We have tested the hypothesis that the inhibition by hydroxyurea of repair patch ligation and chromatin rearrangement during u.v.-induced DNA excision repair results from a reduction in cellular deoxyribonucleotide concentrations and not from a direct effect of hydroxyurea on the repair process. Using permeable human fibroblasts, we have shown that hydroxyurea has no direct effect on either repair synthesis or repair patch ligation. We also have shown that by reducing the deoxyribonucleoside triphosphate concentrations in the permeable cell reaction mixture, we can mimic the inhibition of repair patch ligation and chromatin rearrangement seen when u.v.-damaged intact confluent fibroblasts are treated with hydroxyurea. Our results are consistent with the concept that hydroxyurea inhibits DNA repair in intact cells by inhibiting deoxyribonucleotide synthesis through its effect on ribonucleotide reductase and, conversely, that continued deoxyribonucleotide synthesis is required for the excision repair of u.v.-induced DNA damage even in resting cells.