II. Science Issues and Opportunities

A. Stars and Planetary Systems

We have identified four critical areas in this field, each of which is at, or simply beyond, the limiting light-grasp, image quality and sensitivity of 4-m class telescopes and in which the particular strengths of the Gemini telescopes will allow them to play major roles:

1. defining the stellar initial mass function by finding and characterizing candidate brown-dwarfs and extra-solar system giant-planets,
2. studying the physics of nearby stars in similar detail to the Sun,
3. establishing analogs between the stars in our own and other galaxies,
4. mapping the structure of magnetic fields, circumstellar discs and accretion phenomena in interacting binary systems with white dwarfs and stellar-mass black hole primaries.

1. Brown Dwarfs and Giant Planets

1. Direct imaging: speckle-noise severely limits the ability of ground-based telescopes to detect faint, close companions. To limit speckles and compete with the NICMOS camera dictates an AO system giving infra-red Strehl ratios ~1. Only then would it be feasible to search for gas-giant companions between a fraction of an arc second and several arc seconds from nearby (<10 pc) stars. Neighboring filters on and off a methane band-head may provide the best sensitivity for this search. Modeling has demonstrated that deep methane absorption leads to wide ranging near-infrared colors for brown dwarfs as they age and cool. The blue near-infrared colors (e.g. J-K < 0) have confused a large number of past searches predicated on detection of red objects.

   A high performance coronagraph on Gemini would allow a statistically significant census of brown dwarf companions to distances of ~100 pc. Any detections of low-mass companions would provide precise astrometry which, with radial velocities, would yield masses, absolute magnitudes, and temperatures, all basic to understanding the sub-stellar mass function, and ultimately the transfer of angular momentum in star formation, and the universal proportion of matter bound up in sub-stellar companions.

2. Precise radial velocities offer the greatest sensitivity at the moment in the detection of, or setting limits to, low-mass companions. For gas-giants, program stars must be observed regularly with a precision of a few m/s at R>100,000. But there must be parallel checks for, and calibration of, intrinsic line variability, plus monitoring of chromospherically active lines, which requires R>300,000 to resolve the stellar lines.
3. Within one or two tenths of an arcsec of the parent star, planetary or Brown Dwarf companions cannot be imaged directly but their spectra may be detectable using phase sensitive spectroscopy because of their large velocity excursion and distinct spectrum. The technique demands exact registration and subtraction of the primary spectrum with "perfect" flat-field calibration. In combination with the coronagraphic AO system, spectra could be obtained of close but spatially resolved faint companions.

2. Stellar Properties

1. Convection. There is considerable interest in comparing the level of convection in stars at various stages of evolution, metallicity, rotation, magnetic field strength, etc. with the Sun where our understanding of convection still remains imperfect. Observations covering a range of spectral types would also allow a realistic theory for the changes which depend on evolution, and differences in mass. Convective motions can be estimated from asymmetries in line bisectors derived from high resolution spectra. These would provide new insights into turbulence and dredge up rates providing important data for the interpretation of surface abundances and for understanding the importance of rotation in mixing.
2. Magnetic Fields. For stars with weak integrated magnetic fields like the Sun, Zeeman splitting cannot be resolved. On the other hand, stars with strong fields such as active M-dwarfs are generally too faint for observation with adequate spectral resolution. Thus, results to date are inconclusive. In particular the question of field strength vs. area coverage is ambiguous since it requires detailed modeling of the line profiles, particularly of many lines with different temperature sensitivities and continuum contrasts. Figure 2 illustrates the type of measurements required. Combined optical and IR Zeeman measurements would be invaluable.
3. The connection between rotation, magnetic field strength, angular momentum loss, age and chromospheric/coronal activity is poorly understood at present. For low-mass (often very active) stars, rotation rates are clearly of the order of, or less than, 1 km/sec; the lines can only be resolved with R>300,000. Further, such stars are too faint to achieve the necessary S/N for profile modeling with 4-M telescopes.
4. Basic atomic/molecular data such as the TiO laboratory line lists are frequently in error even among the well studied optical bands. Consequently, attempts to model the atmospheres of cool stars to determine metallicity will fail. In addition, better line data are required to understand how a two component model which includes starspots will alter basic parameters such as temperature and gravity, compared to one component models. Improved TiO data could be obtained from the empirical study of cool stars at very high resolution.
5. Isotopic abundances, where line splitting is very small and the isotope abundance very low, requires careful modeling of very high resolution data.

For this wide range of problems in the physics of stellar atmospheres, stellar line profiles must be resolved which, for late-type stars, demands <1 km/s or R>300,000, at optical and near-IR wavelengths.

3. Star-Galaxy Connection

1. The age-abundance relation in nearby galaxies. HROS observations at R=50,000 can provide detailed abundances out to 1 Mpc by classical techniques at moderate to good S/N. At larger distances, very low S/N HROS observations with good sky subtraction can yield precise radial velocities and thereby mean heavy-element abundances as shown in Figure 3. Long- and multi-slit HROS observations would be particularly suitable in crowded fields of nearby galaxies.
2. Eclipsing spectroscopic binaries in about twenty of the nearest globular clusters (preferentially in the South) would be best observed at R=50,000 for accurate velocity curves. Such velocities would impose unique constraints on stellar evolutionary models and cluster ages.
3. Abundances and luminosities of supergiants measured at R~120,000 in the nearest galaxies will impose limitations on estimates of fundamental stellar properties in different environments, particularly the Magellanic Clouds where accurate abundances and calibrated mass-loss rates as a function of luminosity can be determined. With Gemini the technique could be extended to nearby galaxy groups.

4. Surface Mapping

Flickering (0.1-1 mag on time scales of seconds to minutes) is a fundamental signature of accretion (Figure 4), yet it is the most poorly studied and badly understood phenomenon of the accretion processes. Time-resolved spectroscopy can be used to study flickering in the continuum and emission lines by measuring correlations and delays at different wavelengths. At high spectral resolution, velocity-resolved flickering can be studied across emission lines. The power spectrum of flickering may show a high-frequency break corresponding to the frequency of the innermost Keplerian orbit in the accretion disk (~10 s for a disk around a white dwarf). Spectroscopy at ~1 sec time resolution could test this hypothesis. Flickering studies have so far been restricted to Cataclysmic Variables because most low-mass X-ray binaries (LMXB) are very faint (<18 mag). Accretion disks in LMXBs have far smaller inner disk radii than those of CVs and are subject to much stronger irradiation effects. Gemini will be particularly important in extending flickering studies to LMXB and to compare flickering behavior for different accretion environments and regimes.

Much of the optical emission in LMXB arises from reprocessing of X-rays and, for a distant observer, its arrival is delayed relative to the X-rays by light travel times of order several seconds. By analyzing simultaneous optical and X-ray light curves one can recover the range of delays present between the two (echo mapping), producing a one dimensional map that resolves the reprocessing sites on the iso-delay surfaces which are nested paraboloids around the line of sight to the X-ray source. Phase-resolved echo-mapping will generate two dimensional maps of X-ray reprocessing sites and accretion flows projected onto the binary orbit (echo tomography). The spatial resolution is set by the time resolution of the observations. This technique was recently applied to Sco X-1 by combining XTE observations with 1-sec time-resolved spectroscopy on 4-m class telescopes. A larger facility such as Gemini will be fundamental to extend echo tomography to fainter LMXBs. Simultaneous X-Ray observations (with AXAF or XTE) and time-resolved spectroscopy at high time-resolution (0.1-1 sec) will resolve the structure of the reprocessing sites in LMXBs at different wavelengths to ~ 0.1-1 light seconds.

b) Stellar discs and ejecta: Young and old luminous B stars (e.g. Herbig Be stars, embedded BNtype sources, classical Be stars and B[e] supergiants) tend to have circumstellar discs and/or outflows of uncertain characteristics and origin. Star formation and the formation of planetary systems are central concerns at this time, and there is particular interest in Herbig Be stars and BN-type objects, the massive counterparts of T Tauri stars. The majority of these objects are too distant for their ionized circumstellar discs to be resolved in direct images, even at the 0.1 arcsec level, but there are some exceptions and it will, in addition, be possible to use spectropolarimetry to indirectly "image" (on the scale of a few stellar radii) and to study more extended faint nebulosities where present.

Spectropolarimetry is highly 'photon-starved' since the percentages of linear polarization are typically a percent or so even for Be stars presenting nearly edge-on discs. In the past this has meant that polarimetry has mostly been applied at broad-band rather than high spectral resolution. Spectropolarimetry at a resolving power of ~5000 on 4-metre class telescopes is only feasible to a magnitude of ~12 which excludes all but a handful of embedded sources. Spectropolarimetry at echelle resolution is not a routine option anywhere, but could be implemented at little cost with Gemini/HROS.

5. Instrumentation Requirements

• Adaptive Optics with coronagraphic capability on GS with equivalent performance to system on GN
• Adaptive Optics optimised for coronographic imaging and with high (>0.9) near-infrared Strehl ratios
• Near-infrared camera optimized for AO coronagraphic use
• Stable bench-mounted optical spectrograph, R=120,000, ~100nm simultaneous wavelength coverage usable with an absorption cell in 500-600 nm range, covering through the H & K lines, plus a R>300,000 mode from 380 to 1000 nm
• High resolution (~100,000) near-infrared spectrograph capable of accurate flat-field calibration. Equivalent to Phoenix with multi-order capability
• Spectrographic capability with R=20,000 to 40,000 covering the HROS/GMOS gap
• Pulse counting detectors with high quantum efficiency

6. Upgrades

B. Star Formation and the ISM

The Gemini telescopes will play a central role in our quest to understand stellar and planetary birth more deeply -- by virtue of their potential to provide diffraction-limited images from 1.6 microns to 20 microns, and light gathering power sufficient to permit moderate- to ultrahigh resolution infrared and optical spectroscopy of both deeply embedded and optically-revealed young stars and their circumstellar environs.

2. The Stellar Initial Mass Function

The advent of sensitive infrared array detectors and their deployment on large telescopes over the past decade has enabled astronomers to penetrate the birthplaces of stars -- molecular clouds and dense molecular cores. Observations with these arrays have revealed young stars forming under a variety of environmental conditions: in isolation, in aggregates of a few tens of stars and in rich, dense clusters of stars. Because these forming stars can be observed at ages t << 1 Myr, we can be certain that (1) they are located close to their birthplaces; and (2) are relatively isolated from other star-forming events in the cloud. Owing to these favorable circumstances, we can for the first time provide the empirical basis for answering the following fundamental questions: What is the frequency distribution of stellar masses characterizing an isolated star formation site and how does it vary from one site to the next? What initial conditions favor the production of high and low mass stars? Do high and low mass stars form at the same time within a star-forming core? Does the formation of high mass stars require special environmental conditions? Does the frequency of binary star formation differ as a function of environment?

To address these issues require observations which are in principle simple: near-infrared (JHK) photometry and moderate resolution near-infrared spectroscopy (R-2000-5000) which enable determination of stellar effective temperature, interstellar extinction and stellar luminosity. In practice, these quantities are difficult to determine, particularly in the rich, dense stellar clusters which are presently thought to dominate star formation in the Milky Way, because of (1) source confusion; (2) irregular background (owing to the combined effect of complex reflection and emission nebulosity.). To probe the initial mass function over the full range of masses -- from the hydrogen burning to the Eddington limit -- requires (1) light gathering power sufficient to penetrate optically opaque molecular cores in which extinction values may range from 20 < Av < 100 mag. and detect a representative sample of intrinsically faint stars near the hydrogen burning limit: (2) high spatial resolution in order to obtain accurate photometry and spectra in confused regions with high background.

As an example of what we might expect to do with these tools, Figure 5 shows a plot of the HR diagram constructed from OPTICAL photometry and multi-object spectroscopy (multislit in the inner regions; fibers in the outer regions) of the Orion Nebula (Trapezium) Cluster -- recently emerged from its natal core.

That the shape of the IMF may depend critically on initial and environmental conditions is illustrated in Figures 6 and 7.

3. Physical and Chemical State of Molecular Clouds and Cores

The ionization fraction within a molecular cloud is critical in determining the ambipolar diffusion time and therefore the rate of formation of magnetic-flux-free molecular cores. Fundamental to determining the ionization fraction is the probability that an interstellar molecule is ionized by an encounter with a cosmic ray, which is known only very poorly (estimated to have a value of ~10^-17/sec). Ionization of molecular hydrogen, H₂, by cosmic rays creates the electrons which support magnetic fields, and subsequent reaction of H₂+ with H₂ creates H₃+, is believed to initiate nearly all of the chains of reactions which lead to the production of complex molecules observed in dark clouds. Now that detection of H₃+ in dark clouds has been shown to be feasible, measurements of absorption lines of it, H₂, and CO (the dominant destroyer of H₃+) will allow accurate determinations of the ionization rate.

Gemini, equipped with high resolution infrared spectrographs operating in the 1 to 20 micron region will play a critical role in determining the ionization state in molecular clouds. High spectral resolving powers (R > 100,000) is imperative for detecting the weak and narrow lines of H₂ and H₃+ species in molecular clouds (characterized by temperatures T < 30 K). High angular resolution (~0.1") will be crucial as, e.g. for H₂, line absorption in the quiescent cloud against the pointlike embedded or background source is known in some cases to be contaminated or filled in by extended emission from fluorescent or shocked H₂ - this effect will be most easily corrected with good image quality and a narrow slit. High sensitivity will enable more deeply embedded sources to be measured, as well as multiple sources in a single cloud. The most accurate value of the ionization rate will come from statistical studies of many clouds combined with knowledge of their sizes from radio/submm mapping.

Gemini can also provide the observational basis for understanding the role of grain chemistry in clouds and determining the abundances of molecules and ices that may later be incorporated into cometary bodies. The relative abundances of molecules in the gas and solid states can be determined by similar absorption measurements of gas sampling the same line of sight. The ISO mission has provided the first extensive observations of solid state absorption features (ices of CO₂, CH₄, etc.) and measurements of gas-phase water abundance in a variety of molecular environments. Ground-based observations of molecular absorption lines in the 3-20µm range for these sources would provide far deeper understanding of the gas and solid phase chemistry which is obtained in these regions. High velocity resolution is required to improve detectivity of complex molecular species, to resolve multiple velocity/temperature components, and (in some cases) to separate telluric from interstellar lines. Examples of molecular abundances that can be measured are CH₄, C₂H₂ (both not observable in the radio), CO, H₂, HCN, NH₃, CS and OCS -- all critical to developing a quantitative understanding of the relationship between gas and solid phase chemistry.

High-resolution spectroscopy (R > 30,000) of molecular transitions in the 1-20µm range towards embedded and background sources in molecular clouds and cores are central to these studies. Near IR wavefront sensors would enable 3-20µm imaging and spectroscopy during early daylight hours.

Magnetic fields almost certainly play a central role in the star formation process. They are important in all stages from the collapse of molecular clouds to the formation and evolution of disks and envelopes. The component of the field in the plane of the sky can be traced through the effects of aligned non-spherical dust grains on the transmission, emission and/or scattering of light. The study of magnetic field directions at high (arcsec) resolution has been confined to the brightest star forming regions. For example, dichroic extinction measurements have been made with imaging polarimetry towards the diffuse unpolarized H₂ emission at 2µm in Orion and through observations of polarized emission and absorption at 17 microns in Orion. As with all polarimetric measurements, Gemini offers very substantial gains but the advantages at these infrared wavelengths are particularly large.

4. Evolution of Disks and Envelopes

Gemini will enable detailed study of the structure of YSO envelopes and the processes of infall and outflow in the immediate vicinity (distances ranging from 10 to 1000 AU) of the star/disk system from analysis of adaptively-corrected infrared images and high resolution infrared spectroscopy.

A. Outflows

High spatial resolution imaging in spectral lines probing shock-ionized gas (e.g. [FeII] 1.6 microns) will enable observation of energetic jets emerging from the inner regions of accretion disks such as that associated with HH30 (Figure 8), and the shocked regions which mark the interaction between jets and/or neutral winds and circumstellar material. Such observations are critical to understanding the role played by high energy outflows in terminating infall and determining stellar masses. High angular resolution is crucial in order to probe regions 5-10 AU from the stellar surface -- where the wind/cloud interactions should be revealed.

The origin (disk or star?) and initial collimation region of [FeII] jets can be probed through long-slit spectroscopy using the technique pioneered by Solf and Böhm (1993) to remove the underlying stellar continuum. For a YSO located in a nearby (d ~ 150 pc) star-forming region, the diffraction limit of Gemini is ~8 AU at 1.65 microns. By obtaining R ~ 20,000 infrared spectra at a variety of position angles, it should be possible to observe jets to within 0. 4 AU of the stellar surface.

B. Accretion

High spatial resolution imaging and polarimetry of scattered light regions are required to probe the infalling envelope and disk regions. The 1-5 micron region is critical to these studies with extension to shorter wavelengths for more evolved objects and into the mid-infrared for measurements of the thermal emission. Ideally, the observations would be carried out with a high precision imaging polarimeter sampling stable AO-corrected images. A near IR WFS would be valuable as these objects lie in dark clouds.

The potential of imaging with high Strehl ratios (SR > 0.5) down to wavelengths as short as 1.6 microns in principle enables imaging of circumstellar disks with resolutions of 0.05" (7.5 AU in the nearest star-forming regions; 25 AU in Orion). Observations of scattered light, particularly from the inner regions of these disks, may prove challenging even with coronagraphic rejection of light from the central YSO and polarimetric imaging (owing to the high contrast ratios expected in the case of flat, optically thick disks). However, with AO-fed coronagraphs and polarimetric capabilities, we should be able to probe basic disk parameters: size, surface density falloff at large distances. Moreover, in regions such as the Orion nebula, with sufficiently high backgrounds from nebular emission or scattered light, it should be possible to carry out IR surveys of disk sizes using the "silhouette" technique where the disks are observed in absorption against the bright nebular background. With carefully chosen samples, it should be possible to develop an empirical relationship between disk size and stellar age -- critical for constraining our understanding of disk accretion physics.

Statistical studies of near and mid-infrared excess from young stars provide a means to understand the evolution of dusty circumstellar disks at planetary formation distances of several AU. Sensitive mid-infrared imaging (5 microns to 20 microns) will enable detection of optically thin emission from post-planet-building debris disks. By carrying out such measurements for large samples of stars spanning an age range from 1 Myr to 1 Gyr, we can trace the evolution of emission arising from such disks from their origin to the end of the epoch of maximum bombardment. Gemini -- with its diffraction-limited imaging capability and low emissivity -- will be uniquely well suited to making such measurements. Low-moderate resolution spectroscopy of resolved images of disks would enable detection of features arising from the solid constituents as well (e.g. water, ammonia ices).

However, study of the gas properties of accretion disks requires other techniques. At radial distances of <10 AU, the expected temperatures are sufficient to excite rotational and vibrational transitions of molecules, making infrared spectroscopy well suited to investigate the kinematics, chemistry and evolution of gas in the inner regions of circumstellar accretion disks. Infrared molecular spectroscopy is likely to offer the only method to address various issues such as the radial temperature distribution and vertical structure of disk atmospheres, the timescale for clearing of gas in the inner disk, the formation of giant planets through kinematic detection of disk gaps, and the chemistry in protoplanetary disks. This requires sensitive spectroscopic measurements of the strengths and velocity profiles of emission/absorption lines in the near and mid-infrared (2 - 20 microns) regions. In order to resolve lines originating at distances of a few AU, resolutions of about 100,000 are necessary. Important features to observe include the fundamental transitions of CO near 4.6 microns, the rotational lines of H₂ in the mid-infrared, and various molecular trace species (e.g., H₂O, HCN, CH₄) in the 3-20 micron region.

In the last few years, important kinematic evidence for accretion disks around both low-mass T Tauri stars and higher mass YSOs has come from high-resolution spectroscopy of CO overtone emission which reveals the velocity signature of Keplerian rotation; for example, Figure 9. In T Tauri stars, the CO traces gas of 1500-4000 K within 0.3 AU of the central star. The radial temperature and column density distribution of the hot gas can be derived from modeling the CO overtone data. The ability to study the physical properties and kinematics of gas at these radii is important since this is the region in which magnetospheric accretion, disk truncation, and the generation of YSO winds are believed to occur. Much of the high-resolution infrared spectroscopy has concentrated on the brighter T Tauri stars over limited wavelength regions. Great gains in the analysis of circumstellar gas can be made with a combination of greater sensitivity from large telescopes and increased wavelength coverage from a crossed-dispersed echelle. For example, observation and modeling of the entire CO overtone band, rather than a single bandhead, can place strong constraints on the gas properties. Current spectrographs with R>=40,000 (e.g. CSHELL, Phoenix) require 10-20 grating settings to cover this region.

The ability to study gas dynamics in such systems suggests the exciting possibility of kinematic detection of newly formed giant planets through their influence on the protoplanetary accretion disk. The data in Figure 9 required 1-2 hours at the IRTF. Both higher S/N and higher spectral resolution are required to clearly define the velocity profiles. In addition, large wavelength coverage is desirable in order to simultaneously observe a number of CO lines and thus derive gas temperatures and column densities. With the large aperture and low emissivity of the Gemini telescopes, a R ~ 100,000 spectrograph will provide about a 3 magnitude increase in sensitivity over current observations, allowing easy observation of solar mass T Tauri stars in nearby star forming regions. The large sensitivity increase will also provide the ability to measure trace species other than CO and open up the study of protoplanetary disk chemistry.

5. Substellar Mass Companions and Debris Disks

A. Substellar Companions

B. Debris Disks

Gemini provides the imaging quality and light gathering power to enable high Strehl ratio, adaptively-corrected coronagraphic imaging of post-planet building disks analogous to that surrounding Pictoris (see Figure 10). For Pic, such images would enable mapping and dust surface density and characterization of dust properties from a few to nearly 1000 AU -- providing thereby a unique "snapshot" of a system thought to be similar to the solar system during the epoch of maximum bombardment. Coronagraphic optical imaging will enable surface density and morphology observations to distances of a few AU. Low resolution near- and mid- IR spectroscopy, aimed at detecting ice, silicate and PAH like features, will allow studies of the chemical properties of the solid constituents. Ultra-high resolution spectroscopy of -Pic-like objects offers the possibility of detecting and characterizing cometesimals as they spiral into the central object (Figure 11). Observations at ultra-high resolution (R > 500000) which enables tracing the release of molecular and metallic species from the putative comets (NaI, CaII, CN, CH, CH+...) as these bodies approach the parent star.

These observations would provide keys to understanding the chemistry of the solar nebulae, as well as characteristics of the cometary bodies populating debris disks, both key insights into the early evolution of solar systems.

6. Required Instrumental Capabilities

1. A laser-guide star AO system capable of delivering high Strehl ratios at 1.6 microns and longward (for studying crowded regions of forming clusters; for resolving envelopes and disks).
2. An AO optimized IR imager with coronagraphic capability (for observing high contrast ratio scenes: e.g. envelopes and superplants around forming stars).
3. Polarimetric capability (for diagnosing scattered light patterns produced by infalling envelopes and disks).
4. An IR spectrograph with R~5000 which enables multi-object spectroscopy over the isoplanatic patch at 2 microns; slits and optimally-designed integral field unit(s) are required (for obtaining spectra of embedded YSOs in crowded regions with complex, irregular background emission).
5. Very high spatial resolution (<0.1") and spectral resolution (R >> 30,000) in the 2-25 micron region. GNIRS in its highest reolution mode and/or Phoenix will be needed. Phoenix would provide an important near-term high resolution capability for the near-infrared. In the intermediate/long term, a high resolution 1-5 micron spectrograph with greater simultaneous wavelength coverage is highly desirable. In the mid-infrared, Michelle would provide critical capability which will allow a large number of pioneering observations in this largely unexplored wavelength region. In the long-term, resolution R >= 100,000 may be required (for determining kinematics of envelopes, disks, jets; for probing chemistry of gas and solid phase for envelopes and disks).
6. Ultra High Resolution (R>500,000) bench-mounted optical spectrograph. Initial studies would benefit from a R=120,000 mode (for studies of infalling cometary bodies in post-planet-building disks).
7. Near-IR Peripheral wavefront sensor (PWFS) would enable both imaging and spectroscopy beyond 3 microns during early daylight hours -- thus expanding the time available for many of the studies outlined above.

C. Galactic Structure and Nearby Galaxies

2. New Views of Massive Stars

A. Mass Loss and Stellar Evolution

By obtaining optical spectra of OB stars in different metallicity environments, it will be possible to determine mass loss rates by using model atmosphere techniques to fit the H line profiles and to parameterize mass loss as a function of metallicity. High angular resolution (~0.2") is required to locate single OB stars and a spectral resolution of 25,000 is necessary to resolve the line profiles and to remove nebular contamination.

This project requires that we extend our understanding of stellar mass loss to a wider range of conditions, and that we more effectively combine results from UV space observations with ground-based measurements. Local Group galaxies such as M31 (metallicity > solar) and M33 (steep metallicity gradient down to metallicities < SMC) are ideal candidates for this program (see Figure 13). Expected V magnitudes for normal massive stars are in the range 18-24 mag and a S/N of ~50 is required. Furthermore, for the most luminous stars this program can be observed to well beyond the Local Group with similar S/N.

B. Super Star Clusters

Gemini with its unique near and mid-infrared imaging and spectroscopic capability allows the study of dust-embedded SSCs. High spatial resolution and thermal infrared imaging of SSCs will give unique views of the SSCs and surrounding interstellar medium, while near infrared spectroscopy with an IFU will provide identification of the massive winds of pre-Wolf-Rayet and Wolf-Rayet stars (e.g., broad HI, HeI, HeII and NIII emission lines in the K-band).

While evolving through their stellar wind and final supernova phases, massive stars inject large amounts of momentum, kinetic energy, and copious amounts of newly synthesized heavy elements into their surroundings. These concentrated energy flows from numerous massive stars in SSCs will have an influence extending well beyond the cluster and into the surrounding galaxy, e.g., through the production of supershells or galactic fountains.

Another consequence of SSC formation may be short-lived (<10⁷ yr) zones of metal over-abundances (e.g. C, N, O, S, Ne) which should be observable from optical/infrared emission lines on 10-100 pc scales if these new elements are mixed into a cool (T~10⁴K) phase of the interstellar medium. Spectroscopic (0.37 - 1.00 microns with R ~ 10,000) mapping of starburst regions in nearby galaxies in the principal nebular diagnostic emission lines, together with H velocity maps (R ~ 30,000) with 0.2 arcsec angular resolutions should allow the input and dispersal history of enriched gases produced by evolving SSCs to be quantified.

These programs will need a combination of AO and IFUs (or multiple IFUs?) to feed GMOS giving a field 5.0 arcsec x 0.1 arcsec with proper sampling in the optical and the near IR.

3. The Formation of Galactic Spheroids

A. Kinematics

Most of the mass in galaxies is thought to reside in their invisible halos of dark matter. Planetary nebulae and globular clusters provide excellent tracers of the halo gravitational potential and can be studied well beyond the furthest extent of the visible stellar spheroid to measure the total mass. Detailed kinematic studies of the resolved late-type stellar population in the nearest galaxies, combined with the photometric and spectroscopic analysis discussed above, will also help to establish the still obscure relationship of the bulge, disk and halo and their role in the galaxy formation process.

B. Stellar Populations

Among the most challenging, but also the most scientifically lucrative, targets for detailed study are the innermost regions of galaxies. These areas are best observed at wavelengths longward of 1 micron to (1) optimize image quality, (2) probe the component that dominates the mass distribution, and (3) provide a direct comparison with the nucleus of the Milky Way, which cannot be observed at optical wavelengths. Nevertheless, when combined with optical or ultraviolet data from the HST, it will be possible, for example, to probe the structures, stellar populations, and interstellar medium in nearby galactic nuclei with unprecedented resolution. An illustration constructed by combining HST plus AO corrected ground-based imaging of the nucleus of M31 is shown in Figure 14.

C. Stellar Mass Functions

Studies of this nature are ideally suited for Gemini, as they require near-diffraction limited image quality and large light-gathering power. The diagnostic measurements, which provide information on the stellar mass function are near-and mid-infrared spectroscopy of lines of H, He, various ionic fine structure lines, and the stellar CO absorption feature.

4. Nuclei of Galaxies

A. Origins of Nuclear Starbursts

The dynamics of the ionized gas throughout the starburst region gives us a way to examine the response of gas to the gravitational potential. To some degree, these gas motions will deviate from that expected for purely gravitational forces primarily because the starburst itself will energize the ISM by its radiation pressure and mass outflows from the stars and their supemova progeny. A completely self consistent picture might be constructed which demonstrates how the gravitational perturbation (stellar bar) causes redistribution of the ISM, its clumping, and the resultant star formation.

One approach to these issues is to determine (1) the spatial distribution of the old stars by near-IR imaging, (JHK); (2) their dynamics by R = 20,000 2D spectroscopy of the 2.3 micron CO overtone (i.e. the CO-band edge's velocity); (3) the distribution of the youngest stars by mid-IR imaging; and (4) the dynamics of the gas by 2-D spectroscopy of mid-IR emission lines, particularly the [NeII] 12.8 micron emission line which is ubiquitous in intensely star-forming regions.

The bulges of some nearby spiral galaxies also display intense star-formation in a ring of "hotspots" originally described by Sersic and Pastoriza in 1968. Such galaxies are generally Hubble type SB or SAB galaxies, with the size of the ring being about 1 kpc. These features likely form when an inner Lindblad resonance stops the flow of gas towards a galactic center. Some of the nearer nuclear rings have been studied in detail, as, for example, in NGC 3310, which shows a conspicuous ring of giant HII regions. The detection of the near-IR CaII triplet in absorption on the nucleus and in one of the HII regions and the Wolf-Rayet feature at 4686Å in the biggest HII region, implies that star formation started in the ring only a few million years ago.

HST images of hot-spot galaxies further reveal the existence of large populations of super star clusters (SSCs). Observations of these SSCs, e.g. those of NGC 2903 shown in Figure 15, offer us the unique opportunity to study massive star formation processes in a dense and high-metallicity environment, which would be physically distinct from normal massive star formation in the moderate-to-low metallicity, low density OB associations in galactic disks.

Studies of these complex starburst regions require a variety of observations. Using 2D optical and near IR spectroscopy, we can determine the chemical abundances of the gas from the emission line fluxes of [OII], [OIII], [SII], [SIII], [NI], [NeIII] etc. at a spectral resolution of R=5,000. Stellar populations can be explored from the equivalent widths of near-IR absorption lines, particularly the near IR CaII triplet and Paschen series of HI, as well as markers of hotter stars such as OI or HeI lines. In addition, infrared-optical colors and the infrared CO lines can allow us to separate red giants from red supergiants. The identification of the Wolf-Rayet stars can be done from the detection of HeII and HeI emission lines in the K band and in the optical from the standard Wolf-Rayet emission lines from C, N and He. It will therefore be possible to obtain high quality maps of the distributions of massive stars in and around the "hot spot" nuclei and to thereby directly test models for the formation of these distinctive structures.

B. The AGN-Starburst Connection

One such issue is the so-called "starbursts-monsters" controversy. In many galaxies where the central regions are dusty, virtually all of the luminosity detected externally is from thermal emission by dust which is heated by an underlying source. This power source could be either a burst of star formation or a "monster" AGN. High angular resolution imaging in the mid-infrared, which showed the infrared source is resolved, would strongly suggest that a starburst is the dominant energy source. Infrared spectroscopic diagnostics can also provide evidence for the presence or absence of a buried AGN and its associated H band lines emission.

There has been growing appreciation of the role of galaxy-galaxy interactions and mergers in stimulating starburst activity, and interactions are also associated with Seyfert (i.e. AGN) activity. There are also simple astrophysical arguments suggesting the starbursts and AGN are causally related and may co-exist in many galaxies, although one of the other may appear dominant. Starbursts can provide the mass flows required to "feed the monster", presumably an accreting high mass black hole, and likewise the outflows driven by an AGN could compress gas and stimulate a burst of star formations. Near- and mid-infrared spectroscopy provide a number of key diagnostics of each type of activity. Measuring the spatial extent of these two groups of diagnostics will reveal the relative contributions of AGN and starburst activity and provide clues to causal relationships. Measurements of the velocity fields to determine such flows will also illuminate this question.

C. Investigations of Massive Black Holes in Galaxy Centers

The enhanced flux collection and angular resolution provided by Gemini offers the opportunity to extend the search for massive black holes to a significantly larger sample of galaxies. This will provide a more comprehensive picture of the incidence of massive black holes and permit generalizations on the implications for extra galactic astrophysics, such as what triggers AGN activity, what quenches it, the likelihood that a given galaxy as been in an AGN phase sometime in its past, etc.

The measurements required are stellar absorption line spectroscopy with a long slit or IFU at the highest angular resolution and Strehl ratio, and imaging to obtain the radial surface brightness profile.

5. Recommendations

• The majority of programs in this area would benefit from enhanced angular resolution. Adaptive optics (AO) systems which are capable of at least moderate Strehl ratio imaging in the near IR are a high priority for both Gemini Telescopes. The very encouraging and impressive results from AO systems which are now operating on the CFHT and NTT have illustrated the gains in image quality that can be achieved at IR wavelengths utilizing this technique. The next phase of Gemini instruments should take advantage of AO, and we therefore place AO systems on both telescopes as a top priority. We also noted the importance of instruments to make use of the potentially significant 'low Strehl' improvements in optical wavelength image quality that may be provided by an AO system.
• High spectral resolution measurements will extend our understanding of the physics of luminous stars via slit spectroscopy or small FOV IFUs with R ~ 25,000 in the North and South. This capability will also address the related problem of the stellar content and evolution of your super star clusters, as well as stellar and emission line properties of galactic nuclei. Since all of these programs target objects with small angular sizes, operation at angular resolutions in the 1/4 arcsec range or better would be a great advantage. The spectroscopic capabilities can likely be achieved by an upgrade to the GMOS to operate at higher spectral resolution.
• Many of the most luminous astrophysical sources are deeply embedded in dust clouds, and therefore radiate their power almost exclusively in the thermal infrared. Mid-IR spectroscopic capabilities are needed in the south to properly study unique classes of embedded objects, ranging from protostars in the Magellanic Clouds to some of the best nearby examples of galactic starbursts. This objective might initially be achieved by sharing of Michelle.
• A proper exploration of compact, dusty astrophysical systems also requires the best-possible angular resolution near-IR imaging. A near IR imager is needed at both Gemini sites to produce properly sampled, moderate Strehl images when the AO system is in operation.
• Equally important is a complimentary near IR IFU/multi-slit spectroscopic capability for small fields at high angular resolutions for both sites.

A multi-object capability for HROS would allow a substantial multiplex advantage for work on specific projects in dense star fields, such as those in the Galactic bulge, Magellanic Clouds, or globular clusters.

D. Formation and Evolution of Galaxies and Cosmologies

2. Galaxy Evolution

A. Clusters of galaxies exhibiting the Sunyaev-Zeldovich effect at z>1

The new radio techniques should produce new samples of rich clusters at z>1, and 4-m telescopes can be used to identify galaxies which are plausible cluster members. An 8-m equipped with a multi-object near-IR spectrometer will be required to measure redshifts and velocity dispersions since the key spectral features ([OII] 3727, 4000Å break, [OIII]5007) will all be redshifted out of the optical passband. The angular sizes of these systems will be at most a few arcmin, and so well suited to the near-IR field of view achievable with the Gemini Telescopes.

These new radio techniques provide an argument for situating the best multi-object near-IR instrumentation on the Mauna Kea Gemini Telescope. The discovery of distant clusters via distant radio galaxies will be pursued most vigorously in the North since it is here that the new generation of deep low-frequency radio surveys are concentrated (6C, 7C, 8C, WENSS, NVSS, FIRST). The discovery of distant clusters via the SZ effect, both decrement work at centimeter wavelengths, and increment work at millimeter wavelengths, will also concentrate in the North. The two groups which are routinely producing SZ detections work with the Ryle Telescope in Cambridge and with the CSO on MK. Although facilities such as COBRAS/SAMBA will survey the whole sky it is important to realize that the positional accuracy required for optical-follow-up will require follow-up CMB observations, and existing (and most planned) facilities are in the North (Cambridge, MK, Hat Creek, the planned Arcmin-Microkelvin Imager on Tenerife).

B. Lensed Galaxies and Lenses

Superb, AO corrected images will be extremely advantageous to disentangle the complicated structure of objects and spectra- good images are required in the visible as well as IR since the spectral features of high z galaxies may be in the blue. The IFU should have ~0.1" pixels extending over a ~5" diameter area to simultaneously get spectra of all components or of arcs (see Figure 17). Spectral resolutions from 500 to 5000 with S/N from 5 to 30 are required, depending on the scientific goals. Visible to near-IR wavelengths may be required depending on the redshifts involved, although for many projects wavelengths from 3650Å to 1.4µm will suffice. The lensing galaxies will typically be V ~ 20-23, the arcs or images will have V ~ 21-26.

Queue scheduling will be very advantageous for time-critical observations and combined with good images may make Gemini unbeatable. Since lenses offer one of the few ways to get quantitative observations of the dark matter (which comprises 95% of our Universe), and measurements of cosmological parameters on truly cosmological scales, this area may provide some of Gemini's most important contributions in the next decade.

C. Galaxies Very Close to Quasar Lines of Sight

If the redshifts of the objects are those of the absorbers, such information will be very valuable to determine the relationship between the absorber and the host galaxy. Moreover, it would be possible to link the information available on the gas (abundances, dynamics, etc from the absorption lines) with those of the stellar populations (ages, histories, etc from colors and spectra), building a picture of the properties of these very high redshift objects.

If the candidates are associated with the QSOs, the effect of its presence on the evolution of close galaxies could be studied, providing information of galaxy formation and evolution in dense environments at early times.

Near-infrared spectroscopy of very faint galaxies very close to bright QSOs at 1<z<4 can be achieved with Gemini. Measurement of Balmer lines (H or H) and/or [OII] will determine redshifts and star formation rates. Integral field spectroscopy in the near infrared (JHK bands) with high spatial resolution (0.2" or better) is needed to explore lines of sight as close as possible to the QSOs (the impact parameter of damped Lyman-alpha absorbers is expected to be very small). Adaptive optics is therefore required. Spectral resolution of 5000 is needed to suppress the OH atmospheric lines post detection. A S/N of 5-10 or so in the emission lines will be enough for redshift determination and to estimate star formation rates. Color information can be obtained by integrating the spectra over wide bands.

3. The Astrophysics of Active Galactic Nuclei

A. The Nuclei of Nearby AGNs

High Strehl (>0.1) near IR imaging and spectroscopy in J-K bands would be required to tackle this problem. J-K colors would constrain stellar populations over central tens of parsecs (10 pc=0.1 arcsec at closest AGN). IFU spectroscopy would map out stellar and gas dynamics. We would need a sample of around 100 AGN. Such a sample is available for Natural Guide Star AO. The Veron Catalog contains galaxies with 250 AGN with total magnitudes less than 12 and north of -40 degrees, so even correcting for the fact that one can only guide on the nucleus, there are sufficient targets. This application needs adaptive optics and is helped by a coronagraphic capability, and an integral field unit with 5 arcsec FOV minimum, with sampling better than or equal to 0.1 arcsec. For this particular science case, the Natural Guide Star system would be sufficient as one can guide on nucleus. Examples of the advantages of 0."1 image quality are illustrated in Figure 18.

Ironically the problem arises in observing a comparison sample as we do not expect to be able to guide on 'normal' spiral nuclei. The comparison sample would need either a good sample of galaxies at distances around 1/10 of that of the AGN, or alternatively a Laser Guide Star system. A spectral resolution greater than or equal to 5000 is needed to measure velocities of ~10 km/s.

B. The geometry and emission mechanism of AGNs

While polarimetry is a powerful diagnostic tool, it requires a large number of photons when applied to extragalactic targets. The gains that can be realized by coupling the Gemini telescopes to efficient polarimeters should be quite spectacular both reducing the time that many observations currently require and making other observations possible for the first time. However to fully realize the potential of polarimetry of extragalactic objects with Gemini will require both multi-wavelength polarimetry capabilities (waveplates in the near-UV, optical, and IR and analyzers for GMOS and the IR instruments) and imaging and spectropolarimetry (low to moderate spectral resolution) capabilities. The wide wavelength coverage is required to allow the comparison of the same intrinsic wavelengths of objects at different redshifts and, most critically, to allow adequate wavelength baseline on individual objects where scattering is playing a role to allow the nature of the scatters (dust or electrons) to be determined. Spectropolarimetry allows more detailed studies of sources. For example, the geometries and relative positions of the various entitling and absorbing regions of AGN can be studied in detail (e.g. broad absorption line quasars). Polarimetry capabilities for the Gemini telescopes will allow the full information content of the radiation from extragalactic sources to be extracted.

4. Galaxies as Probes of Large Scale Structure

A. Determining q_o using Supernovae

At peak brightness, SN can outshine their host galaxies. If galaxies form in a hierarchical scenario from smaller building blocks, SN may be the easiest tracer of the baryonic matter at z>l since they may outshine all the galactic scale objects at high redshift. Since the surface brightness of protodisks or the stellar component of damped Lyman-alpha galaxies is likely to be low, detection of the host galaxies via their SN may be the only viable approach of detecting galaxies in an unbiased manner at high redshift. Such an approach is independent of (1+z)⁴ surface brightness effects.

Even if the integrated magnitudes of SN are fainter than the host galaxy, since they are unresolved objects, AO will be of great benefit. Synoptic scheduling is important to confirm them within 7 days of discovery and then follow the light curves. Searches will benefit from well sampled images at the expense of field of view.

B. Optically Selected Galaxy Clusters at z>1

Observations are required in two steps. First, the clusters need to be found in objective, well defined ways, so that sizeable complete samples are built, and second, the candidate clusters have to be characterised dynamically and the evolutionary properties of their galaxies studied. Deep wide-field, multi-color red and near-infrared imaging can be used to select the cluster candidates. Given the relatively low abundance of these objects, several square degrees need to be surveyed to build large enough samples. Sensible photometric filter combinations will enhance the contrast over field galaxies: passively-evolving stellar populations formed at high redshift (such as those in early-type galaxies, abundant in clusters) would have developed substantial 4000Å breaks by z~1.5, thus red (I-J) colors would be a good discriminator at that redshift. Photometry at longer wavelengths would sample higher redshifts. Observations in multiple photometric bands will give an indication of the cluster redshifts. 4m-class telescopes with wide-field optical and near-infrared cameras are ideal for this phase.

An alternative strategy would be the targeted search for galaxies clustered around known high redshift objects (radio galaxies and QSOs). Very deep narrow band imaging targeting the [OII], [OIII] and Balmer emission lines would be needed over moderate (a few square arc-minutes) fields. The choice of wavelength windows should take advantage of the "gaps" existing in the OH sky emission to significantly reduce the background. Again 4-m telescopes would be ideal for this stage. This method can potentially reach much larger redshifts, since galaxies get bluer with z, so multicolor techniques lose sensitivity until the Lyman-break becomes accessible. It also needs a much smaller survey area. However, broad band, "blind" imaging will provide a less biased determination of the abundance of clusters, thus both strategies should be complementary.

Once the cluster candidates have been found, the role of Gemini becomes paramount. Near-infrared spectroscopy of the cluster galaxy candidates will determine galaxy redshifts and cluster velocity dispersions. Galaxy colors and spectroscopic properties will characterize the evolutionary stage of the galaxies. Spectroscopy provides a crucial diagnostic of recent star formation in galaxies. It may be possible to discriminate starburst, poststarburst, red and normal phases like E/S0s. High resolution (~5Å) spectra of galaxies in pairs, groups and clusters at different redshifts, are needed in order to measure equivalent widths and determine star formation rates of galaxies as a function of z, environment and type. Multi-object near-infrared spectroscopy over a field of view of several square arc-minutes is needed. Resolutions of 5000 or so are required to effectively subtract the OH sky lines, and a S/N of 10-15 per resolution element would be adequate for redshift and velocity dispersion determination and to obtain information on star formation rates and galaxy ages. Good image quality essential for good quality spectra. Galaxy surveys now show that the 1 < z < 3 region is critical to our understanding of galaxy evolution (as it appears to be for QSOs), hence multi-object J band and UV spectroscopy over as large a field as possible (>5') is vital to identify and study the characteristics of the galaxy population at these redshifts.

Redshifts of galaxies with z < 1 can be determined from regular "CCD" spectra, but the dearth of features between 160-370nm means that near-IR spectra are essential for z > 1. Recent work by Steidel and others indicates that galaxies at z > 2 typically have high rates of star formation, so that redshifts of these galaxies may be measured in the ultraviolet (Lyot and the "OB lines" become visible at z =2 at 365nm). This means that it is only the 1 < z < 2 region that is problematical but, fortunately, enough spectral features (370-450nm at a minimum) are available in the J band where multi-slit spectroscopy is relatively easy.

An important aspect of this work which has a bearing on the instrumentation plans is the interpretation of emission line strengths in terms of star formation rate (SFR). The role environment plays in the star formation cycle in galaxies is still unclear, and it is yet unknown whether there are links between the history of the SFR in galaxies and the evolution of the luminosity function (LF) for different galaxy types. In most cases the strongest emission line from star-forming galaxies is H, ([OIII] is sometimes stronger, but its interpretation in terms of SFR is highly uncertain, depending critically on assumptions about the physical conditions). Moreover H provides the best measurement of current SFR ([OII] is affected by excitation/metallicity: H is affected by stellar absorption; UV fluxes are strongly affected by uncertainties in stellar population synthesis modeling, and reddening becomes a much more important issue at shorter wavelengths). Having access to H at z's as high as possible is of great importance for studies of SFR history. The GMOS upgrade would only reach z=1.5, probably missing the peak of the SFR history of the Universe (z=2?). Having access to the K band MOS would cover that all important era. Careful consideration of the resolution required for effective post-detection OH suppression will enable the best use to be made of the IR array pixels that could be made available in a GMOS upgrade.

Eventually, "cooled-spectrograph" multi-object IR spectroscopy will be required to get longer wavelength coverage on targets of particular interest, particularly when H moves into the K-band, ie. for z>1.8. The excellent images delivered by Gemini will give a substantial gain, especially in the K since the half-light radii of these high redshift galaxies are ~0.3".

5. Quasars and the High Z Universe

A. Temperature, Velocity & Abundance Measurements

The gas temperatures in the interstellar medium in the Galaxy are typically up to about 10,000K. The low-ionization heavy element absorption systems found in the spectra of high redshift quasars are likely to arise in galactic disks at high redshift, and if this is the case similar temperatures would be expected. For thermally broadened absorption lines the full-width-half-maximum is 21.4 (T/10,000A)^1/2 km/s, where T is the gas temperature and A the atomic weight of the ion doing the absorbing. Thus, for SiII at 10,000K, FWHM =5.3 km/s, and FeII FWHM is 2.9 km/s. To unscramble this sort of temperature from any turbulent motions requires us to resolve the lines at about the level of the narrower ones, so R=100,000 or more is indicated. If these temperatures can be determined they provide a direct comparison with local conditions, and constraints on the heating/cooling at high redshifts. Knowledge of the true line widths also helps firm up abundance estimates.

Keck HIRES has been used at R=50,000 on quasars, and at high S/N one might hope to see a marginal broadening of the lighter element absorption lines. However, multiple velocity component fits to what are sometimes quite complex profiles so far suggest only that the temperatures could be anything less than about 20,000K, with a significant turbulent component. Apparent turbulence may instead be unresolved velocity structure. One cannot claim that higher spectral resolution will result in the systems resolving further into component clouds -- we don't know. However, each previous significant improvement in resolution has revealed more velocity components, and we would expect this to continue until the thermal widths are reached.

An example can be based on a 6 km/s resolution Keck spectrum of the redshift z=4.11 quasar 0000-26 seen in Figure 20.

The case for spectral resolution R=500,000 for quasar absorption has a common thread with that for the R=120,000, to resolve the lines and so measure temperatures and abundances. There is an expectation that there will be cold clouds at high redshifts because stars must have formed then, and there are already indicators in one or two cases of cold material in quasar absorption systems. The presence of molecular hydrogen at redshift z=2.8 towards 0528-250 with inferred spin temperatures of 175K and higher, and the presence of 21cm absorption in some objects with widths of as little as 4km/s, point towards temperatures of a few 100K. Thus we need resolutions in the half - one million range if we wish to attempt to resolve the Fe lines in those systems. It is possible for significant amounts of material to be missed if the velocity widths are low compared with the resolution, and so the abundances inferred are lower than reality. If we are trying to trace the chemical evolution in galaxies through quasar absorption lines it is important to establish whether or not we are being misled by low resolution spectroscopic investigations.

Figures 21 and 22 show how this can happen. The first shows a 6 km/s resolution Keck spectrum of the SiII 1808 line at redshift z=1.776 towards 1331+170 (provided by Art Wolfe), with a best fit to the data including a 300K component. In this example the column density in the narrow component is about half of the total. The second is the R=500,000 simulated spectrum at SN=25 per resolution element, along with the fitted profile. A fit to the Keck data with a broader component instead of the narrow one yields a column density for that component about a factor 15 lower.

6. Recommendations

A. New Capabilities

1. The highest priority capability is the provision of multi-object spectroscopy (MOS) at the JHK wavelengths. The need is for MOS over the full IR field of the telescope. The spectrograph would require R=5000 to enable the OH lines to be effectively suppressed post detection (although the exact resolution needed for effective suppression may be lower). The provision of an IFU was seen as essential, more than one IFU may be required to work in all three bands. An IFU with a FOV of 5-10" and 0.05-0.1" sampling was identified as being required for studies in crowded regions and of galactic nuclei. This capability supports a wide range of scientific applications, many of which follow the themes of GMOS science to higher redshift or probe dusty environments. For example: studies of lenses and lensed objects, AGN and normal galaxy nuclei, the properties of groups and clusters of galaxies as f(z) and the study of galaxies close to QSO lines of sight. It was felt that this instrument would have wide applicability and get substantial use at both sites and is therefore required in both hemispheres. The scientific link with deep radio surveys suggest that the initial implementation should be in the north.

2. Laser Guide Stars. Superb Adaptive Optics corrected images in the near IR bands and extending into the visible, and feeding both imagers and spectrographs, will be required to disentangle the complicated structure of objects and spectra of lensed extragalactic systems and galaxies very close to Quasar lines-of sight. Moderate Strehl ratios >0.3 over most of the sky is needed; a Laser Beacon system will be required to achieve this sky coverage.

   AO corrected Imaging- and spectroscopic observations will be powerful diagnostics of the emission mechanism(s) and geometry of Active Galactic Nuclei. Good AO correction is required in the J-K bands. A Natural Guide star system will be adequate for the brighter AGN's, while an LBS will be required for the fainter objects and for the comparison sample of 'normal' galactic nuclei.

3. High resolution spectroscopy in the near infrared (JHK). For investigating the interstellar medium abundances in the earliest galaxies, HROS capabilities (R=50,000 and R=120,000) in the infrared are desirable.

   Such a capability would also provide a uniform way of investigating FeII/MgII velocity structure, ratio and strength evolution over a wide redshift range, since the MgII lines are accessible from the ground with HROS for z>0.15, and Fell starts to become available at z=3.25.

   Ultimately this capability will be needed over a much larger wavelength range than can be provided simultaneously by Phoenix however the group felt that Phoenix would offer an adequate capability for this work in the short term and supported the projects efforts to provide this. The provision of a new instrument for this work that would provide extended wavelength coverage in a single exposure should appear later in the program.

B. Upgrades

1. AO Polarimetric analyzers for both GMOS's and for the IR instruments. This capability was widely supported as a technique for studying the geometry and emission mechanism in AGNs. This capability should be provided as soon as possible.

2. The R=120,000 camera for HROS is required to disentangle the effects of temperature and velocity in absorbing clouds along the line of sight to quasars.

3. Grating Improvements: There is a broad need to obtain identification spectra in both the optical and IR of VERY FAINT sources that have been discovered in either very deep images (e.g. the Hubble Deep Field) or identified at other wavelengths (e.g. ROSAT, x-ray, or FIRST, radio, surveys). GMOS and the other instruments planned for Gemini are very efficient, however the largest single source of loss in GMOS is the grating efficiency (at 65-70%). Possible high efficiency innovative grating developments would push the capabilities of GMOS to the faintest possible limits.

4. Upgrade GMOS to work in J and part of H. This was considered as a way of providing near term capability for near infrared, full field, MOS, which was the group's highest priority capability overall. Many of the science applications can be addressed with a J-band IFU on GMOS. A significant constraint on the upgrade path is the requirement to retain GMOS optical spectroscopy and imaging at both sites. A number of important and detailed technical questions, such as the exact cut-off wavelength in H (where the thermal background makes uncooled masks uncompetitive) needs to be evaluated before an assessment of the utility of the GMOS upgrade can be assessed compared to a new near IR MOS instrument. The relative timing of the two such developments was identified as a strong scientific driver; if a GMOS upgrade can provide Gemini with a IR MOS capability before it is available at other observatories the group would put this at a very high priority.