194 DR CARGILL G. KNOTT ON THE 



where t is, as before, the absolute temperature. This expression will likewise 

 be found to suit the numbers given in the last column of table B. 



There is no very obvious mode for obtaining a similar formula for the thick 

 nickel. 



It may be remarked that this mode of representing the temperature rela- 

 tions of resistance by a power of the absolute temperature — a power which may 

 be fractional — includes as a special case the well-known statement that, for 

 pure metals, the resistance is directly as the absolute temperature. For small 

 ranges of temperature the equation 



R = CT* 

 may be easily thrown into the approximate form 



where T is absolute temperature, t centigrade, and the other quantities are 

 constants. In this case a is to a first approximation equal to k times the 

 reciprocal of 274. 



We now pass to the discussion of the second series of experiments. In 

 these the temperature was raised to a fairly bright red heat by means of a 

 charcoal furnace. The four stout copper rods, with the attached wires which 

 were to be tested, dipped into a porcelain vessel through suitable holes in the 

 lid. The vessel itself stood inside a small charcoal furnace, and was heated by 

 red charcoal dropped in around it. After reaching its highest temperature the 

 charcoal and wires gradually cooled; and during this cooling the resistances of 

 the two wires were measured in rapid alternation. 



To obtain what might be regarded as simultaneous values of the resistances, 

 means of successive pairs of readings for the one metal were interpolated. 

 In every case the one wire was the same piece of platinum, whose indications 

 served the purpose of a thermometer. In terms of its resistances, the resist- 

 ances of the other wires could be expressed, graphically or otherwise. Many 

 experiments were made with each kind of wire, and a vast number of observa- 

 tions accumulated. These I have not thought necessary to reproduce in the 

 form in which they were obtained. The five curves of Plate XII. (II.), however, 

 which tell their own tale clearly enough, are drawn from the observations, 

 conveniently reduced, of the five best experiments. The reductions were the 

 same as those made in the earlier series of experiments ; that is, the resistance 

 at 0° C. for each metal was reduced to 100, and the other resistances changed 

 proportionally. 



Although the numbers themselves are not reproduced, their essence is 

 given in table E, which is really a comparative table of the resistances of 

 certain wires at various temperatures from 0° C. to a fairly bright red heat. 

 The series of platinum resistances, as shown in the first column, rises from 100 



