sugar-cane. Difficulties which were then considered insuperable have, however, since been overcome, and the purest loaf-sugar is now found in our markets prepared from beet. In 1858 Johnston calculated the total beetroot sugar manufactured in the world at 159,821 tons. In 1869 it was estimated at 689,500 tons for Europe alone; in 1877-78, at 1,420,800 tons; in 1878-79, at 1,574,100 tons; in 1879-80-there being a failure in the crops— 1,275,000 tons.

The production of glucose from starch was first accomplished by M. Kirchoff, of St. Petersburg, in 1702. He recommended boiling the starch with 1 per cent. of sulphuric acid for thirtysix hours. The manufacture had been carried on for some years on the Continent before its introduction into this country, and a liquid form of it is imported and used to some extent by confectioners. The solid form is, however, chiefly used by brewers, and comes into direct competition with the glucose made in this country by one of the methods we have described.

CHEMICAL COMPOSITION.

Pure cane-sugar has been analysed by many chemists. One of these results is as follows:

This analysis agrees very closely with the recognised formula for sugar, CH2O. The specific gravity of the crystals of pure cane-sugar, according to Joule and Playfair, is 1593 at 39° F. (4° C.) It melts at about 320° F. (160° C.) without losing weight to a clear pale yellow liquid, which is a mixture of dextro-glucose and levolusan.

Heated to 410° F. (210° C.) sugar is converted into a dark brown substance called Caramel, which is extensively used for colouring liquids.

Cane-sugar, unlike glucose, is decomposed by strong sulphuric acid, with copious formation of carbonaceous matter. It is not turned brown when treated with alkalis, and is insoluble in cold absolute alcohol. It dissolves in one-third of its weight of water at mean temperature, and in all proportions in boiling water. From a solution containing 5 parts of sugar to 1 part of water, threefifths of the sugar crystallizes on cooling in four or six sided rhomboidal prisms.

ESTIMATION OF SUGAR.

Cane-sugar does not precipitate the suboxide of copper from alkaline solutions of cupric tartrate, but it is very readily converted by boiling with dilute acid into invert-sugar, which does possess that property. Advantage is taken of this in what is generally called "Fehling's test." A solution is made by dissolving 86 grams of tartaric acid in crystals with 104 grams of caustic soda. To this is added 29 grams of sulphate of copper dissolved in water. The bulk is then made up by additional water to 1 litre. This is Fehling's solution, and in its application for the estimation of sugar it may be used either "volumetrically" or "gravimetrically:" in either case it is necessary in the first place to have a standard. In the volumetric process, which is the easier, 625 gram of pure cane-sugar is for this purpose boiled for ten minutes with about 4 ounces of water acidulated with 5 drops of concentrated sulphuric acid. The solution is then cooled, neutralised with solution of caustic soda, and made up to a bulk of 250 cubic centimetres. Twenty-five cubic centimetres of the copper solution are then heated in a white glass flask to the boiling point, and the sugar solution is run into it from a burette, care being taken not to add more than will reduce the whole of the copper. It will generally be found that 40 cubic centimetres of the sugar solution, which correspond to 'I gram cane-sugar, or 105 gram glucose, will be required to reduce the copper or decolourise 25 cubic centimetres of copper solution. If more or less than

40 cubic centimetres are required, a corresponding difference will have to be made in the quantities of cane-sugar and glucose represented respectively. This result is applied in the examination of saccharine substances or solutions in the following way: If a known weight-say 8 grams—of a liquid which contains glucose and cane-sugar be taken and made up to 250 cubic centimetres, and if it be found that 45 cubic centimetres of this diluted solution are required to reduce the copper in 25 cubic centimetres of Fehling's solution, the percentage of glucose is thus found:

250 x 100 x 'I -- 6'94 per cent., the cane-sugar equivalent,

45 × 8

or 7.30 per cent. glucose.

It is then necessary to make a second experiment to find the total amount of sugar present. A less weight than beforesay 4 grams-is taken and boiled for four minutes with about 4 ounces of water and 5 cubic centimetres of normal sulphuric acid to invert the cane-sugar. It is then neutralised with soda and made up as before to 250 cubic centimetres at 60° F. (15·5° C.), and if it be then found that 50 cubic centimetres of this solution are necessary to reduce the copper in 25 cubic centimetres of Fehling's solution, the total sugar in the liquid, calculated as cane-sugar, is as follows:

250 × 100 × 1 =12'5; and 12'5-6′94 = 5'56, the percentage of

In the gravimetric method, the standard is the quantity of cuprous oxide precipitated by a given quantity of sugar-solution. The cuprous oxide being either ignited with a little nitric acid, and weighed as cupric oxide, or, as recommended by Pavey, the suboxide of copper is dissolved, and the copper precipitated from it by electrolysis and weighed. Fehling's test, although fairly accurate, where the percentage exceeds o ̊5 per cent., is not well adapted for cases in which it falls below that quantity. Knapp's method, based upon the decomposition of an alkaline solution of cyanide of mercury, has been suggested where the quantity of sugar is very small. The standard liquid is

prepared by dissolving 10 grams pure dry mercuric cyanide in water, adding 100 cubic centimetres of sodium hydrate solutionspecific gravity 1145, and diluting to 1,000 cubic centimetres. Ten cubic centimetres of this solution are equal to 25 milligrams of glucose. To apply the test, 10 cubic centimetres of the solution, diluted with from 20 to 30 cubic centimetres of water, are heated to the boiling point. The sugar solution is run in from a burette until the whole of the mercury is precipitated. When the precipitate has settled, a drop of the supernatant liquid, which has a more or less yellow tint, is transferred by means of a capillary tube to a thin pure white Swedish filter-paper. This paper is held, first over a bottle containing strong hydrochloric acid, and then over a saturated sulphuretted hydrogen solution. The slightest trace of mercury is shown by the production of a light brown or yellow stain. It is well to place a drop of the original liquid beside that which has been subjected to the action of hydrochloric acid and sulphuretted hydrogen for comparison.

Cane-sugar, when exposed to the action of yeast, is rapidly changed, first into invert-sugar, and then into alcohol and carbonic dioxide, the following being the reactions:

A process based upon the quantity of alcohol produced by fermentation from a given quantity of sugar has been long in use for estimating the percentage of sugar in substances to which, owing either to their colour or to the fact that they contain matter other than sugar capable of reducing salts of copper, Fehling's method cannot be applied. In determining the amount of sugar by the fermentation process, the quantity taken, in order to insure complete fermentation, should not exceed 100 grains. Assuming that 100 grains of the sample to be analysed, when dissolved in about a quart of water and fermented with 200 grains of pressed yeast, yield a distillate of 1,000 fluid grains of a density of 990*3

at 60° F. (15.5° C.), and that 200 grains of yeast similarly fermented yield a distillate of the same volume, having a density of 998.3, the following calculations will give the percentage of canesugar or glucose present. By Gilpin's tables it will be found that mixtures of alcohol and water of a density of 990°3 and 998*3 contain respectively 5.52 grains and 88 grain by weight of absolute alcohol in each 100 fluid grains, and therefore 55°2 and 8.8 grains respectively will be contained in each distillate. Deducting the latter from the former, there femain 46'4 grains of absolute alcohol as having been produced from the sugar. By the equations given above 342 parts by weight of cane-sugar, or 360 parts of glucose, are seen to be necessary to produce theoretically 184 parts of absolute alcohol; hence

the percentage of glucose in the substance analysed.

Should the rectifying power of the distilling apparatus used not be sufficient to insure the collection of the whole of the alcohol in the first distillate of 1,000 fluid grains, a second similar bulk must be distilled over, and the amount of alcohol found added to that obtained in the first distillate.

As small quantities of glycerin and succinic acid are formed during fermentation, the amount of sugar, calculated from the alcohol produced, is invariably less than the true quantity, even under the most favourable conditions. It is sometimes desirable, therefore, in practice, to make experiments with pure cane-sugar, and to use the highest alcoholic result obtained as a factor for calculating the amount of sugar contained in saccharine substances submitted to the fermentation test.

Another process is to estimate the sugar from the loss of carbonic dioxide in the course of fermentation. Its application requires very great care in manipulation, and it is not likely to be