Yo and Wilson (1992 Vision Research 32 135 – 147) reported that Type II plaids move in the vector sum (VS) direction at short durations. Bowns (Vision Research in press) showed that the result did not generalise to all Type II plaids. Computational analysis of the stimuli revealed an alternative explanation of the result in terms of features that were shown to move in a similar direction to the VS. When features (or VS) moved in a different direction to the intersection of constraints (IOC) the plaids were often perceived to move in the direction of the feature. It is hypothesised that a feature will contribute only when there is sufficient projection of the IOC in the feature direction. The direction of the features and the IOC for the stimuli used by Bowns was computed for a set of plaids that shifted from being perceived in the feature direction to the IOC. If projection is critical, then it should decrease as the stimuli shift direction. This was confirmed. A different set of plaids comprising a feature that moved in the opposite direction to the IOC and varied in terms of the magnitude of projection was presented to subjects in a forced-choice task. The plaids moved as predicted from the hypothesis. These results show a correlation between the amount of projection of the IOC in the direction of a feature and the incidence of perceived movement in the direction of the feature. The hypothesis that perceived movement of a plaid is influenced by the degree of projection of the IOC onto another available motion signal such as a displaced feature is supported.
