438 
CAPTAIN ABNEY AND MAJOR-GENERAL FESTING 
He describes it as blue, as if the red element which to us is present in violet were 
unperceived by him. 
Having measured and plotted our luminosity curve of the candle, the area of which 
should represent the strength of the comparison-light as seen by us, we proceeded to 
derive from it a curve whose area should represent that as seen by him. This we did 
by reducing the ordinates in the same proportion that those of G’s curve mentioned 
above were less than those of the normal curve. The areas of these two last curves 
were determined to be in the proportion of 1 : ‘6 87. It would thus appear that the 
measurements made by G would be exaggerated as compared to ours in this propor¬ 
tion ; and that, therefore, to compare his original curve with ours his ordinates should 
be multiplied by ’687. The factor wdrich we actually used, as we have said above, 
was '667, differing only - 02 from that obtained by the method described. This may, 
we think, be looked upon as a fair confirmation of our method of proceeding. Similarly 
in reducing the curves of H and H we used the factors '465 and ’425, and the factor 
found as above was ‘455 and ’408. For Dr. Pole’s curve the factors were respectively 
•952 and '93. 
The bearing of this on practical photometry of unanalysed light is worthy of atten¬ 
tion. Colour-blind people can compare two lights of nearly the same tint without 
apparent error. But when there is any appreciable difference in colour their com¬ 
parisons will differ from those of people with normal sight. 
As an illustration—the ratio of the area of G’s reduced electric-light curve to that 
of the normal curve is ‘838 : 1, while the ratio of the areas of the candle curve is 
’687 : 1 ; in comparing the electric light with that of a candle we should evidently 
838 
estimate the former light as being or 1 - 22 times stronger than a normal-eyed 
person woidd ; and this is the proportion of the area of his original curve to that of 
the normal curve. 
