I would like to start with a quick overview of the O stars - their significance and role in the galaxy and in astrophysics - just to remind ourselves of why we are here and what we hope to talk about. In Table 1 I show a rough outline of the contribution of O stars to what happens in the galaxy as a whole. Because of their extreme luminosity, they contribute a large fraction of the radiation of the galaxy, while forming a very tiny group of objects and mass. Because of their short lifetime they are a population that has gone through 104 generations in the life of the galaxy. Their high mass loss rates may account for a large fraction of the new matter injected into the interstellar medium, and they probably power some significant fraction of the hard X-ray sources in the galaxy, by virtue of the fact that a companion can become a neutron star a) without disrupting the binary and b) while the companion is still a mass losing O star.

The aim of this paper is to report quite preliminary results obtained with a new multichannel detector put into operation at the Haute Provence Observatory. The observations have been made at the Cassegrain focus of the 1.93 m. The dispersion of the spectrograph was 17 Å. mm−1 at Hα and the slit was 300μ wide. The detector was a cooled SIT TV camera tube commercially available under the name of “Nocticon”. Details on the observing technique have been given elsewhere (Adrianzyk and al., 1978). We will only point out that all the wavelengths are recorded simultaneously : this recording technique is different from the scanning technique the result of which have been recently reviewed by Lacy (1977).

In recent years many observations of short time scale variability in a variety of spectral classes have been published. Some of these observations relate to phenomena which are reasonably well understood, as for example the Beta Canis Majoris stars, which are short period pulsators. In other cases, such as the Of and WR stars, the mechanism responsible for the variation is not fully understood. Models proposed to explain this variability include:- the binary hypothesis which ascribes the variations to fluctuations in gas streams in and around the members of a close binary system, and:- the pulsation hypothesis, which ascribes the variation in line strength to the presence of pulsational instability of massive carbon burning cores of evolved objects. The choice of the correct model is made difficult by the lack of an extended set of homogeneous observations. We must fulfil two conditions before making this choice. It is necessary to observe individual objects over extended periods of time at high dispersion to establish whether or not all WR's are binary systems. It is also required firmly to establish the temporal nature of the variations, since short term periodicities would tend to favor the core pulsation model.

Apart from the known P-Cygni profiles of Hβ, HeI 5876, 4471 in the spectrum of HD 153919, also HeII 3923 and a line at 3759 A which is tentatively identified as OIII show an emission component. These lines indicate that the Of star is losing mass at a high rate. A detailed analysis of the radial velocities obtained from 76 spectrograms indicates that the outflow of the wind is not spherically symmetric: at the side where the Of star faces its companion the velocity of the wind has increased.

The superluminous B supergiants are loosing mass and their progenitors are supposed to be O stars between 03 and 06 and hence it is justified to talk about B1Ia-O supergiants in this Symposium. Extensive high dispersion spectroscopic observations of four luminous B1 supergiants of the southern hemisphere have been carried out during 1972 and 1975 at ESO, La Silla. Some characteristic data of the program stars are summarized in Table 1.

The attempts to interpret the values of the observed radio and infrared fluxes of P Cygni via simple mass outflow models (constant outflow velocity or usual radiative acceleration outflow) lead to a strong discrepancy between the observed and the calculated values.

In our general program of research on the velocity field of emission nebulae, using the photographic Fabry-Pérot technique, we have included: NGC 6164-5, an H II region with striking symmetry, NGC 2359 a “ring nebula” and Ml-67. The exciting stars in all three are centrally located and are of spectral types 06f (showing P Cygni profiles), WN 5 and WN 8 respectively.

This high luminosity (5 × 106 L⊙) star since 1840 is losing mass at the rate of 7.5 × 10−2 M⊙ per year. The large mass loss could be the result of vibrational instabilities produced in the CNO hydrogen burning phase of a very massive (160 M⊙) star. The presence of high excitation lines in the ultraviolet spectrum of Eta Car suggests the idea of a hot zone excited by dissipation of the supersonic turbulent flow.

Up to now 4 WR stars were known in the SMC (Breysacher and Westerlund, 1978) and 76 in the LMC (Fehrenbach et al., 1976). Because no systematic search for WR stars in the SMC had ever been made, a survey was carried out with the ESO Objective Prism Astrograph which resulted in the identification of 4 new WR stars of the WN type, afterwards confirmed by slit spectroscopy. In the LMC, two fields were also observed and 13 new faint WR stars detected. The results presented here mainly concern the SMC WR stars.

On the basis of arguments presented by Roberts (1972) and of Shu et al. (1972), Minn and Greenberg (1973) argued that the velocity differences between newly formed hot stars and the surrounding interstellar medium are sufficiently different so that typical H II regions should consist of material which is continually being replaced by the ambient medium and which should therefore possess the velocity of the medium rather than that of the star. Obviously, the critical test of this hypothesis will be a comparison of nebular velocities with the velocities of the exciting stars.

Ultraviolet observations of mass-loss effects in O stars have, over the past decade, revealed a broad picture of a phenomenon whose extent was only partially evident from earlier ground-based observations. Ultraviolet resonance lines of a variety of ionization stages of several common elements provide a comprehensive probe of the low-density, extended winds. Three general types of information have been derived from ultraviolet spectroscopy of mass-loss profiles: (1) the nature of the stars which experience mass loss via radiatively-driven winds; (2) the physical conditions in the winds; and (3) variability in the outflow, which in turn may yield clues to the origins of the winds. Observations and results in each of these areas are reviewed, and some new results are included. A good correlation of mass loss rate and luminosity is indicated by the data, in agreement with theoretical predictions. Time variations in the P Cygni profiles may be quite common, with variability on times of hours or longer. Anticipated new observations, which should be possible with existing and planned instrumentation, are described.

The envelopes of early type stars can be studied by comparing observed and theoretical P Cygni profiles of UV resonance lines. The lines are formed by scattering in the expanding envelope and complete frequency redistribution is a reasonable assumption. The Sobolev approximation can be used for the prediction of the profiles, except within a Doppler shift of 2xvthermal from the line center. The profiles are sensitive to the optical depth τrad(v) ∝L ni(r) (dr/dv), but insensitive to the velocity law. We present a fairly simple method to calculate the effects of a photospheric profile or partly overlapping doublets.

The region of the O VI resonance doublet around 1030 Å has been observed in 9 stars with the high resolution (0.051 Å FWHM) detector on the Copernicus satellite spectrometer. Although the spectra are confused with interstellar Lyman β and H2 absorption, P Cygni profiles or shifted absorption lines indicating mass flow were found in ζ Pup (O4If), 15 Mon (O7V), HD 151804 (O81af), 10 Lac (09.4V), υ Ori (BOV) and τ Sco (BOV). However, no O VI could be seen in ρ Leo (BlIab) or γ Ara (BlIb). The O VI profiles display a considerable variety that is not correlated with spectral type. Zeta Pup has a broad emission and deep absorption ranging over all velocities to about −2800 km s−1 while HD 151804 has only narrow absorption from −1060 to −1660 km s−1, and 15 Mon has a narrower feature from −1700 to −2100 km s−1. Both 10 Lac and τ Sco have broad relatively shallow absorptions extending from positive velocities to −1380 and −1000 km s−1 respectively. In contrast, the absorptions are narrower with steeper wings in ζ Oph and υ Ori. The latter star also has emission, while in ζ Oph the absorption is split into two components from about −1550 to −1280 and −1280 to −980 km s−1.

The S-019 experiment on Skylab (cf. Henize et al. 1975; “Paper I”) recorded far UV spectra in about 160 4° x 5° fields, covering 10% of the sky, on 101 film with a 15 cm aperture objective-prism telescope. Several hundred early-type stars were observed in the vicinity of 1550 Å with a resolution between 3 and 4 Å, as well as thousands of stars at longer wavelengths and correspondingly lower resolution. An atlas of spectra for types 04 to B5 is illustrated in Paper I. That figure shows that the P Cygni profile is a characteristic of all supergiants earlier than B3 and main sequence stars earlier than 08.

Recent observations with the International Ultraviolet Explorer (IUE) satellite show that two very different types of hot stars have stellar winds: not only do the young, massive OB stars (the subjects of our discussion in this symposium, so far) undergo high-velocity mass-loss, but so also do hot evolved, solar mass stars, among them the central star of planetary nebulae. In this talk, I would like to show ultraviolet spectra of two central stars, the nuclei of NGC 6826 and Abell 78, and to describe how the characteristics of these spectra may be used to derive information concerning winds in these stars and in hot stars in general.

Angular diameters have been estimated for 18 O and 142 B stars using absolute intermediate-band photometry in the near infrared and they have been combined with integrated fluxes to yield effective temperatures. The effective temperatures of the O stars lie in the range 30000 K to about 47000 K. For a given subtype, the luminosity class I stars have lower effective temperatures than the main-sequence stars by about 1000 K. The absorption-line spectral types of the supergiants of types O and B reflect electron temperatures which are higher than can be maintained by the integrated flux which flows through the stellar atmosphere. Distances have been estimated for all the stars and linear diameters found. The average radius for an 08 to 09.5 supergiant is about 23.3 R⊙; the radii for luminosity class III and Class V O stars lie in the range 6.8 to 10.7⊙ R.