Previous reports have suggested that Lewis (Le) antigens may exert a significant effect on cadaver renal allograft (CRA) survival, especially in black recipients in whom there is a higher frequency of Le-negative phenotypes. We review our experience with this problem in 70 donor-recipient pairs of CRA who underwent prospective Le typing and received conventional immunosuppression between 1980 and 1983. Recipient typing alone yielded the following graft survival (GS) and patient survival (PS) at 2 years by life table analysis: (a+,b;-) (n = 12) 51% GS, 93% PS; (a-, b+) (n = 44) 57% GS, 88% PS; and (a-,b-) (n = 14) 51% GS, 93% PS(P-ns for GS, PS). Recipient racial characteristics did not effect ultimate graft survival, as whites and blacks had similar two-year GS in all phenotypic groups. When Le matching was considered, no significant differences in one-year graft survivals could be ascertained between Le-matched and Le-mismatched donor-recipient pairs, and this effect persisted despite stratification for race and HLA-A,B and DR histoincompatibilities. In light of these results, we do not recommend using Lewis compatibility as a criterion for donor selection in cadaver renal allografting, as this substantially increases the difficulty in finding suitable matches, especially in the (a-,b-) recipient group.