The role of attention in rivalry remains controversial (Mitchell et al, 2004; Meng &Tong 2004; Paffen et al, 2005). To test whether rivalry can occur without attention, we used an array of 8 rivaling gratings, each flickering out of phase between eyes to create successive rivalry (at 3 to 4 Hz, O'Shea & Crassini, 1984), arranged in a circle around a central fixation point. In a control condition, observers reported 1) changes in dominance when attending to a single patch and 2) the pattern of dominance across multiple patches attended at the same time. Result: approximately equal dominance durations of the two rivaling patterns and an absence of any obvious synchrony across patches. However, moving attention rapidly from one patch to the next around the array (guided by a pointer) gave a completely different outcome. Alternation slowed or stopped entirely and one or the other pattern was strongly dominant (average of 90% of the time). If rivalry was occurring on the unattended locations, then the dominance of each location would be sampled by attention and should show relatively random alternation from location to location. This would be true even if the alternation at each patch had stopped in the absence of attention, leaving the dominance frozen on one pattern or the other. These results suggest that the mutual inhibition underlying rivalry (at least of the successive variety) does not occur in the absence of attention. The competing patterns have not just stopped alternating, they have stopped competing.