Sheep are commonly used as an animal model for the human lumbar spine, but similarities in trunk muscle activity of humans and sheep during functional tasks, such as locomotion have not been investigated. Therefore, the aim of the study was to evaluate trunk and pelvic limb muscle activity during walk and run/trot gaits in man and sheep. Electromyography of the muscles erector spinae (ES), gluteus maximus (GM), rectus abdominis (RA), obliquus externus (OE) and obliquus internus (OI) were collected in 24 humans and 15 sheep during treadmill walk and run/trot. Kinematic data from the tarsus (human) or metatarsus (sheep) were obtained to define motion cycles and determine stride characteristics. Mean and range of normalised muscle activity were calculated. In phasic muscles, the occurrence of the maximum was reported. At walk, mean activity was greater in humans for all three abdominal muscles (all P<0.01). At the run/trot, mean activity of ES was significantly greater in sheep (P<0.05) and mean activity of right OI was greater in humans (P=0.016). At the walk, range of ES activity was significantly greater in humans compared to sheep (P<0.01), but significantly smaller in humans in RA and right OE (P<0.05). At the run/trot, range of activity was significantly greater in humans compared to sheep in all muscles (P<0.05), except right RA and OI. Compared to humans, occurrence of maximum activity was earlier in sheep for ES right during walk (P=0.005), and later for GM during walk and run/trot (P<0.001). The results suggest that numerous differences in trunk muscle activity exist between man and sheep during treadmill walk and run/trot, and that these differences are muscle-and gait-specific. Trunk muscle activity should therefore be regarded as species-specific which suggests differences in stabilisation strategies. This should be taken into consideration when extrapolating animal model findings to the human spine.