Recent developments in the theory of element production and the chemical evolution of the galaxy are presented. Following this, observational data and their interpretation are given. A case by case analysis of results for D, He, Li and CNO isotope data in the disk and center of our galaxy is presented; previous results for element gradients are also summarized. The primordial abundances of D and He cannot be directly obtained from observations; corrections for stellar processing are discussed. From these data and the Li abundances, it appears that the abundance of the light elements is consistent with the standard big bang. In agreement with previous results, the range of % the baryon to photon ratio, is 5-8 10-l~ If the amount of non-baryonic matter is small, these results indicate an open universe, in the standard big bang model. New data show a gradient in the (12C/13C) and (160/180) ratios with galactocentric distance, Dec. The presence of a gradient in the (14N/15N) ratio is less clear and there is no measurable gradient in the (32S/34S) ratio. In the interstellar medium near the sun, the carbon isotope ratio is _~ 20 percent lower than the solar system ratio. This indicates that there has been only a moderate amount of enrichment of the nearby interstellar medium since the formation of the solar system. These results and previously determined galactic element gradients are interpreted in the framework of chemical evolution models. Delayed recycling of nucleosynthesis products is essential for the correct interpretation of the results. Comparisons of data with galactic evolution models are discussed.

This series of papers comprises a systematic exploration of the hypothesis that the far ultraviolet radiation from star clusters and elliptical galaxies originates from extremely hot horizontal-branch (HB) stars and their post-HB progeny. This first paper presents an extensive grid of calculations of stellar models from the Zero Age Horizontal Branch through to a point late in post-HB evolution or a point on the white dwarf cooling track. The grid will be used to produce synthesized UV fluxes for the interpretation of existing and future short wavelength (900-3000Å) observations. Our sequences have been computed for a range of masses which concentrates on models that begin their HB evolution very close to the hot end of the ZAHB. We have calculated tracks for three metal-poor compositions ([Fe=H] = 2:26; 1:48; 0:47 with [O=Fe] > 0), for use with globular cluster observations. We have also chosen three metal rich compositions (Z = 0:017 = Z , Z = 0:04; 0:06) for use in the study of elliptical galaxy populations. For each of the two supermetal-rich compositions, for which the helium abundance is unconstrained by observation, we have computed two sets of sequences: one assuming no additional helium, and a second with a large enhancement (Y HB = 0:29 and 0.36 for Z = 0:04), and (Y HB = 0:29 and 0.46 for Z = 0:06). For each set of sequences our lowest ZAHB envelope masses (M 0 env ) are in the range 0:002 < M 0 env < 0:006 M . We use the term 'Extreme Horizontal Branch' (EHB) to refer to HB sequences of constant mass that do not reach the thermally-pulsing stage on the AGB. These models evolve after core helium exhaustion into Post-Early Asymptotic Giant Branch (AGB) stars, which leave the AGB before thermal pulsing, and AGB-Manqué stars, which never reach the AGB. We describe various features of the evolution of post-HB stars, discussing the correspondence between slow phases of evolution at high temperature and the Early-AGB evolution. We note that the relationship between core mass and luminosity for stars on the upper AGB is not straightforward, because stars arrive on the ZAHB with a range of masses and subsequently burn different amounts of fuel. We determine from our models an upper bound to the masses of EHB stars, finding that it varies little for [Fe=H] < 0, but that it is sensitive to the helium abundance. We show that for each composition there is a range of M 0 env (at least a few hundredths M ) in which the models have a slow phase of evolution at high temperature. The duration of this phase is found to increase with the metallicity, but its luminosity is lower, so that total UV energy output is not significantly different from metal-poor sequences. The properties of very metal rich stars are, however, made uncertain by our lack of knowledge of the helium abundance for [Fe=H] > 0; the range of stellar masses in which high temperatures are attained for significant periods of time increases with Y . There is no intrinsic composition dependence of the peak UV output from evolved stars; the output from a ste...

Primordial nucleosynthesis after the Big Bang can be constrained by the abundances of the light elements and isotopes 2H, 3He, 4He and 7Li (ref. 1). The standard theory of stellar evolution predicts that 3He is also produced by solar-type stars, so its abundance is of interest not only for cosmology, but also for understanding stellar evolution and the chemical evolution of the Galaxy. The 3He abundance in star-forming (H II) regions agrees with the present value for the local interstellar medium, but seems to be incompatible with the stellar production rates inferred from observations of planetary nebulae, which provide a direct test of stellar evolution theory. Here we develop our earlier observations, which, when combined with recent theoretical developments in our understanding of light-element synthesis and destruction in stars, allow us to determine an upper limit for the primordial abundance of 3He relative to hydrogen: 3He/H = (1.1 +/- 0.2) x 10(-5). The primordial density of all baryons determined from the 3He data is in excellent agreement with the densities calculated from other cosmological probes. The previous conflict is resolved because most solar-mass stars do not produce enough 3He to enrich the interstellar medium significantly.

We have made a detailed study of the color distribution of the main sequence of the globular cluster (GC) NGC 2808, based on new deep HST WFPC2 photometry of a field in the uncrowded outskirts of the cluster. The observed color distribution of main-sequence stars is not Gaussian and is wider than expected for a single stellar population, given our (carefully determined) measurement errors. About 20% of the sample stars are much bluer than expected and are most plausibly explained as a population having a much larger helium abundance than the bulk of the main sequence. Using synthetic color-magnitude diagrams based on new stellar models, we estimate that the helium mass fraction of these stars is Y $ 0:4. The newly found anomaly on the main sequence gives credence to the idea that GCs like NGC 2808 have undergone self-enrichment and that different stellar populations were born from the ejecta of the intermediate-mass asymptotic giant branch (AGB) stars of the first generation. The enhancement and spread of helium among the stars in NGC 2808 have recently been suggested as a simple way to explain the very peculiar morphology of its horizontal branch. We find that if in addition to the Y ¼ 0:40 stars, roughly 30% of the stars have Y distributed between 0.26-0.29, while 50% have primordial Y, this leads to a horizontal-branch morphology similar to that observed. In this framework, three main stages of star formation are identified, the first with primordial helium content Y ' 0:24, the second born from the winds of the most massive AGBs of the first stellar generation ($6-7 M ), with Y $ 0:4, and a third born from the matter ejected from less massive AGBs ($3.5-4.5 M ), with Y $ 0:26-0.29. There could have been a long hiatus (several times 10 7 yr), between the second and third generation in which no star formed in the protocluster. We suggest that during this period, star formation has been inhibited by the explosion of late Type II supernovae deriving from binary evolution.

We have recently obtained a set of high-resolution images of Terzan 5 in the K and J bands by using MAD 6 , a Multi-Conjugate Adaptive Optics demonstrator instrument installed at the Very Large Telescope (VLT) of the European Southern Observatory (ESO). MAD operates at near-infrared wavelengths, thus revealing the only component of stellar radiation that can efficiently cross the thick clouds of dust obscuring the Galactic bulge. It is able to perform exceptionally good and uniform adaptive optics correction over its entire field of view (1'x1'), thus compensating for the degradation effects to the astronomical images induced by the Earth's atmosphere. In particular, we have obtained a set of K-band (2.2m) images of Terzan 5 close to the diffraction limit (Fig. 1). The sharpness and uniformity of the images yields very high quality photometry, resulting in accurate (K, JK) colour-magnitude diagram (CMD) even for the very central region of the cluster, and leading to a surprising discovery. We have detected two well-defined red horizontal branch clumps, separated in luminosity: a bright horizontal branch (BHB) at K = 12.85 and a faint horizontal branch (FHB) at K = 13.15, the latter having a bluer (JK) colour (Fig. 2).We have carefully considered whether the double horizontal branch could be spurious. It is neither due to instrumental effects ( Fig. 2), nor to differential reddening 7,8 3 (as the two horizontal branch clumps in the CMD are separated in a direction which is essentially orthogonal to the reddening vector), nor to field contamination (while field stars are expected to be almost uniformly distributed over the MAD field of view, the radial distributions of the stars belonging to the two horizontal branch clumps are significantly concentrated toward the cluster centre and are inconsistent with a uniform distribution at more than 5 level; see Fig. 3a and Supplementary Information). We have also found that the radial distributions of the two horizontal branch populations are different ( Fig. 3a): according to a Kolmogorov-Smirnov test, the BHB population is significantly (at > 3.5level) more centrally concentrated than that of the FHB. The stars belonging to the BHB are also substantially more numerous than those of the FHB near the cluster centre (that is, at distances r < 20''), becoming progressively more rare at larger radii (Fig. 3b).Once alerted to the existence of the double horizontal branch, we have also (Fig. 4a).To date, apart from a significant spread in the abundance patterns of a few light elements (such as Na and O) 1 , the chemical composition of all globular clusters in the Galaxy is known to be extremely uniform in terms of iron content, with the only exception being  Centauri 4,5 in the Galactic halo. Hence, Terzan 5 is the first stellar aggregate discovered in the Galactic bulge that has globular-cluster-like properties but also with the signatures of a much more complex star formation history.To further investigate this issue, we have performed a differential reddening correc...