Appendix B: Interpolation corrections for the Luck et al. data

If authors do not interpolate their model-atmosphere grids to force equality between model and derived metallicities, small corrections to the latter may be required. Because Castro et al. (1996) discuss such corrections for the Luck et al. results, those corrections receive special attention in this analysis. Castro et al. point out that for the solar-abundance model used by Luck et al., $\log \epsilon({\rm Fe}) = 7.50$ (see Bell et al. 1976). Since the solar abundance used by Luck et al. is 7.67, the effective value of [Fe/H] for their nominal solar model is -0.17 dex. For the purpose at hand, this number is adopted as the zero point of the Bell et al. grid.

Castro et al. also calculate a scale factor for the interpolation correction. That scale factor is set aside here, and a counterpart given by Luck et al. themselves is used (see Table 7 of Luck & Challener 1995). The complete correction may be expressed as follows: if N is an integer and (0.5N - 0.25) dex $\leq$ [Fe/H]$_{\rm LC95}$ < (0.5N +0.25) dex,

$${\rm [Fe/H](corr)} = 1.12{\rm [Fe/H]}_{\rm LC95} + 0.02 + 0.06N.$$ (B1)

This correction is applied before the zero-point adjustments given in Table 3 are calculated. It may be noted that to first order, the effect of the correction is a simple change in zero-point adjustment.