The long-term success of massive osteochondral allografts depends not only on the incorporation of the transplanted articular cartilage. Osteochondral allografts are immunogenic, and, once an immune response is stimulated by exposure to donor cellular antigens, the cartilage becomes vulnerable to direct injury by cytotoxic antibodies or by lymphocytes, or to indirect injury by inflammatory mediators and enzymes induced by the immune response. To clarify the role of histocompatibility antigen-matching on the health of transplanted articular cartilage, we orthotopically implanted canine leukocyte antigen-matched and mismatched proximal osteochondral allografts of the radius, both fresh and cryopreserved, in beagles. Four groups of dogs received: (1) canine leukocyte antigen-mismatched frozen allografts, (2) canine leukocyte antigen-mismatched fresh allografts, (3) canine leukocyte antigen-matched fresh allografts, or (4) canine leukocyte antigen-matched frozen allografts. In twelve of the dogs, the contralateral leg was subjected to a sham operation, and in ten of the dogs, the proximal part of the radius was removed and replaced as an autogenous graft control. All animals were followed for eleven months after the operation and then were killed. The cartilage of the grafts was evaluated grossly, histologically, and biochemically. The biochemical analysis consisted of measurement of dry weight, content of glycosaminoglycan and hydroxyproline, and galactosamine-to-glucosamine ratios. Analyses of variance were used to study the effect of tissue antigen-matching and freezing on degradation of cartilage. During the study, no dog had grossly obvious clinical abnormalities, all host-graft interfaces healed, and no joints dislocated. The gross appearance of the cartilage was normal for both the joints that had an autogenous graft and those that were subjected to the sham operation. The cartilage of all allografts was thinned, dull, and roughened. The synovial membrane of all of the joints that had been operated on was mildly fibrotic and hyperplastic, but only that of the dogs that had an allograft was severely fibrotic and hyperplastic and demonstrated an inflammatory response. The inflammatory response was most severe in joints that had received a fresh canine leukocyte antigen-mismatched allograft. Invasive pannus was more frequent in joints that had received a fresh graft, particularly those that had received a canine leukocyte antigen-mismatched allograft, and cartilage was sometimes eroded to subchondral bone. Freezing was harmful to the cartilage. Very few cells survived the freezing procedure, and frozen grafts received s significantly worse histological scores had significantly less glycosaminoglycans and had a lower ratio of galactosamine to glucosamine than fresh grafts.