SUPPLEMENTARY OBSERVATIONS

In order to include a greater number of observers in the acuity
test, the investigation, which was necessarily discontinued in June,
1910, was resumed in May, 1911. The procedure was the same as
in the earlier experiments, except that a change was made in the
source of the colored illumination. In the course of the work with
the colored liquids it was found that these were subject to continual
and irregular changes in density, due to evaporation and heating
from the glower during the observations. A satisfactory substitute
for the green liquid was found in the photographic color filter pre¬
pared by the Cramer Dry Plate Company, of St. Louis, Mo. This
filter transmits a band in the spectrum lying between the wave¬
lengths of 490 and 560/jijli.

To secure red illumination a piece of ruby glass was substituted
for the poncean red solution. This yielded a pure red with approxi¬
mately the same limits of wave-length as the liquid. Observations
with blue illumination were discontinued, inasmuch as the work
previously done indicated that so far as acuity was concerned the
blue illumination might be regarded merely as a modification of the
green.

In view of the fact that the general direction of the acuity curve
was determined with sufficient definiteness in previous observations,
it was not considered necessary in this series to use more than four
variations of intensity with the colored lights, two points being
located in the rapidly ascending portion of the curve, and the other
two in that portion of the curve in which large increases in intensity
are attended with but slight variations in acuity. For the uncolored
illumination only two degrees of intensity were used, as these were
sufficient to afford a satisfactory basis of comparison.

Table XYI. gives the results of these additional acuity observa¬
tions for the several individuals, and Table XVII. gives the same
values reduced to equal intensities, for purposes of comparison.
The values given in the latter table were gotten by plotting the
curves (Figs. 10, 11, 12, 13, 14) from Table XVI., and measuring
the ordinates for the given abscissas, after smoothing out the curves.
The figures given are, therefore, only approximate.

31