It is possible to find galaxies that are dark matter dominated across the size spectrum - from giant low surface brightness disks to gas-rich dwarfs. Perhaps there are dark halos where the baryonic component is negligible. There have been claims of ``dark galaxies'' based on the incidence of close quasar pairs where no lens is visible, even though a massive galaxy would be required to cause the observed image splitting (Hawkins 1997). Counter to this claim is the fact that radio and optical surveys do not show the same incidence of close quasars with no visible lensing galaxy (Kochanek, Falco, & Munoz 1998). The alternative explanation is that the quasars are bound pairs, with interaction fuelling the activity in both AGN. At the other end of the mass scale, there is a clear expectation that dark halos exist, since the mass function in hierarchical clustering models is much steeper than the faint end of the galaxy luminosity function. Small halos may not be able to make or retain substantial numbers of stars.