Abstract. In the present paper, the size of the soil particle and the weight of the aliquot submitted to gamma measurement or radiochemical analysis, while fulfilled the requirement of representativeness, were investigated experimentally. The radioactive contaminated soil sample was pulverized, 239 Pu was separated and determined by isotope dilution introduced coupled plasma mass spectroscopy, and gamma emitters, 241 Am and 137 Cs etc were measured by the gamma spectrometry. The dependence of the time taken during the milling and consequent mesh size, and the relationship between the weight of aliquot for radiochemical analysis and the mesh size of the soil were established, under the supposed representativeness criteria of (1) the relative standard deviation (RSD) less than 20% and (2) the deviation from certificated value (DCV) less than 10%. An empirical equation of the aliquot weight and size was fitted, by which the optimum size is 40mesh when circa 180 g is used for gamma measurement, and the preference size is 150 mesh when the aliquot for the radioanalysis around 2 g soil is scooped for assay. These results could be used in the assay of the soil samples in the radiological survey.