We've generated the following MC datasets, all located in /star/institutions/mit/corliss/magic_single_particle/

All of these sets are 5000 events, pt 5-40, phi 0-6.28.  The barrel files range in eta from -.2 to 1.2, the endcap files range from .8 to 2.2.  All contain a single thrown particle in the eta range described, and four muons thrown at eta=-.9 (phi=0,pi/2,pi,3pi/2) to constrain the primary vertex.

electron_barrel.MuDst.root
electron_endcap.MuDst.root
gamma_barrel.MuDst.root
gamma_endcap.MuDst.root
mu_barrel.MuDst.root
mu_endcap.MuDst.root
piminus_barrel.MuDst.root
piminus_endcap.MuDst.root
piplus_barrel.MuDst.root
piplus_endcap.MuDst.root
pizero_barrel.MuDst.root
pizero_endcap.MuDst.root
proton_barrel.MuDst.root
proton_endcap.MuDst.root
single_gamma_endcap.MuDst.root

We've been looking more closely at the electron_barrel set, and are seeing secondary tracks coming from specific regions in the detector:

This first plot is a histogram of the eta of all tracks (primary and not):

Looking at the correlation between eta and pt, we see that the majority of particles in this eta region are fairly low-energy:

When we look at only primary tracks, we don't see this feature, which confirms that it must come from secondary tracks:

Looking at the R vs Z position of secondary vertices, the TPC support seems to be the culprit: