4 Individual results

In this section we describe the behaviour of the above mentioned chromospheric activity indicators for each star of the sample. The profiles of H$\alpha$, H$\beta$, Ca  II H & K and Ca  II IRT are displayed from Figs. 1 to 23. For each system we have plotted the observed spectrum (solid-line) and the synthesized spectrum (dashed-line) in the left panel and the subtracted spectrum (dotted line) in the right panel. The name of the star, the observing run (NOT96, NOT98, McD98, INT99) and the orbital phase ($\varphi$) of each spectrum are given in every figure. The He  I D₃ line, for selected stars of the sample, is displayed in Fig. 24.

4.1 UX Ari (HD 21242)

This double-lined spectroscopic binary (G5V/K0IV) is a well known RS CVn system and extensively studied in the literature (Carlos & Popper 1971; Bopp & Talcott 1978; Huenemoerder et al. 1989; Raveendran & Mohin 1995). Recently, Duemmler & Aarum (2000) have given a new orbit determination, which we have adopted in Table 2. Our previous H$\alpha$ observations (Montes et al. 1995a,b) showed clear H$\alpha$ emission above the continuum from the cool component. This emission was superimposed to the weak absorption of the hot component. Our spectrum in the Ca  II H & K region (Montes et al. 1995c) showed strong emission from the cool component and a weak H$\epsilon$ emission line. We also detected a flare in this system through simultaneous H$\alpha$, Na  I D₁, D₂and He  I D₃ observations (Montes et al. 1996b; Paper I).

In the new observation (NOT96) we observe intense emission in the Ca  II H & K, H$\epsilon$, H$\alpha$ and Ca  II $\lambda$8662 lines and a filled-in absorption H$\beta$ line from the cool component (see Fig. 1). The excess emission measured in all these activity indicators is larger than in our previous observations of this system in quiescent state in 1992 and 1995. The small emission we observe in He  I D₃ (Fig. 24) confirms the high level of activity of UX Ari in this observation. The He  I D₃ line has been observed as clear emission in this system in other occasions associated to flare-like events (Montes et al. 1996b and references therein).

4.2 12 Cam (BM Cam, HD 32357)

This single-lined spectroscopy binary was classified by Bidelman (1964) as a K0 giant. He also noticed Ca  II H & K emission. Later, Hall et al. (1995) revised the system spectroscopically and photometrically and obtained new values of orbital parameters, given in Table 2. Eker et al. (1995) observed that H$\alpha$ profiles showed an asymmetric shape with a round shoulder in the red wing and a steeper blue wing. They also confirmed the variable H$\alpha$ filling, which was suspected by Strassmeier et al. (1990) too. In our previous observations of this system (FFMCC) we found the H$\alpha$ line filled-in and strong Ca  II H & K emission.

In the new spectrum (Fig. 2) we find strong emission in the Ca  II H & K lines. Thanks to the higher resolution of this spectrum it is possible to see now that both lines exhibit self-absorption with red asymmetry. Small H$\epsilon$ emission is observed, but it was impossible to deblend it from the Ca  II H line. Both H$\alpha$ and H$\beta$appear in absorption with a slight filling-in, the H$\alpha$ filling-in is smaller than the one corresponding to other epochs. The H$\alpha$ line shows excess absorption in the red wing similar to the asymmetric shape of H$\alpha$ observed by Eker et al. (1995). The application of the spectral subtraction technique reveals that the He I D₃line appears as an absorption feature (Fig. 24). This fact is more frequent in giants than in dwarfs (Paper I). Finally, the Ca  II IRT ($\lambda$$\lambda$ 8542, 8662 Å) absorption lines are clearly filled-in.

4.3 V1149 Ori (HD 37824)

This single-lined spectroscopic binary, classified as K1III + F by Bidelman & MacConnell (1973), was listed as a G5IV star by Hirshfeld & Sinnott (1982). Our previous observations revealed clear excess H$\alpha$ emission (Montes et al. 1995a,b; Paper I), strong Ca  II H & K and H$\epsilon$ emission (Montes et al. 1995c) and clear absorption in the He  I D₃ line in the subtracted spectrum (Paper I).

In the new observation (NOT98, Fig. 3), the H$\alpha$ and H$\beta$ lines show a filled-in absorption profile, and clear absorption is observed in the blue wing of both lines. This excess absorption in the blue wing, not observed in our previous observations of this system, could be indicative of variable mass motion. To confirm this behaviour we took a new spectrum with the ESA-MUSICOS spectrograph in January 2000 (forthcoming paper) in which the blue wing of the H$\alpha$ line is in emission, confirming the high variability of the H$\alpha$ line profile in this system. Strong emission is observed in the Ca  II H & K lines but the H$\epsilon$ line is not detected. The Ca  II IRT lines ($\lambda$$\lambda$ 8542, 8662 Å) show a strong filling-in. The He  I D₃ line appears in absorption (Fig. 24).

4.4 OU Gem (HD 45088)

OU Gem is a bright (V = 6.79, Strassmeier et al. 1990) and nearby (d = 14.7 pc, ESA 1997) BY Dra-type SB2 system (K3V/ K5V) with an orbital period of 6.99 days and a noticeable eccentricity (Griffin & Emerson 1975). Both components show Ca  II H & K emission, though the primary shows slightly stronger emission than the secondary. The H$\alpha$ line is in absorption for the primary and filled-in for the secondary (Bopp 1980; Bopp et al. 1981a,b; Strassmeier et al. 1990; Montes et al. 1995a,b, 1996). Dempsey et al. (1993a) observed that the Ca  II IRT lines were filled-in. This binary was detected by the WFC on board the ROSAT satellite during the all-sky survey (Pounds et al. 1993; Pye et al. 1995). OU Gem has 1.7 10²⁹ erg s^-1X-ray luminosity, typical value of the BY Dra systems (Dempsey et al. 1993b, 1997). The photometric variability was discovered by Bopp et al. (1981a) and they also computed a 7.36-day photometric period. It is interesting that the orbital and rotational periods differ in 5% due to the appreciable orbital eccentricity (e = 0.15), according to Bopp (1980). Although BY Dra systems are main-sequence stars, their evolutionary stage is not clear. OU Gem has been listed by Soderblom et al. (1990) and Montes et al. (2000a, 2000c) as a possible member of the UMa moving group (300 Myr), indicating that it may be a young star.

The H$\alpha$ line: In the observed spectra, we see an absorption line for the primary star and a nearly complete filling-in for the secondary star. After applying the spectral subtraction technique, clear excess H$\alpha$ emission is obtained for the two components, being stronger for the hot one (see Fig. 4 upper panel). The excess H$\alpha$ emission EW is measured in the subtracted spectrum and corrected for the contribution of the components to the total continuum. We took one spectrum in this region in Dec-92 (Montes et al. 1995b). At the orbital phase of this observation ($\varphi$ = 0.48) we could not separate the emission from both components and we measured the total excess H$\alpha$ emission EW relative to the combined continuum. We obtained a similar value to Mar-96, Apr-98 and Jan-99 values obtained adding up the excess emission EW from the two components. The H$\beta$ line: Looking at the observed spectra, we only see the H$\beta$ line for the primary, in absorption. After applying the spectral subtraction technique small excess H$\beta$ emission is obtained for the two components (see Fig. 4 lower panel). We have obtained, in general, $\frac{E_{{\rm H}\alpha}}{E_{{\rm H}\beta}}$ values larger than three for the two components, so the emission can come from prominences.

The Ca  II H $\&$ K and H$\epsilon$ lines: We observe that both components of this binary have the Ca  II H & K and H$\epsilon$ lines in emission. We can also see that the excess Ca  II H & K emission of the hot star is larger than the one of the cool star (Fig. 5 upper panel). The measured excess Ca  II H & K emission of both components is larger in the two spectra of the NOT98 observing run than in the NOT96 spectrum. Overlapping between the H$\epsilon$ line of one star and the Ca  II H line of the other only allows to see the H$\epsilon$ line of the cool star at orbital phase 0.19, and H$\epsilon$ of the hot star otherwise.

The Ca  II IRT lines: In the observed spectra, we can see that both components of OU Gem show the Ca  II IRT lines in emission superimposed to the corresponding absorption. After applying the spectral subtraction technique, clear excess emission appears for the two components, being clearly stronger for the hot one (see Fig. 5 lower panel).

4.5 $\sigma$ Gem (HD 62044)

This single-lined spectroscopic binary belongs to the long-period group of RS CVn binary systems. It was classified as K1III, but the radius obtained by Duemmler et al. (1997) seems to be too small for a giant star. In Table 2 we have adopted the orbital and physical parameters updated by Duemmler et al. (1997). Strong and variable Ca  II H & K emission always centered at the absorption line has been reported by Bopp (1983), Strassmeier et al. (1990), FFMCC, Montes 1995, Montes et al. 1996b. Our previous observation in the H$\alpha$ line region (Montes et al. 1995a,b) revealed small excess emission, similar to that reported by Strassmeier et al. (1990) and Frasca & Catalano (1994). Variable excess H$\alpha$ emission anti-correlated with spot regions has been found by Zhang & Zhang (1999).

The new observations (NOT96, NOT98) show strong Ca  II H & K emission lines and small emission in the H$\epsilon$ line. After applying the spectral subtraction, a small filling-in is observed in the H$\alpha$, H$\beta$ and Ca  II IRT lines (Fig. 6). We observe in this system notable He  I D₃ absorption (Fig. 24). All the activity indicators show an increase in the 1998 observation ($\varphi$ = 0.88) with respect to 1996 one ($\varphi$ = 0.81). The emission in the Ca  II H & K lines in these two observations is noticeably larger than in our previous observations at different epochs and orbital phases.

4.6 BF Lyn (HD 80715)

This double-lined spectroscopic binary with spectral types K2V/[dK] shows variable H$\alpha$ emission and strong Ca  II H & K, H$\epsilon$ and Ca  II IRT emission from both components (Barden & Nations 1985; Strassmeier et al. 1989b). In our previous observation (Montes et al. 1995c) we found strong emission in the Ca  II H & K lines from both components with very similar intensity and the H$\epsilon$ line in emission. The orbital period is 3.80406 days (Barden & Nations 1985), and Strassmeier et al. (1989b), from photometric observations, found that BF Lyn is a synchronized binary with a circular orbit.

In the four runs analysed in this paper we have obtained 11 spectra of BF Lyn at different orbital phases. We have used the original heliocentric Julian date on conjunction ( $T_{\rm conj}$) given by Barden & Nations (1985) to calculate the orbital phases since we discovered a mistake in the Strassmeier et al. (1993) catalog where the orbital data from Barden & Nations (1985) are compiled. In the original paper the epoch was determined using Modified Julian Date (MJD), that is why 0.5 days must be added to the Strassmeier et al. (1993) date, who used the 2440000.0 Julian day as a reference date.

The H$\alpha$ line: We took several spectra of BF Lyn in the H$\alpha$ line region in four different epochs and at different orbital phases. In all the spectra (Fig. 7) we can see the H$\alpha$ line in absorption from both components. The spectral subtraction reveals that both stars have excess H$\alpha$ emission. At some orbital phases, near to the conjunction, it is impossible to separate the contribution of both components. The excess H$\alpha$ emission of BF Lyn shows variations with the orbital phase for both components, but the hot star is the most active in H$\alpha$. In Fig. 8 we have plotted for the McD 98 observing run the excess H$\alpha$ emission EW versus the orbital phase for the hot and cool components. The highest excess H$\alpha$ emission EW for the hot component has been reached at about 0.4 orbital phase and the lowest value is placed at about 0.9 orbital phase, whereas the cool component shows the highest excess H$\alpha$ emission EW at near 0.9 orbital phase and the lowest value between 0.2 and 0.4 orbital phases. The variations of the excess H$\alpha$ emission EW for both components are anti-correlated, which indicates that the chromospheric active regions are concentrated on faced hemispheres of both components, but at about 0.4 and 0.9 orbital phases for the hot and cool component, respectively. The same behaviour is also found in Ca  II IRT. The excess H$\alpha$ emission EW also shows seasonal variations, for instance, the values of the INT99 observing run are very different, specially for the cool component, from McD 98 values at very similar orbital phase.

The H$\beta$ line: Five spectra in the H$\beta$ line region are available. In all of them the H$\beta$ line appears in absorption from both components. The application of the spectral subtraction technique reveals clear excess H$\beta$ emission from both stars (see Fig. 9). The ( $\frac{E_{{\rm H}\alpha}}{E_{{\rm H}\beta}}$) values that we have found for this star allow us to say that the emission comes from extended regions viewed off the limb.

The Ca  II H $\&$ K and H$\epsilon$ lines: We took four spectra in the Ca  II H & K region during the NOT (96 & 98) observing runs (Fig. 10). Another spectrum was taken in 1993 with the 2.2 m telescope at the German Spanish Astronomical Observatory (CAHA) (Montes et al. 1995c). These spectra exhibit clear and strong Ca  II H & K and H$\epsilon$ emission lines. At 0.02, 0.43, and 0.54 orbital phases the emission from both components can be deblended using a two-Gaussian fit (see Fig. 10). In the case of CAHA 93 run, the H$\epsilon$ emission line from the hot component is overlapped with the Ca  II H emission of the cool component. The excess Ca  II H $\&$ K and H$\epsilon$ emission changes with the orbital phase during the NOT 98 run in the same way as the corresponding excess Ca  II $\lambda$8542 and H$\alpha$ emission. The excess Ca  II H $\&$ K emission EW also shows seasonal variations, for instance, the values of CAHA 93 observing run are lower than NOT 96 & 98 values.

The Ca  II IRT lines: In all the spectra we can see the Ca  II IRT lines in emission from both components (Fig. 11). As in the case of H$\alpha$, the Ca  II IRT emission shows variations with the orbital phase for both components. In Fig. 8 we have plotted, for the McD 98 observing run, the excess Ca  II $\lambda$8542 emission EW versus the orbital phase for the hot and cool components. The variations of the excess Ca  II emission EW for both components are anti-correlated and they show the same behaviour as the excess H$\alpha$emission EW.

4.7 IL Hya (HD 81410)

IL Hydrae is a typical RS CVn star with very strong Ca  II H & K emission (Bidelman & MacConnell 1973). The 12.86833-day photometric period derived by Raveendran et al. (1982) is very close to the orbital period. It was found to be an X-ray source and a microwave emitter (Mitrou et al. 1996). From a photometric analysis, Cutispoto (1995) estimated the secondary to be a G8V star. Later, Donati et al. (1997) detected the secondary component in the optical range and they calculated a 1.0 $R_{\odot}$ radius for it. Weber & Strassmeier (1998) gave a K0III-IV type for the primary and computed the first double-lined orbit of IL Hya. Later, Raveendran & Mekkaden (1998) gave a new orbital solution and just recently, Fekel et al. (1999) have presented updated SB2 orbital elements which we have adopted and they are given in Table 2 (we have corrected the heliocentric Julian date on conjunction ( $T_{\rm conj}$) so that the primary is in front). The multiwavelength Doppler images presented by Weber & Strassmeier (1998) revealed a cool polar spot and several features at low latitudes. These authors also found that the H$\alpha$ EW showed sinusoidal variation which was in phase with the photospheric light curve.

We have taken one spectrum of IL Hya (NOT98) with 0.02 orbital phase which is very close to the conjunction, so the very weak lines of the secondary can not be detected in any wavelength. In the observed spectrum (Fig. 12), the H$\alpha$ line can be seen as a filled-in absorption line with a superimposed 1.04 Å$\$ (48 km s^-1) red-shifted absorption feature, as obtained from a two-Gaussian fit. After applying the spectral subtraction, clear excess emission is observed. The excess emission shows an asymmetric profile due to the red-shifted absorption feature. Similar H$\alpha$ profiles were observed by Weber & Strassmeier (1998) in this system, but the red-shifts measured in their spectra were larger (1.24 Å) and remained constant during a rotational cycle. This behaviour could be due to mass motions that change from one epoch to another, but a combination of several dynamical processes may be involved. A filling-in is also found in the H$\beta$ line. According to the value obtained for the corrected ratio of the excess emission EW of both lines, the emission may be ascribed to an extended region viewed off the limb. The He I D₃ line appears in absorption (Fig. 24), but no filling-in is detected in the Na I D₁, D₂ lines. The Ca  II H $\&$ K lines are observed in emission. Furthermore, the Ca II IRT lines show clear central emission reversal.

4.8 FG UMa (HD 89546)

FG UMa is the least studied star of our sample. Bidelman (1981) included it in his Catalogue of stars with Ca  II H & K emission. This star is identified as a single-lined binary by CABS. A 21.50-day photometric period has been obtained from the automated monitoring that Henry et al. (1995a) carried out. From spectroscopic measurements, they confirmed $V\sin{\it i}$ = (15 $\pm$ 2) km s^-1 and a G8IV spectral type. These authors also mentioned that, according to an unpublished orbital analysis, the system is synchronized and circularized. We have also taken from them the orbital period and the radius. Some indication of possible eclipses is noted in the Hipparcos Catalogue (ESA 1997). Fluxes of the Ca  II H & K emission lines have been calculated by Strassmeier (1994b) and a filled-in and variable H$\alpha$ line has been reported by Henry et al. (1995a).

We have not got enough data to be able to compute the orbital phases corresponding to the two spectra (NOT98) that we present here. However, there is a change of 0.2 in the photometric phase between both observations. We have not found any evidence of the secondary star through the whole spectral range. Moreover, according to the appearance of some Ti I and Fe I lines (Paper II) we suggest that the observed spectra correspond to a luminosity class more evolved than subgiant (in agreement with the radius calculated by Henry et al. 1995a, who suggested a luminosity class III-IV). The presence of a notable He I D₃absorption line (Fig. 24) encourages this conclusion. In the observed spectra of the H$\alpha$ and H$\beta$ lines (Fig. 13), both absorption lines show a clear filling-in. The spectral subtraction allows us to compare the two observations. As it can be read in Table 5 the ratio of the excess emission EW is typical of extended regions viewed off limb, and a significant variation, mainly due to H$\beta$, is obtained for the two different nights. Furthermore, excess emission is detected in the blue wing of the H$\alpha$ line. Similar behaviour was mentioned by Henry et al. (1995a). Strong filling-in is observed in the Ca  II IRT lines. The Ca  II H & K lines present strong emission with clear self-absorption with blue-shifted asymmetry in both observations.

4.9 LR Hya (HD 91816)

It is a double-lined spectroscopic binary, classified as a BY Dra-type system, that contains two almost identical K-type dwarf components (Bopp et al. 1984). The orbital parameters were determined by Fekel et al. (1988) who suggested a K0V spectral type for both components. The photometric period of 3.1448 days, given in CABS, was reported by Bopp et al. (1984), but following observation campaigns could not confirm that value. In fact, the results obtained are not consistent (Strassmeier 1989; Cutispoto 1991, 1993) and point out low-amplitude rotational modulation due to the development and decline of small active regions at different longitudes of both components.

We only have got one spectrum (NOT98) of this system at 0.58 orbital phase, so that the chromospheric activity indicators from both components can be easily analyzed. The H$\alpha$ absorption line (Fig. 14) shows a weak filling-in for both components, as it has been previously mentioned by Strassmeier et al. (1990). No filling-in is detected in the H$\beta$ line. Although the S/N ratio in the Ca II H & K region is very low in this observation and the synthetic and observed spectra are not well matched, we can clearly see moderate emission in the Ca II H & K lines from both components. The Ca II IRT lines of the two stars exhibit a clear filling-in. The measured excess emission in the different lines are very similar in both components, although a bit larger in the red-shifted one.

Figure 15: H$\alpha$ observed (left panel) and subtracted (right panel) spectra of HU Vir. We have superposed the two Gaussian components fit (solid-line). The sort-dashed-line represents the broad component and the large-dashed-line the narrow one

Figure 16: H$\beta$, and Ca  II H & K observed (left panel) and subtracted (right panel) spectra of HU Vir

Figure 17: Ca  II IRT $\lambda$8542 and $\lambda$8662 observed (left panel) and subtracted (right panel) spectra of HU Vir

Figure 18: H$\alpha$, H$\beta$, Ca  II H & K, and Ca  II IRT spectra of DK Dra. Observed and synthetic spectra in the left panel and subtracted spectra in the right panel. We have also plotted the Gaussian fit to the subtracted spectrum used to deblend the contribution of the hot (H) and cool (C) components in the Ca  II H & K lines at 0.84 orbital phase (short- and long-dashed lines)

Figure 19: H$\alpha$, H$\beta$, Ca  II H & K, and Ca  II IRT spectra of BQ CVn. Observed and synthetic spectra in the left panel and subtracted spectra in the right panel

Figure 20: H$\alpha$, H$\beta$, Ca  II H & K, and Ca  II IRT spectra of IS Vir. Observed and synthetic spectra in the left panel and subtracted spectra in the right panel

Figure 21: H$\alpha$, H$\beta$, Ca  II H & K, and Ca  II IRT spectra of BL CVn. Observed and synthetic spectra in the left panel and subtracted spectra in the right panel

Figure 22: H$\alpha$, H$\beta$, Ca  II H & K, and Ca  II IRT spectra of BM CVn. Observed and synthetic spectra in the left panel and subtracted spectra in the right panel

Figure 23: H$\alpha$, H$\beta$, Ca  II H & K, and Ca  II IRT spectra of MS Ser. Observed and synthetic spectra in the left panel and subtracted spectra in the right panel

4.10 HU Vir (HD 106225)

HU Vir is a late-type star (K0III-IV) with strong Ca II H $\&$ K emission noted by Bidelman (1981) for the first time. Recently, Fekel et al. (1999) have discovered that HU Vir is a triple system with a long period of about 6.3 years and we have taken from them the spectral type, $T_{\rm conj}$ and the rotational period. The B-V colour index has been taken from Hipparcos Catalogue (ESA 1997) and $V\sin{\it i}$ from Fekel (1997). Fekel et al. (1986) observed the H$\alpha$ line in emission and Strassmeier & Fekel (1990) found the H$\epsilon$ line in emission. Such emission lines are seen in the most active RS CVn type systems. Strassmeier (1994a) found a big, cool polar spot from Doppler imaging and two hot plages 180$^\circ$ apart from the H$\alpha$ and Ca II line-profile analysis. The chromospheric plages seemed to be spatially related to two large appendages of the polar spot. Broadening in the H$\alpha$ profile suggested mass flow in a coronal loop connecting the two plage regions. Hatzes (1998) used the Doppler imaging technique to derive the cool spot distribution. He found a large spot at latitude 45$^\circ$ and a weak polar spot with an appendage. The polar spot was considerably smaller than similar features found on other RS CVn stars. From an analysis of the H$\alpha$ variations he also found evidence for a plage located at high ($\approx$ 70$^\circ$) latitude, near the polar extension.

The H$\alpha$ line: Strassmeier (1994a) identified three distinctive features in the H$\alpha$ line: blue-shifted emission, central sharp absorption and red-shifted broad absorption. Hatzes (1998) found similar behaviour in this line. In our spectra (Fig. 15), the H$\alpha$ line always appears in emission, although its intensity is variable. Moreover, the emission is blue-shifted at some orbital phases. Hall & Ramsey (1992, 1994) explained the blue-shifted emission as prominence-like structures. We can see red-shifted broad absorption at the 0.54 orbital phase but we never observe central sharp absorption. Walter & Byrne (1998) said that there was growing evidence for continuous low-level mass in-fall, seen as red-shifted absorption in H$\alpha$ line profiles. We can also notice that the subtracted H$\alpha$ profiles have broad and variable extended wings which are not well matched using a single-Gaussian fit. These profiles have been fitted using two Gaussian components. The parameters of the broad and narrow components used in the two-Gaussian fit are given in Table 4 and the corresponding profiles are plotted in the right panel of Fig. 15. These broad wings are observed at different orbital phases and in different epochs. In some cases the blue wing is noticeable stronger than the red wing and the fit is better matched when the broad component is blue-shifted with respect to the narrow component. We have interpreted these broad components as microflaring activity that occurs in the chromosphere of this very active star (Papers I, II; Montes et al. 1998b). The contribution of the broad component to the total EW of the line ranges from 32% to 66% which is in the range observed in the stars analysed in Papers I and II. Strassmeier (1994a), in 1991, and Hatzes (1998), in 1995, observed strong emission between the 0.27 and 0.51 orbital phases. We have obtained the strongest emission at the 0.44 orbital phase, in January 1998, and at 0.29, in April 1998 (see Table 3). Thus, our 0.29-0.44 orbital phase interval is similar to Strassmeier and Hatzes's orbital phase interval, so we can conclude that HU Vir has an active longitude area (which corresponds to that orbital phase interval) since 1991. Similar active longitudes have been found by other authors in several chromospherically active binaries (Oláh et al. 1991; Henry et al. 1995b; Jetsu 1996; Berdyugina & Tuominen 1998). The H$\beta$ line: A nearly complete filling-in is observed (Fig. 16 upper panel). After applying the spectral subtraction technique, clear excess H$\beta$ emission is obtained. We have obtained $\frac{E_{{\rm H}\alpha}}{E_{{\rm H}\beta}}$ values larger than three (see Table 5), so the emission can come from prominences. The Ca II H $\&$ K and H$\epsilon$ lines: We can observe very strong Ca II H $\&$ K emission and an important H$\epsilon$ emission line superimposed to the wide Ca  II H line (Fig. 16 lower panel). The H$\epsilon$ line in emission indicates that HU Vir is a very active system. Similar strong emission was found in our previous observation obtained in March 1993 at 0.71 orbital phase (Montes et al. 1995c). We can also notice that the largest values of the excess Ca II H $\&$ K emission EW appear at the 0.29 and 0.38 orbital phases (see Table 6). It is in agreement with the H$\alpha$ line behaviour. The Ca II IRT lines: HU Vir shows the Ca II IRT lines in emission above the continuum (Fig. 17). We observe that the emission is centered at its corresponding absorption. We also notice that the subtracted profiles have broad wings due to microflares according to Montes et al. (1997). The excess Ca  II $\lambda$8542 emission EW (see Table 7) behaves like the excess H$\alpha$ emission EW. The He I D₃ line: We have not detected any significant absorption or emission for He  I D₃(Fig. 24), contrary to the absorption observed in other giants. This total filling-in of the He  I D₃ line could be explained (Saar et al. 1997; Paper II) as a filling-in due to the low-level flaring (microflares) that takes place in this very active star according to the H$\alpha$ broad component that we have found.

The Li  I $\lambda$6707.8 line: The Li  I absorption line is clearly observed in the eight spectra of the McD98 observing run (see Fig. 25). The mean EW obtained is 56 $\pm$ 11 (mÅ). At this spectral resolution the Li  I line is blended with the nearby Fe  I $\lambda$6707.41 Å line. We have corrected the total EW measured, EW(Li  I+Fe  I), by subtracting the EW of Fe  I calculated from the empirical relationship with (B-V) given by Soderblom et al. (1990) (EW(Fe  I) = 24 (mÅ)). Finally, the corrected EW(Li  I) $_{\rm corr}$ = 32 (mÅ) was converted into abundances by means of the curves of growth computed by Pallavicini et al. (1987), obtaining log N(Li  I) = 1.2 (on a scale where log N(H) = 12.0) with an accuracy of the $\approx$ 0.30 dex. This value is larger than the lower limit reported by Barrado et al. (1998) for this star.

4.11 DK Dra (HD 106677)

This is a double-lined spectroscopic binary with almost identical components of spectral types K1III and Ca  II H & K emission from both components (Bopp et al. 1979; Fekel et al. 1986; Strassmeier 1994b). Eker et al. (1995) reported variable nature of H$\alpha$ and, using a subtracted spectrum, found emission of similar intensity from both components. In our previous observations (Montes et al. 1995a,b; FFMCC) we found a broad excess H$\alpha$ emission line and the Ca  II H & K lines in emission, but all of them were taken at orbital phases near to the conjunction, so it was impossible to distinguish the contribution of each component.

The spectra analysed in this paper were taken at 0.90 (NOT98) and 0.84 (NOT98) orbital phases (see Fig. 18). In both spectra we observe the H$\alpha$, H$\beta$ and Ca  II $\lambda$8662 lines filled-in, and the Ca  II H & K lines in emission. At 0.90, we cannot separate the contribution of each component in H$\alpha$, H$\beta$ and Ca  II H & K lines, but due to the large wavelength of the Ca  II $\lambda$8662 line, a double peak is clearly observed in the subtracted spectrum of this line. In the 0.84 spectrum, we can see that the H$\alpha$ line is filled-in for both stars. Moreover, we can observe the excess H$\alpha$emission of the blue-shifted component is bigger than that of the red-shifted one. The excess H$\beta$ emission seems to come only from the blue-shifted component. Although the S/N ratio in the Ca  II H & K lines region is low, we can clearly see broad and two-piked emission in the H & K lines. In the subtracted spectrum of the Ca  II IRT $\lambda$$\lambda$ 8542 and 8662 lines the emission from both components is clearly separated, being the emission of the blue-shifted one slightly larger. All of this indicates that the blue-shifted component is slightly more active than the other component. Very small absorption is observed in the He  I D₃ line in the expected position for both components (see Fig. 24), contrary to the notable absorption observed in other giants of the sample. This is probably due to the blend with other photospheric lines of both stars in this SB2 system.

4.12 BQ CVn (HD 112859)

A quarter of century ago Schild (1973) classified this star as a peculiar G8III-IV. Henry et al. (1995a) detected its Ca II H & K emission and noticed that at red wavelengths the spectrum was double-lined although the intensity of the lines of both components was very different. They suggested a K0III primary and a late-F spectral type for the secondary. We have taken from these authors both the orbital and rotational period and the radius. On the other hand, Strassmeier (1994b) confirmed the strong Ca II H & K emission.

Orbital parameters have not been published, so we cannot calculate the orbital phase of our spectrum (NOT98). However, some conclusions can be obtained looking at the observed spectrum (Fig. 19): at infrared and red wavelengths, the secondary lines are not clear. A very weak blue-shifted absorption feature can be ascribed to the secondary in the H$\alpha$ region. At shorter wavelengths, as in the Na I D₁, D₂ lines region, the spectral lines of the secondary are more conspicuous. In the Ca II H & K lines region the contribution of the F star is evident in the broad Ca II absorption lines, where a clear red-shift of the emission can be seen. Taking into account what has been said above, we have calculated the synthesized spectrum using a F8V star as a template one for the secondary. In the observed spectrum, the H$\alpha$ line shows a filling-in for the primary star. The H$\beta$ line is slightly filled-in. The presence of the He I D₃ absorption line is detected (Fig. 24). In our spectrum we observe strong Ca II H & K emission and a weak H$\epsilon$ emission line from the cool star. Finally, the Ca II IRT lines exhibit reversal emission.

4.13 IS Vir (HD 113816)

IS Vir is a single-lined spectroscopic binary classified as K0III by Henry et al. (1995a). We have taken from them the orbital and rotational periods and the radius given in Table 2. Strong Ca  II H & K emission was observed (Buckley et al. 1987; Strassmeier 1994b; Montes et al. 1995c). In our previous observation of this system in the Ca  II H & K lines region in Mar-93 at 0.68 orbital phase (Montes et al. 1995c) we found strong emission in the H & K lines with intensity above the continuum at 3950 Å, but lower than reported by Strassmeier (1994b).

In the new spectrum (NOT 98) (Fig. 20), the H$\alpha$ line and the Ca  II IRT lines show intense filled-in absorption, whereas the H$\beta$ line only shows a slight filling-in. In the Ca II H & K lines region the S/N ratio is low and the synthetic and observed spectra are not well matched, but a strong emission in the H & K lines well above the continuum is observed. The He  I D₃ line appears in absorption (Fig. 24).

4.14 BL CVn (HD 115781)

This double-lined spectroscopy binary was classified by Hall (1990) as K1III + FIV, and later as K1III + G5IV by Stawikowski & Glebocki (1994). Lines et al. (1985) found this system to have a photometric period of 9.31 $\pm$ 0.06 days and an amplitude of 0.16 mag. They attributed the light variability to the ellipticity effect because the orbital period was twice the photometric period and times of maximum brightness occurred at times of maximum positive radial velocity. It is a nearly-synchronous binary: its orbital period is 18.6917 $\pm$ 0.0011 days (Griffin & Fekel 1988) and its rotational period is 18.70 days (Stawikowski & Glebocki 1994). The orbital eccentricity cannot be far from zero (Griffin & Fekel 1988). Fekel et al. (1986) found $V\sin{\it i}$ values of (35 $\pm$ 2) km^-1 and (7 $\pm$ 2) km^-1 for the primary and secondary, respectively. The great line broadening of BL CVn and the ellipsoidal light variations might suggest that the K giant star is close to filling its Roche lobe (Griffin & Fekel 1988). Moderate H$\alpha$ absorption is found by Fekel et al. (1986). Strassmeier et al. (1990) found strong Ca II H & K emission.

In our present observations (NOT98, Fig. 21) the H$\alpha$ and H$\beta$ lines appears in absorption in the observed spectrum. After applying the spectral subtraction technique, we only obtain small excess H$\alpha$ emission. Broad emission is observed in the Ca II H & K lines. We observe small absorption in the He I D₃ line (Fig. 24). A small filling-in is obtained in the Ca II $\lambda$8542 and $\lambda$8662 lines.

4.15 BM CVn (HD 116204)

This single-lined spectroscopy binary was classified by Keenan (1940) as K1III. It is a nearly-synchronous binary: its orbital period is (20.6252 $\pm$ 0.0018) days (Griffin & Fekel 1988) and its rotational period is (20.66 $\pm$ 0.03) days (Strassmeier et al. 1989a). The orbit is judged to be circular (Griffin & Fekel 1988). It is also a relatively fast rotator, Fekel et al. (1986) found rotationally broadened lines with $V\sin i$ = (15 $\pm$ 2) km^-1. This system shows strong Ca II H & K emission (Bidelman 1983), together with an emission-line spectrum typical of RS CVn stars in the IUE ultraviolet region, but H$\alpha$ is an absorption line (Griffin & Fekel 1988).

In our present spectrum (NOT98, Fig. 22) we observe the H$\alpha$ line as nearly total filled-in absorption. After applying the spectral subtraction technique, strong excess H$\alpha$ emission is obtained. The H$\beta$ line appears as an absorption line in the observed spectrum and clear excess emission is observed in the subtracted spectrum. The $\frac{E_{{\rm H}\alpha}}{E_{{\rm H}\beta}}$ ratio obtained is larger than 3, which indicates that the emission would arise from extended regions viewed off the limb. The Ca II H & K lines show strong emission, with a blue-shifted self-absorption feature, and small emission is also detected in the H$\epsilon$ line. The He I D₃ line appears in absorption (Fig. 24). A clear emission reversal is observed in the Ca II IRT $\lambda$8542 and $\lambda$8662 lines.

4.16 MS Ser (HD 143313)

Griffin (1978) first observed the binary nature of MS Ser calculating the orbital elements for this system (see Table 2). Griffin gave its T₀ in MJD, and this yielded Strassmeier et al. (1993) to a bad calculation of the T₀ in HJD. Griffin also proposed K2V/K6V as spectral types of the components, based on photometric arguments for the secondary star. Bopp et al. (1981b) observed a variable filling-in of the H$\alpha$ line and calculated a photometric period of 9.60 days, slightly different from the orbital period. Miller & Osborn (1996) confirmed the value of the photometric period and Strassmeier et al. (1990) observed strong emission in the Ca  II H & K composite spectrum. Dempsey et al. (1993a) noted some filling-in in the Ca IRT lines for MS Ser, but not reverse emission. Alekseev (1999) made a photometric and polarimetric study of MS Ser, calculating a spot area of 15% of the total stellar surface, and observed some seasonal variations. Finally, Osten & Saar (1998) revised the stellar parameters for MS Ser and suggested K2IV/G8V as a better classification. We have also found that the primary component may have a luminosity class IV or higher based on our analysis of some metallic lines, like the Ti  I lines, the Hipparcos data and the Wilson Bappu effect (see Sanz-Forcada et al. 1999).

Two spectra of MS Ser were taken in the NOT98 observing run. Moreover, we analyse here another spectrum taken on 12th June 1995 with the 2.2 m telescope at the German Spanish Observatory (CAHA) in Calar Alto (Almería, Spain), using a Coudé spectrograph with the f/3 camera, CCD RCA #11 covering two ranges: H$\alpha$ (from 6510 to 6638 Å) and H$\beta$(from 4807 to 4926 Å), with a resolution of $\Delta\lambda$ 0.26 in both cases.

The H$\alpha$ and H$\beta$ lines: In both lines, we observe a nearly total filling-in in the 1995 spectra and small absorption in the spectra taken in the NOT98 observing run. After applying the spectral subtraction, a clear filling-in in the H$\alpha$ and H$\beta$ lines is observed in the three spectra (see Fig. 23). The H$\alpha$ line of this system is highly variable. We have found, in the present spectra, a variable filling-in whereas Bopp et al. (1981b) found H$\alpha$ was a weak emission line. In the three new spectra we have taken of this system with the FOCES echelle spectrograph in July 1999 (forthcoming paper) we observe variable H$\alpha$ emission well above the continuum.

The Ca  II H $\&$ K and H$\epsilon$ lines: Strong emission in these lines and the H$\epsilon$ line in emission arising from the primary component was observed in our previous observation of this system in March 1993 at 0.16 orbital phase (Montes et al. 1995c). In the present observations (Fig. 23) we have deblended the emission arising from both components in the spectrum taken near to the quadratures ($\varphi$ = 0.21). The strongest emission, centered at the absorption line, arises from the K2IV component, which is the component with larger contribution to the continuum. The red-shifted and less intense emission corresponds to the G8V secondary component. In the $\varphi$ = 0.55 observation we cannot separate the contribution from both stars. The H$\epsilon$ line appears in emission in both spectra. The emission intensity observed in our 1993 and 1998 spectra is larger than the emission intensity observed in the 1988 spectrum presented by Strassmeier et al. (1990).

The Ca  II IRT lines: A clear emission reversal is observed in the core of the Ca  II IRT absorption lines $\lambda$8542 and $\lambda$8662 in both spectra (Fig. 23). After applying the spectral subtraction technique, we can see small excess emission arising from the secondary component, in the 0.21 spectrum, as in the case of the Ca  II H & K lines. This emission reversal observed here clearly contrasts with the only filling-in in these lines reported by Dempsey et al. (1993a).

The He  I D₃ line: We can distinguish (Fig. 24) the He  I D₃ line as a very small absorption from the primary star. A slight variation is observed between the spectra at 0.21 and 0.55 orbital phases. The luminosity class of this star (IV-III) and the SB2 nature of the system could be the reason of this small absorption in the He  I D₃ line in relation with that observed in other giants of the sample.

Figure 24: He  I D3 line for selected stars of the sample. We plot observed spectrum (solid-line) and the synthesized spectrum (dashed-line). The expected position of the He  I D3 line for each star is marked with a short vertical line