URLs in Webmaster Tools and not site:

[Tonearm]

Very old URLs have suddenly appeared in Webmaster Tools as 404s. Their sudden appearance may have to do with the fact that I started returning proper 404s about a month ago, but the interesting part is these URLs were NOT in site:www.example.com.

[tedster]

Sounds like Google dug back into their back office for historical URL records. They will do that from time to time, just to be sure that an old URL hasn't been brought back. However, URLs where Google gets a 404 response should not appear in the publicly available site: query - glad to hear that they don't.

[Tonearm]

But Google wasn't getting a 404 until recently and the pages were not visible with site:www.example.com before that change.

[tedster]

Right - but Google stores WAY more information than we can access on the public side. Their history and age data patent application [webmasterworld.com] and others make that very clear. They have a long memory about your domain's history.

So they had the urls from somewhere or other, stored along with whatever HTTP server response they were getting - and from time to time they were re-checking them. When they recently got a 404 response, they let you know in GWT.

[Tonearm]

OK, I was under the impression that all HTTP 200 pages were displayed with site:www.example.com if there are less than 1000. This really muddies the duplicate content waters.

[g1smd]

If Google is telling you the URLs are 404, then they already know the pages are gone. That is good. There is nothing to worry about.

They will keep on checking to see they remain gone. That, too, isn't a problem.