Setting the HTTP response header: Content-Type: text/html; charset=iso-8859-1 or adding the tag: <meta http-equiv="Content-Type" content="text/html; charset=iso-8859-1"> or even both - still does not deter IE from scanning the contents and interpreting them as UTF-7 when Encoding=Auto-Select. (observed on w2k with IE 6.0.2800.1106 SP1 + Q867801 + Q823353 + Q833989 + Q903235) It appears there is little that web servers can do to thwart this, short of changing all '+' characters to %2B. That seems excessive. -tom-