It is commonly believed that galaxies use, throughout the Hubble time, a
very small fraction of the baryons associated to their dark matter halos
to form stars. The fact that galaxies have typically very low
baryon-to-stars conversion efficiencies both at low-mass & high-mass is
typically interpreted as a consequence of stellar & AGN feedback
respectively. While this is robustly established for the average galaxy
population, is this true for all galaxies?
In this talk I will show that if we can measure accurately the
kinematics of the cold gas in disc galaxies, which traces very well the
circular velocity, then we can measure these baryon-to-stars conversion
efficiencies quite reliably for individual objects. High-quality radio
interferometric data can be used to collect a fairly large (>100) sample
of nearby discs of all masses, from dwarfs to massive discs, from which
we find a couple of surprises. 1) the population of massive (>~10^11)
discs has systematically higher efficiencies than spheroids of similar
masses; 2) the population of field, gas-rich ultra diffuse galaxies also
has systematically higher efficiencies than a typical dwarf. These are
two cases where AGN and stellar feedback, respectively, has failed to
lower the baryon-to-stars conversion efficiencies of these galaxy
population. I will finally discuss the implications of these results in
our understating of galaxy evolution.