slightly soluble in chloroform, but dissolved readily in warm benzene, from which it was precipitated by light petroleum. It commenced to melt at 115°, although it was not completely fused until some degrees higher. Possibly the substance was a mixture of the allo- and orthocompounds.

0.2243 gave 8.6 c.c. moist nitrogen at 15° and 767 mm. N = 4.4. C16H19NO requires N = 4.3 per cent.

Succinic anhydride heated with the sodium derivative of orthonitrophenol also gave no reaction. Probably the method just described may be generally applied to nitrophenols to distinguish the metafrom the ortho- and para-compounds.

14

6

Action of Camphoric Anhydride on Sodium Bromophenates. Parabromophenyl Hydrogen Camphorate, COOH CH11*COO·C ̧H1Bг. -Sodium readily dissolves in a warm solution of parabromophenol in xylene, forming a white derivative; this goes into solution on adding camphoric anhydride, and the mixture deposits sodium bromophenyl camphorate on standing. On acidifying the aqueous solution of this salt with mineral acids, the acid separates almost immediately in a crystalline form. It is very readily soluble in chloroform, and when precipitated from solution in this solvent by light petroleum, separates in radiating masses of hard, thick needles, which are not perfectly white and melt at 111°.

0.2421 gave 0.1303 AgBr. Br=22.9.

C16H19Bro, requires Br=22.6 per cent.

4

14

8

2:4-Dibromophenyl Hydrogen Camphorate, COOH CH14 COO·CH.Br. -2: 4-Dibromophenol was prepared by the action of bromine on a solution of phenol in carbon bisulphide in the presence of small quantities of aluminium chloride. The carbon bisulphide solution of the bromophenol thus formed was washed, first with water, then with dilute alkali, and dried; after removing the carbon bisulphide by evapora. tion, the residue was fractionally distilled in a vacuum.

From the pure dibromo-compound thus obtained, 2 : 4-dibromophenyl hydrogen camphorate was prepared in exactly the same way as the corresponding monobromo-derivative. It very readily crystallises, and is easily soluble in hot, although somewhat sparingly in cold, chloroform, and separates in the form of dendritic masses of small crystals melting at 173°.

01711 gave 01467 AgBr.

Br=36.5.

C16H18Br2O, requires Br=36.8 per cent,

2

Action of Camphoric Anhydride on Sodium 2: 6-Dibromophenate.

2: 6-Dibromophenol was obtained in small quantity by distilling tetrabromophenolphthalein with concentrated sulphuric acid (Baeyer, Annalen, 1880, 202, 138); the yield, however, was extremely small. It was finally prepared from Fischer and Hepp's diorthobromamidophenol (Ber., 1888, 21, 674).

This was dissolved in four times its weight of alcohol, and to the solution rather more than the theoretical quantity of ethylic nitrite was added in 16 per cent. alcoholic solution; after some time, the whole was dropped gradually into an equal bulk of boiling alcohol, and kept in the state of constant ebullition for about one hour. The alcohol was then distilled off, and the residual oil distilled in steam. The dibromophenol thus obtained, being somewhat coloured, was purified by fractional distillation in a vacuum, and then had a constant boiling point not far removed from that of the 2 : 4-dibromo-compound.

2: 6-Dibromophenol did not react with camphoric anhydride, either when treated under the same conditions as the 2 : 4-bromo-compound, or when the mixture was heated in sealed tubes at 150°.

The 24: 6-tribromophenol was prepared in the usual way, and recrystallised several times from aqueous alcohol. It reacted with the anhydride neither at the temperature of a water-bath nor at 180° in sealed tubes.

Action of Camphoric Anhydride on Sodium Nitrosophenates.

Sodium derivatives were made from nitrosophenol and nitrosothymol by the action of sodium ethoxide on the alcoholic solutions and precipitation by ether. The sodium derivative of nitrosothymol prepared in this way did not separate immediately but only after some little time; it crystallised in dichroic golden plates reflecting a green light, and when dried in mass formed a brick-red powder,

When warmed with a 10 per cent. solution of camphoric anhydride in xylene on a water-bath, the sodium derivative of nitrosophenol charred before any reaction occurred. A similar result was obtained with nitrosothymol, although the latter, when left for a long time. with a more dilute solution of the anhydride in xylene, partially disappeared, forming a white sodium salt which was not further investigated.

THE WELLCOME CHEMICAL RESEARCH LABORATORIES.

By JAMES J. DOBBIE, D.Sc., M. A., and ALEXANDER LAUDER.

IN a former paper (Trans., 1895, 67, 17), we gave an account of the separation from the products of oxidation of corydaline with potassium permanganate of a compound having the composition CH,NO(OCH3)2 which we provisionally named corydaldine. By altering the conditions of oxidation, we have obtained a larger yield of this substance, and have examined its reactions and decomposition products more fully. In the paper referred to, we gave analyses of corydaldine together with the results of the determination of the methoxy-groups, and we showed that this substance has a neutral reaction in aqueous solution, and that it does not combine, at all events readily, with phenylhydrazine.

Freund and Josephi (Annalen, 1893, 277, 10), from the similarity in behaviour of methylcorydaline and hydrohydrastinine, and especially from the stability of their alkylic haloid addition compounds towards alkali, infer that corydaline, like hydrastine, contains a methyl group attached to the nitrogen atom. J. Herzig and H. Meyer (Monatsh., 1897, 18, 385), on the other hand, found that corydaline only contains four methyl groups in all, which can be split off by the action of hydrogen iodide, and we had previously shown that it contains four methoxy-groups. It follows, therefore, that methyl is not attached to the nitrogen atom in this alkaloid. We have arrived at the same conclusion by an entirely different method. By acting on corydaldine with nitrous acid, we obtained a highly characteristic nitrosoderivative, C11H2N2O4. When gently warmed with a dilute solution of sodium hydrate, this substance goes into solution, evolving nitrogen and forming a sodium salt, C1H1O,Na, which, on treatment with hydrochloric acid, yields the anhydride of the acid C11H1405. The anhydride C1H1204 contains two methoxy-groups. On oxidation with potassium permanganate, it yields metahemipinic acid, and when heated under pressure with dilute hydrochloric acid, it gives a phenolic compound which we believe to be identical with a-hydroxyethylcatecholcarboxylic anhydride, obtained by Perkin from berberine (Trans., 1890, 57, 1027).

11

12

Corydaldine is therefore closely allied to Perkin's w-amidoethylpiperonylcarboxylic anhydride, from which it differs only in containing two methoxy-groups in place of the piperonyl or CH,<O: group. Like Perkin's substance, it is a secondary base, and the derivatives of the one bear a very close resemblance to the correspond

2

ing derivatives of the other, the relation between the two series of compounds being shown by the following formulæ :

Corydaline is thus brought into close relation, not only with berberine, but also with hydrastine, narcotine, and papaverine, in common with which it contains an isoquinoline nucleus. From oxyhydrastinine, corydaldine only differs in containing two methoxygroups in place of the piperonyl group, and in having a hydrogen atom in place of a methyl group united with the nitrogen atom. That the arrangement of the groups attached to the benzene nucleus is the same in both substances, is confirmed by the oxidation on the one hand of the anhydride C11H1204 to metahemipinic acid, and on the other hand by the conversion of hydrastic acid into 4: 5-dihydroxyphthalic acid (normetahemipinic acid) by the action of phosphorus pentachloride and water.

Prof. W. H. Perkin has recently suggested to us that it might be possible to convert w-amidoethylpiperonylcarboxylic anhydride into corydaldine, by first preparing its phenolic derivative and then methylating, and he has very kindly furnished us with a quantity of the anhydride for the purpose of making the experiment.

In the paper already referred to, we gave an account of the separation of hemipinic acid [(COOH)2: (OCH3)2 = 1:2:3:4], from the products of oxidation of corydaline with potassium permanganate, and stated fully our reasons for considering the acid so obtained to be identical with the hemipinic acid from narcotine and berberine. In a later paper (Dobbie and Marsden, Trans., 1897, 71, 657), it was

stated that metahemipinic acid [(COOH),: (OCH3)2=1:2:4:5], identical with the hemipinic acid from papaverine, had also been obtained amongst the products of oxidation of corydic acid, CH17NO6 with potassium permanganate. We have now succeeded in separating metahemipinic acid, as well as hemipinic acid, from the products of the direct oxidation of corydaline with potassium permanganate.

The formation of metahemipinic acid by the oxidation of the anhydride, CH12O4, obtained from corydaldine shows that this acid is derived from the benzene ring of the isoquinoline nucleus. We thus obtain an important clue to the constitution of corydic acid which must also contain the isoquinoline nucleus, since it likewise yields metahemipinic acid on oxidation. It follows, therefore, that the ring which is destroyed when corydaline is oxidised with dilute nitric acid is the ring which yields hemipinic acid when the alkaloid is oxidised with potassium permanganate.

It is thus established that corydaline contains an isoquinoline nucleus and a benzene ring, to each of which two methoxy-groups are attached. We have therefore accounted satisfactorily for 19 of the 22 carbon atoms of the alkaloid.

In a former paper (Trans., 1894, 65, 57), we described one of the products of the oxidation of corydaline with a boiling solution of potassium permanganate, under the name of corydalinic acid. Had our original view of the nature of this substance been correct, a nitrogenous acid should have been found amongst its products of oxidation. As every attempt to obtain such an acid failed, we have made a complete reinvestigation of this substance, and find that it is, in reality, a sparingly soluble ammonium hydrogen metahemipinate, C10H1008C10H90, NH4+3H,0; the analyses of the substance agree perfectly with this formula, and those of its supposed salts with the formulæ of the corresponding metahemipinates. The error arose through the earlier oxidations with permanganate having been effected without the addition of alkali, preliminary experiments having led us to believe that the potassium of the permanganate was more than sufficient for the neutralisation of all the acid produced. In the absence of any reason for suspecting the presence of ammonia, the analyses of the supposed acid and its derivatives naturally led to the formula which we assigned to it; the sparing solubility of the substance in water and the strong acid reaction of its solution, moreover, seemed to confirm the belief that we were dealing with an acid. The discovery of the error has removed the inconsistencies which appeared to exist between the results obtained by oxidising corydaline with potassium permanganate and nitric acid respectively.