Don't overlook Mike's other point. Why not sample all quakes greater than 7M since instrumentation has been put in place? That would, I think, give a more representative sample, but a smaller pool from which to draw your conclusions. Still, might be interesting.

Cathryn