DETERMINATION OF THE THERMAL CONDUCTIVITY OP MARBLE. 267 



heated board and ring- seemed to retard the cooHng of the ball. 



At first, it wa.s attempted to determine the temperature of the 

 centre by the method of compensation ; that is to say, the electro- 

 motive force in the thermometric circuit was balanced by an equal and 

 opposite electromotive force produced by a second junction inserted 

 into the circuit. The bath int<i which this circuit was put was heated 

 or cooled until the galvanometer gave no current. The temperature 

 of this second junction would then be the same as temperature at the 

 centi'e of the sphere. It was found, however, very difficult to mani- 

 pulate cooling or heating of this external bath so as to keep pace with 

 the changing temperature at the centre of the ball. The results 

 obtained were quite irregular. The second method tried was to 

 balance the electromotive force of the circuit (the junction not in the 

 ball being always kept at 0° C.) by a jwrtion of the electromotive 

 force of a Daniel's cell. A long wire was stretched to and fro on a 

 board a considerable number of times, and, with an additional resist- 

 ance of 40 ohms, was put in circuit with the cell. The electromotive 

 force of the thermo-junction could then be balanced by the ditference 

 of potentials between the two extremities of a portion of the wire. 

 The wire was gauged inunediately afterwards, so that the temperatures 

 at the centre could be at once deduced from the measured lengths of 

 the portions of wire. One of the specimens given below was carried 

 out by this method. 



The third method used was that of deflection ; that is to say, 

 the external junction Avas maintained at a constant temperature ; a 

 sensitive galvanometer was put in the circuit ; the position of the 

 spot of light in the galvanometer-scale was read from time to 

 time, the change of the zero point l)eing also observed. Directly 

 after an experiment, the galvanometer was gauged, so that the 

 temperature corresponding to any particular galvanometer reading 



