Object-based effects of visual attention were reported by Egly Driver and Rafal (1994) using a modified Posner cueing paradigm. In a two-rectangle (i.e., “object”) display, they cued one end of one rectangle on each trial. Targets were most likely to appear at the cued location. In invalid-within conditions, the target appeared in the same object as the cue, but at the opposite end. In invalid-between conditions, the target appeared in the object opposite the cue. Importantly, invalid targets appeared the same distance away from the cue in both within- and between- conditions, holding space constant. When comparing the two invalid conditions, within-object targets were detected faster than between-object targets, i.e., RTs showed object-based facilitation. In a set of similar experiments, we replicated these effects when the objects were white but, surprisingly, not when they were black. To investigate the disappearance of the object-based effect, the present study attempted to manipulate the rectangles so that they were either perceived as “holes” or as objects whether they were black or white. The logic was that black may more likely be perceived as background and thus processed as “holes”. The data partially supported our hypothesis: we found a trend for an interaction between object-based effects and the holes/no-holes manipulation. However, debriefing questions suggested that not all participants perceived the stimuli as expected. To further investigate this issue, we manipulate stereoscopic depth cues (using stereographic goggles) to test the same hypothesis: that the perceptual organization of objects as “holes” or “wholes” can determine the presence of object-based attention effects.