1888.] Mr G. N. Stewart on Intermittent Light. 
457 
Flame 2 feet from mirror . — Same appearances as with flame at 
2J. feet distance, but the greenish phase is better 
marked. The reddish edges are in no case so distinct 
as the coloured edges in the other experiments. 
The above will suffice as examples of the observations with 
coloured light. With white light I have made more than a hundred 
similar experiments, using besides the small mirror described, which 
was silvered on the back, a larger one silvered on the front, which 
was kindly lent me by Professor Schuster of Owens College. The 
result has been always the same, although I think that the small 
mirror showed the phenomena best. 
The explanation of the main part of the appearances, viz., those 
in the broad central part of the image, which I have given above, 
must now be considered in relation to the phenomena seen with the 
monochromatic light. And it will be well to discuss at the same 
time the changes at the edges. 
Take Experiment 3. — Here the pure red light gives a golden 
yellow with the longer period of stimulation, and a red with the 
shorter. We must explain this as follows : — We know that red 
light, if very intense, gives the impression of orange, or even 
ultimately of yellowish-white. It is, therefore, assumed that red 
light has the power of exciting not only the red fibres, but also the 
violet and the green. When the light is of moderate intensity, the 
violet and green fibres are only feebly excited, but, as the intensity 
is increased, the excitation of the red fibres reaches its maximum, 
and after this the excitation of the green and violet increases with- 
out any further increase in the red. 
In the first stage of Experiment 3, we may suppose that the time 
of stimulation is more than enough, in the bright central part, to 
allow the red fibres to reach their maximum, and therefore the green 
fibres are strongly excited, and the yellow colour is the result. At 
the more feebly illuminated edges the stimulation of the red fibres 
may not have reached its maximum, and with the long stimulation 
time the green may preponderate (see fig. 4), for, of course, the curve 
of excitation will be the same whether the stimulus be white light 
or coloured. It might seem simpler to assume that the appearances 
at the edges are contrast phenomena. That effects of contrast mix 
themselves with the other effects is not only probable, but, I think. 
