The kinetics of DNA strand break repair was studied in exponentially-growing CHO cells after X-irradiation with doses of 3, 9, 30, 60 and 90 Gy. DNA strand breaks were measured using the alkaline unwinding technique. For all X-ray doses applied the kinetics of DNA strand break repair consisted of fast, intermediate and slow phases. The latter, which was interpreted as the repair kinetic of DNA double-strand breaks, was best described by an exponential decline. The actual repair half-time of double-strand break repair, tau dsb, was obtained from the slope of the slow component after subtracting the number of non-reparable breaks measured 24 h after irradiation. This half-time was found to be independent of the dose applied with a mean value of tau dsb = 168 +/- 10 min. This result indicated that the repair of double-strand breaks was unsaturated for doses up to 90 Gy. The repair kinetics of the breaks of the fast and intermediate phases were found to be dependent on the dose applied. These kinetics were associated with the induction and repair of primary and secondary single-strand breaks, the latter possibly generated by enzymatic incision at damaged bases. Analysis of these curves using the Michaelis-Menten equation showed that the half-time of enzymatic incision, tau in, and the half-time at which both primary and secondary single-strand breaks were rejoined, tau rep, varied with the amount of damage present in the cell. tau in Increased from a minimum value tau in,min = 13 +/- 2 min proportionally to the number of base damage with a rate of alpha in = 0.138 +/- 0.015 min Gy-1, and tau rep from a minimum value tau rep,min = 1.4 +/- 0.2 proportionally to the number of single-strand breaks with a rate of alpha rep = 0.038 +/- 0.001 min Gy-1. The enzymatic incision was unsaturated for doses up to about 30 Gy, whereas the repair of single-strand breaks was unsaturated only for doses up to about 10 Gy. Up to these doses the increase in the half-time tau in and tau rep was so small that, within the range of experimental errors, the parameters may be approximated by constant values. From the results it is concluded that for CHO cells the continuously-bending dose-response curve obtained for radiation-induced killing cannot be attributed to a saturated repair.