In the investigation of radiation accidents, it is of great importance to estimate the dose absorbed by exposed persons in order to plan their therapy. Although occasionally in these situations physical dose measurements are possible, most often biological methods are required for dose estimation. The aim of this investigation was to assess the suitability of the cytokinesis blocked (CB) micronucleus assay as a biodosimetric method using lymphocytes irradiated in vivo. The approach adopted to achieve this was to estimate whole body doses by relating micronuclei yields in patients undergoing radiotherapy treatment with an in vitro radiation dose-response curve. These biologically derived estimates were then compared with the corresponding doses obtained by physical measurement and calculation. As a first approach a study was performed of the in vitro dose-response of gamma-ray induced micronuclei following cytokinesis-block in the lymphocytes of peripheral blood samples obtained from 4 healthy donors. The results indicated that the distribution of the induced micronuclei were overdispersed. Furthermore, a linear dose-response relationship was established when a curve was fitted to the data by an iteratively reweighted least squares method. By means of an analysis of covariance it was demonstrated that this result is in agreement with the dose-response relationships found by various other workers (Fenech et al., 1985; Fenech et al., 1986; Fenech et al., 1989; Balasem et al., 1992, and Slabbert, 1993). To assess the suitability and accuracy of dose assessment using the CB micronucleus assay for in vivo exposure of lymphocytes, blood samples obtained from 8 patients undergoing radiotherapy before, during and after treatment were examined. The physical doses of these patients were determined according to conventional radiation treatment plans and cumulative dose-volume histograms. The dose-volume histograms permitted calculation of integral doses and subsequently the estimate of equivalent whole-body doses. The results of the CB micronucleus assay applied to peripheral blood lymphocytes of 6 patients undergoing fractionated partial-body irradiation showed a dose-related increase in micronucleus frequency in each of the patients studied. This demonstrated that micronuclei analysis may serve as a quantitative biological measure of such exposures. The pooled data of these patients compared to the pooled data of the healthy donors show that there was no statistically significant difference between in vitro and in vivo results, however a slightly lower induced micronuclei frequency was observed after in vivo exposure. When the biological dose estimates for equivalent whole-body doses obtained from the in vitro dose response curve were compared with calculated physical doses, it was found that: biologically estimated dose = 0.936 physical dose. However, there was inadequate statistical evidence to discard the hypothesis that the gradient of the equation was equal to one. Therefore, the analysis of micronuclei induced in lymphocytes in vivo yields highly quantitative information on the equivalent whole-body dose. The negative binomial method was used for analysing the micronucleus data from two patients who received single, relatively larger tumour doses of 10 Gy each, with the objective to obtain estimates of the exposed body fraction and the dose to this fraction. The dose estimates to the irradiated volume were found to be within 30% of the physical tumour dose. The irradiated volume estimates seemed to be higher than the physically calculated volumes but by discarding the correction for the loss of cells due to interphase death the agreement was good between the physically and biologically determined integral doses. This study has revealed that the CB micronucleus assay appears to offer a reliable, consistent and relatively rapid biological method of whole body dose estimation. It is recognised that further corroborative work using the techniques described in this thesis is required for estimating localized exposure.