Inhibition of cell growth and DNA synthesis by hydroxyurea is thought to occur via an effect on the enzyme ribonucleotide reductase leading to a block of deoxyribonucleotide synthesis. Earlier attempts to bypass such a block by delivering deoxyribonucleosides to the medium of cultured cells have given equivocal results. Complications arise in such experiments from the specificity of the phosphorylating enzymes since 3 of the 4 deoxyribonucleosides are substrates for the same enzyme, with widely differing Km values, and from allosteric effects exerted by deoxyribonucleotides. We simplify this situation by using a mutant hamster V79 line that lacks the enzyme dCMP deaminase. The cells contain a 20-fold enlarged dCTP pool and require thymidine for optimal growth. Concentrations of hydroxyurea (50 or 100 microM) that in short-term experiments inhibited DNA synthesis depleted the dATP pool without seriously affecting pyrimidine deoxyribonucleotide pools. The dATP pool could be restored by addition of deoxyadenosine but this depleted the dGTP pool. This depletion could be counteracted by the simultaneous addition of deoxyguanosine but then critically depended on the relative concentrations of the two purine deoxyribonucleosides, with optimal results at 1 microM deoxyadenosine + 100 microM deoxyguanosine. Under those conditions the inhibition of DNA synthesis by hydroxyurea was partially reversed.