The aims of this study were to test analytical procedures for the determination of Cd, Cu, Mn, Pb, Se and Zn in breast milk and to establish optimum sampling conditions for monitoring purposes. Two population groups were analysed: (1) Seven women from Prague whose breast milk was sampled on days 1,2, 3, 4, 10, 20 and 30 after delivery; (2) 200 women from four (two industrial and two rural) regions whose breast milk was sampled at defined intervals. All samples were mineralised in a microwave oven in the mixture of HNO3 + H2O2 and analysed by atomic absorption spectrometry. Conditions for the measurement of the elements under study (i.e. those for the electrothermal atomisation for Cd, Mn and Pb, flame technique for Cu and Zn, and hydride generation technique for Se) were optimized. Using optimized parameters the analysis was performed and the following conclusion has been made: the concentrations of zinc and manganese decreased very sharply over the first days, that of copper slightly increased within the first two days and then slightly decreased, that of selenium did not change significantly. Partial "stabilisation" was achieved after the second decade. No correlation among the elements was found. A significant difference between whole and skim milk was only found for selenium (26% rel. higher in whole milk). The majority concentrations of cadmium and lead were below the detection limit of the method (0.3 microg x l(-1) and 8.2 microg x l(-1), respectively, as calculated for the original sample). To provide biological monitoring, the maintenance of sampling conditions and especially the time of sampling is crucial.