methods for betadisper objects fail when bias.adjust=TRUE

[jrcunning]

Using the sample code from the permutest.betadisper help text, if the bias.adjust=TRUE is included when building the betadisper object "mod", downstream anova and permutest methods fail, giving this error:

Error in eval(expr, envir, enclos) : object 'Distances' not found

data(varespec)

## Bray-Curtis distances between samples
dis <- vegdist(varespec)

## First 16 sites grazed, remaining 8 sites ungrazed
groups <- factor(c(rep(1,16), rep(2,8)), labels = c("grazed","ungrazed"))

## Calculate multivariate dispersions
mod <- betadisper(dis, groups, bias.adjust=TRUE)
mod

## Perform test
anova(mod)

## Permutation test for F
pmod <- permutest(mod, permutations = 99, pairwise = TRUE)

[jarioksa]

This is a subtle issue: R objects are not always what they look like. We needed counts of observations for groups, and we used table() for that purpose. Although the output of table looks like an ordinary vector, it is a vector of class "table" and has a specific handling in data.frame. When we then built an internal data.frame for anova.betadisper() or permutest.betadisper the result was not like we expected.

I have a fix for this (commit a0e16ea), but we got to see that this fix does not break anything else before merging this.

Thanks for reporting this problem. We are going to have this fix in vegan release 2.4-2 before northern winter solstice.

(cc @adrianstier @bbolker)

[bbolker]

OK, I was surprised by this (thought it was an evaluation-environment problem), but I see it's just a renaming issue. Sorry about the trouble.

[jarioksa]

Fixed with commit e70050e