Subjects were presented with the printed word above or below and a configuration consisting of a reference shape (a square) and a smaller shape of variable location (a blacked-in circle) which might appear above or below the square. The instruction was to report ‘yes’ if the word and the location of the circle relative to the square were congruent, and ‘no’ if they were not. The latency of this judgement was measured from the simultaneous onset of the displays to the initiation of the verbal response. The pair above/above was classified as congruent significantly more rapidly than the pair below/below or either of the mismatched pairs. The order in which word and configuration were fixated was a significant factor, response times being faster where the word was fixated first.