OK. Now after looking into the code, I could see the problem and it was like expected. Your files are coded in UTF-8 format with Unix line ending, which is also correct specified in your files with
<meta http-equiv="Content-Type" content="text/html; charset=utf-8">
So when you open your html files and auto detect UTF-8 is enabled in the general configuration dialog, you will see the "ü" in "zurück" correctly in UltraEdit, although it is coded in the file as two byte character with hex code C3 BC.
But to see, that it is coded with these hex codes, you have to disable first the auto detect UTF-8 feature, then open the html file, go to "zurÃ¼ck" and switch to hex mode.
Because the "Replace In Files" has no detection of file format, it will never find a "ü" entered in the find field (with hex code FC). To do the replace with UltraEdit, you would have to search for "Ã¼" instead of "ü".
However, you have now done the replaces with Dreamweaver, so this is just for info for you and maybe for other users in future with the same problem.
For all users, a little table for UTF-8 to ANSI to OEM conversion table for all special German characters.
ANSI: ä | Ä | ö | Ö | ü | Ü | ß --- hex: E4 | C4 | F6 | D6 | FC | DC | DF
OEM: „ | Ž | ” | ™ | | š | á | --- hex: 84 | 8E | 94 | 99 | 1 | 9A | E1
UTF8: Ã¤ | Ã„ | Ã¶ | Ã– | Ã¼ | Ãœ | ÃŸ --- hex: C3 A4 | C3 84 | C3 B6 | C3 96 | C3 BC | C3 9C | C3 9F
Best regards from Austria