Mistracking in alignment illusions.

A number of anomalies have been pointed out in opposition to attempts to account for the Poggendorff illusion in terms of the expansion of small or medium-sized acute angles. A principle under which these anomalies could be subsumed was proposed and subjected to experimental testing using simplified displays. The following theory emerged from these experiments. If a straight line (the pointer) is aligned with a do (the target) situated at the end of a second line (the induction line) and lying in a different orientation from the pointer, then (depending on the angle between pointer and induction line) the pointer will appear to be aligned with a point further along the induction line. The amount of misalignment varies inversely with increase in size, up to about 135 degrees, of the angle formed by the pointer with the induction line, after which there is no displacement attributable to the induction line. The effect appears not to be due to neural interaction. An explanation in terms of eye movement is discussed.