1. Black transplant recipients showed little or no matching effect when HLA-A, B, and DR antigens matched. This may have been due in part to differences in serologically defined antigens found in Black populations. 2. There was a statistically significant (p less than 0.001 for best vs worst match) matching effect in White transplant recipients that occurred at 3 months following grafting and continued through 3 years follow-up. These patients had better graft survival for both 5 and 6, compared to 3 and 4, or as compared to 0-, 1-, and 2-match transplants; the corollary groups for mismatched transplants were also significantly different. At 3 years, there was a 20% difference in graft survival from the best- to the worst-matched transplants. Earlier this difference was 7-10% but significantly different. 3. When the HLA Class I parent antigens as defined in the Materials and Methods section were eliminated, there was an increase in graft survival for the best-matched transplants. This was postulated to occur because of better matching obtained by matching the well-defined HLA specificities. 4. With hierarchy matching, possibly there was a stronger effect of matching HLA Class II antigens at 3 months; however, this effect was quickly lost at 1 and 3 years, and data that support early graft survival advantages by matching Class II antigens were inconclusive. Most importantly, there was a threshold matching effect that showed transplant recipients with 4 or more HLA matches or 2 or less mismatches had significantly better short- and long-term outcome than the poorer-matched groups.(ABSTRACT TRUNCATED AT 250 WORDS)