The thorax of WAG/Rij rats was irradiated with fractionated doses of X rays. Irradiation schedules were designed either to allow virtually complete repair of sublethal damage between subsequent fractions by fractionating at 6-h intervals, or to result in incomplete repair by allowing only 1-h intervals between subsequent fractions. Combination of the data from both experimental series permitted the calculation of alpha/beta ratios and values for the repair halftime T1/2. The animals were monitored by assessment of the breathing frequency and by recording deaths. At the end of the experiments, 18 months after treatment, the hydroxyproline content of the lung tissue was determined as a biochemical indicator of radiation-induced fibrosis, and an histopathological analysis was performed. Early endpoints, indicative of radiation-induced pneumonitis, resulted in an alpha/beta ratio of 3.5 Gy and a T1/2 value of 0.95 h. Late endpoints were presumed to be indicative of radiation-induced fibrosis. Based on the combined analysis of data from three different late endpoints, the mean alpha/beta ratio was 2.3 Gy, and the T1/2 value was 1.13 h. The difference in alpha/beta ratio and T1/2 value between early and late endpoints was not significant, since the 95% confidence limits were overlapping. For each individual early or late endpoint as well as for the two early or the three late endpoints combined, there was a trend for lower alpha/beta ratios and higher T1/2 values associated with low doses per fraction. However, widely overlapping confidence limits indicated that again the differences were not significant.