Measure and Integration PDF

Textbooks in Mathematics
Dietmar A. Salamon
Dietmar A. Salamon Dietmar A. Salamon
Measure and Integration
Measure

The book is intended as a companion to a one semester introductory lecture course
on measure and integration. After an introduction to abstract measure theory it
proceeds to the construction of the Lebesgue measure and of Borel measures on
and
Integration
locally compact Hausdorff spaces, Lp spaces and their dual spaces and elementary
Hilbert space theory. Special features include the formulation of the Riesz
Representation Theorem in terms of both inner and outer regularity, the proofs of
the Marcinkiewicz Interpolation Theorem and the Calderon–Zygmund inequality as
applications of Fubini’s theorem and Lebesgue differentiation, the treatment of the
generalized Radon–Nikodym theorem due to Fremlin, and the existence proof for Haar
measures. Three appendices deal with Urysohn’s Lemma, product topologies, and the
inverse function theorem.
The book assumes familiarity with first year analysis and linear algebra. It is suitable
for second year undergraduate students of mathematics or anyone desiring an
introduction to the concepts of measure and integration.
ISBN 978-3-03719-159-0
www.ems-ph.org
Salamon Cover | Font: Frutiger_Helvetica Neue | Farben: Pantone 116, Pantone 287 | RB 31 mm
EMS Textbooks in Mathematics
EMS Textbooks in Mathematics is a series of books aimed at students or professional mathemati-

cians seeking an introduction into a particular field. The individual volumes are intended not only to
provide relevant techniques, results, and applications, but also to afford insight into the motivations
and ideas behind the theory. Suitably designed exercises help to master the subject and prepare
the reader for the study of more advanced and specialized literature.
Jørn Justesen and Tom Høholdt, A Course In Error-Correcting Codes

Markus Stroppel, Locally Compact Groups
Peter Kunkel and Volker Mehrmann, Differential-Algebraic Equations
Dorothee D. Haroske and Hans Triebel, Distributions, Sobolev Spaces, Elliptic Equations
Thomas Timmermann, An Invitation to Quantum Groups and Duality
Oleg Bogopolski, Introduction to Group Theory
Marek Jarnicki and Peter Pflug, First Steps in Several Complex Variables: Reinhardt Domains
Tammo tom Dieck, Algebraic Topology
Mauro C. Beltrametti et al., Lectures on Curves, Surfaces and Projective Varieties
Wolfgang Woess, Denumerable Markov Chains
Eduard Zehnder, Lectures on Dynamical Systems. Hamiltonian Vector Fields and Symplectic
Capacities
Andrzej Skowroński and Kunio Yamagata, Frobenius Algebras I. Basic Representation Theory
Piotr W. Nowak and Guoliang Yu, Large Scale Geometry
Joaquim Bruna and Juliá Cufí, Complex Analysis
Eduardo Casas-Alvero, Analytic Projective Geometry
Fabrice Baudoin, Diffusion Processes and Stochastic Calculus
Olivier Lablée, Spectral Theory in Riemannian Geometry
Dietmar A. Salamon
Measure
and
Integration
Author:
Dietmar A. Salamon
Departement Mathematik
ETH Zürich
Rämistrasse 101
8092 Zürich
Switzerland
E-mail: dietmar.salamon@math.ethz.ch
2010 Mathematics Subject Classification: Primary: 28-01; Secondary: 28A05, 28A10, 28A12,
28A15, 28A20, 28A25, 28A30, 28A33, 28A35, 28C05, 28C10, 28C15, 35J05, 43A05, 44A35,
46B22, 46C05, 46E27, 46E30
Key words: sigma-Algebra, Lebesgue monotone convergence, Caratheodory criterion, Lebesgue

measure, Borel measure, Dieudonné’s measure, Riesz Representation Theorem, Alexandrov Double
Arrow Space, Sorgenfrey Line, separability, Cauchy–Schwarz inequality, Jensen’s inequality, Egoroff’s
theorem, Hardy’s inequality, absolutely continuous measure, truly continuous measure, singular
measure, signed measure, Radon–Nikodym Theorem, Lebesgue Decomposition Theorem, Hahn
Decomposition Theorem, Jordan Decomposition Theorem, Hardy–Littlewood maximal inequality,
Vitali’s Covering Lemma, Lebesgue point, Lebesgue Differentiation Theorem, Banach-Zarecki
Theorem, Vitali–Caratheodory Theorem, Cantor function, product sigma-algebra, Fubini’s Theorem,
convolution, Young’s inequality, mollifier, Marcinciewicz interpolation, Poisson identity, Green’s
formula, Calderon–Zygmund inequality, Haar measure, modular character.
ISBN 978-3-03719-159-0
The Swiss National Library lists this publication in The Swiss Book, the Swiss national bibliography,
and the detailed bibliographic data are available on the Internet at http://www.helveticat.ch.
This work is subject to copyright. All rights are reserved, whether the whole or part of the material is
concerned, specifically the rights of translation, reprinting, re-use of illustrations, recitation, broad-
casting, reproduction on microfilms or in other ways, and storage in data banks. For any kind of use
permission of the copyright owner must be obtained.
© 2016 European Mathematical Society
Contact address:
European Mathematical Society Publishing House

Seminar for Applied Mathematics
ETH-Zentrum SEW A27
CH-8092 Zürich
Switzerland
Phone: +41 (0)44 632 34 36

Email: info@ems-ph.org
Homepage: www.ems-ph.org
Typeset using the author’s TEX files: M. Zunino, Stuttgart, Germany

Printing and binding: Beltz Bad Langensalza GmbH, Bad Langensalza, Germany
∞ Printed on acid free paper
987654321
Preface
This book is based on notes for the lecture course Measure and Integration held
at ETH Zürich in the spring semester 2014. Prerequisites are the first-year courses
on analysis and linear algebra, including the Riemann integral [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 Rudin’s book, the lecture notes by
Urs Lang [10, 11], the five volumes on measure theory by David H. Fremlin [4],
the paper by Heinz König [8] on the generalized Radon–Nikodým theorem,
the lecture notes by C. E. Heil [7] on absolutely continuous functions, Dan Ma’s
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.
The text 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 generalized Radon–Nikodým
theorem, Sections 7.6 and 7.7 on Marcinkiewicz interpolation and the Calderón–
Zygmund 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 en-
lightening discussions and for pointing out the book by Fremlin, Dan Ma’s topology
blog, and the paper by Pedersen. Thanks to Urs Lang for his insightful comments
on the construction of the Haar measure.
ETH, Zürich Dietmar A. Salamon

1 August 2015
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1 Abstract measure theory . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1.1 -algebras . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Measurable functions . . . . . . . . . . . . . . . . . . . . . . . . 11
1.3 Integration of nonnegative functions . . . . . . . . . . . . . . . . . 18
1.4 Integration of real valued functions . . . . . . . . . . . . . . . . . 32
1.5 Sets of measure zero . . . . . . . . . . . . . . . . . . . . . . . . . 36
1.6 Completion of a measure space . . . . . . . . . . . . . . . . . . . 43
1.7 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47
2 The Lebesgue measure . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

2.1 Outer measures . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
2.2 The Lebesgue outer measure . . . . . . . . . . . . . . . . . . . . . 64
2.3 The transformation formula . . . . . . . . . . . . . . . . . . . . . 77
2.4 Lebesgue equals Riemann . . . . . . . . . . . . . . . . . . . . . . 86
2.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90
3 Borel measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95
3.1 Regular Borel measures . . . . . . . . . . . . . . . . . . . . . . . 95
3.2 Borel outer measures . . . . . . . . . . . . . . . . . . . . . . . . 107
3.3 The Riesz representation theorem . . . . . . . . . . . . . . . . . . 113
3.4 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 126
4 Lp spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 133
4.1 Hölder and Minkowski . . . . . . . . . . . . . . . . . . . . . . . . 133
4.2 The Banach space Lp ./ . . . . . . . . . . . . . . . . . . . . . . . 136
4.3 Separability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 141
4.4 Hilbert spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147
4.5 The dual space of Lp ./ . . . . . . . . . . . . . . . . . . . . . . . 152
4.6 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 168
viii Contents
5 The Radon–Nikodým theorem . . . . . . . . . . . . . . . . . . . . . . . 179

5.1 Absolutely continuous measures . . . . . . . . . . . . . . . . . . . 179
5.2 The dual space of Lp ./ revisited . . . . . . . . . . . . . . . . . . 187
5.3 Signed measures . . . . . . . . . . . . . . . . . . . . . . . . . . . 197
5.4 Radon–Nikodým generalized . . . . . . . . . . . . . . . . . . . . 205
5.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 214
6 Differentiation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 221
6.1 Weakly integrable functions . . . . . . . . . . . . . . . . . . . . . 221
6.2 Maximal functions . . . . . . . . . . . . . . . . . . . . . . . . . . 227
6.3 Lebesgue points . . . . . . . . . . . . . . . . . . . . . . . . . . . 233
6.4 Absolutely continuous functions . . . . . . . . . . . . . . . . . . . 239
6.5 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 243
7 Product measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

7.1 The product -algebra . . . . . . . . . . . . . . . . . . . . . . . . 249
7.2 The product measure . . . . . . . . . . . . . . . . . . . . . . . . . 255
7.3 Fubini’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . 261
7.4 Fubini and Lebesgue . . . . . . . . . . . . . . . . . . . . . . . . . 272
7.5 Convolution . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 277
7.6 Marcinkiewicz interpolation . . . . . . . . . . . . . . . . . . . . . 286
7.7 The Calderón–Zygmund inequality . . . . . . . . . . . . . . . . . 292
7.8 Exercises . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 304
8 The Haar measure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309

8.1 Topological groups . . . . . . . . . . . . . . . . . . . . . . . . . . 309
8.2 Haar measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312
Appendices
A Urysohn’s lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333
B The product topology . . . . . . . . . . . . . . . . . . . . . . . . . . . 339
C The inverse function theorem . . . . . . . . . . . . . . . . . . . . . . . 343
Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 347
Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 351
Introduction
We learn already in high school that integration plays a central role in mathemat-
ics 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 func-
tions 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 functions 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
drawbacks 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 Fatou lemma,
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 < 1. Chapter 5 is devoted to the proof of the Radon–Nikodým theorem
about absolutely continuous measures and to the proof that Lq is naturally isomor-
phic to the dual space of Lp when 1=p C1=q D 1 and 1 < p < 1. Chapter 6 deals
with differentiation. Chapter 7 introduces product measures and contains a proof of
Fubini’s Theorem, an introduction to the convolution product on L1 .Rn /, and a
proof of the Calderón–Zygmund inequality. Chapter 8 constructs Haar measures on
locally compact Hausdorff groups.
Despite the overlap with the book of Rudin [17], there are some differences in
exposition and content. A small expository difference is that in Chapter 1 measur-
able 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 repre-
2 Introduction
sentation 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 regular-
ity together with inner regularity on open sets. This leads to a somewhat different
formulation of the Riesz representation theorem (which includes the result as for-
mulated 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) when-
ever 1 < p < 1 and 1=p C 1=q D 1. It is also shown that L1 ./ 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 Radon–Nikodým theorem for signed measures,
due to Fremlin [4], which does not require that the underlying measure is -finite.
In the formulation of König [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 ad-
dition, the present book includes a self-contained proof of the Calderón–Zygmund
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 theory 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 Malli-
avin [13] which covers many additional topics including Fourier analysis, limit theo-
rems in probability theory, Sobolev spaces, and the stochastic calculus of variations.
Many other fields of mathematics require the basic notions of measure and integra-
tion. They include functional analysis and partial differential equations (see, e.g.,
Gilbarg–Trudinger [5]), geometric measure theory, geometric group theory, ergodic
theory and dynamical 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 measure

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 under 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 subsets of a given domain. For-
mally 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 additivity 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 observation 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.
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.
4 1. Abstract measure theory
One way out of this dilemma is to give up on the idea of countable additivity and
replace it by the weaker requirement of countable subadditivity. This leads to the
notion of an outer measure which will be discussed in Chapter 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 subtlety 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 addi-
tivity of the integral and the proof is based on the Lebesgue monotone convergence
theorem. An important inequality is the Fatou lemma. 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 dis-
cusses 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
For any fixed set X denote by 2X the set of all subsets of X and, for any subset
A X , denote by
Ac WD X n A
its complement.
1.1. -algebras 5
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 2 A;
(b) if A 2 A, then Ac 2 A;
(c) every countable union of elements of A is again an element of A, i.e., if Ai 2 A
for i D 1; 2; 3; : : : , then 1iD1 Ai 2 A.
S
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) ; 2 A;
Sn
(e) if n 2 N and A1 ; : : : ; An 2 A, then iD1 Ai 2 A;
(f) every finite or countable intersection of elements of A is an element of A;
(g) if A; B 2 A, then A n B 2 A.
Proof. Condition (d) follows from (a), (b) because X c D ;, and (e) follows from (c),
(d) by taking A WD ; for i > n. Condition (f) follows from (b), (c), (e) because
T c Si c
A
i i D i Ai , and (g) follows from (b), (f) because A n B D A \ B . This
c
proves Lemma 1.2.

Example 1.3. The sets A WD ¹;; Xº and A WD 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 ¹Ai ºi2I be a partition of X, i.e., Ai is a
S
nonempty subset of X for each i 2 I , Ai \ Aj D ; for i ¤ j , and X D i 2I Ai .
Then [
A WD ¹AJ WD Aj j J I º
j 2J
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 n B; B n A; A \ B; X n .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 D k for k D 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 i 2 I a -algebra Ai 2X is given. Prove that the intersection
A WD i 2I Ai D ¹A X j A 2 Ai for all i 2 I º is a -algebra.
T
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
X
A D B. To prove existence, denote by S 22 the collection of all -algebras
B 2X that contain E and define
ˇ if B 2X is a -algebra
\ ² ˇ ³
A WD BD AX ˇ :
such that E B; then A 2 B
ˇ
B2S
Thus A is a -algebra by Exercise 1.7. Moreover, it follows directly from the defi-
nition 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 2 U;
Tn
(b) if n 2 N and U1 ; : : : ; Un 2 U, then Ui 2 U;
iD1
(c) if I is any index set and Ui 2 U for i 2 I , then i 2I Ui 2 U.

S
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 2 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 per-
mits 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 courses. Here is a recollection of the definition.
Definition 1.10 (metric space). A metric space is a pair .X; d / consisting of a set
X and a function
dW X X ! R
satisfying the following axioms:
(a) d.x; y/ 0 for all x; y 2 X, with equality if and only if x D y;
(b) d.x; y/ D d.y; x/ for all x; y 2 X ;
(c) d.x; z/ d.x; y/ C d.y; z/ for all x; y; z 2 X .
A function d W 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 2 U there exists a constant
" > 0 such that the open ball
B" .x/ WD B" .x; d / WD ¹y 2 X j d.x; y/ < "º
(centered at x and of radius ") is contained in U . The collection of d -open subsets
of X will be denoted by
U.X; d / WD ¹U X j U is d -openº:
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 Defini-
tion 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 ! RW x 7 ! kxk
satisfying the following:
(a) kxk 0 for all x 2 X , with equality if and only if x D 0;
(b) kxk D jj kxk for all x 2 X and 2 R;
(c) kx C yk kxk C kyk for all x; y 2 X.
Let .X; kk/ be a normed vector space. Then the formula
d.x; y/ WD 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 D R of real numbers is a metric space with the standard
distance function
d.x; y/ WD jx yj:
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
x WD Œ 1; 1 WD R [ ¹ 1; 1º:
R
For a; b 2 R define
.a; 1 WD .a; 1/ [ ¹1º; Œ 1; b/ WD . 1; b/ [ ¹ 1º:
Call a subset U Rx open if it is a countable union of open intervals in R and sets

of the form .a; 1 or Œ 1; b/ for a; b 2 R.
x satisfies the axioms of a topology. This

(i) Show that the set of open subsets of R
x
is called the standard topology on R.
x is induced by the distance function
(ii) Prove that the standard topology on R
x R
dW R x !R
defined by the following formulas for x; y 2 R:
2je x y e y x j
d.x; y/ WD ;
e xCy C ex y C ey x C e x y
2e x
d.x; 1/ WD d.1; x/ WD ;
ex C e x
2e x
d.x; 1/ WD d. 1; x/ WD ;
ex C e x
d. 1; 1/ WD d.1; 1/ WD 2:
1.1. -algebras 9
(iii) Prove that the map

x ! Œ 1; 1
f WR
defined by
ex e x
f .x/ WD tanh.x/ WD ; x 2 R;
ex C e x
f .˙1/ WD ˙1;
is a homeomorphism. Prove that it is an isometry with respect to the metric
x and the standard metric on the interval Œ 1; 1. Deduce that .R
in (ii) on R x; d/
is a compact metric space.
x by
Exercise 1.14. Extend the total ordering of R to R x.
1 a 1 for a 2 R
Extend addition by
1 C a WD 1 for 1 < a 1,
and by
1 C a WD 1 for 1 a < 1.
(The sum a C b is undefined when ¹a; bº D ¹ 1; 1º.) Let a1 ; a2 ; a3 ; : : : and

x.
b1 ; b2 ; b3 ; : : : be sequences in R
(i) Define lim sup an and lim inf an and show that they always exist.
n!1 n!1
(ii) Show that

lim sup. an / D lim inf an :
n!1 n!1
(iii) Assume ¹an ; bn º ¤ ¹ 1; 1º so the sum an C bn is defined for n 2 N. Prove

the inequality
lim sup.an C bn / lim sup an C lim sup bn ;

n!1 n!1 n!1
whenever the right-hand side exists. Find an example where the inequality is
strict.
(iv) If an bn for all n 2 N, show that
lim inf an lim inf bn :

n!1 n!1
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 1
S
i D1 Fi of closed subsets Fi X is a Borel set.
(These are sometimes called F -sets.)
(iii) Every countable intersection 1
T
iD1 Ui of open subsets Ui X is a Borel set.
(These are sometimes called Gı -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 Gı -set.
Consider for example the Borel -algebra on the real axis R with its standard
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 speaking, 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 through-
out 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 is a
collection of open sets ¹Ui ºi2I , indexed by a set I , such that
[
K Ui :
i 2I
The set K is called compact if every open cover of K has a finite subcover, i.e.,
if for every open cover ¹Ui ºi2I of K there exist finitely many indices i1 ; : : : ; in 2 I
such that
K Ui1 [ [ Uin :
1.2. Measurable functions 11
When .X; d / is a metric space and U D 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 D ¹;; Xº, 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 2 X there exist open sets U; V 2 U such that x 2 U ,
y 2 V , and U \ V D ;), then every compact subset of X is closed (Lemma A.2).
Next, recall that a map f W X ! Y between two metric spaces .X; dX / and
.Y; dY / is continuous (i.e., for every x 2 X and every " > 0 there is a ı > 0 such
that f .Bı .x; dX // B" .f .x/; dY /) if and only if the pre-image
1
f .V / WD ¹x 2 X j f .x/ 2 V º
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 W X ! Y between topological spaces
.X; UX / and .Y; UY / is called continuous if
V 2 UY H) f 1
.V / 2 UX :
It follows directly from the definition that topological spaces form a category, in that
the composition g ı f W X ! Z of two continuous maps f W X ! Y and gW Y ! Z
between topological spaces is again continuous. Another basic observation is that
if f W 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

In analogy to continuous maps between topological spaces, one can define mea-
surable 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 above discussion of continuity. For any map f W X ! Y
between two sets X and Y and any subset B Y , the pre-image
1
f .B/ WD ¹x 2 X j f .x/ 2 Bº
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 W Y ! X . The pre-image defines a
map from 2Y to 2X . It satisfies
1 1
f .Y / D X; f .;/ D ;; (1.1)
and preserves union, intersection, and complement. Thus

1 1
f .Y n B/ D X n f .B/ (1.2)
for every subset B Y , and

[ [ \ \
f 1 Bi D f 1
.Bi /; f 1
Bi D f 1
.Bi / (1.3)
i2I i 2I i 2I i2I
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 W X ! Y is called measurable if the pre-image of every measur-
able subset of Y under f is a measurable subset of X, i.e.,
B 2 AY H) f 1
.B/ 2 AX :
(ii) Let .X; AX / be a measurable space. A function f W X ! R x is called mea-

x associated to
surable if it is measurable with respect to the Borel -algebra on R
the standard topology in Exercise 1.13 (see Definition 1.15).
(iii) Let .X; UX / and .Y; UY / be topological spaces. A map f W 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 W X ! R
defined by
´
1 if x 2 A;
A .x/ WD (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). Theorem 1.20
below shows that every continuous function between topological spaces is Borel
measurable. It also shows that a function from a measurable 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 measur-
able spaces.
(i) The identity map idX W X ! X is measurable.
(ii) If f W X ! Y and gW Y ! Z are measurable maps, then so is the composition
g ı f W X ! Z.
(iii) Let f W X ! Y be any map. Then the set
f AX WD ¹B Y j f 1
.B/ 2 AX º (1.5)
is a -algebra on Y , called the pushforward of AX under f .

(iv) A map f W X ! Y is measurable if and only if AY f AX .
Proof. Parts (i) and (ii) follow directly from the definitions. That f AX 2Y
defined by (1.5) is a -algebra follows from (1.1) (for axiom (a)), (1.2) (for ax-
iom (b)), and (1.3) (for axiom (c)). This proves part (iii). Moreover by Defini-
tion 1.17, f is measurable if and only if f 1 .B/ 2 AX for every B 2 AY ,
and this means that AY f AX . This establishes (iv) and completes the proof
of 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 W X ! Y is measurable if an only if the pre-image of every open

subset V Y under f is measurable, i.e.,
V 2 UY H) f 1
.V / 2 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 W X ! Y is (Borel) measurable.
Proof. By part (iv) of Theorem 1.19, a map f W 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 2 UY H) f 1
.V / 2 AX :
This proves part (i). If, in addition, AX is the Borel -algebra of a topology UX
on X and f W .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 measur-

x be any function. Then the following are equivalent:
able space and let f W X ! R
(i) f is measurable;
1
(ii) f ..a; 1/ is a measurable subset of X for every a 2 R;
1
(iii) f .Œa; 1/ is a measurable subset of X for every a 2 R;
1
(iv) f .Œ 1; b// is a measurable subset of X for every b 2 R;
1
(v) f .Œ 1; b/ is a measurable subset of X for every b 2 R.
Proof. That (i) implies (ii), (iii), (iv), and (v) follows directly from the definitions.
We prove that (ii) implies (i). Thus let f W X ! R x be such that f 1 ..a; 1/ 2 AX
for every a 2 R and define
x jf x
B WD f AX D ¹B R 1
.B/ 2 AX º 2R :
x by part (iii) of Theorem 1.19 and .a; 1 2 B for every

Then B is a -algebra on R
a 2 R by assumption. Therefore Œ 1; b D R x n .b; 1 2 B for every b 2 R by
axiom (b) and hence
[
Œ 1; b/ D Œ 1; b n1 2 B
n2N
by axiom (c) in Definition 1.1. Hence it follows from (f) in Lemma 1.2 that
.a; b/ D Œ 1; b/ \ .a; 1 2 B
for every pair of real numbers a < b. Since every open subset of R x is a count-
able union of sets of the form .a; b/, .a; 1, Œ 1; b/, it follows from axiom (c) in
Definition 1.1 that every open subset of Rx 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 func-
tions 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 D .f1 ; : : : ; fn /W X ! Rn be a function. Then f is measurable if
and only if fi W X ! R is measurable for each i .
Proof. For i D 1; : : : ; n define the projection i W Rn ! R by i .x/ WD xi for

x D .x1 ; : : : ; xn / 2 R. Since i is continuous, it follows from Theorems 1.19
and 1.20 that if f is measurable, so is fi D i ı f for all i. Conversely, suppose
that fi is measurable for i D 1; : : : ; n. Let ai < bi for i D 1; : : : ; n and define
Q.a; b/ WD ¹x 2 Rn j ai < xi < bi , for all i º D .a1 ; b1 / .an ; bn /:
Then
n
\
1 1
f .Q.a; b// D fi ..ai ; bi // 2 A
iD1
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 / 2 A for every open set U Rn , and so f
is measurable. This proves Theorem 1.22.
Lemma 1.23. Let .X; A/ be a measurable space and let u; vW X ! R be measur-

able functions. If W R2 ! R is continuous, then the function hW X ! R defined
by
h.x/ WD .u.x/; v.x// for x 2 X
is measurable.
Proof. The map f WD .u; v/W X ! R2 is measurable (with respect to the Borel
-algebra on R2 ) by Theorem 1.22 and the map W R2 ! R is Borel measurable
by Theorem 1.20. Hence the composition h D ı f W 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; gW X ! R are measurable functions, then so are the functions
f C g; fg; max¹f; gº; min¹f; gº; jf j:
(ii) Let fk W X ! R x , k D 1; 2; 3; : : : , be a sequence of measurable functions.

Then the following functions from X to R x are measurable:
inf fk ; sup fk ; lim sup fk ; lim inf fk :

k k k!1 k!1
Proof. (i) The functions W R2 ! R defined by .s; t/ WD s C t, .s; t/ WD st ,

.s; t/ WD max¹s; tº, .s; t/ WD min¹s; tº, or .s; t/ WD jsj are all continuous.
Hence assertion (i) follows from Lemma 1.23.
(ii) Define
x
g WD sup fk W X ! R
k
and let a 2 R. Then the set
1
g ..a; 1/ D ¹x 2 X j sup fk .x/ > aº
k
D ¹x 2 X j there exists k 2 N such that fk .x/ > aº
[
D ¹x 2 X j fk .x/ > aº
k2N
[
D fk 1 ..a; 1/
k2N
is measurable. Hence it follows from Theorem 1.21 that g is measurable. It also fol-
lows from part (i) (already proved) that fk is measurable, hence so is supk . fk /
by what we have just proved, and then so is the function inf k fk D supk . fk /.
With this understood, we conclude that the functions
lim sup fk D inf sup fk ; lim inf fk D sup inf fk

k!1 `2N k` k!1 `2N k`
are also measurable. This proves Theorem 1.24.

It follows from Theorem 1.24 that the pointwise limit of a sequence of measur-
able functions, if it exists, is again measurable. This is in sharp contrast 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 sW 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 sW X ! R be a step function, write
s.X/ D ¹˛1 ; : : : ; ˛` º with ˛i ¤ ˛j for i ¤ j ,
and define
1
Ai WD s .˛i / D ¹x 2 X j s.x/ D ˛i º
for i D 1; : : : ; `. Then the sets A1 ; : : : ; A` form a partition of X , i.e.,
`
[
XD Ai ; Ai \ Aj D ; for i ¤ j: (1.6)
iD1
(See Example 1.5.) Moreover,
`
X
sD ˛i A ; (1.7)
i
i D1
where A W X ! R is the characteristic function of the set Ai for i D 1; : : : ; `

i
(see (1.4)). In this situation s is measurable if and only if the set Ai X is mea-
surable for each i. For later reference we prove the following.
Theorem 1.26 (approximation). Let .X; A/ be a measurable space and let
f W X ! Œ0; 1
be a function. Then f is measurable if and only if there exists a sequence of mea-

surable step functions sn W X ! Œ0; 1/ such that
0 s1 .x/ s2 .x/ f .x/; f .x/ D lim sn .x/ for all x 2 X:

n!1
Proof. If f can be approximated by a sequence of measurable step functions,

then f is measurable by Theorem 1.24. Conversely, suppose that f is measurable.
For n 2 N define
n W Œ0; 1 ! R
by
´
n n n
k2 if k2 t < .k C 1/2 ; k D 0; 1; : : : ; n2n 1;
n .t/ WD (1.8)
n if t n:
These functions are Borel measurable and satisfy n .0/ D 0 and n .1/ D n for
all n, as well as t 2 n n .t/ nC1 .t/ t whenever n t > 0. Thus
lim n .t/ D t for all t 2 Œ0; 1:

n!1
Hence the functions sn WD n ı f satisfy the requirements of the theorem.
1.3 Integration of nonnegative functions
Our next goal is to define the integral of a measurable step function and then the inte-
gral of a general nonnegative measurable function via approximation. 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 Œ0; 1. Addition to 1 and
multiplication by 1 are defined by
´
1 if a ¤ 0;
a C 1 WD 1 C a WD 1; a 1 WD 1 a WD
0 if a D 0:
With this convention, addition and multiplication are commutative, associative, and
distributive. Moreover, if ai and bi are nondecreasing sequences in Œ0; 1, then the
limits
a WD lim ai and b WD lim bi

i!1 i!1
exists in Œ0; 1 and satisfy the familiar rules

a C b D lim .ai C bi / and ab D lim .ai bi /:
i !1 i!1
1.3. Integration of nonnegative functions 19
These rules must be treated with caution. The product rule does not hold when
the sequences are not nondecreasing. For example ai WD i converges to a D 1,
bi WD 1=i converges to b D 0, but ai bi D 1 does not converge to ab D 0.
(Exercise. Show that the sum of two convergent sequences in Œ0; 1 always con-
verges to the sum of the limits.) Also, for all a; b; c 2 Œ0; 1,
a C b D a C c; a < 1 H) b D c;
ab D ac; 0 < a < 1 H) b D c:
Neither of these assertions extend to the case a D 1.
Definition 1.27 (measure). Let .X; A/ be a measurable space. A measure on .X; A/

is a function
W A ! Œ0; 1
satisfying the following axioms.
(a) is -additive, i.e., if Ai 2 A, i D 1; 2; 3; : : : , is a sequence of pairwise

disjoint measurable sets, then
1
[ X1
Ai D .Ai /:
i D1 i D1
(b) There exists a measurable set A 2 A such that .A/ < 1.
A measure space is a triple .X; A; / consisting of a set X , a -algebra A 2X ,

and a measure W A ! Œ0; 1.
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) .;/ D 0.
(ii) If n 2 N and A1 ; : : : ; An 2 A are such that Ai \ Aj D ; for i ¤ j , then
.A1 [ [ An / D .A1 / C C .An /:
(iii) If A; B 2 A are such that A B, then
.A/ .B/:
(iv) Let Ai 2 A be a sequence such that Ai AiC1 for all i. Then

1
[
Ai D lim .Ai /:
i !1
iD1
(v) Let Ai 2 A be a sequence such that Ai AiC1 for all i. Then

1
\
.A1 / < 1 H) Ai D lim .Ai /:
i !1
i D1
Proof. We prove (i). Choose A1 2 A such that .A1 / < 1 and define Ai WD ; for
i > 1. Then it follows from -additivity that
X
.A1 / D .A1 / C .;/;
i>1
and hence .;/ D 0. This proves part (i).

Part (ii) follows from (i) and -additivity by choosing Ai WD ; for i > n.
We prove (iii). If A; B 2 A are such that A B, then B nA 2 A by property (g)
in Lemma 1.2 and hence .B/ D .A/C.B nA/ .A/ by part (ii). This proves
part (iii).
We prove (iv). Assume Ai AiC1 for all i and set
´
A1 for i D 1,
Bi WD
Ai n Ai 1 for i > 1.
Then Bi is measurable for all i and, for n 2 N,
n
[ 1
[ 1
[
An D Bi ; A WD Ai D Bi :
iD1 iD1 iD1
Since Bi \ Bj D ; for i ¤ j , it follows from -additivity that

1
X n
X
.A/ D .Bi / D lim .Bi / D lim .An /:
n!1 n!1
iD1 i D1
Here the last equality follows from part (ii). This proves part (iv).
We prove (v). Assume Ai Ai C1 for all i and define Ci WD Ai n AiC1 . Then
Ci is measurable for all i and, for n 2 N,
1
[ 1
\
An D A [ Ci ; A WD Ai :
iDn i D1
Since Ci \ Cj D ; for i ¤ j , -additivity yields

1
X
.An / D .A/ C .Ci /
iDn
P1
for all n 2 N. Since .A1 / < 1. It follows that iD1 .Ci / < 1 and hence
1
X
lim .An / D .A/ C lim .Ci / D .A/:
n!1 n!1
iDn
This proves part (v) and Theorem 1.28.
Exercise 1.29. Let .X; A; / be a measure

Pspace and let Ai 2 A be a sequence of
S
measurable sets. Prove that i Ai i .Ai /.
Example 1.30. Let .X; A/ be a measurable space. The counting measure
W A ! Œ0; 1
is defined by
.A/ WD #A for A 2 A.
As an example, consider the counting measure W 2N ! Œ0; 1 on the natural num-
bers. Then the sets An WD ¹n; nC1; º all have infinite measure and their intersec-
tion is the empty set and hence has measure zero. Thus the hypothesis .A1 / < 1
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 2 X.

The Dirac measure at x0 is the measure ıx0 W A ! Œ0; 1 defined by
´
1 if x0 2 A;
ıx0 .A/ WD for A 2 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 W A ! Œ0; 1 defined by .A/ WD 0 when A is countable and by
.A/ WD 1 when Ac is countable is a measure.
Example 1.33. Let X D i2I Ai be a partition and let A 2X be the -algebra

S
in Example 1.5. Then any function I ! Œ0; 1W i 7! ˛i determines a measure

W A ! Œ0; 1 via .AJ / WD j 2J ˛j for J I and AJ D j 2J Aj 2 A.
P S
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 2 A be a measurable set.
(i) Let sW X ! Œ0; 1/ be a measurable step function of the form
n
X
sD ˛i A (1.9)
i
iD1
with ˛i 2 Œ0; 1/ and RAi 2 A for i D 1; : : : ; n. The (Lebesgue) integral of s

over E is the number E s d 2 Œ0; 1 defined by
Z Xn
s d WD ˛i .E \ Ai /: (1.10)
E i D1
(ii) Let f W X ! Œ0; 1 beR a measurable function. The (Lebesgue) integral of f

over E is the number E f d 2 Œ0; 1 defined by
Z Z
f d WD sup s d;
E sf E
where the supremum is taken over all measurable step function sW X ! Œ0; 1/
that satisfy s.x/ f .x/ for all x 2 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 WD ¹A 2 A j A Eº is a -algebra on E
Rand E WD jAE 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/ WD 0 for x 2 X n E.
Theorem 1.35 (basic properties of the Lebesgue integral). Let .X; A; / be a mea-
sure space, f; gW X ! Œ0; 1 be measurable functions, and E 2 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 D X f E d;
R
(iii) if f .x/ D 0 for all x 2 E, then E f d D 0;
R
(iv) if .E/ D 0, then E f d D 0;
(v) if A 2 A and E A, then E f d A f d;
R R
R R
(vi) if c 2 Œ0; 1/, then E cf d D c E f d.
Proof. To prove (i), assume f g on E. If sW X ! Œ0; 1/ is a measurable step

function such that s f , then sE g, so
Z Z Z
s d D sE d g d
E E E
by the definition of the integral of g. Now take the supremum over all measurable
step functions s f to obtain
Z Z
f d g d:
E E
This proves (i).

We prove (ii). It follows from the definitions that
Z Z
f d D sup s d
E sf E
Z
D sup sE d
sf X
Z
D sup t d
t f E X
Z
D f E d:
X
Here the supremum is over all measurable step functions sW X ! Œ0; 1/, respec-
tively tW X ! Œ0; 1/, that satisfy s f , respectively t f E . The second
equality Rfollows fromR the fact that every measurable step function sW X ! Œ0; 1/
satisfies E s d D X sE d by the definition of the integral. The third equa-
tion follows from the fact that a measurable step function t W X ! Œ0; 1/ satisfies
t f E if and only if it has the form t D sE for some measurable step function
sW X ! Œ0; 1/ such that s f .
R
Part (iii) is a consequence of (i) with g D 0 Rand the fact that E f d 0
by definition. Part (iv) follows from the fact that E s d D 0 for every measur-
able step function s when .E/ D 0. Part (v) follows from parts (i) and (ii) and
R fact that fRE f A whenever E A. Part (vi) follows from the fact that
the
E cs d D c E s d for every c 2 Œ0; 1/ and every measurable step function s,
by the commutativity, associativity, and distributivity rules for calculations with
numbers in Œ0; 1. 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 Definition 1.34 indeed satisfies
this crucial condition requires some preparation. First one verifies this property for
integrals of step functions and then one uses the Lebesgue monotone convergence
Theorem 1.37.
Lemma 1.36 (additivity for step functions). Let .X; A; / be a measure space and
let s; tW X ! Œ0; 1/ be measurable step functions.
(i) For every measurable set E 2 A
Z Z Z
.s C t/ d D s d C t d:
E E E
(ii) If E1 ; E2 ; E3 ; : : : is a sequence of pairwise disjoint measurable sets, then

Z X1 Z [
s d D s d; E WD Ek :
E Ek
kD1 k2N
Proof. Write the functions s and t in the form

m
X n
X
sD ˛i A ; tD ˇj B
i j
i D1 j D1
where ˛i ; ˇj 2 Œ0; 1/ and Ai ; Bj 2 A are such that Ai \ Ai 0 D ; for i ¤ i 0 ,

Bj \ Bj 0 D ; for j ¤ j 0 , and X D m
S Sn
iD1 Ai D j D1 Bj . Then
m X
X n
sCt D .˛i C ˇj /A ;
i \Bj
iD1 j D1
and hence
Z m X
X n
.s C t/ d D .˛i C ˇj /.Ai \ Bj \ E/
E iD1 j D1
m
X n
X n
X m
X
D ˛i .Ai \ Bj \ E/ C ˇj .Ai \ Bj \ E/
iD1 j D1 j D1 i D1
m
X n
X
D ˛i .Ai \ E/ C ˇj .Bj \ E/
iD1 j D1
Z Z
D s d C t d:
E E
To prove (ii), let E1 ; E2 ; E3 ; : : : be a sequence of pairwise disjoint measurable sets

and define E WD 1
S
kD1 Ek : Then
Z m
X
s d D ˛i .E \ Ai /
E iD1
m
X 1
X
D ˛i .Ek \ Ai /
iD1 kD1
m
X n
X
D ˛i lim .Ek \ Ai /
n!1
iD1 kD1
m
X n
X
D lim ˛i .Ek \ Ai /
n!1
iD1 kD1
n X
X m
D lim ˛i .Ek \ Ai /
n!1
kD1 iD1
n Z
X
D lim s d
n!1 Ek
kD1
1 Z
X
D s d:
Ek
kD1
Theorem 1.37 (Lebesgue monotone convergence theorem). Let .X; A; / be a

measure space and let
fn W X ! Œ0; 1
be a sequence of measurable functions such that
fn .x/ fnC1 .x/ for all x 2 X and all n 2 N:
Define f W X ! Œ0; 1 by
f .x/ WD lim fn .x/ for x 2 X:

n!1
Then f is measurable and

Z Z
lim fn d D f d:
n!1 X X
Proof. By part (i) of Theorem 1.35, we have

Z Z
fn d fnC1 d
X X
for all n 2 N and hence the limit

Z
˛ WD lim fn d (1.11)
n!1 X
exists in Œ0; 1. Moreover, f D supn fn is a measurable function on X , by part (ii)

of Theorem 1.24, and satisfies fn f for all n 2 N. Thus it follows from part (i)
of Theorem 1.35 that
Z Z
fn d f d for all n 2 N;
X X
and hence Z
˛ f d: (1.12)
X
Now fix a measurable step function sW X ! Œ0; 1/ such that s f . Define
s W A ! Œ0; 1
by Z
s .E/ WD s d for E 2 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 .;/ D 0).
Now fix a constant 0 < c < 1 and define
En WD ¹x 2 X j cs.x/ fn .x/º for n 2 N:
Then En 2 A is a measurable set and En EnC1 for all n 2 N. Moreover,

1
[
En D X: (1.14)
nD1
(To spell it out, choose an element x 2 X . If f .x/ D 1, then fn .x/ ! 1 and

hence cs.x/ s.x/ fn .x/ for some n 2 N, which means that x belongs to
one of the sets En . If f .x/ < 1, then fn .x/ converges to f .x/ > cf .x/, hence
fn .x/ > cf .x/ cs.x/ for some n 2 N, and for this n we have x 2 En .) Since
cs fn on En , it follows from parts (i) and (vi) of Theorem 1.35 that
Z Z Z Z
cs .En / D c s d 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 2 N: (1.15)
c
Since s W A ! Œ0; 1 is a measure, by part (i) of Theorem 1.35, it follows
from (1.14) and part (iv) of Theorem 1.28 that
˛
Z
s d D s .X/ D lim s .En / : (1.16)
X n!1 c
Here the last inequality
R follows from (1.15). Since (1.16) holds for every constant
0 < c < 1, we have X s d ˛ for every measurable step function sW X ! Œ0; 1/
such that s f . Take the supremum over all such s to obtain
Z Z
f d D sup s d ˛:
X sf X
R
Combining this with (1.12) we obtain X f d 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; gW X ! Œ0; 1 are measurable and E 2 A, then

Z Z Z
.f C g/ d D f d C g d: (1.17)
E E E
(ii) Let fn W X ! Œ0; 1 be a sequence of measurable functions and define

1
X
f .x/ WD fn .x/ for x 2 X:
nD1
Then f W X ! Œ0; 1 is measurable and, for every E 2 A,

Z 1 Z
X
f d D fn d: (1.18)
E nD1 E
(iii) If f W X ! Œ0; 1 is measurable and E1 ; E2 ; E3 ; : : : is a sequence of pairwise

disjoint measurable sets, then
Z 1 Z
X [
f d D f d; E WD Ek : (1.19)
E Ek
kD1 k2N
Proof. (i) By Theorem 1.26, there exist sequences of measurable step functions
sn ; tn W X ! Œ0; 1/ such that
sn snC1 and tn tnC1
for all n 2 N and
f .x/ D lim sn .x/ and g.x/ D lim tn .x/
n!1 n!1
for all x 2 X . Then sn C tn is a monotonically nondecreasing sequence of measur-

able step functions converging pointwise to f C g. Hence,
Z Z
.f C g/ d D lim .sn C tn / d
X n!1 X
Z Z
D lim sn d C tn d
n!1 X X
Z Z
D lim sn d C lim tn d
n!1 X n!1 X
Z Z
D f d C g d:
X X
Here the first and last equalities follow from Theorem 1.37, while the second equal-
ity follows from part (i) of Lemma 1.36. This proves (i) for E D X. To prove it in
general, replace f; g by f E ; gE and use part (ii) of Theorem 1.35.
(ii) Define gn W X ! Œ0; 1 by
n
X
gn WD fk :
kD1
This is a nondecreasing sequence of measurable functions, by part (i) of Theo-

rem 1.24, and it converges pointwise to f by definition. Hence it follows from
part (ii) of Theorem 1.24 that f is measurable and how the Lebesgue monotone
convergence theorem (Theorem 1.37) shows that
Z Z
f d D lim gn d
X n!1 X
n
Z X
D lim fk d
n!1 X
kD1
n Z
X
D lim fk d
n!1 X
kD1
1 Z
X
D fn d:
nD1 X
Here the second equality follows from the definition of gn while the third one fol-
lows from part (i) of the present theorem (already proved). This proves (ii) for
E D X. To prove it in general, replace f; fn by f E ; fn E and use part (ii) of
Theorem 1.35.
(iii) Let f W X ! Œ0; 1 be a measurable function and let Ek 2 A be a sequence
of pairwise disjoint measurable sets. Define
1
[ n
X
E WD Ek ; fn WD f E :
k
kD1 kD1
Then it follows from part (i) of the present theorem (already proved) and part (ii) of
Theorem 1.35 that
Z n
Z X
fn d D f E d
k
X X
kD1
n Z
X
D f E d
k
X
kD1
n Z
X
D f d:
Ek
kD1
Now fn W X ! Œ0; 1 is a nondecreasing sequence of measurable functions con-

verging pointwise to f E . Hence it follows from the Lebesgue monotone conver-
gence theorem (Theorem 1.37) that
Z Z
f d D f E d
E X
Z
D lim fn d
n!1 X
n Z
X
D lim f d
n!1 Ek
kD1
1 Z
X
D f d:
Ek
kD1
This proves Theorem 1.38.

Exercise 1.39. Let W 2N ! Œ0; 1 be the counting measure on the natural num-
bers. Show that in this case equation (1.18) in part (ii) of Theorem 1.38 is equivalent
to the formula
X1 X 1 X 1 X 1
aij D aij (1.20)
iD1 j D1 j D1 iD1
for every map N N ! Œ0; 1W .i; j / 7! aij .
The next theorem shows that every measurable function f W X ! Œ0; 1 induces
another measure f on .X; A/.
Theorem 1.40. Let .X; A; / be a measure space and let f W X ! Œ0; 1 be a

measurable function. Then the function
f W A ! Œ0; 1;
defined by Z
f .E/ WD f d for E 2 A (1.21)
E
is a measure, and Z Z
g df D fg d (1.22)
E E
for every measurable function gW X ! Œ0; 1 and every E 2 A.
Proof. f is -additive by part (iii) of Theorem 1.38 and f .;/ D 0 by part (iv)
of Theorem 1.35. Hence f is a measure (see Definition 1.27). Now let g WD A
be the characteristic function of a measurable set A 2 A. Then
Z Z Z
A df D f .A/ D f d D f A d:
X A X
Here the first equality follows from the definition of the integral for measurable step
functions in Definition 1.34, the second equality from the definition of f , and the
last equality from part (ii) of Theorem 1.35. Thus equation (1.22) (with E D 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 D X) continues to hold for all measurable step functions g D sW X ! Œ0; 1/.
Now approximate an arbitrary measurable function gW X ! Œ0; 1 by a sequence
of measurable step functions via Theorem 1.26 and use the Lebesgue monotone
convergence theorem (Theorem 1.37) to deduce that (1.22) holds with E D X for
all measurable functions gW X ! Œ0; 1. 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 W A ! Œ0; 1 can be expressed in the form f for some measurable
function f W X ! Œ0; 1. 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 (Theorem 1.45).
Theorem 1.41 (Fatou’s lemma). Let .X; A; / be a measure space. Consider a
sequence of measurable functions fn W X ! Œ0; 1. Then
Z Z
lim inf fn d lim inf fn d:
X n!1 n!1 X
Proof. For n 2 N define gn W X ! Œ0; 1 by
gn .x/ WD inf fi .x/

i n
for x 2 X. Then gn is measurable, by Theorem 1.24, and
g1 .x/ g2 .x/ g3 .x/ ; lim gn .x/ D lim inf fn .x/ DW f .x/

n!1 n!1
for all x 2 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 i n X
for all n 2 N. Thus, by the Lebesgue monotone convergence theorem (Theo-

rem 1.37),
Z Z Z Z
f d D lim gn d lim inf fi d D lim inf fn d:
X n!1 X n!1 in X n!1 X

Example 1.42. Let .X; A; / be a measure space and E 2 A be a measurable set
such that 0 < .E/ < .X/. Define fn WD E when n is even and fn WD 1 E
when n is odd. Then lim inf fn D 0 and so
n!1
Z Z
lim inf fn d D 0 < min¹.E/; .X n E/º D lim inf fn d:
X n!1 n!1 X
Thus the inequality in Theorem 1.41 can be strict.

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 naturally to the concepts
of integrability and the Lebesgue integral. The basic properties of the Lebesgue
integral are summarized in Theorem 1.44 below. The main result of this section is
the Lebesgue dominated convergence theorem (Theorem 1.45).
Definition 1.43 (Lebesgue integrable functions). Let .X; A; / be a measure space.
A function fR W X ! R is called (Lebesgue) integrable or -integrable if f is mea-
surable and X jf j d < 1: Denote the set of -integrable functions by
L1 ./ WD L1 .X; A; / WD ¹f W X ! R j f is -integrableº:
The Lebesgue integral of f 2 L1 ./ over a set E 2 A is the real number

Z Z Z
C
f d WD f d f d; (1.23)
E E E
where the functions f ˙ W X ! Œ0; 1/ are defined by
f C .x/ WD max¹f .x/; 0º; f .x/ WD max¹ f .x/; 0º (1.24)
The functions f ˙ are measurable by Theorem 1.24 and 0 f ˙ jf j. 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 a real

R vector space and, for every E 21 A, the function
L ./ ! RW f 7! E f d is linear, i.e., if f; g 2 L ./ and c 2 R,
1
then f C g; cf 2 L1 ./, and

Z Z Z Z Z
.f C g/ d D f d C g d; cf d D c f d: (1.25)
E E E E E
(ii) For all f; g 2 L1 ./ and all E 2 A,

Z Z
f g on E H) f d g d: (1.26)
E E
(iii) If f 2 L1 ./, then jf j 2 L1 ./ and, for all E 2 A,

ˇZ ˇ Z
ˇ ˇ
ˇ f dˇ jf j d: (1.27)
ˇ ˇ
E E
1.4. Integration of real valued functions 33
(iv) If f 2 L1 ./ and E1 ; E2 ; E3 ; : : : is a sequence of pairwise disjoint measur-

able sets, then
Z X1 Z [
f d D f d; E WD Ek : (1.28)
E Ek
kD1 k2N
(v) For all E 2 A and all f 2 L1 ./,

Z Z
f d D f E d: (1.29)
E X
(vi) Let E 2 A and f 2 L1 ./. If .E/ D 0 or f jE D 0, then

Z
f d D 0:
E
Proof. (i) Let f; g 2 L1 ./ and R let c 2 R. Then f C g 2 L ./ because

1
jf C gj jf j C jgj and hence X jf C gj d < 1 by part R(i) of Theorem 1.38.

Likewise, cf 2 L1 ./ because jcf j D jcjjf j and hence X jcf j d < 1 by
part (vi) of Theorem 1.35. To prove the second equation in (1.25) assume first that
c 0. Then .cf /˙ D cf ˙ and hence
Z Z Z
cf d D cf C d cf d
E E E
Z Z
C
Dc f d c f d
E E
Z
Dc f d:
E
Here the second equation follows from part (vi) of Theorem 1.35. If c < 0,
then .cf /C D . c/f and .cf / D . c/f C and hence, again using part (iv)
of Theorem 1.35, we obtain
Z Z Z
cf d D . c/f d . c/f C d
E E E
Z Z
D . c/ f d . c/ f C d
E E
Z
Dc f d:
E
Now let h WD f C g. Then hC h DfC f C gC g and hence

hC C f C g D h C f C C gC :
Hence it follows from part (i) of Theorem 1.38 that

Z Z Z Z Z Z
C C
h d C f d C g d D h d C f d C g C d:
E E E E E E
Therefore,
Z Z Z
C
h d D h d h d
E E E
Z Z Z Z
C C
D f d C g d f d g d
E E E E
Z Z
D f d C g d
E E
and this proves (i).

(ii) Assume fR D f C f g D C C
R g C g on E. Then f C g g C f
C
C
on E and hence E .f C g / d E .g C 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
C C
f d C g d g d C f d:
E E E E
This implies (1.26).

(iii) Since jf j f jf j it follows from (1.25) and (1.26) that
Z Z Z Z
jf j d D . jf j/ d f d jf j d
E E E E
and this implies (1.27).

(iv) Equation (1.28) holds for f ˙ by part (iii) of Theorem 1.38 and hence holds
for f by Definition 1.43.
R R
(v) The formula E f d D X f E d in (1.29) follows from part (ii) of
Theorem 1.35 since f ˙ E D .f E /˙ .
(vi) If f vanishes on E, then f ˙ also vanish on E ˙
R
R and˙ hence E f d D 0
by part (iii) of Theorem 1.35. If .E/ D 0, then E f d D 0 by part (iv)
of Theorem 1.35. In R either case it follows from the definition of the integral in
Definition 1.43 that E f d D 0. This proves Theorem 1.44.
1.4. Integration of real valued functions 35
Theorem 1.45 (Lebesgue dominated convergence theorem). Let .X; A; / be a

measure space, gW X ! Œ0; 1/ be an integrable function, let fn W X ! R be a
sequence of integrable functions satisfying
jfn .x/j g.x/ for all x 2 X and n 2 N; (1.30)
and converging pointwise to f W X ! R, i.e.,
f .x/ D lim fn .x/ for all x 2 X: (1.31)

n!1
Then f is integrable and, for every E 2 A,

Z Z
f d D lim fn d: (1.32)
E n!1 E
Proof. f is measurable by part (ii) of Theorem 1.24 and jf .x/j g.x/ for all
x 2 X by (1.30) and (1.31). Hence it follows from part (i) of Theorem 1.35 that
Z Z
jf j d g d < 1;
X X
and so f is integrable. Moreover
jfn f j jfn j C jf j 2g:
Therefore, by Fatou’s lemma (Theorem 1.41), that

Z Z
2g d D lim inf .2g jfn f j/ d
X X n!1
Z
lim inf .2g jfn f j/ d
n!1 X
Z Z
D lim inf 2g d jfn f j d
n!1 X X
Z Z
D 2g d lim sup jfn f j d:
X n!1 X
Here the penultimate step follows from part (i) of Theorem 1.44. This implies
Z
lim sup jfn f j d 0:
n!1 X
Hence
Z
lim jfn f j d D 0:
n!1 X
Since
ˇZ Z ˇ Z Z
ˇ ˇ
ˇ
ˇ fn d f dˇˇ jfn f j d jfn f j d;
E E E X
part (iii) of Theorem 1.44 yields
ˇZ Z ˇ
ˇ ˇ
lim ˇˇ fn d f dˇˇ D 0;
n!1 E E
which is equivalent to (1.32). This completes the proof of 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 2 A such that .N / D 0. Let P stand
for some property that a point x 2 X may have, or not have, depending on x.
For example, if f W X ! Œ0; 1 is a measurable function on X, then P could stand
for the condition f .x/ > 0, or for the condition f .x/ D 0, or for the condition
f .x/ D 1. Or if fn W X ! R is a sequence of measurable functions, 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 2 X n N has the property P. It is not required that the set of all
elements x 2 X that have the property P is measurable, although that may often be
the case.
Example 1.46. Let fn W X ! R be any sequence of measurable functions. Then
the set
E WD ¹x 2 X j .fn .x//1
nD1 is a Cauchy sequenceº
\ [ \
k
D ¹x 2 X j jfn .x/ fm .x/j < 2 º
k2N n0 2N n;mn0
is measurable. If N WD X n E is a set of measure zero, then fn converges almost

everywhere to a function f W X ! R. This function can be chosen measurable by
defining
f .x/ WD lim fn .x/ for x 2 E
n!1
and
f .x/ WD 0 for x 2 N .
This is the pointwise limit of the sequence of measurable functions gn WD fn E
and hence is measurable by part (ii) of Theorem 1.24.
The first observation is that every nonnegative function with finite integral is
almost everywhere finite.
1.5. Sets of measure zero 37
R
Lemma 1.47. Let f W X ! Œ0; 1 be a measurable function. If X f d < 1,
then f < 1 almost everywhere.
R N WD ¹xR 2 X j f .x/ D 1º and h WD 1N : Then h f and so

Proof. Define
1.N / D X h d X f d < 1 by Theorem 1.35. Hence .N / D 0.
The second observation is that if two integrable, or nonnegative measurable,

functions agree almost everywhere, then their integrals agree over every measurable
set.
Lemma 1.48. Assume either that f; gW X ! Œ0; 1 are measurable functions that
agree almost everywhere, or that f; gW X ! R are -integrable functions that agree
almost everywhere. Then
Z Z
f d D g d for all A 2 A: (1.33)
A A
Proof. Fix a measurable set A 2 A and define N WD ¹x 2 X j f .x/ ¤ g.x/º.

Then N is measurable and .N / D 0 by assumption. Hence .A \ N / D 0 by
part (iii) of Theorem 1.28. This implies
Z Z Z
f d D f d C f d
A AnN A\N
Z
D f d
AnN
Z
D f AnN d:
X
Here the first equality follows from part (iii) of Theorem 1.38 in the nonnegative
case and from part (iv) of Theorem 1.44 in the integrable case. The second equality
follows from part (iv) of Theorem 1.35 in the nonnegative case and from part (vi)
of Theorem 1.44 in the integrable case. The third equality follows from part (ii)
of Theorem 1.35 in the nonnegative case and from part (v) of Theorem 1.44 in the
integrable case. Since f AnN D gAnN , 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/ D 1, 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 D 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 W X ! Œ0; 1 is measurable, or that f W X ! R

is -integrable. Then the following are equivalent:
(i) f D 0 almost everywhere;
(ii) A f d D 0 for all A 2 A;
R
R
(iii) X jf j d D 0.
Proof. That (i) implies (ii) is the content of Lemma 1.48. That (ii) implies (iii) is
obvious in the nonnegative case. In the integrable case, define
AC WD ¹x 2 X j f .x/ 0º; A WD ¹x 2 X j f .x/ < 0º:
Then f C D f AC and f D f A by (1.24). Hence
Z Z Z Z Z
C
jf j d D f d C f d D f d f d D 0
X X X AC A
by Theorem 1.44 and (ii).

It remains to prove
R that (iii) implies (i). Let f W X ! Œ0; 1 be a measurable
function such that X f D 0 and define the measurable sets
n
An WD ¹x 2 X j f .x/ > 2 º for n 2 N:
Then Z Z
n n
2 .An / D 2 An d f d D 0
X X
for all n 2 N by Theorem 1.35. Hence .An / D 0 for all n 2 N, and so
1
[
N WD ¹x 2 X j f .x/ ¤ 0º D An
nD1
is a set of measure zero. In the integrable case apply this argument to the function
jf jW X ! Œ0; 1/. This proves Lemma 1.49.
Lemma 1.50. Let f 2 L1 ./. Then
ˇZ ˇ Z
ˇ ˇ
ˇ f dˇ D jf j d (1.34)
ˇ ˇ
X X
if and only if f D jf j almost everywhere or f D jf j almost everywhere.

R R R R
Proof. Assume R(1.34). Then X f d D X jf j d or X f d D X jf j d.
In the first case X .jf j f / dRD 0, and so jf j f D 0 almost everywhere by
Lemma 1.49. In the second case X .jf j C f / d D 0, and so jf j C f D 0 almost
everywhere. This proves Lemma 1.50.
Definition 1.51 (the Banach space L1 ./). Define an equivalence relation on the
real vector space of all measurable function from X to R by
def
f g () the set ¹x 2 X j f .x/ ¤ g.x/º (1.35)
has measure zero:
Thus two functions are equivalent if and only if 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; gW X ! R are measurable,

f 2 L1 ./, and f g, then g 2 L1 ./. Moreover, the set ¹f 2 L1 ./ j f 0º
is a linear subspace of L1 ./ and hence the quotient space

L1 ./ WD 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
L ./ ! RW f 7 ! jf j d DW kf kL1
1
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 mea-
sure space and let fn W X ! R be a sequence of -integrable functions such that
X1 Z
jfn j d < 1: (1.36)
nD1 X
Then there are a set N of measure zero and a function f 2 L1 ./ such that
1
X 1
X
jfn .x/j < 1 and f .x/ D fn .x/ for all x 2 X n N; (1.37)
nD1 nD1
Z 1 Z
X
f d D fn d for all A 2 A; (1.38)
A nD1 A
and Z ˇ n
X
ˇ
ˇ ˇ
lim ˇf fk ˇˇ d D 0: (1.39)
n!1 X ˇ
kD1
Proof. Define
1
X
.x/ WD jfk .x/j
kD1
for x 2 X . This function is measurable by part (ii) of Theorem 1.24. Moreover,

it follows from the Lebesgue monotone convergence theorem (Theorem 1.37) and
from part (i) of Theorem 1.38 that
Z n
Z X
d D lim jfk j d
X n!1 X
kD1
n
X Z
D lim jfk j d
n!1 X
kD1
1 Z
X
D jfk j d
X
kD1
< 1:
Hence the set

N WD ¹x 2 X j .x/ D 1º
has measure zero by Lemma 1.47 and

1
X
jfk .x/j < 1 for all x 2 X n N .
kD1
Define the function f W X ! R by

8
ˆ
ˆ0 for x 2 N ,
<
1
f .x/ WD X
ˆ
ˆ fk .x/ for x 2 X n N:
:
kD1
Then f satisfies (1.37). Define the functions gW X ! R and gn W X ! R by

n
X
g WD XnN ; gn WD fk X nN for n 2 N:
kD1
These functions are measurable by part (i) of Theorem 1.24. Moreover,

Z Z
g d D d < 1
X X
by Lemma 1.48. Since jgn .x/j g.x/ for all n 2 N and gn converges pointwise to
f , it follows from the Lebesgue dominated convergence theorem (Theorem 1.45)
that f 2 L1 ./ and, for all A 2 A,
Z Z n
Z X 1 Z
X
f d D lim gn d D lim fk d D fn d:
A n!1 A n!1 A
kD1 nD1 A
Here the second step follows from Lemma 1.48 because gn D nkD1 fk almost
P
everywhere. The last step follows by interchanging sum and integral, using part (i)
of Theorem 1.44. This proves (1.38). To prove (1.39) note that
n
X
f fk D f gn
kD1
almost everywhere, that f .x/ gn .x/ converges to zero for all x 2 X , and that
jf gn j jf j C g, where jf j C g is integrable. Hence, by Lemma 1.48 and the
Lebesgue dominated convergence theorem (Theorem 1.45),
Z ˇ n
X ˇ Z
ˇ ˇ
lim ˇf fk ˇˇ d D lim jf gn j d D 0;
n!1 X ˇ n!1 X
kD1
This proves (1.39) and Theorem 1.52.
Theorem 1.53 (completeness of L1 ). Let .X; A; / be a measure space and let

fn 2 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 2 N such that, for
all m; n 2 N, Z
n; m n0 H) jfn fm j d < ": (1.40)
X
Then there exists a function f 2 L1 ./ such that

Z
lim jfn f j d D 0: (1.41)
n!1 X
Moreover, there is a subsequence fni that converges almost everywhere to f .

Proof. By assumption, there is a sequence ni 2 N such that

Z
jfniC1 fni j d < 2 i ; ni < niC1 ; for all i 2 N:
X
Then the sequence

gi WD fniC1 fni 2 L1 ./
satisfies (1.36). Hence, by Theorem 1.52, there exists a function g 2 L1 ./ such
that
X1 1
X
gD gi D .fniC1 fni /
i D1 iD1
almost everywhere and
Z ˇ kX1 ˇ Z
ˇ ˇ
0 D lim ˇ gi g ˇˇ d D lim jfnk fn1 gj d: (1.42)
k!1 k!1 X
ˇ
X i D1
Define
f WD fn1 C g:
Then
i 1
X
fni D fn1 C gj
j D1
converges almost everywhere to f . We prove (1.41). Let " > 0. By (1.42), there is
an ` 2 N such that
Z
jfnk f j d < "=2 for all k `.
X
By (1.40), the integer ` can be chosen such that

Z
jfn fm j d < "=2 for all n; m n` .
X
Then Z Z Z
jfn f j d jfn fn` j d C jfn` f j d < "
X X X
for all n n` . This proves (1.41) and Theorem 1.53.

1.6. Completion of a measure space 43
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 2 A; .N / D 0; E N H) E 2 A:
Theorem 1.55. Let .X; A; / be a measure space and define

ˇ there exist measurable sets A; B 2 A such that
² ˇ ³
A WD E X ˇ
ˇ :
A E B and .B n A/ D 0
Then the following holds.
(i) A is a -algebra and A A .
(ii) There exists a unique measure W A ! Œ0; 1 such that
jA D :
(iii) The triple .X; A ; / is a complete measure space. It is called the completion
of .X; A; /.
(iv) If f W X ! R is -integrable, then f is -integrable and, for E 2 A,

Z Z

f d D f d (1.43)
E E
This continues to hold for all A-measurable functions f W X ! Œ0; 1.

x is A -measurable, then there exists an A-measurable function
(v) If f W X ! R
x
f W X ! R such that the set
N WD ¹x 2 X j f .x/ ¤ f .x/º 2 A
has measure zero, i.e., .N / D 0.

Proof. (i) First, X 2 A because A A . Second, let E 2 A and choose

A; B 2 A such that A E B and .B n A/ D 0. Then B c E c Ac and
Ac n B c D Ac \ B D B n A. Hence .Ac n B c / D 0 and so E c 2 A . Third,
let Ei 2 A for i 2 N and choose Ai ; Bi 2 A such that Ai Ei Bi and
.Bi n Ai / D 0. Define
[ [ [
A WD Ai ; E WD Ei ; B WD Bi :
i i i
S S
Then A E B and B n A D i .Bi n A/ i .Bi n Ai /. Hence
X
.B n A/ .Bi n Ai / D 0
i
and this implies E 2 A . Thus we have proved (i).

(ii) For E 2 A define
.E/ WD .A/ where A; B 2 A; (1.44)

A E B;
.B n A/ D 0:
This is the only possibility for defining a measure W A ! Œ0; 1 that agrees
with on A because .A/ D .B/ whenever A; B 2 A such that A B
and .B n A/ D 0. To prove that is well defined, let E 2 A and A; B 2 A as
in (1.44). If A0 ; B 0 2 A is another pair such that A0 E B 0 and .B 0 n A0 / D 0,
then A n A0 E n A0 B 0 n A0 and hence .A n A0 / D 0. This implies
.A/ D .A \ A0 / D .A0 /, where the last equality follows by interchanging
the roles of the pairs .A; B/ and .A0 ; B 0 /. Thus the map W A ! Œ0; 1 in (1.44)
is well defined.
We prove that is a measure. Let Ei 2 A be a sequence of pairwise disjoint
sets and choose sequences Ai ; Bi 2 A such that Ai Ei Bi for all i . Then the
Ai are pairwise disjoint and .Ei / D .Ai / for all i . Moreover, A WD i Ai 2 A,
S
B WD i Bi 2 A, A E B, and .B n A/ D 0 as we have seen in the proof of

S
part (i). Hence .E/ D .A/ D i .Ai / D i .Ei /. This proves (ii).
P P
(iii) Let E 2 A such that .E/ D 0 and let E 0 E. Choose A; B 2 A

such that A E B and .B n A/ D 0. Then .A/ D .E/ D 0 and hence
.B/ D .A/ C .B n A/ D 0. Since E 0 E B, this implies that E 0 2 A
(by choosing B 0 WD B and A0 WD ;). This shows that .X; A ; / is a complete
measure space.
1.6. Completion of a measure space 45
(iv) Assume f W X ! Œ0; 1 is A-measurable. By Theorem 1.26, there exists a

sequence of A-measurable step functions sn W X ! R such that
0 s1 s2 f
and
f .x/ D lim sn .x/ for all x 2 X .
n!1
Since jA D we have
Z Z
sn d D sn d
X X
for all n and hence it follows from the Lebesgue monotone convergence theorem
(Theorem 1.37) for both and that
Z Z Z Z
f d D lim sn d D lim sn d D f d :
X n!1 X n!1 X X
This proves (1.43) for E D X and all A-measurable functions f W X ! Œ0; 1.
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 W X ! Œ0; 1. That it
continues to hold for all f 2 L1 ./ follows directly from Definition 1.43. This
proves (iv).
(v) If f D E for E 2 A , choose A; B 2 A such that
A E B; .B n A/ D 0;
and define f WD A . Then
N D ¹x 2 X j f .x/ ¤ f .x/º D E n A B n A:
Hence .N / .B n A/ D .B n A/ D 0. This proves (v) for characteris-

tic functions of A -measurable sets. For A -measurable step functions the asser-
tion follows by multiplication with real numbers and taking finite sums. Now let
f W X ! Œ0; 1 be an arbitrary A -measurable function. By Theorem 1.26, there
exists a sequence of A -measurable step functions si W X ! Œ0; 1/ such that si
converges pointwise to f . For each i 2 N choose an A-measurable step function
si W X ! Œ0; 1/ and a set Ni 2 A such that si D si on X n Ni and .Ni / D 0.
Then there is a sequence of sets Ni 2 A such that Ni Ni and .Ni / D 0 for
all i . Define f W X ! Œ0; 1 by
´
f .x/ if x … N; [
f .x/ WD N WD Ni :
0 if x 2 N; i
Then N 2 A, .N / D 0, and the sequence of A-measurable functions si XnN

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 N by definition. Now let
f W X ! R x be A -measurable. Then so are .f /˙ WD max¹˙f ; 0º. Construct
f W X ! Œ0; 1 as above. Then f .x/ D 0 whenever f C .x/ > 0 and vice versa.
˙
Thus f WD f C 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 com-
pletion. Denote the equivalence class of a -integrable function f 2 L1 ./ under
the equivalence relation (1.35) in Definition 1.51 by
Œf WD ¹g 2 L1 ./ j .¹x 2 X j f .x/ ¤ g.x/º/ D 0º:
Then the map

L1 ./ ! L1 . /W Œf 7 ! Œf (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 W E ! R, defined on
a measurable subset E X, measurable if .X nE/ D 0 and the set f 1 .B/ E
is measurable for every Borel set B R. Call f integrable if the function on all
of X , obtained by setting f jX nE D 0, is integrable.
If .X; A; / is complete our integrable function f W E ! R can be extended in
any manner whatsoever to all of X , and the extended function on X is then inte-
grable in the original sense, regardless of the choice of the extension. Moreover,
its integral over any measurable set A 2 A is unaffected by the choice of the exten-
sion (see Lemma 1.48).
With this extended notion of integrability we see that the Lebesgue dominated
convergence theorem (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 Theo-
rem 1.52, where the series 1
P
nD1 fn only converges on the complement of a set N
of measure zero, and the function f can only be naturally defined on E WD X n N .
Our choice in the proof of Theorem 1.52 was to define f jN WD 0, but this choice
1.7. Exercises 47
does not affect any of the statements of the theorem. Moreover, when working with

the quotient space L1 ./ D 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
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/ WD
1 if Ac is countable;
for A 2 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/ < 1 and let
fn W X ! Œ0; 1/ be a sequence of bounded measurable functions that converges
uniformly to f W X ! Œ0; 1/. Prove that
Z Z
f d D lim fn d: (1.46)
X n!1 X
Find an example of a measure space .X; A; / with .X/ D 1 and a sequence

of bounded measurable functions fn W X ! Œ0; 1/ converging uniformly to f such
that (1.46) does not hold.
Exercise 1.59. (i) Let fn W Œ0; 1 ! Œ 1; 1 be a sequence of continuous functions

that converges uniformly to zero. Show that
Z 1
lim fn .x/ dx D 0:
n!1 0
(ii) Let fn W Œ0; 1 ! Œ 1; 1 be a sequence of continuous functions such that
lim fn .x/ D 0 for all x 2 Œ0; 1:

n!1
Prove that Z 1
lim fn .x/ dx D 0;
n!1 0
without using Theorem 1.45. A good reference is Eberlein [3].

(iii) Construct a sequence of continuous functions fn W Œ0; 1 ! Œ 1; 1 that con-

verges pointwise, but not uniformly, to zero.
R 1 (iv) Construct a sequence of continuous functions fn W Œ0; 1 ! Œ 1; 1 such that

0 fn .x/ dx D 0 for all n and fn .x/ does not converge for any x 2 Œ0; 1.
Exercise 1.60. Let .X; A; / be a measure space and f W X ! Œ0; 1 be a measur-

able function such that Z
0 < c WD f d < 1:
X
Prove that
8
Z f ˛ <1 if ˛ < 1;
ˆ
lim n log 1 C d D c if ˛ D 1; for 0 < ˛ < 1:
n!1 X n˛ ˆ
0 if ˛ > 1;
:
Hint. The integrand can be estimated by ˛f when ˛ 1.
Exercise 1.61. Let X WD N and A WD 2N and let W 2N ! Œ0; 1 be the counting

measure (Example 1.30). Prove that a function f W N ! R is -integrable if and
only if the sequence .f .n//n2N of real numbers is absolutely summable and that in
this case Z 1
X
f d D f .n/:
N nD1
Exercise 1.62. Let .X; A/ be a measurable space and let n W A ! Œ0; 1 be a

sequence of measures. Show that the formula
1
X
.A/ WD n .A/
nD1
for A 2 A defines a measure W A ! Œ0; 1. Let f W X ! R be a measurable

function. Show that f is -integrable if and only if
1 Z
X
jf j dn < 1:
nD1 X
If f is -integrable, prove that

Z 1 Z
X
f d D f dn :
X nD1 X
1.7. Exercises 49
Exercise 1.63. Let .X; A; / be a measure space such that .X/ < 1 and let
f W X ! R be a measurable function. Show that f is integrable if and only if
1
X
j.¹x 2 X j jf .x/j > nº/j < 1:
nD1
Exercise 1.64. Let .X; A; / be a measure space and let f W X ! R be a

-integrable function.
(i) Prove that for every " > 0 there exists a ı > 0 such that, for any A 2 A,
ˇZ ˇ
ˇ ˇ
.A/ < ı H) ˇˇ f dˇˇ < ":
A
Hint. Argue indirectly. See Lemma 5.21.
(ii) Prove that for every " > 0 there exists a measurable set A 2 A such that, for
any B 2 A, ˇZ ˇ
Z
ˇ ˇ
B A H) ˇ f d
ˇ f dˇˇ < ":
X B
Exercise 1.65. Let .X; A/ be a measurable space and define

´
0 if A D ;;
.A/ WD
1 if A 2 A and A ¤ ;:
Determine the completion .X; A ; / and the space L1 ./.
Exercise 1.66. Let .X; A; / be a measure space such that D ıx0 is the
Dirac measure at some point x0 2 X (Example 1.31). Determine the completion
.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 jA D . Prove that L1 ./ L1 .0 / and
Z Z
f d D f d0
X X
for every f 2 L1 ./.

Hint. Prove the following.
(i) Let f W X ! Œ0; 1 be A-measurable and define

8
ˆ
ˆ 0 if f .x/ ı;
<
fı .x/ WD f .x/ if ı < f .x/ ı 1
;
ˆ
ˆ
: 1
ı if f .x/ > ı 1 :
Then fı is A-measurable for every ı > 0 and

Z Z
lim fı d D f d:
ı!0 X X
(ii) Let 0 < c < 1, let f W X ! Œ0; c be A-measurable, and assume that
.¹x 2 X j f .x/ > 0º/ < 1:
Then Z Z
f d D f d0 :
X X
(Consider also the function c f .)
Exercise 1.69 (pushforward of a measure). Let .X; A; / be a measure space,

let Y be a set, and let W X ! Y be a map. The pushforward of A is the -algebra
A WD ¹B Y j 1
.B/ 2 Aº 2Y : (1.47)
The pushforward of is the function
W A ! Œ0; 1
defined by
1
. /.B/ WD . .B//; for B 2 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/ D . 1

.E// for all E 2 . A/ A : (1.49)
Deduce that .Y; A; / is complete whenever .X; A; / is complete. Find

an example where . A/ ¨ A .
1.7. Exercises 51
(iii) Fix a function f W Y ! Œ0; 1. Prove that f is A-measurable if and only if
f ı is A-measurable. If f is A-measurable, prove that
Z Z
f d. / D .f ı / 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 ./ D 1:
The underlying set is called the sample space, the -algebra F 2 is called the
set of events, and the measure
P W F ! Œ0; 1
is called a probability measure. Examples of finite sample spaces are
• the set D ¹h; tº for tossing a coin,
• the set D ¹1; 2; 3; 4; 5; 6º for rolling a dice,
• the set D ¹00; 0; 1; : : : ; 36º for spinning a roulette wheel, and
• the set D ¹2; : : : ; 10; j; q; k; aº ¹}; ~; ; |º for drawing a card from a

deck.
Examples of infinite sample spaces are the set D N [ ¹1º 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.
A random variable is an integrable function
XW ! R:
Its expectation E.X/ and variance V.X/ are defined by

Z
E.X/ WD X dP;

and Z
V.X/ WD .X E.X//2 dP D E.X 2 / E.X/2 ;

respectively.
Given a random variable X W ! 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 prob-
ability of the event that the random variable X takes its value in the set B and is
denoted by
P .X 2 B/ WD P .X 1 .B// D .X P /.B/:
Here X P denotes the pushforward of the probability measure P to the Borel
-algebra B 2R (Exercise 1.69). By (1.50) the expectation and variance of X
are given by
Z
E.X/ D x d.X P /.x/
R
and Z
V.X/ D .x E.X//2 d.X P /.x/;
R
respectively.
The (cumulative) distribution function of a random variable X is the function
FX W R ! Œ0; 1
defined by
FX .x/ WD P .X x/ D P .¹! 2 j X.!/ xº/ D .X P /.. 1; x/:
It is nondecreasing and right continuous, satisfies
lim FX .x/ D 0; lim FX .x/ D 1;

x! 1 x!1
and the integral of a continuous function on R with respect to the pushforward mea-
sure X P agrees with the Riemann–Stieltjes integral (Exercise 6.20) with respect
to FX . Moreover,
FX .x/ lim FX .t/ D P .X 1 .x//
t !x
by Theorem 1.28. Thus FX is continuous at x if and only if P .X 1 .x// D 0. This

leads to the following notions of convergence. Let X W ! R be a random variable.
1.7. Exercises 53
A sequence .Xi /i2N of random variables is said to
• converge in probability to X if
lim P .jXi X j "/ D 0 for all " > 0,

i!1
• converge in distribution to X if
FX .x/ D lim FXi .x/

i!1
for every x 2 R such that FX is continuous at x.
We prove that that convergence almost everywhere implies convergence in prob-

ability. Let " > 0 and define
Ai WD ¹! 2 j jXi .!/ X.!/j "º
Let E be the set of all ! 2 such that the sequence Xi .!/ does not con-
verge to X.!/. This set is measurable by Example 1.46 and has measure zero by
convergence almost everywhere. Moreover,
\ [
Aj E
i2N j i
and so [
lim P Aj D P .E/ D 0
i!1
j i
by Theorem 1.28. Thus

lim P .Ai / D 0:
i !1
We prove that convergence in probability implies convergence in distribution.

Let x 2 R such that FX is continuous at x. Let " > 0 and choose ı > 0 such that
" "
FX .x/ < FX .x ı/ FX .x C ı/ < FX .x/ C :
2 2
Now choose i0 2 N such that
"
P .jXi X j ı/ < for all i i0 .
2
Then
FX .x ı/ P .jXi Xj ı/ FXi .x/
FX .x C ı/ C P .jXi X j ı/
and hence
FX .x/ " < FXi .x/ < FX .x/ C " for all i i0 .
This shows that
lim FXi .x/ D FX .x/
i !1
as claimed.
A finite collection of random variables X1 ; : : : ; Xn is called independent if,
for every collection of Borel sets B1 ; : : : ; Bn R,
n
\ Yn
P Xi 1 .Bi / D P .Xi 1 .Bi //:
iD1 i D1
In Chapter 7 we shall see that this condition asserts that the pushforward of P under
the map
X WD .X1 ; : : : ; Xn /W ! 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 W ! R
that satisfy the following:
(a) the random variables X1 ; : : : ; Xn are independent for all n;

(b) the Xk have expectation E.Xk / D 0;
(c) the Xk are identically distributed, i.e., FXk D FX` for all k and `.
For n 2 N define
Sn WD X1 C C Xn :
Kolmogorov’s strong law of large numbers asserts that, under these assumptions,
the sequence Sn =n converges almost everywhere to zero. (This continues to hold
when (c) is replaced by the assumption 1 1
P
kD1 k 2 V.Xk / < 1.) If, in addition,
2
V.Xk / D for all k and some positive real number , then the central limit the-
orem of Lindeberg and Lévy asserts that the sequence
p
Tn WD Sn = n
1.7. Exercises 55
converges in distribution to a so-called standard normal random variable with ex-

pectation zero and variance one, i.e.,
Z x
1 2
lim FTn .x/ D p e t =2 dt for all x 2 R.
n!1 2 1
The hypotheses listed above are quite restrictive and in modern probability
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 func-
tions FX W R ! Œ0; 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 Radon–Nikodým theorem.
Chapter 2
The Lebesgue measure
This chapter introduces the most important example, namely the Lebesgue mea-
n
sure on Euclidean space. Let n 2 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 C x WD ¹y C x j y 2 Bº 2 B for all B 2 B and all x 2 Rn ;
because the translation Rn ! Rn W y 7! y C x is a homeomorphism. A measure
W B ! Œ0; 1 is called translation-invariant if it satisfies
.B C x/ D .B/ for all B 2 B and all x 2 Rn : (2.1)
Theorem 2.1. There exists a unique measure W B ! Œ0; 1 that is translation-
invariant and satisfies the normalization condition .Œ0; 1n / D 1.
Proof. See page 74.
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 ˇ there exist Borel sets B0 ; B1 2 B
² ˇ ³
A WD A R ˇ
ˇ
(2.2)
such that B0 A B1 and .B1 n B0 / D 0
and
m.A/ WD .B0 / for A 2 A,
where B0 ; B1 2 B are chosen such that B0 A B1 and .B1 n B0 / D 0.
The elements of A are called Lebesgue measurable subsets of Rn , the function
mW A ! Œ0; 1
is called the Lebesgue measure, and the triple .Rn ; A; m/ is called the Lebesgue
measure space. A function
f W 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).
58 2. The Lebesgue measure
2.1 Outer measures

In preparation for the proof of Theorem 2.1 we now take up the idea, announced 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
W 2X ! Œ0; 1
is called an outer measure if it satisfies the following three axioms:
(a) .;/ D 0;
(b) if A B X , then
.A/ .B/I
(c) if Ai X for i 2 N, then

1
[ X1
Ai .Ai /:
i D1 i D1
Let W 2X ! Œ0; 1 be an outer measure. A subset A X is called -measurable

if it satisfies
.D/ D .D \ A/ C .D n A/ (2.3)
for every subset D X .
The inequality .D/ .D \ A/ C .D n 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. Carathéodory’s
theorem (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 Carathéodory
criterion) characterizes outer measures on metric spaces such that every Borel set
is -measurable.
Theorem 2.4 (Carathéodory). Let X be a set, let W 2X ! Œ0; 1 be an outer
measure, and define
A WD A./ WD ¹A X j A is -measurableº (2.4)
Then A is a -algebra, the function
WD jA W A ! Œ0; 1
is a measure, and the measure space .X; A; / is complete.
2.1. Outer measures 59
Proof. The proof is carried out in six steps.
Step 1. X 2 A.
For every subset D X , we have
.D \ X/ C .D n X/ D .D/ C .;/ D .D/
by condition (a) in Definition 2.3. Hence X 2 A.
Step 2. If A 2 A, then Ac 2 A.
Let A 2 A. Since
D \ Ac D D n A; D n Ac D D \ A;
it follows from (2.3) that
.D/ D .D \ Ac / C .D n Ac /
for every subset D X . Hence Ac 2 A.
Step 3. If A; B 2 A, then A [ B 2 A.
Let A; B 2 A. Then, for every subset D X ,
.D/ D .D \ A/ C .D n A/

D .D \ A/ C .D \ Ac /
D .D \ A/ C .D \ Ac \ B/ C ..D \ Ac / n B/
..D \ A/ [ .D \ Ac \ B// C .D \ Ac \ B c /
D .D \ .A [ B// C .D \ .A [ B/c /
D .D \ .A [ B// C .D n .A [ B//:
Here the inequality follows from axioms (a) and (c) in Definition 2.3. Using ax-
ioms (a) and (c) again, we obtain
.D/ D .D \ .A [ B// C .D n .A [ B//
for every subset D X and hence A [ B 2 A.

Step 4. Let Ai 2 A for i 2 N such that Ai \ Aj D ; for i ¤ j . Then

1
[ 1
X
A WD Ai 2 A; .A/ D .Ai /:
iD1 i D1
For k 2 N define
Bk WD A1 [ A2 [ [ Ak :
Then Bk 2 A for all k 2 N by Step 3. Now let D X . Then, for all k 2,
.D \ Bk / D .D \ Bk \ Ak / C ..D \ Bk / n Ak /
D .D \ Ak / C .D \ Bk 1/
and so, by induction on k,

k
X
.D \ Bk / D .D \ Ai /:
iD1
Since Bk 2 A, this implies

.D/ D .D \ Bk / C .D n Bk /
k
X
D .D \ Ai / C .D n Bk /
iD1
k
X
.D \ Ai / C .D n A/:
iD1
Here the last inequality follows from axiom (b) in Definition 2.3. Since this holds
for all k 2 N and D \ A D 1
S
iD1 .D \ Ai /, it follows that
1
X
.D/ .D \ Ai / C .D n A/
i D1
.D \ A/ C .D n A/
.D/:
Here the last two inequalities follow from axiom (c). Hence
1
X
.D/ D .D \ Ai / C .D n A/ D .D \ A/ C .D n A/ (2.5)
iD1
for all D X . This shows that A 2 A. Now take D D A to obtain D n A D ; and

D \ Ai D Ai . Then it follows from (2.5) that .A/ D 1
P
iD1 .Ai /.
Step 5. Let Ai 2 A for i 2 N. Then

1
[
A WD Ai 2 A:
iD1
Define ´
A1 for i D 1,
Bi D
Ai n .A1 [ [ Ai 1/ for i 2.
Then Bi \ Bj D ; for i ¤ j and Bi D .A1 [ [ Ai 1 [ Aci /c 2 A for all i by
Steps 2 and 3. Hence A D 1i D1 Bi 2 A by Step 4.
S
Step 6. .X; A; / is a complete measure space.

It follows from Steps 1, 2, 4, and 5 that .X; A; D jA / is a measure space.
We prove that it is complete. To see this, let A X and suppose that A N
where N 2 A satisfies .N / D 0. Then it follows from axiom (b) in Definition 2.3
that .A/ .N / D .N / D 0 and therefore .A/ D 0. Now use axioms (a), (b)
and (c) to obtain
.D/ .D \ A/ C .D n A/ .A/ C .D/ D .D/;
and so .D/ D .D \ A/ C .D n A/ for all D X, which shows that A 2 A.
This yields Step 6 and completes the proof of Theorem 2.4.
Theorem 2.5 (Carathéodory criterion). Let .X; d / be a metric space. Consider an
outer measure W 2X ! Œ0; 1. Let A./ 2X be the -algebra given by (2.4)
and B 2X be the Borel -algebra of .X; d /. Then the following statements are
equivalent:
(i) B A./;
(ii) if A; B X satisfy d.A; B/ WD inf d.a; b/ > 0; then
a2A; b2B
.A [ B/ D .A/ C .B/:
Proof. We prove that (i) implies (ii). Thus assume that satisfies (i). Let A; B X
such that " WD d.A; B/ > 0. Define
[
U WD ¹x 2 X j there exists a 2 A such that d.a; x/ < "º D B" .a/:
a2A
Then U is open, A U , and U \ B D ;. Hence U 2 B A./ by assumption

and hence .A [ B/ D ..A [ B/ \ U / C ..A [ B/ n U / D .A/ C .B/: Thus
the outer measure satisfies (ii).
We prove that (ii) implies (i). Thus assume that satisfies (ii). We show that
every closed set A X is -measurable, i.e., .D/ D .D \ A/ C .D n A/ for all
D X . Since .D/ .D \ A/ C .D n 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/ < 1. Then
.D/ .D \ A/ C .D n A/:
To see this, replace the set D n A by the smaller set D n Uk , where
[
Uk WD ¹x 2 X j there exists a 2 A such that d.a; x/ < 1=kº D B1=k .a/:
a2A
For each k 2 N, the set Uk is open and d.x; y/ 1=k for all x 2 D \ A and all
y 2 D n Uk . Hence
1
d.D \ A; D n Uk / :
k
By (ii) and axiom (b) this implies
.D \ A/ C .D n Uk / D ..D \ A/ [ .D n Uk // .D/ (2.6)
for every subset D X and every k 2 N. We will prove the following.
Claim 2. lim .D n Uk / D .D n A/.
k!1
Claim 1 follows directly from Claim 2 and (2.6). To prove Claim 2 note that
1
\
AD Ui ;
iD1
because A is closed. (If x 2 Ui for all i 2 N, then there exists a sequence ai 2 A

such that d.ai ; x/ < 1=i and hence x D lim ai 2 A.) This implies
i !1
1
[ 1
[
Uk n A D .Uk n Ui / D Ui n Ui C1
i D1 iDk
and hence
D n A D .D n Uk / [ .D \ .Uk n A//
1
[
D .D n Uk / [ .D \ .Ui n UiC1 //:
i Dk
Thus
1
[
D n A D .D n Uk / [ Ei ; Ei WD .D \ Ui / n Ui C1 : (2.7)
i Dk
Claim 3. The outer measures of the Ei satisfy

1
X
.Ei / < 1:
iD1
Claim 3 H) Claim 2. It follows from Claim 3 that the sequence

1
X
"k WD .Ei /
iDk
converges to zero. Moreover, it follows from equation (2.7) and axiom (c) in Defi-
nition 2.3 that
1
X
.D n A/ .D n Uk / C .Ei / D .D n Uk / C "k :
i Dk
Hence it follows from axiom (b) in Definition 2.3 that
.D n A/ "k .D n Uk / .D n A/
for every k 2 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 C 2.
Claim 4 H) Claim 3. It follows from Claim 4, axiom (b), and (ii) that
n
X n
[
.E2i / D E2i .D/
i D1 iD1
and
n
X n
[
.E2i 1/ D E2i 1 .D/
i D1 iD1
for every n 2 N. Hence

1
X
.Ei / 2.D/ < 1
i D1
and this shows that Claim 4 implies Claim 3.

Proof of Claim 4. We show that

1
d.Ei ; Ej / for j i C 2:
.i C 1/.i C 2/
To see this, fix indices i; j with j i C 2. Let x 2 Ei and y 2 X such that
1
d.x; y/ < :
.i C 1/.i C 2/
Then x … UiC1 because Ei \ UiC1 D ;. (See (2.7).) Hence
1
d.a; x/ for all a 2 A:
i C1
This implies
d.a; y/ d.a; x/ d.x; y/
1 1
>
i C1 .i C 1/.i C 2/
1
D
i C2
1

j
for all a 2 A. Hence y … Uj and consequently 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 WD Q.a; b/
WD Œa1 ; b1 Œa2 ; b2 Œan ; bn (2.8)
n
D ¹x D .x1 ; : : : ; xn / 2 R j aj xj bj for j D 1; : : : ; nº
for a1 ; : : : ; an ; b1 ; : : : ; bn 2 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// WD Voln .Q.a; b// WD .bj aj /: (2.9)
j D1
2.2. The Lebesgue outer measure 65
The volume of the open cuboid

n
Y
U WD int.Q/ D .ai ; bi /
iD1
is defined by
Vol.U / WD Vol.Q/:
The set of all closed cuboids in Rn will be denoted by
² ˇ ³
ˇ a ; : : : ; an ; b1 ; : : : ; bn 2 R;
Qn WD Q.a; b/ ˇˇ 1 :
aj < bj for j D 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` 2 Qn such that
`
[ `
X
A Qi ; Vol.Qi / < ":
iD1 iD1
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 2 Qn , i 2 N, such that
1
[ 1
X
A Qi ; Vol.Qi / < ":
iD1 iD1
Definition 2.9. The Lebesgue outer measure on Rn is the function

n
D n W 2 R ! Œ0; 1
defined by
1
°X ˇ 1
[ ±
.A/ WD inf Voln .Qi / ˇ Qi 2 Qn ; A Qi for A Rn :
ˇ
(2.10)
i D1 i D1
n
Theorem 2.10 (the Lebesgue outer measure). Let W 2R ! Œ0; 1 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 2 Rn
.A C x/ D .A/;
(iii) if A; B Rn are such that d.A; B/ > 0, then .A [ B/ D .A/ C .B/;
(iv) .int.Q// D .Q/ D Vol.Q/ for all Q 2 Qn .
Proof. We prove (i). The empty set is contained in every cuboid Q 2 Qn .

Since there are cuboids with arbitrarily small volume, it follows that .;/ D 0.
If A B Rn it follows directly from Definition 2.9 that .A/ .B/. Now let
Ai Rn for i 2 N, define
1
[
A WD Ai ;
iD1
and fix a constant " > 0. Then it follows from Definition 2.9 that, for i 2 N, there
exists a sequence of cuboids Qij 2 Qn , j 2 N, such that
1 1
[ X "
Ai Qij ; Vol.Qij / < C .Ai /:
2i
j D1 j D1
Hence
1 1
[ X X " X
A Qij ; Vol.Qij / < C .A i / D " C .Ai /:
2i
i;j 2N i;j 2N i D1 iD1
This implies
1
X
.A/ < " C .Ai /
iD1
for every " > 0, and thus .A/ 1

P
i D1 .Ai /. This proves part (i).S
We prove (ii). If A iD1 Qi with Qi 2 Qn , then A C x 1
S1
iD1 .Qi C x/
n
for every x 2 R and Vol.Qi C x/ D Vol.Qi /, by the 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 2 Qn such that
1
[ 1
X
A[B Qi ; Vol.Qi / < .A [ B/ C ":
iD1 i D1
Subdividing each Qi into finitely many smaller cuboids, if necessary, we may

assume without loss of generality that
d.A; B/
diam.Qi / WD sup jx yj < :
x;y2Qi 2
Here jj denotes the Euclidean norm on Rn . Then, for every i 2 N, we have either
Qi \ A D ; or Qi \ B D ;. This implies
I \ J D ;; I WD ¹i 2 N j Qi \ A ¤ ;º; J WD ¹i 2 N j Qi \ B ¤ ;º:
Hence
X X
.A/ C .B/ Vol.Qi / C Vol.Qi /
i 2I i 2J
1
X
Vol.Qi /
iD1
< .A [ B/ C ":
Thus .A/ C .B/ < .A [ B/ C " for all " > 0, so .A/ C .B/ .A [ B/,
and hence .A/ C .B/ D .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 D I1 In ; Ii D Œai ; bi :
We claim that
Vol.Q/ .Q/: (2.11)
Equivalently, if Qi 2 Qn , i 2 N, is a sequence of closed cuboids, then
1
[ 1
X
Q Qi H) Vol.Q/ Vol.Qi /: (2.12)
iD1 i D1
For a closed interval I D Œa; b R with a < b define
jI j WD b a:
Then
jI j 1 #.I \ Z/ jI j C 1:
Hence
N jI j 1 #.NI \ Z/ N jI j C 1;
and thus
1 1 1 1
jI j # I \ Z jI j C
N N N N
for every integer N 2 N. Take the limit N ! 1 to obtain
1 1
jI j D lim # I\ Z :
N !1 N N
Thus
n
Y 1 1
Vol.Q/ D lim # Ij \ Z
N !1 N N
j D1 (2.13)
1 1 n
D lim # Q \ Z :
N !1 N n N
Now suppose Qi 2 Qn , i 2 N, is a sequence of closed cuboids such that
1
[
Q Qi :
iD1
Fix a constant " > 0 and choose a sequence of open cuboids Ui Rn such that
"
Qi Ui ; Vol.Ui / < Vol.Qi / C i :
2
Since Q is compact and the Ui form an open cover of Q, there exists a constant
k 2 N such that
[k
Q Ui :
iD1
This implies
k
1 1 n X 1 1 n
# Q \ Z # U i \ Z
Nn N Nn N
i D1
k
X 1 x 1 n
# U i \ Z :
Nn N
i D1
Take the limit N ! 1 and use (2.13) to obtain

k
X
Vol.Q/ Vol.Ui /
i D1
1
X
Vol.Ui /
i D1
1
X "
C Vol.Qi /
2i
i D1
1
X
D"C Vol.Qi /:
iD1
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/ D Vol.Q/. To prove that
.int.Q// D Vol.Q/, fix a constant " > 0 and choose a closed cuboid P 2 Qn such
that
P int.Q/; Vol.Q/ " < Vol.P /:
Then
Vol.Q/ " < Vol.P / D .P / .int.Q//:
Thus Vol.Q/ " < .int.Q// for all " > 0. Therefore, by axiom (b),
Vol.Q/ .int.Q// .Q/ D Vol.Q/;
and hence .int.Q// D Vol.Q/. This proves part (iv) and Theorem 2.10.
Rn
Definition 2.11. Let W 2 ! Œ0; 1 be the Lebesgue outer measure. A subset
A Rn is called Lebesgue measurable if A is -measurable, i.e.,
.D/ D .D \ A/ C .D n A/ for all D Rn :
The set of all Lebesgue measurable subsets of Rn will be denoted by
A WD ¹A Rn j A is Lebesgue measurableº:
The function
m WD jA W A ! Œ0; 1
is called the Lebesgue measure on Rn . A function f W 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 2 A and x 2 Rn , then
A C x 2 A and m.A C x/ D m.A/:
(iii) Every Borel set in Rn is Lebesgue measurable.

(iv) If Q 2 Qn , then
Q; int.Q/ 2 A; and m.int.Q// D m.Q/ D Vol.Q/:
Proof. Assertion (i) follows from Theorem 2.4 and part (i) of Theorem 2.10. Asser-
tion (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. Finally, assertion (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). The Lebesgue outer
n
measure W 2R ! Œ0; 1 satisfies the following.
(i) For every subset A Rn ,
.A/ D inf¹.U / j A U Rn and U is openº:
(ii) If A Rn is Lebesgue measurable, then
.A/ D sup¹.K/ j K A and K is compactº:
(iii) If Ai is a sequence of subsets of Rn such that Ai Ai C1 for all i 2 N,

then their union
[1
A WD Ai
i D1
has Lebesgue outer measure
.A/ D lim .Ai /:

i!1
Proof. (i) Fix a subset A Rn and a constant " > 0. The assertion is obvious when
.A/ D 1. Hence assume .A/ < 1 and choose a sequence of closed cuboids
Qi 2 Qn such that
1 1
[ X "
A Qi ; Vol.Qi / < .A/ C :
2
iD1 iD1
Now choose a sequence of open cuboids Ui Rn such that

"
Q i Ui ; Vol.Ui / < Vol.Qi / C :
2i C1
S1
Then U WD i D1 Ui is an open subset of Rn containing A and
1 1 1
X X X "
.U / .Ui / D Vol.Ui / < Vol.Qi / C < .A/ C ":
2iC1
iD1 iD1 iD1
(ii) Assume first that A Rn is Lebesgue measurable and bounded. Choose

r > 0 so large that A Br WD ¹x 2 Rn j jxj < rº: Fix a constant " > 0. By (i),
there exists an open set U Rn such that Bxr n A U and .U / .Bxr n A/ C ".
Hence K WD Bxr n U is a compact subset of A and
.K/ D .Bxr / .U / .Bxr / .Bxr n A/ " D .A/ ":
Here the first equality uses the fact that K and U are disjoint Lebesgue measurable
sets with union Bxr and the last equality uses the fact that A and Bxr n A are disjoint
Lebesgue measurable sets with union Bxr . This proves (ii) for bounded Lebesgue
measurable sets. If A 2 A is unbounded, then
.A/ D sup .A \ Bxr /
r>0
D sup sup¹.K/ j K .A \ Bxr / and K is compactº

r>0
D sup¹.K/ j K A and K is compactº:
(iii) If .Ai / D 1 for some i , then the assertion is obvious. Hence assume
.Ai / < 1 for all i and fix a constant " > 0. By (i), there is a sequence of open
sets Ui Rn such that .Ui / < .Ai / C "2 i for all i. Since Ai Ui \ UiC1 ,
this yields
.UiC1 n Ui / D .UiC1 / .Ui \ Ui C1 /
i 1
< .AiC1 / .Ai / C "2
for all i 2 N. It follows that
[ k k
X1
Ui D .U1 / C .UiC1 n Ui / < .Ak / C ":
i D1 iD1
Take the limit k ! 1 to obtain

[1
Ui lim .Ak / C ":
k!1
iD1
Thus
.A/ lim .Ak / C "; for all " > 0,
k!1
and so
.A/ lim .Ak /:
k!1
The converse inequality is obvious. This proves Theorem 2.13.
n
Theorem 2.14 (the Lebesgue measure as a completion). Let W 2R ! Œ0; 1 be
the Lebesgue outer measure in Definition 2.9, let
m D jA W A ! Œ0; 1
be the Lebesgue measure, let B A be the Borel -algebra of Rn , and define
WD jB W B ! Œ0; 1:
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 2 A, i.e.,
.D/ D .D \ A/ C .D n A/ for all D Rn .
(II) A 2 B , i.e., there exist Borel measurable sets B0 ; B1 2 B such that
B0 A B1 and .B1 n B0 / D 0:
If the set A satisfies both (I) and (II), then
.A/ .B1 / D .B0 / C .B1 n B0 / D .B0 / .A/;
and hence
m.A/ D .A/ D .B0 / D .A/:
This shows that A D B and m D . Thus it remains to prove the claim. Fix a
subset A Rn .
We show that (II) implies (I). Thus assume that A 2 B and choose Borel mea-
surable sets B0 ; B1 2 B such that
B0 A B1 ; .B1 n B0 / D 0:
Then .A n B0 / .B1 n B0 / D 0 and hence .A n B0 / D 0. Since is an outer

measure, by part (i) of Theorem 2.10, it follows from Theorem 2.4 that A n B0 2 A
and hence A D B0 [ .A n B0 / 2 A.
We now show that (I) implies (II). Thus assume that A 2 A. Suppose first that
.A/ < 1. By Theorem 2.13, there exist 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/ C :
i i
Define
1
[ 1
\
B0 WD Ki ; B1 WD Ui :
i D1 i D1
These are Borel sets satisfying B0 A B1 and
1 1
.A/ .Ki / .B0 / .B1 / .Ui / .A/ C :
i i
Take the limit i ! 1 to obtain
.A/ .B0 / .B1 / .A/;
hence
.B0 / D .B1 / D .A/ < 1;
and hence
.B1 n B0 / D .B1 / .B0 / D 0:
This shows that A 2 B for every A 2 A with .A/ < 1.
Now suppose that our set A 2 A satisfies .A/ D 1 and define
Ak WD ¹x 2 A j jxi j k for i D 1; : : : ; nº for k 2 N:
Then Ak 2 A and .Ak / .2k/n for all k. Hence Ak 2 B for all k and so there
exist sequences of Borel sets Bk ; Bk0 2 B such that
Bk Ak Bk0 and .Bk0 n Bk / D 0:
Define
1
[ 1
[
B WD Bk and B 0 WD Bk0 :
kD1 kD1
Then B; B 0 2 B, B A B 0 , and
1
X 1
X
0
.B n B/ .Bk0 n B/ .Bk0 n Bk / D 0:
kD1 kD1
This shows that A 2 B for every A 2 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 W B ! Œ0; 1 is a translation-invariant measure such that 0 .Œ0; 1n / D 1.
Define WD 0 .Œ0; 1/n /. Then 0 1. We show in five steps that D 1 and
0 D .
Step 1. For x D .x1 ; : : : ; xn / and k 2 N0 WD N [ ¹0º define
k k
R.x; k/ WD Œx1 ; x1 C 2 / Œxn ; xn C 2 /:
Then 0 .R.x; k// D 2 nk
D .R.x; k//.
Fix an integer k 2 N0 . Since R.x; k/ D R.0; k/ C x for every x 2 Rn , it follows
from the translation invariance of 0 that there is a constant ck 0 such that
0 .R.x; k// D ck for all x 2 Rn :
Since R.x; 0/ can be expressed as the disjoint union
[
k
R.x; 0/ D R.x C 2 `; k/;
`2Zn ; 0`j 2k 1
this implies
X
D 0 .R.x; 0// D 0 .R.x C 2 k
`; k// D 2nk ck :
`2Zn ; 0`j 2k 1
Hence ck D 2 nk D .R.x; k//. Here the last equation follows from the fact
that .0; 2 k /n R.0; k/ Œ0; 2 k n and hence .R.x; k// D .R.0; k// D 2 k
by part (iv) of Corollary 2.12.
Step 2. 0 .U / D .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 D R.xi ; ki / as in Step 1. To see this, define
R0 WD ¹R.x; 0/ j x 2 Zn ; R.x; 0/ U º;
ˇ x 2 2 1 Zn ; R.x; 1/ U;
² ˇ ³
R1 WD R.x; 1/ ˇ ;
R.x; 1/ 6 R, for all R 2 R0
ˇ
ˇ x 2 2 k Zn ; R.x; k/ U;
² ˇ ³
Rk WD R.x; k/ ˇˇ for k 2,
R.x; k/ 6 R, for all R 2 R0 [ R1 [ [ Rk 1
and denote
1
[
R WD Rk :
kD0
R and 0 .R/ D .R/

S
Then U can be expressed as the disjoint union U D R2R
for all R 2 R by Step 1. Hence
X X
0 .U / D 0 .R/ D .R/ D .U /:
R2R R2R
Step 3. 0 .K/ D .K/ for every compact set K Rn .
Let K Rn be compact. Choose r > 0 so large that K U WD . r; r/n : Then U

and U n K are open. Hence, by Step 2,
0 .K/ D 0 .U / 0 .U n K/ D .U / .U n K/ D .K/:
Step 4. 0 .B/ D .B/ for every Borel set B 2 B.
Let B 2 B. It follows from Step 2, Step 3, and Theorem 2.13 that
0 .B/ inf¹0 .U / j B U Rn and U is openº

D inf¹.U / j B U Rn and U is openº
D .B/
D sup¹.K/ j K B and K is compactº
D sup¹0 .K/ j K B and K is compactº
0 .B/:
Step 5. D 1 and 0 D .
By Step 4, we have D .Œ0; 1n / D 0 .Œ0; 1n / D 1 and hence 0 D D .

This proves Step 5 and Theorem 2.1.
We have given two definitions of the Lebesgue measure mW A ! Œ0; 1:

The first one, in Definition 2.2, uses the existence and uniqueness of a normalized
translation-invariant Borel measure W B ! Œ0; 1; established in Theorem 2.1,
and then defines .Rn ; A; m/ as the completion of that measure. The second one,
n
in Definition 2.11, uses the Lebesgue outer measure W 2R ! Œ0; 1 of Defini-
tion 2.9 and Theorem 2.10, and defines the Lebesgue measure as the restriction of
to the -algebra of -measurable subsets of Rn (see Theorem 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
x y () x y2Q
for x; y 2 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 2 R there exists a rational number q 2 Q such that x q 2 E;

(II) if x; y 2 E and x ¤ y, then x y … Q.
For q 2 Q define the set
Bq WD A \ .E C q/ D ¹x 2 A j x q 2 Eº:
S
Then it follows from (I) that A D q2Q Bq :
Fix a rational number q 2 Q. We prove that if Bq is Lebesgue measurable, then
m.Bq / D 0. Indeed, let n 2 N, and define
Bq;q 0 ;n WD .Bq \ Œ n; n/ C q 0 D ¹x C q 0 j x 2 Bq ; jxj nº for q 0 2 Q:
This set is Lebesgue measurable, its Lebesgue measure is independent of q 0 , and

Bq;q 0 ;n \ Bq;q 00 ;n D ; for all q 0 ; q 00 2 Q with q 0 ¤ q 00 by condition (II). Since
Bq;q 0 ;n Œ n; nC1 for q 0 2 Œ0; 1\Q, we have q 0 2Œ0;1\Q m.Bq;q 0 ;n / 2n C 1:
P
This sum is infinite and all summands agree, so m.Bq \ Œ n; n/ D 0. This holds
for all n 2 N, and hence m.Bq / D 0, as claimed.
S
If Bq is Lebesgue measurable for all q 2 Q, it follows that A D q2Q Bq
is a Lebesgue null set, a contradiction. Thus one of the sets Bq is not Lebesgue
measurable and this proves Lemma 2.15.
Remark 2.16. (i) Using Lemma 2.15 one can construct a continuous function
f W R ! R and a Lebesgue measurable function gW R ! R such that the com-
position g ı 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 \ Œ n; n/ C q/ Œ0; 1. On the other hand,
the set E ¹0º 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 W x 7! .x; 0/ is the original set E and hence is not a Borel set.
2.3. The transformation formula 77
2.3 The transformation formula

The transformation formula describes how the integral of a Lebesgue measurable
function transforms under composition with a C 1 diffeomorphism. Fix a positive
integer n 2 N and denote by .Rn ; A; m/ the Lebesgue measure space. For any
Lebesgue measurable set X Rn denote by
AX WD ¹A 2 A j A X º
the restricted Lebesgue -algebra and by
mX WD mjAX W AX ! Œ0; 1
the restriction of the Lebesgue measure to AX .
Theorem 2.17 (transformation formula). Suppose W U ! V is a C 1 diffeomor-

phism between open subsets of Rn .
(i) If f W V ! Œ0; 1 is Lebesgue measurable, then f ı W U ! Œ0; 1 is

Lebesgue measurable and
Z Z
.f ı /jdet.d/j d m D f d m: (2.14)
U V
(ii) If E 2 AU and f 2 L1 .mV /, then .E/ 2 AV , .f ı /jdet.d/j 2 L1 .mU /,

and Z Z
.f ı /jdet.d/j d m D f d m: (2.15)
E .E /
Proof. See page 82.
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 formula (2.14) for the characteristic functions of open sets.
Lemma 2.18. Let ˆW Rn ! Rn be a linear transformation and let A Rn be a

Lebesgue measurable set. Then ˆ.A/ is a Lebesgue measurable set and
m.ˆ.A// D jdet.ˆ/jm.A/: (2.16)

Proof. If det.ˆ/ D 0, then ˆ.A/ is contained in a proper linear subspace of Rn

and hence is a Lebesgue null set for every A 2 A. In this case both sides of equa-
tion (2.16) vanish. Hence it suffices to assume that ˆ is a vector space isomor-
phism. For vector space isomorphisms we prove the assertion in six steps. Denote
n
by B 2R the Borel -algebra and by
WD mjB
the restriction of the Lebesgue measure to the Borel -algebra. Thus is the unique
translation-invariant Borel measure on Rn that satisfies the normalization condi-
tion .Œ0; 1n / D 1 (Theorem 2.1) and .Rn ; A; m/ is the completion of .Rn ; B; /
(Theorem 2.14).
Step 1. There exists a unique map
W GL.n; R/ ! .0; 1/
such that
.ˆ.B// D .ˆ/.B/ (2.17)
for every ˆ 2 GL.n; R/ and every Borel set B 2 B.
Fix a vector space isomorphism ˆW Rn ! Rn . Since ˆ is a homeomorphism of Rn
with its standard topology, it follows that ˆ.B/ 2 B for every B 2 B. Define the
number .ˆ/ 2 Œ0; 1 by
.ˆ/ WD .ˆ.Œ0; 1/n //: (2.18)
Since ˆ.Œ0; 1/n / has nonempty interior, it follows that .ˆ/ > 0 and since ˆ.Œ0; 1/n /
is contained in the compact set ˆ.Œ0; 1n /, it follows that .ˆ/ < 1. Now define
the map ˆ W B ! Œ0; 1 by
.ˆ.B//
ˆ .B/ WD for B 2 B:
.ˆ/
Then ˆ is a normalized translation-invariant Borel measure. The -additivity
follows directly from the -additivity of , the formula ˆ .;/ D 0 is obvious
from the definition, that compact sets have finite measure follows from the fact
that ˆ.K/ is compact if and only if K Rn is compact, the translation invari-
ance follows immediately from the translation invariance of and the fact that
ˆ.B C x/ D ˆ.B/ C ˆ.x/ for all B 2 B and all x 2 Rn , and the normaliza-
tion condition ˆ .Œ0; 1/n / D 1 follows directly from the definition of ˆ . Hence
ˆ D by Theorem 2.1.
Step 2. Let be as in Step 1 and let A 2 A and ˆ 2 GL.n; R/. Then ˆ.A/ 2 A
and m.ˆ.A// D .ˆ/m.A/.
By Theorem 2.14, there exist Borel sets B0 ; B1 2 B such that B0 A B1 and
.B1 n B0 / D 0. Then ˆ.B0 / ˆ.A/ ˆ.B1 / and, by Step 1,
.ˆ.B1 / n ˆ.B0 // D .ˆ.B1 n B0 // D .ˆ/.B1 n B0 / D 0:
Hence ˆ.A/ is a Lebesgue measurable set and
m.ˆ.A// D .ˆ.B0 // D .ˆ/.B0 / D .ˆ/m.A/
by Theorem 2.14 and Step 1.
Step 3. Let be as in Step 1 and let ˆ D diag.1 ; : : : ; n / be a diagonal matrix
with nonzero diagonal entries i 2 R n ¹0º. Then .ˆ/ D j1 j jn j.
Define I WD Œ 1; 1 and Ii WD Œ ji j; ji j for i D 1; : : : ; n. Then Q WD I n has
Lebesgue measure m.Q/ D 2n and the cuboid ˆ.Q/ D I1 In has Lebesgue
measure m.ˆ.Q// D 2n j1 j jn j by part (iv) of Corollary 2.12. Hence Step 3
follows from Step 2.
Step 4. The map W GL.n; R/ ! .0; 1/ 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 ˆ; ‰ 2 GL.n; R/. Then it follows from (2.17) with B WD ‰.Œ0; 1/n / and from
the definition of .‰/ in (2.18) that
.ˆ‰/ D .ˆ‰.Œ0; 1/n // D .ˆ/.‰.Œ0; 1/n // D .ˆ/.‰/:
Thus is a group homomorphism as claimed and this proves Step 4.
Step 5. The map W GL.n; R/ ! .0; 1/ in Step 1 is continuous with respect to the
standard topologies on GL.n; R/ and .0; 1/.
It suffices to prove continuity at the identity. Define the norms
kˆxk1
kxk1 WD max jxi j ; kˆk1 WD sup (2.19)
iD1;:::;n 0¤x2Rn kxk1
for x 2 Rn and a linear map ˆW Rn ! Rn . Denote the closed unit ball in Rn by
Q WD ¹x 2 Rn j kxk1 1º D Œ 1; 1n . Fix a constant 0 < ı < 1 and a linear
map ˆW Rn ! Rn such that kˆ 1k1 < ı. Then ˆ 2 GL.n; R/ and
1
1
X 1
ˆ D .1 ˆ/k ; kˆ 1
k1 < :
1 ı
kD0
1
Thus ˆ.Q/ .1 C ı/Q and .1 ı/ˆ .Q/ Q. Hence
.1 ı/Q ˆ.Q/ .1 C ı/Q:
Since .ˆ/ D m.ˆ.Q//=m.Q/ by Step 2 and m.rQ/ D r n m.Q/ for r > 0 by

Steps 2 and 3, this shows that .1 ı/n .ˆ/ .1 C ı/n . Given " > 0, choose a
constant 0 < ı < 1 such that 1 " < .1 ı/n < .1 C ı/n < 1 C ". Then
kˆ 1k1 < ı H) j.ˆ/ 1j1 < "
for all ˆ 2 GL.n; R/.
Step 6. .ˆ/ D jdet.ˆ/j for all ˆ 2 GL.n; R/.
If ˆ 2 GL.n; R/ is diagonalizable with real eigenvalues, then .ˆ/ D jdet.ˆ/j

by Step 3 and Step 4. If ˆ 2 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 .ˆ/ D jdet.ˆ/j. 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

d1 W Rn Rn ! Œ0; 1/
by
d1 .x; y/ WD kx yk1 for x; y 2 Rn ,
where kk1 is as in (2.19). The open ball of radius r > 0 about a point
a D .a1 ; : : : ; an / 2 Rn with respect to this metric is the open cube
Br .a/ WD .a1 r; a1 C r/ .an r; an C r/
and its closure is Bxr .a/ D Œa1 r; a1 C r Œan r; an C r:
Lemma 2.19. Let U Rn be an open set and let K U be a compact subset. Let
W U ! Rn be a continuously differentiable map such that det.d.x// ¤ 0 for all
x 2 K. For every " > 0 there exists a constant ı > 0 such that the following holds:
if 0 < r < ı, a 2 Rn , and R Rn satisfy Br .a/ R Bxr .a/ K, then
jm..R// jdet.d.a//j m.R/j < " m.R/: (2.20)

Proof. The maps K ! RW x 7! kd.x/ 1 k1 and K ! RW x 7! jdet.d.x//j are

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

d.x/
1
c; jdet.d.x//j c for all x 2 K: (2.21)
Let " > 0 and choose a constant 0 < ˛ < 1 so small that
" "
1 < .1 ˛/n < .1 C ˛/n < 1 C : (2.22)
c c
Choose ı > 0 so small that, for all x; y 2 Rn ,
˛
x; y 2 K; kx yk1 < ı H) kd.x/ d.y/k1 < : (2.23)
c
Such a constant exists because the map dW U ! Rnn is uniformly continuous on
the compact set K U . We prove that the assertion of Lemma 2.19 holds with this
constant ı.
Choose a 2 Rn and 0 < r < ı such that Bxr .a/ K. Then ka xk1 < ı for
all x 2 Bxr .a/. By (2.23) with ˆ WD d.a/ this implies
˛ ˛
kd.x/ ˆk1 < for all x 2 Bxr .a/:
c kˆ 1 k1
Here the first step follows from (2.23) and the second step follows from (2.21).
Define the map
W U ! Rn
by
1
.x/ WD ˆ ..x/ .a//:
1
Then .a/ D 0 and d .x/ D ˆ d.x/ and hence, by (2.23),
1 1

kd .x/ 1k1 D ˆ .d.x/ ˆ/ 1 ˆ 1
kd.x/ ˆk1 ˛
for all x 2 Bxr .a/. By Theorem C.1, this implies
B.1 ˛/s .0/ .Br .a// .Bxr .a// Bx.1C˛/s .0/ (2.24)
Now fix a subset R Rn such that Br .a/ R Bxr .a/. Then by (2.24)
.1 ˛/ˆ.Br .0// .R/ .a/ .1 C ˛/ˆ.Bxr .0//:

Since m.R/ D m.Br .0// D m.Bxr .0// by part (iv) of Corollary 2.12, it follows
from Lemma 2.18 and the inequalities (2.21) and (2.22) that
"
jdet.ˆ/j m.R/ " m.R/ 1 jdet.ˆ/j m.R/
c
< .1 ˛/n jdet.ˆ/j m.R/
D m..1 ˛/ˆ.Br .0///
m..R//
m..1 C ˛/ˆ.Bxr .0///
D .1 C ˛/n jdet.ˆ/j m.R/
"
< 1C jdet.ˆ/j m.R/
c
jdet.ˆ/j m.R/ C " m.R/:
This proves (2.20) and Lemma 2.19.
Proof of Theorem 2.17. The proof has seven steps. The first four steps establish
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 WS U , then

Z
m..W // D jdet.d/j d m:
W
Fix a constant " > 0. Then there exists a constant ı > 0 that satisfies the following
two conditions.
(a) If a 2 Rn , 0 < r < ı, R Rn satisfy Br .a/ R Bxr .a/ W

S , then
" m.R/
jm..R// jdet.d.a//j m.R/j < :
2 m.W /
S
(b) For all x; y 2 W
"
kx yk1 < ı H) jdet.d.x// det.d.y//j < :
2 m.W /
That ı > 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/W U ! R is uniformly continuous on the compact set W S . Now write W as a
countable union of pairwise disjoint half-open cubes Ri Rn centered at ai 2 Rn
with side lengths 2ri such that 0 < ri < ı. (See page 74.) Then
Bri .ai / Ri Bxri .ai / W
S for all i
and X X
m.W / D m.Ri /; m..W // D m..Ri //: (2.25)
i i
It follows from (2.25) and (a) that
ˇ X ˇ "
m..W // jdet.d.a //j m.Ri ˇ< :
/
ˇ ˇ
i (2.26)
2
ˇ
i
It follows from (b) that

ˇ X ˇ "
ˇ j det.d/j jdet.d.ai //j R ˇ < on W .
ˇ ˇ
i 2 m.W /
i
Integrate this inequality over W to obtain

ˇZ ˇ
ˇ X ˇ "
ˇ jdet.d/j d m jdet.d.ai //j m.Ri /ˇˇ < : (2.27)
ˇ
W 2
i
By (2.26) and (2.27), we have

ˇ Z ˇ
ˇ ˇ
ˇm..W //
ˇ jdet.d/j d mˇˇ < ":
W
Since this holds for all " > 0, Step 1 follows.

Step 2. K U is compact, then
Z
m..K// D jdet.d/j d m:
K
S U . Then
Choose an open set W K with compact closure W
m..K// D m..W // m..W n K//
Z Z
D jdet.d/j d m jdet.d/j d m
W W nK
Z
D jdet.d/j d m:
K
Here the second equality follows from Step 1.

Step 3. If B 2 B has compact closure Bx U , then .B/ 2 B and

Z
m..B// D jdet.d/j d m:
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 W V ! U (Theorem 1.20). Abbreviate
b WD m..B//:
Assume first that b < 1 and fix a constant " > 0. Then it follows from Theo-
rem 2.13 that there exist an open set W 0 Rn with compact closure W 0 V
such that .B/ W 0 and m.W 0 / b ". Define K WD 1 .K 0 / and W WD 1 .W 0 /. Then K is compact,
W is open, WS U is compact, and
K B W; b " < m..K// m..W // 0,
B
and so Z
jdet.d/j d m D b D m..B//:
B
If b D 1, then by Theorem 2.13, there exists a sequence of compact sets
Ki0 .B/ such that .Ki0 / > i. Hence
1
Ki WD .Ki0 /
is compact and Z
jdet.d/j d m D ..Ki // > i
Ki
by Step 2. Since Ki B, this implies

Z
jdet.d/j d m > i for all i 2 N,
B
and hence Z
jdet.d/j d m D 1 D m..B//:
B
Step 4. If A 2 A has compact closure Ax U , then .A/ 2 A and

Z
m..A// D jdet.d/j d m:
A
Let A 2 A. By Theorem 2.14, there exist Borel sets B0 ; B1 2 B, with compact

closure contained in U , such that B0 A B1 and m.B1 n B0 / D 0. Then
.B0 / .A/ .B1 / and it follows from Step 3 that .B0 / and .B1 / are
Borel sets and
Z
m..B1 / n .B0 // D m..B1 n B0 // D jdet.d/j d m D 0:
B1 nB0
Now Theorem 2.14 shows that .A/ is a Lebesgue measurable set and
Z Z
m..A// D m..B0 // D jdet.d/j d m D jdet.d/j d m:
B0 A
Here the last equation follows from the fact that the set A n B0 is Lebesgue measur-
able and has Lebesgue measure zero.
Step 5. Assertion (i) of Theorem 2.17 holds for every Lebesgue measurable step
function f D sW V ! R whose support is a compact subset of V .
Write
`
X
sD ˛i A
i
iD1
with ˛i 2 R and Ai 2 A such that Ai is a compact subset of V for all i . Then

1 .Ai / is a Lebesgue measurable set with compact closure in U by Step 4. Hence
s ı D ìD1 ˛i 1 .A / is a Lebesgue measurable step function and
P
i
Z `
X Z
.s ı /jdet.d/j d m D ˛i jdet.d/j d m
U 1 .A
iD1 i/
`
X
D ˛i m.Ai /
iD1
Z
D s d m:
V
Here the second equality follows from Step 4.

Step 6. We prove (i).
By Theorem 1.26, there is a sequence of Lebesgue measurable step functions
si W V ! Œ0; 1/
such that
0 s1 s2
and
f .x/ D lim si .x/ for every x 2 V .
i!1
Choose an exhaustive sequence of compact sets Ki V such that Ki KiC1 for

S
all i and i Ki D V , and replace si by si K . Then part (i) follows from Step 5
i
and the Lebesgue monotone convergence theorem (Theorem 1.37).
Step 7. We prove (ii).
For E D U part (ii) follows from part (i) and the fact that .f ı /˙ D f ˙ ı .
If F 2 AV , then f WD F jV is Lebesgue measurable, hence f ı D 1 .F / jU
is Lebesgue measurable by part (i), and so 1 .F / 2 AU . Replace by 1 to
deduce that if E 2 AU , then .E/ 2 AV . Then (ii) follows for all E 2 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 function on
Rn agrees with the Riemann integral whenever the latter is defined and the func-
tion 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
k n
R.x; k/ WD x C Œ0; 2 / for x 2 Rn and k 2 N,
which was used in the proof of Theorem 2.1 on page 74. The closure of R.x; k/ is
the closed cube R.x; k/ D x C Œ0; 2 k n : The sets R.`; k/, with ` ranging over the
countable set 2 k Zn , form a partition of the Euclidean space Rn .
2.4. Lebesgue equals Riemann 87
Definition 2.20. Let f W Rn ! R be a bounded function whose support
supp.f / WD ¹x 2 Rn j f .x/ ¤ 0º
is a bounded subset of Rn . For k 2 N define the lower sum S.f; k/ 2 R and the
x k/ 2 R by
upper sum S.f; x
X X
S.f; k/ WD x k/ WD
inf f 2 nk ; S.f; sup f 2 nk : (2.28)
x k n R.`;k/ k n R.`;k/
`22 Z `22 Z
These are finite sums and satisfy

x k/:
sup S.f; k/ inf S.f;
k x k
The function f W Rn ! R is called Riemann integrable if

x k/:
sup S.f; k/ D inf S.f;
k x k
The Riemann integral of a Riemann integrable function f W Rn ! R is the real

number
Z
R.f / WD x k/ D lim S.f; k/: (2.29)
f .x/ dx WD sup S.f; k/ D inf S.f;
Rn k2N x k2N k!1 x
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 is called Jordan measurable if its char-
acteristic function A W Rn ! R is Riemann integrable. The Jordan measure of a
Jordan measurable set A Rn is the real number
J .A/ WD R.A /
Z
D A .x/ dx (2.30)
Rn
nk k
D lim 2 #¹` 2 2 Zn j R.`; k/ \ A ¤ ;º:
k!1
Exercise 2.23. (i) Prove the last equality in (2.30).

(ii) Prove that a bounded set A Rn is Jordan measurable if and only if its
boundary @A D Ax n 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 W Rn ! R is Riemann integrable, then f 2 L1 .m/ and its
Lebesgue integral agrees with the Riemann integral, i.e.,
Z
f d m D R.f /:
Rn
(ii) If A Rn is Jordan measurable, then A is Lebesgue measurable and
m.A/ D J .A/:
Proof. Assertion (ii) follows from (i) by taking f D A . Thus it remains to

prove (i). Let f W Rn ! R be a Riemann integrable function. Then f is bounded
and has bounded support. Define the functions
f k ; fNk W Rn ! R
N
by
f k .x/ WD inf f; fNk .x/ WD sup f for x 2 R.`; k/; ` 2 2 k

Zn : (2.31)
N R.`;k/ R.`;k/
These are Lebesgue measurable step functions and

Z Z
f k d m D S.f; k/; fNk d m D S.f;
x k/:
Rn N x Rn
They also satisfy

f k f kC1 f fNkC1 fNk
N N
for all k 2 N. Define the functions
f ; fNW Rn ! R
N
by
f .x/ WD lim f k .x/; fN.x/ WD lim fNk .x/ for x 2 Rn :
N k!1 N k!1
Then
f .x/ f .x/ fN.x/
N
for every x 2 Rn . Moreover, jfNk j and jf k j are bounded above by the Lebesgue
integrable function cA , where N
c WD sup jf .x/j
x2Rn
2.4. Lebesgue equals Riemann 89
and
A WD Œ N; N n ;
with N 2 N chosen such that supp.f / Œ N; N n . Hence it follows from
the Lebesgue dominated convergence theorem (Theorem 1.45) that f and fN are
Lebesgue integrable and N
Z Z
f d m D lim f dm
Rn N k!1 Rn N k
D lim S.f; k/
k!1 x
D R.f /
x k/
D lim S.f;
k!1
Z
D lim fNk d m
k!1 Rn
Z
D fN d m:
Rn
By Lemma 1.49, with f replaced by fN f , this implies that f D f D fN Lebesgue

N
almost everywhere. Hence f 2 L1 .m/ and N
Z Z
f dm D f d m D R.f /:
Rn Rn N
The discussion in this section is restricted to Riemann integrable functions
f W Rn ! R with compact support, and Theorem 2.24 asserts that for such func-
tions the Riemann integral agrees with the Lebesgue integral. When f does not
have compact support and is locally Riemann integrable, the improper Riemann
integral is defined by
Z Z
f .x/dx WD lim f .x/ dx; (2.32)
Rn r!1 B
r
provided that the limit exists. Here Br Rn denotes the ball of radius r centered
at the origin. There are
R many examples where the limit (2.32) exists even though
the Lebesgue integral Rn jf j d m is infinite and so the Lebesgue integral of f does
not exist. An example is the function f W R ! R given by f .x/ WD x 1 sin.x/
for x 2 R n ¹0º and f .0/ WD 1. This function is continuous and is not Lebesgue
integrable, but the improper Riemann integral exists and is equal to (see Exam-
ple 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\Œ0; 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/ D 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 C B WD ¹x C y j x 2 A; y 2 Bº
is a Lebesgue null set.
Exercise 2.28. Let .X; A; / be a measure space and define the function
W 2X ! Œ0; 1
by
.B/ WD inf¹.A/ j A 2 A; B Aº: (2.33)
(i) Prove that is an outer measure and that A A./.
(ii) Assume .X/ < 1. Prove that the measure space .X; A./; jA./ / is the
completion of .X; A; /. Hint. Show that for every subset B X there exists
a set A 2 A such that B A and .B/ D .A/.
(iii) Let X be a set and A ¨ X be a nonempty subset. Define
A WD ¹;; A; Ac ; Xº; .;/ WD .A/ WD 0; .Ac / WD .X/ WD 1:
Prove that .X; A; / is a measure space. Given B X , prove that .B/ D 0
whenever B A and .B/ D 1 whenever B 6 A. Prove that A./ D 2X
and that the completion of .X; A; / is the measure space .X; A ; / with
A D ¹B X j B A or Ac Bº and D jA . (Thus the hypothesis
.X/ < 1 cannot be removed in part (ii).)
Exercise 2.29. Let f W R ! R be continuously differentiable and define
A WD ¹x 2 R j f 0 .x/ D 0º:
Prove that f .A/ is a Lebesgue null set. Hint. Consider the sets
An;" WD ¹x 2 R j jxj < n; jf 0 .x/j < 2 n
"º:
2.5. Exercises 91
Exercise 2.30. Find a continuous function f W Œ0; 1/ ! R such that f is not

RT
Lebesgue integrable, but the limit lim 0 f .t/ dt exists.
T !1
Exercise 2.31. Determine the limits of the sequences

n n
x n x=2 x n
Z Z
2x
an WD 1 e dx; bn WD 1C e dx; n 2 N:
0 n 0 n
Hint. Use the Lebesgue dominated convergence theorem (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 C e t / c C t for all t 0:
Does the limit

1
1
Z
lim log.1 C e nf .x/ / dx
n!1 n 0
exist for every Lebesgue integrable function f W Œ0; 1 ! R? Determine the limit
when it does exist.
Exercise 2.34. Let .Rn ; A; m/ be the Lebesgue measure space and let W Rn ! Rn
be a C 1 -diffeomorphism. Prove that A D A and that
1
Z
. m/.A/ D d m for all A 2 A:
A jdet.d/ ı 1j
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/ WD sup .x; y/: (2.34)

x;y2A
For " > 0 define the function
d;" W 2X ! Œ0; 1
by
8 ˇ 9
<X ˇ I is finite or countably infinite; =
diam.Di /d ˇˇ Di X; diam.Di / < " for i 2 I; :
ˇ
d;" .A/ WD inf (2.35)
:
i2I ˇ and A S Di ;
i2I
for A X. Thus d;" .;/ D 0 and d;" .A/ D 1 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 Haus-
dorff outer measure is the function
d W 2X ! Œ0; 1
defined by
d .A/ WD sup d;" .A/ D 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/ WD inf¹.x; y/ j x 2 A; y 2 Bº > 0;
then
d .A [ B/ D d .A/ C d .B/:
Hence, by Theorems 2.4 and 2.5, the set
Ad WD ¹A X j A is d -measurableº
is a -algebra containing the Borel sets and
d WD d jAd W Ad ! Œ0; 1
is a measure. It is called the d -dimensional Hausdorff measure on X . Haus-

dorff measures play a central role in geometric measure theory.
(iii) If d D 0, then A0 D 2X and 0 D 0 is the counting measure.
(iv) The n-dimensional Hausdorff measure on Rn agrees with the Lebesgue mea-
sure up to a factor (the Lebesgue measure of the ball of radius 1=2).
2.5. Exercises 93
(v) Let A X be nonempty. The Hausdorff dimension of A is the number
dim.A/ WD sup¹r 0 j r .A/ D 1º D inf¹s 0 j s .A/ D 0º: (2.37)
The second equality follows from the fact that d .A/ > 0 implies r .A/ D 1
for 0 r < d , and d .A/ < 1 implies s .A/ D 0 for s > d .
(vi) The Hausdorff dimension of a smooth embedded curve Rn is d 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 D log.2/= log.3/.

Chapter 3
Borel measures
The regularity properties established for the Lebesgue (outer) measure in Theo-
rem 2.13 play an important role in much greater generality. The present chapter is
devoted to the study of Borel measures on locally compact Hausdorff spaces that
satisfy similar regularity properties. The main result is the Riesz representation the-
orem (Theorem 3.15). We begin with some further recollections about topological
spaces.
Let .X; U/ be a topological space (see Definition 1.9). A neighborhood of a
point x 2 X is a subset A X that contains x in its interior, i.e., x 2 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 2 X with x ¤ y there exist open sets
U; V X such that x 2 U , y 2 V , and U \ V D ;. X is called locally compact if
every point in X has a compact neighborhood. It is called -compact if there exists
a sequence of compact sets Ki X, i 2 N, such that Ki Ki C1 for all i and
XD 1
S
iD1 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 con-
tains 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 (Exer-
cise 2.35), where one allows for compact sets to have infinite measure, but these are
not discussed here.
Definition 3.1. A measure W B ! Œ0; 1 is called a Borel measure if
.K/ < 1
for every compact set K X. A measure W B ! Œ0; 1 is called outer regular if
.B/ D inf¹.U / j B U X and U is openº (3.1)

96 3. Borel measures
for every Borel set B 2 B, is called inner regular if
.B/ D sup¹.K/ j K B and K is compactº (3.2)
for every Borel set B 2 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 D Rn to the Borel
-algebra is a regular Borel measure by Theorem 2.13.
Example 3.3. The counting measure on X D N with the discrete topology
U D B D 2N is a regular Borel measure.
Example 3.4. Let .X; U/ be any locally compact Hausdorff space and fix a point
x0 2 X . Then the Dirac measure D ıx0 at x0 is a regular Borel measure (Exam-
ple 1.31).
Example 3.5. Let X be an uncountable set equipped with the discrete topology
U D B D 2X . Define
W B ! Œ0; 1
by
´
0 if B is countable;
.B/ WD
1 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/ D 1 and .K/ D 0 for every compact set K X . Thus is
not a Radon measure.
The next example occupies four 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 2 X such that the open set
U WD X n ¹º
is not -compact and satisfies .U / D 1 and .K/ D 0 for every compact sub-
set K U . This example is a kind of refinement of Example 3.5. It is due to
Dieoudonné.
Example 3.6 (Dieudonné’s measure). (i) Let .X; 4/ be an uncountable well-ordered
set with a maximal element 2 X such that every element x 2 X n ¹º has only
countably many predecessors. Here a set is called countable if it is finite or count-
ably 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.
3.1. Regular Borel measures 97
(a) If x; y; z 2 X satisfy x 4 y and y 4 z, then x 4 z.
(b) If x; y 2 X satisfy x 4 y and y 4 x, then x D y.
(c) If x; y 2 X , then x 4 y or y 4 x.
(d) If ; ¤ A X , then there is an a 2 A such that a 4 x for all x 2 A.
(e) If x 2 X n ¹º, then x 4 and the set ¹y 2 X j y 4 xº is countable.
Define the relation on X by x y if x 4 y and x ¤ y. For ; ¤ A X denote

by min.A/ 2 A the unique element of A that satisfies min.A/ 4 x for all x 2 A.
(See conditions (b) and (d).) For x 2 X define
Sx WD ¹y 2 X j x yº; Px WD ¹y 2 X j y xº:
Thus Px is the set of predecessors of x and Sx is the set of successors of x.

If x 2 X n ¹º, then Px is countable and Sx is uncountable. Define the map
sW X n ¹º ! X n ¹º; s.x/ WD min.Sx /:
Then X n Sx D Ps.x/ D Px [ ¹xº for all x 2 X . Let U 2X be the smallest

topology that contains the sets Px and Sx for all x 2 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 2 X such that x ¤ y
and suppose without loss of generality that x y. Then Ps.x/ and Sx are disjoint
open sets such that x 2 Ps.x/ and y 2 Sx .
(iii) We prove that every nonempty compact set K X contains a largest
element max.K/ 2 K such that K \ Smax.K/ D ;. This is obvious when 2 K
because S D ;. Thus assume … K and define
V WD ¹x 2 X j K Px º:
Since 2 V , this set is nonempty and min.X/ min.V / DW v because K ¤ ;.

Since X n K is open and v 2 X n K, there exist elements a; b 2 X such that
a v b and Sa \ Pb \ K D ;. This implies
K Pv n .Sa \ Pb / Pb n .Sa \ Pb / X n Sa D Ps.a/ :
Hence K n ¹aº Ps.a/ n ¹aº D Pa and K 6 Pa because a v and so a … V .

This implies a 2 K Ps.a/ , and hence K \ Sa D K n Ps.a/ D ;.
(iv) We prove that .X; U/ is compact. Let ¹Ui ºi2I be an open cover of X.
We prove by induction that there exist finite sequences
x1 ; : : : ; x ` 2 X and i1 ; : : : ; i` 2 I
such that
xk 2 Uik n Uik 1
and
Sxk Ui1 [ [ Uik 1
S`
for k 2, and X D j D1 Uij . Define
x1 WD
and choose i1 2 I such that 2 Ui1 . If Ui1 D X the assertion holds with ` D 1.
Now suppose, by induction, that x1 ; : : : ; xk and i1 ; : : : ; ik have been constructed
such that xj 2 Uij for j D 1; : : : ; k and Sxk Ui1 [ [ Uik 1 . If
Ui1 [ [ Uik D X;
then we are done with ` D k. Otherwise
Ck WD X n Ui1 [ [ Uik
is a nonempty compact set and we define
xkC1 WD max.Ck /
by part (iii). Then xkC1 2 Ck and Ck \SxkC1 D ;. Hence SxkC1 Ui1 [ [Uik .
Choose ikC1 2 I such that xkC1 2 UikC1 . This completes the induction argument.
The induction must stop because xkC1 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 2 N, be a sequence of uncountable compact sets. We prove
that the compact set \
K WD Ki
i2N
is uncountable. To see this, we first prove that
K n ¹º ¤ ;: (3.3)
Choose a sequence xn 2 X n¹º such that xn xnC1 for all n 2 N and x2k Ci 2 Ki
for 1 i 2k 1 and k 2 N. That such a sequence exists follows by induction
from the fact that the set X n Sxn D Ps.xn / is countable for each n, while the sets
Ki are uncountable for all i. Now the set
[
P WD Pxn
n2N
is countable and hence the set
S WD X n P
[
DXn Pxn
n2N
[
DXn Ps.xn /
n2N
\
D .X n Ps.xn / /
n2N
\
D Sxn
n2N
is uncountable. Hence
x WD min.S/ :
We prove that x 2 Ki for all i 2 N. Assume by contradiction that x … Ki for
some i . Then there are elements a; b 2 X such that a x b and
U WD Pb \ Sa X n Ki :
If xn 4 a for all n 2 N, then P Pa and so a 2 X n P D S , which is impossible
because a x D min.S/. Thus there must be an integer n0 2 N such that a xn0 .
This implies a xn x b and hence xn 2 U X n Ki for all n n0 ,
contradicting the fact that x2k Ci 2 Ki for all k 2 N. This contradiction shows that
our assumption that x … Ki for some i 2 N must have been wrong. Thus x 2 K
and this proves (3.3).
We prove that K is uncountable. Assume by contradiction that K is countable
and choose a sequence xi 2 K such that K n ¹º D ¹xi j i 2 Nº. Then s.xi /
and
Ki0 WD Ki \ Sxi D Ki n Ps.xi /
is an uncountable compact set for every i 2 N. Moreover,
\
K 0 WD Ki0 K n ¹xi j i 2 Nº D ¹º;
i2N
contradicting the fact that K 0 n ¹º ¤ ; by (3.3). This contradiction shows that K
is uncountable as claimed.
(vi) Define A 2X by
² ˇ ³
ˇ A [ ¹º contains an uncountable compact set;
A WD A X ˇ ˇ :
or Ac [ ¹º contains an uncountable compact set:
We prove that this is a -algebra. To see this, note first that X 2 A and that A 2 A
implies Ac 2 A by definition. Now choose a sequence Ai 2 A and denote
[
A WD Ai :
i2N
If one of the sets Ai [ ¹º contains an uncountable compact set, then so does the
set A [ ¹º. If none of the sets Ai [ ¹º contains an uncountable compact set, then
the set Aci [ ¹º contains an uncountable compact set for all i 2 N and hence so
T
does the set i2N .Aci [ ¹º/ D Ac [ ¹º by part (v). In both cases it follows that
A 2 A.
(vii) Define the map
W A ! Œ0; 1
by
´
1 if A [ ¹º contains an uncountable compact set,
.A/ WD
0 if Ac [ ¹º contains an uncountable compact set.
This map is well defined because the sets A [ ¹º and Ac [ ¹º cannot both contain
uncountable compact sets by part (v). It satisfies .;/ D 0. Moreover, if Ai 2 A is
a sequence of pairwise disjoint measurable sets, then at most one of the sets Ai [¹º
S P
can contain an uncountable compact set and hence . i 2N Ai / D i2N .Ai /.
Therefore, 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 [ ¹º is an uncountable
compact set and hence U 2 A. If U c is countable choose a sequence xi 2 U c such
that U c n ¹º D ¹xi j i 2 Nº and define
\
S WD Sxi :
i2N
S S
Then X n S D i2N .X n Sxi / D i2N Ps.xi / is a countable set and hence
s WD min.S/ :
Since xi s for all i 2 N, it follows that U c n ¹º Ps . Hence X n Ps is an

uncountable compact subset of U [ ¹º and so U 2 A.
(ix) The set

U WD X n ¹º
is uncountable and every compact subset of U is countable by part (v). Hence
.K/ D 0 for every compact subset K U and .U / D 1 because U [ ¹º D X
is an uncountable compact set. Thus jB W B ! Œ0; 1 is a Borel measure, but not
a Radon measure.
The next lemma and theorem are included here in preparation for the Riesz rep-
resentation theorem (Theorem 3.15). They explain the relation between the various
regularity properties of Borel measures.
Lemma 3.7. Let W B ! Œ0; 1 be an outer regular Borel measure that is inner
regular on open sets, i.e.,
.U / D sup¹.K/ j K U and K is compactº (3.4)
for every open set U X. Then the following holds:
(i) every Borel set B X with .B/ < 1 satisfies (3.2);
(ii) if X is -compact, then is regular.
Proof. (i) Fix a Borel set B X with .B/ < 1 and a constant " > 0. Since is
outer regular, there exists an open set U X such that
"
B U; .U / < .B/ C :
2
Thus U n B is a Borel set and .U n B/ D .U / .B/ < "=2. Use the outer
regularity of again to obtain an open set V X such that
"
U n 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 WD K n V . Since X is a Hausdorff space, K is closed, hence C is a closed
subset of K, and so C is compact (see Lemma A.2). Moreover,
C U n V B; B n C .B n K/ [ V .U n K/ [ V;
and hence .B n C / .U n K/ C .V / < ".

(ii) Choose a sequence of compact sets Ki X such that Ki Ki C1 for

all i 2 N and X D 1 i D1 Ki . Fix a Borel set B 2 B. If .B/ < 1, then B
S
satisfies (3.2) by (i). Hence assume .B/ D 1. Then it follows from part (iv) of
Theorem 1.28 that lim .B \ Ki / D 1. For each integer n 2 N choose in 2 N
i!1
such that
.B \ Kin / > n:
Since .B \ Kin / .Kin / < 1, 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 Lemma 3.7.
Theorem 3.8. Let 1 W B ! Œ0; 1 be an outer regular Borel measure that is inner
regular on open sets. Define
0 W B ! Œ0; 1
by
0 .B/ WD sup¹1 .K/ j K B and K is compactº for B 2 B: (3.5)
(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 2 B.
(ii) If X is -compact, then 0 D 1 .
(iii) If f W X ! R is a compactly supported continuous function, then

Z Z
f d0 D f d1 : (3.6)
X X
(iv) Let W B ! Œ0; 1 be a Borel measure that is inner regular on open sets. Then
Z Z
f d D f d1
X X
for every compactly supported continuous function f W X ! R if and only if
0 .B/ .B/ 1 .B/ for all B 2 B.

Proof. We prove that 0 is a measure. It follows directly from the definition that
0 .;/ D 0. Now assume that Bi 2 B is a sequence of pairwise disjoint Borel sets
and define
[1
B WD Bi :
i D1
Choose any compact set K B. Then 1 .Bi \ K/ < 1, and so, from part (i) of
Lemma 3.7,
0 .Bi \ K/ D 1 .Bi \ K/
for all i 2 N. Consequently,
1
X 1
X 1
X
1 .K/ D 1 .Bi \ K/ D 0 .Bi \ K/ 0 .Bi /:
iD1 iD1 iD1
Take the supremum over all compact sets K B to obtain

1
X
0 .B/ 0 .Bi /: (3.7)
iD1
To prove the converse inequality, it suffices to assume that 0 .B/ < 1. Then
0 .Bi / 0 .B/ < 1 for all i 2 N. Fix a constant " > 0 and choose a sequence
of compact sets Ki Bi such that 1 .Ki / > 0 .Bi / 2 i " for all i. Then, for
every n 2 N, the union K1 [ [ Kn is a compact subset of B and
n
X n
X
0 .B/ 1 .K1 [ [ Kn / D 1 .Ki / > 0 .Bi / ":
iD1 iD1
Now take the limit n ! 1 to obtain

1
X
0 .B/ 0 .Bi / ":
iD1
Since this holds for all " > 0, it follows that 0 .B/ 1
P
P1 i D1 0 .Bi / and hence
0 .B/ D i D1 0 .Bi / by (3.7). This shows that 0 is a measure. Moreover,
it follows directly from the definition of 0 that 0 .K/ D 1 .K/ for every compact
set K X . Since 1 is inner regular on open sets it follows that 0 .U / D 1 .U /
for every open set U X. Since 0 .K/ D 1 .K/ for every compact set K X ,
the definition of 0 in (3.5) shows that 0 is inner regular and hence is a Radon
measure. The inequality 0 .B/ 1 .B/ for B 2 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 sW X ! R is a Borel measurable step

function with compact support. Then
`
X
sD ˛i B ;
i
iD1
where ˛i 2 R and Bi 2 B with 1 .Bi / < 1. Hence 0 .Bi / D 1 .Bi / by part (i)
of Lemma 3.7 and hence
Z `
X Z
s d0 D ˛i 0 .Bi / D s d1 :
X i D1 X
Now let f W X ! Œ0; 1 be a Borel measurable function with compact support.

By Theorem 1.26, there exists a sequence of Borel measurable step functions
sn W X ! Œ0; 1/ such that
0 s1 .x/ s2 .x/ and f .x/ D lim sn .x/

n!1
for all x 2 X. Thus sn has compact support for each n. By the Lebesgue monotone
convergence theorem (Theorem 1.37) this implies
Z Z Z Z
f d0 D lim sn d0 D lim sn d1 D f d1 :
X n!1 X n!1 X X
If f W X ! R is a 1 -integrable function with compact support, then by what we

have just proved, Z Z
f ˙ d0 D f ˙ d1 < 1;
X X
so f is 0 -integrable and satisfies (3.6). This proves part (iii).

We prove part (iv) in four steps.
Step 1. Let W B ! Œ0; 1 be a Borel measure such that

Z Z
f d D f d1 (3.8)
X X
for every compactly supported continuous function f W 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 Urysohn’s lemma

(Theorem A.1) asserts that there exists a compactly supported continuous function
f W X ! R such that
f jK 1; supp.f / U; 0 f 1:
Hence it follows from (3.8) that

Z Z
.K/ f d D f d1 1 .U /
X X
and likewise Z Z
1 .K/ f d1 D 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/ inf¹1 .U / j K U X and U is openº D 1 .K/:
Since 1 .K/ .U / for every compact set K U and 1 is inner regular on
open sets, we obtain
1 .U / D sup¹1 .K/ j K U and K is compactº .U /:
Step 2. Let be as in Step 1 and assume in addition that is inner regular on open
sets. Then .K/ D 1 .K/ for every compact set K X and .U / D 1 .U / for
every open set U X.
If U X is an open set, then
.U / D sup¹.K/ j K U and K is compactº

sup¹1 .K/ j K U and K is compactº
D 1 .U / .U /:
Here the two inequalities follow from Step 1. It follows that .U / D 1 .U /.
Now let K be a compact set. Then 1 .K/ < 1. Since 1 is outer regular, there
exists an open set U X such that K U and 1 .U / < 1. Since and 1
agree on open sets it follows that
.K/ D .U / .U n K/ D 1 .U / 1 .U n K/ D 1 .K/:

Step 3. Let be as in Step 2. Then
0 .B/ .B/ 1 .B/ for all B 2 B: (3.9)
Fix a Borel set B 2 B. Then, by Step 2,
0 .B/ D sup¹1 .K/ j K B and K is compactº

D sup¹.K/ j K B and K is compactº
.B/
inf¹.U / j B U X and U is openº
D inf¹1 .U / j B U X and U is openº
D 1 .B/:
Step 4. Let W B ! Œ0; 1 be a Borel measure that satisfies (3.9). Then

Z Z Z
f d D f d0 D f d1
X X X
for every continuous function f W X ! R with compact support.

By the definition of the integral and part (iii),
Z Z Z Z
f d0 f d f d1 D f d0
X X X X
for every compactly supported continuous function f W X ! Œ0; 1/. Hence

Z Z Z
f d D f d0 D f d1
X X X
for every compactly supported continuous function f W X ! Œ0; 1/, and hence also
for every compactly supported continuous function f W 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
W B ! Œ0; 1 be Dieudonné’s measure.
(i) Take 1 WD and define the function 0 W B ! Œ0; 1 by (3.5). Then
0 .X/ D 1; 0 .¹º/ D 0; 0 .X n ¹º/ D 0;
and so 0 is not a measure. Hence the assumptions on 1 cannot be removed

in part (i) of Theorem 3.8.
3.2. Borel outer measures 107
(ii) Take 1 WD ı to be the Dirac measure at the point 2 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 W X ! R with
respect to the Dieudonné measure is given by
Z Z Z
f d D f ./ D f d0 D f d1 :
X X X
Moreover, the compact set K D ¹º satisfies .K/ D 0 < 1 D 1 .K/ and
the open set U WD X n ¹º satisfies 1 .U / D 0 < 1 D .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 measur-
able sets and then restrict it to B. Thus let A 2X be a -algebra containing B
and let W A ! Œ0; 1 be a measure. Call outer regular if it satisfies (3.1) for all
B 2 A, call it inner regular if it satisfies (3.2) for all B 2 A, and call it regular if
it is both outer and inner regular. If is regular and .X; B ; / denotes the com-
pletion of .X; B; jB /, 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 ; D jA : (3.10)
To see this, let A 2 A such that .A/ < 1. Choose a sequence of compact sets
Ki X and a sequence of open sets Ui X such that Ki A Ui and
.A/ 2 i .Ki / .Ui / .A/ C 2 i for all i 2 N. Then B0 WD 1
S
T1 iD1 Ki
and B1 WD i D1 Ui are Borel sets such that B0 A B1 and .B1 n B0 / D 0.
Thus every set A 2 A with .A/ < 1 belongs to B and .A/ D .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/ < 1 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 com-
pact 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 Sec-
tion 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
W 2X ! Œ0; 1
that satisfies the following axioms:
(a) if K X is compact, then .K/ < 1;
(b) if K0 ; K1 X are disjoint compact sets, then .K0 [ K1 / D .K0 / C .K1 /;
(c) .A/ D inf¹.U / j A U X; U is openº for every subset A X ;
(d) .U / D sup¹.K/ j K U; K is compactº for every open set U X .
Theorem 3.12. Let W 2X ! Œ0; 1 be a Borel outer measure. Then jB is an outer
regular Borel measure and is inner regular on open sets.
One can deduce Theorem 3.12 from Carathéodory’s theorem (Theorem 2.4) and
use axioms (a) and (b) (instead of the Carathéodory 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
Ae WD ¹E X j .E/ D sup¹.K/ j K E; K is compactº < 1º (3.11a)
and
A WD ¹A X j A \ K 2 Ae for every compact set K X º: (3.11b)
Here the subscript “e” stands for “endlich” and indicates that the elements of Ae are
sets of finite measure. We prove in seven steps that A is a -algebra containing B,
that
WD jA W A ! Œ0; 1
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.
Step 1. Let E1 ; E2 ; E3 ; : : : be a sequence of pairwise disjoint sets in Ae and define

1
[
E WD Ei :
iD1
Then the following holds:

(i) .E/ D 1
P
i D1 .Ei /;
(ii) if .E/ < 1, then E 2 Ae .

The assertions are obvious when .E/ D 1 because .E/ 1
P
iD1 .Ei /. Hence
assume .E/ < 1. We argue as in the proof of Theorem 3.8. Fix a constant " > 0.
Since Ei 2 Ae for all i , there is a sequence of compact sets Ki Ei such that
.Ki / > .Ei / 2 i " for all i . Then for all n 2 N
.E/ .K1 [ [ Kn /
D .K1 / C C .Kn / (3.12)
.E1 / C C .En / ":
Here the equality follows from condition (b) in Definition 3.11. Take the limit
n ! 1 to obtain
1
X
.Ei / .E/ C ":
i D1
Since this holds for all " > 0, it follows that
1
X 1
X
.Ei / .E/ .Ei /
i D1 iD1
and hence
1
X
.Ei / D .E/: (3.13)
i D1
Now (3.12) and (3.13) yield
.E/ .K1 [ [ Kn /
n
X
.Ei / "
i D1
1
X
D .E/ .Ei / "
iDnC1
for all n 2 N.
By (3.13), there exists an n" 2 N such that

1
X
.Ei / < ":
iDn" C1
Hence the compact set
K" WD K1 [ [ Kn" E
satisfies
.E/ .K" / .E/ 2":
Since this holds for all " > 0, we obtain
.E/ D sup¹.K/ j K E; K is compactº
and hence E 2 Ae . This proves Step 1.
Step 2. If E0 ; E1 2 Ae , then
E0 [ E1 2 A e ; E0 \ E1 2 Ae ; E0 n E1 2 Ae :
We first prove that E0 n E1 2 Ae . Fix a constant " > 0. Since E0 ; E1 2 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 / C "; i D 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 n Ki 2 Ae
for i D 0; 1 and it follows from Step 1 that
.Ei n Ki / .Ui n Ki / D .Ui / .Ki / 2"; (3.14a)
.Ui n Ei / .Ui n Ki / D .Ui / .Ki / 2"; (3.14b)
for i D 0; 1. Define
K WD K0 n U1 E0 n E1 : (3.15)
Then K is a compact set and
E0 n E1 .E0 n K0 / [ .K0 n U1 / [ .U1 n E1 /:
By the definition of an outer measure, this implies
.E0 n E1 / .E0 n K0 / C .K0 n U1 / C .U1 n E1 / .K/ C 4":
Here the last inequality follows from the definition of K in (3.15) and the inequali-
ties in (3.14). Since " > 0 was chosen arbitrarily, it follows that
.E0 n E1 / D sup¹.K/ j K E0 n E1 ; K is compactº
and hence E0 n E1 2 Ae . With this understood, it follows from Step 1 that
E0 [ E1 D .E0 n E1 / [ E1 2 Ae ; E0 \ E1 D E0 n .E0 n E1 / 2 Ae :
This proves Step 2.

Step 3. A is a -algebra.
First, X 2 A because K 2 Ae for every compact set K X .
Second, assume A 2 A and let K X be a compact set. Then by definition
A \ K 2 Ae . Moreover K 2 Ae and hence, by Step 2,
Ac \ K D K n .A \ K/ 2 Ae :
Since this holds for every compact set K X , we have Ac 2 Ae .

Third, let Ai 2 A for i 2 N and denote
1
[
A WD Ai :
iD1
Fix a compact set K X . Then
Ai \ K 2 Ae
for all i by the definition of A. Hence, by Step 2,
Bi WD Ai \ K 2 Ae
for all i and then, again by Step 2,
Ei WD Bi n .B1 [ [ Bi 1/ 2 Ae
for all i .
The sets Ei are pairwise disjoint and

1
[ 1
[
Ei D Bi D A \ K:
i D1 iD1
Since .A \ K/ .K/ < 1 by condition (a) in Definition 3.11, it follows from
Step 1 that A\K 2 Ae . This holds for every compact set K X and hence A 2 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 2 Ae for every
compact subset K X and so F 2 A. Therefore 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 2 Ae if and only if A 2 A and .A/ < 1.
If A 2 Ae , then A \ K 2 Ae for every compact set K X by Step 2 and hence
A 2 A. Conversely, let A 2 A such that .A/ < 1. Fix a constant " > 0.
By condition (c) in Definition 3.11, there exists an open set U X such that A U
and .U / < 1. By condition (d) in Definition 3.11, there exists a compact set
K X such that
K U; .K/ > .U / ":
Since K; U 2 Ae and U D .U n K/ [ K, it follows from Step 1 that
.U n K/ D .U / .K/ < ":
Moreover, A \ K 2 Ae because A 2 A. Hence it follows from the definition of Ae
that there exists a compact set H A \ K such that
.H / .A \ K/ "
D .A n .A n K// "
.A/ .A n K/ "
.A/ .U n K/ "
.A/ 2":
Since " > 0 was chosen arbitrarily, we conclude that
.A/ D sup¹.K/ j K A; K is compactº
and hence A 2 Ae . This proves Step 5.
3.3. The Riesz representation theorem 113
Step 6. WD jA is an outer regular extended Borel measure and is inner regular
on open sets.
We prove that is a measure. By definition, .;/ D 0. Now let Ai 2 A be a
sequence of pairwise disjoint measurable sets and define
1
[
A WD Ai :
iD1
If .Ai / < 1 for all i , then Ai 2 Ae by Step 5 and hence .A/ D 1

P
i D1 .Ai / by
Step 1. If .Ai / D 1 for some i, then .A/ .Ai / and so .A/ D 1. Thus is
a measure. Moreover, B A by Step 4, .K/ < 1 for every compact set K X
by condition (a) in Definition 3.11, is outer regular by condition (c) in Defini-
tion 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/ D 0, then E 2 Ae by the definition of Ae , and hence
E 2 A by Step 5. This proves Step 7 and Theorem 3.12.
3.3 The Riesz representation theorem

Let .X; U/ be a locally compact Hausdorff space and B be its Borel -algebra.
A function f W X ! R is called compactly supported if its support
supp.f / WD ¹x 2 X j f .x/ ¤ 0º
is a compact subset of X. The set of compactly supported continuous functions on

X will be denoted by
² ˇ ³
ˇ f is continuous and
Cc .X/ WD f W X ! R ˇ ˇ :
supp.f / is a compact subset of X
Thus a continuous function f W X ! R belongs to Cc .X/ if and only if there exists

a compact set K X such that f .x/ D 0 for all x 2 X n K. The set Cc .X/ is a
real vector space.
Definition 3.13. A linear functional ƒW Cc .X/ ! R is called positive if
f 0 H) ƒ.f / 0
for all f 2 Cc .X/.

The next lemma shows that every positive linear functional on Cc .X/ is con-
tinuous 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 ƒW Cc .X/ ! R be a positive linear functional and let

fi 2 Cc .X/ be a sequence of compactly supported continuous functions that con-
verges uniformly to f 2 Cc .X/. If there exists a compact set K X such that
supp.fi / K for all i 2 N,
then
ƒ.f / D lim ƒ.fi /:
i !1
Proof. Since fi converges uniformly to f , the sequence
"i WD sup jfi .x/ f .x/j

x2X
converges to zero. By Urysohn’s lemma (Theorem A.1), there exists a compactly

supported continuous function W X ! Œ0; 1 such that .x/ D 1 for all x 2 K.
This function satisfies
"i fi f "i for all i .
Hence
"i ƒ./ ƒ.fi / ƒ.f / "i ƒ./;
because ƒ is positive, and hence jƒ.fi / ƒ.f /j "i ƒ./ for all i . Since "i
converges to zero, so does jƒ.fi / ƒ.f /j and this proves Lemma 3.14.
Let W B ! Œ0; 1 be a Borel measure. Then every continuous function

f W X ! R with compact support is integrable with respect to . Define the map
ƒ W Cc .X/ ! R
by Z
ƒ .f / WD f d: (3.16)
X
Then ƒ is a positive linear functional. The Riesz representation theorem asserts

that every positive linear functional on Cc .X/ has this form. It also asserts unique-
ness under certain regularity hypotheses on the Borel measure. The following the-
orem includes two versions of the uniqueness statement.
Theorem 3.15 (Riesz representation theorem). Let ƒW Cc .X/ ! R be a positive

linear functional. Then the following holds.
(i) There exists a unique Radon measure
0 W B ! Œ0; 1
such that
ƒ0 D ƒ:
(ii) There exists a unique outer regular Borel measure
1 W B ! Œ0; 1
such that 1 is inner regular on open sets and
ƒ1 D ƒ:
(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 2 B.
(iv) Let W B ! Œ0; 1 be a Borel measure that is inner regular on open sets. Then
ƒ D ƒ () 0 .B/ .B/ 1 .B/ for all B 2 B.
Proof. The proof has nine steps. Step 1 defines a function W 2X ! Œ0; 1, 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 ƒ1 D ƒ.
Step 8 defines 0 and Step 9 proves uniqueness.
Step 1. Define the function
U W U ! Œ0; 1
by
U .U / WD sup¹ƒ.f / j f 2 Cc .X/; 0 f 1; supp.f / U º (3.17)
for every open set U X and define
W 2X ! Œ0; 1
.A/ WD inf¹U .U / j A U X; U is openº (3.18)
for every subset A X . Then
.U / D 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 / D inf¹U .V / j U V X; V is openº D U .U / for every open set
U X and this proves Step 1.
Step 2. The function W 2X ! Œ0; 1 in Step 1 is an outer measure.
By definition, .;/ D U .;/ D 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 / C U .V /: (3.19)
To see this, let f 2 Cc .X/ such that 0 f 1 and
K WD supp.f / U [ V:
By Theorem A.4, there exist functions ; 2 Cc .X/ such that
supp./ U; supp. / V; ; 0; C 1; . C /jK 1:
Hence f D f C f and so
ƒ.f / D ƒ.f C f / D ƒ.f / C ƒ. f / U .U / C U .V /:
This proves (3.19).

Now choose a sequence of subsets Ai X and define
1
[
A WD Ai :
iD1
We must prove that

1
X
.A/ .Ai /: (3.20)
iD1
If there exists an i 2 N such that .Ai / D 1, then .A/ D 1 because Ai A

and hence 1
P
iD1 .Ai / D 1 D .A/. Hence assume .Ai / < 1 for all i . Fix a
constant " > 0. By the definition of in (3.18), there exists a sequence of open sets
Ui X such that
Ai Ui ; U .Ui / < .Ai / C 2 i ":
Define
1
[
U WD Ui :
iD1
Let f 2 Cc .X/ such that 0 f 1 and supp.f / U . Since f has compact

support, there exists an integer k 2 N such that
k
[
supp.f / Ui :
i D1
By the definition of U and (3.19), this implies
ƒ.f / U .U1 [ [ Uk /
U .U1 / C C U .Uk /
< .A1 / C C .Ak / C ":
Hence
1
X
ƒ.f / .Ai / C "
iD1
for every f 2 Cc .X/ such that 0 f 1 and supp.f / U . Consequently
1
X
.A/ U .U / .Ai / C "
i D1
by the definition of U .U / in (3.17). Thus

1
X
.A/ .Ai / C " for every " > 0,
iD1
and hence
1
X
.A/ .Ai /:
iD1
This proves (3.20) and Step 2.
Step 3. Let U X be an open set. Then
U .U / D sup¹.K/ j K U; K is compactº: (3.21)
Let f 2 Cc .X/ such that
0 f 1; K WD supp.f / U:
Then it follows from the definition of U in (3.17) that ƒ.f / U .V / for every
open set V X with K V . Hence it follows from the definition of in (3.18)
that
ƒ.f / .K/:
Hence
U .U / D sup¹ƒ.f / j f 2 Cc .X/; 0 f 1; supp.f / U º

sup¹.K/ j K U; K is compactº
.U /
D U .U /:
Hence U .U / D sup¹.K/ j K U; K is compactº and this proves Step 3.

Step 4. Let K X be an compact set. Then
.K/ D inf¹ƒ.f / j f 2 Cc .X/; f 0; f jK 1º: (3.22)
In particular, .K/ < 1.

Define
a WD inf¹ƒ.f / j f 2 Cc .X/; f 0; f jK 1º:
We prove that a .K/. Let U X be any open set containing K. By Urysohn’s
lemma (Theorem A.1), there exists a function f 2 Cc .X/ such that
0 f 1; supp.f / U; f jK 1:
Hence,
a ƒ.f / U .U /:
This shows that a U .U / for every open set U X containing K. Take the infi-
mum 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 2 Cc .X/ such that f 0 and
f .x/ D 1 for all x 2 K. Fix a constant 0 < ˛ < 1 and define
U˛ WD ¹x 2 X j f .x/ > ˛º:
Then U˛ is open and K U˛ . Hence
.K/ U .U˛ /:
Moreover, every function g 2 Cc .X/ with 0 g 1 and supp.g/ U˛ satisfies

˛g.x/ ˛ f .x/ for x 2 U˛ , hence ˛g f , and so ˛ƒ.g/ ƒ.f /. Take the
supremum over all such g to obtain ˛U .U˛ / ƒ.f / and hence
1
.K/ U .U˛ / ƒ.f /:
˛
This shows that .K/ ˛1 ƒ.f / for all ˛ 2 .0; 1/, and hence
.K/ ƒ.f /:
Since this holds for every function f 2 Cc .X/ with f 0 and f jK 1, it follows
that .K/ a. This proves Step 4.
Step 5. Let K0 ; K1 X be compact sets such that K0 \ K1 D ;. Then
.K0 [ K1 / D .K0 / C .K1 /:
The inequality .K0 [ K1 / .K0 / C .K1 / holds because is an outer measure

by Step 2. To prove the converse inequality, choose f 2 Cc .X/ such that
0 f 1; f jK0 0; f jK1 1:
That such a function exists follows from Urysohn’s lemma (Theorem A.1) with
K WD K1 and U WD X n K0 :
Now fix a constant " > 0. Then it follows from Step 4 that there exists a function
g 2 Cc .X/ such that
g 0; gjK0 [K1 1; ƒ.g/ < .K0 [ K1 / C ":
It follows also from Step 4 that
.K0 / C .K1 / ƒ..1 f /g/ C ƒ.fg/ D ƒ.g/ < .K0 C K1 / C ":
Hence
.K0 / C .K1 / < .K0 C K1 / C " for every " > 0,
and therefore
.K0 / C .K1 / .K0 C K1 /:
This proves Step 5.
Step 6. The function

1 WD jB W B ! Œ0; 1
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 Defini-

tion 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
ƒ1 D ƒ:
We will prove that Z
ƒ.f / f d1 (3.23)
X
for all f 2 Cc .X/. Once this is understood, it follows that
Z Z
ƒ.f / D ƒ. f / . f / d1 D f d1
X X
and hence
Z
f d1 ƒ.f / for all f 2 Cc .X/.
X
Thus Z
ƒ.f / D f d1 for all f 2 Cc .X/,
X
and this proves Step 7.

Thus it remains to prove the inequality (3.23). Consider a continuous function
f W X ! R with compact support and denote
K WD supp.f /; a WD inf f .x/; b WD sup f .x/:
x2X x2X
Fix a constant " > 0 and choose real numbers

y0 < a < y1 < y2 < < yn 1 < yn D b
such that
yi yi 1 < "; i D 1; : : : ; n:
For i D 1; : : : ; n define
Ei WD ¹x 2 K j yi 1 < f .x/ yi º:
Then Ei is the intersection of the open set f 1 ..yi 1 ; 1// with the closed set
f 1 .. 1; yi / and hence is a Borel set. Moreover, Ei \ Ej D ; for i ¤ j
and
[n
KD Ei :
i D1
Since 1 is outer regular, there exist open sets U1 ; : : : ; Un X such that

"
Ei Ui ; 1 .Ui / < 1 .Ei / C ; sup f < yi C " (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 .. 1; yi C "//.) By Theo-
rem A.4, there exist functions 1 ; : : : ; n 2 Cc .X/ such that
n
X n
X
i 0; supp.i / Ui ; i 1; i jK 1: (3.25)
iD1 iD1
It follows from (3.24), (3.25), and Step 4 that

n
X
f D i f; i f .yi C "/i ;
iD1
and
n
X "
1 .K/ ƒ.i /; ƒ.i / 1 .Ui / < 1 .Ei / C :
n
iD1
Hence,
n
X
ƒ.f / D ƒ.i f /
i D1
X n
.yi C "/ƒ.i /
i D1
X n n
X
D .yi C jaj C "/ƒ.i / jaj ƒ.i /
i D1 iD1
n
X "
.yi C jaj C "/ 1 .Ei / C jaj1 .K/
n
i D1
n n
X "X
D .yi C "/1 .Ei / C .yi C jaj C "/
n
i D1 iD1
n
X
.yi "/1 .Ei / C ".21 .K/ C b C jaj C "/
ZiD1
f d1 C ".21 .K/ C b C jaj C "/:
X
Here we have used the inequality yi C Rjaj C " 0. Since " > 0 can be chosen
arbitrarily small it follows that ƒ.f / X f d1 . This proves (3.23).
Step 8. Define
0 W B ! Œ0; 1
by
0 .B/ WD sup¹.K/ j K B; K is compactº
Then 0 is a Radon measure, ƒ0 D ƒ, 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 ƒ0 D ƒ1 D ƒ. 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/ D .K/ D 1 .K/ for every compact set K X. Second,
it follows from and Steps 1 and 3 that 0 .U / D U .U / D .U / D 1 .U / for
every open set U X. Third, Steps 7 and 8 assert that ƒ0 D ƒ1 D ƒ. Hence it
follows from part (iv) of Theorem 3.8 that every Borel measure W B ! Œ0; 1 that
is inner regular on open sets and satisfies ƒ D ƒ agrees with on all compact
sets and on all open sets. Hence, every Radon measure W B ! Œ0; 1 that satisfies
ƒ D ƒ is given by
.B/ D sup¹.K/ j K B; K is compactº D 0 .B/
for every B 2 B. Likewise, every outer regular Borel measure W B ! Œ0; 1 that
is inner regular on open sets and satisfies ƒ D ƒ is given by
.B/ D inf¹.U / j B U X; U is openº D .B/ D 1 .B/
for every B 2 B. This proves Step 9 and Theorem 3.15.
The following corollary is the converse of Theorem 3.8.
Corollary 3.16. Let 0 W B ! Œ0; 1 be a Radon measure and define
1 .B/ WD inf¹0 .U / j B U X; U is openº for all B 2 B: (3.26)
Then 1 is an outer regular Borel measure, is inner regular on open sets, and
0 .B/ D sup¹1 .K/ j K B; K is compactº for all B 2 B: (3.27)
Proof. Let 1 be the unique outer regular Borel measure on X that is inner regular
on open sets and satisfies ƒ1 D ƒ0 . 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 D K compact and 0 .K/ D 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. (i) Let W B ! Œ0; 1 be a Borel measure that is inner regular on open
sets and let 0 ; 1 W B ! Œ0; 1 be the Borel measures associated to ƒ WD ƒ
in parts (i) and (ii) of the Riesz representation theorem (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 2 B. Since X is -compact, it follows from
part (ii) of Theorem 3.8 that 0 D D 1 . Hence is regular.
(ii) Let W B ! Œ0; 1 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 KiC1 for all i 2 N and U D 1
S
iD1 Ki .
Hence
.U / D lim .Ki /
i!1
by Theorem 1.28, so
.U / D sup¹.K/ j K U and K is compactº:
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 Theo-
rem 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 ƒW Cc .X/ ! R
there exists a unique Borel measure such that
ƒ D ƒ:
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/ D 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 hypothe-
ses 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 2 X, there is a sequence of open sets Wi , i 2 N, such
that x 2 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. (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 2 V such that Vx U and Vx is
compact. Let x 2 U . By Lemma A.3 there is an open set W X with compact
closure such that x 2 W W S U . Since V is a basis of the topology, there is an
element V 2 V such that x 2 V W . Hence Vx is a closed subset of the compact
set WS and so is compact by Lemma A.2. Thus V 2 V.U / and x 2 V . This shows
that [
U D V:
V 2V.U /
Since V is countable, so is V.U /. Choose a bijection N ! V.U /W i 7! Vi and define
Ki WD Vx1 [ [ Vxi
for i 2 N. Then Ki Ki C1 for all i and U D 1

S
iD1 Ki . Hence U is -compact.
(ii) Fix an element x 2 X. Since X is a Hausdorff space, the set X n ¹xº is open
and hence is -compact by assumption. Choose a sequence of compact sets
Ki X n ¹xº
such that
Ki KiC1 for all i 2 N
and
1
[
Ki D X n ¹xº:
iD1
Then each set

Ui WD X n 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 2 Vi Vxi Ui D X n Ki for all i 2 N.
Define
Wi WD V1 \ \ Vi for i 2 N.
Then
i
\
Si
W .X n Kj / D X n Ki
j D1
and hence
1
\
Si D ¹xº:
W
iD1
This implies that each open set U X that contains x also contains one of the
Si . Namely, if x 2 U and U is open, then W
sets W S1 n U is a compact set contained
in
[1
X n ¹xº D Si /;
.X n W
iD1
hence there exists a j 2 N such that
j
[
S1 n U
W Si / D X n W
.X n W Sj ;
iD1
and so WSj U . Thus the sets Wj form 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 WD Œ0; 1 ¹0; 1º
and denotes the lexicographic ordering

´
s < t or
.s; i/ .t; j / ()
s D t and i D 0 and j D 1:
The topology U 2X is defined as the smallest topology containing the sets
Sa WD ¹x 2 X j a xº; Pb WD ¹x 2 X j x bº; a; b 2 X:
It has a basis consisting of the sets Sa , Pb , Sa \ Pb for all a; b 2 X.

This topological space .X; U/ is a compact Hausdorff space and is perfectly nor-
mal, i.e., for any two disjoint closed subsets F0 ; F1 X there exists a continuous
function f W X ! Œ0; 1 such that
1 1
F0 D f .0/; F1 D f .1/:
(For a proof see Dan Ma’s topology blog [12].) This implies that every open subset
of X is -compact. Moreover, the subsets
Y0 WD .0; 1/ ¹0º; Y1 WD .0; 1/ ¹1º
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
Œa; 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 WD R2 and
´
0 if x1 D x2 ;
d..x1 ; y1 /; .x2 ; y2 // WD jy1 y2 j C
1 if x1 ¤ x2 :
Let B 2X be the Borel -algebra of .X; d /.

3.4. Exercises 127
(i) Show that .X; d / is locally compact.

(ii) Show that for every compactly supported continuous function f W X ! R
there exists a finite set Sf R such that supp.f / Sf R.
(iii) Define the positive linear functional ƒW Cc .X/ ! R by
X Z 1
ƒ.f / WD f .x; y/ dy:
x2Sf 1
(Here the integrals on the right are understood as the Riemann integrals or,
equivalently by Theorem 2.24, as the Lebesgue integrals.) Let W B ! Œ0; 1
be a Borel measure such that
Z
f d D ƒ.f / for all f 2 Cc .X/:
X
Prove that every one-element subset of X has measure zero.

(iv) Let be as in (iii) and let E WD R ¹0º. This set is closed. If is inner
regular, prove that .E/ D 0. If is outer regular, prove that .E/ D 1.
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 WD N [ ¹1º and
U WD ¹U X j U N or #U c < 1º:
Thus .X; U/ is the (Alexandrov) one-point compactification of the set N of natural

numbers with the discrete topology. (If 1 2 U , then the condition #U c < 1 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 D 2X .
(ii) Let W 2X ! Œ0; 1 be the counting measure. Prove that is inner regular,
but not outer regular.
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 W X ! Y be a
continuous map and let X W BX ! Œ0; 1 be a measure.
(i) Prove that BY BX (see Exercise 1.69).

(ii) If X is inner regular, show that X jBY is inner regular.
(iii) Find an example where X is outer regular and X jBY 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 W X ! Œ0; 1 such that F0 D f 1 .0/ and F1 D f 1 .1/. Compare this with
Urysohn’s lemma (Theorem A.1). Hint. An explicit formula for f is given by
d.x; F0 /
f .x/ WD ;
d.x; F0 / C d.x; F1 /
where
d.x; F / WD inf d.x; y/
y2F
for x 2 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 Œa; 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 WD Œ0; 1 ¹0; 1º
with the topology induced by the lexicographic ordering. Prove that B X is a

Borel set for this topology if and only if there exist a Borel set E Œ0; 1 and two
countable sets F; G X such that
B D ..E ¹0; 1º/ [ F / n G: (3.28)
Hint 1. Show that the projection f W X ! Œ0; 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 Œ0; 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
² ˇ ³
ˇ K is compact and there is a sequence of open sets
Ka WD K X ˇ :
Ui such that UiC1 Ui for all i and K D 1
ˇ T
iD1 Ui
3.4. Exercises 129
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 W X ! R is called Baire measurable if f 1 .U / 2 Ba for
every open set U R. A Baire measure is a measure W Ba ! Œ0; 1 such that
.K/ < 1 for all K 2 Ka .
(i) Let f W X ! R be a continuous function with compact support. Prove that

f 1 .c/ 2 Ka for every nonzero real number c.
(ii) Prove that Ba is the smallest -algebra such that every continuous function
f W X ! R with compact support is Ba -measurable.
(iii) If every open subset of X is -compact, prove that Ba D 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 WD 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
W Ba ! Œ0; 1
by
´
0 if B is countable;
.B/ WD
1 if B c is countable:
R
Show that X f d D 0 for every f 2 Cc .X/. Thus positive linear functionals
ƒW 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 n ¹º and their complements.
(iii) Let X be the Stone–Čech 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 D 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 WD ¹y 2 R j .x; y/ 2 Bº is a Borel set in R for every x 2 R and one of
the sets S0 WD ¹x 2 R j Bx ¤ ;º and S1 WD ¹x 2 R j Bx ¤ Rº is countable.
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 W X ! R. For F F.X/ consider the following conditions:
(a) Cc .X/ F;
(b) If fi 2 F is a sequence converging pointwise to f 2 F.X/, then f 2 F.
Let Fa F.X/ be the intersection of all subsets F F.X/ that satisfy condi-
tions (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 measurable
functions on X satisfies (a) and (b).
(iii) If f 2 Fa and g 2 Cc .X/, then f C g 2 Fa . Hint. Let g 2 Cc .X/. Then the

set Fa g satisfy (a) and (b) and hence contains Fa .
(iv) If f; g 2 Fa , then f Cg 2 Fa . Hint. Let g 2 Fa . Then the set Fa g satisfy (a)

and (b) and hence contains Fa .
(v) If f 2 Fa and c 2 R, then cf 2 Fa . Hint. Fix a real number c ¤ 0. Then the

set c 1 Fa satisfy (a) and (b) and hence contains Fa .
(vi) If f 2 Fa and g 2 Cc .X/, then fg 2 Fa . Hint. Fix a real number c such that
c C g.x/ > 0 for all x 2 R. Then the set .c C g/ 1 Fa satisfy (a) and (b) and
hence contains Fa . Now use (iv) and (v).
(vii) If A X such that A 2 Fa and f 2 Fa , then f A 2 Fa . Hint. The set

.1 C A / 1 Fa satisfy (a) and (b) and hence contains Fa .
(viii) The set

A WD ¹A X j A 2 Fa or X nA 2 Fa º
is a -algebra. Hint. If A ; B 2 FA , then A[B D A C B A B 2 Fa .
If XnA ; XnB 2 FA , then X n.A[B/ D XnA X nB 2 Fa . If A ; XnB 2 FA ,
then X n.A[B/ D .X nA/\.X nB/ D XnB A X nB 2 Fa . Thus
A; B 2 A H) A [ B 2 A:
3.4. Exercises 131
(ix) A D Ba . Hint. Let K 2 Ka . Use Urysohn’s lemma (Theorem A.1) to con-

struct a sequence gi 2 Cc .X/ that converges pointwise to K .
(x) For every f 2 Fa there exists a sequence of compact sets Ki 2 Ka such that
S
Ki KiC1 for all i and supp.f / i2N Ki . Hint. The set of functions
f W 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 / D 0 and 0 .B/ D 1 .B/ for every Baire set B X n N .
Hint 1. Show first that
0 .B/ D sup¹0 .K/W K 2 Ka ; K Bº; (3.29)
where Ka is as in Exercise 3.29. To see this, prove that the right-hand side of equa-
tion (3.29) defines a Borel measure on X that is inner regular on open sets and
satisfies 0 and ƒ D ƒ0 .
Hint 2. Suppose there exists a Baire set N X such that 0 .N / < 1 .N /.
Show that 1 .N / D 1 and that N can be chosen such that 0 .N / D 0. Next
show that XnN 2 Fa , where Fa is as in Exercise 3.31, and deduce that X n N is
contained in a countable union of compact sets.
Example 3.33. Let X be the Stone–Čech 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 n B N. (See part (iii)
of Exercise 3.30.) For a Borel set B X, define
X1
0 .B/ WD ;
n
n2B
² ˇ ³
ˇ B U X;
1 .B/ WD inf 0 .U / ˇ
ˇ :
U is open
As in Example 4.60, denote by X0 X the union of all open sets U X with

0 .U / < 1. 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 N is a Baire set in X0 and 0 .X0 n N/ D 0
while 1 .X0 n N/ D 1. Thus we can choose N WD X0 n N in Exercise 3.32 and
0 and 1 do not agree on the Baire -algebra.
Example 3.34. Let X D R2 be the locally compact Hausdorff space in Exam-

ple 3.23 and let 0 ; 1 be the Borel measures of Theorem 3.15 associated to the
linear functional ƒW Cc .X/ ! R in that example. Then it follows from part (iv)
of Exercise 3.30 that 0 .B/ D 1 .B/ for every Baire set B X. Thus we can
choose N D ; in Exercise 3.32. However, there does not exist any Borel set N X
such that 0 .N / D 0 and 0 agrees with 1 on all Borel subsets of X n N . (A set
N X is a Borel set with 0 .N / D 0 if and only if Nx WD ¹y 2 R j .x; y/ 2 N º
is a Borel set and m.Nx / D 0 for all x 2 R.)
Exercise 3.35. Let Z be the disjoint union of the locally compact Hausdorff spaces
X0 in Example 3.33 and X D 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 1. Section 4.1 introduces the inequalities of
Hölder and Minkowski and Section 4.2 shows that Lp ./ is complete. In Sec-
tion 4.3 we prove that, when X is a locally compact Hausdorff space, is a Radon
measure, and 1 p < 1, the subspace of continuous functions with compact sup-
port is dense in Lp ./. If in addition X is second countable, it follows that Lp ./
is separable. When 1 < p < 1 (or p D 1 and the measure space .X; A; / is
localizable) the dual space of Lp ./ is isomorphic to Lq ./, where 1=pC1=q D 1.
For p D 2 this follows from elementary Hilbert space theory and is proved in Sec-
tion 4.4. For general p the proof requires the Radon–Nikodým theorem and is de-
ferred to Chapter 5. Some preparatory results are proved in Section 4.5.
4.1 Hölder and Minkowski
Assume throughout that .X; A; / is a measure space and that p; q are real numbers
such that
1 1
C D 1; 1 < p < 1; 1 < q < 1: (4.1)
p q
Then any two nonnegative real numbers a and b satisfy Young’s inequality
1 p 1 q
ab a C b ; (4.2)
p q
and equality holds in (4.2) if and only if ap D b q .

Exercise. Prove this by examining the critical points of the function
1 p
.0; 1/ ! RW x 7 ! x xb:
p
134 4. Lp spaces
Theorem 4.1. Let f; gW X ! Œ0; 1 be measurable functions. Then f and g satisfy

the Hölder inequality
Z Z 1=p Z 1=q
p q
fg 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 C g/ d f d C g d : (4.4)
X X X
Proof. Define
Z 1=p Z 1=q
p q
A WD f d ; B WD g d :
X X
If A D 0, then f D 0 almostR everywhere by Lemma 1.49, hence fg D 0 almost
everywhere, and hence X fg d D 0 by Lemma 1.48. This proves the Hölder
inequality (4.3) in the case A D 0. If A D 1 and B > 0, then AB D 1 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 < 1
and 0 < B < 1. Then it follows from (4.2) that
R
X fg d f g
Z
D d
AB X AB
Z p
1 f 1 g q
C d
X p A q B
1 X f p d 1 X g q d
R R
D C
p Ap q Bq
1 1
D C
p q
D 1:
This proves the Hölder inequality. To prove the Minkowski inequality, put
Z 1=p
a WD f p d ;
X
Z 1=p
p
b WD g d ;
X
Z 1=p
c WD .f C g/p d :
X
4.1. Hölder and Minkowski 135
We must prove that c a C b. This is obvious when a D 1 or b D 1. Hence

assume a; b < 1. We first show that c < 1. This holds because
f .f p C g p /1=p and g .f p C g p /1=p ;
hence
f C g 2.f p C g p /1=p ;
therefore
.f C g/p 2p .f p C g p /;
and integrating this inequality and raising the integral to the power 1=p we obtain
c 2.ap C b p /1=p < 1:
With this understood, it follows from the Hölder inequality that

Z Z
p p 1
c D f .f C g/ d C g.f C g/p 1
d
X X
Z 1=p Z 1=q
f p d .f C g/.p 1/q
d
X X
Z 1=p Z 1=q
C g p d .f C g/.p 1/q
d
X X
Z 1 1=p
p
D .a C b/ .f C g/ d
X
p 1
D .a C b/c :
Here we have used the identity pq q D p. It follows that c a C b and this

proves Theorem 4.1.
Exercise 4.2. (i) Assume 0 < X f p d < 1 and 0 < X g q d < 1. Prove
R R
that equality holds in (4.3) if and only if there exists a constant ˛ > 0 such that
g q D ˛f p almost everywhere. Hint. Use the proof of the Hölder inequality and the
fact that equality holds in (4.2) if and only ap D b q .
(ii) Assume 0 < X f p d < 1 and 0 < X g p d < 1. Prove that equality
R R
holds in (4.4) if and only if there is a real number > 0 such that g D f almost
everywhere. Hint. Use part (i) and the proof of the Minkowski inequality.
136 4. Lp spaces
4.2 The Banach space Lp ./
Definition 4.3. Let .X; A; / be a measure space and let 1 p < 1. Consider a
measurable function f W X ! R x . The Lp -norm of f is the number
Z 1=p
kf kp WD jf jp d : (4.5)
X
A function f W X ! R is called p-integrable or an Lp -function if it is measurable

and kf kp < 1. The space of Lp -functions is denoted by
Lp ./ WD ¹f W X ! R j f is A-measurable and kf kp < 1º: (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 ./ ! Œ0; 1/W f 7 ! kf kp
satisfies the triangle inequality
kf C gkp kf kp C kgkp
for all f; g 2 Lp ./ by (4.4), and
kf kp D jj kf kp ;
for all 2 R and f 2 Lp ./ by definition. However, in general kkp is not a

norm on Lp ./ because kf kp D 0 if and only if f D 0 almost everywhere by
Lemma 1.49. We can turn the space Lp ./ into a normed vector space by identify-
ing two functions f; g 2 Lp ./ whenever they agree almost everywhere. Thus we
introduce the equivalence relation

f g () f D g -almost everywhere: (4.7)
Denote the equivalence class of a function f 2 Lp ./ under this equivalence

relation by Œf and the quotient space by

Lp ./ WD Lp ./= : (4.8)
4.2. The Banach space Lp ./ 137
This is again a real vector space. (For p D 1 see Definition 1.51.) The Lp -norm
in (4.5) depends only on the equivalence class of f and so the map
Lp ./ ! Œ0; 1/W Œf 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 < 1. It is often convenient to abuse notation
and write f 2 Lp ./ instead of Œf 2 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 < 1 and let f; g 2 Lp ./ such that
kf C gkp D kf kp C kgkp ; kf kp ¤ 0:
Then it follows from part (ii) of Exercise 4.2 that there exists a real number 0
such that g D f almost everywhere.
Example 4.5. If .Rn ; A; m/ is the Lebesgue measure space we write
Lp .Rn / WD Lp .m/:
(See Definition 2.2 and Definition 2.11.)

Example 4.6. If W 2N ! Œ0; 1 is the counting measure we write
`p WD Lp ./:
Thus the elements of `p are sequences .xn /n2N of real numbers such that
1
X 1=p
k.xn /kp WD jxn jp < 1:
pD1
If we define
f WN ! R
by
f .n/ WD xn for n 2 N,
then Z 1
X
jf jp d D jxn jp :
N pD1
For p D 1 there is a similar normed vector space L1 ./, defined next.

138 4. Lp spaces
Definition 4.7. Let .X; A; / be a measure space and let f W X ! Œ0; 1 be a mea-
surable function. The essential supremum of f is the number ess sup f 2 Œ0; 1
defined by
ess sup f WD inf¹c 2 Œ0; 1 j f c almost everywhereº (4.9)
A function f W X ! R is called an L1 -function if it is measurable and
kf k1 WD ess sup jf j < 1 (4.10)
The set of L1 -functions on X will be denoted by
L1 ./ WD ¹f W X ! R j f is measurable and ess supjf j < 1º
and the quotient space by the equivalence relation (4.7) by

L1 ./ WD L1 ./= : (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 2 L1 ./ there exists a measurable set E 2 A such that
.E/ D 0 and sup jf j D kf k1 :

X nE
Proof. The set

En WD ¹x 2 X j jf .x/j > kf k1 C 1=nº
has measure zero for all n. Hence
[
E WD En
n2N
is also a set of measure zero and jf .x/j kf k1 for all x 2 X n E. Therefore

supX nE jf j D kf k1 . This proves Lemma 4.8.
Theorem 4.9. Lp ./ is a Banach space for 1 p 1.

Proof. Assume first that 1 p < 1. In this case the argument is a refinement of
the proof of Theorem 1.52 and Theorem 1.53 for the case p D 1. Let fn 2 Lp ./
be a Cauchy sequence with respect to the norm (4.5). Choose a sequence of positive
integers n1 < n2 < n3 < such that
i
kfni fni C1 kp < 2
for all i 2 N.
4.2. The Banach space Lp ./ 139
Define
k
X 1
X
gk WD jfniC1 fni j; g WD jfniC1 fni j D lim gk :
k!1
iD1 iD1
Then it follows from Minkowski’s inequality (4.4) that

k
X k
X
i
kgk kp kfni fniC1 kp < 2 1
iD1 i D1
for all k 2 N. Moreover,
gkp gkC1
p
for all k 2 N
and the sequence of functions gkp W X ! Œ0; 1 converges pointwise to the integrable
function g p . Hence it follows from the Lebesgue monotone convergence theorem
(Theorem 1.37) that
kgkp D lim kgk kp 1:
k!1
Hence, by Lemma 1.47, there is a measurable set E 2 A such that
.E/ D 0; g.x/ < 1 for all x 2 X n E:

P1
Hence the series iD1 .fni C1 .x/ fni .x// converges absolutely for x 2 X n E.
Define the function
f WX ! R
by
1
X
f .x/ WD fn1 .x/ C .fniC1 .x/ fni .x//
iD1
for x 2 X n E, and by f .x/ WD 0 for x 2 E. Then the sequence

k
X1
fnk X nE D fn1 XnE C .fniC1 fni /XnE
iD1
converges pointwise to f . Hence f is A-measurable by Theorem 1.24.

We must prove that f 2 Lp ./ and that
lim kf fn kp D 0:
n!1
To see this, fix a constant " > 0. Then there exists an integer n0 2 N such that
kfn fm kp < " for all n; m n0 .

140 4. Lp spaces
By Fatou’s lemma (Theorem 1.41), this implies

Z Z
jfn f jp d D lim inf jfn fnk XnE jp d
X X k!1
Z
lim inf jfn fnk XnE jp d
k!1 X
Z
D lim inf jfn fnk jp d
k!1 X
"p
for all n n0 . Hence kfn f kp " for all n n0 , and so
kf kp kfn0 kp C kf fn0 kp kfn0 kp C " < 1:
Thus f 2 Lp ./ and lim kf fn kp D 0 as claimed. This shows that Lp ./ is

n!1
a Banach space for p < 1.
The proof for p D 1 is simpler. Let fn 2 L1 ./ such that the Œfn form a
Cauchy sequence in L1 ./. Then there is a set E 2 A such that
.E/ D 0; kfn k1 D sup jfn j; kfm fn k1 D sup jfm fn j (4.12)

X nE X nE
for all m; n 2 N. To see this, use Lemma 4.8 to find null sets En ; Em;n 2 A such
that
sup jfn j D kfn k1 and sup jfm fn j D kfm fn k1 for all m; n 2 N.

XnEn XnEm;n
Then the union E of the sets En and Em;n is measurable and satisfies (4.12). Since
Œfn is a Cauchy sequence in L1 ./, we have
lim "n D 0; "n WD sup kfm fn k1 :

n!1 mn
Since jfm .x/ fn .x/j "n for all m n and all x 2 X n E, it follows that
.fn .x//n2N is Cauchy sequence in R and hence converges for every x 2 X n E.
Define
f WX ! R
by 8
< lim fn .x/ for x 2 X n E,
n!1
f .x/ WD
:0 for x 2 E.
4.3. Separability 141
Then
kf fn k1 sup jf .x/ fn .x/j D sup lim jfm .x/ fn .x/j "n

x2X nE x2XnE m!1
for all n 2 N. Hence
kf k1 kf1 k1 C "1 < 1 and lim kf fn k1 D 0:

n!1
Corollary 4.10. Let .X; A; / be a measure space and, given 1 p 1, let

f 2 Lp ./ and fn 2 Lp ./ be a sequence such that
lim kfn f kp D 0:
n!1
• If p D 1, then fn converges almost everywhere to f .

• If p < 1, then there exists a subsequence fni that converges almost every-
where to f .
Proof. For p D 1 this follows directly from the definitions. For p < 1 choose
a sequence of integers 0 < n1 < n2 < n3 < such that kfni fniC1 kp < 2 i
for all i 2 N. Then the proof of Theorem 4.9 shows that fni converges almost
everywhere to an Lp -function g such that limn!1 kfn gkp D 0. Since the limit
is unique in Lp ./, it follows that g D f almost everywhere.
4.3 Separability
Definition 4.11. Let X be a topological space. A subset S X is said to be dense
(in X ) if its closure is equal to X or, equivalently, U \ S ¤ ; 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 D Rn with its standard topology is separable (Qn is a countable dense subset) and
hence is second countable. The next lemma gives a condition for a linear subspace
to be dense in Lp ./.
142 4. Lp spaces
Lemma 4.12. Let .X; A; / be a measure space and let 1 p < 1. Let X be
a linear subspace of Lp ./ such that ŒA 2 X for every measurable set A 2 A
with .A/ < 1. 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 D Lp ./.
Step 1. If s 2 Lp ./ is a measurable step function, then Œs 2 Y.

P`
Write s D i D1 ˛i A where ˛i 2 R n ¹0º and Ai D s 1
.˛i / 2 A. Then
i
Z Z
p p
j˛i j .Ai / D j˛i A j d jsjp d < 1
i
X X
and hence .Ai / < 1 for all i . This implies ŒA 2 Y for all i . Since Y is a
i
linear subspace of Lp ./, it follows that Œs 2 Y.
Step 2. If f 2 Lp ./ and f 0, then Œf 2 Y.
By Theorem 1.26, there is a sequence of measurable step functions si W X ! R such

that 0 s1 s2 and si converges pointwise to f . Then si 2 Lp ./ and
hence Œsi 2 Y for all i by Step 1. Moreover, jf si jp f p , f p is integrable, and
jf si jp converges pointwise to zero. Hence it follows from the Lebesgue dom-
inated convergence theorem (Theorem 1.45) that limi!1 kf si kp D 0. Since
Œsi 2 Y for all i and Y is a closed subspace of Lp ./, it follows that Œf 2 Y.
Step 3. Y D Lp ./.
Let f 2 Lp ./. Then f ˙ 2 Lp ./, hence Œf ˙ 2 Y by Step 2, and hence

Œf D Œf C Œf 2 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,
W B ! Œ0; 1 is a Borel measure, and fix a constant 1 p < 1.
Theorem 4.13. If X is second countable, then Lp ./ is separable.
Proof. See page 144.
Example 4.14. If X is an uncountable set with the discrete topology U D 2X

and W 2X ! Œ0; 1 is the counting measure, then X is not second countable and
Lp ./ D Lp ./ is not separable.
Theorem 4.15. Assume is outer regular and is inner regular on open sets. Define
² ˇ ³
ˇ s is a Borel measurable step function
Sc .X/ WD sW X ! R ˇˇ : (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 con-
tinues to hold when is a Radon measure.
Example 4.16. Let .X; U/ be the compact Hausdorff space constructed in

Example 3.6, let W A ! Œ0; 1 be the Dieudonné measure constructed in that
example, let ıW 2X ! Œ0; 1 be the Dirac measure at the point 2 X , and de-
fine 0 WD jB C ıjB W B ! Œ0; 2. Then Lp .0 / is a 2-dimensional vector space
0
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 D 1 is outer regular and is inner regular on open sets.
Let 0 W B ! Œ0; 1 be the unique Radon measure such that ƒ1 D ƒ0 . Then
Lp .1 / Lp .0 / and the linear map
Lp .1 / ! Lp .0 /W Œf 1 7 ! Œf 0 (4.14)
is a Banach space isometry.
Proof. Since 0 .B/ 1 .B/ for all B 2 B by Theorem 3.15, it follows that
Z Z
p
jf j d0 jf jp d1
X X
for every Borel measurable function f W X ! R. Hence Lp .1 / Lp .0 /.

We prove that
Z Z
jf jp d0 D jf jp d1 for all f 2 Lp .1 /: (4.15)
X X
Thus the map (4.14) is injective and has a closed image. To prove (4.15), define
E" WD ¹x 2 X j jf .x/j > "º for " > 0.

144 4. Lp spaces
Then 1 .E" / < 1 and hence 1 and 0 agree on all Borel subsets of E" by
Lemma 3.7. This implies
Z Z
jf jp d0 D jf jp d1 ;
E" E"
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 2 B such
that 0 .B/ < 1. By (3.5), there is a sequence of compact sets Ki B such that
Ki Ki C1 and 1 .Ki / D 0 .Ki / > 0 .B/ 2 i for all i . Define
[
A WD Ki B:
i2N
Then
1 .A/ D 0 .A/ D lim 0 .Ki / D 0 .B/:
i !1
This implies A 2 Lp .1 / and ŒB 0 D ŒA 0 2 X. By Lemma 4.12, it follows
that X D Lp .0 / and this proves Lemma 4.17.
Proof of Theorem 4.13. Let V U be a countable basis for the topology. Assume
without loss of generality that Vx is compact for all V 2 V. (If W U is any
countable basis for the topology, then the set V WD ¹V 2 W j Vx is compactº is also
a countable basis for the topology by Lemma A.3.) Choose a bijection
N ! VWi 7 ! Vi
and let
I WD ¹I N j #I < 1º
be the set of finite subsets of N. Then the map

X
I ! NWI 7 ! 2i 1
i2I
is a bijection, so the set I is countable. For I 2 I define

[
VI WD Vi :
i 2I
Define the set V Lp ./ by

° `
X ˇ ±
V WD s D ˛j V ˇ ` 2 N and ˛j 2 Q; Ij 2 I for j D 1; : : : ; ` :
ˇ
Ij
j D1
This set is contained in Lp ./ because Vx is compact for all V 2 V. It is countable

and its closure
X WD V x
in Lp ./ is a closed linear subspace. By Lemma 4.12, it suffices to prove that

ŒB 2 X for every B 2 B with .B/ < 1. To see this, fix a Borel set B 2 B
with .B/ < 1 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.
Therefore, there exist a compact set K X and an open set U X such that
K B U; .U n K/ < "p :
Define
I WD ¹i 2 N j Vi U º:
Since V is a basis of the topology, we have K U D i 2I Vi . Since K is compact,
S
there is a finite set I I such that
K VI U:
Next, since B V vanishes on X n .U n K/ and jB V j 1,

I I
kB V kp .U n K/1=p < ":

I
x
Since V 2 V and " > 0 was chosen arbitrary, it follows that ŒB 2 X D V.
I
Proof of Theorem 4.15. By Corollary 3.16 and Lemma 4.17, it suffices to consider
the case where is outer regular and is inner regular on open sets. Define
S WD ¹Œf 2 Lp ./ j for all " > 0, there exists s 2 Sc .X/

such that kf skp < "º;
and
C WD ¹Œf 2 Lp .1 / j for all " > 0, there exists g 2 Cc .X/
such that kf gkp < "º:
We must prove that Lp ./ D S D C. Since S and C are closed linear subspaces
of Lp ./, it suffices to prove that ŒB 2 S \ C for every Borel set B 2 B
with .B/ < 1 by Lemma 4.12. Let B 2 B with .B/ < 1 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 n K/ < "p . By Urysohn’s lemma (Theorem A.1), there
146 4. Lp spaces
is a function f 2 Cc .X/ such that 0 f 1, f jK 1, and supp.f / U .

This implies 0 f K U nK and 0 B K U nK : Hence,
kB K kp kU nK kp D .U n K/1=p < "
and likewise kf K kp < ". By Minkowski’s inequality (4.4), this implies
kB f kp kB K kp C kK f kp < 2":
This shows that ŒB 2 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 W B ! Œ0; 1 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 ƒ0 D ƒ1 D ƒ . 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 2 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, equivalently,
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/ D 1. Thus the quest for a counterexample can be
rephrased as follows.
Question. Do there exist a locally compact Hausdorff space .X; U/ and Borel mea-
sures
0 ; 1 ; W B ! Œ0; 1
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 2 B, and
0 D 0 .B/ < .B/ < 1 .B/ D 1
for some Borel set B 2 B?

4.4. Hilbert spaces 147
This leads to deep problems in set theory. A probability measure on a measur-

able space .X; A/ is a measure W A ! Œ0; 1 such that .X/ D 1. A measure
W A ! Œ0; 1 is called nonatomic if countable sets have measure zero. Now
consider the measure on X D R2 in Exercise 3.23 with 0 .R ¹0º/ D 0 and
1 .R ¹0º/ D 1, and define W R ! R2 by .x/ WD .x; 0/. If there is a nonatomic
probability measure W 2R ! Œ0; 1, then the measure 0 C provides a posi-
tive 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 gen-
eralized continuum hypothesis asserts that, if X is any set, then each subset of 2X
whose cardinality is strictly larger than that of X admits a bijection to 2X . It is in-
dependent of the other axioms of set theory and implies that nonatomic probability
measures W 2X ! Œ0; 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

This section introduces some elementary Hilbert space theory. It serves two pur-
poses. 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 Radon–Nikodým
theorem for -finite measure spaces in the next chapter.
Definition 4.19. Let H be a real vector space. A bilinear map
H H ! RW .x; y/ 7 ! hx; yi (4.16)
is called an inner product if it is symmetric, i.e., hx; yi D hy; xi for all x; y 2 H

and positive definite, i.e., hx; xi > 0 for all x 2 H n ¹0º. The norm associated to an
inner product (4.16) is the function
p
H ! RW x 7 ! kxk WD hx; xi: (4.17)
Lemma 4.20. Let H be a real vector space equipped with an inner product (4.16)
and the associated norm (4.17). The inner product and norm satisfy the Cauchy–
Schwarz inequality
jhx; yij kxk kyk (4.18)
and the triangle inequality
kx C yk kxk C kyk (4.19)
for all x; y 2 H . Thus (4.17) is a norm on H .

148 4. Lp spaces
Proof. The Cauchy–Schwarz inequality is obvious when x D 0 or y D 0.

Hence assume x ¤ 0 and y ¤ 0, and put WD kxk 1 x and WD kyk 1 y.
Then kk D kk D 1. Hence
0 k h; ik2 D h; h; ii D 1 h; i2 :
This implies jh; ij 1 and hence jhx; yij kxk kyk. In turn it follows from the
Cauchy–Schwarz inequality that
kx C yk2 D kxk2 C 2hx; yi C kyk2

kxk2 C 2 kxk kyk C kyk2
D .kxk C 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 WD L2 ./ is a Hilbert

space. The inner product is induced by the bilinear map
Z
L2 ./ L2 ./ ! RW .f; g/ 7 ! hf; gi WD fg d: (4.20)
X
It is well defined because the product of two L2 -functions f; gW X ! R is integrable

by (4.3) with p D q D 2. That it is bilinear follows from Theorem 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 ./ ! RW .Œf ; Œg / 7 ! hf; gi D fg 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 p
2 2
L ./ ! RW Œf 7 ! kf k2 D f d D hf; f i: (4.22)
X
This is the L2 -norm in (4.5) with p D 2. By Theorem 4.9, L2 ./ is complete 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
ƒW V ! R is called bounded if there exists a constant c 0 such that
jƒ.x/j c kxk for all x 2 V:
The norm of a bounded linear functional ƒW V ! R is the smallest such constant

c and will be denoted by
jƒ.x/j
kƒk WD sup : (4.23)
0¤x2V 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.)
Theorem 4.26 (Riesz). Let H be a Hilbert space and let ƒW H ! R be a bounded

linear functional. Then there exists a unique element y 2 H such that
ƒ.x/ D hy; xi for all x 2 H: (4.24)
This element y 2 H satisfies

jhy; xij
kyk D sup D kƒk : (4.25)
0¤x2H kxk
Thus the map H ! H W 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 2 E such that kx0 k kxk
for all x 2 E.
Theorem 4.27 implies Theorem 4.26. We prove existence. If ƒ D 0, then y D 0
satisfies (4.24). Hence assume ƒ ¤ 0 and define
E WD ¹x 2 H j ƒ.x/ D 1º:
Then E ¤ ; because there exists an element 2 H such that ƒ./ ¤ 0 and hence
x WD ƒ./ 1 2 E. The set E is a closed because ƒW H ! R is continuous,
150 4. Lp spaces
and it is convex because ƒ is linear. Hence Theorem 4.27 yields an element x0 2 E

such that
kx0 k kxk for all x 2 E:
We prove that
x 2 H; ƒ.x/ D 0 H) hx0 ; xi D 0: (4.26)
To see this, fix an element x 2 H such that ƒ.x/ D 0. Then x0 C tx 2 E for all
t 2 R. This implies
kx0 k2 kx0 C txk2 D kx0 k2 C 2thx0 ; xi C t 2 kxk2 for all t 2 R:
Thus the differentiable function t 7! kx0 C txk2 attains its minimum at t D 0,

and so its derivative vanishes at t D 0. Hence,
ˇ
d ˇˇ
0D kx0 C txk2 D 2hx0 ; xi;
dt ˇ tD0
and this proves (4.26).
Now put
x0
y WD :
kx0 k2
Fix an element x 2 H and put WD ƒ.x/. Then ƒ.x x0 / D ƒ.x/ D 0.
0 D hx0 ; x x0 i D hx0 ; xi kx0 k2 :
This implies
hx0 ; xi
hy; xi D D D ƒ.x/:
kx0 k2
Thus y satisfies (4.24).
We prove (4.25). Assume y 2 H satisfies (4.24). If y D 0, then ƒ D 0,
and so kyk D 0 D kƒk. Hence assume y ¤ 0. Then
kyk2 ƒ.y/ jƒ.x/j

kyk D D sup D kƒk :
kyk kyk 0¤x2H kxk
Conversely, it follows from the Cauchy–Schwarz inequality that
jƒ.x/j D jhy; xij kykkxk
for all x 2 H and hence kƒk kyk. This proves (4.25).

We prove uniqueness. Assume y; z 2 H satisfy

hy; xi D hz; xi D ƒ.x/
for all x 2 H . Then hy z; xi D 0 for all x 2 H . Take x WD y z to obtain
ky zk2 D hy z; y zi D 0;
and hence y z D 0. This proves Theorem 4.26, assuming Theorem 4.27.
ı WD inf¹kxk j x 2 Eº:
We prove uniqueness. Let x0 ; x1 2 E such that
kx0 k D kx1 k D ı:
Then 21 .x0 C x1 / 2 E because E is convex, and so kx0 C x1 k 2ı. Thus
kx0 x1 k2 D 2 kx0 k2 C 2 kx1 k2 kx0 C x1 k2 D 4ı 2 kx0 C x1 k2 0

and therefore x0 D x1 .
We prove existence. Choose a sequence xi 2 E such that

lim kxi k D ı:
i!1
We claim that xi is a Cauchy sequence. Indeed, fix a constant " > 0. Then there
exists an integer i0 2 N such that
"
i 2 N; i i0 H) kxi k2 < ı 2 C :
4
Let i; j 2 N such that i i0 and j i0 . Then 12 .xi C xj / 2 E because E is
convex, and hence kxi C xj k 2ı. This implies
kxi xj k2 D 2kxi k2 C 2kxj k2 kxi C xj k2
"
< 4 ı2 C 4ı 2
4
D ":
Thus xi is a Cauchy sequence. Since H is complete, the limit
x0 WD lim xi
i !1
exists. Moreover, x0 2 E because E is closed and kx0 k D ı because the norm

function (4.17) is continuous. This proves Theorem 4.27.
152 4. Lp spaces
Corollary 4.28. Let .X; A; / be a measure space and let ƒW L2 ./ ! R be a

bounded linear functional. Then there exists a function g 2 L2 ./, unique up to
equality almost everywhere, such that
Z
ƒ.Œf / D fg d for all f 2 L2 ./:
X
Moreover kƒk D 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 equa-
tion (4.8) (for 1 p < 1) and in Definition 4.7 (for p D 1). When 1 < p < 1
it turns out that the dual space of Lp ./ is always isomorphic to Lq ./, where
1=p C 1=q D 1. For p D 1 the natural homomorphism L1 ./ ! L1 ./ is
an isometric embedding, however, in most cases the dual space of L1 ./ is much
larger than L1 ./. For p D 1 the situation is more subtle. The natural homo-
morphism L1 ./ ! 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 2 A such that
1
[
XD Xi ; Xi Xi C1 ; .Xi / < 1 for all i 2 N: (4.27)
iD1
It is called semi-finite if every measurable set A 2 A satisfies
.A/ > 0 H) there exists E 2 A such that E A (4.28)

and 0 < .E/ < 1:
It is called localizable if it is semi-finite and, for every collection of measurable sets

E A, there is a set H 2 A satisfying the following two conditions:
(L1) .E n H / D 0 for all E 2 E;

(L2) if G 2 A satisfies .E n G/ D 0 for all E 2 E, then .H n G/ D 0.
A measurable set H satisfying (L1) and (L2) is called an envelope of E.

4.5. The dual space of Lp ./ 153
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 2 E up to a set of measure zero and, if any other mea-
surable set G covers each set E 2 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/ D sup¹.E/ j E 2 A; E A; .E/ < 1º (4.29)
for every measurable set A 2 A;
(ii) if .X; A; / is -finite, then it is semi-finite.
Proof. (i) Assume .X; A; / is semi-finite, let A 2 A, and define
a WD sup¹.E/ j E 2 A; E A; .E/ < 1º:
Then a .A/ and we must prove that a D .A/. This is obvious when a D 1.
Hence assume a < 1. Choose a sequence of measurable sets Ei A such that
.Ei / < 1 and .Ei / > a 2 i for all i . Define
1
[ 1
[
Bi WD E1 [ [ Ei ; B WD Bi D Ei :
i D1 iD1
Then Bi 2 A, Ei Bi A, and .Bi / < 1. Hence .Ei / .Bi / a for all

i 2 N, and hence
.B/ D lim .Bi / D a < 1:
i !1
If .A n B/ > 0, then since .X; A; / is semi-finite, there exists a measurable set
F 2 A such that F A n B and 0 < .F / < 1, and hence
B [ F A; a < .B [ F / D .B/ C .F / < 1;
contradicting the definition of a. This shows that .A n B/ D 0 and hence
.A/ D .B/ C .A n B/ D 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).
(ii) Assume that .X; A; / is -finite and choose a sequence of measurable sets
Xi 2 A that satisfies (4.27). If A 2 A, then it follows from Theorem 1.28 that
.A/ D limi!1 .A \ Xi /. Since .A \ Xi / < 1 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.
154 4. Lp spaces
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 WD ¹a; bº;
A WD 2X ;
.¹aº/ WD 1; .¹bº/ WD 1:
This measure space is not semi-finite. Thus the linear map L1 ./ ! L1 ./ in
Theorem 4.33 below is not injective. In fact, L1 ./ 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 W A ! Œ0; 1 be the count-
ing measure. Then .X; A; / is semi-finite, but it is not localizable. For example,
let H X be an uncountable set with an uncountable complement 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 L1 ./ ! L1 ./ is injective, while Theorem 4.35 below shows that it is not
surjective. An example of a bounded linear functional
ƒW L1 ./ ! R
that cannot be represented by an L1 -function is given by

X
ƒ.f / WD f .x/
x2H
for f 2 L1 ./ D L1 ./.
The next theorem is the first step towards understanding the dual space of Lp ./
and is a fairly easy consequence of the Hölder inequality. It asserts that for
1=p C 1=q D 1 every element of Lq ./ determines a bounded linear functional
on Lp ./ and that the resulting map Lq ./ ! Lp ./ is an isometric embedding
(for p D 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 < 1 (and for p D 1 under the localizable hypothesis) is the con-
tent of Theorem 4.35 below. This is a much deeper theorem whose proof for p ¤ 2
requires the Radon–Nikodým 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; 1 q 1; C D 1: (4.30)
p q
(i) Let g 2 Lq ./. Then the formula

Z
ƒg .Œf / WD fg d for f 2 Lp ./ (4.31)
X
defines a bounded linear functional
ƒg W Lp ./ ! R
and ˇZ ˇ
ˇ ˇ
ˇ fg dˇ
ˇ ˇ
X
kƒg k D sup kgkq : (4.32)
f 2Lp ./; kf kp ¤0 kf kp
(ii) The map g 7! ƒg in (4.31) descends to a bounded linear operator
Lq ./ ! Lp ./ W Œg 7 ! ƒg : (4.33)
(iii) Assume 1 < p 1 Then
kƒg k D kgkq for all g 2 Lq ./.
(iv) Assume p D 1. Then the map L1 ./ ! 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 Minkowski and Calderón–Zygmund inequalities in
Theorems 7.19 and 7.43 below.
156 4. Lp spaces
Lemma 4.34. Let .X; A; / be a measure space and let p; q be as in (4.30). Let
gW X ! Œ0; 1 be a measurable function and suppose that there exists a constant
c 0 such that
Z
f 2 L ./; f 0 H)
p
fg d c kf kp : (4.34)
X
(i) if q D 1, then kgk1 c;
(ii) if 1 < q < 1 and kgkq < 1, then kgkq c;
(iii) if 1 < q < 1 and .X; A; / is semi-finite, then kgkq c;
(iv) if q D 1 and .X; A; / is semi-finite, then kgk1 c.
Proof. (i) If q D 1 take f 1 in (4.34) to obtain kgk1 c.

(ii) Assume 1 < q < 1 and kgkq < 1. Then it follows from Lemma 1.47 that
the set
A WD ¹x 2 X j g.x/ D 1º
has measure zero. Define the function
hW X ! Œ0; 1/
by
´
g.x/ for x 2 X n A,
h.x/ WD
0 for x 2 A.
Then h is measurable and
Z Z
khkq D kgkq < 1; f h d D fg d c kf kp
X X
for all f 2 Lp ./ with f 0 by Lemma 1.48. Define
f W X ! Œ0; 1/
by
f .x/ WD h.x/q 1
for x 2 X .
Then
f p D hp.q 1/
D hq D f h
and hence
Z 1 1=q Z
kf kp D q
h d D khkqq 1 ; f h d D khkqq :
X X
Thus f 2 Lp ./ and so

Z
khkqq D f h d c kf kp D c khkqq 1
:
X
Since khkq < 1, it follows that kgkq D khkq c, and this proves part (ii).
(iii) Assume .X; A; / is semi-finite and 1 < q < 1. Suppose, by contra-
diction, that kgkq > c. We will prove that there exists a measurable function
hW X ! Œ0; 1/ such that
0 h g; c < khkq < 1: (4.35)
By (4.34), this function h satisfies

Z Z
f h d fg d c kf kp
X X
for all f 2 L ./ with f 0. Since khkq < 1 it follows from part (ii) that
p
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 Def-
inition 1.34 that Rthere exists a measurable step function sW X ! Œ0; 1/ such that
0 s g and X s q d > c q . If kskq < 1, take h WD s. If kskq D 1, there
exist a measurable set A X and a constant ı > 0 such that .A/ D 1 and
ıA s g. Since .X; A; / is semi-finite, Lemma 4.30 asserts that there exists
a measurable set E 2 A such that E A and c q < ı q .E/ < 1. Then the
function h WD ıE W X ! Œ0; 1/ satisfies 0 h g and khkq D ı.E/1=q > c,
as required. This proves part (iii).
(iv) Let q D 1 and assume .X; A; / is semi-finite. Suppose, by contradiction,
that kgk1 > c. Then there exists a constant ı > 0 such that the set
A WD ¹x 2 X j g.x/ c C ıº
has positive measure. Since .X; A; / is semi-finite there exists a measurable

R E A such that 0 < .E/ < 1. Hence, f WD E 2 L ./ and we get
1
set
X fg d .c C ı/.E/ > c.E/ D c kf k1 , in contradiction to (4.34). This
proves (iv) and Lemma 4.34.
158 4. Lp spaces
Proof of Theorem 4.33. The proof has four steps.

Step 1. Let f 2 Lp ./, g 2 Lq ./. Then fg 2 L1 ./ and kfgk1 kf kp kgkq .
R
If 1 < p < 1, then X jfgj d kf kp kgkq by the Hölder inequality (4.3).
If p D 1, then jfgj jf j kgk1 almost everywhere by Lemma 4.8, so fg 2 L1 ./
and kfgk1 kf k1 kgk1 . If p D 1, 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 de-
pends only on the equivalence class of f under equality almost
everywhere. Hence
ƒg is well defined. It is linear by Theorem 1.44 and ƒg kgkq by Step 1.
This proves (i). The bounded linear functional ƒg W 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/ WD 1 for t < 0, and sign.0/ D 0. If f 2 Lp ./ is nonnegative, then the
function f sign.g/W X ! R is p-integrable and
Z
f jgj d D ƒg .f sign.g// kƒg k kf sign.g/kp kƒg k kf kp :
X
Hence kgkq kƒg k by Lemma 4.34 and so kƒg k D kgkq by Step 2.

Step 4. If the map L1 ./ ! L1 ./ is injective, then .X; A; / is semi-finite.
Let A 2 A such that .A/ > 0 and define g WD A . Then ƒg W L1 ./ ! R is
nonzero by assumption. Hence there is an f 2 L1 ./ such that
Z Z
0 < ƒg .f / D fg d D f d: (4.36)
X A
For i 2 N define Ei WD ¹x 2 A j f .x/ > 2 i º. Then Ei 2 A, Ei A, and

Z
i
.Ei / 2 f d 2i kf k1 < 1:
Ei
S1
Moreover, E WD iD1 Ei D ¹x 2 A j f .x/ > 0º is not a null set by (4.36). Hence
one of the sets Ei has positive measure. Thus .X; A; / is semi-finite. This proves
Step 4 and Theorem 4.33.
The next theorem asserts that, for 1 < p < 1, every bounded linear functional
on Lp ./ has the form (4.31) for some g 2 Lq ./. For p ¤ 2 this is a much
deeper result than Corollary 4.28. The proof requires the Radon–Nikodým 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; 1 < q 1; C D 1:
p q
(i) Assume 1 < p < 1. Then the map Lq ./ ! Lp ./ W Œg 7! ƒg ; defined
by (4.31) is bijective and hence is a Banach space isometry.
(ii) Assume p D 1. Then the map L1 ./ ! L1 ./ W Œg 7! ƒg 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 D 1 (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 ./ ! L1 ./ , however, the dual space of L1 ./ is typically much larger
than L1 ./.
Example 4.36. Let W 2N ! Œ0; 1 be the counting measure on the positive inte-
gers. Then
`1 WD L1 ./ D L1 ./
is the Banach space of bounded sequences x D .xn /n2N of real numbers, equipped
with the supremum norm
kxk1 WD sup jxn j:
n2N
1
An interesting closed subspace of ` is
c WD ¹x D .xn /n2N 2 `1 j x is a Cauchy sequenceº:
It is equipped with a bounded linear functional ƒ0 W c ! R, defined by
ƒ0 .x/ WD lim xn for x D .xn /n2N 2 c:

n!1
160 4. Lp spaces
The Hahn–Banach theorem, one of the fundamental principles of functional anal-

ysis, asserts that every bounded linear functional on a linear subspace 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
ƒW `1 ! R such that ƒjc D ƒ0 . This linear functional cannot have the form (4.31)
for any g 2 L1 ./. To see this, note that `1 WD L1 ./ D L1 ./ is the space of
summable sequences of real numbers. Let y D .yn /n2N 2 `1 be a sequence of real
numbers such that 1
P
nD1 jyn j < 1, and define the linear functional
ƒy W `1 ! R
by
1
X
ƒy .x/ WD xn y n for x D .xn /n2N 2 `1 :
nD1
P1
Choose N 2 N such that nDN jyn j DW ˛ < 1 and define x D .xn /n2N 2 c by
xn WD 0 for n < N and xn WD 1 for n N . Then ƒy .x/ ˛ < 1 D ƒ.x/ and
hence ƒy ¤ ƒ. This shows that ƒ does not belong to the image of the isometric
inclusion `1 ,! .`1 / .
Exercise 4.37. Let ƒ0 W c ! R be the functional in Example 4.36 and denote

c0 WD ker ƒ0 . Thus c0 is the set of all sequences of real numbers that converge
to zero, i.e.,
c0 D ¹x D .xn /n2N 2 `1 j lim xn D 0º:
n!1
1
Prove that c0 is a closed linear subspace of ` and that `1 is naturally isomorphic
to the dual space of c0 . Thus
`1 Š .c0 / ; c0 ¨ `1 Š .`1 / Š .c0 / ; `1 ¨ .`1 / Š .`1 / :
In the language of functional analysis, this means that the Banach spaces c0 and `1
are not reflexive, and neither is `1 .
We close this section with two results that will be needed in the proof of The-
orem 4.35. The first asserts that every bounded linear functional on Lp ./ can be
written as the difference of two positive bounded linear functionals (Theorem 4.39).
The second asserts that every positive bounded linear functional on Lp ./ is sup-
ported on a -finite subset of X (Theorem 4.40). When ƒW Lp ./ ! R is a
bounded linear functional it will be convenient to abuse notation and write
ƒ.f / WD ƒ.Œf / for f 2 Lp ./.

Definition 4.38. Let .X; A; / be a measure space and let 1 p < 1. A bounded
linear functional ƒW Lp ./ ! R is called positive if
f 0 H) ƒ.f / 0
for all f 2 Lp ./.
Theorem 4.39. Let .X; A; / be a measure space and let 1 p < 1. Consider a
bounded linear functional ƒW Lp ./ ! R. Define
˙ W A ! Œ0; 1
by
˙ .A/ WD sup¹ƒ.˙E / j E 2 A; E A; .E/ < 1º (4.37)
Then the maps ˙ are measures, Lp ./ L1 .C / \ L1 . /, and the formulas
Z
˙
ƒ .f / WD f d˙ for f 2 Lp ./ (4.38)
X
define positive bounded linear functionals ƒ˙ W Lp ./ ! R such that
ƒ D ƒC ƒ ; kƒC k C kƒ k D kƒk: (4.39)
Proof. The proof is divided into four steps.
Step 1. The maps ˙ W A ! Œ0; 1 in (4.37) are measures.
It follows directly from the definition that ˙ .;/ D 0. We must prove that C is
-additive. That is then also -additive follows by reversing the sign of ƒ.
Thus, let Ai 2 A be a sequence of pairwise disjoint measurable sets and set
1
[
A WD Ai :
iD1
Let E 2 A such that E A and .E/ < 1. Then it follows from the definition
of C that
ƒ.E \A / C .Ai / for all i 2 N: (4.40)
i
Moreover, the sequence of measurable functions

n
X
fn WD E E \A 0
i
iD1
162 4. Lp spaces
converges pointwise to zero and satisfies 0 fnp E for all n. Since .E/ < 1,
the function E is integrable, and so it follows from the Lebesgue dominated con-
vergence theorem (Theorem 1.45) that
Z
lim fnp d D 0;
n!1 X
i.e.,
n
X
lim E E \A D 0:

n!1 i p
i D1
Then (4.40) yields that

n
X 1
X 1
X
ƒ.E / D lim ƒ.E \A / D ƒ.E \A / C .Ai /:
n!1 i i
iD1 i D1 i D1
Take the supremum over all E 2 A with E A and .E/ < 1 to obtain
1
X
C
.A/ C .Ai /:
iD1
To prove the converse inequality, assume first that C .Ai / D 1 for some i ; since
Ai A, this implies C .A/ D 1 D 1 C
P
i D1 .Ai /. Hence it suffices to assume
C
.Ai / < 1 for all i . Fix a constant " > 0 and choose a sequence of measurable
sets Ei 2 A such that Ei Ai and ƒ.E / > C .Ai / 2 i " for all i . Since
i
E1 [ [ En A, it follows from the definition of C that
n
X n
X
C .A/ ƒ.E [[En
/D ƒ.E / > C .Ai / ":
1 i
i D1 iD1
Take the limit n ! 1 to obtain

1
X
C
.A/ C .Ai / " for all " > 0,
iD1
so
1
X
C .A/ C .Ai /;
i D1
as claimed. Thus C is -additive and this proves Step 1.

Step 2. Let c WD kƒk. Then every measurable function f W X ! R satisfies

Z Z
jf j dC C jf j d c kf kp : (4.41)
X X
In particular, Lp ./ L1 .C / \ L1 . /.

Assume first that f D sW X ! Œ0; 1/ is a measurable step function in Lp ./.
Then there exist real numbers ˛i > 0 and measurable sets Ai 2 A for i D 1; : : : ; `
such that Ai \ Aj D ; for i ¤ j , .Ai / < 1 for all i , and s D ìD1 ˛i A :
P
i
Now fix a real number " > 0 and choose "i > 0 such that
`
X "
˛i "i D :
2
i D1
For i D 1; : : : ; ` choose Ei˙ 2 A such that
Ei˙ Ai ; ƒ.E C / C .Ai / "i ; ƒ.E / .Ai / "i :

i i
Then
Z Z `
X
C
s d C s d D ˛i .C .Ai / C .Ai //
X X iD1
`
X
˛i .ƒ.E C / ƒ.E / C 2"i /
i i
iD1
`
X
Dƒ ˛i .E C E / C "
i i
iD1
X`
c ˛i .E C E / C "

i i p
iD1
`
X 1=p
Dc ˛ip ..EiC n Ei / C .Ei n EiC // C"
iD1
`
X 1=p
c ˛ip .Ai / C"
iD1
D c kskp C ":
164 4. Lp spaces
Take the limit " ! 0 to obtain (4.41) for f D s. To prove (4.41) in general it suffices
to assume that f 2 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 2 L1 ./.
Hence
lim kf sn kp D 0
n!1
by the Lebesgue dominated convergence theorem (Theorem 1.45), and

Z Z
lim sn d˙ D f d˙
n!1 X X
by the Lebesgue monotone convergence theorem (Theorem 1.37). This proves

the inequality (4.41). It follows from (4.41) that Lp ./ L1 .C / \ L1 . / and
this proves Step 2.
Step 3. If A 2 A and .A/ < 1, then
˙ .A/ < 1; ƒ.A / D C .A/ .A/ (4.42)
It follows from the inequality (4.41) in Step 2 that

Z Z
C C
.A/ C .A/ D A d C A d ckA kp D c.A/1=p < 1:
X X
Now let " > 0 and choose E 2 A such that E A and ƒ.E / > C .A/ ".
Since ƒ.AnE / .A/, this implies
ƒ.A / D ƒ.E / C ƒ.AnE / > C .A/ .A/ ":
Since this holds for all " > 0, we obtain ƒ.A / C .A/ .A/. Reversing the
sign of ƒ we also obtain ƒ.A / .A/ C .A/ and this proves Step 3.
Step 4. If f 2 Lp ./, then
Z Z
ƒ.f / D f dC f d : (4.43)
X X
Let sW X ! R be a p-integrable step function. Then there are real numbers ˛i

and measurable sets Ai 2 A for i D 1; : : : ; ` such that .Ai / < 1 for all i and
s D ìD1 ˛i A . Hence it follows from Step 3 that
P
i
`
X `
X Z Z
C C
ƒ.s/ D ˛i ƒ.A / D ˛i . .Ai / .Ai // D s d s d :
i
iD1 iD1 X X
This proves (4.43) for p-integrable step functions. Now let f 2 Lp ./ and as-
sume 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 2 L1 ./. Hence
lim kf sn kp D 0
n!1
by the Lebesgue dominated convergence theorem and so

lim ƒ.sn / D ƒ.f /:
n!1
Moreover, Z
f d˙ c kf kp < 1
X
by Step 2, and Z Z
lim sn d˙ D f d˙
n!1 X X
by the Lebesgue monotone convergence theorem. Thus every nonnegative Lp -func-
tion f W X ! Œ0; 1/ satisfies (4.43). If f 2 Lp ./, then f ˙ 2 Lp ./
satisfy (4.43) by what we have just proved, and hence so does f D f C f .
This proves Step 4.
It follows from Steps 2 and 4 that the linear functionals ƒ˙ W 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 and let 1 < p < 1. Consider a
positive bounded linear functional ƒW Lp ./ ! R. Define
.A/ WD sup¹ƒ.E / j E 2 A; E A; .E/ < 1º (4.44)
for A 2 A. Then the map W A ! Œ0; 1 is a measure, Lp ./ L1 ./, and
Z
ƒ.f / D f d for all f 2 Lp ./: (4.45)
X
Moreover, there are measurable sets N 2 A and Xn 2 A for n 2 N such that

1
[
X nN D Xn ; .N / D 0; .Xn / < 1; Xn XnC1 (4.46)
nD1
for all n 2 N.
Proof. That is a measure satisfying Lp ./ L1 ./ and (4.45) follows from
Theorem 4.39 and the fact that C D and D 0 because ƒ is positive.
Now define c WD kƒk. We prove in three steps that there exist measurable sets
N 2 A and Xn 2 A for n 2 N satisfying (4.46).
166 4. Lp spaces
Step 1. For every " > 0 there exist a measurable set A 2 A and a measurable
function
f W X ! Œ0; 1/
such that
f jX nA D 0; inf f > 0; kf kp D 1; ƒ.f / > c ": (4.47)

A
In particular,
p
.A/ .inf f / < 1:
A
Choose h 2 Lp ./ such that khkp D 1 and ƒ.h/ > c ". Assume without loss of
generality that h 0. (Otherwise replace h by jhj.) Define
Ai WD ¹x 2 X j h.x/ > 2 i º:
Then .h hA /p converges pointwise to zero as i ! 1 and is bounded by the

i
integrable function hp . Hence it follows from the Lebesgue dominated convergence
theorem (Theorem 1.45) that
lim kh hA kp D 0;
i !1 i
and therefore
lim ƒ.hA / D ƒ.h/ > c ":
i !1 i
Choose i 2 N such that

ƒ.hA / > c "
i
and define
hA
i
A WD Ai ; f WD :
khA kp
i
Then A and f satisfy (4.47) and so

Z
p
.A/ .inf f / f p d D .inf f / p
:
A X A
This proves Step 1.

Step 2. Let "; A; f be as in Step 1 and let E 2 A. Then
ƒ.E / c
.E \ A D ;; .E/ < 1/ H) < "1=q C1 ; (4.48)
.E/1=p p
where 1 < q < 1 is chosen such that 1=p C 1=q D 1.
Define
" 1=p
g WD f C E :
.E/
Then
Z 1=p
p
kgkp D f d C " D .1 C "/1=p
X
and, by (4.47),
" 1=p ƒ.E /
ƒ.g/ D ƒ.f / C ƒ.E / > c " C "1=p :
.E/ .E/1=p
Since ƒ.g/ c kgkp , it follows that
ƒ.E /
c " C "1=p < c.1 C "/1=p :
.E/1=p
Further since .1 C "/1=p 1 "=p for all " 0, this implies
ƒ.E / c
"1=p < c..1 C "/1=p 1/ C " " C1 :
.E/1=p p
Finally since "1 1=p

D "1=q , Step 2 is proved.
Step 3. There exist measurable sets N; X1 ; X2 ; X3 ; : : : satisfying (4.46).
Choose An 2 A and fn 2 Lp ./ as in Step 1 with " D 1=n. For n 2 N, set

1
[ 1
[
Xn WD A1 [ [ An ; N WD X n An D X n Xn :
nD1 nD1
By Step 2, every measurable set E N with .E/ < 1 satisfies
ƒ.E / 1 c
< C 1
.E/1=p n1=q p
for all n 2 N, and hence ƒ.E / D 0. This implies .N / D 0 by (4.44). More-
over, .Xn / niD1 .Ai / < 1 for every n by Step 1. This proves Step 3 and
P
Theorem 4.40.
168 4. Lp spaces
4.6 Exercises
Many of the exercises in this section are taken from Rudin [17, pages 71–75].
Exercise 4.41. Let .X; A; / be a measure space and let
f D .f1 ; : : : ; fn /W X ! Rn
be a measurable function such that
Z
jfi j d < 1
X
for i D 1; : : : ; n. Define
Z Z Z
f d WD f1 d; : : : ; fn d 2 Rn :
X X X
n n
Let R ! Œ0; 1/W v 7! kvk ; be any norm on R . Prove that the function
X ! Œ0; 1/W x 7 ! kf .x/k
is integrable and Z Z

f d kf k d: (4.49)

X X
Hint. First prove the inequality first for vector-valued integrable step functions
sW X ! Rn . Then show that forR all " > 0 there is a vector-valued integrable step
n
R
function sW X ! R 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/ D 1. Let
f 2 L1 ./, and let W R ! R be convex. Prove Jensen’s inequality
Z Z
f d . ı f / d: (4.50)
X X
(In particular, show that ı f is necessarily integrable, so the right-hand side is

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

n n n

X Y X
i D 1 H) ai i i ai (4.52)
iD1 i D1 iD1
for all positive real numbers i and ai . In particular, ab ap =p C b q =q for all

positive real numbers a; b; p; q such that 1=p C 1=q D 1.
4.6. Exercises 169
Exercise 4.43. Let .X; A; / be a measure space, choose p; q; r 2 Œ1; 1 such that
1 1 1
C D ;
p q r
and let f 2 Lp ./ and g 2 Lq ./. Prove that fg 2 Lr ./ and
kfgkr kf kp kgkq : (4.53)
Exercise 4.44. Let .X; A; / be a measure space, choose real numbers
1 r < p < s < 1;
and let 0 < < 1 such that

1 1
C D :
r s p
Prove that every measurable function f W X ! R satisfies the inequality
kf kp kf kr kf k1s
: (4.54)
Deduce that Lr ./ \ Ls ./ Lp ./.

Exercise 4.45. Let .X; A; / be a measure space and let f W X ! R be a measurable
function. Define
If WD ¹p 2 R j 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 ¤ ; prove that
lim kf kp D kf k1 :
p!1
Exercise 4.46. For each of the following three conditions find an example of mea-
sure space .X; A; / that satisfies it for all p; q 2 Œ1; 1.
(a) If p < q, then Lp ./ ¨ Lq ./.

(b) If p < q, then Lq ./ ¨ Lp ./.
(c) If p ¤ q, then Lp ./ 6 Lq ./ and Lq ./ 6 Lq ./.
170 4. Lp spaces
Exercise 4.47. Let .X; U/ be a locally compact Hausdorff space and define
8 ˇ 9
< ˇ f is continuous and =
ˇ
C0 .X/ WD f W X ! R ˇ for all " > 0, there exists K X such that :
ˇ
XnK jf j < "
: ˇ K is compact and sup ;
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/ D 1 and let
f; gW X ! Œ0; 1 be measurable functions such that fg 1. Prove that
kf k1 kgk1 1:
Exercise 4.49. Let .X; A; / be a measure space such that .X/ D 1 and let
f W X ! Œ0; 1 be a measurable function. Prove that
q Z q
1 C kf k21 1 C f 2 d 1 C 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 D Œ0; 1 and f D F 0 is the derivative of a
continuously differentiable function F W Œ0; 1 ! R. Under which conditions does
equality hold in either of the two inequalities in (4.55)?
Exercise 4.50. Let .X; A; / be

R a measure space and let f W X ! R be a measurable
function such that f > 0 and X f d D 1. Let E X be a measurable set such
that 0 < .E/ < 1. Prove that
Z 1
log.f / d .E/ log (4.56)
E .E/
and Z
f p d .E/1 p
for 0 < p < 1: (4.57)
E
Exercise 4.51. Let f W Œ0; 1 ! .0; 1/ 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
4.6. Exercises 171
Exercise 4.52. Fix two constants 1 0.
(i) Let f W .0; 1/ ! R be Lebesgue measurable and suppose that the function
.0; 1/ ! RW x 7! x p 1 a jf .x/jp ; is integrable. Show that the restriction of
f to each interval .0; x is integrable and prove Hardy’s inequality
Z 1
ˇZ x
ˇp 1=p Z 1 1=p
1 a p
xp 1 a p
ˇ ˇ
x ˇ f .t/ dt ˇ dx
ˇ jf .x/j dx :
0
ˇ
0 a 0
(4.59)
Show that equality holds in (4.59) if and only if f D 0 almost everywhere.
Hint. Assume first that f is nonnegative with compact support and define
1 x
Z
F .x/ WD f .t/ dt for x > 0.
x 0
Use integration by parts to obtain
Z 1
p 1 p
Z
p 1 a p 1 a
x F .x/ dx D x F .x/p 1
f .x/ dx:
0 a 0
Use Hölder’s inequality.
(ii) Show that the constant p=a in Hardy’s inequality is sharp. Hint. Choose
< 1 a=p and take f .x/ WD x for x 1 and f .x/ WD 0 for x > 1.
(iii) Prove that every sequence .an /n2N of positive real numbers satisfies
1 N 1
X 1 X p p p X p
an a : (4.60)
N nD1 p 1 nD1 n
N D1
Hint. If an is nonincreasing, then (4.60) follows from (4.59) with a D p 1

for a suitable function f . Deduce the general case from the special case.
(iv) Let f W .0; 1/ ! R be Lebesgue measurable and suppose that the function
.0; 1/ ! RW x 7! x p 1Ca jf .x/jp ; is integrable. Show that the restriction
of f to each interval Œx; 1/ is integrable and prove the inequality
Z 1
ˇZ 1
ˇp 1=p Z 1 1=p
a 1 p
xp 1Ca p
ˇ ˇ
x ˇ f .t/ dt ˇ dx
ˇ jf .x/j dx :
0
ˇ
x a 0
(4.61)
2 1
Hint. Apply the inequality (4.59) to the function g.x/ WD x f .x /.
172 4. Lp spaces
Exercise 4.53. Let .X; U/ be a locally compact Hausdorff space and let
W B ! Œ0; 1 be an outer regular Borel measure on X that is inner regular on
open sets. Let g 2 L1 ./. Prove that the following are equivalent:
(i) the function g vanishes -almost everywhere;

R
(ii) X fg d D 0 for all f 2 Cc .X/.
Hint. Assume (ii). Let K X be compact. Use Urysohn’s lemma (Theorem A.1)
to show that there is a sequence fn 2 CcR.X/ such that 0 fn 1 and Rfn converges
almost everywhere to K . DeduceR that K g d D 0. Then prove that U g d D 0
for every open set U X and B g d D 0 for all B 2 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.)
Exercise 4.54. Egoroff’s theorem. Let .X; A; / be a measure space such

that .X/ < 1 and let fn W X ! R be a sequence of measurable functions that
converges pointwise to f W X ! R. Fix a constant " > 0. Then there exists a
measurable set E 2 A such that
.X n E/ < "
and fn jE converges uniformly to f jE .

Hint. Define
S.k; n/ WD ¹x 2 X j jfi .x/ fj .x/j < 1=k, for all i; j > nº for k; n 2 N:
Prove that
lim .S.k; n// D .X / for all k 2 N:
n!1
T
Deduce that there is a sequence nk 2 N such that E WD k2N S.k; nk / satisfies
the required conditions. Show that Egoroff’s theorem does not extend to -finite
measure spaces.
Exercise 4.55. Let .X; A; / be a measure space and let 1 < p < 1. Let
f 2 Lp ./ and let fn 2 Lp ./ be a sequence such that
lim kfn kp D kf kp
n!1
and fn converges to f almost everywhere. Prove that
lim kf fn kp D 0:
n!1
4.6. Exercises 173
Prove that the hypothesis

lim kfn kp D kf kp
n!1
cannot be removed.
Hint 1. Fix a constant " > 0. Use Egoroff’s
R theorem to construct disjoint mea-
surable sets A; B 2 A such that X D A [ B, A jf jp d < ", .B/ < 1, and fn
R to fp uniformly on B. Use Fatou’s lemma (Theorem 1.41) to prove that
converges
lim sup A jfn j d < ".
n!1
Hint 2. Let gn WD 2p 1 .jfn jp C jf jp / jf fn jp and use Fatou’s lemma
(Theorem 1.41) as in the proof of the Lebesgue dominated convergence theorem
(Theorem 1.45).
Exercise 4.56. Let .X; A; / be a measure space, let fn W X ! R be a sequence of

measurable functions, and let f W X ! R be a measurable function. The sequence
fn is said to converge in measure to f if
lim .¹x 2 X j jfn .x/ f .x/j > "º/ D 0

n!1
for all " > 0. (On page 53 this is called convergence in probability.) Assume
.X/ < 1 and prove the following.
(i) If fn converges to f almost everywhere, then fn converges to f in measure.

Hint. See page 53.
(ii) If fn converges to f in measure, then a subsequence of fn converges to f

almost everywhere.
(iii) If 1 p 1 and fn ; f 2 Lp ./ satisfy
lim kfn f kp D 0;
n!1
then fn converges to f in measure.
Exercise 4.57. Let .X; U/ be a compact Hausdorff space and W B ! Œ0; 1 be a

Borel measure. Let C.X/ D Cc .X/ be the space of continuous real-valued func-
tions 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 2 X such that every open
neighborhood U of x0 satisfies .U \ E/ > 0 and .U n E/ > 0;
(c) is outer regular and is inner regular on open sets.

174 4. Lp spaces

(i) Assume (a). Then the map C.X/ ! L1 ./ in (b) is an isometric embedding
and hence its image is a closed linear subspace of L1 ./.
(ii) Assume (a) and (b). Then there is a nonzero bounded linear functional
ƒW L1 ./ ! R
that vanishes on the image of the inclusion C./ ! L1 ./. Hint. If f D E
almost everywhere, then f is discontinuous at x0 .
(iii) Assume (a), (b), and (c). Then the isometric embedding L1 ./ ! L1 ./
of Theorem 4.33 is not surjective. Hint. Use part (ii) and Exercise 4.53.
(iv) The Lebesgue measure on Œ0; 1 satisfies (a), (b), and (c).
Exercise 4.58. Prove that every -finite measure space .X; A; / is localizable.
Hint. Assume first that .X/ < 1. Let E A and define
c WD sup¹.E1 [ [ En / j n 2 N; E1 ; : : : ; En 2 Eº:
Show that there is a sequence Ei 2 E such that . 1
S
S1 iD1 Ei / D c. Prove that
H WD i D1 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 W Af ! R, each defined on
a measurable set Af 2 A. Suppose that any two functions f1 ; f2 2 F agree
almost everywhere on Af1 \ Af2 . Then there exists a measurable function
gW X ! R such that gjAf D f almost everywhere for all f 2 F.
We will see in the next chapter that condition (F) is equivalent to localizability
for semi-finite measure spaces. Hint. Let F be a collection of measurable functions
as in (F). For a 2 R and f 2 F define
Afa WD ¹x 2 Af j f .x/ < aº:
For q 2 Q, let H q 2 A be an envelope of the collection
Eq WD ¹Afq j f 2 Fº:
Define the measurable sets
[
X a WD H q; a 2 R:
q2Q
q<a
4.6. Exercises 175
(i) If a < b, then X a X b .
(ii) For every a 2 R the measurable set X a is an envelope of the collection
Ea WD ¹Afa j f 2 Fº:
Thus .Afa n X a / D 0 for all f 2 F and, if G 2 A, then
.Afa n G/ D 0 for all f 2 F H) .X a n G/ D 0:
(iii) If a 2 R and E 2 A, then
.Afa \ E/ D 0 for all f 2 F H) .X a \ E/ D 0:
(iv) .X a \ Af n Afa / D 0 for all f 2 F and all a 2 R.
(v) Define \ [
E0 WD Xr [ X n Xs :
r2R s2R
Then E0 is measurable and .Af \ E0 / D 0 for all f 2 F.
(vi) For f 2 F define the measurable set Ef Af by

[ q
.X q \ Af n Afq /:
[
Ef WD .Af \ E0 / [ .Af n X q / [
q2Q q2Q
Then .Ef / D 0.
(vii) Define gW X ! R by
´
0 if x 2 E0 ;
g.x/ WD (4.62)
a if x 2 X s for all s > a and x … X r for all r < a:
Then g is well defined and measurable and g D f on Af n Ef for all f 2 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 Stone–Čech compactification
X of the natural numbers. This is a compact 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 .
176 4. Lp spaces
The space C.X/ of continuous functions on X can be naturally identified with the
space `1 . Hence the space of positive bounded linear functionals on `1 is isomor-
phic to the space of Radon measures on X by Theorem 3.15. Thus the Stone–Čech
compactification of N can be used to understand the dual space of `1 . 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 ! .`1 / W n 7 ! ƒn
which assigns to each natural number n 2 N the bounded linear functional
ƒn W `1 ! R
defined by
ƒn ./ WD n for D .i /i 2N 2 `1 .
This functional has norm one. Now the space of all bounded linear functionals on
`1 of norm at most one, i.e., the unit ball in .`1 / , is compact with respect to the
weak- topology by the Banach–Alaoglu theorem. Define X to be the closure of
the set ¹ƒn j n 2 Nº in .`1 / with respect to the weak- topology. Thus
ˇ For all finite sequences c 1 ; : : : ; c ` 2 R
8 ˇ 9
ˆ >
ˇ and 1 D .i1 /i2N ; : : : ; ` D .i` /i2N 2 `1>
ˆ
ˆ ˇ >
ˆ
< >
=
X WD ƒ 2 .`1 / ˇˇ satisfying ƒ. j / < c j for j D 1; : : : ; `;
ˇ
:
ˇ there exists an n 2 N such that
ˆ
ˆ ˇ >
>
ˆ
ˆ >
>
: ˇ j < c j for j D 1; : : : ; ` ;
n
The weak- topology U 2X is the smallest topology such that the map
f W X ! R; f .ƒ/ WD ƒ./;
is continuous for each 2 `1 . The topological space .X; U/ is a separable compact

Hausdorff space, called the Stone–Čech 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 ƒn is not -compact. The only continuous functions on X are those
of the form f , so the map
`1 ! C.X/W 7 ! f
is a Banach space isometry. (Verify that kf k WD sup jf .ƒ/j D kk1 for all
ƒ2X
2 `1 .) Thus the dual space of `1 can be understood in terms of the Borel
measures on X.
4.6. Exercises 177
By Theorem 3.18, every Radon measure on X is regular. However, the Borel

-algebra B 2X does carry -finite measures W B ! Œ0; 1 that are inner
regular, but not outer regular (and must necessarily satisfy .X/ D 1). Here is an
example pointed out to me by Theo Buehler. Define
X 1
.B/ WD
n
n2N
ƒn 2B
for every Borel set B X. This measure is -finite and inner regular, but not
outer regular. (The set U WD ¹ƒn j n 2 Nº is open, its complement K WD X n 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 Radon–Nikodým theorem
Recall from Theorem 1.40 that every measurable function f W X ! Œ0; 1/ on a

measure space .X; A; / determines a measure f W A ! Œ0; 1 defined by
Z
f .A/ WD f d for A 2 A.
A
By Theorem 1.35 it satisfies f .A/ D 0 whenever .A/ D 0. A measure with this
property is called absolutely continuous with respect to . The Radon–Nikodým
theorem asserts that, when is -finite, every -finite measure that is absolutely
continuous with respect to has the form f for some measurable function
f W X ! Œ0; 1/. It was proved by Johann Radon in 1913 for the Lebesgue measure
space and extended by Otton Nikodým in 1930 to general -finite measure spaces.
A proof is given in Section 5.1. Consequences of the Radon–Nikodým theorem in-
clude 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 Radon–Nikodým 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 W A ! Œ0; 1/ is
absolutely continuous with respect to if
.A/ D 0 H) .A/ D 0 for all A 2 A.
is singular with respect to if there exists a measurable set A such that
.A/ D 0 and .Ac / D 0:
In this case we also say that and are mutually singular. We write

if is absolutely continuous with respect to and ? if and are mutually
singular.
180 5. The Radon–Nikodým theorem
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 C 2 ? ;
(ii) if 1 and 2 , then 1 C 2 ;
(iii) if 1 and 2 ? , then 1 ? 2 ;
(iv) if and ? , then D 0.
Proof. (i) Suppose that 1 ? and 2 ? . Then there exist measurable sets
Ai 2 A such that i .Ai / D 0 and .Aci / D 0 for i D 1; 2. Define A WD A1 \ A2 .
Then Ac D Ac1 [ Ac2 is a null set for and A is a null set for both 1 and 2 , and
hence also for 1 C 2 . Thus 1 C 2 ? and this proves (i).
(ii) Suppose that 1 and 2 . If A 2 A satisfies .A/ D 0,
then 1 .A/ D 2 .A/ D 0 and so .1 C 2 /.A/ D 1 .A/ C 2 .A/ D 0. Thus
1 C 2 and this proves (ii).
(iii) Suppose that 1 and 2 ? . Since 2 ? , there exists a measurable
set A 2 A such that 2 .A/ D 0 and .Ac / D 0. Since 1 it follows that
1 .Ac / D 0 and hence 1 ? 2 .This proves (iii).
(iv) Suppose that and ? . Since ? , there exists a measurable set
A 2 A such that .A/ D 0 and .Ac / D 0. Since , it follows that .Ac / D 0
and hence .X/ D .A/ C .Ac / D 0. This proves (iv) and Lemma 5.2.
Theorem 5.3 (Lebesgue decomposition theorem). Let .X; A; / be a -finite mea-
sure space and let be a -finite measure on A. Then there exist unique measures
a ; s W A ! Œ0; 1 such that
D a C s ; a ; s ? : (5.1)
Theorem 5.4 (Radon–Nikodým). Let .X; A; / be a -finite measure space and let
W A ! Œ0; 1 be a measure. The following are equivalent:
(i) is -finite and absolutely continuous with respect to ;
(ii) there exists a measurable function f W X ! Œ0; 1/ such that
Z
.A/ D f d for all A 2 A: (5.2)
A
If (i) holds, then equation (5.2) determines f uniquely up to equality -almost

everywhere. Moreover, f 2 L1 ./ if and only if .X/ < 1.
5.1. Absolutely continuous measures 181
Proof. The last assertion follows by taking A D X in (5.2).

We prove that (ii) H) (i). Thus assume that there exists a measurable function
f W X ! Œ0; 1/ 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 of
measurable sets X1 X2 X3 such that .Xn / < 1 and X D 1
S
nD1 Xn .
Define
An WD ¹x 2 Xn j f .x/ nº:
Then An AnC1 and .An / n.Xn / < 1 for all n and X D 1
S
nD1 An .
Thus is -finite and this shows that (ii) H) (i).
It remains to prove that (i) H) (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/ < 1 and
.X/ < 1, and the last step establishes existence in general.
Step 1. Let .X; A; / be a measure space, let W A ! Œ0; 1 be a -finite measure,

and let f; gW X ! Œ0; 1/ be two measurable functions such that
Z Z
.A/ D f d D g d for all A 2 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 measurable

sets A1 A2 A3 such that .An / < 1 for all n 2 N and X D 1
S
nD1 An .
For n 2 N put
An WD ¹E 2 A j E An º; n WD jAn :
Take A D An in (5.3) to obtain f; g 2 L1 .n / for all n. Thus
Z
f g 2 L .n /;
1
.f g/ dn D 0 for all E 2 An :
E
Hence f g vanishes n -almost everywhere by Lemma 1.49. Thus the set
En WD ¹x 2 An j f .x/ ¤ g.x/º
satisfies .En / D n .En / D 0, and hence the set

1
[
E WD ¹x 2 X j f .x/ ¤ g.x/º D En
nD1
satisfies .E/ D 0. This proves Step 1.

Step 2. Let .X; A/ be a measurable space and let ; W A ! Œ0; 1 be measures

such that .X/ < 1, .X/ < 1, and . Then there exists a measurable
function
hW X ! Œ0; 1/
such that Z
.A/ D h d for all A 2 A.
A
By assumption, C W A ! Œ0; 1 is a finite measure defined by
. C /.A/ WD .A/ C .A/ for A 2 A:
Since . C /.X/ < 1, it follows from the Cauchy–Schwarz inequality that
H WD L2 . C / L1 . C /:
Namely, if f 2 L2 . C /, then
Z sZ
p
jf j d. C / c jf j2 d. C / < 1; c WD .X/ C .X/:
X X
Define
ƒW L2 . C / ! R
by Z
ƒ.f / WD f d:
X
for f 2 L2 .C/. (Here we abuse notation and use the same letter f for a function
in L2 . C / and its equivalence class in L2 . C /.) Then
Z Z
jƒ.f /j jf j d jf j d. C / c kf kL2 .C/
X X
for all f 2 L . C /. Thus ƒ is a bounded linear functional on L2 . C / and it

2
follows from Corollary 4.28 that there exists an L2 -function g 2 L2 . C / such

that Z Z
f d D fg d. C / (5.4)
X X
for all f 2 L2 . C /. This implies
Z Z Z
f .1 g/ d. C / D f d. C / fg d. C /
X X X
Z Z
D f d. C / f d (5.5)
X X
Z
D f d
X
for all f 2 L2 . C /.

We claim that the inequalities 0 g < 1 hold . C /-almost everywhere.

To see this, consider the measurable sets
E0 WD ¹x 2 X j g.x/ < 0º; E1 WD ¹x 2 X j g.x/ 1º:
Then it follows from (5.4) with f WD E that

0
Z Z
0 .E0 / D E d D E g d. C / 0:
0 0
X X
Hence
Z
E g d. C / D 0
0
X
and Lemma 1.49 shows that the function f WD E g vanishes . C /-almost

0
everywhere. Hence . C /.E0 / D 0. Likewise, (5.5) with f WD E yields that
1
Z Z
.E1 / D E d D .1 g/ d. C / 0:
1
X E1
Hence .E1 / D 0. Since is absolutely continuous with respect to , it follows

that .E1 / D 0 and hence . C /.E1 / D 0 as claimed. Assume from now on that
0 g.x/ < 1 for all x 2 X . (Namely, redefine g.x/ WD 0 for x 2 E0 [ E1 without
changing the identities (5.4) and (5.5).)
Apply identity (5.5) to the characteristic function f WD A 2 L2 . C / of a
measurable set A to obtain the identity
Z
.A/ D .1 g/ d. C / for all A 2 A:
A
By Theorem 1.40 this implies that (5.5) continues to hold for every measurable
function f W X ! Œ0; 1/, whether or not it belongs to L2 . C /. Now define the
measurable function
hW X ! Œ0; 1/
by
g.x/
h.x/ WD for x 2 X:
1 g.x/
By (5.4) with f D A and (5.5) with f D A h it satisfies

Z
.A/ D A d
X
Z
D A g d. C /
X
Z
D A h.1 g/ d. C /
X
Z
D A h d
X
Z
D h d
A
for all A 2 A. This proves Step 2.
Step 3. We prove that (i) H) (ii).
Since and are -finite measures, there exist sequences of measurable sets
An ; Bn 2 A such that An AnC1 , .An / < 1, Bn BnC1 , .Bn / < 1
for all n, and X D 1
S S1
nD1 An D nD1 Bn . Define Xn WD An \ Bn . Then
Xn XnC1 ; .Xn / < 1; .Xn / < 1
for all n, and X D 1

S
nD1 Xn . Thus, by Step 2, there exists a sequence of measurable
functions fn W Xn ! Œ0; 1/ such that
Z
.A/ D fn d for all n 2 N and all A 2 A such that A Xn : (5.6)
A
It follows from Step 1 that the restriction of fnC1 to Xn agrees with fn -almost
everywhere. Thus, modifying fnC1 on a set of measure zero if necessary, we may
assume without loss of generality that fnC1 jXn D fn for all n 2 N. With this
understood, define
f W X ! Œ0; 1/
by
f jXn WD fn for n 2 N:
This function is measurable because

1
[ 1
[
1 1
f .Œ0; c/ D .Xn \ f .Œ0; c// D fn 1 .Œ0; c/ 2 A
nD1 nD1
for all c 0. Now let E 2 A and define En WD E \ Xn 2 A for n 2 N. Then

1
[
E1 E2 E3 ; ED En :
nD1
Hence it follows from part (iv) of Theorem 1.28 that

.E/ D lim .En /
n!1
Z
D lim f d
n!1 E
n
Z
D lim En f d
n!1 X
Z
D E f d
X
Z
D f d:
E
Here the last but one equality follows from the Lebesgue monotone convergence
theorem (Theorem 1.37). This proves Step 3 and Theorem 5.4.
Example 5.5. Let X be a one-element set and let A WD 2X . Define the measure
W 2X ! Œ0; 1 by .;/ WD 0 and .X/ WD 1.
(i) Choose .;/ WD 0 and .X/ WD 1. Then , R but there does not exist a
(measurable) function f W X ! Œ0; 1 such that X f d D .X/. Thus the
hypothesis that .X; A; / is -finite cannot be removed in Theorem 5.4.
R
(ii) Choose WD . Then we get .A/ D A f d for every nonzero func-
tion f W X ! Œ0; 1/. 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 ; ; W A ! Œ0; 1 by
´
0 if A is countable;
.A/ WD .A/ WD #.A \ H /; .A/ WD #A:
1 if Ac is countable;
Then and and are not -finite.

R There does not exist any measurable
function f W X ! Œ0; 1 such that .X/R D X f d. Nor is there any measurable
function hW X ! R such that .A/ D A h d for all A 2 A. (The only possible
such function would be h WD H which is not measurable.)
Proof of Theorem 5.3. We prove uniqueness. Let a ; s ; 0a ; 0s W A ! Œ0; 1 be
measures such that
D a C s D 0a C 0s ; a ; 0a ; s ? ; 0s ? :
Then there exist measurable sets A; A0 2 A such that
s .A/ D 0; .X n A/ D 0; 0s .A0 / D 0; .X n A0 / D 0:
Since X n .A \ A0 / D .X n A/ [ .X n A0 /, this implies
.X n .A \ A0 // D 0:
Let E 2 A. Then
s .E \ A \ A0 / D 0 D 0s .E \ A \ A0 /
and hence
a .E \ A \ A0 / D .E \ A \ A0 / D 0a .E \ A \ A0 /:
Moreover .E n .A \ A0 // D 0, hence
a .E n .A \ A0 // D 0 D 0a .E n .A \ A0 //
and so
s .E n .A \ A0 // D .E n .A \ A0 // D 0s .E n .A \ A0 //:
This implies
a .E/ D a .E \ A \ A0 / D 0a .E \ A \ A0 / D 0a .E/
s .E/ D s .E n .A \ A0 // D 0s .E n .A \ A0 // D 0s .E/:
This proves uniqueness.
We prove existence. The measure
WD C W A ! Œ0; 1
is -finite. Hence it follows from the Radon–Nikodým theorem (Theorem 5.4) that
there exist measurable functions f; gW X ! Œ0; 1/ such that
Z Z
.E/ D f d; .E/ D g d for all E 2 A: (5.7)
E E
5.2. The dual space of Lp ./ revisited 187
Define
A WD ¹x 2 X j g.x/ > 0º (5.8)
and
a .E/ WD .E \ A/; s .E/ WD .E \ Ac / for E 2 A: (5.9)
Then it follows directly from (5.9) that the maps a ; s W A ! Œ0; 1 are measures
and satisfy a C s D . Moreover, by (5.9),
s .A/ D .A \ Ac / D .;/ D 0
and from (5.8) that gjAc D 0, so by (5.7)

Z
.Ac / D g d D 0:
Ac
This shows that s ? . It remains to prove that a is absolutely continuous with

respect to . To see this, let E 2 A such that .E/ D 0. Then by (5.7)
Z Z
E g d D g d D .E/ D 0:
X E
Hence it follows from Lemma 1.49 that E g vanishes -almost everywhere.

Thus E \A g D A E g vanishes -almost everywhere. Since g.x/ > 0 for all
x 2 E \ A, this implies
.E \ A/ D 0:
Hence Z
a .E/ D .E \ A/ D f d D 0:
E \A
This shows that a and completes the proof of Theorem 5.3.
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; 1 < q 1; C D 1: (5.10)
p q
As in Section 4.5 we abuse notation and write
ƒ.f / WD ƒ.Œf /
for the value of a bounded linear functional ƒW Lp ./ ! R on the equivalence

class of a function f 2 Lp ./. Recall from Theorem 4.33 that every g 2 Lq ./
determines a bounded linear functional
ƒg W Lp ./ ! R
via Z
ƒg .f / WD fg d for f 2 Lp ./:
X
The next result proves Theorem 4.35 in -finite case.
Theorem 5.7. Assume .X; A; / is -finite and let ƒW Lp ./ ! R be a bounded
linear functional. Then there exists a function g 2 Lq ./ such that
ƒg D ƒ:
Proof. Assume first that ƒ is positive. We prove in six steps that there exists a
function g 2 Lq ./ such that g 0 and ƒg D ƒ.
Step 1. Define
.A/ WD sup¹ƒ.E / j E 2 A; E A; .E/ < 1º (5.11)
for A 2 A. Then the map W A ! Œ0; 1 is a measure, Lp ./ L1 ./, and

Z
ƒ.f / D f d for all f 2 Lp ./.
X
This follows directly from Theorem 4.40.

Step 2. Let be as in Step 1 and define c WD kƒk. Then .A/ c.A/1=p for all
A 2 A.
By assumption, ƒ.f / c kf kp for all f 2 Lp ./. Take f WD E to obtain
ƒ.E / c.E/1=p c.A/1=p for all E 2 A with E A and .E/ < 1.
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. Then there exists a measurable function
gW X ! Œ0; 1/
such that Z
.A/ D g d for all A 2 A.
A
By Step 2, is -finite and . Hence Step 3 follows from the Radon–Nikodým

theorem (Theorem 5.4) for -finite measure spaces.
Step 4. Let be as in Step 1 and g be as in Step 3. Then
Z Z
fg d D f d
X X
for every measurable function f W X ! Œ0; 1/.

This follows immediately 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 2 Lp ./ such that f 0. Then

Z Z
Step 4 Step 1
fg d D f d D ƒ.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
ƒ D ƒg :
Since g 2 Lq ./ by Step 5, the function g determines a bounded linear functional

ƒg W L ./ ! R via ƒg .f / WD X fg d for f 2 Lp ./. By (5.12), it satisfies
p
R
ƒg .f / D ƒ.f / for all f 2 Lp ./ with f 0. Apply this identity to the functions
f ˙ W X ! Œ0; 1/ for all f 2 Lp ./ to obtain ƒ D ƒg . This proves the assertion
of Theorem 5.7 for every positive bounded linear functional ƒW Lp ./ ! R.
Let ƒW Lp ./ ! R be any bounded linear functional. By Theorem 4.39 there
exist positive bounded linear functionals ƒ˙ W Lp ./ ! R such that ƒ D ƒC ƒ .
Hence, by what we have just proved, there exist functions g ˙ 2 Lq ./ such that
g ˙ 0 and ƒ˙ D ƒg ˙ . Define
g WD g C g :
Then g 2 Lq ./ and
ƒg D ƒg C ƒg D ƒC ƒ D ƒ:

The next result proves Theorem 4.35 in the case p D 1.

Theorem 5.8. Assume p D 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 Exer-
cise 4.59, i.e., if F is a collection of measurable functions f W Af ! R, each
defined on a measurable set Af 2 A, such that any two functions f1 ; f2 2 F
agree almost everywhere on Af1 \ Af2 , then there exists a measurable func-
tion gW X ! R such that gjAf D f almost everywhere for all f 2 F;
(iii) the linear map
L1 ./ ! L1 ./ W g 7 ! ƒg (5.13)
is bijective.
Proof. The proof that (i) H) (ii) is outlined in Exercise 4.59.

We prove that (ii) H) (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 first that
ƒW L1 ./ ! R is a positive bounded linear functional. Define
E WD ¹E 2 A j .E/ < 1º
and, for E 2 E, define
AE WD ¹A 2 A j A Eº; E WD jAE : (5.14)
Then there is an extension operator
E W L1 .E / ! L1 ./
defined by
´
f .x/ for x 2 E;
E .f /.x/ WD (5.15)
0 for x 2 X n E:
It descends to a bounded linear operator from L1 .E / to L1 ./ which will still be
denoted by E . Define
ƒE D ƒ ı E W L1 .E / ! R:
This is a positive bounded linear functional for every E 2 E. Hence it follows

from Theorem 5.7 (and the Axiom of Choice) that there is a collection of bounded
measurable functions
gE W E ! Œ0; 1/; E 2 E;
such that
Z
ƒE .f / D fgE dE for all E 2 E and all f 2 L1 .E /:
E
If E; F 2 E, then E \ F 2 E and the functions gE jE \F , gF jE \F , and gE \F all

represent the same bounded linear functional ƒE \F W L1 .E \F / ! R. Hence they
agree almost everywhere by Theorem 4.33. This shows that the collection
F WD ¹gE j E 2 Eº
satisfies the hypotheses of condition (F) on page 174. Thus it follows from (ii) that
there exists a measurable function gW X ! R such that, for all E 2 E, the restriction
gjE agrees with gE almost everywhere on E.
We claim that g 0 almost everywhere. Suppose on the contrary that the set
A WD ¹x 2 X j g.x/ < 0º
has positive measure. Since .X; A; / is semi-finite, there exists a set E 2 E such
that E A and .E/ > 0. Since g.x/ < 0 gE .x/ for all x 2 E it follows that
gjE does not agree with gE almost everywhere, a contradiction. This shows that
g 0 almost everywhere.
We prove that g kƒk almost everywhere. Suppose otherwise that the set
AC WD ¹x 2 X j g.x/ > kƒkº
has positive measure. Since .X; A; / is semi-finite, there exists a set E 2 E such
that E AC and .E/ > 0. Since kgE k1 D kƒE k kƒk, it follows from
Lemma 4.8 that gE .x/ kƒk < g.x/ for almost every x 2 E. Hence gjE does
not agree with gE almost everywhere, a contradiction. This shows that g kƒk
almost everywhere and we may assume without loss of generality that
0 g.x/ kƒk
for all x 2 X.
We prove that ƒg D ƒ. Fix a function f 2 L1 ./ such that f 0. Then there
exists a sequence Ei 2 E such that
E1 E2 E3
and E f converges pointwise to f . Namely, by Theorem 1.26, there exists a

i
sequence of measurable step functions
si W X ! Œ0; 1/
such that
0 s1 s2
and si converges pointwise to f . Since

Z Z
si d f d < 1 for all i ,
X X
the sets
Ei WD ¹x 2 X j si .x/ > 0º
have finite measure and
0 si E f f for all i .
i
Thus the Ei are as required. Since the sequence jf E f j converges point-

i
wise to zero and is bounded above by the integrable function f , it follows from
the Lebesgue dominated convergence theorem (Theorem 1.45) that
lim kf E f k1 D 0:
i!1 i
Hence
ƒ.f / D lim ƒ.E f /

i !1 i
D lim ƒEi .f jEi /

i !1
Z
D lim fgEi d
i !1 E
i
Z
D lim fg d
i !1 E
i
Z
D fg d:
X
Here the last step follows from the Lebesgue monotone convergence theorem
(Theorem 1.37). This shows that ƒ.f / D ƒg .f / for every nonnegative integrable
function f W X ! Œ0; 1/. Consequently,
ƒ.f / D ƒ.f C / ƒ.f / D ƒg .f C / ƒg .f / D ƒg .f /
for all f 2 L1 ./. Thus ƒ D ƒg as claimed.

This shows that every positive bounded linear functional on L1 ./ belongs 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 Theorem 4.39, it follows
that the map (5.13) is surjective. Thus we have proved that (ii) H) (iii).
We prove that (iii) H) (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 col-
lection of measurable sets. Assume without loss of generality that
E1 ; : : : ; E` 2 E H) E1 [ [ E` 2 E:
(Otherwise, replace E by the collection E0 of all finite unions of elements of E;

then every measurable envelope of E0 is also an envelope of E.) For E 2 E define
AE and E by (5.14) and define the bounded linear functional
ƒE W L1 .E / ! R
by Z
ƒE .f / WD f dE for f 2 L1 .E /: (5.16)
E
Then for all E; F 2 A and f 2 L1 ./
E F; f 0 H) ƒE .f / ƒF .f /: (5.17)
Define
ƒW L1 ./ ! R
by
ƒ.f / WD sup ƒE .f C jE / sup ƒE .f jE /: (5.18)
E 2E E 2E
We prove that this is a well defined bounded linear functional with kƒk 1. To see
this, note that Z
ƒE .f jE / f d
X
for every nonnegative function f 2 L1 ./, and so

Z Z
jƒ.f /j f C d C f d D kf k1 for all f 2 L1 ./.
X X
Moreover, it follows directly from the definition that ƒ.cf / D cƒ.f / for all c 0
and ƒ. f / D ƒ.f /. Now let f; g 2 L1 ./ be nonnegative integrable functions.
Then
ƒ.f C g/ D sup ƒE .f jE C gjE /

E 2E
sup ƒE .f jE / C sup ƒE .gjE /

E 2E E 2E
D ƒ.f / C ƒ.g/:
To prove the converse inequality, let " > 0 and choose E; F 2 E such that
ƒE .f jE / > ƒ.f / "; ƒF .gjF / > ƒ.g/ ":
Then E [ F 2 E and it follows from (5.17) that
ƒE [F ..f C g/jE [F / D ƒE [F .f jE [F / C ƒE [F .gjE [F /

ƒE .f jE / C ƒF .gjF /
> ƒ.f / C ƒ.g/ 2":
Hence
ƒ.f C g/ > ƒ.f / C ƒ.g/ 2" for all " > 0
and so
ƒ.f / C ƒ.g/ ƒ.f C g/ ƒ.f / C ƒ.g/:
This shows that
ƒ.f C g/ D ƒ.f / C ƒ.g/
for all f; g 2 L1 ./ such that f; g 0. If f; g 2 L1 ./, then
.f C g/C C f C g D .f C g/ C f C C g C ;
and hence
ƒ..f C g/C / C ƒ.f / C ƒ.g / D ƒ..f C g/ / C ƒ.f C / C ƒ.g C /
by what we have just proved. Since ƒ.f / D ƒ.f C / ƒ.f / by definition,

it follows that
ƒ.f C g/ D ƒ.f / C ƒ.g/ for all f; g 2 L1 ./.
This shows that ƒW L1 ! R is a positive bounded linear functional of norm

kƒk 1.
With this understood, it follows from (iii) that there exists a function g 2 L1 ./
such that ƒ D ƒg . Define
H WD ¹x 2 X j g.x/ > 0º:
We prove that H is an envelope of E. Fix a set E 2 E and suppose, by contradiction,

that .E n H / > 0. Then, since .X; A; / is semi-finite, there exists a measurable
set A 2 A such that 0 < .A/ < 1 and A E n H . Since g.x/ 0 for all x 2 A,
it follows that Z
0 < .A/ D ƒE .A jE / D g d 0;
A
a contradiction. This shows that our assumption .E n H / > 0 was wrong. Hence
.E n H / D 0 for all E 2 E, as claimed.
Now let G 2 A be any measurable set such that .E n G/ D 0 for all E 2 E.
We must prove that .H n G/ D 0. Suppose, by contradiction, that .H n G/ > 0.
Since .X; A; / is semi-finite there exists a measurable set A 2 A such that
0 < .A/ < 1 and A H n G. Then
Z
g d D ƒ.A / D sup ƒE .A jE / D sup .E \ A/ D 0:
A E 2E E 2E
Here the second equality follows from (5.18), the third follows from (5.16), and the
last follows from the fact that E \ A E n G for all E 2 E. Since g > 0 on A,
it follows from Lemma 1.49 that .A/ D 0, a contradiction. This shows that our
assumption that .H n G/ > 0 was wrong and so .H n G/ D 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 D 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 ./ W g 7! ƒg is surjective. Let ƒW Lp ./ ! R
be a positive bounded linear functional and define
.A/ WD sup¹ƒ.E / j E 2 A; E A; .E/ < 1º for A 2 A:
Then W A ! Œ0; 1 is a measure by Theorem 4.40 and

Z
L ./ L ./; ƒ.f / D
p 1
f d for all f 2 Lp ./:
X
Theorem 4.40 also asserts that there exists a measurable set N 2 A such that
.N / D 0 and the restriction of to X n N is -finite. Define
X0 WD X n N; A0 WD ¹A 2 A j A X0 º; 0 WD jA0
as in (5.14), let 0 W Lp .0 / ! Lp ./ be the extension operator as in (5.15), and

define
ƒ0 WD ƒ ı 0 W Lp .0 / ! R:
Then ƒ0 is a positive bounded linear functional on Lp .0 / and
Z Z
ƒ.f / D f d D f d D ƒ0 .f jX0 / for all f 2 Lp ./:
X X nN
Since .X0 ; A0 ; 0 / is -finite, Theorem 5.7 shows that there exists a function
g0 2 Lq .0 / such that g0 0 and
Z
ƒ0 .f0 / D f0 g0 d0 for all f0 2 Lp .0 /:
X0
Define
gW X ! Œ0; 1/
by
´
g0 .x/ for x 2 X0 D X n N ,
g.x/ WD
0 for x 2 N .
Then, by Theorem 4.33,
kgkLq ./ D kg0 kLq .0 / D kƒ0 k D kƒk
and, for all f 2 Lp ./,

Z Z
ƒ.f / D ƒ0 .f jX0 / D fg0 d0 D fg d:
X0 X
This proves the assertion for positive bounded linear functionals. Since every
bounded linear functional ƒW 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 L1 ./ ! L1 ./ W g 7! ƒg 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.
5.3. Signed measures 197
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 W A ! R is called a signed measure if it is -additive,
i.e., for every sequence Ei 2 A of pairwise disjoint measurable sets, it holds that
X1 [ 1 X 1
j.Ei /j < 1; Ei D .Ei /: (5.19)
iD1 i D1 iD1
Lemma 5.11. Every signed measure W A ! R satisfies the following:

(1) .;/ D 0;
(2) if E1 ; : : : ; E` 2 A are pairwise disjoint, then
`
[ X`
Ei D .Ei /:
i D1 i D1
Proof. To prove (i) take Ei WD ; in (5.19). To prove (ii) take Ei WD ; for all i > `.

Given a signed measure W A ! R, it is a natural question to ask whether it can
be written as the difference of two measures ˙ W A ! Œ0; 1/. Closely related to
this is the question whether there exists a measure W A ! Œ0; 1/ that satisfies
j.A/j .A/ for all A 2 A: (5.20)
If such a measure exists, it must satisfy
.E; F 2 A; E \ F D ;/ H) .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 W A ! R be a signed measure and define
ˇ E; F 2 A; =
8 ˇ 9
<
for A 2 A:
ˇ
jj.A/ WD sup .E/ .F / ˇˇ E \ F D ;; (5.21)
: ˇ E [F DA;
Then j.A/j jj.A/ < 1 for all A 2 A and jjW A ! Œ0; 1/ is a measure,
called the total variation of .
Proof. We prove that jj is a measure. If follows directly from the definition that
jj.;/ D 0 and jj.A/ j.A/j 0 for all A 2 A. We must prove that the
function jjW A ! Œ0; 1 is -additive. Let Ai 2 A be a sequence of pairwise
disjoint measurable sets and put
1
[
A WD Ai :
i D1
Let E; F 2 A are measurable sets such that
E \ F D ;; E [ F D A: (5.22)
Then
1
[ 1
[
ED .E \ Ai /; F D .F \ Ai /:
iD1 i D1
Hence
1
X 1
X
.E/ .F / D .E \ Ai / .F \ Ai /
iD1 i D1
1
X
D ..E \ Ai / .F \ Ai //
iD1
1
X
jj.Ai /:
iD1
Take the supremum over all pairs of measurable sets E; F satisfying (5.22) to obtain
1
X
jj.A/ jj.Ai / (5.23)
i D1
To prove the converse inequality, fix a constant " > 0. Then there are sequences of
measurable sets Ei ; Fi 2 A such that
"
Ei \ Fi D ;; Ei [ Fi D Ai ; .Ei / .Fi / > jj.Ai /
2i
for all i 2 N. The sets
1
[ 1
[
E WD Ei and F WD Fi
iD1 i D1
satisfy (5.22) and so

1
X 1
X
jj.A/ .E/ .F / D ..Ei / .Fi // > jj.Ai / ":
iD1 i D1
Hence
1
X
jj.A/ > jj.Ai / " for all " > 0.
iD1
Thus
1
X
jj.A/ jj.Ai /
iD1
and so
1
X
jj.A/ D jj.Ai /
iD1
by (5.23). This shows that jj is a measure.
It remains to prove that jj.X/ < 1. Suppose, by contradiction, that
jj.X/ D 1:
Claim. Let A 2 A such that jj.X n A/ D 1. Then there exists a measurable set
B 2 A such that A B, j.B n A/j 1, and jj.X n B/ D 1.
There exist measurable sets E; F such that E \ F D ;, E [ F D X n A, and
.E/ .F / 2 C j.X n A/j;
.E/ C .F / D .X n A/:

Take the sum, respectively the difference, of these (in)equalities to obtain
2.E/ 2 C j.X n A/j C .X n A/ 2;
2.F / .X n A/ 2 j.X n A/j 2;

and hence j.E/j 1 and j.F /j 1. Since jj.E/ C jj.F / D jj.X n A/ D 1,
it follows that jj.E/ D 1 or jj.F / D 1. If jj.E/ D 1 choose B WD A [ F ,
and if jj.F / D 1 choose B WD A [ E. This proves the claim.
It follows from the claim by induction that there exists a sequence of measurable
sets ; WD A0 A1 A2 such that j.An n An 1 /j 1 for all n 2 N. Hence
En WD An n An 1 is a sequence of pairwise disjoint measurable sets such that
P1
nD1 j.En /j D 1, in contradiction to Definition 5.10. This contradiction shows
that the assumption that jj.X/ D 1 was wrong. Hence jj.X/ < 1 and thus
jj.A/ < 1 for all A 2 A. This proves Theorem 5.12.
Definition 5.13. Let W A ! R be a signed measure and let jjW A ! Œ0; 1/ 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 jj. The measures
˙ W A ! Œ0; 1/
are defined by
jj.A/ ˙ .A/
˙ .A/ WD D sup¹˙.E/ j E 2 A; E Aº (5.24)
2
for A 2 A and they satisfy
C D ; C C D jj: (5.25)
Exercise 5.14. Let .X; A; / be a measure space, let f 2 L1 ./, and define
Z
.A/ WD f d for A 2 A.
A
Prove that is a signed measure and

Z Z
jj.A/ D jf j d; ˙ .A/ D f ˙ d for all A 2 A: (5.26)
A A
Definition 5.15. Let W A ! Œ0; 1 be a measure and let ; 1 ; 2 W A ! R be

signed measures.
(i) is called absolutely continuous with respect to (notation “ ”)
if .E/ D 0 implies .E/ D 0 for all E 2 A.
(ii) is called concentrated on A 2 A if .E/ D .E \ A/ for all E 2 A.
(iii) is called singular with respect to (notation “ ? ”) if there exists a
measurable set A such that .A/ D 0 and is concentrated on A.
(iv) 1 and 2 are called mutually singular (notation “1 ? 2 ”) if there are mea-
surable sets A1 ; A2 such that A1 \ A2 D ;, A1 [ A2 D X, and i is concen-
trated on Ai for i D 1; 2.
Lemma 5.16. Let be a measure on A and let ; 1 ; 2 be signed measures on A.
(i) if and only if jj ;
(ii) ? if and only if jj ? ;
(iii) 1 ? 2 if and only if j1 j ? j2 j.
Proof. The proof has four steps.
Step 1. Let A 2 A. Then jj.A/ D 0 if and only if .E/ D 0 for all measurable
sets E A.
If jj.A/ D 0, then j.E/j jj.E/ jj.A/ D 0 for all measurable sets E A.

The converse implication follows directly from the definition.
Step 2. is concentrated on A 2 A if and only if jj.X n A/ D 0.
The signed measure is concentrated on A if and only if .E/ D .E \ A/ for all
E 2 A, or equivalently .E n A/ D 0 for all E 2 A. By Step 1, this holds if and
only if jj.X n A/ D 0.
Step 3. We prove (i).
Assume . If E 2 A satisfies .E/ D 0, then every measurable set F 2 A

with F E satisfies .F / D 0 and so .F / D 0; hence jj.E/ D 0 by Step 1.
Thus jj . The converse follows from the fact that jj.
Step 4. We prove (ii) and (iii).
? if and only if there is a measurable sets A 2 A such that .A/ D 0 and

is concentrated on A. By Step 2, the latter holds if and only if jj.X n A/ D 0 or,
equivalently, jj ? . This proves (ii). Assertion (iii) follows from Step 2 by the
same argument and this proves Lemma 5.16.
Theorem 5.17 (Lebesgue decomposition). Let .X; A; / be a -finite measure

space and let W A ! R be a signed measure. Then there exists a unique pair
of signed measures
a ; s W A ! R
such that
D a C s ; a ; s ? : (5.27)
Proof. We prove existence. Let ˙ W A ! Œ0; 1/ be the measures defined by (5.24).

By Theorem 5.3, there exist measures ˙a W A ! Œ0; 1/ and s W A ! Œ0; 1/ such
˙
˙ ˙ ˙ ˙ ˙
that a , s ? , and D a C s . Hence the signed measures
a WD C
a a and s WD C
s s
satisfy (5.27).
We prove uniqueness. Assume D a C s D 0a C 0s where a ; s ; 0a ; 0s are

signed measures on A such that a ; 0a and s ; 0s ? . Then ja j; j0a j
and js j; j0s j ? by Lemma 5.16. This implies ja jCj0a j and js jCj0s j ?
by parts (i) and (ii) of Lemma 5.2. Moreover,
ja 0a j ja j C j0a j; j0a a j D js 0s j js j C j0s j:
Hence ja 0a j and ja 0a j ? by part (iii) of Lemma 5.2. Thus, by
part (iv) of Lemma 5.2, ja 0a j D 0, and therefore a D 0a and s D 0s . This
proves Theorem 5.17.
Theorem 5.18 (Radon–Nikodým). Let .X; A; / be a -finite measure space and

let W A ! R be a signed measure. Then if and only if there exists a
-integrable function f W X ! R such that
Z
.A/ D f d for all A 2 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 2 L1 ./, then by part (vi) of

Theorem 1.44. Conversely, assume and let jj; C ; W A ! Œ0; 1/ be the
measures defined by (5.21) and (5.24). Then jj by part (i) of Lemma 5.16,
and so ˙ . Hence it follows from TheoremR 5.4 that there exist -integrable
functions f ˙ W A ! Œ0; 1/ such that ˙ .A/ D A f ˙ d for all A 2 A. Hence
the function f WD f C f 2 L1 ./ satisfies (5.28). The uniqueness of f ,
up to equality -almost everywhere, follows from Lemma 1.49. This proves Theo-
rem 5.18.
Theorem 5.19 (Hahn decomposition). Let W A ! R be a signed measure. Then

there exists a measurable set P 2 A such that
.A \ P / 0; .A n P / 0 for all A 2 A: (5.29)
Moreover, there exists a measurable function hW X ! ¹1; 1º such that

Z
.A/ D h d jj for all A 2 A: (5.30)
A
Proof. By Theorem 5.12, the function WD jjW A ! Œ0; 1/ in (5.21) is a finite

measure and satisfies j.A/j .A/ for all A 2 A. Hence and it follows
from Theorem 5.18 that there exists a function h 2 L1 ./ such that (5.30) holds.
We prove that h.x/ 2 ¹1; 1º for -almost every x 2 X. To see this, fix a real
number 0 < r < 1 and set
Ar WD ¹x 2 X j jh.x/j rº:
If E; F 2 A such that E \ F D ; and E [ F D Ar , then

Z Z
.E/ .F / D h d h d
E F
Z Z
jhj d C jhj d
E F
r.Ar /:
Take the supremum over all pairs E; F 2 A such that E \ F D ; and E [ F D Ar

to obtain .Ar / r.Ar /, and hence .Ar / D 0. Since this holds for all r < 1,
it follows that jhj 1 -almost everywhere. Modifying h on a set of measure
zero, if necessary, we may assume without loss of generality that jh.x/j 1 for all
x 2 X. Define
P WD ¹x 2 X j h.x/ 1º;
N WD ¹x 2 X j h.x/ 1º:
Then P \ N D ;, P [ N D X, and
Z
.P / h d D .P / .P /;
P
Z
.N / .N / D h d .N /:
N
Hence
Z
.h 1/ d D .P / .P / D 0;
P
Z
.h C 1/ d D .N / C .N / D 0:
N
By Lemma 1.49 this implies h D 1 -almost everywhere on P and h D 1

-almost everywhere on N . Modify h again on a set of measure zero, if neces-
sary, to obtain h.x/ D 1 for all x 2 P and h.x/ D 1 for all x 2 N . This proves
Theorem 5.19.
Theorem 5.20 (Jordan decomposition). Let .X; A/ be a measurable space, let

W A ! R be a signed measure, and let ˙ W A ! Œ0; 1/ be finite measures such
that D C . Then the following are equivalent:
(i) C C D jj;
(ii) C ? ;
(iii) there exists a measurable set P 2 A such that C .A/ D .A \ P / and
.A/ D .A n P / for all A 2 A.
Moreover, for every signed measure , there is a unique pair of measures ˙

satisfying D C and these equivalent conditions.
Proof. (i) H) (ii). By Theorem 5.19, there exists a measurable function
hW X ! ¹˙1º
such that Z
.A/ D h d jj
A
for all A 2 A. Set

P WD ¹x 2 X j h.x/ D 1º:
Then it follows from (i) that
jj.P c / C .P c / 1Ch
Z
C .P c / D D d jj D 0;
2 Pc 2
jj.P / .P / 1 h
Z
.P / D D d jj D 0:
2 P 2
Hence C ? .
(ii) H) (iii). By (ii), there exists a measurable set P 2 A such that
C .P c / D 0 and .P / D 0:
Therefore,
C .A/ D C .A \ P / D C .A \ P / .A \ P / D .A \ P /;
.A/ D .A n P / D .A n P / C .A n P / D .A n P /
for all A 2 A.
5.4. Radon–Nikodým generalized 205
(iii) H) (i). Assume (iii) and fix a set A 2 A. Then
C .A/ C .A/ D .A \ P / .A n P / jj.A/:
Now choose E; F 2 A such that E \ F D ; and E [ F D A. Then
.E/ .F / D .E \ P / C .E n P / .F \ P / .F n P /

.E \ P / .E n P / C .F \ P / .F n P /
D .A \ P / .A n P /
D C .A/ C .A/:
Take the supremum over all such pairs E; F 2 A to obtain the inequality
jj.A/ C .A/ C .A/
for all A 2 A and hence jj D C C :

Thus we have proved that assertions (i), (ii), and (iii) are equivalent. Existence
and uniqueness of ˙ now follows from (i) with ˙ D 21 .jj ˙ /. This proves
Theorem 5.20.
5.4 Radon–Nikodým generalized
This section discusses an extension of the Radon–Nikodým theorem (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 continuity of with
respect to is not sufficient for obtaining the conclusion of the Radon–Nikodým
theorem and a stronger condition is needed. In [4, Theorem 232B] Fremlin intro-
duces the notion “truly continuous”, which is equivalent to “absolutely continuous”
whenever is -finite. In [8] König reformulates Fremlin’s criterion in terms of
“inner regularity of with respect to ”. We shall discuss both conditions below,
show that they are equivalent, and prove the generalized Radon–Nikodým theorem.
As a warmup we rephrase absolute continuity in the familiar "-ı language of anal-
ysis.
Standing assumption. Throughout this section .X; A; / is a measure space and

W 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 ı > 0 such that
A 2 A; .A/ < ı H) j.A/j < ":
Proof. That (ii) implies (i) is obvious. Conversely, assume (i). Then jj by
Lemma 5.16. Assume by contradiction that (ii) does not hold. Then there exist a
constant " > 0 and a sequence of measurable sets Ai 2 A such that
.Ai / 2 i ; j.Ai /j " for all i 2 N:
For n 2 N define
1
[ 1
\
Bn WD Ai ; B WD Bn :
iDn nD1
Then
1
Bn BnC1 ; .Bn / ; jj.Bn / jj.An / j.An /j "
2n 1
for all n 2 N. Hence .B/ D 0 and jj.B/ D limn!1 jj.Bn / " by part (v) of
Theorem 1.28. This contradicts the fact that jj and shows that our assumption
that (ii) does not hold was wrong. Thus (i) implies (ii) and this proves Lemma 5.21.

Definition 5.22. The signed measure is called truly continuous with respect to
if, for every " > 0, there exist a constant ı > 0 and a measurable set E 2 A such
that .E/ < 1 and
A 2 A; .A \ E/ < ı H) j.A/j < ": (5.31)
Lemma 5.23. The following are equivalent:
(i) is truly continuous with respect to ;
(ii) jj is truly continuous with respect to ;
(iii) C and are truly continuous with respect to .
Proof. Assume (i), fix a constant " > 0, and choose ı > 0 and E 2 A such
that .E/ < 1 and (5.31) holds. Let A 2 A such that .A \ E/ < ı. Then
.B/ .A n B/ < 2" for every measurable set B A, and hence jj.A/ 2"
by Theorem 5.12. This shows that (i) H) (ii). That (ii) H) (iii) and (iii) H) (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 mea-
sure 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 . Fremlin’s condition was reformulated
by König [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 WD ¹E 2 A j .E/ < 1º:
Lemma 5.24. The following are equivalent:
(i) for all A 2 A,
.A \ E/ D 0 for all E 2 E H) .A/ D 0;
(ii) for all A 2 A,
jj.A \ E/ D 0 for all E 2 E H) jj.A/ D 0;
(iii) for all A 2 A,
jj.A/ D sup jj.A \ E/ D sup jj.E/:

E 2E E 2E
E A
Proof. By Theorem 5.19, there exists a set P 2 A such that
.A \ P / 0; .A n P / 0; jj.P / D .A \ P / .A n P / (5.32)
for all A 2 A. Such a measurable set P will be fixed throughout the proof.
We prove that (i) implies (ii). Fix a set A 2 A such that jj.A \ E/ D 0 for all
E 2 E. Then it follows from (5.32) that .A \ E \ P / D .A \ E n P / D 0 for
all E 2 E. By (i), this implies .A \ P / D .A n P / D 0 and hence jj.A/ D 0
by (5.32). This shows that (i) implies (ii).
We prove that (ii) implies (i). Fix a set A 2 A such that .A \ E/ D 0
for all E 2 E. Since E \ P 2 E and E n P 2 E for all E 2 E, this implies
.A \ E \ P / D .A \ E n P / D 0 for all E 2 E. Hence it follows from (5.32)
that jj.A \ E/ D 0 for all E 2 E. By (ii), this implies jj.A/ D 0 and hence
.A/ D 0 because j.A/j jj.A/. This shows that (ii) implies (i).
We prove that (ii) implies (iii). Fix a set A 2 A and define
c WD sup jj.E/ jj.A/: (5.33)

E 2E
E A
Choose a sequence Ei 2 E such that Ei A for all i and
lim jj.Ei / D c:
i!1
For i 2 N define
Fi WD E1 [ E2 [ [ Ei :
Then
Fi 2 E; Fi Fi C1 A; jj.Ei / jj.Fi / c (5.34)
for all i , and hence
lim jj.Fi / D c: (5.35)
i !1
Put
1
[
B WD A n F; F WD Fi : (5.36)
iD1
Then
jj.F / D lim jj.Fi / D c
i!1
by part (iv) of Theorem 1.28 and hence
jj.B/ D jj.A/ jj.F / D jj.A/ c: (5.37)
Let E 2 E such that E B. Then E \ Fi D ;, E [ Fi 2 E, and E [ Fi A for

all i by (5.36). Hence
jj.E/ C jj.Fi / D jj.E [ Fi / c
for all i by (5.33). This implies
jj.E/ lim .c jj.Fi // D 0

i!1
by (5.35). Hence jj.E/ D 0 for all E 2 E with E B and it follows from (ii)
that jj.B/ D 0. Therefore, (5.37) yields jj.A/ D c. This shows that (ii) H) (iii).
That (iii) H) (ii) is obvious and this proves Lemma 5.24.
Definition 5.25. The signed measure is called inner regular with respect to if
it satisfies the equivalent conditions of Lemma 5.24.
Theorem 5.26 (generalized Radon–Nikodým theorem). Let .X; A; / be a measure

space and let W 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 2 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 König [8]. It has the advantage
that it reduces the proof of the generalized Radon–Nikodým theorem to the standard
Radon–Nikodým theorem (Theorem 5.18) for -finite measure spaces.
(i) H) (ii). To see that is absolutely continuous with respect to , fix a
measurable set A 2 A such that .A/ D 0 and fix a constant " > 0. Choose
ı > 0 and E 2 A such that .E/ < 1 and (5.31) holds. Then we obtain
.A \ E/ .A/ D 0 < ı, and hence j.A/j < " by (5.31). Thus j.A/j < " for
all " > 0, and hence .A/ D 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 2 A such that
E 2 A; .E/ < 1 H) .A \ E/ D 0:
We must prove that .A/ D 0. Let " > 0 and choose ı > 0 and E 2 A such that
.E/ < 1 and (5.31) holds. Then ..A n E/ \ E/ D 0 < ı, hence j.A n E/j < "
by (5.31), so
j.A/j D j.A n E/ C .A \ E/j D j.A n E/j < ":
This shows that j.A/j < " for all " > 0 and so .A/ D 0, as claimed. Thus we
have proved that (i) H) (ii).
(ii) H) (iii). Since is inner regular with respect to , there exists a sequence
of measurable sets Ei 2 A such that Ei Ei C1 and .Ei / < 1 for all i 2 N and
jj.X/ D lim jj.Ei /. Define
i !1
1
[
X0 WD Ei ; A0 WD ¹A 2 A j A X0 º; 0 WD jA0 ; 0 WD jA0 :
i D1
Then .X0 ; A0 ; 0 / is a -finite measure space and 0 W A0 ! R is a signed measure

that is absolutely continuous with respect to 0 . Hence the Radon–Nikodým theo-
rem (Theorem 5.18) for -finite measure spaces asserts that there exists a function
f0 2 L1 .0 / such that

Z
0 .A/ D f0 d0 for all A 2 A0 :
A
Define
f WX ! R
by
f jX0 WD f0 and f jX nX0 WD 0:
Then f 2 L ./. Choose a measurable set A 2 A. Then it follows from part (v)
1
of Theorem 1.28 that

j.A n X0 /j jj.A n X0 / jj.X n X0 / D lim jj.X n Ei / D 0:
i !1
Hence Z Z
.A/ D 0 .A \ X0 / D f0 d0 D f d
A\X0 A
for all A 2 A. This shows that (i) H) (iii). The uniqueness of f up to equality
-almost everywhere follows immediately from Lemma 1.49.
(iii) H) (i). Choose f 2 L1 ./ such that (5.28) holds. Set
Z
c WD jj.X/ D jf j d
X
and
n
En WD ¹x 2 X j 2 jf .x/j 2n º;
[
E1 WD ¹x 2 X j f .x/ ¤ 0º D En :
n2N
Then 2 .En / jj.En / c and hence .En / 2n c < 1 for all n 2 N.

n
Moreover,
c D jj.X/ D jj.E1 / D lim jj.En /:
n!1
Now fix a constant " > 0. Choose n 2 N such that jj.En / > c "=2 and put
n 1
ı WD 2 ":
If A 2 A such that .A \ En / < ı, then
jj.A/ D jj.A n En / C jj.A \ En /
jj.X n En / C 2n .A \ En /
"
< C 2n ı D ":
2
This shows that is truly continuous with respect to . This completes the first
proof of Theorem 5.26.
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 (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 W A ! Œ0; 1/
is a finite measure that is truly continuous with respect to .
Consider the set
² ˇ ³
ˇ f is measurable and
F WD f W X ! Œ0; 1/ ˇˇ R :
A f d .A/ for all A 2 A
This set is nonempty because 0 2 F. Moreover,

Z
f d .X/ < 1 for all f 2 F,
X
and1
f; g 2 F H) max¹f; gº 2 F: (5.38)
Now put Z
c WD sup f d .X/
f 2F X
and choose a sequence gi 2 F such that

Z
lim gi d D c:
i !1 X
Then it follows from (5.38) that
fi WD max¹g1 ; g2 ; : : : ; gi º 2 F
and Z Z
gi d fi d c for all i 2 N.
X X
1
Let f; g 2 F and A 2 A and define the sets
Af WD ¹x 2 A j f .x/ > g.x/º
and
Ag WD ¹x 2 A j g.x/ f .x/ºI
then Af ; Ag 2 A, Af \ Ag D ;, and Af [ Ag D A; hence
Z Z Z
max¹f; gº d D f d C g d .Af / C .Ag / D .A/:
A Af Ag
Hence fi fiC1 for all i and

Z
lim fi d D c:
i!1 X
Define the function

f W X ! Œ0; 1
by
f .x/ WD lim fi .x/ for x 2 X .
i!1
Then it follows from the Lebesgue monotone convergence Theorem 1.37 that
Z Z
f d D lim fi d D c;
X i !1 X
and Z Z
f d D lim fi d .A/ for all A 2 A:
A i !1 A
Hence f < 1 -almost everywhere by Lemma 1.47 and we may assume without
loss of generality that f .x/ < 1 for all x 2 X . Thus f 2 F.
We claim that Z
f d D .A/ for all A 2 A.
A
Suppose on the contrary that there exists a set A0 2 A such that
Z
f d < .A0 /:
A0
Then the formula

Z
0 .A/ WD .A/ f d for A 2 A (5.39)
A
defines a finite measure by Theorem 1.40.

We prove that there is a measurable function hW X ! Œ0; 1/ such that
Z Z
h d > 0; h d 0 .A/ for all A 2 A: (5.40)
X A
Define
0 .A0 /
" WD > 0: (5.41)
3
Since is truly continuous with respect to , so is 0 . Hence there are a ı > 0 and
a set E 2 A such that .E/ < 1 and
A 2 A; .A \ E/ < ı H) 0 .A/ < ": (5.42)

Take A WD X n E to obtain 0 .X n E/ < " and hence
0 .E/ 0 .A0 \ E/ D 0 .A0 / 0 .A0 n E/ D 3" 0 .A0 n E/ > 2":
Then take A WD A0 . Since 0 .A0 / D 3" " by (5.41), it follows from (5.42) that
.E/ .A0 \ E/ ı > 0:
Define the signed measure

00 W A ! R
by
.A \ E/
00 .A/ WD 0 .A/ " for A 2 A. (5.43)
.E/
Then
00 .E/ D 0 .E/ " ":
By the Hahn decomposition theorem (Theorem 5.19), there exists a measurable set
P 2 A such that
00 .A \ P / 0; 00 .A n P / 0 for all A 2 A:
Since 00 .E n P / 0, it follows that
" 00 .E/ 00 .E \ P / 0 .E \ P /:
Hence .E \ P / ı by (5.42). Now put

"
h WD : (5.44)
.E/ E \P
Then Z
h d > 0:
X
Moreover, if A 2 A, then 00 .A \ P / 0 and so, by (5.43),
.A \ P \ E/
Z
0
.A \ P / " D h d:
.E/ A
Thus Z
h d 0 .A/ for all A 2 A,
A
and so h satisfies (5.40), as claimed.

It follows from (5.40) that

Z Z
.f C h/ d f d C 0 .A/ D .A/
A A
for all A 2 A and hence f C h 2 F. Since

Z Z
.f C h/ d D c C h d > c;
X X
R
this contradicts the definition of c. Thus we have proved that A f d D .A/
for all A 2 A, and hence f satisfies (5.28). This completes the second proof of
Theorem 5.26.
5.5 Exercises
Exercise 5.27. Let .X; A; / be a measure space such that .X/ < 1. Define
.A; B/ WD .A n B/ C .B n A/ for A; B 2 A: (5.45)
Define an equivalence relation on A by A B if .A; B/ D 0. Prove that

descends to a function
W A= A= ! Œ0; 1/
(denoted by the same letter) and that the pair .A=; / is a complete metric space.
Prove that the function
Z
A ! RW A 7 ! f d
A
descends to a continuous function on A= for every f 2 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
ı > 0 such that, for all E 2 A and all f 2 F,
ˇZ ˇ
ˇ ˇ
.E/ < ı H) ˇ f dˇˇ < ":
ˇ
E
5.5. Exercises 215
(i) Every finite subset of L1 ./ is uniformly integrable. Hint. Lemma 5.21.
(ii) Vitali’s theorem. Assume .X/ < 1, let f W X ! R be measurable,

and let fn 2 L1 ./ be a uniformly integrable sequence that converges almost
everywhere to f . Then f 2 L1 ./ and
Z
lim jf fn j d D 0:
n!1 X
Hint. Use Egoroff’s theorem in Exercise 4.54.
(iii) The hypothesis .X/ < 1 cannot be omitted in Vitali’s theorem.

Hint. Consider the Lebesgue measure on R. Find a uniformly integrable
sequence fn 2 L1 .R/ that converges pointwise to the constant function
f 1.
(iv) Vitali’s theorem implies the Lebesgue dominated convergence theorem

(Theorem 1.45) under the assumption .X/ < 1.
(v) Find an example where Vitali’s 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/ < 1 and a sequence
fn 2 L1 ./ that is not uniformly integrable, converges pointwise to zero, and
satisfies Z
lim fn d D 0:
n!1 X
Hint. Consider the Lebesgue measure on X D Œ0; 1.
(vii) Converse of Vitali’s theorem. Assume R .X/ < 1 and let fn be

a sequence in L1 ./ such that the limit lim A fn d exists for all A 2 A.
n!1
Then the sequence fn is uniformly integrable.
Hint. Let " > 0. Prove that there exist a constant ı > 0, an integer n0 2 N,
and a measurable set E0 2 E such that, for all E 2 A and all n 2 N,
ˇZ ˇ
ˇ ˇ
.E; E0 / < ı; n n0 H) ˇ .fn fn0 / dˇˇ < ":
ˇ (5.46)
E
(Here .E; E0 / is defined by (5.45) as in Exercise 5.27.)

If A 2 A satisfies .A/ < ı, then the sets
E WD E0 n A and E WD E0 [ A
both satisfy .E; E0 / < ı. Deduce that, for all A 2 A and all n 2 N,
ˇZ ˇ
ˇ ˇ
.A/ < ı; n n0 H) ˇ .fn fn0 / dˇˇ < 2":
ˇ (5.47)
A
Now use part (i) to find a constant ı 0 > 0 such that, for all A 2 A,
ˇZ ˇ
0
ˇ ˇ
.A/ < ı H) sup ˇ fn dˇˇ < 3":
ˇ (5.48)
n2N A
Exercise 5.29 (Rudin [17, page 134]). Let .X; A; / be a measure space such that
.X/ < 1 and fix a real number p > 1. Let f W X ! R be a measurable function
and let fn 2 L1 ./ be a sequence that converges pointwise to f and satisfies
Z
sup jfn jp d < 1:
n2N X
Prove that Z
f 2 L1 ./; lim jf fn j d D 0:
n!1 X
Hint. Use Vitali’s theorem in Exercise 5.28.
Exercise 5.30. Let X WD R, denote by B 2X the Borel -algebra, and let

W B ! Œ0; 1 be the restriction of the Lebesgue measure to B. Let W B ! Œ0; 1
be a measure. Prove the following.
(1) If B 2 B and 0 < c < .B/, then there exists a Borel set A B such that
.A/ D c. Hint. Show that the function f .t/ WD .B \Œ t; t/ is continuous.
(2) If there exists a constant 0 < c < 1 such that
.B/ D c H) .B/ D c:
for all B 2 B, then .
Exercise 5.31. Let X WD R, denote by B 2X the Borel -algebra,

let W B ! Œ0; 1 be the restriction of the Lebesgue measure to B, and let
W B ! Œ0; 1 be the counting measure.
5.5. Exercises 217
(i) .
(ii) is not inner regular with respect to .
(iii) There does not exist any measurable function f W X ! Œ0; 1 such that
Z
.B/ D f d for all B 2 B.
B
Exercise 5.32. Let X WD Œ1; 1/, denote by B 2X the Borel -algebra, and let
W B ! Œ0; 1 be the restriction of the Lebesgue measure to B. Let W B ! Œ0; 1
be a Borel measure such that
.B/ D ˛.˛B/ for all ˛ 1 and all B 2 B: (5.49)
Prove that there exists a real number c 0 such that

Z
.B/ WD f d for all B 2 B; (5.50)
B
where f W Œ1; 1/ ! Œ0; 1/ is the function given by

c
f .x/ WD for x 1: (5.51)
x2
Hint. Show that .Œ1; 1// < 1, and then that .
Exercise 5.33. Let X WD Œ0; 1/, denote by B 2X the Borel -algebra, and
let W B ! Œ0; 1 be the restriction of the Lebesgue measure to B. Define the
measures
1 ; 2 W B ! Œ0; 1
by2
1
1
X Z
1 .B/ WD x dx;
nD1
n3 B\Œn;nC1
1
Z
2 .B/ WD dx
B\Œ1;1/ x2
for B 2 B.
2
Here we denote by Z Z
f .x/ dx WD f d
B B
the Lebesgue integral of a Borel measurable function f W Œ0; 1/ ! Œ0; 1/ over a Borel set B 2 B.
Prove that 1 and 2 are finite measures that satisfy
1 ; 2 ; 1 2 ; 2 1 ;
and
6 1 ; 6 2 :
Exercise 5.34. Let .X; A; / be a measure space. Show that the signed measures
W A ! R form a Banach space M D M.X; A/ with the norm
kk WD jj.X/:
Show that the map

L1 ./ ! MW Œf 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/ WD Cc .X/
the space of continuous real-valued functions on X . This is a Banach space equipped

with the supremum norm
kf k WD sup jf .x/j:
x2X
Let M.X/ denote the space of signed Borel measures as in Exercise 5.34.
For 2 M.X/ define the linear functional
ƒ W C.X/ ! R
by Z
ƒ .f / WD f d:
X
(i) kƒ k D kk. Hint. Use the Hahn decomposition theorem (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 2 C.X/ with f 0 prove that
ƒC .f / WD sup¹ƒ.hf / j h 2 C.X/; 0 h 1º
(5.52)
D sup¹ƒ.g/ j g 2 C.X/; 0 g f º:
5.5. Exercises 219
Here the second supremum is obviously greater than or equal to the first.
To prove the converse inequality, show that, for all g 2 C.X/ with
0 g f and all " > 0 there is an h 2 C.X/ such that 0 h 1 and
jƒ.g hf /j < ". Namely, find 2 C.X/ such that 0 1, .x/ D 0
when f .x/ "=2 kƒk and .x/ D 1 when f .x/ "= kƒk; then define
h WD g=f . Once (5.52) is established, show that ƒC extends to a positive
linear functional on C.X/.
(iii) The map M.X/ ! C.X/ W 7! ƒ is bijective. Hint. Use the Riesz repre-
sentation theorem (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.
Exercise 5.36. Let .X; A; / be a measure space and let f W X ! Œ0; 1/ be a
measurable function. Define the measure f W A ! Œ0; 1 by
Z
f .A/ WD f d for A 2 A: (5.53)
A
(See Theorem 1.40.)

(i) If is -finite, then so is f .
(ii) If is semi-finite, then so is f .
(iii) If is localizable, then 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 1. Find an example of a measure space
.X; A; / and a measurable function f W X ! Œ0; 1 that violates assertions (i),
(ii), and (iii).
Hint 1. To prove (ii), fix a set A 2 A, define Af WD ¹x 2 A j f .x/ > 0º, and
choose a measurable set E 2 A such that E Af and 0 < .E/ < 1. Consider
the sets En WD ¹x 2 E j f .x/ nº.
Hint 2. To prove (iii), let E A be any collection of measurable sets and choose
a measurable -envelope H 2 A of E. Prove that the set
Hf WD ¹x 2 H j f .x/ > 0º
is a measurable f -envelope of E. In particular, if N 2 A is a measurable set such
that f .E \ N / D 0 for all E 2 E, define Nf WD ¹x 2 N j f .x/ > 0º, show that
.H \ Nf / D 0, and deduce that f .Hf \ N / D f .H \ Nf / D 0.
Chapter 6
Differentiation
This chapter returns to the Lebesgue measure on Euclidean space Rn introduced in

Chapter 2. The main result is the Lebesgue differentiation theorem (Section 6.3).
It implies that if f W Rn ! R is a Lebesgue integrable function, then for almost
every element x 2 Rn , the mean value of f over a ball centered at x converges to
f .x/ as the radius tends to zero. Essential ingredients in the proof are the
Vitali covering lemma and the Hardy–Littlewood maximal inequality (Section 6.2).
One of the consequences of the Lebesgue differentiation theorem is the fundamen-
tal theorem of calculus for absolutely continuous functions of one real variable
(Section 6.4). The Lebesgue differentiation theorem also plays a central role in
the proof of the Calderón–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 W X ! R be a measur-
able function. Define
f W Œ0; 1/ ! Œ0; 1
by
f .t/ WD .t; f / WD .A.t; f //; A.t; f / WD ¹x 2 X j jf .x/j > tº; (6.1)
for t 0. The function f is nonincreasing and hence Borel measurable. Define

the function f W Œ0; 1/ ! Œ0; 1 by
f .˛/ WD inf¹t 0 j .t; f / ˛º for 0 ˛ < 1: (6.2)
Thus f .0/ D kf k1 and f is nonincreasing and hence Borel measurable.

By definition, the infimum of the empty set is infinity. Thus f .˛/ D 1 if and
only if .A.t; f // > ˛ for all t > 0. When f .˛/ < 1, it is the smallest number
t such that the domain A.t; f / (on which jf j > t ) has measure at most ˛. This is
spelled out in the next lemma.
222 6. Differentiation
Lemma 6.1. Let 0 ˛ < 1 and 0 t < 1. Then the following holds:
(i) f .˛/ D 1 if and only if f .s/ > ˛ for all s 0;
(ii) f .˛/ D 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 .˛/ D 1 if and
only if .s; f / > ˛ for all s 2 Œ0; 1/, and this proves (i).
To prove (ii), fix a constant 0 t < 1. 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 .˛/ D t by definition. Conversely, suppose that f .˛/ D 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

1
[
A.t; f / D A.t C 1=n; f /:
nD1
Hence it follows from part (iv) of Theorem 1.28 that
f .t/ D .A.t; f //
D lim .A.t C 1=n; f //
n!1
D lim .t C 1=n; f /

n!1
˛:
This proves (ii).

If f .˛/ t, then f .t/ f .f .˛// ˛ by (ii). If f .t/ ˛, then
f .˛/ t by the definition of f . This proves (iii) and Lemma 6.1.

6.1. Weakly integrable functions 223
Lemma 6.2. Let f; gW X ! R be measurable functions and let c 2 R. Then
kf k1;1 WD sup ˛f .˛/ D sup tf .t/ kf k1 ; (6.3)

˛>0 t>0
kcf k1;1 D jcj kf k1;1 ; (6.4)
kf k1;1 kgk1;1
kf C gk1;1 C for 0 < < 1; (6.5)
1
q q q
kf C gk1;1 kf k1;1 C kgk1;1 : (6.6)
Moreover, kf k1;1 D 0 if and only if f vanishes almost everywhere. The inequal-

ity (6.6) is called the weak triangle inequality.
Proof. For 0 < t; c < 1 it follows from part (iii) of Lemma 6.1 that
1
t.t; f / c () .t; f / ct
() f .ct 1
/t
1
() ct f .ct 1
/ c:
This shows that

sup t.t; f / D sup ˛f .˛/:
t>0 ˛>0
Moreover,
Z Z
t.t; f / D t.A.t; f // jf j d jf j d for all t > 0.
A.t;f / X
This proves (6.3).

For c > 0 equation (6.4) follows from the fact that A.t; cf / D A.t=c; f / and
hence .t; cf / D .t=c; f / for all t > 0. Since k f k1;1 D kf k1;1 by definition,
this proves (6.4).
To prove (6.5), observe that A.t; f C g/ A.t; f / [ A..1 /t; g/, hence
.t; f C g/ .t; f / C ..1 /t; g/; (6.7)
and so
kf k1;1 kgk1;1
t.t; f C g/ t.t; f / C t..1 /t; g/ C
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 p p
inf C D a C b for a; b 0: (6.8)
0<<1 1
This is obvious when a D 0 or b D 0. Hence assume a and b are positive and
define the function
hW .0; 1/ ! .0; 1/
by
a b
h./ WD C :
1
It satisfies
b a
h0 ./ D
.1 /2 2
and hence has a unique critical point at
p
a
0 WD p p :
aC b
p p
Since h.0 / D . a C b/2 , this proves (6.8). The inequality (6.6) then follows
by taking a WD kf k1;1 and b WD kgk1;1 .
The last assertion follows from the fact that kf k1;1 D 0 if and only if f .0/ D 0
if and only if the set A.0; f / D ¹x 2 X j f .x/ ¤ 0º 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; gW R ! R by
1 1
f .x/ WD ; g.x/ WD for 0 < x < 1;
x 1 x
and f .x/ WD g.x/ WD 0 for x 0 and for x 1. Then
kf k1;1 D kgk1;1 D 1; kf C gk1;1 D 4:
Definition 6.4. Let .X; A; / be a measure space. A measurable function
f WX ! R
is called weakly integrable if

kf k1;1 < 1:
The space of weakly integrable functions will be denoted by
L1;1 ./ WD ¹f W X ! R j f is measurable and kf k1;1 < 1º:

6.1. Weakly integrable functions 225
The quotient space

L1;1 ./ WD L1;1 ./=

under the equivalence relation f g if f D g -almost everywhere is called
the weak L1 space. It is not a normed vector space because the function
L1;1 ./ ! Œ0; 1/W Œf 7! kf k1;1 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;1 .Œf ; Œg / WD kf gk1;1 for f; g 2 L1;1 ./: (6.9)
For the Lebesgue measure space .Rn ; A; m/ we write
L1;1 .Rn / WD L1;1 .m/ and L1;1 .Rn / WD L1;1 .m/:
A subset of L1;1 ./ is open in the topology determined by the metric (6.9) if
and only if it is a union of sets of the form ¹Œg 2 L1;1 ./ j kf gk1;1 < rº
with f 2 L1;1 ./ and r > 0. A sequence Œfi 2 L1;1 ./ converges to Œf
in this topology if and only if lim kfi f k1;1 D 0. The inequality (6.3) in
i!1
Lemma 6.2 shows that
L1 ./ L1;1 ./
for every measure space .X; A; /. In general, L1;1 ./ is not equal to L1 ./.
For example, the function
f WR ! R
defined by
81
< for x > 0,
f .x/ WD x
:
0 for x 0,
is weakly integrable, but not integrable.
Theorem 6.5. The metric space .L1;1 ./; d1;1 / is complete.
Proof. Choose a sequence of weakly integrable functions
fi W X ! R
whose equivalence classes Œfi form a Cauchy sequence in L1;1 ./ with respect
to the metric (6.9). Then there is a subsequence
i1 < i2 < i3 <

such that
2k
kfik fikC1 k1;1 < 2 for all k 2 N.
For k; ` 2 N put
k
Ak WD A.2 ; fik fikC1 /
and
1
[ 1
\
E` WD Ak ; E WD E` :
kD` `D1
Then
k 2k

2 .Ak / fik fikC1 1;1 < 2 for all k 2 N,
hence
1
X 1
X
k
.E` / .Ak / 2 D 21 `
kD` kD`
for all ` 2 N, and so .E/ D 0. If x 2 X n E, then there exists an ` 2 N such that

x … Ak for all k ` and so
k
jfik .x/ fikC1 .x/j 2 for all k `.
This shows that the limit

f .x/ WD lim fik .x/
k!1
exists for all x 2 X n E. Extend f to a measurable function on X by setting
f .x/ WD 0 for x 2 E.
We prove that
lim kfi f k1;1 D 0
i!1
and hence also f 2 L1;1 ./. To see this, fix a constant " > 0 and choose an integer
i0 2 N such that

.i; j 2 N; i; j i0 / H) 4 fi fj 1;1 < ":
Now fix a constant t > 0 and choose ` 2 N such that
i` i0 ; 22 ` t "; 22 `
t:
6.2. Maximal functions 227
If x … E` , then x … Ak for all k `, hence

k
jfik .x/ fikC1 .x/j 2 for k `,
and so
1
X 1
X
k
jfi` .x/ f .x/j jfik .x/ fikC1 .x/j 2 D 21 `
t=2:
kD` kD`
This shows that A.t=2; fi` f / E` and hence
tfi f .t=2/ D t.A.t=2; fi` f // t.E` / t21 `

"=2:
`
With this understood, it follows from (6.7) with D 1=2 that
tfi f .t/ tfi fi` .t=2/ C tfi f .t=2/ 2kfi fi` k1;1 C "=2 < "
`
for all i 2 N with i i0 . Hence
kfi f k1;1 D 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 differentiability of a
function that is obtained by integrating a signed measure.
Theorem 6.6. Let W A ! R be a signed measure and define
f WR ! R
by
f .x/ WD .. 1; x// for x 2 R: (6.10)
Fix two real numbers x; A 2 R. Then the following are equivalent:
(i) f is differentiable at x and f 0 .x/ D A.
(ii) For every " > 0 there is a ı > 0 such that, for every open interval U R,
ˇ ˇ
ˇ .U / ˇ
x 2 U; m.U / < ı H) ˇ ˇ Aˇˇ ": (6.11)
m.U /
Proof. (i) H) (ii). Fix a constant " > 0. Since f is differentiable at x and we have
f 0 .x/ D A, there exists a constant ı > 0 such that, for all y 2 R,
ˇ ˇ
ˇ f .x/ f .y/ ˇ
0 < jx yj < ı H) ˇ ˇ Aˇˇ ": (6.12)
x y
Let a; b 2 R such that a < x < b and b a < ı. Then, by (6.12),
ˇ ˇ ˇ ˇ
ˇ f .x/ f .a/ ˇ ˇ f .b/ f .x/ ˇ
ˇ Aˇˇ "; ˇ Aˇˇ ";
ˇ x a ˇ b x
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 .Œa; b// D f .b/ f .a/ and m.Œa; b// D b a it follows that
ˇ ˇ
ˇ .Œa; b// ˇ
ˇ A ˇ ":
ˇ m.Œa; b// ˇ
k
Replace a by a C 2 and take the limit k ! 1 to obtain
ˇ ˇ
ˇ ..a; b// ˇ
ˇ A ˇ ":
ˇ m..a; b// ˇ
Thus we have proved that (i) H) (ii).

(ii) H) (i). Fix a constant " > 0. Choose ı > 0 such that (6.11) holds for every
open interval U R. Choose y 2 R such that x < y < x C ı. Choose k 2 N
such that y x C 2 k < ı. Then
k
Uk WD .x 2 ; y/
is an open interval of length m.Uk / < ı containing x and hence
ˇ ˇ
ˇ .Uk / ˇ
ˇ m.U / Aˇ "
ˇ ˇ
k
by (6.11). Take the limit k ! 1 to obtain

ˇ ˇ ˇ ˇ ˇ ˇ
ˇ f .y/ f .x/ ˇ ˇ .Œx; y//
ˇ D lim ˇ .Uk /
ˇ ˇ ˇ
ˇ A ˇDˇ A Aˇˇ ":
ˇ y x ˇ ˇ m.Œx; y// ˇ k!1 m.Uk /
ˇ
Thus (6.12) holds for x < y < x C ı and an analogous argument proves the
inequality for x ı < y < x. Thus (ii) implies (i) and this proves Theorem 6.6.
The main theorem of this chapter will imply that, when is absolutely con-
tinuous with respect to m, the derivative of the function f in (6.10) exists almost
everywhere, defines a Lebesgue integrable function f 0 W R ! R, and that
Z
.A/ D f 0 dm
A
for all Lebesgue measurable sets A 2 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
W A ! R
at a point x 2 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 2 N. Let .Rn ; A; m/ denote the Lebesgue mea-
sure 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 W Rn ! R. An element of L1 .Rn /
need not be Borel measurable, but differs from a Borel measurable function on a
Lebesgue null set by Theorem 2.14 and part (v) of Theorem 1.55. For x 2 Rn and
r > 0, denote the open ball of radius r, centered at x, by
Br .x/ WD ¹y 2 Rn j jx yj < rº:
Here q
jj WD 12 C C n2
denotes the Euclidean norm of D .1 ; : : : ; n / 2 Rn .
Definition 6.7 (Hardy–Littlewood maximal function). Let W B ! Œ0; 1/ be a
finite Borel measure. The maximal function of is the function
MW Rn ! R
defined by
.Br .x//
.M/.x/ WD sup : (6.13)
r>0 m.Br .x//
The maximal function of a signed measure W B ! R is defined as the maximal

function
M WD M jjW Rn ! R
of its total variation jjW B ! Œ0; 1/.
Theorem 6.8 (Hardy–Littlewood maximal inequality). Let W B ! R be a signed

Borel measure. Then the function M W Rn ! R in Definition 6.7 is lower semi-
continuous, i.e., the pre-image of the open interval .t; 1/ under M is open for all
t 2 R. Hence M is Borel measurable. Moreover,
kM k1;1 3n jj.Rn /; (6.14)
and so M 2 L1;1 .Rn /.
The proof of Theorem 6.8 relies on the following two lemmas.
Lemma 6.9. Let W B ! Œ0; 1/ be a finite Borel measure. Then the maximal
function MW Rn ! R is lower semi-continuous and hence is Borel measurable.
Proof. Let t > 0 and define
U t WD A.t; M/ D ¹x 2 Rn j .M/.x/ > tº: (6.15)
We prove that U t is open. Fix an element x 2 U t . Since .M/.x/ > t, there exists
a number r > 0 such that
.Br .x//
t< :
m.Br .x//
Choose ı > 0 such that
.r C ı/n .Br .x//
t n
< :
r m.Br .x//
Further, choose y 2 Rn such that jy xj < ı. Then Br .x/ BrCı .y/, and hence
.BrCı .y// .Br .x//

.r C ı/n
>t m.Br .x//
rn
.r C ı/n
Dt m.Br .y//
rn
D t m.BrCı .y//:
This implies
.BrCı .y//
.M/.y/ > t;
m.BrCı .y//
and hence y 2 U t . This shows that U t is open for all t > 0. It follows that
n
S
U0 D t >0 U t is open and U t D R is open for t < 0. Thus M is lower
semi-continuous, as claimed. This proves Lemma 6.9.
The Hardy–Littlewood estimate on the maximal function M is equivalent to

an upper bound for the Lebesgue measure of the set U t in (6.15). The proof relies
on the next lemma about coverings by open balls.
Lemma 6.10 (Vitali’s covering lemma). Let ` 2 N and, for i D 1; : : : ; `, let

xi 2 Rn and ri > 0. Define
`
[
W WD Bri .xi /:
iD1
Then there exists a set

S ¹1; : : : ; `º
such that
Bri .xi / \ Brj .xj / D ; for all i; j 2 S with i ¤ j (6.16)
and [
W B3ri .xi /: (6.17)
i2S
Proof. Abbreviate Bi WD Bri .xi / and choose the ordering such that
r1 r2 r` :
Choose i1 WD 1 and let i2 > 1 be the smallest index such that Bi2 \ Bi1 D ;.
Continue by induction to obtain a sequence
1 D i1 < i2 < < ik `
such that
Bij \ Bij 0 D ; for j ¤ j 0

and
Bi \ .Bi1 [ [ Bij / ¤ ; for ij ik when j D k).

Then,
Bi B3ri1 .xi1 / [ [ B3rij .xij / for ij 0. Then the set
U t WD A.t; M /
is open by Lemma 6.9. Choose a compact set K U t . If x 2 K U t , then

.M /.x/ > t and so there exists a number r.x/ > 0 such that
jj.Br.x/ .x//
> t: (6.18)
m.Br.x/ .x//
Since K is compact, there exist finitely many points
x1 ; : : : ; x ` 2 K
such that
`
[
K Bri .xi /;
iD1
where ri WD r.xi /. By Lemma 6.10, there is a subset S ¹1; : : : ; `º such that

S
the balls Bri .xi / for i 2 S are pairwise disjoint and K i2S B3ri .xi /. Since
m.B3r / D 3n m.Br / by Theorem 2.17, this gives
X 3n X 3n
m.K/ 3n m.Bri .xi // < jj.Bri .xi // jj.Rn /:
t t
i2S i2S
Here the second step follows from (6.18) with ri D r.xi / and the last step fol-
lows from the fact that the balls Bri .xi / for i 2 S are pairwise disjoint. Take the
supremum over all compact sets K U t to obtain
3n
m.A.t; M // D m.U t / jj.Rn /: (6.19)
t
(See Theorem 2.13.) Multiply the inequality (6.19) by t and take the supremum over
all real numbers t > 0 to obtain kM k1;1 3n jj.Rn /. This proves Theorem 6.8.

6.3. Lebesgue points 233
Definition 6.11. Let f 2 L1 .Rn /. The maximal function of f is the function
Mf W Rn ! Œ0; 1/
defined by
1
Z
.Mf /.x/ WD sup jf j d m for x 2 Rn : (6.20)
r>0 m.Br .x// Br .x/
Corollary 6.12. Let f 2 L1 .Rn / and define the signed Borel measure f on Rn by
Z
f .B/ WD f dm
B
for every Borel set B Rn . Then
Mf D Mf 2 L1;1 .Rn /

and
kMf k1;1 3n kf k1 :
Proof. The formula jf j.B/ D B jf j d m for B 2 B shows that Mf D Mf .

R
Hence the assertion follows from Theorem 6.8.
Corollary 6.12 shows that the map f 7! Mf descends to an operator (denoted

by the same letter) from the Banach space L1 .Rn / to the topological vector space
L1;1 .Rn /. Corollary 6.12 also shows that the resulting operator
M W L1 .Rn / ! L1;1 .Rn /
is continuous (because jMf Mgj M.f g/). Note that it is not linear.
By Theorem 6.8, M extends naturally to an operator 7! M from the Banach
space of signed Borel measures on Rn to L1;1 .Rn /. (See Exercise 5.34.)
6.3 Lebesgue points

Definition 6.13. Let f 2 L1 .Rn /. An element x 2 Rn is a called a Lebesgue point
of f if
1
Z
lim jf f .x/j d m D 0: (6.21)
r!0 m.Br .x// Br .x/
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 2 L1 .Rn /. Then there

exists a Borel set E Rn such that m.E/ D 0 and every element of Rn n E is a
Lebesgue point of f .
Proof. For f 2 L1 .Rn / and r > 0 define the function
Tr f W Rn ! Œ0; 1/
by
1
Z
.Tr f /.x/ WD jf f .x/j d m for x 2 Rn : (6.22)
m.Br .x// Br .x/
One can prove via an approximation argument that Tr f is Lebesgue measurable

for every r > 0 and every f 2 L1 .Rn /. However, we shall not use this fact in the
proof. For f 2 L1 .Rn / define the function
Tf W Rn ! Œ0; 1
by
.Tf /.x/ WD lim sup.Tr f /.x/ for x 2 Rn : (6.23)
r!0
We must prove that Tf D 0 almost everywhere for every f 2 L1 .Rn /.

To see this, fix a function f 2 L1 .Rn / and assume without loss of generality
that f is Borel measurable. (See Theorem 2.14 and part (v) of Theorem 1.55.)
By Theorem 4.15, there exists a sequence of continuous functions gi W Rn ! R
with compact support such that
1
kf gi k 1 < for all i 2 N:
2i
Since gi is continuous, we have T gi D 0. Moreover, the function
hi WD f gi
is Borel measurable and satisfies, for all x 2 Rn ,

1
Z
.Tr hi /.x/ D jhi hi .x/j d m
m.Br .x// Br .x/
1
Z
jhi j d m C jhi .x/j
m.Br .x// Br .x/
.M hi /.x/ C jhi .x/j:
Thus
Tr hi M hi C jhi j for all i and all r > 0.
Take the limit superior as r tends to zero to obtain
T hi M hi C jhi j for all i .
Moreover, it follows from the definition of Tr that
Tr f D Tr .gi C hi / Tr gi C Tr hi for all i and all r > 0.
Take the limit superior as r tends to zero to obtain
Tf T gi C T hi D T hi M hi C jhi j for all i .
This implies
A."; Tf / A."=2; M hi / [ A."=2; hi / for all i and all " > 0. (6.24)
(See (6.1) for the notation A."; Tf /, etc.) Since hi and M hi are Borel measurable
(see Theorem 6.8) the set
Ei ."/ WD A."=2; M hi / [ A."=2; hi / (6.25)
is a Borel set. Since khi k1 < 2 i , we have
2 2 1
m.A."=2; hi // khi k1;1 khi k1 i 1
" " 2 "
and, by Theorem 6.8,
2 2 3n 3n
m.A."=2; M hi // kM hi k1;1 khi k1 i 1 :
" " 2 "
Thus
3n C 1
m.Ei ."// :
2i 1 "
Since this holds for all i 2 N, it follows that the Borel set
1
\
E."/ WD Ei ."/
iD1
has Lebesgue measure zero for all " > 0.

Hence the Borel set

1
[
E WD E.1=k/
kD1
has Lebesgue measure zero. By (6.24) and (6.25), we have
A.1=k; Tf / E.1=k/
for all k 2 N, and hence

1
[
¹x 2 Rn j .Tf /.x/ ¤ 0º D ¹x 2 Rn j .Tf /.x/ > 1=kº
kD1
1
[
D A.1=k; Tf /
kD1
1
[
E.1=k/
kD1
D E:
This shows that

.Tf /.x/ D 0 for all x 2 Rn n E,
and hence every element of Rn n E is a Lebesgue point of f . This proves Theo-

rem 6.14.
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 mea-
sure. The following theorem provides a formula for the function f in Theorem 5.18
(also called the Radon–Nikodým derivative of ).
Theorem 6.15. Let W B ! R be a signed Borel measure on Rn that is absolutely

continuous with respect to the Lebesgue
R measure. Choose a Borel measurable func-
tion f 2 L .R / such that .B/ D B f d m for all B 2 B. Then there exists a
1 n
Borel set E Rn such that m.E/ D 0 and
.Br .x//
f .x/ D lim for all x 2 Rn n E: (6.26)
r!0 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 n E is a Lebesgue point. Since
ˇ ˇ ˇZ ˇ
ˇ .Br .x// ˇ 1 ˇ ˇ
ˇ f .x/ ˇD ˇ f f .x/ d m ˇ
ˇ m.B .x// ˇ m.Br .x// Br .x/
ˇ ˇ
r
1
Z
jf f .x/j d m
m.Br .x// Br .x/
for all r > 0 and all x 2 Rn , it follows that (6.26) holds for all x 2 Rn n E.
Theorem 6.16. Let f 2 L1 .Rn / and let x 2 Rn be a Lebesgue point of f .

Fix two constants 0 < ˛ < 1 and " > 0. Then there exists a ı > 0 such that,
for every Borel set E 2 B and every r > 0,
8
<r < ı;
ˆ ˇ
ˇ 1
Z ˇ
ˇ
E Br .x/; H) ˇ
ˇ f dm f .x/ˇˇ < ": (6.27)
ˆ m.E/ E
m.E/ > ˛m.Br .x//
:
Proof. Since x is a Lebesgue point of f , there exists a constant ı > 0 such that
Z
1
m.Br .x// jf f .x/j d m < ˛" for 0 < r < ı.
Br .x/
Assume 0 < r < ı and let E Br .x/ be a Borel set such that m.E/ ˛m.Br .x//.
Then
ˇ ˇ
ˇ 1 1
Z Z
ˇ
ˇ
ˇ m.E/ f d m f .x/ˇ
ˇ jf f .x/j d m
E m.E/ E
1
Z
jf f .x/j d m
˛m.Br .x// Br .x/
< ":
The Lebesgue differentiation theorem (Theorem 6.14) can be viewed as a the-

orem 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 W B ! R be a signed Borel

measure on Rn such that ? m. Then there exists a Borel set E Rn such that
m.E/ D 0 D jj.Rn n E/ and
jj.Br .x//
lim D 0 for all x 2 Rn n E: (6.28)
r!0 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/ D 0; jj.Rn n A/ D 0:
For " > 0, define the set
° ˇ jj.Br .x// ±
A" WD x 2 Rn n A ˇ lim sup >" :
ˇ
r!0 m.Br .x//
We claim that A" is a Lebesgue null set for every " > 0. To see this, fix two constants
" > 0 and ı > 0. Since the Borel measure jj is regular by Theorem 3.18, there
exists an open set Uı Rn such that
A Uı ; jj.Uı / < ı:
For x 2 A" choose a radius r D r.x/ > 0 such that
jj.Br.x/ .x//
> "; Br.x/ .x/ Uı ;
m.Br.x/ .x//
and consider the open set
[
Wı WD Br.x/ .x/ Uı :
x2A"
Fix a compact subset K Wı and cover K by finitely many of the balls Br.x/ .x/
with x 2 A" . By Vitali’s covering Lemma 6.10, there are elements x1 ; : : : ; xN 2 A"
such that the balls Br.xi / .xi / are pairwise disjoint and K N
S
i D1 B3r.xi / .xi /. Thus
N
X
m.K/ 3n m.Br.xi / .xi //
i D1
N
3n X
< jj.Br.xi / .xi //
"
iD1
N
3n X
jj.Uı /
"
iD1
n
3 ı
:
"
6.4. Absolutely continuous functions 239
This holds for every compact set K Wı and hence m.Wı / 3n ı=". Since ı > 0
was arbitrary, the set A" is contained in an open set of arbitrarily small Lebesgue
measure and so is a Lebesgue null set as claimed. This implies that
1
[
E WD A [ A1=k
kD1
is a Lebesgue null set. It satisfies jj.Rn n E/ D 0 and (6.28) by definition, and this
6.4 Absolutely continuous functions

Definition 6.18. Let I R be an interval. A function f W I ! R is called abso-
lutely continuous if for every " > 0 there exists a ı > 0 such that, for every finite
sequence s1 t1 s2 t2 s` t` in I ,
`
X `
X
jsi ti j < ı H) jf .si / f .ti /j < ": (6.29)
i D1 i D1
Every absolutely continuous function is continuous.
The equivalence of (i) and (iii) in the following result is the fundamental theo-
rem of calculus for Lebesgue integrable functions. The equivalence of (i) and (ii)
is known as the Banach–Zarecki theorem. For functions of bounded variation see
Exercise 6.20 below.
Theorem 6.19 (fundamental theorem of calculus). Let I D Œa; b R be a com-
pact interval, let B 2I be the Borel algebra, and let mW B ! Œ0; 1 be the
restriction of the Lebesgue measure to B. Let f W 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 gW I ! R such that I jgj d m < 1 and,
for all x; y 2 I with x < y,
Z y
f .y/ f .x/ D g.t/ dt: (6.30)
x
The right-hand side denotes the Lebesgue integral of g over the interval Œx; y.
If (iii) holds, then there exists a Borel set E I such that m.E/ D 0 and,
for all x 2 I n E, f is differentiable at x and f 0 .x/ D g.x/.
Proof. We prove that (iii) implies the last assertion of the theorem. Thus assume that
there exists a function g 2 L1 .I / that satisfies (6.30) for all x; y 2 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 n E is a Lebesgue point of g. By Theorem 6.16
with ˛ D 1=2, every element x 2 I n E satisfies condition (ii) in Theorem 6.6 with
A WD g.x/. Hence Theorem 6.6 asserts that the function f is differentiable at every
point x 2 I n E and satisfies f 0 .x/ D g.x/ for x 2 I n E.
(iii) H) (i). Thus assume f satisfies (iii) and define the signed measure
W B ! R
by Z
.B/ WD g dm (6.31)
B
for every Borel set B I . Then is absolutely continuous with respect to the
Lebesgue measure and Z
jj.B/ D jgj d m
B
for ever Borel set B I . Now let " > 0. Since jj m, it follows from
Lemma 5.21 that there exists a constant ı > 0 such that jj.B/ < " for every
Borel set B I with m.B/ < ı. Choose a sequence
s1 t1 s` t`
in I such that
`
X
jti si j < ı
iD1
and define
Ui WD .si ; ti / for i D 1; : : : ; `.
Then the Borel set
`
[
B WD Ui
iD1
has Lebesgue measure

`
X
m.B/ D jti si j < ı:
i D1
Hence
jj.B/ < ":
6.4. Absolutely continuous functions 241
Since ˇZ ˇ Z
ˇ ˇ
jf .ti / f .si /j D ˇˇ g d mˇˇ jgj d m D jj.Ui /
Ui Ui
for all i , it follows that

`
X `
X
jf .ti / f .si /j jj.Ui / D jj.B/ < ":
iD1 i D1
Hence f is absolutely continuous.

(i) H) (ii). Assume f is absolutely continuous. It follows directly from the def-
inition that f is continuous; that it has bounded variation is part (v) of Exercise 6.20.
Now let E I be a Lebesgue null set and assume without loss of generality that
a; b … E. Fix any constant " > 0 and choose ı > 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 / < ı. Choose a (possibly finite) sequence
S
of pairwise disjoint open intervals Ui I such that U D i Ui . Choose si ; ti 2 Ui
such that f .si / D inf Ui f and f .ti / D supUi f . Then it follows from (6.29) that
P P
i m.f .Ui // D Pi .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// D 0.
(ii) H) (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 W B ! R such that f .y/ f .x/ D ..x; y/ for x; y 2 I with x < y.
Since C and are regular by Theorem 3.18 and f is continuous, we have
f .y/ f .x/ D .Œx; y/ D ..x; y/ D .Œx; y// D ..x; y// (6.32)
for all x; y 2 I with x < y. By the Lebesgue decomposition theorem (Theo-

rem 5.17), there exist two signed Borel measures a ; s W B ! R such that
D a C s ; a m; s ? m: (6.33)
Since a m, it follows from Theorem 5.18 that there is an integrable function

g 2 L1 .I / such that Z
a .B/ D g dm (6.34)
B
for every Borel set B I . Define the functions
fa ; fs W I ! R
by
Z x
fa .x/ WD f .a/ C a .Œa; x/ D f .a/ C g.t/ dt; fs .x/ WD s .Œa; x/:
a
Then f D fa C 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 continu-
ous. Moreover fs D f fa is continuous.
It remains to prove that fs 0. The proof given below follows an argument in
Heil [7, Section 3.5.4]. By Theorem 6.17, there exists a Lebesgue null set Es I
such that a; b 2 Es (without loss of generality) and
js ..x r; x C r//j
js j.I n Es / D 0; lim D0 for all x 2 I n Es :
r!0 r
This implies that every element x 2 I n Es satisfies condition (ii) in Theorem 6.6
with A D 0 and f replaced by fs . Hence fs is differentiable at every point x 2 I nEs
and fs0 .x/ D 0 for x 2 I n 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 / D ¹f .x/ fa .x/ j x 2 Es º ¹f .x/ fa .y/ j x; y 2 Es º
is a Lebesgue null set.
We claim that fs .I n Es / is also a Lebesgue null set. To see this, fix a constant
" > 0. For n 2 N define the set
° ˇ 1 ±
An WD x 2 I n Es ˇ y 2 I; jx yj < H) jfs .x/ fs .y/j "jx yj :
ˇ
n
Then
An AnC1 for all n 2 N
and [
I n Es D An :
n2N
Here the last assertion follows from the fact that fs is differentiable on I n Es with
derivative zero. We prove that the Lebesgue outer measure of the set fs .I n Es /
satisfies the estimate
.fs .I n Es // ".b a C "/: (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 / C "
i
6.5. Exercises 243
and each interval Ui contains an element of An . Then f .Ui / is contained in an

interval of length at most "m.Ui /, by the definition of An . Hence
X
.f .An // " m.Ui / "..An / C "/ ".b a C "/:
i
Since the Lebesgue outer measure is continuous from below by part (iii) of Theo-
rem 2.13 and [
fs .I n Es / D fs .An /;
n2N
it follows that
.fs .I n Es // D lim .f .An // ".b a C "/:

n!1
This proves (6.35). Since " > 0 was chosen arbitrary, this implies that fs .I n 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. Finally, since fs is continuous
and fs .0/ D 0 by definition, this implies fs 0. Hence f D fa is absolutely
continuous and this proves Theorem 6.19.
6.5 Exercises
Exercise 6.20. Let I D Œa; b R be a compact interval and let B 2I be the
Borel -algebra. A function f W I ! R is said to be of bounded variation if
`
X
V .f / WD sup jf .ti / f .ti 1 /j < 1: (6.36)
aDt0 <t1 <<t` Db i D1
Denote by BV.I / the set of all functions f W 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. Prove the following.
(i) Every monotone function f W I ! R has bounded variation.

(ii) Let f 2 BV.I / be right continuous. Then there exist right continuous mono-
tone functions f ˙ W I ! R such that f D f C f .
Hint. Define
`
X
F .x/ WD V .f jŒa;x / D sup jf .ti / f .ti 1 /j (6.37)
aDt0 <t1 <<t` Dx
i D1
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 W I ! R be right continuous. Then f 2 BV.I / if and only if there

exists a signed Borel measure D f W B ! R such that .¹aº/ D 0 and
f .x/ f .a/ D .Œa; x/ for a x b: (6.38)
Hint. Assume f is monotone. For h 2 C.I / define

Z b
ƒf .h/ WD h df
a
` (6.39)
X
WD sup . inf h/ .f .ti / f .ti 1 //:
aDt0 <t1 <<t` Db iD1 Œti 1 ;ti
(This is the Riemann–Stieltjes integral. See Körner [9] and compare it with the
Riemann integral [9, 18, 21].) Prove that ƒf W C.I / ! R is a positive linear
functional. Use the Riesz representation theorem R(Theorem 3.15) to find a
Borel measure f W B ! Œ0; 1/ such that ƒf .h/ D I h df for all h 2 C.I /.
Use the fact that f is right continuous to prove that f satisfies (6.38).
(v) If f 2 BV.I / is right continuous and f is as in (iii), then V .f / D jf j.I /.
(vi) Every absolutely continuous function f W I ! R has bounded variation.
Exercise 6.21. Let .R; A; m/ be the Lebesgue measure space and fix a constant
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 WD E \Œ 1;1
and define the measure f W B ! R by

Z
f .B/ WD f d m D m.B \ E \ Œ 1; 1/:
B
Examine the Lebesgue points of f . (Compare this with Exercise 2.32.)

6.5. Exercises 245
Exercise 6.22. Prove the Vitali–Carath éodory theorem. Let .X; U/ be a

locally compact Hausdorff space and let B 2X be its Borel -algebra.
Let W B ! Œ0; 1 be an outer regular Borel measure that is inner regular on open
sets. Let f 2 L1 ./ and let " > 0. Then there exists an upper semi-continuous
function uW X ! R that is bounded above and a lower semi-continuous function
vW 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 measur-
able sets Ei 2 A, not necessarily disjoint, and a sequence of real numbers ci > 0
such that .Ei / < 1 for all i and
1
X
f D ci E :
i
iD1
Thus
1
X Z
ci .Ei / D f d < 1:
iD1 X
Choose a sequence of compact sets Ki X and a sequence of open sets Ui X

such that Ki Ei Ui and ci .Ui n Ki / < "2 i 1 for all i. Choose n 2 N such
that 1
P
i DnC1 ci .Ei / < "=2 and define
n
X 1
X
u WD ci K ; v WD ci U :
i i
i D1 iD1
Show that X .v u/ d < ", v is lower semi-continuous (i.e., v 1 ..t; 1// is open
R
for all t 2 R), and u is upper semi-continuous (i.e., u 1 .. 1; t// is open for all
t 2 R).
Exercise 6.23. Fix two real numbers a < b and prove the following.
(i) If f W Œa; b ! R is everywhere differentiable, then f 0 W Œa; b ! R is Borel
measurable.
Rb
(ii) If f W Œa; b ! R is everywhere differentiable and a jf 0 .t/j dt < 1, then f
is absolutely continuous.
Hint. Fix a constant " > 0. By the Vitali–Carathéodory theorem in Exer-
cise 6.22, there is a lower semi-continuous function gW Œa; b ! R such that
Z b Z b
0
g>f ; g.t/ dt < f 0 .t/ dt C ":
a a
For > 0 define the function
F W Œa; b ! R
by
Z x
F .x/ WD g.t/ dt f .x/ C f .a/ C .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 ıx > 0 such that
f .t/ f .x/
g.t/ > f 0 .x/; < f 0 .x/ C for x < t < x C ıx :
t x
Deduce that
F .t/ > F .x/ for x < t < x C ıx :
Since F .a/ D 0, there exists a maximal element x 2 Œa; b such that F .x/ D 0.
If x < b, it follows from the previous discussion that F .t/ > 0 for x < t b.
In either case F .b/ 0 and hence
Z b Z b
f .b/ f .a/ g.t/ dt C .b a/ < f 0 .t/ dt C " C .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 relation f .b/ f .a/ D a f 0 .t/ dt. Now deduce
that Z x
f .x/ f .a/ D f 0 .t/ dt
a
for all x 2 Œa; b.
Example 6.24. (i) The Cantor function is the unique monotone function
f W Œ0; 1 ! Œ0; 1
that satisfies
1 1
X ai X ai
f 2 D
3i 2i
iD1 i D1
6.5. Exercises 247
for all sequences ai 2 ¹0; 1º. It is continuous and nonconstant, and its derivative
exists and vanishes on the complement of the standard Cantor set
1 n n
\ [ h X ai X ai 1i
C WD 2 ; 2 C :
nD1
3i 3i 3n
ai 2¹0;1º iD1 iD1
This Cantor set has Lebesgue measure zero. Hence f is almost everywhere differ-
entiable 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
gW Œ0; 1 ! Œ0; 2 and hW Œ0; 2 ! Œ0; 1
by
1
g.x/ WD f .x/ C x; h WD g :
The image g.Œ0; 1nC / is a countable union of disjoint open intervals of total length
one and hence has Lebesgue measure one. Thus its complement
K WD g.C / Œ0; 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 WD h.E/ Œ0; 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 D h 1 .F / is not Lebesgue measurable.
This shows that the function hW Œ0; 2 ! Œ0; 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 The-
orem 6.19 that every Lebesgue-Lebesgue measurable homeomorphism hW I ! J
has an absolutely continuous inverse.
(iv) Let hW Œ0; 2 ! Œ0; 1 and F C Œ0; 1 be as in part (ii). Then the
characteristic function F W R ! R is Lebesgue measurable and hW Œ0; 2 ! R is
continuous. However, the composition F ı hW Œ0; 2 ! R is not Lebesgue measur-
able because the set .F ı h/ 1 .1/ D E is not Lebesgue measurable.
(v) By contrast, if I; J R are intervals, f W J ! R is Lebesgue measurable,
and hW I ! J is a C 1 diffeomorphism, then f ı hW 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 Fubini’s
Theorem (Section 7.3). The archetypal example is the Lebesgue measure on
RkC` D 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 convo-
lution (Section 7.5), Marcinkiewicz interpolation (Section 7.6), and the Calderón–
Zygmund 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 smallest

-algebra on the product space
X Y WD ¹.x; y/ j x 2 X; y 2 Y º
that contains all subsets of the form A B, where A 2 A and B 2 B. It will be

denoted by A ˝ B 2X Y .
x be an .A ˝ B/-measurable
Lemma 7.2. Let E 2 A ˝ B and let f W X Y ! R
function. Then the following holds.
x , defined by fx .y/ WD f .x; y/ for

(i) For every x 2 X, the function fx W Y ! R
y 2 Y , is B-measurable and
Ex WD ¹y 2 Y j .x; y/ 2 Eº 2 B: (7.1)
x , defined by f y .x/ WD f .x; y/ for

(ii) For every y 2 Y , the function f y W X ! R
x 2 X , is A-measurable and
E y WD ¹x 2 X j .x; y/ 2 Eº 2 A: (7.2)
250 7. Product measures
Proof. Define 2XY by
WD ¹E X Y j Ex 2 B for all x 2 X º:
We prove that is a -algebra. To see this, note first that X Y 2 . Second,

if E 2 , then Ex 2 B for all x 2 X , hence
.E c /x D ¹y 2 Y j .x; y/ … Eº D .Ex /c 2 B
for all x 2 X , and hence E c 2 . Third, if Ei 2 is a sequence and

E WD 1
S1
iD1 .Ei /x 2 B for all x 2 X , and hence E 2 .
S
i D1 Ei , then Ex D
This shows that is a -algebra. Since A B 2 for all A 2 A and all B 2 B,
it follows that A ˝ B . This proves (7.1) for all x 2 X.
Now fix an element x 2 X . If V R x is open, then E WD f 1 .V / 2 A ˝ B and
hence .fx / 1 .V / D Ex 2 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:
S1
(a) if Ai 2 M for i 2 N such that Ai Ai C1 for all i, then i D1 Ai 2 M;
T1
(b) if Bi 2 M for i 2 N such that Bi BiC1 for all i, then i D1 Bi 2 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 2 A and B 2 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 251
Step 1. For every set P X Y the collection
.P / WD ¹Q X Y j P n Q; Q n P; P [ Q 2 Mº
is a monotone class.
This follows immediately from the definition of monotone class.
Step 2. If P; Q X Y , then Q 2 .P / if and only if P 2 .Q/.
This follows immediately from the definition of .P / in Step 1.
Step 3. If P; Q 2 E, then P \ Q; P n Q; P [ Q 2 E.
For the intersection this follows from the fact that
.A1 B1 / \ .A2 B2 / D .A1 \ A2 / .B1 \ B2 / :
For the complement it follows from the fact that
.A1 B1 / n .A2 B2 / D ..A1 n A2 / B1 / [ ..A1 \ A2 / .B1 n B2 //:
For the union this follows from the fact that P [ Q D .P n Q/ [ Q.

Step 4. If P 2 E, then M .P /.
Let P 2 E. Then P n Q; Q n P; P [ Q 2 E M for all Q 2 E by Step 3.
Hence Q 2 .P / for all Q 2 E by the 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 2 M, then M .P /.
Fix a set P 2 M. Then P 2 .Q/ for all Q 2 E by Step 4. Hence Q 2 .P /
for all Q 2 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 2 M, then P n Q; P [ Q 2 M.
If P; Q 2 M, then Q 2 M .P / by Step 5, and hence P n Q; P [ Q 2 M by
the definition of .P / in Step 1.
Step 7. M is a -algebra.
By definition, X Y 2 E M. If P 2 M, then P c D .X Y / n P 2 M by
Step 6. If Pi 2 M for i 2 N, then Qn WD niD1 Pi 2 M for all n 2 N by Step 6
S
and hence 1
S1
nD1 Qn 2 M because M is a monotone class. This proves
S
iD1 Pi D
Step 7 and Lemma 7.5.
Lemma 7.6. Let .X; UX / and .Y; UY / be topological spaces, let UX Y be the prod-
uct topology on X Y (see Appendix B), and let BX , BY , BX Y 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 D BX Y : (7.4)
Proof. The projections X W X Y ! X and Y W X Y ! Y are continuous, and

hence Borel measurable by Theorem 1.20. Thus X 1 .A/ D A Y 2 BXY for all
A 2 BX and Y 1 .B/ D X B 2 BX Y for all B 2 BY . Hence A B 2 BX Y
for all A 2 BX and all B 2 BY , and this implies (7.3).
Now assume .X; UX / is a second countable locally compact Hausdorff space
and choose a countable basis ¹Ui j i 2 Nº of UX such that Uxi is compact for all
i 2 N. Fix an open set W 2 UX Y and, for i 2 N, define
Vi WD ¹y 2 Y j .x; y/ 2 W for all x 2 Ux i º:
We prove that Vi is open. Let y0 2 Vi . Then .x; y0 / 2 W for all x 2 Ux i . Hence,

for every x 2 Ux i , there exist open sets U.x/ 2 UX and V .x/ 2 UY such that
.x; y0 / 2 U.x/V .x/ W . Since Ux i is compact, there are finitely many elements
x1 ; : : : ; x` 2 Ux i such that Ux i U.x1 / [ [ U.x` /. Define
V WD V .x1 / \ \ V .x` /:
Then V is open and Ux i V W , so y0 2 V Vi . This shows that Vi is open for

all i 2 N. Next we prove that
1
[
W D .Ui Vi /: (7.5)
iD1
Let .x0 ; y0 / 2 W . Then there exist open sets U 2 UX and V 2 UY such that
.x0 ; y0 / 2 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 2 U 0 U 0 U . Since the sets Ui form
a basis of the topology, there exists an integer i 2 N such that x0 2 Ui U 0 and
hence x0 2 Ux i U 0 U . Thus
Ux i ¹y0 º U V W ;
7.1. The product -algebra 253
hence y0 2 Vi , and so .x0 ; y0 / 2 Ui Vi W . Since the element .x0 ; y0 / 2 W

was chosen arbitrarily, this proves (7.5). Thus we have shown that
UX Y BX ˝ BY
and this implies
BX Y BX ˝ BY :
Hence (7.4) follows from (7.3). This proves Lemma 7.6.
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 WD ¹.x; x/ j x 2 X º 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 2 .E/, then there exists a sequence Ei 2 E, i 2 N, such that
D 2 .¹Ei j i 2 Nº/:
The union of the sets .E0 / over all countable subsets E0 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 D 2 .E/. Then there are a sequence
Ei 2 E and a set I 2N such that
[\ \
DD Ei \ Eic :
I 2I i2I i2NnI
By Step 1 there exists a sequence Ei 2 E such that D 2 .¹Ei j i 2 Nº/. For

I N define \ \
EI WD Ei \ Eic :
i 2I i2NnI
These sets form a partition of Y . Hence the collection

°[ ˇ ±
F WD EI ˇ I 2N
ˇ
I 2I
is a -algebra on Y . Since Ei 2 F for each i 2 N, it follows that D 2 F. This

proves Step 2.
Step 3. … A ˝ A.
Let E 2X X be the collection of all sets of the form A B with A; B 2 A.

Let D 2 A ˝ A. By Step 2, there are sequences Ai ; Bi 2 A and a set I 2N such
that
[
DD EI ;
I 2I
where \
\
EI WD .Ai Bi / \ .Ai Bi /c :
i 2I i2NnI
Thus
[
EI D AIJ BIJ ;
J NnI
where
\ \
AIJ WD Ai \ .X n Aj /
i2I j 2J
and
\ \
BIJ WD Bi \ .X n Bj / :
i2I j 2Nn.I [J /
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. This proves Lemma 7.7.
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 D UX D 2X . Then is an
open subset of X X with respect to the product topology (which is also discrete
because points are open). Hence 2 BXX D 2X X . 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 W X X ! R defined by
d.x; y/ WD 1 for x ¤ y and d.x; x/ WD 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 255
7.2 The product measure

The definition of the product measure on the product -algebra is based on the
following theorem. For a measure space .X; A; / and a measurable function
W X ! Œ0; 1 we use the notation
Z Z
.x/ d.x/ WD d:
X X
Theorem 7.9. Let .X; A; / and .Y; B; / be -finite measure spaces and let
Q 2 A ˝ B. Then the functions
X ! Œ0; 1W x 7 ! .Qx / and Y ! Œ0; 1W y 7 ! .Qy / (7.6)
are measurable and

Z Z
.Qx / d.x/ D .Qy / d.y/: (7.7)
X Y
Definition 7.10. Let .X; A; / and .Y; B; / be -finite measure spaces. The prod-
uct measure of and is the map
˝ W A ˝ B ! Œ0; 1
defined by
Z Z
. ˝ /.Q/ WD .Qx / d.x/ D .Qy / d.y/ (7.8)
X Y
for Q 2 A ˝ B. That ˝ is -additive, and hence is a measure, follows from

P1
Theorem 1.38 and the fact that .Qx / D iD1 ..Qi /x / for every sequence of
pairwise disjoint sets Qi 2 A ˝ B. The product measure satisfies
. ˝ /.A B/ D .A/ .B/ (7.9)
for A 2 A and B 2 B and hence is -finite.

² ˇ ³
ˇ the functions (7.6) are measurable
WD Q 2 A ˝ B ˇˇ :
and equality (7.7) holds
We prove in five steps that D A ˝ B.

Step 1. If A 2 A and B 2 B, then Q WD A B 2 .
By assumption
´ ´
B if x 2 A; y
A if y 2 B;
Qx D Q D
; if x … A; ; if y … B:
Define the function

W X ! Œ0; 1
by
.x/ WD .Qx / D .B/A .x/ for x 2 X
and the function
WY ! Œ0; 1
by
.y/ WD .Qy / D .A/B .y/ for y 2 Y .
Then ; are measurable and
Z Z
d D .A/.B/ D d:
X Y
Thus Q 2 .
Step 2. If Q1 ; Q1 2 and Q1 \ Q2 D ;, then Q WD Q1 [ Q2 2 .
Define
i .x/ WD ..Qi /x /; .x/ WD .Qx /; (7.10a)
i .y/ WD ..Qi /y /; .y/ WD .Qy / (7.10b)
for x 2 X , y 2 Y and i D 1; 2. Then D 1 C 2 and D 1 C 2. Moreover,

Z Z
i d D i d
X Y
for i D 1; 2 because Qi 2 . Hence

Z Z
d D d
X Y
and so Q 2 .
S1
Step 3. If Qi 2 for i 2 N and Qi QiC1 for all i, then Q WD i D1 Qi 2 .
Define i ; W X ! Œ0; 1 and i; W Y ! Œ0; 1 by (7.10) for i 2 N. Since

1
[ 1
[
y
Qx D .Qi /x ; Q D .Qi /y
iD1 i D1
and .Qi /x 2 B and .Qi /y 2 A for all i , it follows from Theorem 1.28 (iv) that
.x/ D .Qx / D lim ..Qi /x / D lim i .x/ for all x 2 X;

i !1 i !1
.y/ D .Qy / D lim ..Qi /y / D lim i .y/ for all y 2 Y:

i !1 i!1
By the Lebesgue monotone convergence theorem (Theorem 1.37) this implies

Z Z Z Z
d D lim i d D lim i d D d:
X i !1 X i !1 Y Y
Thus Q 2 and this proves Step 3.
Step 4. Let A 2 A and B 2 B such that .A/ < 1 and .B/ < 1. If Qi 2
for i 2 N such that A B Q1 Q2 , then Q WD 1
T
i D1 Qi 2 .
Let i ; ; i ; be as in the proof of Step 3. Since .Qi /x B and .B/ < 1,

it follows from part (v) of Theorem 1.28 that i converges pointwise to . More-
over, i .B/A for all i and the function .B/A W X ! Œ0; 1/ is integrable
because .A/ < 1 and .B/ < 1. Hence it follows from the Lebesgue dominated
convergence theorem (Theorem 1.45) that
Z Z
d D lim i d:
X i!1 X
The same argument shows that

Z Z
d D lim i d:
Y i!1 Y
Since Qi 2 for all i, this implies

Z Z
d D d
X Y
and hence Q 2 . This proves Step 4.

Step 5. D A ˝ B.
Since .X; A; / and .Y; B; / are -finite, there exist sequences of measurable sets
Xn 2 A and Yn 2 B such that
Xn XnC1 ; Yn YnC1 ; .Xn / < 1; .Yn / < 1

S1 S1
for all n 2 N and X D nD1 Xn and Y D nD1 Yn . Define
M WD ¹Q 2 A ˝ B j Q \ .Xn Yn / 2 for all n 2 Nº:
Then M is a monotone class by Steps 3 and 4, E M by Steps 1 and 2,

and M A ˝ B by definition. Hence it follows from Lemma 7.5 that M D A ˝ B.
In other words, Q \ .Xn \ Yn / 2 for all Q 2 A ˝ B. By Step 3, this implies
1
[
for all Q 2 A ˝ B:

QD Q \ .Xn Yn / 2
nD1
Thus A ˝ B A ˝ B and so D A ˝ B, as claimed. This proves Step 5

and theorem (Theorem 7.9).
Examples and exercises

Example 7.11. Let X D Y D Œ0; 1, let A 2X be the -algebra of Lebesgue
measurable sets, let B WD 2Y , let W A ! Œ0; 1 be the Lebesgue measure, and let
W B ! Œ0; 1 be the counting measure. Consider the diagonal
n h
1 [
\ i 1 i i2
WD ¹.x; x/ j 0 x 1º D ; 2 A ˝ B:
nD1 i D1
n n
Its characteristic function
f WD W X Y !R
is given by
´
1 if x D y;
f .x; y/ WD
0 if x ¤ y:
Hence
Z
y
. / D f .x; y/ d.x/ D 0 for 0 y 1;
X
Z
.x / D f .x; y/ d.y/ D 1 for 0 x 1;
Y
and so Z Z
y
. / d D 0 ¤ 1 D .x / d.x/:
X Y
Thus the hypothesis that .X; A; / and .Y; B; / are -finite cannot be removed in
Theorem 7.9.
Example 7.12. Let X WD Y WD Œ0; 1, let A D B 2Œ0;1 be the -algebra of

Lebesgue measurable sets, and let D be the Lebesgue measure.
Claim 1. Assume the continuum hypothesis. Then there is a set Q Œ0; 12 such
that Œ0; 1 n Qx is countable for all x and Qy is countable for all y.
Let Q be as in Claim 1 and define
f WD Q W Œ0; 12 ! R:
Then
Z
.Qy / D f .x; y/ d.x/ D 0 for 0 y 1;
X
Z
.Qx / D f .x; y/ d.y/ D 1 for 0 x 1;
Y
and hence Z Z
.Qy / d D 0 ¤ 1 D .Q/ d.x/:
X Y
The sets Qx and Qy are all measurable and the integrals are finite, but the set Q is
not A ˝ B-measurable. This shows that the hypothesis Q 2 A ˝ B in Theorem 7.9
cannot be replaced by the weaker hypothesis that the sets Qx and Qy are all mea-
surable, 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 W Œ0; 1 ! W with values in a well-ordered set .W; / such that the set
¹w 2 W j w zº is countable for all z 2 W .
Claim 2 H) Claim 1. Let j be as in Claim 2 and put
Q WD ¹.x; y/ 2 Œ0; 12 j j.x/ j.y/º:
Then the set Qy D ¹x 2 Œ0; 1 j j.x/ j.y/º is countable for all y 2 Œ0; 1 and
the set Œ0; 1 n Qx D ¹y 2 Œ0; 1 j j.y/ 4 j.x/º is countable for all x 2 Œ0; 1.
Proof of Claim 2. By Zorn’s lemma, every set admits a well ordering. Choose any
well ordering on A WD Œ0; 1 and define
B WD ¹b 2 A j the set ¹a 2 A j a bº is uncountableº:
If B D ; choose W WD A D Œ0; 1 and j D id. If B ¤ ;, then by the well ordering

axiom, B contains a smallest element b0 . Since b0 2 B, the set
W WD B n A D ¹w 2 A j w b0 º
is uncountable. Since W \ B D ; the set ¹w 2 W j w zº is countable for

all z 2 W . Since W is an uncountable subset of Œ0; 1, the continuum hypothesis
asserts that there exists a bijection j W Œ0; 1 ! W . This proves Claim 2.
Example 7.13. Let X and Y be countable sets, let A D 2X and B D 2Y ,

and let W 2X ! Œ0; 1 and W 2Y ! Œ0; 1 be the counting measures. Then
A ˝ B D 2X Y and ˝ W 2XY ! Œ0; 1 is the counting measure.
Example 7.14. Let .X; A; / and .Y; B; / be probability measure spaces, so that
.X/ D .Y / D 1. Then ˝ W A ˝ B ! Œ0; 1 is also a probability measure.
A trivial example is A D ¹;; Xº and B D ¹;; Y º. In this case the product -algebra
is A ˝ B D ¹;; X Y º and the product measure is given by . ˝ /.;/ D 0 and
. ˝ /.X Y / D 1.
Exercise 7.15. Let .X; A; /, .Y; B; / be -finite measure spaces and let
W X ! X; WY !Y
be bijections. Define the bijection W X Y ! X Y by
. /.x; y/ WD ..x/; .y//
for x 2 X and y 2 Y . Prove that
. / .A ˝ B/ D A ˝ B; . / . ˝ / D ˝ :
Hint. Use Theorem 1.19 to show that A ˝ B . / .A ˝ B/.

See also Exercise 2.34.
7.3. Fubini’s Theorem 261
Exercise 7.16. For n 2 N let Bn Rn be the Borel -algebra and let
n W Bn ! Œ0; 1
be the restriction of the Lebesgue measure to Bn . Let k; ` 2 N and n WD k C `.

Identify Rk R` with Rn in the obvious manner. Then
Bk ˝ B` D Bn
by Lemma 7.6. Prove that the product measure k ˝ ` is translation-invariant and

satisfies .k ˝ ` /.Œ0; 1n / D 1. Deduce that
k ˝ ` D n :
Hint. Use Exercise 7.15. We return to this example in Section 7.4.
7.3 Fubini’s Theorem
There are three versions of Fubini’s Theorem. The first concerns nonegative func-
tions that are measurable with respect to the product -algebra (Theorem 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 ˝ W A ˝ B ! Œ0; 1 be the product measure in Def-
R 1 be an A ˝ B-measurable function. Then
inition 7.10. Let f W X Y ! Œ0;
the function X ! RŒ0; 1W x 7! Y f .x; y/ d.y/ is A-measurable, the function
Y ! Œ0; 1W y 7! X f .x; y/ d.x/ is B-measurable, and
Z Z Z
f d. ˝ / D f .x; y/ d.y/ d.x/
X Y X Y
Z Z (7.11)
D 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 D Y D N.
Proof of Theorem 7.17. Let
fx .y/ WD f y .x/ WD f .x; y/ for .x; y/ 2 X Y
and define the functions
W X ! Œ0; 1 and W Y ! Œ0; 1
by Z Z
.x/ WD fx d; .y/ WD f y d (7.12)
Y X
for x 2 X and y 2 Y . We prove in three steps that is A-measurable,
is B-measurable, and and satisfy (7.11).
Step 1. The assertion holds when f W X Y ! Œ0; 1/ is the characteristic function
of an A ˝ B-measurable set.
Let Q 2 A ˝ B and f D Q . Then fx D Qx and f y D Qy , and so
.x/ D .Qx /; .y/ D .Qy /
for all x 2 X and all y 2 Y . Hence it follows from Theorem 7.9 that
Z Z Z
d D d D . ˝ /.Q/ D f d. ˝ /:
X Y X
Here the third equality follows from the definition of the measure ˝. This proves
Step 1.
Step 2. The assertion holds when f W X Y ! Œ0; 1/ is an A ˝ B-measurable
step-function.
This follows immediately from Step 1 and the linearity of the integral.
Step 3. The assertion holds when f W X Y ! Œ0; 1 is A ˝ B-measurable.
By Theorem 1.26, there exists a sequence of A ˝ B-measurable step-functions
sn W X Y ! Œ0; 1/
such that sn snC1 for all n 2 N and sn converges pointwise to f . Define

Z
n .x/ WD sn .x; y/ d.y/ for x 2 X;
Y
Z
n .x/ WD sn .x; y/ d.x/ for y 2 Y:
X
Then
n nC1 ; n nC1 for all n 2 N
by part (i) of Theorem 1.35. Moreover, it follows from the Lebesgue monotone
convergence theorem (Theorem 1.37) that
.x/ D lim n .x/; .y/ D lim n .y/
n!1 n!1
for all x 2 X and all y 2 Y . Use the Lebesgue monotone convergence theorem
(Theorem 1.37) again as well as Step 2 to obtain
Z Z
d D lim n d
X n!1 X
Z
D lim sn d. ˝ /
n!1 X Y
Z
D f d. ˝ /
XY
Z
D lim n d
n!1 Y
Z
D d:
Y
This proves Step 3 and Theorem 7.17.
A first application of Fubini’s Theorem 7.17 is Minkowski’s 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 < 1. Let .X; A; / and
.Y; B; / be -finite measure spaces and let
f WX Y ! Œ0; 1
be A ˝ B-measurable. Then
Z Z p 1=p Z Z 1=p
f .x; y/ d.y/ d.x/ f .x; y/p d.x/ d.y/:
X Y Y X
In the notation
fx .y/ WD f y .x/ WD f .x; y/;
Minkowski’s 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 W Y ! Œ0; 1 is B-measurable for all x 2 X and

fy W X ! Œ0; 1 is A-measurable for all y 2 Y . Moreover, by Theorem 7.17,
p
the function X ! Œ0; 1W x 7! kfx kL 1 ./ is A-measurable and the function
Y ! Œ0; 1W y 7! kf kLp ./ is B-measurable. Hence both sides of the in-
y
equality (7.13) are well defined. Theorem 7.17 also shows that for p D 1 equality
holds in (7.13). Hence assume 1 < p < 1 and a choose 1 < q < 1 such that
1=p C 1=q D 1. It suffices to assume
Z
c WD kf y kLp ./ d.y/ < 1:
Y
Define W X ! Œ0; 1 by
Z
.x/ WD fx d for x 2 X
Y
and let g 2 Lq ./. Then the function
X Y ! Œ0; 1W .x; y/ 7! f .x; y/jg.x/j
is A ˝ B-measurable. Hence it follows from Theorem 7.17 that

Z Z Z
jgj d D f .x; y/jg.x/j d.y/ d.x/
X X Y
Z Z
D f .x; y/jg.x/j d.x/ d.y/
Y X
Z
kf y kLp ./ kgkLq ./ d.y/
Y
D c kgkLq ./ :
Here the third step follows from Hölder’s 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:

Theorem 7.20 (Fubini for integrable functions). Let .X; A; / and .Y; B; / be
-finite measure spaces, let ˝ W A ˝ B ! Œ0; 1 be the product measure,
and let f 2 L1 . ˝ /. Define
fx .y/ WD f y .x/ WD f .x; y/ for x 2 X and y 2 Y .
(i) fx 2 L1 ./ for -almost every x 2 X and the map
W X ! R
defined by
8Z
< fx d if fx 2 L1 ./;
.x/ WD Y (7.14)
if fx … L1 ./;
:
0
is -integrable.
(ii) f y 2 L1 ./ for -almost every y 2 Y and the map
WY !R
defined by
8Z
< f y d if f y 2 L1 ./;
.y/ WD X (7.15)
y
… L ./;
1
:
0 if f
is -integrable.
(iii) Let 2 L1 ./ and 2 L1 ./ be as in (i) and (ii). Then

Z Z Z
d D f d. ˝ / D d: (7.16)
X X Y Y
Proof. We prove part (i) and the first equality in (7.16). The functions
f ˙ WD max¹˙f; 0ºW X Y ! Œ0; 1/
are A B-measurable by Theorem 1.24. Hence the functions
fx˙ WD max¹˙fx ; 0º D f ˙ .x; /W Y ! Œ0; 1/
are B-measurable by Lemma 7.2.

Define
ˆ˙ W Y ! Œ0; 1
by Z
˙
ˆ .x/ WD fx˙ d for x 2 X:
Y
By Theorem 7.17, the functions ˆ˙ W X ! Œ0; 1 are A-measurable and

Z Z Z
ˆ˙ d D f ˙ d. ˝ / jf j d. ˝ / < 1: (7.17)
X XY X Y
Now Lemma 1.47 asserts that the A-measurable set

² ˇZ ³
E WD x 2 X ˇ jfx j d D 1 D ¹x 2 X j ˆC .x/ D 1 or ˆ .x/ D 1º
ˇ
ˇ
Y
has measure .E/ D 0. Moreover, for all x 2 X,
x 2 E () fx … L1 ./:
Define
˙ W X ! Œ0; 1/
by
´
˙
ˆ˙ .x/ if x … E;
.x/ WD for x 2 X:
0 if x 2 E;
Then it follows from (7.17) that ˙ 2 L1 ./ and
Z Z
˙
d D f ˙ d. ˝ /:
X X Y
Hence D C 2 L1 ./ and

Z Z Z
d D C d d
X X X
Z Z
C
D f d. ˝ / f d. ˝ /
X Y X Y
Z
D f d. ˝ /:
X Y
This proves (i) and the first equality in (7.16). An analogous argument proves (ii)
and the second equality in (7.16). This proves Theorem 7.20.
Example 7.21. Let .X; A; / D .Y; B; / be the Lebesgue measure space in the
unit interval Œ0; 1 as in Example 7.12. Let gn W Œ0; 1 ! Œ0; 1/ be a sequence of
smooth functions such that
Z 1
gn .x/ dx D 1;
0
and
n 1 n
gn .x/ D 0 for x 2 Œ0; 1 n Œ2 ;2
for all n 2 N. Define

f W Œ0; 12 ! R
by
1
X
f .x; y/ WD .gn .x/ gnC1 .x//gn .y/:
nD1
The sum on the right is finite for every pair .x; y/ 2 Œ0; 12 . Then
Z
f .x; y/ dx D 0;
X
Z 1
X
f .x; y/ dy D .gn .x/ gnC1 .x// D g1 .x/;
Y nD1
and hence
Z 1 Z 1 Z 1 Z 1
f .x; y/ dx dy D 0 ¤ 1 D f .x; y/ dy dx:
0 0 0 0
Thus the hypothesis f 2 L1 . ˝ / cannot be removed in Theorem 7.20.
Example 7.22. This example shows that the product measure is typically not com-
plete. Let .X; A; / and .Y; B; / be two complete -finite measure spaces. Sup-
pose .X; A; / admits a nonempty null set A 2 A and B ¤ 2Y . Choose B 2 2Y nB.
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 Fubini’s 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/ ; . ˝ / / denote the completion of
the product space, and let f 2 L1 .. ˝ / /. Define
fx .y/ WD f y .x/ WD f .x; y/ for x 2 X and y 2 Y .
(i) fx 2 L1 ./ for -almost every x 2 X and the map
W X ! R
defined by 8Z
< fx d if fx 2 L1 ./;
.x/ WD Y (7.18)
if fx … L1 ./;
:
0
is -integrable.
(ii) f y 2 L1 ./ for -almost every y 2 Y and the map
WY !R
defined by 8Z
< f y d if f y 2 L1 ./;
.y/ WD X (7.19)
if f y … L1 ./;
:
0
is -integrable.
Z Z Z

d D f d. ˝ / D d: (7.20)
X X Y Y
Proof. By part (v) of Theorem 1.55, there exists a function g 2 L1 . ˝ / such

that the set N WD ¹.x; y/ 2 X Y j f .x; y/ ¤ g.x; y/º 2 .A ˝ B/ has measure
zero, i.e., . ˝ / .N / D 0. By, the definition of the completion, there exists a set
Q 2 A ˝ B such that N Q and . ˝ /.Q/ D 0. Thus
Z Z
.Qx / d.x/ D .Qy / d.y/ D 0:
X Y
Hence, by Lemma 1.49,

.E/ D 0; E WD ¹x 2 X j .Qx / ¤ 0º;
.F / D 0; F WD ¹y 2 Y j .Qy / ¤ 0º:

Since f D g on .X Y / n Q, we have fx D gx on Y n Qx for all x 2 X and

f y D g y on X n Qy for all y 2 Y . By Theorem 7.20, for g 2 L1 . ˝ / there are
measurable sets E 0 2 A and F 0 2 B such that .E 0 / D .F 0 / D 0 and
gx 2 L1 ./ for all x 2 X n E 0 ;
g y 2 L1 ./ for all y 2 Y n F 0 :
If x 2 X n .E [ E 0 /, then .Qx / D 0 and fx D gx on Y n Qx . Since .Y; B; /

is complete and gx 2 L1 ./, every function that differs from gx on a set of
measure zero is also B-measurable and -integrable. Hence fx 2 L1 ./ for all
x 2 X n .E [ E 0 /. The same argument shows that f y 2 L1 ./ for all
y 2 Y n .F [ F 0 /. Define the
W X ! R
by
8Z
< fx d for x 2 X n .E [ E 0 /;
.x/ WD Y
0 for x 2 E [ E 0
:
and
WY !R
by
8Z
< f y d for y 2 Y n .F [ F 0 /;
.y/ WD X
0 for y 2 F [ F 0 :
:
Since Z
.x/ D gx d for all x 2 X n .E [ E 0 /,
Y
it follows from part (i) of Theorem 7.20 for g that 2 L1 ./. The same argument,
using part (ii) of Theorem 7.20 for g, shows that 2 L1 ./. Moreover, the three
integrals in (7.20) for f agree with the corresponding integrals for g because
.E [ E 0 / D .F [ F 0 / D . ˝ /.Q/ D 0:
Hence equation (7.20) for f follows from part (iii) of Theorem 7.20 for g. This
Example 7.24. Assume .X; A; / is not complete. Then there exist a E 2 2X n A

and N 2 A such that E N and .N / D 0. In this case E Y is a null set in the
completion .X Y; .A ˝ B/ ; . ˝ / /. Therefore,
f WD E Y 2 L1 .. ˝ / /:
However, the function f y D E is not measurable for every y 2 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 the map
f W X Y ! Œ0; 1 is .A ˝ B/ -measurable. Prove that fx is B-measurable
for -almost all x 2 X, that f y is A-measurable for -almost all y 2 Y , and
that (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 mea-
sures 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
W 2X Y ! Œ0; 1
by
1
ˇ A 2 A; Bn 2 B for n 2 N
²X ˇ ³
.W / WD inf .An / .Bn / ˇˇ n S1 (7.21)
and W nD1 .An Bn /
nD1
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 WD jC W C ! Œ0; 1
the primitive product measure. By Carathéodory’s theorem (Theorem 2.4), the mea-
sure space .X Y; C; 1 / is complete. By definition,
1 .A B/ D .A/ .B/
for all A 2 A and all B 2 B. Fremlin then defines the complete locally determined
(CLD) product measure
0 W C ! Œ0; 1
by
ˇ E 2 A; F 2 B;
² ˇ ³
0 .W / WD sup 1 .W \ .E F // ˇ
ˇ : (7.22)
.E/ < 1; .F / < 1
He shows that .X Y; C; 0 / is a complete measure space, that 0 1 , and
1 .W / < 1 H) 0 .W / D 1 .W / for all W 2 C.
(See [4, Theorem 251I].) One can also prove that a measure W C ! Œ0; 1 satisfies
.E F / D .E/ .F / for all E 2 A and F 2 B with .E/ .F / < 1
if and only if 0 1 . With these definitions Fubini’s 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 second)
S
is either strictly localizable (i.e., there is a partition X D i 2I Xi into measur-
able sets with .Xi / < 1 such that a set A X is A-measurable if and only
if A \ Xi 2 A for all i 2 I and, moreover, .A/ D
P
i 2I .A \ Xi / for all
A 2 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 2 A for all E 2 A with .E/ < 1).
See Fremlin [4, Theorem 252B] for details.
If the measure spaces .X; A; / and .Y; B; / are both -finite, then the mea-
sures 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 (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 W BX ! Œ0; 1 and Y W BY ! Œ0; 1 be Borel measures.
Define
ƒW Cc .X Y / ! R
by
Z Z
ƒ.f / WD f .x; y/ d.y/ d.x/
X Y
Z Z (7.23)
D f .x; y/ d.x/ d.y/
Y X
for f 2 Cc .X Y /. That the two integrals agree for every continuous function with
compact support follows from Fubini’s Theorem (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 WD ¹x 2 X j .¹xº Y / \ K ¤ ;º and
KY WD ¹y 2 Y j .X ¹yº/ \ K ¤ ;º.) Since ƒ is a positive linear functional,

the Riesz representation theorem (Theorem 3.15) asserts that there exists a unique
outer regular Borel measure 1 W BXY ! Œ0; 1 that is inner regular on open sets
and a unique Radon measure 0 W BXY ! Œ0; 1 such that
Z Z
ƒ.f / D f d0 D f d1
XY X Y
for all f 2 Cc .X Y /. It turns out that in this situation the Borel -algebra BX Y
is contained in the -algebra C 2X Y of Remark 7.26 and
0 D 0 jBXY ; 1 D 1 jBXY :
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 D 1 D X ˝ Y
is the product measure of Definition 7.10 and 0 D 1 W C ! Œ0; 1 is its comple-
tion. (See Theorem 3.15 and Remark 7.26.)
7.4 Fubini and Lebesgue

For n 2 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; ` 2 N we identify RkC` 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` D BkC` ; .mk jBk / ˝ .m` jB` / D mkC` jBkC` : (7.24)
(See Exercise 7.16.) Thus Theorem 7.17 has the following consequence.
Theorem 7.28 (Fubini and Borel). Let k; ` 2 N and n WD kC`. Let f W Rn ! Œ0; 1
be Borel measurable. Then
fx WD f .x; /W R` ! Œ0; 1

and
f y WD f .; y/W Rk ! Œ0; 1
are Borel measurable for all x 2 Rk and all y 2 R` . Moreover, the functions
Z
k
R ! Œ0; 1W x 7 ! f .x; y/ d m` .y/
R`
7.4. Fubini and Lebesgue 273
and
Z
R` ! Œ0; 1W y 7 ! f .x; y/ d mk .x/
Rk
are Borel measurable, and

Z Z Z
f d mn D f .x; y/ d m` .y/ d mk .x/
Rn Rk R`
Z Z (7.25)
D f .x; y/ d mk .x/ d m` .y/:
R` Rk
Proof. The assertion follows directly from (7.24) and Theorem 7.17.
For Lebesgue measurable functions f W Rn ! Œ0; 1 the analogous statement is

considerably more subtle. In that case the function fx , respectively f y , need not
be Lebesgue measurable for all x, respectively all y. However, they are Lebesgue
measurable for almost all x 2 Rk , respectively almost all y 2 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 D 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; ` 2 N, define n WD k C `, 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 D .Ak ˝ A` / and mn D .mk ˝ m` / :
Proof. Define
Cn WD ¹Œa1 ; b1 / Œan ; bn / j ai ; bi 2 R and ai < bi for i D 1; : : : ; nº;

n
so that Cn Bn An 2R for all n. We prove the assertion in three steps.
Step 1. Bn Ak ˝ A` and mn .B/ D .mk ˝ m` /.B/ for all B 2 Bn .
By Lemma 7.6, we have Bn D Bk ˝B` Ak ˝A` . It then follows from the unique-
ness of a normalized translation-invariant Borel measure on Rn in Theorem 2.1 that
mn jBn D .mk ˝ m` /jBn . Here is a more direct proof.
First, assume B D E D Œa1 ; b1 / Œan ; bn / 2 Cn . Define
E 0 WD Œa1 ; b1 / Œak ; bk /; E 00 WD ŒakC1 ; bkC1 / Œan ; bn /:
Thus E 0 2 Ck Ak , E 00 2 C` A` , and so E D E 0 E 00 2 Ak ˝ A` . Moreover

n
Y
mn .E/ D .bi ai / D mk .E 0 / m` .E 00 / D .mk ˝ m` /.E/:
iD1
Second, assume B D U Rn is open. Then there is a sequence of pairwise disjoint

sets Ei 2 Cn such that U D 1
i D1 Ei . Hence U 2 Ak ˝ A` and
S
1
X 1
X
.mk ˝ m` /.U / D .mk ˝ m` /.Ei / D mn .Ei / D mn .U /:
i D1 iD1
Thus every open set is an element of Ak ˝ A` and so Bn Ak ˝ A` . Third,

assume B D K Rn is compact. Then there is an open set U Rn such that
K U and mn .U / < 1. Hence the set V WD U n K is open. This implies that
K D U n V 2 Ak ˝ A` and
.mk ˝ m` /.K/ D .mk ˝ m` /.U / .mk ˝ m` /.V /

D mn .U / mn .V /
D mn .K/:
Now let B Rn be any Borel set. Then B 2 Ak ˝ A` , as we have seen above.

Moreover, it follows from Theorem 2.13 that
mn .B/ D inf mn .U / D inf .mk ˝ m` /.U / .mk ˝ m` /.B/

U B U B
U is open U is open
and
mn .B/ D inf mn .K/ D inf .mk ˝ m` /.K/ .mk ˝ m` /.B/:

KB KB
K is compact K is compact
Hence, mn .B/ D .mk ˝ m` /.B/, and this proves Step 1.

Step 2. Ak ˝ A` An .
We claim that
E 2 Ak H) E R` 2 An : (7.26)
To see this, fix a set E 2 Ak . Then there exist Borel sets A; B 2 Bk such that
A E B and mk .B n A/ D 0. Let W Rn Rk denote the projection onto
7.4. Fubini and Lebesgue 275
the first k coordinates. This map is continuous and hence Borel measurable by
Theorem 1.20. Thus A R` D 1 .A/ and B R` D 1 .B/ are Borel sets in
Rn . Moreover, by Step 1,
mn ..B R` / n .A R` // D mn ..B n A/ R` /
D .mk ˝ m` /..B n A/ R` /
D mk .B n A/ m` .R` /
D 0:
Since A R` E R` B R` , it follows that E R` 2 An , as claimed

in (7.26). A similar argument shows that
F 2 A` H) Rk F 2 An :
Hence E F D .E R` / \ .Rk F / 2 An for all E 2 Ak and all F 2 A` . Thus

Ak ˝ A` An and this proves Step 2.
Step 3. .Ak ˝ A` / D An and .mk ˝ m` / D mn .
Let A 2 An . Then there are Borel sets B0 ; B1 2 Bn such that B0 A B1 and

mn .B1 n B0 / D 0. By Step 1, B0 ; B1 2 Ak ˝ A` and .mk ˝ m` /.B1 n B0 / D 0.
Hence A 2 .Ak ˝ A` / and
.mk ˝ m` / .A/ D .mk ˝ m` /.B0 / D mn .B0 / D mn .A/:
Thus we have proved that
An .Ak ˝ A` / ; .mk ˝ m` / jAn D mn :
Since Ak ˝ A` An by Step 2, it follows that
mn jAk ˝A` D .mk ˝ m` / jAk ˝A` D mk ˝ m` :
Now let A 2 .Ak ˝ A` / . Then there are sets A0 ; A1 2 Ak ˝ A` such that

A0 A A1 and .mk ˝ m` /.A1 n A0 / D 0. Hence A0 ; A1 2 An by Step 2 and
mn .A1 n A0 / D 0. Since .Rn ; An ; mn / is complete, it follows that A n A0 2 An
and so A D A0 [ .A n A0 / 2 An . Hence An D .Ak ˝ A` / . This proves Step 3
and Theorem 7.29.
The next result specializes Theorem 7.23 to the Lebesgue measure.
Theorem 7.30 (Fubini and Lebesgue). Let k; ` 2 N and n WD kC`. Let f W Rn ! R

be Lebesgue integrable and define
fx .y/ WD f y .x/ WD f .x1 ; : : : ; xk ; y1 ; : : : ; y` /
for x D .x1 ; : : : ; xk / 2 Rk and y D .y1 ; : : : ; y` / 2 R` . Then there are Lebesgue

null sets E Rk and F R` such that the following holds.
(i) fx 2 L1 .R` / for every x 2 Rk n E and the map
W Rk ! R
defined by
8Z
< fx d m` for x 2 Rk n E;
.x/ WD R` (7.27)
0 for x 2 E;
:
is Lebesgue integrable.
(ii) f y 2 L1 .Rk / for every y 2 R` n F and the map
W R` ! R
defined by
8Z
ˆ
< f y d mk for y 2 R` n F;
.y/ WD Rk (7.28)
ˆ
0 for y 2 F;
:
is Lebesgue integrable.

Z Z Z
d mk D f d mn D d m` : (7.29)
Rk Rn R`
Proof. This follows directly from Theorem 7.23 and Theorem 7.29.
7.5. Convolution 277
7.5 Convolution
An application of Fubini’s Theorem is the convolution product on the space of
Lebesgue integrable functions on Euclidean space. Fix an integer n 2 N and
let .Rn ; A; m/ be the Lebesgue measure space. The convolution of two Lebesgue
integrable functions f; g 2 L1 .Rn / is defined by
Z
.f g/.x/ WD f .x y/g.y/ d m.y/ for almost all x 2 Rn :
Rn
Here the function Rn ! RW y 7! f .x y/g.y/ is Lebesgue integrable for al-

most every x 2 Rn and the resulting almost everywhere defined function f g is
again Lebesgue integrable. This is the content of Theorem 7.33. The convolution
descends to a bilinear map
W 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 func-
tions f; gW Rn ! R to mean that they agree almost everywhere with respect to the
Lebesgue measure.
Definition 7.31. Let f; gW Rn ! R be Lebesgue measurable and define

² ˇ n ³
ˇ R ! RW y 7! f .x y/g.y/
E.f; g/ WD x 2 Rn ˇ
ˇ is not Lebesgue integrable : (7.30)
The convolution of f and g is the function
f gW Rn ! R
defined by
8Z
< f .x y/g.y/ d m.y/ for x 2 Rn n E.f; g/;
.f g/.x/ WD Rn (7.31)
0 for x 2 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 W 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 2 Rn ;
(ii) if f 0 f and g 0 g, then E.f 0 ; g 0 / D E.f; g/ and f 0 g 0 D f g;
(iii) E.f; g/ is a Borel set and f g is Borel measurable;
(iv) E.g; f / D E.f; g/ and g f D f g;
(v) if m.E.f; g// D m.E.g; h// D 0, then
E WD E.jf j; jgj jhj/ D E.jf j jgj; jhj/
and f .g h/ D .f g/ h on Rn n E.
An example with n D 1, where E.f; g h/ D E.f g; h/ D ; and E D R in

part (v) of Theorem 7.32, is discussed in Exercise 7.53 below.
Proof of Theorem 7.32. (i) For x 2 Rn define
fx W Rn ! R and x W Rn ! Rn
by
fx .y/ WD f .x y/ and x .y/ WD x y:
Then x is a diffeomorphism and jdet.dx /j 1. Hence, Theorem 2.17 asserts

that fx D f ı x is Lebesgue measurable for all x 2 Rn .
(ii) By assumption,
A WD ¹y 2 Rn j f .y/ ¤ f 0 .y/º
and
B WD ¹y 2 Rn j g.y/ ¤ g 0 .y/º
are Lebesgue null sets. Hence so are the sets
Cx WD x .A/ [ B D ¹y 2 Rn j f .x y/ ¤ f 0 .x y/ or g.y/ ¤ g 0 .y/º
for all x 2 Rn . Hence the functions fx g and fx0 g 0 agree on the complement of a
Lebesgue null set for every x 2 Rn . Hence they are either both integrable or both
not integrable and when they are their integrals agree.
(iii) By (ii) and Theorem 1.55, it suffices to assume that f and g are Borel mea-
surable. Now define
F; GW R2n ! R
and
W R2n ! R2n
by
F .x; y/ WD f .x y/g.y/;
G.x; y/ WD f .x/g.y/;
.x; y/ WD .x y; y/
for x; y 2 Rn . Then G is Borel measurable and is a diffeomorphism. Hence

preserves the Borel -algebra and this implies that
F DGı
is Borel measurable. Hence the function

Z
n
R ! Œ0; 1W x 7 ! jF .x; y/j d m.y/
Rn
is Borel measurable, by Fubini’s Theorem (Theorem 7.28). Thus the set E.f; g/
where this function takes on the value 1 is a Borel set. Moreover, the functions
F ˙ WD max¹˙F; 0º
are Borel measurable and so are the functions

˙
Fz W R2n ! Œ0; 1/
defined by
´
˙ F ˙ .x; y/ if x 2 Rn n E.f; g/;
Fz .x; y/ WD for .x; y/ 2 R2n :
0 if x 2 E.f; g/;
Since Z Z
C
.f g/.x/ D Fz .x; y/ d m.y/ Fz .x; y/ d m.y/
Rn Rn
for all x 2 Rn , it follows from Theorem 7.28 that f g is Borel measurable.

(iv) Since gx f D .fx g/ ı x , it follows from Theorem 2.17 that
E.g; f / D ¹x 2 Rn j gx f 2 L1 .Rn /º D E.f; g/
and
Z
.f g/.x/ D fx g d m
Rn
Z
D .fx g/ ı x d m
Rn
Z
D gx f d m
Rn
D .g f /.x/
for all x 2 Rn n E.f; g/.

(v) By (ii) and Theorem 1.55, it suffices to assume that f , g, and h are Borel
measurable. Let x 2 Rn and define
Fx W R2n ! R
by
Fx .y; z/ WD f .z/g.x y z/h.y/:
Thus Fx is the composition of the maps R2n ! R3n W .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 2 E.jf j; jgj jhj/ () jFx j D 1 () x 2 E.jf j jgj; jhj/: (7.32)
R2n
It follows from Theorem 7.28 that

Z Z Z
jFx j d m2n D jFx .y; z/j d m.y/ d m.z/:
R2n Rn Rn
This integral is finite if and only if Fx 2 L1 .R2n /. Moreover,

Z Z
jFx .y; z/j d m.y/ D jf .z/j jg.x y z/jjh.y/j d m.y/
Rn Rn
D jf .z/j.jgj jhj/.x z/
for z 2 Rn n .x E.g; h//. Since E.g; h/ is a Lebesgue null set, it follows that
Z
kFx kL1 .R2n / D jf .z/j.jgj jhj/.x z/ d m.z/:
Rn
The integral on the right is infinite if and only if x 2 E.jf j; jgj jhj/. This proves
the first equivalence in (7.32). The proof of the second equivalence is analogous,
with y and z interchanged.
Now let x 2 Rn n E. Then Fx 2 L1 .R2n / and x 2 Rn n E.f; g h/.
Moreover, for z 2 Rn , the function Rn ! RW 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/ d m.y/ D f .z/ g.x y z/h.y/ d m.y/
Rn Rn
D f .z/.g h/.x z/:
Since E.g; h/ is a Lebesgue null set, Theorem 7.30 shows that

Z Z
Fx d m2n D f .z/.g h/.x z/ d m.z/
R2n Rn
D .f .g h//.x/
The last equality holds because x … E.f; g h/. A similar argument with y and z
interchanged shows that
Z
Fx d m2n D ..f g/ h/.x/ for all x 2 Rn n E.
R2n
This proves (v) and Theorem 7.32.
Theorem 7.33. Let 1 p; q; r 1 such that 1=p C 1=q D 1 C 1=r and let
f 2 Lp .Rn / and g 2 Lq .Rn /. Then m.E.f; g// D 0 and
kf gkr kf kp kgkq : (7.33)
Thus f g 2 Lr .Rn /. The estimate (7.33) is called Young’s inequality.

Proof. Define the function
hW Rn ! Œ0; 1
by Z
h.x/ WD jf .x y/g.y/j d m.y/ for x 2 Rn :
Rn
Then jf gj h and E.f; g/ D ¹x 2 Rn j h.x/ D 1º. Hence it suffices to
prove that khkr kf kp kgkq . For r D 1 this follows from Hölder’s inequality.
So assume r < 1. Then 1 p; q < 1. Define
p p p
WD 1 Dp ; q 0 WD :
r q
Then 0 < 1 and 1=q C 1=q 0 D 1. Also D 0 if and only if q D 1. If > 0,

then Hölder’s inequality in Theorem 4.1 shows that
Z
h.x/ D jfx j jfx j1 jgj d m kjfx j kq 0 kjfx j1 jgjkq ;
Rn
where fx .y/ WD f .x y/. Since q 0 D p, this implies

Z q=qZ0
q q 0
h.x/ jfx j d m jfx j.1 /q jgjq d m
Rn Rn (7.34)
Z
q
D kf kp jf .x y/j.1 /q jg.y/jq d m.y/
Rn
n
for all x 2 R . This continues to hold for D 0. Now it follows from Minkowski’s
inequality in Theorem 7.19 with the exponent s WD r=q 1 that
Z q=r
q r
khkr D h dm
Rn
Z 1=s
qs
D h dm
Rn
Z Z s 1=s
kf kpq jf .x y/j.1 /q
jg.y/jq d m.y/ d m.x/
Rn Rn
Z Z 1=s
kf kpq jf .x .1 /qs
y/j jg.y/j qs
d m.x/ d m.y/
Rn Rn
D kf kpq kf kp.1 /q

kgkqq :
Here the last equality follows from the fact that .1 /qs D .1 /r D 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 /W .f; g/ 7 ! f g: (7.35)
This map is bilinear by Theorem 1.44, it is associative by part (v) of Theorem 7.32,
and satisfies kf gk1 kf k1 kgk1 by Young’s inequality in Theorem 7.33.
Hence L1 .Rn / is a Banach algebra. By part (iv) of Theorem 7.32, L1 .Rn / is
commutative, and, by Theorem 7.33 with q D 1 and r D p, it acts on Lp .Rn /.
(A Banach algebra is a Banach space .X; kk/ equipped with an associative bilinear
map X X ! XW .x; y/ 7! xy that satisfies the inequality kxyk kxk kyk for all
x; y 2 X.)
Definition 7.34. Fix a constant 1 p < 1. A Lebesgue measurable function

f W Rn ! R is called locally p-integrable if
Z
jf jp d m < 1
K
for every compact set K Rn . It is called locally integrable if it is locally 1-inte-

grable.
Theorem 7.33 carries over to locally integrable functions as follows.

If 1=p C 1=q D 1 C 1=r, f is locally p-integrable, and g 2 Lq .Rn / has com-
pact 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 ˇ such that ˇjK 1. Then ˇf 2 Lp .Rn / and .ˇf / g agrees
with f g on the set ¹x 2 Rn j x supp.g/ Kº. In the following theorem
C01 .Rn / denotes the space of compactly supported smooth functions on Rn .
Theorem 7.35. Let 1 p < 1 and 1 < q 1 such that 1=p C 1=q D 1.
(i) If f W Rn ! R is locally p-integrable, then

Z
lim jf .x C / f .x/jp d m.x/ D 0
!0 B
for every bounded Lebesgue measurable subset B Rn . If f 2 Lp .Rn / this

continues to hold for B D Rn .
(ii) If f 2 Lp .Rn / and g 2 Lq .Rn /, then f g is uniformly continuous. If f is

locally p-integrable and g 2 Lq .Rn / has compact support (or if f 2 Lp .Rn /
has compact support and g is locally q-integrable), then f g is continuous.
(iii) If f W Rn ! R is locally integrable and g 2 C01 .Rn /, then f g is smooth

and @˛ .f g/ D f @˛ g for every multi-index ˛.
(iv) C01 .Rn / is dense in Lp .Rn / for 1 p < 1.
Proof. (i) Assume first that f 2 Lp .Rn / and fix a constant " > 0.
By Theorem 4.15, there is a function g 2 Cc .Rn / such that kf gkp < "1=p =3.
Since g is uniformly continuous, there is a ı > 0 such that, for all 2 Rn ,
" 1=p
jj < ı H) sup jg.x C / g.x/j < :
x2Rn 3p m.supp.g//
Take 2 Rn such that jj < ı. Then

Z 1=p
p
jf .x C / f .x/j d m.x/
Rn
Z 1=p
2 kf gkp C jg.x C / g.x/jp d m.x/
Rn
2"1=p
C .m.supp.g// sup jg.x C / g.x/jp /1=p
3 x2Rn
< "1=p :
This proves (i) for f 2 Lp .Rn /. To prove the result in general, choose a compact
set K Rn such that B1 .x/ K for all x 2 B and multiply f by a smooth
compactly supported cutoff function to obtain a function f 0 2 Lp .Rn / that agrees
with f on K. Then (i) holds for f 0 and hence also for f .
(ii) Assume first that f 2 Lp .Rn / and g 2 Lq .Rn / and fix a constant " > 0.
By part (i), there exists a ı > 0 such that, for all 2 Rn ,
Z " p
jj < ı H) jf .y C / f .y/jp d m.y/ <
Rn kgkq
Fix two elements x; 2 Rn such that jj < ı and denote
fx .y/ WD f .x y/:
Then, by Hölder’s inequality in Theorem 4.1,

ˇZ ˇ
ˇ ˇ
j.f g/.x C / .f g/.x/j D ˇˇ .fxC fx /g d mˇˇ
Rn
kfxC fx kp kgkq
Z 1=p
D jf .y C / f .y/jp d m.y/ kgkq
Rn
< ":
This shows that f g is uniformly continuous. If f is locally p-integrable and

g 2 Lq .Rn / has compact support, continuity follows by taking the integral over a
suitable compact set. In the opposite case continuity follows by taking the Lq -norm
of g over a suitable compact set.
(iii) Fix an index i 2 ¹1; : : : ; nº and denote by ei 2 Rn the i th unit vector.

Fix an element x 2 Rn and choose a compact set K Rn such that B1 .y/ K
whenever x y 2 supp.g/. Let " > 0. Since @i g is continuous, there is a constant
0 < ı < 1 such that
"
j@i g.y C hei / @i g.y/j < Z
jf j d m
K
n
for all y 2 R and all h 2 R with jhj < ı. Hence the fundamental theorem of
calculus asserts that
ˇ ˇ
ˇ g.y C hei y/ g.y/ ˇ "
sup ˇ ˇ @i g.y/ˇˇ < Z
y2R n h
jf j d m
K
for all h 2 R with 0 < jhj < ı. Take h 2 Rn with 0 < jhj < ı. Then
ˇ ˇ
ˇ .f g/.x C hei / .f g/.x/ ˇ
ˇ .f @i g/.x/ˇ
ˇ h ˇ
ˇZ g.x C he ˇ
ˇ i y/ g.x y/ ˇ
D ˇˇ f .y/ @i g.x y/ d m.y/ˇˇ
Rn h
ˇ ˇ
ˇ g.x C hei y/ g.x y/
Z
ˇ
jf .y/j ˇˇ @i g.x y/ d m.y/ˇˇ
Rn h
< ":
By part (ii), the function @i .f g/ D f @i g is continuous for i D 1; : : : ; n.

For higher derivatives the assertion follows by induction.
(iv) Let f 2 Lp .Rn / and choose a compactly supported smooth function
W Rn ! Œ0; 1/ such that
Z
supp./ B1 ; d m D 1:
Rn
Define
ı W R n ! R
by
1 x
ı .x/ WD
ın ı
for ı > 0 and x 2 Rn . Then
Z
supp.ı / Bı ; ı d m D 1
Rn
by Theorem 2.17. By part (iii), the function
fı WD ı f W Rn ! R
is smooth for all ı > 0. Now fix a constant " > 0. By part (i), there exists a constant
ı > 0 such that, for all y 2 Rn ,
Z
jyj < ı H) jf .x y/ f .x/jp d m.x/ < "p :
Rn
Hence, by Minkowski’s inequality in Theorem 7.19,

Z ˇZ ˇp 1=p
ˇ ˇ
kfı f kp D ˇ .f .x y/ f .x//ı .y/ d m.y/ˇ d m.x/
ˇ
n Rn
ˇ
R
Z Z 1=p
jf .x y/ f .x/jp ı .y/p d m.x/ d m.y/
Rn Rn
Z 1=p
p
sup jf .x y/ f .x/j d m.x/
jyj<ı Rn
":
If f has compact support, then so does fı . If not, choose a function g 2 Lp .Rn /

with compact support such that kf gkp < "=2 and then a smooth function
hW 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 ı can be viewed as approximate Dirac delta
functions that concentrate near the origin as ı tends to zero.
7.6 Marcinkiewicz interpolation
Another interesting application of Fubini’s Theorem is Marcinkiewicz interpola-

tion, 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 mea-
sure 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
Calderón–Zygmund inequality in Section 7.7.
7.6. Marcinkiewicz interpolation 287
Let .X; A; / be a measure space. For a measurable function f W X ! R define

the function
f W Œ0; 1/ ! Œ0; 1
by (6.1), i.e.,
f .t/ WD .A.t; f //; A.t; f / WD ¹x 2 X j jf .x/j > tº; for t 0.
The function f is nonincreasing and hence Borel measurable.

Lemma 7.36. Let 1 p < 1 and let f; gW X ! R be measurable. Then
f Cg .t/ f .t=2/ C g .t=2/; (7.36)

Z Z 1
t p f .t/ jf jp d D p s p 1 f .s/ ds (7.37)
X 0
for all t 0.
Proof. Inequality (7.36) was established in the proof of Lemma 6.2. We prove (7.37)
in four steps.
Step 1. For all t 0, Z
p
t f .t/ jf jp d:
X
Since t p A.t;f / jf jp , it follows that
Z Z
t p f .t/ D t p A.t;f / d jf jp d for all t 0.
X X
This proves Step 1.

Step 2. If f .t/ D 1 for some t > 0, then
Z Z 1
p
jf j d D 1 D tp 1
f .t/ dt:
X 0
By Step 1, we have Z
jf jp d D 1:
X
Moreover, t p 1
f .t/ D 1 for t > 0 sufficiently small, and hence
Z 1
t p 1 f .t/ dt D 1:
0
This proves Step 2.

Step 3. Assume .X; A; / is -finite and f .t/ < 1 for all t > 0. Then equa-
tion (7.37) holds.
Let B 2Œ0;1/ be the Borel -algebra and denote by mW B ! Œ0; 1 the restriction
of the Lebesgue measure to B. Let .X Œ0; 1/; A ˝ B; ˝ m/ be the product
measure space of Definition 7.10. We prove that
Q.f / WD ¹.x; t/ 2 X Œ0; 1/ j 0 t < jf .x/jº 2 A ˝ B:
To see this, assume first that f is an A-measurable step-function. Then there exist
finitely many pairwise disjoint measurable sets A1 ; : : : ; A` 2 A and positive real
numbers ˛1 ; : : : ; ˛` such that
`
X
jf j D ˛i A :
i
iD1
In this case Q.f / D ìD1 Ai Œ0; ˛i / 2 A ˝ B. Now consider the general case.
S
Then Theorem 1.26 asserts that there is a sequence of A-measurable step-functions

fi W X ! Œ0; 1/ such that 0 f1 f2 and fi converges pointwise to jf j.
Then Q.fi / 2 A ˝ B for all i, and so Q.f / D 1 iD1 Q.fi / 2 A ˝ B.
S
Now define
hW X Œ0; 1/ ! Œ0; 1/
by
h.x; t/ WD pt p 1
:
This function is A ˝ B-measurable, and so is hQ.f / . Hence, by Fubini’s Theorem

(Theorem 7.17),
!
Z Z Z jf .x/j
jf jp d D pt p 1
dt d.x/
X X 0
Z Z 1
D .hQ.f / /.x; t/ d m.t/ d.x/
X 0
Z 1 Z
D .hQ.f / /.x; t/ d.x/ d m.t/
0 X
Z 1
D pt p 1
.A.t; f // dt:
0
This proves Step 3.

Step 4. Assume f .t/ < 1 for all t > 0. Then (7.37) holds.
Define
X0 WD ¹x 2 X j f .x/ ¤ 0º;
A0 WD ¹A 2 A j A X0 º;
and
0 WD jA0 :
Then the measure space .X0 ; A0 ; 0 / is -finite because Xn WD A.1=n; f / is a
sequence of An -measurable sets such that 0 .Xn / D f .1=n/ < 1 for all n and
X0 D 1 nD1 Xn . Moreover, f0 WD f jX0 W X0 ! R is A0 -measurable and f D f0 .
S
Hence it follows from Step 3 that

Z Z Z 1 Z 1
p p p 1
jf j d D jf0 j d0 D t f0 .t/ dt D t p 1 f .t/ dt:
X X0 0 0
This proves Step 4 and Lemma 7.36.

Fix real numbers 1 p q. Then the inequality
q p q.p 1/
kf kp kf k1p.q 1/
kf kqp.q 1/
(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 Józef Marcinkiewicz
(a Ph.D. 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 W Lq ./ ! Lq ./ be a linear
operator. Suppose that there exist constants c1 > 0 and cq > 0 such that
kTf k1;1 c1 kf k1 ; kTf kq cq kf kq (7.39)
for all f 2 L1 ./ \ Lq ./. Fix a constant 1 < p < q. Then
p.q 1/ 1=p q p q.p 1/
kTf kp cp kf kp ; cp WD 2 c1p.q 1/ cqp.q 1/
; (7.40)
.q p/.p 1/
for all f 2 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 2 L1 ./ \ L2 ./. For t 0 define

´
f .x/ if jf .x/j > ct;
f t .x/ WD
0 if jf .x/j ct;
and
´
0 if jf .x/j > ct;
g t .x/ WD
f .x/ if jf .x/j ct:
Then
´
A.s; f / if s > ct;
A.s; f t / D
A.ct; f / if s ct;
´
; if s ct;
A.s; g t / D
A.s; f / n A.ct; f / if s < ct;
and
´
f .s/ if s > ct;
f t .s/ D
f .ct/ if s ct;
´
0 if s ct;
g t .s/ D
f .s/ f .ct/ if s < ct:
By Lemma 7.36 and Fubini’s Theorem (Theorem 7.28), this implies that
Z 1 Z 1 Z 1
p 2 p 2
t kf t k1 dt D t f t .s/ ds dt
0 0 0
Z 1 Z 1
p 2
D t ctf .ct/ C f .s/ ds dt
0 ct
Z 1 Z 1 Z s=c
D c1 p
tp 1
f .t/ dt C tp 2
dt f .s/ ds
0 0 0
1 1
.s=c/p 1
Z Z
D c1 p
tp 1
f .t/ dt C f .s/ ds
0 0 p 1
1 p 1
c p
Z
D tp 1
f .t/ dt
p 1 0
c1 p Z
D jf jp d;
p 1 X
and
Z 1
tp q 1
kg t kqq dt
0
Z 1 Z 1
p q 1 q 1
D t qs g t .s/ ds dt
0 0
Z 1 Z ct
p q 1 q 1
D t qs .f .s/ f .ct// ds dt
0 0
Z 1 Z 1 Z 1
p q 1 q 1 q
Dq t dt s f .s/ ds c tp 1
f .ct/ dt
0 s=c 0
1 1
sp 1cq p
Z Z
Dq f .s/ ds cq p
tp 1
f .t/ dt
0 q p 0
cq p p 1 p 1
Z
D t f .t/ dt
q p 0
cq p
Z
D jf jp d:
q p X
Moreover, f D f t C g t for all t 0. Hence, by Lemma 7.36 and (7.39),
Tf .t/ Tf t .t=2/ C T g t .t=2/

2 2q
kTf t k1;1 C q kT g t kqq
t t
2c1 .2cq /q
kf t k1 C kg t kqq :
t tq
Hence, by Lemma 7.36,
Z Z 1
p
jTf j d D p tp 1
Tf .t/ dt
X 0
Z 1 Z 1
p2c1 t p 2
kf t k1 dt C p.2cq / q
tp q 1
kg t kqq dt
0 0
p2c c 1 p
p.2cq / c q q pZ
1
D C jf jp d
p 1 q p X
p.q 1/2p c1.q p/=.q 1/ .qp q/=.q 1/

cq
Z
D jf jp d:
.q p/.p 1/ X
Here the last equality follows with the choice of c WD .2c1 /1=.q 1/
=.2cq /q=.q 1/
.
7.7 The Calderón–Zygmund inequality

The convolution product discussed in Section 7.5 has many applications, notably in
the theory of partial differential equations. One such application is the Calderón–
Zygmund inequality, which plays a central role in the regularity theory for elliptic
equations. Its proof requires many results from measure theory, including Fubini’s
Theorem, convolution, Marcinkiewicz interpolation, Lebesgue’s’ differentiation
theorem, and the dual space of Lp . Denote the standard Laplace operator
on Rn by
n
X @2
WD 2
(7.41)
i D1
@x i
and, for i D 1; : : : ; n, denote the partial derivative with respect to the i th coordinate
by @i D @=@xi . Denote the open ball of radius r > 0 centered at the origin by
Br WD ¹x 2 Rn j jxj < rº. Call a function uW Rn ! R smooth if all its partial
derivatives exist and are continuous. Denote by C01 .Rn / the space of compactly
supported smooth functions on Rn .
Definition 7.38. Fix an integer n 2. The fundamental solution of Laplace’s

equation is the function KW Rn n ¹0º ! R defined by
8
<.2/ 1 log.jxj/ if n D 2;
K.x/ WD (7.42)
:.2 n/ 1 ! 1 jxj2 n if n > 2:
n
Here !n denotes the area of the unit sphere S n 1

Rn or, equivalently,
!n =n WD m.B1 /
denotes the Lebesgue measure of the unit ball in Rn . The first and second partial
derivatives
Ki WD @i K and Kij WD @i @j K
of the fundamental solution are given by
xi nxi xj jxj2 nxi2

Ki .x/ D ; Kij .x/ D ; Ki i .x/ D (7.43)
!n jxjn !n jxjnC2 !n jxjnC2
for 1 i; j n with i ¤ j . Extend the functions K, Ki , Kij to all of Rn by

setting
K.0/ WD Ki .0/ WD Kij .0/ WD 0 for all i; j .
7.7. The Calderón–Zygmund inequality 293
Exercise 7.39. Prove that K D 0. Prove that K and Ki are locally integrable,
while Kij is not Lebesgue integrable over any neighborhood of the origin.
Hint. Use Fubini’s Theorem in polar coordinates (Exercise 7.47).
Exercise 7.40. Prove that

m.B1 / D !n =n
and that 8
ˆ 2 n=2
ˆ
ˆ
< .n=2 1/Š if n is even;
2 n=2
ˆ
!n D D (7.44)
.n=2/ ˆ
ˆ 2.nC1/=2 .n 1/=2

if n is odd:
ˆ
ˆ
1 3 .n 2/
:
2
Hint. Use Fubini’s Theorem to prove that Rn e jxj d m.x/ D 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 2 C01 .Rn /. Then
f D K f: (7.45)
Moreover, the function

uW Rn ! R;
defined by Z
u.x/ WD .K f /.x/ D K.x y/f .y/ d m.y/ (7.46)
Rn
for x 2 Rn , is smooth and satisfies
u D f; @i u D Ki f for i D 1; : : : ; n: (7.47)
The equations (7.45) and (7.47) are called Poisson’s identities.
Proof. The proof relies on Green’s formula

Z Z @v @u
.uv vu/ d m D u v d (7.48)
@ @ @
for a bounded open set Rn with smooth boundary @ and two smooth func-
tions u; vW Rn ! R. The term
n
@u X @u
.x/ WD i .x/ .x/
@ @xi
i D1
where x 2 @ denotes the outward normal derivative and W @ ! S n 1

denotes the outward pointing unit normal vector field on the boundary. The in-
tegral 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 defi-
nition because the boundary integral will only be needed here when the boundary
component is a sphere (see Exercise 7.47 below). Formula (7.48) can be viewed as
a higher-dimensional analogue of the fundamental theorem of calculus.
Now let f 2 C01 .Rn / and choose r > 0 so large that supp.f / Br . Fix an
element 2 supp.f / and a constant " > 0 such that Bx" ./ Br . Choose
WD Br n Bx" ./; u.x/ WD K .x/ WD K. x/; v WD f:
Then @ D @Br [ @B" ./ and the functions v and @v=@ vanish on @Br . Moreover,
K 0. Hence Green’s formula (7.48) asserts that
Z Z @K @f

K f d m D f K d: (7.49)
Rn nB" ./ @B" ./ @ @
Here the reversal of sign arises from the fact that the outward unit normal vector on
@B" ./ is inward pointing with respect to . Moreover,
1
.x/ D jx j .x / for x 2 @B" ./,
so
@K =@.x/ D !n 1 "1 n
by (7.43). Also, by (7.42),

´
1
2 log."/ if n D 2;
K .x/ D DW ."/ for x 2 @B" ./:
1
.2 n/ !n 1 "2 n
if n > 2;

1
Z Z Z
K f d m D n 1
u d ."/ f d m: (7.50)
n
R nB" ./ ! n " @B" ./ B" ./
The last summand is obtained from (7.48) with u D 1, v D f , D 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 D .K f / D K f D f:
To prove the second equation in (7.47), fix an index i 2 ¹1; : : : ; nº and a point
2 Rn . Then the divergence theorem on
WD Br n Bx" ./
asserts that
Z
.Ki . x/f .x/ K. x/@i f .x// d m.x/
Rn nB" ./
Z
D ..@i K /f C K @i f / d m
Rn nB" ./
Z
D @i .K f / d m
Rn nB" ./
Z
D i K f d
@B" ./
xi i
Z
D ."/ f .x/ d.x/:
@B" ./ "
The last term converges to zero as " tends to zero. Hence
.Ki f /./ D .K @i f /./ D @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 W Rn ! R and asserts that K f is C 2 . However, this does not hold for contin-
uous functions with compact support. A counterexample is u.x/ D jxj3 , which is
not C 2 and satisfies f WD u D 3.n C 1/jxj. It then follows that K ˇf (for any
ˇ 2 C01 .Rn / equal to 1 near the origin) cannot be C 2 .
Theorem 7.43 (Calderón–Zygmund). Fix both an integer n 2 and a number

1 0 such that
n
X
@i @j u c kuk (7.51)
p p
i;j D1
for all u 2 C01 .Rn /.
Proof. See page 304. The proof is based on the exposition in Gilbarg and
Trudinger [5].
The Calderón–Zygmund inequality is a beautiful and deep result in the the-

ory of partial differential equations. It extends to all functions u D K f with
f 2 C01 .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 Laplace’s equation are not locally integrable and,
nevertheless, the Calderón–Zygmund inequality still asserts that the second deriva-
tives of its convolution u D K f with a p-integrable function f are p-integrable.
Despite this subtlety, the proof is elementary in the case p D 2. Denote by
ru WD .@1 u; : : : ; @n u/W Rn ! Rn
the gradient of a smooth function uW Rn ! R.
Lemma 7.44. Fix an integer n 2 and let f 2 C01 .Rn /. Then
kr.Kj f /k2 kf k2 for j D 1; : : : ; n: (7.52)
Proof. Define u WD Kj f . This function is smooth by Theorem 7.35, but it need
not have compact support. By the divergence theorem,
n
@u
Z Z Z X Z
2
jruj d m C uu d m D @i .u@i u/ d m D u d (7.53)
Br Br Br iD1 @Br @
for all r > 0. By Poisson’s identities (7.45) and (7.47), we have

u D .Kj f / D @j .K f / D @j .K f / D @j f:
Since f has compact support, it follows from (7.43) that there is a constant
c > 0 such that ju.x/j C j@u=@.x/j cjxj1 n for jxj sufficiently large. Hence
the integral on the right in (7.53) tends to zero as r tends to infinity. Thus
Z
2
kruk2 D jruj2 d m
Rn
Z
D u@j f d m
Rn
Z
D .@j u/f d m kruk2 kf k2 :
Rn
By Theorem 7.35, the space C01 .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 Calderón–
Zygmund inequality is the following delicate argument which shows that this oper-
ator also extends to a continuous linear operator from the Banach space L1 .Rn / to
the topological vector space L1;1 .Rn / of weakly integrable functions introduced
in Section 6.1. This argument occupies the next six pages. Recall the definition
kf k1;1 WD sup tf .t/;

t>0
where
f .t/ WD m.A.t; f //; A.t; f / WD ¹x 2 Rn j jf .x/j > tº:
(See equation (6.1).)
Lemma 7.45. Fix an integer n 2. Then there is a constant c D c.n/ > 0 such
that
k@k .Kj f /k1;1 ckf k1 (7.54)
for all f 2 C01 .Rn / and all indices j; k D 1; : : : ; n.
Proof. Fix two integers j; k 2 ¹1; : : : ; nº and let T W L2 .Rn / ! L2 .Rn / be the
unique bounded linear operator that satisfies
Tf D @k .Kj f / (7.55)
for f 2 C01 .Rn /. This operator is well defined by Lemma 7.44. We prove in three
steps that there is a constant c D c.n/ > 0 such that kTf k1;1 c kf k1 for all
f 2 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 D c.n/ 1 such that the following holds. If B Rn

is a countable union of closed cubes Qi Rn with pairwise disjoint interiors and
if h 2 L2 .Rn / \ L1 .Rn / satisfies
Z
hjR nB 0;
n h d m D 0 for all i 2 N; (7.56)
Qi
then
1
T h .t/ c m.B/ C khk1 (7.57)
t
for all t > 0.
For i 2 N define
hi W Rn ! R
by
´
h.x/ if x 2 Qi ;
hi .x/ WD
0 if x … Qi :
Denote by qi 2 Qi the center of the cube Qi and by 2ri > 0 its side length. Then
p
jx qi j nri for all x 2 Qi . Fix an element x 2 Rn n Qi . Then Kj is smooth
on x Qi and so Theorem 7.35 asserts that
.T hi /.x/ D .@k Kj hi /.x/

Z (7.58)

D @k Kj .x y/ @k Kj .x qi / hi .y/ d m.y/:
Qi
This identity is more delicate than it looks at first glance. To see this, note that
equation (7.55) only holds for compactly supported smooth functions, 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 D supp.hi /
one can approximate hi in L2 .Rn / by a sequence of compactly supported smooth
functions that vanish near x (by using the mollifier method in the proof of Theo-
rem 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 equality in (7.58) follows
by taking the limit. The second equality follows from (7.56). It follows from (7.58)
that
Z
j.T hi /.x/j j@k Kj .x y/ @k Kj .x qi /jjhi .y/j d m.y/
Qi
sup j@k Kj .x y/ @k Kj .x qi /j khi k1

y2Qi
p
nri sup jr@k Kj .x y/j khi k1
y2Qi
1
c1 ri sup khi k1
y2Qi jx yjnC1
c1 ri
D khi k1 :
d.x; Qi /nC1
Here d.x; Qi / WD inf jx yj and

y2Qi
n.n C 3/
c1 D c1 .n/ WD max sup jyjnC1 jr@k Kj .y/j :
j;k y2Rn n¹0º !n
Here the last inequality follows by differentiating equation (7.43).
Now define
p
Pi WD ¹x 2 Rn j jx qi j < 2 nri º Qi :
Then p
d.x; Qi / jx qi j nri for all x 2 Rn n Pi .
Hence
1
Z Z
jT hi j d m c1 ri p d m.x/ khi k1
Rn nPi Rn nPi .jx qi j nri /nC1
1
Z
D c1 ri p p d m.y/ khi k1
jyj>2 nri .jyj nri /nC1
1
!n s n 1 ds
Z
D c1 ri p p khi k1
2 nri .s nri /nC1
Z 1 p
.s C nri /n 1 ds
D c1 !n ri p khi k1
nri s nC1
Z 1
n 1 ds
c1 2 !n ri p 2
khi k1
nri s
D c2 khi k1 :
p
Here c2 D c2 .n/ WD c1 .n/2n 1 !n n 2n 1 n3=2 .n C 3/. The third step in
the above computation follows from Fubini’s Theorem in polar coordinates (Exer-
cise 7.47). Thus we have proved that
Z
jT hi j d m c2 khi k1 for all i 2 N: (7.59)
Rn nPi
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 2 L2 .Rn /. We prove that there is a Lebesgue null set
E Rn such that
1
X
jT h.x/j jT hi .x/j for all x 2 Rn n E: (7.60)
iD1
To see this, note that the sequence ìD1 hi converges to h in L2 .Rn / as ` tends to
P
P`
infinity. So the sequence iD1 T hi converges to T h in L2 .Rn / by Lemma 7.44.
By Corollary 4.10, a subsequence converges almost everywhere. Hence there exist
a Lebesgue null set E Rn and a sequence of integers 0 < `1 < `2 < `3 <
such that the sequence ì D1
P
T hi .x/ converges to T h.x/ as tends to infinity for
n
P` P1 n
all x 2 R n E. Since j iD1 T hi .x/j iD1 jT hi .x/j for all x 2 R , this
proves (7.60).
Now put
1
[
A WD Pi :
i D1
Then it follows from (7.59), (7.60), and Theorem 1.38 that
Z Z 1
X
jT hj d m jT hi j d m
Rn nA Rn nA iD1
1 Z
X
D jT hi j d m
i D1 Rn nA
1 Z
X
jT hi j d m
i D1 Rn nPi
1
X
c2 khi k1
i D1
D c2 khk1 :
Moreover,
1
X 1
X
m.A/ m.Pi / D c3 m.Qi / D c3 m.B/;
i D1 iD1
where
m.B2pn /
c3 D c3 .n/ WD D m.Bpn / D !n nn=2 1
:
m.Œ 1; 1n /
Hence
tT h .t/ tm.A/ C tm.¹x 2 Rn n A j jT h.x/j > tº/
Z
tm.A/ C jT hj d m
Rn nA
c3 tm.B/ C c2 khk1
c4 .tm.B/ C khk1 /
for all t > 0, where
c4 D c4 .n/ WD max¹c2 .n/; c3 .n/º max¹2n 1 3=2

n .n C 3/; !n nn=2 1
º:
This proves Step 1.

Step 2 (Calderón–Zygmund decomposition). Let f 2 L2 .Rn /\L1 .Rn / and t > 0.
Then there exists a countable collection of closed cubes Qi Rn with pairwise
disjoint interiors such that
1
Z
m.Qi / < jf j 2n m.Qi / for all i 2 N (7.61)
t Qi
and
jf .x/j t for almost all x 2 Rn n B; (7.62)
S1
where B WD iD1 Qi .
For 2 Zn and ` 2 Z define
Q.; `/ WD ¹x 2 Rn j 2 ` i xi 2 ` .i C 1/º:
Let
Q WD ¹Q.; `/ j 2 Zn ; ` 2 Zº
and define the subset Q0 Q by
ˇ tm.Q/ < Q jf j d m and, for all Q0 2 Q;
² ˇ R ³
Q0 WD Q 2 Q ˇˇ :
Q ¨ Q0 H) Q0 jf j d m tm.Q0 /
R
Then every decreasing sequence of cubes in Q contains at most one element of Q0 .

Hence every element of Q0 satisfies (7.61) and any two cubes in Q0 have disjoint
S
interiors. Define B WD Q2Q0 Q. We prove that
1
Z
x 2 Rn n B; x 2 Q 2 Q H) jf j d m t: (7.63)
m.Q/ Q
Suppose, by contradiction, that there existsRan element x 2 Rn n B and a cube
Q 2 Q such that x 2 Q and tm.Q/ < Q jf j d m. Then, since kf k1 < 1,
there is a maximal cube Q 2 Q such that x 2 Q and tm.Q/ < Q jf j d m. Such
R
a maximal cube would be an element of Q0 and hence x 2 B, a contradiction.

This proves (7.63). Now Theorem 6.14 asserts that there exists a Lebesgue null set
E Rn n B such that every element of Rn n .B [ E/ is a Lebesgue point of f .
By (7.63), every point x 2 Rn n .B [ E/ is the intersection point of a decreasing
sequence of cubes over which jf j has mean value at most t. Hence it follows from
Theorem 6.16 that jf .x/j t for all x 2 Rn n .B [ E/. This proves Step 2.
Step 3. Let c D c.n/ 1 be the constant in Step 1. Then

kTf k1;1 .2nC1 C 6c/ kf k1 (7.64)
for all f 2 L2 .Rn / \ L1 .Rn /.
Fix a function f 2 L2 .Rn / \ L1 .Rn / and a constant t > 0. Let the Qi be as in

Step 2 and define [
B WD Qi :
i
1
R
Then m.Qi / < t Qi jf
j d m for all i by Step 2, and hence
1X 1
X Z
m.B/ D m.Qi / jf j d m kf k1 :
t Qi t
i i
Define
g; hW Rn ! R
by Z
f dm
Qi
X
g WD f Rn nB C Q ; h WD f g:
m.Qi / i
i
Then
kgk1 kf k1 ; khk1 2 kf k1 :
Further, h vanishes on Rn n B and Qi h d m D 0 for all i. Hence it follows from
R
Step 1 that
1 3c
T h .t/ c m.B/ C khk1 kf k1 : (7.65)
t t
Moreover, it follows from Step 2 that jg.x/j t for almost every x 2 Rn n B and
jg.x/j 2n t for every x 2 int.Qi /. Thus jgj 2n t almost everywhere. Hence it
follows from Lemma 7.36 that
1 2n 2n
Z Z
2
T g .t/ 2 jgj d m jgj d m 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
2nC1 C 6c
Tf .2t/ T g .t/ C T h .t/ kf k1 :
2t
Here the splitting f D g C h depends on t , but the constant c does not. Multiply
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 (Calderón–Zygmund). Fix both an integer n 2 and a number

1 0 such that
k@i .Kj f /kp ckf kp (7.67)
for all f 2 C01 .Rn / and all i; j D 1; : : : ; n.
Proof. For p D 2 this estimate was established in Lemma 7.44 with c D 1.

Second, suppose 1 < p < 2 and let c1 .n/ be the constant of Lemma 7.45.
For i; j D 1; : : : ; n denote by
Tij W L2 .Rn / ! L2 .Rn /
the unique bounded linear operator that satisfies
Tij f D @i .Kj f / for f 2 C01 .Rn /.
Then kTij f k1;1 c1 .n/ kf k1 for all f 2 C01 .Rn / and all i; j by Lemma 7.45.
Since C01 .Rn / is dense in L2 .Rn / \ L1 .Rn / by Theorem 7.35, it follows that
kTij f k1;1 c1 .n/ kf k1 for all f 2 L2 .Rn / \ L1 .Rn /. Hence Theorem 7.37
(with q D 2) asserts that (7.67) holds with
p 1=p
c D c.n; p/ WD 2 c1 .n/2=p 1
:
.2 p/.p 1/
Third, suppose 2 < p < 1 and choose 1 < q < 2 such that 1=p C 1=q D 1. Then
it follows from Theorem 7.41, integration by parts, Hölder’s inequality, and from
what we have just proved that, for all f; g 2 C01 .Rn /,
Z Z
.@i .Kj f //g d m D .@i @j f /g d m
Rn Rn
Z
D f .@i @j g/ d m
Rn
Z
D f .@i .Kj g// d m
Rn
kf kp k@i .Kj g/kq

c.n; q/ kf kp kgkq :
Since C01 .Rn / is dense in Lq .Rn / by Theorem 7.35, and the Lebesgue measure is
semi-finite, it follows from Lemma 4.34 that k@i .Kj f /kp c.n; q/ kf kp for all
f 2 C01 .Rn /. This proves Theorem 7.46.
Proof of Theorem 7.43. Fix both an integer n 2 and a number 1 < p < 1.
Let c D c.n; p/ be the constant of Theorem 7.46 and let u 2 C01 .Rn /. Then
@j u D @j .K u/ D Kj u
by Theorem 7.41. Hence it follow from Theorem 7.46 with f D u that
k@i @j ujp D k@i .Kj u/jp

c.n; p/ kukp
for i; j D 1; : : : ; n. This proves Theorem 7.43.
7.8 Exercises
Exercise 7.47 (Lebesgue measure on the sphere). For n 2 N, let .Rn ; An ; mn / the
Lebesgue measure space, denote the open unit ball by B n WD ¹x 2 Rn j jnj < 1º,
and the unit sphere by
Sn 1
WD @B n D ¹x 2 Rn j jxj D 1º:
For A S n 1
define
p
A˙ WD ¹x 2 B n 1
j .x; ˙ 1 jxj2 / 2 Aº:
Prove that the collection
AS WD ¹A S n 1
j AC ; A 2 An 1º
is a -algebra and that the map
W AS ! Œ0; 1
defined by
1 1
Z Z
.A/ WD p d mn 1 .x/ C p d mn 1 .x/
AC 1 jxj2 A 1 jxj2
for A 2 AS is a measure. Prove Fubini’s Theorem in polar coordinates as stated

below. Use it to prove that
!n WD .S n 1
/ < 1:
7.8. Exercises 305
Fubini’s Theorem for polar coo rdinates. Let f W Rn ! R be

Lebesgue integrable. For r 0 and x 2 S n 1 define
f r .x/ WD fx .r/ WD f .rx/:
Then there is a set E 2 AS such that
.E/ D 0
and
fx 2 L1 .Œ0; 1// for all x 2 S n 1
n E,
and there is a Lebesgue null set F Œ0; 1/ such that
f r 2 L1 ./ for all r 2 Œ0; 1/ n F .
Define
gW S n 1
!R and hW Œ0; 1/ ! R
by
8Z
rn 1
fx .r/ d m1 .r/ for x 2 S n 1
n E,
ˆ
<
g.x/ WD Œ0;1/ (7.68a)
ˆ
:0 for x 2 E,
8 Z
<r n 1
f r d for r 2 Œ0; 1/ n F ,
h.r/ WD Sn 1 (7.68b)
0 for r 2 F .
:
Then
g 2 L1 ./; h 2 L1 .Œ0; 1//;
and Z Z Z
f d mn D g d D h d m1 : (7.69)
Rn Sn 1 Œ0;1/
Hint. Define the diffeomorphism

W B n 1
.0; 1/ ! ¹x 2 Rn j xn > 0º
by
.x; r/ WD .rx; r.1 jxj2 /1=2 /:
Prove that
det.d.x; r// D .1 jxj2 / 1=2 n 1
r for x 2 B n 1
and r > 0.
Use Theorem 2.17 and Fubini’s Theorem (Theorem 7.30).
Exercise 7.48 (divergence theorem). Let f W Rn ! R be a smooth function. Prove

that Z Z
@i f d mn D xi f .x/ d.x/: (7.70)
Bn Sn 1
Hint. Assume first that i D n. Use the fundamental theorem of calculus and Fubini’s
Theorem (Theorem 7.30) for Rn D Rn 1 R.
Exercise 7.49. Prove that

Z 1 Z 1="
sin.x/ sin.x/
dx WD lim dx D :
0 x "!0 " x 2
Hint. Use the identity

Z 1
rt
e dt D 1=r for r > 0
0
and Fubini’s Theorem.
Exercise 7.50. Define the function
f W R2 ! R
by
8
ˆ
ˆ 1 if z > 0;
sign.xy/ <
f .x; y/ WD 2 ; sign.z/ WD 0 if z D 0; for .x; y/ ¤ 0,
x C y2 ˆ
ˆ
1 if z < 0;
:
and by
f .0; 0/ WD 0:
Prove that fx ; f y W R ! R
R are Lebesgue integrable for allR x; y 2 R. Prove that the
functions R ! RW x 7! R fx d m1 and R ! RW y 7! R f y d m1 are Lebesgue
integrable and
Z Z Z Z
f .x; y/ d m1 .x/ d m1 .y/ D f .x; y/ d m1 .y/ d m1 .x/:
R R R R
Prove that f is not Lebesgue integrable.

7.8. Exercises 307
Exercise 7.51. Let .X; A; / and .Y; B; / be two -finite measure spaces and let
f 2 L1 ./ and g 2 L1 ./. Define
hW X Y !R
by
h.x; y/ WD f .x/g.y/; for x 2 X and y 2 Y:
Prove that h 2 L1 . ˝ / and
Z Z Z
h d. ˝ / D f d g d:
X Y X Y
Exercise 7.52. Let .X; A; / and .Y; B; / be two -finite measure spaces and let
W A ˝ B ! R be any measure such that .A B/ D .A/.B/ for all A 2 A
and all B 2 B. Prove that D ˝ .
Exercise 7.53. Define
W R ! R
by
´
1 cos.x/ for 0 x 2;
.x/ WD
0 otherwise:
Define the functions
f; g; hW R ! R
by Z x
0
f .x/ WD 1; g.x/ WD .x/; h.x/ WD .t/ dt
1
for x 2 R. Prove that .f g/ h D 0 and f .g h/ > 0. Thus the convolution
need not be associative on nonintegrable functions. Compare this with part (v) of
Theorem 7.32. Prove that
E.jf j jgj; jhj/ D E.jf j; jgj jhj/ D R
while
E.f g; h/ D E.f; g h/ D ;:
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
W B ! Œ0; 1/ with the norm kk WD jj.R/. (See Exercise 5.34.) The convolu-
tion of two signed measures ; 2 M is the map
W B ! R
defined by
. /.B/ WD . ˝ /.¹.x; y/ 2 R2 j x C y 2 Bº/ (7.71)
for B 2 B, where
. ˝ / WD C ˝ C C ˝ C ˝ ˝ C:
(See Definition 5.13 and Theorem 5.20.)

(i) If ; 2 M, then 2 M and
k k kk kk :
(ii) There exists a unique element ı 2 M such that
ıD
for all 2 M.
(iii) The convolution product on M is commutative, associative, and distributive.
Thus M is a commutative Banach algebra with unit.
(iv) If f 2 L1 .R/ and
f W B ! R
is defined by Z
f .B/ WD f d m for B 2 B;
B
then f 2 M and kf k D kf k1 .
(v) If f; g 2 L1 .R/, then
f g D f g :
(vi) Let ; ; 2 M. Then D if and only if

Z Z
f d D f .x C y/d. ˝ /.x; y/
R R2
for all bounded Borel measurable functions f W R ! R.

(vii) If ; 2 M and B 2 B, then
Z
. /.B/ D .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 nor-
malized 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 ! GW .x; y/ 7 ! xy (8.1)
the unit 1 2 G, and the inverse map
1
G ! GW x 7 ! x : (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 continuous.
Here the continuity of the group multiplication (8.1) is understood with respect to
the product topology on G G (see Appendix B). A locally compact Hausdorff
group is a topological group .G; U/ such that the topology is locally compact and
Hausdorff (see page 95).
Example 8.1. Let G be any group and define U WD ¹;; Gº. Then .G; U/ is a compact
topological group, but is not Hausdorff unless G D ¹1º.
Example 8.2 (discrete groups). Let G be any group. Then the pair .G; U/ with
the discrete topology U WD 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 multiplicative group SL.n; Z/ of integer
310 8. The Haar measure
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/ n 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 D 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 2 Rnn such that x T x D 1, the special orthogonal group
SO.n/ WD O.n/ \ SL.n; R/, the unitary group U.n/ of matrices x 2 Cnn such
that x x D 1, the special unitary group SU.n/ WD U.n/ \ SL.n; C/, the group
Sp.1/ of the unit quaternions, the unit circle S 1 D U.1/ in the complex plane, the
torus Tn WD S 1 S 1 (n times), or, for any multi-linear form W .Cn /k ! C,
the group of all matrices x 2 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 discon-
nected (every connected component is a single point), but does not have the discrete
topology. It is not locally compact.
Example 8.6 (p-adic integers). Fix a prime number p 2 N and denote by
N0 WD N [ ¹0º
the set of nonnegative integers. For x; y 2 Z define
k
dp .x; y/ WD jx yjp WD inf¹p j k 2 N0 ; x y 2 p k Zº: (8.3)
8.1. Topological groups 311
Then the function

dp W Z Z ! Œ0; 1
is a distance function and so .Z; dp / is a metric space. It is not complete. Its com-
pletion 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
kC1 k k 1 3 2 1
! Z=p k Z ! Z=p k 1
Z ! ! Z=p 2 Z ! Z=p Z ! ¹1º:
The inverse limit of this sequence of maps is the set of sequences
Zp WD ¹x D .xk /k2N0 j xk 2 Z=p k Z; k .xk / D xk 1 for all k 2 Nº:
This set is a commutative ring with unit. Addition and multiplication are defined
term by term, i.e.,
x C y WD .xk C yk /x2N0 ; xy WD .xk yk /x2N0
for x D .xk /k2N0 2 Zp and y D .yk /k2N0 2 Zp . The ring of p-adic integers is a
compact metric space with
k
dp .x; y/ WD inf¹p j k 2 N0 ; xk D yk º: (8.4)
The inclusion of Z into the p-adic integers is given by
p W Z ! Zp ; p .x/ WD .x mod p k /k2N :
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 Hausdorff 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 2 N. Write a nonzero

rational number x 2 Q in the form x D p k a=b where k 2 Z and the numbers
a 2 Z and b 2 N are relatively prime to p, and define jxjp WD p k . For x D 0,
define j0jp WD 0. Define the function
dp W Q Q ! Œ0; 1/
by
dp .x; y/ WD jx yjp
° ˇ a ± (8.5)
WD inf p k ˇ k 2 Z; x y D p k ; a 2 Z; b 2 N n p N :
ˇ
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 2 I , let Gi be a
compact Hausdorff group. Then the product
Y
G WD Gi
i2I
F
is a compact Hausdorff group. Its elements are maps I ! i 2I Gi W i 7! xi
such that xi 2 Gi for all i 2 I . Write such a map as x D .xi /i2I . The product
topology on G is defined as the smallest topology such that the obvious projections
Q
i W G ! Gi are continuous. Thus the (infinite) products U D i2I Ui of open
sets Ui Gi , such that Ui D Gi for all, but finitely many i , form a basis for the
topology of G. (See Appendix B for #I D 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 1 and the product
X X ! XW .x; y/ 7! xy. (See page 282.) Then the group of invertible ele-
ments G WD ¹x 2 X j there exists y 2 X such that xy D yx D 1º is a Hausdorff
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 van-
ishing real-valued continuous functions on a compact topological space. In general
G is not locally compact.
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 2 Cc .G/ and fix a constant " > 0. Then there exists an open
neighborhood U of 1 such that, for all x; y 2 G,
1
x y 2 U H) jf .x/ f .y/j < ": (8.6)
8.2. Haar measures 313
Proof. Choose an open neighborhood U0 G with compact closure and define
K WD ¹ab 1
j a 2 supp.f /; b 2 Ux0 º:
This set is compact because the maps (8.1) and (8.2) are continuous. Also,
1
x … K; x y 2 U0 H) f .x/ D f .y/ D 0 (8.7)
for all x; y 2 G. (If y 2 supp.f / and x 1 y 2 U0 , then x D y.x 1 y/ 1 2 K.)

Since f is continuous there exists, for each x 2 K, an open neighborhood
V .x/ G of x such that
"
y 2 V .x/ H) jf .x/ f .y/j < : (8.8)
2
Since the map G ! GW y 7! x 1 y; is a homeomorphism, the set x 1 V .x/ is an

open neighborhood of 1 for every x 2 K. Since the map (8.1) is continuous it
follows from the definition of the product topology in Appendix B that, for every
x 2 K, there exists an open neighborhood U.x/ G of 1 such that the prod-
uct neighborhood U.x/ U.x/ of the pair .1; 1/ is contained in the pre-image of
x 1 V .x/ under the multiplication map (8.1). In other words,
a; b 2 U.x/ H) xab 2 V .x/: (8.9)
Since the map G ! GW y 7! xy; is a homeomorphism, the set xU.x/ is an open

neighborhood of x for every x 2 K. Since K is compact, there exist finitely many
elements x1 ; : : : ; x` 2 K such that K ìD1 xi U.xi /. Define
S
`
\
U WD U0 \ U.xi /
iD1
and let x; y 2 G such that x 1 y 2 U . If x … K, then f .x/ D f .y/ D 0

by (8.7). Hence assume x 2 K. Then there exists an index i 2 ¹1; : : : ; `º such that
x 2 xi U.xi / and therefore xi 1 x 2 U.xi /. Hence x D xi .xi 1 x/1 2 V .xi / and
y D xi .xi 1 x/.x 1 y/ 2 V .xi / by (8.9). Now, it follows from (8.8) that
jf .x/ f .y/j jf .x/ f .xi /j C jf .xi / f .y/j < ":

For x 2 G define the homeomorphisms
Lx ; Rx W G ! G
by
Lx .a/ WD xa; Rx .a/ WD ax for x 2 G: (8.10)
They satisfy
Lx ı Ly D Lxy ; Rx ı Ry D Ryx ; Lx ı Ry D Ry ı Lx : (8.11)
For A G and x 2 G define

1 1
xA WD ¹xa j a 2 Aº; Ax WD ¹ax j a 2 Aº; A WD ¹a j a 2 Aº:
Thus xA D Lx .A/ and Ax D Rx .A/.
Definition 8.11. A measure W B ! Œ0; 1 is called left-invariant if
.xB/ D .B/ for all B 2 B and x 2 G.

It is called right-invariant if
.Bx/ D .B/ for all B 2 B and x 2 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 the fol-
lowing holds.
(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/ D 1. This measure satisfies .B 1 / D .B/ for all B 2 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 D U.1/ or on S 3 D Sp.1/ in Exer-
cise 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 (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
ƒW Cc .G/ ! R is called left-invariant if
ƒ.f ı Lx / D ƒ.f / (8.12)
for all f 2 Cc .G/ and all x 2 G. It is called right-invariant
ƒ.f ı Rx / D ƒ.f / (8.13)
for all f 2 Cc .G/ and all x 2 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, positive, 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 ƒW Cc .G/ ! R is a left Haar
integral and f 2 Cc .G/ is a nonnegative function that does not vanish identically,
then ƒ.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 sim-
plifications due to Urs Lang [11]. In preparation for the proof it is convenient to
introduce some notation. Let
CcC .G/ WD ¹f 2 Cc .G/ j f 0; f 6 0º (8.14)
be the space of nonnegative continuous functions with compact support that do not
vanish identically. A function ƒW CcC .G/ ! Œ0; 1/ is called
• additive if ƒ.f C g/ D ƒ.f / C ƒ.g/ for all f; g 2 CcC .G/,
• subadditive if ƒ.f C g/ ƒ.f / C ƒ.g/ for all f; g 2 CcC .G/,
• homogeneous if ƒ.cg/ D cƒ.f / for all f 2 CcC .G/ and all c > 0,
• monotone if f g implies ƒ.f / ƒ.g/ for all f; g 2 CcC .G/,
• left-invariant if ƒ.f ı Lx / D ƒ.f / for all f 2 CcC .G/ and all x 2 G.
Every additive functional ƒW CcC .G/ ! Œ0; 1/ is necessarily homogeneous and

monotone. Moreover, every positive linear functional on Cc .G/ restricts to an ad-
ditive functional ƒW CcC .G/ ! Œ0; 1/ and, conversely, every additive functional
ƒW CcC .G/ ! Œ0; 1/ extends uniquely to a positive linear functional on Cc .G/.
This is the content of the next lemma.
Lemma 8.15. Let ƒW CcC .G/ ! Œ0; 1/ be an additive functional. Then there is a
unique positive linear functional on Cc .G/ whose restriction to CcC .G/ agrees with
ƒ. If ƒ is left-invariant, then so is its linear extension to Cc .G/.
Proof. We prove that ƒ is monotone. Let f; g 2 CcC .G/ such that f g.
If f ¤ g, then g f 2 CcC .G/ and hence
ƒ.f / ƒ.f / C ƒ.g f / D ƒ.g/
by additivity. If f D g there is nothing to prove.
We prove that ƒ is homogeneous. Let f 2 CcC .G/. Then ƒ.nf / D nƒ.f / for
all n 2 N by additivity and induction. If c D m=n is a positive rational number,
then ƒ.f / D nƒ.f =n/ and hence ƒ.cf / D mƒ.f =n/ D cƒ.f /. If c > 0 is
irrational, then it follows from monotonicity that
aƒ.f / D ƒ.af / ƒ.cf / ƒ.bf / D bƒ.f /
for all a; b 2 Q with 0 < a < c < b, and this implies ƒ.cf / D cƒ.f /.
Now define
ƒ.0/ WD 0
and, for f 2 Cc .G/,
ƒ.f / WD ƒ.f C / ƒ.f /:
If f; g 2 Cc .G/, then f C C g C C .f C g/ D f C g C .f C g/C , hence
ƒ.f C / C ƒ.g C / C ƒ..f C g/ / D ƒ.f / C ƒ.g / C ƒ..f C g/C /
by additivity, and hence ƒ.f / C ƒ.g/ D ƒ.f C g/. Moreover, . f /C D f and
. f / D f C , and so
ƒ. f / D ƒ.f / ƒ.f C / D ƒ.f /:
Hence it follows from homogeneity that ƒ.cf / D cƒ.f / for all f 2 Cc .G/ and
all c 2 R. This shows that the extended functional is linear.
If the original functional ƒW CcC .G/ ! Œ0; 1/ is left-invariant, then so is
the extended linear functional on Cc .G/ because .f ı Lx /˙ D f ˙ ı Lx for all
f 2 Cc .G/ and all x 2 G. This proves Lemma 8.15.
Consider the space

² ˇ ³
ˇ ƒ is subadditive, monotone,
L WD ƒW Cc .G/ ! .0; 1/ ˇˇ
C
: (8.15)
homogeneous, and left-invariant
The strategy of the proof of Theorem 8.14 is to construct certain functionals
ƒg 2 L associated to functions g 2 CcC .G/ supported near the identity element
and to construct the required positive linear functional ƒW Cc .G/ ! R as a suitable
limit where the functions g converge to the Dirac ı-function at the identity. The
precise definition of the ƒg involves the following construction.
Denote by P the set of all Borel measures W B ! Œ0; 1/ of the form
k
X
WD ˛i ıxi (8.16)
iD1
where k 2 N, ˛1 ; : : : ; ˛k are positive real numbers, x1 ; : : : ; xk 2 G, and ıxi is

the Dirac measure at xi (see Example 1.31). The norm of a measure 2 P of the
form (8.16) is defined by
k
X
kk WD .G/ D ˛i > 0: (8.17)
i D1
If WD j`D1 ˇj ıyj 2 P is any other such measure, define the convolution product
P
of and by
Xk X `
WD ˛i ˇj ıxi yj :
i D1 j D1
This product is not commutative in general. It satisfies k k D kk kk. Asso-

ciated to a measure 2 P of the form (8.16) are two linear operators
L ; R W Cc .G/ ! Cc .G/
defined by
k
X k
X
.L f /.a/ WD ˛i f .xi a/; .R f /.a/ WD ˛i f .axi / (8.18)
i D1 iD1
for f 2 Cc .G/ and a 2 G. The next two lemmas establish some basic properties of
the operators L and R . Denote by
kf k1 WD sup jf .x/j
x2G
the supremum norm of a compactly supported function f W G ! R.

Lemma 8.16. Let ; 2 P, f 2 Cc .G/, and x 2 G. Then
L ı L D L f ı Lx D Lıx f; kL f k1 kk kf k1 ;

R ı R D R ; f ı Rx D Rıx f; kR f k1 kk kf k1 ;
L ı R D R ı L :
Proof. The assertions follow directly from the definitions.
Lemma 8.17. Let f; g 2 CcC .G/. Then there exists a 2 P such that
f L g:
Proof. Fix an element y 2 G such that g.y/ > 0. For x 2 G define

² ˇ ³
ˇ f .x/ C 1 1
Ux WD a 2 G ˇˇ 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 2 G such that supp.f / Ux1 [ [ Uxk . Define
k
X f .xi / C 1
WD ıyx 1 :
g.y/ i
i D1
Then
k
X f .xi / C 1
f .a/ g.yxi 1 a/ D .L g/.a/
g.y/
iD1
for all a 2 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 construction
of a subadditive functional
Mg W CcC .G/ ! .0; 1/
associated to a function g 2 CcC .G/. Step 2 shows that Mg is asymptotically linear

as g concentrates near the unit 1. 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 2 CcC .G/ define
Mg .f / WD M.f I g/ WD inf¹kk j 2 P; f L gº: (8.19)
(i) M.f I g/ > 0 for all f; g 2 CcC .G/.
(ii) For every g 2 CcC .G/, the functional Mg W CcC .G/ ! .0; 1/ is subadditive,
homogeneous, monotone, and left-invariant, and hence is an element of L.
(iii) Let ƒ 2 L. Then
ƒ.f / M.f I g/ƒ.g/ for all f; g 2 CcC .G/: (8.20)
In particular, M.f I h/ M.f I g/M.gI h/ for all f; g; h 2 CcC .G/.
(iv) M.f I f / D 1 for all f 2 CcC .G/.
We prove part (ii). Monotonicity follows directly from the definition. Homogeneity
follows from the identities Lc g D cL g and kck D c kk. To prove left-
invariance, observe that
.L g/ ı Lx D Lıx g; k ıx k D kk
for all 2 P by Lemma 8.16. Since f L g if and only if f ı Lx .L g/ ı Lx ,

this proves left-invariance. To prove subadditivity, fix a constant " > 0 and choose
; 0 2 P such that
" < M.f 0 I g/ C " :
f 0 L0 g;
0
f L g; kk < M.f I g/ C ;
2 2
Then f C f 0 L g C L0 g D LC0 g and hence
M.f C f 0 I g/ C 0 D kk C 0 < M.f I g/ C M.f 0 I g/ C ":

Thus
M.f C f 0 I g/ < M.f I g/ C M.f 0 I g/ C " for all " > 0.
This proves subadditivity and part (ii).

We prove part (iii). Fix a functional ƒ 2 L. We prove first that
ƒ.L f / kk ƒ.f / (8.21)

Pk
for all f 2 CcC .G/ and all 2 P. To see this write D iD1 ˛i ıxi . Then
L f D kiD1 ˛i .f ı Lxi /, and hence
P
k
X
ƒ.L f / ƒ.˛i .f ı Lxi //
i D1
k
X
D ˛i ƒ.f ı Lxi /
i D1
k
X
D ˛i ƒ.f /
i D1
D kk ƒ.f /:
Here the first step follows from subadditivity, the second step follows from homo-
geneity, 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 2 CcC .G/. By Lemma 8.17,
there is a 2 P such that f L g. Since ƒ is monotone this implies
ƒ.f / ƒ.L g/ kk ƒ.g/ by (8.21). Now take the infimum over all 2 P
such that f L g to obtain ƒ.f / M.f I g/ƒ.g/.
We prove parts (i) and (iv). Since the map CcC .G/ ! .0; 1/W f 7! kf k1 ; is
an element of L, it follows from part (iii) that
kf k1 M.f I g/ kgk1 (8.22)
and hence M.f I g/ > 0 for all f; g 2 CcC .G/. Next observe that M.f I f / 1
by (8.22) and M.f I f / 1 because f D Lı1 f . This proves Step 1.
Step 2. Let f; f 0 2 CcC .G/ and let " > 0. Then there is an open neighborhood
U G of 1 such that every g 2 CcC .G/ with supp.g/ U satisfies
Mg .f / C Mg .f 0 / < .1 C "/Mg .f C f 0 /: (8.23)
By Urysohn’s lemma (Theorem A.1), there is a function 2 Cc .G/ such that

.x/ D 1 for all x 2 supp.f / [ supp.f 0 /. Choose a constant 0 < ı 1=2
such that
2ı C 2ıkf C f 0 k1 M.I f C f 0 / < ": (8.24)
Define
h WD f C f 0 C ıkf C f 0 k1 :
Then f = h and f 0 = h extend to continuous functions on G with compact support by

setting them equal to zero on G n supp./. By Lemma 8.10, there exists an open
neighborhood U G of 1 such that
ˇ f .x/ f .y/ ˇ ˇ f 0 .x/ f 0 .y/ ˇ

ˇ ˇ ˇ ˇ
1
x y 2 U H) ˇˇ ˇCˇ ˇ<ı
h.x/ h.y/ ˇ ˇ h.x/ h.y/ ˇ
P`
for all x; y 2 G. Let g 2 CcC .G/ with supp.g/ U . If D iD1 ˛i ıxi 2 P such
that h L g, then for all a 2 supp.f /,
` X f .x 1 / `
L g.a/ X f .a/
i

f .a/ f .a/ D ˛i g.xi a/ ˛i C ı g.xi a/:
h.a/ h.a/
iD1
h.xi 1 / iD1
Thus f L g, where
` f .x 1
X
i/
WD ˛i C ı ıx i :
iD1
h.xi 1 /
This implies
` f .x 1
X /i

Mg .f / ˛i C ı :
i D1
h.xi 1 /
The same inequality holds for f 0 . Since f C f 0 h, we find
` 1
0
X f .x
i / C f 0 .xi 1 /
Mg .f / C Mg .f / ˛i C 2ı kk .1 C 2ı/:
iD1
h.xi 1 /
Now take the infimum over all 2 P such that h L g to obtain
Mg .f / C Mg .f 0 / .1 C 2ı/Mg .h/
.1 C 2ı/.Mg .f C f 0 / C ıkf C f 0 k1 Mg .//
.1 C 2ı C 2ıkf C f 0 k1 M.I f C f 0 //Mg .f C f 0 /
.1 C "/Mg .f C f 0 /:
Here we have used the inequalities 1 C 2ı 2 and (8.24). This proves Step 2.
Step 3. Let f 2 CcC .G/ and let " > 0. Then there is an open neighborhood U G
of 1 with the following significance. For every g 2 CcC .G/ such that
1
supp.g/ U; g.x/ D g.x / for all x 2 G; (8.25)
there exists an open neighborhood W G of 1 such that every h 2 CcC .G/ with
supp.h/ W satisfies the inequality
M.f I g/Mh .g/ .1 C "/Mh .f /: (8.26)
This inequality continues to hold with Mh replaced by any left-invariant positive
linear functional ƒW Cc .G/ ! R.
By Urysohn’s lemma (Theorem A.1) there is a function 2 CcC .G/ such that
.x/ D 1 for all x 2 K WD supp.f /. Choose "0 and "1 such that
1 C "0 "0
0 < "0 < 1; 1 C "; "1 WD : (8.27)
1 "0 2M.I f /
By Lemma 8.10, there exists an open neighborhood U G of 1 such that
1
x y 2 U H) jf .x/ f .y/j < "1 (8.28)
for all x; y 2 G. We prove that the assertion of Step 3 holds with this neighbor-
hood U . Fix a function g 2 CcC .G/ that satisfies (8.25). Put
"1
"2 WD : (8.29)
2M.f I g/
Use Lemma 8.10 to find an open neighborhood V G of 1 such that
1
xy 2 V H) jg.x/ g.y/j < "2 (8.30)
for all x; y 2 G. Then the sets xV for x 2 K form an open cover of K. Hence
S`
there exist finitely many points x1 ; : : : ; x` 2 K such that K i D1 xi V .
By Theorem A.4, there exist functions 1 ; : : : ; ` 2 CcC .G/ such that
`
X
0 i 1; supp.i / xi V; i jK 1: (8.31)
i D1
It follows from Step 2 by induction that there exists an open neighborhood W G

of 1 such that every h 2 CcC .G/ with supp.h/ W satisfies
`
X
Mh .i f / < .1 C "0 / Mh .f /: (8.32)
iD1
We prove that every h 2 CcC .G/ with supp.h/ W satisfies (8.26).

For x 2 G define the function Fx 2 Cc .G/ by

1
Fx .y/ WD f .y/g.y x/ for y 2 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 2 G. Since g.y 1 x/ D g.x 1 y/ D .gıLx 1 /.y/ by (8.25), this implies
f .x/g ı Lx 1 Fx C "1 g ı Lx 1 : Hence, for all x 2 G and all h 2 CcC .G/,
f .x/Mh .g/ Mh .Fx / C "1 Mh .g/ (8.34)
Now fix a function h 2 CcC .G/ with supp.h/ W . By (8.30) and (8.31),
1
i .y/Fx .y/ D i .y/f .y/g.y x/ i .y/f .y/.g.xi 1 x/ C "2 /
P
for all x; y 2 G and all i D 1; : : : ; `. Since Fx D i i Fx , this implies
X
Mh .Fx / Mh .i Fx /
i
X
Mh .i f /.g.xi 1 x/ C "2 / (8.35)
i
X
Mh .i f /g.xi 1 x/ C 2"2 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.xi 1 x/ C 2"2 Mh .f / C "1 Mh .g/
i
X
Mh .i f /g.xi 1 x/ C 2"1 Mh .g/:
i
Here the second step uses (8.29) and the inequality
Mh .f / M.f I g/Mh .g/:
Thus
.f 2"1 /C Mh .g/ L g;
where X
WD Mh .i f /ıx 1 :
i
i
This implies
X
Mg ..f 2"1 /C /Mh .g/ Mh .i f / .1 C "0 /Mh .f /:
i
Here the second step uses (8.32). Since f .f 2"1 /C C 2"1 , we have
Mg .f /Mh .g/ Mg ..f 2"1 /C /Mh .g/ C 2"1 Mg ./Mh .g/

.1 C "0 /Mh .f / C 2"1 M.I f /Mg .f /Mh .g/
D .1 C "0 /Mh .f / C "0 Mg .f /Mh .g/:
Hence
1 C "0
Mg .f /Mh .g/ Mh .f / .1 C "/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 ƒW Cc .G/ ! R.
Step 4. We prove uniqueness.
Let ƒ; ƒ0 W Cc .R/ ! R be two left-invariant positive linear functionals that do not
vanish identically. Then there exists a function f 2 CcC .G/ such that ƒ.f / > 0 by
Lemma 8.15. Hence
ƒ.g/ M.f I g/ 1 ƒ.f / > 0
for all g 2 CcC .G/ by (8.20). The same argument shows that ƒ0 .g/ > 0 for all
g 2 CcC .G/.
Now fix two functions f; f0 2 CcC .G/ and a constant " > 0. Choose an open
neighborhood U G of 1 that satisfies the requirements of Step 3 for both f and
f0 and this constant ". By Urysohn’s lemma (Theorem A.1), there exists a function
g 2 CcC .G/ such that
1
g.1/ > 0; supp.g/ ¹x 2 G j x 2 U and x 2 U º:
Replacing g by the function x 7! g.x/ C g.x 1 /, if necessary, we may assume that

g satisfies (8.25). Hence it follows from Step 1 and Step 3 that
ƒ.f / M.f I g/ƒ.g/ .1 C "/ƒ.f /
and
.1 C "/ƒ.f0 / M.f0 I g/ƒ.g/ ƒ.f0 /:
Take the quotient of these inequalities to obtain
1 ƒ.f / M.f I g/ ƒ.f /

.1 C "/ .1 C "/ :
ƒ.f0 / M.f0 I g/ ƒ.f0 /
The same inequality holds with ƒ replaced by ƒ0 . Hence
2 ƒ.f / ƒ0 .f / ƒ.f /
.1 C "/ 0 .1 C "/2 :
ƒ.f0 / ƒ .f0 / ƒ.f0 /
Since this holds for all " > 0, it follows that

ƒ0 .f0 /
ƒ0 .f / D cƒ.f /; c WD :
ƒ.f0 /
Since this holds for all f 2 CcC .G/, ƒ0 and cƒ agree on CcC .G/. Hence ƒ0 D cƒ
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 2 CcC .G/ and, for g 2 CcC .G/, define
ƒg W CcC .G/ ! .0; 1/
by
M.f I g/
ƒg .f / WD for f 2 CcC .G/: (8.36)
M.f0 I g/
Then ƒg 2 L for all g 2 CcC .G/ by Step 1. It follows also from Step 1 that
M.f0 I g/ M.f0 I f /M.f I g/ and M.f I g/ M.f I f0 /M.f0 I g/ and hence
1
M.f0 I f / ƒg .f / M.f I f0 / (8.37)
for all f; g 2 CcC .G/. Fix a function f 2 CcC .G/ and a number " > 0. Define
8 ˇ 9
< ˇ ƒ.f0 / D 1 and there exists a neighborhood =
L" .f / WD ƒ 2 L ˇˇ W G of 1 such that for all h 2 CcC .G/
ˇ
:
: ˇ supp.h/ W H) ƒ.f / .1 C "/ƒh .f / ;
We prove that L" .f / ¤ ;. To see this let U G be the open neighborhood

of 1 constructed in Step 3 for f and ". Choose a function g 2 CcC .G/ that sat-
isfies (8.25) and choose an open neighborhood W G of 1 associated to g that
satisfies the requirements of Step 3. Let h 2 CcC .G/ with supp.h/ W . Then
M.f I g/M.gI h/ .1 C "/M.f I h/ and M.f0 I g/M.gI h/ M.f0 I h/ by Step 3
and Step 1. Take the quotient of these inequalities to obtain ƒg .f / .1C"/ƒh .f /.
Since ƒg .f0 / D 1, it follows that ƒg 2 L" .f /. This shows that L" .f / ¤ ;,
as claimed. Next we observe that
ƒ.f / M.f I f0 /ƒ.f0 / D M.f I f0 /
for all ƒ 2 L" .f / by Step 1. Hence the supremum
ƒ" .f / WD sup¹ƒ.f / j ƒ 2 L" .f /º (8.38)

is a real number, bounded above by M.f I f0 /. Since L" .f / contains an element of

the form ƒg for some g 2 CcC .G/, it follows from (8.37) that
1
M.f0 I f / ƒ" .f / M.f I f0 / (8.39)
for all f 2 CcC .G/ and all " > 0. Moreover, the function " 7! ƒ" .f / is nonde-
creasing by definition. Hence the limit
ƒ0 .f / WD lim ƒ" .f / D inf ƒ" .f / (8.40)
"!0 ">0
exists and is a positive real number for every f 2 CcC .G/.
We prove that the functional ƒ0 W CcC .G/ ! .0; 1/ is left-invariant. To see this,
fix a function f 2 CcC .G/ and an element x 2 G. Then
L" .f / D L" .f ı Lx /
for all " > 0. Namely, if W G is an open neighborhood of 1 such that
ƒ.f / .1 C "/ƒh .f / for all h 2 CcC .G/ with supp.h/ W , then the same in-
equality holds with f replaced by f ı Lx because both ƒ and ƒh are left-invariant.
Hence ƒ" .f / D ƒ" .f ı Lx / for all " > 0, and so ƒ0 .f / D ƒ0 .f ı Lx /.
We prove that the functional ƒ0 W CcC .G/ ! .0; 1/ is additive. To see this,
fix two functions f; f 0 2 CcC .G/. We show that
1
.1 C "/ ƒ" .f C f 0 / ƒ" .f / C ƒ" .f 0 / .1 C "/ƒ" .f C f 0 / (8.41)
for all " > 0. To establish the first inequality in (8.41), choose any functional
ƒ 2 L" .f C f 0 /. Then there exists an open neighborhood W G of 1 such that
ƒ.f Cf 0 / .1C"/ƒh .f Cf 0 / for all h 2 CcC .G/ with supp.h/ W . Moreover,
we have seen above that h 2 CcC .G/ can be chosen such that supp.h/ W and
also ƒh 2 L" .f / \ L" .f 0 /. Any such h satisfies
1
.1 C "/ ƒ.f C f 0 / ƒh .f C f 0 / ƒh .f / C h .f 0 / ƒ" .f / C ƒ" .f 0 /:
Take the supremum over all ƒ 2 L" .f C 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 ƒ 2 L" .f / and ƒ0 2 L" .f 0 /. Then there exists an open neighborhood
W G of 1 such that ƒ.f / .1 C "/ƒh .f / and ƒ0 .f 0 / .1 C "/ƒh .f 0 / for
all h 2 CcC .G/ with supp.h/ W . By Step 2, the function h 2 CcC .G/ can be
chosen such that supp.h/ W and also ƒh .f / C ƒh .f 0 / ˛ƒh .f C f 0 / and
ƒh 2 L" .f C f 0 /. Any such h satisfies
1
.1 C "/ .ƒ.f / C ƒ0 .f 0 // ƒh .f / C ƒh .f 0 /
˛ƒh .f C f 0 /
˛ƒ" .f C f 0 /:
Take the supremum over all pairs of functionals ƒ 2 L" .f / and ƒ0 2 L" .f 0 /
to obtain .1 "/ 1 .ƒ" .f / C ƒ" .f 0 // ˛ƒ" .f C f 0 / for all ˛ > 1: This proves
the second inequality in (8.41). Take the limit " ! 0 in (8.41) to obtain that ƒ0 is
additive. Moreover, it follows directly from the definition that ƒ0 .f0 / D 1. Hence
by Lemma 8.15, ƒ0 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 WD ¹g 2 CcC .G/ j 0 g 1; g.1/ D 1º
is a directed set equipped with a map g 7! ƒg that takes values in the space
ˇ M.f0 I f / 1 ƒ.f / M.f0 I f /

² ˇ ³
L WD ƒW Cc .G/ ! R ˇ
C ˇ :
for all f 2 CcC .G/
The map G ! LW g 7! ƒg 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! ƒg converges and so there is no
need to choose a universal subnet. That this can be proved with a refinement of
the uniqueness argument (ƒ in Step 3) is pointed out in Pedersen [16]. That paper
also contains two further uniqueness proofs. One is based on Fubini’s Theorem and
the other on the Radon–Nikodým theorem. Another existence proof for compact
second countable Hausdorff groups is due to Pontryagin. It uses the Arzelà–Ascoli
theorem to establish the existence of a sequence i 2 P with ki k D 1 such that
Li f converges to a constant function whose value is then taken to be ƒ.f /.
Proof of Theorem 8.12. Existence and uniqueness in (i) follow directly from
Theorem 8.14 and the Riesz representation theorem (Theorem 3.15). That nonempty
open sets have positive measure follows from Urysohn’s lemma (Theorem A.1).
To prove (ii), consider the map
W G ! G
defined by
1
.x/ WD x for x 2 G.
Since is a homeomorphism, it preserves the Borel -algebra B. Since
ı Rx D Lx 1 ı ;
a measure W B ! Œ0; 1 is a left Haar measure if and only if the measure
W B ! Œ0; 1
defined by
1
.B/ WD ..B// D .B /
is a right Haar measure. Hence assertion (ii) follows from (i).

We prove (iii). Assume G is compact and let W B ! Œ0; 1 be the unique left
Haar measure such that .G/ D 1. For x 2 G define
x W B ! Œ0; 1
by
x .B/ WD .Rx .B// for B 2 B.
Since Rx commutes with Ly for all y by (8.11), x is a left Haar measure. Since
x .G/ D .Rx .G// D .G/ D 1, it follows that x D for all x 2 G. Hence
is right-invariant. Therefore, the map
B ! Œ0; 1W B 7 ! .B/ WD ..B// D .B 1

/
is a left Haar measure and, since .G/ D 1, it agrees with . This proves Theo-
rem 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 W G ! .0; 1/ such that
.Rx .B// D .x/.B/ (8.42)
for all x 2 G and all B 2 B. The map W G ! .0; 1/ 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 if its modular character is
trivial or, equivalently, if 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 2 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 W B ! Œ0; 1 be a right Haar measure. Show that the modular
character is characterized by the condition .Lx 1 .B// D .x/.B/ for all x 2 G
and all B 2 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 ! XW .g; x/ 7 ! g x (8.43)
one can use the Haar measure to produce G-invariant continuous functions by aver-
aging. Namely, if f W X ! R is any continuous function, and is the Haar measure
on G with .G/ D 1, then the function F W X ! R defined by
Z
F .x/ WD f .a x/ d.a/ (8.44)
G
for x 2 X is G-invariant in that
F .g x/ D F .x/
for all x 2 X and all g 2 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 W G ! GL.V / be a homomorphism from a compact Hausdorff

group to the general linear group of automorphisms of a finite-dimensional vector
space. (Such a homomorphism is called a representation of G.) Prove that V ad-
mits 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 282.) Find conditions under which f g D g f .
Show that the convolution is not commutative in general. Hint. See Section 7.5 for
G D Rn . See also Step 3 in the proof of Theorem 8.14.
Appendices
Appendix A
Urysohn’s lemma
Theorem A.1 (Urysohn’s 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 W X ! Œ0; 1
such that
f jK 1; supp.f / D ¹x 2 X j f .x/ ¤ 0º U: (A.1)
Lemma A.2. Let X be a topological space and let K X be compact. Then the
following holds:
(i) every closed subset of K is compact;
(ii) if X is Hausdorff, then for every y 2 X n K, there exist open sets U; V X

such that K U , y 2 V , and U \ V D ;;
(iii) if X is Hausdorff, then K is closed.
Proof. (i) Let F K be closed and let ¹Ui ºi2I be an open cover of F . Then the
sets ¹Ui ºi 2I together with V WD X n F form an open cover of K. Therefore, 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.
(ii) Assume X is Hausdorff and let y 2 X n K. Define
U WD ¹U X j U is open and y … Ux º:
334 A. Urysohn’s lemma
Since X is Hausdorff, the collection U is an open cover of K. Since K is compact,

there exists finitely many set U1 ; : : : ; Un 2 U such that
K U WD U1 [ [ Un :
Since y … Uxi for all i , it follows that y 2 V WD X n Ux and U \ V D ;. Hence the

sets U and V satisfy the requirements of (ii).
(iii) Assume X is Hausdorff. Then it follows from part (ii) that, for every y 2
X n K, there exists an open set V X such that y 2 V and V \ K D ;. Hence
X n K is the union of all open sets in X that are disjoint from K. Thus X n 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 Vx is compact and
K V Vx U: (A.2)
Proof. We first prove the assertion in the case where K D ¹xº consist of a single
element. Choose a compact neighborhood B X of x. Then F WD B n 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 2 W , F W 0 and W \ W 0 D ;. Hence V WD int.B/ \ W is an open
neighborhood of x, its closure is a closed subset of B and hence compact, and
Vx B \ W
S B n W 0 B n F U:
This proves the lemma in the case #K D 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 sets V1 ; : : : ; Vn such that Vxi is a compact
subset of U for all i and K niD1 Vi . Hence V WD niD1 Vi is an open set
S S
Sn
containing K and its closure Vx D i D1 Vxi is a compact subset of U . This proves
Lemma A.3.
Proof of Theorem A.1. The proof has three steps. The first step requires the Axiom
of Dependent Choice.1
1
See http://en.wikipedia.org/wiki/Axiom_of_dependent_choice.
335
Step 1. There exists a family of open sets Vr X with compact closure, para-
metrized by r 2 Q \ Œ0; 1, such that
K V1 Vx1 V0 Vx0 U (A.3)
and
s > r H) Vxs Vr (A.4)
for all r; s 2 Q \ Œ0; 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 \ Œ0; 1W i 7! qi such that
q0 D 0 and q1 D 1. Suppose by induction that the open sets Vi D Vqi have been
constructed for i D 0; 1; : : : ; n such that (A.4) holds for r; s 2 ¹q0 ; q1 ; : : : ; qn º.
Choose k; ` 2 ¹0; 1; : : : ; nº such that
qk WD max¹qi j 0 i n; qi < qnC1 º;
q` WD min ¹qi j 0 i n; qi > qnC1 º:

Then Vx` Vk . Hence it follows from Lemma A.3 that there exists an open set
VnC1 D VqnC1 X with compact closure such that Vx` VnC1 VxnC1 Vk .
This completes the induction argument and Step 1 then follows from the Axiom of
Dependent Choice. (Denote by V the collection of all open sets V X such that
K V Vx U . Denote by V the set of all finite sequences v D .V0 ; : : : ; Vn /
in V that satisfy (A.3) and qi < qj H) Vxj Vi for all i; j . Define a relation
on V by v D .V1 ; : : : ; Vn / v0 D .V10 ; : : : ; Vn00 / if n < n0 and Vi D Vi0 for
i D 0; : : : ; n. Then V is nonempty and for every v 2 V there is a v0 2 V such
that v v0 . Hence, by the Axiom of Dependent Choice, there exists a sequence
vj D .Vj;0 ; : : : ; Vj;nj / 2 V such that vj vj C1 for all j 2 N. Define the map
Q \ Œ0; 1 ! VW q 7! Vq by Vqi WD Vj;i for i; j 2 N with nj i . This map is well
defined and satisfies (A.3) and (A.4) by the definition of V and .)
Step 2. Let Vr X be as in Step 1 for r 2 Q \ Œ0; 1. Then
f .x/ WD sup¹r 2 Q \ Œ0; 1 j x 2 Vr º
(A.5)
D inf ¹s 2 Q \ Œ0; 1 j x … Vxs º
for all x 2 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 2 X and define
a WD sup¹r 2 Q \ Œ0; 1 j x 2 Vr º;
b WD inf¹s 2 Q \ Œ0; 1 j x … Vxs º:

336 A. Urysohn’s lemma
We prove that a b. If b D 1 this follows directly from the definitions. Hence

assume b < 1 and choose an element s 2 Q \ Œ0; 1 such that
x … Vxs :
If r 2 Q \ Œ0; 1 is such that x 2 Vr , then Vr n Vxs ¤ ;, hence Vxs Vr , and hence
r s. Thus we have proved that
x 2 Vr H) r s
for all r 2 Q \ Œ0; 1. Take the supremum over all r 2 Q \ Œ0; 1 with x 2 Vr
to obtain a s. Then take the infimum over all s 2 Q \ Œ0; 1 with x … Vxs to
obtain a b. Now suppose, by contradiction, that a < b. Choose rational numbers
r; s 2 Q \ Œ0; 1 such that a < r < s < b. Since a < r, it follows that x … Vr , since
s < b, it follows that x 2 Vxs , and since r < s, it follows from Step 1 that Vxs Vr .
This is a contradiction and shows that our assumption that a < b was wrong. Thus
a D b and this proves Step 2.
Step 3. The function f W X ! Œ0; 1 in Step 2 is continuous and
´
0 for x 2 X n V0 ;
f .x/ D (A.6)
1 for x 2 Vx1 :
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 2 R. Then f .x/ < c if and only if
there exists a rational number s 2 Q \ Œ0; 1 such that s < c and x … Vxs . Likewise,
f .x/ > c if and only if there exists a rational number r 2 Q \ Œ0; 1 such that r > c
and x 2 Vr . Hence,
[ [
f 1 ..c; 1// D Vr ; f 1 .. 1; c// D .X n Vxs /:
r2Q\.c;1 s2Q\Œ0;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.
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 W X ! R with compact support such that
n
X
fi 0; fi 1; supp.fi / Ui
i D1
Pn
for all i and iD1 fi .x/ D 1 for all x 2 K.
337
Proof. Define the set

ˇ V is open; Vx is compact; and there exists
² ˇ ³
V WD V X ˇ an index i 2 ¹1; : : : ; nº such that Vx Ui :
ˇ
If x 2 K, then x 2 Ui for some index i 2 ¹1; : : : ; nº and, by Lemma A.3, there is an

open set V X such that Vx is compact and x 2 V Vx Ui . Thus V is an open
cover of K. Since K is compact there exist finitely many open sets V1 ; : : : ; V` 2 V
such that K V1 [ [ V` : For i D 1; : : : ; n define
[
Ki WD Vxj :
1j `; Vxj Ui
Then K K1 [ [ Kn and Ki is a compact subset of Ui for each i. Hence

it follows from Urysohn’s lemma (Theorem A.1) that, for each i, there exists a
compactly supported continuous function gi W X ! R such that
0 gi 1; supp.gi / Ui ; gi jKi 1:
Define
f1 WD g1 ;
f2 WD .1 g1 /g2 ;
f3 WD .1 g1 /.1 g2 /g3 ;
::
:
fn WD .1 g1 / .1 gn 1 /gn :
Then supp.fi / supp.gi / Ui for each i and

n
X n
Y
1 fi D .1 gi /:
i D1 iD1
Since K K1 [ [ Kn and the factor 1 gi vanishes on Ki , this implies

Pn
iD1 fi .x/ D 1 for all x 2 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 WD ¹.x; y/ j x 2 X; y 2 Y º;
and let X W X Y ! X and Y W X Y ! Y be the projections onto the first

and second factor. Consider the following universality property for a topology
U 2X Y on the product space.
(F) Let .Z; UZ / be any topological space and let hW Z ! X Y be any map.
Then hW .Z; UZ / ! .X Y; U/ is continuous if and only if the maps
f WD X ı hW .Z; UZ / ! .X; UX /; (B.1a)
g WD Y ı hW .Z; UZ / ! .Y; UY / (B.1b)
are continuous.
Theorem B.1. (i) There is a unique topology U on X Y that satisfies (P).

(ii) Let U 2X Y be as in (i). Then W 2 U if and only if there are open sets
Ui 2 UX and Vi 2 UY , indexed by any set I , such that W D i2I .Ui Vi /.
S
(iii) Let U 2X Y 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 W .Y; UY / ! .X Y; U/; x .y/ WD .x; y/ for y 2 Y;
is continuous for every x 2 X and the inclusion
y W .X; UX / ! .X Y; U/; y .x/ WD .x; y/ for x 2 X;
is continuous for every y 2 Y .

340 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 2X Y :
Proof of Theorem B.1. The proof has five steps.
Step 1. If U 2X Y is a topology satisfying (P), then the projections X and Y

are continuous.
Take h WD idW X Y ! X Y so that f D X ı h D X and g D Y ı h D Y .
Step 2. We prove uniqueness in (i).
Let U; U0 2X Y be two topologies satisfying (P) and consider the map
h WD idW .X Y; U/ ! .X Y; U0 /:
Since f D X W .X Y; U/ ! .X; UX / and g D Y W .X Y; U/ ! .Y; UY / are

continuous by Step 1, and U0 satisfies (P), it follows that h is continuous and hence
U0 U. Interchange the roles of U and U0 to obtain U0 D U.
Step 3. We prove (ii) and existence in (i).
Define U 2X Y as the collection of all sets of the form W D i 2I .Ui Vi /,

S
where I is any index set and Ui 2 UX , Vi 2 UY for i 2 I . Then U is a topology

and the projections X W .X Y; U/ ! .X; UX / and Y W .X Y; U/ ! .Y; UY / are
continuous. We prove that U satisfies (P). To see this, let .Z; UZ / be any topological
space, let hW Z ! X Y be any map, and define f WD X ı h and g WD Y ı h
as in (B.1). If h is continuous, then so are f and g. Conversely, if f and g are
continuous, then h 1 .U V / D f 1 .U / \ g 1 .V / is open in Z for all U 2 UX
and all V 2 UY , and hence it follows from the definition of U that h 1 .W / is open
for all W 2 U. Thus h is continuous.
Step 4. We prove (iii).
Let U be the topology in (i) and let U0 be any topology on X Y with respect
to which X and Y are continuous. If U 2 UX and V 2 UY , then U V D
X 1 .U / \ Y 1 .V / 2 U0 . Hence U U0 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.
341
Step 5. We prove (iv).

Fix an element x 2 X and consider the map h WD x W Y ! X Y . Then the map
f WD X ı hW Y ! X is constant and g WD Y ı hW 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 2 Y .
Appendix C
The inverse function theorem
This appendix contains a proof of the inverse function theorem. The result is for-
mulated 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 inverse function theorem is used on
page 81 in the proof of Lemma 2.19, which is a key step in the proof of the transfor-
mation formula for the Lebesgue measure (Theorem 2.17). Assume throughout that
.X; kk/ is a Banach space. When ˆW X ! X is a bounded linear operator denote
its operator norm by
kˆxk
kˆk WD kˆkL.X/ WD sup :
x2X n¹0º kxk
For x 2 X and r > 0 denote by Br .x/ WD ¹y 2 X j kx yk < rº the open ball of

radius r about x. For x D 0 abbreviate Br WD Br .0/.
Theorem C.1 (inverse function theorem). Fix an element x0 2 X and two real num-
bers r > 0 and 0 < ˛ < 1. Let W Br .x0 / ! X be a continuously differentiable
map such that
kd .x/ 1kL.X/ ˛ for all x 2 Br .x0 /: (C.1)
Then
B.1 ˛/r . .x0 // .Br .x0 // B.1C˛/r . .x0 //: (C.2)
1
Moreover, the map is injective, its image is open, the map is continuously
differentiable, and
1 1 1
d .y/ D d . .y// for all y 2 .Br .x0 //.
344 C. The inverse function theorem
Proof. Assume without loss of generality that x0 D .x0 / D 0.

Step 1. is a homeomorphism onto its image and .Br / B.1C˛/r .
Define
WD id W Br ! X:
Then kd.x/k ˛ for all x 2 Br . Hence
k.x/ .y/k ˛kx yk: (C.3)
for all x; y 2 Br , and so
.1 ˛/ kx yk k .x/ .y/k .1 C ˛/ kx yk : (C.4)
The second inequality in (C.4) shows that .Br / B.1C˛/r and the first inequality
1
in (C.4) shows that is injective and is Lipschitz continuous.
Step 2. B.1 ˛/r .Br /.
Let 2 B.1 ˛/r and define " > 0 by kk DW .1 ˛/.r "/. Then, by (C.3) with
y D 0, we have k.x/k ˛kxk for all x 2 Br . If kxk r " this implies
k.x/ C k r ". Thus the map x 7! .x/ C is a contraction of the closed
ball Bxr " . By the contraction mapping principle, it has a unique fixed point x and
the fixed point satisfies .x/ D x .x/ D . Hence 2 .Br /.
Step 3. .Br / is open.
Let x 2 Br and define y WD .x/. Choose " > 0 such that B" .x/ Br . Then,
by Step 2, B.1 ˛/" . .x// .B" .x// .Br /.
1
Step 4. is continuously differentiable.
Let x0 2 Br and define y0 WD .x0 / and ‰ WD d .x0 /. Then k1 ‰k ˛, so ‰
is invertible, ‰ 1 D 1 ‰/k , and k‰ 1 k .1 ˛/ 1 . We prove that 1
P
kD0 .1
1 1
is differentiable at y0 and d .y0 / D ‰ . Let " > 0. Since is differentiable
at x0 and d .x0 / D ‰, there is a constant ı > 0 such that, for all x 2 Br with
kx x0 k < ı.1 ˛/ 1 , we have
k .x/ .x0 / ‰.x x0 /k ".1 ˛/2 kx x0 k:
Shrinking ı, if necessary, we may assume, by Step 3, that Bı .y0 / .Br /.

1
Now suppose ky y0 k < ı and denote x WD .y/ 2 Br . Then, by (C.4),
1 1
kx x0 k .1 ˛/ ky y0 k < ı.1 ˛/ ;
345
and hence
1 1 1 1
k .y/ .y0 / ‰ .y y0 /k D k‰ .y y0 ‰.x x0 //k
1
k .x/ .x0 / ‰.x x0 /k
1 ˛
".1 ˛/ kx x0 k
"ky y0 k:
1
Hence is differentiable at y0 and d 1 .y0 / D ‰ 1 D d . 1
.y0 // 1
.
1
Thus d is continuous by Step 1. This proves Theorem C.1.
Bibliography
[1] R. B. Ash, Basic probability theory. John Wiley & Sons, New York, 1970.
Reprint, Dover Publications, Mineola, N,Y., 2008. John Wiley & Sons 1970,
Dover Publications, 2008.
MR 0268923 Zbl 1234.60001 Zbl 0206.18001 (reprint)
http://www.math.uiuc.edu/~r-ash/BPT/BPT.pdf
[2] V. I. Bogachev, Measure Theory. Volumes I and II. Springer, Berlin etc., 2006.
MR 2267655 Zbl 1120.28001
[3] W. F. Eberlein, Notes on integration I: The underlying convergence theorem.

Comm. Pure Appl. Math. 10 (1957), 357–360.
MR 0088537 Zbl 0077.31302
[4] D. H. Fremlin, Measure theory. Volumes 1–5. Torres Fremlin, Colchester,

2000–2008.
MR 2462519 (I) MR 2462280 (II) MR 2459668 (III) MR 2462372 (IV)
Zbl 1162.28001 (I) Zbl 1165.28001 (II) Zbl 1165.28002 (III)
Zbl 1166.28001 (IV) Zbl 1166.28002 (V)
https://www.essex.ac.uk/maths/people/fremlin/mt.htm
[5] D. Gilbarg and N. S. Trudinger, Elliptic partial differential equations of

second order. Second edition. Springer, Berlin etc., 1983.
MR 0737190 Zbl 0562.35001
[6] P. R. Halmos, Measure theory. University Series in Higher Mathematics.

D. Van Nostrand, New York, and Macmillan & Co., London, 1950. Reprint,
Graduate Texts in Mathematics, 18. Springer, Berlin etc., 1974.
MR 0033869 Zbl 0040.16802 Zbl 0283.28001 (reprint)
[7] C. E. Heil, Lecture notes on real analysis. Georgia Institute of Technology,

Atlanta, Fall 2007.
http://people.math.gatech.edu/~heil/6337/fall07/section3.4.pdf
http://people.math.gatech.edu/~heil/6337/fall07/section3.5b.pdf
348 Bibliography
[8] H. König, New versions of the Radon–Nikodým theorem. Arch. Math.

(Basel) 86 (2006), no. 3, 251–260.
MR 2215314 Zbl 1091.28002
http://www.math.uni-sb.de/service/preprints/preprint132.pdf
[9] T. W. Körner, A companion to analysis. A second first and first second course
in analysis. Graduate Studies in Mathematics, 62. American Mathematical
Society, Providence, R.I., 2004.
MR 2015825 Zbl 1061.26003
https://www.dpmms.cam.ac.uk/~twk/
[10] U. Lang, Mass & Integral, Vorlesungsmanuskript. ETH, Zürich, September

2006.
http://www.math.ethz.ch/~lang/mi.pdf
[11] U. Lang, Existence and uniqueness of Haar integrals. Preprint. ETH Zürich,
January 2015.
http://www.math.ethz.ch/~lang/HaarIntegral.pdf
[12] D. Ma, The lexicographic order and the double arrow space. Online survey,
2009.
http://dantopology.wordpress.com/2009/10/07/
[13] P. Malliavin, Integration and probability. With the collaboration of H. Airault,

L. Kay, and G. Letac. Edited and translated from the French by L. Kay.
With a foreword by M. Pinsky. Graduate Texts in Mathematics, 157. Springer,
New York, 1995.
MR 1335234 Zbl 0874.28001
[14] J.R. Munkres, Topology. Second edition. Prentice Hall, Upper Saddle River,
N.J., 2000.
Zbl 0951.54001
[15] G. Nagy, Real Alalysis. Version 2.0. Section IV.6, Duals of Lp spaces. Lecture
notes. Kansas State University, Manhattan, KS.
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, Copenhagen,
2000.
http://www.math.ku.dk/kurser/2004-2/mat3re/haarintegral.pdf
Bibliography 349
[17] W. Rudin, Real and complex analysis. Third edition. McGraw-Hill, New York,
1987.
MR 0924157 Zbl 0925.00005
[18] D. A. Salamon, Das Riemannsche Integral. Lecture Notes, ETH, Zürich, 2014.
http://www.math.ethz.ch/~salamon/PREPRINTS/ana1-int.pdf
[19] D. A. Salamon, Mehrfache Integrale. Lecture notes, ETH, Zürich, 2014.

http://www.math.ethz.ch/~salamon/PREPRINTS/ana2-int.pdf
[20] D. A. Salamon, Funktionentheorie. Grundstudium Mathematik. Birkhäuser,

Basel, 2012.
Zbl 1236.30001
http://www.math.ethz.ch/~salamon/PREPRINTS/cxana.pdf
[21] W. F. Trench, Introduction to real analysis. Prentice Hall/Pearson Education,

Upper Saddle River, N.J., 2003.
Zbl 1204.00023
http://ramanujan.math.trinity.edu/wtrench/misc/index.shtml
Index
—A— —C—
absolutely continuous Calderón–Zygmund inequality, 295
function, 239, 245 Carathéodory criterion, 61
measure, 179 Cauchy–Schwarz inequality, 147
signed measure, 200, 206 characteristic function, 12
almost everywhere, 36 closed set, 6
compact set, 10
compactly supported function, 113
—B— complete measure space, 43
Baire measurable complete metric space, 8
completion of a measure space, 43
function, 129
continuous map, 11
set, 129
continuum hypothesis, 147
Baire measure, 129
convergence in measure, 173
Baire -algebra, 129
convolution, 277
Banach algebra, 282
of signed measures, 307
Banach space, 8 counting measure, 21
Banach–Zarecki theorem, 239 cuboid, 64
basis of a topological space, 124
bilinear form
positive definite, 147
symmetric, 147 —D—
Boolean algebra, 6 dense subset, 141
Borel measurable Dieudonné’s measure, 96
function, 12 Dirac measure, 21
set, 9 divergence theorem, 306
Borel measure, 95 double arrow space, 126
extended, 107 dual space
left-invariant, 314 of a Hilbert space, 149
right-invariant, 314 of a normed vector space, 149
Borel outer measure, 108 of C.X /, 124, 218
Borel set, 9 of L2 ./, 152
Borel -algebra, 9 of Lp ./, 159
bounded linear functional, 149 of `1 , 160, 176
bounded variation, 243
352 Index
—E— —I—
Egoroff’s Thorem, 172 inner product, 147
elementary set, 250 on L2 ./, 148
envelope, 152 inner regular, 96, 208
essential supremum, 138 on open sets, 101
integrable function
Lebesgue, 32
—F— locally, 283
first countable, 124 partially defined, 46
Fubini’s Theorem Riemann, 87
for integrable functions, 265 weakly, 224
for positive functions, 261 integral
for the completion, 268 Haar, 315
Lebesgue, 22, 32
for the Lebesgue measure, 276
Riemann, 87
in polar coordinates, 305
Riemann–Stieltjes, 244
fundamental solution
invariant
of Laplace’s equation, 292
linear functional, 315
fundamental theorem
measure, 314
of calculus, 239 inverse limit, 311
—G— —J—
Jensen’s inequality, 168
Green’s formula, 293
Jordan decomposition, 200, 204
group
Jordan measurable set, 87
discrete, 309
Jordan measure, 87
Lie, 310
Jordan null set, 65
topological, 309
unimodular, 328
—L—
—H— Laplace operator, 292
Lebesgue
Haar integral, 315
differentiation theorem, 234, 238
Haar measure, 314
Lebesgue decomposition, 180
Hahn decomposition, 202
for signed measures, 201
Hahn–Banach theorem, 160
Lebesgue dominated
Hardy’s inequality, 171
convergence theorem, 35
Hardy–Littlewood Lebesgue integrable, 32
maximal inequality, 230 Lebesgue integral, 22, 32
Hausdorff dimension, 93 Lebesgue measurable
Hausdorff measure, 92 function, 57, 69
Hausdorff space, 11, 95 set, 57, 69
Hilbert space, 148 Lebesgue measure, 57, 69
Hölder inequality, 134 on the sphere, 304
Index 353
Lebesgue monotone measure, 19

convergence theorem, 25 absolutely continuous, 179
Lebesgue null set, 65 Baire, 129
Lebesgue outer measure, 65 Borel, 95
continuous from below, 70 Borel outer, 108
regularity, 70 counting, 21
Lebesgue point, 233 Dirac, 21
left-invariant Haar, 314
linear functional, 315 Hausdorff, 92
measure, 314 inner regular, 96, 101
lexicographic ordering, 126 Jordan, 87
Lie group, 310 Lebesgue, 57, 69
Lebesgue outer, 65
linear functional
left-invariant, 314
left/right-invariant, 315
localizable, 152, 174
localizable, 152, 174
locally determined, 271
strictly, 271
nonatomic, 147
locally compact, 95
outer, 58
Hausdorff group, 309
outer regular, 95
locally determined, 270 probability, 147
locally integrable, 283 product, 255
lower semi-continuous, 245 Radon, 96
lower sum, 87 regular, 96
Lp ./, 136 right-invariant, 314
Lp .Rn /, 137 semi-finite, 152
L1 ./, 138 -finite, 152
`p , 137 signed, 197
`1 , 159 singular, 179
strictly localizable, 271
translation-invariant, 57
—M— truly continuous, 206
measure space, 19
Marcinkiewicz interpolation, 289
complete, 43
maximal function, 229, 233 Lebesgue, 57, 69
measurable function, 12 localizable, 152, 174
Baire, 129 locally determined, 271
Borel, 12 semi-finite, 152
Lebesgue, 57, 69 -finite, 152
partially defined, 46 strictly localizable, 271
measurable set, 5 metric space, 7
Baire, 129 Minkowski inequality, 134, 263
Borel, 9 modular character, 328
Jordan, 87 mollifier, 286
Lebesgue, 57, 69 monotone class, 250
w.r.t. an outer measure, 58 mutually singular, 179
measurable space, 5 signed measures, 200
354 Index
—N— —R—
neighborhood, 95 Radon measure, 96
nonatomic measure, 147 Radon–Nikodým derivative, 236
norm of a bounded Radon–Nikodým theorem, 180
linear functional, 149 for signed measures, 202
normed vector space, 7 generalized, 209
null set, 36 regular measure, 96
Jordan, 65 Riemann integrable, 87
Lebesgue, 65 Riemann integral, 87
Riemann–Stieltjes integral, 244
Riesz representation theorem, 115
—O— right-invariant
one-point compactification, 127 linear functional, 315
open ball, 7 measure, 314
open set, 6
in a metric space, 7 —S—
outer measure, 58 second countable, 124
Borel, 108 semi-finite, 152
Lebesgue, 65 separability of Lp ./, 142
translation-invariant, 65 separable, 141
outer regular, 95 set of measure zero, 36
-additive
measure, 19
—P— signed measure, 197
p-adic integers, 311 -algebra, 5
p-adic rationals, 312 Baire, 129
partition of a set, 5 Borel, 9
partition of unity, 336 Lebesgue, 57, 69
perfectly normal, 126 product, 249
metric spaces are, 128 -compact, 95
Poisson identity, 293 -finite, 152
positive linear functional signed measure, 197
on Cc .X /, 113 absolutely continuous, 200, 206
on Lp ./, 161 concentrated, 200
pre-image, 11 Hahn decomposition, 202
probability theory, 51–55, 147 inner regular
product measure, 255 w.r.t. a measure, 208
complete locally determined, 270 Jordan decomposition, 200, 204
primitive, 270 Lebesgue decomposition, 201
product -algebra, 249 mutually singular, 200
product topology, 312, 340 Radon–Nikodým theorem, 202
pushforward total variation, 197
of a measure, 50 truly continuous, 206
of a -algebra, 13, 50 simple function, 17
singular measure, 179
Index 355
smooth function, 292 transformation formula, 77

Sorgenfrey line, 126, 128 translation-invariant
step function, 17 measure, 57
Stone–ech compactification outer measure, 65
of N, 176 triangle inequality, 7, 147
weak, 223
—T— truly continuous, 206
topological group, 309 Tychonoff’s theorem, 312
invariant measure, 314
left-invariant measure, 314 —U—
locally compact Hausdorff, 309 uniformly integrable, 214
right-invariant measure, 314 upper semi-continuous, 245
topology, 6 upper sum, 87
basis, 124 Urysohn’s lemma, 333
first countable, 124
group, 309 —V—
Hausdorff, 11, 95
Vitali’s covering lemma, 231
locally compact, 95
Vitali’s theorem, 215
perfectly normal, 126
Vitali–Carathéodory theorem, 245
product, 312, 340
second countable, 124
—W—
separable, 141
-compact, 95 weak triangle inequality, 223
standard on R x, 8 weakly integrable, 224
standard on R, 8
total variation —Y—
of a signed measure, 197 Young’s inequality, 133, 281
Textbooks in Mathematics
Dietmar A. Salamon
Dietmar A. Salamon Dietmar A. Salamon
Measure

The book is intended as a companion to a one semester introductory lecture course
on measure and integration. After an introduction to abstract measure theory it
proceeds to the construction of the Lebesgue measure and of Borel measures on
and
Integration
locally compact Hausdorff spaces, Lp spaces and their dual spaces and elementary
Hilbert space theory. Special features include the formulation of the Riesz
Representation Theorem in terms of both inner and outer regularity, the proofs of
the Marcinkiewicz Interpolation Theorem and the Calderon–Zygmund inequality as
applications of Fubini’s theorem and Lebesgue differentiation, the treatment of the
generalized Radon–Nikodym theorem due to Fremlin, and the existence proof for Haar
measures. Three appendices deal with Urysohn’s Lemma, product topologies, and the
inverse function theorem.
The book assumes familiarity with first year analysis and linear algebra. It is suitable
for second year undergraduate students of mathematics or anyone desiring an
introduction to the concepts of measure and integration.
ISBN 978-3-03719-159-0
www.ems-ph.org
Salamon Cover | Font: Frutiger_Helvetica Neue | Farben: Pantone 116, Pantone 287 | RB 31 mm

Measure and Integration PDF

Diunggah oleh

Informasi Dokumen

Judul Asli

Hak Cipta

Format Tersedia

Bagikan dokumen Ini

Bagikan atau Tanam Dokumen

Opsi Berbagi

Apakah menurut Anda dokumen ini bermanfaat?

Apakah konten ini tidak pantas?

Hak Cipta:

Format Tersedia

Measure and Integration PDF

Diunggah oleh

Hak Cipta:

Format Tersedia

Textbooks in Mathematics

Measure and Integration

EMS Textbooks in Mathematics is a series of books aimed at students or professional mathemati-

Jørn Justesen and Tom Høholdt, A Course In Error-Correcting Codes

Key words: sigma-Algebra, Lebesgue monotone convergence, Caratheodory criterion, Lebesgue

© 2016 European Mathematical Society

European Mathematical Society Publishing House

Phone: +41 (0)44 632 34 36

Typeset using the author’s TEX files: M. Zunino, Stuttgart, Germany

ETH, Zürich Dietmar A. Salamon

1 Abstract measure theory . . . . . . . . . . . . . . . . . . . . . . . . . . 3

2 The Lebesgue measure . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

5 The Radon–Nikodým theorem . . . . . . . . . . . . . . . . . . . . . . . 179

7 Product measures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 249

8 The Haar measure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 309

A Urysohn’s lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 333

B The product topology . . . . . . . . . . . . . . . . . . . . . . . . . . . 339

C The inverse function theorem . . . . . . . . . . . . . . . . . . . . . . . 343

sentation theorem. The notion of a Radon measure on a locally compact Hausdorff

The purpose of this first chapter is to introduce integration on abstract measure

Definition 1.1 (measurable space). Let X be a set. A collection A  2X of subsets

A measurable space is a pair .X; A/ consisting of a set X and a  -algebra

proves Lemma 1.2.

(c) if I is any index set and Ui 2 U for i 2 I , then i 2I Ui 2 U.

A topological space is a pair .X; U/ consisting of a set X and a topology

Let .X; kk/ be a normed vector space. Then the formula

defines a distance function on X. X is called a Banach space if the metric space

Exercise 1.13. Consider the set

.a; 1 WD .a; 1/ [ ¹1º; Œ 1; b/ WD . 1; b/ [ ¹ 1º:

Call a subset U  Rx open if it is a countable union of open intervals in R and sets

x satisfies the axioms of a topology. This

defined by the following formulas for x; y 2 R:

(iii) Prove that the map

(The sum a C b is undefined when ¹a; bº D ¹ 1; 1º.) Let a1 ; a2 ; a3 ; : : : and

(ii) Show that

(iii) Assume ¹an ; bn º ¤ ¹ 1; 1º so the sum an C bn is defined for n 2 N. Prove

lim sup.an C bn /  lim sup an C lim sup bn ;

lim inf an  lim inf bn :

(i) Every closed subset F  X is a Borel set.

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 bijective,

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 W X ! R x is called mea-

Example 1.18. Let X be a set. The characteristic function of a subset A  X is

is a -algebra on Y , called the pushforward of AX under f .

(i) A map f W X ! Y is measurable if an only if the pre-image of every open

(ii) Assume UX  2X is a topology on X such that AX is the Borel -algebra of

Proof. By part (iv) of Theorem 1.19, a map f W X ! Y is measurable if and only if

Theorem 1.21 (characterization of measurable functions). Let .X; A/ be a measur-

x by part (iii) of Theorem 1.19 and .a; 1 2 B for every

Theorem 1.22 (vector valued measurable functions). Let .X; A/ be a measurable

Proof. For i D 1; : : : ; n define the projection i W Rn ! R by i .x/ WD xi for

Q.a; b/ WD ¹x 2 Rn j ai < xi < bi , for all i º D .a1 ; b1 / .an ; bn /:

Lemma 1.23. Let .X; A/ be a measurable space and let u; vW X ! R be measur-

Theorem 1.24 (properties of measurable functions). Let .X; A/ be a measurable

(i) If f; gW X ! R are measurable functions, then so are the functions

f C g; fg; max¹f; gº; min¹f; gº; jf j:

(ii) Let fk W X ! R x , k D 1; 2; 3; : : : , be a sequence of measurable functions.

inf fk ; sup fk ; lim sup fk ; lim inf fk :

Proof. (i) The functions W R2 ! R defined by .s; t/ WD s C t, .s; t/ WD st ,

Definition 1.1 (measurable space). Let X be a set. A collection A 2X of subsets

A measurable space is a pair .X; A/ consisting of a set X and a -algebra

.a; 1 WD .a; 1/ [ ¹1º; Œ 1; b/ WD . 1; b/ [ ¹ 1º:

Call a subset U Rx open if it is a countable union of open intervals in R and sets

lim sup.an C bn / lim sup an C lim sup bn ;

lim inf an lim inf bn :

(i) Every closed subset F X is a Borel set.

for every subset B Y , and

for every collection of subsets Bi Y , indexed by a set I .

Example 1.18. Let X be a set. The characteristic function of a subset A X is

is a -algebra on Y , called the pushforward of AX under f .

(ii) Assume UX 2X is a topology on X such that AX is the Borel -algebra of

x by part (iii) of Theorem 1.19 and .a; 1 2 B for every

Proof. For i D 1; : : : ; n define the projection i W Rn ! R by i .x/ WD xi for

Proof. (i) The functions W R2 ! R defined by .s; t/ WD s C t, .s; t/ WD st ,

where A W X ! R is the characteristic function of the set Ai for i D 1; : : : ; `

0 s1 .x/ s2 .x/ f .x/; f .x/ D lim sn .x/ for all x 2 X:

lim n .t/ D t for all t 2 Œ0; 1:

Hence the functions sn WD n ı f satisfy the requirements of the theorem.

exists in Œ0; 1 and satisfy the familiar rules

(a) is -additive, i.e., if Ai 2 A, i D 1; 2; 3; : : : , is a sequence of pairwise

(b) There exists a measurable set A 2 A such that .A/ < 1.

A measure space is a triple .X; A; / consisting of a set X , a -algebra A 2X ,

Theorem 1.28 (properties of measures). Let .X; A; / be a measure space. Then

.A1 [ [ An / D .A1 / C C .An /:

(iii) If A; B 2 A are such that A B, then

(iv) Let Ai 2 A be a sequence such that Ai AiC1 for all i. Then

(v) Let Ai 2 A be a sequence such that Ai AiC1 for all i. Then

and hence .;/ D 0. This proves part (i).

Since Bi \ Bj D ; for i ¤ j , it follows from -additivity that

Since Ci \ Cj D ; for i ¤ j , -additivity yields

Exercise 1.29. Let .X; A; / be a measure

Example 1.33. Let X D i2I Ai be a partition and let A 2X be the -algebra

in Example 1.5. Then any function I ! Œ0; 1W i 7! ˛i determines a measure

(ii) Let f W X ! Œ0; 1 beR a measurable function. The (Lebesgue) integral of f

Proof. To prove (i), assume f g on E. If sW X ! Œ0; 1/ is a measurable step

Theorem 1.37 (Lebesgue monotone convergence theorem). Let .X; A; / be a

fn .x/ fnC1 .x/ for all x 2 X and all n 2 N:

exists in Œ0; 1. Moreover, f D supn fn is a measurable function on X , by part (ii)

En WD ¹x 2 X j cs.x/ fn .x/º for n 2 N:

Then En 2 A is a measurable set and En EnC1 for all n 2 N. Moreover,

Theorem 1.38 ( -additivity of the Lebesgue integral). Let .X; A; / be a measure

(i) If f; gW X ! Œ0; 1 are measurable and E 2 A, then

(ii) Let fn W X ! Œ0; 1 be a sequence of measurable functions and define

Then f W X ! Œ0; 1 is measurable and, for every E 2 A,

(iii) If f W X ! Œ0; 1 is measurable and E1 ; E2 ; E3 ; : : : is a sequence of pairwise

Now fn W X ! Œ0; 1 is a nondecreasing sequence of measurable functions con-

for every map N N ! Œ0; 1W .i; j / 7! aij .

Theorem 1.40. Let .X; A; / be a measure space and let f W X ! Œ0; 1 be a

for every measurable function gW X ! Œ0; 1 and every E 2 A.