The Conductivity of Heat Insulators. 



285 



power expended on internal heating was measured. The outer pot 

 was immersed in a tank kept overflowing from the water main, the 

 lid of the pot being made into a sort of saucer, into which the incoming 

 water ran, thus the outer pot's surface was kept at a uniform and 

 constant temperature. The resulting temperature differences were 

 measured by means of thermo-electric junctions of copper and iron, in 

 a way shortly to be described. The current was allowed to pass steadily 

 into the inner pot, driving the motor and fan, until the inner thermo- 

 electric junction arrived at a steady value ; this usually occurred in 

 about three hours or so; when this was the case, the supply of energy 

 by the current was just equal to the heat conducted through the insu- 

 lator and carried off by the water. Knowing the temperature gradient 

 and the number of watts supplied and the dimensions of the system, 

 we can deduce the specific conductivity of the material. 



The general arrangement is shown in Fig. 1. (P) is the outer pot 

 which stood in the tank j its approximate dimensions were 8 inches in 

 diameter and 1 6 inches high ; (Q) is the inner pot, with one inch clearance 

 Tjetween it and the outer one ; outside are shown the voltmeter (V), 

 .ammeter (A), and adjustable resistance (?•) ; inside the pot (Q) are a 

 fixed resistance (R), the motor (M), and fan (F). To a point about the 

 middle of each pot, inside the outer, and outside the inner, are soldered 

 the copper-iron junctions B, C, which are brought outside to a three- 

 way plug and a galvanometer, a is a junction placed in a vessel of 

 water at a known temperature ; g is a galvanometer of the Crompton- 

 D'Arsonval pattern. The junctions used were always made of exactly 

 the same length to keep the total resistance constant, and on calibra- 

 tion were found to give a linear relation between temperature difference 

 and deflection on the galvanometer within the ranges which were to 

 be used, and during each experiment a check calibration at two known 

 temperatures was made by means of a third junction in water at 

 another known temperature. When a steady state was attained, the 

 temperatures of B and C above A were measured by the deflections 

 and calibration tests. The reason for the above indirect method of 

 measuring the difference of temperature between B and C was to avoid 

 possible leakage currents. The flow of heat per second from the inside 

 pot to the outside one will be given by the expression H = XcO, where c 

 is the specific conductivity, $ is the temperature difference, X is a constant 

 depending on the size of the pots, being the area in square centimetres 

 of two plane surfaces, distant one centimetre apart, that would permit 

 the same flux of heat as the actual arrangement employed under the 

 same conditions of heat transference and temperature gradient. The 

 value of A. was calculated from careful measurements of their dimen- 

 sions, on the assumption that the flow of heat could be taken as radial 

 for the sides, and from the top and bottom of the inner, to the bottom 

 and top of the outer, pot ; this leads to the expression 



