ON COLOUR PHOTOMETRY. 
429 
into one of a more convenient range for plotting.] The collecting lens was then so 
placed that the whole spectrum formed a patch about 4 inches square on the screen. 
§ VI. The Receiving Screen. 
In our earlier experiments we used white cardboard as a receiving screen and for 
ordinary purposes this answered admirably; but the question arose whether card of 
the same whiteness could under all circumstances be obtained. On consideration we 
came to the conclusion that zinc oxide could be used with advantage as a standard 
white. It is to be found in every laboratory, and can be used in distemper. The 
medium with which we mixed the zinc oxide was a very pure white gelatine dissolved 
in water, which we used as sparingly as possible. On comparing the illuminating 
values of different parts of the spectrum on the cardboard with those of a spectrum 
on a surface coated with zinc oxide, we found a trace of deficiency in the violet, and 
a smaller one in the green in the former. This was a matter of some concern, as it 
rendered necessary a repetition with the zinc oxide screen of the series of experiments 
which we had made with the cardboard. Other white surfaces were tried, but on the 
whole the zinc oxide answered best. 
A card, or board, was brushed over with a cream of the substance, and when dry a 
second coat was applied and then flatted down with a brush. An ordinary white 
card placed alongside it appeared yellowish, while the coated card seemed bluish 
by contrast. 
The portion of the screen used for receiving the lights to be compared was limited 
to a space about 2 inches square by a sheet of black paper, with a hole of that size 
cut in it placed close in front of and in contact with the screen. 
§ VII. The Comparison-Light. 
In selecting a comparison-light it should be remembered that it has to be moved 
and is not fixed in position. An electric glow lamp, with a constant current through 
it, would no doubt be very convenient were it not for the fact that it does not 
give out light symmetrically in all directions, owing to the different surface of carbon 
exposed; and the slightest deviation in angular position makes a considerable dif¬ 
ference in the illuminating value of such a source at the screen. The flexible connexion 
necessary for a gas flame is a source of error, as it is liable to get pinched, and the supply 
of gas to be thus made irregular, and although many readings were taken with a gas 
flame we found it inferior, with the appliances at our command, to an ordinary candle, 
or to the Siemens’s standard unit lamp burning amyl acetate, which has the value of 
about '9 of such a candle. 
In the experiments recorded in the immediately following pages a candle was used 
as the comparison light. Burnt as we burnt it, open to the air but not exposed to 
