reaction took place, steam was given off, and the substance turned very dark. The dark brown product was boiled with water, with alcoholic sodium hydrate, again with water, then with alcohol alone, and finally with a small quantity of benzene. In this way, an almost white residue was obtained, which was recrystallised several times from solvent naphtha, and afterwards from benzene. It forms tufts of colourless, slender needles melting at 360°. During the purification, fluorescent solutions were obtained; but solutions of the pure substance are non-fluorescent. It is almost insoluble in alcohol, glacial acetic acid, and ethylic acetate; but chloroform dissolves it readily. Analysis gave figures agreeing with the formula of benzohexaphenyl trifurfuran, C2[C(CH2):Ç•C2H‚]. 0.1747 gave 0.5642 CO2 and 0·0751 H2O. C=88·08; H=4·78. 0.1488 0.4787 CO2 0·0642 H2O. C=87·74; H=4·79. C48H [3003 requires C=88·07; H=4.59 per cent. CHEMICAL DEpartment, UNIVERSITY OF ABERDeen. CIII.-Interaction of Benzoin with Phenylenediamines. By FRANCIS R. JAPP, F.R.S., and ANDREW N. MELDRUM, B.Sc. By heating benzoin with aniline and a little aniline hydrochloride, Japp and Murray (Trans., 1894, 65, 892) obtained 2': 3'-diphenylindole. The desylanilide, CH, CO-CH(NH C2H) C&H,, which is formed in the first instance, and which is the sole product if aniline alone is used, parts with water under the influence of the hydrochloric acid, yielding the indole. We have now studied this reaction with the phenylenediamines, in order, if possible, to obtain the corresponding benzodipyrrhole derivatives; but only in the case of metaphenylenediamine did the reaction take place in the desired sense. 1. Benzoin and Orthophenylenediamine. 6 By heating benzoin with orthophenylenediamine hydrochloride, NHphenylbenzimidazole hydrochloride, CHA >C.CH, HCl, was ob tained, the benzoin molecule breaking up in the process. The free phenylbenzimidazole melted at 285° (instead of 291°). Benzoin and free orthophenylenediamine, on the other hand, yielded N:C.CH 6 diphenylquinoxaline, C.HN:C-CH,, oxidation taking place during the process. The formation of the latter compound in this reaction had, as we afterwards found, been already observed by O. Fischer (Ber., 1891, 24, 720). 2. Benzoin and Metaphenylenediamine. Ten grams of metaphenylenediamine hydrochloride were dissolved in water, and the base was liberated by caustic soda and extracted with ether. The ethereal solution was dried with potassium carbonate and the ether distilled off. To the base thus obtained 23 grams of benzoin were added, and the mixture was heated to 180°. No reaction occurred; but on adding a little metaphenylenediamine hydrochloride, there was a vigorous effervescence; water was given off, and the mass solidified in spite of the high temperature. By extracting the melt with boiling alcohol and recrystallising the residue from solvent naphtha, the new compound was obtained in slender needles melting at 282°. It is almost insoluble in alcohol and benzene, soluble in chloroform. Dilute acids do not dissolve it; but it is soluble in concentrated sulphuric acid, giving a faint red solution with a greenish fluorescence. Analysis gave figures agreeing with the formula of a metabenzo 0.1806 gave 0.5858 CO2 and 0·0871 H2O. C-88·46; H=5·36. 0.2604 13.1 c.c. moist nitrogen at 14° and 751 mm. N = 5.85. C34H24N2 requires C=88.70; H=5·22; N=6·09 per cent. Two constitutions are possible for this compound, and at present there is apparently no means of deciding between them. The same compound may be obtained by heating together benzoin and metaphenylenediamine hydrochloride, thus obviating the necessity of preparing the free base; but the yield is bad, and the substance difficult to purify. 3. Benzoin and Paraphenylenediamine. The only method that yielded a definite product was that described in the case of the meta-compound: namely, of heating together benzoin and the free base, and adding a little of the hydrochloride. Twelve grams of paraphenylenediamine hydrochloride were taken; the base was liberated and extracted with ether, of which a large quantity was required on account of the sparing solubility of the para-compound. The base which remained after distilling off the ether was heated along with 23 grams of benzoin until the whole melted. No action took place; but on gradually adding 3 grams of paraphenylenediamine hydrochloride, the fused mass became pasty and ultimately solidified. The reaction was not so vigorous as in the case of the metacompound, and less water was given off. The product was boiled with alcohol, and the undissolved portion, which was yellow, was washed with boiling water, dried, and recrystallised from benzene. It was obtained pure by a single crystallisation, and formed bright yellow, minute lamina which, when heated, showed signs of softening about 230°, and melted completely at 257°. Analysis pointed to the formula CH2N2O2. 34 28 0.1846 gave 0.5588 CO2 and 0.0954 H2O. C=82·55; H=5·74. 0.2144 10.4 c.c. moist nitrogen at 12.5° and 743 mm. N = 5.62. C34H28N2O2 requires C=82·26; H=5·64; N=5.64 per cent. The compound is formed according to the equation C2H(NH2)2 + 2CH(OH)·CH ̧ Paraphenylenediamine. 6 5 == C2H‚[N] NH CH•CH 6 + 2H2O, and may receive the name didesylparaphenylenediamine. It corresponds with desylanilide (anilbenzoin) and the other compounds obtained by Voigt by heating benzoin with primary benzenoid amines (compare Trans., 1894, 65, 890). Diacetyl Derivative.-Two grams of the foregoing compound were boiled with excess of acetic anhydride for 10 minutes. The solution, which was light red at first and afterwards became darker, was allowed to cool, and ether was added. A white, crystalline substance separated, which was recrystallised from amylic alcohol, and was thus obtained in colourless, slender needles melting at 279°. Analysis gave figures agreeing with those required for didesylparaphenylene 6 5 diacetdiamine, C¿HN(C2H2O)·CH·C2H2 0.1217 gave 0.3507 CO2 and 0.0616 H2O. C=78.59; H=5·62. 0.1166 5.0 c.c. moist nitrogen at 14° and 758 mm. N=5.03. C38H32N2O4 requires C=78.62; H=5·52; N=4.83 per cent. 88 We tried to convert didesylparaphenylenediamine into the corresponding benzodipyrrhole derivative by heating it with zinc chloride. It is possible that this transformation does take place, but the product was a mixture of two substances (needles melting at 318° and prisms melting at 335°-both yellow coloured) of almost equal solubility; so that, with the small quantity at our disposal, we were unable to separate them. The foregoing five papers form a continuation of a general investigation of the reactions of ketonic compounds (compare Trans., 1897, 71, 123), and the expenses incidental to the work have for some years past been defrayed by various grants from the Government Grant Fund of the Royal Society. CHEMICAL Department, UNIVERSITY OF ABERDEEN. CIV.-A Series of Substituted Nitrogen Nitrogen Chlorides and their Relation to the Substitution of Halogen in Anilides and Anilines. By F. D. CHATTAWAY and K. J. P. ORTON. COMPOUNDS in which a halogen is attached to nitrogen have been little studied, although a number of isolated examples are known. Among such may be mentioned nitrogen chloride, nitrogen iodide, the chlorine derivatives of a few aliphatic amines (Tscherniak, Ber., 1876, 9, 143; Norton and Tscherniak, Bull. Soc. Chim., 1878, [ii], 30, 106), of acetanilide (Bender, Ber., 1886, 19, 2272), of benzamide (ibid.), of succinimide (ibid.), and of phenylnitramine (Bamberger, Ber., 1894, 27, 376), the bromine derivatives of the aliphatic amides (Hofmann, Ber., 1882, 15, 407 and 752), and of benzanilide (Linebarger, Amer. Chem. J., 1894, 16, 218), the iodine derivatives of succinimide (Bunge, Annalen Suppl., 1870, 7, 119), of acetamide (Seliwanow, Ber., 1893, 26, 985), and of formanilide (Comstock and Kleeberg, Amer. Chem. J., 1890, 12, 500), methylic and ethylic chlorimidocarbonates (Sandmeyer, Ber., 1886, 9, 862), and the chlorimide obtained from the oxime of benzophenone (Beckmann, Ber., 1886, 19, 988). Some years ago, the attention of one of us was drawn to a substance obtained by Witt (Ber., 1875, 8, 1226) by the action of hypochlorous acid on acetanilide, and this was employed in a simple process for preparing metadichlorobenzene in quantity (Chattaway and Evans, Trans., 1896, 69, 848). In continuation of our work upon nitrogen iodide, we recently took up the study of this compound, and have obtained a series of substituted nitrogen chlorides which undergo remarkable intramolecular transformations and are of extreme interest from their bearing on the theory of substitution. We find that, by the action of hypochlorous acid under carefully regulated conditions, compounds of this nature can be obtained from most substances containing hydrogen attached to nitrogen. In the present paper are considered the disubstitution products of nitrogen chloride containing formyl, acetyl, or benzoyl, together with a phenyl or chlorophenyl residue. These form a well-defined group of compounds, phenyl acetyl nitrogen chloride, discovered by Bender, being one of the simplest members. They are readily obtained from the corresponding formanilide, acetanilide, or benzanilide by interaction with hypochlorous acid. R.CO.NR'H+HOCI = R.CO·NR'Cl + H2O. They are all stable compounds of low melting point, crystallising well in large, colourless, transparent prisms or plates, and reacting very readily with alcohol, hydrochloric acid, hydrocyanic acid, potassium iodide, and alkaline hydrogen peroxide. (a) With alcohol, the anilide is reformed, and aldehyde, ethylic chloride, and other substances are liberated, ethylic hypochlorite, which breaks up into aldehyde and hydrochloric acid, being probably first produced. 2 5 R.CO.NR'Cl+C2H ̧·OH=R.CO.NR'H+C2H ̧•OCI. 2 (b) With strong hydrochloric acid, chlorine is set free and the anilide regenerated. R.CO.NR'CI+ HCl=R.CO.NR'H+Cl2. In some cases, however, under the influence of the acid, a portion of the compound undergoes an intramolecular transformation similar to that which takes place when it is heated. (c) With hydrocyanic acid, the anilide and cyanogen chloride are formed. R.CO.NR'CI+HCN = CNCI+R.CO.NR'H. (d) With an acid solution of potassium iodide, the anilide is reproduced and iodine liberated. R.CO.NR'Cl + 2HI = R•CO·NR'H + HCl + I2. (e) With alkaline hydrogen peroxide, oxygen is liberated and the anilide reformed. R.CO.NR'Cl + H2O2 = R·CO·NR'H + HCl + 02. 2 2 When warmed with various anilines, a vigorous reaction takes place, and the anilide is regenerated, whilst a chlorine substitution product of the aniline is formed. These compounds all undergo a remarkable isomeric change. When an unsubstituted phenyl residue is attached to the nitrogen, the |