1. To compare the basic hindlimb synergies for backward (BWD) and forward (FWD) walking, electromyograms (EMG) were recorded from selected flexor and extensor muscles of the hip, knee, and ankle joints from four cats trained to perform both forms of walking at a moderate walking speed (0.6 m/s). For each muscle, EMG measurements included burst duration, burst latencies referenced to the time of paw contact or paw off, and integrated burst amplitudes. To relate patterns of muscle activity to various phases of the step cycle, EMG records were synchronized with kinematic data obtained by digitizing high-speed ciné film. 2. Hindlimb EMG data indicate that BWD walking in the cat was characterized by reciprocal flexor and extensor synergies similar to those for FWD walking, with flexors active during swing and extensors active during stance. Although the underlying synergies were similar, temporal parameters (burst latencies and durations) and amplitude levels for specific muscles were different for BWD and FWD walking. 3. For both directions, iliopsoas (IP) and semitendinosus (ST) were active as the hip and knee joints flexed at the onset of swing. For BWD walking, IP activity decreased early, and ST activity continued as the hip extended and the knee flexed. For FWD walking, in contrast, ST activity ceased early, and IP activity continued as the hip flexed and the knee extended. For both directions, tibialis anterior (TA) was active throughout swing as the ankle flexed and then extended. A second ST burst occurred at the end of swing for FWD walking as hip flexion and knee extension slowed for paw contact. 4. For both directions, knee extensor (vastus lateralis, VL) activity began at paw contact. Ankle extensor (lateral gastrocnemius, LG) activity began during midswing for BWD walking but just before paw contact for FWD walking. At the ankle joint, flexion during the E2 phase (yield) of stance was minimal or absent for BWD walking, and ankle extension during BWD stance was accompanied by a ramp increase in LG-EMG activity. At the knee joint, the yield was also small (or absent) for BWD walking, and increased VL-EMG amplitudes were associated with the increased range of knee extension for BWD stance. 5. Although the uniarticular hip extensor (anterior biceps femoris, ABF) was active during stance for both directions, the hip flexed during BWD stance and extended during FWD stance.(ABSTRACT TRUNCATED AT 400 WORDS)