posthoc.NMDS: Two-by-two post-hoc tests for NMDSs using PERMANOVA
In mrelnoob/airpoumpoum: Multi-use package for ecological studies

Description Usage Arguments Value Note Examples

This function performs as many NMDSs and PERMANOVAs as there are possible combinations between the input factor levels. In other words, the function will compute (using the comb_mod function) all possible combinations of the modalities (levels) contained in FAC, and run a separate NMDS and associated PERMANOVA between each pair of those modalities (i.e. treatments) and the community data in REL.
The post-hoc tests are corrected using Holm-Bonferroni method to control the family-wise error rate. By default, the functions only performs NMDSs and PERMANOVAs using Bray-Curtis distances, and keep only 2 dimensions in each ordination.

1	posthoc.NMDS(REL, FAC)

`REL`	A contingency table. Typically, a species-site matrix.
`FAC`	A factor containing the different modalities/treatments we want to combine and test.

A table containing the p-values and corrected p-values for each PERMANOVA test.

This function should only be used when a preliminary PERMANOVA detected a significant effect of FAC on the centroid position of REL, and after ensuring that there was no multivariate spread in the data groups (see adonis).

## For a simulated matrix with 20 lines (sites) and
## 30 columns (species), split in 2 treatments:
library(vegan)
data(dune)
mytreatment <- gl(2, 10)
aa <- posthoc.NMDS(REL = dune, FAC = mytreatment)
aa
## Here, since there is only 2 levels in the factor, there
## will only be one combination and thus, one test.
## It's a silly example allright.