The influence of overall treatment time on the radiation response of the mouse kidney was studied in an experiment in which 16 fractions were administered either evenly distributed over 20, 40 or 80 days, or as a split course (8 F/3 days; 74 days rest; 8 F/3 days). Urine output and an isotope assay of glomerular filtration were used to test the mice sequentially. The data were used both to obtain dose-response curves and also to determine the latent period before a chosen level of injury was expressed functionally. Prolonging the overall time from 20 to 80 days increased the isoeffect dose by 2-5 Gy (4-9%) for the isotope assay, and by 4-9 Gy (7-18%) for the urine output assay. This additional recovery as the interval between fractions was prolonged from 1 to 5 days is consistent with slow repair and can be expressed as a small "T" exponent of 0.02-0.12. (One analysis gave a result consistent with negative repair, but the errors on this result were unusually wide.) When the radiation was given as a split course, at the rate of 2 fractions per day, with a large gap of 10.5 weeks between courses, there was no additional sparing compared with 16 fractions over 20 days. This indicates that any sparing that might have resulted from slow repair or stimulated repopulation in the gap has been counterbalanced by having less time for repair of sublethal injury when intervals of 6-12 h are used instead of 24-48 h. Clearly no great increase in the tolerance dose for mouse kidney resulted from the split course.