Math 521 Spring 2010

521
May 12, 2010

Contents
1 Pep Talk 2

2 Notation 5

3 Logic 6

4 Sets 8

5 Functions and Maps 11

6 Composition and Inverse 13

7 Mathematical Induction 15

8 Infinite Sets 16

9 Axioms for the Real Numbers 18

10 Distance 22

11 Limits 24

12 Convergence of Sequences 26

13 Continuity 32

14 Open Sets and Closed Sets 34

1
15 Connected Sets 37

16 Compact Sets 39

17 Derivatives 43

18 The Integral 44

19 Taylors Formula 49

20 Series 50

21 Uniform Convergence 56

22 Power Series - I 59

23 Analytic Functions 61

24 Power Series - II 64

25 The Heat Equation 66

26 Fourier Series - I 68

27 Fourier Series - II 68

A Uniqueness of the Real Numbers 68

B Additional Problems 71

1 Pep Talk
What does not kill me, makes me stronger.
Friedrich Nietzsche (1844 - 1900)
in The Twilight of the Idols (1899)

It is generally acknowledged in the math department that the two hardest

undergraduate courses we teach are 222 and 521. The former is hard because

2
for many students it is their first college math course, and although they did
well in high school, they are unprepared for what is expected in college. The
latter is hard for many reasons. For many students it is the first course where
they are expected to understand and write proofs. (See 3.4.) There are a lot
of hard words like if, for all, there exists, (see Chapter 3) and the order
in which they appear is crucial (see 13.3). Inequalities (see 9.2) play a crucial
role. The language of set theory (see Chapter 4) is used heavily and some
students have a hard time visualizing what set is. On top of all this there
is some strange lingo like connected, compact, open cover, etc. The first
time you think you understand and discover that I disagree you may well
decide that it is easier to get hit by a truck.
The subject is hard. Calculus was invented by Leibniz and Newton in
the seventeenth century, but the ideas presented in these notes were not fully
developed until the early twentieth century. Some of the best mathematical
minds in human history contributed to the subject. But you have one ad-
vantage over these luminaries: a teacher who understands the subject and
wants to help you understand.
Given the task before you, why should you make the effort? The above
quote of Nietzsche may inspire you. The ability to reason carefully will
serve you well in any future endeavor. You may be driven (like I am) to
understand the patterns in mathematical reasoning. These patterns help us
see the similarities between apparently different problems and enable us to
use our understanding of one problem to solve the others. Finally, you may
actually need to use the material someday. Many mathematical problems
have no explicit solution: we must prove that there is a solution and use
that proof both to guide us to an approximate solution and to estimate the
accuracy of that solution. If we get to the end of these notes in this course
you will get a glimpse of how this works when we solve the heat equation in
problem 25.2.
Here are some tips:

1. Read. It is a bad idea to try to learn just from lectures. A really good
lecturer can make the material look easy, but just listening is rarely
enough to lead to true understanding.

2. Focus on definitions. It is impossible to understand why a compact

set is closed and bounded if you dont know what the words mean.
Use the index and cross references in these notes or the text you are

3
 reading to recall definitions and previous theorems. If you are reading
these notes on a computer and have a sufficiently recent version of
Adobe Acrobat Reader or some other pdf viewer, you will notice that
this document contains embedded hypertext links. Some viewers even
allow text searches and links to URLs. (If you are online try clicking
on the link to the official description below.) This makes the task of
jumping around the text much easier.

3. Be active not passive. As you read or listen, make up examples to test

your understanding. When you see a definition make up an example of
something which satisfies it and something which doesnt.

4. Ask questions. If you can formulate a question when you are confused,
you may discover that you become unconfused and dont need to ask
the question. But dont hesitate to ask me questions in class or office
hours.

5. Be concise when you write. Excess verbiage is hard to follow and may
conceal confusion.

6. Dont fall behind. If you postpone understanding till the night before
the exam, you may find that you are still lost. For each lecture there
will be a short prequiz on Moodle (see below) designed to get you to
read the material before I lecture on it. In addition there will be a
short post quiz almost every day (I hope).

We will use an online course management system called Moodle. To use

it go to https://www.math.wisc.edu/moodle and click on Math 521. You
will be prompted to authenticate. Use your net id and net password the same
as if you were logging on the MyUW or WiscMail. If you are not enrolled in
the course (according to the registrar) you will not be allowed in.
The text references in these notes are to the following texts:

(Buck) R.Creighton Buck, Advanced Calculus, 3rd edition, Waveland Press.

(Lang) Serge Lang, Undergraduate Analysis, 2nd edition, Springer Under-

graduate Texts in Mathematics.

(Morgan) Frank Morgan, Real Analysis, American Mathematical Society

4
(Rudin) Walter Rudin, Principles of Mathematical Analysis, McGraw Hill.
There are three sections of 521 this semester and each has chosen one of the
first three. The instructor I am replacing chose Buck. I used this book the last
time I taught 521. The terminology, content, even the title is are somewhat
dated. (The math department will soon change the name of 521 from Ad-
vanced Calculus to Real Analysis I.) Lang is good but somewhat ponderous.
It contains more material than can be comfortably treated in a two semester
course. My favorite is from the first three is Morgan. It is student friendly
and contains almost exactly the material I hope to cover. (See the official de-
scription of 521 at http://www.math.wisc.edu/521-advanced-calculus.)
It is also the least expensive. The Rudin text is the best for someone who
already thinks like a mathematician, but is difficult for the apprentice.
I have arranged these notes so you can use them with any of these texts
there are remarks and footnotes which explain slight differences in termi-
nology but the notes are designed with the Morgan text in mind. Possibly
you wont even need to buy a text: all three texts are on reserve in the Math
Library. The prequizzes will only require you to read the course notes.

2 Notation
The following standard notations are used:1

Z+ the set of positive integers (natural numbers).

N the set of nonnegative integers.
Z the set of integers.
Q the set of rational numbers.
R the set of real numbers.
C the set of complex numbers.

The usual interval notation is used, e.g.

[a, b) := {x R : a x < b}, (, b] := {x R : x b}, etc.

We write := to indicate that two objects are equal by definition. We also

signal definitions by writing iff instead of if and only if.
1
Morgan uses N denote the positive integers while Lang uses N to denote the nonneg-
ative integers.

5
3 Logic
3.1. In these notes we shall often use the abbreviations

= for implies,
for if and only if,
for there exists,
for for all.

These abbreviations help clarify the logical structure of the definitions and
theorems. They usually arent used in textbooks but are commonly used in
lectures.

3.2. The logical operations are defined by the following table:

P Q P and Q P or Q P = Q P Q not P
T T T T T T F
T F F T F F
F T F T T F T
F F F F T T

It is particularly important to notice that the fifth column asserts that false
implies anything is true. Note also that the use of the word or is inclusive
as in either x > 0 or else x < 5 (true) not exclusive as in soup or salad?
(not both).

3.3. The following are synonymous:

P = Q.

P implies Q.

If P, then Q.

If not Q, the not P.

Q, if P.

P only if Q.

6
Do not confuse the statement if Q, then P with its converse if P, then Q:
it can happen that one is true and the other is not as in x = 2 = x2 = 4.
The statement If not Q, the not P is called the contrapositive of the
statement if P, then Q. A statement and its contrapositive are equivalent.
This provides the justification for proof by contradiction.

3.4. One of the principal aims of this course is to teach you how to read and
write proofs. A proof is an argument intended to convince the reader that a
general principle is true in all situations. The amount of detail that an author
supplies in a proof should depend on the audience. Too little detail leaves
the reader in doubt; too much detail may leave the reader unable to see the
forest for the trees. As a general principle, the author of a proof should be
able to supply the reader with additional detail on demand. When a student
writes a proof for a teacher, the aim is usually not to convince the teacher of
the truth of some general principle (the teacher already knows that), but to
convince the teacher that the student understands the proof and can write
it clearly.
The theorems below show the proper format for writing a proof. In
each of them you are supposed to imagine that the theorem to be proved
has the indicated form. Notice how the key words choose, assume, let, and
therefore are used in the proof. In these sample formats, the phrase Blah
Blah Blah indicates a sequence of steps, each one justified by earlier steps.
Theorem If P, then Q.
Proof. Assume P. Blah Blah Blah. Therefore Q.

Theorem P if and only if Q.

Proof. Assume P. Blah Blah Blah. Therefore Q. Conversely, assume Q. Blah
Blah Blah. Therefore P.

Theorem P(x) for all x.

Proof. Choose x. Blah Blah Blah. Therefore P(x).

Theorem There is an x such that P(x).

Proof. Let x = . . .. Blah Blah Blah. Therefore P(x).

7
4 Sets
4.1. A set A divides the mathematical universe into two parts: those objects
x that belong to A and those that dont. The notation x A means x belongs
to A. The notation x / A means that x does not belong to A. The objects
that belong to A are sometimes called the elements of A but we will often call
them points or numbers. Other words roughly synonymous with the word set
are class, collection, and aggregate. These longer words are generally used to
avoid using the word set twice in one sentence. The situation typically arises
when an author wants to talk about sets whose elements are themselves sets.
One might write the collection of all finite sets of integers, rather than the
set of all finite sets of integers.
4.2. For two sets A and B, the notation A B means that A is a subset
of B, i.e. for all x we have x A = x B. By definition, two sets are
equal if each is a subset of the other:

A = B A B and B A.

The notation {x : P (x)} denotes the set of all x for which the property P (x)
is true. The notation {x A : P (x)} denotes the set of all x A for which
the property P (x) is true. Finite sets may be defined by enumerating their
elements as in

x {a1 , a2 , . . . , an } x = a1 or x = a2 or or x = an

and often infinite sets as well as in

N = {0, 1, 2, . . .}, Z = {. . . , 2, 1, 0, 1, 2, . . .}, Z+ = {1, 2, 3, . . . , }.

4.3. If A and B are sets, then the sets

A B := {x : x A or x B}, A B := {x : x A and x B},

are called respectively the union and intersection of A and B. The empty
set is denoted : For all x it is true that

x
/ .

Two sets are disjoint iff they have no elements in common, i.e iff A B = .
The set
X \ A := {x X : x
/ A}

8
is called the complement of A in X.2 The set
A B := {(x, y) : x A and y B}
of all ordered pairs (x, y) with x A and y B is called the Cartesian
product of A and B. The term direct product is a synonym. We also use
the notation
An := A
| A {z A}
n
In particular,
Rn := {(x1 , x2 , . . . , xn ) : xi R}
denotes the vector space of all ntuples of real numbers, so R1 = R, R2 =
R R, R3 = R R R, etc.3
Remark 4.4. Do not confuse an ntuple (finite sequence of length n) with
a finite set. For the former order is important: {3, 7} = {7, 3} but (3, 7) 6=
(7, 3); for the latter repetitions dont matter: {2, 2, 3} = {2, 3} but (2, 2, 3) 6=
(2, 3).
4.5. An indexed family of sets is a function which assigns a set Ai to each
element i of a set I. The set I is called the index set of the family and
the family is usually denoted (Ai )iI . The union and intersection of the
indexed family are defined by
[
x Ai i I such that x Ai .
iI
\
x Ai i I we have x Ai .
iI

The notation i I is an abbreviation for there exists i I, and the

notation i I is an abbreviation for for all i I. In these definitions the
set I can be infinite. For finite sets I we recover the earlier definitions, e.g.
for I = {1, 2} we have
[ \
Ai = A1 A2 , Ai = A1 A2 .
i{1,2} i{1,2}

This illustrates the logical principle that is like an infinite or and is

like an infinite and.
2
Morgan uses the notation A{ for Rn \ A (when A is a subset of Rn ).
3
Buck uses the term n space as a synonym for Rn .

9
Remark 4.6. Set theory is simply a way of formalizing logic. Simple set
theoretic identities may be proved by truth tables. For example, consider
the following distributive law

(A B) C = (A C) (B C).

To show that x (A B) C x (A C) (B C) we can simply

consider all the possibilities:
A B C (A B) C (A C) (B C)
T T T T T
T T F F F
T F T T T
T F F F F
F T T T T
F T F F F
F F T F F
F F F F F

It is usually not necessary to show so much detail in your written work, but
it will be hard for you to decide just how much detail is appropriate. A
good rule of thumb is that you should be prepared to supply more detail if
challenged.
4.7. From logic we know that

not not, and not not,

so for any set X and any indexed family (Ai )iI we have
[ \ \ [
X \ Ai = (X \ Ai ), X \ Ai = (X \ Ai ).
iI iI iI iI

Logicians call these De Morgans Laws. In particular, for I = {1, 2} we

have
X \ (A1 A2 ) = (X \ A1 ) (X \ A2 ),
X \ (A1 A2 ) = (X \ A1 ) (X \ A2 ).
Also by logic we have set theoretic distributive laws
[ [ \ \
X Ai = (X Ai ), X Ai = (X Ai ).
iI iI iI iI

10
In particular, for I = {1, 2} we have

X (A1 A2 ) = (X A1 ) (X A2 ),
X (A1 A2 ) = (X A1 ) (X A2 ).

The latter was proved above in Remark 4.6.

5 Functions and Maps

5.1. A function is a rule which assigns a value f (x) to every point x from
a set called the domain of the function. The set

graph(f ) := {(x, y) : y = f (x)}

of all pairs (x, y) such that y = f (x) is called the graph of the function f .
Two functions are equal iff they have the same graph.
5.2. Let X and Y be sets. We say that f is a map from X to Y and write
f : X Y when f is a function which assigns a point y = f (x) Y to each
point x X. Two maps f : X Y and f 0 : X 0 Y 0 are said to be equal
when X = X 0 , Y = Y 0 , and f (x) = f 0 (x) for all x X. Thus if f and f 0
equal maps, then graph(f ) = graph(f 0 ) but not conversely (because Y = Y 0
is part of the definition of equality for maps). However most authors would
say that two functions are equal iff they have the same graph.
Remark 5.3. Some authors use the notation x 7 f (x) to define a map.
This allows them to avoid introducing a name for the map. Thus instead of
writing
Consider the map f : R R defined by f (x) = x5 + x.
they may write
Consider the map R R : x 7 x5 + x.
5.4. When A X, B Y , and f : X Y , the sets

f (A) := {y Y : x A such that y = f (x)},

f 1 (B) := {x X : f (x) B},

are called respectively the image of A by f and inverse image of B by f .

11
Remark 5.5. The sets X and Y are sometimes called the source and target
of a map f : X Y . The image f (X) of the source is what is called the
range of the function f in calculus. Thus the domain of a map is the same
as its source while the range is a subset of its target.

Remark 5.6. There are slight variations in terminology among authors.

Morgan avoids the word map and Lang sometimes uses the word mapping.
Buck uses the term the preimage instead of inverse image. Lang (page 4)
uses the 7 notation but the other authors apparently avoid it. For Lang
a function is a map whose target is R. In precalculus courses a function is
usually defined by an expression and the domain is implicitly taken to be the
largest set of numbers for which the expression is meaningful, but in advanced
mathematics authors usually make the domain explicit. Examples 6.7 below
shows why the distinction is important.

Problem 5.7. Let f : X Y , (Ai )iI be a family of subsets of X, and

(Bi )iI be a family of subsets of Y . Which of the following are (always)
true? Hint: When there is a counterexample, there is a counterexample with
the index set I = {1, 2}.
! !
[ [ [ [
f 1 Bi f 1 (Bi )? f 1 (Bi ) f 1 Bi ?
iI iI iI iI
! !
\ \ \ \
f 1 Bi f 1 (Bi )? f 1 (Bi ) f 1 Bi ?
iI iI iI iI
! !
[ [ [ [
f Ai f (Ai )? f (Ai ) f Ai ?
iI iI iI iI
! !
\ \ \ \
f Ai f (Ai )? f (Ai ) f Ai ?
iI iI iI iI

Problem 5.8. Proof or counterexample:

(1) If f : X Y and A1 , A2 X, then f (A1 \ A2 ) f (A1 ) \ f (A2 )?

(2) If f : X Y and A1 , A2 X, then f (A1 ) \ f (A2 ) f (A1 \ A2 )?

(3) If f : X Y and B1 , B2 Y , then f 1 (B1 \ B2 ) f 1 (B1 ) \ f 1 (B2 )?

12
(4) If f : X Y and B1 , B2 Y , then f 1 (B1 ) \ f 1 (B2 ) f 1 (B1 \ B2 )?

Problem 5.9. Proof or counterexample:

(1) If f : X Y and A X, then f 1 (f (A)) A?

(2) If f : X Y and A X, then A f 1 (f (A))?

(3) If f : X Y and B Y , then f (f 1 (B)) B?

(4) If f : X Y and B Y , then B f (f 1 (B))?

6 Composition and Inverse

6.1. If f : X Y and g : Y Z, then the composition of f and g is the
map g f : X Z defined by

(g f )(x) = g(f (x))

for x X. For any set X the identity map of X is the map idX : X X
defined by
idX (x) = x
for x X. Clearly
idY f = f and f idX = f
for f : X Y .

6.2. A map g : Y X is said to be a left inverse for the map f : X Y

iff g f = idX , i.e. iff g(f (x)) = x for all x X. A map g : Y X is
said to be a right inverse for the map f : X Y iff f g = idY , i.e. iff
f (g(y)) = y for all y Y . A map g : Y X is said to be a (two sided)
inverse to the map f : X Y iff it is both a left inverse and a right inverse
to f . If g is a left inverse to f and g 0 is a right inverse to f then g = g 0 .
(Proof: g = g idX = g f g 0 = idY g 0 = g 0 .) In this case there is a unique
inverse and it is denoted f 1 . So if f : X Y has an inverse f 1 : Y X,
then
y = f (x) x = f 1 (y)
for x X and y Y .

13
Definition 6.3. A map f : X Y is said to be injective iff

x1 , x2 X [f (x1 ) = f (x2 ) = x1 = x2 ]

and it is said to be surjective iff Y = f (X), i.e.

y Y x X y = f (x).

A map is bijective iff it is both injective and surjective. Thus

(1) A map is injective if and only if it has a left inverse;

(2) A map is surjective if and only if it has a right inverse;

(3) A map is bijective if and only if it has a (two-sided) inverse.

In the lingo used in Remark 5.5, a map is surjective if and only if its range
equals its target.
Remark 6.4. Both Lang and Morgan use the more modern terms injective,
surjective, bijective, but Buck and Rudin use the older terminology one-one
instead of injective, onto instead of surjective, and one-one onto instead of
bijective. The older terminology is used in teaching calculus.
Remark 6.5. The only if part of item (2) is called the Axiom of Choice.
It was once controversial because one can imagine a situation where one can
prove that a map f is surjective but where one cannot give an explicit formula
for a right inverse.
Example 6.6. The assertions (1-3) in Definition 6.3 are false if continuity
(defined later) is required: There is a continuous bijective map whose inverse
is not continuous (and hence continuous injective map which does not have a
continuous left inverse, and a continuous surjective map which does not have
a continuous right inverse). Let S := {(x, y) R2 : x2 + y 2 = 1} denote the
unit circle in R2 and define f : [0, 2) S by f () = (cos , sin ). Then f
bijective and continuous but f 1 is not continuous. In Theorem 16.9 below
we will see that the inverse if a continuous bijective map is continuous if its
domain is compact (defined later), but [0, 2) is not compact.
Example 6.7. Consider the four maps

f1 : R R, f2 : [0, ) R, f3 : R [0, ), f4 : [0, ) [0, )

14
defined by fi (x) = x2 . Then f1 is not injective and not bijective, f2 is injective
but not surjective, f3 is surjective but not injective, and f4 is bijective. Any

map g2 : R [0, ) such that g2 (y) = y for y 0 is a left inverse to f2 ,

and any map g3 : [0, ) R such that g3 (y) = y (the can depend on

y) is a right inverse to f3 The inverse map to f4 is f41 (y) = y.
Problem 6.8. Which (if any) of the false formulas in problem 5.7 become
true if we assume that the map f is injective? surjective? (Proof or counter
example.)

7 Mathematical Induction
7.1. The principle of mathematical induction is the following axiom.
Let S N be a set of nonnegative integers. Assume that 0 S
and that n S = n + 1 S. Then S = N.
It is usually taught in College Algebra as a means of proving identities like
n n
X n(n + 1) X n(n + 1)(2n + 1)
k= and k2 = .
k=0
2 k=0
6

For example to prove the first identity let S denote the set of all n N for
which the statement is true. Then 0 S and if n S then
n+1 n
!
X X n(n + 1) (n + 1)(n + 2)
k= k + (n + 1) = + (n + 1) =
k=0 k=0
2 2

so n + 1 S. Hence by the principle of mathematical induction, S = N, i.e.

the identity is true for all n N.
7.2. We will often use inductive definitions where a sequence is defined
by specifying the first few values and then giving a recurrence relation for
determining the remaining values from the earlier ones. For example, the
Fibonacci numbers are defined by

F0 = 0, F1 = 1, Fn = Fn1 + Fn2 .

Thus F2 = 1, F3 = 3, F4 = 5, F5 = 8, F9 = 55 etc.
Problem 7.3. Prove the sum of squares formula in 7.1

15
8 Infinite Sets
Definition 8.1. A set S is said to be

infinite iff there is a injective map f : N S,

finite iff it is not infinite,

countable iff there is a bijective map J S for some J N, and

uncountable iff it is not countable.4

Proposition 8.2. If there is a bijective map {1, 2, . . . , m} {1, 2, . . . , n},

then m = n. Hence if S is finite, there is a unique n (called the cardinality
of S) such that there is a bijective map {1, 2, . . . , n} S.

Problem 8.3. Suppose that S and T are finite sets of cardinality s and t
respectively? What is the cardinality of S T ? Of the set T S of maps from
S to T ? Of the set 2S of subsets of S? Of S T (assuming that S T = )?

The following four propositions are proved in Chapter 2 of Morgan. See

also Lang pages 12-15. The main value of these propositions is to give the
student experience in constructing maps. The constructions are not needed
in the rest of the course, but learning them will give you some confidence.

Proposition 8.4. A subset of a countable set is countable.

Proof. A subset of a finite set is finite so it is enough to consider a subset

S N of the nonnegative integers. Define sn inductively (see 7.2) by sn+1 =
the least element of S \ {s1 , . . . , sn }. If for some n the set S \ {s1 , . . . , sn }
is empty then the induction stops and the set S is finite of cardinality n.
Otherwise, the map N S : n 7 sn is injective. But it is also surjective: if
k S then k = sn for some n k.

Proposition 8.5. If S and T are countable, so is S T .

4
This this the terminology used by Morgan. For Buck and Rudin countable means
countable and infinite which is equivalent to saying that there is a bijective map N S.
Lang uses the terms denumerable and nondenumerable for countable and infinite and
uncountable.

16
Proof. It is enough to prove that N N is countable. Consider the following
enumeration of N N:
f (0) = (0, 0) f (2) = (0, 1) f (5) = (0, 2) f (9) = (0, 3) f (14) = (0, 4)
f (1) = (1, 0) f (4) = (1, 1) f (8) = (1, 2) f (13) = (1, 3)
f (3) = (2, 0) f (7) = (2, 1) f (12) = (2, 2)
f (6) = (3, 0) f (11) = (3, 1)
f (10) = (4, 0)
.. .. .. .. .. ..
. . . . . .

This defines a bijection f : N N N. (Compare with Morgan page 10.)

On each diagonal the sum of the coordinates is constant and the input n to
the function f (n) increases as you go from left to right up the diagonal.

Problem 8.6. For f as in Proposition 8.5, find f (1307) and f 1 (58, 19).
Hint: The formula for the sum of the first n natural numbers from 7.1 will
be helpful.

Proposition 8.7. The set Q of rational numbers is countable.

Proposition 8.8. The set R of real numbers is uncountable.

Theorem 8.9. The set 2N of all subsets of N is uncountable.

Proof. Suppose not. Then there is a bijective map N 2N : n 7 An . Let

B = {n N : b / An } Then B = Ak for some k. Either k Ak or k / Ak .
If the former, then k
/ B so Ak 6= B. If the latter, k B, so again Ak 6= B.
Either way we have a contradiction to the assumption that the map n 7 An
is bijective.

Theorem 8.10 (Dedekind). A set S is infinite if and only if there is an

injective map from S to itself which is not surjective.

Theorem 8.11 (Cantor Schroeder Bernstein). If there is an injective map

from S to T and there is an injective map from T to S, then there is a
bijective map from S to T .

Proof. Buck page 552.

17
9 Axioms for the Real Numbers
We state here the axioms for the real number system R. We shall accept
these axioms without proof but it can proved (from more general axioms)
that there is an essentially unique structure satisfying them.

9.1. Algebraic Axioms. The set R of real numbers is equipped with two
operations

R R R : (a, b) 7 a + b, R R R : (a, b) 7 a b

such that the usual laws of grade school arithmetic hold:

(Commutative Laws) a + b = b + a and a b = b a.

(Associative Laws) (a + b) + c = a + (b + c) and (a b) c = a (b c).

(Distributive Law) (a + b) c = (a c) + (b c).

(Zero,One) There are (necessarily unique) distinct elements 0, 1 R such

that a + 0 = a and a 1 = a for all a R.

(Inverses) For every a R there is a (necessarily unique) element a such

that a + (a) = 0. For every a R \ {0} there is a (necessarily unique)
element a1 such that a a1 = 1.

The standard notations from high school algebra are used: in particular,
ab := a b, a b := a + (b), and a/b = a b1 .

9.2. Order Axioms. The set R has an order relation denoted a < b satis-
fying the following laws for all a, b, c R:

(Trichotomy) Exactly one of the alternatives a < b, a = b, b < a, holds.

(Transitivity) a < b, b < c = a < c.

(Addition) a < b = a + c < b + c.

(Multiplication) 0 < a, b = 0 < ab.

The other order notations are defined as usual, i.e. a < b b > a and
a b b a either a < b or a = b.

18
Remark 9.3. All the rules of algebra used in College Algebra (Math 112)
follow from the Algebraic Axioms 9.1 and Order Axioms 9.2. For example,
(a + b)2 = a2 + 2ab + b2 , a2 0, etc.
Problem 9.4. Does it follow from the axioms in 9.1 and 9.2 that a1 > 0 if
a > 0? Explain.
Definition 9.5. A set S of real numbers is said to be bounded above iff
there is a number b R such that x b for all x S; the number b is then
called an upper bound for S. A number b R is called a least upper
bound for S iff it is an upper bound for S and b b0 for every other upper
bound b0 for S. Similarly the set S is said to be bounded below iff there
is an number a R such that a x for all x S; the element b is then
called a lower bound for S. An element a R is called a greatest lower
bound iff it is an lower bound for S and a0 a for every other lower bound
a0 for S. The words infimum and greatest lower bound are synonymous as
are the words supremum and least upper bound. The least upperbound of
the set S will be denoted sup(S) and the greatest lower bound of the set S
will be denoted inf(S). We write sup(S) = when S is not bounded above
and inf(S) = when S is not bounded below.
9.6. Completeness Axiom. Every set S of real numbers which is bounded
above has a least upper bound, i.e.
if x b for all x S, then x sup(S) b for all x S.
Because multiplication by 1 reverses the order it is the same to say that
every set which is bounded below has a greatest lower bound. Thus
if a x for all x S, then a inf(S) x for all x S.
Remark 9.7. For Morgan the completeness axiom is a theorem (see page 44
of his book) but most authors of undergraduate texts dont prove it. Of
course it cant be proved until we give a precise construction (definition) of
the real numbers. Morgan takes the view that a real number is a number
with a decimal expansion, but avoids certain subtle issues connected with
this view. Actually making his definition precise would involve defining how
to do the arithmetic operations with decimal expansions and would be a bit
tedious. To see why, imagine that a + b = c and that

X
X
X
a= an 10n , b= bn 10n , c= cn 10n ,
n=1 n=1 n=1

19
where the coefficients, an , bn , cn are integers between 0 and 9 and try to
express cn in terms the an s and the bn s. In an appendix to his book, Buck
sketches a construction the real numbers by something called Dedekind
cuts but leaves out many details. Rudin explains cuts in an appendix to
his chapter 1. (The book Foundations of Analysis by Edmund Landau gives
all the details.) The completeness axiom is an easy consequence of this
construction. One can also define the real numbers using equivalence classes
of Cauchy sequences (see 12.15) and again the completeness axiom is a
consequence; if we get to the topic of completion of metric spaces, Ill explain
it then.
The crucial point is the uniqueness up to isomorphism which we will
prove in Appendix A. It means that if we start from the axioms, it doesnt
matter what definition of the real numbers we use. We will not use the
completeness axiom until the proof of Theorem 12.7 and the exposition is
arranged in such a way that it will be clear that the reasoning is not circular.
The following exercises will help you understand some of the issues.
Problem 9.8. Prove the following Archimedean Property of the real
numbers: There is neither an infinite real number nor an infinitesimal real
number. More precisely,
(1) There is no real number which is larger than every integer.
(2) For every positive real number > 0 there is a positive integer n such
that 1/n < .
Hint: If > n for every integer n what about 1? The proof will use the
completeness axiom.
Problem 9.9. Let R denote the set of real valued rational functions, i.e.
f R iff f (x) = p(x)/q(x) where p(x) and q(x) are polynomials with real
coefficients (and q(x) is not the zero polynomial). For f, g R define an
order relation by the condition that f > g iff there exists an M such that
f (x) > g(x) for all x > M . Then the set R satisfies the algebraic axioms
and order axioms given above. View R (and hence Z) as a subset of R
by identifying the real number c with the constant function whose value is
always c. Exhibit (in the lingo of Problem 9.8) an infinite element f R
and an infinitesimal element g R. Hint: What is limx p(x)
q(x)
?
Problem 9.10. Prove that there is a rational number in every nonempty
open interval (a, b) R.

20
Solution: By the Archimedean property from Problem 9.8 there are integers
m and n with m < a < b < n. Again by the Archimedean property there
is a positive integer k with 0 < 1/k < b a. Define qj = m + j/k for
j = 0, 1, . . . , k(n m). Then qj = m when j = 0, qj = n when j = k(n m),
and qj qj1 = 1/k < b a. If j is the largest integer such that qj1 a
then a < qj < b as qj qj1 = 1/k < b a.

Problem 9.11. Let Q( 2) denote the set of all numbers of form x = a+b 2
where a and b are rational.Show that Q( 2) is closed under the
algebraic
operations, i.e if x, y Q( 2), then x
y Q( 2), xy Q( 2), and (if
y 6= 0) x/y Q( 2). Show further that 3 / Q( 2). Does Q( 2) satisfy
the completeness axiom?
Problem 9.12. Assume that

X
x= xk 10k
k=1

is the decimal expansion of a real number x so that the xk are integers

between 0 and 9. How large must n be to ensure that
 
 n
X 
k 
x x 10  < 107 ?

 k
 
k=1

Assume that a + b = c as in Remark 9.7 and that

n
X n
X
k
A= ak 10 , B= bk 10k .
k=1 k=1

How large must n be to ensure that |c (A + B)| < 107 ?

Solution: For any n we have
 

X 
X
X
k  k
x x 10 = x 10 9 10k = 10n

 k  k
 
k=n k=n+1 k=n+1

We have equality if xk = 9 for k > n so to be sure that the strict inequality

holds we should take n = 8. For the second part we have

|(a + b) (A + B)| |a A| + |b B| 2 108

21
by the first part and the triangle inequality. Since 2108 < 107 we have that
n = 8 works here as well. Note: The point of this problem is to convince you
that it is awkward to work with decimal expansion if you want to be super
careful. Also the wording of the problem (How large must n be to ensure
that) suggests that I want the smallest possible n that works. Usually I
dont care about that. I could have replaced the phrase How large must n
be to ensure that by Find N so that for n N we have. In that case the
answer n = 13472 would be correct (but overkill).

10 Distance
10.1. The distance d(p, q) between two points p = (x1 , x2 , . . . , xn ) and q =
(y1 , y2 , . . . , yn ) in Rn is defined by5
p
d(p, q) = (x1 y1 )2 + (x2 y2 )2 + + (xn yn )2 .

The distance d(v, 0) from a vector v Rn to the origin is called the norm
of v denoted by |v| so
d(p, q) := |p q|.
The norm satisfies the following laws for v, w Rn :

(zero norm) |v| = 0 v = 0,

(homogeneity) |av| = a |v| if a > 0,

(symmetry) | v| = |v|,

(triangle inequality) |v + w| |v| + |w|.

The zero norm law holds because a sum of squares vanishes only if each
summand vanishes and the triangle inequality is proved in problem 10.3
below.6 The laws for the norm imply that the distance function satisfies the
following:

(zero distance) d(p, q) = 0 p = q,

5
In Chapters I-V Buck writes |p q| instead of d(p, q) but uses the notation d(p, q)
starting in Chapter VI (page 304) in a more general setting.
6
See also the Corollary on page 14 of Buck or Theorem 2.1 on page 134 of Lang.

22
(symmetry) d(p, q) = d(q, p),

(triangle inequality) d(p, r) d(p, q) + d(q, r).

These are proved by reading v = p q and w = q r in the corresponding

law for the norm.

10.2. Define the inner product of two vectors in Rn by

hu, vi = u1 v1 + u2 v2 + + un vn

for u = (u1 , u2 , . . . , un ), v = (v1 , v2 , . . . , vn ). In freshman calculus you called

this the dot product and learned that the angle between the two vectors
u and v satisfies
hu, vi = cos |u| |v|.
Since the cosine takes values between -1 and +1, it follows that

|hu, vi| |u| |v|.

This inequality is called the Schwarz inequality.

Problem 10.3. In this problem you will prove the triangle inequality with-
out using trigonometry in two steps.

1. Prove the Schwarz inequality. (Hint: The quadratic polynomial f (t) =

|u + tv|2 = At2 + Bt + C is always nonnegative so its discriminant
B 2 4AC is nonpositive.)

2. Derive the triangle inequality as a corollary. (Hint: Compare |u v|2

and (|u| + |v|)2 .)

Definition 10.4. For p Rn and > 0 the set

B(p, ) := {q Rn : d(p, q) < }

is called the open ball centered at p with radius . When X Rn we often

use the abbreviation

BX (p, ) := B(p, ) X := {q X : d(p, q) < }

23
Remark 10.5. When n = 1 the open ball is an open interval:

B(a, ) = {x R : |x a| < }
= {x R : a < x < a + }
= (a , a + )

for a R

Definition 10.6. A set S is bounded iff it is contained in some large ball,

i.e. iff there exists M > 0 such that |p| < M for all p S. Thus a set of real
numbers is bounded (see 9.5) if and only if it is bounded above and bounded
below.

11 Limits
The intuitive idea of the notation

lim F (p) = L
pp0

is that F (p) is very close to L when p is very close to p0 . Some authors write
F (p) L as p p0 ; others write F (p) L when p p0 . In this chapter
we give a more precise definition. The following lingo is helpful.

11.1. A set U is called a neighborhood of the point p if U contains some

open ball B(p, ) centered at p. A punctured neighborhood of p is a set of
form U \ {p} where U is a neighborhood of p. A point p is a accumulation
point7 of a set S iff every punctured neighborhood of p contains a point of
S. The following equivalent definition appears in some books.

Proposition 11.2. A point p is an accumulation point of the set S if and

only if every neighborhood of p contains infinitely many points of S.

Proof. If is easy: an infinite set is nonempty and at most one of the points
in an infinite set is p. For only if assume p is an accumulation point of
the set S and choose a neighborhood U of p. By definition there is a point
p0 U S \ {p}. Let 0 = |p p0 |. For n > 0 define n > 0 and pn S
inductively (see 7.2) by n = min(|pn p|, 1/n) and pn+1 B(p, n ) S. The
7
Buck uses the term cluster point and some authors use the term limit point.

24
map n 7 pn is injective as |pn p| < |pm p| for n > m. Choose > 0
so that B(p, ) U (by the definition of neighborhood). Then n < for
1/n < so pn B(p, ) S U S. Hence U contains the infinite set
{pn : n > 1/}.
11.3. Let p0 be a accumulation point of a set S and F be a function defined
on S (but possibly not at p0 ). The notation

lim F (p) = L
pp0

means that for every neighborhood V of L of there is a punctured neighbor-

hood U \ {p} of L such that f (S U \ {p}) V . When p0 S and p0 is a
accumulation point of S we have that a function f defined on S is continuous
at p0 (see Definition 13.1 below) if and only if

lim f (p) = f (p0 )

pp0

(and the function is trivially continuous at a point p0 S which is not an

accumulation point of S). However, the limit notation is usually used in
situations where (p0 is a accumulation point of S but) p0
/ S. For example,
the derivative of a real valued function f : I R defined on an open
interval I R is defined by
f (x) f (x0 )
f 0 (x0 ) := lim .
xx0 x x0
The ratio in the limit is undefined when x = x0 but is defined for nearby
values of x.
11.4. For a real valued function f defined on a subset of R we can extend
the definition of the notation limxa F (x) = L to include the cases where
a = and/or L = as follows. Let
:= {} R {}
R

consist of the set of real numbers together with two additional points which
we think of as located at infinity. The set R is sometimes called the set of
extended real numbers. Extend the usual order relation on R to R in the
obvious way. For a R, a set U R is called neighborhood of a iff

either a R and U contains an open interval (a , a + ) for some > 0,

25
 or else a = and U contains an interval (M, ] for some M > 0,

or else a = and U contains an interval [, M ) for some M > 0.

Because B(a, ) = (a , a + ) this definition agrees with the definition

is called a accumulation point of a subset
in 11.3 for a R. A point a R
S R iff every punctured neighborhood of a intersects S. If f : S R and
a is a accumulation point of S, then the notation

lim F (x) = L
xa

means that for every neighborhood V of L there is a punctured neighborhood

U of a such that f (U S) V .

Remark 11.5. Unraveling the above definitions we see that for a, L R

we have that limxa F (x) = L iff > 0 > 0 such that x S we have
that 0 < |x a| < = |F (x) L| < . Also limx F (x) = L iff
> 0 M > 0 such that x S we have that M < x = |F (x) L| <
with similar definitions for the other cases where a, L {}. The various
definitions given in 11.3 and 11.4 are easier to understand because the lingo
makes them look the same and because there arent so many symbols. This
is why the terminology was invented.

12 Convergence of Sequences
12.1. A sequence is a function defined on a subset of the integers. (Usually
this subset is the set Z+ := {n Z : n > 0} of positive integers or the set
N := {n Z : n 0} of nonnegative inregers. ) It is customary to denote
the value of a sequence at an integer n with a subscript rather than with
parentheses and to denote a sequence with a notation like (pn )n or (pn )nZ+ .

Definition 12.2. The sequence (pn )n of points of Rm is said to converge

to the point p Rm iff
lim pn = p
n

This is sometimes abbreviated as pn p as n . We say a sequence

converges or is convergent iff it converges to p for some p Rm . A
sequence is said to diverge when it does not converge. (A sequence in R
whose limit is infinite is also said to diverge.)

26
Remark 12.3. Using the lingo introduced in 11.4 this may be stated as
limn pn = p iff for every > 0 there exists N = N () > 0 such that
n N = |pn p| < .

Theorem 12.4. Assume that (an )n and (bn )n are convergent sequences of
real numbers:
lim an = a, lim bn = b.
n n

Then
lim an + bn = a + b, lim an bn = ab.
n n

Moreover, if b 6= 0 then bn 6= 0 for sufficiently large n and

an a
lim = .
n bn b
Proof. (I) We prove limn an + bn = a + b. Choose > 0. By hypothesis
there exists N1 and N2 such that

n > N1 = |an a| < , n > N2 = |bn b| < .
2 2
Let N = max(N1 , N2 ). Then for n > N we have

|(an + bn ) (a + b)| = |(an a) + (bn b)|

|an a| + |bn n|

< + =
2 2
as required.
(II) We prove limn an bn = ab. Choose > 0. By hypothesis there exists
N0 ,N1 ,N2 such that

n > N0 = |an a| < 1,

n > N1 = |bn b| < ,
2(|a| + 1)

n > N2 = |an a| < .
2|b|

27
Let N = max(N0 , N1 , N2 ). Then for n > N we have

|an bn ab| = |an (bn b) + (an a)b|

|an | |bn b| + |an a| |b|
< (|a| + 1)|bn b| + |an a| |b|

< (|a| + 1) + |b|
2(|a| + 1) 2|b|

= + = .
2 2
(III) We prove limn 1/bn = 1/b if b 6= 0. Choose > 0. By hypothesis
there exists N1 such that for n > N1 we have that |bn b| < 21 |b| (and hence
that 12 |b| < |bn |) and there exists N2 such that |bn b| < 21 |b|2 . Then for
n > max(N1 , N2 ) we have
 
 1  = |bn b| < |bn b| < .
1 
1
 bn b |bbn | 2
|b|2

(IV) That limn an /bn = a/b follows immediately from (II) and (III) by
substituting 1/bn for bn and 1/b for b

Definition 12.5. A real valued function f defined on a subset of the real

numbers R is called

increasing iff x1 < x2 = f (x1 ) < f (x2 ),

decreasing iff x1 > x2 = f (x1 ) > f (x2 ) ,

monotonic iff it is either increasing or decreasing.

Remark 12.6. If < is replaced by in this definition, the meaning changes:

a constant function satisfies the modified definition. When we want to use
the weaker form we use the term nondecreasing instead of increasing, the
term nonincreasing instead of decreasing, and the term weakly mono-
tonic instead of monotonic. Thus in freshman calculus it is correct to say
that a differentiable function is nondecreasing if and only if its derivative is
everywhere nonnegative but incorrect to say that a differentiable function is
increasing if and only if its derivative is everywhere positive. (If f (x) = x3 ,
then f is increasing but f 0 (0) = 0.) To avoid this confusion some authors
insert the word strictly in 12.5.

28
Theorem 12.7. A bounded weakly monotonic sequence is convergent. In
fact
lim an = sup an
n n

if the sequence (an )n is nondecreasing, and

lim an = inf an
n n

if the sequence {an } is nonincreasing. (See 9.5.)

Proof. (Compare Buck page 47 or Lang page 35 or Morgan page 38.) In this
proof we use the completeness axiom 9.6 for the first time.
Assume that the sequence (an )n is nondecreasing and let a = sup{an :
n N}. Then an a for all n as a is an upperbound for the set {an : n N}.
Choose > 0. Then a < a so a is not an upperbound for the set
{an : n N}. Hence there is an N with a < aN . For n > N we have
aN an as the sequence (an )n is nondecreasing so

a < aN an a < a +

so |an a| < for n > N as required. The nonincreasing case is proved by

an analogous argument, or alternatively by applying the nondecreasing case
to the seqeunce (an )n .

12.8. We introduce some handy notation. For any sequence (ak )k of real
numbers we have {ak : k n} {ak : k m} for m < n. If the se-
quence (ak )k is bounded above, then the sequence sn := sup{ak : k n} is
nonincreasing. The limit of the latter sequence is denoted

lim sup an := lim sup{ak : k n}.

n n

Similarly for a sequence which is bounded below,

lim inf an := lim inf{ak : k n}.

n n

Definition 12.9. When n1 < n2 < n3 < is an increasing sequence of

positive integers, the sequence (pnk )k is called a subsequence of the sequence
(pn )n .

29
Remark 12.10. If the sequence (pn )n converges to p, then every subsequence
(pnk )k also converges to p. This follows immediately from the definition of
convergence: nk k so if N = N () satisfies |pn p| < for n > N () then
in particular we have |pnk p| < for k > N ().

Theorem 12.11 (Bolzano-Weierstrass). Every bounded sequence in Rm

has a convergent subsequence.

Proof. Compare Buck Theorem 22 page 62 or Lang page 38-39 or Mor-

gan page 38. We first do the case m = 1. Let (an )n be a bounded sequence
of real numbers. Then there is a number M such that m an M for all
M . For each n define

An := {an+1 , an+2 , . . . , }, bn := inf An .

Since An An+1 we have that bn bn+1 so the sequence (bn )n converges to its
supremum b := sup{bn } := lim inf n an . We will show that a subsequence of
the sequence (an )n also converges to b. For n N we have that bn < bn + n1
and bn is the greatest lower bound for An so bn + n1 is not a lower bound
for An so there is a cn An with bn cn < bn + n1 . As bn converges to b
by Theorem 12.7 we have that fore every > 0 there is an N = N () such
that |bn b| < /2 for n > N (). Hence for n > max(2/, N () we have that
|cn b| |cn bn | + |b + n b| < n1 + 2 < which shows that (cn )n converges
to bn .
Now cn An so cn = aj for some j = j(n) n + 1, but we arent quite
done because the definition of subsequence requires that the subscripts j(n)
increase and there is no reason for that to be true. However we can extract a
further subsequence by induction. Namely if n1 < n2 < < nk have been
defined, define nk+1 by nk+1 = j(nk ). Then nk+1 = j(nk ) nk + 1 > nk as
required. (The further subsequence still converges by Remark 12.10.
Now we prove the theorem for a sequence of points in Rm by induction on
m. Assume the theorem holds for Rm and choose a bounded sequence (pn )n
of points in Rm+1 . Then pn = (qn , an ) where qn Rm and an R. That
the sequence (pn )n is bounded means that there is an M such that |pn | M
for all n, As |pn |2 = |qn |2 + a2n it follows that |qn | M and |an | M for
all n, i.e. the sequence (qn )n and (an )n are also bounded. By the inductive
hypothesis the sequence (qn )n has a subsequence (qnk )k converging to q. By
replacing the sequence (pn )n by the sequence (pnk )k we may assume that the

30
sequence (qn )n converges. (Remark 12.10.) Now by the case m = 1 (already
proved) the sequence (an )n contains a convergent subsequence (ank )k . Hence

lim qnk = q, lim anl = a.

k k

By the triangle inequality |(q 0 , a0 ) (q, a)| |q 0 a| + |a0 a| so

lim pnk = lim (qnk , ank ) = (q, a)

k k

as required.
Corollary 12.12. For a subset S Rm the following conditions are equiv-
alent.
(1) For every sequence (pn )n of points of S there is a subsequence (pnk )k
which converges to p S.
(2) The set S closed and bounded.
Proof. Assume (1). The certainly every convergent sequence pn S has limit
in S so S is closed by Theorem 14.11. Also S is bounded as otherwise for
every n there would be a point pn S with |pn | > n and this sequence cannot
have a convergent subsequence. Conversely assume (2) and let (pn )n be a
sequence of points in S. As S is bounded there is a convergent subsequence
(by Theorem 12.11) and as S is closed the limit of this subsequence is a point
of S (by Theorem 14.11).
Corollary 12.13. Every bounded infinite subset of Rm has an accumulation
point.
Proof. By Theorem 12.11 and Proposition 11.2.
Remark 12.14. The image8 of the sequence (pn )nN (when viewed as a map
n 7 pn ) is the set
S = {pn : n N}.
The set S can be finite. For example for the sequence pn = (1)n , the
set S is the two element set S = {1, 1}. If the image of a sequence is
finite then there must be at least one constant subsequence and a constant
subsequence is trivially convergent. By definition only an infinite set can
have an accumulation point.
8
Buck calls the set S the trace of the sequence, but that terminology is uncommon.

31
Definition 12.15. A sequence {pn } is called Cauchy iff

lim |pn pm | = 0
m,n

i.e. iff for every > 0 there exists N > 0 such that |pn pm | < for
n, m N .
Theorem 12.16 (Cauchy Convergence Criterion). A sequence in Rn
converges if and only if it is a Cauchy sequence.
Proof. (Buck Theorem 23 and its corollary on pages 62-63.)

13 Continuity
Throughout this chapter f : X Y where X Rn and Y Rm .
Definition 13.1. The map f is said to be continuous at a point p X iff
for every > 0 there exists > 0 such that f (BX (p, )) B(f (p), ).
Theorem 13.2. The map f is continuous at p X if and only if for every
sequence {pn } of points in X we have

lim pn = p = lim f (pn ) = f (p). (1)

n n

Proof. We prove only if . Assume f is continuous at p. Choose sequence

{pn } of points in X. Assume

lim pn = p. (2)
n

Choose > 0. Because f is assumed to be continuous at p there is a > 0

such that for all q X

|q p| < = |f (q) f (p)| < . (3)

By (2) there is an N such that |pn p| < for n > N . Hence by (3)
|f (pn ) f (q)| < for n > N . This proves

lim f (pn ) = f (p). (4)

n

as required.

32
 We prove if . Assume that f is not continuous at p X. Then there is
an > 0 such that for every > 0 there is a q X such that
|q p| < but |f (q) f (p)| .
In particular, for each n Z+ there is a qn such that
1
|qn p| < but |f (qn ) f (p)| .
n
But then (2) holds but (4) fails. This proves that (1) is false as required.
Definition 13.3. The map f is said to be continuous iff it is continuous
at every point of X, i.e. iff
p X > 0 > 0 such that f (B(p, )) B(f (p), ).
The map f is said to be uniformly continuous iff
> 0 > 0 such that p X we have f (B(p, )) B(f (p), ).
(For continuity = (p, ); for uniform continuity = ().)
Proposition 13.4. If f : X Y and g : Y Z are both continuous, then
so is the composition g f : X Z.
Proof. Choose p0 X and > 0. As g is continuous there exists > 0
such that |g(q) g(f (p0 ))| < whenever q Y and |q f (p0 )| < . As f is
continuous, there exists > 0 such that |f (p) f (p0 )| < whenever p X
and |p p0 | < . For p X we have q = f (p) Y so
|p p0 | < = |f (p) = f (p0 )| < = |(g f )(p) (g f )(p0 )| <
as required.
13.5. The map f is said to be Lipschitz iff there is a constant M such that
|f (p) f (q)| M |p q|
for all p, q X. A Lipschitz function is uniformly continuous. (Proof:
= /M .)
Problem 13.6. Let f (x) = xp . Show that f is Lipschitz on every closed
interval [a, b] (0, ). For which values of p is f uniformly continuous
on (0, )? Hint: Use the Mean Value Theorem from calculus. (See Theo-
rem 17.4 below.) Theorem 16.4 below may also help.

33
14 Open Sets and Closed Sets
In all the following definitions the term set means subset of Rm .
Definition 14.1. A set U is open iff for every p U there exists a > 0
such that B(p, ) U .
Problem 14.2. Prove that the ball B(p, ) is open. Hint: You must choose
an arbitrary q B(p, ) and then find an > 0 so that B(q, ) B(p, ).
Use the triangle inequality.
Theorem 14.3. The collection of all open sets in Rm satisfies the following
conditions:
(i) The set Rm and the empty set are both open.
(ii) The intersection of a finite collection of open sets is open.
(ii) The union of an arbitrary collection of open sets is open.
Proof. The set Rm is open because B(p, 1) Rm for p Rm . The empty set
is open because for every p satisfies the required condition or any other
condition since false implies anything is true. See 3.3). To prove (ii)
assume U is open. Then for every point p U there is a = p such that
B(p, p ) U .9 It follows that
[
U= B(p, p ), ()
pU

i.e. that U is an union of balls. A union of unions is a union:

[[ [
Bij = Bij , K := {(i, j) : i I, j Ij }
iI jIj (i,j)K

so (ii) follows. To prove (iii)

T assume that U1 , U2 , . . . Um are open and choose
an arbitrary point p m i=1 Ui . Then p Ui so there is a i > 0 with
B(p, i ) Ui . Let = min(1 , . . . , m ). Then
m
\ m
\
B(p, ) B(p, i ) Ui
i=1 i=1

as required.
9
This is actually an example of an application of the Axiom of Choice. See Remark 6.5.

34
Definition 14.4. A set W X is called relatively open in X iff for every
p W there exists a > 0 such that BX (p, ) W . (See 10.4.)

Remark 14.5. A set U Rm is open if and only if it is relatively open

in Rm . For this reason many theorems can be generalized by systematically
replacing Rm by X, B(p, ) by BX (p, ), and the word open by the phrase
relatively open in X. In Math 551 you will learn to generalize further by
replacing Rn by something called a topological space.

Corollary 14.6. A set W is relatively open in X if and only if W = X U

for some open set U Rn .

Proof. Equation () in the last proof and the distributive law from 4.7.

Proposition 14.7. A map f : X Y is continuous if and only if the inverse

image f 1 (V ) of every relatively open subset V of Y is a relatively open subset
of X.

Proof. Exercise.

Corollary 14.8. Assume that f : X Y and g : Y Z are continuous.

Then g f : X Z is continuous.

Proof. Exercise. Hint: Prove that (g f )1 (W ) = f 1 g 1 (W ) .

Definition 14.9. A set X is closed iff its complement Rn \ X is open.

Corollary 14.10. The collection of all closed sets in Rn satisfies the follow-
ing conditions:

(i) The set Rn and the empty set of both closed.

(ii) The intersection of an arbitrary collection of closed sets is closed.

(iii) The union of a finite collection of closed sets is closed.

Proof. Apply the De Morgan laws from 4.7 to Theorem 14.3.

Theorem 14.11. A set S is closed if and only if it is closed under limits of

sequences, i.e. whenever limn pn = p and each pn S we have p S.

35
Proof. (See Buck Theorem 5 page 40.) To prove only if assume that S is
closed, that limn pn = p, and that each pn S. If p / S then p Rm \ S.
As this set is open there is a > 0 such that B(p, ) Rm \ S. As the
sequence converges to p there is an N such that pn B(p, ) for n > N
contradicting the hypothesis that pn S. To prove if assume that S is not
closed. Then Rm \ S is not open so there is a point p Rm \ S such that
B(p, ) 6 Rm \ S for every > 0. In particular for = 1/n there is a point
/ Rm \ S, i.e. pn S. Thus
pn B(p, 1/n) (i.e. |pn p| < 1/n) such that pn
limn pn = p and p / S as desired.
14.12. Let S Rn . For any point p Rn exactly one of the following
alternatives holds:
(i) B(p, ) S for some > 0.

(ii) B(p, ) Rn \ S for some > 0.

(iii) B(p, ) S 6= and B(p, ) (Rn \ S) 6= for all > 0.

The interior of S is the set int(S) of all points p where (i) holds, the exterior
of S is the set ext(S) of all points p where (ii) holds, and the boundary of
a set S is the set bdry(S) of all points p where (ii) holds. The ambient space
Rn may be written as the pairwise disjoint union

Rn = int(S) ext(S) bdry(S).

The notations
S := int(S), S := bdry(S)
are commonly used.
Example 14.13. For the half open interval S = [a, b) R we have

int(S) = (a, b), ext(S) = (, a) (b, ), bdry(S) = {a, b}.

Definition 14.14. A set S Rn is closed iff its complement Rn \ S is open.

The closure of the set S is the set

cl(S) := S := S bdry(S).

Proposition 14.15. The interior int(S) of S is the largest open set contained
in S and closure S of S is the smallest closed set containing S.

36
Proof. Exercise. Hint: You must show (1a) int(S) is open, (1b) S is closed,
(2a) int(S) S, (2b) S S, (3a) if U is open and U S then U int(S),
and (3b) if T is closed and S T then S T .
Problem 14.16. Prove that if U Rm is an open set and p U , then there
is a point q with rational coordinates and a positive rational number such
that p B(q, ) U .
Solution. As U is open there is an r > 0 such that B(p, r) U . Let pi
denote the ith coordinate of p so that p = (p1 , p2 , . . . , pm ). By Problem 9.10
there are rational numbers q1 , q2 , . . . , qm such that
r r
pi < qi < pi + .
2m 2m
Let q = (q1 , q1 , . . . , qm ). By the Triangle Inequality |p q| < r/2. By
Problem 9.10 there is rational number with |p q| < < r/2. Then
p B(q, ). Choose x B(q, ). Then |p x| < |p q| + |q x| < r so
x B(p, r) U . This shows that B(q, ) U as required.

15 Connected Sets
Definition 15.1. A set S is disconnected iff there are disjoint open sets
U and V such that S U V and both S U and S V are nonempty. A
set is connected iff it is not disconnected.
Theorem 15.2. A subset S R of the real line is connected if and only if
S is an interval, i.e. [a, b] S whenever a, b S.
Proof. We prove only if. Assume S is not an interval, i.e. that there exist
a, b S with [a, b] 6 S. Then there is a c [a, b] with c / S. Let U =
(, c) and V = (c, ). The point c lies in the open interval (a, b) as
a, b S so a U and b V . Hence both S U and S V are nonempty
and clearly S U V (as c / S). Hence the open sets U and V separate S
so S is disconnected as required.
We prove if. Assume that S is disconnected, i.e. that there exist open
sets U, V R with S U V , S U 6= , S V 6= , and U V = . We
must show that S is not an interval. Choose a S U and b S V . Then
a 6= b as U V = . Assume without loss of generality that a < b. (The case
b < a is the same.)

37
 The set [a, b] U is nonempty (it contains a) and bounded above (b is
an upper bound). Let c = sup([a, b] U ). Since a U there is an > 0
with (a , a + ) U . Making smaller we also have a + < b. Therefore
[a, a + ) [a, b] U ) so a + = sup [a, a + ) sup [a, b] U = c. Since
b V there is an(other) > 0 with (b , b + ) V . Making smaller we
also have a < b . Therefore (b , b] [a, b] V ) so [b , b] V =
so b is an upperbound for [a, b] U , so c b . We have proved that
a < c < b. If c U there is an > 0 with a < c < c < c + < b and
(c , c + ) U contradicting the fact that c is an upper bound of [a, b] U .
If c V there is an > 0 with a < c < c < c + < b and (c , c + ) V
so c is an upperbound for [a, b] U contradicting the fact that c is the
least upper bound of [a, b] U . Hence c / U V so (as S U V ) c / S.
Thus a < c < b, a S, b S, c / S, so S is not an interval.

Theorem 15.3. The continuous image of a connected set is connected: If

f : X Rm is continuous and X is connected, then f (X) is connected.

Proof. (Buck Theorem 15 on page 94.)

Corollary 15.4 (Intermediate Value Theorem). Assume that S is con-

nected and that f : S R is continuous. Suppose that a, b f (S) and that
a < c < b. Then c f (S).

Proof. (Buck Theorem 14 on page 93.)

Remark 15.5. The Intermediate Value Theorem from calculus is a special

case. It says that if f : [, ] R is a real valued continuous function on
the closed interval [, ] R, {a, b} = {f (), f ()}, and a c b, then the
equation f (x) = c has a solution x [, ].

Theorem 15.6. A continuous function f : I R defined on an interval I

R is injective if and only if it is strictly monotonic. When these equivalent
conditions hold, the image J = f (I) is again an interval and the inverse
function is continuous.

Problem 15.7. Prove Theorem 15.6. (This theorem is proved in Buck The-
orem 18 page 96 and Theorem 25 page 114, but Buck assumes that the
intervals are closed and bounded. This assumption can be removed.)

38
Theorem 15.8. Let I R be an interval and f : I R be f is continuous.
Then the set
graph(f ) := {(x, y) I R : y = f (x)}
is connected.
Proof. Define F : I R by F (x) = (x, f (x)) so that F (I) = graph(f ).
Clearly f is continuous if and only if F is continuous. We will assume
that I is an open interval; the case where I contains one of its endpoints
is similar. Assume that F (I) is not connected. Then there are open sets
U, V R2 with F (I) U V , U V = , F (I) U 6= , F (I) V 6=
. Then F 1 (U ), F 1 (V ) R2 are open, I F 1 (U ) F 1 (V ), and
F 1 (U ) F 1 (V ) = F 1 (U V ) = . This contradicts the fact that I
is an interval and therefore connected.
Example 15.9. The converse is false. Consider the function f : R R
defined by 
sin(1/x) if x > 0,
f (x) =
0 if x 0.
This function is not continuous as follows. Let xn = (2n + /2)1 . Then
f (xn ) = 1, limn xn = 0, but limn f (xn ) = 1 6= 0 = f (0). However, the
graph of f is connected. To see this suppose U and V are open subsets of
R2 and graph(f ) U V with U V = . Suppose that (0, 0) U . Then
(x, f (x)) U for x 0 as f is continuous on (, 0] and (x, f (x)) U
in U for x > 0 as f is continuous on (0, ). But then graph(f ) U so
graph(f ) V = .

16 Compact Sets
Definition 16.1. AnS open cover of a set S is a collection (U )A of open
sets such that S A U . The subset S is compact iff every open cover
(U )A of S has finite subcover, i.e. there are indices 1 , 2 , . . . , n A
such that S U1 U2 Un .
Theorem 16.2 (Heine Borel). The following are equivalent conditions on
a set S Rm :
(1) For every sequence (pn )n of points of S there is a subsequence (pnk )k
which converges to p S.

39
(2) The set S closed and bounded.
(3) The set S is compact.
Proof. (This is Theorem 9.2 on page 41 of Morgan. See also Buck Theorem 25
page 65.) The equivalence of (1) and (2) is Corollary 12.12.
We prove (3) = (2). Assume that (2) is false. Then either S is not
closed or S is not bounded. In the former case by Theorem 14.11 there is a
convergent sequence of points pn S S whose limit p := limnmpn is not in
S. This implies S Rm \ {p} = k=1 Uk where Uk := {q R : |q p| >
1/k}. The sets Uk are open but we cannot have S U U2 UN as as
U1 U2 UN and pn / UN for n > N so we have an open cover of S
with no finite subcover so S is not compact. In the latterScase the open balls
Vn := B(0, n) := {q Rm : |q| < n} cover S as Rm = n Vn but no finite
collection subcollection covers so again S is not compact.
We prove (1) = (3). Choose an open cover (U )A of S. We first
construct a countable subcover. Consider the set
I := {(q, ) Qm Q : > 0 and B(q, ) U for some A}
of open balls with rational center and rational radius which are contained in
some element of the open cover. For every point p S there is an A
with p U (because the sets U cover S), and so by Problem 14.16 there
is a point q Qm so that p B(q, ) U . In other words there is a point
(q, ) I with p B(q, ). We have proved that
[
S B(q, ).
(q,)I

The set I is countable because I is a subset of the countable set Qm Q. Let

V1 , V2 , V3 , . . . be an enumeration of the sets (B(q, ))(q,)I . By construction
each Vi is a subset of some U . Hence if finitely many of the sets Vi cover S so
do finitely many of the sets U from the original cover. But (1) implies that
finitely many of the sets Vi cover S. If not, we have S 6 V1 V2 Vn
for all n so for each n there is a point pn S \ (V1 V2 Vn ). By
condition (1) the sequence (pn )n contains a subsequence (pnk )k converging to
a point p S. The point p lies in some set Vi and this set must contain pnk
for sufficiently large k which is a contradiction if nk > i.
Corollary 16.3. The closed interval [a, b] is compact. (Buck Theorem 24
page 65.)

40
Theorem 16.4. Assume that X is compact and f is continuous. Then f is
uniformly continuous.

Proof. Choose > 0. Then, because f is continuous, for every p X there

is a = (p) > 0 such that

|q p| < (p) = |f (q) f (p)| < .
2
S
Let Up := B(p, (p)/2). The sets Up cover X (since p Up ), i.e. X p Up .
As X is compact, finitely many of these sets cover X, i.e.

X Up1 Up2 Upn . (5)

Define
:= 12 min{(p1 ), (p2 ), . . . , (pn )}.
Choose p, q X. Assume |q p| < . By (5) we have that p Upk for some
k. Hence
|p pk | < (pk )/2. (6)
But (pk )/2 by its definition so

(pk ) (pk ) (pk )

|q pk | |q p| + |p pk | + + = (pk ). (7)
2 2 2
Hence, by the definition of () and Equations (6) and (7) we have

|f (p) f (q)| |f (p) f (pk )| + |f (q) f (pk )| < + =
2 2
as required.

Remark 16.5. A proof using the Bolzano-Weierstrass theorem instead on

the Heine-Borel theorem is given in Buck page 85. The proof given here is
like the proof sketched in Exercises 11 and 12 in Buck page 89.

Theorem 16.6. The continuous image of a compact set is compact: If f :

X Rm is continuous and X is compact, then f (X) is compact.

Proof. See Buck Theorem 13 on page 93 or

Corollary 16.7. The continuous image of compact set is bounded.

41
Proof. See Buck Theorem 10 on page 90 or
Theorem 16.8. If f : X R is continuous and X is compact, then f
assumes its maximum on X, i.e. there exists p X such that f (q) f (p)
for all q X. Similarly for the minimum.
Proof. See Buck Theorem 11 page 91 or
Theorem 16.9. Let f : X Y be bijective and continuous, and assume
X (and hence by Theorem 16.6 also Y ) is compact. Then f 1 : Y X is
continuous.
Proof. Choose a convergent sequence (qn )n in E and a let
q := lim qn
n

be its limit. We will show that

f 1 (q) = lim f 1 (qn ); (#)
n

the theorem will then follow by Theorem 13.2. By Bolzano Weierstrass and
Heine Borel there is a convergent subsequence (f 1 (qnk ))k . Let
p := lim f 1 (qnk )
k

be its limit. Now f is assumed to be continuous so

 
f (p) = f lim f 1 (qnk ) = lim f (f 1 ((qnk )) = lim qnk = q.
k k k

But f (p) = q = p = f 1 (q) so f 1 (q) = limk f 1 ((qnk ). If (#) fails,

then there is a neighborhood U of f 1 (q) such that for every N there ex-
ists n > N with f 1 (qn )
/ U , i.e. there is a subsequence f 1 (qmj ) with
f 1 (qmj )
/ U . As before choose a further subsequence (again denoted
1
f (qmj ) which converges and let
p0 := lim f 1 (qmj )
j

denote the limit. Then p0 / U (else we would have f 1 (qmj ) U for suffi-
ciently large j) so p0 6= p. But as before
 
f (p ) = f lim f (qmj ) = lim f (f 1 (qmj )) = lim qmj = q.
0 1
j j j

But now f (p) = q = f (p0 ) which contradicts the fact that f is injective.

42
Theorem 16.10. Let S Rn be and f : S Rm be uniformly continu-
ous.Then the function f can be continuously extended to the closure S of S.
i.e. there is a continuous function F : S Rm such that F (p) = f (p) for
p S.

Proof. (Buck Theorem 25 on page 109.)

Example 16.11. The function f : (0, 1] R defined by f (x) = sin(1/x)

cannot be extended to a continuous function on the closure [0, 1] of (0, 1].

17 Derivatives
Definition 17.1. The function f : I R is said to be differentiable at
the point x0 I iff the limit

f (x) f (x0 )
f 0 (x0 ) := lim
xx0 x x0
exists; we say that f is differentiable on a set iff it is differentiable at each
point x0 in the set. The function f 0 is called the derivative of f .

Theorem 17.2. A differentiable function is continuous.

Theorem 17.3. If I is an open interval, f : I R is differentiale on I,

and f assumes its maximum (or minimum) at c I then f 0 (c) = 0.

Theorem 17.4 (Mean Value Theorem). Suppose that f is differentiable

of (a, b) and continuous on [a, b]. The there exists c (a, b) such that

f (b) f (a)
f 0 (c) = .
ba
Corollary 17.5. Assume that f is differentiable on I. Then the derivative
f 0 vanishes identically on I if and only if f is constant on I.

Problem 17.6. The Intermediate Value Theorem for Derivatives. Assume

that f is differentiable on an open interval I, that a, b I with a < b, and
that f 0 (a) < w < f 0 (b). Show that there is a c with a < c < b and f 0 (c) = w.
Hint: Consider the function g(x) = f (x) wx.

43
18 The Integral
18.1. A partition of the closed interval [a, b] is an increasing finite sequence
P = (xk )0kn with x0 = a and xn = b. For any bounded function f defined
on [a, b] and any partition P = (xk )0kn of [a, b] define the upper sum by
S(f, P ) and the lower sum S(f, P ) by
n
X n
X
S(f, P ) := y k (xk xk1 ), S(f, P ) := y k (xk xk1 ),
k=1 k=1

where we have used the abbreviations

y k := sup{f (x) : xk1 x xk }, y k := inf{f (x) : xk1 x xk },

Theorem 18.2. Assume that f : [a, b] R is continuous. Then there is a

Rb
unique number a f called the definite integral of f on the interval [a, b]
such that the inequality
Z b
S(f, P ) f S(f, P )
a

holds for every partition P of the interval [a, b].

Proof. We say that a partition P of [a, b] refines the partition Q of [a, b]

iff Q is a subsequence of P , i.e. iff P = (xk )0kn , Q = (xkj )0jm where
0 = k0 < k1 < < km = n. The mesh of a partition P = (xk )0kn to be
the maximum of the positive numbers k := xk xk1 , k = 1, 2, . . . , n. We
prove the theorem in five steps.
Step 1. For any partition P of [a, b] we have

y (b a) S(f, P ) S(f, P ) y (b a)

where y is the infimum of f (x) for x [a, b] and y is the supremum. The
middle inequality follows the inequality y k y k which in turn is an immediate
consequence of the fact that the infimum of a bounded nonempty set is less
than or equal to the supremum of that set. For the inequality on the left
note that
y (xk xk1 ) y k (xk xk1 )

44
since the infimum on the left is over a larger set. From the collapsing sum

b a = (x1 x0 ) + (x2 x1 ) + + (xk xk1 )

we obtain the inequality y (ba) S(y, P ) by summing on k. The inequality

The inequality S(f, P ) y (b a) is proved similarly.
Step 2. If the partition P of [a, b] refines the partition Q of [a, b], then

S(f, Q) S(f, P ) S(f, P ) S(f, Q).

Since xkj1 < xkj1 +1 < xkj1 +2 < < xkj the partition P determines a
partition Pj of the interval [xkj1 , xkj ]. Applying Step 1 to this partition gives

y k (xkj xkj1 ) S(f, Pj ) S(f, Pj ) y (xkj xkj1 ).

j

Step 2 follows by summing on j.

Step 3. For every > 0 there is a > 0 such that for every partition P
whose mesh is less than we have

S(f, P ) < S(f, P ) + .

To see this choose > 0. By Theorem 16.4 there is a > 0 such that
|f (x) f (x0 )| < /(b a) whenever |x x0 | < . If the partition P has mesh
less than it follows that y k y k + /(b a) for k = 1, 2, . . . , n. The desired
inequality follows by multiplying by xk xk1 and summing on k. (Use the
collapsing sum from Step 1.)
Step 4. Given any two partitions of the interval [a, b] there is a third partition
which refines both of them. To construct this third partition we simply take
the union of the underlying sets of the two partitions and list the elements
of this union in increasing order.
Step 6. S(f, P 0 ) S(f, P 00 ) for any two partitions P 0 and P 00 of the interval
[a, b]. This is because by Step 4 the partitions P 0 and P 00 have a common
refinement P so Step 2 we have

S(f, P 0 ) S(f, P ) S(f, P ) S(f, P 00 ).

Step 7. supP S(f, P ) inf P S(f, P ). To see this take the supremum over
P 0 is Step 4 to get supP S(f, P ) S(f, P 00 ) for every P 00 . Then take the
infimum over P 00 .

45
Step 8. inf P S(f, P ) supP S(f, P 0 ). To see this choose > 0 and let > 0
be given as in Step 3. There certainly are partitions with mesh less than ,
for example the partition Pn = (xk )k with xk = a + k(b a)/n where n is so
large that (b a)/n < . From Step 3 we conclude that

inf S(f, P ) S(f, Pn ) S(f, Pn ) + sup S(f, P ) + .

P P

Step 7 follows since this inequality is true for any > 0.

By Steps 6 and 7 we have that supP S(f, P ) = inf P S(f, P ) This common
Rb
value satisfies the conclusion of the theorem and is the definite integral a f .

Remark 18.3. A Riemann sum for a partition P = (xk )k of the interval

[a, b] is a sum of form
n
X
S= f (ck )(xk xk1 )
k=1

where ck [xk1 , xk ]. Since f k f (ck ) f k it follows that

S(f, P ) S S(f, P )

for any Riemann sum for the partition P . It follows from the proof of Theo-
rem 18.2 that the definite integral is the limit of the Riemann sums S as the
mesh of the partition P tends to zero, i.e. for every > 0 there exists a > 0
R b 
such that S a f  < whenever the mesh of P is less than . In Math 221

the definite integral is usually defined as the limit of Riemann sums in this
sense.
Theorem 18.4. The definite integral of a continuous function satisfies the
following properties.
Z b
(1) (Normalization). 1 = b a.
a
Z b Z b Z b
(2) (Linearity). (f + g) = f+ g
a a a
Z b Z b
(3) (Linearity). cf = c f for c R.
a a

46
 Z b Z c Z c
(4) (Additivity). f+ f= f.
a b a
Z b Z b
(5) (Order). If f (x) g(x) for all x [a, b], then f g.
a a
Z b  Z b
 
(6) (Triangle Inequality).  f  |f |.
a a

In (3) c is any real number, and in (4) a, b, c I satisfy a b c.10

Proof. These properties hold because Riemann sums satisfy similar proper-
ties:
X
(1) k = b a.
k=1

n
X n
X n
X
(2) (f (ck ) + g(ck ))k = f (ck )k + g(ck )k
k=1 k=1 k=1

n
X n
X
(3) cf (ck )k = c f (ck )k
k=1 k=1

n
X m
X m
X
(4) f (ck )k + f (ck )k = f (ck )k
k=1 k=n+1 k=1

n
X n
X
(5) f (ck )k g(ck )k if f (ck ) g(ck )
k=1 k=1
 n  n
X  X
(6) f (c ) |f (ck )|k
 
 k k
 
k=1 k=1

In these formulas (xk )nk=0 is a partition of [a, b] = [x0 , xn ] (in (4) it is extended
to a partition (xk )m
k=0 of [a, c]), ck is in the kth interval (i.e. xk1 ck xk ),
and we have used the abbreviation

k := xk xk1 .
10
Rb Ra
It is customary to define a f = b f if b < a. With this definition the additivity
formula holds without the restriction that a b c.

47
These formulas and Remark 18.3 imply the theorem by taking the limit as
the mesh of the partition goes to zero.
Theorem 18.5 (Fundamental Theorem of Calculus). Assume that I is
an open interval, that f : I R is continuous, that a I, and that F (x) is
defined by Z x
F (x) := f.
a
Then
(I) F is differentiable on I and its derivative is f and hence
Z b
(II) f = F (b) F (a) for a, b I with b a.
a
Proof. We must show that for every > 0 there is a > 0 such that
F (x + h) F (x)
< f (x) <
h
whenever 0 < |h| < . We will only use the properties in Theorem 18.4 to
prove this. Assume h > 0 (the case h < 0 is similar). Then
Z x+h Z x Z x+h
F (x + h) F (x) = f f= f
a a x

by property (4). Choose > 0. By Theorem 16.4 f is uniformly continuous

so there is a > 0 such that 0 < h < implies that
f (x) < f (t) < f (x) +
for all t in the interval [x, x + h]. This implies that
Z x+h
(f (x) )h < f < (f (x) + )h
x

by (1), (3), and (5). Dividing by h and subtracting f (x) and gives
 Z x+h 
1
< f f (x) < .
h x
whenever 0 < h < , i.e.
F (x + h) F (x)
< f (x) <
h
as required.

48
 Rb
Remark 18.6. Henceforth we use the more traditional notation a f (x) dx
Rb
(rather than the notation a f ) for the integral. The reader is reminded that
the variable x in this expression is a dummy variable, i.e.
Z b Z b
f (x) dx = f (t) dt.
a a

19 Taylors Formula
Theorem 19.1 (Taylors Formula Lagrange Form). Let I be an in-
terval, a I, and f : I Rm be of class C n+1 . Then

f (x) = Pn (a, x) + Rn (a, x)

for any x I where

n x
f (k) (a) f (n+1) (t)
X Z
k
Pn (a, x) := (x a) , Rn (a, x) := (x t)n dt.
k=0
k! a n!

The polynomial Pn (a, x) is called the Taylor polynomial of degree n at a

and Rn (a, x) is called the nth Taylor remainder.
Proof. By induction on n. For n = 0 this is the Fundamental Theorem of
Calculus. We assume the formula for n and integrate by parts, viewing x as
a constant and t as a variable:
(x t)n+1
v= , u = f (n+1) (t),
(n + 1)!
(x t)n
dv = dt, du = f (n+2) (t) dt,
n!
Z x x Z x

Rn (a, x) = u dv = uv  v du
a a a
(n+1) Z x
f (a) n+1 (x t)n+1 (n+2)
= (x a) + f (t) dt
(n + 1)! a (n + 1)!
f (n+1) (a)
= (x a)n+1 + Rn+1 (a, x).
(n + 1)!
Adding Pn (a, c) to both sides gives f (x) = Pn+1 (a, x) + Rn+1 (a, x).

49
Corollary 19.2. If |f (n+1) (t)| M for t I then
M |x a|n+1
|Rn (a, x)|
(n + 1)!
for x I.
Proof. Assume that x > a. (The case x < a is similar.)
Z x (n+1) 
 f (t) n

|Rn (a, x)| = 
 (x t) dt
a n!
Z x  (n+1) 
f (t) n

 n! (x t)  dt
 
a
Z x
(x t)n (x a)n+1
M =M .
a n! (n + 1)!

Remark 19.3. In Math 222 it is shown that that there is a number c between
a and x such that
f (n+1) (c)
Rn (a, x) = (x a)n+1 .
(n + 1)!
This form of the remainder has the advantage that it is easy to remember: the
remainder is the next term in the series with f (n+1) (a) replaced by f (n+1) (c).
However, this version of the theorem only holds when f is real valued, i.e.
when m = 1. For m > 1 there will be a different value of c for each component
of f .

20 Series
20.1. A sequence determines a series and a series determines a sequence.
More precisely, a sequence (ak )k determines a series whose partial sums
are n
X
Sn = ak := a1 + a2 + + an ,
k=1

and the terms of the series may be recovered from the sequence of partial
sums via the formula

an = Sn Sn1 , a1 = S1 .

50
Convergence of the series is synonymous with convergence of the sequence of
partial sums:
X n
X
ak := lim ak .
n
k=1 k=1

Since the limit of a difference is the difference of the limits and the limit of
a constant is the constant we have the useful formula

X n
X
X
ak ak = ak .
k=1 k=1 k=n+1

Here the notation on the right has the obvious definition, namely

X m
X
ak := lim ak .
m
k=n+1 k=n+1

20.2. The notation

X
ak =
k=1

means
Pn that for every M > 0 there exists an integer N > 0 such that
k=1 k > M for n > N . If ak 0 for all k then the sequence ofP
a partial
sums is monotonic increasing so by Theorem 12.7 either Pthe limit nk=1 ak
exists (i.e. the sequence of partial sums is bounded) or
k=1 = (i.e. the
sequence of partial sums is unbounded).
Example 20.3. The nth partial sum
n
X
xk = 1 + x + x2 + + xn
k=0

of the geometric series is easy to compute:

n
X n
X
k
(1 x) x = (xk xk+1 ) = 1 xn+1
k=0 k=0

(as the sum telescopes) so dividing by (1 x) gives

n
X 1 xn+1
xk =
k=0
1x

51
Hence if |x| < 1 we have the formula

X 1
xk =
k=0
1x

for the infinite sum.

P
Theorem 20.4 (nth Term Test). If the series k ak converges, then the
nth term converges to zero.

Proof. (See Buck Theorem 2 page 230.) Let Sn = nk=0 ak be the nth partial
P
sum so an = Sn Sn1 . If the series converges then limn Sn = limn Sn1
so limn an = limn Sn limn Sn1 = 0.

Example 20.5. (Harmonic series) Since

   
X 1 1 1 1 1
=1+ + + + +
k=1
k 2 10 11 100
P 1
and each sum in parentheses is 9/10 we see that k=1 k
= . But
limn 1/n = 0 so the converse to Theorem 20.4 is false.

Theorem
P 20.6 (Cauchy Convergence Criterion for Series). A series
k ak converges if and only if

n
X
lim ak = 0,
m,n
k=m+1

i.e. for every > 0 there exists N = N () such that  nk=m+1 ak  < for
P 

n > m > N.

Proof. This follows immediately from the formulas

n
X n
X n
X
ak = Sn Sm , Sn := ak , ak := am+1 + + an .
k=m+1 k=1 k=m+1

and the Cauchy Convergence Criterion for Sequences (Theorem 12.16).

52
 P
Definition 20.7. The series k ak is said to converge absolutely iff

X
|ak | < .
k=1

A series which converges but does not converge absolutely is said to converge
conditionally.
Theorem 20.8. If a series converges absolutely, then it converges.
Proof. This is an immediate consequence of the inequality
 
 X n  Xn
ak  |ak |
 

 
k=m+1 k=m+1

and Theorem 20.6.

Remark 20.9. Of course, if ak 0 then |ak | = ak so convergence and
absolute convergence coincide in this case. When ak 0 the sequence of
partial sums is monotonic increasing so either the series converges or diverges
to infinity.
Corollary 20.10. Assume that a0 a1 a2 0. Then the series
k
P
k (1) ak converges if and only if limn an = 0. (A series like this whose
terms alternate sign is called an alternating series.)
Proof. Only if follows by the nth Term Test (Theorem 20.4). For if
assume that limn an = 0 and let
n
X
Sn := (1)k ak
k=0

denote the nth partial sum. Then the sequence

m
X
S2m+1 = (a2k a2k+1 )
k=0

is monotonic non decreasing (since (a2k a2k+1 0) so it either converges

or tends to infinity. But the latter cannot happen as
m1
X
S2m+1 = a0 (a2k+1 a2k+2 ) a2m+1 a0 a2m+1
k=0

53
and the sequence (a2m+1 )m is bounded (as it converges to zero). Hence the
sequence (S2m+1 )m converges by Theorem 12.7. But the sequence (S2m )m
converges to the same limit as S2m+1 = S2m + a2m+1 and we have assumed
that limn an = 0. Hence the sequence (Sn )n converges as claimed.
Example 20.11. (Alternating harmonic series) We will see later (see
Remark 24.3) that

X (1)k
ln 2 = .
k=1
k
The convergence is conditional but (by Example 20.5) not absolute.
Theorem 20.12P (Comparison Test). If 0 akP bk for sufficiently large
k and the series k bk converges, then the series k ak does.
Proof. (Buck Theorem 5 page 231.) The partial sums satisfy the inequality
n
X n
X
0 ak bk .
k=m+1 k=m+1
P Pn
Because the k bP
k is assumed to converge we have lim m,n k=m+1 bk = 0.
n P
Hence limm,n k=m+1 ak = 0 so the series k ak converges by Theo-
rem 12.16

R f : [1, )
Theorem 20.13 (Integral Test). Assume that ak = f (k) where
[0, ) is monotonic decreasing. Then
P the improper integral 1 f (x) dx <
converges if and only if the sum k=1 ak < does.

Proof. (Buck Theorem 10 page 233.) We have ak+1 = f (k + 1) f (x)

f (k) = ak for k x k + 1 so
n+1
X Z n n
X
ak f (x) dx ak .
k=2 1 k=1

This inequality shows that the infinite sum is finite if and only if the integral
is finite.
Theorem 20.14 (Root Test). Let
R = lim sup |ak |1/k
k
P
Then the series k ak converges absolutely if R < 1 and diverges (does not
converge) if R > 1.

54
 
Proof. (Buck Theorem 9 page 232.) Let Sn = |ak |1/k : k = n, n + 1, . . . ,
sn = sup Sn so Sn+1 Sn and hence sn+1 sn . (We are using the convention
that sup S if S is not bounded above.) Either sn = for all n (this falls
under the case R > 1) or else (sn )n is a monotonic sequence decreasing
(i.e non-increasing) converging to R. If R < 1 then R < r < 1 where
r = (1 + R)/2 so there is an N such that sn < r for n > N and hence
|ak |1/k sn < r for k n > N . From this we deduce that |ak | rk and
hence
X X
|ak | rk
k=N +1 k=N +1

so the series convergence by the Comparison Test (Theorem 20.12, see also)
and the fact that the geometric series converges for r < 1. Conversely if
R > 1 then sn R > 1 for all n so for every n there exists a k n with

k
|ak | = |ak |1/k 1. This means that it is not the case that limn an = 0
so the series diverges by the nth term test (Theorem 20.4).
Remark 20.15. Theorem 20.14 gives no information when R =P1. If ak =
1/k, then R = 1 but k ak = . If ak = 1/k 2 , then R = 1 and k ak < .
P

Remark 20.16. The proofs the convergence tests tell us how to estimate
the error, i.e. the difference between a partial
P sum and the infinite sum. For
2
example, by the Integral Test, the series k 1/k converges and
    Z
n
X 1 X 1   X 1  dx 1
= = .

k 2 2 2
k  k=n+1 k  x 2 n
   

k=1 k=1 n

Similarly if R < 1 in the Root Test and R < r < 1 there is an N such that
|ak | < rk for k > N and hence
 n

X X  X X
k rn+1 rN
a a = |a | r =
 
k k k
1r 1r

 
k=1 k=1 k=n+1 k=n+1

for n > N .
Problem 20.17. Show that for p > 1 the series p
P
k=1 k converges and
that the estimate
 n

X X  (n + 1)1p
p p 
k k 

p1

 
k=1 k=1

holds for the difference (error) between the nth partial sum and the limit.

55
 P
Definition
P 20.18. A series k=1 bk is 11 said to be a rearrangement of the
series k=1 ak iff there is a permutation : Z+ Z+ such that bk = a(k) .

Theorem 20.19. (1) Any rearrangement of an absolutely convergent series

converges absolutely to the same limit.

(2) Assume that the series

Pn
L R
k=1 ak converges
Pconditionally and P
n
(see 11.4). Then there is a rearrangement k=1 bk of the series k=1 ak
such that
X
bk = L.
k=1

Proof. Part (1) is Theorem 13 page 239 of Buck. Part (2) is proved on
pages 238-9 of Buck in the special case where an = (1)n /n and L = 10; the
general argument is much the same but uses Theorem 20.4.

21 Uniform Convergence
Definition 21.1. A sequence (fn )n of functions with common domain X
said to converge pointwise to the function f : X R iff

lim fn (p) = f (p)

n

for all p X, i.e. iff

 
p X > 0 N n n > N = |fn (p) f (p)| < .

The sequence is said to converge uniformly to the function f iff

lim sup |fn (p) f (p)| = 0,

n pU

i.e. iff  
> 0 N p U n n > N = |fn (p) f (p)| < .
For a sequence (uk : X Rm )k of functions the series k uk ofPfunctions
P
is said to converge pointwise or uniformly iff the sequence fn = nk=0 uk of
partial sum does.
11
A permutation is bijective map from a set itself. This terminology is most often
used for finite sets, but here it is used for a infinite set.

56
Example 21.2. Define fn : [0, 1] R by fn (x) = xn . Then the sequence
(fn )n converges pointwise but not uniformly to the function

0 for 0 x < 1
f (x) =
1 for x = 1.

Theorem 21.3. If fn : X Y is continuous for each n and the sequence

(fn )n converges uniformly to f : X Y , then the limit f is also continuous.
Proof. (Buck Theorem 3 page 266.) Choose x0 and > 0. By uniform
convergence there exists an n such that |f (x) fn (x)| < /3 for all x and all
n > N . Let n = N + 1. As fn is continuous at x0 there exists > 0 such
that fn (x) fn (x0 )| < /3 whenever |x x0 | < . Then

|f (x) f (x0 )| |f (x) fn (x)| + |fn (x) fn (x0 )| + |fn (x0 ) f (x0 )| <

whenever |x x0 | < .
Theorem 21.4 (Weierstrass Comparison Test). Assume that the func-
m
P
tions
P u k : X R satisfy |u k (p)| Mk where k Mk < . Then the series

k uk converges uniformly.

Proof. (Buck Theorem 2 page 266.) This is an immediate consequence of the

inequality  
 X  X X
uk (p) |uk (p)| Mk
 

 
k=n+1 k=n+1 k=n+1

as follows. Since the series on the right converges we have that for every
> 0 there is an N = N () such that the right hand side is < if n > N
and hence the left hand side is < for all p.
Remark 21.5. The proof doesnt require that the inequality |uk (p)| Mk
hold for all k but only for all sufficiently
P large k. This inequality isPoften
expressed by saying that the series k uk is dominated by the series k Mk
Theorem 21.6. Assume that the sequence (fn : [a, b] R)n converges uni-
formly to a function f and that each fn is continuous. Then the limit of the
integrals is the integral of the limit, i.e.
Z b Z b
lim f (x) dx = f (x) dx.
n a a

57
Proof. Choose > 0. Then there is an N such that |f (x)fn (x)| < /(ba)
for all x [a, b] and all n > N . For n > N we have, by the various properties
listed in Theorem 18.4, that
Z b Z b  Z b 
   

 f (x) dx f n (x) dx  =  (f (x) fn (x)) dx
  
a a a
Z b
|f (x) fn (x)| dx
a
(b a)
=
ba
as required.
Corollary 21.7. Let (fn )n be a sequence of of functions defined on an open
interval I. Assume that
(i) Each fn is differentiable;

(ii) Each derivative fn0 is continuous;

(iii) The sequence (fn )n converges uniformly on I;

(iv) The sequence (fn0 )n converges uniformly on I.

Then the limit f is differentiable and the limit of the derivatives is the deriva-
tive of the limit, i.e.
lim fn0 (x) = f 0 (x)
n

for x I.
Proof. Define g(x) := limn fn0 (x). By Theorem 21.3 g is continuous. By
the Fundamental Theorem of Calculus part (II) we have
Z x
fn (x) fn (a) = fn0 (t) dt
a

so Theorem 21.6 we have

Z x Z x Z x
f (x) f (a) = lim fn0 (t) dt = lim fn0 (t) dt = g(t) dt.
n a a n a

Hence f 0 = g by the Fundamental Theorem of Calculus part (I).

58
22 Power Series - I
22.1. A series of form k
P
k=0 ck (x a) is called a power series centered at
a. The radius of convergence of the power series is the number R defined
by
1
:= lim sup |ck |1/k .
R k
(If the lim sup is infinite, then R := 0 and if the lim sup is zero, then R := .)
Problem 22.2. (A formula for the radius of convergence). Assume that the
coefficients ck are nonzero. Show that
|ck+1 |
lim sup |ck |1/k = lim
k k |ck |

if the limit on the right exists.

Theorem 22.3. Let k
P
k=0 ck (x a) be a power series and R be its radius
of convergence. Then the series converges if |x a| < R and diverges if
|x a| > R. More precisely,
(i) If 0 < r < R the series converges uniformly on the interval [a r, a + r].
(ii) If |xa| > R then the nth term of the series is unbounded and hence does
not converge to zero (so the series does not converge by Theorem 20.4).
Proof. (Buck Theorem 14 page 240.) Let ak = ck (x a)k . Then by the Root
Test (Theorem 20.14) the series converges if lim supk |ak |1/k is less than
one and diverges if it is greater than one. But
|x a|
lim sup |ak |1/k = lim sup |ck |1/k |x a| =
k k R
so the series converges if |x a| < R and diverges if |x a| > R. To prove
that the convergence is uniform on the interval [a r, a + r] we need to
repeat the proof of the Root Test. Let = (r + R)/2 so 0 < r < < R.
As lim supk |ck |1/k = 1/R < 1/ it follows that |ck | < (1/)k for sufficiently
large k (say k > N ) so
 k
k r
|ck (x a) | < .

Since r/ < 1 this says that the power series is dominated by a convergent
geometric series if |x a| < r. Hence the power series converges uniformly
by Theorem 21.4. (See also Remark 21.5.)

59
Example 22.4. If a = 0 and ck = 1, then R = 1 and

1 X
= xk
1 x k=0
where the convergence is uniform on each interval [r, r] with 0 r < 1.
The convergence is not uniform on the interval (r, 1) or on the interval
(1, r) and the series does not converge at x = 1.
Example 22.5. If a = 0 and ck = 1/k!, then R = and

x
X xk
e =
k=0
k!
where the convergence is uniform on each interval [r, r] with r < .
Remark 22.6. A power series never converges uniformly on an unbounded
interval unless it is a polynomial. To see this we argue by contradiction. Let
n
X
Pn (x) = ck (x a)k
k=0

denote the nth partial sum of the power series k ck (xa)k and assume that
P
(Pn )n converges uniformly to a function f on some interval I. Let = 1.
Then there exists N Psuch that |Pn (x)f (x)| < 1 whenever n > N and x I.
If the power series k ck (x a)k has infinitely many nonzero terms, there
exist m > n > N with cm 6= 0 and cn 6= 0 and hence Pn Pm is a nonconstant
polynomial. But then |Pn (x) Pm (x)| |Pn (x) f (x)| + |f (x) Pm (x)| 2
for all x I. Hence the interval I must be unbounded as a nonconstant
polynomial becomes infinite as x .
Theorem 22.7. Let k
P
k=0 ck (x a) be a power series and R be its radius
of convergence. Denote the sum by
X
f (x) := ck (x a)k
k=0

for |x a| < R. Then f is differentiable on the interval (a R, a + R), its

derivative is given by term-wise differentiation, i.e.
X
0
f (x) = kck (x a)k1 ,
k=1

and the radius of convergence of this last power series is also R.

60
Corollary 22.8 (Taylor Series). Continue the hypotheses of Theorem 22.7.
Then f is infinitely differentiable12 on the interval (a R, a + R), the nth
derivative f (n) of f is

X X k!ck
f (n) (x) = k(k 1) (k n + 1)ck (x a)kn = (x a)kn ,
k=n k=n
(k n)!

so that cn = f (n) (a)/n!, i.e.

X f (k) (a)
f (x) = (x a)k .
k=0
k!

Corollary 22.9. The nth derivative of the sum of the geometric series is

X k! n!
xkn =
k=n
(k n)! (1 x)n

for 1 < x < 1.

23 Analytic Functions
Definition 23.1. A function f is called analytic iff for every a in its domain
there is a power series with

X
f (x) = ck (x a)k
k=0

for all x in some interval about a. From Corollary 22.8 it follows that an
analytic function is infinitely differentiable and that it equals its Taylor series
at each point a in its domain, i.e. ck = f (k) (a)/k! and hence

X f (k) (a)
f (x) = (x a)k .
k=0
k!

Problem 23.2. Show that function f : R R defined by

 1/x
e if x > 0
f (x) =
0 if x 0
12
A function is called infinitely differentiable iff it has derivatives of all orders.

61
is infinitely differentiable but not analytic. (All its derivatives vanish at zero
so it cannot equal its Taylor series.) Hint: Show inductively that the kth
derivative of f has form

P (x) 1/x
f (k) (x) = e
Q(x)

for x > 0 where P (x) and Q(x) are polynomials. Then use the definition of
the derivative to show that f (k+1) (0) = 0. Then

P (x) 1/x P (y) y 1

e = e , y=
Q(x)
Q(y) x

where P (y) and Q(y)

are also polynomials. The variable x approaches 0 as
y approaches . Use lHo pitals rule.

Problem 23.3. Give two proofs that the function f (x) = x1 defined for
x > 0 is analytic. First: Write f (x) = Pn (x, a) + Rn (x, a) where
n
X f (k) (a)(x a)k
Pn (x, a) :=
k=0
k!

and show that limn Rn (x, a) = 0 for x sufficiently near a. Second: Use
the identity
1 1
=
x a(1 + y)
where y = (x a)/a and the formula for the sum of a geometric series.

Theorem 23.4. If a function f is represented by a power series in an open

interval I about of some point in its domain, it is analytic in that interval,
i.e. if a I and

X f (k) (a)(x a)k
f (x) = ()
k=0
k!
for x I, then

X f (k) (b)(x b)k
f (x) =
k=0
k!
for x, b I.

62
Proof. This is normally proved in a course on complex variables (Math 623
at UW). Here is a proof that doesnt use complex numbers. Choose b I
and write Taylors Formula (Theorem 19.1) centered at b
f (x) = Pn (b, x) + Rn (b, x)
n Z x (n+1)
X f (k) (b) k f (t)
Pn (b, x) := (x b) , Rn (b, x) := (x t)n dt.
k=0
k! b n!
We must show that limn Rn (b, x) = 0 for x in an open interval about b.
Let R be the radius of convergence of () so that
f (k) (a)
R1 = lim sup |ck |1/k , ck := .
k k!
By Theorem 22.3 the series () diverges for |x a| > R so we may as well
assume that I = (a R, a + R). Since b I we have |b a| < R. Let
r = (R + |b a|)/2 and = (R + r)/2 so that |b a| < r < < R. As in the
proof of that theorem we have the inequality
 k
k
|ck | |x a|
r
for sufficiently large k and x (a r, a + r). Since this last inequality holds
for sufficiently large k there is an M such that
 k
k r
|ck | |x a| M

for all k and all x (a r, a + r). Take the absolute value of the series for
the nth derivative:
 
X k!c 
k
|f (n) (x)| =  (x a)kn 
 

k=n
(k n)! 
 
X  k!ck 
kn 

 (k n)! (x a) 
k=n
 kn
X k! r
M
k=n
(k n)!
 n
r
= M n! 1

63
by Corollary 22.9. Now choose > 0 so small that (b, b+) (ar, a+r)
and  1
:= 1 < 1.
r
Then for x (b , b + ) and t between b and x we have |x t| < so
Z x (n+1) 
 f (t) n

|Rn (b, x)| =  (x t) dt
b n!
 n1 n
r
|b x| M 1
n!
 n1 n+1
r
M (n + 1)! 1
n!
n+1
= M (n + 1) .

But < 1 so limn |Rn (b, x)| = M limn nn = 0 as required.

24 Power Series - II
Theorem 24.1. Assume that the power series

X
f (x) = ck (x a)k
k=0

has radius of convergence R and converges pointwise on the closed interval

[a, a + R]. Then it converges uniformly on that interval. Similarly for the
closed interval [a R, a].

Proof. (See Buck page 279.) We assume that R = 1 and a = 0 to simplify

the notation: the proof in general is similar but messier. Thus define

X
X
k
f (x) = ck x , 0 x < 1; f (1) = ck .
k=0 k=0

Define Bn :=
P
k=n ck . Then ck = Bk Bk+1 . Since the series f (1) converges,
it follows that limn Bn = 0. Choose > 0. There exists N such that

64
|Bn | < /2 for n > N . For x [0, 1) we have

X
X
k
ck x = (Bk Bk+1 )xk
k=n k=n

X
n
= Bn x + Bk+1 (xk+1 xk )
k=n+1

X
n
= Bn x + (x 1) Bk+1 xk
k=n

so for n > N
 
X 
X
k n
ck x  |Bn | x + (1 x) |Bk+1 | xk


 
k=n k=n

xn (1 x)x n X
+ xj
2 2(1 x) j=0

=
When x = 1 this inequality holds because |Bn | < /2 < .
Corollary 24.2. Suppose the power series

X
f (x) = ck (x a)k
k=0

converges pointwise on the interval [a, a + R]. Then the function f is contin-
uous on the interval [a, a + R]. Similarly for the closed interval [a R, a].
Proof. By Theorem 24.1 the series converges uniformly so the function f is
continuous by Theorem 21.3.
Remark 24.3. The formula

1 X
= tk
1 t k=0

was proved in 20.3. By Theorem 21.6 we may integrate and get

Z x
dt X xk+1 X xk
ln(1 x) = = =
0 1t k=0
k+1 k=1
k

65
which holds for 1 < x < 1. Let f (x) denote the right hand side. When
x = 1 the right hand side is the Alternating Harmonics Series from Exam-
ple 20.11 and it converges The series converges for x = 1 by Theorem 20.10.
Hence f is continuous on [1, 0) by Theorem 24.1. But ln(1 x) is also con-
tinuous on [1, 0). Since f (x) = ln(1 x) on (1, 1) this must remain true
on [1, 1) so taking x = 1 gives

X (1)k
ln(2) = .
k=1
k

This may stated more dramatically as

n n
X (x)k X (x)k
lim lim = lim lim
n x1
k=1
k x1 n
k=1
k

in contrast to
lim lim xn = 1 6= 0 = lim lim xn .
n x1 x1 n

25 The Heat Equation

Problem 25.1. Assume that the sequence (bn )n is dominated by the se-
quence (np )n i.e there is an N and C such that

|bn | Cnp

for n > N .

(1) Show that, if p > 1, the series

X
f (x) := bn sin nx (25.1-1)
n=1

converges uniformly.

(2) Show that, if p > 1, then

Z
2
bn = f (x) sin(nx) dx. (25.1-2)
0

66
(3) Show that, if p > 2, then f is differentiable and that

X
0
f (x) := nbn cos nx.
n=1

Hint: See Problem 20.17 above. You may use any of the theorems stated
above but state which theorems you are using and verify that the hypotheses
of the theorems are satisfied.

Problem 25.2. Continue the notation of Problem 25.1. Show that the series

2
X
u(t, x) := en t bn sin nx (25.2-1)
n=1

converges uniformly on [0, ) [0, ] if p > 1 and that the limit satisfies the
partial differential equation13

u 2u
= (25.2-2)
t x2
on the open set (0, ) (0, ). Show also that u is continuous on the closed
set [0, ) [0, ], that it satisfies the initial condition

u(0, x) = f (x) (25.2-3)

and the boundary condition

u(t, 0) = u(t, ) = 0. (25.2-4)

(You may use any of the theorems stated above but state which theorems
you are using and verify that the hypotheses of the theorems are satisfied.
2
Hint: en t is very small if t > 0 and n large.) It looks like this exercise
proves that the solution of the partial differential equation (25.2-2) subject
to the intial condition (25.2-3) and the boundary condition (25.2-4) is given
by (25.2-1) where the coefficients are defined by (25.1-2). Is there anything
missing for a rigorous proof?
13
This PDE is called the Heat Equation.

67
26 Fourier Series - I
27 Fourier Series - II
Theorem 27.1. Assume that the function f : R C is 2 periodic and
Lipshitz. Then the Fourier series for f converges uniformly to f , i.e.
n Z
X 1
f (x) = lim ikx
ck e , ck := f (x)eikx dx
n
k=n
2

and the convergence is uniform.

Proof. See Rudin page 189.

A Uniqueness of the Real Numbers

Definition A.1. A field is a set R equipped two binary operations

R R R : (x, y) 7 x + y, R R R : (x, y) 7 x y,

two distinct special elements 0 and 1, and two unary operations

R R : a 7 a, R \ {0} R \ {0} : a 7 a1 ,

satisfying the following (additive and multiplicative) commutative, associa-

tive, identity, and inverse laws:

Addition Multiplication
commutative a+b=b+a ab=ba
associative (a + b) + c = a + (b + c) (a b) c = a (b c)
identity a+0=a a1=a
inverse a + (a) = 0 a a1 = 1

and the distributive law:

(a + b) c = (a c) + (b c).

68
The operations of subtraction and division are then defined by
a 1
a b := a + (b) := a b1 = a .
b b
The standard abbreviations ab := a b, a b := a + (b), and a/b = a b1
are used. The above axioms are those which appear in 9.1.
A.2. The rational numbers Q, the real numbers R, and the complex numbers
C are the most important examples of fields, but there are many others, e.g.
the field R of rational functions from problem 9.9 and the field Q( 2) from
problem 9.11. In Math 441 (or 541) you will even meet finite fields. (The
simplest example of a finite field is the set {0, 1, 2, . . . , p 1} where p is a
prime, and addition and multiplication are done modulo p.)
A.3. The axioms in the column headed Addition above are the axioms for
an abelian group in additive notation, and the axioms in the column headed
Multiplication above are the axioms for an abelian group in multiplicative. If
you replace 0 by 1, a + b by a b, and a by a1 in the former column you get
the axioms in the latter column. Because both addition and multiplication
satisfy the axioms for an abelian group there are further analogies.

Addition Multiplication

(0) a + b = 0 = b = a (0) a b = 1 = b = a1

(i) (a) = a (i) (a1 )1 = a

(ii) (a + b) = a b (ii) (a b)1 = a1 b1

 a 1 b
(iii) (a b) = b a (iii) =
b a
a c ac
(iv) (a b) + (c d) = (a + c) (b + d) (iv) =
b d bd
a ac
(v) a b = (a + c) (b + c) (v) =
b bc
a/b a c
(vi) (a b) (c d) = (a b) + (c d) (vi) =
c/d b d

Line (0) explains the phrase necessarily unique used in 9.1. The last line
explains why we invert and multiply to divide fractions.

69
A.4. Using the distributive law one can also prove the following familiar
identities.

(vii) a 0 = 0 (viii) a = (1)a

(ix) a(b) = ab (x) (a)(b) = ab

a c ad + cb
(xi) + = (xii) (a + b)(c + d) = ab + ad + bc + bd
b d bd
(xiii) (a + b)2 = a2 + 2ab + b2 (xiv) (a + b)(a b) = a2 b2

It follows from these laws that if a product is zero, then one of its factors
must be zero. (Proof: if ab = 0 and a 6= 0 then b = 1 b(a1 a)b = a1 (ab) =
a1 0 = 0.)
Definition A.5. A homomorphism from a field R to a field R0 is a map
: R R0 such that
(a + b) = (a) + (b), (a b) = (a) (b)
for a, b R. (In each equation the operation on the left is the one for R
and the operation on the right is the one for R0 .) It follows easily that
(0) = 0, (1) = 1, (a) = (a), and (a1 ) = (a)1 . An isomorphism is
a bijective homomorphism.
Remark A.6. The inverse of an isomorphism is an isomorphism. For ex-
ample, if a0 , b0 , c0 := a0 + b0 R and a = 1 (a0 ), b = 1 (b0 ), c = 1 (c0 ),
then (a + b) = a0 + b0 (as is a homomorphism) and (c) = c0 = a0 + b0 , so
(c) = (a + b) so c = a + b (as is injective).
Proposition A.7. A homomorphism of fields is injective.
Proof. Assume that (a) = (b); we must show that a = b. If a 6= b, then
c := a b 6= 0 so 1 = (1) = (c1 c) = (c)1 (c) = (c)1 (a b) =
(c)1 (0) = (c)1 0 = 0 contradicting 1 6= 0.
Definition A.8. An ordered field is a field equipped with an order relation
satisfying the conditions in 9.2. An order preserving homomorphism from
an ordered field R to an ordered field R0 is a homomorphism : R R0 such
that
a < b = (a) < (b).

70

Example A.9. The inclusions Q Q( 2), Q R, $ R are all order
preserving homomorphisms. The map

Q( 2) 7 Q( 2) : a + b 2 7 a b 2
is a field homomorphism but is not order preserving.
Lemma A.10. For any ordered field R, there is a unique order preserving
homomorphism : Q R from the set Q of rational numbers into R.
Proof. The homomorphism must send 0 Q to 0 R, 1 Q to 1 R,
n Z+ to
(n) := (1) + (1) + + (1) R
| {z }
n

a negative integer n to (n) = (n), and therefore a ratio m/n of integers

to (m/n) = (m)/(n). Since (1) > 0 and R obeys the same algebraic and
order laws as Q, it is immediate that is an order preserving homomorphism.

Definition A.11. An ordered field R is complete iff every subset S of R

which is bounded above has a least upper bound sup(S). (The definitions
are mutatis mutandis the same as in 9.5.)
Theorem A.12. For any complete ordered field R, there is a unique order
preserving isomorphism : R R from the set R of real numbers onto R.
Proof. For a R we have a = sup{q Q : q < a} so we must have
(a) = sup{(q) : q Q, q < a}.
We must show that is an order preserving homomorphism and that it is
surjective. We omit the details.

B Additional Problems
Problem B.1. Fix a positive number a R. The purpose of this problem
is to define the exponential ax for x R. Define a0 := 1 and for n a positive
integer define
an := |a a{z a}, an := 1/an .
n
Then

71
(1) Prove that for every nonzero integer n there is a unique solution b > 0
to the equation bn = a. Define a1/n to be this unique solution, i.e.
a1/n = b bn = a.

(2) For a rational number q define aq by aq = (am )1/n where q = m/n. Prove
that this definition is independent of the choice of the integers m and
n such that q = m/n.

(3) Prove that there is a unique continuous function

R (0, ) : x 7 ax

such that ax = aq when x is a rational number q.

In your proof make clear which theorems from these notes you are appealing
to. Also make your proof self contained so that a person who doesnt have
access to the statement of the problem can follow it. (You neednt provide
proofs for the theorems you use, but do provide references to them.) In your
proof of (3) you may use the inequality

|ap aq | M |p q|

which is true when a > 1, N is a positive integer, 1 p, q N + 1, and

N
!
X 1
M := aN .
k=1
k

You need not prove this inequality but use calculus to show where it comes
from. Hint: What is the definition of ln x, ex , and ax used in calculus?
Consider the Integral Test 20.13. The natural logarithm function ln x is
usually defined as an integral. How do you bound an integral by a sum? How
does the Mean Value Theorem from calculus (see 17.4)) give inequalities like
this?

72
Index
n space, 9 contrapositive, 7
nth Term Test, 52 converge, 26
ntuple, 9 converge absolutely, 53
converge conditionally, 53
accumulation point, 24, 26 converge pointwise, 56
Algebraic Axioms, 18 converge uniformly, 56
Alternating harmonic series, 54 convergent, 26
alternating series, 53 converges, 26
analytic, 61 converse, 7
Archimedean Property, 20 countable, 16
Axiom of Choice, 14
De Morgans Laws, 10
bijective, 14 decreasing, 28
Bolzano-Weierstrass, 30 Dedekind cuts, 20
boundary, 36 definite integral, 44
bounded, 24 denumerable, 16
bounded above, 19 derivative, 25, 43
bounded below, 19 differentiable, 43
cardinality, 16 direct product, 9
Cartesian product, 9 disconnected, 37
Cauchy, 32 disjoint, 8
Cauchy Convergence Criterion, 32 distance, 22
Cauchy Convergence Criterion for Se- diverge, 26
ries, 52 domain, 11
closed, 35, 36 dominated, 57
closure, 36 dot product, 23
cluster point, 24 dummy variable, 49
compact, 39 empty set, 8
Comparison Test, 54 equal, 8, 11
complement, 9 equality of sets, 11
complete, 71 extended real numbers, 25
Completeness Axiom, 19 exterior, 36
composition, 13
connected, 37 Fibonacci numbers, 15
continuous, 32, 33 field, 68

73
finite, 16 lower sum, 44
function, 11
Fundamental Theorem of Calculus, 48 map, 11
mathematical induction, 15
geometric series, 51 Mean Value Theorem, 43
graph, 11 mesh, 44
greatest lower bound, 19 monotonic, 28

Harmonic series, 52 natural numbers, 5

Heat Equation, 67 neighborhood, 24, 25
Heine Borel, 39 nondecreasing, 28
homomorphism, 70 nondenumerable, 16
nonincreasing, 28
identity map, 13 norm, 22
image, 11
increasing, 28 one-one, 14
index set, 9 one-one onto, 14
indexed family of sets, 9 onto, 14
inductive definitions, 15 open, 34
infimum, 19 open ball, 23
infinite, 16 open cover, 39
infinitely differentiable, 61 Order Axioms, 18
injective, 14 order preserving, 70
inner product, 23 ordered field, 70
Integral Test, 54
interior, 36 partial sums, 50
Intermediate Value Theorem, 38 partition, 44
intersection, 8, 9 permutation, 56
interval notation, 5 power series, 59
inverse, 13 preimage, 12
inverse image, 11 proof, 7
isomorphism, 70 proof by contradiction, 7
punctured neighborhood, 24
least upper bound, 19
left inverse, 13 radius of convergence, 59
limit point, 24 range, 12
Lipschitz, 33 real numbers, 18
logical operations, 6 rearrangement, 56
lower bound, 19 refines, 44

74
relatively open, 35
Riemann sum, 46
right inverse, 13
Root Test, 54

Schwarz inequality, 23
sequence, 26
series, 50
source, 12
strictly, 28
subsequence, 29
subset, 8
supremum, 19
surjective, 14

target, 12
Taylor polynomial, 49
Taylor remainder, 49
Taylor Series, 61
Taylors Formula Lagrange Form,
49
terms, 50
trace, 31
truth tables, 10

uncountable, 16
uniformly continuous, 33
union, 8, 9
upper bound, 19
upper sum, 44

weakly monotonic, 28
Weierstrass Comparison Test, 57

75