We tested the hypothesis that poor performance on the Piagetian invisible displacement task is related to increased memory requirements. Rhesus monkeys and orangutans received 3 types of problems (invisible, visible, and no transfer problems) each containing a number of steps equivalent to that of standard invisible displacements. If failure to solve invisible displacements was due to increased memory requirements, then the primates should perform at chance level on all 3 problems. However, rhesus monkeys solved visible and no transfer problems, but not invisible transfer problems. Half of the orangutans solved all 3 transfer problems, although their performance on invisible transfer problems was lower than that on the other problems. A subsequent cuing phase led to improved performance, and a few monkeys solved invisible transfer problems.