introduced into the melting vessel as negative electrodes, as described with reference to Fig. 11. 3. To diminish the electric resistance within the apparatus first described, as well as for the enrichment of the smelting bath as it becomes impoverished, the use of plates or rods, M (Figs. 9 and 11), of alumina and carbon, or of magnesia and carbon, in equivalent quantities respectively, which should be placed within the vessel, G, independently of the electrode.

The only actual novelty in the whole patent was the introduction of the plates or rods of magnesia and carbon, or of alumina and carbon, to be used in the production of magnesium and aluminium respectively, and this has not proved successful; its use was discontinued in the aluminium and magnesium factory at Bremen even at the time that Graetzel was managing the works. The effect of these plates, like that of Deville's oxidecarbon anodes, was only to introduce impurities into the bath, and thus to cause irregularity and inconvenience in the working. In order to bring about the desired reaction, as formulated in the equation

MgO + C + Cl2

=

MgCl2 + CO

the temperature of the bath would have to be maintained much higher than is desirable either for the production of magnesium, for the preservation of the apparatus, or for preventing the destruction of the plates themselves through the removal of the oxide.

The introduction of reducing gases into the cathode chamber was made a prominent claim in the specification. But this proposal had been made a year earlier by F. Fischer, as we have seen above. Hiller* also had arranged for the use of a reducing atmosphere in the cathode chamber of his apparatus for the reduction of strontium and lithium. Moreover, this precautionary measure is entirely unnecessary in the electrolysis of carnallite. Practically the whole of the magnesium remains adhering to the immersed electrode and to the walls of the crucible, provided that a current of suitable strength be used and that the fused salt be not connected up with the battery until it has been melted sufficiently long, and at a sufficiently high temperature, to drive off the last traces of water, which are retained by it with the most obstinate pertinacity. The use of reducing gas is not, therefore, necessary on this account; but if the crucible be immersed in the fire almost to the top rim it will be found that in the absence of reducing gases the walls of the vessel become strongly corroded above the level of the fused mass within. This is due to the metal of the crucible becoming chloridised, owing to the combined effects of the hot fire-gases without and the acid-gases within, the latter being always present above the surface of

* F. Hiller, Lehrbuch der Chemie, 1863.

carnallite when it is melted in the presence of even a small proportion of air. The crucible walls, therefore, above the level of the fusion become rapidly corroded through, whilst, in addition, the melted carnallite tends to creep over the rim of the vessel on to the outer surface, and there, aided by the furnace gases, it exerts a most destructive influence. This action may, however, be prevented by adopting the construction of crucible used by Borchers. This crucible (Fig. 12) is provided, at a distance of about 2 to 2 in. from the top, with a flange, which serves to support it with its lower portion im

mersed in the fire. The upper part of the crucible being thus exposed freely to the air remains sufficiently cool to reduce to a minimum both the formation of acid gases and the corrosion of the vessel. Thus the furnace gases come in contact only with those portions of the crucible walls which are in contact with fused substance within, and which serve as cathode surfaces. The tendency of the melted salts to creep up the side of the crucible is completely prevented because, even if the uppermost margin of the fused mass within the crucible should not be crystallised in contact with the comparatively cool walls, yet any thin layer of fused salt that might reach above this level would be at once solidified.

Finally, the use of the melting vessel as cathode, claimed by Graetzel as his invention, was anticipated in actual practice by Davy so long ago as the year 1808.

Napier's Electrolytic Vessel.-An account may be given here of an apparatus patented by Napier* in 1844. It was originally intended for the extraction of copper, but, as in the case of Graetzel's and other apparatus, several elements may be treated in it.

Napier used a large crucible or other convenient vessel made of son e conducting material, of which the inner side, with the exception of the bottom, was lined with a coating of clay. The copper ore, roasted as free as possible from sulphur, was mixed with the necessary fluxes and fused in the crucible; the melted mass was then submitted to the action of the galvanic current in such a way that the crucible itself formed the cathode, whilst the anode was an iron rod united to a plate at the lower end. It is thus clear that even in 1844 an apparatus for the reduction of metals had been patented,† which consisted of a melting vessel

English Patents 10,362, 1844, and 684, 1845. Cf. Houston, Journ. Franklin Inst., 1889, vol. cxxv., p. 376.

+ Cf. Graetzel's Patent claims, pp. 23, 25.

made of some conducting material and serving as the negative electrode, in combination with a non-conducting and fire-resisting vessel, open at the bottom, and surrounding the positive electrode for the purpose of insulating the electrodes and facilitating the removal of the anion.

Borchers' Magnesium Process.-In accordance with the principles explained above, and following the pattern of Troost's*

Fig. 13.-Borchers' experimental furnace for the electrolysis
of fused salts.

apparatus, the cell surrounding the carbon electrode being reduced in size, whilst at the same time taking into account the experience which led to the adoption of the shape of crucible depicted in Fig. 13, the author has designed an apparatus suitable for a large number of experiments in the electrolysis of fused salts. Within the iron crucible, K, which forms the cathode, is suspended the carbon rod, A, as anode, surrounded by the porcelain

tube, C. The anode is connected to the conducting wire from the dynamo by means of the screw clamp, V, and is supported by the ring-shaped porcelain cover, L, whilst the porcelain tube rests by means of a collar on another annular porcelain plate, d; the weight of the whole crucible is in turn borne by the firebrick cover, D, of a Perrot furnace, upon which it is held by the flange, F; and this, by an extension on one side, makes electrical connection with the negative pole of the dynamo through the binding screw, N. The furnace consists of a wide fire-clay cylinder, O, protected externally by a sheet-iron jacket, M, which may be supported either by feet attached to it, or by a temporary stand; and the whole is closed beneath by a fireclay plate, B, provided with a central opening. The internal fireclay cylinder, W, serves to convey the heated gases from a sufficiently powerful gas-burner, first of all upwards and around the crucible, and thence downwards through the annular space between W and O to the flue, Z.

In using the apparatus, the empty crucible and its appurtenances should first be heated for a considerable time, while the carnallite is being melted, preferably, in a second crucible. During the heating of the former, a piece of charcoal may be placed in the crucible to prevent the oxidation of the inner surface, which should be perfectly clean, and to minimise the loss of anode material by combustion; but the charcoal must of course be removed before introducing the melted salt.

During electrolysis at a dull red heat (about 700° C.), the magnesium deposits upon the walls of the vessel in globules, which rapidly increase in size, while the chlorine, mounting upwards through C, finds an escape by the side tube, R.

Working with a current density of at least 1000 amperes per square metre [0.64 ampere per sq. in.] of cathode surface, the current density at the surface of the anode will amount to about ten times as much, even if an unusually thick carbon rod be employed. In spite of this, an electromotive force of only 7 to 8 volts is required, which may be reduced to 1 to 2 volts (where an economical installation is sought) by increasing the size of the anode.

After continuing the electrolysis for a sufficiently long time, an inspection of the contents of the crucible through the clear fused salt will show when the required quantity of metal has been accumulated. The current is then broken, the screw connections are unfastened, and the cover, d, together with all that it supports, is removed from the melting vessel. The flame

is now somewhat increased, and the mass of metal clinging to the walls is detached with the aid of an iron scraper shaped to correspond with the inner surface of the crucible. The whole contents of the vessel are then poured into a flat box made of sheet iron, which must be cold and perfectly dry, and any metal

that still adheres to the walls of the vessel must be rapidly scraped off. The solidified melt, after cooling, is broken up, and the globules of magnesium are picked out. The larger pure shot may be melted together without flux in a plumbago crucible, but the less pure metal must be fused and refined by a process which will be described later.

The apparatus just described may be used for operations on a manufacturing scale almost without any alteration other than that necessitated by the proportional enlargement of the various parts. Under these circumstances it will only be necessary to

discuss briefly the special points to be observed in the use of a plant of this nature.

After the above description, Fig. 14 requires no further explanation, so far, at least, as concerns the various parts of the apparatus. In a plant of about twenty times the size of the figure the melting vessel and its appurtenances are suitable to the application of a current of 250 to 300 amperes, which corresponds to an outturn of 0.199 to 0.238 kg. [0.44 to 0.5 lb.] of magnesium per hour. It is not advisable to use an apparatus much larger than that which is here figured, not