Recently developed powered prostheses are capable of producing near-physiological joint torque at the knee and/or ankle joints. Based on previous studies of biological joint impedance and the mechanics of able-bodied gait, an impedance-based controller has been developed for a powered knee and ankle prosthesis that integrates knee swing initiation and powered plantar flexion in late stance with increasing ankle stiffness throughout stance. In this study, five prosthesis configuration conditions were tested to investigate the individual contributions of each sub-strategy to the overall walking mechanics of four unilateral transfemoral amputees as they completed a clinical 10-m walk test using a powered knee and ankle prosthesis. The baseline condition featured constant ankle stiffness and no swing initiation or powered plantar flexion. The four remaining conditions featured knee swing initiation alone (SI) or in combination with powered plantar flexion (SI+PF), increasing ankle stiffness (SI+IK), or both (SI+PF+IK). Self-selected walking speed did not significantly change between conditions, although subjects tended to walk the slowest in the baseline condition compared to conditions with swing initiation. The addition of powered plantar flexion resulted in significantly higher ankle power generation in late stance irrespective of ankle stiffness. The inclusion of swing initiation resulted in a significantly more flexed knee at toe off and a significantly higher average extensor knee torque following toe off. Identifying individual contributions of intrinsic control strategies to prosthesis biomechanics could help inform the refinement of impedance-based prosthesis controllers and simplify future designs of prostheses and lower-limb assistive devices alike.