Robert_Charlton - 1:19 am on Jul 23, 2013 (gmt 0)
I haven't checked this in quite a while, but a related:example.com/pagefile search used to return domains/pages that shared an inbound link source with the url being checked.
As the web has grown, it appears that Google is looking at the results in ways that go beyond a simple shared link source. It's probably now a statistical view of what, on a smaller web, was a simple list. Google maxes out on related: searches at about 190 results, at least on those I tried.
Thinking out loud, related: returns mostly domains, probably to be expected... but perhaps a way of merging a group of pages.