In the local Universe roughly 60% of disk galaxies exhibit a large-scale stellar bar when observed at near-infrared wavelengths (and in roughly half these cases the stellar bar is also clearly visible at optical wavelengths). N-body simulations indicate that stellar bars can only exist in 'dynamically-cold' disks, i.e., disks that are primarily rotationally-supported with low stellar velocity dispersions. However, in the Cold Dark Matter model of our Universe nearby disk galaxies are expected to have experienced many mergers with much smaller galaxies, which can lead to a certain degree of 'disk-heating' depending on the fraction of cold gas available to stabilise the disk. Hence, by measuring the fraction of barred disks split into those with early-type (roughly, gas-poor) and late-type (roughly, gas-rich) morphologies over a wide range of redshifts (cosmic time) one can gain insights into the processes that govern the formation and evolution of disk galaxies.

In recent work with Marcella Carollo and the COSMOS team (soon-to-be-published) I have measured the bar fraction in massive, early-type (red, bulge-dominated, smooth morphologies) and late-type (blue, disk-dominated, clumpy morphologies) disks in the COSMOS field. This work has revealed that a substantial fraction of intermediate mass (10.5 < log M < 11) early-type disks are barred since as long ago as 3.3 Gyr (redshift 0.6), while the fraction of barred, late-type disks increases significantly over this time. While at high stellar mass (log M > 11) few early-type disks are barred, yet there is a significant fraction of barred, late-type disks displaying negligible evolution over these epochs. These results confirm that the Hubble sequence we observe in the local Universe is already in place for high mass galaxies at redshift 0.6, and offer a strong constraint on the possible role of minor mergers in building up the bulges of early-type disks over the past 3.3 Gyr. These key results are illustrated in the Figure displayed above.