7.

1 The Riemann Integral

We begin with the definition of a partition.

Def. 7.1.1: Partitions

Let a, b ∈ R and suppose a < b . By a partition of the interval [a, b] we mean a

collection of intervals

P = {[x 0 , x 1 ], [x 1 , x 2 ], … , [x n−1 , x n ]}

such that a = x0 < x1 < x2 < ⋯ < xn = b and where n ∈ N .

Hence, a partition P defines a finite collection of non-overlapping intervals

I k = [x k−1 , x k ], where k = 1, … , n. The norm of a partition P is defined as

∥P∥ = max{x 1 − x 0 , x 2 − x 1 , … , x n − x n−1 }.

In other words, ∥P∥ is the maximum length of the intervals in P . To ease our notation,
we will denote a partition as P = {[x k−1 , x k ]} n k=1
. Let P = {[x k−1 , x k ]} n
k=1
be a
partition of [a, b]. If t k ∈ I k = [x k−1 , x k ] then we say that t k is a sample of I k and the
set of ordered pairs

n
Ṗ = {([x k−1 , x k ], t k )}
k=1

will be called a sampled partition.

Example 7.1.2
Examples of sampled partitions are mid-points, right-end points, and left-end points
partitions.

Now consider a function f : [a, b] → R and let Ṗ = {([x k−1 , x k ], t k )}

n

k=1
be a
sampled partition of the interval [a, b]. The Riemann sum of f corresponding to P
˙ is the

number
 n

S(f ; Ṗ) = ∑ f (t k )(x k − x k−1 ).

k=1

When f (x) > 0 on the interval [a, b], the Riemann sum S(f ; Ṗ) is the sum of the areas
of the rectangles with height f (t k ) and width (x k − x k−1 ). We now define the notion of
Riemann integrability.

Def. 7.1.3: Riemann Integrability

The function f : [a, b] → R is said to be Riemann integrable if there exists a number

L ∈ R such that for every ε > 0 there exists δ > 0 such that for any sampled

partition Ṗ that satisfies ∥Ṗ∥ < δ it holds that |S(f ; Ṗ) − L| < ε.

The set of all Riemann integrable functions on the interval [a, b] will be denoted by
R[a, b].

Thm. 7.1.4

If f ∈ R[a, b] then the number L in the definition of Riemann integrability is unique.

Proof: Let L 1 and L 2 be two real numbers satisfying the definition of Riemann
integrability and let ε > 0be arbitrary. Then there exists δ > 0 such that
|S(f ; Ṗ) − L 1 | < ε/2 and |S(f ; Ṗ) − L 2 | < ε/2, for all sampled partitions Ṗ with

˙
∥P∥ < δ . Then, if ∥P∥
˙
< δ it holds that

|L 1 − L 2 | ≤ |S(f ; Ṗ) − L 1 | + |S(f ; Ṗ) − L 2 |

< ε.

By Theorem 2.2.7 this proves that L 1 = L2 . QED

If f ∈ R[a, b] , we call the number L the integral of f over [a, b] and we denote it by

L = ∫ f
a

Example 7.1.5
Show that a constant function on [a, b] is Riemann integrable.

Proof: Let f : [a, b] :→ R be such that f (x) = C for all x ∈ [a, b] and let
Ṗ = {([x k−1 , x k ], t k )} be a sampled partition of [a, b]. Then
 n

S(f ; Ṗ) = ∑ f (t k )(x k − x k−1 )

k=1

= C ∑(x k − x k−1 )

k=1

= C(x n − x 0 )

= C(b − a).

Hence, with L = C(b − a) , we obtain that |S(f ; Ṗ) − L| = 0 < ε for any ε > 0 and
b
therefore ∫ a f = C(b − a) . This proves that f is Riemann integrable. QED

Example 7.1.6
Prove that f (x) = x is Riemann integrable on [a, b].

Proof: We consider the special case that [a, b] , the general case is similar. Let
= [0, 1]

˙
Q = {([x k−1 , x k ], q k )} be a sampled partition of [0, 1] chosen so that
+ x k−1 ), i.e., q k is the midpoint of the interval [x k−1 , x k ]. Then
1
qk = (x k
2

˙
S(f ; Q) = ∑ f (q k )(x k − x k−1 )

k=1

1
= ∑ (x k + x k−1 )(x k − x k−1 )
2

k=1

n
1 2 2
= ∑(x k − x k−1 )
2
k=1

1 2 2
= (x − x )
n 0
2

1 2 2
= (1 − 0 )
2

1
= .
2
Now let Ṗ
suppose that ∥Ṗ∥
˙
= {([x k−1 , x k ]), t k }

Q = {([x k−1 , x k ], q k )}

|t k − q k | < δ
n

. Therefore,

The next result shows that if f

[a, b]

x ∈
< δ

˙
|S(f ; P) − S(f ; Q)| =

does not affect the value of ∫ a

Thm. 7.1.7

Let f ∈ R[a, b]

First, if c
c = tk ∉

If c
≠ tk

= t k = t k−1
= ∫
a

Ṗ = {([x k−1 , x k ], t k )}

and c
f

≠ xk
k=1

{x 0 , x 1 , … , x n } for some k then

n

k=1

, so that (x k

Hence, we have proved that for arbitrary Ṗ that satisfies ∥Ṗ∥

˙
|S(f ; P) − 1/2| < δ

|S(f ; Ṗ) − 1/2| < ε

. Hence, given ε
. Therefore, ∫ 0
1

∈ R[a, b]

˙
b
∣
be an arbitrary sampled partition of [0, 1] and
− x k−1 ) ≤ δ

f
k=1

n
n

.
2
for all k

≤ ∑ |t k − q k |(x k − x k−1 )

k=1

< δ(1 − 0)

= δ.

f =
> 0 we let δ
1
. QED
= ε
= 1, 2, … , n

is the corresponding midpoint sampled partition then

∑ t k (x k − x k−1 ) − q k (x k − x k−1 )

< δ
. If

it holds that
and then if ∥P∥
˙
< δ

then changing f at a finite number of points in

and let g : [a, b] → R be a function such that g(x) = f (x) for all
[a, b] except possibly at a finite number of points in [a, b]. Then g ∈ R[a, b] and

in fact ∫ a

Proof: Let L
b
g = ∫
a

b
b
f .

. Suppose that g(x) = f (x)

S(g; Ṗ) − S(f ; Ṗ) = (f (c) − g(c))(x k − x k−1 ).

for some k then necessarily c

except at one point x
be a sampled partition. We consider mutually exclusive cases.
for all k then S(f ; P)

= x k−1
˙
= S(g; P) . If

and then
= c
then

. Let
 S(g; Ṗ) − S(f ; Ṗ) = (f (c) − g(c))(x k − x k−1 ) + (f (c) − g(c))(x k−1 − x k−2 ).

Hence, in any case, by the triangle inequality

˙ ˙ ˙
|S(g; P) − S(f ; P)| ≤ 2(|f (c)| + |g(c)|)∥P∥

= M ∥Ṗ∥.

where M = 2(|f (c)| + |g(c)|) . Let ε > 0 be arbitrary. Then there exists δ 1 > 0 such
that |S(f ; P)
˙
− L| < ε/2 for all partitions P
˙ such that ˙
∥P∥ < δ1 . Let
δ = min{δ 1 , ε/(2M )} . Then if ∥Ṗ∥ < δ then

∥S(g; f ) − L∥ ≤ |S(g; Ṗ) − S(f ; Ṗ)| + |S(f ; Ṗ) − L|

< M ∥Ṗ| + ε/2

< M ε/(2M ) + ε/2

= ε.

b b
This proves that g ∈ R[a, b] and ∫ a g = L = ∫ a f . Now suppose by induction that if
g(x) = f (x) for all x ∈ [a, b] except at a j ≥ 1 number of points in [a, b] then
b b
g ∈ R[a, b] and ∫ a g = ∫ a f . Now suppose that h : [a, b] → R is such that
h(x) = f (x) for all x ∈ [a, b] except at the points c 1 , c 2 , … , c j , c j+1 . Define the

function g by g(x) = h(x) for all x ∈ [a, b] except at x = c j+1 and define
g(c j+1 ) = f (c j+1 ). Then g and f differ at the points c 1 , … , c j . Then by the induction
b b
hypothesis, g ∈ R[a, b] and ∫ a g = ∫
a
f . Now g and h differ at the point c j+1 and
b b b
therefore h ∈ R[a, b] and ∫ a h = ∫
a
g = ∫
a
f . This ends the proof. QED

We now state some properties of the Riemann integral.

Thm. 7.1.8: Properties of the Riemann Integral

Suppose that f , g ∈ R[a, b] . The following hold.

b b
(i) If k ∈ R then (kf ) ∈ R[a, b] and ∫ a kf = k ∫
a
f .
b b b
(ii) (f + g) ∈ R[a, b] and ∫ a (f + g) = ∫
a
f + ∫
a
.
g

b b
(iii) If f (x) ≤ g(x) for all x ∈ [a, b] then ∫ a f ≤ ∫
a
g .
 ∣
Proof: If k
k ≠ 0 . Let ε
˙
|S(f ; P) − ∫

there exists δ
whenever ∥Ṗ∥
= 0

> 0
b

S(kf ; Ṗ) = kS(f ; Ṗ)

> 0

S(f + g; Ṗ) − (∫

To prove (c), let ε

then (kf )(x)

f ∥ < ε/|k|

< δ

> 0

Now, by assumption, S(f ; Ṗ)

Therefore,

Real Analysis
Cesar O. Aguilar

7. Riemann Integration
∫
a
S(kf ; Ṗ) − k ∫

To prove (b), it is easy to see that S (f

b
a

∫
a
f + ∫

b
= 0

be given. Then there exists δ

such that |S(f ; Ṗ) − ∫ a

. Therefore, if ∥Ṗ∥

be arbitrary and let δ

∫
a
b
for all x and then clearly ∫ a

. Now for any partition P

. Therefore, if ∥Ṗ∥

g)
a
b

< δ

f − ε/2 < S(f ; Ṗ) < ∫

g − ε/2 < S(g; Ṗ) < ∫

≤ S(g; Ṗ)

f − ε/2 < S(f ; Ṗ) ≤ S(g; Ṗ) < ∫

∫
a
b
f

< ε.

f < ∫
< δ

> 0
> 0

˙ , it holds that

then
such that if ∥Ṗ∥

= |k| S(f ; Ṗ) − ∫

< |k|(ε/|k|)

= ε.

+ g; Ṗ) = S(f ; Ṗ) + S(g; Ṗ)

b
f | < ε/2

we have that

S(f ; Ṗ) − ∫

and therefore

a
b
a

g + ε.
b
a
b

be such that if ∥Ṗ∥

f + ε/2

g + ε/2

a
b

b
kf = 0

and |S(g; Ṗ) − ∫ a

+
b

g + ε/2.
f
, so assume that
< δ then

. Given ε
b

S(g; Ṗ) − ∫

< δ then
> 0

g| < ε/2

a
b

g
,
7. Riemann Integration

that ∫ a
b
f ≤ ∫

|S(f ; Ṗ) − L| < 1

b

a
g. QED

∣
Since ε is arbitrary, we can choose ε n

Thm. 7.1.9: Integrable Functions are Bounded

If f ∈ R[a, b]

Proof: Let f ∈ R[a, b]

t k = x k for k ≠ j and t j is such that

|S(f ; Ṗ)| =
= 1/n

then f is bounded on [a, b].

and put L = ∫
b

a
f
and then passing to the limit we deduce

Properties (i), (ii), and (iii) in Theorem 7.1.8 are known as homogeneity, additivity, and
monotonicity, respectively. We now give a necessary condition for Riemann integrability.

. There exists δ > 0 such that if ∥Ṗ∥

and therefore |S(f ; Ṗ)| < |L| + 1. Suppose by contradiction that
f is unbounded on [a, b]. Let P be a partition of [a, b], with sets I 1 , … , I n , and with

∥P∥ < δ. Then f is unbounded on some I j , i.e., for any M > 0 there exists

x ∈ I j = [x j−1 , x j ] such that f (x) > M . Choose samples in P by asking that

|f (t j )(x j − x j−1 )| > |L| + 1 +

Therefore, (using |a| = |a + b − b| ≤ |a + b| + |b|

∑ f (t k )(x k − x k−1 ) .

k≠j

implies that |a + b|

f (t j )(x j − x j−1 ) + ∑ f (t k )(x k − x k−1 )

≥ |f (t j )(x j − x j−1 )| −

> |L| + 1.

This is a contradiction and thus f is bounded on [a, b]. QED

Example 7.1.10 (Thomae)

k≠j

∑ f (t k )(x k − x k−1 )

k≠j

Consider Thomae's function h : [0, 1] → R defined as h(x) = 0 if x is irrational and

h(m/n) = 1/n for every rational m/n ∈ [0, 1], where gcd(m, n) = 1. In Example

5.1.7, we proved that h is continuous at every irrational but discontinuous at every

rational. Prove that h is Riemann integrable.
< δ

≥ |a| − |b|
then

)
Proof: Let ε > 0 be arbitrary and let E = {x ∈ [0, 1] : h(x) ≥ ε/2} . By definition of
h, the set E is finite, say consisting of n elements. Let δ = ε/(4n) and let Ṗ be a
sampled partition of [0, 1] with ∥P∥
˙
< δ . We can separate the partition P
˙
into two
sampled partitions Ṗ 1 and Ṗ 2 where Ṗ 1 has samples in the the set E and Ṗ 2 has no
samples in E . Then the number of intervals in Ṗ 1 can be at most 2n, which occurs when
all the elements of E are samples and they are at the endpoints of the subintervals of
Ṗ 1 . Therefore, the total length of the intervals in Ṗ 1 can be at most 2nδ = ε/2. Now

0 < h(t k ) ≤ 1 for every sample t k in Ṗ 1 and therefore S(f ; Ṗ 1 ) ≤ 2nδ = ε/2 . For
samples t k in Ṗ 2 we have that h(t k ) < ε/2 . Therefore, since the sum of the lengths of
the subintervals of Ṗ 2 is ≤ 1, it follows that S(f ; Ṗ 2 ) ≤ ε/2 . Hence
b
0 ≤ S(f ; Ṗ) = S(f ; Ṗ 1 ) + S(f ; Ṗ 2 ) < ε . Thus ∫ a h = 0 . QED

Exercises

Exercise 7.1.1
Suppose that f , g ∈ R[a, b] and let α, β ∈ R . Prove by definition that
(αf + βg) ∈ R[a, b] .

Exercise 7.1.2
If f is Riemann integrable on [a, b] and |f (x)| ≤ M for all x ∈ [a, b] , prove that
b
|∫
a
. Hint: The inequality |f (x)| ≤ M is equivalent to
f | ≤ M (b − a)

−M ≤ f (x) ≤ M . Then use the fact that constants functions are Riemann integrable

whose integrals are easily computed. Finally, apply a theorem from this section.

Exercise 7.1.3
If f is Riemann integrable on [a, b] and (Ṗ n ) is a sequence of tagged partitions of [a, b]
such that ∥P
˙
n∥ → 0 prove that

∫ f = lim S(f ; Ṗ n )
n→∞
a

Hint: For each n ∈ N we have the real number s n ˙

= S(f ; P n ) , and we therefore have a
b
sequence (s n ). Let L = ∫
a
f . We therefore want to prove that lim s n = L .
n→∞

Exercise 7.1.4
Give an example of a function f : [0, 1] → R that is Riemann integrable on [c, 1] for
every c ∈ (0, 1) but which is not Riemann integrable on [0, 1]. Hint: What is a necessary
condition for Riemann integrability?

7.2 Riemann Integrable Functions

To ease our notation, if I is a bounded interval with end-points a < b we denote by μ(I )
the length of I , that is μ(I ) = b − a. Hence, if I = [a, b], I = [a, b), I = (a, b], or
I = (a, b) then μ(I ) = b − a. Thus far, to establish the Riemann integrability of f , we
b
computed a candidate integral L and showed that in fact L = ∫ a f . The following
theorem is useful when a candidate integral L is unknown. The proof is omitted.

Thm. 7.2.1: Cauchy Criterion

A function f : [a, b] → R is Riemann integrable if and only if for every ε > 0 there
exists δ > 0 such that if Ṗ and Q
˙
are sampled partitions of [a, b] with norm less than
δ then

˙
|S(f ; Ṗ) − S(f ; Q)| < ε.

Using the Cauchy Criterion, we show next that the Dirichlet function is not Riemann
integrable.

Example 7.2.2 (Non-Riemann integrable function)

Let f : [0, 1] → R be defined as f (x) = 1 if x is rational and f (x) = 0 if x is
irrational. Show that f is not Riemann integrable.

Proof: To show that f is not in R[0, 1], we must show that there exists ε 0 > 0 such that
for all δ > 0 there exists sampled partitions Ṗ and Q
˙
with norm less than δ but
˙ ˙
|S(f ; P) − S(f ; Q)| ≥ ε 0 . To that end, let ε 0 = 1/2, and let δ > 0 be arbitrary. Let n
be sufficiently large so that 1/n < δ . Let Ṗ be a sampled partition of [0, 1] with
intervals all of equal length 1/n < δ and let the samples of Ṗ be rational numbers.
Similarly, let Q
˙
be a partition of [0, 1] with intervals all of equal length 1/n and with
samples irrational numbers. Then S(f ; Ṗ) = 1 and S(f ; Q)
˙
= 0 , and therefore
˙ ˙
|S(f ; Q) − S(f ; Q)| ≥ ε 0 . QED
We now state a sort of squeeze theorem for integration.

Thm. 7.2.3: Squeeze Theorem

Let f be a function on [a, b]. Then f ∈ R[a, b] if and only if for every ε > 0 there
exist functions α and β in R[a, b] with α(x) ≤ f (x) ≤ β(x) for all x ∈ [a, b] and
b
∫
a
(β − α) < ε .

b
Proof: If f ∈ R[a, b]then let α(x) = β(x) = f (x). Then clearly ∫ a (β − α) = 0 < ε
for all ε > 0. Now suppose the converse and let ε > 0 be arbitrary. Let α and β satisfy
b
the conditions of the theorem, with ∫ a (β − α) . Now, there exists δ such that
ε
< > 0
3

if ∥P∥
˙
< δ then

b b
ε ε
˙
∫ α − < S(α; P) < ∫ α +
a
3 a
3

and

b b
ε ε
˙
∫ β − < S(β; P) < ∫ β + .
a
3 a
3

For any sampled partition P

˙ it holds that ˙
S(α; P)
˙ ˙
≤ S(f ; P) ≤ S(β; P) , and therefore

b b
ε ε
˙ (1)
∫ α − < S(f ; P) < ∫ β + .
a
3 a
3

If Q
˙
is another sampled partition with ∥Q∥
˙
< δ then also

b b
ε ε
˙ (2)
∫ α − < S(f ; Q) < ∫ β + .
a
3 a
3

Subtracting the two inequalities (???)-(2), we deduce that

b b
ε ε
˙
−∫ (β − α) − 2 < S(f ; Ṗ) − S(f ; Q) < ∫ (β − α) + 2 .
a
3 a
3

b
Therefore, since ∫ a (β − α) it follows that
ε
<
3
 ˙ ˙
−ε < S(f ; P) − S(f ; Q) < ε.

By the Cauchy criterion, this proves that f ∈ R[a, b] . QED

Step-functions, defined below, play an important role in integration theory.

Def. 7.2.4

A function s : [a, b] → R is called a step-function on [a, b] if there is a finite number

n
of disjoint intervals I 1 , I 2 , … , I n contained in [a, b] such that [a, b] = ⋃ k=1 I k and
such that s is constant on each interval.

In the definition of a step-function, the intervals I k may be of any form, i.e., half-closed,
open, or closed.

Lem. 7.2.5

Let J be a subinterval of [a, b] and define φ J on [a, b] as φ J (x) = 1 if x ∈ J and

b
φ J (x) = 0 otherwise. Then φ J ∈ R[a, b] and ∫ a φ J = μ(J ) .

Thm. 7.2.6

If φ : [a, b] → R is a step function then φ ∈ R[a, b] .

Proof: Let I 1 , … , I n be the intervals where φ is constant, and let c 1 , … , c n be the

constant values taken by φ on the intervals I 1 , … , I n , respectively. Then it is not hard
n
to see that φ = ∑ k=1 c k φ I . Then φ is the sum of Riemann integrable functions and
k

b n
therefore is also Riemann integrable. Moreover, ∫ a φ = ∑
k=1
c k μ(I k ). QED

We will now show that any continuous function on [a, b] is Riemann integrable. To do
that we will need the following.

Lem. 7.2.7: Continuity and Step-Functions

Let f : [a, b] → R be a continuous function. Then for every ε > 0 there exists a step-
function s : [a, b] → R such that |f (x) − s(x)| < ε for all x ∈ [a, b].

Proof: Let ε be arbitrary. Since f is uniformly continuous on [a, b] there exists

> 0

δ > 0 such that if |x − u| < δ then |f (x) − f (u)| < ε. Let n ∈ N be sufficiently large

so that (b − a)/n < δ. Partition [a, b] into n subintervals of equal length (b − a)/n,
and denote them by I 1 , I 2 , … , I n , where I 1 = [x 0 , x 1 ] and I k = (x k−1 , x k ] for
1 < k ≤ n . Then for x, u ∈ I k it holds that |f (x) − f (u)| < ε. For x ∈ I k define
s(x) = f (x k ). Therefore, for any x ∈ I k it holds that
n
|f (x) − s(x)| = |f (x) − f (x k )| < ε. Since ⋃ I k = [a, b], it holds that
k=1

|f (x) − s(x)| < ε for all x ∈ [a, b]. QED

We now prove that continuous functions are integrable.

Thm. 7.2.8: Continuous Functions are Integrable

A continuous function on [a, b] is Riemann integrable on [a, b].

Proof: Suppose that f : [a, b] → R is continuous. Let ε > 0 be arbitrary and let
~
ε = (ε/4)/(b − a) . Then there exists a step-function s : [a, b] → R such that
~
|f (x) − s(x)| < ε for all x ∈ [a, b]. In other words, for all x ∈ [a, b] it holds that

~ ~
s(x) − ε < f (x) < s(x) + ε.

~ ~
The functions α(x) := s(x) − ε and β(x) := s(x) + ε are Riemann integrable
b
~
integrable on [a, b], and ∫ a (β − α) = 2ε(b − a) = ε/2 < ε . Hence, by the Cauchy
criterion, f is Riemann integrable. QED

Recall that a function is called monotone if it is decreasing or increasing.

Thm. 7.2.9: Monotone Functions are Integrable

A monotone function on [a, b] is Riemann integrable on [a, b].

Proof: Assume that f : [a, b] → R is increasing and that M = f (b) − f (a) > 0 (if
M = 0 then f is the zero function which is clearly integrable). Let ε > 0 be arbitrary.
M (b−a)
Let n ∈ N be such that n
. Partition [a, b] into subintervals of equal length
< ε

(b−a)
Δx =
n
, and as usual let a = x 0 < x 1 < ⋯ < x n−1 < x n = b denote the
resulting points of the partition. On each subinterval [x k−1 , x k ], it holds that
f (x k−1 ) ≤ f (x) ≤ f (x k ) for all x ∈ [x k−1 , x k ] since f is increasing. Let

α : [a, b] → R be the step-function whose constant value on the interval [x k−1 , x k ) is

f (x k−1 ) and similarly let β : [a, b] → R be the step-function whose constant value on

the interval [x k−1 , x k ) is f (x k ), for all k = 1, … , n. Then α(x) ≤ f (x) ≤ β(x) for
all x ∈ [a, b]. Both α and β are Riemann integrable and
 b n

∫ (β − α) = ∑[f (x k ) − f (x k−1 )]Δx

a
k=1

= (f (x n ) − f (x n−1 )Δx

(b − a)
= M
n

< ε.

Hence by the Squeeze theorem for integrals (Theorem 7.2.3), f ∈ R[a, b] . QED

Our las theorem is the additivity property of the integral, the proof is omitted.

Thm. 7.2.10: Additivity Property

Let f : [a, b] → R be a function and let c ∈ (a, b). Then f ∈ R[a, b] if and only if its
restrictions to [a, c] and [c, b] are both Riemann integrable. In this case,

b c b

∫ f = ∫ f + ∫ f
a a c

Exercises

Exercise 7.2.1
Suppose that f : [a, b] → R is continuous and assume that f (x) > 0 for all x .
∈ [a, b]
b
Prove that ∫ a f > 0. Hint: A continuous function on a closed and bounded interval
achieves its minimum value.

Exercise 7.2.2
Suppose that f is continuous on [a, b] and that f (x) ≥ 0 for all x ∈ [a, b] .
b
(a) Prove that if ∫ a f = 0 then necessarily f (x) = 0 for all x ∈ [a, b] .

(b) Show by example that if we drop the assumption that f is continuous on [a, b] then
it may not longer hold that f (x) = 0 for all x ∈ [a, b].

Exercise 7.2.3
Show that if f : [a, b] → R is Riemann integrable then |f | : [a, b] → R is also Riemann
integrable.

7.3 The Fundamental Theorem of Calculus

Thm. 7.3.1: FTC Part I

Let f : [a, b] → R be a function. Suppose that there exists a finite set E ⊂ [a, b] and
a function F : [a, b] → R such that F is continuous on [a, b] and F ′ (x) = f (x) for
b
all x ∈ [a, b]∖
E . If f is Riemann integrable then ∫ a f = F (b) − F (a) .

Proof: Assume for simplicity that E := {a, b} . Let ε > 0 be arbitrary. Then there exists
b
δ > 0 such that if ∥Ṗ∥ < ε then |S(f ; Ṗ) − ∫ a f | < ε. For any Ṗ , with intervals
I k = [x k−1 , x k ] for k = 1, 2, … , n, there exists, by the Mean Value Theorem applied

to F on I k , a point u k ∈ (x k−1 , x k ) such that

F (x k ) − F (x k−1 ) = F (u k )(x k − x k−1 ). Therefore,
′

F (b) − F (a) = ∑ F (x k ) − F (x k−1 )

k=1

= ∑ f (u k )(x k − x k−1 )

k=1

˙
= S(f ; P u )

where Ṗ u has the same intervals as Ṗ but with samples u k . Therefore, if ∥Ṗ∥ < δ then

b b

F (b) − F (a) − ∫ f = S(f ; Ṗ u ) − ∫ f

a a

< ε.

b
Hence, for any ε we have that F (b) − F (a) − ∫
a
f < ε and this shows that
b
∫
a
f = F (b) − F (a) . QED

Def. 7.3.2: Indefinite Integral

 Let f

for x

for all x

QED
∈ R[a, b]

F : [a, b] →

∈ [a, b]

Thm. 7.3.3
.

∈ [w, z]
R defined by

Under the additional hypothesis that f

f is differentiable.

Thm. 7.3.4: FTC Part 2

z
F (x) := ∫

Let f ∈ R[a, b]. The indefinite integral F : [a, b] → R of f

Lipschitz function on [a, b], and thus continuous on [a, b].

Proof: For any w, z ∈ [a, b] such that w

F (z) = ∫

= ∫
≤ z

a
z

= F (w) + ∫

≤ ∫

|F (z) − F (w)| =

∈ R[a, b]
w
a

∣
. The indefinite integral of f with basepoint a is the function

it holds that

f + ∫

f ≤ K(z − w)

≤ K|z − w|
w
z

w
z

z
f

and therefore F (z) − F (w) = ∫ w f . Since f is Riemann integrable on [a, b] it is

bounded and therefore |f (x)| ≤ K for all x ∈ [a, b]. In particular, −K ≤ f (x) ≤
and thus −K(z − w)
z

f
: [a, b] → R

, and thus
is a

is continuous, the indefinite integral of

 Let f ∈ R[a, b] and let f be continuous at a point c ∈ [a, b]. Then the indefinite
integral F of f is differentiable at c and F ′ (c) = f (c).

7.4 Riemann-Lebesgue Theorem

In this section we present a complete characterization of Riemann integrability for a

bounded function. Roughly speaking, a bounded function is Riemann integrable if the set
of points were it is discontinuous is not too large. We first begin with a definition of “not
too large”.

Def. 7.4.1

A set E ⊂ R is said to be of measure zero if for every ε > 0 there exists a countable
collection of open intervals I k such that

∞ ∞

E ⊂ ⋃ I k and ∑ μ(I k ) < ε.

k=1 k=1

Example 7.4.2
Show that a subset of a set of measure zero also has measure zero. Show that the union
of two sets of measure zero is a set of measure zero.

Example 7.4.3
Let S ⊂ R be a countable set. Show that S has measure zero.

Solution: Let S = {s 1 , s 2 , s 3 , …} . Consider the interval

ε ε
I k = (s k − , sk + ).
k+1 k+1
2 2

Clearly, s k ∈ Ik and thus S ⊂ ⋃ Ik . Moreover,

∞ ∞
ε
∑ μ(I k ) = ∑ = ε.
k
2
k=1 k=1

As a corollary, Q has measure zero. ∎

However, there exists uncountable sets of measure zero.

Example 7.4.4
The Cantor set is defined as follows. Start with I 0 = [0, 1] and remove the middle third
J1 = (
1

3
yielding the set I 1 = I 0 ∖J 1 = [0, 3 ] ∪ [ 3 , 1]. Notice that μ(J 1 )
,
2

3
)
1 2
=
1

3
.
Now remove from each subinterval of I 1 the middle third resulting in the set

1 2 7 8 1 2 3 6 7 8
I 2 = I 1 ∖(( , ) ∪ ( , )) = [0, ] ∪ [ , ] ∪ [ , ] ∪ [ , 1]
9 9 9 9 9 9 9 9 9 9

The two middle thirds J 2 = ( 9 , 9 ) ∪ ( 9 , 9 ) removed have total length μ(J 2 ) = 2 9 .

1 2 7 8 1

By induction, having constructed I n which consists of the union of 2 n closed

subintervals of [0, 1], we remove from each subinterval of I n the middle third resulting in
the set I n+1 = I n ∖J n+1 , where J n+1 is the union of the 2 n middle third open intervals
and I n+1 now consists of 2 n+1 disjoint closed-subintervals. By induction, the total
n

length of J n+1 is μ(J n+1 ) = 32 . The Cantor set is defined as

n+1

C = ⋂ In .

n=1

We now state the Riemann-Lebesgue theorem.

Thm. 7.4.5: Riemann-Lebesgue

Let f : [a, b] → R be a bounded function. Then f is Riemann integrable if and only if

the points of discontinuity of f forms a set of measure zero.