It is readily soluble in hot water, cold alcohol, benzene, chloroform, hot carbon bisulphide, or cold acetone, dissolves sparingly in light petroleum, and crystallises from carbon bisulphide in extremely welldeveloped prisms melting at 110°.

0.2529 gave 0.2864 AgBr. Br=48.19.

0.7

3

CH2Br2 COO⚫NH(CH), requires Br=47·81; acid = 71.5 per cent. Tribenzylammonium Tribromobenzoate.-When ethereal solutions of tribenzylamine and the acid are mixed and left for some time, a crystalline separation of the salt is obtained, which, on recrystallisation from benzene, forms short, brownish prisms. It is practically insoluble in water, but dissolves sparingly in light petroleum, moderately in carbon bisulphide, ether, or alcohol, and readily in acetone or chloroform. When heated, it appears to melt partially at about 95°, to resolidify, and then to melt at 142°.

0.3094 gave 0.2684 AgBr. Br=36.91.

CH2Br2 COO NH(CH), requires Br=37·15; acid=55.6 per cent. Phenylammonium Tribromobenzoate was obtained by triturating the acid with rather more than the required amount of aniline until a hard, solid mass was formed; this, after being well broken up, was washed with light petroleum, and crystallised from benzene. It dissolves readily in chloroform, acetone, or warm alcohol, somewhat sparingly in water, light petroleum, or ether, and is almost insoluble in carbon bisulphide. It crystallises from alcohol, light petroleum, or benzene in small, colourless plates, and melts at 141-142°.

0.2012 gave 0.2504 AgBr. Br=52.96.

CH2Br2 COO NH, CH, requires Br=53.09; acid=79.4 per cent.

Metabromophenylammonium Tribromobenzoate was prepared by mixing ethereal solutions of metabromaniline and the acid; it is sparingly soluble in boiling water, and separates on cooling in small, colourless needles. It crystallises from dilute alcohol in small plates, dissolves readily in alcohol, acetone, or boiling chloroform, but only sparingly in carbon bisulphide or light petroleum, and melts at 141°.

0.17 gave 0.2401 AgBr. Br=60.1.

CH2Br2 COO NH, CH,Br requires Br=60-26; acid = 67.6 per cent.

Salts of Metanitrobenzoic Acid.

The acid employed is obtained from Kahlbaum; it was crystallised twice from benzene, and melted at 141°.

Trimethylammonium Metanitrobenzoate was obtained by passing trimethylamine into an ethereal solution of metanitrobenzoic acid. It crystallises from water, alcohol, or benzene in colourless needles, melts at 129°, dissolves readily in cold alcohol, water, chloroform, acetone, or warm benzene, and is almost insoluble in carbon bisulphide or light petroleum.

0.2162 gave 22-8 c.c. of moist nitrogen at 10° and 752 mm. N = 12.5. 0.7 0.5154 acid. Acid=73.6.

NO2 CH COO NH(CH), requires N=12.39; acid = 73.9 per cent.

Phenylammonium Metanitrobenzoate.-Small crystals of this salt separate when ethereal solutions of aniline and the acid are mixed and left for about 3 hours. The salt is readily soluble in boiling water, from which it crystallises in long, feathery needles melting at 115°; it also dissolves readily in hot alcohol, chloroform, or cold acetone, but only sparingly in carbon bisulphide or light petroleum. 0.3211 gave 29.3 c.c.of moist nitrogen at 10° and 764 mm. N=10·98. 0.7 0.4469 acid. Acid = 63.8.

NO2 CH, COO NH, CH, requires N = 10-77; acid = 64.2 per cent.

Metabromophenylammonium Metanitrobenzoate is deposited when ethereal solutions of meta bromaniline and the acid are mixed and allowed to remain. It is somewhat unstable, and when boiled with benzene for some time or with carbon bisulphide, is decomposed and metanitrobenzoic acid crystallises from the solutions; it may, however, be crystallised from water, dilute alcohol, chloroform, or light petroleum, and then melts at 76-77°.

0.4 gave 28.7 c.c. of moist nitrogen at 14° and 748 mm. N=8.31. 0.7 0.3451 acid. Acid = 49.3. NO, CH, COO NH, CH,Br requires N = 8-26; acid = 49-26 per cent.

6 4

a-Naphthylammonium Metanitrobenzoate.-Solutions of a-naphthylamine and the acid in benzene were employed; when mixed, they turned red, and after remaining overnight deposited warty masses consisting of long, hair-like needles; on recrystallisation from benzene, the salt was obtained in slender, colourless, silky needles melting at 105-106°. It is readily soluble in cold acetone, alcohol, or chloroform, also in boiling water, but a slight odour of a-naphthylamine is observable, and on boiling with carbon bisulphide, it is decomposed, as, on cooling, crystals of metanitrobenzoic acid are obtained.

0.3 gave 23·4 c.c. of moist nitrogen at 13° and 758 mm. N=9.19. 0.7 0.3746 acid. Acid=53.5.

0.7 0.3236 a-naphthylamine. Base = 46.2.

NO, CH, COO NH, C1H, requires N = 9.03; acid = 53·9;

10

B-Naphthylammonium Metanitrobenzoate is obtained when ethereal solutions of B-naphthylamine and the acid are mixed and evaporated to dryness, and separates in long needles when benzene solutions of the amine and acid are mixed and left for a short time. It crystallises in long, pale yellow needles, melts at 119°, and is readily soluble in alcohol or cold acetone, moderately in boiling water, hot benzene, or chloroform, but only sparingly in carbon bisulphide or light petroleum. 0.3832 gave 30-3 c.c. of moist nitrogen at 14° and 748 mm. N = 9.15. 0.7 0.3788 acid. Acid=53.7.

2:4:5-Trimethylphenylammonium Metanitrobenzoate.-An immediate white precipitate of this salt is obtained when ethereal solutions of pseudocumidine and the acid are mixed. It crystallises from benzene in silky needles, or from boiling water in slender needles, melts at 140-141°, and is readily soluble in cold alcohol, acetone or chloroform, or hot benzene, but very sparingly in light petroleum or carbon bisulphide.

0.2736 gave 22 c.c. of moist nitrogen at 10° and 744 mm. N=9·42. 0.7 0.3858 acid. Acid=55.1.

NO, CH, COO NH, CH(CH), requires N=9-27; acid=55.3

per cent.

Metanitrophenylammonium Metanitrobenzoate is obtained in long, yellow needles when the mixed ethereal solutions of metanitraniline and the acid are allowed to evaporate. It melts at 88-89°, and is readily soluble in hot water, alcohol, benzene, chloroform, or acetone, but practically insoluble in carbon bisulphide or light petroleum.

0.25 gave 29.5 c.c. of moist nitrogen at 15° and 752 mm. N = 13.68. 0.7 0.3808 acid. Acid=54.4.

NO, CH, COO NH, CH, NO, requires N = 13.77; acid=54.75

2 6

per cent.

Salts could not be obtained by mixing ethereal solutions of the acid and of diethylaniline, or of tribenzylamine, and the residues left on evaporating the ether gave the theoretical amounts of acid and base on extraction with carbon bisulphide.

Salts of Benzoic Acid.

Trimethylammonium Benzoate.-Benzoic acid readily dissolved in an aqueous solution of trimethylamine, but the salt was not isolated. Tripropylammonium Benzoate.-Benzoic acid dissolved in warm tripropylamine; when allowed to cool, the solution separated into two layers, but after several days a homogeneous syrup was obtained, which did not solidify when kept over sulphuric acid for some weeks. That combination had taken place was indicated by the fact that after a short time the mixture ceased to smell of the amine.

Phenylammonium Benzoate.-Aniline and benzoic acid do not appear to combine when their ethereal solutions are mixed and evaporated. When equivalent quantities of the amine and acid are stirred in a test-tube with a thermometer, the temperature rises from 15-30° as the mixture turns solid, thus indicating that combination has probably taken place; the resulting salt is unstable, however, and is decomposed partially when crystallised from light petroleum. This decomposition can be avoided by employing an excess of aniline in preparing the salt, or by adding 5 per cent. of aniline to the solvent. From light petroleum containing aniline, the salt crystallises in long, colourless, silky needles melting at 90°; it dissolves with the greatest readiness in cold acetone, ether, alcohol, benzene, or carbon bisulphide, and is moderately soluble in warm water or light petroleum.

0.4996 gave 28.7 c.c. of moist nitrogen at 14° and 754 mm. N=6·51. 0.7 gave 0.3959 acid. Acid=56.6.

CH COO NH, CH, requires N = 6.7. Acid=56.7 per cent.

All attempts to isolate salts of tribenzylamine, metabromaniline, pseudocumidine, or diethylaniline with benzoic acid proved fruitless, and in each case the unaltered amine and acid were recovered in nearly theoretical quantities.

The investigations recorded in this and the preceding papers have been carried out by the aid of grants from the Government Grant Fund of the Royal Society and the Research Fund of the Chemical Society, and to both of these we wish to express our indebtedness.

UNIVERSITY COLLEGE,

NOTTINGHAM.

LVIII-A New Compound of Arsenic and Tellurium.

By E. C. SZARVASY, Ph.D., and C. MESSINGER, Ph.D.

In making an experimental investigation of the compounds which arsenic forms with the various elements of the oxygen group, a certain regularity in their properties was observed, which led to the discovery of a new telluride. The investigation, however, aimed chiefly at an elucidation of the behaviour of the compounds at high temperatures and the determination of their molecular weights by means of their vapour densities.

Up to the present time, the following compounds of arsenic with the elements of the oxygen group are known.

6

Of these, As40, is stable at 1560°, As,S2 begins to dissociate at 1200° (Szarvasy and Messinger, Ber., 1897, 30, 1344), and As,Se at about 1150° (Szarvasy, Ber., 1897, 30, 1247), the remaining oxide, sulphides, or selenides being dissociated at lower temperatures into the stable derivative of the series and oxygen, sulphur, or selenium, as the case may be.

If, now, the compounds which are most stable at high temperatures are selected from each group of those examined, we find, on comparison, a certain numerical regularity existing between their molecular weights. For simplicity's sake, the simple molecular formulæ, As,O,, As,S2, As,Se, may be considered, and on comparing the weights of the negative components contained in the molecular weights: 3048, 2864, Se79, the differences between them are 16 and 15 units respectively. By extending the principle, we are able to determine the formula of the tellurium compound next in the series by the following calculation: the mean value of the difference is 15.5; this, when added to 79, the preceding term, gives 94.5, a number which has to be divided by the atomic weight of tellurium in order to obtain the number of tellurium atoms capable of combining with two atoms of arsenic, 94.5/127=0·744. We have now to find the smallest round number the product of which, with 0·744, will give an even number of atoms. This minimum number is 4, 0756 x 4 = 2.976, consequently three atoms of tellurium will be in combination with As, × 4, or eight atoms of arsenic. Hence the formula of the telluride is As,Te, supposing it to be capable of existence.