1. Seeing-frequency functions (ψ[S]) determined uniocularly for small (1.6'), brief images at the thoroughly dark-adapted human central fovea take the form of log-Gaussian integrals, for intensity or for exposure-time as the independent variable. They cannot be of Poisson type, since mean (τ') and standard deviation (σ) are quite independently modifiable.
2. The properties of τ', the abscissa of inflection in the log-Gaussian integrals, and of σ, are discussed in relation to customary determinations of spectral visibilities; agreements and certain necessary divergences are considered.
3. The values of σlogΔIlogΔI0, are distributed periodically, with 5 maxima in the range λ383 to λ712. Outside the fovea this is not the case. These maxima, and the intervening minima, are definitely correlated with "irregularities" in the values of τ' as function of λ, with data on color threshold intensities, with data on "saturation," and quite specifically with minima and maxima in the data of λ discrimination. They are unmistakably of analytical significance.
4. The interpretation of these properties is taken to require (consistent with the indications of other evidence) that under specified conditions the magnitude of σ is a measure of the "size" of the population of excitable elements of neural effect, in the range of 0 to 100 per cent response-frequency as a function of wave number. This means that under the procedure used there is detectable the spectral distribution of local, specific, photosensitization. The σ's refer to a property of the excitation process which is not affected by the action of irrelevant absorbers.
5. A preliminary test of this has been made by the use of mixed "monochromatic" lights. The total energy is empirically the significant quantity, although the separate effects of λ's are sharply exhibited.
6. The λ distribution of photosensitization indices, σlogΔIlogΔI0, is necessarily taken to have the nature of "action spectra." Careful search discloses, surprisingly enough, precise and apparently unique correspondence with the maxima and minima (and details of shape) required for the participation of: cytochrome-c reductase; the reduced forms of cytochromes-a and -c; and the often overlooked ferric-prophyrin bands of oxidized cytochrome-c.
7. For the central fovea, in which blood vessels are absent, and in which resistance to anoxia is at a maximum, the presence of an enzyme system of the sort here indicated is entirely reasonable. It is pointed out that a number of specific experiments can be based upon the general finding, whatever the causal mechanism for the sensitized λ regions may in the end prove to be. The data cannot be reconciled to the demands of existing color primary theories, although apparently quite accurately consistent with varied observational data.