We show the events passing timing and angular cut. After unblinding and
seing no events, these cuts have been loosen in order to check that no
unexpected features are seen (see below). Only the events in the
indicated windows are shown. In all cases, the observed number of
events seems to be compatible with the expected background.