Abstract. Natural radioactivity is associated with the vast mineral resources in South Africa in such concentrations that the radionuclides from the natural uranium and thorium decay series are found to pose concern for public exposure to communities living around these areas. Consumption of water and food is usually the most important route by which natural radionuclides can enter the human body and assessment of natural radionuclide levels in different foods and diets is therefore important to estimate the intake of these radionuclides by man. Sensitive measurement of three major radionuclides (in addition to 238 227 Ac and 228 Ra also require improved sensitivity. In order to evaluate the yearly dose due to an individual source at a screening level of 25 Sv/a, one is faced with a required lower limit of determination (LLD) of 0.1 to 0.5 Bq/kg for certain foodstuffs. In this study the potential radiological impact of foodstuffs grown in a catchment area influenced by mining and mineral processing industries in South Africa, was determined by measuring the natural radionuclides in a number of foodstuffs collected from the area. The radionuclides were measured by non-destructive techniques such as Instrumental Neutron Activation Analysis (INAA) and low background gamma spectrometry. The assessment of natural radionuclides in foods allowed us to evaluate the items that present the highest risk to the population, and compare this to the limits established by the National Nuclear Regulator (NNR). From the public's point of view it is important to ensure the population that the contaminant levels in specific food as a result of mining activities, do not exceed the permissible limits. For the majority of the analysed foodstuffs, the estimated dose was less than 250 Sv/a. Overestimation of dose, due to poor measurement detection limits, was clearly indicated for some of the samples. For most of the analysed nuclides, suitable data for evaluation of the yearly dose at the screening level of 25 Sv/a was obtained by broad energy gamma analysis of ashed samples. However, nuclides such as 230 Th, 210 Po and 231 Pa has to be analysed by radiochemical separation through acid destruction of dried foodstuffs followed by individual element separations to provide suitable data with a low enough LLD not to result in an overestimation of the calculated dose.