1887.] 



Conduction of Heat in Liquids. 



301 



temperature, was suddenly poured into the metal dish, and the time 

 noted. In one set of experiments this water was after a given 

 interval siphoned from the dish, in another set it was left undis- 

 turbed. In either case the variation in the temperature of the 

 platinum wire, as indicated by the change in its resistance, was 

 determined by observations of the readings of a delicate galvano- 

 meter, which was affected by the variation in an electrical current 

 traversing the platinum wire. The galvanometer readings supplied 

 data from which could be calculated the interval that elapsed after 

 the application of heat before the temperature in the liquid surround- 

 ing the platinum wire was rising fastest. An independent series of 

 experiments gave the rate at which heat passed into each liquid from 

 the dish. 



To calculate the conductivity a mathematical investigation is 

 carried out, which leads to an equation connecting the conductivity, 

 density, and specific heat of a liquid with the time elapsed after the 

 application of heat at the surface before the heating at a given depth 

 should be most rapid. Though this equation cannot be directly 

 solved, solutions of a close degree of approximation can be obtained. 

 The density and specific heat beug known, these solutions enable the 

 conductivity to be calculated in absolute measure. 



The liquids examined were water, paraffin and turpentine oils, 

 bisulphide of carbon, methylated spirit, and solutions of various 

 strengths of sulphuric acid and water. In the case of paraffin oil, 

 methylated spirit, and water, the two different methods were 

 employed, and the results agreed fairly well. In the case of turpen- 

 tine the water was never siphoned, and in the case of the remaining 

 liquids the siphon was always used. It was found that the con- 

 ductivity of the various sulphuric acid solutions, some of consider- 

 able strength, differed very slightly from that of water, and thus there 

 is a marked distinction between conducting powers for heat and for 

 electricity. The presence of small impurities in the liquids, such as 

 small quantities of salt, had no appreciable effect on the conductivity. 



The intervals that elapsed after the application of heat before the 

 temperature at the given depth was rising fastest did not differ very 

 largely for the various liquids. It was shortest for the bisulphide and 

 longest for turpentine. Owing to the comparatively small variation 

 in this interval the value of the conductivity depends largely on the 

 product of the density into the specific heat, a quantity to which it is 

 directly proportional. 



The values actually obtained are the following : — In those under 

 column 1 the water was siphoned from the dish, in column 2 it was 

 not. The units are centimetre and minute. 



z 2 



