Nice catch, Gavin - I missed that part of the original post. The
nucleic distances need to be included as the left-hand-side of
the formula, not as the "distance" argument.
"comm" is still optional, though, but it's not a good idea to omit
it if there's any way you can provide the original data. From the
help:
"If this
is not supplied, the ``species scores'' are the axes of
initial metric scaling ('cmdscale') and may be confusing."
I don't know if it's true in this case, but there are applications
where there is no data matrix - the distances themselves are the
original data. I don't know offhand of any other constrained ordination
functions in R that will easily accomodate a precalculated distance
matrix, but I expect there are some somewhere.
My usual approach is to use metric or nonmetric multidimensional
scaling plus vector fitting, but the assumptions behind that are
different than those underlying a constrained ordination.
Sarah
--
Sarah Goslee
http://www.functionaldiversity.org
______________________________________________
[email protected] mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
