Experimental results concerning the behavior of the contrast threshold are given as obtained from about 100,000 responses of the same subject.
The circular test spot of size d between 6.6′ and 8° was projected 7° nasal from the fovea. Green light was used for background and test spot. The stimulus exposure t was varied between 0.02 and 1 second.
So long as d and t are small, the behavior of ΔB/B agrees with the explanation of de Vries: ΔB~B½. The dependence of the sensitivity of the eye on B is calculated from the precise shape of the curve for ΔB as a function of B for these combinations of d and t. The result does not contradict direct measurements of the sensitivity.
For combinations of d and t values in which d or t are large, deviations from quanta explanations similar to those published in previous papers on the absolute threshold (J. Opt. Soc. Am. 37, 908 (1947)) are found.
It appeared that these deviations are due to the fact that ΔB can never decrease below a limiting value depending on B. In a region of the B values studied, this limit for ΔB is proportional to B, so that ΔB/B is constant for large d and t values in this region.
The duplicity of the retina becomes apparent in the curves representing the sensitivity of the eye as a function of B and in the behavior of the limiting value of ΔB mentioned.
M. A. BOUMAN, "Peripheral Contrast Thresholds of the Human Eye," J. Opt. Soc. Am. 40, 825-832 (1950)