In this paper I discuss our basic picture of the structure of AGB stars, and how their earlier evolution led them to this situation.

We present the results of the First and Second Abastumani Objective Prism Spectral Surveys along with other surveys for carbon stars carried out in recent years.

We are attempting to investigate the space distribution of carbon stars in our Galaxy by evaluating the distances of individual carbon stars. We estimate the distance by evaluating the total radiant energy of each star and assuming a value for the bolometric magnitude of carbon stars. Since the flux maxima of carbon stars are mostly in the near-infrared (NIR), NIR photometric data are most useful for evaluating the total energy. We have compiled NIR photometric data for 694 carbon stars at all galactic longitudes and have made new observations of 470 carbon stars in the galactic longitude region between 20° and 160°. We discuss the space distribution of 1164 galactic carbon stars.

Most stars in our Galaxy with photospheric C/O > 1 (carbon stars) are not giants but dwarfs. The newly-recognized class of dwarf carbon stars joins the growing family of stars with peculiar abundances that are now recognized as products of mass-transfer binary (MTB) evolution. The dozen examples now known span a wide range of evolutionary histories, ages, and abundances. These stars can already provide some much-needed constraints on the formation of AGB C stars in the disk and spheroid populations, and on the parameters characterizing binary evolution there. A larger sample, with some bright members, would hasten our progress.

We report on an effort to identify further members of the dwarf carbon class and new members of a putative dwarf S group in the general field through determination of proper motions of catalogued stars. Examination of nearly 1500 C stars and over 300 S stars reveals some interesting false alarms but no new dwarf members of these classes.

We discuss new results based on RI and JHK photometry for 888 and 204 carbon stars, respectively, of the 1035 C stars found by Blanco and his collaborators in the Large Magellanic Cloud (LMC). Bolometric magnitudes and effective temperatures for these stars are calculated and compared with theoretical predictions. We find a spatial gradient in the transition luminosity between M and C type stars. This has implications for the age of the most recent major epoch of star formation in the LMC.

Prominent molecular absorption bands in the spectra of carbon stars make their detection possible, even in relatively distant external galaxies. Although extensive surveys for carbon stars have been carried out mainly in the Magellanic Clouds and the dwarf spheroidals in the Galactic halo, more distant galaxies in the Local Group and beyond have been successfully searched for this kind of object.

We have used a four-band photometric system capable of distinguishing C, S, and M stars to undertake a survey of AGB stars in M31. We discuss the results from this survey and from follow-up spectroscopy.

I review the current status of model atmospheres for red giants, with special emphasis on recent progress and newer grids. I draw attention to some specific problems regarding opacity sources and present current and forthcoming efforts in cool-star atmospheric modeling.

We present the results of a quantitative analysis of 12C/13C ratios in 62 N-type, 15 SC-type, and 26 J-type carbon stars. The 12C/13C ratios are determined from lines of the CN Red System around 8000 Å, based on the iso-intensity method and line-blanketed model atmospheres. The resulting 12C/13C ratios in N- and SC-type carbon stars are consistent with a scenario in which M giants evolve through SC- to N-type carbon stars, as 12C produced during thermal pulses is added to the envelope.

High-resolution and high signal-to-noise-ratio spectroscopy is used to derive 12C/13C ratios from the Red System of the CN molecule at 8000 Å in several galactic low-mass C stars previously analyzed for lithium. It is found that Li-rich C stars usually have low isotopic ratios (< 15). This suggests that the Li and 13C enrichments in C stars might be produced by a similar mechanism. However, due to the large uncertainties in the abundance derivations for both Li and carbon isotope ratios, this result must be confirmed with further studies.

We report on our investigations in progress on the importance of polyatomic molecules – especially of carbon and hydrocarbon – in the atmospheres of very cool helium-rich white dwarfs. For two sets of abundance ratios, model atmospheres with Teff = 5000 K, log g = 8.0 have been computed and the dependence of C3, C2H and C2H2 on abundance variations of H/He, C/He, O/He and N/He has been calculated in detail, with the absorption of carbon molecules included quantitatively. We discuss the possibility that C2H features can explain the observed IR flux deficiency for LHS 1126 at 1.6 μm and 2.0–2.3 μm.

An abundance analysis of the yellow symbiotic system AG Draconis reveals it to be a metal-poor K giant ([Fe/H] = −1.3) which is enriched in the heavy s-process elements. This star thus provides a link between the symbiotic stars and the binary barium and CH stars which are also s-process enriched. These binary systems, which exhibit overabundances of the heavy elements, owe their abundance peculiarities to mass transfer from thermally-pulsing asymptotic giant branch stars, which have since evolved to become white-dwarf companions of the cool stars we now view as the chemically peculiar primaries. A comparison of the heavy-element abundance distribution in AG Dra with theoretical nucleosynthesis calculations shows that the s-process is defined by a relatively large neutron exposure (τ = 1.3 mb−1), while an analysis of the rubidium abundance suggests that the s-process occurred at a neutron density of about 2 × 108 cm−3. The derived spectroscopic orbit of AG Dra is similar to the orbits of barium and CH stars. Because the luminosity function of low-metallicity K giants is skewed towards higher luminosities by about 2 magnitudes relative to solar-metallicity giants, it is argued that the lower metallicity K giants have larger mass-loss rates. It is this larger mass-loss rate that drives the symbiotic phenomena in AG Dra and we suggest that the other yellow symbiotic stars are probably low-metallicity objects as well.

Most oxygen-rich late-type giant stars show evidence for chromospheres in their visual spectra (e.g. Ca ii H & K emission features). Cool (i.e. N-type) non-Mira carbon stars, however, have never been observed to have chromospheric emission in their Ca ii H & K lines. However, faint Mg ii h & k lines were detected in emission in low-dispersion IUE spectra of the brightest cool carbon stars in the early 1980s. May 1984 saw the first (and only) successful high-dispersion IUE spectrum taken of a cool carbon star, TX Psc (N0; C6,2). Armed with this high-dispersion spectrum, as well as low-dispersion IUE and ground-based spectra, Luttermoser et al. (1989) made the first detailed attempt to semiempirically model the chromosphere of a cool carbon star. This model was successful in reproducing the Mg ii lines, but it was not well constrained due to the lack of other observed high-resolution chromospheric profiles for comparison. Modeling carbon star chromospheres can now be addressed more accurately with HST/GHRS high-resolution spectra. New fluoresced emission features have been discovered in the GHRS spectra of carbon stars that are not present in their oxygen-rich counterparts.

Spectroscopic properties of RV Tau variables are summarized. We report on our detailed spectroscopic investigation of a sample of RV Tau stars covering all three RV Tau subgroups. Though the observed abundance pattern is similar to those of a few post-AGB stars, we find considerable variation in the observed chemical compositions indicating that there is some non-homogeneity among subgroups of RV Tau variables. Possible scenarios of dust-gas separation in the material ejected by the dusty wind are discussed to explain the observed abundances. This mechanism would work better if the star happens to possess a binary companion. Interestingly, one of our sample stars, EP Lyr, shows long-term variations in radial velocity possibly caused by the presence of a companion.