Theory and Measurement of Visual Mechanisms

Flicker response contours (F vs. log I(m)) for a square image subtending 0.602 degrees on a side, located in the fovea, are simplex probability integrals for a "white" and for four (five) spectral regions filtered from this white, and with different light-time fractions in the flash cycle. The subjective phenomena (the appearance of the field, the intensity threshold for color, and others) at the fusion points along these contours parallel in a variety of ways those obtained on duplex flicker contours resulting from the use of larger or eccentrically placed flickered images. These phenomena therefore cannot be held to indicate involvements of "rod" excitation. The scatter of the index of variation of I(1) is such as to demonstrate the full participation of all the potentially excitable neural units at all levels of flash frequency, for each kind of light. The magnitude of this scatter, a measure of neural integration in visual performance, is a function of the number of these units (with F(max.) nearly constant); the two quantities vary together when wave-length composition of light is altered. The properties of the contours for a white light and for the spectral regions filtered from it show that, for the image within the fovea, different numbers of units are excitable in flicker recognition according to the wave-length band used, and different mean frequencies of elements of effect under fixed conditions. The changes in the mean intensity for activation of these units as a function of the light-time fraction in the flash cycle are correlated with the numbers of these units; when this is corrected for, it is pointed out that despite the differences in shape of F vs. log I it cannot be concluded that the mechanism of excitation differs for different wave-lengths. It is indicated that "white" must be regarded as a synthesis, not a mere summation, of effects due to different spectral regions. Certain differences are pointed to as between foveal and more peripheral regions tested, and as between observers differing in the degree of the "yellow spot effect," with regard to the relative effects of wave-length and of image area. A general consequence is the outlining of conditions required for the precise comparison of excitabilities as a function of wave-length in the multivariate visual system.