3 Reductions and measurements

Figure 1: Comparison of measurements of $W_{\lambda }$ (Å) in this work and in Feltzing & Gustafsson (1998) for HD 32147 and HD 182572

The spectra were reduced with the standard software available within the CCDRED and ECHELLE packages of NOAO IRAF. The steps included bias subtraction, flat fielding, extraction of one-dimensional spectra, wavelength calibration, and continuum normalization. Additional details concerning the quality of the data resulting from the Sandiford spectrograph can be found in Gonzalez & Lambert (1996) and Gonzalez (1998).

Equivalent widths ( $W_{\lambda }$) were measured using the SPLOT task in IRAF. The lines were measured both by simply integrating the line and also by fitting a Gaussian to the line profile. Most lines were measured twice and some up to four times due to overlap of the spectral orders. As the final adopted value of $W_{\lambda }$ we used the mean of the measurements. In these cases the measurement errors are typically no more than a few percent.

In Fig. 1 we compare the measured values of $W_{\lambda }$ for HD 32147 and HD 182572 with those measured by Feltzing & Gustafsson (1998). For HD 182572 the agreement is good, while for HD 32147, our coolest star, we measure significantly larger $W_{\lambda }$. This difference is most likely due to the lower resolution used in this work. See also the two examples of stellar spectra shown in Fig. 3 from which it is clear that HD 32147, but also to some extent HD 145675, shows a much richer spectrum than the other stars. Since these stars are cool, there will naturally be more molecular lines and low-excitation atomic lines that will cause blending problems.