At the low temperatures of cool carbon star atmospheres polyatomic molecules form. Most abundant of these for the temperature range 2500–3000 are HCN and C2H2 (Fig. 1).

Early S stars occur between M and C stars in the colour magnitude diagrams of intermediate age globular clusters in the Magellanic Clouds. Most have −4.2 ≥ Mbol ≥ −4.8 and are probably brighter in younger or more metal rich clusters. The galactic globular cluster NGC 6723 contains two marginal S stars at Mbol ∼ −3.3. The rare CS stars have Mbol ∼ −6, with no faint examples.

Most of our present knowledge about central stars of planetary nebulae (CPN) is obtained by indirect methods using the emission line spectra of the surrounding nebula. These methods, which apply the beautiful recombination theory, can provide us with information about temperature, distance and radius of the CPN. However, we know that these methods are subject to severe problems, which then lead to strong discrepancies in the parameters of the CPN.

During the last years some progress has been made in the determination of temperatures of central stars of planetary nebulae (CPN). The main reasons are the successful deployment of space crafts which made the more temperature sensitive spectral region in the UV accessible, and the application of NLTE-spectral analyses (e.g. Mendez et al., 1981). Therefore, a re-examination of the H.R.-diagram of CPN's seems to be in order.

Up to date, 1518 true, possible, and probable planetary nebulae (PN) are known in our Galaxy; 463 nuclei (NPN) were observed. For about 150 stars, a spectral study has been done; only about 30 NPN have a well studied spectrum. (Acker, Gleizes et al, 1982 “Catalogue of the central stars of true and possible PN”) It must be remembered that this is a difficult study, as the PN are very distant objects (the nearest, Helix, is about 120 pc), and that their central stars are very faint (the 35 brightest have a magnitude of from 9 to 12), are generally of a particular spectral type, and have a spectrum in which nebular lines interfere with stellar features.

Among the faint blue halo stars the sdO's form a group whose evolutionary status is still not fully understood. Phenomenologically they appear to be related to the horizontal branch. The knowledge of stellar parameters and abundances may give a clue for understanding their evolutionary status. As the sdO's are hot (Teff ≥ 35000K), the assumption of LTE leads to substantial errors. Hence detailed NLTE analyses are inevitable.

A spectroscopic survey by Drilling (1983) of non-emission OB+ stars in the Case-Hamburg survey and its extension to b = ±30° for 1 = ±60° has turned up 12 hot, relatively bright O-type subdwarfs (V<13). Drilling's survey covers more than 13 times the area of the south galactic polar survey of Slettebak and Brundage (1971), but has revealed only the hottest and/or hydrogen-poor objects because only the weakest-lined 0-and B-type stars were observed. We have observed all of these stars in the large-aperture, low-resolution mode of the IUE satellite. A full description of their spectral appearance is given elsewhere (Schönberner and Drilling, 1984). The spectra were corrected for interstellar absorption using the Seaton (1979) reddening law to nullify the 2200 å interstellar feature, and in all cases EB-V < 0.25.

8 sdB stars have been analyzed for effective temperatures, gravity and helium content by means of LTE and NLTE model atmospheres (Heber et al., 1983). The analyses are based on the following observations:

1. 1.) B&C IDS spectrograms obtained with ESO's 3.6m and 1.5m telescopes (4250–4800å at 29 å/mm, or, 4000–5000å at 59 å/mm).

2. 2.) IUE spectra in low resolution (6å) through the large aperture.

3. 3.) Johnson and Strömgren colours, which are taken from the literature.

This is a summarizing discussion about detailed studies of stellar winds from hot, subluminous stars of different classes: two sdO's (Hamann et al., 1981), three extreme helium stars (Hamann et al.,1982) and one CPN (Hamann et al.,1983). All program stars have been analyzed previously in Kiel by non-LTE techniques with respect to effective temperature, gravity and abundances. For comparison, we have also included the Of-supergiant Pup and the B-main sequence star Sco, because their winds have been investigated in detail by the same methods as applied in this paper (Hamann, 1980, 1981).

The recent observation of white dwarfs in the open cluster NGC 2516 and the determination of their surface gravity and effective temperatures (Reimers and Koester, 1982) has enabled the establishment of the initial-final mass relation for low and intermediate mass stars which was published a few months ago and is presented here (Fig. 1 of Weidemann and Koester, 1983a). The most important conclusions drawn are:

1. 1. The limiting mass for white dwarf progenitors is 8–9 M⊙ rather than 5–6 M⊙, with supernova production beyond;

2. 2. The rather flat run of the initial-final mass relation in the main range of star production, 1–5 M⊙, explains the observed narrow mass distribution of white dwarfs and central stars of planetary nebulae around 0.6 M⊙;

3. 3. High mass white dwarfs exist, as shown in the rich, young cluster NGC 2516, but are extremely rare in general.

The sources of information and uncertainties in the intrinsic stellar parameters of luminosity, effective temperature, mass, composition and mass loss rates are discussed. These are used to compare the observed positions of massive stars in the Hertzsprung-Russell Diagram (HRD) with evolutionary tracks. The current status of this effort is briefly reviewed. A short summary of the kinematic properties of massive stars is made. A preliminary but fairly extensive discussion of the distributions and numbers of O-type and Wolf-Rayet stars in the galaxy and other members of the local group is then given.

We have undertaken a systematic survey of 115 0-B0 type spectra between 1200 and 1900å, by means of International Ultraviolet Explorer high-dispersion data rebinned to a constant wavelength resolution of 0.25å and uniformly normalized (Walborn and Panek 1983). The ultraviolet spectral features, both photospheric and stellar wind, are found to be smoothly correlated with the optical spectral types for the majority of normal stars. In particular, the C IV (λλ1548,1551) and N V (λλ1239,1243) resonance doublets show similar strong P Cygni profiles, which decline along the main sequence beginning at type 07 but remain strong in all higher luminosity spectra. In contrast, the stellar wind effect in the Si IV (λλ1394,1403) doublet displays a pronounced luminosity dependence, being absent from the main sequence spectra at all types and developing gradually through the intermediate luminosity classes into a full P Cygni profile in the supergiants. Relative to this consistent framework provided by the normal stars, a number of peculiar spectra can be succinctly described, including several of types OBN and OBC (Walborn 1976).

Eta Carinae has provided more incisive clues to the evolution of the most massive stars than any other individual object. Its luminosity is comparable to that of a very massive O3 star, but it appears to have evolved beyond the O3 stage, to the limit where its surface is rather unstable. We think that it has begun to lose mass semi-catastrophically, in outbursts which may recur at intervals of perhaps several hundred years. The suspected binary nature of Eta Carinae introduces some conceptual uncertainty but is most likely not crucial.

Non-LTE analyses of five very massive O-stars yield effective temperature, gravity and helium abundance of their photospheres. From these results together with photometric data an estimation of the stellar mass is possible. The comparison of the position in the log g — log Teff diagram relative to theoretical evolutionary tracks allows three compatibility-checks: (1) masses, (2) helium-abundance and (3) age of the stars. From these checks it is inferred, that (a) the stars are in an advanced main-sequence-phase, and (b) they have suffered from considerable mass-loss. Furthermore it seems to be likely that (c) mass-loss-rates do not scale with N > 200, (d) the age of the stars in the η Carinae region is less than 2 mio yrs., and (e) HD 93128 is in the background of Tr14.

With a semiempirical NLTE-model in spherical geometry we investigate the hydrogen abundance in Wolf-Rayet stars. We find that the flux contribution of the H-Balmer to the He-Pickering lines lies between the usually applied approximate formulae for the optically thin and optically thick case. This result can be understood with our finding that there are no optically thin helium and hydrogen lines. The influence of the temperature structure on the relative intensity of the hydrogen-helium blends to the helium lines is vanishingly small, even for peculiar temperature laws. The conclusion is that Wolf-Rayet stars are hydrogen deficient objects.

In a recent IR photometric survey of late-type WC and WN stars, it was discovered that not only most WC8-10 stars have circumstellar dust shells, but that two extreme late-type WN stars also have strong IR excesses from circumstellar dust. The latter shells appear to have significantly different density distributions. In this paper the possibility of an evolutionary sequence is suggested.

In this review I will primarily be discussing the observational data relevant to understanding the process of stellar evolution in galaxies of different types. This discussion will focus on the stellar content of the nearer galaxies; those galaxies in which the brightest individual stars are resolved and can be observed.

Massive stars (say M ≳ 9 M⊙) play a major role in Astrophysics. They are the main agents of nucleosynthesis and galactic enrichment and present many interesting properties, like high stellar winds, chemical surface enhancements, instabilities etc. which deserve further investigation. Furthermore, massive stars also are the powering sources of giant H II regions and have important radiative, mechanical and chemical interactions with the interstellar medium.

R136a is the luminous object in the core of the 30 Dor nebula in the LMC, embedded in the extremely young star cluster NGC 2070 with the most recent star forming event less 106 yr ago. R136a-1 (V = 11.08) lies in the center of this cluster. The companion or foreground star R136a-2 (ca. 1.2m fainter than R 136a-1) 0′.5 apart does not impair the dominating role of R136a-1. The light distribution is strongly peaked to the center of R136a and disagrees completely with Moffat and Seggewiss's (1983) description and their interpretation as a dense core of a cluster (Chu and Wolfire, 1983). This is strenghtened by the UV observations (Cassinelli et al. 1981, Feitzinger et al. 1983, Savage et al. 1983): the spatial spread function of the IUE satellite is dominated by an object of extremely small ultraviolet extent, less than that of a gaussian with 1′.5. Speckle observations (Meaburn, 1982; Weigelt, 1981), are discordant in the interpretation of the faint background, but concordant in revealing the existence of a small bright component, with an upper limit of the diameter < 0.02 pc. This is also found, with somewhat lower resolution, by Chu (1934).

To look for other supermassive stars in other galaxies means to answer the question: what is the spectroscopic and morphological signature of such objects? The interpretation of WR sources found in other supergiant HII complexes depends basically on our picture of 30 Dor and in particular of R136a.