4 Discussion

We intended a classification of the red component, based on the fact that the TiO bands are good indicators of the spectral type (Kenyon & Fernandez-Castro 1987). We measured the [TiO]₁, [TiO]₂and [VO] indices defined by Kenyon & Fernandez-Castro (1987), obtaining 0.115(5), 0.37(0) and 0.22(0) respectively. These indices yield spectral types of M3, M5 and M4, using Eqs. (5-7) of Kenyon & Fernandez-Castro (1987). The spectral type $\sim$M4 is close to the mean of a sample of 100 symbiotic stars (see Mürset & Schmid 1999).

The absence of Ca II absorption and other luminosity indicators (e.g. Na II and Fe I absorption lines, Zhu et al. 1999) makes any estimation of the luminosity class almost impossible. However, in the case where the secondary is a giant, like in most symbiotic stars, the calcium triplet in emission could not be understood as chromospheric activity, as sometimes occurs in dwarfs. An alternative explanation is an origin in the outer cool parts of an accretion disk as seen in some cataclysmic variables (e.g. Persson 1988).

We can use the emission line ratios to estimate the interstellar extinction. Assuming a theoretical ratio $F_{{\rm H}\alpha}/F_{{\rm H}\beta}$ = 2.8 (nebulae case B, Osterbrock 1974), we derive, from the observed ratio of 4.13, a selective extinction E(B-V) = 0.42 (see Seaton 1979 for details of the method). This figure yields A_V = 1.33. This deduced extinction is similar to the total galactic extinction in the direction of V589 Sgr ( $N_{\rm H} = 1.8~10^{21}$ Dickey & Lockman 1990), when using the conversion according to Predehl & Schmitt (1995). This fact implies that V589 Sgr is beyond all the galactic column. At a galactic latitude of $b_{\rm II} = -6.5$ deg this yields a distance larger than 1 kpc. Based on the intensity in the spectra we deduce a brightness of $m_{V} \approx 15.6$ mag, corresponding to a stage of intermediate brightness. The distance, inferred by assuming M_V = -0.3 for a M4 giant and the extinction above, is d = 8 kpc. On the contrary, when assuming a dwarf companion with $M_{V} \sim 11.3$ mag, the deduced distance would have to be of the order of 40 pc only, inconsistent with the deduced extinction of the Balmer lines.

A check of the ROSAT All-Sky survey archive reveals no X-ray source at the position of V 589 Sgr. This corresponds to an unabsorbed flux limit of < 10^-12 erg cm^-2 s^-1 in the 0.1-2.4 keV range. Given the above extinction, this is no surprise.