Obviously the foregoing methods will give results of only an approximate character. The following, however, is a tolerably accurate process for estimating alcohol in such articles as wine or beer, and is expeditious, especially where a number of determinations are required. It depends upon the principle that the boiling point of an alcoholic liquid is lower than that of water in proportion to the quantity of alcohol present. The special instrument devised for taking the boiling point is known as the ebullioscope. It was originally used for the purpose by the Abbe Broussard-Vidal, of Toulon, and later was improved by Malligand and Mlle. E. Broussard-Vidal.1 Field in 1847 patented an "alcoholmeter" based on the same principle, and Ure subsequently made some improvements in Field's apparatus.

Essentially, Malligand's form of the ebullioscope consists of a boiling vessel fitted with a condenser and a special thermometer, and heated by a spirit lamp. The thermometer is graduated, not in temperature degrees, but in figures showing directly the percentage of alcohol by volume. A few years ago J. C. Cain tested Malligand's ebullioscope by the determination of alcohol in beer, comparing the results with those obtained by the distillation process; he found a satisfactory agreement.2 Cain's description is essentially as follows: mark and boiled as described in the case of the water. Readings of the mercury are conveniently taken at fifteen seconds intervals, beginning when the regular boiling commences. The true boiling point is not reached until the beer has been boiling steadily for some time, when the mercury remains stationary for a period. The steady boiling is easily recognised by placing the ear near the top of the condenser, when a continuous bubbling is heard. If the boiling is continued, the mercury advances again after a time, and vapour escapes from the top of the condenser. The following readings, taken at fifteen seconds intervals, show the course of an experiment: -

In making an experiment, distilled water is first placed in the lower part of the vessel (up to the mark inside), the lid screwed on, the thermometer inserted, the condenser filled with cold water and screwed in, and the water heated to boiling. The spirit lamp should be protected from draughts. When the water has boiled steadily for a short time, the mercury thread of the thermometer remains at one point, and the scale of the thermometer is then moved by means of the controlling screw until the zero point is coincident with the position of the end of the mercury thread. The water may be boiled until the steam begins to escape through the condenser, by which time the zero point will have been accurately ascertained.

This "setting" of the instrument renders unnecessary any other correction for barometric pressure. It is required on each day when the instrument is to be used.

After the zero point has been fixed and the vessel emptied of the water, the beer (at any ordinary temperature) is filled up to the

1 Compt. rend,, 1874, 78, 1470.

2 Chem. News, 1914, 109, 37.

The result was therefore 5 8 per cent. of alcohol (by volume). Distillation gave the same result.

The solids present in beers and wines exert no practical influence on the boiling point, as their molecular weights are so great. Abbe Vidal had originally found that there was no sensible effect produced on the temperature of ebullition by such ingredients as sugar, if not present in greater amount than 15 per cent.

An adaptation of the instrument for use in determining the "original gravity " of beer has been described by T. H. Pope1

Another form of ebullioscope for estimating alcohol in wines has been described by Ph. Malvezin2, which claims to obviate certain sources of error, notably those arising from the presence of a high proportion of extract. The wine is contained in a conical vessel immersed in a conical water-bath, and there are two thermometers, one giving the temperature of the wine and the other that of the water-bath. The readings are referred to tables constructed from actual determinations. A figure of the instrument, but not the tables, is given in the J. Soc. Chem. Ind. for 1915, p. 846.

Using the following table, which is due to P. N. Evans,3 it is possible to estimate approximately the percentage of alcohol in mixtures of alcohol and water by a simple determination of the boiling point. The table was constructed from observations made when 500 c.c. of alcohol was distilled at a uniform rate of 1 drop per second. The temperature was noted when 7.5 c.c. had distilled over, and the distillation continued until 15 c.c. was collected. The mean of the percentages of alcohol in the liquid in the flask before and after distillation, as ascertained from the density, was taken to represent the composition of the liquid, and the percentage of alcohol in the distillate was taken to represent the composition of the vapour at the temperature observed when 7 5 c.c. of distillate had been collected. The temperatures given are corrected for barometric pressure and for the exposed mercury column of the thermometer.

1 Brewers' J., 1914, 50, 239. 2 Bull. Assoc. Chim. Sucr., 1915, 32, 104. 3 J. Ind. Eng. Chem., 1916, 8, 260-262.

Boiling points and condensing points of aqueous alcohol.