1) Zeitschr. Untersuch. der Nahrungs- u. Genubmittel 6 (1903), 817. Comp. also Report of Schimmel & Co. April 1904, 19.

2) Report of Schimmel & Co. April 1904, 21.

3) As to the method of preparation comp. Kerp and Unger, Berl. Berichte 30 (1897), 585; also Weddige, Journ. f. prakt. Chem. 10 (1874), 196.

This method can also be used for determining the amount of cinnamic aldehyde in cinnamon bark. For this purpose Hanus gives special directions.

For the determination of small amounts of aldehyde neither the bisulphite nor the sulphite method is adapted since both yield results that may vary from 1 to 2 p. c. Hence the error is too great for small amounts of aldehyde. For cinnamic aldehyde the method of Hanus affords a reliable assay in such cases.

Small amounts of benzaldehyde can be determined, according to Herissey,1) as phenylhydrazone according to the following method:

50 cc. of the solution to be examined, which should contain only so much aldehyde as to yield 0,1 to 0,25 g. of phenyl-hydrazone, are mixed with a solution of 0,5 cc. freshly distilled phenylhydrazine and 0,25 cc. glacial acetic acid in 50 cc. of water and the mixture heated for 20 to 30 minutes on a water bath. After 12 hours the phenylhydrazone formed is filtered through a weighed Gooch crucible, washed with 20 cc. water and dried in a vacuum desiccator.

Very similar is the method of Denis and Dunbar.2) Woodman and Lyford3) have suggested a colorimetric method for this purpose which need not be considered here.

The problem of the citral assay of lemon oil, which contains about 4 to 7 p. c. of this aldehyde, has received much attention from chemists, particularly in recent years. Of the many methods proposed not one has thus far been found practicable since all fall short in the desired degree of accuracy. (For details see under oil of lemon.) If approximate results are sufficient the method proposed by Walther1) and as improved by A. H. Bennet'-) can be recommended: To a mixture of 20 cc. lemon oil with ,20 cc. alcoholic (80 p. c. alcohol) half-normal hydroxylamine hydrochloride solution, 8 cc. of alcoholic normal potassium hydroxide solution and 20 cc. of strong, aldehyde-free alcohol are added and the whole boiled for1/2 hour, the flask being connected with a reflux condenser. After cooling the contents are diluted with 250 cc. of water, part of which is used to rinse the condenser. The hydrochloric acid still combined with hydroxylamine is then neutralized, phenolphthalein being used as indicator. The hydroxylamine not combined with citral can now be titrated with half-normal sulphuric acid, the end of the reaction being ascertained by testing a drop at a time with dilute methylorange solution. A blank test is made in like manner without lemon oil for the purpose of determining the hydroxylamine factor. The difference between the number of cc. of half-normal sulphuric acid used in both tests reveals the amount of hydroxylamine that has entered into reaction. By multiplying this with 0,076 the amount of citral can be computed.

1) Joum. de Pharm. et Chim. VI. 23 (1906), 60.

2) Journ. Ind. and Eng. Chem. 1909, 256. According to Chem. Ztg. Repert. 33 (1909), 281.

3) Journ. Americ. chem. Soc. 30 (1908), 1607.

As has been ascertained by Schimmel & Co.3) the results obtained by this method average about 10 p. c. too low as compared with the amount of citral present. For lemon oils, with their low citral content, the error may not be too great, but for oils, such as lemongrass, with a high percentage of aldehyde the method cannot be used.

Vanillin. For the assay of vanillin, its compounds with B-naphthylhydrazine und p-bromphenylhydrazine may, according to Hanus,4) be utilized. For every part of vanillin 2 to 3 parts of hydrazine are taken. After 5 hours the reaction products are collected in a Gooch crucible, washed, and dried at 90 to 100° until of constant weight.

1) Pharm. Zentralh. 40 (1899), 621; 41 (1900), 614. 2) Analyst 34 (1909), 14. According to Zentralbl. 1909, I. 593. 3) Report of Schimmel & Co. October 1909, 153.

4) Zeitschr. Untersuch. der Nahrungs- u. Genufimittel 3 (1900), 531. According to Chem. Zentralbl. 1900, II. 693.

For the determination of vanillin in the presence of piperonal Hanus1) has, likewise, worked out a method. It is based on the condensation of vanillin with platinum hydrogen chloride, whereby presumably dehydrovanillin results. If the conditions of the reaction, as determined by Hanus are carefully observed, vanillin reacts quantitatively whereas piperonal remains unchanged.

A quantitative method for the separation of vanillin, cumarin and acetanilid has been published by Winton and Bailey.2)

The assay of citronellal is acomplished by means of acetylation (comp. p. 579). The bisulphite method is not applicable in this case because the citronellal sulphonate is very difficultly soluble in the bisulphite solution. Hence it remains suspended in both the aqueous and oily layers and renders even an approximate reading impossible. Neither is the sulphite method serviceable although it is possible to bring the citronellal into solution gradually if the mixture is heated for a longer period and if acetic acid is added from time to time without reference to any alkalinity of the solution.

ketones. As already pointed out, carvone and pulegone can be assayed by means of the sulphite method. This is particularly important in the case of carvone, since the absence of a practicable method was formerly keenly felt.

The method proposed some years ago by Kremers and Schreiner3) yields very inaccurate results. According to this method, the carvone is converted into the oxime, this is separated from the other constituents by steam distillation and weighed. The difficulty lies in determining the point when the distillation should be interrupted, hence a sharp separation of the slightly volatile carvoxime from the other constituents of the oil is impossible. A titrimetric method proposed by Walther,4) which also makes use of hydroxylamine, has likewise not attained practical significance.

1) Zeitschr. Untersuch. der Nahrungs- u. Genufimittel 3 (1900), 657. According to Chem. Zentralbl. 1900, II. 1165.

2) Pharmaceutical Journ. 75 (1905), 476.

3) Pharm. Review 14 (1896), 76. Comp. also Alden and Nolte, Pharmaceutical Archives 2 (1899), 81 and Kremers, Journ. Soc. chem. Industry 20 (1901), 16.

4) Chem. Ztg. Repert. 23 (1899), 264.

Menthone. Thus far the direct assay of menthone has not been a success. Hence one is still dependent on the indirect method first proposed by Power and Kleber,1) whereby the ketone is reduced to menthol by means of sodium and alcohol, and the menthol determined quantitatively. As has been shown in connection with mixtures of known menthone content, approximate results only are obtained. For practical purposes they may, however, suffice in most instances. The assay is carried out as follows: 15 cc. of oil, contained in a round bottomed flask, are diluted with about four times their volume of absolute alcohol,2) the flask connected with a reflux condenser and the solution heated to boiling. To the boiling solution 5 to 6 g. metallic sodium are added very gradually. After all of the sodium has been consumed, the mixture is allowed to cool, diluted with water and acidulated with acetic acid. By means of a separating funnel the oil is separated from the aqueous liquid, washed repeatedly with sodium chloride solution for the purpose of removing completely the ethyl alcohol, and dried with anhydrous sodium sulphate. By means of acetylation the menthol content of the original as well as that of the reduced oil is ascertained, thus revealing the amount of menthol produced by the reduction. From this the amount of menthone contained in the original oil can be computed. If the original oil contained m1 % total menthol, and the reduced oil m2 %, the percentage of menthone of the original oil follows from the formula

Determination Of Aldehydes And Ketones 230 lonone. For the quantitative determination of ionone see p. 468.