THE THRESHOLD OF VISION FOR DIFFERENT COLOURED LIGHTS. 
95 
to those obtained with the fixation spot at 2‘5 degrees from the fovea. At the 
second attempt the fixation was better but still not central. At the third attempt, 
the start in each case being made at the red end of the spectrum, the fixation was 
central down to SSN 40, after which the eye wandered slightly. During the fourth, 
fifth, and sixth sets of readings, the fixation was kept central throughout, and all the 
points obtained are in good agreement. 
On the supposition that we start with a spectrum of such an intensity that 
the illumination on the disc when the slit is at the D line is one lux (metre-candle), 
we have calculated from the annulus readings by how much the light has to 
be reduced to reach the threshold for the different parts of the spectrum. This 
.SSN 
Fig. 2. 
reduction, in conformity with the nomenclature used by one of us in previous papers, 
we shall call the extinction. The actual numbers obtained for the extinction, except 
of course at the D line, will depend on the distribution of light in the spectrum employed, 
that is on the source of light and on the dispersion curve of the prisms. In order 
to obtain numbers which do not involve 'these quantities we have determined 
the distribution of energy in the spectrum we have used, so that we can calculate 
what is the energy of the radiation falling on unit area of the disc. I, at the threshold. 
Since our measures of the energy are relative they give the comparative distribution 
of energy throughout the spectrum, but do not give the absolute values, as we have 
to fix an arbitrary unit. Since it is convenient to tabulate and plot the figures in the 
logarithms of the energy, it is advisable to avoid the use of negative characteristics, 
