The declaration Greg pointed out is only necessary if your *identifiers* need unicode. If all your unicode is safely tucked away inside of strings, use u"Pe\\xf1asco" or u"Pe\\u00f1asco". The u-prefix makes them unicode objects, as opposed to string objects. Almost all builtin and library funcs will handle them properly, including .replace().

[link|http://docs.python.org/lib/typesseq.html|http://docs.python.org/lib/typesseq.html]

[edit: added link to docs]