Your posts:

"If so, it's not a poisson. I calculated expected values that way and did a chi-square on the numbers and it's WAY off."

"John;
Here's the numbers. Each number represents the number of times a given time interval between the main quake and the next similar quake over 1000 km away.
I stopped at 30 but there were not all that many further out.
623 464 356 281 258 198 175 131 108 104 84 76 51 39 30 36 35 17 15 12 11 7 5 4 6 2 4 3 0 0"

I didn't do anything formal, but these look like the numbers expected from a Poisson distribution.

As Chris pointed out, if the average spread of M6+ events is 4 days, the rate should diminish by 25% each day until one reaches intervals so long that they never are encountered in the 30-year catalog. These number look by eye to match such a geometric progression.