68 



matter was independently studied by A. M. Doyle and P. B. Dimbar, of the Bureau 

 of Chemistry. Both reported that the color given by ferric chlorid with the material 

 extracted from highly colored malt by ether was quite different from the salicylic 

 acid and one should not be mistaken for the other, although the presence of a small 

 amount of salicylic acid may readily be masked by the material extracted from highly 

 colored malt and similar material. 



Experiments on malt-nutrine alone and on malifc-nutrine containing salicylic acid 

 (100 mg per liter) indicate that there is some possibility of being deceived by the 

 color when ferric chlorid is added directly to the dish containing the dried residue 

 obtained by evaporating the ether solution. In this case a color is sometimes devel- 

 oped which slightly resembles the salicylic acid reaction. If this color is examined 

 in a good light and compared with the color developed by salicylic acid and ferric 

 chlorid there is little danger of being deceived. It is better, however, to carry the 

 evaporation of the extract to about 5 cc on the steam bath and then to complete the 

 evaporation by means of a blast of air, since heating to dryness may darken the residue. 

 The dry residue should be dissolved in a little hot water and ferric chlorid added to 

 this solution. Under these conditions it seems impossible to mistake the salicylic 

 acid reaction. Millon's reagent, freshly prepared, which has been suggested for the 

 detection of salicylic acid in such substances, was not found by either Miss Doyle 

 or Mr. Dunbar to be as satisfactory as ferric chlorid. In the absence of salicylic acid 

 this reagent gives a light pink color, whereas in its presence a deep red is given. The 

 intensity of the color seems to vary so much with the time of boiling, however, as to 

 render the reaction uncertain and unsatisfactory. Both methods were also applied to 

 highly caramelized breakfast foods, to which these observations also apply. 



DETERMINATION OF BENZOIC ACID. 



Several methods for the quantitative determination of benzoic acid have been sug- 

 gested recently. Among these the following have been studied by the referee and his 

 collaborators: (1) La Wall's method; (2) La Wall's method modified by mixing a 

 definite weight of tomato ketchup with sufficient saturated sodium chlorid solution to 

 make a definite volume, filtering, and extracting an aliquot portion of the filtrate with 

 chloroform; (3) precipitation as copper benzoate; (4) precipitation as silver benzoate; 

 (5) distillation with steam after decomposing organic matter with sulphuric acid and 

 extracting with ether. It will be noted that the first four methods depend on extract- 

 ing the benzoic acid from the food or an aliquot extract of the food by means of an im- 

 miscible solvent, whereas the fifth depends on separating the preservative from the 

 food by distilling with steam. 



Before applying the first four methods a preliminary study was made of the relative 

 advantages of several solvents. This question has been greatly altered by the intro- 

 duction of the principle of "salting out" the benzoic acid by means of a saturated 

 solution of sodium chlorid . The benzoic acid is thus rendered much less soluble in the 

 solution from which it is extracted and consequently much more readily extracted by 

 the immiscible solvent. The relation of the immiscible solvent to the preservative 

 thus more nearly coincides with the direct solubility of the former in the latter. An 

 approximate determination was therefore made of benzoic acid in several solvents. 

 It was found that the solvents more commonly employed dissolved benzoic acid as 

 follows, 100 cc of the solvent being used in each case: Ether, 25.70 grams; chloroform, 

 17.34 grams; carbon tetrachlorid, 7.65 grams; toluene, 7.58 grams; benzol, 6.40 grams. 



The chief desiderata in the extraction of a substance of this kind are, first, com- 

 pleteness of extraction; second, freedom of the extract from interfering substances; 

 and third, noninnammability. Of the four solvents mentioned, it would appear from 

 the solubility that by far the most complete extraction can be obtained by means of 

 ether, and this is known to be true. Ether is objectionable, however, because of its 

 property of dissolving water and its consequent tendency to extract tannin, salts, min- 



