Damage from occupational or accidental exposure to ionising radiation is often assessed by monitoring chromosome aberrations in peripheral blood lymphocytes, and these procedures have, in several cases, assisted physicians in the management of irradiated persons. Thereby, circulating lymphocytes, which are in the Go stage of the cell cycle are stimulated with a mitogenic agent, usually phytohaemagglutinin, to replicate in vitro their DNA and enter cell division, and are then observed for abnormalities. Comparison with dose-response relationships obtained in vitro allows an estimate of exposure based on scoring: Unstable aberrations by the conventional, well-established analysis of metaphases for chromosome abnormalities or for micronuclei; So-called stable aberrations by the classical G-banding (Giemsa-Stain-banding) technique or by the more recently developed fluorescent in situ hybridisation (FISH) method using fluorescent-labelled probes for centromeres and chromosomes. Three factors need to be considered in applying such biological dosimetry: (1) Radiation doses in the body are often inhomogeneous. A comparison of the distribution of the observed aberrations among cells with that expected from a normal poisson distribution can allow conclusions to be made with regard to the inhomogeneity of exposure by means of the so-called contaminated poisson distribution method; however, its application requires a sufficiently large number of aberrations, i.e. an exposure to a rather large dose at a high dose rate. (2) Exposure can occur at a low dose rate (e.g. from spread or lost radioactive sources) rendering a comparison with in vitro exposure hazardous. Dose-effect relationships of most aberrations that were scored, such as translocations, follow a square law. Repair intervening during exposure reduces the quadratic component with decreasing dose rate as exposure is spread over a longer period of time. No valid solution for this problem has yet been developed, although, in theory, both deterministic damage and aberrations might be repaired to a similar degree; a comparison of aberrations following a linear dose relationship might also help when the doses have been sufficiently large. (3) Investigations might have been possible only a certain time after the exposure. The relatively rapid disappearance of lymphocytes carrying unstable aberrations limits their use in retrospective dosimetry, years after exposure. Scoring stable aberrations, thought to persist in the circulating lymphocytes, might appear more appropriate in such situations. However, the examination of a representative number of cells by G-banding is extremely laborious, and the FISH method is not only expensive but has not yet been fully validated in different laboratories. In conclusion, biological dosimetry has serious limitations exactly for situations where the need for information is most urgent. It renders its most useful results when an individual has been exposed to a rather homogeneous high-level radiation over a short time interval, i.e. accidents at high-intensity radiation devices. On the other hand, it yielded less satisfactory information even when the most recent techniques were used for situations, where a low level, low dose rate exposure has occurred at some time in the past, for example for persons living in areas contaminated from the Chernobyl accident. Such negative experiences should be kept in mind in order to avoid futile and expensive investigations in the case of populations exposed from radioactivity and, notably, also from potentially clastogenic chemical agents.