DICKINSON AND MUELLER: RESISTANCE THERMOMETER SCALE 179 



The agreement of the transfer methods is complete. The differ-' 

 ence of 0?17 between the transfer and the direct determination 

 at the boiling point of sulfur may be due to (1) difference in the 

 sulfur baths used, (2) an unusual accumulation of accidental errors, 

 (3) a systematic error affecting all transfer measurements. 



TABLE III 



1. The two sulfur baths were compared side by side by means of 

 thermometer "R." The Day and Sosman sulfur bath was oper- 

 ated by Dr. Sosman. Exploration with the resistance thermome- 

 ter showed no temperature differences as large as 0?05 within the 

 aluminium shield which had been used with the gas thermometer. 

 The temperature in this shield was found to be 0?03 to 0?04 lower 

 than in the Bureau of Standards gas-heated apparatus, a differ- 

 ence which is hardly significant. 



2. The accidental errors of either the gas thermometer or the 

 resistance thermometer are not as large as the difference found. 

 The three values found by transfer with the resistance thermome- 

 ter as well as the two values found by transfer with the thermo- 

 couples, give the same value of 8 (1.484) for the resistance ther- 

 mometer "R." 



3. The above agreement, together with the fact that the value 

 of 8 computed from the value 444.55 for the sulfur boiling point is 

 1.494, may be taken as indicating a possible systematic error 

 affecting all the transfer values. On account of the absence of a 

 direct transfer with the thermocouples, however, this cannot be 

 considered as proven. The indirect determination of the sulfur 

 boiling point would appear to have less weight than the direct 

 determination, since the sulfur bath used for the latter was found 

 to define the same termperature as that obtained in the conven- 

 tional form of sulfur boiling apparatus. 



