Advocates of vascularized bone grafts believe that these grafts should have a decreased time to graft-host union, and that they should be mechanically stronger than conventional (non-vascularized) grafts. The objectives of the present study were to determine the rate and pattern of repair in vascularized autogenous cortical bone grafts, to determine the mechanical strength of the grafts, and to correlate the mechanical strength with the biological repair. Forty-nine adult male mongrel dogs were divided into six groups to evaluate conventional (non-vascularized), cuff (periosteal-encased, non-vascularized), and vascularized segmental grafts. The fibula was the site of experimentation and all grafts were four-centimeter cortical segments. The vascularized and conventional grafts were analyzed at two, six, twelve, and twenty-four weeks. The cuff grafts were analyzed at twenty-four weeks and were compared with conventional grafts to assess the effect of the periosteal soft tissue. Roentgenograms were made every two weeks to evaluate the time to union. The mechanical strength of each graft was assessed by determining rapid torsional load to failure. Biological repair was assessed by tetracycline labeling for new-bone formation and by microradiographic techniques for porosity and cross-sectional areas. The study showed that conventional and cuff grafts were similar in terms of mechanical and biological repair at six months. At all sampling intervals, the vascularized grafts exhibited histological findings that were consistent with viability. The conventional and vascularized grafts underwent different mechanisms of repair. The conventional, non-vascularized grafts healed by peripheral and internal resorption followed by callus encasement and osteonal remodeling.(ABSTRACT TRUNCATED AT 250 WORDS)