Although most recent interest and development in radiation dosimetry has been centered on higher energy radiations, calibrations of conventional x-ray machines for radiotherapy are still based on free-air ionisation chamber measurements. Such measurements require a correction for ionisation produced by secondary photons scattered out of the primary beam, and although the correction is relatively small (generally less than 1%) the data for deriving it are neither as complete or satisfactory as might be wished. This paper describes the derivation of theoretical expressions which allow the calculation of corrections, and compares calculated values with experimental results. It is apparent from the study that in determining the photon gain correction, account should be taken not only of the chamber geometry but also of photon scattering from the wall materials and, for higher energy beams in particular, of contributions from diaphragm scatter and possibly stray radiation penetration of the chamber.