1 Introduction

The Burst and Transient Source Experiment (BATSE) on board the Compton Gamma Ray Observatory has been very effective in recording gamma-ray bursts (GRB). Soon after the observatory's launch on April 19, 1991, the results obtained with BATSE showed convincingly that the sky distribution of GRBs was isotropic (Meegan et al. 1992a) and that the line-of-sight distribution was incompatible with a homogeneous distribution in euclidean space (Meegan et al. 1992b). The proposition based on these findings that the distance scale of GRBs is cosmological (Paczynski 1992) was eventually confirmed by the observation of a large redshift (Metzger 1997).

The GRB data from BATSE have been published primarily in a succession of catalogs (cf. Meegan et al. 1997). These data are based on an on-board trigger, which acts when certain conditions are fulfilled. Usually these require that the counts in the energy range 50 - 300 keV exceed the background by $5.5 \sigma$ on a time scale of 64, 256 or 1024 msec in at least two of the eight detectors. Variations of these criteria, in the channels and the signal-to-noise limits used, have been in effect at various times (cf. Meegan et al. 1997).

BATSE produces data in various modes in archival form. The DISCLA data provide a continuous record of the counts in channels 1, 2, 3, and 4 for the eight detectors on a time scale of 1024 msec. The DISCLA data allow the detection of GRBs a posteriori in a way similar to the on-board trigger (except for the availability of only one time scale).

There are distinct advantages to searching for GRBs in archival data: the detection parameters can be set independently of those that are hard-wired in the on-board trigger mechanism; the search can be repeated with a different detection algorithm; each search can be carried out on the full data set, etc. A search for GRBs not detected by the on-board BATSE trigger has been conducted by Kommers et al. (1997).

We have used the DISCLA data to search for GRBs in the time period TJD 8365-10528. In the following sections we describe the search, the classification of the triggers, and the derivation of $<V/V_{\rm max}\gt$ of the resulting sample.