MEASUREMENTS AT TELEPHONIC FREQUENCIES. 121 



If we plot the square of the received current strength of Fig. 12, 

 as scalar ordinates, to L(d/R as abscissas, we obtain the resonance 

 curve of Fig. 14. It follows from the linear relations of Lw/R 

 along the straight line AB, in Fig. 13, that the ordinates of the reso- 

 nance curve in Fig. 14 will lie symmetrically at equally spaced ab- 

 scissas on each side of the maximum. If a condenser had been used 

 as a receiving-end load beyond the point L = o, the part of the reso- 

 nance curve, Fig. 14, on the left-hand of the origin, might have been 

 covered. 



Prevalence of Circular Vector Loci. 



It may be noted that in the cases of Figs. 2, 4, 5, 7, 9 and 12, cir- 

 cular graphs of vector potential loci have presented themselves, and 

 many measurements in the laboratory, at constant frequency, lead 

 to such circular graphs. 



If we consider any fixed network of conductors, we may assume 

 that all joints are electrically good, all leaks are steady, and all iron 

 cores removed from the inductances ; or, in other words, assume 

 that each and all of the elements of the network obey Ohm's law. 



Let us select arbitrarily any pair of terminals A B in the net- 

 work, at which a constant simple alternating e.m.f. is applied, in the 

 steady state, at a fixed frequency, and also select any other pair of 

 terminals CD in the network, at which the effects of the applied 

 e.m.f. are to be looked for. It will be shown, in the Appendix, that 

 if we introduce, between C and D, an impedance which varies along 

 a circular locus in the impedance plane, a straight line variation being 

 included as a particular case, then both the current and voltage will 

 be caused to vary along a circular locus at CD. Moreover, the cur- 

 rent and sending-end impedance at the terminals A B will be caused 

 to vary along a circular locus. Again, if a circularly varied impe- 

 dance is inserted between the impressed e.m.f. and the AB termi- 

 nals, the voltage and current at those terminals will be caused to 

 vary circularly, and also the voltage and current at CD. A further 

 development of this remarkable phenomenon will be found in the 

 Appendix. It means that a conducting network has a wonderful 

 propensity for reproducing impressed circular variations. A net- 

 work of n elements, in one of which a circular variation of inipe- 



