G.—ENGINEERING, 181 
from his allegiance to direct currents, for at his last appearance at the 
Institution of Electrical Engineers, in 1907, he said, ‘ I have never swerved 
from the opinion that the right system for long-distance transmission of 
power by electricity is the direct-current system.’ 
The development of the dynamo during the seventies and the simul- 
taneous development of the incandescent lamp led to the general introduc- 
tion of electric light during the eighties. Attempts to make incandescent 
electric lamps had been made as early as 1841, when de Moleyns patented 
one having a spiral platinum filament, and in 1847 Grove illuminated the 
lecture theatre of the Royal Institution with such lamps, the source of 
power being primary batteries; but it was not until 1878 that the commercial 
development of the incandescent electric lamp was begun by Edison 
and Swan. 
One of the earliest complete house-lighting installations was put in by 
Kelvin in 1881. A Clerk gas-engine was used to drive a Siemens dynamo, 
a battery of Faure cells was fitted up, and every gas-light in his house and 
laboratory at Glasgow University was replaced by 16 candle-power Swan 
_ lamps for 85 volts. He had to design his own switches and fuses, etc., 
for such things were almost unknown. 
For about twenty years the carbon-filament lamp held the field without 
a rival for interior illumination, and, although attempts were made to 
improve its efficiency by coating the filament with silicon, the plain carbon 
filament only gave way finally to the metal-filament lamp. One of the most 
interesting developments in the history of electric lighting was the Nernst 
lamp, which was introduced in 1897 ; the filament consisted of a mixture 
of zirconia and yttria, and not only had to be heated before it became 
conducting but also had to be connected in series with a ballast resistance 
in order that it might burn stably. The way in which these difficulties 
were surmounted and the lamp, complete with heater, ballast resistance, 
and automatic cut-out, put on the market in a compact form occupying 
little more space than the carbon-filament lamp was, in my opinion, a 
triumph of applied science and industrial research. The efficiency was 
about double that of the carbon lamp. About this time, however, a return 
was made to the long-neglected metal filament. The osmium lamp invented 
by Welsbach in 1898 was put on the market in 1902, to be followed two 
years later by the tantalum and tungsten lamps. The latter was greatly 
improved by the discovery in 1909 of the method of producing ductile 
tungsten and by the subsequent development of gas-filled lamps in which 
the filament can be run at such a temperature without undue volatilisation 
that the consumption is reduced in the larger sizes to 0°6 watt per mean 
spherical candle-power. This improvement of eight times as compared with 
the efficiency of the carbon-filament lamp has led to the gradual replacement 
of the arc lamp even for outdoor illumination. The arc lamp was intro- 
duced at about the same time as the carbon-filament lamp, the Avenue 
de l’Opéra having been lit with Jablochkoff candles in 1878. The open 
are was developed during the eighties; the enclosed arc, giving long burning 
hours and thus reducing the cost of re-carboning, was introduced in 1893, 
and the flame arc in 1899. During the first few years of this century the 
flame arc was brought to a high stage of development and the consumption 
brought down to about 0°25 watt per candle-power, but the necessity 
of frequent cleaning to prevent the reduction of efficiency by dirt and the 
