Contradictory evidence has been presented in the literature on the existence of a right-hemisphere mechanism subserving global stereopsis. In the present study evoked responses to dynamic random-dot stereograms were recorded from International 10-20 scalp sites O2, P4, T6, T4 and homotopic sites over the left hemisphere. Stroke patients with unilateral lesions were selected on the basis of satisfactory performance on ocular screening tests and on clinical or objective indications that dementia was absent. Control subjects were found among the relatives of patients. The stereopsis test schedule included left, right and centre field presentations of 30 arc min, disparities. The dot density of the disparity, relative to that of the remainder of the random dot field, was manipulated to provide a full contrast stimulus and a no-contrast uniform field stimulus, thus varying the availability of monocular cues provided by the contrast factor. Latency and amplitude of the evoked potentials were computed and submitted to analyses of variance. No significant results were found which supported the proposition of a lateralized mechanism for global stereopsis. This conclusion is in conflict with the findings of some earlier reports. It is suggested that the techniques used in the earlier experiments may not have been adequate for the study of the questions involved.