Content explicitly disallowed by a domain's robots.txt file is excluded from the Google index.
Websites that are not linked from other websites that Google already knows. That is, there are probably a lot of websites that do not get linked from visible pages, those websites are never going to be found by the Google spider unless they're manually submitted to Google via the Webmaster Tools.
Websites that are behind web forms that you need to fill out.
Census images. Since the content are images that are often manually index, they usually found on paid-for sites like ancestry.com.
Aside from Twitter, Google does not index Tumblr all that well. Blog posts on Tumblr are easier to find using Tumblr search. Also everything on Google Sites isn't (or is hardly) indexed. If you start a Google site, get your own domain.
Smaller blogs that aren't regularly updated are often dumped from search results. Plus anything that they think is a splog.