The aim of the present study was to evaluate the influence of the exclusion of cells with unstable aberrations in the elaboration of dose-effect curves for translocations and their implications in biological dosimetry of past exposures. To establish dose-effect curves, peripheral blood samples were irradiated with 60CO γ rays at ten different doses and the yield of translocations analyzed by FISH was considered in all cells and in stable cells (those without dicentrics, acentrics or rings). To discriminate transmissible translocations, the dose-effect curve for total apparently simple translocations in stable cells was chosen as the reference. In stable cells, dose-effect curves for apparently simple translocations without pseudosimple and complex-derived one-way patterns, t(Ab)t(Ba) and total translocations were obtained. None of these curves differed from the reference curve. When all cells were considered, only the curve for total translocations was significantly different from the reference curve. From the results obtained it can be concluded that the use of dose-effect curves for apparently simple translocations in stable cells and in all cells will give similar dose estimates in retrospective biological dosimetry studies. However, the use of dose-effect curves for total translocations in all cells will lead to underestimations of the dose mainly at high doses. © 2004 by Radiation Research Society.