the same piece of metal in contact with the galvanometer, but suddenly, towards the cold end of the other piece of metal. Consequently, in heated platinum, the conducting power is most diminished for positive electricity; since this kind of electricity is least conducted by the platinum. when cold; in antimony, the power of conducting negative electricity suffers the greatest diminution. The heat accumulated at one end of the piece of metal endeavours to diffuse itself uniformly through the whole metallic circuit, but can only do so gradually as a whole; its two elements, the two electricities, would flow rapidly round the circuit. Since now the conducting power is different on the opposite sides of the hottest point, the heat is decomposed; and in the case of platinum, the positive electricity goes from the hot end to the cold, and thence into the galvanometer, while the negative electricity goes from the hot piece of metal (which from that cause has less power of conducting positive electricity) into the galvanometer; so that the two electricities recombine and form heat in the colder part of the metallic circuit. With antimony, in which the conducting power is greatest for positive electricity, the direction of the current is the reverse of that just described. This action continues till, by the passage of heat into the cold end, the rate of decrease of the heat becomes the same on both sides of the hottest point, and consequently, there is no longer any inducement for the positive electricity to go one way, and the negative the other; for this reason, no current is produced when a continuous metallic arc is heated at any point whatever.]— Vid. p. 316.

Even when the heated end is separated from the cold end by a body in the state of igneous fusion, an electric current is produced in the same direction. If borax, carbonate of soda, carbonate of potash, chloride of potassium, iodide of potassium, sulphate of soda, or chloride of strontium in the state of igneous fusion, be placed between two platinum wires separated by a little distance from one another, and one of the wires be heated more strongly than the other, the positive electricity will go from the colder wire through the galvanometer to the hotter wire. The current is sufficiently strong to decompose iodide of potassium under favourable circumstances. If five platinum wires be joined together by four balls of fused borax, and the four ends of the wires heated always to the right of the borax, a current will be produced sufficient to evolve gas when passed through dilute sulphuric acid. The borax and the other salts suffer no chemical change. Even boracic acid, which conducts less readily than borax, produces a distinct electric current. When the carbonate of soda is not perfectly fused, the direction of the current is reversed. If the platinum wires are separated by glass, the positive electricity passes from the hot platinum through the galvanometer to the cold, when the plate of glass is thin, or when it is thick and strongly heated; but in the opposite direction when the glass is thick and moderately heated. A strongly heated film of mica interposed between the platinum wires likewise allows a little positive electricity to pass from the cold wire through the galvanometer to the hot wire (Andrews). Also in fused (but not red-hot) phosphate of soda and ammonia, sulphate of copper and potash, anhydrous acetate of soda, nitrate of ammonia, borax, bichromate of potash, acetate of potash, and nitrate of soda, two unequally heated platinum wires produce an electric current, strongest in the first, and weakest in the last-mentioned of these salts (R. Böttger, Pogg. 50, 53). If a hot and a cold platinum wire connected with a galvanometer be dipped in water, or if, when the two wires are placed in the same

vessel, hot water be poured on the one and cold on the other, positive electricity will pass for a short time from the cold wire through the galvanometer to the hot wire. (Nobili.)

b. With Two Metals.

When two pieces of different metals connected with a galvanometer are united by soldering or by intimate contact, an electric current is set up, as soon as the point of junction is brought to a temperature different (either higher or lower) from that of the rest of the circuit. (Seebeck.) The strength of the current appears to be in direct proportion to the difference of temperature; its direction depends upon the nature of the metals employed. The metals may be arranged in a thermo-electric series, such, that each metal-when connected with the one on its left handtransmits negative electricity, and when connected with that on its right hand, positive electricity, from its heated point to the galvanometer,-the point of junction being supposed to be heated, and vice versa when it is cooled. The strongest current is produced by a circuit of bismuth and antimony, these metals standing at the extremities of the series.

According to Yelin, the order is: Bismuth, silver, platinum, copper, gold, tin, lead, zinc, iron, antimony.

According to Becquerel: Bismuth, platinum, lead, tin, gold, silver, copper,-zinc, iron, antimony.

According to Cumming: Galena, bismuth, mercury and nickel, platinum, palladium, cobalt, silver alloyed with copper and manganese, tin, lead, brass, rhodium, gold, copper, osmium-iridium, silver,-zinc, charcoal, graphite, iron, arsenic, antimony.

The dash separates the thermo-positive from the thermo-negative metals; the conjunction and is placed between two metals of equal thermoelectric power. The discrepancies in the statements of different experimenters induce the supposition that the direction of the current is affected by impurities in the metals, and varies with the difference of temperature. That the latter may really be the case, will appear from the following.

When iron is moderately heated in contact with copper, silver, gold, brass, or zinc, positive electricity proceeds from the iron through the galvanometer to the copper, &c., but when the heat is stronger from the copper, &c., to the iron. (Cumming.) The reversal of the direction of the current takes place at a dull red heat, when copper and iron are the metals employed. (Becquerel.)-Zinc and gold produce at 70° a feeble current, which passes from the zinc through the galvanometer to the gold; at 150°, this current ceases; at 180° an opposite current sets in, and becomes very strong at 275°. On the contrary, with zinc and silver at a low temperature, positive electricity passes from the silver through the galvanometer into the zinc. This current attains its greatest strength at 120°, diminishes at a higher temperature, ceases at 225°, and is reversed at still higher temperatures. (Becquerel.)

The farther two metals are separated in the thermo-electric series, and the higher the temperature of their point of junction, the more powerful is the electric current; but it is always small in quantity, and of very feeble tension.

[Since bismuth conducts negative electricity better than positive, and antimony conducts positive electricity better than negative, it follows, from the hypothesis laid down on page 321, that, on heating the point of junction of these metals, the negative electricity will pass through the bismuth, and the positive through the antimony, towards the colder part of the metallic

circuit. When, on the other hand, the point of junction is made colder than the rest of the circuit, heat is decomposed in the warmer part, sexing negative electricity through the bismuth, and positive through the antimony, towards the point of junction; hence a reverse current is excited. The effect produced in a circuit of bismuth and antimony is likewise obtained with any other pair of metals, one of which, like platinum or copper, has greater conducting power for positive electricity, the other, like zine or iron, for negative electricity. In the case of two metals, both of which conduct one kind of electricity-e. g., negative clectricity-better than the other, it must be observed that this power belongs to the two metals in different degrees, e. g., to bismuth more than platinum, and to platinum more than copper. In such a case, the negative electricity will always travel through that metal which has relatively the greater power of transmitting it, and the positive electricity through the other, which relatively offers it the least resistance. The reversal of the direction of the current at certain temperatures seems to imply, that the power of a metal to conduct one kind of electricity in preference to the other varies with the temperature.]

By combining a number of thermo-electric pairs into a thermo-electric Battery, Chain, or Pile, a powerful thermo-electric current may be obtained: e. g., a number of bars of bismuth and antimony, or of platinum and iron, alternately soldered together, and heat applied to the first, third, fifth, &c., points of junction, whilst the second, fourth, &c., are kept cold. In this manner, the thermo-electric current produces, not only deflection of the needle, but likewise chemical decompositions of liquids (Botto, Pogg. 28, 238), and heating effects,—so that even when one of the conducting wires of the battery is cooled by immersion in ice, the point of junction of that wire situated without the ice becomes sensibly warmed. (Watkins, Phil. Mag. J. 14, 82; also Pogg. 46, 497.)-Melloni's Thermomultiplier or Thermoscope (p. 214) is also a pile formed in this manner of bars of bismuth and antimony, and connected with a galvanometer. The rays of heat fall on the 1st, 3rd, 5th, &c., points of junction, whilst the 2nd, 4th, &c., are kept cool. The smallest difference of temperature between the two sets of junctions suffices to produce deviation of the needle.

A thermo-electric pile of bismuth and antimony, which exhibited no action on the magnet, when heated, acquired this power permanently when previously dipped for a second in nitric acid of 14 specific gravity. (Döbereiner, Pogg. 49, 588.)

G. Electricity by Friction: Common Electricity.

Solid bodies become oppositely electrified when rubbed on one another, and to a certain extent when merely pressed together and afterwards separated. Homogeneous bodies for the most part exhibit this result only when their surfaces are in different states, or when they differ in temperature; with heterogeneous bodies, the development of electricity is much greater.

When two parallel faces, natural or artificial, of a mineral are pressed between the fingers, the mineral frequently becomes electrical on these surfaces, generally showing positive electricity. Calcspar retains the electricity thus developed from three to eleven days, topaz and fluor-spar several hours, mica one or two hours, and rock-crystal for a shorter time. Tale must be insulated in order to render it electrical; heavy spar and gypsum are not sensibly electrified. (Hauy.)-According to Becquerel, not a trace

of electricity is perceptible while bodies are being pressed together; it is not till they are separated, that the one appears positively, the other negatively electrified. A slice of cork becomes positively electrified when pressed against a slice or plate of caoutchouc, orange-peel, retinasphalt, coal, amber, zinc, copper, silver, cyanite, or heated double refracting spar, these substances at the same time becoming negative. On the contrary, the slice of cork becomes negative with all dry animal substances, with heavy spar, gypsum (which must be freed by drying from hygroscopic water), fluor-spar, and double refracting spar, not heated,-those substances at the same time becoming positive. Two good conductors pressed together exhibit, when separated, no other electricity than that developed by mere contact. Two similar bodies do not become electrical by pressure unless one of them is at a higher temperature than the other; and then the hotter body always becomes negative, the colder positive. The strength of the electricity thus developed depends upon the nature of the bodies, the state of their surfaces, the intensity of pressure, and the rapidity of separation. Cork produces more electricity with calespar, when the pressure is exerted on one of the faces parallel to the cleavage-planes, than with heavy spar, with the latter more than with polished rockcrystal, and with this more than with gypsum, or with the polished surface of calcspar, and-when pressure, temperature, dryness, and polish, of the cleft surfaces are equal,-three times as much with calcspar as with gypsum. The intensity of the electricity varies directly as the pressure,-so that when the latter is doubled, the former is doubled also. Lastly, if the two bodies which have been pressed together are slowly separated, the two electricities have time to reunite, and a much smaller quantity remains in the free state after the separation. Bodies rendered electrical by pressure retain the electricity for a longer time, in proportion as their insulating or non-conducting power is greater.

These phenomena are ascribed by Becquerel to the same cause as that which produces the following effects, observed by himself. When two laminæ of a crystal of mica are suddenly torn asunder, there is not only an appearance of light produced, but one lamina becomes positively, the other negatively electrified: if again pressed together and subsequently separated, they again appear electrified. Also, on the cleavage of calcspar, fluor-spar, heavy spar, topaz, talc, and dry warmed gypsum, and on tearing a playing-card into its two sheets, the separated lamina appear oppositely electrified. In topaz, whose cleavage takes place parallel to the terminal faces of the prism, the cleft surface belonging to one end of the prism exhibits, sometimes one kind of electricity, sometimes the other. When melted shellac is poured upon glass and pulled off after cooling, both become electrified.

Non-crystalline bodies, such as sealing-wax or glass, exhibit no electricity when broken. (Vid. Crell. Ann. 1786, 1, 325.) In the former cases, the separation of bodies united by pressure produced electricity; in the present case, it is a union produced by cohesion that is overcome. (Becquerel.)

If sulphate of copper and potash be brought into a state of red-hot fusion in a platinum crucible-connected, by a ring of wire on which it rests, with a Bohnenberger's electrometer-and then left to cool, no electricity will be apparent during the crystallization; but as soon as the crystalline mass begins to contract with an audible decrepitation, and falls to pieces gradually and spontaneously, the formation of each new fissure is accompanied by a development of electricity, and the effect gees on till

the whole mass is crumbled into dust. (R. Böttger, Pogg. 50, 43.)-Two sheets of unsized paper, impregnated with hot wax, stick fast together when rubbed with a leather cushion; and, on being pulled asunder while warm, are found to be electrified sufficiently to produce a spark. (Simon, Br. Archiv. 31, 216.)

According to Hauy, the following substances become positively electrical when rubbed with woollen cloth, and are likewise insulators: Diamond, sulphate of potash, nitre, common salt, witherite, heavy spar, strontianite, double refracting spar, arragonite, apatite, anhydrite, gypsum, glauberite, fluor-spar, boracite, bitter spar, sulphate of magnesia, sapphire, spinell, chrysoberyl, quartz, zircon, emerald, euclase, topaz, iolite, cyanite, chiastolite, axinite, tourmalin, garnet, cinnamon stone, vesuvian, felspar, prehnite, mica, apophyllite, hornblende, actynolite, tremolite, augite, diopside, epidote, nephelin, mesotype, stilbite, analcime, and, in short, probably all combinations of earths, alkalis, and mineral acids among themselves occurring in the mineral kingdom, with the exception of tale: moreover, electric calamine, zinc-spar, tinstone, lead-spar, sulphate of lead, sphene, and tungstate of lime.-The following become positively electric, and are conductors: Bismuth, zinc, lead, copper, brass, silver, and silveramalgam. Non-conductors becoming negatively electrical, are: Sulphur, talc, anatase, titanschorl, uranite, cobalt-bloom, orpiment, blende, phosphate of lead, molybdate of lead, chromate of lead, specular iron, green vitriol, prussian blue, cube-ore, red oxide of copper, malachite (which sometimes becomes positive), blue carbonate of copper, pseudo-malachite, blue vitriol, dioptase, chrysocolla, arseniate of copper, cinnabar, subchloride of mercury or horn-quicksilver, red silver, amber, retinasphalt, elastic petroleum (Erdharz), honey-stone, and anthracite.-Conductors becoming negatively electrical are: Ilvaite, allanite, tantalite, yttro-tantalite, wolfram, sulphuret of molybdenum, chrome iron ore, pitch-blende, peroxide of manganese, sulphuret of manganese (mangan-glanz), earthy cobalt, arsenic, antimony, native sulphuret of antimony, black tellurium, bismuth-glance, tin, tin-pyrites, galena, iron, graphite, magnetic iron ore, red hæmatite, magnetic pyrites, native ferroso-ferric sulphate, iron pyrites, arsenical pyrites, copper-glance, grey copper, nickel, copper-nickel, native amalgam of tin, vitreous silver, antimonial silver, gold, platinum, and palladium.— When the crystalline structure is imperfect, the property of becoming electrical by friction is often absent, and so likewise is the insulating power; e. g., calc-spar and statuary marble. Want of polish on the surface also frequently interferes with the insulating power, and produces negative instead of positive electricity, when the substance is rubbed, e. g., in quartz.

Insulated metals rubbed with the dry hand, or with horn, ivory, wood, or cork, become electrical,-rhodium, platinum, palladium, gold, cobalt, nickel, tellurium, and antimony, always negative, the last very strongly, silver, copper, brass, and tin, mostly negative; zinc and iron, sometimes negative, sometimes positive; lead and bismuth (the latter in the highest degree) positive. It is to be supposed that all metals, if perfectly clean, would become negative; but since the more oxidable soon become covered with a thin film of oxide, a portion of this oxide often adheres to the rubber, and the metal becomes positive by being rubbed with it; or, if the film of oxide holds faster, the friction takes place between this oxide and the wood, &c., in consequence of which, the metal becomes negative. Freshly polished metals rubbed with new wood always become negative, except lead and bismuth, which are sometimes rendered positive. (De la Rive.)