In this substation arrangement four coils of wire—­1, 2, 3, and 4—­are so arranged as to be connected in the circuit of the line, two in series and two in multiple.  The current flowing from the battery at the central office, after passing through the transmitter, divides between the two paths containing, respectively, the coils 1 and 3 and the coils 2 and 4.  The receiver is connected between the junction of the coils 2 and 4 and that of 1 and 3.  The resistances of the coils are so chosen that the drop of potential through the coil 2 will be equal to that through the coil 1, and likewise that through the coil 4 will be equal to that through the coil 3.  As a result, the receiver will be connected between two points of equal potential, and no direct current will flow through it.  How, then, do voice currents find their way through the receiver, as they evidently must, if the circuit is to fulfill any useful function?  The coils 2 and 3 are made to have high impedance, while 1 and 4 are so wound as to be non-inductive and, therefore, offer no impedance save that of their ohmic resistance.  What is true, therefore, of direct currents does not hold for voice currents, and as a result, the voice currents, instead of taking the divided path which the direct currents pursued, are debarred from the coils 2 and 3 by their high impedance and thus pass through the non-inductive coil 1, the receiver, and the non-inductive coil 4.

This circuit employs a Wheatstone-bridge arrangement, adjusted to a state of balance with respect to direct currents, such currents being excluded from the receiver, not because the receiver circuit is in any sense opaque to such direct currents, but because there is no difference of potential between the terminals of the receiver circuit, and, therefore, no tendency for current to flow through the receiver.  In order that fluctuating currents may not, for the same reason, be caused to pass by, rather than through, the receiver circuit, the diametrically-opposed arms of the Wheatstone bridge are made to possess, in large degree, self-induction, thereby giving these two arms a high impedance to fluctuating currents.  The conditions which exist for direct currents do not, therefore, exist for fluctuating currents, and it is this distinction which allows alternating currents to pass through the receiver and at the same time excludes direct currents therefrom.

In practice, the coils 1, 2, 3, and 4 of the Dean substation circuit are wound on the same core, but coils 1 and 4—­the non-inductive ones—­are wound by doubling the wire back on itself so as to neutralize their self-induction.