Like some other problems in telephony, this one seems simpler at first sight than it proves to be after more exhaustive study.  It is possible for any amateur to produce at once a repeating device which will relay telephone circuits in one direction.  It is required, however, that in practice the voice currents be relayed in both directions, and further, that the relay actually augment the energy which passes through it; that is, that it will send on a more powerful current than it receives.  Most of the devices so far invented fail in one or the other of these particulars.  Several ways have been shown of assembling repeating devices which will talk both ways, but not many assembling repeating devices have been shown that will talk both ways and augment in both directions.

[Illustration:  Fig. 36.  Shreeve Repeater and Circuit]

Practical repeaters have been produced, however, and at least one type is in daily successful use.  It is not conclusively shown even of it that it augments in the same degree all of the voice waves which reach it, or even that it augments some of them at all.  Its action, however, is distinctly an improvement in commercial practice.  It is the invention of Mr. Herbert E. Shreeve and is shown in Fig. 39.  Primarily it consists of a telephone receiver, of a particular type devised by Gundlach, associated with a granular carbon transmitter button.  It is further associated with an arrangement of induction coils or repeating coils, the object of these being to accomplish the two-way action, that is, of speaking in both directions and of preventing reactive interference between the receiving and transmitting elements.  The battery 1 energizes the field of the receiving element; the received line current varies that field; the resulting motion varies the resistance of the carbon button and transforms current from battery 2 into a new alternating line current.

By reactive interference is meant action whereby the transmitter element, in emitting a wave, affects its own controlling receiver element, thus setting up an action similar to that which occurs when the receiver of a telephone is held close to its transmitter and humming or singing ensues.  No repeater is successful unless it is free from this reactive interference.

[Illustration:  Fig. 37.  Mercury-Arc Telephone Relay]

Enough has been accomplished by practical tests of the Shreeve device and others like it to show that the search for a method of relaying telephone voice currents is not looking for a pot of gold at the end of the rainbow.  The most remarkable truth established by the success of repeaters of the Shreeve type is that a device embodying so large inertia of moving parts can succeed at all.  If this mean anything, it is that a device in which inertia is absolutely eliminated might do very much better.  Many of the methods already proposed by inventors attack the problem in this way and one of the most recent and most promising ways is that of Mr. J.B.  Taylor, the circuit of whose telephone-relay patent is shown in Fig. 37.  In it, 1 is an electromagnet energized by voice currents; its varying field varies an arc between the electrodes 2-2 and 3 in a vacuum tube.  These fluctuations are transformed into line currents by the coil 4.