BURGESS AND KELLBERG : CRITICAL RANGES OF IRON 437 



ture and waiting for equilibrium to be established. It soon 

 became evident that, although great sensitiveness and accuracy 

 could be obtained, nevertheless it would take an infinite time 

 to plot an entire resistance-temperature curve satisfactorily. 

 The method adopted in the final series, and which satisfies all the 

 above requirements, depends on the use of the cooling curve 

 apparatus described in Bureau Reprint 213 (loc. cit.), together 

 with a very sensitive, quickly manipulated and accurate Wheat- 

 stone bridge, by means of which the resistances of an iron wire 

 and one of platinum wound on the same support and enclosed in 

 vacuo in quartz glass may be exactly compared every few seconds 

 by the intermediary of a drmn-chronograph recording the times 

 at which the resistances are measured. In other words, we have 

 used an electrical resistance cooling curve outfit of the highest 

 attainable accuracy and sensitiveness. The temperatures are 



Fig. 1. Construction of platinum and iron thermometers. 



given in tei'ms of the resistance of the platinum wire which 

 serves as a thermometer integrating the temperature of the iron 

 wire exactly. 



The construction of the combined platinum and iron thermometers 

 is shown in figure 1. The platinum and iron wires of 0.2 mm. and 0.24 

 mm. diameter, respectively, are wound on thin walled, unglazed hard 

 porcelain insulators 6 cm. in length and separated by a strip of mica. 

 The thermometers are of the compensated three-lead type with one 

 common lead and a common battery lead, all four leads being of plati- 

 num and provided with porcelain insulators. After winding the coils 

 and before sealing off, the quartz containing tube was evacuated and, 

 with the coils, heated to a bright red, thus partly annealing the wires 

 and expelling gases. After sealing, the thermometers were again 

 annealed to about 1000°C. in the electric furnace. Several platinum 

 and iron thermometers were made in this way, the values of the resist- 

 ances at 0°C. usually being about 1.5 ohms for the platinum and 1 

 ohm for the. iron. The length of the thermometer was about one-tenth 

 that of the specially wound platinum resistance furnace used in takirg 

 the resistance observations. The iron was from samples of the purest 



