Li and Rogers describe a small but significant problem in the EGS4 Mon
te Carlo program. EGS4 normally overestimates the electron scattering
power by redundantly taking atomic orbital electron scattering into ac
count in two different ways. This paper shows that the two scattering
models EGS4 uses can be combined consistently by setting the parameter
$FLIDGEMS to a fractional value. For 10 MeV electron beams in water,
inconsistent use of the scattering power can result in differences of
about +/-1.5% in the broad beam maximum dose.