Planetary and lunar ephemerides continue to improve in accuracy as they continue to be adjusted to newer and more accurate observational data. An additional improvement will be that of the orientation of the ephemerides; in the future, the ephemerides produced at JPL will be based upon the reference frame of the radio source catalogues. Recent planetary observations have been made directly with respect to the radio reference frame, and these observations have shown a satisfying degree of absolute accuracy and internal consistency; they enable the automatic orientation of the ephemerides onto the radio reference system during the ephemeris adjustment process.

The new extragalactic reference frame will be based primarily on the positions of a globally selected number of about 500 compact radio sources, which almost all display optical counterparts, mainly quasars or BL Lac objects. Precise radio positions of these objects on the mas level have been determined by VLBI techniques and will be monitored and updated on a regular basis. In parallel an extensive optical observing program is underway to determine precise optical positions of these objects. The present optical reference frame is based on the FK5 fundamental catalog, containing 4652 brighter stars. The transfer of the VLBI radio frame to the optical domain will be achieved now by the Hipparcos stellar net, containing about 120,000 stars with comparable accuracy. The paper will address details and consequences of the implemenation of the new system for astrometry and general astronomical applications.

High-accuracy astrometric VLBI observations of 7 radio stars are presented with applications to the connection between the Hipparcos and VLBI extragalactic reference frames, to the identification of the radio emitting region in the ternary system Algol and the detectabity of Jupiter-size planet orbiting the radio star σ2 CrB.

In conformity with the IAU resolutions on reference frames adopted in 1991, the Hipparcos catalogue will represent, in the visible spectrum, the celestial reference system defined by fixed positions of extra-galactic radio-sources. This will be realized by the strongest possible link between the IERS celestial reference frame with positions and/or proper motions of the largest possible number of Hipparcos stars determined also with respect to extragalactic objects. The data which will be used must be available before April 1995. It will include the following: positions and proper motions of radio stars observed by VLBI, VLA and MERLIN; photographic positions in fields including quasars; proper motions with respect to galaxies of the Lick, Yale, and Kiev programs, proper motions derived from pairs of photographic plates taken at large time intervals; and possibly data acquired by Hubble Space Telescope and from Earth's rotation data. The organization of the tasks within the working group is briefly described. The final accuracy of the link is expected to be of the order of, or better than, half a milliarcsecond.

Uncertainties on stellar ages due to the physical description of the stellar material entering the models and to measurements of global parameters and chemical composition are estimated in the case of A/F stars.

With combined efforts on atmosphere modeling (to improve global parameters), on asteroseismology (to improve the physical description of the stellar interior) and on distance determination at the level of the sub-milliarcsecond (to improve luminosity), an accuracy on age determination of about 10% is foreseeable.

The impact of high-precision astrometry on modelling of stellar spectra and determination of fundamental stellar parameters for single stars is discussed. It is found to be considerable, but improvements in the understanding of stellar atmospheres, and in particular of the physical background of luminosity (i.e., surface-gravity) criteria will be needed in order to match the possibilities of stellar parallaxes with accuracies on the order of a few percents or better.

Although stellar interferometers are capable of measuring the angular diameters of stars, with longer baselines they may also be used to measure brightness variations across a star's surface and to provide constraints on models of stellar envelopes. In this paper we will look at the interpretation of visibility data and some of the more exciting prospects within the reach of current interferometers.

Precise trigonometric parallax measurements orders of magnitude more accurate than a milliarcsecond will contribute greatly to our understanding of peculiar, low luminosity stellar objects of several types. First, the volume of space out to which luminosities may be determined to the accuracy of the best, very-nearby stars will be increased greatly. For the relatively rare field Population II stars, this will lead to the first accurate empirical calibrations of the main sequence at the low mass end, for comparison with globular clusters of various metallicities. Parallaxes at 1 kpc or farther will be adequate to help in the discovery or confirmation of the rare carbon dwarfs - main sequence stars with carbon-rich atmospheres. For cool white dwarfs, luminosities accurate to a few per cent or better will identify unresolved binaries, and objects of unusually high and low mass. For our most numerous solar neighbors, the M dwarfs and especially those near the stellar mass limit, accurate luminosities can help in the determinations of the chemical composition and age distributions.

We present the technique of long-term, high-precision timing of millisecond pulsars as applied to precision astrometry. We provide a tutorial on pulsars and pulsar timing, as well as up-to-date results of long-term timing observations of two millisecond pulsars, PSRs B1855+09 and B1937+21. We consider the feasibility of tying the extragalactic and optical reference frames to that defined by solar system objects, and we conclude that precision astrometry from millisecond pulsar timing has a bright future.

A review is given of the use of high-accuracy astrometry on research on white dwarfs and the hot subdwarfs and central stars of planetary nebulae (CSPN). Predictions are made about the expected impact of HIPPARCOS, and the possible impacts of GAIA and ROEMER. Discovery of large numbers of new white dwarfs is expected, and, for the more distant hot subdwarfs and CSPN, important refinements of our current understanding of these objects. For white dwarfs independent values of mass and radius may be accurate enough to allow new understanding of the internal composition.

The primordial helium abundance YP is important for cosmology and the ratio δY/δZ constrains models of stellar evolution. While the most accurate estimates of both quantities now come from emission lines in HII regions, significant information comes from effects of helium content on stellar structure including in particular the location of the main sequence as a function of metallicity and age. HIPPARCOS parallaxes with 1 or 2 mas accuracy will naturally lead to great advances in this type of study for stars with metallicities down to about 0.1 solar, but sub-mas accuracy will be needed in order to extend it to stars of still lower metallicity.

The state of the art in accurate mass determination for binary stars is reviewed, and the angular sizes and their errors are computed for a typical system from the existing high-precision sample. It appears that sub-μas (microarcsecond) absolute astrometry will be needed in order to improve the accuracy substantially by astrometry alone. The types of system, and the kinds of data, where precision astrometry appears most promising are outlined. Finally, astrophysical applications of such accurate stellar masses, and the auxiliary data required in them, are briefly reviewed.

Long-baseline optical interferometry has made it possible to measure the visual orbits of binary stars with major axes as small as 5 mas and errors of ≲ 100 μas. Interferometers now nearing completion will extend these values to ≳ 500 μas and σa ∼ 10 μas. Observations of double-lined spectroscopic binaries with current interferometers have already yielded some mass estimates with precisions rivaling those from fitting the light curves of eclipsing double-lined systems. Luminosity estimates based on combined visual interferometric observations and velocity curves are often more precise than those from more indirect methods based on estimates of Teff. New interferometers now coming into operation will make it possible to measure fundamental parameters in dozens to hundreds of binary systems.

The HST Astrometry Science Team is using the Fine Guidance Sensors (FGS) in the Transfer Function (TF) Scan mode to search for binaries among the faint members of the Hyades cluster. To date (March 1994), nine binaries have been discovered among 24 stars examined. The closest pair (total V=13.5) has a separation of 0.051 arcsec; the faintest (sep=0.287 arcsec) has magnitudes V=15.0 and 16.5; neither object posed a challenge to the capabilities of FGS. For another pair, two observations 152 days apart show a 13 deg change in position angle, indicating rapid orbital motion. One decade should suffice to define the orbit with angular dimensions of sub-millisecond of arc accuracy.

Clearly, this work will soon permit mass determinations for low-luminosity members of the Hyades cluster. Moreover, information on the frequency of binaries will provide insight into the role of duplicity in star formation and in the dynamic evolution of the cluster. To be truly useful, a census of binaries in the Hyades (and other clusters) must ultimately reach cluster members fainter than those currently under investigation, requiring astrometry with sub-millisecond of arc accuracy at near-infrared wavelengths.

All periodic variables with stable periods such as pulsating stars, eclipsing binaries and pulsars offer the possibility of detecting unseen companion(s) by means of the light-time effect. We discuss the limitations of the method (the visibility of the effect) for different types of periodic variables. Special attention is paid to the ranges of mass ratios and orbital periods in which unseen companions can be found. We also indicate several systems with light-time effect in which hypothetical companions can be detected by speckle interferometry or precise astrometric observations. In these cases, the detection of the companions may lead to the determination of the components' masses.

A galactic counterpart of the Sloan Digital Sky Survey, consisting of two dedicated 2–3-m class telescopes (one in each hemisphere), could provide precisely those stellar data necessary to complement the information furnished by a future astrometric space mission: in particular, an automated acquisition and analysis system could provide spectral types, metallicities, and radial velocities of several tens of millions of stars down to about 15–16 mag. Such a dedicated spectroscopic stellar survey would provide considerably more information about the details of the stellar distribution within our Galaxy than is known at present. It is suggested that a galactic version of the Sloan Digital Sky Survey, utilising the GSC-II as an observing list, and capitalising on recent developments in multi-fibre spectroscopic survey capabilities, would be a timely initiative.

An accumulation of high precision astrometric data in conjunction with high-precision monitoring of the Earth's orientation, motivates “Galactic Astronomy”. As regards local kinematics, all of the three components of both the vorticity and the shear of stars can be completely determined, in addition to the velocity ellipsoid. We can now be released from the constraint of the “axisymmetric” galaxy. The determination of the proper motion of the LMC will be crucial to understanding the global structure and dynamics of the Galaxy with the dark halo and MACHO's motions.

Some of the possibilities created by sub–milliarcsecond astrometry in the study of both nearby and distant star clusters are presented.

We discuss the inclusion of a warp structure (as discussed by Miyamoto et al. 1993) to a realistic galactic model. We investigate how the proper motions are effected by changes in the inclination, phase angle and spin velocity of the warp. It is shown that accurate proper motions of disk stars do reflect the variation of warp parameters. A simulation of OB stars with an apparent magnitude limit of 7.0, mimicing the early type content of the HIPPARCOS catalogue, indicates that variations of phase are probably not observable while detection of velocity variation is within the limit of the current precision.