Vertex Miscreconstruction in Z->ee Events
                                  S. Glenn

   The fraction of multi-vertex events in which the primary vertex was 
misidentified was estimated using Z->ee candidates from a sample corresponding
to about 60/pb.  Loose kinematic and quality cuts were applied to the W/Z group
dielectron sample: isolation<0.15, H-Matrix chi-square < 200.0, ET>15 GeV.  The
resulting subsample had an average of 2.0 vertices per event, and the number of
multi-vertex events was roughly equal to the number of single vertex events.
   In order to identify events with misidentified primary vertices, a 
cluster/track matching algorithm was used which ignored all vertex information.
Each track in the event was projected to the cluster centroid in EM3 and the
azimuthal and longitudinal distances of closest approach were computed.  For
clusters in the both in the CDC/CC and FDC/EC, these distances were required to
be less than 10 cm for a good cluster/track match.  For events in which two
clusters had matching tracks, the best match was project to the beamline using
the polar angle measured by the respective tracking systems.  The distance of
the average of these two projections from the vertex was used to discriminate
between "correct" and "incorrect" primary vertices.  (In a subsample of CC/CC
events containing only one vertex, this techihique yielded an RMS of 3.9 cm 
for the projected distance from the vertex.)  Events in which the projected
distance from the first vertex was greater than 10 cm and the distance to the
second vertex was less than 10 cm were considered to be misreconstructed
events.
    For CC/CC events with two or more vertices, the dielectron mass was
recalculated using polar angles measured with respect to the second vertex.
The misreconstruction probability was then calculated by dividing the number of
Z's with misreconstructed vertices by the total number of Z's in the
multivertex sample.  This yielded a primary vertex misreconstruction 
probability of 23%.  This number is valid only for Z->ee events and there is
some evidence for jet multiplicity dependence.