[R] Complex lme model

[Emilio A. Laca]

I would like to request help for coding the lme for the following:

Soils from 14 Regions (fixed) were sampled every Year (fixed) during three
years. Each Year each Region was sampled at several Points (random), and two
Cores (random) were taken at each Point. Each Core was split into five
Depths (fixed, crossed with all other factors) and C (response) was measured
at each Depth of each Core of each Point of each Region in each Year.

First, I want to assume that Points were randomly selected each year, and I
want to test an AR(1) over Depth.

Second, I want to assume that Points were selected only once during the
first year. Each near, new cores were randomly selected from each of the
original points. I would like to test and AR(1) for year.

Help will be greatly appreciated and knowledge will be passed on to other
students. Thank you.

PS:  lme(C ~ Region*Year*Depth, data=soil, random=~1|Region/Point/Core) just
returned:

Error in MEEM(object, conLin, control$niterEM) :
    Singularity in backsolve at level 0, block 1


-- 
Emilio A. Laca     
One Shields Avenue, 2306 PES Building
Agronomy and Range Science                    ealaca at ucdavis.edu
University of California                      fax: (530) 752-4361
Davis, California  95616                            (530) 754-4083