Measure

MEASURE AND INTEGRATION
Dietmar A. Salamon
ETH Z urich
9 September 2016
ii
Preface
This book is based on notes for the lecture course Measure and Integration
held at ETH Z urich in the spring semester 2014. Prerequisites are the first
year courses on Analysis and Linear Algebra, including the Riemann inte-
gral [9, 18, 19, 21], as well as some basic knowledge of metric and topological
spaces. The course material is based in large parts on Chapters 1-8 of the
textbook Real and Complex Analysis by Walter Rudin [17]. In addition
to Rudins book the lecture notes by Urs Lang [10, 11], the five volumes on
measure theory by David H. Fremlin [4], the paper by Heinz Konig [8] on
the generalized RadonNikod ym theorem, the lecture notes by C.E. Heil [7]
on absolutely continuous functions, Dan Mas Topology Blog [12] on exotic
examples of topological spaces, and the paper by Gert K. Pedersen [16] on
the Haar measure were very helpful in preparing this manuscript.
This manuscript also contains some material that was not covered in the
lecture course, namely some of the results in Sections 4.5 and 5.2 (concerning
the dual space of Lp () in the non -finite case), Section 5.4 on the Gen-
eralized RadonNikod ym Theorem, Sections 7.6 and 7.7 on Marcinkiewicz
interpolation and the CalderonZygmund inequality, and Chapter 8 on the
Haar measure.
I am grateful to many people who helped to improve this manuscript.
Thanks to the students at ETH who pointed out typos or errors in earlier
drafts. Thanks to Andreas Leiser for his careful proofreading. Thanks to
Theo Buehler for many enlightening discussions and for pointing out the
book by Fremlin, Dan Mas Topology Blog, and the paper by Pedersen.
Thanks to Urs Lang for his insightful comments on the construction of the
Haar measure.
1 August 2015 Dietmar A. Salamon
iii
iv
Contents
Introduction 1
1 Abstract Measure Theory 3

1.1 -Algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
1.2 Measurable Functions . . . . . . . . . . . . . . . . . . . . . . . 11
1.3 Integration of Nonnegative Functions . . . . . . . . . . . . . . 17
1.4 Integration of Real Valued Functions . . . . . . . . . . . . . . 29
1.5 Sets of Measure Zero . . . . . . . . . . . . . . . . . . . . . . . 33
1.6 Completion of a Measure Space . . . . . . . . . . . . . . . . . 39
1.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
2 The Lebesgue Measure 49

2.1 Outer Measures . . . . . . . . . . . . . . . . . . . . . . . . . . 50
2.2 The Lebesgue Outer Measure . . . . . . . . . . . . . . . . . . 56
2.3 The Transformation Formula . . . . . . . . . . . . . . . . . . . 67
2.4 Lebesgue Equals Riemann . . . . . . . . . . . . . . . . . . . . 75
2.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78
3 Borel Measures 81
3.1 Regular Borel Measures . . . . . . . . . . . . . . . . . . . . . 81
3.2 Borel Outer Measures . . . . . . . . . . . . . . . . . . . . . . . 92
3.3 The Riesz Representation Theorem . . . . . . . . . . . . . . . 97
3.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
4 Lp Spaces 113
4.1 Holder and Minkowski . . . . . . . . . . . . . . . . . . . . . . 113
4.2 The Banach Space Lp () . . . . . . . . . . . . . . . . . . . . . 115
4.3 Separability . . . . . . . . . . . . . . . . . . . . . . . . . . . . 120
v
vi CONTENTS
4.4 Hilbert Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

4.5 The Dual Space of Lp () . . . . . . . . . . . . . . . . . . . . . 129
4.6 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143
5 The RadonNikod ym Theorem 151

5.1 Absolutely Continuous Measures . . . . . . . . . . . . . . . . . 151
5.2 The Dual Space of Lp () Revisited . . . . . . . . . . . . . . . 159
5.3 Signed Measures . . . . . . . . . . . . . . . . . . . . . . . . . 166
5.4 RadonNikod ym Generalized . . . . . . . . . . . . . . . . . . 173
5.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180
6 Differentiation 185
6.1 Weakly Integrable Functions . . . . . . . . . . . . . . . . . . . 185
6.2 Maximal Functions . . . . . . . . . . . . . . . . . . . . . . . . 190
6.3 Lebesgue Points . . . . . . . . . . . . . . . . . . . . . . . . . . 196
6.4 Absolutely Continuous Functions . . . . . . . . . . . . . . . . 201
6.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 205
7 Product Measures 209

7.1 The Product -Algebra . . . . . . . . . . . . . . . . . . . . . . 209
7.2 The Product Measure . . . . . . . . . . . . . . . . . . . . . . . 214
7.3 Fubinis Theorem . . . . . . . . . . . . . . . . . . . . . . . . . 219
7.4 Fubini and Lebesgue . . . . . . . . . . . . . . . . . . . . . . . 228
7.5 Convolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . 231
7.6 Marcinkiewicz Interpolation . . . . . . . . . . . . . . . . . . . 239
7.7 The CalderonZygmund Inequality . . . . . . . . . . . . . . . 243
7.8 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 255
8 The Haar Measure 259

8.1 Topological Groups . . . . . . . . . . . . . . . . . . . . . . . . 259
8.2 Haar Measures . . . . . . . . . . . . . . . . . . . . . . . . . . 263
A Urysohns Lemma 279
B The Product Topology 285
C The Inverse Function Theorem 287
References 289
Introduction
We learn already in high school that integration plays a central role in math-
ematics and physics. One encounters integrals in the notions of area or
volume, when solving a differential equation, in the fundamental theorem of
calculus, in Stokes theorem, or in classical and quantum mechanics. The
first year analysis course at ETH includes an introduction to the Riemann
integral, which is satisfactory for many applications. However, it has certain
drawbacks, in that some very basic functions are not Riemann integrable,
that the pointwise limit of a sequence of Riemann integrable functions need
not be Riemann integrable, and that the space of Riemann integrable func-
tions is not complete with respect to the L1 -norm. One purpose of this book
is to introduce the Lebesgue integral, which does not suffer from these draw-
backs and agrees with the Riemann integral whenever the latter is defined.
Chapter 1 introduces abstract integration theory for functions on measure
spaces. It includes proofs of the Lebesgue Monotone Convergence Theorem,
the Lemma of Fatou, and the Lebesgue Dominated Convergence Theorem.
In Chapter 2 we move on to outer measures and introduce the Lebesgue
measure on Euclidean space. Borel measures on locally compact Hausdorff
spaces are the subject of Chapter 3. Here the central result is the Riesz
Representation Theorem. In Chapter 4 we encounter Lp spaces and show
that the compactly supported continuous functions form a dense subspace of
Lp for a regular Borel measure on a locally compact Hausdorff space when
p < . Chapter 5 is devoted to the proof of the RadonNikod ym theorem
q
about absolutely continuous measures and to the proof that L is naturally
isomorphic to the dual space of Lp when 1/p + 1/q = 1 and 1 < p < .
Chapter 6 deals with differentiation. Chapter 7 introduces product measures
and contains a proof of Fubinis Theorem, an introduction to the convolu-
tion product on L1 (Rn ), and a proof of the CalderonZygmund inequality.
Chapter 8 constructs Haar measures on locally compact Hausdorff groups.
1
2 CONTENTS
Despite the overlap with the book of Rudin [17] there are some differ-
ences in exposition and content. A small expository difference is that in
Chapter 1 measurable functions are defined in terms of pre-images of (Borel)
measurable sets rather than pre-images of open sets. The Lebesgue measure
in Chapter 2 is introduced in terms of the Lebesgue outer measure instead of
as a corollary of the Riesz Representation Theorem. The notion of a Radon
measure on a locally compact Hausdorff space in Chapter 3 is defined in
terms of inner regularity, rather than outer regularity together with inner
regularity on open sets. This leads to a somewhat different formulation of
the Riesz Representation Theorem (which includes the result as formulated
by Rudin). In Chapters 4 and 5 it is shown that Lq () is isomorphic to
the dual space of Lp () for all measure spaces (not just the -finite ones)
whenever 1 < p < and 1/p + 1/q = 1. It is also shown that L () is
isomorphic to the dual space of L1 () if and only if the measure space is
localizable. Chapter 5 includes a generalized version of the RadonNikod ym
theorem for signed measures, due to Fremlin [4], which does not require that
the underying measure is -finite. In the formulation of Konig [8] it asserts
that a signed measure admits a -density if and only if it is both absolutely
continuous and inner regular with respect to . In addition the present
book includes a self-contained proof of the CalderonZygmund inequality in
Chapter 7 and an existence and uniqueness proof for (left and right) Haar
measures on locally compact Hausdorff groups in Chapter 8.
The book is intended as a companion for a foundational one semester
lecture course on measure and integration and there are many topics that it
does not cover. For example the subject of probability theory is only touched
upon briefly at the end of Chapter 1 and the interested reader is referred to
the book of Malliavin [13] which covers many additional topics including
Fourier analysis, limit theorems in probability theory, Sobolev spaces, and
the stochastic calculus of variations. Many other fields of mathematics re-
quire the basic notions of measure and integration. They include functional
analysis and partial differential equations (see e.g. GilbargTrudinger [5]),
geometric measure theory, geometric group theory, ergodic theory and dy-
namical systems, and differential topology and geometry.
There are many other textbooks on measure theory that cover most or
all of the material in the present book, as well as much more, perhaps from
somewhat different view points. They include the book of Bogachev [2]
which also contains many historical references, the book of Halmos [6], and
the aforementioned books of Fremlin [4], Malliavin [13], and Rudin [17].
Chapter 1
Abstract Measure Theory
The purpose of this first chapter is to introduce integration on abstract mea-

sure spaces. The basic idea is to assign to a real valued function on a given
domain a number that gives a reasonable meaning to the notion of area un-
der the graph. For example, to the characteristic function of a subset of the
domain one would want to assign the length or area or volume of that subset.
To carry this out one needs a sensible notion of measuring the size of the sub-
sets of a given domain. Formally this can take the form of a function which
assigns a nonnegative real number, possibly also infinity, to each subset of
our domain. This function should have the property that the measure of a
disjoint union of subsets is the sum of the measures of the individual subsets.
However, as is the case with many beautiful ideas, this naive approach does
not work. Consider for example the notion of the length of an interval of real
numbers. In this situation each single point has measure zero. With the ad-
ditivity requirement it would then follow that every subset of the reals, when
expressed as the disjoint union of all its elements, must also have measure
zero, thus defeating the original purpose of defining the length of an arbitrary
subset of the reals. This reasoning carries over to any dimension and makes
it impossible to define the familiar notions of area or volume in the manner
outlined above. To find a way around this, it helps to recall the basic obser-
vation that any uncountable sum of positive real numbers must be infinity.
Namely, if we are given a collection of positive real numbers whose sum is
finite, then only finitely many of these numbers can be bigger than 1/n for
each natural number n, and so it can only be a countable collection. Thus it
makes sense to demand additivity only for countable collections of disjoint
sets.
3
4 CHAPTER 1. ABSTRACT MEASURE THEORY
Even with the restricted concept of countable additivity it will not be

possible to assign a measure to every subset of the reals and recover the
notion of the length of an interval. For example, call two real numbers
equivalent if their difference is rational, and let E be a subset of the half
unit interval that contains precisely one element of each equivalence class.
Since each equivalence class has a nonempty intersection with the half unit
interval, such a set exists by the Axiom of Choice. Assume that all translates
of E have the same measure. Then countable additivity would imply that
the unit interval has measure zero or infinity.
One way out of this dilemma is to give up on the idea of countable ad-
ditivity and replace it by the weaker requirement of countable subadditivity.
This leads to the notion of an outer measure which will be discussed in Chap-
ter 2. Another way out is to retain the requirement of countable additivity
but give up on the idea of assigning a measure to every subset of a given
domain. Instead one assigns a measure only to some subsets which are then
called measurable. This idea will be pursued in the present chapter. A sub-
tlety of this approach is that in some important cases it is not possible to give
an explicit description of those subsets of a given domain that one wants to
measure, and instead one can only impose certain axioms that the collection
of all measurable sets must satisfy. By contrast, in topology the open sets
can often be described explicitly. For example the open subsets of the real
line are countable unions of open intervals, while there is no such explicit
description for the Borel measurable subsets of the real line.
The precise formulation of this approach leads to the notion of a -algebra
which is discussed in Section 1.1. Section 1.2 introduces measurable functions
and examines their basic properties. Measures and the integrals of positive
measurable functions are the subject of Section 1.3. Here the nontrivial part
is to establish additivity of the integral and the proof is based on the Lebesgue
Monotone Convergence Theorem. An important inequality is the Lemma of
Fatou. It is needed to prove the Lebesgue Dominated Convergence Theorem
in Section 1.4 for real valued integrable functions. Section 1.5 deals with sets
of measure zero which are negligible for many purposes. For example, it is
often convenient to identify two measurable functions if they agree almost
everywhere, i.e. on the complement of a set of measure zero. This defines
an equivalence relation. The quotient of the space of integrable functions by
this equivalence relation is a Banach space and is denoted by L1 . Section 1.6
discusses the completion of a measure space. Here the idea is to declare every
subset of a set of measure zero to be measurable as well.
1.1. -ALGEBRAS 5
1.1 -Algebras
For any fixed set X denote by 2X the set of all subsets of X and, for any
subset A X, denote by Ac := X \ A its complement.
Definition 1.1 (Measurable Space). Let X be a set. A collection A 2X
of subsets of X is called a -algebra if it satisfies the following axioms.
(a) X A.
(b) If A A then Ac A.
(c) Every countable union of elements of A is again an element of A, i.e. if
Ai A for i = 1, 2, 3, . . . then
S
i=1 i A.
A
A measurable space is a pair (X, A) consisting of a set X and a -algebra
A 2X . The elements of a -algebra A are called measurable sets.
Lemma 1.2. Every -algebra A 2X satisfies the following.
(d) A.
(e) If n N and A1 , . . . , An A then ni=1 Ai A.
S
(f ) Every finite or countable intersection of elements of A is an element

of A.
(g) If A, B A then A \ B A.
Proof. Condition (d) follows from (a), (b) because X c = , and (e) follows
from (c), (d) by taking Ai :=S for i > n. Condition (f) follows from (b),
(c), (e) because ( i Ai )c = i Aci , and (g) follows from (b), (f) because
T
A \ B = A B c . This proves Lemma 1.2.
Example 1.3. The sets A := {, X} and A := 2X are -algebras.
Example 1.4. Let X be an uncountable set. Then the collection A 2X
of all subsets A X such that either A or Ac is countable is a -algebra.
(Here countable means finite or countably infinite.)
Example 1.5. Let X be a set and let {Ai }iI be a partition of X, i.e.
a nonempty subset of X forSeach i I, Ai Aj = for i 6= j, and
Ai is S
X = iI Ai . Then A := {AJ := jJ Aj | J I} is a -algebra.
Exercise 1.6. (i) Let X be a set and let A, B X be subsets such that
the four sets A \ B, B \ A, A B, X \ (A B) are nonempty. What is the
cardinality of the smallest -algebra A X containing A and B?
(ii) How many -algebras on X are there when #X = k for k = 0, 1, 2, 3, 4?
(iii) Is there an infinite -algebra with countable cardinality?
Exercise 1.7. Let X be any set and let I be any nonempty index set.
Suppose that for every
T i I a -algebra Ai 2X is given. Prove that the
intersection A := iI Ai = {A X | A Ai for all i I} is a -algebra.
Lemma 1.8. Let X be a set and E 2X be any set of subsets of X. Then
there is a unique smallest -algebra A 2X containing E (i.e. A is a -
algebra, E A, and if B is any other -algebra with E B then A B).
Proof. Uniqueness follows directly from the definition. Namely, if A and B
are two smallest -algebras containing E, we have both B A and A B
and hence A = B. To prove existence, denote by S 22 the collection of
X
all -algebras B 2X that contain E and define

\ if B 2X is a -algebra
A := B = A X .
such that E B then A B
BS
Thus A is a -algebra by Exercise 1.7. Moreover, it follows directly from the

definition of A that E A and that every -algebra B that contains E also
contains A. This proves Lemma 1.8.
Lemma 1.8 is a useful tool to construct nontrivial -algebras. Before doing
that let us first take a closer look at Definition 1.1. The letter stands for
countable and the crucial observation is that axiom (c) allows for countable
unions. On the one hand this is a lot more general than only allowing for
finite unions, which would be the subject of Boolean algebra. On the other
hand it is a lot more restrictive than allowing for arbitrary unions, which one
encounters in the subject of topology. Topological spaces will play a central
role in this book and we recall here the formal definition.
Definition 1.9 (Topological Space). Let X be a set. A collection U 2X
of subsets of X is called a topology on X if it satisfies the following axioms.
(a) , X U.
(b) If n N and U1 , . . . , Un U then ni=1 Ui U.
T
S
(c) If I is any index set and Ui U for i I then iI Ui U.
A topological space is a pair (X, U) consisting of a set X and a topology
U 2X . If (X, U) is a topological space, the elements of U are called open
sets, and a subset F X is called closed if its complement is open, i.e.
F c U. Thus finite intersections of open sets are open and arbitrary unions
of open sets are open. Likewise, finite unions of closed sets are closed and
arbitrary intersections of closed sets are closed.
1.1. -ALGEBRAS 7
Conditions (a) and (b) in Definition 1.9 are also properties of every -
algebra. However, condition (c) in Definition 1.9 is not shared by -algebras
because it permits arbitrary unions. On the other hand, complements of
open sets are typically not open. Many of the topologies used in this book
arise from metric spaces and are familiar from first year analysis. Here is a
recollection of the definition.
Definition 1.10 (Metric Space). A metric space is a pair (X, d) con-
sisting of a set X and a function d : X X R satisfying the following
axioms.
(a) d(x, y) 0 for all x, y X, with equality if and only if x = y.
(b) d(x, y) = d(y, x) for all x, y X.
(c) d(x, z) d(x, y) + d(y, z) for all x, y, z X.
A function d : X X R that satisfies these axioms is called a distance
function and the inequality in (c) is called the triangle inequality. A
subset U X of a metric space (X, d) is called open (or d-open) if, for
every x U , there exists a constant > 0 such that the open ball
B (x) := B (x, d) := {y X | d(x, y) < }
(centered at x with radius ) is contained in U . The collection of d-open

subsets of X will be denoted by U(X, d) := {U X | U is d-open} .
It follows directly from the definitions that the collection U(X, d) 2X
of d-open sets in a metric space (X, d) satisfies the axioms of a topology in
Definition 1.9. A subset F of a metric space (X, d) is closed if and only if
the limit point of every convergent sequence in F is itself contained in F .
Example 1.11. A normed vector space is a pair (X, kk) consisting of a
real vector space X and a function X R : x 7 kxk satisfying the following.
(a) kxk 0 for all x X, with equality if and only if x = 0.
(b) kxk = || kxk for all x X and R.
(c) kx + yk kxk + kyk for all x, y X.
Let (X, kk) be a normed vector space. Then the formula
d(x, y) := kx yk
defines a distance function on X. X is called a Banach space if the metric

space (X, d) is complete, i.e. if every Cauchy sequence in X converges.
Example 1.12. The set X = R of real numbers is a metric space with the
standard distance function
d(x, y) := |x y|.
The topology on R induced by this distance function is called the standard

topology on R. The open sets in the standard topology are unions of open
intervals. Exercise: Every union of open intervals is a countable union of
open intervals.
Exercise 1.13. Consider the set
R := [, ] := R {, }.
For a, b R define
(a, ] := (a, ) {}, [, b) := (, b) {}.
Call a subset U R open if it is a countable union of open intervals in R

and sets of the form (a, ] or [, b) for a, b R.
(i) Show that the set of open subsets of R satisfies the axioms of a topology.
This is called the standard topology on R.
(ii) Prove that the standard topology on R is induced by the distance function
d : R R R, defined by the following formulas for x, y R:
2|exy eyx |
d(x, y) :=
ex+y + exy + eyx + exy
2ex
d(x, ) := d(, x) := x ,
e + ex
2ex
d(x, ) := d(, x) := x ,
e + ex
d(, ) := d(, ) := 2.
(iii) Prove that the map f : R [1, 1] defined by

ex ex
f (x) := tanh(x) := , f () := 1,
ex + ex
for x R is a homeomorphism. Prove that it is an isometry with respect
to the metric in (ii) on R and the standard metric on the interval [1, 1].
Deduce that (R, d) is a compact metric space.
1.1. -ALGEBRAS 9
Exercise 1.14. Extend the total ordering of R to R by a

for a R. Extend addition by + a := for < a and by
+ a := for a < . (The sum a + b is undefined when
{a, b} = {, }.) Let a1 , a2 , a3 , . . . and b1 , b2 , b3 , . . . be sequences in R.
(i) Define lim supn an and lim inf n an and show that they always exist.
(ii) Show that lim supn (an ) = lim inf n an .
(iii) Assume {an , bn } 6= {, } so the sum an + bn is defined for n N.
Prove the inequality
lim sup(an + bn ) lim sup an + lim sup bn ,
n n n
whenever the right hand side exists. Find an example where the inequality
is strict.
(iv) If an bn for all n N show that lim inf n an lim inf n bn .
Definition 1.15. Let (X, U) be a topological space and let B 2X be the
smallest -algebra containing U. Then B is called the Borel -algebra of
(X, U) and the elements of B are called Borel (measurable) sets.
Lemma 1.16. Let (X, U) be a topological space. Then the following holds.
(i) Every closed subset F X is a Borel set.
(ii) Every countable union
S
i=1 Fi of closed subsets Fi X is a Borel set.
(These are sometimes called F -sets.)
(iii) Every countable intersection
T
i=1 Ui of open subsets Ui X is a Borel
set. (These are sometimes called G -sets.)
Proof. Part (i) follows from the definition of Borel sets and condition (b) in
Definition 1.1, part (ii) follows from (i) and (c), and part (iii) follows from (ii)
and (b), because the complement of an F -set is a G -set.
Consider for example the Borel -algebra on the real axis R with its stan-
dard topology. In view of Lemma 1.16 it is a legitimate question whether
there is any subset of R at all that is not a Borel set. The answer to this
question is positive, which may not be surprising, however the proof of the
existence of subsets that are not Borel sets is surprisingly nontrivial. It will
only appear much later in this book, after we have introduced the Lebesgue
measure (see Lemma 2.15). For now it is useful to note that, roughly speak-
ing, every set that one can construct in terms of some explicit formula, will
be a Borel set, and one can only prove with the Axiom of Choice that subsets
of R must exist that are not Borel sets.
Recollections About Point Set Topology

We close this section with a digression into some basic notions in topology
that, at least for metric spaces, are familiar from first year analysis and will be
used throughout this book. The two concepts we recall here are compactness
and continuity. A subset K X of a metric space (X, d) is called compact
if every sequence in K has a subsequence that converges to some element
of K. Thus, in particular, every compact subset is closed. The notion of
compactness carries over to general topological spaces as follows.
Let (X, U) be a topological space and let K X. An open cover
of K S is a collection of open sets {Ui }iI , indexed by a set I, such that
K iI Ui . The set K is called compact if every open cover of K has a
finite subcover, i.e. if for every open cover {Ui }iI of K there exist finitely
many indices i1 , . . . , in I such that K Ui1 Uin . When (X, d) is
a metric space and U = U(X, d) is the topology induced by the distance
function (Definition 1.10), the two notions of compactness agree. Thus, for
every subset K X, every sequence in K has a subsequence converging to an
element of K if and only if every open cover of K has a finite subcover. For a
proof see for example Munkres [14] or [20, Appendix C.1]. We emphasize that
when K is a compact subset of a general topological space (X, U) it does not
follow that K is closed. For example a finite subset of X is always compact
but need not be closed or, if U = {, X} then every subset of X is compact
but only the empty set and X itself are closed subsets of X. If, however,
(X, U) is a Hausdorff space (i.e. for any two distinct points x, y X there
exist open sets U, V U such that x U , y V , and U V = ) then every
compact subset of X is closed (Lemma A.2).
Next recall that a map f : X Y between two metric spaces (X, dX )
and (Y, dY ) is continuous (i.e. for every x X and every > 0 there is a
> 0 such that f (B (x, dX )) B (f (x), dY )) if and only if the pre-image
f 1 (V ) := {x X | f (x) V } of every open subset of Y is an open subset
of X. This second notion carries over to general topological spaces, i.e. a
map f : X Y between topological spaces (X, UX ) and (Y, UY ) is called
continuous if V UY = f 1 (V ) UX . It follows directly from the
definition that topological spaces form a category, in that the composition
g f : X Z of two continuous maps f : X Y and g : Y Z between
topological spaces is again continuous. Another basic observation is that
if f : X Y is a continuous map between topological spaces and K is a
compact subset of X then its image f (K) is a compact subset of Y .
1.2. MEASURABLE FUNCTIONS 11
1.2 Measurable Functions

In analogy to continuous maps between topological spaces one can define
measurable maps between measurable spaces as those maps under which pre-
images of measurable sets are again measurable. A slightly different approach
is taken by Rudin [17] who defines a measurable map from a measurable
space to a topological space as one under which pre-images of open sets are
measurable. Both definitions agree whenever the target space is equipped
with its Borel -algebra.
As a warmup we begin with some recollections about pre-images of sets
that are also relevant for the discussion on page 10. For any map f : X Y
between two sets X and Y and any subset B Y , the pre-image
f 1 (B) := {x X | f (x) B}
of B under f is a well defined subset of X, whether or not the map f is

bijective, i.e. even if there does not exist any map f 1 : Y X. The
pre-image defines a map from 2Y to 2X . It satisfies
f 1 (Y ) = X, f 1 () = , (1.1)
and preserves union, intersection, and complement. Thus
f 1 (Y \ B) = X \ f 1 (B) (1.2)
for every subset B Y and

! !
[ [ \ \
f 1 Bi = f 1 (Bi ), f 1 Bi = f 1 (Bi ) (1.3)
iI iI iI iI
for every collection of subsets Bi Y , indexed by a set I.

Definition 1.17 (Measurable Function). (i) Let (X, AX ) and (Y, AY ) be
measurable spaces. A map f : X Y is called measurable if the pre-image
of every measurable subset of Y under f is a measurable subset of X, i.e.
B AY = f 1 (B) AX .
(ii) Let (X, AX ) be a measurable space. A function f : X R is called

measurable if it is measurable with respect to the Borel -algebra on R
associated to the standard topology in Exercise 1.13 (see Definition 1.15).
(iii) Let (X, UX ) and (Y, UY ) be topological spaces. A map f : X Y is

called Borel measurable if the pre-image of every Borel measurable subset
of Y under f is a Borel measurable subset of X.
Example 1.18. Let X be a set. The characteristic function of a subset
A X is the function A : X R defined by

1, if x A,
A (x) := (1.4)
0, if x
/ A.
Now assume (X, A) is a measurable space, consider the Borel -algebra on R,
and let A X be any subset. Then A is a measurable function if and only
if A is a measurable set.
Part (iii) in Definition 1.17 is the special case of part (i), where AX 2X
and AY 2Y are the -algebras of Borel sets (see Definition 1.15). The-
orem 1.20 below shows that every continuous function between topological
spaces is Borel measurable. It also shows that a function from a measur-
able space to a topological space is measurable with respect to the Borel
-algebra on the target space if and only if the pre-image of every open set is
measurable. Since the collection of Borel sets is in general much larger than
the collection of open sets, the collection of measurable functions is then also
much larger than the collection of continuous functions.
Theorem 1.19 (Measurable Maps).
Let (X, AX ), (Y, AY ), and (Z, AZ ) be measurable spaces.
(i) The identity map idX : X X is measurable.
(ii) If f : X Y and g : Y Z are measurable maps then so is the
composition g f : X Z.
(iii) Let f : X Y be any map. Then the set
f AX := B Y | f 1 (B) AX

(1.5)
is a -algebra on Y , called the pushforward of AX under f .
(iv) A map f : X Y is measurable if and only if AY f AX .
Proof. Parts (i) and (ii) follow directly from the definitions. That the set
f AX 2Y defined by (1.5) is a -algebra follows from equation (1.1) (for
axiom (a)), equation (1.2) (for axiom (b)), and equation (1.3) (for axiom (c)).
This proves part (iii). Moreover, by Definition 1.17 f is measurable if and
only if f 1 (B) AX for every B AY and this means that AY f AX .
This proves part (iv) and Theorem 1.19.
Theorem 1.20 (Measurable and Continuous Maps). Let (X, AX ) and

(Y, AY ) be measurable spaces. Assume UY 2Y is a topology on Y such
that AY is the Borel -algebra of (Y, UY ).
(i) A map f : X Y is measurable if an only if the pre-image of every open
subset V Y under f is measurable, i.e.
V UY = f 1 (V ) AX .
(ii) Assume UX 2X is a topology on X such that AX is the Borel -algebra

of (X, UX ). Then every continuous map f : X Y is (Borel) measurable.
Proof. By part (iv) of Theorem 1.19 a map f : X Y is measurable if

and only if AY f AX . Since f AX is a -algebra on Y by part (iii) of
Theorem 1.19, and the Borel -algebra AY is the smallest -algebra on Y
containing the collection of open sets UY by Definition 1.15, it follows that
AY f AX if and only if UY f AX . By the definition of f AX in (1.5),
this translates into the condition V UY = f 1 (V ) AX . This proves
part (i). If in addition AX is the Borel -algebra of a topology UX on X and
f : (X, UX ) (Y, UY ) is a continuous map then the pre-image of every open
subset V Y under f is an open subset of X and hence is a Borel subset
of X; thus it follows from part (i) that f is Borel measurable. This proves
part (ii) and Theorem 1.20.
Theorem 1.21 (Characterization of Measurable Functions).

Let (X, A) be a measurable space and let f : X R be any function. Then
the following are equivalent.
(i) f is measurable.
(ii) f 1 ((a, ]) is a measurable subset of X for every a R.
(iii) f 1 ([a, ]) is a measurable subset of X for every a R.
(iv) f 1 ([, b)) is a measurable subset of X for every b R.
(v) f 1 ([, b]) is a measurable subset of X for every b R.
Proof. That (i) implies (ii), (iii), (iv), and (v) follows directly from the def-
initions. We prove that (ii) implies (i). Thus let f : X R be a function
such that f 1 ((a, ]) AX for every a R and define
B := f AX = B R | f 1 (B) AX 2R .

Then B is a -algebra on R by part (iii) of Theorem 1.19 and (a, ] B for

every a R by assumption. Hence [, b] = R \ (b, ] B for every b R
by axiom (b) and hence
[
[, b) = [, b n1 ] B
nN
by axiom (c) in Definition 1.1. Hence it follows from (f) in Lemma 1.2 that
(a, b) = [, b) (a, ] B
for every pair of real numbers a < b. Since every open subset of R is a
countable union of sets of the form (a, b), (a, ], [, b), it follows from
axiom (c) in Definition 1.1 that every open subset of R is an element of B.
Hence it follows from Theorem 1.20 that f is measurable. This shows that (ii)
implies (i). That either of the conditions (iii), (iv), and (v) also implies (i) is
shown by a similar argument which is left as an exercise for the reader. This
proves Theorem 1.21.
Our next goal is to show that sums, products, and limits of measurable
functions are again measurable. The next two results are useful for the proofs
of these fundamental facts.
Theorem 1.22 (Vector Valued Measurable Functions). Let (X, A) be
a measurable space and let f = (f1 , . . . , fn ) : X Rn be a function. Then f
is measurable if and only if fi : X R is measurable for each i.
Proof. For i = 1, . . . , n define the projection i : Rn R by i (x) := xi for
x = (x1 , . . . , xn ) R. Since i is continuous it follows from Theorems 1.19
and 1.20 that if f is measurable so is fi = i f for all i. Conversely, suppose
that fi is measurable for i = 1, . . . , n. Let ai < bi for i = 1, . . . , n and define
Q(a, b) := {x Rn | ai < xi < bi i} = (a1 , b1 ) (an , bn ).
Then n
\
f 1
(Q(a, b)) = fi1 ((ai , bi )) A
i=1
by property (f) in Lemma 1.2. Now every open subset of Rn can be expressed
as a countable union of sets of the form Q(a, b). (Prove this!) Hence it follows
from axiom (c) in Definition 1.1 that f 1 (U ) A for every open set U Rn
and hence f is measurable. This proves Theorem 1.22.
Lemma 1.23. Let (X, A) be a measurable space and let u, v : X R

be measurable functions. If : R2 R is continuous then the function
h : X R, defined by h(x) := (u(x), v(x)) for x X, is measurable.
Proof. The map f := (u, v) : X R2 is measurable (with respect to the
Borel -algebra on R2 ) by Theorem 1.22 and the map : R2 R is Borel
measurable by Theorem 1.20. Hence the composition h = f : X R is
measurable by Theorem 1.19. This proves Lemma 1.23.
Theorem 1.24 (Properties of Measurable Functions).
Let (X, A) be a measurable space.
(i) If f, g : X R are measurable functions then so are the functions
f + g, f g, max{f, g}, min{f, g}, |f |.
(ii) Let fk : X R, k = 1, 2, 3, . . . , be a sequence of measurable functions.
Then the following functions from X to R are measurable:
inf fk , sup fk , lim sup fk , lim inf fk .
k k k k
Proof. We prove (i). The functions : R2 R defined by (s, t) := s + t,

(s, t) := st, (s, t) := max{s, t}, (s, t) := min{s, t}, or (s, t) := |s| are all
continuous. Hence assertion (i) follows from Lemma 1.23.
We prove (ii). Define g := supk fk : X R and let a R. Then the set

1

g ((a, ]) = x X sup fk (x) > a
k
= {x X | k N such that fk (x) > a}
[ [
= {x X | fk (x) > a} = fk1 ((a, ])
kN kN
is measurable. Hence it follows from Theorem 1.21 that g is measurable.

It also follows from part (i) (already proved) that fk is measurable, hence
so is supk (fk ) by what we have just proved, and hence so is the function
inf k fk = supk (fk ). With this understood, it follows that the functions
lim sup fk = inf sup fk , lim inf fk = sup inf fk
k `N k` k `N k`
are also measurable. This proves Theorem 1.24.

It follows from Theorem 1.24 that the pointwise limit of a sequence of
measurable functions, if it exists, is again measurable. This is in sharp con-
trast to Riemann integrable functions.
Step Functions
We close this section with a brief discussion of measurable step functions.
Such functions will play a central role throughout this book. In particular,
they are used in the definition of the Lebesgue integral.
Definition 1.25 (Step Function). Let X be a set. A function s : X R
is called a step function (or simple function) if it takes on only finitely
many values, i.e. the image s(X) is a finite subset of R.
Let s : X R be a step function, write s(X) = {1 , . . . , ` } with i 6= j
for i 6= j, and define Ai := s1 (i ) = {x X | s(x) = i } for i = 1, . . . , `.
Then the sets A1 , . . . , A` form a partition of X, i.e.
`
[
X= Ai , Ai Aj = for i 6= j. (1.6)
i=1
(See Example 1.5.) Moreover,

`
X
s= i A i , (1.7)
i=1
where Ai : X R is the characteristic function of the set Ai for i = 1, . . . , `

(see equation (1.4)). In this situation s is measurable if and only if the set
Ai X is measurable for each i. For later reference we prove the following.
Theorem 1.26 (Approximation). Let (X, A) be a measurable space and
let f : X [0, ] be a function. Then f is measurable if and only if there
exists a sequence of measurable step functions sn : X [0, ) such that
0 s1 (x) s2 (x) f (x), f (x) = lim sn (x) for all x X.
n
Proof. If f can be approximated by a sequence of measurable step func-

tions then f is measurable by Theorem 1.24. Conversely, suppose that f is
measurable. For n N define n : [0, ] R by
n
k2 , if k2n t < (k + 1)2n , k = 0, 1, . . . , n2n 1,
n (t) := (1.8)
n, if t n.
These functions are Borel measurable and satisfy n (0) = 0 and n () = n
for all n as well as t 2n n (t) n+1 (t) t whenever n t > 0. Thus
lim n (t) = t for all t [0, ].
n
Hence the functions sn := n f satisfy the requirements of the theorem.

1.3. INTEGRATION OF NONNEGATIVE FUNCTIONS 17
1.3 Integration of Nonnegative Functions

Our next goal is to define the integral of a measurable step function and
then the integral of a general nonnegative measurable function via approxi-
mation. This requires the notion of volume or measure of a measurable set.
The definitions of measure and integral will require some arithmetic on the
space [0, ]. Addition to and multiplication by are defined by

, if a 6= 0,
a + := + a := , a := a :=
0, if a = 0.
With this convention addition and multiplication are commutative, associa-
tive, and distributive. Moreover, if ai and bi are nondecreasing sequences
in [0, ] then the limits a := limi ai and b := limi bi exists in [0, ]
and satisfy the familiar rules a + b = limi (ai + bi ) and ab = limi (ai bi ).
These rules must be treated with caution. The product rule does not hold
when the sequences are not nondecreasing. For example ai := i converges
to a = , bi := 1/i converges to b = 0, but ai bi = 1 does not converge to
ab = 0. (Exercise: Show that the sum of two convergent sequences in [0, ]
always converges to the sum of the limits.) Also, for all a, b, c [0, ],
a + b = a + c, a< = b = c,
ab = ac, 0<a< = b = c.
Neither of these assertions extend to the case a = .
Definition 1.27 (Measure). Let (X, A) be a measurable space. A measure
on (X, A) is a function
: A [0, ]
satisfying the following axioms.
(a) is -additive, i.e. if Ai A, i = 1, 2, 3, . . . , is a sequence of pairwise
disjoint measurable sets then

!
[ X
Ai = (Ai ).
i=1 i=1
(b) There exists a measurable set A A such that (A) < .

A measure space is a triple (X, A, ) consisting of a set X, a -algebra
A 2X , and a measure : A [0, ].
The basic properties of measures are summarized in the next theorem.

Theorem 1.28 (Properties of Measures).
Let (X, A, ) be a measure space. Then the following holds.
(i) () = 0.
(ii) If n N and A1 , . . . , An A such that Ai Aj = for i 6= j then
(A1 An ) = (A1 ) + + (An ).
(iii) If A, B A such that A B then (A) (B).
(iv) Let Ai A be a sequence such that Ai Ai+1 for all i. Then

!
[
Ai = lim (Ai ).
i
i=1
(v) Let Ai A be a sequence such that Ai Ai+1 for all i. Then

!
\
(A1 ) < = Ai = lim (Ai ).
i
i=1
Proof. We prove (i). Choose A1 A such that (A1 ) < and define Ai :=
for i > 1. Then it follows from -additivity that
X
(A1 ) = (A1 ) + ()
i>1
and hence () = 0. This proves part (i).

Part (ii) follows from (i) and -additivity by choosing Ai := for i > n.
We prove (iii). If A, B A such that A B then B \ A A by
property (g) in Lemma 1.2 and hence (B) = (A) + (B \ A) (A) by
part (ii). This proves part (iii).
We prove (iv). Assume Ai Ai+1 for all i and define B1 := A1 and
Bi := Ai \ Ai1 for i > 1. Then Bi is measurable for all i and, for n N,
n
[
[
[
An = Bi , A := Ai = Bi .
i=1 i=1 i=1
Since Bi Bj = for i 6= j it follows from -additivity that

X n
X
(A) = (Bi ) = lim (Bi ) = lim (An ).
n n
i=1 i=1
Here the last equation follows from part (ii). This proves part (iv).
We prove (v). Assume Ai Ai+1 for all i and define Ci := Ai \ Ai+1 .

Then Ci is measurable for all i and, for n N,

[
\
An = A Ci , A := Ai .
i=n i=1
Since Ci Cj = for i 6= j it follows from -additivity that

X
(An ) = (A) + (Ci )
i=n
P
for all n N. Since (A1 ) < it follows that i=1 (Ci ) < and hence

X
lim (An ) = (A) + lim (Ci ) = (A).
n n
i=n
This proves part (v) and Theorem 1.28.

Exercise 1.29. Let (X, A, ) be a measure S space P and let Ai A be a
sequence of measurable sets. Prove that ( i Ai ) i (Ai ).
Example 1.30. Let (X, A) be a measurable space. The counting measure
: A [0, ] is defined by (A) := #A for A A. As an example, consider
the counting measure : 2N [0, ] on the natural numbers. Then the sets
An := {n, n + 1, } all have infinite measure and their intersection is the
empty set and hence has measure zero. Thus the hypothesis (A1 ) <
cannot be removed in part (v) of Theorem 1.28.
Example 1.31. Let (X, A) be a measurable space and fix an element x0 X.
The Dirac measure at x0 is the measure x0 : A [0, ] defined by

1, if x0 A,
x0 (A) := for A A.
0, if x0
/ A,
Example 1.32. Let X be an uncountable set and let A be the -algebra
of all subsets of X that are either countable or have countable complements
(Example 1.4). Then the function : A [0, 1] defined by (A) := 0 when
A is countable and by (A) := 1 when Ac is countable is a measure.
Example 1.33. Let X = iI Ai be a partition and let A 2X be the -
S
P I [0, ] : i 7 i determines
algebra in Example 1.5. Then any function S a
measure : A [0, ] via (AJ ) := jJ j for J I and AJ = jJ Aj .
With these preparations in place we are now ready to introduce the

Lebesgue integral of a nonnegative measurable function
Definition 1.34 (Lebesgue Integral). Let (X, A, ) be a measure space
and let E A be a measurable set.
(i) Let s : X [0, ) be a measurable step function of the form
n
X
s= i Ai (1.9)
i=1
with i [0, ) and Ai R A for i = 1, . . . , n. The (Lebesgue) integral of

s over E is the number E s d [0, ] defined by
Z Xn
s d := i (E Ai ). (1.10)
E i=1
(ii) Let f : X [0, ] be a measurable

R function. The (Lebesgue) integral
of f over E is the number E f d [0, ] defined by
Z Z
f d := sup s d,
E sf E
where the supremum is taken over all measurable step function s : X [0, )
that satisfy s(x) f (x) for all x X.
The reader may verify that the right hand side of (1.10) depends only on s
and not on the choice of i and Ai . The same definition can be used if f is
only defined on the measurable set E X. Then AE := {A A | A E} is
a -algebra on E and RE := |AE is a measure. So (E, AE , E ) is a measure
space and the integral E f dE is well defined. It agrees with the integral of
the extended function on X, defined by f (x) := 0 for x X \ E.
Theorem 1.35 (Basic Properties of the Lebesgue Integral).
Let (X, A, ) be a measure space and let f, g : X [0, ] be measurable
functions and let E A. Then the following holds.
R R
(i) If f g on E then E f d E g d.
R R
(ii) E f d = X f E d.
R
(iii) If f (x) = 0 for all x E then E f d = 0.
R
(iv) If (E) = 0 then E f d = 0.
R R
(v) If A A and E A then E f d A f d.
R R
(vi) If c [0, ) then E cf d = c E f d.
Proof. To prove (i), assume f g on E. If s R: X [0,R) is a measurable

R
step function such that s f then sE g, so E s d = E sE d E g d
by definition of the integral of g.R Now takeRthe supremum over all measurable
step functions s f to obtain E f d E g d. This proves (i).
We prove (ii). It follows from the definitions that
Z Z Z Z Z
f d = sup s d = sup sE d = sup t d = f E d.
E sf E sf X tf E X X
Here the supremum is over all measurable step functions s : X [0, ),

respectively t : X [0, ), that satisfy s f , respectively t f E . The
second equation follows fromR the fact
R that every measurable step function
s : X [0, ) satisfies E s d = X sE d by definition of the integral.
The third equation follows from the fact that a measurable step function
t : X [0, ) satisfies t f E if and only if it has the form t = sE for
some measurable step function s : X [0, ) such that s fR .
R the fact that E f d 0 by
Part (iii) follows from part (i) with g = 0 and
definition. Part (iv) follows from the fact that E s d = 0 for every measur-
able step function s when (E) = 0. Part (v) follows from parts (i) and (ii)
R that f ER f A whenever E A. Part (vi) follows from the
and the fact
fact that E cs d = c E s d for every c [0, ) and every measurable step
function s, by the commutative, associative, and distributive rules for calcu-
lations with numbers in [0, ]. This proves Theorem 1.35.
Notably absent from the statements of Theorem 1.35 is the assertion
that the integral of a sum is the sum of the integrals. This is a fundamental
property that any integral should have. The proof that the integral in Defi-
nition 1.34 indeed satisfies this crucial condition requires some preparation.
The first step is to verify this property for integrals of step functions and the
second step is the Lebesgue Monotone Convergence Theorem 1.37.
Lemma 1.36 (Additivity for Step Functions). Let (X, A, ) be a mea-
sure space and let s, t : X [0, ) be measurable step functions.
(i) For every measurable set E A
Z Z Z
(s + t) d = s d + t d.
E E E
(ii) If E1 , E2 , E3 , . . . is a sequence of pairwise disjoint measurable sets then

Z X Z [
s d = s d, E := Ek .
E k=1 Ek kN
Proof. Write the functions s and t in the form

m
X n
X
s= i A i , t= j Bj
i=1 j=1
where i , j [0, ) and Ai , Bj S A suchSthat Ai Ai0 = for i 6= i0 ,

Bj Bj 0 = for j 6= j 0 , and X = m
i=1 Ai =
n
j=1 Bj . Then
m X
X n
s+t= (i + j )Ai Bj
i=1 j=1
and hence
Z X n
m X
(s + t) d = (i + j )(Ai Bj E)
E i=1 j=1
Xm X n n
X m
X
= i (Ai Bj E) + j (Ai Bj E)
i=1 j=1 j=1 i=1
Xm n
X Z Z
= i (Ai E) + j (Bj E) = s d + t d.
i=1 j=1 E E
To prove (ii), let E1 , ES2 , E3 , . . . be a sequence of pairwise disjoint measurable

sets and define E := k=1 Ek . Then
Z Xm m
X
X
s d = i (E Ai ) = i (Ek Ai )
E i=1 i=1 k=1
m
X n
X
= i lim (Ek Ai )
n
i=1 k=1
m
X Xn
= lim i (Ek Ai )
n
i=1 k=1
n
XX m
= lim i (Ek Ai )
n
k=1 i=1
n Z
X Z
X
= lim s d = s d.
n Ek Ek
k=1 k=1
This proves Lemma 1.36.

Theorem 1.37 (Lebesgue Monotone Convergence Theorem).

Let (X, A, ) be a measure space and let fn : X [0, ] be a sequence of
measurable functions such that
fn (x) fn+1 (x) for all x X and all n N.
Define f : X [0, ] by
f (x) := lim fn (x) for x X.

n
Then f is measurable and

Z Z
lim fn d = f d.
n X X
Proof. By part (i) of Theorem 1.35 we have

Z Z
fn d fn+1 d
X X
for all n N and hence the limit

Z
:= lim fn d (1.11)
n X
exists in [0, ]. Moreover, f = supn fn is a measurable function on X, by

part (ii) of Theorem 1.24, and satisfies fn f for all n N. Thus it follows
from part (i) of Theorem 1.35 that
Z Z
fn d f d for all n N
X X
and hence Z
f d. (1.12)
X
Now fix a measurable step function s : X [0, ) such that s f . Define

s : A [0, ] by
Z
s (E) := s d for E A. (1.13)
E
This function is a measure by part (ii) of Lemma 1.36 (which asserts that s is
-additive) and by part (iv) of Theorem 1.35 (which asserts that s () = 0).
Now fix a constant 0 < c < 1 and define
En := {x X | cs(x) fn (x)} for n N.
Then En A is a measurable set and En En+1 for all n N. Moreover,

[
En = X. (1.14)
n=1
(To spell it out, choose an element x X. If f (x) = , then fn (x)

and hence cs(x) s(x) fn (x) for some n N, which means that x belongs
to one of the sets En . If f (x) < , then fn (x) converges to f (x) > cf (x),
hence fn (x) > cf (x) cs(x) for some n N, and for this n we have x En .)
Since cs fn on En , it follows from parts (i) and (vi) of Theorem 1.35 that
Z Z Z Z
cs (En ) = c s d = cs d fn d fn d .
En En En X
Here the last inequality follows from the definition of in (1.11). Hence

s (En ) for all n N. (1.15)
c
Since s : A [0, ] is a measure, by part (i) of Theorem 1.35, it follows
from equation (1.14) and part (iv) of Theorem 1.28 that
Z

s d = s (X) = lim s (En ) . (1.16)
X n c
Here the last inequality follows

R from (1.15). Since (1.16) holds for every
constant 0 < c < 1, we have X s d for every measurable step function
s : X [0, ) such that s f . Take the supremum over all such s to obtain
Z Z
f d = sup s d .
X sf X
R
Combining this with (1.12) we obtain X f d = and hence the assertion
of Theorem 1.37 follows from the definition of in (1.11).
Theorem 1.38 (-Additivity of the Lebesgue Integral).

Let (X, A, ) be a measure space.
(i) If f, g : X [0, ] are measurable and E A then
Z Z Z
(f + g) d = f d + g d. (1.17)
E E E
(ii) Let fn : X [0, ] be a sequence of measurable functions and define

X
f (x) := fn (x) for x X.
n=1
Then f : X [0, ] is measurable and, for every E A,

Z X Z
f d = fn d. (1.18)
E n=1 E
(iii) If f : X [0, ] is measurable and E1 , E2 , E3 , . . . is a sequence of

pairwise disjoint measurable sets then
Z X Z [
f d = f d, E := Ek . (1.19)
E k=1 Ek kN
Proof. We prove (i). By Theorem 1.26 there exist sequences of measurable

step functions sn , tn : X [0, ) such that sn sn+1 and tn tn+1 for
all n N and f (x) = limn sn (x) and g(x) = limn tn (x) for all x X.
Then sn + tn is a monotonically nondecreasing sequence of measurable step
functions converging pointwise to f + g. Hence
Z Z
(f + g) d = lim (sn + tn ) d
X n X
Z Z
= lim sn d + tn d
n X X
Z Z
= lim sn d + lim tn d
n X n X
Z Z
= f d + g d.
X X
Here the first and last equations follow from Theorem 1.37 and the second
equation follows from part (i) of Lemma 1.36. This proves (i) for E = X. To
prove it in general, replace f, g by f E , gE and use part (ii) of Theorem 1.35.
We prove (ii). Define gn : X [0, ] by gn := nk=1 fk . This is a nonde-

P
creasing sequence of measurable functions, by part (i) of Theorem 1.24, and it
converges pointwise to f by definition. Hence it follows from part (ii) of The-
orem 1.24 that f is measurable and it follows from the Lebesgue Monotone
Convergence Theorem 1.37 that
Z Z
f d = lim gn d
X n X
n
Z X
= lim fk d
n X k=1
n
XZ
= lim fk d
n X
k=1
Z
X
= fn d.
n=1 X
Here the second equation follows from the definition of gn and the third
equation follows from part (i) of the present theorem (already proved). This
proves (ii) for E = X. To prove it in general replace f, fn by f E , fn E and
use part (ii) of Theorem 1.35.
We prove (iii). Let f : X [0, ] be a measurable function and let
Ek A be a sequence of pairwise disjoint measurable sets. Define

[ n
X
E := Ek , fn := f Ek .
k=1 k=1
Then it follows from part (i) of the present theorem (already proved) and
part (ii) of Theorem 1.35 that
Z Z Xn n Z
X n Z
X
fn d = f Ek d = f Ek d = f d.
X X k=1 k=1 X k=1 Ek
Now fn : X [0, ] is a nondecreasing sequence of measurable functions

converging pointwise to f E . Hence it follows from the Lebesgue Monotone
Convergence Theorem 1.37 that
Z Z Z Xn Z X Z
f d = f E d = lim fn d = lim f d = f d.
E X n X n Ek Ek
k=1 k=1
This proves Theorem 1.38.

Exercise 1.39. Let : 2N [0, ] be the counting measure on the natural

numbers. Show that in this case equation (1.18) in part (ii) of Theorem 1.38
is equivalent to the formula

!
!
X X X X
aij = aij (1.20)
i=1 j=1 j=1 i=1
for every map N N [0, ] : (i, j) 7 aij .

The next theorem shows that every measurable function f : X [0, ]
induces another measure f on (X, A).
Theorem 1.40. Let (X, A, ) be a measure space and let f : X [0, ] be
a measurable function. Then the function f : A [0, ], defined by
Z
f (E) := f d for E A (1.21)
E
is a measure and Z Z
g df = f g d (1.22)
E E
for every measurable function g : X [0, ] and every E A.
Proof. f is -additive by part (iii) of Theorem 1.38 and f () = 0 by
part (iv) of Theorem 1.35. Hence f is a measure (see Definition 1.27). Now
let g := A be the characteristic function of a measurable set A A. Then
Z Z Z
A df = f (A) = f d = f A d.
X A X
Here the first equation follows from the definition of the integral for measur-
able step functions in Definition 1.34, the second equation follows from the
definition of f , and the last equation follows from part (ii) of Theorem 1.35.
Thus equation (1.22) (with E = X) holds for characteristic functions of
measurable sets. Taking finite sums and using part (vi) of Theorem 1.35 and
part (i) of Theorem 1.38 we find that (1.22) (with E = X) continues to hold
for all measurable step functions g = s : X [0, ). Now approximate an
arbitrary measurable function g : X [0, ] by a sequence of measurable
step functions via Theorem 1.26 and use the Lebesgue Monotone Conver-
gence Theorem 1.37 to deduce that equation (1.22) holds with E = X for
all measurable functions g : X [0, ]. Now replace g by gE and use
part (ii) of Theorem 1.35 to obtain equation (1.22) in general. This proves
Theorem 1.40.
It is one of the central questions in measure theory under which conditions

a measure : A [0, ] can be expressed in the form f for some mea-
surable function f : X [0, ]. We return to this question in Chapter 5.
The final result in this section is an inequality which will be used in the proof
of the Lebesgue Dominated Convergence Theorem 1.45.
Theorem 1.41 (Lemma of Fatou). Let (X, A, ) be a measure space and
let fn : X [0, ] be a sequence of measurable functions. Then
Z Z
lim inf fn d lim inf fn d.
X n n X
Proof. For n N define gn : X [0, ] by

gn (x) := inf fi (x)
in
for x X. Then gn is measurable, by Theorem 1.24, and

g1 (x) g2 (x) g3 (x) , lim gn (x) = lim inf fn (x) =: f (x)
n n
for all x X. Moreover, gn fi for all i n. By part (i) of Theorem 1.35

this implies Z Z
gn d fi d
X X
for all i n, and hence
Z Z
gn d inf fi d
X in X
for all n N. Thus, by the Lebesgue Monotone Convergence Theorem 1.37,

Z Z Z Z
f d = lim gn d lim inf fi d = lim inf fn d.
X n X n in X n X

Example 1.42. Let (X, A, ) be a measure space and E A be a measur-
able set such that 0 < (E) < (X). Define fn := E when n is even and
fn := 1 E when n is odd. Then lim inf n fn = 0 and so
Z Z
lim inf fn d = 0 < min{(E), (X \ E)} = lim inf fn d.
X n n X
Thus the inequality in Theorem 1.41 can be strict.

1.4. INTEGRATION OF REAL VALUED FUNCTIONS 29
1.4 Integration of Real Valued Functions

The integral of a real valued measurable function is defined as the difference
of the integrals of its positive and negative parts. This definition makes sense
whenever at least one of these numbers is not equal to infinity. It leads natu-
rally to the following concept of integrability and the Lebesgue integral. The
basic properties of the Lebesgue integral are summarized in Theorem 1.44 be-
low. The main result of this section is the Lebesgue Dominated Convergence
Theorem 1.45.
Definition 1.43 (Lebesgue Integrable Functions). Let (X, A, ) be a
measure space. A function f : X R isR called (Lebesgue) integrable
or -integrable if f is measurable and X |f | d < . Denote the set of
-integrable functions by
L1 () := L1 (X, A, ) := {f : X R | f is -integrable} .
The Lebesgue integral of f L1 () over a set E A is the real number

Z Z Z
f d := +
f d f d, (1.23)
E E E
where the functions f : X [0, ) are defined by
f + (x) := max{f (x), 0}, f (x) := max{f (x), 0} (1.24)
The functions f are measurable by Theorem 1.24 and 0 f |f |. Hence

their integrals over E are finite by part (i) of Theorem 1.35.
Theorem 1.44 (Properties of the Lebesgue Integral).
Let (X, A, ) be a measure space. Then the following holds.
(i) The set L1 () is Ra real vector space and, for every E A, the function
L1 () R : f 7 E f d is linear, i.e. if f, g L1 () and c R then
f + g, cf L1 () and
Z Z Z Z Z
(f + g) d = f d + g d, cf d = c f d. (1.25)
E E E E E
(ii) For all f, g L1 () and all E A

Z Z
f g on E = f d g d. (1.26)
E E
(iii) If f L1 () then |f | L1 () and, for all E A,

Z Z

f d |f | d. (1.27)

E E
(iv) If f L1 () and E1 , E2 , E3 , . . . is a sequence of pairwise disjoint mea-

surable sets then
Z X Z [
f d = f d, E := Ek . (1.28)
E k=1 Ek kN
(v) For all E A and all f L1 ()

Z Z
f d = f E d. (1.29)
E X
R
(vi) Let E A and f L1 (). If (E) = 0 or f |E = 0 then E
f d = 0.
1 1
Proof. We prove (i). Let f, g R L () and c R. Then f +g L () because
|f + g| |f | + |g| and hence X |f + g| d < by part (i)R of Theorem 1.38.
Likewise, cf L1 () because |cf | = |c||f | and hence X |cf | d < by
part (vi) of Theorem 1.35. To prove the second equation in (1.25) assume
first that c 0. Then (cf ) = cf and hence
Z Z Z
cf d = +
cf d cf d
E EZ EZ
= c f d c f d
+
ZE E
= c f d.
E
Here the second equation follows from part (vi) of Theorem 1.35. If c < 0
then (cf )+ = (c)f and (cf ) = (c)f + and hence, again using part (iv)
of Theorem 1.35, we obtain
Z Z Z

cf d = (c)f d (c)f + d
E E Z E Z

= (c) f d (c) f + d
Z E E
= c f d.
E
1.4. INTEGRATION OF REAL VALUED FUNCTIONS 31
Now let h := f + g. Then h+ h = f + f + g + g and hence
h+ + f + g = h + f + + g + .
Hence it follows from part (i) of Theorem 1.38 that

Z Z Z Z Z Z
+ +
h d + f d + g d = h d + f d + g + d.
E E E E E E
Hence
Z Z Z
h d = +
h d h d
E ZE ZE Z Z

= +
f d + g d +
f d g d
ZE Z E E E
= f d + g d
E E
and this proves (i).

We prove (ii). Thus assume f = fR+ f g = g + R g on E. Then
f + + g g + + f on E and hence E (f + + g ) d E (g + + f ) d by
part (i) of Theorem 1.35. Now use the additivity of the integral in part (i)
of Theorem 1.38 to obtain
Z Z Z Z

+
f d + g d +
g d + f d.
E E E E
This implies (1.26).

We prove (iii). Since |f | f |f | it follows from (1.25) and (1.26)
that Z Z Z Z
|f | d = (|f |) d f d |f | d
E E E E
and this implies (1.27).
We prove (iv). Equation (1.28) holds for f by part (iii) of Theorem 1.38
and hence holds for f by definition
R of the integral
R in Definition 1.43.
We prove (v). The formula E f d = X f E d in (1.29) follows from
part (ii) of Theorem 1.35 since f E = (f E ) .
R We prove (vi). If f vanishes on E then f also vanish on R E and hence

E
f d = 0 by part (iii) of Theorem 1.35. If (E) = 0 then E f d = 0 by
part (iv) of Theorem 1.35. In either
R case it follows from the definition of the
integral in Definition 1.43 that E f d = 0. This proves Theorem 1.44.
Theorem 1.45 (Lebesgue Dominated Convergence Theorem).

Let (X, A, ) be a measure space, let g : X [0, ) be an integrable function,
and let fn : X R be a sequence of integrable functions satisfying
|fn (x)| g(x) for all x X and n N, (1.30)
and converging pointwise to f : X R, i.e.
f (x) = lim fn (x) for all x X. (1.31)

n
Then f is integrable and, for every E A,

Z Z
f d = lim fn d. (1.32)
E n E
Proof. f is measurable by part (ii) of Theorem 1.24 and |f (x)| g(x) for all
x X by (1.30) and (1.31). Hence it follows from part (i) of Theorem 1.35
that Z Z
|f | d g d <
X X
and so f is integrable. Moreover
|fn f | |fn | + |f | 2g.
Hence it follows from the Lemma of Fatou (Theorem 1.41) that

Z Z
2g d = lim inf 2g |fn f | d
X X n Z
lim inf 2g |fn f | d
n
XZ Z
= lim inf 2g d |fn f | d
n X
Z Z X
= 2g d lim sup |fn f | d.
X n X
Here penultimate step follows from part (i) of Theorem 1.44. This implies
Z
lim sup |fn f | d 0.
n X
1.5. SETS OF MEASURE ZERO 33
Hence Z
lim |fn f | d = 0.
n X
Since Z Z Z Z

fn d f d |fn f | d |fn f | d

E E E X
by part (iii) of Theorem 1.44 it follows that
Z Z

lim fn d
f d = 0,
n E E
which is equivalent to (1.32). This proves Theorem 1.45.
1.5 Sets of Measure Zero

Assume throughout this section that (X, A, ) is a measure space. A set of
measure zero (or null set) is a measurable set N A such that (N ) = 0.
Let P be a name for some property that a point x X may have, or not
have, depending on x. For example, if f : X [0, ] is a measurable
function on X, then P could stand for the condition f (x) > 0 or for the
condition f (x) = 0 or for the condition f (x) = . Or if fn : X R
is a sequence of measurable functions the property P could stand for the
statement the sequence fn (x) converges. In such a situation we say that
P holds almost everywhere if there exists a set N X of measure zero
such that every element x X \ N has the property P. It is not required
that the set of all elements x X that have the property P is measurable,
although that may often be the case.
Example 1.46. Let fn : X R be any sequence of measurable functions.
Then the set
E := {x X | (fn (x))
n=1 is a Cauchy sequence}
\ [ \
x X | |fn (x) fm (x)| < 2k

=
kN n0 N n,mn0
is measurable. If N := X \ E is a set of measure zero then fn converges

almost everywhere to a function f : X R. This function can be chosen
measurable by defining f (x) := limn fn (x) for x E and f (x) := 0 for
x N . This is the pointwise limit of the sequence of measurable functions
gn := fn E and hence is measurable by part (ii) of Theorem 1.24.
The first observation is that every nonnegative function with finite inte-
gral is almost everywhere finite.
R
Lemma 1.47. Let f : X [0, ] be a measurable function. If X f d <
then f < almost everywhere.
R N := {x
Proof. Define R X | f (x) = } and h := N . Then h f and so
(N ) = X h d X f d < by Theorem 1.35. Hence (N ) = 0.
The second observation is that if two integrable, or nonnegative measur-
able, functions agree almost everywhere, then their integrals agree over every
measurable set.
Lemma 1.48. Assume either that f, g : X [0, ] are measurable functions
that agree almost everywhere or that f, g : X R are -integrable functions
that agree almost everywhere. Then
Z Z
f d = g d for all A A. (1.33)
A A
Proof. Fix a measurable set A A and define N := {x X | f (x) 6= g(x)}.

Then N is measurable and (N ) = 0 by assumption. Hence (A N ) = 0
by part (iii) of Theorem 1.28. This implies
Z Z Z Z Z
f d = f d + f d = f d = f A\N d.
A A\N AN A\N X
Here the first equation follows from part (iii) of Theorem 1.38 in the non-
negative case and from part (iv) of Theorem 1.44 in the integrable case.
The second equation follows from part (iv) of Theorem 1.35 in the non-
negative case and from part (vi) of Theorem 1.44 in the integrable case.
The third equation follows from part (ii) of Theorem 1.35 in the nonnega-
tive case and from part (v) of Theorem 1.44 in the integrable case. Since
f A\N = gA\N it follows that the integrals of f and g over A agree. This
proves Lemma 1.48.
The converse of Lemma 1.48 fails for nonnegative measurable functions.
For example, if X is a singleton and (X) = then the integrals of any two
positive functions agree over every measurable set. However, the converse of
Lemma 1.48 does hold for integrable functions. Since the difference of two
integrable functions is again integrable, it suffices to assume g = 0, and in
this case the converse also holds for nonnegative measurable functions. This
is the content of the next lemma.
Lemma 1.49. Assume either that f : X [0, ] is measurable or that

f : X R is -integrable. Then the following are equivalent.
(i) f = 0 almost everywhere.
R
(ii) A f d = 0 for all A A.
R
(iii) X |f | d = 0.
Proof. That (i) implies (ii) is the content of Lemma 1.48. That (ii) im-
plies (iii) is obvious in the nonnegative case. In the integrable case define
A+ := {x X | f (x) 0} , A := {x X | f (x) < 0} .
Then f + = f A+ and f = f A by (1.24). Hence
Z Z Z Z Z
+
|f | d = f d + f d = f d f d = 0
X X X A+ A
by Theorem 1.44 and (ii).

R implies (i). Let f : X [0, ] be a
It remains to prove that (iii)
measurable function such that X f = 0 and define the measurable sets
An := x X | f (x) > 2n

for n N.
Then Z Z
n n
2 (An ) = 2 An d f d = 0
X X
for all n N by Theorem 1.35. Hence (An ) = 0 for all n N and so

[
N := {x X | f (x) 6= 0} = An
n=1
is a set of measure zero. In the integrable case apply this argument to the
function |f | : X [0, ). This proves Lemma 1.49.
Lemma 1.50. Let f L1 (). Then
Z Z

f d =
|f | d (1.34)
X X
if and only if f = |f | almost everywhere or f = |f | almost everywhere.

R R R R
Proof. AssumeR (1.34). Then X f d = X |f | d or X f d = X |f | d. In
the first case X (|f | f ) d = 0Rand so |f | f = 0 almost everywhere by
Lemma 1.49. In the second case X (|f | + f ) d = 0 and so |f | + f = 0 almost
everywhere. This proves Lemma 1.50.
Definition 1.51 (The Banach Space L1 ()). Define an equivalence rela-

tion on the real vector space of all measurable function from X to R by
def the set {x X | f (x) 6= g(x)}
f g (1.35)
has measure zero.
Thus two functions are equivalent iff they agree almost everywhere. (Verify
that this is an equivalence relation!) By Lemma 1.48 the subspace L1 () is
invariant under this equivalence relation, i.e. if f, g : X R are measurable,

f L1 (), and f g then g L1 (). Moreover, the set {f L1 () | f 0}
is a linear subspace of L1 () and hence the quotient space

L1 () := L1 ()/
is again a real vector space. It is the space of all equivalence classes in L1 ()
under the equivalence relation (1.35). Thus an element of L1 () is not a
function on X but a set of functions on X. By Lemma 1.48 the map
Z
1
L () R : f 7 |f | d =: kf kL1
X
takes on the same value on all the elements in a given equivalence class and
so descends to the quotient space L1 (). By Lemma 1.49 it defines a norm
on L1 () and Theorem 1.53 below shows that L1 () is a Banach space with
this norm (i.e. a complete normed vector space).
Theorem 1.52 (Convergent Series of Integrable Functions).
Let (X, A, ) be a measure space and let fn : X R be a sequence of
-integrable functions such that
X Z
|fn | d < . (1.36)
n=1 X
Then there is a set N of measure zero and a function f L1 () such that

X
X
|fn (x)| < and f (x) = fn (x) for all x X \ N, (1.37)
n=1 n=1
Z Z
X
f d = fn d for all A A, (1.38)
A n=1 A
Z n
X
lim f fk d = 0. (1.39)

n X k=1
Proof. Define

X
(x) := |fk (x)|
k=1
for x X. This function is measurable by part (ii) of Theorem 1.24. More-
over, it follows from the Lebesgue Monotone Convergence Theorem 1.37 and
from part (i) of Theorem 1.38 that
Z Z X n n Z
X Z
X
d = lim |fk | d = lim |fk | d = |fk | d < .
X n X k=1 n X X
k=1 k=1
HenceP the set N := {x X | (x) = } has measure zero by Lemma 1.47

and k=1 |fk (x)| < for all x X \ N . Define the function f : X R by
f (x) := 0 for x N and by

X
f (x) := fk (x) for x X \ N.
k=1
Then f satisfies (1.37). Define the functions g : X R and gn : X R by

n
X
g := X\N , gn := fk X\N for n N.
k=1
These
R functions
R are measurable by part (i) of Theorem 1.24. Moreover,
X
g d = X
d < by Lemma 1.48. Since |gn (x)| g(x) for all n N
and gn converges pointwise to f it follows from the Lebesgue Dominated
Convergence Theorem 1.45 that f L1 () and, for all A A,
Z Z Z X n X Z
f d = lim gn d = lim fk d = fn d.
A n A n A k=1 A
n=1
Here the second step follows from Lemma 1.48 because gn = nk=1 fk almost
P
everywhere. The last step follows by interchanging sum and integral, using
part (i) of
PTheorem 1.44. This proves (1.38). To prove equation (1.39) note
n
that f k=1 fk = f gn almost everywhere, that f (x) gn (x) converges to
zero for all x X, and that |f gn | |f |+g where |f |+g is integrable. Hence,
by Lemma 1.48 and the Lebesgue Dominated Convergence Theorem 1.45
Z X n Z
lim f fk d = lim |f gn | d = 0,

n X n X
k=1
This proves (1.39) and Theorem 1.52.

Theorem 1.53 (Completeness of L1 ). Let (X, A, ) be a measure space

and let fn L1 () be a sequence of integrable functions. Assume fn is a
Cauchy sequence with respect to the L1 -norm, i.e. for every > 0 there is an
n0 N such that, for all m, n N,
Z
n, m n0 = |fn fm | d < . (1.40)
X
Then there exists a function f L1 () such that

Z
lim |fn f | d = 0. (1.41)
n X
Moreover, there is a subsequence fni that converges almost everywhere to f .

Proof. By assumption there is a sequence ni N such that
Z
|fni+1 fni | d < 2i , ni < ni+1 , for all i N.
X
Then the sequence gi := fni+1 fni L1 () satisfies (1.36). Hence, by

Theorem 1.52, there exists a function g L1 () such that

X
X
g= gi = fni+1 fni
i=1 i=1
almost everywhere and

Z Xk1 Z
0 = lim gi g d = lim |fnk fn1 g| d. (1.42)

k X i=1 k X
Define
f := fn1 + g.
Pi1
Then fni = fn1 + j=1 gj converges almost everywhere to f . We prove (1.41).
R
Let > 0. By (1.42) there is an ` N such that X |fnkR f | d < /2 for all
k `. By (1.40) the integer ` can be chosen such that X |fn fm | d < /2
for all n, m n` . Then
Z Z Z
|fn f | d |fn fn` | d + |fn` f | d <
X X X
for all n n` . This proves (1.41) and Theorem 1.53.

1.6. COMPLETION OF A MEASURE SPACE 39
1.6 Completion of a Measure Space

The discussion in Section 1.5 shows that sets of measure zero are negligible
in the sense that the integral of a measurable function remains the same if
the function is modified on a set of measure zero. Thus also subsets of sets
of measure zero can be considered negligible. However such subsets need not
be elements of our -algebra A. It is sometimes convenient to form a new
-algebra by including all subsets of sets of measure zero. This leads to the
notion of a completion of a measure space (X, A, ).
Definition 1.54. A measure space (X, A, ) is called complete if
N A, (N ) = 0, EN = E A.
Theorem 1.55. Let (X, A, ) be a measure space and define

there exist measurable sets A, B A such that

A := E X .
A E B and (B \ A) = 0
Then the following holds.

(i) A is a -algebra and A A .
(ii) There exists a unique measure : A [0, ] such that
|A = .
(iii) The triple (X, A , ) is a complete measure space. It is called the

completion of (X, A, ).
(iv) If f : X R is -integrable then f is -integrable and, for E A,
Z Z

f d = f d (1.43)
E E
This continues to hold for all A-measurable functions f : X [0, ].

(v) If f : X R is A -measurable then there exists an A-measurable
function f : X R such that the set
N := {x X | f (x) 6= f (x)} A
has measure zero, i.e. (N ) = 0.

Proof. We prove (i). First X A because A A . Second, let E A

and choose A, B A such that A E B and (B \ A) = 0. Then
B c E c Ac and Ac \ B c = Ac B = B \ A. Hence (Ac \ B c ) = 0 and
so E c A . Third, let Ei A for i N and choose Ai , Bi A such that
Ai Ei Bi and (Bi \ Ai ) = 0. Define
[ [ [
A := Ai , E := Ei , B := Bi .
i i i
S S
Then A E B and B \ A = \ A) i (Bi \ Ai ). Hence
i (Bi
X
(B \ A) (Bi \ Ai ) = 0
i
and this implies E A . Thus we have proved (i).

We prove (ii). For E A define
A, B A,
(E) := (A) where A E B, (1.44)
(B \ A) = 0.
This is the only possibility for defining a measure : A [0, ] that

agrees with on A because (A) = (B) whenever A, B A such that
A B and (B \ A) = 0. To prove that is well defined let E A and
A, B A as in (1.44). If A0 , B 0 A is another pair such that A0 E B 0
and (B 0 \ A0 ) = 0, then A \ A0 E \ A0 B 0 \ A0 and hence (A \ A0 ) = 0.
This implies (A) = (A A0 ) = (A0 ), where the last equation follows
by interchanging the roles of the pairs (A, B) and (A0 , B 0 ). Thus the map
: A [0, ] in (1.44) is well defined.
We prove that is a measure. Let Ei A be a sequence of pairwise
disjoint sets and choose sequences Ai , Bi A such that Ai Ei Bi for
all i. Then
S the Ai are pairwise
S disjoint and (Ei ) = (Ai ) for all i. Moreover
A := i Ai A, B := i Bi A, A E B, and (B P \ A) = 0 as P we have
seen in the proof of part (i). Hence (E) = (A) = i (Ai ) = i (Ei ).
This proves (ii).
We prove (iii). Let E A such that (E) = 0 and let E 0 E. Choose
A, B A such that A E B and (B \ A) = 0. Then (A) = (E) = 0
and hence (B) = (A) + (B \ A) = 0. Since E 0 E B, this implies that
E 0 A (by choosing B 0 := B and A0 := ). This shows that (X, A , ) is
a complete measure space.
1.6. COMPLETION OF A MEASURE SPACE 41
We prove (iv). Assume f : X [0, ] is A-measurable. By Theo-

rem 1.26 there exists a sequence of A-measurable step functions sn : X R
such that 0 s1 s2 R f and
R f (x) = limn sn (x) for all x X.
Since |A = we have X sn d = X sn d for all n and hence it follows
from the Lebesgue Monotone Convergence Theorem 1.37 for both and
that Z Z Z Z

f d = lim sn d = lim sn d = f d .
X n X n X X
This proves (1.43) for E = X and all A-measurable functions f : X [0, ].
To prove it for all E replace f by f E and use part (ii) of Theorem 1.35. This
proves equation (1.43) for all A-measurable functions f : X [0, ]. That
it continues to hold for all f L1 () follows directly from Definition 1.43.
This proves (iv).
We prove (v). If f = E for E A , choose A, B A such that
A E B, (B \ A) = 0,
and define f := A . Then
N = {x X | f (x) 6= f (x)} = E \ A B \ A.
Hence (N ) (B \ A) = (B \ A) = 0. This proves (v) for charac-
teristic functions of A -measurable sets. For A -measurable step functions
the assertion follows by multiplication with real numbers and taking finite
sums. Now let f : X [0, ] be an arbitrary A -measurable function.
By Theorem 1.26 there exists a sequence of A -measurable step functions
si : X [0, ) such that si converges pointwise to f . For each i N
choose an A-measurable step function si : X [0, ) and a set Ni A
such that si = si on X \ Ni and (Ni ) = 0. Then there is a sequence of sets
Ni A such that Ni Ni and (Ni ) = 0 for all i. Define f : X [0, ] by

f (x), if x / N, [
f (x) := N := Ni .
0, if x N,
i
Then N A, (N ) = 0, and the sequence of A-measurable functions si X\N

converges pointwise to f as i tends to infinity. Hence f is A-measurable
by part (ii) of Theorem 1.24 and agrees with f on X \ N by definition.
Now let f : X R be A -measurable. Then so are (f ) := max{f , 0}.
Construct f : X [0, ] as above. Then f (x) = 0 whenever f + (x) > 0
and vice versa. Thus f := f + f is well defined, A-measurable, and agrees
with f on the complement of a -null set. This proves Theorem 1.55.
Corollary 1.56. Let (X, A, ) be a measure space and let (X, A , ) be its
completion. Denote the equivalence class of a -integrable function f L1 ()
under the equivalence relation (1.35) in Definition 1.51 by
n o
[f ] := g L1 () {x X | f (x) 6= g(x)} = 0 .

Then the map

L1 () L1 ( ) : [f ] 7 [f ] (1.45)
is a Banach space isometry.
Proof. The map (1.45) is linear and injective by definition. It preserves

the L1 -norm by part (iv) of Theorem 1.55 and is surjective by part (v) of
Theorem 1.55.
As noted in Section 1.5, sets of measure zero can be neglected when

integrating functions. Hence it is sometimes convenient to enlarge the notion
of integrability. It is not even necessary that the function be defined on all
of X, as long as it is defined on the complement of a set of measure zero.
Thus let (X, A, ) be a measure space and call a function f : E R,
defined on a measurable subset E X, measurable if (X \E) = 0 and the
set f 1 (B) E is measurable for every Borel set B R. Call it integrable
if the function on all of X, obtained by setting f |X\E = 0, is integrable.
If (X, A, ) is complete our integrable function f : E R can be ex-
tended in any manner whatsoever to all of X, and the extended function
on X is then integrable in the original sense, regardless of the choice of the
extension. Moreover, its integral over any measurable set A A is unaffected
by the choice of the extension (see Lemma 1.48).
With this extended notion of integrability we see that the Lebesgue Dom-
inated Convergence Theorem 1.45 continues to hold if (1.31) is replaced by
the weaker assumption that fn only converges to f almost everywhere.
That such an extended terminology might be useful can also be seen in
Theorem 1.52, where the series
P
f
n=1 n only converges on the complement of
a set N of measure zero, and the function f can only be naturally defined on
E := X \ N . Our choice in the proof of Theorem 1.52 was to define f |N := 0,
but this choice does not affect any of the statements of the theorem. More-

over, when working with the quotient space L1 () = L1 ()/ we are only
interested in the equivalence class of f under the equivalence relation (1.35)
rather than a specific choice of an element of this equivalence class.
1.7. EXERCISES 43
1.7 Exercises
Exercise 1.57. Let X be an uncountable set and let A 2X be the set of
all subsets A X such either A or Ac is countable. Define

0, if A is countable,
(A) :=
1, if Ac is countable,
for A A. Show that (X, A, ) is a measure space. Describe the measurable

functions and their integrals. (See Examples 1.4 and 1.32.)
Exercise 1.58. Let (X, A, ) be a measure space such that (X) < and
let fn : X [0, ) be a sequence of bounded measurable functions that
converges uniformly to f : X [0, ). Prove that
Z Z
f d = lim fn d. (1.46)
X n X
Find an example of a measure space (X, A, ) with (X) = and a sequence

of bounded measurable functions fn : X [0, ) converging uniformly to f
such that (1.46) does not hold.
Exercise 1.59. (i) Let fn : [0, 1] [1, 1] be a sequence of continuous
functions that converges uniformly to zero. Show that
Z 1
lim fn (x) dx = 0.
n 0
(ii) Let fn : [0, 1] [1, 1] be a sequence of continuous functions such that
lim fn (x) = 0 for all x [0, 1].

n
Prove that Z 1
lim fn (x) dx = 0,
n 0
without using Theorem 1.45. A good reference is Eberlein [3].

(iii) Construct a sequence of continuous functions fn : [0, 1] [1, 1] that
converges pointwise, but not uniformly, to zero.
(iv) RConstruct a sequence of continuous functions fn : [0, 1] [1, 1] such
1
that 0 fn (x) dx = 0 for all n and fn (x) does not converge for any x [0, 1].
Exercise 1.60. Let (X, A, ) be a measure

R space and f : X [0, ] be a
measurable function such that 0 < c := X f d < . Prove that

Z
f
, if < 1,
lim n log 1 + d = c, if = 1, for 0 < < .
n X n
0, if > 1,

Hint: The integrand can be estimated by f when 1.
Exercise 1.61. Let X := N and A := 2N and let : 2N [0, ] be the

counting measure (Example 1.30). Prove that a function f : N R is -
integrable if and only if the sequence (f (n))nN of real numbers is absolutely
summable and that in this case
Z X
f d = f (n).
N n=1
Exercise 1.62. Let (X, A) be a measurable space and let n : A [0, ]

be a sequence of measures. Show that the formula

X
(A) := n (A)
n=1
for A A defines a measure : A [0, ]. Let f : X R be a measurable

function. Show that f is -integrable if and only if
Z
X
|f | dn < .
n=1 X
If f is -integrable prove that

Z Z
X
f d = f dn .
X n=1 X
Exercise 1.63. Let (X, A, ) be a measure space such that (X) < and
let f : X R be a measurable function. Show that f is integrable if and
only if
X
|({x X | |f (x)| > n})| < .
n=1
1.7. EXERCISES 45
Exercise 1.64. Let (X, A, ) be a measure space and let f : X R be a

-integrable function.
(i) Prove that for every > 0 there exists a > 0 such that, for all A A,
Z

(A) < = f d < .

A
Hint: Argue indirectly. See Lemma 5.21.

(ii) Prove that for every > 0 there exists a measurable set A A such
that, for all B A,
Z Z

BA = f d
f d < .

X B
Exercise 1.65. Let (X, A) be a measurable space and define

0, if A = ,
(A) :=
, if A A and A 6= .
Determine the completion (X, A , ) and the space L1 ().
Exercise 1.66. Let (X, A, ) be a measure space such that = x0 is the
Dirac measure at some point x0 X (Example 1.31). Determine the com-
pletion (X, A , ) and the space L1 ().
Exercise 1.67. Let (X, A, ) be a complete measure space. Prove that
(X, A, ) is equal to its own completion.
Exercise 1.68. Let (X, A, ) and (X, A0 , 0 ) be two measure spaces with
A A0 and 0 |A = . Prove that L1 () L1 (0 ) and
Z Z
f d = f d0
X X
for every f L1 (). Hint: Prove the following.

(i) Let f : X [0, ] be A-measurable and define

0, if f (x) ,
f (x) := f (x), if < f (x) 1 ,
1
, if f (x) > 1 .
R R
Then f is A-measurable for every > 0 and lim0 X f d = X f d.
(ii) Let 0 < c < , let f : X [0, R c] be A-measurable, and assume that
({x X | f (x) > 0}) < . Then X f d = X f d0 . (Consider also the
R
function c f .)
Exercise 1.69 (Pushforward of a Measure).

Let (X, A, ) be a measure space, let Y be a set, and let : X Y be a
map. The pushforward of A is the -algebra
A := B Y | 1 (B) A 2Y .

(1.47)
The pushforward of is the function : A [0, ] defined by
( )(B) := (1 (B)), for B A. (1.48)
(i) Prove that (Y, A, ) is a measure space.

(ii) Let (X, A , ) be the completion of (X, A, ) and let (Y, ( A) , ( ) )
be the completion of (Y, A, ). Prove that
( ) (E) = (1 (E)) for all E ( A) A . (1.49)
Deduce that (Y, A, ) is complete whenever (X, A, ) is complete. Find

an example where ( A) ( A .
(iii) Fix a function f : Y [0, ]. Prove that f is A-measurable if and
only if f is A-measurable. If f is A-measurable, prove that
Z Z
f d( ) = (f ) d. (1.50)
Y X
(iv) Determine the pushforward of (X, A, ) under a constant map.
The following extended remark contains a brief introduction to some of

the basic concepts and terminology in probability theory. It will not be used
elsewhere in this book and can be skipped at first reading.
Remark 1.70 (Probability Theory). A probability space is a measure
space (, F, P ) such that P () = 1. The underlying set is called the sam-
ple space, the -algebra F 2 is called the set of events, and the measure
P : F [0, 1] is called a probability measure. Examples of finite sample
spaces are the set = {h, t} for tossing a coin, the set = {1, 2, 3, 4, 5, 6}
for rolling a dice, the set = {00, 0, 1, . . . , 36} for spinning a roulette wheel,
and the set = {2, . . . , 10, j, q, k, a} {, , , } for drawing a card from
a deck. Examples of infinite sample spaces are the set = N {} for
repeatedly tossing a coin until the first tail shows up, a compact interval of
real numbers for random arrival times, and a disc in the plane for throwing
a dart.
1.7. EXERCISES 47
A random variable is an integrable function X : R. Its expecta-

tion E(X) and variance V(X) are defined by
Z Z
E(X) := X dP , V(X) := (X E(X))2 dP = E(X 2 ) E(X)2 .

Given a random variable X : R one is interested in the value of the prob-

ability measure on the set X 1 (B) for a Borel set B R. This value is the
probability of the event that the random variable X takes its value in the set B
and is denoted by P (X B) := P (X 1 (B)) = (X P )(B). Here X P denotes
the pushforward of the probability measure P to the Borel -algebra B 2R
(ExerciseR 1.69). By (1.50) the expectation
R and variance of X are given by
E(X) = R x d(X P )(x) and V(X) = R (x E(X))2 d(X P )(x).
The (cumulative) distribution function of a random variable X is
the function FX : R [0, 1] defined by
FX (x) := P (X x) = P ({ | X() x}) = (X P )((, x]).
It is nondecreasing and right continuous, satisfies
lim FX (x) = 0, lim FX (x) = 1,
x x
and the integral of a continuous function on R with respect to the pushfor-

ward measure X P agrees with the RiemannStieltjes integral (Exercise 6.20)
with respect to FX . Moreover,
FX (x) lim FX (t) = P (X 1 (x))
tx
by Theorem 1.28. Thus FX is continuous at x if and only if P (X 1 (x)) = 0.

This leads to the following notions of convergence. Let X : R be a
random variable. A sequence (Xi )iN of random variables is said to
converge in probability to X if limi P (|Xi X| ) = 0 for all > 0,
converge in distribution to X if FX (x) = limi FXi (x) for every x R
such that FX is continuous at x.
We prove that that convergence almost everywhere implies convergence
in probability. Let > 0 and define Ai := { | |Xi () X()| }.
Let E be the set of all such that the sequence Xi () does not
converge to X(). This set is measurable by Example
T 1.46S and has measure
zero by convergence
S almost everywhere. Moreover, iN ji Aj E and so
limi P ( ji Aj ) = P (E) = 0 by Theorem 1.28. Thus limi P (Ai ) = 0.
We prove that convergence in probability implies convergence in distri-

bution. Let x R such that FX is continuous at x. Let > 0 and choose
> 0 such that FX (x) 2 < FX (x ) FX (x + ) < FX (x) + 2 .
Now choose i0 N such that P (|Xi X| ) < 2 for all i i0 . Then
FX (x ) P (|Xi X| ) FXi (x) FX (x + ) + P (|Xi X| )
and hence FX (x) < FXi (x) < FX (x) + for all i i0 . This shows that
limi FXi (x) = FX (x) as claimed.
A finite collection of random variables X1 , . . . , Xn is called independent
if, for every collection of Borel sets B1 , . . . , Bn R, it satisfies
\ n Y n
1
P Xi1 (Bi ) .

P Xi (Bi ) =
i=1 i=1
In Chapter 7 we shall see that this condition asserts that the pushforward of P
under the map X := (X1 , . . . , Xn ) : Rn agrees with the product of the
measures (Xi ) P . Two foundational theorems in probability theory are the
law of large numbers and the central limit theorem. These are results about
sequences of random variables Xk : R that satisfy the following.
(a) The random variables X1 , . . . , Xn are independent for all n.
(b) The Xk have expectation E(Xk ) = 0.
(c) The Xk are identically distributed, i.e. FXk = FX` for all k and `.
For n N define Sn := X1 + + Xn . Kolmogorovs strong law of large
numbers asserts that, under these assumptions, the sequence Sn /n con-
verges almost everywhere P to zero. (This continues to hold when (c) is re-
placed by the assumption 1
k=1 k2 V(Xk ) < .) If, in addition, V(Xk ) =
2
for all k and some positive real number then the central limit theorem
of LindebergLevy asserts that the sequence Tn := Sn / n converges in dis-
tribution to a so-called standard normal randomR variable with expectation
x 2
zero and variance one, i.e. limn FTn (x) = 12 et /2 dt for all x R.
The hypotheses listed above are quite restrictive and in modern probabil-
ity theory these theorems are often needed in much greater generality. For
proofs, many examples, and comprehensive expositions of probability theory
see Ash [1], Fremlin [4, Chapter 27], Malliavin [13].
An important class of random variables are those where the distribution
functions FX : R [0, 1] are absolutely continuous (Theorem 6.19). This
means that the pushforward measures X P on the Borel -algebra B 2R
admit densities as in Theorem 1.40 with respect to the Lebesgue measure.
The Lebesgue measure is introduced in Chapter 2 and the existence of a
density is the subject of Chapter 5 on the RadonNikod ym Theorem.
Chapter 2
The Lebesgue Measure
This chapter introduces the most important example, namely the Lebesgue
n
measure on Euclidean space. Let n N and denote by B 2R the -algebra
of all Borel sets in Rn , i.e. the smallest -algebra on Rn that contains all open
sets in the standard topology (Definition 1.15). Then
B + x := {y + x | y B} B for all B B and all x Rn ,
because the translation Rn Rn : y 7 y + x is a homeomorphism. A
measure : B [0, ] is called translation invariant if it satisfies
(B + x) = (B) for all B B and all x Rn . (2.1)
Theorem 2.1. There exists a unique measure : B [0, ] that is trans-
lation invariant and satisfies the normalization condition ([0, 1]n ) = 1.
Proof. See page 64.
Definition 2.2. Let (Rn , B, ) be the measure space in Theorem 2.1 and
denote by (Rn , A, m) its completion as in Theorem 1.55. Thus

n there exist Borel sets B0 , B1 B

A := A R (2.2)
such that B0 A B1 and (B1 \ B0 ) = 0
and m(A) := (B0 ) for A A, where B0 , B1 B are chosen such that
B0 A B1 and (B1 \ B0 ) = 0. The elements of A are called Lebesgue
measurable subsets of Rn , the function m : A [0, ] is called the
Lebesgue measure, and the triple (Rn , A, m) is called the Lebesgue mea-
sure space. A function f : Rn R is called Lebesgue measurable if it
is measurable with respect to the Lebesgue -algebra A on Rn (and the Borel
-algebra on the target space R).
49
50 CHAPTER 2. THE LEBESGUE MEASURE
2.1 Outer Measures

In preparation for the proof of Theorem 2.1 we now take up the idea, an-
nounced in the beginning of Chapter 1, of assigning a measure to every subset
of a given set but requiring only subadditivity. Here is the basic definition.
Definition 2.3. Let X be a set. A function : 2X [0, ] is called an
outer measure if it satisfies the following three axioms.
(a) () = 0.
(b) If A B X then (A) (B).
(c) If Ai X for i N then (
S P
i=1 Ai ) i=1 (Ai ).
Let : 2X [0, ] be an outer measure. A subset A X is called -
measurable if it satisfies
(D) = (D A) + (D \ A) (2.3)
for every subset D X.
The inequality (D) (D A) + (D \ A) holds for every outer measure
and any two subsets A, D X by (a) and (c). However, the outer measure
of a disjoint union need not be equal to the sum of the outer measures.
Caratheodorys Theorem 2.4 below asserts that the -measurable sets form
a -algebra A and that the restriction of to A is a measure. Theorem 2.5
(the Caratheodory Criterion) characterises outer measues on metric spaces
such that every Borel set is -measurable.
Theorem 2.4 (Carath eodory). Let X be a set, let : 2X [0, ] be an
outer measure, and define

A := A() := A X A is -measurable (2.4)
Then A is a -algebra, the function
:= |A : A [0, ]
is a measure, and the measure space (X, A, ) is complete.
Proof. The proof has six steps.
Step 1. X A.
For every subset D X, we have
(D X) + (D \ X) = (D) + () = (D)
by condition (a) in Definition 2.3. Hence X A.
2.1. OUTER MEASURES 51
Step 2. If A A then Ac A.
Let A A. Since
D Ac = D \ A, D \ Ac = D A,
it follows from equation (2.3) that (D) = (D Ac ) + (D \ Ac ) for every

subset D X. Hence Ac A.
Step 3. If A, B A then A B A.
Let A, B A. Then, for every subset D X,
(D) = (D A) + (D \ A)
= (D A) + (D Ac )
= (D A) + (D Ac B) + ((D Ac ) \ B)
((D A) (D Ac B)) + (D Ac B c )
= (D (A B)) + (D (A B)c )
= (D (A B)) + (D \ (A B)).
Here the inequality follows from axioms (a) and (c) in Definition 2.3. Using
axioms (a) and (c) again we obtain (D) = (D (A B)) + (D \ (A B))
for every subset D X and hence A B A.
Step 4. Let Ai A for i N such that Ai Aj = for i 6= j. Then

[
X
A := Ai A, (A) = (Ai ).
i=1 i=1
For k N define
Bk := A1 A2 Ak .
Then Bk A for all k N by Step 3. Now let D X. Then, for all k 2,
(D Bk ) = (D Bk Ak ) + ((D Bk ) \ Ak )
= (D Ak ) + (D Bk1 )
and so, by induction on k,

k
X
(D Bk ) = (D Ai ).
i=1
Since Bk A, this implies
(D) = (D Bk ) + (D \ Bk )
k
X
= (D Ai ) + (D \ Bk )
i=1
Xk
(D Ai ) + (D \ A).
i=1
Here the last inequality follows from

S axiom (b) in Definition 2.3. Since this
holds for all k N and D A = i=1 (D Ai ), it follows that

X
(D) (D Ai ) + (D \ A) (D A) + (D \ A) (D).
i=1
Here the last two inequalities follow from axiom (c). Hence

X
(D) = (D Ai ) + (D \ A) = (D A) + (D \ A) (2.5)
i=1
for all D X. This shows that A A. Now take D = A toPobtain D \A =

and D Ai = Ai . Then it follows from (2.5) that (A) =
i=1 (Ai ).
S
Step 5. Let Ai A for i N. Then A := i=1 Ai A.
Define B1 := A1 and Bi := Ai \ (A1 Ai1 ) for i 2. Then Bi Bj =
c c
for i 6= j and
SBi = (A1 Ai1 Ai ) A for all i by Steps 2 and 3.
Hence A = i=1 Bi A by Step 4. This proves Step 5.
Step 6. (X, A, ) is a complete measure space.
It follows from Steps 1, 2, 4, and 5 that (X, A, = |A ) is a measure space.
We prove that it is complete. To see this, let A X and suppose that
A N where N A satisfies (N ) = 0. Then it follows from axiom (b) in
Definition 2.3 that (A) (N ) = (N ) = 0 and therefore (A) = 0. Now
use axioms (a), (b) and (c) to obtain
(D) (D A) + (D \ A) (A) + (D) = (D)
and so (D) = (D A) + (D \ A) for all D X, which shows that A A.

This proves Step 6 and Theorem 2.4.
Theorem 2.5 (Carath eodory Criterion). Let (X, d) be a metric space

and : 2 [0, ] be an outer measure. Let A() 2X be the -algebra
X
given by (2.4) and let B 2X the Borel -algebra of (X, d). Then the
following are equivalent.
(i) B A().
(ii) If A, B X satisfy d(A, B) := inf aA, bB d(a, b) > 0 then
(A B) = (A) + (B).
Proof. We prove that (i) implies (ii). Thus assume that satisfies (i). Let
A, B X such that := d(A, B) > 0. Define
[
U := x X a A such that d(a, x) < = B (a).
aA
Then U is open, A U , and U B = . Hence U B A() by assumption

and hence (A B) = ((A B) U ) + ((A B) \ U ) = (A) + (B). Thus
the outer measure satisfies (ii).
We prove that (ii) implies (i). Thus assume that satisfies (ii). We prove
that every closed set A X is -measurable, i.e. (D) = (D A)+(D \A)
for all D X. Since (D) (D A) + (D \ A), by definition of an outer
measure, it suffices to prove the following.
Claim 1. Fix a closed set A X and a set D X such that (D) < .
Then (D) (D A) + (D \ A).
To see this, replace the set D \ A by the smaller set D \ Uk , where
[
Uk := x X a A such that d(a, x) < 1/k = B1/k (a).
aA
For each k N the set Uk is open and d(x, y) 1/k for all x D A and
all y D \ Uk . Hence
1
d(D A, D \ Uk ) .
k
By (ii) and axiom (b) this implies
(D A) + (D \ Uk ) = ((D A) (D \ Uk )) (D) (2.6)
for every subset D X and every k N. We will prove the following.

Claim 2. limk (D \ Uk ) = (D \ A).

Claim 1 follows directly from Claim 2 and (2.6). To prove Claim 2 note that

\
A= Ui
i=1
because A is closed. (If x Ui for all i N then there exists a sequence

ai A such that d(ai , x) < 1/i and hence x = limi ai A.) This implies

[
[
Uk \ A = (Uk \ Ui ) = Ui \ Ui+1
i=1 i=k
and hence
D \ A = (D \ Uk ) (D (Uk \ A))
[
= (D \ Uk ) (D (Ui \ Ui+1 )).
i=k
Thus
[
D \ A = (D \ Uk ) Ei , Ei := (D Ui ) \ Ui+1 . (2.7)
i=k
P
Claim 3. The outer measures of the Ei satisfy i=1 (Ei ) < .
Claim 3 implies Claim 2. It follows from Claim 3 that the sequence

X
k := (Ei )
i=k
converges to zero. Moreover, it follows from equation (2.7) and axiom (c) in
Definition 2.3 that

X
(D \ A) (D \ Uk ) + (Ei ) = (D \ Uk ) + k .
i=k
Hence it follows from axiom (b) in Definition 2.3 that

(D \ A) k (D \ Uk ) (D \ A)
for every k N. Since k converges to zero, this implies Claim 2. The proof
of Claim 3 relies on the next assertion.
Claim 4. d(Ei , Ej ) > 0 for i j + 2.

Claim 4 implies Claim 3. It follows from Claim 4, axiom (b), and (ii) that
n n
!
X [
(E2i ) = E2i (D)
i=1 i=1
and !
n
X n
[
(E2i1 ) = E2i1 (D)
i=1 i=1
P
for every n N. Hence i=1 (Ei ) 2(D) < and this shows that
Claim 4 implies Claim 3.
Proof of Claim 4. We show that
1
d(Ei , Ej ) for j i + 2.
(i + 1)(i + 2)
To see this, fix indices i, j with j i + 2. Let x Ei and y X such that

1
d(x, y) < .
(i + 1)(i + 2)
Then x
/ Ui+1 because Ei Ui+1 = . (See equation (2.7).) Hence
1
d(a, x) for all a A.
i+1
This implies
d(a, y) d(a, x) d(x, y)

1 1
>
i + 1 (i + 1)(i + 2)
1
=
i+2
1

j
for all a A. Hence y

/ Uj and hence y
/ Ej because Ej Uj . This proves
Claim 4 and Theorem 2.5.
2.2 The Lebesgue Outer Measure

The purpose of this section is to introduce the Lebesgue outer measure
on Rn , construct the Lebesgue measure as the restriction of to the -algebra
of all -measurable subsets of Rn , and prove Theorem 2.1.
Definition 2.6. A closed cuboid in Rn is a set of the form
Q := Q(a, b)
:= [a1 , b1 ] [a2 , b2 ] [an , bn ] (2.8)

= x = (x1 , . . . , xn ) Rn aj xj bj for j = 1, . . . , n

for a1 , . . . , an , b1 , . . . , bn R with aj < bj for all j. The (n-dimensional)

volume of the cuboid Q(a, b) is defined by
n
Y
Vol(Q(a, b)) := Voln (Q(a, b)) := (bj aj ). (2.9)
j=1
The volume of the open cuboid U := int(Q) = ni=1 (ai , bi ) is defined by

Q
Vol(U ) := Vol(Q). The set of all closed cuboids in Rn will be denoted by

a1 , . . . , an , b1 , . . . , bn R,
Qn := Q(a, b) .
aj < bj for j = 1, . . . , n
Definition 2.7. A subset A Rn is called a Jordan null set if, for every
> 0, there exist finitely many closed cuboids Q1 , . . . , Q` Qn such that
`
[ `
X
A Qi , Vol(Qi ) < .
i=1 i=1
Definition 2.8. A subset A Rn is called a Lebesgue null set if, for

every > 0, there is a sequence of closed cuboids Qi Qn , i N, such that

[
X
A Qi , Vol(Qi ) < .
i=1 i=1
Definition 2.9. The Lebesgue outer measure on Rn is the function

n
= n : 2R [0, ] defined by
(

)
X [
(A) := inf Voln (Qi ) Qi Qn , A Qi for A Rn . (2.10)

i=1 i=1
2.2. THE LEBESGUE OUTER MEASURE 57
n
Theorem 2.10 (The Lebesgue Outer Measure). Let : 2R [0, ]
be the function defined by (2.10). Then the following holds.
(i) is an outer measure.
(ii) is translation invariant, i.e. for all A Rn and all x Rn
(A + x) = (A).
(iii) If A, B Rn such that d(A, B) > 0 then (A B) = (A) + (B).

(iv) (int(Q)) = (Q) = Vol(Q) for all Q Qn .
Proof. We prove (i). The empty set is contained in every cuboid Q Qn .

Since there are cuboids with arbitrarily small volume it follows that () = 0.
If A B Rn it follows directly from Definition 2.9 that (A) (B). Now
let Ai Rn for i N, define

[
A := Ai ,
i=1
and fix a constant > 0. Then it follows from Definition 2.9 that, for i N,
there exists a sequence of cuboids Qij Qn , j N, such that

[ X
Ai Qij , Vol(Qij ) < + (Ai ).
j=1 j=1
2i
Hence

[ X X X
A Qij , Vol(Qij ) < + (Ai ) = + (Ai ).
i,jN i,jN i=1
2i i=1
This implies

X
(A) < + (Ai )
i=1
P
for every > 0 and thus (A) i=1 (Ai ). This proves part S
(i).
We prove (ii). If A i=1 Qi with Qi Qn , then A + x
S
i=1 (Qi + x)
for every x Rn and Vol(Qi + x) = Vol(Qi ) by definition of the volume.
Hence part (ii) follows from Definition 2.9.
We prove (iii). Let A, B Rn such that d(A, B) > 0. Choose a sequence

of closed cuboids Qi Qn such that

[
X
AB Qi , Vol(Qi ) < (A B) + .
i=1 i=1
Subdividing each Qi into finitely many smaller cuboids, if necessary, we may

assume without loss of generality that
d(A, B)
diam(Qi ) := sup |x y| < .
x,yQi 2
Here || denotes the Euclidean norm on Rn . Then, for every i N, we have

either Qi A = of Qi B = . This implies
I J = , I := {i N | Qi A 6= }, J := {i N | Qi B 6= }.
Hence
X X
(A) + (B) Vol(Qi ) + Vol(Qi )
iI iJ

X
Vol(Qi )
i=1
< (A B) + .
Thus (A) + (B) < (A B) + for all > 0, so (A) + (B) (A B),
and hence (A) + (B) = (A B), by axioms (a) and (c) in Definition 2.3.
This proves part (iii).
We prove (iv) by an argument due to von Neumann. Fix a closed cuboid
Q = I1 In , Ii = [ai , bi ].
We claim that
Vol(Q) (Q). (2.11)
Equivalently, if Qi Qn , i N, is a sequence of closed cuboids then

[
X
Q Qi = Vol(Q) Vol(Qi ). (2.12)
i=1 i=1
For a closed interval I = [a, b] R with a < b define
|I| := b a.
Then
|I| 1 #(I Z) |I| + 1.
Hence
N |I| 1 #(N I Z) N |I| + 1
and thus
1 1 1 1
|I| # I Z |I| +
N N N N
for every integer N N. Take the limit N to obtain

1 1
|I| = lim # I Z .
N N N
Thus
n
Y 1 1
Vol(Q) = lim # Ij Z
N
j=1
N N
(2.13)
1 1 n
= lim n # Q Z .
N N N
Now Ssuppose Qi Qn , i N, is a sequence of closed cuboids such that

Q i=1 Qi . Fix a constant > 0 and choose a sequence of open cuboids
Ui Rn such that

Qi Ui , Vol(Ui ) < Vol(Qi ) + .
2i
Since Q is compact, and the Ui form an open cover of Q, there exists a
constant k N such that
[k
Q Ui .
i=1
This implies
Xk Xk
1 1 n 1 1 n 1 1 n
# Q Z # Ui Z # Ui Z .
Nn N i=1
Nn N i=1
Nn N
Take the limit N and use equation (2.13) to obtain

k
X
Vol(Q) Vol(Ui )
i=1
X
Vol(Ui )
i=1

X
+ Vol(Qi )
i=1
2i

X
= + Vol(Qi ).
i=1
Since > 0 can be chosen arbitrarily small, this proves (2.12) and (2.11).
Thus we have proved that (Q) Vol(Q) (Q) and so (Q) = Vol(Q).
To prove that (int(Q)) = Vol(Q), fix a constant > 0 and choose a closed
cuboid P Qn such that
P int(Q), Vol(Q) < Vol(P ).
Then
Vol(Q) < Vol(P ) = (P ) (int(Q)).
Thus Vol(Q) < (int(Q)) for all > 0. Hence, by axiom (b),
Vol(Q) (int(Q)) (Q) = Vol(Q),
and hence (int(Q)) = Vol(Q). This proves part (iv) and Theorem 2.10.
n
Definition 2.11. Let : 2R [0, ] be the Lebesgue outer measure. A
subset A Rn is called Lebesgue measurable if A is -measurable, i.e.
(D) = (D A) + (D \ A) for all D Rn .
The set of all Lebesgue measurable subsets of Rn will be denoted by

A := A Rn A is Lebesgue measurable .

The function
m := |A : A [0, ]
is called the Lebesgue measure on Rn . A function f : Rn R is called
Lebesgue measurable if it is measurable with respect to the Lebesgue -
algebra A on Rn (and the Borel -algebra on the target space R).
Corollary 2.12. (i) (Rn , A, m) is a complete measure space.

(ii) m is translation invariant, i.e. if A A and x Rn then A + x A
and m(A + x) = m(A).
(iii) Every Borel set in Rn is Lebesgue measurable.
(iv) If Q Qn then Q, int(Q) A and m(int(Q)) = m(Q) = Vol(Q) .
Proof. Assertion (i) follows from Theorem 2.4 and part (i) of Theorem 2.10.
Assertion (ii) follows from the definitions and part (ii) of Theorem 2.10.
Assertion (iii) follows from Theorem 2.5 and part (iii) of Theorem 2.10. As-
sertion (iv) follows from (iii) and part (iv) of Theorem 2.10.
The restriction of the measure m in Corollary 2.12 to the Borel -algebra
of Rn satisfies the requirements of Theorem 2.1 (translation invariance and
normalization) and hence settles the existence problem. The uniqueness
proof relies on certain regularity properties of the measure m which are
established in the next theorem along with continuity from below for the
Lebesgue outer measure . Theorem 2.14 shows that m is the completion of
its restriction to the Borel -algebra of Rn and, with that at hand, we can
then prove uniqueness in Theorem 2.1.
Theorem 2.13 (Regularity of the Lebesgue Outer Measure).

n
The Lebesgue outer measure : 2R [0, ] satisfies the following.
(i) For every subset A Rn

(A) = inf (U ) A U Rn and U is open .

(ii) If A Rn is Lebesgue measurable then

(A) = sup (K) K A and K is compact .
n
S of subsets of R such that Ai Ai+1 for all i N then
(iii) If Ai is a sequence
their union A := i=1 Ai has Lebesgue outer measure (A) = limi (Ai ).
Proof. We prove (i). Fix a subset A Rn and a constant > 0. The

assertion is obvious when (A) = . Hence assume (A) < and choose a
sequence of closed cuboids Qi Qn such that

[ X
A Qi , Vol(Qi ) < (A) + .
i=1 i=1
2
Now choose a sequence of open cuboids Ui Rn such that

Qi Ui , Vol(Ui ) < Vol(Qi ) + i+1 .
2
S n
Then U := i=1 Ui is an open subset of R containing A and

X X X
(U ) (Ui ) = Vol(Ui ) < Vol(Qi ) + i+1 < (A) + .
i=1 i=1 i=1
2
This proves part (i).
To prove (ii), assume first that A Rn is Lebesgue measurable and
bounded. Choose r > 0 so large that A Br := {x Rn | |x| < r} . Fix a
constant > 0. By (i) there exists an open set U Rn such that B r \ A U
and (U ) (B r \ A) + . Hence K := B r \ U is a compact subset of A and
(K) = (B r ) (U ) (B r ) (B r \ A) = (A) .
Here the first equation uses the fact that K and U are disjoint Lebesgue
measurable sets with union B r and the last equation uses the fact that A and
B r \ A are disjoint Lebesgue measurable sets with union B r . This proves (ii)
for bounded Lebesgue measurable sets. If A A is unbounded then
(A) = sup (A B r )
r>0

= sup sup (K) K (A B r ) and K is compact
r>0

= sup (K) K A and K is compact .
This proves part (ii).
We prove (iii). If (Ai ) = for some i then the assertion is obvious.
Hence assume (Ai ) < for all i and fix a constant > 0. By part (i) there
is a sequence of open sets Ui Rn such that (Ui ) < (Ai ) + 2i for all i.
Since Ai Ui Ui+1 this implies
(Ui+1 \ Ui ) = (Ui+1 ) (Ui Ui+1 ) < (Ai+1 ) (Ai ) + 2i1
for all i N. This implies
k
! k1
[ X
Ui = (U1 ) + (Ui+1 \ Ui ) < (Ak ) + .
i=1 i=1
Take the limit k to obtain (

S
i=1 Ui ) limk (Ak ) + . Thus
(A) limk (Ak ) + for all > 0 and so (A) limk (Ak ). The
converse inequality is obvious. This proves part (iii) and Theorem 2.13.
Theorem 2.14 (The Lebesgue Measure as a Completion).

n
Let : 2R [0, ] be the Lebesgue outer measure in Definition 2.9, let
m = |A : A [0, ] be the Lebesgue measure, let B A be the Borel
-algebra of Rn , and define := |B : B [0, ]. Then (Rn , A, m) is the
completion of (Rn , B, ).
Proof. Let (Rn , B , ) denote the completion of (Rn , B, ).
Claim. Let A Rn . Then the following are equivalent.
(I) A A, i.e. (D) = (D A) + (D \ A) for all D Rn .
(II) A B , i.e. there exist Borel measurable sets B0 , B1 B such that
B0 A B1 and (B1 \ B0 ) = 0.
If the set A satisfies both (I) and (II) then
(A) (B1 ) = (B0 ) + (B1 \ B0 ) = (B0 ) (A)
and hence m(A) = (A) = (B0 ) = (A). This shows that A = B and
m = . Thus it remains to prove the claim. Fix a subset A Rn .
We prove that (II) implies (I). Thus assume that A B and choose Borel
measurable sets B0 , B1 B such that
B0 A B1 , (B1 \ B0 ) = 0.
Then (A \ B0 ) (B1 \ B0 ) = 0 and hence (A \ B0 ) = 0. Since is an
outer measure, by part (i) of Theorem 2.10, it follows from Theorem 2.4 that
A \ B0 A and hence A = B0 (A \ B0 ) A.
We prove that (I) implies (II). Thus assume that A A. Suppose first
that (A) < . By Theorem 2.13 there exists a sequence of compact sets
Ki Rn and a sequence of open sets Ui Rn such that
1 1
Ki A Ui , (A) (Ki ) (Ui ) (A) + .
i i
Define
[ \
B0 := Ki , B1 := Ui .
i=1 i=1
These are Borel sets satisfying B0 A B1 and
1 1
(A) (Ki ) (B0 ) (B1 ) (Ui ) (A) + .
i i
Take the limit i to obtain (A) (B0 ) (B1 ) (A), hence
(B0 ) = (B1 ) = (A) < , and hence (B1 \ B0 ) = (B1 ) (B0 ) = 0.
This shows that A B for every A A with (A) < .
Now suppose that our set A A satisfies (A) = and define

Ak := {x A | |xi | k for i = 1, . . . , n} for k N.
Then Ak A and (Ak ) (2k)n for all k. Hence Ak B for all k and so
sets Bk , Bk0 B such
there exist sequences of Borel S that Bk Ak Bk0 and

(Bk0 \ Bk ) = 0. Define B := k=1 Bk and B 0 := k=1 Bk0 . Then B, B 0 B,
S
B A B 0 , and

X
X
0
(B \ B) (Bk0 \ B) (Bk0 \ Bk ) = 0.
k=1 k=1
This shows that A B for every A A. Thus we have proved that (I)
implies (II) and this completes the proof of Theorem 2.14.
Proof of Theorem 2.1. The existence of a translation invariant normalized
Borel measure on Rn follows from Corollary 2.12. We prove uniqueness.
Thus assume that 0 : B [0, ] is a translation invariant measure such
that 0 ([0, 1]n ) = 1. Define := 0 ([0, 1)n ). Then 0 1. We prove in
five steps that = 1 and 0 = .
Step 1. For x = (x1 , . . . , xn ) and k N0 := N {0} define
R(x, k) := [x1 , x1 + 2k ) [xn , xn + 2k ).
Then 0 (R(x, k)) = 2nk = (R(x, k)).
Fix an integer k N0 . Since R(x, k) = R(0, k) + x for every x Rn it follows
from the translation invariance of 0 that there is a constant ck 0 such that
0 (R(x, k)) = ck for all x Rn .
Since R(x, 0) can be expressed as the disjoint union
[
R(x, 0) = R(x + 2k `, k),
`Zn , 0`j 2k 1
this implies
X
= 0 (R(x, 0)) = 0 (R(x + 2k `, k)) = 2nk ck .
`Zn , 0`j 2k 1
Hence ck = 2nk = (R(x, k)). Here the last equation folows from the fact
that (0, 2k )n R(0, k) [0, 2k ]n and hence (R(x, k)) = (R(0, k)) = 2k
by part (iv) of Corollary 2.12. This proves Step 1.
Step 2. 0 (U ) = (U ) for every open set U Rn .

Let U Rn be open. We prove that U can be expressed as a countable union
of sets Ri = R(xi , ki ) as in Step 1. To see this, define
R0 := R(x, 0) x Zn , R(x, 0) U ,

x 21 Zn , R(x, 1) U,

R1 := R(x, 1) ,
R(x, 1) 6 R R R0

x 2k Zn , R(x, k) U,

Rk := R(x, k)
R(x, k) 6 R R R0 R1 Rk1

S
for k 2 and denote R := k=0 Rk . Then U can be expressed as the disjoint
union U = RR R and (R) = (R) for all R R by Step 1. Hence
0
S
X X
0 (U ) = 0 (R) = (R) = (U )
RR RR
and this proves Step 2.

Step 3. 0 (K) = (K) for every compact set K Rn .
Let K Rn be compact. Choose r > 0 so large that K U := (r, r)n .
Then U and U \ K are open. Hence, by Step 2,
0 (K) = 0 (U ) 0 (U \ K) = (U ) (U \ K) = (K).
This proves Step 3.
Step 4. 0 (B) = (B) for every Borel set B B.
Let B B. It follows from Step 2, Step 3, and Theorem 2.13 that
0 (B) inf {0 (U ) | B U Rn and U is open}
= inf {(U ) | B U Rn and U is open}
= (B)
= sup {(K) | K B and K is compact}
= sup {0 (K) | K B and K is compact}
0 (B).
This proves Step 4.
Step 5. = 1 and 0 = .
By Step 4 we have = ([0, 1]n ) = 0 ([0, 1]n ) = 1 and hence 0 = = .
We have given two definitions of the Lebesgue measure m : A [0, ].

The first in Definition 2.2 uses the existence and uniqueness of a normalized
translation invariant Borel measure : B [0, ], established in Theo-
rem 2.1 and then defines (Rn , A, m) as the completion of that measure. The
n
second in Definition 2.11 uses the Lebesgue outer measure : 2R [0, ]
of Definition 2.9 and Theorem 2.10 and defines the Lebesgue measure as the
restriction of to the -algebra of -measurable subsets of Rn (see Theo-
rem 2.4). Theorem 2.14 asserts that the two definitions agree. The next
lemma uses the Axiom of Choice to establish the existence of subsets of Rn
that are not Lebesgue measurable.
Lemma 2.15. Let A R be a Lebesgue measurable set such that m(A) > 0.
Then there exists a set B A that is not Lebesgue measurable.
Proof. Consider the equivalence relation on R defined by
def
xy xy Q
for x, y R. By the Axiom of Choice there exists a subset E R which
contains precisely one element of each equivalence class. This means that E
satisfies the following two conditions.
(I) For every x R there exists a rational number q Q such that xq E.
(II) If x, y E and x 6= y then x y / Q.
For q Q define the set
Bq := A (E + q) = {x A | x q E} .
S
Then it follows from (I) that A = qQ Bq .
Fix a rational number q Q. We prove that if Bq is Lebesgue measurable
then m(Bq ) = 0. Assume Bq is Lebesgue measurable, let n N, and define
Bq,q0 ,n := (Bq [n, n]) + q 0 = {x + q 0 | x Bq , |x| n} for q 0 Q.
This set is Lebesgue measurable, its Lebesgue measure is independent of q 0 ,
and Bq,q0 ,n Bq,q00 ,n = for all q 0 , q 00 Q with P
q 0 6= q 00 by condition (II). Since
Bq,q0 ,n [n, n+1] for q 0 [0, 1]Q, we have q0 [0,1]Q m(Bq,q0 ,n ) 2n + 1.
This sum is infinite and all summands agree, so m(Bq [n, n]) = 0. This
holds for all n N and hence m(Bq ) = 0 as claimed. S
If Bq is Lebesgue measurable for all q Q it follows that A = qQ Bq is
a Lebesgue null set, a contradiction. Thus one of the sets Bq is not Lebesgue
measurable and this proves Lemma 2.15.
2.3. THE TRANSFORMATION FORMULA 67
Remark 2.16. (i) Using Lemma 2.15 one can construct a continuous func-
tion f : R R and a Lebesgue measurable function g : R R such that
the composition g f is not Lebesgue measurable (see Example 6.24).
(ii) Let E R be the set constructed in the proof of Lemma 2.15. Then
the set E R R2 is not Lebesgue measurable. This follows from a similar
argument as in Lemma 2.15 using the sets ((E [n, n]) + q) [0, 1]. On the
other hand, the set E {0} R2 is Lebesgue measurable and has Lebesgue
measure zero. However, it is not a Borel set, because its pre-image in R
under the continuous map R R2 : x 7 (x, 0) is the original set E and
hence is not a Borel set.
2.3 The Transformation Formula

The transformation formula describes how the integral of a Legesgue measur-
able function transforms under composition with a C 1 diffeomorphism. Fix a
positive integer n N and denote by (Rn , A, m) the Lebesgue measure space.
For any Lebesgue measurable set X Rn denote by AX := {A A |A X}
the restricted Lebesgue -algebra and by mX := m|AX : AX [0, ] the
restriction of the Lebesgue measure to AX .
Theorem 2.17 (Transformation Formula).
Suppose : U V is a C 1 diffeomorphism between open subsets of Rn .
(i) If f : V [0, ] is Lebesgue measurable then f : U [0, ] is
Lebesgue measurable and
Z Z
(f )|det(d)| dm = f dm. (2.14)
U V
(ii) If E AU and f L1 (mV ) then (E) AV , (f )|det(d)| L1 (mU ),

and Z Z
(f )|det(d)| dm = f dm. (2.15)
E (E)
Proof. See page 72.

The proof of Theorem 2.17 relies on the next two lemmas. The first lemma
is the special case where is linear and f is the characteristic function of a
Lebesgue measurable set. The second lemma is a basic estimate that follows
from the linear case and implies the formula (2.14) for the characteristic
functions of open sets.
Lemma 2.18. Let : Rn Rn be a linear transformation and let A Rn

be a Lebesgue measurable set. Then (A) is a Lebesgue measurable set and
m((A)) = |det()|m(A). (2.16)
Proof. If det() = 0 then (A) is contained in a proper linear subspace of

Rn and hence is a Lebesgue null set for every A A. In this case both
sides of equation (2.16) vanish. Hence it suffices to assume that is a vector
space isomorphism. For vector space isomorphisms we prove the assertion
n
in six steps. Denote by B 2R the Borel -algebra and by := m|B
the restriction of the Lebesgue measure to the Borel -algebra. Thus
is the unique translation invariant Borel measure on Rn that satisfies the
normalization condition ([0, 1]n ) = 1 (Theorem 2.1) and (Rn , A, m) is the
completion of (Rn , B, ) (Theorem 2.14).
Step 1. There exists a unique map : GL(n, R) (0, ) such that
((B)) = ()(B) (2.17)
for every GL(n, R) and every Borel set B B.

Fix a vector space isomorphism : Rn Rn . Since is a homeomorphism
of Rn with its standard topology it follows that (B) B for every B B.
Define the number () [0, ] by
() := (([0, 1)n )). (2.18)
Since ([0, 1)n ) has nonempty interior it follows that () > 0 and since
([0, 1)n ) is contained in the compact set ([0, 1]n ) it follows that () < .
Now define the map : B [0, ] by
((B))
(B) := for B B.
()
Then is a normalized translation invariant Borel measure. The -addi-
tivity follows directly from the -additivity of , the formula () = 0 is
obvious from the definition, that compact sets have finite measure follows
from the fact that (K) is compact if and only if K Rn is compact, the
translation invariance follows immediately from the translation invariance
of and the fact that (B + x) = (B) + (x) for all B B and all x Rn ,
and the normalization condition ([0, 1)n ) = 1 follows directly from the
definition of . Hence = by Theorem 2.1. This proves Step 1.
Step 2. Let be as in Step 1 and let A A and GL(n, R). Then

(A) A and m((A)) = ()m(A).
By Theorem 2.14 there exist Borel sets B0 , B1 B such that B0 A B1
and (B1 \ B0 ) = 0. Then (B0 ) (A) (B1 ) and, by Step 1,
((B1 ) \ (B0 )) = ((B1 \ B0 )) = ()(B1 \ B0 ) = 0.
Hence (A) is a Lebesgue measurable set and
m((A)) = ((B0 )) = ()(B0 ) = ()m(A)
by Theorem 2.14 and Step 1. This proves Step 2.
Step 3. Let be as in Step 1 and let = diag(1 , . . . , n ) be a diagonal
matrix with nonzero diagonal entries i R \ {0}. Then () = |1 | |n |.
Define I := [1, 1] and Ii := [|i |, |i |] for i = 1, . . . , n. Then Q := I n
has Lebesgue measure m(Q) = 2n and the cuboid (Q) = I1 In has
Lebesgue measure m((Q)) = 2n |1 | |n | by part (iv) of Corollary 2.12.
Hence Step 3 follows from Step 2.
Step 4. The map : GL(n, R) (0, ) in Step 1 is a group homomorphism
from the general linear group of automorphisms of Rn to the multiplicative
group of positive real numbers.
Let , GL(n, R). Then it follows from (2.17) with B := ([0, 1)n ) and
from the definition of () in (2.18) that
() = (([0, 1)n )) = ()(([0, 1)n )) = ()().
Thus is a group homomorphism as claimed and this proves Step 4.
Step 5. The map : GL(n, R) (0, ) in Step 1 is continuous with respect
to the standard topologies on GL(n, R) and (0, ).
It suffices to prove continuity at the identity. Define the norms
kxk
kxk := max |xi | , kk := sup (2.19)
i=1,...,n 06=xRn kxk
for x Rn and a linear map : Rn Rn . Denote the closed unit ball in
Rn by Q := {x Rn | kxk 1} = [1, 1]n . Fix a constant 0 < < 1 and a
linear map : Rn Rn such that k 1lk < . Then GL(n, R) and

1
X 1
= (1l )k , k1 k < .
k=0
1
Thus (Q) (1 + )Q and (1 )1 (Q) Q. Hence

(1 )Q (Q) (1 + )Q.
Since () = m((Q))/m(Q) by Step 2 and m(rQ) = rn m(Q) for r > 0 by
Steps 2 and 3, this shows that (1 )n () (1 + )n . Given > 0
choose a constant 0 < < 1 such that 1 < (1 )n < (1 + )n < 1 + .
Then
k 1lk < = |() 1| <
for all GL(n, R). This proves Step 5.
Step 6. () = |det()| for all GL(n, R).
If GL(n, R) is diagonalizable with real eigenvalues then () = |det()|
by Step 3 and Step 4. If GL(n, R) has only real eigenvalues then it
can be approximated by a sequence of diagonalizable automorphisms with
real eigenvalues and hence it follows from Step 5 that () = |det()|. Since
every automorphism of Rn is a finite composition of automorphisms with real
eigenvalues (elementary matrices) this proves Step 6. Lemma 2.18 follows
immediately from Step 2 and Step 6.
Define the metric d : Rn Rn [0, ) by d (x, y) := kx yk for
x, y Rn , where kk is as in (2.19). The open ball of radius r > 0 about a
point a = (a1 , . . . , an ) Rn with respect to this metric is the open cube
Br (a) := (a1 r, a1 + r) (an r, an + r)
and its closure is B r (a) = [a1 r, a1 + r] [an r, an + r].
Lemma 2.19. Let U Rn be an open set and let K U be a compact
subset. Let : U Rn be a continuously differentiable map such that
det(d(x)) 6= 0 for all x K. For every > 0 there exists a constant > 0
such that the following holds. If 0 < r < , a Rn , and R Rn satisfy
Br (a) R B r (a) K then

m((R)) |det(d(a))| m(R) < m(R). (2.20)

Proof. The maps K R : x 7 kd(x)1 k and K R : x 7 |det(d(x))|

are continuous by assumption. Since K is compact these maps are bounded.
Hence there is a constant c > 0 such that
d(x)1 c,

|det(d(x))| c for all x K. (2.21)
Let > 0 and choose a constant 0 < < 1 so small that

1 < (1 )n < (1 + )n < 1 + . (2.22)
c c
n
Choose > 0 so small that, for all x, y R ,

x, y K, kx yk < = kd(x) d(y)k < . (2.23)
c
Such a constant exists because the map d : U Rnn is uniformly contin-
uous on the compact set K U . We prove that the assertion of Lemma 2.19
holds with this constant .
Choose a Rn and 0 < r < such that B r (a) K. Then ka xk <
for all x B r (a). By (2.23) with := d(a) this implies

kd(x) k < 1
for all x B r (a).
c k k
Here the first step follows from (2.23) and the second step follows from (2.21).
1
n

Define the map : U R by (x) := (x) (a) . Then (a) = 0
and d(x) = 1 d(x) and hence, by (2.23),
kd(x) 1lk = 1 (d(x) ) 1 kd(x) k

for all x B r (a). By Theorem C.1 this implies

B(1)s (0) (Br (a)) (B r (a)) B (1+)s (0) (2.24)
Now fix a subset R Rn such that Br (a) R B r (a). Then by (2.24)
(1 )(Br (0)) (R) (a) (1 + )(B r (0)).
Since m(R) = m(Br (0)) = m(B r (0)) by part (iv) of Corollary 2.12, it follows
from Lemma 2.18 and the inequalities (2.21) and (2.22) that

|det()| m(R) m(R) 1 |det()| m(R)
c
< (1 )n |det()| m(R)
= m((1 )(Br (0)))
m((R))
m((1 + )(B r (0)))
= (1 + )n |det()| m(R)

< 1+ |det()| m(R)
c
|det()| m(R) + m(R).
This proves (2.20) and Lemma 2.19.
Proof of Theorem 2.17. The proof has seven steps. The first four steps es-
tablish equation (2.14) for the characteristic functions of open sets, compact
sets, Borel sets, and Lebesgue measurable sets with compact closure in U .
Step 1. If W Rn is an open set with compact closure W U then
Z
m((W )) = |det(d)| dm.
W
Fix a constant > 0. Then there exists a constant > 0 that satisfies the
following two conditions.
(a) If a Rn , 0 < r < , R Rn satisfy Br (a) R B r (a) W then
m(R)
m((R)) |det(d(a))| m(R) < .

2 m(W )
(b) For all x, y W

kx yk < = |det(d(x)) det(d(y))| < .
2 m(W )
That > 0 can be chosen so small that (a) holds follows from Lemma 2.19
and that it can be chosen so small that (b) holds follows from the fact
that the function det(d) : U R is uniformly continuous on the com-
pact set W . Now write W as a countable union of pairwise disjoint half-open
cubes Ri Rn centered at ai Rn with side lengths 2ri such that 0 < ri < .
(See page 64.) Then Bri (ai ) Ri B ri (ai ) W for all i and
X X
m(W ) = m(Ri ), m((W )) = m((Ri )). (2.25)
i i
It follows from (2.25) and (a) that

X
m((W )) |det(d(ai ))| m(Ri ) < . (2.26)

i
2

P
It follows from (b) that ||det(d)| i |det(d(ai ))| Ri | < 2 m(W )
on W .
Integrate this inequality over W to obtain
Z
X
|det(d)| dm |det(d(ai ))| m(Ri ) < . (2.27)

W i
2
R
By (2.26) and (2.27) we have |m((W )) W
|det(d)| dm| < . Since this
holds for all > 0, Step 1 follows.
Step 2. If K U is compact then

Z
m((K)) = |det(d)| dm.
K
Choose an open set W K with compact closure W U . Then

m((K)) = m((W )) m((W \ K))
Z Z Z
= |det(d)| dm |det(d)| dm = |det(d)| dm.
W W \K K
Here the second equation follows from Step 1. This proves Step 2.
Step 3. If B B has compact closure B U then (B) B and
Z
m((B)) = |det(d)| dm.
B
That (B) is a Borel set follows from the fact that it is the pre-image of
the Borel set B under the continuous map 1 : V U (Theorem 1.20).
Abbreviate b := m((B)). Assume first that b < and fix a constant > 0.
Then it follows from Theorem 2.13 that there exists an open set W 0 Rn
with compact closure W 0 V such that (B) W 0 and m(W 0 ) b . Define K := 1 (K 0 ) and
W := 1 (W 0 ). Then K is compact, W is open, W U is compact, and
K B W, b < m((K)) m((W )) 0 and so
Z
|det(d)| dm = b = m((B)).
B
If b = then, by Theorem 2.13, there exists a sequence of compact sets

0 0 1 0
R i (B) such that (Ki ) > i. Hence Ki := (Ki ) is compact and
K
|det(d)| dm = ((Ki )) > i by Step 2. Since Ki B this implies
RKi R
B
|det(d)| dm > i for all i N and hence B |det(d)| dm = = m((B)).
This proves Step 3.
Step 4. If A A has compact closure A U then (A) A and

Z
m((A)) = |det(d)| dm.
A
Let A A. By Theorem 2.14 there exist Borel sets B0 , B1 B, with compact

closure contained in U , such that B0 A B1 and m(B1 \ B0 ) = 0. Then
(B0 ) (A) (B1 ) and it follows from Step 3 that
R (B0 ) and (B1 ) are
Borel sets and m((B1 ) \ (B0 )) = m((B1 \ B0 )) = B1 \B0 |det(d)| dm = 0.
Hence it follows from TheoremR 2.14 that (A) is
R a Lebesgue measurable set
and m((A)) = m((B0 )) = B0 |det(d)| dm = A |det(d)| dm. Here the last
equation follows from the fact that the set A \ B0 is Lebesgue measurable
and has Lebesgue measure zero. This proves Step 4.
Step 5. Assertion (i) of Theorem 2.17 holds for every Lebesgue measurable
step function f = s : V R whose support is a compact subset of V .
Write s = ì=1 i Ai with i R and Ai A such that Ai is a compact
P
subset of V for all i. Then 1 (Ai ) is a Lebesgue
P measurable set with com-
pact closure in U by Step 4. Hence s = ì=1 i 1 (Ai ) is a Lebesgue
measurable step function and
Z `
X Z
(s )|det(d)| dm = i |det(d)| dm
U i=1 1 (Ai )
X` Z
= i m(Ai ) = s dm.
i=1 V
Here the second equation follows from Step 4. This proves Step 5.
Step 6. We prove (i).
By Theorem 1.26 there is a sequence of Lebesgue measurable step functions
si : V [0, ) such that 0 s1 s2 and f (x) = limi si (x) for
every x V . Choose an exhausing
S sequence of compact sets Ki V such
that Ki Ki+1 for all i and i Ki = V and replace si by si Ki . Then part (i)
follows from Step 5 and the Lebesgue Monotone Convergence Theorem 1.37.
Step 7. We prove (ii).
For E = U part (ii) follows from part (i) and the fact that (f ) = f .
If F AV then f := F |V is Lebesgue measurable, hence f = 1 (F ) |U
2.4. LEBESGUE EQUALS RIEMANN 75
is Lebesgue measurable by part (i), and so 1 (F ) AU . Replace by 1

to deduce that if E AU then (E) AV . Then (ii) follows for all E AU
by replacing f with f (E) . This proves Step 7 and Theorem 2.17.
2.4 Lebesgue Equals Riemann

The main theorem of this section asserts that the Lebesgue integral of a func-
tion on Rn agrees with the Riemann integral whenever the latter is defined
and the function in question has compact support. The section begins with
a recollection of the definition of the Riemann integral. (For more details
see [9, 19, 21].)
The Riemann Integral

Recall the notation R(x, k) := x + [0, 2k )n for x Rn and k N, which was
used in the proof of Theorem 2.1 on page 64. The closure of R(x, k) is the
closed cube R(x, k) = x + [0, 2k ]n . The sets R(`, k), with ` ranging over the
countable set 2k Zn , form a partition of the Euclidean space Rn .
Definition 2.20. Let f : Rn R be a bounded function whose support

supp(f ) := x Rn f (x) 6= 0
is a bounded subset of Rn . For k N define the lower sum S(f, k) R and

the upper sum S(f, k) R by
!
X
S(f, k) := inf f 2nk ,
R(`,k)
`2k Zn
! (2.28)
X
S(f, k) := sup f 2nk .
`2k Zn R(`,k)
These are finite sums and satisfy supk S(f, k) inf k S(f, k). The function
f : Rn R is called Riemann integrable if supk S(f, k) = inf k S(f, k).
The Riemann integral of a Riemann integrable function f : Rn R is the
real number
Z
R(f ) := f (x) dx := sup S(f, k) = inf S(f, k) = lim S(f, k). (2.29)
Rn kN kN k
Remark 2.21. The Riemann integral can also be defined by allowing for
arbitrary partitions of Rn into cuboids (see [19, Definition 2.3]) or in terms of
convergence of the so-called Riemann sums (see [21, Definition 7.1.2]). That
all three definitions agree is proved in [19, Satz 2.8] and [21, Theorem 7.1.8]).
Definition 2.22. A bounded set A Rn us called Jordan measurable if
its characteristic function A : Rn R is Riemann integrable. The Jordan
measure of a Jordan measurable set A Rn is the real number
J (A) := R(A )
Z
= A (x) dx (2.30)
Rn
n o
nk k n

= lim 2 # ` 2 Z R(`, k) A 6= .
k
Exercise 2.23. (i) Prove the last equation in (2.30).

(ii) Prove that a bounded set A Rn is Jordan measurable if and only if its
boundary A = A \ int(A) is a Jordan null set. (See Definition 2.7.)
(iii) Prove that the closure of a Jordan null set is a Jordan null set.
The Lebesgue and Riemann Integrals Agree

Theorem 2.24. (i) If f : Rn R is Riemann integrable then f L1 (m)
and its Lebesgue integral agrees with the Riemann integral, i.e.
Z
f dm = R(f ).
Rn
(ii) If A Rn is Jordan measurable then A is Lebesgue measurable and

m(A) = J (A).
Proof. Assertion (ii) follows from (i) by taking f = A . Thus it remains to
prove (i). Let f : Rn R be a Riemann integrable function. Then f is
bounded and has bounded support. Define the functions f k , f k : Rn R by
f k (x) := inf f, f k (x) := sup f for x R(`, k), ` 2k Zn . (2.31)

R(`,k) R(`,k)
These are Lebesgue measurable step functions and

Z Z
f k dm = S(f, k), f k dm = S(f, k).
Rn Rn
2.4. LEBESGUE EQUALS RIEMANN 77
They also satisfy

f k f k+1 f f k+1 f k
for all k N. Define the functions f , f : Rn R by
f (x) := lim f k (x), f (x) := lim f k (x) for x Rn .

k k
Then
f (x) f (x) f (x)
for every x Rn . Moreover, |f k | and |f k | are bounded above by the Lebesgue
integrable function cA , where c := supxRn |f (x)| and A := [N, N ]n with
N N chosen such that supp(f ) [N, N ]n . Hence it follows from the
Lebesgue Dominated Convergence Theorem 1.45 that f and f are Lebesgue
integrable and
Z Z
f dm = lim f dm = lim S(f, k) = R(f )
Rn k Rn k k
Z Z
= lim S(f, k) = lim f k dm = f dm.
k k Rn Rn
By Lemma 1.49, with f replaced by f f , this implies that f = f = f

Lebesgue almost everywhere. Hence f L1 (m) and
Z Z
f dm = f dm = R(f ).
Rn Rn

The discussion in this section is restricted to Riemann integrable functions
f : Rn R with compact support and Theorem 2.24 asserts that for such
functions the Riemann integral agrees with the Lebesgue integral. When f
does not have compact support and is locally Riemann integrable, the im-
proper Riemann integral is defined by
Z Z
f (x)dx := lim f (x) dx, (2.32)
Rn r Br
provided that the limit exists. Here Br Rn denotes the ball of radius r
centered at the origin. There are many
R examples where the limit (2.32) exists
even though the Lebesgue integral Rn |f | dm is infinite and so the Lebesgue
integral of f does not exist. An example is the function f : R R given by

f (x) := x1 sin(x) for x R \ {0} and f (0) := 1. This function is continuous
and is not Lebesgue integrable, but the improper Riemann integral exists
and is equal to (see Example 7.49). Improper integrals play a central
role in Fourier analysis, probability theory, and partial differential equations.
However, this topic will not be pursued any further in this book
2.5 Exercises
Exercise 2.25. Show that the Cantor set in R is a Jordan null set. Show
that Q [0, 1] is a Lebesgue null set but not a Jordan null set. Show that
A Rn is a Lebesgue null set if and only if (A) = 0. Find an open set
U R whose boundary has positive Lebesgue measure.
Exercise 2.26. Prove that every subset of a proper linear subspace of Rn
is Lebesgue measurable and has Lebesgue measure zero. Find a Jordan
measurable subset of Rn that is not a Borel set. Find a bounded Lebesgue
measurable subset of Rn with positive Lebesgue measure that is neither a
Borel set nor Jordan measurable.
Exercise 2.27. Let A, B Rn be Lebesgue null sets. Prove that their sum
A + B := {x + y | x A, y B} is a Lebesgue null set.
Exercise 2.28. Let (X, A, ) be a measure space and define the function
: 2X [0, ] by

(B) := inf (A) A A, B A . (2.33)
(i) Prove that is an outer measure and that A A().
(ii) Assume (X) < . Prove that the measure space (X, A(), |A() ) is
the completion of (X, A, ). Hint: Show that for every subset B X there
exists a set A A such that B A and (B) = (A).
(iii) Let X be a set and A ( X be a nonempty subset. Define
A := {, A, Ac , X}, () := (A) := 0, (Ac ) := (X) := .
Prove that (X, A, ) is a measure space. Given B X, prove that (B) = 0
whenever B A and (B) = whenever B 6 A. Prove that A() = 2X
and that the completion of (X, A, ) is the measure space (X, A , ) with
A = {B X | B A or Ac B} and = |A . (Thus the hypothesis
(X) < cannot be removed in part (ii).)
2.5. EXERCISES 79
Exercise 2.29. Let f : R R be continuously differentiable and define

A := {x R | f 0 (x) = 0} .
Prove that f (A) is a Lebesgue null set. Hint: Consider the sets
An, := x R | |x| < n, |f 0 (x)| < 2n .

Exercise 2.30. Find a continuous function f : [0, ) R such that f is

RT
not Lebesgue integrable but the limit limT 0 f (t) dt exists.
Exercise 2.31. Determine the limits of the sequences
Z n Z n
x n x/2 x n 2x
an := 1 e dx, bn := 1+ e dx, n N.
0 n 0 n
Hint: Use the Lebesgue Dominated Convergence Theorem 1.45.
Exercise 2.32. Let (R, A, m) be the Lebesgue measure space. Construct a
Borel set E R such that
m(E I)
0< <1
m(I)
for every nonempty bounded open interval I R. (See also Exercise 6.21.)
Exercise 2.33. Find the smallest constant c such that
log(1 + et ) c + t for all t 0.
Does the limit
1 1
Z
log 1 + enf (x) dx

lim
n n 0
exist for every Lebesgue integrable function f : [0, 1] R? Determine the

limit when it does exist.
Exercise 2.34. Let (Rn , A, m) be the Lebesgue measure space and let
: Rn Rn
be a C 1 -diffeomorphism. Prove that A = A and that
Z
1
( m)(A) = 1
dm for all A A.
A |det(d) |
Hint: See Exercise 1.69 and Theorems 1.40 and 2.17.

Exercise 2.35 (Hausdorff Measure). Let (X, ) be a metric space and

fix a real number d 0. The diameter of a subset A X is defined by
diam(A) := sup (x, y). (2.34)
x,yA
For > 0 define the function d, : 2X [0, ] by

I is finite or countably infinite,

X
d, (A) := inf diam(Di )d Di X, diam(D i ) < for i I . (2.35)

S
and A iI Di

iI
for A X. Thus d, () = 0 and d, (A) = whenever A does not admit a

countable cover by subsets of diameter less than . Moreover, the function
7 d, (A) is nonincreasing for every subset A X. The d-dimensional
Hausdorff outer measure is the function d : 2X [0, ] defined by
d (A) := sup d, (A) = lim d, (A) for A X. (2.36)
>0 0
Prove the following.

(i) d is an outer measure.
(ii) If A, B X satisfy (A, B) := inf {(x, y) | x A, y B} > 0 then
d (A B) = d (A) + d (B). Hence, by Theorems 2.4 and 2.5, the set

Ad := A X A is d -measurable
is a -algebra containing the Borel sets and
d := d |Ad : Ad [0, ]
is a measure. It is called the d-dimensional Hausdorff measure on X.
Hausdorff measures play a central role in geometric measure theory.
(iii) If d = 0 then A0 = 2X and 0 = 0 is the counting measure.
(iv) The n-dimensional Hausdorff measure on Rn agrees with the Lebesgue
measure up to a factor (the Lebesgue measure of the ball of radius 1/2).
(v) Let A X be nonempty. The Hausdorff dimension of A is the number
dim(A) := sup {r 0 | r (A) = } = inf {s 0 | s (A) = 0} . (2.37)
The second equality follows from the fact that d (A) > 0 implies r (A) =
for 0 r < d, and d (A) < implies s (A) = 0 for s > d.
(vi) The Hausdorff dimension of a smooth embedded curve Rn is d = 1
and its 1-dimensional Hausdorff measure 1 () is the length of the curve.
(vii) The Hausdorff dimension of the Cantor set is d = log(2)/ log(3).
Chapter 3
Borel Measures
The regularity properties established for the Lebesgue (outer) measure in

Theorem 2.13 play an important role in much greater generality. The present
chapter is devoted to the study of Borel measures on locally compact Haus-
dorff spaces that satisfy similar regularity properties. The main result is the
Riesz Representation Theorem 3.15. We begin with some further recollec-
tions about topological spaces.
Let (X, U) be a topological space (see Definition 1.9). A neighborhood
of a point x X is a subset A X that contains x in its interior, i.e.
x U A for some open set U . X is called a Hausdorff space if any
two distinct points in X have disjoint neighborhoods, i.e. for all x, y X
with x 6= y there exist open sets U, V X such that x U , y V , and
U V = . X is called locally compact if every point in X has a compact
neighborhood. It is called -compact if there exists a sequence
S of compact
sets Ki X, i N, such that Ki Ki+1 for all i and X = i=1 Ki .
3.1 Regular Borel Measures

Assume throughout that (X, U) is a locally compact Hausdorff space and
denote by B 2X the Borel -algebra. Thus B is the smallest -algebra on X
that contains all open sets. In the context of this chapter it is convenient to
include local finiteness (compact sets have finite measure) in the definition
of a Borel measure. There are other geometric settings, such as the study
of Hausdorff measures (Exercise 2.35), where one allows for compact sets to
have infinite measure, but these are not discussed here.
81
82 CHAPTER 3. BOREL MEASURES
Definition 3.1. A measure : B [0, ] is called a Borel measure if

(K) < for every compact set K X. A measure : B [0, ] is called
outer regular if

(B) = inf (U ) B U X and U is open (3.1)
for every Borel set B B, is called inner regular if

(B) = sup (K) K B and K is compact (3.2)
for every Borel set B B, and is called regular if it is both outer and inner
regular. A Radon measure is an inner regular Borel measure.
Example 3.2. The restriction of the Lebesgue measure on X = Rn to the

Borel -algebra is a regular Borel measure by Theorem 2.13.
Example 3.3. The counting measure on X = N with the discrete topology

U = B = 2N is a regular Borel measure.
Example 3.4. Let (X, U) be any locally compact Hausdorff space and fix
a point x0 X. Then the Dirac measure = x0 at x0 is a regular Borel
measure (Example 1.31).
Example 3.5. Let X be an uncountable set equipped with the discrete

topology U = B = 2X . Define : B [0, ] by

0, if B is countable,
(B) :=
, if B is uncountable.
This is a Borel measure. Moreover, a subset K X is compact if and only

if it is finite. Hence (X) = and (K) = 0 for every compact set K X.
Thus is not a Radon measure.
The next example occupies three pages and illustrates the subtlety of
the subject (see also Exercise 18 in Rudin [17, page 59]). It constructs a
compact Hausdorff space (X, U) and a Borel measure on X that is not a
Radon measure. More precisely, there is a point X such that the open
set U := X \ {} is not -compact and satisfies (U ) = 1 and (K) = 0
for every compact subset K U . This example is a kind of refinement of
Example 3.5. It is due to Dieudonne.
3.1. REGULAR BOREL MEASURES 83
Example 3.6 (Dieudonn es measure). (i) Let (X, 4) be an uncountable

well ordered set with a maximal element X such that every element
x X \ {} has only countably many predecessors. Here a set is called
countable iff it is finite or countably infinite. (Think of this as the uncountable
Mount Everest; no sequence reaches the mountain peak .) Thus the relation
4 on X satisfies the following axioms.
(a) If x, y, z X satisfy x 4 y and y 4 z then x 4 z.
(b) If x, y X satisfy x 4 y and y 4 x then x = y.
(c) If x, y X then x 4 y or y 4 x.
(d) If =
6 A X then there is an a A such that a 4 x for all x A.
(e) If x X \ {} then x 4 and the set {y X | y 4 x} is countable.
Define the relation on X by x y iff x 4 y and x 6= y. For =
6 AX
denote by min(A) A the unique element of A that satisfies min(A) 4 x for
all x A. (See conditions (b) and (d).) For x X define
Sx := {y X | x y} , Px := {y X | y x} .
Thus Px is the set of predecessors of x and Sx is the set of successors of x.
If x X \ {} then Px is countable and Sx is uncountable. Define the map
s : X \ {} X \ {}, s(x) := min(Sx ).
Then X \ Sx = Ps(x) = Px {x} for all x X. Let U 2X be the smallest
topology that contains the sets Px and Sx for all x X. A set U X is
open in this topology if it is a union of sets of the form Pb , Sa and Sa Pb .
(ii) We prove that (X, U) is a Hausdorff space. Let x, y X such that x 6= y
and suppose without loss of generality that x y. Then Ps(x) and Sx are
disjoint open sets such that x Ps(x) and y Sx .
(iii) We prove that every nonempty compact set K X contains a largest
element max(K) K such that K Smax(K) = . This is obvious when
K because S = . Thus assume / K and define
V := {x X | K Px } .
Since V this set is nonempty and min(X) min(V ) =: v because K =
6 .
Since X \ K is open and v X \ K there exist elements a, b X such that
a v b and Sa Pb K = . This implies
K Pv \ (Sa Pb ) Pb \ (Sa Pb ) X \ Sa = Ps(a) .
Hence K \ {a} Ps(a) \ {a} = Pa and K 6 Pa because a v and so a
/ V.
This implies a K Ps(a) and hence K Sa = K \ Ps(a) = .
(iv) We prove that (X, U) is compact. Let {Ui }iI be an open cover of X.
We prove by induction that there exist finite sequences x1 , . . . , x` X and
i1 , . . . , i` I such that xk Uik \ Uik1 and Sxk Ui1 Uik1 for k 2,
and X = `j=1 Uij . Define x1 := and choose i1 I such that Ui1 .
S
If Ui1 = X the assertion holds with ` = 1. Now suppose, by induction,
that x1 , . . . , xk and i1 , . . . , ik have been constructed such that xj Uij for
j = 1, . . . , k and Sxk Ui1 Uik1 . If Ui1 Uik = X we are done with
` = k. Otherwise Ck := X \ Ui1 Uik is a nonempty compact set and we
define xk+1 := max(Ck ) by part (iii). Then xk+1 Ck and Ck Sxk+1 = .
Hence Sxk+1 Ui1 Uik . Choose ik+1 I such that xk+1 Uik+1 .
This completes the induction argument. The induction must stop because
xk+1 xk for all k and every strictly decreasing sequence in X is finite by
the well ordering axiom (d). This shows that (X, U) is compact.
(v) Let Ki X, i N, be a sequence of uncountable compact sets. We
prove that the compact set \
K := Ki
iN
is uncountable. To see this, we first prove that
K \ {} =
6 . (3.3)
Choose a sequence xn X \ {} such that xn xn+1 for all n N and

x2k +i Ki for 1 i 2k 1 and k N. That such a sequence exists follows
by induction from the fact that the set X \ Sxn = Ps(xn ) is countable
S for each
n while the sets Ki are uncountable for all i. Now the set P := nN Pxn is
countable and hence the set
[ [ \ \
S := X \ P = X \ Pxn = X \ Ps(xn ) = X \ Ps(xn ) = S xn
nN nN nN nN
is uncountable. Hence x := min(S) . We prove that x Ki for all i N.

Assume by contradiction that x / Ki for some i. Then there are elements
a, b X such that a x b and U := Pb Sa X \ Ki . If xn 4 a for
all n N then P Pa and so a X \ P = S, which is impossible because
a x = min(S). Thus there must be an integer n0 N such that a xn0 .
This implies a xn x b and hence xn U X \ Ki for all n n0 ,
contradicting the fact that x2k +i Ki for all k N. This contradiction
shows that our assumption that x / Ki for some i N must have been
wrong. Thus x K and this proves (3.3).
We prove that K is uncountable. Assume by contradiction that K is

countable and choose a sequence xi K such that K \ {} = {xi | i N}.
Then s(xi ) and Ki0 := Ki Sxi = KT i \ Ps(xi ) is an uncountable compact
set for every i N. Moreover, K 0 := iN Ki0 K \ {xi | i N} = {},
contradicting the fact that K 0 \ {} 6= by (3.3). This contradiction shows
that K is uncountable as claimed.
(vi) Define A 2X by

A {} contains an uncountable compact set,
A := A X .
or Ac {} contains an uncountable compact set.
We prove that this is a -algebra. To see this note first that X A and that
c
A A implies S A A by definition. Now choose a sequence Ai A and
denote A := iN Ai . If one of the sets Ai {} contains an uncountable
compact set then so does the set A{}. If none of the sets Ai {} contains
an uncountable compact set then the set Aci {}Tcontains an uncountable
compact set for all i N and hence so does the set iN (Aci {}) = Ac {}
by part (v). In both cases it follows that A A.
(vii) Define the map : A [0, ] by

1, if A {} contains an uncountable compact set,
(A) :=
0, if Ac {} contains an uncountable compact set.
This map is well defined because the sets A {} and Ac {} cannot
both contain uncountable compact sets by part (v). It satisfies () = 0.
Moreover, if Ai A is a sequence of pairwise disjoint measurable sets then
at most one P Ai {} can contain an uncountable compact set and
S of the sets
hence ( iN Ai ) = iN (Ai ). Hence is a measure.
(viii) The -algebra B 2X of all Borel sets in X is contained in A. To
see this, let U X be open. If U c is uncountable then U c {} is an
uncountable compact set and hence U A. If U c is countable T choose a
sequence xi U cSsuch that U c \ {}S= {xi | i N} and define S := iN Sxi .
Then X \ S = iN (X \ Sxi ) = iN Ps(xi ) is a countable set and hence
s := min(S) . Since xi s for all i N it follows that U c \ {} Ps .
Hence X \ Ps is an uncountable compact subset of U {} and so U A.
(ix) The set U := X \ {} is uncountable and every compact subset of U
is countable by part (v). Hence (K) = 0 for every compact subset K U
and (U ) = 1 because U {} = X is an uncountable compact set. Thus
|B : B [0, ] is a Borel measure but not a Radon measure.
The next lemma and theorem are included here in preparation for the
Riesz Representation Theorem 3.15. They explain the relation between the
various regularity properties of Borel measures.
Lemma 3.7. Let : B [0, ] be an outer regular Borel measure that is
inner regular on open sets, i.e.

(U ) = sup (K) K U and K is compact (3.4)
for every open set U X. Then the following holds.
(i) Every Borel set B X with (B) < satisfies (3.2).
(ii) If X is -compact then is regular.
Proof. We prove (i). Fix a Borel set B X with (B) < and a constant
> 0. Since is outer regular, there exists an open set U X such that

B U, (U ) < (B) + .
2
Thus U \ B is a Borel set and (U \ B) = (U ) (B) < /2. Use the outer
regularity of again to obtain an open set V X such that

U \ B V, (V ) < .
2
Now it follows from (3.4) that there exists a compact set K X such that

K U, (K) > (U ) .
2
Define C := K \ V . Since X is a Hausdorff space, K is closed, hence C is a
closed subset of K, and hence C is compact (see Lemma A.2). Moreover,
C U \ V B, B \ C (B \ K) V (U \ K) V,
and hence (B \ C) (U \ K) + (V ) < . This proves (i).
We prove (ii). Choose a sequenceSof compact sets Ki X such that
Ki Ki+1 for all i N and X = i=1 Ki . Fix a Borel set B B. If
(B) < then B satisfies (3.2) by (i). Hence assume (B) = . Then
it follows from part (iv) of Theorem 1.28 that limi (B Ki ) = . For
each integer n N choose in N such that
(B Kin ) > n.
Since (B Kin ) (Kin ) < it follows from (i) that (3.2) holds with B
replaced by B Kin . Hence there exists a compact set Cn B Kin such
that (Cn ) > n. This proves (ii) and Lemma 3.7.
Theorem 3.8. Let 1 : B [0, ] be an outer regular Borel measure that

is inner regular on open sets. Define 0 : B [0, ] by

0 (B) := sup 1 (K) K B and K is compact for B B. (3.5)
Then the following holds
(i) 0 is a Radon measure, it agrees with 1 on all compact sets and all open
sets, and 0 (B) 1 (B) for all B B.
(ii) If X is -compact then 0 = 1 .
(iii) If f : X R is a compactly supported continuous function then
Z Z
f d0 = f d1 . (3.6)
X X
R : B R[0, ] be a Borel measure that is inner regular on open sets.

(iv) Let
Then X f d = X f d1 for every compactly supported continuous function
f : X R if and only if 0 (B) (B) 1 (B) for all B B.
Proof. We prove that 0 is a measure. It follows directly from the definition
that 0 () = 0. Now assumeS that Bi B is a sequence of pairwise disjoint
Borel sets and define B := i=1 Bi . Choose any compact set K B. Then
1 (Bi K) < and hence it follows from part (i) of Lemma 3.7 that
0 (Bi K) = 1 (Bi K)
for all i N. This implies

X
X
X
1 (K) = 1 (Bi K) = 0 (Bi K) 0 (Bi ).
i=1 i=1 i=1
Take the supremum over all compact sets K B to obtain

X
0 (B) 0 (Bi ). (3.7)
i=1
To prove the converse inequality, it suffices to assume that 0 (B) < .

Then 0 (Bi ) 0 (B) < for all i N. Fix a constant > 0 and choose a
sequence of compact sets Ki Bi such that 1 (Ki ) > 0 (Bi ) 2i for all i.
Then, for every n N, the set K1 Kn is a compact subset of B and
n
X n
X
0 (B) 1 (K1 Kn ) = 1 (Ki ) > 0 (Bi ) .
i=1 i=1
Now take the limit n to obtain

X
0 (B) 0 (Bi ) .
i=1
holds for all > 0 it follows that 0 (B)

P
Since thisP i=1 0 (Bi ) and hence

0 (B) = i=1 0 (Bi ) by (3.7). This shows that 0 is a measure. Moreover
it follows directly from the definition of 0 that 0 (K) = 1 (K) for every
compact set K X. Since 1 is inner regular on open sets it follows that
0 (U ) = 1 (U ) for every open set U X. Since 0 (K) = 1 (K) for every
compact set K X it follows from the definition of 0 in (3.5) that 0 is
inner regular and hence is a Radon measure. The inequality 0 (B) 1 (B)
for B B follows directly from the definition of 0 . This proves part (i).
Part (ii) follows directly from the definition of 0 and part (ii) of Lemma 3.7.
We prove part (iii). Assume first that s : X R is a Borel measurable
step function with compact support. Then
`
X
s= i Bi
i=1
where i R and Bi B with 1 (Bi ) < . Hence 0 (Bi ) = 1 (Bi ) by

part (i) of Lemma 3.7 and hence
Z `
X Z
s d0 = i 0 (Bi ) = s d1 .
X i=1 X
Now let f : X [0, ] be a Borel measurable function with compact

support. By Theorem 1.26 there exists a sequence of Borel measurable
step functions sn : X [0, ) such that 0 s1 (x) s2 (x) and
f (x) = limn sn (x) for all x X. Thus sn has compact support for each n.
By the Lebesgue Monotone Convergence Theorem 1.37 this implies
Z Z Z Z
f d0 = lim sn d0 = lim sn d1 = f d1 .
X n X n X X
If f : X R is a 1 -integrable function with compact support then, by what

we have just proved, X f d0 = X f d1 < , so f is 0 -integrable and
R R
satisfies (3.6). This proves part (iii).

We prove part (iv) in four steps.

Step 1. Let : B [0, ] be a Borel measure such that
Z Z
f d = f d1 (3.8)
X X
for every compactly supported continuous function f : X R. Then
(K) 1 (K), 1 (U ) (U )
for every compact set K X and every open set U X.

Fix an open set U X and a compact set K U . Then Urysohns
Lemma A.1 asserts that there exists a compactly supported continuous func-
tion f : X R such that
f |K 1, supp(f ) U, 0 f 1.
Hence it follows from equation (3.8) that

Z Z
(K) f d = f d1 1 (U )
X X
and likewise Z Z
1 (K) f d1 = f d (U ).
X X
Since (K) 1 (U ) for every open set U X containing K and 1 is outer

regular we obtain
(K) inf {1 (U ) | K U X and U is open} = 1 (K).
Since 1 (K) (U ) for every compact set K U and 1 is inner regular on

open sets we obtain
1 (U ) = sup {1 (K) | K U and K is compact} (U ).
This proves Step 1.

Step 2. Let be as in Step 1 and assume in addition that is inner
regular on open sets. Then (K) = 1 (K) for every compact set K X and
(U ) = 1 (U ) for every open set U X.
If U X is an open set then

(U ) = sup {(K) | K U and K is compact}
sup {1 (K) | K U and K is compact}
= 1 (U ) (U ).
Here the two inequalities follow from Step 1. It follows that (U ) = 1 (U ).
Now let K be a compact set. Then 1 (K) < . Since 1 is outer regular,
there exists an open set U X such that K U and 1 (U ) < . Since
and 1 agree on open sets it follows that
(K) = (U ) (U \ K) = 1 (U ) 1 (U \ K) = 1 (K).
This proves Step 2.
Step 3. Let be as in Step 2. Then
0 (B) (B) 1 (B) for all B B. (3.9)
Fix a Borel set B B. Then, by Step 2,

0 (B) = sup {1 (K) | K B and K is compact}
= sup {(K) | K B and K is compact}
(B)
inf {(U ) | B U X and U is open}
= inf {1 (U ) | B U X and U is open}
= 1 (B).
This proves Step 3.
R : B R [0, ] be a Borel measure that satisfies (3.9). Then

RStep 4. Let
X
f d = X
f d0 = X f d1 for every continuous function f : X R with
compact support.
It follows from the definition of the integral and part (iii) that
Z Z Z Z
f d0 f d f d1 = f d0
X X X X
for
R every compactly
R supported
R continuous function f : X [0, ). Hence
X
f d = X
f d 0 = X
f d 1 for every compactly supported continuous
function f : X [0, ) and hence also for every compactly supported
continuous function f : X R. This proves Step 4 and Theorem 3.8.
Example 3.9. Let (X, U) be the compact Hausdorff space in Example 3.6
and let : B [0, ] be Dieudonnes measure.
(i) Take 1 := and define the function 0 : B [0, ] by (3.5). Then
0 (X) = 1, 0 ({}) = 0, 0 (X \ {}) = 0,
and so 0 is not a measure. Hence the assumptions on 1 cannot be removed
in part (i) of Theorem 3.8.
(ii) Take 1 := to be the Dirac measure at the point X. This is a
regular Borel measure and so the measure 0 in (3.5) agrees with 1 . It is an
easy exercise to show that the integral of a continuous function f : X R
with respect to the Dieudonne measure is given by
Z Z Z
f d = f () = f d0 = f d1 .
X X X
Moreover, the compact set K = {} satisfies (K) = 0 < 1 = 1 (K) and

the open set U := X \ {} satisfies 1 (U ) = 0 < 1 = (U ). This shows
that the inequalities in Step 1 in the proof of Theorem 3.8 can be strict and
that the hypothesis that is inner regular on open sets cannot be removed
in part (iv) of Theorem 3.8.
Remark 3.10. As Example 3.6 shows, it may sometimes be convenient to
define a Borel measure first on a -algebra that contains the -algebra of
all Borel measurable sets and then restrict it to B. Thus let A 2X be
a -algebra containing B and let : A [0, ] be a measure. Call
outer regular if it satisfies (3.1) for all B A, call it inner regular if it
satisfies (3.2) for all B A, and call it regular if it is both outer and inner
regular. If is regular and (X, B , ) denotes the completion of (X, B, |B ),
it turns out that the completion is also regular (exercise). If in addition
(X, A, ) is -finite (see Definition 4.29 below) then
A B , = |A . (3.10)
To see this, let A A such that (A) < . Choose a sequence of compact
sets Ki X and a sequence of open sets Ui X such that Ki A SUi and
(A) 2i T (Ki ) (Ui ) (A) + 2i for all i N. Then B0 :=
i=1 Ki

and B1 := i=1 Ui are Borel sets such that B0 A B1 and (B1 \ B0 ) = 0.
Thus every set A A with (A) < belongs to B and (A) = (A).
This proves (3.10) because every A-measurable set is a countable union of
A-measurable sets with finite measure. Note that if X is -compact and
(K) < for every compact set K X then (X, A, ) is -finite.
3.2 Borel Outer Measures

This section is of preparatory nature. It discusses outer measures on a locally
compact Hausdorff space that satisfy suitable regularity properties and shows
that the resulting measures on the Borel -algebra are outer/inner regular.
The result will play a central role in the proof of the Riesz Representation
Theorem. As in Section 3.1 we assume that (X, U) is a locally compact
Hausdorff space and denote by B the Borel -algebra of (X, U).
Definition 3.11. A Borel outer measure on X is an outer measure
: 2X [0, ] that satisfies the following axioms.
(a) If K X is compact then (K) < .
(b) If K0 , K1 X are disjoint compact sets then (K0 K1 ) = (K0 )+(K1 ).
(c) (A) = inf {(U ) | A U X, U is open} for every subset A X.
(d) (U ) = sup {(K) | K U, K is compact} for every open set U X.
Theorem 3.12. Let : 2X [0, ] be a Borel outer measure. Then |B is
an outer regular Borel measure and is inner regular on open sets.
One can deduce Theorem 3.12 from Caratheodorys Theorem 2.4 and use
axioms (a) and (b) (instead of the Caratheodory Criterion in Theorem 2.5) to
show that the -algebra of -measurable sets contains the Borel -algebra.
That the resulting Borel measure has the required regularity properties is
then obvious from axioms (c) and (d). We choose a different route, following
Rudin [17], and give a direct proof of Theorem 3.12 which does not rely
on Theorem 2.4. The former approach is left to the reader as well as the
verification that both proofs give rise to the the same -algebra, i.e. the -
algebra A in (3.11) agrees with the -algebra of -measurable subsets of X.
Proof of Theorem 3.12. Define
n o
Ae := E X (E) = sup {(K) | K E, K is compact} < ,

n o (3.11)
A := A X A K Ae for every compact set K X .

Here the subscript e stands for endlich and indicates that the elements
of Ae have finite measure. We prove in seven steps that A is a -algebra
containing B, that := |A : A [0, ] is a measure, and that (X, A, ) is
a complete measure space. That is outer regular and is inner regular on
open sets follows immediately from conditions (c) and (d) in Definition 3.11.
3.2. BOREL OUTER MEASURES 93
Step 1. LetS E1 , E2 , E3 , . . . be a sequence of pairwise disjoint sets in Ae and

define E := i=1 Ei . Then the following holds.
P
(i) (E) = i=1 (Ei ).
(ii) If (E) < then E Ae .
P
The assertions are obvious when (E) = because (E) i=1 (Ei ).
Hence assume (E) < . We argue as in the proof of Theorem 3.8. Fix a
constant > 0. Since Ei Ae for all i there is a sequence of compact sets
Ki Ei such that (Ki ) > (Ei ) 2i for all i. Then for all n N
(E) (K1 Kn )
= (K1 ) + + (Kn ) (3.12)
(E1 ) + + (En )
Here the equality follows from condition (b) in Definition 3.11. Take the
limit n to obtain
X
(Ei ) (E) + .
i=1
Since this holds for all > 0 it follows that

X
X
(Ei ) (E) (Ei )
i=1 i=1
and hence
X
(Ei ) = (E). (3.13)
i=1
Now it follows from (3.12) and (3.13) that
n
X
X
(E) (K1 Kn ) (Ei ) = (E) (Ei )
i=1 i=n+1
for all n N. By (3.13) there exists an n N such that

P
i=n +1 (Ei ) < .
Hence the compact set K := K1 Kn E satisfies
(E) (K ) (E) 2.
Since this holds for all > 0 we obtain
(E) = sup {(K) | K E, K is compact}
and hence E Ae . This proves Step 1.
Step 2. If E0 , E1 Ae then E0 E1 Ae , E0 E1 Ae , and E0 \ E1 Ae .

We first prove that E0 \ E1 Ae . Fix a constant > 0. Since E0 , E1 Ae ,
and by condition (c) in Definition 3.11, there exist compact sets K0 , K1 X
and open sets U0 , U1 X such that
Ki Ei Ui , (Ei ) < (Ki ) (Ui ) < (Ei ) + , i = 0, 1.
Moreover, every compact set with finite outer measure is an element of Ae by

definition and every open set with finite outer measure is an element of Ae
by condition (d) in Definition 3.11. Hence
Ki , Ui , Ui \ Ki Ae
for i = 0, 1 and it follows from Step 1 that
(Ei \ Ki ) (Ui \ Ki ) = (Ui ) (Ki ) 2,

(3.14)
(Ui \ Ei ) (Ui \ Ki ) = (Ui ) (Ki ) 2
for i = 0, 1. Define
K := K0 \ U1 E0 \ E1 . (3.15)
Then K is a compact set and
E0 \ E1 (E0 \ K0 ) (K0 \ U1 ) (U1 \ E1 ).
By definition of an outer measure this implies
(E0 \ E1 ) (E0 \ K0 ) + (K0 \ U1 ) + (U1 \ E1 ) (K) + 4.
Here the last inequality follows from the definition of K in (3.15) and the
inequalities in (3.14). Since > 0 was chosen arbitrarily it follows that
(E0 \ E1 ) = sup {(K) | K E0 \ E1 , K is compact}
and hence E0 \ E1 Ae . With this understood it follows from Step 1 that
E0 E1 = (E0 \ E1 ) E1 Ae , E0 E1 = E0 \ (E0 \ E1 ) Ae .
This proves Step 2.

3.2. BOREL OUTER MEASURES 95
Step 3. A is a -algebra.
First, X A because K Ae for every compact set K X.
Second, assume A A and let K X be a compact set. Then by
definition A K Ae . Moreover K Ae and hence, by Step 2,
Ac K = K \ (A K) Ae .
Since this holds for every compact set K X we have Ac Ae .

Third, let Ai A for i N and denote

[
A := Ai .
i=1
Fix a compact set K X. Then
Ai K Ae
for all i by definition of A. Hence, by Step 2
Bi := Ai K Ae
for all i and hence, again by Step 2
Ei := Bi \ (B1 Bi1 ) Ae
for all i. The sets Ei are pairwise disjoint and

[
[
Ei = Bi = A K.
i=1 i=1
Since (A K) (K) < by condition (a) in Definition 3.11, it follows

from Step 1 that A K Ae . This holds for every compact set K X and
hence A A. This proves Step 3.
Step 4. B A.
Let F X be closed. If K X is compact then F K is a closed subset
of a compact set and hence is compact (see Lemma A.2). Thus F K Ae
for every compact subset K X and so F A. Thus we have proved that
A contains all closed subsets of X. Since A is a -algebra by Step 3, it also
contains all open subsets of X and thus B A. This proves Step 4.
Step 5. Let A X. Then A Ae if and only if A A and (A) < .

If A Ae then A K Ae for every compact set K X by Step 2 and
hence A A. Conversely, let A A such that (A) < . Fix a constant
> 0. By condition (c) in Definition 3.11, there exists an open set U X
such that A U and (U ) < . By condition (d) in Definition 3.11, there
exists a compact set K X such that
K U, (K) > (U ) .
Since K, U Ae and U = (U \ K) K it follows from Step 1 that
(U \ K) = (U ) (K) < .
Moreover, A K Ae because A A. Hence it follows from the definition
of Ae that there exists a compact set H A K such that
(H) (A K)
= (A \ (A \ K))
(A) (A \ K)
(A) (U \ K)
(A) 2.
Since > 0 was chosen arbitrarily it follows that
(A) = sup {(K) | K A, K is compact}
and hence A Ae . This proves Step 5.
Step 6. := |A is an outer regular extended Borel measure and is inner
regular on open sets.
We prove that is a measure. By definition () = 0. Now let S Ai A be
a sequence of pairwise disjoint measurable sets and define A := Pi=1 Ai . If
(Ai ) < for all i then Ai Ae by Step 5 and hence (A) = i=1 (Ai )
by Step 1. If (Ai ) = for some i then (A) (Ai ) and so (A) = .
Thus is a measure. Moreover, B A by Step 4, (K) < for every
compact set K X by condition (a) in Definition 3.11, is outer regular
by condition (c) in Definition 3.11, and is inner regular on open sets by
condition (d) in Definition 3.11. This proves Step 6.
Step 7. (X, A, ) is a complete measure space.
If E X satisfies (E) = 0 then E Ae by definition of Ae and hence
E A by Step 5. This proves Step 7 and Theorem 3.12.
3.3. THE RIESZ REPRESENTATION THEOREM 97
3.3 The Riesz Representation Theorem

Let (X, U) be a locally compact Hausdorff space and B be its Borel -algebra.
A function f : X R is called compactly supported if its support

supp(f ) := x X f (x) 6= 0
is a compact subset of X. The set of compactly supported continuous func-
tions on X will be denoted by

f is continuous and
Cc (X) := f : X R .
supp(f ) is a compact subset of X
Thus a continuous function f : X R belongs to Cc (X) if and only if there
exists a compact set K X such that f (x) = 0 for all x X \ K. The set
Cc (X) is a real vector space.
Definition 3.13. A linear functional : Cc (X) R is called positive if
f 0 = (f ) 0
for all f Cc (X).
The next lemma shows that every positive linear functional on Cc (X)
is continuous with respect to the topology of uniform convergence when
restricted to the subspace of functions with support contained in a fixed
compact subset of X.
Lemma 3.14. Let : Cc (X) R be a positive linear functional and let
fi Cc (X) be a sequence of compactly supported continuous functions that
converges uniformly to f Cc (X). If there exists a compact set K X such
that supp(fi ) K for all i N then (f ) = limi (fi ).
Proof. Since fi converges uniformly to f the sequence
i := sup|fi (x) f (x)|
xX
converges to zero. By Urysohns Lemma A.1 there exists a compactly sup-

ported continuous function : X [0, 1] such that (x) = 1 for all x K.
This function satisfies i fi f i for all i. Hence
i () (fi ) (f ) i (),
because is positive, and hence |(fi ) (f )| i () for all i. Since i
converges to zero so does |(fi ) (f )| and this proves Lemma 3.14.
Let : B [0, ] be a Borel measure. Then every continuous function

f : X R with compact support is integrable with respect to . Define the
map : Cc (X) R by
Z
(f ) := f d. (3.16)
X
Then is a positive linear functional. The Riesz Representation Theorem

asserts that every positive linear functional on Cc (X) has this form. It also
asserts uniqueness under certain regularity hypotheses on the Borel measure.
The following theorem includes two versions of the uniqueness statement.
Theorem 3.15 (Riesz Representation Theorem). Let : Cc (X) R

be a positive linear functional. Then the following holds.
(i) There exists a unique Radon measure 0 : B [0, ] such that 0 = .
(ii) There exists a unique outer regular Borel measure 1 : B [0, ] such
that 1 is inner regular on open sets and 1 = .
(iii) The Borel measures 0 and 1 in (i) and (ii) agree on all compact sets
and on all open sets. Moreover, 0 (B) 1 (B) for all B B.
(iv) Let : B [0, ] be a Borel measure that is inner regular on open
sets. Then = if and only if 0 (B) (B) 1 (B) for all B B.
Proof. The proof has nine steps. Step 1 defines a function : 2X [0, ],
Step 2 shows that it is an outer measure, and Steps 3, 4, and 5 show that it
satisfies the axioms of Definition 3.11. Step 6 defines 1 and Step 7 shows
that 1 = . Step 8 defines 0 and Step 9 proves uniqueness.
Step 1. Define the function U : U [0, ] by

U (U ) := sup (f ) f Cc (X), 0 f 1, supp(f ) U (3.17)
for every open set U X and define : 2X [0, ]
(A) := inf {U (U ) | A U X, U is open} (3.18)
for every subset A X. Then (U ) = U (U ) for every open set U X.

If U, V X are open sets such that U V then U (U ) U (V ) by definition.
Hence (U ) = inf {U (V ) | U V X, V is open} = U (U ) for every open
set U X and this proves Step 1.
Step 2. The function : 2X [0, ] in Step 1 is an outer measure.

By definition () = U () = 0. Since U (U ) U (V ) for all open sets
U, V X with U V , it follows also from the definition that (A) (B)
whenever A B X. Next we prove that for all open sets U, V X
U (U V ) U (U ) + U (V ). (3.19)
To see this, let f Cc (X) such that 0 f 1 and K := supp(f ) U V.
By Theorem A.4 there exist functions , Cc (X) such that
supp() U, supp() V, , 0, + 1, ( + )|K 1.
Hence f = f + f and hence
(f ) = (f + f ) = (f ) + (f ) U (U ) + U (V ).
This proves (3.19).
Now choose a sequence of subsets Ai X and define A :=
S
i=1 Ai . We
must prove that
X
(A) (Ai ). (3.20)
i=1
P an i N such that (Ai ) = then (A) = because Ai A

If there exists
and hence i=1 (Ai ) = = (A). Hence assume (Ai ) < for all i. Fix
a constant > 0. By definition of in (3.18) there exists a sequence of open
sets Ui X such that
Ai Ui , U (Ui ) < (Ai ) + 2i .
Define U :=
S
i=1 Ui . Let f Cc (X) such that 0 f 1 and supp(f ) U .
S compact support, there exists an integer k N such that

Since f has
supp(f ) ki=1 Ui . By definition of U and (3.19) this implies
(f ) U (U1 Uk )
U (U1 ) + + U (Uk )
< (A1 ) + + (Ak ) + .
Hence (f )
P
i=1 (Ai ) + for every f Cc (X) such that 0 f 1 and
supp(f ) U . This implies
X
(A) U (U ) (Ai ) +
i=1
P
by definition of U (U
P) in (3.17). Thus (A) i=1 (Ai ) + for every > 0
and hence (A) i=1 (Ai ). This proves (3.20) and Step 2.
Step 3. Let U X be an open set. Then

U (U ) = sup (K) K U, K is compact . (3.21)
Let f Cc (X) such that
0 f 1, K := supp(f ) U.
Then it follows from the definition of U in (3.17) that (f ) U (V ) for

every open set V X with K V . Hence it follows from the definition of
in (3.18) that
(f ) (K).
Hence

U (U ) = sup (f ) f Cc (X), 0 f 1, supp(f ) U

sup (K) K U, K is compact
(U )
= U (U ).
Hence U (U ) = sup {(K) | K U, K is compact} and this proves Step 3.

Step 4. Let K X be an compact set. Then

(K) = inf (f ) f Cc (X), f 0, f |K 1 . (3.22)
In particular, (K) < .

Define
a := inf (f ) f Cc (X), f 0, f |K 1 .
We prove that a (K). Let U X be any open set containing K. By
Urysohns Lemma A.1 there exists a function f Cc (X) such that
0 f 1, supp(f ) U, f |K 1.
Hence
a (f ) U (U ).
This shows that a U (U ) for every open set U X containing K. Take
the infimum over all open sets containing K and use the definition of in
equation (3.18) to obtain a (K).
We prove that (K) a. Choose a function f Cc (X) such that f 0

and f (x) = 1 for all x K. Fix a constant 0 < < 1 and define
U := {x X | f (x) > } .
Then U is open and K U . Hence
(K) U (U ).
Moreover, every function g Cc (X) with 0 g 1 and supp(g) U

satisfies g(x) f (x) for x U , hence g f , and so (g) (f ).
Take the supremum over all such g to obtain U (U ) (f ) and hence
1
(K) U (U ) (f ).

This shows that (K) 1 (f ) for all (0, 1) and hence
(K) (f ).
Since this holds for every function f Cc (X) with f 0 and f |K 1 it

follows that (K) a. This proves Step 4.
Step 5. Let K0 , K1 X be compact sets such that K0 K1 = . Then
(K0 K1 ) = (K0 ) + (K1 ).
The inequality (K0 K1 ) (K0 ) + (K1 ) holds because is an outer

measure by Step 2. To prove the converse inequality choose f Cc (X) such
that
0 f 1, f |K0 0, f |K1 1.
That such a function exists follows from Urysohns Lemma A.1 with K := K1
and U := X \ K0 . Now fix a constant > 0. Then it follows from Step 4
that there exists a function g Cc (X) such that
g 0, g|K0 K1 1, (g) < (K0 K1 ) + .
It follows also from Step 4 that
(K0 ) + (K1 ) ((1 f )g) + (f g) = (g) < (K0 + K1 ) + .
Hence (K0 ) + (K1 ) < (K0 + K1 ) + for every > 0 and therefore
(K0 ) + (K1 ) (K0 + K1 ). This proves Step 5.
Step 6. The function 1 := |B : B [0, ] is an outer regular Borel

measure that is inner regular on open sets.
The function is an outer measure by Step 2. It satisfies condition (a) in
Definition 3.11 by Step 4, it satisfies condition (b) by Step 5, it satisfies
condition (c) by Step 1, and it satisfies condition (d) by Step 3. Hence is
a Borel outer measure. Hence Step 6 follows from Theorem 3.12.
Step 7. Let 1 be as in Step 6. Then 1 = .
We will prove that Z
(f ) f d1 (3.23)
X
for all f Cc (X). Once this is understood, it follows that

Z Z
(f ) = (f ) (f ) d1 = f d1
X X
R R
and hence X f d1 (f ) for all f Cc (X). Thus (f ) = X f d1 for all
f Cc (X), and this proves Step 7.
Thus it remains to prove the inequality (3.23). Fix a continuous function
f : X R with compact support and denote
K := supp(f ), a := inf f (x), b := sup f (x).

xX xX
Fix a constant > 0 and choose real numbers
y0 < a < y1 < y2 < < yn1 < yn = b
such that
yi yi1 < , i = 1, . . . , n.
For i = 1, . . . , n define

Ei := x K yi1 < f (x) yi .
Then Ei is the intersection of the open set f 1 ((yi1 , )) with the closed set
f 1 ((, yi ]) and hence is a Borel set. Moreover Ei Ej = for i 6= j and
n
[
K= Ei .
i=1
Since 1 is outer regular there exist open sets U1 , . . . , Un X such that

Ei Ui , 1 (Ui ) < 1 (Ei ) + , sup f < yi + (3.24)
n Ui
for all i. (For each i, choose first an open set that satisfies the first two
conditions in (3.24) and then intersect it with the open set f 1 ((, yi +)).)
By Theorem A.4 there exist functions 1 , . . . , n Cc (X) such that
n
X n
X
i 0, supp(i ) Ui , i 1, i |K 1. (3.25)
i=1 i=1
It follows from (3.24), (3.25), and Step 4 that

n
X
f= i f, i f (yi + )i ,
i=1
n
X
1 (K) (i ), (i ) 1 (Ui ) < 1 (Ei ) + .
i=1
n
Hence
n
X
(f ) = (i f )
i=1
Xn
(yi + )(i )
i=1
n
X n
X

= yi + |a| + (i ) |a| (i )
i=1 i=1
n
X
yi + |a| + 1 (Ei ) + |a|1 (K)
i=1
n
n n
X X
= yi + 1 (Ei ) + yi + |a| +
i=1
n i=1
n
X
yi 1 (Ei ) + 21 (K) + b + |a| +
Zi=1

f d1 + 21 (K) + b + |a| + .
X
Here we have used the inequality yi + |a|R + 0. Since > 0 can be chosen
arbitrarily small it follows that (f ) X f d1 . This proves (3.23).
Step 8. Define 0 : B [0, ] by

0 (B) := sup (K) K B, K is compact
Then 0 is a Radon measure, 0 = , and 0 and 1 satisfy (iii) and (iv).
It follows from Step 6 and part (i) of Theorem 3.8 that 0 is a Radon measure
and it follows from Step 7 and part (iii) of Theorem 3.8 that 0 = 1 = .
That the measures 0 and 1 satisfy assertions (iii) and (iv) follows from
parts (i) and (iv) of Theorem 3.8.
Step 9. We prove uniqueness in (i) and (ii).
By definition 0 (K) = (K) = 1 (K) for every compact set K X. Second,
it follows from and Steps 1 and 3 that 0 (U ) = U (U ) = (U ) = 1 (U ) for
every open set U X. Third, Steps 7 and 8 assert that 0 = 1 = .
Hence it follows from part (iv) of Theorem 3.8 that every Borel measure
: B [0, ] that is inner regular on open sets and satisfies = agrees
with on all compact sets and on all open sets. Hence every Radon measure
: B [0, ] that satisfies = is given by

(B) = sup (K) K B, K is compact = 0 (B)
for every B B. Likewise, every outer regular Borel measure : B [0, ]
that is inner regular on open sets and satisfies = is given by

(B) = inf (U ) B U X, U is open = (B) = 1 (B)
for every B B. This proves Step 9 and Theorem 3.15.
The following corollary is the converse of Theorem 3.8.
Corollary 3.16. Let 0 : B [0, ] be a Radon measure and define
1 (B) := inf {0 (U ) | B U X, U is open} for all B B. (3.26)
Then 1 is an outer regular Borel measure, is inner regular on open sets, and
0 (B) = sup {1 (K) | K B, K is compact} for all B B. (3.27)
Proof. Let 1 be the unique outer regular Borel measure on X that is inner
regular on open sets and satisfies 1 = 0 . Then Theorem 3.15 asserts
that 0 and 1 agree on all compact sets and all open sets. Since 1 is outer
regular, it follows that 1 is given by (3.26). Since 0 is inner regular it
follows that 0 satisfies (3.27). This proves Corollary 3.16.
Corollary 3.17. Every Radon measure is outer regular on compact sets.

Proof. Equation (3.26) with B = K compact and 0 (K) = 1 (K).
The next theorem formulates a condition on a locally compact Hausdorff
space which guarantees that all Borel measures are regular. The condition
(every open subset is -compact) is shown below to be strictly weaker than
second countability.
Theorem 3.18. Let X be a locally compact Hausdorff space.
(i) Assume X is -compact. Then every Borel measure on X that is inner
regular on open sets is regular.
(ii) Assume every open subset of X is -compact. Then every Borel measure
on X is regular.
Proof. We prove (i). Let : B [0, ] be a Borel measure that is inner reg-
ular on open sets and let 0 , 1 : B [0, ] be the Borel measures associated
to := in parts (i) and (ii) of the Riesz Representation Theorem 3.15.
Since is inner regular on open sets, it follows from part (iii) of Theorem 3.15
that 0 (B) (B) 1 (B) for all B B. Since X is -compact, it follows
from part (ii) of Theorem 3.8 that 0 = = 1 . Hence is regular.
We prove (ii). Let : B [0, ] be a Borel measure. We prove that
is inner regular on open sets. Fix an open set U X. Since U is -compact,
there exists a sequence of compact sets Ki U such that Ki Ki+1 for all
i N and U =
S
i=1 K i Hence (U ) = limi (Ki ) by Theorem 1.28, so
.
(U ) = sup {(K) | K U and K is compact} .
This shows that is inner regular on open sets and hence it follows from (i)
that is regular. This proves Theorem 3.18.
Example 3.9 shows that the assumption that every open set is -compact
cannot be removed in part (ii) of Theorem 3.18 even if X is compact. Note
also that Theorem 3.18 provides another proof of regularity for the Lebesgue
measure, which was established in Theorem 2.13.
Corollary 3.19. Let X be a locally compact Hausdorff space such that every
open subset of X is -compact. Then for every positive linear functional
: Cc (X) R there exists a unique Borel measure such that = .
Proof. This follows from Theorem 3.15 and part (ii) of Theorem 3.18.
Remark 3.20. Let X be a compact Hausdorff space and let C(X) = Cc (X)
be the space of continuous real valued functions on X. From a functional
analytic viewpoint it is interesting to understand the dual space of C(X),
i.e. the space of all bounded linear functionals on C(X) (Definition 4.23).
Exercise 5.35 below shows that every bounded linear functional on C(X) is
the difference of two positive linear functionals. If every open subset of X
is -compact it then follows from Corollary 3.19 that every bounded linear
functional on C(X) can be represented uniquely by a signed Borel measure.
(See Definition 5.10 in Section 5.3 below.)
An important class of locally compact Hausdorff spaces that satisfy the

hypotheses of Theorem 3.18 and Corollary 3.19 are the second countable
ones. Here are the definitions. A basis of a topological space (X, U) is a
collection V U of open sets such that every open set U X is a union of
elements of V. A topological space (X, U) is called second countable if it
admits a countable basis. It is called first countable if, for every x X,
there is a sequence of open sets Wi , i N, such that x Wi for all i and
every open set that contains x contains one of the sets Wi .
Lemma 3.21. Let X be a locally compact Hausdorff space.

(i) If X is second countable then every open subset of X is -compact.
(ii) If every open subset of X is -compact then X is first countable.
Proof. We prove (i). Let V be a countable basis of the topology and let
U X be an open set. Denote by V(U ) the collection of all sets V V such
that V U and V is compact. Let x U . By Lemma A.3 there is an open
set W X with compact closure such that x W W U . Since V is a
basis of the topology, there is an element V V such that x V W . Hence
V is a closed subset of the compact set W and so is compact by Lemma A.2.
Thus V V(U ) and x V . This shows that
[
U= V.
V V(U )
Since V is countable so is V(U ). Choose a bijection N V(U ) : i 7 Vi and

define
Ki := V 1 V i
for i N. Then Ki Ki+1 for all i and U =
S
i=1 Ki . Hence U is -compact.
We prove (ii). Fix an element x X. Since X is a Hausdorff space,

the set X \ {x} is open and hence is -compact by assumption. Choose a
sequence
S of compact sets Ki X \ {x} such that Ki Ki+1 for all i N and
i=1 K i = X \ {x}. Then each set Ui := X \ Ki is open and contains x. By
Lemma A.3 there exists a sequence of open sets Vi X with compact closure
such that x Vi V i TUi = X \ Ki for all i N. Define WTi := V1 Vi
for i N. Then W i j=1 (X \ Kj ) = X \ Ki and hence
i
i=1 W i = {x}.
This implies that each open set U X that contains x also contains one of
S
the sets W i . Namely, if x

U and U is open, then W 1 \ U is a compact set
contained in X \ {x} = i=1 (X \ W i ), hence there exists a j N such that
W 1 \ U ji=1 (X \ W i ) = X \ W j , and so W j U . Thus the sets Wj form
S
a countable neighborhood basis of x and this proves Lemma 3.21.
Example 3.22. The Alexandrov Double Arrow Space is an example
of a compact Hausdorff space in which every open subset is -compact and
which is not second countable. It is defined as the ordered space (X, ),
where X := [0, 1] {0, 1} and denotes the lexicographic ordering

s < t or
(s, i) (t, j)
s = t and i = 0 and j = 1.
The topology U 2X is defined as the smallest topology containing the sets
Sa := {x X | a x} , Pb := {x X | x b} , a, b X.
It has a basis consisting of the sets Sa , Pb , Sa Pb for all a, b X.
This topological space (X, U) is a compact Hausdorff space and is per-
fectly normal, i.e. for any two disjoint closed subsets F0 , F1 X there
exists a continuous function f : X [0, 1] such that
F0 = f 1 (0), F1 = f 1 (1).
(For a proof see Dan Mas Topology Blog [12].) This implies that every open
subset of X is -compact. Moreover, the subsets
Y0 := (0, 1) {0}, Y1 := (0, 1) {1}
are both homeomorphic to the Sorgenfrey line, defined as the real axis
with the (nonstandard) topology in which the open sets are the unions of
half open intervals [a, b). Since the Sorgenfrey line is not second countable
neither is the double arrow space (X, U). (The Sorgenfrey line is Hausdorff
and perfectly normal, but is not locally compact because every compact
subset of it is countable.)
3.4 Exercises
Exercise 3.23. This exercise shows that the measures 0 , 1 in Theorem 3.15
need not agree. Let (X, d) be the metric space given by X := R2 and

0, if x1 = x2 ,
d((x1 , y1 ), (x2 , y2 )) := |y1 y2 | +
1, if x1 6= x2 .
Let B 2X be the Borel -algebra of (X, d).
(i) Show that (X, d) is locally compact.
(ii) Show that for every compactly supported continuous function f : X R
there exists a finite set Sf R such that supp(f ) Sf R.
(iii) Define the positive linear functional : Cc (X) R by
XZ
(f ) := f (x, y) dy.
xSf
(Here the integrals on the right are understood as the Riemann integrals or,
equivalently by Theorem 2.24, as the Lebesgue integrals.) Let : B [0, ]
be a Borel measure such that
Z
f d = (f ) for all f Cc (X).
X
Prove that every one-element subset of X has measure zero.

(iv) Let be as in (iii) and let E := R {0}. This set is closed. If is inner
regular prove that (E) = 0. If is outer regular, prove that (E) = .
Exercise 3.24. This exercise shows that the Borel assumption cannot be
removed in Theorem 3.18. (The measure in part (ii) is not a Borel measure.)
Let (X, U) be the topological space defined by X := N {} and

U := U X U N or #U c < .

Thus (X, U) is the (Alexandrov) one-point compactification of the set N

of natural numbers with the discrete topology. (If U then the condition
#U c < is equivalent to the assertion that U c is compact.)
(i) Prove that (X, U) is a compact Hausdorff space and that every subset
of X is -compact. Prove that the Borel -algebra of X is B = 2X .
(ii) Let : 2X [0, ] be the counting measure. Prove that is inner
regular, but not outer regular.
3.4. EXERCISES 109
Exercise 3.25. Let (X, UX ) and (Y, UY ) be locally compact Hausdorff spaces
and denote their Borel -algebras by BX 2X and BY 2Y . Let : X Y
be a continuous map and let X : BX [0, ] be a measure.
(i) Prove that BY BX (See Exercise 1.69).
(ii) If X is inner regular show that X |BY is inner regular.
(iii) Find an example where X is outer regular and X |BY is not outer
regular. Hint: Consider the inclusion of N into its one-point compactification
and use Exercise 3.24. (In this example X is a Borel measure, however, X
is not a Borel measure.)
Exercise 3.26. Let (X, d) be a metric space. Prove that (X, d) is perfectly
normal, i.e. if F0 , F1 X are disjoint closed subsets then there is a continuous
function f : X [0, 1] such that F0 = f 1 (0) and F1 = f 1 (1). Compare
this with Urysohns Lemma A.1. Hint: An explicit formula for f is given by
d(x, F0 )
f (x) := ,
d(x, F0 ) + d(x, F1 )
where
d(x, F ) := inf d(x, y)
yF
for x X and F X.
Exercise 3.27. Recall that the Sorgenfrey line is the topological space
(R, U), where U 2R is the smallest topology that contains all half open
intervals [a, b) with a < b. Prove that the Borel -algebra of (R, U) agrees
with the Borel -algebra of the standard topology on R.
Exercise 3.28. Recall from Example 3.22 that the Double Arrow Space is
X := [0, 1] {0, 1}
with the topology induced by the lexicographic ordering. Prove that B X
is a Borel set for this topology if and only if there is a Borel set E [0, 1]
and two countable sets F, G X such that
B = ((E {0, 1}) F ) \ G. (3.28)
Hint 1: Show that the projection f : X [0, 1] onto the first factor is
continuous with respect to the standard topology on the unit interval.
Hint 2: Denote by B 2X the set of all sets of the form (3.28) with E [0, 1]
a Borel set and F, G X countable. Prove that B is a -algebra.
Exercise 3.29 (The Baire -algebra).

Let (X, U) be a locally compact Hausdorff space and define

K is compact and there is a sequence of open sets

Ka := K X .
Ui such that Ui+1 Ui for all i and K =
T
i=1 Ui
Let
Ba 2X
be the smallest -algebra that contains Ka . It is contained in the Borel -
algebra B 2X and is called the Baire -algebra of (X, U). The elements of
Ba are called Baire sets. A function f : X R is called Baire measurable
if f 1 (U ) Ba for every open set U R. A Baire measure is a measure
: Ba [0, ] such that (K) < for all K Ka .
(i) Let f : X R be a continuous function with compact support. Prove
that f 1 (c) Ka for every nonzero real number c.
(ii) Prove that Ba is the smallest -algebra such that every continuous func-
tion f : X R with compact support is Ba -measurable.
(iii) If every open subset of X is -compact prove that Ba = B. Hint:
Show first that every compact set belongs to Ka and then that every open
set belongs to Ba .
Exercise 3.30. (i) Let X be an uncountable set and let U := 2X be the
discrete topology. Prove that B X is a Baire set if and only if B is
countable or has a countable complement. Define : Ba [0, 1] by

0, if B is countable,
(B) :=
1, if B c is countable.
R
Show that X f d = 0 for every f Cc (X). Thus positive linear functionals
: Cc (X) R need not be uniquely represented by Baire measures.
(ii) Let X be the compact Hausdorff space of Example 3.6. Prove that the
Baire sets in X are the countable subsets of X \ {} and their complements.

(iii) Let X be the StoneCech compactification of N in Example 4.60 below.
Prove that the Baire sets in X are the subsets of N and their complements.
(iv) Let X = R2 be the locally compact Hausdorff space in Exercise 3.23
(with a nonstandard topology). Show that B X is a Baire set if and only if
the set Bx := {y R | (x, y) B} is a Borel set in R for every x R and one
of the sets S0 := {x R | Bx 6= } and S1 := {x R | Bx 6= R} is countable.
3.4. EXERCISES 111
Exercise 3.31. Let (X, U) be a locally compact Hausdorff space and let
Ba B 2X
be the Baire and Borel -algebras. Let F (X) denote the real vector space of
all functions f : X R. For F F (X) consider the following conditions.
(a) Cc (X) F.
(b) If fi F is a sequence converging pointwise to f F (X) then f F.
Let Fa F (X) be the intersection of all subsets F F (X) that satisfy
conditions (a) and (b). Prove the following.
(i) Fa satisfies (a) and (b).
(ii) Every element of Fa is Baire measurable. Hint: The set of Baire mea-
surable functions on X satisfies (a) and (b).
(iii) If f Fa and g Cc (X) then f + g Fa . Hint: Let g Cc (X). Then
the set Fa g satisfy (a) and (b) and hence contains Fa .
(iv) If f, g Fa then f + g Fa . Hint: Let g Fa . Then the set Fa g
satisfy (a) and (b) and hence contains Fa .
(v) If f Fa and c R then cf Fa . Hint: Fix a real number c 6= 0.
Then the set c1 Fa satisfy (a) and (b) and hence contains Fa .
(vi) If f Fa and g Cc (X) then f g Fa . Hint: Fix a real number c such
that c + g(x) > 0 for all x R. Then the set (c + g)1 Fa satisfy (a) and (b)
and hence contains Fa . Now use (iv) and (v).
(vii) If A X such that A Fa and f Fa then f A Fa . Hint: The
set (1 + A )1 Fa satisfy (a) and (b) and hence contains Fa .
(viii) The set

A := A X | A Fa or X\A Fa
is a -algebra. Hint: If A , B FA then AB = A + B A B Fa . If
X\A , X\B FA then X\(AB) = X\A X\B Fa . If A , X\B FA then
X\(AB) = (X\A)(X\B) = X\B A X\B Fa . Thus
A, B A = A B A.
(ix) A = Ba . Hint: Let K Ka . Use Urysohns Lemma A.1 to construct a
sequence gi Cc (X) that converges pointwise to K .
(x) For every f Fa there exists a sequence
S of compact sets Ki Ka such
that Ki Ki+1 for all i and supp(f ) iN Ki . Hint: The set of functions
f : X R with this property satisfies conditions (a) and (b).
Exercise 3.32. Show that, for every locally compact Hausdorff space X and
any two Borel measures 0 , 1 as in Theorem 3.8, there is a Baire set N X
such that 0 (N ) = 0 and 0 (B) = 1 (B) for every Baire set B X \ N .
Hint 1: Show first that

0 (B) = sup 0 (K) K Ka , K B , (3.29)
where Ka is as in Exercise 3.29. To see this, prove that the right hand side
of equation (3.29) defines a Borel measure on X that is inner regular on
open sets and satisfies 0 and = 0 .
Hint 2: Suppose there exists a Baire set N X such that 0 (N ) < 1 (N ).
Show that 1 (N ) = and that N can be chosen such that 0 (N ) = 0. Next
show that X\N Fa , where Fa is as in Exercise 3.31, and deduce that X \N
is contained in a countable union of compact sets.
Example 3.33. Let X be the StoneCech compactification of N discussed
in Example 4.60 below and denote by Ba B 2X the Baire and Borel -
algebras. Thus B X is a Baire set if and only if either B N or X \B N.
(See part (iii) of Exercise 3.30.) For a Borel set B X define

X1 B U X,
0 (B) := , 1 (B) := inf 0 (U )
.
n U is open
nB
As in Example 4.60 denote by X0 X the union of all open sets U X

with 0 (U ) < . Then the restriction of 0 to X0 is a Radon measure, the
restriction of 1 to X0 is outer regular and is inner regular on open sets, and
0 is given by (3.5) as in Theorem 3.8. Moreover, X0 \ N is a Baire set in X0
and 0 (X0 \ N) = 0 while 1 (X0 \ N) = . Thus we can choose N := X0 \ N
in Exercise 3.32 and 0 and 1 do not agree on the Baire -algebra.
Example 3.34. Let X = R2 be the locally compact Hausdorff space in
Example 3.23 and let 0 , 1 be the Borel measures of Theorem 3.15 associated
to the linear functional : Cc (X) R in that example. Then it follows from
part (iv) of Exercise 3.30 that 0 (B) = 1 (B) for every Baire set B X.
Thus we can choose N = in Exercise 3.32. However, there does not exist
any Borel set N X such that 0 (N ) = 0 and 0 agrees with 1 on all Borel
subsets of X \ N . (A set N X is a Borel set with 0 (N ) = 0 if and only if
Nx := {y R | (x, y) N } is a Borel set and m(Nx ) = 0 for all x R.)
Exercise 3.35. Let Z be the disjoint union of the locally compact Hausdorff
spaces X0 in Example 3.33 and X = R2 in Example 3.34. Find Baire sets
B0 X0 and B X whose (disjoint) union is not a Baire set in Z.
Chapter 4
Lp Spaces
This chapter discusses the Banach space Lp () associated to a measure space

(X, A, ) and a number 1 p . Section 4.1 introduces the inequalities
of Holder and Minkowski and Section 4.2 shows that Lp () is complete. In
Section 4.3 we prove that, when X is a locally compact Hausdorff space, is
a Radon measure, and 1 p < , the subspace of continuous functions with
compact support is dense in Lp (). If in addition X is second countable it
follows that Lp () is separable. When 1 < p < (or p = 1 and the measure
space (X, A, ) is localizable) the dual space of Lp () is isomorphic to Lq (),
where 1/p + 1/q = 1. For p = 2 this follows from elementary Hilbert space
theory and is proved in Section 4.4. For general p the proof requires the
RadonNikod ym theorem and is deferred to Chapter 5. Some preparatory
results are proved in Section 4.5.
4.1 H
older and Minkowski
Assume throughout that (X, A, ) is a measure space and that p, q are real
numbers such that
1 1
+ = 1, 1 < p < , 1 < q < . (4.1)
p q
Then any two nonnegative real numbers a and b satisfy Youngs inequality
1 1
ab ap + bq (4.2)
p q
and equality holds in (4.2) if and only if ap = bq . (Exercise: Prove this by
examining the critical points of the function (0, ) R : x 7 p1 xp xb.)
113
114 CHAPTER 4. LP SPACES
Theorem 4.1. Let f, g : X [0, ] be measurable functions. Then f and

g satisfy the H
older inequality
Z Z 1/p Z 1/q
p q
f g d f d g d (4.3)
X X X
and the Minkowski inequality

Z 1/p Z 1/p Z 1/p
p p p
(f + g) d f d + g d . (4.4)
X X X
Proof. Define
Z 1/p Z 1/q
p q
A := f d , B := g d .
X X
If A = 0 then f = 0 almost R everywhere by Lemma 1.49, hence f g = 0 almost

everywhere, and hence X f g d = 0 by Lemma 1.48. This proves the Holder
inequality (4.3) in the case A = 0. If A = and B > 0 then AB = and
so (4.3) holds trivially. Interchanging A and B if necessary, we find that (4.3)
holds whenever one of the numbers A, B is zero or infinity. Hence assume
0 < A < and 0 < B < . Then it follows from (4.2) that
R
f g d
Z
X f g
= d
AB X AB
Z p
1 f 1 g q
+ d
X p A q B
R p R q
1 X f d 1 X g d
= +
p Ap q Bq
1 1
= +
p q
= 1.
This proves the Holder inequality. To prove the Minkowski inequality, define
Z 1/p Z 1/p Z 1/p
p p p
a := f d , b := g d , c := (f + g) d .
X X X
We must prove that c a + b. This is obvious when a = or b = .

Hence assume a, b < . We first show that c < . This holds because
4.2. THE BANACH SPACE LP () 115
f (f p + g p )1/p and g (f p + g p )1/p , hence f + g 2(f p + g p )1/p , therefore

(f +g)p 2p (f p +g p ), and integrating this inequality and raising the integral
to the power 1/p we obtain c 2(ap + bp )1/p < . With this understood, it
follows from the Holder inequality that
Z Z
p p1
c = f (f + g) d + g(f + g)p1 d
X X
Z 1/p Z 1/q
p (p1)q
f d (f + g) d
X X
Z 1/p Z 1/q
p (p1)q
+ g d (f + g) d
X X
Z 11/p
p
= (a + b) (f + g) d
X
p1
= (a + b)c .
Here we have used the identity pq q = p. It follows that c a + b and this
proves Theorem 4.1.
R R
Exercise 4.2. (i) Assume 0 < X f p d < and 0 < X g q d < . Prove
that equality holds in (4.3) if and only if there exists a constant > 0
such that g q = f p almost everywhere. Hint: Use the proof of the Holder
inequality and the fact that equality holds in (4.2) if and only ap = bq .
R R
(ii) Assume 0 < X f p d < and 0 < X g p d < . Prove that equality
holds in (4.4) if and only if there is a real number > 0 such that g = f
almost everywhere. Hint: Use part (i) and the proof of the Minkowski
inequality.
4.2 The Banach Space Lp()

Definition 4.3. Let (X, A, ) be a measure space and let 1 p < . Let
f : X R be a measurable function. The Lp -norm of f is the number
Z 1/p
p
kf kp := |f | d . (4.5)
X
A function f : X R is called p-integrable or an Lp -function if it is
measurable and kf kp < . The space of Lp -functions is denoted by
n o
p
L () := f : X R f is A-measurable and kf kp < .
(4.6)
It follows from the Minkowski inequality (4.4) that the sum of two Lp -
functions is again an Lp -function and hence Lp () is a real vector space.
Moreover, the function
Lp () [0, ) : f 7 kf kp
satisfies the triangle inequality
kf + gkp kf kp + kgkp
for all f, g Lp () by (4.4) and
kf kp = || kf kp
for all R and f Lp () by definition. However, in general kkp is not
a norm on Lp () because kf kp = 0 if and only if f = 0 almost everywhere
by Lemma 1.49. We can turn the space Lp () into a normed vector space
by identifying two functions f, g Lp () whenever they agree almost every-
where. Thus we introduce the equivalence relation

f g f =g -almost everywhere. (4.7)
Denote the equivalence class of a function f Lp () under this equivalence
relation by [f ] and the quotient space by

Lp () := Lp ()/ . (4.8)
This is again a real vector space. (For p = 1 see Definition 1.51.) The
Lp -norm in (4.5) depends only on the equivalence class of f and so the map
Lp () [0, ) : [f ] 7 kf kp
is well defined. It is a norm on Lp () by Lemma 1.49. Thus we have defined
the normed vector space Lp () for 1 p < . It is sometimes convenient to
abuse notation and write f Lp () instead of [f ] Lp (), always bearing
in mind that then f denotes an equivalence class of p-integrable functions. If
(X, A , ) denotes the completion of (X, A, ) it follows as in Corollary 1.56
that Lp () is naturally isomorphic to Lp ( ).
Remark 4.4. Assume 1 < p < and let f, g Lp () such that
kf + gkp = kf kp + kgkp , kf kp 6= 0.
Then it follows from part (ii) of Exercise 4.2 that there exists a real number
0 such that g = f almost everywhere.
Example 4.5. If (Rn , A, m) is the Lebesgue measure space we write

Lp (Rn ) := Lp (m).
(See Definition 2.2 and Definition 2.11.)
Example 4.6. If : 2N [0, ] is the counting measure we write
`p := Lp ().
Thus the elements of `p are sequences (xn )nN of real numbers such that

!1/p
X
k(xn )kp := |xn |p < .
p=1
R P
If we define f : N R by f (n) := xn for n N then N
|f |p d = p
p=1 |xn | .
For p = there is a similar normed vector space L () defined next.

Definition 4.7. Let (X, A, ) be a measure space and let f : X [0, ]
be a measurable function. The essential supremum of f is the number
ess sup f [0, ] defined by

ess sup f := inf c [0, ] f c almost everywhere (4.9)
A function f : X R is called an L -function if it is measurable and
kf k := ess sup |f | < (4.10)
The set of L -functions on X will be denoted by
L () := f : X R f is measurable and ess sup|f | <

and the quotient space by the equivalence relation (4.7) by

L () := L ()/ . (4.11)
This is a normed vector space with the norm defined by (4.10), which depends
only on the equivalence class of f .
Lemma 4.8. For every f L () there exists a measurable set E A such
that (E) = 0 and supX\E |f | = kf k .
Proof. The set En S := {x X | |f (x)| > kf k + 1/n} has measure zero for
all n. Hence E := nN En is also a set of measure zero and |f (x)| kf k
for all x X \ E. Hence supX\E |f | = kf k . This proves Lemma 4.8.
Theorem 4.9. Lp () is a Banach space for 1 p .

Proof. Assume first that 1 p < . In this case the argument is a refine-
ment of the proof of Theorem 1.52 and Theorem 1.53 for the case p = 1. Let
fn Lp () be a Cauchy sequence with respect to the norm (4.5). Choose a
sequence of positive integers n1 < n2 < n3 < such that
fn fn < 2i

i i+1 p
for all i N. Define

k
X
X
gk := |fni+1 fni |, g := |fni+1 fni | = lim gk .
k
i=1 i=1
Then it follows from Minkowskis inequality (4.4) that

k
X k
X
2i 1

kgk kp fni fni+1 p <

i=1 i=1
for all k N. Moreover, gkp p

for all k N and the sequence of functions
gk+1
gkp : X [0, ] converges pointwise to the integrable function g p . Hence it
follows from the Lebesgue Monotone Convergence Theorem 1.37 that
kgkp = lim kgk kp 1.

k
Hence, by Lemma 1.47, there is a measurable set E A such that
(E) = 0, g(x) < for all x X \ E.
Hence the series

P
i=1 (fni+1 (x) fni (x)) converges absolutely for x X \ E.
Define the function f : X R by

X
f (x) := fn1 (x) + (fni+1 (x) fni (x))
i=1
for x X \ E and by f (x) := 0 for x E. Then the sequence

k1
X
fnk X\E = fn1 X\E + (fni+1 fni )X\E
i=1
converges pointwise to f . Hence f is A-measurable by Theorem 1.24.

We must prove that f Lp () and that limn kf fn kp = 0. To see

this fix a constant > 0. Then there exists an integer n0 N such that
kfn fm kp < for all n, m n0 . By the Lemma of Fatou 1.41 this implies
Z Z
p
|fn f | d = lim inf |fn fnk X\E |p d
X X k Z
lim inf |fn fnk X\E |p d
k
ZX
= lim inf |fn fnk |p d
k X
p

for all n n0 . Hence kfn f kp for all n n0 and hence
kf kp kfn0 kp + kf fn0 kp kfn0 kp + < .
Thus f Lp () and limn kf fn kp = 0 as claimed. This shows that

Lp () is a Banach space for p < .
The proof for p = is simpler. Let fn L () such that the [fn ] form
a Cauchy sequence in L (). Then there is a set E A such that
(E) = 0, kfn k = sup |fn |, kfm fn k = sup |fm fn | (4.12)

X\E X\E
for all m, n N. To see this, use Lemma 4.8 to find null sets En , Em,n A
such that supX\En |fn | = kfn k and supX\Em,n |fm fn | = kfm fn k for
all m, n N. Then the union E of the sets En and Em,n is measurable and
satisfies (4.12). Since [fn ] is a Cauchy sequence in L () we have
lim n = 0, n := sup kfm fn k .

n mn
Since |fm (x) fn (x)| n for all m n and all x X \ E it follows that
(fn (x))nN is Cauchy sequence in R and hence converges for every x X \ E.
Define f : X R by f (x) := limn fn (x) for x X \ E and by f (x) := 0
for x E. Then
kf fn k sup |f (x) fn (x)| = sup lim |fm (x) fn (x)| n

xX\E xX\E m
for all n N. Hence kf k kf1 k + 1 < and limn kf fn k = 0.

Corollary 4.10. Let (X, A, ) be a measure space and let 1 p . Let

f Lp () and let fn Lp () be a sequence such that limn kfn f kp = 0.
If p = then fn converges almost everywhere to f . If p < then there
exists a subsequence fni that converges almost everywhere to f .
Proof. For p = this follows directly from the definitions. For p < choose
a sequence of integers 0 < n1 < n2 < n3 < such that kfni fni+1 kp < 2i
for all i N. Then the proof of Theorem 4.9 shows that fni converges almost
everywhere to an Lp -function g such that limn kfn gkp = 0. Since the
limit is unique in Lp () it follows that g = f almost everywhere.
4.3 Separability
Definition 4.11. Let X be a topological space. A subset S X is called
dense (in X) if its closure is equal to X or, equivalently, U S 6= for every
nonempty open set U X. A subset S X of a metric space is dense if and
only if every element of X is the limit of a sequence in S. The topological
space X is called separable if it admits a countable dense subset.
Every second countable topological space is separable and first countable
(see Lemma 3.21). The Sorgenfrey line is separable and first countable but is
not second countable (see Example 3.22). A metric space is separable if and
only if it is second countable. (If S is a countable dense subset then the balls
with rational radii centered at the points of S form a basis of the topology.)
The Euclidean space X = Rn with its standard topology is separable (Qn is
a countable dense subset) and hence is second countable. The next lemma
gives a criterion for a linear subspace to be dense in Lp ().
Lemma 4.12. Let (X, A, ) be a measure space and let 1 p < . Let X
be a linear subspace of Lp () such that [A ] X for every measurable set
A A with (A) < . Then X is dense in Lp ().
Proof. Let Y denote the closure of X in Lp (). Then Y is a closed linear
subspace of Lp (). We prove in three steps that Y = Lp ().
Step 1. If s Lp () is a measurable step function then [s] Y .
Write s = Rì=1 i Ai where Ri R \ {0} and Ai = s1 (i ) A. Then
P
|i |p (Ai ) = X |i Ai |p d X |s|p d < and hence (Ai ) < for all i.
This implies [Ai ] Y for all i. Since Y is a linear subspace of Lp () it
follows that [s] Y . This proves Step 1.
4.3. SEPARABILITY 121
Step 2. If f Lp () and f 0 then [f ] Y .

By Theorem 1.26 there is a sequence of measurable step functions si : X R
such that 0 s1 s2 and si converges pointwise to f . Then si Lp ()
and hence [si ] Y for all i by Step 1. Moreover, |f si |p f p , f p is
integrable, and |f si |p converges pointwise to zero. Hence it follows from the
Lebesgue Dominated Convergence Theorem 1.45 that limi kf si kp = 0.
Since [si ] Y for all i and Y is a closed subspace of Lp (), it follows that
[f ] Y . This proves Step 2.
Step 3. Y = Lp ().
Let f Lp (). Then f Lp (), hence [f ] Y by Step 2, and hence
[f ] = [f + ] [f ] Y . This proves Step 3 and Lemma 4.12.
Standing Assumption. Assume throughout the remainder of this section
that (X, U) is a locally compact Hausdorff space, B 2X is its Borel -
algebra, : B [0, ] is a Borel measure, and fix a constant 1 p < .
Theorem 4.13. If X is second countable then Lp () is separable.
Proof. See page 122.
Example 4.14. If X is an uncountable set with the discrete topology U = 2X
and : 2X [0, ] is the counting measure then X is not second countable
and Lp () = Lp () is not separable.
Theorem 4.15. Assume is outer regular and is inner regular on open sets.
Define

s is a Borel measurable step function
Sc (X) := s : X R
. (4.13)
and supp(s) is a compact subset of X

Then the linear subspaces Sc (X)/ and Cc (X)/ are dense in Lp (). This
continues to hold when is a Radon measure.
Example 4.16. Let (X, U) be the compact Hausdorff space constructed in
Example 3.6, let : A [0, 1] be the Dieudonne measure constructed in
that example, let : 2X [0, 1] be the Dirac measure at the point X,
and define 0 := |B + |B : B [0, 2]. Then Lp (0 ) is a 2-dimensional
0
vector space and Cc (X)/ is a 1-dimensional subspace of Lp (0 ) and hence
is not dense. Thus the regularity assumption on cannot be removed in
Theorem 4.15.
Lemma 4.17. Assume = 1 is outer regular and is inner regular on open

sets. Let 0 : B [0, ] be the unique Radon measure such that 1 = 0 .
Then Lp (1 ) Lp (0 ) and the linear map
Lp (1 ) Lp (0 ) : [f ]1 7 [f ]0 (4.14)
is a Banach space isometry.

Proof.
R Since R0 (B) 1 (B) for all B B by Theorem 3.15 it follows that
p p
X
|f | d0 X |f | d1 for every Borel measurable function f : X R.
Hence Lp (1 ) Lp (0 ). We prove that
Z Z
p
|f | d0 = |f |p d1 for all f Lp (1 ). (4.15)
X X
Thus the map (4.14) is injective and has a closed image. To prove (4.15),
define E := {x X | |f (x)| > } for > 0. Then 1 (E ) < and hence
R 1 and 0 agree
R onp all Borel subsets of E by Lemma 3.7. This implies
p
E
|f | d0 = E |f | d1 , and (4.15) follows by taking the limit 0.
We prove that the map (4.14) is surjective. Denote its image by X . This
is a closed linear subspace of Lp (0 ), by what we have just proved. Let B B
such that 0 (B) < . By (3.5) there is a sequence of compact sets Ki B
i
such thatS Ki Ki+1 and 1 (Ki ) = 0 (Ki ) > 0 (B) 2 for all i. Define
A := iN Ki B. Then 1 (A) = 0 (A) = limi 0 (Ki ) = 0 (B). This
implies A Lp (1 ) and [B ]0 = [A ]0 X . By Lemma 4.12, it follows
that X = Lp (0 ) and this proves Lemma 4.17.
Proof of Theorem 4.13. Let V U be a countable basis for the topology. As-
sume without of generality that V is compact for all V V. (If W U is any
countable basis for the topology then the set V := V W | V is compact
is also a countable basis for the topology by Lemma A.3.) Choose a bijection
N V : i 7 Vi and let I := {I N P| #I < } be the set of finite subsets
of N. Then the map I N : I 7S iI 2 i1
is a bijection, so the set I is
countable. For I I define VI := iI Vi . Define the set V Lp () by
( `
)
X
V := s = j VIj ` N and j Q, Ij I for j = 1, . . . , ` .

j=1
This set is contained in Lp () because V is compact for all V V . It

is countable and its closure X := V in Lp () is a closed linear subspace.
4.3. SEPARABILITY 123
By Lemma 4.12 it suffices to prove that [B ] X for every B B with

(B) < . To see this, fix a Borel set B B with (B) < and a constant
> 0. Since X is second countable every open subset of X is -compact
(Lemma 3.21). Hence is regular by Theorem 3.18. Hence there exists a
compact set K X and an open set U X such that
K B U, (U \ K) < p .
Define I :=
S {i N | Vi U }. Since V is a basis of the topology we have
K U = iI Vi . Since K is compact there is a finite set I I such that
K VI U.
Since B VI vanishes on X \ (U \ K) and |B VI | 1 it follows that
kB VI kp (U \ K)1/p < .
Since VI V and > 0 was chosen arbitrary it follows that [B ] X = V .

Proof of Theorem 4.15. By Corollary 3.16 and Lemma 4.17 it suffices to con-
sider the case where is outer regular and is inner regular on open sets.
Define
n o
S := [f ] L () | > 0 s Sc (X) such that kf skp < ,
p
n o
C := [f ] Lp (1 ) | > 0 g Cc (X) such that kf gkp < .
We must prove that Lp () = S = C . Since S and C are closed linear

subspaces of Lp () it suffices to prove that [B ] S C for every Borel set
B B with (B) < by Lemma 4.12. Let B B with (B) < and let
> 0. By Lemma 3.7 there is a compact set K X and an open set U X
such that K B U and (U \ K) < p . By Urysohns Lemma A.1 there
is a function f Cc (X) such that 0 f 1, f |K 1, and supp(f ) U .
This implies 0 f K U \K and 0 B K U \K . Hence

kB K kp U \K p = (U \ K)1/p <
and likewise kf K kp < . By Minkowskis inequality (4.4) this implies

kB f kp kB K kp + kK f kp < 2.
This shows that [B ] S C . This proves Theorem 4.15.
Remark 4.18. The reader may wonder whether Theorem 4.15 continues to
hold for all Borel measures : B [0, ] that are inner regular on open sets.
To answer this question one can try to proceed as follows. Let 0 , 1 be the
Borel measures on X in Theorem 3.15 that satisfy 0 = 1 = . Then
0 is a Radon measure, 1 is outer regular and is inner regular on open sets,
and 0 (B) (B) 1 (B) for all B B. Thus Lp (1 ) Lp () Lp (0 )
and one can consider the maps
Lp (1 ) Lp () Lp (0 ).
Their composition is a Banach space isometry by Lemma 4.17. The question

is now whether or not the first map Lp (1 ) Lp () is surjective or, equiv-
alently, whether the second map Lp () Lp (0 ) is injective. If this holds

then the subspace Cc (X)/ is dense in Lp (), otherwise it is not. The proof
of Lemma 4.17 shows that the answer is affirmative if and only if every Borel
set B X with 0 (B) < (B) satisfies (B) = . Thus the quest for a
counterexample can be rephrased as follows.
Question. Does there exist a locally compact Hausdorff space (X, U) and
Borel measures 0 , 1 , : B [0, ] on its Borel -algebra B 2X such
that all three measures are inner regular on open sets, 1 is outer regular,
0 is given by (3.5), 0 (B) (B) 1 (B) for all Borel sets B B, and
0 = 0 (B) < (B) < 1 (B) = for some Borel set B B?
This leads to deep problems in set theory. A probability measure on a mea-
surable space (X, A) is a measure : A [0, 1] such that (X) = 1. A mea-
sure : A [0, ] is called nonatomic if countable sets have measure zero.
Now consider the measure on X = R2 in Exercise 3.23 with 0 (R {0}) = 0
and 1 (R {0}) = , and define : R R2 by (x) := (x, 0). If there is
a nonatomic probability measure : 2R [0, 1] then the measure 0 +
provides a positive answer to the above question, and thus Theorem 4.15
would not extend to all Borel measures that are inner regular on open sets.
The question of the existence of a nonatomic probability measure is related
to the continuum hypothesis. The generalized continuum hypothesis
asserts that, if X is any infinite set, then each subset of 2X whose cardinality
is strictly larger than that of X admits a bijection to 2X . It is independent
of the other axioms of set theory and implies that nonatomic probability
measures : 2X [0, 1] do not exist on any set X. This is closely related
to the theory of measure-free cardinals. (See Fremlin [4, Section 4.3.7].)
4.4. HILBERT SPACES 125
4.4 Hilbert Spaces

This section introduces some elementary Hilbert space theory. It serves two
purposes. First, it shows that the Hilbert space L2 () is isomorphic to its
own dual space. Second, this result in turn will be used in the proof of the
RadonNikod ym Theorem for -finite measure spaces in the next chapter.
Definition 4.19. Let H be a real vector space. A bilinear map
H H R : (x, y) 7 hx, yi (4.16)
is called an inner product if it is symmetric, i.e. hx, yi = hy, xi for all
x, y H and positive definite, i.e. hx, xi > 0 for all x H \ {0}. The
norm associated to an inner product (4.16) is the function
p
H R : x 7 kxk := hx, xi. (4.17)
Lemma 4.20. Let H be a real vector space equipped with an inner prod-
uct (4.16) and the associated norm (4.17). The inner product and norm
satisfy the CauchySchwarz inequality
|hx, yi| kxk kyk (4.18)
and the triangle inequality
kx + yk kxk + kyk (4.19)
for all x, y H. Thus (4.17) is a norm on H.
Proof. The CauchySchwarz inequality is obvious when x = 0 or y = 0.
Hence assume x 6= 0 and y 6= 0 and define := kxk1 x and := kyk1 y.
Then kk = kk = 1. Hence
0 k h, ik2 = h, h, ii = 1 h, i2 .
This implies |h, i| 1 and hence |hx, yi| kxk kyk. In turn it follows from
the CauchySchwarz inequality that
kx + yk2 = kxk2 + 2hx, yi + kyk2
kxk2 + 2 kxk kyk + kyk2
= (kxk + kyk)2 .
This proves the triangle inequality (4.19) and Lemma 4.20.
Definition 4.21. An inner product space (H, h, i) is called a Hilbert space

if the norm (4.17) is complete, i.e. every Cauchy sequence in H converges.
Example 4.22. Let (X, A, ) be a measure space. Then H := L2 () is a
Hilbert space. The inner product is induced by the bilinear map
Z
2 2
L () L () R : (f, g) 7 hf, gi := f g d. (4.20)
X
2
It is well defined because the product of two L -functions f, g : X R is
integrable by (4.3) with p = q = 2. That it is bilinear follows from Theo-
rem 1.44 and that it is symmetric is obvious. In general, it is not positive
definite. However, it descends to a symmetric bilinear form
Z
2 2
L () L () R : ([f ] , [g] ) 7 hf, gi = f g d. (4.21)
X
by Lemma 1.48 which is positive definite by Lemma 1.49. Hence (4.21) is

an inner product on L2 (). It is called the L2 inner product. The norm
associated to this inner product is
Z 1/2
2 2
p
L () R : [f ] 7 kf k2 = f d = hf, f i. (4.22)
X
This is the L -norm in (4.5) with p = 2. By Theorem 4.9, L2 () is complete

2
with the norm (4.22) and hence is a Hilbert space.

Definition 4.23. Let (V, kk) be a normed vector space. A linear functional
: V R is called bounded if there exists a constant c 0 such that
|(x)| c kxk for all x V.
The norm of a bounded linear functional : V R is the smallest
such constant c and will be denoted by
|(x)|
kk := sup . (4.23)
06=xV kxk
The set of bounded linear functionals on V is denoted by V and is called the
dual space of V .
Exercise 4.24. Prove that a linear functional on a normed vector space is
bounded if and only if it is continuous.
Exercise 4.25. Let (V, kk) be a normed vector space. Prove that the dual
space V with the norm (4.23) is a Banach space. (See Example 1.11.)
4.4. HILBERT SPACES 127
Theorem 4.26 (Riesz). Let H be a Hilbert space and let : H R be a

bounded linear functional. Then there exists a unique element y H such
that
(x) = hy, xi for all x H. (4.24)
This element y H satisfies
|hy, xi|
kyk = sup = kk . (4.25)
06=xH kxk
Thus the map H H : y 7 hy, i is an isometry of normed vector spaces.
Theorem 4.27. Let H be a Hilbert space and let E H be a nonempty
closed convex subset. Then there exists a unique element x0 E such that
kx0 k kxk for all x E.
Theorem 4.27 implies Theorem 4.26. We prove existence. If = 0 then
y = 0 satisfies (4.24). Hence assume 6= 0 and define
E := {x H | (x) = 1} .
Then E 6= because there exists an element H such that () 6= 0
and hence x := ()1 E. The set E is a closed because : H R is
continuous, and it is convex because is linear. Hence Theorem 4.27 asserts
that there exists an element x0 E such that
kx0 k kxk for all x E.
We prove that
x H, (x) = 0 = hx0 , xi = 0. (4.26)
To see this, fix an element x H such that (x) = 0. Then x0 + tx E for
all t R. This implies
kx0 k2 kx0 + txk2 = kx0 k2 + 2thx0 , xi + t2 kxk2 for all t R.
Thus the differentiable function t 7 kx0 + txk2 attains its minimum at t = 0
and so its derivative vanishes at t = 0. Hence

d
0 = kx0 + txk2 = 2hx0 , xi
dt t=0
and this proves (4.26).

Now define
x0
y := .
kx0 k2
Fix an element x H and define := (x). Then (xx0 ) = (x) = 0.
Hence it follows from (4.26) that
0 = hx0 , x x0 i = hx0 , xi kx0 k2 .
This implies
hx0 , xi
hy, xi = = = (x).
kx0 k2
Thus y satisfies (4.24).
We prove (4.25). Assume y H satisfies (4.24). If y = 0 then = 0 and
so kyk = 0 = kk. Hence assume y 6= 0. Then
kyk2 (y) |(x)|
kyk = = sup = kk .
kyk kyk 06=xH kxk
Conversely, it follows from the CauchySchwarz inequality that

|(x)| = |hy, xi| kykkxk
for all x H and hence kk kyk. This proves (4.25).
We prove uniqueness. Assume y, z H satisfy
hy, xi = hz, xi = (x)
for all x H. Then hy z, xi = 0 for all x H. Take x := y z to obtain
ky zk2 = hy z, y zi = 0
and hence y z = 0. This proves Theorem 4.26, assuming Theorem 4.27.

:= inf kxk x E .
We prove uniqueness. Let x0 , x1 E such that
kx0 k = kx1 k = .
Then 12 (x0 + x1 ) E because E is convex and so kx0 + x1 k 2. Thus
kx0 x1 k2 = 2 kx0 k2 + 2 kx1 k2 kx0 + x1 k2 = 4 2 kx0 + x1 k2 0
and therefore x0 = x1 .
4.5. THE DUAL SPACE OF LP () 129
We prove existence. Choose a sequence xi E such that

lim kxi k = .
i
We prove that xi is a Cauchy sequence. Fix a constant > 0. Then there

exists an integer i0 N such that

i N, i i0 = kxi k2 < 2 + .
4
1
Let i, j N such that i i0 and j i0 . Then 2 (xi + xj ) E because E is
convex and hence kxi + xj k 2. This implies
kxi xj k2 = 2 kxi k2 + 2 kxj k2 kxi + xj k2

< 4 2 + 4 2 = .
4
Thus xi is a Cauchy sequence. Since H is complete the limit x0 := limi xi
exists. Moreover x0 E because E is closed and kx0 k = because the Norm
function (4.17) is continuous. This proves Theorem 4.27.
Corollary 4.28. Let (X, A, ) be a measure space and let : L2 () R be
a bounded linear functional. Then there exists a function g L2 (), unique
up to equality almost everywhere, such that
Z
([f ] ) = f g d for all f L2 ().
X
Moreover kk = kgk2 . Thus L2 () is isomorphic to L2 ().

Proof. This follows immediately from Theorem 4.26 and Example 4.22.
4.5 The Dual Space of Lp()

We wish to extend Corollary 4.28 to the Lp -spaces in Definition 4.3 and
equation (4.8) (for 1 p < ) and in Definition 4.7 (for p = ). When
1 < p < it turns out that the dual space of Lp () is always isomorphic
to Lq () where 1/p + 1/q = 1. For p = the natural homomorphism
L1 () L () is an isometric embedding, however, in most cases the dual
space of L () is much larger than L1 (). For p = 1 the situation is more
subtle. The natural homomorphism L () L1 () need not be injective or
surjective. However, it is bijective for a large class of measure spaces and one
can characterize those measure spaces for which it is injective, respectively
bijective. This requires the following definition.
Definition 4.29. A measure space (X, A, ) is called -finite if there exists

a sequence of measurable subsets Xi A such that

[
X= Xi , Xi Xi+1 , (Xi ) < for all i N. (4.27)
i=1
It is called semi-finite if every measurable set A A satisfies

E A such that E A
(A) > 0 = (4.28)
and 0 < (E) < .
It is called localizable if it is semi-finite and, for every collection of measur-
able sets E A, there is a set H A satisfying the following two conditions.
(L1) (E \ H) = 0 for all E E.
(L2) If G A satisfies (E \ G) = 0 for all E E then (H \ G) = 0.
A measurable set H satisfying (L1) and (L2) is called an envelope of E.
The geometric intuition behind the definition of localizable is as follows.
The collection E A will typically be uncountable so one cannot expect its
union to be measurable. The envelope H is a measurable set that replaces
the union of the sets in E. It covers each set E E up to a set of measure
zero and, if any other measurable set G covers each set E E up to a set of
measure zero, it also covers H up to a set of measure zero. The next lemma
clarifies the notion of semi-finiteness.
Lemma 4.30. Let (X, A, ) be a measure space.
(i) (X, A, ) is semi-finite if and only if
(A) = sup {(E) | E A, E A, (E) < } (4.29)
for every measurable set A A.
(ii) If (X, A, ) is -finite then it is semi-finite.
Proof. We prove (i). Assume (X, A, ) is semi-finite, let A A, and define
a := sup {(E) | E A, E A, (E) < } .
Then a (A) and we must prove that a = (A). This is obvious when
a = . Hence assume a < . Choose a sequence of measurable sets Ei A
such that (Ei ) < and (Ei ) > a 2i for all i. Define

[
[
Bi := E1 Ei , B := Bi = Ei .
i=1 i=1
Then Bi A, Ei Bi A, and (Bi ) < . Hence (Ei ) (Bi ) a for

all i N and hence
(B) = lim (Bi ) = a < .
i
If (A \ B) > 0 then, since (X, A, ) is semi-finite, there exists a measurable

set F A such that F A \ B and 0 < (F ) < , and hence
B F A, a < (B F ) = (B) + (F ) < ,
contradicting the definition of a. This shows that (A \ B) = 0 and hence
(A) = (B) + (A \ B) = a, as claimed. Thus we have proved that every
semi-finite measure space satisfies (4.29). The converse is obvious and this
proves part (i).
We prove (ii). Assume that (X, A, ) is -finite and choose a sequence of
measurable sets Xi A that satisfies (4.27). If A A then it follows from
Theorem 1.28 that (A) = limi (A Xi ). Since (A Xi ) < for all
i this shows that every measurable set A satisfies (4.29) and so (X, A, ) is
semi-finite. This proves Lemma 4.30.
It is also true that every -finite measure space is localizable. This can be
derived as a consequence of Theorem 4.35 (see Corollary 5.9 below). A more
direct proof is outlined in Exercise 4.58.
Example 4.31. Define (X, A, ) by
X := {a, b}, A := 2X , ({a}) := 1, ({b}) := .
This measure space is not semi-finite. Thus the linear map L () L1 ()
in Theorem 4.33 below is not injective. In fact, L () has dimension two
and L1 () has dimension one.
Example 4.32. Let X be an uncountable set, let A 2X be the -algebra of
all subsets A X such that A or Ac is countable, and let : A [0, ] be
the counting measure. Then (X, A, ) is semi-finite, but it is not localizable.
For example, let H X be an uncountable set with an uncountable com-
plement and let E be the collection of all finite subsets of H. Then the only
possible envelope of E would be the set H itself, which is not measurable.
Thus Theorem 4.33 below shows that the map L () L1 () is injective
and Theorem 4.35 below shows that it is not surjective. An example of a
bounded linear functional : L1 ()
P R that cannot1 be represented by an
1
L -function is given by (f ) := xH f (x) for f L () = L ().
The next theorem is the first step towards understanding the dual space
of Lp () and is a fairly easy consequence of the Holder inequality. It asserts
that for 1/p + 1/q = 1 every element of Lq () determines a bounded lin-
ear functional on Lp () and that the resulting map Lq () Lp () is an
isometric embedding (for p = 1 under the semi-finite hypothesis). The key
question is then whether every bounded linear functional on Lp () is of that
form. That this is indeed the case for 1 < p < (and for p = 1 under the
localizable hypothesis) is the content of Theorem 4.35 below. This is a much
deeper theorem whose proof for p 6= 2 requires the RadonNikod ym theorem
and will be carried out in Chapter 5.
Theorem 4.33. Let (X, A, ) be a measure space and fix constants
1 1
1 p , 1 q , + = 1. (4.30)
p q
(i) Let g Lq (). Then the formula
Z
g ([f ] ) := f g d for f Lp () (4.31)
X
defines a bounded linear functional g : Lp () R and

R
| X f g d|
kg k = sup kgkq . (4.32)
f Lp (), kf kp 6=0 kf kp
(ii) The map g 7 g in (4.31) descends to a bounded linear operator
Lq () Lp () : [g] 7 g . (4.33)
(iii) Assume 1 < p Then kg k = kgkq for all g Lq ().

(iv) Assume p = 1. Then the map L () L1 () in (4.33) is injective if
and only if it is an isometric embedding if and only if (X, A, ) is semi-finite.
The heart of the proof is the next lemma. It is slightly stronger than
what is required to prove Theorem 4.33 in that the hypothesis on g to be
q-integrable is dropped in part (iii) and replaced by the assumption that
the measure space is semi-finite. In this form Lemma 4.34 is needed in the
proof of Theorem 4.35 and will also be useful for proving the inequalities of
Minkowski and CalderonZygmund in Theorems 7.19 and 7.43 below.
Lemma 4.34. Let (X, A, ) be a measure space and let p, q be as in (4.30).

Let g : X [0, ] be a measurable function and suppose that there exists a
constant c 0 such that
Z
p
f L (), f 0 = f g d c kf kp . (4.34)
X

(i) If q = 1 then kgk1 c.
(ii) If 1 < q < and kgkq < then kgkq c.
(iii) If 1 < q < and (X, A, ) is semi-finite then kgkq c.
(iv) If q = and (X, A, ) is semi-finite then kgk c.
Proof. We prove (i). If q = 1 take f 1 in (4.34) to obtain kgk1 c.

We prove (ii). Assume 1 < q < and kgkq < . Then it follows from
Lemma 1.47 that the set A := {x X | g(x) = } has measure zero. Define
the function h : X [0, ) by h(x) := g(x) for x X \ A and by h(x) := 0
for x A. Then h is measurable and
Z Z
khkq = kgkq < , f h d = f g d c kf kp
X X
for all f Lp () with f 0 by Lemma 1.48. Define f : X [0, ) by

f (x) := h(x)q1 for x X. Then f p = hp(q1) = hq = f h and hence
Z 11/q Z
kf kp = q
h d = khkq1
q , f h d = khkqq .
X X
Thus f Lp () and so khkqq = X f h d c kf kp = c khkqq1 . Since

R
khkq < it follows that kgkq = khkq c and this proves part (ii).
We prove (iii). Assume (X, A, ) is semi-finite and 1 < q < . Suppose,
by contradiction, that kgkq > c. We will prove that there exists a measurable
function h : X [0, ) such that
0 h g, c < khkq < . (4.35)

R R
By (4.34) this function h satisfies X f h d X f g d c kf kp for all
f Lp () with f 0. Since khkq < it follows from part (ii) that khkq c,
which contradicts the inequality khkq > c in (4.35).
It remains to prove the existence of h. Since kgkq > c it follows from

Definition 1.34 that there exists
R qa measurable step function s : X [0, )
q
such that 0 s g and X s d > c . If kskq < take h := s. If
kskq = there exists a measurable set A X and a constant > 0 such
that (A) = and A s g. Since (X, A, ) is semi-finite, Lemma 4.30
asserts that there exists a measurable set E A such that E A and
cq < q (E) < . Then the function h := E : X [0, ) satisfies
0 h g and khkq = (E)1/q > c as required. This proves part (iii).
We prove (iv). Let q = and assume (X, A, ) is semi-finite. Suppose,
by contradiction, that kgk > c. Then there exists a constant > 0 such that
the set A := {x X | g(x) c + } has positive measure. Since (X, A, ) is
semi-finite there exists a measurable
R set E A such that 0 < (E) < .
1
Hence f := E L () and X f g d (c + )(E) > c(E) = c kf k1 , in
contradiction to (4.34). This proves (iv) and Lemma 4.34.
Proof of Theorem 4.33. The proof has four steps.
Step 1. Let f Lp (), g Lq (). Then f g L1 () and kf gk1 kf kp kgkq .
R
If 1 < p < then X |f g| d kf kp kgkq by the Holder inequality (4.3). If
p = 1 then |f g| |f | kgk almost everywhere by Lemma 4.8, so f g L1 ()
and kf gk1 kf k1 kgk . If p = interchange the pairs (f, p) and (g, q).
Step 2. We prove (i) and (ii).
By Step 1 the right hand side of (4.31) is well defined and by Lemma 1.48 it
depends only on the equivalence class of f under equality almost everywhere.
Hence g is well defined. It is linear by Theorem 1.44 and kg k kgkq by
Step 1. This proves (i). The bounded linear functional g : Lp () R
depends only on the equivalence class of g, again by Lemma 1.48. Hence
the map (4.33) is well defined. By Theorem 1.44 and (4.32) it is a bounded
linear operator of norm less than or equal to one. This proves (ii).
Step 3. If 1 0,
sign(t) := 1 for t < 0, and by sign(0) = 0. If f Lp () is nonnegative then
the function f sign(g) : X R is p-integrable and
Z
f |g| d = g (f sign(g)) kg k kf sign(g)kp kg k kf kp .
X
Hence kgkq kg k by Lemma 4.34 and so kg k = kgkq by Step 2.
Step 4. If the map L () L1 () is injective then (X, A, ) is semi-finite.

Let A A such that (A) > 0 and define g := A . Then g : L1 () R is
nonzero by assumption. Hence there is an f L1 () such that
Z Z
0 < g (f ) = f g d = f d. (4.36)
X A
For i N define Ei := {x A | f (x) > 2i }. Then Ei A, Ei A, and

Z
i
(Ei ) 2 f d 2i kf k1 < .
Ei
Moreover E :=
S
i=1 Ei = {x A | f (x) > 0} is not a null set by (4.36).
Hence one of the sets Ei has positive measure. Thus (X, A, ) is semi-finite.
The next theorem asserts that, for 1 < p < , every bounded linear
functional on Lp () has the form (4.31) for some g Lq (). For p 6= 2
this is a much deeper result than Corollary 4.28. The proof requires the
RadonNikod ym Theorem and will be deferred to the next chapter.
Theorem 4.35 (The Dual Space of Lp ). Let (X, A, ) be a measure space

and fix constants
1 1
1 p < , 1 < q , + = 1.
p q
(i) Assume 1 < p < . Then the map Lq () Lp () : [g] 7 g defined
by (4.31) is bijective and hence is a Banach space isometry.
(ii) Assume p = 1. Then the map L () L1 () : [g] 7 g defined
by (4.31) is bijective if and only if (X, A, ) is localizable.

This next example shows that, in general, Theorem 4.35 does not extend
to the case p = (regardless of whether or not the measure space (X, A, )
is -finite). By Theorem 4.33 the Banach space L1 () is equipped with an
isometric inclusion L1 () L () , however, the dual space of L () is
typically much larger than L1 ().
Example 4.36. Let : 2N [0, ] be the counting measure on the positive

integers. Then ` := L () = L () is the Banach space of bounded
sequences x = (xn )nN of real numbers equipped with the supremum norm
kxk := supnN |xn |. An interesting closed subspace of ` is
c := {x = (xn )nN ` | x is a Cauchy sequence} .
It is equipped with a bounded linear functional 0 : c R, defined by
0 (x) := lim xn for x = (xn )nN c.
n
The HahnBanach Theorem, one of the fundamental principles of Func-

tional Analysis, asserts that every bounded linear functional on a linear sub-
space of a Banach space extends to a bounded linear functional on the entire
Banach space (whose norm is no larger than the norm of the original bounded
linear functional on the subspace). In the case at hand this means that there
is a bounded linear functional : ` R such that |c = 0 . This lin-
ear functional cannot have the form (4.31) for any g L1 (). To see this,
note that `1 := L1 () = L1 () is the space of summable sequences of real
numbers.
P Let y = (yn )nN `1 be a sequence of real numbers such that

n=1 |yn | < and define the linear functional y : ` R by

X
y (x) := xn y n for x = (xn )nN ` .
n=1
that
P
Choose N N such n=N |yn |
=: < 1 and define x = (xn )nN c by
xn := 0 for n < N and xn := 1 for n N . Then y (x) < 1 = (x)
and hence y 6= . This shows that does not belong to the image of the
isometric inclusion `1 , (` ) .
Exercise 4.37. Let 0 : c R be the functional in Example 4.36 and
denote its kernel by c0 := ker 0 . Thus c0 is the set of all sequences of real
numbers that converge to zero, i.e.

c0 = x = (xn )nN ` lim xn = 0 .
n
Prove that c0 is a closed linear subspace of ` and that `1 is naturally iso-

morphic to the dual space of c0 . Thus
`1
= (c0 ) , c0 ( `
= (`1 )
= (c0 ) , `1 ( (` )
= (`1 ) .
In the language of Functional Analysis this means that the Banach spaces c0
and `1 are not reflexive, and neither is ` .
We close this section with two results that will be needed in the proof
of Theorem 4.35. The first asserts that every bounded linear functional on
Lp () can be written as the difference of two positive bounded linear function-
als (Theorem 4.39). The second asserts that every positive bounded linear
functional on Lp () is supported on a -finite subset of X (Theorem 4.40).
When : Lp () R is a bounded linear functional it will be convenient to
abuse notation and write (f ) := ([f ] ) for f Lp ().
Definition 4.38. Let (X, A, ) be a measure space and let 1 p < . A
bounded linear functional : Lp () R is called positive if
f 0 = (f ) 0
for all f Lp ().

Theorem 4.39. Let (X, A, ) be a measure space, let 1 p < , and let
: Lp () R be a bounded linear functional. Define : A [0, ] by
(A) := sup {(E ) | E A, E A, (E) < } (4.37)
Then the maps are measures, Lp () L1 (+ ) L1 ( ), and the formulas

Z

(f ) := f d for f Lp () (4.38)
X
define positive bounded linear functionals : Lp () R such that
= + , k+ k + k k = kk. (4.39)
Proof. The proof has four steps.

Step 1. The maps : A [0, ] in (4.37) are measures.
It follows directly from the definition that () = 0. We must prove that
+ is -additive. That is then also -additive follows by reversing the
S Ai A be a sequence of pairwise disjoint measurable sets
sign of . Thus let
and define A := i=1 Ai . Let E A such that E A and (E) < . Then
it follows from the definition of + that
(EAi ) + (Ai ) for all i N. (4.40)
Moreover the sequence of measurable functions fn := E ni=1 EAi 0

P
converges pointwise to zero and satisfies 0 fnp E for all n. Since
(E) < the function E is integrable and so it follows

R pfrom the Lebesgue
Dominated Convergence Theorem 1.45 that limn X fn d = 0, i.e.

Xn

lim E EA
= 0.
i
n

i=1 p

n
X
X
X
(E ) = lim (EAi ) = (EAi ) + (Ai ).
n
i=1 i=1 i=1
Take the supremum over all E A with E A and (E) < to obtain

X
+ (A) + (Ai ).
i=1
To prove the converse inequality, assume first that + (Ai ) = for some i;

since Ai A this implies + (A) = = i=1 + (Ai ). Hence it suffices to
P
assume + (Ai ) < for all i. Fix a constant > 0 and choose a sequence of
measurable sets Ei A such that Ei Ai and (Ei ) > + (Ai ) 2i for
all i. Since E1 En A it follows from the definition of + that
n
X n
X
+
(A) (E1 En ) = (Ei ) > + (Ai ) .
i=1 i=1
P
Take the limit n to obtain + (A) i=1 + (Ai ) for all > 0, so

X
+
(A) + (Ai )
i=1
as claimed. Thus + is -additive and this proves Step 1.

Step 2. Let c := kk. Then every measurable function f : X R satisfies
Z Z
|f | d + |f | d c kf kp .
+
(4.41)
X X
In particular, Lp () L1 (+ ) L1 ( ).
Assume first that f = s : X [0, ) is a measurable step function in
Lp (). Then there exist real numbers i > 0 and measurable sets Ai A
for i = 1, . . . , ` such that Ai Aj = for i 6= j, (Ai ) < for all i, and

s = ì=1 i Ai . Now fix a real number > 0 and choose i > 0 such that
P
`
X
i i = .
i=1
2
For i = 1, . . . , ` choose Ei A such that
Ei Ai , (Ei+ ) + (Ai ) i , (Ei ) (Ai ) i .
Then
Z Z `
X

+
i + (Ai ) + (Ai )

s d + s d =
X X i=1
X`
i (Ei+ ) (Ei ) + 2i
i=1
`
!
X
= i Ei+ Ei +
i=1
`
X

c
i Ei
+ +
Ei
i=1 p
`
!1/p
X
= c ip (Ei+ \ Ei ) + (Ei \ Ei+ ) +
i=1
`
!1/p
X
c ip (Ai ) +
i=1
= c kskp + .
Take the limit 0 to obtain (4.41) for f = s. To prove (4.41) in gen-

eral it suffices to assume that f Lp () is nonnegative. By Theorem 1.26
there is a sequence of measurable step functions 0 s1 s2 that
converges pointwise to f . Then (f sn )p converges pointwise to zero and is
bounded above by f p L1 (). Hence limn kf sn kp = 0 by the Lebesgue
Dominated Convergence Theorem 1.45 and limn X sn d = X f d by
R R
the Lebesgue Monotone Convergence Theorem 1.37. This proves (4.41). It

follows from (4.41) that Lp () L1 (+ ) L1 ( ) and this proves Step 2.
Step 3. If A A and (A) < then

(A) < , (A ) = + (A) (A) (4.42)
It follows from the inequality (4.41) in Step 2 that

Z Z

+
(A) + (A) = +
A d + A d c kA kp = c(A)1/p < .
X X
Now let > 0 and choose E A such that E A and (E ) > + (A) .
Since (A\E ) (A) this implies
(A ) = (E ) + (A\E ) > + (A) (A) .

Since this holds for all > 0 we obtain (A ) + (A) (A). Reversing
the sign of we also obtain (A ) (A)+ (A) and this proves Step 3.
Step 4. If f Lp () then
Z Z
(f ) = f d +
f d . (4.43)
X X
Let s : X R be a p-integrable step function. Then there are real numbers

i and measurable sets Ai A for i = 1, . . . , ` such that (Ai ) < for all i
and s = ì=1 i Ai . Hence it follows from Step 3 that
P
`
X `
X Z Z

+ +
s d .

(s) = i (Ai ) = i (Ai ) (Ai ) = s d
i=1 i=1 X X
This proves (4.43) for p-integrable step functions. Now let f Lp () and
assume f 0. By Theorem 1.26 there is a sequence of measurable step
functions 0 s1 s2 that converges pointwise to f . Then (f sn )p
converges pointwise to zero and is bounded above by f p L1 (). Hence
limn kf sn kp = 0 by the Lebesgue Dominated Convergence Theorem
and hence limn (sn ) = (f ). Moreover, X f d c kf kp < by
R
Step 2 and limn X sn d = X f d by the Lebesgue Monotone Con-

R R
vergence Theorem. Thus every nonnegative Lp -function f : X [0, )

satisfies (4.43). If f Lp () then f Lp () satisfy (4.43) by what we have
just proved and hence so does f = f + f . This proves Step 4.
It follows from Steps 2 and 4 that the linear functionals : Lp () R
in (4.38) are bounded and satisfy (4.39). This proves Theorem 4.39.
Theorem 4.40. Let (X, A, ) be a measure space, let 1 < p < , and let
: Lp () R be a positive bounded linear functional. Define
(A) := sup {(E ) | E A, E A, (E) < } (4.44)
for A A. Then the map : A [0, ] is a measure, Lp () L1 (), and
Z
(f ) = f d for all f Lp (). (4.45)
X
Moreover, there are measurable sets N A and Xn A for n N such that

[
X \N = Xn , (N ) = 0, (Xn ) < , Xn Xn+1 (4.46)
n=1
for all n N.
Proof. That is a measure satisfying Lp () L1 () and (4.45) follows from
Theorem 4.39 and the fact that + = and = 0 because is positive.
Now define c := kk. We prove in three steps that there exist measurable
sets N A and Xn A for n N satisfying (4.46).
Step 1. For every > 0 there exists a measurable set A A and a measur-
able function f : X [0, ) such that
f |X\A = 0, inf f > 0, kf kp = 1, (f ) > c . (4.47)
A
In particular, (A) (inf A f )p < .

Choose h Lp () such that khkp = 1 and (h) > c . Assume without
loss of generality that h 0. (Otherwise replace h by |h|.) Define
Ai := x X h(x) > 2i .

Then (h hAi )p converges pointwise to zero as i and is bounded by

the integrable function hp . Hence it follows from the Lebesgue Dominated
Convergence Theorem 1.45 that limi kh hAi kp = 0 and therefore
lim (hAi ) = (h) > c .
i
Choose i N such that (hAi ) > c and define

hAi
A := Ai , f := .
khAi kp
Then A and f satisfy (4.47) and so (A) (inf A f )p f p d = (inf A f )p .
R
X
This proves Step 1.
Step 2. Let , A, f be as in Step 1 and let E A. Then

EA= (E ) 1/q c
= < +1 , (4.48)
(E) < (E)1/p p
where 1 < q < is chosen such that 1/p + 1/q = 1.
Define 1/p

g := f + E .
(E)
Then Z 1/p
p
kgkp = f d + = (1 + )1/p
X
and, by (4.47),
1/p
(E )
(g) = (f ) + (E ) > c + 1/p .
(E) (E)1/p
Since (g) c kgkp it follows that
(E )
c + 1/p < c(1 + )1/p .
(E)1/p
Since (1 + )1/p 1 /p for all 0 this implies

1/p (E ) c

1/p
< c (1 + ) 1 + +1 .
(E)1/p p
Since 11/p = 1/q this proves Step 2.
Step 3. There exist measurable sets N, X1 , X2 , X3 , . . . satisfying (4.46).
Choose An A and fn Lp () as in Step 1 with = 1/n. For n N define

[
[
Xn := A1 An , N := X \ An = X \ Xn .
n=1 n=1
By Step 2 every measurable set E N with (E) < satisfies

(E ) 1 c
< 1/q +1
(E)1/p n p
for all n N and hence
Pn (E ) = 0. This implies (N ) = 0 by (4.44).
Moreover (Xn ) i=1 (Ai ) < for every n by Step 1. This proves
Step 3 and Theorem 4.40.
4.6. EXERCISES 143
4.6 Exercises
Many of the exercises in this section are taken from Rudin [17, pages 7175].
Exercise 4.41. Let (X, A, ) be a measure space and let
f = (f1 , . . . , fn ) : X Rn
R
be a measurable function such that X |fi | d < for i = 1, . . . , n. Define
Z Z Z
f d := f1 d, . . . , fn d Rn .
X X X
Let Rn [0, ) : v 7 kvk be any norm on Rn . Prove that the function

X [0, ) : x 7 kf (x)k is integrable and
Z Z

f d kf k d. (4.49)

X X
Hint: Prove the inequality first for vector valued integrable step functions
s : X Rn . Show that for all R> 0 there is a vector valued
R integrable step
function s : X Rn such that k X (f s) dk < and X kf sk d < .
Exercise 4.42. Let (X, A, ) be a measure space such that (X) = 1. Let
f L1 () and let : R R be convex. Prove Jensens inequality
Z Z
f d ( f ) d. (4.50)
X X
(In particular, show that f is necessarily integrable so the right hand

side is well defined, even if f is not integrable.) Deduce that
Z Z
exp f d exp(f ) d. (4.51)
X X
Deduce also the inequality

n
X n
Y n
X
i = 1 = ai i i ai (4.52)
i=1 i=1 i=1
for all positive real numbers i and ai . In particular, ab ap /p + bq /q for all

positive real numbers a, b, p, q such that 1/p + 1/q = 1.
Exercise 4.43. Let (X, A, ) be a measure space, choose p, q, r [1, ]

such that
1 1 1
+ = ,
p q r
and let f L () and g L (). Prove that f g Lr () and
p q
kf gkr kf kp kgkq . (4.53)

Exercise 4.44. Let (X, A, ) be a measure space, choose real numbers
1 r < p < s < ,
and let 0 < < 1 such that
1 1
+ = .
r s p
Prove that every measurable function f : X R satisfies the inequality
kf kp kf kr kf ks1 . (4.54)
Deduce that Lr () Ls () Lp ().
Exercise 4.45. Let (X, A, ) be a measure space and let f : X R be a
measurable function. Define
If := {p R | 1 1.
Prove that f is continuous and that the restriction of f to the interior of If
is convex. Find examples where If is closed, where If is open, and where If
is a single point. If If 6= prove that
lim kf kp = kf k .
p
Exercise 4.46. For each of the following three conditions find an example
of measure space (X, A, ) that satisfies it for all p, q [1, ].
(a) If p < q then Lp () ( Lq ().
(b) If p < q then Lq () ( Lp ().
(c) If p 6= q then Lp () 6 Lq () and Lq () 6 Lq ().
4.6. EXERCISES 145
Exercise 4.47. Let (X, U) be a locally compact Hausdorff space and define

f is continuous and

C0 (X) := f : X R > 0 K X such that

K is compact and supX\K |f | <

Prove that X is a Banach space with respect to the sup-norm. Prove that
Cc (X) is dense in C0 (X).
Exercise 4.48. Let (X, A, ) be a measure space such that (X) = 1 and
let f, g : X [0, ] be measurable functions such that f g 1. Prove that
kf k1 kgk1 1.
Exercise 4.49. Let (X, A, ) be a measure space such that (X) = 1 and
let f : X [0, ] be a measurable function. Prove that
q Z p
2
1 + kf k1 1 + f 2 d 1 + kf k1 . (4.55)
X
Find a geometric interpretation of this inequality when is the restriction

of the Lebesgue measure to the unit interval X = [0, 1] and f = F 0 is the
derivative of a continuously differentiable function F : [0, 1] R. Under
which conditions does equality hold in either of the two inequalities in (4.55)?
Exercise 4.50. Let (X, A, ) be a measureRspace and let f : X R be

a measurable function such that f > 0 and X f d = 1. Let E X be a
measurable set such that 0 < (E) < . Prove that
Z
1
log(f ) d (E) log (4.56)
E (E)
and Z
f p d (E)1p for 0 < p < 1. (4.57)
E
Exercise 4.51. Let f : [0, 1] (0, ) be Lebesgue measurable. Prove that

Z 1 Z 1 Z 1
f (s) ds log(f (t)) dt f (x) log(f (x)) dx. (4.58)
0 0 0
Exercise 4.52. Fix two constants 1 0.

(i) Let f : (0, ) R be Lebesgue measurable and suppose that the function
(0, ) R : x 7 xp1a |f (x)|p is integrable. Show that the restriction of f
to each interval (0, x] is integrable and prove Hardys inequality
Z Z x p 1/p Z 1/p
1a
p p1a p
x f (t) dt dx
x |f (x)| dx . (4.59)
0

0 a 0
Show that equality holds in (4.59) if and only if f = 0 almost everywhere.

Hint: Assume Rfirst that f is nonnegative with compact support and de-
x
fine (x) := x1 0 f (t) dt
R Fp1a R for x > 0. p1 Use integration by parts to obtain
p p p1a
0
x F (x) dx = a 0 x F (x) f (x) dx. Use Holders inequality.
(ii) Show that the constant p/a in Hardys inequality is sharp. Hint: Choose
< 1 a/p and take f (x) := x for x 1 and f (x) := 0 for x > 1.
(iii) Prove that every sequence (an )nN of positive real numbers satisfies
N
!p p X
X 1 X p
an apn . (4.60)
N =1
N n=1
p 1 n=1
Hint: If an is nonincreasing then (4.60) follows from (4.59) with a = p 1

for a suitable function f . Deduce the general case from the special case.
(iv) Let f : (0, ) R be Lebesgue measurable and suppose that the
function (0, ) R : x 7 xp1+a |f (x)|p is integrable. Show that the
restriction of f to each interval [x, ) is integrable and prove the inequality
Z Z p 1/p Z 1/p

a1
p p1+a p
x f (t) dt dx x |f (x)| dx . (4.61)
0 x a 0
Hint: Apply the inequality (4.59) to the function g(x) := x2 f (x1 ).

Exercise 4.53. Let (X, U) be a locally compact Hausdorff space and let
: B [0, ] be an outer regular Borel measure on X that is inner regular
on open sets. Let g L1 (). Prove that the following are equivalent.
(i) The
R function g vanishes -almost everywhere.
(ii) X f g d = 0 for all f Cc (X).
Hint: Assume (ii). Let K X be compact. Use Urysohns Lemma A.1
to show that there is a sequence fn Cc (X) suchRthat 0 fn 1 and fn
converges
R almost everywhere to K . Deduce that R K g d = 0. Then prove
that U g d = 0 for every open set U X and B g d = 0 for all B B.
Warning: The regularity hypotheses on cannot be removed. Find an
example of a Borel measure where (ii) does not imply (i). (See Example 4.16.)
4.6. EXERCISES 147
Exercise 4.54. Prove Egoroff s Theorem: Let (X, A, ) be a measure

space such that (X) < and let fn : X R be a sequence of measurable
functions that converges pointwise to f : X R. Fix a constant > 0.
Then there exists a measurable set E A such that (X \ E) < and fn |E
converges uniformly to f |E . Hint: Define
S(k, n) := {x X | |fi (x) fj (x)| < 1/k i, j > n} for k, n N.
Prove that
lim (S(k, n)) = (X) for all k N.
n
T
Deduce that there is a sequence nk N such that E := kN S(k, nk ) satisfies
the required conditions. Show that Egoroffs theorem does not extend to -
finite measure spaces.
Exercise 4.55. Let (X, A, ) be a measure space and let 1 < p < . Let
f Lp () and let fn Lp () be a sequence such that limn kfn kp = kf kp
and fn converges to f almost everywhere. Prove that limn kf fn kp = 0.
Prove that the hypothesis limn kfn kp = kf kp cannot be removed.
Hint 1: Fix a constant > 0. Use Egoroffs Theorem R to construct disjoint
p
measurable sets A, B A such that X = A B, A |f | d < , (B) < ,
and fn convergesR to f uniformly on B. Use Fatous Lemma 1.41 to prove
that lim supn A |fn |p d < .
Hint 2: Let gn := 2p1 (|fn |p + |f |p ) |f fn |p and use Fatous Lemma 1.41
as in the proof of the Lebesgue Dominated Convergence Theorem 1.45.
Exercise 4.56. Let (X, A, ) be a measure space and let fn : X R be
a sequence of measurable functions and let f : X R be a measurable
function. The sequence fn is said to converge in measure to f if

lim x X |fn (x) f (x)| > = 0
n
for all > 0. (On page 47 this is called convergence in probability.) Assume
(X) < and prove the following.
(i) If fn converges to f almost everywhere then fn converges to f in measure.
Hint: See page 47.
(ii) If fn converges to f in measure then a subsequence of fn converges to f
almost everywhere.
(iii) If 1 p and fn , f Lp () satisfy limn kfn f kp = 0 then fn
converges to f in measure.
Exercise 4.57. Let (X, U) be a compact Hausdorff space and : B [0, ]

be a Borel measure. Let C(X) = Cc (X) be the space of continuous real
valued functions on X. Consider the following conditions.
(a) Every nonempty open subset of X has positive measure.
(b) There exists a Borel set E X and an element x0 X such that every
open neighborhood U of x0 satisfies (U E) > 0 and (U \ E) > 0.
(c) is outer regular and is inner regular on open sets.
(i) Assume (a). Then the map C(X) L () in (b) is an isometric em-
bedding and hence its image is a closed linear subspace of L ().
(ii) Assume (a) and (b). Then there is a nonzero bounded linear functional
: L () R that vanishes on the image of the inclusion C() L ().
Hint: If f = E almost everywhere then f is discontinuous at x0 .
(iii) Assume (a), (b), (c). Then the isometric embedding L1 () L ()
of Theorem 4.33 is not surjective. Hint: Use part (ii) and Exercise 4.53.
(iv) The Lebesgue measure on [0, 1] satisfies (a), (b), and (c).
Exercise 4.58. Prove that every -finite measure space (X, A, ) is localiz-
able. Hint: Assume first that (X) < . Let E A and define

c := sup (E1 En ) n N, E1 , . . . , En E .
that there is a sequence Ei E such that (
S
Show S i=1 Ei ) = c. Prove that

H := i=1 Ei is an envelope of E.
Exercise 4.59. Let (X, A, ) be a localizable measure space. Prove that it
satisfies the following.
(F) Let F be a collection of measurable functions f : Af R, each defined
on a measurable set Af A. Suppose that any two functions f1 , f2 F agree
almost everywhere on Af1 Af2 . Then there exists a measurable function
g : X R such that g|Af = f almost everywhere for all f F .
We will see in the next chapter that condition (F) is equivalent to localizabil-
ity for semi-finite measure spaces. Hint: Let F be a collection of measurable
functions as in (F). For a R and f F define Aaf := {x Af | f (x) < a} .
For q Q let H q A be an envelope of the collection E q := Aqf | f F .

Define the measurable sets

[
X a := Hq, a R.
qQ
q<a
4.6. EXERCISES 149

(i) If a < b then X a X b .
(ii) For every a R the measurable set X a is an envelope of the collection
E a := Aaf | f F .

Thus (Aaf \ X a ) = 0 for all f F and, if G A, then

(Aaf \ G) = 0 f F = (X a \ G) = 0.
(iii) If a R and E A then
(Aaf E) = 0 f F = (X a E) = 0.
(iv) (X a Af \ Aaf ) = 0 for all f F and all a R.
T r
S s

(v) Define E0 := rR X X \ sR X . Then E0 is measurable and
(Af E0 ) = 0 for all f F .
(vi) For f F define the measurable set Ef Af by
[ q [
Ef := (Af E0 ) (Af \ X q ) (X q Af \ Aqf ).
qQ qQ
Then (Ef ) = 0.
(vii) Define g : X R by

0, if x E0 ,
g(x) := (4.62)
a, if x X s for all s > a and x
/ X r for all r < a.
Then g is well defined and measurable and g = f on Af \ Ef for all f F .
Example 4.60. This example is closely related to Exercise 3.24, however,
it requires a considerable knowledge of Functional Analysis and the details
go much beyond the scope of the present manuscript. It introduces the

StoneCech compactification X of the natural numbers. This is a com-
pact Hausdorff space containing N and satisfying the universality property
that every continuous map from N to another compact Hausdorff space Y
extends uniquely to a continuous map from X to Y . The space C(X) of
continuous functions on X can be naturally identified with the space ` .
Hence the space of positive bounded linear functionals on ` is isomorphic
to the space of Radon measures on X by Theorem 3.15. Thus the Stone
Cech compactification of N can be used to understand the dual space of ` .

Moreover, it gives rise to an interesting example of a Radon measure which
is not outer regular (explained to me by Theo Buehler).
Consider the inclusion N (` ) : n 7 n which assigns to each nat-

ural number n N the bounded linear functional n : ` R defined by
n () := n for = (i )iN ` . This functional has norm one. Now the
space of all bounded linear functionals on ` of norm at most one, i.e. the
unit ball in (` ) , is compact with respect to the weak- topology by the
BanachAlaoglu theorem. Define X to be the closure of the set {n | n N}
in (` ) with respect to the weak- topology. Thus
1 `

For all finite sequences c , . . . , c R

1 1 ` `
and = ( ) , . . . , = ( ) `

i iN i iN

j j
X := (` ) satisfying ( ) < c for j = 1, . . . , ` .
there exists an n N such that

nj < cj for j = 1, . . . , `

The weak- topology U 2X is the smallest topology such that the map
f : X R, f () := (),
is continuous for each ` . The topological space (X, U) is a separable
compact Hausdorff space, called the StoneCech compactification of N.
It is not second countable and one can show that the complement of a point
in X that is not equal to one of the n is not -compact. The only continuous
functions on X are those of the form f , so the map ` C(X) : 7 f is
a Banach space isometry. (Verify that kf k := supX |f ()| = kk for all
` .) Thus the dual space of ` can be understood in terms of the Borel
measures on X.
By Theorem 3.18 every Radon measure on X is regular. However, the
Borel -algebra B 2X does carry -finite measures : B [0, ] that are
inner regular but not outer regular (and must necessarily satisfy (X) = ).
Here is an example pointed out to me by Theo Buehler. Define
X 1
(B) :=
nN
n
n B
for every Borel set B X. This measure is -finite and inner regular but
is not outer regular. (The set U := {n | n N} is open, its complement
K := X \ U is compact and has measure zero, and every open set containing
K misses only a finite subset of U and hence has infinite measure.) Now let
X0 X be the union of all open sets in X with finite measure. Then X0
is not -compact and the restriction of to the Borel -algebra of X0 is a
Radon measure but is not outer regular.
Chapter 5
The RadonNikod
ym Theorem
Recall from Theorem 1.40 that every measurable function f : X [0, )

on a measure Rspace (X, A, ) determines a measure f : A [0, ] defined
by f (A) := A f d for A A. By Theorem 1.35 it satisfies f (A) = 0
whenever (A) = 0. A measure with this property is called absolutely con-
tinuous with respect to . The RadonNikod ym Theorem asserts that, when
is -finite, every -finite measure that is absolutely continuous with re-
spect to has the form f for some measurable function f : X [0, ). It
was proved by Johann Radon in 1913 for the Lebesgue measure space and
extended by Otton Nikod ym in 1930 to general -finite measure spaces. A
proof is given in Section 5.1. Consequences of the RadonNikod ym Theorem
include the proof of Theorem 4.35 about the dual space of Lp () (Section 5.2)
and the decomposition theorems of Lebesgue, Hahn, and Jordan for signed
measures (Section 5.3). An extension of the RadonNikod ym Theorem to
general measure spaces is proved in Section 5.4.
5.1 Absolutely Continuous Measures

Definition 5.1. Let (X, A, ) be a measure space. A measure : A [0, )
is called absolutely continuous with respect to if
(A) = 0 = (A) = 0
for all A A. It is called singular with respect to if there exists a
measurable set A such that (A) = 0 and (Ac ) = 0. In this case we also say
that and are mutually singular. We write iff is absolutely
continuous with respect to and iff and are mutually singular.
151
152 THEOREM
CHAPTER 5. THE RADONNIKODYM
Lemma 5.2. Let (X, A) be a measurable space and let , , 1 , 2 be measures

on A. Then the following holds.
(i) If 1 and 2 then 1 + 2 .
(ii) If 1 and 2 then 1 + 2 .
(iii) If 1 and 2 then 1 2 .
(iv) If and then = 0.
Proof. We prove (i). Suppose that 1 and 2 . Then there exist
measurable sets Ai A such that i (Ai ) = 0 and (Aci ) = 0 for i = 1, 2.
Define A := A1 A2 . Then Ac = Ac1 Ac2 is a null set for and A is a null
set for both 1 and 2 and hence also for 1 + 2 . Thus 1 + 2 and this
proves (i).
We prove (ii). Suppose that 1 and 2 . If A A satisfies
(A) = 0 then 1 (A) = 2 (A) = 0 and so (1 + 2 )(A) = 1 (A) + 2 (A) = 0.
Thus 1 + 2 and this proves (ii).
We prove (iii). Suppose that 1 and 2 . Since 2 there
exists a measurable set A A such that 2 (A) = 0 and (Ac ) = 0. Since
1 it follows that 1 (Ac ) = 0 and hence 1 2 . This proves (iii).
We prove (iv). Suppose that and . Since there exists
a measurable set A A such that (A) = 0 and (Ac ) = 0. Since it
follows that (Ac ) = 0 and hence (X) = (A) + (Ac ) = 0. This proves (iv)
and Lemma 5.2.
Theorem 5.3 (Lebesgue Decomposition Theorem). Let (X, A, ) be a
-finite measure space and let be a -finite measure on A. Then there exist
unique measures a , s : A [0, ] such that
= a + s , a , s . (5.1)
Theorem 5.4 (RadonNikod ym). Let (X, A, ) be a -finite measure
space and let : A [0, ] be a measure. The following are equivalent.
(i) is -finite and absolutely continuous with respect to .
(ii) There exists a measurable function f : X [0, ) such that
Z
(A) = f d for all A A. (5.2)
A
If (i) holds then equation (5.2) determines f uniquely up to equality -almost

everywhere. Moreover, f L1 () if and only if (X) < .
5.1. ABSOLUTELY CONTINUOUS MEASURES 153
Proof. The last assertion follows by taking A = X in (5.2).

We prove that (ii) implies (i). Thus assume that there exists a measurable
function f : X [0, ) such that is given by (5.2). Then is absolutely
continuous with respect to by Theorem 1.35. Since is -finite, there exists
a sequence S of measurable sets X1 X2 X3 such that (Xn ) <
and X = n=1 Xn . Define An := {x Xn | f S(x) n}. Then An An+1 and

(An ) n(Xn ) < for all n and X = n=1 An . Thus is -finite and
this shows that (ii) implies (i).
It remains to prove that (i) implies (ii) and that f is uniquely determined
by (5.2) up to equality -almost everywhere. This is proved in three steps.
The first step is uniqueness, the second step is existence under the assumption
(X) < and (X) < , and the last step establishes existence in general.
Step 1. Let (X, A, ) be a measure space, let : A [0, ] be a -finite
measure, and let f, g : X [0, ) be two measurable functions such that
Z Z
(A) = f d = g d for all A A. (5.3)
A A
Then f and g agree -almost everywhere.

Since (X, A, ) is a -finite measure space there exists a sequence of mea-
S sets A1 A2 A3 such that (An ) < for all n N and
surable
X= n=1 An . For n N define
An := {E A | E An } , n := |An .
Take A = An in (5.3) to obtain f, g L1 (n ) for all n. Thus

Z
1
f g L (n ), (f g) dn = 0 for all E An .
E
Hence f g vanishes n -almost everywhere by Lemma 1.49. Thus the set
En := {x An | f (x) 6= g(x)}
satisfies (En ) = n (En ) = 0 and hence the set

[
E := {x X | f (x) 6= g(x)} = En
n=1
satisfies (E) = 0. This proves Step 1.

154 THEOREM
Step 2. Let (X, A) be a measurable space and let , : A [0, ] be

measures such that (X) < , (X) < , and . R Then there exists a
measurable function h : X [0, ) such that (A) = A h d for all A A.
By assumption + : A [0, ] is a finite measure defined by
( + )(A) := (A) + (A) for A A.
Since ( + )(X) < it follows from the CauchySchwarz inequality that
H := L2 ( + ) L1 ( + ).
Namely, if f L2 ( + ) then
Z sZ
p
|f | d( + ) c |f |2 d( + ) < , c := (X) + (X).
X X
Define : L2 ( + ) R by
Z
(f ) := f d.
X
for f L2 ( + ). (Here we abuse notation and use the same letter f for a
function in L2 ( + ) and its equivalence class in L2 ( + ).) Then
Z Z
|(f )| |f | d |f | d( + ) c kf kL2 (+)
X X
for all f L2 ( + ). Thus is a bounded linear functional on L2 ( + ) and

it follows from Corollary 4.28 that there exists an L2 -function g L2 ( + )
such that Z Z
f d = f g d( + ) (5.4)
X X
2
for all f L ( + ). This implies
Z Z Z
f (1 g) d( + ) = f d( + ) f g d( + )
X ZX ZX
= f d( + ) f d (5.5)
ZX X
= f d
X
for all f L2 ( + ).
We claim that the inequalities 0 g < 1 hold ( + )-almost everywhere.

To see this, consider the measurable sets

E0 := x X g(x) < 0 , E1 := x X g(x) 1 .
Then it follows from (5.4) with f := E0 that
Z Z
0 (E0 ) = E0 d = E0 g d( + ) 0.
X X
R
Hence X E0 g d(+) = 0 and it follows from Lemma 1.49 that the function
f := E0 g vanishes ( + )-almost everywhere. Hence ( + )(E0 ) = 0.
Likewise, it follows from (5.5) with f := E1 that
Z Z
(E1 ) = E1 d = (1 g) d( + ) 0.
X E1
Hence (E1 ) = 0. Since is absolutely continuous with respect to it follows

that (E1 ) = 0 and hence ( + )(E1 ) = 0 as claimed. Assume from now on
that 0 g(x) < 1 for all x X. (Namely, redefine g(x) := 0 for x E0 E1
without changing the identities (5.4) and (5.5).)
Apply equation (5.5) to the characteristic function f := A L2 ( + )
of a measurable set A to obtain the identity
Z
(A) = (1 g) d( + ) for all A A.
A
By Theorem 1.40 this implies that equation (5.5) continues to hold for every
measurable function f : X [0, ), whether or not it belongs to L2 ( + ).
Now define the measurable function h : X [0, ) by
g(x)
h(x) := for x X.
1 g(x)
By equation (5.4) with f = A and equation (5.5) with f = A h it satisfies
Z Z
(A) = A d = A g d( + )
ZX X Z
= A h(1 g) d( + ) = A h d
ZX X
= h d
A
for all A A. This proves Step 2.

156 THEOREM
Step 3. We prove that (i) implies (ii).

Since and are -finite measures, there exist sequences of measurable sets
An , Bn A such
S that An An+1 , (An ) < , Bn Bn+1 , (Bn ) < for
all n and X =
S
A
n=1 n = n=1 Bn . Define Xn := An Bn . Then
Xn Xn+1 , (Xn ) < , (Xn ) <

S
for all n and X = n=1 Xn . Thus it follows from Step 2 that there exists a
sequence of measurable functions fn : Xn [0, ) such that
Z
(A) = fn d for all n N and all A A such that A Xn . (5.6)
A
It follows from Step 1 that the restriction of fn+1 to Xn agrees with fn
-almost everywhere. Thus, modifying fn+1 on a set of measure zero if
necessary, we may assume without loss of generality that fn+1 |Xn = fn for
all n N. With this understood, define f : X [0, ) by
f |Xn := fn for n N.
This function is measurable because
[
1 1
[
f ([0, c]) = Xn f ([0, c]) = fn1 ([0, c]) A
n=1 n=1
for all c 0. Now let E A and define En := E Xn A for n N. Then

[
E1 E2 E3 , E= En .
n=1
Hence it follows from part (iv) of Theorem 1.28 that

(E) = lim (En )
n
Z
= lim f d
n E
n
Z
= lim En f d
n X
Z
= E f d
ZX
= f d.
E
Here the last but one equation follows from the Lebesgue Monotone Conver-
gence Theorem 1.37. This proves Step 3 and Theorem 5.4.
Example 5.5. Let X be a one element set and let A := 2X . Define the
measure : 2X [0, ] by () := 0 and (X) := .
(i) Choose () := 0 and (X) := 1. Then R but there does not exist a
(measurable) function f : X [0, ] such that X f d = (X). Thus the
hypothesis that (X, A, ) is -finite cannot be removed in Theorem 5.4.
R
(ii) Choose := . Then (A) = A f d for every nonzero function
f : X [0, ). Thus the hypothesis that (X, A, ) is -finite cannot be
removed in Step 1 in the proof of Theorem 5.4.
Example 5.6. Let X be an uncountable set and denote by A 2X the

set of all subsets A X such that either A or Ac is countable. Choose
an uncountable subset H X with an uncountable complement and define
, , : A [0, ] by

0, if A is countable,
(A) := (A) := #(A H), (A) := #A.
1, if Ac is countable,
Then and and are not -finite. There R does not exist any
measurable function f : X [0, ] such that (X) = R X f d. Nor is there
any measurable function h : X R such that (A) = A h d for all A A.
(The only possible such function would be h := H which is not measurable.)
Proof of Theorem 5.3. We prove uniqueness. Let a , s , 0a , 0s : A [0, ]

be measures such that
= a + s = 0a + 0s , a , 0a , s , 0s .
Then there exist measurable sets A, A0 A such that
s (A) = 0, (X \ A) = 0, 0s (A0 ) = 0, (X \ A0 ) = 0.
Since X \ (A A0 ) = (X \ A) (X \ A0 ), this implies (X \ (A A0 )) = 0.

Let E A. Then s (E A A0 ) = 0 = 0s (E A A0 ) and hence
a (E A A0 ) = (E A A0 ) = 0a (E A A0 ).
Moreover (E \ (A A0 )) = 0, hence a (E \ (A A0 )) = 0 = 0a (E \ (A A0 ))
and hence
s (E \ (A A0 )) = (E \ (A A0 )) = 0s (E \ (A A0 )).
158 THEOREM
This implies
a (E) = a (E A A0 ) = 0a (E A A0 ) = 0a (E)
s (E) = s (E \ (A A0 )) = 0s (E \ (A A0 )) = 0s (E).
This proves uniqueness.
We prove existence. The measure
:= + : A [0, ]
is -finite. Hence it follows from the RadonNikod
ym Theorem 5.4 that there
exist measurable functions f, g : X [0, ) such that
Z Z
(E) = f d, (E) = g d for all E A. (5.7)
E E
Define
A := x X g(x) > 0 (5.8)
and
a (E) := (E A), s (E) := (E Ac ) for E A. (5.9)
Then it follows directly from (5.9) that the maps a , s : A [0, ] are
measures and satisfy a + s = . Moreover, it follows from (5.9) that
s (A) = (A Ac ) = () = 0
and from (5.8) that g|Ac = 0, so by (5.7)
Z
c
(A ) = g d = 0.
Ac
This shows that s . It remains to prove that a is absolutely continuous
with respect to . To see this, let E A such that (E) = 0. Then by (5.7)
Z Z
E g d = g d = (E) = 0.
X E
Hence it follows from Lemma 1.49 that E g vanishes -almost everywhere.
Thus EA g = A E g vanishes -almost everywhere. Since g(x) > 0 for all
x E A, this implies
(E A) = 0.
Hence Z
a (E) = (E A) = f d = 0.
EA
This shows that a and completes the proof of Theorem 5.3.
5.2. THE DUAL SPACE OF LP () REVISITED 159
5.2 The Dual Space of Lp() Revisited

This section is devoted to the proof of Theorem 4.35. Assume throughout
that (X, A, ) is a measure space and fix two constants
1 1
1 p < , 1 < q , + = 1. (5.10)
p q
As in Section 4.5 we abuse notation and write (f ) := ([f ] ) for the value
of a bounded linear functional : Lp () R on the equivalence class
of a function f Lp (). Recall from Theorem 4.33 that every g Lq ()
determines a bounded linear functional g : Lp () R via
Z
g (f ) := f g d for f Lp ().
X
The next result proves Theorem 4.35 in -finite case.

Theorem 5.7. Assume (X, A, ) is -finite and let : Lp () R be a
bounded linear functional. Then there exists a function g Lq () such that
g = .
Proof. Assume first that is positive. We prove in six steps that there exists
a function g Lq () such that g 0 and g = .
Step 1. Define

(A) := sup (E ) E A, E A, (E) < (5.11)
for A RA. Then the map : A [0, ] is a measure, Lp () L1 (), and
(f ) = X f d for all f Lp ().
This follows directly from Theorem 4.40.
Step 2. Let be as in Step 1 and define c := kk. Then (A) c(A)1/p
for all A A.
By assumption (f ) c kf kp for all f Lp (). Take f := E to obtain
(E ) c(E)1/p c(A)1/p for all E A with E A and (E) < .
Take the supremum over all such E to obtain (A) c(A)1/p by (5.11).
Step 3. Let be as in Step 1. R Then there exists a measurable function
g : X [0, ) such that (A) = A g d for all A A.
By Step 2, is -finite and . Hence Step 3 follows from the Radon
Nikod
ym Theorem 5.4 for -finite measure spaces.
160 THEOREM
R R
Step 4. Let be as in Step 1 and g be as in Step 3. Then X
f g d = X
f d
for every measurable function f : X [0, ).
This follows immediatey from Step 3 and Theorem 1.40.
Step 5. Let c be as in Step 2 and g be as in Step 3. Then kgkq c.
Let be as in Step 1 and let f Lp () such that f 0. Then
Z Z
Step 4 Step 1
f g d = f d = (f ) c kf kp . (5.12)
X X
Moreover, the measure space (X, A, ) is semi-finite by Lemma 4.30. Hence

it follows from parts (iii) and (iv) of Lemma 4.34 that kgkq c.
Step 6. Let g be as in Step 3. Then = g .
Since g Lq () by Step 5, the function R g determines a pbounded linear
p
functional g : L () R via g (f ) := X f g d for f L (). By (5.12)
it satisfies g (f ) = (f ) for all f Lp () with f 0. Apply this identity
to the functions f : X [0, ) for all f Lp () to obtain = g .
This proves the assertion of Theorem 5.7 for every positive bounded linear
functional : Lp () R.
Let : Lp () R be any bounded linear functional. By Theorem 4.39
there exist positive bounded linear functionals : Lp () R such that
= + . Hence, by what we have just proved, there exist functions
g Lq () such that g 0 and = g . Define g := g + g . Then
g Lq () and g = g+ g = + = . This proves Theorem 5.7.
The next result proves Theorem 4.35 in the case p = 1.
Theorem 5.8. Assume p = 1. Then the following are equivalent.
(i) The measure space (X, A, ) is localizable.
(ii) The measure space (X, A, ) is semi-finite and satisfies condition (F) in
Exercise 4.59, i.e. if F is a collection of measurable functions f : Af R,
each defined on a measurable set Af A, such that any two functions
f1 , f2 F agree almost everywhere on Af1 Af2 , then there exists a measur-
able function g : X R such that g|Af = f almost everywhere for all f F .
(iii) The linear map
L () L1 () : g 7 g (5.13)
is bijective.
Proof. The proof that (i) implies (ii) is outlined in Exercise 4.59.
We prove that (ii) implies (iii). Since (X, A, ) is semi-finite, the linear
map (5.13) is injective by Theorem 4.33. We must prove that it is surjective.
Assume firstthat : L1 () R is a positive bounded linear functional.
Define E := E A (E) < and, for E E, define
AE := {A A | A E} , E := |AE . (5.14)
Then there is an extension operator E : L1 (E ) L1 () defined by

f (x), for x E,
E (f )(x) := (5.15)
0, for x X \ E,
It descends to a bounded linear operator from L1 (E ) to L1 () which will

still be denoted by E . Define
E = E : L1 (E ) R.
This is a positive bounded linear functional for every E E. Hence it follows

from Theorem 5.7 (and the Axiom of Choice) that there is a collection of
bounded measurable functions gE : E [0, ), E E, such that
Z
E (f ) = f gE dE for all E E and all f L1 (E ).
E
If E, F E then E F E and the functions gE |EF , gF |EF , and gEF

all represent the same bounded linear functional EF : L1 (EF ) R.
Hence they agree almost everywhere by Theorem 4.33. This shows that the
collection
F := gE E E

satisfies the hypotheses of condition (F) on page 148. Thus it follows from (ii)
that there exists a measurable function g : X R such that, for all E E,
the restriction g|E agrees with gE almost everywhere on E.
We prove that g 0 almost everywhere. Suppose otherwise that the
set A := {x X | g(x) < 0} has positive measure. Since (X, A, ) is semi-
finite there exists a set E E such that E A and (E) > 0. Since
g(x) < 0 gE (x) for all x E it follows that g|E does not agree with gE
almost everywhere, a contradiction. This contradiction shows that g 0
almost everywhere.
162 THEOREM
We prove that g kk almost everywhere. Suppose otherwise that the

set A+ := {x X | g(x) > kk} has positive measure. Since (X, A, ) is
semi-finite there exists a set E E such that E A+ and (E) > 0. Since
kgE k = kE k kk it follows from Lemma 4.8 that gE (x) kk < g(x)
for almost every x E. Hence g|E does not agree with gE almost everywhere,
a contradiction. This contradiction shows that g kk almost everywhere
and we may assume without loss of generality that
0 g(x) kk
for all x X.
We prove that g = . Fix a function f L1 () such that f 0. Then
there exists a sequence Ei E such that E1 E2 E3 and Ei f
converges pointwise to f . Namely, by Theorem 1.26 there exists a sequence
of measurable step functions si : X R [0, ) such R that 0 s1 s2
and si converges pointwise to f . Since X si d X f d < for all i the
sets Ei := {x X | si (x) > 0} have finite measure and 0 si Ei f f for
all i. Thus the Ei are as required. Since the sequence |f Ei f | converges
pointwise to zero and is bounded above by the integrable function f it follows
from the Lebesgue Dominated Convergence Theorem 1.45 that
lim kf Ei f k1 = 0.
i
Hence
(f ) = lim (Ei f ) = lim Ei (f |Ei )
i i
Z Z Z
= lim f gEi d = lim f g d = f g d.
i Ei i Ei X
Here the last step follows from the Lebesgue Monotone Convergence Theo-
rem 1.37. This shows that (f ) = g (f ) for every nonnegative integrable
function f : X [0, ). It follows that
(f ) = (f + ) (f ) = g (f + ) g (f ) = g (f )
for all f L1 (). Thus = g as claimed.
This shows that every positive bounded linear functional on L1 () be-
longs to the image of the map (5.13). Since every bounded linear functional
on L1 () is the difference of two positive bounded linear functionals by The-
orem 4.39, it follows that the map (5.13) is surjective. Thus we have proved
that (ii) implies (iii).
We prove that (iii) implies (i). Assume that the map (5.13) is bijective.
Then (X, A, ) is semi-finite by part (iv) of Theorem 4.33. Now let E A
be any collection of measurable sets. Assume without loss of generality that
E1 , . . . , E` E = E1 E` E.
(Otherwise, replace E by the collection E 0 of all finite unions of elements
of E; then every measurable envelope of E 0 is also an envelope of E.) For
E E define AE and E by (5.14) and define the bounded linear functional
E : L1 (E ) R by
Z
E (f ) := f dE for f L1 (E ). (5.16)
E
Then for all E, F A and f L1 ()

E F, f 0 = E (f ) F (f ). (5.17)
Define : L1 () R by
(f ) := sup E (f + |E ) sup E (f |E ). (5.18)
EE EE
We prove that this is a well defined bounded

R linear functional with kk 1.
To see this, note that E (fR|E ) X fRd for every nonnegative function
f L1 () and so |(f )| X f + d + X f d = kf k1 for all f L1 ().
Moreover, it follows directly from the definition that (cf ) = c(f ) for
all c 0 and (f ) = (f ). Now let f, g L1 () be nonnegative inte-
grable functions. Then
(f + g) = sup E (f |E + g|E )
EE
sup E (f |E ) + sup E (g|E )
EE EE
= (f ) + (g).
To prove the converse inequality, let > 0 and choose E, F E such that
E (f |E ) > (f ) , F (g|F ) > (g) .
Then E F E and it follows from (5.17) that
EF ((f + g)|EF ) = EF (f |EF ) + EF (g|EF )
E (f |E ) + F (g|F )
> (f ) + (g) 2.
164 THEOREM
Hence (f + g) > (f ) + (g) 2 for all > 0 and so

(f ) + (g) (f + g) (f ) + (g).
This shows that (f +g) = (f )+(g) for all f, g L1 () such that f, g 0.
If f, g L1 () then (f + g)+ + f + g = (f + g) + f + + g + and hence
((f + g)+ ) + (f ) + (g ) = ((f + g) ) + (f + ) + (g + )
by what we have just proved. Since (f ) = (f + ) (f ) by definition it
follows that (f + g) = (f ) + (g) for all f, g L1 (). This shows that
: L1 R is a positive bounded linear functional of norm kk 1.
With this understood, it follows from (iii) that there exists a function
g L () such that = g . Define
H := {x X | g(x) > 0} .
We prove that H is an envelope of E. Fix a set E E and suppose, by
contradiction, that (E \ H) > 0. Then, since (X, A, ) is semi-finite, there
exists a measurable set A A such that 0 < (A) < and A E \ H.
Since g(x) 0 for all x A it follows that
Z
0 < (A) = E (A |E ) = g d 0,
A
a contradiction. This contradiction shows that our assumption (E \ H) > 0

must have been wrong. Hence (E \ H) = 0 for all E E as claimed.
Now let G A be any measurable set such that (E \ G) = 0 for all
E E. We must prove that (H \ G) = 0. Suppose, by contradiction, that
(H \ G) > 0. Since (X, A, ) is semi-finite there exists a measurable set
A A such that 0 < (A) < and A H \ G. Then
Z
g d = (A ) = sup E (A |E ) = sup (E A) = 0.
A EE EE
Here the second equation follows from (5.18), the third follows from (5.16),
and the last follows from the fact that E A E \ G for all E E. Since
g > 0 on A it follows from Lemma 1.49 that (A) = 0, a contradiction. This
contradiction shows that our assumption that (H \ G) > 0 must have been
wrong and so (H \ G) = 0 as claimed. Thus we have proved that every
collection of measurable sets E A has a measurable envelope, and this
completes the proof of Theorem 5.8.
Now we are in a position to prove Theorem 4.35 in general.

Proof of Theorem 4.35. For p = 1 the assertion of Theorem 4.35 follows from
the equivalence of (i) and (iii) in Theorem 5.8. Hence assume p > 1. We
must prove that the linear map Lq () Lp () : g 7 g is surjective. Let
: Lp () R be a positive bounded linear functional and define
(A) := sup {(E ) | E A, E A, (E) < } for A A.
Then : A [0, ] is a measure by Theorem 4.40 and
Z
p 1
L () L (), (f ) = f d for all f Lp ().
X
Theorem 4.40 also asserts that there exists a measurable set N A such
that (N ) = 0 and the restriction of to X \ N is -finite. Define
X0 := X \ N, A0 := {A A | A X0 } , 0 := |A0
as in (5.14), let 0 : Lp (0 ) Lp () be the extension operator as in (5.15),
and define 0 := 0 : Lp (0 ) R. Then 0 is a positive bounded linear
functional on Lp (0 ) and
Z Z
(f ) = f d = f d = 0 (f |X0 ) for all f Lp ().
X X\N
Since (X0 , A0 , 0 ) is -finite it follows from Theorem 5.7 that there exists a
function g0 Lq (0 ) such that g0 0 and
Z
0 (f0 ) = f0 g0 d0 for all f0 Lp (0 ).
X0
Define g : X [0, ) by g(x) := g0 (x) for x X0 = X \ N and g(x) := 0

for x N . Then kgkLq () = kg0 kLq (0 ) = k0 k = kk by Theorem 4.33, and
R R
(f ) = 0 (f |X0 ) = X0 f g0 d0 = X f g d for all f Lp (). This proves
the assertion for positive bounded linear functionals. Since every bounded
linear functional : Lp () R is the difference of two positive bounded
linear functionals by Theorem 4.39, this proves Theorem 4.35.
Corollary 5.9. Every -finite measure space is localizable.
Proof. Let (X, A, ) be a -finite measure space. Then (X, A, ) is semi-
finite by Lemma 4.30. Hence the map L () L1 () : g 7 g in (4.31) is
injective by Theorem 4.33 and is surjective by Theorem 5.7. Hence it follows
from Theorem 5.8 that (X, A, ) is localizable.
166 THEOREM
5.3 Signed Measures

Throughout this section (X, A) is a measurable space, i.e. X is a set and
A 2X is a -algebra. The following definition extends the notion of a
measure on (X, A) to a signed measure which can have positive and negative
values. As an example from physics one can think of electrical charge.
Definition 5.10. A function : A R is called a signed measure if it is
-additive, i.e. every sequence Ei A of pairwise disjoint measurable sets
satisfies

!
X [ X
|(Ei )| < , Ei = (Ei ). (5.19)
i=1 i=1 i=1
Lemma 5.11. Every signed measure : A R satisfies the following.

(i) () = 0.
(ii) If E1 , . . . , E` A are pairwise disjoint then ( ì=1 Ei ) = ì=1 (Ei ).
S P
Proof. To prove (i) take Ei := in equation (5.19). To prove (ii) take Ei :=

for all i > `.
Given a signed measure : A R it is a natural question to ask whether
it can be written as the difference of two measures : A [0, ). Closely
related to this is the question whether there exists a measure : A [0, )
that satisfies
|(A)| (A) for all A A. (5.20)
If such a measure exists it must satisfy
E, F A, EF = = (E) (F ) (E F )
Thus a lower bound for (A) is the supremum of the numbers (E) (F )
over all decompositions of A into pairwise disjoint measurable sets E and F .
The next theorem shows that this supremum defines the smallest measure
that satisfies (5.20).
Theorem 5.12. Let : A R be a signed measure and define

E, F A,

||(A) := sup (E) (F ) E F = , for A A. (5.21)

EF =A

Then |(A)| ||(A) < for all A A and || : A [0, ) is a measure,

called the total variation of .
5.3. SIGNED MEASURES 167
Proof. We prove that || is a measure. If follows directly from the definition

that ||() = 0 and ||(A) |(A)| 0 for all A A. We must prove that
the function || : A [0, ] is -additive. Let Ai A be a sequence of
pairwise disjoint measurable sets and define

[
A := Ai .
i=1
Let E, F A are measurable sets such that

E F = , E F = A. (5.22)
Then
[ [
E= (E Ai ), F = (F Ai ).
i=1 i=1
Hence

X
X
(E) (F ) = (E Ai ) (F Ai )
i=1 i=1
X
= (E Ai ) (F Ai )
i=1

X
||(Ai ).
i=1
Take the supremum over all pairs of measurable sets E, F satisfying (5.22)
to obtain
X
||(A) ||(Ai ) (5.23)
i=1
To prove the converse inequality, fix a constant > 0. Then there are
sequences of measurable sets Ei , Fi A such that

Ei Fi = , Ei Fi = Ai , (Ei ) (Fi ) > ||(Ai ) i
2
S S
for all i N. The sets E := i=1 Ei and F := i=1 Fi satisfy (5.22) and so

X X
||(A) (E) (F ) = (Ei ) (Fi ) > ||(Ai ) .
i=1 i=1
Hence ||(A) > P

P P
i=1 ||(Ai ) for all > 0. Thus ||(A) i=1 ||(Ai )

and so ||(A) = i=1 ||(Ai ) by (5.23). This shows that || is a measure.
168 THEOREM
It remains to prove that ||(X) < . Suppose, by contradiction, that

||(X) = . We prove the following.
Claim. Let A A such that ||(X \A) = . Then there exists a measurable
set B A such that A B, |(B \ A)| 1, and ||(X \ B) = .
There exist measurable sets E, F such that E F = , E F = X \ A, and
(E) (F ) 2 + |(X \ A)|,
(E) + (F ) = (X \ A).
Take the sum, respectively the difference, of these (in)equalities to obtain
2(E) 2 + |(X \ A)| + (X \ A) 2,
2(F ) (X \ A) 2 |(X \ A)| 2,
and hence |(E)| 1 and |(F )| 1. Since ||(E)+||(F ) = ||(X \A) =
it follows that ||(E) = or ||(F ) = . If ||(E) = choose B := A F
and if ||(F ) = choose B := A E. This proves the claim.
It follows from the claim by induction that there exists a sequence of
measurable sets := A0 A1 A2 such that |(An \ An1 )| 1 for
all n N. Hence PE n := An \An1 is a sequence of pairwise disjoint measurable

sets such that n=1 |(En )| = , in contradiction to Definition 5.10. This
contradiction shows that the assumption that ||(X) = must have been
wrong. Hence ||(X) < and thus ||(A) < for all A A. This proves
Theorem 5.12.
Definition 5.13. Let : A R be a signed measure and let || : A [0, )
the measure in Theorem 5.12. The Jordan decomposition of is the
representation of as the difference of two measures whose sum is equal
to ||. The measures : A [0, ) are defined by
||(A) (A)
(A) := = sup {(E) | E A, E A} (5.24)
2
for A A and they satisfy
+ = , + + = ||. (5.25)
ExerciseR 5.14. Let (X, A, ) be a measure space, let f L1 (), and define
(A) := A f d for A A. Prove that is a signed measure and
Z Z

||(A) = |f | d, (A) = f d for all A A. (5.26)
A A
Definition 5.15. Let : A [0, ] be a measure and let , 1 , 2 : A R

be signed measures.
(i) is called absolutely continuous with respect to (notation )
if (E) = 0 implies (E) = 0 for all E A.
(iii) is called concentrated on A A if (E) = (E A) for all E A.
(iii) is called singular with respect to (notation ) if there exists
a measurable set A such that (A) = 0 and is concentrated on A.
(iv) 1 and 2 are called mutually singular (notation 1 2 ) if there
are measurable sets A1 , A2 such that A1 A2 = , A1 A2 = X, and i is
concentrated on Ai for i = 1, 2.
Lemma 5.16. Let be a measure on A and let , 1 , 2 be signed measures
on A. Then the following holds.
(i) if and only if || .
(ii) if and only if || .
(iii) 1 2 if and only if |1 | |2 |.
Proof. The proof has four steps.
Step 1. Let A A. Then ||(A) = 0 if and only if (E) = 0 for all
measurable sets E A.
If ||(A) = 0 then |(E)| ||(E) ||(A) = 0 for all measurable sets
E A. The converse implication follows directly from the definition.
Step 2. is concentrated on A A if and only if ||(X \ A) = 0.
The signed measure is concentrated on A if and only if (E) = (E A)
for all E A, or equivalently (E \ A) = 0 for all E A. By Step 1 this
holds if and only if ||(X \ A) = 0.
Step 3. We prove (i).
Assume . If E A satisfies (E) = 0 then every measurable set
F A with F E satisfies (F ) = 0 and hence (F ) = 0; hence ||(E) = 0
by Step 1. Thus || . The converse follows from the fact that ||.
Step 4. We prove (ii) and (iii).
if and only if there is a measurable sets A A such that (A) = 0 and
is concentrated on A. By Step 2 the latter holds if and only if ||(X \A) = 0
or, equivalently, || . This proves (ii). Assertion (iii) follows from Step 2
by the same argument and this proves Lemma 5.16.
170 THEOREM
Theorem 5.17 (Lebesgue Decomposition). Let (X, A, ) be a -finite

measure space and let : A R be a signed measure. Then there exists a
unique pair of signed measures a , s : A R such that
= a + s , a , s . (5.27)
Proof. We prove existence. Let : A [0, ) be the measures defined

by (5.24). By Theorem 5.3 there exist measures a : A [0, ) and
s : A [0, ) such that
a ,
s , and
=
a + s . Hence the

signed measures a := + +
a a and s := s s satisfy (5.27).
We prove uniqueness. Assume = a + s = 0a + 0s where a , s , 0a , 0s
are signed measures on A such that a , 0a and s , 0s . Then
|a |, |0a | and |s |, |0s | by Lemma 5.16. This implies |a | + |0a |
and |s | + |0s | by parts (i) and (ii) of Lemma 5.2. Moreover,
|a 0a | |a | + |0a |, |0a a | = |s 0s | |s | + |0s |.
Hence |a 0a | and |a 0a | by part (iii) of Lemma 5.2. Thus

|a 0a | = 0 by part (iv) of Lemma 5.2 and therefore a = 0a and s = 0s .
Theorem 5.18 (RadonNikod ym). Let (X, A, ) be a -finite measure

space and let : A R be a signed measure. Then if and only if
there exists a -integrable function f : X R such that
Z
(A) = f d for all A A. (5.28)
A
f is determined uniquely by (5.28) up to equality -almost everywhere.
Proof. If is given by (5.28) for some f L1 () then by part (vi) of

Theorem 1.44. Conversely, assume and let ||, + , : A [0, )
be the measures defined by (5.21) and (5.24). Then || by part (i) of
Lemma 5.16 and so . Hence it follows from Theorem 5.4R that there ex-
ist -integrable functions f : A [0, ) such that (A) = A f d for all
A A. Hence the function f := f + f L1 () satisfies (5.28). The unique-
ness of f , up to equality -almost everywhere, follows from Lemma 1.49. This
proves Theorem 5.18.
Theorem 5.19 (Hahn Decomposition). Let : A R be a signed

measure. Then there exists a measurable set P A such that
(A P ) 0, (A \ P ) 0 for all A A. (5.29)
Moreover, there exists a measurable function h : X {1, 1} such that
Z
(A) = h d|| for all A A. (5.30)
A
Proof. By Theorem 5.12 the function := || : A [0, ) in (5.21) is a

finite measure and satisfies |(A)| (A) for all A A. Hence and
it follows from Theorem 5.18 that there exists a function h L1 () such
that (5.30) holds. We prove that h(x) {1, 1} for -almost every x X.
To see this, fix a real number 0 < r < 1 and define

Ar := x X |h(x)| r .
If E, F A such that E F = and E F = Ar then
Z Z Z Z
(E) (F ) = h d h d |h| d + |h| d r(Ar )
E F E F
Take the supremum over all pairs E, F A such that E F = and

E F = Ar to obtain (Ar ) r(Ar ) and hence (Ar ) = 0. Since this holds
for all r < 1 it follows that |h| 1 -almost everywhere. Modifying h on a
set of measure zero, if necessary, we may assume without loss of generality
that |h(x)| 1 for all x X. Define

P := x X h(x) 1 , N := x X h(x) 1 .
Then P N = , P N = X, and
Z Z
(P ) h d = (P ) (P ), (N ) (N ) = h d (N ).
P N
Hence
Z Z
(h 1) d = (P ) (P ) = 0, (h + 1) d = (N ) + (N ) = 0.
P N
By Lemma 1.49 this implies h = 1 -almost everywhere on P and h = 1

-almost everywhere on N . Modify h again on a set of measure zero, if
necessary, to obtain h(x) = 1 for all x P and h(x) = 1 for all x N .
172 THEOREM
Theorem 5.20 (Jordan Decomposition). Let (X, A) be a measurable

space, let : A R be a signed measure, and let : A [0, ) be finite
measures such that = + . Then the following are equivalent.
(i) + + = ||.
(ii) + .
(iii) There exists a measurable set P A such that + (A) = (A P ) and
(A) = (A \ P ) for all A A.
Moreover, for every signed measure , there is a unique pair of measures
satisfying = + and these equivalent conditions.
Proof. We prove that (i) implies (ii). By Theorem 5.19 R there exists a mea-
surable function h : X {1} such that (A) = A h d|| for all A A.
Define P := {x X | h(x) = 1}. Then it follows from (i) that
||(P c ) + (P c )
Z
+ c 1+h
(P ) = = d|| = 0,
2 Pc 2
||(P ) (P ) 1h
Z

(P ) = = d|| = 0.
2 P 2
Hence + .
We prove that (ii) implies (iii). By (ii) there exists a measurable set
P A such that + (P c ) = 0 and (P ) = 0. Hence
+ (A) = + (A P ) = + (A P ) (A P ) = (A P ),
(A) = (A \ P ) = (A \ P ) + (A \ P ) = (A \ P )
for all A A.
We prove that (iii) implies (i). Assume (iii) and fix a set A A. Then
+ (A) + (A) = (A P ) (A \ P ) ||(A).
Now choose E, F A such that E F = and E F = A. Then
(E) (F ) = (E P ) + (E \ P ) (F P ) (F \ P )
(E P ) (E \ P ) + (F P ) (F \ P )
= (A P ) (A \ P ) = + (A) + (A).
Take the supremum over all such pairs E, F A to obtain the inequality
||(A) + (A) + (A) for all A A and hence || = + + .
Thus we have proved that assertions (i), (ii), and (iii) are equivalent.
Existence and uniqueness of now follows from (i) with = 21 (|| ).
GENERALIZED
5.4. RADONNIKODYM 173
5.4 RadonNikod
ym Generalized
This section discusses an extension of the RadonNikod ym Theorem 5.18 for
signed measures to all measure spaces. Thus we drop the hypothesis that
is -finite. In this case Examples 5.5 and 5.6 show that absolute conti-
nuity of with respect to is not sufficient for obtaining the conclusion
of the RadonNikod ym Theorem and a stronger condition is needed. In [4,
Theorem 232B] Fremlin introduces the notion truly continuous, which is
equivalent to absolutely continuous whenever is -finite. In [8] Konig re-
formulates Fremlins criterion in terms of inner regularity of with respect
to . We shall discuss both conditions below, show that they are equiva-
lent, and prove the generalized RadonNikod ym Theorem. As a warmup we
rephrase absolute continuity in the familiar - language of analysis.
Standing Assumption. Throughout this section (X, A, ) is a measure
space and : A R is a signed measure.
Lemma 5.21 (Absolute Continuity). The following are equivalent.
(i) is absolutely continuous with respect to .
(ii) For every > 0 there exists a constant > 0 such that
A A, (A) < = |(A)| < .
Proof. That (ii) implies (i) is obvious. Conversely, assume (i). Then ||
by Lemma 5.16. Assume by contradiction that (ii) does not hold. Then there
exists a constant > 0 and a sequence of measurable sets Ai A such that
(Ai ) 2i , |(Ai )| for all i N.
For n N define

[
\
Bn := Ai , B := Bn .
i=n n=1
Then
1
Bn Bn+1 , (Bn ) , ||(Bn ) ||(An ) |(An )|
2n1
for all n N. Hence (B) = 0 and ||(B) = limn ||(Bn ) by part (v)
of Theorem 1.28. This contradicts the fact that || . This contradiction
shows that our assumption that (ii) does not hold must have been wrong.
Thus (i) implies (ii) and this proves Lemma 5.21.
174 THEOREM
Definition 5.22. The signed measure is called truly continuous with

respect to if, for every > 0, there exists a constant > 0 and a
measurable set E A such that (E) < and
A A, (A E) < = |(A)| < . (5.31)
Lemma 5.23. The following are equivalent.
(i) is truly continuous with respect to .
(ii) || is truly continuous with respect to .
(iii) + and are truly continuous with respect to .
Proof. Assume (i), fix a constant > 0, and choose > 0 and E A such
that (E) < and (5.31) holds. Let A A such that (A E) < . Then
(B) (A \ B) < 2 for every measurable set B A and hence ||(A) 2
by Theorem 5.12. This shows that (i) implies (ii). That (ii) implies (iii)
and (iii) implies (i) follows directly from the definitions.
Definition 5.22 is due to Fremlin [4, Chapter 23]. If the measure space
(X, A, ) is -finite then is truly continuous with respect to if and only if it
is absolutely continuous with respect to by Theorem 5.26 below. However,
for general measure spaces the condition of true continuity is stronger than
absolute continuity. The reader may verify that, when (X, A, ) and are as
in part (i) of Example 5.5 or Example 5.6, the finite measure is absolutely
continuous with respect to but is not truly continuous with respect to .
Fremlins condition was reformulated by Konig [8] in terms of inner regularity
of with respect to . This notion can be defined in several equivalent ways.
To formulate the conditions it is convenient to introduce the notation
E := {E A | (E) < } .
Lemma 5.24. The following are equivalent.
(i) For all A A
(A E) = 0 for all E E = (A) = 0.
(ii) For all A A
||(A E) = 0 for all E E = ||(A) = 0.
(iii) For all A A
||(A) = sup||(A E) = sup ||(E).
EE EE
EA
GENERALIZED
Proof. By Theorem 5.19 there exists a set P A such that
(A P ) 0, (A \ P ) 0, ||(P ) = (A P ) (A \ P ) (5.32)
for all A A. Such a measurable set P will be fixed throughout the proof.
We prove that (i) implies (ii). Fix a set A A such that ||(AE) = 0 for
all E E. Then it follows from (5.32) that (A E P ) = (A E \ P ) = 0
for all E E. By (i) this implies (A P ) = (A \ P ) = 0 and hence
||(A) = 0 by (5.32). This shows that (i) implies (ii).
We prove that (ii) implies (i). Fix a set A A such that (A E) = 0
for all E E. Since E P E and E \ P E for all E E this implies
(A E P ) = (A E \ P ) = 0 for all E E. Hence it follows from (5.32)
that ||(A E) = 0 for all E E. By (ii) this implies ||(A) = 0 and hence
(A) = 0 because |(A)| ||(A). This shows that (ii) implies (i).
We prove that (ii) implies (iii). Fix a set A A and define
c := sup ||(E) ||(A). (5.33)

EE
EA
Choose a sequence Ei E such that Ei A for all i and limi ||(Ei ) = c.

For i N define Fi := E1 E2 Ei . Then
Fi E, Fi Fi+1 A, ||(Ei ) ||(Fi ) c (5.34)
for all i and hence

lim ||(Fi ) = c. (5.35)
i
Define
[
B := A \ F, F := Fi . (5.36)
i=1
Then ||(F ) = limi ||(Fi ) = c by part (iv) of Theorem 1.28 and hence
||(B) = ||(A) ||(F ) = ||(A) c. (5.37)
Let E E such that E B. Then E Fi = , E Fi E, and E Fi A

for all i by (5.36). Hence ||(E) + ||(Fi ) =
||(E Fi ) c for all i by (5.33).
This implies ||(E) limi c ||(Fi ) = 0 by (5.35). Hence ||(E) = 0
for all E E with E B and it follows from (ii) that ||(B) = 0. Hence
it follows from (5.37) that ||(A) = c. This shows that (ii) implies (iii).
That (iii) implies (ii) is obvious and this proves Lemma 5.24.
176 THEOREM
Definition 5.25. The signed measure is called inner regular with re-
spect to if it satisfies the equivalent conditions of Lemma 5.24.
Theorem 5.26 (Generalized RadonNikod ym Theorem).
Let (X, A, ) be a measure space and let : A R be a signed measure.
Then the following are equivalent.
(i) is truly continuous with respect to .
(ii) is absolutely continuous and inner regular with respect to .
(iii) There exists a function f L1 () such that (5.28) holds.
If these equivalent conditions are satisfied then the function f in (iii) is
uniquely determined by (5.28) up to equality -almost everywhere.
First proof of Theorem 5.26. This proof is due to Konig [8]. It has the advan-
tage that it reduces the proof of the generalized RadonNikod ym Theorem
to the standard RadonNikod ym Theorem 5.18 for -finite measure spaces.
We prove that (i) implies (ii). To see that is absolutely continuous
with respect to , fix a measurable set A A such that (A) = 0 and fix a
constant > 0. Choose > 0 and E A such that (E) < and (5.31)
holds. Then (A E) (A) = 0 < and hence |(A)| < by (5.31). Thus
|(A)| < for all > 0 and hence (A) = 0. This shows that .
We prove that is inner regular with respect to by verifying that
satisfies condition (i) in Lemma 5.24. Fix a set A A such that
E A, (E) < = (A E) = 0.
We must prove that (A) = 0. Let > 0 and choose > 0 and E A
such that (E) < and (5.31) holds. Then ((A \ E) E) = 0 < , hence
|(A \ E)| < by (5.31), and hence
|(A)| = |(A \ E) + (A E)| = |(A \ E)| < .
This shows that |(A)| < for all > 0 and so (A) = 0 as claimed. Thus
we have proved that (i) implies (ii).
We prove that (ii) implies (iii). Since is inner regular with respect to
there exists a sequence of measurable sets Ei A such that Ei Ei+1 and
(Ei ) < for all i N and ||(X) = limi ||(Ei ). Define

[
X0 := Ei , A0 := {A A | A X0 } , 0 := |A0 , 0 := |A0 .
i=1
GENERALIZED
Then (X0 , A0 , 0 ) is a -finite measure space and 0 : A0 R is a signed

measure that is absolutely continuous with respect to 0 . Hence the Radon
Nikodym Theorem 5.18 for -finite measure spaces asserts that there exists
a function f0 L1 (0 ) such that
Z
0 (A) = f0 d0 for all A A0 .
A
Define f : X R by f |X0 := f0 and f |X\X0 := 0. Then f L1 (). Choose

a measurable set A A. Then it follows from part (v) of Theorem 1.28 that
|(A \ X0 )| ||(A \ X0 ) ||(X \ X0 ) = lim ||(X \ Ei ) = 0.

i
Hence Z Z
(A) = 0 (A X0 ) = f0 d0 = f d
AX0 A
for all A A. This shows that (i) implies (iii). The uniqueness of f up to
equality -almost everywhere follows immediately from Lemma 1.49.
We prove that (iii) implies (i). Choose f L1 () such that (5.28) holds.
Define Z
c := ||(X) = |f | d
X
and
En := x X | 2n |f (x)| 2n ,

[
E := {x X | f (x) 6= 0} = En .
nN
Then 2n (En ) ||(En ) c and hence (En ) 2n c < for all n N.

Moreover, c = ||(X) = ||(E ) = limn ||(En ). Now fix a constant > 0.
Choose n N such that ||(En ) > c /2 and define := 2n1 . If A A
such that (A En ) < then
||(A) = ||(A \ En ) + ||(A En )

||(X \ En ) + 2n (A En )

< + 2n = .
2
This shows that is truly continuous with respect to . This completes the
first proof of Theorem 5.26.
178 THEOREM
Second proof of Theorem 5.26. This proof is due to Fremlin [4, Chapter 23].
It shows directly that (i) implies (iii) and has the advantage that it only uses
the Hahn Decomposition Theorem 5.19. It thus also provides an alternative
proof of Theorem 5.18 (assuming the Hahn Decomposition Theorem) which
is of interest on its own. By Lemma 5.23 it suffices to consider the case where
: A [0, ) is a finite measure that is truly continuous with respect to .
Consider the set

f is measurable and
F := f : X [0, ) R .
A
f d (A) for all A A
This set is nonempty because 0 F . Moreover, X f d (X) < for

R
all f F and
f, g F = max{f, g} F . (5.38)
(Let f, g F and A A and define the sets Af := {x A | f (x) > g(x)} and
Ag := R{x A | g(x) f (x)};
R then AfR, Ag A, Af Ag = , and Af Ag = A;
hence A max{f, g} d = Af f d + Ag g d (Af ) + (Ag ) = (A).)
Now define Z
c := sup f d (X)
f F X
and choose a sequence gi F such that limi X giRd = c. Then

R
R it follows
from (5.38) that fi := max{g1 , g2 , . . . , gi } F and R X gi d X fi d c
for all i N. Hence fi fi+1 for all i and limi X fi d = c. Define the
function f : X [0, ] by f (x) := limi fi (x) for x X. Then it follows
from the Lebesgue Monotone Convergence Theorem 1.37 that
Z Z Z Z
f d = lim fi d = c, f d = lim fi d (A) for all A A
X i X A i A
Hence f < -almost everywhere by Lemma 1.47 and we may assume

without loss of generality
R that f (x) < for all x X. Thus f F .
We prove that A f d = (A) for R all A A. Suppose otherwise that
there exists a set A0 A such that A0 f d < (A0 ). Then the formula
Z
0
(A) := (A) f d for A A (5.39)
A
defines a finite measure by Theorem 1.40.

GENERALIZED
We prove that there is a measurable function h : X [0, ) such that

Z Z
h d > 0, h d 0 (A) for all A A. (5.40)
X A
Define
0 (A0 )
:= > 0. (5.41)
3
Since is truly continuous with respect to so is 0 . Hence there is a > 0
and a set E A such that (E) < and
A A, (A E) < = 0 (A) < . (5.42)
0
Take A := X \ E to obtain (X \ E) < and hence
0 (E) 0 (A0 E) = 0 (A0 ) 0 (A0 \ E) = 3 0 (A0 \ E) > 2.
Then take A := A0 . Since 0 (A0 ) = 3 by (5.41) it follows from (5.42)
that (E) (A0 E) > 0. Define the signed measure 00 : A R by
(A E)
00 (A) := 0 (A) (5.43)
(E)
for A A. Then 00 (E) = 0 (E) . By the Hahn Decomposition
Theorem 5.19 there exists a measurable set P A such that
00 (A P ) 0, 00 (A \ P ) 0 for all A A.
Since 00 (E \ P ) 0 it follows that 00 (E) 00 (E P ) 0 (E P ).
Hence (E P ) by (5.42). Now define

h := EP . (5.44)
(E)
Then X h d > 0. Moreover, if A A then 00 (A P ) 0 and so, by (5.43),
R
(A P E)
Z
0
(A P ) = h d.
(E) A
Thus A h d 0 (A) for all A A and so h satisfies (5.40) as claimed.

R
It follows from (5.40) that

Z Z
(f + h) d f d + 0 (A) = (A)
A A
for all A A and hence f + h F . Since X (f + h) d = cR+ X h d > c,

R R
this contradicts the definition of c. Thus we have proved that A f d = (A)

for all A A and hence f satisfies (5.28). This completes the second proof
of Theorem 5.26.
180 THEOREM
5.5 Exercises
Exercise 5.27. Let (X, A, ) be a measure space such that (X) < .
Define
(A, B) := (A \ B) + (B \ A) for A, B A. (5.45)
Define an equivalence relation on A by A B iff (A, B) = 0. Prove that
descends to a function : A/ A/ [0, ) (denoted by the same
letter) and that the pair (A/,
R ) is a complete metric space. Prove that the
function A R : A 7 A f d descends to a continuous function on A/
for every f L1 ().
Exercise 5.28 (Rudin [17, page 133]). Let (X, A, ) be a measure space.
A subset F L1 () is called uniformly integrable if, for every > 0,
there is a constant > 0 such that, for all E A and all f F ,
Z

(E) < = f d < .

E

(i) Every finite subset of L1 () is uniformly integrable. Hint: Lemma 5.21.
(ii) Vitalis Theorem. Assume (X) < , let f : X R be measurable,
and let fn L1 () be a uniformly integrable sequence
R that converges almost
1
everywhere to f . Then f L () and limn X |f fn | d = 0.
Hint: Use Egoroffs Theorem in Exercise 4.54.
(iii) The hypothesis (X) < cannot be omitted in Vitalis Theorem.
Hint: Consider the Lebesgue measure on R. Find a uniformly integrable
sequence fn L1 (R) that converges pointwise to the constant function f 1.
(iv) Vitalis Theorem implies the Lebesgue Dominated Convergence Theo-
rem 1.45 under the assumption (X) < .
(v) Find an example where Vitalis Theorem applies although the hypotheses
of the Lebesgue Dominated Convergence Theorem are not satisfied.
(vi) Find an example of a measure space (X, A, ) with (X) < and a
sequence fn L1 () that is notR uniformly integrable, converges pointwise
to zero, and satisfies limn X fn d = 0. Hint: Consider the Lebesgue
measure on X = [0, 1].
(vii) Converse of Vitalis Theorem. Assume R (X) < and let fn be a
sequence in L1 () such that the limit limn A fn d exists for all A A.
Then the sequence fn is uniformly integrable.
5.5. EXERCISES 181
Hint: Let > 0. Prove that there is a constant > 0, an integer n0 N,

and a measurable set E0 E such that, for all E A and all n N,
Z

(E, E0 ) < , n n0 =
(f n f n0 ) d < .
(5.46)
E
(Here (E, E0 ) is defined by (5.45) as in Exercise 5.27.) If A A satisfies

(A) < then the sets E := E0 \A and E := E0 A both satisfy (E, E0 ) < .
Deduce that, for all A A and all n N,
Z

(A) < , n n0 =
(f n f n0 ) d < 2.
(5.47)
A
Now use part (i) to find a constant 0 > 0 such that, for all A A,
Z
0

(A) < = sup fn d < 3.
(5.48)
nN A
Exercise 5.29 (Rudin [17, page 134]). Let (X, A, ) be a measure space
such that (X) < and fix a real number p > 1. Let f : X R be
a measurable function and let fn L1 () be a sequence that converges
pointwise to f and satisfies
Z
sup |fn |p d < .
nN X
Prove that Z
1
f L (), lim |f fn | d = 0.
n X
Hint: Use Vitalis Theorem in Exercise 5.28.
Exercise 5.30. Let X := R, denote by B 2X the Borel -algebra, and
let : B [0, ] be the restriction of the Lebesgue measure to B. Let
: B [0, ] be a measure. Prove the following.
(i) If B B and 0 < c < (B) then there exists a Borel set A B such that
(A) = c. Hint: Show that the function f (t) := (B [t, t]) is continuous.
(ii) If there exists a constant 0 < c < such that
(B) = c = (B) = c.
for all B B, then .

182 THEOREM
Exercise 5.31. Let X := R, denote by B 2X the Borel -algebra, let

: B [0, ] be the restriction of the Lebesgue measure to B, and let
: B [0, ] be the counting measure. Prove the following.
(i)
(ii) is not inner regular with respect to .
R does not exist any measurable function f : X [0, ] such that
(iii) There
(B) = B f d for all B B.
Exercise 5.32. Let X := [1, ), denote by B 2X the Borel -algebra,
and let : B [0, ] be the restriction of the Lebesgue measure to B. Let
: B [0, ] be a Borel measure such that
(B) = (B) for all 1 and all B B. (5.49)
Prove that there exists a real number c 0 such that

Z
(B) := f d for all B B, (5.50)
B
where f : [1, ) [0, ) is the function given by

c
f (x) := for x 1. (5.51)
x2
Hint: Show that ([1, )) < and then that .
Exercise 5.33. Let X := [0, ) denote by B 2X the Borel -algebra, and
let : B [0, ] be the restriction of the Lebesgue measure to B. Define
the measures 1 , 2 : B [0, ] by
Z Z
X 1 1
1 (B) := 3
x dx, 2 (B) := dx
n=1
n B[n,n+1] B[1,) x2
R R
for B B. (Here we denote by B f (x) dx := B f d the Lebesgue integral
of a Borel measurable function f : [0, ) [0, ) over a Borel set B B.)
Prove that 1 and 2 are finite measures that satisfy
1 , 2 , 1 2 , 2 1 ,
and
6 1 , 6 2 .
5.5. EXERCISES 183
Exercise 5.34. Let (X, A, ) be a measure space. Show that the signed
measures : A R form a Banach space M = M(X, A) with norm
kk := ||(X).
Show that the map
L1 () M : [f ] 7 f
defined by (5.53) is an isometric linear embedding and hence L1 () is a closed
subspace of M.
Exercise 5.35. Let (X, U) be a compact Hausdorff space such that every
open subset of X is -compact and denote by B 2X its Borel -algebra.
Denote by C(X) := Cc (X) the space of continuous real valued functions
on X. This is a Banach space equipped with the supremum norm
kf k := sup|f (x)|.
xX
Let M(X) denote the space of signed Borel measures as in Exercise 5.34.
For M(X) define the linear functional : C(X) R by
Z
(f ) := f d.
X
(i) k k = kk. Hint: Use the Hahn Decomposition Theorem 5.19 and the
fact that every Borel measure on X is regular by Theorem 3.18.
(ii) Every bounded linear functional on C(X) is the difference of two positive
linear functionals. Hint: For f C(X) with f 0 prove that

+ (f ) := sup (hf ) h C(X), 0 h 1

(5.52)
= sup (g) g C(X), 0 g f .
Here the second supremum is obviously greater than or equal to the first. To
prove the converse inequality show that, for all g C(X) with 0 g f
and all > 0 there is an h C(X) such that 0 h 1 and |(g hf )| < .
Namely, find C(X) such that 0 1, (x) = 0 when f (x) /2 kk
and (x) = 1 when f (x) / kk; then define h := g/f . Once (5.52) is
established show that + extends to a positive linear functional on C(X).
(iii) The map M(X) C(X) : 7 is bijective. Hint: Use the Riesz
Representation Theorem 3.15.
(iv) The hypothesis that every open subset of X is -compact cannot be
removed in part (i). Hint: Consider Example 3.6.
184 THEOREM
Exercise 5.36. Let (X, A, ) be a measure space and let f : X [0, ) be

a measurable function. Define the measure f : A [0, ] by
Z
f (A) := f d for A A. (5.53)
A
(See Theorem 1.40.) Prove the following.

(i) If is -finite so is f .
(ii) If is semi-finite so is f .
(iii) If is localizable so is f .
Note: See Theorem 5.4 for (i) and [4, Proposition 234N] for (ii) and (iii).
It is essential that f does not take on the value . Find an example of
a measure space (X, A, ) and a measurable function f : X [0, ] that
violates the assertions (i), (ii), (iii).
Hint 1: To prove (ii), fix a set A A, define Af := {x A | f (x) > 0},
and choose a measurable set E A such that E Af and 0 < (E) < .
Consider the sets En := {x E | f (x) n}.
Hint 2: To prove (iii), let E A be any collection of measurable sets and
choose a measurable -envelope H A of E. Prove that the set

Hf := x H f (x) > 0
is a measurable f -envelope of E. In particular, if N A is a measurable

set such that f (E N ) = 0 for all E E, define Nf := {x N | f (x) > 0},
show that (H Nf ) = 0, and deduce that f (Hf N ) = f (H Nf ) = 0.
Chapter 6
Differentiation
This chapter returns to the Lebesgue measure on Euclidean space Rn intro-

duced in Chapter 2. The main result is the Lebesgue Differentiation Theorem
(Section 6.3). It implies that if f : Rn R is a Lebesgue integrable func-
tion then, for almost every element x Rn , the mean value of f over a ball
centered at x converges to f (x) as the radius tends to zero. Essential ingre-
dients in the proof are the Vitali Covering Lemma and the HardyLittlewood
Maximal Inequality (Section 6.2). One of the consequences of the Lebesgue
Differentiation Theorem is the Fundamental Theorem of Calculus for abso-
lutely continuous functions of one real variable (Section 6.4). The Lebesgue
Differentiation Theorem also plays a central role in the proof of the Calderon
Zygmund inequality (Section 7.7). The chapter begins with a discussion of
weakly integrable functions on general measure spaces.
6.1 Weakly Integrable Functions

Assume throughout that (X, A, ) is a measure space. Let f : X R be a
measurable function. Define the function f : [0, ) [0, ] by

f (t) := (t, f ) := (A(t, f )), A(t, f ) := x X |f (x)| > t , (6.1)
for t 0. The function f is nonincreasing and hence Borel measurable.
Define the function f : [0, ) [0, ] by
f () := inf {t 0 | (t, f ) } for 0 < . (6.2)
Thus f (0) = kf k and f is nonincreasing and hence Borel measurable.
By definition, the infimum of the empty set is infinity. Thus f () = if
185
186 CHAPTER 6. DIFFERENTIATION
and only if (A(t, f )) > for all t > 0. When f () < it is the smallest
number t such that the domain A(t, f ) (on which |f | > t) has measure at
most . This is spelled out in the next lemma.
Lemma 6.1. Let 0 < and 0 t < . Then the following holds.
(i) f () = if and only if f (s) > for all s 0.
(ii) f () = t if and only if f (t) and f (s) > for 0 s < t.
(iii) f () t if and only if f (t) .
Proof. It follows directly from the definition of f in (6.2) that f () =
if and only if (s, f ) > for all s [0, ) and this proves (i).
To prove (ii), fix a constant 0 t < . Assume first that (t, f )
and (s, f ) > for 0 s < t. Since f is nonincreasing this implies
(s, f ) (t, f ) for all s t and hence f () = t by definition.
Conversely, suppose that f () = t. Then it follows from the definition of
f that (s, f ) for s > t and (s, t) > for 0 s < t. We must prove
that (t, f ) . To see this observe that

[
A(t, f ) = A(t + 1/n, f ).
n=1
Hence it follows from part (iv) of Theorem 1.28 that
f (t) = (A(t, f )) = lim (A(t + 1/n, f )) = lim (t + 1/n, f ) .

n n
This proves (ii). If f () t then f (t) f (f ()) by (ii). If f (t)

then f () t by definition of f . This proves (iii) and Lemma 6.1.
Lemma 6.2. Let f, g : X R be measurable functions and let c R. Then
kf k1, := sup f () = sup tf (t) kf k1 , (6.3)

>0 t>0
kcf k1, = |c| kf k1, , (6.4)

kf k1, kgk1,
kf + gk1, + for 0 < < 1, (6.5)
1
q q q
kf + gk1, kf k1, + kgk1, . (6.6)
Moreover kf k1, = 0 if and only if f vanishes almost everywhere. The
inequality (6.6) is called the weak triangle inequality.
6.1. WEAKLY INTEGRABLE FUNCTIONS 187
Proof. For 0 < t, c < it follows from part (iii) of Lemma 6.1 that
t(t, f ) c (t, f ) ct1 f (ct1 ) t ct1 f (ct1 ) c.
This shows that supt>0 t(t, f ) = sup>0 f (). Moreover,
Z Z
t(t, f ) = t(A(t, f )) |f | d |f | d
A(t,f ) X
for all t > 0. This proves (6.3).

For c > 0 equation (6.4) follows from the fact that A(t, cf ) = A(t/c, f )
and hence (t, cf ) = (t/c, f ) for all t > 0. Since kf k1, = kf k1, by
definition, this proves (6.4).
To prove (6.5), observe that A(t, f + g) A(t, f ) A((1 )t, g), hence
(t, f + g) (t, f ) + ((1 )t, g), (6.7)
and hence
kf k1, kgk1,
t(t, f + g) t(t, f ) + t((1 )t, g) +
1
for all t > 0. Take the supremum over all t > 0 to obtain (6.5).
The inequality (6.6) follows from (6.5) and the identity

r
a b
inf + = a+ b for a, b 0. (6.8)
0<<1 1
This is obvious when a = 0 or b = 0. Hence assume a and b are positive and
define the function h : (0, 1) (0, ) by h() := a + 1
b
. It satisfies
b a
h0 () = 2
2
(1 )
and hence has a unique critical point at

a
0 := .
a+ b

Since h(0 ) = ( a+ b)2 , this proves (6.8). The inequality (6.6) then follows
by taking a := kf k1, and b := kgk1, .
The last assertion follows from the fact that kf k1, = 0 if and only if
f (0) = 0 if and only if the set A(0, f ) = {x X | f (x) 6= 0} has measure
zero. This proves Lemma 6.2.
Example 6.3. This example shows that the weak triangle inequality (6.6) is
sharp. Let (R, A, m) be the Lebesgue measure space and define f, g : R R
by
1 1
f (x) := , g(x) := for 0 < x < 1
x 1x
and f (x) := g(x) := 0 for x 0 and for x 1. Then
kf k1, = kgk1, = 1, kf + gk1, = 4.
Definition 6.4. Let (X, A, ) be a measure space. A measurable function

f : X R is called weakly integrable if kf k1, < . The space of weakly
integrable functions will be denoted by
n o
L1, () := f : X R f is measurable and kf k1, < .
The quotient space

L1, () := L1, ()/

under the equivalence relation f g iff f = g -almost everywhere is called
the weak L1 space. It is not a normed vector space because the function
L1, () [0, ) : [f ] 7 kf k1, does not satisfy the triangle inequality, in
general, and hence is not a norm. However, it is a topological vector space
and the topology is determined by the metric
q
d1, ([f ] , [g] ) := kf gk1, for f, g L1, (). (6.9)
For the Lebesgue measure space (Rn , A, m) we write L1, (Rn ) := L1, (m)
and L1, (Rn ) := L1, (m).
A subset of L1, () is open in the topology determined by the metric (6.9)

if and only if it is a union of sets of the form {[g] L1, () | kf gk1, < r}
with f L1, () and r > 0. A sequence [fi ] L1, () converges to [f ] in
this topology if and only if limi kfi f k1, = 0. The inequality (6.3) in
Lemma 6.2 shows that
L1 () L1, ()
for every measure space (X, A, ). In general, L1, () is not equal to L1 ().
For example the function f : R R defined by f (x) := 1/x for x > 0 and
f (x) := 0 for x 0 is weakly integrable but is not integrable.
6.1. WEAKLY INTEGRABLE FUNCTIONS 189
Theorem 6.5. The metric space (L1, (), d1, ) is complete.

Proof. Choose a sequence of weakly integrable functions fi : X R whose
equivalence classes [fi ] form a Cauchy sequence in L1, () with respect to
the
metric (6.9).
Then there is a subsequence i1 < i2 < i3 < such that
fi fi
k k+1 1,
< 22k for all k N. For k, ` N define

[
\
k
Ak := A(2 , fik fik+1 ), E` := Ak , E := E` .
k=` `=1

Then 2k (Ak ) fik fik+1 1, < 22k for all k N, hence

X
X
(E` ) (Ak ) 2k = 21`
k=` k=`
for all ` N, and hence (E) = 0. If x X \ E then there exists an ` N

such that x / Ak for all k ` and so |fik (x) fik+1 (x)| 2k for all k `.
This shows that the limit f (x) := limk fik (x) exists for all x X \ E.
Extend f to a measurable function on X by setting f (x) := 0 for x E.
We prove that limi kfi f k1, = 0 and hence also f L1, (). To
see this, fix a constant > 0 and choose an integer i0 N such that
i, j N, i, j i0 = 4 kfi fj k1, < .
Now fix a constant t > 0 and choose ` N such that
i` i0 , 22` t , 22` t.
If x
/ E` then x / Ak for allPk `, hence |fik (x) fiP k+1
(x)| 2k for k `,
and hence |fi` (x) f (x)| k=` |fik (x) fik+1 (x)|

k=` 2
k
= 21` t/2.
This shows that A(t/2, fi` f ) E` and hence
tfi` f (t/2) = t(A(t/2, fi` f )) t(E` ) t21` /2.
With this understood, it follows from (6.7) with = 1/2 that
tfi f (t) tfi fi` (t/2) + tfi` f (t/2) 2 kfi fi` k1, + /2 <
for all i N with i i0 . Hence
kfi f k1, = sup tfi f (t)
t>0
for every integer i i0 and this proves Theorem 6.5.

6.2 Maximal Functions

Let (R, A, m) be the Lebesgue measure space on R. In particular, the length
of an interval I R is m(I). As a warmup we characterize the differentia-
bility of a function that is obtained by integrating a signed measure.
Theorem 6.6. Let : A R be a signed measure and define f : R R by
f (x) := ((, x)) for x R. (6.10)
Fix two real numbers x, A R. Then the following are equivalent.

(i) f is differentiable at x and f 0 (x) = A.
(ii) For every > 0 there is a > 0 such that, for every open interval U R,

(U )
x U, m(U ) < = m(U ) A .
(6.11)
Proof. We prove that (i) implies (ii). Fix a constant > 0. Since f is
differentiable at x and f 0 (x) = A, there exists a constant > 0 such that,
for all y R,

f (x) f (y)
0 < |x y| < =
xy A . (6.12)
Let a, b R such that a < x < b and b a < . Then, by (6.12),

f (x) f (a) f (b) f (x)

xa A
,
bx A ,

or, equivalently,
(x a) f (x) f (a) A(x a) (x a),

(b x) f (b) f (x) A(b x) (b x).
Add these inequalities to obtain
(b a) f (b) f (a) A(b a) (b a).
Since ([a, b)) = f (b) f (a) and m([a, b)) = b a it follows that

([a, b))

m([a, b)) A .

6.2. MAXIMAL FUNCTIONS 191
Replace a by a + 2k and take the limit k to obtain

((a, b))
m((a, b)) A .

Thus we have proved that (i) implies (ii).

Conversely, assume (ii) and fix a constant > 0. Choose > 0 such
that (6.11) holds for every open interval U R. Choose y R such that
x < y < x+. Choose k N such that yx+2k < . Then Uk := (x2k , y)
is an open interval of length m(Uk ) < containing x and hence

(Uk )
m(Uk ) A

by (6.11). Take the limit k to obtain

f (y) f (x) ([x, y)) (Uk )

yx A=
m([x, y)) A = lim
k m(Uk ) A .

Thus (6.12) holds for x < y < x + and an analogous argument proves
the inequality for x < y < x. Thus (ii) implies (i) and this proves
Theorem 6.6.
The main theorem of this chapter will imply that, when is absolutely
continuous with respect to m, the derivative of the function f in (6.10) exists
almost everywhere, defines a Lebesgue integrable function f 0 : R R, and
that Z
(A) = f 0 dm
A
for all Lebesgue measurable sets A A. It will then follow that an absolutely
continuous function on R can be written as the integral of its derivative. This
is the fundamental theorem of calculus in measure theory (Theorem 6.19).
The starting point for this program is the assertion of Theorem 6.6. It
suggests the definition of the derivative of a signed measure
:AR
at a point x R as the limit of the quotients (U )/m(U ) over all open

intervals U containing x as m(U ) tends to zero, provided that the limit
exists. This idea carries over to all dimensions and leads to the concept of a
maximal function which we explain next.
Notation. Fix a natural number n N. Let (Rn , A, m) denote the

Lebesgue measure space and let
n
B 2R
denote the Borel -algebra of Rn with the standard topology. Thus L1 (Rn )
denotes the space of Lebesgue integrable functions f : Rn R. An element of
L1 (Rn ) need not be Borel measurable but differs from a Borel measurable func-
tion on a Lebesgue null set by Theorem 2.14 and part (v) of Theorem 1.55.
For x Rn and r > 0 denote the open ball of radius r, centered at x, by

Br (x) := y Rn |x y| < r .

Here q
|| := 12 + + n2
denotes the Euclidean norm of = (1 , . . . , n ) Rn .
Definition 6.7 (HardyLittlewood Maximal Function).
Let : B [0, ) be a finite Borel measure. The maximal function of
is the function
M : Rn R
defined by
(Br (x))
(M )(x) := sup . (6.13)
r>0 m(Br (x))
The maximal function of a signed measure : B R is defined as the
maximal function
M := M || : Rn R
of its total variation || : B [0, ).
Theorem 6.8 (HardyLittlewood Maximal Inequality).
Let : B R be a signed Borel measure. Then the function M : Rn R
in Definition 6.7 is lower semi-continuous, i.e. the pre-image of the open in-
terval (t, ) under M is open for all t R. Hence M is Borel measurable.
Moreover,
kM k1, 3n ||(Rn ) (6.14)
and so M L1, (Rn ).
The proof of Theorem 6.8 relies on the following two lemmas.
Lemma 6.9. Let : B [0, ) be a finite Borel measure. Then the

maximal function M : Rn R is lower semi-continuous and hence is Borel
measurable.
Proof. Let t > 0 and define
Ut := A(t, M ) = {x Rn | (M )(x) > t} . (6.15)
We prove that Ut is open. Fix an element x Ut . Since (M )(x) > t there

exists a number r > 0 such that
(Br (x))
t< .
m(Br (x))
Choose > 0 such that

(r + )n (Br (x))
t n
< .
r m(Br (x))
Choose y Rn such that |y x| < . Then Br (x) Br+ (y) and hence
(Br+ (y)) (Br (x))

(r + )n
> t m(Br (x))
rn
(r + )n
= t m(Br (y))
rn
= t m(Br+ (y)).
This implies
(Br+ (y))
(M )(y) >t
m(Br+ (y))
S y Ut . This shows thatn Ut is open for all t > 0. It follows that

and hence
U0 = t>0 Ut is open and Ut = R is open for t < 0. Thus M is lower
semi-continuous as claimed. This proves Lemma 6.9.
The HardyLittlewood estimate on the maximal function M is equiva-

lent to an upper bound for the Lebesgue measure of the set Ut in (6.15). The
proof relies on the next lemma about coverings by open balls.
Lemma 6.10 (Vitalis Covering Lemma). Let ` N and, for i = 1, . . . , `,

let xi Rn and ri > 0. Define
`
[
W := Bri (xi ).
i=1
Then there exists a set

S {1, . . . , `}
such that
Bri (xi ) Brj (xj ) = for all i, j S with i 6= j (6.16)
and [
W B3ri (xi ). (6.17)
iS
Proof. Abbreviate Bi := Bri (xi ) and choose the ordering such that
r1 r2 r` .
Choose i1 := 1 and let i2 > 1 be the smallest index such that Bi2 Bi1 = .
Continue by induction to obtain a sequence
1 = i1 < i2 < < ik `
such that
Bij Bij0 = for j 6= j 0
and
Bi (Bi1 Bij ) 6= for ij ik when j = k). Then
Bi B3ri1 (xi1 ) B3rij (xij ) for ij 0. Then the set Ut := A(t, M ) is
open by Lemma 6.9. Choose a compact set K Ut . If x K Ut then
(M )(x) > t and so there exists a number r(x) > 0 such that
||(Br(x) (x))
> t. (6.18)
m(Br(x) (x))
Since K isScompact there exist finitely many points x1 , . . . , x` K such

that K ì=1 Bri (xi ), where ri := r(xi ). By Lemma 6.10 there is a subset
S {1,
S . . . , `} such that the balls Brin(xi ) for i S are pairwise disjoint and
K iS B3ri (xi ). Since m(B3r ) = 3 m(Br ) by Theorem 2.17, this gives
X 3n X 3n
m(K) 3n m(Bri (xi )) < ||(Bri (xi )) ||(Rn ).
iS
t iS t
Here the second step follows from (6.18) with ri = r(xi ) and the last step
follows from the fact that the balls Bri (xi ) for i S are pairwise disjoint.
Take the supremum over all compact sets K Ut to obtain
3n
m(A(t, M )) = m(Ut ) ||(Rn ). (6.19)
t
(See Theorem 2.13.) Multiply the inequality (6.19) by t and take the supre-
mum over all real numbers t > 0 to obtain kM k1, 3n ||(Rn ). This
proves Theorem 6.8.
Definition 6.11. Let f L1 (Rn ). The maximal function of f is the

function M f : Rn [0, ) defined by
Z
1
(M f )(x) := sup |f | dm for x Rn . (6.20)
r>0 m(Br (x)) Br (x)
Corollary 6.12. Let

R f L1 (Rn ) and define the signed Borel measure f
on Rn by f (B) := B f dm for every Borel set B Rn . Then
M f = M f L1, (Rn ), kM f k1, 3n kf k1 .

R
Proof. The formula |f |(B) = B |f | dm for B B shows that M f = M f .
Hence the assertion follows from Theorem 6.8.
Corollary 6.12 shows that the map f 7 M f descends to an operator

(denoted by the same letter) from the Banach space L1 (Rn ) to the topological
vector space L1, (Rn ). Corollary 6.12 also shows that the resulting operator
M : L1 (Rn ) L1, (Rn )
is continuous (because |M f M g| M (f g)). Note that it is not linear. By

Theorem 6.8 it extends naturally to an operator 7 M from the Banach
space of signed Borel measures on Rn to L1, (Rn ). (See Exercise 5.34.)
6.3 Lebesgue Points

Definition 6.13. Let f L1 (Rn ). An element x Rn is a called a
Lebesgue point of f if
Z
1
lim |f f (x)| dm = 0. (6.21)
r0 m(Br (x)) B (x)
r
In particular, x is a Lebesgue point of f whenever f is continuous at x.
The next theorem is the main result of this chapter.
Theorem 6.14 (Lebesgue Differentiation Theorem).

Let f L1 (Rn ). Then there exists a Borel set E Rn such that m(E) = 0
and every element of Rn \ E is a Lebesgue point of f .
Proof. For f L1 (Rn ) and r > 0 define the function Tr f : Rn [0, ) by

Z
1
(Tr f )(x) := |f f (x)| dm for x Rn . (6.22)
m(Br (x)) Br (x)
One can prove via an approximation argument that Tr f is Lebesgue measur-

able for every r > 0 and every f L1 (Rn ). However, we shall not use this
fact in the proof. For f L1 (Rn ) define the function T f : Rn [0, ] by
(T f )(x) := lim sup(Tr f )(x) for x Rn . (6.23)

r0
We must prove that T f = 0 almost everywhere for every f L1 (Rn ).

6.3. LEBESGUE POINTS 197
To see this, fix a function f L1 (Rn ) and assume without loss of gener-
ality that f is Borel measurable. (See Theorem 2.14 and part (v) of Theo-
rem 1.55.) By Theorem 4.15 there exists a sequence of continuous functions
gi : Rn R with compact support such that
1
kf gi k1 < for all i N.
2i
Since gi is continuous we have T gi = 0. Moreover, the function
hi := f gi
is Borel measurable and satisfies

Z
1
(Tr hi )(x) = |hi hi (x)| dm
m(Br (x)) Br (x)
Z
1
|hi | dm + |hi (x)|
m(Br (x)) Br (x)
(M hi )(x) + |hi (x)|
for all x Rn . Thus

Tr hi M hi + |hi |
for all i and all r > 0. Take the limit superior as r tends to zero to obtain
T hi M hi + |hi |
for all i. Moreover, it follows from the definition of Tr that
Tr f = Tr (gi + hi ) Tr gi + Tr hi
for all i and all r > 0. Take the limit superior as r tends to zero to obtain
T f T gi + T hi = T hi M hi + |hi |
for all i. This implies
A(, T f ) A(/2, M hi ) A(/2, hi ). (6.24)
for all i and all > 0. (See equation (6.1) for the notation A(, T f ) etc.)
Since hi and M hi are Borel measurable (see Theorem 6.8) the set
Ei () := A(/2, M hi ) A(/2, hi ) (6.25)

is a Borel set. Since khi k1 < 2i we have
2 2 1
m(A(/2, hi )) khi k1, khi k1 i1
2
and, by Theorem 6.8,
2 2 3n 3n
m(A(/2, M hi )) kM hi k1, khi k1 i1 .
2
Thus
3n + 1
m(Ei ()) .
2i1
Since this holds for all i N it follows that the Borel set

\
E() := Ei ()
i=1
has Lebesgue measure zero for all > 0. Hence the Borel set

[
E := E(1/k)
k=1
has Lebesgue measure zero. By (6.24) and (6.25), we have
A(1/k, T f ) E(1/k)
for all k N and hence

[
n

x Rn (T f )(x) > 1/k

x R (T f )(x) 6= 0 =

k=1
[
= A(1/k, T f )
k=1
[
E(1/k)
k=1
= E.
This shows that (T f )(x) = 0 for all x Rn \ E and hence every element of
Rn \ E is a Lebesgue point of f . This proves Theorem 6.14.
6.3. LEBESGUE POINTS 199
The first consequence of Theorem 6.14 discussed here concerns a signed

Borel measure on Rn that is absolutely continuous with respect to the
Lebesgue measure. The following theorem provides a formula for the function
f in Theorem 5.18 (also called the RadonNikod ym derivative of ).
Theorem 6.15. Let : B R be a signed Borel measure on Rn that is
absolutely continuous with respect to the Lebesgue measure.
R Choose a Borel
measurable function f L1 (Rn ) such that (B) = B f dm for all B B.
Then there exists a Borel set E Rn such that m(E) = 0 and
(Br (x))
f (x) = lim for all x Rn \ E. (6.26)
r0 m(Br (x))
Proof. By Theorem 6.14 there exists a Borel set E Rn of Lebesgue measure

zero such that every element of X \ E is a Lebesgue point. Since
Z
(Br (x)) 1

m(Br (x)) f (x) = f f (x) dm
m(Br (x)) Br (x)
Z
1
|f f (x)| dm
m(Br (x)) Br (x)
for all r > 0 and all x Rn , it follows that (6.26) holds for all x Rn \ E.
Theorem 6.16. Let f L1 (Rn ) and let x Rn be a Lebesgue point of f .
Fix two constants 0 < < 1 and > 0. Then there exists a > 0 such that,
for every Borel set E B and every r > 0,
r < ,
1
Z

E Br (x), =
m(E) f dm f (x) < . (6.27)

m(E) > m(Br (x)) E
Proof. Since x is Ra Lebesgue point of f , there exists a constant > 0 such

that m(Br (x))1 Br (x) |f f (x)| dm < for 0 < r < . Assume 0 < r <
and let E Br (x) be a Borel set such that m(E) m(Br (x)). Then
Z Z
1 1

m(E) f dm f (x)
|f f (x)| dm
E m(E) E
Z
1
|f f (x)| dm
m(Br (x)) Br (x)
< .
The Lebesgue Differentiation Theorem 6.14 can be viewed as a theorem

about signed Borel measures that are absolutely continuous with respect to
the Lebesgue measure. The next theorem is an analogous result for signed
Borel measures that are singular with respect to the Lebesgue measure.
Theorem 6.17 (Singular Lebesgue Differentiation).
Let : B R be a signed Borel measure on Rn such that m. Then
there exists a Borel set E Rn such that m(E) = 0 = ||(Rn \ E) and
||(Br (x))
lim =0 for all x Rn \ E. (6.28)
r0 m(Br (x))
Proof. The proof follows an argument in Heil [7, Section 3.4]. By assumption
and Lemma 5.16, there exists a Borel set A Rn such that
m(A) = 0, ||(Rn \ A) = 0.
For > 0, define the set

n
||(Br (x))
A := x R \ A lim sup
> .
r0 m(Br (x))
We prove that A is a Lebesgue null set for every > 0. To see this, fix
two constants > 0 and > 0. Since the Borel measure || is regular by
Theorem 3.18 there exists an open set U Rn such that
A U , ||(U ) < .
For x A choose a radius r = r(x) > 0 such that
||(Br(x) (x))
> , Br(x) (x) U ,
m(Br(x) (x))
S
and consider the open set W := xA Br(x) (x) U . Fix a compact subset
K W and cover K by finitely many of the balls Br(x) (x) with x A . By
Vitalis Covering Lemma 6.10 there are elements S x1 , . . . , xN A such that
the balls Br(xi ) (xi ) are pairwise disjoint and K N i=1 B3r(xi ) (xi ). Thus
N N N
X
n 3n X 3n X 3n
m(K) 3 m(Br(xi ) (xi )) < ||(Br(xi ) (xi )) ||(U ) .
i=1
i=1 i=1
This holds for every compact set K W and hence m(W ) 3n /. Since
> 0 was arbitrary, the set A is contained in an open set of arbitrarily small
Lebesgue measureSand so is a Lebesgue null set as claimed. This implies that
the set E := A n
k=1 A1/k is a Lebesgue null set. It satisfies ||(R \ E) = 0
and (6.28) by definition and this proves Theorem 6.17.
6.4. ABSOLUTELY CONTINUOUS FUNCTIONS 201
6.4 Absolutely Continuous Functions

Definition 6.18. Let I R be an interval. A function f : I R is called
absolutely continuous if for every > 0 there exists a > 0 such that,
for every finite sequence s1 t1 s2 t2 s` t` in I,
`
X `
X
|si ti | < = |f (si ) f (ti )| < . (6.29)
i=1 i=1
Every absolutely continuous function is continuous.

The equivalence of (i) and (iii) in the following result is the Fundamental
Theorem of Calculus for Lebesgue integrable functions. The equivalence
of (i) and (ii) is known as the BanachZarecki Theorem. For functions
of bounded variation see Exercise 6.20 below.
Theorem 6.19 (Fundamental Theorem of Calculus).
Let I = [a, b] R be a compact interval, let B 2I be the Borel algebra,
and let m : B [0, ] be the restriction of the Lebesgue measure to B. Let
f : I R be a function. Then the following are equivalent.
(i) f is absolutely continuous.
(ii) f is continuous, it has bounded variation, and if E I is a Lebesgue
null set then so is f (E).
R
(iii) There is a Borel measurable function g : I R such that I |g| dm <
and, for all x, y I with x < y,
Z y
f (y) f (x) = g(t) dt. (6.30)
x
The right hand side denotes the Lebesgue integral of g over the interval [x, y].
If (iii) holds then there exists a Borel set E I such that m(E) = 0 and,
for all x I \ E, f is differentiable at x and f 0 (x) = g(x).
Proof. We prove that (iii) implies the last assertion of the theorem. Thus
assume that there exists a function g L1 (I) that satisfies (6.30) for all
x, y I with x < y. Then Theorem 6.14 asserts that there exists a Borel
set E I of Lebesgue measure zero such that every element of I \ E is a
Lebesgue point of g. By Theorem 6.16 with = 1/2, every element x I \ E
satisfies condition (ii) in Theorem 6.6 with A := g(x). Hence Theorem 6.6
asserts that the function f is differentiable at every point x I \ E and
satisfies f 0 (x) = g(x) for x I \ E.
We prove that (iii) implies (i). Thus assume f satisfies (iii) and define
the signed measure : B R by
Z
(B) := g dm (6.31)
B
for every Borel set B I. Then is absolutely continuous with respect to

the Lebesgue measure and
Z
||(B) = |g| dm
B
for ever Borel set B I. Now let > 0. Since || m it follows from
Lemma 5.21 that there exists a constant > 0 such that ||(B) < for every
Borel set B PI with m(B) < . Choose a sequence s1 t1 s` t`
in I such that ì=1 |ti si | < and define Ui := (si , ti ) for i = 1, . . . , `. Then
the Borel set B := ì=1 Ui has Lebesgue measure m(B) = ì=1 |ti si | < .
S P
Hence ||(B) < . Since
Z Z

|f (ti ) f (si )| =
g dm |g| dm = ||(Ui )
Ui Ui
for all i it follows that

`
X `
X
|f (ti ) f (si )| ||(Ui ) = ||(B) < .
i=1 i=1
Hence f is absolutely continuous.

We prove that (i) implies (ii). Assume f is absolutely continuous. It
follows directly from the definition that f is continuous, and that it has
bounded variation is part (v) of Exercise 6.20. Now suppose that E I
is a Lebesgue null set and assume without loss of generality that a, b / E.
Fix any constant > 0 and choose > 0 such that (6.29) holds. Since the
Lebesgue measure is outer regular by Theorem 2.13, there exists an open
set U int(I) such that E U and m(U ) < . Choose a (possibly finite) S
sequence of pairwise disjoint open intervals Ui I such that U = i Ui .
Choose si , ti Ui such P
that f (si ) = inf
PUi f and f (ti ) = supUi f . Then it
follows from (6.29) that i m(f (Ui )) = P i (f (ti ) f (si )) < for every finite
sum. Take the limit to obtain m(f (U )) i m(f (Ui )) . Since > 0 was
chosen arbitrary and the Lebesgue measure is complete, it follows that f (E)
is a Lebesgue measurable set and m(f (E)) = 0.
6.4. ABSOLUTELY CONTINUOUS FUNCTIONS 203
We prove that (ii) implies (iii). Assume f satisfies (ii). Then f has
bounded variation and under this assumption Exercise 6.20 below outlines
a proof that there exists a signed Borel measure : B R such that
f (y) f (x) = ((x, y]) for x, y I with x < y. Since + and are regular
by Theorem 3.18 and f is continuous, we have
f (y) f (x) = ([x, y]) = ((x, y]) = ([x, y)) = ((x, y)) (6.32)
for all x, y I with x < y. By the Lebesgue Decomposition Theorem 5.17
there exist two signed Borel measures a , s : B R such that
= a + s , a m, s m. (6.33)
Since a m it follows from Theorem 5.18 that there is an integrable
function g L1 (I) such that
Z
a (B) = g dm (6.34)
B
for every Borel set B I. Define the functions fa , fs : I R by

Z x
fa (x) := f (a) + a ([a, x]) = f (a) + g(t) dt, fs (x) := s ([a, x]).
a
Then f = fa +fs by (6.32) and (6.33). Since (iii) implies (i) and (i) implies (ii)
(already proved) both functions f and fa satisfy (ii) and fa is absolutely
continuous. Moreover fs = f fa is continuous.
It remains to prove that fs 0. The proof given below follows an ar-
gument in Heil [7, Section 3.5.4]. By Theorem 6.17, there exists a Lebesgue
null set Es I such that a, b Es (without loss of generality) and
|s ((x r, x + r))|
|s |(I \ Es ) = 0, lim = 0 for all x I \ Es .
r0 r
This implies that every element x I \ Es satisfies condition (ii) in Theo-
rem 6.6 with A = 0 and f replaced by fs . Hence fs is differentiable at every
point x I \ Es and fs0 (x) = 0 for x I \ Es .
Since Es is a Lebesgue null set so are the sets f (Es ) and fa (Es ), because
the functions f and fa satisfy (ii). Since the difference of two Lebesgue null
sets in R is a Lebesgue null set, by Exercise 2.27, it follows that
fs (Es ) = {f (x) fa (x) | x Es } {f (x) fa (y) | x, y Es }
is a Lebesgue null set.
We claim that fs (I \ Es ) is also a Lebesgue null set. To see this, fix a

constant > 0. For n N define the set

1
An := x I \ Es y I, |x y| < = |fs (x) fs (y)| |x y| .
n
Then [
An An+1 for all n N and I \ Es = An .
nN
Here the last assertion follows from the fact that fs is differentiable on I \ Es
with derivative zero. We prove that the Lebesgue outer measure of the set
fs (I \ Es ) satisfies the estimate
(fs (I \ Es )) (b a + ). (6.35)
To see this, cover the set An by at most countably many open intervals Ui ,
each of length less than 1/n, such that
X
m(Ui ) (An ) +
i
and each interval Ui contains an element of An . Then f (Ui ) is contained in

an interval of length at most m(Ui ), by definition of An . Hence
X
(f (An )) m(Ui ) ((An ) + ) (b a + ).
i
Since the Lebesgue outer measure is continuous from below by part (iii) of
Theorem 2.13 and [
fs (I \ Es ) = fs (An ),
nN
it follows that
(fs (I \ Es )) = lim (f (An )) (b a + ).

n
This proves (6.35). Since > 0 was chosen arbitrary, this implies that
fs (I \ Es ) is a Lebesgue null set as claimed. Since fs (Es ) is also a Lebesgue
null set, as noted above, it follows that fs (I) is a Lebesgue null set. Since fs
is continuous and fs (0) = 0 by definition, this implies fs 0. Hence f = fa
is absolutely continuous and this proves Theorem 6.19.
6.5. EXERCISES 205
6.5 Exercises
Exercise 6.20. Let I = [a, b] R be a compact interval and let B 2I
be the Borel -algebra. A function f : I R is said to be of bounded
variation if
`
X
V (f ) := sup |f (ti ) f (ti1 )| < . (6.36)
a=t0 <t1 <<t` =b i=1
Denote by BV(I) the set of all functions f : I R of bounded variation.

This is a real vector space. Functions of bounded variation have at most
countably many discontinuities and the left and right limits exist everywhere.
(i) Every monotone function f : I R has bounded variation.
(ii) Let f BV(I) be right continuous. Then there exist right continuous
monotone functions f : I R such that f = f + f . Hint: Define
`
X
F (x) := V (f |[a,x] ) = sup |f (ti ) f (ti1 )| (6.37)
a=t0 <t1 <<t` =x
i=1
for a x b. Prove that F is right continuous and F f are monotone.

(iii) If f is continuous then the function F in (6.37) is continuous.
(iv) Let f : I R be right continuous. Then f BV(I) if and only if there
exists a signed Borel measure = f : B R such that ({a}) = 0 and
f (x) f (a) = ([a, x]) for a x b. (6.38)
Hint: Assume f is monotone. For h C(I) define
Z b X`

f (h) := h df := sup inf h f (ti ) f (ti1 ) . (6.39)
a a=t0 <t1 <<t` =b i=1 [ti1 ,ti ]
(This is the RiemannStieltjes integral. See Korner [9] and compare it

with the Riemann integral [9, 18, 21].) Prove that f : C(I) R is a positive
linear functional. Use the Riesz RepresentationR Theorem 3.15 to find a Borel
measure f : B [0, ) such that f (h) = I h df for all h C(I). Use
the fact that f is right continuous to prove that f satisfies (6.38).
(v) If f BV(I) is right continuous and f is as in (iii) then V (f ) = |f |(I).
(vi) Every absolutely continuous function f : I R has bounded variation.
Exercise 6.21. Let (R, A, m) be the Lebesgue measure space and fix a con-
stant 0 < < 1/2. Prove that there does not exist a Lebesgue measurable
set E R such that
m(E I)
< <1
m(I)
for every nonempty bounded open interval I R. Hint: Consider the
function f := E[1,1] and define the measure f : B R by
Z
f (B) := f dm = m(B E [1, 1]).
B
Examine the Lebesgue points of f . (Compare this with Exercise 2.32.)
Exercise 6.22. Prove the Theorem of VitaliCarath eodory:
Let (X, U) be a locally compact Hausdorff space and let B 2X be its Borel
-algebra. Let : B [0, ] be an outer regular Borel measure that is
inner regular on open sets. Let f L1 () and let > 0. Then there exists
an upper semi-continuous function u : X R that is bounded above and a
lower semi-continuous function v : X R that is bounded below such that
Z
u f v, (v u) d < . (6.40)
X
Hint: Assume first that f 0. Use Theorem 1.26 to find a sequence of
measurable sets Ei A, not necessarily disjoint, and a sequence of real
numbers ci > 0 such that (Ei ) < for all i and
X
f= ci Ei .
i=1
Thus Z
X
ci (Ei ) = f d < .
i=1 X
Choose a sequence of compact sets Ki X and a sequence of open sets

Ui X such thatPKi Ei Ui and ci (Ui \ Ki ) < 2i1 for all i. Choose
n N such that i=n+1 ci (Ei ) < /2 and define
n
X
X
u := ci Ki , v := ci Ui .
i=1 i=1
Show that X (v u) d < , v is lower semi-continuous (i.e. v 1 ((t, ))

R
is open for all t R), and u is upper semi-continuous (i.e. u1 ((, t))
is open for all t R).
6.5. EXERCISES 207
Exercise 6.23. Fix two real numbers a < b and prove the following.
(i) If f : [a, b] R is everywhere differentiable then f 0 : [a, b] R is Borel
measurable.
Rb
(ii) If f : [a, b] R is everywhere differentiable and a |f 0 (t)| dt < then f
is absolutely continuous.
Hint: Fix a constant > 0. By the VitaliCaratheodory Theorem in Exer-
cise 6.22 there is a lower semi-continuous function g : [a, b] R such that
Z b Z b
0
g>f, g(t) dt < f 0 (t) dt + .
a a
For > 0 define the function F : [a, b] R by

Z x
F (x) := g(t) dt f (x) + f (a) + (x a)
a
for a x b. Consider a point a x < b. Since g(x) > f 0 (x) and g is lower
semi-continuous, find a number x > 0 such that
f (t) f (x)
g(t) > f 0 (x), < f 0 (x) + for x < t < x + x .
tx
Deduce that
F (t) > F (x) for x < t < x + x .
Since F (a) = 0 there exists a maximal element x [a, b] such that F (x) = 0.
If x 0 for x < t b.
In either case F (b) 0 and hence
Z b Z b
f (b) f (a) g(t) dt + (b a) < f 0 (t) dt + + (b a).
a a
Since this holds for all > 0 and all > 0 it follows that
Z b
f (b) f (a) f 0 (t) dt.
a
Rb
Replace f by f to obtain the equation f (b) f (a) = a
f 0 (t) dt. Now
deduce that Z x
f (x) f (a) = f 0 (t) dt
a
for all x [a, b].
Example 6.24. (i) The Cantor function is the unique monotone function
f : [0, 1] [0, 1] that satisfies
X X
ai ai
f 2 i
=
i=1
3 i=1
2i
for all sequences ai {0, 1}. It is continuous and nonconstant and its deriva-
tive exists and vanishes on the complement of the standard Cantor set
[ X n n
\ ai X ai 1
C := 2 i
,2 i
+ n .
n=1 i=1
3 i=1
3 3
ai {0,1}
This Cantor set has Lebesgue measure zero. Hence f is almost everywhere
differentiable and its derivative is integrable. However, f is not equal to the
integral of its derivative and therefore is not absolutely continuous.
(ii) The following construction was explained to me by Theo Buehler. Define
the homeomorphisms g : [0, 1] [0, 2] and h : [0, 2] [0, 1] by
g(x) := f (x) + x, h := g 1 .
The image g([0, 1] \ C) is a countable union of disjoint open intervals of
total length one and hence has Lebesgue measure one. Thus its complement
K := g(C) [0, 2] is a modified Cantor set of Lebesgue measure one. Hence,
by Theorem 6.19, g is not absolutely continuous. Moreover, by Lemma 2.15
there exists a set E K which is not Lebesgue measurable. However, its
image F := h(E) [0, 1] under h is a subset of the Lebesgue null set C
and hence is a Lebesgue null set. Thus F is a Lebesgue measurable set
and E = h1 (F ) is not Lebesgue measurable. This shows that the function
h : [0, 2] [0, 1] is not measurable with respect to the Lebesgue -algebras on
both domain and target (i.e. it is not Lebesgue-Lebesgue measurable).
(iii) Let I, J R be intervals. Then it follows from Lemma 2.15 and
Theorem 6.19 that every Lebesgue-Lebesgue measurable homeomorphism
h : I J has an absolutely continuous inverse.
(iv) Let h : [0, 2] [0, 1] and F C [0, 1] be as in part (ii). Then the
characteristic function F : R R is Lebesgue measurable and h : [0, 2] R
is continuous. However, the composition F h : [0, 2] R is not Lebesgue
measurable because the set (F h)1 (1) = E is not Lebesgue measurable.
(v) By contrast, if I, J R are intervals, f : J R is Lebesgue measurable,
and h : I J is a C 1 diffeomorphism, then f h : I R is again Lebesgue
measurable by Theorem 2.17.
Chapter 7
Product Measures
The purpose of this chapter is to study products of two measurable spaces

(Section 7.1), introduce the product measure (Section 7.2), and prove Fubinis
Theorem (Section 7.3). The archetypal example is the Lebesgue measure on
Rk+` = Rk R` ; it is the completion of the product measure associated to
the Lebesgue measures on Rk and R` (Section 7.4). Applications include the
convolution (Section 7.5), Marcinkiewicz interpolation (Section 7.6), and the
CalderonZygmund inequality (Section 7.7).
7.1 The Product -Algebra

Assume throughout that (X, A) and (Y, B) are measurable spaces.
Definition 7.1. The product -algebra of A and B is defined as the small-
est -algebra on the product space X Y := {(x, y) | x X, y Y } that con-
tains all subsets of the form A B, where A A and B B. It will be
denoted by A B 2XY .
Lemma 7.2. Let E AB and let f : X Y R be an (AB)-measurable
function. Then the following holds.
(i) For every x X the function fx : Y R, defined by fx (y) := f (x, y) for
y Y , is B-measurable and

Ex := y Y (x, y) E B. (7.1)
(ii) For every y Y the function f y : X R, defined by f y (x) := f (x, y)
for x X, is A-measurable and

E y := x X (x, y) E A.

(7.2)
209
210 CHAPTER 7. PRODUCT MEASURES
Proof. Define 2XY by

:= E X Y Ex B for all x X .
We prove that is a -algebra. To see this, note first that X Y .

Second, if E then Ex B for all x X, hence
(E c )x = {y Y | (x, y)
/ E} = (Ex )c B
for allSx X, and henceSE c . Third, if Ei is a sequence and

E := i=1 Ei , then Ex =

i=1 (Ei )x B for all x X, and hence E .
This shows that is a -algebra. Since A B for all A A and all
B B it follows that A B . This proves (7.1) for all x X.
Now fix an element x X. If V R is open then E := f 1 (V ) A B
and hence (fx )1 (V ) = Ex B by (7.1). Thus fx is B-measurable. This
proves (i). The proof of (ii) is analogous and this proves Lemma 7.2.
Definition 7.3. Let Z be a set. A collection of subsets M 2Z is called a
monotone class if it satisfies the following two axioms
(a) If Ai M for i N such that Ai Ai+1 for all i then
S
Ai M.
Ti=1

(b) If Bi M for i N such that Bi Bi+1 for all i then i=1 Bi M.
Definition 7.4. A subset Q X Y is called elementary if it is the union
of finitely many pairwise disjoint subsets of the form A B with A A and
B B.
The next lemma is a useful characterization of the product -algebra.
Lemma 7.5. The product -algebra A B is the smallest monotone class
in X Y that contains all elementary subsets.
Proof. Let E 2XY denote the collection of all elementary subsets and
define M 2XY as the smallest monotone class that contains E. This is
well defined because the intersection of any collection of monotone classes is
again a monotone class. Since every -algebra is a monotone class and every
elementary set is an element of A B it follows that
M A B.
Since E M by definition, the converse inclusion follows once we know

that M is a -algebra. We prove this in seven steps.
7.1. THE PRODUCT -ALGEBRA 211
Step 1. For every set P X Y the collection

(P ) := Q X Y P \ Q, Q \ P, P Q M
is a monotone class.
This follows immediately from the definition of monotone class.
Step 2. Let P, Q X Y . Then Q (P ) if and only if P (Q).
This follows immediately from the definition of (P ) in Step 1.
Step 3. If P, Q E then P Q, P \ Q, P Q E.
For the intersection this follows from the fact that
(A1 B1 ) (A2 B2 ) = (A1 A2 ) (B1 B2 ) .
For the complement it follows from the fact that

(A1 B1 ) \ (A2 B2 ) = (A1 \ A2 ) B1 (A1 A2 ) (B1 \ B2 ) .
For the union this follows from the fact that P Q = (P \ Q) Q.

Step 4. If P E then M (P ).
Let P E. Then P \ Q, Q \ P, P Q E M for all Q E by Step 3.
Hence Q (P ) for all Q E by definition of (P ) in Step 1. Thus we have
proved that E (P ). Since (P ) is a monotone class by Step 1 it follows
that M (P ). This proves Step 4.
Step 5. If P M then M (P ).
Fix a set P M. Then P (Q) for all Q E by Step 4. Hence Q (P )
for all Q E by Step 2. Thus E (P ) and hence it follows from Step 1
that M (P ). This proves Step 5.
Step 6. If P, Q M then P \ Q, P Q M.
If P, Q M then Q M (P ) by Step 5 and hence P \ Q, P Q M
by the definition of (P ) in Step 1.
Step 7. M is a -algebra.
c
By definition X Y E M. If P M Sn then P = (X Y ) \ P M by
i M for S
Step 6. If PS i N then Qn := i=1 Pi M for all n N by Step 6
and hence P
i=1 i =
n=1 Qn M because M is a monotone class. This
proves Step 7 and Lemma 7.5.
Lemma 7.6. Let (X, UX ) and (Y, UY ) be topological spaces, let UXY be
the product topology on X Y (see Appendix B), and let BX , BY , BXY be
the associated Borel -algebras. Then
BX BY BXY . (7.3)
If (X, UX ) is a second countable locally compact Hausdorff space then
BX BY = BXY . (7.4)
Proof. The projections X : X Y X and Y : X Y Y are continuous

1
and hence Borel measurable by Theorem 1.20. Thus X (A) = AY BXY
for all A BX and Y1 (B) = X B BXY for all B BY . Hence
A B BXY for all A BX and all B BY , and this implies (7.3).
Now assume (X, UX ) is a second countable locally compact Hausdorff
space and choose a countable basis {Ui | i N} of UX such that U i is compact
for all i N. Fix an open set W UXY and, for i N, define

Vi := y Y | (x, y) W for all x U i .
We prove that Vi is open. Let y0 Vi . Then (x, y0 ) W for all x U i .

Hence, for every x U i , there exist open sets U (x) UX and V (x) UY
such that (x, y0 ) U (x) V (x) W . Since U i is compact there are finitely
many elements x1 , . . . , x` U i such that U i U (x1 ) U (x` ). Define
V := V (x1 ) V (x` ). Then V is open and U i V W , so y0 V Vi .
This shows that Vi is open for all i N. Next we prove that

[
W = (Ui Vi ). (7.5)
i=1
Let (x0 , y0 ) W . Then there exist open sets U UX and V UY such

that (x0 , y0 ) U V W . Since (X, UX ) is a locally compact Hausdorff
space, Lemma A.3 asserts that there exists an open set U 0 X such that
x0 U 0 U 0 U . Since the sets Ui form a basis of the topology, there
exists an integer i N such that x0 Ui U 0 and hence x0 U i U 0 U .
Thus U i {y0 } U V W , hence y0 Vi , and so (x0 , y0 ) Ui Vi W .
Since the element (x0 , y0 ) W was chosen arbitrarily, this proves (7.5). Thus
we have proved that UXY BX BY and this implies BXY BX BY .
Hence (7.4) follows from (7.3). This proves Lemma 7.6.
7.1. THE PRODUCT -ALGEBRA 213
Lemma 7.7. Let (X, A) be a measurable space such that the cardinality of
X is greater than that of 2N . Then the diagonal := {(x, x) | x X} is not
an element of A A.
Proof. The proof has three steps.
Step 1. Let Y be a set. For E 2Y denote by (E) 2Y the smallest
-algebra containing E. If D (E) then there exists a sequence Ei E for
i N such that D ({Ei | i N}).
The union of the sets (E 0 ) over all countable subsets E 0 E is a -algebra
that contains E and is contained in (E). Hence it is equal to (E).
Step 2. Let Y be a set, let E 2Y , and let DS (E). T ThenTthere is a
sequence Ei E and a set I 2N such that D = II iI Ei iN\I Eic .
By Step 1 there exists T Ei Ec such that D ({Ei | i N}). For
T a sequence
I N define EI := iI Ei iN\I Ei . These sets form a partition of Y .
S
Hence the collection F := II E I | I 2 N
is a -algebra on Y . Since
Ei F for each i N it follows that D F. This proves Step 2.
Step 3. / A A.
Let E 2XX be the collection of all sets of the form A B with A, B A.
Let D A A.SBy Step 2 there are sequences Ai , Bi ATand a set I 2N
(A Bi )c .
T
such that D = II EI , where EI := iI (ATi Bi )
S TiN\I i
Thus EI = JN\I AIJ BIJ , where AIJ := iI Ai jJ (X \ Aj ) and
T T
BIJ := iI Bi jN\(IJ) (X \ Bj ). If D then, for all I and J, we
have AIJ BIJ and so AIJ BIJ is either empty or a singleton. Thus
the cardinality of D is at most the cardinality of the set of pairs of disjoint
subsets of N, which is equal to the cardinality of 2N . Since the cardinality of
the diagonal is bigger than that of 2N it follows that / A A as claimed.
Example 7.8. Let X be an uncountable set, of cardinality greater than
that of 2N , and equipped with the discrete topology so that BX = UX = 2X .
Then is an open subset of X X with respect to the product topology
(which is also discrete because points are open). Hence BXX = 2XX .
However, / BX BX by Lemma 7.7. Thus the product BX BX of the
Borel -algebras is not the Borel -algebra of the product. In other words,
the inclusion (7.3) in Lemma 7.6 is strict in this example. Note also that
the distance function d : X X R defined by d(x, y) := 1 for x 6= y and
d(x, x) := 0 is continuous with respect to the product topology but is not
measurable with respect to the product of the Borel -algebras.
7.2 The Product Measure

The definition of the product measure on the product -algebra is based
R (X, A, ) and Ra measurable
on the following theorem. For a measure space
function : X [0, ] we use the notation X (x) d(x) := X d.
Theorem 7.9. Let (X, A, ) and (Y, B, ) be -finite measure spaces and let
Q A B. Then the functions
X [0, ] : x 7 (Qx ), Y [0, ] : y 7 (Qy ) (7.6)
are measurable and
Z Z
(Qx ) d(x) = (Qy ) d(y). (7.7)
X Y
Definition 7.10. Let (X, A, ) and (Y, B, ) be -finite measure spaces. The
product measure of and is the map : A B [0, ] defined by
Z Z
( )(Q) := (Qx ) d(x) = (Qy ) d(y) (7.8)
X Y
for Q AB. That is -additive, andPhence is a measure, follows from

Theorem 1.38 and the fact that (Qx ) = i=1 ((Qi )x ) for every sequence of
pairwise disjoint sets Qi A B. The product measure satisfies
( )(A B) = (A) (B) (7.9)
for A A and B B and hence is -finite.

the functions (7.6) are measurable
:= Q A B .
and satisfy equation (7.7)
We prove in five steps that = A B.
Step 1. If A A and B B then Q := A B .
By assumption

B, if x A, y A, if y B,
Qx = Q =
, if x
/ A, , if y
/ B.
Define the function : X [0, ] by (x) := (Qx ) = (B)A (x) for x X
y
and the function : Y [0, ]R by (y) := (Q ) = (A)
R B (y) for y Y .
Then , are measurable and X d = (A)(B) = Y d. Thus Q .
7.2. THE PRODUCT MEASURE 215
Step 2. If Q1 , Q1 and Q1 Q2 = then Q := Q1 Q2 .

Define
i (x) := ((Qi )x ), (x) := (Qx ),
(7.10)
i (y) := ((Qi )y ), (y) := (Qy )
for x X, y Y and i = 1, 2. Then = 1 + 2 and = 1 + 2 . Moreover,
Z Z
i d = i d
X Y
R R
for i = 1, 2 because Qi . Hence X
d = Y
d and so Q .
for all i then Q :=
S
Step 3. If Qi for i N and Qi Qi+1 i=1 Qi .
Define i , : X [0, ] and i , : Y [0, ] by (7.10) for i N. Since

[
[
Qx = (Qi )x , Qy = (Qi )y
i=1 i=1
y
and (Qi )x B and (Qi ) A for all i it follows from Theorem 1.28 (iv) that
(x) = (Qx ) = lim ((Qi )x ) = lim i (x) for all x X,
i i
(y) = (Qy ) = lim ((Qi )y ) = lim i (y) for all y Y.
i i
By the Lebesgue Monotone Convergence Theorem 1.37 this implies

Z Z Z Z
d = lim i d = lim i d = d.
X i X i Y Y
Thus Q and this proves Step 3.

Step 4. Let A A and B B such that (A) < and (B) T < . If
Qi for i N such that A B Q1 Q2 then Q :=
i=1 Qi .
Let i , , i , be as in the proof of Step 3. Since (Qi )x B and (B) <
it follows from part (v) of Theorem 1.28 that i converges pointwise to .
Moreover, i (B)A for all i and the function (B)A : X [0, ) is
integrable because (A) < and (B) < . Hence it follows from the
Lebesgue Dominated Convergence Theorem 1.45 that
Z Z
d = lim i d.
X i X
R R
The same argument R shows that
R Y
d = limi Y
i d. Since Qi for
all i, this implies X d = Y d and hence Q . This proves Step 4.
Step 5. = A B.
Since (X, A, ) and (Y, B, ) are -finite, there exist sequences of measurable
sets Xn A and Yn B such that
Xn Xn+1 , Yn Yn+1 , (Xn ) < , (Yn ) <
for all n N and X =
S S
n=1 Xn and Y = n=1 Yn . Define

M := Q A B Q (Xn Yn ) for all n N .
Then M is a monotone class by Steps 3 an 4, E M by Steps 1 and 2, and
M AB by definition. Hence it follows from Lemma 7.5 that M = AB.
In other words Q (Xn Yn ) for all Q A B. By Step 3 this implies

[
Q= Q (Xn Yn ) for all Q A B.
n=1
Thus A B A B and so = A B as claimed. This proves Step 5

and Theorem 7.9.
Examples and exercises

Example 7.11. Let X = Y = [0, 1], let A 2X be the -algebra of Lebesgue
measurable sets, let B := 2Y , let : A [0, 1] be the Lebesgue measure,
and let : B [0, ] be the counting measure. Consider the diagonal
[ n 2
\ i 1 i
:= (x, x) 0 x 1 = , A B.
n=1 i=1
n n
Its characteristic function f := : X Y R is given by

1, if x = y,
f (x, y) :=
0, if x 6= y.
Hence
Z
y
( ) = f (x, y) d(x) = 0 for 0 y 1,
ZX
(x ) = f (x, y) d(y) = 1 for 0 x 1,
Y
R R
and so X (y ) d = 0 6= 1 = Y (x ) d(x). Thus the hypothesis that
(X, A, ) and (Y, B, ) are -finite cannot be removed in Theorem 7.9.
7.2. THE PRODUCT MEASURE 217
Example 7.12. Let X := Y := [0, 1], let A = B 2[0,1] be the -algebra of

Lebesgue measurable sets, and let = be the Lebesgue measure.
Claim 1. Assume the continuum hypothesis. Then there is a set Q [0, 1]2
such that [0, 1] \ Qx is countable for all x and Qy is countable for all y.
Let Q be as in Claim 1 and define f := Q : [0, 1]2 R. Then
Z
y
(Q ) = f (x, y) d(x) = 0 for 0 y 1,
ZX
(Qx ) = f (x, y) d(y) = 1 for 0 x 1,
Y
and hence Z Z
y
(Q ) d = 0 6= 1 = (Q) d(x).
X Y
y
The sets Qx and Q are all measurable and the integrals are finite, but the
set Q is not A B-measurable. This shows that the hypothesis Q A B
in Theorem 7.9 cannot be replaced by the weaker hypothesis that sets Qx
and Qy are all measurable, even when the integrals are finite. It also shows
that Lemma 7.2 does not have a converse. Namely, fx and f y are measurable
for all x and y, but f is not A B-measurable.
Claim 2. Assume the continuum hypothesis. Then there exists a bijection
j : [0, 1] W with values in a well ordered set (W, ) such that the set
{w W | w z} is countable for all z W .
Claim 2 implies Claim 1. Let j be as in Claim 2 and define
Q := (x, y) [0, 1]2 | j(x) j(y) .

Then the set Qy = {x [0, 1] | j(x) j(y)} is countable for all y [0, 1] and
the set [0, 1] \ Qx = {y [0, 1] | j(y) 4 j(x)} is countable for all x [0, 1].
Proof of Claim 2. By Zorns Lemma every set admits a well ordering.
Choose any well ordering on A := [0, 1] and define
B := {b A | the set {a A | a b} is uncountable} .
If B = choose W := A = [0, 1] and j = id. If B 6= then, by the well
ordering axiom, B contains a smallest element b0 . Since b0 B, the set
W := B \ A = {w A | w b0 } is uncountable. Since W B = the set
{w W | w z} is countable for all z W . Since W is an uncountable
subset of [0, 1], the continuum hypothesis asserts that there exists a bijection
j : [0, 1] W . This proves Claim 2.
Example 7.13. Let X and Y be countable sets, let A = 2X and B = 2Y ,

and let : 2X [0, ] and : 2Y [0, ] be the counting measures. Then
A B = 2XY and : 2XY [0, ] is the counting measure.
Example 7.14. Let (X, A, ) and (Y, B, ) be probability measure spaces

so that (X) = (Y ) = 1. Then : A B [0, 1] is also a probability
measure. A trivial example is A = {, X} and B = {, Y }. In this case the
product -algebra is A B = {, X Y } and the product measure is given
by ( )() = 0 and ( )(X Y ) = 1.
Exercise 7.15. Let (X, A, ), (Y, B, ) be -finite measure spaces and let
: X X, :Y Y
be bijections. Define the bijection : X Y X Y by
( )(x, y) := ((x), (y))
for x X and y Y . Prove that
( ) (A B) = A B, ( ) ( ) = .
Hint: Use Theorem 1.19 to show that A B ( ) (A B). See

also Exercise 2.34.
Exercise 7.16. For n N let Bn Rn be the Borel -algebra and let
n : Bn [0, ]
be the restriction of the Lebesgue measure to Bn . Let k, ` N and n := k +`.

Identify Rk R` with Rn in the obvious manner. Then
Bk B` = Bn
by Lemma 7.6. Prove that the product measure k ` is translation invari-

ant and satisfies (k ` )([0, 1]n ) = 1. Deduce that
k ` = n .
Hint: Use Exercise 7.15. We return to this example in Section 7.4.

7.3. FUBINIS THEOREM 219
7.3 Fubinis Theorem

There are three versions of Fubinis Theorem. The first concerns nonegative
functions that are measurable with respect to the product -algebra (Theo-
rem 7.17), the second concerns real valued functions that are integrable with
respect to the product measure (Theorem 7.20), and the third concerns real
valued functions that are integrable with respect to the completion of the
product measure (Theorem 7.23).
Theorem 7.17 (Fubini for Positive Functions). Let (X, A, ), (Y, B, )
be -finite measure spaces and let : A B [0, ] be the product mea-
sure in Definition 7.10. Let f : X Y [0,R] be an AB-measurable func-
tion. Then the function X [0, R ] : x 7 Y f (x, y) d(y) is A-measurable,
the function Y [0, ] : y 7 X f (x, y) d(x) is B-measurable, and
Z Z Z
f d( ) = f (x, y) d(y) d(x)
XY X Y
Z Z (7.11)
= f (x, y) d(x) d(y).
Y X
Example 7.18. Equation (1.20) is equivalent to equation (7.11) for the

counting measure on X = Y = N.
Proof of Theorem 7.17. Let fx (y) := f y (x) := f (x, y) for (x, y) X Y and
define the functions : X [0, ] and : Y [0, ] by
Z Z
(x) := fx d, (y) := f y d (7.12)
Y X
for x X and y Y . We prove in three steps that is A-measurable, is
B-measurable, and and satisfy equation (7.11).
Step 1. The assertion holds when f : X Y [0, ) is the characteristic
function of an A B-measurable set.
Let Q A B and f = Q . Then fx = Qx and f y = Qy , and so
(x) = (Qx ), (y) = (Qy )
for all x X and all y Y . Hence it follows from Theorem 7.9 that
Z Z Z
d = d = ( )(Q) = f d( ).
X Y X
Here the third equation follows from the definition of the measure .
This proves Step 1.
Step 2. The assertion holds when f : X Y [0, ) is an AB-measurable

step-function.
This follows immediately from Step 1 and the linearity of the integral.
Step 3. The assertion holds when f : X Y [0, ] is A B-measurable.
By Theorem 1.26 there exists a sequence of A B-measurable step-functions
sn : X Y [0, )
such that sn sn+1 for all n N and sn converges pointwise to f . Define
Z
n (x) := sn (x, y) d(y) for x X,
ZY
n (x) := sn (x, y) d(x) for y Y.

X
Then
n n+1 , n n+1 for all n N
by part (i) of Theorem 1.35. Moreover, it follows from the Lebesgue Mono-
tone Convergence Theorem 1.37 that
(x) = lim n (x), (y) = lim n (y)
n n
for all x X and all y Y . Use the Lebesgue Monotone Convergence

Theorem 1.37 again as well as Step 2 to obtain
Z Z
d = lim n d
X n X
Z Z
= lim sn d( ) = f d( )
n XY XY
Z
= lim n d
n Y
Z
= d.
Y

A first application of Fubinis Theorem 7.17 is Minkowskis inequality
for a measurable function on a product space that is p-integrable with respect
to one variable such that the resulting Lp norms define an integrable function
of the other variable.
Theorem 7.19 (Minkowski). Fix a constant 1 p < . Let (X, A, )

and (Y, B, ) be -finite measure spaces and let f : X Y [0, ] be A B-
measurable. Then
Z Z p 1/p Z Z 1/p
p
f (x, y) d(y) d(x) f (x, y) d(x) d(y).
X Y Y X
y
In the notation fx (y) := f (x) := f (x, y) Minkowskis inequality has the form
Z 1/p Z
p
kfx kL1 () d(x) kf y kLp () d(y). (7.13)
X Y
Proof. By Lemma 7.2 fx : Y [0, ] is B-measurable for all x X and

fy : X [0, ] is A-measurable for all y Y . Moreover, by Theorem 7.17,
the function X [0, ] : x 7 kfx kpL1 () is A-measurable and the function
Y [0, ] : y 7 kf y kLp () is B-measurable. Hence both sides of the
inequality (7.13) are well defined. Theorem 7.17 also shows that for p = 1
equality holds in (7.13). Hence assume 1 < p < and a choose 1 < q <
such that 1/p + 1/q = 1. It suffices to assume
Z
c := kf y kLp () d(y) < .
Y
Define : X [0, ] by
Z
(x) := fx d for x X
Y
and let g Lq (). Then the function X Y [0, ] : (x, y) 7 f (x, y)|g(x)|
is A B-measurable. Hence it follows from Theorem 7.17 that
Z Z Z
|g| d = f (x, y)|g(x)| d(y) d(x)
X X Y
Z Z
= f (x, y)|g(x)| d(x) d(y)
Y X
Z
kf y kLp () kgkLq () d(y)
Y
= c kgkLq () .
Here the third step follows from Holders inequality in Theorem 4.1. Since
(X, A, ) is semi-finite by part (ii) of Lemma 4.30, it follows from Lemma 4.34
that kkLp () c. This proves Theorem 7.19.
Theorem 7.20 (Fubini for Integrable Functions). Let (X, A, ) and

(Y, B, ) be -finite measure spaces, let : A B [0, ] be the product
measure, and let f L1 ( ). Define fx (y) := f y (x) := f (x, y) for x X
and y Y . Then the following holds.
(i) fx L1 () for -almost every x X and the map : X R defined by
R
f d, if fx L1 (),
Y x
(x) := (7.14)
0, / L1 (),
if fx
is -integrable.
(ii) f y L1 () for -almost every y Y and the map : Y R defined by
R y
X
f d, if f y L1 (),
(y) := (7.15)
0, if f y
/ L1 (),
is -integrable.
(iii) Let L1 () and L1 () be as in (i) and (ii). Then
Z Z Z
d = f d( ) = d. (7.16)
X XY Y
Proof. We prove part (i) and the first equation in (7.16). The functions
f := max{f, 0} : X Y [0, ) are AB-measurable by Theorem 1.24.
Hence the functions fx := max{fx , 0} = f (x, ) : Y [0, ) are B-
measurable by Lemma 7.2. Define : Y [0, ] by
Z

(x) := fx d for x X.
Y
By Theorem 7.17 the functions : X [0, ] are A-measurable and

Z Z Z

d = f d( ) |f | d( ) < . (7.17)
X XY XY
Now Lemma 1.47 asserts that the A-measurable set

Z

+
E := x X |fx | d = = x X (x) = or (x) =

Y
has measure (E) = 0. Moreover, for all x X,
xE / L1 ().
fx
Define : X [0, ) by

(x), if x
/ E,
(x) := for x X.
0, if x E,
Then it follows from (7.17) that L1 () and
Z Z

d = f d( ).
X XY
Hence = + L1 () and
Z Z Z
d = +
d d
X ZX X Z
= +
f d( ) f d( )
ZXY XY
= f d( ).
XY
This proves (i) and the first equation in (7.16). An analogous argument
proves (ii) and the second equation in (7.16). This proves Theorem 7.20.
Example 7.21. Let (X, A, ) = (Y, B, ) be the Lebesgue measure space
in the unit interval [0, 1] as in Example 7.12. Let gn : [0, 1] [0, ) be a
sequence of smooth functions such that
Z 1
gn (x) dx = 1, gn (x) = 0 for x [0, 1] \ [2n1 , 2n ]
0
for all n N. Define f : [0, 1]2 R by

X
f (x, y) := gn (x) gn+1 (x) gn (y).
n=1
The sum on the right is finite for every pair (x, y) [0, 1]2 . Then
Z Z
X
f (x, y) dx = 0, f (x, y) dy = gn (x) gn+1 (x) = g1 (x),
X Y n=1
and hence
Z 1 Z 1 Z 1 Z 1
f (x, y) dx dy = 0 6= 1 = f (x, y) dy dx.
0 0 0 0
Thus the hypothesis f L1 ( ) cannot be removed in Theorem 7.20.

Example 7.22. This example shows that the product measure is typically
not complete. Let (X, A, ) and (Y, B, ) be two complete -finite measure
spaces. Suppose (X, A, ) admits a nonempty null set A A and B 6= 2Y .
Choose B 2Y \ B. Then A B / A B. However, A B is contained in
the -null set A Y and so belongs to the completion (A B) .
In the first version of Fubinis Theorem integrability was not an issue. In
the second version integrability of fx was only guaranteed for almost all x.
In the third version the function fx may not even be measurable for all x.
Theorem 7.23 (Fubini for the Completion). Let (X, A, ) and (Y, B, )
be complete -finite measure spaces, let (X Y, (A B) , ( ) ) de-
note the completion of the product space, and let f L1 (( ) ). Define
fx (y) := f y (x) := f (x, y) for x X and y Y . Then the following holds.
(i) fx L1 () for -almost every x X and the map : X R defined by
R
f d, if fx L1 (),
Y x
(x) := (7.18)
0, / L1 (),
if fx
is -integrable.
(ii) f y L1 () for -almost every y Y and the map : Y R defined by
R y
X
f d, if f y L1 (),
(y) := (7.19)
0, if f y
/ L1 (),
is -integrable.
Z Z Z

d = f d( ) = d. (7.20)
X XY Y
Proof. By part (v) of Theorem 1.55 there exists a function g L1 ( )

such that the set N := {(x, y) X Y | f (x, y) 6= g(x, y)} (A B) has
measure zero, i.e. ( ) (N ) = 0. By definition of the completion there
exists a set Q A B such that N Q and ( )(Q) = 0. Thus
Z Z
(Qx ) d(x) = (Qy ) d(y) = 0.
X Y
Hence, by Lemma 1.49,

(E) = 0, E := x X (Qx ) 6= 0 ,

F := y Y (Qy ) 6= 0 .

(F ) = 0,
Since f = g on (X Y ) \ Q we have fx = gx on Y \ Qx for all x X and

f y = g y on X \ Qy for all y Y . By Theorem 7.20 for g L1 ( ) there
are measurable sets E 0 A and F 0 B such that (E 0 ) = (F 0 ) = 0 and
gx L1 () for all x X \ E 0 ,
g y L1 () for all y Y \ F 0 .
If x X \ (E E 0 ) then (Qx ) = 0 and fx = gx on Y \ Qx . Since (Y, B, )

is complete and gx L1 (), every function that differs from gx on a set of
measure zero is also B-measurable and -integrable. Hence fx L1 () for
all x X \ (E E 0 ). The same argument shows that f y L1 () for all
y Y \ (F F 0 ). Define the functions : X R and : Y R by
f d, for x X \ (E E 0 ),
R
(x) := Y x
0, for x E E 0 ,
f y d, for y Y \ (F F 0 ),
R
(y) := X
0, for y F F 0 .
Since (x) = Y gx d for all x X \ (E E 0 ) it follows from part (i) of The-

R
orem 7.20 for g that L1 (). The same argument, using part (ii) of Theo-
rem 7.20 for g, shows that L1 (). Moreover, the three integrals in (7.20)
for f agree with the corresponding integrals for g because
(E E 0 ) = (F F 0 ) = ( )(Q) = 0.
Hence equation (7.20) for f follows from part (iii) of Theorem 7.20 for g.
Example 7.24. Assume (X, A, ) is not complete. Then there exists a set
E 2X \ A and a set N A such that E N and (N ) = 0. In this case
the set E Y is a null set in the completion (X Y, (A B) , ( ) ).
Hence f := EY L1 (( ) ). However, the function f y = E is not
measurable for every y Y . This shows that the hypothesis that (X, A, )
and (Y, B, ) are complete cannot be removed in Theorem 7.23.
Exercise 7.25. Continue the notation of Theorem 7.23 and suppose that
f : X Y [0, ] is (A B) -measurable. Prove that fx is B-measurable
for -almost all x X, that f y is A-measurable for -almost all y Y , and
that equation (7.11) continues to hold. Hint: The proof of Theorem 7.23
carries over verbatim to nonnegative measurable functions.
We close this section with two remarks about the construction of product
measures in the non -finite case, where the story is considerably more subtle.
These remarks are not used elsewhere in this book and can safely be ignored.
Remark 7.26. Let (X, A, ) and (Y, B, ) be two arbitrary measure spaces.
In [4, Chapter 251] Fremlin defines the function : 2XY [0, ] by
( )
X An A, Bn B for n N
(W ) := inf (An ) (Bn ) (7.21)
and W
S
n=1 n=1 An Bn
for W X Y and proves that it is an outer measure. He shows that

the -algebra C 2XY of -measurable sets contains the product -algebra
A B and calls the measure
1 := |C : C [0, ]
the primitive product measure. By Caratheodorys Theorem 2.4 the
measure space (X Y, C, 1 ) is complete. By definition
1 (A B) = (A) (B)
for all A A and all B B. Fremlin then defines the complete locally
determined (CLD) product measure 0 : C [0, ] by

E A, F B,
0 (W ) := sup 1 W (E F ) . (7.22)
(E) < , (F ) <
He shows that (X Y, C, 0 ) is a complete measure space, that 0 1 , and
1 (W ) < = 0 (W ) = 1 (W )
for all W C. (See [4, Theorem 251I].) One can also prove that a measure
: C [0, ] satisfies (E F ) = (E) (F ) for all E A and F B
with (E) (F ) < if and only if 0 1 . With these definitions
Fubinis Theorem holds for 0 whenever the factor (Y, B, ) (over which the
integral is performed first) is -finite and the factor (X, A, ) (over which
the integral is performed
S second) is either strictly localizable (i.e. there
is a partition X = iI Xi into measurable sets with (Xi ) < such that
a set A X is A-measurable
P if and only if A Xi A for all i I
and, moreover, (A) = iI (A Xi ) for all A A) or is complete and
locally determined (i.e. it is semi-finite and a set A X is A-measurable
if and only if A E A for all E A with (E) < ). See Fremlin [4,
Theorem 252B] for details.
If the measure spaces (X, A, ) and (Y, B, ) are both -finite then the
measures 0 and 1 agree and are equal to the completion of the product
measure on A B (see [4, Proposition 251K]).
Remark 7.27. For topological spaces yet another approach to the product
measure is based on the Riesz Representation Theorem 3.15. Let (X, UX )
and (Y, UY ) be two locally compact Hausdorff spaces, denote by BX and BY
their Borel -algebras, and let X : BX [0, ] and Y : BY [0, ] be
Borel measures. Define : Cc (X Y ) R by
Z Z
(f ) := f (x, y) d(y) d(x)
X Y
Z Z (7.23)
= f (x, y) d(x) d(y)
Y X
for f Cc (X Y ). That the two integrals agree for every continuous function
with compact support follows from Fubinis Theorem 7.20 for finite measure
spaces. (To see this, observe that every compact set K X Y is contained
in the product of the compact sets KX := {x X | ({x} Y ) K 6= } and
KY := {y Y | (X {y}) K 6= }.) Since is a positive linear functional,
the Riesz Representation Theorem 3.15 asserts that there exists a unique
outer regular Borel measure 1 : BXY [0, ] that is inner regular on
open sets and a unique Radon measure 0 : BXY [0, ] such that
Z Z
(f ) = f d0 = f d1
XY XY
for all f Cc (X Y ). It turns out that in this situation the Borel -algebra
BXY is contained in the -algebra C 2XY of Remark 7.26 and
0 = 0 |BXY , 1 = 1 |BXY .
Recall from Lemma 7.6 that the product -algebra BX BY agrees with the
Borel -algebra BXY whenever one of the spaces X or Y is second countable.
If they are both second countable then so is the product space (X Y, UXY )
(Appendix B). In this case
0 = 1 = X Y
is the product measure of Definition 7.10 and 0 = 1 : C [0, ] is its

completion. (See Theorem 3.15 and Remark 7.26.)
7.4 Fubini and Lebesgue

For n N denote by (Rn , An , mn ) the Lebesgue measure space on Rn and by
Bn An the Borel -algebra on Rn with respect to the standard topology.
For k, ` N we identify Rk+` with Rk R` in the standard manner. Since Rn
is second countable for all n it follows from Lemma 7.6 and Theorem 2.1 that
Bk B` = Bk+` , (mk |Bk ) (m` |B` ) = mk+` |Bk+` . (7.24)
(See Exercise 7.16.) Thus Theorem 7.17 has the following consequence.
Theorem 7.28 (Fubini and Borel). Let k, ` N and n := k + `. Let
f : Rn [0, ] be Borel measurable. Then fx := f (x, ) : R` [0, ] and
f y := f (, y) : Rk [0, ] are Borel measurable for R all x Rk and all
y R` . Moreover,R the functions Rk [0, ] : x 7 R` f (x, y) dm` (y) and
R` [0, ] : y 7 Rk f (x, y) dmk (x) are Borel measurable and
Z Z Z
f dmn = f (x, y) dm` (y) dmk (x)
Rn Rk R`
Z Z (7.25)
= f (x, y) dmk (x) dm` (y).
R` Rk
Proof. The assertion follows directly from (7.24) and Theorem 7.17.
For Lebesgue measurable functions f : Rn [0, ] the analogous state-
ment is considerably more subtle. In that case the function fx , respec-
tively f y , need not be Lebesgue measurable for all x, respectively all y.
However, they are Lebesgue measurable for almost all x Rk , respectively
almost all y R` , and the three integrals in (7.25) can still be defined and
agree. The key result that one needs to prove this is that the Lebesgue
measure on Rn = Rk R` is the completion of the product of the Lebesgue
measures on Rk and R` . Then the assertion follows from Exercise 7.25.
Theorem 7.29. Let k, ` N, define n := k + `, and identify Rn with the
product space Rk R` in the canonical way. Denote the completion of the
product space (Rk R` , Ak A` , mk m` ) by (Rk R` , (Ak A` ) , (mk m` ) ).
Then An = (Ak A` ) and mn = (mk m` ) .
Proof. Define
n o
Cn := [a1 , b1 ) [an , bn ) ai , bi R and ai < bi for i = 1, . . . , n

n
so that Cn Bn An 2R for all n. We prove the assertion in three steps.
7.4. FUBINI AND LEBESGUE 229
Step 1. Bn Ak A` and mn (B) = (mk m` )(B) for all B Bn .

By Lemma 7.6 we have Bn = Bk B` Ak A` . It then follows from
the uniqueness of a normalized translation invariant Borel measure on Rn in
Theorem 2.1 that mn |Bn = (mk m` )|Bn . Here is a more direct proof.
First, assume B = E = [a1 , b1 ) [an , bn ) Cn . Define
E 0 := [a1 , b1 ) [ak , bk ), E 00 := [ak+1 , bk+1 ) [an , bn ).
Thus E 0 Ck Ak , E 00 C` A` , and so E = E 0 E 00 Ak A` . Moreover

n
Y
mn (E) = (bi ai ) = mk (E 0 ) m` (E 00 ) = (mk m` )(E).
i=1
Second, assume B = U Rn is open. SThen there is a sequence of pairwise

disjoint sets Ei Cn such that U = i=1 Ei . Hence U Ak A` and

X
X
(mk m` )(U ) = (mk m` )(Ei ) = mn (Ei ) = mn (U ).
i=1 i=1
Thus every open set is an element of Ak A` and so Bn Ak A` . Third,

assume B = K Rn is compact. Then there is an open set U Rn such
that K U and mn (U ) < . Hence the set V := U \ K is open. This
implies that K = U \ V Ak A` and
(mk m` )(K) = (mk m` )(U ) (mk m` )(V )

= mn (U ) mn (V )
= mn (K).
Now let B Rn be any Borel set. Then B Ak A` as we have seen above.

Moreover, it follows from Theorem 2.13 that
mn (B) = inf mn (U ) = inf (mk m` )(U ) (mk m` )(B)

U B U B
U is open U is open
and
mn (B) = inf mn (K) = inf (mk m` )(K) (mk m` )(B).

KB KB
K is compact K is compact
Hence mn (B) = (mk m` )(B) and this proves Step 1.

Step 2. Ak A` An .
We prove that
E Ak = E R` An . (7.26)
To see this, fix a set E Ak . Then there exist Borel sets A, B Bk such that
A E B and mk (B \ A) = 0. Let : Rn Rk denote the projection onto
the first k coordinates. This map is continuous and hence Borel measurable
by Theorem 1.20. Thus the sets A R` = 1 (A) and B R` = 1 (B) are
Borel sets in Rn . Moreover, by Step 1
mn ((B R` ) \ (A R` )) = mn ((B \ A) R` )
= (mk m` )((B \ A) R` )
= mk (B \ A) m` (R` )
= 0.
Since A R` E R` B R` it follows that E R` An . This
proves (7.26). A similar argument shows that
F A` = Rk F An .
Hence E F = (E R` ) (Rk F ) An for all E Ak and all F A` .
Thus Ak A` An and this proves Step 2.
Step 3. (Ak A` ) = An and (mk m` ) = mn .
Let A An . Then there are Borel sets B0 , B1 Bn such that B0 A B1
and mn (B1 \B0 ) = 0. By Step 1, B0 , B1 Ak A` and (mk m` )(B1 \B0 ) = 0.
Hence A (Ak A` ) and
(mk m` ) (A) = (mk m` )(B0 ) = mn (B0 ) = mn (A).
Thus we have proved that
An (Ak A` ) , (mk m` ) |An = mn .
Since Ak A` An by Step 2 it follows that
mn |Ak A` = (mk m` ) |Ak A` = mk m` .
Now let A (Ak A` ) . Then there are sets A0 , A1 Ak A` such that
A0 A A1 and (mk m` )(A1 \ A0 ) = 0. Hence A0 , A1 An by Step 2 and
mn (A1 \ A0 ) = 0. Since (Rn , An , mn ) is complete it follows that A \ A0 An
and so A = A0 (A \ A0 ) An . Hence An = (Ak A` ) . This proves Step 3
and Theorem 7.29.
7.5. CONVOLUTION 231
The next result specializes Theorem 7.23 to the Lebesgue measure.

Theorem 7.30 (Fubini and Lebesgue). Let k, ` N and n := k + `.
Let f : Rn R be Lebesgue integrable and, for x = (x1 , . . . , xk ) Rk and
y = (y1 , . . . , y` ) R` , define fx (y) := f y (x) := f (x1 , . . . , xk , y1 , . . . , y` ). Then
there are Lebesgue null sets E Rk and F R` such that the following holds.
(i) fx L1 (R` ) for every x Rk \ E and the map : Rk R defined by
R
f dm` , for x Rk \ E,
R` x
(x) := (7.27)
0, for x E,
is Lebesgue integrable.
(ii) f y L1 (Rk ) for every y R` \ F and the map : R` R defined by
R
Rk
f y dmk , for y R` \ F,
(y) := (7.28)
0, for y F,
is Lebesgue integrable.
Z Z Z
dmk = f dmn = dm` . (7.29)
Rk Rn R`
Proof. This follows directly from Theorem 7.23 and Theorem 7.29.
7.5 Convolution
An application of Fubinis Theorem is the convolution product on the space
of Lebesgue integrable functions on Euclidean space. Fix an integer n N
and let (Rn , A, m) be the Lebesgue measure space. The convolution of two
Lebesgue integrable functions f, g L1 (Rn ) is defined by
Z
(f g)(x) := f (x y)g(y) dm(y) for almost all x Rn .
Rn
Here the function Rn R : y 7 f (x y)g(y) is Lebesgue integrable for al-
most every x Rn and the resulting almost everywhere defined function f g
is again Lebesgue integrable. This is the content of Theorem 7.33. The con-
volution descends to a bilinear map : L1 (Rn ) L1 (Rn ) L1 (Rn ). This
map is associative and endows L1 (Rn ) with the structure of a Banach alge-
bra. Throughout we use the notation f g for two Lebesgue measurable
functions f, g : Rn R to mean that they agree almost everywhere with
respect to the Lebesgue measure.
Definition 7.31. Let f, g : Rn R be Lebesgue measurable and define

n

n the function R R : y 7 f (x y)g(y)

E(f, g) := x R . (7.30)
is not Lebesgue integrable
The convolution of f and g is the function f g : Rn R defined by
Z
(f g)(x) := f (x y)g(y) dm(y) for x Rn \ E(f, g) (7.31)
Rn
and by (f g)(x) := 0 for x E(f, g).
The next theorem shows that the convolution is very robust in that f g
is always Borel measurable and depends only on the equivalence classes of f
and g under equality almost everywhere.
Theorem 7.32. Let f, g, h, f 0 , g 0 : Rn R be Lebesgue measurable. Then
the following holds.
(i) The function y 7 f (x y)g(y) is Lebesgue measurable for all x Rn .
(ii) If f 0 f and g 0 g then E(f 0 , g 0 ) = E(f, g) and f 0 g 0 = f g.
(iii) E(f, g) is a Borel set and f g is Borel measurable.
(iv) E(g, f ) = E(f, g) and g f = f g.
(v) If m(E(f, g)) = m(E(g, h)) = 0 then
E := E(|f |, |g| |h|) = E(|f | |g|, |h|)
and f (g h) = (f g) h on Rn \ E.
An example with n = 1, where E(f, g h) = E(f g, h) = and E = R
in part (v) of Theorem 7.32, is discussed in Exercise 7.53 below.
Proof of Theorem 7.32. We prove (i). For x Rn define fx : Rn R
and x : Rn Rn by fx (y) := f (x y) and x (y) := x y. Then x
is a diffeomorphism and |det(dx )| 1. Hence Theorem 2.17 asserts that
fx = f x is Lebesgue measurable for all x Rn and this proves (i).
We prove (ii). By assumption the sets
A := {y Rn | f (y) 6= f 0 (y)} , B := {y Rn | g(y) 6= g 0 (y)} .
are Lebesgue null sets. Hence so are the sets
Cx := x (A) B = {y Rn | f (x y) 6= f 0 (x y) or g(y) 6= g 0 (y)}
for all x Rn . Hence the functions fx g and fx0 g 0 agree on the complement of
a Lebesgue null set for every x Rn . Hence they are either both integrable or
both not integrable and when they are their integrals agree. This proves (ii).
We prove (iii). By (ii) and Theorem 1.55 it suffices to assume that f and
g are Borel measurable. Now define F, G : R2n R and : R2n R2n by
F (x, y) := f (x y)g(y),
G(x, y) := f (x)g(y),
(x, y) := (x y, y)
for x, y Rn . Then G is Borel measurable and is a diffeomorphism. Hence

preserves the Borel -algebra and this implies that
F =G
is Borel measurable. Hence the function

Z
n
R [0, ] : x 7 |F (x, y)| dm(y),
Rn
is Borel measurable by Fubinis Theorem 7.28. Thus the set E(f, g) where
this function takes on the value is a Borel set. Moreover, the functions
F := max{F, 0}
are Borel measurable and so are the functions Fe : R2n [0, ) defined by

F (x, y), if x Rn \ E(f, g),
Fe (x, y) := for (x, y) R2n .
0, if x E(f, g),
Since Z Z
(f g)(x) = Fe+ (x, y) dm(y) Fe (x, y) dm(y)
Rn Rn
for all x Rn it follows from Theorem 7.28 that f g is Borel measurable.
This proves (iii).
We prove (iv). Since gx f = (fx g) x it follows from Theorem 2.17 that
E(g, f ) = x Rn | gx f L1 (Rn ) = E(f, g)

and
Z Z Z
(f g)(x) = fx g dm = (fx g) x dm = gx f dm = (g f )(x)
Rn Rn Rn
for all x Rn \ E(f, g). This proves (iv).

We prove (v). By (ii) and Theorem 1.55 it suffices to assume that f , g,

and h are Borel measurable. Let x Rn and define Fx : R2n R by
Fx (y, z) := f (z)g(x y z)h(y).
Thus Fx is the composition of the maps R2n R3n : (y, z) 7 (z, x y z, y)
and R3n R : (, , ) 7 (f (), g(), h()). Since the first map is continuous
and the second is Borel measurable it follows that Fx is Borel measurable.
We claim that
Z
x E(|f |, |g| |h|) |Fx | = x E(|f | |g|, |h|). (7.32)
R2n
It follows from Theorem 7.28 that
Z Z Z
|Fx | dm2n = |Fx (y, z)| dm(y) dm(z).
R2n Rn Rn
This integral is finite if and only if Fx L1 (R2n ). Moreover,

Z Z
|Fx (y, z)| dm(y) = |f (z)| |g(x y z)||h(y)| dm(y)
Rn Rn
= |f (z)|(|g| |h|)(x z)
for z Rn \ (x E(g, h)). Since E(g, h) is a Lebesgue null set it follows that
Z
kFx kL1 (R2n ) = |f (z)|(|g| |h|)(x z) dm(z).
Rn
The integral on the right is infinite if and only if x E(|f |, |g| |h|). This
proves the first equivalence in (7.32). The proof of the second equivalence is
analogous with y and z interchanged.
Now let x Rn \ E. Then Fx L1 (R2n ) and x Rn \ E(f, g h).
Moreover, for z Rn , the function Rn R : y 7 Fx (y, z) is integrable if
and only x z / E(g, h) and in that case its integral is equal to
Z Z
Fx (y, z) dm(y) = f (z) g(x y z)h(y) dm(y) = f (z)(g h)(x z).
Rn Rn
Since E(g, h) is a Lebesgue null set, it follows from Theorem 7.30 that
Z Z
Fx dm2n = f (z)(g h)(x z) dm(z) = (f (g h))(x)
R2n Rn
R x
The last equation holds because / E(f, g h). A similar argument with y
and z interchanged shows that R2n Fx dm2n = ((f g)h)(x) for all x Rn \E.
This proves (v) and Theorem 7.32.
Theorem 7.33. Let 1 p, q, r such that 1/p + 1/q = 1 + 1/r and let
f Lp (Rn ) and g Lq (Rn ). Then m(E(f, g)) = 0 and
kf gkr kf kp kgkq . (7.33)
Thus f g Lr (Rn ). The estimate (7.33) is called Youngs inequality.
Proof. Define the function h : Rn [0, ] by
Z
h(x) := |f (x y)g(y)| dm(y) for x Rn .
Rn
Then |f g| h and E(f, g) = {x Rn | h(x) = }. Hence it suffices to
prove that khkr kf kp kgkq . For r = this follows from Holders inequality.
So assume r < . Then 1 p, q < . Define
p p p
:= 1 = p , q 0 := .
r q
0
Then 0 < 1 and 1/q + 1/q = 1. Also = 0 if and only if q = 1. If > 0
then Holders inequality in Theorem 4.1 shows that
Z

h(x) = |fx | |fx |1 |g| dm |fx | q0 |fx |1 |g| q ,
Rn
where fx (y) := f (x y). Since q 0 = p this implies
Z q/q0 Z
q q 0
h(x) |fx | dm |fx |(1)q |g|q dm
n n
R
Z R (7.34)
q
= kf kp |f (x y)|(1)q |g(y)|q dm(y)
Rn
n
for all x R . This continues to hold for = 0. Now it follows from
Minkowskis inequality in Theorem 7.19 with the exponent s := r/q 1 that
Z q/r Z 1/s
q r qs
khkr = h dm = h dm
Rn Rn
Z Z s 1/s
kf kq
p |f (x y)| (1)q q
|g(y)| dm(y) dm(x)
Rn Rn
Z Z 1/s
kf kq
p |f (x y)| (1)qs qs
|g(y)| dm(x) dm(y)
RnRn
q (1)q
= kf kp kf kp kgkqq .
Here the last equation follows from the fact that (1 )qs = (1 )r = p.
It follows from Theorem 7.33 and part (ii) of Theorem 7.32 that the
convolution descends to a map
L1 (Rn ) L1 (Rn ) L1 (Rn ) : (f, g) 7 f g. (7.35)
This map is bilinear by Theorem 1.44, it is associative by part (v) of Theo-

rem 7.32, and satisfies kf gk1 kf k1 kgk1 by Youngs inequality in Theo-
rem 7.33. Hence L1 (Rn ) is a Banach algebra. By part (iv) of Theorem 7.32
the Banach algebra L1 (Rn ) is commutative and by Theorem 7.33 with q = 1
and r = p it acts on Lp (Rn ). (A Banach algebra is a Banach space (X , kk)
equipped with an associative bilinear map X X X : (x, y) 7 xy that
satisfies the inequality kxyk kxk kyk for all x, y X .)
Definition 7.34. Fix a constant 1 p < . A Lebesgue
R measurable func-
n p
tion f : R R is called locally p-integrable if K |f | dm < for ev-
ery compact set K Rn . It is called locally integrable if it is locally
p-integrable for p = 1.
Theorem 7.33 carries over to locally integrable functions as follows. If
1/p + 1/q = 1 + 1/r, f is locally p-integrable, and g Lq (Rn ) has compact
support, then E(f, g) is a Lebesgue null set and f g is locally r-integrable. To
see this, let K Rn be any compact set and choose a compactly supported
smooth function such that |K 1. Then f Lp (Rn ) and (f ) g agrees
with f g on the set {x Rn | x supp(g) K}. In the following theorem
C0 (Rn ) denotes the space of compactly supported smooth functions on Rn .
Theorem 7.35. Let 1 p < and 1 < q such that 1/p + 1/q = 1.
(i) If f : Rn R is locally p-integrable then
Z
lim |f (x + ) f (x)|p dm(x) = 0
0 B
for every bounded Lebesgue measurable subset B Rn . If f Lp (Rn ) this

continues to hold for B = Rn .
(ii) If f Lp (Rn ) and g Lq (Rn ) then f g is uniformly continuous. If f
is locally p-integrable and g Lq (Rn ) has compact support (or if f Lp (Rn )
has compact support and g is locally q-integrable) then f g is continuous.
(iii) If f : Rn R is locally integrable and g C0 (Rn ) then f g is smooth
and (f g) = f g for every multi-index .
(iv) C0 (Rn ) is dense in Lp (Rn ) for 1 p < .
Proof. We prove (i). Assume first that f Lp (Rn ) and fix a constant > 0.
By Theorem 4.15 there is a function g Cc (Rn ) such that kf gkp < 1/p /3.
Since g is uniformly continuous there is a > 0 such that, for all Rn ,
1/p

|| < = sup |g(x + ) g(x)| <
xRn 3p m(supp(g)
Take Rn such that || < . Then
Z 1/p
p
|f (x + ) f (x)| dm(x)
Rn
Z 1/p
p
2 kf gkp + |g(x + ) g(x)| dm(x)
Rn
1/p
21/p

p
+ m(supp(g)) sup |g(x + ) g(x)| < 1/p .
3 xRn
This proves (i) for f Lp (Rn ). To prove the result in general choose a
compact set K Rn such that B1 (x) K for all x B and multiply f by a
smooth compactly supported cutoff function to obtain a function f 0 Lp (Rn )
that agrees with f on K. Then (i) holds for f 0 and hence also for f .
We prove (ii). Assume first that f Lp (Rn ) and g Lq (Rn ) and fix a
constant > 0. By part (i) there exists a > 0 such that, for all Rn ,
Z !p
p
|| < = |f (y + ) f (y)| dm(y) <
Rn kgkq
Fix two elements x, Rn such that || < and denote fx (y) := f (x y).
Then, by Holders inequality in Theorem 4.1,
Z

|(f g)(x + ) (f g)(x)| = (fx+ fx )g dm
n R
kfx+ fx kp kgkq
Z 1/p
p
= |f (y + ) f (y)| dm(y) kgkq
Rn
< .
This shows that f g is uniformly continuous. If f is locally p-integrable and
g Lq (Rn ) has compact support continuity follows by taking the integral
over a suitable compact set. In the converse case continuity follows by taking
the Lq -norm of g over a suitable compact set. This proves (ii).
We prove (iii). Fix an index i {1, . . . , n} and denote by ei Rn the ith

unit vector. Fix an element x Rn and choose a compact set K Rn such
that B1 (y) K whenever xy supp(g). Let > 0. Since i g is continuous,
R
there is a constant 0 < < 1 such that |i g(y + hei ) i g(y)| < / K |f | dm
for all y Rn and all h R with |h| < . Hence the fundamental theorem
of calculus asserts that

g(y + hei y) g(y)
sup
i g(y) < R
yRn h K
|f | dm
for all h R with 0 < |h| < . Take h Rn with 0 < |h| < . Then

(f g)(x + hei ) (f g)(x)
(f i g)(x)
h
Z
g(x + hei y) g(x y)
= f (y) i g(x y) dm(y)
n h
ZR
g(x + hei y) g(x y)
|f (y)|
i g(x y) dm(y) < .
Rn h
By part (ii) the function i (f g) = f i g is continuous for i = 1, . . . , n.
For higher derivatives the assertion follows by induction. This proves (iii).
We prove (iv). Let f Lp (Rn ) and choose a compactly supported smooth
function : Rn [0, ) such that
Z
supp() B1 , dm = 1.
Rn
Define : Rn R by
1 x
(x) := n

for > 0 and x Rn . Then
Z
supp( ) B , dm = 1
Rn
by Theorem 2.17. By part (iii) the function

f := f : Rn R
is smooth for all > 0. Now fix a constant > 0. By part (i) there exists a
constant > 0 such that, for all y Rn ,
Z
|y| < = |f (x y) f (x)|p dm(x) < p .
Rn
7.6. MARCINKIEWICZ INTERPOLATION 239
Hence, by Minkowskis inequality in Theorem 7.19,

Z Z p 1/p

kf f kp =
f (x y) f (x) (y) dm(y) dm(x)
Rn Rn
Z Z 1/p
p p
|f (x y) f (x)| (y) dm(x) dm(y)
Rn Rn
Z 1/p
p
sup |f (x y) f (x)| dm(x) .
|y|< Rn
If f has compact support then so does f . If not, choose a function g Lp (Rn )

with compact support such that kf gkp < /2 and then a smooth function
h : Rn R with compact support such that kg hkp < /2. This proves (iv)
and Theorem 7.35.
The method explained in the proof of part (iv) of Theorem 7.35 is called
the mollifier technique. The functions can be viewed as approximate
Dirac delta functions that concentrate near the origin as tends to zero.
7.6 Marcinkiewicz Interpolation

Another interesting application of Fubinis Theorem is Marcinkiewicz inter-
polation which provides a criterion for a linear operator on L2 () to induce a
linear operator on Lp () for 1 < p < 2. Marcinkiewicz interpolation applies
to all measure spaces, although it is used here only for the Lebesgue measure
space on Rn . In particular, Marcinkiewicz interpolation plays a central role
in the proof of the CalderonZygmund inequality in Section 7.7.
Let (X, A, ) be a measure space. For a measurable function f : X R
define the function f : [0, ) [0, ] by (6.1), i.e.

f (t) := (A(t, f )), A(t, f ) := x X |f (x)| > t ,
for t 0. The function f is nonincreasing and hence Borel measurable.
Lemma 7.36. Let 1 p < and let f, g : X R be measurable. Then
f +g (t) f (t/2) + g (t/2), (7.36)
Z Z
p p
t f (t) |f | d = p sp1 f (s) ds (7.37)
X 0
for all t 0.
Proof. The inequality (7.36) was established in the proof of Lemma 6.2. We
prove (7.37) in four steps.
R
Step 1. tp f (t) X |f |p d for all t 0.
R R
Since tp A(t,f ) |f |p it follows that tp f (t) = X tp (A(t,f ) d X |f |p d for
all t 0. This proves Step 1.
R R
Step 2. If f (t) = for some t > 0 then X |f |p d = = 0 tp1 f (t) dt.
R
By Step 1, we have X |f |p d p1
R =p1. Moreover, t f (t) = for t > 0
sufficiently small and hence 0 t f (t) dt = . This proves Step 2.
Step 3. Assume (X, A, ) is -finite and f (t) < for all t > 0. Then
equation (7.37) holds.
Let B 2[0,) be the Borel -algebra and denote by m : B [0, ] the
restriction of the Lebesgue measure to B. Let (X [0, ), A B, m) be
the product measure space of Definition 7.10. We prove that

Q(f ) := (x, t) X [0, ) 0 t < |f (x)| A B.
To see this, assume first that f is an A-measurable step-function. Then
there exist finitely many pairwise disjoint measurable P sets A1 , . . . , A` A
`
and positive real numbers 1 , . . . , ` such that |f | = i=1 i Ai . In this
S`
case Q(f ) = i=1 Ai [0, i ) A B. Now consider the general case. Then
Theorem 1.26 asserts that there is a sequence of A-measurable step-functions
fi : X [0, ) such that 0 f1 f2 andSfi converges pointwise
to |f |. Then Q(fi ) A B for all i and so Q(f ) = i=1 Q(fi ) A B.
Now define h : X [0, ) [0, ) by h(x, t) := ptp1 . This function is
A B-measurable and so is hQ(f ) . Hence, by Fubinis Theorem 7.17,
!
Z Z Z |f (x)|
|f |p d = ptp1 dt d(x)
X X 0
Z Z
= (hQ(f ) )(x, t) dm(t) d(x)
X 0
Z Z

= (hQ(f ) )(x, t) d(x) dm(t)
0 X
Z
= ptp1 (A(t, f )) dt.
0
This proves Step 3.

7.6. MARCINKIEWICZ INTERPOLATION 241
Step 4. Assume f (t) < for all t > 0. Then (7.37) holds.

Define X0 := {x X | f (x) 6= 0}, A0 := A A A X0 , and 0 := |A0 .
Then the measure space (X0 , A0 , 0 ) is -finite because Xn := A(1/n, f ) is
a sequence S of An -measurable sets such that 0 (Xn ) = f (1/n) < for all n
and X0 = n=1 Xn . Moreover, f0 := f |X0 : X0 R is A0 -measurable and
f = f0 . Hence it follows from Step 3 that
Z Z Z Z
p p p1
|f | d = |f0 | d0 = t f0 (t) dt = tp1 f (t) dt.
X X0 0 0
This proves Step 4 and Lemma 7.36.
Fix real numbers 1 p q. Then the inequality

qp q(p1)
kf kp kf k1p(q1) kf kqp(q1) (7.38)
in Exercise 4.44 shows that
L1 () Lq () Lp ().
Since the intersection L1 () Lq () contains (the equivalences classes of) all

characteristic functions of measurable sets with finite measure, it is dense in
Lp () by Lemma 4.12. The following theorem was proved in 1939 by Jozef
Marcinkiewicz (a PhD student of Antoni Zygmund). To formulate the result
it will be convenient to slightly abuse notation and use the same letter f to
denote an element of Lp () and its equivalence class in Lp ().
Theorem 7.37 (Marcinkiewicz). Let q > 1 and let T : Lq () Lq () be

a linear operator. Suppose that there exist constants c1 > 0 and cq > 0 such
that
kT f k1, c1 kf k1 , kT f kq cq kf kq (7.39)
for all f L1 () Lq (). Fix a constant 1 < p < q. Then
1/p
p(q 1) qp q(p1)
kT f kp cp kf kp , cp := 2 c1p(q1) cqp(q1) , (7.40)

(q p)(p 1)
for all f L1 () Lq (). Thus the restriction of T to L1 () Lq () extends

(uniquely) to a bounded linear operator from Lp () to itself for 1 0 and let f L1 () L2 (). For t 0 define

f (x), if |f (x)| > ct, 0, if |f (x)| > ct,
ft (x) := gt (x) :=
0, if |f (x)| ct, f (x), if |f (x)| ct.
Then

A(s, f ), if s > ct, , if s ct,
A(s, ft ) = A(s, gt ) =
A(ct, f ), if s ct, A(s, f ) \ A(ct, f ), if s < ct,

f (s), if s > ct, 0, if s ct,
ft (s) = gt (s) =
f (ct), if s ct, f (s) f (ct), if s < ct.
By Lemma 7.36 and Fubinis Theorem 7.28 this implies
Z Z Z
p2 p2
t kft k1 dt = t ft (s) ds dt
0 0 0
Z Z
p2
= t ctf (ct) + f (s) ds dt
0 ct
Z Z Z s/c
1p p1
=c t f (t) dt + tp2 dt f (s) ds
Z0 Z0 0
(s/c)p1
= c1p tp1 f (t) dt + f (s) ds
0 0 p1
1p Z
c p
= tp1 f (t) dt
p1 0
c1p
Z
= |f |p d,
p1 X
Z Z Z
pq1 q pq1 q1
t kgt kq dt = t qs gt (s) ds dt
0 0 0
Z Z ct
pq1 q1
= t qs (f (s) f (ct)) ds dt
0 0
Z Z Z
pq1 q1 q
=q t dt s f (s) ds c tp1 f (ct) dt
0 s/c 0
Z p1 qp Z
s c
=q f (s) ds cqp tp1 f (t) dt
0 qp 0
cqp p p1
Z
= t f (t) dt
qp 0
cqp
Z
= |f |p d.
qp X

7.7. THE CALDERONZYGMUND INEQUALITY 243
Moreover, f = ft + gt for all t 0. Hence, by Lemma 7.36 and (7.39),

T f (t) T ft (t/2) + T gt (t/2)
2 2q
kT ft k1, + q kT gt kqq
t t
2c1 (2cq )q
kft k1 + q
kgt kqq .
t t
Hence, by Lemma 7.36 and the identities on page 242,
Z Z
p
|T f | d = p tp1 T f (t) dt
X 0
Z Z
p2c1 p2
t kft k1 dt + p(2cq ) q
tpq1 kgt kqq dt
0 0
p2c1 c1p p(2cq )q cqp
Z
= + |f |p d
p1 qp X
p (qp)/(q1) (qpq)/(q1) Z
p(q 1)2 c1 cq
= |f |p d
(q p)(p 1) X
Here the last equation follows with the choice of c := (2c1 )1/(q1) /(2cq )q/(q1) .
7.7 The Calder

onZygmund Inequality
The convolution product discussed in Section 7.5 has many applications, no-
tably in the theory of partial differential equations. One such application is
the CalderonZygmund inequality which plays a central role in the regularity
theory for elliptic equations. Its proof requires many results from measure
theory, including Fubinis Theorem, convolution, Marcinkiewicz interpola-
tion, Lebesgues differentiation theorem, and the dual space of Lp . Denote
the standard Laplace operator on Rn by
n
X 2
:= (7.41)
i=1
x2i
and, for i = 1, . . . , n, denote the partial derivative with respect to the ith
coordinate by i = /xi . Denote the open ball of radius r > 0 centered at
the origin by Br := {x Rn | |x| < r}. Call a function u : Rn R smooth
if all its partial derivatives exist and are continuous. Denote by C0 (Rn ) the
space of compactly supported smooth functions on Rn .
Definition 7.38. Fix an integer n 2. The fundamental solution of

Laplaces equation is the function K : Rn \ {0} R defined by
(2)1 log(|x|),

if n = 2,
K(x) := 1 1 2n (7.42)
(2 n) n |x| , if n > 2.
Here n denotes the area of the unit sphere S n1 Rn or, equivalently,

n /n := m(B1 ) denotes the Lebesgue measure of the unit ball in Rn . The
first and second partial derivatives Ki := i K and Kij := i j K of the
fundamental solution are given by
xi nxi xj |x|2 nx2i
Ki (x) = , Kij (x) = , Kii (x) = (7.43)
n |x|n n |x|n+2 n |x|n+2
for 1 i, j n with i 6= j. Extend the functions K, Ki , Kij to all of Rn by
setting K(0) := Ki (0) := Kij (0) := 0 for all i, j.
Exercise 7.39. Prove that K = 0. Prove that K and Ki are locally
integrable while Kij is not Lebesgue integrable over any neighborhood of the
origin. Hint: Use Fubinis Theorem in polar coordinates (Exercise 7.47).
Exercise 7.40. Prove that m(B1 ) = n /n. Prove that
(
n/2 2 n/2
2 (n/21)!
, if n is even,
n = = 2 (n+1)/2 (n1)/2 (7.44)
(n/2) 13(n2)
, if n is odd.
2
Hint: Use Fubinis Theorem to prove that Rn e|x| dm(x) = n/2 . Use polar
R
coordinates to express the integral in terms of n (Exercise 7.47).

Theorem 7.41. Fix an integer n 2 and let f C0 (Rn ). Then
f = K f. (7.45)
Moreover, the function u : Rn R, defined by

Z
u(x) := (K f )(x) = K(x y)f (y) dm(y) (7.46)
Rn
for x Rn , is smooth and satisfies
u = f, i u = Ki f for i = 1, . . . , n. (7.47)
The equations (7.45) and (7.47) are called Poissons identities.

Proof. The proof relies on Greens formula

Z Z
v u
(uv vu) dm = u v d (7.48)

for a bounded open set Rn with smooth boundary and two smooth
functions u, v : Rn R. The term
n
u X u
(x) := i (x) (x)
i=1
xi
for x denotes the outward normal derivative and : S n1 denotes

the outward pointing unit normal vector field on the boundary. The integral
over the boundary is understood with respect to the Borel measure induced
by the geometry of the ambient Euclidean space. We do not give a precise
definition because the boundary integral will only be needed here when the
boundary component is a sphere (see Exercise 7.47 below). Equation (7.48)
can be viewed as a higher-dimensional analogue of the fundamental theorem
of calculus.
Now let f C0 (Rn ) and choose r > 0 so large that supp(f ) Br . Fix
an element supp(f ) and a constant > 0 such that B () Br . Choose
:= Br \ B (), u(x) := K (x) := K( x), v := f.
Then = Br B () and the functions v, v/ vanish on Br . More-
over, K 0. Hence Greens formula (7.48) asserts that
Z Z
K f
K f dm = f K d. (7.49)
Rn \B () B ()
Here the reversal of sign arises from the fact that the outward unit nor-
mal vector on B () is inward pointing with respect to . Moreover,
(x) = |x |1 (x ) for x B (), so K /(x) = n1 1n by (7.43).
Also, by (7.42),
1
2 log(), if n = 2,
K (x) = =: () for x B ().
(2 n)1 n1 2n , if n > 2,
Z Z Z
1
K f dm = u d () f dm. (7.50)
Rn \B () n n1 B () B ()
The last summand is obtained from (7.48) with u = 1, v = f , = B ().

Now take the limit 0. Then the first term on the right in (7.50) converges
to f () and the second term converges to zero. This proves (7.45). It follows
from Theorem 7.35 and equation (7.45) that
u = (K f ) = K f = f.
To prove the second equation in (7.47) fix an index i {1, . . . , n} and a point
Rn . Then the divergence theorem on := Br \ B () asserts that
Z

Ki ( x)f (x) K( x)i f (x) dm(x)
Rn \B ()
Z

= (i K )f + K i f dm
Rn \B ()
Z
= i (K f ) dm
Rn \B ()
Z
= i K f d
B ()
xi i
Z
= () f (x) d(x)
B ()
The last term converges to zero as tends to zero. Hence
(Ki f )() = (K i f )() = i (K f )()
by Theorem 7.35. This proves Theorem 7.41.
Remark 7.42. Theorem 7.41 extends to compactly supported C 1 -functions
f : Rn R and asserts that K f is C 2 . However, this does not hold for
continuous functions with compact support. A counterexample is u(x) = |x|3
which is not C 2 and satisfies f := u = 3(n + 1)|x|. It then follows that
K f (for any C0 (Rn ) equal to one near the origin) cannot be C 2 .
Theorem 7.43 (Calder onZygmund). Fix an integer n 2 and a number
1 0 such that
n
X
ki j ukp c kukp (7.51)
i,j=1
for all u C0 (Rn ).

Proof. See page 254. The proof is based on the exposition in Gilbarg
Trudinger [5].

The CalderonZygmund inequality is a beautiful and deep theorem in the

theory of partial differential equations. It extends to all functions u = K f
with f C0 (Rn ) and thus can be viewed as a result about the convolution
operator f 7 K f . Theorem 7.35 shows that a derivative of a convolution
is equal to the convolution with the derivative. This extends to the case
where the derivative only exists in the weak sense and is locally integrable.
For the function K this is spelled out in equation (7.47) in Theorem 7.41.
Thus the convolution of an Lp function with a function whose derivatives
are integrable has derivatives in Lp . The same holds for second derivatives.
(The precise formulation of this observation requires the theory of Sobolev
spaces.) The remarkable fact is that the second derivatives of the fundamental
solution K of Laplaces equation are not locally integrable and, nevertheless,
the CalderonZygmund inequality still asserts that the second derivatives of
its convolution u = K f with a p-integrable function f are p-integrable.
Despite this subtlety the proof is elementary in the case p = 2. Denote by
u := (1 u, . . . , n u) : Rn Rn
the gradient of a smooth function u : Rn R.
Lemma 7.44. Fix an integer n 2 and let f C0 (Rn ). Then
k(Kj f )k2 kf k2 for j = 1, . . . , n. (7.52)
Proof. Define u := Kj f . This function is smooth by Theorem 7.35 but it
need not have compact support. By the divergence theorem
Z Z Z X n Z
2 u
|u| dm + uu dm = i (ui u) dm = u d (7.53)
Br Br Br i=1 Br
for all r > 0. By Poissons identities (7.45) and (7.47), we have

u = (Kj f ) = j (K f ) = j (K f ) = j f
Since f has compact support it follows from (7.43) that there is a constant
c > 0 such that |u(x)| + |u/(x)| c|x|1n for |x| sufficiently large. Hence
the integral on the right in (7.53) tends to zero as r tends to infinity. Thus
Z Z Z
2 2
kuk2 = |u| dm = uj f dm = (j u)f dm kuk2 kf k2 .
Rn Rn Rn

By Theorem 7.35 the space C0 (Rn ) is dense in L2 (Rn ). Thus Lemma 7.44
shows that the linear operator f 7 k (Kj f ) extends uniquely to a bounded
linear operator from L2 (Rn ) to L2 (Rn ). The heart of the proof of the
CalderonZygmund inequality is the following delicate argument which shows
that this operator also extends to a continuous linear operator from the Ba-
nach space L1 (Rn ) to the topological vector space L1, (Rn ) of weakly inte-
grable functions introduced in Section 6.1. This argument occupies the next
six pages. Recall the definition
kf k1, := sup tf (t),
t>0
where

A(t, f ) := x Rn |f (x)| > t .

f (t) := m(A(t, f )),
(See equation (6.1).)
Lemma 7.45. Fix an integer n 2. Then there is a constant c = c(n) > 0
such that
kk (Kj f )k1, c kf k1 (7.54)
for all f C0 (Rn ) and all indices j, k = 1, . . . , n.
Proof. Fix two integers j, k {1, . . . , n} and let T : L2 (Rn ) L2 (Rn ) be
the unique bounded linear operator that satisfies
T f = k (Kj f ) (7.55)
for f C0 (Rn ). This operator is well defined by Lemma 7.44. We prove in
three steps that there is a constant c = c(n) > 0 such that kT f k1, c kf k1
for all f L1 (Rn )L2 (Rn ). Throughout we abuse notation and use the same
letter f to denote a function in L2 (Rn ) and its equivalence class in L2 (Rn ).
Step 1. There is a constant c = c(n) 1 with the following significance. If
B Rn is a countable union of closed cubes Qi Rn with pairwise disjoint
interiors and if h L2 (Rn ) L1 (Rn ) satisfies
Z
h|Rn \B 0, h dm = 0 for all i N (7.56)
Qi
then
1
T h (t) c m(B) + khk1 (7.57)
t
for all t > 0.

For i N define hi : Rn R by

h(x), if x Qi ,
hi (x) :=
0, if x
/ Qi .
Denote by qi Qi the center of the cube Qi and by 2ri > 0 its side length.
Then |x qi | nri for all x Qi . Fix an element x Rn \ Qi . Then Kj
is smooth on x Qi and so Theorem 7.35 asserts that
(T hi )(x) = (k Kj hi )(x)
Z
(7.58)
= k Kj (x y) k Kj (x qi ) hi (y) dm(y).
Qi
This identity is more delicate than it looks at first glance. To see this, note
that the formula (7.55) only holds for compactly supported smooth func-
tions but is not meaningful for all L2 functions f because Kj f may not be
differentiable. The function hi is not smooth so care must be taken. Since
x / Qi = supp(hi ) one can approximate hi in L2 (Rn ) by a sequence of com-
pactly supported smooth functions that vanish near x (by using the mollifier
method in the proof of Theorem 7.35). For the approximating sequence
part (iii) of Theorem 7.35 asserts that the partial derivative with respect to
the kth variable of the convolution with Kj is equal to the convolution with
k Kj near x. Now the first equation in (7.58) follows by taking the limit.
The second equation follows from (7.56). It follows from (7.58) that
Z
|(T hi )(x)| |k Kj (x y) k Kj (x qi )||hi (y)| dm(y)
Qi
sup |k Kj (x y) k Kj (x qi )| khi k1
yQi

nri sup |k Kj (x y)| khi k1
yQi
1
c1 ri sup khi k1
yQi |x y|n+1
c1 r i
= khi k1 .
d(x, Qi )n+1
Here d(x, Qi ) := inf yQi |x y| and
n(n + 3)
c1 = c1 (n) := max sup |y|n+1 |k Kj (y)| .
j,k yRn \{0} n
Here the last inequality follows by differentiating equation (7.43).
Now define

Pi := x Rn |x qi | < 2 nri Qi .

Then d(x, Qi ) |x qi | nri for all x Rn \ Pi . Hence
Z Z
1
|T hi | dm c1 ri n+1 dm(x) khi k1
Rn \Pi Rn \Pi (|x qi | nri )
Z
1
= c1 r i n+1 dm(y) khi k1
|y|>2 nri (|y| nri )
Z
n sn1 ds
= c1 ri n+1 khi k1
2 nri (s nri )
Z n1
(s + nri ) ds
= c1 n ri n+1
khi k1
nri s
Z
n1 ds
c1 2 n ri 2
khi k1
nri s
= c2 khi k1 .

Here c2 = c2 (n) := c1 (n)2n1 n n 2n1 n3/2 (n + 3). The third step in
the above computation follows from Fubinis Theorem in polar coordinates
(Exercise 7.47). Thus we have proved that
Z
|T hi | dm c2 khi k1 for all i N. (7.59)
Rn \Pi
Recall that T h and T hi are only equivalence classes in L2 (Rn ). Choose

square integrable functions on Rn representing these equivalence classes and
denote them by the same letters T h, T hi L2 (Rn ). We prove that there is
a Lebesgue null set E Rn such that

X
|T h(x)| |T hi (x)| for all x Rn \ E. (7.60)
i=1
To see this, note that the sequence ì=1 hi converges to h in L2 (Rn ) as `

P
tends to infinity. So the sequence ì=1 T hi converges to T h in L2 (Rn ) by

P
Lemma 7.44. By Corollary 4.10 a subsequence converges almost everywhere.
Hence there exists a Lebesgue null set E P Rn and a sequence of integers
0 < `1 < `2 < `3 < such that the sequence ì=1

T hi (x) coverges to T h(x)
as tends to infinity for all x R \ E. Since | i=1 T hi (x)|
n `
P P
i=1 |T hi (x)|
n
for all x R , this proves (7.60).

Now define
[
A := Pi .
i=1
Then it follows from (7.59), (7.60), and Theorem 1.38 that

Z Z
X
|T h| dm |T hi | dm
Rn \A Rn \A i=1

XZ
= |T hi | dm
i=1 Rn \A
Z
X
|T hi | dm
i=1 Rn \Pi

X
c2 khi k1
i=1
= c2 khk1 .
Moreover,

X
X
m(A) m(Pi ) = c3 m(Qi ) = c3 m(B),
i=1 i=1
where
m(B2n )
c3 = c3 (n) := n
= m(Bn ) = n nn/21 .
m([1, 1] )
Hence

x Rn \ A |T h(x)| > t

tT h (t) tm(A) + tm
Z
tm(A) + |T h| dm
Rn \A
c3 tm(B) + c2 khk1

c4 tm(B) + khk1
for all t > 0, where
c4 = c4 (n) := max{c2 (n), c3 (n)} max{2n1 n3/2 (n + 3), n nn/21 }.
This proves Step 1.

Step 2 (Calder onZygmund Decomposition).

Let f L (R ) L1 (Rn ) and t > 0. Then there exists a countable collection
2 n
of closed cubes Qi Rn with pairwise disjoint interiors such that

Z
1
m(Qi ) < |f | 2n m(Qi ) for all i N (7.61)
t Qi
and
|f (x)| t for almost all x Rn \ B, (7.62)
S
where B := i=1 Qi .
For Zn and ` Z define
Q(, `) := x Rn 2` i xi 2` (i + 1) .

Let
Q := Q(, `) Zn , ` Z

and define the subset Q0 Q by

tm(Q) < Q |f | dm and, for all Q0 Q,
R
Q0 := Q Q .
Q ( Q0 = |f | dm tm(Q0 )
R
Q0
Then every decreasing sequence of cubes in Q contains at most one element

of Q0 . Hence every element of Q0 satisfies
S (7.61) and any two cubes in Q0
have disjoint interiors. Define B := QQ0 Q. We prove that
Z
1
x R \ B, x Q Q
n
= |f | dm t. (7.63)
m(Q) Q
Suppose, by contradiction, that there exists an element x Rn \B and a cube

Q Q such that x Q and tm(Q) < Q |f | dm. Then, since kf k1 < , there
R
is a maximal cube Q Q such that x Q and tm(Q) < Q |f | dm. Such a

R
maximal cube would be an element of Q0 and hence x B, a contradiction.

This proves (7.63). Now Theorem 6.14 asserts that there exists a Lebesgue
null set E Rn \ B such that every element of Rn \ (B E) is a Lebesgue
point of f . By (7.63), every point x Rn \ (B E) is the intersection point
of a decreasing sequence of cubes over which |f | has mean value at most t.
Hence it follows from Theorem 6.16 that |f (x)| t for all x Rn \ (B E).
This proves Step 2.

Step 3. Let c = c(n) 1 be the constant in Step 1. Then

kT f k1, 2n+1 + 6c kf k1

(7.64)
for all f L2 (Rn ) L1 (Rn ).
Fix a function f L2 (Rn ) L1 (Rn ) and a constant t > 0. Let the Qi be as
in Step 2 and define [
B := Qi .
i
1
R
Then m(Qi ) < t
|f | dm for all i by Step 2 and hence
Qi
Z
X 1X 1
m(B) = m(Qi ) |f | dm kf k1 .
i
t i Qi t
Define g, h : Rn R by
R
X Qi
f dm
g := f Rn \B + Qi , h := f g.
i
m(Qi )
Then
kgk1 kf k1 , khk1 2 kf k1 .
R
Moreover, h vanishes on Rn \ B and Qi h dm = 0 for all i. Hence it follows
from Step 1 that

1 3c
T h (t) c m(B) + khk1 kf k1 . (7.65)
t t
Moreover, it follows from Step 2 that |g(x)| t for almost every x Rn \ B
and |g(x)| 2n t for every x int(Qi ). Thus |g| 2n t almost everywhere.
Hence it follows from Lemma 7.36 that
2n 2n
Z Z
1 2
T g (t) 2 |g| dm |g| dm kf k1 . (7.66)
t Rn t Rn t
Now combine (7.65) and (7.66) with the inequality (7.36) in Lemma 7.36 to
obtain the estimate
2n+1 + 6c
T f (2t) T g (t) + T h (t) kf k1 .
2t
Here the splitting f = g + h depends on t but the constant c does not. Mul-
tiply the inequality by 2t and take the supremum over all t to obtain (7.64).
This proves Step 3 and Lemma 7.45.
Theorem 7.46 (Calder onZygmund). Fix an integer n 2 and a number

1 0 such that
ki (Kj f )kp c kf kp (7.67)
for all f C0 (Rn ) and all i, j = 1, . . . , n.
Proof. For p = 2 this estimate was established in Lemma 7.44 with c = 1.
Second, suppose 1 < p < 2 and let c1 (n) be the constant of Lemma 7.45.
For i, j = 1, . . . , n denote by Tij : L2 (Rn ) L2 (Rn ) the unique bounded
linear operator that satisfies Tij f = i (Kj f ) for f C0 (Rn ). Then
kTij f k1, c1 (n) kf k1 for all f C0 (Rn ) and all i, j by Lemma 7.45.
Since C0 (Rn ) is dense in L2 (Rn ) L1 (Rn ) by Theorem 7.35 it follows that
kTij f k1, c1 (n) kf k1 for all f L2 (Rn ) L1 (Rn ). Hence Theorem 7.37
(with q = 2) asserts that (7.67) holds with
1/p
p
c = c(n, p) := 2 c1 (n)2/p1 .
(2 p)(p 1)
Third, suppose 2 < p < and choose 1 < q < 2 such that 1/p + 1/q = 1.
Then it follows from Theorem 7.41, integration by parts, Holders inequality,
and from what we have just proved that, for all f, g C0 (Rn ),
Z Z
(i (Kj f ))g dm = (i j f )g dm
Rn ZRn
= f (i j g) dm
Rn
Z
= f (i (Kj g)) dm
Rn
kf kp ki (Kj g)kq
c(n, q) kf kp kgkq .
Since C0 (Rn ) is dense in Lq (Rn ) by Theorem 7.35, and the Lebesgue measure
is semi-finite, it follows from Lemma 4.34 that ki (Kj f )kp c(n, q) kf kp
for all f C0 (Rn ). This proves Theorem 7.46.
Proof of Theorem 7.43. Fix an integer n 2 and a number 1 < p < .
Let c = c(n, p) be the constant of Theorem 7.46 and let u C0 (Rn ). Then
j u = j (K u) = Kj u by Theorem 7.41. Hence it follow from Theo-
rem 7.46 with f = u that ki j ukp = ki (Kj u)kp c(n, p) kukp for
i, j = 1, . . . , n. This proves Theorem 7.43.
7.8. EXERCISES 255
7.8 Exercises
Exercise 7.47 (Lebesgue Measure on the Sphere).
For n N let (Rn , An , mn ) the Lebesgue measure space, denote the open
unit ball by B n := {x Rn | |n| < 1}, and the unit sphere by
S n1 := B n = {x Rn | |x| = 1} .
p
For A S n1 define A := {x B n1 | (x, 1 |x|2 ) A}. Prove that
the collection
AS := A S n1 | A+ , A An1

is a -algebra and that the map : AS [0, ] defined by

Z Z
1 1
(A) := p dmn1 (x) + p dmn1 (x)
A+ 1 |x|2 A 1 |x|2
for A AS is a measure. Prove Fubinis Theorem in Polar Coordinates
stated below. Use it to prove that n := (S n1 ) < .
Fubinis Theorem for Polar Coordinates: Let f : Rn R be Lebesgue
integrable. For r 0 and x S n1 define f r (x) := fx (r) := f (rx). Then
there exists a set E AS such that (E) = 0 and fx L1 ([0, )) for all
x S n1 \ E, and there exists a Lebesgue null set F [0, ) such that
f r L1 () for all r [0, ) \ F . Define g : S n1 R and h : [0, ) R
by g(x) := 0 for x E, h(r) := 0 for r F , and
Z Z
n1 n1
g(x) := r fx (r) dm1 (r), h(r) := r f r d, (7.68)
[0,) S n1
for x S n1 \ E and r [0, ) \ F . Then g L1 (), h L1 ([0, )), and

Z Z Z
f dmn = g d = h dm1 . (7.69)
Rn S n1 [0,)
n1
Hint: Define the diffeomorphism : B (0, ) {x Rn | xn > 0} by
(x, r) := (rx, r(1 |x|2 )1/2 ). Prove that det(d(x, r)) = (1 |x|2 )1/2 rn1
for x B n1 and r > 0. Use Theorem 2.17 and Fubinis Theorem 7.30.
Exercise 7.48 (Divergence Theorem). Let f : Rn R be a smooth
function. Prove that
Z Z
i f dmn = xi f (x) d(x). (7.70)
Bn S n1
Hint: Assume first that i = n. Use the fundamental theorem of calculus
and Fubinis Theorem 7.30 for Rn = Rn1 R.
Exercise 7.49. Prove that

Z Z 1/
sin(x) sin(x)
dx := lim dx = .
0 x 0 x 2
R rt
Hint: Use the identity 0 e dt = 1/r for r > 0 and Fubinis Theorem.
Exercise 7.50. Define the function f : R2 R by

sign(xy) 1, if z > 0,
f (x, y) := 2 , sign(z) := 0, if z = 0,
x + y2
1, if z < 0,

for (x, y) 6= 0 and by f (0, 0) := 0. Prove that fx , f y : R R are RLebesgue

integrable for all x,R y R. Prove that the functions R R : x 7 R fx dm1
and R R : y 7 R f y dm1 are Lebesgue integrable and
Z Z Z Z
f (x, y) dm1 (x) dm1 (y) = f (x, y) dm1 (y) dm1 (x).
R R R R
Prove that f is not Lebesgue integrable.

Exercise 7.51. Let (X, A, ) and (Y, B, ) be two -finite measure spaces
and let f L1 () and g L1 (). Define h : X Y R by
h(x, y) := f (x)g(y), for x X and y Y.
R R R
Prove that h L1 ( ) and XY h d( ) = X f d Y g d.
Exercise 7.52. Let (X, A, ) and (Y, B, ) be two -finite measure spaces
and let : A B R be any measure such that (A B) = (A)(B) for
all A A and all B B. Prove that = .
Exercise 7.53. Define : R R by

1 cos(x), for 0 x 2,
(x) :=
0, otherwise.
Define the functions f, g, h : R R by
Z x
0
f (x) := 1, g(x) := (x), h(x) := (t) dt

for x R. Prove that (f g) h = 0 and f (g h) > 0. Thus the convolu-

tion need not be associative on nonintegrable functions. Compare this with
part (v) of Theorem 7.32. Prove that E(|f | |g|, |h|) = E(|f |, |g| |h|) = R
while E(f g, h) = E(f, g h) = .
7.8. EXERCISES 257
Exercise 7.54. Let (R, A, m) be the Lebesgue measure space, let B A be

the Borel -algebra, and denote by M the Banach space of all signed Borel
measures : B [0, ) with the norm kk := ||(R). (See Exercise 5.34.)
The convolution of two signed measures , M is the map
:B R
defined by

(x, y) R2 x + y B

( )(B) := ( ) (7.71)
for B B, where
( ) := + + + + + .
(See Definition 5.13 and Theorem 5.20.) Prove the following.
(i) If , M then M and
k k kk kk .
(ii) There exists a unique element M such that
=
for all M.
(iii) The convolution product on M is commutative, associative, and dis-
tributive. Thus M is a commutative Banach algebra with unit.
(iv) If f L1 (R) and f : B R is defined by
Z
f (B) := f dm for B B
B
then f M and kf k = kf k1 .
(v) If f, g L1 (R) then
f g = f g .
(vi) Let , , M. Then = if and only if
Z Z
f d = f (x + y)d( )(x, y)
R R2
for all bounded Borel measurable functions f : R R.
(vii) If , M and B B then
Z
( )(B) = (B t) d(t).
R
Chapter 8
The Haar Measure
The purpose of this last chapter is to prove the existence and uniqueness of a
normalized invariant Radon measure on any compact Hausdorff group. In the
case of a locally compact Hausdorff group the theorem asserts the existence
of a left invariant Radon measure that is unique up to a scaling factor. An
example is the Lebesgue measure on Rn . A useful exposition is the paper by
Gert K. Pedersen [16] which also discusses the original references.
8.1 Topological Groups

Let G be a group, in multiplicative notation, with the group operation
G G G : (x, y) 7 xy, (8.1)
the unit 1l G, and the inverse map
G G : x 7 x1 . (8.2)
A topological group is a pair (G, U) consisting of a group G and a topology
U 2G
such that the group multiplication (8.1) and the inverse map (8.2) are con-
tinuous. Here the continuity of the group multiplication (8.1) is understood
with respect to the product topology on G G (see Appendix B). A lo-
cally compact Hausdorff group is a topological group (G, U) such that
the topology is locally compact and Hausdorff (see page 81).
259
260 CHAPTER 8. THE HAAR MEASURE
Example 8.1. Let G be any group and define U := {, G}. Then (G, U) is
a compact topological group but is not Hausdorff unless G = {1l}.
Example 8.2 (Discrete Groups). Let G be any group. Then the pair
(G, U) with the discrete topology U := 2G is a locally compact Hausdorff
group, called a discrete group. Examples of discrete groups (where the
discrete topology appears naturally) are the additive group Zn , the multi-
plicative group SL(n, Z) of integer n n-matrices with determinant one, the
mapping class group of isotopy classes of diffeomorphisms of any manifold,
and every finite group.
Example 8.3 (Lie Groups). Let G GL(n, C) be any subgroup of the

general linear group of invertible complex n n-matrices that is closed as a
subset of GL(n, C) with respect to the relative topology, i.e. GL(n, C) \ G is
an open set in Cnn . Let U 2G be the relative topology on G, i.e. U G is
open if and only if there is an open subset V Cnn such that U = G V .
Then (G, U) is a locally compact Hausdorff group. In fact, it is a basic
result in the theory of Lie groups that every closed subgroup of GL(n, C) is a
smooth submanifold of Cnn and hence is a Lie group. Specific examples of
Lie groups are the general linear groups GL(n, R) and GL(n, C), the special
linear groups SL(n, R) and SL(n, C) of real and complex n n-matrices with
determinant one, the orthogonal group O(n) of matrices x Rnn such that
xT x = 1l, the special orthogonal group SO(n) := O(n)SL(n, R), the unitary
group U(n) of matrices x Cnn such that x x = 1l, the special unitary group
SU(n) := U(n) SL(n, C), the group Sp(1) of the unit quaternions, the unit
circle S 1 = U(1) in the complex plane, the torus Tn := S 1 S 1 (n
times), or, for any multi-linear form : (Cn )k C, the group of all matrices
x GL(n, C) that preserve . The additive groups Rn and Cn are also Lie
groups. Lie groups form an important class of locally compact Hausdorff
groups and play a central role in differential geometry.
Example 8.4. If (V, kk) is a normed vector space (Example 1.11) then the
additive group V is a Hausdorff topological group. It is locally compact if
and only if V is finite-dimensional.
Example 8.5. The rational numbers Q with the additive structure form a
Hausdorff topological group with the relative topology as a subset of R. It is
totally disconnected (every connected component is a single point) but does
not have the discrete topology. It is not locally compact.
8.1. TOPOLOGICAL GROUPS 261
Example 8.6 (p-adic Integers). Fix a prime number p N and denote by
N0 := N {0}
the set of nonnegative integers. For x, y Z define

n o
dp (x, y) := |x y|p := inf pk k N0 , x y pk Z . (8.3)

Then the function

dp : Z Z [0, 1]
is a distance function and so (Z, dp ) is a metric space. It is not complete.
Its completion is denoted by Zp and called the ring of p-adic integers.
Here is another description of the p-adic integers. Consider the sequence of
projections
k+1 k1
Z/pk Z
k
Z/pk1 Z
3
Z/p2 Z
2 1
Z/pZ {1}.
The inverse limit of this sequence of maps is the set of sequences

n o
Zp := x = (xk )kN0 xk Z/pk Z, k (xk ) = xk1 for all k N .

This set is a commutative ring with unit. Addition and multiplication are
defined term by term, i.e.
x + y := (xk + yk )kN0 , xy := (xk yk )kN0
for x = (xk )kN0 Zp and y = (yk )kN0 Zp . The ring of p-adic integers is
a compact metric space with
n o
k
dp (x, y) := inf p k N0 , xk = yk . (8.4)
The inclusion of Z into the p-adic integers is given by
p : Z Zp , p (x) := (x mod pk )kN .
This is an isometric embedding with respect to the distance functions (8.3)

and (8.4). The additive p-adic integers form an uncountable compact Haus-
dorff group (with the topology of a Cantor set) that is not a Lie group.
Example 8.7 (p-adic Rationals). Fix a prime number p N. Write a

nonzero rational number x Q in the form x = pk a/b where k Z and the
numbers a Z and b N are relatively prime to p, and define |x|p := pk .
For x = 0 define |0|p := 0. Define the function dp : Q Q [0, ) by
dp (x, y) := |x y|p
n
k
ka
o (8.5)
:= inf p k Z, x y = p , a Z, b N \ pN .
b
Then (Q, dp ) is a metric space. The completion of Q with respect to dp is
denoted by Qp and is called the field of p-adic rational numbers. It
can also be described as the quotient field of the ring of p-adic integers in
Example 8.6. The multiplicative group of nonzero p-adic rationals is a locally
compact Hausdorff group that is not a Lie group. One can also consider
groups of matrices whose entries are p-adic rationals. Such groups play an
important role in number theory.
Example 8.8 (Infinite Products). Let I be any index set and, for i I,
let Gi be a compact Hausdorff group. Then the product
Y
G := Gi
iI
is a compact Hausdorff group. Its elements are maps I tiI Gi : i 7 xi

such that xi Gi for all i I. Write such a map as x = (xi )iI . The
product topology on G is defined as the smallest topology such that the
Q projections i : G Gi are continuous. Thus the (infinite) products
obvious
U = iI Ui of open sets Ui Gi , such that Ui = Gi for all but finitely
many i, form a basis for the topology of G. (See Appendix B for #I = 2.)
The product topology is obviously Hausdorff and Tychonoff s Theorem
asserts that it is compact (see Munkres [14]). An uncountable product of
nontrivial groups Gi is not first countable.
Example 8.9. Let (X , kk) be a Banach algebra with a unit 1l and the
product X X X : (x, y) 7 xy. (See page 236.) Then the group of
invertible elements G := {x X | y X such that xy = yx = 1l} is a Haus-
dorff topological group. Examples include the group of nonzero quaternions,
the general linear group of a finite-dimensional vector space, the group of
bijective bounded linear operators on a Banach space, and the multiplicative
group of nowhere vanishing real valued continuous functions on a compact
topological space. In general G is not locally compact.
8.2. HAAR MEASURES 263
8.2 Haar Measures

Throughout let G be a locally compact Hausdorff group, in multiplicative
notation, and denote by B 2G its Borel -algebra. We begin our discussion
with a technical lemma about continuous functions on G.
Lemma 8.10. Let f Cc (G) and fix a constant > 0. Then there exists an
open neighborhood U of 1l such that, for all x, y G,
x1 y U = |f (x) f (y)| < . (8.6)
Proof. Choose
an open neighborhood U 0 G with compact closure and
1
define K := ab | a supp(f ), b U 0 . This set is compact because the
maps (8.1) and (8.2) are continuous. Also,
/ K, x1 y U0
x = f (x) = f (y) = 0 (8.7)
for all x, y G. (If y supp(f ) and x1 y U0 then x = y(x1 y)1 K.)
Since f is continuous there exists, for each x K, an open neighborhood
V (x) G of x such that

y V (x) = |f (x) f (y)| < . (8.8)
2
Since the map G G : y 7 x1 y is a homeomorphism, the set x1 V (x) is an
open neighborhood of 1l for every x K. Since the map (8.1) is continuous
it follows from the definition of the product topology in Appendix B that,
for every x K, there exists an open neighborhood U (x) G of 1l such that
the product neighborhood U (x) U (x) of the pair (1l, 1l) is contained in the
pre-image of x1 V (x) under the multiplication map (8.1). In other words,
a, b U (x) = xab V (x). (8.9)
Since the map G G : y 7 xy is a homeomorphism the set xU (x) is
an open neighborhood of x for every x K. Since K is compact there
exist finitely many elements x1 , . . . , x` K such that K ì=1 xi U (xi ).
S
Define U := U0 ì=1 U (xi ) and let x, y G such that x1 y U . If x

T
/K
then f (x) = f (y) = 0 by (8.7). Hence assume x K. Then there exists
an index i {1, . . . , `} such that x xi U (xi ) and therefore x1 i x U (xi ).
Hence x = xi (x1i x)1 l V (x i ) and y = x (x
i i
1
x)(x 1
y) V (x i ) by (8.9).
|f (x) f (y)| |f (x) f (xi )| + |f (xi ) f (y)| < .
For x G define the homeomorphisms Lx , Rx : G G by

Lx (a) := xa, Rx (a) := ax for x G. (8.10)
They satisfy
Lx Ly = Lxy , Rx Ry = Ryx , Lx Ry = Ry Lx . (8.11)
For A G and x G define
A1 := a1 a A .

xA := xa a A , Ax := ax a A ,
Thus xA = Lx (A) and Ax = Rx (A).
Definition 8.11. A measure : B [0, ] is called left invariant if
(xB) = (B) for all B B and all x G. It is called right invariant if
(Bx) = (B) for all B B and all x G. It is called invariant if it is
both left and right invariant. A left Haar measure on G is a left invariant
Radon measure that does not vanish identically. A right Haar measure on
G is a right invariant Radon measure that does not vanish identically. An
invariant Haar measure on G is an invariant Radon measure that does
not vanish identically.
Theorem 8.12 (Haar). Let G be a locally compact Hausdorff group. Then
(i) G admits a left Haar measure , unique up to a positive factor. Every
such measure satisfies (U ) > 0 for every nonempty open set U G.
(ii) G admits a right Haar measure , unique up to a positive factor. Every
such measure satisfies (U ) > 0 for every nonempty open set U G.
(iii) Assume G is compact. Then G admits a unique invariant Haar measure
such that (G) = 1. This measure satisfies (B 1 ) = (B) for all B B
and (U ) > 0 for every nonempty open set U G.
Examples of Haar measures are the restriction to the Borel -algebra of
the Lebesgue measure on Rn (where the group structure is additive), the
restriction to the Borel -algebra of the measure on S 1 = U(1) or on
S 3 = Sp(1) in Exercise 7.47, the measure dt/t on the multiplicative group of
the positive real numbers, and the counting measure on any discrete group.
The proof of Theorem 8.12 rests on the Riesz Representation Theorem 3.15
and the following result about positive linear functionals.
Definition 8.13. Let G be a locally compact Hausdorff group. A linear

functional : Cc (G) R is called left invariant if
(f Lx ) = (f ) (8.12)
for all f Cc (G) and all x G. It is called right invariant if
(f Rx ) = (f ) (8.13)
for all f Cc (G) and all x G. It is called invariant if it is both left and
right invariant. It is called a left Haar integral if it is left invariant, posi-
tive, and nonzero. It is called a right Haar integral if it is right invariant,
positive, and nonzero.
Theorem 8.14 (Haar). Every locally compact Hausdorff group G admits a
left Haar integral, unique up to a positive factor. Moreover, if : Cc (G) R
is a left Haar integral and f Cc (G) is a nonnegative function that does not
vanish identically then (f ) > 0.
The proof of Theorem 8.14 given below follows the notes of Pedersen [16]
which are based on a proof by Weil. Our exposition benefits from elegant
simplifications due to Urs Lang [11]. In preparation for the proof it is con-
venient to introduce some notation. Let

Cc+ (G) := f Cc (G) f 0, f 6 0

(8.14)
be the space of nonnegative continuous functions with compact support that

do not vanish identically. A function : Cc+ (G) [0, ) is called
additive iff (f + g) = (f ) + (g) for all f, g Cc+ (G),
subadditive iff (f + g) (f ) + (g) for all f, g Cc+ (G),
homogeneous iff (cg) = c(f ) for all f Cc+ (G) and all c > 0,
monotone iff f g implies (f ) (g) for all f, g Cc+ (G),
left invariant iff (f Lx ) = (f ) for all f Cc+ (G) and all x G.
Every additive functional : Cc+ (G) [0, ) is necessarily homogeneous
and monotone. Moreover, every positive linear functional on Cc (G) restricts
to an additive functional : Cc+ (G) [0, ) and, conversely, every addi-
tive functional : Cc+ (G) [0, ) extends uniquely to a positive linear
functional on Cc (G). This is the content of the next lemma.
Lemma 8.15. Let : Cc+ (G) [0, ) be an additive functional. Then there
is a unique positive linear functional on Cc (G) whose restriction to Cc+ (G)
agrees with . If is left invariant then so is its linear extension to Cc (G).
Proof. We prove that is monotone. Let f, g Cc+ (G) such that f g. If
f 6= g then g f Cc+ (G) and hence
(f ) (f ) + (g f ) = (g)
by additivity. If f = g there is nothing to prove.

We prove that is homogeneous. Let f Cc+ (G). Then (nf ) = n(f )
for all n N by additivity and induction. If c = m/n is a positive ratio-
nal number then (f ) = n(f /n) and hence (cf ) = m(f /n) = c(f ).
If c > 0 is irrational then it follows from monotonicity that
a(f ) = (af ) (cf ) (bf ) = b(f )
for all a, b Q with 0 < a < c < b, and this implies (cf ) = c(f ).
Now define (0) := 0 and, for f Cc (G), define (f ) := (f + ) (f ).
If f, g Cc (G) then f + + g + + (f + g) = f + g + (f + g)+ , hence
(f + ) + (g + ) + ((f + g) ) = (f ) + (g ) + ((f + g)+ )
by additivity, and hence (f ) + (g) = (f + g). Moreover, (f )+ = f

and (f ) = f + and so (f ) = (f ) (f + ) = (f ). Hence it follows
from homogeneity that (cf ) = c(f ) for all f Cc (G) and all c R. This
shows that the extended functional is linear.
If the original functional : Cc+ (G) [0, ) is left-invariant then so is
the extended linear functional on Cc (G) because (f Lx ) = f Lx for all
f Cc (G) and all x G. This proves Lemma 8.15.
Consider the space

is subadditive, monotone,
L := : Cc (G) (0, )
+ . (8.15)
homogeneous, and left invariant
The strategy of the proof of Theorem 8.14 is to construct certain functionals

g L associated to functions g Cc+ (G) supported near the identity ele-
ment and to construct the required positive linear functional : Cc (G) R
as a suitable limit where the functions g converge to a Dirac -function at the
identity. The precise definition of the g involves the following construction.
Denote by P the set of all Borel measures : B [0, ) of the form

k
X
:= i xi (8.16)
i=1
where k N, 1 , . . . , k are positive real numbers, x1 , . . . , xk G, and xi is

the Dirac measure at xi (see Example 1.31). The norm of a measure P
of the form (8.16) is defined by
k
X
kk := (G) = i > 0. (8.17)
i=1
If := `j=1 j yj P is any other such measure define the convolution

P
product of and by
k X
X `
:= i j x i y j .
i=1 j=1
This product is not commutative in general. It satisfies k k = kk kk.

Associated to a measure P of the form (8.16) are two linear operators
L , R : Cc (G) Cc (G) defined by
k
X k
X
(L f )(a) := i f (xi a), (R f )(a) := i f (axi ) (8.18)
i=1 i=1
for f Cc (G) and a G. The next two lemmas establish some basic
properties of the operators L and R . Denote by
kf k := sup|f (x)|
xG
the supremum norm of a compactly supported function f : G R.

Lemma 8.16. Let , P, f Cc (G), and x G. Then
L L = L f Lx = Lx f, kL f k kk kf k ,
R R = R , f Rx = Rx f, kR f k kk kf k ,
L R = R L .
Proof. The assertions follow directly from the definitions.

Lemma 8.17. Let f, g Cc+ (G). Then there exists a P such that
f L g.
Proof. Fix an element y G such that g(y) > 0. For x G define

f (x) + 1 1
Ux := a G f (a) <
g(yx a)
g(y)
This set is an open neighborhood of x. Since f has compact support there
exist finitely many points x1 , . . . , xk G such that supp(f ) Ux1 Uxk .
Define
k
X f (xi ) + 1
:= yx1 .
i=1
g(y) i
Then
k
X f (xi ) + 1
f (a) g(yx1
i a) = (L g)(a)
i=1
g(y)
for all a supp(f ) and hence f L g. This proves Lemma 8.17.
Proof of Theorem 8.14. The proof has five steps. Step 1 is the main con-
struction of a subadditive functional Mg : Cc+ (G) (0, ) associated to a
function g Cc+ (G). Step 2 shows that Mg is asymptotically linear as g
concentrates near the unit 1l. The heart of the convergence proof is Step 3
and is due to Cartan. Step 4 proves uniqueness and Step 5 proves existence.
Step 1. For f Cc+ (G) define
Mg (f ) := M (f ; g) := inf kk P, f L g .

(8.19)

(i) M (f ; g) > 0 for all f, g Cc+ (G).
(ii) For every g Cc+ (G) the functional Mg : Cc+ (G) (0, ) is subadditive,
homogeneous, monotone, and left invariant and hence is an element of L .
(iii) Let L . Then
(f ) M (f ; g)(g) for all f, g Cc+ (G). (8.20)
In particular, M (f ; h) M (f ; g)M (g; h) for all f, g, h Cc+ (G).

(iv) M (f ; f ) = 1 for all f Cc+ (G).
We prove part (ii). Monotonicity follows directly from the definition. Homo-
geneity follows from the identities Lc g = cL g and kck = c kk. To prove
left invariance observe that
(L g) Lx = Lx g, k x k = kk
for all P by Lemma 8.16. Since f L g if and only if f Lx (L g)Lx

this proves left invariance. To prove subadditivity, fix a constant > 0 and
choose , 0 P such that

f L g, f 0 L0 g, kk < M (f ; g) + , k0 k < M (f 0 ; g) + .
2 2
Then f + f 0 L g + L0 g = L+0 g and hence
M (f + f 0 ; g) k + 0 k = kk + k0 k < M (f ; g) + M (f 0 ; g) + .
Thus M (f + f 0 ; g) < M (f ; g) + M (f 0 ; g) + for all > 0. This proves

subadditivity and part (ii).
We prove part (iii). Fix a functional L. We prove first that
(L f ) kk (f ) (8.21)
for all f Cc+ (G) and all P. To see this write =

Pk
i=1 i xi . Then
L f = ki=1 i (f Lxi ) and hence
P
k
X k
X k
X
(L f ) (i (f Lxi )) = i (f Lxi ) = i (f ) = kk (f )
i=1 i=1 i=1
Here the first step follows from subadditivity, the second step follows from
homogeneity, the third step follows from left invariance, and the last step
follows from the definition of kk. This proves (8.21). Now let f, g Cc+ (G).
By Lemma 8.17 there is a P such that f L g. Since is monotone
this implies (f ) (L g) kk (g) by (8.21). Now take the infimum
over all P such that f L g to obtain (f ) M (f ; g)(g).
We prove parts (i) and (iv). Since the map Cc+ (G) (0, ) : f 7 kf k
is an element of L it follows from part (iii) that
kf k M (f ; g) kgk (8.22)
and hence M (f ; g) > 0 for all f, g Cc+ (G). Next observe that M (f ; f ) 1
by (8.22) and M (f ; f ) 1 because f = L1l f . This proves Step 1.
Step 2. Let f, f 0 Cc+ (G) and let > 0. Then there is an open neighborhood
U G of 1l such that every g Cc+ (G) with supp(g) U satisfies
Mg (f ) + Mg (f 0 ) < (1 + )Mg (f + f 0 ). (8.23)
By Urysohns Lemma A.1 there is a function Cc (G) such that (x) = 1
for all x supp(f ) supp(f 0 ). Choose a constant 0 < 1/2 such that
2 + 2 kf + f 0 k M (; f + f 0 ) < . (8.24)
Define
h := f + f 0 + kf + f 0 k .
Then f /h and f 0 /h extend to continuous functions on G with compact sup-
port by setting them equal to zero on G \ supp(). By Lemma 8.10 there
exists an open neighborhood U G of 1l such that
f (x) f (y) f 0 (x) f 0 (y)

1
x yU = h(x) h(y) + h(x) h(y) <

for all x, y G. Let g Cc+ (G) with supp(g) U . If = ì=1 i xi P

P
such that h L g then, for all a supp(f ),
` `
f (x1

L g(a) X f (a) X
i )
f (a) f (a) = i g(xi a) i 1 + g(xi a).
h(a) i=1
h(a) i=1
h(x i )
1
f (xi )
Thus f L g, where := ì=1 i h(x1
P
)
+ xi . This implies
i
`
f (x1

i )
X
Mg (f ) i + .
i=1
h(x1
i )
The same inequality holds for f . Since f + f 0 h we find

0
`
f (x1 0 1

i ) + f (xi )
X
0
Mg (f ) + Mg (f ) i + 2 kk (1 + 2).
i=1
h(x1
i )
Now take the infimum over all P such that h L g to obtain

Mg (f ) + Mg (f 0 ) (1 + 2)Mg (h)
(1 + 2) Mg (f + f 0 ) + kf + f 0 k Mg ()

1 + 2 + 2 kf + f 0 k M (; f + f 0 ) Mg (f + f 0 )

(1 + )Mg (f + f 0 ).
Here we have used the inequalities 1 + 2 2 and (8.24). This proves Step 2.
Step 3. Let f Cc+ (G) and let > 0. Then there is an open neighborhood
U G of 1l with the following significance. For every g Cc+ (G) such that
supp(g) U, g(x) = g(x1 ) for all x G, (8.25)
there exists an open neighborhood W G of 1l such that every h Cc+ (G)
with supp(h) W satisfies the inequality
M (f ; g)Mh (g) (1 + )Mh (f ). (8.26)
This inequality continues to hold with Mh replaced by any left invariant pos-
itive linear functional : Cc (G) R.
By Urysohns Lemma A.1 there is a function Cc+ (G) such that (x) = 1
for all x K := supp(f ). Choose 0 and 1 such that
1 + 0 0
0 < 0 < 1, 1 + , 1 := . (8.27)
1 0 2M (; f )
By Lemma 8.10 there exists an open neighborhood U G of 1l such that
x1 y U = |f (x) f (y)| < 1 (8.28)
for all x, y G. We prove that the assertion of Step 3 holds with this
neighborhood U . Fix a function g Cc+ (G) that satisfies (8.25). Define
1
2 := . (8.29)
2M (f ; g)
Use Lemma 8.10 to find an open neighborhood V G of 1l such that
xy 1 V = |g(x) g(y)| < 2 (8.30)
for all x, y G. Then the sets xV for x K form an open coverSof K. Hence
there exist finitely many points x1 , . . . , x` K such that K ì=1 xi V . By
Theorem A.4 there exist functions 1 , . . . , ` Cc+ (G) such that
`
X
0 i 1, supp(i ) xi V, i |K 1. (8.31)
i=1
It follows from Step 2 by induction that there exists an open neighborhood

W G of 1l such that every h Cc+ (G) with supp(h) W satisfies
`
X
Mh (i f ) < (1 + 0 ) Mh (f ). (8.32)
i=1
We prove that every h Cc+ (G) with supp(h) W satisfies (8.26).

For x G define the function Fx Cc (G) by

Fx (y) := f (y)g(y 1 x) for y G. (8.33)
It follows from (8.25) and (8.28) that f (x)g(y 1 x)f (y)g(y 1 x) 1 g(y 1 x)
for all x, y G. Since g(y 1 x) = g(x1 y) = (g Lx1 )(y) by (8.25), this
implies f (x)gLx1 Fx +1 gLx1 . Hence, for all x G and all h Cc+ (G),
f (x)Mh (g) Mh (Fx ) + 1 Mh (g) (8.34)
Now fix a function h Cc+ (G) with supp(h) W . By (8.30) and (8.31),
i (y)Fx (y) = i (y)f (y)g(y 1 x) i (y)f (y) g(x1

i x) + 2
P
for all x, y G and all i = 1, . . . , `. Since Fx = i i Fx this implies
X X
Mh (i f ) g(x1

Mh (Fx ) Mh (i Fx ) i x) + 2
i i
X (8.35)
Mh (i f ) g(x1
i x) + 22 Mh (f ).
i
Here the last step uses (8.32). By (8.34) and (8.35),

X
f (x)Mh (g) Mh (i f ) g(x1
i x) + 22 Mh (f ) + 1 Mh (g)
i
X
Mh (i f ) g(x1
i x) + 21 Mh (g).
i
Here the second step uses (8.29) and the inequality Mh (f ) M (f ; g)Mh (g).
+ P
Thus (f 21 ) Mh (g) L g, where := i Mh (i f ) x1 i
. This implies
X
Mg ((f 21 )+ )Mh (g) Mh (i f ) (1 + 0 )Mh (f ).
i
Here the second step uses (8.32). Since f (f 21 )+ + 21 we have

Mg (f )Mh (g) Mg ((f 21 )+ )Mh (g) + 21 Mg ()Mh (g)
(1 + 0 )Mh (f ) + 21 M (; f )Mg (f )Mh (g)
= (1 + 0 )Mh (f ) + 0 Mg (f )Mh (g).
Hence
1 + 0
Mg (f )Mh (g) Mh (f ) (1 + )Mh (f )
1 0
and this proves Step 3 for Mh . This inequality and its proof carry over to
any left invariant positive linear functional : Cc (G) R.
Step 4. We prove uniqueness.

Let , 0 : Cc (R) R be two left invariant positive linear functionals that
do not vanish identically. Then there exists a function f Cc+ (G) such that
(f ) > 0 by Lemma 8.15. Hence
(g) M (f ; g)1 (f ) > 0
for all g Cc+ (G) by (8.20). The same argument shows that 0 (g) > 0 for
all g Cc+ (G).
Now fix two functions f, f0 Cc+ (G) and a constant > 0. Choose an
open neighborhood U G of 1l that satisfies the requirements of Step 3 for
both f and f0 and this constant . By Urysohns Lemma A.1 there exists a
function g Cc+ (G) such that
supp(g) x G | x U and x1 U .

g(1l) > 0,
Replacing g by the function x 7 g(x) + g(x1 ), if necessary, we may assume

that g satisfies (8.25). Hence it follows from Step 1 and Step 3 that
(f ) M (f ; g)(g) (1 + )(f )
and
(1 + )(f0 ) M (f0 ; g)(g) (f0 ).
Take the quotient of these inequalities to obtain
(f ) M (f ; g) (f )
(1 + )1 (1 + ) .
(f0 ) M (f0 ; g) (f0 )
The same inequality holds with replaced by 0 . Hence
2 (f ) 0 (f ) (f )
(1 + ) 0 (1 + )2 .
(f0 ) (f0 ) (f0 )
Since this holds for all > 0 it follows that

0 (f0 )
0 (f ) = c(f ), c := .
(f0 )
Since this holds for all f Cc+ (G) it follows that 0 and c agree on Cc+ (G).
Hence 0 = c by Lemma 8.15. This proves Step 4.
Step 5. We prove existence.

The proof follows the elegant exposition [11] by Urs Lang. Fix a reference
function f0 Cc+ (G) and, for g Cc+ (G), define g : Cc+ (G) (0, ) by
M (f ; g)
g (f ) := for f Cc+ (G). (8.36)
M (f0 ; g)
Then g L for all g Cc+ (G) by Step 1. It follows also from Step 1 that
M (f0 ; g) M (f0 ; f )M (f ; g) and M (f ; g) M (f ; f0 )M (f0 ; g) and hence
M (f0 ; f )1 g (f ) M (f ; f0 ) (8.37)
for all f, g Cc+ (G). Fix a function f Cc+ (G) and a number > 0. Define

(f0 ) = 1 and there exists a neighborhood
L (f ) := L W G of 1l such that for all h Cc+ (G)

.
supp(h) W = (f ) (1 + )h (f )

We prove that L (f ) 6= . To see this let U G be the open neighborhood

of 1l constructed in Step 3 for f and . Choose a function g Cc+ (G) that
satisfies (8.25) and choose an open neighborhood W G of 1l associated to g
that satisfies the requirements of Step 3. Let h Cc+ (G) with supp(h) W .
Then M (f ; g)M (g; h) (1 + )M (f ; h) and M (f0 ; g)M (g; h) M (f0 ; h)
by Step 3 and Step 1. Take the quotient of these inequalities to obtain
g (f ) (1 + )h (f ). Since g (f0 ) = 1 it follows that g L (f ). This
shows that L (f ) 6= as claimed. Next we observe that
(f ) M (f ; f0 )(f0 ) = M (f ; f0 )
for all L (f ) by Step 1. Hence the supremum
(f ) := sup {(f ) | L (f )} (8.38)
is a real number, bounded above by M (f ; f0 ). Since L (f ) contains an
element of the form g for some g Cc+ (G) it follows from (8.37) that
M (f0 ; f )1 (f ) M (f ; f0 ) (8.39)
for all f Cc+ (G) and all > 0. Moreover, the function 7 (f ) is
nondecreasing by definition. Hence the limit
0 (f ) := lim (f ) = inf (f ) (8.40)
0 >0
exists and is a positive real number for every f Cc+ (G).

We prove that the functional 0 : Cc+ (G) (0, ) is left invariant. To

see this, fix a function f Cc+ (G) and an element x G. Then
L (f ) = L (f Lx )
for all > 0. Namely, if W G is an open neighborhood of 1l such that

(f ) (1 + )h (f ) for all h Cc+ (G) with supp(h) W , then the same
inequality holds with f replaced by f Lx because both and h are left
invariant. Hence (f ) = (f Lx ) for all > 0 and so 0 (f ) = 0 (f Lx ).
We prove that the functional 0 : Cc+ (G) (0, ) is additive. To see
this, fix two functions f, f 0 Cc+ (G). We prove that
(1 + )1 (f + f 0 ) (f ) + (f 0 ) (1 + ) (f + f 0 ) (8.41)
for all > 0. To prove the first inequality in (8.41) choose any functional
L (f + f 0 ). Then there exists an open neighborhood W G of 1l such
that (f + f 0 ) (1 + )h (f + f 0 ) for all h Cc+ (G) with supp(h) W .
Moreover, we have seen above that h Cc+ (G) can be chosen such that
supp(h) W and also h L (f ) L (f 0 ). Any such h satisfies
(1 + )1 (f + f 0 ) h (f + f 0 ) h (f ) + h (f 0 ) (f ) + (f 0 ).
Take the supremum over all L (f + f 0 ) to obtain the first inequality

in (8.41). To prove the second inequality in (8.41) fix a constant > 1
and choose two functionals L (f ) and 0 L (f 0 ). Then there exists
an open neighborhood W G of 1l such that (f ) (1 + )h (f ) and
0 (f 0 ) (1 + )h (f 0 ) for all h Cc+ (G) with supp(h) W . By Step 2,
the function h Cc+ (G) can be chosen such that supp(h) W and also
h (f ) + h (f 0 ) h (f + f 0 ) and h L (f + f 0 ). Any such h satisfies
(1 + )1 (f ) + 0 (f 0 ) h (f ) + h (f 0 ) h (f + f 0 ) (f + f 0 ).

Take the supremum over all pairs of functionals L (f ) and 0 L (f 0 )

to obtain (1 )1 (f ) + (f 0 ) (f + f 0 ) for all > 1. This proves
the second inequality in (8.41). Take the limit 0 in (8.41) to obtain
that 0 is additive. Moreover, it follows directly from the definition that
0 (f0 ) = 1. Hence it follows from Lemma 8.15 that 0 extends to a nonzero
left invariant positive linear functional on Cc (G). This proves Step 5 and
Theorem 8.14.
If one is prepared to use some abstract concepts from general topology

then the existence proof in Theorem 8.14 is essentially complete after Step 2.
This approach is taken in Pedersen [16]. In this language the space

G := g Cc+ (G) 0 g 1, g(1l) = 1

is a directed set equipped with a map g 7 g that takes values in the space

M (f0 ; f )1 (f ) M (f0 ; f )

+
L := : Cc (G) R .
for all f Cc+ (G)
The map G L : g 7 g is a net. A net can be thought of as an uncountable
analogue of a sequence and a subnet as an analogue of a subsequence. The
existence of a universal subnet is guaranteed by the general theory and its
convergence for each f by the fact that the target space is compact. Instead
Step 3 in the proof of Theorem 8.14 implies that the original net g 7 g
converges and so there is no need to choose a universal subnet. That this
can be proved with a refinement of the uniqueness argument ( in Step 3) is
pointed out in Pedersen [16]. That paper also contains two further uniqueness
proofs. One is based on Fubinis Theorem and the other on the Radon
Nikodym Theorem. Another existence proof for compact second countable
Hausdorff groups is due to Pontryagin. It uses the ArzelàAscoli theorem to
establish the existence of a sequence i P with ki k = 1 such that Li f
converges to a constant function whose value is then taken to be (f ).
Proof of Theorem 8.12. Existence and uniqueness in (i) follow directly from
Theorem 8.14 and the Riesz Representation theorem 3.15. That nonempty
open sets have positive measure follows from Urysohns Lemma A.1. To
prove (ii) consider the map : G G defined by (x) := x1 for x G.
Since is a homeomorphism it preserves the Borel -algebra B. Since
Rx = Lx1 , a measure : B [0, ] is a left Haar measure if and
only if the measure : B [0, ] defined by (B) := ((B)) = (B 1 ) is
a right Haar measure. Hence assertion (ii) follows from (i).
We prove (iii). Assume G is compact and let : B [0, 1] be the unique
left Haar measure such that (G) = 1. For x G define x : B [0, 1]
by x (B) := (Rx (B)) for B B. Since Rx commutes with Ly for all y
by (8.11), x is a left Haar measure. Since x (G) = (Rx (G)) = (G) = 1
it follows that x = for all x G. Hence is right invariant. Therefore
the map B [0, 1] : B 7 (B) := ((B)) = (B 1 ) is a left Haar measure
and, since (G) = 1, it agrees with . This proves Theorem 8.12.
In the noncompact case the left and right Haar measures need not agree.
The above argument then shows that the measure x differs from by a
positive factor. Thus there exists a unique map : G (0, ) such that
(Rx (B)) = (x)(B) (8.42)
for all x G and all B B. The map : G (0, ) in (8.42) is a continuous
group homomorphism, called the modular character. It is independent of
the choice of . A locally compact Hausdorff group is called unimodular
iff its modular character is trivial or, equivalently, iff its left and right Haar
measures agree. Thus every compact Hausdorff group is unimodular.
Exercise 8.18. Prove that is a continuous homomorphism.
Exercise 8.19. Prove that the group of all real 2 2-matrices of the form

a b
, a, b R, a > 0,
0 1
is not unimodular. Prove that the additive group Rn is unimodular. Prove
that every discrete group is unimodular.
Exercise 8.20. Let : B [0, ] be a right Haar measure. Show that the
modular character is characterized by the condition (Lx1 (B)) = (x)(B)
for all x G and all B B.
Haar measures are extremely useful tools in geometry, especially when the
group in question is compact. For example, if a compact Hausdorff group G
acts on a topological space X continuously via
G X X : (g, x) 7 g x, (8.43)
one can use the Haar measure to produce G-invariant continuous functions
by averaging. Namely, if f : X R is any continuous function, and is the
Haar measure on G with (G) = 1, then the function F : X R defined by
Z
F (x) := f (a x) d(a) (8.44)
G
for x X is G-invariant in that

F (g x) = F (x)
for all x X and all g G.
Exercise 8.21. Give a precise definition of what it means for a topological

group to act continuously on a topological space.
Exercise 8.22. Show that the map F in (8.44) is continuous and G-invariant.
Exercise 8.23. Let : G GL(V ) be a homomorphism from a com-

pact Hausdorff group to the general linear group of automorphisms of a
finite-dimensional vector space. (Such a homomorphism is called a repre-
sentation of G.) Prove that V admits a G-invariant inner product. This
observation does not extend to noncompact groups. Show that the standard
representation of SL(2, R) on R2 does not admit an invariant inner product.
Exercise 8.24. Show that the Haar measure on a discrete group is a multiple
of the counting measure. Deduce that for a finite group the formula (8.44)
defines F (x) as the average (with multiplicities) of the values of f over the
group orbit of x.
Exercise 8.25. Let G be a locally compact Hausdorff group and let be

a left Haar measure on G. Define the convolution product on L1 (). Show
that L1 () is a Banach algebra. (See page 236.) Find conditions under which
f g = g f . Show that the convolution is not commutative in general. Hint:
See Section 7.5 for G = Rn . See also Step 3 in the proof of Theorem 8.14.
Appendix A
Urysohns Lemma
Theorem A.1 (Urysohns Lemma). Let X be a locally compact Hausdorff

space and let K X be a compact set and U X be an open set such that
K U.
Then there exists a compactly supported continuous function
f : X [0, 1]
such that

f |K 1, supp(f ) = x X f (x) 6= 0 U. (A.1)

Lemma A.2. Let X be a topological space and let K X be compact. Then
(i) Every closed subset of K is compact.
(ii) If X is Hausdorff then, for every y X \ K, there exist open sets
U, V X such that K U , y V , and U V = .
(iii) If X is Hausdorff then K is closed.
Proof. We prove (i). Let F K be closed and let {Ui }iI be an open cover
of F . Then the sets {Ui }iI together with V := X \ F form an open cover
of K. Hence there exist finitely many indices i1 , . . . , in such that the sets
Ui1 , . . . , Uin , V cover K. Hence F Ui1 Uin . This shows that every
open cover of F has a finite subcover and so F is compact.
279
280 APPENDIX A. URYSOHNS LEMMA
We prove (ii). Assume X is Hausdorff and let y X \ K. Define
U := U X | U is open and y

/U .
Since X is Hausdorff the collection U is an open cover of K. Since K is

compact, there exists finitely many set U1 , . . . , Un U such that
K U := U1 Un .
Since y / U i for all i it follows that y V := X \ U and U V = . Hence

the sets U and V satisfy the requirements of (ii).
We prove (iii). Assume X is Hausdorff. Then it follows from part (ii)
that, for every y X \ K, there exists an open set V X such that y V
and V K = . Hence X \ K is the union of all open sets in X that
are disjoint from K. Thus X \ K is open and so K is closed. This proves
Lemma A.2.
Lemma A.3. Let X be a locally compact Hausdorff space and let K, U be

subsets of X such that K is compact, U is open, and K U . Then there
exists an open set V X such that V is compact and
K V V U. (A.2)
Proof. We first prove the assertion in the case where K = {x} consist of
a single element. Choose a compact neighborhood B X of x. Then
F := B \ U is a closed subset of B and hence is compact by part (i) of
Lemma A.2. Since x / F it follows from part (ii) of Lemma A.2 that there
exist open sets W, W 0 X such that x W , F W 0 and W W 0 = .
Hence V := int(B) W is an open neighborhood of x, its closure is a closed
subset of B and hence compact, and
V B W B \ W 0 B \ F U.
This proves the lemma in the case #K = 1.

Now consider the general case. By the first part of the proof the open
sets whose closures are compact and contained in U form an open cover of K.
Since K is compact there exist finitely any open
Sn sets V1 , . . . , Vn such
Snthat V i
is a compact subset of U for all i and K i=1 S Vi . Hence V := i=1 Vi is
an open set containing K and its closure V = ni=1 V i is a compact subset
of U . This proves Lemma A.3.
281
Proof of Theorem A.1. The proof has three steps. The first step requires the
Axiom of Dependent Choice.
Step 1. There exists a family of open sets Vr X with compact closure,
parametrized by r Q [0, 1], such that
K V1 V 1 V0 V 0 U (A.3)
and
s>r = V s Vr (A.4)
for all r, s Q [0, 1].
The existence of open sets V0 and V1 with compact closure satisfying (A.3)
follows from Lemma A.3. Now choose a bijective map N0 Q [0, 1] : i 7 qi
such that q0 = 0 and q1 = 1. Suppose by induction that the open sets
Vi = Vqi have been constructed for i = 0, 1, . . . , n such that (A.4) holds for
r, s {q0 , q1 , . . . , qn }. Choose k, ` {0, 1, . . . , n} such that
qk := max {qi | 0 i n, qi < qn+1 } ,
q` := min {qi | 0 i n, qi > qn+1 } .
Then V ` Vk . Hence it follows from Lemma A.3 that there exists an open set
Vn+1 = Vqn+1 X with compact closure such that V ` Vn+1 V n+1 Vk .
This completes the induction argument and Step 1 then follows from the
Axiom of Dependent Choice. (Denote by V the set of all open sets V X
such that K V V U . Denote by V the set of all finite sequences
v = (V0 , . . . , Vn ) in V that satisfy (A.3) and qi < qj = V j Vi for all i, j.
Define a relation on V by v = (V1 , . . . , Vn ) v0 = (V10 , . . . , Vn00 ) iff n < n0 and
Vi = Vi0 for i = 0, . . . , n. Then V is nonempty and for every v V there is
a v0 V such that v v0 . Hence, by the Axiom of Dependent Choice, there
exists a sequence vj = (Vj,0 , . . . , Vj,nj ) V such that vj vj+1 for all j N.
Define the map Q [0, 1] V : q 7 Vq by Vqi := Vj,i for i, j N with nj i.
This map is well and satisfies (A.3) and (A.4) by definition of V and .)
Step 2. Let Vr X be as in Step 1 for r Q [0, 1]. Then
f (x) := sup {r Q [0, 1] | x Vr }
(A.5)
= inf s Q [0, 1] | x / Vs
for all x X. (Here the supremum of the empty set is zero and the infimum
over the empty set is one.)
We prove equality in (A.5). Fix a point x X and define

a := sup {r Q [0, 1] | x Vr } ,

b := inf s Q [0, 1] | x / Vs .
We prove that a b. If b = 1 this follows directly from the definitions.
Hence assume b < 1 and choose an element s Q [0, 1] such that
x
/ V s.
If r Q [0, 1] such that x Vr then Vr \ V s 6= , hence V s Vr , and hence
r s. Thus we have proved that
x Vr = rs
for all r Q [0, 1]. Take the supremum over all r Q [0, 1] with x Vr
to obtain a s. Then take the infimum over all s Q [0, 1] with x / Vs
to obtain a b. Now suppose, by contradiction, that a < b. Choose rational
numbers r, s Q [0, 1] such that a < r < s < b. Since a < r it follows that
x / Vr , since s < b it follows that x V s , and since r < s it follows from
Step 1 that V s Vr . This is a contradiction and shows that our assumption
that a < b must have been wrong. Thus a = b and this proves Step 2.
Step 3. The function f : X [0, 1] in Step 2 is continuous and

0, for x X \ V0 ,
f (x) = (A.6)
1, for x V 1
Thus f satisfies the requirements of Theorem A.1.
That f satisfies (A.6) follows directly from the definition of f in (A.5). We
prove that f is continuous. To see this fix a constant c R. Then f (x) < c
if and only if there exists a rational number s Q [0, 1] such that s < c
and x / Vs . Likewise, f (x) > c if and only if there exists a rational number
r Q [0, 1] such that r > c and x Vr . Hence
[ [
f 1 ((c, )) = Vr , f 1 ((, c)) = (X \ V s ).
rQ(c,1] sQ[0,c)
This implies that the pre-image under f of every open interval in R is an

open subset of X. Hence also the pre-image under f of every union of
open intervals is open in X and so f is continuous. This proves Step 3 and
Theorem A.1.
283
Theorem A.4 (Partition of Unity). Let X be a locally compact Hausdorff

space, let U1 , . . . , Un X be open sets, and let K U1 Un be a compact
set. Then there exist continuous functions f1 , . . . , fn : X R with compact
support such that
Xn
fi 0, fi 1, supp(fi ) Ui
i=1
Pn
for all i and i=1 fi (x) = 1 for all x K.
Proof. Define the set

V is open, V is compact, and there exists
V := V X .
an index i {1, . . . , n} such that V Ui
If x K then x Ui for some index i {1, . . . , n} and, by Lemma A.3,
there is an open set V X such that V is compact and x V V Ui .
Thus V is an open cover of K. Since K is compact there exist finitely many
open sets V1 , . . . , V` V such that K V1 V` . For i = 1, . . . , n define
[
Ki := V j.
1j`, V j Ui
Then K K1 Kn and Ki is a compact subset of Ui for each i. Hence it

follows from Urysohns Lemma A.1 that, for each i, there exists a compactly
supported continuous function gi : X R such that
0 gi 1, supp(gi ) Ui , gi |Ki 1.
Define
f1 := g1 ,
f2 := (1 g1 )g2 ,
f3 := (1 g1 )(1 g2 )g3 ,
..
.
fn := (1 g1 ) (1 gn1 )gn .
Then supp(fi ) supp(gi ) Ui for each i and
Xn Yn
1 fi = (1 gi ).
i=1 i=1
Since K K1 Kn and the factor 1 gi vanishes on Ki , this implies

P n
i=1 fi (x) = 1 for all x K. This proves Theorem A.4.
Appendix B
The Product Topology
Let (X, UX ) and (Y, UY ) be topological spaces, denote the product space by

X Y := (x, y) x X, y Y ,
and let X : X Y X and Y : X Y Y be the projections onto
the first and second factor. Consider the following universality property for
a topology U 2XY on the product space.
(P) Let (Z, UZ ) be any topological space and let h : Z X Y be any map.
Then h : (Z, UZ ) (X Y, U) is continuous if and only if the maps
f := X h : (Z, UZ ) (X, UX ),
(B.1)
g := Y h : (Z, UZ ) (Y, UY )
are continuous.
Theorem B.1. (i) There is a unique topology U on X Y that satisfies (P).
(ii) Let U 2XY be as in (i). Then W U if and only if there
S are open sets
Ui UX and Vi UY , indexed by any set I, such that W = iI (Ui Vi ).
(iii) Let U 2XY be as in (i). Then U is the smallest topology on X Y
with respect to which the projections X and Y are continuous.
(iv) Let U 2XY be as in (i). Then the inclusion
x : (Y, UY ) (X Y, U), x (y) := (x, y) for y Y,
is continuous for every x X and the inclusion
y : (X, UX ) (X Y, U), y (x) := (x, y) for x X,
is continuous for every y Y .
285
286 APPENDIX B. THE PRODUCT TOPOLOGY
Definition B.2. The product topology on X Y is defined as the unique

topology that satisfies (P) or, equivalently, as the smallest topology on X Y
such that the projections X and Y are continuous. It is denoted by
UXY 2XY .
Proof of Theorem B.1. The proof has five steps.
Step 1. If U 2XY is a topology satisfying (P) then the projections X
and Y are continuous.
Take h := id : X Y X Y so that f = X h = X and g = Y h = Y .
Step 2. We prove uniqueness in (i).
Let U, U 0 2XY be two topologies satisfying (P) and consider the map
h := id : (X Y, U) (X Y, U 0 ). Since f = X : (X Y, U) (X, UX )
and g = Y : (X Y, U) (Y, UY ) are continuous by Step 1, and U 0
satisfies (P), it follows that h is continuous and hence U 0 U. Interchange
the roles of U and U 0 to obtain U 0 = U.
Step 3. We prove (ii) and existence in (i).
Define U 2XY as the collection of all sets of the form W = iI (Ui Vi ),
S
where I is any index set and Ui UX , Vi UY for i I. Then U is a topology
and the projections X : (X Y, U) (X, UX ) and Y : (X Y, U) (Y, UY )
are continuous. We prove that U satisfies (P). To see this, let (Z, UZ ) be any
topological space, let h : Z X Y be any map, and define f := X h and
g := Y h as in (B.1). If h is continuous then so are f and g. Conversely,
if f and g are continuous, then h1 (U V ) = f 1 (U ) g 1 (V ) is open in Z
for all U UX and all V UY , and hence it follows from the definition of U
that h1 (W ) is open for all W U. Thus h is continuous.
Step 4. We prove (iii).
Let U be the topology in (i) and let U 0 be any topology on X Y with
respect to which X and Y are continuous. If U UX and V UY then
1
U V = X (U ) Y1 (V ) U 0 . Hence U U 0 by (ii). Since X and Y
are continuous with respect to U it follows that U is the smallest topology
on X Y with respect to which X and Y are continuous.
Step 5. We prove (iv).
Fix an element x X and consider the map h := x : Y X Y . Then
the map f := X h : Y X is constant and g := Y h : Y Y is
the identity. Hence f and g are continuous and so is h by condition (P). An
analogous argument shows that y is continuous for all y Y .
Appendix C
The Inverse Function Theorem
This appendix contains a proof of the inverse function theorem. The result
is formulated in the setting of continuously differentiable maps between open
sets in a Banach space. Readers who are unfamiliar with bounded linear
operators on Banach spaces may simply think of continuously differentiable
maps between open sets in finite-dimensional normed vector spaces. The in-
verse function theorem is used on page 71 in the proof of Lemma 2.19, which
is a key step in the proof of the transformation formula for the Lebesgue mea-
sure (Theorem 2.17). Assume throughout that (X, kk) is a Banach space.
When : X X is a bounded linear operator denote its operator norm by
kxk
kk := kkL(X) := sup .
xX\{0} kxk
For x X and r > 0 denote by Br (x) := {y X | kx yk < r} the open

ball of radius r about x. For x = 0 abbreviate Br := Br (0).
Theorem C.1 (Inverse Function Theorem). Fix an element x0 X

and two real numbers r > 0 and 0 < < 1. Let : Br (x0 ) X be a
continuously differentiable map such that
kd(x) 1lkL(X) for all x Br (x0 ). (C.1)
Then
B(1)r ((x0 )) (Br (x0 )) B(1+)r ((x0 )). (C.2)
Moreover, the map is injective, its image is open, the map 1 is contin-
uously differentiable, and d 1 (y) = d( 1 (y))1 for all y (Br (x0 )).
287
288 APPENDIX C. THE INVERSE FUNCTION THEOREM
Proof. Assume without loss of generality that x0 = (x0 ) = 0.

Step 1. is a homeomorphism onto its image and (Br ) B(1+)r .
Define := id : Br X. Then kd(x)k for all x Br . Hence
k(x) (y)k kx yk. (C.3)
for all x, y Br and so
(1 ) kx yk k(x) (y)k (1 + ) kx yk . (C.4)
The second inequality in (C.4) shows that (Br ) B(1+)r and the first
inequality in (C.4) shows that is injective and 1 is Lipschitz continuous.
Step 2. B(1)r (Br ).
Let B(1)r and define > 0 by kk =: (1 )(r ). Then, by (C.3)
with y = 0, we have k(x)k kxk for all x Br . If kxk r this implies
k(x) + k r . Thus the map x 7 (x) + is a contraction of the closed
ball B r . By the contraction mapping principle it has a unique fixed point
x and the fixed point satisfies (x) = x (x) = . Hence (Br ).
Step 3. (Br ) is open.
Let x Br and define y := (x). Choose > 0 such that B (x) Br . Then,
by Step 2, B(1) ((x)) (B (x)) (Br ).
Step 4. 1 is continuously differentiable.
Let x0 Br and define Py0 := (xk0 ) and := d(x0 ). Then k1lk , so
is invertible, = k=0 (1l) , and k k (1)1 . We prove that 1
1 1
is differentiable at y0 and d 1 (y0 ) = 1 . Let > 0. Since is differentiable

at x0 and d(x0 ) = , there is a constant > 0 such that, for all x Br with
kxx0 k < (1)1 , we have k(x)(x0 )(xx0 )k (1)2 kxx0 k.
Shrinking , if necessary, we may assume, by Step 3, that B (y0 ) (Br ).
Now suppose ky y0 k < and denote x := 1 (y) Br . Then, by (C.4),
kx x0 k (1 )1 ky y0 k < (1 )1 and hence
1
(y) 1 (y0 ) 1 (y y0 ) = 1 y y0 (x x0 )

1
k(x) (x0 ) (x x0 )k
1
(1 ) kx x0 k
ky y0 k .
Hence 1 is differentiable at y0 and d 1 (y0 ) = 1 = d( 1 (y0 ))1 . Thus
d 1 is continuous by Step 1. This proves Theorem C.1.
Bibliography
[1] R.B. Ash, Basic Probability Theory, John Wiley & Sons 1970, Dover Publications, 2008.
http://www.math.uiuc.edu/~r-ash/BPT/BPT.pdf
[2] V.I. Bogachev, Measure Theory, Volumes 1 & 2, Springer Verlag, 2006.
[3] W.F. Eberlein, Notes on integration I: The underlying convergence theorem. Comm. Pure Appl.
Math. 10 (1957), 357360.
[4] D.H. Fremlin, Measure Theory, Volumes 15, 2000.
https://www.essex.ac.uk/maths/people/fremlin/mt.htm
[5] D. Gilbarg and N.S. Trudinger, Elliptic Partial Differential Equations of the Second Order, Springer
Verlag, 1983.
[6] P.R. Halmos, Measure Theory, Springer-Verlag, 1974.
[7] C.E. Heil, Lecture Notes on Real Analysis, Georgia Institute of Technology, Fall 2007.
http://people.math.gatech.edu/~heil/6337/fall07/section3.4.pdf
http://people.math.gatech.edu/~heil/6337/fall07/section3.5b.pdf
[8] H. Konig, New versions of the RadonNikod ym theorem. Arch. Math. 86 (2006), 251260.
http://www.math.uni-sb.de/service/preprints/preprint132.pdf
[9] T.W. K orner, A Companion to Analysis. Graduate Studies in Mathematics 62, AMS 2004. https:
//www.dpmms.cam.ac.uk/~twk/
[10] U. Lang, Mass & Integral, Vorlesungsmanuskript, ETH Z urich, September 2006.
http://www.math.ethz.ch/~lang/mi.pdf
[11] U. Lang, Existence and uniqueness of Haar integrals. ETH Z urich, January 2015.
http://www.math.ethz.ch/~lang/HaarIntegral.pdf
[12] Dan Ma, The Lexicographic Order and The Double Arrow Space. 7 Oct. 2009.
http://dantopology.wordpress.com/2009/10/07/
[13] P. Malliavin, Integration and Probability (in cooperation with H el`
ene Airault, Leslie Kay, G
erard
Letac), Springer-Verlag, 1995.
[14] J.R. Munkres, Topology, Second Edition. Prentice Hall, 2000.
[15] G. Nagy, Real Analysis IV.6, Duals of Lp spaces.
http://www.math.ksu.edu/~nagy/real-an/4-06-dual-lp.pdf
[16] G.K. Pedersen, The existence and uniqueness of the Haar integral on a locally compact topological
group. Preprint, University of Copenhagen, November 2000.
http://www.math.ku.dk/kurser/2004-2/mat3re/haarintegral.pdf
[17] W. Rudin, Real and Complex Analysis, 3rd Ed. McGrawHill, 1987.
[18] D.A. Salamon, Das Riemannsche Integral. Lecture Notes, ETHZ, 2014.
http://www.math.ethz.ch/~salamon/PREPRINTS/ana1-int.pdf
[19] D.A. Salamon, Mehrfache Integrale. Lecture notes, ETHZ, 2014.
http://www.math.ethz.ch/~salamon/PREPRINTS/ana2-int.pdf
[20] D.A. Salamon, Funktionentheorie. Birkhauser, 2012.
http://www.math.ethz.ch/~salamon/PREPRINTS/cxana.pdf
[21] W.F. Trench, Introduction to Real Analysis, Prentice Hall, 2003.
http://ramanujan.math.trinity.edu/wtrench/misc/index.shtml
289
Index
absolutely continuous CalderonZygmund inequality, 246

function, 201, 207 Caratheodory Criterion, 53
measure, 151 CauchySchwarz inequality, 125
signed measure, 169, 173 characteristic function, 12
almost everywhere, 33 closed set, 6
compact set, 10
Baire measurable
compactly supported function, 97
function, 110
complete measure space, 39
set, 110
complete metric space, 7
Baire measure, 110
completion of a measure space, 39
Baire -algebra, 110
Banach algebra, 236 continuous map, 10
Banach space, 7 continuum hypothesis, 124
BanachZarecki Theorem, 201 convergence in measure, 147
basis of a topological space, 106 convolution, 232
bilinear form of signed measures, 257
positive definite, 125 counting measure, 19
symmetric, 125 cuboid, 56
Boolean algebra, 6
Borel measurable dense subset, 120
function, 12 Dieudonnes measure, 83
set, 9 Dirac measure, 19
Borel measure, 82 Divergence Theorem, 255
extended, 91 double arrow space, 107
left invariant, 264 dual space
right invariant, 264 of a Hilbert space, 127
Borel outer measure, 92 of a normed vector space, 126
Borel set, 9 of C(X), 106, 183
Borel -algebra, 9 of L2 (), 129
bounded linear functional, 126 of Lp (), 135
bounded variation, 205 of ` , 136, 149
290
INDEX 291
Egoroffs Thorem, 147 inner regular, 82, 176

elementary set, 210 on open sets, 86
envelope, 130 integrable function
essential supremum, 117 Lebesgue, 29
locally, 236
first countable, 106 partially defined, 42
Fubinis Theorem Riemann, 75
for integrable functions, 222 weakly, 188
for positive functions, 219 integral
for the completion, 224 Haar, 265
for the Lebesgue measure, 231 Lebesgue, 20, 29
in polar coordinates, 255 Riemann, 75
fundamental solution RiemannStieltjes, 205
of Laplaces equation, 244 invariant
Fundamental Theorem linear functional, 265
of Calculus, 201 measure, 264
Greens formula, 245 inverse limit, 261
group
Jensens inequality, 143
discrete, 260
Jordan decomposition, 168, 172
Lie, 260
Jordan measurable set, 76
topological, 259
Jordan measure, 76
unimodular, 277
Jordan null set, 56
Haar integral, 265
Haar measure, 264 Laplace operator, 243
Hahn decomposition, 171 Lebesgue
HahnBanach Theorem, 136 Differentiation Theorem, 196, 200
Hardys inequality, 146 Lebesgue decomposition, 152
HardyLittlewood for signed measures, 170
maximal inequality, 192 Lebesgue Dominated
Hausdorff dimension, 80 Convergence Theorem, 32
Hausdorff measure, 80 Lebesgue integrable, 29
Hausdorff space, 10, 81 Lebesgue integral, 20, 29
Hilbert space, 126 Lebesgue measurable
Holder inequality, 114 function, 49, 60
set, 49, 60
inner product, 125 Lebesgue measure, 49, 60
on L2 (), 126 on the sphere, 255
292 INDEX
Lebesgue Monotone Jordan, 76

Convergence Theorem, 23 Lebesgue, 49, 60
Lebesgue null set, 56 w.r.t. an outer measure, 50
Lebesgue outer measure, 56 measurable space, 5
continuous from below, 61 measure, 17
regularity, 61 absolutely continuous, 151
Lebesgue point, 196 Baire, 110
left invariant Borel, 82
linear functional, 265 Borel outer, 92
measure, 264 counting, 19
lexicographic ordering, 107 Dirac, 19
Lie group, 260 Haar, 264
linear functional Hausdorff, 80
left/right invariant, 265 inner regular, 82, 86
localizable, 130, 148 Jordan, 76
strictly, 226 Lebesgue, 49, 60
locally compact, 81 Lebesgue outer, 56
Hausdorff group, 259 left invariant, 264
locally determined, 226 localizable, 130, 148
locally integrable, 236 locally determined, 226
lower semi-continuous, 206 nonatomic, 124
lower sum, 75 outer, 50
Lp (), 115 outer regular, 82
Lp (Rn ), 117 probability, 124
L (), 117 product, 214
`p , 117 Radon, 82
` , 136 regular, 82
right invariant, 264
Marcinkiewicz interpolation, 241 semi-finite, 130
maximal function, 192, 195 -finite, 130
measurable function, 11 signed, 166
Baire, 110 singular, 151
Borel, 12 strictly localizable, 226
Lebesgue, 49, 60 translation invariant, 49
partially defined, 42 truly continuous, 174
measurable set, 5 measure space, 17
Baire, 110 complete, 39
Borel, 9 Lebesgue, 49, 60
INDEX 293
localizable, 130, 148 positive linear functional

locally determined, 226 on Cc (X), 97
semi-finite, 130 on Lp (), 137
-finite, 130 pre-image, 11
strictly localizable, 226 probability theory, 4648, 124
metric space, 7 product measure, 214
Minkowski inequality, 114, 220 complete locally determined, 226
modular character, 277 primitive, 226
mollifier, 239 product -algebra, 209
monotone class, 210 product topology, 262, 286
mutually singular, 151 pushforward
signed measures, 169 of a measure, 46
of a -algebra, 12, 46
neighborhood, 81
nonatomic measure, 124 Radon measure, 82
norm of a bounded RadonNikod ym derivative, 199
linear functional, 126 RadonNikod ym Theorem, 152
normed vector space, 7 for signed measures, 170
null set, 33 generalized, 176
Jordan, 56 regular measure, 82
Lebesgue, 56 Riemann integrable, 75
one-point compactification, 108 Riemann integral, 75
open ball, 7 RiemannStieltjes integral, 205
open set, 6 Riesz Representation Theorem, 98
in a metric space, 7 right invariant
outer measure, 50 linear functional, 265
Borel, 92 measure, 264
Lebesgue, 56
second countable, 106
translation invariant, 57
outer regular, 82 semi-finite, 130
separability of Lp (), 121
p-adic integers, 261 separable, 120
p-adic rationals, 262 set of measure zero, 33
partition of a set, 5 -additive
partition of unity, 283 measure, 17
perfectly normal, 107 signed measure, 166
metric spaces are, 109 -algebra, 5
Poisson identity, 244 Baire, 110
294 INDEX
Borel, 9 second countable, 106

Lebesgue, 49, 60 separable, 120
product, 209 -compact, 81
-compact, 81 standard on R, 8
-finite, 130 standard on R, 8
signed measure, 166 total variation
absolutely continuous, 169, 173 of a signed measure, 166
concentrated, 169 transformation formula, 67
Hahn decomposition, 171 translation invariant
inner regular measure, 49
w.r.t. a measure, 176 outer measure, 57
Jordan decomposition, 168, 172 triangle inequality, 7, 125
Lebesgue decomposition, 170 weak, 186
mutually singular, 169 truly continuous, 174
RadonNikod ym Theorem, 170 Tychonoffs Theorem, 262
total variation, 166
truly continuous, 174 uniformly integrable, 180
simple function, 16 upper semi-continuous, 206
upper sum, 75
singular measure, 151
Urysohns Lemma, 279
smooth function, 243
Sorgenfrey line, 107, 109 Vitalis Covering Lemma, 194
step function, 16 Vitalis Theorem, 180

StoneCech compactification VitaliCaratheodory Theorem, 206
of N, 150
weak triangle inequality, 186
topological group, 259 weakly integrable, 188
invariant measure, 264
left invariant measure, 264 Youngs inequality, 113, 235
locally compact Hausdorff, 259
right invariant measure, 264
topology, 6
basis, 106
first countable, 106
group, 259
Hausdorff, 10, 81
locally compact, 81
perfectly normal, 107
product, 262, 286

Measure

Uploaded by

Copyright:

Available Formats

Measure

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Measure

Uploaded by

Copyright:

Available Formats

MEASURE AND INTEGRATION

1 August 2015 Dietmar A. Salamon

1 Abstract Measure Theory 3

2 The Lebesgue Measure 49

4.4 Hilbert Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . 125

5 The RadonNikod ym Theorem 151

7 Product Measures 209

8 The Haar Measure 259

A Urysohns Lemma 279

B The Product Topology 285

C The Inverse Function Theorem 287

Abstract Measure Theory

The purpose of this first chapter is to introduce integration on abstract mea-

Even with the restricted concept of countable additivity it will not be

(f ) Every finite or countable intersection of elements of A is an element

all -algebras B 2X that contain E and define

Thus A is a -algebra by Exercise 1.7. Moreover, it follows directly from the

B (x) := B (x, d) := {y X | d(x, y) < }

(centered at x with radius ) is contained in U . The collection of d-open

defines a distance function on X. X is called a Banach space if the metric

The topology on R induced by this distance function is called the standard

(a, ] := (a, ) {}, [, b) := (, b) {}.

Call a subset U R open if it is a countable union of open intervals in R

(iii) Prove that the map f : R [1, 1] defined by

Exercise 1.14. Extend the total ordering of R to R by a

Recollections About Point Set Topology

1.2 Measurable Functions

of B under f is a well defined subset of X, whether or not the map f is

and preserves union, intersection, and complement. Thus

for every subset B Y and

for every collection of subsets Bi Y , indexed by a set I.

(ii) Let (X, AX ) be a measurable space. A function f : X R is called

(iii) Let (X, UX ) and (Y, UY ) be topological spaces. A map f : X Y is

Theorem 1.20 (Measurable and Continuous Maps). Let (X, AX ) and

(ii) Assume UX 2X is a topology on X such that AX is the Borel -algebra

Proof. By part (iv) of Theorem 1.19 a map f : X Y is measurable if

Theorem 1.21 (Characterization of Measurable Functions).

Then B is a -algebra on R by part (iii) of Theorem 1.19 and (a, ] B for

Lemma 1.23. Let (X, A) be a measurable space and let u, v : X R

Proof. We prove (i). The functions : R2 R defined by (s, t) := s + t,

is measurable. Hence it follows from Theorem 1.21 that g is measurable.

are also measurable. This proves Theorem 1.24.

(See Example 1.5.) Moreover,

where Ai : X R is the characteristic function of the set Ai for i = 1, . . . , `

Proof. If f can be approximated by a sequence of measurable step func-

Hence the functions sn := n f satisfy the requirements of the theorem.

1.3 Integration of Nonnegative Functions

(b) There exists a measurable set A A such that (A) < .

The basic properties of measures are summarized in the next theorem.

(v) Let Ai A be a sequence such that Ai Ai+1 for all i. Then

and hence () = 0. This proves part (i).

Since Bi Bj = for i 6= j it follows from -additivity that

We prove (v). Assume Ai Ai+1 for all i and define Ci := Ai \ Ai+1 .

Since Ci Cj = for i 6= j it follows from -additivity that

This proves part (v) and Theorem 1.28.

With these preparations in place we are now ready to introduce the

with i [0, ) and Ai R A for i = 1, . . . , n. The (Lebesgue) integral of

(ii) Let f : X [0, ] be a measurable

Proof. To prove (i), assume f g on E. If s R: X [0,R) is a measurable