MAR. 4, 1923 CRITTENDEN: MEASUREMENT OF LIGHT 85 
There is always doubt as to the certainty with which the values can 
be reestablished by reference to the fundamental standards even 
though the color differences involved are the relatively small ones 
represented by incandescent lamps of different efficiencies. Further- 
more, this method is, of course, not at all applicable to color differences 
which can not be reached step by step or for which there is no suitable 
secondary reference standard available. 
Flicker Photometry.— Various methods for meeting or avoiding these 
difficulties have been proposed. The only direct method which has 
given any evidence of practicability is the use of the flicker photom- 
eter. This type of photometer differs from the ordinary one in 
that it presents the two photometric surfaces to view alternately in 
the same place instead of simultaneously side by side. One of the 
peculiarities of vision is the fact that the two colored surfaces can 
be interchanged at such a rate that the disturbing color differences 
_ are no longer seen, although differences of brightness still show in 
the form of a perceptible flickering of the light in the field. Conse- 
quently a speed of alternation can be chosen to get rid of the color 
sense, and then a setting for equal brightness can be made by finding 
the point at which the flicker disappears or reaches a minimum. 
Whether the “brightness” thus found is exactly the same thing as 
that found by the ordinary photometer is a disputed question. As 
we have already seen, however, brightness must be rather arbitrarily 
defined in either case, and it appears that a suitable choice of con- 
ditions will give a very close agreement between the two methods. 
The results obtained by the flicker photometer are not free from 
variation with conditions such as field size, absolute brightness, and 
individual characteristics of the observer. The advantages of the 
instrument are simply that definite results can be obtained on any 
color difference by any observer and that the results at various times 
and on related color differences are far more consistent than when 
‘measurements are made by the ordinary photometer. This consis- 
tency makes it possible to establish the characteristics of an observer 
and to predict rather accurately what results he will obtain in com- 
parison with an average observer. As an example of the possibilities 
in this kind of procedure, the Bureau of Standards has for a consider- 
able time made use of a simple test originally due to Ives which 
indicates the relative values an observer may be expected to obtain 
when comparing lights of certain types. The test measurement 
is the determination of the transmission of two solutions, one of 
which transmits the blue end of the spectrum while the other transmits 
the yellow end. 
