From a consideration of these phenomena we were led to discuss the functions of the rods and cones in the retina of the eye, and the suggestion was made that differences in color-vision were due to central rather than retinal processes, and that in many cases of partial or total color-blindness the retina would be found normal and the defect in vision due to a lesion in some more central structure. The various forms of visual irradiation which have been described by a number of different writers we found to be all forms of one rather simple process. Resistance, removal of resistance upon further stimulation, and recurrence of resistance through fatigue in some part of the optic tract, together with the spreading of stimulation over the retina (probably through the molecular layers) from one afferent nerve to another are assumed as the minimal requirements which are sufficient to explain the five forms of irradiation which have been considered in this paper. FOOTNOTES: [Footnote 23: Image no. 5 appeared with the concavity in front. In the centre of the figure appeared a dark grayish splotch of light, very much darker than the rest of the image. This is due, most probably, to the presence of Charpentier's phenomenon of recurrent bands. If this happens in this figure the ends of the recurrent bright images would overlap while the centres would not, so that the black bands appearing, as it were, through interstices in the central part of the figure, would seem like a dark splotch, especially since the outlines of the bands are vague and hazy. The back end of the figure had the effect of being vertical or nearly so. This is probably due to the same cause as that which made the circular figure to appear as it did, namely, the negative after-image overlapping the positive after-image. The front of this black image is usually of about the same shape as the front of the real figure which it follows. If this is so, then, in this case, it would make the back part of the image pretty nearly vertical.] [Footnote 24: The intensity of the objective stimulation may be represented by the line _AB_. If there were no reënforcement of stimulation the whole figure would be flat on top and of this height. The difference between _AB_ and _AC_, or _BC_, represents the increase of intensity due to irradiation at the most favorably situated portion of the figure. The other portions receive increments proportional to their location, as indicated in the diagram.] [Footnote 25: Favorable localization will of course depend largely on the shape of the figure in which the point is situated. Thus one in the angle of a triangle or at the horn of a crescent would have much less reënforcement of excitation than another point, say halfway down the side.]