[R-sig-eco] plotting capscale results using ordiplot in Vegan

Jari Oksanen jari.oksanen at oulu.fi
Wed Apr 20 18:56:25 CEST 2011

On 19/04/11 20:47 PM, "ckellogg" <ctekellogg at gmail.com> wrote:

> Hi,
> I used capscale to examine relationships between relative species abundances
> at different sites with environmental variables.  When I attempt to plot the
> results using ordiplot all of my site points are squished to the center of
> the biplot.  I have tried changing the scaling and that only makes things
> worse.  Is there anyway I can get the sites to spread out a bit in the
> ordination?

Vegan makes some automatic decisions for the user, but the design philosophy
is that you overrule those decisions and make your own. It seems that you
know the 'scaling' argument (you say you used that) which helps often.
However, in all choices of 'scaling' the relation of species and site scores
is fixed in certain ways. If those alternatives (that is scaling -3, -2, -1,
0, 1, 2, or 3) are not sufficient, you can also use a special 'const'
algorithm which sets the general scaling constant for the scores, or if you
give it two values (like const = c(2,1)) you can have different scaling for
species and site scores (but then they may not be biplot scores in Gabriel
sense). See docs and vegan vignette on design decision for details.

Cheers, Jari Oksanen

More information about the R-sig-ecology mailing list