No CrossRef data available.
Published online by Cambridge University Press: 04 August 2017
It is shown that a great breakthrough has occurred in the accuracy of spectroscopic abundance analyses with the introduction of solid state light detectors, such as Reticons and CCDs. Because of uncontrolled systematic errors in photographic photometry, abundances derived from high dispersion photographic spectra can hardly be known with an accuracy better than 0.3 dex. This is well exemplified by the recent finding that the observational scatter is large even in the equivalent widths of the Utrecht Solar Atlas. A fortiori these uncertainties are present in the [Fe/H] stellar abundance Catalogue, chiefly based in its present form on photographic material. For the future the calibration of the [Fe/H] Catalogue with spectra taken with Reticon detectors is recommended. A signal/noise ratio of 300 to 500 is more important than an improvement in spectral resolution with a low signal/noise ratio. Then, the remaining uncertainties in the abundances will mostly reflect inaccuracies in atmospheric parameter determinations of the models and in the assumptions underlying model computations.