We present a detailed abundance analysis based on high-resolution and high signal-to-noise spectra of eight extremely metal-poor ( EMP) stars with [Fe/H] less than or similar to 3.5 dex, four of which are new. Only stars with 4900 K < T-eff < 5650 K are included. Two stars of the eight are outliers in each of several abundance ratios. The most metal-poor star in this sample, HE 1424-0241, has [Fe/H] similar to -4 dex and is thus among the most metal-poor stars known in the Galaxy. It has highly anomalous abundance ratios unlike those of any other known EMP giant, with very low Si, Ca, and Ti relative to Fe, and enhanced Mn and Co, again relative to Fe. Only ( low) upper limits for C and N can be derived from the nondetection of the CH and NH molecular bands. HE 0132-2429, another sample star, has excesses of N and Sc with respect to Fe. The strong outliers in abundance ratios among the Fe-peak elements in these C-normal stars, not found at somewhat higher metallicities ([Fe/H] similar to -3 dex), are definitely real. They suggest that at such low metallicities we are beginning to see the anticipated and long sought stochastic effects of individual supernova events contributing to the Fe-peak material within a single star. With spectra reaching well into the near-UV we are able to probe the behavior of copper abundances in such extreme EMP stars. A detailed comparison of the results of the analysis procedures adopted by our 0Z project compared to those of the First Stars VLT Large Project finds a systematic difference for [Fe/H] of similar to 0.3 dex, our values always being higher.

We present chemical abundances for O, Na, Mg, Al, Si, Ca, Ti, and Fe in eight red giants and one turnoff star in the metal-rich globular cluster 47 Tuc, based on spectroscopy with the Magellan Inamori Kyocera Echelle high-resolution spectrograph on the Magellan 6.5 m Clay telescope. A robust line by a line differential abundance analysis technique, relative to the K-giant Arcturus, was used to reduce systematic errors from atmospheric and atomic parameters. Our derived mean LTE [Fe/H] of -0.76 +/- 0.01 +/- 0.04 dex (random and systematic error, respectively) is more metal poor by about 0.1 dex than recent literature results. The chemical element ratios in this nearby globular cluster most closely resemble those of the Galactic bulge, although there is a non-negligible overlap with the composition of thick-disk stars. We find that the [Al/Fe] and [Na/Fe] ratios coincide with the upper boundary of the trends seen in the bulge and thick disk. There is only a small intrinsic scatter in the majority of the abundance ratios, indicating that 47 Tuc is mostly a rather chemically homogeneous system.

We show that the Galactic bulge and disk share a similar, strong, decline in [O/Mg] ratio with [Mg/H]. The similarity of the [O/Mg] trend in these markedly different populations suggests that the strong decline is due to a metallicity-dependent modulation of the stellar yields from massive stars by mass loss from winds, related to the Wolf-Rayet phenomenon, as proposed by McWilliam & Rich in 2004. We have modified existing models for the chemical evolution of the Galactic bulge and the solar neighborhood with the inclusion of metallicity-dependent oxygen yields from theoretical predictions for massive stars that include mass loss by stellar winds. Our results significantly improve the agreement between predicted and observed [O/Mg] ratios in the bulge and disk above solar metallicity; however, a small zero-point normalization problem remains to be resolved. The zero-point shift indicates that either the semi-empirical yields of Francois et al. obtained in 2004 need adjustment, or that the bulge initial mass function ( IMF) is not quite as flat as found by Ballero et al. in 2007. Our result removes a previous inconsistency between the interpretation of [O/Fe] and [Mg/Fe] ratios in the bulge, and confirms the conclusion that the bulge formed more rapidly than the disk, based on the overabundances of elements produced by massive stars. We also provide an explanation for the long-standing difference between [alpha/Fe] and [O/Fe] trends among disk stars more metal rich than the Sun.

To compare the globular clusters (GCs) associated with the Sagittarius Galaxy (Sgr) we report the results obtained from new high-resolution spectra of red giant stars in Terzan 8 and Arp 2, collected with the Magellan Inamori Kyocera Echelle (MIKE) spectrograph at Las Campanas Observatory. For Ter 8 we find [Fe I/H] = -2.37 +/- 0.04 and [Fe II/H] = -2.30 +/- 0.07, while for Arp 2 [Fe I/H] = -1.77 +/- 0.04 and [Fe II/H] = -1.89 +/- 0.07. Other elements, usually of interest, show the following results. For oxygen, we obtain [O/Fe] = 0.71 +/- 0.17 for Ter 8 and [O/Fe] = 0.21 +/- 0.22 for Arp 2. For the alpha-elements (Mg, Si, Ca, Ti), the mean values are 0 ''.37 +/- 0.14 dex for Ter 8 and 0.31 +/- 0.11 for Arp 2. These values are not too different from the Galactic GCs at a similar [Fe/H] value. Regarding Cu, we find quite underabundant results for both clusters: [Cu/Fe] = -0.96 for Ter 8 and [Cu/Fe] = -0.93 for Arp 2. The heavy s-process elements Ba and Nd show no excesses in Ter 8 and Arp 2. Our only r-process element, Eu, shows an excess of 0.45 dex in Arp 2, based on the only useful line at 6645 angstrom.

Aims. We model the chemical evolution of manganese relative to iron in three different stellar systems: the Solar neighbourhood, the Galactic bulge, and the Sagittarius dwarf spheroidal galaxy, and we compare our results with recent and homogeneous observational data sets.

We report on the abundance analysis of two red giants in the faint Hercules dwarf spheroidal (dSph) galaxy. These stars show a remarkable deficiency in the neutron-capture elements, while the hydrostatic alpha-elements (O, Mg) are strongly enhanced. Our data indicate [Ba/Fe] and [Mg/Fe] abundance ratios of less than or similar to - 2 and similar to + 0.8 dex, respectively, with essentially no detection of other n-capture elements. In contrast to the only other dSph star with similar abundance patterns, Dra 119, which has a very low metallicity at [Fe/H] = -2.95 dex, our objects, at [Fe/H] similar to -2.0 dex, are only moderately metal-poor. The measured ratio of hydrostatic/explosive alpha-elements indicates that high-mass (similar to 35 M-circle dot) Type II supernovae progenitors are the main, if not only, contributors to the enrichment of this galaxy. This suggests that star formation and chemical enrichment in the ultrafaint dSphs proceeds stochastically and inhomogeneously on small scales, or that the IMF was strongly skewed to high-mass stars. The neutron capture deficiencies and the [Co/Fe] and [Cr/Fe] abundance ratios in our stars are similar to those in the extremely low metallicity Galactic halo. This suggests that either our stars are composed mainly of the ejecta from the first, massive, Population III stars ( but at moderately high [Fe/H]), or that SN ejecta in the Hercules galaxy were diluted with similar to 30 times less hydrogen than typical for extreme metal-poor stars.

Context. The evolution of C and O abundances in the Milky Way can impose strong constraints on stellar nucleosynthesis and help in understanding the formation and evolution of our Galaxy.

We present LTE chemical abundances for five red giants and one AGB star in the Galactic globular cluster (GC) M5 based on high-resolution spectroscopy using the Magellan Inamori Kyocera Echelle spectrograph on the Magellan 6.5 m Clay telescope. Our results are based on a line-by-line differential abundance analysis relative to the well-studied red giant Arcturus. The stars in our sample that overlap with existing studies in the literature are consistent with published values for [Fe/H] and agree to within typically 0.04 dex for the alpha-elements. Most deviations can be assigned to varying analysis techniques in the literature. This strengthens our newly established differential GC abundance scale and advocates future use of this method. In particular, we confirm a mean [Fe I/H] of -1.33 +/- 0.03 (stat.) +/- 0.03 (sys.) dex and also reproduce M5's enhancement in the alpha-elements (O, Mg, Si, Ca, Ti) at +0.4 dex, rendering M5 a typical representative of the Galactic halo. Over-ionization of Fe I in the atmospheres of these stars by non-LTE effects is found to be less than 0.07 dex. Five of our six stars show O-Na-Al-Mg abundance patterns consistent with pollution by proton-capture nucleosynthesis products.

We present high-resolution Magellan/MIKE spectroscopy of the brightest star in the ultra-faint dwarf galaxy Leo IV. We measure an iron abundance of [Fe/H] = -3.2, adding to the rapidly growing sample of extremely metal-poor (EMP) stars being identified in Milky Way satellite galaxies. The star is enhanced in the a elements Mg, Ca, and Ti by similar to 0.3 dex, very similar to the typical Milky Way halo abundance pattern. All of the light and iron-peak elements follow the trends established by EMP halo stars, but the neutron-capture elements Ba and Sr are significantly underabundant. These results are quite similar to those found for stars in the ultra-faint dwarfs Ursa Major II, Coma Berenices, Bootes I, and Hercules, suggesting that the chemical evolution of the lowest-luminosity galaxies may be universal. The abundance pattern we observe is consistent with predictions for nucleosynthesis from a Population III supernova explosion. The extremely low metallicity of this star also supports the idea that a significant fraction (similar to 10%) of the stars in the faintest dwarfs have metallicities below [Fe/H] = -3.0.