roasted chicory, it varies from 12 to 18 per cent. The quantity of sugar in a sample can be determined by Fehling's method as follows:

A standard solution of pure cupric sulphate is first prepared by dissolving 34 64 grammes of the crystals (previously ground and dried by pressing between bibulous paper) in about 200 c.c. of distilled water; 173 grammes of pure Rochelle salt are separately dissolved in 480 c.c. of a solution of sodium hydroxide of sp. gr. 114. The solutions are then mixed and diluted with distilled water to one litre. Each c.c. of the above solution represents 0.05 gramme of grape sugar. The test is applied by taking 10 c.c. of the copper solution, adding about four times its volume of water, and bringing it to the boiling point. The coffee infusion is then gradually added from a burette, until the copper salt is completely reduced to the red sub-oxide, which point is recognised by the disappearance of its blue colour, and can be more accurately determined by acidulating the filtered fluid with acetic acid and testing it (while still hot) for any remaining trace of copper with potassium ferrocyanide. In preparing the coffee solution for the foregoing test, it is advisable to exhaust a weighed quantity of the sample with hot water. The infusion is treated with basic plumbic acetate so long as a precipitate forms; it is then filtered, the precipitate being well washed, and the lead contained is removed by conducting sulphuretted hydrogen gas through the fluid which is subsequently again filtered and boiled until the dissolved gas is expelled. The sugar determination is now made. Wanklyn employs the following equation to estimate the amount of chicory in an adulterated sample:

where E is the percentage of chicory, and S the percentage of sugar.

According to the analysis of König, the proportions of sugar and other constituents in some of the adulterants of coffee, are as follows:

Estimations of the amount of sugar obtained upon boiling the suspected coffee with water containing a little sulphuric acid (see p. 37), and the proportion of the sample which is soluble in hot water should be made. The presence of chicory is shown by a decided increase in the amount of soluble substances; that of rye, by the notable quantity of sugar produced by the inversion with acid, due to the starch contained in the grain.

In this connection, the following determinations of Krausch are of interest :

The presence of roasted rye, corn, and other grains in coffee, may be qualitatively recognised by testing the cold infusion of the sample with iodine solution for starch, which

is not contained in a ready formed state in coffee. Caffeine is absent in chicory and the other usual adulterants of coffee, and the estimation of this alkaloid is of decided service (see p. 21). Roasted coffee contains about I per cent. of caffeine.

A popular brand of ground coffee received by the author for examination, and labelled "Prepared Java Coffee," had the following approximate composition :-Coffee, 38; peas, 52; rye, 2; and chicory, 7 per cent.

A sample of "acorn" coffee, analysed by König, gave the following results :

The non-nitrogenous constituents contained from 20 to 30 per cent. of starch, and from 6 to 8 per cent. of tannic acid.

The composition of the well-known German coffeesubstitutes, prepared by Behr Bros., is stated to be as follows::

The raw coffee bean is sometimes subjected to a process termed "sweating," which consists in treating it with moist steam, the object being to artificially reproduce the conditions present in the holds of vessels, by means of which the bean is increased in size, and also somewhat improved in colour and flavour. Another form of manipulation, analogous to the facing of tea, is to moisten the raw bean with water containing a little gum, and agitate it with various pigments, such as indigo, Prussian blue, Persian berries, turmeric, alkanet, Venetian red, soap-stone, chromeyellow, and iron ochre. Mexican coffees are sometimes made to resemble the more expensive Java in appearance. The chemist of the New York City Board of Health has found in the quantity of such treated coffee commonly taken to make a cup of the beverage 0.0014 gramme of cupric arsenite. Indigo may be detected in the artificially coloured product by treating a considerable portion of the sample with dilute nitric acid, filtering and saturating the filtrate with sulphuretted hydrogen. If indigo be present, it can now be extracted upon agitating the solution with chloroform. Alkanet root and Prussian blue are separated by warming the coffee with solution of potassium carbonate, from which these pigments are precipitated upon addition of hydrochloric acid.

(c) Microscopic Examination.-Great aid to the chemical investigation is afforded by the microscopic examination of ground coffee. It is necessary to first become familiar with the appearance of the genuine article-low magnify

ing powers being employed-and then make comparative examinations of the adulterant suspected to be present.

The coffee bean mainly consists of irregular cells inclosed in very thick walls which are distinguished by uneven projections. The cells contain globules of oil. Most of the roots added to coffee exhibit a conglomeration of cells (provided with thin walls) and groups of jointed tubes, often quite similar to one another in structure. The microscopic appearance of some of the starch granules, occasionally met with in coffee mixtures, is represented on p. 100.

Of 151 samples of ground coffee recently purchased at random and tested by various American chemists, 69 (45 7 per cent.) were found to be adulterated.