Thermo-electric Powers of Metals and Alloys. 99 



temperatures intermediate between fixed and known tempe- 

 ratures, such as that of the boiling-point of liquid oxygen and 

 the mixture of solid carbonic acid and ether, was to embed the 

 thermo-junctions in a mass of paraffin wax, and this mass 

 having been cooled down to about — 200°, was allowed very 

 slowly to heat up again. 



If a platinum thermometer with copper leading-in wires is 

 immersed in this way in paraffin wax, to determine the tem- 

 perature of the junctions, then, as the thermal conductivity of 

 the copper is greatly increased by cooling it, heat will very 

 rapidly flow into the thermometer along the leading-in wires ; 

 and it was found by preliminary experiments quite impossible 

 under these conditions to keep the temperature of the platinum 

 wire used as a thermometer down to the same temperature as 

 the thermo-junction in contact with it. By employing an 

 alloy like platinoid of a rather high specific resistance, and by 

 making the cross section of these platinoid leading wires rela- 

 tively somewhat small, it was found possible to greatly hinder 

 the passage of heat along the legs of the thermometer, and yet 

 by the employment of an identical blank conducting circuit to 

 determine the exact temperature of the platinum wire itself 

 from its measured resistance. 



Returning, then, to the construction of the thermometer, we 

 may mention that it was constructed of a length of platinum 

 wire given to one of us by Professor Callendar, and that this 

 specimen was remarkable for its relatively large temperature- 

 coefficient. In order to determine the resistance, and therefore 

 the temperature, of this platinum thermometer, a slide-wire 

 bridge was constructed. A length of two metres of uniformly 

 drawn manganin wire was stretched over a two-metre scale. 

 This wire constituted the bridge slide-wire. It had a diameter 

 of '0193 centim. (No. 36 s.w.G.) and a resistance at 15° C. of 

 26'576 ohms. It was very carefully calibrated for resistance 

 per unit of length. Over this slide-wire moved a contact- 

 maker ; and the other two arms of the bridge consisted of a 

 coil of manganin wire wound on a small bobbin and having a 

 resistance of 5*000 ohms at 15° C, and the above-described 

 platinum thermometer or the resistance representing its 

 leading-in wires. The battery employed with the bridge was 

 a two-cell lithanode secondary battery, and the galvanometer 

 a Holden-Pitkin suspended-coil galvanometer of 4 ohms 

 resistance. Battery- and galvanometer-keys were inserted in 

 the circuits as usual. 



The first step in the investigation consisted in determining 

 the resistance of the platinum thermometer attached to tin* 

 bridge, which we shall speak of as the working thermometer 

 P 2 , at temperatures defined by our standard platinum thermo- 



H 2 



