The Wolf-Rayet (WR) star EZ CMa=HD 50896=WR6 is one of the brightest WR stars in the sky. It started to receive a great deal of attention soon after van den Heuvel (1976) proposed that the evolution of massive stars is expected to lead to the formation of binary systems in which WR stars have a closely orbiting compact companion (cc). The periodic variability discovered in this star (Firmani et al., 1980), its large distance from the galactic plane, and its surrounding ring nebula made HD 50896 an excellent candidate for such a WR+cc system. However, as further observations were made, it was discovered that the variability seemed not to be coherent over long timescales (Robert et al.1992). In addition, analyses were performed which indicated that the observed levels of X-ray emission were not consistent with the hypothesis of an accreting compact companion (Vanbeveren et al.1982; Stevens & Willis 1988). Furthermore, ROSAT IPCS variations reported by Willis & Stevens (1996) do not show any modulation on the 3.76 day cycle. Hence, the very binary nature of HD 50896 began to be questioned, and numerous alternative scenarios for producing the variability in the system were put forth. These alternative scenarios include wind instabilities and non-isotropic wind structures. Reviews of the different scenarios and the observations which have been accumulated over the years of this system can be found in Morel et al. (1997, 1998) and St-Louis et al.(1995).
EZ CMa is a WN5 star (van der Hucht et al.1981) with very strong emission lines present in its spectrum. The notable emission line profile variability lead Firmani et al. (1980) to identify a period of 3.76 days for these variations, a period which was later confirmed to be present in the photometry (Cherepashchuk1981) and polarimetric observations as well (McLean1980). Although this basic period is maintained within most data sets, a coherent phase dependence is apparently not present from one data set to another one, when timescales are larger than a few weeks. This is the most damaging argument against the binary scenarios, since variability due to orbital motion would not be expected to present phase jumps between different observing epochs.
Recently, however, Ivanov at al.(1999) have argued that some of the features in the spectrum of HD 50896 undergo exactly the same variations at all epochs, and discuss the implications of this for the various non-binary scenarios. Also, Morel et al.(1997) confirm the persistent nature of the variations in the strength of the P Cyg absorptions of the NV 4604-20 lines. In this paper we combine different data sets, and using an improved period for the variability in HD 50896, we show the variations to be coherent over long timescales thus reopening the possibility of a binary scenario for this intriguing object.
© European Southern Observatory (ESO) 1999
Online publication: June 30, 1999