24 Dr. W. S. Day on a Comparison of 



meters of each kind, there were nine possible comparisons. 

 These nine were all made, and finally an additional one on 

 26th April, 1897, between Tonnelot 11801 and Baudin 

 6163. When this was made, the Baudin thermometer had 

 not yet recovered its zero from a depression caused by frequent 

 use up to within six days of the time of the comparison. Its 

 correction-curve therefore differed from the others, but in 

 the way one would expect on account of this depression. 

 It was, however, rejected. The corrections given in Table VIII. 

 were finally used to calculate the changes to be made in 

 Rowland's value of the mechanical equivalent to reduce it to 

 the Paris hydrogen scale. 



In order to get a further check on the comparisons, the 

 three Baudin thermometers were placed in the tank at the 

 same time, and compared together. Without going into 

 the details here, it may be said that these comparisons 

 verified the results, so far as it was possible, of the other 

 comparisons. 



8. Calculation of the effect on Rowland's Value 

 of the Mechanical Equivalent of Heat. 



In order to estimate the effect of these corrections to the 

 Baudin thermometers on the value of the mechanical equi- 

 valent of heat, the experiments made by Rowland with each 

 thermometer were considered separately. Rowland has a 

 table on p. 192 of his paper, giving the mechanical equivalent 

 in kilogramme-metres at Baltimore, each value calculated from 

 a rise of 10° in temperature, as deduced from each of his 

 experiments. From this table the mean value of the 

 mechanical equivalent at each temperature was calculated, as 

 obtained by each thermometer separately. Rowland's value 

 was deduced from an equation of the form 



I- W 



where W is a certain number of ergs, found by experiment, 

 required to raise the temperature of 1 gram of water 10° C. 

 J at 12°, for example, was calculated from the value of W in 

 going from 7° to 17° C, That is 



J - W 



17°- 7° 



If On and C 7 are the corrections found for the readings of 

 the Baudin thermometer to reduce them to the hydrop-en 

 scale, these corrections applied to the value of J 12 will 



