Previous
abstract 
Next
abstract
A bimodal burst duration distribution is now confirmed by BATSE, with the minimum near 1.5 s, and with long bursts having significantly lower integral hardness ratios than do the short bursts (Kouveliotou et al. 1993, ApJ. 413, L101). Yet other analyses are consistent with the two groups having the same parent population: their celestial distributions are isotropic, their number vs. intensity relationships are similar, and they have a common range in peak intensity.
The appearance of a bimodal duration distribution -- with all bursts arising from the same process -- can be constructed if there is a non-zero mode in the distribution of intervals between major pulses and if a substantial fraction of short bursts consist of a single pulse. Measurements at 64-ms resolution indicate that, for the longer bursts, the interval distribution exhibits a broad maximum around 1 s. From examination of short bursts with higher timing resolution we determine that approximately one-third of short bursts are representable by single pulses. Not all short bursts are required to be single outbursts since the duration distribution is partially ($\sim$50\%) filled in between the shortward and longward modes. The difference in spectral hardness ratios (HR) could be explained if the onset -- rather than integral -- HR is calculated for the longer bursts. In fact, the highest HR ($\approx$1.3) measured from the average trigger-aligned profile of long bursts occurs near event onset, and since peaks tend to be harder, a hypothetical ``peak-aligned near onset" HR might be comparable to the integral HR ($\approx$1.5) found for short bursts. \
