We present the NLTE abundances of 10 chemical species in 65 very metal-poor stars in eight dSphs and the Milky Way halo. The classical dSphs Sculptor, Ursa Minor, Sextans, and Fornax reveal a similar plateau at [α/Fe] = 0.3 for each of Mg, Ca, and Ti, similarly to the MW halo. We provide the evidence for a decline in α/Fe in the Boötes I UFD, that is probably due to the ejecta of SNeIa. The dichotomy in the [Sr/Ba] versus [Ba/H] diagram is observed in the classical dSphs, similarly to the MW halo, calling for two different nucleosynthesis channels for Sr. The Boötes I and UMa II UFDs reveal very similar ratios of [Sr/Mg] = −1.3 and [Ba/Mg] = –1. The stars in the Coma Berenices and Leo IV UFDs are even poorer in Sr and Ba. The subsolar Sr/Ba ratios of Boötes I and UMa II indicate a common r-process origin of their n-capture elements.

Stars in low-mass dwarf galaxies show a larger range in their chemical properties than those in the Milky Way halo. The slower star formation efficiency make dwarf galaxies ideal systems for testing nucleosynthetic yields. Not only are alpha-poor stars found at lower metallicities, and a higher fraction of carbon-enhanced stars, but we are also finding stars in dwarf galaxies that appear to be iron-rich. These are compared with yields from a variety of supernova predictions.

Based on high-resolution observed spectra, the non-local thermodynamic equilibrium (NLTE) line formation, and precise stellar atmosphere parameters, we present the first complete sample of dwarf spheroidal galaxies (dSphs) with accurate chemical abundances in the very metal-poor (VMP) regime. The obtained stellar elemental ratios are compared with chemical enrichment models, and we show that NLTE is a major step forward for studies of the dSph and the Milky Way (MW) chemical evolution.

The Fornax dSph is an interesting case as it contains five old globular clusters and its field stars, although predominantly of intermediate (3-8 Gyr old) age, cover a wide range of age and metallicity. Detailed abundance analysis is crucial to our understanding of the earliest star formation epoches, where classic CMD analysis fails to provide a unique answer. It also allows us to measure the chemical evolution of the stellar population following tracers of different enrichment mechanisms through time, e.g. SN type II (alpha elements); AGB stars (s-process elements) etc. With our large sample of abundance measurements we will obtain a detailed picture of the evolution of Fornax and of the role played by small galaxies in the building up of larger ones.

We have obtained HST-NICMOS observations of five of M31's most metal-rich globular clusters: G1, G170, G174, G177 & G280. For the two clusters farthest from the nucleus, G1 and G280, we statistically subtract the field population and estimate metallicities using K-(J - K) color-magnitude diagrams (CMDs). Based on the slopes of their infrared giant branches we estimate [Fe/H] = −1.22 ± 0.43 dex for G1 and −0.15 ± 0.37 dex for G280. We combine our infrared observations of G1 with two epochs of optical HST-WFPC2 F-band data and identify at least one LPV based on color and variability. The location of G1's giant branch in the K, (V - K) CMD is very similar to that of M107, indicating a somewhat higher metallicity than our purely infrared CMD; [Fe/H]= −0.9 ± 0.2 dex.

Population synthesis is a powerful tool to study stellar populations where the analysis of the stellar content of a composite system is based on the results of breaking down into components (of a given base) the spectrum of the observed system. Such process constitutes an inverse problem which can have a multitude of possible or “acceptable” solutions. This degenerative character of the synthesis rises mainly from observable errors and from the base of components itself with respect to its internal consistency and its (in)capacity to fully embrace all the free parameters involved.