436 
CAPTAIN ABNEY AND MAJOR-GENERAL FESTING 
§ XV. Testing the Perceptions of Colour in Other People. 
The slit slide was placed in the spectrum focus, and we placed the comparison light 
ourselves, requesting the observer to oscillate the slide until he obtained a balance. 
The position of the slit in the spectrum was then noted by one of us by reading the 
scale at the back of the slide, the observer himself not being able to see the scale. 
The observation was repeated three times for each position of the candle. We have 
often found the first two or three measurements to be rather wild, but the observa¬ 
tions soon get more accurate. We find it well, as a rule, to place the comparison light 
so that the first observations shall be a little on the D side of E and then to work up 
towards the maximum on both sides. By this plan the accuracy of the judgment 
increases as the maximum is reached. The observer then works back into the violet. 
This plan ensures a double comparison in the most difficult part of the spectrum, and 
in case of much difference between the first and second sets of readings a third set is 
taken ; this we have invariably found to agree better with the second than with the 
first set. It should be recollected that few persons can use an ordinary Bumford 
photometer with much accuracy until they have had some practice with it; it is not 
therefore surprising that we have found that those who have had some practice in 
ordinary photometry are, as a rule, much better observers than those who have had no 
such practice. In some cases we have caused observations to be repeated on two or 
three separate occasions, and have always found that with untrained eyes the ease 
and accuracy of observation are much increased in the later trials. 
On comparing the “ curves of luminosity ” plotted from the observations of different 
persons, we have found some to be identical, or nearly so, with our own ; but in some 
cases, even of persons not actually colour-blind, this is not the case. The variable 
part may be in the green near the yellow, or in a want of normal perception of red, 
the appreciation of these colours not being so strong as with us. 
Observer K. — We append a curve, fig. 4, Plate 24, obtained from a very careful 
observer wdiom we will call K. 
K’s Curve. 
This carve is the same as the normal curve, except at the following scale numbers :— 
Scale reading. 
Intensity. 
47-0 
100 
47-8 
87 
48-3 
80 
49-0 
68 
49-4 
60 
49-6 
54 
49-8 
47 
50-0 
42 
50-1 
40 
