[R] Problems running IsoMDS using vegdist with pres-abs data and two sites with zero distance

[Gavin Simpson]

On Sat, 2006-10-21 at 13:19 +0200, Leonard Sandin wrote:
> Hi
> I have just (finally) started to poke around in R and wanted to analyse 
> a stream fish dataset with 28 sites and 18 species. When trying to 
> follow the Vegan manual to run nmds from distance measures calculated by 
> the vegdist function it turns out that I have two sites (streams) with 
> the exactly the same four species (I have used pres-abs data in this 
> case). When I try to run isoMDS I get an error message saying that:
> Error in isoMDS(vare.dis) : zero or negative distance between objects 14 
> and 15
> i.e. the two sites with the same fish species. I can run the decorana 
> function with "sensible" results so the dataset seems to be ok.
> My next step would be to compare the results of the fish ordination with 
> some other datasets from the same sites (i.e. through procrustes 
> rotation) - so I wouldn´t like to remove one of the "twin" sites.
> Any suggestion on how to circumvent this or am I simply using the wrong 
> method?
> 
> Best wishes, Leonard
> 

Hi Leonard,

I assume you are running isoMDS (package MASS) directly? If so, take a
look at metaMDS function in vegan which has an argument zerodist which
if you set it to "add" adds a small distance so you don't get this
error.

HTH

G

-- 
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%
 Gavin Simpson                 [t] +44 (0)20 7679 0522
 ECRC & ENSIS, UCL Geography,  [f] +44 (0)20 7679 0565
 Pearson Building,             [e] gavin.simpsonATNOSPAMucl.ac.uk
 Gower Street, London          [w] http://www.ucl.ac.uk/~ucfagls/cv/
 London, UK. WC1E 6BT.         [w] http://www.ucl.ac.uk/~ucfagls/
%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%~%