Abstract: Normal speed videography was used to determine the angular parameters of 28 Spanish Thoroughbreds at trot. Horses were divided into 3 groups: Group UT, comprising 9 animals (provided by the VII National Stud, Cordoba, Spain) which had undergone no specific training programme and which were hand led at the trot; Group T, formed by 19 horses considered to be highly bred and trained, and which were also hand led; and Group RT, comprising the same horses as the latter group but this time trotted by a rider. Each animal was filmed 6 times from the right-hand side, using a Hi8 (25 Hz) video camera. Angular parameters for fore- and hindlimb joints were measured in each stride from computer-grabbed frames and entered into a spreadsheet for calculation; parameters included maximum and minimum angles, range of motion, and angles at landing, lift off and maximum hoof height; the times at which maximum angle, minimum angle, lift off and maximum hoof height occurred were calculated as percentages of total stride duration. Stride velocity (mean [s.d.]) was 4.01 (0.62), 3.60 (0.34) and 3.07 (0.36) m/s for Groups UT, T and RT, respectively. Data were then compared between Groups UT-T and Groups T-RT. Compared with Group UT, horses from Group T featured a shorter stance percentage (P<0.001) in both fore- and hindlimbs. The range of motion in forelimbs was smaller (P<0.05), due to lower retraction (P<0.001); moreover, maximum retraction appeared earlier (P<0.05). Greater scapular inclination was in evidence (P<0.05) and the shoulder joint extended further (P<0.05). Fore- and hind fetlock joints revealed a relatively shorter hyperextension period during the stance phase (P<0.01). Compared with Group T, horses from Group RT had a longer stance percentage, with belated maximum retraction of the fore- and hindlimbs. The range of movement in scapular inclination was greater (P<0.05), due to a smaller minimum angle (P<0.01), and the shoulder joint flexed more (P<0.05). The elbow joint extended more and for longer during the stance phase. Initial extension of the hip joint (P<0.05) and tarsus (P<0.001) lasted longer. The carpal and fore and hind fetlock joints recorded relatively longer hyperextension times, in addition to greater hyperextension during the stance phase. The results from the present study suggest that rider-effect must be taken in consideration when well gaited horses are selected for dressage purposes.