In two experiments we investigated visual functioning of visually impaired persons at different illumination levels. Often, substantial improvements in visual performance can occur at light levels where normal subjects show no or only very little improvement. Although there is reasonable agreement on the preference for a certain light level and the improvement in visual performance that can be obtained by adapted lighting, not all subjects adjust light levels to such an extent that they achieve maximal performance. Subjective measures, therefore, do not always suffice for optimizing illumination. We further found, that subjects appear to choose preferred levels in such a way that they optimize their vision of detail, that optimal light levels for near and distance vision should be determined separately, and that the light tolerance of visually impaired subjects is often reduced.