[Rd] plot.TukeyHSD (PR#8229)

[ehlers@math.ucalgary.ca]

Full_Name: Peter Ehlers
Version: "R version 2.2.0, 2005-10-19"
OS: Windows XP
Submission from: (NULL) (136.159.71.162)


The newly added column of adjusted p-values in TukeyHSD output causes a problem
with plotting the confidence intervals; an extraneous vertical line segment is
plotted. 

plot.TukeyHSD expects a list of n-by-3 matrices. These are now n-by-4.

Suggested fix:
Replace:
plot.TukeyHSD <- function (x, ...)
{
    for (i in seq(along = x)) {
        xi <- x[[i]]

with:
plot.TukeyHSD <- function (x, ...)
{
    for (i in seq(along = x)) {
        xi <- x[[i]]
        xi <- xi[, -4]

I apologize if this is a known bug; I didn't see anything in my search.

Peter