Phylomorphospace projection with a fixed aspect ratio

Normally, the argument asp can be used to set the aspect ratio (basically, the relationship between real width and length in our plotting device and the units of the x and y variables that are being plotted.

Unfortunately, phylomorphospace does not allow the user to supply an argument for asp, nor can asp be set using graphics::par.

Fortunately, an easy work-around exists and that involves first opening a plotting device and setting asp to the desired value (the most common non-default value of asp desired by users is asp = 1), and then using the phylomorphospace argument add=TRUE to add our tree projection to the open device.

First, we can see the default plot and how that looks. For this example, I'll use a tree & dataset that comes from a study by Dave Collar & I (Revell & Collar 2009).

data(sunfish.tree)
data(sunfish.data)
cols<-setNames(palette()[c(4,2)],
    levels(sunfish.data$feeding.mode))
phylomorphospace(sunfish.tree,sunfish.data[,3:2],
    bty="n",colors=cols,lwd=4,
    ftype="off",node.by.map=TRUE,
    node.size=c(0,1.2),
    xlab="relative buccal length",
    ylab="relative gape width",las=1)

Next, let's redo this but setting asp=1. To do that, I'll first open an empty plot with asp=1, then I'll add my phylomorphospace projection to that device.

While I'm at it I'll also show how to add a background grid to the plot.

plot(NA,xlim=range(sunfish.data[,3]),
    ylim=range(sunfish.data[,2]),asp=1,
    axes=FALSE,xlab="relative buccal length",
    ylab="relative gape width")
axis(1,at=round(seq(-0.15,0.2,by=0.05),2),
    cex.axis=0.8)
axis(2,at=seq(-0.2,0.1,by=0.05),cex.axis=0.8)
abline(v=seq(-0.25,0.25,by=0.05),col="grey",
    lty="dotted")
abline(h=seq(-0.25,0.25,by=0.05),col="grey",
    lty="dotted")
phylomorphospace(sunfish.tree,
    sunfish.data[,3:2],colors=cols,lwd=2,
    ftype="off",node.by.map=TRUE,
    node.size=c(0,1.2),add=TRUE)
legend("bottomright",c("non-piscivorous","piscivorous"),
    pch=21,pt.bg=cols,bty="n",pt.cex=1.2,cex=0.8)
title(main="Phylomorphospace of buccal morphology\nin Centrarchidae",
    font.main=3,line=0)

The grid helps us see that we succeeded in setting asp=1 since our grid lines (which are spaced evenly) form perfect squares in our plot.

As an addendum, it is not necessary to turn of axes (axes=FALSE) and then add the x and y axes separately using axis unless you want to specify the locations of the tick-marks, as I decided to do here.

That's OK, though.

Lastly, since our data (when graphed this way) are taller than wide, let's flip the x and y axes, widen our plotting device, but still maintain asp=1.

plot(NA,xlim=range(sunfish.data[,2]),
    ylim=range(sunfish.data[,3]),asp=1,
    axes=FALSE,xlab="relative gape width",
    ylab="relative buccal length")
axis(1,at=round(seq(-0.25,0.15,by=0.05),2),
    cex.axis=0.8)
axis(2,at=seq(-0.05,0.1,by=0.05),cex.axis=0.8)
abline(v=seq(-0.25,0.25,by=0.05),col="grey",
    lty="dotted")
abline(h=seq(-0.25,0.25,by=0.05),col="grey",
    lty="dotted")
phylomorphospace(sunfish.tree,
    sunfish.data[,2:3],colors=cols,lwd=2,
    ftype="off",node.by.map=TRUE,
    node.size=c(0,1.2),add=TRUE)
legend("topleft",c("non-piscivorous","piscivorous"),
    pch=21,pt.bg=cols,bty="n",pt.cex=1.2,cex=0.8)

Cool, it worked.