[ESS] latin-1 in windows

[pleydell at supagro.inra.fr]

Hi, I recently had to returned to Windows XP after 4 years of Debian (sobs). I
have a French database with accented characters. Using Rgui / RWinEdt the
following allows me to work correctly with the accented characters

read.table(file="myFile.txt", encoding="latin1")

however, emacs is displaying accented characters in code (e.g. "P\352cher"
instead of "Pêcher" in my imported data) and logical arguments such as
myData[mydata$species==Pêcher,] do not produce the desired results. I imagine
there are two issues: first, the internal coding of the character; second,
providing a correct display.

If I do M-x set-input-method latin-1-prefix and then try to define an object in
R using accented character...

> pêcher==1
Error: unexpected input in "p\201"

how might I pass this stumbling block?
David