tannin ; but we have failed to find any confirmation of this view from the results of our experiments. On the contrary, our results point to the presence of a sugar peculiar to coffee, and bearing somewhat the same relation to cane sugar as melezitose, mycose, and similar compounds. When the saccharine extract of the beans is boiled with a little dilute sulphuric acid, a quantity of glucose is obtained nearly equivalent to the weight of dry extract operated upon-a result which is inconsistent with the existence of sugar as a glucoside. The sugar is also but slightly affected by boiling even for a considerable time with acetic acid. If it existed as cane-sugar it would, under such conditions, be converted into invert sugar.

Gum.-This appears to be present in coffee in the form of dextrin, and can be recovered and identified in the water extract after the coffee has been thoroughly exhausted with ether and alcohol. The amount of gum is but little altered in the roasting. The quantity present is from 84 per cent. in the raw to 1.38 per cent. in the roasted berries.

Albumin.-The amount of soluble albumin or legumin is greater in coffee than in tea. The amount found is 2.53 per cent. in the raw and 1'47 per cent. in the roasted coffee.

Cellulose. The cellulose of coffee berries is of a very hard and horny character, and said by some not to be converted into sugar by boiling with sulphuric acid. We found, however, that by treating the cellulose with a sulphuric acid solution of the strength of 1 in 2, a quantity of sugar was obtained, as indicated by a standard copper solution, equivalent to nearly 36 per cent. of the quantity of coffee analysed.

Ash.-The following table, which is taken from the Coffee Report by Messrs. Graham, Stenhouse, and Campbell, exhibits the composition of the ash of seven kinds of coffee:

The structure of the coffee berry is comparatively simple.

FIG. II.-TRANSVERSE SECTION. FIG. 12.-LONGITUDINAL SECTION.

Transverse and longitudinal sections of it are shown in Figs. II and 12. The husk is represented at A, the parchment at B, the skin at c, and the bean at D.

The bean is the only portion that is imported into this

country, the three outer membranes being removed in the preparation of the coffee for market. A portion, however, of the membrane, called the skin, is always more or less entangled in the folds of the bean, and is present in variable proportions in all genuine coffee. We have, therefore, to deal only with the microscopic structure of the skin and bean.

The skin is a thin delicate membrane, lined with peculiar

spindle-shaped bodies, lying in rather a confused manner, and forming a kind of plate, as shown at A in Fig. 13. These cells are very thick-sided, and have a central canal, their edges presenting the appearance of being marked by minute transverse bars. When more highly magnified, the bars are seen to resemble open ducts, as represented at c.

The structure of the bean is represented at B. It is composed

entirely of a mass of quadrangular, hexangular, and irregular cells, with very thick walls or sides enclosing oil globules. These walls are somewhat sinuous, and are characterised by a quantity of irregular projections on their sides, which, when examined in the fresh state, present a luminous appearance, but, when roasted, appear as opaque masses, imparting a rugged character to the cell walls. The projections are due in some cases to very minute starch granules, while in others they appear to belong to the solid sides of the cell walls. These are the only vegetable structures that occur in commercial coffee. The tissues are not altered by the roasting, and, excepting that they are partially charred, still preserve their characteristic structure. The few starch granules which are present in the fresh bean are, however, entirely destroyed.

ANALYSIS.

Fat. One hundred grains of dry and finely rasped or ground coffee are repeatedly exhausted with ether. The ether is evaporated in a weighed beaker, and the weight of the dry residue ascertained. The non-fatty substances, which may have been also removed by the ether, are dissolved out with boiling water. The aqueous portion is evaporated and weighed, and the amount deducted from the first weight. The difference is the proportion of fat present in the coffee.

Caffeine. The caffeine is estimated in the same way as that described under "theine" on page 16.

Caffeic or Caffeo-tannic Acid.-One hundred grains of coffee are exhausted with alcohol. The alcohol is evaporated, and to the aqueous residue subacetate of lead is added. The precipitate is thrown on a filter, washed and decomposed with sulphuretted hydrogen. The filtrate from the plumbic sulphide is evaporated to dryness, when the caffeic acid is obtained as a yellowish brittle mass.

A larger amount of astringent matter is obtained by extracting, with ether in presence of phosphoric acid, the non-fatty portion of the alcohol extract of the coffee. We have obtained by this means from 8:4 to 9'5 per cent. in the raw coffee.

Sugar. The estimation of the sugar in coffee is sometimes made by the copper test described on pages 56 and 106, but the quantity is more frequently determined by what is termed the fermentation test. In applying the latter method it is usual to take from 1,000 to 2,000 grains of the sample, exhausting at least four times with hot water, and adding the extracts together. To the solution 250 grains of pressed yeast are then added, and the whole set to ferment for 48 hours, at a temperature of from 70° to 80° F. 250 grains of the yeast are next added to a quantity of distilled water equal in bulk to the coffee extract, and fermented under like conditions. The solutions are then distilled, and the specific gravity and bulk of the distillates taken. The amount of alcohol in the yeast solution is deducted from the amount found in the coffee extract, and the sugar calculated from the remainder, as described on page 108, under the head of "Sugar."

Albumin. The albumin is determined by making a nitrogen combustion of the dry residue of the coffee berries after they have been exhausted with ether, alcohol, and water. The soluble albumin is found with the dextrin and gummy precipitate given by alcohol in the water extract, and is estimated by finding the amount of nitrogen, and multiplying the product by 6'3.

Gum or Dextrin.-The ground coffee is exhausted with ether, then with alcohol, and the residuet reated several times with boiling water. The aqueous extract is boiled for some time to precipitate albumin, filtered, and to the filtrate is added 200 grains of concentrated hydrochloric acid. The solution is now treated with alcohol to precipitate the gum. The precipitate is thrown on a tared filter washed with alcohol, and then dried and weighed.

Cellulose. This is estimated indirectly, as in the case of the cellulose of tea. The coffee is treated successively with ether, alcohol, and water, and the residue is afterwards dried and weighed. From the dry residue thus obtained is deducted the amount of albumin and ash found therein, and the remainder represents the percentage of cellulose in the coffee.

E