Multicellular tumour spheroids (MTS) are a useful in vitro model of human cancer. An experiment was designed to assess the likely therapeutic advantage of hyperfractionation--a proposed strategy in radiotherapy. A cell line (NB1-G) derived from human neuroblastoma was grown as MTS. This MTS line is radiosensitive with low capacity for repair of sublethal radiation damage. These properties make NB1-G a suitable line to test the theoretical advantage of hyperfractionation. MTS were irradiated using alternative fractionated regimens, with fraction sizes varying from 0.5 to 4 Gy. In each experiment, the total dose was chosen to make the regimens theoretically isoeffective for damage to late-responding normal tissues (calculated using the linear-quadratic mathematical model with alpha/beta = 3 Gy). The radiation responses of MTS were evaluated using the end-points of regrowth delay and "proportion cured". Regimens using smaller doses per fraction were found to be markedly more effective in causing damage to neuroblastoma MTS, as assessed by either end-point. These experimental findings support the proposal that hyperfractionation should be a therapeutically advantageous strategy in the treatment of tumours whose radiobiological properties are similar to those of the MTS neuroblastoma line NB1-G.