Circumstantial evidence suggests that merging played an important role in galactic evolution long before the present epoch. The key points of the argument can be summed up as follows:
Starbursts and AGN are signposts of high-redshift mergers; the high
incidence of such objects at z 
2 to 4 reflect frequent
merging of juvenile galaxies.
The bulk of the Milky Way's halo merged more than 10 Gyr ago as part of this activity.
Cluster ellipticals merged before z 2; their immediate
progenitors were few and only moderately gassy.
The metal-rich globular cluster systems of these ellipticals are relics of their final mergers.
Finally, direct observations of high-redshift events are complemented
by archeological investigation of nearby systems. Both approaches are
needed to discover what happened at redshifts z = 2 to 5.
I thank Alex Stephens and Hector Velázquez for communicating results
in advance of publication. I also thank Jun Makino and the University
of Tokyo for hospitality while I prepared this article. This research
made use of NASA's
Astrophysics Data System Abstract Service. Travel
to the conference was covered by air miles accumulated while following
the Grateful Dead.