ON COLOUR PHOTOMETRY. 
433 
been unable to trace any variation from the curve referred to above. We are aware 
that statements have been made regarding the luminosity of the red of the spectrum, 
to the effect that when very dim it loses proportionally more luminosity than do the 
other colours; but in the experiments which we have carried out we have not detected 
this difference. 
§ XIII. The Illuminating Values of Colour Mixtures compared with those of their 
Components. 
It has been assumed, but, as far as we know, never been experimentally proved, 
that the impression on the eyes of a mixed light is equal to the sum of the 
impressions of each of the components of the light. 
If this law be correct, and our observations did not confirm it, it is evident that our 
method must be untrustworthy. If, on the other hand, our observations, made under 
varying circumstances of intensity of illumination, obey this law, then the proba¬ 
bility is that our method is sound, and the law correct. To test the illuminating 
value of colour mixtures, instead of the slide with the single slit, one with three 
slits, of different widths, any of which could be closed at pleasure, was used, so that 
three slices of the spectrum could be dealt with simultaneously. The collecting lens 
being used as before, the light from each slit was thrown on to the same patch on the 
screen. The illuminating value of each slice was then determined, as well as that of 
the slices taken in pairs, and of the three together. In these measurements it was 
necessary to move the comparison-light, # as the slits must necessarily be stationary, 
and, as we have stated before, this method is not quite so accurate as the other. 
We append some examples. The slide was so placed that one slit occupied a 
position in the violet, another in the green, and the third in the red. Measures were 
taken of these colours singly, in pairs, and altogether, and equations obtained by the 
method of least squares from these six measurements, with the following results :— 
Observed. 
Calculated. 
R 
203 
204-25 
(R+-G) 
242 
241-75 
G 
38-5 
37-50 
(G + Y) 
45 
46-00 
V 
8-5 
8-50 
(R + V) 
214-0 
212-50 
(R + G + Y) 
250-0 
250-25 
The accuracy of the results may perhaps be best shown by adding the values of the 
single colours together, and the pairs and single colours, and comparing these values 
with that obtained by the three colours combined. 
* It will be seen in tlie addendum that this is now unnecessary if a rotating disc with moving' sectors 
be employed in front of the comparison-light. 
MDCCCLXXXVI. 3 K 
