The question of the origins of cosmic systems at all scales is at the heart of modern astrophysics. In the ΛCDM galaxy formation paradigm, the first galaxies to form are the smallest, and the dwarf galaxies we observe today are their closest surviving relics. Searches in the Sloan Digital Sky Survey have already noticeably increased the number of known dwarf galaxies around the Milky Way and have revealed a population of ultra-faint objects with lower luminosities than typical globular clusters. These faint objects are the most ancient and metal-poor Galactic systems known.

A potentially much larger population, including streams from recently disrupted systems, will be revealed and characterised by new optical and near-infrared surveys and the upcoming Gaia space mission. We believe

that it is now time to thoroughly explore and assess the potential for discoveries in this field. To exploit these advances requires a range of cross-disciplinary expertise: from numerical simulations, via survey data

analysis, through to detailed chemical modelling. The immediate goals will be to enhance the derivation of accurate chemical abundances in the extremely low stellar metallicity and low gravity regime found in dwarf

galaxies and ultra-faint objects and to optimise the next target selection strategies. These are key requirements for identifying the chemical patterns of statistically meaningful samples of stars and hence to understanding the nucleosynthesis and the nature of the first stars in dwarf spheroidal galaxies.