376 Dr. Norman Campell : Experiments on 



the oscillations, but the increase found here is considerably 

 less than that found by Prof. Jones *. 



The agreement in the value of c is not very close in most 

 cases, but the variation of c found by the oscillation method 

 is always in the right direction ; as the uncoupled secondary 

 inductance increases, c increases also. The measurement of 

 C 2 is not capable of much accuracy, and the differences 

 between the various figures can hardly be said to exceed 

 experimental error. The values shown include the 11*8 mmf. 

 added to the circuit for measuring purposes. It is perhaps 

 rather surprising that the difference between C 2 for these 

 armatures and that for the armature previously used is not 

 greater ; for in the inductor machine the windings are 

 not surrounded on all sides by earthed conductors. On the 

 other hand, it is to be expected that the addition of un- 

 coupled turns should not increase C 2 materially, for the 

 housing in which those turns were contained was insulated 

 from earth, and the capacity to earth of the windings in it 

 should be less than that of the housing, which was estimated 

 to be 4 or 5 mmf. In what follows it will be assumed that 

 C 2 may be taken the same for all the combinations. 



In general, then, it may be said again that the oscillation 

 method gives results which are consistent among themselves, 

 and agree, as well as can be expected, with the values obtained 

 by other methods. The theory which is involved in the 

 measurements is reliable so far as the determination of 

 the constants is concerned. 



The Damping Coefficients. 



15. Fig. 8. corresponding in all particulars to fig. 5 of 

 the first part, shows p plotted against r. The full line refers 

 to the oscillations in the secondary circuit; the dotted line to 

 oscillations in the primary circuit. The suspicion previously 

 noted that the damping in the two circuits is not strictly the 

 same is thus confirmed. The lines refer equally to all 

 the combinations examined ; it could not be established 

 that, at the same frequency, fi was different for different 

 combinations of the armatures. 



It was found, again, that, for the same frequency, yj is 



* For the larger values of c the estimate of Lj by the oscillation 

 method may be too low. For that method involves the substitution of 

 (l'l) for (1), which is only legitimate when r is very large — a condition 

 which is not fulfilled when c is as great as 0*4. An error of 10 per cent, 

 might arise iu this way. 



