We report evidence from the Third BATSE Gamma-Ray Burst Catalog that long s) and (T 90 [ 10 short s) gamma-ray bursts (GRBs) represent distinct source populations. Their spatial distribu-(T 90 \ 10 tions are signiÐcantly di †erent, with long bursts having but short bursts SV /V max T \ 0.282^0.014 having di †ering by 0.103^0.024, signiÐcant at the 4.3 p level. This implies SV /V max T \ 0.385^0.019, di †erent spatial origin and physical processes for long and short bursts. Long bursts may be explained by accretion-induced collapse. Short bursts require another mechanism, for which we suggest neutron star collisions. These are capable of producing neutrino bursts as short as a few milliseconds, consistent with the shortest observed timescales in GRBs. We brieÑy investigate the parameters of clusters in which neutron star collisions may occur, and discuss the nuclear evolution of expelled and accelerated matter.

We report evidence from the 3B Catalogue that short (T90 < 10 s) and long (T90 > 10 s) GRB represent different populations and processes: Their spectral behavior is qualitatively different, with short bursts harder in the BATSE range, but chiefly long bursts detected at higher photon energies; V /Vmax = 0.385 ± 0.019 for short GRB but V /Vmax = 0.282 ± 0.014 for long GRB, differing by 0.103 ± 0.024. Long GRB may be the consequence of accretion-induced collapse, but this mechanism fails for short GRB, for which we suggest colliding neutron stars.

No abstract