572 HEAT solar spectrum (thoso of dark heat, those of luminous heat, and those of actinism), is re- flected from polished surfaces without any de- composition, we have another proof of the fact that the laws of reflection are the same for each. The reflecting power also of a surface for heat and for light is found by experiment to be the same, allowance being made for errors ; and the reflecting power of different substan- ces varies with the angle of incidence in the same degree for heat as for light. In glass it increases rapidly with the angle of incidence, while in metals it increases slowly. It is also found that heat is diffused and scattered by the same surfaces in the same proportion that light is. II. CONDUCTION OF HEAT. If a copper rod, fig. 7, is placed on supports and a flame applied at one end, heat will flow along it to- ward the other end, and the rate may be measured by thermometers having their bulbs placed in cup-shaped holes containing mercury along the upper side of the bar. This was the method of . Despretz, who made 'the first im- portant series of experiments on the subject. FIG. 7. If an iron bar is substituted for the copper, the rate of flow will be much less, and a bar of platinum will be found a poorer conductor than one of iron. But the results of Despretz have been found by the later experiments of Wiedemann and Franz not to be perfectly ac- curate. The results obtained by these investi- gators are given in the following table, which also gives the electric conductions of the same metals, according to Riess and Lenz, these being very nearly the same as for heat, a fact which was first shown by Forbes : KATES OF THERMAL AND ELECTRIC CONDUCTIVITY. METALS. TlU.l'.M M CONDUCTIVITY. ELECTRIC CONDUCTIVITY. Wiedemann and Franz. Rie. Lenz. Stiver 100-0 73-6 53-2 23-6 14-5 11-9 8-5 6-4 1-8 100-0 66-7 59-0 18-4 10-0 12-0 7-0 10-5 100-0 78-3 58-5 21-5 22-6 18-0 10-7 10-8 1-9 Copper. . . Gold Brass Tin Iron . Lead Platinum Bismnth One of the sources of error in Despretz's ex- periments was the employment of the holes in the bar containing mercury, and another, a want of sensitiveness in the thermometer. Wiedemann and Franz used smooth rods, and measured the temperature with a thermo-elec- tric pile and galvanometer. Marble and mine- ral substances generally are poorer conductors than any of the metals, and porcelain and glass are still poorer. The property of the thermal conductivity of metals is the basis of the inven- tion of Davy's safety lamp for miners. (See LAMP.) The unequal conductivity of metals and other bodies is the cause of an interest- ing phenomenon, which is beautifully exhibit- ed by what is known as Trevelyan's experi- ment, but which had been previously observed when a hot metal of good conducting power was laid against a cold one, a comparatively poor conductor, particularly if it had consider- able expansion, as a copper brazing iron laid upon a block of cold lead. When the heated metal can readily acquire a slight rocking mo- tion, the experiment succeeds the best. Tre- velyan's apparatus consists of a "rocker" made of brass, having a longitudinal groove, and lying upon the cylindrical surface of a block of lead. When the rocker is heated and placed upon the lead, the ridges on each side of the groove are alternately thrown upward by the expansion in those parts of the sur- face of the lead which are heated by coming in contact with the hot brass, and thus a series of vibrations having a musical tone is produced. The reason why the heated metal should be a good conductor is that its surface may be kept hot uniformly with the mass, and thus be in a condition to impart sufficient heat to the surface of the lead at every moment. The advantage of employing lead as the other metal consists in its being capable of consider- able expansion by heat, and in its being a poor conductor, so that in a moment the superficial portions may acquire enough heat to cause the requisite expansion to throw the rocker into vibrations. The same effect may be produced if, instead of a block of lead, one of stone is cov- ered with a thin sheet of metal which is a good conductor, the condition required being one fa- vorable to the rapid expansion of the surface, as was shown by Faraday. Other materials be- sides metals may be used, as various rocks and minerals. Liquids are almost non-conductors of heat, as may be shown by pouring a small quantity of alcohol upon the surface of water in a tumbler and igniting it ; a long time will elapse before the upper layers of the water become sensibly heated. Despretz employed an apparatus which consisted of a cylindrical wooden vessel about 3 ft. in height and 8 or 10 in. in diameter, which was filled with water. Through the sides of the cylinder 12 ther- mometers were placed, with their bulbs one above another in the axis of the column of water. A metal box, which was kept filled with water at 212 F., rested upon the top of the column of water. In this manner he found that the conductivity of heat for liquids follows
