Abstract: In order to objectively quantify the effect of manipulation on back-related locomotion anomalies in the horse, a recently developed kinematic measuring technique for the objective quantification of thoracolumbar motion in the horse was applied in a dressage horse that was suffering from a back problem. In this horse, clinically, a right-convex bending (scoliosis) from the 10th thoracic vertebra to the second lumbar vertebra was diagnosed. As a result, there was a marked asymmetric movement of the thoracolumbar spine. Functionally, there was severe loss of performance. Thoracolumbar motion was measured in terms of ventrodorsal flexion, lateral flexion, and axial rotation using an automated gait analysis system. Measurements were repeated before and 2 days after treatment, before the second treatment 3 weeks later, and at 4 weeks and 8 months after the second treatment to assess long-term effect. At the same time, performance of the horse was assessed subjectively by the trainer as well. Symmetry of movement improved dramatically after the first treatment. After this, there was a slight decrease in symmetry, but 8 months after the last treatment the symmetry indexes for the various joints were still considerably better than during the first (pre-treatment) measuring session. Subjectively, the trainer did not notice improvement until after measurement session 4. Between sessions 4 and 5 (at 4 weeks and 8 months after the second treatment) there was a change of trainer. The new trainer did not report any back problem, and succeeded in bringing the horse back to its former level in competition. It is concluded that manipulation had a measurable influence on the kinematics of the thoracolumbar spine. However, it is recognized that an improvement in symmetry of motion is not equivalent to clinical improvement and that other measures, such as changes in management, may be more decisive.