80 
Proceedings of the Royal Society of Edinburgh. [Sess. 
adjusted to pure red, and steps taken along the line red-A. Similarly, steps 
were taken along the lines red-B, red-C, and red-green. The marks in 
fig. 8 indicate in each case the ends of the steps. These marks were then 
joined by contour lines ; the closer the contour lines, the greater the 
power of discriminating difference of tint. 
Filters 6, 7, 8, and 9 were then fitted, and the steps made on the trans- 
verse paths, blue-green side, DE, FG, and HJ, noted. In this case the 
start was always made from the median, and steps taken down to blue ; 
then the filters were set on the median again, and steps taken up to green. 
This procedure gave a clearer diagram than when the start was made 
from a side. The steps made on the transversals are indicated in fig. 9 ; 
they were afterwards joined by broken contour lines. The completed 
GREEN 
Fig. 8. 
GREEN 
diagram of this observer is given in fig. 10. His colour vision was normal. 
The complete set of observations he made is given below, as reduced 
in the record of tests ; the numbers are in each case percentages : — 
Red to blue 
28 
39 
53 
70 
79 83 
100 
„ i green 
28 
42 
53 
63 
73 83 
90 96 
1 
2 ” 
13 
28 
35 
47 
55 70 
78 89 
99 
3 
55 4 ” 
20 
28 
39 
49 
60 69 
79 88 
99 
,, green . 
19 
29 
41 
55 
68 86 
95 100 
Median to blue 
. -20 
32 
41 
50 
to green 
1 -1-14 
28 
3 
55 4 55 
9 
18 
27 
35 
42 
55 I „ 
18 
43 
i 
9 
19 
31 
41 
55 J 55 
21 
41 
1 
55 4 ” 
. 26 
55 i 5, 
21 
46 
A complete test takes anything from thirty to sixty-five minutes. Half 
the time is occupied in changing and fitting the filters. Observers with 
