The aim of this study was to evaluate the fixation strength of press-fit-implanted osteochondral grafts with respect to graft size (length and diameter), the effect of repeated insertion after pullout, and harvesting technique. Experiments were performed using the Osteochondral Autograft Transfer System on porcine femoral condyles. Failure loads of 10-mm-long grafts (mean, 47 N) were significantly lower than failure loads of 15-mm-long grafts (mean, 93 N) and 20-mm-long grafts (mean, 110 N) (all grafts, 11 mm in diameter). Reinsertion of the 15-mm-long grafts after initial pullout resulted in a significant reduction of failure loads (mean, 93 N versus 44 N). Failure loads of 8-mm-diameter grafts (mean, 41 N) were significantly lower than those of 11-mm-diameter grafts (mean, 92 N) (all 15 mm long). Levering of the tubular chisel during graft harvest significantly decreased press-fit stability as compared with simple turning of the chisel (mean, 32 N versus 52 N) (8-mm diameter and 15-mm length). These results suggest that primary fixation strength of press-fit-inserted osteochondral grafts depends on the size of the grafts and that repeated pullout and reinsertion of grafts as well as a nonoptimal harvesting technique (levering) will reduce primary stability.