1888.] Mr G. N. Stewart on Intermittent Light. 
463 
take first the case of decreasing brightness. Let the curves of 
excitation be drawn as in fig. 6, viz., and for redg 
corresponding respectively to times of stimulation and 0 ^-^ ; 
similarly for green and violet. Further, let the declining part of 
the curve, after stimulation has ceased, be drawn for time O^h^, viz., 
for red, <^2^2 green, and for violet. 
\ - 
/ 
C; 
% 
^ A 2 
C2 
0 , 
h 
^3 
0 / 0 ^ bj 
Fig. 6. 
I^ow, if the time of stimulation remained steadily equal to 
the curve after the stimulation had risen to its maximum would be 
represented by the interrupted line a^d-^ for the red, ^26^2 green, 
and for violet. The proportion of the three ordinates must be 
about that corresponding to white light, since, at complete fusion, 
the original brightness is only diminished by half. 
Suppose now that the time of stimulation is reduced to O2&2) and 
consider the state of affairs a short time after, say in the position of 
the vertical line passing through Cy If the excitation came imme- 
diately to correspond to the new time, the new curves would at once 
be 02A^D^, O2A2D2, and O2A3D3. But we know that the excitation 
takes some time to fall away, and that it falls away more slowly in 
the violet and green fibres than in the red (at least at first, and that is 
what is important here), so that a very short time after the change of 
speed has been made the curve of red has fallen nearly to the height 
of that corresponding to the new time, while the violet and green 
of the old stimulatian are still present in excess. For a sensible time 
after the change of speed the curves of violet and green must have 
higher ordinates than the interrupted lines A3D3, A2r>2. There will 
thus be an excess of green and violet, and the star will be bluish. 
VOL. XV. 12/12/88 2 G 
