Injae Hong^1,2 (), Jeremy Wolfe^1,2; ¹Brigham & Women's Hospital, ²Harvard Medical School

In visual foraging, decisions on when to leave a patch have been studied using the Marginal Value Theorem (MVT, Charnov, 1976), which suggests leaving a patch when its instantaneous rate of return drops below the environment’s average rate. While lab studies often assume rewards in every patch, real-world scenarios include target-absent patches, such as apple trees without ripe apples but with unripe apples only. This study investigated how the presence of empty patches influenced patch-leaving decisions in target-present patches. The task was to collect as many good targets as possible in 10 minutes, moving from one visual patch to another at will. Travel time between patches was either 3 or 6 seconds, during which no targets could have been collected. Three conditions were tested: ‘Always-Present’ condition - good targets were present in every patch; ‘Half-Present’ condition - good targets were present in half the patches; remaining patches contained only bad targets; ‘Extended-travel’ condition - good targets were present in every patch, but the travel time was increased to match the average time lost in the Half-Present condition. The result shows that observers did not reject bad patches immediately and picked some bad targets, presumably to confirm that the patch was empty. Every pick in an empty patch, by definition, had a rate of return below the overall rate. Ideally, observers should have moved on from those patches before picking anything. In apparent compensation, observers tended to leave good patches earlier, while the rate of return was still high. This “underharvesting” seems to violate the MVT prediction but the overall result was broadly in line with MVT. The result seems to reflect complex, (probably implicit) reasoning about balancing behavior in bad and good patches in the Half-Present situation.

Acknowledgements: NSF-2146617