The infrared colours of the RV Tauri stars fall in a well-defined and thinly-populated region of the IRAS [12] – [25], [25] – [60] diagram. This provides an alternative method of searching for such stars. Stars with these IRAS colours as well as the K and L excesses which are also typical of RV Tauri stars have spectral types from B to M. The characteristic variability of RV Tauri-type has been found for several stars with spectral types typical of known RV Tauri stars, while stars of different spectral types do not vary in this way. RV Tauri stars are the stars with these colours, probably indicative of a dusty circumstellar disk, which lie within the Cepheid instability strip.

A review of the main results from the MACHO project for RV Tauri stars and Type II Cepheids in the LMC is presented. In light and colour curve behaviour, the RV Tauri stars appear to be a direct extension of the Type II Cepheids to longer periods. A single period–luminosity–colour relationship is seen to describe both the Type II Cepheids and RV Tauri stars in the LMC.

Follow-up low resolution spectroscopy of the LMC RV Tauri stars has been undertaken. The spectra of the longest period RV Tauri star in the MACHO sample show that this star has strong C2 bands with a pronounced phase dependence, their strength being greater at minimum light. This and other spectral features are similar to those observed in proto-planetary nebulae. This LMC star therefore appears to combine the main spectral anomaly of the proto-planetary nebulae with RV Tauri-like photometric variability. This is an intriguing result which indicates a possible link between RV Tauri stars and proto-planetary nebulae, and strengthens the post-AGB interpretation of the RV Tauri stars.

This review focuses on the evolution of single and binary optically bright post-AGB stars and the relation between the photospheric chemical abundances and the circumstellar material.

Surface abundance determinations of optically bright post-AGB stars offer a powerful tool, not only to clarify the evolutionary nature of the objects themselves, but also to constrain theoretical calculations for chemical evolution. However, a survey of recent results shows that the chemical patterns observed are very diverse and that several different classes can be distinguished. The most striking result is that only a minor fraction, namely the 21 μm objects, are conform to standard post third dredge-up theory.

In recent years it became clear that binarity can influence the late evolutionary stages considerably. A second part therefore discusses the varied and fundamental impact of the presence of a companion on the evolution of post-AGB systems.

On the basis of C/O ratios the evolutionary connection between C-rich AGB/post-AGB stars and C-rich central stars of PNe is discussed. The proportion of carbon-rich post-AGB stars is comparable to that of carbon-rich PNe. The post-AGB stars with 21μm emission are found to be over abundant in carbon and s-process elements indicating that they have gone through the third dredge-up and carbon star phase on the AGB. As these stars evolve to higher temperature they will form carbon-rich PNe with carbon-rich central stars. Analysis of C/O ratios of PNe suggests that 68 percent of PNe have C/O > 1 indicating that the progenitors have gone through the third dredge-up and carbon-rich phase on the AGB. The Type I PNe can be divided into two groups. Group I Type I PNe are those with C/O < 1 indicating that their massive progenitors have gone through the third dredge-up and hot bottom burning on the AGB. Group II Type I PNe are those with C/O > 1. Their progenitors have gone through third dredge-up only and may not have experienced hot bottom burning. Carbon stars with silicate dust and or carbon-rich post-AGB stars (such as post-AGB stars with 21μm emission) may be the progenitors of PNe with [WC] central stars. The evolutionary sequence seems to be: carbon-rich AGB → carbon-rich post-AGB → [WC late type] → [WC early type] → [WC-PG 1159] → PG 1159.

The born-again giant Sakurai's object is currently experiencing its second stage as an AGB-star. Furthermore, Sakurai's object has shown an unprecedented rapid stellar evolution since discovery in 1996, both in terms of a continued cooling of the photosphere and spectacular changes in chemical composition on a time-scale of a mere few months. The surface cooling and abundance alterations of H, Li and the s-elements are the direct consequences of the expansion, mixing and nucleosynthesis which has ensued as a result of a final He-shell flash occurring while the star was descending the white dwarf cooling track. Sakurai's object shows striking similarities with the R CrB stars in both chemical composition and visual variability.

Distinctive features of dust condensation events in the circumstellar shells of the born-again AGB stars FG Sge and V4334 Sgr are described together with updated optical, IR and bolometric light curves.

We discuss the origin of hot hydrogen-deficient post-AGB stars. Evidence is growing that the “born-again AGB star scenario” as a consequence of a late He shell flash cannot explain the diverse observations. Analyses of the stars and their associated planetary nebulae suggest that the H-deficiency is not formed during the second post-AGB phase but already while the star is still on the AGB. There are even hints that the H-deficiency can occur before the first departure from the AGB.

AGB stars, seen as a stellar population, can be used to probe the dynamical state of galaxies. The relevant data are mostly positions and line-of-sight velocities, sometimes together with information on chemical composition and/or age. As of now, dynamical models have been made for OH/IR stars and Planetary Nebulae. Other candidates are C stars, S stars, and Miras. We review the methods used and the results obtained so far, for the Milky Way and for (relatively nearby) extragalactic stellar systems.

First results on AGB stars with the ISOGAL mid-infrared survey are presented. We show that M6 and later spectral type giants are detected at 7 and 15 μm up to the distance of the Galactic Centre. Analysis of the observed magnitudes and colours indicates that most of the detected AGB stars in the Galactic Bulge are losing mass. ISOCAM CVF spectra on a field in the Bulge confirm the mass loss and also show a large diversity in the shape of the silicate feature.

The properties of the carbon and oxygen chemical balance in planetary nebulae (PNe) are analysed through mid infrared spectroscopy of warm dust emission features in a sample of 72 objects. The adoption of a statistical distance scale shows that the galactic disk distribution of warm dust types in PNe is rather homogeneous with height over the galactic plane, and that there is a trend for the proportion of PNe with O rich dust signatures to decrease with galactocentric radius. Models of the galactic distribution of PNe compositions require a minimum progenitor mass of 1.2M⊙, although the observational constraints suffer from the smallness of the sample. This initial investigation is however an incentive to pursue the use of warm dust emission in PNe to study their progenitor population in various galactic environments.

A second report on our 4-year survey for long-period variables (LPVs) in the central 24 × 24 arcmin2 of the Galactic Bulge is presented. All four seasons of data have now been reduced and ∼ 350 variables have been found. Preliminary periods have been obtained for most of them. Fifty-five out of 103 known OH/IR stars in the field have been recovered in the K-band.

We are measuring Fe abundances of cool, luminous stars within 30 pc of the center of the Milky Way. Our sample contains both AGB stars and M supergiants. Low-resolution (λ/Δλ = 500) H and K band spectra are used to estimate temperatures and gravities. Stellar Fe abundances are determined from high-resolution (λ/Δλ = 40 000) K band spectra obtained on the IRTF using CSHELL. We find that Fe abundances of stars in the Galactic Center are consistent with the solar Fe abundance.

Our current knowledge of carbon stars in the Local Group and beyond, is discussed. Although many carbon stars and late M-stars have been identified in external galaxies, a coherent understanding in terms of the chemical evolution- and star-formation-rate history of a galaxy is still largely lacking. Issues that need to be addressed are: 1) for some of the larger galaxies only a small fraction in area has been surveyed so far, 2) surveys have been conducted using different techniques, and may be incomplete in bolometric magnitude, 3) only for some galaxies is there information about the late M-star population, 4) not all galaxies in the Local Group have been surveyed, 5) only for a sub-set of stars are bolometric magnitudes available.

From the existing observations one can derive the following: the formation of carbon stars is both a function of metallicity and star formation. In galaxies with a similar star-formation-rate history, there will be relatively more carbon stars formed in the system with the lower metallicity. On the other hand, the scarcity of AGB-type carbon stars in some systems with the lowest metallicity indicates that these galaxies have had a low, if any, star-formation-rate history over the last few Gyrs.

The Bulge carbon stars have been a mystery since their discovery by Azzopardi et al. (1991), because they are about too faint to be regarded as genuine AGB stars, if located inside the metal-rich Bulge . Part of the mystery can be solved if these carbon stars are related to the Sagittarius dwarf galaxy . They are in that case not old and metal-rich, but young, ∼0.1 Gyr, with SMC-like metallicity (Ng 1998).

The σRV = 113 ± 14 km s−1 (Tyson & Rich 1991) radial velocity dispersion of the stars appears to be consistent with Bulge membership. On the other hand, a similar velocity dispersion could be the result from an induced star formation event when the SDG crosses the galactic midplane. It is suggested that the carbon stars are tracers of such an event and that they therefore are located at distances related to the SDG. However, the majority of the carbon stars are not member of the SDG, nor are they similar to the C-stars which are member of the SDG.

The radial velocities can be used to determine a possible membership to the SDG. However, they do not give information about the distance of the stars. In particular, if the stars are located at a distance comparable to the SDG. This implies that only the period-luminosity relation (Groenewegen & Whitelock 1996) can be used to distinguish unambiguously if the carbon stars are located at Bulge-like or SDG-like distances. Thus far only carbon stars with reliable periods have been identified at a SDG related distance (Ng & Schultheis 1997; Whitelock 1998).

The effect of metallicity on the AGB mass loss is reviewed. Observations have mainly been limited to the Magellanic Clouds but are observationally feasible throughout the Local Group. Expansion velocities are predicted to depend strongly on Z but the peak mass-loss rates appear not to. The Mira PL relation shows no evidence for a Z dependence, giving a powerful potential distance indicator. I derive a distance modulus to the LMC of m − M = 18.63 ± 0.09 as well as a bright calibration for the Horizontal Branch in globular clusters. Finally, the predicted initial-final mass relation at low Z is shown to give rise to higher mass remnants. This would result in an increased supernova rate in young, low-Z populations, such as found in the early Universe.

We surveyed 0.5 square degrees in the Bar of the LMC with ISOCAM at 4.5 and 12 μm, and with DENIS in the I, J, and Ks bands. Our goal was to build a complete sample of Thermally-Pulsing AGB stars. Here we present the first analysis of 0.14 square degrees. In total we find about 300 TP-AGB stars. Among these TP-AGB stars, 9% are obscured AGB stars (high mass-loss rates); 9 of them were detected by IRAS, and only 1 was previously identified. Their luminosities range from 2 500 to 14 000 L⊙, with a distribution very similar to the one of optical TP-AGB stars (i.e. those with low mass-loss rates). Such a luminosity distribution, as well as the percentage of obscured stars among TP-AGB stars, is in very good agreement with the evolutionary models of Vassiliadis & Wood (1993) if most of the TP-AGB stars that we find have initial masses smaller than 1.5 to 2 M⊙.

We report on some recent advances in the study and understanding of heavily obscured AGB stars in the Magellanic Clouds.

We performed systematic infrared observations of the intermediate-age Magellanic Clouds clusters NGC 419, NGC 1783 and NGC 1978. Mid-infrared stars discovered in NGC 419 and NGC 1978 are very red and must be undergoing intense mass loss (comparable to superwinds). They are probably carbon stars but do not seem to show any FIR excesses. Three optically visible carbon stars as well as (at least) 2 near-infrared carbon stars observed with ISOPHOT show 60 μm excesses which may indicate mass loss in the past. It seems that the MIR stars are fainter than the AGB tip luminosity and that their Mbols are close to those of the transition luminosity from M type to C stars. Therefore, these MIR stars may not be in the final stage of the AGB phase. This may suggest that AGB stars lose mass heavily at some other time, possibly during the transition from M type to C stars.

Asymptotic giant branch stars are essential contributors to the near and mid-IR emission of intermediate age (108-109 yr old) stellar populations. Detecting this light will set constraints on the star formation history in galaxies and, conversely, the search for AGB signatures in well studied populations will help us reduce some of the still large uncertainties in AGB models. This paper reviews how AGB stars are currently included in population synthesis models and which spectral features can be used to identify their emission in galaxy light; targets for observational tests are suggested, and some observational and theoretical difficulties are discussed.

My assignment was to listen to all of the talks at the symposium and to discover and summarize results that have appeared since the last conference on AGB stars. Since the entire meeting was devoted to new results, discovery was a simple task. Because of the large volume of new results, an adequate summary is impossible. I therefore do not attempt to provide a comprehensive survey, but confine my remarks to a small subset of the discoveries and advances which have enriched and modified our understanding of AGB stars over the past decade.