Sequences and Series

Contents
PrefaceIII
1. Preliminaries and notation 2
1.1 Basic set notation2
1.2 Upper and lower bounds of sets2
1.3 Relations and functions3
1.4 Limits and derivatives of functions5
1.5 Mathematical induction8
2. Sequences 11
2.1 General properties of sequences11
2.2 Arithmetic progressions16
2.3 Geometric progressions18
2.4 Convergent sequences21
2.5 Techniques for evaluating limits28
2.5.1 Indeterminate form ∞ − ∞31
2.5.2 Indeterminate form ∞
∞ 34
0
2.5.3 Indeterminate form 39
0
2.5.4 Indeterminate form 0 × ∞40
2.5.5 Indeterminate forms 00, ∞0 and 1∞42
2.6 Suggested exercises44
3. Series 52
3.1 Basic definitions and convergence52
3.2 Series of non-negative terms61
3.3 Alternating series86
3.4 Function sequences and series91
3.4.1 Sequences of functions91
3.4.2 Function series100
3.5 Power series103
3.5.1 Definition and properties104
3.5.2 Representation of a function as a power series110
3.6 Suggested exercises115
Solutions to suggested exercises 124
II
Preface
The idea behind this book was mainly to put on paper and release to the public much
of the contents of the lecturing notes from personal teaching experience, from
secondary to university level, regarding the subject of sequences and series.
From the point of view of the mathematician, the subject of sequences and series is
one the most interesting chapters in the curriculum of the young apprentice, as it
deals with the biggest addiction a creative mind has: patterns! Moreover, learning
about sequences and series constitutes one of the first moments that the student is
taught how to deal with such phenomena, how to mathematically tame them and
predict their behavior.
Throughout many years of teaching, I found that many students struggle with the
subject of sequences and series, and thus I began, independently, drawing these
lectures notes, which were mainly a strategy to present the subject in the most clear
and understandable way possible. The notes were revised many times throughout the
years, including additions, examples, and corrections.
I feel obligated to thank all the students we taught throughout the years for their
direct and indirect input in the classes. That feedback received was, in the end,
responsible for the contents and the development of the teaching notes and, of
course, the present text.
Some of the sections were expanded in an attempt to make the reading of this book
more appealing for the interested reader. The subject is presented in a clear and logical
manner for both beginner and intermediate students at secondary or university levels
and whether they have no or pre-existing knowledge of the subject. It is an avenue to
learn more about the subject or refresh past learnings.
I sincerely hope that this text will prove to be a helpful and useful tool for understand-
ing sequences and series as well as other subjects one may study on their academic path.
Luís Vieira
University of Porto,
Porto, Portugal
Sequences and Series:
An Introduction for Beginners
Luís Vieira
Abstract
This mathematics book presents the main definitions and properties of sequences,
series of real numbers, sequences, and series of functions. We combine a rigorous
exposition with a set of intuitive examples to achieve an easily readable final text that
is approachable by intermediate and more advanced calculus students from the sec-
ondary to the university level. This book is divided into three sections: an introductory
one where we present the basic mathematical concepts and notation we will follow; a
section devoted to sequences of real numbers; and a section dedicated to series of real
numbers, sequences, and functions. These two last sections end with exercises with
solutions for the readers to test their learning and knowledge. Our objective was to
provide students and curious minds on the subject of the book with a tool for
autonomous learning.
Keywords: calculus, real analysis, sequences of functions, sequences of real numbers,

series of functions, series of real numbers, Taylor series
Sequences and Series – An Introduction for Beginners
1. Preliminaries and notation
We decided to begin this book with a brief exposition of the main basic
mathematical concepts we will use and assume the reader is familiar with. Also, the
notation used in this book will be presented in the following paragraphs. For further
reading on the topics addressed in this book, we recommend the detailed expositions
presented in Refs. [1–10]. Refs. [1–7] consist of theoretical fundamental texts while
Refs. [8–10] present more practical approaches.
1.1 Basic set notation
In this book, we will deal with real numbers. Let us start by recalling the basic set
notation.
We denote by  the set of all natural numbers excluding 0, that is,
 ¼ f1, 2, 3, … g,
and the set of integers is denoted by ,
 ¼ f … , 3, 2, 1, 0, 1, 2, 3, … g:
Next, the set of rational numbers is denoted by :

na o
¼ : a, b ∈ ∧b 6¼ 0 :
b
Finally, we denote by  the set of real numbers, that is,
 ¼ ∪fx : x is irrationalg,
where it is sufficient to say that being irrational is not being able to be written as
the division of two integers.
Some additional symbols may be used to represent special subsets of the main sets
presented above like, for example,
0 ¼ ∪f0g;
p ¼ fn ∈  : n ≥ pg
 ¼ fx ∈  : x < 0g;
þ
0 ¼ fx ∈  : x ≥ 0g:
1.2 Upper and lower bounds of sets
Some sets of real numbers are limited as their elements will not be bigger or
smaller than some values.
Definition 1.1 (Bounded sets). Let X be a non-empty subset of .
1. X is said to be bounded from above if there is a real number M such that

x ≤ M, ∀x ∈ X: In these conditions, M is called an upper bound of X.
2
Sequences and Series: An Introduction for Beginners
ITexLi.113863
2. X is said to be bounded from below if there is a real number m such that

x ≥ m, ∀x ∈ X: In these conditions, m is called a lower bound of X.
3. X is said to be bounded if it is bounded from above and below and unbounded if it is

not bounded.
Definition 1.2 (Maximum, minimum, supremum, and infimum of sets). Let X be a

non-empty subset of . A real number is
1. the maximum of X if it is an element of X and if it is an upper bound of X.
2. the minimum of X if it is an element of X and if it is a lower bound of X.
3. the supremum of X if it is the smaller of all upper bounds of X.
4. the infimum of X if it is the bigger of all lower bounds of X.
Note that, for a given set, the existence of any of the numbers presented in
Definition 1.2 is not guaranteed. Also, observe that, when the maximum and
minimum of X do exist, they coincide with the supremum and infimum of X,
respectively.
Example 1.1. Let X ¼0, 1. For this set, the following statements are true.
• This set is bounded from above, as ½1, þ ∞½ is the set of all upper bounds of X.
• The set X is bounded from below, as ∞, 0 is the set of all lower bounds of X.
• Hence, X is a bounded set.
• The maximum of X is 1.
• X does not have a minimum.
• The supremum of X is 1.
• The infimum of X is 0.
The concept of neighborhood of a number will also be used ahead. We define it in a

simple way as follows.
Definition 1.3 (Neighborhood of a number). Let a be a real number. A neighborhood
of a is a set that contains an interval a ε, a þ ε½, with ε > 0.
1.3 Relations and functions
Given two sets X and Y, the Cartesian product X Y is defined as
X Y ¼ fðx, yÞ : x ∈ X∧y ∈ Yg
and its elements are called ordered pairs.

3
A binary relation or correspondence, R, between X and Y is simply a subset of X Y.

That subset is often called the graph of R. If ðx, yÞ ∈ R, it means that x is related to y
through R, we say that x is R-related to y, and we write xRy. The order is important, as
xRy does not necessarily imply that yRx.
The sets X and Y are usually called the departure set and the destination set of R,
respectively. The domain of R is the set of elements of X that are R-related to at least
one element of Y, and the image of R is the set of elements of Y for which there is at
least one element of x such that xRy.
There are different types of binary relations.
Definition 1.4 (Types of binary relations). Let R be a binary relation between the sets
X and Y.
1. If, for all x ∈ X and y1 , y2 ∈ Y we have xRy1 ∧ xRy2 ) y1 ¼ y2 , then R is said to be a

functional relation.
2. If, for all x ∈ X, there exists at least one y ∈ Y such that xRy, then R is said to be
left-total.
3. If R is functional and left-total, it is called a function.
4. If, for all x1 , x2 ∈ X and y ∈ Y we have x1 Ry ∧ x2 Ry ) x1 ¼ x2 , then R is said to be

injective.
5. If, for all y ∈ Y there exists at least one x ∈ X such that xRy, then R is said to be
surjective or right-total.
6. A function that is injective and surjective is said to be bijective or a bijection.
Example 1.2. Let X ¼ f1, 2, 3g, Y ¼ fa, b, cg, and Z ¼ fα, βg. Then:
• R1 ¼ fð1, aÞ, ð2, bÞg is a functional and injective relation between X and Y that is not
a function;
• R2 ¼ fð1, αÞ, ð2, βÞg is a functional, injective, and surjective relation between X and
Z that is not a function because it is not left-total;
• R3 ¼ fðα, aÞ, ðα, bÞ, ðβ, cÞg is a left-total, surjective, but not injective, relation between
Z and Y that is not a function because it is not functional;
• R4 ¼ fð1, aÞ, ð2, aÞ, ð3, aÞg is function between X and Y that is not injective or
surjective;
• R5 ¼ fðα, aÞ, ðβ, bÞg is an injective, but not surjective, function between Z and Y;
• R6 ¼ fð1, αÞ, ð2, αÞ, ð3, βÞg is a surjective, but not injective, function between X and Z;
• R7 ¼ fð1, cÞ, ð2, aÞ, ð3, bÞg is a bijective function between X and Y.
Usually, functions are denoted by the letter f and, if x and y are f-related, instead of
writing xfy, we write y ¼ f ðxÞ.
4
ITexLi.113863
1.4 Limits and derivatives of functions
In this section, we make a brief summary of two concepts that might be of interest
for the reader to understand a few examples presented in this text.
The concept of limit of a function can actually be formally defined with the
help of the limit of sequences, which will be surveyed in detail in Section 2.4
of Section 2. The idea behind both of them is, however, the same. Given a
function f, we say that the limit of f, when x approximates to the value a, is b, and we
write
lim f ðxÞ ¼ b,
x!a
if the images of the elements of the domain, which are converging to a, are
converging to b. Note that a and b can be ∞. In the special case of sequences, that we
will study in the next section, the limits are always calculated when a ¼ þ∞ because
of the nature of this special kind of functions.
The derivative of a function, f, of one variable, x, is essentially an instantaneous
rate of change of the images of f with the values of x. It is denoted by f 0 , Lagrange’s
df
notation, or by dx , Leibniz’s notation, and can be defined by the limit (if the limit
exists):
f ðx þ hÞ f ðxÞ
f 0 ðxÞ ¼ lim : (1)
h!0 h
If the limit in (1) is a real number, then f is said to be differentiable at x.

From (1), one can easily deduce formulae for the derivative of many
functions. Two basic rules that can be deduced are very useful while
differentiating are the product and the quotient rules: given two functions f and g,
both differentiable, then
ðfgÞ0 ðxÞ ¼ f 0 ðxÞg ðxÞ þ f ðxÞg0 ðxÞ

and
0
f f 0 ðxÞg ðxÞ f ðxÞg0 ðxÞ
ðxÞ ¼ ,
g g2 ðxÞ
provided, in the latter case, that gðxÞ 6¼ 0.

To differentiate other more complex analytic expressions of functions one makes
use of the so-called chain rule. Indeed, if we have a composite function f ∘g, then we
have:
ð f ∘ g Þ0 ðxÞ ¼ f 0 ðg ðxÞÞ g0 ðxÞ, (2)
provided that g is differentiable at x and f is differentiable at g(x).

For the reader to be able to follow some of the examples presented ahead, we will
provide the most common derivatives in Table 1. All of the formulae presented can be
obtained by making use of (1) and/or (2). For more insight on this derivative
5
f(x) f 0 ðxÞ f(x) f 0 ðxÞ
a 0 u u0
ax a au au0
axn naxn1 aun anun1 u0

pffiffiffi 1 ffiffi
pffiffiffi u0 ffiffi
x p u p
2 x 2 u
p ffiffiffi p ffiffiffi
pn ffiffiffiffiffiffi
1 ffi 0
n
x n
u pnuffiffiffiffiffiffiffi
n xn1 n un1
ln(x) 1 ln(u) u0
x u
log b ðxÞ 1
x ln b
log b ðuÞ u0
u ln b
ex ex eu u 0 eu
x x u 0 u
b b lnðbÞ b u b lnðbÞ
sin(x) cos(x) sin(u) u0 cosðuÞ
cos(x) sin(x) cos(u) u0 sinðuÞ
tan(x) sec 2 ðxÞ tan(u) u0 sec 2 ðuÞ

pffiffiffiffiffiffiffi
1 ffi 0
arcsin(x) arcsin(u) pffiffiffiffiffiffiffi
u ffi
1x2 1u2
pffiffiffiffiffiffiffi
1 ffi 0
arccos(x)
1x2
arccos(u) pffiffiffiffiffiffiffi
u ffi
1u2
arctan(x) 1 arctan(u) u0
x2 þ1 u2 þ1
sinh(x) cosh(x) sinh(u) u0 coshðuÞ
cosh(x) sinh(x) cosh(u) u0 sinhðuÞ
tanh(x) sech2 ðxÞ tanh(u) u0 sech2 ðuÞ
Table 1.
Most common derivatives ða ∈ , b ∈ þ nf1g and u is a function with respect to x).
operator, including its operation properties, please consult any calculus book; as for
the scope of this book, what was enlightened is enough, and we do not pretend to go
deeper on this subject.
Finally, we will make use, at some point, of a technique, called l’Hôpital’s Rule,
for computing limits of functions that are the division of two other functions.
The technique has two cases of application that are presented next, but both work
basically in the same way; that is, you just have to consider the derivatives of the
numerator function and the denominator function and recalculate the limit.
Usually, differentiating a function makes it simpler. This rule made it is first appari-
tion in a book of the French mathematician Guillaume de l’Hôpital.
Theorem 1.1 (l’Hôpital’s Rule for 0/0). Let f and g be two differentiable functions on
an open interval I containing a ∈  and g is such that g 0 ðxÞ 6¼ 0 in Infag: If
lim x!a f ðxÞ ¼ lim x!a g ðxÞ ¼ 0, then
f ðxÞ f 0 ðxÞ
lim ¼ lim 0 ,
x!a g ðxÞ x!a g ðxÞ
so long as the limit is finite or ∞.

Theorem 1.2 (l’Hôpital’s Rule for ∞=∞). Let f and g be two differentiable functions on
an open interval I containing a ∈  and g is such that g 0 ðxÞ 6¼ 0 in Infag. If lim x!a f ðxÞ ¼
∞ and lim x!a g ðxÞ ¼ ∞, then
6
ITexLi.113863
f ðxÞ f 0 ðxÞ
lim ¼ lim 0 ,
x!a g ðxÞ x!a g ðxÞ
so long as the limit is finite or ∞.

Note that, in both cases, similar results hold for a ¼ ∞.
In the following, we present some practical examples of the effectiveness of
l’Hôpital’s Rule on calculating limits.
Example 1.3. Let’s use l’Hôpital’s Rule to compute
ex þ 2
lim :
x!0 3x
Indeed, we have

0 d
2 2e x 0 ð2 2ex Þ
lim ¼ lim dx
x!0 3x x!0 d
ð3xÞ
dx
2ex
¼ lim
x!0 3
2e0
¼
3
2
¼ :
3
Example 1.4. Now we will compute
sin x
lim :
x!0 5x
using l’Hôpital’s Rule. We have

0 d
sin x 0 ðsin xÞ
lim ¼ lim dx
x!0 5x x!0 d
ð5xÞ
dx
cos x
¼ lim
x!0 5
1
¼ :
5
Example 1.5. In this example, we will apply l’Hôpital’s Rule to calculate
ex1
lim :
x!þ∞ 2x3
7
We have
∞
d x1
ex1 ∞ e
lim ¼ lim dx
x!þ∞ 2x3 x!þ∞ d 3
2x
dx
ex1
¼ lim
x!þ∞ 6x2
∞
d x1
e
∞
¼ lim dx
x!þ∞ d
6x2
dx
ex1
¼ lim
x!þ∞ 12x
∞
d x1
e
∞
¼ lim dx
x!þ∞ d
ð12xÞ
dx
ex1
¼ lim
x!þ∞ 12
þ∞
¼
12
¼ þ∞:
1.5 Mathematical induction
Since sequences are essentially, as we will shortly see, functions with domain , it
is important to recall the principle of mathematical induction, an important technique
for proving properties among the naturals.
Proposition 1.1 (Principle of Mathematical Induction). Let P(n) be a statement
involving the natural number n. The proposition
∀n ∈ , PðnÞ
is true if:
1. the statement P(n) is true for n ¼ 1, that is, P(1) is true;
2. if the statement P(n) is true for n ¼ m ∈ , then it is also true for n ¼ m þ 1, that is,
∀m ∈ , PðmÞ ) Pðm þ 1Þ:
The principle of mathematical induction comes from the axiomatic definition of 

which is defined as the smallest subset, S, of  such that, if x ∈ S, then x þ 1 ∈ S. If the
statement has this inductive property, that is, if it is valid for a number m, then it is
8
ITexLi.113863
also valid for it is successor, m þ 1, then verifying that the statement is valid for 1
originates a chain reaction that proves the statement for all-natural numbers.
Note that Proposition 1.1 can be generalized for sets p , with p ∈ . In this case, the
first step is to verify if the statement is valid for n ¼ p and not n ¼ 1. The rest is
analogous.
Example 1.6. Using mathematical induction, let’s prove that, for each n ∈ , we have
nðn þ 1Þð2n þ 1Þ
12 þ 22 þ 32 þ ⋯ þ n2 ¼ :
6
For n ¼ 1, we have only one element in the sum, therefore,
1 ð1 þ 1Þ ð2 1 þ 1Þ
12 ¼
6
123
⇔ 1¼
6
⇔ 1 ¼ 1,
and the formula is valid.

Now, suppose that, for a certain m ∈ , we have
mðm þ 1Þð2m þ 1Þ
12 þ 22 þ ⋯ þ m2 ¼ :
6
Under this assumption (usually called the hypothesis of induction), let’s prove that the
formula is valid for m þ 1 (usually called the thesis of induction), that is, that is valid the
equality
ðm þ 1Þðm þ 2Þð2m þ 3Þ
12 þ 22 þ ⋯ þ ðm þ 1Þ2 ¼
6
2m3 þ 9m2 þ 13m þ 6
¼ :
6
Indeed, using the hypothesis of induction, we have
mðm þ 1Þð2m þ 1Þ
12 þ 22 þ ⋯ þ m2 þ ðm þ 1Þ2 ¼ þ ðm þ 1Þ2
6
mðm þ 1Þð2m þ 1Þ þ 6ðm þ 1Þ2
¼
6
2m3 þ 3m2 þ m þ 6m2 þ 12m þ 6
¼
6
2m3 þ 9m2 þ 13m þ 6
¼
6
ðm þ 1Þðm þ 2Þð2m þ 3Þ
¼ :
6
This means that, if the formula is valid for m ∈ , then it is also valid for m þ 1. Since
we already proved that the formula is valid for n ¼ 1, we can conclude, by mathematical
induction, that the formula is valid for all n ∈ .
9
Example 1.7. In this example, we will prove, using induction, that for each
natural n, we have:
ðn þ 3Þ! ≥ 2nþ3 : (3)
Indeed, for n ¼ 1, we have
ð1 þ 3Þ! ¼ 24 ≥ 16 ¼ 21þ3 ,
which means that (3) is valid.

Let’s suppose that, for a certain natural m, we have
ðm þ 3Þ! ≥ 2mþ3 :
Let’s prove that
ðm þ 1 þ 3Þ! ≥ 2mþ1þ3 ⇔ðm þ 4Þ! ≥ 2mþ4 :
We have:
ðm þ 4Þ! ¼ ðm þ 4Þðm þ 3Þ!

≥ ðm þ 4Þ2mþ3
> 2 2mþ3
¼ 2mþ4 ,
using the induction hypothesis in the second step and, in the third step, the fact that
m þ 3 > 2:Hence,
ðm þ 4Þ! ≥ 2mþ4
and we can conclude, by induction, that (3) is valid for all n in .
10
ITexLi.113863
2. Sequences
In this section, we present a general survey on the topic of sequences. A sequence is

just a function of arbitrary real numbers but whose domain is the set of natural
numbers, . The origin of such functions can be traced back farther until at least
ancient Greece, particularly to the grounds of the mythical Pythagoras’ school, sixth
century BC, and the practice of its followers.
This special kind of functions has many practical applications, since many real-life
situations are mainly a discrete succession of outputs, which can be modeled by a
sequence of real numbers. For example, consider the sequence of perfect squares,
also known simply as the sequence of squares: 1, 4, 9, 16, 25, 36, and so on. This
sequence is the function of domain  that takes each natural number and maps it into
its square.
Additionally, sequences are able to open up a window toward some rather complex
mathematical concepts, such as the different types of infinity, as well as the concepts
of convergence and limit. In particular, sequences allow us to establish, in a rather
easy and elegant manner, the concept of limit, which is a structural stone in the field
of mathematical analysis.
This section is organized in the following way. First, in Section 2.1, we
present some general properties of sequences such as monotony and bounds. In
Sections 2.2 and 2.3, we present and characterize a couple of special kinds of
sequences, which are the arithmetic and the geometric progressions,
respectively. Then, in Section 2.4, we address the fundamental problem of
sequence convergence, presenting the most relevant results in the subject and
examples, and finally, in Section 2.5, we present useful techniques for computing
the limits of sequences. We finish the section with some suggested exercises in Section
2.6.
2.1 General properties of sequences
We will start this section with a formal definition of sequence.

Definition 2.1 (Sequence). Let A be a non-empty set. A sequence, u, in A is any
function that can be defined from  into A, that is,
u:  ! A
:
n ↦ uðnÞ
The function values of the sequence u, that is, u(1), u(2), u(3), and so on, are
called the terms of the sequence u, with u(1) being the first term, u(2) the second term, u
(3) the third term, etc. The set of terms of a sequence u is precisely the image of the
map u. The object of a term is also referred to as the order of the term. For instance,
the fourth term of the sequence, u(4), is the term of order 4, and the nth term, u(n), is
the term of order n.
Along this text, we will consider only real sequences, that is, sequences defined in
. We will also adopt the following standard notation for sequences. Instead of
11
writing u(1), u(2), … , u(n), … , we will denote the terms of the sequence as
u1 , u2 , … , un , … , and the sequence u, itself, will be represented as ðun Þn ∈  , or simply
by ðun Þ.
In frequent cases, the term of order n of a sequence can be obtained by a formula
involving n. Let us take another look at the example of the square numbers presented
at the beginning of this section.
Example 2.1. Consider the sequence, ðun Þ, of squares number (see Figure 1).
Figure 1.
Figure of squares.
The term of order n of this sequence is given by un ¼ n2 .

In cases similar to Example 2.1, where the term of order n is given by a certain
formula involving n, un assumes a greater level of importance within the sequence
ðun Þ because it basically yields an algebraic expression that is capable of generating all
the terms of ðun Þ, given the value of n. In view of this, the term of order n of ðun Þ is
often referred to as the general term of the sequence.
The terms of a sequence ðun Þ can also be plotted out in a Cartesian coordinate
system. The graph of the sequence ðun Þ is the set
fð1, u1 Þ, ð2, u2 Þ, ð3, u3 Þ, … , ðn, un Þ, … g:
Example 2.2. Consider the sequence ðvn Þ defined by its general term
3n þ 2
vn ¼ :
n
The graph of ðvn Þ is the set

11 7 17
ð1, 5Þ, ð2, 4Þ, 3, , 4, , 5, ,… ,
3 2 5
which can be graphed as in Figure 2:
12
ITexLi.113863
Figure 2.
Graphically representation of a succession 2.
One can also define a sequence by recurrence, that is, after defining one, or
more, of the terms of the sequence, all the other terms are obtained by solving a
so-called recurrence equation involving terms already determined before. In general,
each term is obtained by making use of the term or terms that precede it. A famous
example of a sequence defined in such a manner is the sequence of the Fibonacci
numbers.
Example 2.3. The Fibonacci numbers are the following: 1, 1, 2, 3, 5, 8, 13, 21, and so
on. After the first two numbers, each new number is obtained by adding the previous two
Fibonacci numbers. This sequence of numbers, ðF n Þn ∈  , can be defined as follows:
8
>
> F1 ¼ 1
>
<
F2 ¼ 1
>
>
>
:
F n ¼ Fn1 þ Fn2 , ∀n ∈ nf1, 2g
All the basic algebraic operations for real numbers are still valid for real sequences.
Indeed, if ðan Þn ∈  and ðbn Þn ∈  are two real sequences, then one can define the
sequence of the sum, ða þ bÞn ∈  , the sequence of the difference, ða bÞn ∈  , the sequence of

the product, ðabÞn ∈  , and the sequence of the quotient, ab n ∈  , providing, in the latter
case, that bn 6¼ 0, ∀n ∈ , whose general terms are defined as ða þ bÞn ¼ an þ

bn , ða bÞn ¼ an bn , ðabÞn ¼ an bn and ab n ¼ abnn , respectively.
We proceed the exposition with some general properties of some sequences.
Definition 2.2 (Monotonic sequence). A sequence ðun Þ is:
1. increasing if, ∀n ∈ , un ≤ unþ1 ;
2. decreasing if, ∀n ∈ , un ≥ unþ1 ;
13
3. strictly increasing if, ∀n ∈ , un < unþ1 ;
4. strictly decreasing if, ∀n ∈ , un > unþ1 ;
5. monotonic if it is increasing or decreasing;
6. strictly monotonic if it is strictly increasing or strictly decreasing.
We follow up Definition 2.2 with some examples.

Example 2.4. Consider the sequence ðwn Þ with general term given by
wn ¼ 2 3n:
Since
wnþ1 wn ¼ 2 3ðn þ 1Þ ð2 3nÞ

¼ 2 3n 3 2 þ 3n
¼ 3 < 0,
we conclude that wn > wnþ1 holds for each n ∈ , thus ðwn Þ is strictly decreasing.
Example 2.5. Let ðun Þ be the sequence whose general term is
2
un ¼ :
2n
In this case, we have
2 2
unþ1 un ¼
2 ðn þ 1Þ 2 n
2 2
¼
1n 2n
2ð2 nÞ 2ð1 nÞ
¼
ð1 nÞð2 nÞ
2
¼ > 0,
ð1 nÞð2 nÞ
since the formula in the denominator always yields positive numbers, regardless of the
value of n. Therefore, we can conclude that un < unþ1 holds for each n ∈ , thus ðun Þ is
strictly increasing.
Example 2.6. In this example, consider ðvn Þ to be the sequence defined by
vn ¼ 4n þ 2ð1Þn :
For this sequence, we have

14
ITexLi.113863
vnþ1 vn ¼ 4ðn þ 1Þ þ 2ð1Þnþ1 ð4n þ 2ð1Þn Þ

¼ 4 þ 2 ð1Þnþ1 ð1Þn :
Observe that this expression yields the values 0 and 8, successively, when n is even or odd,
respectively. This means that vn ¼ vnþ1 when n is even and vn < vnþ1 when n is odd.
Accordingly, ðvn Þ is an increasing, but not strictly increasing, sequence.
Example 2.7. To finish this set of examples, consider the sequence ðwn Þ, whose general
term is given by
wn ¼ n2 6n:
In this case, we have

wnþ1 wn ¼ ðn þ 1Þ2 6ðn þ 1Þ n2 6n
¼ n2 þ 2n þ 1 6n 6 n2 þ 6n
¼ 2n 5:
This expression yields negative values for n ∈ f1, 2g and positive values in the remaining
cases. This means that wn > wnþ1 when n ¼ 1, 2 and wn < wnþ1 when n > 2. Accordingly,
ðwn Þ is a non-monotonic sequence.
In the following, we introduce the concept of bounded sequences.
Definition 2.3 (Bounded sequence). A sequence ðun Þ is:
1. bounded from above if there is a M ∈ , such that un ≤ M, ∀n ∈ ;
2. bounded from below if there is a m ∈ , such that un ≥ m, ∀n ∈ ;
3. bounded if it is bounded from above and below.
From Definition 2.3, we conclude that a bounded sequence, ðun Þ, is one in which all
terms are bounded by certain real numbers m and M, that is, m ≤ un ≤ M. It is
straightforward to see, though, that this assertion can be slightly modified. In fact,
considering k ¼ maxf m jj, M j g, we have that ∣ un ∣ ≤ k, ∀n ∈ .
There is a couple of immediate consequences relating to Definitions 2.2 and 2.3. An
increasing sequence is clearly bounded from below by its first term, and a decreasing
sequence is clearly bounded from above also by its first term.
One might be tempted to conclude that a strictly increasing sequence cannot be
bounded from above, but such a conclusion is false. The same applies to strictly
decreasing sequences bounded from below. The classical examples are the sequences
of general terms un ¼ n1 and vn ¼ n1 . The sequence ðun Þ is strictly increasing and ðvn Þ
is strictly decreasing, yet we have 1 ≤ un ≤ 0 and 0 ≤ vn ≤ 1, thus both sequences are
bounded.
Example 2.8. Consider the sequence ðwn Þ, whose general term is
2n þ 5
wn ¼ :
nþ1
15
By making use of Euclid’s division algorithm, we can rewrite the general term’s
formula as
3
wn ¼ 2 þ :
nþ1
Now, we observe that the sequence

1
nþ1
is strictly decreasing, bounded above by its first term and below by 0, and we proceed as
it follows:
1 1
0< ≤
nþ1 2
3 3
⇔ 0< ≤
nþ1 2
3 3
⇔ 2<2 þ ≤2 þ ,
nþ1 2
which allows us to conclude that 2 < wn ≤ 72, that is, ðwn Þ is a bounded sequence.
2.2 Arithmetic progressions
In this section we will survey a class of sequences on which each term is obtained
from its previous by adding a constant.
Definition 2.4 (Arithmetic progression). An arithmetic progression ðun Þ is a sequence
defined by recurrence as

u1 ¼ a
unþ1 ¼ un þ d, ∀n ∈ ,
where a and d are real constants.

The constant a in Definition 2.4 corresponds to the first term of the arithmetic
progression. It is immediate to observe that the difference between any two consecu-
tive terms of an arithmetic progression is constant and equal to d:
∀n ∈ , unþ1 un ¼ d
and, in view of this property, d is normally called the common difference of the
arithmetic sequence ðun Þ.
Example 2.9. Consider the sequence ðun Þ whose first terms are:
2, 5, 8, 11, 14, …
ðun Þ is the arithmetic sequence of first term 2 and common difference 3.

16
ITexLi.113863
The value of the common difference of an arithmetic sequence contains all the
information we need to analyze the sequence’s monotony. In fact, if the common
difference is positive, that means that the arithmetic sequence is strictly increasing,
since each term increases relatively to its predecessor by adding a positive constant.
Similarly, if the value of the common difference is negative, then, by an analogous
argument, the sequence is strictly decreasing. The particular case in which we have a
null common difference yields a trivial constant arithmetic sequence.
One can easily derive a formula for the general term of an arithmetic sequence.
Consider the arithmetic sequence of the first term a and common difference d. Then,
we have
u1 ¼ a,
u2 ¼ u1 þ d ¼ a þ d,
u3 ¼ u2 þ d ¼ ðk þ dÞ þ c ¼ a þ 2d,
(4)
u4 ¼ u3 þ d ¼ ðk þ 2dÞ þ c ¼ a þ 3d,
⋮
un ¼ a þ ðn 1Þ d,
which is the general term of the arithmetic sequence of first term a and common
difference d.
Example 2.10. Continuing Example 2.9, the general term of the arithmetic sequence
ðun Þ, of first term 2 and common difference 3, is
un ¼ 2 þ ðn 1Þ 3 ¼ 2 þ 3n 3 ¼ 3n 1:
The general term formula (4) can be generalized for any given term of the
sequence:
un ¼ um þ ðn mÞ d:
This means that we do not need the first term of the sequence to compute the
general term. Any other term can do the trick.
Example 2.11. Continuing Examples 2.9 and 2.10, the general term of the arithmetic
sequence ðun Þ, of first term 2 and common difference 3, can also be computed by using the
third term u3 ¼ 8:
un ¼ 8 þ ðn 3Þ 3 ¼ 8 þ 3n 9 ¼ 3n 1:
We will finish this section with a formula for calculating the sum of a finite
number of terms of an arithmetic sequence. Let ðun Þ be an arithmetic sequence of
common difference d. Then, we have
X
n
2 ui ¼ 2ðu1 þ u2 þ u3 þ ⋯ þ un2 þ un1 þ un Þ
i¼1
(5)
¼ ðu1 þ un Þ þ ðu2 þ un1 Þ þ ðu3 þ un2 Þ þ ⋯þ
þ ðun2 þ u3 Þ þ ðun1 þ u2 Þ þ ðun þ u1 Þ
17
Observe that, for any natural m we have:
um þ unðm1Þ ¼ u1 þ ðm 1Þ d þ u1 þ ðn ðm 1Þ 1Þ d
¼ u1 þ md d þ u1 þ nd md
(6)
¼ u1 þ u1 þ ðn 1Þ d
¼ u1 þ un :
Then, by making use of property (6), one can rewrite equality (5) as:
X
n
2 ui ¼ ðu1 þ un Þ n, (7)
i¼1
which, in turn, allows us to conclude that the sum, Sn , of the first n terms of an
arithmetic sequence, ðun Þ, regardless of its common difference, can be given by:
X
n
u1 þ un
Sn ¼ ui ¼ n: (8)
i¼1
2
This formula is usually associated with the German mathematician Carl Friedrich
Gauss, who allegedly surprised the teacher by computing the sum of the first
hundred natural numbers, 5050, by adding 1 to 100, multiplying the result by 100,
and dividing by 2.
Formula (8) can be generalized in a straightforward way to calculate the sum of
any consecutive terms of the sequence ðun Þ between terms um and un as:
X
n
um þ un
Sm,n ¼ ui ¼ ðn m þ 1Þ: (9)
i¼m
2
Example 2.12. Continuing Examples 2.9, 2.10, and 2.11, the sum of the ten first terms
of the arithmetic sequence ðun Þ, of first term 2 and common difference 3, is:
2 þ 3 10 1 310
S10 ¼ 10 ¼ ¼ 155:
2 2
The sum of the ten consecutive terms of ðun Þ starting from u3 is:
3 3 1 þ 3 12 1 43
S3:12 ¼ 10 ¼ 10 ¼ 215:
2 2
2.3 Geometric progressions
In this section, a different type of sequence is presented. Contrary to the arithmetic

progressions presented in Section 2.2, which have a linear growth, the sequences
presented next have an exponential growth, as each of its terms is obtained by
multiplying, not adding, its predecessor by a constant number.
18
ITexLi.113863
Definition 2.5 (Geometric progression). A geometric progression ðun Þ is a sequence

defined by recurrence as

u1 ¼ a
unþ1 ¼ un r, ∀n ∈ ,
where a and r are real constants.

The value a is the first term of the sequence. If a ¼ 0, then the geometric progres-
sion is the null sequence. Also, if a 6¼ 0 but r ¼ 0, then all the terms of the geometric
progression from the second onwards will be null. In the remaining case, that is, with
a 6¼ 0 and r 6¼ 0, all the terms of the geometric progression will not be null. In this
latter case, as an immediate consequence, we observe that the ratio between any two
consecutive terms of a geometric progression is constant and equal to r:
unþ1
∀n ∈ , ¼ r,
un
and because of this r is usually referred to as the common ratio of the geometric
progression.
Example 2.13. Consider the sequence ðun Þ of the numbers
3, 6, 12, 24, 48, …
The sequence ðun Þ is the geometric progression of first term 3 and common ratio 2.
As we did for the arithmetic progressions, we can also derive a formula for the
general term of a geometric progression. Consider the geometric sequence of the first
term a and common ratio r. Then, we have
u1 ¼ a,
u2 ¼ u1 r ¼ a r,
u3 ¼ u2 r ¼ ða rÞ r ¼ a r2 ,
(10)
u4 ¼ u3 r ¼ a r2 r ¼ a r3 ,
⋮
un ¼ a rn1 ,
which is the general term of the geometric sequence of the first term a and
common ratio r.
Example 2.14. Continuing Example 2.13, we have that the general term of the geomet-
ric progression ðun Þ of first term 3 and common ratio 2 is:
un ¼ 3 2n1 :
One can obtain the general term’s formula by making use of any term of the
geometric progression and not necessarily the first term. In fact, it is easy to obtain that:
un ¼ um rnm , (11)
for any given term of order m ∈ .

19
Example 2.15. Continuing Examples 2.13 and 2.14, we can also obtain the general
term of the geometric progression ðun Þ of first term 3 and common ratio 2, by using the term
u3 ¼ 12 and formula (11):
un ¼ 12 2n3 ¼ 3 22 2n3 ¼ 3 22þn3 ¼ 3 2n1 :
The analysis of the monotony of geometric progressions is not as straightforward

as in the case of the arithmetic progressions. Let’s exclude the trivial cases with first
term or common ratio equal to 0 which yield null sequences at least from the second
term onwards.
Proposition 2.1. Let ðun Þ be a geometric progression with u1 6¼ 0 and common ration
r 6¼ 0. The following hold:
1. If r ¼ 1, then ðun Þ is constant and all terms are equal to u1 .
2. If r > 1, then
a. if u1 > 0, then ðun Þ is increasing.
b. if u1 < 0, then ðun Þ is decreasing.
3. If 0 < r < 1, then
a. if u1 > 0, then ðun Þ is decreasing.
b. if u1 < 0, then ðun Þ is increasing.
4. If r < 0, then ðun Þ is not monotonic, as it has alternating sign terms.
The proof of Proposition 2.1 is left as an exercise for the reader.

As in the previous section, there is also a formula for calculating the sum, Sn , of n
consecutive terms of a geometric progression ðun Þ:
The case where r ¼ 1 is trivial. Since, in this case, all terms are equal to u1 , then
Sn ¼ n u1 .
Let’s consider the more general case with r 6¼ 1. We have that
X
n
Sn ¼ ui ¼ u1 þ u2 þ u3 þ ⋯ þ un1 þ un : (12)
i¼1
Multiplying both members of equality (12) by r, we obtain
Sn r ¼ u1 r þ u2 r þ u3 r þ ⋯ þ un1 r þ un r
(13)
¼ u2 þ u3 þ u4 þ ⋯ þ un þ unþ1 :
Now, subtracting equalities (12) and (13), we obtain:
20
ITexLi.113863
Sn Sn r ¼ u1 þ u2 þ u3 þ ⋯ þ un1 þ un
ðu2 þ u3 þ u4 þ ⋯ þ un þ unþ1 Þ
⇔ Sn ð1 rÞ ¼ u1 unþ1
⇔ Sn ð1 rÞ ¼ u1 un r
⇔ Sn ð1 rÞ ¼ u1 u1 rn1 r (14)
⇔ Sn ð1 rÞ ¼ u1 u1 rn
⇔ Sn ð1 rÞ ¼ u1 ð1 rn Þ
u1 ð1 rn Þ
⇔ Sn ¼ ,
1r
which gives the sum of the n first terms of the geometric progression ðun Þ with first
term u1 and common ratio r. Note that formula (14) can be easily applied to calculate
the sum of any consecutive terms of ðun Þ between um and un :
um ð1 rnmþ1 Þ
Sm:n ¼ : (15)
1r
Example 2.16. Continuing Examples 2.13, 2.14, and 2.15, we have that the sum the ten
consecutive terms of the geometric progression ðun Þ of first term 3 and common ratio 2,
starting at term u4 ¼ 24 is:

24 1 210
S4:13 ¼ ¼ 24552:
12
2.4 Convergent sequences
Some sequences have an interesting property: their terms approximate to a certain

value as the order increases. This kind of sequence is called convergent, and we will
survey it along this section.
Definition 2.6 (Limit of a sequence, convergent sequence). Let ðun Þ be a sequence.
A real number a is said to be the limit of ðun Þ if
∀ε > 0, ∃m ∈  : ∀n ∈ , n ≥ m ) ∣ un a ∣ < ε,
and we write lim n!þ∞ un ¼ a, lim n un ¼ a or simply lim un ¼ a. If a number a exists

in these conditions, ðun Þ is said to be convergent. Otherwise, ðun Þ is divergent.
When we say that lim un ¼ a the idea is that we can always find an order from
which all the terms are as close as we want, or as we previously defined, from the
value a. Therefore, ðun Þ gets closer and closer to that value a, that is, it converges to a.
Example 2.17. Consider the sequence ðun Þ of general term
2n þ 1
un ¼ :
n
Let’s prove that lim un ¼ 2.

Let ε > 0. We have
21
2n þ 1
jun 2j < ε ⇔ 2 <ε
n
2n þ 1 2n
⇔ <ε
n
1
⇔ <ε
n
1
⇔ <ε
n
1
⇔ n> :
ε
Then, for all ε > 0, there is a natural number m such that for all n ∈ , we have
n ≥ m ) ∣ un a ∣ < ε. That number m can be chosen as any natural number bigger that
1=ε. Hence, lim un ¼ 2.
An expected property of the definition of limit of a sequence is that when the limit
exists, it is unique; that is, a convergent sequence cannot have two distinct limits.
There are two particular cases of divergent sequences that should be highlighted.
Definition 2.7 (Infinite limit of a sequence). Let ðun Þ be a sequence.
1. It is said that ðun Þ has limit þ∞, and we write lim un ¼ þ∞, if
∀L > 0, ∃m ∈  : ∀n ∈ , n ≥ m ) un > L:
2. It is said that ðun Þ has limit ∞, and we write lim un ¼ ∞, if
∀L > 0, ∃m ∈  : ∀n ∈ , n ≥ m ) un < L:
Example 2.18. Let ðun Þ and ðvn Þ be the sequences defined by un ¼ 3n þ 2 and
vn ¼ 3 2n, respectively. Let’s prove that lim un ¼ þ∞ and lim vn ¼ ∞.
Let L > 0.
Regarding ðun Þ, we have
un > L ⇔ 3n þ 2 > L
L2
⇔ n> :
3
Then, there is a order m ∈  such that, for all n ∈ , we have
n ≥ m ) un > L:
We just have to choose m to be a natural number bigger that ðL 2Þ=3. Hence,

lim un ¼ þ∞.
Regarding ðvn Þ, we have
vn < L ⇔ 3 2n < L
Lþ3
⇔ n> :
2
22
ITexLi.113863
Then, there is a order m ∈  such that, for all n ∈ , we have
n ≥ m ) vn < L:
In this case, it is enough to choose m to be a natural number bigger that ðL þ 3Þ=2.

Hence, lim vn ¼ ∞.
There are some interesting and well-known results involving the convergence of
sequences.
Theorem 2.1. Let X be a finite subset of  and ðun Þ and ðvn Þ be two sequences such that,
for all n ∈ nX, we have un ¼ vn . Then, ðun Þ is convergent if and only if ðvn Þ is convergent.
Also, lim un ¼ lim vn .
Theorem 2.1 basically states that two different sequences that differ only in a finite
number of terms converge or diverge in the same manner. We leave the proof as an
exercise for the reader.
pffiffiffi
Example 2.19. Let ðun Þ be the sequence whose general term is un ¼ 2 n and ðvn Þ the
sequence defined by:
8
<1
þ 2, if n < 25
vn ¼ n :
:
un , if n ≥ 25
Since ðun Þ and ðvn Þ only differ in a finite number of terms (the first 24 terms), then
lim vn ¼ lim un ¼ 2 ∞ ¼ ∞.
As it is implied in the resolution of Example 2.19, let it be clear that the first step in
evaluating a sequence’s limit is to replace every instance of n, in the sequence’s general
term, with þ∞ and calculate. Operating with ∞ is, therefore, of great importance. We
will not elaborate on this topic for now, as this section deals with a more theoretical
part of convergence, but it will be one of the main topics surveyed in Section 2.5.
Theorem 2.2. Let ðun Þ be a sequence. If ðun Þ is convergent, then ðun Þ is bounded.
Proof. Let ðun Þ be a convergent sequence and suppose that lim un ¼ a: Then, by
definition, for any fixed ε > 0, there is an order m ∈  such that ∣ un a ∣ < ε, for all
n ≥ m. This means that the set of terms of ðun Þ with n > m belongs to V a , the neigh-
borhood of a. We observe that a þ ε and a ε constitute an upper bound and a lower
bound of that set of terms of ðun Þ, respectively; thus, that set is bounded.
As for the remaining terms of ðun Þ, that is, the terms with n < m they constitute a
finite set that is naturally bounded. We can, therefore, conclude that all terms of ðun Þ
are bounded, hence ðun Þ is a bounded sequence. □
Observe that the converse from Theorem 2.2 is not true. Consider the trivial
alternating sequence, that is, a sequence whose terms alternate between positive and
negative, with general term ð1Þn . It is clearly a bounded sequence, but it does not
converge as it keeps alternating from 1 to 1.
Theorem 2.3 (Squeeze Theorem for Convergent Sequences). Let ðun Þ, ðvn Þ and
ðwn Þ be sequences such that
∃m ∈  : ∀n ∈ , n > m ) un ≤ vn ≤ wn :
If lim un ¼ lim wn ¼ a, where a ∈ , then lim vn ¼ a.

Proof. Let ε > 0. Then, since lim un ¼ lim wn ¼ a, there are natural numbers m1
and m2 such that
23
n ≥ m1 ) ∣ un a ∣ < ε ⇔ a ε < un < a þ ε,

n ≥ m2 ) ∣ wn a ∣ < ε ⇔ a ε < wn < a þ ε:
Then, considering m ¼ maxfm1 , m2 g, we have
n ≥ m ) a ε < un ≤ vn ≤ wn < a þ ε,
which means that
n ≥ m ) ∣ vn a ∣ < ε:
Hence, lim vn ¼ a. □
Note that the result of Theorem 2.3 can be straightforwardly extended to the cases
where a ¼ ∞.
Example 2.20. Let ðun Þ be the sequence whose general term is
sin n
un ¼ :
n
Since the sine yields values in the interval [1, 1], we can write
1 sin n 1
≤ ≤ :
n n n

Now, observing that lim n1 ¼ lim n1 ¼ 0, we can conclude, by Theorem 2.3, that
lim un ¼ 0.
From Theorem 2.3, one can prove the following result.
Theorem 2.4. Let ðun Þ be a sequence such that lim∣ un ∣ ¼ 0. Then, lim un ¼ 0.
Theorem 2.4 is especially useful for computing certain limits involving alternating
sequences.
Example 2.21. Let ðun Þ be the sequence defined by
ð1Þn
un ¼ :
2n þ 1
Since
1 1
lim∣un ∣ ¼ lim ¼ ¼ 0,
2n þ 1 þ∞
we can conclude, by Theorem 2.4, that lim un ¼ 0.

Theorem 2.5. Let ðun Þ and ðvn Þ be two convergent sequences such that
lim un < lim vn :
Then, there is a m ∈  such that
24
ITexLi.113863
n ≥ m ) un < vn :
Proof. Let ðun Þ and ðvn Þ be two convergent sequences with lim un ¼ a and lim vn ¼
b such that a < b. Let ε ¼ b2 a. Note that . By the definition of limit, there are natural
numbers m1 and m2 such that
n ≥ m1 ) ∣ un a ∣ < ε⇔a ε < un < a þ ε,
n ≥ m2 ) ∣ vn b ∣ < ε⇔b ε < vn < b þ ε:
Now, observing that a þ ε ¼ b ε ¼ aþ2 b, and taking m ¼ maxfm1 , m2 g, we have

that
n > m ) un < a þ ε ¼ b ε < vn :
□
Theorem 2.6. Let ðun Þ be a sequence.
1. If ðun Þ is increasing and bounded from above, then it is convergent.
2. If ðun Þ is decreasing and bounded from below, then it is convergent.
3. If ðun Þ is monotonic and bounded, then it is convergent.
Proof. We will just present proof for assertion 1 from Theorem 2.6. Assertion 2 is
analogous, and assertion 3 is a direct consequence from the previous two.
Let ðun Þ be an increasing sequence bounded from above. We can immediately
conclude that ðun Þ is bounded, that is, that the set of terms of ðun Þ is a bounded set.
Since it is bounded and it is a non-empty set, then we can conclude that it has a
supremum. Let a be the supremum of the set of terms of ðun Þ. We will prove that
lim un ¼ a.
Let ε > 0. By definition of supremum, we have that, for all n ∈ , un ≤ a. We
also have that there is at least one m ∈  such that um > a ε: Since ðun Þ is an
increasing sequence, we also must have that, for all n > m, un ≥ um > a ε:
Therefore, we can conclude that there is an order m ∈  such that, for all
n ∈ , we have ∣un a ∣ < ε, if n > m. Thus, lim un ¼ a and ðun Þ is a convergent
sequence. □
The result presented above has direct application as it is shown in the following
example.
Example 2.22. Consider the sequence ðun Þ with general term
2n 1
un ¼ :
n
Let’s show that ðun Þ is convergent.

Firstly, since
25
2ðn þ 1Þ 1 2n 1
unþ1 un ¼
nþ1 n
2n þ 1 2n 1
¼
nþ1 n
ð2n þ 1Þn ð2n 1Þðn þ 1Þ
¼
nðn þ 1Þ
1
¼
nðn þ 1Þ
> 0,
then ðun Þ is increasing.

Now, observe that un ¼ 2 1=n and that
1
0< ≤1
n
1
⇔ 1 ≤ <0
n
1
⇔ 1≤2 < 2:
n
Hence, ðun Þ is bounded. Then, from Theorem 2.6, we can conclude that ðun Þ is
convergent.
Note that being convergent does not imply being monotonic. For instance, the
sequence with general term ð1Þn =n is convergent to 0, but it is not monotonic as it is
an alternating sequence.
Theorem 2.7. Let ðun Þ and ðvn Þ be two sequences such that un is bounded and
lim vn ¼ 0. Then lim ðun vn Þ ¼ 0:
Proof. Let ðun Þ be a bounded sequence and ðvn Þ be a sequence such that lim vn ¼ 0.
Let wn ¼ un vn and ε > 0.
Since ðun Þ is bounded, then there is a k ∈  such that ∣un ∣ ≤ k, for all n ∈ . Then,
we can write
∣ wn ∣ ¼ ∣ un vn ∣ ¼ ∣ un ∣ ∣ vn ∣ ≤ k ∣ vn ∣:
Now, since lim vn ¼ 0, there is an order m ∈  such that, for all n > m, vn < ε=k:
Then, we can conclude that, for all n > m, we have
∣ wn ∣ ≤ k ∣ vn ∣ < ε,
that is, lim wn ¼ 0: □

Some potentially complicated limits can be computed using the result presented in
Theorem 2.7.
Example 2.23. Let’s compute the value of
cosðn 2Þ
lim :
nþ1
26
ITexLi.113863
Observe that
cosðn 2Þ 1
¼ cosðn 2Þ :
nþ1 nþ1
The sequence whose general term is cosðn 2Þ is clearly bounded, as the cosine yields
values in the interval [1, 1].
As for the sequence whose general term is 1=ðn þ 1Þ, we have that
1 1
lim ¼ ¼ 0:
n þ 1 þ∞
Then, we can conclude, by Theorem 2.7, that
cosðn 2Þ
lim ¼0
nþ1
In the following, we present a set of operating properties regarding limits of

sequences.
Proposition 2.2. Let ðun Þ and ðvn Þ be convergent sequences such that lim un ¼ a and
lim vn ¼ b: Let α ∈  and k ∈ . We have:
1. lim ðun þ vn Þ ¼ a þ b;
2. lim ðun vn Þ ¼ a b;
3. lim ðun vn Þ ¼ ab;
4. lim ðαun Þ ¼ αa;
5. lim ðun Þk ¼ ak ;
6. lim∣ un ∣ ¼ ∣ a ∣;

7. lim 1
un ¼ a1, if un 6¼ 0, ∀n ∈  and a 6¼ 0;

8. lim un
vn ¼ ab, if vn 6¼ 0, ∀n ∈  and b 6¼ 0;
pffiffiffiffiffi pffiffiffi
9. lim k un ¼ k a, with if un ≥ 0, ∀n ∈ , if k is an even number.
We finish this survey on sequences by presenting a very important class of

sequences called the Cauchy sequences, named after the French mathematician
Augustin-Louis Cauchy. Their relevance and their implications on convergence make
them frequently referred to as fundamental sequences.
Definition 2.8 (Cauchy sequence). The sequence ðun Þ is a Cauchy sequence if
∀ε > 0, ∃p ∈  : ∀m, n ∈ , m, n > p ) ∣ um un ∣ < ϵ:
27
Observe that the above formal definition simply states that the terms of a Cauchy
sequence become as close as we want as the sequence progresses.
Theorem 2.8. Let ðun Þ be a Cauchy sequence. Then ðun Þ is bounded.
Proof. Let ðun Þ be a Cauchy sequence. By definition, for any fixed ε > 0, there is
an order p ∈  such that ∣ um un ∣ < ε, for all natural numbers m, n > p. In
particular, if n ¼ p þ 1, we have that ∣ um upþ1 ∣ < ε, for all m > p, which
means that the set of terms of ðun Þ with n > p is bounded as ε þ upþ1 is a
lower bound and ε þ upþ1 is an upper bound for this set. As for the set of terms of
ðun Þ with n < p, since it is a finite set, it is also bounded. Hence, ðun Þ is a bounded
sequence.
□
Using Theorem 2.8, one can prove the following result that states that being a
convergent sequence and being a Cauchy sequence are the same thing.
Theorem 2.9. Let ðun Þ be a sequence. Then, ðun Þ is convergent if and only if ðun Þ is a
Cauchy sequence.
2.5 Techniques for evaluating limits
In this section, we will present a set of relevant tools for computing the limits of
certain sequences. We will also analyze the main indeterminate forms that one can
attain when computing limits, and we will discuss techniques to evaluate these inde-
terminate forms.
We will start this section with a result that can be of extreme importance in all
domains of sequence analysis.
Theorem 2.10. Let ðun Þ be a sequence and f a real function such that,
∀n ∈ , f ðnÞ ¼ un . Then we have:
1. If lim f ðxÞ ¼ a, for a ∈ , then lim un ¼ a.

x!þ∞
2. If f is increasing (decreasing), then ðun Þ is increasing (decreasing).
This result allows us to apply functional analysis to solve some problems involving
sequences by considering a corresponding real function.
Example 2.24. Calculating the limit of the sequence ðun Þ defined by
3n
un ¼
lnðnÞ þ 2n
may prove to be difficult at first glance, but if one considers the real function f
defined by
3x
f ðxÞ ¼ ,
lnðxÞ þ 2x
then we obtain, by applying the well-known L’Hôpital’s Rule,
28
ITexLi.113863
3x
lim f ðxÞ ¼ lim
x!þ∞ lnðxÞ þ 2x
x!þ∞
∞
d
ð3xÞ
∞
¼ lim dx
x!þ∞ d
ðlnðxÞ þ 2xÞ
dx
3
¼ lim
x!þ∞ 1
þ2
x
3
¼
0þ2
3
¼ :
2
By Theorem 2.10, we can conclude that lim un ¼ 3=2.

Example 2.24 brings out again the need to clarify operations involving ∞. The
following proposition presents a list of relevant properties. Before, however, let us
introduce the following notation.
Let ðun Þ be a convergent sequence with limit equal to a. If ðun Þ converges to the
value a through values which are strictly smaller than a, then we write lim un ¼ a . If
ðun Þ converges to the value a through values which are strictly bigger than a, then we
write lim un ¼ aþ .
As a trivial example, consider the sequences ðun Þ and ðvn Þ whose general terms are
un ¼ 1=n and vn ¼ 1=n: It is easy to verify that lim un ¼ 0þ and lim vn ¼ 0 .
Proposition 2.3. While evaluating a limit, the following equalities hold.
1. a ∞ ¼ ∞, ∀a ∈ ;
2. þ∞ þ ∞ ¼ þ∞;
3. ∞ ∞ ¼ ∞;
4. a ð∞Þ ¼ ∞, ∀a ∈ þ or a ¼ þ∞;
5. a ð∞Þ ¼ ∓∞, ∀a ∈  or a ¼ ∞;
a
6. ∞ ¼ 0, ∀a ∈ ;
a. a
þ∞ ¼ 0þ , ∀a ∈ þ ;
b. a
∞ ¼ 0 , ∀a ∈ þ ;
c. a
þ∞ ¼ 0 , ∀a ∈  ;
d. a
∞ ¼ 0þ , ∀a ∈  ;
7. 0a ¼ ∞, ∀a ∈ nf0g;
29
a. a
0þ
¼ þ∞, ∀a ∈ þ or a ¼ þ∞;
b. a
0þ
¼ ∞, ∀a ∈  or a ¼ ∞;
c. a
0 ¼ ∞, ∀a ∈ þ or a ¼ þ∞;
d. a
0 ¼ þ∞, ∀a ∈  or a ¼ ∞;
8. ð∞Þk ¼ ∞, ∀k ∈ ;
a. ðþ∞Þk ¼ þ∞, ∀k ∈ ;
b. ð∞Þk ¼ þ∞, ∀k ∈  and k even;
c. ð∞Þk ¼ ∞, ∀k ∈  and k odd;

pffiffiffiffiffiffiffiffi
9. k þ∞ ¼ þ∞, ∀k ∈ nf1g;
10. k ∞ ¼ ∞, ∀k ∈ nf1g and k is odd.
The properties presented are invaluable to operate with ∞ and compute the values
of limits by direct substitution of the variable n in the general term by þ∞. The
following example presents some cases.
Example 2.25. Let’s compute the value of the following limits.
1. limð2n 5Þ
limð2n 5Þ ¼ 2 ðþ∞Þ 5
¼ þ∞ 5
¼ þ∞:
2
2. lim 33n
2 2
lim ¼
3 3n 3 3 ðþ∞Þ
2
¼
3∞
2
¼
∞
¼ 0:
3. lim½n2 ð1 nÞ
lim n2 ð1 nÞ ¼ ðþ∞Þ2 ½1 ðþ∞Þ
¼ þ∞ ð1 ∞Þ
¼ þ∞ ð∞Þ
¼ ∞:
30
ITexLi.113863
Left out from the list presented in Proposition 2.3 are the cases of
∞
þ∞ ∞, ∞ ∞ , 0 , 0 ∞, ∞ , 0 and 1 . These are called indeterminate forms as they can
0 0 0
actually be equal to different real values or ∞ depending on the situation at hand.

We will observe this particular behavior in the following paragraphs, on which we
will survey some of the techniques one can apply to evaluate different types of
indeterminate forms.
2.5.1 Indeterminate form ∞ ∞
This type of indeterminate form, ∞ ∞, usually arises from evaluating the limit of
sequences constructed in the form of a polynomial of variable n, although it can also
appear in other contexts. For example, if one attempts to evaluate the limit of the
sequence ðun Þ whose general term is un ¼ n2 2n þ 1 by direct substitution, one will
obtain this indeterminate form.
In the context of sequences defined by polynomials, this difficulty is very easy to
surpass. Indeed, one needs only to collect the higher degree term in ðun Þ and substitute
again. The indeterminate form will disappear.
Example 2.26. Consider the sequence ðun Þ with general term un ¼ n2 2n þ 1.
We have:

lim un ¼ lim n2 2n þ 1

ð∞∞Þ 2 2 1
¼ n 1 þ 2
n n

2 1
¼ þ∞ 1 þ
þ∞ þ∞
¼ þ∞ð1 0 þ 0Þ
¼ þ∞:
Example 2.26 leads to a more general property.

Proposition 2.4. Let p(x) be a non-null polynomial with leading coefficient a and ðun Þ
be a sequence such that, ∀n ∈ , un ¼ pðnÞ. Then, the following statements hold.
1. If a > 0, then lim un ¼ þ∞;
2. If a < 0, then lim un ¼ ∞.
When defining sequences with the help of the concept of a polynomial, it is

implied that the degree of the powers of n are greater or equal than 0. The next result
lists clarifies all possibilities regarding the limits of rational powers of n.
Proposition 2.5. Let q ∈  and ðun Þ be the sequence whose general term is un ¼ nq .
Then, the following hold.
1. If q > 0, then lim un ¼ þ∞;
2. If q ¼ 0, then lim un ¼ 1;
3. If q < 0, then lim un ¼ 0.
31
Proof. Let q ∈ .
1. If q > 0, let L > 0. Then, we have
un > L ⇔ nq > L
p ffiffiffi
⇔ n > L:
q
p
q
ffiffiffi
Therefore, choosing any natural m bigger than L, we have that
∀n ∈ , n ≥ m ) un > L,
which is the same as saying that lim un ¼ þ∞.
2. If q ¼ 0, then un ¼ 1, ∀n ∈ . Hence, lim un ¼ 1.
3. If q < 0, let ε > 0. Then, we have
∣ un 0 ∣ < ε ⇔ ∣ nq ∣ < ε
1
⇔ q < ε
n
1
⇔ q < ε
n
1
⇔ nq >
pεffiffiffi
⇔ n > q ε:
pffiffiffi
Therefore, choosing any natural m bigger than q ε, we have that
∀n ∈ , n ≥ m ) ∣ un 0 ∣ > ε,
which is the same as saying that lim un ¼ 0.

□
The next example presents another occasion where the indeterminate form appears.
pffiffiffiffiffiffiffiffiffiffiffi pffiffiffi
Example 2.27. Consider the sequence ðun Þ with general term un ¼ n þ 1 n. The
limit of ðun Þ can be obtained by making use of the well-known multiplication identity
ðA BÞðA þ BÞ ¼ A2 B2 in the following manner:
lim un ¼ lim n þ 1 n
pffiffiffiffiffiffiffiffiffiffiffi pffiffiffipffiffiffiffiffiffiffiffiffiffiffi pffiffiffi
ð∞∞Þ nþ1 n nþ1þ n
¼ lim pffiffiffiffiffiffiffiffiffiffiffi pffiffiffi
nþ1þ n
nþ1n
nþ1þ n
1
nþ1þ n
1
¼
þ∞ þ ∞
1
¼
þ∞
¼ 0:
32
ITexLi.113863
We will finish this subsection by presenting one type of sequence for which the
limit evaluation might result in indeterminate forms, particularly the one that is under
the scope in these paragraphs.
Proposition 2.6. Let a ∈ þ and ðun Þ be the sequence whose general term is un ¼ an .
Then, the following statements hold.
1. If a > 1, then lim un ¼ þ∞;
2. If a ¼ 1, then lim un ¼ 1;
3. If 0 < a < 1, then lim un ¼ 0.
Proof. Let a ∈ þ .
1. If a > 1, let L > 0. We have
un > L ⇔ an > L
⇔ n > log a L:
Then, choosing any natural number bigger than log a L, we have that
∀n ∈ , n ≥ m ) un > L,
which means that lim un ¼ þ∞.
2. If a ¼ 1, then all terms of un are equal to 1. Hence, lim un ¼ 1
3. If 0 < a < 1, then a1 > 1, which allows us to write
lim un ¼ lim an
1
¼ lim
ða1 Þn
1
¼
lim ða1 Þn
1
¼
þ∞
¼ 0:
□
In the following, we present an example of an indeterminate form
∞ ∞ that arises from subtracting two sequences of the form presented in
Proposition 2.6.
Example 2.28. Let ðun Þ be the sequence defined by un ¼ 5n 2n .
Computing the limit of ðun Þ gives us, by Proposition 2.6, the indeterminate form ∞ ∞.
We can evaluate this limit by rewriting the general term, collecting the biggest power, and
applying Proposition 2.6:
33
lim un ¼ lim ð5n 2n Þ

ð∞∞Þ 2n
¼ lim 5 1 n
n
5
n
2
¼ lim 5n 1
5
¼ þ∞ð1 0Þ
¼ þ∞:
∞
2.5.2 Indeterminate form ∞
The type of indeterminate form represented by ∞ ∞ can appear when working out
limits of sequences whose general term is a division of polynomials in the variable n,
but it is not restricted to these cases. Trivial examples, which can be solved just by
simplifying the fraction, are presented next.
Example 2.29. Let ðun Þ, ðvn Þ and ðwn Þ be the sequences whose general terms are
n2 n n
un ¼ , vn ¼ and wn ¼ ,
3n 3n 3n2
respectively.
By simplifying the general terms, we obtain
n2
lim un ¼ lim
3n
∞
∞
¼ lim
n
3
þ∞
¼
3
¼ þ∞,
n
lim vn ¼ lim
3n
∞
∞
¼ lim
1
3
1
¼ ,
3
n
lim wn ¼ lim 2
3n
∞
∞
¼ lim
1
3n
1
¼
3 ðþ∞Þ
1
¼
þ∞
¼ 0:
Example 2.29 enlightens in a very simple way the concept of indeterminate form.
Observe how the three cases of the indeterminate form ∞
∞, after simplification, yield
three different outcomes.
34
ITexLi.113863
Some cases, however, cannot be solved through direct simplification only. Simi-
larly to what we did in the previous section, collecting the higher degree terms of
polynomials can be useful before simplifying. In the following example, three cases
are presented using this technique.
Example 2.30. Consider the sequence ðun Þ, ðvn Þ and ðwn Þ with general terms defined by
2n3 n2 þ 1 3n2 þ n 2 3 2n
un ¼ , vn ¼ and wn ¼ ,
n3 n2 3n þ 1 n2 þ n þ 1
respectively.
In each case, direct substitution ends with the indeterminate form ∞
∞. The idea is to collect
the term with a higher degree in the fraction’s numerator and repeat the process with the
denominator. Then, simplify and, finally, substitute. We have:
2n3 n2 þ 1
lim un ¼ lim
n3
∞
n2 1
2n 1 3 þ 3
3
∞
¼ lim
2n

2n
3
n 1
n

1 1
2n2 1 þ 3
2n 2n
¼ lim
3
1
n
1 1
þ∞ 1 þ
þ∞ þ∞
¼
3
1
þ∞
þ∞ð1 0 þ 0Þ
¼
10
¼ þ∞,
3n2 þ n 2
lim vn ¼ lim
n2 3n þ 1
∞
n 2
3n2 1 þ 2 2
∞
¼ lim
3n 3n

3n 1
n 1
2 þ
n2 n2

1 2
3 1 þ 2
3n 3n
¼ lim
3 1
1þ 2
n n
1 2
3 1 þ
þ∞ þ∞
¼
3 1
1þ
þ∞ þ∞
3ð1 þ 0 0Þ
¼
1þ00
¼ 3,
35
3 2n
lim wn ¼ lim
n2þnþ1
∞
3
2n þ1
∞ 2n
¼ lim
n 1
n 1þ 2þ 2
2
n n

3
2 þ 1
2n
¼ lim
1 1
n 1þ þ 2
n n

3
2 þ1
þ∞
¼
1 1
þ∞ 1 þ þ
þ∞ þ∞
2ð0 þ 1Þ
¼
þ∞ð1 þ 0 þ 0Þ
2
¼
þ∞
¼ 0:
The three different cases presented in Example 2.30 allow us to straightforwardly

establish the following more general result.
Proposition 2.7. Let p(x) and q(x) be two non-null polynomials with degrees k and l
and leading coefficients a and b, respectively. Let ðun Þ be the sequence defined by
pðnÞ
un ¼ , ∀n ∈ :
qðnÞ
Then, the following statements hold.
1. If k > l and ab > 0, then lim un ¼ þ∞;
2. If k > l and ab < 0, then lim un ¼ ∞;
3. If k ¼ l, then lim un ¼ ab ;
4. If k < l, then lim un ¼ 0.
Next, we present two examples where an indeterminate form ∞ ∞ appears when

evaluating the limit of sequences defined with radicals.
Example 2.31. Let ðun Þ and ðvn Þ be the sequences defined by:
pffiffiffi
n2 2n þ 1
un ¼ pffiffiffi and vn ¼ pffiffiffiffiffiffiffiffiffiffiffi :
2 nþ1 nþ2
When evaluating the limit either of un or vn by direct substitution we come upon the
indeterminate form ∞∞. In each case, we will solve the indeterminate form by collecting terms
and rewriting the fraction. Indeed, we have
36
ITexLi.113863
pffiffiffi
n2
lim un ¼ lim pffiffiffi
2 nþ1
∞
pffiffiffi 2
n 1 pffiffiffi
∞
¼ lim
n

pffiffiffi 1
n 2 þ pffiffiffi
n
2
1 pffiffiffi
n
¼ lim
1
2 þ pffiffiffi
n
2
1
¼ þ∞
1
2þ
þ∞
10
¼
2þ0
1
¼ ,
2
as well as
2n þ 1
lim vn ¼ lim pffiffiffiffiffiffiffiffiffiffiffi
nþ2
∞
1
2n 1 þ
∞ 2n
¼ lim sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ffi
n 2
n 2 þ
n2 n2

1
2n 1 þ
2n
¼ lim rffiffiffiffiffiffiffiffiffiffiffiffiffi
1 2
n þ
n n2

1
2 1þ
2n
¼ lim rffiffiffiffiffiffiffiffiffiffiffiffiffi
1 2
þ
n n2

1
2 1þ
þ∞
¼ rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
1 2
þ
þ∞ þ∞
2ð1 þ 0Þ
¼ pffiffiffiffiffiffiffiffiffiffiffiffi
0þ0
2
¼ þ
0
¼ þ∞:
37
Note that the square root will always attain 0 through values bigger than 0 as it is not
defined in  .
In the following, we present an example of an indeterminate form ∞ ∞ that arises
from dividing sequences of the form presented in Proposition 2.6.
Example 2.32. Consider the sequence ðun Þ with general term given by
4n
un ¼ :
2n þ 3n
Clearly, evaluating the limit of ðun Þ by simple substitution results in the indeterminate
form ∞∞ : In a similar manner than that of Example 2.28, we will solve this problem by
collecting the higher power of the numerator and denominator like it follows.
4n
lim un ¼ lim n
2 þ 3n
∞
∞ 4n
¼ lim
2n
3n þ 1
3n
" #
4n 1
¼ lim n 2n
3 3 þ1
" #
n
4 1
¼ lim 2n
3 3 þ1

1
¼ þ∞
0þ1
¼ þ∞
Example 2.33. Let ðvn Þ be the sequence defined by

n
vn ¼ :
7n
Attempting to evaluate the limit of ðvn Þ directly gives us the indeterminate form ∞
∞.
Applying L’Hôpital’s Rule, to the function f, such that
x
f ðx Þ ¼ ,
7x
allows us to obtain the limit of ðvn Þ: Indeed,
x
lim f ðxÞ ¼ lim x
x!þ∞ x!þ∞ 7
∞
d
ðxÞ
∞
¼ lim dx
x!þ∞ d
ð7 x Þ
dx
1
¼ lim
x!þ∞ 7 x ln 7
1
¼
þ∞
¼ 0:
Hence, by Theorem 2.10, lim vn ¼ 0.

38
ITexLi.113863
2.5.3 Indeterminate form 00
While working out the limits of sequences, the indeterminate form 00 arises mainly
when dividing two sequences that converge to 0 and can be solved by simplifying the
expression at hand, eventually obtaining a ∞
∞ indeterminate form and applying some
technique surveyed in the previous subsection.
Example 2.34. Consider the sequences ðun Þ and ðvn Þ with general terms given by
1 3
un ¼ and vn ¼ :
nþ1 n2 2
We have that
1
lim
un
¼ lim n þ 1
vn 3
n2 2

0
0 n2 2
¼ lim
3ð n þ 1 Þ
∞
2
n2 1 2
∞
¼ lim
n

1
3n 1 þ
n

2
n 1 2
n
¼ lim
1
3 1þ
n

2
þ∞ 1
þ∞
¼
1
3 1þ
þ∞
þ∞ð1 0Þ
¼
3ð1 þ 0Þ
¼ þ∞:
Example 2.35. Let ðwn Þ be the sequence defined by

ln 1 n12
wn ¼ :
ln 1 n22
If we try to evaluate the limit of ðwn Þ by direct substitution, we obtain the indeterminate
form 00.
Consider the function f such that
39

ln 1 x12
f ðxÞ ¼ :
ln 1 x22
Applying L’Hôpital’s Rule, we have

1
ln 1 2
x
lim f ðxÞ ¼ lim
x!þ∞ x!þ∞ 2
ln 1 2
x

0 d 1
ln 1 2
0 dx x
¼ lim
x!þ∞ d 2
ln 1 2
dx x
2
x3
1
1
¼ lim x2
x!þ∞ 4
x3
2
1
x2
2
1 2
¼ lim x
x!þ∞ 1
2 1 2
x
10
¼
2ð1 0Þ
1
¼ :
2
Hence, by Theorem 2.10, lim wn ¼ 21.
2.5.4 Indeterminate form 0 ∞
The case of the indeterminate form is solved using the techniques presented in the
two previous subsections as it can always be converted into either the indeterminate
form ∞ 0
∞ or 0.
Example 2.36. Let ðun Þ, ðvn Þ and ðwn Þ be the sequences defined by
n
3 1
un ¼ , vn ¼ and wn ¼ n 2:
n 2
We have:
40
ITexLi.113863
3
limðun wn Þ ¼ lim ðn 2 Þ
n
ð0∞Þ 3n 6
¼ lim
n

∞ 6
3n 1
∞
¼ lim
3n
n
2
¼ lim 3 1
n

2
¼ 3 1
þ∞
¼ 3ð 1 0 Þ
¼ 3:
Also,
n
1
limðvn wn Þ ¼ lim ðn 2Þ
2
ð0∞Þ n2
¼ lim n
2
∞

∞ n
¼ lim n n
2
2 2 (16)
n
¼ lim n lim 21n
2
n
¼ lim n 0
2
n
¼ lim n :
2
Now consider the real function f defined by

x
f ðxÞ ¼ ,
2x
applying L’Hôpital’s Rule, we have

x
lim f ðxÞ ¼ lim x
x!þ∞ x!þ∞ 2
∞
∞
¼ lim
1
x!þ∞ 2x ln 2
1
¼ lim
x!þ∞ þ∞
¼ 0:
Therefore, we can conclude, by Theorem 2.10, that (16) is equal to 0.

Example 2.37. Let ðun Þ and ðvn Þ be the sequences defined by
41
1
un ¼ lnðnÞ and vn ¼ :
n
We have that
limðun vn Þ ¼ limðlnðnÞ 1=nÞ
ð0∞Þ lnðnÞ (17)
¼ lim :
n
∞
The limit (17) results in the indeterminate form ∞. As we did in previous examples, let
us consider the function f such that
lnðxÞ
f ðxÞ ¼ :
x
Applying L’Hôpital’s Rule, we obtain
lnðxÞ
lim f ðxÞ ¼ lim
x!þ∞ x!þ∞ x
∞
d
ðlnðxÞÞ
∞
¼ lim dx
x!þ∞ d
ðx Þ
dx
1
¼ lim x
x!þ∞ 1
1
¼ lim
x!þ∞ x
1
¼
þ∞
¼ 0:
Then, by Theorem 2.10, we can conclude that limðun vn Þ ¼ 0:
2.5.5 Indeterminate forms 00 , ∞0 and 1∞
In this subsection, we present some examples for each of the remaining indeter-
minate forms, 00 , ∞0 , and 1∞ , which were not addressed before. We apply the natural
logarithm to the expressions to solve the indeterminate forms.
Example 2.38. In this example, we will calculate the value of
n1
1
lim :
n
We have:
n1 0
1 ð0 Þ 1
¼ lim eln ðnÞ

1 n
lim
n
h 1 1 i
¼ lim enlnðnÞ
h 1 i
¼ lim enð lnðnÞÞ
¼ e lim½
lnðnÞ
n
¼ e ¼10
42
ITexLi.113863
using Example 2.37.

1
lim nn :
We have:

ð∞0 Þ
1 1
lim n ¼ lim eln n
n
n
h 1 i
¼ lim enlnðnÞ
h lnðnÞ i
¼ lim e n
¼ e lim½
lnðnÞ
n
¼ e0 ¼ 1
using Example 2.37.


1 n
lim 1 þ 2 :
n
We have:
n
1 n ð1∞ Þ 1
lim 1 þ 2 ¼ lim eln 1þn2
n

n ln 1þ 12
¼ lim e n :
Considering the real function f, defined by

1
f ðxÞ ¼ x ln 1 þ 2 ,
x
we have

1
lim f ðxÞ ¼ lim xln 1 þ 2
x!þ∞ x!þ∞ x

ð∞0Þ 1
¼ lim ln 1 þ y2
y!0 y

0 d
ln 1 þ y2
0 dy
¼ lim
y!0 d
ð yÞ
dy
2y
1 þ y2
¼ lim
y!0 1
¼ 0,
43
changing variable y ¼ 1=x and applying l’Hôpital’s Rule.

Hence, we conclude that the limit at hand is equal to e0 ¼ 1.
We finish this subsection with two results containing two notable cases for solving
indeterminations treated here, 1∞ and ∞0 .
Proposition 2.8. Let ðun Þ be a sequence of real numbers such that un 6¼ 0 for n ≥ m,
with m ∈ , converging to ∞ and a ∈ . Then,

a un
lim 1 þ ¼ ea :
un
Proposition 2.9. Let a ∈ þ and n ∈ . Then,

pffiffiffi
1. lim n a ¼ 1;
pffiffiffi
2. lim n n ¼ 1.
2.6 Suggested exercises
General properties of sequences
1. Let ðun Þ be the sequence of real numbers 0, 2, 6, 12, 20, 30, …
a. Find the general term of ðun Þ:
b. Verify that 156 is a term of ðun Þ: Find its order.
c. Solve the condition 400 < un < 600.
2. Let ðvn Þ be the sequence whose general term is given by:
6n þ 8
vn ¼ :
3n þ 2
a. Find the first five terms of vn .
b. Verify if 49 is a term of vn .
c. Prove that ðvn Þ is a decreasing sequence.
d. Show that ðvn Þ is bounded. Determine a lower bound and an upper bound
of ðvn Þ.
e. Determine the order from which vn is not greater than 2,1.
3. Consider the sequence ðwn Þ whose first five terms are:
1 8 27 64 125
; ; ; ;
4 9 16 25 36
a. Find the general term of ðwn Þ:
44
ITexLi.113863
b. Determine w10 .
c. Show that ðwn Þ is unbounded.
4. Let ðun Þ be the sequence defined by recurrence as

u1 ¼ 5
:
unþ1 ¼ 3un þ 10, ∀n ∈ 
a. Find the first four terms of ðun Þ:
b. Using mathematical induction, show that
5
un ¼ ð2 3n 3Þ, ∀n ∈ :
3
c. Show that ðun Þ is monotonic.
d. Prove that ðun Þ is not bounded from above.
5. Let ðvn Þ be the sequence defined by

nπ
vn ¼ sin :
4
a. Find the image of ðvn Þ:
b. Justify that ðvn Þ is bounded.
c. Justify that ðvn Þ is not monotonic.
6. Let ðwn Þ be the sequence defined by

8
< n 2, if n ≤ 4
wn ¼ 5n 4 :
: , if n > 4
n
a. Study the monotony of ðwn Þ:
b. Show that ðwn Þ is bounded.
7. Consider the sequence ðun Þ of general term
2n 7
un ¼ :
nþ1
Determine the number of positive terms of ðun Þ that are smaller than 75.
8. Let ðvn Þ be the sequence whose general term is

45
1 2n
vn ¼ :
2n þ 3
a. Obtain the term of order 5 of ðvn Þ:
b. Verify if 29
25
is a term of ðvn Þ
c. Study the monotony of ðvn Þ:
d. Show that ðvn Þ is bounded.
9. Study the monotony of the sequence ðwn Þ defined by wn ¼ ð1Þn þ 3n.
Arithmetic an geometric progressions
10.Let ðun Þ be the arithmetic sequence defined by:

u1 ¼ 2
:
unþ1 ¼ un þ 3, ∀n ∈ 
a. Obtain the general term of ðun Þ:
b. Determine the value of u31 u10 .
c. Determine the sum of the first 20 terms of the ðun Þ:
11. Let ðvn Þ be the sequence whose general term is
1 5n
vn ¼ :
3
a. Show that ðvn Þ is an arithmetic progression.
b. Justify that ðvn Þ is a decreasing sequence.
c. Determine the sum of the 15 consecutive terms of ðvn Þ after v22 .
12. Let ðwn Þ be an arithmetic progression, such that w5 ¼ 12 and w13 ¼ 44.
a. Determine the common difference of ðwn Þ and classify its monotony.
b. Obtain the general term of ðwn Þ:
c. The sum of the first m terms of ðwn Þ is 360. Determine the value of m.
13. Consider the geometric progression ðun Þ defined by

(
u1 ¼ 4
un :
unþ1 ¼ , ∀n ∈ 
5
46
ITexLi.113863
a. Obtain the general term of ðun Þ:
b. Determine the monotony of ðun Þ:
c. Prove that ðun Þ is bounded.
d. Compute the exact value of the sum of the first 10 terms of ðun Þ:
14.Consider the sequence ðvn Þ whose general term is
5
vn ¼ n1
:
2
a. Show that ðvn Þ is a geometric progression.
b. Determine the monotony of ðvn Þ:
c. Show that ðvn Þ is bounded.
d. Obtain the exact value of v3 þ v4 þ ⋯ þ v9 .
15. Consider the arithmetic progression ðwn Þ defined by

w1 ¼ 50
:
wnþ1 ¼ wn þ 2, ∀n ∈ 
For a certain p ∈ , the sum of the first p terms of ðwn Þ is equal to p. Determine
the value of p.
Limits and convergence
16. Let ðun Þ be the sequence defined by
2n
un ¼ :
nþ3
Using the definition of convergent sequence, show that lim un ¼ 2.
17. Consider the sequence ðvn Þ such that
2 cosðnπ Þ
vn ¼ pffiffiffiffiffiffiffiffiffiffiffi :
nþ1
Using the Squeeze Theorem for Convergent Sequences, show that
lim vn ¼ 0:
47
18. Evaluate lim wn where
a. wn ¼ 4n2 þ n 5
b. wn ¼ 7n5 100n2 þ 500
2n þ3n7
5 2
c. wn ¼ 3n
2n3 n2 þ2n3
þn 5
3 2
d. wn ¼ 2n3n
3 3n2 nþ6
þ3n5
2
e. wn ¼ 2nn3 5n 2 þnþ2
pffiffiffiffiffiffiffiffi pffiffi
2nþ5 n
f. wn ¼ nþ5
g. wn ¼ pffiffiffiffiffiffiffi
2nffi
n2 þ1þ3n
pffiffi 3 pffiffi
3n
h. wn ¼ n2n4 þ4n3
pffiffiffiffiffiffiffiffiffiffiffiffiffi
i. wn ¼ n2 1 n þ 3
lnð2nþ5Þ
j. wn ¼ n2 1
sinðn2Þ
k. wn ¼ nþ1
l. wn ¼ 7 n 2n
19. Consider the sequences ðun Þ and ðvn Þ whose general terms are given by
3
un ¼ ð1Þn and vn ¼ :
2n
Calculate limit of the sequence ðun vn Þ:
20.Let ðwn Þ be a monotonic geometric progression such that w5 ¼ 125 and w11 ¼ 125
1
.
a. Show that w1 ¼ 78125 and r ¼ 15, where r is the common ratio of ðwn Þ:
b. Let Sn be the sum of the first n terms of ðwn Þ: Calculate lim Sn .
21. Consider the sequence ðun Þ whose general term is given by
3
un ¼ :
ðn þ 1Þ!5n
Show that ðun Þ is a Cauchy sequence.
48
ITexLi.113863
Miscellaneous problems
22. Determine the monotony of the sequence ðvn Þ as defined below.
a. vn ¼ 2nþ1
1
;
b. vn ¼ nnþ 2
þ1 ;
c. vn ¼ enn ;
d. vn ¼ n2nþ1 :
23. Consider the sequence ðwn Þ such that
n!
wn ¼
nn
for all n ∈ .
a. Show that ðwn Þ is decreasing.
b. Justify that ðwn Þ is convergent.
24.Using the definition of limit, prove the following equalities.
a. lim nþ1
2nþ1 ¼ 21 ;
n4 1
b. lim n4 þ1 ¼ 1;
sinðnÞ
c. lim n2 ¼ 0;
lnðnÞ
d. lim n ¼ 0;
1þnn
e. lim n ¼ e:
25. Let ðun Þ be a sequence of real numbers converging to 1. Consider the new
sequence ðvn Þ such that
ðun Þ3 1
vn ¼ :
un 1
Analyze the convergence of ðvn Þ:
26. Show that
12 þ 22 þ ⋯ þ n2
lim ¼ 0:
n4
49
27. Calculate the following limits:

n
nþ1
a. lim nþ2 ;
b. lim nþ1 n ;
pffiffiffi
c. lim n n;
2n
d. lim n ;

e. lim p31ffiffi þ p3 ffiffiffiffiffiffi
1
þ ⋯ þ p31ffiffiffi
ffi ;
n nþ1 2n
pffiffiffi
f. lim n 3;
2n þ3n
g. lim 4n þ7n ;
7n þ8n
h. lim 3n þ4n ;
pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
i. lim n 3n þ 9n :
28. Show that, if ðwn Þ is a sequence of positive real numbers such that
pffiffiffiffiffiffi
lim wwnþ1
n
¼ L ∈ , then lim n wn ¼ L:
29. Prove that, given ðun Þ a convergent sequence of real numbers, if lim un ¼ L,
then
u1 þ u2 þ ⋯ þ un
lim ¼ L:
n
pffiffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffi
30.Consider the sequence ðvn Þ such that v1 ¼ 2 and vnþ1 ¼ 2 þ vn for n ≥ 2.
a. Show that ðvn Þ is an increasing sequence.
b. Show that ðvn Þ is bounded from above.
c. Show that lim vn ¼ 2.

pffiffiffi
31. Show that the sequence ðwn Þ of positive terms, where w1 ¼ 2 and wnþ1 ¼
2wn for n ∈ nf1g, is convergent. Find its limit.
32. Show that

1 1 1
lim pffiffiffiffiffiffiffiffiffiffiffi þ pffiffiffiffiffiffiffiffiffiffiffi þ ⋯ þ pffiffiffiffiffiffiffiffiffiffiffiffiffi ¼ þ∞:
nþ1 nþ2 2n þ 1
33. Show that the sequence ðun Þ, where
1 1 1
un ¼ pffiffiffiffiffiffiffiffiffiffiffiffiffi þ pffiffiffiffiffiffiffiffiffiffiffiffiffi þ ⋯ þ pffiffiffiffiffiffiffiffiffiffiffiffiffi ,
n2 þ 1 n2 þ 2 n2 þ n
50
ITexLi.113863
for all n ∈ , is convergent.
34.Show that
p ffiffiffiffiffiffiffiffiffiffiffi p ffiffiffi
lim 3
n þ 1 3 n ¼ 0:
35. Use mathematical induction to show that, ∀n ∈ , we have:
a. 1 þ 3 þ 5 þ ⋯ þ 2n 1 ¼ n2 ;
2
P
n
3 P
n
b. k ¼ k ;
k¼1 k¼1
P
n
nðnþ3Þ
c. 1
kðkþ1Þðkþ2Þ ¼ 4ðnþ1 Þðnþ2Þ :
k¼1
36.Consider a convergent sequence ðun Þ such that un 6¼ 2, for all n ∈  and

lim un ¼ 2: Calculate
un þ u2n þ u3n 14
lim :
un 2
37. Show that the sequences ðvn Þ are decreasing, where:
lnðnþ2Þ
a. vn ¼ nþ2 ;
b. vn ¼ 9nn :
51
3. Series
This section is devoted to the subject of series, which are essentially generalized
infinite sums of numbers. In view of this, it is easy to understand the relevance of
having addressed the subject of sequences in the previous section. Indeed, we can say
that the following survey on series will be made on the shoulders of all that was
presented about sequences of real numbers.
The subject of series plays a major role in mathematical analysis, and it is widely
applied to many different areas like economics and finance, statistics, physics, or
computer science.
This section is organized as follows. We start by presenting, on Section 3.1, the
basic definitions and results involving series, including convergence. Then, from
Sections 3.2 to 3.5 we present some important types of series, including series of non-
negative terms, alternating series, function and power series, along with the applica-
tion to the representation of functions. As in the previous section, we propose, in
Section 3.6, a set of exercises for the reader to grasp their acquired knowledge on this
subject, with which we finish this section.
3.1 Basic definitions and convergence
We start this section by defining the concept of series by means of the concept of
sequence. Before, however, note that along this text, we will only consider series of
real numbers.
Definition 3.1 (Series). Given a sequence of real numbers, ðun Þ, a series is an expres-
sion of the form
X
þ∞
uk ¼ u1 þ u2 þ ⋯,
k¼1
where each uk , k ∈ , is simultaneous a term of the series and a term of the sequence ðun Þ.
The sequence ðun Þ, of Definition 3.1, is sometimes called the generating sequence of
the series, due to its nature.
One can also come to an equivalent definition of series if, from ðun Þ, we build a
new sequence ðsn Þ, with general term
sn ¼ u1 þ u2 þ ⋯ þ un ,
which is called the sequence of partial sums of ðun Þ: Note that the value of the series
Pþ∞
k¼1 uk of Definition 3.1, also referred to as the sum of the series, is the same as
calculating lim n!þ∞ sn . The next definition is, therefore, natural.
Definition 3.2 (Convergent series). Let ðun Þ be a sequence and ðsn Þ be its
P
corresponding partial sums sequence. Then, we say that the series þ∞ k¼1 uk is convergent
(divergent) if and only if the sequence ðsn Þ is convergent (divergent).
Next, we present some examples where Definition 3.2 is put to effective use.
P
Example 3.1. The series þ∞ k¼1 1 is divergent. Indeed, it is only necessary to observe that
the sequence of partial sums of the constant sequence with all terms equal to 1 is given by the
general term sn ¼ n and, for this sequence, we have that lim sn ¼ lim n ¼ þ∞. Since ðsn Þ is
P
divergent, then the corresponding series þ∞ k¼1 1 is divergent.
Example 3.2. Let’s study the convergence of the series
52
ITexLi.113863
þ∞ n
X 1
:
n¼1
2
The generating sequence of this series is a geometric sequence with the first term and
common ratio both equal to 21. Then, by applying the sum formula for a geometric sequence,
the partial sums sequence is given by
n !
1 1 21
sn ¼ :
2 1 21
Then, we have
0 1
1n
1 B1 2 C
lim sn ¼ lim@
2 1 A
1
2
n
1
¼ lim 1
2
¼ 1:
Pþ∞ 1n
Therefore, ðsn Þ is convergent and, consequently, the series n¼1 2 is also convergent
and we write
þ∞ n
X 1
¼ 1:
n¼1
2
P
Example 3.3. Consider the series þ∞ n
n¼1 5 .
Like in Example 3.2, this series also has a generating sequence that is a geometric
progression. In this case, the first term and the common ratio are both equal to 5. Thus, we
have

1 5n
lim sn ¼ lim 5
15
5
¼ limð1 5n Þ
4
5
¼ ½1 ðþ∞Þ
4
5
¼ ð∞Þ:
4
¼ þ∞:
P
Since ðsn Þ is divergent, the series þ∞ n
n¼1 5 is also divergent. Some authors write
Pþ∞ n
n¼1 5 ¼ þ∞, though.
Next, we present an example of a series whose generating sequence is an alternat-
ing sequence.
P
Example 3.4. Consider the series þ∞ n¼1 ð1Þ
nþ1
.
For the corresponding sequence, the partial sums are:
53
s1 ¼ 1
s2 ¼ 1 1 ¼ 0
s3 ¼ 1 1 þ 1 ¼ 1
s4 ¼ 1 1 þ 1 1 ¼ 0
s5 ¼ 1 1 þ 1 1 þ 1 ¼ 1
⋮
s2n1 ¼ 1
s2n ¼ 1:
Since lim s2n1 ¼ 1 and lim s2n ¼ 0, lim sn does not exist and, therefore, the series
Pþ∞ nþ1
n¼1 ð1Þ is divergent.
Example 3.5. Let’s analyze the convergence of the series of the form
X
þ∞
rn1 ,
n¼1
with r ∈ , also known as a geometric series.

The corresponding sequence of partial sums, ðsn Þ has its general term given by
X
n
sn ¼ rk1 :
k¼1
We will break down this analysis into different situations.
1. Let r 6¼ 1. Then, we have the following equalities:
sn ¼ 1 þ r þ r2 þ ⋯ þ rn1 (18)
and
rsn ¼ r þ r2 þ ⋯ þ rn , (19)
Subtracting (19) from (18) yields

sn ð1 rÞ ¼ 1 rn
1 rn
⇔ sn ¼ :
1r
Observe that the convergence of ðsn Þ, and, therefore, of the series, depends now on
the fact that ∣r∣ is smaller or bigger than 1.
i. If ∣r∣ < 1, we have that lim rn ¼ 0. Then,

1 rn
lim sn ¼ lim
1r
1
¼ :
1r
1
Then the series is convergent and its sum is equal to 1r .
54
ITexLi.113863
ii. If ∣r∣ > 1, there are two cases two consider, r > 1 and r < 1.
a. If r > 1, then lim rn ¼ þ∞ and we have
1 rn
lim sn ¼ lim
1r
1 ðþ∞Þ
¼
1r
1∞
¼
1r
¼ þ∞,
and we conclude that the series is divergent.
b. If r < 1, then lim rn does not exist since lim r2n1 ¼ ∞ and
lim r2n ¼ þ∞: Therefore,
1 rn
lim sn ¼ lim
1r
1 lim rn
¼
1r
Does not exist and the series is divergent.
iii. We are missing to consider one single sub-case from the general case r 6¼ 1,
that is, the case when r ¼ 1. If r ¼ 1, then
1 ð1Þn
lim sn ¼ lim
1 ð1Þ
1 ð1Þn
¼ lim :
2
Now, since
1 ð1Þ
lim s2n1 ¼ lim ¼1
1 ð1Þ
and
11
lim s2n ¼ lim ¼ 0,
1þ1
we conclude that lim sn does not exist and the series is divergent.
2. The final case to consider is, of course, the case when r ¼ 1. If r ¼ 1, we have
X
n
lim sn ¼ lim 1k
k¼1
¼ lim n
¼ þ∞:
55
Thus, the series diverges.

P
Hence, we conclude that the series þ∞
n¼1 r
n1
is convergent if and only if ∣r∣ < 1 and, in
this case, we have
X
þ∞
1
rn1 ¼ :
n¼1
1r
The general case of the geometric series is presented next.

Proposition 3.1. Let a, r ∈ . If ∣r∣ < 1, then the geometric series of first term a and
common ratio r,
X
þ∞
arn1 ,
n¼1
a
is convergent and its sum is 1r .
In some cases, the properties of the generating sequence allow us to deduce easily
an expression for the partial sums sequence.
P pffiffiffi pffiffiffiffiffiffiffiffiffiffiffi
Example 3.6. Consider the series þ∞ n¼1 n nþ1 .
Observe that
pffiffiffi pffiffiffi
s1 ¼ 1 2
pffiffiffi pffiffiffi pffiffiffi pffiffiffi
s2 ¼ 1 2 þ 2 3
pffiffiffi pffiffiffi pffiffiffi pffiffiffi pffiffiffi pffiffiffi
s3 ¼ 1 2 þ 2 3 þ 3 4
⋮
pffiffiffi pffiffiffi pffiffiffi pffiffiffi pffiffiffi pffiffiffi
sn ¼ 1 2 þ 2 3 þ 3 4
pffiffiffi pffiffiffiffiffiffiffiffiffiffiffi
þ⋯ þ n n þ 1
pffiffiffiffiffiffiffiffiffiffiffi
¼ 1 nþ1
Pþ∞ pffiffiffi pffiffiffiffiffiffiffiffiffiffiffi
Hence, lim sn ¼ ∞ and, therefore, the series n¼1 n n þ 1 is divergent.
Example 3.7. Consider the series
X
þ∞
1
:
n¼1
nðn þ 1Þ
Firstly, note that
1 1 1
¼ :
nðn þ 1Þ n n þ 1
Then, we have
56
ITexLi.113863

1
s1 ¼ 1
2

1 1 1
s2 ¼ 1 þ
2 2 3

1 1 1 1 1
s3 ¼ 1 þ þ
2 2 3 3 3
⋮

1 1 1 1 1
sn ¼ 1 þ þ þ
2 2 3 3 4

1 1 1 1
þ⋯ þ þ
n1 n n nþ1
1
¼ 1 :
nþ1
P
Hence, lim sn ¼ lim 1 nþ1
1
¼ 1 and the series þ∞ 1
n¼1 nðnþ1Þ is convergent with sum
equal to 1.
We will now introduce a special kind of series.
P
Definition 3.3 (Mengoli’s series). A series þ∞ n¼1 un such that un ¼ vn vnþm , with
m ∈ , is called Mengoli’s series.
Mengoli’s series is also sometimes referred to as Telescopian series. The convergence
of Mengoli’s series is easy to analyze.
P
Proposition 3.2. A Mengoli’s series þ∞ n¼1 ðvn vnþm Þ, m ∈ , is convergent if and only
if the sequence ðvn Þ is convergent. Additionally,
X
þ∞ X
m
ðvn vnþm Þ ¼ vk m lim vn :
n¼1 k¼1
Pþ∞
Proof. For Mengoli’s series n¼1 ðvn vnþm Þ, we have that the partial sums
sequence, ðsn Þ, is given by
X
n
sn ¼ ðvk vkþm Þ
k¼1
Xn X
n
¼ vk vkþm
k¼1 k¼1
Xn X
nþm
¼ vk vk
k¼1 k¼mþ1
X
m X
n X
n X
n þm
¼ vk þ vk vk vk
k¼1 k¼mþ1 k¼mþ1 k¼nþ1
X
m X
nþm
¼ vk vk :
k¼1 k¼nþ1
Now, let ðwn Þ be the sequence whose general term is
wn ¼ vnþ1 þ vnþ2 þ ⋯ þ vnþm :
57
Pþ∞
Observe that the series n¼1 ðvn vnþm Þ converges if and only if ðwn Þ converges
and, in this case, we have
lim wn ¼ limðvnþ1 þ vnþ2 þ ⋯ þ vnþm Þ

¼ lim vnþ1 þ lim vnþ2 þ ⋯ þ lim vnþm
¼ m lim vn :
Hence,
X
þ∞
ðvn vnþm Þ ¼ lim sn
n¼1
!
X
m X
n þm
¼ lim vk vk
k¼1 k¼nþ1
X
m X
n þm
¼ vk lim vk
k¼1 k¼nþ1
X
m
¼ vk m lim vn :
k¼1
□
The particular case of a Mengoli’s series with m ¼ 1 is worth considering.
P
Corollary 3.1. A Mengoli’s series þ∞ n¼1 ðvn vnþ1 Þ is convergent if and only if the
sequence ðvn Þ is convergent. Additionally,
X
þ∞
ðvn vnþ1 Þ ¼ v1 lim vn :
n¼1
We proceed with some relevant properties involving series convergence.

P
Theorem 3.1. Let α be a real non null number. Then þ∞n¼1 un converges if and only if
Pþ∞
ð αu Þ converges.
n¼1 n
P
Proof. Let s1n and s2n denote the sequences of partial sums of the series þ∞ n¼1 un
Pþ∞
and n¼1 ðαun Þ, respectively.
Since
X
n
s2n ¼ ðαuk Þ
k¼1
X
n
¼ α uk
k¼1
¼ αs1n ,
then we have that
lim s2n ¼ α lim s1n
and, therefore,
58
ITexLi.113863
X
þ∞ X
þ∞
ðαun Þ ¼ α un :
n¼1 n¼1
□
P Pþ∞
Theorem 3.2. Let α and β be non null real numbers. If þ∞ un and v n are
P n¼1 n¼1
convergent series, then the series þ∞ ðαun þ βvn Þ is also a convergent series.
n¼1 P
Proof. Let s1n and s2n denote the partial sums sequences of the series þ∞ n¼1 un and
Pþ∞
n¼1 nv , respectively. Also, since both series are convergent, let lim s 1
n ¼ S 1 and
lim sn ¼ S2 , for some S1 , S2 ∈ .
2
We have:
X
n X
n X
n
ðαuk þ βvk Þ ¼ αuk þ βvk
k¼1 k¼1 k¼1
X n X
n
¼α uk þ β vk
k¼1 k¼1
¼ αs1n þ βs2n :
Hence,

lim αs1n þ βs2n ¼ α lim s1n þ β lim s2n
¼ αS1 þ βS2 :
P
We conclude that the series þ∞ n¼1 ðαun þ βvn Þ is convergent and that the sum of this
Pþ∞ P
series is αS1 þ βS2 where S1 and S2 are the sum of the series n¼1 un and þ∞ n¼1 vn ,
respectively. □
P Pþ∞
Note that if we consider two divergent series, þ∞ u n and vn , we cannot
P n¼1 n¼1
conclude that the series þ∞ ðαun þ βvn Þ is also divergent, for any real numbers α and
n¼1 P Pþ∞ 1
β. Indeed, we have the series þ∞ 1
n¼1 n and n¼1 n are both divergent and yet the

Pþ∞ 1 1
series n¼1 n n is clearly convergent to 0.
P Pþ∞
Theorem 3.3. Let l be a natural number. Then, the series þ∞ n¼1 un and n¼lþ1 un are
either both convergent
or both
divergent.
P
Proof. Let s1n and s2n denote the partial sums sequences: s1n ¼ nk¼1 uk and s2n ¼
Pn
k¼lþ1 uk and let S1 ¼ lim sn and S2 ¼ lim sn .
1 2
If n is a natural number greater than l, we have
X
n
s1n ¼ uk
k¼1
X
l X
n
¼ uk þ uk
k¼1 k¼lþ1
X
l
¼ uk þ s2n :
k¼1
Applying limits, we obtain
59
X
l
lim s1n ¼ uk þ lim s2n (20)
k¼1
P
From equality (20), we straightforwardly conclude that the series þ∞ n¼1 un and
Pþ∞
n¼lþ1 un are both convergent or both divergent, and, in the first case, we have that
X
l
S1 ¼ ak þ S2 :
k¼1
□
It is a rather common situation to find ourselves with a series at hand for which there
is not easy way to obtain an expression for the respective partial sums sequence. In these
situations it is important to establish a criterium that allows us to identify convergent
series without having to resort to the partial sums sequence. Such a criterium, that is, a
necessary condition for the convergence of series, is presented next.
P
Theorem 3.4. If the series þ∞ lim un ¼ 0.
n¼1 un is convergent, then P
Pþ∞
Proof. Let n¼1 un be a convergent series, such that þ∞ n¼1 un ¼ l, l ∈ , and let ðsn Þ
be its sequence of partial sums.
Since un ¼ sn sn1 and lim sn ¼ lim sn1 ¼ l, we have
lim un ¼ lim sn lim sn1
¼ ll
¼ 0:
□
Example 3.8. Let’s prove that the series
X
þ∞
2n
n¼1
2n þ 1
is divergent.
Let ðun Þ be the sequence whose general term is given by
2n
un ¼ :
2n þ 1
We have that
2n
lim un ¼ lim
2 þ1
n
2n
¼ lim
1
2 1þ n
n
2
1
¼ lim
1
1þ n
2
1
¼
1þ0
¼ 1 6¼ 0:
Pþ∞ 2n
Then, by Theorem 3.4, we conclude that the series n¼1 2n þ1 is divergent.
60
ITexLi.113863
3.2 Series of non-negative terms

P
In this section we focus on series þ∞ n¼1 un , such that un ≥ 0, ∀n ∈ , which are called
series of non-negative terms, and we survey the most relevant techniques for studying
their convergence.
P Pþ∞
Theorem 3.5. Let þ∞ n¼1 un be a sequence of non-negative terms. Then n¼1 un is
convergent if and only if its sequence of partial sums is bounded from above.
P
Proof. Let þ∞ n¼1 un be a series of non-negative terms with ðsn Þ being the respective
partial sums sequence.
P
Let þ∞ n¼1 un be convergent with lim sn ¼ L: Note that snþ1 ¼ sn þ unþ1 and unþ1 ≥ 0
by hypothesis. Then, clearly, ðsn Þ is an increasing sequence and sn ≤ L, that is, ðsn Þ is
bounded from above.
Conversely, let ðsn Þ be bounded from above. Since ðsn Þ is an increasing sequence
bounded from above, then it is convergent to the supremum of its range, L ∈ . Then,
P
lim sn ¼ L and þ∞ n¼1 un is a convergent series. □
Pþ∞ 1
Example 3.9. Let’s show that the Harmonic series, n¼1 n, is divergent by proving that
the corresponding partial sums sequence ðsn Þ is divergent.
We have that
X
2n
1
s2n ¼
k¼1
k
X
n
1 X2n
1
¼ þ
k¼1
k k¼nþ1
k
1 1 1
¼ sn þ þ þ⋯þ
nþ1 nþ2 2n
1 1 1
≥ sn þ þ þ ⋯ þ
2n 2n 2n
n
¼ sn þ
2n
1
¼ sn þ
2
and, therefore,
1
s2n sn ≥
2
Now, suppose that lim sn ¼ lim s2n ¼ L, with L ∈ . Then, we would have
1
lim s2n lim sn ¼ L L ¼ 0 ≥ ,
2
which is absurd. Therefore, we conclude that such an L does not exist and, accordingly,
the sequence ðsn Þ is divergent. Hence, the harmonic series is also divergent.
P Pþ∞
Theorem 3.6 (Comparison test). Let þ∞ n¼1 un and n¼1 vn be two series of non-
negative terms such that un ≤ vn , ∀n ∈ . Then we have:
Pþ∞ Pþ∞
a. If n¼1 vn converges then n¼1 un converges.
Pþ∞ Pþ∞
b. If n¼1 un diverges then n¼1 vn diverges.
61
P Pþ∞
Proof. Let þ∞ n¼1 un and n¼1 v be two series of non-negative terms such that
n
un ≤ vn , ∀n ∈ : Denote by s1n and s2n denote the partial sums sequences of the series
Pþ∞ Pþ∞
n¼1 un and n¼1 vn , respectively.
P
a. If þ∞
n¼1 vn is convergent, let lim sn ¼ S, with S ∈ . Then, we have the following
2
inequalities:
s1n ≤ s2n ≤ S:

Since s2n is an increasing sequence that is bounded from above, we have that s1n is
Pþ∞
convergent and, consequently, the series n¼1 un is convergent.
P
b. If þ∞
n¼1 un is divergent, since it is a series of non-negative terms, then lim s1n ¼ þ∞:
P
Therefore, lim s2n ¼ þ∞ and the series þ∞ n¼1 vn is divergent.
□
Theorem 3.6 is usually referred to as the comparison test. Note that it is enough
that there exists a m ∈  such that un ≤ vn , ∀n ≥ m, for Theorem 3.6’s assertions to
be valid.
We will now present a couple of examples where we apply Theorem 3.6.
Example 3.10. In this example, we will study the convergence of the series:
X
þ∞
lnðn þ 2Þ
:
n¼1
nþ2
Firstly, note that the following inequality holds for every natural n:
lnðn þ 2Þ lnð3Þ
≥ :
nþ2 nþ2
Secondly, we have that
X
þ∞
lnð3Þ X
þ∞
1 X
þ∞
1
¼ lnð3Þ ¼ lnð3Þ :
n¼1
n þ 2 n¼1
n þ 2 n¼3
n
Pþ∞ 1
Since is just the Harmonic series starting at the third term, which we already saw
n¼3 n
P lnð3Þ
on Example 3.9, that it is not convergent, we can conclude that the series þ∞ n¼1 nþ2 is also
divergent.
Finally, using the comparison test, Theorem 3.6-(b), we can conclude that the series
X
þ∞
lnðn þ 2Þ
n¼1
nþ2
is divergent.
Example 3.11. Let’s analyze the convergence of the series:
þ∞
X
3n
:
n¼1 32n þ 1
62
ITexLi.113863
Observe that
3n 3n
≤
32n þ 1 32n
3n
≤ 2n
3
3n
¼ 2 n
3
n
1
¼ :
3
P 1n
Now, since the series þ∞n¼1 3 converges because it is a geometric series with ratio 31 and
absolute value smaller than 1, then, using the comparison test, Theorem 3.6-(a), we con-
clude that the series
þ∞
X
3n
n¼1 32n þ 1
is also convergent.
Theorem 3.7 (Cauchy condensation test). Let ðun Þ be a decreasing sequence of non-
P Pþ∞ n
negative terms. Then the series þ∞
n¼1 un converges if and only if the series n¼0 2 u2
n
converges.
Proof. Let ðun Þ be a decreasing sequence of non-negative terms.
Pþ∞ n n
1. Suppose that the series n¼0 2 u2 converges.
Since the sequence ðun Þ is decreasing, then the following inequalities hold:
u2 þ u3 ≤ u2 þ u2 ¼ 21 u21
u4 þ u5 þ u6 þ u7 ≤ u4 þ u4 þ u4 þ u4 ¼ 22 u22
u8 þ u9 þ ⋯ þ u15 ≤ u8 þ u8 þ ⋯ þ u8 ¼ 23 u23
u16 þ u17 þ ⋯ þ u31 ≤ 216 u216
⋮
u2n þ ⋯ þ u2nþ1 1 ≤ 2n u2n :
Pþ∞
Then, denoting by ðsn Þ the sequence of partial sums of n¼1 un , we can write the
following:
X
n
sn ¼ uk
k¼1
¼ u1 þ u2 þ u3 þ ⋯ þ un
≤ u1 þ ðu2 þ u3 Þ þ ðu4 þ u5 þ u6 þ u7 þ u8 Þ þ ⋯
þðu2n þ u2n þ1 þ ⋯ þ u2nþ1 1 Þ
≤ u1 þ 2u2 þ 4u4 þ 8u8 þ ⋯ þ 2n u2n
X
þ∞
¼ 2n u2n :
n¼0
63
We conclude that ðsn Þ is bounded from above and, since it is an increasing

P
sequence, we can conclude that it converges. Hence, the series þ∞ n¼1 un is also
convergent.
Pþ∞
2. Suppose that the series n¼1 un is convergent.
Pþ∞
Let ðsn Þ denote the sequence of partial sums of the series n¼0 2
n
u2n . We have
that
1 1X n
sn ¼ 2k u2k
2 2 k¼0
1
¼ ðu1 þ 2u2 þ 4u4 þ 8u8 þ ⋯ þ 2n u2n Þ
2
1
¼ u1 þ u2 þ 2u4 þ 4u8 þ ⋯ þ 2n1 u2n
2
≤ u1 þ u2 þ ðu3 þ u4 Þ þ ðu5 þ u6 þ u7 þ u8 Þ þ ⋯

þ u2n1 þ1 þ u2n1 þ2 þ ⋯ þ u2n 1 þ u2n
X
þ∞
≤ un :
n¼1
P
Therefore, we have that sn ≤ 2 þ∞ n¼1 un , which means that ðsn Þ is bounded from
P
above and, since it is an increasing sequence, it is convergent. Hence, þ∞ n n
n¼0 2 u2
is also a convergent series.
□
We will now present a special kind of series.
Definition 3.4 (p-series). For any real number p, a p-series is a series defined as
X
þ∞
1
p
:
n¼1
n
The convergence of a p-series is determined by the value of p as it is shown next.

P
Theorem 3.8. Let p be a real number. The series þ∞n¼1 np is convergent if p > 1 and
1
divergent if p ≤ 1.
P
Proof. Let p be a real number and consider the p-series þ∞ n¼1 np : The proof will be
1
separated into different cases.
1. If p ¼ 1, then the p-series corresponds to the Harmonic series which we already

know, from Example 3.9, that is divergent.
2. If 0 < p < 1, then we consider the following trivial inequalities:
1 1
np ≤ n⇔ p
≥ :
n n
P 1
Therefore, since þ∞ n is divergent, see (1.), by the comparison test, we have
Pþ∞n¼11
that the series n¼1 np is also divergent.
64
ITexLi.113863
3. If p is a natural number greater or equal to 2, consider the trivial inequalities
1 1
np > nðn 1Þ⇔ < :
np nðn 1Þ
Now, observe that
X
þ∞ þ∞
X
1 1 1
¼ þ
n¼2
nðn 1Þ n¼2
n n1
þ∞
X
1 1
¼
n¼2
n1 n
þ∞
X
1 1
¼ ,
n¼1
n nþ1
P
which is a Mengoli’s series with sum 1. Then, since þ∞ 1
n¼2 nðn1Þ is convergent, by
P
the comparison test, we have that þ∞ 1
n¼1 np is a convergent series.
4. If p is a real number greater or equal than 2 and not natural, choose k to be the
greatest natural number smaller or equal than p. We have that nk < np and,
therefore,
1 1
> :
nk np
P
Since the series þ∞ 1
n¼1 nk is convergent, see 4., applying the comparison test, we
P
conclude that the series þ∞ 1
n¼1 np is also convergent.
5. If p is a real number such that p ∈ 1, 2½, consider the sequence un whose general
term is un ¼ n1p . Then, we have
1
2n a2n ¼ 2n
ð2n Þp
1
¼ npn
2
1
¼
2nðp1Þ

1 n
¼ :
2p1
Since p ∈ 1, 2½, we have that
65
1
0< <1
2p1
n
1
and, therefore, the series 2 p1 is convergent, since it is a geometric series with
1
ratio p1 < 1. Finally, by the Cauchy condensation test, we have that the series
Pþ∞ 2 1
n¼1 np is convergent
P Pþ∞
6. If p ¼ 0, then þ∞
n¼1 np ¼
1
n¼1 1 which is a divergent series because lim 1 ¼ 1 6¼
0 (Theorem 3.4).
P Pþ∞ ∣ p ∣
7. If p < 0, then þ∞
n¼1 np ¼
1
n¼1 n , which is a divergent series because lim n
∣p∣
¼
þ∞ 6¼ 0 (Theorem 3.4).
□
Example 3.12. In this example, we will show that the series
X
þ∞
n
n¼1
7n
converges using the comparison test.

As a preliminary observation, since 7 n is much bigger than n, as n grows, it is natural to
expect that the series will converge. Let us see that, indeed, this happens.
Observe that:
7 n ¼ ð1 þ 6Þn
!
Xn n
¼ 1nj 6j
j¼0 j
! ! ! !
n n 0 n n1 1 n n
¼ 1 6 þ 1 6 þ 1 6 þ⋯þ
n2 2
10 6n
0 1 2 n
!
n n3 3
> 1 6
3
n!
¼ 216
3!ðn 3Þ!
216nðn 1Þðn 2Þ
¼
3!
¼ 36nðn 1Þðn 2Þ
> 36ðn 2Þ2 :
Then, we also have that
1 1
< ,
7 n 36ðn 2Þ2
and
66
ITexLi.113863
n n
< ,
7 n 36nðn 2Þ2
for all n ≥ 3.
P Pþ∞ 1
Now, since the p-series þ∞ 1
n¼1 n2 converges, by Theorem 3.8, then the series n¼3 ðn2Þ2
1
Pþ∞ 1
and 36 n¼3 ðn2Þ2 also converge.
P
Finally, by the comparison test, we conclude that the series þ∞ n
n¼1 7n is convergent.
In view of the fact that we have a good understanding of their natures, the
geometric series and the p-series are of great importance in the context of the usage of
the comparison test to study the convergence of many series.
We will now present another test that is called the limit comparison test.
P
Theorem 3.9 (Limit comparison test). Let þ∞ n¼1 un be a series of non-negative terms
Pþ∞
and n¼1 vn be a series of positive terms. If lim uvnn exists, then the following statements hold.
Pþ∞ Pþ∞
a. If lim uvnn ¼ r, with r ∈ þ , then n¼1 un converges (diverges) if and only if n¼1 vn
converges (diverges).
b. If lim un
vn ¼ 0, then:
Pþ∞ Pþ∞
i. if n¼1 vn converges, then n¼1 un converges.
Pþ∞ Pþ∞
ii. if n¼1 un diverges, then n¼1 vn diverges.
c. If lim un
vn ¼ þ∞, then:
Pþ∞ Pþ∞
i. if n¼1 vn diverges, then n¼1 un diverges.
Pþ∞ Pþ∞
ii. if n¼1 un converges, then n¼1 vn converges.
P Pþ∞
Proof. Let þ∞n¼1 un be a series of non-negative terms and n¼1 vn be a series of
positive terms.
a. Suppose that lim un

vn ¼ r, with r ∈ þ . Then, there is a m ∈  such that:
r un r
n>m ) r < <r þ :
2 vn 2
Therefore, for n > m we have the following inequalities

un r r
> ⇔un > vn (21)
vn 2 2
and
un 3r 3r
< ⇔un < vn : (22)
vn 2 2
67
P Pþ∞
If þ∞n¼1 un converges, so does the series n¼mþ1 un . Then, considering inequality
P
(21) and applying the comparison test, we conclude that the series þ∞ r
n¼mþ1 2 vn
Pþ∞
is convergent and, therefore, n¼1 vn is also convergent.
P Pþ∞ 3r
If þ∞n¼1 vn is convergent, n¼mþ1 2 vn is also convergent. Then, taking into
consideration the inequality (22) and applying the comparison test, we conclude
P Pþ∞
that þ∞n¼mþ1 un converges and so does the series n¼1 un .
b. Suppose that lim un

vn ¼ 0. Then, there is a m ∈  such that:
un
n > m ) 1 < < 1:
vn
This means that, for n > m we have that un < vn .

P Pþ∞
If þ∞n¼1 vn converges, so does the series vn . Then, applying the
Pþ∞ n¼mþ1
comparison test, we conclude that n¼mþ1 un is convergent and, therefore,
Pþ∞
n¼1 un is also convergent. This proves assertion (i.). Assertion (ii.) is just the
contrapositive implication of (i.), which is logically equivalent.
c. Suppose that lim un

vn ¼ þ∞. Then, there is a m ∈  such that:
n > m ) un > vn :
P Pþ∞
If þ∞n¼1 vn is divergent, n¼mþ1 vn is also divergent. Then, by the comparison test,
Pþ∞ P
we conclude that n¼mþ1 un is divergent. Therefore, the series þ∞ n¼1 un is also diver-
gent. This proves assertion (i.). Assertion (ii.) is just the contrapositive implication of
(i.), which is logically equivalent.
□
Much of the analysis of a series convergence is done through the medium of the
comparison test or the limit comparison test together with the support of series whose
nature we are familiar with, such as the geometric series or the p-series. Next, we
present some examples.
Example 3.13. In this example, we will show that the series
X
þ∞
nþ1
n¼1
n3 þ n2 þ n þ 1
is convergent.
As a preliminary observation, note that if n is sufficiently big, we have that
nþ1 n 1
≈ ¼ :
n3 þ n2 þ n þ 1 n3 n2
P Pþ∞ 1
It is, therefore, natural to compare the series þ∞ nþ1
n¼1 n3 þn2 þnþ1 to the series n¼1 n2 which
is a convergent p-series, since p ¼ 2 > 1.
Considering un ¼ n3 þnnþ1
2 þnþ1 and vn ¼ n2 , we have:
1
68
ITexLi.113863
nþ1
¼ lim n þ n þ n þ 1
un 3 2
lim
vn n!þ∞ 1
n2
n3 þ n2
¼ lim 3
n þ n2 þ n þ 1

1
n 1þ
3
n
¼ lim
1 1 1
n3 1 þ þ 2 þ 3
n n n
1
1þ
¼ lim n
1 1 1
1þ þ 2þ 3
n n n
1þ0
¼
1þ0þ0þ0
¼ 1 > 0:
P
Then, applying the limit comparison test, we conclude that the series þ∞ 1
n¼1 n2 and
Pþ∞ nþ1
n3 þn2 þnþ1 have the same nature of convergence and, therefore, the series
Pn¼1
þ∞ nþ1
n¼1 n3 þn2 þnþ1 is convergent.
Example 3.14. Let’s show that the series
X
þ∞
n
n¼1
n2 þnþ1
is divergent.
Consider the sequence ðun Þ whose general term is
n
un ¼
n2 þ n þ 1
and the sequence ðvn Þ defined by vn ¼ n1. Then, we have:

n
un n 2þnþ1
lim ¼ lim
vn 1
n
n2
¼ lim 2
n!þ∞ n þ n þ 1
n2
¼ lim
n!þ∞ 1 1
n 1þ þ 2
2
n n
1
¼ lim
n!þ∞ 1 1
1þ þ 2
n n
1
¼
1þ0þ0
¼ 1 > 0:
69
P 1
Now, since the series þ∞ n¼1 n diverges, it’s the harmonic series, then, by the limit com-
P
parison test, we conclude that þ∞ n
n¼1 n2 þnþ1 is a divergent series.
Example 3.15. In this example, we will again apply the limit comparison test to show
that the series
X
þ∞
n
n¼1
2 ð
n 5
n þ 1Þ
is convergent.
In this case, we consider the sequences ðun Þ defined by
n
un ¼
2n ðn5 þ 1Þ
and ðvn Þ defined by vn ¼ n14 . Then, we have:
n
un 2n ðn5 þ 1Þ
lim ¼ lim
vn 1
n4
n5
¼ lim n 5
2 ðn þ 1 Þ
n5
¼ lim
1
n 2 1þ 5
5 n
n
1
¼ lim
1
2 1þ 5
n
n
1
¼
þ∞
¼ 0:
P
Since þ∞ n¼1 n4 is a convergent series, because it is a p-series with p ¼ 4 > 1, then, by the
1
P
limit comparison test, we conclude that the series þ∞ n
n¼1 2n ðn5 þ1Þ is also convergent.
Note that we could have also used the comparison test to prove the convergence of
Pþ∞ n
n¼1 2n ðn5 þ1Þ. Indeed, since we have, for all n ∈ , that

2n n5 þ 1 > n5
1 1
⇔ n 5 <
2 ðn þ 1Þ n5
n n
⇔ n 5 <
2 ðn þ 1Þ n5
n 1
⇔ n 5 < 4
2 ðn þ 1Þ n
70
ITexLi.113863
P
and since þ∞ 1
4 is convergent, as we already observed, then, by the comparison test, we
Pþ∞ n¼1 nn
have that n¼1 2n ðn5 þ1Þ is convergent.
We will now present another test for verifying the convergence of series of posi-
tive numbers.
P
Theorem 3.10 (Ratio test). Let þ∞ unþ1
n¼1 un be a series of positive terms. If lim un exists,
then the following statements hold.
unþ1 Pþ∞
a. If lim un is a real number smaller than 1, then n¼1 un is a convergent series.
unþ1 Pþ∞
b. If lim un is a real number greater than 1, then n¼1 un is a divergent series.
Pþ∞
c. If lim unþ1
un ¼ 1þ , then n¼1 un is a divergent series.
Pþ∞
d. If lim unþ1
un ¼ 1 , then no conclusion can be drawn about the convergence of n¼1 un .
unþ1 Pþ∞
e. If lim un ¼ þ∞, then n¼1 un is a divergent series.
Pþ∞
Proof. Let n¼1 un be a series of positive terms.
a. Let lim uunþ1

n
¼ r, with r being a real number smaller than 1. Then, there is a
m ∈  such that
unþ1 r1 rþ1

n>m ) <r ¼ : (23)
un 2 2
Since 0 < r < 1, we have that rþ1

2 < 1. Let s ¼ 2 . We can, then, rewrite (23) as
rþ1
unþ1
n>m ) < s,
un
which means that, for n > m, we have:

umþ2
< s⇔umþ2 < umþ1 s:
umþ1
We can, therefore, deduce the following inequalities:
umþ2 < umþ1 s

umþ3 < umþ2 s < umþ1 s2
⋮
umþkþ1 < umþk s < umþ1 sk
Hence, we can conclude that:

umþ1 n
un < umþ1 snm1 ¼ s ,
smþ1
71
for all n ≥ m þ 2.
P
The series þ∞ umþ1 n
n¼mþ2 smþ1 s is a convergent series since it is a geometric series with
ratio s and such that ∣ s ∣ < 1. Then, by the comparison test, we conclude that the
P Pþ∞
series þ∞ n¼mþ1 un is convergent and, therefore, the series n¼1 un is also
convergent.
b. Let lim uunþ1

n
¼ r, with r being a real number bigger than 1. Then, there is a m ∈ 
such that
unþ1 r1 rþ1

n>m ) >r ¼ : (24)
un 2 2
Since r > 1, we have that rþ1

2 > 1. Let s ¼ 2 > 1. Then, we can rewrite (24) as
rþ1
unþ1
n>m ) > s,
un
which means that, for n > m, we have
unþ1
> s⇔unþ1 > sun :
un
We can, therefore, conclude that, for n > m, we have unþ1 > un . But then, the
following statements are true:
umþ2 > umþ1

umþ3 > umþ2 > umþ1
⋮
umþkþ1 > umþk > ⋯ > umþ1 :
Hence, umþkþ1 > umþ1 , ∀k ∈ : Then un > umþ1 > 0 for all n ≥ m þ 2: But then
P
lim un > 0 and, therefore, the series þ∞
n¼1 un diverges.
c. Let lim unþ1

un ¼ 1þ . Then, there exists a m ∈  such that for n > m we have
unþ1
> 1:
un
But that means that un > um , ∀n > m: Then, lim un 6¼ 0 and, therefore, the series
Pþ∞
n¼1 un is divergent.
d. In the case that lim uunþ1 ¼ 1 nothing can be concluded about the convergence
Pþ∞ n
of the series n¼1 un . To show this, we will present two examples.

P 1
On the one hand, if we consider the divergent harmonic series þ∞ n¼1 n, we have
that un ¼ n and
1
72
ITexLi.113863
1
lim
unþ1
¼ ¼ lim n þ 1
un 1
n
n
¼ lim
nþ1
n
¼ lim
1
n 1þ
n
1
¼ lim
1
1þ
n
¼ 1 :
Pþ∞
On another hand, if we consider the convergent series 1
n¼1 n2 , we have un ¼ n12
and
1
unþ1 ðn þ 1Þ2
lim ¼ lim
un 1
n2
n2
¼ lim
ðn þ 1Þ2
n2
¼ lim
2 1
n 1þ þ 2
2
n n
1
¼ lim
2 1
1þ þ 2
n n
¼ 1 :
Thus, nothing can be concluded if lim unþ1

un ¼ 1 .
unþ1
e. Let lim un ¼ þ∞. Then, there is a natural number m such that
unþ1
n>p ) > 1:
un
Therefore, we have that unþ1 > un for n > m. But this means that
umþ2 > umþ1 , umþ3 > umþ2 , … , umþkþ1 > umþk , ∀k ∈ , and, therefore, we
conclude that un > umþ1 for n > m þ 1: Thus,
lim un ≥ umþ1 > 0:
Pþ∞
Hence, lim un 6¼ 0 and n¼1 un is divergent.
□
73
The ratio test is often referred to as the d’Alembert criterion in honor of the French
mathematician Jean-Baptiste le Rond d’Alembert, who is credited for publishing it
first. Some applications are presented next.
Example 3.16. In this example, we will study the convergence of the series
X
þ∞ 3
n
:
n¼1
5n
We start by considering ðun Þ to be the sequence whose general term is
n3
un ¼ :
5n
Now, applying the ratio test, we have
ðn þ 1 Þ3
unþ1 nþ1
lim ¼ ¼ lim 5 3
un n
5n
ðn þ 1Þ3
¼ lim
5n3
3
n3 1 þ n1
¼ lim
5n3
3
1 þ n1
¼ lim
5
1
¼ < 1:
5
P n3
Therefore, we conclude that the series þ∞ n¼1 5n converges.
Example 3.17. In this example, we will again apply the ratio test in order to prove that
the series
X
þ∞
n!
n¼1
n2
is divergent.
Let ðun Þ be the sequence defined by
n!
un ¼ :
n2
Then, we have
74
ITexLi.113863
ðn þ 1Þ!
unþ1 ðn þ 1Þ2
lim ¼ ¼ lim
un n!
" n2 #
ðn þ 1Þ! n2
¼ lim
n! ðn þ 1Þ2
" #
n2
¼ lim ðn þ 1Þ 2
n2 1 þ n1
" #
1
¼ lim ðn þ 1Þ 2
1 þ n1
¼ ðþ∞Þ 1
¼ þ∞:
P
Then, by the ratio test, we conclude that the series þ∞ n!
n¼1 n2 diverges.
Example 3.18. In this example, we will use the ratio test to show that
n
lim ¼ 0:
2n

n
un ¼ :
2n
Then, we have
nþ1
unþ1 nþ1
lim ¼ lim 2 n
un
2n
n þ 1 2n
¼ lim
n 2nþ1
nþ11
¼ lim
n 2
1
n 1þ
1 n
¼ lim
2 n
1
1 1þ
¼ lim n
2 1
1 1
¼ lim 1 þ
2 n
1
¼ < 1:
2
P
Therefore, by the ratio test, we conclude that the series nn¼1 2nn converges and, therefore,
we have that lim 2nn ¼ 0.
75
We will now present another test for the convergence of sequences that is known
as Cauchy’s root test or simply the root test.
P
Theorem 3.11 (Root test). Let þ∞ n¼1 un be a series of non-negative terms such that there
pffiffiffiffiffi
is an order from which all terms of the series are positive. If lim n un exists, then the
following statements are true.
pffiffiffiffiffi P
a. If lim n un is a real number smaller than 1, then þ∞
n¼1 un is a convergent series.
pffiffiffiffiffi P
b. If lim n un is a real number greater than 1, then þ∞
n¼1 un is a divergent series.
pffiffiffiffiffi P
c. If lim n un ¼ 1þ , then þ∞
pffiffiffiffiffi P
d. If lim n un ¼ 1 , then nothing can be concluded about the convergence of þ∞
n¼1 un .
pffiffiffiffiffi P
e. If lim n un ¼ þ∞, then þ∞
P
Proof. Let þ∞n¼1 un be a series of non-negative terms such that there is an order
from which all terms of the series are positive.
pffiffiffiffiffi
a. Let lim n un ¼ r, where r is a real number smaller than 1. Then, there is a
natural number m such that
p ffiffiffiffiffi r1 rþ1

n>m ) n
un < r ¼ :
2 2
Let s ¼ rþ1
2 . note that s is such that 0 < s < 1. Then, for n > m, we have
p ffiffiffiffiffi
n
un < s ⇔ un < sn :
Hence, we have the following inequalities:
umþ1 < smþ1

umþ2 < smþ2
⋮
umþk < smþk , ∀k ∈ 
un < sn , ∀n ≥ m þ 1:
P
Now, since the series þ∞ n
n¼1 s converges, as it is P
a geometric series with a ratio
smaller than 1 in absolute value, then the series þ∞ n
n¼mþ1 s also converges. Then,
Pþ∞
by the comparison test, we conclude that the series n¼mþ1 un is convergent
P
and, therefore, the series þ∞ n¼1 un is also convergent.
pffiffiffiffiffi
b. Let lim n un ¼ r, where r is a real number bigger than 1. Then, there is a natural
number m such that
p ffiffiffiffiffi r1 rþ1

n>m ) n
un > r ¼ :
2 2
76
ITexLi.113863
p ffiffiffiffiffi
Let s ¼ rþ1
2 . Note that s is such that 1 < s < un . Therefore, we have, for
n
n > m, that un > sn : But then we have un > 1 for n > m and, therefore,
P
lim un ≥ 1, that is, lim un 6¼ 0, which, in turn, implies that the series þ∞
n¼1 un is
divergent.
pffiffiffiffiffi
c. Let lim n un ¼ 1þ . Then, there is a natural number m such that
p ffiffiffiffiffi
n>m ) n
un ≥ 1:
Hence, for n > m, we have that un > 1. Then, we can conclude that lim un 6¼ 0
P
and, therefore, the series þ∞
pffiffiffiffiffi
d. In the case that lim n un ¼ 1 nothing can be concluded about the convergence
Pþ∞
of the series n¼1 un . To show this, we will present two examples.
We first consider the series
þ∞
X n
n
:
n¼1
nþ1
In this case, ðun Þ is the sequence defined by

n
n
un ¼ :
nþ1
Pþ∞ n
n
The series n¼1 nþ1 is divergent since we have
n
n
lim un ¼ lim
nþ1
nn
¼ lim
ðn þ 1Þn
nn
¼ lim n n
n 1 þ n1
1
¼ lim n
1 þ n1
1
¼ n
lim 1 þ n1
1
¼ 6¼ 0:
e
Yet, the root test yields:
77
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
n
pffiffiffiffiffi n n
lim n un ¼ lim
nþ1
n
¼ lim
nþ1
n
¼ lim
1
n 1þ
n
1
¼ lim
1
1þ
n
¼ 1 :
Now, we consider the series
X
þ∞
nþ1
:
n¼1
n3
In this case, ðun Þ is the sequence whose general term is
nþ1
un ¼ :
n3
In the case of this series, the root test also yields 1 :

rffiffiffiffiffiffiffiffiffiffiffi
pffiffiffiffiffi n n þ 1
lim n un ¼ lim
n3
p ffiffiffiffiffiffiffiffiffiffiffi
n
nþ1
¼ lim p n
ffiffiffiffiffi
n3
p ffiffiffiffiffiffiffiffiffiffiffi
n
nþ1
¼ lim pffiffiffi 3
ð n nÞ
p ffiffiffi n 1
n
n 1þ
n
¼ lim pffiffiffi pffiffiffi 2
n
nð n nÞ
n 1
1þ
n
¼ lim pffiffiffi 2
ð n nÞ
¼ 1 :
P nþ1
But the series þ∞
n¼1 n3 is a convergent series. Indeed, considering ðvn Þ to be
the sequence defined by vn ¼ n12 , we have
78
ITexLi.113863
nþ1
un 3
lim ¼ lim n
vn 1
n2
n ðn þ 1Þ
2
¼ lim 3
n
1
n3 1 þ
n
¼ lim 3
n
1
¼ lim 1 þ
n
¼ 1 6¼ 0:
P
Since the series þ∞
n¼1 n2 is convergent, it is a p-series with p ¼ 2 > 1, then, by the
1
P nþ1
limit comparison test, we conclude that þ∞ n¼1 n3 is also a convergent series.
pffiffiffiffiffi
Thus, nothing can be concluded if lim n un ¼ 1 .
e. The proof of this assertion is analogous to the one produced for (b).
□
Next, we present a couple of examples where we apply the root test.
þ∞
X
1 þ 2n n
:
n¼1
n

1 þ 2n n
un ¼ :
n
Then, we have
sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi

pffiffiffiffiffi n 1 þ 2n n
lim n un ¼ lim
n
1 þ 2n
¼ lim
n
1
n þ2
n
¼ lim
n
1
þ2
n
¼ lim
1
1
¼ lim þ 2
n
¼ 2 > 1:
79
Pþ∞ 1þ2nn
Then, by the root test, we conclude that n¼1 n is a divergent series.
X
þ∞
1
:
n¼3
ðln nÞn
Let ðun Þ be the sequence whose general term is
1
un ¼ :
ðln nÞn
Then, we have
sffiffiffiffiffiffiffiffiffiffiffiffiffiffi
pffiffiffiffiffi 1
lim n un ¼ lim
n
ðln nÞn
1
¼ lim
ln n
¼ 0 < 1:
P
Then, by the root test, we conclude that the series þ∞ 1
n¼1 ðln nÞn is convergent.
It can be proved that the root test is stronger than the previously presented ratio
test in the sense that whenever the ratio test is conclusive, so is the root test, but the
converse is not true.
P
Proposition 3.3. Let þ∞ n¼1 un be a series of positive terms. If the ratio test proves that
Pþ∞ Pþ∞
n¼1 un is convergent,
P
then the root test also proves that n¼1 un is convergent.
Proof. Let þ∞ n¼1 n be a series
u of positive terms.
Pþ∞
If the ratio test proves that n¼1 un is convergent, then we have that lim uunþ1 n
¼ r,
where r is a real number smaller than 1. Then, there is a natural m such that:
unþ1 r1 rþ1

n>m ) <r ¼ :
un 2 2
Let s ¼ rþ1
2 . Note that 0 < s < 1. Then, for n > m, we have:
unþ1
< s⇔unþ1 < un s:
un
We can, therefore, write the following inequalities
umþ2 < umþ1 s

⋮
umþkþ1 < umþ1 sk , ∀k ∈ :
This means that, for n > m, we have
un < umþ1 snm1 : (25)
80
ITexLi.113863
From (25), we can deduce the following:

pffiffiffi ffiffi pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
n
un < n umþ1 snm1
pffiffiffiffiffiffiffiffiffiffi nm1
¼ n umþ1 s n
pffiffiffiffiffiffiffiffiffiffi mþ1
¼ n umþ1 s1 n :
p ffiffiffiffiffi p ffiffiffiffiffiffiffiffiffiffi mþ1
Applying limits to both sides of the inequality n
un < n umþ1 s1 n , we obtain
pffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffi mþ1

lim n un < lim n umþ1 s1 n
¼ 1 s10
¼ s:
pffiffiffiffiffi
We can, therefore, conclude that lim n un < 1 which means that the root test also
Pþ∞
proves that n¼1 un is a convergent series. □
Proposition 3.3 shows that whenever the ratio test provides information on a series
convergence, the root test will also provide that information. Additionally, there are
some examples of series for which we can determine its convergence with the root test
but not with the ratio test. We present one of those next.
P
Example 3.21. Consider the series þ∞ n¼1 un , where ðun Þ is the sequence defined by
8
> 1
>
< nþ1 , if n is odd
3
un ¼
>
>
: 1 , if n is even:
3n
If we apply the ratio test, we obtain
1
unþ1 nþ1
lim ¼ lim 3
un 1
3n
3n
¼ lim nþ1
3
1
¼
3
if n is even and
1
unþ1 n
lim ¼ lim 3
un 1
3nþ1
3nþ1
¼ lim
3n
¼ 3
81
if n is odd, which means that the limit does not exist rendering the test inconclusive.
However, if we apply the root test, we obtain
rffiffiffiffiffi
pffiffiffiffiffi n 1
lim n un ¼ lim
3n
1
¼
3
if n is even and
rffiffiffiffiffiffiffiffiffi
pffiffiffiffiffi n 1
lim n un ¼ lim nþ1
3
1
¼ lim p ffiffiffiffiffiffiffiffiffi
3nþ1
n
1
¼ lim nþ1
3n
1
¼ lim 1þ1
3 n
1
¼ lim 1
3 3n
1
¼ lim
31
1
¼
3
pffiffiffiffiffi P
if n is odd. Thus, lim n un ¼ 31 < 1 and we conclude that the series þ∞
n¼1 un is convergent.
Note that it is not our intention, with Proposition 3.3 and Example 3.21, to suggest
that the ratio test is useless. Despite the root test being a stronger test, the ratio test, as
some of our previous examples show, proves to be, in many situations, easier to apply
than the root test. Thus, its pertinence is justified.
We will now present the last test in this section that can be applied to series of
positive terms.
Theorem 3.12 (Integral test). Let f : ½1, þ ∞½! þ be a monotonically decreasing
and integrable function and ðun Þ be a sequence such that un ¼ f ðnÞ, ∀n ∈ . Then, the series
Pþ∞ Ð þ∞
n¼1 un is convergent if and only if the improper integral 1 f ðxÞdx converges.
Proof. Let f : ½1, þ ∞½! þ be a monotonically decreasing function and ðun Þ be a
sequence such that un ¼ f ðnÞ, ∀n ∈ . Also, let ðsn Þ denote the sequence of partial sums
P
of þ∞ n¼1 un and ðvn Þ denote the sequence whose general term is
ðn
vn ¼ f ðxÞdx:
1
Before we start the proof itself, we need to deduce some preliminary relations
between the terms of the sequences ðsn Þ and ðvn Þ. Since f is a decreasing function, we
have, for any natural m, that
82
ITexLi.113863
ð mþ1
f ðm þ 1 Þ ≤ f ðxÞdx ≤ f ðmÞ
m
which can be rewritten as:
ð mþ1
umþ1 ≤ f ðxÞdx ≤ um : (26)
m
Applying the principle of (26) recursively, we obtain:
u2 þ u3 þ ⋯ þ un ≤ vn ≤ u1 þ u2 þ ⋯ þ un1 : (27)
From (27), we conclude two things: that
vn ≤ sn1 (28)
and that
u2 þ u3 þ ⋯ þ un ≤ vn
⇔ u1 þ u2 þ u3 þ ⋯ þ un ≤ u1 þ vn (29)
⇔ sn ≤ u1 þ vn :
P
We are now in conditions to prove Theorem 3.12. Suppose that þ∞ n¼1 un is a
convergent series. Then, ðsn Þ is a convergent sequence. Since ðsn Þ converges, then, by
(28), the sequence ðvn Þ is bounded from above. Note that ðvn Þ is an increasing
sequence because f is a positive function. Since ðvn Þ is increasing and is bounded from
above we can conclude that ðvn Þ is also a convergent sequence. Thus, the improper
Ð þ∞
integral 1 f ðxÞdx converges.
Ð þ∞
Conversely, suppose that 1 f ðxÞdx converges. Then, ðvn Þ is a convergent
sequence. Then, by (29), we conclude that ðsn Þ is bounded from above and, therefore,
since it is an increasing sequence, it is also convergent. □
Note that the integral test can be straightforwardly generalized for functions
defined in the interval ½a, þ ∞½, ∀a ∈ . The integral test proves to be very useful for
series whose general term converts immediately into a decreasing and positive func-
tion defined in the interval ½1, þ ∞½ whose primitive is known. For instance, the
integral test can be used to easily determine the nature of convergence of the series
Pþ∞ 1 Pþ∞ 1
n¼1 np , with p ∈ , n lnðnÞ or
1
n¼1 n2 þ1, as we will see in the following examples, with
which we finish this section.
P
Example 3.22. Let us show, using the integral test, that the p-series þ∞ n¼1 np , with p ∈ ,
1
is convergent if p > 1 and divergent if 0 < p ≤ 1.

Let p be a real number greater than 0 and f the function defined by
1
f ðxÞ ¼ :
xp
We start by observing that f is a continuous, positive, and decreasing function in the

interval ½1, þ ∞½. Indeed, since
83
f 0 ðxÞ ¼ ðxp Þ0
¼ pxp1
p
¼
xpþ1
then we conclude that f 0 ðxÞ < 0, ∀x ∈ ½1, þ ∞½, and, therefore, f is decreasing on the
interval ½1, þ ∞½.
Suppose that p > 1. Then, we have
ð þ∞ ðb
1
f ðxÞdx ¼ lim dx
1 b!þ∞ 1 xp
ðb
¼ lim xp dx
b!þ∞ 1
b
xpþ1
¼ lim
b!þ∞ p þ 1 1
" #
1 1
¼ lim
b!þ∞ ðp þ 1Þbp1 p þ 1
1
¼ 0þ
p1
1
¼ :
p1
Ð þ∞
Therefore, we conclude that the improper integral 1 f ðxÞdx is convergent and, by the
P
integral test, the series þ∞
n¼1 np is convergent when p > 1.
1
Now, suppose that p is such that 0 < p < 1. In this case, we have
ð þ∞ ðb
1
f ðxÞdx ¼ lim dx
1 b!þ∞ 1 xp
ðb
¼ lim xp dx
b!þ∞ 1
b
xpþ1
¼ lim
b!þ∞ p þ 1 1
" #
1 1
¼ lim
b!þ∞ ðp þ 1Þbp1 p þ 1
1
¼ þ∞ þ
p1
¼ þ∞:
Ð þ∞
So, in this case, the improper integral 1 f ðxÞdx diverges and, therefore, the series
Pþ∞ 1
n¼1 np , with p ∈ 0, 1½, is divergent by the integral test.
Finally, let’s suppose that p ¼ 1. In this case we have
84
ITexLi.113863
ð þ∞ ðb
f ðxÞdx ¼ lim f ðxÞdx
1 b!þ∞ 1
ðb
1
¼ lim dx
b!þ∞ 1 x
¼ lim ½lnðxÞb1
b!þ∞
¼ lim ðlnðbÞ lnð1ÞÞ

b!þ∞
¼ lim lnðbÞ
b!þ∞
¼ þ∞:
Ð þ∞
Similarly to the last case studied, the improper integral 1 f ðxÞdx diverges and, there-
Pþ∞ 1
fore, the integral test allows us to conclude that the series n¼1 np , with p ¼ 1, is divergent.
Example 3.23. In this last example, we will analyze the convergence of the series
X
þ∞
1
n¼1
ðn þ 2Þ ln ðn þ 2Þ
using the integral test.

We start by considering the function f defined by
1
f ðxÞ ¼ ,
ðx þ 2Þ lnðx þ 2Þ
for x ∈ ½1, þ ∞½.

Note that f ðxÞ > 0, ∀x ∈ ½1, þ ∞½. Additionally, we have
0
f 0 ðxÞ ¼ ððx þ 2Þ lnðx þ 2ÞÞ1

2 xþ2
¼ ð1Þðððx þ 2Þ lnðx þ 2ÞÞ lnðx þ 2Þ þ
xþ2

¼ ð1Þ ððx þ 2Þ lnðx þ 2ÞÞ2 ðlnðx þ 2Þ þ 1Þ
and, therefore, f 0 ðxÞ < 0, ∀x ∈ ½1, þ ∞½ and f is a decreasing function in the interval
½1, þ ∞½:
Now, since
ð þ∞ ðb 1
1
dx ¼ lim x þ 2 dx
1 ðx þ 2Þ lnðx þ 2Þ b!þ∞ 1 lnðx þ 2Þ
¼ ½lnðlnðx þ 2ÞÞb1
¼ lim ½lnðlnðb þ 2ÞÞ lnðlnð3ÞÞ
b!þ∞
¼ þ∞ lnð ln ð3ÞÞ
¼ þ∞,
85
Ð þ∞ 1
we have that the improper integral 1 ðxþ2Þ lnðxþ2Þ dx diverges and, therefore, by the
Pþ∞ 1
integral test, we conclude that the series n¼1 ðnþ2Þ lnðnþ2Þ is divergent.
3.3 Alternating series
In this section, we will focus on the study of the convergence of alternating series.
Informally, an alternating series is just a series whose consecutive terms have opposite
signs. The formal definition is as follows.
Definition 3.5 (Alternating series). An alternating series is a series of the form
X
þ∞
ð1Þn1 un ,
n¼1
where ðun Þ is a sequence of positive terms.

The study of the convergence of alternating series often starts with a test most
commonly known as the alternating series test.
P
Theorem 3.13 (Alternating series test). Let þ∞ n¼1 ð1Þ
n1
un , with ðun Þ being a
sequence of positive terms, be an alternating series. If ðun Þ is a decreasing sequence and
P
lim un ¼ 0, then the alternating series þ∞ n¼1 ð1Þ
n1
un is convergent.
Proof. Let ðun Þ be a decreasing sequence of positive terms such that:
lim un ¼ 0:
Since ðun Þ is decreasing, we have that un > unþ1 , ∀n ∈ . Let ðsn Þ denote the
P
sequence of partial sums of the series þ∞ n¼1 ð1Þ
n1
un . Now, consider the sequence
ðs2n Þ. We have:
s2 ¼ u1 u2 ;
s4 ¼ ðu1 u2 Þ þ ðu3 u4 Þ
s6 ¼ ðu1 u2 Þ þ ðu3 u4 Þ þ ðu5 u6 Þ
⋮
s2n ¼ ðu1 u2 Þ þ ðu3 u4 Þ þ ðu5 u6 Þ þ ⋯ þ ðu2n1 u2n Þ
and, therefore, we conclude that
s2 < s4 < s6 < ⋯ < s2n2 < s2n ,
that is, ðs2n Þ is an increasing sequence.

Additionally, we have
s2 ¼ u1 u2 < u1
s4 ¼ u1 ðu2 u3 Þ u4 < u1
s6 ¼ u1 ðu2 u3 Þ ðu4 u5 Þ u6 < u1
⋮
s2n ¼ u1 ðu2 u3 Þ ðu4 u5 Þ ⋯ ðu2n2 u2n1 Þ u2n Þ < u1 :
86
ITexLi.113863
This means that s2n < u1 , for all n ∈ , and, therefore, we conclude that s2n is a
convergent sequence since it is an increasing sequence bounded from above by u1 .
Now observe that, since
s2n ¼ ðu1 u2 Þ þ ðu3 u4 Þ þ ðu5 u6 Þ þ ⋯ þ ðu2n1 u2n Þ
and
s2n1 ¼ ðu1 u2 Þ þ ðu3 u4 Þ þ ðu5 u6 Þ þ ⋯ þ ðu2n3 u2n2 Þ þ u2n1 ,
we have that s2n ¼ s2n1 u2n . Then, we have
lim s2n ¼ lim s2n1 lim u2n
and since lim un ¼ 0 we conclude that
lim s2n ¼ lim s2n1 :
We can, therefore, conclude that the sequence of partial sums ðsn Þ converges and,
P ∞
thus, the series þ n¼1 ð1Þ
n1
un is convergent. □
The alternating series test is normally credited to the philosopher and mathemati-
cian Gottfried Leibniz and, therefore, it is frequently referred to as Leibniz’s test,
Leibniz’s rule, or even the Leibniz criterion.
When dealing with an alternating series, it is a good strategy to study the series of
the absolute values of the terms as there is a relationship between the convergence of
these two series. That relationship is observed in the next result.
P ∞ Pþ∞
Theorem 3.14. If the series þ n¼1 ∣ un ∣ is convergent, then the series n¼1 un is
convergent.
P ∞
Proof. Let þ n¼1 ∣ un ∣ be a convergent series.
Firstly, observe that
∣ un ∣ ≤ un ≤ ∣ un ∣
⇔ 0 ≤ un þ ∣ un ∣ ≤ 2 ∣ un ∣,
for all n ∈ .
P ∞
Now, note that þ n¼1 ð2 j un j Þ is a convergent series due to our hypothesis that the
Pþ∞
series n¼1 ∣ un ∣ converges.
P ∞
Then, by the comparison test, we have that the series þ n¼1 ðun þj un j Þ is
convergent.
P ∞
Finally, observe that the series þ n¼1 un can be written as
X
þ∞ X
þ∞
un ¼ ðun þj un j j un j Þ
n¼1 n¼1
X
þ∞ X
þ∞
¼ ðun þj un j Þ ∣ un ∣:
n¼1 n¼1
P
Then, since þ∞ n¼1 un is equal to the difference between two convergent series, we
conclude that it is also convergent. □
87
Given a certain series, Theorem 3.14 exhibits the importance of the convergence of
the series of the absolute values of the terms to conclude on the convergence of the
original series. Also, since the result contains only one implication and not an equiva-
lence, it points out to a certain division in the series’ universe between the convergent
series whose series of absolute values also converge and those convergent series whose
series of absolute values do not converge. Such a division is clarified with the next
definition.
P
Definition 3.6 (Absolute and conditional convergence). A series þ∞ n¼1 un is
said to be:
Pþ∞
1. absolutely convergent if n¼1 ∣ un ∣ converges.
Pþ∞ Pþ∞
2. conditionally convergent if n¼1 un converges but n¼1 ∣ un ∣ diverges.
Next, we present a couple of examples.

X
þ∞
1
ð1Þn :
n¼1
n
Let ðun Þ be the sequence whose general term is un ¼ n1 . We have that ðun Þ is a decreasing
sequence of positive terms and that lim un ¼ 0. Then, by the alternating series test, the series
Pþ∞ n
n¼1 ð1Þ un converges. P Pþ∞ 1
The corresponding series of absolute values of þ∞ n1
n¼1 ð1Þ n is the series n¼1 n which is
the already studied divergent harmonic series.
P
Therefore, the series þ∞ n1
n¼1 ð1Þ n is conditionally convergent.
X
þ∞
sinðnπ Þ
:
n¼1
n4
Note that
sinðnπ Þ 1
≤ 4,
n4 n
for all n ∈ , and that the series
X
þ∞
1
n 4
n¼1
is convergent because it is a p-series with p ¼ 4 > 1. Then, by the comparison test, we

conclude that the series
X
þ∞
sinðnπ Þ
n¼1
n4
88
ITexLi.113863
is convergent and, therefore, we can conclude that the series
X
þ∞
sinðnπ Þ
n¼1
n4
is absolutely convergent.
The ratio and root tests presented in Section 3.2 for series of non-negative terms
can be straightforwardly adapted for series with terms with variable sign, including
series of alternating terms, as we present next.
P
Theorem 3.15 (Generalized ratio test). Let þ∞ n¼1 un be a series such that there is a
m ∈  for which we have un 6¼ 0 for all n > m. If lim ∣ ∣uunþ1
n∣
∣
exists, then the following
statements are true.
Pþ∞
a. If lim ∣ u∣ unþ1
n∣
∣
is a real number smaller than 1, then the series n¼1 un is absolutely
convergent.
Pþ∞ Pþ∞
b. If lim ∣ u∣ unþ1
n∣
∣
is a real number greater than 1, then the series n¼1 ∣ un ∣ and n¼1 un
are divergent.
∣ unþ1 ∣ Pþ∞
c. If lim ∣ un ∣ ¼ 1þ , then the series n¼1 un is divergent.
d. If lim ∣ u∣ unþ1
n∣
∣
¼ 1 , then nothing can be concluded about the convergence of the series
Pþ∞ Pþ∞
n¼1 ∣ un ∣ and n¼1 un .
∣ unþ1 ∣ Pþ∞ Pþ∞

e. If lim ∣ un ∣ ¼ þ∞, then the series n¼1 ∣ un ∣ and n¼1 un are divergent.
P
Theorem 3.16 (Generalized root test). Let þ∞ n¼1 un be a series of real numbers. If
p ffiffiffiffiffiffiffiffiffi
lim n ∣ un ∣ exists, then the following statements are true.
p ffiffiffiffiffiffiffiffiffi P
a. If lim n
∣ un ∣ is a real number smaller than 1, then þ∞
n¼1 un is absolutely convergent.
b. If lim n
∣ un ∣ is a real number greater than 1, then þ∞
c. If lim n
∣ un ∣ ¼ 1þ , then þ∞
pffiffiffiffiffiffiffiffiffi
d. If lim n ∣ un ∣ ¼ 1 , nothing can be concluded about the convergence of the series
Pþ∞
n¼1 un .
e. If lim n
∣ un ∣ ¼ þ∞, then þ∞
In the following paragraphs we present two examples with which we finish this
section.
Example 3.26. In this example we will study the convergence of the series
X
þ∞
n!
xn ,
n¼1
2n
89
where x is a real number.

Let ðun ðxÞÞ be the sequence whose general term is given by
n! n
un ðxÞ ¼ x :
2n
Applying the ratio test, we have, for x 6¼ 0,
ðn þ 1Þ! nþ1
x
unþ1 ðxÞ 2nþ1
lim ¼ lim
un ðxÞ n! n
x
2n
ðn þ 1Þ! nþ1 2n
¼ lim x
2nþ1 n!xn
xðn þ 1Þ
¼ lim
2
x
¼ limðn þ 1Þ
2
x
¼ ðþ∞Þ
2
¼ þ∞
P
and, therefore, we conclude that the series þ∞n¼1 2n x is divergent for x 6¼ 0.
n! n
Pþ∞ n! n
If x ¼ 0, then clearly we have that the series n¼1 2n x is convergent with sum equal to
0, since it is an infinite sum of null terms.
Example 3.27. In this example we will study the convergence of the series
X
þ∞
nðx 1Þn ,
n¼1
where x is a real number.

Let ðun ðxÞÞ be the sequence defined by un ðxÞ ¼ ∣ nðx 1Þn ∣. Then, applying the root test,
we have
p ffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
un ðxÞ ¼ lim n ∣ ðx 1Þn ∣
n n
lim
pffiffiffiq ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
¼ lim n n j x 1 jn
n
pffiffiffi
¼ lim n n ∣ x 1 ∣
¼ 1 ∣x 1∣
¼ ∣ x 1 ∣:
Hence, we have:
Pþ∞
• if ∣ x 1 ∣ < 1, that is, if x ∈ 0, 2 ½, then n¼1 nðx 1Þn is absolutely convergent;
Pþ∞
• if ∣ x 1 ∣ > 1, that is, if x ∈ ∞, 0½∪2, þ ∞½, then n¼1 nðx 1Þn diverges.
90
ITexLi.113863
For x ¼ 0 then, we have
X
þ∞ X
þ∞
nðx 1Þn ¼ nð1Þn :
n¼1 n¼1
In this case, let ðun Þ be the sequence defined by un ¼ nð1Þn . Now, if we consider the
sequences of even and odd terms of ðun Þ, ðu2n Þ and ðu2n1 Þ, respectively, then their general
terms are u2n ¼ 2n and u2n1 ¼ 1 2n and we have that lim u2n ¼ þ∞ and
lim u ¼ ∞. Therefore, lim n!þ∞ un does not exist and, consequently, the series
Pþ∞ 2n1 n
n¼1 nð1Þ is divergent.
For x ¼ 2 we have that
X
þ∞ X
þ∞
nðx 1Þn ¼ n
n¼1 n¼1
which is divergent since lim n!þ∞ n ¼ þ∞.

P
Therefore, the series þ∞ n
n¼1 nðx 1Þ is convergent if and only if x ∈ 0, 2 ½.
3.4 Function sequences and series
In this section, we aim to study the convergence of a special type of series whose
terms are functions themselves. These series have a great importance in several
branches of Mathematics such as differential equations or approximation theory.
Before, however, we must also study briefly the corresponding sequences of functions
which are the generating sequences of those series.
3.4.1 Sequences of functions
A sequence of functions is just a sequence whose terms are functions instead of

numbers.
Definition 3.7 (Sequence of functions). A set f n n ∈  , or simply f n , is called a
sequence of functions if, for each natural n, f n is a function f n : A ! , with A, the

common domain of all terms of f n , being a subset of .
Unlike sequences of real numbers, a sequence of functions can converge in differ-
ent manners. The first type of convergence, somewhat the most natural one, is called
pointwise convergence, and it is presented next.
Definition 3.8 (Pointwise convergence). Let f n be a sequence of functions with

common domain A ⊂ . Then, f n converges pointwise to the function f : A !  if, for all
x ∈ A, we have:
lim f n ðxÞ ¼ f ðxÞ:

n!þ∞
In the study of sequences of functions, we deal with two variables: n, the natural
variable of the sequence, and x, the real variable of the functions, which are the terms
of the sequences. In view of this and to avoid confusion, along this section, we will
write lim n!þ∞ f n ðxÞ and not lim f n ðxÞ, although all limits will be evaluated when n
tends to þ∞.
An example of pointwise convergence is presented next.
91

Example 3.28. Let f n be the sequence of functions defined as
f n : ½0, 1 ! 
:
x ↦ xn

The sequence of functions f n converges pointwise to the function
(
0 if x 6¼ 1
f ðx Þ ¼ :
1 if x¼1
Indeed, we have
lim f ðxÞ ¼ lim xn ¼ 0 ¼ f ðxÞ

n!þ∞ n n!þ∞
if x ∈ ½0, 1½ and
lim f ð1Þ ¼ 1 ¼ f ð1Þ:

n!þ∞ n

Saying that a sequence of functions f n converges pointwise to a function f, that
is, that lim n!þ∞ f n ðxÞ ¼ f ðxÞ, ∀x ∈ A, is equivalent to say that
∀x ∈ A, ∀ε > 0, ∃m ∈  : ∀n ∈  : ðn ≥ mÞ ) ∣ f n ðxÞ f ðxÞ ∣ < ε:
This means that, for each x and ε, we can find an order m from which f n ðxÞ differs
from f(x) no more than ε.
A stronger form of convergence is presented next.

Definition 3.9 (Uniform convergence). Let f n be a sequence of functions with

common domain A ⊂ . Then, f n converges uniformly to the function f : A !  if:
∀ε > 0, ∃m ∈  : ∀n ∈ , ∀x ∈ A, ðn > mÞ ) ∣ f n ðxÞ f ðxÞ ∣ < ε:

Example 3.29. Let f n be the sequence of functions whose general term is
fn : A ! 
x,
x ↦
n
where A ⊂ .
The sequence f n converges pointwise to the function f ðxÞ ¼ 0, since, for each x ∈ ,
we have:
x
lim f n ðxÞ ¼ lim
n!þ∞ n!þ∞ n
x
¼
þ∞
¼ 0:

However, uniform convergence of f n is not guaranteed as it depends on the domain A.
92
ITexLi.113863

If A ¼ ½0, 1, for instance, then the series of functions f n converges uniformly to the
function
f : ½0, 1 ! 
:
x ↦ 0
Indeed, let ε > 0. Then

x
∣ f n ðxÞ f ðxÞ ∣ < ε ⇔ 0 <ε
n
∣x∣
⇔ <ε
n
∣x∣
⇔ n> :
ε
Since x ∈ ½0, 1, then
∣x∣ 1
≤ :
ε ε
Therefore, choosing m ¼ 1ε þ 1, for example, we have that
∀ε > 0, ∃m ∈  : ∀n ∈ , ðn > mÞ ) ∣ f n ðxÞ f ðxÞ ∣ < ε, ∀x ∈ ½0, 1,

which means that f n is uniformly convergent.
If A ¼ , nonetheless, it is easy to see that, given an ε > 0, we cannot find an m such that
∣ f n ðxÞ f ðxÞ ∣ < ε, for each x ∈ . In fact, if ε ¼ 1, we would have
x
∣ f n ðxÞ f ðxÞ ∣ < 1 ⇔ 0 <1
n
∣x∣
⇔ <1
n
⇔ n > ∣ x ∣,
which means that there is not any order m such that ∣ f n ðxÞ f ðxÞ ∣ < 1 for each x ∈ .

Hence, f n does not converge uniformly, although it is pointwise convergent.
As it follows from the previous definitions, we have the following relationship
between these two typesof convergence of sequences of functions.

Proposition 3.4. Let f n be a sequence of functions. If f n is uniformly convergent,

then f n is pointwise convergent.

Given a sequence of functions f n and a function f, all of them bounded functions
on a domain A, consider the sequence of real numbers ðMn Þ defined by:
Mn ¼ sup ∣ f n ðxÞ f ðxÞ ∣:

x∈A

Then, the sequence of functions f n converges uniformly to f if and only if
lim Mn ¼ 0.
In the cases that, for all n ∈ , f n and f are continuous functions and A ¼
½a, b, ðMn Þ can be defined by
93
Mn ¼ max ∣ f n ðxÞ f ðxÞ ∣

x∈A

and the sequence f n converges uniformly to the function f if and only if
lim Mn ¼ 0.
Note that in Example 3.29, when A ¼ ½0, 1, we have
Mn ¼ sup f n ðxÞ f ðxÞ

x ∈ ½ 0, 1
¼ max f n ðxÞ f ðxÞ

x ∈ ½ 0, 1
x
¼ max
x ∈ ½ 0, 1 n
1
¼
n

and, therefore, since lim Mn ¼ 0, the sequence f n converges uniformly to the
function f ðxÞ ¼ 0:

Example 3.30. Let f n be the sequence of functions defined by:
f n : ½1, þ ∞½ ! 
:
↦ en x
2
x

We will show that f n is uniformly convergent to the function f defined by
f : ½1, þ ∞½ ! 
:
x ↦ 0
Indeed, we have that
Mn ¼ sup ∣ f n ðxÞ f ðxÞ ∣

x ∈ ½1, þ∞½
sup ∣ en x 0 ∣
2
¼
x ∈ ½1, þ∞½
1
¼ :
en2
Since
1
lim Mn ¼ lim
en2
1
¼
þ∞
¼ 0,

then the sequence of functions f n converges uniformly to the function f.
Note, however, that if the domain of the functions considered was the set 0, þ ∞½,
instead of ½1, þ ∞½, we would have
94
ITexLi.113863
Mn ¼ sup ∣ f n ðxÞ f ðxÞ ∣

x ∈ 0, þ∞½
sup ∣ en x 0 ∣
2
¼
x ∈ 0, þ∞½
¼1

and, therefore, lim Mn ¼ 1 6¼ 0. Hence, f n would not be uniformly convergent.
The next result relates the convergence of a sequence of functions with the conti-
nuity of the functions involved.
Theorem 3.17. Let A be a subset of , a be an interior point of A and f n be a
sequence of functions with domain A that is uniformly convergent to the function f also of
domain A. Then, if for all n ∈ , f n is a continuous function at a, then f is continuous at a.

Proof. Let A be a subset of , a be an interior point of A and f n be a sequence
of functions with domain A that is uniformly convergent to the function f also of
domain A.
Suppose that, for each natural n, f n is a continuous function at a. Since the sequence

f n converges uniformly to f let m be the natural number such that for n > m we have
ε
∣ f n ðxÞ f ðxÞ ∣ < ,
3
∀x ∈ A. Let n ¼ m þ 1. The function f mþ1 is a continuous function at a by hypoth-

esis. Then, let δ be a real positive number such that a δ, a þ δ ½ ⊂ A and
ε
∣ f mþ1 ðxÞ f mþ1 ðaÞ ∣ < ,
3
∀x ∈ a δ, a þ δ ½: Then, for x ∈ a δ, a þ δ ½, we have the following:
∣ f ðxÞ f ðaÞ ∣ < ∣ f ðxÞ f mþ1 ðxÞ þ f mþ1 ðxÞ f mþ1 ðaÞ þ f mþ1 ðaÞ
f ðaÞ ∣
≤ ∣ f ðxÞ f mþ1 ðxÞ ∣ þ ∣ f mþ1 ðxÞ f mþ1 ðaÞ ∣þ
þ∣ f mþ1 ðaÞ f ðaÞ ∣
ε ε ε
≤ þ þ
3 3 3
≤ ε:
Hence, we conclude that f is continuous at a. □

Example 3.31. Consider the sequence of functions f n such that
f n : ½0, 1 ! 
,
x ↦ xn
∀n ∈ , and the real function f also of domain [0, 1] defined by
0, if x ∈ ½0, 1½
f ðxÞ ¼ :
1, if x¼1
95

The sequence f n converges pointwise to the function f. Indeed, for x ∈ ½0, 1½,
we have that
lim f n ðxÞ ¼ lim xn ¼ 0
n!þ∞ n!þ∞
and, for x ¼ 1, we have

lim f n ð1Þ ¼ lim 1n ¼ 1:
n!þ∞ n!þ∞

However, by Theorem 3.17, f n does not converge uniformly to the function f since all
the functions f n are continuous and f is a discontinuous function.
Theorem 3.17 and Example 3.31 provide a powerful criterium for a sequence of
continuous functions not being uniformly convergent to a function. If that function is
not continuous, then uniform convergence does not occur.
Example 3.32. Consider the sequence of functions f n such that
f n : ½0, þ ∞½ ! 
x ,
x ↦
1 þ nx

∀n ∈ . Let us show that f n converges uniformly to the function f of domain ½0, þ ∞½,
such that f ðxÞ ¼ 0.
Firstly, we note that
x
lim f n ðxÞ ¼ lim
n!þ∞ n!þ∞ 1 þ nx
x
¼
þ∞
¼ 0:
Now, since, for x 6¼ 0, we have

x x
≤ ,
1 þ nx nx
then, for each n ∈ , we have
1
0 ≤ f n ðxÞ ≤ :
n
We can, therefore, write that
1
0 ≤ f n ðxÞ f ðxÞ ≤ ,
n
∀x ∈ ½0, þ ∞½, which implies that
1
∀ε > 0, ∃m ¼ ⌊ ⌋ þ 1, n > m ! ∣ f n ðxÞ f ðxÞ ∣ < ε, ∀x ∈ ½0, þ ∞½,
ε

that is, f n converges uniformly to f.
Next, we present a necessary
and sufficient condition for the uniform convergence
of a sequence of functions f n known as the Cauchy criterion.
96
ITexLi.113863

Theorem 3.18 (Cauchy criterion – version 1). The sequence of functions f n , all of
domain A, converges uniformly if and only if:
∀ε > 0, ∃p ∈  : ðn ≥ p∧m ≥ pÞ ) ∣ f n ðxÞ f m ðxÞ ∣ < ε, ∀x ∈ A:
Sometimes Theorem 3.18 is presented in the following manner.

Theorem 3.19 (Cauchy criterion – version 2). The sequence of functions f n , all of
domain A, converges uniformly if and only if:
∀ε > 0, ∃m ∈  : ∀k ∈ , ðn > mÞ ) ∣ f nþk ðxÞ f n ðxÞ ∣ < ε, ∀x ∈ A:
Now, we present a result that deals with sequences of Riemann integrable func-
tions. Informally, a Riemann integrable function is just a function for which we can
approximate the area under its curve using step functions, that is, rectangles with an
equal given length, and that approximation can be as tight as we want. The Riemann
integral, named after the German mathematician Bernhard Riemann, is credited to be
the first rigorous definition of the integral ofa function
on a given interval.
Theorem 3.20. Let A be a subset of  and g n a sequence of Riemann integral
functions in ½a, b ⊂ A converging uniformly to a function g of domain [a, b]. Then, g is
integrable on [a, b] and we have the following equality:
ðb ðb
g ðxÞdx ¼ lim g ðxÞdx:
a n!þ∞ a n
Proof. Since it is out of the scope of our book and it requires some more advanced
calculus tools, we will skip theproof
that g is integrable.
Let A be a subset of  and g n a sequence of Riemann integral functions in
½a, b ⊂ A converging uniformly to a function
g of domain [a, b].
Let ε be a positive real number. Since g n converges uniformly to g, then let m be a
natural number such that, for n > m we have
ε
∣ g n ðxÞ g ðxÞ ∣ < ,
ba
∀x ∈ ½a, b. But then we have

ðb ðb ðb

g n ðxÞdx gðxÞdx ¼ gn ðxÞ g ðxÞ dx
a a a
ðb
≤ ∣ g n ðxÞ gðxÞ ∣ dx
a
ðb
ε
≤ dx
aba
ðb
ε
≤ dx
ba a
ε
≤ ½xb dx
ba a
ε
≤ ðb aÞ
ba
≤ ε:
97
Since we proved that
ðb ðb
∃m ∈  : ðn > mÞ ) g n dx gðxÞdx < ε,
a a
we conclude that
ðb ðb
lim g ðxÞdx ¼ g ðxÞdx:
n!þ∞ a n a
□
We now present an example of an application of Theorem 3.13.
Example 3.33. In this example, we will show that the sequence of functions f n of
domain [0, 1] defined by
8
> 1
>
> 4n2 x , if x ∈ 0,
>
> 4n
>
>
< 1 1
f n ðxÞ ¼ 4n2 x þ 2n , if x ∈ , ,
>
> 4n 2n
>
>
>
> 1
>
:0 , if x ∈ ,1
2n
converges pointwise to the null function, f, defined on [0, 1] but does not converge
uniformly to f.
Let x ∈ 0, 1 . Next, we choose m ∈  such that m ¼ ⌊2x1
⌋ þ 1. Then, for n > m we have
n > 2x, which means that, for n > m, we have 2n < x. Therefore, for n > m, we have that
1 1
f n ðxÞ ¼ 0 and, consequently,
lim f n ðxÞ ¼ 0 ¼ f ðxÞ:

n!þ∞
Now, let x ¼ 0. Then, since f n ð0Þ ¼ 0, we have
lim f n ð0Þ ¼ 0 ¼ f ðxÞ

n!þ∞

and, from all of the above, we conclude that f n converges pointwise to f.
As for uniform convergence, observe that
ð1
1
f n ðxÞdx ¼
0 4
and
ð1 ð1
f ðxÞdx ¼ 0 dx ¼ 0:
0 0
98
ITexLi.113863
Then, by Theorem 3.13, since
ð1 ð1
1
lim f n ðxÞ ¼ 6¼ 0 ¼ f ðxÞdx,
n!þ∞ 0 4 0

we conclude that f n does not converge uniformly to the null function f.
We finish this section with
a result on sequences of differentiable functions.
Theorem 3.21. Let f n be a sequence of differentiable functions with domain [a, b]. If

the sequence f 0n , of the derivatives of f n , converges uniformly to the function g of domain

[a, b] and if there is a c ∈ ½a, b such that f n ðcÞ converges, then the sequence f n converges
uniformly to a function f of domain [a, b] such that f is differentiable and, for all x ∈ ½a, b,
we have that
f 0 ðxÞ ¼ lim f n ðxÞ ¼ g ðxÞ:

n!þ∞

Proof. Let x ∈ ½a, b and c ∈ ½a, b such that the sequence f n ðcÞ converges. By
Theorem 3.13 we can write that
ðx ðx
g ðuÞdu ¼ lim f 0n ðuÞdu:
c n!þ∞ c
Since g is a continuous function, then we have

ðx

g ðuÞdu ¼ lim f n ðxÞ f ðcÞ :
c n!þ∞

Therefore, we conclude that f n converges to the function f defined by
ðx
f ðxÞ ¼ gðuÞdu þ f ðcÞ,
c
∀x ∈ ½a, b. Since g is a continuous function, we can, therefore, conclude that

f 0 ðxÞ ¼ g ðxÞ.

Now, since f 0n converges uniformly to g, let m1 be a natural number such that,
for n > m1 , we have
ε
g ðuÞ f 0n ðuÞ < :
ba

Since f n ðcÞ converges pointwise to f(c), then let m2 be a natural number such
that, for n > m2 , we have
ε
∣ f n ðcÞ f ðcÞ ∣ < :
2
Then, for n > m ¼ maxfm1 , m2 g, we have
99
ðx ð x
∣ f ðxÞ f n ðxÞ ∣ ¼ g ðuÞdu þ f ðcÞ f 0n ðuÞdu þ f n ðcÞ
c a
ð x ðx

< gðuÞdu f 0n ðuÞdu þ f ðcÞ f n ðcÞ
c c
ðx ðx
≤ g ðuÞdu f 0n ðuÞdu þ ∣ f ðcÞ f n ðcÞ ∣
c c
ðx
≤ ∣ gðuÞ f 0n ðuÞ ∣ du þ ∣ f ðcÞ f n ðcÞ ∣
c
ðx
ε ε
≤ du þ
c 2ðb aÞ 2
ε ε
≤ ðx cÞ þ
2ðb aÞ 2
ε ε
≤ ðb aÞ þ
2ðb aÞ 2
ε ε
≤ þ
2 2
ε
≤ :
2

Therefore, for n > m, we have that ∣ f ðxÞ f n ðxÞ ∣ < ε, for all x ∈ ½a, b: Hence, f n
converges uniformly to f. □
3.4.2 Function series
In the following, we define a function series and present some important results.
P
Definition 3.10 (Function series). A function series is an infinite sum þ∞ n¼1 f n , where

f n is a sequence of functions defined in a set A ⊂ .
As we did with series of real numbers, for each function series, we consider P the
correspondent sequence of partial sums of order n, ðSn Þ, defined by: Sn ¼ nk¼1 f k .
Now, we extend the concepts of pointwise and uniform convergence from
sequences of functions to function series.
Definition 3.11 (Pointwise convergence). Let f n be a sequence of functions
P
defined in a set A ⊂  and f : A ! . The function series þ∞ n¼1 f n is said to be pointwise
Pþ∞
convergent to the sum function f in a set A, and we write n¼1 f n ¼ f , if the corresponding
sequence of partial sums ðSn Þ is pointwise convergent to f in A, that is, if
lim Sn ðxÞ ¼ f ðxÞ, ∀x ∈ A.

Definition 3.12 (Uniform convergence). Let f n be a sequence of functions defined
P
in a set A ⊂  and f : A ! . The function series þ∞ n¼1 f n is said to be uniformly convergent
to the function f in A if the corresponding sequence of partial sums ðSn Þ is uniformly
convergent to f.
An example is presented
next.
Example 3.34. Let f n be the sequence of real functions with common domain 1, 1 ½
such that, for each n ∈ , f n ðxÞ ¼ xn .
100
ITexLi.113863
The corresponding sequence of partial sums is:
X
n
Sn ðxÞ ¼ xk1
k¼1
¼ 1 þ x þ x2 þ ⋯ þ xn1
1 xn
¼ :
1x
Hence, for x ∈ 1, 1½, we have
SðxÞ ¼ lim Sn ðxÞ

1 xn
¼ lim
1x
1 lim xn
¼
1x
10
¼
1x
1
¼ :
1x
P
Therefore, we conclude that the series þ∞
n¼1 f n is pointwise convergent to the function:
f : 1, 1½ !

1 :
x ↦
1x

Definition 3.13 (Domain of convergence). Let f n be a sequence of functions of
P
domain A ⊂ : The set of all elements x ∈ A for which the series of real numbers þ∞
n¼1 f n ðxÞ
converges is called the domain of convergence.
Example 3.35. Consider the series of functions defined by
X
þ∞
x
3 sin n :
n
n¼1
5
We start by observing that:

x x
n
3 sin n ≤ ∣ 3 ∣ sin n
n
5 5
x
≤ 3n
5n
n
3
≤ ∣x∣ n
5
n
3
≤ ∣x∣ ,
5
for all x ∈  and for all n ∈ .

101
P 3n
Now, since the series þ∞ n¼1 ∣ x ∣ 5 converges for all x ∈ , then the series
Pþ∞ n x
3 sin is absolutely convergent for any x ∈ . Therefore, the domain of convergence
Pþ∞ n x
n¼1 n
5
of the function series n¼1 3 sin 5n is .
We continue this survey by presenting two relevant results for determining uni-
form convergence of function series.
Theorem
3.22 (Cauchy’s uniform convergence criterion for series of functions).
Let f n be a sequence of functions with common domain A ⊂ . Then, the series of
P
functions þ∞ n¼1 f n is uniformly convergent in A if
kX
þm
∀ε > 0, ∃p ∈  : ∀m ∈ , ∀x ∈ A, ðk > pÞ ) f i ðxÞ < ε:
i¼kþ1

Theorem 3.23 (Weierstrass M-Test). Let f n be a sequence of functions with
common domain A ⊂ . Then, if there is a sequence ðMn Þ of real non-negative numbers
such that
Pþ∞
i. n¼1 Mn is convergent and
ii. ∀n ∈ , ∀x ∈ A, ∣ f n ðxÞ ∣ ≤ Mn ,
P
the series of functions þ∞ n¼1 f n is absolutely and uniformly convergent in A.
Pþ∞
Proof. Let n¼1 Mn be a convergent series of real numbers and f n be a sequence
of functions with common domain A such that ∣ f n ðxÞ ∣ ≤ Mn , for all x ∈ A and all
n ∈ .
Let x be an element of A. Since ∣ f n ðxÞ ∣ ≤ Mn , for each n ∈ , then the series
Pþ∞ Pþ∞
n¼1 ∣ f n ðxÞ ∣ is convergent in A. Thus, n¼1 f n ðxÞ is absolutely convergent for any
x ∈ A.
P
Regarding uniform convergence, since þ∞ n¼1 Mn is convergent, then the sequence
of partial sums
X
n
Sn ¼ Mk
k¼1
is also convergent and, therefore, it is a Cauchy sequence. Then, for all ε > 0, there
exists a m ∈  such that, for all n > m and for all k ∈ , we have
Mnþ1 þ Mnþ2 þ ⋯ þ Mnþk < ε:
We can, therefore, write:
∣ f nþ1 ðxÞ þ f nþ2 ðxÞ þ ⋯ þ f nþk ðxÞ ∣

≤ ∣ f nþ1 ðxÞ ∣ þ ∣ f nþ2 ðxÞ ∣ þ ⋯ þ ∣ f nþk ðxÞ ∣
≤ Mnþ1 þ Mnþ2 þ ⋯ þ Mnþk
< ε,
for all x ∈ A. Hence, by Cauchy’s uniform convergence criterion for series of

P
functions, we can conclude that the series þ∞
n¼1 f n is uniformly convergent. □
102
ITexLi.113863
Example 3.36. In this example, we will show that the series of functions defined as
X
þ∞
cosðnxÞ
n¼1
n3
converges uniformly in .
Indeed, we have that
cosðnxÞ 1
≤ 3,
n3 n
∀x ∈ .
P
Now, since þ∞ 1
n¼1 3 is convergent, then, by the Weierstrass M-Test, we can conclude that
Pþ∞ cosðnxÞn
the series n¼1 n3 is uniformly convergent in .
We finish this subsection by presenting three results that we consider to be
important to refer to in the context of this survey on function series.
Theorem 3.24. Let f n be a sequence of continuous functions in the domain
P
½a, b ⊂ . If þ∞ n¼1 f n is uniformly convergent to a function f in [a, b], then f is continuous
in [a, b].
Theorem 3.25. Let f n be a sequence of integrable functions in the domain ½a, b ⊂ .
P
If þ∞ n¼1 f n is uniformly convergent to a function f in [a, b], then f is integrable in [a, b] and
we have that
ðb þ∞ ð b
X
f ðxÞdx ¼ f n ðxÞdx:
a n¼1 a

Theorem 3.26. Let f n be a sequence of class C1 functions in the domain ½a, b ⊂ 
such that:
Pþ∞
i. there is a c ∈ ½a, b such that the series n¼1 f n ðcÞ is convergent and
Pþ∞ 0
ii. the series n¼1 f n is uniformly convergent to a function g in [a, b].
P
Then, the function series þ∞
n¼1 f n is uniformly convergent to a differentiable function f
0
in [a, b] such that f ¼ g on [a, b] and
X
þ∞
f 0 ðxÞ ¼ f 0n ðxÞ, ∀x ∈ ½a, b:
n¼1
3.5 Power series
In this section, we will study a particular class of function series that hold a
great deal of importance for their applications, specifically regarding function
analysis.
103
3.5.1 Definition and properties
In a very simple way, we can define a power series as an infinite sum like the
following:
X
þ∞
an xn ¼ a0 þ a1 x þ a2 x2 þ ⋯ þ an xn þ ⋯,
n¼0
where each ai , i ∈ , is a real number.

The type of power series presented above is normally referred to as a power series
centered at 0. A more general definition of a power series centered at any real number
x0 is given next.
Definition 3.14 (Power series). A power series centered at x0 is a function series
Pþ∞
n¼0 f n , where each function f n can be defined as
f n ðxÞ ¼ an ðx x0 Þn ,
with an ∈ , ∀n ∈ 0 .
Given a power series centered at x0 , it is easy to observe that x0 belongs to its
convergence domain. However, what can we conclude on the convergence of a power
series that we know that is convergent at a certain point x ¼ b? The following result
sheds some light on this matter.
P
Theorem 3.27. Let þ∞ n¼0 an x be a power series centered at 0, with an ∈ , ∀n ∈ .
n
Then:
P Pþ∞
1. if þ∞ n
n¼0 an x converges at a non-zero value c, then
n
n¼0 an x converges absolutely at
any value x such that ∣ x ∣ < ∣ c ∣;
P Pþ∞
2. if þ∞ n
n¼0 an x diverges at a certain non-zero value c, then
n
n¼0 an x diverges for any x
such that ∣ x ∣ > ∣ c ∣.
Pþ∞
Proof. Let n¼0 an x
n
be a power series centered at 0, with an ∈ , ∀n ∈ . Let
c ∈ nf0g.
Pþ∞ n
1. If the series n¼0 an c is convergent, then we have that
lim an cn ¼ 0:
n!þ∞
But since the sequence whose general term is an cn is convergent, then it is

bounded. Therefore, there exists a real positive number L such that
∣ an cn ∣ < L, ∀n ∈ :
Now, let b ∈  such that ∣ b ∣ < ∣ c ∣ and denote by Sn ðbÞ the sequence of partial
104
ITexLi.113863
Pþ ∞ n
sums of the series n¼0 ∣ an b ∣: Then, we have:
X
n
Sn ðbÞ ¼ ∣ ak bk ∣
k¼0
X
n
¼ ∣ ak ∣ j b jk
k¼0
X
n
j b jk
¼ ∣ ak ∣ ∣ ck ∣
k¼0 j c jk
X
n
b k
¼ ∣ ak ck ∣
k¼0
c
X
n
b k
≤ L
k¼0
c
Xn
b k
¼ L
k¼0
c
Hence, we have
b n
b 1
Sn ðb Þ ≤ L c
c 1 b
c
and, consequently,
b 1
Sn ðbÞ ≤ L :
c 1 b
c
Now, since Sn ðbÞ is an increasing sequence of positive terms that is bounded from
above, then we conclude that Sn ðbÞ is convergent and, therefore, the series
Pþ∞ k Pþ∞
k¼0 ∣ ak b ∣ is convergent which means that the series
n
n¼0 an x is absolutely
convergent for any x ∈ , such that ∣ x ∣ < ∣ c ∣:
Pþ∞
2. If n¼0 an c is a divergent series, let d ∈  be such that ∣ d ∣ > ∣ c ∣.
n
Pþ∞
If n¼0 an dn is convergent, then, by assertion (1.) proved before, the series
Pþ∞ n
n¼0 an x would be absolutely convergent on all values x such that ∣ x ∣ < ∣ d ∣. But,
Pþ∞
since ∣ c ∣ < ∣ d ∣, this would mean that the series n¼0 an cn would be convergent,
which is absurd because we are supposing precisely the contrary. Therefore, we
P
conclude that þ∞ n¼0 an x is divergent for any x ∈  such that ∣ x ∣ > ∣ c ∣.
n
□
The assertions of Theorem 3.27 allow us to conclude that:
P
• if the power series þ∞
n¼0 an x is convergent at c ∈ nf0g, then it is also absolutely
n
convergent in the interval ∣c∣, ∣c∣½ and

105
P
• if the power series þ∞n¼0 an x diverges at a point c ∈ nf0g, then it is divergent in
n
the union of intervals ∞, ∣ c ∣ ½∪ ∣ c ∣, þ ∞½.

P
If there exists a real positive number R such that the power series þ∞ n
n¼0 an x con-
verges absolutely in the interval ] R, R[ and is divergent in the interval ∞,
R½∪R, þ ∞½ (where at the points –R and R the convergence of the series needs to be
checked), then this interval ] R, R[ is called the interval of convergence while the number
P
R is called the radius of convergence both regarding the power series þ∞ n
n¼0 an x in x. If
there is not such a real number, then we define the radius of convergence, R, as þ∞.
Note that, although we studied the particular case of a power series centered at 0,
analogous conclusions can be drawn in the general case of a power series centered at
any given point x0 ∈ :
Next we will show how we can obtain the radius of convergence of power series.
Theorem 3.28. Let ðan Þ be a sequence of real numbers such that:
∃m ∈  : ∀n ∈ , n > m ) an 6¼ 0
Pþ∞
and n¼0 an ðx x0 Þn be a power series centered at x0 ∈ . If lim an
anþ1 exists, then the
radius of convergence of this power series, R, is given by
an
R ¼ lim :
anþ1
Pþ∞
Proof. Let n¼0 an ðx x0 Þn be a power series satisfying the conditions of Theorem
an
3.28 and suppose that lim anþ1 exists.
We will divide this proof in three cases.
Pþ∞
1. If lim anþ1
an ¼ 0þ , then the power series n¼0 an ðx x0 Þn is absolutely
convergent for any x. Indeed, in this case we have:
!
anþ1 ðx x0 Þnþ1 anþ1 jx x0 jnþ1
lim ¼ lim
jan ðx x0 Þn j an jx x0 jn

anþ1
¼ lim jx x0 j
an
¼ 0jx x0 j
¼ 0 < 1:
P
Therefore, by the ratio test, the series þ∞ n
n¼1 an ðx x0 Þ is absolutely convergent
for any x ∈  and, consequently, the radius of convergence of the power series is
R ¼ þ∞: Observe that
an 1 1
lim ¼ lim ¼ ¼ þ∞:
anþ1 anþ1 0þ
an
Hence, R ¼ lim an
anþ1 .
106
ITexLi.113863
anþ1
2. If lim an ¼ þ∞, then we have:
!
lim ¼ lim

anþ1
¼ lim jx x0 j
an
¼ þ∞ jx x0 j
¼ þ∞,
if x 6¼ x0 . If x ¼ x0 , then
anþ1 xnþ1
lim ¼ 0:
jan xn j
Therefore, the power series converges only at x ¼ x0 . Hence,
an 1
R ¼ lim ¼ ¼ 0:
anþ1 þ∞
anþ1
3. If lim an ¼ L ∈ , then we have:
!
lim ¼ lim
anþ1
¼ lim jx x0 j
an
¼ Ljx x0 j:
Therefore, if
1
L ∣ x x0 ∣ < 1⇔∣ x x0 ∣ < ,
L
P
then the power series þ∞ n
n¼0 an ðx x0 Þ is absolutely convergent.
Hence, the radius of convergence of the power series is
1 an
R¼ ¼ lim :
L n!þ∞ anþ1
□
Pþ∞ n
If R is the radius of convergence of a power series n¼0 an ðx x0 Þ in x, then
the interval of convergence of the power series is the interval x0 R, x0 þ R½.
Note that, in general, the interval of convergence and the domain of
convergence of the power series are not the same, as the latter may include
the extreme points of the former in the case that the power series converges in
such points.
107
Example 3.37. In this example, we will find the domain of convergence of the power
series defined as:
X
þ∞
5n x n :
n¼0
Applying Theorem 3.28, we have:
jan j
R ¼ lim
janþ1 j
5n
¼ lim nþ1
5
1
¼ lim
5
1
¼ :
5

Hence, the interval of convergence of the given power series is: 15, 15 .
Now let’s analyze the power series for x ¼ 15 and x ¼ 15.
P
If x ¼ 15 , the power series becomes the series ∞n¼0 1 which is divergent since
lim an ¼ lim 1 ¼ 1 6¼ 0.
P
If x ¼ 15, the power series becomes the series ∞ n
n¼0 ð1Þ which is also divergent since
n
lim an ¼ lim ð1Þ does not exist.
P
Hence, the domain of convergence of the power series þ∞ n¼0 5 x is the interval 5, 5 ,
n n 1 1
that is, in this case, the domain of convergence corresponds to the interval of convergence.
We can also compute the radius of convergence through the root test presented
before.
Theorem 3.29. Let ðan Þ be a sequence of real numbers such that:
∃m ∈  : ∀n ∈ , n > m ) an 6¼ 0
Pþ∞
and n¼0 an ðx x0 Þn be a power series centered at x0 ∈ . If lim p ffiffiffiffiffiffiffi
1
exists, then the
n
∣ an ∣
radius of convergence of this power series, R, is given by
1
R ¼ lim p ffiffiffiffiffiffiffiffiffi :
n
∣ an ∣
Proof. The proof is similar to the one given for Theorem 3.28. We leave it as an
exercise for the reader. □
Example 3.38. In this example, we will determine the domain of convergence of the
power series defined by
X
þ∞
ðx 1Þn
:
n¼0
nn
By Theorem 3.29, we have:
108
ITexLi.113863
1
R ¼ lim pffiffiffiffiffiffiffiffiffi
n
∣ an ∣
1
¼ lim rffiffiffiffiffi
n 1
nn
1
¼ lim
1
n
1
¼
0þ
¼ þ∞:
Therefore, the domain of convergence of the given power series is the interval of conver-
gence ∞, þ ∞½ or simply the set .
The following two results relate power series convergence with the concepts of
differentiation and integration. The proofs are left out because they require some
mathematical tools outside the scope of this book.
P
Theorem 3.30. Let þ∞ n
n¼0 an ðx x0 Þ be a power series whose radius of convergence is
R > 0, convergent to the function f in x0 R, x0 þ R ½. Then, the power series
Pþ∞ n1
n¼1 nan ðx x0 Þ has radius of convergence equal to R and converges to the derivative of
0
f, f , in x0 R, x0 þ R½: Additionally, this series converges uniformly in any closed interval
½a, b ⊂ x0 R, x0 þ R½.
A couple of examples are presented next.
P
Example 3.39. The series þ∞ n¼0 x converge to the function f ðxÞ ¼ 1x and the radius of
n 1
Pþ∞ n1
this series is R ¼ 1. Hence, the series n¼1 nx is convergent to the function f 0 ðxÞ ¼ ð1x
1
Þ2
in 1, 1½. Therefore, we can write
X
þ∞
f 0 ðxÞ ¼ nxn1 , ∀x ∈ 1, 1½:
n¼1
As for the extreme points of the interval 1, 1½, it is easy to conclude that the series
Pþ∞ n1
n¼1 nx is divergent if x ¼ 1 or x ¼ 1.
Example 3.40. In this example, we will consider the power series defined by:
X
þ∞
xn
:
n¼2
nðn 1Þ
We have that the radius of convergence of this series is R ¼ 1 and the domain of
convergence is [–1, 1].
However, the series
X
þ∞
nxn1 X
þ∞
xn1
¼
n¼2
nðn 1Þ n¼2 ðn 1Þ
with radius of convergence 1 is convergent at x ¼ 1 but diverges at x ¼ 1. Thus, its

domain of convergence is ½1, 1½.
109
P
Theorem 3.31. Let þ∞ n
n¼0 an ðx x0 Þ be a power series whose radius of convergence is
P
R > 0, convergent to the function f in x0 R, x0 þ R½: Then, the power series þ∞ xnþ1
n¼0 an nþ1
has radius of convergence equal to R and converges to the primitive of f, such that its value at
x0 is 0, in x0 R, x0 þ R½: Additionally, this series is uniformly convergent in any closed
interval ½a, b ⊂ x0 R, x0 þ R½ and we have
ð X
þ∞
an
f ðxÞdx ¼ ðx x0 Þnþ1 , ∀x ∈ x0 R, x0 þ R½:
n¼1
n þ 1
3.5.2 Representation of a function as a power series
Arguably, the most notable application of power series is the representation of

functions that are infinitely differentiable at a real number of its domain. By
representing a function by its power series, we can easily obtain approximations of
that function at a given point with an associated error as small as we want. It was,
therefore, natural not to finish this section on power series without summarizing the
key aspects on this important application.
Indeed, in some situations, given a function f, it is important to obtain a power
series that converges to f at least at a certain domain. Such a power series is said to
represent the function f in the domain of convergence.
Before we proceed any further, let’s establish that this representation of a function
P
by a power series is unique. In fact, if we have simultaneously that f ðxÞ ¼ þ∞ n¼0 an x
n
Pþ∞
and f ðxÞ ¼ n¼0 bn x for all x in a certain domain D, then we have
n
X
þ∞ X
þ∞
an xn ¼ bn xn
n¼0 n¼0
(30)
X
þ∞
⇔ ðan bn Þ x ¼ 0
n
n¼0
⇔ a0 b0 þ ða1 b1 Þx þ ða2 b2 Þx2 þ ⋯ (31)
Since 0 naturally belongs to the domain of convergence D, if we make x ¼ 0 in

(31), we obtain:
a0 b0 ¼ 0⇔a0 ¼ b0 :
Now, differentiating equality (30) established above, with respect to x, we obtain
P
þ∞
nðan bn Þxn1 ¼ 0
n¼1 (32)
⇔ a1 b1 þ 2ða2 b2 Þx þ 3ða3 b3 Þx þ ⋯ 2
Again, since 0 ∈ D, substituting x ¼ 0 in (32) we obtain
a1 b1 ¼ 0⇔a1 ¼ b1 :
110
ITexLi.113863
Applying this reasoning successively, we can conclude that an ¼ bn , ∀n ∈ , that is,

the two developments of f as a power series are equal. A similar conclusion would be
drawn if the series were centered in a given non-zero point.
It is known that, if f admits derivatives until the order n þ 1 in an open interval
I ¼a, b½, with a < b, centered at x ¼ x0 , then f can be written in the following way:
0 00
f 0 ðx0 Þ f 0 ðx0 Þ f 0 ðx0 Þ
f ðxÞ ¼ f ðx0 Þ þ ðx x0 Þ þ ðx x0 Þ2 þ ðx x0 Þ3
1! 2! 3!
f ðnÞ ðx0 Þ
þ ⋯þ ðx x0 Þn þ rn ðx0 , xÞ (33)
n!
X
n
f ðkÞ ðx0 Þ
¼ ðx x0 Þk þ rn ðx0 , xÞ,
k¼0
k!
for x ∈ I, where rn ðx0 , xÞ is a remainder of this development of f given by
f ðnþ1Þ ðcÞ
rn ðx0 , xÞ ¼ ðx x0 Þnþ1 ,
ðn þ 1Þ!
for some c ∈ x0 , x ½ if x > x0 , or c ∈ x, x0 ½ if x < x0 .

In the cases that we have that
lim rn ða, xÞ ¼ 0, (34)

n!þ∞
then, applying limits to equality (33), we deduce
X
n
f ðkÞ ðaÞ
lim f ðxÞ ¼ lim ðx aÞk þ lim rn ða, xÞ
n!þ∞ n!þ∞
k¼0
k! n!þ∞
(35)
X
þ∞ ðnÞ
f ðx 0 Þ
⇔ f ðxÞ ¼ ðx x0 Þn :
n¼0
n!
The development of the function f presented in (35) is called the Taylor series of the
function f at the point x ¼ x0 . In the particular case where the series is centered at
the point x0 ¼ 0, it is called a Maclaurin series instead of a Taylor series centered at 0.
The names of these series were given after the British mathematicians who developed
and applied them in their respective works: the English mathematician Brook Taylor
(1685–1731) and the Scottish mathematician Colin Maclaurin (1698–1746).
From what was presented above, it should be clear that the Taylor series represents
the function f if and only if equality (34) holds. However, by reading this very brief
survey on the Taylor series, one might assume that it is enough for the function f to be
infinitely differentiable to be able to be represented by a Taylor series. In general, this
is not true, that is, although we can always construct the Taylor series of f if f is
infinitely differentiable, the convergence
X
þ∞ ðnÞ
f ðx0 Þ
f ðxÞ ¼ ðx x0 Þn
n¼0
n!
111
is not guaranteed. A classic example that enlightens this situation is the function f
such that f ð0Þ ¼ 0 and f ðxÞ ¼ ex2 , if x 6¼ 0. It is easy to see that
1
X
þ∞ ðnÞ
f ð0Þ
f ðxÞ 6¼ xn ,
n¼0
n!
for all x 6¼ 0, since that in this case the Taylor series is the null series because all the
derivatives of f at 0 are null.
The functions, f, that can be represented by a Taylor series at a certain domain D
are, therefore, required to have a “very good behavior” that goes beyond being
infinitely differentiable. These functions are usually called analytic which means
simply that they can be written, for any point x ¼ x0 ∈ D, as
X
þ∞
f ðxÞ ¼ an ðx x0 Þn
n¼0
for some coefficients an and with a positive radius of convergence.

As we suggested in the beginning of this subsection, given a Taylor series at x0 that
represents a function f on a certain interval I containing x0 , we are in conditions to
obtain an approximation of f at a point x ∈ I by considering a finite partial sum of
the Taylor series at x0 up to the order that will provide the desired approximation
significance. This partial sum up to the order, let’s say, m is called the Taylor
polynomial of order m of f at the point x0 , which we denote by T m ðx0 , xÞ, and we can
write, for x ∈ I:
f ðxÞ ≈ T m ðx0 , xÞ
Xm ð kÞ
f ðx0 Þ
¼ ðx x0 Þk
k¼0
k!
f 0 ðx 0 Þ f ðmÞ ðx0 Þ
¼ f ðx 0 Þ þ ðx x0 Þ1 þ ⋯ þ ðx x0 Þm :
1! m!
Besides adding summands to the Taylor polynomial by increasing the value of m,

we can also improve the quality of the approximation by considering a center of the
power series, x0 , closer to the x in I on which we are evaluating the function f. By
doing so, we make the successive powers ðx x0 Þk smaller and, consequently, the
Taylor polynomial will provide a more accurate approximation to f(x).
Example 3.41. In this example, we will illustrate the approximation technique
presented before with the function f ðxÞ ¼ ex .
For this function, we have the following Taylor series centered at 0:
e0 2 e0 2
f ðxÞ ¼ e0 þ e0 x þ x þ x þ⋯
2! 3!
x2 x3
¼ 1 þ x þ þ þ ⋯,
2 6
with the first Taylor polynomials of orders 1 to 3 being:

112
ITexLi.113863
T 1 ð0, xÞ ¼ 1 þ x;
x2
T 2 ð0, xÞ ¼ 1 þ x þ ;
2
x2 x3
T 3 ð0, xÞ ¼ 1 þ x þ þ :
2 6
In Figure 3, we show graphically how these approximations improve as they come closer
to the function f with the increase of the order of the polynomial.
P
Let þ∞ n
n¼0 an ðx x0 Þ be a power series with radius of convergence R. Using Theo-
Pþ∞
rem 3.30, we know that, if f ðxÞ ¼ n¼0 an ðx x0 Þn , then the series
Pþ∞ n1
n¼1 nan ðx x0 Þ as also radius of convergence equal to R and
X
þ∞
f 0 ðx Þ ¼ an nxn1 , ∀x ∈ x0 R, x0 þ R½:
n¼1
If we keep applying Theorem 3.30 recursively, we conclude that, for

x ∈ x0 R, x0 þ R½, we have
0
f 0 ðxÞ ¼ nðn 1Þ an ðx bÞn2 ;
00
f 0 ðxÞ ¼ nðn 1Þ ðn 2Þ an ðx bÞn3 ;
⋮
ðpÞ
f ðxÞ¼ nðn 1Þ⋯ðn p þ 1Þ an ðx bÞnp , ∀p ∈ :
P
Therefore, a function f ðxÞ ¼ þ∞ n
n¼0 an ðx x0 Þ , for x ∈ x0 R, x0 þ R½, is infinitely
differentiable at x0 and its derivatives are such that f ðpÞ ðbÞ ¼ p!ap , ∀p ∈ . We can,
therefore, conclude that
f ðpÞ ðx0 Þ
ap ¼ , ∀p ∈ :
p!
P
Hence, we conclude that if f ðxÞ ¼ þ∞ n
n¼0 an ðx x0 Þ , for x ∈ x0 R, x0 þ R½, then
the development of f as a Taylor series at the point x0 is precisely f ðxÞ ¼
Figure 3.
Approximations with Taylor polynomials to the function ex .
113
Pþ∞
n¼0 an ðx x0 Þn and this development is valid in x0 R, x0 þ R½, while for x < x0
R or x > x0 þ R the Taylor series is divergent. The points x ¼ x0 R and x ¼ x0 þ R
must be objective of a particular analysis.
Next, we present an example where we deduce the Maclaurin series of a function.
Example 3.42. Consider the function f defined by
1
f ðxÞ ¼
ð1 xÞ2
for any x belonging to the interval ∞, 1½. Let’s determine the Maclaurin series of f and
its domain of convergence.
We know that
1 X
þ∞
¼ 1 þ x þ x2 þ ⋯þ ¼ xn ,
1x n¼0
for x ∈ 1, 1½. The radius of convergence of this series is

p ffiffiffi
R ¼ lim 1 ¼ 1:
n
Now, applying Theorem 3.30, we have that

0 X
þ∞
1
¼ nxn1
1x n¼1
1 X
þ∞
⇔ ¼ nxn1 ,
ð1 xÞ2 n¼1
for x ∈ 1, 1½.
P
Therefore, the Maclaurin series of f is þ∞ nxn1 and this development is valid
Pþ∞ n1
n¼1
for x ∈ 1, 1½: Additionally, since the series n¼1 nx is divergent at x ¼ 1 and x ¼ 1,
then the function f can be developed as a Maclaurin series only in the interval 1, 1½.
We finish this section with a list of the most commonly used Maclaurin series
presented in Table 2.
f(x) Maclaurin series Conv. domain
ex 2 3 P
þ∞
xn 
1 þ x þ x2! þ x3! þ ⋯ ¼ n!
n¼0
ln(1 + x) P
þ∞ ] 1, 1]
ð1Þnþ1 xn
2 3 n
x x2 þ x3 ⋯ ¼
n¼1
ln(1 x) 2 3 P
þ∞
xn [1, 1[
x x2 x3 ⋯ ¼ n
n¼1
1 P
þ∞ ] 1, 1[
1x 1 þ x þ x2 þ x 3 þ ⋯ ¼ xn
n¼0
1 P
þ∞ ] 1, 1[
ð1xÞ2 1 þ 2x þ 3x2 þ ⋯ ¼ nxn1
n¼1

ð 1 þ xÞ n P
þ∞ n k ] 1, 1[
1 þ nx þ nðn1Þ
2! þ ⋯ ¼ x
k¼0 k
114
ITexLi.113863
f(x) Maclaurin series Conv. domain
sin(x) 3 5 P
þ∞
ð1Þn 2nþ1 
x x3! þ x5! ⋯ ¼ ð2nþ1Þ! x
n¼0
cos(x) 2 4 P
þ∞
ð1Þn 
1 x2! þ x4! ⋯ ¼ ð2nÞ! x2n
n¼0

π2, π
3 5 7
tan(x) x þ x3 þ 2x
15 þ 315 þ ⋯
17x
2
arcsin(x) 3 5 P
þ∞
ð2nÞ! [1, 1]
x þ x6 þ 3x
40 þ ⋯ ¼ 4n ðn!Þ2 ð2nþ1Þ
x2nþ1
n¼0
arccos(x) π 3
π 5 P
þ∞
ð2nÞ! [1, 1]
2 x x6 3x
40 ⋯ ¼ 2 4n ðn!Þ2 ð2nþ1Þ
x2nþ1
n¼0
arctan(x) 3 5 P
þ∞
ð1Þn [1, 1]
x x3 þ x5 ⋯ ¼ 2nþ1 x2nþ1
n¼0
sinh(x) 3 5 P
þ∞ 
x þ x3! þ x5! þ ⋯ ¼ 1
ð2nþ1Þ! x
2nþ1
n¼0
cosh(x) 2 4 P
þ∞ 
1 þ x2! þ x4! þ ⋯ ¼ 1
ð2nÞ! x
2n
n¼0

π2, π
3 5 7
tanh(x) x x3 þ 2x
15 315 þ ⋯
17x
2
Table 2.
Most common Maclaurin series.
3.6 Suggested exercises
Convergence of series
1. Classify, using the definition of convergent series, the following series.

Pþ∞ 1n
a. n¼1 5 ;
Pþ∞ pffiffiffiffiffiffiffiffiffiffiffi pffiffiffiffiffiffiffiffiffiffiffi
b. n¼1 nþ2 nþ1 ;
Pþ∞ 1
c. n¼1 n2 ðnþ1
1
Þ2 ;
Pþ∞ nþ1
d. n¼2 ln n ;
Pþ∞
e. 1
n¼1 ðnþ2Þ! n!1 :
2. Classify the following series and, if possible, obtain their sum.

Pþ∞ 1
a. n¼1 3n ;
Pþ∞ 1
b. n¼1 nðnþ1Þ ;
Pþ∞ 1
c. n¼1 2n þ ðnþ1Þ1ðnþ2Þ ;
115
Pþ∞ 1

d. n¼1 nðnþ3Þ ;
Pþ∞ nþ1

e. n¼1 n2 ðnþ2Þ2 ;
Pþ∞ 3
f. n¼1 2n ;
Pþ∞ nþ1 3nþ1
g. n¼1 ð1Þ 5n ;
Pþ∞ 1
h. n¼1 ðnþ1Þ ðnþ2Þ ðnþ3Þ ;
Pþ∞ 3n þ4n þ5n

i. n¼1 5n :
3. Show that
Pþ∞
a. 1
n¼1 ð2n1Þð2nþ5Þ ¼ 90
23
;
Pþ∞ pffiffiffiffiffiffi pffiffi

n
b. n¼1
nþ1
p ffiffiffiffiffiffiffiffi ¼ 1;
2 n þn
Pþ∞ 1
n2 þ2n
c. n¼1 ln ðnþ1Þ2
¼ ln 2 :
Pþ∞
4. Obtain the general term of the series n¼1 un knowing that
n
Sn ¼ :
nþ1
5. Find the general term of the following series (suppose that the terms follow the
same rule of formation) and analyze their convergence. Also, in each case where
convergence occurs, calculate the series sum.
a. 2 þ 4 þ 8 þ 16 þ 32 þ ⋯;
b. 2 31 þ 91 271 þ ⋯;
c. 1
13 þ 35
1
þ 57
1
þ ⋯;
d. 1 þ 31 þ 312 þ 313 þ ⋯:
Series of non-negative terms

6. Prove Theorem 3.8 using the Integral Test.
7. Use the comparison test to analyze the convergence of the following series.
Pþ∞ 1
a. n¼1 n2 þnþ1 ;
116
ITexLi.113863
Pþ∞
b. pffiffiffiffiffiffiffi
1 ffi
;
n¼1 n3 þ1
Pþ∞ lnðnþ1Þ
c. n¼2 n ;
Pþ∞ 1
d. n¼1 ðnþ1Þ ðnþ2Þ ðnþ3Þ ;
Pþ∞ 3n þ2n
e. n¼1 7n þ5n ;
Pþ∞
f. pffiffiffiffiffiffi
1
:
n¼1 nþ1
8. Show that the series
X
þ∞
lnð3Þ lnð4Þ lnð5Þ⋯ lnðnÞ
n¼1
ðn þ 2Þ!
is convergent.
9. Use the limit comparison test to analyze the convergence of the following series
of real numbers.
Pþ∞ n2 þ1
a. n¼1 n4 þn2 þ1 ;
Pþ∞
b. ffiffi
pnþ1
n¼1 n nþnþ1 ;
Pþ∞ 3n þ4n
c. n¼1 7 n þ6n ;
Pþ∞
d. p3 ffiffiffiffiffiffiffiffiffiffiffiffi
1
;
n¼1 n2 þnþ1
Pþ∞ n
e. n¼1 2n2 þ1 ;
Pþ∞ p3 ffiffi
npffiffi
f. n¼1 ðnþ2Þ n ;
Pþ∞ 1
g. n¼1 ðð1Þn þ5Þ5n ;
Pþ∞
n¼1 nðnþ1Þðnþ2Þðnþ3 :
1
h.
10.Analyze the convergence of the following series by applying the ratio test.
Pþ∞ 1
a. n¼1 n! ;
Pþ∞ n
b. n¼1 5n ;
Pþ∞ pffiffiffiffiffiffi
nþ1
c. n¼1 6n ;
117
Pþ∞ 5n
d. n¼1 n2 ;
Pþ∞ nn
e. n¼2 n! ;
n4
f. ðln 5Þn
;
Pþ∞ n4 5n
g. n¼1 n! ;
Pþ∞ 5n
h. n¼1 7 n þ8n ;
Pþ∞
i. nþ3
n¼1 5n ðnþ1Þ2 :
11. Use the root test to analyze the convergence of the following series.
Pþ∞ nn
a. n¼1 52nþ1 ;
Pþ∞ n
b. n¼1 n ;
Pþ∞ n
n2
c. n¼1 nþ1 ;
Pþ∞ n2
2n
d. n¼1 2n2 þ1 ;
Pþ∞ nþ1n
e. n¼1 n ;
Pþ∞ 1
f. n¼1 ðln nÞn ;
Pþ∞
g. 1
n¼1 ðnþ1Þn :
12. Analyze the convergence of the following series by applying the integral test.
Pþ∞ 1
a. 3 n ln n ;
Pþ∞ n
b. n¼1 ð4n3Þð4n1Þ ;
Pþ∞ 1
c. n¼1 n2 þ1 ;
Pþ∞ n
d. n¼1 1þn2 ;
Pþ∞
e. pffiffiffiffiffiffiffiffi
1
;
n¼1 2nþ1
Pþ∞ 1
f. n¼1 n ln 4 n ;
118
ITexLi.113863
Pþ∞
g. n¼1 ðnþ1Þ
p1ffiffiffiffiffiffiffiffiffiffiffiffi ;
lnðnþ1Þ
Pþ∞ 3 n4
h. n¼1 n e ;
Pþ∞
i. n¼1 ðn þ 1Þeðnþ1Þ ;
13. Find the nature of convergence of the following series analyzing its general
term.
Pþ∞ n2 þ2nþ1
a. n¼1 n2 þ2nþ3 ;
Pþ∞ nþ12n
b. n¼1 n ;
Pþ∞
c. pffiffiffiffiffiffiffiffiffi
n
;
n¼1 2n2 þ1
Pþ∞ p ffiffiffi
d. n¼1 n;
n
Pþ∞ 2n
e. n¼1 n :
Pþ∞
14.Show that, if the series of positive terms n¼1 un converges, then the series
þ∞
X
nþ1
un
n¼1
2n þ 1
also converges.
Series of alternating terms
15. Show that the following series are convergent using the alternating series test
and study their absolute convergence.
Pþ∞ ð1Þn
a. n¼1 n2 ;
Pþ∞ ð1Þn1
b. p5 ffiffi ;
n¼1 n
Pþ∞ ð1Þn
c. n¼1 n5n ;
Pþ∞ ð1Þn
d. n¼1 nðnþ1Þð2nþ1Þ ;
Pþ∞ ð1Þn
e. n¼1 ð3nÞ! ;
16. Study the convergence of each of the following series stating if the convergence
is absolute or conditional.
119
Pþ∞
a. n¼1 sinðnπ Þ 2nþ1
n ;
Pþ∞ n
b. n¼1 ð1Þ cosðnπ Þ;
Pþ∞ n nþ1
c. n¼1 ð1Þ 3n ;
Pþ∞ nþ2 2n
d. n¼1 ð1Þ 3nþ1
;
Pþ∞ n n2
e. N¼1 ð1Þ n4 þn2 þ1 ;
Pþ∞ sinðnπ Þ
f. n¼1 n4 ;
Pþ∞
g. n¼1 cosðnπ Þnen ;
Pþ∞ cosðnÞ
h. n¼1 1n3 :
Function series

17. In each of the following cases, study if the given sequence of functions f n is
pointwise or uniformly convergent on the given domain D.
2
a. f n ðxÞ ¼ xn , D ¼ ∞, þ ∞½;
b. f n ðxÞ ¼ 1þn14 x2 , D ¼ ½2, 3;
c. f n ðxÞ ¼ 2þnx
x
, D ¼ ½2, þ ∞½;
sinðnxÞ
d. f n ðxÞ ¼ n4 , D ¼ ;
qffiffiffiffiffiffiffiffiffiffiffiffiffiffi
e. f n ðxÞ ¼ x4 þ n12 , D ¼ ;

18. Show that the sequence of functions f n , where f n ðxÞ ¼ xn ð1 xn Þ, is
pointwise convergent but not uniformly convergent to the null function on the
interval [0, 1].

19. Show that the sequence of functions f n , where, for each n ∈  we have f n :
½0, 1 !  and
8
>
> n 1
< x, if x ∈ 0, 1
n1 n
f n ðxÞ ¼ ,
>
>
: nðx 1Þ, if x ∈ 1 1 , 1
n
is pointwise convergent but not uniformly convergent to the function f such

that f ðxÞ ¼ x, ∀x ∈ ½0, 1½ and f ð1Þ ¼ 0.
120
ITexLi.113863

20.Consider the sequence of functions f n such that
f n : ½0, 1 ! 
:
x ↦ ð1 xÞxn

a. Show that, for all x ∈ ½0, 1, the sequence of functions f n is pointwise
convergent to the function f such that f ðxÞ ¼ 0.
b. Show that
n
n 1
max f n ðxÞ f ðxÞ ¼ :
x ∈ ½0, 1 nþ1 nþ1

c. Show that the sequence f n is uniformly convergent to the function f
presented in (a).

21. Consider the sequence of functions f n such that
f n : ½0, 1 ! 
n2 x2 :
x ↦
1 þ n4 x4

a. Show that f n is pointwise convergent to the null function f on [0, 1].

b. Verify that f n n1 ¼ 21 and justify why the sequence of functions f n does
not converge uniformly to the null function f.
22. Using the Weierstrass M-test, show that the following series of functions are
uniformly convergent on the domain presented in each item.
Pþ∞
a. 1
n¼1 n4 þx8 , D ¼ ;
Pþ∞ sinðnxÞ
b. n¼1 n4 , D ¼ ;
Pþ∞ xn
c. n¼1 3
n
sin 5n , D ¼ ½0, 1;
Pþ∞ ð1Þn
d. n¼1 nðxþ3Þ3 , D ¼ ½1, þ ∞½:

23. Consider the sequence of functions f n defined as f n : ½0, 1 !  such that
8
> 1
> 2
> 2n x, if x ∈ 1,
>
> 2n
>
>
< 1 1
f n ðxÞ ¼ 2nðnx 1Þ, if x ∈ , :
>
> 2n n
>
>
>
> 1
>
: 0, if x ∈ , 1
n
121

Show that f n is pointwise convergent but not uniformly convergent to the
null function on [0, 1].
[Suggestion: start by showing that the area between the null function and f n is
equal to 21.]
24.Show that the series

X
þ∞
2nx2
n¼1
1 þ n4 x4
is uniformly convergent on the interval [0, 1].

P
25. Show that if þ∞ n¼1 an is absolutely convergent, then the series of functions
Pþ∞ an x4n
n¼1 1þx4n is uniformly convergent on the interval [0, 1].
Power series
26. Find the radius of convergence and the interval of convergence of the following
power series.
Pþ∞ ðx1Þn
a. n¼1 n! ;
Pþ∞ n n
b. n¼1 5 x ;
Pþ∞ xn
c. n¼1 n ;
Pþ∞ xn
d. n¼1 n2n ;
Pþ∞ x2nffiffi
e. p ;
n¼1 n
Pþ∞ xn
f. n¼1 n2 :
27. Show that
1 X
þ∞
¼ ðn þ 1Þxn
ð1 xÞ2 n¼0
for ∣x∣ < 1.
28. Show that
6x X
þ∞
¼ ½ð5Þn 1xn
5x2 4x 1 n¼1
for ∣x∣ < 15.
122
ITexLi.113863
29. Show that
X
þ∞
xn
lnð1 þ xÞ ¼ ð1Þn1
n¼1
n
for all x ∈ 1, 1.
30.Obtain the interval of convergence of the following power series.
Pþ∞ ðx3Þn
a. n¼0 2n ;
Pþ∞ 4n
b. n¼0 ð2nÞ! ðx 1Þn ;
Pþ∞ 1 n
c. n¼1 n x ;
Pþ∞
d. n
n¼1 2n ðx 2Þn ;
Pþ∞
n¼1 nþ1 x :
n n
e.
31. Find the domain of convergence of the following series.

Pþ∞ n n
a. n¼1 ð1Þ x ;
Pþ∞ xnffiffi
np
b. n¼1 ð1Þ n
;
Pþ∞ xn
c. n¼1 n! ;
Pþ∞ n n2 xn
d. n¼1 ð1Þ 3n ;
Pþ∞ n!ðx1Þn
e. n¼1 135⋯ð2n1Þ! :
32. Consider the function f ðxÞ ¼ ex .

2
a. Determine the Maclaurin series of f.
b. Using the Taylor polynomial of order 4 of the series determined in (a),

obtain an approximation of f(2).
33. Consider the function g ðxÞ ¼ sinð2x 8Þ.
a. Determine the Taylor series of g at x0 ¼ 4.
b. Using the Taylor polynomial of order 5 of the series determined in (a),

obtain an approximation of g(5).
123
Solutions to suggested exercises
Sequences
1. (a) un ¼ n2 n
(b) 156 ¼ u13
(c) n ∈ f21,22,23,24g
2. (a) v1 ¼ 14
5 , v2 ¼ 2 , v3 ¼ 11 , v4 ¼ 7 , v5 ¼ 17
5 26 16 38
(b) No, it is not a term of ðvn Þ since n ¼ 14

3 ∉ℕ
(c) vnþ1 vn ¼ ð3nþ512

Þð3nþ2Þ < 0
(d) 2 < vn ≤ 14
5
(e) n ¼ 13
n 3
3. (a) wn ¼ ð1Þ n2
ðnþ1Þ
(b) w10 ¼ 1000

121
(c) lim wn ¼ þ∞, for even n and lim wn ¼ ∞, for odd n
4. (a) u1 ¼ 5, u2 ¼ 25, u3 ¼ 85, u4 ¼ 265

nþ1
(c) ðun Þ is increasing unþ1 un ¼ 10
3 3 3n > 0
(d) lim un ¼ þ∞
n pffiffi pffiffi o
5. (a) 1, 22, 0, 22 , 1
(b) 1 ≤ vn ≤ 1
(c) v1 < v2 > v3
6. (a) ðwn Þ is increasing
(b) 1 ≤ wn < 5
7. 10
8. (a) v5 ¼ 13
9
(b) Yes, it is the term v13

124
ITexLi.113863

(c) ðvn Þ is decreasing vnþ1 vn ¼ ð2nþ5Þð2nþ3Þ
8
<0
(d) 1 < vn ≤ 15
9. ðwn Þ is increasing, since wnþ1 wn ¼ ð1Þnþ1 ð1Þn þ 3 > 1 > 0
10. (a) un ¼ 3n 1
(b) 63
(c) 610
11. (a) vnþ1 vn ¼ 35
(b) vnþ1 vn < 0
(c) –745
12. (a) d ¼ 4, ðwn Þ is increasing
(b) wn ¼ 4n 8
(c) m = 15
13. (a) un ¼ 4 51n
(b) un is decreasing
(c) 0 < un ≤ 4
(d) 5 519
14. (a) vvnþn 1 ¼ 21
(b) ðvn Þ is decreasing
(c) 0 < vn ≤ 5
(d) 625365
15. p ¼ 52
18. (a) ∞
(b) þ∞
(c) þ∞
(d) 32
125
(e) 0
(f) 0
(g) 21
(h) 0
(i) 3
(j) 0
(k) 0
(l) þ∞
19. 0
20. (b) 390625

4
22. They are all decreasing sequences.
23. (a) Show that wnþ1 wn < 0
(b) Show that 0 < wn < n1 and apply the Squeeze Theorem
25. Converges to 3
Pn
26. Use the fact that k¼1 k
2
¼ nðnþ1Þ6ð2nþ1Þ
27. (a) e1
(b) 0
(c) 1
(d) þ∞
(e) þ∞
(f) 1
(g) 0
(h) þ∞
(i) 9
31. lim wn ¼ 2
36. 17
126
ITexLi.113863
Series
1. (a) Convergent
(b) Divergent
(c) Convergent
(d) Divergent
(e) Convergent
2. (a) Convergent; S ¼ 21
(b) Convergent; S ¼ 1
(c) Convergent; S ¼ 32
(d) Convergent; S ¼ 1181
(e) Convergent; S ¼ 16
5
(f) Convergent; S ¼ 3
(g) Convergent; S ¼ 98
(h) Convergent; S ¼ 121
(i) Divergent; lim an ¼ 1
4. un ¼ nðnþ1Þ
1
5. (a) an ¼ 2n. The series diverges since lim an ¼ þ∞

n1
(b) an ¼ 2 31 . The series converges and its sum is S ¼ 32
(c) an ¼ ð2n1Þ1ð2nþ1Þ. The series converges and its sum is S ¼ 21

1n
(d) an ¼ 3 . The series converges and its sum is S ¼ 32
7. (a) Convergent
(b) Convergent
(c) Divergent
(d) Convergent
127
(e) Convergent
(f) Divergent
9. (a) Convergent
(b) Divergent
(c) Convergent
(d) Divergent
(e) Divergent
(f) Convergent
(g) Convergent
(h) Convergent
10. (a) Convergent
(b) Convergent
(c) Convergent
(d) Divergent
(e) Divergent
(f) Convergent
(g) Convergent
(h) Convergent
(i) Convergent
11. (a) Divergent
(b) Divergent
(c) Convergent
(d) Convergent
(e) Divergent
(f) Convergent
(g) Convergent
128
ITexLi.113863
12. (a) Divergent
(b) Divergent
(c) Convergent
(d) Divergent
(e) Divergent
(f) Convergent
(g) Divergent
(h) Convergent
(i) Convergent
13. (a) Divergent
(b) Divergent
(c) Divergent
(d) Divergent
(e) Divergent
15. (a) Absolutely convergent
(b) Conditionally convergent
(c) Absolutely convergent
(d) Absolutely convergent
(e) Absolutely convergent
16. (a) Divergent
(b) Divergent
(c) Absolutely convergent
(d) Absolutely convergent
(e) Absolutely convergent
(f) Absolutely convergent
(g) Divergent
(h) Absolutely convergent
17. (a) Pointwise convergent
(b) Uniformly convergent

129
(c) Uniformly convergent
(d) Uniformly convergent
(e) Uniformly convergent
26. (a) R ¼ þ∞ and I ¼ ∞, þ ∞½

(b) R ¼ 15 and I ¼ 15, 15
(c) R ¼ 1 and I ¼ ½1, 1½
(d) R ¼ 2 and I ¼ ½2, 2½
(e) R ¼ 1, and I ¼ 1, 1½
(f) R ¼ 1 and I ¼ ½1, 1
30. (a) I ¼1, 5½
(b) I ¼ ∞, þ ∞½
(c) I ¼ ½1, 1½
(d) I ¼0, 4½
(e) I ¼ 1, 1½
31. (a) D ¼ 1, 1½
(b) D ¼ 1, 1
(c) D ¼ 
(d) D ¼ 3, 3½
(e) D ¼ 2, 2
Pþ∞ ð1Þn x2n
32. (a) n¼0 n!
(b) 5
Pþ∞ ð1Þnþ1 22n1 ðx4Þ2n1
33. (a) n¼1 ð2n1Þ!
(d) 14
15
130
ITexLi.113863
Acknowledgements
Luís Vieira was partially supported by CMUP (UID/MAT/00144/2013), which is

funded by FCT (Portugal) with national (MEC) and European structural funds
through the programs FEDER under the partnership agreement PT2020.
References
[1] Apostol T. Calculus. 2nd ed. Vol. 1.

New York: John Wiley & Sons; 1967
[2] Courant R, John F. Introduction to

Calculus and Analysis. Vol. 1. New York:
Interscience Publishers; 1974
[3] Marsden J. Elementary Classical

Analysis. San Francisco: W. H. Freeman
and Company; 1974
[4] Nikolsky S. A Course of Mathematical

Analysis. Moskow: MIR; 1977
[5] Piskounov N. Calcul Différentiel et

Intégral. Moskow: MIR; 1974
[6] Shilov G. Analyse Mathématique.

Moscow: MIR; 1973
[7] Smirnov VI. A Course of Higher

Mathematics. Vol. 1. Reading: Pergamon
Press; 1964
[8] Spivak M. Calculus. 4th ed.

Cambridge: Publish or Perish, Inc.; 2008
[9] Stewart J. Calculus. 4th ed. Boston:

Brooks/Cole Publishing Company; 1999
[10] Trench W. Introduction to Real

Analysis. San Antonio: Pearson
Education; 2003

Sequences and Series

Uploaded by

Copyright:

Available Formats

Sequences and Series

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Sequences and Series

Uploaded by

Copyright:

Available Formats

Contents

Keywords: calculus, real analysis, sequences of functions, sequences of real numbers,

1. Preliminaries and notation

1.1 Basic set notation

and the set of integers is denoted by ,

 ¼ f … , 3, 2, 1, 0, 1, 2, 3, … g:

Next, the set of rational numbers is denoted by :

Finally, we denote by  the set of real numbers, that is,

1.2 Upper and lower bounds of sets

1. X is said to be bounded from above if there is a real number M such that

2. X is said to be bounded from below if there is a real number m such that

3. X is said to be bounded if it is bounded from above and below and unbounded if it is

Definition 1.2 (Maximum, minimum, supremum, and infimum of sets). Let X be a

1. the maximum of X if it is an element of X and if it is an upper bound of X.

2. the minimum of X if it is an element of X and if it is a lower bound of X.

3. the supremum of X if it is the smaller of all upper bounds of X.

4. the infimum of X if it is the bigger of all lower bounds of X.

• Hence, X is a bounded set.

• X does not have a minimum.

The concept of neighborhood of a number will also be used ahead. We define it in a

1.3 Relations and functions

Given two sets X and Y, the Cartesian product X Y is defined as

and its elements are called ordered pairs.

A binary relation or correspondence, R, between X and Y is simply a subset of X Y.

1. If, for all x ∈ X and y1 , y2 ∈ Y we have xRy1 ∧ xRy2 ) y1 ¼ y2 , then R is said to be a

3. If R is functional and left-total, it is called a function.

4. If, for all x1 , x2 ∈ X and y ∈ Y we have x1 Ry ∧ x2 Ry ) x1 ¼ x2 , then R is said to be

6. A function that is injective and surjective is said to be bijective or a bijection.

1.4 Limits and derivatives of functions

If the limit in (1) is a real number, then f is said to be differentiable at x.

ðfgÞ0 ðxÞ ¼ f 0 ðxÞg ðxÞ þ f ðxÞg0 ðxÞ

provided, in the latter case, that gðxÞ 6¼ 0.

ð f ∘ g Þ0 ðxÞ ¼ f 0 ðg ðxÞÞ  g0 ðxÞ, (2)

provided that g is differentiable at x and f is differentiable at g(x).

f(x) f 0 ðxÞ f(x) f 0 ðxÞ

axn naxn1 aun anun1 u0

sin(x) cos(x) sin(u) u0 cosðuÞ

cos(x) sin(x) cos(u) u0 sinðuÞ

tan(x) sec 2 ðxÞ tan(u) u0 sec 2 ðuÞ

sinh(x) cosh(x) sinh(u) u0 coshðuÞ

cosh(x) sinh(x) cosh(u) u0 sinhðuÞ

tanh(x) sech2 ðxÞ tanh(u) u0 sech2 ðuÞ

so long as the limit is finite or ∞.

so long as the limit is finite or ∞.

Example 1.4. Now we will compute

using l’Hôpital’s Rule. We have

Example 1.5. In this example, we will apply l’Hôpital’s Rule to calculate

1.5 Mathematical induction

1. the statement P(n) is true for n ¼ 1, that is, P(1) is true;

∀m ∈ , PðmÞ ) Pðm þ 1Þ:

The principle of mathematical induction comes from the axiomatic definition of 

For n ¼ 1, we have only one element in the sum, therefore,

and the formula is valid.

Indeed, using the hypothesis of induction, we have

ðn þ 3Þ! ≥ 2nþ3 : (3)

Indeed, for n ¼ 1, we have

which means that (3) is valid.

Let’s prove that

ðm þ 1 þ 3Þ! ≥ 2mþ1þ3 ⇔ðm þ 4Þ! ≥ 2mþ4 :

ðm þ 4Þ! ¼ ðm þ 4Þðm þ 3Þ!

ð f ∘ g Þ0 ðxÞ ¼ f 0 ðg ðxÞÞ g0 ðxÞ, (2)

so long as the limit is finite or ∞.

so long as the limit is finite or ∞.