432 
CAPTAIN ABNEY AND MAJOR-GENERAL FESTING 
§ X. Contrast Colours. 
The change of apparent colour by contrast was remarkably shown in the course of 
our measurements. It must be recollected that the true colour of the shadow 
illuminated by the candle is yellowish as compared with daylight. At only one part 
of the spectrum, viz., near D, does it appear to us of its true colour. The apparent 
change in colour in passing along the spectrum is very significant. In the red part of 
the spectrum it assumes a bluish tint, and in the green part a ruddy tint. It must 
be remembered that the field on which these colours lie is compounded of the light of 
the candle and of that of the particular part of the spectrum employed. 
We have endeavoured to describe, as they appeared to us at different parts of the 
spectrum, the colours of the two lights :— 
Candlelight. 
Spectrum. 
At 55 of the spectrum scale 
Pale oak stain. 
Deep violet. 
53 
Dirty orange. 
Deep cobalt. 
521 „ 
Seville orange. 
Artificial ultramarine. 
52 ,, ,, 
Indian yellow and salmon mixed 
Prussian blue. 
501 „ 
Dark yellow ochre. 
Green-blue. 
491 „ 
Raw sienna. 
Laurel leaf. 
49 ,, „ 
Burnt sienna. 
Emerald-green. 
46i „ „ 
Yellow (of candle flame) . . 
Yellow (of candle flame). 
45 
Grey. 
Scarlet in sunshine. 
441 „ 
Slate grey. 
Payne’s grey. 
Cochineal. 
44 
Ruby velvet. 
43f „ 
Slightly blue-grey. 
Cherry-red. 
43i 
Blue-grey. 
Morelia cherry. 
431 „ 
Dark blue-grey. 
Dirty red. 
§ XI. Has the Colour of the Comparison-Light any Effect on the Results ? 
This is one of the first questions which presented itself to our minds, and one to 
which we got a satisfactory answer. A cell containing a solution of fuchsine was 
placed in front of the comparison-candle, the light of which was thus rendered quite 
red. Although the individual readings taken under these circumstances differed more 
from the mean than when the naked candle was used, the general result was the 
same, as it was when green or blue glass or a blue solution in the cell were used to 
mitigate the colour of the comparison-light. These measurements alone would be 
sufficient to arrive at the conclusion that the colour of the comparison-light has no 
effect on the results ; but an additional proof will be seen further on. 
§ XII. Does the Quantity of Light Admitted to Form the Spectrum Cause a Difference 
in the Results ? 
This question is one of supreme importance, and we have experimented largely to 
obtain an answer to it. Observations were made with variations in the width of the 
opening of the spectroscope slit between ToVoth ami M-th of an inch, yet we have 
