2006ApJ...653.1049W

We use kinematic measurements of a large sample of galaxies from the Team Keck Redshift Survey in the GOODS-N field to measure evolution in the optical and near-IR Tully-Fisher (TF) relations to z=1.2. We construct TF relations with integrated line-of-sight velocity widths of ∼1000 galaxies in B and ∼670 in J; these relations have large scatter, and we derive a maximum likelihood least-squares method for fitting in the presence of scatter. The B-band TF relations, from z=0.4 to 1.2, show evolution of ∼1.0-1.5 mag internal to our sample without requiring calibration to a local TF relation. There is evolution in both TF intercept and slope, suggesting differential luminosity evolution. In J band, there is evolution in slope but little evolution in overall luminosity. The slope measurements imply that bright, massive blue galaxies fade more strongly than fainter blue galaxies from z∼1.2 to now. This conclusion runs counter to some previous measurements and to our naive expectations, but we present a simple set of star formation histories to show that it arises naturally if massive galaxies have shorter timescales of star formation, forming most of their stars before z∼1, while less massive galaxies form stars at more slowly declining rates. This model predicts that the higher global star formation rate at z∼1 is mostly due to higher star formation rate in massive galaxies. The amount of fading in B constrains star formation timescale more strongly than redshift of formation. TF and color-magnitude relations can provide global constraints on the luminosity evolution and star formation history of blue galaxies.