11)0 Prof. J. Perry on Mr. Blakesley's Method of 



a cycle. ^ If I had time, I could show that in alternating-current 

 calculations there are other very important uses of the fact 



r*r+t 

 that I xx.dt — 0, if x is any periodic function of the 

 time. 



Added May 23rd, 1891. 



In the discussion of this paper it was obvious that I had 

 not at sufficient length made known what I meant by magnetic 

 "_ leakage." _ It was owing to this, no doubt, that my introduc- 

 tion of the idea of the importance of leakage was looked upon 

 as academic merely. Again, my use of the symbols L, M, and 

 1/ did not seem to be understood, nor what they had to do 

 with a transformer. It is therefore necessary that I should 

 say more fully, but not more definitely than in the paper, 

 that as M is always less than VLL' owing to magnetic leak- 

 age, I define leakage as the value of y where 



M=(l- y ) VLL'. 



Hence, if L = P 2 ^^ or P 2 m, say, then L'=S 2 7tz and- 



M = PSm(l-?/). 



Again, the method of treatment to which I object is to 

 state the equations as 



V = BO + P§, d) 



= R'C' + S^| (2) 



I affirm that the induction I of equation (1) is a. very 

 different thing from the I of equation (2). As the old Max- 

 well method of writing the equations does not seem to be 

 understood, I wish to make it clear that if I did use the 

 induction I would use I in equation (1) and I s in equation (2), 

 where 



I p = PmC + Sm(l-y)C, 



I 8 = Pm(l-y)C+SroC. 



This is assuming that the number of ampere-turns which 

 produces the effective induction through the primary is 

 PC + S0'(1 — y) ; and the number of ampere-turns which 

 produces the effective induction through the secondary is 



PO(l-y) + SC. 



In fact, as I stated clearly when reading the paper, y is the 



