Mechanical fatigue is the predominant etiology of stress fracture, a known contributor to atypical femoral fracture, and may also play a critical role in fragility fracture. While these fatigue-related fractures are well-documented in humans, they are poorly understood. Extensive research has attempted to characterize the fatigue behavior of cortical bone;​ however, owing to the inherent variability in bone tissue, samples that appear identical in macrostructure can exhibit a large degree of scatter in fatigue life. The overarching hypothesis of this thesis is that the variance in fatigue-life data can be attributed to intracortical microarchitecture, including the size, spacing, and density of vascular canals and osteocyte lacunae. A series of studies were conducted that utilized ex vivo mechanical testing, high-resolution imaging, and finite element modeling to establish the relationship between intracortical microarchitecture and the fatigue life of bone in compression. Both porosity and canal diameter demonstrated a strong negative relationship with fatigue life, whereas lacunar density was positively correlated. The reduced fatigue life associated with higher porosity was a result of larger, rather than more abundant canals, indicating that canals act as stress concentrators that may impair the fatigue resistance of bone beyond increasing overall porosity. The stress concentrations caused by vascular canals were quantified as stressed volume (i.e., the volume of material above yield) which was positively correlated to porosity and canal diameter. Furthermore, stressed volume proved to be a strong predictor of fatigue-life variance across multiple loading magnitudes. The findings from thisthesis suggest that a majority of the fatigue-life variance of cortical bone in compression is driven by intracortical microarchitecture, and fatigue failure may be predicted by quantifying the stress concentrations associated with vascular canals.