The separation of the total scatter correction factor Sc, p in a collimator scatter component, Sc, and a phantom scatter component, Sp, has proven to be an useful concept in megavoltage photon beam dose calculations in situations which differ from the standard treatment geometry. A clinically applicable method to determine Sc is described. Measurements are carried out with an ionization chamber, placed at a depth beyond the range of contaminant electrons, in a narrow cylindrical polystyrene phantom with a diameter of 4 cm of which the axis coincides with the beam axis. Sc, p is measured in a full-scatter phantom and Sp can be derived from Sc, p and Sc. In order to obtain a reliable separation, i.e. excluding the influence of contaminant electrons, measurements of Sc, p have been carried out at depths of 5 cm for photon beams with a quality index (QI) up to and including 0.75 and a depth of 10 cm with QI larger than 0.75. These depths are in accordance with recommendations given in recent dosimetry protocols. The consistency of the method was checked by comparing calculated and measured values of Sc, p for a set of blocked fields for a range of photon beam energies from 60Co up to 25 MV showing a maximum deviation of 2%. The method can easily be implemented in existing procedures for the calculation of the number of monitor units to deliver a specified dose to a target volume.