R help - Jul 2007 - constraining residual variance to zero in lme or lmer

Hi All

I havent seen that lme or lmer allows to specify certain constraints (although I
heard one could program ones own covariance structure). Specific constraints, 
including one on the residual variance can be specified in sas proc mixed. In my
model I want to fix the residual variance to the known value of 0. If anyone 
having an idea about that, I would appreciate letting me know.

Thanks Toby