The photometric data are given in Table 2, and the light curve in the V-band for the period 1965-1997 is displayed in Fig. 1. A major part of these data (before JD 2 448 000) were included in the catalogue by Herbst et al. (1994). The last brightening started in October 1996 (JD 2 450 360). Within about 40 days, the star raised up to V and stayed at this level during three months, with small fluctuations of (Fig. 2). The subsequent fading of brightness down to V lasted for about one month, to the end of the observational season in March 1997. At the maximum brightness the average B-V colour was redder by about as compared to that at the minimum brightness. The same change can be noticed in the U-B colour, but with larger scatter. The V versus B-V diagram is shown in Fig. 3: like was observed for previous photometric changes, the star becomes slightly redder when fading from V to , but then turns to become bluer when even fainter, with some intrinsic dispersion in the colours (Fig. 3).
RY Tau is located in a region of large and non-uniform interstellar extinction. The value of toward RY Tau was estimated as to by different observers (Cohen & Kuhi 1979, Kuhi 1974, Cernicharo et al. 1985).
Therefore, one may expect a large degree of interstellar polarization, comparable to the intrinsic polarization of RY Tau. Efimov (1980) found an interstellar polarization of % with a position angle of PA = for the wavelength of maximum polarization 0.55 µm, using observations of different stars in the vicinity of RY Tau and applying the normal law of interstellar extinction with . After that publication, a large set of polarimetric UBVRI observations of RY Tau was carried out at the Crimean Astrophysical Observatory in the period of 1981-1987. Using the new method for determination of interstellar polarization proposed by Shakhovskaya et al. (1987), we made a new estimation of the interstellar polarization toward RY Tau, based on more than 140 observations of the star: %, PA = . The Serkowski law of polarization was adopted with the maximum polarization wavelength of 0.55 µm (Whittet 1977).
During the latest event of brightening of RY Tau in 1996, we started polarimetric observations in December 1996, when the star was already at high brightness, and followed the decline of brightness to the end of the observational season. The results of the observations are given in Table 3 and shown in Fig. 4. After correction for the interstellar polarization, the intrinsic polarization of RY Tau and its position angle were found to be variable with the brightness of the star. In the V, R and I bands the intrinsic polarization was increasing from 0.5-1.0% at high brightness to about 2% at low brightness. In the U and B bands this tendency is not well expressed. The position angle was progressively rotating as the star fades.
3.3.1. Photospheric spectrum
The depths of the photospheric lines in the 6 échelle spectra taken at high brightness in 1996 are the same; thus here we use the average photospheric spectrum.
The spectra of RY Tau at high and low brightness levels (Fig. 5) show remarkable constancy of the photospheric absorption lines within the accuracy of our measurements, which indicates the constancy of the photospheric temperature within about 40 K, in spite of the one magnitude difference in stellar brightness. From the region of 5000-7000 Å we determined the radial velocity RV km s-1 and the projected rotational velocity km s-1. Both values are consistent with previous estimates (see e.g. Hartmann & Stauffer 1989 and Bouvier 1990).
The photospheric line spectrum of RY Tau is very similar to the solar spectrum, taken with the same spectrograph (day sky spectrum) and spun up to km s-1. Another suitable standard star is 84 Her: G1 III, K, (Berdyugina 1994). The photospheric line depths and ratios in the spectrum of RY Tau are in the range given by spectra of the Sun and 84 Her (Fig. 6), in the spectral interval of 4500-8800 Å. An upper limit for the veiling factor can be set to . Therefore, we estimate the spectral type of RY Tau as G1-G2 IV with the intrinsic B-V colour . At the bright state, RY Tau had V and B-V , that is and . With the distance to RY Tau of 140 pc (Elias 1978), we get , which is in accordance with the spectral classification of G1-G2 IV.
3.3.2. Emission line spectrum
The most prominent emission lines in our spectra of RY Tau are those of H and the IR-triplet of Ca ii. Weaker emission components are also present in H, He i 5876 Å and Na i D. Of the forbidden emission lines only [O i] 6300 Å is clearly seen.
It is interesting that the relative intensities of the emission components of H and the IR Ca ii lines drop by a factor of 2-3 between Dec 95 and Nov 96, when the star has brightened by one magnitude (see Fig. 9). This means that the flux radiated in these emission lines has remained about the same.
In the region around 5180 Å, the depths of the absorption lines of Mg i and Fe ii at high brightness are consistent with the spectral type of the star (G1-2), but at low brightness these lines were shallower, as if the star were of late F-type. Since the photospheric temperature was the same at the two brightness levels, we conclude that at low brightness the Mg i and Fe ii lines were partially filled in with emission. These emissions are shown in the differential spectrum: normalized spectrum at low brightness minus normalized spectrum at high brightness (Fig. 10).
In the high brightness state, the emission components of the sodium doublet are not prominent, while there are variable absorptions in the blue and red wings of the lines. The average equivalent width of the sodium lines is noticeably larger than in the comparison G stars (Fig. 7). The He i 5876 Å line is only slightly variable, with an average profile of inverse P Cyg-type, with the deepest absorption at +15 km s-1 and the red absorption wing extending to +100 km s-1 (the radial velocities are referred to the stellar restframe).
Since the H region was most frequently observed and the photometric history of RY Tau is well documented, we can check if the flux radiated in H relates to the stellar brightness. The equivalent widths of the H emissions and the fluxes radiated in H are given in Table 1. In addition to the spectra listed in Table 1, we also used several older spectra in the H region obtained at Crimea in 1988-89 (Petrov & Vilhu 1991).
Fig. 11 shows that the flux in H is variable by a factor of three but does not reveal any dependence on the stellar brightness. On average, when the star is fainter, the H emission appears stronger, because the range of the continuum variability is comparable to the range of intrinsic variability of the emission line flux.
The H line profile consists of a broad emission with symmetric wings extending to 400 km s-1, and absorption components which appear and disappear on a time scale of days. There are two preferential velocities of these absorptions: one is around zero velocity and the other is at -10050 km s-1; both are clearly visible in the average H profile in Fig. 9. This kind of H profile variability was observed also in SU Aur (Giampapa et al. 1993, Johns & Basri 1995, Petrov et al. 1996).
The method of the correlation matrix is now commonly used to reveal a correlation in variability between different parts of a line profile or between two different lines. The correlation matrix H versus H plotted in Fig. 12 includes the 19 spectra listed in Table 1. The absolute flux profiles of H were used to compute the matrix, i.e. the variations of the continuum brightness do not affect the correlation matrix.
The two absorption components of H (at 0 and -100 km s-1) vary independently from each other and independently from the rest of the emission profile. There is also no correlation between the blue ( km s-1) and red ( km s-1) wings of the line. The appearance of the correlation matrix is not very different from that published by Johns & Basri (1995).
© European Southern Observatory (ESO) 1999
Online publication: December 4, 1998