Periosteal autograft were used for repair of large osteochondral defects in 10 horses aged 2 to 3 years old. In each horse, osteochondral defects measuring 1.0 X 1.0 cm2 were induced bilaterally on the distal articular surface of each radial carpal bone. Control and experimental defects were drilled. Periosteum was harvested from the proximal portion of the tibia and was glued into the principal defects, using a fibrin adhesive. Control defects were glued, but were not grafted. Sixteen weeks after the grafting procedure, the quality of the repair tissue of control and grafted defects was assessed biochemically. Total collagen content and the proportion of type-II collagen were determined. Galactosamine and glucosamine contents also were determined. From these measurements, contents of chondroitin and keratan sulfate and total glycosaminoglycan, and galactosamine-to-glucosamine ratio were calculated. All biochemical variables were compared with those of normal equine articular cartilage taken from the same site in another group of clinically normal horses. Total collagen content was determined on the basis of 4-hydroxyproline content, using a colorimetric method. The proportions of collagen types I and II in the repair tissue were assessed by electrophoresis of their cyanogen bromide-cleaved peptides on sodium dodecyl sulfate slab gels. Peptide ratios were computed and compared with those of standard mixtures of type-I and type-II collagens. Galactosamine and glucosamine contents were determined by use of ion chromatography. In general, the biochemical composition of repair tissue of grafted and nongrafted defects was similar, but clearly differed from that of normal articular cartilage. Total glycosaminoglycan content, galactosamine and glucosamine contents, and galactosamine-to-glucosamine ratio of grafted and nongrafted defects were all significantly (P < 0.05) less than corresponding values in normal equine articular cartilage. By contrast, total collagen content of neocartilaginous tissues of grafted and nongrafted defects was greater than that of normal articular cartilage, although the difference was not significant. The proportion of type-I and type-II collagens in repair tissue in grafted and nongrafted defects was 70 and 30%, respectively. The fibrous nature of the repair tissue reported in a companion morphologic and histochemical study was substantiated by the biochemical results. We concluded that use of periosteal autograft did not improve the healing of osteochondral defects.