I argue that supervenience is an inadequate device for representing relations between different levels of phenomena. I then provide six criteria that emergent phenomena seem to satisfy. Using examples drawn from macroscopic physics, I suggest that such emergent features may well be quite common in the physical realm.