[R-sig-eco] different symbols to NMDS graph: it works PARTIALLY!??!?

cristabel.duran at waldbau.uni-freiburg.de cristabel.duran at waldbau.uni-freiburg.de
Thu Jan 13 08:50:14 CET 2011


Dear list, 

I have a problem adding points with different symbols to my
NMDS ordination graph.

I performed a NMDS ordination with this vegan function
(metaMSD) as follow:
	metaf <- metaMDS(spp, distance = "bray", k = 3, trymax =
1000, wascores = T, autotrasform = F, shrink=T, noshare =
1) 
spp is my species matrix.

Then, I want to assign different colors and symbols to my
plots ("sites") according to three categories. These
categories are in the vector called pos with the categories
"gap", "border" and "forest".

For that I am using the following functions and commands:

	ScoresNMDS<-scores(metaf)
	plot(metaf,type="none",choices = c(1,3))
	points.metaMDS(ScoresNMDS[pos$for_pos_anl=="border",],
display="sites",choices = c(1,3), pch=16, cex=0.8,
col="green")
	points.metaMDS(ScoresNMDS[pos$for_pos_anl=="gap",],display="sites",choices
= c(1,3),pch=17, cex=0.8, col="black")
	points.metaMDS(ScoresNMDS[pos$for_pos_anl=="forest",],display="sites",choices
= c(1,3), pch=4,col="red",cex=0.8, bg="red")

The problem:
I get a empty ordination graph, were I am able to add the
points corresponding to the first two categories ("border"
and "gap").
The strange thing, is that it is impossible to add the
third category, I get this error message:

	Error in dimnames(x) <- dn : 
	length of 'dimnames' [1] not equal to array extent

Already tried:
- The length of my explanatory variable (pos) have the same
length that the species matrix.
- Changing the name from my open office spreadsheet, in the
case there is a strange character.
- Change colors and symbols.
- etc...etc.

What I also do not understand, is why it works for some
categories of the same vector and why not for the last
category.
The same happens with ordiellipse and ordihull.

I would appreciate your help! 
Thanks in advance and have a nice day!
Cristabel.


Cristabel Durán Rangel. PhD Student.
Institute of Silviculture. Faculty of Forest and
Environmental Sciences. University of Freiburg.
Germany
Telf: +49 (761) 203 8604 (ofc) 

„Man lernt die Physiognomie einer Landschaft desto besser
kennen, je genauer man die einzelnen Züge auffaßt, sie
untereinander vergleicht und so auf dem Wege der Analysis
den Quellen der Genüsse nachgeht, die uns das große
Naturgemälde bietet.“
				Alexander von Humboldt, 1799



More information about the R-sig-ecology mailing list