At this stage the cost of 1 lb. of glycerine is about 1 1/2d. The liquor is now ready for osmosis, by which process the ashes it contains are so far reduced that after further evaporation it can be distilled either by itself or with crude glycerine from stearine works. The loss on distilling is small, and the product pure enough for dynamite works. When evaporated to 1.26 specific gravity, and therefore free from water, it contains no sodium chloride. The lyes do not attack the parchment paper, as they contain no lime compounds, the paper having been found in good condition after 6 months' usage. The patentee calculates the cost of 1 lb. of 60 per cent, glycerine at 2 3/4 d. The water of osmosis contains a considerable quantity of glycerine. In one experiment, 1400 lb. of the water gave 23 lb. liquor, containing 16 lb. glycerine, at a cost of about 2 3/4 d. per lb. This product, contaminated with salt, and therefore not as pure as that obtained in stearine works, was still found as pure as a common Dutch glycerine of 63 per cent, which cost Is. per lb.

In the patentee's opinion, the proposal of using only fatty acids in soapmaking, and of decomposing the fats for that purpose in an autoclave, promises but little success.

In a subsequent patent, Fleming obviates the diffusion of the glycerine by replacing the ordinary parchment paper by a guttapercha membrane, which is proof against the passage of glycerine.

O'Farrell evaporates the soap lyes to the saturation point of sodium chloride, and uses it again to precipitate fresh soap. This is repeated till the lye is very rich in glycerine. The solution is next evaporated till all the salt crystallizes out, and is then distilled in vacuo, with a steam jet at about 392° F.

(200° C). The distillate is heated a second time with low surface-heat only, to separate the water as steam.

Dr. Versman, of New Charlton, separates a large percentage of the salts in soap lyes by simply boiling down the soap-lye and raking out the salts as they become insoluble; he then allows the concentrated solution to cool, after which carbonic acid gas is passed through it until the whole of the carbonate and caustic soda is converted into bicarbonate of soda, which is much less soluble in glycerine than either the carbonate of soda or the caustic soda, and may readily be removed by filtration or other convenient means. The liquid from which the bicarbonate of soda has been removed is rich in glycerine, but it still retains sensible quantities of chloride of sodium and other salts, the presence of which may act injuriously in the subsequent application of the glycerine to certain purposes. These salts the patentee separates by submitting {he liquid, either hot or cold, to the process of osmosis in an apparatus known as the "osmogene," such as is used in the separation of saline compounds from solutions of beet-root sugar; by this process nearly all the salts are separated from the glycerine; but as the latter becomes diluted with water, it is concentrated by evaporation, when it will be ready for the market as crude glycerine.

Instead of commencing the treatment by boiling down the lye, carbonic acid gas may be passed through the original soap-lye, or, if preferred, the use of carbonic acid gas may be dispensed with; but the patentee finds the most convenient and economical way is to first reduce the bulk of the liquid by boiling down, so as to separate large quantities of the salts, and then to treat the liquid with carbonic acid; the subsequent treatment in the osmogene is essential in all cases. This osmogene is a modification of Graham's dialyser, and consists of 40 or more cells formed by sheets of parchment paper laid flat and connected at the edges all round, the space between each pair of sheets being fully 1/2 in.

Each sheet is supported by a cross piece of wood and a network of string, and the whole affair is about 4 ft. long by 3 ft. high. Water is sent through alternate spaces, and the substance from which the .salts are to be extracted through the others, the water by osmosis carrying off the salts and leaving a small quantity of itself in the glycerine or other substance.

F. H. T. Allen's process is as follows: First, if necessary, he neutralizes the soap lye with any ordinary mineral acid, and agitates; after settling, he adds a solution of alum, or solid chloride of lime (bleaching-powder), or crude pyrolig-neous acid, and stirs thoroughly; or evaporates to nearly salting point, before adding any of the substances named above. He allows the precipitate to fall, leaving a clear liquor, and after settling, draws off the clear upper liquor, and evaporates to concentration in pans (in which the heat is only applied at the sides). Finally, he distils this liquor in a glycerine retort, having a current of superheated steam within, and provided with an exit pipe at the bottom, which carries off the precipitated salt as it accumulates.

J. P. Battershall, New York, heats the liquors with steam, neutralizes with sulphuric acid, and concentrates to 1/10 their original bulk. A slight excess of acid is added, if necessary, to separate the resin and fat. Carbonate of lime is then added to again neutralize the liquor; the latter is cooled, and the liquor separated from the crystallized salts in a centrifugal machine. The liquid is then distilled to purify the glycerine from the residual salts. Alcohols can be used instead of distillation, the precipitated salts (insoluble in alcohol), filtered or fished out, and the glycerine and alcohol separated by distillation. The crude glycerine can then be decolorized by filtering through bone-black.

Although evaporation and distillation are the usual methods of purifying glycerine, the action of cold upon more or less dilute glycerine is sometimes employed in conjunction with them, especially by Sarg, at Vienna. When an aqueous solution of glycerine partially freezes, the frozen mass contains more water than the remaining liquid; hence some amount of concentration may be thus effected. The following table gives the freezing-points of such mixtures: -