Infants search for an object hidden by an occluder in the light months later than one hidden by darkness. One explanation attributes this décalage to easier action demands in darkness versus occlusion, whereas another attributes it to easier representation demands in darkness versus occlusion. However, search tasks typically confound these two types of demands. This paper presents a search task that unconfounds them to better address these two explanations of the 'dark advantage'. Objects were hidden by submersion in liquid instead of occlusion with a screen, allowing infants to search with equally simple actions in light versus dark. In Experiment 1, 6-month-olds unexpectedly showed a dark disadvantage by discriminating when an object was hidden in the light but not the dark. Experiment 2 addressed the possibility that representation demands were higher in the dark than the light and showed that infants’ search in the dark increased to match that in the light, but not exceed it. Six-month-olds can thus search for a hidden object both when action demands are simplified and when a non-cohesive substance rather than a cohesive occluder hides the object, supporting aspects of both action-demand and representation-demand explanations of décalage in search behavior.