6 $\zeta$ And = HR 215 = HD 4502

6.1 Brief history

Campbell (1911) was the first to report the variable velocity of $\zeta$ And and also commented on the obvious broadening of its absorption lines. Shortly thereafter, Cannon (1915) computed the initial set of orbital elements for this bright star. Spencer Jones (1928) and Gratton (1950) obtained additional velocities, and each determined a new orbital solution. The latter's remained the standard for nearly 50 years. Recently however, Kaye et al. (1995) recomputed a new orbital solution from 135 velocities from the literature.

The CaII emission lines of $\zeta$ And were independently discovered by Joy & Wilson (1949) and Gratton (1950). The latter also noted that the absorption lines are diffuse, suggesting a noticeable rotational velocity. From new spectrograms Hendry (1980) discussed the variability of the emission lines.

Light variations, initially discovered by Stebbins (1928), are primarily the result of the ellipticity effect (e.g. Stebbins 1928; Strassmeier et al. 1989). Variability due to starspots with a period similar to the orbital period and an amplitude as great as 0$.\!\!^{\rm m}$04 has also been detected (Kaye et al. 1995). We note that the maximum spot amplitude of 0$.\!\!^{\rm m}$083, highlighted by Kaye et al. (1995), appears to be a misprint.

6.2 Orbital elements

For $\zeta$ And two new sets of velocities have been obtained, 53 velocities at NSO during an extensive 1996-1997 observing run, and 14 velocities at KPNO during a single run in 1997-1998. With the orbital period fixed at Gratton's (1950) value and his other elements used as starting values, independent orbital solutions of the two data sets were carried out with the SB1 computer program. The results indicated no need for a zero-point shift, and the two sets were combined with weights of 1.0 for the KPNO velocities and 0.5 for the NSO velocities. Independent solutions were also computed for the large data sets of Cannon (1915) and Spencer Jones (1928), resulting in weights of 0.02 and 0.03, respectively, and a velocity offset of 6 kms^-1, which was added to the data of Cannon (1915). Additional sets of velocities in the literature, including those of Gratton (1950), do not warrant inclusion. They generally contain just a few velocities and have very low weights, providing no additional leverage on the period or other elements.

An SB1 orbital solution with the four sets of appropriately weighted velocities resulted in an eccentricity of $0.0068\pm 0.0056$. Because of this very low eccentricity, a circular-orbit solution for the same data was computed with SB1C. According to the precepts of Lucy & Sweeney (1971), the circular orbit is to be preferred. The two old sets of velocities have such low weights that they do not significantly improve the other elements. Thus, with the period fixed from the previous circular-orbit solution and two NSO velocities given zero weight because of their large residuals, a final circular orbit was computed with only the NSO and KPNO data. Those orbital elements are listed in Table 5. Since the eccentricity is zero, the time of periastron is undefined and has been replaced with the time of maximum positive velocity.

Our velocities and residuals to the computed fit are given in Table A5 in the electronic Appendix. The computed radial-velocity curve is compared with our velocities in Fig. 4. The standard error of an observation of unit weight is 0.6 kms^-1, quite good for such a broad lined star.

 

Figure 4: Radial-velocity curve of $\zeta$ And. Open symbols are NSO McMath-Pierce data, filled symbols are from KPNO. Zero phase is the time of maximum positive velocity