le, an experimental bulb L is shown, which is
provided with a neck n on the top for the application of an external
tinfoil coating, which may be connected to a body of larger surface.
Such a lamp as illustrated in Fig. 25 may also be lighted by
connecting the tinfoil coating on the neck n to the terminal, and the
leading-in wire w to an insulated plate. If the bulb stands in a
socket upright, as shown in the cut, a shade of conducting material
may be slipped in the neck n, and the action thus magnified.

A more perfected arrangement used in some of these bulbs is
illustrated in Fig. 26. In this case the construction of the bulb is
as shown and described before, when reference was made to Fig. 19. A
zinc sheet Z, with a tubular extension T, is slipped over the metallic
socket S. The bulb hangs downward from the terminal t, the zinc sheet
Z, performing the double office of intensifier and reflector. The
reflector is separated from the terminal t by an extension of the
insulating plug P.

[Illustration: FIG. 27.--PHOSPHORESCENT TUBE WITH INTENSIFYING
REFLECTOR.]

A similar disposition with a phosphorescent tube is illustrated in
Fig. 27. The tube T is prepared from two short tubes of a different
diameter, which are sealed on the ends. On the lower end is placed an
outside conducting coating C, which connects to the wire w. The wire
has a hook on the upper end for suspension, and passes through the
centre of the inside tube, which is filled with some good and tightly
packed insulator. On the outside of the upper end of the tube T is
another conducting coating C_1 upon which is slipped a metallic
reflector Z, which should be separated by a thick insulation from the
end of wire w.

The economical use of such a reflector or intensifier would require
that all energy supplied to an air condenser should be recoverable,
or, in other words, that there should not be any losses, neither in
the gaseous medium nor through its action elsewhere. This is far from
being so, but, fortunately, the losses may be reduced to anything
desired. A few remarks are necessary on this subject, in order to make
the experiences gathered in the course of these investigations
perfectly clear.

Suppose a small helix with many well insulated turns, as in experiment
Fig. 17, has one of its ends connected to one of the terminals of the
induction coil, and the other to a metal plate, or, for the sake of
simplicity, a sphere, insulated in space. When the