Resistance of Metals and Alloys. 281 



lower the temperature of the resistance-coils to about — 182° C. 

 The liquid gas was contained in vacuum-jacketed flasks or test- 

 tubes in which the resistance-coil was placed. 



Temperatures lower than the above limits were obtained 

 by closing the mouth of the vacuum-jacketed test-tube with 

 a tightly fitting indiarubber cork and producing a partial 

 vacuum above the liquid oxygen. In these experiments the 

 resistance-coils were therefore immersed in liquid oxygen 

 boiling under a very reduced pressure, as low as 14 milli- 

 metres of mercury. At all times great care is necessary to 

 preserve the coil entirely immersed in the liquid gas. If the 

 coil is only partially immersed thermoelectric effects are set 

 up, in some cases of such a magnitude as to entirely vitiate 

 the measurements of resistance. Hence the necessity for an 

 abundant supply of the liquid gas. In some of the sets of 

 experiments several litres of liquid oxygen were used up. In 

 the above set of measurements about fifty coils in all were 

 wound, and the measurements of each coil often taken several 

 times at the same temperature. In those instances in which 

 the temperature of the bath was below zero Centigrade the 

 resistance of one of the standard platinum wires was measured 

 in the same bath before and after each measurement of the 

 resistance of the wire. The platinum-thermometer wire was 

 thus placed under identical conditions with the wire under 

 examination. In all the following Tables the temperatures 

 in the columns headed pt are platinum temperatures taken 

 with the standard platinum wire. 



§ 7. We proceed in the next place to describe some of the 

 results obtained from the comparison of resistances effected as 

 above. 



I. Platinum. 



The samples of platinum wire were prepared for us by 

 Mr. George Matthey with great care. Two wires were chiefly 

 employed, one having a diameter of about 10 mils, called 

 the i£ thick " wire, and one having a diameter of about 3 mils, 

 called the " thin *" wire. The thick wire was the one taken 

 as the standard thermometer. 



Thick Platinum wire. 

 Length =150 centim. Mean diameter =0*025945 centim. 

 Mean cross-sectional area =0*00052868 sq. centim. 



Let R*= resistance of wire corrected for connexions and 

 bridge temperature taken at any temperature f C, and 

 measured in true ohms ; 

 Phil Mag. S. 5. Vol. 36. No. 220. Sept. 1893. U 



Log 10 £ : 



