Is attention is automatically drawn to figures? Some evidence suggests “yes.” Nelson and Palmer (2007) reported faster responses for targets shown on the familiar vs. complementary side of bipartite, equal area displays, although targets were equiprobable on both sides. Familiarity is a figural cue; hence they concluded that attention is automatically drawn to figures. We investigated whether similar results were obtained with the figural cue of convexity (Mojica et al., VSS2009). In 3 experiments we found no evidence that attention is automatically drawn to convexity, or to convex regions perceived as figures: Subjects were equally fast to discriminate targets presented on the convex side of a bipartite display, p > 0.99. Our displays were smaller than Nelson & Palmer's (3°× 6° vs. 18° × 20°). Were their results due to familiarity, per se, or to the strategic allocation of attention to familiar regions when displays were large and targets far from fixation? This year, we used 16 small (5.3° × 5.5°), bipartite equal-area displays depicting a familiar shape on one side. Each display appeared twice, once upright, once inverted. Participant's primary task was to report whether a target shown equiprobably on each side was an “X” or a “Y.” For upright and inverted displays, response latency was not different for targets appearing on the familiar versus the complementary side, all ps > 0.11. Participants also reported figure-ground perception. Subjects perceived the familiar region as the figure more often than chance in the upright displays, p < 0.0001, yet we found no evidence that attention was automatically drawn to the figural side. We suggest that the previous results may have been due to the strategic allocation of attention when targets are located far from fixation. Other evidence that attention is automatically allocated to figures may also reflect strategic attentional deployment.