undergo hardening while at rest at the ordinary temperature. If the temperature is raised so as to bring about " tempering," the solid solution breaks down in precisely the same way as indicated above and the steel becomes hard (and also magnetic). It would thus seem that hardening as the result direct br indirect of quenching is due to the separation from solid solution, in a state of extremely fine division, of a phase the formation of which had been suppressed by quenching. According to the theory of amorphous metal (see below) each of the minute crystallites of the phase thus separated will be surrounded by a zone of amorphous metal, which is itself very hard. If the minute crystals thus separated are sufficiently small and numerous, the result will be that a considerable proportion of the whole alloy will be thrown into the amorphous state, extreme hard- ness resulting. On this view, the martensite of hardened steel should consist mainly of minute crystallites of alpha-iron embedded in an amorphous matrix consisting of iron and carbon (or carbide) in solution in it. This suggested constitution of martensite readily accounts for its hardness and for the fact that it is magnetic, and in view of the intimate manner in which the minute crystallites of alpha-iron are embedded in unyielding and un-magnetisable amor- phous metal for the magnetic hardness of the martensitic steel. This view is further confirmed by the observation that the chemical behaviour of quench-hardened steel is in certain respects closely similar to that of the same steel hardened by cold work and thus rendered partially amorphous (Whiteley). Finally, it has recently been shown by X-ray methods, that the space-lattice typical of alpha-iron is present in martensitic steel (Westgren).
The theory of amorphous metal, just mentioned, has played a very important part in scientific metallurgical thought during the period under review. The conception that metal could be rendered amor- phous by mechanical disturbance of its crystalline structure was originated by Beilby, in the first instance, to account for the phenom- ena observed by him and others in connexion with the polishing of metals and other substances. Beilby further applied the concep- tion to explain the hardening which metals undergo as the result of plastic deformation (cold work) by suggesting that layers of amorphous metal are formed on the surfaces of internal slip which occurs during plastic straining. Both these theories are now widely, if not universally, accepted in England and America, but still find opposition on the Continent, where the weight of the experimental data is not sufficiently appreciated. More recently Rosenhain has brought forward a conception which has already been present in the minds of many other investigators (notably Osmond) in a less definite form, that a film or thin layer of amorphous metal exists in the inter-crystalline boundaries of all metals, quite apart from any effects of strain. This view has been vigorously contested, but ex- perimental evidence in its confirmation has been steadily accumu- lated from a very great variety of sources, until at the present time the " intercrystalline amorphous cement " theory is at least as firmly established as that of the ' amorphous theory of strain-hardening. The most striking series of facts supporting the " amorphous cement " theory is connected with the behaviour of the inter- crystalline boundaries under stress. It is well established that in normal circumstances these boundaries are stronger than the crystals themselves, so that fractures of metals generally occur by breaking through the crystals and not by pulling them apart. It has, however, been shown that at a high temperature near to, but definitely below, the melting point, pure metals can be easily caused to break with a perfectly inter-crystalline fracture (Rosenhain and Ewen). This is to be ascribed to the greatly decreased viscosity at such temperatures of the inter-crystalline amorphous metal, which is regarded as pos- sessing the properties of a viscous under-cooled liquid. The actual viscosity, however, depends very much upon the nature of the metal and upon the temperature the farther a metal is below its normal melting-point the higher the viscosity of the amorphous phase. Accordingly, in some of the softer metals and alloys the amorphous material is sufficiently mobile to allow of sensible movement in relatively short times. Thus, an alloy of zinc with copper and alu- minium has been discovered which, in the cold-worked state when it is partially amorphous, behaves very much like pitch; it will bend to any desired extent if allowed to do so gradually, but breaks short if rapid bending is attempted. Similarly, the inter-crystalline cement in certain metals and alloys, although it proves stronger than the crystals when the metal is loaded at any normal rate, appears to be capable of giving way by some form of viscous or visco-elastic move- ment under very prolonged loading such as that due to internal stresses. Much attention has been devoted to the study of fractures occurring in various metals as the result of the application of internal or other prolonged stresses. In brass these phenomena have become known by the misleading term " season cracking," but it has recently been discovered that strikingly similar phenomena are to be found in a number of other metals, including lead, certain alloys of aluminium, platinum and steel (Rosenhain and Archbutt). It has been shown that in the case of brass, lead, steel and aluminium alloys, certain types of chemical reagents which appear to act preferentially upon the material in the crystal boundaries, contribute to the occurrence of such fractures, which are typically inter-crystalline (Moore and Beckinsale). At the same time it has been clearly shown that in the case of the aluminium alloys at all events such chemical action serves to accelerate the fractures, but is not essential to it, since it occurs,
although more slowly, in high vacuum or in an atmosphere of pur dry hydrogen (Rosenhain and Archbutt). In the case of brass i seems probable, although it has not yet been finally demonstrate^ that " season cracking ' can occur without the intervention of any chemical action. Similar types of cracking which have been dis covered in mild steel, however, appear to be very closely associates with the effects of certain chemicals, such as concentrated solution of alkalies, fused ammonium nitrate, etc. While there are still son metallurgists who refuse to think in terms of an amorphous intei crystalline cement (Hatfield, Tammann), the great majority of in vestigators are agreed that, directly or indirectly, this concepts serves to explain the occurrence not only of inter-crystalline fra tures under prolonged loading but also a number of other phenomen associated with the crystal boundaries.
Intimately connected, also, with the nature of inter-crystallin boundaries is the whole of the important phenomena of re-crystalliz tion and crystal growth which are of such fundamental importan with all annealing and heat -treatment operations. These have 1 studied in great detail in recent years. One of the most strikin = features is the relatively rapid formation of large crystals in certain conditions. Thus in an oblong piece of metal which has been severely strained, and is then heated in such a way as to be well above the usual temperature of re-crystallization at one end and well below it at the other, a zone is found in which very large crystals are formed ; this may occur either as the result of a temperature-gradient being applied to a uniformly strained piece of metal or of the application of a suitable uniform temperature to a piece of metal in which there is i strain-gradient. The explanation appears to be that for a given degre< of previous plastic strain there is a temperature most favourable to rapid crystal growth (Jeffries). An interesting practical applicatioi of the ideas derived from the study of these phenomena is the pro duction of wires of certain metals, notably tungsten, which have beei so treated as to consist, for considerable lengths, of single long crystals This result is achieved by drawing the cold-worked wire into a annealing furnace at a suitable temperature at precisely the righ rate. The tungsten wire thus produced is particularly valuable fo the manufacture of electric lamp filaments and it has also been shown to possess interesting elastic properties (Wartenberg), whic are readily accounted for by the absence in such material of any amorphous inter-crystalline material the viscous or visco-elastic properties of which affect the behaviour of the wire in this respect. Much study has also been devoted particularly to the re-crystalliz tion of aluminium after cold- working, but new theoretical views hav not yet been advanced. Here, as in almost every direction, th progress of research upon metals and alloys tends to open up ne avenues for further research and further advance. In a subject which is showing such rapid and vigorous growth, such a summary as that here given cannot hope to deal with more than a few outstanding points which appear to be of primary importance, but development is so rapid and on such wide lines that it is impossible to foresee what trend it may follow in the near future.
LITERATURE. For iron and steel metallurgy, industrial as well as scientific, the journal of the Iron and Steel Institute should be consulted, not only for original publications but for abstracts which cover the literature of the whole world on this subject. In addition, excellent abstracts will also be found in the metallurgical section i the journal of the Society of Chemical Industry, while such journal: as Stahl und Risen, the Revue de Metallurgie and Chemical and Metallurgical Engineering also deal with this branch of the subject. In addition to Iron Age, The Iron and Coal Trade Review and similar journals may also be mentioned, but adequate references to these can be found in the abstracting journals already mentioned. As regards general metallurgy, the annual volumes of Mineral Industry contain detailed reviews of progress, year by year, while the journal of the Institution of Mining and Metallurgy contains important original papers. For the non-ferrous metals, apart from their reduc- tion from the ore, the most valuable reference is to the journal of the Institute of Metals, including particularly extensive abstracts as well as original papers. The Revue de Metallurgie and several German journals, also formerly the International Journal of Metallography (now Mclallographie) , may be named, as well as the appropriate section of the American Institution of Mining and Metallurgical Engineers (American Institute of Metals). The publications of the Bureau of Standards (Washington, U.S.A.) and of the National Physical Laboratory (Teddinjjton, England) are also of first-rate importance. The Faraday Society (London) has also published in its Transactions several "General Discussions " relating to metallurgical subjects, including particularly one on The Failure of Metals under Internal and Prolonged Stress, another relating to metallurgical microscopy, and one on the application of X-rays. (W. R. N.)
METEOROLOGY (see 18.264*). Since 1910 considerable ad- vances in meteorological knowledge have been made both on the observational and the theoretical sides. The World War emphasized the importance of meteorology, more particular!; in regard to a knowledge of the density and of both the direction and velocity of the wind in the overlying air strata, and th meteorological services of the combatant nations were largely
- These figures indicate the volume and page number of the previous article.
