1894.] Thermal Conductivity of Metals, with Applications. 201 



always allowed for by applying an approximate formula to each 

 different length. 



In order to make a complete test of a metal it is only necessary to 

 take a wire of 5 or 6 cm. length and solder it firmly, the one end into 

 the bottom of the heating box, the other into the calorimeter ball. 

 The water in the heating box is kept boiling briskly, and readings 

 are taken every half minute from the thermometer in the ball. 

 These readings are then put upon a curve as ordinates, with the 

 time in minutes as absciss. From this curve the rise of temperature 

 per unit time can then be accurately read off, and, the thermal 

 capacity of the ball being already determined, the flow of heat per 

 unit time is obtained. 



In order to eliminate radiation from the surface of the calorimeter 

 ball, the latter is, at the beginning of the experiment, cooled to about 

 6 or 7 C. below the temperature of the air, or rather of the water- 

 jacket which surrounds the ball. 



Let a, be the quantity of heat that passes from the surface of the 

 ball, when the latter is 6 above or below the temperature of the 

 water-jacket ; Q! the quantity of heat that flows into the ball at the 

 temperature above that of the water-jacket ; Q z the quantity that 

 flows in when the ball is 6 below that of the jacket ; T the tempera- 

 ture of the hot end of the wire. 



Then if *: is the mean conductivity, 



If, therefore, the rise in temperature per half minute at 6 above 

 that of the water-jacket be taken from the curve and added to the 

 rise for 6 below the temperature of the jacket, the quantity ^(Qi + Qa) 

 is obtained, and is the flow of heat when the temperatures of the ends 

 of the wire are T J and t C., the radiation from the ball being thus 

 eliminated. If ten or fifteen of these values be taken from the curve 

 and the mean found, a very accurate result is obtained. It is thus 

 immaterial whether the surface of the ball changes between each test, 

 as long as it remains constant during the test. 



The metal which was chiefly used for the exhaustive tests of the 

 method was copper wire, of diameter 0'21 cm., density 8'85, volume 

 specific (electrical) resistance at about 13 C. 1834 in absolute units. 



The number obtained for the absolute value of the thermal conduc- 

 tivity was 0'88838 C.G.S. units, which was the mean of the values for 

 different lengths of from 4 to 7 cm., the greatest variation being a 

 little over 1 per cent. 



