25 



and Ferranti, it is almost negligible ; in 

 machines containing iron in the arma- 

 ture, which is not encircled by the coils, 

 such as the Westinghouse and Lowrie- 

 Parker, it is still very small ; and even in 

 machines in which the coils are wound 

 round the armature-core, as in Kennedy's 

 machine and that designed by the author, 

 the coefficient, although appreciable, is 

 yet not so large as to cause a sensible 

 error being introduced by the assump- 

 tion that it is a constant. 



The method originally devised by Mr. 

 Joubert, for taking the self-induction of 

 an alternator into account, requires the 

 solution of a differential equation, but 

 the problem can be treated graphically 

 in a much more simple manner. Both 

 methods have this in common, that it is 

 assumed that the current and all elec- 

 tromotive forces are sine-functions of the 

 time. Strictly speaking, this is not cor- 

 rect, but the error is probably not very 

 great, and it is in a certain sense un- 

 avoidable; because if this assumption 

 were discarded, and each case treated on 



