The technique of separating dose into primary and scatter components for calculating photon dose distributions is widely used. The primary and scatter dose models ignore the fact that electrons have a finite range. This may be a good approximation for60Co photons but not for higher energies. For the latter, the range of electrons may be several centimeters in soft tissue and even longer in lungs and will lead to errors in computed dose in regions where electronic equilibrium does not exist. Ignoring the finite range of electrons will affect dose at points such as those near the beam boundaries, near inhomogeneity boundaries, and at bone–soft‐tissue interfaces. Other possible problems associated with the definition and use of “primary” and “scatter” dose in dose distribution calculations result from extrapolation of measured data to obtain data for zero and very large field sizes and from the use of these quantities, which are defined for central axis, for points at large distances from the central axis. This paper examines the limits of the validity of these assumptions.