9. Effect of increasing the number of Sections.-In order to find what improvement is effected by increasing the number of sections, distillations of 100 grams of the mixture were carried out with Young, and Thomas dephlegmators with 3, 6, 12, and 18 sections. Finally, to see how complete a separation could be effected with the dephlegmator with 18 sections, a similar distillation was carried out at the rate of 30 instead of 60 drops a minute. The results are given in Table V, and for the sake of comparison the results of the distillation with the short plain still-head are added.

The improvement is perhaps better seen in the diagram, Fig. 8,

where the total weights of distillate are plotted against the temperature. In this diagram, curves are also given for the distillation with the "evaporator" still-head with 5 sections as well as with the short plain still-head. The dotted line shows the form that the curves would take if the separation were absolutely perfect.

* Mean percentage calculated from values with 400, 50, and 25 grams.

In conclusion, I would say that, when only a moderate degree of efficiency is required, comparable with that of the ordinary 3-bulb Glinsky or Le Bel-Henninger dephlegmators, I would recommend either a "rod and disc" or a "pear" still-head of about the size 3 B

VOL. LXXV.

described in this paper, in preference to either of them, or, indeed, to the Young and Thomas dephlegmator, on the grounds (1) that they are simpler, (2) that they are better suited for the distillation of small quantities, (3) that after the distillation is stopped the liquid returns almost completely from the still-head to the still. Of the two, I would prefer the " pear" still-head. A "rod and disc" or "pear still-head of efficiency equal to that of the 12 or 18 column Young and Thomas dephlegmator would, however, be of altogether excessive length; therefore, if great efficiency is required, I would recommend either a Young and Thomas dephlegmator or an "evaporator" stillhead. Of the two, the "evaporator" is superior, inasmuch as the amount of liquid in the still-head is considerably less, and also, after the distillation is over, the liquid in the still-head returns almost completely to the still.

UNIVERSITY COLLEGE,

BRISTOL.

LXVI. The Salts of Dimethylpyrone, and the
Quadrivalence of Oxygen.

By J. N. COLLIE, Ph.D., F.R.S., and THOMAS TICKLE, Salters' Company's Research Fellow in the Research Laboratory of the Pharmaceutical Society of Great Britain.

SOME years ago, one of us (Trans., 1891, 59, 617) noticed that when dehydracetic acid was boiled with fuming hydrochloric acid, a crystalline compound containing hydrogen chloride was obtained. This substance at the time was considered to be a compound of diacetylacetone, as it was formed from the acid by addition of water and subsequent elimination of carbon dioxide:

C2H2O4 + H2O+ HCl = C,H12O2Cl + CO2.

8

11

Feist (Ber., 1892, 25, 1067) pointed out later, however, that when fully dried over sulphuric acid this compound has the formula C,H,O,C1; still he also suggested that it was a derivative of diacetylacetone :

CH, CO CH2 CO-CH, CO-CH,, diacetylacetone.

3

CH, CCI.CH CO CH2 CO CH2, compound (m. p. 154°).

The properties of the substance, however, are somewhat anomalous. It does not give a red coloration with ferric chloride, a reaction which is very characteristic of diacetylacetone; moreover, when dissolved in water, it reacts exactly as if it were a mixture of dimethylpyrone and hydrogen chloride. That a compound having the formula suggested

by Feist, with chlorine directly bound to a carbon atom, should be so easily decomposed by dissolution in water seemed unlikely, and for this as well as other reasons, we thought it worth while to investigate the substance further. It was therefore prepared from dehydracetic acid in considerable quantity, and recrystallised from dilute hydrochloric acid, from which it separated in transparent plates melting at 84-85°. In aqueous solution, the compound is strongly acid, and the amount of hydrochloric acid present can be determined by titration with alkali of known strength. As already pointed out, the solution probably does not contain diacetylacetone, and this observation is further confirmed by the fact that barium hydroxide solution gives no yellow precipitate when added in excess. Moreover, the aqueous solution, when shaken with chloroform, yields dimethylpyrone, and, like a solution of this substance, is decolorised by bromine water, a milkiness being produced owing to the separation of one of the bromacetones.

The statement in the former paper (loc. cit., 620), that dimethylpyrone did not combine with hydrogen chloride, has been found to be incorrect, for both diacetylacetone and dimethylpyrone yield this additive product when allowed to evaporate with excess of hydrochloric acid. The substance has already been analysed by one of the authors, and partly by Feist, but further analyses were thought necessary to determine whether the hydrogen chloride was combined in the same manner as water of crystallisation, and whether the compound was one of constant composition. When dried in the air, it gave the following numbers on analysis, the water being determined. by leaving the substance in a desiccator over sulphuric acid:

Found, HCl = 184; H20=18.3.

C2H2O,,HCl + 2H,O requires HCl= 18.5; H,0-18.3 per cent.

The salt was not changed in composition by recrystallising four times from water, but when it was left for some days slight decomposition occurred, and the amount of hydrochloric acid fell to 18.1, and finally to 17.9, per cent. In the same manner, the salt, when left in a desiccator for a month, lost a little more than the theoretical amount of water (loss = 20.1; 2H2O= 18.3 per cent.), and the amount of hydrochloric acid in the dried salt was found to be proportionately low : Found HCl = 20.8. CH ̧O2,HCl requires HCl = 22.1 per cent. The dehydrated salt has been described by Feist. It is very hygroscopic; if left in the air, it rapidly absorbs water, passing first into the monohydrated salt (loc. cit., 619), and after a considerable time finally returns to the fully hydrated condition.

From these results, it appears that this curious compound behaves

much as an easily dissociated salt of a strong acid and a feeble base might be expected to do, and that the basic substance is dimethylpyrone. As it seemed somewhat remarkable that dimethylpyrone should possess such properties, a series of experiments was made in the hope that other salts than the hydrochloride might be isolated. Preliminary experiments soon showed that dimethylpyrone formed crystalline compounds with several other acids. The method employed when the acids, such as nitric or hydrobromic acids, were volatile, was to add an excess of the acid to the dimethylpyrone in an evaporating basin and allow the solution to evaporate in a vacuum over solid caustic soda; the crystalline separation was then collected and also dried in a vacuum over caustic soda. By this method, the hydrobromide, hydriodide, and nitrate were prepared. The other salts examined were the oxalate, tartrate, salicylate, chloracetate, picrate, and platinochloride. Several attempts to prepare the acetate and the sulphate were made, but without success.

8

Dimethylpyrone Hydrobromide, CH,O2, HBr, was prepared by evaporating a solution of dimethylpyrone in fuming hydrobromic acid in a vacuum over solid caustic soda. It crystallises in thin, transparent, slightly deliquescent plates, which were dried on a porous tile and left in a vacuum over caustic soda. The hydrogen bromide present was estimated by titration of the aqueous solution with decinormal soda and also with decinormal silver nitrate.

Found, HBr=394; Br=39.0.

CHO,HBr requires HBr = 39.5; Br= 390 per cent.

Dimethylpyrone Hydriodide, C,H,O,HI, formed in a manner exactly similar to that employed in the preparation of the hydrobromide, crystallises in long needles and decomposes somewhat in air into iodine and other substances. When freshly prepared, quickly dried on porous porcelain, and then in a vacuum over caustic soda, it gave the following result, when titrated with decinormal silver nitrate: Found, I=51.5. CHO,HI requires I=50.8 per cent.

Dimethylpyrone Platinochloride, (C‚H ̧O2)2,H,PtCl ̧.—When moderately strong solutions of dimethylpyrone hydrochloride and platinic chloride are added to one another, a copious precipitate of this salt is at once obtained, and its formation and appearance forcibly recall the action of platinic chloride on the hydrochlorides of bases such as those of the pyridine series. The platinochloride was recrystallised from water, and after drying did not lose weight when left for a considerable time in a vacuum desiccator.

Found, Pt=29.4 and 29.5. C=26·1; H=3·0.

(CHO),H,PtCl requires Pt=29-6; C-25·5; H= 2.7 per cent,