In the analysis a fairly simple pattern recognition algorithm was used.
The time of flight was
taken as known because the real background track will be charged and will
have its time of flight measured by the hodoscope system. The
shower was rejected if any tower within 19.5 cm of the candidate impact point
had a time outside of the correct time by 
ns. The shower was
rejected if any tower neighboring the struck tower (the tower to which the candidate
points) showed an energy
deposit in excess of 
where
where r is the distance between the neighboring tower center and the hit
location, and 
is the energy of the struck tower plus the energies
of any neighboring towers within 1.5 cm of the hit location. The cluster
energy was taken (for the surviving showers) as the sum of all the tower
energies in towers that were within 19.5 cm of the hit location. This
algorithm is rather crude (e.g. it does not make any detailed fit to the shower
shape, or make use of available information on nearby showers), however it
is adequate for the purpose of rejecting background for the charged particle search.
As will be discussed below, this simple algorithm is not adequate for rejecting
low mass backgrounds for neutral particle searches.