[R] fit.contrast or some other for interactions and stratified

[John Christie]

I was trying to do a bunch of planned contrast but ran into a snag.  
fit.contrast won't work with an Error term in aov.  Ok, then I use lme. 
  Only, now it seems there is no way to enter the interaction terms 
(something like "x1:x2").  I couldn't find a way to use the interaction 
term in the help.  Is there an undocumented method or some other 
command for planned contrasts?