I THINK I have solved the problem.
Google Reader and the aggregator I was talking about had and used their cached info for the post ID’s.
So when I wrote a post “17”, it overwrote the info they had for their link to post “17”, and gave it the date of the original post. This is all speculative, but I think it worked that way.
My solution – I used phpmyadmin and changed the post ID of the most recent post to 100. The next one will be 101, and so on. Since my first database only had approx. 80 entries, I won’t have any more “repeats”.
Does this make sense to the people who know the code?
