as the "regenerative circuit"
because the tube keeps on generating an alternating current. The little
coil which is used to feed back into the grid circuit some of the
effects from the plate circuit is sometimes called a "tickler" coil.

It is not necessary to use a grid condenser in a feed-back circuit but
it is perhaps the usual method of detecting where the regenerative
circuit is used. The whole value of the regenerative circuit so far as
receiving is concerned is in the high efficiency which it permits. One
tube can do the work of two.

We can get just as loud signals by using another tube instead of making
one do all the work. In the regenerative circuit the tube is performing
two jobs at once. It is detecting but it is also amplifying.[9] By
"amplifying" we mean making an e. m. f. larger than it is without
changing the shape of its picture, that is without changing its "wave
form."

To show just what we mean by amplifying we must look again at the audion
and see how it acts. You know that a change in the grid potential makes
a change in the plate current. Let us arrange an audion in a circuit
which will tell us a little more of what happens. Fig. 93 shows the
circuit.

This circuit is the same as we used to find the audion characteristic
except that there is a clip for varying the number of batteries in the
plate circuit and a voltmeter for measuring their e. m. f. We start with
the grid at zero potential and the usual number of batteries in the
plate circuit. The voltmeter tells us the e. m. f. We read the ammeter
in the plate circuit and note what that current is. Then we shift the
slider in the grid circuit so as to give the grid a small potential. The
current in the plate circuit changes. We can now move the clip on the
B-batteries so as to bring the current in this circuit back to its
original value. Of course, if we make the grid positive we move the clip
so as to use fewer cells of the B-battery. On the other hand if we make
the grid negative we shall need more e. m. f. in the plate circuit. In
either case we shall find that we need to make a very much larger change
in the voltage of the plate circuit than we have made in the voltage of
the grid circuit.

[Illustration: Fig 93]

Usually we perform the experiment a little differently so as to get more
accurate results. We read the voltmeter in the plate circuit and the
ammeter in that circuit. Then we change the number of batteries which we
are usin