Movement of line-like samples with lengths from 5 to 50 mm along the x-axis of the double TE104 rectangular cavity has been analyzed. The observed dependencies of the EPR signal intensity versus sample position showed: (i) a sharp maximum for sample lengths from 5 to 20 mm; (ii) a plateau, over which the EPR signal intensity remained constant within experimental errors of 0.26-1.07%, for lengths from 30 to 40 mm; and (iii) a "sloping plateau," which could be approximated by the linear function (correlation, r = 0.98) for sample length 50 mm. Theoretical values of the experimentally observed dependencies of the intensity versus sample position were calculated using the modified sine-squared function and the correlation between observed and theoretically predicted dependencies is very good. The experimental dependence of the EPR signal intensity versus the sample length for samples situated at the same point in the cavity was nonlinear with a maximum for the 40-mm sample. The dependence of the EPR signal intensity upon the movement of a large cylindrical sample (o.d. 4 mm and length 100 mm) along the x-axis of the cavity was similar to that found for the 50-mm sample. However, an additional oscillating signal superimposed on the sloping plateau was observed. The presence of a large sample fixed in the complementary cavity of the double TE104 cavity caused an additional deformation of the signal intensity for a 30-mm sample which was moving in the first cavity. The primary effect was that the plateau was replaced by a region in which the intensity increased linearly with sample position, r = 0.99. Each of the above phenomena may be a source of significant errors in quantitative EPR spectroscopy. Cylindrical samples to be compared should be of identical length and internal diameter. Accurate and precise positioning of each sample in the microwave cavity is essential. Copyright 1997 Academic Press. Copyright 1997Academic Press