Mathematics Learning Centre

Introduction to Dierential Calculus

Christopher Thomas
c 1997 University of Sydney
Acknowledgements
Some parts of this booklet appeared in a similar form in the booklet Review of Dieren-
tiation Techniques published by the Mathematics Learning Centre.
I should like to thank Mary Barnes, Jackie Nicholas and Collin Phillips for their helpful
comments.
Christopher Thomas
December 1996
Contents
1 Introduction 1
1.1 An example of a rate of change: velocity . . . . . . . . . . . . . . . . . . . 1
1.1.1 Constant velocity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1.2 Non-constant velocity . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Other rates of change . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2 What is the derivative? 6
2.1 Tangents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 The derivative: the slope of a tangent to a graph . . . . . . . . . . . . . . 7
3 How do we nd derivatives (in practice)? 9
3.1 Derivatives of constant functions and powers . . . . . . . . . . . . . . . . . 9
3.2 Adding, subtracting, and multiplying by a constant . . . . . . . . . . . . . 12
3.3 The product rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.4 The Quotient Rule . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
3.5 The composite function rule (also known as the chain rule) . . . . . . . . . 15
3.6 Derivatives of exponential and logarithmic functions . . . . . . . . . . . . . 18
3.7 Derivatives of trigonometric functions . . . . . . . . . . . . . . . . . . . . . 21
4 What is dierential calculus used for? 24
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2 Optimisation problems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
4.2.1 Stationary points - the idea behind optimisation . . . . . . . . . . . 24
4.2.2 Types of stationary points . . . . . . . . . . . . . . . . . . . . . . . 25
4.2.3 Optimisation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
5 The clever idea behind dierential calculus (also known as dierentiation
from rst principles) 31
6 Solutions to exercises 35
2. 00 4. 00 6. 00 8. 00
100
200
300
(metres)
Di stance
t ime (seconds)
Mathematics Learning Centre, University of Sydney 1
1 Introduction
In day to day life we are often interested in the extent to which a change in one quantity
aects a change in another related quantity. This is called a rate of change. For example,
if you own a motor car you might be interested in how much a change in the amount of
fuel used aects how far you have travelled. This rate of change is called fuel consumption.
If your car has high fuel consumption then a large change in the amount of fuel in your
tank is accompanied by a small change in the distance you have travelled. Sprinters are
interested in how a change in time is related to a change in their position. This rate
of change is called velocity. Other rates of change may not have special names like fuel
consumption or velocity, but are nonetheless important. For example, an agronomist
might be interested in the extent to which a change in the amount of fertiliser used on a
particular crop aects the yield of the crop. Economists want to know how a change in
the price of a product aects the demand for that product.
Dierential calculus is about describing in a precise fashion the ways in which related
quantities change.
To proceed with this booklet you will need to be familiar with the concept of the slope
(also called the gradient) of a straight line. You may need to revise this concept before
continuing.
1.1 An example of a rate of change: velocity
1.1.1 Constant velocity
Figure 1 shows the graph of part of a motorists journey along a straight road. The
vertical axis represents the distance of the motorist from some xed reference point on
the road, which could for example be the motorists home. Time is represented along the
horizontal axis and is measured from some convenient instant (for example the instant an
observer starts a stopwatch).
Figure 1: Distance versus time graph for a motorists journey.
Mathematics Learning Centre, University of Sydney 2
Exercise 1.1
How far is the motorist in Figure 1 away from home at time t = 0 and at time t = 6?
Exercise 1.2
How far does the motorist travel in the rst two seconds (ie from time t = 0 to time t = 2)?
How far does the motorist travel in the two second interval from time t = 3 to t = 5? How far
do you think the motorist would travel in any two second interval of time?
The shape of the graph in Figure 1 tells us something special about the type of motion
that the motorist is undergoing. The fact that the graph is a straight line tells us that the
motorist is travelling at a constant velocity.
At a constant velocity equal increments in time result in equal changes in distance.
For a straight line graph equal increments in the horizontal direction result in the
same change in the vertical direction.
In Exercise 1.2 for example, you should have found that in the rst two seconds the
motorist travels 50 metres and that the motorist also travels 50 metres in the two seconds
between time t = 3 and t = 5.
Because the graph is a straight line we know that the motorist is travelling at a constant
velocity. What is this velocity? How can we calculate it from the graph? Well, in this
situation, velocity is calculated by dividing distance travelled by the time taken to travel
that distance. At time t = 6 the motorist was 250 metres from home and at time t = 2
the motorist was 150 metres away from home. The distance travelled over the four second
interval from time t = 2 to t = 6 was
distance travelled = 250 150 = 100
and the time taken was
time taken = 6 2 = 4
and so the velocity of the motorist is
velocity =
distance travelled
time taken
=
250 150
6 2
=
100
4
= 25 metres per second.
But this is exactly how we would calculate the slope of the line in Figure 1. Take a look
at Figure 2 where the above calculation of velocity is shown diagramatically.
The slope of a line is calculated by vertical rise divided by horizontal run and if we were
to use the two points (2, 150) and (6, 250) to calculate the slope we would get
slope =
rise
run
=
250 150
6 2
= 25.
To summarise:
The fact that the car is travelling at a constant velocity is reected in the fact that the
distance-time graph is a straight line. The velocity of the car is given by the slope of this
line.
2. 00 4. 00 6. 00 8. 00
100
200
300
(metres)
Di stance
t ime (seconds)
250 - 150 = 100
6 - 2 = 4
Time in seconds
Distance in metres
59 60 61 62 63 64 65
1020
1040
1060
1080
Mathematics Learning Centre, University of Sydney 3
Figure 2: Calculation of the velocity of the motorist is the same as the calculation of the slope
of the distance - time graph.
1.1.2 Non-constant velocity
Figure 3 shows the graph of a dierent motorists journey along a straight road. This
graph is not a straight line. The motorist is not travelling at a constant velocity.
Exercise 1.3
How far does the motorist travel in the two seconds from time t = 60 to time t = 62?
How far does the motorist travel in the two second interval from time t = 62 to t = 64?
Since the motorist travels at dierent velocities at dierent times, when we talk about
the velocity of the motorist in Figure 3 we need to specify the particular time that we
mean. Nevertheless we would still like somehow to interpret the velocity of the motorist
as the slope of the graph, even though the graph is curved and not a straight line.
Figure 3: Position versus time graph for a motorists journey.
Fer ti l i ser Useage (Tonnes)
slope = 50
slope = 2 5
1 0 0
150
2 0 0
1 2 3
4
5 0
Crop Yield (Tonnes)
Mathematics Learning Centre, University of Sydney 4
What do we mean by the slope of a curve? Suppose for example that we are interested
in the velocity of the motorist in Figure 3 at time t = 62. In Figure 3 we have drawn in
a dashed line. Notice that this line just grazes the curve at the point on the curve where
t = 62. The dashed line is in fact the tangent to the curve at that point. We will talk
more about tangents to curves in Section 2. For now you can think of the dashed line
like this: if you were going to draw a straight line through this point on the curve, and if
you wanted that straight line to look as much like the curve near that point as it possibly
could, this is the line that you would draw. This solves our problem about interpreting
the slope of the curve at this point on the curve.
The slope of the curve at the point on the curve where t = 62 is the slope of the tangent
to the curve at that point: that is the slope of the dashed line in Figure 3.
The velocity of the motorist at time t = 62 is the slope of the dashed line in that gure.
Of course if we were interested in the velocity of the motorist at time t = 64 then we
would draw the tangent to the curve at the point on the curve where t = 64 and we would
get a dierent slope. At dierent points on the curve we get dierent tangents having
dierent slopes. At dierent times the motorist is travelling at dierent velocities.
1.2 Other rates of change
The situation above described a car moving in one direction along a straight road away
from a xed point. Here, the word velocity describes how the distance changes with time.
Velocity is a rate of change. For these type of problems, the velocity corresponds to the
rate of change of distance with respect to time. Motion in general may not always be in
one direction or in a straight line. In this case we need to use more complex techniques.
Velocity is by no means the only rate of change that we might be interested in. Figure 4
shows a graph representing the yield a farmer gets from a crop depending on the amount
of fertiliser that the farmer uses.
The shape of this graph makes good sense. If no fertiliser is used then there is still some
crop yield (50 tonnes to be precise). As more fertiliser is used the crop yield increases,
Figure 4: Crop yield versus fertiliser useage for a hypothetical crop.
Mathematics Learning Centre, University of Sydney 5
as you would expect. Note though that at a certain point putting on more fertiliser does
not improve the yield of the crop, but in fact decreases it. The soil is becoming poisoned
by too much fertiliser. Eventually the use of too much fertiliser causes the crop to die
altogether and no yield is obtained.
On the graph the tangents to the curve corresponding to fertiliser usage of 1 tonne (the
dotted line) and of 1.5 tonnes (the dashed line) are drawn. The slope of these tangents
give the rate of change of crop yield with respect to fertiliser usage.
The slope of the dotted tangent is 50. This means that if fertiliser usage is increased from
1 tonne by a very small amount then the crop yield will increase by 50 times that small
change. For example an increase in fertiliser usage from 1 tonne (1000 kg) to 1005 kg
will increase the crop yield by approximately 50 5 = 250 kg. If we are using 1 tonne
of fertiliser then the rate of change of crop yield with respect to fertiliser useage is quite
high. On the other hand the slope of the dashed tangent is 25. The same increase (by 5
kg) in fertiliser useage from 1500 kg (1.5 tonnes) to 1505 kg will increase the crop yield
by about 25 5 = 125 kg.
B
A
Mathematics Learning Centre, University of Sydney 6
2 What is the derivative?
If you are not completely comfortable with the concept of a function and its graph then
you need to familiarise yourself with it before continuing. The booklet Functions published
by the Mathematics Learning Centre may help you.
In Section 1 we learnt that dierential calculus is about nding the rates of change of
related quantities. We also found that a rate of change can be thought of as the slope of
a tangent to a graph of a function. Therefore we can also say that:
Dierential calculus is about nding the slope of a tangent to the graph of a function, or
equivalently, dierential calculus is about nding the rate of change of one quantity with
respect to another quantity.
If we are going to go to all this trouble to nd out about the slope of a tangent to a graph,
we had better have a good idea of just what a tangent is.
2.1 Tangents
Look at the curve and straight line in Figure 5.
Figure 5: The line is tangent to the curve at point A but not at point B.
Imagine taking a very powerful magnifying glass and looking very closely at this gure near
the point A. Figure 6 shows two views of this curve at successively greater magnications.
The closer we look at the curve near the point A the straighter the curve appears to be.
The more we zoom in the more the curve begins to look like the straight line. This straight
line is called the tangent to the curve at the point A. If we want to draw a straight line
that most resembles the curve near the point A, the tangent line is the one that we would
draw. It is pretty clear from Figure 5 that no matter how closely we look at the curve
near the point B the curve is never going to look like the straight line we have drawn in
here. That line is tangent to the curve at A but not at B. The curve does have a tangent
at B, but it is not shown on Figure 5.
Note that it is not necessarily true that the tangent line only cuts the curve at one point
or that curve lies entirely on one side of the line. These properties hold for some special
curves like circles, but not for all curves, and certainly not for the one in Figure 5.
A
A
A (0.5, f (0.5))
C (1.5, f (1.5))
B (0.5, f (0.5))
x 1.5 0.5 1. 0 0.5
Mathematics Learning Centre, University of Sydney 7
Figure 6: Two close up views of the curve in Figure 5 near the point A. The closer we look
near the point A the more the curve looks like the tangent.
2.2 The derivative: the slope of a tangent to a graph
Terminology The slope of the tangent at the point (x, f(x)) on the graph of f is called
the derivative of f at x and is written f

(x).
Look at the graph of the function y = f(x) in Figure 7. Three dierent tangent lines have
Figure 7: Tangent lines to the graph of f(x) drawn at three dierent points on the graph.
been drawn on the graph, at A, B and C, corresponding to three dierent values of the
independent variable, x = 0.5, x = 0.5 and x = 1.5.
If we were to make careful measurements of the slopes of the three tangents shown we
would nd that f

(0.5) 2.75, f

(0.5) 1.25 and f

(1.5) 0.75. Here the symbol

means is approximately. We can only say approximately here because there is no
way that we can make completely accurate measurements from a graph, and no way
even to draw a completely accurate graph. However this graphical approach to nding
the approximate derivative is often very useful, and in some situations may be the only
technique that we have.
At dierent points on the graph we get dierent tangents having dierent slopes. The
slope of the tangent to the graph depends on where on the graph we draw the tangent.
Because we can specify a point on the graph by just giving its x coordinate (the other
Mathematics Learning Centre, University of Sydney 8
coordinate is then f(x)), we can say that the slope of the tangent to the graph of a function
depends on the value of the independent variable x, or the value of f

(x) depends on x.
In other words, f

is a function of x.
Terminology The function f

is called the derivative of f.

Terminology The process of nding the derivative is called dierentiation.
The derivative of a a function f is another function, called f

, which tells us about the

slopes of tangents to the graph of f. Because there are several dierent ways of writing
functions, there are several dierent ways of writing the derivative of a function. Most of
the ways that are commonly used are expressed in the following table.
Function Derivative
f(x) f

(x) or
df(x)
dx
f f

or
df
dx
y y

or
dy
dx
y(x) y

(x) or
dy(x)
dx
Exercise 2.1 (You will nd this exercise easier to do if you use graph paper.)
Draw a careful graph of the function f(x) = x
2
. Draw the tangents at the points x = 1,
x = 0 and x = 0.5. Find the slopes of these lines by picking two points on them and
using the formula
slope =
y
2
y
1
x
2
x
1
.
These slopes are the (approximate) values of f

(1), f

(0) and f

(0.5) respectively.
Exercise 2.2
Repeat Exercise 2.1 with the function f(x) = x
3
.
Mathematics Learning Centre, University of Sydney 9
3 How do we nd derivatives (in practice)?
Dierential calculus is a procedure for nding the exact derivative directly from the for-
mula of the function, without having to use graphical methods. In practise we use a few
rules that tell us how to nd the derivative of almost any function that we are likely to
encounter. In this section we will introduce these rules to you, show you what they mean
and how to use them.
Warning! To follow the rest of these notes you will need feel comfortable manipulating
expressions containing indices. If you nd that you need to revise this topic you may nd
the Mathematics Learning Centre publication Exponents and Logarithms helpful.
3.1 Derivatives of constant functions and powers
Perhaps the simplest functions in mathematics are the constant functions and the func-
tions of the form x
n
.
Rule 1 If k is a constant then
d
dx
k = 0.
Rule 2 If n is any number then
d
dx
x
n
= nx
n1
.
Rule 1 at least makes sense. The graph of a constant function is a horizontal line and a
horizontal line has slope zero. The derivative measures the slope of the tangent, and so
the derivative is zero.
How you approach Rule 2 is up to you. You certainly need to know it and be able to use
it. However we have given no justication for why Rule 2 works! In fact in these notes we
will give little justication for any of the rules of dierentiation that are presented. We
will show you how to apply these rules and what you can do with them, but we will not
make any attempt to prove any of them.
Examples If f(x) = x
7
then f

(x) = 7x
6
.
If y = x
0.5
then
dy
dx
= 0.5x
1.5
.
d
dx
x
3
= 3x
4
.
If g(x) = 3.2 then g

(x) = 0.
If f(t) = t
1
2
then f

(t) =
1
2
t

1
2
.
If h(u) = 13.29 then h

(u) = 0.
0. 50 1. 00 1. 50 - 0. 50 - 1. 00
1. 00
2. 00
3. 00
slope = f ' (-0.5) = - 1
( 1, 1 )
( -0.5, 0.25)
slope = f ' (1) = 2
Mathematics Learning Centre, University of Sydney 10
In the examples above we have used Rules 1 and 2 to calculate the derivatives of many
simple functions. However we must not lose sight of what it is that we are calculating
here. The derivative gives the slope of the tangent to the graph of the function.
For example, if f(x) = x
2
then f

(x) = 2x. To nd the slope of the tangent to the graph

of x
2
at x = 1 we substitute x = 1 into the derivative. The slope is f

(1) = 2 1 = 2.
Similarly the slope of the tangent to the graph of x
2
at x = 0.5 is found by substituting
x = 0.5 into the derivative. The slope is f

(0.5) = 2 0.5 = 1. This is illustrated

in Figure 8.
Figure 8: Slopes of tangents to the graph of y = x
2
.
Example
Find the slope of the tangent to the graph of the function g(t) = t
4
at the point on the
graph where t = 2.
Solution
The derivative is g

(t) = 4t
3
, and so the slope of the tangent line at t = 2 is g

(2) =
4 (2)
3
= 32.
Example
Find the equation of the line tangent to the graph of y = f(x) = x
1
2
at the point x = 4.
Solution
f(4) = 4
1
2
=

4 = 2, so the coordinates of the point on the graph are (4, 2). The
derivative is
f

(x) =
x

1
2
2
=
1
2

x
and so the slope of the tangent line at x = 4 is f

(4) =
1
4
. We therefore know the slope of
the line and we know one point through which the line passes.
Any non vertical line has equation of the form y = mx+b where m is the slope and b the
vertical intercept.
Mathematics Learning Centre, University of Sydney 11
In this case the slope is
1
4
, so m =
1
4
, and the equation is y =
x
4
+ b. Because the line
passes through the point (4, 2) we know that y = 2 when x = 4.
Substituting we get 2 =
4
4
+ b, so that b = 1. The equation is therefore y =
x
4
+ 1.
Notice that in the examples above the independent variable is not always called x. We
have also used u and t, and in fact we can and will use many dierent letters for the
independent variable. Notice also that we might not stick to the symbol f to stand
for function. Many other symbols are used. Some of the common ones are g and h.
Throughout this booklet we will use a variety of symbols for functions and variables to
get you used to the fact that our choice of symbols makes no dierence to the ideas that
we are introducing. On the other hand, we can make life easier for ourselves if we make
sensible choices of symbols. For example if we were discussing the revenue obtained by
a manufacturer who sells articles for a certain price it might be sensible for us to choose
the symbol p to mean price, and r to mean revenue, and to write r(p) to express the
fact that the revenue is a function of the price. In this way the symbols we have chosen
remind us of their meaning, much more than if we had chosen x to represent price and f
to represent revenue and written f(x). On the other hand, because the symbol d has a
special use in calculus, to express the derivative
df(x)
dx
, we almost never use d for any other
purpose. For this reason you will often see the letter s used to represent diSplacement.
We now know how to dierentiate any function that is a power of the variable. Examples
are functions like x
3
and t
1.3
. You will come across functions that do not at rst appear
to be a power of the variable, but can be rewritten in this form. One of the simplest
examples is the function
f(t) =

t,
which can also be written in the form
f(t) = t
1
2
.
The derivative is then
f

(t) =
t

1
2
2
=
1
2

t
.
Similarly, if
h(s) =
1
s
= s
1
then
h

(s) = s
2
=
1
s
2
.
Examples
If f(x) =
1
3

x
= x

1
3
then f

(x) =
1
3
x

4
3
.
If y =
1
x

x
= x

3
2
then
dy
dx
=
3
2
x

5
2
.
Mathematics Learning Centre, University of Sydney 12
Exercises 3.1
Dierentiate the following functions:
a. f(x) = x
4
b. y = x
7
c. f(u) = u
2.3
d. f(t) = t

1
3
e. f(t) = t
22
7
f. g(z) = z

3
2
g. y = t
3.8
h. z = x
3
7
Exercise 3.2
Express the following as powers and then dierentiate:
a.
1
x
2
b. t

t c.
3

x
d.
1
x
2

x
e.
1
x
4

x
f.
s
3

s
3

s
g.
1
u
3
h.
t
t
2

t
i. x
1
2

x
x
Exercise 3.3
Find the equation of the line tangent to the graph of y =
3

x when x = 8.
3.2 Adding, subtracting, and multiplying by a constant
So far we know how to dierentiate powers of the independent variable. Many of the
functions that you will encounter are made up in simple ways from powers. For example,
a function like 3x
2
is just a constant multiple of x
2
. However neither Rule 1 nor Rule 2
tell us how to dierentiate 3x
2
. Nor do they tell us how to dierentiate something like
x
2
+ x
3
or x
2
x
3
.
Rules 3 and 4 specify how to dierentiate combinations of functions that are formed by
multiplying by constants, or by adding or subtracting functions.
Rule 3 If f(x) = cg(x), where c is a constant, then f

(x) = cg

(x).
Rule 4 If f(x) = g(x) h(x) then f

(x) = g

(x) h

(x).
Examples If f(x) = 3x
2
then f

(x) = 3
d
dx
x
2
= 6x.
If g(t) = 3t
2
+ 2t
2
then g

(t) =
d
dt
3t
2
+
d
dt
2t
2
= 6t 4t
3
.
If y =
3

x
2x
3

x = 3x

1
2
2x
4
3
then
dy
dx
=
3
2
x

3
2

8
3
x
1
3
.
If y = 0.3x
0.4
then
dy
dx
= 0.12x
1.4
.
d
dx
2x
0.3
= 0.6x
0.7
.
Mathematics Learning Centre, University of Sydney 13
Warning! Although Rule 4 tells us that
d
dx
(f(x) g(x)) = f

(x) g

(x), the same is not

true for multiplication or division. To dierentiate f(x) g(x) or f(x) g(x) we cannot
simply nd f

(x) and g

(x) and multiply or divide them. Be careful of this! The methods

of dierentiating products of functions or quotients of functions are discussed in Sections
3.3 and 3.4.
Exercise 3.4
Dierentiate the following functions:
a. f(x) = 5x
2
2

x b. y = 2x
7
+
3
x
2
c. f(t) = 2.5t
2.3
+
t

t
d. h(z) = z

1
3
+ 5z e. f(u) = u
5
3
3u
7
f. g(z) = 8z
2

5
z
g. y = 5t
8
+
t

t
h. z = 4x
1
7
+ 2x

1
2
3.3 The product rule
Another way of combining functions to make new functions is by multiplying them to-
gether, or in other words by forming products. The product rule tells us how to dieren-
tiate functions like this.
Rule 5 (The product rule) If f(x) = u(x)v(x) then
f

(x) = u(x)v

(x) + u

(x)v(x).
Examples
If y = (x + 2)(x
2
+ 3) then y

= (x + 2)2x + 1(x
2
+ 3).
If f(x) =

x(x
3
3x
2
+ 7) then f

(x) =

x(3x
2
6x) +
1
2
x

1
2
(x
3
3x
2
+ 7).
If z = (t
2
+ 3)(

t + t
3
) then
dz
dt
= (t
2
+ 3)(
1
2
t

1
2
+ 3t
2
) + 2t(

t + t
3
).
Mathematics Learning Centre, University of Sydney 14
Exercise 3.5
Use the product rule to dierentiate the functions below:
a. f(x) = (4x
3
+ 2)(1 3x)
b. g(x) = (x
2
+ x + 2)(x
2
+ 1)
c. h(x) = (3x
3
2x
2
+ 8x 5)(x
2
2x + 4)
d. f(s) = (1
1
2
s
2
)(3s + 5)
e. g(t) = (

t +
1
t
)(2t 1)
f. h(y) = (2

y + y
2
)(1 3y
2
)
Exercise 3.6
If r = (t +
1
t
)(t
2
2t + 1), nd the rate of change of r with respect to t when t = 2.
Exercise 3.7
Find the slope of the tangent to the curve y = (x
2
2x + 1)(3x
3
5x
2
+ 2) at x = 2.
3.4 The Quotient Rule
This rule allows us to dierentiate functions which are formed by dividing one function
by another, ie by forming quotients of functions. An example is such as
f(x) =
2x + 3
3x 5
.
Rule 6 (The quotient rule)
f(x) =
u(x)
v(x)
f

(x) =
v(x)u

(x) u(x)v

(x)
[v(x)]
2
=
vu

uv

v
2
.
Warning! Because of the minus sign in the numerator (ie in the top line) it is important
to get the terms in the numerator in the correct order. This is often a source of mistakes,
so be careful. Decide on your own way of remembering the correct order of the terms.
Mathematics Learning Centre, University of Sydney 15
Examples
If y =
2x
2
+ 3x
x
3
+ 1
, then
dy
dx
=
(x
3
+ 1)(4x + 3) (2x
2
+ 3x)3x
2
(x
3
+ 1)
2
.
If g(t) =
t
2
+ 3t + 1

t + 1
then g

(t) =
(

t + 1)(2t + 3) (t
2
+ 3t + 1)(
1
2
t

1
2
)
(

t + 1)
2
.
Exercise 3.8
Use the Quotient Rule to nd derivatives for the following functions:
a. f(x) =
x 1
x + 1
b. g(x) =
2x + 3
3x 2
c. h(x) =
x
2
+ 2
x
2
+ 5
d. f(t) =
2t
1 + 2t
2
e. f(s) =
1 +

s
1

s
f. h(x) =
x
2
1
x
3
+ 4
g. f(u) =
u
3
+ u 4
3u
4
+ 5
h. g(t) =
t(t + 6)
t
2
+ 3t + 1
3.5 The composite function rule (also known as the chain rule)
Have a look at the function f(x) = (x
2
+ 1)
17
. We can think of this function as being
the result of combining two functions. If g(x) = x
2
+ 1 and h(t) = t
17
then the result of
substituting g(x) into the function h is
h(g(x)) = (g(x))
17
= (x
2
+ 1)
17
.
Another way of representing this would be with a diagram like
x
g
x
2
+ 1
h
(x
2
+ 1)
17
.
We start o with x. The function g takes x to x
2
+ 1, and the function h then takes
x
2
+ 1 to (x
2
+ 1)
17
. Combining two (or more) functions like this is called composing the
functions, and the resulting function is called a composite function. For a more detailed
discussion of composite functions you might wish to refer to the Mathematics Learning
Centre booklet Functions.
Using the rules that we have introduced so far, the only way to dierentiate the function
f(x) = (x
2
+1)
17
would involve expanding the expression and then dierentiating. If the
function was (x
2
+1)
2
= (x
2
+1)(x
2
+1) then it would not take too long to expand these
two sets of brackets. But to expand the seventeen sets of brackets involved in the function
f(x) = (x
2
+1)
17
(or even to expand using the binomial theorem) would take a long time.
The composite function rule shows us a quicker way.
Rule 7 (The composite function rule (also known as the chain rule))
If f(x) = h(g(x)) then f

(x) = h

(g(x)) g

(x).
Mathematics Learning Centre, University of Sydney 16
In words: dierentiate the outside function, and then multiply by the derivative of the
inside function.
To apply this to f(x) = (x
2
+ 1)
17
, the outside function is h() = ()
17
and its derivative
is 17()
16
. The inside function is g(x) = x
2
+ 1 which has derivative 2x. The composite
function rule tells us that f

(x) = 17(x
2
+ 1)
16
2x.
As another example let us dierentiate the function 1/(z
3
+ 4z
2
3z 3)
6
. This can be
rewritten as (z
3
+ 4z
2
3z 3)
6
. The outside function is ()
6
which has derivative
6()
7
. The inside function is z
3
+4z
2
3z 3 with derivative 3z
2
+8z 3. The chain
rule says that
d
dz
(z
3
+ 4z
2
3z 3)
6
= 6(z
3
+ 4z
2
3z 3)
7
(3z
2
+ 8z 3).
There is another way of writing down, and hence remembering, the composite function
rule.
Rule 7 (The composite function rule (alternative formulation))
If y is a function of u and u is a function of x then
dy
dx
=
dy
du

du
dx
.
This makes the rule very easy to remember. The expressions
dy
du
and
du
dx
are not really
fractions but rather they stand for the derivative of a function with respect to a variable.
However for the purposes of remembering the chain rule we can think of them as fractions,
so that the du cancels from the top and the bottom, leaving just
dy
dx
.
To use this formulation of the rule in the examples above, to dierentiate y = (x
2
+ 1)
17
put u = x
2
+ 1, so that y = u
17
. The alternative formulation of the chain rules says that
dy
dx
=
dy
du

du
dx
= 17u
16
2x
= 17(x
2
+ 1)
16
2x.
which is the same result as before. Again, if y = (z
3
+ 4z
2
3z 3)
6
then set u = z
3
+ 4z
2
3z 3 so that y = u
6
and
dy
dx
=
dy
du

du
dx
= 6u
7
(3z
2
+ 8z 3).
You select the formulation of the chain rule that you nd easiest to use. They are equiv-
alent.
Mathematics Learning Centre, University of Sydney 17
Example
Dierentiate (3x
2
5)
3
.
Solution
The rst step is always to recognise that we are dealing with a composite function and
then to split up the composite function into its components. In this case the outside
function is ()
3
which has derivative 3()
2
, and the inside function is 3x
2
5 which has
derivative 6x, and so by the composite function rule,
d(3x
2
5)
3
dx
= 3(3x
2
5)
2
6x = 18x(3x
2
5)
2
.
Alternatively we could rst let u = 3x
2
5 and then y = u
3
. So
dy
dx
=
dy
du

du
dx
= 3u
2
6x = 18x(3x
2
5)
2
.
Example
Find
dy
dx
if y =

x
2
+ 1.
Solution
The outside function is

= ()
1
2
which has derivative
1
2
()

1
2
, and the inside function is
x
2
+ 1 so that
y

=
1
2
(x
2
+ 1)

1
2
2x.
Alternatively, if u = x
2
+ 1, we have y =

u = u
1
2
. So
dy
dx
=
1
2
u

1
2
2x =
1
2
(x
2
+ 1)

1
2
2x.
Exercise 3.9
Dierentiate the following functions using the composite function rule.
a. (2x + 3)
2
b. (x
2
+ 2x + 1)
12
c. (3 x)
21
d. (x
3
1)
5
e. f(t) =

t
2
5t + 7 f. g(z) =
1

2z
4
g. y = (t
3

t)
3.8
h. z = (x +
1
x
)
3
7
Mathematics Learning Centre, University of Sydney 18
Exercise 3.10
Dierentiate the functions below. You will need to use both the composite function rule
and the product or quotient rule.
a. (x + 2)(x + 3)
2
b. (2x 1)
2
(x + 3)
3
c. x

(1 x)
d. x
1
3
(1 x)
2
3
e.
x

1 x
2
3.6 Derivatives of exponential and logarithmic functions
If you are not familiar with exponential and logarithmic functions you may wish to consult
the booklet Exponents and Logarithms which is available from the Mathematics Learning
Centre.
You may have seen that there are two notations popularly used for natural logarithms,
log
e
and ln. These are just two dierent ways of writing exactly the same thing, so that
log
e
x ln x. In this booklet we will use both these notations.
The basic results are:
d
dx
e
x
= e
x
d
dx
(log
e
x) =
1
x
.
We can use these results and the rules that we have learnt already to dierentiate functions
which involve exponentials or logarithms.
Example
Dierentiate log
e
(x
2
+ 3x + 1).
Solution
We solve this by using the chain rule and our knowledge of the derivative of log
e
x.
d
dx
log
e
(x
2
+ 3x + 1) =
d
dx
(log
e
u) (where u = x
2
+ 3x + 1)
=
d
du
(log
e
u)
du
dx
(by the chain rule)
=
1
u

du
dx
=
1
x
2
+ 3x + 1

d
dx
(x
2
+ 3x + 1)
=
1
x
2
+ 3x + 1
(2x + 3)
=
2x + 3
x
2
+ 3x + 1
.
Mathematics Learning Centre, University of Sydney 19
Example
Find
d
dx
(e
3x
2
).
Solution
This is an application of the chain rule together with our knowledge of the derivative of
e
x
.
d
dx
(e
3x
2
) =
de
u
dx
where u = 3x
2
=
de
u
du

du
dx
by the chain rule
= e
u

du
dx
= e
3x
2

d
dx
(3x
2
)
= 6xe
3x
2
.
Example
Find
d
dx
(e
x
3
+2x
).
Solution
Again, we use our knowledge of the derivative of e
x
together with the chain rule.
d
dx
(e
x
3
+2x
) =
de
u
dx
(where u = x
3
+ 2x)
= e
u

du
dx
(by the chain rule)
= e
x
3
+2x

d
dx
(x
3
+ 2x)
= (3x
2
+ 2) e
x
3
+2x
.
Example
Dierentiate ln (2x
3
+ 5x
2
3).
Solution
We solve this by using the chain rule and our knowledge of the derivative of lnx.
d
dx
ln (2x
3
+ 5x
2
3) =
d ln u
dx
(where u = (2x
3
+ 5x
2
3)
=
d ln u
du

du
dx
(by the chain rule)
=
1
u

du
dx
=
1
2x
3
+ 5x
2
3

d
dx
(2x
3
+ 5x
2
3)
=
1
2x
3
+ 5x
2
3
(6x
2
+ 10x)
=
6x
2
+ 10x
2x
3
+ 5x
2
3
.
Mathematics Learning Centre, University of Sydney 20
There are two shortcuts to dierentiating functions involving exponents and logarithms.
The four examples above gave
d
dx
(log
e
(x
2
+ 3x + 1)) =
2x + 3
x
2
+ 3x + 1
d
dx
(e
3x
2
) = 6xe
3x
2
d
dx
(e
x
3
+2x
) = (3x
2
+ 2)e
3x
2
d
dx
(log
e
(2x
3
+ 5x
2
3)) =
6x
2
+ 10x
2x
3
+ 5x
2
3
.
These examples suggest the general rules
d
dx
(e
f(x)
) = f

(x)e
f(x)
d
dx
(ln f(x)) =
f

(x)
f(x)
.
These rules arise from the chain rule and the fact that
de
x
dx
= e
x
and
d ln x
dx
=
1
x
. They can
speed up the process of dierentiation but it is not necessary that you remember them.
If you forget, just use the chain rule as in the examples above.
Exercise 3.11
Dierentiate the following functions.
a. f(x) = ln(2x
3
) b. f(x) = e
x
7
c. f(x) = ln(11x
7
)
d. f(x) = e
x
2
+x
3
e. f(x) = log
e
(7x
2
) f. f(x) = e
x
g. f(x) = ln(e
x
+ x
3
) h. f(x) = ln(e
x
x
3
) i. f(x) = ln

x
2
+ 1
x
3
x

Mathematics Learning Centre, University of Sydney 21

3.7 Derivatives of trigonometric functions
To understand this section properly you will need to know about trigonometric functions.
The Mathematics Learning Centre booklet Introduction to Trigonometric Functions may
be of use to you.
There are only two basic rules for dierentiating trigonometric functions:
d
dx
sin x = cos x
d
dx
cos x = sin x.
For dierentiating all trigonometric functions these are the only two things that we need
to remember.
Of course all the rules that we have already learnt still work with the trigonometric
functions. Thus we can use the product, quotient and chain rules to dierentiate functions
that are combinations of the trigonometric functions.
For example, tan x =
sin x
cos x
and so we can use the quotient rule to calculate the derivative.
f(x) = tan x =
sin x
cos x
,
f

(x) =
cos x.(cos x) sin x.(sin x)
(cos x)
2
=
cos
2
x + sin
2
x
cos x
=
1
cos
2
x
(since cos
2
x + sin
2
x = 1)
= sec
2
x
Note also that
cos
2
x + sin
2
x
cos
2
x
=
cos
2
x
cos
2
x
+
sin
2
x
cos
2
x
= 1 + tan
2
x
so it is also true that
d
dx
tan x = sec
2
x = 1 + tan
2
x.
Mathematics Learning Centre, University of Sydney 22
Example
Dierentiate f(x) = sin
2
x.
Solution
f(x) = sin
2
x is just another way of writing f(x) = (sin x)
2
. This is a composite function,
with the outside function being ()
2
and the inside function being sin x.
By the chain rule, f

(x) = 2(sin x)
1
cos x = 2 sin x cos x. Alternatively using the other
method and setting u = sin x we get f(x) = u
2
and
df(x)
dx
=
df(x)
du

du
dx
= 2u
du
dx
= 2 sin x cos x.
Example
Dierentiate g(z) = cos(3z
2
+ 2z + 1).
Solution
Again you should recognise this as a composite function, with the outside function being
cos() and the inside function being 3z
2
+ 2z + 1. By the chain rule g

(z) = sin(3z
2
+
2z + 1) (6z + 2) = (6z + 2) sin(3z
2
+ 2z + 1).
Example
Dierentiate f(t) =
e
t
sin t
.
Solution
By the quotient rule
f

(t) =
e
t
sin t e
t
cos t
sin
2
t
=
e
t
(sin t cos t)
sin
2
t
.
Example
Use the quotient rule or the composite function rule to nd the derivatives of cot x, sec x,
and cosec x.
Solution
These functions are dened as follows:
cot x =
cos x
sin x
sec x =
1
cos x
csc x =
1
sin x
.
Mathematics Learning Centre, University of Sydney 23
By the quotient rule
d cot x
dx
=
sin
2
x cos
2
x
sin
2
x
=
1
sin
2
x
.
Using the composite function rule
d sec x
dx
=
d(cos x)
1
dx
= (cos x)
2
(sin x) =
sin x
cos
2
x
.
d csc x
dx
=
d(sin x)
1
dx
= (sin x)
2
cos x =
cos x
sin
2
x
.
Exercise 3.12
Dierentiate the following:
a. cos 3x b. sin(4x + 5) c. sin
3
x d. sin x cos x e. x
2
sin x
f. cos(x
2
+ 1) g.
sin x
x
h. sin
1
x
i. tan(

x) j.
1
x
sin
1
x
Mathematics Learning Centre, University of Sydney 24
4 What is dierential calculus used for?
4.1 Introduction
The development of mathematics stands as one of the most important achievements of
humanity, and the development of the calculus, both the dierential calculus and integral
calculus is one of most important achievements in mathematics. The practical applications
of dierential calculus are so wide ranging that it would be impossible to mention them
all here. Suce to say that dierential calculus is an indispensable tool in every branch
of science and engineering.
In elementary mathematics there are two main applications of dierential calculus. One is
to help in sketching curves, and the other is in optimisation problems. For a treatment of
the uses of calculus in curve sketching see the Mathematics Learning Centre publication
Curve Sketching. In this section we will give a brief introduction to how dierential
calculus is used in optimisation problems.
4.2 Optimisation problems
There are many practical situations in which we would like to make a quantity as small
as we possibly can or as large as we possibly can. For example, a manufacturer of bicycles
trying to decide how much to charge for a model of bicycle would think that if he charges
too little for the bicycles then he will probably sell a lot of bicycles but that he wont
make much prot because the price is too low, and that if he charges too much for the
bicycle then he wont make much prot because not many people will buy his bicycles.
The manufacturer would like to nd just the right price to charge to maximise his prot.
Similarly a farmer might realise that if she uses too little fertiliser on her crops then her
yield will be very low, and if she uses too much fertiliser then she will poison the soil and
her yield will be low. The farmer might like to know just how much fertiliser to use to
maximise the crop yield. A manufacturer of sheet metal cans that are meant to hold one
litre of liquid might like to know just what shape to make the can so that the amount of
sheet metal that is used is a minimum. These are all examples of optimisation problems.
If we were to draw a graph of the prot versus price for the bicycle manufacturer mentioned
above then nding the maximum prot is equivalent to nding the highest point on the
graph. Similarly a minimisation problem may be thought of geometrically as nding the
lowest point on the graph of a funcion.
4.2.1 Stationary points - the idea behind optimisation
As a thought experiment, let us imagine that a person wearing a blindfold is walking
along a road, and that the road has a hill on it. Let us imagine also that the blindfolded
person is searching for the highest point on the road. How would this person be able to
decide when they were at the top of the hill? Well, while they were walking uphill the
person would know that this wasnt the top of the hill - because they are still going up!
And of course while they are walking downhill they would know that they are not at the
top of the hill because they are going down. In other words, while they are on a sloping
bit of the road the blindfolded person would know that this is not the top of the hill.
Mathematics Learning Centre, University of Sydney 25
Right at the top of the hill there would be a little bit of level road. The slope at the
top of the hill would be zero! Without even being able to see the road, the blindfolded
person would know that they could not possibly be at the top of the hill unless they were
standing on level ground. The same idea would apply if the road had a valley in it, and
the person was searching for the lowest piece of road. Right at the lowest point of the
valley the slope of the road would be zero.
So if we are searching for the highest (or lowest) point on a road, of all the possible
places we only have to consider those places where the road has slope zero. This is the
idea behind using calculus for optimisation. If we are searching for the highest or lowest
points on the graph of a function we have to look for those places where the graph has
slope zero. These points are called stationary points.
Denition For a function y = f(x) the points on the graph where the graph has zero
slope are called stationary points. In other words stationary points are where f

(x) = 0.
To nd the stationary points of a function we dierentiate, set the derivative equal to
zero and solve the equation.
Example Find the stationary points of the function f(x) = 2x
3
+ 3x
2
12x + 17.
Solution f

(x) = 6x
2
+ 6x 12. Setting f

(x) = 0 and solving we obtain

6x
2
+ 6x 12 = 0
x
2
+ x 2 = 0
(x 1)(x + 2) = 0
x = 1, 2.
This gives us the values of x for which the function f is stationary. The corresponding
values of the function are found by substituting 1 and 2 into the function.
They are f(1) = 2 1
3
+ 3 1
2
12 1 + 17 = 10 and
f(2) = 2(2)
3
+3(2)
2
12(2) +17 = 37. The stationary points are therefore
(1, 10) and (2, 37).
Example Find the stationary points of the function g(t) = e
t
2
.
Solution Dierentiating and setting the derivative equal to zero we obtain the equa-
tion g

(t) = 2te
t
2
= 0. Since e
t
2
is never zero, the only solution to this equation is
where 2t = 0, ie t = 0. Substituting into the formula for g we obtain the function value
g(0) = e
0
2
= 1. Thus the stationary point is (0, 1).
4.2.2 Types of stationary points
In our thought experiment above we mentioned two types of stationary points: one was
the top of the hill and the other was the bottom of the valley. The top of the hill is called
a local maximum, and the bottom of the valley is called a local minimum. The word
local conveys the fact that at the top of the hill the blindfolded person is not necessarily
at the highest point in the world, but merely at the highest point in the local vicinity.
Sometimes you will see local maxima and local minima called relative maxima and relative
local maximum
local minimum
s t ati onary point of infl ection
Mathematics Learning Centre, University of Sydney 26
Figure 9: Graph of a function showing a local maximum and a local minimum.
minima. Figure 9 shows a function with a local maximum and a local minimum. Note
that at each of these points the slope of the curve is zero.
Local maxima and local minima are not the only types of stationary points. There is a
third kind. Figure 10 shows a stationary point that is neither a local maximum nor a
local minimum. This type of stationary point is called a stationary point of inection.
Dont worry about why it is given this name. That is beyond the scope of this booklet.
You just need to be aware of the fact that stationary points exist that are neither local
maxima nor local minima.
Figure 10: Graph of a function showing a stationary point of inection.
Let us now return to the rst of the examples in the previous section. We found that
the function f(x) = 2x
3
+ 3x
2
12x + 17 had stationary points at (1, 10) and (2, 37).
What type of stationary points are they? At the moment you probably have no idea just
what the graph of f(x) = 2x
3
+ 3x
2
12x + 17 looks like. How can you tell what type
of stationary points these are? If you could see the graph you would be able to tell what
types of stationary points they were, but it takes a lot of work to draw the graph of a
function. What we need is a way of testing a stationary point that will tell us whether we
have found a local maximium, a local minimum or neither (in other words a stationary
point of inection) without drawing the graph. There are several ways of doing this, but
in this booklet we will look at only one of them. This is called the rst derivative test.
f ' < 0
f ' < 0
f ' = 0
f' > 0
local maximum
local minimum
f ' > 0
f' = 0
f ' > 0
f ' > 0
f ' = 0
s t ati onary point of infl ection
Mathematics Learning Centre, University of Sydney 27
Really we are pretty much in the shoes of the blindfolded person now. We cant see the
whole graph, so how can we tell what type of stationary point we have got? Imagine the
blindfolded person standing on a piece of level ground, and wanting to know whether this
was the top of a hill (a local maximum), the bottom of a valley (a local minimum) or
neither (a stationary point of inection). One thing the person could do is take a step
backwards from the level spot. Which way does the ground slope here? And then take a
step forwards from the level spot. Which way does the ground slope here? If the person
took a step backward and found that the ground in front of them sloped up, then returned
to the original position and took a step forward and the ground sloped down, then the
level spot must have been the top of the hill. On the other hand if the person took a step
backward and the ground sloped down, and a step forward and the ground sloped up,
then the level spot must have been the bottom of a valley. You should be able to gure
out what the blindfolded person would nd for a stationary point of inection. This idea
is the basis of the rst derivative test.
The rst derivative test
If x
0
is a stationary point of the function f, so that f

(x
0
) = 0 then to nd out the nature
of the stationary point check the sign (ie positive or negative) of f

just either side of x

0
.
If f

< 0 to the left of x

0
(ie for x < x
0
) and f

> 0 to the right of x

0
(ie for x > x
0
) then
x
0
is a local minimum. If f

> 0 to the left of x

0
and f

< 0 to the right of x

0
then x
0
is a
local maximum. Otherwise x
0
is a stationary point of inection. Have a look at Figures
11 and 12.
Figure 11: First derivative test for local minima and local maxima.
Figure 12: First derivative test for point of inection.
1. 00 2. 00 3. 00 4. 00 5. 00 6. 00 7. 00 8. 00
5. 0
10. 0
15. 0
20. 0
25. 0
30. 0
35. 0
l ocal maximum
t he minimum f or the r egion unde r consider at i on
t he maximum for the regi on under consi de rat ion
Mathematics Learning Centre, University of Sydney 28
4.2.3 Optimisation
Okay, now we are in a position to be able to do some optimisation problems. To maximise
a function f(x) in a certain region of the x values we are looking for the greatest value
that f(x) can possibly take for x in the region that we are interested in. This may or
may not be at a stationary point. Figure 13 illustrates this. In this gure, we are looking
for the maximum and minimum of the function in the region 2 x 7. In this region
there are two stationary points, one a local maximum and one a local minimum. However
notice that the maximum value of the function does not occur at the local maximum, but
at the endpoint of the region, ie where x = 7. This point is not at the top of the hill, so it
is not a stationary point, but it is still the maximum value of the function for 2 x 7
because we are ignoring any x which is bigger than 7. On the other hand, in this case the
minimum value of this function for 2 x 7 is found at a stationary point. Now we are
in a position to tell you exactly how to nd the maximum or minimum of a function.
Figure 13: The maximum is found at the endpoint of the region under consideration, and not at
a stationary point. The minimum is found inside the region under consideration at a stationary
point
The location of maxima and minima
A function f(x) may or may not have a maximum or minimum value in a particular region
of x values. However, if they do exist the maximum and the minimum values must occur
at one of three places:
1. At the endpoints (if they exist) of the region under consideration.
2. Inside the region at a stationary point.
3. Inside the region at a point where the derivative does not exist.
Notes
1. It is easy to nd an example of a function which has no maximum or minimum in
a particular region. For example the function f(x) = x has neither a maximum nor a
Mathematics Learning Centre, University of Sydney 29
minimum value for < x < . Its graph simply keeps increasing as the values of x
increase. Referring to Point 1 above, if for example the region under consideration was
< x < then this region has no endpoints. As another example, the region x 1
has only one endpoint, x = 1.
2. A note about Point 3 above: in this booklet we will not treat points where the derivative
does not exist. However you should be aware that there may be such points, and that
the maximum or minimum may be found at one. For more information consult a more
comprehensive calculus text.
Now that we know exactly where the maxima or minima can occur, we can give a proce-
dure for nding them.
Procedure for nding the maximum or minimum values of a function.
1. Find the endpoints of the region under consideration (if there are any).
2. Find all the stationary points in the region.
3. Find all points in the region where the derivative does not exist.
4. Substitute each of these into the function and see which gives the greatest (or
smallest) function value.
Example Find the minimum value and the maximum value of the function f(x) = x
2
e
x
for 4 x 1.
Solution We will follow the procedure outlined above. The endpoints are 4 and 1.
Dierentiating we obtain f

(x) = x
2
e
x
+2xe
x
= x(x+2)e
x
. Setting f

(x) = 0 and solving

we get stationary points at x = 0 and x = 2. There are no points where the derivative
does not exist. Therefore the maximum and minimum values will be found at one of the
points x = 4, 2, 0, 1. Substituting we obtain f(4) 0.29, f(2) 0.54, f(0) = 0
and f(1) = e 2.7. therefore the maximum value occurs at x = 1 and is equal to e, and
the minimum value occurs at x = 0 and is 0.
Example Find the maximum and minimum values of the function g(t) =
1
3
t
3
t + 2 for
0 t 3.
Solution The endpoints are t = 0 and t = 3. Dierentiating and equating to zero
we get g

(t) = t
2
1 = (t 1)(t + 1) = 0 so the stationary points are at t = 1, 1. Since
1 is not in the region, the possible locations of the maximum and the minimum are
t = 0, 1, 3. Substituting into g we obtain g(0) = 2, g(1) =
4
3
and g(3) = 8. The maximum
is therefore g(3) = 8 and the minimum is g(1) =
4
3
.
Example A farmer is to make a rectangular paddock. The farmer has 100 metres of fenc-
ing and wants to make the rectangle that will enclose the greatest area. What dimensions
should the rectangle be?
Solution There are many rectangular paddocks that can be made with 100 metres of
fencing. If we call one side of the rectangle x, then because the perimeter is 100, the other
side of the rectangle is 50 x. The area of the paddock is then A(x) = x(50 x). We
must maximise the function A(x) for 0 x 50 (since the sides of the rectangle cannot
Mathematics Learning Centre, University of Sydney 30
have negative length). Now
dA
dx
= 50 2x which is zero when x = 25. Thus x = 25 is
the only stationary point and the maximum is found at one of the points x = 0, 25, 50.
Substituting these values into A(x) we nd that the maximum occurs when x = 25. The
rectangular paddock with the maximum area is a square.
Exercise 4.1 Find the maximum and the minimum of the function f(x) = x
4
2x
2
for
1 x 2
Exercise 4.2 Maximise the function g(t) = te
t
2
for 2 < t < 2.
Exercise 4.3 Find the minimum value of h(u) = 2u
3
+ 3u
2
12u + 5 in the region
3 u 2.
Exercise 4.4 A farmer wishes to make a rectangular chicken run using an existing wall
as one side. He has 16 metres of wire nettting. Find the dimensions of the run which will
give the maximum area. What is this area?
0. 500 1. 000
0. 200
0. 400
0. 600
y
(0.8, 0.64)
(0.5, 0.25)
x
Mathematics Learning Centre, University of Sydney 31
5 The clever idea behind dierential calculus (also
known as dierentiation from rst principles)
In this section we will have a look at the idea behind dierential calculus. While it is
important that you at least see this idea once, in practice you calculate the derivative of
a function using the procedures explained in Section 3. These procedures work because
of the clever idea that we are going to describe now, but in practice we just use them
without keeping in mind the whole time where they came from.
Figure 14 shows a portion of the graph of the function f(x) = x
2
.
Figure 14: Graph of y = x
2
. The solid line is the tangent to the graph at x = 0.5, and the
dotted line is an approximate tangent line drawn through the points (0.5, 0.25) and (0.8, 0.64)
which both lie on the graph of the function.
The tangent to the graph at the point (x, y) = (0.5, (0.5)
2
) is represented by the solid
line. We are going to nd the exact slope of this tangent.
To work out the slope of a line we need to know two points on the line. If we know the
points (x
1
, y
1
) and (x
2
, y
2
) on the line then the rise between these two points is y
2
y
1
,
and the run between them is x
2
x
1
, and so the slope of the line is given by
slope =
rise
run
=
y
2
y
1
x
2
x
1
We cannot use this formula directly to work out the slope of the tangent, because we only
know the exact location of one point on the tangent line, the point (0.5, 0.25). If we were
to pick another point from the diagram that looks like it is on the line, then we would
be back to using the approximate graphical methods from Section 2. If we want to get
the exact answer, we must use another way. This is the clever idea behind dierential
calculus.
We look at another line which which has slope nearly equal to the slope of the tangent,
and on which we do know two points. In Figure 14 we have drawn such a line (the dotted
(0.6, 0.36)
(0.5, 0.25)
Mathematics Learning Centre, University of Sydney 32
line) going through the points (0.5, (0.5)
2
), which is the actual point on the curve where
we are trying to nd the tangent, and (0.8, (0.8)
2
) which is another point on the curve.
This second point is not far from the point (0.5, (0.5)
2
), and so the slope of the line joining
them is not too dierent from the slope of the tangent at (0.5, (0.5)
2
). Because we know
two points on the dotted line, we can work out its slope. It is
slope =
(0.8)
2
(0.5)
2
0.8 0.5
= 1.3
The slope of the tangent is therefore about (but certainly not exactly) 1.3.
To get a better approximation we might try taking the second point closer to the point
(0.5, (0.5)
2
). In Figure 15 we have done this.
Figure 15: Graph of y = x
2
. The solid line is the tangent to the graph at x = 0.5, and the
dotted line is an approximate tangent line drawn through the points (0.5, 0.25) and (0.6, 0.36)
which both lie on the graph of the function.
Here the second point is just 0.1 units to the right of (0.5, (0.5)
2
). The slope of the dotted
line in this gure is
slope =
(0.6)
2
(0.5)
2
0.6 0.5
= 1.1
The exact slope of the tangent is closer to 1.1 than it is to 1.3, though we still dont know
its precise value.
If we wanted an even better approximation then we could choose the second point to be
even closer to (0.5, (0.5)
2
). For example we could try the second point to be (0.49, (0.49)
2
).
Notice that this point is to the left of (0.5, (0.5)
2
), whereas previously we had chosen points
to the right of the point. This is not important. What is important is that it is just 0.01
Mathematics Learning Centre, University of Sydney 33
units to the left of this point. The line joining these two points is very close to the actual
tangent, and the slope of this line is
slope =
(0.49)
2
(0.5)
2
0.49 0.5
= 0.99
It seems that the closer the second point gets to the point (0.5, (0.5)
2
) the closer the slope
of the line joining the two points gets to 1. We might guess that the slope of the tangent
to the curve at the point (0.5, (0.5)
2
) must be 1. We can be sure of this with the following
calculation.
Suppose that the second point is just h units to the right (or left if h < 0) of x = 1.
We can think of h as being a very small number. For example we have used h = 0.3,
h = 0.1, and h = 0.01 in our examples above. The coordinates of the second point
will be (0.5 + h, (0.5 + h)
2
). We can work out the slope of the line joining the points
(0.5, (0.5)
2
) and (0.5 + h, (0.5 + h)
2
) in the same way that we did above. It is
Slope =
(0.5 + h)
2
0.5
2
(0.5 + h) 0.5
=
(0.5 + h)
2
0.5
2
h
=
(.25 + h + h
2
) .25
h
=
h + h
2
h
= 1 + h
Moving the second point closer and closer to the rst is the same as making h closer and
closer to zero. But the slope is 1 +h so the closer that h gets to zero the closer the slope
gets to 1. The slope of the tangent is therefore exactly 1. This puts the matter beyond
doubt. We are no longer relying on approximations or guesses. We have shown that the
slope of the tangent to the graph of y = x
2
at the point (0.5, 0.25) is exactly 1. In symbols,
d
dx
x
2

x=0.5
= 1.
With the same method we could have found the slope of the tangent to the curve when
x = 1, or x = 0.37, or indeed at any value of the independent variable x.
Exercise 5.1 (a) Using the same technique as above, nd the slope of the tangent
to the graph of x
2
at x = 2. Check that this agrees with the
answer that you would have obtained using the results of Section
3.
(b) Using the idea introduced above, nd the slope of the tangent
to the graph of x
3
at x = 1. Check your answer by using the
techniques from Section 3.
Mathematics Learning Centre, University of Sydney 34
We dont really have to specify any particular value of x, but can leave it as unknown.
Slope of line through (x, x
2
) and (x + h, (x + h)
2
) =
(x + h)
2
x
2
x + h x
=
x
2
+ 2xh + h
2
x
2
h
= 2x + h
and as h 0 the line through (x, x
2
) and (x + h, (x + h)
2
) gets closer to the tangent
at x and the slope of this line gets closer and closer to 2x. The slope of the tangent is
therefore exactly 2x. This works no matter what the value of x is. For example the slope
of the tangent at x = 0.37 is 2 (0.37) = 0.74. We can say that the derivative of
the function x
2
is 2x. In symbols,
d
dx
x
2
= 2x.
We have chosen the function f(x) = x
2
for this example, because it is perhaps the simplest
function that gets across the idea. The same method works for any function, though the
resulting algebra will often be more dicult.
Exercise 5.2 (a) Using the ideas of this section, nd the derivative of x
3
.
(b) Use the same ideas to nd the derivative of x
3
+ 2x .
By now you probably have little doubt that the derivative of x
2
is 2x, and that the
derivative of x
3
is 3x
2
. Hopefully you are now willing to believe that the derivative of x
n
is nx
n1
, no matter what value n has. It would not be too dicult for us to prove this
fact, but the proof is beyond the scope of this booklet. However what you have seen in
this section is the basic idea that underlies all of dierential calculus, and all of the rules
and techniques of Section 3 come from it.
1. 00 2. 00 3. 0 - 1. 00 - 2. 00 3. 00
1. 00
2. 00
3. 00
4. 00
Mathematics Learning Centre, University of Sydney 35
6 Solutions to exercises
Exercise 1.1
From the graph, at time t = 0 the motorist is 100 metres from home, and at time t = 6
the motorist is 250 metres from home.
Exercise 1.2
At time t = 0 the motorist is 100 metres from home and at time t = 2 the motorist is
150 metres from home, so in the rst 2 seconds the motorist has travelled 150 100 = 50
metres. At time t = 3 the motorist is 175 metres from home and at time t = 5 the
motorist is 225 metres from home so in the time from t = 3 to t = 5 the motorist has
travelled 225 175 = 50 metres.
Exercise 1.3
A time t = 60 the motorist is 1008 metres from home and at time t = 62 the motorist is
1032 metres from home so in the 2 second interval from time t = 60 to time t = 62 the
motorist travelled 1032 1008 = 24 metres. A time t = 64 the motorist is 1072 metres
from home so in the 2 second interval from time t = 62 to time t = 64 the motorist has
travelled 1072 1032 = 40 metres.
Exercise 2.1
Refer to Figure 16. We have used the indicated points on the lines to calculate the slopes.
You may have chosed dierent points, but your answers should be close to those here.
Remember this is only an approximate way of nding the slopes, so you shouldnt consider
yourself wrong if you dont get exactly the same answers as here.
Figure 16: Tangents to graph of f(x) = x
2
.
Slope of tangent to f(x) = x
2
at x = 1 is
f

(1)
3 1
2 1
= 2.
0. 50 1. 00 1. 50 - 0. 50 - 1. 00
- 0. 50
0. 50
1. 00
1. 50
Mathematics Learning Centre, University of Sydney 36
The tangent at x = 0 is the x-axis, which has slope 0, so f

(0) = 0.
Slope of tangent to f(x) = x
2
at x = 0.5 is
f

(0.5)
2.75 0.75
3 (1)
= 1.
Exercise 2.2
Refer to Figure 17.
Figure 17: Tangents to graph of f(x) = x
3
.
Slope of tangent to f(x) at x = 1 is
f

(1)
1 (0.5)
1 0.5
= 3.
As in Exercise 2.1, the tangent at x = 0 is the x-axis which has slope 0, so f

(0) = 0.
Slope of tangent to f(x) at x = 0.5 is
f

(0.5)
0.25 (0.5)
0 (1)
= 0.75.
Exercise 3.1
(a) f

(x) = 4x
3
(b)
dy
dx
= 7x
8
(c) f

(u) = 2.3u
1.3
(d) f

(t) =
1
3
t

4
3
(e) f

(t) =
22
7
t
15
7
(f ) g

(z) =
3
2
z

5
2
(g)
dy
dt
= 3.8t
4.8
(h)
dz
dx
=
3
7
x

4
7
Mathematics Learning Centre, University of Sydney 37
Exercise 3.2
(a)
1
x
2
= x
2
so
d
dx

1
x
2

= 2x
3
(b) t

t = t
3
2
so
d
dt

=
3
2
t
1
2
(c)
d
dx
3

x =
d
dx
x
1
3
=
1
3
x

2
3
(d)
d
dx

1
x
2

=
d
dx
x

5
2
=
5
2
x

7
2
(e)
d
dx

1
x
4

=
d
dx
x
5/4
=
5
4
x

9
4
(f )
d
dx

s
3

s
3

=
d
ds
s
19
6
=
19
6
s
13
6
(g)
d
du

1
u
3

=
du
3
du
= 3u
4
(h)
d
dt

t
t
2

=
d
dt
t

3
2
=
3
2
t

5
2
(i)
d
dx

x
1
2

x
x

=
d
dx
1 = 0
Exercise 3.3
When x = 8 we have y =
3

8 = 2 so the point (8, 2) is on the line. Now

dy
dx
=
1
3
x

2
3
and
so
dy
dx
=
1
12
when x = 8. The tangent therefore has equation
y =
1
12
x + b.
Substituting x = 8 and y = 2 into this equation we obtain
2 =
1
12
8 + b
so that b =
4
3
. The equation is therefore y =
x
12
+
4
3
.
Exercise 3.4
(a) f

(x) = 10x x

1
2
(b)
dy
dx
= 14x
8
6x
3
(c) f

(t) = 5.75t
1.3
+
1
2
t
1
2
(d) h

(z) =
1
3
z

4
3
+ 5 (e) f

(u) =
5
3
u
2
3
+ 21u
8
(f ) g

(z) = 16z
3
+ 5z
2
(g)
dy
dt
= 40t
9
+
1
2
t

1
2
(h)
dz
dx
=
4
7
x

6
7
x

3
2
Mathematics Learning Centre, University of Sydney 38
Exercise 3.5
(a) f

(x) = 12x
2
(1 3x) 3(4x
3
+ 2)
(b) g

(x) = (2x + 1)(x

2
+ 1) + (x
2
+ x + 2)2x
(c) h

(x) = (9x
2
4x + 8)(x
2
2x + 4) + (3x
3
2x
2
+ 8x 5)(2x 2)
(d) f

(s) = s(3s + 5) + 3(1

s
2
2
)
(e) g

(t) = (
t

1
2
2
t
2
)(2t 1) + 2(

t +
1
t
)
(f ) h

(y) = (
y

1
2
2
+ 2y)(1 3y
2
) 6y(2

y + y
2
)
Exercise 3.6
The rate of change of r with respect to t is
dr
dt
= (1 t
2
)(t
2
2t + 1) + (t +
1
t
)(2t 2).
Substituting t = 2 we obtain (1
1
4
)(4 4 + 1) + (2 +
1
2
)(4 2) =
23
4
.
Exercise 3.7
The gradient of the tangent is given by
dy
dx
= (2x 2)(3x
3
5x
2
+ 2) + (x
2
2x + 1)(9x
2
10x).
Substituting x = 2 we obtain 28.
Exercise 3.8
(a) f

(x) =
(x + 1) (x 1)
(x + 1)
2
=
2
(x + 1)
2
(b) g

(x) =
(3x 2)2 (2x + 3)3
(3x 2)
2
=
13
(3x 2)
2
(c) h

(x) =
(x
2
+ 5)2x (x
2
+ 2)2x
(x
2
+ 5)
2
=
6x
(x
2
+ 5)
2
(d) f

(t) =
(1 + 2t
2
)2 8t
2
(1 + 2t
2
)
2
=
2 4t
2
(1 + 2t
2
)
2
(e) f

(s) =
(1

s)
1
2
s

1
2
+ (1 +

s)
1
2
s
1/2
(1

s)
2
=
s

1
2
(1

s)
2
(f ) h

(x) =
(x
3
+ 4)2x (x
2
1)3x
2
(x
3
+ 4)
2
=
x
4
+ 3x
2
+ 8x
(x
3
+ 4)
2
(g) f

(u) =
(3u
4
+ 5)(3u
2
+ 1) (u
3
+ u 4)12u
3
(3u
4
+ 5)
2
=
3u
6
9u
4
+ 48u
3
+ 15u
2
+ 5
(3u
4
+ 5)
2
(h) g

(t) =
(t
2
+ 3t + 1)(2t + 6) (t
2
+ 6t)(2t + 3)
(t
2
+ 3t + 1)
2
=
3t
2
+ 2t + 6
(t
2
+ 3t + 1)
2
Mathematics Learning Centre, University of Sydney 39
Exercise 3.9
(a)
d
dx

(2x + 3)
2

= 8x + 12
(b)
d
dx

(x
2
+ 2x + 1)
12

= 12(x
2
+ 2x + 1)
11
(2x + 2)
(c)
d
dx

(3 x)
21

= 21(3 x)
20
(d)
d
dx

(x
3
1)
5

= 5(x
3
1)
4
3x
2
= 15x
2
(x
3
1)
4
(e)
d
dt

t
2
5t + 7 =
d
dt
(t
2
5t + 7)
1
2
=
1
2
(t
2
5t + 7)

1
2
(2t 5)
(f )
d
dz

2 z
4

=
d
dz

(2 z
4
)

1
2

= 2z
3
(2 z
4
)

3
2
(g)
d
dt

(t
3

t)
3.8

= 3.8(t
3

t)
4.8
(3t
2

1
2

t
)
(h)
d
dx

(x +
1
x
)
3
7

=
3
7
(x +
1
x
)

4
7
(1
1
x
2
)
Exercise 3.10
(a)
d
dx

(x + 2)(x + 3)
2

= (x + 3)
2
+ 2(x + 2)(x + 3)
(b)
d
dx

(2x 1)
2
(x + 3)
3

= 4(2x 1)(x + 3)
3
+ 3(2x 1)
2
(x + 3)
2
(c)
d
dx

1 x

1 x
x
2

1 x
(d)
d
dx

x
1
3
(1 x)
2
3

=
1
3
x

2
3
(1 x)
2
3

2
3
x
1
3
(1 x)

1
3
(e)
d
dx

1 x
2

1 x
2
+ x
2
(1 x
2
)

1
2
1 x
2
Exercise 3.11
(a) f

(x) =
6x
2
2x
3
=
3
x
Alternatively write f(x) = ln 2 + 3 ln x so that f

(x) = 3
1
x
.
(b) f

(x) = 7x
6
e
x
7
(c) f

(x) =
7
x
(d) f

(x) = (2x + 3x
2
)e
x
2
+x
3
(e) Write f(x) = log
e
7 2 log
e
x so that f

(x) =
2
x
.
(f ) f

(x) = e
x
(g) f

(x) =
e
x
+ 3x
2
e
x
+ x
3
Mathematics Learning Centre, University of Sydney 40
(h) Write f(x) = ln e
x
+ 3 ln x so that f

(x) = 1 +
3
x
.
(i) Write f(x) = ln(x
2
+ 1) ln(x
3
x) so that f

(x) =
2x
x
2
+ 1

3x
2
1
x
3
x
.
Exercise 3.12
(a)
d
dx
cos 3x = 3 sin 3x
(b)
d
dx
sin(4x + 5) = 4 cos(4x + 5)
(c)
d
dx
sin
3
x = 3 sin
2
x cos x
(d)
d
dx
sin x cos x = cos
2
x sin
2
x
(e)
d
dx
x
2
sin x = 2x sin x + x
2
cos x
(f )
d
dx
cos(x
2
+ 1) = 2x sin(x
2
+ 1)
(g)
d
dx

sin x
x

=
x cos x sin x
x
2
(h)
d
dx
sin
1
x
=
1
x
2
cos
1
x
(i)
d
dx
tan

x =
1
2

x
sec
2

x
(j)
d
dx

1
x
sin
1
x

=
1
x
2
sin
1
x

1
x
3
cos
1
x
Exercise 4.1
f

(x) = 4x
3
4x so f

(x) = 0 at x = 0, 1 and the maxima and minima must occur at

the points x = 1, 0, 1, 2. Substituting these values into f(x) we nd that the maximum
occurs at x = 2 and the minimum occurs at x = 1 and at x = 1.
Exercise 4.2
g

(t) = (1 2t
2
)e
t
2
. Setting this equal to zero and solving we nd that the stationary
points are at t =
1

2
and the maximum must occur at one of the points t = 2,
1

2
, 2.
Substituting into g(t) we nd that the maximum value occurs at t =
1

2
.
Exercise 4.3
h

(u) = 6u
2
+ 6u 12 = 6(u
2
+ u 2). The stationary points are at u = 2, 1 and the
minimum value occurs at one of the points u = 3, 2, 1, 2. Substituting into h(u) we
nd that the minimum occurs at u = 1.
x
8 - x /2
Mathematics Learning Centre, University of Sydney 41
Exercise 4.4
If we let the side of the run that is opposite the existing wall have length x, then the other
side of the run has length 8
x
2
.
The area of the run is A(x) = x(8
x
2
) and we must maximise this function in the region
0 x 16. Dierentiating gives A

(x) = 8 x so the only stationary point is at x = 8.

The maximum occurs at one of x = 0, 8, 16. Substituting, we see that the maximum
occurs when x = 8, giving an area of 32 square metres.
Figure 18: A chicken run built against the side of an existing wall, with 16 metres of netting.
Exercise 5.1
b.
Slope =
(1 + h)
3
1
3
(1 + h) 1
=
1 + 3h + 3h
2
+ h
3
1
h
=
3h + 3h
2
+ h
3
h
= 3 + 3h + h
2
.
So, the slope of the tangent to the graph of x
3
at x = 1 is 3.
Exercise 5.2
b. Slope of the line through (x, x
3
+ 2x) and ((x + h), (x + h)
3
+ 2(x + h)) is
Slope =
((x + h)
3
+ 2(x + h)) (x
3
+ 2x)
(x + h) x
=
x
3
+ 3x
2
h + 3xh
2
+ h
3
+ 2x + 2h x
3
2x
h
=
3x
2
h + 3xh
2
+ h
3
+ 2h
h
= 3x
2
+ 3xh + h
2
+ 2.
As h 0 the slope of this line 3x
2
+ 2.
So, the slope of the tangent to the curve is exactly 3x
2
+ 2.