Oxygen abundances have been derived in a sample of very metal-poor stars using the O I triplet at λλ 7771-5 Å and OH lines in the near UV. A detailed NLTE analysis of iron lines has been carried out for one of the observed stars, BD +23° 3130, providing consistent values of effective temperature and surface gravity that are in very good agreement with independent estimates from the infrared flux method and Hipparcos parallaxes, respectively. These parameters, especially the higher gravity obtained with respect to previous analyses, reduce the discrepancies claimed between the oxygen abundances determined from OH, O I triplet and [O I] λ 6300 Å lines, and give consistent abundances to within 0.16 dex for BD +23° 3130 ([Fe/H]NLTE = −2.43). The oxygen abundances derived for this new sample confirm previous findings for a progressive linear increase in the oxygen-to-iron ratio with a slope −0.33 ± 0.02 (including NLTE corrections to the iron abundances for all the stars considered) from solar metallicity to [Fe/H]∼ −3, and [O/Fe] values as high as ∼ 1.1 for stars with [Fe/H]≾ −2.5. These results can be interpreted as evidence for oxygen overproduction in the very early epoch of the formation of the Galactic halo, possibly associated with supernova events with very massive progenitor stars.

We have made observations of 24 unevolved stars of the OH lines in the ultraviolet spectral region at high spectral resolution and high signal-to-noise, typically 60 − 110 with Keck + HIRES. The O abundances have been computed by spectrum synthesis. For these cool, unevolved stars there is a linear relation between [O/H] and [Fe/H] over three orders of magnitude with very little scatter and a slope of +0.66 ±0.02. The relation between [O/Fe] and [Fe/H] is robustly linear with [O/Fe] = −0.35 (±0.03) [Fe/H]. There is no sign of a break at metallicities between −1.0 and −2.0.

Oxygen abundances are derived from IR-OH lines and compared to published results from UV-OH lines.

The effect of temperature inhomogeneities on the formation of two oxygen lines and one Fe II line has been investigated based on 2-D radiation hydrodynamics simulations of convection in the atmospheres of metal-poor stars. It was expected that the IR O I triplet lines would be enhanced by the temperature inhomogeneities, more than the forbidden [O I] line or Fe II lines. The actual computations done on two snapshots under the assumption of LTE have not confirmed this expectation, instead suggesting that the main difference between ‘classical’ and ‘dynamical’ atmospheres of metal-poor dwarfs is the significantly cooler mean temperature of the dynamical upper photosphere (τRoss ≾ 0.01).

The sensitivity of various methods of oxygen-abundance determination to the structure of the model atmosphere in a classical analysis is described, with emphasis on the treatment of convection. Oxygen-abundance results presented during this meeting are outlined. Because the oxygen diagnostics are sensitive in different ways to the choice of model stellar parameters and to the treatment of convection, it is not surprising that different methods applied by different groups give different results. Most are consistent with an increase in [O/Fe] as [Fe/H] declines, but the slope is uncertain still, and the scatter is large at very low abundances. The trend may continue to low [Fe/H], but more stars and more checks are needed.

The NLTE effects affecting oxygen-abundance determinations of solar-type stars are discussed. LTE is safe for the forbidden lines. The permitted triplet at 777 nm is expected to show NLTE effects so that assuming LTE overestimates the abundance, but the magnitude of the effects is dependent on the poorly known cross sections of collisional excitation by collisions with neutral hydrogen atoms. Little is known about the NLTE effects on molecular line formation.

The formation of [Oɪ], Oɪ and OH lines in metal-poor stars has been studied by means of 3D hydrodynamical model atmospheres. For Oɪ detailed 3D non-LTE calculations have been performed. While the influence of 3D model atmospheres is minor for [Oɪ] and Oɪ lines, the very low temperatures encountered at low metallicities have a drastic impact on OH. As a result, the derived O abundances are found to be systematically overestimated in 1D analyses, casting doubts on the recent claims for a monotonic increase in [O/Fe] towards lower metallicities.

We discuss the evolution of the [O/Fe] as a function of [Fe/H] in the Galaxy. In particular, we discuss the influence of different stellar yields as well as different SN (II, I) rates on the [O/Fe] ratio. Possible explanations for the steep rise of the [O/Fe] ratio at low metallicities, as indicated by recent observations, are critically discussed.

Stellar evolutionary models for [Fe/H] = −2.27 have been constructed to examine how predicted C-M diagrams, age versus turnoff-luminosity relations, the luminosities of the red-giant “bump” and tip, and the Teff/color of a zero-age horizontal branch model of fixed mass are affected by varying the assumed abundance of oxygen (with and without enhancements in the other α-elements).

Oxygen is a much better evolutionary index than iron to follow the history of Lithium-Beryllium-Boron (LiBeB) since it is the main producer of these light elements at least in the early Galaxy. The O-Fe relation is crucial to the determination of the exact physical process responsible for the LiBeB production. Calculated nucleosynthetic yields of massive stars, estimates of the energy cost of Be production, and above all recent observations reported in this meeting seem to favor a mechanism in which fast nuclei enriched in He, C and O arising from supernovae are accelerated in superbubbles and fragment on H and He in the interstellar medium.

The current dispute over the oxygen abundance in metal-poor stars has been viewed from a distance by this observer who would claim a measure of independence despite an intermittent interest in ox-gen abundance determinations over 3 decades. This summary attempts to persuade all participants whether they advocate a ‘high’ or ‘low’ oxygen abundance or are simply bemused by the dispute that in resolving the present disagreements we shall learn not only about the oxygen abundance of these oldest Galactic stars but shall achieve a deeper understanding of the atmospheres of the stars.

The number of spectroscopically confirmed galaxies at high redshift is increasing rapidly, with many being found efficiently by deep optical imaging and color selection. In parallel, a confluence of technical developments is bringing rapid progress to the domain of observations of cold gas and dust at high redshift. Large telescopes operating at high, dry sites, with a new generation of sensitive detectors, together with recent satellite missions, are opening up new areas of study through observations of dust continuum emission and associated atomic and molecular emission and absorption lines. These data bear directly on fundamental questions of cosmic evolution by probing the ordinary cool material that forms stars and fuels active galactic nuclei. Analysis of data from the COBE satellite confirms the presence of a diffuse far-infrared background from a widespread population of distant dusty objects. The global energetics of the optical and far-infrared backgrounds suggest that perhaps half of distant activity may be enshrouded by dust. Understanding the nature and redshifts of the sources responsible for these emissions is profoundly important. The intent of Joint Discussion 9, between Division X (Radio Astronomy) and Division XIII (Galaxies and the Universe), was to provide a forum to present observations from this newly accessible realm and to consider the astrophysical implications.

This article reviews three related topics: the extragalactic background light and its sources, evolution models for the dust absorption and emission in galaxies, and empirical constraints on these transfer processes in nearby starburst galaxies. It is intended that the material presented here will serve as an introduction to this Joint Discussion.

I will first review the observational evidence relating dust emission and the energy production in the far-IR/submm range. This latter contains crucial information on the global baryons transformation and the related stellar activity in the Universe. Present knowledge on this topic relies mainly on the far-IR local surveys of IRAS and ISO missions and on submm/mm surveys performed with SCUBA and MAMBO arrays. Further constraints are provided by the measurements of the Cosmic far-IR Background (CFIRB). Our scanty knowledge of galaxy formation and evolution is mainly caused by the difficulties of unveiling stellar activity at redshifts larger than 1 and at present we may only have detected massive objects in a transient hyper-luminous phase. We still lack an unbiased census of the much more numerous population of lower luminosity dusty objects. It will soon be possible to disclose the entire history of evolving dusty objects, and therefore of the stellar activity, selecting unbiased samples out of far-IR imaging and photometry in deep far-IR surveys.

After the production of the ionizing background by the first generation of stars, neutral gas must be confined to sufficiently high density to be self-shielding and remain neutral. Neutral gas is an identifier of the presence of confining gravitational potentials and a tracer of the kinematics of the potential. Kinematical studies are being extended to neutral atomic gas at high redshift.

Deep submillimeter (submm) surveys offer an unobscured view of dust-enshrouded star formation or AGN activity at high red-shifts. SCUBA observations above 2 mJy have resolved 20 – 30% of the far-infrared (FIR) background into discrete sources and have revealed the existence of a distant population of galaxies with properties similar to those of local ultraluminous infrared galaxies. A large fraction of the submm sources have extremely faint optical/near-infrared (NIR) counterparts and hence are inaccessible to optical spectroscopy. Millimetric redshift estimation places the submm population at z = 1 to 3. While the cumulative surface density of the submm sources is low, they are so luminous that if powered mainly by star formation, they dominate the high redshift star formation history. Recent combined SCUBA submm and Chandra hard X-ray studies suggest that the majority of the submm sources are star formers with only a small admixture of obscured AGN.

We discuss results from sensitive, wide-field imaging of the millimeter extragalactic background using the Max-Planck Millimeter Bolometer array (MAMBO) at the IRAM 30 m telescope.

High-z mm and submm dust sources are thought to be similar to ultraluminous infrared galaxies, where merger interactions have driven most of the gas and dust into circumnuclear molecular disks or rings with typical radii of 500 pc. Within these disks, CO and radio data indicate a new class of extreme starbursts, with characteristic sizes of 100 pc, gas masses of 109 M⊙ and IR luminosities of ∼ 3 × 1011L⊙ from OB stars. Some of these regions, however, harbor powerful AGNs, even though the optical IR diagnostics may classify them as starbursts. Trentham (2000) argues that averaged over time, most of the high-z dust sources must be powered by AGNs.

We review observations of molecular absorption line systems at high redshift toward red quasars and gravitational lenses.

The steep slope of the ISOCAM 15 μm number counts indicates that infrared galaxies have strongly evolved since z ∼ 1 − 1.5. A nearly complete spectroscopic sample of ISOCAM galaxies in the Hubble Deep Field North (HDFN) shows that their redshift distribution is peaked around z ∼ 0.8. We show that the 7 μm luminosity of local galaxies is correlated with their 8-1000 μm luminosity, and therefore star formation rate (SFR). We use this correlation in the rest-frame of the ISOCAM galaxies to deduce their IR luminosities (∼ 4 × 1011L⊙), SFR (∼ 80 M⊙yr−1) and contribution to the peak of the cosmic IR background (CIRB) at 140 μm. We find that they most probably produce the bulk of the CIRB.