I have been running a code to simulate energy deposition by xrays in the 50 keV - 10 MeV energy
range for a medical physics problem.  Up until now, I have always run the beam in along the z-axis.
Yesterday, I introduced variable direction beams (so I could sum over multiple doses) and got a
surprise.  The energy deposition around beams not running along te z-axis is very asymmetrical with
the +z axis being preferred to the -z axis in G4PenelopeCompton events.  Changing
G4PenelopeCompton to G4LowEnergyCompton produces the expected result with the energy deposition
symmetrical about the beam axis.

I have posted images of the energy deposition for each case at the URL listed below. The files of
interest are image_lowenergy.jpg and image_penelope.jpg.  The problem is immediately visible in the
Penelope image.

 I will try to start poking around PenelopeCompton for this, but someone who is really familiar with the
Penelope physics package might have a better shot at finding it quickly.