262 Proceedings of the Royal Society of Edinburgh. [Sess. 
occasion may require. A simple variant of the mirror wheel just described 
is a complete glass disc silvered, however, only over a semicircle, the 
beam X being transmitted through the clear half. This arrangement has 
the curious property of being unsymmetrical except when the two beams 
are of equal intensity , as will readily be discovered. The result is that 
it will give true readings, but with a slight loss of sensitiveness. 
IY. It remains to call attention to some general principles regarding 
flicker spectrophotometers. 
(1) In substituting one beam for the other, each ray of the beam just 
withdrawn must have its place exactly taken by the corresponding ray of 
the new beam, otherwise a sensation of flickering will persist when the 
intensities of the two beams are made equal. Before I discovered this I 
had a great deal of trouble in a certain case where the two beams came 
from different light-sources, because, as I found afterwards, the image of 
the aperture 0, fig. 3, formed on the retina of the eye by the two beams 
respectively was in the one case formed by a narrower cone of rays than 
in the other. The arrangement of a suitable diaphragm in the path of the 
wider beam at once got rid of the difficulty. 
(2) Not only must the corresponding rays of the two beams agree 
exactly in position, but they must agree also in brightness if flicker is 
to cease entirely at the position of balance. For example, if one of the 
two beams has its central rays somewhat less bright than its marginal 
rays, while the other beam is quite homogeneous, then it is impossible to 
make a satisfactory measurement. It is often easier to discover than to 
remedy such a difficulty. Two beams which would give quite good results 
with an ordinary spectrophotometer, in which they are merely brought 
into contact edge to edge, may prove unworkable with a flicker instrument 
because of their non-uniformity of brightness. It may perhaps be urged 
that such beams ought not to be used in any circumstances, and that 
therefore this peculiarity of the instrument should not be considered a 
drawback. 
(3) It will be noticed that all the above instruments have been designed 
so that the transition from the one beam to the other is very sudden. 
This I have done in consequence of some experiments which I made to 
find whether a sudden or a gradual replacement of the one beam by the 
other gave rise to the more decided flickering. The evidence was con- 
clusively in favour of a transition as sudden as could be arranged for, and 
a psychologist whom I questioned on the subject informed me that a 
change of brightness may be expected to be more noticeable if quickly 
accomplished. 
