An important concept in bone mechanics is that osteons influence mechanical properties in several ways, including contributing to toughness and fatigue strength by debonding from the interstitial matrix so as to "bridge" developing cracks. Observations of "pulled out" osteons on fracture surfaces are thought to be indicative of such behavior. We tested the hypothesis that osteon pullout varies with mode of loading (fatigue vs. monotonic), cortical region, elastic modulus, and fatigue life. Mid-diaphseal beams from the dorsal, medial, and lateral regions of the equine third metacarpal bone were fractured in four point bending by monotonic loading to failure under deflection control, with or without lo5 cycles of previous fatigue loading producing 5000 microstrain (15-20'%) of the expected failure strain) on the first cycle; or sinusoidal fatigue loading to failure, under load or deflection control, with the initial cycle producing 10,000 microstrain (3040% of the expected failure strain). Using scanning electron microscopy, percent fracture surface area exhibiting osteon pullout (%OP.Ar) was measured. Monotonically loaded specimens and the compression side of fatigue fracture surfaces exhibited no osteon pullout. In load-controlled fatigue, pullout was present on the tension side of fracture surfaces, was regionally dependent (occurring to a greater amount dorsally), and was correlated negatively with elastic modulus and positively with fatigue life. Regional variation in '%OP.Ar was also significant for the pooled (load and deflection controlled) fatigue specimens. '%)OP.Ar was nearly significantly greater in deflection controlled fatigue specimens than in load-controlled specimens (p = 0.059). The data suggest that tensile fatigue loading of cortical bone eventually introduces damage that results in osteonal debonding and pullout, which is also associated with increased fatigue life via mechanisms that are not yet clear.