I just got my thick copy of the Bulletin of the Seismological Society of America. This time they have a big section on a hundred years after the 1906 San Francisco earthquake.
Most of the work has been in precisely modeling the ground motion of the earthquake. There were a few seismometers working at the time, there were reports of seismic shaking, and there were some land surveys done before and after.
One paper confirmed that there was 'super shear' rupture, and this may now be considered 'common' for large earthquakes. This means that the rupture proceeds down the fault plane at a speed greater than the shear wave velocity. As I have mentioned before, there was no great reason for having a shear wave limit on the rupture velocity, since it all depends on the critical displacement (amount of motion before the strength drop).
Although the rupture along the fault plane is shear, and the far field seismic wave is shear, you can have a compressive wave whipping up one side of the fault, and a tension (P-wave) on the other side. This makes for shear along the fault, and the speed of the rupture approaches the P-wave velocity as the critical displacement goes down to zero.
It all means that there is greater seismic ground motion near the fault. As I have said, it's a lot worse for thrust faults, where you get hammered by this 'Fist of God'.
They also did a lot more on local soil amplification, but I always feel they haven't done enough on this.
There is also a great field mapping of the giant Pakistan earthquake.