Roger, John,

Roger-your logic sounds correct. I don't think this is a small population-not withh 400+ quakes/bin. If they were random it should be much smoother IMHO. John-are you around? Any idea what we are seeing in Roger's numbers? Need an explanation of what we are doing?
I don't think that the earthquakes 20-30 days apart are related (although Lowell had those sorts of relationships). But, if the program is working and the catalogue does not have some strange incompleteness, then would we see this result if there was a periodicity in the quakes-increased seismicity on a 20-30 day period....Hmm, the solar rotation period falls in that range.... I don't believe this right now, but it can be thrown out there as a hypothesis. If there was such a period you should be able to see it, roughly, by just plotting quakes vs time (quakes within 500 km removed to get rid of aftershocks and swarms), with a running average to smooth it out.
Chris