In three experiments, the distribution and malleability of infant visual attention were studied in 5-month-olds (N= 72) while they inspected large geometric designs. In Experiment 1, we established that infants maintained their distribution of attention from a pretest to a familiarization phase. We also replicated and extended our previous findings that infants who examined targets with briefer, more numerous looks and shifts —short lookers —had novelty scores above chance, whereas long lookers demonstrated chance responding. In Experiment 2, different portions of the display were successively illuminated with red light. This manipulation induced long lookers to scan like short lookers during familiarization; they then showed novelty scores well above chance. A third experiment ruled out the simple presence of a red light as the source of this effect. In sum, then, these results suggest that the distribution of attention is malleable, and that a broader distribution of attention, as reflected in briefer and more numerous looks and shifts, can improve processing.