Thirty-five patients with known cerebral lesions complained of recently acquired blurred vision. None of them had evident oculomotor or pupillary abnormalities, and each had intact central fields and normal visual acuity by conventional standards. Examination of spatial contrast sensitivity was carried out with sinusoidal grating patterns, by determining the minimum contrast between alternate light and dark bars required to distinguish the pattern from a homogeneous field. This was done at each of several spatial frequencies which were established by varying the width of the bars and hence the coarseness of the pattern. The contrast sensitivities were plotted as "visuograms" which, by analogy to audiograms, record the sensitivities in comparison to normal standards. Of the 35 patients, most showed significant losses, amounting to greater than 50 per cent elevation of contrast thresholds. Eighteen showed high frequency losses; 11 had uniform reductions over the entire visible spatial frequency range and 6 had selective frequency losses in the intermediate frequency ranges. These defects in spatial contrast sensitivity, which were not predictable from standard acuity scores, indicate that the visual symptoms in our patients may have been caused by damage to frequency-selective neural elements in the central visual systems. The method may be used to advantage in clinical investigations as well as in physiological investigations of the functional pathways subserving central vision.