and of a theine-like alkaloid, from the subjoined varieties of

Theobromine, when quite pure, crystallizes in well-defined rhombic prisms. It is almost insoluble in benzol and petroleum spirit, very slightly soluble in cold water, alcohol, and chloroform, but easily soluble in boiling water and alcohol. It contains 31'1 per cent. of nitrogen, being richer in this substance than theine, which it otherwise closely resembles. Its subliming temperature has been fixed at about 554° F. (290° C.), but it has been found that some of the theobromine is volatilised below that point. It forms, with acids, well-defined crystalline compounds, among which may be mentioned the nitrate and hydrochlorate. Like theine, it forms murexide by the action of nitric acid and ammonia.

The alkaloid which has been found associated with theobromine crystallizes in fine silky needles resembling theine, with which it also corresponds in its solubility in benzol. It is present in both kernel and husk-in Trinidad cocoa to the extent of o°25 per cent. of the kernel, and 0.33 per cent of the husk; but Surinam nibs yielded only o'02 per cent. of this substance. contains 25'48 per cent. of nitrogen.

It

Cocoa Red. This substance gives the characteristic colour to cocoa. It does not appear to be present in the freshly-gathered bean, but arises from the oxidation of the astringent principle or natural tannin of the cocoa berry. This may account for the variable proportions of this substance found in different cocoas. It has the character of a resin, and is partly soluble in hot water,

but more soluble in alcohol. The different degrees of solubility may probably coincide with the extent of oxidation it has undergone. A portion of this substance remains insoluble in the cocoa, even after treatment with boiling alcohol and water.

Ash.-In four descriptions of cocoa-nibs we have found the percentage of ash to vary from 4'47 to 5'71, but the amount of inorganic matter in the husk is much higher. In the husk of the finest Trinidad beans we have found 10'19 per cent.

A complete analysis of the ash of cocoa will be found on page 86.

MICROSCOPIC STRUCTURE.

The cocoa bean has been already described as consisting of a husk or shell, and a kernel, the latter, when broken into fragments, forming the cocoa nibs of commerce. In Fig. 23, at A, B, and C,

G

is given the microscopic structure of the tissues of the husk, and at D and E that of the tissues of the kernel.

The outermost tissue of the husk represented at a consists of two layers of elongated quadrangular cells, one of which layers lying upon the other in a slanting direction imparts to the tissue a striated appearance. Towards the ends of the bean the cells of this tissue become smaller and somewhat hexangular. The second tissue, seen at B, constitutes the main substance of the husk, and consists chiefly of large circular or slightly angular cells containing mucilage and a considerable number of sphæraphides. Several rows of spiral vessels run through this tissue from end to end. The third tissue, represented at c, is very characteristic, and consists of small hexangular cells, some of which are filled with dark granular matter.

The kernel consists of a number of lobes irregular in size and shape, each of which is enveloped in a membrane composed of angular cells filled with granular matter, as seen at D. The substance of these lobes consists of a mass of hexangular cells, represented at E, containing small round starch granules, having a central hilum.

ANALYSIS.

Fat. This is best estimated by the use of petroleum spirit, or benzol. Fifty grains of the dry cocoa are bruised in a mortar, with the solvent employed, and the dissolved fat carefully separated by the aid of a good filter-paper. This operation is repeated several times, or until all traces of fat are removed. On the evaporation of the benzol, the fat is obtained as a perfectly white residue. We have obtained by this means from 470 to 530 per cent. of fat from the cocoa nibs.

Starch. In the estimation of the starch and other constituents of the cocoa, the non-fatty portion is used which is insoluble in the benzol and left as an almost impalpable powder. 50 grains of

this powder are exhausted with alcohol; the insoluble residue is then dried and boiled in water for four hours, or until all the starch is rendered soluble. The portion insoluble in boiling water should not give any colouration with iodine. The dissolved starch is boiled for 6 hours, with 20 cubic centimetres of normal sulphuric acid. Basic acetate of lead is then added, the filtrate decomposed with sulphuretted hydrogen, and the sulphide of lead removed by filtration. The filtrate or solution is next made up to a given bulk, and the sugar formed by the action of the sulphuric acid on the starch estimated by Fehling's solution. From the proportion of sugar obtained the amount of starch is calculated. As cocoa contains matter which very readily reduces copper solutions, the precipitation with basic acetate of lead must always be resorted to. Any estimation of sugar performed without this precaution will give fallacious results.

Albumin.-The nitrogen in the portion insoluble in boiling water is estimated by combustion with copper oxide, and the percentage thus obtained multiplied by 6'3, gives the amount of albumin. The proportion of nitrogenous matter thus obtained is only about one-half the entire quantity present in the cocoa bean, exclusive of theobromine; the remainder is dissolved out by alcohol and boiling water. The amount of nitrogen in the soluble portion is also estimated by copper oxide, but its conversion into an equivalent of albumin by the usual formula may, in the uncertainty attaching to the composition of some of these nitrogenous bodies, give rise to an erroneous result.

Astringent Matter.-This is estimated by digesting the nonfatty cocoa first with alcohol and then with boiling water. Keeping. the two filtrates distinct, the latter is boiled down to a small bulk and alcohol added to it. Any precipitate that falls is filtered off, and the filtrate added to the primary alcoholic solution. To the mixed alcoholic solutions is next added acetate of lead, and the precipitate that then falls is separated, washed, diffused in water, and the lead precipitated by passing a current of sulphuretted hydrogen

through the liquid and filtering off the lead sulphide. The clear and colourless filtrate is then evaporated to dryness, and the residue weighed, the ash being subsequently determined and deducted from the total dry residue. As the solution evaporates it assumes a bright red hue, and part of the colouring matter of the residue becomes insoluble in water.

Gum.-Five grams of the cocoa are exhausted first with benzol and then with alcohol. The residue is next treated with cold water for two hours, and filtered; the aqueous filtrate is evaporated to a small bulk, filtered, and to the filtrate alcohol of 90 per cent. is added to precipitate the gum. The precipitate is thrown on a filter, washed with alcohol, and re-dissolved in a small quantity of water. The gum is again precipitated with alcohol, dried, and weighed on a tared filter. The amount of mineral matter in the gum is determined by ignition, and deducted from the total weight.

Cellulose. After the starch has been rendered soluble, and the solution filtered off, the residue is boiled with 20 cubic centimetres of normal sulphuric acid for six hours, and the resulting sugar estimated with the precautions observed in the case of starch. In addition to the direct estimation before described, the starch may be determined by difference—that is, by deducting the amount equivalent to the cellulose from the total sugar formed after conversion of the starch and cellulose in the part insoluble in alcohol.

Theobromine.-One of the earliest methods adopted for the separation of this alkaloid very much resembles that employed for the estimation of caffeine. An alcoholic extract of the bean was made, the alcohol evaporated, basic acetate of lead added, and the precipitate separated. Sulphuretted hydrogen was then passed through the filtrate, the precipitated sulphide of lead separated, and the solution evaporated nearly to dryness. Finally, the mass was extracted with alcohol, the filtrate partly evaporated, and then left