Small < 10 m s−1 variations of radial velocity (RV) with multi-year periods in solar-type stars may be indistinguishable from the effects induced on lines by stellar activity cycles (Dravins 1985; Saar & Donahue 1997). Dravins (1992) recommended a resolving power R > 3 × 105 to measure accurately the subtle changes in the shapes of bisectors of photospheric absorption lines driven by changes of granular convection in slowly rotating dwarf stars. Butler et al. (1996) measure impressively small amplitudes of RVs by using echelle spectrographs that cover a broad spectrum. However, to cover a broad spectrum the resolving power is typically limited to < 7 × 104, and the necessary presence of the iodine absorption spectrum may make it difficult to measure convective line shifts contemporaneously with the RV time series. Furthermore, to reach an RV accuracy of ± 3 ms−1 the whole profile of each line is used, thus maximizing the possibility that changes in the shapes of the lines’ C-bisectors could induce an apparent variation of RV.
Dravins (1985) recommended the exclusive use of the steep flanks of photospheric absorption lines to minimize the effects of convection on apparent RV. McMillan et al. (1993, 1994) demonstrated that such RV measurements made with a Fabry-Perot etalon (FPE) interferometer in transmission can be stable in the presence of stellar line variations seen by other investigators whose measurements were not based exclusively on line flanks. Dravins also prescribed high resolving power, high signal-to-noise ratio, high instrumental contrast, and low instrumental wings to analyze the rest of the line profile for convectively-driven changes (Dravins 1978, 1987, 1992). A double- or multiple-pass FPE scanning whole line profiles can provide high R, high contrast, low wings, and a stable, symmetrical line spread function with small (portable) optics, although the low photon efficiency will restrict its use to a few carefully selected stars. The spectral classes of these stars should span the spectral classes of the stars being monitored by others for planets. We describe a possible implementation of this concept that has the potential for adequate photon flux: observing symbiotically through another instrument on a 6-m to 10-m class telescope.