Comparisons of Mercury and Platinum Thermometers. 1 3 



Table II. — Correction to reduce the Bridge-wire Reading 

 to the Mean Box Unit at 20° C. 



Bridge-wire 



Calibration 



Bridge-wire 



Calibration 



Reading. 



Correction. 



Beading. 



Correction. 



+ 05 



-•004 



-0-5 



+■004 



+ 1-0 



-•007 



-10 



+ •007 



+ 15 



-•011 



-1-5 



+ -on 



+2-0 



-•015 



-20 



-4 014 



+2-5 



-•018 



-2-5 



+ •018 



+30 



-•022 



-3-0 



+•022 



+3o 



-•025 



-3-5 



+•026 



+4-0 



-•028 



-4-0 



+ •029 



+4-5 



-•031 



-45 



+•033 



+5-0 



-•034 



-5-0 



+ 036 



+55 



- -037 



-5-5 



+ •040 



+60 



-•041 



-60 



+ •043 



+6-5 



-044 



-6-5 



+ •046 



+7-0 



-•047 



-7-0 



+ •049 



+ 7-5 



-•051 



-75 



+ •051 



+8-0 



-054 



-8-0 



+ 054 



+8-5 



-■057 



-8-5 



+ •057 



+9-0 



-060 



-90 



+ 060 



+9-5 



-•063 



-9-5 



+ 064 



The bridge- wire vernier could be easily read to the nearest 

 yJ- - ram., and this with considerable confidence. Obser- 

 vations of any constant resistance taken with the same com- 

 binations of coils and bridge-wire rarely differed by as much 

 as 000002 ohm; when several different combinations were 

 used to measure a constant resistance, as will be seen on 

 referring to the measurement of E (the resistance of the 

 platinum thermometer at 0° 0.), the several measurements 

 rarely differed by as much as 0*00005 ohm. This is a severe 

 test of the accuracy of the standardization. Even these 

 slight discrepancies are not entirely due to errors in cali- 

 bration, but could, we believe, be still further diminished by 

 stirring the water in the outer tank surrounding the oil-bath 

 in which the coils are immersed. Such great accuracy in 

 the standardization of a resistance-box is, of course, easily 

 obtained so long as we are only concerned with the relative 

 values of the coils, which is all that we care about in the 

 measurement of platinum temperature. But if the absolute 

 values of the coils were required, a similar degree of accuracy 

 would involve a vast amount of most painstaking work, 

 mainly on account of the unscientific construction of most 

 standard coils, which renders the accurate determination of 

 their temperature impossible, and of their having undergone 

 considerable changes in resistance with time. 



