e "paste" process already described. These methods
produced fragile filaments, but their luminous efficiency was higher
than that of previous ones. However, in this country ductile tungsten
was soon on its way. An ingot of tungsten is subjected to vigorous
swaging until it takes the form of a rod. This is finally drawn into
wire.

Much of this development work was done by the laboratories of the
General Electric Company and they were destined to contribute another
great improvement. The blackening of the lamp bulbs was due to the
evaporation of tungsten from the filament. All filaments up to this time
had been confined in evacuated bulbs and the low pressure facilitates
evaporation, as is well known. It had long been known that an inert gas
in the bulb would reduce the evaporation and remedy other defects;
however, under these conditions, there would be a considerable loss of
energy through conduction of heat by the gases. In the vacuum lamp
nearly all the electrical energy is converted into radiant energy, which
is emitted by the filament and any dissipation of heat is an energy
loss. A high vacuum was one of the chief aims up to this time, but a
radical departure was pending.

If an ordinary tungsten-lamp bulb be filled with an inert gas such as
nitrogen, the filament may be operated at a very much higher temperature
without any more deterioration than takes place in a vacuum at a lower
temperature. This gives a more efficient _light_ but a less efficient
_lamp_. The greater output of light is compensated by losses by
conduction of heat through the gas. In other words, a great deal more
energy is required by the filament in order to remain at a given
temperature in a gas than in a vacuum. However, elaborate studies of the
dependence of heat-losses upon the size and shape of the filament and of
the physics of conduction from a solid to a gas, established the
foundation for the gas-filled tungsten lamp. The knowledge gained in
these investigations indicated that a thicker filament lost a relatively
less percentage of energy by conduction than a thin one for equal
amounts of emitted light. However, a practical filament must have
sufficient resistance to be used safely on lighting circuits already
established and, therefore, the large diameter and high resistance were
obtained by making a helical coil of a fine wire. In fact, the
gas-filled tungsten lamp may be thought of as an ordinary lamp with its
long filament made