It is well recognized, particularly from the work of Hecht (1937), that measurements of visual function at different levels of illumination generally fall into two categories, a low intensity scotopic portion and a high intensity photopic portion. Thus in the time course of dark-adaptation, and in the curves where log. increment threshold, log. acuity or fficker fusion frequency is plotted against log. field intensity, the results are found to fall upon two distinct branches. The low-intensity branch has the spectral sensitivity of rhodopsin, and it is not seen in measurements made upon the fovea centralis; it is therefore considered to be due to rod function. The other branch has a-different spectral sensitivity, it is found upon the rod-free fovea and hence it is considered to be due to cone function. Now though there are considerable difficulties in distinguishing the functions of the various types of cone, it is easy to separate cone curves from rod curves over the whole intensity range of the cones. For not only may measurements be made upon the fovea where rods are absent, but even elsewhere it turns out that cones have an increment sensitivity and fusion frequency so much higher than rods that the plotted curves generally show the cone branch substantially in its entirety. It is otherwise with rod curves, for not only is there no part of the human retina which contains rods without cones, but the very features which make cone measurements outstanding in a mixed population make rod measurements unavailable. In fact, we can only measure rods at intensities below the cone threshold. This does not mean that rods are necessarily inactive at higher intensities, but their activity cannot be measured by ordinary threshold procedures, and we are left wondering what course the rod branch may take after it disappears behind the cone branch.This question could be answered if we were able to investigate a subject who had no cones but whose rods and rod pathways to the brain were * Instituto Neurologico, Via Celoria 11, Milano. t Present address: Physiological Laboratory, University of Cambridge.12-2