This study determined whether walking at the preferred stride frequency minimizes muscle activity compared with other cadences at the same speed. Anthropometric measurements were recorded from 10 subjects and used to estimate their predicted resonant stride frequency. The preferred walking speed and stride frequency were determined from freely adopted walking on a treadmill. For the experimental trials the treadmill was set at each individual's preferred walking speed. Participants walked for 6 min at eight cadences prescribed by an auditory metronome: preferred stride frequency and -35, -25, -15, 0, +15, +25, +35% of predicted resonant stride frequency. Oxygen consumption was measured via gas analysis. Muscle activity of the right leg gastrocnemius (GA), tibialis anterior (TA), biceps femoris (BF) and rectus femoris (RF) muscles was recorded via electromyography (EMG). On average, participants preferred to walk with a stride frequency .07 Hz lower than their predicted resonant stride frequency, however a strong positive correlation was observed between these variables. Stride frequency had a significant and large quadratic effect on VO2 (RLR(2)=.76), and activity of the GA (RLR(2)=.66), TA (RLR(2)=.83), BF (RLR(2)=.70) and RF (RLR(2)=.78) muscles. VO2, GA and TA activity were all minimal at the preferred stride frequency and increased for faster or slower cadences. BF and RF activity were minimal across a broad range of slow frequencies including the preferred stride frequency and increased for faster frequencies. The preferred stride frequency that humans readily adopt during walking minimizes the activation of the GA, TA, BF and RF muscles, which in turn minimizes the overall metabolic cost.