purity added to a mass of metal bears a close relation to the atomic volume of the added element.

Density. The density of a metal is dependent on the intimacy of the contact between the molecules. It is dependent, therefore, on the crystalline structure, and is influenced by the temperature of casting, by the rate of cooling, by the mechanical treatment, and by the purity of the metal. With the exception of bismuth, all metals are lighter when molten than when in the solid state. In the case of cast iron, which passes through a pasty state on solidification, the density is less in that state than in the fluid or solid. The density of a metal is augmented by wire-drawing, hammering, and any other physical method of treatment in which a compressing stress is employed. Mere traction, however, may diminish the density by tending to develop cavities in the metal. Pressure on all sides of a piece of metal increases its density. The density of standard gold, for example, by compression between dies is increased by 0.9, and cast discs of platinum, having a density of 21.21, may have the density increased to 21.46 by striking; whilst annealing such struck discs will again diminish their density. This shows that the compression is not permanent, and is solely due to the closing of pores. W. Spring W. Spring has even shown by careful experiments on lead, tin, bismuth, antimony, cadmium, aluminium, and zinc, that a pressure of 20,000 atmospheres continued for many days is insufficient to effect the obliteration of all the pores. A metal can only be really compressed if the result of the application of pressure is to cause it to pass to an allotropic state, that is denser than that which it originally possessed. The specific gravities of the various metals are given in the table on p. 58. Lithium is the lightest metal, and iridium the heaviest, the specific gravity of the former being o.6, and that of the latter 22.38.

So early as 1845, Joule + recognised the importance of determining the specific gravity of melted metals, seeing that "this condition would completely obviate the influence of cohesion, or that of any particular molecular arrangement." His method, which was essentially that afterwards adopted by Mallet ‡ and by the author,§ may be described as follows:-It consists in filling with molten metal a vessel, the capacity of which may be calculated for the particular temperature at which the molten metal is introduced. The weight of the metal when cold, divided by the *Bull. Soc. Chim. Paris, vol. xxxix. (1883), p. 515; and Gray, Proc. Roy. Soc., vol. liv. (1893), p. 283.

† Collected Papers. Published by the Physical Society, vol ii. p. 136. Proc. Roy. Soc., vol. xxii. (1873), p. 366; and vol. xxiii. (1874), p. 209. Ibid., vol. xxiii. (1875), p. 481.

weight of water which the expanded vessel is capable of holding, gives the fluid density of the molten metal. Subsequently, working in conjunction with T. Wrightson, the author determined, by the aid of an instrument called the oncosimeter, devised by the former, the fluid density of several metals. The results which have as yet been obtained may be briefly summarised in the following table :—

The question of the fluid densities has also been investigated by Nies and Winkelmann,† who have adopted another method, determining the liquid density by observing the weights of blocks which just sink and just swim. With regard to bismuth, C. Lüdeking finds that this metal, like water, attains a maximum density just before becoming solid, the expansion at the moment of solidification being about 3 per cent. of the volume.

Fracture. The appearance of the fractured surface of a metal depends partly on the nature of the metal and partly on the manner in which solidification occurred. Sudden cooling to a great extent prevents the formation of crystals, whilst slow cooling facilitates their development. Long-continued hammering, frequent vibrations, and intense cold will produce the latter result. Any condition that affects either the cohesion or the crystalline structure of a metal affects its fracture. Thus, lead broken when hot has a columnar structure; when broken cold, this structure is not exhibited.

* Phil. Mag., vol. xi. (1881), p. 295; vol. xiii. (1882), p. 360.
+ Sitzungsber. der Acad. der Wissen. zu München, 1881, p. 63.
† Ann. Phys. Chem., vol. xxxiv. (1888), p. 21.

The ordinary mineralogical terms regarding colour and frac ture are used in relation to metals. Practice, however, can alone enable the student to accurately describe these appearances.

Malleability. This is the property of permanently extending in all directions, without rupture, by pressure produced by slow stress or by impact. As a rule, crystalline metals are not malleable, and any circumstance that tends to produce crystallisation must affect the malleability. Thus, in nearly all metals the malleability becomes impaired when they are subjected to rolling or long-continued hammering; but this property may be regained by annealing, which consists in raising the metal to a high temperature and allowing it to cool, either rapidly or slowly, usually the latter. At different temperatures metals behave in different ways; some are malleable when at a red heat, but not so when cold. These are defined as being cold-short. Others are malleable when cold, but not when at a red heat. These are described as being red-short. Some metals are malleable at all temperatures, others are not malleable at all. Zinc is brittle when cold and when hot, but at a temperature of 150° it is malleable. The malleability of a metal is dependent on its purity. Relative malleability may be determined by the degree of thinness of the sheets that can be produced by beating or rolling the metals, without annealing.

Ductility is the property that enables metals to be drawn into wire. It generally decreases with an increase in the temperature of the wire at the time of drawing, but there is no regular ratio between the two. Iron is less ductile at 100°, and more ductile at 200°, than it is at o°. Malleable metals are also ductile, but they do not possess the two properties in the same order. Arranged according to their malleability, the more important metals follow this order :-1. Gold; 2. Silver; 3. Copper; 4. Tin; 5. Platinum; 6. Lead; 7. Zinc; 8. Iron; 9. Nickel. The order of ductility, on the other hand, is:-1. Gold; 2. Silver; 3. Platinum; 4. Iron; 5. Nickel; 6. Copper; 7. Zine; 8. Tin; 9. Lead. The rate at which the traction is applied, exercises an important influence on the properties of malleability and ductility.

Tenacity is the property possessed by metals, in varying degrees, of resisting the separation of their molecules by the action of a tensile stress.

Toughness is the property of resisting the separation of the molecules after the limit of elasticity has been passed.

Hardness is the resistance offered by the molecules of a sub. stance to their separation by the penetrating action of another sub

stance. Great differences are observable between the hardness of the different metals. The results of the experiments of Bottone gave valuable information. In his scale the hardness of the diamond was found to be 3010, whilst the relative hardness of twenty metals was determined with the following results :

In these determinations the time necessary to produce a cut of definite depth was taken as a measure of the hardness of the material, and Bottone concluded that the hardness so obtained was proportional to the specific gravity of the metal divided by its atomic weight. Metals that possess high limits of elasticity are usually very hard. The softness increases with an increase of temperature. Mr. T. Turner,* of the Mason College, Birmingham, a former pupil of the author, has also investigated the hardness of metals and devised a useful instrument for the purpose.

Brittleness is the sudden interruption of molecular cohesion when the substance is subjected to the action of some extraneous force, such as a blow or a change of temperature. It is largely influenced by the purity of the metal.

Elasticity, Extensibility, and Strength of Metals.-At first sight it might seem that testing the mechanical properties of metals is more within the province of the engineer than that of the metallurgist. The latter has, however, not only to extract metals from their ores, but also to fit them for use. He must therefore know what mechanical properties † are possessed by the more important metals and alloys, and be able to submit them to experimental tests instead of merely trusting to statements recorded by others.

Elasticity is the power a body possesses of resuming its original form after the removal of an external force which has produced a change in that form. The point at which the elasticity and

* Proc. Birmingham Phil. Soc., vol. v. part 2.

The more important recent works dealing with this subject are:Kennedy, Proc. Inst. C. E., vol. lxxxviii. (1887), p. 1; Unwin, Testing of Materials of Construction, 1888; Lebasteur, Les Métaux à l'Exposition Universelle, 1878. J. Résal, Constructions Métalliques: Elasticité et résistance des matériaux, 1892.

B

the applied stress exactly counterbalance each other, is termed the limit of elasticity. If the applied stress were then removed, the material acted upon would resume its original form. If, however, the stress were increased, the change in form would become permanent, and permanent set would be effected. Within the limit of elasticity, a uniform rod of metal lengthens or shortens equally under equal additions of stress. If this were the case beyond that limit, it is obvious that there would be some stress that would stretch the bar to twice its original length, or shorten it to zero. This stress, expressed in lbs. or tons for a bar of 1 inch square cross section is termed the modulus of elasticity. As an illustration, let it be supposed that a bar of steel 1 inch square is stretched to its limit of elasticity by a force of 139,000 lbs., and to have elongated under the action of this stress 0.00418 inch, the modulus of elasticity of this bar would be the force that would be required to elongate it by 1 inch, and this would be

0.00418 1.0 :: 139,000: x x=33,253,588 lbs. per square inch. Hence the modulus of elasticity is a stress that bears the same proportion to the original length of a uniform bar as the stress that will produce any given amount of strain bears to the length of this strain, the term stress meaning an equilibrating application of force to a body, and the term strain meaning any definite alteration of form or dimensions sustained by that body. The modulus of elasticity may thus be defined as being the number obtained by dividing the number expressing the stress by that expressing the strain that it produced. Unwin expresses this as follows:

Let p be the stress reckoned on unit of area, and λ the extension or compression reckoned per unit of length.

the coefficient of direct elasticity, or Young's modulus. It has the same value for tension and compression.

Thus, to take the above case of a hardened sample of steel in which the extension up to the limit of elasticity was small, the load, however, being considerable :

Load up to limit of elasticity, 139,000 lbs. per square inch.
Extension up to limit of elasticity, 0.00418 inch.

E=

139,000

0.00418

=33,250,000 lbs.

per square inch.

Professor Kennedy finds the specific extension to be a quantity most useful in works. This is the extension in thousandths of an inch on a length of 10 inches under a stress of 1000 lbs.