UNIT II

1 Monotonic Functions
Definition 1.1. Let f : I → R be a real valued function.

(i) Let x, y ∈ I such that x < y =⇒ f (x) ≤ f (y), then f is called monotonically increasing.

(ii) Let x, y ∈ I such that x < y =⇒ f (x) ≥ f (y), then f is called monotonically decreasing.

Properties :

(i) A function is called monotonic if it is either monotonically increasing or monotonically

decreasing.

(ii) If f is monotonically increasing then f −1 is monotonically decreasing.

(iii) If f is differentiable on (a, b), then

(a) f is monotonically increasing if

f ′ (x) ≥ 0, ∀x ∈ (a, b).

(b) f is monotonically decreasing if

f ′ (x) ≤ 0, ∀x ∈ (a, b).

(iv) If f is monotonic on [a, b] then f is bounded on [a, b].

(a) If f is monotonically increasing on [a,b] then

f (a) ≤ f (x) ≤ f (b), ∀x ∈ [a, b]

(b) If f is monotonically decreasing on [a,b] then

f (b) ≤ f (x) ≤ f (a), ∀x ∈ [a, b]

(v) Sum and product of two monotonic functions need not be a monotonic function.

(vi) If f ′ (x) exists and is continuous on I and f ′ (x) ̸= 0 at all interior points of I then f is
either increasing or decreasing.

(vii) If f is differentiable on I then f is measurable on I.

(viii) Monotonic(decreasing or increasing) function on I has a finite derivative almost every-

where in I.

1
2 Functions of Bounded Variation
Definition 2.1. Let f : [a, b] → R be a real valued function and P = {x0 , x1 , · · · , xn } be a
partition of [a, b] with each partition of P of [a, b], we associate a sum
n
X
|f (xi ) − f (xi−1 )| ,
i=1

and the sum is called Variation of f over [a, b] with respect to the partition P of [a, b] and is
b
denoted by V(P, f ).
a
Definition 2.2. A non-negative extended real valued function
( )
b b
V(f ) = sup V(P, f ), P is partition of [a, b]
a P a

is called total Variation of f .

b
Definition 2.3. A function f is called a function of bounded variation if V(f ) < +∞ and f
a
b
is not a function of bounded variation if V(f ) = +∞.
a
The set of functions of bounded variation defined on a set [a, b] is denoted by BV [a, b].
Remarks :
b
(i) Let f be a function of bounded variation on [a, b]. If V(f ) = 0 then f is a constant
a
function.

(ii) A polynomial function is function of bounded variation on any finite interval.

(iii) Some examples :

(a) sin x and cos x are examples of functions of bounded variation on a finite interval.
(b) Let P = {x0 , x1 , · · · , xn } be a partition of [a, b]
(
b 2, if x = 0 is partitioning point of P
V(P, f ) =
a 0, if x = 0 is not partitioning point of P
( )
b b
Therefore, V(f ) = sup V(P, f )
a P a
= sup{0, 2} = 2 < +∞
P
=⇒ f ∈ BV [a, b].

2
 (c) Let f be a function defined by
(
1, x ∈ Q
f (x) =
0, x ∈
/Q
b
In this case, V(f ) = +∞,
a
which means f is not a function of bounded variation.

(iv) A function of bounded variation is always bounded. However, the converse is not always
true, e.g., (
1 x∈Q
f (x) =
0 x∈ /Q

(v) A continuous function need not be a function of bounded variation.

(vi) A step function is a function of bounded variation.
(vii) A function of bounded variation need not be continuous.
(viii) If f is continuous on [a, b] and f ′ exists on [a, b] and is bounded on (a, b), i.e.,

|f ′ (x)| ≤ M, M > 0, ∀x ∈ (a, b),

then f is a function of bounded variation.

(ix) A monotonic function (increasing or decreasing) f : [a, b] → R is a function of bounded
variation and
b
V(f ) = |f (b) − f (a)|
a
.
Example : f : [0, 1] → R defined as
(
1


x2 cos x
, x ̸= 0
f (x) =
0, x=0

(i) f is continuous.
As, f is continuous when x ̸= 0. For the case when x = 0,
Left Hand Limit (LHL) :

lim (0 + h) = lim (0 + h)2 cos(1/h)

h→0 h→0
= lim h2 cos(1/h) = 0
h→0

Right Hand Limit (RHL) :

lim (0 − h) = lim (0 − h)2 cos(1/h)

h→0 h→0
= lim h2 cos(1/h) = 0
h→0

Since LHL=RHL, therefore f is continuous at x = 0.

3
 (ii) For differentiability of f , consider
f (x) − f (0) h2 cos(1/h) − 0
lim = lim
h→0 h h→0 h
= lim h cos(1/h)
h→0
=0
Hence, f is differentiable at x = 0.Now,
f (x) − f (0)
|f ′ | = lim
h→0 h
2
h cos(1/h)
= lim
h→0 h
2
|h |
= lim cos(1/h) = 0
h→0 h

i.e., |f ′ | exists. Therefore, f is a function of bounded variation.

(x) If [a′ , b′ ] is a subinterval of [a, b] then
b′ b
V(f ) ≤ V(f ).
a′ a

(xi) For a point c ∈ (a, b),

b c b
V(f ) = V(f ) + V.
a a c
b b b
(xii) V(f ± g) ≤ V(f ) ± V(g).
a a a
b b b
(xiii) V(f g) ≤ AV(f ) + BV(g), where A = sup|g(x)|, B = sup|f (x)|.
a a a
(xiv) Let f, g : [a, b] → R be two functions which are monotonically increasing on [a, b] then
f − g is a function of bounded variation on [a, b].
(xv) If f, g : [a, b] → R is a function of bounded variation on [a, b] then f can be expressed as
a difference of two monotonically increasing functions.
Definition 2.4. A function f : [a, b] → R is said to satisfy Lipschitz condition, if
|f (x) − f (y)| ≤ M |x − y|, M > 0.
Exercise : If f : [a, b] → R is a function of bounded variation on [a, b] then f ′ exists almost
everywhere on [a, b].
Proof. Let f ∈ BV [a, b] and let f1 and f2 are monotonically increasing functions, then f can
be expressed as,
f = f1 − f2 .
We know every monotonic increasing function has a finite derivative almost everywhere, so f1′
and f2′ exist almost everywhere on [a, b]. And thus, f ′ = f1′ − f2′ exists almost everywhere on
[a, b].

4
Exercise : If f : [a, b] → R is a function of bounded variation on [a, b] then f is measurable
on [a, b].
Proof. We know f is continuous on [a, b] then f is measurable on [a, b]. Also, if f is continuous
almost everywhere on [a, b], then it is measurable on [a, b].
Since f is a function of bounded variation on [a, b]. =⇒ f is differentiable almost everywhere
on [a, b]
=⇒ f is differentiable almost everywhere on [a, b].
=⇒ f is continuous almost everywhere on [a, b].
=⇒ f is measurable on [a, b].

3 Absolute Continuity
Definition 3.1. A real-valued function f is said to be continuous at the point x0 , if for a given
ϵ > 0, there is a δ > 0 such that
|f (x) − f (x0 )| < ϵ, whenever |x − x0 | < δ.
Definition 3.2. A function f : [a, b] → R is said to be absolutely continuous on [a, b], if for a
given ϵ > 0 there is a δ > 0 such that
n
X
|f (bi ) − f (ai )| < ϵ,
i=1

for every finite collection of mutually disjoint sub-intervals (ai , bi ), i = 1, 2, · · · , n of [a, b] with
n
X
(bi − ai ) < δ.
i=1
Pn
Note : i=1 (bi − ai ) is the total length of all the sub-intervals (ai , bi ), i = 1, 2, · · · , n of [a, b].
Theorem 3.1. An absolutely continuous function is continuous.
Proof. Let f be absolutely continuous on [a, b]. Then, given ϵ > 0, there is a δ > 0 such that
n
X
|f (bi ) − f (ai )| < ϵ,
i=1

for every finite collection of mutually disjoint sub-intervals (ai , bi ), i = 1, 2, · · · , n of [a, b] with
n
X
(bi − ai ) < δ.
i=1

Let us take the case n = 1 i.e., we choose a single open interval (a1 , b1 ) in [a, b], then
|f (b1 ) − f (a1 )| < ϵ, whenever (b1 − a1 ) < δ.
Let x0 ∈ (a1 , b1 ), then
|f (x) − f (x0 )| < ϵ, whenever |x − x0 | < δ, for x ∈ (a1 , b1 ).
Therefore, f is continuous at the point x0 and as x0 is arbitrary hence f is continuous on
[a, b].

5
Remark : The converse of the above theorem is not true in general.
Theorem 3.2. If f is absolutely continuous on [a, b], then it is of bounded variation on [a, b].
Proof. Let f be absolutely continuous on [a, b]. Then, given ϵ = 1 > 0, there is a δ > 0 such
that n
X
|f (bi ) − f (ai )| < 1,
i=1

for any finite collection of mutually disjoint sub-intervals (ai , bi ), i = 1, 2, · · · , n of [a, b] with
n
X
(bi − ai ) < δ.
i=1

Let P = {c0 , c1 , c2 , · · · , ck−1 , ck , · · · , cm } be a partition of [a, b] such that the length of each
sub-interval is less than δ, that is,

(ck − ck−1 ) < δ, k = 1, 2, · · · , m.

Let P ′ = {x0 , x1 , · · · , xi−1 , xi , · · · , xn } be a partition of [ck−1 , ck ], then

n
X
|f (xi ) − f (xi−1 )| < 1
i=1
( n )
X
sup f (xi ) − f (xi−1 )| ≤1
P′ i=1
ck
or V (f ) ≤ 1, k = 1, 2, 3, . · · · , m
ck−1
and hence
b c1 c2 ck
V(f ) = V(f ) + V(f ) + · · · + V (f )
a c0 c1 ck−1
= 1 + 1 + ··· + 1 (upto m-times)
=m<∞
b
i.e., V(f ) < ∞
a
Hence, f is a funtion of bounded variation on [a, b].
Remarks :
1. The above theorem is not true when a or b is infinite.
2. The converse of the above theorem need not be true, i.e., there are functions of bounded
variations which are not absolutely continuous. For example, Let f : [0, 4] → R defined
by
f (x) = [x].
Then f is a function of bounded variation on [0, 4] but is is not absolutely continuous on
[0, 4] as it is not continuous therein.

6
Theorem 3.3. If f is absolutely continuous then it has a derivative almost everywhere.
Proof. f is absolutely continuous on [a, b]. Therefore, by previous theorem, f is of bounded
variation on [a, b]. Also, we can write f as
f = f1 − f2 ,
where f1 and f2 are monotonic increasing functions on [a, b]. Therefore, f1′ and f2′ exist almost
everywhere on [a, b] and hence
f ′ (x) = f1′ (x) − f2′ (x), almost everywhere on [a, b].
i.e., f has a derivative almost everywhere in [a, b]. Hence, an absolutely continuous function
has a derivative almost everywhere.
Definition 3.3. A function f : [a, b] → R is said to satisfy Lipschitz condition of order α > 0
on [a, b] if there is a constant M > 0 such that
|f (x) − f (y)| ≤ M |x − y|α , ∀x, y ∈ [a, b]
We write Lip α to denote the set of all funtions satisfying Lipschitz condition of order α.
Properties :
1. If α > 1, then f is a constant function.
b
2. If α = 1, then f is of bounded variation on [a, b] and V(f ) ≤ M (b − a).
a
3. If α, then f is differentiable on [a, b].
4. If f has a continuous derivative on [a, b], then f ∈ Lip 1 on [a, b].
5. If α = 1, then f is absolutely continuous on [a, b].
Definition 3.4. A function f : [a, b] → R is said to satisfy Lipschitz condition on [a, b] if there
exists a number M > 0 such that
|f (x) − f (y)| ≤ M |x − y|, ∀x, y ∈ [a, b].
Problem : If f is Lipschitzian, then show that it is absolutely continuous on [a, b].
Proof. Let f be Lipschitz function on [a, b], then
|f (x) − f (y)| ≤ M |x − y|, ∀x, y ∈ [a, b]
where M > 0 is a constant.
Let ϵ > 0 andPlet (ai , bi ), i = 1, 2, · · · , n be a finite collection of mutually disjoint sub-intervals
of [a, b] with ni=1 (bi − ai ) < ϵ/M = δ > 0. Then
n
X n
X
|f (bi ) − f (ai )| ≤ M |bi − ai |
i=1 i=1
ϵ
<M· = ϵ,
M
n
X Xn
Thus, |f (bi ) − f (ai )| < ϵ, whenever (bi − ai ) < δ.
i=1 i=1
Therefore, f is absolutely continuous on [a, b].

7
Proposition. Let f : [a, b] → R be differentiable and its derivative is bounded, then show that
f is Lipschitz function (hence is an absolutely continuous function).

Proof. Let f is differentiable on [a, b] such that

|f ′ (x)| ≤ M, ∀x ∈ [a, b].

Let x, y ∈ [a, b], then

f (x) − f (y)
= f ′ (c)
x−y
f (x) − f (y)
and = |f ′ (c)|
x−y
|f (x) − f (y)| = |f ′ (c)||x − y|
≤ M |x − y|

Therefore, f satisfies Lipschitz condition on [a, b] and hence is absolutely continuous on [a, b].

Example : Let f : [0, 1] → R be a function defined by

(
x2 sin x1 , 0 < x ≤ 1
f (x) =
0, x=0

 
′ 2 1 1 1
f (x) = x cos − 2 + 2x sin
x x x
 
1 1
= 2x sin + − cos
x x
′
|f (x)| ≤ 3, for all x ∈ [0, 1]

Therefore, f satisfies Lipschitz condition on [0, 1] and hence is absolutely continuous on [0, 1].

Theorem 3.4. If f and g are absolutely continuous functions on [a, b] then the following are
absolutely continuous functions on [a, b] :
(i) λf, λ ∈ R, (ii) f + g, (iii) f − g, (iv) f · g,
(v) f /g, g(x) ̸= 0 on [a, b].

Proof. (ii). f and g are absolutely continuous on [a, b], therefore for a given ϵ > 0, there exists
a δ > 0 such that n
X ϵ
|f (bi ) − f (ai )| <
i=1
2
n
X ϵ
and |g(bi ) − g(ai )| <
i=1
2

8
for every finite collection of mutually disjoint sub-intervals (ai , bi ), i = 1, 2, · · · n of [a, b] with
n
X
(bi − ai ) < δ.
i=1

Now

|f (bi ) + g(bi ) − (f (ai ) + g(ai ))| = |(f (bi ) − f (ai )) + (g(bi ) − g(ai ))|
≤ |f (bi ) − f (ai )| + |g(bi ) − g(ai )|.

n
X n
X n
X
Therefore, |(f (bi ) + g(bi )) − (f (ai ) + g(ai ))| ≤ |f (bi ) − f (ai )| + |g(bi ) − g(ai )|
i=1 i=1 i=1
ϵ ϵ
< + = ϵ,
2 2
n
X
whenever (bi − ai ) < δ
i=1

Therefore, f + g is absolutely continuous on [a, b].

Similarly, we can prove that f − g is absolutely continuous on [a, b].
Proof. (iv).

|f (bi )g(bi ) − f (ai )g(ai )| = |f (bi )g(bi ) − f (ai )g(ai ) + f (bi )g(ai ) − f (bi )g(ai )|
= |f (bi ) (g(bi ) − g(ai )) + g(ai ) (f (bi ) − f (ai ))|
≤ |f (bi )||g(bi ) − g(ai )| + |g(ai )||f (bi ) − f (ai )|
≤ M1 |g(bi ) − g(ai )| + M2 |f (bi ) − f (ai )|

where |f (x)| ≤ M1 , ∀x ∈ [a, b] and |g(x)| ≤ M2 , ∀x ∈ [a, b] (since every absolutely continuous
function is bounded).
Hence
n
X n
X
|f (bi )g(bi ) − f (ai )g(ai )| ≤ M1 |g(bi ) − g(ai )| + M2 |f (bi ) − f (ai )|
i=1 i=1
n
ϵ ϵ X
< M1 + M2 = ϵ(say), whenever (bi − ai ) < δ.
2 2 i=1

Therefore, f · g is an absolutely continuous function on [a, b].

Proof. (v). First we will show that if g is absolutely continuous on [a, b], then g1 , where g ̸= 0
on [a, b], is absolutely continuous on [a, b].
Now, as g(x) ̸= 0, ∀x ∈ [a, b], therefore ∃ an M > 0 such that

|g(x)| > M, ∀x ∈ [a, b]

1 1
=⇒ < , ∀x ∈ [a, b].
g(x) M

9
Now,
n n
X g(ai ) − g(bi )
X 1 1
− =
i=1
g(bi ) g(ai ) i=1
g(ai )g(bi )
n
1 1 X
< · |g(bi ) − g(ai )|
M M i=1
n
ϵ ′
X
< 2 = ϵ (say), whenever (bi − ai ) < δ
M i=1

1
Therefore, g
is absolutely continuous function on [a, b].
Now,
f 1
=f·
g g
where g ̸= 0 on [a, b]. Hence fg , g ̸= 0 on [a, b] is an absolutely continuous function on [a, b].

Problem : Show that a function f is absolutely continuous if and only if both f + and f − are
absolutely continuous.
Proof. Let f be a real-valued function defined on [a, b]. Then
(
f, if f ≥ 0
f+ =
0, if f < 0
(
0, if f ≥ 0
f− =
−f, if f < 0
Thus
f = f+ − f− and |f | = f + + f −
or
1 1
f + = |f | + f
2 2
Therefore,
1 1
f + (bi ) = |f (bi )| + f (bi )
2 2
1 1
f + (ai ) = |f (ai )| + f (ai )
2 2

1 1 1 1
|f + (bi ) − f + (ai )| = |f (bi )| + f (bi ) − |f (ai )| − f (ai )
2 2 2 2
1
= |(|f (bi )| − |f (ai )|) + (f (bi ) − f (ai ))|
2
1 1
≤ |(|f (bi )| − |f (ai )|)| + |f (bi ) − f (ai )|
2 2
1 1
≤ |f (bi ) − f (ai )| + |f (bi ) − f (ai )|
2 2
= |f (bi ) − f (ai )|

10
 n
X n
X n
X
+ +
Therefore, f (bi ) − f (ai ) ≤ |f (bi ) − f (ai )| < ϵ whenever (bi − ai ) < δ.
i=1 i=1 i=1
+
Therefore, f is an absolutely continuous function on [a, b].
Again, let f − = 12 |f | − 12 f
n
X n
X n
X
− −
Therefore, f (bi ) − f (ai ) ≤ |f (bi ) − f (ai )| < ϵ whenever (bi − ai ) < δ.
i=1 i=1 i=1
−
Hence, f is an absolutely continuous function on [a, b].

Conversely, let f + and f − are absolutely continuous function on [a, b]. Therefore, f + − f −
is also absolutely continuous function on [a, b]. Hence, f = f + − f − is an absolutely continuous
function on [a, b].
Definition 3.5. A function ϕ : [a, b] → R is called a step function if there is a partition P =
{x0 , x1 , · · · , xn } of [a, b] such that ϕ is constant on every open interval (xi−1 , xi ), i = 1, 2, · · · , n.
That is,
ϕ(x) = ai , x ∈ (xi−1 , xi ), i = 1, 2, · · · , n.
Example 1. A function f : [a, b] → R defined by
(
1 if a ≤ x ≤ c
f (x) =
2 if c ≤ x ≤ b

is a step function on [a, b].

Properties :
1. The image of [a, b] under ϕ, ϕ ([a, b]) = {a1 , a2 , · · · , an } is a finite subset of R. Hence
every step function is bounded.

2. Every step function is measurable.

We know that

(i) A constant function is measurable.

(ii) Let {Ei } be a sequence of disjoint measurable sets. If f is measurable on each
Ei , i = 1, 2, · · · , n then f is also measurable on ∪ni=1 Ei .
Here we have a sequence {(xi−1 , xi ), i = 1, 2, · · · , n} of disjoint measurable intervals
such that ϕ is constant on each interval (xi−1 , xi ) and hence measurable in each
(xi−1 , xi ). Therefore, ϕ is measurable on [a, b] = ∪ni=1 (xi−1 , xi ).

Definition 3.6. A function ϕ : E → R, where E is a measurable set, is said to be a simple

function if there is a finite collection {E1 , E2 , · · · , En } of disjoint measurable subsets of E such
that E = ∪ni=1 Ei and a finite set {a1 , a2 , · · · , an } of real numbers such that

ϕ(x) = ai , x ∈ Ei , i = 1, 2, · · · , n.

Hence, the sets Ei , i = 1, 2, · · · , n form a partition of E.

Remarks :

11
 1. The image of E under ϕ, ϕ(E) = {a1 , a2 , · · · , an } is a finite subset of R and therefore
every simple function is bounded.

2. Every simple function is measurable.

3. The characteristic function of a measurable set is a simple function, i.e.,

(
1, x ∈ A
χ(x) =
0, x ∈ /A

4. Every step funciton is simple function, but the converse is not true. For example, let
f : R → R be defined as (
1, x is rational
f (x) =
0, x is irrational
is a simple function but not a step function.

5. If ψ and ϕ are simple functions on E, then ϕ+c, cϕ, ψ ±ϕ and ψϕ are also simple functions
on E.

6. Let ϕ be a simple function defined in a measurable set E. Let {a1 , a2 , · · · , an } be its

range, where ai is a non-zero real number, then
n
X
ϕ(x) = ai χEi (x),
i=1

where the sets Ei = {x ∈ E : ϕ(x) = ai , i = 1, 2, · · · , n} are disjoint measurable subsets

of E and χEi , the characteristic function of the set Ei . This representation of ϕ as a lin-
ear combination of characteristic functions of measurable subsets of E is called canonical
representation of ϕ.

The elementary integral or simply integral of ψ over [a, b] is defined as

Z b Xn
ψ(x)dx = ai (xi , xi−1 )
a i=1
Xn
= ai ∆xi , where ψ is a step function.
i=1

Now, we will define the Riemann integral of a bounded function through step functions as
follows :
Let f be a bounded function defined on [a, b]. Let ψ and ϕ be step functions on [a, b]. Then
we define
Z b Z b
R f (x)dx = inf ψ(x)dx
a ψ≥f a
Z b Z b
and R f (x)dx = sup ϕ(x)dx
a ϕ≤f a

12
 Z b Z b
If R f (x)dx = R f (x)dx
a a

then, we say that f is Riemann integrable on [a, b] and the common value is denoted by
Z b
R f (x)dx.
a

Let f : E → R be a bounded function and E a measurable set with m(E) < ∞. Let ψ and ϕ
be simple functions defined on E. Let us consider the numbers
Z
inf ψ(x)dx, for all simple functions ψ defined on E
ψ≥f E
Z
and sup ϕ(x)dx, for all simple functions ϕ defined on E
ϕ≤f E

which always exist. But the question is whether these two numbers are equal? The answer is
yes, if f is measurable. The answer is given by the following theorem.

Theorem 3.5. Let f be a real-valued bounded function defined on a measurable set E with
m(E) < ∞. In order that Z Z
inf ψ(x)dx = sup ϕ(x)dx
ψ≥f E ψ≤f E

for all simple functions ψ and ϕ, defined on E, it is necessary and sufficient that f to be
measurable on E.

4 Lebesgue integral of a bounded function :

Definition 4.1. Let f be a bounded measurable function defined on a measurable set E with
m(E) < ∞. The Lebesgue integral of f over E is defined as
Z Z
f (x)dx = inf ψ(x)dx
E ψ≥f E

for all simple functions ψ defined on E.

The Lebesgue integral of f over E can also be defined as
Z Z
f (x)dx = sup ϕ(x)dx
E ϕ≤f E

for all simple functions ϕ defined on E.

Remark :
R R
1. We also denote the integral E
f (x)dx as E
f.
R Rb
2. As a special case, when E = [a, b], we write the integral [a,b]
f (x)dx as a
f (x)dx.
Rb Rb
3. supϕ≤f a
ϕ(x)dx ≤ inf ψ≥f a
ψ(x)dx, for all simple functions ϕ and ψ.

13
Theorem 4.1 (Relation between Riemann integrals and Lebesgue integrals). Let f be a bounded
funtion defined in [a, b]. If f is Riemann integrable on [a, b], then it is measurable in [a, b] and
Z b Z b
f (x)dx = R f (x)dx.
a a

Proof. Let ψ1 and ϕ1 be step functions defined on [a, b]. Then

Z b̄ Z b
R f (x)dx = inf ψ1 (x)dx
a ψ1 ≥f a
Z b Z b
and R f (x)dx = sup ϕ1 (x)dx
ā ϕ1 ≤f a

Every step function is a simple function. Therefore

Z b Z b
inf ψ1 (x)dx ≥ inf ψ(x)dx
ψ≥f a ψ≥f a
Z b Z b
and sup ϕ1 (x)dx ≤ inf ϕ(x)dx,
ϕ1 ≤f a ϕ≤f a

where ψ and ϕ are simple functions defined on [a, b]. Therefore,

Z b̄ Z b
R f (x)dx ≥ inf ψ(x)dx
a ψ≥f a
Z b Z b
and R f (x)dx ≤ sup ϕ(x)dx.
ā ϕ≤f a

Now,
Z b Z b̄
R f (x)dx ≤ R f (x)dx.
ā a
Therefore Z b Z b Z b Z b̄
R f (x)dx ≤ sup ϕ(x)dx ≤ inf ψ(x)dx ≤ R f (x)dx.
ā ϕ≤f a ψ≥f a a

As f is Riemann integrable on [a, b], so

Z b Z b̄ Z b
R f (x)dx = R f (x)dx = R f (x)dx.
ā a a

Therefore from above, we have

Z b Z b Z b Z b
R f (x)dx ≤ sup ϕ(x)dx ≤ inf ψ(x)dx ≤ R f (x)dx
a ϕ≤f a ψ≥f a a

Z b Z b Z b
=⇒ sup ϕ(x)dx = inf ψ(x)dx = R f (x)dx
ϕ≤f a ψ≥f a a

14
i.e., f is measurable on [a, b].
Thus, f is bounded and measurable on [a, b]. Therefore f is Lebesgue integrable on [a, b] and
Z b Z b
f (x)dx = inf ψ(x)dx
a ψ≥f a
Z b
=R f (x)dx
a

Thus, Z b Z b
f (x)dx = R f (x)dx.
a a

Remark : The above theorem says that any function which is Riemann integrable is also
Lebesgue integrable and that the two integrals are equal. However, there are bounded measur-
able functions which are Lebesgue integrable but nor Riemann integrable. For example, let us
consider the Dirichlet functions f : [a, b] → R defined by
(
1, x is rational in [a, b] or x ∈ Q ∩ [a, b]
f (x) =
0, x is irrational in [a, b] or x ∈ (R − Q) ∩ [a, b]
Rb
is bounded and measurable in [a, b] and so it is Lebesgue integrable on [a, b]. Also, a f (x)dx =
0. On the other hand, f is not Riemann integrable on [a, b]. Thus we can say that the class
of Lebesgue integrable functions is wider than that of Riemann integrable functions. However,
the situation is entirely different in the case of improper Riemann integrable functions. There
are improper Riemann integrable functions which are not Lebesgue integrable.
Properties of the Lebesgue integral for bounded measurable functions :
Let f and g be bounded measurable functions defined on a set E of finite measure, then
R R
1. E af = a E f, for all real numbers a.
R R R
2. E (f + g) = E f + E g.

3. if f = g almost everywhere in E, then

Z Z
f= g.
E E

4. if f ≤ g almost everywhere in E, then

Z Z
f≤ g.
E E

Also Z Z
f ≤ |f |.
E E

15
 5. (Mean Value Theorem) Let α ≤ f (x) ≤ β for all x ∈ E, then
Z
αm(E) ≤ f (x)dx ≤ βm(E)
E

where α = inf f (x) and β = sup f (x).

x∈E x∈E

Proof. Let α ≤ f (x) ≤ β for all x ∈ E. Then from 4, we have

Z Z Z
αdx ≤ f (x)dx ≤ βdx
Z E ZE EZ

or α 1 · dx ≤ f (x)dx ≤ β 1 · dx
E ZE E
R
or αm(E) ≤ f (x)dx ≤ βm(E), (since, E 1 · dx = m(E))
E

6. If E1 and E2 are disjoint measurable subsets of E, then

Z Z Z Z
f= f= f+ f
E E1 ∪E2 E1 E2

Corollaries :

1. If f (x) ≥ 0 on E, then Z
f (x)dx ≥ 0.
E

Proof. If f (x) ≥ 0 on E, then

Z
0 · m(E) ≤ f (x)dx ≤ βm(E)
E
Z
or 0 ≤ f (x)dx ≤ βm(E).
E
Therefore, Z
f (x)dx ≥ 0.
E

2. If f (x) ≤ 0 on E, then Z
f (x)dx ≤ 0.
E

16
 Proof. Let f (x) ≤ 0 on E, then
Z
αm(E) ≤ f (x)dx ≤ 0 · m(E)
E
Z
or αm(E) ≤ f (x)dx ≤ 0
E
Therefore, Z
f (x)dx ≤ 0.
E

3. If m(E) = 0, then Z
f (x)dx = 0.
E

Proof. Let m(E) = 0, then Z

α·0≤ f (x)dx ≤ β · 0
E
Z
or 0 ≤ f (x)dx ≤ 0
E
Therefore, Z
f (x)dx = 0.
E

Hence, the integral of a function over a countable set is always zero. Thus,
Z
f (x)dx = 0.
Q

4. If f is constant on E, say f (x) = K, for all x ∈ E, then

f (x)dx = Km(E).

Proof. Let f (x) = K, for all x ∈ E, then K ≤ f (x) ≤ K for all x ∈ E. Therefore,
Z
Km(E) ≤ f (x)dx ≤ Km(E)
E
Z
=⇒ f (x)dx = Km(E).
E

(i) If f (x) = 1 for all x ∈ E, then

Z Z
f (x)dx = 1dx = m(E).
E E

17
 (ii) If f (x) = 0 for all x ∈ E, then
Z
f (x)dx = 0 · m(E) = 0.
E

5. If f (x) = K almost everywhere in E, then

Z
f (x)dx = Km(E)
E

Proof. We know that if f = g almost everywhere in E, then

Z Z
f (x)dx = g(x)dx
E E

Therefore, Z Z Z
f (x)dx = Kdx = K dx = Km(E).
E E E

(a) If f (x) = 1 almost everywhere in E, then

Z
f (x)dx = 1 · m(E)
E
Z
i.e. dx = m(E).
E

(b) If f (x) = 0 almost everywhere in E, then

Z
f (x)dx = 0 · m(E) = 0
E
Z
i.e., f (x)dx = 0.
E

5 Pointwise convergence and Uniform convergence

5.1 Pointwise convergence
Definition 5.1. Let {fn } be a sequence of real-valued funcitons defined in [a, b]. We say that
the sequence {fn } converges pointwise (i.e., at each point of [a, b]) to a function f defined on
[a, b], if
lim fn (x) = f (x), for each point x ∈ [a, b]
n→∞
and we write
fn → f pointwise on [a, b]
p
or fn →
− f on [a, b].
This means that the sequence {fn } converges pointwise to a function f on [a, b] if, given ϵ > 0
and x ∈ [a, b], there exists a positive integer N (depending on ϵ and x) such that for all n ≥ N ,
|fn (x) − f (x)| < ϵ.

18
5.2 Uniform Convergence
Definition 5.2. Let {fn } be a sequence of real-valued functions defined on [a, b]. We say that
the sequence {fn } converges uniformly to a function f defined on [a, b], if given ϵ > 0 and for
all x ∈ [a, b] there exists a positive integer N (depending on ϵ but independent of x) such that
for all n ≥ N ,
|fn (x) − f (x)| < ϵ
and we write
u
lim fn (x) = f (x) uniformly on [a, b] or fn →
− f on [a, b].
n→∞

Remarks :
1. The difference between the concept of pointwise convergence and the uniform convergence
is that in case of pointwise convergence, for each ϵ > 0 and for each x ∈ [a, b], we choose
an integer N so that for all n ≥ N,
|fn (x) − f (x)| < ϵ,
whereas in uniform convergence we choose an ϵ > 0 and for all x ∈ [a, b], we choose one
interger N so that for all n ≥ N ,
|fn (x) − f (x)| < ϵ.

2. Uniform convergence implies pointwise convergence but pointwise convergence does not
imply uinform convergence in general. For example, let
fn (x) = xn , x ∈ [0, 1]
p
then fn →
− f on [0, 1], where (
0, x ∈ [0, 1)
f (x) =
1, x = 1
u
but fn →
−
̸ f on [0, 1].
3. A sequence which is not pointwise convergent can not be uniformly convergent. For
example, let √
fn (x) = n cos nx, x ∈ R
p u
then fn →
−
̸ to any function on R =⇒ fn →
−
̸ to any function on R.
Definition 5.3. Let {fn } be a sequence of measurable functions defined on a measurable set E
of finite measure and let f be a measurable function on E.
(i) We say that {fn } converges pointwise to f , if given ϵ > 0 and x ∈ E there exists a positive
integer N such that for all n ≥ N ,
|fn (x) − f (x)| < ϵ
and we write
lim fn (x) = f (x), for each x ∈ E
n→∞
p
or fn → f pointwise on E or fn →
− on E.

19
 (ii) We say {fn } converges pointwise to f almost everywhere in E if there exists a subset E1
of E with m(E1 ) = 0 such that
p
fn →
− f on E − E1
p
and we write fn →
− f almost everywhere in E.

(iii) We say {fn } converges uniformly to f on E, if, given any ϵ > 0 and for all x ∈ E
there exists a positive integer N (depending upon ϵ but independent of x) such that for all
n ≥ N,
|fn (x) − f (x)| < ϵ
and we write
lim fn (x) = f (x) uniformly on E
n→∞
u
or fn →
− f on E.
u p p
Remark : fn → − f =⇒ fn → − f =⇒ fn →
− f almost everywhere, but all the reverse
implications need not be true.

(iv) We say that {fn } converges uniformly to f almost everywhere in E if there exists a subset
E1 of E with m(E1 ) = 0 such that
u
fn →
− f on E − E1
u
and we write fn → − f almost everywhere in E.
u p
Remark : fn → − f almost everywhere =⇒ fn → − f almost everywhere, but the reverse
implications need not be true.

We know that uniform convergence implies pointwise convergence, but the converse may not
be true always. For example, let {fn } be a sequence, where

fn (x) = xn , x ∈ [0, 1]

then the sequence {fn } converges pointwise to the function f on [0, 1], where
(
0, x ∈ [0, 1)
f (x) =
1, x = 1
u
but fn →
−
̸ f on [0, 1].

Theorem 5.1 (Egoroff’s Theorem). Let {fn } be a sequence of measurable functions defined on
E with finite measure. Let f be a measurable function defined on E such that
p
fn →
− f almost everywhere on E.

Then, given η > 0, there is a measurable subset A of E such that m(A) < η and the sequence
{fn } converges uniformly to f on E − A.

Remark : The Egoroff’s theorem motivates the introduction of the concept of almost uniform
convergence.

20
Definition 5.4. A sequence {fn } of measurable functions is said to converge almost uniformly
to a measurable functions f defined on a measurable set E if, for a given ϵ > 0, ∃ a measurable
set A ⊂ E with m(A) < ϵ such that {fn } converges to f uniformly on E − A and we write
a.u.
fn −−→ f on E or fn → f almost uniformly on E.

In view of the above definition, Egroff’s theorem can be rewritten as :

Let {fn } be a sequence of measurable functions defined on E with finite measure. Let f be
p
a measurable function defined on E such that fn → − f almost everywhere in E, then fn → f
almost uniformly on E.
Let {fn } be a sequence of Lebesgue integrable functions over E(i.e., each fn is bounded and
measurable on E with m(E) < ∞) such that
p
fn →
− f on E.

or lim fn (x) = f (x) for each x ∈ E

n→∞
R ∞ R
Now, the question is , does f (x)dx n=1 converges to E f (x)dx?
E n
In other words, if
lim fn (x) = f (x), for each x ∈ E,
n→∞
Z Z
Is lim fn (x)dx = f (x)dx ?
n→∞ E E
Z Z
Is lim fn (x)dx = lim fn (x)dx ?
n→∞ E E n→∞
R
or is it possible to interchange limn→∞ and E operations ?
The answer to the above questions is given by the following theorem, known as (Lebesgue)
Bounded convergence theorem.

Theorem 5.2 (Bounded convergence theorem). Let {fn } be a sequence of measurable functinos
defined on a measurable set E with m(E) < ∞. Suppose there is real number M such that
|fn (x)| ≤ M for all n and all x ∈ E. If

lim fn (x) = f (x) for each x ∈ E,

n→∞
Z Z
then lim fn (x)dx = f (x)dx
n→∞ E E
Z Z
p
i.e., fn →
− f.
E E

Remark : The Bounded convergence theorem is not true for Riemann integrals.
p
Proof. Let {fn } is a sequence of measurable functions defined on E such that fn →
− f on E or,

lim fn (x) = f (x), for each x ∈ E

n→∞

21
∴ f is a measurable function on E.Now,

f (x) = lim fn (x), x ∈ E

n→∞

∴ |f (x)| = lim fn (x)

n→∞
= lim |fn (x)|
n→∞
≤ M, x ∈ E

∴ f is bounded on E.
Thus, f is a bounded measurable function defined on a measurable set E with m(E) < ∞,
hence it is Lebesgue integrable over E.
Also, from above, we have

|fn (x) − f (x)| ≤ |fn (x)| + |f (x)| , x ∈ E

≤M +M
= 2M, x ∈ E

∴ In particular,
|fn (x) − f (x)| ≤ 2M, for x ∈ A ⊂ E.
We know that, if {fn } is a sequence of measurable functions defined on a measurable set E
with m(E) < ∞ and if f is a measurable function on E such that

lim fn (x) = f (x), for each x ∈ E,

n→∞

then, given ϵ > 0 and δ > 0 there is a measurable set A ⊂ E with m(A) < δ and a positive
integer N such that for all n ≥ N ,

|fn (x) − f (x)| < ϵ, for all x ∈ E − A

ϵ ϵ
Thus, for a given ϵ > 0 and δ = 4M , there is a measurable set A ⊂ E with m(A) < 4M
and a
positive integer N such that for all n ≥ N ,
ϵ
|fn (x) − f (x)| < , for all x ∈ E − A.
2m(E)

Now,
Z Z Z
fn (x)dx − f (x)dx = (fn (x) − f (x)) dx
E E E
Z
= |fn (x) − f (x)| dx
ZE Z
= |fn (x) − f (x)| dx + |fn (x) − f (x)| dx,
E−A A
where E and A − E are disjoint measurable sets
Z Z
ϵ
< 1 · dx + 2M 1 · dx, n ≥ N
2m(E) E−A A

22
 ϵ
= m(E − A) + 2M · m(A), n ≥ N
2m(E)
ϵ ϵ
< · m(E) + 2M · ,n ≥ N (⋆)
2m(E) 4M
ϵ ϵ
= + = ϵ, n ≥ N.
2 2
(⋆) : ∵ E − A ⊂ E, m(E − A) ≤ m(E).

Thus, for a given ϵ > 0, there exists a positive integer N such that for all n ≥ N ,
Z Z
fn (x)dx − f (x)dx < ϵ, x ∈ E
E E
Z Z Z
i.e., lim fn (x)dx = f (x)dx = lim fn (x)dx.
n→∞ E E E n→∞

Theorem 5.3. If f is an integrable function on [a, b], then show that the function F defined by
Z x
F (x) = f (t)dt, x ∈ [a, b]
a

is a continuous function of bounded variation on [a, b].

Proof. A function f is said to be continuous at a point x0 if given ϵ > 0 there exists a δ > 0
such that
|x − x0 | < δ =⇒ |f (x) − f (x0 )| < ϵ.
1. Let x0 be any point in [a, b]. Then
Z x0
F (x0 ) = f (t)dt
a
Z x Z x0
∴ F (x) − F (x0 ) = f (t)dt − f (t)dt
a a
Z x
= f (t)dt
x0
Z x
|F (x) − F (x0 )| = f (t)dt
x0
Z x
≤ |f (t)|dt
x0

We know that if f is a non-negative integrable function on a set E, then given ϵ > 0 there
exists a δ > 0 such that for every set A ⊂ E with m(A) < δ
Z
f < ϵ.
A

Since |f | ≥ 0 and is integrable on E,

Z
|f | < ϵ.
A

23
 In particular, let E = [a, b], A = [x0 , x] ⊂ [a, b] then m(A) = |x − x0 | < δ and therefore
Z x
|f (t)|dt < ϵ, whenever |x − x0 | < δ.
x0

∴ From above, we have

|F (x) − F (x0 )| < ϵ, whenever |x − x0 | < δ.
∴ F is continuous at the point x0 and hence F is continuous on [a, b].
2. Let P = {x0 , x1 , · · · , xn } be a partition of [a, b], then
Z xi Z xi−1
F (xi ) = f (t)dt and F (xi−1 ) = f (t)dt
a a
Z xi Z xi−1
∴ F (xi ) − F (xi−1 ) = f (t)dt − f (t)dt
a A
Z xi
= f (t)dt
xi−1
Z xi Z xi
∴ |F (xi ) − F (xi−1 )| = f (t)dt ≤ |f (t)|dt
xi−1 xi−1

n
X n Z
X xi
|F (xi ) − F (xi−1 )| ≤ |f (t)|dt
i=1 i=1 xi−1
Z b
= |f (t)|dt
a
( n )
X Z b
sup |F (xi ) − F (xi−1 )| ≤ |f (t)|dt < ∞
P a
i=1
b Z b
or V(F ) ≤ |f (t)|dt < ∞
a a
Therefore, F is a function of bounded variation on [a, b].

6 Lebesgue Integral of a non-negative measurable func-

tion over any measurable set
Let f be a bounded measurable function defined on E with m(E) < ∞. Then we define
Z Z
f (x)dx = inf ψ(x)dx
E ψ≥f E
Z
= sup ϕ(x)dx,
ϕ≤f E

where ψ and ϕ are simple functions defined on E.

Now we will define the Lebesgue integral of a non-negative measurable function over any
measurable set.

24
Definition 6.1. Let f be a non-negative measurable function defined on a measurable set E,
where m(E) is not necessarily finite, we define
Z Z
f (x)dx = sup h(x)dx,
E h≤f E

where h is a bounded measurable function on E which vanishes outside a set E1 ⊂ E of finite

measure.

Theorem 6.1. Let f and g be non-negative measurable functions defined on a set E, then
R R
1. E cf = c E f, c > 0.
R R R
2. E (f + g) = E f + E g.

Proof. 2. Let h and k be bounded and measurable functions on E such that h ≤ f and k ≤ g
and they vanish outside sets of finite measure i.e.,

m(E1 ) = m{x ∈ E : h(x) ̸= 0} < ∞,

m(E2 ) = m{x ∈ E : k(x) ̸= 0} < ∞.

Then h(x) + k(x) ≤ f (x) + g(x), x ∈ E
Z Z Z
Now, h= h+ h
E E1 E−E1
Z Z
= h+0= h, where m(E1 ) < ∞.
E1 E1

Z Z
Similarly, k= k, where m(E2 ) < ∞.
E E2

Therefore, h and k are Lebesgue integrable over E.

As h + k ≤ f + g on E, therefore
Z Z Z Z
h+ k = (h + k) ≤ (f + g)
E E E E
Z Z Z
sup h + sup k≤ (f + g)
h≤f E k≤g E E
Z Z Z
or f+ g≤ (f + g). (1)
E E E
Now, we will show that Z Z Z
(f + g) ≤ f+ g.
E E E

Let ψ be a bounded measurable function on E such that ψ(x) ≤ f (x) + g(x), x ∈ E and ψ
vanishes outside a set E1 ⊂ E of finite measure. Now we define the functions h and k by setting

h(x) = min (f (x), ψ(x)), x ∈ E

25
 and k(x) = ψ(x) − h(x), x ∈ E,
then 0 ≤ h(x) ≤ f (x), x ∈ E
and 0 ≤ k(x) ≤ g(x), x ∈ E.
Also, h and k are measurable and bounded on E by the boundedness of ψ and vanish outside
the set E1 ⊂ E of finite measure. Therefore h and k are Lebesgue integrable over E.Therefore
Z Z
ψ = (h + k)
E ZE Z
= h+ k
ZE ZE

≤ f+ g, h ≤ f, k ≤ g.
Z ZE ZE
∴ sup ψ≤ f+ g
ψ≤f +g E E E
Z Z Z
or (f + g) ≤ f+ g (2)
E E E

Z Z Z
(1) and (2) =⇒ (f + g) = f+ g.
E E E

Fatou’s Lemma : Let {fn } be a sequence of non-negative measurable functions defined on E

such that fn converges pointwise to f almost everywhere in E then
Z Z
f ≤ lim fn .
E n→∞ E

Proof. Since the integrals over sets of measure zero are zero, therefore, we can assume that
{fn } converges to f on E.
Let h be a bounded measurable function on E such that h ≤ f on E and vanishes outside a
set E1 ⊂ E of finite measure i.e.,
m(E1 ) = m{x ∈ E : h(x) ̸= 0} < ∞.
Let |h(x)| ≤ M, for all x ∈ E and let {hn (x)} be a sequence of functions on E such that
hn (x) = min{h(x), fn (x)}, x ∈ E
then the sequence {hn } is bounded by M on E and hn vanishes outside the set E1 ⊂ E of finite
measure.
Now, for each x ∈ E1 ,
hn (x) = min{h(x), fn (x).}
∴ lim hn (x) = lim min{h(x), fn (x)}
n→∞ n→∞
= min{h(x), lim fn (x)}
n→∞
= min{h(x), f (x)}
= h(x), x ∈ E1 .

26
Therefore, {hn } converges pointwise to h on E1 .
Thus, the sequence {hn } is a sequence of bounded measurable functions on a measurable set E1
of finite measure such that {hn } converges pointwise to h on E1 . Then, by Lebesgue bounded
convergence theorem, we have
Z Z Z
lim hn = h= h
n→∞ E E1 E
1
Z Z
or h = lim hn
E n→∞ E
1
Z
= lim hn
n→∞ E1
Z
≤ lim fn
n→∞ E1
Z Z
Now, fn ≤ fn (∵ fn ≥ 0 on E)
E1 E
Z Z
∴ lim fn ≤ lim fn .
n→∞ E1 n→∞ E

Thus, from above, we have Z Z

h ≤ lim fn
E n→∞ E
Z Z
∴ sup h ≤ lim fn
h≤f E n→∞ E
Z Z
or f ≤ lim fn .
E n→∞ E

7 Differentiation of an Integral
Theorem 7.1. Let f be a bounded and measurable function defined on [a, b]. If
Z x
F (x) = f (t)dt + F (a), x ∈ [a, b]
a

then F ′ (x) = f (x) almost everywhere in [a, b].

For the proof, we need the following result:

“Let f be an integrable function on [a, b]. If
Z x
f (t)dt = 0, ∀x ∈ [a, b]
a

then f = 0 a.e. in [a, b]”.

27
Proof. f is bounded and measurable on [a, b]. Therefore f is integrable on [a, b].
Z x
Let F (x) = f (t)dt + F (a), x ∈ [a, b]
a
then
(i) F is continuous function on [a, b]. Therefore F is a measurable function on [a, b].

(ii) F is a function of bounded variation on [a, b]. Therefore F ′ exist almost everywhere in
[a, b].
f is bounded on [a, b], i.e., |f (x)| ≤ k for all x ∈ [a, b].
Let {fn (x)} be a sequence defined on [a, b] where we take (or set)
F (x + 1/n) − F (x)
fn (x) = , x ∈ [a, b]
1/n
where each fn is measurable on [a, b].
F (x + 1/n) − F (x)
fn (x) = , x ∈ [a, b]
1/n
F (x + h) − F (x)
= , h = 1/n
h
Z x+h
1 x
Z
1
= f (t)dt − f (t)dt
h a h a
1 x+h
Z
= f (t)dt
h x
1 x+h
Z
∴ |fn (x)| ≤ |f (t)|dt
h x
Z x+h
1
≤ k dt = k.
h x
∴ |fn (x)| ≤ k for each n and for all x ∈ [a, b].
Also
F (x + 1/n) − F (x)
lim fn (x) = lim
n→∞ n→∞ 1/n
F (x + h) − F (x)
= lim , h = 1/n
n→∞ h
= F ′ (x) almost everywhere in [a, b].

i.e., {fn } converges to F ′ almost everywhere in [a, b].

Then, {fn } is a sequence of measurable functions defined on [a, b] such that |fn (x)| ≤ k for
each n and for all x ∈ [a, b] and {fn } converges to F ′ almost everywhere in [a, b]. Then by
Lebesgue Bounded convergence theorem, if c ∈ [a, b] is arbitrary,

Z c Z c
′
F (x)dx = lim fn (x)dx
a n→∞ a

28
 c
F (x + 1/n) − F (x)
Z
= lim dx
n→∞ a 1/n
Z c
F (x + h) − F (x)
= lim dx
n→∞ a h
 Z c
1 c
Z 
1
= lim F (x + h)dx − F (x)dx
n→∞ h a h a
 Z c+h
1 c
Z 
1
= lim F (x)dx − F (x)dx
n→∞ h a+h h a
 Z c
1 c+h 1 a+h 1 c
Z Z Z 
1
= lim F (x)dx + F (x)dx − F (x)dx − F (x)dx
n→∞ h a h c h a h a
 Z c+h
1 a+h
Z 
1
= lim F (x)dx − F (x)dx
n→∞ h c h a
1 c+h 1 a+h
Z Z
= lim F (x)dx − lim F (x)dx
n→∞ h c n→∞ h a

= F (c) − F (a) (F is continuous on [a, b])

Z c
= f (x)dx
a
<∞ (∵ f is integrable on [a, b])

Therefore, F ′ is integrable on [a, c]. Hence, (F ′ − f ) is integrable on [a, c].

Now, from above, we have
Z c
(F ′ (x) − f (x)) dx = 0, ∀c ∈ [a, b]
a

Therefore F ′ (x) − f (x) = 0 almost everywhere in [a, b] or F ′ (x) = f (x) almost everywhere in
[a, b].

Now, we will prove that

Z a+h
1
lim F (x)dx = F (a), F is continuous on [a, b].
h→0 h a

Proof. (First Mean value theorem for Riemann integrals)“If f is continuous on [a, b] then there
exists a point c ∈ [a, b] such that
Z b
f (x)dx = (b − a)f (c)
a
= (b − a)f (a + θ(b − a)), 0 ≤ θ ≤ 1

Let 0 ≤ θ ≤ 1
0 ≤ θ(b − a) ≤ (b − a)
∴ 0 ≤ a + θ(b − a) ≤ b

29
Let c = a + θ(b − a), then a ≤ c ≤ b i.e., c ∈ [a, b].”
Since F is continuous on [a, b].
∴ F is R-integrable on [a, b].
∴ F is Lebesgue integrable on [a, b] and
Z b Z b
F (x)dx = R F (x)dx.
a a

Then

Z a+h Z a+h
1 1
lim F (x)dx = lim R F (x)dx
h→0 h a h→0 h a
1
lim (a + h − 1)F (a + θ(a + h − a)), 0 ≤ θ ≤ 1
h→0 h
1
= lim · hF (a + θh)
h→0 h
= lim F (a + θh)
h→0
h i
= F lim (a + θh) (∵ F is continuous on [a, b])
h→0
= F (a).

Similarly, we can show that Z c+h

1
lim F (x)dx = F (c).
h→0 h c

30