We find that the multiple ionization and recombination scenario has difficulties in reconciling the low temperature required to have a fast recombination with the large heating due to the burst flux. On the other hand, we believe that the other two alternatives (i.e. thermal emission and reflection) are promising and not mutually exclusive.
In the reflection scenario, afterglow spectra should show the typical hardening of the spectrum above a few keV, and the line duration should be short.
Copyright The European Southern Observatory (ESO)