The XML encoding declaration of this file says "UTF-8".

The actual encoding is "WINDOWS-1251", however there is no encoding label that says so.

Therefore, if the XML encoding declaration is respected, then, then the fist heading will contain question marks.

NB: If you are in a locale which uses "Windows-1251" as the legacy locale default,then the two headings will probably be identical.

PS: Whether the user agent should completely ignore the XML encoding declaration is, to me, an open question. As I see it, then — normally — an XML encoding declaration (and an XML declaration without the encoding declaration as well), might be a reason for the user agent to guess that the page is UTF-8 encoded. As such it fits into what HTML5 describes as “information on the likely encoding“. At least when combined with encooding pattern matching, then it could be useful to the user agent — I think. (However, for this particular file, then pattern matching would reveal the file to not be UTF-8 encoded. As such, the XML encoding declaration should, in my view, be ignored for the current file.)