Purpose: To evaluate by the fluorescent in-situ hybridization (FISH) technique the dose-response and intercellular distribution of a-particle-induced chromosome aberrations. In particular, the validity of using the yield of characteristic types of chromosome abnormalities in stable cells as quantitative indicators for retrospective dose reconstruction has been evaluated. Material and methods: Monolayers of human peripheral lymphocytes were exposed at doses from 0.02 to 1 Gy to α-particles emitted from a source of americium-241. The most probable energy of the a-particles entering the cells was 2.7 MeV. FISH painting was performed using DNA probes for chromosomes 2, 4 and 8 in combination with a pan-centromeric probe. In complete first-division cells, identified by harlequin staining, aberrations involving painted target chromosomal material were recorded as well as aberrations involving only unpainted chromosomal material. Results: In total, the percentage of complex aberrations was about 35% and no dose dependence was observed. When complex-type exchanges were reduced to simple base types, the different cell distributions were clearly over-dispersed, and the linear coefficients of the dose-effect curves for translocations were significantly higher than for dicentrics. For past dose reconstruction, only a few complex aberrations were in stable cells. The linear coefficient obtained for transmissible aberrations in stable cells was more than seven times lower than that obtained in all analysed cells, i.e. including unstable cells. Conclusion: FISH-based analysis of complex rearrangements allows discrimination between partial-body exposures to low-linear energy transfer radiation and high-linear energy transfer exposures. In assessing past or chronic exposure to α-particles, the use of a dose-effect curve obtained by FISH-based translocation data, which had not excluded data determined in unstable cells, would underestimate the dose. Insertions are ineffective biomarkers because their frequency is too low.