R help - Jul 2004 - Dendrogram plotting options?

Hi, I was wondering if there is more flexibility in the output of 
dendrograms when plotting a hclust object. I can't seem to find 
information on how to change the default output of a "hanging" style 
tree with the axis on the right to a left-to-right plot with and axis 
on the bottom. Example  code follows:

library(vegan) #loads the "vegan" module that compuptes ANOSIM
library(cluster)
community <- read.csv("ANOSIMmites.csv", header=TRUE) #reads the
data
file into memory under the name "community"
enviro <- read.csv("ANOSIMenv.csv", header=TRUE) #reads the
grouping
variables into memory
attach(enviro)
attach(community)
community.dist <- vegdist(community)
community.ano <- anosim(community.dist, plant)
summary(community.ano)
clust <- hclust(community.dist, method = "complete", members=NULL)
plot(clust, labels = sample, hang = 0.05, axes = TRUE, frame.plot = 
FALSE, ann = F, main = "Dendrogram of sample similarity", sub =
"based
on associated soil mite species", xlab = NULL, ylab = "Height")

Thanks! Mark

Mark> Hi, I was wondering if there is more flexibility in the output of 
    Mark> dendrograms when plotting a hclust object. 

This is becoming a real FAQ.
{Have you tried to find the answer in the searchable "R-help"
archives?}

    Mark> dendrograms when plotting a hclust object. I can't seem to find
    Mark> information on how to change the default output of a
"hanging" style
    Mark> tree with the axis on the right to a left-to-right plot with and
axis
    Mark> on the bottom. Example  code follows:

    Mark> library(vegan) #loads the "vegan" module that compuptes
ANOSIM
    Mark> library(cluster)
    Mark> community <- read.csv("ANOSIMmites.csv", header=TRUE)
#reads the data

we don't have that file. So why don't you rather use a
*reproducible* example as the posting guide (bottom line of
every R-help message) tells you to ?

    Mark>  <<... hence irrelevant ...>

    Mark> clust <- hclust(community.dist, method = "complete",
members=NULL)

    Mark> plot(clust, labels = sample, hang = 0.05, axes = TRUE, frame.plot =
    Mark>      FALSE, ann = F, main = "Dendrogram of sample
similarity",
    Mark>      sub = "based on associated soil mite species", 
    Mark>      xlab = NULL, ylab = "Height")

"The" answer in principle is

    d.cl <- as.dendrogram(clust, hang = 0.05)## <-- 'hang' must be
here
    plot(d.cl, ..................., horiz = TRUE)

and the "R-devel" version allows even quite a few more options to
dendrogram plotting than "current R".

Martin Maechler, Seminar fuer Statistik, ETH Zurich