1893.] the Mechanical Equivalent of Heat, $c. 



13 



#!, Temperature. — At the time of writing (December, 1892) our 

 thermometry is based on measurements made by platinum thermo- 

 meters. We propose to make, at an early date, a direct comparison 

 of our standard thermometer with the air thermometer by means of 

 the apparatus described by Mr. Callendar,* who has been so kind as 

 to promise his assistance. 



In a previous paperf the details of a careful comparison of the 

 platinum and the air thermometer have been given. It was then shown 

 that the platinum-air difference curve 6(£|1Q0| 2 — £|100) gave values 

 of t at all temperatures from 0° to 100° C. accurate within 0*01° C. 

 of the real value of t, and that discrepancies appeared to have an 

 experimental origin. The experience of two years' work with 

 platinum thermometers has but increased our confidence in them.| 

 Should a direct comparison with the air thermometer modify our 

 determinations of differences of temperature (and it is only differ- 

 ences which are important to us), our results will be modified 

 accordingly. This will not, however, necessitate any repetition of 

 the experimental work, as the corrections involved will be of a 

 numerical order only. - 



A mercury thermometer by Hicks (labelled E„j§) was standardised 

 by direct comparison with different platinum thermometers, the 

 observations being taken under conditions similar to those prevalent 

 during our " J experiments," and the temperatures corresponding to 

 the readings given in column 1, Table XVII, were thus ascertained. 



We found it necessary to calibrate the mercury thermometer not 

 only for irregularities in the bore, but for certain strictly recurrent 

 changes in " lag," which we have found to be invariably associated 

 with a rising mercury thermometer. We are unable to briefly describe 

 the method adopted, and we confine ourselves to stating that it was 

 based on observations of time. The results of this investigation, 

 which extended over several months, prove that a calibration of the 

 ordinary kind is insufficient if accurate observations have to be ob- 

 tained with a rising mercury thermometer. As an illustration we 

 give the following table, which shows the difference in the value of 

 certain ranges on thermometer E w when steady and when rising at 

 the normal rate of our experiments. 



* ' Eoy. Soc. Proc.,' January, 1891. 

 t See 'Phil. Trans.,' 1891, A, p. 155. 

 t ' Phil. Mag.,' December, 1892. 



§ The stem of this thermometer was graduated in mm. : about 40 mm. = 1° C. 



