Introduction

It has always been ESO's aim to operate the VLT in an interferometric mode (VLTI) which allows the coherent combination of stellar light beams collected by the four 8m telescopes (UTs) and by several smaller auxiliary telescopes (ATs). In December 1993, in response to financial difficulties, the ESO Council decided to postpone implementation of the VLTI, Coudé trains and associated adaptive optics for all the UTs but included provisions for continuing technological and development programmes devoted to the aim of reintroducing these capabilities at the earliest possible date (see Messenger No.74,December 1993). In July 1994,the ESO council approved a revised VLTI implementation plan to provide at least the VLT interferometric sub-array (VISA) consisting of three ATs by the year 2003 and, provided additional funds could be obtained, integration of the UTs by 2006.

The desirability of carrying out the full VLTI programme as originally envisaged at the earliest possible moment has not,however,diminished especially in view of the of VLTI's exceptional capabilities and resulting potential for new and exciting discoveries. In recent years, interferometric projects have begun to play a central role in ground-based high-resolution astronomy, and numerous instruments have been completed or are in the process of construction (see Table 1 for a summary of the present situation in this regard). Several large-aperture interferometers will probably come on-line near the turn of the century. The impending presence of these new instruments represents an important incentive both for clarifying the scientific cases for various VLTI implementation plans and for ensuring VLTI's competitiveness in the international context over the next 10-20 years.

The complexity and ambitious scope of VLTI mean that its astrophysical repercussions are difficult to define fully, even for many of its most vocal supporters. However, the primary scientific issues that it seeks to address are well defined, although there remains a need to present these coherently to the wider community in order to justify the significant resources which the project requires. Another pressing need is to develop an implementation plan that will optimally exploit the various technological stages of the project and ensure their compatibility with a vigorous, yet realistic and timely, astrophysical programme.

In order to study these issues and to establish a clear set of guiding principles for the development of VLTI, a new Interferometry Science Advisory Committee (ISAC) was established in April 1995. This committee has met twice to review the present technical status of VLTI, its scientific rationale as elaborated by past advisory panels (ESO VLT Reports 59 & 65), and the present recovery plans. The committee has now begun to define and prioritise the key science drivers for the program and the technical specifications that flow from them. This article briefly presents these science goals as they currently stand. The list is not meant to be frozen or complete, but rather is intended to stimulate community reflection and comment. The preliminary recommendations of the committee are discussed in the last section of this article.

To provide a forum for the discussion of the ideas presented here, ESO has decided to host a workshop on ``Science with the VLTI,'' in Garching on June 18-21 of this year (see Messenger, No.82 and the announcement in this issue). It is hoped that this will allow the whole of the ESO community to further refine the concepts outlined in this article and to make a case for any capability or role omitted here.

Table 1