202 Dr-G. J: Burch! [ Mar. 7, 
darkness I could not see the bands continuously, but could only catch 
glimpses of them from time to time, as one does when looking for stars in 
early twilight. It is very difficult to tell in total darkness in what direction 
to look, even with the eyebrow touching the eye-piece, so that the band first 
seen was seldom the one I expected to find. Yet the colour was instantly 
recognised. 
Each colour produces a characteristic effect quite apart from tone or tint. 
The red band catches the eye suddenly as it sweeps to and fro through the 
darkness. Once seen, it is easily held for some seconds—and lost as easily. 
Its edges are perfectly definite against the darkness. 
The blue band first shows as a blue haze, persistently appearing in one 
spot, and refusing to follow the eye as it sweeps across. On looking steadily 
the haze resolves itself into a band with fairly definite outlines. (It should 
be noted that the method adopted, of reducing the intensity of the light before 
it enters the spectroscope, effectually does away with the false light that is so 
troublesome in some cases. There is no visible “ field of view.”) 
Violet appears first as a haze of violet after-effect, which seems to have a 
cumulative action with regard to the sensation it excites. It is difficult to 
see the outline of the band, owing apparently to its strong tendency to 
produce after-effects even with such low intensity. But if the eye is kept 
perfectly still the violet is held, when once found, even more easily than 
the red, * 
There is a curious phenomenon associated with the green that I do not 
understand, but have seen so often that I here call attention to it. The 
outline is definite though not so definite as that of the red, but the band 
seems to sparkle as I look at it. That is to say, it is not continuously equally 
bright all over, but spots of slightly greater luminosity of perhaps 5’ or 
8’ diameter appear and disappear in various parts of it. Any portion of 
the spectrum from D to 0 will do this, but beyond 0 the steady light of the 
blue sensation seems to mask the effect, which is only visible when the eye is 
completely rested (after two hours darkness at least) and with very faint 
ilumination—either the minimum visible or a little more. 
Lxpervment 4.—This experiment is the converse of the preceding. I 
arranged a weak continuous spectrum of lamplight which, after half-an-hour’s 
stay in darkness appeared white, or rather colourless, from end to end. After 
setting the broad pointer to mark the extreme limit of this spectrum towards 
the red, I took away the paper reflector and put the lamp in its place. What 
I had taken for the extreme limit of the spectrum proved to be the beginning 
of the red. I concluded, therefore, that the ratio of the light sensation to 
