VCAA-AMSI

Υ𝓍𝓍βmaths
⅓⅝5 modules
85� 134
345 1238 9 8 3 85
=2 57 3 3 0 98 1 2 3 7 21
3δ
9012 x

348 9012 x = 2345

3
ХФУШЩЫЅ
73487
56 ЂЁϟ 92 573 <𝜒𝜒≠14�57
⅙ 583 8 ϝϚό ≌𝑎𝑎 812
⋓ ⊚ ύҖ 7

��
8

47∆
4

⍰
⍬
Ѽ
ᴂ

873 32δ¼
⤮ ԅ

Ӡӓ
⟼

ӕ
⋟⤍

Ӂ
123

Ҷҵ
09

3 4 49

ӛ
82

89

₴₦
45

€€
57

⅘⅔
029

12
30
3

ℨℶℜ
98 72
793

85

ⅈↄ⅞
123
3 34

∭
76€-20ύҖҦѼᴂԅӠ468

83 85�Υ𝓍𝓍β⅓⅝5 12

5984
3751-087634ΩÅℭ℗ℋℤ∬√
51-08

A guide for teachers – Years 11 and 12

7

∜∱≄
43

∾
8

⋞6
9

Geometry
5

⋃
∭

Transformations
⋟⤍
ↄ⅞

₦€
⍰
₴
⟼
ⅈ

⋓
ℜ

⊚
⤮⍬
ℨℶ

348
⋟ӛ 58
38
⑦
756Ϛ
6⋃ 7
⋞
όύҖ
∾ ≄ ∱
∬√∜ ℤ
ҦѼᴂԅӠӓӕӁҶҵ

F T

&
Years DR A

11 12
 A F T
DR
Transformations – A guide for teachers (Years 11–12)

Dr Daniel V. Mathews, Monash University

Illustrations and web design: Catherine Tan

© VCAA and The University of Melbourne on behalf of AMSI 2015.

Authorised and co-published by the
Victorian Curriculum and Assessment Authority
Level 1, 2 Lonsdale Street
Melbourne VIC 3000

This publication may be used in accordance with the VCAA Copyright Policy:
[Link]/Pages/aboutus/policies/[Link]

Copyright in materials appearing at any sites linked to this document rests

with the copyright owner/s of those materials. The VCAA recommends you
refer to copyright statements at linked sites before using such materials.

Australian Mathematical Sciences Institute

Building 161
The University of Melbourne
VIC 3010
Email: enquiries@[Link]
Website: [Link]
Assumed knowledge . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Motivation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4

Content . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Transformations of the plane . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
Geometric examples of transformations . . . . . . . . . . . . . . . . . . . . . . . . . 6
Linear transformations and matrices . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
Describing geometric transformations algebraically . . . . . . . . . . . . . . . . . . 18
Transformations of graphs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

Links Forward . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Areas and determinants . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
Isometries . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
Three-dimensional and higher-dimensional transformations . . . . . . . . . . . . . 41

Answers to exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42
Transformations

Assumed knowledge
• Familiarity with functions, including injective, surjective and inverse functions;
• Familiarity with linear equations, including solving simultaneous equations.
• Familiarity with vectors and 2£2 matrices, including position vectors and matrix mul-
tiplication and inverses;

• Familiarity with plane Euclidean geometry.

Motivation
Give me a place to stand and I will move the earth.

– Archimedes

As in many aspects of life, it’s often useful in mathematics to see things from different
perspectives. Changing your point of view involves some kind of transformation.

You see geometric transformations every minute of every day: whenever you walk around
an object, you see a rotation. When you see an image of an object, it is a projection. When
you move an object, you see a translation. When you look in a mirror, you see a reflection.

In this module we focus on transformations of the plane. Plane transformations encom-

pass a great deal of geometry, including rotations, reflections, projections, translations,
and many other operations.

We mostly focus on transformations of a very particular type: linear transformations.

Linear transformations form a narrow, but useful, class of plane transformations. We
will see how such transformations can be described by some relatively simple algebra.

Studying plane transformations brings together algebra and geometry in a coherent and
elegant way. We will see that the algebra of linear functions and matrices efficiently de-
scribes a wide array of geometric transformations.
 A guide for teachers – Years 11 and 12 • {5}
Content

Transformations of the plane

Not everything that is faced can be changed. But nothing can be changed
until it is faced. – James Baldwin

A plane transformation is just a function from the plane to itself, i.e. a function F :
R2 ! R2 . Such an F takes a point (x, y) to another point, which we can denote (x 0 , y 0 )
or F (x, y).1 You can think of x, y as inputs to f , and x 0 , y 0 as outputs of f .

By our definition, any function R2 ! R2 whatsoever is a plane transformation. E.g.:

F (x, y) = (0, 5), F (x, y) = (2x + y, 5x ° 3y), F (x, y) = (x ° 2y + 3, °4x + 5y ° 6),

p
F (x, y) = (x 2 + y 3 , e y 3 x), F (x, y) = (bx yc, sin(x + y)).

(However, in this module we will not deal with functions as complicated as the last two.)

It’s not possible to draw a graph of a function F : R2 ! R2 in the same way we draw the
graph y = f (x) of a function f : R ! R. For a function f : R ! R, its graph consists of a plot
of all points (x, y) such that y = f (x). To draw a similar graph of a function F : R2 ! R2 ,
we would need to plot all (x, y, x 0 , y 0 ) such that (x 0 , y 0 ) = F (x, y), which would require four
coordinates — it would have to be drawn in four-dimensional space! As we will see,
however, there are other ways to visualise a plane transformation.

In this module, we will do the following.

• We will briefly examine some geometry, introducing a few examples of plane transfor-
mations arising from geometric operations.

• Then, we will examine some algebra of vectors and matrices, and discuss linear trans-
formations, which arise from this algebra.

• Next, we will examine some geometric transformations in depth: rotations, transla-

tions, projections, dilations, and more. We will investigate the relationship between
the geometry and algebra of these operations.

• Finally, we will discuss transformations of graphs, discussing what happens to a graph

y = f (x) when it is subject to a transformation.

• In the Links Forward section, we will discuss some further concepts such as deter-
minants, isometries, and how the ideas of this module can be generalised from the
2-dimensional plane to 3-dimensional space, and even to higher dimensions.

1 Note the primes on x 0 , y 0 have nothing to do with derivatives!

{6} • Transformations AMSI module
Warning regarding notation. Throughout this module, we often blur the distinction be-
2 3
x
tween a point (x, y) in the plane, and its position vector 4 5. This is a slight abuse of
y
notation, but it may be useful, and we hope it will not be too confusing.

Geometric examples of transformations

Some transformations arise from very natural and intuitive geometric operations.

Identity

The laziest plane transformation of all is the one that does nothing. This transformation
is known as the identity function I : R2 ! R2 . It sends every point to itself, I (x, y) = (x, y).

y y0

A A

x x0

Rotations

Given a point P in the plane, and an angle ¡,2 we can consider ro-
tating points around P by the angle ¡. Any point X can be rotated in
this way, giving rise to a plane transformation R2 ! R2 . X0

X
2 This is the Greek letter phi.
¡

P
 A guide for teachers – Years 11 and 12 • {7}

y A0

Rotate

x x0
¡
P P

Translations
Given a vector v, we can consider translating points by v. Any X0
v
point X can be moved along by the vector v to obtain a point
X 0 , and in this way we obtain a plane transformation. X

y y0
A0
A

Translate
v
x x0

Projections
Given a line L in the plane, we can consider projecting points
onto L. From a point X , draw the line through X perpendic- X
L
ular to L; the intersection of this line with L is the projection
X 0 of X onto L. Projecting all points of the plane onto L, we X0
obtain a plane transformation.
{8} • Transformations AMSI module
y y0

Project
A0
x x0

L L

Exercise 1
Show that the projection X 0 of X onto L is the closest point on L to X .

Reflections
Take a line L and think of it as a mirror; we can then reflect
points in L. Given a point X not on L, suppose you are stand-
ing at X ; your reflection X 0 in L is where you appear to be X
L
when you look in the mirror. Reflecting each point in this way
we obtain a plane transformation.

To find X 0 , we can draw the line perpendicular to L through X ; X0

0
X is the point on this perpendicular, on the other side of L, at
the same distance from L as X .

y y0

Reflect

x x0
A0

L L
 A guide for teachers – Years 11 and 12 • {9}
Dilations

Given a line L and a real number k, we can dilate points from X0

L by a factor k. From a point X not on L, once more we draw
the perpendicular from X to L. The dilation X 0 of X from L : k
1
with factor k is the point X 0 on the perpendicular which is k X L
times as far away from L as X is. In other words, if P is the foot
°°! °°!
of the perpendicular from X to L, then P X 0 = k P X .
P

Again we have a plane transformation, which stretches the plane out from L by a factor
of k. Note that if k is positive, then X 0 lies on the same side of L as X ; if k is negative, then
X 0 lies on the opposite side of L from X .

y y0

A0
Dilate

x x0
A

L L

Linear transformations and matrices

We will now take a more algebraic approach to transformations of the plane.

As it turns out, matrices are very useful for describing transformations. Whenever we
have a 2 £ 2 matrix of real numbers
2 3
a b
M =4 5,
c d

we can naturally define a plane transformation T M : R2 ! R2 by

T M (v) = M v.

That is, T M takes a vector v and multiplies it on the left by the matrix M . If v is the position
vector of the point (x, y), then
02 31 2 3 2 32 3 2 3
x x a b x ax + b y
T M (v) = T M @4 5A = M 4 5 = 4 54 5 = 4 5
y y c d y cx + d y
{10} • Transformations AMSI module
or equivalently, T M (x, y) = (ax + b y, cx + d y).
2 3
x
(Note that here we used the notation (x, y) and 4 5 interchangeably; we will be doing
y
this throughout this module.)

As it turns out, matrices give us a powerful systematic way to describe a wide variety of
transformations: they can describe rotations, reflections, dilations, and much more.
Example
2 3
1 2
Let M = 4 5.
3 7

1 Write an expression for T M .

2 Find T M (1, 0) and T M (0, 1).
3 Find all points (x, y) such that T M (x, y) = (1, 0).

Solution
2 32 3 2 3
1 2 x x + 2y
1 T M (x, y) = 4 54 5 = 4 5 = (x + 2y, 3x + 7y).
3 7 y 3x + 7y
2 Using the formula from the previous part, T M (1, 0) = (1, 3) and T M (0, 1) = (2, 7)
3 We have T M (x, y) = (x + 2y, 3x + 7y) = (1, 0), hence the simultaneous equations

x + 2y = 1, 3x + 7y = 0.

Solving these equations yields x = 7, y = °3; and this is the only solution. So the only
point (x, y) such that T M (x, y) = (1, 0) is (x, y) = (7, °3).

Note that T M (1, 0) and T M (0, 1) are precisely the columns of the matrix M . This is an
important fact, which we will discuss later.

While every matrix describes a plane transformation, not every plane transformation can
be described by a matrix. Matrices correspond to a specific type of plane transformation
which sends (x, y) to (ax + b y, cx + d y), for some real numbers a, b, c, d .

A transformation T M arising from a matrix M obeys some “distributive laws”. For any
2 £ 2 matrix M and vectors v and w in R2 , it is true that

M (v + w) = M v + M w and hence T M (v + w) = T M (v) + T M (w).

 A guide for teachers – Years 11 and 12 • {11}
Moreover, for any real number (scalar) c,

M (cv) = cM v and hence T M (cv) = cT M (v).

In fact, if a function F : R2 ! R2 satisfies these two distributive laws, then it must arise
from a matrix. Why? If F (v + w) = F (v) + F (w) and F (cv) = cF (v), then for any point (x, y),
2 3 02 3 2 31 2 3 2 3
x x 0 x 0
F 4 5 = F @4 5 + 4 5A = F 4 5 + F 4 5 (using F (v + w) = F (v) + F (w))
y 0 y 0 y
2 3 2 3
1 0
= xF 4 5 + yF 4 5 (using F (cv) = cF (v))
0 1

Letting F (1, 0) = (a, c) and F (0, 1) = (b, d ) then, we have

2 3 2 3 2 3 2 3 2 32 3
x a b ax + b y a b x
F 4 5 = x 4 5+ y 4 5 = 4 5=4 54 5,
y c d cx + d y c d y

so F corresponds to a matrix.

Therefore, we have two equivalent ways to define linear transformations. A plane trans-
formation F is linear if either of the following equivalent conditions holds:

• F (x, y) = (ax + b y, cx + d y) for some real a, b, c, d . That is, F arises from a matrix.
• For any scalar c and vectors v, w, F (cv) = cF (v) and F (v + w) = F (v) + F (w).

Exercise 2
Show that any linear transformation sends the origin to the origin.

The effect of a linear transformation

Suppose you are given a matrix M . You now know that M determines a linear transfor-
mation T M . But what does T M do, geometrically? We will now see a method to analyse
this question.

Let’s begin by restricting our attention to two vectors: (1, 0) and (0, 1). These two vectors
are sometimes called the standard basis for R2 .
2 3
a b
Multiplying any matrix M = 4 5 by the standard basis vectors gives its columns:
c d
2 3 2 32 3 2 3 2 3 2 32 3 2 3
1 a b 1 a 0 a b 0 b
M4 5=4 54 5 = 4 5 and M 4 5 = 4 54 5 = 4 5.
0 c d 0 c 1 c d 1 d
{12} • Transformations AMSI module
So, just by looking at a matrix M , reading down its columns, you can see where T M sends
(1, 0) and (0, 1). Once you know where T M sends (1, 0) and (0, 1), you can use the “dis-
tributive” property of linear transformations to see where any other point goes.

For instance, if you want to know where T M sends (2, 3), you could note that (2, 3) =
2(1, 0) + 3(0, 1), so that

T M (2, 3) = 2T M (1, 0) + 3T M (0, 1).

Hence T M (2, 3) is given by 2 times the first column of M , plus 3 times the second column.

In the x y-plane we can consider the unit square, with vertices (0, 0), (1, 0), (0, 1), (1, 1).
Unit squares tiles the plane, giving a tessellation whose vertices are precisely the points
with integer coordinates. Where does T M send these points?

• T M sends the origin to the origin (exercise ??).

• T M sends (1, 0) and (0, 1) to the vectors given by the first and second columns of M .
• T M sends (1, 1) = (1, 0)+(0, 1), to T M (1, 0)+T M (0, 1), the sum of the two columns of M .
• In general, T M sends (a, b) to aT M (1, 0) + bT M (0, 1).

(0, 1) (1, 1) T M (1, 1)

TM T M (0, 1)

(0, 0) (1, 0) (0, 0) T M (1, 0)

So the unit square is sent by T M to a parallelogram. The distributive property implies

that the whole tessellation of the plane by unit squares is sent by T M to a tessellation of
the plane by parallelograms. We can then draw a picture of the entire transformation T M .

y y0

TM

x x0

Let us now use these ideas to describe some specific linear transformations.
 A guide for teachers – Years 11 and 12 • {13}
Example
2 3
1 1
Let M = 4 5. Describe the linear transformation T M geometrically.
0 1

Solution

Reading the columns of T M tells us that T M (1, 0) = (1, 0) and T M (0, 1) = (1, 1); the trans-
formation T M thus turns the unit square into a parallelogram with base 1 and height 1
as shown. This transformation is called a shear.

y y0

TM

x x0

In the above figure, we draw a parallelogram (shaded grey) with two vectors v = T M (1, 0)
(green) and w = T M (0, 1) (blue) tail-to-tail at the origin. We call this the parallelogram
spanned by v and w.
Example
2 3
1 0
Let M = 4 5. Describe the linear transformation T M geometrically.
0 °1

Solution

Reading the columns of M , we have T M (1, 0) = (1, 0) and T M (0, 1) = (0, °1). So T M fixes
(1, 0) and “flips” (0, 1) to its negative (0, °1). Thus vertical directions are “flipped”, so that
(x, y) is sent to (x, °y), and T M is reflection in the x-axis.

y y0
TM

x x0

Note that in the two examples above, the parallelograms spanned by T M (1, 0) and T M (0, 1)
have different orientations. In the first example, if we consider sweeping around the ori-
gin, through the parallelogram, from T M (1, 0) to T M (0, 1), we go anticlockwise; in the sec-
{14} • Transformations AMSI module
ond example, we go clockwise. Accordingly, we say the first parallelogram is positively
oriented, and the second is negatively oriented.

In the standard unit square, sweeping

around the origin from (1, 0) to (0, 1) goes
(0, 0) T M (1, 0)
anticlockwise, so it is positively oriented.
T M (0, 1)
So the first example preserved this orien-
tation, and the second example reversed
T M (0, 1)
it. In general, we say a linear transfor- (0, 0) T M (1, 0)
mation preserves or reverse orientation ac-
cordingly as the parallelogram spanned
by T M (1, 0) and T M (0, 1) is positively or negatively oriented.

Exercise 3
For each matrix M below, describe the linear transformation T M geometrically.
2 p 3 2 3 2 3 2 3 2 3
1 3
° 7 0 °1 0 2 0 0 1
4 p2 2 5
, 4 5, 4 5, 4 5, 4 5.
3 1
2 2 0 7 0 °1 0 3 1 0

Composing linear transformations and matrix multiplication

Let’s now suppose we have two matrices M , N representing two linear transformations
T M , T N . What happens if we apply one transformation then the other? What is the com-
position T M ± T N ?
2 3 2 3
a b e f
Let M = 4 5 and N = 4 5. So
c d g h

T M (x, y) = (ax + b y, cx + d y) and T N (x, y) = (ex + f y, g x + h y).

Composing these two functions gives

T M ± T N (x, y) = T M (T N (x, y)) = T M (ex + f y, g x + h y)

° ¢
= a(ex + f y) + b(g x + h y), c(ex + f y) + d (g x + h y)
° ¢
= (ae + bg )x + (a f + bh)y, (ce + d g )x + (c f + d h)y .

Remembering that a, b, c, d , e, f , g , h are just real constants, this is another linear trans-
formation, and the corresponding matrix is
2 3
ae + bg a f + bh
4 5.
ce + d g c f + dh
 A guide for teachers – Years 11 and 12 • {15}
This matrix is precisely the product of the two matrices M and N :
2 32 3 2 3
a b e f ae + bg a f + bh
MN = 4 54 5=4 5.
c d g h ce + d g c f + dh

We conclude that the composition of the linear transformations of M and N is the linear
transformation of M N , proving the following theorem.

Theorem
For any matrices M and N , T M ± T N = T M N .

This theorem gives us a very quick way to compose linear transformations: just multiply
the corresponding matrices!
Example

Find the matrix for the composition g ± f of the two linear transformations f (x, y) =
(x + y, y) and g (x, y) = (y, x + y).

Solution
2 3 2 3
1 1 0 1
We have f = T M and g = T N where M = 4 5 and N = 4 5. So the matrix of the
0 1 1 1
composition g ± f = T N ± T M = T N M is the product N M :
2 32 3 2 3
0 1 1 1 0 1
NM = 4 54 5=4 5.
1 1 0 1 1 2

Exercise 4
Show that if the linear transformation f (x, y) = (y, °x + y) is composed with itself six
times, the result is the identity transformation.

The identity transformation and identity matrix

Recall the identity transformation I : R2 ! R2 sends every point to itself, I (x, y) = (x, y).
We can now note I is a linear transformation, corresponding to the identity matrix
2 3
1 0
Id = 4 5 . In other words, TId = I .
0 1
{16} • Transformations AMSI module
The inverse of a linear transformation and inverse matrices

Recall that a function has an inverse if and only if it is bijective, i.e. injective and surjec-
tive. (A function f : X ! Y is injective if x 6= y implies f (x) 6= f (y); and f is surjective if
its image is all of Y , i.e. for all y 2 Y there exists x 2 X such that f (x) = y.)

Some transformations of the plane are bijective, and some are not; so some have in-
verses, and others do not. As we proceed, we will see many examples of transformations
that have inverses, and many that don’t.

A bijective transformation F : R2 ! R2 , has an inverse F °1 : R2 ! R2 . These functions F

and F °1 “undo” each other. So if F (x) = y then F °1 (y) = x, and vice versa.

Recall that matrices can have inverses too. If

2 3 2 3
a b 1 d °b
M =4 5 then M °1 = 4 5 (provided ad ° bc 6= 0).
c d ad ° bc °c a

We say M °1 is the inverse of M because M and M °1 multiply to give the identity.

Exercise 5
Verify that M M °1 = M °1 M = Id.

If we compose the linear transformations T M and T M °1 , we obtain the identity:

T M ± T M °1 = T M M °1 = TId = I , T M °1 ± T M = T M °1 M = TId = I .

Hence T M and T M °1 undo each other; they are inverse transformations.

Example

What is the inverse of the transformation F : R2 ! R2 given by F (x, y) = (x + 3y, x + 5y)?

Solution

The transformation F is linear and corresponds to the matrix

2 3 2 3 2 3
5 °3
1 3 1 5 °3
M =4 5, which has inverse M °1 = 4 5=4 2 2 5
.
1 5 1 · 5 ° 3 · 1 °1 1 °1 1
2 2

The inverse of F = T M is then F °1 = T M °1 ,

µ ∂
°1 5 3 °1 1
F (x, y) = x ° y, x+ y .
2 2 2 2
 A guide for teachers – Years 11 and 12 • {17}
To show that a function does not have an inverse, one can show that it is not injec-
tive, or that it is not surjective. In the following example we do both, giving two solu-
tions.
Example

Show that the transformation F : R2 ! R2 given by F (x, y) = (x, 0) has no inverse.

Solution

Solution 1. Since (for instance) F (0, 0) = F (0, 1) = (0, 0), F is not injective. Hence F has
no inverse.

Solution 2. Any point in the image of F has second coordinate zero. So there is no (x, y)
such that (for instance) F (x, y) = (0, 1). Hence F is not surjective, and has no inverse.

Exercise 6
Show that F : R2 ! R2 given by F (x, y) = (x + 2y, 2x + 4y) has no inverse.

The following example illustrates a useful technique for finding a linear transformation,
from its value at two points.
Example

Find a linear transformation R2 ! R2 that maps (1, 1) to (°1, 4) and (°1, 3) to (°7, 0)

Solution

Let M be the matrix of the desired linear transformation. We have

2 3 2 3 2 3 2 3
1 °1 °1 °7
M 4 5 = 4 5 and M 4 5 = 4 5 .
1 4 3 0

In fact, we can put these two equations together into a single matrix equation
2 3 2 3
1 °1 °1 °7
M4 5=4 5
1 3 4 0

which we can then solve for M :

2 32 3°1 2 32 3 2 3 2 3
°1 °7 1 °1 1 °1 °7 3 1 1 4 °8 1 °2
M =4 54 5 = 4 54 5= 4 5=4 5
4 0 1 3 4 4 0 °1 1 4 12 4 3 1

Hence the only such transformation is T M (x, y) = (x ° 2y, 3x + y).

{18} • Transformations AMSI module
Exercise 7
Find the linear transformation that sends (3, 1) to (1, 2) and (°1, 2) to (2, °3)

Describing geometric transformations algebraically

We have now seen several geometric plane transformations. We have also seen how ma-
trices can describe some transformations (linear ones) in a simple algebraic way.

We now investigate these geometric operations from an algebraic point of view. We ask
whether, and how, they can be represented by matrices. This gives us a very interesting
way to relate everyday geometric operations to algebra.

Rotations

To everything — turn, turn, turn

– Pete Seeger

As discussed previously, given a point P in the plane, and an angle ¡, we can consider
the rotation around P by angle ¡. By convention, we measure angles anticlockwise and
in radians. We denote3 this rotation by

RotP, ¡ : R2 ! R2 .

Note the following preliminary facts about rotations:

1 A rotation by angle 0 is just the identity transformation, as is rotation by 2º.

2 Rotation by ¡ + 2º is the same transformation as rotation by ¡.
3 Rotations of angles µ and ¡ about the same point P give a rotation of µ + ¡ about P .
4 Any rotation is bijective. Indeed, a rotation by ¡ about P can be undone by rotating
by °¡, so the transformation RotP, ¡ has inverse RotP, °¡ .

That is,

RotP, 0 = RotP, 2º = I , RotP, ¡+2º = RotP, ¡ ,

° ¢°1
RotP, ¡ ± RotP, µ = RotP, µ+¡ , RotP, ¡ = RotP, °¡ .

3 Please note: This is not a standard notation. We find it useful, but it may not be used by others!
 A guide for teachers – Years 11 and 12 • {19}
Exercise 8
Consider Rot0, º/2 , rotation about the origin by º/2. Show that Rot0, º/2 (x, y) = (°y, x).

Now, when we rotate around the origin 0, it turns out that Rot0, ¡ is a linear transforma-
tion. We will show why this is true, and find the corresponding matrix.

To show that Rot0, ¡ is linear, we will show that it satisfies the “distributive” laws

Rot0,¡ (cv) = cRot0,¡ (v) and Rot0,¡ (v + w) = Rot0,¡ (v) + Rot0,¡ (w) ,

for any vectors v, w and real number c.

For the first law, note that the two vectors cv and v point in the same direction; cv is c
times longer than v. After rotation by ¡ about 0, the two resulting vectors still point in
the same direction and one is still c times longer than the other, so

Rot0,¡ (cv) = cRot0,¡ (v) as desired.

y y

Rot0, ¡ (cv) v+w w

cv Rot0, ¡ (v)
¡ Rot0, ¡ (w) ¡
Rot0, ¡ (v) v v
x x
Rot0, ¡ (v + w)

For the second law, place v at the origin and place w head-to-tail with it, as shown, to
obtain the vector v + w. After rotating these vectors all by ¡, we have head-to-tail vectors
Rot0, ¡ (v) and Rot0, ¡ (w), summing to Rot0, ¡ (v + w). Thus

Rot0, ¡ (v) + Rot0, ¡ (w) = Rot0, ¡ (v + w) as desired.

Now that we know Rot0, ¡ is a linear transformation, we know that it corresponds to a

matrix: the next question is to find the matrix.

To find the matrix, consider where the standard basis vectors y

(1, 0) and (0, 1) go under the rotation: this will give us the two
(cos ¡, sin ¡)
columns of the matrix, as discussed earlier.

Rotating (1, 0) by an angle of ¡ about the origin, we obtain the

point (cos ¡, sin ¡), by elementary trigonometry in the right- (1, 0)
¡
x
angle triangle shown.
{20} • Transformations AMSI module
Exercise 9
Show that rotating (0, 1) by ¡ about the origin yields the point (° sin ¡, cos ¡).

Thus we have found the columns of the matrix of Rot0, ¡ . The matrix is
2 3
cos ¡ ° sin ¡
4 5.
sin ¡ cos ¡

Such a matrix is often called a rotation matrix with angle ¡.

Example

Check directly that the inverse of the rotation matrix with angle ¡ is the rotation matrix
° ¢°1
with angle °¡. Hence verify algebraically that Rot0, ¡ = Rot0, °¡ .

Solution

The inverse or the rotation matrix with angle ¡ is

2 3°1 2 3 2 3
cos ¡ ° sin ¡ 1 cos ¡ sin ¡ cos(°¡) ° sin(°¡)
4 5 = 4 5=4 5,
cos2 ¡ + sin2 ¡
sin ¡ cos ¡ ° sin ¡ cos ¡ sin(°¡) cos(°¡)

which is the rotation matrix with angle °¡. So the inverse of Rot0, ¡ is Rot0, °¡ .

Exercise 10
Show that the product of the rotation matrices with angles µ and ¡ is the rotation matrix
with angle µ + ¡. Hence verify algebraically that Rot0, µ ± Rot0, ¡ = Rot0, µ+¡ .

The formula for a rotation about the origin can y

also be derived directly. We can do this using (r cos µ, r sin µ)

polar coordinates. Instead of describing a point r

by its Cartesian coordinates (x, y), we can use µ x
its distance from the origin r , and its direction µ
from the origin. The angle µ is measured anti-
clockwise from the positive x-axis.

Basic trigonometry in the triangle shown gives x = r cos µ and y = r sin µ. When we ap-
ply a rotation about the origin of angle ¡, the point (x, y) moves an angle of ¡ around
the origin, but its distance from the origin r is unchanged. Hence the point (x, y) =
(r cos µ, r sin µ) is mapped to (x 0 , y 0 ) = (r cos(µ + ¡), r sin(µ + ¡)). We can now use the sine
 A guide for teachers – Years 11 and 12 • {21}
and cosine addition formulae to rewrite this expression:
2 3 2 3 2 3
x r cos(µ + ¡) r cos µ cos ¡ ° r sin µ sin ¡
Rot0, ¡ 4 5 = 4 5=4 5
y r sin(µ + ¡) r sin µ cos ¡ + r cos µ sin ¡
2 3 2 32 3
(cos ¡)x ° (sin ¡)y cos ¡ ° sin ¡ x
=4 5=4 54 5.
(sin ¡)x + (cos ¡)y sin ¡ cos ¡ y

This gives us a formula for Rot0, ¡ and the rotation matrix directly.

While a rotation about the origin gives a linear transformation, a rotation about any other
point is not linear, and hence is not given by a matrix. It is a plane transformation, but
not a linear one. You can prove this in the next exercise.

Exercise 11
If ¡ is not a multiple of 2º, and P is not the origin, show that RotP, ¡ (0) 6= 0. Using exercise
??, conclude that RotP, ¡ (0) is not linear.

In the section on Translations, you can find a formula for a rotation about any point.

Translations

I like to move it, move it

– Reel 2 Real4

Picking up an object and moving it is a very familiar concept. As discussed previously,

given a vector v, we can consider the translation of the plane by the vector v. We denote
this transformation by5

Transv : R2 ! R2 .

Let v = (a, b). When we translate a point, with position vector x = (x, y), by v, we simply
add the two vectors, as shown below. Thus

Transv (x) = x + v, or equivalently, Transv (x, y) = (x + a, y + b).

4 Not to be confused with R ! R!

5 Again, this is not a standard notation.

{22} • Transformations AMSI module
y y0
v

x Transv Transv (x)

x x0

Translation by 0 does not move points very far! It is the identity: Trans0 = I .

When v 6= 0, translation by v is not a linear transformation. The constant terms a, b in

the formula (x + a, y + b) are not allowed under our definition of linear transformation.
Another way to see that Transv is not linear is to note that Transv (0) = v, but a linear
transformation must take the origin 0 to 0 (exercise ??). (Transformations which allow
both linear and constant terms are called affine transformations.)

Exercise 12
Show that for any vector v, the translation Transv : R2 ! R2 is bijective. Find the inverse
transformation, and give it an geometric interpretation.

Exercise 13
(This exercise may be easier after reading the rest of this module.) Let P = (a, b) be a point
in the plane with position vector v. Show that translation by °v, followed by rotation by
angle ¡ about the origin 0, followed by translation by v, gives the plane transformation
which rotates by angle ¡ about P . That is, show that

Transv ± Rot0, ¡ ± Trans°v = RotP, ¡ .

Hence give a formula for RotP, µ (x, y), the rotation of (x, y) by angle µ around P = (a, b).

Projections

As discussed earlier, given a line L on the plane, we can consider the plane transforma-
tion which projects points onto L. We can denote6 this transformation by

ProjL : R2 ! R2 .

Such projections are very familiar from everyday life. If the plane is a map, L is a road,
and P is your position, then ProjL (P ) is the point on the road closest to you. If the line L
is horizontal, representing the ground, P is above ground, and the sun is directly above,
then ProjL (P ) is the shadow of P on the ground.

6 Again, beware that this notation is not standard.

 A guide for teachers – Years 11 and 12 • {23}
We have some preliminary facts about projections:

1 A projection ProjL is not surjective. Its image consists only of points on L.

2 A projection ProjL is not injective. Many points project to any given point of L.
3 If P is a point on L, then ProjL (P ) = P .
4 After projecting a point onto L, projecting onto L again has no further effect. That is,
ProjL ± ProjL = ProjL .

A function f such that f ± f = f is called an idempotent. The last property above says a
projection is an idempotent.

Exercise 14
Give proofs of the above facts.

As it turns out, when the line L passes through the origin, ProjL is a linear transformation.
We will show why, and find the corresponding matrix.

First of all, suppose L is the x-axis. Projection onto L in this case just sets the y-coordinate
of a point to 0, and leaves the x-coordinate unchanged: Projx-axis (x, y) = (x, 0). Similarly
Proj y-axis (x, y) = (0, y). These are linear transformations:
2 3 2 3 2 32 3 2 3 2 3 2 32 3
x x 1 0 x x 0 0 0 x
Projx-axis 4 5 = 4 5 = 4 54 5, Proj y-axis 4 5 = 4 5 = 4 54 5.
y 0 0 0 y y y 0 1 y

Now suppose that L is any line through the origin, making an angle of ¡ with the positive
x-axis (as always, measured anticlockwise in radians), and let v = (x, y).

The key idea to calculate ProjL (v) is to perform the projection in several steps. We first
rotate the plane by °¡ about the origin. This rotates L to the x-axis, and rotates v too.
Then, we project onto the x-axis. Finally, we rotate back by ¡ around the origin. The
result is that v has been projected onto L. In symbols,

ProjL = Rot0,¡ ± Projx-axis ± Rot0,°¡ .

{24} • Transformations AMSI module
y y0

v ProjL ProjL (v)

¡ ¡
x x0

Rot0, °¡ Rot0, ¡

y y

x x
Projx-axis

Why did we do this? Because we know all the matrices for the transformations on the
right hand side! The matrix for Projx-axis is given above, and
2 3 2 32 3 2 3 2 32 3
x cos ¡ ° sin ¡ x x cos ¡ sin ¡ x
Rot0,¡ 4 5 = 4 54 5, Rot0,°¡ 4 5 = 4 54 5.
y sin ¡ cos ¡ y y ° sin ¡ cos ¡ y

Hence ProjL is linear, and its matrix is given by multiplying these matrices together:
2 32 32 3 2 3
cos ¡ ° sin ¡ 1 0 cos ¡ sin ¡ cos2 ¡ sin ¡ cos ¡
4 54 54 5=4 5.
sin ¡ cos ¡ 0 0 ° sin ¡ cos ¡ sin ¡ cos ¡ sin2 ¡

So projection onto L is a linear transformation corresponding to this matrix:

2 3 2 32 3
x cos2 ¡ sin ¡ cos ¡ x
ProjL 4 5 = 4 54 5
y sin ¡ cos ¡ sin2 ¡ y

When ¡ = 0, this matrix agrees with our calculation for projection onto the x-axis.
Example

Let L be the line y = x. Find a formula for projection onto L.

Solution
 A guide for teachers – Years 11 and 12 • {25}
The line L passes through the origin and makes an angle ¡ = º/4 with the positive x-axis.
Hence ProjL is a linear transformation with matrix
2 3 2 3
1 1
cos2 º4 sin º4 cos º4
4 5= 42 25
.
1 1
sin º4 cos º4 sin2 º4 2 2

Therefore,
2 3 2 32 3 2 3
1 1 x+y
x 2 2 5 4x 5 2 5
ProjL 4 5 = 4 = 4 .
1 1 x+y
y 2 2 y 2

Exercise 15
Verify the above formula for projection onto y = x directly as follows. Given a point
(x 0 , y 0 ), find the intersection of the line through this point with gradient °1, with the
x 0 +y 0 x 0 +y 0
line y = x. Conclude that ProjL (x 0 , y 0 ) = ( 2 , 2 ) as above.

Exercise 16
What is the projection of the point (8, 3) onto the line y = 2x?

An alternative derivation of the projection matrix is given in the following exercise.

Exercise 17
Let L be a line through the origin, making an angle ¡ with the positive real axis.

a Show that ProjL (cv) = cProjL (v) and ProjL (v+w) = ProjL (v)+ProjL (w), for any vectors
v, w and real number c.

b Show that the projection of the point (1, 0) onto L is (cos2 ¡, sin ¡ cos ¡), and the pro-
jection of (0, 1) onto L is (sin ¡ cos ¡, sin2 ¡).

c Conclude that ProjL is as given above.

It’s now possible to see some facts about projections from the algebraic properties of a
projection matrix.

Exercise 18
Let L be the line through the origin making an angle of ¡ with the positive x-axis.
{26} • Transformations AMSI module
a Let P = (x, y) be a point on L. Show that
2 32 3 2 3
cos2 ¡ sin ¡ cos ¡ x x
4 54 5 = 4 5 and hence ProjL (P ) = P.
sin ¡ cos ¡ sin2 ¡ y y

b Show that ProjL ± ProjL = ProjL by demonstrating that

2 32 2 3
cos2 ¡ sin ¡ cos ¡ cos2 ¡ sin ¡ cos ¡
4 5 =4 5.
sin ¡ cos ¡ sin2 ¡ sin ¡ cos ¡ sin2 ¡

While projection onto a line through the origin is linear, projection onto a line not through
the origin is not linear, and hence not given by a matrix, as you can prove now.
Exercise 19
If L does not pass through the origin, show ProjL is not linear. (Hint: exercise ??.)

Reflections

As discussed earlier, given a line L in the plane, we can consider the plane transformation
which reflects points in L. We denote7 this transformation by

RefL : R2 ! R2 .

Here are some elementary facts about reflections:

1 Reflecting a point P in L, and reflecting it in L again, returns you to P . Hence,

RefL ± RefL = I .

2 A reflection is bijective; in fact, the inverse of a reflection is itself :

(RefL )°1 = RefL .

3 In P is a point on L, then its reflection in L is just P again; RefL (P ) = P .

A function f such that f ± f is the identity — equivalently, a function which is its own
inverse — is called an involution. So reflections are involutions.

When L is the x-axis, reflection in L fixes the x-coordinate and “flips” the y-coordinate.
Similarly, reflection in the y-axis “flips” the x-coordinate:
2 3 2 3 2 32 3 2 3 2 3 2 32 3
x x 1 0 x x °x °1 0 x
Refx-axis 4 5 = 4 5 = 4 5 4 5 , Ref y-axis 4 5 = 4 5 = 4 54 5.
y °y 0 °1 y y y 0 1 y

7 Again, notation not standard.

 A guide for teachers – Years 11 and 12 • {27}
So reflection in either axis is a linear transformation.

In fact, for any line L passing through the origin, RefL is a linear transformation. Letting
¡ be the angle between L and the positive x-axis, the corresponding matrix is given by
2 3 2 32 3
x cos(2¡) sin(2¡) x
RefL 4 5 = 4 54 5.
y sin(2¡) ° cos(2¡) y

In the following two exercises, we outline two methods to derive this result, similar to
those seen previously for projections.

Exercise 20
Show that rotation of angle °¡ about the origin 0, followed by reflection in the x-axis,
followed by rotation of ¡ about 0, is the same as reflection in L, i.e.

RefL = Rot0,¡ ± Refx-axis ± Rot0,°¡ .

Hence show that RefL is a linear transformation and find its matrix.

Exercise 21 a Show that RefL (cv) = cRefL (v) and RefL (v + w) = RefL (v) + RefL (w), for
any vectors v, w and real number c. Hence show RefL is a linear transformation.

b Show that the reflection of the point (1, 0) in L is (cos(2¡), sin(2¡)), and that the re-
flection of (0, 1) in L is (sin(2¡), ° cos(2¡)).

c Hence find the matrix for RefL .

Reflection in one particular line y = x has a useful effect: it swaps coordinates.

Example

Let L be the line y = x. Show that reflection in L takes a point (x, y) to (y, x).

Solution

Since L passes through the origin with angle ¡ = º/4, RefL is linear, with matrix
2 32 3
cos º2 sin º2 0 1
4 5=4 5.
sin º2 º
° cos 2 1 0
{28} • Transformations AMSI module
Thus
2 3 2 32 3 2 3
x 0 1 x y
RefL 4 5 = 4 54 5 = 4 5 as desired.
y 1 0 y x

Exercise 22
Give a formula for reflection in the line y = 2x.

The following exercise gives an interesting relationship between reflections and projec-
tions, and leads to a third way to derive the reflection matrix.

Exercise 23
Let L be a line in the plane, and x a vector. Show that

RefL (x) = 2ProjL (x) ° x, or equivalently, RefL = 2ProjL ° I .

° ¢
(Hint: Show x ° ProjL (x) = ° RefL (x) ° ProjL (x) .)

(The above exercise holds for any line L — it need not pass through the origin!)

Exercise 24
Let L be a line through the origin, making an angle of ¡ with the positive x-axis. Using
the previous exercise, and the matrix for ProjL , deduce the formula for RefL .

In the next exercise, we observe properties of reflections in the algebra of reflection ma-
trices.

Exercise 25
Let L be the line through the origin making an angle of ¡ with the positive x-axis.

a Let P = (x, y) be a point on L. Show that

2 32 3 2 3
cos(2¡) sin(2¡) x x
4 54 5 = 4 5 and hence RefL (P ) = P.
sin(2¡) ° cos(2¡) y y
 A guide for teachers – Years 11 and 12 • {29}
b Show that
2 32
cos(2¡) sin(2¡)
4 5 = Id and hence show RefL ± RefL = I .
sin(2¡) ° cos(2¡)

We have seen that reflection in a line passing through the origin is linear. But, as with
projections, if L does not pass through the origin, then RefL is not linear.

Exercise 26
Let d be a real number, and let L be the line x = d . Show that RefL (x, y) = (2d ° x, y).

(Hint: This can be shown directly, or by finding the projection and using exercise ??.)

Dilations

[R]ound the neck of the bottle was a paper label, with the words ‘DRINK ME’
beautifully printed on it in large letters...

Alice ventured to taste it, and finding it very nice, (it had, in fact, a sort of
mixed flavour of cherry-tart, custard, pine-apple, roast turkey, toffee, and
hot buttered toast,) she very soon finished it off...

‘What a curious feeling!’ said Alice; ‘I must be shutting up like a telescope.’

And so it was indeed: she was now only ten inches high

– Lewis Carroll, Alice’s Adventures in Wonderland

The plane transformation of a dilation has a similar effect to Alice’s “DRINK ME” potion
or “EAT ME” cake: it stretches or shrinks objects in the plane.

As discussed previously, given a line L and a real number k, we can consider the dilation
from L by factor k. We write8 this transformation as

DilL,k : R2 ! R2 .

The effect of such a dilation is to stretch lengths by a factor of k in the direction perpen-
dicular to L. If k is negative, then points are flipped to the other side of L.

Here are some basic properties of dilations:

1 Dilating from L with factor m, and then k, gives a dilation from L with factor km.

DilL,k ± DilL,m = DilL,km

8 Again, not a standard notation.

{30} • Transformations AMSI module
2 If k 6= 0, then dilation from L with factor k is bijective. Its inverse is given by dilation
from L with factor 1/k.
° ¢°1
DilL,k = DilL, 1
k

3 If P is a point on L, then for any k, DilL,k (P ) = P .

These facts are true even for dilation by negative factors.

We will analyse dilations similarly to projections and reflections. First, dilation from the
x-axis with factor k leaves the x-coordinate unchanged, and multiplies the y-coordinate
by k: Dilx-axis,k (x, y) = (x, k y). Similarly for the y-axis, Dil y-axis,k (x, y) = (kx, y). Hence
2 3 2 3 2 32 3 2 3 2 3 2 32 3
x x 1 0 x x kx k 0 x
Dilx-axis,k 4 5 = 4 5 = 4 54 5, Dil y-axis,k 4 5 = 4 5 = 4 54 5.
y ky 0 k y y y 0 1 y

Thus a dilation from either axis is a linear transformation.

In fact, for any line L passing through the origin and any real k, DilL,k is a linear transfor-
mation. Letting ¡ denote the angle L makes with the positive x-axis, we have
2 3 2 32 3
x 1 + (k ° 1) sin2 ¡ (1 ° k) sin ¡ cos ¡ x
DilL,k 4 5 = 4 54 5.
2
y (1 ° k) sin ¡ cos ¡ 1 + (k ° 1) cos ¡ y

You can check that ¡ = 0 or º/2 gives the matrices above for dilations in the axes.

Exercise 27 a Show that DilL,k = Rot0,¡ ± Dilx-axis,k ± Rot0,°¡ .

b Hence show that DilL,k is a linear transformation, with matrix as claimed above.

Example

Find a formula for dilation from the line y = °x with factor 3.

Solution

This line L has angle ¡ = °º/4, so sin ¡ = °1

p and cos ¡ = p1 . Hence
2 2

2 3 2 2 3 3
1 + (k ° 1) sin2 ¡
(1 ° k) sin ¡ cos ¡ 1 + 2 · 12 °2 · °1
2 2 1
4 5=4 5=4 5,
1
(1 ° k) sin ¡ cos ¡ 1 + (k ° 1) cos2 ¡ °2 · °1
2 1+2· 2 1 2

so the dilation is given by DilL,3 (x, y) = (2x + y, x + 2y).

 A guide for teachers – Years 11 and 12 • {31}
Exercise 28
What is the dilation of the point (3, 0) from the line y = 2x by factor 5?

Dilation matrices are more complicated than rotation, reflection or projection matrices.
Nonetheless, geometric properties of dilations are reflected in the algebra of dilation ma-
trices.
Exercise 29
Suppose k 6= 0 and L passes through the origin. Show that the matrices for dilation with
° ¢°1
factor k and k1 from L are inverses. Hence conclude that DilL,k = DilL, 1 .
k

Now for some particular values of k, dilation from a line L by factor k reduces to some
transformations we have seen previously.

1 Dilation from L with factor k = °1 flips every point to the corresponding point on the
opposite side of L: it is reflection in L.
2 Dilation from L with factor k = 0 is just projection onto L. “Stretching from L by factor
zero” simply projects points onto L.
3 Dilation from L with factor k = 1 leaves every point where it is.

In this sense, dilations form a generalisation of reflections and projections. We have

DilL,0 = ProjL , DilL,°1 = RefL , DilL,1 = I .

Exercise 30 a Show that a dilation matrix with k = 0 is a projection matrix.

b Show that a dilation matrix with k = °1 is a reflection matrix.

Exercise 31
Use the fact that DilL,k ± DilL,m = DilL,km to show the following.

a A reflection is an involution, i.e. (RefL )2 = I .

° ¢2
b A projection is an idempotent, i.e. ProjL = ProjL .

Exercise 32
We can characterise dilation, projection and reflection matrices as follows. Let M =
2 3
a b
4 5. Show that the linear transformation corresponding to M is:
c d

a a projection onto a line if and only if M is symmetric (i.e. M equals its transpose
M T ), ad ° bc = 0, and a + d = 1;
{32} • Transformations AMSI module
b a reflection in a line if and only if M is symmetric, a + d = 0, and ad ° bc = °1;
c a dilation if and only if M is symmetric and a + d = 1 + ad ° bc.

Spiral symmetries

A spiral symmetry is another interesting type of plane transformation. The spiral sym-
metry about a point P with factor r and angle µ is given by by dilating the whole plane by
a factor r from P , followed by a rotation by µ around P .

A spiral symmetry about the origin thus dilates a point (x, y) to (r x, r y) and then rotates
it by µ, to give the result
2 32 2 3 3 2 32 3
cos µ ° sin µ rx
r x cos µ ° r y sin µ r cos µ °r sin µ x
4 54 5 = 4 5=4 54 5.
sin µ cos µ ry r x sin µ + r y cos µ r sin µ r cos µ y

Example
2 3
1 °1
Let M = 4 5. Describe the linear transformation T M geometrically.
1 1

Solution

Form the parallelogram spanned by T M (1, 0) = (1, 1) and T M (0, 1) = (°1, 1), and view T M
p
as in previous examples. We see T M sends the unit square to a square of side length 2,
rotated by º/4.

y y0

TM
x x0

p
Thus T M is a spiral symmetry about the origin with factor r = 2 and angle µ = º/4;
indeed, M agrees with the matrix above for these values of r and µ:
2 3 2p p 3 2 3
1 °1 2 cos º4 ° 2 sin º4 r cos µ °r sin µ
M =4 5=4
p p
5=4 5.
1 1 2 sin º4 2 cos º4 r sin µ r cos µ

One reason spiral symmetries are interesting is because they arise naturally with complex
numbers. Consider a complex number w in polar form w = r cisµ = r (cos µ + i sin µ),
 A guide for teachers – Years 11 and 12 • {33}
where r = |w| and µ = arg w. As the complex numbers form the Argand plane C ª
= R2 ,
we may regard multiplication by w as a plane transformation Multw : R2 ! R2 , given by
Multw (z) = w z. That is, Multw multiplies complex numbers by w. If z = x + yi then
° ¢ ° ¢
w z = r (cos µ + i sin µ)(x + yi ) = r x cos µ ° r y sin µ + i r x sin µ + r y cos µ .

If we write out the real and imaginary parts as components of a vector, this means
2 3 2 3 2 32 3
x r x cos µ ° r y sin µ r cos µ °r sin µ x
Multw 4 5 = 4 5=4 54 5.
y r x sin µ + r y cos µ r sin µ r cos µ y

This is exactly a spiral symmetry! Hence, the effect of multiplication by a complex num-
ber w on the complex plane is a spiral symmetry about the origin 0 with factor |w| and
angle arg µ.

Transformations of graphs

A common type of problem asks you to transform a graph y = f (x), and find the equation
of the transformed graph. We illustrate with an example.
Example

Consider the graph y = x 2 . Suppose the graph is dilated from the x-axis by a factor of 2,
and then translated 3 units to the right. What is the equation of the resulting graph?

Solution

The transformation given is Trans(3,0) ± Dilx-axis,2 . Denote by (x 0 , y 0 ) the image of a point

(x, y) under this transformation. A point (x, y) is sent to (x, 2y) by the dilation, and then
to (x + 3, 2y) by the translation, so (x 0 , y 0 ) = (x + 3, 2y).

The original graph consists of points (x, y) such that y = x 2 . Then it is transformed so
that (x, y) 7! (x 0 , y 0 ) = (x + 3, 2y). We must find what equation is satisfied by x 0 and y 0 .

Since x 0 = x + 3 and y 0 = 2y, we have x = x 0 ° 3 and y = y 0 /2. Thus y = x 2 implies

y0
= (x 0 ° 3)2 or equivalently y 0 = 2(x 0 ° 3)2 .
2

Thus, after the transformation, the graph consists of points (x, y) satisfying y = 2(x ° 3)2 .
This is the equation of the transformed graph.

Exercise 33
{34} • Transformations AMSI module
In the above example, suppose the translation was applied first, then the dilation. Would
the answer have been different?

The method in this example can be used generally to find an equation for a graph subject
to a transformation F : R2 ! R2 .

• Write (x 0 , y 0 ) = F (x, y), and invert this equation to write x and y as expressions in terms
of x 0 and y 0 .
• Substitute these expressions for x and y into the equation of the graph to obtain an
equation in x 0 and y 0 .
• The transformed graph then consists of points (x 0 , y 0 ) satisfying this equation in x 0 and
y 0 . Dropping the primes, we have an equation in x and y for the transformed graph.
Example

Consider the graph y = e x . Suppose the graph is dilated from the y-axis by a factor of 3,
reflected in the x-axis, and then translated 1 unit to the left and 2 units down. What is
the equation of the resulting graph?

Solution

A point (x, y) maps to (3x, y) under the dilation, then to (3x, °y) under the reflection,
then to (3x ° 1, °y ° 2) under the translation. So (x 0 , y 0 ) = (3x ° 1, °y ° 2). Inverting,
x 0 +1
we obtain x = 3 and y = °y 0 ° 2. Substituting these expressions into y = e x gives
x 0 +1
°y 0 ° 2 = e 3 . So the equation of the transformed graph is
1
y = °2 ° e 3 (x+1) .

Exercise 34
Suppose that in the above example the transformations were applied in the reverse order:
translations, then reflection, then dilation. Is the answer any different?

Commonly, a graph y = f (x) is transformed by a combination of dilations from the axes,

and a translation. The following theorem gives a formula for the resulting graph.

Theorem
Suppose the graph of y = f (x) is transformed by a dilation of factor k from the x-axis,
factor h 6= 0 from the y-axis, and then translated by the vector (c, d ) (i.e. c units to the
right and d units up). The resulting graph has equation
µ ∂
1
y =kf (x ° c) + d .
h
 A guide for teachers – Years 11 and 12 • {35}
Note that if there is a reflection in the x-axis, then we can take k to be negative; and if
there is a reflection in the y-axis, we can take h to be negative.

Proof
The point (x, y) is mapped to (hx, k y) under the dilations, then to (hx + c, k y +
d ) under translation, so x 0 = hx + c and y 0 = k y + d . Inverting these equations
(assuming h, k 6= 0) gives

1° 0 ¢ 1° 0 ¢
x= x °c and y= y °d .
h k
The graph with equation y = f (x) is thus taken under the transformation to points
(x 0 , y 0 ) such that
µ ∂ µ ∂
1° 0 ¢ 1° 0 ¢ 1° 0 ¢
y °d = f x °c , which simplifies to y0 = k f x ° c + d.
k h h

We may now drop the primes and we obtain the desired equation.

There is also the possibility k = 0. In this case dilation projects the graph to y = 0,
then translation takes it to y = d , agreeing with the formula.

While questions involving translations and dilations in the axes are most common, the
same method applies to general transformations.
Example

The hyperbola given by y = 1/x is rotated by 45± clockwise about the origin. What is the
equation of the resulting curve?

Solution

The point (x 0 , y 0 ) is obtained by rotating (x, y) by ° º4 , and (x, y) is obtained by rotating

(x 0 , y 0 ) by º
4 about the origin. Thus
2 3 2 32 3 2 32 3 2 3
x cos º4 ° sin º4 x0 p1 °1
p x0 0
1 4x ° y 5
0
4 5=4 54 5 = 4 2 254 5 =
p .
y sin º4 cos º4 y0 p1 p1 y0 2 x0 + y 0
2 2

Hence the curve y = 1/x, or x y = 1, is mapped to points (x 0 , y 0 ) such that

1 0
(x ° y 0 )(x 0 + y 0 ) = 1, that is, x 02 ° y 02 = 2.
2

Thus the equation of the rotated hyperbola is x 2 ° y 2 = 2.

{36} • Transformations AMSI module
y y0

1
y= x
Rotate
x 02 ° y 02 = 2
x x0

Exercise 35
Take the graph y = e x . Consider either translating it k units to the left, or dilating it by e k
from the x-axis. Show that either transformation yields the same graph.

Exercise 36
The unit circle x 2 + y 2 = 1 is dilated by a factor of 2 from the x-axis, then rotated by º/4
about the origin. What sort of curve results? What is its equation?

Exercise 37
Let f : R ! R be a bijective function. Show that after reflection in the line y = x, the graph
of y = f (x) is sent to the graph of y = f °1 (x).

Links Forward

Areas and determinants

The determinant of a 2 £ 2 matrix is given by

2 3
a b
det 4 5 = ad ° bc.
c d

The determinant has a nice geometric interpretation in terms of areas, and this interpre-
tation can also be applied to linear transformations.

We discussed earlier (in “The effect of a linear transformation”) how two vectors v and
w can be used to construct a parallelogram,
h i the parallelogram spanned by v and w. We
discussed how, if the matrix M = v w has v and w as its columns, then the linear
transformation T M sends the unit square to this parallelogram. We also discussed the
orientation of this parallelogram and how T M can preserve or reverse orientation.
 A guide for teachers – Years 11 and 12 • {37}
We now ask: what is the area of this parallelogram? It is closely related to the determinant
of the matrix M .

For instance, if v = (1, 0) and w = (0, 1), then v and w span the y
unit square, and these two vectors form the matrix
2 3
1 0 w
M =4 5, which has determinant 1.
0 1 x
v

Similarly, if v = (2, 0) and w = (3, °5), then we obtain a paral-

lelogram with base length 2 and height 5, hence with area 10;
and the corresponding matrix is y
v
2 3 x
2 3
M =4 5, which has determinant °10.
0 °5

The determinant equals the negative area. But in this case w

the parallelogram is negatively oriented and the linear trans-

formation reverses orientation.
h i
In fact, the determinant of M = v w gives the signed area of the parallelogram spanned
by v and w, i.e. positive or negative area according to orientation.

The linear transformation T M sends the unit square to this parallelogram, and in fact
takes the tessellation of the plane by unit squares by a tessellation by these parallelo-
grams. So unit squares of area 1 are taken to parallelograms of signed area det M . Thus
T M expands areas by a factor of | det M |, and preserves or reverses orientation accordingly
as det M is positive or negative. For any region R in the plane, its image T M (R) under T M
has area | det M | times the area of R.

Exercise 38
h i
In this exercise we prove that, if M = v w , then det M is signed area of the parallelo-
gram spanned by v and w. Let A(v, w) be this signed area.

a Show that for any real number k, A(v, w) = A(v, w + kv) and A(v, w) = A(v + kw, w).
b Show that A(v, w) = °A(w, v).
02 3 2 31 02 3 2 31
a b a b
c Show that A @4 5 , 4 5A = ad and A @4 5 , 4 5A = °bc.
0 d c 0
02 3 2 31 02 3 2 31
a b a ° bc
d 5 4 b 5A
d Show that, if d 6= 0, then A @4 5 , 4 5A = A @4 , = ad ° bc.
c d 0 d
{38} • Transformations AMSI module
h i
e Conclude that A(v, w) = det M , where M = v w .

Example

Find the determinant of the matrix corresponding to the rotation Rot0, º3 of º/3 about the
origin, and explain what this means for areas and orientations.

Solution

The corresponding rotation matrix (with ¡ = º/3) is

2 3 2 p 3
1 3
cos ¡ ° sin ¡ 2 2 5
M= 4 5 = 4 p , which has det M = 1.
sin ¡ cos ¡ ° 23 1
2

Thus the linear transformation Rot0,º/3 preserves areas and orientations.

In fact, any rotation preserves areas, so any rotation matrix has determinant 1. Simi-
larly, reflections preserve area, but reverse orientation, so reflection matrices have de-
terminant °1. Projections collapse all areas down to zero, so have determinant zero.

Exercise 39
Prove these facts. Show any rotation matrix has determinant 1, any projection matrix has
determinant 0, and any reflection matrix has determinant °1.

What about dilation? Dilation from a line L by factor k expands lengths by a factor of k
in the direction perpendicular to L, but leaves lengths unchanged in the direction of L.
Thus, areas expand by a factor of k.
Example

Let L be the line y = x. Find the determinant of the matrix corresponding to dilation
from L by a factor of 3. Explain what this means for areas and orientations.

Solution

Using the formula for a dilation matrix, with k = 3 and ¡ = º/4, DilL,3 has matrix
2 3 2 3
1 + (k ° 1) sin2 ¡
(1 ° k) sin ¡ cos ¡ 2 °1
4 5=4 5 which has determinant 3.
(1 ° k) sin ¡ cos ¡ 1 + (k ° 1) cos2 ¡ °1 2

Hence DilL,k expands areas by a factor of 3, and preserves orientation.

 A guide for teachers – Years 11 and 12 • {39}
Exercise 40
Show that the determinant of a general dilation matrix is k.

Isometries
An isometry is a plane transformation which preserves lengths. B F (B )
That is, a plane transformation F : R2 ! R2 is an isometry if,
for any points A and B in the plane, the distance between A F
F (A)
and B equals the distance between F (A) and F (B ). A

Isometries are important types of transformations, as they preserve geometry. All lengths
must remain the same, and hence angles and areas must also remain the same.

Some of the transformations we have seen are isometries, and others are not.

Lemma
Any rotation is an isometry. That is, for any point P and any angle µ, RotP,µ is an isometry.

Proof

Take two distinct points A and B . Let A 0 , B 0

be their images under rotation. We must show A0
AB = A 0 B 0 .
B0 ¡ B
Suppose A, B are distinct from P . Consider tri-
A
angles P AB and P A 0 B 0 . As A 0 is obtained from P
A by a rotation about P , P A = P A 0 . Similarly
P B = P B 0 . Since A 0 , B 0 are obtained by rotating
A, B by ¡ about P , \ AP B = \ A 0 P B 0 . Hence 4P AB ª
= 4P A 0 B 0 , so AB = A 0 B 0 .

In the case A = P , we have A 0 = P so P B = P B 0 implies AB = A 0 B 0 . The case B = P

is similar.

Exercise 41
Show that any translation or reflection is an isometry, but that any projection onto a line,
or dilation from a line by factor k 6= ±1 is not an isometry.

An isometry must preserve areas, so if we have an isometry which is a linear transfor-

mation corresponding to a matrix M , then det M = ±1. However, the converse is not
true. There are transformations which preserve areas, but which are not isometries.
{40} • Transformations AMSI module
Exercise 42
2 3
2 0
Let M = 4 5. Show that T M preserves areas, but is not an isometry.
1
0 2

If you compose two isometries, the result must be another isometry. It is interesting to
think abut what you get by composing various types of isometries.

For instance, composing a reflection in L with itself results in the identity. But what if we
take two reflections in different lines L 1 and L 2 ?
Example

Let L 1 be the line x = 0, and L 2 the line x = d . Show that reflection in L 1 followed by L 2 is
translation by (2d , 0).

Solution

We note that RefL 1 (x, y) = (°x, y) and RefL 2 (x, y) = (2d ° x, y) (see exercise ??), so

RefL 2 ± RefL 1 (x, y) = RefL 2 (°x, y) = (x + 2d , y) = (x, y) + (2d , 0),

which is translation by (2d , 0).

In general, if L 1 and L 2 are parallel, distance d apart, then the composition RefL 2 ± RefL 1
is a translation by distance 2d in a direction perpendicular to L 1 and L 2 .

What if L 1 and L 2 intersect at a point P ? Then it turns out that reflection in L 1 , followed
by reflection in L 2 , gives a rotation.
Example

Let L 1 be the x-axis and L 2 the line through the origin making an angle of ¡ with the
x-axis. Show RefL 2 ± RefL 1 is a rotation, and find its angle.

Solution

Denote points by polar coordinates (r, µ). Then reflection in L 1 sends a point (r, µ) to
(r, °µ), and reflection in L 2 sends (r, µ) to (r, 2¡ ° µ). Performing the two reflections in
order sends (r, µ) 7! (r, °µ) 7! (r, µ + 2¡). So the composition is a rotation by angle 2¡
about the origin:

RefL 2 ± RefL 1 = Rot0, 2¡ .

 A guide for teachers – Years 11 and 12 • {41}
Exercise 43
Give an alternative proof of this statement by showing that the product of the matrices
for reflection in L 1 and L 2 is a rotation matrix of angle 2¡.

If you compose a translation of some distance along a line L, with reflection in L, you
obtain a type of isometry known as a glide reflection.

y y0
L

Glide reflection
A0

x x0

We have now seen many examples of isometries: rotations, translations reflections, glide
reflections, and the identity. We might ask if there are any others. The following theorem
says that there are not: we have now seen them all.

Theorem (Classification of isometries of the plane)

Suppose F : R2 ! R2 is an isometry. Then F is a rotation, translation, reflection, glide
reflection or the identity.

We will not prove this theorem. A proof can be found, for instance, in I. M. Yaglom,
Geometric Transformations I.

Three-dimensional and higher-dimensional transformations

All of our discussion in this module has been about two-dimensional transformations —
transformations of the plane. But much of our reasoning applies to three dimensions,
and even to higher dimensions!

Just as a plane transformation is a function F : R2 ! R2 , a spatial transformation is a

function F : R3 ! R3 . Just as a linear plane transformation is given by a 2 £ 2 matrix, a
linear spatial transformation is given by a 3 £ 3 matrix.
{42} • Transformations AMSI module
In three dimensions, we can consider translations by 3-dimensional vectors, rotations
about a line, reflections in a plane, projections onto a plane, and dilations from a plane.

Similar methods apply as for the two-dimensional case. Just as there is a standard basis
{(1, 0), (0, 1)} of R2 , there is a standard basis {(1, 0, 0), (0, 1, 0), (0, 0, 1)} of R3 . A linear trans-
formation T M takes the three standard basis vectors to the three columns of the matrix
M . We can visualise T M as taking the tessellation of R3 by unit cubes, to a tessellation
of R3 by parallelepipeds (the 3-d version of parallelograms), where each parallelepiped is
spanned by the three column vectors of M . The determinant of M now gives the volume
expansion of T M ; its sign again tells us if T M preserves or reverses orientation.

In essence, everything we have said in this module can be generalised to R3 . And in

principle it can be generalised to any dimension Rn — it’s just a little harder to visualise!

Answers to exercises
Exercise 1
If X is on L then X 0 = X , and the closest point on L to X is X itself. So assume L is on
X . For any point Y on L other than X 0 , X X 0 Y forms a triangle with a right angle at X 0 , so
X Y > X X 0 . Hence X 0 is the closest point to X on L.

Exercise 2
Take any vector v and c = 0 in F (cv) = cF (v) to obtain F (0) = 0.

Exercise 3 a Rotation º/3 anticlockwise about the origin.

b Dilation by factor 7 in all directions from the origin.
c Rotation by º about the origin.
d Dilation from the x-axis with factor 3, and dilation from the y-axis with factor 2.
e Reflection in the line y = x.

Exercise 4
We compute f ± f (x, y) = f (y, °x + y) = (°x + y, °x) and then f 3 (x, y) = f ± f ± f (x, y) =
f (°x+y, °x) = (°x, °y). Now f 6 is obtained by composing f 3 with itself twice, so f 6 (x, y) =
f 3 ( f 3 (x, y)) = f 3 (°x, °y) = (x, y). Thus f 6 is the identity.

Exercise 5
 A guide for teachers – Years 11 and 12 • {43}
2 32 3 2 3 2 3
1 a b d °b 1 ad ° bc 0 1 0
M M °1 = 4 54 5= 4 5=4 5 = Id
ad ° bc c d °c a ad ° bc 0 ad ° bc 0 1

The computation for M °1 M is similar.

Exercise 6
Note that x 0 = x + 2y, y 0 = 2x + 4y implies y 0 = 2x 0 , so the image of F lies on the line y 0 =
2x 0 . So F is not surjective: for instance, there is no (x, y) such that F (x, y) = (0, 1); if there
were, then x + 2y = 0 and 2x + 4y = 1, but then 2x + 4y = 2(x + 2y) = 0, a contradiction.
Alternatively, F is not injective as, for instance, F (0, 0) = F (°2, 1) = (0, 0).

Exercise 7
2 3 2 3 2 32 3°1
3 °1 1 2 1 2 3 °1
Let M be the desired matrix, so M 4 5=4 5. Thus M = 4 54 5 =
1 2 2 °3 2 °3 1 2
2 32 3 2 3 2 3
14
1 2 2 1 0 7 0 1
54 5= 14 5=4 5. Hence T M (x, y) = (y, x ° y).
7 7
2 °3 °1 3 7 °7 1 °1

Exercise 8
The x and y-axes are sent to the y-axis and negative x-axis respectively, so a point (x, y)
is mapped to (°y, x).

Exercise 9
The point (0, 1) is on the unit circle at an angle of º/2 from the positive x-axis. Rotating
by ¡ gives the point (cos(¡ + º2 ), sin(¡ + º2 )) = (° sin ¡, cos ¡).

Exercise 10
We verify
2 32 3 2 3
cos µ ° sin µ cos ¡ ° sin ¡ cos µ cos ¡ ° sin µ sin ¡ ° cos µ sin ¡ ° sin µ cos ¡
4 54 5=4 5
sin µ cos µ sin ¡ cos ¡ sin µ cos ¡ + cos µ sin ¡ ° sin µ sin ¡ + cos µ cos ¡

which by sine and cosine addition formulas is the rotation matrix for angle µ + ¡. Hence
the composition of rotations by µ and ¡ is rotation by µ + ¡.

Exercise 11
If ¡ is not a multiple of 2º, then a rotation of ¡ about P sends no point to itself except P .
{44} • Transformations AMSI module
So if P 6= 0 then RotP, ¡ (0) 6= 0. By exercise ?? a linear transformation must send the origin
to the origin, so RotP, ¡ cannot be linear.

Exercise 12
Translations by v and °v undo each other; they are inverse transformations. Thus Transv
is bijective, with inverse Trans°v .

Exercise 13
Let x be a point in the plane. Translation by °v takes P to the origin and carries x along to
x°v. Then rotating by ¡ about 0 preserves 0 and rotates x°v around it. Then translating
by v moves the origin back to P and takes our point to the point rotated by ¡ about P .
Thus
2 3 2 3 2 32 3
x x cos ¡ ° sin ¡ x ° a
RotP, ¡ 4 5 = Transv ± Rot0, ¡ ± Trans°v 4 5 = Transv 4 54 5
y y sin ¡ cos ¡ y °b
2 3 2 3
(x ° a) cos ¡ ° (y ° b) sin ¡ (x ° a) cos ¡ ° (y ° b) sin ¡ + a
= Transv 4 5=4 5.
(x ° a) sin ¡ + (y ° b) cos ¡ (x ° a) sin ¡ + (y ° b) cos ¡ + b

Exercise 14 a Let Y be a point not on L. Then there is no X such that ProjL (X ) = Y ,

so ProjL is not surjective.

b Let L 0 be a line perpendicular to L, intersecting L at a point Y . Then for any point X

on L 0 , ProjL (X ) = Y . Hence ProjL is not injective.

c If P is on L, then the perpendicular to L through P intersects L at P , so ProjL (P ) = P .

d For any point X , ProjL (X ) lies on L, so using the previous part with P = ProjL (X ) we
° ¢
have ProjL ProjL (X ) = ProjL (X ).

Exercise 15
The line through (x 0 , y 0 ) with gradient °1 has equation y ° y 0 = °(x ° x 0 ), or equivalently
y = °x + x 0 + y 0 . At the intersection with the line y = x we have y = °x + x 0 + y 0 = x so
x = 12 (x 0 + y 0 ) and then y = x = 12 (x 0 + y 0 ) as desired.

Exercise 16
The line L given by y = 2x makes an angle ¡ with the positive x-axis such that tan ¡ = 2,
hence sin ¡ = p2 and cos ¡ = p1 . Thus
5 5
2 3 2 32 3 2 3
1 2 14
8 5 5 5 485
ProjL 4 5 = 4 = 455 .
2 4 28
3 5 5 3 5
 A guide for teachers – Years 11 and 12 • {45}
Exercise 17 a A proof may be given by explaining the following pictures.

cv v+w w

v
v
b Dropping a perpendicular from (1, 0) to L as shown we obtain a right-angled triangle
with hypotenuse 1 and angle ¡; so the projection of (1, 0) to L is a distance of cos ¡
along L from the origin. Lying on a ray at angle ¡ at a distance cos ¡ from the origin,
then, ProjL (1, 0) = cos ¡(cos ¡, sin ¡) = (cos2 ¡, sin ¡ cos ¡). Similarly ProjL (0, 1) lies at
distance sin ¡ from the origin on L, so equals (sin ¡ cos ¡, sin2 ¡).

c The first part shows ProjL is linear, hence given by a matrix. The second part finds
the columns of this matrix. So Projl is as claimed.

Exercise 18 a If P = (x, y) is on L, then (x, y) = (s cos ¡, s sin ¡) for some s. Hence

2 32 3 2 3 2 3
cos2 ¡ sin ¡ cos ¡ s cos ¡ s cos ¡(cos2 ¡ + sin2 ¡) s cos ¡
ProjL (P ) = 4 54 5=4 5=4 5 = P.
2 2 2
sin ¡ cos ¡ sin ¡ s sin ¡ s sin ¡(cos ¡ + sin ¡) s sin ¡

b We compute
2 32 2 3
cos2 ¡ sin ¡ cos ¡ cos2 ¡(cos2 ¡ + sin2 ¡) sin ¡ cos ¡(cos2 ¡ + sin2 ¡)
4 5 =4 5
sin ¡ cos ¡ sin2 ¡ sin ¡ cos ¡(cos2 ¡ + sin2 ¡) sin2 ¡(cos2 ¡ + sin2 ¡)
2 3
cos2 ¡ sin ¡ cos ¡
=4 5.
2
sin ¡ cos ¡ sin ¡

As this is the matrix for ProjL , we conclude ProjL ± ProjL = ProjL as desired.

Exercise 19
Under ProjL , points on L are sent to themselves, but any point not on L is sent to a differ-
ent point. So if L does not pass through the origin 0, then ProjL (0) 6= 0. By exercise ??, a
linear transformation must send 0 to 0. Hence ProjL is not linear.

Exercise 20 a Rotating by °¡ about the origin brings L to the x-axis, carrying along a
point P . Then reflecting in the x-axis flips the point over the x-axis; rotating back by
¡ sends the x-axis back to L, and carries our point to the reflection of P in L. Hence
RefL is given by the desired composition of transformations.

b As Rot0,¡ , Refx-axis and Rot0,°¡ are all linear, and we know their matrices, we can
multiply them together to find a matrix for RefL , which must also be linear. The
{46} • Transformations AMSI module
matrix is
2 32 32 3 2 32 3
cos ¡ ° sin ¡ 1 0 cos ¡ sin ¡ cos ¡ ° sin ¡ cos ¡ sin ¡
4 54 54 5=4 54 5
sin ¡ cos ¡ 0 °1 ° sin ¡ cos ¡ sin ¡ cos ¡ sin ¡ ° cos ¡
2 3 2 3
cos2 ¡ ° sin2 ¡ 2 sin ¡ cos ¡ cos(2¡) sin(2¡)
=4 5=4 5.
2 sin ¡ cos ¡ sin2 ¡ ° cos2 ¡ sin(2¡) ° cos(2¡)

Exercise 21 a A proof may be given by explaining the following diagrams.

w
cv v+w
v v

b Since (1, 0) is distance 1 from the origin on an angle of 0, reflection in a line at angle ¡
sends (1, 0) to a point distance 1 from the origin on an angle of 2¡. Thus RefL (1, 0) =
(cos(2¡), sin(2¡)). Similarly one can show RefL (0, 1) = (sin(2¡), ° cos(2¡)).

c The first part shows RefL is linear; the second part calculates the columns of the
corresponding matrix, which is as claimed.

Exercise 22
The line L given by y = 2x passes through the origin on an angle ¡ such that tan ¡ = 2,
hence sin ¡ = p2 and cos ¡ = p1 . Hence cos(2¡) = cos2 ¡ ° sin2 ¡ = °3
and sin(2¡) =
5 5 5
4
2 sin ¡ cos ¡ = 5 . So reflection in L is given by
2 3 2 32 3
°3 4
x 5 5 5 4x 5
RefL 4 5 = 4 .
4 3
y 5 5 y

Exercise 23
The vector from ProjL (x) to x, and the vector from ProjL (x) to RefL (x), have the same
length and point in opposite directions. Hence
° ¢
x ° ProjL (x) = ° RefL (x) ° ProjL (x) ,

which rearranges to RefL (x) = 2ProjL (x) ° x, as desired.

Exercise 24
From the previous exercise, the matrix for RefL is given by twice the matrix for ProjL ,
 A guide for teachers – Years 11 and 12 • {47}
minus the identity matrix, which is
2 3 2 3 2 3 2 3
cos2 ¡ sin ¡ cos ¡ 1 0 2 cos2 ¡ ° 1 2 sin ¡ cos ¡ cos(2¡) sin(2¡)
24 5°4 5=4 5=4 5.
sin ¡ cos ¡ sin2 ¡ 0 1 2 sin ¡ cos ¡ 2 sin2 ¡ ° 1 sin(2¡) ° cos(2¡)

Exercise 25 a If P = (x, y) lies on L then (x, y) = (s cos ¡, s sin ¡) for some s, so

2 32 3 2 32 3
cos(2¡) sin(2¡) x cos(2¡) sin(2¡) s cos ¡
4 54 5 = 4 54 5
sin(2¡) ° cos(2¡) y sin(2¡) ° cos(2¡) s sin ¡
2 3 2 3 2 3
s cos(2¡ ° ¡) s cos ¡ x
=4 5=4 5=4 5
s sin(2¡ ° ¡) s sin ¡ y

2 32 3
cos(2¡) sin(2¡) cos(2¡) sin(2¡)
b We verify 4 54 5 is diagonal, with both diago-
sin(2¡) ° cos(2¡) sin(2¡) ° cos(2¡)
nal entries given by cos2 (2¡) + sin2 (2¡) = 1, hence is the identity matrix.

Exercise 26
We note that projection onto L is given by ProjL (x, y) = (d , y). Using exercise ?? then
RefL (x, y) = 2ProjL (x, y) ° (x, y) = 2(d , y) ° (x, y) = (2d ° x, y).

Exercise 27 a Rotating by °¡ about the origin brings L to the x-axis, and brings a
point P along with it; then dilating from the x-axis and rotating back by ¡ brings our
line back to L and sends our point to its dilation from L.
b The three transformations in the previous part are all linear; multiplying them will
give a matrix for DilL,k , which must be linear.
2 32 32 3 2 32 3
cos ¡ ° sin ¡ 1 0 cos ¡ sin ¡ cos ¡ ° sin ¡ cos ¡ sin ¡
4 54 54 5=4 54 5
sin ¡ cos ¡ 0 k ° sin ¡ cos ¡ sin ¡ cos ¡ °k sin ¡ k cos ¡
2 3 2
cos2 ¡ + k sin2 ¡
(1 ° k) sin ¡ cos ¡ 1 + (k ° 1) sin2 ¡ (1 ° k) sin ¡ cos ¡
=4 5=4
(1 ° k) sin ¡ cos ¡ sin2 ¡ + k cos2 ¡ (1 ° k) sin ¡ cos ¡ 1 + (k ° 1) cos2 ¡

Exercise 28
The line L given by y = 2x makes angle ¡ with the x-axis, where tan ¡ = 2, so sin ¡ = p2
5
and cos ¡ = p1 . Hence
5
2 3 2 32 3 2 32 3 2 3
3 1 + 4 · 45 °4 · 25 3 21 °8
5 5 435
63
DilL,5 4 5 = 4 54 5 = 4 5 =4
5 5.
0 °4 · 25 1 + 4 · 15 0 °8
5
9
5 0 °24
5
{48} • Transformations AMSI module
Exercise 29
We first note that DilL,k has determinant
° ¢° ¢
1 + (k ° 1) sin2 ¡ 1 + (k ° 1) cos2 ¡ °(1°k)2 sin2 ¡ cos¡ = 1+(k°1) sin2 ¡+(k°1) cos2 ¡ = k.

Hence the inverse of the matrix for DilL,k is

2 3 2 ° ¢ ° ¢ 3
2 1 1 2 1
14 1 + (k ° 1) cos ¡ (k ° 1) sin ¡ cos ¡ + 1 ° cos ¡ 1 ° sin ¡ cos ¡
5=4k k k 5.
k (k ° 1) sin ¡ cos ¡ 1 + (k ° 1) sin2 ¡ ° ¢ ° ¢
1 ° k sin ¡ cos ¡ k + 1 ° k1 sin2 ¡
1 1

° ¢
Now k1 +(1° k1 ) cos2 ¡ = k1 (1°cos2 ¡)+cos2 ¡ = k1 sin2 ¡+1°sin2 ¡ = 1+ k1 ° 1 sin2 ¡ and,
° ¢ ° ¢
similarly, k1 + 1 ° k1 sin2 ¡ = 1+ k1 ° 1 cos2 ¡. Hence this matrix is the matrix for dilation
by factor 1/k from L.

Exercise 30 a When k = 0, the matrix for DilL,k becomes

2 3 2 3
1 ° sin2 ¡sin ¡ cos ¡ cos2 ¡ sin ¡ cos ¡
4 5 =4 5,
2 2
sin ¡ cos ¡ 1 ° sin ¡ sin ¡ cos ¡ cos ¡

which is the matrix for projection onto L.

b When k = °1 we have
2 3 2 3
1 ° 2 sin2 ¡2 sin ¡ cos ¡ cos(2¡) sin(2¡)
4 5=4 5
2 sin ¡ cos ¡ 1 ° 2 cos2 ¡ sin(2¡) ° cos(2¡)

which is the matrix for reflection in L.

Exercise 31 a A reflection is a dilation by factor °1, so applying it twice gives a dila-

tion with factor (°1)2 = 1, i.e. the identity.

b A projection is a dilation by factor 0, so applying it twice gives a dilation with factor

02 = 0, i.e. the same projection.

Exercise 32 a A projection matrix clearly has the desired properties. For the con-
verse, suppose M is symmetric (b = c), ad = b 2 and a + d = 1. Since a and d have
sum and product ∏ 0, both are ∏ 0; since they sum to 1, both lie in [0, 1]. Hence we
have a = cos2 ¡ and b = sin2 ¡ for some ¡. Then b 2 = cos2 ¡ sin2 ¡ so b = sin ¡ cos ¡
or ° sin ¡ cos ¡. In the first case we have the matrix for projection onto a line at angle
¡; in the second case for projection onto a line at angle °¡.

b A reflection matrix clearly has the desired properties. For the converse, suppose M
is symmetric (b = c), a = °d and ad ° b 2 = °1. Substituting a = °d then gives °a 2 °
b 2 = °1, so (a, b) lies on he unit circle and hence a = cos(2¡), b = sin(2¡) for some ¡.
We then have a reflection matrix.
 A guide for teachers – Years 11 and 12 • {49}
c A dilation matrix can be verified to have the desired properties. For the converse,
suppose b = c and a + d = 1 + ad ° b 2 . Let ad ° b 2 = k, so a + d = k + 1. Since (a °
1) + (d ° 1) = k ° 1 and (a ° 1)(d ° 1) = ad ° a ° d + 1 = b 2 , both a ° 1 and d ° 1 have
the same sign as k ° 1. Since they sum to k ° 1, they lie between 0 and k ° 1. Thus
we may write a ° 1 = (k ° 1) sin2 ¡ and d ° 1 = (k ° 1) cos2 ¡ for some ¡. We then have
b 2 = ad ° k = [1 + (k ° 1) sin2 ¡][1 + (k ° 1) cos2 ¡] ° k = (k ° 1)2 sin2 ¡ cos2 ¡. Thus
b = ±(1 ° k) sin ¡ cos ¡. Replacing ¡ with °¡ if necessary, we have a dilation matrix.

Exercise 33
Applying the translation we have (x, y) 7! (x + 3, y), then applying the dilation we have
(x + 3, y) 7! (x + 3, 2y) = (x 0 , y 0 ). So the transformation is the same, and the answer is the
same.

Exercise 34
Applying the translation we have (x, y) 7! (x + 1, y + 3); then applying the reflection we
have (x + 1, y + 3) 7! (°x ° 1, y + 3); then applying the dilation we have (°x ° 1, y + 2) 7!
(°3x ° 3, y + 2). The result is a different transformation, giving a different answer.

Exercise 35
Translating k units to the left gives y = e x+k . Dilating by e k form the x-axis gives y = e k e x .
These are the same graph since e x+k = e k e x .

Exercise 36
From (x, y), dilating and then rotating gives an ellipse; the transformation is
2 3 2 32 32 3 2 p 32 3
x0 p1 °1
p 1 0 x p1 ° 2 x
4 5=4 2 254 54 5 = 4 2 54 5
p
y0 p1 p1 0 2 y p1 2 y
2 2 2

and hence, inverting,

2 3 2 p 3°1 2 0 3 2p p 3 2 03 2 3
x p1 ° 2 x 1 2 2 x 1 x0 + y 0
4 5=4 2 5 4 5= 4 54 5 = p 4 5.
p 2 p
y p1 2 y0 °1 p1 y0 2 °1 x 0 + 1 y 0
2 2 22 2

Thus the expression x 2 + y 2 becomes

µ ∂
1 0 0 2 1 1 0 1 0 2 5 02 5 02 3 0 0
(x + y ) + ° x + y = x + y + x y .
2 2 2 2 8 8 4

The equation of the transformed circle is thus 58 x 2 + 58 y 2 + 34 x y = 1, which is an ellipse.

Exercise 37
{50} • Transformations AMSI module
Reflection in the line y = x takes (x, y) to (x 0 , y 0 ) = (y, x). The equation y = f (x) becomes
x 0 = f (y 0 ), and hence the transformed graph is given by x = f (y), or equivalently, y =
f °1 (x).

Exercise 38 a The parallelograms spanned by v, w and v, w + kv both have base v the

same height above this base, and the same orientation, so have the same signed area.
The same applies to v, w and v + kw, w above a base of w.

b If we reverse the order of v and w the parallelogram is unchanged but has its orien-
tation reversed, hence its signed area changes sign.

c The vectors (a, 0) and (b, d ) span a parallelogram with base a and perpendicular
height d , hence area |ad |. The parallelogram is positively oriented if a and d have
the same sign, and negatively oriented otherwise, so the signed area is ad . A similar
argument applies to (a, c) and (b, 0).

d Applying the first part to (a, c) and (b, d ), the signed area they span is equal to the
signed area spanned by (a, c) ° dc (b, d ) and (b, d ).

e For any vectors v = (a, c) and w = (b, d ), if d 6= 0 then the previous part applies and
gives the signed area as det M . If d = 0 then part c applies and gives the signed area
as °bc = ad ° bc = det M .

Exercise 39
2 3 2 3
2
cos ¡ ° sin ¡ cos ¡ sin ¡ cos ¡
We compute det 4 5 = cos2 ¡ + sin2 ¡ = 1, det 4 5=
sin ¡ cos ¡ sin ¡ cos ¡ sin2 ¡
2 3
cos(2¡) sin(2¡)
cos2 ¡ sin2 ¡ ° cos2 ¡ sin2 ¡ = 0 and det 4 5 = ° cos2 (2¡) ° sin2 (2¡) =
sin(2¡) ° cos(2¡)
°1.

Exercise 40
2 3
1 + (k ° 1) sin2 ¡
(1 ° k) sin ¡ cos ¡ £ §£ §
Compute det 4 5 = 1 + (k ° 1) sin2 ¡ 1 + (k ° 1) cos2 ¡ °
(1 ° k) sin ¡ cos ¡ 1 + (k ° 1) cos2 ¡
(1°k)2 sin2 ¡ cos2 ¡, which simplifies to 1+(k °1)(sin2 ¡+cos2 ¡)+(k °1)2 sin2 ¡ cos2 ¡°
(1 ° k)2 sin2 ¡ cos2 ¡ = 1 + (k ° 1) = k.

Exercise 41
Let A, B be points in the plane and let A 0 B 0 be their images under transformation; to
demonstrate an isometry we show AB = A 0 B 0 .

Under a translation, AB B 0 A 0 forms a parallelogram so AB = A 0 B 0 .

 A guide for teachers – Years 11 and 12 • {51}
For a reflection, let X be a point on the line of reflection. Then triangles X AB and X A 0 B 0
are congruent, so AB = A 0 B 0 .

For a projection, take A, B to be distinct points projecting onto the same point. Then
A 0 = B 0 so A 0 B 0 = 0 < AB . As AB 6= A 0 B 0 , a projection is not an isometry.

For a dilation, take A to be a point on the line of dilation and B a point which projects
onto A. Then A 0 B 0 = |k|AB 6= AB . Hence this transformation is not an isometry.

Exercise 42
Since det M = 1, T M preserves areas. But, for instance, let A = (0, 0) and B = (1, 0), so
A 0 = (0, 0) and B 0 = (2, 0). Then AB = 1 but A 0 B 0 = 2 so T M is not an isometry.

Exercise 43
We compute the matrix for the composition RefL 2 ± RefL 1 as
2 32 3 2 3
cos(2¡) sin(2¡) 1 0 cos(2¡) ° sin(2¡)
4 54 5=4 5,
sin(2¡) ° cos(2¡) 0 °1 sin(2¡) cos(2¡)

which is the matrix for rotation by 2¡ about the origin.

 3751-087634ΩÅℭ℗ℋℤ∬√
5984
ⅈↄ⅞
∭ ∜∱≄
ℨℶℜ ∾ ⋞6
⅘⅔
€€ ⋃ ⑦
ӛ ₴₦ ⋟⤍
Ҷҵ ⟼
ӕ Ӂ ⤮⍬
Ӡӓ

11 12
ԅ

⍰
ᴂ

⋓
Ѽ

⊚
⅙
Ҧ

58
ύҖ

38
348 7
ϝϚό

&
ЂЁϟ

756Ϛ
ХФУШЩЫЅ

όύҖ
ҦѼᴂԅӠӓӕӁҶҵ
56
73487

Years
∬√∜ ℤ
83 8

∱
5

≄
⅙

∾
⊚

⋞
⋓

6⋃
⑦
⍰

⋟ӛ
ℨℶ
ℜⅈ

₴
⟼
ↄ

₦€
⅞

⤮
⋟⤍
09 ∭
82 5 598
294 30 43 751-
6879 0876
ύҖҦѼᴂԅӠ4 €-20