As there is no statistical evidence that saddle fit influences the load exerted on a horse"s back, this study was performed to assess the hypothesis that the width of the tree significantly alters the pressure distribution on the back beneath the saddle. Nineteen sound horses were ridden at walk and trot on a treadmill with three saddles differing only in tree width. Kinetic data were recorded by a sensor mat. A minimum of 14 motion cycles were used in each trial. The saddles were classified into four groups depending on fit. For each horse, the saddle with the lowest overall force (LOF) was determined. Saddles were classified as "too-narrow" if they were one size (2 cm) narrower than the LOF saddle, and "too-wide" if they were one size (2 cm) wider than the LOF saddle. Saddles two sizes wider than LOF saddles were classified as "very-wide". In the group of narrow saddles, the pressure in the caudal third (walk 0.63 N/cm(2)+/-0.10; trot 1.08 N/cm(2)+/-0.26) was significantly higher compared to the LOF saddles (walk 0.50 N/cm(2)+/-0.09; trot 0.86 N/cm(2)+/-0.28). In the middle transversal third, the pressure of the wide saddles (walk 0.73 N/cm(2)+/-0.06; trot 1.52 N/cm(2)+/-0.19) and very-wide saddles (walk 0.77 N/cm(2)+/-0.06; trot 1.57 N/cm(2)+/-0.19) was significantly higher compared to LOF saddles (walk 0.65 N/cm(2)+/-0.10/ 0.63 N/cm(2)+/-0.11; trot 1.33 N/cm(2)+/-0.22/1.27 N/cm(2)+/-0.20). This study demonstrates that the load under poorly fitting saddles is distributed over a smaller area than under properly fitting saddles, leading to potentially harmful pressures peaks.