According to the soil-to-plant transfer concept generally used in dose assessment modeling, the plant uptake of a radionuclide should depend linearly on its concentration in the soil. In order to validate this concept for (90)Sr in a semi-natural ecosystem, plant and soil samples were taken at 100 plots of a 100 x 100 m(2) area within an alpine pasture near Berchtesgaden, Germany. At three plots, the vertical distribution of (90)Sr in the soil was determined in addition. A statistically significant correlation between the soil and plant concentration of (90)Sr was not detectable (Spearman correlation coefficient R=-0.116, p>0.05) within the range of the Sr-concentration covered (15-548 Bq kg(-1) dry soil and 17-253 Bq kg(-1) dry plant material). Thus, the prerequisite of the soil-to-plant transfer concept was not fulfilled for (90)Sr at this site. Organic carbon and total nitrogen were also determined in the soil samples. Both elements were highly correlated (R=0.912, p<0.001), their ratio being C/N=10.9+/-0.7. While C was positively correlated with the (90)Sr concentrations in the soil (R=0.342, p<0.001), negative correlations were observed for the plant concentrations (R=-0.286, p<0.01) and the concentration ratios (R=-0.444, p<0.001) of (90)Sr. These results are compared with those recently obtained for (137)Cs by Bunzl et al. (J Environ Radioactiv 48 (2000) 145).