mineral constituents of the latter seriously affect specific gravity, containing in a given specific gravity about twice as much substance in solution as a sugar solution of the same gravity), it is necessary to subtract from the percentage of extract thus estimated, the percentage of ash found in the same wine; or if the amount of extract without the ash is required, twice the percentage of ash has to be subtracted from the percentage found. Dupré and Thudichum give the following examples:

Specific gravity of de-alcoholised wine, .

Percentage of extract from specific gravity (see

table, p. 522),

Percentage of ash found,

1017.86

4.467

0.515

3.952

0.515

3.437

Total solid constituents, .

H. Hager,* after evaporating off the alcohol, and making up the wine to its original volume by means of water, determines the amount of extract from the following table, which is based on his own experiments, and differs a little from the malt extract table p. 522.

The extract and amount of alcohol being known, it is, in certain instances, possible to detect the watering of wine, although such a diagnosis can only be made when the analyst is intimately acquainted with the kind of wine under examination, and in some cases with the characters of the particular vintage. The Bordeaux wines, according to Girardin and Pressier, give almost always the same amount of extract, varying only within the limits of 20 to 20-8 grms. the litre; and the proportion of alcohol also is fairly constant-viz., from 005 to 015, the mean being 010 per litre. From these data they calculated the

* Chem. Centrbl., 1878, 415.

[N.B. The specific gravity increases or diminishes '00024 for each degree.] amount of genuine wine present in any samples. Thus, supposing the extract in a Bordeaux wine to be 14.5, then

i.e., the litre contains 725 cc. of wine, the rest being alcohol and water. To know the quantity of alcohol added, it is necessary to ascertain how much the 72-5 parts of wine contain of absolute alcohol;

If the absolute alcohol is found, for example, to be 0.11, then, subtracting 7.25 from 11, it is supposed that 3.75 of alcohol has been added.

That this process, as applied to the Bordeaux wines, is in the main correct, is supported by the fact that the Rouen wine-merchants have frequently paid duty on the excess of alcohol, &c., which Girardin and Pressier found in their wines.*

In Paris a commercial standard has been arrived at, based on the analysis of 6,000 samples, and it is laid down that "the amount of added water in all wine which is not sold as being of any special brand, shall be calculated on the basis of 12 per cent. of alcohol by volume, and 24 grms. of dry extract per litre."

In the Municipal Laboratory, Paris, the chemists determine the watering and fortifying of wine by a calculation of the relationship of what they call "the reduced extract" to alcohol.

=

The "reduced extract" is the total extract diminished by the number of grammes less 1 of the potassic sulphate and reducing sugar; thus, if the total extract should be 297, the potassic sulphate 3.1, and the sugar 4·5; 2·1 + 3·5 5-6; and subtracting 56 from the total extract gives the reduced extract as 24-1. The weight of the alcohol for red wines they consider should not exceed four and a-half times the extract, and when this relation is exceeded the wine has been fortified. To determine this relation, the alcohol by weight is divided by the reduced extract.

For white wines the relation between alcohol and reduced extract is fixed at 6.5. Wines watered down and then the alcoholic strength brought up by the addition of alcohol are detected in the French Laboratory as follows:

In all normal wines the sum of the alcohol per cent. by volume added to the total acidity per litre (alcohol-acid number), calculated as sulphuric acid, is seldom below 12.5. Water lowers this number, alcohol increases it. First, then, the relation of the alcohol to the reduced extract is obtained; if it exceed 4-5, then by calculation on the standard of 4·5, the amount of alcohol that the natural wine may be supposed to have had originally is obtained; and the difference between this and the amount found represents the alcohol added. Next, the alcohol-acid number is obtained, and if this is below 12.5 the presumption is that the wine has been watered.

An example will make this clear. following:--

Dry extract per litre,

Acidity per litre,

Alcohol per cent. (volume),

The relation by weight of alcohol-extract,

The alcohol-acid number,

Now calculating on a standard of

Natural alcohol in the wine,

A red wine gave the

14·2 × 4·5 = 63 9

Reducing this by dividing by 08, the volume of

alcohol is equal to

Alcohol in excess,

Alcohol-acid figure,

7.99

Hence, the alcohol-extract number being superior to 4.5 and the corrected acid-alcohol number being below 125, there is a presumption of both watering and fortifying.

This calculation being the results of prolonged experience with regard to ordinary wines is a fairly safe guide to the

analyst; but with regard to certain exceptional wines caution must be exercised in the interpretation of results.*

§ 280. Estimation of Succinic Acid and Glycerin.—Half a litre to a litre of wine is decolorised with animal charcoal, filtered, and the charcoal well washed with water; the filtrate and washings are then evaporated down in the water-bath, and the drying finished in a vacuum. The residue, when dry, is treated with a inixture of 1 part of strong alcohol and 2 parts of rectified ether. The latter is driven off by floating the dish in warm water, and the whole evaporated again on a water-bath. The residue is now neutralised with lime-water, which combines with the succinic acid, and forms succinate of calcium. The glycerin is dissolved out by alcohol and ether, and weighed either directly or by loss. The succinate of calcium remaining behind is impure, and should be well washed with spirit before weighing. Every 100 parts of calcic succinate equals 75 64 of succinic acid (H,CH ̧ ̧); and since Pasteur has shown that 112-8 parts of grape-sugar (107 of cane) yield about 3-6 of glycerin and 0.6 part of succinic acid, it follows that in a natural wine the glycerin would amount to about one-fourteenth part of the alcohol present.†

4

It has, indeed, hitherto been generally accepted that for every 100 parts of alcohol there should be not less than 7 nor more than 14 of glycerin, and it has been held that deviations from this standard mean addition of alcohol or glycerin; but the highest quality of the Rhine wines (Table XLIVa) vary between 16 and 31 parts of glycerin to 100 of alcohol, and, therefore, the views hitherto held demand modification.

Schmidt evaporates a known bulk of the wine with hydrated calcium oxide, extracts the residue with 96 per cent. alcohol, evaporates the clear filtrate, then dissolves this last residue in 15 cc. absolute alcohol, precipitates with 25 cc. of ether, filters and evaporates the alcohol-ether solution, and, after one and a-half hours' drying, weighs.

Stierling evaporates the liquid without any addition to one-fifth or one-sixth of its volume, extracts with hot absolute alcohol, and uses this alcoholic extract for the estimation of the sugar, non-volatile acids, bitter matters,

E. Egger detects the watering of wine by the presence of nitrates, the grape being stated to be absolutely destitute of nitrates; white wines are evaporated to a syrup and absolute alcohol added; so long as it produces a cloud the mixture is filtered, decolorised, and tested with diphenylamine and sulphuric acid. Red wines are precipitated with lead acetate and then white magnesium sulphate before evaporation (Chem. Centr., 1885, 71, 72).

According to E. Borgmann (Zeitsch. für. anal. Chemie, xxii., 58-60) the ratio of glycerin to alcohol in pure wines is never less than 73: 100. Analyses of white wines by R. Fresenius and E. Borgmann give the following ratios of glycerin and alcohol:

Maximum,

Minimum,

Mean,

Alcohol. Glycerin.

§ Stierlin, "Das Bier," &c. Bern, 1879.

alkaloids, and glycerin. For the estimation of the last, a measured portion of the alcoholic extract is freed from alcohol by evaporation, and then evaporated down to dryness with slight excess of caustic lime. The glycerin is extracted either with alcohol and ether (2 : 3), or with alcohol and chloroform. (See also the process for extracting glycerin from beer, p 531.)

Raynaud has pointed out that although the processes in use for the estimation of glycerin are fairly exact, yet with plastered wines too high results are obtained; for if there is any considerable amount of sulphate of potash, it is decomposed by lime, and hydrate of potash is formed, which is dissolved by glycerin in the presence of alcohol, and is weighed with it. He therefore recommends the following process:-The liquid operated upon is evaporated to about one-fifth of its volume, and the potash precipitated by hydrofluosilicic acid and filtered. The filtrate is made weakly alkaline by the addition of hydrate of baryta; sand is also added, and the mass is evaporated to dryness in a vacuum; the dry residue is then extracted with a very large volume of absolute alcohol and ether, as much as 300 cc. for 250 cc. of wine being recommended. With the improved processes of extraction which we now possess, however, this is quite unnecessary, and 50 to 100 cc. in a Soxhlet's apparatus (see p.67) will have quite the same effect as a much larger quantity, On the evaporation of the alcohol and ether, the glycerin is allowed to stand for twenty-four hours in a vacuum over phosphoric anhydride; finally, it is put into a tube, a perfect vacuum formed, and distilled into the cool part of the tube by a temperature of 180°.

Probably the best method of estimating glycerin is to separate it from most volatile substances by distillation in a vacuum, and then to oxidise it into oxalic acid, as described under the article Butter (p. 371). A. Partheil* effects this in the following manner:-50 cc. of the liquid to be examined, first neutralised by adding a little calcium carbonate, are evaporated down to 15 cc. and introduced into a small retort. This retort is enclosed in an air bath, the bottom of the bath being made of sheet iron, the sides and top of asbestos card. The neck is connected with a globular receiver, the second opening of the receiver being joined to an inverted condenser, and then to a pump. The receiver is also kept cool. The liquid is first distilled almost to dryness, at ordinary pressure, at a temperature of 120°. It is then cooled to about 60°, and the pressure reduced by means of the pump, the temperature raised to 180°, and the distillation continued for one and a half hours; the pressure at the end of that time is released, the retort cooled, 10 cc. of water added, and distillation again proceeded with at the ordinary pressure at a temperature in the bath of 120°. The distillate is diluted to about 200 cc., 8 to 10 grms. of caustic soda dissolved in it, and 5 per cent. potassic permanganate added until the colour remains a decided blue-black. The whole is heated for an hour, decolorised with SO2, 20 cc. of acetic acid added, the SO2 driven off by heat, and the oxalic acid precipitated by calcium chloride.

§ 281. Estimation of Tartaric Acid and Bitartrate of Potash.This is best estimated by the method suggested by Berthelot 20 cc. of wine are mixed with 100 cc. of equal volumes of alcohol and ether in a well-stoppered flask. The same process is employed to another 20 cc., but with the addition of potash in sufficient quantity to neutralise about one-fifth of the free acid present. Both bottles are allowed to stand two or three days, and at the

*Arch. Pharm., 1895, ccxxxiii., 391.