3 Individual stars

3.1 HD22920 (= HR1121 = 22 Eri)

According to Maitzen (1976), the silicon star HD22920 has a low value of the photometric peculiarity index $\Delta a$ (= 0.011). Photometric observations have been carried out by Bartholdy (1988) who found this star to be variable with a period of 3.95 d. North (1990 priv. comm.) found two possible periods almost equally probable: 3.96 d, very close to Bartholdy's (1988), and 1.33 d.

No evidence of variability has been found in our spectra of the HeI5876 Å line. The mean value of the equivalent width is: $\langle EW\rangle$ = 125 $\pm$ 13 mÅ. The effective temperature of HD22920 resulting from Napiwotzki et al. (1993) relation is $T_{\rm eff} = 13700$ K. Figure9 shows that the HeI5876 Å line equivalent width of HD22920 is smaller than expected for a main sequence star of the same effective temperature.

3.2 HD24155 (= HR1194 = V766 Tau)

The UBV photometric variability of HD24155 has been studied by Winzer (1974), who reported a possible period of 2.5352 d. Renson & Manfroid (1981) found P = 2.53465 $\pm$ 0.00015 d. The observed light curves show a quite large amplitude (0.10 mag) with very sharp minima and quite broad maxima, hence this star is the fourth largest amplitude silicon star known, exceeded only by HD215441, CU Vir and HR7058.

Assuming Renson & Manfroid's (1981) period, our nine spectra are well distributed in phase, but none of them shows a measurable HeI5876 Å line. Because of its effective temperature of 13700 K, HD24155 is an extremely helium weak star (Fig.9).

3.3 HD24587 (= HR1213 = $\tau^8$ Eri)

HD24587 is listed in the General Catalogue of Ap and Am stars by Renson et al. (1991) as a suspected CP star. Feinstein (1978) used this star as standard for his measurements of hydrogen lines in He weak stars. HD24587 has in fact been considered as a standard for uvby (Garnier 1972 - personal communication to Mathys et al. 1986) and $\beta$ photometry (Strauss & Ducati 1981). Mathys et al. (1986) found this star to be a light variable with a period of 1.728 d and concluded that the light curves resemble those of many CP stars. Recently Leone & Catanzaro (1998) have performed a spectroscopic study and concluded that this star presents chemical elements which are slightly underabundant with respect to main sequence stars.

Our measurements of the HeI5876 Å line do not show any variation of the equivalent width; the mean value is: $\langle EW\rangle$ = 325 $\pm$ 20 mÅ. From Moon's algorithm we find that $T_{\rm eff}$ = 14100 K and the EW of the HeI5876 Å line is close to the value expected for a main sequence star (Fig.9). These facts confirm Leone & Catanzaro's (1998) conclusion that HD24587 is not a peculiar star.

3.4 HD26571 (= HR1297)

The peculiarity of HD26571 was first noted by Gulliver (1971) and independently confirmed by Bond (1972). On the basis of his spectra, Gulliver (1971) described this star as a spectrum variable. Photometric observations were obtained by Winzer (1974), who found HD26571 to vary with a period of 1.0646 d.

Figure 1 shows the EW variation of the HeI5876 Å line versus the phase computed assuming the initial epoch coincident with the light maximum as given by Winzer (1974):

$${\rm JD}(UBV \; {\rm max}) = 2441246.81 + 1.0646 \ E .$$ (3)

Winzer (1974) has not published the uncertainty in the period determination, hence we have estimated the error by applying the Horne & Baliunas (1986) relation to our data and have found that equivalent widths are phased with an expected error $\Delta \Phi$ = 0.3. This means that no phase relation can be determined between the photometric and our spectral variations.

  

Figure 1: Equivalent width variations of HD26571. The errorbar is equal to the error in the equivalent width as given by Eq.(1). Photometric data are from Winzer (1974). Phases have been computed by Eq.(3). Open circles refer to the 1997 data, stars to the 1995 data. The fitting of EW's performed by Eq.(2) is reported with a solid line. The phase relation between EW and photometric variations may be not real because of the large time gap between both data sets and the relatively low accuracy of the variability period

3.5 HD28843 (= HR1441 = DZ Eri)

HD 28843 was classified as B9IV Si He-wk by Davis (1977) and it is classified as B9 He wk in the General catalogue of Ap and Am Stars. The photometric variability of HD28843 had been detected for the first time by Cousins & Stoy (1966) while its peculiar character had been confirmed by Jaschek et al. (1969). Photometric observations of this star have been carried out by Pedersen & Thomsen (1977) who found variability with a period of 1.374 $\pm$ 0.006 d. This value was improved by Pedersen (1979) to the value 1.37375 $\pm$ 0.00035 d. Manfroid et al. (1984) also used Pedersen & Thomsen's (1977) data to improve the period, their most probable value being 1.373813 $\pm$ 0.000012 d. Mathys et al. (1986) concluded that the ambiguity in the choice of the best peak in the periodogram could be removed by inclusion of the measurements of Dean (1980), confirming the value obtained by Manfroid et al. (1984). Further photometric observations have been carried out by Waelkens (1985), by the team of the ESO Long-Term Photometry of Variable Project (Manfroid et al. 1994; Sterken et al. 1995), and by the team of Hipparcos (ESA, 1997).

Our spectroscopic data are plotted in Fig. 2, versus the phase computed from the ephemeris elements of Mathys et al. (1986):  

$${\rm JD}(uvby \; {\rm max}) = 2442777.5 + 1.373813\ E.$$ (4)

The amplitude of the equivalent line width variations is of the order of 75 mÅ. From Fig. 2 a clear anti-correlation is evident between the HeI5876 Å equivalent line width and all the Hipparcos and uvby light curves, in the sense that light minima occur at the phase of maximum HeI. Because of the period error determined by Manfroid et al. (1984), the expected phase error in our EW variations is $\Delta \Phi$ = 0.03. EW variations of the HeI5876 Å line are then out of phase with respect to light variations.

Even if most of our equivalent widths periodically vary with the ephemeris computed with Eq. (4), we have found several (5 out of 21) spectra where the HeI5876 line is absent (Fig.2).

  

Figure 2: Equivalent width variations of HD28843. The bar is equal to the error in the equivalent width as given by Eq.(1). Open circles refer to the 1997 data, stars to the 1995 data. Photometric observations are from Hipparcos (ESA, 1997) and from the ESO LTPV Project (Manfroid et al. 1994; Sterken et al. 1995)

3.6 HD36589 (= HR1860)

In the General Catalogue of Ap and Am stars, the star HD36589 is a suspected CP star. Bossi & Guerrero (1989) and Hao et al. (1996) have used it as a comparison star for photometric observations. Leone & Catanzaro (1998) derived chemical abundances and found that HD36589 shows nearly solar values and no evidence of spectral variability.

From our spectra we confirm this result: no evidence of variation is found in the HeI5876 Å equivalent line width. On the hypothesis that HD36589 is not a CP star, we have determined $T_{\rm eff}$ = 14000 K by mean of Moon's relations and found that the average value of the equivalent widths ($\langle W \rangle$ = 350 $\pm$ 8 mÅ) is very close to that of normal main sequence stars of the same spectral type (Fig.9).

3.7 HD41269 (= HR2139)

This star has been classified as B9p by Cowley et al. (1969) who described it as a mild silicon star. On the basis of a single observing run in UBV, Winzer (1974) found a period of 1.68 d, although he could not rule out the resonance period of 2.47 d, because of the few observed points.

In our spectra the HeI5876 Å line is too weak to be measured. According to Napiwotzki et al. (1993) relation, $T_{\rm eff}$ = 10800 K. Figure 9 shows that the helium abundance is lower than the expected value for a main sequence star of this temperature.

3.8 HD43819 (= HR2258 = HIP30019 = V 1155 Ori)

Cowley (1972) classified this star as B9IIIp Si. Photometric measurements of HD43819 were performed in the UBV system by Winzer (1974) who found a light variation with a period of 1.0785 d. Later on Maitzen (1980) found the light variation to occur with two possible periods: 0.93 d and 1.077 d. A spectroscopic study of this sharp-lined star ($v_{\rm e} \sin i = 14$ km s^-1, Wolff & Preston 1978) was carried out by Lopez-Garcia & Adelman (1994), who found iron peak elements ten times overabundant and rare earths 1000 times overabundant with respect to solar values. From photometric uvby observations Adelman (1997) has deduced a period of 15.0305 $\pm$ 0.0003 d, longer than Winzer's (1974) and more consistent with the low rotational velocity of this star. This period is also confirmed by the Hipparcos observations (Fig.3).

Our HeI5876 Å equivalent line widths are plotted in Fig. 3 versus the phase computed by means of Adelman's (1997) ephemeris elements:

$${\rm JD}(U \; {\rm max}) = 2441254.16 + 15.0305\ E.$$ (5)

The observed EW variation has an amplitude of the order of 80 mÅ. From Fig. 3 we see a clear in-phase correlation between light and spectral variations. This correlation is expected to be real, the phase error being $\Delta\Phi\sim$ 0.01.

  

Figure 3: Equivalent width variations of HD43819. Errorbar is equal to the error in the equivalent width as given by Eq. (1). Photometric data are from Hipparcos (ESA, 1997)

3.9 HD49606(= HR2519 = 33 Gem)

The star HD49606 is classified as a B8HgMnSi star by Renson et al. (1991). Photometric observations of HD49606 were performed by Chunakova et al. (1981), who found the light variations to occur with a period of 3.099 d, and by Glagolevskii et al. (1985), who found two possible period values, namely 3.3546 d and 1.41864 d.

The HeI5876 Å equivalent line width observed in our spectra does not show any detectable variation, so that we consider this line does not vary with time. The average equivalent width is $\langle W \rangle = 145 \pm 7$ mÅ. This result confirms the one obtained by Hubrig & Launhardt (1993) who searched for variations in the equivalent width of helium and some metallic lines and did not find any evidence of variability.

According to an elemental abundances analysis performed by Adelman et al. (1996) our observations show that helium is underabundant with respect to solar composition (Fig.9).

3.10 HD171247 (= HR6967 = HIP90971)

The photometric variability of this star was detected by North (1992), who found the period to be 3.9124 d. This value of the period is confirmed by the Hipparcos photometry (1997) which gives an error on the period equal to 0.0004 d applying Horne & Baliunas (1986) formula.

Computing the phase of the measured equivalent widths by means of North's (1992) ephemeris:

$${\rm JD}([U]{\rm \, Geneva\, max}) = 2447178.245 + 3.9124\ E$$ (6)

we find a sinusoidal variation of the HeI5876 Å line strength (Fig.4) with an amplitude of the order of 45 mÅ.

Converting the period error to a phase error, we get $\Delta \Phi$ = 0.07. We can thus conclude that $H_{\rm p}$ and EW variability are in phase for the silicon star HD171247.

  

Figure 4: Equivalent width variations of HD171247. the bar is equal to the error in the equivalent width as given by Eq. (1). Photometry is from Hipparcos (ESA, 1997)

3.11 HD176582 (= HR7185 = HIP93210)

This star is classified as a silicon star (Renson et al. 1991). Spectroscopic observations of the HeI4026 Å line strength were carried out by Pedersen (1976), who found a variation with the period 0.8143 d. The period is not representative of the variability of Hipparcos photometry and HeI5876 Å equivalent width.

By using our spectroscopic data and Hipparcos photometry we found a period of 1.5817 $\pm$ 0.0003 d. The observations are plotted in Fig. 5 versus the phase computed by means of the ephemeris elements:

$${\rm JD}(EW \; {\rm min}) = 2450624.6410 + 1.5817\ E.$$ (7)

From this figure we see that both curves show a clear evidence of a double-wave variation. The observed EW amplitude is of the order of 40 mÅ. The expected phase error is $\Delta \Phi$ = 0.05, and the Hipparcos photometry appears to vary in phase with the equivalent width variations of the HeI5876 Å line.

  

Figure 5: Equivalent width variations of HD176582. Errorbar is equal to the error in the equivalent width as given by Eq. (1). Photometry is from Hipparcos (ESA, 1997)

3.12 HD177003 (= HR7210 = HIP93299)

Schöneich & Zelwanowa (1984) from their photometric observations in the UV filters found two possible periods: 0.66 d and 2.1 d. From UBVRI photometric observations, Vetö (1993) found this star to be light variable with a period of 0.724 d and amplitudes of about 0.1 mag in all filters. An analysis of Hipparcos photometric data does not give a clear variability period.

Our spectroscopic data are not consistent with the periods given in the literature. A period search of our data yelds two possible values: 1.835 $\pm$ 0.004 d and 2.186 $\pm$ 0.005 d. The HeI5876Å line equivalent width variation is plotted in Fig. 6 versus the phase computed by means of the ephemeris elements:

$${\rm JD}(EW \; {\rm max}) = 2450629.4099 + 1.835\ E$$ (8)

where we have adopted the shorter value of the period which has a smaller $\chi^{2}$ value. The variation shown in Fig. 6 has an amplitude of the order of 75 mÅ.

  

Figure 6: Equivalent width variations of HD177003. Errorbar is equal to the error in the equivalent width as given by Eq. (1)

The photometric variability is not clear for the $H_{\rm p}$ filter assuming this period (Fig.6) and no conclusion can be drawn concerning a possible phase relation between photometric and spectral variations of the HeI5876Å line.

3.13 HD182255 (= HR7358 = HIP95260 = 3 Vul)

According to Hube & Aikman (1991) this star is a nonradial pulsator. It has also been observed by Hipparcos, from whose photometry a period of 1.26239 d has been derived. However this value of the period is not perfectly consistent with our spectroscopic observations; instead, by using both sets of data the most probable value appears to be 1.26263 $\pm$ 0.00005 d. Adopting this period, the measured EW of the HeI5876Å line are plotted in Fig. 7 versus the phase computed by means of the ephemeris elements:

$${\rm JD}(EW \; {\rm max}) = 2450650.4729 + 1.26263\ E .$$ (9)

The observed EW amplitude is of the order of 65 mÅ. Since the period error corresponds to a phase error $\Delta \Phi$ = 0.09, the reported out of phase relation between photometric and helium line variations is expected to be real (Fig.7).

  

Figure 7: The light and HeI5876 Å equivalent line width variations of HD182255. Errorbar is equal to the error in the equivalent width as given by Eq. (1). Photometry is from Hipparcos (ESA, 1997)

3.14 HD209515 (= HR8407 = V1942 Cyg)

This star was classified as A0p by Osawa (1965) and as A0 IV by Cowley et al. (1969). From his photometric observations, Winzer (1974) found a period of 0.63703 d, concluded that the observed photometric variation is typical for a silicon star and suggested that the correct classification should be A0p Si.

The equivalent width of the HeI5876Å line of the cool CP star HD209515 is constant: 40 $\pm$ 5 mÅ. This value of equivalent width is consistent with the helium abundance of a main sequence star (Fig.9).

3.15 HD220825 (= HR8911 = $\kappa$ Psc)

The variability of HD220825 had been detected for the first time by Rakosch (1962) who found a period of 0.5805 d. Recently, Ryabchikova et al. (1996) determined the period to be 1.418 d and magnetic observations performed by Borra & Landstreet (1980) are also consistent with this value.

The HeI$\lambda$5876 Å line is too weak to be measured. Assuming $T_{\rm eff}$ = 10400 K, Fig.9 shows that helium is underabundant in HD220825 with respect to main sequence stars.

3.16 HD223640 (= HR9031 = HIP 117629 = 108 Aqr = ET Aqr)

The photometric variability of this star has been studied by several authors. Morrison & Wolff (1971) found HD223640 to be variable in the Strömgren system with a period of 3.73 d and noted that light curves show quite the same behaviour in all filters. Spectroscopic observations were carried out by Megessier & Garnier (1972) who found strongly variable the Ti and Sr lines and constant the Fe lines. Moreover the Ti lines correlate with photometric variations in the sense that Ti lines are strongest when the star is brightest. This correlation has been interpreted by Megessier (1974, 1975) in terms of the oblique rotator model taking also into account the sign changes of the magnetic field measurements by Babcock (1958). Photometric observations in the Geneva system have been performed by North et al. (1992); they found a period of 3.735239 $\pm$ 0.000024 d which is consistent with the magnetic data. This period has been confirmed by photometric observations in the uvby system performed by Adelman & Knox (1994) and Adelman (1997).

According to North et al. (1992), we phased the measured equivalent widths of the HeI5876 Å line by means of the ephemeris elements:

$${\rm JD}(uvby \; {\rm max}) = 2444696.820 + 3.735239\ E.$$ (10)

The period uncertainty corresponds to a phase error $\Delta \Phi$ = 0.004. There is evidence of an anti-correlation between light and spectroscopic curves. The HeI line is strongest in coincidence with the light minimum: the helium distribution on the surface of HD223640 is then not coincident with the Ti distribution.

  

Figure 8: Equivalent line width variations of HD223640. Errorbar is equal to the error in the equivalent width as given by Eq. (1). Photometry is from Hipparcos (ESA, 1997)