Bone is a stiff, porous, hierarchical structure living tissue that provides structural support to the body via the skeletal system. Bone structural integrity is crucial for the quality of life. However, bones can only bear a load until a certain limit, without experiencing permanent deformation or failure. Therefore, investigation of fracture behaviour of weight-bearing bones of the lower limbs such as femur is of great interest. Additionally, due to the anatomical location and geometry of femur neck, it is the most common fractured location of femur bone particularly in the older adults and patients who have musculoskeletal disorders that often cause to permanent disability or mortality. On the other hand, the mechanical and microstructural properties of cortical bone of femur neck are not adequately understood. This originates from limitations in specimen size for mechanical testing. Therefore, the size of the specimen is an essential matter in mechanical bone behaviour studies. Hence, this thesis aim to investigate the size limitation for maintaining the material continuum assumption. This aim can be accomplished by examining the structural and mechanical properties of cortical bone such as fracture toughness which explains the resistance of bone to crack beginning and propagation. In this thesis, fracture behaviour of cortical bone was studied under three-point bending test which was conducted on single-edge-notch bending (SENB) specimens with different size, 1 mm x 1 mm, 1.7 mm x 1.7 mm, and 3 mm x 3 mm, by the transverse notched direction from different cortex positions, called anterior, lateral, medial, and posterior, from mid-shaft of two equine femurs with different ages. Due to anisotropic behaviour and hierarchical structure of cortical bone, in this thesis, elastic plastic fracture mechanics approach is the best method to evaluate the fracture toughness and fracture energy of cortical bone that was conducted using J-integral approach. Fracture toughness of both equine cortical bone was found to be sensitive to beam size particularly in the young femur. However, most of the significant differences of fracture energy in the largest beam size still present in the smallest beam size with the lower detectability differences except for one case. Therefore, it should be feasible to use fracture energy for smaller samples where it is not possible to obtain larger samples.