4 Discussion

The present work follows a long series of observations to track the history of P Cyg's giant eruptions in 1600 and 1655. Following the first attempts in this direction ([Feibelman 1995]), it was clear that the star to its nebula brightness ratio, constitutes a challenging obstacle for the optical mapping of the nebular morphology. Indeed only few observations, with heterogeneous angular resolutions are reported in the literature. Leitherer and Zickgraf ([Leitherer et al. 1987]) first published the detection of P Cyg's extended nebulosity using CCD imaging. Later, Johnson et al. ([Johnson et al. 1992]) reported the detection of forbidden emission lines due to nitrogen enriched material at 9 arcsec. The first coronagraphic imaging of P Cyg from the ground was obtained by Barlow et al. ([Barlow et al. 1994]) from high resolution long-slit data. Barlow and co-workers discovered the presence of a 22 arcsec nearly circular shell which presents bright condensations of about 2 arcsec wide, mostly in the North (their plate scale is of 0.255 arcsec per pixel where a 4.3 arcsec occulting strip was used). STScl coronagraph imaging, using a 4.8 arcsec occulting disk, shows almost the same features ([Nota et al. 1995]). From another point of view P Cyg constitutes an ideal target for testing new and different imaging techniques. Among these there are the spectral-line image sharpening techniques SCASIS ([de Vos 1994]), the AMOS adaptive optics observations ([Morossi et al. 1996]), and occulting mask imaging (4 arcsec) with the new MOMI instrument for wide field imaging ([O'Connor et al. 1998]). In the latter case the authors suggest highly asymmetrical features at 3-4 pc from the star (7') probably associated with previous mass-loss events ([Meaburn et al. 1999]). At the same time, radio imaging now offers approximately the same panel of field and spatial resolution as optical imaging making the comparison of optical and radio maps possible. Indeed, sounding different scales in radio wavelength can easily be achieved by changing the baseline configuration of radio arrays ([Skinner et al. 1998]). Using this possibility, Skinner and co-workers have compared radio maps to Barlow's coronagraphic image of P Cyg ([Barlow et al. 1994]). These authors claim that the emissive regions in radio and visible are roughly the same, although this comparison is further complicated by the coronagraphic mask and the telescope diffraction pattern. They conclude that both radio and optical maps exhibit the same details having the same physical origin: i.e. dense clumps overtaken by the faster wind and heated by shocks.

Low and intermediate spatial resolution images suggest a global spherical expanding envelope but clumpiness is present in each case. This trend is also present at higher spatial resolution of the 250 km MERLIN centimetric network ([Skinner et al. 1997]) approaching the submilliarcsecond observations of the GI2T optical interferometric ([Vakili et al. 1997]).

In this context our present reconstructed image of P Cyg's environment in H$_{\alpha }$ presents, as expected, strongly clumped features within the 1 arcsec field of view (0.011 pc with D = 1.8 Kpc). More than 6 enhanced emission clumps are counted with our 0.05 arcsec spatial resolution in a nearly 0.6 arcsec region superimposed to the unresolved central star. The mean size of the clumps is roughly 0.08 arcsec which is the angular limitation of the 152 cm OHP telescope. These results agree well both in size and morphology with MERLIN observations, at nearly the same resolution. The typical diameter of emitting regions for MERLIN is 0.4 arcsec (0.13 arcsec for the core), and amazingly comparable to an optical structures (Fig. 9). In this same figure, a North-East/South-West preferential axis appears in the H$_{\alpha }$ image due to the grouping of the clumps, the distribution being otherwise rather uniform. The same orientation was also pointed out in SCASIS observations at a lower spatial resolution ([de Vos 1994]). Note that in our reconstructed image a bright feature is located at 80 mas South-East of the central star. We can speculate on its relation to the local strong emission discovered at 0.8 mas from the star in August 1994 by the GI2T interferometer ([Vakili et al. 1997]) although the E-W absolute position of the latter emission was not given by GI2T. If this scenario holds, this position, some 3.2 years later, implies a projected velocity around 110 km s^-1. Taking into account the radial velocity and uncertainties $208\pm$ 78 km s^-1 obtained by Vakili et al. ([Vakili et al. 1997]), this projected velocity is to be expected for a clump nearly on the line of sight ejected three years earlier with the terminal velocity, and thus, compatible with the GI2T observations. Although the possible physical relation of 1994 blob and 1997 clump remains to be robustly settled, interferometric and AO imaging repeated in the future, should enlighten such scenarios.

At present, only radio observations by Skinner et al. ([Skinner et al. 1997]) present confident temporal variations. In the two 6-cm MERLIN images taken in a 40 day interval, impressive changes were observed, corroborated by VLA observations. In the observed region, the wind velocity suggests a 2 year dynamical time scale, which can hardly be compared to the 6-cm flux variations. On the other hand, the recombination time scale for hydrogen atoms ( $1.2 10^5/ n_{\rm e}$ in years) is shorter, but not sufficient, $\sim$160 days considering a characteristic $n_{\rm e}$ of 2.8 10⁵ cm^-3 at 0.07 arcsec. This short time scale puts strong constraints on the electron density, which has to be four times larger than the surrounding envelope material. The clumpiness can explain such a time evolution if the structures are sufficiently small and dense, or if the shocks between the wind and the clumps are strong enough. The question is whether increasing resolution would reveal the same clumpiness, and if activity observed in optical and radio wavelengths is closely correlated.

Some questions arise. How can these small scale clumps be related to the 2 arcsec ones observed in the Barlow's images at 3 arcsec from the star? How can they survive over such a long distance? Do they reappear at the location we detect them?

A challenging issue is now understanding the connection between the different spatially resolved structures and their scales which needs the monitoring of the clumpiness from the star to the interstellar medium, and constrains the mass ejection dynamics. Therefore, our present observations are the first attempt to prove that an optical monitoring of the clumpiness in inner region of P Cyg's mass loss is observationally feasible. As previously pointed out ([Vakili et al. 1997]), the intermediate regions of P Cyg's wind, from a few stellar radii to a few parsecs can be sounded by means of AO plus coronagraphic imaging from the ground, in relation with radio observations. The dynamical time scale for optical interferometry is approximately one month, but a temporal monitoring of P Cyg by this technique requires both higher sensitivity and larger numbers of baseline orientations due to the complex structures which occur at different scales. For larger distances, the recombination time scale in the clumpy wind should produce large effects as detected in radio, and AO becomes the perfect technique to follow such activities. The brightness of P Cyg and its evolutionary time scale allow the development of a multi-site and multi-wavelength observations campaign, using AO, optical and radio interferometry to get a unified picture of P Cyg's environment physics.