What sort of evidence would you like? My server log file is full of traffic from page 1 of Google's SERPS pointing to pages in the SI.

Are you saying that not using modification date information could cause SI issues?

Also, I have to wonder how accurate our calculations of Google's Supplemental Index size are today. Consider our method for determining what pages are in the main index, since SI pages are no longer marked - we pass a false query to Google that causes an internal error and causes some documents to not be included in the results. I have always suspected that the false query might only cause results from the current datacenter to be reported. This would block SI pages, since they are not in the same index, but if this is the case it would also block newly indexed, updated and modified pages that have not been pushed out to other data centers from showing up. This means that our calculations could, potentially, be inaccurate. Any ideas or suggestions how to test that the SI ratio calculation ONLY takes into account SI pages?

__________________
The best way to learn anything, is to question everything.