Dear all, I want to predict a new observation with nlme using fixed and random effects estimates from a previous fit. How can I do this? (My problem is different from the Loblolly example) Thanks a lot! Peter