Red supergiant stars that are members of the nearby Sc galaxy NGC 300 have been isolated by spectrophotometry and infrared photometry from a blink comparator survey of red stars in the general field of this highly resolved stellar system. Our spectroscopic criterion is the strength of the CaII triplet near λ8500 which is more than twice as strong in red supergiants as in red dwarfs. Red supergiants and dwarf stars also occupy distinct regions in a two-color (J-H) vs (H-K) diagram. Both methods clearly segregate the supergiants which, with visual magnitudes close to 19, must be members of NGC 300. Of the 18 red stars we examined, only 6 are probable members of NGC 300. The other 12 are likely foreground stars belonging to the Galaxy. A principal finding is that, unless special care is taken, contamination by foreground stars can be an important factor when the red stars of an external galaxy are selected only on the basis of their colors in the visual and near infrared spectral region.

The brightest stars in galaxies have been often considered useful as distance indicators. However, one may expect that the absolute magnitude of the brightest stars in galaxies should depend on the IMF and on the present SFR, and that statistical effects may be important, although different in different photometric bands, due to the different numbers of objects effectively sampled.

From the existing literature data have been collected on the stellar rate of mass loss for 189 stars with known or derived values of the effective temperature Teff and luminosity L. It appears that Ṁ depends only on Teff and L for the O- through M-type stars brighter that about 3 × 103 L⊙. This is shown in Figure 1, where we have plotted for each star the value of - log(-Ṁ) with Ṁ expressed in solar masses per year. An interpolation formula has been derived for the Ṁ(Teff, L) dependence. If we define X = log Teff-4, and Y = log(L/L⊙-5, then

The spectral type - effective temperature calibration is a cornerstone of our understanding of massive stars and their environment. Any uncertainty in determining their effective temperature is directly reflected in an uncertainty in defining the bolometric luminosity and the number of ionizing photons. We suggest that previous calibrations of bolometric luminosity and the number of ionizing photons as a function of spectral type may be uncertain by at least a factor of 3 for the early 0-type stars (Table 1).

The presence in the far-UV spectra of early-type stars of spectral lines of superionized atoms is argument of controversial debate among astronomers. Presently there is agreement on the non-radiative origin of these ions but not on the proposed mechanisms for their production nor on the proposed locations in the stellar atmosphere where they are abundant. Cassinelli et al. (1978) suggest that the Auger mechanism is operative in a cool wind blowing above a narrow corona to produce these ions; Lucy and White (1980) introduce radiative instabilities growing into hot blobs distributed across the stellar wind; Doazan and Thomas (1982) make these ions to be formed in both pre- and post-coronal, high temperature regions at low and high velocity respectively.

This is a progress report on a study of the far-UV spectra of hot stars observed with the Voyager ultraviolet spectrometers (UVS); we discuss UV flux distribution and variability of a set of early type high luminosity stars.

The composite HR diagram for luminous stars in our galaxy by Humphreys and McElroy(1984) is the only source which provides enough data for the comparison with the results of massive stars theoretical models. The luminosity limit for completeness of Humphreys' and McElroy's catalog(l984) is about MBol=−8 mag, when restricted to 3 Kpc from the Sun. Inside these limits there are 378 luminous stars in galactic associations and clusters, whose distances are known with sufficient reliability. The correspondent observational HR diagram is reproduced in Fig. 1. Synthetic HR diagrams for massive stars were constructed with the aim of reproducing the observed features for galactic luminous stars. By means of a Montecarlo technique, stars were randomly distributed in the MBol−Log Teff plane, weighted on the Salpeter's initial mass function and with a constant star formation rate.

While there have been many surveys for luminous, blue galactic stars and their numbers can be considered somewhat complete, such is not the case for red supergiants (see e.g. Humphreys and McElroy 1984). One result of this incompleteness is that the ratios B/R and WR/R, often used as diagnostics for evolutionary models of massive stars and the variation of the ratios with galactocentric distance, are not well known for the Galaxy. In an attempt to improve the statistics, the first author began an objective-prism survey within 6 deg of the southern galactic plane using I-N plates. The dispersion is 3400 A/mm at the A-band, and the spectra cover the range 6800–8800 A; the deepest plates reach ir mag ∼13. The detection of possible M supergiants on such plates was first discussed by Nassau, et al. (1954) and depends on the presence of TiO at 7054 A and a spectrum sharply tapered to the blue. For supergiants, this shape results from integration of interstellar dust over a long path-length, but any sample of red stars with tapered spectra contains M giants in heavily-obscured regions and S stars; thus follow-up observations of the candidate stars are necessary.

Dynamical, vibrational, and thermal instabilities of massive blue stars are discussed as possible mechanisms for the observed brightness variations of such objects. Relaxation oscillations (on local thermal time scales) due to dynamical instabilities of the stellar wind flows appear to be the most likely mechanism, at least for the S Dor variables. Very massive main-sequence stars with M > 103 M⊙ should be violently vibrationally unstable and therefore should differ significantly from stable main-sequence stars of lower mass.

The luminous S Dor variables or Hubble-Sandage variables have gained considerable interest during the past few years for their exceptional (if compared to normal luminous hot stars) stellar wind properties. They seem to represent a short intermezzo in the evolution of the very massive stars prior to becoming WR stars. More recent results on S Dor variables (mainly of the LMC) are briefly summarized. Since most of the data have been or are going to be published elsewhere only an outline will be presented here.

The class of P Cygni Type (PCT) stars is defined and the brightest members in our galaxy, LMC, M31 and M33 are identified. They are located near the upper luminosity limit in the HRD in the range of 8500 ≤ Teff ≤ 27,000 K and −11.0 ≤ Mbol ≤ −7.8. We suggest that all PCT stars are S Dor variables, and that the reverse may also be true. The basic parameters of the PCT stars are derived and compared with those of normal supergiants: the effective gravity is a factor 3 to 10 lower, the mass loss rate is a factor 3 to 10 higher, the terminal velocity is about a factor 10 lower. This results in a wind density which is a factor 30 higher and thus produces the P Cygni lines in the visual spectrum. The history and the physical processes in the star P Cygni are discussed. The photometric variability and the shell ejections on a timescale of about a month are probably due to non-radial pulsations. The acceleration of the wind is due to radiation pressure by numerous (∼103) metal lines in the Balmer continuum. This can also explain the S Dor type of variability. The proposed mechanism will automatically lead to a P Cygni phase in the evolution of the most massive stars and to the existence of an upper limit in the HR diagram. The expected lifetime of the PCT phase is about 104 to 105 years. Some unresolved problems and recommendations for future research are discussed.

Three areas of current progress relevant to the theme of this Symposium will be discussed. (1) New spectroscopic observations of the 30 Doradus central cluster, obtained independently by the author and by J. Melnick, confirm the presence of numerous very early O-type members, including several of type O3. In combination with sophisticated new direct imagery of the luminous central object R136 by A. Walker and by G. Weigelt, these results have evident implications for understanding the ionization of the supergiant H II region, as well as for the interpretation of R136 itself and of the apparently similar regions seen in more distant galaxies. In particular, no evidence remains for a supermassive object in 30 Doradus, but its central cluster is revealed as a spectacular grouping of very massive hot stars. (2) A further member of the Ofpe/WN9 category in the LMC has been identified, bringing their number to seven, with no exact spectroscopic counterparts yet known in the Galaxy. One of these objects is currently in a state of outburst and has been interpreted by O. Stahl et al. as the hottest known Hubble-Sandage variable. (3) An extensive survey of IUE high-resolution data has revealed a strong correlation between the ultraviolet stellar wind features and the optical spectral classifications for the majority of normal O stars. These results are relevant to future studies with the High Resolution Spectrograph on the Hubble Space Telescope, which may observe restricted UV wavelength ranges in faint extragalactic OB stars lacking optical data of comparable quality.

Wolf-Rayet (W-R) stars are objects with strong emission lines in the optical region due to a substantial stellar wind. They are found among the extreme Population I stars which are the topic of this Symposium and also this talk, and as central stars of planetary nebulae which I will not discuss here. The mass loss rates, Ṁ, are typically several 10−5 solar masses per year. W-R stars come in two major subtypes, the WN, in which strong lines of helium and nitrogen are seen, and the WC, in which strong lines of carbon and oxygen are found, along with the helium ions. In a few WN stars hydrogen appears to be present but the spectra of W-R stars are notable for the absence of this element. Absorption lines are generally not found in W-R stars, except in some instances due to a binary companion, or a more distant companion, or as a P-Cygni profile associated with certain emission features. An absorption spectrum appears to be an intrinsic feature in a very few WN stars.

Bolometric corrections and absolute magnitudes for WR stars have been evaluated by various authors. A review of the results may be found in De Jager (1980). The calibrations, relating BC or Mv with the spectral subtype, however strongly differ from author to author. Moreover, even with the same observational method a large scatter is found for the same subtype, as was recognised by Conti et al. (1983). The only trend found is the increase of Mv with the subtype from WN3 to WN7, and for the WC subtypes. In this study BC and Mv are computed for a number of WR+O binaries.

Frequently the models proposed to explain some observational data are in conflict with the other observational data. Our purpose was to find out the empirical models of the envelopes of WN5 and WN6 stars which are not in conflict with the continuum runs in the IR spectral region and with the emission line intensities and profiles. In our recent study (Nugis, 1984) we proposed for Wolf-Rayet stars the spherically symmetric outflow model having an increasing velocity zone (with some coronal heating near the stellar surface) which is followed by the decreasing velocity zone extending up to the distance where the matter starts flowing with constant velocity. Let us call this model the AD-model (Acceleration-Deceleration).

A number of surveys for emission-line objects have been made in the Magellanic Clouds (see Westerlund, 1983). Hα-emission-line objects in the Small Magellanic Cloud (SMC) have been identified mainly by Henize (1956) and Lindsay (1961). Since this pioneering work took place, no other extensive survey for this kind of object - such as the one by Bohannan and Epps (1974) in the Large Cloud - has been carried out in the SMC. Consequently the detection of Hα-emission-line objects in the SMC is still rather incomplete.

The paths that massive stars follow in their evolution can potentially be determined by studying a homogeneous collection of luminous stars. In this investigation, mass-losing stars in a galaxy are studied through a sample of 59 H alpha emission-line stars in the Large Magellanic Cloud (LMC). The most luminous, most massive stars in the LMC, like their counterparts in the Galaxy, are losing mass at a rate that significantly alters their spectroscopic appearance and that affects their evolution.

The recent study by Caillault et al. (1985) has shown the emission-line star AS 431 (= WR 147) to be a strong X-ray source and moderately strong radio emitter. Combining optical, near-infrared and radio data, they deduced that its mass loss rate was ∼ 4 × 10−4 M⊙yr−1, an order of magnitude greater than is normal for WR stars. They also suggested that it would show up in the IRAS survey and that such data would help to elucidate whether its extreme reddening had a significant circumstellar component. To investigate this possibility and to study the general properties of the object, we have raided the IRAS Point Source Catalogue.