Carina and S Dor, as one of the prototypes of the Luminous
Blue Variables (LBVs). LBVs form a rather heterogeneous class of very massive, unstable stars
located close to the Humphreys-Davidson limit on the HR diagram. The stars are characterized
by high mass-loss rates, episodes of shell ejection and large eruptions. Though considered to
be in a quiescent phase, the LBV P Cygni has demonstrated several aspects of wind activity
such as line-profile variability (lpv) at UV and optical wavelengths; radio-flux variability
(van den Oord et al. 1985; Skinner et al. 1997) and polarization
variability (Hayes 1985; Taylor et al. 1991). The most widely studied is the lpv of
optical lines, which manifests itself by variations in emission and absorption-line
strength, by changes in velocity of absorption cores and emission peaks, as well as
by the appearance of Discrete Absorption Components (DACs).
DACs are the most prominent feature of P Cygni's wind variability. For a long time these features were thought to be a permanent characteristic of the spectrum since they were readily recognized in almost all snapshot observations. In addition, DACs were detected not only in several unsaturated resonance and subordinate lines of several highly ionized species in the UV as well as in Hff which is the situation for other luminous OB stars (Prinja et al. 1992; Henrichs et al. 1994; Kaper et al. 1997) - but also in a variety of lines of different excitation and ionization in the UV and optical domain (Israelian & de Groot 1999). In view of this apparent ubiquity of the phenomenon the result reported by Stahl et al. (1994) was really a surprise: no DAC was found in any of the CCD spectra taken over the years 1990 to 1992. Motivated by this finding Israelian et al. (1996) studied the variability of UV metal lines and found that these lines still showed DACs in 1990. To bring all these results into line one must assume, either that the majority of DACs detected in the optical by previous investigators were not real but due to unrecognised photographic noise, or that the activity of the mechanism that causes the appearance of DACs in the optical has changed and that, as a result, the components disappeared. It is also possible that for some reasons, e.g. relatively low spectral resolution (R = 12000), Stahl et al. failed to detect any DACs.
Another kind of lpv in P Cygni consists of variations in velocity of the absorption troughs of optical lines. Markova (1993) found from 1981 observations that high-excitation lines, such as OII, NII, SiIII, exhibited smooth variation in the position of the maximum absorption depth with an amplitude of 30 to 50 km s-1 and a peak-to-peak timescale of 100 to 120 days. Similar variations, but on the shorter time-scale of 50 to 60 days, were established by Kolka (1989, 1991) for weaker lines of HI and HeI in 1982 and 1989. Long-term observations suggest that variations on different time-scales may coexist in the lines (Stahl et al. 1994). The nature and the origin of this kind of lpv is not known at present.
Variations of up to 30 per cent in emission and absorption-line strength (i.e. equivalent width, EW) were measured on various occasions (Luud 1967a,b; de Groot 1969). Long-term observations with high time resolution (Markova 1986a, 1993; Stahl et al. 1994) showed that line-strength variations, at least in the strongest Balmer and HeI lines, do not seem to be accidental but follow a definite pattern of behaviour. Neither the origin of these variations nor how they affect other spectral lines formed deeper in P Cygni's wind, are known at present.
Thus, we see that our knowledge of lpv in P Cygni is far from complete. Most of the above-quoted studies share one and the same failure: they focus only on a few particular aspects of lpv in particular ions, e.g. DAC propagation, radial-velocity or line-strength variations, etc. For this reason, none of the investigators were able to make a direct comparison between different kinds of lpv in order to establish possible relationships between them and, thus, to look for the origin of the variability.
The purpose of our work is to investigate the temporal behaviour of a large sample of optical lines of various excitation and ionization in order to classify the different kinds of lpv, to specify their properties, and to search for possible relationships between them. Studying lines formed in different layers of P Cygni's wind simultaneously we hope to gain a better understanding of the nature and the origin of P Cygni's wind variability. Additionally, we aim to clarify the situation for the optical DACs in 1990. In Sect. 3.1 some characteristics of the DAC phenomenon specific to P Cygni are highlighted and the appearance of DACs in the 1990 photographic spectra is examined. In Sects. 3.2, 3.3 and 3.4 we elaborate on the simultaneous behaviour of different lines in the spectrum. In Sect. 4 we discuss the nature of the different kinds of lpv found, with special emphasis on the "swaying'' variability and the DAC-induced variability. The relationship between the "swaying'' variability and the photometric variability is examined. Several qualitative similarities between the wind variability of P Cygni and the B-type supergiant HD 64760 are noted.
Copyright The European Southern Observatory (ESO)