424 REPORT—1883. 
(This last expression is correct for all practical purposes, but requires a few further 
experiments to ascertain the correctness of the law with greater exactitude.) 
a, b, m, n, and a are of course constants, which require determination for each 
lamp. 
The mode of ascertaining these constants was then described, by means of which 
a curve of potential and current can be plotted, and the constants a and 6 calculated 
by the method of least squares, if necessary. From the observed currents and 
potentials the Watts can be calculated, and also from the corrected currents and 
potential, the latter being found to be more accurately observed than the former. 
The same curve is adopted for the resistances. The resistances, current, and Watts 
will be found to be nearly coincident when calculated from the direct observations 
or from the corrected curves of current and potential. To find the constant » the 
observations of corrected Watts and deflections are plotted, the one as ordinates, 
and the other as the abscissse to the curves, when it will be found that the curve 
at any temperature above 530° C. is a straight line, and is thus readily obtained 
either by calculation or by a graphic method, as is also m. 
The constant / can be obtained by observing the resistance at 530°, the tempe- 
rature when luminous radiation just commences. 
By this plan all constants are known, and any required temperature can be 
obtained by increasing the potential, and if necessary introducing a known resist- 
ance in the circuit. In choosing an incandescent solid, however, there are certain 
conditions that require attention. In the first place the section of the radiating 
body should be uniform, and also homogeneous, The carbon threads such as those 
prepared by Edison meet this condition as fully as practicable. Thismay be readily 
ascertained by passing a current of such an intensity through the filament as just 
to. cause it to be at a red glow when seen in a darkened room. If the filament he 
uniform in section and homogeneous, the glow will be seen to be equally bright in 
every part of its length, no dark patches being apparent. Another condition which 
also should be fulfilled theoretically is that the body should radiate on to matter 
which is everywhere of uniform temperature, or nearly so. In an ordinary in- 
candescence lamp this is not quite the case, for if the filament be of the form of a 
simple loop, the two legs must radiate one on to the other, and the inner surfaces 
should have a higher temperature. At the distance apart at which these legs are 
placed this difficulty does not arise, but in making a standard lamp it is proposed 
that it should radiate from a single thread. The best method of construction of 
such a lamp the Committee propose to submit in a subsequent report. 
The light which it is proposed to employ as a standard of quality is as follows. 
Taking the colour of Mr. Vernon Harcourt’s standard as a comparison light—the 
red (at the C line of the solar spectrum) being taken as equal in the two lights, the 
light at E in the new standard should be 1°5 times that of the gas-light; the increase 
in intensity of the higher radiations will then follow of necessity. Compared with 
the electric light this increase in the green is small, as the increase in the green of 
the crater light (positive pole) is very nearly three times that of gas-light. 
When possessed of one lamp of which the necessary constants for the production 
of the standard temperature are known, any other lamp which has a uniform fila- 
ment may be standardised by direct comparison with it by increasing or diminishing 
the current till the shadows as thrown by the Rumford photometer ona white screen 
appear of the same tint. It will be found that a very slight alteration in current 
from the point at which the shadows appear equal in brightness and similar in colour 
will alter the latter. By this plan the original standard may be preserved for a 
considerable period, the second lamp taking its place in all photometric or other 
experiments. 
The method of obtaining an exact quality of light has now been indicated, and 
the quantity of light radiated can easily be proved by direct experiment. It is pro- 
posed that the amount of candle-light (so called) be obtained by measuring with a 
photometer the standard light proposed by Mr. Vernon Harcourt with the lamp at 
the given temperature, the observation being made through a cell, the plates of 
which are 1 mm. apart, filled with an aqueous solution of iodine and iodide of 
potassium made as follows :— 
