The potential of the Milky Way is almost certainly not axisymmetric: the centre is believed to be dominated by a bar, and beyond the solar radius some non-axisymmetric feature of the potential appears to warp the disk. There are grounds for believing that the outer potential is mildly elliptical. Sub mas yr–1 proper motions of objects that lie near the plane several kiloparsecs from the Sun would play a crucial role in refining our understanding of these non-axisymmetries.

Improving the astrometric accuracy by one or two orders of magnitude over ground-based techniques will not only change our raw knowledge about the Galaxy, but it will also modify 1) the fundamental questions that can be addressed, and 2) the stellar dynamical concepts used so far. More detail in Galactic structure, such as the shape and flow in its putative bar, will be accessible. Also, with the instruments of the next generation the large scale dark matter distribution in the Galaxy, whether distributed in a spheroidal smooth halo or a massive outer disc made of cold clumpy gas, will be measurable. Techniques used for mapping the cosmic flow and mass distribution at Mpc scales and more might be applied to the solar neighbourhood to find the degree of clumpiness of the local matter distribution.

There are many fundamental aspects of Galactic structure and evolution which can be studied best or exclusively with high quality three dimensional kinematics. Amongst these we note as examples determination of the orientation of the stellar velocity ellipsoid, and the detection of structure in velocity-position phase space. The first of these is the primary limitation at present to reliable and accurate measurement of the Galactic gravitational potential. The second is a critical test of current standard models of Galactic formation and evolution.

Absolute proper motions of seven globular clusters have been determined with respect to extragalactic references and with accuracies of ∼ 1 mas/yr. Derived quantities and qualitative implications are described.

It is generally agreed that the Hubble parameter lies in the range 50 < Ho (km s−1 Mpc−1) < 100. Recent observations, which are discussed in the present paper, favor Ho ≳ 75 km s−1 Mpc−1. Hubble Space Telescope observations of Cepheids in Virgo cluster galaxies will probably reduce the uncertainty in Ho to ∼ 20%. It should be possible to lower this remaining uncertainty in the Hubble parameter by strengthening the calibration of the Cepheid period-luminosity relation, the MV(RR) versus [Fe/H] relation of RR Lyrae stars, and the luminosity calibration of the subdwarf main sequence, by using parallaxes obtained with Hipparcos.

It is essential for our understanding of the evolution of the Magellanic System, comprising the Large and the Small Magellanic Cloud, the Intercloud or Bridge region and the Magellanic Stream, to know its motions in the past. The Clouds have a common envelope of neutral hydrogen; this indicates that they have been bound to each others for a long time. The Magellanic System moves in the gravitational potential of our Galaxy; it is exposed to ram pressure through its movement in the galactic halo. Both effects ought to be noticeable in their present structure and kinematics. It is generally assumed, but not definitely proven, that the Clouds have been bound to our Galaxy for at least the last 7 Gyr. Most models assume that the Clouds lead the Magellanic Stream. The interaction between the Clouds has influenced their structure and kinematics severely. The effects should be possible to trace in the motions of their stellar and gaseous components as pronounced disturbances. Recent astrometric contributions in this field show a great promise for the future if still higher accuracy can be achieved.

The best current approximation to an inertial reference frame is provided by Very Long Baseline Interferometry (VLBI) observations of extragalactic radio sources with red shifts (z) up to 3.8. The stability of the resulting reference frame directly depends on the amount of any secular changes in the observed source positions.

Two types of potentially observable secular motions should be present in extragalactic source positions. Gravitational accelerations of the solar system will cause secular motions through aberration, amounting to, e.g., about 4 microarcsec (μas) year–1 due to the mass of the galaxy. Extragalactic mass concentrations will cause gravitational deflections in the apparent positions of more distant radio sources, and these will change with time as the mass concentrations evolve. This effect could easily cause secular motions of order 1 μas year–1 in some, or even most, radio sources with z ≥ 1.

The present astrometric VLBI data set contains about one million observations over a 15 year period, with current source proper motion formal errors being as small as 2.5 μas year–1. Proper motion estimates from these data reveal many sources with statistically significant proper motion estimates of order 30 μas year–1, about an order of magnitude larger than expected. Work continues to determine if the observed motions are due to systematic errors or reflect true secular changes in source positions. The results from a continued proper motion analysis of the complete astrometric VLBI data set will be presented.

Very Long Baseline Interferometry (VLBI) is at present the most powerful technique to construct the best approximation to an inertial reference frame. After more than a decade of VLBI observations several hundreds of extragalactic objects have positions known within ±0.0003″. Since 1988 the International Earth Rotation Service (IERS) elaborates a global extragalactic celestial reference frame (IERS Celestial Reference Frame, ICRF) that is tied to the international terrestrial reference frame through the high precision monitoring of the Earth's rotation. The direction of the ICRF axes relative to the IAU definitions are known within better than ±0.001″ for the polar axis and ±0.003″ for the origin of right ascensions. The FK5 axes are consistent with the ICRF ones within their uncertainties (0.050″-0.100″). The maintenance of this high accuracy extragalactic frame will be necessary for the long term programs, such as the future monitoring of the tie of the Hipparcos galactic frame as well as of the dynamical planetary frame (millisecond pulsars, lunar laser ranging).

Sub-milliarcsecond astrometry often requires an accurate characterization of the orientation of the Earth in a quasi-inertial reference frame. The International Earth Rotation Service (IERS) standards provide the current state of the art in the transformation between celestial and terrestrial reference systems. Improvements in the determination of Earth orientation parameters which describe this transformation continue to be made. Current and future capabilities are given.

The high precision of present observations makes it reasonable to clear up a question about GRT (general relativity theory) corrections in the problem of Earth's rotation. The answer is that one may almost forget about GRT corrections when dealing in an adequate reference system (RS). The problem of Earth's rotation may be related to the relativistic hierarchy of RS started in (Brumberg and Kopejkin, 1989) and completed in (Klioner, 1993). Let letters B, G and T be related to barycentric, geocentric and topocentric RS, respectively. Let DRS and KRS be dynamically nonrotating or kinematically nonrotating RS, respectively. From the dynamical equations of rotation it follows that the most adequate system for studying the Earth's rotation is DGRS. Apart from the geophysical factors the rotation of the Earth in this system is fairly well approximated by the rigid-body rotation with some angular velocity . The same rotation of the Earth as considered in BRS and DTRS may be also approximated by the rigid-body rotation but with some additive relativistic corrections and with other angular velocities ωi and , respectively. Substituting these three rotation relations into four-dimensional BRS-DGRS and DGRS-DTRS transformations one may express ωi and in terms of and determine the additive relativistic corrections in BRS and BTRS. These corrections are of importance for treating kinematics problems in various coordinate systems and for obtaining physically meaningful solutions of the dynamical equations of rotation in the barycentric reference system.

The complete text will be published in Journal of Geodynamics.

The relativistic effects in positioning of distant objects at microarcsecond level are studied. The main points are: statistics of random variations of the image position of a distant radiation source due to the gravitational field of moving stars, motion of the image due to individual invisible gravitators, and possibilities to obtain information on their masses and velocities. The gravitators shifting the object image may be stars of our Galaxy or of a lensing galaxy.

Achieving angular accuracies at milli- and microarcsecond levels for objects within the solar system presents real challenges concerning the meaning of the observations due to the effects of phase and due to the sizes of the objects being observed. The use of retroreflectors or transmitters located on solar system objects may be the most promising technique.

However, if these accurate observations can be achieved, this provides the opportunity for significant improvements in the orbits of the planets, satellites, and minor planets, and will provide, or require, improvements in the models for the tidal interactions, secular accelerations, and masses. The observations will provide a test, or improvement, of the ties between radar, optical, and radio observations. This should provide an opportunity for improving the reductions of pulsar timing observations and improve our determination of the solar system reference system. This will also test the accuracy of the equivalence between Atomic Time and Dynamical Time. These accuracy levels should permit the discovery of other solar systems and the resulting increase in our knowledge, based on more than a single solar system.

Improvements in the accuracy level of observations provide increases in our knowledge and raises new challenges to explain the signatures detected by the observations.

We showed that it is feasible to measure the mass of a single star by observing the variation of gravitational deflection caused by the orbital motion of the Earth. When the distance of a star is less than 60 pc and some appropriate sources are within 1 arcsec. in its background, not only the distance but also the mass of the star may be determined by measuring the deflection with an accuracy of 10 μ arcsec. In the case of photometric microlensing by a MACHO, the observation of astrometric gravitational deflection is also useful. By measuring the separation between the primary image and the secondary image, the ratio of mass to distance of the MACHO will be obtained. Further, the orbital motion of the Earth modifying the light curve of the source is discussed.

Reference Systems include the reference frames and their relationships, time arguments, ephemerides, and the standard constants and algorithms.

The extragalactic, or radio, reference frame will be the basic frame. Achieving milli to microarcsecond accuracies at optical wavelengths will reduce the disparity between optical, radar, and radio reference frame determinations. Thus, the relationships and identifications of common sources should be much more accurate. Another significant change should be the ability to determine distances, and thus space motions on a three-dimensional basis, rather than the current two-dimensional basis of proper motions.

Improvements in ephemerides provide the opportunity to investigate the difference between atomic and dynamical time, the relationship between the dynamical and extragalactic reference frame and the values of precession and nutation.

Also, the relationships between the bright and faint optical catalogs, the infrared, and extragalactic reference frames should be better determined. Reference frames at other wavelengths will become determinable.

A ground-based project for a small telescope about 20 cm aperture is proposed. It would be able to obtain 500 million astrometric observations per year of all stars between V=7 and 18 mag, based on the Hipparcos-Tycho reference net of one million stars. In addition, V and I magnitudes for all stars from V=7 to 17 mag would be obtained. — A new space mission is proposed, capable of obtaining 10 microarcsecond (μas) precision for parallaxes and proper motions of stars of V = 11 mag, and a precision of 0.5 millimagnitude (mmag) for intermediate-band photometry. The mission is here called Roemer Plus (or Roemer+) and the design is derived from Hipparcos and Roemer, but a larger telescope aperture and high-precision metrology is applied. At magnitude V = 20 a precision of 1.0 milliarcsecond (mas) would be achieved.

The researches on the AIST-project are directed to the design of the focal plane assembly. The concept of the chip-mosaic and numeric simulation of the two-pixels mode of the signal sampling are presented.

The arguments are given of the choice for the Vilnius seven color medium band photometric system as the most effective one for the AIST space project. A prediction is made of the limiting stellar magnitudes V for every Vilnius passband and the integral band T at σΣ ≤ 30mmag from a 3 year AIST mission.

Newcomb is a design concept for an astrometric optical interferometer satellite with a nominal single measurement accuracy of 100 microarcseconds. In a 30-month mission life, it will make scientifically interesting measurements of O stars, RR Lyrae and Cepheid distances, probe dark matter in our Galaxy via parallax measurements of K giants in the disk, establish a reference grid with internal consistency better than 50 microarcseconds, and lay the groundwork for the larger optical interferometers that are expected to produce a profusion of scientific results during the next century.

The OSI project at JPL is a study of a series of concepts for space based astrometric interferometry. The various concepts are designed for slightly different types of astrometry. The original concept OSI was designed for wide angle/global astrometry by look at stars about 30 degrees apart. A derivative of OSI, called SONATA was designed for narrow angle astrometry (10 arcmin) at sub-microarcsec levels for exo-planet searches. This paper describes optical concepts of theses instruments with regard to metrology and systematic error. In addition to conceptual designs, system error budgets and covariance analysis the OSI project has a technology development program in place to investigate the most difficult problems.

Technology is being persued in three areas. One is laser metrology at the picometer level where sub-picometer accuracy for null metrology has been achieved and < 2 picometer accuracy has been achieved for relative metrology. A second area is the control of vibration on a large space structure. A 5 m truss structure has been built at JPL with an interferometer on the structure to test vibration isolation and control technologies. Tests with this structure will begin in winter of 1994. The key to vibration control for OSI will be the delay lines. The third technology project is a shuttle flight experiment, Stellar Interferometer Technology Experiment (SITE), currently in a phase A study, (part of NASA's INSTEP program). The goal is to fly the experiment on the Shuttle in 1997-1998.

Scientific and technical problems of accuracy improvement of stellar position measurements in the LOMONOSOV project are discussed. In contrast to the HIPPARCOS project, in our project we use a steerable satellite with a three-axis spatial stabilization. A Cassegrain-type 1-meter mirror telescope will be mounted on board of the satellite. We discuss scientific goals in astrophysics, stellar astronomy and astrometry, for which the knowledge of stellar positions with sub-milliarcsecond accuracy is necessary.