3 Results

3.1 Optical variability and period determination

For the stars showing optical variability, we have determined their photometric periods by fitting sine waves to the V-band data and searching for the minimum $\chi^2$. See Cutispoto et al. (1995) for further details. In fact, one of the principal results of this study is the discovery of the photometric variability of the following 15 stars:
HD35114, HD36869, HD41824, HD43162, HD48189, HD71285, HD75997, HD78644, HD96064, HD124672, HD141943, SAO111210, SAO150508, SAO150676, SAO196024.

3.2 Radial velocities

The properties inferred from the strength and characteristics of the Li I (6708Å), H$\alpha$ (6563Å) and Ca II K (3933Å) lines are discussed in full detail in Tagliaferri et al. (1999). Here we present the results of radial velocity (RV) determinations, that we use to ascertain the single or binary nature of the stars in our sample. The 6705Å region contains up to about 20 unblended lines of various strengths suitable for accurate radial velocities measurements as well as the 3935Å region. The 6560Å region is half-filled by the H$\alpha$ line but moderately strong, unblended metallic lines suitable for RV measurements are still available. The spectral regions, resolution and the wavelength of the lines are given in Tables 3 and 4. The individual lines used for the RV computation are listed in Table 4. The RV data reduction was performed by standard procedures within the IRAF package, fitting, by Gaussian profiles, the strongest lines present in each spectrum. The resulting RVs from the individual lines have been averaged and heliocentric correction has been applied. The typical RV accuracy for sharp-lined stars for each line is of the order of $\pm$1 kms^-1. The final results are listed in Table 5.

  

Table 3: Spectral region and resolution

  

Table 4: List of the lines used for RV computations

  

Table 5: Star name, HJD (2449000.0 +), heliocentric RV and error, number of lines used (N), spectral region (SR), notes

 

Table 5: continued

 

Table 5: continued

3.3 Rotational velocities

We also computed the v$\sin$i values from the spectra in the 6705Å region by using the cross correlation task "FXCORR" of the IRAF package. The FWHM of the lines thus obtained can be used to estimate the v$\sin$i (Soderblom et al. 1989). We calibrated the method by using seven stars of known v$\sin$i. The cross correlation method gives reliable results in the 5-60 kms^-1 range. For values higher than 60 kms^-1 the Gaussian fit we used is no longer adequate and the rotational broadening of the lines represent a large fraction of the observed spectral range. For values smaller than 5 kms^-1 the intrinsic lines width is larger than the rotational broadening. The error for high signal-to-noise spectra is of the order of $\pm$2 kms^-1. Further details on the method are given in Tagliaferri et al. (1999). The rotational data are given in Table 2.

3.4 Inferred spectral classification

As in the case of the EXOSAT and Einstein samples, for most of the observed stars the spectral classification is not well-defined. In order to infer or further constrain the spectral type and the luminosity class, we used our multicolour photometry, our estimate of RV and the intensity of the Ca I 6717.7Å line from our high resolution spectra, while the distances were taken from the HIPPARCOScatalogue (Perryman et al. 1997). Color indices of active stars have to be taken prudently when used for spectral classification, as the presence of activity phenomena can modify them by an unknown amount. Cutispoto et al. (1996) developed a method, hereafter referred to as Active Star Colors (ASC), to infer the spectral classification from the observed colors. The ASC method was improved taking into account the effects of stellar activity on the U-B index and considering the luminosity calibration of the HR diagram obtained by using the HIPPARCOSdata (see Cutispoto 1998 and references therein for details). For stars later than K5 we used the luminosity calibration reported by Henry et al. (1998). The ASC method is better suitable for statistical purposes and for single stars. In the case of binary stars, two or more solutions can often reproduce the observed colors. However, if an accurate stars' distance measurement is available, it is possible to exclude most of the multiple solutions. Moreover, by using our spectra we can further constrain the allowed solutions. In particular, we used the Ca I 6717.7Å line as a spectral classification indicator. We first calibrated the equivalent width (EW) values of the Ca I 6717.7Å line vs. B-V. For this calibration we used the stars in our sample that are either single or have a WD companion, which give a negligible contribution to the continuum in the spectral region around the Ca I 6717.7Å line. Following these criteria, we selected 21 stars, 20 of which belonging to our sample plus HD 143937B (B-V=1.30, EW(Ca) = 332, K7V). The latter star is the third component of the visual binary HD 143937, which is part of our sample. The measured Ca I 6717.7Å EW for the stars in our sample are reported in the last column of Table 2, where the value used for the EW vs. B-V calibration are marked with "e". A double value means that we have measured both EWs of the SB2 components. We obtained a very good correlation between the EW and the color index B-V (Fig. 1), being the scatter due either to errors in the measured EWs or to slight differences in metallicity among the stars. From the data in Fig. 1 we get the calibration: $B-V = 0.00381 * {\rm EW(Ca)} + 0.0577$ with a linear correlation coefficient $r\,=\,0.978$. We estimated a statistical error of $\pm 15$mÅ for the EW values. The spectral classification is thus determined with an accuracy of $\pm 2$ spectral subclasses. The use of the Ca I 6717.7Å line EW was particularly important for the classification of the binary stars, allowing us to discern, by choosing the one that better fit the combined EW, between different possible solutions for both the ASC method and the distance. We are confident that the spectral types reported in Table 1 are very reliable, with uncertainties of the order of few spectral subclasses.

  

Figure 1: Plot of the measured Ca I 6717.7 Å line EW (mÅ) against the B-V for a sample of 21 stars. The continuous line is the fit of the data