I use Emacs on both Linux and Windows to edit the same files (that I keep in Dropbox, which I hope is irrelevant, but who knows). The files are encoded in utf-8 with Unix line endings. Occasionally, I will notice that when editing them on Windows accented characters display incorrectly which I can correct by telling
M-x recode-region that the file was really in utf-8 but was interpreted as iso-latin-1.
Why is Emacs incorrectly guessing my file is in iso-latin-1? Why does it only do that on Windows?