I have a relatively good understanding about ordination and constrained vs unconstrained analyses.
I ran an NMDS on my species data and the superimposed habitat type with colours in R. It shows a nice linear trend from Habitat A to Habitat C which can be explained ecologically.
I then wanted to understand the variation explained by the habitat type and so I ran
rda(species ~ habitat.type)
This shifted the linear trend into a triangular relationship.
How should I interpret this ordination?
Should i say the the shift in composition among habitats in linear or non-linear? Does it mean there is a component in Habitat B that is less important (i.e RDAaxis2) than something between A:C (i.e RDA axis 1) that is not seen in the NDMS?
How do i explain the difference between the two graphs ecologically?
Thanks Tania