considerable difficulties, for until the year 1882 this problem is scarcely referred to in literature. In that year, Jablochkoff* devised the apparatus shown in Fig. 20. Here the salt was introduced through the charging

funnel D into the clay crucible A, where it was melted and electrolysed. The electrodes a

and b were surrounded with the tubes c and c, for the removal of chlorine and sodium vapour respectively.

The practical difficulties which arise in applying such an arrangement to the electrolysis of alkaline chlorides are unquestionably great. They lie chiefly in the designing of a sufficiently durable apparatus. The tube for the removal of the chlorine must not be made of metal; and porcelain, especially in closegrained pieces of the thickness required for the purpose, is very easily broken at the high temperature employed. Then, again, if the tube arranged to convey the metallic vapour to the condenser be made of porcelain or any other material prepared from clay, it will very soon become perforated by the action of the alkali metals in the liquid or vaporous condition. But a metal tube is equally unsuitable. Even if the electrode b were insulated from the surrounding tube, the metal sheath would be of little value; because, granted even that it might be possible, by the use of sufficient heat, to maintain so high a temperature that the separation of fluid metal-and therefore the formation of a metallic contact between b and c-could be avoided, there would necessarily be a deposition of sodium on the outside of the tube throughout the process. This results from the fact that insulated metallic substances placed in the electrolyte tend to form intermediate electrolysing surfaces between the poles, the side of the metal next the anode functioning as a cathode, and that next the cathode acting as an anode. The consequence of this is that intermediate reactions are produced, which cannot but be prejudicial to the working of the process as a whole. Under these circumstances the durability of the metallic sheath is naturally of the shortest. Later modifications of the original construction, consisting in the arrangement of the electrodes and of their separating partitions in the form of concentric tubes, have failed to overcome the difficulty.

Hoepfner's Process.-In the year 1884, C. H. W. Hoepfner t Dingler's Polytechnisches Journal, 1884, vol. ccli., p. 422.

+ German Patent 30,414.

patented a process, of which the following is an account derived from his specification:- "Sodium chloride is melted in a crucible, the bottom of which is covered with a layer of copper on silver. Instead of the latter, other heavy metals may be employed, with the exception of mercury, which boils at too low a temperature. The side-walls of the crucible must be made of some non-conducting material. The layer of metal at the bottom of the crucible serves as an anode, and it is, therefore, connected with the corresponding pole of the galvanic battery, or of the dynamo, by means of an iron or copper wire introduced either from below or at one side. If the circuit be now completed by immersing a carbon or metal cathode in the fused chloride from above, a rapid deposition of metallic sodium at once sets in; the sodium burns in the presence of air, but if the latter be excluded, the metal may be collected or distilled. While the light metal separates at the upper surface, the chlorine attacks the metal of the anode, forming a chloride which fuses at the high temperature of the operation, so that by continuing the current the anode undergoes electrolytic solution; but on account of its weight the chloride remains at the bottom of the crucible."

Hoepfner must have found that by continuing the current the course of the separation did not run so smoothly as was the case at the beginning, for the metallic chlorides do not remain so quietly at the bottom of the vessel as the patent specification suggests. With the density of current required for the deposition of sodium, the copper, silver, or other metal dissolved at the anode would, in a short time, be re-deposited at the cathode in such quantities, and in so slightly coherent a condition, that, in spite of every precaution, it would be washed off again by the fused salt, and (especially in this form of apparatus) would be returned to the anode. Not only would there be no chlorine evolution, which the invention aims at avoiding, but in a short time there would be no deposition of alkali metal either.

The

Rogers' Process.-A. J. Rogers* has patented the apparatus shown in Fig. 21, which is a longitudinal sectional elevation of the whole apparatus along the line y y (Fig. 22), and in Fig. 22, which is a plan of one portion of the apparatus along the line xx (Fig. 21). A' shows the masonry, B the fire-grate. melting-pot, C, is provided with a cover, C', which may be fastened down; on the middle of this cover is a funnel, c', which closes automatically by a ball-float valve, c, when the crucible is sufficiently charged. The safety-valve, C2, is provided to guard against risk from a sudden evolution of gas. The decomposing-vessel, D, is divided into two compartments by the wall, d, which is made of some porous material, such as fire-brick, and reaches nearly to the bottom of the chamber.

The tube, E, forms a connection between C and D, and a cock, * United States Patent 296,357, April 8, 1884.

e, serves to regulate the flow of melted material from C to D. N is the negative, and P the positive electrode; the wire connections for the electrodes within the cells are insulated, and at the same time protected from the attack of acid, gases, &c., by the fire-clay tubes n and p, in which they are respectively enclosed. The tightly-fitting cover, d', is connected by two necks, F F', with separate condensers, and is fitted with four closed doors at the points ff and f'f' shown in Fig. 22. The

Fig. 22.

Rogers' apparatus for the extraction of sodium.

separating partition, d, must be gas-tight even up to the top of the roof D'. Through the roof of the negative compartment is passed the pipe, K, with the valve, k, by which hydrogen or other reducing gas may be introduced. The vessel, D, is set in thin -masonry, and must be constructed of a non-conducting and fireresisting material, and should be about 1 ft. wide, 3 ft. long, and 4 ft. deep. The decomposing vessel bears a strong resemblance to that of Watt's, but it is not so faulty in the details of its construction. Nevertheless, it is scarcely to be expected that it could be advantageously used for the reduction of the alkali

metals; and the material of which it is made is incompatible with great durability. The distillation of sodium in clay vessels is an undertaking of very doubtful expediency. The iron pre-heating vessel, C, filled at a strong red heat with melted salt, would, by its own weight, become unserviceable after a few hours. To attach the tube, E, sufficiently firmly to the clay vessel, D, and to maintain a tight joint when the whole apparatus is charged with fused sodium chloride at a red heat, would be no easy task. Rogers has, meanwhile, had an opportunity to discover the faults of the apparatus.

The

Omholt's Apparatus.-Omholt's plant* (Figs. 23-26) for the continuous production of the light metals, appears also to belong to that class of inventions which are first patented and then, when opportunity offers, are put to the practical test. The patentee writes: "The apparatus shown in longitudinal section (both in elevation and in plan), in Figs. 23 and 24, consists of a reverberatory furnace, with the hearth divided into two separate compartments by the bridges, a a. In each of these compartments are two half retorts, b and c, borne horizontally and parallel to one another, on supports, d d, of fire-resisting material, placed side by side, in such a way that the half retorts are separated by a short space from the sole of the hearth. half retorts marked bb contain the negative electrodes, e e, whilst those marked c c enclose the positive electrodes, ff. The halogen compound to be electrolysed lies melted on the hearth at such a height that the electrodes in the half retorts are completely immersed. The metal separating at the cathode, and the halogen set free at the anode, collect within their respective half retorts, and are there protected from contact with the furnacegases, owing to the liquid seal which is formed by the dipping of the half retorts into the melted charge. The half retorts are made of a fire-resisting and non-conducting substance, and are coated on the interior with a lining of carbonaceous material. The latter material consists for the most part of carbon (graphite or wood-charcoal); the mixture used, for example, in the manufacture of plumbago crucibles being suitable. The admixed clay serves only to give plasticity to the carbon.

"The electrodes are made of carbon or other resisting material. They lie lengthwise within the half retorts, as shown in the illustration, whilst at the one side (at g in Fig. 25) they penetrate the masonry of the furnace, and there make the necessary connections for the circulation of the electric current. To prevent the separation of halogen or metal on the electrodes outside the half retorts, the exposed portions between the face of the retorts and the brickwork of the furnace are insulated by a covering of fire-clay. By this construction, I [the patentee] allow for the * German Patent 34,727, June 6, 1885.

Fig. 25.-Omholt's Sodium Plant (cross section).

renewal of broken half retorts without damaging the electrodes or altering their position.

"The half retorts, b b, are connected with the fire-resisting fore