Experiments that compare monocular and binocular visual performance of human psychophysical Os on a variety of visual tasks are reviewed, The review attempts to include all experiments published in English in this century, excluding work on stereopsis, rivalry, and evoked potentials, The concept of probability summation as a baseline for assessing the presence of neural summation is discussed, and the assumptions of several models for estimating probability summation are considered. Experiments are classified in terms of visual task, major categories being increment detection, flicker fusion, brightness magnitude, and contour resolution. A major conclusion is that binocular performance is superior for essentially all task categories and in most cases by a magnitude greater than that predicted by appropriate probability summation models.The cardinal feature of the binocular visual system in man is the unification of the two separate monocular views into a binocular cyclopean view that betrays little trace of its monocular origins. Relatively little is known about the mechanism responsible for unification, although an examination of the physiology of the visual system reveals a number of features that seem expressly designed to promote unification. Some of the more prominent features are the partial crossing of the optic tracts at the optic chiasma and the elaborate system for the maintenance of conjugate eye movements. The existence of these features in man, and in binocular vertebrates in general, suggests an adaptive function of binocularity of sufficient magnitude to justify the survival of this complicated physiological organization. Yet the immediate advantage of having two eyes functioning as one is not obvious. The simple experiment of closing one eye does not result in any dramatic change in the appearance of the visual world. Rather, only by more stringent experimentation are the subtle advantages of binocularly joined eyes revealed.The purpose of this paper is to review those experiments undertaken in the 20th century that have compared monocular and binocular performance on a variety of visual tasks. The rubric under which such experiments are typically grouped is "binocular summation." These experiments compare the relative performance on some visual task (such as detection or form resolution) by a human S using one and both eyes. In the literature, this kind of inquiry into binocular vision is treated separately from investigations of stereopsis and of binocular rivalry; following that distinction, the voluminous research on these binocular phenomena is not included. A further exclusion is data