Astron. Astrophys. 356, 873-887 (2000)

4. Observational data

As shown in Table 1, the observational data were selected from various high-resolution studies. Elements marked with "c" were adjusted to the same solar abundance scale by S. G. Ryan (Ryan et al. 1996and private communication), entries marked with "x" remained unaltered. Unmarked entries were not observed in the corresponding study. All observations of very metal-poor stars published after 1989 were taken into account. In the case where multiple observations of stars exist, we used the most recent data. If they were published in the same year or after 1995, the values were averaged.

Table 1. Reference list of the observational data. Elements marked with "c" were adjusted to the same solar abundance scale by S. G. Ryan (Ryan et al. 1996and private communication), unaltered observations are marked with "x". Unmarked entries were not observed in the corresponding study. Newer publications (after 1995) are listed separately.

Typical abundance errors given for these observations are about 0.1 dex (see Ryan et al. 1996). These may not fully account for systematic errors, which could, e.g., be caused by the choice of the employed stellar atmosphere models and parameters, or by the employed solar abundance values. If there are systematic offsets between different subsamples, this could enhance the scatter in the combined sample.

Special attention has to be paid to Cr, Mn and O. In the case of Cr and Mn, data published after 1995 show a decrease in the [Cr/Fe] and [Mn/Fe] ratios for lower metallicities. Such trends are not present in older observations. Regardless of this, it is not possible to reproduce metallicity-dependent trends with our model, since we only use the stellar yields of Thielemann et al. (1996), which assume constant solar progenitor star metallicity. A possible explanation of the [Cr/Fe] and [Mn/Fe] trends is the dependency of stellar yields on the metallicity of the progenitor star of a SN Type II (Samland 1997) or a progenitor mass dependent mixing of SN II yields with the ISM (Nakamura et al. 1999).

In the data set of Israelian et al. (1998), the [O/Fe] ratio increases at lower metallicities. This behaviour is apparently different from previous determinations and is not yet understood, but almost identical results were published by a different group of observers (Boesgaard et al. 1999). The -elements oxygen and magnesium are produced mainly during the hydrostatic burning phase of a high mass star and are only slightly affected by the actual explosion. Furthermore, their yields depend in almost the same way on the mass of the progenitor star (Thielemann et al. 1996; Woosley & Weaver 1995). Thus observations of Mg for the objects discussed in the O abundance determinations by Israelian et al. (1998) and Boesgaard et al. (1999) would be very interesting.

© European Southern Observatory (ESO) 2000

Online publication: April 17, 2000
helpdesk.link@springer.de