Repopulation in mouse tongue epithelium was determined during unconventional fractionation schedules, i.e., hyperfractionation (2 x 1.5 and 2 x 1.75 Gy/day) and accelerated treatment (2 x 3 Gy/day). The residual tolerance of the epithelium at defined days of the fractionated treatment was tested by graded single test doses (top-up design). The dose required to induce complete epithelial denudation in 50% of the animals (ED50) was used to calculate the number of fractions repopulated during the preceding treatment. After the first week of hyperfractionation, tolerance was reduced compared to untreated epithelium However, subsequently no further change was observed, indicating complete compensation of the weekly dose with all doses per fraction used. Epithelial cell density, defined by histological examination in additional experiments, in all fractionation arms decreased similarly by similar to 40% during the first week and remained constant at 60-80% in the subsequent 2 weeks. During accelerated fractionation, the residual mucosal tolerance decreased continuously with treatment time and resulted in epithelial denudation after 12 fractions. However, a substantial repopulation effect was observed, compensating 1.5 fractions by day 2, and 5 fractions by day 5, respectively. After cessation of the therapy the repopulation rate clearly decelerated to compensate a dose equivalent to about 0.5 fractions per day. Cell density decreased linearly during the treatment with 5, 10 or 12 fractions at a rate close to normal cell loss. Marked cell production, dependent on the total fractionated dose, was seen from one day after the last fraction in each experimental arm. These results indicate that maximum stem cell repopulation occurs predominantly during treatment, while major production of differentiating cells takes place in treatment splits.