Humans have been shown to falsely remember seeing the details just beyond the edges of a pictured scene. This constructive memory error is known as boundary extension. Either the traditional visual-cognitive model or the multisource model, which differ in their distinction between scene perception and representation, can explain boundary extension. Five experiments assessed boundary extension in humans (Homo sapiens), rhesus monkeys (Macaca mulatta), and capuchin monkeys (Cebus apella) using identical and equivalent delayed match-to-sample tasks. The present study replicated boundary extension in human adults. However, neither monkey species demonstrated boundary extension when viewing human-unique or monkey-relevant scenes. Unlike humans, monkeys may not have demonstrated boundary extension because they are local visual processors. This would have limited their view of the stimuli as scenes, allowing them to rely on direct visual input. This species discontinuity reflects the potentially human-unique qualities of boundary extension.