Laterality can be observed as side biases in locomotory behaviour which, in the horse, manifest inter alia as forelimb preferences, most notably in the gallop. The current study investigated possible leading-leg preferences at the population and individual level in Thoroughbred racehorses (n = 2095) making halt-to-gallop transitions. Videos of flat races in the UK (n = 350) were studied to record, for each horse, the lead-leg preference of the initial stride into gallop from the starting stalls. Races from clockwise (C) and anti-clockwise (AC) tracks were chosen alternately at random to ensure equal representation. Course direction, horse age and sex, position relative to the inside rail and finishing position were also noted. On C courses, the left/right ratio was 1.15, which represents a significant bias to the left (z = –2.29, p = 0.022), while on AC courses it was 0.92 (z = 0.51, p = 0.610). In both course directions, there was no significant difference between winning horses that led with the left leading leg versus the right (C courses, z = –1.32, p = 0.19 and AC courses, z = –0.74, p = 0.46). Of the 2,095 horses studied 51.26% led with their L fore and 48.74% with their R, with no statistically significant difference (z = -1.16, p = 0.25). Therefore, there was no evidence of a population level motor laterality. Additionally, 22 male and 22 female horses were randomly chosen for repeated measures of leading leg preference. A laterality index was calculated for each of the 44 horses studied using the repeated measures: 22 exhibited right laterality (of which two were statistically significant) and 21 exhibited left laterality (eight being statistically significant); one horse was ambilateral. Using these data, left lateralized horses were more strongly lateralized on an individual level than the right lateralized horses (t = 2.28, p = 0.03, DF = 34) and mares were more left lateralized than males (t = 2.4, p = 0.03, DF = 19).