3.4. The importance of galaxy mergers since z ~ 1
Three independent methods have been brought to bear on this problem - the evolution of close galaxy pairs, the evolution of morphologically-disturbed galaxies, and the evolution of early-type galaxies as a proxy for plausible merger remnants. All three methods suffer from important systematic and interpretive difficulties. Close galaxy pairs, even supplemented with spectroscopy to isolate galaxies within 500 km s-1 of each other, remain contaminated by galaxies in the same local structures. Some measures of morphological disturbance are susceptible to this source of error to a lesser extent, and a clear consensus on the meaning of the different automated and visual measures of morphological disturbance is yet to emerge. Early-type galaxies will be the result of only a subset of galaxy mergers and interactions, and the role of disk re-accretion and fading in driving early-type galaxy evolution is frustratingly unclear.
Yet, despite these difficulties some broad features are clear. Mergers between ~ L* galaxies are almost certainly more frequent at higher redshift than at the present day, but this does not imply by any means that galaxy mergers are unimportant at z ~ 1: indeed, it is possible that an average ~ L* galaxy undergoes roughly 1 merger between z = 1 and the present day ([Le Fèvre et al. (2000)]). This is supported by the substantial build-up in stellar mass in red-sequence and early-type galaxies since z ~ 1. Substantial uncertainties remain and important questions, like the stellar mass dependence of the merger rate, or the fraction of dissipationless vs. dissipational mergers, are completely open. Nonetheless, these first, encouraging steps imply that the massive galaxy population is strongly affected by late galaxy mergers, in excellent qualitative agreement with our understanding of galaxy evolution in a CDM Universe.