good bibliography concludes the work.

He has confined himself thus far to analyses of the pure tannins and does not as yet propose formulas. In addition to the monograph just mentioned, he has been a frequent contributor on this subject in the periodical literature of chemistry.

LATE WORK OF SCHIFF ON DIGALLIC ACID.

In the present year, Hugo Schiff,' having already found that natural tannic acid is possessed of a slight optical activity, namely dextrorotatory, inquires as to its containing in some way asymmetric carbon, and infers a ketonic structure with six hydroxyl groups. He does not obtain a hydrazone from tannic acid. From experimental results he proposes this formula:

Irrespective of orientation, and for comparison of a kolatannin formula offered for consideration further on, the above formula may be held simply as

CAFFEOTANNIC AND CAFFEIC ACIDS.

The literature of caffeotannic acid, of the formula C,,H,,O., dates from Rochleder, in 1846. Its production of caffeic acid (dihydroxyphenylacrylic acid), along with sugar, was set forth by Hlasiwetz in 1867.3

A few years ago, caffeic acid was obtained from the tannin of coffee, also from the tannin of the maté, by H. Kunz-Krause.* Very recently the same author has investigated caffeotannic

1 Gazz. chim, ital., 1897, 27, i, 90.

2 Ann. Chem. (Liebig), 59, 300.

8 Ibid, 142, 220.

4 Arch. Pharm., 1893, 231, 613.

acid.' In his report he finds confirmation of the composition of caffeic acid as a dihydroxyphenylacrylic acid. His analyses obtain the carbon dioxide given off when the caffeic acid is heated to 200° C., and are confined to this analytic method. A residue of a vinyl pyrocatechol is predicted. The author recognizes a class of bodies which he terms glucotannoids.

"FREE" KOLATANNIN.

PREPARATION OF THE SAMPLE.

The amorphous character of tannins, together with their extreme liability to decomposition, renders their separation and purification very difficult. The method used by us last year," in the very brief examination of kolatannin, was based on that of Allen, being fractional precipitation with lead acetate.

On taking up the work this year, it soon became evident that whatever may be said of lead-acetate methods as applied to other tannins, they certainly are unsuited to this one, for various reasons. The lead kolatannate, being bulky, is extremely slow and difficult of filtration, and the tannin in this state exhibits a marked tendency to pass into insoluble forms, principally anhydrides. During the washing, the lead salt becomes gradually darker, and this tendency is favored by the moisture present. If, on the other hand, the washing be less prolonged, the resulting product will be contaminated with various other constituents of kola extract, such as sugar, alkaloids, inorganic salts, coloring-matter, etc., etc. It is indeed very doubtful if any usual amount of washing will remove these impurities, hence the purity of a tannin, separated by precipitation with lead acetate, would be regarded with too much suspicion to be employed in analytical work, even if its sensible properties were not changed during the treatment. Kolatannin was at first supposed and since determined to be soluble in water. But on evaporating the solution of tannin after removing the lead from combination by hydrogen sulphide, the end-product was a red-brown

11 am indebted to Prof. Kremers for bringing Kunz-Krause's article in the July Berichte to my attention, as I read this paper in the August meeting of the American Pharmaceutical Association.-A. B. P.

2 Ber. d. chem. Ges., 1897, 1617.

3 Knox and Prescott, loc. cit.

4 Allen: Com. Org. Anal., 3, pt. I., 76.

substance, insoluble in water, and possessing a not particularly astringent taste. So by the use of lead acetate a tannin is obtained whose purity is a matter of doubt, and whose properties have undergone radical changes during separation, both as to color and solubility. Furthermore, we desired to effect a perfect separation of the free tannin from that existing in combination with caffein (as caffein kolatannate) in order to determine by analysis whether they were identical tannins or not, the analysis of them last year having shown somewhat different values in each for carbon and hydrogen. Caffein kolatannate is sparingly soluble in water, more soluble in solutions of tannin and of caffein, so that the aqueous infusion of kola contains caffein kolatannate as well as free tannin, and the lead tannate obtained from it would be a mixture of the lead salts of both tannins, or both forms of tannin, and no delicate separation could be had in this way.

Plainly, another method was necessary. The objections just noted were for the most part neither new, nor limited to kolatannin. As previously stated, Löwe, in 1872, who had encountered many difficulties in the use of lead acetate for the separation of tannins, and found it necessary to employ other means to obtain a pure product, resorted to the use of ethyl acetate. This immiscible solvent, proving satisfactory in many cases, has since been used considerably, either alone or mixed with ether, in the separation of tannins for analytical purposes, and for some tannins at any rate is unquestionably the best separative solvent which has yet been given publicity as such. At the same time a single shaking out with ethyl acetate does not suffice for the preparation of a pure sample. More or less coloring-matter will be found in the residue left on evaporating the first portion of the solvent, so that several repetitions of the process are required. Care must be used not to carry the method of purification too far, for if this be done the decomposition of the tannin is apt to prove a greater obstacle than the original impurities. We find that four or five repetitions of the "shaking out" give the best results. In detail the manipulations are as follows, and the method has given satisfactory results:

The fresh kola nuts are sliced into boiling alcohol,' removed after a few moments boiling, and dried in a current of warm air, then ground up to a number twenty powder and packed firmly in a percolator. The alcohol so used in sterilizing the drug is diluted to about fifty per cent. strength, and employed as a menstruum, with addition of sufficient dilute alcohol to complete the extraction. The highly-colored extract of kola thus obtained is concentrated by distillation in vacuo, until the alcohol is all removed. This can be done with a good pump at between 18° and 20° C. The contents of the flask are then filtered, the insoluble portion being chiefly caffein kolatannate, while the solution contains caffein, kolatannin, caffein kolatannate, glucose, traces of fatty matter, and more or less colored derivatives of tannin according to whether the sterilization by boiling alcohol has been properly done. Common salt is then added to this filtrate to saturation and the caffein kolatannate is completely precipitated. It is filtered out and added to the first residue of caffein kolatannate. The reddish-colored filtrate is then transferred to a large separatory funnel and agitated with chloroform repeatedly to remove alkaloids and traces of fat. The dissolved chloroform is then separated by agitation with small portions of

Kolatannin is insoluble in chloroform, and only very sparingly soluble in ether, so this treatment does not remove any considerable amount of it. Ethyl acetate is now added to the liquid in the separatory funnel and the liquid extracted with it as long as any tannin is removed. Usually eight, ten, or even more successive portions are required for this. The united solutions of tannin in ethyl acetate are transferred to a distilling flask and concentrated to dryness under reduced pressure. To avoid loss of the ethyl acetate we use a condenser about a meter in length, and pack the receiver in ice. Sometimes also a condenser is placed between the pump and the receiver. With these precautions the waste is very slight. Before using the recovered ethyl acetate as a solvent again, however, it is washed with a weak solution of sodium carbonate and redistilled, to avoid the presence of any free acetic acid formed by decomposition of the ether. The statement of Etti already alluded to, concern1 To prevent the formation of the colored body, which would otherwise appear.

ing the decomposition of ethyl acetate by tannin, has not been found to apply to kolatannin, but the washing was nevertheless adopted as a precautionary measure.

The tannin residue in the distilling flask is a porous, pinkish mass, very friable, and easily and completely soluble in water. It is redissolved in a convenient quantity of saturated salt solution, filtered if need be and again shaken out with ethyl acetate, which is distilled off in the way just described. The tannin residue is next dissolved in cold distilled water and shaken out as before, this process being repeated once or twice. The final tannin residue, after the ethyl acetate has been removed as far as possible by distillation, is treated in the flask with a small quantity of ether, which after thoroughly permeating the mass is removed by distillation, using only the partial force of the pump and gentle heat. This is repeated several times until the mass has no odor of ethyl acetate. Then transferring to a vacuum desiccator the tannin is kept for several days over sulfuric acid. If this does not remove the last traces of ether, gentle heat (60°65°) will do so.'

PROPERTIES OF KOLATANNIN.

Kolatannin is a cream-colored powder with a slight pinkish tinge. It is freely and completely soluble in water, alcohol, acetone, ethyl acetate, sparingly soluble in ether, insoluble in chloroform and in petroleum benzin. The following tabulated comparative statement of its reactions will serve to point out some of the important differences between it and gallotannic acid, and to indicate the close resemblance between this tannin and that of oak bark.

1 We desire to take this means of expressing our appreciation of the liberality and courtesy of the firm of Frederick Stearns & Company, Manufacturing Pharmacists, Detroit, Michigan, who supplied us with a large quantity of True African Kola nuts, for our research this year. The drug was quite fresh and in perfect preservation, very necessary conditions for work of this kind.