330 



M. Edlund on the Thermal Phenomena which 



same quantity as in the preceding case, in raising the same weight ; 

 3rd, the quantity of heat disengaged when the wire became 

 shortened without raising any weight. From the mechanical 

 theory of heat, the two first quantities ought to be equal, but less 

 than the third. 



To estimate these quantities of heat, M. Edlund used a thermo- 

 electric couple of a peculiar construction, which exactly measured 

 the diminution or elevation of temperature of the metallic wire 

 to which it was applied. It consisted of a crystal of antimony 

 and a crystal of bismuth (not represented in the figure) fixed in 

 two pieces of ivory, h and h 1 (fig. 1), which were themselves fixed 

 in two large brass 

 springs, m m, m l m*. 

 The crystals commu- 

 nicated on the inside 

 with the metal screws 

 I and /', and by means 

 of these screws with 

 the galvanometer. The 

 three screws n, ^'served 

 to bring together the 

 two springs mm and 

 m! m 1 , and thus to press the metal wire between the two crystals 

 without its being otherwise in contact with any part of the appa- 

 ratus. M. Edlund considers the substitution of defined crystals 

 for the melted and confusedly crystalline bars commonly used, as 

 an essential improvement. As the electromotive force of a bis- 

 muth-antimony couple may vary greatly with the nature of the 

 crystalline faces in contact, if in two successive experiments the 

 two crystals are not in contact with the metal wire in exactly the 

 same points, the results could not be compared. A Weber's 

 reflecting galvanometer was used with an astatic needle. 



The metal wire used was fixed at the top to the end of a short 

 bar of iron, a (fig. 2), which was itself supported by a strong 

 vertical beam A, fixed in a recess. At the bottom there was 

 a steel pincer /, perforated by a horizontal hole, which fitted 

 in the two arms of a metal fork supported by a wooden lever 

 a 1 a*, moveable about a horizontal axis. When a steel rod 

 was passed through two holes in the prongs of the fork, and 

 through the hole of the pincer /, the lower end of the wire 

 was virtually part of the lever. A weight q was suspended 

 to the lever on the other side of the axis b, and was placed in 

 such a position that the preceding operation could be effected 

 without deranging the lever, and consequently without elon- 

 gating the wire at all. A piece of brass, c, which supported a 

 weight d, produced elongations by the intervention of the lever 



