This paper presents the final discussion of the meeting, following introductory remarks by A. Ardeberg, B. Ellerbroek and C. Cunningham.

Areas surveyed for ELT projects using satellite data are described. A synopsis of the methodology used is given and selected results from recent studies are presented.

A tiny fraction (<1%) of very metal-deficient (12+log(O/H)≤7.6) blue compact dwarf (BCD) galaxies exhibits a nearly galaxy-wide starburst activity and no signatures of an old stellar host galaxy. The evolutionary status and formation history of these most metal-deficient BCDs are still a subject of debate. Various lines of evidence suggest, however, that these systems do not contain a substantial population of stars older than $\sim$1 Gyr and hence qualify as nearby young-galaxy candidates. Elaborated multiwavelength studies of these rare, most metal-deficient BCDs may therefore provide crucial insights into the formation and starburst-driven evolution of low-mass galaxies in the early universe.

ELTs will bring galaxies within 5 Mpc or more “as close as the Magellanic Clouds”. This is why stellar populations is such an exciting subject for ELTs. We can resolve galaxies into stars and learn their history (1) from the “fossil record” of old stars and (2) looking back in time. The confrontation of these two views will bring important new insights in the star formation history of galaxies and stellar evolution, at the same time.

IAU Symposium 232 allows a snapshot of ELTs at a stage when design work in several critical mass projects has been seriously underway for two to three years. The status and some of the main initial design choices are reviewed for the North American Giant Magellan Telescope (GMT) and the Thirty Meter Telescope (TMT) projects and the European Euro-50 and the Overwhelmingly Large (OWL) projects. All the projects are drawing from the same “basket” of science requirements, although each project has somewhat different ambitions. The role of the project offices in creating the balance between project scope, timeline and cost, the “iron triangle” of project management, is emphasized with the OWL project providing a striking demonstration at this meeting. There is a reasonable case that the very broad range of science would be most efficiently undertaken on several complementary telescopes.

A forthcoming step in the study of extrasolar planetary systems is the direct detection and characterization of Earth-like planets. An asset of the ELTs in that context is their very high angular resolution and their collecting area. The luminosity ratio between a terrestrial planet and its star ($10^{-10}$) is such an ambitious goal that a thorough study needs to be carried out. We started with a simple analysis of the fundamental limitations for the detection of extraterrestrial planets with ELTs. Here, we considered an extreme adaptive optics device upstream of a perfect coronagraph. Even with high Strehl ratios, the coronagraphic halo level is only $10^{-6}$ to $10^{-7}$ at typical exo-Earth angular distances. A calibration device is therefore mandatory to reach the contrast between a terrestrial planet and its star in the near infra-red. We considered a simple but realistic model taking into account dynamic aberrations left uncorrected by the adaptive optics system, static aberrations of optical system and differential static aberrations due to the calibration channel itself. Numerical simulations prove that, after the calibration, the limitations are set by the static aberrations which cannot be neglected anymore.

Adaptive Optics (AO) will be essential for accomplishing many, if not most, of the science objectives currently planned for Extremely Large Telescopes including GMT, OWL, and TMT. AO will be needed to support a range of instrumentation, including near infrared (IR) imagers and spectrometers, mid IR imagers and spectrometers, “planet finding” instrumentation and wide-field optical spectrographs. Multiple advanced AO systems, utilizing the full range of concepts currently under development, will need to be combined into an integrated architecture to meet a broad range of requirements for field-of-view, spatial resolution and spectral bandpass.

In this paper, we describe several of the possible options for these systems and outline the range of issues, trade studies and component development activities which must be addressed. Some of these challenges include very high-order, large-stroke wavefront correction, tip-tilt sensing with faint natural guide stars to maximize sky coverage, laser guide star wavefront sensing on a very large aperture and achieving extremely high contrast ratios for the detection of extra-solar planet and other faint companions of nearby bright stars.

We show that contrary to what is expected from 1D stationary model atmospheres, 3D hydrodynamical modeling predicts a considerable influence of convection on the spectral properties of late-type giants. This is due to the fact that convection overshoots into the formally stable outer atmospheric layers producing a notable granulation pattern in the 3D hydrodynamical models, which has a direct influence on the observable spectra and colors. Within the framework of standard 1D model atmospheres the average thermal stratification of the 3D hydro model can not be reproduced with any reasonable choice of the mixing length parameter and formulation of the turbulent pressure. The differences in individual photometric colors – in terms of 3D versus 1D – reach up to $\sim0.2$ mag, or $\Delta T_{\rm eff}\sim70$ K. We discuss the impact of full 3D hydrodynamical models on the interpretation of observable properties of late-type giants, briefly mentioning problems and challenges which need to be solved for bringing these models to a routine use within the astronomical community in 5–10 years from now.

The enrichment of the intergalactic medium (IGM) with heavy elements provides us with a record of past star formation and with an opportunity to study the interactions between galaxies and their environments. We discuss what future observations with extremely large telescopes could do for this field. We conclude that a further increase in the quality of the spectra of bright, $z\sim 3$ quasars will be useful, but may not lead to dramatic progress. In contrast, the ability to obtain high-quality spectra of $z>5$ quasars will be extremely exciting because they will allow us to extend studies of the distribution of metals to early times and to lower density contrasts, and because they may enable us to study the end of reionization. Spectacular progress in our understanding of the interactions between galaxies and the IGM can be made at $z\sim 3$, by obtaining accurate redshifts for large numbers of faint galaxies in the fields of bright quasars, and by obtaining low and intermediate resolution absorption spectra of some of the brighter galaxies.

Current models for dark energy include a cosmological constant and scalar field models such as quintessence and k–essence. By measuring the dark energy sound speed, $c_{\hbox{\scriptsize\it eff}}$, and its equation of state parameter, $w$, one can distinguish between these. Here we investigate the possibility of measuring $c_{\hbox{\scriptsize\it eff}}$ and $w$ using combined observations from the SKA and CMB experiments. We present theoretical predictions for the cross–power spectrum of ISW fluctuations in the CMB and the expected field of HI galaxies detectable with the SKA.

I present a brief summary of the parallel session on exoplanets and star formation, and how their study drives the science requirements for extremely large ground-based telescopes. I also offer a few thoughts on the development of these highly ambitious and highly expensive new astronomical facilities, as viewed from the perspective of someone also involved in the James Webb Space Telescope.

Future extremely large telescopes will be the engines of major progress in the fields of star and planet formation, brown dwarfs, and extrasolar planets. Their throughput will enable spectroscopic studies of the structure of brown dwarf atmospheres; reveal the composition and kinematics of protoplanetary disks; extend radial velocity searches for extrasolar planets to fainter stars and lower masses; and characterize the surfaces of the most distant Kuiper Belt objects. Their resolution will allow us to resolve and track the orbits of close binary young stars and brown dwarfs, establishing their dynamical masses and calibrating their evolutionary tracks; probe the inner region of young stellar object disks, resolving the jet collimation region in accreting systems and disk inner holes in more evolved ones; and move the horizon for stellar companion searches inward to 40 mas, enabling the direct detection of hot young planets orbiting 10 AU from young T Tauri stars. The high contrast imaging capability of future ELTs is uncertain, and depends on future developments in extreme adaptive optics. Wavefront control considerations suggest a fundamental atmospheric contrast limit of 10$^{-8}$ for companion searches to nearby solar-type stars, below which detections are unlikely to be possible. To reach this performance level, continuing investments will be needed in extreme adaptive optics work, in addition to careful attention to the specialized requirements of ultra-high contrast imaging in ELT design. Above the 10$^{-8}$ contrast limit, a 30m telescope has the potential to directly image about a dozen of the currently known radial velocity planets.

We review the need for technology developments to meet the science cases that have been assembled for optical and infrared telescopes from 20 to 100 metres. Novel technologies can make an impact on the scientific capabilities of these ELTs, make them more affordable and decrease the operating costs. We consider those technologies highlighted by design studies and technology roadmapping exercises on both sides of the Atlantic, with particular emphasis on instrumentation. We do not consider adaptive optics (AO), which is covered by Ellerbroek & Hubin (this volume). Finally, we recommend a joint technology development programme to enable a world-wide suite of ELTs to be built.

Simulations of galaxy formation require strong feedback to produce realistic galaxies. Observations of starburst galaxies both at low and high redshift often show signatures of strong galaxy-wide winds. Are such winds common, and are they the mechanism which pollutes the intergalactic medium with metals? Which are the sources of ionising photons at higher redshift? Detailed observations of galaxies and their surroundings and comparison with cosmological simulations to eliminate biases and test models are required to address these questions.

The next generation of ground-based telescopes will give access to resolved stellar populations in previously unexplored environments. To highlight the potential for stellar spectroscopy using an extremely large telescope (ELT), we present data from an ongoing project with the Very Large Telescope (VLT) to calibrate a novel method of distance determination. An ELT would extend the reach of this method and would allow quantitative studies of individual stars well beyond the Local Group, sampling a wide range of galaxy morphologies and metallicities.

Over the coming two decades astronomers will plan, construct and exploit extremely large telescopes operating at both optical (ELTs) and radio (SKA) wavelengths. In studies of galaxy evolution and cosmology the combination of the two types of facility are likely to be extremely powerful. In this paper I will review the key questions that the ELT/SKA combination seem set to solve, focussing on the unique advantages afforded by optical and radio observations against the backdrop of the other theoretical and observational information to be available by 2020.

The star formation, mass assembly and chemical enrichment histories of galaxies, and their present distributions of dark matter, remain encoded in their stellar populations. Distinguishing the actual distribution functions of stellar age, metallicity and kinematics at several locations in a range of galaxies, sampling across Hubble types and representative environments, is the information required for a robust description of galaxy histories. Achieving this requires large aperture, to provide the sensitivity to reach a range of environs and Hubble types beyond the Local Group, to provide high spatial resolution, since the fields are crowded, and preferably with optical performance since age-sensitivity is greatest near the main-sequence turn-off, and metallicity-sensitivity for these warm stars is greatest in the optical.

Metallicity may play an important role in the symbiotic phenomenon. Unfortunately, chemical abundances of symbiotic stars have been thus far poorly studied. Ongoing abundance analysis of a sample of over 30 symbiotic stars based on high-resolution, near-infrared spectra obtained with the Phoenix spectrometer on Gemini South telescope will allow for the first time to address properly the metallicity problem as well as provide important information about the past history of these binaries.

The future planning working group of the optical and infrared astronomy community of Japan completed its two-year study to yield a concrete plan for ground based telescopes and space missions for the coming decades. The R&D issues towards the realization of extremely large telescopes reported in this paper include a novel optical design, development of new ceramic mirrors, high precision grinding approach to reduce the involved cost and time for fabricating aspheric segmented mirrors, and conceptual studies of instruments.

The present paper is essentially a slightly updated version of the paper presented in the workshop “Instrumentation for Extremely Large Telescopes” held at Ringberg Castle, Bavaria, 25-29 July 2005.

We report an on-going blank-field multi-wavelength deep and wide survey of the Subaru/XMM-Newton Deep Survey Field (SXDF). The SXDF has been the focus of a wide range of multi-wavelength observing programs spanning the X-ray to the radio. These observations cover a large enough area (the initial optical imaging covers $\sim$1.3 deg$^{2}$) and depth ($B=28.2$) that they are not affected by large-scale structures (which exist on tens of Mpc scales) and allow us to study the distribution and evolution of high-$z$ galaxies and AGNs, and thus constrain theories for their formation. Our early results include: i) an indication of the primeval Large Scale Structure (LSS) at z $\sim$ 5.7, ii) an indication for the down-sizing of galaxy formation at z $\sim$ 1, iii) identifications of passively evolving systems, and evidence for early formation and the passive evolution of present-day early-type galaxies, and iv) discovery of a large number of optically obscured QSOs. As for the next step forward - we express our hope to use the next generation optical/IR Extremely Large Telescopes (ELTs) to obtain larger and deeper spectroscopy samples of the high-$z$ objects.