We evaluate the significance of the line candidates identified by a vi
sual search of burst spectra from BATSE's Spectroscopy Detectors. None
of the candidates satisfy our detection criteria: an F-test probabili
ty less than 10(-4) for a feature in one detector and consistency amon
g the detectors that viewed the burst. Most of the candidates are not
very significant and are likely to be fluctuations. Because of the exp
ectation of finding absorption lines, the search was biased toward abs
orption features. We do not have a quantitative measure of the complet
eness of the search, which would enable a comparison with previous mis
sions. Therefore, a more objective computerized search has begun.