Fortunately we have a particularly delicate and characteristic test for this substance, by means of which its presence can be detected in the minute quantity of 1 part to 100,000. This is the well-known characteristic violet color it gives with ferric salts. The test can sometimes be applied directly in the case of very clear beers, but in most samples it is obscured, either by the original color of the solution, or by the color produced by the combination of the iron with other constituents of the liquid. There are various procedures given for the separation of the salicylic acid from these constituents, so as to apply the test to it when in a pure state. Blas1 has investigated various methods, including the previous precipitation of the beer with lead acetate, and the removal of the excess of lead with sulphuric acid;2 shaking out the acidified solution with ether, evaporation of the ether and testing of the residue; treatment with bone-black, washing out the latter with alcohol and applying the test to the alcoholic solution; - none of which methods, he thinks, gives so good results as the application of the test to the urine of a person who has drank some of the beer in question. By this peculiar method of making the human body a medium of separation, he claims to make the test five times as delicate as where it is applied directly to the beer. Borntrager3 arrived at similar conclusions, and I can testify myself to its superior delicacy to the direct application to the beer. Muter4 was probably the first to use dialysis for the separation, also recommended by Aubry.5 Portele6 precipitates the tannin, etc, by a solution of gelatine and shakes up the filtrate with ether. Weigert7 uses amyl alcohol as a solvent, as follows: "50cc. wine1 (or beer) are shaken up with 5cc. amyl alcohol in a small flask for a few minutes and allowed to separate. The clear amyl alcohol is then drawn off and an equal quantity of alcohol added, with which it makes a colorless solution. To this solution is added a few drops of dilute solution of chloride of iron, which produces the violet color."

1 Jour. prak. Cheni., 19, 43. 2 E. Robinet, Cornpt., rend. 84, 1321, 3 Zeit. Anal. Chem. 1881, 87. 4 The Analyst 1, 193.

5 Zeit. Anal. Chem. 1880, 105.

6 Weinlaube 1879, 389.

7 Zeit. Anal. Chem. 1880, 45.

The use of chloroform as a solvent is recommended by the German Imperial Commission for the establishment of methods for wine analysis.2

The Paris Municipal Laboratory allows the choice of four different methods for applying the test in case the ether extract does not afford satisfactory results:

(1) By adding a small quantity of a dilute solution of chloride of iron to precipitate the tannin, and subsequent extraction with ether.

(2) By precipitation of the tannin with gelatine or albumen.

(3) By treatment with ether in the ordinary way and after the ether is evaporated the residue is again treated with a few cubic centimeters of perfectly pure benzine, the solution evaporated, and the residue from this evaporation subjected to the test.

(4) By treatment with pure chloroform.

The union of Bavarian chemists has3 adopted the method of Rose4 who uses a mixture of equal parts of ethylic and petroleum ether for extracting the beer or wine, as follows: "50cc. of the beer are shaken up in a separatory funnel with equal parts of ether and petroleum ether, after acidulating with 5cc. of dilute sulphuric acid. The separation follows very quickly, when the watery part is allowed to flow through the stop cock and the ethereal is poured out through the neck into a small dish. After the ether has been evaporated, and also the greater part of the petroleum ether except a few cubic centimeters, 3-4 cc. of water are brought into the still warm dish. This is well stirred, a few drops of a very dilute solution of chloride of iron added, and the whole filtered through a moistened filter, which allows of the passage only of the watery part of the solution. On the addition of the chloride of iron the petroleum ether solution assumes a deep yellow color, due to certain compounds of the iron with the resin of the hops. In the absence of salicylic acid the filtrate is nearly water-clear with a slight tinge of yellow; if present, even in traces, the solution takes on the well known violet color.5

To test the efficiency of these various methods two samples of a very dark beer known to be free from salicylic acid were taken, and to one was added salicylic acid in the proportion of .05 grams to the liter, and to the other .005 grams. These two samples were then subjected to treatment by eight different methods, as follows :

1 Some of the processes given are intended for wine, but are equally applicable to beer, milk, fruit juices, etc. 2 Die Weinanalyse, Kommentar, u. s. w. 3 Op. cit.

4Arch; f. Hygien. Analyst. 1886, 133. 5The same test is given by H. Taffe, Bull. de la Soc. de Chim. de Paris, 46, No. 12.

(1) Extracted with chloroform.

(2) Extracted with ether simply.

(3) Extracted with amyl alcohol.

(4) Extracted with ether after previous precipitation with lead subacetate.

(5) Extracted with ether after previous precipitation with gelatine.

(6) Extracted with ether after previous precipitation with ferric chloride.

(7) Extracted with equal parts of ethylic and petroleum ether.

(8) Extracted with ether, the solution allowed to evaporate spontaneously, and the residue extracted with benzine. These tests, tried on the beer containing .05 grams to the liter, gave results which ranged them in the following order of succession, according to the strength and brilliancy of the test: 8, 5, 7, 6, 2, 1, 4, 3. In methods 5 and 6 the length of time required to filter the solutions constitutes an objection to their use. The same tests applied to the beer containing .005 grams to the liter gave the test only in the case of Nos. 8 and 7, the former being the better of the two. The same series of tests were applied to a red wine, with very similar results, so I have adopted method No. 8 in all the work done on beers and wines, and would recommend either that or No. 7, both of which give a perfectly bright, waterclear solution, in which the slightest tinge of violet color is plainly visible.