Mark was not completely happy with my previous explanation for the time delay
between the mainshock and associated far-field aftershock activity. So, I'll
try again. The nice thing about physics is there is usually more than one mechanism
involved in complicated phenomena and it keeps us thinking all the time.

The one thing we know (as a fact) is that after strong global earthquakes, ground
water levels are disturbed throughout the world often by meters in elevation. This
first occurs when the p- and s-waves pass through. The stronger the wave motion,
the more the elevation in the ground water table. If some of that water (or other
fluid) is pushed into cracks and fault zones, as it will be, it will immediately
begin the process of corrosion. Some studies have shown that with the
addition of water, fault-like processes and slippage speed up by orders of
magnitude. The chemical erosion on the surface of the fault can eventually,
given enough time, eat through asperities and locked points allowing earthquakes
to occur. However, the corrosion only lasts for a few days before the water
either drains away or is incorporated into the surrounding material. This leaves
a harder bond between the two surfaces than before (kind of like drying cement).
This mechanism would explain several observational aspects of FFT. 1)
the FFT earthquake is delayed by up to 10 days; 2) a period of seismic quiet
follows the initial burst of FFT activity.
There, happy? I hope not, because there are undoubtedly other mechansism
and associated explanations as well - don't stop searching, every answer should
lead to more questions!!