When human lymphocytes are treated with [3H]thymidine ([ 3H]dThd), the observed number of chromosomal aberrations induced by [3H]dThd and subsequent doses of X-rays is less than the number induced by X-rays alone. Experiments in which cells were examined at various times after exposure to the X-rays showed that this phenomenon, termed an adaptive response to the endogenous radiation from tritium, is not an artefact caused by radiation-induced mitotic delays, which could result in the sampling of metaphase cells that were irradiated in different parts of the G2 phase of the cell cycle, where sensitivity to X-rays changes dramatically. Reconstruction experiments in which labelled female cells were co-cultured with unlabelled male cells now show that labelled and unlabelled cells progress to metaphase equally, and therefore that the adaptive response is not the result of selection against a radiosensitive population of cells that have incorporated [3H]dThd. Measurements of chromosomal aberrations induced in the labelled female cells and unlabelled male cells that had been co-cultured show that the adaptive response is restricted to those cells exposed to radiation from the incorporation of [3H]dThd and that diffusible factors are not involved. The results are consistent with the proposal that this adaptive response is the result of the induction of a hitherto unknown chromosomal repair mechanism. It has now been found that this repair mechanism is induced at levels of radiation from [3H]dThd that in themselves are too low to induce any discernible chromosomal aberrations and that its activity is dependent on the enzyme poly(ADP-ribose) polymerase, because 3-aminobenzamide, an inhibitor of poly(ADP-ribosyl)ation, prevents the adaptive response.