2.

Convergence theorems
In this section we analyze the dynamics of integrabilty in the case when se-
quences of measurable functions are considered. Roughly speaking, a convergence
theorem states that integrability is preserved under taking limits. In other words,
if one has a sequence (f
n
)

n=1
of integrable functions, and if f is some kind of a
limit of the f
n
s, then we would like to conclude that f itself is integrable, as well
as the equality
_
f = lim
n
_
f
n
.
Such results are often employed in two instances:
A. When we want to prove that some function f is integrable. In this case
we would look for a sequence (f
n
)

n=1
of integrable approximants for f.
B. When we want to construct and integrable function. In this case, we will
produce rst the approximants, and then we will examine the existence
of the limit.
The rst convergence result, which is somehow primite, but very useful, is the
following.
Lemma 2.1. Let (X, A, ) be a nite measure space, let a (0, ) and let
f
n
: X [0, a], n 1, be a sequence of measurable functions satisfying
(a) f
1
f
2
0;
(b) lim
n
f
n
(x) = 0, x X.
Then one has the equality
(1) lim
n
_
X
f
n
d = 0.
Proof. Let us dene, for each > 0, and each integer n 1, the set
A

n
= {x X : f
n
(x) }.
Obviously, we have A

n
A, > 0, n 1. One key fact we are going to use is the
following.
Claim 1: For every > 0, one has the equality
lim
n
(A

n
) = 0.
Fix > 0. Let us rst observe that, using (a), we have the inclusions
(2) A

1
A

2
. . .
Second, using (b), we clearly have the equality

k=1
A

k
= . Since is nite,
using the Continuity Property (Lemma III.4.1), we have
lim
n
(A

n
) =
_

n=1
A

n
_
= () = 0.
Claim 2: For every > 0 and every integer n 1, one has the inequality
0
_
X
f
n
d a(A

n
) +(X).
300
2. Convergence theorems 301
Fix and n, and let us consider the elementary function
h

n
= a
A

n
+
B

n
,
where B

n
= X A

. Obviously, since (X) < , the function h

n
is elementary
integrable. By construction, we clearly have 0 f
n
h

n
, so using the properties
of integration, we get
0
_
X
f
n
d
_
X
h

n
d = a(A

n
) +(B

) a(A

) +(X).
Using Claims 1 and 2, it follows immediately that
0 liminf
n
_
X
f
n
d limsup
n
_
X
f
n
d (X).
Since the last inequality holds for arbitrary > 0, the desired equality (1) immedi-
ately follows.
We now turn our attention to a weaker notion of limit, for sequences of mea-
surable functions.
Definition. Let (X, A, ) be a measure space, let K be a one of the symbols

R, R, or C. Suppose f
n
: X K, n 1, are measurable functions. Given
a measurable function f : X K, we say that the sequence (f
n
)

n=1
converges
-almost everywhere to f, if there exists some set N A, with (N) = 0, such that
lim
n
f
n
(x) = f(x), x X N.
In this case we write
f = -a.e.- lim
n
f
n
.
Remark 2.1. This notion of convergence has, among other things, a certain
uniqueness feature. One way to describe this is to say that the limit of a -a.e. con-
vergent sequence is -almost unique, in the sense that if f anf g are measurable func-
tions which satisfy the equalities f = -a.e.- lim
n
f
n
and g = -a.e.- lim
n
f
n
,
then f = g, -a.e. This is quite obvious, because there exist sets M, N A, with
(M) = (N) = 0, such that
lim
n
f
n
(x) = f(x), x X M,
lim
n
f
n
(x) = g(x), x X N,
then it is obvious that (M N) = 0, and
f(x) = g(x), x X [M N].
Comment. The above denition makes sense if K is an arbitrary metric space.
Any of the spaces

R, R, and C is in fact a complete metric space. There are instances
where the requirement that f is measurable is if fact redundant. This is somehow
claried by the the next two exercises.
Exercise 1*. Let (X, A, ) be a measure space, let K be a complete separable
metric space, and let f
n
: X K, n 1, be measurable functions.
(i) Prove that the set
L =
_
x X :
_
f
n
(x)
_

n=1
K is convergent
_
belongs to A.
302 CHAPTER IV: INTEGRATION THEORY
(ii) If we x some point K, and we dene : X K by
(x) =
_
lim
n
f
n
(x) if x L
if x X L
then is measurable.
In particular, if (X L) = 0, then = -a.e.- lim
n
f
n
.
Hints: If d denotes the metric on K, then prove rst that, for every > 0 and every m, n 1,
the set
D

mn
=
_
x X : d
_
fm(x), fn(x)
_
<
_
belongs to A (use the results from III.3). Based on this fact, prove that, for every p, k 1, the
set
E
p
k
=
_
x X ; d
_
fm(x), fn(x)
_
<
1
p
, m, n k
_
belongs to A. Finally, use completeness to prove that
L =

p=1
_

_
k=1
E
p
k
_
.
Exercise 2. Use the setting from Exercise 1. Prove that Let (X, A, ), K, and
(f
n
)

n=1
be as in Exercise 1. Assume f : X K is an arbitrary function, for which
there exists some set N A with (N) = 0, and
lim
n
f
n
(x) = f(x), x X N.
Prove that, when is a complete measure on A (see III.5), the function f is auto-
matically measurable.
Hint: Use the results from Exercise 1. We have X N L, and f(x) = (x), x X N.
Prove that, for a Borel set B K, one has the equality f
1
(B) =
1
(B)M, for some M N.
By completeness, we have M A, so f
1
(B) A.
The following fundamental result is a generalization of Lemma 2.1.
Theorem 2.1 (Lebesgue Monotone Convergence Theorem). Let (X, A, ) be
a measure space, and let (f
n
)

n=1
L
1
+
(X, A, ) be a sequence with:
f
n
f
n+1
, -a.e., n 1;
sup
__
X
f
n
d : n 1
_
< .
Assume f : X [0, ] is a measurable function, with f = -a.e.- lim
n
f
n
. Then
f L
1
+
(X, A, ), and
_
X
f d = lim
n
_
X
f
n
d.
Proof. Dene
n
=
_
X
f
n
d, n 1. First of all, we clearly have
0
1

2
. . . ,
so the sequence (
n
)

n=1
has a limit = lim
n

n
, and we have in fact the
equality
= sup
_

n
: n 1
_
< .
With these notations, all we need to prove is the fact that f L
1
+
(X, A, ), and
that we have
(3)
_
X
f d = .
Fix a set M A, with (M) = 0, and such that lim
n
f
n
(x) = f(x),
x X M. For each n, we dene the set
M
n
= {x X : f
n
(x) > f
n+1
(x)}.
2. Convergence theorems 303
Obviously M
n
A, and by assumption, we have (M
n
) = 0, n 1. Dene the
set N = M
_

n=1
M
n
_
. It is clear that (N) = 0, and
0 f
1
(x) f
2
(x) f(x), x X N;
f(x) = lim
n
f
n
(x), x X N.
So if we put A = X N, and if we dene the measurable functions g
n
= f
n

A
,
n 1, and g = f
A
, then we have
(a) 0 g
1
g
2
g (everywhere!);
(b) lim
n
g
n
(x) = g(x), x X;
(c) g
n
= f
n
, -a.e., n 1;
(d) g = f, -a.e.;
Notice that property (c) gives g
n
L
1
+
(X, A, ) and
_
X
g
n
d =
n
, n 1. By
property (d), we see that we have the equivalence
f L
1
+
(X, A, ) f L
1
+
(X, A, ).
Moreover, if g L
1
+
(X, A, ), then we will have
_
X
g d =
_
X
f d. These observa-
tions show that it suces to prove the theorem with gs in place of the fs. The
advantage is now the fact that we have the slightly stronger properties (a) and (b)
above. The rst step in the proof is the following.
Claim 1: For every t (0, ), one has the inequality
_
g
1
((t, ])
_


t
.
Denote the set g
1
((t, ]) simply by A
t
. For each n 1, we also dene the set
A
n
t
= g
1
n
((t, ]). Using property (a) above, it is clear that we have the inclusions
(4) A
t
1
A
2
t
A
t
.
Using property (b) above, we also have the equality A
t
=

n=1
A
n
t
. Using the
continuity Lemma 4.1, we then have
(A
t
) = lim
n
(A
n
t
),
so in order to prove the Claim, it suces to prove the inequalities
(5) (A
n
t
)

n
t
, n 1.
But the above inequality is pretty obvious, since we clearly have 0 t
A
n
t
g
n
,
which gives
t(A
n
t
) =
_
X
t
A
n
t
d
_
X
g
n
d =
n
.
Claim 2: For any elementary function h A-Elem
R
(X), with 0 h g,
one has
(i) h L
1
R,elem
(X, A, );
(ii)
_
X
hd .
Start with some elementary function h, with 0 h g. Assume h is not identically
zero, so we can write it as
h =
1

B1
+ +
p

Bp
,
with (B
j
)
p
j=1
A pairwise disjoint, and 0 <
1
< <
p
. Dene the set
B = B
1
B
p
. It is obvious that, if we put t =
1
/2, we have the inclusion
B g
1
_
(t, ]
_
, so by Claim 1, we get (B) < . This gives, of course (B
j
) < ,
j = 1, . . . , p, so h is indeed elementary integrable. To prove the estimate (ii), we
dene the measurable functions h
n
: X [0, ] by h
n
= min{g
n
, h}, n 1.
304 CHAPTER IV: INTEGRATION THEORY
Since 0 h
n
g
n
, n 1, it follows that, h
n
L
1
+
(X, A, ), n 1, and we have
the inequalities
(6)
_
X
h
n
d
_
X
g
n
d =
n
, n 1.
It is obvious that we have
() 0 h
1
h
2
h
p

B
(everywhere);
() h(x) = lim
n
h
n
(x), x X.
Let us restrict everything to B. We consider the -algebra B = A

B
, and the
measure =

B
. Consider the elementary function = h

B
B-Elem
R
(B), as
well as the measurable functions
n
= h
n

B
: B [0, ], n 1. It is clear that
L
1
R,elem
(B, B, ), and we have the equality
(7)
_
B
d =
_
X
hd.
Likewise, using (), which clearly forces h
n

XB
= 0, it follows that, for each n 1,
the function
n
belongs to L
1
+
(B, B, ), and by (6), we have
(8)
_
B

n
d =
_
X
h
n
d, n 1.
Let us analyze the dierences
n
=
n
. On the one hand, using (), we
have
n
(x) [0,
p
], x B, n 1. On the other hand, again by (), we have

1

2
. . . . Finally, by () we have lim
n

n
(x) = 0, x B. We can
apply Lemma 2.1, and we will get lim
n
_
B

n
d = 0. This clearly gives,
_
B
d = lim
n
_
B

n
d,
and then using (7) and (8), we get the equality
_
X
hd = lim
n
_
X
h
n
d.
Combining this with (6), immediately gives the desired estimate
_
X
hd .
Having proven Claim 2, let us observe now that, using the denition of the
positive integral, it follows immediately that g L
1
+
(X, A, ), and we have the
inequality
_
X
g d .
The other inequality is pretty obvious, because the inequality g g
n
forces
_
X
g d
_
X
g
n
d =
n
, n 1,
so we immediately get
_
X
g d sup{
n
; n 1} = .
Comment. In the previous section we introduced the convention which denes
_
X
f d = , if f : X [0, ] is measurable, but f L
1
+
(X, A, ). Using this
convention, the Lebesgue Monotone Convergence Theorem has the following general
version.
2. Convergence theorems 305
Theorem 2.2 (General Lebesgue Monotone Convergence Theorem). Let
(X, A, ) be a measure space, and let f, f
n
: X [0, ], n 1, be measurable
functions, such that
f
n
f
n+1
, -a.e., n 1;
f = -a.e.- lim
n
f
n
.
Then
(9)
_
X
f d = lim
n
_
X
f
n
d.
Proof. As before, the sequence (
n
)

n=1
[0, ], dened by
n
=
_
X
f
n
d,
n 1, is non-decreasing, and is has a limit
= lim
n

n
= sup
__
X
f
n
d : n 1
_
[0, ].
There are two cases to analyze.
Case I: = .
In this case the inequalities f f
n
0, -a.e. will force
_
X
f d
_
X
f
n
d =
n
, n 1,
which will force
_
X
f d , so we indeed get
_
X
f d = = .
Case II: < .
In this case we apply directly Theorem 2.1.
The following result provides an equivalent denition of integrability for non-
negative functions (compare to the construction in Section 1).
Corollary 2.1. Let (X, A, ) be a measure space, and let f : X [0, ] be
a measurable function. The following are equivalent:
(i) f L
1
+
(X, A, );
(ii) there exists a sequence (h
n
)

n=1
L
1
R,elem
(X, A, ), with
0 h
1
h
2
. . . ;
lim
n
h
n
(x) = f(x), x X;
sup
__
X
h
n
d : n 1
_
< .
Moreover, if (h
n
)

n=1
is as in (ii), then one has the equality
(10)
_
X
f d = lim
n
_
X
h
n
d.
Proof. (i) (ii). Assume f L
1
+
(X, A, ). Using Theorem III.3.2, we know
there exists a sequence (h
n
)

n=1
A-Elem
R
(X), with
(a) 0 h
1
h
2
f;
(b) lim
n
h
n
(x) = f(x), x X.
Note the (a) forces h
n
L
1
R,elem
(X, A, ), as well as the inequalities
_
X
h
n
d
_
X
f d < , n 1, so the sequence (h
n
)

n=1
clearly satises condition (ii).
The implication (ii) (i), and the equality (10) immediately follow from the
General Lebesgue Monotone Convergence Theorem.
306 CHAPTER IV: INTEGRATION THEORY
Corollary 2.2 (Fatou Lemma). Let (X, A, ) be a measure space, and let
f
n
: X [0, ], n 1, be a sequence of measurable functions. Dene the function
f : X [0, ] by
f(x) = liminf
n
f
n
(x), x X.
Then f is measurable, and one has the inequality
_
X
f d liminf
n
_
X
f
n
d.
Proof. The fact that f is measurable is already known (see Corollary III.3.5).
Dene the sequence (
n
)

n=1
[0, ] by
n
=
_
X
f
n
d, n 1.
Dene, for each integer n 1, the function g
n
: X [0, ] by
g
n
(x) = inf
_
f
k
(x) : k n
_
, x X.
By Corollary III.3.4, we know that g
n
, n 1 are all measurable. Moreover, it is
clear that
0 g
1
g
2
. . . ;
f(x) = lim
n
g
n
(x), x X.
By the General Lebesgue Monotone Convergence Theorem 2.2, it follows that
(11)
_
X
f d = lim
n
_
X
g
n
d.
Notice that, if we dene the sequence (
n
)

n=1
[0, ], by
n
=
_
X
g
n
d, n 1,
then the obvious inequalities 0 g
n
f
n
give
_
X
g d
_
X
f
n
d, so we get

n

n
, n 1.
Using (11), we then get
_
X
f d = lim
n

n
= liminf
n

n
liminf
n

n
.
The following is an important application of Theorem 2.1, that deals with
Riemann integration.
Corollary 2.3. Let a < b be real numbers. Denote by the Lebesgue measure,
and consider the Lebesgue space
_
[a, b], M

([a, b]),
_
, where M

([a, b]) denotes the

-algebra of all Lebesgue measurable subsets of [a, b]. Then every Riemann inte-
grable function f : [a, b] R belongs to L
1
R
([a, b], M

([a, b]), ), and one has the

equality
(12)
_
[a,b]
f d =
_
b
a
f(x) dx.
Proof. We are going to use the results from III.6. First of all, the fact that f
is Lebesgue integrable, i.e. f belongs to L
1
R
_
[a, b], M

([a, b]),
_
, is clear since f is
Lebesgue measurable, and bounded. (Here we use the fact that the measure space
_
[a, b], M

([a, b]),
_
is nite.)
Next we prove the equality between the Riemann integral and the Lebesgue
integral. Adding a constant, if necessary, we can assume that f 0. For every
partition = (a = x
0
< x
1
< < x
n
= b) of [a, b], we dene the numbers
m
k
= inf
t[x
k1
,x
k
]
f(t), k = 1, . . . , n,
2. Convergence theorems 307
and we dene the function
f

= m
1

[x0,x1]
+m
2

(x1,x2]
+ +m
m

(xn1,xn]
.
Fix a sequence of partitions (
p
)

p=1
, with
1

2
. . . , and lim
p
|
p
| = 0,
We know (see III.6) that we have
f = -a.e.- lim
p
f
p
.
Clearly we have 0 f
1
f
2
f, so by Theorem 2.1, we get
(13)
_
[a,b]
f d = lim
p
_
[a,b]
f
p
d.
Notice however that
_
[a,b]
f
p
d = L(f,
p
), p 1,
where L(f,
p
) denotes the lower Darboux sum. Combining this with (13), and with
the well known properties of Riemann integration, we immediately get (12).
The following is another important convergence theorem.
Theorem 2.3 (Lebesgue Dominated Convergence Theorem). Let (X, A, ) be
a measure space, let K be one of the symbols

R, R, or C, and let (f
n
)

n=1

L
1
K
(X, A, ). Assume f : X K is a measurable function, such that
(i) f = -a.e.- lim
n
f
n
;
(ii) there exists some function g L
1
+
(X, A, ), such that
|f
n
| g, -a.e., n 1.
Then f L
1
K
(X, A, ), and one has the equality
(14)
_
X
f d = lim
n
_
X
f
n
d.
Proof. The fact that f is integrable follows from the following
Claim: |f| g, -a.e.
To prove this fact, we dene, for each n 1, the set
M
n
=
_
x X : |f
n
(x)| > g(x)
_
.
It is clear that M
n
A, and (M
n
) = 0, n 1. If we choose M A such that
(M) = 0, and
f(x) = lim
n
f
n
(x), x X M,
then the set N = M
_

n=1
M
n
_
A will satisfy
(N) = 0;
|f
n
(x)| g(x), x X N;
f(x) = lim
n
f
n
(x), x X N.
We then clearly get
|f(x)| g(x), x X N,
and the Claim follows.
Having proven that f is integrable, we now concentrate on the equality (14).
Case I: K =

R.
308 CHAPTER IV: INTEGRATION THEORY
First of all, without any loss of generality, we can assume that 0 g(x) < ,
x X. (See Lemma 1.2.) Let us dene the functions g
n
= min{f
n
, g} and
h
n
= max{f
n
, g}, n 1. Since we have g f
n
g, -a.e., we immediately get
(15) g
n
= h
n
= f
n
, -a.e., n 1,
thus giving the fact that g
n
, h
n
L
1

R
(X, A, ), n 1, as well as the equalities
(16)
_
X
g
n
d =
_
X
h
n
d =
_
X
f
n
d, n 1.
Dene the measurable functions , : X

R by
(x) = liminf
n
h
n
(x) and (x) = limsup
n
g
n
(x), x X.
Using (15), we clearly have f = = , -a.e., so we get
(17)
_
X
f d =
_
X
d =
_
X
d.
Remark also that we have equalities
(18)
g(x)(x) = liminf
n
[g(x)g
n
(x)] and g(x)+(x) = liminf
n
[g(x)+h
n
(x)], x X.
Since we clearly have
g g
n
0 and g +h
n
0, n 1,
using (18), and Fatou Lemma (Corollary 2.2) and we get the inequalities
_
X
(g ) d liminf
n
_
X
(g g
n
) d,
_
X
(g +) d liminf
n
_
X
(g +h
n
) d,
In other words, we get
_
X
g d
_
X
d liminf
n
_ _
X
g d
_
X
g
n
d
_
=
_
X
g d limsup
n
_
X
g
n
d,
_
X
g d +
_
X
d liminf
n
_ _
X
g d +
_
X
h
n
d
_
=
_
X
g d + liminf
n
_
X
h
n
d.
Using the equalities (16) and (17), the above inequalities give
_
X
f d =
_
X
d limsup
n
_
X
g
n
d = limsup
n
_
X
f
n
d,
_
X
f d =
_
X
d liminf
n
_
X
h
n
d = liminf
n
_
X
f
n
d.
In other words, we have
_
X
f d liminf
n
_
X
f
n
d limsup
n
_
X
f
n
d
_
X
f d,
thus giving the equality (14)
The case K = R is trivial (it is in fact contained in case K =

R).
2. Convergence theorems 309
The case K = C is also pretty clear, using real and imaginary parts, since for
each n 1, we clearly have
|Re f
n
| g, -a.e.,
|Imf
n
| g, -a.e.,.
Exercise 3. Give an example of a sequence of continuous functions f
n
: [0, 1]
[0, ), such that
(a) lim
n
f
n
(x) = 0, n 1;
(b)
_
[0,1]
f
n
d = 1, n 1.
(Here denotes the Lebesgue measure). This shows that the Lebesgue Dominated
Convergence Theorem fails, without the dominance condition (ii).
Hint: Consider the functions fn dened by
fn(x) =
_
_
_
n
2
x if 0 x 1/n
n(2 nx) if 1/n x 2/n
0 if 2/n x 1
The Lebesgue Convergence Theorems 2.2 and 2.3 have many applications. They
are among the most important results in Measure Theory. In many instances, these
theorem are employed during proofs, at key steps. The next two results are good
illustrations.
Proposition 2.1. Let (X, A, ) be a measure space, and let f : X [0, ] be
a measurable function. Then the map
: A A
_
A
f d [0, ]
denes a measure on A.
Proof. It is clear that () = 0. To prove -additivity, start with a pairwise
disjoint sequence (A
n
)

n=1
A, and put A =

n=1
A
n
. For each integer n 1,
dene the set B
n
=

n
k=1
A
k
, and the measurable function g
n
= f
Bn
. Dene also
the function g = f
A
. It is obvious that
0 g
1
g
2
g (everywhere),
lim
n
g
n
(x) = g(x), x X.
Using the General Lebesgue Monotone Convergence Theorem, it follows that
(19) (A) =
_
X
f
A
d =
_
X
g d = lim
n
_
X
g
n
d.
Notice now that, for each n 1, one has the equality
g
n
= f
A1
+ +f
An
,
so using Remark 1.7.C, we get
_
X
g
n
d =
n

k=1
_
X
f
A
k
d =
n

k=1
(A
k
),
so the equality (19) immediately gives (A) =

n=1
(A
n
).
The next result is a version of the previous one for K-valued functions.
Proposition 2.2. Let (X, A, ) be a measure space, let K be one of the symbols

R, R, or C, and let (A
n
)

n=1
A be a pairwise disjoint sequence with

n=1
A
n
= X.
For a function f : X K, the following are equivalent.
310 CHAPTER IV: INTEGRATION THEORY
(i) f L
1
K
(X, A, );
(ii) f

An
L
1
K
(A
n
, A

An
,

An
), n 1, and

n=1
_
An
|f| d < .
Moreover, if f satises these equivalent conditions, then
_
X
f d =

n=1
_
An
f d.
Proof. (i) (ii). Assume f L
1
K
(X, A, ). Applying Proposition 2.1, to
|f|, we immediately get

n=1
_
An
|f| d =
_
X
|f| d < ,
which clearly proves (ii).
(ii) (i). Assume f satises condition (ii). Dene, for each n 1, the set
B
n
= A
1
A
n
. First of all, since (A
n
)

n=1
A, and

n=1
A
n
= X, it follows
that f is measurable. Consider the the functions f
n
= f
Bn
and g
n
= f
An
, n 1.
Notice that, since f

An
L
1
K
(A
n
, A

An
,

An
), it follows that g
n
L
1
K
(X, A, ),
n 1, and we also have
_
X
g
n
d =
_
An
f d, n 1.
In fact we also have
_
X
|g
n
| d =
_
An
|f| d, n 1.
Notice that we obviously have f
n
= g
1
+ +g
n
, and |f
n
| = |g
1
| + +|g
n
|, so if
we dene
S =

n=1
_
An
|f| d,
we get
_
X
|f
n
| d =
n

k=1
_
A
k
|f| d S < , n 1.
Notice however that we have 0 |f
1
| |f
2
| . . . |f|, as well as the equality
lim
n
f
n
(x) = f(x), x X. On the one hand, using the General Lebesgue
Monotone Convergence Theorem, we will get
_
X
|f| d = lim
n
_
X
|f
n
| d = lim
n
_
n

k=1
_
A
k
|f| d
_
=

n=1
_
An
|f| d = S < ,
which proves that |f| L
1
+
(X, A, ), so in particular f belongs to L
1
K
(X, A, ). On
the other hand, since we have |f
n
| |f|, by the Lebesgue Dominated Convergence
Theorem, we get
_
X
f d = lim
n
_
X
f
n
d = lim
n
_
n

k=1
_
A
k
f d
_
=

n=1
_
An
f d.
2. Convergence theorems 311
Corollary 2.4. Let (X, A, ) be a measure space, let K be one of the symbols

R, R, or C, and let (X
n
)

n=1
A be sequence with

n=1
X
n
= X, and X
1

X
2
. . . . For a function f : X K be a measurable function, the following are
equivalent.
(i) f L
1
K
(X, A, );
(ii) f

Xn
L
1
K
(X
n
, A

Xn
,

Xn
), n 1, and
sup
__
Xn
|f| d : n 1
_
< .
Moreover, if f satises these equivalent conditions, then
_
X
f d = lim
n
_
Xn
f d.
Proof. Apply the above result to the sequence (A
n
)

n=1
given by A
1
= X
1
and A
n
= X
n
X
n1
, n 2.
Remark 2.2. Suppose (X, A, ) is a measure space, K is one of the elds R or
C, and f L
1
K
(X, A, ). By Proposition 2.2, we get the fact that the map
: A A
_
A
f d K
is a K-valued measure on A. By Proposition 2.1, we also know that
: A A
_
A
|f| d K
is a nite honest measure on A. Using Proposition 1.6, we clearly have
|(A)| =

_
A
f d

_
A
|f| d = (A), A A,
which by the results from III.8 gives the inequality || . (Here || denotes the
variation measure of .) Later on (see Section 4) we are going to see that in fact
we have the equality || = .
Comment. It is important to understand the sequential nature of the con-
vergence theorems discussed here. If we examine for instance the Mononotone
Convergence Theorem, we could easily formulate a series version, which states
the equality
_
X
_

n=1
f
n
_
d =

n=1
_
X
f
n
d,
for any sequence measurable functions f
n
: X [0, ].
Suppose now we have an arbitrary family f
j
: X [0, ], j J of measurable
functions, and we dene
f(x) =

jJ
f
j
(x), x X.
(Here we use the summability convention which denes the sum as the supremum
of all nite sums.) In general, f is not always measurable. But if it is, one still
cannot conclude that
_
X
f d =

jJ
_
X
f
i
d.
312 CHAPTER IV: INTEGRATION THEORY
The following example illustrates this anomaly.
Example 2.1. Take the measure space ([0, 1], M

([0, 1]), ), and x J [0, 1]

and arbitrary set. For each j J we consider tha characteristic function f
j
=
{j}
.
It is obvious that the function f : X [0, ], dened by
f(x) =

jJ
f
j
(x), x [0, 1],
is equal to
J
If J is non-measurable, this already gives an example when f =

jJ
f
j
is non-measurable. But even if J were measurable, it would be impossible
to have the equality
_
X
f d =

jJ
_
X
f
j
d,
simply because the right hand side is zero, while the left hand side is equal to (J).
The next two exercises illustrate straightforward (but nevertheless interesting)
applications of the convergence theorems to quite simple situations.
Exercise 4. Let A be a -algebra on a (non-empty) set X, and let (
n
)

n=1
be
a sequence of signed measures on A. Assume that, for each A A, the sequence
_

n
(A)
_

n=1
has a limit denoted (A) [, ]. Prove that the map : A
[0, ] denes a measure on A, if the sequence (
n
)

n=1
satises one of the following
hypotheses:
A. 0
1
(A)
2
(A) . . . , A A;
B. there exists a nite measure on A, such that |
n
(A)| (A), n 1,
A A.
Hint: To prove -additivity, x a pairwise disjoint sequence (A
k
)

k=1
A, and put A =

k=1
A
k
.
Treat the problem of proving the equality (A) =

k=1
(A
k
) as a convergence problem on
the measure space (N, P(N), ) - with the counting measure - for the sequence of functions
fn : N [0, ] dened by fn(k) = n(A
k
), k N.
Exercise 5*. Let A be a -algebra on a (non-empty) set X, and let (
j
)
jJ
be
a family of signed measures on A. Assume either of the following is true:
A.
j
(A) 0, j J, A A.
B. There exists a nite measure on A, such that

jJ
|
j
(A)| (A),
A A.
Dene the map : A [0, ] by (A) =

jJ

j
(A), A A. (In Case A, the
sum is dened as the supremum over nite sums. In case B, it follows that the
family
_

j
(A)
_
jJ
is summable.) Prove that is a measure on A.
Hint: To prove -additivity, x a pairwise disjoint sequence (A
k
)

k=1
A, and put A =

k=1
A
k
.
To prove the equality (A) =

k=1
(A
k
), analyze the following cases: (i) There is some k 1,
such that (A
k
) = ; (ii) (A
k
) < , k 1. The rst case is quite trivial. In the second
case reduce the problem to the previous exercise, by observing that, for each k 1, the set
J(A
k
) = {j J :
j
(A
k
) > 0
_
must be countable. Then the set J(A) = {j J :
j
(A) > 0} is
also countable.
Comment. One of the major drawbacks of the theory of Riemann integration
is illustrated by the approach to improper integration. Recall that for a function
2. Convergence theorems 313
h : [a, b) R the improper Riemann integral is dened as
_
b
a
h(t) dt = lim
xb
_
x
a
f(t) dt,
provided that
(a) h

[a,x]
is Riemann integrable, x (a, b), and
(b) the above limit exists.
The problem is that although the improper integral may exist, and the function is
actually dened on [a, b], it may fail to be Riemann integrable, for example when
it is unbounded.
In contrast to this situation, by Corollary 2.4, we see that if for example h 0,
then the Lebesgue integrability of h on [a, b] is equivalent to the fact that
(i) h

[a,x]
is Lebesgue integrable, x (a, b), and
(ii) lim
xb
_
[a,x]
hd exists.
Going back to the discussion on improper Riemann integral, we can see that
a sucient condition for h : [a, b) R to be Riemann integrable in the improper
sense, is the fact that h has property (a) above, and h is Lebesgue integrable on
[a, b). In fact, if h 0, then by Corollary 2.4, this is also necessary.
Notation. Let a < b , and let f be a Lebesgue integrable function,
dened on some interval J which is one of (a, b), [a, b), (a, b], or [a, b]. Then the
Lebesgue integral
_
J
f d will be denoted simply by
_
b
a
f d.
Exercise 6*. Let (X, A, ) be a nite measure space. Prove that for every
f L
1
+
(X, A, ), one has the equality
_
X
f d =
_

0

_
f
1
([t, ])
_
dt,
where the second term is dened as improper Riemann integral.
Hint: The function : [0, ) [0, ) dened by (t) =
_
f
1
([t, ])
_
, t 0, is non-
increasing, so it is Riemann integrable on every interval [0, a], a > 0. Prove the inequalities
_
Xa
f d
_
a
0
(t) dt
_
X
f d, a > 0,
where Xa = f
1
([0, a)), by analyzing lower and upper Darboux sums of

[0,a]
. Use Corollary
2.4 to get lima
_
Xa
f d =
_
X
f d.