R help - Apr 2008 - How do I specify a model with nested random terms?

Dear R-users,

I have problems specifying a model with nested random terms. I want to 
reproduce a variance component model made in Mintab, the model 
specification in Minitab is:

Response = Prod P Prod*P Sample(Prod P) Extract(Prod P Sample)
        Random Prod Sample Extract

The dataset is balanced, and "P" is a fixed effect (treatment) while 
"Prod", "Sample" and "Extract" are random effects.
"Sample" is nested under the "Prod*P" interaction, and
"Extract" is nested
under "Sample". 

I tried using the lme function with the following formula specification:

lme(Response~Prod*P, Random=~1|Prod/P/Sample/Extract)

but the results do not look right. I don't understand how to specify the 
"fixed" and "random" arguments correctly. Could someone
explain these
arguments to me? I have tried to read the help files and manuals but I 
still don't get it :-S

Would it also be possible to use the aov function with Error() strata for 
this problem?

Thank you and regards,
Ingrid Måge 
	[[alternative HTML version deleted]]