The role of intraocular stray light upon contrast threshold of two skilled human observers was studied by perimetric methods using the automatic perimeter Octopus. Stray light falsifies the contrast sensitivity profile of the blind spot when a critical test stimulus luminance, which differs for various targets, is exceeded. At still higher luminance levels, because stray light effects increase, the blind spot shrinks and finally disappears. A series of high resolution measurements of the blind spot with the automatic perimeter Octopus provide a quantitative answer concerning the commencement and amount of this disturbance as a function of target size and target luminanee. The amount of stray light, when using targets varying from 0 to 5 (Goldmann standard) is related to luminous power (target luminance x target area). Using first order assumptions about the stray light emitting characteristics of the disc and empirical data, one may conclude that an increase in luminance of target 0 from 10 3 to 10 5 asb only increases the effective dynamic range by about 2 dB (= 0.2 log u nits 3, the standard target as compared with target) size used in Octopus perimetry, at a luminance level of 10 s ash. Falsification of sensitivity gradients and underestimates of depths of scotomata due to stray light effects may be an ever present danger in perimetric determinations. The useful dynamic range in perimetry appears to be limited by photon noise and noise in the neurovisual system on the one hand and by stray light interference on the other.