Abstract
1. The effect of spatial uncertainty on line orientation thresholds was studied in normal observers. Vertical lines, 5 min of arc long, built up a matrix in which one of the line elements could be tilted to the left or right. The orientation thresholds depended strongly on the number of alternative test positions. There was a linear relation between log (threshold) and log (P), where P is the probability that a particular line element was the one being tested. 2. The uncertainty effect was shown to be time dependent. The effect was more marked for the shortest stimulus duration (1 s). However, even with a 6 s stimulus duration, allowing several re-fixations, the thresholds were significantly higher in the presence of uncertainty, compared to the situation in which the test position was fixed and known to the observer. 3. When the measurements were restricted to the centre line in a matrix, thresholds were more than twice as high when the test line could be in any of the centre 3 x 3 positions, compared to the case in which there was no uncertainty as to the test position. Foreknowledge of location of the test line within the matrix improved the threshold further, even if the whole matrix was displaced to different retinal positions. 4. It is concluded that the physiological mechanism mediating threshold improvement probably operates on a cortical processing apparatus more central than V1.