Starburst galaxies are currently forming massive stars at prodigious rates. I discuss the star-formation histories and the shape of the initial mass function, with particular emphasis on the high- and on the low-mass end. The classical Salpeter IMF is consistent with constraints from observations of the most massive stars, irrespective of environmental properties. The situation at the low-mass end is less clear: direct star counts in nearby giant H II regions show stars down to ~1 M⊙, whereas dynamical arguments in some starburst galaxies suggest a deficit of such stars.
Destruction of wide binary stars in low-mass elliptical galaxies: implications for initial mass function estimates