Gait transitions during unrestrained locomotion in dogs

Gait transitions during long distance, unrestrained locomotion were studied in 22 mongrel dogs. Spatial and temporal limb movement parameters were collected and the phase relationships between limb movements based upon a 2-dimensional (2-D) gait diagram were computed. During most of the trials, the dogs trotted within a relatively narrow velocity range. Gait transitions were observed during radical changes of the movement velocity. In most cases the gait switches were abrupt and completed within 2 strides of the gait cycle. The dogs walked, depending on the animal size, within the upper velocity range of 0.93-1.21 m/s. Most of the walk-trot transitions were observed within this range. All of them had a typical pattern that involved changes of the phase shift between diagonal limb movements from 0.31 +/- 0.02 (a typical value for a walking dog) down to 0.02 +/- 0.03. These changes appeared abruptly within one stride cycle for each diagonal pair of limbs; therefore, the transition was completed in 2 strides of the gait cycle. The switch involved momentary shortening of the hindlimb amplitudes. During the next gait cycle, all limb movement amplitudes were reduced with a concomitant increase in limb movement frequencies. In contrast to the clear border between the symmetrical gaits, the dogs switched to gallop at any speed within the trot range (most frequently between 1.5-2.6 m/s). The transitions were usually completed within one stride of the diagonal limbs. In most cases, the switch from trot to gallop had a similar pattern; while maintaining synchronous movement of one diagonal pair of limbs, the other pair movement control was modified accordingly. The typical transition pattern involved the shortening of the swing phase in the front limb with simultaneous lengthening of the swing phase in the diagonal hindlimb. These transient modifications had their equivalent in the analogous limb movement amplitude changes. A mirror-image pattern of phase changes was observed in the majority (82%) of the gallop-trot transitions. Three of the dogs, however, besides the typical gallop-trot switch, occasionally employed a second pattern of transition.