118 Proceedings of Royal Society of Eclinhiirgh. [sess. 
These facts are corroborated by another test of a different 
character, viz., the quantity, of blue required for neutralisa- 
tion. For me vermilion required 0’37 proportion of blue, for 
j\[r Parry OTl were sufficient; while emerald-green required for 
me 0-37, for Mr Parry 0.66. 
These two cases are the extremes, and if we look at the other 
cases we find them irregularly intermediate. For example, the 
chromic strength varies in the five columns thus : — 
For vermilion, . . 235 130 85 80 80 
For emerald-green, . 23 33 28 42 43 
and, similarly for other data — strong presumptive proof against 
grouping. 
I may call attention also to what I think is a new feature in these 
data, viz., the distinction between the effects of the saturation and 
the luminosity on the colour impression. 
The impressions given by the red ra}^s are nearly all perfectly 
saturated ; whereas those given by the green rays are only saturated 
to about 80 per cent. This accords perfectly with the spectral 
explanation given in my figure 5, and page 111. Then, again, the 
saturation remains constant, or nearly so, for each colour through 
the whole series ; the variation being in the luminosity only. This 
points to a distinct cause of the variability. 
The luminosity of the neutral mixture of red and blue is very 
small, and varies to some extent with the power of the red 
impression. That of the neutral blue-green is much greater, and 
nearly constant, owing to the large mixture of white with the green 
rays. 
It is a striking feature, in contrast with the red and green 
variations, that the relations of the four fundamental colours — yellow, 
blue, white, and black — with each other, as shown in Equation I., 
are essentially the same in all the cases, showing, I think, a general 
uniformity of the whole of the experiments. I was not aware, at 
that early date, of the extreme precautions now thought necessary 
in regard to the colour of the daylight, &c., but I think the series 
shows that the care taken was sufficient to get fairly good results. 
But although the fundamental colours agree so well, there is 
clearly a tendency to variation generally, wdiich is usually slight, 
