S 266: a ring nebula around a Galactic B[e] supergiant?

Cesar Esteban1 and Matilde Fernandez2

1Instituto de Astrofısica de Canarias, 38200 La Laguna, Tenerife, Spain2Max Planck Institut fur Astronomie, Konigstuhl 17, D-69117, Heidelberg, Germany

Accepted 1998 March 2. Received 1998 February 10; in original form 1997 November 3

A B S T R A C TA narrow-band Ha image and high-resolution spectroscopy have been obtained in order toinvestigate the nature of S 266 and its central star MWC 137. The analysis of the stellar andnebular spectra suggests that MWC 137 is a B[e] supergiant located $ 6 kpc away from theSun and not a Herbig Ae/Be star, as it has been traditionally classified. Moreover, themorphology and other properties of the nebula suggest that S 266 is a ring nebula, probablyproduced by the interaction of stellar winds with the ambient interstellar medium orunprocessed ejected matter. This result would imply that S 266 is the first ring nebulaaround a B[e] supergiant known in the Galaxy.

Key words: circumstellar matter – stars: evolution – stars: mass-loss – ISM: bubbles – H iiregions – ISM: individual: S 266.

1 I N T RO D U C T I O N

The nebula S 266 was discovered by Sharpless (1959) and firstdescribed as a possible planetary nebula (PN) with the designationPK 195¹0◦1 in the catalogue of Perek & Kohoutek (1967). Thisobject has not drawn much interest in the past because of its locationtowards the Galactic anticentre and its modest optical appearance.The first study devoted to S 266 was by Frogel, Persson &Kleinmann (1972) who obtained near-infrared (NIR) photometricobservations and optical spectra of the object and reported the lowexcitation of the nebula and an excess IR emission around thecentral star, MWC 137. These authors suggested that S 266 isprobably not a PN but a nebulosity associated with an early-typeemission-line star originally classified by Merrill & Burwell (1933).Herbig & Rao (1972) included MWC 137 in their catalogue of starsof the Orion population, albeit with a cautionary note that it may notbe a member of this class. Allen & Swings (1976) obtainedphotographic spectra of a sample of peculiar Be stars with IRexcess finding a large number of emission lines in the spectrum ofMWC 137, including [O i], He i, Fe ii and possibly [Fe ii] amongothers. These authors classify the star as B[e]. However, since thework by Finkenzeller & Mundt (1984), MWC 137 has traditionallybeen considered as a Herbig Ae/Be (HAEBE) star in extensivestudies of large samples of these objects, (e.g. Hillenbrand et al.1992; Hamann & Persson 1992; The, de Winter & Perez 1994;Bergner et al. 1995) although its membership in this class has neverbeen discussed in depth. On the other hand, Fich & Blitz (1984)performed a study of the distribution of optical H ii regions in theouter Galaxy, compiling photometric distances and CO velocities.These authors found that S 266 was the most distant H ii region fromthe centre of the Galaxy in their sample. Fich & Silkey (1991), andmore recently Vılchez & Esteban (1996), have studied the chemicalabundances of the nebula in the context of the abundance gradient

towards the Galactic anticentre, finding consistent results with thebehaviour of the expected abundances at large galactocentricdistances. We present a new narrow-band image as well as high-resolution spectroscopy of the object in order to determine themorphology and dynamics of the nebula and to elucidate the natureof its central star.

2 O B S E RVAT I O N S A N D DATA R E D U C T I O N

2.1 Direct imaging

Direct images of S 266 were obtained in 1995 February with the1.5-m telescope of the Observatorio Astronomico Nacional (OAN)at San Pedro Martir (Baja California, Mexico), using a 1024×1024Thompson CCD. The pixel scale was 0.2 arcsec pixel¹1 and theseeing was 2.8 arcsec. A narrow-band image in Ha was obtainedwith a filter of 10 A full width at half maximum (FWHM) from threeexposures of 300, 420 and 900 s of integration time. Two exposuresof 420 s of the adjacent Ha continuum were also obtained using afilter of 127 A FWHM and centred at 6451 A. Unfortunately, it wasnot possible to flux-calibrate the images because of bad weatherconditions.

2.2 Echelle spectroscopy

These observations were carried out in 1994 October with the 2.1-mtelescope of the OAN at San Pedro Martir (Baja California,Mexico). The spectra were obtained using the OAN echelle spectro-graph, which gives a reciprocal linear dispersion of 9.738 A mm¹1

at Ha when used with the UCL camera. A 1024×1024 ThompsonCCD detector was used with a 19 mm2 pixel size. At Ha, each pixelcorresponds to 0.185 A. The effective spectral resolution was 0.33A FWHM. The cross-dispersion mode was used to obtain spectra

Mon. Not. R. Astron. Soc. 298, 185–192 (1998)

q 1998 RAS

covering from 4500 to 6850 A, including 17 spectral orders. A slitcovering 26.6×1 arcsec2 was used in order to avoid overlappingbetween two consecutive spectral orders in the spatial direction.Two different north–south slit positions were observed. A first 600-s exposure was taken passing through the central star. Two addi-tional 1200-s consecutive exposures were taken 10 arcsec to the eastof the central star.

2.3 Additional long-slit high-resolution spectroscopy

These observations were carried out in 1989 January with the 2.5-mIsaac Newton Telescope of the Observatorio del Roque de LosMuchachos (La Palma, Spain). A 578×400 pixel GEC CCD of 22mm2, attached to the 500-mm camera of the Intermediate Disper-sion Spectrograph (IDS), was used to obtain a spectrum coveringfrom 6500 to 6625 A. A 1800 line mm¹1 grating was used in the firstorder, giving a reciprocal dispersion of 10 A mm¹1 at Ha. Theeffective spectral resolution was 0.44 A FWHM. The slit was300 arcsec long (larger than the optical diameter of the nebula ofabout 1 arcmin) and 1 arcsec wide. The position angle was 1408.Two different exposures were taken passing through the central star,one of 100 s and a second one of 1500 s of integration time.

All the data were reduced using the standard iraf reductionpackage to perform bias corrections, flatfielding, cosmic rayrejection and wavelength calibration (in the case of the spectraldata).

3 R E S U LT S A N D D I S C U S S I O N

The analysis of the echelle spectra was made extracting a single 1Dspectrum for each of the two slit positions observed. The first one –labelled position 1 – corresponds to the central 8 arcsec andincludes the emission of the star and the inner part of the nebula.For the second 1D spectrum – position 2 – we binned the original2D spectrum taken 10 arcsec to the east of the central star along allthe spatial direction. This second extraction corresponds to purenebular emission. Sections of the spectra taken at these two slit

positions are shown in Fig. 1. In the case of the high-resolution IDSspectrum we extracted 11 different zones of 5 or 8 arcsec wide tomap the spatial variation of the Ha profile over the nebula. Radialvelocities, FWHM and equivalent width (EW) of the emission lineswere obtained via single or multiple Gaussian fitting proceduresusing the Starlink dipso software package (Howarth & Murray1990).

3.1 The controversial nature of MWC 137

In Table 1, we show the intrinsic FWHM (corrected for instrumentalprofile) of the different emission lines measured in both echellespectra as well as the EWs of the lines detected in position 1. As canbe seen in the table, many lines are observed only in the spectrum ofthe central star (position 1) and are therefore of stellar origin. Thesestellar lines are all in emission, have different FWHM and havebeen identified as Fe ii, Mg i, Si ii, He i, [O i], H i and Na i (Na i linesare not included in Table 1 because they are strongly affected byinterstellar absorption). All these lines are common in B[e] stars(e.g. Jaschek & Jaschek 1987) as well as in HAEBE stars (e.g.Hamann & Persson 1992). The EW of Ha is remarkably large (254A) and its profile is very broad but symmetric (see Fig. 2). Reason-able fits of the Ha line profile can only be obtained with thecombination of four or more Gaussians of different width, whichis obviously an artificial solution arising from the complexity of theprofile. The maximum FWHM of the different Gaussians fitted isabout 1200 km s¹1 and indicates strong stellar winds. An importantbroad component is also observed in the Ha profile in position 2, aswell as in all the 11 zones extracted from the INT long-slitspectrum. The FWHM of this broad component is similar all overthe nebula and its relative contribution to the total Ha flux decreaseswith increasing the radial distance to the central star. Both factsindicate that this spectral feature is the result of reflection of thestellar emission by dust contained in the nebula. An interestingresult is that the shape of the Ha profile of the stellar emissionappears to be almost identical in both periods of observations, 1989January and 1994 October (see Fig. 2), although the observations

186 C. Esteban and M. Fernandez

q 1998 RAS, MNRAS 298, 185–192

Figure 1. Section of the red spectral range of position 1 (top; star+nebula emission) and position 2 (bottom; nebula emission) of S 266.

are of somewhat different spectral resolution. This fact suggests thestability of the shape, at least in an almost six-year period.

The radial velocities measured for the different nebular lines ofposition 2, the previous INT data and the narrow components ofposition 1 give very similar values without any apparent differencebetween ions. The mean radial velocity is þ1862 km s¹1 whichcorresponds to a velocity with respect to the local standard of rest ofvLSR=þ3062 km s¹1. This value is consistent with previous vLSR

estimates obtained by Fich & Blitz (1984) from the CO velocity ofthe object (31 km s¹1) and Fich, Treffers & Dahl (1990) from Fabry-Perot Ha observations (26.160.5 km s¹1). The kinematic distanceof S 266 can be derived by comparing its vLSR with the establishedvelocity field of the Galactic interstellar gas in the direction of thenebula. We have used the compilation of spectrophotometricdistances and radial velocities of H ii regions and reflection nebulaeby Brand & Blitz (1993). These authors show a complete map of theobserved radial velocities projected on to the Galactic plane (theirfig. 2) and covering the four quadrants, up to 7 kpc towards theanticentre. From this comparison a lower limit of 6 kpc can beestimated due to the tangential behaviour of the isovelocity lines inthe direction of the nebula up of that distance. This result is broadlyconsistent with the distance obtained previously by Fich & Blitz(1984).

The Na i D lines are also present in the spectrum of MWC 137.Na i D2 shows a P-Cygni-type structure, with a component inemission of stellar origin (much brighter than the sky emission)and an absorption component probably of interstellar origin. Onthe other hand, Na i D1 is only seen in absorption. The FWHM ofthe absorption component of this last line is 35 km s¹1 and,therefore, is consistent with the expected behaviour of the velocityfield of the Galactic interstellar gas for large distances towardsS 266, the vLSR of which ranges from 0 to almost 30 km s¹1

Brand & Blitz (1993).

S 266: a Galactic ring nebula? 187

q 1998 RAS, MNRAS 298, 185–192

Table 1. Intrinsic FWHM and EW of emission lines in the echelle spectra.

Emission line Multiplet Component Position 1 Position 2(A) Central star + nebula Nebula

FWHM EW FWHM(km s¹1) (A) (km s¹1)

Hb 4861 3865 7.261.1 –Fe II 5276 49 89624 0.1160.03 –Fe II 5317 49,48 146611 0.5760.06 –[N II] 5754 (3F) 4765 0.2860.04 –He I 5876 11 302613 2.0860.10 –Fe II 6248a 74 124619 0.2560.06 –[O I] 6300 (1F) 5464 0.3260.04 –Mg I 6318 23 175625 0.3360.06 –Si II 6347 2 274635 0.3760.07 –[O I] 6364 (1F) 5464 0.1460.04 –Fe II 6385a 203 171612 0.5260.06 –Fe II 6456 74 228629 0.5260.10 –Fe II 6516 40 161640 0.3060.10 –[N II] 6548 (1F) – – 2564Ha 6563 narrow 3161 3761

broad 1200b 210620integrated 254617

[N II] 6584 (1F) 2961 0.4360.03 3363[S II] 6717 (2F) 2364 0.1560.04 3765[S II] 6731 (2F) 3868 0.1860.04 3766

a Blend of several Fe II lines.b Maximum width of the broad component.

Figure 2. Ha profiles of MWC 137 observed in 1989 January with the INTand 1994 October at SPM. The peak fluxes are normalized to unity. The INTdata have been displaced +0.2 units in ordinates to permit the visualcomparison of both line profiles. The lower line represents the subtractionof both normalized profiles (SPM–INT).

The comparison between the visual extinction of S 266 and thegeneral trend of the interstellar extinction in the line of sight to theobject provides another estimation of the distance. The reddeningcoefficient, C(Hb), has been obtained by Vılchez & Esteban (1996)and is 1.86 for the external parts of the nebula, where the contam-ination of the Balmer lines by stellar emission as well as the internaldust contribution are expected to be lower. Using the relationbetween EðB ¹ VÞ and C(Hb) computed by Kaler & Lutz (1985)and the standard extinction law AV ¼ 3:1 × EðB ¹ VÞ, we obtain AV

= 3.77. From the data of Neckel & Klare (1980), this value seems tobe consistent with the general trend of the interstellar extinction inthe line of sight to S 266 at a distance of about 6 kpc. On the otherhand, the O/H abundance measured by Vılchez & Esteban (1996)for S 266 is consistent with the expected behaviour at such largedistances, and abnormally low if we consider that the object islocated in the Solar vicinity.

We have searched the astrometric data of MWC 137 in theHipparcos Catalogue, finding that the trigonometric parallaxobtained for this object is not useful because of its negative sign(¹4.5464.50 mas), although its measured proper motions are large(¹31.1764.79 mas yr¹1 and 23.7962.57 mas yr¹1, in right ascen-sion and declination, respectively). However, the goodness-of-fitstatistic parameter of the astrometric solution to the accepted datafor MWC 137 (F2, field H30 in the catalogue) is quite high, 1.96.That means the quality of the astrometric solution is poor. Weconsider that the faintness of the object (mV ¼ 11:84, almost at thelimiting magnitude of Hipparcos) and the possible confussion dueto the presence of the nebula may have contributed to the uncer-tainty of the results. In fact, the difficulty of parallax determinationsby Hipparcos in nebulous regions has been pointed out by van denAncker et al. (1997). These authors obtain inconsistent results forthe distance of the HAEBE star BD+40◦4124, which is embeddedin a reflection nebula.

A distance of 900 or 1300 pc has been traditionally adopted forMWC 137 in many works assuming the HAEBE status of this star(Hillenbrand et al. 1992; Berrilli et al. 1992; Hamann & Persson1992; Skinner, Brown & Stewart 1993; Pezzuto, Strafella &Lorenzetti 1997; Di Francesco et al. 1997; Testi et al. 1997). It isnecessary to emphasize that neither of those distance estimates of900 pc (Higgs 1971) and 1300 pc (Maciel & Pottasch 1980) areappropriate. In fact, they have been derived using statisticalmethods only applicable for true PNe. However, it has been welldemonstrated that S 266 is not a member of that group of objects, infact it is included in several lists of misidentified PNe (Acker et al.1987; Zijlstra, Pottasch & Bignell 1990; Acker et al. 1992). Makinguse of the inappropriate distance of 1300 pc, Hillenbrand et al.(1992) derive a luminosity of 28 800 L( for MWC 137, one of thehighest values found for typical HAEBE stars.

We have made a reassessment of the parameters of the central starusing the available observational data and adopting our lower limitof 6 kpc for the distance of MWC 137. An upper limit of MV can bederived from the observed mV = 11.84 mag (taken from thecompilation by Hillenbrand et al. 1992), the lower limit of thedistance obtained from our echelle spectra and the visual extinction.From these data we obtain MV # ¹5.82 mag. Using the calibrationon spectral classification by Schmidt-Kaler (1982), we estimate thatthe spectral type of MWC 137 should be an early B supergiant. Weobtain a bolometric magnitude of Mbol # ¹8.66 mag, making use ofthe bolometric correction for B supergiants given by Underhill &Doazan (1982). The corresponding luminosity is L ¼ 2:36 × 105

(D/6 kpc)2 L( (where D is the distance in kpc). On the other hand,the presence of ionized gas of low ionization degree imposes

restrictions on the effective temperature of the star and thereforeon its spectral subtype. From the results of the analysis of thenebular spectrum of S 266 with photoionization models, Vılchez &Esteban (1996) estimate an effective temperature of <30 000 K forMWC 137. From the estimated value of effective temperature andluminosity and the theoretical HR diagrams by Schaller et al.(1992), we obtain a minimum initial mass of about 30 M( forMWC 137.

As was noted in Section 1, MWC 137 has been traditionallyclassified as an HAEBE star since the work by Finkenzeller &Mundt (1984). However, these authors include a cautionary noteabout its strong and broad Ha emission and unusual radio emission.The lower limits of the absolute magnitude and initial mass we haveobtained for MWC 137, are not consistent with HAEBE status forthe star but are similar with those derived for B[e] supergiants of theMagellanic Clouds (Zickgraf et al. 1986) and the very few GalacticB[e] stars that are certainly supergiants (see Zickgraf 1992 andreferences therein). HAEBE and B[e] supergiant stars share severalcommon observational charactersitics which are present inMWC 137 as, for example, the presence of lines of H i, He i,Fe ii, [Fe ii] and [O i] in the stellar spectrum as well as strong IRexcess as a result of thermal radiation of circumstellar dust.However, the spectrum of MWC 137 resembles most closelythose of the B0 supergiants R126 (Zickgraf et al. 1985), Hen 134(Zickgraf et al. 1986) and S111 (Stahl et al. 1989), all in the LargeMagellanic Cloud (LMC). All these stars show pure emission linespectra with similar lines and widths. Moreover, our EW(Ha) = 254A (quite similar to the value of 240 A obtained by Swings &Andrillat 1981; and somewhat larger than the 133 A obtained byFinkenzeller & Mundt 1984) is larger than the EW(Ha) reported byFinkenzeller & Mundt (1984) and Fernandez et al. (1995) forsamples of HAEBE stars, but comparable to the typical values of102 –103 A obtained for B[e] supergiants (Zickgraf 1992).

Bergner et al. (1995) report a variability of D V ¼ 0:15 mag forMWC 137. This behaviour fits the usual variability of B[e] super-giants (Zickgraf 1992) and Herbig Be stars (The et al. 1994), whichis of the order of 0.1 to 0.2 mag in both cases. In contrast, Bergner etal. (1995) suggest that the variability of 0.6 mag detected in the NIRfrom JHK photometry is more typical for binary objects amongB[e]-type stars. As it was commented above, the spectroscopicvariability seems not to be striking in MWC 137. The reportedEW(Ha) seems not to change by more than a factor of 2 and theshape of the Ha line profile appears to be stable, as shown in Fig. 2.This result also makes MWC 137 more similar to B[e] supergiants.In fact, according to Zickgraf (1992), the spectroscopic variationsof the B[e] supergiants are generally weak. In the case of HAEBEstars, spectroscopic variability seems to be a common behaviourand can even be spectacular in some cases (Catala 1994). Forinstance, Ha line profiles can change drastically on a time-scale of afew months, going from a double-peak emission to a P Cygni profileor from a P Cygni profile to a single-peak emission.

Very recently, Testi et al. (1997) have obtained NIR images of thefield around MWC 137 and other HAEBE stars. They find apopulation of faint NIR sources associated with MWC 137 (unde-tectable at shorter wavelengths); this population has the largestspatial density in all their sample (1.5 × 103 stars pc¹3). However,Testi et al. (1997) adopt the traditional distance of 1300 pc forMWC 137. In any case, we do not consider this result as a proof ofthe pre-main sequence status of MWC 137. In fact, the shorttimelife of a Minitial $ 30 M( star (less than 10 Myr, Schaller etal., 1992) would permit the presence of a post-main sequencemassive star in a young cluster. In this sense, we are interested in

188 C. Esteban and M. Fernandez

q 1998 RAS, MNRAS 298, 185–192

the result obtained by Hillenbrand et al. (1995) who identify astellar aggregate associated with the HAEBE star BD +40◦4124.This cluster is characterized by an apparent age spread of approxi-mately 3 Myr, larger than the time-scale for a B star (or even an Astar) to reach the main-sequence phase (about 105 years), and of theorder of the lifetime of the most massive stars.

In summary, we consider that the analysis and comparison of theobservational results gathered on MWC 137, the central star of

S 266, favour the B[e] supergiant nature of the object and, therefore,its post-main-sequence status.

3.2 The nebula

In Fig. 3, we show the Ha continuum-subtracted image of S 266.The nebula has a clear filamentary shell morphology with a fainterdiffuse emission filling its inner part and apparently not detached

S 266: a Galactic ring nebula? 189

q 1998 RAS, MNRAS 298, 185–192

Figure 3. Continuum-subtracted Ha image of S 266. North is at the top and east is on the right.

from the star. In fact, the Ha stellar emission of MWC 137dominates the integrated flux of the object and prevents the inner-most parts of the nebula from being seen. S 266 has an ellipticalshape elongated towards the north. The angular size of the shell is77×51 arcsec2 which corresponds to a linear dimensions of 2.2 (D/6 kpc) pc × 1.5 (D/6 kpc) pc. The surface brightness is not uniform;there is a dominant arc-shaped condensation to the NE and anotherbright condensation to the SW of the southern shell. These twobright condensations are aligned with respect the central star. Thepresence of an arched filament to the NW of MWC 137, whichseems not to be connected with the main shell, is striking. Thiscurious feature suggests a complex formation process of the shell ora highly inhomogeneous spatial distribution of the circumstellarmatter. It is interesting to note that the morphology of the nebuladoes not resemble the typical ionization fronts and bright rimsfrequently associated with young star-forming regions. Thistypical morphology can be seeing, for example, around theHAEBE stars BD +40◦4124 (Hillenbrand et al. 1995) and AFGL4029 (Deharveng et al. 1997).

Two of the traditional membership criteria to classify an object asa HAEBE star are its location in an obscured region and thepresence of a bright reflection nebula in its inmediate vicinity(e.g. Catala 1989; The et al. 1994). A close inspection of thePOSS plates some degrees around S 266, does not reveal anyevidence for the presence of dark clouds or other bright nebulae.High-resolution radio continuum observations of S 266 (Israel1976; Fich 1993) show that the radio emission fits closely theoptical morphology of the nebula. Moreover, the results of the COsurvey of optical H ii regions of Blitz, Fich & Stark (1982) indicatethat the source is unresolved and centred on the optical position ofMWC 137 with a diameter less than 1 arcmin, the optical angularsize of the nebula. As we can see, the observational evidencessuggest that the nebula is isolated and not inmersed in a large-scalestar-forming region. However, the optical morphology of S 266 issimilar to nebulae surrounding Galactic LBV stars: AG Car (Nota etal. 1992), WRA 751 (Hutsemekers & Van Drom 1991), the LBVcandidate He 3-519 (Smith, Crowther & Prinja 1994); as well as tothe ring nebula RCW 58 surrounding a WN8 star (Smith et al.1988). All these objects show an extended and almost completebright ring around a diffuse emission centred on the ionizing centralstar. On the other hand, the low-ionization spectrum of S 266(Vılchez & Esteban 1996) is very similar to those observed in theWR ring nebula M1-67 (Esteban et al. 1991) and the LBV nebulaAG Car (Smith et al. 1997). Moreover, some contribution of dustscattering in the recombination lines of the stellar spectrum in thenebula is also reported in AG Car (Mitra & Dufour 1990; Smith etal. 1997).

The narrow nebular lines observed in all the spectra do not showany evidence of line splitting in their profiles, but their FWHM of<35 km s¹1 is clearly superthermal. This result indicates that thenebula is stalled or very slowly expanding with velocities notlarger than <15 km s¹1. In fact, previous photographic redechellograms of S 266 taken by Sabbadin & Hamzaoglu (1981)reveal partial splitting of the nebular emission, which indicates anexpansion velocity of 10–15 km s¹1. Assuming that the nebulaexpands at such a velocity and considering the linear dimensionsquoted above, we estimate a tentative dynamical age of about 105

(D/6 kpc) yr for the shell of S 266, of the order of the lower limit ofthe lifetime of the B[e] phase, 105 yr, estimated by Zickgraf (1992)for B[e] supergiants of the MCs. This result suggests that thenebula could have been formed just before or at the beginning ofthe B[e] phase.

The ionized mass of S 266 can be derived from the integratedradio continuum flux of 0.09 Jy at 5.89 GHz obtained by Fich(1993); the electron temperature of 9900 K obtained by Vılchez &Esteban (1996); the rms density of 160 cm¹3 from Israel (1976) andassuming the standard He/H fraction of 0.10. With all these data, weobtain a nebular ionized mass of 7.2 (D/6 kpc)2 M(. This value is ofthe order of the typical mass of an ejecta nebula such as the oneassociated with AG Car, 4.2 M(, the nebulae around He 3-519, 2M(, and ejecta nebulae around Galactic Wolf–Rayet (WR) stars,between 1 and 20 M( (Esteban et al. 1992). The correspondingionizing flux turns out to be 8.2 × 1047 (D/6 kpc)2 s¹1, which is afactor of 4 lower than the expected ionizing flux for a B0 supergiant(Panagia 1973).

The most remarkable difference between S 266 and other ejectanebulae around massive stars is their chemical content. Smith et al.(1997) and Hutsemekers & Van Drom (1991) find strong Nenrichment in the chemical abundances of the nebulae associatedwith AG Car and WRA 751. In fact, theoretical calculations for theinteraction of a 60-M( star with its circumstellar medium from theO-star to the LBV phase by Garcıa-Segura et al. (1996a) predictvery high N and He content in the material of the ejecta. In the caseof S 266, Vılchez & Esteban (1996) do not find compelling evidenceof N enrichment in the shell. The same situation has been found forthe ring nebula around the LMC B2 supergiant Sk¹69 271. Weis etal. (1997) find a normal [N ii]/Ha ratio in the nebula; they concludethat the shell consists mainly of interstellar material and classify itas an interstellar bubble. The nebula around Sk¹69 271 has a lineardiameter of 5.3 pc and an expansion velocity of 24 km s¹1,corresponding to a dynamical age of 2 × 105 yr. As can be seen,S 266 may be the Galactic analogy of this object. On the other hand,it is surprising that only one nebula around a B supergiant is knownin the Galaxy. This object has been discovered very recently(Brandner et al. 1997). The central star is the B1.5 supergiantSher 25, located in the giant Hii region NGC 3603. The nebulaconsists of a clumpy ring and a bipolar outflow. The ring has asemimajor axis of 6.9 arcsec, which corresponds to a lineardiameter of 0.4 pc; much smaller than S 266 but very similar tothe inner ring around SN1987 A. Its high [N ii]/Ha ratio indicatesan enhanced N abundance. These abundance differences amongnebulae associated with B supergiants are also common in the classof ring nebulae around WR stars (Esteban et al. 1992) and areprobably related to different evolutionary stages and/or ambientinterstellar medium conditions.

The status of B[e] supergiants in the stellar evolution scheme isstill not clear. B[e] supergiants and LBVs are located nearly in thesame region of the HR diagram of the MCs, but it is not clearwhether an evolutionary link exists between both kind of objects.Accordingly to Zickgraf (1992), both should represent a short-livedstage in post-main-sequence evolution of massive stars. Moreover,Conti (1997) argues that the presence of a disc is an indication thatthe B[e] supergiants have still not passed the red supergiant (RSG)phase and proposes that B[e] supergiants might evolve to WR stars.In this context, we have two possibilities for the origin of the shell,depending on the evolutionary status of the central star. First, ifMWC 137 is a young post-main-sequence object at presentmoving redward out of the main sequence (accordingly to Conti’ssuggestion), it has not experienced massive ejections of stellarmaterial and the shell should be composed of swept-up ambientinterstellar medium. In this case, the shell appears to be too small tobe the fossil of a bubble created by the central star during the main-sequence phase (that should be acting during several Myr if the starhas Minitial $ 30 M(). In fact, Garcıa-Segura et al. (1996b) obtain a

190 C. Esteban and M. Fernandez

q 1998 RAS, MNRAS 298, 185–192

final radius of about 38 pc for the main-sequence shell created by a35-M( star (assuming an ambient density of 20 cm¹3), much largerthan the radius of 1 (D/6 kpc) pc found for S 266. This apparentcontradiction could be explained if there were an abnormal highdensity or pressure in the ambient gas around the nebula; this wouldproduce the early stalling of the main-sequence bubble. Secondly, ifMWC 137 is an evolved post-main-sequence object and is movingblueward in the HR diagram, it should have experienced massiveejections in its previous RSG phase (if Minitial < 50 M() or LBVphase (if Minitial > 50 M(). In both cases, we expect a high N contentin the nebulae, which is not observed. This chemical anomaly canonly be avoided if the star has suffered massive ejections ofunprocessed stellar material during the RSG stage or at the begin-ning of the B[e] phase. This possibility is in contradiction to currentmodels of ring nebula formation and stellar evolution. In any case,there have been no reported composition anomalies in B[e] stars(Conti 1997), a fact that supports the non-post-RSG status of thesekinds of stars.

4 C O N C L U S I O N S

New narrow-band Ha image and high-resolution spectroscopy ofS 266 have been obtained in order to investigate the nature of thenebula and its central star, MWC 137. A lower limit of 6 kpc hasbeen derived for the distance to the object from the comparison ofits vLSR with the velocity field of the Galactic interstellar gas. Afteranalysis and discussion of the observational data, it is concludedthat MWC 137 should be a B[e] supergiant with L ¼ 2:36 × 105 (D/6 kpc)2 L(, probably of spectral subclass B0, and not an HAEBEstar, as it has been traditionally classified. The narrow-band Ha

image shows that S 266 is a filamentary nebula with a shellmorphology of linear dimensions 2.2 (D/6 kpc) pc × 1.5 (D/6kpc) pc. The analysis of the spectrum and the morphology of thenebula suggests that S 266 is a ring nebula around a Galactic B[e]supergiant, probably produced by the interaction of stellar windswith the ambient interstellar medium or unprocessed ejected matter.This result would imply that S 266 is the first ring nebula around aB[e] supergiant known in the Galaxy.

AC K N OW L E D G M E N T S

We thank D. I. Mendez for his help in the reduction of the echellespectra and C. Abia, J. M. Vılchez and R. Rebolo for taking the INThigh-resolution spectra. We thank L. Cuesta for his help andpermission to use the graficos program. We are grateful to adiligent anonymous referee who made interesting suggestions forimproving the paper. This research was partially funded throughgrant PB90-0570 from the Spanish Direccion General de Investi-gacion Cientıfica y Tecnica of the Spanish Ministerio de Educaciony Ciencia and by the Mexican DGAPA-UNAM project IN101794.The observations were made at the Observatorio AstronomicoNacional at San Pedro Martir, Baja California, Mexico.

R E F E R E N C E S

Acker A., Chopinet M., Pottasch S. R., Stenholm B., 1987, A&AS, 71, 163Acker A., Ochsenbein F., Stenholm B., Tylenda R., Marcout J., Schohn C.,

1992, Strasbourg-ESO Catalogue of Galactic Planetary Nebulae. ESO,Garching bei Munchen

Allen D. A., Swings J. P., 1976, A&A, 47, 293Bergner Y. K., Miroshnichenko A. S., Yudin R. V., Kuratov K. S., Mukanov

D. B., Shejkina T. A., 1995, A&AS, 112, 221

Berrilli F., Corciulo G., Ingrosso G., Lorenzetti D., Nisini B., Strafella F.,1992, ApJ, 398, 254

Blitz L., Fich M., Stark A. A., 1982, ApJS, 49, 183Brand J., Blitz L., 1993, A&A, 275, 67Brandner W., Grebel E. K., Chu Y.-H., Weis K., 1997, ApJ, 475, L45Catala C., 1989, in Reipurth B., ed., ESO Workshop, Low Mass Star

Formation and Pre-Main-Sequence Objects. ESO, p. 471Catala C., 1994, in The P. S., Perez M. R., van den Heuvel P. J., eds, ASP

Conf. Ser. 62, The Nature and Evolutionary Status of Herbig Ae/BeStars. Astron. Soc. Pac., San Francisco p. 91

Conti. P. S., 1997, in Nota A., Lamers H. J. G. L. M., eds, ASP Conf. Ser.120, Luminous Blue Variables: Massive Stars in Transition. Astron. Soc.Pac., San Francisco p. 161

Deharveng L., Zavagno A., Cruz-Gonzalez I., Salas L., Caplan J., CarrascoL., 1997, A&A, 317, 459

Di Francesco J., Evans N. J., II, Harvey P. M., Mundy L. G., Guilloteau S.,Chandler C. J., 1997, ApJ, 482, 433

Esteban C., Vılchez J. M., Smith L. J., Manchado A., 1991, A&A, 244,205

Esteban C., Vılchez J. M., Smith L. J., Clegg R. E. S., 1992, A&A, 259,629

Fernandez M., Ortız E., Eiroa C., Miranda L. F., 1995, A&AS, 114, 439Fich M., 1993, ApJS, 86, 475Fich M., Blitz L., 1984, ApJ, 279, 125Fich M., Silkey M., 1991, ApJ, 366, 107Fich M., Treffers R. R., Dahl G. P., 1990, AJ, 99, 622Finkenzeller U., Mundt R., 1984, A&AS, 55, 109Frogel J. A., Persson S. E., Kleinmann D. E., 1972, Astrophys. Lett., 11,

95Garcıa-Segura G., MacLow M.-M., Langer N., 1996a, A&A, 305, 229Garcıa-Segura G., Langer N., MacLow M.-M., 1996b, A&A, 316, 133Hamann F., Persson S. E., 1992, ApJS, 82, 285Herbig G. H., Rao N. K., 1972, ApJ, 174, 401Higgs L. A., 1971, Catalog of radio observations of planetary nebulae and

related optical data, Pub. Ap. Branch N. R. C. 1, No.1Hillenbrand L. A., Strom S. E., Vrba F. J., Keene J., 1992, ApJ, 397, 613Hillenbrand L. A., Meyer M. R., Strom S. E., Skrutskie M. F., 1995, AJ, 109,

280Howarth I. D., Murray J., 1990, SERC Starlink User Note No. 50Hutsemekers D., Van Drom E., 1991, A&A, 251, 620Israel F. P., 1976, A&A, 52, 175Jaschek C., Jaschek M., 1987, The classification of stars. Cambridge Univ.

Press, CambridgeKaler J. B., Lutz J. H., 1985, PASP, 97, 700Maciel W. J., Pottasch S. R., 1980, A&A, 88, 1Merrill P. W., Burwell C. G., 1933, ApJ, 78, 87Mitra P. M., Dufour R. J., 1990, MNRAS, 242, 98Neckel T., Klare G., 1980, A&AS, 42, 251Nota A., Leitherer C., Clampin M., Greenfield P., Golimowsky D. A., 1992,

ApJ, 398, 621Panagia N., 1973, AJ, 78, 929Perek L., Kohoutek L., 1967, Catalogue of Galactic Planetary Nebulae.

Academy of Sciences, PraguePezzuto S., Strafella F., Lorenzetti D., 1997, ApJ, 485, 290Sabbadin F., Hamzaoglu E., 1981, A&A, 94, 25Schaller G., Schaerer D., Meynet G., Maeder A., 1992, A&AS, 96, 269Schmidt-Kaler Th., 1982, Landolt-Borstein, N.S, Vol. 2b. Springer, Berlin,

p. 21Sharpless S., 1959, ApJS, 4, 257Skinner S. L., Brown A., Stewart R. T., 1993, ApJS, 87, 217Smith L. J., Pettini M., Dyson J. E., Hartquist T. W., 1988, MNRAS,

234, 625Smith L. J., Crowther P. A., Prinja R. K., 1994, A&A, 281, 833Smith L. J., Stroud M. P., Esteban C., Vılchez J. M., 1997, MNRAS,

290, 265Stahl O., Smolinski J., Wolf B., Zickgraf F.-J., 1989, in Davidson K., Moffat

A. F. J., Lamers H. J. G. L. M., eds, Physics of Luminous Blue Variables.Kluwer, Dordrecht p. 285

S 266: a Galactic ring nebula? 191

q 1998 RAS, MNRAS 298, 185–192

Swings J. P., Andrillat Y., 1981, A&A, L3Testi L., Palla F., Prusti T., Natta A., Maltagliati S., 1997, A&A, 320, 159The P. S., de Winter D., Perez M. R., 1994, A&AS, 104, 315Underhill A., Doazan V., 1982, B Stars with and without Emission Lines.

NASA SP-456van den Ancker M. E., The P. S., Tjin A Djie H. R. E., Catala C., de Winter

D., Blondel P. F. C., Waters L. B. F. M., 1997, A&A, 324, L33Vılchez J. M., Esteban C., 1996, MNRAS, 280, 720Weis K., Chu Y.-H., Duschly W. J., Bomans D. J., 1997, A&A, 325,

1157

Zickgraf F.-J., 1992, in Drissen L., Leitherer C., Nota A., eds, ASP Conf. Ser.22, Nonisotropic and Variable Outflows from Stars. Astron. Soc. Pac.,San Francisco p. 75

Zickgraf F.-J., Wolf B., Stahl O., Leitherer C., Klare G., 1985, A&A, 143,421

Zickgraf F.-J., Wolf B., Stahl O., Leitherer C., Appenzeller I., 1986, A&A,163, 119

Zijlstra A., Pottasch S. R., Bignell C., 1990, A&AS, 82, 273

This paper has been typeset from a TEX=LATEX file prepared by the author.

192 C. Esteban and M. Fernandez

q 1998 RAS, MNRAS 298, 185–192