208 Prof. Ayrton and Dr. Sumpner on Alternate 



then, taking the difference, we get the mean value of 



K-^3)(>l-^)> 



which, as before, is equal to r times the mean watts given 



Fi£. 6. 



to ab. 



Mr. Blakesley has also pointed out, in his paper published 

 in the Phil. Mag. for April of this year, that the current 

 measurements may be varied and the apparatus arranged as 

 seen in fig. 7. The two mea- 

 surements now give respectively Fig. 7. 

 the mean values of 



and 



c 2 a 3 



therefore the difference gives the 

 mean value of 



a 3 (« 3 — a 2 ), 



that is, the mean value of 



which is equal to - times the mean watts given to ab. 



While, however, our second method of using the electro- 

 meter (fig. 6) gives the answer with the same accuracy as 

 the first method (fig. 4), Mr. Blakesley's second method of 

 joining up the dynamometers in figure 7 introduces self- 

 induction into a circuit which ought to be entirely non- 

 inductive, and so it does not give the answer with the same 

 accuracy as his first method (fig. 5). 



IV. 



The modification of our electrometer method suggested to 

 one of us by Mr. L. Atkinson, while he was a pupil at the 

 Pinsbury Technical College, and afterwards carried out by 



