1 Introduction

The different evolutive stages of massive stars have strong influence on their surroundings and, consequently, on the galactic evolution. During their lives, the massive stars inject large amounts of matter, energy and momentum into the interstellar medium (ISM). In the pre-main sequence, massive stars undergo a phase of very energetic mass loss giving rise to molecular outflows. UV radiation, expansion of HII regions and fast stellar winds are present in the main sequence stage. In the short-lived red supergiant (RSG) and luminous blue variable (LBV) stages, the winds become slower and denser and ejections of up to some solar masses may be produced. The Wolf-Rayet (WR) stage, prior to SN explosions, is characterized by a copious mass loss (typically 10^-5 $M_\odot$ yr^-1) driven by a fast (about 2000 km s^-1) and chemically enriched stellar wind. Massive stars are therefore continuously heating, ionizing, shocking and blowing-up the ISM surrounding them. García-Segura & Mac Low (1995) have modeled the evolution of the gas which surrounds massive stars, from the main sequence to the WR stage. They only considered simple models for the mass ejection in the form of stellar winds in the different stages of evolution and predict the presence of large scale structures around the massive stars. Both the blown-up gas and the ejecta form multiple shells as a consequence of the different shockfronts produced inside the gas.

During the main sequence phase, the gas around massive stars is ionized, heated and highly evacuated by the fast, long-lived stellar wind and the Lyman continuum radiation. A hot cavity is surrounded by an expanding shell. The ionization front may get trapped within it, the shell may start to recombine and become observable in the line radiation emitted by neutral atoms or molecules. These large bubbles have been successfully detected in the 21-cm line of atomic hydrogen (Arnal 1992; Rizzo & Bajaja 1994; Cappa & Benaglia 1998; Rizzo & Arnal 1998; Arnal et al. 1999), in the form of huge shells of 30-70 pc, expanding at 6-20 km s^-1, with masses of hundreds (or even thousands) of solar masses. Some of the large IRAS shells found by Marston (1996) might also be related to this type of structures.

Once the massive star finishes the hydrogen burning in its nucleus and becomes a RSG or LBV, its stellar wind suddenly becomes more dense, and a significant amount of the stellar mass is deposited into the ISM at velocities around 20 km s^-1. Very probably, the formation of the optical ring nebulae starts in this brief stage. When the star comes to a WR stage, the wind is accelerated to hypersonic velocities and rapidly reaches the RSG/LBV wind. Many of the optical nebulae possess arcs and have high inhomogeneities in the abundances of heavy elements (Chu et al. 1983, 1999).

However, the effects of the WR phase in the neutral (neither atomic nor molecular) gas have not been easily detected. The structures formed during this phase would not be as large as those developed in the O-phase. Just in few cases, the neutral atomic hydrogen has been detected in interaction with ring nebulae (Cappa et al. 1996; Arnal & Cappa 1996). Vibrationally excited molecular hydrogen has only been detected towards NGC2359. The CO J=  $1 \rightarrow 0$ emission has been observed around WR16 (Marston et al. 1999) and in NGC2359 (Schneps et al. 1981).

NGC2359 is a very interesting object, and many aspects of which has been thoroughly studied in the recent years. This optical nebula is excited by HD56925 (WR7 in the catalogue of van der Hucht (2000), a WN4 star located in the outer Milky Way. Its distance varies from 3.5 to 6.9 kpc depending on the authors (see Goudis et al. 1994 for a discussion). The mass-loss rate ( $<7\; 10^{-5}$ $M_\odot$ yr^-1) and the terminal wind velocity (1545 km s^-1) have been measured by Schneps et al. (1981) and Rochowicz & Niedzielski (1995), respectively. NGC2359, a prototypical wind-blown bubble (Chu et al. 1983), is nearly spherical with several small filaments inside. The nebula is not highly enriched of heavy elements as a whole, with the exception of the filaments (Esteban et al. 1990). This may indicate that the nebula was largely produced in the RSG/LBV and the WR stages of the exciting star. From the optical emission lines (Goudis et al. 1983) and radio recombination lines (Lockman 1989; Fich & Silkey 1991), we know that the ionized gas has several components at velocities between 30 and 70 km s^-1 (all velocities in this paper are referred to LSR). An HII region partially surrounds the wind-blown bubble and is apparently limited by obscuring material. Schneps et al. (1981) have taken several profiles in the J=  $1 \rightarrow 0$ line of CO and detected the presence of three clouds at 37, 54 and 67 km s^-1. The presence of broad CO profiles at 54 km s^-1 and the morphology of the emission, just following the south-eastern edge of the nebula, suggests that this component has been interacting with the nebula. St-Louis et al. (1998) have analyzed the H₂ emission in this region. They have found the 1-0 S(1) line towards the southern border of the HII region, but they could not establish the nature of the excitation (fluorescence or shocks) of the H₂. Cappa et al. (Cappa et al. 1999) made a complete map of the ionized component at the 1465 MHz continuum, and traced the HI emission presumably connected with the nebula. These HI features are probably associated with the WR stage of HD56925; so far, the large atomic main-sequence bubble predicted by the theory has not been yet reported.

The goal of this paper is to look into the history of the interaction of the wind and the radiation of HD56925 and its surroundings. We have analyzed a public HI survey and studied for the first time the large scale interaction of HD56925 in the O-phase with the ambient ISM. Based on the mapping of the J=  $1 \rightarrow 0$ and J=  $2 \rightarrow 1$ lines of CO over the whole optical nebula, we study in detail the morphology and kinematics of the CO around NGC2359. We also present several ¹³CO profiles towards selected positions in order to determine the global physical parameters of the molecular gas, such as kinetic temperature, density and opacity of the lines. This study has allowed us to shed some light on the dominant excitation mechanism of the neutral gas in the WR stage.