[R-sig-eco] Non-parametric test for unbalanced complete block data

[Narkis Morales San Martin]

Hi all

I need some help with some data from a germination experiment,
specifically hypocotyl and radicle elongation data. I wanted to
determine if the concentration of different leaf leachates (3
different concentrations) and the removal the mesocarp of the seeds
(with mesocarp/without mesocarp) had some effect on the germination
and hypocotyl and radicle elongation. The experiment consisted in 8
seeds per leachate treatment (control, 1:50, 1:1) with mesocarp and
without mesocarp. Each of these replicated 3 times.

The problem is that in some cases I didn't get a response (elongation)
as some seeds didn't germinate because they were attacked by a moth
larvae. I decided to take out the zeros as they didn't represent a
response to the treatment. So now I have an unbalanced complete block
with a non normal dataset. I did some research and, as I can't use a
Friedman test because of the missing data, I found a test call The
Skillings–Mack test but I couldn't find a package in R to do this.

Can any one tell me if this is the right test? and also does any one
know how to do it in R?

Regards

Narkis


-- 
Narkis Morales San Martin - PhD Candidate


School of Environment
The University of Auckland | Te Whare Wananga o Tamaki Makaurau
Private Bag 92019
Auckland 1001
New Zealand