mt1174 ch1-4

Calculus
J.M. Ward
MT1174, 2790174
2011
Undergraduate study in
Economics, Management,
Finance and the Social Sciences
This is an extract from a subject guide for an undergraduate course offered as part of the
University of London International Programmes in Economics, Management, Finance and
the Social Sciences. Materials for these programmes are developed by academics at the
London School of Economics and Political Science (LSE).
For more information, see: www.londoninternational.ac.uk
This guide was prepared for the University of London International Programmes by:
J.M. Ward, Department of Mathematics, London School of Economics and Political Science.
This is one of a series of subject guides published by the University. We regret that due to
pressure of work the author is unable to enter into any correspondence relating to, or arising
from, the guide. If you have any comments on this subject guide, favourable or unfavourable,
please use the form at the back of this guide.
University of London International Programmes

Publications Office
Stewart House
32 Russell Square
London WC1B 5DN
United Kingdom
Website: www.londoninternational.ac.uk
Published by: University of London
University of London 2011
The University of London asserts copyright over all material in this subject guide except where
otherwise indicated. All rights reserved. No part of this work may be reproduced in any form,
or by any means, without permission in writing from the publisher.
We make every effort to contact copyright holders. If you think we have inadvertently used
your copyright material, please let us know.
Contents
Contents
Preface
1 Introduction
1.1
This subject . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2
Reading . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3
Online study resources . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.1
The VLE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3.2
Making use of the Online Library . . . . . . . . . . . . . . . . . .
1.4
Using this guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.5
Examination advice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.6
The use of calculators . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
2 Functions
2.1
Introduction: What is a function? . . . . . . . . . . . . . . . . . . . . . .
2.1.1
Some elementary functions and their graphs . . . . . . . . . . . .
11
2.1.2
Combinations of functions . . . . . . . . . . . . . . . . . . . . . .
15
2.1.3
Inverse functions . . . . . . . . . . . . . . . . . . . . . . . . . . .
20
2.1.4
Identities
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
24
2.1.5
Applications of functions . . . . . . . . . . . . . . . . . . . . . . .
26
Conic sections . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
2.2.1
Parabolae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
30
2.2.2
Circles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
32
2.2.3
Ellipses . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
33
2.2.4
Hyperbolae . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
34
Learning outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
36
Solutions to activities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
36
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
45
Solutions to exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
46
2.2
3 Differentiation
3.1
Introduction: What is differentiation? . . . . . . . . . . . . . . . . . . . .
53
53
Contents
3.2
How to find derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . .
55
3.2.1
Standard derivatives . . . . . . . . . . . . . . . . . . . . . . . . .
56
3.2.2
The rules of differentiation . . . . . . . . . . . . . . . . . . . . . .
57
3.2.3
Higher-order derivatives . . . . . . . . . . . . . . . . . . . . . . .
65
Using derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
66
3.3.1
The meaning of the derivative . . . . . . . . . . . . . . . . . . . .
66
3.3.2
Tangent lines and linear approximations . . . . . . . . . . . . . .
68
3.3.3
Applications of derivatives . . . . . . . . . . . . . . . . . . . . . .
72
3.3.4
Existence of derivatives . . . . . . . . . . . . . . . . . . . . . . . .
74
Using higher-order derivatives . . . . . . . . . . . . . . . . . . . . . . . .
78
3.4.1
Maclaurin series . . . . . . . . . . . . . . . . . . . . . . . . . . . .
78
3.4.2
Taylor series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
85
87
88
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
96
97
3.3
3.4
4 One-variable optimisation
4.1
Introduction: What is optimisation? . . . . . . . . . . . . . . . . . . . . .
103
4.2
Using first-order derivatives . . . . . . . . . . . . . . . . . . . . . . . . .
104
4.2.1
Increasing and decreasing functions . . . . . . . . . . . . . . . . .
104
4.2.2
Stationary points . . . . . . . . . . . . . . . . . . . . . . . . . . .
106
4.2.3
An application: Elasticities revisited . . . . . . . . . . . . . . . . .
109
Using second-order derivatives . . . . . . . . . . . . . . . . . . . . . . . .
110
4.3.1
Second-derivatives and stationary points . . . . . . . . . . . . . .
110
4.3.2
Convex and concave functions . . . . . . . . . . . . . . . . . . . .
111
4.3.3
Points of inflection . . . . . . . . . . . . . . . . . . . . . . . . . .
113
Curve sketching . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
114
4.4.1
Sketching curves defined by polynomials . . . . . . . . . . . . . .
115
4.4.2
Sketching curves defined using other elementary functions . . . .
119
4.4.3
Asymptotes and cusps . . . . . . . . . . . . . . . . . . . . . . . .
121
Optimisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
123
4.5.1
Constrained optimisation . . . . . . . . . . . . . . . . . . . . . . .
125
4.5.2
What happens when differentiability fails? . . . . . . . . . . . . .
126
4.5.3
Applications of optimisation . . . . . . . . . . . . . . . . . . . . .
127
130
4.3
4.4
4.5
ii
103
Contents
131
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
136
138
5 Integration
145
5.1
Introduction: What is integration? . . . . . . . . . . . . . . . . . . . . . .
145
5.2
How to find indefinite integrals . . . . . . . . . . . . . . . . . . . . . . .
147
5.2.1
Standard integrals . . . . . . . . . . . . . . . . . . . . . . . . . .
147
5.2.2
The basic rules of integration . . . . . . . . . . . . . . . . . . . .
149
5.2.3
Integration by substitution . . . . . . . . . . . . . . . . . . . . . .
150
5.2.4
Integration by parts
. . . . . . . . . . . . . . . . . . . . . . . . .
158
5.2.5
Using partial fractions to simplify integrands . . . . . . . . . . . .
162
5.2.6
Using trigonometric identities to simplify integrands . . . . . . . .
167
5.3
. . . . . . . . . . . . . . . . . . . . . . . . .
170
5.3.1
Definite integrals and what they represent . . . . . . . . . . . . .
170
5.3.2
Definite integrals and the other rules of integration . . . . . . . .
178
Applications of integrals . . . . . . . . . . . . . . . . . . . . . . . . . . .
182
5.4.1
Marginal functions revisited . . . . . . . . . . . . . . . . . . . . .
182
5.4.2
Consumer and producer surpluses . . . . . . . . . . . . . . . . . .
183
186
187
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
195
196
5.4
Definite integrals and areas
6 Functions of several variables
201
6.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
201
6.2
Surfaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
202
6.2.1
Planes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
203
6.2.2
Contours and sections . . . . . . . . . . . . . . . . . . . . . . . .
204
Partial differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
210
6.3.1
Sections and partial derivatives . . . . . . . . . . . . . . . . . . .
211
6.3.2
Finding partial derivatives . . . . . . . . . . . . . . . . . . . . . .
212
6.3.3
The chain rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
214
6.3.4
An application: Homogeneous functions . . . . . . . . . . . . . . .
220
6.3.5
Second-order partial derivatives . . . . . . . . . . . . . . . . . . .
224
Using partial derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . .
226
6.3
6.4
iii
Contents
6.4.1
Tangent planes . . . . . . . . . . . . . . . . . . . . . . . . . . . .
226
6.4.2
Gradient vectors . . . . . . . . . . . . . . . . . . . . . . . . . . .
230
6.4.3
Directional derivatives . . . . . . . . . . . . . . . . . . . . . . . .
232
6.4.4
Implicitly defined functions of two variables . . . . . . . . . . . .
234
6.4.5
Taylor series . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
238
241
242
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
253
255
7 Two-variable optimisation
7.1
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
261
7.2
Unconstrained optimisation . . . . . . . . . . . . . . . . . . . . . . . . .
261
7.2.1
Stationary points . . . . . . . . . . . . . . . . . . . . . . . . . . .
262
7.2.2
Classifying stationary points . . . . . . . . . . . . . . . . . . . . .
264
7.2.3
Convex and concave functions . . . . . . . . . . . . . . . . . . . .
269
7.2.4
Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
272
Constrained optimisation . . . . . . . . . . . . . . . . . . . . . . . . . . .
275
7.3.1
Finding optimal points on the boundary of a region . . . . . . . .
277
7.3.2
The method of Lagrange multipliers . . . . . . . . . . . . . . . . .
279
7.3.3
The meaning of the Lagrange multiplier . . . . . . . . . . . . . .
282
7.3.4
Applications . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
284
289
290
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
294
296
7.3
8 Differential equations
303
8.1
Introduction: What is a differential equation? . . . . . . . . . . . . . . .
303
8.2
First-order ODEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
306
8.2.1
Separable first-order ODEs . . . . . . . . . . . . . . . . . . . . . .
307
8.2.2
Linear first-order ODEs . . . . . . . . . . . . . . . . . . . . . . .
308
8.2.3
Homogeneous first-order ODEs . . . . . . . . . . . . . . . . . . .
310
Second-order ODEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
312
8.3.1
Homogeneous second-order ODEs . . . . . . . . . . . . . . . . . .
312
8.3.2
Non-homogeneous second-order ODEs . . . . . . . . . . . . . . .
314
8.3
iv
261
Contents
8.4
Systems of first-order ODEs . . . . . . . . . . . . . . . . . . . . . . . . .
318
8.4.1
Simple systems of first-order ODEs . . . . . . . . . . . . . . . . .
319
8.4.2
Other systems of first-order ODEs . . . . . . . . . . . . . . . . . .
321
Applications of ODEs . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
323
8.5.1
Determining demand functions from elasticities . . . . . . . . . .
323
8.5.2
Continuous price adjustment . . . . . . . . . . . . . . . . . . . . .
324
8.5.3
Continuous cash flows . . . . . . . . . . . . . . . . . . . . . . . .
325
8.5.4
Market trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
327
327
328
Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
334
334
8.5
A Sample examination paper
339
B Solutions to the sample examination paper
341
Contents
vi
Preface
This subject guide is not a course text. It sets out a logical sequence in which to study
the topics in this subject. Where coverage in the main texts is weak, it provides some
additional background material.
I am grateful to Mark Baltovic for his careful reading of a draft of this guide and for his
many helpful comments.
Preface
Chapter 1
Introduction
In this very brief introduction, we aim to give you an idea of the nature of this subject
and to advise you on how best to approach it. We give general information about the
contents and use of this subject guide, and on recommended reading and how to use the
textbooks.
1.1
This subject
Calculus, as studied in this Level 1 course is primarily the study of derivatives and
integrals of functions of one variable and partial derivatives of functions of several
variables.
Our approach here is not just to help you acquire proficiency in techniques and
methods, but also to help you understand some of the theoretical ideas behind these.
For example, after completing this course, you will hopefully understand why the
derivatives of a function allow you to determine where a function of one variable is
optimised. In addition to this, we try to indicate the uses of some of the methods in
applications to economics, finance and related disciplines.
Aims of the course
The broad aims of this course are as follows:
to enable students to acquire skills in the methods of calculus (including
multivariate calculus), as required for their use in further mathematics subjects and
economics-based subjects;
to prepare students for further courses in mathematics and/or related disciplines.
As emphasised above, however, we do also want you to understand why certain
methods work: this is one of the skills that you should acquire. Indeed, the
examination will not simply test your ability to perform routine calculations, it will also
probe your knowledge and understanding of the principles that underlie the material.
Learning outcomes
We now state the broad learning outcomes of this course, as a whole. At the end of this
course and having completed the essential reading and activities, you should be able to:
1. Introduction
use the concepts, terminology, methods and conventions covered in the course to
solve mathematical problems in this subject;
solve unseen mathematical problems involving understanding of these concepts and
application of these methods;
see how calculus can be used to solve problems in economics and related subjects;
demonstrate knowledge and understanding of the underlying principles of calculus.
There are a couple of things that we should stress at this point. Firstly, note the
intention that you will be able to solve unseen problems. This means simply that you
will be expected to be able to use your knowledge and understanding of the material to
solve problems that are not completely standard. This is not something you should
worry unduly about: all topics in mathematics expect this, and you will never be
expected to do anything that cannot be done using the material of this course.
Secondly, we expect you to be able to demonstrate knowledge and understanding and
you might well wonder how you would demonstrate this in the examination. Well, it is
precisely by being able to grapple successfully with unseen, non-routine, questions that
you will indicate that you have a proper understanding of the topic.
Topics covered
Descriptions of the topics to be covered appear in the relevant chapters. However, it is
useful to give a brief overview at this stage.
We start by revising some of the basic ideas that are needed for the study of this course
and, in particular, the idea of a function of one variable. We then introduce derivatives
of such functions and how to find them using the techniques of differentiation. This
enables us to see how such functions are behaving and, in particular, enables us to see
where such functions are optimised. We then introduce integrals of such functions and
how to find them using the techniques of integration. In particular, this will enable us
to see how to relate functions to areas. We then introduce functions of several variables
and develop techniques for finding their partial derivatives. In particular, we will see
how we can use these ideas to see where these slightly more complicated functions are
optimised. Lastly, we introduce the idea of a differential equation and examine methods
for solving them.
Throughout this subject guide, the emphasis will be on the theory as much as on the
methods. That is to say, our aim in this subject is not only to provide you with some
useful techniques and methods from calculus, but to also enable you to understand why
these techniques work.
1.2
Reading
There are many books that would be useful for this subject. We recommend two in
particular, and a couple of others for additional, further reading. (You should note,
however, that there are very many books suitable for this course. Indeed, almost any
text on first-year university calculus will cover the majority of the material.)
1.2. Reading
Textbook reading is essential as textbooks will provide you with more in-depth
explanations than you will find in this subject guide, and they will also provide many
more examples to study and exercises to work through. The books listed are the ones
we have referred to in this guide.
Essential reading
Detailed reading references in this subject guide refer to the editions of the set
textbooks listed below. New editions of one or more of these textbooks may have been
published by the time you study this course. You can use a more recent edition of any
of the books; use the detailed chapter and section headings and the index to identify
relevant readings. Also check the virtual learning environment (VLE) regularly for
updated guidance on readings.
+ Binmore, K. and J. Davies Calculus: Concepts and Methods. (Cambridge:
Cambridge University Press, 2002, second revised edition) [ISBN 9780521775410].
+ Anthony, M. and N. Biggs Mathematics for economics and finance: methods and
modelling. (Cambridge: Cambridge University Press, 1996) [ISBN 9780521559133].
By and large we will be following Binmore and Davies but, sometimes, we will follow
the simpler treatment found in Anthony and Biggs. Both texts, when used wisely, will
provide you with a large number of examples for you to study and exercises for you to
attempt. It is recommended that you purchase both of these. Another thing you might
like to bear in mind is that some of the material from Binmore and Davies that we omit
here will be useful if you go on to study 176 Further Calculus.
Further reading
Once you have covered the essential reading you are then free to read around the
subject area in any text, paper or online resource. You will need to support your
learning by reading as widely as possible and by thinking about how these principles
apply in the real world. To help you read extensively, you have free access to the VLE
and University of London Online Library (see Section 1.3.2). However, two useful
textbooks that we have referred to in this subject guide are the following.
+ Simon, C.P. and L. Blume Mathematics for economists. (New York and London:
W.W. Norton and Company, 1994) [ISBN 9780393957334].
+ Adams, R.A. and C. Essex Calculus: a complete course. (Toronto: Pearson, 2010,
seventh edition) [ISBN 9780321549280].
Simon and Blume is a useful supplementary text with an emphasis on applications of
the material to economics; whereas Adams and Essex (which is merely an example from
a large range of very similar calculus textbooks) is a detailed calculus textbook which
contains much material which is beyond the scope of this course. Both of these texts are
suitable as sources of additional explanation, examples and exercises, but they are
probably not worth purchasing.
1. Introduction
1.3
Online study resources
In addition to the subject guide and the essential reading, it is crucial that you take
advantage of the study resources that are available online for this course, including the
virtual learning environment (VLE) and the Online Library.
You can access the VLE, the Online Library and your University of London email
account via the Student Portal at
http://my.londoninternational.ac.uk
You should receive your login details in your study pack. If you have not, or you have
forgotten your login details, please email uolia.support@london.ac.uk quoting your
student number.
1.3.1
The VLE
The VLE, which complements this subject guide, has been designed to enhance your
learning experience, providing additional support and a sense of community. It forms an
important part of your study experience with the University of London and you should
access it regularly.
The VLE provides a range of resources for EMFSS courses:
Self-testing activities: Doing these allows you to test your own understanding of
subject material.
Electronic study materials: The printed materials that you receive from the
University of London are available to download, including updated reading lists
and references.
Past examination papers and Examiners commentaries: These provide advice on
how each examination question might best be answered.
A student discussion forum: This is an open space for you to discuss interests and
experiences, seek support from your peers, work collaboratively to solve problems
and discuss subject material.
Videos: There are recorded academic introductions to the subject, interviews and
debates and, for some courses, audio-visual tutorials and conclusions.
Recorded lectures: For some courses, where appropriate, the sessions from previous
years Study Weekends have been recorded and made available.
Study skills: Expert advice on preparing for examinations and developing your
digital literacy skills.
Feedback forms.
Some of these resources are available for certain courses only, but we are expanding our
provision all the time and you should check the VLE regularly for updates.
1.4. Using this guide
1.3.2
Making use of the Online Library
The Online Library contains a huge array of journal articles and other resources to help
you read widely and extensively.
To access the majority of resources via the Online Library at
http://tinyurl.com/ollathens
you will either need to use your University of London Student Portal login details, or
you will be required to register and use an Athens login.
The easiest way to locate relevant content and journal articles in the Online Library is
to use the Summon search engine.
If you are having trouble finding an article listed in a reading list, try removing any
punctuation from the title, such as single quotation marks, question marks and colons.
For further advice, please see the online help pages at
www.external.shl.lon.ac.uk/summon/about.php
1.4
Using this guide
We have already mentioned that this guide is not a textbook. It is important that you
read textbooks in conjunction with the guide and that you try problems from the
textbooks. The exercises at the end of the main chapters of this subject guide are a very
useful resource and you should try them once you think you have mastered the material
from the chapter. You should really try these exercises before consulting the solutions,
as simply reading the solutions provided will not help you at all. Sometimes, the
solutions we provide will just be an overview of what is required, i.e. an indication of
how you should answer the questions, but in the examination, you must always show all
of your calculations. It is vital that you develop and enhance your problem-solving skills
and the only way to do this is to try lots of exercises.
1.5
Examination advice
Important: the information and advice given here are based on the examination
structure used at the time this guide was written. Please note that subject guides may
be used for several years. Because of this we strongly advise you to always check both
the current Regulations for relevant information about the examination, and the virtual
learning environment (VLE) where you should be advised of any forthcoming changes.
You should also carefully check the rubric/instructions on the paper you actually sit
and follow those instructions.
Remember, it is important to check the VLE for:
Up-to-date information on examination and assessment arrangements for this
course.
1. Introduction
Where available, past examination papers and Examiners commentaries for the
course which give advice on how each question might best be answered.
This course is assessed by a three hour unseen written examination. There are no
optional topics in this subject: you should study them all and this is reflected in the
structure of the examination paper. There are five questions (each worth 20 marks) and
all questions are compulsory. A sample examination paper may be found in an appendix
to this subject guide.
Please do not think that the questions in your real examination will necessarily be very
similar to the exercises in this subject guide or those in the sample examination paper.
The examination is designed to test you. You will get examination questions unlike the
questions in this subject guide. The whole point of examining is to see whether you can
apply your knowledge in familiar and unfamiliar settings. The Examiners (nice people
though they are) have an obligation to surprise you! For this reason, it is important
that you try as many examples as possible, from the subject guide and from the
textbooks. This is not so that you can cover any possible type of question the
Examiners can think of! It is so that you get used to confronting unfamiliar questions,
grappling with them, and finally coming up with the solution.
Do not panic if you cannot completely solve an examination question. There are many
marks to be awarded for using the correct approach or method.
1.6
The use of calculators
You will not be permitted to use calculators of any type in the examination. This is not
something that you should worry about: the Examiners are interested in assessing that
you understand the key concepts, ideas, methods and techniques, and will set questions
which do not require the use of a calculator.
Chapter 2
Functions
Essential reading
(For full publication details, see Chapter 1.)
Binmore and Davies (2002) Sections 2.12.6, 2.14 and part of 7.1.2.
Anthony and Biggs (1996) Chapters 1, 2 and parts of 7.
Further reading
Simon and Blume (1994) Sections 2.1, part of 2.2, 5.1, 5.3, and 5.4, Appendices
A1.1, parts of A1.2 and A2.16.
Adams and Essex (2010) Preliminaries parts of P.1P.7, parts of Sections 3.13.3
and 3.5.
Aims and objectives
The objectives of this chapter are as follows.
To introduce functions in general and the elementary functions and their graphs in
particular.
To see how to find combinations of functions and the inverse of a function (if it
exists).
To see how functions can be used in economics-based subjects.
To introduce conic sections and see how to draw them.
Specific learning outcomes can be found near the end of this chapter.
2.1
Introduction: What is a function?
NOTE: Before you start this chapter, you should make sure that you have
covered the background material in Chapter 1 of 173 Algebra.
Given two sets A and B, a function, f , from A to B is a rule which takes each element
of A and gives us a unique (or exactly one) element of B. We often express the fact that
the function f takes elements from A and gives us elements of B by writing
2. Functions
f : A B. In such cases, we call the sets A and B the domain and co-domain of the
function respectively.
One way of visualising a function f : A B is to think of it as a black box that takes

any x A, the domain, and applies the rule given by f to it to get the unique output
f (x) B, the co-domain, i.e.
x A f f (x) B.
Here, we have used x to denote the independent variable as we are free to choose any
element, x A from the domain. But, of course, the choice of x here is not essential as
it is just a dummy variable we could have used any other letter instead and said
that the function f : A B is a black box that takes any p A and applies the rule
given by f to it to get the unique output f (p) B.
It is often convenient to introduce another variable, called the dependent variable, to

stand for the elements of B that the function f : A B gives us. For instance, we
could say that this function takes any x A, the domain, and applies the rule given by
f to it to get the unique output y B, the co-domain, where y = f (x). Of course, here
the independent variable, x, via the rule given by f , will determine the value of y and
this is why we think of y as the dependent variable.
For now, we will only be interested in functions whose domain and co-domain are
certain sets of real numbers. In particular, they will either be R itself or certain subsets
of R called intervals. Typically, we think of R as the points on a line so that intervals
are described by line segments. Indeed, for a, b R, we will have finite intervals like
(a, b) = {x R | a < x < b}
and
[a, b] = {x R | a x b},
which only differ according to whether the end-points, i.e. the elements a and b, are in
the set. Of course, we can also have finite intervals where one end-point, but not the
other, is in the set and we denote these by
(a, b] = {x R | a < x b}
and
[a, b) = {x R | a x < b}.
There are also infinite intervals which will have one finite end-point, say a R, and we
denote these by
(, a] = {x R | x a}
and
[a, ) = {x R | a x},
and
(a, ) = {x R | a < x},
if this finite end-point is in the set, or by

(, a) = {x R | x < a}
if it isnt. Of course, as we can see by looking at the sets involved when writing these
infinite intervals, the symbols and are not end-points as they are not real
numbers, they are just a notational convenience.
Putting these ideas together, we find that another way of visualising a function
f : A B is its graph which is the set of all points (x, y) R2 such that y = f (x).
Indeed, as a function f : A B must give a unique output y B for each x A, its
graph could look like the one illustrated in Figure 2.1(a) but not like the one in
Figure 2.1(b).
10
2.1. Introduction: What is a function?
y
c
x
y = f (x)
a x
(a)
(b)
Figure 2.1: In (a) we have the graph of a function f : [0, a] [b, c] as each input, x [0, a],
gives a unique output y [b, c]. In (b), we do not have the graph of a function from [0, a]
to [b, c] as each input, x [0, a], gives two outputs y [b, c].
2.1.1
Some elementary functions and their graphs
We now revise some elementary functions that will be useful in this course and look at
their graphs.
Power functions
A power function is a function f : R R given by
f (x) = xn ,
where n N. Depending on the value of n, the graphs of these functions look very much
like the ones illustrated in Figure 2.2. In addition to this, we also include the power
function f (x) = x0 = 1 as the function whose graph is a horizontal straight line that
goes through the point (0, 1).
y
y = xn
y=x
O
y=
O
(a) n = 1
(b) n is even
xn
(c) n 3 is odd
Figure 2.2: (a) When n = 1, the graph of the function f (x) = xn is just the straight line
y = x. (b) The graph of the function f (x) = xn when n is even. (c) The graph of the
function f (x) = xn when n 3 is odd. Of course, in (b) and (c) we are only looking at
the shape of the graph for different values of n without any regard to the scales on the
axes.
In particular, if we let x mean that x is positive and getting arbitrarily large (i.e.
we are considering what happens as x takes values far to the right on the x-axis) and
11
2. Functions
x means that x is negative but getting arbitrarily large in magnitude (i.e. we are
considering what happens as x takes values far to the left on the x-axis), we see that:
If n is even, xn as x and as x .
If n is odd, xn as x whereas xn as x .
This insight will be important in Section 4.4 when we consider how to sketch the graphs
of more complicated functions.
Exponential functions
An exponential function with base a is a function f : R (0, ) given by
f (x) = ax ,
where a 6= 1 is a positive real number. Depending on the value of a, the graphs of these
functions look very much like the ones illustrated in Figure 2.3.
y
y = ax
y = ax
(a) 0 < a < 1
(b) a > 1
Figure 2.3: (a) The graph of the function f (x) = ax when 0 < a < 1. (b) The graph of the
function f (x) = ax when a > 1. Of course, in both of these graphs we are only looking
at the shape of the graph for different values of a without any regard to the scales on the
axes.
Indeed, looking at these graphs we see that
If 0 < a < 1, ax 0 as x and ax as x .
If a > 1, ax as x and ax 0 as x .
And, as a0 = 1 for any positive a 6= 1, the graphs of these functions always go through
the point (0, 1).
Trigonometric functions
The two elementary trigonometric functions that we will be using are the sine and
cosine functions but, unlike what you may have seen before, we will always be using
them for angles that are given in radians instead of degrees. As you may know, we can
easily convert between these two units by using the formula
angle in radians =
12
2
angle in degrees,
360
so, in particular, 360 is 2 radians, 180 is radians and 90 is /2 radians. Then,

measuring in radians, we can define the sine and cosine functions for 0 /2 by
using the right-angled triangle in Figure 2.4 to get
opposite
hypotenuse
and
s
nu
e
t
po
hy
adjacent
cos =
adjacent
.
hypotenuse
opposite
sin =
Figure 2.4: Defining the sine and cosine functions, sin and cos , for 0 /2.
In particular, by considering the two special triangles in Figure 2.5, we can see that the
values of these functions for some common angles (in radians) are
sin
cos
6
1
2
3
2
4
1
2
1
3
3
2
1
2
Activity 2.1 Recall that we also have the tangent function which, for 0 /2,
can be defined by using the right-angled triangle in Figure 2.4 to get
tan =
opposite
.
adjacent
Use the triangles in Figure 2.5 to find the values of tan when is /6, /4 and /3
radians. Incidentally, what are these three angles in degrees?
/4
4 1
/6
(a)
(b)
3 1
Figure 2.5: Finding sin and cos when (a) = /4 radians and (b) when = /6 or
= /3 radians.
At this point, well stop saying that an angle is in radians as, unless explicitly stated
otherwise, this will always be the case.
13
2. Functions
If we want to extend the definition of the sine and cosine functions to 0 2, we

think of a unit circle and a triangle with an hypotenuse of 1 as illustrated in
Figure 2.6(a) which, for 0 /2 gives us a point (x, y) with
x = cos
and
y = sin ,
which can be found as before. But, if we now have /2 2, we get the situation
illustrated in Figure 2.6(b), where we can find the magnitude of x and y using our
original triangle and their sign by considering where the point lies in the (x, y)-plane.
For instance, in Figure 2.6(b), the angle could be 5/4 and so the angle in the triangle
y
y
(x, y)
1
O
(x, y)
(a)
(b)
Figure 2.6: Finding sin and cos when 0 2 by considering a unit circle.
= 4 as the angle subtended by a straight line in this case

would be /4 (i.e. 5
4
the x-axis is ). This gives x and y a magnitude of 1/ 2 and their signs would be
negative as x, y < 0 so we see that
sin
5
1
=
4
2
and
cos
5
1
= ,
4
2
using the unit circle method.

Activity 2.2 Use the unit circle method to find sin and cos if = 2/3.
Activity 2.3 Use the unit circle method to find the values of sin and cos when
= 0 and = /2.
Hence deduce the values of these functions when = , = 3/2 and = 2.
If we want to extend the definition of the sine and cosine functions to all R, we can
see from the unit triangle method that both of these functions are periodic with a
period of 2, i.e.
sin( + 2) = sin
and
cos( + 2) = cos ,
and their graphs are illustrated in Figure 2.7. In particular, we observe that
cos = sin( + 2 ), i.e. the graph of the cosine function is what we get when we shift the
sine function to the left by /2.
14
Figure 2.7: The graphs of the sine and cosine functions, sin (solid line) and cos (dashed
line), for 4.
2.1.2
Combinations of functions
The elementary functions we have seen can be combined in various ways to make more
complicated functions. Generally, this is straightforward and works in the way you
would expect, but sometimes there are slight complications and so we revise these
different types of combination here.
Linear combinations of functions
If we have two functions with the same domain and co-domain, say f : A B and
g : A B, we can define a new function which is a linear combination of these two
functions. For instance, if k and l are constants, we would have the new function
kf + lg : A B defined by
(kf + lg)(x) = kf (x) + lg(x),
for all x A. In particular, this gives us polynomials, i.e. functions pn : R R which
are a linear combination of power functions of the form
pn (x) = an xn + an1 xn1 + + a1 x + a0 ,
where the ai for 0 i n are real constants. Indeed, if an 6= 0, we say that this is a
polynomial of degree n.
Of course, you have seen polynomials before as, in Chapter 1 of 173 Algebra, you saw
how to solve polynomial equations of the form pn (x) = 0 where n = 1 (a linear
equation), n = 2 (a quadratic equation) and n = 3 (a cubic equation). The information
we get from solving these equations is useful when we come to draw the graphs of
polynomial functions as the next example shows.
15
2. Functions
Example 2.1 Draw the graphs of the functions f : R R and g : R R given by

f (x) = 5 and g(x) = x + 2 on the same axes. At what point(s) do these graphs
intersect?
When we draw graphs, we will often do this by doing a sketch. Indeed, for a sketch
of the simple functions given here, it suffices to indicate their shape (they are both
straight lines) and where they are relative to the x and y-axes (by indicating where
they intersect these axes). So, as we saw in Section 2.1.1, we should expect the graph
of g(x) to be a horizontal line that goes through the point (0, 5) as g(x) = 5 for all
x R whereas for f (x), we would expect a straight line that has an
x-intercept that occurs when f (x) = 0, i.e. when x = 2, and a
y-intercept that occurs when x = 0, i.e. when f (0) = 2.
This information allows us to obtain the sketch illustrated in Figure 2.8.
To find the point(s) at which these two graphs intersect, we are looking for the
value(s) of x that make f (x) = g(x), i.e. where 5 = x + 2. This gives x = 3 and we
know that the values of the functions here must satisfy f (3) = g(3) = 5 which gives
(3, 5) as the required point of intersection.1
y
5
y = g(x)
y = f (x)
2
2
Figure 2.8: The graphs of the functions f (x) = 5 and g(x) = x + 2. Notice that these
graphs intersect at the point (3, 5) which we found in Example 2.1.

We will see how to draw the graphs of polynomial functions where n = 2 in
Section 2.2.1 and we will develop a more general method for dealing with the case
where n 3 in Section 4.4.
Products and quotients of functions
If we have two functions with the same domain but possibly different co-domains, say
f : A B and g : A C, we can define a new function which is the product of these
two functions. For instance, here we would have the new function f g : A D where
D is the possibly new co-domain, defined by
(f g)(x) = f (x)g(x),
for all x A. Of course, we have seen how this works the other way in Chapter 1 of 173
Algebra as the process of factorisation involves writing a polynomial of degree n as the
1
Of course, thinking about the graphs of these functions as the points, (x, y), satisfying the equations
y = f (x) and y = g(x), all we have done here is solve the equations y = 5 and y = x + 2 simultaneously.
16
product of two polynomials, one of degree m and another of degree p, with n = m + p.

However, the quotient of these two functions is slightly more tricky to deal with as the
function f /g defined by
f (x)
(f /g)(x) =
,
g(x)
only makes sense for those x A where g(x) 6= 0 as, of course, we can never divide by
zero. As such, when finding the quotient of two functions, we get a function
f /g : A0 B where A0 is a new domain given by
A0 = {x A | g(x) 6= 0}.
The points at which a quotient are undefined may have interesting consequences for its
graph since they can give rise to vertical asymptotes. But, this neednt be the case as
the next example shows.
Example 2.2 Discuss the behaviour of the functions given by
x+1
f (x) =
x1
x2 + x 2
and g(x) =
,
x1
at the point x = 1.
For f (x), the polynomials in the numerator and denominator of the quotient are
defined for all x R, but f itself is not defined at x = 1 because that would entail
division by zero. As such, f must be a function from {x R | x 6= 1} to R. Indeed, if
we are considering values of x close to one, i.e. x ' 1, we could say that
f (x) '
1+1
2
=
,
x1
x1
and so we see that:

If we let x go to one from values of x that are larger than one (here we say x
goes to 1 from above and write x 1+ ), we see that x 1 is positive and
getting very small, which means that f (x) itself is positive and getting very
large. That is, f (x) is getting arbitrarily large as x goes to 1 from above and we
write this as f (x) as x 1+ .
If we let x go to one from values of x that are smaller than one (here we say x
goes to 1 from below and write x 1 ), we see that x 1 is negative and
getting very small, which means that f (x) itself is negative and getting very
large in magnitude. That is, f (x) is negative but getting arbitrarily large in
magnitude as x goes to 1 from below and we write this as f (x) as
x 1 .
As such, we see that f (x) has a vertical asymptote at the point x = 1 where it is
undefined. The graph of this function is illustrated in Figure 2.9(a) so that you can
see this asymptote and you will understand why its graph looks like this away from
the asymptote after you have covered the material in Section 2.2.4.
For g(x), the polynomials in the numerator and the denominator of the quotient are
defined for all x R, but g itself is not defined at x = 1 because, again, that would
17
2. Functions
entail division by zero. As such, g must also be a function from {x R | x 6= 1} to R.

However, in this case, we notice that x = 1 makes both the numerator and the
denominator equal to zero and so, in particular, x = 1 must be a root of the
numerator. This means that, if we factorise the numerator, we find that
g(x) =
(x + 2)(x 1)
x2 + x 2
=
,
x1
x1
and so, as long as x 6= 1, we have

g(x) = x + 2.
As such, where it is defined (i.e. for x 6= 1) the graph of g is a straight line like the
one sketched in Figure 2.9(b) although, of course, we must exclude the point (1, 3)
from this line as g(x) is not defined there. In particular, note that in this case the
function does not have a vertical asymptote at x = 1 even though it is undefined
there.
We will look at asymptotes in more detail when we see them again in Section 2.2.4 and
Section 4.4.
y
3
2
y
y = f (x)
O
(a)
y = g(x)
(b)
Figure 2.9: The graphs of the functions f (x) and g(x) from Example 2.2. In (a), the
vertical asymptote at x = 1 is indicated by a dashed line. In (b), the point where the
function is undefined is indicated by .
We can also form quotients using trigonometric functions and, in particular, we can use
the triangle in Figure 2.4 to see that
tan =
opposite/hypotenuse
sin
opposite
=
=
,
adjacent
adjacent/hypotenuse
cos
that is, we can think of the tangent function as the quotient

tan =
sin
,
cos
(2.1)
which will be defined for R as long as cos 6= 0, i.e. as long as 6= (2n + 1) 2 for
n Z. At the points where it is undefined this function has vertical asymptotes and its
graph is sketched in Figure 2.10.
18
Figure 2.10: The graph of the tangent function, tan for 4. Note the vertical
asymptotes when = (2n + 1) 2 for n Z.
We can also find the reciprocals of our three trigonometric functions and these are
defined as follows.
The secant function, sec =
1
which is defined when 6= (2n + 1) 2 for n Z.
cos
The cosecant function, cosec =
1
which is defined when 6= n for n Z.
sin
The cotangent function, cot =
1
which is defined when 6= n for n Z.
tan
These functions will be especially useful in Section 2.1.4.

Activity 2.4 Show that we also have cot =
cos
as long as 6= n for n Z.
sin
Compositions of functions
If we have two functions, say f : A B and g : B C, then we can define the
composition g f : A C to be the function
(g f )(x) = g(f (x)),
and here we say that we are applying g after f . That is, thinking of this in terms of
black boxes we have
x A f f (x) B g g(f (x)) C,
i.e. we take an x A and apply f to get the output f (x) B which we then use as the
input for g yielding the final output g(f (x)) C which is the value of (g f )(x).
19
2. Functions
Example 2.3 Let f : R R and g : R R be the functions f (x) = x2 and

g(x) = 2x 1. What are the functions g f and f g?
Here, as the functions both go from R to R, we can find both of these compositions.
In particular,
g f is the function where
(g f )(x) = g(f (x)) = g(x2 ) = 2x2 1,
where (g f ) : R R.
f g is the function where
(f g)(x) = f (g(x)) = f (2x 1) = (2x 1)2 ,
and (f g) : R R.
Indeed, observe that as (2x 1)2 = 4x2 4x + 1, these are certainly not the same
function.
Activity 2.5 Let f : R R and g : R R be the functions f (x) = x2 + 1 and
g(x) = 2x . What are the functions g f and f g?
In particular, we will also need to be able to identify compositions the other way when
we cover the chain rule in Section 3.2.2. For instance, it should be clear that the
function (x2 + 5)3 is the composition of the function x3 after the function x2 + 5.
Activity 2.6 Explain why the function (x2 + 5)3 is the composition of the function
x3 after the function x2 + 5.
2.1.3
Inverse functions
If A and B are sets and we have a function f : A B, we know that this means that
for every x A there is a unique y B such that y = f (x). Now, if we can define
another function g : B A, i.e. for every y B there is a unique x A such that
y = f (x) if and only if x = g(y),
then we call the function, g, the inverse of f and denote it by f 1 . In terms of black
boxes, this means that we have
x A f f (x) B,
for f and, if it exists, we have
y B f 1 f 1 (y) A,
for f 1 , or more usefully,
f (x) B f 1 x A.
20
In particular, this means that if the inverse, f 1 , of f exists, we see that the
composition f 1 after f gives us
x A f f (x) B f 1 x A,
and so (f 1 f )(x) = f 1 (f (x)) = x whereas the composition f after f 1 gives us
y B f 1 f 1 (y) A f y B,
and so (f f 1 )(y) = f (f 1 (y)) = y. That is, the inverse of a function (if it exists)
undoes what the function does and vice versa.
The question, then, is how can we tell whether an inverse function exists? And, if it
does exist, how can we find it? Well, given the function f : A B, the inverse will exist
if we are able to take y = f (x) and solve it to obtain a unique solution, x, in terms of y
for every y B. And, if we can do this, these solutions will tell us what the inverse
function is, i.e. they will allow us to identify the function, f 1 (y), by comparison with
x = f 1 (y). To make this clear, lets look at an example.
Example 2.4 Consider the function f : R R given by f (x) = x + 2. Explain why
this function has an inverse and find it.
Using the graph or common sense, we see that the function f (x) = x + 2 has an
inverse, since every y R where y = f (x) gives rise to a unique x R given by
x = y 2. As such, we can conclude that the inverse of this function exists and we
have x = f 1 (y) = y 2. Of course, we can now write this inverse as f 1 (x) = x 2
if we want it in terms of x.
Indeed, notice that, if we have the function f (x) and its inverse function f 1 (x), the
graph of f 1 is the reflection of the graph of f about the line y = x. This happens
because any point (x, y) on the curve y = f (x) becomes, under a reflection about the
line y = x, a point (y, x) on the curve x = f (y) which is the same as saying that
y = f 1 (x)!
Activity 2.7 Verify that the curve y = f 1 (x) is the reflection about the line y = x
of the curve y = f (x) using the function we saw in Example 2.4.
Of course, not every function has an inverse as the next example shows.
Example 2.5 Consider the function f : R R given by f (x) = x2 . Explain why
this function does not have an inverse.
If we take any y R where y = f (x) this gives us the equation y = x2 and, if we are
considering x R, this gives rise to a problem as far as the inverse of f is concerned
because:
If y < 0, we get no solution for x as we know that x R means that y = x2 0.
If y > 0, we get two solutions for x as we know that we can get x = y R.
That is, we can find no inverse in this case since we cannot guarantee a unique
solution for x R from the equation y = x2 for all y R.
21
2. Functions
Of course, we can usually get around such problems if we are prepared to restrict the
domain and the co-domain of the function. But, in that case, we would be finding the
appropriate local inverses as opposed to its inverse (which, remember, doesnt exist!).
Activity 2.8 By considering the domains (, 0] and [0, ) and suitably
restricting the co-domain of the function in Example 2.5, find its local inverses.
Lets now look at the inverses of the elementary functions we considered in Section 2.1.1.
Power functions: root functions
If we have the power function f (x) = xn where x N and f : [0, ) [0, ) we can
see that the inverse is given by
x = f 1 (y) = y 1/n ,
and this is called a root function. Thus, we have
x = y 1/n
if and only if y = xn ,
provided that x, y 0. In particular, if n = 2, this is the square root function, i.e. we
have y 1/2 = y.
Activity 2.9 Draw the graph of the power function f : [0, ) [0, ) where
f (x) = x2 and its inverse.
This also works for f (x) = xn where f : R R if n is odd. But, if n is even, the
function f (x) = xn where f : R R does not have an inverse as we saw, for n = 2, in
Example 2.5.
Activity 2.10 Explain why we can find an inverse of the function f : R R where
f (x) = xn if n is odd. Why doesnt this work if n is even?
Exponential functions: logarithmic functions
If we have the exponential function f (x) = ax where f : R (0, ) and a 6= 1 is a
positive real number, the inverse is the function f 1 : (0, ) R given by
x = f 1 (y) = loga y,
which is the logarithm to base a. Thus, we have
x = loga y
provided that y > 0.
22
if and only if y = ax ,
Activity 2.11 Draw the graph of the exponential function f : R (0, ) where
f (x) = 2x and its inverse, f 1 (x) = log2 x where f 1 : (0, ) R.
In particular, we see from this that as
(f f 1 )(x) = f (f 1 (x)) = x we have aloga x = x,
and as
(f 1 f )(x) = f 1 (f (x)) = x we have
loga ax = x.
These results will be useful in Section 2.1.4 when we consider the laws of of logarithms.
Trigonometric functions: inverse trigonometric functions
If we want to discuss the inverses of the trigonometric functions sine and cosine, it is
first necessary to restrict their domain due to their oscillatory nature. To do this, we
consider a certain interval of values of , called the principal range, so that each value of
the function corresponds to a unique value of . Indeed, for the:
sine function, we take the principal range to be the interval [ 2 , 2 ] so that the
function sin : [ 2 , 2 ] [1, 1] where y = sin has an inverse. This inverse is
denoted by sin1 (or arcsin) where sin1 : [1, 1] [ 2 , 2 ]. Thus, we have
y = sin
provided that 2
if and only if = sin1 y,
and 1 y 1.
cosine function, we take the principal range to be the interval [0, ] so that the
function cos : [0, ] [1, 1] where y = cos has an inverse. This inverse is
denoted by cos1 (or arccos) where cos1 : [1, 1] [0, ]. Thus, we have
y = cos
if and only if = cos1 y,
provided that 0 and 1 y 1.
It will also be convenient for us to consider the inverse of the tangent function where, as
well as the oscillations, we need to take care to avoid the asymptotes that occur when
this function is undefined. As such, for the
tangent function, we take the principal range to be the interval ( 2 , 2 ) so that the
function tan : ( 2 , 2 ) R where y = tan has an inverse. This inverse is denoted
by tan1 (or arctan) where tan1 : R ( 2 , 2 ). Thus, we have
y = tan
if and only if = tan1 y,
provided that 2 < < 2 .
In particular, observe that sin1 , cos1 and tan1 are the inverses of the functions sin,
cos and tan respectively and not their reciprocals which we denoted by cosec, sec and
cot respectively in Section 2.1.2!
23
2. Functions
Activity 2.12 Find the acute angles 1 , 2 and 3 where 1 = sin1 12 , 2 = cos1
and t3 = tan1 1.
1
2
Also find cosec 1 , sec 2 and cot 3 .
2.1.4
Identities
An expression such as
(x + 1)2 = x2 + 2x + 1,
which is true for all x is called an identity and, as you know, these are useful when we
need to simplify expressions. In particular, in Chapter 1 of 173 Algebra, you saw that
the power laws dictate that
am an = am+n ,
am
= amn
an
and
(am )n = amn ,
and these are identities that work for any values of a, m and n for which both sides are
defined. Indeed, these laws allow us to simplify expressions that may result from
appropriate products, quotients and compositions of power functions or exponential
functions.
Activity 2.13 If f (x) = x3 , g(x) = x4 and h(x) = 2x , find the functions (f g)(x),
(f /g)(x) and (g h)(x) simplifying your answers as far as possible.
We now look at some other identities that will be useful in this course.
The laws of logarithms
For any positive real number a 6= 1, the laws of logarithms state that

x
loga x + loga y = loga (xy), loga x loga y = loga
and y loga x = loga (xy ),
y
provided that all of the terms involved are defined. As you may know, these laws are
easily derived from the power laws we saw above and the fact that
aloga x = x,
which we saw earlier in Section 2.1.3.
Activity 2.14 Derive the laws of logarithms from the power laws.
It is also useful to note that if a, b 6= 1 are positive real numbers, then we have the
change of base formula which states that
loga x =
logb x
,
logb a
and this allows us to write logarithms to base a in terms of logarithms to base b.
24
Activity 2.15 Derive the change of base formula for logarithms.
Trigonometric identities
There are also identities that allow us to simplify various expressions involving the
trigonometric functions. For instance, using the triangle in Figure 2.4, Pythagoras
theorem allows us to see that
2
2

opposite
adjacent
2
2
sin + cos =
+
hypotenuse
hypotenuse
2
2
opposite + adjacent
=
hypotenuse2
hypotenuse2
=
hypotenuse2
= 1,
and so, for acute angles,2 we have shown that
sin2 + cos2 = 1.
(2.2)
In particular, for natural numbers n 2, note that we commonly abbreviate things like
(sin )n by writing them as sinn . Further, dividing both sides of this expression by
sin2 we get
1 + cot2 = cosec2 ,
(2.3)
and this works as long as 6= n for n Z whereas dividing both sides of this
expression by cos2 we get
tan2 + 1 = sec2 ,
(2.4)
and this works as long as 6= (2n + 1) 2 for n Z. We call these three identities the
Pythagorean identities as they are simple consequences of Pythagorass theorem.
Activity 2.16 Use (2.2) to derive the Pythagorean identities (2.3) and (2.4).
Another useful pair of trigonometric identities are the compound-angle formulae given
by
sin( + ) = sin cos + cos sin and
cos( + ) = cos cos sin sin ,
which work for all , R.

Activity 2.17 Observe from the graphs of the sine and cosine functions in
Figure 2.7 that sine is an odd function, i.e. sin() = sin , and cosine is an even
function, i.e. cos() = cos . Use these facts and the compound-angle formulae to
show that we also have
sin( ) = sin cos cos sin and
cos( ) = cos cos + sin sin ,
for , R.
2
Of course, if we consider how we extend the definitions of the sine and cosine functions to all R,
it should be clear that this identity is actually true for all R.
25
2. Functions
Usually, we summarise these four compound-angle formulae by writing them as
sin( ) = sin cos cos sin and
cos( ) = cos cos sin sin , (2.5)
for , R. Indeed, they are especially useful since, setting = , we can use them to
obtain the double-angle formulae
sin(2) = 2 sin cos
and
cos(2) = cos2 sin2 ,
(2.6)
which work for all R. These will be especially useful in Chapter 5.

Activity 2.18 Use the compound-angle formulae to derive the double-angle
formulae given above.
Use the Pythagorean identity sin2 + cos2 = 1 to show that we also have
cos(2) = 1 2 sin2
and
cos(2) = 2 cos2 1,
for all R.
2.1.5
Applications of functions
In economics and related subjects, functions can be used to represent how one quantity
depends on another. For instance, as the profit that a company makes, , would depend
on the quantity of goods sold, q, it makes sense to suppose that there is some function
of q, say f , that tells us the corresponding profit, . In this case, we would use an
equation of the form = f (q) to express this dependency and we would have found a
profit function. Moreover, if f is invertible, we could find its inverse function, f 1 , and
we would use this to find the value of q that corresponds to a given value of . In which
case, the dependency would now be given by an equation of the form q = f 1 (). We
will look at profit functions properly in Section 4.5.3, but for now, we consider another
application of functions in economics, namely how they can be used to represent
information about supply and demand in a market.
Supply and demand functions
In any given market, there is a good which is supplied by the producers (and demanded
by the consumers) and the general idea is that, for both supply (and demand), if
producers are charging (or consumers are buying) at a price of p per-unit, then the level
of supply (or demand) for that good, q, will depend on p. Indeed, since each value of p
will lead the producers to supply (and the consumers to demand) exactly one quantity
q, it makes sense to think of the quantity, q, supplied (or demanded) as a function of
the price, p. This leads us to a description of the market in terms of two kinds of
function, namely:
If the quantity supplied, q, can be written in terms of p then we can identify the
supply function, q S , from the fact that we have q = q S (p). This tells us the
quantity, q, that the producers will supply if the prevailing market price is p.
26
If the quantity demanded, q, can be written in terms of p then we can identify the
demand function, q D , from the fact that we have q = q D (p). This tells us the
quantity, q, that the consumers will demand if the prevailing market price is p.
In particular, note that, although we have q as a function of p in both of these cases we
follow the practice common in economics and use the vertical axis for p and the
horizontal axis for q when drawing the graphs of these functions. As such, any point on
the graph of these functions is of the form (q, p) where q = q S (p) for supply and
q = q D (p) for demand. Also, these functions and their graphs only make economic sense
when p 0 and the quantities they yield, q, are also non-negative.3
Once we have these functions, we are often interested in the the equilibrium point for
the market as this is the point where the supply and demand functions are equal. In
theory, this is the point, (q , p ), where the market stabilises since, at this point, the
per-unit price, p , is such that the levels of supply and demand are equal, i.e. we have
q S (p ) = q D (p ).
As such, we can find the equilibrium price, p , by solving the resulting equation and the
corresponding equilibrium quantity, q , can then be found by, say, using the demand
function as q = q D (p ). Lets look at a simple example.
Example 2.6 The supply and demand functions for a good are
q S (p) = p + 1
and
q D (p) = 3 p,
respectively. Sketch the graphs of these functions and find the equilibrium point.
Here the supply and demand functions are straight lines which can easily be
sketched using the method outlined in Example 2.1 and the results of doing this are
illustrated in Figure 2.11. To find the equilibrium price, p , we have
q S (p ) = q D (p )
p + 1 = 3 p
2p = 4,
and so, p = 2. Then using the demand function, say, we have

q = q D (p ) = 3 p ,
and so the equilibrium quantity is q = 3 2 = 1. Consequently, the equilibrium
point is (q , p ) = (2, 1) which, as indicated in Figure 2.11, is the point at which the
two straight lines intersect.
Usually, the supply and demand functions are invertible and so we can also find the
inverses of these functions. In particular, if they are invertible, we note that:
If the price, p, can be written in terms of q then we can identify the inverse supply
function, pS , from the fact that we have p = pS (q). This tells us the price, p, that
the producers will charge if the quantity being supplied is q.
3
Although, when drawing their graphs, it is often useful to consider all possible values of p and q
before restricting your attention to the economically meaningful ones where p, q 0!
27
2. Functions
p
3
S
1
D
O
1
Figure 2.11: A sketch of the graphs of the supply and demand functions in Example 2.6
indicating the equilibrium point for this market. (Note that this sketch only makes
economic sense when p 0.)
If the price, p, can be written in terms of q then we can identify the inverse
demand function, pD , from the fact that we have p = pD (q). This tells us the price,
p, that the consumers will pay if the quantity being demanded is q.
Activity 2.19 Decide whether the supply and demand functions in the example
above are invertible. If they are, find the inverse supply and demand functions.
The effects of taxation
Sometimes, in order to control a market, a government will impose an excise tax of T
per unit sold. We model such situations by assuming that the tax is paid to the
government by the supplier and so, if the price paid by the consumers in the presence of
this tax is p per unit, the suppliers effectively receive p T for each unit sold as they
must pay T of each p received to the government. As such, the supply and demand
functions in the presence of the tax, lets call them qTS (p) and qTD (p) respectively, will be
given by
qTS (p) = q S (p T )
and
qTD (p) = q D (p).
That is, the consumers still pay a price of p per unit and so the demand function is
unchanged, but the suppliers now only receive an amount p T per unit and so the
supply function is modified by the introduction of an excise tax. Of course, the
introduction of an excise tax will affect the equilibrium price and quantity for a market,
i.e. in the presence of such a tax, the new equilibrium point, lets call it (qT , pT ), will be
the point where
qTS (pT ) = qTD (pT )
or, equivalently,
q S (pT T ) = q D (pT ),
and, using the unchanged demand function qT = qTD (pT ) or, equivalently, qT = q D (pT ).
Lets look at how such a tax would affect the market we considered in Example 2.6.
28
Example 2.7 An excise tax of T per unit is imposed on the market in

Example 2.6. Find the new equilibrium point and, by sketching the graph of the new
supply function on your earlier sketch, comment on how the equilibrium point for
the market has changed. How much of the tax has been passed onto the consumers?
What is the maximum tax, Tm , that can be imposed if this market is to continue
functioning?
If an excise tax of T per unit is imposed, the demand function is still
qTD (p) = q D (p) = 3 p,
but the supply function becomes
qTS (p) = q S (p T ) = p T + 1,
as the suppliers now see an effective price of p T . This means that the equilibrium
price in the presence of the tax, pT , is given by
qTS (pT ) = qTD (pT )
pT T +1 = 3pT
2pT = 2+T
T
p = 1+ ,
2
and so the equilibrium quantity in the presence of tax, qT , is

T
T
D
qT = qT (pT ) = 3 1 +
=2 ,
2
2
if we use the demand function, qTD (p).4 Thus, the new equilibrium point is
(2 T /2, 1 + T /2). Sketching the graph of the new supply function, as in
Figure 2.12, we see that it is parallel to the old one and the p-intercept has increased
by T . Indeed, as the equilibrium price has increased from 1 to 1 + T /2 due to the
presence of the tax, half the tax has been passed on to the consumer. Of course, the
equilibrium quantity in the presence of the tax must be positive and so, for the
market to function, we require that
qT > 0
T
>0
2
T < 4,
i.e. the maximum tax, Tm , that can be imposed is given by Tm = 4.

Alternatively, the government may decide to impose a percentage of the price tax of
100r% (so, for instance, a tax of 5% of the price would correspond to r = 0.05) instead
of the per unit tax that we have considered so far. So, again assuming that the tax is
paid to the government by the supplier, if the price paid by the consumers in the
presence of this tax is p per unit, the suppliers effectively receive p rp for each unit
sold as they must pay rp of each p received to the government. As such, in the presence
of such a tax, the supply and demand functions in the presence of the tax, lets call
4
Alternatively, we could use the supply function

qTS (p) = q S (p T ) = p T + 1,
to find qT . However, we can not use q S (p) = p + 1 as this no longer holds in the presence of the tax!
29
2. Functions
(2 12 T, 1 + 21 T )
new S
S
1
T 1
D
1
1
Figure 2.12: Following on from the sketch in Figure 2.11, if an excise tax of T per unit is
imposed, the supply set changes as shown and the demand set stays the same. Observe
how the introduction of this tax affects the equilibrium point for this market. (Note that
this sketch only makes economic sense when p 0.)
them qrS (p) and qrD (p) respectively, will be given by

qrS (p) = q S (p rp)
and
qrD (p) = q D (p).
That is, once again, the consumers still pay a price of p per unit and so the demand
function is unchanged, but the suppliers now only receive an amount p rp per unit
and so the supply function is modified by the introduction of a percentage of the price
tax. Of course, the introduction of this tax will also affect the equilibrium price and
quantity for the market, i.e. in the presence of such a tax, the new equilibrium point,
lets call it (qr , pr ), will be the point where
qrS (pr ) = qrD (pr )
or, equivalently,
q S (pr rpr ) = q D (pr ),
and, using the unchanged demand function qr = qrD (pr ) or, equivalently, qr = q D (pr ).
See, for example, Exercise 2.3 at the end of this chapter.
2.2
Conic sections
So far, we have been dealing with functions that are explicitly defined in terms of an
independent variable but, sometimes, we may have an equation relating two variables,
say x and y, which implicitly defines y as one or more functions of x. As it will be useful
in various places, we now investigate some important instances of functions defined in
this way and their graphs, the so-called conic sections.5
2.2.1
Parabolae
A parabola is a curve whose equation has the form

y = ax2 + bx + c,
5
See, for example, Binmore and Davies (2002) Section 2.14 for a full discussion of the geometric
aspects of conic sections and where they come from. Although this is interesting, we will not be delving
into these overly geometric aspects of conic sections in this course.
30
2.2. Conic sections
where a 6= 0, b and c are constants. Indeed, if we complete the square, we can write this
in the form
y = a(x p)2 + q,
for some constants p and q. This curve will have a y-intercept which we can find by
setting x = 0 and it may have x-intercepts which, if they exist, we can find by setting
y = 0. It will also have a turning point with coordinates (p, q) which will be a minimum
if a > 0 and a maximum if a < 0. Once we have this information, the parabola should
be easy to draw as the next example shows.
Example 2.8 Sketch the parabolae whose equations are
(a) y = x2 4x + 3, and
(b) y = x2 + 2x + 3.
For (a), we are told that y = x2 4x + 3 and so we find that:
For the y-intercept: Setting x = 0 we get y = 3.
For the x-intercepts: Setting y = 0 we get
x2 4x + 3 = 0
(x 1)(x 3) = 0,
i.e. the x-intercepts are x = 1 and x = 3.

The turning point of the parabola can be found by writing the equation of the
parabola in completed square form and, doing this, we get
y = (x 2)2 1.
Here, a = 1 > 0 and so we get a minimum at the point (2, 1).
Putting this information together, we then get the sketch in Figure 2.13(a).
For (b), we are told that y = x2 + 2x + 3 and so we find that
For the y-intercept: Setting x = 0 we get y = 3.
For the x-intercept: Setting y = 0 we get
x2 + 2x + 3 = 0
x2 2x 3 = 0
(x + 1)(x 3) = 0,

The turning point of the parabola can be found by writing the equation of the
parabola in completed square form and, doing this, we get

y = x2 +2x+3 = x2 2x +3 = (x1)2 1 +3 = (x1)2 +1+3 = (x1)2 +4.
Here, a = 1 < 0 and so we get a maximum at the point (1, 4).
Putting this information together, we then get the sketch in Figure 2.13(b).
31
2. Functions
y
y = x2 4x + 3
4
3
y = x2 + 2x + 3
2
O
x
1
(a)
x
1
(b)
Figure 2.13: In (a) we have a sketch of the parabola from Example 2.8(a). In (b) we have
a sketch of the parabola from Example 2.8(b).

Activity 2.20 Given the equation of a parabola in completed square form, i.e.
y = a(x p)2 + q,
use the fact that (x p)2 0 for all x R to explain why the turning point of this
parabola will be a minimum if a > 0 and a maximum if a < 0.
2.2.2
Circles
A circle of radius, r, centred on the point (a, b) has an equation given by

(x a)2 + (y b)2 = r2 .
Of course, such a circle is easy to draw and its x and y-intercepts can be found by
seeing where y = 0 and where x = 0 respectively. Once we have this information, the
circle should be easy to draw as the next example shows.
Example 2.9 Find the radius and centre of the circle whose equation is given by
Sketch the circle.
x2 6x + y 2 8y = 0.
We are told that

x2 6x + y 2 8y = 0,
is the equation of a circle and so, completing the square in x and y, we find that

2
2
(x 3) 9 + (y 4) 16 = 0
=
(x 3)2 + (y 4)2 = 25,
32
2.2. Conic sections
and so, comparing this with (x a)2 + (y b)2 = r2 we see that we have a circle of
radius 5 centred on the point (3, 4). We also find that:
For the x-intercept: Setting y = 0 we get

(x 3)2 + 16 = 25
(x 3)2 = 9
x 3 = 3,
y 4 = 4,

For the y-intercept: Setting x = 0 we get
9 + (y 4)2 = 25
(y 4)2 = 16
i.e. the y-intercepts are y = 8 and y = 0.

Putting this information together, we then get the sketch in Figure 2.14(a).
y
8
y
3
5
4
O
2
3
(a)
(b)
Figure 2.14: In (a) we have a sketch of the circle from Example 2.9. In (b) we have a
sketch of the ellipse from Example 2.10.
2.2.3
Ellipses
An ellipse has an equation of the form

x2 y 2
+ 2 = 1.
a2
b
In particular, an ellipse of this form is effectively a circle centred on the origin that has
been squashed and it is easy to draw once we have found its x and y-intercepts by
seeing where y = 0 and where x = 0 respectively.
Example 2.10 Sketch the ellipse whose equation is given by
x2 y 2
+
=1
4
9
Given that the equation of the ellipse is

x2 y 2
+
= 1,
4
9
we see that the x-intercepts, which occur when y = 0, are given by
x2
=1
4
x2 = 4
x = 2,
33
2. Functions
whereas the y-intercepts, which occur when x = 0, are given by

y2
=1
9
y2 = 9
y = 3.
Putting this information together, and bearing in mind that this should look like a
circle centred on the origin that has been squashed, we then get the sketch in
Figure 2.14(b).
2.2.4
Hyperbolae
A hyperbola can have an equation of the form

x2 y 2
2 = 1.
a2
b
This curve will have x-intercepts which can be found by setting y = 0, but no
y-intercepts. It will also have slant (or oblique) asymptotes which can be found by
writing the equation as

1
1
y2
2
=b
,
x2
a2 x 2
so that, as x , we have 1/x2 0 and this leaves us with
b2
b
y2
=
=
y = x,
2
2
x
a
a
as the equations of the asymptotes. Once we have this information, the hyperbola
should be easy to draw as the next example shows.
Example 2.11 Sketch the hyperbola whose equation is given by
x2 y 2
= 1.
4
9
Given that the equation of the hyperbola is

x2 y 2
= 1,
4
9
we see that the x-intercepts, which occur when y = 0, are given by
x2
= 1 = x2 = 4 = x = 2,
4
whereas there are no y-intercepts since, setting x = 0, we get
y2
= 1 = y 2 = 9,
9
which has no real solutions. To find the asymptotes, we write the equation as

y2
1
1
=9
,
x2
4 x2
y2
9
3
= y = x,
=
2
x
4
2
as the equations of the asymptotes. Putting this information together, we then get
the sketch in Figure 2.15(a).
34
2.2. Conic sections
y
2
x1
y =1+
x2
4
y2
9
=1
x
O
1
1
=
3
(a)
(b)
Figure 2.15: In (a) we have a sketch of the hyperbola from Example 2.11. In (b) we have
a sketch of the rectangular hyperbola from Example 2.12.

Of course, similar remarks apply to a hyperbola which has an equation of the form
y 2 x2
2 = 1,
b2
a
and, in particular, this curve will have y-intercepts but no x-intercepts.
Activity 2.21 Sketch the hyperbola whose equation is given by
y 2 x2
= 1.
9
4
Lastly, we note that a rectangular hyperbola has an equation of the form

(x a)(y b) = c,
and this arises when the asymptotes turn out to be the horizontal line y = b and the
vertical line x = a as the next example illustrates.
Example 2.12 Sketch the rectangular hyperbola whose equation is given by
(x 1)(y 1) = 2.
Given that (x 1)(y 1) = 2, we can see that:
For the x-intercept: Setting y = 0 we get (x 1)(1) = 2 or x 1 = 2, i.e.the
x-intercept is given by x = 1.
For the y-intercept: Setting x = 0 we get (1)(y 1) = 2 or y 1 = 2, i.e. the
y-intercept is given by y = 1.
Then, by writing the equation as
y =1+
2
,
x1
we can find the asymptotes by noting that:
35
2. Functions
For the vertical asymptote: As x 1+ we have y and as x 1 we have

y .
For the horizontal asymptote: As x we have y 1 from above and as

x we have y 1 from below.
Putting this information together, we then get the sketch in Figure 2.15(b). In
particular, observe that here we have
y =1+
(x 1) + 2
x+1
2
=
=
,
x1
x1
x1
and so this gives us y = f (x) where f (x) is the first function in Example 2.2 which
was illustrated in Figure 2.9(a).
Learning outcomes
At the end of this chapter and having completed the relevant reading and activities, you
should be able to:
identify elementary functions and sketch their graphs;
find combinations of elementary functions and inverses (if they exist);
use identities to rewrite expressions involving powers, logarithms and trigonometric
functions;
solve problems from economics-based subjects that involve functions;
identify and sketch conic sections.
Solutions to activities
Solution to activity 2.1
Using the triangles in Figure 2.5 and the definition of the tangent function, it should be
clear that

tan = , tan = 1 and tan = 3.
6
4
3
3
Indeed, using the fact that
angle in radians =
2
angle in degrees,
360
we can see that an angle of /6, /4 or /3 radians corresponds to an angle of 30, 45 or

60 degrees respectively.
In this case, the unit circle method gives us the situation illustrated in Figure 2.16 and
so the angle in the triangle would be /3 (i.e. 2
= 3 as the angle subtended by a
3
36
2.2. Solutions to activities
straight line in this case the x-axis is ) giving x a magnitude of 1/2 and y a
magnitude of 3/2 whereas their signs would be negative for x (as x < 0) and positive
for y (as y > 0). Thus we see that
3
2
2
1
sin
=
and
cos
= ,
3
2
3
2
using the unit circle method.
(x, y)
y
1
2/3
Figure 2.16: For Activity 2.2, we find sin and cos when = 2/3 by considering a unit
circle.
Using the unit circle in Figure 2.17(a), it should be clear that
sin 0 = 0
and
cos 0 = 1,
whereas using the unit circle in Figure 2.17(b), it should be clear that
and
cos = 0.
sin = 1
2
2
Then, using similar reasoning, we should be able to deduce that
sin
cos
3
2
1
0
2
0
1
are the other values of the functions sin and cos that we seek.
From the definition of cot , we have
1
1
cos
=
=
,
sin
tan
sin
cos
as we know that tan = sin / cos . This function is defined as long as 6= n for n Z
since, at these values of , we have tan = 0 or, equivalently, sin = 0.
cot =

Given the functions f : R R and g : R R where f (x) = x2 + 1 and g(x) = 2x , we
see that
37
2. Functions
1
O
(a)
1
x
(b)
Figure 2.17: For Activity 2.3, we find sin and cos by considering a unit circle when (a)
= 0 and (b) = /2.
g f is the function where

(g f )(x) = g(f (x)) = g(x2 + 1) = 2x
2 +1
where (g f ) : R R.
f g is the function where
(f g)(x) = f (g(x)) = f (2x ) = (2x )2 + 1 = 22x + 1,
and (f g) : R R.
Indeed, observe that as 2x
2 +1
6= 22x + 1, these are certainly not the same function.

If we have f (x) = x3 and g(x) = x2 + 5, then the function (x2 + 5)3 can be written as
(x2 + 5)3 = f (x2 + 5) = f (g(x)) = (f g)(x),
i.e. it is the composition we get from applying f after g or, in terms of x, it is the
composition of the function x3 after the function x2 + 5.
By considering the graphs of the functions f (x) = x + 2 and f 1 (x) = x 2 as
illustrated in Figure 2.18, we see that the latter is indeed the reflection in the line y = x
of the former. Alternatively, we can see that if y = x + 2, a reflection in the line y = x
just means replacing all points (x, y) that satisfy this equation with points given by
(y, x) to get the new equation x = y + 2. But, of course, this gives y = x 2 which is
what we wanted.
When we considered the function f : R R in Example 2.5, there were two problems
that prevented us from finding an inverse. To counteract these so that we can find the
local inverses of this function, we note that:
If we take the co-domain to be the interval [0, ) so that we have y 0, then we
remove the problem that occurs because y = x2 has no solution for y < 0.
38
2
x
y
O
x
=
Figure 2.18: For Activity 2.7, we see that the graph of f 1 (x) = x 2 is the reflection of
the function f (x) = x + 2 about the line y = x.
If we take the two domains given by the intervals (, 0] and [0, ) so that we
have x 0 and x 0 respectively, then we remove the problem that occurs
because y = x2 has two solutions for x R.
Indeed, this means that if we consider the function
f : [0, ) [0, ) given by f (x) = x2 , then we have

y = f (x)
y = x2
x=
y,
1
as x 0 because x [0, ).
Thus, using x = f (y), the inverse of this function is
1
1
f (y) = y or f (x) = x if we want it in terms of x.
f : (, 0] [0, ) given by f (x) = x2 , then we have

y = f (x)
y = x2
x = y,
1
as x 0 because x (, 0]. Thus,
using x = f (y), the inverse of this function
1
1
is f (y) = y or f (x) = x if we want it in terms of x.
In particular,
this means that the local inversesof f : R [0, ) where f (x) = x2 are
f 1 (x) = x when x [0, ) and f 1 (x) = x when x (, 0].

We saw in Activity 2.8 that

the function f : [0, ) [0, ) where f (x) = x2 has an
inverse given by f 1 (x) = x. The graphs of these two

functions are illustrated in
Figure 2.19. In particular, observe that the curve y = x is the reflection about the line
y = x of the curve y = x2 and that all three of these curves intersect at the points (0, 0)
and (1, 1).
From the graphs of the function f (x) = xn where f : R R when n is odd, which we
saw in Figure 2.2(a) and (c), it should be clear that the the equation y = f (x) has a
unique solution, x, for all y R and so the inverse of this function exists. In particular,
we see that
y = xn
=
x = y 1/n = n y,
gives us this unique solution for any y R provided that n
is odd and so we have
1
1
n
n
f (y) = y as the inverse function or, indeed, f (x) = x if we want it in terms of x.
39
x
y
y=
x2
2. Functions
y=
Figure 2.19: For Activity 2.9, we see that the graph of f 1 (x) =
x is the reflection of
the function f (x) = x2 about the line y = x.

However, from the graph of the function f (x) = xn where f : R R when n is even,
which we saw in Figure 2.2(b), it should be clear that when
y < 0, the equation y = f (x) has no solution for x as we know that x R means
that y = xn 0 when n is even.
y > 0, the equation y = f (x) has two solutions for x as we know that we can get
x = n y R when n is even.
As such, we can not find a unique solution, x, for all y R and so the inverse of this
function can not exist.
x
log 2
=
y
1
O
y= x
2
We saw the graph of a function like f : R (0, ) where f (x) = 2x in Figure 2.3(b)
since we have a = 2 > 1 here. As such, we find that the graphs of the function
f : R (0, ) where f (x) = 2x and its inverse, f 1 (x) = log2 x where
f 1 : (0, ) R, are as illustrated in Figure 2.20. In particular, observe that the curve
y = log2 x is the reflection about the line y = x of the curve y = 2x .
Figure 2.20: For Activity 2.11, we see that the graph of f 1 (x) = log2 x is the reflection
of the function f (x) = 2x about the line y = x.
40

To find the acute angles 1 and 2 where 1 = sin1
of values in Section 2.1.1 to see that
1
= sin
2
6
sin 1 =
gives us
1
2
and 2 = cos1 12 , we use the table
1 = sin1
= ,
2
6
and
1
1
= cos
gives us
2 = cos1 = ,
2
3
2
3
1
whereas to find the acute angle 3 where t3 = tan 1, we use the table we found in
Activity 2.1 to see that
cos 2 =
tan 3 = 1 = tan
gives us
3 = tan1 1 =
.
4
We also have
cosec 1 =
1
1
1
1
1
1
=
= 2, sec 2 =
=
= 2 and cot t3 =
= = 1,
sin 1
1/2
cos 2
1/2
tan 3
1
using the definitions of the reciprocals of our three trigonometric functions, which we
saw in Section 2.1.2.
Given that f (x) = x3 , g(x) = x4 and h(x) = 2x , we use the definitions of the
combinations of functions we need from Section 2.1.2, to get
(f g)(x) = f (x)g(x) = (x3 )(x4 ) = x7 ,
f (x)
x3
1
(f /g)(x) =
= 4 = , and
g(x)
x
x
(g h)(x) = g(h(x)) = g(2x ) = (2x )4 = 24x ,
where we have used the power laws to simplify our answers. Indeed, observe that for the
last function, we can also write 24x = (24 )x = 16x .
To derive the laws of logarithms, we note that for the first one, we use the power laws
and the given fact to get
aloga x+loga y = aloga x aloga y = xy = aloga (xy) ,
which means that loga x + loga y = loga (xy), for the second one, we similarly get
aloga xloga y =
x
aloga x
=
= aloga (x/y) ,
aloga y
y
which means that loga x loga y = loga (x/y) and for the third one, we get
y
ay loga x = (aloga x )y = xy = aloga (x ) ,

which means that y loga x = loga (xy ).
41
2. Functions
We take logarithms to the base b on both sides of the given fact to see that

aloga x = x = logb aloga x = logb x = (loga x)(logb a) = logb x,
where we have used the third law of logarithms in the last step. Then, dividing through
on both sides by logb a (which is non-zero as a 6= 1), we get
loga x =
logb x
,
logb a
as required.
Starting with sin2 + cos2 = 1, we divide both sides by sin2 to get

2
2
sin2 + cos2
1
sin2 cos2
1
cos
1
=
=
+
=
= 1+
=
,
sin
sin
sin2
sin2
sin2 sin2
sin2
so that 1 + cot2 = cosec2 if we use the definition of cosec from Section 2.1.2 and the
result from Activity 2.4. Then, again starting with sin2 + cos2 = 1, we divide both
sides by cos2 to get

2

2
sin
sin2 + cos2
1
sin2 cos2
1
1
=
=
+
=
=
+1 =
,
cos2
cos2
cos2 cos2
cos2
cos
cos
so that tan2 + 1 = sec2 if we use the definition of sec and (2.1) from Section 2.1.2.
With the given facts, we can use the compound-angle formula for sin( + ) to see that
sin( ) = sin( + ()) = sin cos() + cos sin() = sin cos cos sin ,
and the compound-angle formula for cos( + ) to see that
cos( ) = cos( + ()) = cos cos() sin sin() = cos cos + sin sin ,
as required.
Using the compound-angle formula
sin( + ) = sin cos + cos sin ,
with = we get
sin( + ) = sin cos + cos sin
sin(2) = 2 sin cos ,
whereas using the compound-angle formula

cos( + ) = cos cos sin sin ,
42
with = we get
cos( + ) = cos cos sin sin
cos(2) = cos2 sin2 ,

2
as required. Indeed, since we also have the Pythagorean identity sin + cos = 1, we
can write this last double-angle formula as
cos(2) = (1 sin2 ) sin2 = 1 2 sin2 ,
in terms of sin2 , or as
cos(2) = cos2 (1 cos2 ) = 2 cos2 1,
in terms of cos2 , as required.
From the graph in Figure 2.11, we can see that the economically meaningful part of the
supply function is q S : [0, ) [1, ) where q S (p) = p + 1 and the economically
meaningful part of the demand function is q D : [0, 3] [0, 3] where q D (p) = 3 p.
Clearly, both of these functions are invertible as each q in the co-domain gives rise to a
unique p in the domain and we find that
q =p+1
p = pS (q) = q 1,
is the inverse supply function, whereas

q =3p
p = pD (q) = 3 q,
is the inverse demand function.

Given that
y = a(x p)2 + q,
we see that
If a > 0, then for any x R,
(x p)2 0
a(x p)2 0
a(x p)2 + q q,
i.e. for all x R, y q and so the smallest value of y occurs when y = q which, in
turn, means that we must have x = p. Thus, the turning point of the parabola is a
minimum and this occurs at the point (p, q).
If a < 0, then for any x R,
(x p)2 0
a(x p)2 0
a(x p)2 + q q,
i.e. for all x R, y q and so the largest value of y occurs when y = q which, in
turn, means that we must have x = p. Thus, the turning point of the parabola is a
maximum and this occurs at the point (p, q).
43
2. Functions
y 2 x2
= 1,
9
4
we see that there are no x-intercepts since, setting y = 0, we get
x2
=1
4
x2 = 4,
which has no real solutions, whereas we see that the y-intercepts, which occur when
x = 0, are given by
y2
= 1 = y 2 = 9 = y = 3.
9
To find the asymptotes, we write the equation as
y2
=9
x2
1
1
+ 2
4 x

9
y2
=
2
x
4
3
y = x,
2
as the equations of the asymptotes. Putting this information together, we then get the
sketch in Figure 2.21.
x2
4
=1
y2
9
Given that the equation of the hyperbola is
y
=
3
Figure 2.21: For Activity 2.21, a sketch of the hyperbola
44
y 2 x2
= 1.
9
4
2.2. Exercises
Exercises
Exercise 2.1
Sketch the graph of the function f : {x R | x 6= 1, 1} R given by
x4 1
.
x2 1
f (x) =
Exercise 2.2
Use the compound-angle formulae to show that
tan( ) =
tan tan
,
1 tan tan
and hence deduce an expression for tan(2).

Exercise 2.3
The supply and demand functions for a good are
q S (p) = p 4
and
q D (p) = 8 p,
respectively. Sketch the graphs of these functions and find the equilibrium point.
A percentage [of the price] tax of 100r% is imposed. Find the new equilibrium point
and, by sketching the graph of the new supply function on your earlier sketch, comment
on how the equilibrium point for the market has changed. How much of the tax has
been passed onto the consumers? What is the maximum tax, rm , that can be imposed if
this market is to continue functioning?
Exercise 2.4
When selling a quantity, q, a firm makes a profit given by
(q) = q 2 + 2q + 2,
and the largest quantity it can produce is 10. Sketch the graph of this profit function
and deduce the value of q that will yield the greatest profit for this firm.
Explain why the inverse profit function exists and find it.
Exercise 2.5
Sketch the circle and the rectangular hyperbola with equations
x2 + y 2 = 1
and
2xy = 1,
respectively. At what points do these two curves intersect?
45
2. Functions
Solutions to exercises
Solution to exercise 2.1

The function f : {x R | x 6= 1, 1} R given by
x4 1
f (x) = 2
,
x 1
is clearly undefined at x = 1 and x = 1 as these values of x would entail division by
zero. However, we notice that factorising the numerator and the denominator we get
f (x) =
(x2 1)(x2 + 1)
,
x2 1
and so, as long as x 6= 1, we have

f (x) = x2 + 1.
This means that, to sketch the graph of f (x), we start by sketching the graph of the
parabola
y = x2 + 1,
which has
a y-intercept when x = 0, i.e. when y = 1,
no x-intercepts as y = 0 gives x2 + 1 = 0 which has no real solutions, and
a turning point which is a minimum at the point (0, 1).
We then exclude the points (1, 2) and (1, 2) on the parabola at which f (x) itself is
undefined to get the sketch in Figure 2.22.
y
y = f (x)
2
1
1 O 1
Figure 2.22: For Exercise 2.1, a sketch of the graph of f (x). (Note that the points at
which f (x) is undefined are marked by a .)

Using (2.1), we have
tan( ) =
46
sin( )
,
cos( )
2.2. Solutions to exercises
and so, using the compound-angle formulae in (2.5), we get

sin
sin
sin cos cos sin

cos cos
tan( ) =
=
,
sin sin
cos cos sin cos
1
cos cos
if we divide the numerator and denominator by cos cos and cancel where
appropriate. Thus, using (2.1) again, we have
tan( ) =
tan tan
,
1 tan tan
as required. Indeed, observe that this only makes sense if , 6= (2n + 1) 2 for n Z as,
if this isnt true, we cant divide through by cos cos or, equivalently, one of tan or
tan wont exist.
To deduce a formula for tan(2), we set = in the formula for tan( + ) to get
tan(2) = tan( + ) =
2 tan
tan + tan
=
.
1 tan tan
1 tan2
Again, we observe that this only makes sense if 6= (2n + 1) 2 for n Z as, if this isnt
true, tan wont exist.
Here the supply and demand functions are straight lines which can be easily sketched
using the method outlined in Example 2.1 and the results of doing this are illustrated in
Figure 2.23(a). To find the equilibrium price, p , we have
q S (p ) = q D (p )
p 4 = 8 p
2p = 12,
and so, p = 6. Then, using the demand function, say, we have

q = q D (p ) = 8 p ,
and so the equilibrium quantity is q = 8 6 = 2. Consequently, the equilibrium point
is (q , p ) = (2, 6) which, as indicated in Figure 2.23(a), is the point at which the two
straight lines intersect.
If a percentage [of the price] tax of 100r% is imposed,6 the demand function is still
qrD (p) = q D (p) = 8 p,
but the supply function becomes
qrS (p) = q S (p rp) = p rp 4,
6
Here we will start by restricting our attention to the case where 0 r 1 as, prima facie, these are
the values that would appear to be economically sensible. Although, as we will soon see, the economically
meaningful values of r will turn out to be 0 r < 1/2!
47
2. Functions
as the suppliers now see an effective price of p rp. This means that the equilibrium
price in the presence of tax, pr , is given by
qrS (pr ) = qrD (pr ) = pr rpr + 4 = 8 pr

= pr (2 r) = 12
12
= p =
,
2r
and so the equilibrium quantity in the presence of tax, qr , is
qr = qrD (pr ) = 8
16 8r 12
4 8r
12
=
=
,
2r
2r
2r
if we use the demand function, qrD (p).7 Sketching the graph of the new supply function,
as in Figure 2.23(b), we see that by writing its equation as
p=
q
4
+
,
1r 1r
and noting that
1
4
1
and
4,
1r
1r
when considering 0 r 1, this means that it is steeper than the old one and that the
p-intercept, which is now
4
4(1 r) + 4r
4r
=
=4+
,
1r
1r
1r
has increased by 4r/(1 r). In this case, as the equilibrium price has increased from 6 to
12
6(2 r) + 6r
6r
=
=6+
,
2r
2r
2r
we see that the consumer pays 6r/(2 r) more. But, as the total tax to be paid by the
supplier is given by
12
12r
rpr = r
=
,
2r
2r
this means that only half of the tax has been passed on to the consumer in this case. Of
course, the equilibrium quantity in the presence of the tax must be positive and so, for
the market to function, we require that
qr > 0
4 8r
>0
2r
4 > 8r
1
r< ,
2
(bearing in mind that 2 r > 0 if 0 r 1), i.e. the maximum tax, rm , that can be
imposed is given by rm = 1/2.
7
Alternatively, we could use the supply function

qrS (p) = q S (p rp) = p rp 4,
to find qr . However, we can not use q S (p) = p 4 as this no longer holds in the presence of the tax!
48
p
8
new S
12 48r
, 2r )
( 2r
4
1r
6
4
D
4
(a)
(b)
Figure 2.23: For Exercise 2.3, a sketch of the graphs of the supply and demand functions
indicating the equilibrium point of the market when (a) there is no tax and (b) a
percentage of the price tax of 100r% is imposed. (Note that these sketches only make
economic sense when q 0.)
The firms profit function is
(q) = q 2 + 2q + 2,
and its domain is the interval [0, 10] as q 0 since it is a quantity and q 10 since the
largest quantity it can produce is 10. So, to sketch the graph of this profit function, we
start by sketching the parabola
y = q 2 + 2q + 2 = (q + 1)2 + 1,
in completed square form. This has
a y-intercept when q = 0, i.e. y = 2,
no q-intercepts as y = 0 gives (q + 1)2 + 1 = 0 which has no real solutions, and
a turning point which is a minimum at the point (1, 1).
We then restrict our attention to the relevant values of q, i.e. those that satisfy
0 q 10, to get a sketch of the graph of the profit function itself as illustrated in
Figure 2.24.
Looking at the graph of the profit function, we see that as it is a function
: [0, 10] [2, 122], its inverse exists since there is a unique q [0, 10] such that
y = (q) for all y [2, 122]. Indeed, solving this equation we find that, using the
completed square form above, we have
p
p
y = (q+1)2 +1 = (q+1)2 = y1 = q+1 = y 1 = q = 1 y 1,
which gives us two values of q for each value of y > 1. However, as we must be getting
values of q [0, 10] from our inverse function, we take the + sign here (i.e. we discard
49
2. Functions
the sign) so that we can get the solutions where q 1 (instead of getting the
solutions where q 1 which we dont want). That is, we have found that
p
q = 1 (y) = 1 + y 1,
is the sought after inverse function in terms of y.
y
122
y = (q)
2
1
1 O
10 q
Figure 2.24: For Exercise 2.4, a sketch of the graph of the profit function, (q). (Note
that dashed parts of the curve are on the parabola but are not part of the graph of the
profit function.)
To sketch the circle and the rectangular hyperbola, we note that:
The circle with equation
x2 + y 2 = 1,
is centred on the origin and has a radius of 1. Indeed, setting x = 0, we find that its
y-intercepts are y = 1 and, setting y = 0, we find that its x-intercepts are x = 1.
The rectangular hyperbola with equation
2xy = 1
y=
1
,
2x
has the x and y-axes, i.e. the lines y = 0 and x = 0, respectively, as its asymptotes
since
For the vertical asymptote: As x 0+ we have y and as x 0 we have
y .
For the horizontal asymptote: As x we have y 0 from above and as
x we have y 0 from below.
We also note that it has no x-intercepts (as no value of x makes y = 0) and no

y-intercepts (as no value of y makes x = 0).
and these curves are illustrated in Figure 2.25.
To find the points of intersection of these two curves we have to solve the equations
x2 + y 2 = 1
and
2xy = 1,
simultaneously. This can easily be done in two different ways which we give here for
completeness.
50
Method I: The equation 2xy = 1 tells us that, say, y = 1/(2x) and substituting
this into the other equation we get
x2 + y 2 = 1
1
=1
4x2
x2 +
4x4 4x2 + 1 = 0,
which is a quadratic equation in x2 . By factorising this, say, we find that

(2x2 1)2 = 0
x2 =
1
2
1
x = .
2
Then, substituting these back into the equation y = 1/(2x) we get
1
2
1
y=
=
= ,
2x
2
2
as the corresponding values of y.
Method II: We note that, using our equations, we have
(x y)2 = x2 2xy + y 2 = (x2 + y 2 ) (2xy) = 1 1 = 0,
and so any solutions we seek must satisfy (x y)2 = 0 or, equivalently, x = y. If we
substitute this into one of our equations, say 2xy = 1, we get
2xy = 1
2y 2 = 1
y2 =
1
2
1
y = .
2
Then, using the equation x = y again, we get x = 1/ 2 as the corresponding

values of x.
So, whichever
method
we find
that the points of intersection of these two
you choose,
curves are (1/ 2, 1/ 2) and (1/ 2, 1/ 2), both of which are indicated in
Figure 2.25.
y
y=
1
2x
1
1
O
1
x2 + y 2 = 1
Figure 2.25: For Exercise 2.5, a sketch of the circle x2 + y 2 = 1 and

the rectangular
hyperbola
2xy
= 1. The indicated points of intersection are (1/ 2, 1/ 2) and

(1/ 2, 1/ 2).
51
2. Functions
52
Chapter 3
Differentiation
3
Essential reading
(For full publication details, see Chapter 1.)
Binmore and Davies (2002) Sections 2.72.13.
Anthony and Biggs (1996) Chapter 6 and parts of Chapter 7.
Further reading
Simon and Blume (1994) Sections 2.32.7 and 3.6, Chapter 4 and Section 5.5.
Adams and Essex (2010) Sections 2.12.7, parts of Sections 3.1 and 3.3, parts of
Sections 4.9 and 4.10.
Aims and objectives
The objectives of this chapter are as follows.
To introduce the idea of a derivative and see how it can be found using various
techniques.
To use derivatives to find tangent lines and approximate functions using various
techniques.
To see how derivatives can be used in economics-based subjects.
Specific learning outcomes can be found near the end of this chapter.
3.1
Introduction: What is differentiation?
Having revised the idea of a function in the previous chapter, we now turn to
differentiation, the process by which we find the derivative of a function. Given a
function, f , its derivative at the point a, which we denote by f 0 (a), is given by the
formula
f (a + h) f (a)
f 0 (a) = lim
,
h0
h
provided that the limit exists. Indeed, when the limit exists, i.e. when we can find a
value for f 0 (a), we say that the function is differentiable at a. Observe that here, we
have introduced the notation
lim g(h),
h0
53
3. Differentiation
to denote the value1 of the function g(h) as h 0 (provided, of course, that there is
such a value) and we call this value the limit of g(h) as h 0 whereas if there is no
such value, we say that this limit does not exist.2 To see how this works in practice, we
can consider a simple example.
Example 3.1 Use the definition to find the derivative of the function f (x) = x2 at
the point x = 3.
We need to find f 0 (3) and, using the formula above with a = 3, we start by looking at
f (3 + h) f (3)
(3 + h)2 32
=
,
h
h
which, looking at the numerator, is easily simplified to give
f (3 + h) f (3)
(9 + 6h + h2 ) 9
6h + h2
=
=
= 6 + h.
h
h
h
This in turn means that

f (3 + h) f (3)
f (3) = lim
= lim 6 + h = 6,
h0
h0
h
0
because, in the limit as h 0, we see that 6 + h goes to 6. Consequently, we can see

that the derivative of f (x) at the point x = 3, i.e. f 0 (3), is 6. Indeed, we can say that
the function f (x) = x2 is differentiable at x = 3.
Activity 3.1 Use the definition to find the derivative of the function f (x) = x2 at
the point x = 1.
More generally, instead of finding the derivative of f at individual points, we can find
its derivative at a general point, x, by finding f 0 (x). Of course, according to our
formula, this would involve finding
f (x + h) f (x)
,
h0
h
and, provided the limit exists, this will give us another function of x. This can then be
used to find the derivative, say f 0 (a), at an individual point, a, by setting x = a in our
result. Lets see how this works.
f 0 (x) = lim
Example 3.2 Use the definition to find the derivative of the function f (x) = x2 at
the general point x and use this to verify that f 0 (3) = 6 as we found in Example 3.1.
We need to find f 0 (x) and, using the formula above, we start by looking at
f (x + h) f (x)
(x + h)2 x2
=
,
h
h
1
That is, a finite real number.

In 176 Further Calculus, you will do limits properly, but this simple explanation of what is going on
should suffice for our purposes here. In particular, we briefly introduced the notation in Example 2.2
and, with the examples below, you should be able to see what is happening for now. Also, in Section
3.3.4, we will see some examples where a limit fails to exist and we will explain what that means there.
2
54
3.2. How to find derivatives

f (x + h) f (x)
(x2 + 2xh + h2 ) x2
2xh + h2
=
=
= 2x + h.
h
h
h

f (x + h) f (x)
f (x) = lim
= lim 2x + h = 2x,
h0
h0
h
because, in the limit as h 0, we see that 2x + h goes to 2x. Consequently, we can

see that the derivative of f (x) at the general point x, i.e. f 0 (x), is 2x which is also a
function of x as we should expect.3
Having found this, we can substitute x = 3 into our result to see that
f 0 (3) = 2(3) = 6 as we found in Example 3.1.
Activity 3.2 Use the result in Example 3.2 to verify your answer to Activity 3.1.
At what point is the derivative of f (x) = x2 equal to (a) 16 and (b) 4?
We have seen that a function, f (x), has a derivative, f 0 (x), which is also a function of x.
The process by which we go from a function to its derivative is called differentiation.
That is, when we have a function, f (x), we differentiate it with respect to x and we
sometimes denote this operation by
d
f (x) which is read as differentiate f (x) with respect to x,
dx
and the outcome of this operation will be the sought after derivative which we can write
as
df
or f 0 (x).
dx
If we then want to evaluate the derivative of f at the point a, we write

df
or f 0 (a),
dx x=a
depending on which notation we are using.
We will see what derivatives tell us about functions in Section 3.3 and, in particular, we
will see that some functions do not have derivatives at certain points as the limit in the
definition may not exist. But, before we do that, we turn our attention to how we can
find the derivative of a function when we dont want to explicitly use the definition.
3.2
How to find derivatives
The previous section told us how to find derivatives from first principles, but now we
want to explore a more convenient way of finding them. The key idea is that we
3
Indeed, as this limit exists for all x R, we can say that the function f (x) = x2 is differentiable for
all x R.
55
3. Differentiation
introduce standard derivatives which tell us how to differentiate the basic functions that
we saw in the previous chapter. Once we know how to differentiate these, the rules of
differentiation will allow us to differentiate combinations of these functions.
3.2.1
Standard derivatives
In Example 3.2, we used the definition of the derivative to show that the function
f (x) = x2 has a derivative given by f 0 (x) = 2x. We now state some results that will
allow us to differentiate other elementary functions.
Power and root functions
If n Z, we can use the definition of the derivative to show that
f (x) = xn
f 0 (x) = nxn1 .
For instance, we note that:

If n = 0, we have
f (x) = x0 = 1
f 0 (x) = 0x1 = 0,
which tells us the derivative of 1.

If n = 1, we have
f (x) = x1 = x
f 0 (x) = 1x0 = 1,
which tells us the derivative of x.

If n = 2, we have
f (x) = x2
f 0 (x) = 2x1 = 2x,
in agreement with what we saw in Example 3.2.

Indeed, we will also use this rule when n Q, and so we also have things like
f (x) = x 2 =
which tells us the derivative of
1 1
1
f 0 (x) = x 2 = ,
2
2 x
x.
Exponential and logarithmic functions

If we are using e and ln, the derivatives are very simple, i.e.
f (x) = ex
f 0 (x) = ex ,
as ex is the special function which is equal to its derivative. We also have

f (x) = ln x
56
f 0 (x) =
1
,
x
which, as we will see in Activity 3.12, follows from the fact that the function ln x is the
inverse of ex .
If we have another base, a, the derivatives are not so simple. We shall see in Activity 3.9
that
f (x) = ax
=
f 0 (x) = ax ln a,
and, using the change of base formula for logarithms, we will see that
f (x) = loga x
1
f (x) =
,
x ln a
f 0 (x) = cos x,
in Section 3.2.2.
Sine and cosine functions
For the sine function we find that
f (x) = sin x
and for the cosine function we have
f (x) = cos x
f 0 (x) = sin x.
Although, we could have used the fact that the sine and cosine functions are
interdefinable, i.e.

cos x = sin x +
and sin x = cos x +
,
2
2
to derive the latter from the former once we have the chain rule (see Exercise 3.2).
Indeed, using these standard derivatives, we can then derive the derivatives of the other
trigonometric functions using their definitions in terms of sine and cosine together with
the rules of differentiation in Section 3.2.2 see, for example, Activity 3.6(c).
3.2.2
The rules of differentiation
In Section 2.1.2, we saw that there are several standard ways of making new functions
from old ones. Here, we see how we can use the standard derivatives, i.e. the derivatives
of our basic functions, and rules of differentiation to differentiate new functions that are
created from these basic ones in these standard ways. We start with the most
straightforward of these which allows us to differentiate linear combinations of functions.
The linear combination rule
If k and l are constants, this allows us to differentiate the linear combination,
kf (x) + lg(x), of two functions f (x) and g(x). It states that

df
dg
d
kf (x) + lg(x) = k
+l ,
dx
dx
dx
or, using our shorthand, (kf + lg)0 (x) = kf 0 (x) + lg 0 (x). Indeed, this gives us three more
basic rules straightaway, i.e. the
57
3. Differentiation
constant multiple rule: If k is a constant and f (x) is a function, then

d
df
kf (x) = k ,
dx
dx
or, using our shorthand, (kf )0 (x) = kf 0 (x).
sum rule: If f (x) and g(x) are functions, then

d
df
dg
f (x) + g(x) =
+
,
dx
dx dx
or, using our shorthand, (f + g)0 (x) = f 0 (x) + g 0 (x).
difference rule: If f (x) and g(x) are functions, then

d
df
dg
f (x) g(x) =
,
dx
dx dx
or, using our shorthand, (f g)0 (x) = f 0 (x) g 0 (x).
Activity 3.3 Derive the constant multiple, sum and difference rules from the linear
combination rule.
Example 3.3 Using these rules we see that:

3
3
1
if f (x) = 3x 2 , then f 0 (x) = 3 12 x 2 = 32 x 2 by the constant multiple
rule;
1
if f (x) = x2 + x 2 , then f 0 (x) = 2x + 12 x 2 by the sum rule;

if f (x) = cos x sin x, then f 0 (x) = sin x cos x by the difference rule;
if f (x) = 3 ln x 4 ex , then f 0 (x) =
3
4 ex by the linear combination rule.
x
So, in the case of simple combinations of functions such as these, we see that the
derivative of the linear combination is given by the linear combination of the derivatives.
Activity 3.4 Use the rules above to differentiate the following functions with
respect to x.
(a) 3 cos x,
(b) ex + cos x,
(c) 3 sin x 3 ln x.
Indeed, we can see that using the change of base formula for logarithms from
Section 2.1.4, we have
ln x
loga x =
,
ln a
58
and so, using the constant multiple rule, we get

d
d ln x
1 d
1
1
1
loga x =
=
ln x =
=
,
dx
dx ln a
ln a dx
ln a x
x ln a
as mentioned in Section 3.2.1. We now look at the other rules of differentiation, i.e. the
ones that will allow us to differentiate products, quotients and compositions of functions.
The product rule
This allows us to differentiate the product of two functions f (x) and g(x). It states that

d
df
dg
f (x)g(x) =
g(x) + f (x) ,
dx
dx
dx
or, using our shorthand, [f (x)g(x)]0 = f 0 (x)g(x) + f (x)g 0 (x)]. Lets have a look at some
examples of how it works.
Example 3.4 Differentiate the function h(x) = x ex with respect to x.
This is the product of the two functions
f (x) = x
and
g(x) = ex ,
f 0 (x) = 1
and
g 0 (x) = ex .
and these give us

As such, the product rule tells us that
h0 (x) = (1)(ex ) + (x)(ex ) = (1 + x) ex ,
is the derivative of the function h(x) = x ex with respect to x.
Example 3.5 Differentiate the function h(x) = x ln x with respect to x.
f (x) = x
and
g(x) = ln x,
and these give us

f 0 (x) = 1
and
g 0 (x) =
1
.
x

1
h (x) = (1)(ln x) + (x)
= ln x + 1.
x
0
is the derivative of the function h(x) = x ln x with respect to x.
59
3. Differentiation
Example 3.6 Differentiate the function h(x) = ex ln x with respect to x.

f (x) = ex
and
g(x) = ln x,
and these give us

f 0 (x) = ex
and
g 0 (x) =
1
.
x

1
1
x
h (x) = (e )(ln x) + (e )
= e ln x +
,
x
x
0
is the derivative of the function h(x) = ex ln x with respect to x.

Activity 3.5 Use the product rule to differentiate the following functions with
respect to x.
(a) x sin x, (b) ex cos x, (c) sin x cos x.
What can you deduce about the derivative of sin(2x) from your answer to (c)?
The quotient rule
This allows us to differentiate the quotient of two functions f (x) and g(x). It states that
df
dg

g(x) f (x)
d f (x)
dx ,
= dx
dx g(x)
[g(x)]2
or, using our shorthand,

f (x)
g(x)
0
f 0 (x)g(x) f (x)g 0 (x)

.
[g(x)]2
Of course, as we saw in Section 2.1.2, this all assumes that the quotient of the two
functions is defined for the values of x that we are working with, i.e. it only works for
values of x in the domain where g(x) 6= 0. Lets have a look at some examples of how it
works.
Example 3.7 For x 6= 0, differentiate the function h(x) =

This is the quotient of the two functions
f (x) = ex
and
g(x) = x,
f 0 (x) = ex
and
g 0 (x) = 1.
and these give us
60
ex
with respect to x.
x
As such, for x 6= 0, the quotient rule tells us that

h0 (x) =
(ex )(x) (ex )(1)

x1 x
=
e ,
2
x
x2
is the derivative of the function h(x) with respect to x.

Example 3.8 For x 6= 1, differentiate the function h(x) =
x3
with respect to x.4
ln x

f (x) = x3
and
g(x) = ln x,
and these give us

f 0 (x) = 3x2
and
g 0 (x) =
1
.
x
As such, for x 6= 1, the quotient rule tells us that

1
2
3
(3x
)(ln
x)
(x
)
x2 (3 ln x 1)
x
h0 (x) =
=
,
[ln x]2
[ln x]2
is the derivative of h(x) with respect to x.
Example 3.9 Differentiate the function h(x) =
ln x
with respect to x.5
ex

f (x) = ln x
and
g(x) = ex ,
and these give us
1
and
g 0 (x) = ex .
x
As such, the quotient rule tells us that
x
1
(e ) (ln x)(ex )
(1 x ln x) ex
1 x ln x
0
x
h (x) =
=
=
,
[ex ]2
x e2x
x ex
f 0 (x) =

Activity 3.6 Use the quotient rule to differentiate the following functions with
respect to x and find the values of x for which the derivatives exist.
(a)
sin x
,
x
(b)
ex
,
cos x
(c)
sin x
.
cos x
What can you deduce about the derivative of tan x from your answer to (c)?
4
5
Here, h(x) is only defined for x 6= 1 since we have ln x = 0 if x = 1.

Observe that as ex > 0 for all x R, we dont have to worry about dividing by zero here.
61
3. Differentiation
The chain rule

This allows us to differentiate the composition of two functions f (x) and g(x). It states
that
d
df dg
[f (g(x))] =
,
dx
dg dx
or, using our shorthand, [f (g(x))]0 = f 0 (g)g 0 (x). Lets have a look at some examples of
how it works.
Example 3.10 Differentiate the function h(x) = (2x + 1)3 with respect to x.
The function h(x) = (2x + 1)3 is the composition of the functions
f (g) = g 3
and
g(x) = 2x + 1.
As such we have
f 0 (g) = 3g 2
and
g 0 (x) = 2,
and so the chain rule tells us that

h0 (x) = (3g 2 )(2) = 6g 2 = 6(2x + 1)2 ,
Activity 3.7 Verify that this is correct by multiplying out the brackets and
differentiating your new expression for h(x) with respect to x.
Example 3.11 Differentiate the function h(x) = 2x + 1 with respect to x.
The function h(x) = 2x + 1 is the composition of the functions

f (g) =
g = g2
and
g(x) = 2x + 1.
As such we have
1 1
f 0 (g) = g 2
and
g 0 (x) = 2,
2

1
1 1
1
0
,
h (x) =
g 2 (2) = g 2 =
2
2x + 1
is the derivative of h(x) with respect to x.6
Example 3.12 Differentiate the function h(x) = ex

6
3 +2
with respect to x.
In particular, observe that here the original function is only defined if x 1/2 whereas the derivative
is only defined if x > 1/2 (as, in the derivative, x = 1/2 would entail division by zero).
62
The function h(x) = ex
3 +2
is the composition of the functions
f (g) = eg
and
g(x) = x3 + 2.
As such we have
f 0 (g) = eg
and
g 0 (x) = 3x2 ,

3 +2
h0 (x) = (eg )(3x2 ) = 3x2 ex

Activity 3.8 Use the chain rule to differentiate the following functions with respect
to x.
(a) sin(2x), (b) ln(cos x), (c) ln(ex ).
Why should your answer to (c) be obvious?
The chain rule can also be used to derive some useful results.
Activity 3.9 (A useful result)
Using the fact that
ax = ex ln a ,
which we saw in Section 2.1.3, show that
dax
= ax ln a.
dx
This was mentioned in Section 3.2.1, but there is no need to remember it as you
should be able to derive this result if it is needed.
Activity 3.10 (Deriving the quotient rule)
Derive the quotient rule by writing the quotient
f (x)
g(x)
as the product f (x)[g(x)]1 ,
and using the product and chain rules to differentiate it with respect to x.
Activity 3.11 (Derivatives of inverse functions)
If the function, f , has an inverse, f 1 , then we can let y = f (x) so that x = f 1 (y).
Use the chain rule to show that

d 1
d
f (y) = 1
f (x) .
dy
dx
63
3. Differentiation
Activity 3.12 We know that if y = ex , then x = ln y. Use the result in

Activity 3.11 and the fact that (ex )0 = ex to show that the derivative of ln y with
respect to y is 1/y.
Using the rules together
Sometimes it will be necessary to apply several of the above rules of differentiation in

order to find a derivative. This is easily done as long as care is taken to recognise what
you are differentiating at each step. Here are two examples that should make the
procedure clear.
Example 3.13 Differentiate the function l(x) = (x3 + 1) ln(x2 + 4) with respect to
x.
f (x) = x3 + 1
and
g(x) = ln(x2 + 4),
and clearly, f 0 (x) = 3x2 . But to differentiate g(x) we need to use the chain rule
because it is a composition. In this case, we have
g(h) = ln h
which gives us
and
h(x) = x2 + 4,
1
and
h0 (x) = 2x,
h

1
2x
2x
0
g (x) =
(2x) =
= 2
,
h
h
x +4
g 0 (h) =
so that
by the chain rule. Now, putting all of this into the product rule gives us

2x
2x(x3 + 1)
0
2
2
3
2
2
=
3x
ln(x
+
4)
+
,
l (x) = (3x ) ln(x + 4) + (x + 1)
x2 + 4
x2 + 4
as the derivative of l(x) with respect to x.
2 +x
Example 3.14 Differentiate the function l(x) = ex
ln(x3 + 1) with respect to x.

f (x) = ex
2 +x
and
g(x) = ln(x3 + 1),
and to differentiate f (x) we need to use the chain rule because it is a composition.
In this case, we have
f (h) = eh
and
h(x) = x2 + x,
f 0 (h) = eh
and
h0 (x) = 2x + 1,
which gives us
64
so that
f 0 (x) = (eh )(2x + 1) = (2x + 1) eh = (2x + 1) ex
2 +x
by the chain rule. Then, to differentiate g(x), we need to use the chain rule again
because it is also a composition. In this case, we have
g(h) = ln h
which gives us
and
h(x) = x3 + 1,
1
and
h0 (x) = 3x2 ,
h

1
3x2
3x2
0
g (x) =
(3x2 ) =
= 3
,
h
h
x +1
g 0 (h) =
so that
by the chain rule. Now, putting all of this into the product rule gives us

x2 +x
3x2
0
x2 +x
3
l (x) = (2x + 1) e
ln(x + 1) + e
x3 + 1

3x2
2
= (2x + 1) ln(x3 + 1) + 3
ex +x ,
x +1
as the derivative of l(x) with respect to x.
Of course, once you can reliably apply the rules, there is no need to show all of the
intermediate working.
Activity 3.13 Use the rules of differentiation to differentiate the following
functions with respect to x.
2
(a) ex ln(sin x),
(b)
sin(cos x)
,
esin x
(c) sin2 (3x) + cos2 (3x).
Why should your answer to (c) be obvious?
3.2.3
Higher-order derivatives
As we have seen above, when we differentiate a function, f (x), we find that its
derivative, f 0 (x), is also a function of x. In this context, we call f 0 (x) the first-order
derivative of f (x) and we can differentiate it again to get the second-order derivative,
i.e. we find

d2 f
d df
and we denote this by
or f 00 (x).
2
dx dx
dx
Of course, the second-order derivative will also be a function of x and so we can
differentiate it again to get the third-order derivative, i.e. we find

d d2 f
d3 f
and
we
denote
this
by
or f 000 (x).
dx dx2
dx3
We can, of course, do this again and again but the shorthand notation we use can
become a bit unwieldy once we pass the third-order derivative. As such, for n 4, we
65
3. Differentiation
often write the nth-order derivative, i.e.

dn f
dxn
as f (n) (x),
when we use our shorthand.
Example 3.15 Find the first four derivatives of f (x) = sin x. What is the
relationship between these derivatives of sin x?
We have f (x) = sin x, and so the first-order derivative of f is given by
f 0 (x) =
d
sin x = cos x.
dx
The second-order derivative of f is then

d df
d
00
f (x) =
=
cos x = sin x.
dx dx
dx
The third-order derivative of f is then

d d2 f
d
000
f (x) =
( sin x) = cos x.
=
2
dx dx
dx
And, finally, the fourth-order derivative of f is

d d3 f
d
(4)
f (x) =
( cos x) = sin x.
=
3
dx dx
dx
So, in particular, we see that f 00 (x) = f (x), f 000 (x) = f 0 (x) and f (4) (x) = f (x).
Activity 3.14 Using the pattern inherent in Example 3.15, what is f (n) (x) for
n 1?
Activity 3.15 Find the first four derivatives of f (x) = x ex . Hence deduce an
expression for f (n) (x) for n 1.
3.3
Using derivatives
Derivatives can be very useful in mathematics and economics, but before we see how,
we need to understand what derivatives represent.
3.3.1
The meaning of the derivative
If we draw the graph of a function, f , we get the curve with equation y = f (x). At any
point on this curve, say the point (a, f (a)), we can draw a chord (or secant line) that
connects the given point to another point on the curve. For instance, in Figure 3.1, the
66
3.3. Using derivatives

y
y = f (x)
f (b)
3
C
f (a)
O
Figure 3.1: The line segment C is the chord joining the points (a, f (a)) and (b, f (b)) on
the curve y = f (x). This is extended using the dotted lines at both ends so that we can
see what line the chord is a line segment of.
line segment C is the chord joining the points (a, f (a)) and (b, f (b)) on the curve
y = f (x). In particular, we see that the gradient of this chord, lets call it mC , can be
found using the formula
f (b) f (a)
,
mC =
ba
which you should know.
To relate this to the derivative, we take some number, h 6= 0, and let b = a + h so that
we now have a chord, C, which is joining the points (a, f (a)) and (a + h, f (a + h)). The
gradient of this chord is then given by
mC (h) =
f (a + h) f (a)
f (a + h) f (a)
=
,
(a + h) a
h
and, for h 6= 0, this is a function of h since the value of mC will depend on the value of
h that we choose. In particular, recalling what we saw in Section 3.1, we can see that
f 0 (a) = lim mC (h),
h0
and we want to see what this is telling us about the function, f .

We now consider the chords that join the point (a, f (a)) to the points
(a + h1 , f (a + h1 )), (a + h2 , f (a + h2 )) and (a + h3 , f (a + h3 )) where h3 > h2 > h1 > 0.
These are the line segments on the lines C1 , C2 and C3 which can be seen in Figure 3.2.
That is, we have three points that are getting successively closer to the point (a, f (a))
so that we can see the effect of letting h 0. In particular, as we let h get smaller, we
see that the gradients of these particular chords are decreasing. The question is, do the
67
3. Differentiation
y
y = f (x)
C3
C2
f (a + h3 )
C1
T
f (a + h2 )
f (a + h1 )
f (a)
O
a + h1
a + h2
a + h3
Figure 3.2: C1 , C2 and C3 are three chords of the curve y = f (x) originating from the
point (a, f (a)). Observe that as the other end of a chord approaches this point, the chords
pivot about it and their gradients get closer to the gradient of the line, T .
gradients of these chords tend to some finite limit as h 0? That is, does the limit in
our expression for f 0 (a) above exist?
Hopefully, in Figure 3.2, you can see that as h gets smaller (i.e. as we consider C3 , then
C2 and then C1 ), the lines are pivoting through the point (a, f (a)) and their gradients
are getting closer to the gradient of the line T . Indeed, in the limit as h 0, the lines
we get from extending an arbitrary chord joining the points (a, f (a)) and
(a + h, f (a + h)) should become the line T . In particular, this means that the limit of
mC (h) as h 0 exists because it should be equal to the gradient of T . This means that
the line T , called the tangent to f at the point (a, f (a))
goes through the point (a, f (a)), and
its gradient is the limit, as h 0, of mC (h), i.e. f 0 (a).
For this reason, we define the gradient of a curve y = f (x) at the point (a, f (a)) to be
the gradient of its tangent line at that point and this, as we have seen, is simply the
value of f 0 (a).
3.3.2
Tangent lines and linear approximations
Now that we know how the tangent lines to a curve are related to derivatives, we can
use derivatives to find the equation of the tangent line to a curve at a given point. This,
in turn, will introduce us to a useful way of performing approximations.
68
Finding tangent lines

Given that f 0 (a) is the gradient of the curve y = f (x) at the point x = a, we can use
this to find the equation of the tangent line at this point. In particular, the formula for
the gradient of a straight line, i.e.
f 0 (a) =
y f (a)
,
xa
(3.1)
gives us the equation of the tangent line as it goes through the point (a, f (a)) and its
gradient is given by f 0 (a). Lets look at a quick example.
Example 3.16 Find the equation of the tangent line to the function f (x) = x2
when x = 3.
When x = 3, the point on the curve y = x2 is (3, 9) and we know that f 0 (3) = 6 as
f 0 (x) = 2x. Consequently, using (3.1), the equation of the tangent line is given by
6=
y9
x3
y 9 = 6x 18
y = 6x 9.
In particular, when written in this form, we see that the gradient of the line is
indeed 6 and the point (3, 9) does indeed lie on it as 6(3) 9 = 9.
Activity 3.16 Find the equation of the tangent line to the function f (x) = ex when
x = 1.
Linear approximations
One use of tangent lines is that they provide us with a simple way of approximating the
value of a function. For instance, if we have the tangent line to the function f (x) at the
point x = a, the equation of its tangent line, i.e.
f 0 (a) =
y f (a)
,
xa
can be rearranged to give us

y = f (a) + (x a)f 0 (a).
Now, if we pick a value of x that is close to a, say x , the value of y when x = x , will be
y = f (a) + (x a)f 0 (a),
and this will be close to the value of f (x ) as illustrated in Figure 3.3. Of course, if we
pick a value of x which is closer to a, the value of y will be closer to the value of f (x )
and we will have a better approximation to the value of f (x) at this point.
As we are approximating the function f by a straight line, we call this a linear
approximation to f around a. In particular, we have
f (x) ' f (a) + (x a)f 0 (a),
69
3. Differentiation
y
y = f (x)
3
T
error
f (x )
f (a)
O
Figure 3.3: When x is close to a we can use the tangent line at a to find y which gives
us an approximate value for f (x ). There will, of course, be an error involved in this

approximation but this can be made smaller if we use values of x which are closer to a.
if x is close to a. In Section 3.4, we will discuss Taylor series and these will allow us to
find better approximations to f around a, but before we do that, lets consider a useful
application of linear approximations.
Example 3.17 Without using a calculator, find an approximate value of 3 e0.1 .
Given that f (x) = 3 ex , we have
f 0 (x) = 3 ex
f 0 (0) = 3 e0 = 3,
and so, using our linear approximation, we get

f (0.1) ' f (0) + (0.1 0)f 0 (0) = 3 e0 +(0.1)(3) = 3 0.3 = 2.7,
i.e. an approximate value of 3 e0.1 is 2.7. We note in passing that, using a
calculator, we can see that the exact value of 3 e0.1 is 2.71 to 2dp and so this is a
pretty good approximation.
Using linear approximations to find changes
As well as allowing us to find approximations to f around a, linear approximations can
give us useful information about how the value of f is changing as we move from a to,
say, a + h. Using our linear approximation, we see that
f (a + h) ' f (a) + hf 0 (a)
70
f 0 (a) '
f (a + h) f (a)
,
h
and so, if we denote the change in f by f and the change in x by x = h, we see that
f
' f 0 (a)
or
f ' f 0 (a)x.
x
That is, we can find the approximate value of the change in f if we change x from a to
a + h. Of course, the smaller x = h is, the better our approximation. This is
illustrated in Figure 3.4.
y
y = f (x)
approx f
f (a)
O
exact f
error
f (a + h)
a+h
x = h
Figure 3.4: Finding an approximate value for f using a linear approximation to f .
Obviously, the smaller the value of the change x = h, the better the approximation for
f will be.
Example 3.18 Without using a calculator, find the approximate change in 3 ex if
x is increased from zero to 0.1. Hence deduce the approximate value of 3 e0.1 .
Given that f (x) = 3 ex , we have
f 0 (x) = 3 ex
f 0 (0) = 3 e0 = 3,
so, if x = 0.1 as x increases from 0 to 0.1, we find that

f ' f 0 (0)x = (3)(0.1) = 0.3,
i.e. the change in f is approximately 0.3. Observe that the minus sign is telling us
that when x increases from 0 to 0.1, f (x) is decreasing by approximately 0.3.
This means that using
f (0.1) ' f (0) + f = 3 e0 +(0.3) = 3 0.3 = 2.7,
we see that the approximate value of 3 e0.1 is 2.7 as we would expect from the linear
approximation in Example 3.17.
71
3. Differentiation
Further, as the derivative of a function gives us information about how f (x) is changing
due to changes in x, we often refer to f 0 (a) as the rate of change of f (x) with respect to
x when x = a.
3.3.3
Applications of derivatives
Derivatives are useful in economics and we now introduce two ways in which they can
arise in that subject. The first is their use when discussing marginal functions and the
second is when they are used in the context of elasticities. At this point, we will just
introduce these ideas and see how they might be useful, but they will also be used when
we consider some applications of the material contained in other chapters of this subject
guide.
Marginal functions
In economics, the term marginal denotes the rate of change of a quantity with respect
to a variable on which it depends. For instance, if a firm has a cost function, C(q), this
tells us the cost of producing q units of their product. The marginal cost of the firm,
which we denote by MC(q), would then be given by
MC(q) =
dC
.
dq
This is useful since, using what we saw above, we can see that the marginal cost is
telling us (approximately) about how changes in the level of production, q, will incur
changes in the costs, C. That is, if the level of production is increased by q, i.e. our
production increases from q to q + q, we find that
MC(q) =
dC
dq
MC(q) '
C
q
C ' MC(q)q,
where C = C(q + q) C(q) is, of course, the resulting increase in costs. In

particular, if q is so large that a change in production of one unit (i.e. q = 1) is small
compared to q, we see that
C = C(q + 1) C(q) ' MC(q).
That is, in these circumstances, the marginal cost tells us (approximately) the extra
cost incurred if the firm wishes to produce one more unit of their good given that they
are already producing q units.
Example 3.19 A firm has a cost function given by
C(q) = 1000 + 5q + q 2 ,
in dollars. Find the marginal cost function for this firm and use it to determine the
approximate cost of producing one more unit if the original level of production is 100
units.
The marginal cost function, MC(q), is given by
MC(q) = C0 (q) = 5 + 2q,
72
and so using the fact that the change in cost, C, is related to the change in
production, q, by
C ' C0 (q)q,
we see that an increase in production of one unit, i.e. q = 1, gives rise to an
increase in costs given by
C ' C0 (100)(1) = (5 + 2(100))(1) = 205.

That is, if the firm is producing 100 units and they increase their production by one
unit, they will incur additional costs of approximately 205 dollars.
Activity 3.17 By using C(q + 1) C(q) directly when q = 100, determine how
good the approximation found in Example 3.19 is.
Generally then, if f is some economically meaningful function, its derivative is referred
to as the marginal of f and we denote this by Mf . For instance, if R(q) is the revenue
function for a firm, the marginal revenue, MR(q), is just R0 (q).
Elasticities
Suppose that, as in Section 2.1.5, we have a market where consumers purchase a good
according to the demand function, q D (p). If the price of this good was to increase from
p to p + p, then there will be a change in the quantity demanded by the consumers
from q to q + q. Indeed, since a rise in price will usually lead to a fall in demand, we
would expect q to be negative here. In these circumstances, we can see how these
changes are related by noting that
q = q D (p + p) q D (p) ' q 0 (p)p
q
' q 0 (p),
p
where we have used q to denote the quantity demanded, i.e. q(p) = q D (p).
Now, suppose that we are interested in the relative change in quantity, q/q, and the
relative change in price, p/p, we can see that the ratio of these two terms is then given
by
q/q
p q
p
=
' q 0 (p).
p/p
q p
q
Indeed, as q is usually negative (whereas the other terms on the left-hand-side, i.e. p,
q and p, are all positive) we would usually expect the right-hand-side to be negative
as well. With this in mind, we define the [price] elasticity of demand, (p), to be
p
(p) = q 0 (p),
q
where q = q D (p) and the minus sign is introduced so that, in the usual case where q is
negative, we can be sure that (p) itself will be positive.7 Then, we can see that using
q
p
' (p)
,
q
p
7
Some books omit the minus sign in their definition of the elasticity of demand, but it will be useful
for us to include it as it is easier to deal with positive quantities.
73
3. Differentiation
we can see how the relative change in quantity is simply related to the relative change
in price via the elasticity of demand.
Example 3.20 Suppose that the demand function for some good is given by
q D (p) = 10pr where r is a constant. Find the elasticity of demand. What does this
tell us about the effect of relative changes in price on relative changes in quantity?
Here we have q = q D (p) = 10pr where r is a constant and so,

q 0 (p) = 10rpr1 ,
which means that the elasticity of demand is given by

p 0
p
r1
= r,
(p) = q (p) =
10rp
q
10pr
i.e. (p) is a constant too. This means that, using
q
p
' (p)
,
q
p
we see that a relative increase in price of, say, x% will lead to a relative decrease in
quantity purchased of (approximately) rx%.
Indeed, we will see, in Section 4.2.3, that elasticities can also give us useful information
about how the revenue, R = pq, generated from selling a quantity, q, at a price of p per
unit will be affected by increases in the price.
3.3.4
Existence of derivatives
Although we will usually be dealing with situations where a function has a derivative at
every point where it is defined, we will occasionally encounter situations where there is
at least one point at which the derivative of a function does not exist. Just so that we
are aware of what this means and the kinds of situation in which it can arise, we
consider some of the most common ways in which a derivative can fail to exist at a
certain point.8
Discontinuous functions
If a function is discontinuous at a point, i.e. there is a point at which the function is not
continuous, then the derivative will not exist at that point as the next example
illustrates.
Example 3.21 Show that the derivative of the function defined by
(
1
x0
f (x) =
,
1 x < 0
8
See, for example, Section 2.8 of Binmore and Davies (2002) for a discussion of some similar cases.
74
does not exist when x = 0.

This function is illustrated in Figure 3.5(a) and, clearly, as the function is a
continuous horizontal line when x 6= 0, its derivative is defined and equal to zero as
long as x 6= 0. However, when x = 0, the function is discontinuous and its derivative,
if it exists, would be given by
f (h) f (0)
.
h0
h
f 0 (0) = lim
However, here we can not just find
f (h) f (0)
,
h
and let h 0 as we did in Section 3.1 since the value of f (h) is different depending
on whether h is positive or negative. In such cases, we say that the limit we seek, i.e.
f (h) f (0)
,
h0
h
lim
exists if, firstly, both of the limits

lim
h0
f (h) f (0)
h
and
lim+
h0
f (h) f (0)
,
h
exist9 and, secondly, if they exist, they must be equal. But, using the given function,
we see that
(1) 0
1
f (h) f (0)
lim
= lim
= lim
= ,
h0
h0
h0
h
h
h
and
(1) 0
1
f (h) f (0)
= lim+
= lim+ = ,
lim+
h0
h0 h
h0
h
h
i.e. neither of these limits exists as is not a value10 but more of a notational
convenience which tells us that a function is getting arbitrarily large in the limit.
Consequently, we see that
f (h) f (0)
,
h0
h
f 0 (0) = lim
fails to exist too and so the derivative of this function does not exist at x = 0.
Of course, the graph of a function can also have a discontinuity due to the presence of a
vertical asymptote. In such cases, the function is not actually defined at the value of x
where the asymptote occurs and so, because of this, the derivative cannot exist at this
point either.11 In both of these cases, as we cant ascribe a gradient to the function at
these points, the function cant have a tangent line at these points.
9
Notice that the former limit allows us to deal with negative h and the latter allows us to deal with
positive h. Also recall that the notation h 0 and h 0+ was explained in Example 2.2.
10
That is, it is not a real number.
11
Well come across this again in Section 4.4.3.
75
3. Differentiation
y
y
y = x1/3
1
y=
y = |x|
1
x0
1 x < 0
(a)
(b)
(c)
Figure 3.5: The graphs of three functions that have no derivative at x = 0 as explained in
(a) Example 3.21, (b) Example 3.22 and (c) Example 3.23. We note however that, unlike
the functions in (a) and (b), the function in (c) does have a tangent line at x = 0 given
by the vertical line with equation x = 0.
Continuous functions with corners
But, even if a function is continuous at every point, the derivative will not exist at
points where the curve changes too sharply, i.e. when the curve has a corner, as the
next example illustrates.
Example 3.22 Show that the derivative of the function f (x) = |x| does not exist
when x = 0.
This function is illustrated in Figure 3.5(b) and, clearly, as the function is the
continuous straight line f (x) = x when x < 0 and f (x) = x when x > 0, its
derivative is defined and equal to 1 when x < 0 and 1 when x > 0. However, when
x = 0, the function has a corner and its derivative, if it exists, would be given by
f (h) f (0)
.
h0
h
f 0 (0) = lim
However here, as in Example 3.21, we can not just find

f (h) f (0)
,
h
and let h 0 as we did in Section 3.1 since the value of f (h) is different depending
on whether h is positive or negative. In such cases, we again say that the limit we
seek, i.e.
f (h) f (0)
lim
,
h0
h
exists if, firstly, both of the limits
lim
h0
f (h) f (0)
h
and
lim+
h0
f (h) f (0)
,
h
exist12 and, secondly, if they exist, they must be equal. But, using the given
function, we see that
lim
h0
76
f (h) f (0)
(h) 0
= lim
= lim 1 = 1,
h0
h0
h
h
and
f (h) f (0)
h0
= lim+
= lim+ 1 = 1,
h0
h0
h0
h
h
i.e. both of these limits exist, but they are clearly not equal. Consequently, we see
that
f (h) f (0)
f 0 (0) = lim
,
h0
h
fails to exist and so the derivative of this function does not exist at x = 0.
lim+
Observe that, in this case, the limits as h 0+ and as h 0 both exist, but the
problem occurs because they are not equal and so we cannot ascribe a value to the
derivative (i.e. the limit as h 0) in such situations. In particular, as this means that
we cant ascribe a gradient to f at this point, the function cant have a tangent line
here either.
Continuous functions with vertical tangent lines
Also, if a function is continuous at every point, the derivative will not exist at points
where the gradient of the curve becomes infinite, i.e. when the curve has a vertical
tangent line, as the next example illustrates.
Example 3.23 Show that the derivative of the function f (x) = x1/3 does not exist
when x = 0.
This function is illustrated in Figure 3.5(c) and, clearly, we can see that its
derivative is given by
1
f 0 (x) = 31 x2/3 = 2/3 ,
3x
which exists as long as x 6= 0. Of course, when x = 0, the derivative cannot exist
since, if we were to use this formula, we would have to divide by zero and this is
never allowed. However, we can see from Figure 3.5(c) that the graph of the function
has a vertical tangent line at x = 0 which is given by the vertical line with equation
x = 0.13 Thus, we have a situation where the derivative of the function does not
exist at x = 0, but it does have a tangent line at that point.
Observe that, in cases where the tangent line to f at a point is a vertical line we cannot
use (3.1) to find its equation as its derivative is not defined.14
12
Again, as in Example 3.21, the former limit allows us to deal with negative h and the latter allows
us to deal with positive h.
13
Notice that the tangent lines of the function are getting steeper as we move towards x = 0 on the
left and shallower as we move away from x = 0 on the right.
14
Well come across this again in Section 4.4.3.
77
3. Differentiation
3.4
Using higher-order derivatives
We have seen that the first derivative of a function, f , can allow us to find a linear
approximation to f around a by using the formula
f (x) ' f (a) + (x a)f 0 (a).

However, if we use higher-order derivatives, we can get better approximations to f
around a by using the formula
f (x) = f (a) + (x a)f 0 (a) +
(x a)2 00
(x a)n (n)
f (a) + +
f (a) + ,
2!
n!
(3.2)
which is called the Taylor series for f (x) about x = a.15 You will notice that the
right-hand-side of this formula is an infinite series and, for reasons beyond the scope of
this course, there will generally be conditions that depend on f and a that determine
whether this infinite series does indeed give us the value of f (x) that we expect to get
on the left-hand-side. For now, we just note that these conditions can be used to find a
set of values of x, that includes the point x = a, for which the formula works. Of course,
if the value of x in question does not lie in this set, the formula does not work!
In this course, we will often just use the first few terms from the Taylor series to get an
approximate value of f (x).16 And, as long as we are considering what this formula tells
us about f (x) when x is close to a, these approximations will generally be more than
adequate. For instance, if we take n = 1 in this formula, i.e. if we take the first two
terms of the Taylor Series, we recover our linear approximation to f around a and, if we
take n = 2, we get
f (x) ' f (a) + (x a)f 0 (a) +
(x a)2 00
f (a),
2!
which is now a quadratic approximation to f around a. Indeed, we have seen how useful
the linear approximation is in Section 3.3.2 and the quadratic approximation will be
useful in the next chapter.
3.4.1
Maclaurin series
Lets start with the relatively simple case of a Maclaurin series which is what we call a
Taylor series about x = 0. That is, the Maclaurin series of the function f (x) is found by
setting a = 0 in (3.2) to get
f (x) = f (0) + xf 0 (0) +
xn
x2 00
f (0) + + f (n) (0) + .
2!
n!
(3.3)
To see how this works, lets start by finding a simple Maclaurin series.
15
See, for instance, Section 2.13 of Binmore and Davies (2002) for an explanation of where this formula
comes from.
16
It will be an approximation since, if we only keep the first few terms from the beginning of the series,
we lose all the information about the value of f (x) that is contained in the terms we are neglecting.
78
3.4. Using higher-order derivatives
Example 3.24 Find the Maclaurin series for ex .

Here we have f (x) = ex so that f (0) = 1. We also note that the first three
derivatives of this function are
f 0 (x) = ex ,
f 00 (x) = ex
and f 000 (x) = ex .
Indeed, it should be clear that f (n) (x) = ex for all n 1. Then, to use these in (3.3),
we need to evaluate these derivatives at x = 0, i.e. we find that
f 0 (0) = e0 = 1,
f 00 (0) = e0 = 1 and f 000 (0) = e0 = 1.
Indeed, it should be clear that f (n) (0) = e0 = 1 for all n 1. Consequently, putting
this into (3.3), we get
x2 x3
xn
e =1+x+
+
+ +
+ ,
2!
3!
n!
x
and this formula works for all x R.

In particular, observe that ex is only equal to the series on the right-hand-side if we
keep all of the terms in this infinite series. Of course, it is not always so easy to find a
Maclaurin series and so lets look at another example.
Example 3.25 Find the Maclaurin series for (1 + x)r when r Q.
Here we have f (x) = (1 + x)r so that f (0) = 1. We also note that the first three
derivatives of this function are
f 0 (x) = r(1 + x)r1 , f 00 (x) = r(r 1)(1 + x)r2 and f 000 (x) = r(r 1)(r 2)(1 + x)r3 .
Indeed, it should be clear that
f (n) (x) = r(r 1) (r [n 1])(1 + x)rn ,
for all n 1. Then, to use these in (3.3), we need to evaluate these derivatives at
x = 0, i.e. we find that
f 0 (0) = r,
f 00 (x) = r(r 1) and f 000 (0) = r(r 1)(r 2).
Indeed, spotting the pattern, it should be clear that

f (n) (0) = r(r 1) (r [n 1]),
for all n 1. Consequently, putting this into (3.3), we get
(1 + x)r = 1 + rx +
r(r 1) 2 r(r 1)(r 2) 3

x +
x +
2!
3!
r(r 1) (r [n 1]) n
+
x + ,
n!
and this formula works when |x| < 1.
79
3. Differentiation
In particular, notice that if r Q but r 6 N, this is always an infinite series as, for any
n N, we will find that r [n 1] 6= 0. However, if r N, we will find a value of n,
namely n = r + 1 that makes r [n 1] = 0 and this will mean that all of the terms
with n r + 1 will be zero, i.e. the Maclaurin series will be finite and will terminate at
the term where n = r. This is a very special Maclaurin series that you may have
encountered before as the binomial theorem and we look at some examples of this
special case in Activity 3.18.
Activity 3.18 Use the Maclaurin series for (1 + x)r which we found in
Example 3.25 to find (1 + x)2 and (1 + x)3 .
As well as the two Maclaurin series derived in Examples 3.24 and 3.25, you should also
remember the following
x3 x5
x2n+1
sin x = x
+
+ +
+
3!
5!
(2n + 1)!
cos x = 1
x2n
x2 x4
+
+ +
+
2!
4!
(2n)!
for x R.
for x R.
x2 x3
xn
+
+ + (1)n+1 + for |x| < 1.
2
3
n
In particular, observe how these series differ in their first term, the presence of terms of
odd and even degree and the absence of factorials in the series for ln(1 + x).
ln(1 + x) = x
Using Maclaurin series as approximations

As we have seen, a Maclaurin series is an infinite series in powers of x and, by taking a
certain number of terms, we can use it to approximate a function. In particular, we say
that we have the nth-order Maclaurin series of a function if we keep all of the terms up
to and including the one in xn and discard the rest.
Example 3.26 Write down the second and fourth-order Maclaurin series for cos x.
As we saw above, the Maclaurin series for cos x is given by the infinite series
x2 x 4
x2n
cos x = 1
+
+ +
+ ,
2!
4!
(2n)!
As such, the second-order Maclaurin series for cos x is
1
x2
,
2!
which, since there is no x3 term in the Maclaurin series for cos x, is also the
third-order Maclaurin series for cos x. Similarly, the fourth-order Maclaurin series for
cos x is
x2 x4
1
+ ,
2!
4!
5
which, since there is no x term in the Maclaurin series for cos x, is also the
fifth-order Maclaurin series for cos x.
80
These nth-order Maclaurin series can be used to approximate a function, f (x), for
values of x close to x = 0. In general, there are two factors that determine how accurate
this approximation will be, namely
the value of x we are considering: the closer this value of x is to x = 0, the better
the approximation will be, and
the order of the Maclaurin series we use: the more terms we keep, the better the
approximation will be.
The precise way of determining the accuracy of such approximations in terms of these
two factors will be dealt with in 176 Further Calculus where you will encounter Taylors
theorem. But, we can see how it works and begin to see how these factors affect the
accuracy of our approximations by considering some examples.
Example 3.27 Use the fourth-order Maclaurin series for cos x to find an
approximate value for cos 1 and cos 2.
The fourth-order Maclaurin series for cos x is
1
x2 x4
+ .
2!
4!
This means that taking x = 1, we see that

cos 1 ' 1
13
12 14
+
= ,
2
24
24
which is 0.5417 to 4dp. Using a calculator we see that the true value of cos 1 is 0.5403
to 4dp and so this is a good approximation as, to 2dp, it gives us 0.54 either way.
Similarly, taking x = 2, we see that
cos 2 ' 1
22 24
2
1
+
=12+ = ,
2
24
3
3
which is 0.3333 to 4dp. Using a calculator we see that the true value of cos 2 is
0.4161 to 4dp and so this is a poor approximation as it isnt even accurate to 1dp.
But, of course, we should expect our approximations to be poor if we move too far away
from x = 0 as, by definition, the Maclaurin series represents how the function is
behaving around x = 0. To see this, consider the curves in Figure 3.6 which illustrate
how the fourth-order Maclaurin series for cos x becomes less accurate at approximating
the function as we move away from x = 0.
The other way in which the accuracy of our approximation to a function can be affected
is the number of terms we take in the Maclaurin series. For instance, the second-order
Maclaurin series for cos x contains less information about the function than the
fourth-order one and so we would expect this to give us a worse approximation. This
can be seen in Figure 3.7, which illustrates how the second-order Maclaurin series is
even less accurate than the fourth-order one as we move away from x = 0.
81
3. Differentiation
Figure 3.6: The solid curve is the graph of the function cos x and the dashed curve is the
graph of the fourth-order Maclaurin series for this function. Observe how the Maclaurin
series moves away from the function as we take values of x further away from x = 0.
Using Maclaurin series to approximate other functions
We now look at some ways of finding Maclaurin series for more complicated functions
and see how we can use these to find approximations.
Example 3.28 Find the fourth-order Maclaurin series for x ex .
There are two ways to do this. We could use (3.3) to see that as f (x) = x ex we have
f (0) = 0 and then, using what we found in Activity 3.15 above, i.e.
f 0 (x) = (1+x) ex ,
f 00 (x) = (2+x) ex ,
f 000 (x) = (3+x) ex ,
and f (4) (x) = (4+x) ex ,
we see that
f 0 (0) = 1,
f 00 (0) = 2,
f 000 (0) = 3,
and f (4) (0) = 4.
So, putting this into (3.3) we see that, keeping terms up to x4 ,

0 + x(1) +
x3
x4
x3 x 4
x2
(2) + (3) + (4) = x + x2 +
+ ,
2!
3!
4!
2
6
is the fourth-order Maclaurin series for x ex .

Alternatively, since we know that the Maclaurin series for ex is given by
ex = 1 + x +
82
x2 x3 x4
+
+
+ ,
2!
3!
4!
Figure 3.7: The solid curve is the graph of the function cos x, the dotted curve is the graph
of its second-order Maclaurin series and the dashed curve is the graph of its fourth-order
Maclaurin series. Observe how the former less accurately tracks the function than the
latter as we take values of x further away from x = 0.
we can see that

x2 x3 x4
x3 x4
xe = x 1 + x +
+
+
+ = x + x2 +
+
+ ,
2!
3!
4!
2
6
x
as before, if we keep terms up to x4 for the fourth-order Maclaurin series for x ex .

This example illustrates a general point: when asked to find a Maclaurin series of a
certain order, we can always use the definition and differentiation. But, if the
derivatives start to become difficult to calculate, it is always easier to use the Maclaurin
series for the elementary functions (which we saw above) and a little algebra to find
what we are looking for. Lets consider another example to see how we can do this in a
slightly harder situation.
Example 3.29 Find the fourth-order Maclaurin series for cos(ln(1 + x)).
Here we have f (x) = cos(ln(1 + x)) which is a composition where f (x) = cos y with
y = ln(1 + x). So we need to look at the Maclaurin series for cos y which is given by
y2 y4
+
+ ,
2!
4!
and y, in turn, will be given by the Maclaurin series for ln(1 + x), i.e.
cos y = 1
y = ln(1 + x) = x
x2 x3 x 4
+
+
+ .
2
3
4!
83
3. Differentiation
So, substituting our series for y into our series for cos y, we can see that

2

4
1
x2 x3 x4
x2 x3 x4
1
f (x) = 1
x
+
+ +
x
+
+ + ,
2!
2
3
4
4!
2
3
4
|
|
{z
}
{z
}
A
and we start by looking at how the terms A and B contribute to the series if we are
only interested in terms up to x4 . For A, we have
2

x2 x3 x4
+
+
A= x
2
3
4

x2 x3 x4
x2 x3 x4
= x
+
+
x
+
+ ,
2
3
4
2
3
4
so we can multiply each term in the second bracket by the appropriate terms in the
first bracket (taking care to include cross-terms) to get
2
3
2 2
x
x
x
x
11
A = (x)(x) 2
(x) + 2
(x) +
+ = x2 x3 + x4 + ,
2
3
2
2
12
where indicates terms we can ignore because their degree is greater than four.
Similarly, for B, we have
4

x2 x3 x4
+
+ ,
B = x
2
3
4
which is the bracketed expression

x2 x3 x 4
x
+
+ ,
2
3
4
multiplied by itself four times. The terms which arise from this product are obtained
by multiplying together four objects, one from each occurrence of the bracketed
expression. Since the term with lowest power of x in each bracket is x, it is only by
taking the x from each bracket that we obtain a term which is at most x4 and so we
get
B = x4 + ,
where indicates terms we can ignore because their degree is greater than four.
Of course, using similar reasoning, we can see that there will be no further terms for
our series as the next term in the cos y series (i.e. the first one we omitted above) is
y 6 /6! and the smallest term this can yield looks like x6 whose degree is greater
than four.
Therefore, putting this all together, we have
A B
+ +
2! 4!

1
11 4
1
2
3
4
=1
x x + x + +
x + +
2
12
24
cos (ln(1 + x)) = 1
5
x2 x3
+
x4 + ,
2
2
12
and this gives us the fourth-order Maclaurin series for cos(ln(1 + x)) as we have kept
all of the terms up to x4 .
=1
84
Activity 3.19 Find the fourth-order Maclaurin series for cos(ln(1 + x)) by using
the definition and differentiation to verify the answer we found in Example 3.29.
(Notice that it is harder to work it out using this method!)
Once we have the Maclaurin series of a function, f (x), we can use it to estimate the
value of the function at some value of x close to zero as we did above.
Example 3.30 Use the Maclaurin series we found in Example 3.29 to find an
approximate value for cos(ln 1.1) and cos(ln 1.9).
To find an approximate value for cos(ln 1.1), we use the Maclaurin series above to
get the approximation
cos(ln(1 + x)) ' 1
x2 x3
5
+
x4 ,
2
2
12
and then set x = 0.1 to get

cos(ln 1.1) = cos(ln(1+0.1)) ' 1
0.12 0.13 5
+
0.14 = 10.005+0.00050.000042,
2
2
12
which is 0.995458 to 6dp. In passing we note that, using a calculator, the true value
is 0.995461 to 6dp and so this is a good approximation as, to 5dp, it gives us 0.99546
either way.
To find an approximate value for cos(ln 1.9), we use the approximation above with
x = 0.9 to get
cos(ln 1.9) = cos(ln(1+0.9)) ' 1
0.92 0.93 5
+
0.94 = 10.405+0.36450.273375,
2
2
12
which is 0.686125 to 6dp. In passing we note that, using a calculator, the true value
is 0.800987 to 6dp and so this is a poor approximation as it isnt even accurate to
1dp.
Observe that this approximation has deteriorated much more quickly than the one we
used when considering approximate values of cos x in Example 3.27. We wont pursue
the nature of this sensitivity here, but we do reiterate that we should expect our
approximations to be poor if we move too far away from x = 0 for, as we have seen, the
Maclaurin series is there to represent how the function is behaving around x = 0.
3.4.2
Taylor series
We now briefly consider what happens when we are looking for the Taylor series for
f (x) around x = a when a 6= 0. In this case, we follow the general method outlined
above, but now we have to use (3.2), i.e.
f (x) = f (a) + (x a)f 0 (a) +
(x a)2 00
(x a)n (n)
f (a) + +
f (a) + ,
2!
n!
that we saw earlier.
85
3. Differentiation
Example 3.31 Find the Taylor series for ex around x = 1.
Here we have f (x) = ex so that f (1) = e. We also note, as in Example 3.24, that
f (n) (x) = ex for n 1. Then, to use these derivatives in (3.2), we need to evaluate
them at x = 1, i.e. we find that f (n) (1) = e for n 1. Consequently, putting this into
(3.2), we get
ex = e +(x 1) e +
(x 1)2
(x 1)3
(x 1)n
e+
e+ +
e+ ,
2!
3!
n!
as the Taylor series for ex around x = 1.

Activity 3.20 We can write ex as e1 ex1 so that values of x around x = 1
correspond to values of x 1 around x = 0. Use this fact and the Maclaurin series for
ex which we found in Exercise 3.24 to derive the result we found in Example 3.31.
Activity 3.21 Find the Taylor series for ex around x = 2.
Using Taylor series as approximations
We can use the Taylor series of a function around x = a to get approximations to the
value of the function for values of x close to x = a in the same way as we used the
Maclaurin series of a function to get approximations to the value of the function for
values of x close to x = 0 in Section 3.4.1. As the ideas are so similar, we will just take a
brief look at how they work.
Example 3.32 Find an approximation to e1.1 using (a) the second-order Maclaurin
series for ex and (b) the second-order Taylor series for ex around x = 1. How do
these approximations compare?
For (a), we know from Example 3.24 that the second-order Maclaurin series for ex is
given by
x2
1+x+ ,
2!
and, using this, we find that
1.12
= 2.705.
2!
Incidentally, the exact value of e1.1 is 3.0042 (to 4dp) and so this approximation
doesnt even agree with this to 1dp.
e1.1 ' 1 + 1.1 +
For (b), we know from Example 3.31 that the second-order Taylor series for ex
around x = 1 is given by
(x 1)2
e +(x 1) e +
e,
2!
and, using this, we find that
e1.1 ' e +(1.1 0.1) e +
86
(1.1 1)2
e = 1.105 e,
2!
3.4. Learning outcomes
which, if we know the value of e, gives us 3.0037 (to 4dp). This agrees with the
exact value of e1.1 to 3dp.
As we should expect, the answer to (b) gives us a better approximation to e1.1 than
the one we found in (a) since x = 1.1 is closer to x = 1 than it is to x = 0. But, on
the other hand, the answer to (a) didnt require us to have any accurate knowledge
of the value of e itself!
Following on from this example, as we can see in Figure 3.8, we observe that the
Maclaurin series for ex is most accurate when x is close to x = 0 whereas the Taylor
series for ex about x = 1 is most accurate when x is close to x = 1. This is, of course,
exactly what we should expect!
Figure 3.8: The solid curve is the graph of the function ex , the dashed curve is the graph
of its second-order Maclaurin series and the dotted curve is the graph of its second-order
Taylor series about x = 1. Observe how, as we might expect, the Maclaurin series is more
accurate around x = 0 and this Taylor series is more accurate around x = 1.
Learning outcomes
At the end of this chapter and having completed the relevant reading and activities, you
should be able to:
find simple derivatives using the definition of the derivative;
find derivatives using standard derivatives and the rules of differentiation;
use the derivative to find tangent lines and use these to approximate functions;
solve problems from economics-based subjects that involve derivatives;
find Maclaurin and Taylor series and use these to approximate functions.
87
3. Differentiation
Solutions to activities
We need to find the derivative of the function f (x) = x2 at the point x = 1, i.e.
f 0 (1). So, using the definition of the derivative with a = 1, we start by looking at
f (1 + h) f (1)
(1 + h)2 (1)2
=
,
h
h
f (1 + h) f (1)
(1 2h + h2 ) 1
2h + h2
=
=
= 2 + h.
h
h
h
f (1 + h) f (1)
f (1) = lim
= lim
h0
h0
h
0

2 + h = 2,
because, in the limit as h 0, we see that 2 + h goes to 2.

In Example 3.2, we showed that if f (x) = x2 , then f 0 (x) = 2x. This means that
f 0 (1) = 2 in agreement with what we saw in Activity 3.1.
To find the point at which the derivative of f (x) = x2 , i.e. f 0 (x) = 2x, is equal to (a) 16,
we see that
f 0 (x) = 2x = 16 when x = 8,
and (b) 4, we see that
f 0 (x) = 2x = 4 when x = 2.
Given the linear combination rule, i.e.

dg
d
df
+l ,
kf (x) + lg(x) = k
dx
dx
dx
we can derive the constant multiple rule by setting l = 0 so that

d
d
df
dg
df
kf (x) =
kf (x) + 0g(x) = k
+0
=k ,
dx
dx
dx
dx
dx
the sum rule by setting k = 1 and l = 1 so that

d
d
df
dg
df
dg
f (x) + g(x) =
1f (x) + 1g(x) = 1
+1
=
+
,
dx
dx
dx
dx
dx dx
and the difference rule by setting k = 1 and l = 1 so that

d
d
df
dg
df
dg
f (x) g(x) =
1f (x) + (1)g(x) = 1
+ (1)
=
.
dx
dx
dx
dx
dx dx
88

For (a), we use the constant multiple rule to see that

d
d
3 cos x = 3
cos x = 3 sin x = 3 sin x.
dx
dx
For (b), we use the sum rule to see that

d
d
d
x
x
x
e + cos x =
e
+
cos x = e + sin x = ex sin x.
dx
dx
dx
For (c), we use the linear combination rule to see that

d
d
d
1
3
3 sin x 3 ln x = 3
sin x 3
ln x = 3 cos x 3
= 3 cos x .
dx
dx
dx
x
x
For (a), h(x) = x sin x is the product of the two functions
f (x) = x
and
g(x) = sin x,
f 0 (x) = 1
and
g 0 (x) = cos x.
and these give us

h0 (x) = (1)(sin x) + (x)(cos x) = sin x + x cos x.
For (b), h(x) = ex cos x is the product of the two functions
f (x) = ex
and
g(x) = cos x,
and these give us

f 0 (x) = ex
and
g 0 (x) = sin x.

h0 (x) = (ex )(cos x) + (ex )( sin x) = ex (cos x sin x).
For (c), h(x) = sin x cos x is the product of the two functions
f (x) = sin x
and
g(x) = cos x,
and these give us

f 0 (x) = cos x
and
g 0 (x) = sin x.

h0 (x) = (cos x)(cos x) + (sin x)( sin x) = cos2 x sin2 x.
Then, using the double angle formulae
sin x cos x = sin(2x) and
cos2 x sin2 x = cos(2x),
89
3. Differentiation
from (2.6), this means that we have

d 1
sin(2x) = cos(2x),
dx 2
so that, using the constant multiple rule, we can deduce that

d
sin(2x) = 2 cos(2x).
dx
This result will make sense once we have seen the chain rule and, in particular,
Activity 3.8(a).
sin x
For (a), h(x) =
is the quotient of the two functions
x
f (x) = sin x
and
g(x) = x,
f 0 (x) = cos x
and
g 0 (x) = 1.
and these give us

h0 (x) =
x cos x sin x
(cos x)(x) (sin x)(1)
=
.
2
x
x2
In this case, the original function and the derivative are only defined if x 6= 0.
For (b), h(x) =
ex
cos x
f (x) = ex
and
g(x) = cos x,
and these give us

f 0 (x) = ex
and

h0 (x) =
g 0 (x) = sin x.
(ex )(cos x) (ex )( sin x)

cos x + sin x x
=
e .
2
[cos x]
cos2 x
In this case, the original function and the derivative are only defined if cos x 6= 0, i.e. if
x 6= (2n + 1) 2 for n Z.
For (c), h(x) =
sin x
cos x
f (x) = sin x
and
g(x) = cos x,
and these give us

f 0 (x) = cos x
and
g 0 (x) = sin x.
(cos x)(cos x) (sin x)( sin x)

cos2 x + sin2 x
h (x) =
=
.
[cos x]2
cos2 x
0
90
In this case, the original function and the derivative are only defined if cos x 6= 0, i.e. if
x 6= (2n + 1) 2 for n Z.
Indeed, using the Pythagorean identity
sin2 x + cos2 x = 1 and the definitions
tan x =
sin x
cos x
and
sec x =
1
,
cos x
from (2.2), (2.1) and Section 2.1.2, we can deduce that

d
1
tan x =
= sec2 x,
2
dx
cos x
as long as x 6= (2n + 1) 2 for n Z.
Given that h(x) = (2x + 1)3 , we can multiply out the brackets to get
h(x) = 8x3 + 12x2 + 6x + 1,
which means that
h0 (x) = 24x2 + 24x + 6 = 6(4x2 + 4x + 1) = 6(2x + 1)2 ,
in agreement with what we saw in Example 3.10.
For (a), h(x) = sin(2x) is the composition of the functions
f (g) = sin g
and
g(x) = 2x.
As such we have
f 0 (g) = cos g
and
g 0 (x) = 2,

h0 (x) = (cos g)(2) = 2 cos(2x),
which agrees with what we found in Activity 3.5(c).
For (b), h(x) = ln(cos x) is the composition of the functions
f (g) = ln g
and
g(x) = cos x.
As such we have
1
and
g 0 (x) = sin x,
g

sin x
1
0
h (x) =
( sin x) =
= tan x.
g
cos x
f 0 (g) =
For (c), h(x) = ln(ex ) is the composition of the functions

f (g) = ln g
and
g(x) = ex .
91
3. Differentiation
As such we have
f 0 (g) =
1
g
and
g 0 (x) = ex ,

1
ex
h (x) =
(ex ) = x = 1.
g
e
0
Of course, this is obvious as ln(ex ) = x and so its derivative with respect to x is

therefore 1.
Given that ax = ex ln a , we use the chain rule with h(g) = eg and g(x) = x ln a to get

d
dax
x ln a
x ln a
ln a = ax ln a,
= e
=
e
dx
dx
as required.
Writing the quotient f (x)/g(x) as the product f (x)[g(x)]1 , the product rule gives us

d
df
1
2 dg
1
[g(x)] + f (x) [g(x)]
f (x)[g(x)]
,
=
dx
dx
dx
where we have used the chain rule to differentiate [g(x)]1 with respect to x. Rewriting
this, we then have
df
dg

g(x) f (x)
d f (x)
dx ,
= dx
dx g(x)
[g(x)]2
which is the quotient rule, as required.
We have y = f (x) so that x = f 1 (y). Thus, differentiating both sides of the latter with
respect to x, we get
dx
df 1 dy
=
,
dx
dy dx
where we have used the chain rule on the right-hand-side as y itself is a function of x
since y = f (x). This gives us

df 1 dy
df 1
df
1=
=
=1
,
dy dx
dy
dx
as required.17 In particular, observe that this formula makes no sense at points where
f 0 (x) = 0.
17
See Section 2.9 of Binmore and Davies (2002) for a geometric view of this result.
92

Here we have y = f (x) = ex and x = f 1 (y) = ln y so, using
Activity 3.11, we see that

1
d
d
ln y = 1
ex = x =
dy
dx
e
the result from

1
,
y
as (ex )0 = ex = y.
There is, generally, no need to apply the rules of differentiation in as much detail as we
have been using. So, lets do the three examples in this activity quickly.
2
For (a), we have h(x) = ex ln(sin x) which is the product of two compositions and so
using the product and chain rules we get

2
cos x
ex
x2
x2
h (x) = 2x e
ln(sin x) + e
=
2x sin x ln(sin x) + cos x .
sin x
sin x
0
For (b), we have

sin(cos x)
,
esin x
which is the quotient of two compositions and so using the quotient and chain rules we
get

cos(cos x)( sin x) esin x sin(cos x) esin x cos x
0
h (x) =
[esin x ]2
sin x cos(cos x) + cos x sin(cos x)
.
=
esin x
h(x) =
For (c), we have h(x) = sin2 (3x) + cos2 (3x) which is the sum of two compositions and so
we can easily use the chain rule to see that
h0 (x) = 2 sin(3x) cos(3x)(3) + 2 cos(3x)[ sin(3x)](3) = 0.
Of course, this is obvious as sin2 (3x) + cos2 (3x) = 1 using (2.2) and so its derivative
with respect to x is zero.
We have seen that the first four derivatives are given by
f 0 (x),
f 00 (x) = f (x),
f 000 (x) = f 0 (x),
and f (4) (x) = f 00 (x) = f (x),
which returns us to our original function. Indeed, we can then see that the next four
derivatives will be given by
f (5) (x) = f 0 (x),
f (6) (x) = f (x),
f (7) (x) = f 0 (x) and f (8) (x) = f 00 (x) = f (x),
93
3. Differentiation
which, again, returns us to our original function. This means that, spotting the pattern,
we can see that
f (x)
n = 4, 8, . . .
f 0 (x)
n = 1, 5, 9, . . .
f (n) (x) =
f (x) n = 2, 6, 10, . . .
0
f (x) n = 3, 7, 11, . . .
for n 1.

To find the first four derivatives of x ex , we use the product rule to see that
f 0 (x) = (1)(ex ) + (x)(ex ) = (1 + x) ex ,
f 00 (x) = (1)(ex ) + (1 + x)(ex ) = (2 + x) ex ,
f 000 (x) = (1)(ex ) + (2 + x)(ex ) = (3 + x) ex , and
f (4) (x) = (1)(ex ) + (3 + x)(ex ) = (4 + x) ex .
Indeed, spotting the pattern, we can deduce that f (n) (x) = (n + x) ex for n 1.
Here we have f (x) = ex so that f 0 (x) = ex . Then using (3.1), we see that when x = 1 we
have
y f (1)
= y e1 = e1 (x 1) = y = e x.
f 0 (1) =
x1
as the equation of the tangent line to the function f (x) = ex at x = 1.
Here we have C(q) = 1000 + 5q + q 2 and so, when operating at q = 100, the increase in
cost given an increase in quantity of one is given by
C = C(101) C(100) = (1000 + 5(101) + 1012 ) (1000 + 5(100) + 1002 ) = 206.
This is pretty close to the approximate answer of 205 that we found in Example 3.19,
especially if we consider this as a relative error of 1/206 = 0.49% (to 2dp) instead of an
absolute error of one.
The Maclaurin series for (1 + x)2 is given by
(1 + x)2 = 1 + 2x +
(2)(1) 2
x = 1 + 2x + x2 ,
2!
as all terms involving xn with n 3 will have a coefficient of zero. Similarly, the
Maclaurin series for (1 + x)3 is given by
(1 + x)3 = 1 + 3x +
94
(3)(2) 2 (3)(2)(1) 3
x +
x = 1 + 3x + 3x2 + x3 ,
2!
3!
as all terms involving xn with n 4 will have a coefficient of zero. Of course, this is
exactly what we would get if we just multiplied out the brackets in the usual way!
To use (3.3), we see that f (x) = cos(ln(1 + x)) gives
f (0) = cos(ln 1) = cos 0 = 1,

and then, finding the first four derivatives of f (x), we get
f 0 (x) =
sin(ln(1 + x))
,
1+x
f 00 (x) =
sin(ln(1 + x)) cos(ln(1 + x))

,
(1 + x)2
f 000 (x) =
3 cos(ln(1 + x)) sin(ln(1 + x))

,
(1 + x)3
f (4) (x) = 10
cos(ln(1 + x))
,
(1 + x)4
after some fairly tortuous differentiation. These then give us

f 0 (0) = 0,
f 00 (0) = 1,
f 000 (0) = 3 and f (4) (0) = 10,
if we evaluate them at x = 0. Consequently, putting these into (3.3), we get

x2
x3
x4
(1) + (3) + (10) +
2!
3!
4!
x2 x3
5 4
=1
+
x + ,
2
3
12
cos(ln(1 + x)) = 1 + x(0) +
and this gives us the fourth-order Maclaurin series for cos(ln(1 + x)) in agreement with
what we saw before in Example 3.29. Notice, however, that this method involved some
fairly complicated differentiation whereas the method in Example 3.29 only involved
some simple algebra!
For values of y around y = 0 we have the Maclaurin series
ey = 1 + y +
yn
y2 y3
+
+ +
+ ,
2!
3!
n!
so that, if we let y = x 1 for values of x around x = 1, we still have values of y around

y = 0, i.e. we can write
ex1 = 1 + (x 1) +
(x 1)2 (x 1)3
(x 1)n
+
+ +
+ ,
2!
3!
n!
95
3. Differentiation
which gives us the Taylor series for ex1 for values of x around x = 1. So, as
ex = e1 ex1 , this means that
ex = e +(x 1) e +
(x 1)3
(x 1)n
(x 1)2
e+
e+ +
e+ ,
2!
3!
n!
is the Taylor series for ex for values of x around x = 1 in agreement with what we found
in Example 3.31.
To find the Taylor series for ex around x = 2, we can either use (3.2) or the method we
saw in Activity 3.20.
Method I: Using (3.2), we have f (x) = ex so that f (2) = e2 . We also note, as in
Example 3.24, that f (n) (x) = ex for n 1. Then, to use these derivatives in (3.2), we
need to evaluate them at x = 2, i.e. we find that f (n) (2) = e2 for n 1. Consequently,
putting these into (3.2), we get
ex = e2 +(x 2) e2 +
(x 2)2 2 (x 2)3 2
(x 2)n 2
e +
e + +
e + ,
2!
3!
n!
as the Taylor series for ex around x = 2.

Method II: Using the method of Activity 3.20, we know that for values of y around
y = 0 we have the Maclaurin series
ey = 1 + y +
yn
y2 y3
+
+ +
+ ,
2!
3!
n!
so that, if we let y = x 2 for values of x around x = 2, we still have values of y around

y = 0, i.e. we can write
ex2 = 1 + (x 2) +
(x 2)2 (x 2)3
(x 2)n
+
+ +
+ ,
2!
3!
n!
which gives us the Taylor series for ex2 for values of x around x = 2. So, as
ex = e2 ex2 , this means that
ex = e2 +(x 2) e2 +
(x 2)n 2
(x 2)2 2 (x 2)3 2
e +
e + +
e + ,
2!
3!
n!
is the Taylor series for ex for values of x around x = 2 in agreement with what we have
just found using the other method.
Exercises
Exercise 3.1
Find the derivatives of the following functions.
(a) esin x cos x,
96
(b)
tan x
,
ex2
(c) sin(x ex ).
Exercise 3.2
Use the compound-angle formulae to show that

cos x = sin x +
2

sin x = cos x +
.
2
and
Hence use the chain rule to derive the derivative of cos x from the derivative of sin x.
Exercise 3.3
Verify that the point (e, e) is on the curve with equation
y = x ln x,
and find the equation of the tangent line to the curve at this point.
Consider, for some constants a and b, the curve with equation
y = ax2 + b.
For what values of a and b does this curve pass through the point (e, e) with the same
tangent line as the one you found above?
Exercise 3.4
Suppose the demand function for a good is
1
q D (p) = p
.
1 + p4
Find the elasticity of demand in terms of p and verify that it is positive if p > 0.
Exercise 3.5
Find the fourth-order Maclaurin series for ln

1 + sin x
.
1+x
Solutions to exercises
We apply the rules of differentiation quickly as we did in Activity 3.13.
(a) The function h(x) = esin x cos x is a product that has the composition esin x as one
of its terms. As such, applying the product rule we get
0
h (x) =
sin x
cos x cos x + e
sin x
sin x
cos x sin x esin x ,
where we have used the chain rule to differentiate the composition.
97
3. Differentiation
2
(b) The function h(x) = (tan x)/ ex is a quotient whose denominator is the
2
composition ex . As such, applying the quotient rule we get

2
x2
x2
(sec x) e (tan x) e (2x)
sec2 x 2x tan x
h0 (x) =
=
,
2
[ex ]2
ex2
where we have used the fact, from Activity 3.6(c), that the derivative of tan x is
sec2 x and the chain rule to differentiate the composition.
Also note that this derivative can be found by writing the function as
2
h(x) = (tan x) ex and, if we do this, we would use the product rule instead of the
quotient rule.
(c) The function h(x) = sin(x ex ) is the composition sin x after x ex where the latter
function is a product. As such, applying the chain rule we get

0
x
x
x
h (x) = cos(x e ) (1) e +x(e ) = (1 + x) ex cos(x ex ),
where we have used the product rule to differentiate the product.
Using the compound-angle formulae from (2.5), we have

= sin x cos + cos x sin = cos x,
sin x +
2
2
2
and

cos x +
= cos x cos sin x sin = sin x,
2
2
2
as required. Indeed, notice that we have used the facts, from Activity 2.3, that
sin(/2) = 1 and cos(/2) = 0.
Now, using chain rule and the derivative of sin x, we see that

d
sin x +
= cos x +
(1) = cos x +
,
dx
2
2
2
which, using the results we showed above, becomes
d
cos x = sin x,
dx
as required.
Substituting x = e into y = x ln x we get
y = e ln e = e(1) = e,
and so the point (e, e) lies on this curve. The gradient of the curve at any point is given
by the derivative of f (x) = x ln x and so, using the product rule, we get

1
0
f (x) = (1) ln(x) + x
= ln(x) + 1.
x
98
Thus, when x = e, the gradient of the curve is given by

f 0 (e) = ln(e) + 1 = 1 + 1 = 2,
which means that, using (3.1), we get
2=
ye
xe
y e = 2(x e)
y = 2x e,
as the equation of the tangent line to the curve y = x ln x at the point (e, e).
The curve y = ax2 + b will have a tangent line at (e, e) which is the same as the one we
have just found if, firstly, the curve goes through the point (e, e), i.e. a and b must satisfy
e = a e2 +b,
and, secondly, it has the same gradient at e, i.e. if the derivative of g(x) = ax2 + b at
x = e is two. That is, as
g 0 (x) = 2ax
we need
g 0 (e) = 2a e,
to be two. Thus, from the second condition, we have

2a e = 2
and, from the first condition, we have

1 2
e=
e +b
e
1
a= ,
e
b = e e = 0.
Consequently, we see that when a = 1/ e and b = 0 the curve y = ax2 + b passes through
the point (e, e) with the same tangent line as the one we found above.
We have the demand function
1
1
q D (p) = p
= (1 + p4 ) 2 ,
1 + p4
and so, setting q = q D (p), we can use the chain rule to get the derivative
3
1
2p3
q 0 (p) = (1 + p4 ) 2 (4p3 ) =
3 .
2
(1 + p4 ) 2
Then, using the definition of the elasticity of demand from Section 3.3.3, we have
!
p 0
p
2p3
2p4
(p) = q (p) =
=
,
1
3
q
1 + p4
(1 + p4 ) 2 (1 + p4 ) 2
in terms of p. Indeed, when p > 0, we have p4 > 0 and 1 + p4 > 0, which means that
(p) > 0 too.
99
3. Differentiation
We start by noticing that it really is much easier to make use of the standard Maclaurin
series rather than trying to use (3.3) directly on the given function. Especially as, in
order to apply (3.3), we would need to find the first four derivatives of the function to
answer this question and this would get very messy very quickly! Indeed, if we decide to
use the standard Maclaurin series, two methods present themselves.
Method I: We start by simplifying the function by using the laws of logarithms from
Section 2.1.4. This gives us

1 + sin x
= ln(1 + sin x) ln(1 + x),
ln
1+x
and so, we can easily use the Maclaurin series for ln(1 + x) from Section 3.4.1, i.e.
x2 x3 x4
+
+ ,
2
3
4
to get the second term in this difference. Then, using the Maclaurin series for sin x, also
from Section 3.4.1, we have
x3
sin x = x
+ ,
3!
which means that the first term in this difference is

x3
+
ln(1 + sin x) = ln 1 + x
3!

2

3

4
x3
1
1
x3
1
= x
+ ,
+
x
+
+
x
x
3!
2
3!
3
4
ln(1 + x) = x
where we have used the Maclaurin series for ln(1 + x) again in the second line. Now, as
we want to keep terms up to x4 , we can see that the brackets in the second term give us

x3
x3
x3
x4
x
+
x
+ = x2 2 x
+ = x2
+ ,
3!
3!
3!
3
where, here, were trying to make it clear that each term that arises from this product is
obtained by multiplying out the relevant brackets. Further, we see that the brackets in
the last two terms will give us x3 and x4 respectively. Overall, then, we have

1
x4
1 3
1 4
x3
2
+
x
+ +
x
x +
ln(1 + sin x) = x
3!
2
3
3
4
x2 x3 x4
=x
+
+ ,
2
6
12
for the first part of our difference. Putting these together in our expression for the
function, we then have

1 + sin x
ln
= ln(1 + sin x) ln(1 + x)
1+x

x2 x3 x4
x2 x3 x4
= x
+
+ x
+
+
2
3
4
2
6
12
x3 x4
= +
,
6
6
100
and this gives us the required fourth-order Maclaurin series.

Method II: We could also use the Maclaurin series for sin x which we saw above to get
1 + sin x = 1 + x
x3
+ ,
3!
and the fact that
3
1
= (1 + x)1 = 1 x + x2 x3 + x4 + ,
1+x
which, with r = 1, follows from a simple application of the Maclaurin series for
(1 + x)r that we saw in Example 3.25. This means that we have

x3
1 + sin x
2
3
4
= 1+x
+
1 x + x x + x +
1+x
3!

x3
2
3
4
2
3
1 x +
= 1 1 x + x x + x + + x 1 x + x x +
3!
x3 x4
=1
+
+ ,
3!
3!
if we want to keep terms up to x4 . Then, using the Maclaurin series for ln(1 + x) which
we saw above, we get

3

1 + sin x
x
x4
x3 x4
ln
= ln 1 + +
+
= +
+ ,
1+x
3!
3!
3!
3!
and this gives us the same fourth-order Maclaurin series as the one we found using the
other method.
101

mt1174 ch1-4

Uploaded by

Copyright:

Available Formats

mt1174 ch1-4

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

mt1174 ch1-4

Uploaded by

Copyright:

Available Formats

What is the main topic of the document?

What is the main topic of the document?

What are some examples of functions discussed in the document?

What are some examples of functions discussed in the document?

Calculus

University of London International Programmes

Online study resources . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Making use of the Online Library . . . . . . . . . . . . . . . . . .

Using this guide . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The use of calculators . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Introduction: What is a function? . . . . . . . . . . . . . . . . . . . . . .

Some elementary functions and their graphs . . . . . . . . . . . .

Introduction: What is differentiation? . . . . . . . . . . . . . . . . . . . .

How to find derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . . .

The rules of differentiation . . . . . . . . . . . . . . . . . . . . . .

The meaning of the derivative . . . . . . . . . . . . . . . . . . . .

Tangent lines and linear approximations . . . . . . . . . . . . . .

Using higher-order derivatives . . . . . . . . . . . . . . . . . . . . . . . .

Introduction: What is optimisation? . . . . . . . . . . . . . . . . . . . . .

Using first-order derivatives . . . . . . . . . . . . . . . . . . . . . . . . .

Increasing and decreasing functions . . . . . . . . . . . . . . . . .

An application: Elasticities revisited . . . . . . . . . . . . . . . . .

Using second-order derivatives . . . . . . . . . . . . . . . . . . . . . . . .

Second-derivatives and stationary points . . . . . . . . . . . . . .

Convex and concave functions . . . . . . . . . . . . . . . . . . . .

Sketching curves defined by polynomials . . . . . . . . . . . . . .

Sketching curves defined using other elementary functions . . . .

Asymptotes and cusps . . . . . . . . . . . . . . . . . . . . . . . .

What happens when differentiability fails? . . . . . . . . . . . . .

Introduction: What is integration? . . . . . . . . . . . . . . . . . . . . . .

How to find indefinite integrals . . . . . . . . . . . . . . . . . . . . . . .

The basic rules of integration . . . . . . . . . . . . . . . . . . . .

Using partial fractions to simplify integrands . . . . . . . . . . . .

Using trigonometric identities to simplify integrands . . . . . . . .

Definite integrals and what they represent . . . . . . . . . . . . .

Definite integrals and the other rules of integration . . . . . . . .

Marginal functions revisited . . . . . . . . . . . . . . . . . . . . .

Consumer and producer surpluses . . . . . . . . . . . . . . . . . .

Definite integrals and areas

6 Functions of several variables

Contours and sections . . . . . . . . . . . . . . . . . . . . . . . .

Sections and partial derivatives . . . . . . . . . . . . . . . . . . .

Finding partial derivatives . . . . . . . . . . . . . . . . . . . . . .

The chain rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

An application: Homogeneous functions . . . . . . . . . . . . . . .

Second-order partial derivatives . . . . . . . . . . . . . . . . . . .

Using partial derivatives . . . . . . . . . . . . . . . . . . . . . . . . . . .

Implicitly defined functions of two variables . . . . . . . . . . . .

Classifying stationary points . . . . . . . . . . . . . . . . . . . . .

Convex and concave functions . . . . . . . . . . . . . . . . . . . .

Finding optimal points on the boundary of a region . . . . . . . .

The method of Lagrange multipliers . . . . . . . . . . . . . . . . .

The meaning of the Lagrange multiplier . . . . . . . . . . . . . .

Introduction: What is a differential equation? . . . . . . . . . . . . . . .

Separable first-order ODEs . . . . . . . . . . . . . . . . . . . . . .

Linear first-order ODEs . . . . . . . . . . . . . . . . . . . . . . .

Homogeneous first-order ODEs . . . . . . . . . . . . . . . . . . .

Homogeneous second-order ODEs . . . . . . . . . . . . . . . . . .

Non-homogeneous second-order ODEs . . . . . . . . . . . . . . .