become filled with the débris of the crushed particles, and a solid block is the result. Finally, it may be urged that this union of powders of solid metals under the influence of pressure-that is to say, the close approximation of the particles-can be compared to the liquefaction of gases by pressure. At the first view this comparison may appear rash or strained, but it is not so if the views of Clausius on the nature of gases and liquids be accepted. In a gas the molecules are free, but if by pressure at a suitable temperature the molecules are brought within the limit of their mutual attraction, the gas may be liquified, and under suitable thermal conditions, solidified. The mechanical pulverisation of a metal merely detaches groups of molecules from other groups, because the mechanical treatment is imperfect, but the analogy between a solid and a gas has, in this sense, been established; filing coarsely gasifies the mass, but pressure solidifies it.

It is possible that in some of the compressed metallic blocks, the particles are not actually united by the pressure, which may, nevertheless, develop the kind of "mutual attraction contemplated by Lord Kelvin as existing between two pieces of matter at distances of less than 10 micro-millimetres.

Occlusion of Gases.-With reference to the absorption of gases by metals, it may be sufficient to point out that Sainte ClaireDeville and Troost discovered that hydrogen would pass through a plate of platinum, prepared from the fused metal or through iron, at a red heat; and it was well known that molten silver had the power of absorbing many times its own volume of oxygen. In Deville's experiments a new kind of porosity was imagined, more minute than that of graphite and earthenware, an intermolecular porosity due entirely to dilatation. Graham showed that when gas penetrates the substance of the metal there is previous absorption and possibly liquefaction of the gas. Since his time it has been abundantly recognised that the presence of an element which is capable of re-appearing with the elastic tension of a gas must materially affect the mechanical properties of a metal. Palladium is known to possess the power of occluding gas-hydrogen-in the most marked degree. By slow cooling from a red-heat in an atmosphere of hydrogen, palladium foil or wire occludes no less than 900 volumes of hydrogen. Similarly, gold is found to occlude, that is, retain when solid, 0.48 of its volume of hydrogen and 0.2 of its volume of nitrogen, silver occludes 0.7 of its volume of oxygen, and wrought copper occludes 0.306 volume of hydrogen.

It is, however, in relation to the metallurgy of iron that the *Proc. Roy. Soc., vol. xvi. p. 422; vol. xvii. p. 212, and p. 500; Trans. Roy. Soc., 1886, pp. 399-439.

occlusion of gases is of importance. It is well known that at the conclusion of the Bessemer process, oxygen from the air blown through the metal becomes intimately associated with iron; but the manner in which the oxygen is held, whether as oxide or as dissolved gas, appears to be still obscure, though Müller* has given strong evidence in support of the view that gases are dissolved in iron. One thing is certain, that the oxygen may readily be removed from the iron by the action of manganese. Hydrogen is usually present in iron, chiefly as gas, sometimes as ammonia,† and in certain cases probably in some non-gaseous state. It does not appear to be in strong chemical combination, as it can easily be expelled. This may happen on solidification of the metal, by heating in vacuo, or by the action of a drill, which appears to release entangled or loosely-held hydrogen. The escape of gas can be prevented by increasing the pressure during solidification, and by the addition, before solidification, of silicon, manganese, or aluminium. The hydrogen probably remains in the cold iron after it is solidified. Cailletet extracted from electrolytic iron, in which the metal probably exists in a distinct molecular form, nearly 250 times its volume of hydrogen by heating in vacuo. Graham proved that carbonic oxide is dissolved by iron, and that that gas probably plays an important part during the conversion of iron into steel in the ordinary process of cementation. It is certain that the presence of silicon and manganese appears to enable the iron to retain carbonic oxide, as well as hydrogen and nitrogen, in solution.

Another analogy between metals and fluids is presented by the power which certain solid metals possess of taking up fluids, sometimes with a rapidity which suggests the miscibility of ordinary fluid substances. In reference to this, an interesting paper was published, so long ago as 1713, by the Dutch chemist Homberg,§"On Substances which Penetrate and which Pass Through Metals, without Melting Them." He enumerates several substances which will pass through the pores of metals without disturbing the particles, and he points out that mercury penetrates metals without destroying them. The rapidity with which mercury will pass through tin is remarkable. A bar 1 inch

wide and inch thick will be penetrated by mercury in thirty seconds, so that it breaks readily, although before the addition

* Iron, 1883, vol. xxi. p. 115, and vol. xxii. p. 244; 1884, vol. xxiii. p. 161. + Recognised by many observers; notably by Regnard, Comptes Rendus, vol. lxxxiv. (1877), p. 260.

+ Comptes Rendus, vol. lxxx. (1875), p. 319.

Mem. de l'Acad. Royale des Sciences, 1713 (vol. for 1739, p. 306).

of the mercury the bar would bend double without any sign of fracture.

With regard to the vaporisation of solid metals, Demarçay* has shown that in vacuo metals evaporate at much lower temperatures than they do at the ordinary atmospheric pressure, and he suggests that even metals of the platinum group will be found to be volatile at comparatively low temperatures. Merget† has shown that the solidification of mercury by extreme cold does not prevent the solid metal evaporating into the atmosphere surrounding it.

In relation to surface tension there is an interesting property belonging to a hard drawn rod or thick wire of 13-carat gold, the gold being alloyed with silver and copper in the following propor

tions:

If such a rod be touched with a solution of chloride of iron, or certain other soluble chlorides, it will, in a short time, varying from a few seconds to some minutes, break away, the fracture rapidly extending for a distance of some inches.

The last property to be considered is diffusion. The author has shown that in the case of molten metals the interdiffusion may be extremely rapid. In regard to solid metals, some experiments conducted by Abel prove that carbon can pass from a plate of richly carburised iron to one of iron free from carbon against which it is tightly pressed. This passage of carbon appears to take place at the ordinary temperature, and it is difficult to explain the transference of matter without admitting the presence of some action closely allied to the diffusion of liquids.

These facts afford additional evidence as to continuity in the properties of all kinds of matter, and serve as a connecting link with the work of the past, the importance of which is too often overlooked.

This chapter may fittingly conclude with a table of the physical constants of metals.

*Comptes Rendus, vol. xcv. (1882), p. 183.

+ Ann. de Chim. et de Phys. (4), vol. xxv. p. 121.

For Table of Physical Constants, see next page.

Notes to the Table of Physical Constants (pp. 58, 59).—For melting points the student must refer to the Physikalisch-chemische Tabellen, by Drs. Landolt and Börnstein, Berlin, 1883,*p. 81, and to papers by Violle (Comptes Rendus, vol. lxxxix. (1879) p. 702), and by Pictet (ibid., vol. lxxxviii. (1879) pp. 855, 1315). The atomic weights and atomic volumes are those given in Lothar Meyer's Modern Theories of Chemistry, London, 1888. With regard to electrical conductivity, it is usual in this country to employ pure copper as a standard of reference for industrial purposes. For scientific purposes, it has been usual to refer to pure silver, and to assign to it the value 1000. In view, however, of the fact that it is very difficult to obtain silver of absolute purity, and that the conductivity varies greatly with the thermal and mechanical treatment to which the metal has been subjected, it has been considered better to adopt mercury at o° as the standard metal, its electrical conductivity being taken as unity. The figures given in the last column of the table may be converted into a series in which silver would be 100, by simply multiplying by 1.75.

NOTE. While this volume is passing through the press, Professor Dewar stated, in a lecture delivered at the Royal Institution Jan. 19, 1894, that when metals are exposed to the very low temperature of 180° C., the tenacity is greatly increased: for instance, steel, which breaks with a stress of 34 tons per square inch, is nearly double as strong at - 180°. The strength of German silver is also nearly doubled. The elongation of the steel does not appear to be changed by this very low temperature.

* There is a later edition of these tables (1893).