INTRODUCTION. W ITH the growth of the electron theory and its application to the flow of heat in metals the measurement of thermal conductivity becomes of greater interest. The determinations, up to the present time, have been confined within narrow limits of temperature and the object of this investigation was to determine the thermal conductivity of a few metals over a wide range. The first measurements at high temperatures were made by Forbes, 1 using the well-known " Forbes bar method," on

more »

... rought iron at temperatures of o°, ioo°, 200 0 and 275 0 C; giving a decrease in thermal conductivity from .207 at o° C. to .124 at 275 0 C. Tait 2 and Mitchell 3 used the same method, but found an increase in conductivity with temperature, which is undoubtedly in error, as shown by the later work of Jager and Diesselhorst. 4 Lorenz 5 heated a rod at one end by contact with a large slab and determined the temperature gradient by means of thermocouples inserted in holes along the rod; the source of heat was then removed and the temperature gradient determined when the rod was cooling. Between o° C. and ioo° C. he found an increase in conductivity with temperature for aluminum, brass, copper and german silver; a decrease for antimony, bismuth, cadmium, iron, lead and zinc; and a constant for magnesium. Baillie, 6 with the Forbes method, determined the thermal conductivity of nickel between o° C. and ioo° C. to be .132 and found this value to be changed only in the third decimal place at a temperature of 200 0 C. Jager and Diesselhorst 4 used a method suggested by Kohlrausch, 7 which consisted in heating a rod with a constant electric current; the ends of the rods being in contact with large baths, a steady flow of heat was maintained from the center to the ends. The temperature was measured at three points along the rod by means of thermocouples inserted in holes. Determinations were made 1