In 2 experiments the authors demonstrated that adaptive locomotion can involve means-ends problem solving. Sixteen-month-old toddlers crossed bridges of varying widths in the presence or absence of a handrail. Babies attempted wider bridges more often than narrow ones, and attempts on narrow bridges depended on handrail presence. Toddlers had longer latencies, examined the bridge and handrail more closely, and modified their gait when bridges were narrow and/or the handrail was unavailable. Infants who explored the bridge and handrail before stepping onto the bridge and devised alternative bridge-crossing strategies were more likely to cross successfully. Results challenge traditional conceptualizations of tools: Babies used the handrail as a means for augmenting balance and for carrying out an otherwise impossible goal-directed task.