6 Conclusion

We have shown that our method of determining detailed abundances for A-F type stars works and that the ELODIE echelle spectrograph can be used to get accurate abundance determinations for Vega and the Sun. We achieve a high level of automation to extract spectra and analyze them. One line list covering the ELODIE spectral range was compiled and will be used for the following study of A-F stars. It includes new gf values for some critical lines determined using the solar spectrum.

The problem of the choice of stellar parameters which arose during the analysis of Vega is also a good justification to analyze a large sample of stars with one given method for determining stellar parameters ( $T_{\rm eff}, \log{g})$. The choice between a photometric or spectroscopic method is not so important since the uncertainty of these methods is comparable. Next it is important to determine abundances of all stars of the sample in a homogeneous way, and that will be possible with the method presented in this paper. This is our final goal for which automation will be crucial. Therefore, even if some uncertainties remain, the resulting errors will be systematic, and will not depend on the author's subjectivity.

Finally, our work is further justified by the commissioning of medium and high resolution multi-fibers spectrographs, because when an observer gets hundreds of spectra each night, he can no longer handle them by hand.

Acknowledgements

We thank Dr. G. M. Hill for providing the main program, and for his help in the early stages of development We thank Dr. Y. Chmielewski for useful and instructive discussions and for giving us some subroutines that proved very useful. We are grateful to Dr. R. O. Gray for his availability, his kindness and the pertinence of his answers to our numerous questions. Finally, the constructive comments of the referee, Dr. J. Landstreet, are gratefully acknowledged.