with extreme rapidity, it emits at the point of fusion less light and
more invisible radiation than it does when melted by a steady current,
though the total energy used up in the process of fusion is the same
in both cases. Or, to cite another example, a lamp filament is not
capable of withstanding as long with currents of extreme frequency as
it does with steady currents, assuming that it be worked at the same
luminous intensity. This means that for rapidly alternating currents
the filament should be shorter and thicker. The higher the
frequency--that is, the greater the departure from the steady
flow--the worse it would be for the filament. But if the truth of this
remark were demonstrated, it would be erroneous to conclude that such
a refractory button as used in these bulbs would be deteriorated
quicker by currents of extremely high frequency than by steady or low
frequency currents. From experience I may say that just the opposite
holds good: the button withstands the bombardment better with currents
of very high frequency. But this is due to the fact that a high
frequency discharge passes through a rarefied gas with much greater
freedom than a steady or low frequency discharge, and this will say
that with the former we can work with a lower potential or with a less
violent impact. As long, then, as the gas is of no consequence, a
steady or low frequency current is better; but as soon as the action
of the gas is desired and important, high frequencies are preferable.

In the course of these experiments a great many trials were made with
all kinds of carbon buttons. Electrodes made of ordinary carbon
buttons were decidedly more durable when the buttons were obtained by
the application of enormous pressure. Electrodes prepared by
depositing carbon in well known ways did not show up well; they
blackened the globe very quickly. From many experiences I conclude
that lamp filaments obtained in this manner can be advantageously used