334 PROCEEDINGS OF THE A:\LERICAN ACADEMY 



or less sharpened crests and troughs. lu no case, however (except 

 for the lowest possible velocities), would the wave approximate to the 

 rectangular form assigned to it by the theory of instantaneous action ; 

 nor, since its form differs at different distances, can it conform to any 

 theory of instantaneous propagation. 



The table also enables us to estimate the length of time that elapses 

 before the wave can reach its mean height within the secondary when 

 placed as near as possible (say two inches) to the centre of the primary. 

 The time is about .058 seconds. This large retardation accounts for 

 certain discrepancies found in earlier determinations of the zero points 

 by the use of insufficiently low velocities. 



The table shows, moreover, the actual velocity of propagation. 

 We find, e.g., that the time occupied by a point at fourteen inches in 

 reaching fifty per cent of its maximum magnetization is about .035 

 seconds greater than for a point at two inches, giving a velocity of 

 about twenty-eight and a half feet per second, which agrees as closely 

 as could be expected with early results. For other percentages of 

 magnetization we should of course obtain different results ; but this 

 particular result is of special significance in that it represents, in a 

 measure, the rate of propagation of the body of the pulse. It may 

 conveniently be termed the. principal velocity of the magnetic wave. 



The last method of measurement is likewise adapted to the use of 

 a dynamometer (the phase-adjustment being a hiuderance rather than 

 a help) ; and is really an independent way of determining the velocity 

 of propagation. 



There is still another method of arriving at this velocity. If in 

 the Table for Phase-Retardation, we subtract the phase for two inches 

 from that for fourteen, and divide by the whole niuuber of degrees 

 described by the commutator in a second, we shall have the length of 

 time occupied by tlie phase of mean height of the wave in travelling 

 over the same twelve inches of the bar. The lowest possible velocity 

 of revolution must be chosen in order to approximate to our former 

 conditions. 



Taking two revolutions per second, we find a difference of 24°, 

 corresponding to one thirtieth of a second ; whence the velocity is 

 thirty feet per second, nearly as before. 



The following is a series of velocities computed in the same way, 

 each over six inches of the bar, beginning at 0, 2, 4, &c., inches : — 



