36 



Mr. E. H. Griffiths and Miss Marshall on the 



Remarks on Experiments at 20° C. 



Exp. XXXV. — This should be regarded as a preliminary- 

 experiment only. We had no idea of the pressure required to 

 produce equilibrium, and had great difficulties in sufficiently 

 reducing the pressure. " 1 was extremely irregular ; the 

 galvanometer -swing was often off the scale for a considerable 

 period." 



Exp. XXXVI. — " This was a very fair experiment ; the 

 thermal balance well under control." 



We regret that we were unable to perform more experiments 

 at 20° C. It was found necessary, however, to work the 

 mercury-pump almost continuously, and this could not be 

 done without further assistance*. As such assistance was 

 bat rarely available, and as our time was drawing to a close, 

 we were compelled to content ourselves with the above 

 experiments. 



Conclusion. 



When O = 19-947, L = 103'82. 



21. Discussion of the Results. 



The values of L over the range 20° to 50° appear to be 

 practically a linear function of the temperature. 

 If we assume 



L = 10705 -0-1581 0, .... (a) 



we obtain a very close approximation to the experimental 

 values, as shown by the following Table : — 



Table IX. 



Tempei'ature. 



Nitrogen-scale. 



L (expressed in terms of a thermal unit 

 "at 15° C"). 



By formula (a). 



Experimental results. 



50014 

 40-045 

 30-066 

 19-947 



99-14 

 100-72 

 102-30 

 103-90 



99- 14 ±-034 

 100-71 ±-057 

 102-30 ±-076 

 103-82 



It is evident, therefore, that the curvature over the above 

 temperature-range must be very slight. 



The value of L cannot continue to be a linear function of 

 at very high temperatures, either in the case of benzene or 



* We return our sincere thanks to Mr. C. T. Heycock, F.R.S., for his 

 help during these experiments at 20° O. 



