4.4. Radial Velocity Zero-Point Errors: Dependence on Wavelength

  Because wavelengths were measured individually for more than 700 lines in the 10 Lac atlas, it was possible to search for wavelength errors as a function of wavelength in both the GHRS and IUE/SWP data for 10 Lac. Figure 12 depicts (small crosses) the differences between the laboratory lines referred to in the Brandt et al. (1998) study and the measurements they made from their atlas lines (these differences have been corrected for the radial velocity of the star). We also plot as a solid line the cross-correlation results between the GHRS atlas and ensemble of SWP IUE spectra at each SWP-order bin (cf. Fig. 9). This line shows mean fluctuations of ± 3 km s^-1. Its excursions from zero are always less than ± 3 km s^-1. From these results we conclude that there are no significant differences between the wavelengths of the GHRS and SWP camera. To some degree, this is not unexpected because both calibrations were made using the same line list. Even so, the calibrations were derived using different algorithms. In particular, the GHRS wavelength solutions were obtained by a block-grouping of lines in the entire spectral domain (Lindler 1993) whereas the IUE solutions were performed for individual echelle orders for wavelengths greater than $\lambda$1350.

Figure 12 shows no trend in the mean velocities for echelle orders grouped above and below the $\lambda$1350 demarcation. We can expect that both random and systematic errors of the more current Pt wavelengths are smaller than the older, heterogeneous sources. For example, one may attribute the increased scatter of points in this figure below $\lambda$1500 to random errors from the older laboratory studies. Similarly, any systematic errors in the older laboratory data would show up as trends away from both the zero-level and the solid line. The dashed line in the plot shows evidence of possible systematics in the laboratory data in the wavelength range $\lambda$ $\lambda$ 1650-1750 and $\lambda$ $\lambda$ 1190-1250. Detailed inspection of the line list from which the GHRS wavelengths are drawn suggests that such descrepancies are not limited to one type of ion. For example, several ions of both light and Fe-like ions are represented in these wavelength regions.