high frequencies, certainly 20 times, if not more, the quantity of
light which is obtained in the present incandescent lamp by the same
expenditure of energy. This estimate may appear to many exaggerated,
but in reality I think it is far from being so. As this statement
might be misunderstood I think it necessary to expose clearly the
problem with which in this line of work we are confronted, and the
manner in which, in my opinion, a solution will be arrived at.

Any one who begins a study of the problem will be apt to think that
what is wanted in a lamp with an electrode is a very high degree of
incandescence of the electrode. There he will be mistaken. The high
incandescence of the button is a necessary evil, but what is really
wanted is the high incandescence of the gas surrounding the button. In
other words, the problem in such a lamp is to bring a mass of gas to
the highest possible incandescence. The higher the incandescence, the
quicker the mean vibration, the greater is the economy of the light
production. But to maintain a mass of gas at a high degree of
incandescence in a glass vessel, it will always be necessary to keep
the incandescent mass away from the glass; that is, to confine it as
much as possible to the central portion of the globe.

In one of the experiments this evening a brush was produced at the end
of a wire. This brush was a flame, a source of heat and light. It did
not emit much perceptible heat, nor did it glow with an intense light;
but is it the less a flame because it does not scorch my hand? Is it
the less a flame because it does not hurt my eye by its brilliancy?
The problem is precisely to produce in the bulb such a flame, much
smaller in size, but incomparably more powerful. Were there means at
hand for producing electric impulses of a sufficiently high frequency,
and for transmitting them, the bulb could be done away with, unless it