2 HD 191226

  The spectroscopic binary nature of HD 191226 is already mentioned in the General Catalogue of Stellar Radial Velocities (GCSRV; Wilson 1953): from 10 measurements obtained at the Mount Wilson and Victoria observatories, the radial velocity of this gM2 star varied by 22  km s$\sp{-1}$. Also classified as M2III by Nassau & McRae (1949), whereas Barbier (1962) assigns it the spectral type K2II:, HD 191226 is finally recognized by Keenan & Boeshaar (1980) as being a weak S star (M1S-M3SIIIa). The spectral class MxS refers to stars with the strongest ZrO bands barely visible. A spectrum obtained in the near infrared at a dispersion of 3.3 nm/mm with the Aurélie spectrograph on the 1.52-m telescope at OHP yields a M0-M1III+ classification from a comparison with spectral standards (Ginestet et al. 1994; Carquillat et al. 1997). This spectrum is unfortunately of no use to confirm the S-type classification of HD 191226, as the typical S spectral features are better seen in the optical domain (Jaschek & Jaschek 1987). The Tc-poor nature of HD 191226 was reported by Little et al. (1987) and Smith & Lambert (1988).

 

Table 1:  Radial velocities of HD 191226. Column 4 ($\epsilon_1$) lists the uncertainty on the individual measurements. The columns labelled "Phase'' and "$\rm O-C$'' refer to the orbital solution listed in Table 2. The column labelled "Obs'' provides the origin of the radial velocity: COR = CORAVEL; Cou = Coudé spectrum obtained at OHP; McD = data obtained by Brown et al. (1990) at McDonald observatory

 

Table 2:  Orbital elements of HD 191226 and HR 363

The orbital elements of HD 191226 have been derived from the 36 radial-velocity measurements listed in Table 1. Among these, 31 were obtained with the CORAVEL spectrovelocimeter at OHP (with an average uncertainty of $\bar{\epsilon} = 0.34 \pm 0.06$ km s$\sp{-1}$). One has been obtained at the Coudé focus of the 1.52-m telescope at OHP, on a baked IIaO plate (# GA 8139), with a 2.0 nm/mm dispersion. Four radial-velocity measurements obtained by Brown et al. (1990) on the 2.1-m telescope at McDonald Observatory equipped with a Reticon detector have also been used in the orbital solution. All these radial velocities, covering 3.37 orbital cycles, are on the system of IAU standards. The orbital elements listed in Table 2 have been computed with the BS1 program (Nadal et al. 1979) by assigning a weight 1 to the CORAVEL data and 0.25 to the five other measurements. The radial-velocity curve, folded with the orbital period, is presented in Fig. 1. Older data, obtained through the years 1912-1934, were not used in the orbital solution, because their accuracy is not good enough, even to improve the orbital period. The Mount Wilson data have (when available) uncertainties of the order of 2.5 to 3.3  km s$\sp{-1}$, and their variation range (-13 to 35  km s$\sp{-1}$;Abt 1973) appears incompatible with the small value of the semi-amplitude (K = 4.8  km s$\sp{-1}$;Table 2) derived from modern, more accurate measurements. As far as the the radial velocities obtained at Victoria by Harper (1934) are concerned, they are of doubtful quality (both plates are qualified as "weak'') and were therefore not used either.

 

Figure 1:  The radial-velocity curve of HD 191226, folded with the orbital period. Filled dots refer to OHP data (CORAVEL and Coudé), and open dots to McDonald data

An estimate of the mass of the companion can be obtained from the mass function listed in Table 2. Adopting a mass of $1.5\pm0.5$ $M\sb{\odot}$ for the S star, which seems reasonable given its M2III spectral type (e.g., Schmidt-Kaler 1982; see also Jorissen et al. 1998), a minimum mass of $0.35\pm0.07$ $M\sb{\odot}$ for the secondary follows from the condition $\sin i \le 1$. Taking M₁ = 1.5 $M\sb{\odot}$ for the red giant and M₂ = 0.6 $M\sb{\odot}$ for its companion yields $i = 40^\circ$,q = M₂/M₁ = 0.4 and A₁ = 121 Gm, or in terms of orbital separation, A = A₁ + A₂ = A₁ (1 + 1/q) = 2.8 AU.

The companion mass is thus compatible with that of a white dwarf (WD). Observations in the UV with the Goddard High Resolution Spectrograph (GHRS) on board the Hubble Space Telescope indeed reveal the presence of a WD with $T \sim 15\thinspace000$ K (Ake 1997). These results will be discussed further in Sect. 4.