3 HR 363

  As for HD 191226, the spectroscopic-binary nature of HR 363 (= HD 7351) is already mentioned in the GCSRV, where it is classified as gM2. Six radial-velocity measurements are mentioned in that catalogue, among which 4 were performed at the David Dunlap Observatory (Young 1945) and 2 at the Mount Wilson Observatory (Wilson & Joy 1952; Abt 1970). These two sets of measurements yield average velocities of +5.8 and -1.8  km s$\sp{-1}$, respectively. It is likely that it is this difference which led to suspect the binary nature of HR 363 in the GCSRV. More recently, three radial-velocity measurements were obtained at the E.W. Fick Observatory (Beavers & Eitter 1986), and 15 more by Brown et al. (1990). The latter measurements, covering 1000 d, confirmed the spectroscopic-binary nature of HR 363, and made it clear that its period was quite long.

HR 363 was classified M2S by Keenan (1954), who noted its strong BaII $\lambda 455.4$ nm line, and M3IIS by Yamashita (1967). Keenan & Boeshaar (1980) reclassified it later on as $\rm S3+/2-$ in their revised classification scheme (where the first digit is a temperature index, and the second one is an index of ZrO strength, 2 corresponding to ZrO < TiO). More recently, Sato & Kuji (1990) have classified HR 363 as M2III. These authors stress that, although HR 363 is often considered as an S star, its salient spectral features are those typical of M giants, except for a strong BaII $\lambda 455.4$ nm line. As for HD 191226, a near-infrared spectrum has been obtained for HR 363 with the Aurélie spectrograph at OHP, and yields a M3III spectral type. Finally, regarding its Tc-poor nature, we refer to Little et al. (1987) and Smith & Lambert (1988).

 

Table 3:  Radial velocities of HR 363 = HD 7351. Symbols are as in Table 1

The orbital solution for HR 363 listed in Table 2 is based on 49 CORAVEL radial-velocity measurements covering 1.1 orbital cycle (from 1983 to 1997), to which one older measurement obtained in 1976 has been added (Table 3). This early measurement (obtained at the 1.52-m telescope of Haute-Provence Observatory on the photographic spectrum GA 2881) substantially improves the period determination, since it brings the orbital coverage to 1.65 cycle. Its accuracy is only 0.8  km s$\sp{-1}$, compared to an average of 0.30  km s$\sp{-1}$ for the CORAVEL measurements. It has therefore been attributed a weight of 0.25 in the orbital solution (compared to 1 for the CORAVEL measurements).

 

Figure 2:  The radial-velocity curve of HR 363 (=HD 7351), folded with the orbital period. Filled dots refer to CORAVEL data, and open circles to radial velocities derived from photographic spectra obtained at the Haute-Provence Observatory (1.52-m telescope). Measurements from Brown et al. (1990) and from Beavers & Eitter (1976) are represented by open squares and open triangles, respectively. Only the CORAVEL data, plus one OHP photographic measurement (open circle with an error bar), have been used to derive the orbital solution

The radial-velocity curve, folded with the orbital period, is presented in Fig. 2. This figure presents as well the other measurements (namely the three radial velocities from the Fick Observatory, the 15 measurements from Brown et al., and the 6 radial velocities we obtained from photographic spectra at OHP) that were not used in the orbital solution, since they would degrade its accuracy. These observations are nevertheless compatible with the computed solution. Note that the $\rm O-C$ residuals are significantly larger than the accuracy of the measurements. This jitter is likely due to envelope pulsations or to atmospheric motions, as discussed by Jorissen et al. (1998; see their Fig. 1).

As for HD 191226, the mass function is compatible with a WD companion, since $M_1 = 1.5\pm0.5$ $M\sb{\odot}$ for the red giant implies $M_2 \gt 0.70\pm0.15$ $M\sb{\odot}$ for the unseen companion, given $\sin i \le 1.0$. If the companion is to be a WD with a mass typical of field WD's (0.58 $M\sb{\odot}$; Reid 1996), the orbital inclination has to be close to 90$^\circ$, and the system may be an eclipsing binary.

This star has been observed with the International Ultraviolet Explorer (IUE) and ROSAT satellites. Although there is only marginal evidence for an UV continuum from a hot companion (Ake et al. 1988), HR 363 is a strong source of hard X-rays (Jorissen et al. 1996). These X-rays are not expected to come from a hot corona, because with B-V = 1.7, HR 363 lies far to the right of the region of the Hertzsprung-Russell diagram populated by class III giants with a hot corona (Hünsch et al. 1996). The hard X-rays in HR 363 are therefore likely powered by mass transfer in the binary system. The same holds true for the weak HeI $\lambda 1083.0$ nm emission line observed in HR 363 (Brown et al. 1990), since that line is generally absent in cool M giants, but is frequent in interacting binary systems like symbiotic stars (Brown et al. 1990 and references therein). It is somewhat surprising, though, that HR 363 behaves as an interacting binary system while having the longest known period (4593 d = 12.6 y) among S stars (Jorissen et al. 1998). That question is discussed in more details in Sect. 4.