LAWS, BISHOP, AND MCJUNKIN. — THERMAL CONDUCTIVITY. 463 



computed; then the E. M. F. of the mean temperature thus obtained is 

 used to again determine the difference of temperature. This method of 

 approximation is carried out until the difference of temperature obtained 

 on successive trials is the same. 



The thickness and diameter of the sphere were carefully measured 

 before it was electroplated by special micrometer calipers. 



The following results show the mean of thirty-six measurements. 



Thickness, 1.4327 inches; Diameter, 6.0022 inches. 



Method of taking Readings. 



The circulating fans on the interior of the sphere and the agitating 

 fans in the bath are first set in motion. Then the different junctions are 

 connected to the galvanometer in order to locate any outside E. M. F., 

 if such exists. The current is now turned on the heating coil and the 

 flow of water through the cooling coil adjusted until the temperature of 

 the bath remains constant. Meanwhile the galvanometer is calibrated 

 as follows: The galvanometer is shunted across a standard ohm which 

 is in series with a large resistance and storage cell. The resistance in 

 the storage cell circuit is varied until the desired deflection is obtained 

 (this being determined by the difference of temperature at which the run 

 is to be made). The E. M. F. of the storage cell is at once obtained by 

 the Poggendorf method, using a standard Carhart-Clark cell. When the 

 temperatures of the outside of the sphere and bath have become constant, 

 the terminals of the galvanometer are connected to junction No. 1. As 

 soon as the deflection is constant the readings of the galvanometer, 

 ammeter, voltmeter, and thermometer in bath are taken, also reading of 

 thermometer in bath with nickel-iron junction, giving the temperature of 

 the outside junction. The galvanometer is then reversed on No. 1 and 

 the readings again taken. 



The same method is followed on each junction, and then the whole 

 series repeated four times, after which the galvanometer is again cali- 

 brated. 



Results. 



The following results show the mean of four series of readings on each 

 junction ; i. e., 29.633 is the mean deflection on the four junctions, while 

 6.472 and 89.55 are the average amperes and volts during the run. 

 This is a typical set of readings. 



