Sees 
, TRANSACTIONS OF SECTION A. 423 
_a standard of white light, in that it is surrounded to a greater or less degree with 
carbon vapour, which, though radiating but little energy, yet radiates that energy 
chiefly as bright bands in the green and the blue of the visible spectrum. Could 
these bands be eliminated there is a temperature which is apparently constant, and 
which, consequently, will radiate also the same proportionate intensity of rays. 
Failing this, the incandescence lights offer the next best standard; and though 
when compared with daylight of an ordinary character they appear yellow even at 
their highest practicable temperature, yet they are much whiter, containing more 
proportionate green, blue, and violet than gas-licht, taking the red near the C line 
as equal in both cases, Again, we have another decided advantage over gas in the 
fact that the body heated is a solid, and, for practical purposes, black. In gas-light 
there is a decided preponderance of yellow and orange, compared with a solid 
heated to the same temperature. Hence the “spectrum range,’ to coin a word, is 
more accurate with the incandescent lamp than with the gas. <A point that re- 
quired investigation was as to whether all carbon-threads emitted the same relative 
proportion of spectrum rays, and it was found that they did so, and that at what 
is believed to be the same temperature, the proportion of these rays remained 
constant. (The proportion was obtained by comparing it with ignited coal gas.) 
Hence we arrive at one step in fixing a standard quality of light. The question 
arises as to what temperature the carbon filament may be heated without endan- 
gering the existence of the lamp. At one stage of heat in the carbon-thread of a 
well-exhausted lamp there is a peculiar glow, which illuminates the bulb of the 
lamp, and if that glow be examined by the spectroscope it will be found to consist 
of four or five bright lines, due to carbon vapour in some shape or another; and if 
that temperature be maintained the carbon is found to be deposited as an impalpable- 
powder on parts of the glass globe, and eventually the thread breaks at the place 
of greatest resistance. Below this heat the thread will remain unaltered for many 
hours without any apparent change, always supposing the thread to have been 
previously heated to such a degree as to give constant resistance at freezing point. 
This is a matter of some importance, as in the experiments made new lamps in- 
creased in resistance after a few hours’ ignition as much as five per cent., and after 
that remained constant, when heated to a temperature below that already indicated. 
An investigation then took place regarding the intensity of radiation from an in- 
candescent carbon filament and the energy and temperature. The results of these 
experiments are given in the ‘ Phil. Mag.’ September 1883, in which it will be seen 
that after a certain temperature (dependent on the thickness of the filament and 
the temperature of the surroundings) the radiation and the energy expended are 
directly proportional. A good fiducial temperature is when the carbon-thread is 
just visible to the eye when examined in a darkened room, and is very nearly 
530° C. If the energy at this temperature be accurately measured by means of 
the current and the electro-motive force, and if the resistance be measured at the 
temperature of melting ice, the temperature of filament at just below the point 
below which the carbon lines appear can be readily obtained by diminishing the 
resistance of the carbon filament by half. This has been found to be approximately 
the temperature required. Another check-method is to note the radiation by 
means of the thermopile at the point of first visible incandescence, and to increase 
the energy expended till the radiation noted is forty times as great. This can be 
effected with great facility, and the quality of light radiated is in this case in- 
variably the same, as it is indeed if any other proportion be taken. 
It may be well to note here the expressions which exist between— Watts, W ; 
radiation, D; potential, P; current, C; resistance, R; temperature, T. 
C=ap+ bp? 
W =p" (a+ bp) 
1l—ar\1 1 
R=(—"). > 
D=m+nW 
a 
= 
