entirely disappears as soon as an alloy is solidified, or whether it is not partially maintained even when the alloy has reached its quite solid state.

To answer this question we must, however, cast a glance upon another wide series of investigations into some physical properties of metals.

II

It is well known that if a rod of lead, or even of steel or of brittle glass, is placed by its two ends on two supports, and is left in that position for a long time, its own weight ultimately gives it a permanent bend. The molecules of the unsupported part of the rod, under the accumulating effects of gravitation, slowly glide past each other, and ultimately re-arrange themselves in their mutual positions, just as if, instead of the metallic rod, a stick of soft sealing-wax had been taken, or some other plastic body, in which the particles easily glide and change places. But the analogy between metals and plastic bodies can be rendered still more apparent if external pressure is resorted to. Suppose we put a lump of plastic clay in a flower-pot, and press it from above. The clay will flow' through the hole at the bottom of the pot, exactly reproducing the flow of a vein of water out of the same pot; the speed only of the flow will be slower, but all the relative movements of the particles will be exactly the same. But now, suppose we take a piece of lead instead of the clay, and, after having placed it in a strong steel cylinder, which also has a hole in its bottom like the flower-pot, exert upon it a strong pressure: a powerful piston, let us say, slowly presses the lead. The lead will then 'flow,' exactly as the clay flowed out of the flower-pot, although it will never cease to remain solid-its temperature being hundreds of degrees below the point at which lead could be molten. The same happens, if we use a still greater pressure, with copper, and even with steel, as was proved some five-and-twenty years ago by a member of the French Academy, Tresca, in his memorable researches on the Flowing out of Solids. All metals, when they are submitted to a sufficient pressure, behave exactly as plastic bodies: their molecules acquire a certain mobility, and glide past each other, exactly as they glide in liquids—the metal remaining in the meantime quite solid, or even brittle.

A still closer analogy between liquids and solids appears from the experiments of the Belgian Professor, W. Spring.10 He shows that, solutions, increases every year. The rejection of pure metal out of solidifying alloys, or of metals combined with a definite number of molecules of the solvent, is quite similar to the crystallisation of salts out of liquid solutions. Also the influence of a third metal for increasing solubility. In a word, all the properties of solutions (they have been analysed in this Review in August, 1892) are known to exist in alloys.

10 They were begun since 1878, and the results were published in the Bulletin de l'Académie de Belgique; the chief memoirs are in 1880, vol. xlix. p. 323; 1883, 3rd series, vol. v. p. 492; 1883, vol. vi. p. 507; and 1894, vol. xxviii. p. 23.

just as two drops of a liquid coalesce when they are brought in contact with each other, so also two pieces of solid metal coalesce, at a temperature very remote from their melting-points, if they are brought into real contact with each other by external pressure. He takes, for instance, two small cylinders prepared of each of the following metals: steel, aluminium, antimonium, bismuth, cadmium, copper, tin, lead, gold, and platinum. Their ends are carefully planed, true toth of an inch, by a tool quite free from grease. One cylinder of each pair is then posed upon the other, the two being pressed upon each other by means of a hand-vice. They are left in this position for a few hours, and ultimately are found solidly welded to each other. If they are heated at the same time to a temperature which is, however, very remote from their fusion-temperature, they are so solidly welded together that all traces of the joint disappear.

Cylinders of different metals, submitted to the same experiment, give still more striking results. They are so well welded together that, when they are afterwards torn asunder by means of a powerful machine, quite new surfaces of tearing are produced. Besides, real alloys are formed between the two cylinders, in a few hours, for a thickness of from th to th of an inch, and more than that for lead and tin. An interpenetration of the molecules of the two metals takes place, although they both remain as solid as solid can be. As to fine filings of various metals, even of such a brittle metal as bismuth, they are easily compressed into solid blocks, as solid as if they had been molten before solidification and having the crystalline fracture characteristic of certain metals. More than that. Alloys of Wood's metal, as well as bronze and brass, have been obtained by pressing together fine filings of the different metals, although it was proved, both by calculation and direct experiment, that the temperature of the filings rose but a few degrees above the temperature of the laboratory." And finally, Spring has proved that solid metals evaporate from their surfaces, exactly as if they were in a liquid state, or as camphor evaporates, while remaining solid, so that, if we were endowed with a finer sense of smell, we could smell a metal at a distance. Zinc requires, as is known, a temperature of 780° Fahr. in order to be fused, and a still higher temperature in order to be brought to the state of vapour. And yet, even at a temperature of from 680° to 750° Fahr., it is volatilised and its molecules set upon a copper cylinder placed very near to it, making a brass alloy on its surface, as if the copper cylinder had been held

11 It is very interesting to note, however, that alloys were not obtained at once. When the filings of two or more metals were compressed into one solid block, the block had to be filed again into a fine powder; and when this powder was thoroughly mixed once more, and compressed for a second time, the alloy was obtained. Spring gives to that operation the characteristic name of 'kneading' (pétrissage).

in vapour of zinc at a high temperature. Strange as it may seem at first sight, we are thus bound to admit that the superficial molecules of a solid piece of metal enjoy the same mobility as if that surface were in the liquid state; and that they can as easily be freed from cohesion with their neighbours, and be projected into space, as if they were gaseous molecules.

The explanation of these most remarkable phenomena is found, as W. Spring points out, in a broad generalisation which we owe to Otto Graham, and which passed unnoticed when it was published, thirty-four years ago. A gas, we have said, consists of molecules dashing in all directions with very great velocities, which are increased when the temperature of the gas is raised. But it seems highly improbable that all the molecules of a gas should have the same velocities. Some of them, in all probability, run at a smaller speed, in consequence of their impacts with other molecules; while others have much greater velocities. One could say, as Spring writes, that some of them are hotter and some others are cooler, and that the thermometer, which gives the temperature of the gas, informs us only about the average velocity of the molecules which bombard it, without giving us an idea of either the maximum or the minimum velocities attained by some of them. Spring concludes therefrom, in conformity with Graham, that while most molecules of a solid move about (or vibrate) with the slower velocities characteristic of the solid state, there are, in addition, a number of molecules which move about with a much greater rapidity, corresponding to the liquid or to the gaseous state. And when a heated metal, on approaching its temperature of fusion, becomes soft, as red-hot iron does, its softness is simply due to an increased proportion of rapidly moving molecules amongst those which still perform the slower movements characteristic of the solid state. The great puzzle of plasticity in the most solid rocks and the most brittle metals thus ceases to be a puzzle.12

As to the fact of evaporation from the surface of solid metals, Spring suggests that each piece of metal (each solid, in fact) has on its surface a number of molecules which, finding more free scope for their oscillatory movements, acquire greater velocities and are torn off the sphere of cohesion with their neighbours so as to be projected into space. In other words, they evaporate like gaseous molecules, although the average temperature of the piece of metal is very much below its temperature of evaporation, or even its temperature of fusion.13 This conclusion of Spring finds a further most remarkable

12 The importance of time in plastic changes of form is well known, although it was so much neglected by Tyndall in his polemics with Forbes. The bearings of Graham's hypothesis upon this feature of plasticity are self-evident, and we must hope that somebody will soon take up this question.

13 'Sur l'apparition, dans l'état solide, de certaines propriétés caractéristiques de

confirmation in the work of G. Van der Mensbrugghe, his colleague in the Belgian Academy, who worked in a quite different direction, but came about the very same time to the same idea; namely, that 'the density of a solid is often, if not always, smaller in its superficial layer than it is in its interior.' 14

However, one step more remained to be made in order to prove by direct experiment that in a solid block of metal certain molecules are really endowed with a greater mobility, and can travel through its mass while the block itself remains solid. And this step was made by Graham's former collaborator, Roberts-Austen, and announced in the Bakerian lecture which he delivered before the Royal Society in February last. Roberts-Austen took a small cylinder of lead (about of an inch long), with either gold, or a rich alloy of lead with gold, at its base. He kept it for thirty-one days at a temperature of 485° Fahr., which is 135 degrees lower than the temperature of fusion of lead. Or else he kept like cylinders at a still lower temperature, down to the temperature of the laboratory rooms. At the end of this time, the lead cylinder was cut into sections and the amount of gold which had diffused through it, in its solid state, was determined. It then appeared that gold had diffused through solid lead, more or less, at all temperatures between 484 and 212 degrees, and there is evidence that diffusion went on, though at a smaller speed, even at the ordinary temperature of our rooms. Molecules of gold had travelled up the cylinder amidst the lead molecules, and they had lodged themselves amongst the latter on their own accord. A decisive proof in favour of Graham's hypothesis was thus produced.

The brilliant hypothesis of Graham, who suggested, so long ago as 1863, that the 'three conditions of matter (solid, liquid, and gaseous) probably always exist in every liquid or solid substance, but that one predominates over the others,' 16 finds now a full confirmation in Spring's and Roberts-Austen's researches, which have themselves been confirmed by other workers in the same field. If these views become generally accepted, as they probably will, their bearings upon the whole domain of molecular physics and chemistry will have a far-reaching and lasting importance. Not only the continuity between the three states of matter, solid, liquid, and gaseous, is demonstrated, but we can understand now why such continuity exists. Moreover, with the aid of Graham's hypothesis we

l'état liquide ou gazeux des métaux,' in Bulletin de l'Académie de Belgigue, 3o série, tome xxviii. pp. 27 sq.

་

14 Remarques sur la constitution de la couche superficielle des corps solides. Ibid., tome xxvii. 1894, p. 877.

15 Transactions of the Royal Society, 1896, vol. clxxxvii., A, p. 383. A summary of the lecture was published in the Proceedings, and in Nature, as also in most continental papers.

16 Quoted from Roberts-Austen's Bakerian lecture.

begin to see our way in the extremely difficult and puzzling subjects of solutions and alloys, of the critical state' of matter, of dissociation, and of a number of other physico-chemical phenomena. From this hypothesis the kinetic theory of gases receives a new, powerful support; and very probably the theories of surface-tension and evaporation, as also, perhaps, of surface-electrification, will receive a new impulse. Seeing that, we are ready to recognise, with RobertsAusten, that 'metals have been sadly misunderstood'; that they probably are never quiescent, and fully deserve that the methods so fruitful for the study of living beings should be applied to them and their alloys.

III

A corner of the veil which for so many centuries concealed from man the North-Polar area has at last been lifted by the NansenSverdrup expedition. All what we formerly knew of that vast realm of ice was its borderlands only; but the bold Norwegians have deeply penetrated into its heart, beyond the eighty-sixth degree of latitude, and the whole aspect of our hypothetical knowledge about these dreary regions is already modified. The vague name of a 'NorthPolar area' can be abandoned, and henceforward we can speak of a 'North-Polar basin.'

If we

This basin is often referred to as if it were a circle, the centre of which is the North Pole; but it has not that circular shape. look at it, keeping the Greenwich meridian before us, we see, first, a broad channel, 900 miles wide, between Greenland and Norway, inclined to the north-east and leading from the Atlantic into the Arctic Ocean. From that wide entrance a long and wide gulf stretches, in a slightly crescent-shaped form, between the shores of Russia and Siberia on the right, and the North-American archipelagoes and Alaska on the left. It widens as it crosses the Pole, and it ends in a wide semi-circle, out of which the Behring Strait is the only outlet. This narrow issue being, however, of little importance, we may neglect it, as well as several wide indentations of the two coasts, and we may say that the Arctic basin is a broad, pear-shaped gulf, 2,500 miles long, 900 miles broad at its entrance, widening to 2,000 miles at its nearly blind Behring Strait end.17

"The Behring Strait is so narrow and so shallow (maximum depth, 60 fathoms) that for oceanic circulation it has but little importance. A warm current flows along its American side, from the Pacific into the Arctic Sea; and a cold current flows in the opposite direction along the coast of Asia-both seemingly varying in intensity with the seasons. As to a permanent cold under-current, the Yukon soundings have rendered it improbable. Cf. the admirable Atlas of the Pacific, published by the Deutsche Seewarte; Otto Petterson's excellent paper, Contributions to the Hydrography of the Siberian Sea' (in English), in Vega Expeditionens Vetenskapliga Iakttagelser, vol. ii. p. 379; Stuxberg's Evertebrat fauna i Sibiriens Ishaf,' same work, vol. i. p. 677; and H. W. Dall, in American Journal of Science, 1881, vol. xxi. quoted by Petterson.