This research examined the development of inhibition in a locomotor context. In a within-subjects design, infants received high- and low-demand locomotor A-not-B tasks. In Experiment 1, walking 13-month-old infants followed an indirect path to a goal. In a control condition, infants took a direct route. In Experiment 2, crawling and walking 13-month-old infants crawled through a tunnel to reach a goal at the other end and received the same control condition as in Experiment 1. In both experiments, perseverative errors occurred more often in the high-demand condition than in the low-demand condition. Moreover, in Experiment 2, walkers perseverated more than crawlers, and extent of perseveration was related to infants' locomotor experience. In Experiment 3, the authors addressed a possible confound in Experiment 2 between locomotor expertise and locomotor posture. Novice crawlers perseverated in the difficult tunnels condition, behaving more like novice walkers than expert crawlers. As predicted by a cognitive capacity account of infant perseveration, overtaxed attentional resources resulted in a cognition-action trade-off. Experts who found the task less motorically effortful than novices had more cognitive resources available for problem solving. (Contains 5 figures.)