2003MNRAS.343..249S

Up to half of the baryons inferred to once have been in our galaxy have not yet been detected. Ejection would seem to provide the most attractive explanation. Previous numerical studies may have underestimated the role of winds. I propose a solution involving a multiphase model of the protogalactic interstellar medium and the possibility of driving a superwind. Simulations do not yet incorporate the small-scale physics that, I argue, drives mass-loading of the cold-phase gas and enhances the porosity, thereby ensuring that winds are driven at a rate that depends primarily on the star-formation rate. The occurrence of hypernovae, as claimed for metal-poor and possibly also for starburst environments, and the possibility of a top-heavy primordial stellar initial mass function are likely to have played important roles in allowing winds to prevail in massive gas-rich starbursting protogalaxies as well as in dwarfs. I discuss why such outflows are generically of the same order as the rate of star formation and may have been a common occurrence in the past.