326 KENNELLY-NABESHIMA— ESTABLISHING A 



upon the circuit. A permanent or steady electromagnetic state can- 

 not therefore contain transients to any appreciable extent; although 

 transients will probably have been involved in the production of 

 that state. It is true that transient phenomena persist with pro- 

 gressive attentuation for an indefinitely long period, and, in that 

 sense, never completely disappear. From an engineering point of 

 view, however, transients diminish into practical negligibility in a 

 period of time that is ordinarily measured in milliseconds ; so that 

 they may be regarded as having vanished, when the steady state is 

 accepted as having been attained. 



For the purposes of discussion, transients are subdivided as 

 shown in Table I. 



Class A includes transients of all kinds, mechanical and elec- 

 trical. Thus, a kick of the pen of a recording ammeter pointer 

 might represent a mechanical transient due to an electrical transient. 



Prominent among transients are electromagnetic-wave transient 

 disturbances, or zvave transients. 



An important class of wave transients are those which accom- 

 pany the establishment of an alternating current over a line. These 

 may be classed collectively as initiating a.-c. transients. This is the 

 principal class of transients discussed in this paper. 



Initiating alternating-current wave transients occur, in general, 

 whenever a switch is closed at the generating end of an a.-c, line. 

 If the switch could be closed in such a manner as to produce no 

 electric " splash," the outgoing wave transient would be the " regu- 

 lar " transient, which accompanies the regular formation of the 

 final alternating-current state. 



Infinite Distortionless Line Closed at Zero E.m.f. {No Tran- 

 sients). — If we assume, for simplicity, a perfectly regulated sine- 

 wave generator, of negligible internal impedance, connected as in 

 Fig. I to an indefinitely long uniform distortionless line,^ then if the 

 switch S is closed at some instant when the alternating e.m.f. is 

 passing through zero, the initial outgoing waves of e.m.f. and cur- 



2 A distortionless line was originally defined by Heaviside as a line in 

 which l/r = c/g. In such a line, the surge impedance s^ must have zero 

 slope at all frequencies, or becomes reactanceless. See Bibliography (i), 

 page 126, Vol. II.; also Bibliography (30), page 154. 



