214 



Mr. E. H. Griffiths. 



[Jan. 17, 



by the expression 2^ the supply of heat during an experiment due to 

 the above sources. 



A source of heat was the supply due to the stirring. Let thermal 

 units per second due to this cause be Q s , and let the stirring be main- 

 tained for a time t t , then total mechanical supply = Q,k. I have 

 given full particulars of the manner in which the values of Q^ for 

 different values of O l were ascertained (Q 5 increased as 6\ diminished, 

 owing to the increase in viscosity of the oil). A study of the results 

 of the stirring experiments led to the following conclusion : " The 

 values of Q* given in the tables are certainly correct to better than 

 1 in 50 at temperatures 30 and 40, for in no case do individual 

 experiments (when reduced to the same rate) differ by 1 in 100. 

 Now an error of 1 in .50 in Q 5 would cause an error of only about 

 1 in 5000 in L. I am less certain, however, about the values of Q s at 

 20 and 50 C." 



The last, and by far the greatest source of heat, is the work done 

 by the electric current. Let Q e be the units per second due to this 

 cause, and let the current be maintained for a time t e . 



JSTo\v 



(1), 



where e is the potential difference of a Clark cell ab 15 C., n the 

 number of cells, and R : the resistance of the coil at temperature 0,, 

 ufter the corrections for the rise in temperature of the wire due to the 

 passage of the current, &c., have been made.* 

 Let m be the mass in vacuo of water evaporated. 



Then 



(2). 



In order to give an idea of the comparative importance of the terms 

 in equation (2), I give the mean value of each term resulting from the 

 experiments. 



Table I. 



* The value of J used in these reductions was 4'199 x 10?. If there is some 

 error in the electrical standards, &c., by which I determined J, this is still the 

 right value to use here, for the standards are the same, and the methods practically 



