Clinical efforts to repair cartilage defects delivering cells or engineered cartilage implants into the lesions have met with limited success. This study used a critical-size chondral defect model in immunocompromised rat xiphoid cartilage to test whether endogenous chondrogenesis could be achieved using human bone matrix scaffolds to deliver human cartilage particles and/or a variant isoform of fibroblast growth factor-2 (FGF2-variant). Seventy-two male athymic RNU rats were enrolled in this study with eight rats per experimental group. Decellularized and demineralized human bone matrix scaffolds loaded with human articular cartilage particles or heat-inactivated cartilage particles were combined with different doses of the FGF2-variant. Scaffolds were implanted into 3-mm-diameter critical-size defects prepared using a biopsy punch through the center of the xiphoid. The samples were evaluated 28 days postsurgery using X-ray, equilibrium partitioning of ionic contrast microcomputed tomography, and safranin O-stained histological sagittal sections. Scaffolds containing cartilage particles plus the FGF2-variant induced dose-dependent increases in the formation of neocartilage (p<0.05), which was distributed homogeneously throughout the defects in comparison to scaffolds containing only the FGF2-variant. These effects were less pronounced when scaffolds with heat-inactivated cartilage particles were used. These results demonstrate that endogenous repair of chondral defects can be achieved in the absence of exogenous cells or bone marrow, suggesting that a similar approach may be successful for treating chondral lesions clinically.