LOW-FREQUENCY INDUCTION 581 



than with low-resistivity earth, but (except for very wide separations) 

 the percentage reduction secured by increasing the separation a given 

 amount is smaller. 



Figure 4 shows a schematic of a demonstration set-up to show some 

 of these efifects. It is, of course, impracticable to employ earth of 

 different resistivities in a demonstration of this kind, but the difference 

 between return current which concentrates near the line and that 

 which is more remote can be roughly indicated. In order to illustrate 

 the fact that when the currents are concentrated closely under the line, 

 the coupling falls off rather rapidly as the separation is increased, 

 returns immediately under the lines are used and the telephone line 

 is moved to change the separation. To illustrate that a wider dis- 

 tribution of current in the earth tends to increase the coupling and to 

 make it less affected by separation, returns on the floor are used for 

 both the power and telephone lines. It can be shown that: 



(a) The induced voltage increases when the connections are changed 

 from the upper to the lower returns. 



{b) When the upper returns are used, the percentage reduction in induced 

 voltage when the separation is increased, is greater than the 

 percentage reduction when the lower returns are used and 

 the separation is increased by the same amount, i.e., when 

 the telephone line is moved between the same positions of 

 minimum and maximum separation. 



If the earth is not homogeneous, that is, has strata of different re- 

 sistivities, the distribution of the earth current is distorted and varying 

 effects are noted. Where local irregularities e.xist, marked and some- 

 times erratic changes in coupling may occur within comparatively 

 short distances. An "effective" earth resistivity can usually be 

 determined by test even where the earth is stratified. 



Another important factor in determining the net coupling between 

 power and telephone circuits is the effect of grounded wires or other 

 linear grounded metallic structures along the inductive e.xposure. 

 Voltages are induced in such grounded metallic structures in the same 

 way as voltages are induced in telephone wires and these voltages 

 cause currents. The magnetic fields accompanying these currents 

 generally oppose those from the power wires and reduce the induction 

 in the telephone circuit. The effect of such currents in grounded 

 structures is generally spoken of as "shielding." 



The amount and phase of the current in a grounded conductor in a 

 given location and hence the shielding provided by it depend on the 

 impedance of the conductor with earth return. Hence the shielding is 



