2000A&AS..144..391M

High-resolution photographic spectra as well as high S/N ratio CCD spectral observations of P Cygni are analyzed in terms of line-profile variability (lpv). Four different kinds of lpv are established: systematic variability in the absorption troughs of low and intermediate excitation lines due to propagating Discrete Absorption Components (DACs); ``swaying'' variability consisting of continuous modulations in velocity and intensity of the absorption cores and emission peaks of lines of intermediate and high excitation; red-emission-wing variability due to travelling ``bumps'', and long-term (LT) variability in HI and HeI lines of relatively large optical depth. DAC propagation is a slow variation of P Cygni's stellar wind. The components probably originate from large-scale, high-density (low-excitation) perturbation(s) which develop in a relatively outer part of the wind (V ≥0.41V_inf) but appear to be maintained, in some indirect way, by photospheric processes. The geometry of the structures is not yet clear but they could be either spherically symmetric or curved, like kinks. The ``swaying'' variability manifests itself by modulations in position and intensity of the absorption cores and emission peaks of almost all lines in the optical. Simultaneous variations in emission and absorption line-strength were also observed. The modulations are at least partially due to variations in the number density which affect all layers of the supersonic wind starting at its base up to layers where the Hff line forms (0.18≤V≤0.95V<sub>inf</sub>). The phenomenon appears to be stable over many years, though on a variable time-scale. Suggestive evidence for a close relationship between the modulations and changes in the stellar brightness and temperature was found, indicating that the ``swaying'' variability is more likely coupled to processes in the photosphere. Non-radial pulsations (NRPs) either of g-mode or of s-mode oscillations are a possible cause for this variability. The LT variability makes up a very slow pattern of variation in both the velocity of the absorption cores and the intensity of the emission peaks of the stronger HI and HeI lines. This variability is found only in the outer part of the wind (V ≤0.82V_inf). The nature of the LT variation is not known at present. The red-emission-wing variability is localized in the high-velocity part of the emission lobes of P Cygni-type profiles, +90≤V≤+230/250km/s. This variability is presumably caused by outward propagating ``bumps'', but its exact nature is still unknown. No indication for any clear relation between different kinds of lpv was found. Even when the variations operate in one and the same region of the wind (in velocity space) it is not obvious whether and how they interact. Stellar rotation does not seem to play a fundamental role in setting the time-scale of either the DAC-induced variability or the LT variability. The relationship between the ``swaying'' variability and the rotation is still not clear but it is possible that this variability is rotationally modulated. The wind variability of P Cygni appears to be qualitatively similar to the wind variability of the early B-type supergiant HD 64760 but different from the variability of O-type star winds.