R Programming

Graphics
Graphics

R has extensive graphics facilities. Some examples of

high-level graphics functions are:

• Pie Charts, Bar Charts and Histograms

• Box-and-Whisker Plots

• Scatter plots

• Time Series Plots

• Surface Plots

The quality of the graphs produced by R is often cited as a

major reason for using it in preference to other statistical
software systems.
Low-Level Graphics

The high-level graphics facilities in R are built on a set of

flexible low-level ones. This makes it possible to modify
existing plot types or to implement completely new ones.

The low level facilities include:

• Page Layout

• Setup of Plotting Coordinates

• Drawing Points and Lines

• Drawing Polygons and Rectangles

• Color Management
Simple Plotting

The plot function provides simple graphics facilities. The

function call

plot(x, y)

produces a scatter plot, with the points located at the

coordinates in the variables x, y.

• If only one argument is provided it is used as the y

variable. The x variable is assumed to have the value
1:length(y).

• If an additional argument, type="l", is provided in the

call, the points are joined with straight lines.
Example

This example show a basic scatterplot with some basic

annotation.

> x = 1:10
> y = c(0.45, 1.46, 1.94, 1.45, 2.35,
2.87, 2.39, 2.69, 3.85, 4.07)

> plot(x, y,
xlab = "A Custom x label",
ylab = "A Custom y label",
main = "A Basic Scatter Plot")
 A Basic Scatter Plot

●
4
3 ●
A Custom y label

●
●

● ●
2

● ●
1

2 4 6 8 10

A Custom x label
Range Customisation

A number of optional arguments to plot can be used to affect

the scaling of the plot and the way in which axes appear.

The arguments xlim and ylim make it possible to specify the

x and y ranges to appear in the plot. (The default behaviour is
to use the ranges of the values in the x and y arguments.)

By default the limits of the plots are expanded by 4% at the

top and bottom so that plotting symbols do not collide with the
box around the plot. This expansion can be (separately) turned
off by specifying xaxs="i" and yaxs="i".

> plot(x, y, ylim = c(0, 5), yaxs = "i",

xlab = "A Custom x label",
ylab = "A Custom y label",
main = "A Basic Scatter Plot")
 A Basic Scatter Plot

●
4

●
A Custom y label

●
●

● ●
2

● ●
1

●
0

2 4 6 8 10

A Custom x label
Different Types of Plot

• By default, the plot function produces a scatterplot by

drawing plotting symbols at the given (x, y) coordinates.
• The optional type argument makes it possible to
produce other types of plot by describing their content.
• The plot types are:
"p" points (i.e. a scatterplot)
"l" lines (i.e. a line plot)
"b" both (points and lines)
"c" just the lines from type="b"
"o" points and lines overplotted
"h" high-density needles
"s" step function, horizontal step first
"S" step function, vertical step first
"n" nothing (i.e. no plot contents)
Some plot Types

type = "p" type = "l"

●

●
● ●
●

●
●
●

type = "b" type = "o"

● ●

● ●
● ● ● ●
● ●

● ●

● ●
● ●
● ●

● ●
More Plot Types

type = "c" type = "h"

type = "s" type = "S"

Plotting Details

The plot function is and all-in-one function which carries out

a number of actions.

• It causes the graphics system to advance to a new

“frame.”

• It sets up a new plotting area using the current settings

for margin size, plot type etc.

• It sets up the transformations which determine how data

units are mapped to the plotting area.

• It plots the content of the graph (points, lines etc.).

• It plots x and y axes and box around the plot and

produces the axis labels and title.
Low-Level Graphics

The plot function produces an entire graph with a single

function call.

It can be useful to access graphics functionality at a lower

level so that it is possible to create graphs in a much more
flexible way.

The functions plot.new and plot.window are the functions

that make it possible to work in this low-level way.

plot.new is used to begin a new plot and plot.window is

used to set up coordinate systems.

Neither function does any actual drawing.

Assembling Plots From Components

Here is how the plot produced by the previous example can be

produced using plot.new, plot.window and other low-level
graphics functions.

> plot.new()
> plot.window(xlim = range(x), ylim = c(0, 6))
> points(x, y)
> axis(1)
> axis(2)
> box()
> title(xlab = "A Custom x label")
> title(ylab = "A Custom y label")

Understanding how to produce plots in this step-by-step way

makes it very easy to produce custom displays.
Example

Consider how you would produce a plot of the function

f (x) = x(2x − 3)(2x + 3)

as you might find it in a a calculus book.

The requirements for such are graph are:

(i) Axes which cross at (0, 0) rather than lying at the sides
of the plotting region.

(ii) Tickmarks chosen to avoid the path of the graph.

(iii) Tick labels which are aligned horizontally on each axis.

 10

−2 −1 1 2

−5

−10

A Cubic Polynomial
Axis Customisation

Custom displays often require customised axes. The axis

function has a number of arguments that enable customisation.

axis(side, at, labels, las)

side : Indicates which side of the graph the axis is to appear:

1 = below, 2 = left, 3 = top, 4=right.

at : A vector of values giving the locations of the tickmarks.

(The default is a set of “pretty” values spaced across the
axis range.)

labels : A vector of labels to appear at the tickmarks.

(Defaults to the values in at.)

las : Label “style,” 0 = parallel to the axis, 1 = horizontal,

2 = perpendicular to the axis, 3 = vertical.
Axis Positioning

By default an axis is positioned at the edge of the box that

surrounds a plot (on the side specified by the side argument).

As an alternative it is possible to place it through the actual

graph at a specified coordinate (y coordinate for sides 1 and 3,
x coordinate for sides 2 and 4).

This is done by specifying a value for the pos argument.

R Code To Produce The Graph

We start by producing the the data for the curve. (1001 points
may be overkill here.)
> x = seq(-1.85, 1.85, length = 1001)
> y = x * (2 * x - 3) * (2 * x + 3)
Now we set up the plot and draw the actual curve.
> plot.new()
> plot.window(xlim=c(-2, 2), ylim = c(-10, 10))
> lines(x, y)
Finally, we draw the custom axes and annotation.
> axis(1, pos = 0, at = c(-2, -1, 1, 2))
> axis(2, pos = 0, at = c(-10, -5, 5, 10),
las = 1)
> title(xlab = "A Cubic Polynomial")
Low-Level Drawing Functions

The following functions provide R’s basic drawing facilities.

points Draws points at specified locations in a plot.

lines Draws a connected set of line segments.
segments Draws (multiple) disconnected line segments.
arrows Draws (multiple) arrows.
rect Draws (multiple) rectangles.
polygon Draws a polygon.
abline Draws lines in slope/intercept form.
text Draws text in a plot.

Many statistical displays can be created using just these few

primitives.
Drawing Points

The basic function call is

> points(x, y)

where x and y contain the coordinates of the points to be

plotted.

Other (optional) arguments make it possible to change the

appearance of the points.

Things that can be changed are the plotting symbols, the size
of the plotting symbols and the colors that are used to draw the
symbols.
Customising Plotting Symbols

An optional argument called pch controls the plotting

symbol(s) used by plot.

• The argument can be either a single character, or a

numerical index into a table of symbols.

• Specifying an index of 0 produces an “invisible”

symbol.

• The pch argument is recycled to as many values as their

are points to be plotted.

The optional argument cex can be used to specify a value that

magnifies the size of the symbols.
Plotting Symbols

pch=1 pch=2 pch=3 pch=4 pch=5

pch=6 pch=7 pch=8 pch=9 pch=10

pch=11 pch=12 pch=13 pch=14 pch=15

● ● ●

pch=16 pch=17 pch=18 pch=19 pch=20

pch=21 pch=22 pch=23 pch=24 pch=25

Customising Plotting Symbol Colours

Plotting symbol colours can be customised with the col

argument.

• Colors can be specified by name with character strings;

e.g. "red", "blue", "brown", "darkseagreen", etc.

• A complete list of colour names can be obtained with

the expression colors() and viewed in the resources
section of the class website.

> xvals = 1:100

> yvals = rnorm(100)
> cols = sample(c("red", "green4", "blue"),
length(xvals), replace = TRUE)
> plot(xvals, yvals,
pch = 16, col = cols)
 3
●

● ●
●
2

●
● ●
● ●
● ● ● ●
●● ●
1

● ●● ●
● ● ●
●
● ●● ● ● ● ● ●
●● ●●
yvals

● ● ● ●
● ●
●
●
0

● ●● ● ●
● ● ● ●
● ● ●
● ● ● ● ●
● ● ●
● ●
● ● ● ●
● ●● ●
−1

● ● ●●
● ● ● ● ●
● ● ●
● ●
● ● ● ●
● ●
●
−2

● ●

● ●
●

0 20 40 60 80 100

xvals
Drawing Connected Line Segments

The function call

> lines(x, y)

connects the points whose coordinates are given in x and y

with a series of straight lines.

Other optional arguments make it possible to change the

appearance of the lines.

The texture, colour and width of the lines can be be changed.

Note: The type argument to the plot function can also be

passed to lines to generate different plot types. (This also
works for points. The two functions are really identical).
Line Types

The line type can be specified with an argument of the form

lty=type. The line type can be specified by name:

"solid", "dashed", "dotted", "dotdash",

"longdash", "twodash",

or as a hexadecimal specification of the form "DUDU", where D

represents a distance drawn with “pen down” and U is a
distance with “pen down.”
"1111" high density dots
"1313" spaced-out dots
"1333" dot-dash
"3373" short-dash, longdash
Line thickness can be set with lwd=w.
Line Type and Thickness Control

The type, lwd, lty and col arguments can be used together
to generate a variety of line effects.

> plot(rnorm(10), type="l",

lwd=4, lty="1111", col="blue",
main = "A Thick Dotted Line")

Note that in this plot, the size of dot and the gap between them
have both been scaled by the value of lwd.
 A Thick Dotted Line

2.0
1.5
1.0
rnorm(10)

0.5
0.0
−0.5

2 4 6 8 10

Index
Drawing Segments and Arrows

The functions segments and arrows take four arguments

which specify the start and end points of the arrows.

• The functions also accept a colour and line property

specification.

• The arrows function also accepts arguments which

affect the way the arrow heads are shown.

One of the most common uses of arrows is for drawing error

bars.
Example: Drawing Error Bars

We’ll randomly generate 5 groups of 20 observations each.

> groups = rep(1:5, each = 20)

> means = c(5, 7, 6, 9, 8)
> y = rnorm(100) + rep(means, each = 20)

One way to display the data is with parallel boxplots.

> boxplot(y ~ groups,

main = "Simulated Grouped Data")
 Simulated Grouped Data

11
10
9

●
8
7
6
5
4

1 2 3 4 5
An Alternative Display

Rather than using the canned boxplot display, we can create

a new display that plots the data values and shows the mean
with error bars.

> ybar = sapply(split(y, groups), mean)

> ysd = sapply(split(y, groups), sd)
> plot(groups, y, xlim = c(1, 5.25),
xlab = "Group", ylab = "Values",
main = "Data with Means and Error Bars")
> points(1:5 + .25, ybar, pch=19)
> arrows(1:5 + .25, ybar - 2 * ysd/sqrt(20),
1:5 + .25, ybar + 2 * ysd/sqrt(20),
code = 3, angle = 90, length = .1)
 Data with Means and Error Bars

11
●

●
●
●
10
●
● ●
● ● ●
●
● ●
● ●
●
9

●
● ●
● ●
● ●
●
● ●
● ● ● ●
●
8

● ●
●
Values

● ●
●
●
● ●
● ● ●
●
● ●
●
7

● ●
●
● ●
● ●
● ●
●
● ● ●
●
●
6

● ●
● ●
●
● ●
● ●
●
●
●
● ●
5

●
● ●
●
●
●
●
●
4

1 2 3 4 5

Group
Jittering the Points

It is hard to see the individual data points because of the

amount of overplotting.

This can be fixed (adding a small amount of random variation

in the x direction).

> plot(groups + runif(100, -.05, .05), y,

xlim = c(1, 5.25),
xlab = "Group", ylab = "Values",
main = "Data with Means and Error Bars")
> points(1:5 + .25, ybar, pch=19)
> arrows(1:5 + .25, ybar - 2 * ysd/sqrt(20),
1:5 + .25, ybar + 2 * ysd/sqrt(20),
code = 3, angle = 90, length = .1)
 Data with Means and Error Bars

11
●
●
●
●●
10
●
●● ●
● ● ●
●● ●
●● ●
●
9

●
● ●
●
● ●
●● ●● ●●
● ● ● ●
●
8

● ●
●
Values

●● ●
●
●
● ● ●
● ● ●●
● ●●
●
7

● ●
●
●●
● ●
● ●
● ●
●● ● ●●
●●
6

●● ●●
●● ●
●● ●
● ● ●
●●
●
● ●
5

●●
●● ●
●
●●
●
●
4

●●

1 2 3 4 5

Group
Area Fill Primitives

The graphical primitives rect and polygon can be used to

draw coloured areas.

• The functions accept arguments col and border which

can be used to specify the interior and border colours of
the regions.

• The functions also accept texture specifications for the

border lines via lty and lwd arguments.

• The rect function expects arguments giving the

locations of diagonally opposite sides of the rectangle
(i.e there are 4 arguments, x0 , y0 and x1 , y1 ).

• The polygon function expects two arguments giving

the x and y coordinates of the polygon vertices.
A Display Created With Rectangles

10
8
6
4
2
0

−10 −5 0 5 10
Creating the Display

The figure at the center of the array is drawn with a single call
to rect. It creates 10 rectangles, going bottom to top.

The full set of rectangles is drawn by the code

xl = -(10:1)
xr = 10:1
yb = 0:9
yt = 1:10
rect(xl, yb, xr, yt, col = "tan")

To draw the complete plot, all that needs to be done is to set up

the plot with plot and then to add the rectangles.
Drawing Rectangles

The following code draws the full display.

> plot(c(-10, 10), c(0, 10), type = "n",

ylab = "", xlab = "")
> xl = -(10:1)
> xr = 10:1
> yb = 0:9
> yt = 1:10
> rect(xl, yb, xr, yt, col = "tan")
Dropping Axes from a Plot

The plot can be improved by dropping the axes by specifying

axes=FALSE and adding a title.

The code can also be shortened by inlining the corner

coordinates.

> par(mar = c(0, 0, 4.1, 0))

> plot(c(-10, 10), c(0, 10), type = "n",
axes = FALSE,
ylab = "", xlab = "",
main = "The Great Ziggurat of Ahr")
> rect(-(10:1), 0:9, 10:1, 1:10, col = "tan")
The Great Ziggurat of Ahr
Polygons

Polygons provide a useful component in many graphical

displays.

Examples:

• As an approximate way of drawing (filled) circles and

ellipses.

• For drawing rotated squares and rectangles.

• For drawing slices in pie graphs.

• As a way of showing areas under curves.

• Etc.
A Custom Time Series Display

The following display shows a time series display where the

area below the series is filled with a solid colour.

There is no built-in way of creating this kind of display in R.

Using the ability to add filled polygons to a plot makes it

possible.
 Global Temperature Anomaly with 1951−1980 Base

0.6
0.4
Temperature Anomaly

0.2
0.0
−0.2

1940 1950 1960 1970 1980 1990 2000

Year
Producing the Display

The graph is produced in two stages:

• Setting up the plot and drawing the axes and annotation.

• Adding the filled polygon area.

The polygon is drawn going below the x axis and the part
below the axis is “clipped” out.

> plot(year, temp, xaxs = "i", type = "n",

xlab = "Year", ylab = "Temperature Anomaly",
main = "Global Temperature Anomaly with 1951-19
> polygon(c(min(year), year, max(year)),
c(-1, temp, -1), col = "darkseagreen")
Choosing Colours to Fill Areas

Choosing a set of colours to fill large areas in a plot is an

important task in statistical graphics.

Choosing colours by hand from a colour chart is a very hard

task unless the person choosing has a real flair for the task
(most combinations look ghastly).

R has a special function (called hcl) which is designed to

make this task easier.

The function allows a user to specify colours by hue

(dominant colour wavelength), chroma (colourfulness), and
lightness (brightness).
Six Reasonable Colours

In the function call

hcl(h)

the parameter h specifies hue as an angle (in degrees) around

the colour circle (0 = red, 60 = yellow, 120 = green, 180 =
cyan, 240 = blue, 300 = magenta).

These equally spaces colours work harmoniously together and

the default chroma and lightness values produce results that
are easy to look at.

It is also possible to specify transparency with an optional

parameter alpha. This makes it possible to “look though”
coloured areas and see the objects behind them. (Here 0
means fully transparent and 1 means opaque.)
0 60 120 180 240 300

hue
Other Ways to Specify Colour

The functions rgb and hc also make it possible to specify

colour programmatically. The calls are

rgb(r, g, b)
hsv(h, s, v)

with all the arguments in the range [0, 1].

In the case of rgb the arguments specify the intensity of red,

green and blue making up the colour.

In the case of hsv the arguments specify the hue, saturation

and value of the colour.

Both of these functions allow the specification of transparency

with an optional alpha argument.
Circles

The equation of a circle with radius R centered at (0, 0) is

(R cos θ , R sin θ ), 0 ≤ θ ≤ 2π.

This circle can be approximated by a polygon which has

vertexes equally spaced around the circle.

By using enough vertexes it is possible to produce a polygon

which is visually indistinguishable from a circle.

Changes of direction greater that 5◦ are visually detectable, so

using at least 72 equally-spaced vertexes is necessary.
Drawing Circles with R

Here is some simple code to draw a coloured circle.

> theta = seq(0, 2 * pi, length = 73)[1:72]

> x = cos(theta)
> y = sin(theta)
> plot.new()
> plot.window(xlim = c(-1, 1),
ylim = c(-1, 1),
asp = 1)
> polygon(x, y, col = "darkseagreen")

The specification asp = 1 is is important. It ensures that

equal coordinate changes in the x and y directions are
represented as equal distances on the page.
Rosettes

A rosette is a figure obtained by taking n points equally spaced

around the circumference of circle, and joining every point to
every other point.

> n = 17
> theta = seq(0, 2 * pi, length = n + 1)[1:n]
> x = sin(theta)
> y = cos(theta)
> plot.new()
> plot.window(xlim = c(-1, 1),
ylim = c(-1, 1),
asp = 1)
> for(i in 2:n)
for(j in 1:(i - 1))
segments(x[i], y[i], x[j], y[j])
> title(main="A 17-Rosette")
A 17−Rosette
Spiral Squares

Here is how to create an interesting figure by rotating and

shrinking a square multiple times.
> plot.new()
> plot.window(xlim = c(-1, 1),
ylim = c(-1, 1), asp = 1)
> x = c(-1, 1, 1, -1)
> y = c(-1, -1, 1, 1)
> n = 51
> i1 = 1:4
> i2 = c(2:4, 1)
> for(i in 1:n) {
polygon(x, y)
x = .9 * x[i1] + .1 * x[i2]
y = .9 * y[i1] + .1 * y[i2]
}
> title(main="A Spiral Square")
A Spiral Square
Adding Color

Using a colour gradient can provide more “visual impact.” In

this case we can use a gradient from red to cyan generated by
hsv function with

col = hsv(seq(0, 2/3, length = n))

and the passing the i-th colour to the polygon function with

polygon(x, y, col = col[i])

(In this example the s and v arguments take their default value
of 1.)
A Spiral Square
Drawing Text in Plots

The text function can be used to draw character strings

within a plot. The basic call to text has the form:

text(x, y, labels)

where x and y specify the locations the strings are to be drawn

at and labels contains the actual strings.

Other optional arguments make it possible to specify how the

strings are drawn; adj controls the string “adjustment,” srt
controls rotation of the string, col controls colour, cex
controls magnification and font specifies which font to use.
String Adjustment and Rotation

The adj parameter is a vector containing one or two values.

The first value specified the x justification and the second the y
justification.

A value of 0 indicates left-justification, a value of .5 indicates

centred text and a value of 1 indicates right justification.
(Intermediate values are also possible, but less used.)

The value of srt specifies the string rotation in degrees

counter-clockwise from horizontal.

String rotation occurs before justification.

 st
)
.5

r=
,0

−4
5
0.

5,
c(

ad
j=

j
ad

=
c(
,

0.
45

5
r=

,0
.5
st

)
str = 0, adj = c(0.5 ,0.5)

str = 0, adj = c(0 ,0.5) str = 0, adj = c(1 ,0.5)

str = 0, adj = c(0.5 ,0)

Drawing Text in Plot Margins

The function mtext can be used to draw text in the margins of

a plot. The basic call has the form

mtext(text, side, line, at)

The argument text contains the strings to be written, side

specifies which margin the text is to appear in (default 3),
line specifies how many lines out form the plot the text is to
appear and at contains the locations that the text is to be
drawn at.

Other optional arguments affect how the text is drawn. These

include cex, col and font.

Other arguments have effects specfic to mtext.

Some Margin-Text Specific Parameters

Several parameters have a effect specific to mtext.

las : This specifies the label orientation. A value of 0 means

always parallel to the axis, 1 means always horizontal, 2
means always perpendicular to the axis and 3 means
always vertical.

adj : the adjustment for each string in reading direction.

padj : the adjustment for each string perpendicular to the

reading direction.

outer : If true, the string is plotted in the outer margin of the

plot.
Marginal Text Examples

Here are a few examples showing what is possible with mtext.

> nums = c("One", "Two", "Three")

> lets = c("ABC", "DEF", "GHI")
> plot.new()
> plot.window(xlim = c(1, 3), ylim = c(1, 3))
> box()
> mtext(nums, side = 1, line = c(1, 2, 1),
at = 1:3)
> axis(2, at = 1:3, labels = FALSE)
> mtext(lets, side = 2, line = 1,
at = 1:3, las = 2)
> axis(3, at = 1:3, labels = FALSE)
> mtext(nums, side = 3, at = 1:3,
line = 1, las = 3)
 Three
One

Two
GHI

DEF

ABC

One Three
Two
Eye-Catching Visual “Tricks”

In commercial presentation graphics it is common to add a

variety of effects to make the the graphics more eye-catching.

Most R graphics do not use these tricks, but by building plots

incrementally it is possible to liven up plots.

We’ll look at just a few such effects.

Filling the Plot Background

It is simple to fill the rectangular background region of a plot.

This is done by drawing a rectangle of the appropriate size.

The function par makes it possible to get (and set) many

properties associated with a plot.

The call par("usr") returns the coordinates of the central

region of a plot as vector of values

xleft , xright , ybottom , ytop .

Example

A simple filled background.

> x = 1:30
> y = rnorm(30) + x/5
> plot.new()
> plot.window(xlim = range(x),
ylim = range(y))
> usr = par("usr")
> rect(usr[1], usr[3],
usr[2], usr[4], col = "lemonchiffon")
> lines(x, y, lwd = 3)
> points(x, y, pch = 21, bg = "white")
> axis(1)
> axis(2)
> title(main = "A Filled Plot Region")
> title(xlab = "X Coordinates")
> title(ylab = "Y Coordinates")
 A Filled Plot Region

● ●
● ●
5
●
●
●
●
●
●
4

●
● ●
●
Y Coordinates

●
3

● ●
2

● ● ● ●
●
●
●
1

● ●
0

●
−1

0 5 10 15 20 25 30

X Coordinates
Drop Shadows

One simple effect is the addition of drop shadows for objects

that appear in the graph.

To create a shadow effect the object casting the shadow is

drawn twice; first as a shadow and then as the real thing.

The shadow is usually drawn in gray, offset down and to the

right.

The functions xinch and yinch can be used to produce

appropriate offsets in the x and y directions.
R Code

As an example, consider drawing a smooth line through a set

of points.

The coordinates of the smooth can be obtained with the lowess

scatterplot smoother.

> smo = lowess(x, y)

The shadow for the line can be drawn with the expression

> lines(smo$x + xinch(.4),

smo$y - yinch(.4),
col="gray", lwd = 5)

and the line itself with

> lines(smo$x, smo$y, lwd = 5)

 A Plot with Drop Shadows

●
●
6 ●
●
●
● ●
●
● ●
5 ●
●
●
●
● ●●● ●
●
4 ●
●
●
●
●
●●●●
●●●
3 ●
● ●
●
● ●
●
● ●
●
2 ●
● ● ●
●
●

1 ●
●
● ●
●
●
0 ●
● ●
●
●
●

0 5 10 15 20 25 30

The points have been magnified and given shadows too.

 New Zealand Meat Consumption

Lamb

Mutton

Pigmeat

Poultry

Beef

0 10 20 30 40 50

Percentage of Meat Eaten

Line Thickness and Colour Manipulation

By manipulating the properties of lines it is possible to create

interesting effects.

In the following plot, a green line is drawn with a black

outline and a white “specular highlight.”

> plot(x, y, type = "n")

> lines(x + xinch(.04), y - yinch(.04), lwd = 7,
col = "gray")
> lines(x, y, lwd = 7)
> lines(x, y, lwd = 4, col = "green4")
> lines(x, y+yinch(.02), lwd = 1, col = "white")
> title(main = "A 3D Line with Drop Shadow")
 A 3D Line with Drop Shadow

6
5
4
3
y

2
1
0

0 5 10 15 20 25 30

x
Textured Line Example

By manipulating the properties of lines it is possible to create

interesting effects.

In the following plot, a green line is drawn with a black

outline and a white dashed line is drawn over the top.

> plot(x, y, type = "n")

> lines(x, y, lwd = 7)
> lines(x, y, lwd = 4, col = "green4")
> lines(x, y, lwd = 4, lty = "33", col = "white")
 6
5
4
3
y

2
1
0

0 5 10 15 20 25 30

x
Colour Gradients

Colour gradients can be produced in R by superimposing

slightly offset polygons with small colour changes.

In the following example the flame effect is obtained by

drawing a series of identical polygons. Each polygon is
slightly lower and slightly redder than the one before.

In this case the colours were generated using hsv.

This example should be filed under “propaganda.”

 Annual Global Temperature Increase (°C)

0.6

0.4

0.2

0.0

−0.2

1960 1970 1980 1990 2000

Time
The par Function

R graphics are controlled by use of the par function.

par makes it possible to control low-level graphics by

querying and setting a large set of graphical parameters.
Graphical parameters control many features such as:

• the layout of figures on the device

• the size of the margins around plots

• the colours, sizes and typefaces of text

• the colour and texture of lines

• the style of axis to be used

• the orientation of axis labels

The Layout of Graphics

Graphs appear on R graphics devices as a series of rectangular

graphical figures.

Each figure consists of a rectangular plot region surrounded by

four margins.

Each element of a figure can be described as being in either

the plot region or the margins.

The sides of a plot are numbered 1 through 4.

 Margin 3

Margin 2

Margin 4
Plot Region

Margin 1
Controlling The Margins

The sizes of the margins can be changed by making a call to

the function par before calling plot.new.

There are several possibilities:

1. Set the margin sizes in inches.

> par(mai=c(2, 2, 1, 1))

2. Set the margin sizes in lines of text.

> par(mar=c(4, 4, 2, 2))

3. Set the plot width and height in inches.

> par(pin=c(5, 4))

Querying the Margin Sizes

The par function can be used for querying the margin sizes.

The margin sizes in lines and inches can be obtained as

follows:

> par("mar")
[1] 5.1 4.1 4.1 2.1

> par("mai")
[1] 1.02 0.82 0.82 0.42

and the the plot dimensions in inches as follows:

> par("pin")
[1] 5.76 5.16
Multifigure Layouts

par can be used to set up arrays of figures on the page. These

arrays are then filled row-by-row or column-by-column.

The following example declares a 2 × 2 array to be filled by

rows and then produces the plots.

> par(mfrow=c(2, 2))

> plot(rnorm(10), type = "p")
> plot(rnorm(10), type = "l")
> plot(rnorm(10), type = "b")
> plot(rnorm(10), type = "o")

A 2 × 2 array to be filled by columns would be declared with

> par(mfcol = c(2, 2))

 ●
●
1
● ●
● ●
rnorm(10)

rnorm(10)

1.0
0 ● ●

0.0
−2

2 4 6 8 10 2 4 6 8 10

Index Index

● ● ●
1

● ● ●
● ●

0.5
● ●
●
rnorm(10)

rnorm(10)
0

●
●
−0.5
●
−2

● ●
●
−1.5

● ● ●

2 4 6 8 10 2 4 6 8 10

Index Index
Outer Margins

Arrays of figures can also be created with a set of outer

margins.

The size of outer margins can be specified with the oma or omi
arguments to par.

Here is an example which shows a 2 × 2 array of figures

surrounded by an outer margin with space for 4 lines of text on
all sides.

> par(oma = c(4, 4, 4, 4),

mfrow = c(2, 2))
 Outer Margin 3

Figure 1 Figure 2
Outer Margin 2

Outer Margin 4
Figure 3 Figure 4

Outer Margin 1
Eliminating Waste Margin Space

It can be useful to eliminate redundant space from multi-way

arrays by trimming the margins a little.

> par(mfrow=c(2, 2))

> par(mar = c(5.1, 4.1, 0.1, 2.1))
> par(oma = c(0, 0, 4, 0))
> plot(rnorm(10), type = "p")
> plot(rnorm(10), type = "l")
> plot(rnorm(10), type = "b")
> plot(rnorm(10), type = "o")
> title(main = "Plots with Margins Trimmed",
outer = TRUE)

Here we have trimmed space from the top of each plot, and
placed a 4 line outer margin at the top of of the layout.
 Plots with Margins Trimmed

0.5
●

1.0 ●
rnorm(10)

rnorm(10)

−0.5
●
● ●
0.0

●
●
●

−1.5
●
−1.5

2 4 6 8 10 2 4 6 8 10

Index Index

1.5
● ● ●
● ●
0.5

● ● ●
●

0.5
●
rnorm(10)

rnorm(10)
●
0.0

●
●

−0.5
●
● ●
●
●
−1.0

−1.5

● ●

2 4 6 8 10 2 4 6 8 10

Index Index
An Example: Scatterplot Matrices

A scatterplot matrix is a multidimensional generalisation of a

simple scatterplot.

Given n variables, a scatterplot matrix presents scatterplots of

every possible variable in the set against every other possible
variable in the set.

The results are displayed in an n × n matrix, with the element

in the i j-th position being the plot of variable i against variable
j.

Since a plot of variable i against variable i is relatively

uninteresting, something more useful can be displayed in that
position.
 Anderson's Iris Data
2.0 3.0 4.0 0.5 1.5 2.5
● ● ●
● ● ● ● ● ● ●● ●●●●
●

4.5 5.5 6.5 7.5

●
●● ● ●
●
● ● ● ●●●● ● ●
● ● ●
●●
● ●●● ● ●● ●● ●
● ●●●●●●●
● ●●●●●● ●●●●● ●
●● ●● ●● ●
●●●
●
● ● ● ●● ●● ● ● ●●● ● ● ●
●●●
● ●●●● ● ●●● ●● ●
● ● ●
●● ●
●●
● ●●●●●
●● ●●●●
●● ● ●●● ●● ●● ● ●
● ● ● ●
● ●●●
●●
●
● ● ●
●
●● ●●●●●
●● ● ● ●●●● ●
●● ●
●
Sepal.Length ●●
●●●
●●
●●
● ● ● ● ● ●● ● ●●●
●
●● ●●
●
●
●
● ●●● ●
●●●
●●
● ●● ●
●
●●●●●● ● ● ●
● ●●
● ● ●
● ●● ●
● ● ● ●
●
●●●
●
●● ● ●● ●
● ● ●● ●
● ● ●● ● ●● ● ●● ● ●●
●●●● ● ●
● ●●● ●●●●●
●
●●●
●●●● ●●
●
●●
●●
●●
● ●
●●● ● ●
●●● ● ● ●
●
●
●●
●●
● ● ●
●●
●
●●
● ●
● ●● ● ● ●●●● ●●
●
●●
● ● ●●
● ●●
● ● ●
● ● ●
4.0

● ● ●
● ● ●
● ●● ●
● ● ●● ●●
●● ●● ●●● ● ●
● ●● ● ● ●
● ●● ● ●● ● ●● ●
●●● ● ●
●●● ●● ●
● ● ●●● ● ● ●● ●●●
●● ● ●● ●●● ● ●●
●● ● ● ●● ● ●● ● ● ● ● ●
● ●● ● ● ●● ●●● ● ●●
●● ●●● ●● ●● ● ● ●● ● ● ●
3.0

● ●● ● ●●
●● ●●● ● ●● ●●● ●●●● ●● ●● Sepal.Width ●●
●
●●
●
●
●●
● ●●● ●
●
●● ●
●●
●
●
●●● ●● ● ●●●
●● ●● ● ● ●●
●●●●●●● ●●●●
● ●● ●●●●● ● ● ● ● ● ●●●● ● ● ● ●●● ●
●●● ●●●●● ● ● ● ●●● ●●●●● ● ● ● ●●●● ●●●●● ●
● ● ● ● ●● ●●● ●●● ● ●●● ● ●●
● ●● ● ● ● ●● ● ● ● ● ● ●
● ● ●●● ● ● ● ●● ●● ● ● ● ● ● ●●●●
● ● ●● ● ●●
● ● ● ● ● ● ●● ● ● ●
●● ● ● ● ● ●
2.0

● ● ●

1 2 3 4 5 6 7
● ● ● ●●●●
●● ● ● ●
●
●●● ● ●●● ● ●●● ● ●
● ● ●● ●● ● ●
●● ●
● ● ●● ●
●● ●
●●●● ● ●
●●● ●● ●●●●●●● ● ● ●
● ●●
● ●●●
●●●
●●●● ●● ●●● ●
●● ●●●
●● ●
●●
●
●●●●
●●
● ●
●●
●●●
●●●●●●●●
● ● ●
● ●●
●
●
●
●
●●
●●
●●
●
●●
●●● ●●●
●●●●
●
● ●●
●
●●● ●
●●● ● ● ●
● ●●●● ●● ●● ●● ●● ●● ●
●●●
●●
● ● ●
●●
●●
● ●
●
● ●●●
●●
● ●●●●● ●
●●
● ●●●● ●
●●
●●●
●●
●●● ●●
●●●
● ●
●●
● ●
●● ● ● Petal.Length ●●●●
●
●● ● ● ●● ● ●
●
● ● ●

● ●●●●●
● ● ●●
●● ●
●● ●
●●●
●●●
●●●●●
●●●
●●
●●
●● ● ●●
●● ●● ●●
●●●●
●●●
● ●
●● ●
●●●●●●●● ●
● ●
●●
●
●●
●●●
●●
●
●●●●● ● ● ●
● ● ● ●●
●
●
●
2.5

● ● ● ● ● ●●●
● ● ● ● ● ● ● ●
● ● ●●● ● ● ●●● ● ●
●●● ●●● ●
●● ● ● ● ● ●● ●
● ●●● ● ● ● ●● ● ●
●●●● ●
●● ● ●● ● ● ● ● ● ●●●
● ●●
● ●● ● ● ●● ●●● ●
●●●●●●● ● ●● ● ●●●●●● ●
●● ● ●●● ●
● ● ● ● ● ●
1.5

● ● ● ● ● ●● ●● ● ●
● ● ●● ●●●● ● ● ● ● ●●●●● ●● ●●●●●
● ● ●●● ● ●●●●●●● ● ●● ●● ●
●●● ●●●● ● ● ● ●●●● ●● ●
●●●
●●
● ●●
● ●● ●
●●
●●● ●
● ●
●
●
●●● ● Petal.Width
●● ● ●● ● ● ●●● ●● ●●● ●●
0.5

● ● ●
● ● ●
●● ● ● ● ●●● ● ●● ●
●●
●● ● ●● ● ● ● ●● ● ●
●●
●●
● ●●●●●●●●●● ● ●●●●
●●●●●●●● ● ● ●●●
●●●
●●
● ●● ● ●● ● ● ●●●

4.5 5.5 6.5 7.5 1 2 3 4 5 6 7

Setting Up a Scatterplot Matrix

To create a scatterplot matrix we need to arrange a series of

plots on the page.

The basic arrangement is to have the n2 plots arranged in an

n × n array.

To ensure that the plots are properly juxtaposed, we need to

eliminate the margins around the individual plots but to create
an outer margin around the whole array to hold the axes and
an overall title.

par(mfrow = c(ncol(x), ncol(x)),

mar = rep(0, 4),
oma = c(4.1, 4.1, 5.1, 4.1))
Saving and Restoring Parameter Values

Since the layout parameters for the scatterplot matrix only

apply to that matrix, it is useful to be able to revert back to the
previous set of values when the plot is complete.

The par function returns the previous values of any

parameters which are changed by a a call to it and these can be
save for later use.

opar = par(mfrow = c(ncol(x), ncol(x)),

mar = rep(0, 4),
oma = rep(4.1, 4))

The values can be restored later with another call to par.

par(opar)
Drawing The Scatterplots

Once the layout has been created it is easy to produce the

individual scatterplots.

for(i in 1:n)
for(j in 1:n) {
xrange = range(x[,j])
yrange = range(x[,i])
plot.new()
plot.window(xlim = xrange, ylim = yrange)
if (i == j)
text(mean(xrange), mean(yrange),
colnames(x)[j])
else points(x[,j], x[,i])
box()
}
Drawing Axes

Clearly, not every scatterplot in the matrix has axes and we

need a computation to determine which (if any) axes should be
drawn.

Only axes at the edge of the plot are drawn, and then only
alternate axes are drawn to avoid label collisions.

if (i == 1 && j %% 2 == 0)
axis(3)
if (i == n && j %% 2 == 1)
axis(1)
if (j == 1 && i %% 2 == 0)
axis(2)
if (j == n && i %% 2 == 1)
axis(4)
Printing a Title

If we want to have an overall title on the plot we need to

ensure that there is enough space in the top outer margin.

opar = par(mfrow = c(ncol(x), ncol(x)),

mar = rep(0, 4),
oma = c(4.1, 4.1, 5.1, 4.1))

We can then produce the title using the mtext function.

mtext(main, side = 3, line = 3,

font = 2, outer = TRUE)

This draws the text in the (middle of the) third line of the third
outer margin, in bold font.
Final Cleanup

For general hygiene it is important to reset any parameters

before leaving the function.

par(opar)
Customisation

The scatterplot matrix function we’ve described is very

inflexible.

We could customise it by adding of optional arguments.

pairs = function(x, col = 1, pch = 1, ...)

and then passing these arguments on to the points function.

points(x[,j], x[,i], col = col, pch = pch)

Panel Functions

The heart of the scatterplot matrix function is the call

points(x[,j], x[,i])

We can make the function highly customisable by making the

function which is invoked at this point be an argument to the
function.

pairs = function(x, y, ...,

panel = function(x, y, ...)
pairs(x, y, ...))

and invoking this function in the body of the function.

Panel Function Example

Here is an example which shows a panel function which plots

the points and then superimposes the regression line between
the pair of variables.

pairs(x,
panel = function(x, y, ...) {
points(x, y)
abline(lm(y ~ x))
})
Using par

The par function can be used to control margins and the

layout of plots on the page, using the parameters mar, mai,
pin, oma, omi, mfrow and mfcol.

This is just the tip of the iceberg; there are 71 graphics

parameters in all.

You can view the settings of the parameters by typing

> par()

and view their names by typing

> names(par())
The Full Set of Graphical Parameters
xlog ylog adj ann ask
bg bty cex cex.axis cex.lab
cex.main cex.sub cin col col.axis
col.lab col.main col.sub cra crt
csi cxy din err family
fg fig fin font font.axis
font.lab font.main font.sub gamma lab
las lend lheight ljoin lmitre
lty lwd mai mar mex
mfcol mfg mfrow mgp mkh
new oma omd omi pch
pin plt ps pty smo
srt tck tcl usr xaxp
xaxs xaxt xpd yaxp yaxs
yaxt
Important Groups of Graphics Parameters

Graphics parameters come in groups (usually with related

names). Important groups correspond to names as follows:
cex An abbreviation for character expansion factor.
This provides a magnification of text relative to the
default.
col The colour to be used for text or graphics elements.
font The font to be used for drawing character strings.
Controlling Font Size

Setting the value of cex controls the default size of text

elements and symbols in graphs. A value of cex=2 makes text
twice its standard size.

More specialized cex values are as follows:

cex.main The expansion factor for the main title.

cex.lab The expansion factor for xlab and ylab.
cex.axis The expansion used for the text printed at axis
tickmarks.
cex.sub The expansion used for plot subtitles (sub).
Controlling Colour

Setting the value of col controls the default colour of text

elements and symbols in graphs.

More specialized col values are as follows:

col.main The colour for the main title.

col.lab The colour for xlab and ylab.
col.axis The colour used for the text printed at axis
tickmarks.
col.sub The colour used for plot subtitles (sub).
Controlling Fonts

Setting the value of font controls the default font of the text
elements in graphs.

More specialized font values are as follows:

font.main The font for the main title.

font.lab The font for xlab and ylab.
font.axis The font used for the text printed at axis tickmarks.
font.sub The font used for plot subtitles (sub).

The value of font is 1 for normal font, 2 for bold, 3 for italic
and 4 for bold-italic.
An Example

Many par values can be passed as arguments to plotting

functions or passed to par to set the values for all future plots.

> plot(rnorm(100), col = "red", cex = 2,

font.axis = 1, col.axis = "blue",
font.main = 4, col.main = "green4",
font.lab = 3, col.lab = "brown",
cex.main = 2, main = "Par Overload")

> par(col = "red", cex = 2,

font.axis = 1, col.axis = "blue",
font.main = 4, col.main = "green4",
font.lab = 3, col.lab = "brown",
cex.main = 2)
 Par Overload
●
2 ● ●
● ●
●● ●
● ●●
● ● ●● ●●
● ● ● ● ● ●
1

● ● ●
●
● ●●● ● ● ●● ● ●
rnorm(100)

● ●● ●● ● ●●
● ●● ● ● ● ●
●●● ●
● ●● ●●
0

●
● ● ● ●● ● ●●● ●
● ●● ● ●● ●
● ● ●
●
●● ● ●
● ●
−1

● ● ● ● ●
● ●
● ●●
●
−2

0 20 40 60 80 100

Index
All Those Other pars . . .

There are many other par values which we have not

discussed.

You get a description of all the par values by looking at the

documentation for par.

> ?par

> help("par")
More Flexible Layouts

The multifigure layouts which can be specified with par are

very rigid.

All the panels in the plot have exacly the same size.

The layout function provides an alternative way of producing

multiple figures.

The function still imposes constraints, but they only require

that the widths of figures in the same column be equal and that
the heights of figures in the same row be equal.
1 2 3

4 5 6

7 8 9
Combining Figures

The arrangements produced by layout are a little more flexible

because adjacent figures can be combined.

In the next arrangement, the 4th and 5th figures of the previous
layout are combined into a single figure.

Beware, however, that attempts to combine non-adjacent

figures produces very strange results.
1 2 3

4 5

6 7 8
Arguments to Layout

There are three main arguments to layout; mat, widths and

heights.

The mat argument associates a figure number with each

element of a rectangular layout.

The first example layout had the following mat argument.

> layout(matrix(1:9, nc = 3, byrow = TRUE),

... )

The second layout combined the 4-th and 5-th figures from
this layout as follows

> layout(matrix(c(1:4,4:8), nc = 3, byrow = TRUE),

... )
Arguments to Layout

The widths of the columns in the arrangement are specied by

the widths argument.

These can either be relative widths - e.g. c(1, 2) specifies

that the second column is twice as big as the first, or they can
be absolute widths specified in centimetres.

The specification

widths = c(lcm(1), 1, 2)

says that the first column is exactly 1cm wide and the third
column is twice the width of the second.

The row heights are specified by heights in a similar fashion.

Layouts and Margins

It is important to note that layout ignores the outer margin

parameters set with par.

All par values which work within figures (such as mar and
mai) still have an effect.

It can be useful to set figure margins to 0 when working with

sets of figures which are to be juxtaposed.

The effect of outer margins can be obtained by specifing extra

rows and columns in the mat argument.

These extra rows and columns can be associated with “figure

0” so that no plotting is done in them.
Example: An Augmented Scatterplot

We’ll illustrate what layouts can be used for with a specific

example.

The example will produce a standard scatterplot which is

enhanced with boxplots of the marginal distributions of the x
and y variables.
 An Enhanced Scatterplot
● ● ●

● ●

3
●
●
2 ● ●
●● ●
● ● ● ●
●
● ●● ● ●
●
1 ● ●
● ● ●●
● ● ●
● ● ● ●
● ● ● ●
y

● ● ● ● ●
● ●● ●●
0 ● ● ● ●●● ● ● ● ● ●
● ●● ●● ●
● ●● ● ● ● ● ●
● ● ● ●
● ● ●
●● ● ● ●
● ● ●
●
−1 ● ●
●
●
●
● ● ●
−2 ●
●

● ●
−3
● ●

−3 −2 −1 0 1 2
Choosing a Layout

We need a big figure for the scatterplot, regions along its top
and right for the scatterplots and a figure at the very top for a
title.

> layout(rbind(c(0,4,4,0),
c(0,2,0,0),
c(0,1,3,0),
c(0,0,0,0)),
height = c(lcm(2), lcm(2), 1, lcm(2)),
width = c(lcm(2), 1, lcm(2), lcm(1)))
> layout.show(4)
> box("outer", lty = "dotted")
 4

1 3
Producing the Graph

We constrain the margins of the plots to be zero so that the

plots are juxtaposed. Then we produce the plots, in the order
specified in mat. The last plot is a simple one which contains
just the title.

> par(mar = rep(0, 4), cex = 1)

> plot(x, y, las = 1)
> boxplot(x, horizontal = TRUE, axes = FALSE)
> boxplot(y, axes = FALSE)
> plot.new()
> plot.window(xlim = c(0, 1), ylim = c(0, 1))
> text(.5, .25, "An Enhanced Scatterplot",
cex = 1.5, font = 2)
 An Enhanced Scatterplot
● ● ●

● ●

3
●
●
2 ● ●
●● ●
● ● ● ●
●
● ●● ● ●
●
1 ● ●
● ● ●●
● ● ●
● ● ● ●
● ● ● ●
y

● ● ● ● ●
● ●● ●●
0 ● ● ● ●●● ● ● ● ● ●
● ●● ●● ●
● ●● ● ● ● ● ●
● ● ● ●
● ● ●
●● ● ● ●
● ● ●
●
−1 ● ●
●
●
●
● ● ●
−2 ●
●

● ●
−3
● ●

−3 −2 −1 0 1 2
A Two-Panel Plot with Labelling

Suppose that (for some reason) we want a plot consisting of

two panels with a strip for labelling between them.

How do we arrange this?

> layout(matrix(c(0,0,0,0,0,
0,1,3,2,0,
0,0,0,0,0), nc = 5, byrow = TRUE),
widths = c(lcm(2), 1, lcm(2), 1, lcm(2)),
heights = c(lcm(2), 1, lcm(2)))
> layout.show(3)
> box("outer", lty = "dotted")

Note that this just a start. You need extra panels for the
labelling at the top and bottom, and some space for the axes to
appear in.
1 3 2
 New Zealand Population (1996 Census)
Male Female
95+
90−94
85−89
80−84
75−79
70−74
65−69
60−64
55−59
50−54
45−49
40−44
35−39
30−34
25−29
20−24
15−19
10−14
5−9
0−4

4 3 2 1 0 0 1 2 3 4
Percent of Population
Layouts and Plots

We’ve seen that layouts provide a way to partition the page

into a number of regions which can be used to draw multiple
figures or a single plot with multiple components.

One simple approach is to use a high-level R plotting function

such as plot, hist, barplot or boxplot, and to let the high
level function take care of all the details.

An alternative, much more flexible, approach is to divide the

page into subregions which each forms a single component of
a graph.

It is useful to create a software component to match each of

the types of region which will appear in the graph.
Handling Margins

Margins are added to plots to provide space for axes and

annotations such as axis labels and titles.

Under the component model, such elements are represented by

separate plotting areas, so the need for margins is gone.

When working with layouts in this way it can be very useful to

remove the margins completely with a specification of the
form

> opar = par(mar = rep(0, 4))

Example: A Simple Scatterplot

We can build a layout for a simple scatterplot as follows:

> layout(matrix(c(0, 0, 0,
0, 1, 0,
0, 0, 0), nc = 3, byrow = TRUE),
widths = c(lcm(2), 1, lcm(2)),
heights = c(lcm(2), 1, lcm(2)))
> layout.show(1)
1
Choice of Margin Size

While it is possible to specify the margin size in centimetres, it

is much more natural to specify it in lines of text (just like the
mar and oma parameters).

By defining the function

> llines =
function(x)
lcm(x * par("csi") * 2.54)

it is possible to specify margin size in lines of text.

A margin of 2.1 lines is just enough to fit the standard axis

(with a small amount of extra space).
Example: A Simple Scatterplot

We can build a layout for a simple scatterplot, using just the

minimum of space needed for the axes as follows:

> par(mar = rep(0, 4))

> layout(matrix(c(0, 0, 0,
0, 1, 0,
0, 0, 0), nc = 3, byrow = TRUE),
widths = c(llines(2.1), 1, llines(1.1)),
heights = c(llines(1.1), 1, llines(2.1)))
> par(cex = 1)
> plot(rnorm(10), ann = FALSE)
 ●

●
1.0

●
●

●
0.5

●
0.0

●
−0.5

2 4 6 8 10
Axis Labels

Now suppose that we want to add axis labels to the scatterplot.

This means adding another line at the top and left of the
layout.

> par(mar = rep(0, 4))

> layout(matrix(c(0, 0, 0, 0,
3, 0, 1, 0,
0, 0, 0, 0,
0, 0, 2, 0),
nc = 4, byrow = TRUE),
widths = c(llines(2.1), llines(2.1),
1, llines(1.1)),
heights = c(llines(1.1), 1,
llines(2.1), llines(2.1)))
> par(cex = 1)
Producing the Labels

The plot labels are produced by moving to the region for the
label (with plot.new) and drawing the text at the centre of the
region.

> plot.new()
> text(0.5, 0.5, "An X Label")
> plot.new()
> text(0.5, 0.5, "A Y Label", srt = 90)
 A Label for the Y Axis
−1.0 −0.5 0.0 0.5 1.0

●
●

2
●
●

4
●

6
A Label for the X Axis
●

8
●
●

10
Other Features

It is possible to add other features to the plot such as

additional axes at the top and right or an overall title by adding
further plotting regions.

With a little willingness to experiment, there is virtually no

limit to the kinds of plot which you can produce in this way.

The following plot shows a dotchart comparison of land

animal speeds produced in exactly this way.
 Land Animal Speeds

Cheetah ●
Pronghorn Antelope ●
Thomson's Gazelle ●
Lion ●
Wildebeest ●
Quarterhorse ●
Cape Hunting Dog ●
Elk ●
Coyote ●
Gray Fox ●
Mongolian Wild Ass ●
Zebra ●
Hyena ●
Greyhound ●
Whippet ●
Jackal ●
Mule Deer ●
Rabbit (domestic) ●
Giraffe ●
Reindeer ●
Cat (domestic) ●
Grizzly Bear ●
Wart Hog ●
White−Tailed Deer ●
Human ●
Elephant ●
Black Mamba Snake ●
Six−Lined Race Runner ●
Wild Turkey ●
Squirrel ●
Pig (domestic) ●
Chicken ●
Spider (Tegenaria atrica) ●
●
Giant Tortoise ● ●
Three−Toed Sloth ● ●
Garden Snail ● ●

0 2 25 50 75 100

Speed (km/hr)