lucy24 - 7:25 pm on Sep 24, 2013 (gmt 0)
You don't really need to know about the stylesheet. I just mentioned it as something I use for information.
All my error documents-- 403, 404, 410-- call on a single stylesheet that isn't used by anything else. Humans landing on an error page will automatically get the css-- and also the favicon or apple-touch-icon, if they haven't already got it. Robots don't normally get stylesheets. But the major search engines do, because they index them along with everything else. So my thinking is that if google has ever physically seen any of my custom error pages, they would have learned that the css exists and they would eventually ask for it.
When your server sends out a 4xx response, it automatically sends out the appropriate error document. But a robot can choose not to look at the document.
If you don't use a separate stylesheet for error documents, and you don't have a "noindex" header, you could try an exact-text search for something from your error document and see if it turns up. This will of course only work if there's something unusual about the wording.
But we still haven't nailed down the original question. Are you asking how to keep search engines from indexing your 404 page? Or are you asking something entirely different?