This paper is a brief overview of the past performance and future promise of interferometry as applied to the study of binary stars. For the sake of brevity, the important results from infrared techniques in recent years are regretfully omitted here. It is probably unnecessary to remind the reader that the analysis of binary star orbital motions is the only direct means for the determination of stellar masses. Historically, stellar masses have resulted primarily from orbits that are sufficiently wide in angular separation to permit the astrometric determination of the motions of the individual components about the center of mass as well as the parallax of the system or from short-period spectroscopic/eclipsing binaries. The advent of long baseline optical interferometers holds the promise of a wealth of new stellar mass data through the direct resolution of spectroscopic binaries.

A method to determine the masses of white dwarfs in magnetic cataclysmic variables from X-ray observations is presented. The method is appropriate for both eclipsing and non-eclipsing systems, for it is insensitive to the orbital inclination of the systems.

The surface gravity of a star (log g) is a fundamental parameter in models of stellar atmospheres. Given suitable spectra, log g can be determined from such models with an accuracy of 0.1dex, at best. Detached eclipsing binary stars can provide values of log g an order of magnitude more accurate than this, though for a more limited range of stars. Naturally, less accurate surface gravities can be obtained for a wider range of eclipsing binary stars.

These facts are well known, so in this short review I will outline the types of stars to which the two methods have be usefully applied and might be applied in the near future. This naturally leads to the question of where the two ranges overlap and the comparison of results from the two methods. Techniques for allowing this comparison to made directly will be described. Surface gravities derived from winds in hot stars and (indirectly) from gravitational redshifts in white dwarf stars will also be covered briefly.

The empirical temperatures scale for cool stars is generally well established. Temperatures are now known with reasonable precision for stars covering the range of spectral types from A to M. In the historical paper by Code, Davis, Bless and Hanbury Brown (Code et al. 1976), six stars between 10000K and 6500K had radii measured by the intensity interferometer and these six, together with the sun formed the basis of the empirical temperature calibration at the time. Since then, many temperatures have been derived for A-K stars (Blackwell & Lynas-Gray 1994; Alonso et al. 1996a) using the Infra-Red Flux Method (see Megessier 1994,5 and this volume), while lunar occultations (Ridgway et al. 1980) and more recently Michelson interferometry (Di Benedetto & Rabbia 1987; Dyck et al. 1996), have been used to measure the radii of K and M giants. It is a tribute to Hanbury Brown's Intensity Interferometer that temperature scales based on its measurements are essentially unchanged by the new data.

We review the effective temperature scale of hot (≤ 10 kK) stars, including results from direct, continuum and ionization equilibrium techniques. We discuss the impact of recent developments in theoretical model atmospheres for OB subdwarfs, dwarfs and supergiants, white dwarfs and Wolf Rayet stars and present a revised Teff scale for OB stars. Direct techniques coupled with Kurucz model atmospheres allow Strömgren photometry to be used as a sensitive Teff indicator for normal stars with Teff ≥ 25 kK. Reliable Teff determinations for hotter, low surface gravity and H-deficient stars require sophisticated ionization equilibrium techniques, generally considering non-LTE and line-blanketing effects.

The current situation for the measurement of effective temperatures from interferometrically determined angular diameters and spectrophometry in the UV, visible and IR is considered. Accuracies and reliabilities of the resulting temperatures are assessed and coverage of stellar types is discussed.

The principle of the method and its advantages are recalled, as well as the works giving such Teff. Although the method is almost model independent, the monochromatic red fluxes play an important part. The resulting Teff depends on the blanketing in the models. The stellar metallicity and gravity have to be known to avoid uncertainties of about 1%. In principle an accuracy around 1% on Teff could be obtained, but the uncertainty on the absolute red flux calibration, due to the unsolved problem of the Vega red flux, prevents this from being achieved.

This paper discusses some tests of synthetic spectra used in the calculation of synthetic colours, as well as some of the problems which arise in these calculations. Synthetic colours are tested in two ways. Firstly, observed angular diameters of stars are compared with ones predicted from stellar temperatures derived from synthetic colours and the infrared flux method. The average difference is less than 2%. Secondly, a very metal-poor isochrone is transformed to the Mv, B–V and Mv, V–I colour planes and compared with observations of stars in the globular cluster M92. The agreement is good in both colour–magnitude diagrams.

Among the mid-Teff (5000–7000 K) range stars, those on or near the main sequence are particularly important, because they provide much information on Galactic evolution and nucleosynthesis of the elements. This is due to the fact that these stars represent an age range from 0 to 15 Gyr, and a metallicity range from [Fe/H] ⋍ −4.0 to +0.3. Furthermore, it is likely that the present composition of elements in their atmospheres is similar to the composition of the interstellar gas from which they were formed, and that the present kinematics of the stars give information about their birthplaces. Hence, by observing selected samples of nearby F and G main sequence and subgiant stars one has the possibility to study the chemical and dynamical history of our Galaxy. Due to space limitation the present review is limited to these F and G stars. Supergiants are not discussed, but it should be emphasized, that their abundances are important for studies of advanced stellar evolutionary phases and determinations of the chemical composition of nearby galaxies. e.g. the Magellanic Clouds (Hill et al. 1995).

A theory for the broadening of spectral lines by collisions with atomic hydrogen developed by Anstee and O'Mara [1] is described and when applied to strong lines of elements in the sun leads to abundances which are independent of non-thermal motions in the photosphere and in excellent agreement with meteoritic values. In the case of the solar abundance of iron the results lead to a resolution of recent conflict concerning possible differences between the abundance of iron in the sun and meteorites. Excellent results obtained with the Mg b-lines permit these lines to be used in abundance determinations in other stars and also makes them particularly useful in determining the temperature structure in cool stars.

The halo stars are very old stars formed at the beginning of the evolution of the Galaxy. Their main characteristic is that their atmosphere is very metal-poor since at the time of their birth, the matter of the Galaxy had been enriched by only a few number of supernovae. The first very metal deficient stars have been studied in the sixties: Baschek (1959), Aller and Greenstein(1960) Wallerstein et al. (1963). The stellar models used in these first studies were rather crude: the atmosphere of the stars was represented by only one layer with a mean temperature and a mean electronic pressure, the continuum opacity was supposed to be due to H− and included in addition only the contribution by Rayleigh scattering. Later rescaled solar models were used and finally, at the end of the seventies, grids of theoretical models more and more sophisticated became available.

The galactic globular clusters presumably formed rapidly as high density concentrations at the centres of extensive star forming regions and the natural expectation is that they will be chemically homogeneous. In general, this is confirmed by observation — most globular clusters are extremely homogeneous with regard to most elements. (There are two exceptions: ω Cen and M22 both have intrinsic internal abundance ranges. While of considerable interest, e.g. Norris et al. 1996, this type of “abundance anomaly” will not be discussed here). But we have known for more than twenty years that the surface abundances, which are what we observe, of elements such as C, N and O can vary substantially from red giant to red giant within an individual globular cluster. Indeed it has become clear that “abundance anomalies” of this type are common in the galactic globular cluster population. Briefly, the observed anomalies can be summarized as: (i) the “anomalous” stars are depleted in C and enhanced in N. Depletions of O also often accompany the depletions in C. (ii) The C, N and O variations are usually accompanied by enhancements of Na and Al and when the O depletion and the Al enhancement are both large, Mg is depleted. No other elements, including r- and s-process indicators, vary.

Globular clusters in the Galactic bulge form a flattened system, extending from the Galactic center to about 4.5 kpc from the Sun (Barbuy et al. 1997). A study of abundance ratios in these clusters is very important for a more complete understanding of the bulge formation. In this work we present a spectroscopic analysis of individual stars in NGC 6553. This cluster is a key one because it is located at d⊙ ≍ 5.1 kpc, therefore relatively close to us, and at the same time it is representative of the Galactic bulge stellar population: (a) Ortolani et al. (1995) showed that NGC 6553 and NGC 6528 show very similar Colour-Magnitude Diagrams (CMDs), and NGC 6528 is located at d⊙ ≍ 7.83 kpc, very close to the Galactic center; (b) the stellar populations of the Baade Window is also very similar to that of NGC 6553 and NGC 6528 as Ortolani et al. (1995) have shown by comparing their luminosity functions.

The modeling of hot star atmospheres falls into two broad classes: those where the plane parallel approximation can be used, and those where the effects of spherical extension and stellar winds are important. In both cases non-LTE modeling is a necessity for reliable spectroscopic analyses.

While simple ions (e.g., H, He I, and He II) have been treated routinely in non-LTE for many years it is only recently that advances in computing power, computational techniques, and the availability of atomic data have made it feasible to perform non-LTE line blanketing calculations. Present models, with varying degrees of approximation and sophistication, are now capable of treating the effects of tens of thousands of lines. We review the latest efforts in incorporating non-LTE line blanketing, highlighting recent advances in the modeling of 0 stars, hot sub-dwarfs, Wolf-Rayet stars, novae, and supernovae.

I discuss errors in theory and in interpreting observations that are produced by the failure to consider resolution in space, time, and energy. I discuss convection in stellar model atmospheres and in stars. One dimensional convective models can never work well, but the errors in predicted diagnostics for temperature, gravity, and abundances can be calibrated. I discuss the variation of microturbulent velocity with depth, effective temperature, and gravity. These variations must be dealt with in computing models and grids and in any type of photometric calibration. I describe the state of the art in computing a model atmosphere and complete spectrum for stars with arbitrary abundances. I have made no significant progress on atomic or molecular line data for the last two years, but I will soon begin large scale production. I will continue to distribute the results on CD-ROMs, and I will make them available on the World Wide Web.

Over the past decade considerable effort, both observational and theoretical, has been directed towards a more accurate determination of the stellar lower main sequence and of the sub-stellar domain covered by Brown Dwarfs and Planets. Astronomers have been looking for brown dwarfs for more than a decade, either with standard astronomical technics or with microlensing experiments. A breakthrough in the search for brown dwarfs was very recently achieved with the discovery of the first cool brown dwarf GL 229B (Nakajima et al. 1995). At the same epoch, the search for planets blossomed with the discovery of a Jupiter - mass companion of the star 51 Pegasi (Mayor and Queloz 199,5). Now, the number of faint, cool stars and substellar objects is rising rapidly.

For the consistent and objective determination of stellar parameters and corresponding errors a constrained minimization approach is described which allows to take additional non-spectroscopic information into account and assures that the atmospheric structure equations are always fulfilled.

This review presents a number of aspects in our quest for an understanding of the detailed structure of cool star stellar atmospheres through the study of atmospheric motion. It includes the observational tools (e.g. direct imaging, detailed analyses of line profiles) used to determine the type and amount of motion, as well as the modelling techniques (e.g. simulated images, velocity and pressure fields) used to interpret and describe these motions.

Linkages will be made through the solar-stellar connection to provide insight into the structure of stars where detailed imaging is not possible, and to the determination of the fundamental parameters of stars, such as effective temperature, surface gravity and elemental abundances.

This brief review presents the current state of observations of stellar activity effects including the fluxes of chromospheric emission lines: CaII H+K, MgII h+k, SiII 1812 Å multiplet, CIV, as well as radio and X-ray fluxes versus B-V colours and luminosity classes, rotation periods, Rossby number and especially versus the mean magnetic flux density 〈fB〉. Results of stellar magnetic field measurements are presented.

We review the basic concepts, present state of theoretical models, and the future prospects for theory and observations of pulsating stellar atmospheres. Our emphasis is on radially pulsating cool stars, which dynamic atmospheres provide a general example for the differences with standard static model atmospheres.