216 



Mr. E. H. Griffiths. 



[Jan. 17, 



it to boil off rapidly by sufficiently diminishing the pressure. The 

 object of this arrangement was to ensure, as far as possible, that the 

 temperature of the evaporating water was that of the flask (0^). 

 Many precautions too numerous to be here referred to had to be 

 taken before this method could be worked successfully, but when 

 once the experimental difficulties had been overcome, the agreement 

 between different experiments left little to be desired, and this 

 continued to be the case although the conditions of the experiments 

 were considerably altered. Different masses of water and different 

 rates of evaporation gave closely concordant values of L. 



Table of Results. 



(An improvement was made in the apparatus after No. 5, and I 

 attach greater weight to all succeeding experiments.) 



Table III. 



I regret that the experiments are comparatively few in number, 

 but they occupied so much time that I was unable to make further 

 repetitions. 



I do not, however, consider that such repetitions would have 

 greatly strengthened the evidence, for the mean probable error of 

 even a small group of experimental results like the above is less than 

 the probable error of some of the constants involved in the reduction 

 of those results. 



For the reason previously stated I attach small value to Nos. I 

 toV. 



