I review our knowledge about effective temperatures and masses of massive OB stars paying special attention to results recently obtained using new atmospheric models including sphericity, mass-loss and line-blanketing. The new temperature scales for O dwarfs and supergiants are lower than previously accepted. No systematic mass discrepancy is found among O supergiants, although individual problematic cases still remain. It is shown that model atmospheres give answers consistent with binary systems. Good agreement is found between theoretical predictions and observations of the Wind Momentum-Luminosity relationship for Galactic O-type supergiants. Finally, I review the status of the WLR in M33.

High-quality, blue-violet digital spectroscopy has been assembled of most previously known (and one new) members of the O3 class that exhibit N iv emission and N v absorption lines, with the collaboration of most specialists in the field. A detailed comparative study has led to the introduction of new spectral types O2 and O3.5 to describe the range of classification criteria present. Primary reliance has been placed on the N iv/N iii selective emission-line ratio, because of various difficulties affecting the interpretation of weak He i absorption lines in some of these spectra, which are discussed. Some representative spectra defining the new types are shown. The majority of the O2 - O3 class members reside in the LMC. Their large ranges of luminosities and masses suggest a close evolutionary connection to the WN class, as do several spectroscopic and spatial relationships. The salient characteristics of O2-O3 spectra in the ultraviolet longward of 1200 Å have been established for some time by IUE and HST data. Recently, the rich 900-1200 Å range has been well covered for the first time by FUSE; in a collaboration with its hot-star team, a comprehensive atlas of OB spectra in the Magellanic Clouds has been prepared. Several O2 stars are included, one of which is newly discovered, and some examples of these data are also shown.

The use of photospheric abundances in OB-type main-sequence stars and A and B-type supergiants as probes of rotation and evolutionary status is reviewed. The abundances of CNO and boron can be compared quantitatively with stellar evolutionary calculations. In particular the abundance ratios of N/O and N/C can be derived in blue supergiants to determine if they are consistent with the stars having gone through a red supergiant phase and dredge-up of core material to the surface. The results from several different studies are reviewed and compared. For blue supergiant stars in the mass-range 5-80 M⊙ the situation appears consistent — there is no evidence for stars of any mass having undergone blue-loops in the HR diagram. The stellar samples show significant signs of having N enriched atmospheres, and the quantitative values are consistent with stellar evolutionary calculations which invoke turbulent diffusive mixing while massive stars are on the main-sequence. Results on the interesting blue supergiant Sher 25 are presented, and linked to Sk-69°202; the B3Ia progenitor of SN 1987A. A spectacular ejection nebula surrounds Sher 25, much like that which was ejected by Sk–69°202 during its final stages of evolution. Both of these were thought to have been formed during a mass-loss event when the stars were in the red supergiant phase. The CNO abundances derived in Sher 25 suggest the nebula was ejected during the blue supergiant phase, and that rotation can explain the N-enrichment found in the stellar photosphere.

The relative abundances of elements in galaxies can provide valuable information on the stellar and chemical evolution of a galaxy. While nebulae can provide abundances for a variety of light elements, stars are the only way to directly determine the abundances of iron-group and s-process and r-process elements in a galaxy. The new 8m and 10m class telescopes and high-efficiency spectrographs now make high-quality spectral observations of bright supergiants possible in dwarf galaxies in the Local Group. We have been concentrating on elemental abundances in the metal-poor dwarf irregular galaxies, NGC 6822, WLM, Sextants A, and GR 8. Comparing abundance ratios to those predicted from their star formation histories, determined from color-magnitude diagrams, and comparing those ratios between these galaxies can give us new insights into the evolution of these dwarf irregular galaxies. Iron-group abundances also allow us to examine the metallicities of the stars in these galaxies directly, which affects their inferred mass loss rates and predicted stellar evolution properties.

Current observations of the S Dor/LBVs and candidates and the implications for their important role in massive star evolution are reviewed. Recent observations of the cool hypergiants are altering our ideas about their evolutionary state, their atmospheres and winds, and the possible mechanisms for their asymmetric high mass loss episodes which may involve surface activity and magnetic fields. Recent results for IRC+10420, ρ Cas and VY CMa are highlighted. S Dor/LBVs in eruption, and the cool hypergiants in their high mass loss phases with their optically thick winds are not what their apparent spectra and temperatures imply; they are then ‘impostors’ on the H-R diagram. The importance of the very most massive stars, like η Carinae and the ‘supernovae impostors’ are also discussed.

Recent results for Galactic and Magellanic Cloud Wolf-Rayet stars are summarised based on line-blanketed, clumped model atmospheres together with UV, optical and IR spectroscopy. The trend towards earlier WN and WC spectral types with decreasing metallicity is explained via the sensitivity of classification diagnostics to abundance/wind density, such that WR mass-loss rates are metallicity dependent. Pre-supernovae masses for WC stars are determined, in reasonable agreement with CO-cores of recent Type Ic SN.

We discuss the connection between wind overloading and discrete absorption components in P-Cygni line profiles from O-type stars. Overloading can create horizontal plateaus in the radial wind speed that cause the extra absorption in the line profile. The upstream propagation speed of these velocity plateaus is analyzed. The second part of the paper deals with X-ray emission from O-type stars. X-ray line profiles observed with Chandra and XMM-Newton are often symmetric, contrary to what is expected for lines from a homogeneous wind. We discuss the influence on line symmetry of photon escape channels in a strongly clumped wind.

We report on a re-analysis of the Galactic O-type star sample presented by Puls et al. (1996) by means of non-LTE model atmospheres including line-blocking and line-blanketing. In particular, we concentrate on the question concerning the dependence of the wind-momentum luminosity relation (WLR) on luminosity class. Owing to the line-blanketing, the derived effective temperatures become significantly lower when compared to previous results, whereas the so-called ‘modified wind-momentum rates’ remain roughly at their former values. Therefore, we obtain a new WLR for O-type stars. By comparing these ‘observational’ results with new theoretical predictions and simulations, we conclude that the Hα forming region for objects with Hα in emission might be considerably clumped and thus a larger mass-loss rate than actually present is mimicked. We suggest that the previously found dependence of the WLR on luminosity class is an artefact.

Due to advances in computer power and numerical techniques non-LTE line-blanketing calculations for Wolf-Rayet (WR) stars are now routine. The incorporation of blanketing has led to significant improvements in spectral analyses, and to a systematic increase in the derived WR luminosities. To make further progress we need to understand the distribution, structure, and strength of inhomogeneities in the stellar wind, and in turn, how these influence diagnostics of WR stellar parameters, and radiative driving. Further, we need to understand the physical process that initiates mass loss in WR stars. Problems with existing wind calculations are examined, and the difficulty of observationally determining the shape of the velocity law around the sonic point is discussed. To determine the wind dynamics around the sonic point, it is essential to include ions with ionization potentials in excess of 300 eV. A recent study of the 07 Iaf+ star AV 83 is discussed. The analysis indicates the existence of a clumped wind, and a relatively slow wind acceleration with the velocity law characterized by β = 2. The importance of extreme Of stars, for understanding both WR and O-type star winds, is stressed. Spectra of such stars show numerous photospheric and wind features, allowing the entire wind to be probed observationally.

The formation of massive stars is one of the major unsolved problems in stellar astrophysics. However, only few if any of these are found as single stars, on average massive stars have more than one companion. Many of them are born in dense stellar clusters and several clusters have an excess of massive short-period spectroscopic binaries, with severe implication for binary-related stellar evolution including mergers, and also for the origin of massive runaway stars. The multiplicity of massive stars seems to increase with increasing primary mass and with increasing density of young star clusters. These observations suggest that massive binary and multiple systems originate mainly from dynamical gravitational interactions and accretion-induced protostellar collisions in dense clusters. If true, the binary properties of massive stars in less dense OB associations should be less extreme. This prediction should be tested by future observations. The paper reviews both the latest observations and theoretical ideas related to the origin of massive binaries. It concludes with a speculation on how the binary properties might change with metallicity (e.g., LMC/SMC).

Binary stars provide us with the means to measure stellar mass. Here I present several lists of known O-type stars with reliable mass estimates that are members of eclipsing, double-lined spectroscopic binaries. The masses of young, unevolved stars in binaries are suitable for testing the predictions of evolutionary codes, and there is good agreement between the observed and predicted masses (based upon temperature and luminosity) if the lower temperature scale from line-blanketed model atmospheres is adopted. A final table lists masses for systems in a wide variety of advanced evolutionary stages.

We present the results of calculations that illustrate the effects, on orbital timescales, produced on the external layer of a star that is in a binary system in which the stellar rotation period is not synchronized with the orbital period. The calculations show how, for a fixed set of parameters, the amplitudes of the oscillations depend on stellar radius and on β, the ratio between the stellar rotation angular velocity and the orbital angular velocity.

With the detection of gamma-ray burst (GRB) afterglows, the cosmological origin of GRBs has been firmly established. Recent observations suggest that (long-duration) GRBs are due to the collapse of a massive star forming a black hole. Besides theoretical arguments, observational evidence supporting this hypothesis comes from the coincidence of several GRBs with a supernova. Also, all accurately located GRBs are contained in the optical (restframe UV) extent of distant, blue galaxies. Some of these host galaxies show relatively high star-formation rates, which is expected when massive stars and GRBs are physically linked. Alternatively, GRBs can be produced by the merging of a binary neutron star system, such as the Hulse-Taylor binary pulsar. Very likely GRBs trace the massive-star populations in distant galaxies. With their enormous brightness, GRBs are powerful probes of the early universe, providing information on the properties of their host galaxies, the cosmic star-formation history, and potentially the first generations of massive stars.

Dust formation by Wolf-Rayet stars is a powerful diagnostic of exceptional activity in their winds. Observed behaviour ranges from stars which make dust continually (e.g., WR 104, WC9d+B0.5V) to those which make dust for short episodes at periodic intervals (e.g., WR 140, WC7pd+04-5). In both examples, dust formation is related to colliding WR and OB stellar winds in binary orbits. We have been using IR photometry primarily from the SAAO and ESO and spectroscopy with ISO-sws to investigate systems having intermediate properties (WR 98a, WC8-9vd+?) or more complex behaviour (WR 48a, WC8ed+?).