The subject of this meeting is the confrontation between precise observational and theoretical determinations of fundamental stellar properties. Its goal is to better define the limits of our present understanding of the structure and evolution of stars, and of our own and other galaxies. That goal is not approached by keeping to the “safe” side of the border: The areas where significant disagreement is found between the best observations and the best theories also show the directions where progress is to be made.

After a brief presentation of the Hipparcos mission, and an overview of the astrometric and photometric results obtained from the main mission and from Tycho, more details are given on the parallaxes. Absolute parallaxes have been obtained for 117 955 entries of the Hipparcos Catalogue, out of a total of 118 218, with a median precision of 0.97 mas for stars brighter than 9. This precision varies with apparent magnitude and ecliptic latitude. The estimated systematic error is smaller than 0.1 mas. The distances of more than 20 000 stars are determined to better than 10 %. Some more statistics are presented.

A few applications of this fantastic amount of new and accurate data are presented, in the fields of absolute magnitude calibrations, stellar physics, distance scale determination.

Several groups, using 1-m to 2-m telescopes on ‘visitor’ arrangements, have employed CCDs to measure stellar trigonometric parallaxes with precisions in the range ± 2–5 mas. More intensive observing efforts now routinely achieve sub-mas results with the best obtained to date in the ± 0.3–0.5 mas range. Selective dimming of bright target stars using neutral density spots now permit CCD parallax measures of stars as bright as R ~ 4.

Galactic and extragalactic distances to Cepheid variable stars according to a modern realization of the geometric Baade-Wesselink method are presented. Reliable results as accurate as few percent are currently achievable, allowing the cosmic distance scale calibration to be significantly improved with respect to previous calibrations.

Stellar angular diameter measurements have been made with a range of interferometric techniques including speckle, aperture masking and long baseline optical/infrared interferometry. The current status of these measurements are summarised in terms of the range of spectral types and luminosity classes measured, the accuracies achieved, the wavelengths used for observations, and the reliability of the results. A number of major long-baseline interferometers are coming on-line, or are under development, and their potential is assessed in terms of wavelength cover, accuracy, angular resolution, and the range of spectral type and luminosity class cover.

Interferometry offers an improvement in the accuracy with which astrometric measurements can be made. Using this technique, radio astronomers together with geodeticists have established a global inertial reference frame that is accurate to 0.1 milliarcseconds. At optical wavelengths, interferometry was first developed by Michelson at the turn of the twentieth century, but due to the complexities of precise beam combination at high speeds, it has lagged in its development. Now, with the availability of lasers, detectors and computers that allow path length compensation on millisecond time scales and distance determination between light collectors with a precision of 0.01 μm, interferometry at optical wavelengths will achieve the results in astrometry comparable to those at radio wavelengths.

Offering an angular resolution which has remained unattained by any other technique for decades, lunar occultations have traditionally been the most productive method for the measurement of stellar angular diameters. Unlike interferometric methods, which are limited in resolution by the size of the aperture or of the baseline between apertures, in a lunar occultation the key to high angular resolution is the phenomenon of diffraction by a straight edge, that occurs at the Moon's limb in a turbulence-free environment. For the reader not familiar with the physics and technical aspects of the lunar occultation (LO) technique, it is sufficient here to show in Fig. 1 some practical examples of occultation lightcurves for sources with different angular diameters. It can be noted that the contrast of the fringes is maximum for a point-like source; it then decreases with the angular diameter, and eventually reaches the regime of a monotonic drop in the signal-as predicted by simple geometrical optics- when the angular extent of the source is large. In practice, the LO method is well suited to measure angular diameters in the range 1 to 50 milliarcseconds (mas). There is no real limitation concerning the wavelength of observation, although at present the near-IR is the region of choice for several different reasons (Richichi 1994).

Observing a stellar radius basically means observing a center-to-limb intensity variation. The significance and properties of center-to-limb variations, common approximations, the correlation with optical-depth radii in extended-photophere stars, and direct measurements of angular (interferometry, lunar occultation) and absolute diameters (binary eclipses) are discussed. Spectrophotometric and doppler techniques of diameter determination are also briefly outlined.

We present a critical review of the available visible and near infrared flux calibrations. In the visible, the accuracy and the good consistency of three independent determinations of Vega monochromatic flux allow one to recommend with confidence f5556-Vega = 3.46 10−11Wm−2nm−1 within 0.7%. In the near infrared, the possible flux excess of Vega, as compared to that derived from the atmosphere models fitting the visible, does not allow such a good accuracy. The agreement between the calibrations, either from a comparison of Vega to blackbodies or from solar analog stars, would question the calibrations relying on models. More work is necessary to conclude with confidence and then to reduce the uncertainty on the near infrared calibrations.

The practical difficulties in performing laboratory calibrations of instruments at ultraviolet wavelengths are considerably greater than at visible wavelengths, and the concommitant uncertainties are greater as well. In recent years theoretical models of white dwarf atmospheres have been adopted as UV flux standards, with impressive results. In this review, I will discuss the methodology of laboratory flux calibrations in the UV, the internal consistency and potential shortcomings of calibrations based on white dwarf model atmospheres, recent laboratory results, and future prospects.

The total energy emitted by a star is one of its basic parameters. It is also one of the most difficult to determine, requiring space-based as well as ground-based measurements of the absolute flux from stars. We review the current status of these observations in the ultraviolet, visible, and near infrared, and estimate the uncertainties of the absolute flux measurements in these spectral regions. We compare recent determinations of the bolometric luminosity of early-type stars with the first (mostly) empirical results of 20 years ago. We close with a few comments on possible improvements of these measurements in the future.

In this paper, arguments are developed for treating (R–I)C as the most important colour to be derived for the Sun. The solar value of (R - I)C is then found to be 0.335 ± 0.002 mag. This result updates a counterpart given by Taylor in 1992.

We present flux-calibrated, synthetic spectra for the calibration stars of the Short Wavelength Spectrometer of the Infrared Space Observatory ISO-SWS (cf. Kessler et al. 1996). ISO-SWS covers a wavelength range of 2.4 to 45 μm (cf. de Graauw et al. 1996), and although in the NIR the flux calibration of Vega can be used, at longer wavelengths the flux calibration has to be extrapolated by other means, because (i) Vega has at λ ≥ 25 μm an IR excess due to circumstellar dust and (ii) there is an, as yet, unresolved controversy about Vega's flux at 10 μm (cf. Rieke et al. 1985). The majority of the standard stars of SWS are of MK class G and K III, as cool giants are amongst the brightest objects in the IR, and model atmospheres for cool giants are available and well-studied (cf. Jørgensen and Gustafsson 1994).

We present semi-empirical models based on the “minimum-temperature-concept” in order to provide a description of the UV spectra of solar type stars consistent with the observational data.

Recent progress in observing and data reduction methods for precise mass and mass-luminosity determinations in binary systems are briefly reviewed. The foundations appear to have been laid for a new burst of accurate data. Detailed model simulations of the individual systems are the best way to use these data to critically test the theoretical models and advance our understanding of the evolution of single and binary stars.