METALS. The earliest evidence of a knowledge and use of metals is found in the prehistoric implements of the so-called Bronze and Iron ages. In the earliest periods of written history, however, we meet with a number of metals in addition to these two. The Old Testament mentions six metals - gold, silver, copper, iron, tin, and lead. The Greeks, in addition to these and to bronze, came also to know mercury ; and the same set of metals, without additions, forms the list of the Arabian chemists of the 8th and of the Western chemists of the 13th century. During the 15th century Basilius Valentinus discovered antimony ; he also speaks of zinc and bismuth, but their individuality was established only at a later period. About 1730-40 the Swede Brand discovered arsenic and cobalt (the former is not reckoned a metal by modern chemists), while the Englishman Ward recognized the individuality of platinum. Nickel was discovered in 1774 by Cronstedt, manganese in 1774 by Scheele. The brothers D'Elhujart, in 1783, prepared tungsten ; Hjelm, in 1782, isolated molybdenum from molybdic oxide, where its existence had been conjecturally asserted by Bergmann in 1781. Uranium, as a new clement, was discovered by Klaproth in 17 89 ; but his metallic " uranium," after having been accepted as a metal by all chemists until 1841, was then recognized as an oxide by Peligot, who subsequently isolated the true metal. Tellurium was discovered by M filler von Reichenbach in 1782 (again by Klaproth in 1798) ; titanium, by Klaproth in 1795 ; chromium, by Vauquelin in 1797 ; tantalum, by Ilatchctt in 1801, and by Ekeberg in 1802. Palladium, rhodium, iridium, and osmium (which four metals always accompany platinum in its ores) were discovered, the first two by Wollaston in 1803, the other two by a number of chemists ; but their peculiarity was established chiefly by Smithson Tennant.

After Davy, in 1807 and 1808, had recognized the alkalies and alkaline earths as metallic oxides, the existence of metals in all basic earths became a fbregone conclusion, which was verified sooner or later in all cases. But the discovery of aluminium by Welder in 1828, and that of magnesium by Bussy in 1829, claim special mention. Cadmium, a by no means rare heavy metal, was discovered only in 1818, by Stromeyer.

Of the large number of discoveries of rare metals which have been made in more recent times only a few can be mentioned, as marking new departures in research or offering other special points of interest. In 1861 Bunsen and Kirchhoff, by means of the method of spectrum analysis, which they had worked out shortly before, discovered two new alkali-metals which they called caesium and rubidium. By means of the same method Crookes, in 1861, discovered thallium ; Reich and Richter, in 1863, indium ; and Lecoq de Boisbaudran, in 1875, gallium. The existence of the last-named metal had been maintained, theoretically, by Mendelejeff, as early as 1871. The existence of vanadium was proved in 1830 by Sefstrom ; but what he, and subsequently Berzelius, looked upon as the element was, in 1867, proved to be really an oxide by Roscoe, who also succeeded in isolating the true metal.

The development of earlier notions on the constitution of metals and their genetic relation to one another forms the most interesting chapter in the history of chemistry (see ALCHEMY). What modern science has to say on the matter is easily stated : all metals properly so called (i.e., all metals not alloys) are elementary substances; hence, chemically speaking, they are not " constituted " at all, and no two can be related to each other genetically in any way whatever. Our scientific instinct shrinks from embracing this proposition as final ; but in the meantime it must be accepted as correctly formulating our ignorance on the subject. All metallic elements agree in this that they form each at least one basic oxide, or, what comes to the same thing, one chloride, stable in opposition to liquid water. This at once suggests an obvious definition of metals as a class of substances, but the definition would be highly artificial and objectionable on principle, because when we speak of metals we think, not of their accidental chemical relations, but of a certain sum of mechanical and physical properties which unites them all into one natural family. What these properties are we shall now endeavour to explain.

All metals, when exposed in an inert atmosphere to a sufficient temperature, assume the form of liquids, which all present the following characteristic properties. They are (at least practically) non-transparent ; they reflect light in a peculiar manner, producing what is called "metallic lustre." When kept in non-metallic vessels they take the shape of a convex meniscus. These liquids, when exposed to higher temperatures, some sooner others later, pass into vapours. What these vapours arc like is not known in many cases, since, as a rule, they can be produced only at very high temperatures, precluding the use of transparent vessels. Silver vapour is blue, potassium vapour is green, many others (mercury vapour, for instance) are colourless. The liquid metals, when cooled down sufficiently, some at lower others at higher temperatures, freeze into compact solids, endowed with the (relative) non-transparency and the lustre of their liquids. These frozen metals in general form compact masses consisting of aggregates of crystals belonging to the regular or rhombic or (more rarely) the quadratic system. But in many cases the crystals are so closely packed as to produce an apparent absence of all structure. Compared with non-metallic solids, they in general are good conductors of heat and of electricity. But their most characteristic, though not perhaps their most general, property is that they combine in themselves the apparently incompatible properties of elasticity and rigidity on the one hand and plasticity on the other. To this remarkable combination of properties more than to anything else the ordinary metals owe their wide application in the mechanical arts. In former times a high specific gravity used to be quoted as one of the characters of the genus; but this no longer holds, since we have come to know of a whole series of metals which float on water. Let us now proceed to see to what degree the mechanical and physical properties of the genus are developed in the several individual metals.

.Yon-Transparency. - This, in the case of even the solid metals, is perhaps only a very low degree of transparency. In regard to gold this has been proved to be so ; gold leaf, or thin films of gold produced chemically on glass plates, transmit light with a green colour. On the other hand, those infinitely thin films of silver which can be produced chemically on glass surfaces are absolutely opaque. Very thin films of liquid mercury, according to Melsens, transmit light with a violet•blue colour ; also thin films of copper are said to be translucent. Other metals, so far as we know, have not been more exactly investigated in this direction.

Cotour. - Gold is yellow ; copper is red ; silver, tin, and some others are pure white ; the majority exhibit some modification or other of grey.

Reflexion of Light. - Polished metallic surfaces, like those of other solids, divide any incident ray into two parts, of which one is refracted while the other is reflected, - with this difference, however, that the former is completely absorbed, and that the latter, in regard to polarization, is quite differently affected.1 The degree of absorption is different for different metals. According to Jamin, the remaining intensity, after one and ten successive perpendicular reflexions respectively from the metal-mirrors named, is as follows (original intensity =1) : - METALS This shows the great superiority of silver as a reflecting medium, especially in the case of repeated reflexion.

Crystalline Forni. - Most (perhaps all) metals are capable of crystallization, and in most cases isolated crystals can be produced by judiciously managed partial freezing. The crystals belong to the following systems : - regular system - silver, gold, palladium, mercury, copper, iron, lead ; quadratic system - tin, potassium ; rhombic system - antimony, bismuth, tellurium, zinc, magnesium.

,Structure. - Perhaps all metals, in the shape which they assume in freezing, are crystalline, only the degree of visibility of the crystalline arrangement is very different in different metals, and even in the same metal varies according to the slowness of solidification and other circumstances.

Of the ordinary metals, antimony, bismuth, and zinc may be mentioned as exhibiting a very distinct crystalline structure : a bar-shaped ingot readily breaks, and the crystal faces are distinctly visible on the fracture. Tin also is crystalline : a thin bar, when bent, " creaks " audibly from the sliding of the crystal faces over one another ; but the bar is not easily broken, and exhibits an apparently non-crystalline fracture. - Class I.

Gold, silver, copper, lead, aluminium, cadmium, iron (pure), nickel, and cobalt are practically amorphous, the crystals (where they exist) being so closely packed as to produce a virtually homogeneous mass. - Class IT.

The great contrast in apparent structure between cooled ingots of Class I. and of Class II. appears, however, to be owing chiefly to the fact that, while the latter crystallize in the regular system, metals of Class I. form rhombic or quadratic crystals. Regular crystals expand equally in all directions; rhombic and quadratic ores expand differently in different directions. Hence, supposing the crystals immediately after their formation to be in absolute contact with one another all round, then, in the case of Class II., such contact will be maintained on cooling, while in the case of Class I. the contraction along a given straight line will in general have different values in any two neighbouring crystals, and the crystals consequently become, however slightly, detached from one another. The crystalline structure which exists on both sides becomes visible only in the metals of the first class, and only there manifests itself as brittleness.

Closely related to the structure of metals is their degree of " plasticity " (susceptibility of being constrained into new forms without breach of continuity). This term of course includes as special cases the qualities of " malleability " (capability of being flattened out under the hammer) and " ductility " (capability of being drawn into wire) ; but it is well at once to point out that these two special qualities do not always go parallel to each other, for this reason amongst others that ductility in a higher degree than malleability is determined by the tenacity of a metal. Hence tin and lead, though very malleable, are little ductile. The quality of plasticity is developed to very different degrees in different metals, and even in the same species it depends on temperature, and may be modified by mechanical or physical operations. A bar of zinc, for instance, as obtained by casting, is very brittle ; but when heated to 100° or 150° C. it becomes sufficiently plastic to be rolled into the thinnest sheet or to be drawn into wire. Such sheet or wire then remains flexible after cooling, the originally only loosely cohering crystals having got intertwisted and forced into absolute contact with one another, - an explanation supported by the fact that rolled zinc has a somewhat higher specific gravity (7-2) than the original ingot (6.9). The same metal, when heated to 205° C., becomes so brittle that it can be powdered in a mortar. Pure iron, copper, silver, and other metals are easily drawn into wire, or rolled into sheet, or flattened under the hammer. But all these operations render the metals harder, and detract from their plasticity. Their original softness can be restored to them by "annealing," i.e., by heating them to redness and then quenching them in cold water. In the case of iron, however, this applies only if the metal is perfectly pure. If it contains a few parts of carbon per thousand, the annealing process, instead of softening the metal, gives it a " temper' " meaning a higher degree of hardness and elasticity (see below).

What we have called plasticity must not be mixed up with the notion of softness, which means the degree of facility with which the plasticity of a metal can be discounted. Thus lead is far softer than silver, and yet the latter is by far the more plastic of the two. The now famous experiments of Tresca (Comptes Rendus, lix. 754) show that the plasticity of certain metals at least goes considerably farther than had before been supposed. He operated with lead, copper, silver, iron, and some other metals. Round disks made of these substances were placed in a closely fitting cylindrical cavity drilled in a block of steel, the cavity having a circular aperture of two or four centimetres below. By means of an hydraulic press, applied to a superimposed piston, a pressure of 100,000 kilos was made to act upon the disks, when the metal was seen to " flow" out of the hole like a viscid liquid. In spite of the immense rearrangement of parts there was no breach of continuity. What came out below was a compact cylinder with a rounded bottom, consisting of so many layers superimposed upon one another. Parallel experiments with layers of dough or sand plus some connecting material proved that the particles in all cases moved along the same tracks as would be followed by a flowing cylinder of liquid. Of the better known metals potassium and sodium are the softest; they can be kneaded between the fingers like wax. After these follow first thallium and then lead, the latter being the softest of the metals used in the arts. Among these the softness decreases in about the following order : - lead, pure silver, pure gold, tin, copper, aluminium, platinum, pure iron. As liquidity might be looked upon as the ne plus ultra of softness, this is the right place for stating that, while most metals, when heated up to their melting points, pass pretty abruptly from the solid to the liquid state, platinum and iron first assume, and throughout a long range of temperatures retain, a condition of viscous semi-solidity which enables two pieces of them to be " welded " together by pressure into one continuous mass. Potassium and sodium might probably be welded if their surfaces could be kept clear of oxide.

According to Prechtl, the ordinary metals, in regard to the degree of facility or perfection with which they can be hammered flat on the anvil, rolled out into sheet, or drawn into wire, form the following descending series: - Hammering. Rolling into Sheet. Drawing into Wire.

Lead. Gold. Platinum.

Tin. Silver. Silver.

Gold. Copper. Iron.

Zinc. Tin. Copper.

Silver. Lead Gold.

Copper. Zinc. Zinc.

Platinum. Platinum. Tin.

Iron. Iron. Lead.

To give an idea of what can be done in this way, it may be stated that gold can be beaten out to leaf of the thickness of 3 810 0 mm.; and that platinum, by judicious work, can be drawn into wire „„,.(1,0---cr mm. thick.

By the hardness of a metal we mean the resistance which it offers to the file or to the engraver's tool. Taking it in this sense, it does not necessarily measure, e.g., the resistance of a metal to abrasion by friction. Thus, for instance, 10 per cent. aluminium bronze is scratched by an edge-tool made of ordinary steel as used for knife-blades. And yet it has been found that the sets of needles used for perforating postage stamps last longer if made of aluminium bronze than they do if made.of steel.

Elasticity. - All metals are elastic to this extent that a change of form, brought about by stresses not exceeding certain limit values, will disappear on the stress being removed. Strains exceeding the "limit of elasticity" result in permanent deformation or (if sufficiently great) in rupture. Where this limit lies is in no ease precisely known. According to Wertheim' (who has done more for our knowledge of the subject than any one else) and Hodgkinson, the real law seems to be pretty much as indicated by the two curves on the accompanying diagram, where, in reference to a metallic wire, stretched by an appended weight, the abscissa always means the numerical value P of the weight, the ordinate of the unner curve 58 the total elongation caused by P, the ordinate of the lower curve that part of the elongation which remains when P is removed, so that the piece of the ordinate between the two curves gives the temporary ("elastic") expansion. From P=0 up to a somewhat indefinite point (a or A) both curves are nearly straight lines, the lower almost coinciding in its beginning with the axis of abscissa; ; from that point onwards these two curves approach each other, and at a short distance from the point of rupture they rapidly converge towards intersection. For any value of P which lies fairly on the safe side of A, we have approximately where A means the elastic (or substantially the total) expansion, the length, and q the square section of the wire or cylindrical bar operated upon. The reciprocal of E (viz. E - 1/e) is called the "modulus of elasticity."





Wertheim has determined this constant for a large number of metals and alloys. He used three methods : one was to measure the elongations produced, in a wire of given dimensions, by a succession of charges ; the other two consisted in causing a measured bar to give oil a musical note by (a) longitudinal and (b) transversal vibration, and counting the vibrations per second. The following table gives some of his results. Column 2 gives the constant E for millimetre and kilogramme. Hence 1000/E is the elongation in millimetres per metre length per kilo. Column 3 shows the charge causing a permanent elongation of 0.05 mm. per metre, - which, for practical purposes, he takes as giving the limit of elasticity ; column 4 gives the breaking strain. Values of E in square brackets [ are derived from vibration experiments; the rest from direct measurements of elongations. Numbers in round brackets ( ) do not necessarily refer to the same specimen as the other data.

58 The above numbers may be assumed to hold for temperatures from 15° to 20° C. Wertheim executed determinations also at other temperatures; but, as his numbers do not appear to reveal the tree relations between E and temperature, we quote the results of Kohlrausch and Loomis, who found the following relations between the modulus Ea for 0° C. and the value E, for + t' C.: - Iron : Et =Eo(1 - • 000483 t - •00000012 (2).

Thus, for these three metals at least, the value of E diminishes, when temperature increases, at pretty much the same rate per degree of temperature.

Specific Gravity. - This varies in metals from '594 (lithium) to 22.48 (osmium), and in one and the same species is a function of temperature and of previous physical and mechanical treatment. It has in general one value for the powdery metal as obtained by reduction of the oxide in hydrogen below the melting point of the metal, another for the metal in the state which it assumes spontaneously on freezing, and this latter value again, in general, is modified by hammering,- rolling, or wire-drawing, &c. These mechanical operations do not necessarily add to the density ; stamping, it is true, does so necessarily, but rolling or drawing occasionally causes a diminution of the density. Thus, for instance, chemically pure iron in the ingot has the specific gravity 7'844 ; when it is rolled out into thin sheet, the value falls to 7'6 ; when drawn into thin wire, to 7'75 (Berzelius). The following table gives the specific gravities of all metals (except a few very rare ones) according to the most trustworthy modern determinations. Where special statements are not made, the numbers may be assumed to hold for the ordinary temperature (15° to 17° or 20° C.), referred to water of the same temperature (specific gravity =1) as a standard, and to hold for the natural frozen metal.

58 58 58 58 Thermic Properties. - The specific heats of most metals have been determined very carefully by Regnault. The general result is that, conformably with Dulong and Petit's law, the " atomic heats " all come to very nearly the same value (of about 6'4); i.e., atomic weight by specific heat-64. Thus we have for silver by theory 6'4/108 - '0593, and by experiment '0570 for 10° to 100° C.

The expansion by heat varies greatly. The following table gives-the linear expansions from 0° to 100° C. according to Fizeau (Comptes Rendus, lxviii. 1125), the length at 0° being taken as unity.

The coefficient of expansion is constant for such metals only as. crystallize in the regular system ; the others expand differently in the directions of the different axes. To eliminate this source of uncertainty these metals were employed as compressed powders. The cubical expansion of mercury from 0° to 100° C. is '018153.

(Regnault).

Fusibility and Volatility. - The fusibility in different metals-is very different, as shown by the following table, which, besides including all the fusing points (ill degrees C.) of metals. which have been determined numerically, indicates those of a selection of other metals by the positions assigned to them in the-table. Of the temperatures given, those above (say) 500° C. must be looked upon as rough approximations.

Of the volatility of metals we have little precise knowledge ; only the following boiling points are known numerically : - For practical purposes the volatility of metals may be stated as. follows: - Distillable below redness: mercury.

Distillable at red heats : cadmium, alkali metals, zinc, magnesium.

Volatilized more or less readily when heated beyond their fusing points in open crucibles : antimony (very readily), lead,, bismuth, tin, silver.

Barely so: gold, (copper).

Practically non-volatile : (copper), iron, nickel, cobalt, aluminium; also lithium, barium, strontium, and calcium.

In the oxyhydrogen flame silver boils, forming a blue vapour, while platinum volatilizes slowly, and osmium, though infusible, very readily.

Latent _Heats of Liguefaelion - Of these we know little. The following numbers are due to Person - ice, it may be stated, being 80.

58 Of the latent heats of vaporization only that of mercury has been determined, - by Marignac, who found it to be 103 to 106 units.

Conductivity. - Conductivity, whether thermic or electric, is very differently developed in different metals ; and, as an exact knowledge of these conductivities is of great scientific and practical importance, much attention has been given to their numerical determination. The following are the modes in which the two conductivities have been defined as quantities.

Thermie. - Imagine one side (1) of a metallic plate, I) units thick, to be kept at the constant temperature ti, the other (II) at t2. After a sufficient time each point between I and II will be at a constant intermediate temperature, and in every unit of time a constant quantity Q of heat will pass from any circumscribed area S on I to the opposite area S on f [ , according to the equation Q=iif; 12) is called the (internal) conductivity of the metal the plate is made of. It is, strictly speaking, a function of C, and d2; but within a given small interval of temperatures it may be taken as a constant.

Electric. - When a given constant battery is closed successively by different wires of the same sort, then, according to experience, the strength I of the current (as measured for instance by the heat-equivalent of the electricity flowing through the circuit in unit of time) is in accordance with the equation , where 1 is the length and s the square section of the wire, while A is a constant which, for our purpose, need not be defined in regard to its physical meaning; r measures the specific resistance of the particular metal. Supposing a certain silver wire on the one hand and a certain copper wire on the other, when substituted for each other, to produce currents of the same strength, we have ?VI /$1=r212 /s2 ri/r2=3121(3210=k i.e., k gives Os the specific resistance of silver, that of copper being taken-1. In this relative manner resistances are usually measured, silver generally being taken as the standard of comparison. Supposing the relative resistance of a metal to be R, the reciprocal 1/11 is called its "electric conductivity." For the same metal R varies with the temperature, the higher temperature corresponding to the higher resistance. The following table gives the electric conductivities of a number of metals as determined by Matthicsen, and the relative internal thermic conductivities of (nominally) the same metals as determined by Wiedemann and Franz, with rods about 5 mm. thick, of which one end was kept at 100° C., the rest of the rod in a " vacuum" (of 5 mm. tension) at 12° C. Matthiesen's results, except in the two cases noted, are from his memoir in Pogg. Ann., 1858, ciii. 428.

Going by Matthiesen's old numbers, we find them to agree fairly with Wiedemann and Franz's thermic conductivities, which supports an obvious and pretty generally received proposition. Matthiesen's new numbers for gold and copper, however, destroy the harmony.

Magnetic Properties. - Iron, nickel, and cobalt are the only metals which are attracted by the magnet and can become magnets themselves. But in regard to their power of retaining their magnetism none of them comes at all up to the compound metal steel. See MAGNETISM.

Chemical Changes.

The chemical changes which metals are liable to may he classified according to the loss of metallicity involved in them. We will adopt this principle and begin with the action of metals on metals, which, as experience shows, always leads to the formation of truly metallic compounds.

Any two or more metals when mixed together in the liquid state unite chemically, or at least molecularly, in this sense that, although the mixture, on standing (hot), may separate into layers, each layer is a homogeneous solution or " alloy " of, in general, all the components in one another. With binary combinations the following two cases may present themselves : - (1) the two metals mix permanently in any proportion ; or (2) either of the two metals refuses to take up more than a certain limit-proportion of the other ; hence a random mixture of the two metals will, in general, part into two layers, - one a solution of A in B, the other a solution of B in A. The first case presents itself very frequently; it holds, for instance, for gold and silver, gold and copper, copper and silver, lead and tin, and any alloy of these two and bismuth. Many other cases might be quoted. A good example of the second case is lead and zinc, either of which dissolves only a very small percentage of the other. In the preparation of an alloy we need not start with the components in the liquid state; the several metals need only be heated together in the same crucible when, in general, the liquid of the more readily fusible part dissolves the more refractory components at temperatures far below their fusing points. Molten lead, for instance, as many a tyro in chemical analysis has come to learn to his cost, readily runs through a platinum crucible at little more than its own fusing point.

A homogeneous liquid alloy, when solidified suddenly, say by pouring it drop by drop into cold water, necessarily yields an equally homogeneous solid. But it may not be so when it is allowed to freeze gradually. If, in this case, we allow the process to go a certain way, and then pour off the still liquid portion, the frozen part generally presents itself in the shape of more or less distinct crystals; whether this happens or not, the rule is that its composition differs from that of the mother liquor, and consequently from that of the original alloy. This phenomenon of "liquation," as it is called, is occasionally utilized in metallurgy for the approximate separation of metals from one another ; 2 but in the manipulation of alloys made to be used as such it may prove very inconvenient. It does so, for instance, in the case of the copper-silver alloy which our coins are made of ; in a large ingot of such sterling silver the core may contain as much as 0.3 per cent. of silver more than the outer shell.

The existence of crystallized alloys, as the phenomenon of liquation generally, strongly suggests the idea that alloys generally are mixtures, not of their elementary components, but of chemical compounds of these elements with one another, associated possibly with uncombined remnants of these. This notion is strongly supported by the fact that the formation of many alloys involves an obvious evolution of heat and a decided modification in what one would presume to be the properties of the corresponding = A good illustration is afforded by the process of Pattinson as used for concentrating the silver in argentiferous lead. See LEAD.

mixture. The case of sodium amalgam may be quoted as a forcible illustration. What goes by this name in laboratories is an alloy of two to three parts of sodium with one hundred parts of mercury, which is easily produced by forcing the two components into contact with each other by means of a mortar and pestle, when they unite, with deflagration, into an alloy which after cooling assumes the form of a grey, hard, brittle solid, although mercury is a liquid, and sodium, though a solid, is softer than wax. Similar evidence of chemical action we have in the cases of brass (copper and zinc), bronze (copper and tin), aluminium bronze (copper and aluminium), and in many others that might be quoted. There are indeed a good many alloys the formation of which is not accompanied by any obvious evolution of heat or any very marked change in the mean properties of the components. But in the absence of all precise thermic researches on the subject we are not in a position to assert the absence of chemical action in any case. Indeed our knowledge of the proximate composition of alloys is in the highest degree indefinite - we do not even know of a single composite metal which has been really proved to be an unitary compound, and hence the important problem of the relation in alloys between properties and composition must be attacked on a purely empirical basis. What has been done in this direction is shortly summarized in the following paragraphs.

Colour. - Most metals are white or grey ; so are the alloys of these metals with one another. Gold alloys generally exhibit something like the shade of yellow which one would expect from their composition ; its amalgams, however, are all white, not yellow. Copper shows little tendency to impart its characteristic red colour to its alloys with white or grey metals. Thus, for instance, the silver alloy up to about 30 per cent. of copper exhibits an almost pure white colour. The alloys of copper with zinc (brass) or tin (bronze) are reddish-yellow when the copper predominates largely. As the proportion of white metal increases, the colour passes successively into dark yellow, pale yellow, and ultimately into white. Aluminium bronze, containing from 5 to 10 per cent. of aluminium, is golden-yellow.

Plasticity. - This quality is most highly developed in certain pure metals, notably in gold, platinum, silver, and copper. Of platinum alloys little is known. The other three, on uniting with one another, substantially retain their elasticities, but the addition of any metal outside the group leads to deterioration. Thus, for instance, according to Karsten, copper, by being alloyed with as little as 0.6 per cent, of zinc, loses its capability of being forged at a red heat ; it cracks under the hammer. Antimony or arsenic to the extent of 0.15 per cent. renders it unfit for being rolled into thin sheet or drawn out into fine wire, and makes it brittle in the heat ; 0.1 per cent. of lead prohibits its conversion into leaf.

Hardness, Elasticity, Tensile Strength. - In reference to these qualities, we shall confine ourselves to some very striking changes for the better which the metals (1) gold, (2) silver, (3) copper suffer when alloyed with moderate proportions (10 per cent. or so) of (1) copper, (2) copper, (3) tin, zinc, or aluminium respectively. Any of these five combinations leads to a considerable increase in the three qualities named, although these are by no means highly developed in the added metals ; most strikingly it does so in the case of aluminium bronze (copper and aluminium), which is so hard as to be very difficult to file, and is said to be equal in tensile strength to wrought iron. To illustrate this we give in the following table, after Matthiesen, the breaking strains of double wires, No. 23 gauge, in It) avoirdupois, for certain alloys on the one hand and their components on the other.





Separate Metals. Alloys, Gold 20--2350 If Silver 40-45 Alloy, of silver, of platinum ....75-80 Platinum 45-50 Specific Gravity. - This subject has been extensively investigated by Matthiesen, Calvert and Johnson, Kuppfer, and others. In discussing the results it is convenient to compare the values (S) found with the values (S0) calculated on the assumption that the volume of the alloy is equal to the sum of the volumes of the components. Let p„ p2, ps...stand for the relative weights of the components, P for their joint weight, S1,S2,S,...for their specific gravities, and we have = i.P2 So SI S2 • where the expression on the right hand obviously means the conjoint volume V, of the components ; but the actual volume of the alloy formed by their union is, in general, V = V,(1 + c), where e means the expansion (or, when negative, the contraction) of unit-volume of mixture. Hence the real value S=s,/(l+e), e=(So -S)/S.

Matthiesen's investigation (Fogy. Annalen for 1860, vol. ex. p. 21) extends over a large number of binary alloys derived from the metals named in the following table. Ile naturally began by procuring pure specimens of these metals and determining their specific gravities. The results (each the mean of a number of determinations) were as follows : - 58 lu these, as in all the subsequent determinations for the alloys, the weighings were reduced to the vacuum' and the values for S referred to water at 4° C. as unity. From eight metals twenty-eight different kinds of binary alloys can be produced ; of these twenty-eight combinations eighteen were selected; in each case the two components were fused together in a variety of properly chosen atomic proportions, and the specific gravities of these alloys were determined. The net results are summarized in the following table, which, for each combination A, 13, in the first two columns gives the composition in multiplies of the " atomie-weights " given in the table just quoted, while column 3 gives the values of c as calculated by the writer from Matthiesen's numbers for S, and S. Hence, for example, in the accompanying entries the first line shows that the union into an alloy of twice 118 parts of tin and once 197 parts of gold in- Tin and Gold.

volves an expansion from 1 volume into Sn I A e 1.004 ; the second that the union of once of gold involves a contraction from 1 4 - -028 58 To make these numbers trustworthy it would he necessary to determine their probable errors ; and this ISlatthiesen has not done. It would appear that any value of e from 0 to (say) b .002 counts for nothing, and anything up to .004 certainly must be taken as not proving much either way. If this is correct, then No contraction or expansion is proved iu the cases Sb, Bi; Cd, Bi ; Cd, Pb ; Au, Ag ; A contraction (from 0.5 to 4.7 per cent.) is proved for Sn, Ag ; Bi, Ag ; Bi, Au ; Pb, Au ; Pb, Bi; lig, Sn ; 11g, Pb ; Sn, Bi (9); Au, Ag (?); An expansion (from •5 to 0-8 per cent.) is proved for Sb, Sn; Sb, Pb ; Sn, Cd (?); Sn, Pb (?); certain cases of Sn, Au and Pb, Ag ; In the two series Sn, An and Pb, Ag, there are cases both of expansion and of contraction.

Thermic and Electric Properties. - The specific heat of an alloy, so far as we know, is always in approximate accordance with Dulong and Petit's law. Thns the specific heat of Cu5A1, is (5+1)x 5.4 with about the same degree of correctness as the "constant" 6.4 can claim for itself.

Expansion. - Mattbiesen, from numerous determinations made with alloys and their components, concludes that the expansion of an alloy (from 0° to 100° C.) is nearly equal to the sum of the expansions of its components. Supposing, for instance, one volume of gold to expand (from 0° to 1) by a, and one volume of silver by 0; then an alloy of four volumes of gold and three volumes of silver expands by (4a + 36)/7 per unit.

Fusibility. - In the ease of an alloy the melting-point and the freezing-point are, in general, separated by a greater or less interval of temperature, and the latter in itself may have two values as shown by lludberg, who found that when a fused alloy of tin and lead is allowed to freeze the thermometer becomes stationary at two successive points, as shown in the following table, where x means the number of atomic weights of tin united with p of lead in the given case, and the temperatures are in centigrade degrees.

First point (325') 280° 240° 187' 187' 210' (228°) Second point (325°) 187° 187° 187° 187' 187° (228°) We see that the first point varies with, while the second, within the range of the experiments, proved independent of, the proportion in which the two metals are -united.

The melting-point of many alloys lies below that of even the most fusible component, as illustrated in the following tables, where the numbers mean parts by weight.

58 All these alloys melt in boiling water.

The electric conductivity of alloys qua alloys has been investigated by Matthiesen. He confined himself to binary alloys derived from a certain set of elementary metals. The main results of his researches are given in ELECTRICITY, VOL viii. p. 51. For the practical electrician it is important to observe how very much the conductivity of copper is impaired by very minute admixtures even of metals that are good conductors, and also by non-metallic contamination, especially with oxygen (present as Cu20).

Hydrogen, as was shown by Graham, is capable of uniting with (always very large proportions of) certain metals, notably with palladium, into metal-like compounds. But those hydrogen alloys, being devoid of metallurgic interest, fall better under the heading PALLADIUM.

Oxygen. - Mercury and copper (perhaps also other metals) are capable of dissolving their own oxides with formation of alloys. Mercury, by doing so, becomes viscid and unfit for its ordinary applications. Copper, when pure to start with, suffers considerable deterioration in plasticity. But the presence of moderate proportions of cuprous oxide has been found to correct the evil influence of small contaminations by arsenic, antimony, lead, and other foreign metals. Most commercial coppers owe their good qualities to this compensating influence.

Arsenic combines readily with all metals into true arsenides, which latter, in general, are soluble in the metal itself. The presence in a metal of even small proportions of arsenide generally leads to considerable deterioration in mechanical qualities.

Phosphorus. - The remark just made might be said to bold for phosphorus were it not for the existence of what is called "phosphorus-bronze," an alloy of copper with phosphorus (i.e., its own phosphide), which possesses valuable properties. According to Abel, the most favourable effect is produced by from I to 4 per cent. of phosphorus. Such an alloy can be east like ordinary bronze, but excels the latter in hardness, elasticity, toughness, and tensile strength. See Puosruottus.

Carbon. - Most metals when in a molten state are capable of dissolving at least small proportions of carbon, which, in general, leads to a deterioration in metallicity, except in the case of iron, which by the addition of small percentages of carbon gains in elasticityand tensile strength with little loss of plasticity (see IRON).

Silicon, so far as we know, behaves to metals pretty much like carbon, but our knowledge of facts is limited. What is known as " cast iron " is essentially an alloy of iron proper with 2 to 6 per cent. of carbon and more or less of silicon (see IRON). Alloys of copper and silicon were prepared by Deville in 1863. The alloy with 12 per cent, of silicon is white, hard, and brittle. When diluted down to 4.8 per cent., it assumes the colour and fusibility of bronze, but, unlike it, is tenacious and ductile like iron.

To avoid repetition, let us state beforehand that the metals to be referred to are always understood to be given in the compact (frozen) condition, and that, wherever a series of metals are enumerated as being similarly attacked, the degree of readiness in the action is (so far as our knowledge goes) indicated by the order in which the several members are named, - the more readily changed metal always standing first.

Water, at ordinary or slightly elevated temperatures, is decomposed more or less readily, with evolution of hydrogen gas and formation of a basic hydrate, by (1) potassium (formation of Kilo), sodium (Na110), lithium (Li011), barium, strontium, calcium (13a021(2, &c.); (2) magnesium, zinc, Manganese (MgO2II2, &c.).

In the case of group 1 the action is more or less violent, and the hydroxides formed are soluble in water and very strongly basylous ; metals of group 2 are only slowly attacked, with formation of relatively feebly basylous and practically insoluble hydrates. Disregarding the rarer elements (as we propose to do in this section), the metals not named so far may be said to be proof against the action of pure water in the absence of free oxygen (air).

By the conjoint action of water and air, thallium, lead, bismuth are oxidized, with formation of more or less sparingly soluble hydrates (Th110, Pb02112, BiO3113), which, in the presence of carbonic acid, pass into still less soluble basic carbonates.

Iron, as everybody knows, when exposed to moisture and air, "rusts," that is, undergoes gradual conversion into a brown ferric hydrate, Fe203xH20 ; but this process never takes place in the absence of air, and it is questionable whether it ever sets in in the absence of carbonic acid. What is known is that iron never rusts in solutions of caustic alkalies or lime (which reagents preclude the presence of free carbonic acid), while it does so readily in ordinary moist air containing CO,. When once started the process proceeds with increasing rapidity, the ferric hydrate produced acting as a carrier of oxygen; it gives up part of its oxygen to the adjoining metal, being itself reduced to (perhaps) Fe304, which latter again absorbs oxygen from the air to become ferric hydrate and so on (Kuhlmann).

Copper, in the present connexion, is intermediate between iron and the following group of metals.

Mercury, if pure, and all the " noble" metals (silver, gold, platinum, and platinum-metals), are absolutely proof against water even in the presence of oxygen and carbonic acid.

The metals grouped together above under 1 and 2 act on steam pretty much as they do on liquid water. Of the rest, the following are readily oxidized by steam at a red heat, with formation of hydrogen gas, - zinc, iron, cadmium, cobalt, nickel, tin. Bismuth is similarly attacked, but slowly, at a white heat. Aluminium is barely affected even at a white heat, if it is pure; the ordinary impure metal is liable to be very readily oxidized.

Aqueous Sulphuric or hydrochloric Acid, of course, readily dissolves groups 1 and 2, with evolution of hydrogen and formation of chlorides or sulphates. The same holds for the following group (A) : - [manganese, zinc, magnesium] iron, aluminium, cobalt, nickel, cadmium. Tin dissolves readily in strong hot hydrochloric acid as SnCl, ; aqueous vitriol does not act on it appreciably in the cold ; at 150° it attacks it more or less quickly, according to the strength of the acid, with evolution of sulphuretted hydrogen or, when the acid is stronger, of sulphurous acid gas and deposition of sulphur (Calvert and Johnson). A group (B), comprising copper, are, substantially, attacked only in the presence of oxygen or air. Lead, in sufficiently dilute acid, or in stronger acid if not too hot, remains unchanged. A group (C) may be formed of mercury, silver, gold, and platinum, which are not touched by either aqueous acid in any circumstances.

Hot (concentrated) oil of vitriol does not attack gold, platinum, and platinum-metals generally ; all other metals (including even silver) are converted into sulphates, with evolution of sulphurous acid. In the case of iron, ferric sulphate, Fe2(SO4)3, is produced ; tin yields a somewhat indefinite sulphate of its binoxide Sn02.

Nitric Acid (Aqueous).--Gold, platinum, iridium, and rhodium only are proof against the action of this powerful oxidizer. Tin and antimony (also arsenic) are converted by it (ultimately) into hydrates of their highest oxides Sn02, S1320, (As205), - the oxides of tin and antimony being insoluble in water and in the acid itself. All other metals, including palladium, are dissolved as nitrates, the oxidizing part of the reagent being generally reduced to nitric oxide, NO, or sometimes to N203 or N204. Iron, zinc, cadmium, also tin under certain conditions, reduce the dilute acid, partially at least, to nitrous oxide, N,O, or nitrate of ammonia, NH4'NO3 = N20 + 21120.

Aqua Regia, a mixture of nitric and hydrochloric acids, converts all metals (even gold, the "king of metals," whence the name) into chlorides, except only rhodium, iridium, and ruthenium, which, when pure, are not attacked.

Caustic Alkalies. - Of metals not decomposing liquid pure water, only a few dissolve in aqueous caustic potash or soda, with evolution of hydrogen. The most important of these are aluminium and zinc, which are converted into aluminate, Al2033(K2 or Na2)0, and zincate, ZnO.RHO, where R = K or Na respectively. But of the rest the majority, when treated with boiling sufficiently strong alkali, are attacked at least superficially; of ordinary metals only gold, platinum, and silver are perfectly proof against the reagents under consideration, and these accordingly are used preferably for the construction of vessels intended for analytical operations involving the use of aqueous caustic alkalies. For preparative purposes iron is universally employed and works well ; but it is not available analytically, because a superficial oxidation of the empty part of the vessel (by the water and air) cannot be prevented. According to the writer's experience basins made of pure malleable nickel are free from this drawback ; they work as well as platinum, and rather better than silver ones do. There is hardly a single metal which holds out against the alkalies themselves when in the state of fiery fusion ; even platinum is most violently attacked. In chemical laboratories fusions with caustic alkalies arc always effected in vessels made of gold or silver, these metals holding out fairly well even in the presence of air. Gold is the better of the two. Iron, which stands so well against aqueous alkalies, is most violently attacked by the fused reagents. Yet tons of caustic soda are fused daily in chemical works in iron pots without thereby suffering contamination, which seems to show that (clean) iron, like gold and silver, is attacked only by the conjoint action of fused alkali and air, the influence of the latter being of course minimized in large-scale operations.

Oxygen or Air. - The noble metals (from silver upwards) do not combine directly with oxygen given as oxygen gas (02), although, like silver, they may absorb this gas largely when in the fused condition, and may not be proof against ozone, 03. Mercury, within a certain range of temperatures situated close to its boiling point, combines slowly with oxygen into the red oxide, which, however, breaks up again at higher temperatures. All other metals, when heated in oxygen or air, are converted, more or less readily, into stable oxides. Potassium, for example, yields peroxide, K202 or K204 ; sodium gives NO, ; the barium-group metals, as well as magnesium, cadmium, zinc, lead, copper, are converted into their monoxides MeO. Bismuth and antimony give (the latter very readily) sesquioxidc (Bi203 and Sb203, the latter being capable of passing into Sb204). Aluminium, when pure and kept out of contact with siliceous matter, is only oxidized at a white heat, and then very slowly, into alumina, Al203. Tin, at high temperatures, passes slowly into binoxide, SnO.,.

Sulphur. - Amongst the better known metals, gold and aluminium are the only ones which, when heated with sulphur or in sulphur vapour, remain unchanged. All the rest, under these circumstances, are converted into sulphides. The metals of the alkalies and alkaline earths, also magnesium, burn in sulphur-vapour as they do in oxygen. Of the heavy meals, copper is the one which exhibits by far the greatest avidity for sulphur, its subsulphide being the stablest of all heavy metallic sulphides in opposition to dry reactions. See METALLURGY.

Chlorine. - All metals, when treated with chlorine gas at the proper temperatures, pass into chlorides. In some cases the chlorine is taken up in two instalments, a lower chloride being produced first, to pass ultimately into a higher chloride. Iron, for instance, is converted first into FeC12, ultimately into Fe2,C16, which practically means a mixture of the two chlorides, or pure Fe.,C16 as a final product. Of the several products, the chlorides of gold and platinum (AuCl, and PtC14) are the only ones which when heated beyond their temperature of formation dissociate into metal and chlorine. The ultimate chlorination product of copper, CuC12, when heated to redness, decomposes into the lower chloride, Cu2C12, and chlorine. All the rest, when heated by themselves, volatilize, some at lower, others at higher temperatures.

Of the several individual chlorides, the following are liquids or solids, volatile enough to be distilled from out of glass vessels : - AsC13, SbC13, SnCl„ BiC13, HgCl„ the chlorides of arsenic, antimony, tin, bismuth, mercury re- spectively. The following are readily volatilized in a current of chlorine, at a red heat : - Al2C1,3, Cr2Cl5, Fe2C16, the chlorides of aluminium, chromium, iron. The following, though volatile at higher temperatures, are not volatilized at dull redness : - KC1, NaCl,. LiC1, Niel, CoC12, MnCl2, ZnC12, MgC12, PbC12, AgCl, the chlorides of potassium, sodium, lithium, nickel, cobalt, manganese, zinc, magnesium, lead, silver. Somewhat less volatile than the last-named group are the chlorides (MCL) of barium, strontium, and calcium.

Metallic chlorides, as a class, are readily soluble in water. The following are the most important exceptions: - chloride of silver, AgCl, and subehloride of mercury, Hg2C12, are absolutely insoluble ; chloride of lead, PliCl„ and subchloride of copper, Cu2C12, are very sparingly soluble in water. The chlorides AsCl„ SbC13, BiC13, are at once decomposed by (liquid) water, with formation of oxide (As203) or oxychlorides (SbC10, BiC10) and hydrochloric acid. The chlorides MgCl2, AI2C10, Cr2C15, Fe2C1G suffer a similar decomposition when evaporated with water in the heat. The same holds in a limited sense for ZnC12, CoCL, NiC12, and even CaCL. All chlorides, except those of silver and mercury (and, of course, those of gold and platinum), are oxidized by steam at high temperatures, with elimination of hydrochloric acid.

The above statements concerning the volatilities and solubilities of metallic chlorides form the basis of a number of important analytical methods for the separation of the respective metals.

For the characters of metals as chemical elements the reader is referred to the article CHEMISTRY and to the special articles on the different metals. (W. D.)