There is one other possible course for the reaction between benzaldehyde and caustic soda, namely, that the two molecules of aldehyde This repre

condense directly to an unsaturated glycol:

CH, C.OH

6

sents a tautomeric form of benzoin, the dibenzoyl and diacetyl derivatives of which are known, but no indication of the formation of such a compound or of its derivatives was obtained. It could yield benzoic acid and benzylic alcohol very simply, thus:

but the formation of benzylic benzoate from it would not depend on the absence of water, or the presence of an excess of aldehyde.

EXPERIMENTAL.

Action of Caustic Soda on Benzaldehyde in Presence of Water. A number of experiments were first tried in which an aqueous solution of sodium hydroxide was added to benzaldehyde in the proportion of one molecule of the former to two of the latter. The precipitate that is formed immediately, increases in bulk rapidly on standing, and the reaction may be completed either by heating the mixture on the water-bath for 3 hours, or by allowing it to stand, with repeated shaking, for a day. To 53 grams of benzaldehyde, 10 grams of caustic soda dissolved in 15 grams of water were added. The final product consisted of a hard cake, which was dissolved in 75 c.c. of warm glacial acetic acid, and the solution diluted with water, made alkaline, and extracted with ether; the oil obtained by evaporation of the ethereal solution was then fractionated. The alkaline solution was acidified and the precipitated benzoic acid collected. Under these conditions, benzylic alcohol and sodium benzoate proved to be the sole products; it is a matter of indifference, therefore, whether water or glacial acetic acid is used in the decomposition of the final product of the reaction.

The yield of benzylic alcohol under these conditions is extremely good, 90 per cent. of the theoretical yield being obtained, and by increasing the proportion of sodium hydroxide to 20 grams, a theoretical yield (98 per cent.) resulted. This is consequently a better, as well as a cheaper, method for the preparation of benzylic alcohol than that usually employed.

Action of Solid Caustic Soda on Benzaldehyde.

10 grams of caustic soda were ground to a fine powder under benzene, washed into a flask with benzene, and a solution of 53 grams

of benzaldehyde in 100 c.c. of benzene added. During the addition, the mixture was well shaken; after a short time, the whole set to a hard mass. This was heated in a reflux apparatus on the water-bath for 2 hours, the mass being broken up and shaken several times during the heating. The solid residue was ground up in one case with ether, and in another with benzene, the extracts filtered, and the residues fractionated after distilling off the solvent, but in both cases only benzylic alcohol together with some unchanged benzaldehyde was left, the yield of the former amounting to 76 per cent. of the theoretical.

In these experiments, no special care had been taken to render the materials anhydrous, but the necessity of this condition became obvious when the structure of a possible intermediate product was taken into account. As shown on p. 1157, a molecular proportion of water is regenerated in the action; this, therefore, will suffice to effect the complete breaking down of the compound into benzylic alcohol and sodium benzoate. In the subsequent experiments, the caustic soda was first ground up under benzene to prevent undue absorption of moisture during the grinding, the greater part of the benzene drained off, and the rest removed by drying in an exsiccator. The final drying was effected in a vacuum over phosphoric oxide repeatedly renewed. Only small quantities of caustic soda, spread over a large surface, were dried at a time. The benzene employed as solvent was distilled over sulphuric acid and phosphoric oxide successively. periments were made with materials thus prepared. In both cases, the mixture became solid very soon, athough in one (No. II) the aldehyde was added very gradually, and the temperature kept down by means of ice previous to the heating on the water-bath. The product of the reaction was first dissolved in glacial acetic acid, the solution diluted with water, and then extracted with ether after the addition of alkali. The following are the details of these two experiments, 200 c.c. of benzene being used in each:

Two ex

The presence of unchanged benzaldehyde in both cases is due to the difficulty of effecting a complete reaction in the solid mass, although it was broken up and thoroughly mixed several times during each experiment.

Claisen obtained from 10 to 40 grams of benzylic benzoate from 106 grams of benzaldehyde by the action of sodium methoxide. It therefore appears that the methylic ether of the ortho-compound breaks up somewhat less readily than the simpler derivative into the alcohol and acid (methylic ester). The benzylic ether is still more stable in this respect.

The benzylic benzoate boiled at 280-320°, and was therefore accompanied by some benzylic alcohol. Carefully fractionated, it yielded a product boiling at 310-320°, which gave the following results on analysis:

0-2089 gave 0.6074 CO2 and 0·1132 H2O. C=79·30; H=6·02.

C14H12O2 requires C = 79.25; H=5.66 per cent.

Pure benzylic benzoate was found to boil at 316.8°, as determined by an Anschütz thermometer with its stem immersed in the vapour (Claisen gives 323-324°). That the product was benzylic benzoate was confirmed by hydrolysis with alcoholic potash, when benzylic alcohol and potassium benzoate were obtained.

Action of Caustic Soda on an Excess of Benzaldehyde.

In the above experiments, the theoretical quantity of caustic soda required by the equation was employed. In order to obtain a condition more favourable to the decomposition of the ortho-compound into benzylic benzoate and caustic soda, an excess of benzaldehyde was used with which the liberated alkali might react until the decomposition was complete. The results and conditions of this series of experiments are tabulated below. The formation of benzylic benzoate when an aqueous solution of caustic soda was used, is a most satisfactory proof of the formation of the intermediate ortho-compound. With dry solid caustic soda, it was found extremely difficult to complete the reaction, since the mass cakes together and the excess of aldehyde thus largely escapes further action. This is doubtless the reason why the extent of the decomposition as well as the quantity of ester found as the product of the decomposition is not greater. The alkali was not always added at once, being sometimes introduced in portions at intervals of several hours. In all cases, the mixtures were heated on the water-bath for the times stated, and the products of decomposition worked up as described above.

In experiments I and II, the total decomposition corresponds approximately to the maximum for the quantity of caustic soda employed, but in experiments III, IV, and V, carried out with dry materials and solid caustic soda, only from 30-50 per cent. of the total decomposition has been effected.

Action of Caustic Soda on Benzylic Benzoate.

If the supposition is correct that an intermediate ortho-compound is formed in the action of caustic soda on benzaldehyde, the same product should result when benzylic benzoate is treated with caustic soda. 26 grams of benzylic benzoate, dissolved in 100 c.c. of benzene, were treated with 5 grams of dry solid caustic soda. No appreciable action took place until the mixture was heated on the water-bath, when, after 6 hours, the whole became nearly solid. The product was dissolved in water and examined, with the result that altogether 20 grams of sodium benzoate and benzylic alcohol were isolated, 10 grams of the ester remaining unchanged, thus proving that benzylic benzoate is not the initial product of the action of caustic soda on benzaldehyde.

Action of Sodium Benzyloxide on Benzaldehyde.

46 grams of sodium were dissolved in 21.6 grams of benzylic alcohol in 125 c.c. of benzene, and 43-2 grams of benzaldehyde were added gradually. After shaking, the mixture was heated on the water-bath for 16 hours, the solid residue dissolved in water, extracted with ether, and the residue from the ethereal solution fractionated. The benzoic acid was separated from the alkaline solution by the addition of hydrochloric acid. The following products were obtained :

In this case, therefore, despite the direct addition of water, decomposition into the ester proceeds to the extent of 54 per cent., and that into benzylic alcohol and benzoic acid only to the extent of 46 per cent. of the theoretical; the formation of these products, however, is a satisfactory confirmation of the correctness of Claisen's view. Also the relative stability of the methyl and benzyl ethers of the ortho-compound, as regards their decomposition into esters, is quite in accord with the influence of these substituting groups.

The decomposition of this benzyl ether by water into benzylic alcohol and sodium benzoate is represented by the equation :

The foregoing results lead naturally to the study of the influence of water in several reactions involving the use of alkalis, such as the two decompositions of ethylic acetoacetate and allied changes; also to the complex action of alkalis upon aliphatic aldehydes. These subjects, as also the action of alkalis on other aromatic aldehydes, are now under investigation.

UNIVERSITY COLLEGE,
LIVERPOOL.

CXX.―The Ultra-violet Absorption Spectrum of Proteids in Relation to Tyrosine.

By A. WYNTER BLYTH.

THE apparatus employed in the investigation of the ultra-violet absorption spectrum of proteids consists of a large quartz prism made of two halves of right and left hand rotation respectively.

A double quartz slit is used, as it possesses the advantage that two spectra are taken one above the other, one with a wide, the other with a narrow slit, so that when it is necessary to use a slit wide enough to blur the lines, the narrower slit will give the metallic lines sharply defined, and thus enable the position of any absorption bands to be measured.

The source of light is a powerful spark produced by a large coil charged either from storage batteries or from the main, the coil is provided with a Wehnelt's break, and one Leyden jar is used as a condenser. The optical train is quartz, and the image is thrown by a lens of 12 inches focal length on to a photographic film placed at a