In a recent Language Log post, Geoff Nunberg explains the metadata problems found in Google Books and why they matter. (Basically, Google Books has millions of volumes of scanned books, but the information that describes these scanned books is often wrong because Google has relied on information taken from the scanned images rather than relying on information provided by the libraries that provided the books for scanning.)
This is a big reason why any search for a downloadable copy of an old book that is out of copyright should not begin and end with Google Books. You should use archive.org as well.
Ernster, the Virtual Library Cat
Above the Law - A Legal Tabloid - News, Gossip, and Colorful Commentary on Law Firms and the Legal Profession