Two effects of context on the presence/absence of connecting segments.

Stimulus A consisted of two proximal brackets with colinear lines separated by small gaps. Stimulus B was a square-like rectangle produced by inserting connecting segments into the gaps. Stimulus A was frequently represented as B (i.e., as closed). Following Pomerantz and Pristach (1989) and Treisman and Paterson (1984), perhaps A produced a closure emergent feature that was salient but otherwise independent of other features. However, adding the same two vertical lines (context) to both A and B produced a contrasting division outcome, even though the putative closure emergent feature was an element of the A+context stimulus, and even though it matched a physical feature of the B+context stimulus. Therefore, this emergent feature did not produce the closure. Two additional experiments indicated that the two context lines made the two connecting segments more visible--a context-produced increase in visibility occurred that is not comparable to other known evidence of perceptual improvement. This greater visibility also indicates that the division was not due to the two context lines inhibiting the perception of the two connecting segments, and the closure was not due to good continuation.