The Evolution of the Tully-Fisher Relation at Redshift z~1

The power of the Tully-Fisher relation as a tool of observational
cosmology has long been acknowledged. However, all attempts of
unravelling the evolution of disc galaxies at high redshift have
proven inconclusive. We demonstrate how rotation curves measured at
the rest-wavelength of H_alpha can successfully shed some light
unto the topic. Our sample of 21 galaxies at a mean redshift of
z=0.9, the most distant sample so far, consists of mainly large
isophotally selected late disc galaxies comparable to the Milky
Way. These are amongst the largest in terms of disc scale lengths
even compared to local galaxies. From this we conclude that the
appearance of large disc galaxies must lie further back than ~8Gyr.
We find that these galaxies have much higher surface brightnesses
(1.44mag) than a local set of galaxies with similar sizes. Their
absolute magnitudes are only slightly brighter, though.
Furthermore, combining the evolution in magnitude, size and surface
brightness, we find that a scenario where galaxies grow inside-out
is more consistent with the data than self-similar evolution.
Moreover, the rotation velocities of the distant sample are
systematically lower than the local sample, resulting in a lower
average mass by a factor of ~2. The combination of those two
effects results in an offset from the local Tully-Fisher relation
of Delta_TF=-1.44mag towards brighter magnitudes. Compared to
similar data in the literature we find a consistent picture for the
evolution of the Tully-Fisher relation with redshift:
Delta_TF=(0.19+-0.19)-(1.40+-0.32)z. Total source numbers in the
individual redshift bins are too small even taking all the
available data together to meaningfully constrain the slope of the
Tully-Fisher relation at high redshifts. The observed offset from
the Tully-­Fisher relation is in good agreement with the offset
derived from numerical simulations by Steinmetz and Navarro (1999)
at a redshift z~1.