PURPOSE. Visual pop-out is the phenomenon in which a unique item is rapidly detected from among a set of homogeneous distractors. Adult studies have attributed the phenomenon of visual pop-out to the functioning of early visual processing that registers all items in an array in parallel. Previous studies with infants have suggested that visual pop-out can be observed in them as well (e.g. Adler, Inslicht, Rovee-Collier, & Gerhardstein, 1998). However, these previous studies have observed pop-out occurring on the order of several seconds to minutes whereas in adults pop-out is observed in milliseconds. It is therefore possible that the pop-out observed in infants is not due to early visual processing. Further, in adults, pop-out is unaffected by the number of homogeneous distractors in the array (set-size effect), an effect that has not been examined in infants. The present study examined whether the visual pop-out observed in infants is similar to that observed in adults. METHODS. Three-month-olds' eye movement latencies were assessed as they viewed either a feature-present, pop-out array consisting of a single unique + among distractor L's or a feature-absent array consisting of all L’s. To determine the effect of array set size on infants' eye movement latencies, both feature-present and feature-absent arrays consisted of either 1, 3, 5, or 8 items. RESULTS. The unique + popped out such that eye movement latency was unaffected by the number of distractors in the feature-present array, but increased with set size in the feature-absent array. Further, eye movement latency to the + was on the order of milliseconds rather than seconds. CONCLUSIONS. These results suggest that the same early visual processing mechanism responsible for pop-out in adults is functional in very young infants. Thus, parallel processing of items in an array may be a fundamental mechanism of selective attention and visual processing, even in immature systems.