Previous
abstract 
Next abstract
The availability of a digital version of the solar line spectrum (Kitt Peak Preliminary Solar Atlas, Brault \& Testerman 1972) has made it possible to carry out detailed comparisons of observed and synthetic spectra. The more accurately the spectrum of the Sun, and other standard stars, can be reproduced, the more likely the line list is to give reliable results in other applications. Detailed comparisons have been made using three lists. The first two are: 1) One which has been used repeatedly by the author and collaborators e.g. Bell, Dickens \& Gustafsson (ApJ,229,604,1979); Tripicco \& Bell (AJ,103,1285,1992); 2) One derived from Kurucz (Stellar Atmospheres: Beyond Classical Models, Kluwer, Dordrecht, p408,1991) for elements between Ca and Ni, supplemented with lines for other elements from Kurucz \& Peytremann (SAO Spec Rept 362,1975) and molecular lines from the author's list (e.g. Bell \& Gustafsson MNRAS,236,653,1989). The Kurucz list predicts many lines in the solar spectrum which are either not seen or are observed to be far weaker. The errors in oscillator strength may exceed a factor of 10. On the other hand, there are not a corresponding number of lines which are observed but which are not present in the synthetic spectra. Needless to say, this excess in the computed line absorption will affect the calculation of both model atmospheres and synthetic magnitudes. For example, the computed U-B colors will be too red. In view of these errors, and the much better fit which spectra calculated using the author's lists give to the solar line spectrum, the Kurucz list has been used only to fill in gaps in the author's list, thereby creating a third list. This list also incorporates new laboratory gf values (e.g. O'Brian et al. JOSA B,8,1185,1991). Detailed comparisons of observed and synthetic solar spectra from the different lists are shown.
