The E864 trigger has two separate modes of operation. The first is optimized for the study of high-mass objects, while the second is for the study of low-mass objects.
First, a pretrigger will be imposed which requires an interaction in the target with sufficient multiplicity to indicate a ``central'' interaction. This will be a common pretrigger to both high-mass and low-mass searches.
A ``late energy'' trigger for the high-mass study will be used to
detect mid-rapidity particles depositing 4 - 5 GeV or more in the
calorimeter. The actual requirement will be that at least 3.3 GeV must be
deposited in the peak tower with a time-of-flight at least 2 ns later than a
v=c particle. Any tower giving such a late energy signal will trigger the
system. HIJET/GEANT simulations determined that 3.3 GeV is the appropriate
cutoff to avoid triggering on commonly-produced central nucleons, while
antideuterons, strangelets, and light nuclei with 
will be
efficiently found by this trigger. Monte-Carlo simulations indicate that such
a late energy trigger will be sufficiently selective so that the system can
operate with an interaction rate of 
Hz.
Low-A nuclei and the multiply-strange dibaryons will not satisfy the late energy trigger because of their small energy deposition. However, such events are expected to be produced at substantial rates on the scale of the E864 sensitivity. We thus plan to run in the low-mass mode with the pretrigger requirement alone.
This two mode strategy will work well if the D/A system can accept roughly 4000
events per spill (spill time 
1 s). A D/A system whose capacity
exceeds this requirement is currently taking data in Fermilab experiment E791
(some members of the Yale group collaborate on E791). In fact, the D/A system
for E864 requires only about 10% of the capability of the Fermilab E791
system. For our sensitivity calculations we have assumed a system with the
required capacity will be available.
A considerable study of backgrounds has been carried out and is outlined in the section on backgrounds. We note that the Bryman committee analyzed our work carefully, and studied the large body of documentation we provided. The committee agreed with our conclusions, and made several helpful criticisms which led to useful changes in the details of our design.
One of the key features of E864 is the simultaneous measurement of the particle trajectories in space and time. Each photomultiplier tube is connected to a TDC and to an ADC, so that time and pulse height information are both available. Each scintillation counter thus provides a unique space-time point on the track. The horizontal coordinate is determined by the counter number, the vertical coordinate by the time difference between the top and bottom tubes, and the time-of-flight by the mean time of the two tubes. We chose the size of the counters to minimize the occupancy rate in each slat. If two particles strike the counter, the resulting time and vertical position will be in error for either of the incident particles. It is also required that the pulse heights in each plane agree within error.
The requirement that the same hits make a good space track AND a good time track, together with the redundancy provided by three hodoscope measurements, means that the only backgrounds of consequence are those which somehow produce a real track with a velocity within the range of interest for E864.
We have found (with Monte Carlo studies) that the largest background is due
to neutrons which interact in the first or second straw tube plane (or the
vacuum window) which produce a proton which makes the ``good'' track and
which has a velocity close to that of the incident neutron. Such an event
must occur in accidental coincidence with an overlapping cluster of neutrons
which strike the same tower that the proton happens to aim at. This
background enters at about the 
level (of the total reaction cross
section). Further, there are several handles left in the analysis which
should reduce it further.
For the purely neutral states our analysis is not quite as far advanced, and
in general the background will enter at a higher level. The more massive the
neutral system, the smaller the background. For the H-dibaryon we have
estimated that the background will enter at the 
level. However,
according to coalescence estimates, which should give fairly reliable lower
limits to the cross section, the H should be produced at about the
level or greater. This would yield a very large number of H particles in
our experiment, and signal-to-noise of the order of unity should be adequate.
There are further analysis techniques (involving recognizing overlapping
neutrons) which should further improve the signal-to-noise. For masses 
GeV, the neutral background level is below 
. More complete
studies are currently in progress.