Machine outputs, peak scatter factors, and central axis percentage depth dose distributions were measured for various phantom sizes in large radiation fields produced at extended distances by cobalt, 6-MV, and 10-MV photon beams. The results can be applied to practical total body irradiation procedures which usually involve treatment volumes smaller than the actual field sizes in order to provide a uniform total body exposure to radiation. Our study addresses the question of the appropriate phantom dimension to be used in the calibration of photon beams employed in total body irradiations. The measurements show that the machine outputs are only slightly dependent on phantom size; the percentage depth dose distributions, however, are strongly dependent on the phantom size, suggesting that machine data for total body irradiations should be measured in phantoms whose dimensions approximate the patient during the total body irradiation. Peak scatter factors measured in large-field/small-phantom configurations link up well with the published small-field/large-phantom data. The finite patient thickness lowers the dose to points close to the beam exit surface by a few percent, when compared to dose measured at the same depths in infinitely thick phantoms. The surface doses in large radiation fields are essentially independent of phantom cross sections and range from 40% for the 10-MV beam, to 65% for the 6-MV beam and 80% for the cobalt beam.