Cells from two human cell lines were irradiated both as multicellular tumor spheroids (MTS) and in monolayer culture. Radiation response of MTS was quantified in terms of specific growth delay and proportion cured, and as clonogenic cell survival for monolayer cells. Radiation was applied either as a single or as a split dose with time intervals of 1, 2, and 4 h to determine the rate of sublethal damage repair. Using as endpoint the fraction of MTS cured at an iso-effect level, in MTS of NB-100 neuroblastoma cells repair of sublethal damage was complete within 1 h, whereas in MTS of HN-1 squamous cell carcinoma cells there was still some unrepaired damage left. At a larger dose for NB-100 MTS the repair curve showed a similar shape as for HN-1 spheroids. Using as endpoint specific growth delay, no difference in repair between the various time intervals was observed. In monolayer cells from both cell lines sublethal damage was not fully repaired in the time intervals used. Polarographic microelectrode measurements of oxygen tension inside MTS showed a marked difference in steepness of oxygen tension profiles between MTS from both cell lines. In HN-1 squamous cell carcinoma MTS with diameters up to 500 microns the central pO2 amounted to about 100 Torr, whereas in NB-100 neuroblastoma MTS with the same diameters central pO2-values lower than 30 Torr were observed. NB-100 MTS were irradiated with doses of 5 and 10 Gy gamma rays and subsequently the oxygen tension was measured 1 and 5 h after irradiation. A reoxygenation effect could not be observed, either after single dose or after split dose irradiation. If spheroids may be regarded as a suitable model for tumor responses in vivo, the results from these experiments indicate that reoxygenation is a process eluding polarographic measurements, or that no dramatic changes in oxygen tension are to be expected shortly after high single doses or early in a fractionation scheme.