A prolonged irradiation facility to study late effects on mice by continuous γ-ray irradiation at low-dose rates was established at the Institute for Environmental Sciences (IES), Aomori prefecture, Japan. Before starting studies of late effects, the absorbed doses to mice were estimated in detail. Three 137Cs γ-sources were individually used to expose mice in three irradiation rooms. The cages for mice were arranged so that the mean absorbed doses in the abdomen of a mouse would reach 8000, 400, 20mGy by exposures for about 400 days in respective rooms. The positions of the cages on the shelves of racks were presumed by using two air-ionization chambers whose ionization volume were about 12 and 1200 ml, respectively. Furthermore, by inserting TLD phosphors into the abdomens of mice, the absorbed doses were directly obtained under the actual exposure conditions.The absorbed dose distributions in air at the source level above the floor never showed an inverse-square law behavior with the distance from the source, and also never showed geometric symmetry in every irradiation room. Except for errors concerning the positions of the racks and cages, the uncertainties in the exposure doses were estimated to be about ±12% for a 8000 mGy group, 17% for a 400 mGy group and 35% for a 20 mGy group.

In a general theory of cavity ionisation T.E. Burlin (Br. J. Radiol., vol.39, p.727-34, 1966) derived the ratio, for an intermediate sized cavity, of absorbed dose in the cavity to that in the surrounding medium in the absence of the cavity. Horowitz and Dubi (1982) and Horowitz et al. (1982) have recently proposed a modification of Burlin's theory and applied their expression to the calculation of gamma dose. In this note Burlin's general cavity theory for photons and Horowitz's modified expression are reconsidered and the theoretical conclusion compared with the experimental results of Ogunleye et al. (1980).