I ran your test script (replacing the character entity codes with actual UTF-8 characters) and it works fine for me. So the problem is with your command prompt console not interpreting UTF-8 characters correctly. Try printing a UTF-8 character manually in the console using echo Ʃ for example, and see what happens.

I also ran your CGI script (adding actual UTF-8 characters for testing) and it worked fine too. Check your browser's character encoding setting (for example, in Firefox it's under View->Character Encoding) - if it's isn't UTF-8 already, then set it manually for testing.

There are also some quick guides on setting up UTF-8 in Perl.