Experiment 4: Results.

Proportion of trials motion is perceived in a direction indicating that the target horizontal surface was grouped with the other horizontal surface. This is graphed separately when the central surface could function as an occluder (A) and when it could only function as a visible connector (B). Included in panels A and B are the proportions of trials motion was perceived in a direction indicating that the decremented horizontal surface was grouped with the vertical surface. There were too few ‘bidirectional’ responses to include in the graph. Vertical lines indicate +/- one standard error. The standard error bars are sometimes displaced so that they don’t over-lap for different central surface heights. They are omitted when smaller than the markers. To minimize clutter that would reduce legibility, standard errors are averaged over the three widths of the central surface.