Some aspects of backward locomotion are similar to forward locomotion, while other aspects are not related to their forward counterpart. The backward preferred transition speed (BPTS) has never been directly compared to the energetically optimal transition speed (EOTS), nor has it been compared to the preferred transition speed (PTS) during forward locomotion. The purpose of this study was to determine whether the BPTS occurs at the EOTS, and to examine the relationship between the backward and forward preferred gait transition speeds. The preferred backward and forward transition speeds of 12 healthy, young subjects (7 males, 5 females) were determined after subjects were familiarized with forward and backward treadmill locomotion. On a subsequent day, subjects walked backward at speeds of 70, 80, 90, 100, and 110% of the BPTS and ran backward at speeds of 60, 75, 90, 100, and 120% of the BPTS while VO₂ and RPE data were collected. After subtracting standing VO₂, exercise VO₂ was normalized to body mass and speed. For each subject, energy-speed curves for walking and running were fit to the normalized data points. The intersection of these curves was defined as the EOTS which was compared to the BPTS using a paired t-test (p < 0.05). RPE and VO₂ at the BPTS were also compared between walking and running conditions, and the correlation between BPTS and PTS was calculated. The EOTS (1.85 ± 0.09 m·s^-1) was significantly greater than the BPTS (1.63 ± 0.11 m·s^-1). Even though RPE was equal for walking and running at the BPTS, VO₂ was significantly greater when running. There was a strong correlation (r = 0.82) between the BPTS and the PTS. Similar to forward locomotion, the determinants of the BPTS must include factors other than metabolic energy. The gait transition during backward locomotion exhibits several similarities to its forward counterpart.