OBJECTIVES: The number of locking screws required per fragment during bridging osteosynthesis has not been fully determined in the dog. The purpose of this study was to assess the survival of two constructs, with either two or three screws per fragment, under cyclic bending. METHODS: A 10-hole, 3.5-mm stainless steel locking compression plate was fixed 1 mm away from a bone surrogate in which the fracture gap was 47 mm. Two groups of 10 constructs, prepared with either two or three bicortical locking screws placed at the extremities of each fragment, were tested in a load-controlled 4-point bending test (range 0.7 to + 7 Nm) until failure. RESULTS: The 3-screw constructs were stiffer than the 2-screw constructs (19.73 ± 0.68 N/mm vs. 15.52 ± 0.51 N/mm respectively) and the interfragmentary relative displacements were higher for the 2-screw constructs (11.17 ± 0.88%) than for the 3-screw constructs (8.00 ± 0.45%). The difference between the number of cycles to failure for the 3-screw constructs (162,448 ± 30,073 cycles) and the 2-screw constructs (143,786 ± 10,103 cycles) was not significant. Failure in all constructs was due to plate fracture at the level of the compression holes. CLINICAL SIGNIFICANCE: Omission of the third innermost locking screw during bridging osteosynthesis subjected to bending forces led to a 20% reduction in construct stiffness and increased relative displacement (+39.6%) but did not change fatigue life.