This paper presents the results of a test programme to validate the Dutch pre-normative protocol NVN 5623 'Radioactivity measurements: Determination of the activity of gamma-ray emitting nuclides in a counting sample by means of semi-conductor gamma-spectrometry'. Batches of two reference materials, water and sand, both spiked with radionuclides ( 1 3 3 Ba, 1 3 4 Cs, 1 3 7 Cs, 4 0 K, 5 7 Co and 6 0 Co) were used in an intra- and interlaboratory test. For the water and sand, 10 and 9 laboratories, respectively, agreed to analyse three samples of each reference material strictly according to the pre-normative protocol. The interlaboratory test programme showed that for both reference materials the values for the reproducibility standard deviation were 1.5-6 times larger than those for the repeatability standard deviation. This extra variation is caused by the parameter 'laboratory'. For both reference materials the repeatability standard deviation was found in the range 0.9-2.0% of the mean activity concentration; the reproducibility standard deviation ranged from 2.5-6.0%. Bias of the test method was found for 1 3 4 Cs (-6.8%) and 1 3 3 Ba (-2.3%) in case of the spiked water and for 1 3 4 Cs (-7.8%) in case of the spiked sand. This bias is attributed to no or insufficient correction for coincidence summing.In general, the interlaboratory test showed satisfactory agreement between laboratories. In the light of regulations on radioactivity in materials it is concluded that for measurement of this activity the Dutch protocol may serve as a blueprint for an (European Community) protocol.