ON COLOUR PHOTOMETRY. 
453 
§ XXIII. Unit Surface Values of Illumination. 
As a matter of curiosity we have made a comparison between the illuminating 
values per unit of surface of the filament of the incandescence lamp with the current 
from 28 boxes and of the crater of the arc light. 
This was done in the following manner : by means of a lens of about 9 inches focal 
length, the image of the crater was accurately focussed on a slit of moderate dimen¬ 
sions, which it well filled. The divervinof beam of light from the slit was made to 
fall on a white screen, and its illuminating power compared with that of a candle by 
the Rum ford method. The relative distances of all parts of the apparatus was care¬ 
fully measured, that of the screen from the slit being in this case 498 inches, and of the 
lens to the slit 89’5. The mean distance of the candle was 38 inches from the screen. 
The positions of the lens and slit being unchanged, the incandescence lamp was sub¬ 
stituted for the arc, and the image of the filament was made to fall on the slit, which 
it more than filled. The screen was then placed at 93‘5 inches from the slit, and the 
light compared as before with that of the same candle, which had to be placed at 
157 inches from the screen. Combining these two measurements it was found that, 
area for area, the light of the crater was 484 times that of the filament in the incan¬ 
descence lamp. We have found this to be a convenient plan of measuring the 
illuminating values of areas, and one of us extensively practised this method in a long 
series of experiments undertaken for the War Department at Chatham six years ago. 
Postscript. 
(Added March 4, 1886.) 
§ XXIV. Practical Application of the Mode of Testing. 
As an example of the practical method of testing luminosity by rapid alternations of 
“ too light and too dark,” we would call attention to the method that can be adopted 
in evaluating the light from a candle and a glow-lamp. Suppose we wish to know 
the current, &c., which in a given glow-lamp will produce say 16 candles, a candle 
is placed say 3 feet and the glow-lamp 12 feet from the screen, before which a rod is 
placed. A current is passed through a glow-lamp and some carbon pellet resistances, # 
which can be screwed together more or less tightly. The resistance is increased or 
reduced as may be found necessary till it gives an equal illumination to the screen 
that the candle does. The pellet resistances are then tightened and slackened alter- 
* Since the above experiments were made we have substituted for the pellet resistances carbonized 
cloth resistance, which was kindly supplied by Messrs. Yarley of Mildmay Park. The change of resistance 
in the circuit which this can effect by screwing up is sufficient to allow an ordinary incandescence lamjo 
of 100 ohms resistance to vary between a black heat and a full white heat.—[July 23.] 
