pull electrons across the vacuum towards itself.

[Illustration: Fig 8]

What happens then is something like this: Electrons are freed at the
filament; the plate and the grid both call them and they start off in a
rush. Some of them are stopped by the wires of the grid but most of them
go on by to the plate. The grid is mostly open space, you know, and the
electrons move as fast as lightning. They are going too fast in the
general direction of the grid to stop and look for its few and small
wires.

When the grid is positive the grid helps the plate to call electrons
away from the filament. Making the grid positive, therefore, increases
the stream of electrons _between filament and plate_; that is,
increases the current in the plate circuit.

We could get the same effect so far as concerns an increased plate
current by using more batteries in series in the plate circuit so as to
pull harder. But the grid is so close to the filament that a single
battery cell in the grid circuit can call electrons so strongly that it
would take several extra battery cells in the plate circuit to produce
the same effect.

[Illustration: Fig 9]

If we reverse the grid battery, as in Fig. 9, so as to make the grid
negative, then, instead of attracting electrons the grid repels them.
Nowhere near as many electrons will stream across to the plate when the
grid says, "No, go back." The grid is in a strategic position and what
it says has a great effect.

When there is no battery connected to the grid it has no possibility of
influencing the electron stream which the plate is attracting to itself.
We say, then, that the grid is uncharged or is at "zero potential,"
meaning that it is zero or nothing in possibility. But when the grid is
charged, no matter how little, it makes a change in the plate current.
When the grid says "Come on," even though very softly, it has as much
effect on the electrons as if the plate shouted at them, and a lot of
extra electrons rush for the plate. But when the grid whispers "Go
back," many electrons which would otherwise have gone streaking off to
the plate crowd back toward the filament. That's how the audion works.
There is an electron stream and a wonderfully sensitive way of
controlling the stream.

LETTER 7

HOW TO MEASURE AN ELECTRON STREAM

(This letter may be omitted on the first reading.)

DEAR YOUTH:

If we are to talk about the audion and how its grid controls the current
in the p