We will construct the first self-consistent models of star formation that follow the galactic scale flowswhere molecular clouds form yet still resolve the star formation and feedback events down to sub-parsec scales.By following the full galactic ecology, the life cycle of gas from the interstellar medium into stars and their radiative and kinematic output back intothe galaxy, we will develop a comprehensive theory of star formation. The link between the large-scale dynamics of the galaxy and thesmall-scale star formation provides the ground-breaking nature of this proposal.Star formation produces a wide rangeof outcomes in nearby molecular clouds yet on large scales yields star formation rates that are strongly correlated to galactic-scale gas densities.These observed properties of star forming galaxies have inspired a plethora of theoretical ideas, but until now there has beenno means of testing these analytical theories.We will use galactic-disc simulations to determine how molecular clouds form through self-gravity, spiral shocks and/orcloud-cloud collisions. We will use these self-consistent models of molecular clouds to follow the local gravitational collapse toform individual stars and stellar clusters.We will include ionisation, stellar winds and supernovae into the ISM to study how feedback can supportor destroy molecular clouds, as well as triggering successive generations of young stars.We will also conduct Galactic bulge scale simulations tomodel how gas flows into, and star formation occurs in, the Galactic centre.The primary goals of this proposal are to understand what determines thelocal and global rates, efficiencies and products of star formation in galaxies, and to developa complete theory of star formation that can be applied to galaxy formation and cosmology.