbonate. (For convenience, it is advisable, before neutralising, to remove the excess of sulphur dioxide and part of the liquid by distillation under reduced pressure). It is then extracted 5 or 6 times, or more, with one of the abovementioned solvents, the solution dried over calcium chloride, and the solvent distilled off, when a yellow or orange oil is left which generally sets to a crystalline mass on cooling, but sometimes remains in a superfused condition, and requires shaking, or the addition of a few drops of water, to promote crystallisation. It is then recrystallised from the smallest possible quantity of boiling water, from which it separates on cooling as a bulky yellow or orange, crystalline precipitate; under the microscope, this is seen to consist of long, transparent needles. After again crystallising from hot water or hot dilute alcohol, with the addition, if necessary, of animal charcoal, the colour is very pale golden-yellow. It is easily soluble in cold. acetone, chloroform, glacial acetic acid, or hot alcohol or benzene, and dissolves sparingly in cold alcohol or ether, or hot water. melts sharply at 116.5-117.5°, and when heated to a higher temperature, a considerable portion vaporises unchanged and condenses as a crystalline sublimate. It is also slightly volatile with steam. Its alcoholic or aqueous solutions give, with caustic potash or soda, a

It

magnificent blue colour, which in deep layers appears red or plumcoloured; this colour soon fades on standing. With hydrogen bromide in ether, it gives an intense but transient purple colour.

With aniline acetate, solutions of this product give immediately a beautiful, bright green colour; a mixture of aniline with the substance in alcoholic solution gives no colour, but on the addition of acetic, tartaric, or other organic acid, the green colour appears at once. Mineral acids appear only to produce a similar effect in certain states of dilution.

The substance, crystallised from hot water and dried at 100°, was analysed.

I. 0.1268 gave 0·2991 CO2 and 0·0455 H2O. C=64·33; H=3·98.

This result appeared to be anomalous, since it could not be reconciled to any C-formula, or multiple thereof, consequently several more analyses were made with specimens which were variously treated. The melting point, however, continued to be quite constant, and the results of analysis practically the same.

II. Recrystallised from hot water and then from hot alcohol and dried in a vacuum. M. p. 116.5-117°.

III. Recrystallised from hot water containing sulphur dioxide. Dried in a vacuum. M. p. 117°.

IV. Recrystallised from hot water and then from hot alcohol, both containing sulphur dioxide. Dried at 100°. M. p. 117.5°. II. Found, C=64.50; H=4.00 per cent.

8

The simplest formula which can satisfactorily explain these numbers appears to be C1HO4, which requires C = 64.70; H=3.92 per cent. The molecular weight was determined by Raoult's method, with acetic acid as solvent.

It must, therefore, be concluded that the formula is C,,H,O,, the molecular weight of which is 204.

The substance is a powerful reducing agent and acts decidedly, though feebly, as a photographic developer; it also reacts with phenylhydrazine, but does not give the aldehydic reaction with Schiff's reagent. It would be easy, of course, to make assumptions

as

3

to its constitution; it might, for example, be methylfuril, CH2O·CO.CO.C1H2O·CH ̧, but this is a mere conjecture, and the authors prefer to reserve the discussion of its nature and mode of formation, for a future communication, when further experiments are completed.

Oxidation with Chromic Acid.-If the original bromo-derivative is finely powdered, covered with dilute sulphuric, acid, and a solution of chromic acid is slowly added to the mixture, a violent action takes place with considerable rise of temperature. By using only 1 atomic proportion of oxygen and keeping the mixture cool, a crystalline acid was obtained which closely resembled the w-bromomethylpyromucic acid of Hill and Jennings. Its solutions gave no precipitate with silver nitrate in the cold, but after boiling the aqueous solution there was copious precipitation of silver bromide. The quantity of the acid obtained, however, was very small, the greater part of the substance remaining unchanged, so that it has not yet been fully examined. But if excess (5 atoms) of oxygen is added, and the temperature allowed to rise, another crystalline acid is obtained which is very sparingly soluble in water, and sublimes on heating.

These acids, together with several other interesting derivatives of the aldehyde, are being investigated.

The authors wish gratefully to acknowledge their indebtedness to the Government Grant Committee of the Royal Society, for the funds which have enabled them to carry out this investigation.

XLII.-A Reaction of some Phenolic Colouring Matters. By ARTHUR GEORGE PERKIN, F.R.S.E.

DURING the examination of certain hydroxyanthraquinone derivatives, it was observed that alcoholic solutions of their salts are completely decolorised only with difficulty by acetic acid, and this suggested the presence of some stable salt. On reversing the experiment, this was found to be the case, for alcoholic solutions of these colouring matters decompose potassium acetate, forming sparingly soluble, crystalline salts in which one hydrogen only had been replaced by the metal. As quercetin possessed a similar property, it was desirable to determine if this reaction was in any way general, and to assign, if possible, a reason for the marked acidity of these colouring matters. Salts of certain yellow colouring matters have already been described, which, at first sight, are analogous to those formed by the above method.

Examination, however, shows that, in the majority of cases, these have no value, for their percentage composition, calculated from the older formulæ assigned to them, will not agree with that required for any salt of the colouring matter having the correct molecular weight; a description of these is given in the experimental portion of this paper. Of the numerous colouring matters here studied, some few were obtained by purchase from Merck of Darmstadt, or prepared synthetically. The majority I have either isolated from their natural sources, or from the commercial preparations which are employed in this department.

The Alizarin Group.

Alizarin, as is well known, forms potassium, barium, calcium, and other derivatives in which two atoms of hydrogen are replaced by the metal, but no acid salts of this colouring matter have been hitherto described; the addition of alcoholic potassium acetate to a solution of alizarin in boiling absolute alcohol causes the separation of a crystalline precipitate, and this may be collected and washed with alcohol, in which it is but sparingly soluble.

0.9495, dried at 160°,* gave 0·2935 K2SO4. K=13·86.

C14H7O4K requires K = 14.02 per cent.

Monopotassium alizarin forms a glistening mass of violet-coloured needles, almost insoluble in cold water, a property distinguishing it from the soluble dipotassium salt, which it considerably resembles in appearance. That it could not be a compound of alizarin with potassium acetate analogous to that yielded by hesperitin (Trans., 1898, 73, 1035) was readily ascertained by a study of its decomposition with hydrochloric acid.

0-9453 gave 0.8190 CH2O4. Found, 86.62.

C14H7O4K requires C1HO1 = 86.33 per cent.

If it is suspended in cold water and the mixture heated to boiling, little is dissolved until the temperature is within a few degrees of boiling, when a violet solution is produced, which is redder in tint than that of the dipotassium salt. In this way, no apparent decomposition of the salt ensues, which, as will later appear, is characteristic only of this group. Alizarin remains almost unaffected on digestion with aqueous potassium acetate, and even in sealed tubes at 160° it is but imperceptibly tinged violet. As the presence of water is inimical to the formation of this and the following salts, absolute alcohol has been employed in every case, for not only is the yield * Although the employment of absolute alcohol should preclude the presence of water of crystallisation in these salts, in each case in this paper, to avoid mistake, they have been dried at this temperature before analysis.