[R-sig-eco] specifying permutations for pseudoreplications in RCBD

[Guillaume Adeux]

Hello everyone,

I am trying to test the effect of tillage on weed communities using CCA.
I would like to make sure my permuting scheme is correct.
My experimental lay-out is the following:

4 blocks
3 plots within each block (one for each tillage system)
2 pseudoreplicates within each plot/tillage system

require(vegan)
require(permute)
data(varespec)
block= c(rep("B1",6), rep("B2",6), rep("B3",6), rep("B4",6))
tillage=rep(c("CT","CT","RT","RT","NT","NT"),4)

CCA=cca(log(varespec+1) ~ tillage + Condition(block))
h2 <- how(within = Within(type = "none"),plots = Plots(strata =
interaction(block,tillage), type = "free"),blocks = block,nperm=999)
anova(CCA,permutations=h2,model="reduced",by="term")

If I understood the {permute} vignette correctly, this should permute the
observations at the plot level within a block (i.e. keeping the two
pseudoreplicates together while respecting the block hierarchy).
Can someone confirm or refute?

Thank you very much for your help.

Sincerely,

Guillaume ADEUX

	[[alternative HTML version deleted]]