UNIVERSITY OF GLASGOW

CALCULUS

1S

S C H O O L O F M AT H E M AT I C S A N D S TAT I S T I C S
Copyright © 2018 AJW, after JJCN, TAW and others

published by university of glasgow

Last updated, January 2018

Course Contents

1 Maclaurin Series. . . . . . . . . . . . . . . . . 7
2 Introduction to Integration . . . . . . . . . . . . 19
3 Techniques of Integration . . . . . . . . . . . . . 41
4 Ordinary differential equations . . . . . . . . . . . 67
5 Numerical methods . . . . . . . . . . . . . . . . 79
6 Limits . . . . . . . . . . . . . . . . . . . . . 85
7 Standard Integrals . . . . . . . . . . . . . . . . 91
TO S E E T H E W O R L D I N A G R A I N O F S A N D,

A N D A H E AV E N I N A W I L D F L O W E R ;

H O L D I N F I N I T Y I N T H E PA L M O F Y O U R H A N D ,

AND ETERNITY IN AN HOUR.

WILLIAM BLAKE

A M AT H E M AT I C I A N , L I K E A PA I N T E R O R A P O E T, I S A M A K E R O F PAT T E R N S .

I F H I S PAT T E R N S A R E M O R E P E R M A N E N T T H A N T H E I R S , I T I S B E C A U S E

THEY ARE MADE WITH IDEAS.

G . H . H A R DY

M AT H E M AT I C S R E V E A L S I T S S E C R E T S O N LY T O T H O S E W H O A P P R O A C H I T

W I T H P U R E L O V E , F O R I T S O W N B E A U T Y.

ARCHIMEDES
Maclaurin Series
1
1.1 Introduction . . . . . . . . . . . . . . . . . . 8
1.1.1 Calculation of Maclaurin series 9
1.1.2 Manipulation of Maclaurin series 12
1.2 Some applications of Maclaurin series . . . . . . . . . . 14
1.2.1 Evaluation of limits 15
 8 calculus

1.1 Introduction

A power series in x is Definition

∞
∑ an x n power series,
n =0 Maclaurin series,
for constants an . This infinite sum is defined as the limit of the trun- range of validity.
cated sums
∞ N
∑ an xn = Nlim
→∞
∑ an x n
n =0 n =0
when the limit exists. If the function f ( x ) can be represented by a
power series for some values of x then
∞
f (x) = ∑ an x n
n =0
and we call the power series the Maclaurin series for f ( x ). The values Maclaurin & Taylor Series
of x for which the infinite sum exists and is equal to f ( x ) belong to For a function f its Taylor series about 0
is another name for its Maclaurin series.
the range of validity of the Maclaurin series. Jump forward to the Taylor series (extra)
section on page 16 to learn more.
Defintion 1.1 Consider Maclaurin’s early academic life was
here at the University of Glasgow; he
f ( x ) = a0 + a1 x + a2 x 2 + · · · + a N x N + R N ( x ) . started his undergraduate degree at age
| {z } | {z }
truncated power series remainder eleven in 1709 — google ‘Maclaurin bi-
ography’.
Letting N → ∞, we get f ( x ) = ∑∞
n =0 a n x
n
for values of x for which
R N ( x ) → 0 as N → ∞ (this determines the range of validity).
The truncated power series a0 + a1 x + a2 x2 + · · · + a N x N is called
the nth -degree Maclaurin1 polynomial of f . 1
Also called the nth -degree Taylor poly-
nomial of f .

Example 1.2 Show that the interval (−1, 1) is included in the range
of validity of the Maclaurin series for (1 − x )−1 .

Solution Recall that the geometric series is:

1 − x N +1
(1 + x + x 2 + · · · + x N ) =
| {z } 1−x
N + 1 terms, common ratio x
1 x N +1
= − , x 6= 1.
1−x 1−x
Thus, on rearranging, we have
1 x N +1
= 1| + x + x2{z+ · · · + x N} + .
1−x 1−x
poly of degree N | {z }
remainder

For | x | < 1,
|x| N → 0 as N → ∞
therefore
R N (x) → 0 as N → ∞
and so
∞
1
= ∑ xn
1−x n =0
for x ∈ (−1, 1). Therefore the range of validity includes (−1, 1).
 maclaurin series 9

1.1.1 Calculation of Maclaurin series

Suppose the Maclaurin series of a function f ( x ) exists. Then we have,
for x in the range of validity,
∞
f ( x ) = a0 + a1 x + a2 x 2 + a3 x 3 + · · · = ∑ an x n (†)
n =0

We need a method to calculate a0 , a1 , a2 , . . . for any particular f ( x ).

We make an assumption that the Maclaurin series for any derivative
of f ( x ) can be obtained by differentiating the Maclaurin series for
f ( x ) term by term. The procedure for calculation of an is then

a) Evaluate (†) at x = 0, a0 = f (0).

b) Differentiate (†) term by term with respect to x to give

∞
f 0 ( x ) = a1 + 2a2 x + 3a3 x2 + 4a4 x3 + · · · = ∑ nan xn−1 . (‡)
n =1

Then evaluation of (‡) at x = 0 gives a1 = f 0 (0).

f ( n ) (0)
c) Repeat this procedure to find that the general term is an = .
n!

Example 1.3 Some standard Maclaurin series are given in Table 1.1.
Observe that for the final series for (1 + x )α , that the range of validity
always includes (−1, 1). If α is a positive integer then the series
terminates (an = 0 when n > k for some k) and the Maclaurin series
is a polynomial. The notation2 2
This notation will be studied in depth
in 1S Algebra §2.2 Combinations and Per-
α(α − 1)(α − 2) · · · (α − r + 1) mutations.
 
α
= ,
r r!
 
α
for r > 0 (the numerator contains r terms). For r = 0, = 1.
0
Observe also that the powers appearing in the Maclaurin series
for cosine are all even and those for sine are odd. These words will
appear again in §3.7.1 when we discuss the symmetry of functions.

Example 1.4 Find the Maclaurin series for the following functions

a) f ( x ) = e x , c) h(t) = (1 − t)1/2 ,
1
b) g( x ) = log (1 + x ) , d) p(y) = .
2+y

Solution

a) Let f ( x ) = e x and recall that the exponential function has the

following property
f (n) ( x ) = e x
10 calculus

Function Maclaurin series Range of validity

x x2 x3 x4
ex = 1+ + + + +··· x∈R
1! 2! 3! 4!
x3 x5 x7
sin x = x− + − +··· x∈R
3! 5! 7!
x2 x4 x6
cos x = 1− + − +··· x∈R
2! 4! 6!
x2 x3 x4 x5
log(1 + x ) = x− + − + −··· x ∈ (−1, 1]
2 3 4 5

 
α
 
α 2
 
α 3 (−1, 1) if α < 0
(1 + x ) α = 1+ x+ x + x +··· x∈
1 2 3  R if α > 0

Table 1.1: Standard Maclaurin series.

Observe that for the final series for (1 +
where f (n) is the nth derivative. Evaluation at x = 0 gives f (n) (0) = x )α , that the range of validity always in-
cludes (−1, 1). If α is a positive integer
e0 = 1, so
then the series terminates (an = 0 when
f ( n ) (0) 1 n > k for some k) and the Maclaurin
an = = .
n! n! series is a polynomial.
The Maclaurin series is therefore
∞
xn x x2 x3
ex = ∑ n!
= 1+ +
1! 2!
+
3!
+··· for all x ∈ R.
n =0

b) For g( x ) = log (1 + x ), recall that

d 1
(log y) =
dy y
and
dn
(log y) = (−1)n+1 (n − 1)!y−n
dyn
so that, applying the chain rule we have

(−1)n+1 (n − 1)!
g(n) ( x ) = , n ≥ 1.
(1 + x ) n
Hence
g(n) (0) = (−1)n+1 (n − 1)!, n≥1
and g(0) = 0. Thus the coefficients are

0 (−1)2 (−1)3 1 (−1)4 1 (−1)n+1

= 0, = 1, =− , = ,..., ,....
0! 1 2 2 3 3 n
Therefore the required Maclaurin series is

x2 x3 x4 x5
log (1 + x ) = x − + − + −··· for some x ∈ R.
2 3 4 5
The techniques to determine the range of validity are introduced
in Level-2 Mathematics. Observe that since the domain of the nat-
ural logarithm is all
n positive real numbers,
o the domain of g( x ) =
log(1 + x ) is D = x ∈ R  x > −1 . It follows that3 the range
 3
since we want the power series to be
equal to the function
of validity must be a subset of D. In fact, the series can be shown4 4
using the ratio test covered in the
to converge and be equal to log(1 + x ) when −1 < x > 1. course 2E: Real Analysis
 maclaurin series 11

c) h(t) = (1 − t)1/2 . Since (1 − t)1/2 = (1 + (−t))1/2 = (1 + x )1/2 ,

we can use the series expansion from Table 1.1. To do this we first
calculate the values of the binomial coefficients,
 
1 1
2 2 1
= = ,
1 1! 2
  
1 1
1
2 2 −1 −1
2 = = ,
2 2! 8
   
1 1 1
1
2 2 −1 2 −2 1
2 = = .
3 3! 16

1
Substituting these into the series expansion for (1 + x ) 2 we have

(1 − t)1/2 = (1 + (−t))1/2 = (1 + x )1/2

1 1 1
= 1 + 2 x + 2 x2 + 2 x3 + · · · (−1 < x < 1)
1 2 3
−1
 
1 1
= 1 + (−t) + (−t)2 + (−t)3 + · · ·
2 8 16
1 1 1
= 1 − t − t2 − t3 − · · ·
2 8 16

Valid for −1 < (−t) < 1, that is for −1 < t < 1.

1
d) p(y) = .
2+y
1 1 1 y  −1 1
Rewriting as  y  = 1 + = (1 + x ) −1
2+y 2 1+ 2 2 2
2
and since
−1
 
= (−1)n
n

we have that

1 1
= (1 + x ) −1
2+y 2
−1 −1 2 −1 3
       
1
= 1+ x+ x + x +···
2 1 2 3
1 
= 1 − x + x2 − x3 + x4 − · · ·
2 
1 y  y 2  y 3  y 4
= 1− + − + −···
2 2 2 2 2
1 y y2 y3 y4
= − + − + −···
2 4 8 16 32

y
Valid for −1 < x = 2 < 1, that is for all y ∈ (−2, 2).
 12 calculus

1.1.2 Manipulation of Maclaurin series

Consider the Maclaurin series for f ( x ) and g( x ) and let I be the inter-
section of the range of validity for both series5 . Then the Maclaurin
series for f ( x ) + g( x ), f ( x ) g( x ) and f ( g( x )) can be found using the
Maclaurin series for f ( x ) and g( x ). Suppose
∞ ∞
∑ ∑ gn x n
f (x) = f n xn , g( x ) =
n =0 n =0
5
Since we wish to create new Maclaurin
series out of known ones, we need both
series to be valid. That is x must belong
to both the range of validity of f ( x ) and
that of g( x ). Recall from 1R Calculus,
that x belonging to one set and another
is equivalent to saying that x belongs to
the intersection of the two sets.

then, for x ∈ I,

∞
f ( x ) + g( x ) = ∑ ( f n + gn ) x n
n =0
∞ ∞
! !
f ( x ) g( x ) = ∑ fn x n
∑ gm x m
n =0 m =0
∞
!
n
= ∑ ∑ f r gn −r xn
n =0 r =0
!n
∞ ∞
f ( g( x )) = ∑ fn ∑ gm x m
n =0 m =0

Example 1.5 Find the following Maclaurin series

e2x
a) up to the term x3 , and
1−x

b) log(cos x ) up to the term x6 .

Solution

a) To begin with6 we have 6

The strategy here is to find Maclaurin
series both for e2x and (1 − x )−1 and
then multiply these series together to
y y2 y3 y4 find the Maclaurin series for
e2x = ey = 1 + + + + +··· , y ∈ R
1! 2! 3! 4! e2x
2x (2x )2 (2x )3 (2x )4 1−x
= 1+ + + + +··· , x ∈ R
1! 2! 3! 4! as far as the term x3 . To obtain
the range of validity for this new se-
4x3 2x4
= 1 + 2x + 2x2 + + +··· , x ∈ R ries we’ll take the intersection of both
3 3 ranges of validity.

and also

(1 − x ) −1 = 1 + x + x 2 + x 3 + x 4 + · · · , −1 < x < 1.

Therefore
 maclaurin series 13

e2x
= e2x · (1 − x )−1
1−x
 4x3 2x4   
= 1 + 2x + 2x2 + + + · · · · 1 + x + x2 + x3 + x4 + · · ·
3 3
4
= 1(1 + x + x2 + x3 + · · · ) + 2x (1 + x + x2 + · · · ) + 2x2 (1 + x + · · · ) + x3 (1 + · · · ) + · · ·
3
4
= 1(1 + x + x2 + x3 ) + 2x (1 + x + x2 ) + 2x2 (1 + x ) + x3 + h.o.t.
3
4
= 1 + x + x2 + x3 + 2x + 2x2 + 2x3 + 2x2 + 2x3 + x3 + h.o.t.
3
2 19 3 The abbreviation ‘h.o.t.’ means
= 1 + 3x + 5x + x + h.o.t. higher order terms.
3

b) Consulting7 Table 1.1 we see that we have Maclaurin series for 7

To find the Maclaurin series for
cos x and for log(1 + x ). Thus we write log(cos x ) up to the term x6 our strat-
egy is again to attempt to write the
required function in terms of func-
log(cos x ) = log(1 + (−1 + cos x )) = log(1 + y), tions for which we already know their
Maclaurin series.
where y = −1 + cos x and since

y2 y3 y4
log(1 + y) = y − + − +···
2 3 4

and has range of validity −1 6 y 6 1. That is

−1 6 −1 + cos x 6 1, i.e. 0 6 cos x 6 2.

We now need the Maclaurin series for cos x − 1,

x2 x4 x6
 
−1 + cos x = −1 + 1 − + − +···
2! 4! 6!
x 2 x 4 x 6
=− + − −···
2! 4! 6!
Hence

log(cos x ) = log(1 + (−1 + cos x ))

= log(1 + y)
1 1 1
= y − y2 + y3 − y4 + · · ·
2 3 4
2 4 6 2 3
x2 x4 x2
   
x x x 1 1
= − + − −··· − − + −··· + − +··· +···
2! 4! 6! 2 2! 4! 3 2!
x2 x4 x6 1 x4 x4 x2 x6
   
1
=− + − − −2 + − + h.o.t.
2! 4! 6! 2 4 4! 2! 3 (2!)3
x2
   
1 1 1 1 1
= − + x4 − + x6 − − + h.o.t.
2! 24 8 48 24 720
x2 x4 x6
=− − − + h.o.t.
2 12 45
 14 calculus

1.2 Some applications of Maclaurin series

See also Stewart’s Calculus §11.11 for
Suppose further applications of Maclaurin se-
ries. For example, in Physics to Ein-
stein’s theory of Special Relativity and
f ( x ) = a0 + a1 x + a2 x 2 + a3 x 3 + · · · + a n x n + R n ( x ). the field of Optics.
| {z }
Truncated Maclaurin series

When x is small, the remainder Rn ( x ) is typically less than k | x |n+1

for some constant k. Therefore when | x | is small the first few terms
of a Maclaurin series — the nth -degree Maclaurin polynomial of f —
gives a good approximation8 for f ( x ). 8
The remainder, Rn ( x ), is the error.
x
For example x − is a reasonable approximation for sin x when x
3!
x3 x5
is small with error < k| x |5 and x − + is a better approximation,
3! 5!
with error < k| x |7 .

Example 1.6 Some applications

a) Show that

x3 x5 x7
   
1+x
log =2 x+ + + +··· ,
1−x 3 5 7

1
for | x | < 1 and take x = 3 to find an approximation of log 2.

 x n
b) Fix x and show that lim 1+ = e x . In particular then we’ll
n→∞ n
have9 9
Another definition for Euler’s con-
stant!
 n
1
lim 1+ = e.
n→∞ n

Solution

a) We have, for y = − x,
 
1+x
log = log(1 + x ) − log(1 − x )
1−x
= log(1 + x ) − log(1 + y)
x2 x3 x4 y2 y3 y4
   
= x− + − +··· − y− + − +···
2 3 4 2 3 4
x2 x3 x4 x2 x3 x4
   
= x− + − + · · · − −x − − − −···
2 3 4 2 3 4
  3  5 
x x
= 2x + 2 +2 +···
3 5

x 3 x 5 x 7 
=2 x+ + + +··· .
3 5 7
 maclaurin series 15

 
1
1 (1 + x ) 1+ 3(4/3)
Take x = so =   = = 2. Then
3 (1 − x ) 1
1− 3 ( 2/3)

 3  5  7 !
1 1 1 1 1 1 1
log 2 = 2 + + + +···
3 3 3 5 3 7 3
  !
1 1 1 3 1 1 5 1 1 7
   
53056
≈2 + + + =
3 3 3 5 3 7 3 76545
≈ 0.69313475

The true value of log 2 is 0.69314718 to eight decimal places, hence

our estimate is correct to four decimal places.10 10
Of course we could have been lazy (as
the question didn’t specify) and taken
b) Consider  
1+x
log ≈ 2x
 x n   x 1−x
log 1 + = n log 1 +
n   n    Then x = 13 gives an estimate of 23 =
x 1 x 2 1  x 3 0.66666667 to eight decimal places —
=n − + −··· which is correct to the true value to one
n 2 n 3 n
decimal place.
x2 x3
= x− + 2 −···
2n 3n
x2 1 x3 1 x4 1
     
= x− + − +···
2 n 3 n2 4 n3
1  x n 
As n → ∞, → 0 for p > 0, so log 1 + → x.
np  n
x n
Hence lim 1 + = ex .
n→∞ n

1.2.1 Evaluation of limits

Example 1.7 Evaluate the following limits

e3x − e x
a) Find lim . Consider
x →0 x
3x (3x )2 x2
   
3x x x
e −e = 1+ + +··· − 1+ + +···
1! 2! 1! 2!
= 2x + 8x2 + · · ·
e3x − e x e3x − e x
So = 2 + 8x + · · · so lim = 2.
x | {z } x →0 x
(?)
(?) goes to 0 as x → 0.
x cos 2x − sin x
b) Find lim . Consider
x →0 x3
(2x )2 x3
   
x cos 2x − sin x = x 1 − +··· − x− + · · · + h.o.t.
2! 3!
x3
= x − 2x3 − x + + h.o.t.
3!
11 3
=− x + h.o.t.
6
16 calculus

x cos 2x − sin x 11
So = − + h.o.t..
x3 6
x cos 2x − sin x 11
Hence lim 3
=− .
x →0 x 6

Complex arguments (extra)

Maclaurin series can be used with a complex argument For example
for z ∈ C,
z z2
ez = 1 + + +···
1! 2!
Euler’s famous formula can be proved using this Maclaurin series.

Theorem 1.1 Euler’s formula

For θ ∈ R,
eiθ = cos θ + i sin θ.

Proof Let z = iθ. Then

iθ (iθ )2
eiθ = 1 + + +···
1! 2!
θ2 iθ 3 θ4 iθ 5
= 1 + iθ − − + + −···
2! 3! 4! 5!
θ2 θ4 θ3 θ5
   
= 1− + −··· +i θ − + −···
2! 4! 3! 5!
= cos θ + i sin θ.

Taylor series (extra)

The Maclaurin series of f ( x ) is an expansion of the function about Definition
x = 0. In general we require f ( x ) near x = a, say. An expansion Taylor series.
of f ( x ) about x = a is called a Taylor series. Define g( x ) = f ( x + a),
then the Maclaurin series for g( x ) is
∞
1 (n)
g( x ) = ∑ n!
g (0) x n .
n =0

Use the chain rule to deduce that

g ( n ) (0) = f ( n ) ( a )

so
∞
1 (n)
g( x ) = f ( x + a) = ∑ n!
f ( a) xn ,
n =0
or
∞
1 (n)
f (x) = ∑ n!
f ( a) ( x − a)n .
n =0
 maclaurin series 17

Hyperbolic cosine and sine (extra)

Consider the Maclaurin series for sine and cosine

x3 x5 x7
sin x = x − + − +...
3! 5! 7!

x2 x4 x6
cos x = 1 − + − +...
2! 4! 6!

and the following question.

Question 1.1
What functions have the following Maclaurin series?

x3 x5 x7
S( x ) = x + + + +...
3! 5! 7!

x2 x4 x6
C(x) = 1 + + + +...
2! 4! 6!

Note that

x2 x3 x4 x5
ex = 1 + x + + + + +...
2! 3! 4! 5!

x2 x3 x4 x5
e− x = 1 − x + − + − +...
2! 3! 4! 5!

addition of e x and e− x gives twice the sum of the terms with even
powers, subtraction gives twice the sum of the odd powers. So

1 x x3 x5 x7
e − e− x = x +


S( x ) = + + +...
2 3! 5! 7!

1 x x2 x4 x6
e + e− x = 1 +


C(x) = + + +...
2 2! 4! 6!

By analogy with sine and cosine we call these functions the hyper-
bolic sine and hyperbolic cosine. Definition
sinh hyperbolic sine,
1 x 1 x cosh hyperbolic cosine.
e − e− x , e + e− x



sinh x = cosh x =
2 2

using the Maclaurin series it is easy to deduce the relations

sinh (ix ) = i sin x, cosh (ix ) = cos x.

Hyperbolic sine and cosine have similar properties to sine and co-
sine, including hyperbolic trigonometric identities and results in dif-
ferentiation and integration.11 These results can be proved form the 11
Hyperbolic sine and cosine are useful
definitions of sinh and cosh. when performing
√ integrals that involve
terms like x2 − 1.
18 calculus

Example 1.8 Show that

a) cosh2 x − sinh2 x = 1, and b) d

dx (cosh x ) = sinh x.

Solution

a) We have that

1 x
2 1 x
2
cosh2 x − sinh2 x = e + e− x − e − e− x
4 4
1  2x −2x
 
= e +2+e − e2x − 2 + e−2x
4
=1
cosh2 x − sinh2 x = 1

and

b)
 
d d 1 x 1 x
e + e− x e − e− x = sinh x.



(cosh x ) = =
dx dx 2 2
Introduction to Integration
2
2.1 The area problem . . . . . . . . . . . . . . . . . 20
2.1.1 Area of a circle 21
2.1.2 The area under a curve 23
2.2 The definite integral . . . . . . . . . . . . . . . . 26
2.2.1 Riemann sum estimates 26
2.2.2 Interpreting integrals as area 28
2.2.3 Properties of definite integral 29
2.3 Fundamental theorems of Integral Calculus (FToC) . . . . . 32
2.3.1 First fundamental theorem of Integral Calculus 32
2.3.2 Antidifferentiation 34
2.3.3 Second fundamental theorem of Integral Calculus 35
2.4 Differentiation and integration as inverse processes . . . . . 36
2.4.1 The indefinite integral 36
2.4.2 Standard indefinite integrals 37
2.4.3 Linearity of indefinite integrals 38
 20 calculus

2.1 The area problem

In this section we’ll consider the area problem.

Question 2.1
Given a suitable curve y = f ( x ) and two values a, b in the do-
main of f , what is the area under the curve between x = a and
x = b?

In Figure 2.1 it is the shaded area whose area we wish to know.

y Figure 2.1: Area under the curve y =

f ( x ) between x = a and x = b.
y = f (x)

This course skims over the meaning of

‘suitable curve’ from Question 2.1. At
Level 2, in the course 2E Introduction to
Real Analysis it is covered in some de-
tail. Roughly, ‘suitable curve’ means
the following: suppose that f is the
function defining the curve, then at ev-
ery point a in its domain the value
a x f ( x ) approaches f ( a) as x approaches
b
a. This is called a continuous function
— a small change in the input results in
a small change in output.
To answer Question 2.1, we first need to understand the following. For example, the following types of
functions are continuous at every point
in their domains
Question 2.2 a) polynomials,
What is area? Or more precisely, given a region in the real plane b) rational functions,
how is its area defined? c) functions involving rational powers,
d) trigonometric functions.

In the remainder of this section we will answer Questions 2.1 and

2.2. Let us start with some simple cases.

Example 2.1 Illustrated in Figures 2.3 and 2.2 below.

a) The area of a rectangle with sides of length b and h is bh.

b) The area of a right-angled triangle with base b and height h is A4

easily seen to be half of the rectangle with sides b and h.

A3
c) For a non-right angled triangle, we can split it into two right- A5
angled triangles and then its area is the sum of the two right-
angled triangles. A2

d) For a polygon, that is a figure in the plane bounded by some A1

number of straight lines, we can apply a similar strategy to that
of the non-right-angled triangle. Namely, given a polygon as in
Figure 2.2 we can split it into a finite number of non-overlapping
triangles and calculate the area of the polygon by summing the
Figure 2.2: Area of a polygon.
 introduction to integration 21

individual areas of the triangles. Thus, the area of the polygon

in Figure 2.2 is a sum of the areas of the triangles. That is, A =
A1 + A2 + A3 + A4 + A5 .

h
1 1 1
bh 2 bh 2 b1 h 2 b2 h

b b b1 b2

Figure 2.3: Areas of rectangles and tri-

angles.

2.1.1 Area of a circle

What about defining the area for regions with curved edges? Let’s
take a look at the case of the circle. The basic idea is the same:
somehow split the circle up into regions for which we can calculate
the area and take the sum of these to find the answer. We’ll look at
three different approaches.
The first1 approach is to observe that a regular polygon, for exam-
ple an equilateral triangle, square or regular pentagon, can be drawn
inside a circle with its vertices on the edge of the circle.2 Now the 1
This approach is attributed to Euclid
area of the circle is only a little larger than that of the regular poly- circa 5th Century B.C.. In fact he
went further in his analysis by consid-
gon. As the number of sides of the regular polygon is increased this ering both an inscribed polygon and an
difference between the area of the inscribed polygon and the area of out-scribed polygon, allowing him to
bound the area of the circle between the
the circle decreases steadily. We can then define the area of the circle areas of the smaller inscribed and larger
as the limit of this process of increasing the number of sides. More circumscribed regular polygons.
formally, let A be the area of the circle and An be the area of the
regular polygon with n sides, as in Figure 2.4. Then we can define 2
We say that this regular polygon has
been inscribed on the circle.
A = lim An .
n→∞

As above, we can calculate each of the values of An by splitting up

the polygon into triangles.
The second approach is to observe that a circle, much like the
growth of a tree in annual rings or the structure of an onion as seen

A4 A6 A8 A16

Figure 2.4: Area of circle with polygons.

22 calculus

in Figure 2.5, can be decomposed into a collection of rings with a

common centre. In Figure 2.6, we see that these rings can all be cut
across a radius and ‘unrolled’. As in the previous approach, if we let
the width of the rings approach zero as they increase in number to-
wards infinity, then they can be unrolled into a right-angled triangle
whose area is easily calculated as 12 (2πr ) r = πr2 .

Figure 2.5: Tree rings.

2πr

Figure 2.6: Area of circle with rings.

The third approach3 to defining the area of the circle is to divide it 3
Of course there’s another very similar
up into vertical rectangular strips each of some small common width. forth approach where we divide up the
circle into horizontal rectangular strips.
This gives us an estimate of the area of the circle as illustrated in Fig-
ure 2.7. Increasing the number of rectangular strips as we decrease
the common width of each strip yields estimates closer and closer to
the true area of the circle. Now, as above, we can define the area of
the circle to be the limit of the sums of the areas of the rectangular
strips as the number of strips tends to infinity.
More formally, suppose there are n rectangular strips and let Rn
be the estimate of the area of circle using those n strips. That is,
R n = S1 + S2 + · · · + S n
where Si is the area of the ith rectangular strip. Then the area of the
circle, A, is defined to be
A = lim Rn
n→∞
Later in the course we’ll have covered enough to calculate rigorously
the area of the circle using this approach.

R4 R8 R32 R64

Figure 2.7: Area of circle with rectan-

For a more general figure with curved sides any of these three ap- gles.

proaches can be used to define the area of the figure. In the following
section we focus on the third approach.
 introduction to integration 23

Observation 2.1
The four approaches outlined above can be thought of in terms
of polar and cartesian coordinates. We outline this in Table 2.1.

Table 2.1: Comparison of the four ap-

Approach Sum along Infinitesimal width
proaches
inscribed polygon angle dθ
concentric rings radius dr
vertical rectangles x-axis dx
horizontal rectangles y-axis dy y
y = x2

1
2.1.2 The area under a curve
Returning to Question 2.1, we use the rectangular strip approach
to estimate and finally calculate precisely the area under a familiar
curve.

Example 2.2 Estimate the area under the curve y = x2 between

x = 0 and x = 1 with four rectangular strips. x
1

Solution The area we wish to estimate is the shaded region in Fig- Figure 2.8: Area under parabola be-
tween 0 and 1.
ure 2.8. Since the estimate is to be made with four strips and the total
width of the shaded region is one unit, each of the strips will have 4
There are other options, for example
a width of 41 unit. If we take the leftmost edge of each rectangular we could take the rightmost edge or
perhaps the midpoint of the strip to be
strip to be touching the curve,4 then the leftmost edges of the four resting on the curve. Indeed, in gen-
strips have x-values x = 0, 41 , 21 and 34 . The square of the x-value at eral a more accurate estimate will be
obtained by taking the midpoint, how-
the leftmost edge is the height of that rectangular strip. This data ever it is easier to calculate if one takes
plus the area of each strip is displayed in Table 2.2. the end points.

Table 2.2: Areas of the rectangular

Strip number x-value edge Height Area
strips.
1
1 0 (0)2 = 0 0· 4 =0
1
2 4 ( 14 )2 = 1
16
1
16 · 1
4 = 1
64
1
3 2 ( 21 )2 = 1
4
1
4 · 1
4 = 1
16
3
4 4 ( 34 )2 = 9
16
9
16 · 1
4 = 9
64

Summing the areas of the four strips yields an estimate, R4 , for

the shaded area,

1 1 9 14
(area under y = x2 between 0 and 1) ≈ R4 = 0 + + + = = 0.21875.
64 16 64 64

This estimate is rather crude,5 let’s repeat the above example with 5
We can see from Figure 2.9 that this is
more strips an under estimation of the true value.
What would happen to this estimate
if we used the rightmost edge of the
Example 2.3 Estimate the area under the curve y = x2 between strips, or the midpoint?
x = 0 and x = 1 with sixteen rectangular strips.
24 calculus

y y y
y = x2 y = x2 y = x2
1 1 1

x x x
1 1 1
n = 16 n = 32 n = 64

Figure 2.9: Estimating the area under

the parabola.
Solution Again taking the rectangular strips with their leftmost
edge resting on the curve, we calculate their areas by multiplying
1
height by the width of 16 . This data is displayed in Table 2.3.

Table 2.3: Areas of the rectangular

Strip number x-value edge Height Area
strips.
1 0 0 0
1 1 1
2 16 256 4096
1 1 1
3 8 64 1024
3 9 9
4 16 256 4096
1 1 1
5 4 16 256
5 25 25
6 16 256 4096
3 9 9
7 8 64 1024
7 49 49
8 16 256 4096
1 1 1
9 2 4 64
9 81 81
10 16 256 4096
5 25 25
11 8 64 1024
11 121 121
12 16 256 4096
3 9 9
13 4 16 256
13 169 169
14 16 256 4096
7 49 49
15 8 64 1024
15 225 225
16 16 256 4096

Summing the areas of the sixteen strips yields an estimate, R16 ,

for the shaded area,

1 15 n2
 
155
16 n∑
(area under y = x2 between 0 and 1) ≈ R16 = · = = 0.302734375.
=1 162 512
 introduction to integration 25

We can continue this process of increasing the number of strips

further to obtain estimates Rn . In Table 2.4 we show the estimated
area for larger numbers of strips.

Number of strips 4 8 16 32 64 128 256 512 1024

Area estimate, Rn (3 dp) 0.219 0.273 0.303 0.318 0.326 0.329 0.331 0.332 0.333
Table 2.4: Estimating the area under the
parabola.
We can see from Table 2.4 that the area estimate Rn appears to
be converging towards 0.333 . . . = 13 as n tends to infinity. In the
following example we verify that this is the true limit and hence the
precise area under y = x2 between x = 0 and x = 1.
1
Example 2.4 Show that lim Rn = .
n→∞ 3

Solution We begin by expressing in terms of k the estimate, Rk , of

the shaded area using k rectangular strips. Since there are k strips
the width of each is 1k and the x-value of the leftmost edge of strip
 2  2
i is ki . Thus the height of strip i is ki and so its area is 1k · ki .
Summing these k strips all together gives
 2  2
k−1 2 k −1
 
1 2 1
Rk = 0 +
k
+
k
+···+
k
= 3
k ∑ i2 .
i =1

From 1R Algebra §5.4, we recall that

n
n(n + 1)(2n + 1)
∑ i2 =
6
.
i =1

Hence (with n = k − 1 from above)

1 (k − 1) ((k − 1) + 1)(2(k − 1) + 1) 1 1 1
Rk = · = − +
k3 6 3 2k 6k2
and therefore
1
Rk −→ as k −→ ∞.
3

Thus we can say that the shaded area in Figure 2.8 is 31 . In gen-
eral, we take can take this approach as the definition of the area —
answering the questions that began this section.
Definition
Defintion 2.5 Let y = f ( x ) be a continuous6 function and a, b two
area under a curve.
values in the domain of f . Then the area under the curve y = f ( x )
between x = a and x = b is the limit of the sum of the areas of the
rectangular strips, that is, the area is
n −1 6
Recall (see the margin note on p20),
f ( x0 )∆x + f ( x2 )∆x + s + f ( xn−1 )∆x ∑ f ( xi )∆x


lim = lim that the precise definition of continuous
n→∞ n→∞ will not be discussed here. For further
i =0
details, consult Stewart’s Calculus §1.8
where ∆x is the width of the strips. or the Level 2 course 2E Introduction to
Real Analysis.
 26 calculus

2.2 The definite integral

In the previous section we saw that the area under a curve y = f ( x )

between x = a and x = b, A is defined to be the limit of the sum of
the areas of the rectangular strips as the number of strips tends to
infinity. We further introduce some notation for this limit saying
that the area A is the integral from x = a to x = b of f ( x ) with respect Definition
Rb
to x and writing a f ( x ) dx definite integral,
Z b
A= f ( x ) dx. limits of integration
a
where a and b are called the limits of integration.

y
f ( xi )

y = f (x)

a ∆x b x
y
An approximation to the area under the graph of y = f ( x ), called the
Riemann sum estimate of f ( x ), is the sum of the areas of the rectangles.
That is
Z b i = n −1 6

a
f ( x ) dx ≈ ∑ f ( xi ) ∆x = Rn
y = x3 − x
i =0

where ∆x = (b − a) /n, n is the number of rectangles and the points

xi = a + i∆x. The area is of course the limit of the Riemann sum
estimate as n → ∞, that is,
" # " #
Z b i = n −1 i = n −1

a
f ( x ) dx = lim
∆x →0
∑ f ( xi ) ∆x = lim
n→∞
∑ f ( xi ) ∆x . x
i =0 i =0 2

The solutions to other problems in mathematics, physics, statistics, y

etc., are naturally expressed as the limit of a sum and can often be
expressed in integral form.
6
2.2.1 Riemann sum estimates y = x3 − x

Example 2.6 Consider the function f ( x ) = x3 − x whose graph is

sketched in Figure 2.10. Find both

a) the Riemann sum estimate for f ( x ) between x = 0 and x = 2 with

six intervals, and A1 x
Z 2 A2 2
b) f ( x ) dx. Figure 2.10: Riemann sum estimate and
0
definite integral of y = f ( x ) = x3 − x.
 introduction to integration 27

Solution

a) The Riemann sum is expressed in terms of the number of rect-

angles, their common width and their individual heights f ( xi ).
We begin by calculating this data. For six intervals in the region
between 0 and 2, each rectangle will need to be of width

(2 − 0) 1
∆x = = .
6 3

The first rectangle has its leftmost edge at x = 0 and each subse-
quent rectangle is 31 further to the right, hence

1 2 4 5
x0 = 0, x1 = , x2 = , x3 = 1, x4 = , and x5 = .
3 3 3 3

The heights of rectangles 0 to 5 are respectively

1 8 2 10
f ( x0 ) = f (0) = 0, f ( x1 ) = f ( ) = − , f ( x2 ) = f ( ) = − ,
3 27 3 27

4 28 5 80
f ( x3 ) = f (1) = 0, f ( x4 ) = f ( ) = , and f ( x5 ) = f ( ) = .
3 27 3 27

Putting this together, the required Riemann sum estimate is

R6 = f ( x0 ) · ∆x + f ( x1 ) · ∆x + f ( x2 ) · ∆x + f ( x3 ) · ∆x + f ( x4 ) · ∆x + f ( x5 ) · ∆x
1 
= f ( x0 ) + f ( x1 ) + f ( x2 ) + f ( x3 ) + f ( x4 ) + f ( x5 )
3
1 8 10 28 80 
= 0− − +0+ +
3 27 27 27 27
10
= .
9

b) From the definition,

Z 2
( x3 − x ) dx = lim Rn
0 n→∞
 28 calculus

where
n −1
Rn = ∑ f ( xi ) · ∆x
i =0
n −1
= ∑ f ( xi ) · ∆x
i =1
n −1    
2i 2
= ∑ f ·
i =1
n n
!  
n −1   3
2i 2i 2
= ∑ − · Since f ( x0 ) = f (0) = 0, we can omit
i =1
n n n this term and begin the sum at i = 1.
!
2 23 n−1 3 2 n−1
n3 i∑
= · i − ∑i Recall from 1R Algebra §5.2 & §5.4 that
n =1
n i =1 For scalars α, β and γ,
! k
2 23 (n − 1) ((n − 1) + 1) 2 2 (n − 1) ((n − 1) + 1)
 
∑


α ui + β vi + γ
= −
n n3 2 n 2 i =1
k k
2 2 ( n − 1)2 n2 = α ∑ ui + β ∑ vi + kγ ,
 3
( n − 1) n

= 2 −2 i =1 i =1
n n2 22 2 2
k 
k ( k + 1)
2  ∑ i3 =

= 2 2( n − 1)2 − ( n − 1) n i =1
2
n  and
1 1 k
k ( k + 1)
= 2 1−3· +2· 2 . ∑i= .
n n i =1
2

Since both 1
n and 1
n2
−→ 0 as n −→ ∞,

Rn −→ 2 as n −→ ∞.

Hence Z 2
( x3 − x ) dx = lim Rn = 2.
0 n→∞

The value of 10 9 our Riemann sum estimate gave in part (a) is

relatively far from the true value of 2 that we calculated in part (b).
Below in Table 2.5 we show the values of the Riemann sum estimate
Rn for larger numbers of rectangular strips.

Number of strips 6 8 16 32 64 128 256 512 1024

Area estimate, Rn (3 dp) 1.111 1.313 1.641 1.816 1.907 1.953 1.977 1.988 1.994
Table 2.5: Riemann sum estimates for
x3 − x.

2.2.2 Interpreting integrals as area

Observe that this integral of the previous example cannot be inter- Definition
preted as an area as f ( x ) = x3 − x takes negative values. From the net area.
graph of f , Figure 2.10, we can see that f takes negative values on
the interval (0, 1) and positive values on (1, 2]. Hence we can inter-
pret the definite integral as a net area that is the difference A1 − A2 of
the areas A1 above the x-axis and areas A2 below the x-axis. In the
next examples we reverse this thinking, that is, evaluating integrals
by considering net area.
 introduction to integration 29

Z π y
Example 2.7 Evaluate the integral sin x dx.
−π y = sin x

Solution We begin by making a sketch of the sine curve in Fig-

ure 2.11. By the symmetry of the sine curve, we know that the area
− π2 A3 A4
−π x
A3 is the same as the areas A1 , A2 and A4 . Hence the area above the
π π
x-axis is Z π A1 A2 2

sin x dx = 2A3
0
and the area below the x-axis is
Z 0
sin x dx = −2A3
−π Figure 2.11: Area under the sine curve
from −π to π split into 4 areas marked
so that the required integral is zero. That is A1 , A2 , A3 and A4 .

Z π
sin x dx = 0.
−π

Z 4
Example 2.8 Evaluate the integral 6 − 2x dx.
0 In §3.7.1 we develop these ideas on sim-
plifying integrals further with the con-
cept of odd and even functions. The
Solution We sketch y = 6 − 2x in Figure 2.12 and observe that the sine function is an example of an odd
function — it is ‘anti-symmetric’ about
integral between x = 0 and x = 4 of y = 6 − 2x with respect to x the y-axis, that is, sin(− x ) = − sin( x ).
can be interpreted in terms of the difference between two triangular
y
regions, marked T1 and T2 . Calculating the areas of the triangles we
have
1 1 y = 6 − 2x
T1 = · 6 · 3 = 9 and T2 = · 1 · 2 = 1. 6
2 2
Thus we conclude that
Z 4
6 − 2x dx = T1 − T2 = 8.
0 T1
3 4 x
T2

−2
Figure 2.12: Area under curve y = 6 −
2.2.3 Properties of definite integral 2x between 0 and 4 split into 2 areas
marked T1 and T2 .
Basic Fact 2.2 Linearity
Integration is linear. That is,
Z b
Z b Z b
a) f ( x ) + g( x ) dx = f ( x ) dx + g( x ) dx, and
a a a
Z b
Z b
b) k · f ( x ) dx = k · f ( x ) dx for all constants k.
a a
30 calculus

Theorem 2.3 Properties of definite integral

Definite integrals have the following properties.
Z b Z a
a) f ( x ) dx = − f ( x ) dx.
a b
Z a
b) f ( x ) dx = 0.
a
Z b Z c Z b
c) f ( x ) dx = f ( x ) dx + f ( x ) dx.
a a c

d) If f ( x ) ≤ 0 for x ∈ [ a, b] then
Z b
f ( x ) dx = − (shaded area in diagram)
a

Z b
e) (Integral of a constant) For k > 0 a constant k dx = k (b − a),
a
the area of a rectangle whose sides are length k and b − a.

f) (Area between two curves) The area between the curves y =

f ( x ) and y = g( x ) and the lines x = a and x = b is
Z b
f ( x ) − g( x ) dx.
a

Proof

a) In §2.2 we implicitly assumed that a < b in computing the Rie-

mann sum estimate and its limit the integral between a and b
Z b n −1

a
f ( x ) dx = lim
n→∞
∑ f ( xi ) ∆x
i =1

b− a
where ∆x = n . For a < b,
Z a n −1

b
f ( x ) dx = lim
n→∞
∑ f ( xi ) ∆x
i =1

a−b
where ∆x = n < 0. Hence
Z a Z b
f ( x ) dx = − f ( x ) dx.
b a
Z a Z a
b) Take a = b in part (a), then f ( x ) dx = − f ( x ) dx and so
Z a a a
2 f ( x ) dx = 0 hence result.
a
y
c) (For f ( x ) > 0 on [ a, b] and a < c < b) Taking a geometric point of y = f (x)
view it is clear that the area under the curve y = f ( x ) between a
and b is the same as the sum of the areas under the curve between
a and c and between c and b as shown in Figure 2.13.
Z b
a c b x
d) The area between a positive function g( x ) and the x-axis is g( x ) dx,
a Figure 2.13: Area under the curve y =
if f ( x ) ≤ 0 then − f ( x ) ≥ 0 and area between − f ( x ) and the x-axis f (t) between a and c and c and b.
 introduction to integration 31

y
is a b x
Z b Z b
− f ( x ) dx = − f ( x ) dx
a a
by a simple reflection this is equal to the shaded area in Fig-
Z b y = f (x)
ure 2.2.3 and so f ( x ) dx = − (shaded area in diagram).
a
Figure 2.14: Area between y = f ( x ) and
x-axis and between x = a and x = b.
Functions that change sign For example, the situation illustrated
Z b
in Figure 2.2.3, f ( x ) dx = area X − area Y + area Z.
a

e) The area of a rectangle is illustrated in Figure 2.2.3. y

f) The shaded area in Figure 2.2.3 is y = f (x)

Z b Z b Z b X
Z
[ f ( x ) − g( x )] dx = f ( x ) dx − g( x ) dx, a b x
a a a Y
since the shaded area is the area under y = f ( x ) minus the area
under y = g( x ).
Figure 2.15: Functions that change sign.


Example 2.9 Find the area bounded by the curves y = f ( x ), y =

g( x ) and the lines x = 1 and x = 4 given that y
Z 5 Z 4 Z 4
k
f ( x ) dx = 10, f ( x ) dx = 1 and g( x ) dx = 5.
1 5 1
a b x
Figure 2.16: Area of a rectangle.

Solution By Theorem 2.3 (f), the required area is the value of the
integral
Z 4

f ( x ) − g( x ) dx,
1
which by linearity is
y
Z 4 Z 4 y = g( x )
f ( x ) dx − g( x ) dx.
1 1 y = f (x)
Since a x
Z 5 Z 4 Z 5 b
f ( x ) dx = f ( x ) dx + f ( x ) dx Figure 2.17: Area of between two
1 1 4 curves.
by Theorem 2.3 (c) and
Z 5 Z 4
f ( x ) dx = − f ( x ) dx
4 5

by Theorem 2.3 (a) we have that

Z 4 Z 5 Z 5 Z 5 Z 4
f ( x ) dx = f ( x ) dx − f ( x ) dx = f ( x ) dx + f ( x ) dx = 10 + 1 = 11.
1 1 4 1 5
Putting this all together yields an area of
Z 4
Z 4 Z 4
f ( x ) − g( x ) dx = f ( x ) dx − g( x ) dx = 11 − 5 = 6.
1 1 1
 32 calculus

2.3 Fundamental theorems of Integral Calculus (FToC)

Let f be a continuous function that takes positive values on the in- Definition
terval [ a, b] and consider the ‘area-so-far’ function A( x ) ‘area-so-far’ function.
Z x
A( x ) = f (t) dt.
a

This function7 gives for any a 6 x 6 b the area under the curve 7
Observe that f varies only with x!
y = f (t) between a and x. This is illustrated in Figure 2.18

y = f (t)

Z x
A( x ) = f (t) dt
a t
a x b

Figure 2.18: The area-so-far function.

2.3.1 First fundamental theorem of Integral Calculus

Theorem 2.4 The First FToC

With A( x ) defined as above,

A 0 ( x ) = f ( x ).

Proof The difference quotient (DQ) for A( x ) is

A( x + h) − A( x ) B (x)
= h ,
h h

where Bh ( x ) is the area under the curve on [ x, x + h] (Figure 2.3.1).

y

y = f (x)

Z x
A( x ) = f (t) dt Bh ( x )
a x
a x x+h b
Figure 2.3.1: Area under the curve y =
f (t) between t = x and t = x + h.
 introduction to integration 33

Let L( x ) and M ( x ) be the minimum and maximum values8 of F on 8

L( x ) and M( x ) can be shown to exist
[ x, x + h]. As illustrated in Figure 2.19, we see that Bh ( x ) lies between by virtue of continuity, not proved in
1S.
the areas of a rectangle of breadth h, height L( x ) and a rectangle of
breadth h, height M( x ), that is

hL( x ) 6 Bh ( x ) 6 hM( x ),

Bh
and so (as the DQ is )
h

A( x + h) − A( x )
L( x ) 6 6 M ( x ).
h

Since f is continuous, L( x ) and M ( x ) both9 tend to f ( x ) as h → 0.

y = f (x)

M
L

Z x
A( x ) = f (t) dt Bh ( x )
a x
a x x+h b

Figure 2.19: The area Bh ( x ) is squeezed

between the areas of the rectangles with
width h and heights M and L.
Therefore the difference quotient is squeezed between two things 9
Prove this by contradiction.
that both tend to f ( x ) as h → 0; and so

DQ → f ( x ) as h→0

.
Now, as h → 0

A( x + h) − A( x )
→ A0 ( x )
h

d
and so A( x ) = f ( x ) as required. 
dx
 34 calculus

2.3.2 Antidifferentiation

We see from the 1st FToC above (Theorem 2.4) that integration is the
inverse of differentiation. Below we discuss the reverse question.10 10
This has been covered previously in
1R Calculus §4.3 and is included here
as revision.
Question 2.3
Given the derivative of some unknown function y( x ), what can
you say about that function? That is, given that

dy
= f (x)
dx
what can be said about y( x )?

Definition
Defintion 2.10 For a given f ( x ), F ( x ) is the antiderivative of f ( x ) on
antiderivative.
the interval I if
d
F ( x ) = f ( x ) for x ∈ I.
dx

1 3
Example 2.11 The function x is an antiderivative of x2 (on R)
3
since
 
d 1 3 1
x = · 3x2 = x2 .
dx 3 3

1 3 1
Note that x + 4, is also an antiderviative of x2 , as is x3 + c for
3 3
any constant c. In general we have the following theorem.

Theorem 2.5
If F ( x ) is an antiderivative of f ( x ) on the interval I, then F ( x ) + c
is also an antiderivative for any constant c. Any antiderivative
of f ( x ) can be written F ( x ) + d for some constant d.

Proof For the first part, suppose that F ( x ) is an antiderivative of

f ( x ) on an interval I. Then for every constant c, F ( x ) + c is an anti-
derivative of f ( x ) on I, since for x ∈ I,

d d
( F ( x ) + c) = F ( x ) + 0 = f ( x ).
dx dx

For the second part, Suppose G ( x ) is also an antiderivative of f ( x )

on I. Consider the difference G ( x ) − F ( x ). Then

d
( G ( x ) − F ( x )) = G 0 ( x ) − F 0 ( x ) = f ( x ) − f ( x ) = 0.
dx

Hence G ( x ) − F ( x ) is a constant on I; that is, G ( x ) = F ( x ) + d. 

 introduction to integration 35

2.3.3 Second fundamental theorem of Integral Calculus

We begin by introducing some notation. For any function11 F, Notation
h ib
h ib F(x) = F ( b ) − F ( a ).
F(x) = F ( b ) − F ( a ). a
a

11
We could equally have used another
Theorem 2.6 The Second FToC variable instead of x, so
iu=b
With f as defined above in §2.3,
h
F (u) = F ( b ) − F ( a ).
u= a
Z b h ib
f ( x ) dx = F ( x ) = F (b) − F ( a)
a a

where F is an antiderivative of f , that is, F 0 ( x ) = f ( x ).

Proof Let A( x ) be the ‘area-so-far’ function from above, that is

Z x
A( x ) = f (t) dt.
a

Then the first FToC (Theorem 2.4) says that A is an antiderivative of

f , that is A0 ( x ) = f ( x ). Furthermore, Theorem 2.5 tells us that if F
is any other antiderivative on [ a, b] then it differs from A by only a
constant. That is,

F ( x ) = A( x ) + c for some constant c

and we may assume12 that this is true for all x ∈ [ a, b]. Thus F ( a) = 12
This assumption rests on the Mean
A( a) + c and F (b) = A(b) + c. By Theorem 2.3 (b), Value Theorem and the continuity of A
and F.
Z a
A( a) = f (t) dt = 0.
a

Hence
h ib


F(x) = F (b) − F ( a) = A(b) + c − A( a) + c
a
= A(b) − A( a)
= A(b)
Z b
= f (t) dt.
a

Following Examples 2.4, 2.6 and 2.11 we see that power of the Fun-
damental Theorem of Integral Calculus in the following examples.

Example 2.12 Determine

Z 1 Z 2
a) x2 dx, and b) ( x3 − x ) dx.
0 0
 36 calculus

Solution

a) By Example 2.11, 13 x3 is an antiderivative of x2 and so by the 2nd

FToC,
Z 1  1    
2 1 3 1 3 1 3 1
x dx = x = 1 − 0 = .
0 3 0 3 3 3
 
d 1 4 1 2
b) Since x − x = x3 − x, 1 4
4x − 21 x2 is an antiderviative
dx 4 2
of x3 − x and hence

Z 2  2    
1 4 1 2 1 4 1 2 1 4 1 2
( x3 − x ) dx = x − x = 2 − 2 − 0 − 0 = 2.
0 4 2 0 4 2 4 2

2.4 Differentiation and integration as inverse processes

Together the first and second FToC (Theorems 2.4 and 2.6) state that
integration and differentiation are inverse operations to one another.
Rx
That is, a f (t) dt is an antiderivative of f ( x ) and the definite integral
Rb 0
a F ( x ) dx is the difference between the function F ( x ) evaluated at
the limits of integration, F (b) − F ( a). It is not very convenient to talk
in these terms and thus we introduce the indefinite integral.13 13
Or general antiderivative.

2.4.1 The indefinite integral

Defintion 2.13 The indefinite integral of a function f ( x ) is its general Definition

Z
indefinite integral,
antiderivative, denoted f ( x ) dx. Thus
integrand,
Z particular antiderivative,
f ( x ) dx = F ( x ) + c constant of integration.

where F 0 ( x ) = f ( x ) and c is a constant. The symbol “ ” is the inte-

R

gral sign, f ( x ) is the integrand and dx tells us what we are integrating

with respect to.

Example 2.14 The indefinite integral of x2 with respect to x is

1 3
Z
x2 dx = x + c a particular antiderivative
3 the constant of integration
where c is a constant. the integrand

Warning 2.7
Z b
The definite integral f ( x ) dx is a number. However, the in-
Z a
definite integral f ( x ) dx is a function (or family of functions).
 introduction to integration 37

2.4.2 Standard indefinite integrals

Here we begin to investigate some standard integrals (SIs).14 Sum- 14
Some SIs are proved here, others in
marised in Chapter 7, you should learn them by heart and be able to later sections and the remainder in ex-
ercises.
use the results in calculations.

Example 2.15 Find the antiderivatives of x −1 and cos x.

Solution
1
a) The function log x is the antiderivative of on (0, ∞) as log x is
x
d 1
only defined on x ∈ (0, ∞) and log x = for x ∈ (0, ∞).
dx x
1
The function log(− x ) is the antiderivative of on (−∞, 0) because
x
log(− x ) is defined for x ∈ (−∞, 0); that is,

d 1 d −1 1
(log(− x )) = · (− x ) = · −1 =
dx −
|
x dx
{z }
x x
chain rule
In summary15 15
This is SI 8.
1
Z
dx = log | x | + c.
x
d
b) The derivative of sin is (sin x ) = cos x. Hence16 16
Note that in all cases the constant of
dx integration is important!
Z
cos x dx = sin x + c.

This proves SI 3 (SI 2 is similar).

Example 2.16 Verify SI 9,

Z
1 x
√ dx = sin−1 + c, a > 0.
a2 − x 2 a

Solution
d h −1  x i 1 d x 1 1
sin = r · = r ·
dx a x 2 dx a x 2 a
1− 2 1− 2
a a
1
= s 
x2

a2 1 − 2
a
1
= √
a2 − x2
It is enough toverify that the derivative of sin−1 x is as claimed.
x
Let y = sin−1 , so that
a
x d dy 1
sin y = =⇒ (sin y) = cos y = .
a dx dx a
 38 calculus

dy 1 1
Thus = · .
dx a cos y
x2
Note that sin2 y = 2 and cos2 y + sin2 y = 1. Then 1 − cos2 y =
r a
x2 x2
, so cos y = 1 − 2 . Thus
a2 a

dy 1 1 1 1 1 1
= · = ·q = q = √ .
dx a cos y a 1− x2
a2 − a2 x 2 a2 − x2
a2 a2

2.4.3 Linearity of indefinite integrals

Theorem 2.8 Linearity

The statements of Theorem 2.8 should
Integration is linear. That is, be interpreted as addition of equiva-
Z Z Z lence classes.
a) ( f ( x ) + g( x )) dx = f ( x ) dx + g( x ) dx.
Z Z
b) For a constant k, k · f ( x ) dx = k · f ( x ) dx.

Proof

a) Consider,
Z  Z  Z 
d d d
Z
f ( x ) dx + g( x ) dx = f ( x ) dx + g( x ) dx
dx dx dx
= f ( x ) + g( x )
R R
so that f ( x ) dx + g( x ) dx is an antiderivative of f ( x ) + g( x )
and so
Z Z Z
( f ( x ) + g( x )) dx = f ( x ) dx + g( x ) dx.

b) Consider,
 Z  Z   Z
d d d
k f ( x ) dx = k f ( x ) dx + k f ( x ) dx
dx dx dx

by the product rule, so

 Z 
d
k f ( x ) dx = k f ( x )
dx
R
so that k f ( x ) dx is an antiderivative of k f ( x ).


 introduction to integration 39

Example 2.17 Using SI 1, SI 4 and SI 10 (see Chapter 7) In the next examples we use the result
that
a) Integration of a constant 1
Z
x a dx = x a +1 + c (SI 1).
a+1
Check that
Z Z Z
k dx = kx0 dx = k x0 dx = kx + c.  
d 1
x a +1 + c = x a
dx a + 1
recall that
b) Integration of powers of x,
d
( x n ) = nx n−1
Z √ Z Z Z dx
( x4 + 3 x − 5) dx = x4 dx + 3 x1/2 dx − 5 x0 dx so
 
d 1
1 3 3/2 x a +1 + c
= x5 + x − 5x + c dx a+1
5 3/2 1
1 = ( a + 1 ) x a +1−1
a+1
= x5 + 2x3/2 − 5x + c.
5 = xa
as required.
c)
Z     x 
4 1
3 sec2 x − dx = 3 tan x − 4 tan−1 +c
x2 + 4 2 2
x
= 3 tan x − 2 tan−1 + c,
2

using SI 4 and SI 10 with a = 2.

d2 y 4 1
Example 2.18 Given x ∈ (0, ∞) and 2
= 3 − 2 and y0 (1) = 0, This is a example of using antidifferen-
dx x x tiation to write down the solution to an
y(1) = 0, find y( x ).
ordinary differential equation, we will re-
turn to this later in the course.

d 0 4 1
Solution (y ) = 3 − 2 . Thus
dx x x

4 1 x −2 x −1
Z
y0 = − dx = 4 − +c
x3 x2 −2 −1
−2 1
= 2 + + c1 .
x x

−2 1
Since y0 (1) = 0, then y0 (1) =
+ + c1 = 0, so c1 = 1.
12 1
−2 1
Thus, y0 = 2 + + 1. Then
x x

−2 1 −2x −1
Z
y( x ) = + + 1 dx = + log | x | + x + c2 .
x2 x −1

2
Since y(1) = 0, then 0 = + log |1| + 1 + c2 , so c2 = −3. Hence
1

2
y( x ) = + log x + x − 3.
x

Note that x ∈ (0, ∞), so log | x | = log x in this case.

40 calculus

Example 2.19

a) Integration of powers of x,
Z √ Z Z
x ( x2 + 2) dx = x1/2 x2 + 2x1/2 dx = x5/2 + 2x1/2 dx

x7/2 2x3/2
= + +c
7/2 3/2
2 4
= x7/2 + x3/2 + c.
7 3

b) Integration of powers of x,
√ √
(1 − x )2 1−2 x+x
Z  
1 2
Z Z
dx = dx = − √ + 1 dx
x x x x
2x1/2
= log | x | − +x+c
1/2
= log | x | − 4x1/2 + x + c.

c) Integration of powers of x,
2
x3 x −1
Z 
1 1
Z
x+ dx = 2 + x2 + dx = 2x + + +c
x x2 3 −1
x3 1
= 2x + − + c.
3 x

1
Z
d) What is dx? First, note that
1 + sin x
1 1 − sin x 1 − sin x
= =
1 + sin x (1 + sin x )(1 − sin x ) cos2 x
1 sin x
= −
cos2 x cos2 x
= sec2 x − sec x tan x,

1
Z Z
so dx = sec2 x − sec x tan x dx = tan x − sec x + c.
1 + sin x
Standard Integrals
7
You should learn the following standard integrals and be able to
use the results in calculations.
x n +1
Z
SI 1 x n dx = + c; n is a constant, n 6= −1.
n+1
Z
SI 2 sin x dx = − cos x + c.
Z
SI 3 cos x dx = sin x + c.
Z
SI 4 sec2 x dx = tan x + c.
Z
SI 5 cosec2 x dx = − cot x + c.
Z
SI 6 sec x tan x dx = sec x + c.
Z
SI 7 e x dx = e x + c.

1
Z
SI 8 dx = log | x | + c.
x
Z
1 x
SI 9 √ dx = sin−1
+ c, a > 0.
a2 − x 2 a
1 1 −1 x
Z  
SI 10 dx = tan + c, a is a positive constant.
x 2 + a2 a a
Z
SI 11 sec x dx = log | sec x + tan x | + c.

f 0 (x)
Z
SI 12 dx = log | f ( x )| + c.
f (x)
Z
SI 13 cot x dx = log | sin x | + c.
Z
SI 14 tan x dx = log | sec x | + c.

1
Z  p 
SI 15 √ dx = log  x + x2 + k + c.
x2 + k