Resistance of Metals and Alloys. 273 



we know the changes in dimensions which the material 

 undergoes when heated or cooled. The true volume-specific 

 resistance of the material at any temperature is the resistance 

 in electromagnetic measure between opposed faces of one 

 cubic centimetre of the material taken at that temperature. 

 Pending, however, the determinations of the temperature- 

 change in dimensions of the metals and alloys employed, we 

 omit the correction (no doubt small) due to the change of 

 dimensions, and state our results in terms of the electrical 

 resistance in electromagnetic measure of one cubic centimetre 

 of the material measured at 15° C, the said electrical resist- 

 ance being determined at the various temperatures to which 

 the material is raised or lowered. In the next place we have 

 to consider the question of temperature-measurement. As 

 the lowest temperatures reached by us in these experiments 

 were temperatures at which air liquefies under ordinary pres- 

 sures, the practical employment of the air- thermometer was 

 surrounded with many difficulties . Ultimately it may be 

 necessary to define the temperatures attained in terms of the 

 absolute thermodynamic scale. But even the employment of 

 the constant reduced-pressure hydrogen thermometer-scale, 

 which may be taken as giving us the nearest practicable 

 approximation to the true thermodynamic scale, is a matter 

 in which extreme care w T ill have to be employed in obtaining 

 and interpreting results. We have been led, however, by 

 one part of our results to see that, for our purpose here, by 

 far the most practical method of defining the temperatures 

 employed is to state them in terms of the variation of the 

 specific resistance of one standard metal. 



In the practical determination of temperature we select a 

 thermometric substance, it may be air, hydrogen, mercury- 

 in-glass, or some other material, such as a double metallic 

 junction of thermoelectric materials, and we select also a 

 thermometric quality, such as the change in pressure, change 

 in apparent volume, or thermoelectric difference, and define 

 our temperature by the temperature-change of the thermo- 

 metric quality of this thermometric material. There is, there- 

 fore, nothing to prevent us from selecting as a measure of 

 temperature the change in electrical resistance of a wire of a 

 standard material and defining temperature in terms of the 

 change. It will afterwards be possible to translate the tem- 

 peratures so measured into their equivalents on the absolute 

 thermodynamic or air-thermometer scales. 



Our experiments at low temperatures have confirmed us in 

 the view that the most convenient method of determining and 

 stating such temperatures is in terms of the change of specific 



