560 ELECTRIC CONDUCTIVITY OF METALS AND ALLOYS. [276. in the metal had a very marked effect in reducing the conductivity (Matthiessen and Holzmann, Phil. Trans. 1860, 85). Indeed, there are few metals more easily affected in conductivity by slight traces of impurity than copper, so that very great differences in conductivity are observed in wires drawn from different samples of what would be regarded as good commercial copper. The conductivity of an alloy is generally below that of the mean of its component metals. This is seen in the alloy of antimony and tin; but the alloys of tin and lead, tin and zinc, zine and cadmium, give a conductivity almost exactly the mean of those of the component metals, allowing for the proportion of each that is present (Matthiessen, Phil. Trans. 1860, 161). A similar fact was observed by Calvert and Johnson (149) in the conductivity for heat of some of these very alloys. An elaborate paper on the influence of temperature on the conductivity of alloys, by Matthiessen and Vogt, will be found in the Phil. Trans. for 1864, 167. Lenz found that all the metals continued to decrease in couductivity as the temperature rose to 200° C., and Dr. Robinson proved that this diminution continued as they were raised progressively to a red and even to a white heat. The non-metallic bodies appear to increase in conductivity as the temperature rises, for Matthiessen found that the conductivity of graphite and of coke was increased by heating them, that of gas coke rising about 12 per cent. between the ordinary atmospheric temperature and a 'light' red heat. Hittorf obtained an analogous result with selenium. Comparing the conductivity at ordinary temperatures of different forms of carbon with that of silver at o° C. as 100, Matthiessen obtained the following values: 277.] ELECTRIC CONDUCTIVITY OF SOLIDS. 561 (277) If equal quantities of electricity, whether obtained from the voltaic battery or from the electrical machine, be made to traverse wires of different metals of equal length and diameter in the same interval of time, the rise of temperature in the wire is inversely proportional to its conductivity, and therefore the better the conductor the less heat does it emit. The general truth of the fact may in the case of voltaic electricity be rudely but strikingly demonstrated by taking a wire of silver and one of platinum, each of exactly the same diameter, and forming them into a compound wire consisting of alternate links of the two metals. A current of electricity may be transmitted through this compound wire, of such a strength as to heat the platinum to visible redness, whilst the silver links will exhibit no such intense heat, though each link of the wire, from the form of the experiment, must transmit equal quantities of electricity in equal times. It has been ascertained that the heat developed at any part of the circuit is proportional to the square of the current multiplied into the resistance at that particular point. For the same wire the rise of temperature is proportioned to the square of the current, and this is true also for liquid conductors. Andrews (Proceed. Roy. Irish Acad. 1841, 465), found that when a fine platinum wire was traversed by a current from one of Daniell's constant batteries, the ignition of the wire varied in intensity by varying the gas with which he surrounded the wire. This wire was enclosed in a glass tube, which could be filled at pleasure with the different gases in succession. It was found that gaseous sulphurous anhydride and hydrochloric acid had a smaller cooling power than atmospheric air. Nitrogen, carbonic oxide, cyanogen, carbonic anhydride, nitric oxide, nitrous oxide, oxygen, and aqueous vapour, had nearly the same effect as atmospheric air. Olefiant gas, ammonia, the vapour of alcohol and of ether had a greater cooling power; and hydrogen a far greater cooling power than any of the others. The same subject has also been investigated by Grove (Phil. Trans. 1849, 49). FIG. 231. H A b The following experiment illustrates the cooling power of hydrogen very clearly. Take three pieces of stout copper wire, bend them into the form shown at w w w, fig. 231, and attach them to a weighted board, by which the lower part of the bends can be sunk beneath the surface of water contained in a shallow vessel. At a 562 UNEQUAL COOLING EFFECTS OF DIFFERENT GASES. [277 a and b, where the wires project above the surface of the water, complete the connexion by means of spirals of fine platinum wire, both spirals being equal in length, and each cut from the same wire. Each spiral will thus oppose an equal resistance to the passage of the current. When a voltaic current of a certain strength is transmitted through the wire, w w w, each spiral, consequently, becomes heated to the same degree of visible ignition. But if two similar jars, one, a, filled with air, the other, H, filled with hydrogen, be inverted over them, the wire in the jar H immediately ceases to be luminous, while that in a becomes more intensely ignited. This superior cooling power of the hydrogen is no doubt mainly due to the superior mobility of the particles of the gas over those of air (152, 160). The experiment was varied by enclosing the wires a and b in separate glass tubes, and sealing them up, one in an atmosphere of air, the other in an atmosphere of hydrogen. Both were then included in the same circuit, so that they should transmit equal currents of electricity. Before transmitting the current, however, each tube was immersed in a separate vessel which contained a given quantity of water, the temperature of which was accurately observed. After the current had been allowed to pass for a certain time, the temperature of the water which surrounded each wire was again observed, and it was found that the water around the tube which contained air was considerably hotter than that surrounding the tube filled with hydrogen. : This result, paradoxical as it appears, and as it seems to have been regarded by Grove, must necessarily follow from the operation of two principles which have already been explained the first of these is, that the resistance offered by a metal to the passage of electricity is diminished by reducing the temperature; and the second is that the heat evolved by a current in passing through a conductor is inversely as the resistance which it experiences. Now, in this experiment, the primary effect of the hydrogen is the cooling of the conducting wire; and the consequence is that this cooled wire, in transmitting the same current as a similar wire in air, offers less resistance, and less heat is therefore evolved by the wire surrounded by the hydrogen than by the wire which is surrounded by air. (278) Electric Conductivity of Liquids.-Liquids are very inferior to solids in conductivity; indeed, the difference between the two classes of bodies is so extreme that it is difficult to institute an accurate comparison between them. The attempt, however, has been made by Pouillet (Comptes Rendus, 1837, iv. 278.] ELECTRIC CONDUCTIVITY OF LIQUIDS. 563 785) assuming as the unit of comparison the conductivity of a solution of cupric sulphate, saturated at 15° C. he gives the following as the relative conductivity of the undermentioned solutions : Saturated solution of cupric sulphate Ditto, diluted with an equal bulk of water. Ditto, with, of nitric acid Ι 064 6.44 0'31 0'0025 0'015 2,500,000'00 The conductivity of a platinum wire, of a diameter and length equal to that of the interposed columns of liquid, is probably estimated too high. Since these results of Pouillet's were published, the subject of the conductivities of liquids has been resumed by E. Becquerel (Ann. de Chimie, 1846 [3], xvii. 242). He states that saline solutions may be divided into two classes; in the first, the conductivity increases progressively in proportion to the strength of the solution, until it becomes saturated; cupric sulphate and sodic chloride affording instances of this kind: whilst in the second class, of which cupric nitrate and zincic sulphate may be taken as examples, the conductivity increases with the degree of concentration up to a certain point, beyond which it diminishes as the solution becomes more nearly saturated. The salts which exhibit this peculiarity are either deliquescent or extremely soluble. The table on p. 564 contains a few of Becquerel's results. The saline liquids are to be considered as saturated unless otherwise specified. It is not surprising that differences so considerable should be observed between the conductivities of liquids and those of solids; for the processes of conduction in the two cases are essentially different. In liquids chemical decomposition and free movement of the component particles are indispensable, whilst nothing of the kind takes place in solids. The effects of heat are even inverted in the two cases; for experiment shows that as the temperature rises, the conductivity of the liquid increases rapidly; according to Becquerel, the conductivity of many solutions at 100° C. is three or four times as great as that of the same solution at o°. These phenomena, therefore, are the reverse of those presented by most solids. Exceptions, however, occur: Faraday has shown that argentic sulphide, when cold, is an 564 ELECTRIC CONDUCTIVITY OF GASES. [278. insulator, but on warming it gently it begins to conduct, and when hot it affords a spark like a metal; at a point a little below redness it conducts sufficiently to maintain its conductivity by the heat produced by the current which it transmits. Plumbic sulphide and fluoride, as well as mercuric iodide, also exhibit the same peculiarity. Glass, when cold, is an excellent insulator of the electricity developed by friction, but when heated it conducts, and when red hot it possesses scarcely any insulation; the same thing is also true of the tourmaline. Gutta percha is a much better insulator at o° C. than at 40°, and the same thing is observed, though to a much less extent, in the case of caoutchouc. Most of these cases have been traced to a partial chemical decomposition of the compound under the influence of softening by heat (Beetz, Phil. Mag. 1854 [4], viii. 191). When liquefied by heat, these compounds nearly all undergo chemical decomposition, and allow the current to pass freely. (279) Electric Conductivity of Gases.-Gases are almost perfect insulators of the voltaic current; although some feeble indications of conductivity have been discovered by Andrews, as well as by Hankel, by E. Becquerel, and by Buff, in a highly rarefied atmosphere, between metallic surfaces strongly ignited and in close approximation; and Magnus finds that small as is the conductivity of gases, they differ in degree in this respect, hydrogen surpassing other gases and vapours. Grove has further shown (Phil. Mag. 1854 [4], vii. 47) that |