In the discussion about Google bots hitting Moodle sites quite hard (http://moodle.org/mod/forum/discuss.php?d=29319), N Hansen pointed out that there were repeated hits to non-existant URLs such as:
You can fix this URL by adding &course=1 to the end of it, so I'm guessing that Google has stripped the last parameter as they say on their site that their bot doesn't like dealing with too many of them.
I'm guessing that (part of) the problem may be that these pages do not return a machine readable error message. Perhaps error404 (or possibly some other) error could be added. I think this might dissuade the various search engine bots and possibly be the 'correct' thing to do.