NEWSLETTER OF CHEMICALLY PECULIAR RED GIANT STARS Number 13 December 1992 Edited by Sandra B. Yorka Denison University I. Message from the Working Group Chairman During the Autumn, the Chairs of IAU Working Groups, along with Commission Presidents and Vice-Presidents, received communications from Dr. Jacqueline Bergeron, the IAU General Secretary, informing us of some of the actions taken by the IAU Executive Committee (EC) at its 63rd meeting in September 1992. Although details will soon be distributed to the IAU membership via the Information Bulletin, I would like here to call attention to a few points that seem of most interest to readers of this Newsletter, especially those relating to the future of our Commis- sions and their Working Groups. 1) Just receiving these communications was, I thought, of some interest, since it demonstrates the EC's awareness of the existence of Working Groups, ours in particular. In fact the EC is aware of over 60 WGs, according to the General Secretary. 2) Plans to streamline the Commission structure of the IAU, first discussed in Buenos Aires, are well under way. It seems likely that the present structure of 40 Commissions, with numbers up to 51, will be replaced by a set of approximately 12 Commissions covering broad areas. 3) An example of one of the proposed new Commissions is "Stellar Atmospheres". This would embrace the two Commissions that sponsor our WG (Comm. 29 - Stellar Spectra, and Comm. 45 - Stellar Classsification) as well as Comm. 36 (Theory of Stellar Atmospheres) and perhaps parts of Comm. 25 (Stellar Photometry and Polarimetry) and Comm. 27 (Variable Stars). 4) Working Groups sponsored by the new Commissions would carry out many of their specific tasks and services. They would be fewer in number than the present 60, and they would continue to exist only so long as they had a task to accomplish (their existence would be reviewed at each General Assembly). Whether a WG such as ours will survive this restructuring (the official IAU word seems to be "restructurization") is not clear. If the only task of a WG is to publish its Newsletter, it probably will not. However, it would seem possible for a Newsletter to continue under the sponsorship of a Commission, without the intermediary of a Working Group. With regard to the next General Assembly (The Hague, August 1994): 1) The next GA will be the last one based on the present structure of 40 Commissions. 2) Commissions will hold business sessions, but scientific sessions by individual Commissions will be discouraged if not altogether elimi- nated. 3) There will be 6 symposia associated with the General Assembly. They will be held as two series of 3 parallel symposia, of 4 days each, contained within the two-week period of the General Assembly. 4) There will also be at least 15 Joint Commission Meetings, lasting either one day or half a day. 5) Topics for the symposia and JCMs will be decided at the next meeting of the EC on June 25-28, 1993; proposals must reach the EC by May 1. It seems clear that future General Assemblies, starting with the next one, will be more tightly organized and carefully planned than those we have known in the past. In particular, the scientific con- tent will be channeled into a relatively small number of symposia and JCMs planned well in advance around specific themes. Gone, I would suppose, are the days when one could ask a Commission President for 5 minutes to slip in a report on an interesting new result at the next day's Commission meeting. A fairly broad area such as "Peculiar Red Giant Stars" might not be represented on the GA program at all unless planned as part of a symposium or Joint Commission Meeting. In other words, if we want our interests to be represented at the next General Assembly, we have to plan ahead (i.e., now). I would be interested to hear the thoughts of readers on these matters. --- Robert F. Wing Chairman, WG on Peculiar Red Giant Stars II. Review MASS LOSS AND UNIFIED MODEL ATMOSPHERES Uffe Graae Jorgensen Niels Bohr Institute, Copenhagen, Denmark The bulk of the mass returned by stars into the interstellar medium comes from stars that end their lives as white dwarfs. Such stars have an initial mass, Minit, below a critical mass which is in the range 5 - 8 solar masses (Msun). The observed masses, Mwd, of white dwarfs are narrowly distributed around 0.55 Msun. A simple integration of the birth-rate function times (Minit - Mwd) shows that about 97% of all mass returned by stars per time unit (in a system that, like our own Galaxy, is old enough that the death rate equals the birth rate) comes from these low- and intermediate-mass stars. Only about 3% comes from supernova ejecta and mass loss from high-mass stars. A fundamental question therefore concerns the degree to which this 97% is enriched with elements produced inside these same stars. If the mass loss occurs early in the stellar lifetime, the composition of the mass-return (the yield) will be almost identical to the composition of the original material (the ISM) from which the star was formed. In that case the enormous amount of mass loss is uninteresting, in the sense that it does not contribute to the chemical enrichment of the universe. If, on the other hand, the mass loss occurs selectively very late in the stellar evolution, it may include material mixed to the surface during the helium shell burning epoch. In this case a significant contribution to the chemical enrichment can be expected from the low and intermediate-mass stars, and most of the carbon, nitrogen and lithium, as well as the s-process elements (i.e. 2/3 of all the elements heavier than iron) may then come from these stars, as may selected isotopes of other ele- ments, for example O-17 and Al-26. The answer to this question can be sought in a variety of ways. Observa- tionally, it is well known that many luminous red giant stars are rich in s-process elements, and that these stars do exhibit mass loss. The amount of mass loss is nevertheless very hard to determine observationally. The best general estimate (de Jager et al. 1988) of the mass loss from red giants is still the one given by Reimers' empirical formula from 1975, multiplied by an efficiency factor, eta -- i.e., dm/dt = eta*1.27e-5*L**1.5/M/T**2 in units of Msun/year. Eta is usually believed to have a value around eta = 0.4, though with some scatter around this value. For example, the morphology of the horizontal branch in globular clusters, if explained as a star-to-star variation in eta, requires that 40% of the stars have eta more than 0.1 below and above, respectively, the average value of eta (Jorgensen and Thejll 1992, submitted to A&A). A number of stars are observed to have mass loss values a factor 100 higher than the one given by Reimers' law. If Reimers' mass loss law describes something about the physics of the mass-loss mechanism, this or these physical mechanism(s) seems to be superseded, at some stage of stellar evolution, by another much more efficient mass-loss mechanism, and the high mass loss rate due to this mechanism may be the one responsible for the major part of the stellar yield. It is obvious that a mass loss rate of, say, 10-5 Msun per year needs only operate over a very short part of the stellar life- time in order to be the dominant source of mass outflow. Another problem with an observational determination of mass loss is that it is very difficult, if not impossible, to determine whether the high observed mass loss rate takes place at a relatively constant value or represents a variable process that only over very short periods reaches the exceptionally high values measured. Finally, substantially higher values may be reached that rapidly obscure the atmosphere, so that the star becomes invisible. In order to interpret the observations properly, and to reach a more detailed theoretical understanding of the mechanisms and role of mass loss in red giants, it is necessary to build a model for the mass loss process. Such attempts are well under way, although many obstacles are still left to be overcome. There are basically three types of stellar winds that can carry the mass loss. They are distinguished by the asymptotic velocity, v-inf, of the wind at great distances from the star (e.g. at 10 stellar radii). The energy needed to produce mass loss is determined by the escape velocity at the sur- face of the star, V-esc, and the asymptotic velocity, V-inf. It requires an energy (m*v-esc**2)/2 to lift the material out of the gravitational field of the star, and the kinetic energy of the mass loss at great distances from the star is (m*v-inf**2)/2. Per mass unit, it therefore requires the energy 0.5 * (V-esc**2 + V-inf**2) to blow material out of a star. In hot stars the wind is supposed to be driven by radiation pressure on the gas caused by the interaction between the intense (high-energy) radiation field and strong atomic lines. In such objects V-inf >> V-esc, i.e. the major part of the energy goes into accelerating the wind up to its high speed. In solar-type stars V-inf and V-esc are of the same order, and such winds can be thermally driven. In red giants, V-inf << V-esc. Typically, V-inf = 10 km/s, and V-esc is larger by an order of magnitude or more. Such winds can neither be thermally driven nor driven by radiation pressure on atomic lines. The reason for the failure of the thermal wind is that the surface of the star is so big that the necessary temperature to drive a thermal wind would cause a strong UV excess, which is not observed (the stars are to the right of the coronal line in the HR diagram). The reason for the failure of the atomic-line radiation-pressure wind is that a major fraction of the radiation field is at longer wavelengths than the majority of the atomic lines, and a substantial fraction of the atoms are bound in molecules. The mechanism that is required in red giants must be one that is capable of "gently" accelerating the wind up to a velocity that is much smaller than the escape velocity, but continues with a "gentle push" until the material is out of the stellar gravitational field. The most popular candidate has for some time been radiation pressure on dust. Although this mechanism has many advantages and seems to be an important part of the accelerating mechanism, it nevertheless also has problems which seem to indicate that it is not the only "motor" in the process. The most common assumptions in the calculation of grain formation are 1) that the grains are formed instantly in a static upper stellar atmosphere (decoupled from the photosphere) whenever the temperature drops below the condensation temperature, 2) that all the available molecules necessary to form grains contribute in the process, and 3) that the grains are spherical, and, often, of only one size. A considerable step forward was recently achieved by Dominik, Gail, Sedlmayr and Winters (1990), when the assumptions of static atmosphere, complete condensation, and grain form were all relaxed in a calculation of mass loss from carbon-rich carbon stars (C/O = 2). Their calculations correctly matched both the observed wind velocities and the mass loss rate. In particular, the degree of completeness in the condensation increased considerably when cooler stars were simulated, and consequently the mass loss rate increased from the amount predicted by Reimers' law up to a value three orders of magnitude higher, when the temperature in the relevant stars decreased from 2800 K to 2100 K. This simulation gives strong support to the idea that radiation pressure on dust is an important ingredient in the red giant mass-loss mechanism. The main complication in the results seems to be that rather high luminosities and low temperatures are required to explain the observed high rates of mass loss. The main simplification used in the computations was the lower boundary, which consisted of an over-simplified model photosphere. Photospheric models that match the infrared part of the spectrum of a well-studied cool carbon star (TX Psc) have previously been constructed (Jorgensen 1989), and it is now the plan to compute, in a collaborational project, such photospheric models as the lower boundaries of dynamic grain-forming models. Whether this will shift the theoretical mass loss lines in the HR diagram to the necessary lower luminosi- ties and higher temperatures will be learned in the near future. One of the big advantages of the dynamic models is that they can predict the detailed relative abundances of the various grains, and also their size distribution (Dominik, Gail and Sedlmayr 1989). These parameters can today be compared directly with measurements. Since 1987 it has been known (Lewis et al. 1987) that carbonaceous chondrites (i.e. dust material from the proto-solar nebula, which was only slightly heated during the formation of the solar system) contain (inter)stellar grains. These are grains that have certain isotopic compositions markedly different from the rest of the solar system. I have suggested (Jorgensen 1988) that the identified meteoritic diamond and SiC grains were condensates from the atmospheres of carbon stars. While the SiC grains fit the observations and model atmospheres of carbon stars well in all respects, there are still fundamental problems concerning the diamond grains. Computations of dynamic carbon-star model atmospheres, with diamond dust and SiC dust included, are now in progress. If this work leads to an observa- tional verification of the origin of the meteoritic grains, it could improve considerably our understanding of the way material is transported from dying stars into newborn stars and planets, as well as improve our understanding of galactic chemical evolution. It will obviously also open up new possibilities to be able to study (pieces of) stars in the electron-microscope instead of in telescopes, as will be possible if specific stellar origins of the diamond and SiC dust in meteorites can be verified. A major problem for the presently computed dust-driven red giant winds was pointed out recently by Johnson (1991). The available static photospheric models of red giants seem to predict that the wind must start already at a height in the atmosphere where the temperature is still too high for grains to condense out of the gas. The density, rho, of the (static) atmosphere, at any height r above the stellar radius R (given by the chosen value of Teff and L as R = sqrt [ L / (4*pi*sigma*Teff**4) ] ), is computed along with the model atmosphere. Since the mass loss rate, dm/dt = rho * v * (R+r)**2, is known from observations (typical values for carbon stars are dm/dt = 2*10-7 Msun/ year), the velocity, v, at any depth in the atmosphere can now be determined for stars of given effective temperature and luminosity. Of course these "moving, static" atmospheres are, strictly speaking, internally inconsistent, but the boundary between where the atmosphere is still static (i.e. where the imposed assumption of hydrostatic equilibrium is still valid) and where this assumption breaks down is where the wind must start. In existing photospheric models, the predicted temperature in the wind-initiating region is higher than the condensation temperature of the relevant grains, which implies that something other than dust is necessary in order to initiate the wind. The mechanism needed to "lift" the atmosphere could be either an actual acceleration of the material (initiation of the wind at low altitudes) or a static levitation in the sense of a mechanism, not presently included in the model atmosphere computations, that is able to increase the (static) density of the atmosphere at a given geometrical depth. Two likely mechan- isms that could possibly levitate or "puff" out the atmosphere to the cooler regions where the grains form are turbulent pressure (waves) and radiation pressure on molecules in the upper photosphere. Bowen (1988) computed pulsa- tional shocks by introducing a "piston" at the bottom of the atmosphere with a pre-specified period. The long-period (hydrodynamic) waves produced by this mechanism were able to extend the atmospheres of red giants for piston- periods set equal to the pulsation periods of Mira-type variables, and in this way to lift the material out to the grain-forming regions. Cuntz (l990) introduced short-period acoustic waves generated by the stochastic motion of convective cells and found that they were able to deliver an appreciable amount of energy into the upper atmosphere. Undoubtedly, a large fraction of this energy would be transformed into heating the chromosphere. Hartmann and MacGregor (1980) proposed mass loss induced by Alfven waves, and Pijpers and Hearn (1989) studied acoustically driven winds, whereas Gustafsson and Plez (1992) considered the effect of radiation pressure on molecules in oxygen- rich stars. For a general review, see Lafon and Berruyer (1991). Jorgensen and Johnson (1992) calculated the radiative force on molecules in carbon-rich as well as oxygen-rich cool red giants, and they found that the radiative force could amount to a substantial fraction of the gravitational force at the surface of such stars. For red giants of solar metallicity and solar C/O ratio, the radiative force, m*a-rad, was found to be typically 1% of the gravitational force, m*g. For increasing C/O ratio the radiative force first decreases due to the disappearance of water vapor in the atmosphere as C/O approaches unity. As C/O increases above 1, the radiative force increases rapidly, due to the increased formation of polyatomic carbon molecules (HCN, C2H2, and C3). At C/O close to 2, typical values of a-rad/g (the levitation) are 10%. As an example it was estimated that a carbon star evolving from C/O = 1.02 and log(g) = 0 at 3500 K to C/O = 2.0 and log(g) = -1 at 2500 K would show an increase in the levitation from 0.3% to 13% (at tau-ross = 10-6, and increasing further outward) during its evolution. The exact number de- pends crucially on the estimated opacity of the gas, and it can be argued that the numbers given above are likely to be lower limits, mainly because several molecules are known to be excluded from our opacity computations (because the necessary data do not yet exist). On the basis of these considerations it is therefore sound to postulate that as C/O increases, during the AGB evolution, molecular levitation becomes increasingly capable of counteracting gravitation and eventually levitating the material sufficiently for it to be blown out of the star. An extensive analysis of 30 of the brightest carbon stars in the sky (Lambert, Eriksson, Gustafsson and Hinkle 1986) showed that all of these stars have C/O < 2. This surprisingly low number may now be understood as a consequence of the rapidly increasing radiative force on molecules as a func- tion of increasing C/O ratio. In this scenario, the AGB stars "die due to their own pollution". When enough carbon is mixed to the surface, the opacity, and therefore also the levitation, increases rapidly, followed by an increased mass loss rate. Soon the star has lost its upper envelope and is consequently left as a post-AGB object approaching the WD cooling sequence. This scenario leads to the state- ment made at the beginning of this article, that if the mass loss from low and intermediate mass stars is strongly concentrated to the very late phases of stellar evolution, then the 97% of all stellar mass-return that comes from these stars may have a strong impact on the chemical evolution of the uni- verse -- at C/O = 2 the yield from elements produced in situ is very great. An estimate of the corresponding molecular levitation effect for stars of lower metallicity has shown that a decrease in the metallicity by an order of magnitude, relative to the solar value, also decreases the radiative force by an order of magnitude. If levitation by radiative force on molecules is (one of) the triggering mechanism(s) for high-efficiency mass loss in red giants, it follows that much higher values of C/O will be reached by carbon stars in low-metallicity galaxies than in the solar neighborhood (and hence that the relative impact on chemical evolution was even greater in the past). This fact may be part of the explanation of why the number ratio of carbon stars to M-type giants is a decreasing function of metallicity, as observed, for example, by Richer and Westerlund (1983). In order to compute the levitation to the desired accuracy, it is criti- cally important to have better knowledge of the molecules that absorb the emitted light. This is also necessary in order to compute realistic lower photospheric boundaries for the dust-driven winds in general. Hence it is unlikely that substantial progress in the detailed modelling of red giant winds can be reached before the molecular opacity problems of cool stars are solved. From a chemical point of view, potentially important molecules like C2H and CH2 are as different as is an O star from an M star to a classifier. Therefore a great variety of problems have to be tackled before we can expect to be able to compute satisfactory molecular opacities throughout the stellar environment. Elsewhere in this Newsletter is described a coming IAU Collo- quium that specifically addresses these problems. With considerably better molecular opacities, on the other hand, we may expect to be able to construct, for the first time, realistic unified atmospheric models that are not merely photospheric models, but dynamic models that include a realistic photosphere, computed all the way up to the region where grains form, and a true dynamical wind. One of the immediate challenges of such a model is to combine existing knowledge of the photospheres and chromospheres of red giants. In fact, no self-consistent photospheric-chromospheric model of a red giant star has so far been constructed. Jorgensen and Johnson (1991) combined an existing photospheric model that was known to fit the observed IR spectrum of a bright cool carbon star (TX Psc) with an existing chromospheric model of the same star, that was known to match the observed UV spectrum. Combining these two models showed that the chromospheric model forces the strongest CO lines in the model spectrum to form emission cores, if the chromosphere is to cover more than 10% of the surface area. Apparently, the chromosphere of a red giant is not a "sphere", but may be pillars or clouds of hot material sticking out into the cool surrounding photosphere. The dust must form in the coolest photospheric extensions around the hot clouds or pillars. A unified, dust- forming, dynamic, cool red-giant model atmosphere must include such complex- ities. On the other hand the far-ranging perspective of the model should make it worth the large effort that will be required. Support from the Carlsberg Foundation and valuable comments from H. R. Johnson and R. F. Wing are gratefully acknowledged. Bowen, G.H. 1988, ApJ 329, 299 Cuntz, M. 1990, ApJ 349, 141 Dominik, C., Gail, H.-P., and Sedlmayr, E. 1989, A&A 223, 227 Dominik, C., Gail, H.-P., Sedlmayr, E., and Winters, J.M. 1990, A&A 240, 365 Gustafsson, B., and Plez, B. 1992, in: Instabilities in Evolved Super- and Hyper-giants, ed. C. de Jager, Royal Netherlands Acad. Arts and Sci., Elsevier, p.86 Hartmann, L., and MacGregor, K.B. 1980, ApJ 242, 260 de Jager, C., Nieuwenhuijzen, H., and van der Hucht, K.A. 1988, A&ASS 72, 259 Johnson, H.R. 1991, A&A 249, 455 Jorgensen, U.G. 1988, Nature 332, 702 Jorgensen, U.G. l989, ApJ 344, 906 Jorgensen, U.G., and Johnson, H.R. 1991, A&A 244, 462 Jorgensen, U.G., and Johnson, H.R. 1992, A&A 265, 166 Lafon, J.-P., and Berruyer, N. 1991, A&AR 2, 249 Lambert, D.L., Gustafsson, B., Eriksson, K., and Hinkle, K.H. 1986, A&ASS 62, 373 Lewis, R.S., Ming, T., Wacker, J.F., Anders, E., and Steel, E. 1987, Nature 326, 160 Pijpers, F.P., and Hearn, A.G. 1989, A&A 209, 198 Richer, H.B., and Westerlund B.E. 1983, ApJ 264, 114 III. Research News M. Querci (Toulouse) has sent the following abstract for a paper entitled "A complementary network to GNAT: an Arabian and French project for 3T1M automated photometric stations" by F.R. Querci, M. Querci, S. Kadiri, and L. de Rancourt: "A project to establish a network of automated photometric stations on very high mountaintops around the north-tropical latitude is under progress. Some Arabian countries are interested in collaborating in this project proposed by French and Morrocan astronomers. The design of the entire station - telescopes, power supply, antenna, etc. - is under way at the G.I.E.-TELAS Company. The data will be transmitted simultaneously by satellite to all scientific centers of the network. The main scientific aims are the monitoring of variable stars with many characteristic time variations, and the search for planets around stars." D.G. Luttermoser (Iowa State Univ.) and H.R. Johnson (Indiana Univ.) have investigated the NLTE effects of the ionization and excitation equi- libria in the atmospheres of red giant stars. This work concentrated on results for hydrogen and helium for both pure photospheric models and photosphere-chromosphere models. Although hydrogen is strongly over- ionized with respect to LTE throughout much of the atmosphere of the photospheric models, almost all the hydrogen is still in the ground state or associated in H2, and the hydrogen contribution to the electron density is small. Thus this overionization has little impact on the atmospheric structure or the emergent spectrum. Similar, even stronger, results are found for helium. Radiative equilibrium models converged with polyatomic molecular opacities have hydrogen ionization and excitation equilibria that are fundamentally different from those of models converged with diatomic molecular opacities alone. The Lyman lines control the ioni- zation and excitation in these model atmospheres. The carbon abundance influences the ionization and excitation of hydrogen through the strong bound-free opacity under the Lyman lines: carbon stars are overionized and overexcited in hydrogen with respect to oxygen-rich stars of similar temperatures and surface gravities. Attaching a chromospheric tempera- ture rise to the outer layers further increases the ionization in the upper photosphere from "chromospheric backwarming." Partial redistribu- tion effects in Ly-alpha have significant consequences in the ionization and excitation of hydrogen in the temperature minimum region of these models. Although the Lyman lines are very optically thick throughout much of these models, detailed balance does not hold for these lines, and detailed radiative transfer plays a fundamental role in the ioni- zation and excitation of hydrogen throughout the atmospheres. Helium is ionized only slightly in these chromospheric models, and its ioni- zation plays essentially no role. Luttermoser has also found that NLTE effects have a substantial influence on the ionization equilibrium in neutral metals in the upper and middle photospheres of late-type giants. As a result, the electron density is very different from values given by the Saha equation. As well, chromospheric (or shock) ultraviolet radiation has a great impact on the ionization and excitation of carbon and magnesium in the upper and middle photosphere. However, this chromospheric radiation has little effect on the results for neutral calcium and sodium. The strong resonance lines of these elements, like those of hydrogen, control much of the ionization and excitation in these stellar atmospheres. Elements with strong resonance lines that lie in wavelength regions where strong bound-free opacities form in a chromosphere (e.g. hydrogen, carbon, magnesium) are greatly over- ionized and overexcited with respect to LTE and to results obtained with the pure photospheric models. Elements with strong resonance lines coincident with background continuous opacities that form in the photosphere (e.g. sodium, calcium) are not influenced by the chromospheric radiation field. The result of this work demonstrates that NLTE effects from the photospheric radiation field play a signi- ficant role in the ionization and excitation of neutral metals and shows that "back-flowing" chromospheric photons into the photosphere can lead to even larger NLTE effects. This will have a major impact on abundance studies for these stellar types. Luttermoser and Brown (JILA) have published results of a survey of optically bright N-type carbon stars at 3.6 cm with the VLA radio telescope (1992, ApJ, 384, 634). The observations set upper limits to the radio flux, and hence to the temperature of their winds (T < 10,000 K). Surprisingly, one of the carbon stars in the sample was actually detected -- the Mira-type variable V Hya. Luttermoser and Brown show that the radio flux in V Hya is consistent with shock models of dynamic atmospheres generated by G. Bowen. Luttermoser continues the investigation of radiative transfer in the shocked atmospheres of Mira-type variables in collaboration with G. Bowen and L.A. Willson. NLTE radiative transfer calculations are being carried out with the Bowen hydrodynamic models using the PANDORA code. These calculations have led to some surprising results. The hydrodynamic models produce hydrogen Balmer lines in which the flux in H-alpha is less than H-beta and the flux in H-beta is less than H- gamma. Such inverted Balmer decrements have been observed in the spectra of Miras and in the past have been attributed to obscuration by overlying absorption. Luttermoser and his colleagues have found that radiative transfer effects in the lines for these shocked atmo- spheres are a significant cause of this phenomenon. They have also found that the peculiar phase shift between the maximum Balmer line and maximum Mg II h & k line fluxes seen in Miras results from the existence of a permanent chromosphere or "calorisphere." The following report is from the Radioastrophyical Observatory, Latvian Academy of Science: Detailed abundance analyses have been carried out for 32 barium and normal G-K giant stars using high-dispersion spectra and model atmospheres. A significant enhancement of s-process elements was found for seventeen stars. The abundances of light and iron-peak elements are in general equal to those in the standard star Epsilon Vir. However, Na, Mg, Mn, and Co are systematically slightly de- ficient (by about 0.2 dex). The elements heavier than Ni are enhanced by up to about 1.5 dex compared to the standards, while the r-process element Eu has roughly normal abundance. No substan- tial differences could be found in the abundances, atmospheric parameters, or luminosities of radial-velocity variable and non- variable barium stars. Therefore it seems that both groups of stars belong to a single family of peculiar giants. Comparisons between the mean observed s-process abundances for our uniform barium star sample and theoretical predictions from various neutron exposures show that an AGB star with a 13C neutron source can best reproduce the abundance data of these stars. Low neutron density single exposures of approx. 1.0 mb-1 are also shown to result in good agreement with the barium-star observations. Mass-transfer scenarios were tested using the chemical compo- sitions and orbital parameters of BaII stars. Since a correlation exists between s-process abundance anomalies and orbital periods for barium-star binaries, it is concluded that a wind accretion scenario is more promising than a Roche lobe overflow model. Abundance patterns for barium and carbon stars have been com- pared. Good agreement was found for the iron-group metals, but carbon stars show higher s-process element abundances (by 0.9 dex in the mean). Therefore, the companions to the BaII stars were perhaps once carbon stars, and the second dilution was roughly D = 0.9. Eglitis and Eglite (Riga) obtained carbon to oxygen abundance ratios in the atmospheres of 343 carbon stars in the Orion Arm of the Galaxy. They found that the R0-R8 stars have C/O ratios from 1.0 to 1.2 and the R9-N stars have C/O ratios in a wide interval from 1.0 to 1.8 in their atmospheres. Most stars in both cases have C/O ratios close to 1.0. A set of carbon stars with a possible evolutionary connection has been singled out. These stars have the following common pro- perties (the first being the most important): (1) their atmospheres have enriched carbon abundances (C/O > 1.2); (2) they are situated along the galactic equator; (3) SiC2 bands are typically present; and they show (4) very strong C2 Swan bands, (5) increased 13C abundances, and (6) long-period light variability. It is possible that these are luminous stars in a pre-planetary-nebula stage of evolution. Now the Riga group is studying a large set of homogeneous obser- vational material, obtained at the Bjurakan Astrophysical Observatory 2.6-m telescope, on spectra of carbon stars in the Perseus Arm and toward the Galactic Anticenter. J. Lattanzio (Monash University) is working with R. Cannon (Cambridge) and C. Frost (Monash University) on hot-bottom burning in AGB stars. They have recently found temperatures over 20 million degrees at the base of the convective envelope of a 3 Msun model with Magellanic Cloud composition. They are also adapting the method used by Cannon (in his study of Thorne-Zytkow objects) to the case of AGB envelopes and interiors. This algorithm allows for upward and downward moving streams of matter to have different compositions, and it also includes some diffusive mixing horizontally between the streams. It incorporates the results of 3D convection calculations, in that the upward-moving matter moves slowly but covers about 80% of the cross-sectional area. Conversely, the downward-moving matter is concentrated in small, fast flows. Preliminary calculations should be completed by the end of the year. The same code can be used to study the ingestion of 13C into the helium shell during ther- mal pulses (as discussed by Bazan and Lattanzio in the ApJ recently). The young planetary nebula IRAS 21282+5050 has recently been imaged at several wavelengths by several groups, and it turns out that the ionized region is not as compact as originally thought. Likkel (U. Illinois), Morris (UCLA), Kastner (Haystack), Omont (IAP, Paris) and Forveille (Grenoble) found that the nebula is roughly 3 x 4 arcseconds in diameter at 2 and 6 cm, and it appears larger (about 6 arcsec) at 2.2 microns. The radio flux indicates an electron density lower than expected for a young planetary nebula, suggesting that IRAS 21282+5050 started ionization so recently that it has not yet reached its peak electron density. B.W. Bopp and W. Asbury (Univ. of Toledo) continue a program of high-resolution spectroscopic observations of S-star binaries and carbon stars (in collaboration with H.R. Johnson and T. Ake). They have moni- tored the H-alpha line in several S stars (including HD 35155, HR 1105, HDE 332077, HR 363, and Omi1 Ori) at roughly weekly intervals since 1990. The Ca II infrared triplet region in several carbon and M-type stars (TX Psc, W Cyg, UU Aur, and R Lyr) was observed in September/October 1992 to search for possible chromospheric variability. All the data were obtained with the University of Toledo 1.0-m telescope, echelle spectrograph, and CCD detectors. D. Barthes (Montpellier) has completed a dissertation entitled "Pseudoperiodicity of Mira-type variable stars: luminosity forecasting, power spectra, and theoretical interpretation". Because of their large-amplitude luminosity variations, Mira variables could not be included in the observational program of the HIPPARCOS satellite unless light ephemerides valid for the whole mission were provided. With this aim in view, Barthes developed a method of reliably forecasting the magnitudes of these stars so that this particularly important region of the H-R diagram can be explored by HIPPARCOS. The forecasting method includes the computation of power spectra of pseudoperiodic chronological series with missing data. The quality of these spectra is high enough to warrant attempting a theoretical interpretation giving the nature of the main pulsation mode of these stars. Moreover, without needing to know their distances, indicative values of the intrinsic physical parameters which determine their evolution can be found. Such an interpretation has been carried out in detail for a few stars. Their atmospheres seem to be responsible for one component of their power spectra. U. Jorgensen and P. Thejll (Niels Bohr Institute) have developed a fast and accurate method for calculation of low-mass stellar evolu- tion with a minimum of model dependence. The method relies on observed relations among L, Teff, and Z, combined with numerical relations among L, M, Z, and M-core at the helium core flash. In a study of globular clusters it was found that an average mass difference of around 0.15 Msun between the RGB and the HB stars can be explained as due to normal Reimers-type mass loss at the RGB, with a value of eta = 0.45. The spread in HB mass needed to explain the HB morphology requires a dis- tribution of eta between 0.0 and 0.7. In another application of the method, an anlaysis was carried out of the possible formation of the AGB "manque stars" suggested to explain the UV excess in UV-bright metal-rich galaxies. The analysis showed that stars with metallicity close to the solar value cannot evolve into the manque phase within the lifetime of the universe. Stars in more metal-rich galaxies can form in a fashion that can explain the behavior of the UV excess as a function of metallicity, if R (the ratio of helium to metallicity enrichment) in these galaxies is close to 1.3 times the solar value. Several types of studies relating to red giants are being done at Tennessee State University by G. Henry and Joel Eaton. Henry (with several collaborators, including S. Baliunas) has analyzed VRI photometry of 10 semiregular variable red giants obtained over an interval of 5 years with the Fairborn 10-inch robotic telescope. His analysis finds that all are multiply periodic. Typical periods range from a few weeks to a few months, and there are typically 3-5 periods per star. Eaton, H. Johnson, and R. Cadmus (Grinnell College) have obtained two additional years of Mg II fluxes for 6 or 7 semi- regular variables with IUE. All are variable in this chromospheric line, to about the same extent as they are in visible light. Eaton has obtained about 300 H-alpha spectra of 250 red giants at the National Solar Observatory. These span the spectral types G-M and luminosity classes I-III. They include several carbon stars and S stars. These spectra show a strong (EW = 1.2 A) chromospheric H-alpha line in all the K and M giants earlier than about M5, although the line is narrower and weaker in the cooler M giants if the continuum is placed at the average of the molecular photospheric spectrum. Eaton has finished analyzing the chromospheres of Zeta Aur and 32 Cyg with a combination of archival IUE data and optical data from the National Solar Observatory. The mass density-temperature relation is similar to those for semiempirical models of cool giants, but the density distribution is different. Densities in these stars seem to be much higher than in homogeneous models, probably as the result of clumping of the chromospheric gas. Eaton and R. D. Chapman have begun observing the 1992-93 atmospheric eclipse of 31 Cyg with IUE and the McMath stellar spectrograph at NSO. They will obtain IUE spectra at about 20 phases of the atmo- spheric eclipse, using these to follow mass density, excitation temperature, and ionization throughout the chromosphere, and using contemporaneous violet spectra to determine the excitation of hy- drogen in the chromosphere. Eaton has also obtained high-dispersion, long-wavelength IUE spectra of AL Vel (K0 III + B8) at three phases of deep chromospheric eclipse in May 1992. These show a very strong, fast wind in Mg II, as well as many chromospheric Fe II absorptions. Eaton has completed an analysis of about 80 IUE spectra of 31 Cyg (B3-4 + K4 Ib) obtained outside the atmospheric eclipse to study the wind and the B star. These give an improved velocity curve for the B star, which makes it possible to derive more precise masses and radii for the two stars. The wind in this system probably has a terminal velocity near 80 km/s, with most wind lines formed in the last 20-30 km/s of its acceleration. In the radiation field of the B star, different ions recombine at different distances from the binary system: Si+2 before Fe+2 before O+. This implies that the hydrogen remains ionized throughout the wind. Absorption features formed by accretion onto the B star are seen sporadically throughout all parts of the orbit now sampled. T. Ake (CSC/GSFC), A. Jorissen (ESO), H. Johnson (Indiana U.), M. Mayor (Geneva Obs.), and B. Bopp (U. Toledo) have obtained low- dispersion IUE spectra of the Tc-poor S star HDE 332077, suspected of having a main-sequence A-type companion and perhaps a WD companion as well. A trace of spectrum appears on a 20-hour exposure with the SWP camera. The spectral slope in the SWP region is indeed that of an A star, but it is difficult to match the LWP and SWP fluxes with a simple addition of a red giant and an early main-sequence star. A. Jorissen (ESO), D.T. Frayer (Univ. Virginia), H.R. Johnson (Indiana Univ.), M. Mayor (Obs. Geneva), and V.V. Smith (Univ. Texas) have studied the 2-micron fluxes (from TMSS) and infrared fluxes (from IRAS) of S stars. They confirm that Tc-poor S stars have values of R = f(12)/f(2) less than 0.1, while Tc-rich S stars have values of R greater than 0.1. Tc-rich (intrinsic) S stars appear to belong to a younger and brighter population. That Tc-poor S stars are "accidental" or "extrinsic" S stars which have received their load of s-process elements from prior mass transfer is confirmed by radial-velocity studies. It appears that more than 90% of S stars (in a volume- limited sample) are accidental S stars. Luttermoser, Johnson, and Eaton have studied ten NLTE chromospheric models for M giant stars, 30 g Her (M6 III) being the prototype. Even with the use of partial redistribution in the Mg II lines, it appears impossible to match all the observations, particularly the Mg II and C II] lines in the UV, with a single-component model in hydrostatic equilibrium. An echelle spectrogram of FG Sge was obtained by G. Gonzalez (Univ. Washington) in June 1992, before the start of its current decline. Another spectrogram was obtained by B. Carney in September after the star had declined by about 1.5 mag. The decline is continuing and additional spectra at any resolution would be very useful to under- standing the unprecedented behavior of FG Sge. Gonzalez is planning to analyze these two spectra and would be pleased to analyze other spectra that are obtained. For the past several years, J. Percy (U. Toronto) and colleagues have been carrying out a systematic study of the photometric variability of M giants. This study has included surveys of suspected variables: Percy and Fleming (PASP 104, 1992), Percy and Shepherd (IBVS No. 3792, 1992), and an ongoing "Project SARV" (small-amplitude red variables) involving the photoelectric observers of the AAVSO. It has also in- cluded detailed analyses of individual stars (e.g. EU Del) using photo- electric observations and combined visual and photoelectric observations. In some respects, the results have been predictable and in accord with previous knowledge: the incidence, amplitude, and period of variation are generally higher in cooler stars, with the coolest being the large- amplitude Mira variables. Percy and colleagues have, however, discovered that many of the small-amplitude variables have two periods: one of typically 30 to 100 days, which, over many seasons, is much more regular than expected and is likely due to pulsation, and another, typically an order of magnitude longer, whose cause is unknown. It is interesting that there are other cool variables (notably the RVb sub-class of the RV Tauri stars) which show a similar phenomenon. Any suggestions about the cause of this behavior would be much appreciated. Percy is also studying long-term changes in Miras, using a 75-year database of visual observa- tions archived by the AAVSO. He finds that almost all stars with long- term changes in mean or maximum magnitude have S, N, C, or R-type spectra. IV. Special Topics MOLECULAR OPACITIES IN THE STELLAR ENVIRONMENT IAU Colloquium 146 Niels Bohr Institute and Nordita, Copenhagen, Denmark May 24 - 29, 1993 PROGRAM: D.R. Alexander (USA): Dust opacities. J. Almlof (USA): Ab initio computations of small grains and clusters. M.S. Bessell (Australia): Molecules in Mira variable stars and related objects. A. Borysow (USA): Pressure induced molecular absorption in stellar atmospheres. M. Costes (France): Dissociation energies and partition functions of small molecules. S.P. Davis (USA): The Berkeley program on molecules of astrophysical interest. C. Demuynck (France): Millimeter laboratory research on small radicals. E.F. van Dishoeck (Holland): Continuous molecular opacities and photo-dissociation. N. Grevesse and J. Sauval (Belgium): Molecular data from solar observations. K.H. Hinkle (USA): Molecules in circumstellar envelopes. P. Jensen (Germany): the MORBID method. H.R. Johnson (USA): The role of NLTE on molecular opacities. U.G. Jorgensen (Denmark): Molecular data bases. R.L. Kurucz (USA): Computation of opacities for diatomic molecules. D.L. Lambert (USA): The role of molecules in stellar atmospheres. S. Langhoff (USA): Computational approaches to determining accurate band strengths. M. Larsson (Sweden): Completeness and accuracy in the computation of the absorption coefficient for diatomic molecules. K.K. Lehman (USA): Intensity measurements of weak molecular transitions. J. Liebert (USA): Molecules in cool dwarf stars. P.-AA. Malmqvist (Sweden): The RASSCF and RASSI methods used on small molecules of astrophysical interest. S. Miller (UK): Computations of line-strengths for polyatomics. M. Morillon (France): Production, detection and study in the infrared of unstable molecules and radicals. D. Muchmore (Canada): Molecular formation and destruction processes. H. Olofsson (Sweden): Millimeter observations of molecules in stellar environments. A. Omont (France): Polyynes and PAHs in stellar environments. S.D. Peyerimhoff (Germany): The absorption coefficient of small carbon hydride molecules. J. Schamps (France): Oscillator strengths in metallic oxides and hydrides. M.S. Seaton (UK): Atomic opacities and the marriage with molecular opacities. E. Sedlmayr (Germany): From clusters to grains. P. Thejll (Denmark): The role of H2+ in white dwarf atmospheres. T. Tsuji (Japan): Astrophysical applications of approximate methods for molecular opacities. R. Wehrse (Germany): Opacity problems in cool low mass stars. H. Yorke (Germany): Opacity problems in proto-stellar objects. SCIENTIFIC ORGANIZING COMMITTEE: G. Graner (France), U.G. Jorgensen (chairman; Denmark), D.L. Lambert (USA), B.O. Roos (Sweden), T. Tsuji (Japan), and R. Wehrse (Germany). LOCAL ORGANIZING COMMITTEE: U.G. Jorgensen, B. Mottelson, B. Pagel, P. Thejll (chairman). SPONSORING IAU COMMISSIONS: Commission 14: Atomic and Molecular Data Commission 29: Stellar Spectra Commission 36: Theory of Stellar Atmospheres Participation and hotel reservation can be guaranteed only if the registration is received no later than January 31, 1993. Registration material can be obtained from: Anne Lumholdt Nordita Blegdamsvej 17 DK-2100 Copenhagen Denmark e-mail: lumholdt@nordita.dk V. Meetings May 24-29, 1993 Molecular Opacities in the Stellar Environment Copenhagen, Denmark Contact: Uffe Jorgensen Niels Bohr Institute Blegdamsvej 17 DK-2100 Copenhagen, Denmark uffegj@nbivax.nbi.dk June 3-6, 1993 Technical Conferences on Optical Spectroscopic Instrumentation and Techniques Albuquerque, New Mexico Contact: Society of Photo-Optical Instrumentation Engineers (SPIE) P.O. Box 10 Bellingham, WA 98227 USA July 12-16, 1993 Circumstellar Media in the Late Stages of Stellar Evolution The 34th Herstmonceux Conference Cambridge, UK Contact: Anne Reynolds Royal Greenwich Observatory Madingley Rd. Cambridge CB3 OEZ, UK areynolds@uk.ac.cam.gxvg September 20-24, 1993 The MK Process at 50 Years: A Powerful Tool for Astrophysical Insight Tucson, Arizona Contact: Christopher J. Corbally, S.J. Vatican Observatory Research Group University of Arizona Tucson, AZ 85721 USA corbally@as.arizona.edu October 11-14, 1993 Cool Stars, Stellar Systems, and the Sun Eighth Cambridge Workshop Univ. of Georgia, Athens, GA Contact: Jean-Pierre Caillault Dept. of Physics & Astronomy Univ. of Georgia Athens, GA 30602 USA jpc@jove.phyast.uga.edu November 5-6, 1993 Hot Stars in the Halo - A Colloquium Honoring A. G. Davis Philip Schenectady, NY Contact: Saul Adelman Dept. of Physics The Citadel Charleston, SC 29409 USA adelmans@citadel.bitnet VI. The WG Organizing Committee Hollis R. Johnson Uffe Grae Jorgensen Antonio Mario Magalhaes Astronomy Dept. Niels Bohr Institute Instituto Astronomico Swain West 319 Blegdamsvej 17 e Geofisico Indiana University DK-2100 Copenhagen Universidade de Sao Paulo Bloomington, IN 47405 Denmark Caixa Postal 9638 USA uffegj@nbivax.nbi.dk Sao Paulo, SP 01065-970 johnsonh@iubacs.bitnet Brazil magalhaes%iagusp. decnet@fapq.fapesp.br Monique Querci Verne V. Smith Robert E. Stencel Obs. Midi-Pyrenees Dept. of Astronomy Center for Astrophysics & 14 Avenue Edouard Belin University of Texas Space Astronomy (CASA) F-31400 Toulouse Austin, TX 78712 Univ. of Colorado France USA Campus Box 391 querci@fromp51.bitnet verne@astro.as. Boulder, CO 80309 utexas.edu USA stencel%galaxy@vaxf. colorado.edu Takashi Tsuji Robert F. Wing Sandra B. Yorka Tokyo Astronomical (WG chairman) (Editor) Observatory Dept. of Astronomy Dept. of Physics & Mitaka Ohio State Univ. Astronomy Tokyo 181 174 W. 18th Avenue Denison University Japan Columbus, OH 43210 Granville, OH 43023 ttsuji@c1.mtk.nao.ac.jp USA USA ts4718@ohstmvsa.acs. yorka@denison.bitnet ohio-state.edu or yorka@cc.denison.edu