52 Mr. H. H. Poole on the Thermal Conductivity 



amalgamate with thom. Only four out of the ten small tubes 

 are employed; but this is do disadvantage, and the key 

 could it' necessary be used for five couples. As the key in 

 effect forms the cold junction of the single couple, a know- 

 ledge of its temperature is necessary, and is obtained by 

 means of a thermometer placed in C. This key has always 

 proved very satisfactory, and no irregularities have ever been 

 traced to it. The other end of the hV\ leading from the key 

 is connected direct to the potentiometer, in using which due 

 care is taken to minimise thermal effects at hind— crew- and 

 contacts as far as possible. The galvanometer employed 

 wa9 an Ayrton and Mather giving a deflexion of about 

 17 scale-divisions per microvolt, but this instrument was 

 inconveniently sensitive. 



In the earlier experiments a sufficiently sensitive potentio- 

 meter not being available, the E.M.F.s of the couple- were 

 measured by balancing them against the P.D. across the 

 ends of a known resistance, either one ohm or ten ohm-. 

 through which a known very small current flowed. The 

 current was calculated from the known voltage of the dry 

 cell driving it and the resistance of the circuit, which was 

 always so large that the resistance of the cell might be 

 neglected. The resistance of the circuit was varied until a 

 balance was attained. The voltage of the cell was measured 

 daily by means of a Crompton potentiometer, and only varied 

 by about 0*2 per cent, during the whole period of nse. 



Shortly after the work was started a Tinsley potentiometer 

 was obtained reading to microvolts, and this was always used 

 subsequently. 



In heating up the furnace in the morning the choking-coil 

 is short-circuited until the required temperature is attained. 

 The supply of energy to the furnace is then adjusted by trial 

 until as steady a temperature as possible is attained. In the 

 case of the higher temperatures this could not be done until 

 the evening. Considerable trouble was experienced owing 

 to the variation in voltage of the city supply, A self 

 regulating arrangement tor overcoming this difficulty will 

 be described in a later paper. In the case of the determi- 

 nations at temperatures below int) C. no current waspa 

 through the furnace when a steady state had hern reached. 

 The lowest temperatures could only be attained by reducing 

 the supply of energy Erom the central wire. An approximate 

 correction for -mall temperature changes can be made as 

 Eollows. Denoting Eor the Euture fl as the average tem- 

 perature of the rock indicated by the single couple, and A 

 as the temperature difference indicated 1»\ the mul 



