52 



tares, and then with different degrees of 

 security. To make this point clear, 

 compare two machines of different types 

 such, for instance, as the Westinghouse 

 or the Lowrie-Parker on the one hand, 

 and the Kennedy or the author's machine 

 on the other hand. In the two first- 

 named machines the coils are simply laid 

 on to the surface of the armature-core 

 on one side, whilst in the other two 

 machines they completely surround the 

 core. The coefficient of self-induction 

 may, in the latter case, be roughly taken 

 as four times that of the former case, 

 other conditions being approximately the 

 same ; and it may therefore be expected 

 that parallel working will be safer and 

 more certain when the armature con- 

 ductor surrounds the core, than when it 

 lies only on one side of the core, though 

 in this case it will still be possible. The 

 experience gained in America with West- 

 inghouse machines is in this respect very 

 instructive. It has been found that the 

 experiment of parallel coupling always 

 succeeds when the machines are loaded 



