The Tolman dimming and, less directly, the cosmic microwave background. On top of these, the fact that the volume densities of assorted objects increase with redshift at least as fast as the (1+z)^3 expected from cosmic expansion (much more rapidly for things which show evidence of strong evolution; the more boring the galaxy type, the closer it comes to this). And related to the Tolman test, the fact that supernova absolute magnitudes are much less constant with redshift if you leave out the photon-rate and energy-shift terms that come from redshift in an expanding spacetime (the former, as far as I know, uniquely such a prediction).