colour, even after frequent recrystallisation and the use of animal charcoal. Phenoxyfumaric acid melts and decomposes at 215°, becoming dark a few degrees below this. On analysis:

0-2000 gave 0.4237 CO, and 0.0727 H,O. C-57-74; H=4.03. C10HO requires C=57·69; H= 3.84 per cent.

8 5

On adding silver nitrate to a solution of the acid in ammonia, a white silver salt is precipitated; this can be dried at 100°, but decomposes suddenly on heating to a higher temperature. On analysis:

0.2630 left, on ignition, 0·1338 Ag. Ag=50.87.

C10H6O, Ag2 requires Ag=51·18 per cent.

Phenoxyfumaramide, CONH2∙C(O·C ̧H2):CH·CONH2, separates as a white solid on allowing ethyl phenoxyfumarate to remain in contact with strong aqueous ammonia for 2 days. It is sparingly soluble in alcohol, but dissolves in a large quantity of boiling water, and crystallises in small, colourless prisms which melt at 235° to a brown liquid. On analysis:

0.1965 gave 24 c.c. moist nitrogen at 25° and 760 mm. N=13-61. C10H10O3N2 requires N = 13.59 per cent.

Phenoxymaleic acid, CO2H-C(O.C,H,):CH CO2H.-If phenoxyfumaric acid is heated in a vacuum, it partially decomposes and evolves carbon dioxide, but a portion distils over at 197° under about 10 mm. pressure; the distillate is an oil which solidifies on cooling. This substance is undoubtedly the anhydride of phenoxymaleic acid; it dissolves in hot water, and the solution, on cooling, deposits colourless needles which melt at 168°, dissolve freely in alcohol or ether, and are slightly soluble in cold water. On analysis:

0.2007 gave 0.4220 CO, and 0-0700 H2O. C-57-34; H=3·87. 0.1994 0.4220 CO2 0.0723 H2O. C=57·71; H= 4·02. C10HO requires C=57·69; H=3.84 per cent.

8

On titration: 0.2080 gram was neutralised by 20 c.c. of N/10 sodium hydroxide, which is the calculated amount for a dibasic acid of the formula C1H,O,. In this and the subsequent titrations, the imide of dicinnamylphenylazimide, obtained by one of us (Trans., 1892, 61, 278), was used as indicator instead of litmus, which proved to be unsatisfactory.

Phenoxymaleic acid, as previously mentioned (p. 1120), is distinguished from its stereoisomeride, not only by its melting point and the colour, but also by its behaviour towards lead acetate, since in aqueous solution it gives with the acetate colourless needles of the lead salt.

The silver salt was obtained by adding silver nitrate to the solution of the acid in ammonia; it explodes on heating.

Action of Sodium Phenolate on Ethyl Chlorofumarate.

Although the behaviour of the ester, which is formed by the union of sodium phenolate with ethyl acetylenedicarboxylate, characterises it as ethyl phenoxyfumarate, yet we have thought it advisable to supply further evidence by a comparison of this ester with the product of the action of the phenolate on ethyl chlorofumarate.

In order to effect this reaction, it is best to use the method which served for the formation of ethyl phenoxyfumarate from ethyl acetylenedicarboxylate, that is, by adding ethyl chlorofumarate to the hot solution of sodium in an excess of phenol, and isolating the ester in the manner previously described. Its identity with the compound formed from acetylenedicarboxylic acid was ascertained from the boiling point and the following analysis:

2

0.1912 gave 0.4449 CO, and 0-1058 H2O. C 63.46; H-6.14. CHO, requires C=63.63; H=6.06 per cent.

16

The properties of the acid which is formed on hydrolysis of this ester agree in every respect with those stated above for phenoxyfumaric acid. It composition was, moreover, verified by the following titration:

0.2077 gram was neutralised by 20 c.c. of N/10 sodium hydroxide, instead of 19.975 c.c. as required for a dibasic acid of the formula C10H805.

We have also distilled this acid in a vacuum and found that the distillate, on recrystallisation from water, yields phenoxymaleic acid, identical with the specimen which we obtained from the product of the union of sodium phenolate with ethyl acetylenedicarboxylate.

These experiments prove that, the esters which are formed by the interaction of sodium phenolate with ethyl acetylenedicarboxylate on the one hand and ethyl chlorofumarate on the other, yield the same acid, namely, phenoxyfumaric acid, on hydrolysis. This fact leads to the conclusion that the esters also are identical. This view is supported by the identity of the boiling points, and the fact that on hydrolysis of the ester a transformation of the maleoid into the fumaroid form does not take place, since phenoxymaleic acid on heating with alcoholic potash on the water-bath remains unchanged.

Ethyl o-Cresoxyfumarate, CO2CH·C(O·C ̧H ̧·CH2):CH·CO2C2H ̧. Sodium o-cresolate reacts with ethyl acetylenedicarboxylate, forming an ester which is isolated in the same manner as in the previous cases. It is a colourless oil which boils at 184-185° under 14 mm. pressure; its density d 26°/26° is 1.1137. On analysis:

0-1980 gave 0.4681 CO, and 0·1170 H2O. C-64-47; H=6·56. C15H1805 requires C = 64.74; H= 6.47 per cent.

On hydrolysis of the ester with alcoholic potash, the acid CO2H.C(O.CH, CH,):CH CO2H is formed; this is readily soluble in ether or alcohol, and sparingly so in boiling water. From the last solvent, the acid crystallises out as yellow spherules which melt at 222° with an evolution of gas. On analysis:

0.1986 gave 0.4332 CO2 and 0·0790 H2O. C=59·48; H=4·41. CH10O5 requires C=59.46; H=4.50 per cent.

Both compounds, the ester and the acid, belong to the fumaroid series, being therefore ethyl o-cresoxy fumarate and o-cresoxyfumaric acid, since the same substances are formed on using ethyl acetylenedicarboxylate and ethyl chlorofumarate (as shown by the identity of their boiling points and densities).

The composition of the ester obtained by the interaction of sodium cresolate and ethyl chlorofumarate was verified by the following analysis:

0.1987 gave 0.4698 CO2 and 0·1190 H2O. C=64·48; H=6·65. C15H1805 requires C = 64.74; H=6.47 per cent.

The acid derived from this ester gave on titration the following results: 0.2056 gram was neutralised by 18.5 c.c. of N/10 sodium hydroxide, instead of by 18.52 c.c. as required for a dibasic acid of the formula CH1005

Ethyl m-Cresoxyfumarate, CO2C2H5°C(O·C2H4 CH ̧):CH CO2C2H ̧.

Since it was found that the same products are formed, whether sodium phenolates react with ethyl acetylenedicarboxylate or with ethyl chlorofumarate, the latter only has been employed for the preparation of ethyl m-cresoxyfumarate. The ester distils at 192° under a pressure of 14 mm., and its density d 26°/26° is 1·1115. On analysis :

0.1850 gave 0.4374 CO2 and 0·1090 H2O. C=64·48; H=6·54. C15H1805 requires C=64.74; H = 6.47 per cent.

m-Cresoxyfumaric acid is obtained by hydrolysis of the ester with

alcoholic potash; it readily dissolves in alcohol and ether, but only sparingly in water. After recrystallisation from a hot aqueous solution, it is obtained in the form of yellowish groups of needles which melt at 240° with an evolution of gas. On analysis :

0.1905 gave 0.4135 CO2 and 0.0778 H2O. C-59.20; H=4.53. C11H10O5 requires C=59.46; H=4.50 per cent.

0.1187 gram was neutralised by 107 c.c. of N/10 sodium hydroxide, the amount calculated for the substance C11H1005 being 10.69 c.c.

m-Cresoxymaleic Acid.-On heating m-cresoxyfumaric acid in a vacuum, it partially decomposes and yields a distillate which crystallises from boiling water in colourless needles melting (not sharply) at 208°. On titration:

0.0765 gram was neutralised by 7 c.c. of N/10 sodium hydroxide, instead of 6.89 c.c. as required for a dibasic acid of the formula C1H1005.

11

Lead acetate gives with a hot aqueous solution of m-cresoxymaleic acid a lead salt which gradually separates in groups of needles. On the other hand, no precipitate is formed with m-cresoxyfumaric acid. These stereoisomeric acids therefore resemble phenoxymaleic and phenoxyfumaric acids respectively in their behaviour towards lead

acetate.

Ethyl p-Cresoxyfumarate, CO2CH·C(O·C2H1·CH ̧):CH·CO2C2H ̧· Sodium p-cresolate yields with ethyl chlorofumarate the ethyl ester of p-cresoxyfumaric acid, which is isolated from the product of the reaction in the usual way. It distils at 191-192° under 12 mm. pressure, and its density d 26°/26° is 1.1132. On analysis:

0.1981 gave 0.4698 CO2 and 0.1186 H2O.

C-64.67; H=6.65. C15H1805 requires C-64.74; H=6.47 per cent.

In conclusion, it may be mentioned that the research on the additive compounds of the esters of the acetylene series on the lines indicated in this paper is still being pursued.

GONVILLE AND CAIUS COLLEGE,

CAMBRIDGE.

CIII.-Vapour Pressures, Specific Volumes, and Critical Constants of Diisopropyl and Diisobutyl.

By SYDNEY YOUNG, D.Sc., F.R.S., and EMILY C. FORTEY, B.Sc. A COMPARISON of the constants for normal and isopentane (Trans., 1897, 71, 446; Proc. Phys. Soc., 1895, 13, 602) on the one hand, and for methyl butyrate and isobutyrate (Trans., 1893, 63, 1191) on the other, seemed to show that the deviations from the generalisations of van der Waals in the case of isomeric substances are intimately connected with their constitution.

In order to obtain further light on this point, it was decided to prepare specimens of diisopropyl and diisobutyl, so as to compare them with the corresponding normal paraffins.

Preparation of the Paraffins.

Specimens of these substances were first prepared by the action of sodium on isopropyl iodide and isobutyl bromide respectively in ethereal solution. No great difficulty was experienced in the preparation of diisobutyl, and the yield was fair. As the boiling point of the paraffin (109-20) is considerably higher than that of isobutyl bromide (92.3°), it could be separated fairly completely from the unaltered bromide by fractional distillation. The final purification was effected by treatment with a mixture of nitric and sulphuric acids, and subsequent fractional distillation through a twelve-column Young and Thomas dephlegmator. Owing to the partial conversion of isobutyl bromide into the tertiary bromide, a small quantity of hexamethylethane is formed, but this is completely removed during the fractional distillation.

It is impossible to prepare a pure specimen of diisopropyl from isopropyl bromide, as the boiling points are almost the same, and on treatment of the mixture with nitric and sulphuric acids, bromine is set free, and at once acts on the paraffin.

Even from isopropyl iodide, the preparation is difficult and the yield very poor. Thus Zander (Annalen, 1882, 214, 167) obtained only 5 per cent. of the theoretical yield of diisopropyl as against 24 per cent. of normal hexane from propyl iodide.

As has been pointed out by Silva (Ber., 1872, 5, 984), sodium has little action on a dry ethereal solution of isopropyl iodide, and a little water has therefore to be added. In this case, however, propylene is evolved in large quantity, and naturally carries away a good deal