1888 .] Mr G. N. Stewart on Intermittent Light. 
453 
tion, the zig-zag denoting that the fusion is not quite complete. 
After the first flash the curve will always lie above the base line, 
and the dotted lines rising from the latter show the hypothetical 
course of the curve for the successive flashes. The horizontal dotted 
lines are supposed to be drawn parallel to the abcissa axis, so that 
as much of the zig-zag curves lie above them as below them. The 
mean ordinates are, therefore, the distances between these dotted 
lines and the abcissa. For any given length of each flash we 
have now to take the proportion of the mean ordinates for the three 
curves. As drawn, the figure would indicate a preponderance of 
red corresponding to The descending portion of the curve 
corresponding to the decline of the excitement between each flash 
may be neglected, since fusion is nearly complete. In one experi- 
ment 0^6^" was about 1/1200 second, while the length of flash 
for the green phase was 1/700 second, and for the violet 1/500 
second. 
3. Is any known fact contradicted by drawing the curves as in 
fig. 4? 
There is one which seems to be contradicted, and that is the 
observation of Kunkel, that violet reaches its maximum sooner than 
green, from which it might appear probable that the violet curve 
