88 JOURNAL OF THE WASHINGTON ACADEMY OF SCIENCES VOL. 13, No. 5 
satisfactory values for these coefficients or for the ‘‘visibility’’ curve 
which represents them. Fortunately it is only a few weeks since 
Dr. Gibson presented to the Society the general results of the most 
recent work on visibility curves, and consequently I shall not burden 
you with details of the methods of determining them. While Gibson 
and Tyndall’s curves were determined by equality-of-brightness 
settings, following a step-by-step method used earlier by Hyde, 
Forsythe, and Cady, the greater number of such determinations have 
been made by use of the flicker principle. The most extensive investi- 
gation of this kind was carried out by Dr. Coblentz a few years ago, 
and was likewise reported here. 
If one plots together the curves obtained for individual observers 
in any of these investigations the result is a diffuse band representing 
the scattering of the curves of different individuals. While there is a 
general similarity in the curves of all observers, individual differences 
are such as to represent a very large percentage variation. Here, as 
in the case of the simpler test-solution measurements already men- 
tioned, any classification of observers as ‘“‘normal’’ or “abnormal” 
must be arbitrary, since a complete gradation is found between the 
most extreme types of curves. The choice of a general average or 
“normal” curve is therefore difficult. 
The results of Gibson and Tyndall and of Hyde, Forsythe, and 
Cady, both made by equality-of-brightness methods, show an agree- 
ment which is really remarkable when the difficulties of the investi- 
gation are considered. In fact it is quite possible that the differences 
found are to be explained by the fact that observing fields of different 
size and form were used in the two laboratories. 
Gibson’s comparisons of his own results with those of Coblentz also 
show a fairly close agreement, but apparently indicate a real differ- 
ence between the results of the two methods. ‘There are two con- 
ditions, however, which make this conclusion somewhat uncertain. 
These are that the energy values used were determined in entirely © 
different ways, and that several years elapsed between the two investi- 
gations so that the observers common to both may have changed 
their characteristics in the meantime. The conditions of field size 
and brightness used were also slightly different. 
My opinion is that most of the earlier determinations of visibility 
may now be given little weight, and that the most reasonable method 
of establishing experimentally a reliable normal, or average, curve}is 
to carry out such measurements as are necessary to answer the un- 
settled questions with regard to the differences between the Coblentz 
