[en] A complicated model describing the movement of a radionuclide in both the natural environment and socioeconomical systems is usually used to estimate the internal dose to the public in terms of collective dose, taking demographic data into account. The result can be certified for reliability in some compartments of the model. One of the compartments is the body burden. In the case of radiocesium, the individual body burden can be measured using a whole-body counter. The measurement must be calibrated with a phantom. The public is composed of individuals of various ages. Accordingly, the whole-body counter should be calibrated with a set of phantoms approximating individuals of different body sizes. A relationship between counting efficiency and body size was analyzed on ¹³⁷Cs incorporated into the whole body using a set of phantoms. Four sizes covering average Japanese physique from infant to adult male, were chosen to prepare an anthropomorphic phantom system. Some hollow containers which simulated body segments were made from polyvinyl chloride and assembled to use as the phantom. The distribution of ¹³⁷Cs in aquatic solution was homogeneous through the phantom. A whole-body counter at the National Institute of Radiological Sciences, which had two 8 x 4 inch NaI(Tl) detectors sandwiching the subject, were used at a rate of 5 cm per minute in a scanning mode. The measurements were carried out in an iron room. Relations were analyzed between counting efficiency and some anthropometric parameters. The best fit was given by a linear equation of both reciprocals of height in cm and weight in kg, with a correlation coefficient of 1.00. The result was comparing with that from another whole-body counter that used 8 detectors in a static mode. (author)