Chemical Preparations and Drugs. 103 



of vanillin into its ^»-nitro-phenylhydrazone). Colorimetric methods have also been 

 elaborated, for instance Fellenberg 1 ) has suggested a process, which depends on the 

 colour-reaction which vanillin undergoes with cone, sulphuric acid and isobutyl alcohol. 

 A. W. Dox and G. P. Plaisance 2 ) have elaborated a new method for testing the alcoholic 

 vanilla extracts which find so much favour in America; they have found, namely, that 

 thiobarbituric acid forms with vanillin in the presence of 12 percent, muriatic acid an 

 only slightly soluble vermilion-coloured precipitate of 3-methoxy-4-hydroxybenzalmalonyl- 

 thiourea. This method is not suitable for the valuation of artificial extracts coloured 

 by caramel, as the latter contains furfurol which also reacts with thiobarbituric acid. 

 Accordingly, the vanilla extracts must be tested for the presence of caramel, for which 

 purpose one adds phloroglucinol to the solution, previously clarified by acetate of lead, 

 and after having removed the excess of lead as chloride of lead. A brown precipitate 

 is formed in cases in which caramel is present; when absent, however, the vanillin only 

 undergoes a light-pink colouring or forms a small quantity of a pink-coloured precipitate. 



The determination is executed in the following manner:— 25 ccm. of the alcoholic 

 vanilla extract are freed from alcohol in the customary manner and filled into a 

 measuring flask of 50 ccm. volume ("Sugar flask"), which is then filled up to the 

 graduation mark with lead acetate solution. The flask is then left alone at 37° for 

 some hours an its contents are then poured on to a dry filter. 40 ccm. of the filtrate 

 which must be straw-coloured (absence of caramel) are then poured into a second 

 50 ccm.-flask and filled up to the mark with concentrated muriatic acid; the acid 

 content must then be about 12 per cent. In order to remove the lead chloride, after 

 a few minutes one filters the liquid through a dry filter, treats 40 ccm. of the filtrate 

 with a solution of thiobarbituric acid in 12 per cent, hydrochloric acid, lets it stand 

 over night, filters through a Gooch crucible, washes with 50 ccm. of 12 per cent, 

 muriatic acid, finally with 20 ccm. of water and dries at 98°; one can then easily 

 calculate the content of the extract in the shape of vanillin from the weight of the 

 precipitate. Clarification with the aid of lead acetate is imperative, as otherwise the 

 results are too high. 



C. Estes 3 ) has described another method for the determinatidn or at least for the 

 qualitative detection of vanillin. If one heats vanillin with mercuric nitrate, a violet 

 colouring is formed whose intensity corresponds to the quantity of vanillin present. 

 One treats 5 ccm. of vanilla-extracts, in a 50 ccm.-flask, with 6 ccm. of water and 

 1.5 ccm. of the reagent (see below); after leaving the flask for 20 minutes in boiling 

 water one cools it down rapidly, the contents are diluted to 50 ccm. and the colouring 

 formed compared with that produced by a known quantity of vanillin under the same 

 conditions. Non-alcoholic extracts are treated just in the same manner, but in such 

 cases one only requires 1 ccm. of the reagent. For the purpose of preparing the latter, 

 one dissolves mercury in the double quantity of nitric acid (specific gravity 1.42) and 

 dilutes the solution with 25 times as much of water. 



*) Mitt. Lebensmittelunters. u. Hyg. 6 (1915), 267; Report October 1916, 92. — -) Americ. Journ Pharm. 

 88 (1916), 481. — 3 ) Journ. ind. eng. Chemistry 9 (1917), 142; Journ. Soc. diem. Industry 36 (1917), 352. 



