CHAPTER 2 : SEQUENCE

2.1: Convergent Sequences

Definition 2.1.1 : A sequence is a function whose domain is the set of positive
integers. A sequence of real numbers is a function whose codomain is the set
R.

f :N R

Examples 2.1.1:
(i)

The function f : N R defined as

f n 2n 1, n N ,

is a sequence of real numbers. This sequence can be written as

2n 1 : n N or
(ii)

3 , 5 , ...., 2n 1, ......

The function g : N R defined as

1
g n
, n N,
n!
is a sequence of real numbers. This sequence can be written as
1

1 1
1
: n N or , ,... , or it can also be written as an : n N
n!
n!

1! 2 !
with
a1 1
a
an1 n , n 1, 2, 3,.......
n 1

Definition 2.1.2 : The sequence an is said to converge to a number a if for

each 0 , there exists a positive integer N such that for n N , we have

an a .
Note : We can rewrite the above statement as

lim an a , or
n

an a , when n .
1

The sequence

an

is convergent if lim an a . If
n

an

does not converge, the

sequence an is divergent (or diverges).

What definition 2.1.2 actually means is that, as we go out farther and farther in
the sequence, the terms get closer and closer to a . The mathematical definition
carefully quantifies the vague terms farther and closer. Given an error tolerance
0 , it must be possible to find a positive integer N such that every term in the
sequence past the N th term is within of a . This means that the set
an : n N is contained in the interval a , a . The positive number
measures how close the terms are to a and the positive integer N records how
far out in the sequence you must go in order to guarantee that the terms are
within of a .
The important point is that can be any positive number. No matter what
positive number is, it is always possible to find a positive integer N and the
choice of N depends on . It is worth pointing out that 0 comes first, then an
appropriate N is sought. Different value of , may gives different N . For
1
instance, if we have a sequence an that converges to 0 . If 0.1 , then
n

1
1
0 0.1 . We then obtained, n
10 . Now, we can choose N 11 . Thus
0 .1
n
for n 11 , we have

1
0 0 .1 .
n

If we change the value of to 0.01, then we have

for n

1
0 0.01 . It follows that
n

1
100 . Then we can choose N 101 . Therefore, for n 101, we
0.01

1
0 0.01 . Another way to express definition 2.1.2 in a more
n
understandable way is that, when lim an a , then for each 0 , there exists

have

an integer, N such that an a , a

illustrated as follows.

a1 , a2 , a3 ....a N 1

for n N and this can be

a N a N 1 a N 2 ......

Figure 2.1 : Terms of

an when lim
an a .
n

When this happens, we may have finitely many terms of an located outside the
interval a , a .

Example 2.1.2:
Use the definition of convergent to prove that the following sequence converges.
(i)

lim
n

1
0, for p 0 ,
np

(ii)

lim

1 2n 2
2
.
2
4n 3n
3

Solutions :
(i)

Remember that 0 is given first, then N must be chosen such that

1
for all n N , we have p 0 .
n
1

p
1
1
1
p
0 n n .
np

From the Archimedean property, there exist an integer N such that

1

1
1
0 whenever n N .
N . Thus for this N , we have

np
(ii)

Given 0 , then N must be chosen such that when n N , we have

1 2n 2 2

.
4n 3n 2 3
1 2n 2 2
3 8n
8 8n
8

,
2
2
2
4n 3n 3
3 4n 3n
3 3n 3n
9n

and
8
8
.
n
9n
9

Choose N such that N

8
, then for n N ,
9

1 2n 2 2

.
4n 3n 2 3
Definition 2.1.1 also implies that if the sequence an does not converge to a
number c R , then there exist 0 such that for each integer positive N , there
exists an integer n N for which an c .

Example 2.1.3 : Consider the sequence an 1

converge.

show that a

does not

Solution : Assume that the sequence converges, that is, assume that there is an
n
a R such that lim 1 a . Given that
n

1,
an
1,
Using

n even
n odd

, then the length of the interval

a , a is 1, but the distance

between two consecutives terms, 1 and -1 is 2, that is 1 1 2 . Thus both 1

and -1 cannot be in the interval a , a . That is an and an1 cannot be both

in this interval. Therefore for each n N , there exist an integer N 0 n such that

an a and we conclude
a N a . This contradicts the assumption that lim
n
0

that the sequence an does not converge or an diverges.

Theorem 2.1.1: The limit of a convergent sequence is unique.

Proof : Let an be a convergent sequence and suppose that an converges to
both the numbers a and b , that is lim an a and lim an b . We need to show
n

that a b . For each 0 , lim an a implies that there exist a positives integer
n

N1 , such that an a for n N1 . Similarly, there exists a positive integer N 2

2

an b

such that

for n N 2 . Now for N max N1 , N 2 , we have the

2

followings:
a b a an an b a an an b

2 2

ab .

Since 0 is arbitrary, it follows that, a b 0 , and it only make sense when

a b 0 . Hence a b .

Definition 2.1.3: A sequence an is said to be bounded if there exist a number

M 0 such that an M for all n N .

Examples 2.1.4:
(i)

n
n
The sequence
1 for all n N .
is bounded because
n 1
n 1

(ii)

The sequence

e
n

is not bounded. This is because e n as

n . Then for any m 0 , there exist n0 N such that e

e is not bounded from above and therefore a

n

no

m . Thus

is not bounded.

However an is bounded below, since e 0 for all n N .

n

Theorem 2.1.2 : A convergent sequence is bounded.

Proof : Let an be a convergent sequence and let be the limit of this
sequence. Then for 1 , there exist a positive integer N such that

an 1 for all n N .
From the fact that

an an ,
it follow that

Define M maks

an 1 ,

for n N .

, a2 ,....., aN 1 , 1 .

Then,

an M ,

for all n N ,

and therefore the sequence an is bounded.

Note : The converse may not be true, because there are bounded sequences
n
that do not converge. The sequence 1 is one simple example.

1 1, 1, 1, 1,.........
n

and

an 1 for all n N , but

1
n

does not

converge.
Theorem 2.1.2 also implies that an unbounded sequence does not converge.

Theorem 2.1.3: Suppose that

Then
(i)
(ii)
(iii)
(iv)

an

converge to a and

bn

converges to b .

The sequence an bn converges to a b .

The sequence an bn converges to a b .
The sequence an bn converges to ab .
a
The sequence n converges to a / b , provided that bn 0 for all n
bn
and b 0 .

Proof :
(i) For 0 , lim an a implies that there exist N1 N such that for all n N1 ,
n

an a

,
2

and lim bn b implies that there exist N 2 N such that for all n N ,
n

bn b .
2

Define N maks N1 , N 2 . Then for n N ,

an bn a b

an a bn b

,
2 2

that is

lim an bn a b .

For 0, we have to find N such that for all n N ,

(iii)

an bn ab .
First we have to make an assumption that,

anbn ab anbn anb anb ab an bn b b an a .

The sequence

an

converges, then it is bounded. Then there exist M 0

an M for all n N . Define L maks M , b 0 , since

lim an a and lim bn b , then there exist N1 , N 2 N such that for all

such that
n

n N1 , a n a

and for all n N 2 , bn b

.
2L
2L

Define N maks N1 , N 2 , and for all n N ,

anbn ab an bn b b an a L

2L
2L

Therefore lim anbn ab .

Example 2.1.5: Find lim

n

n 2 5n 2 n .

Solution :

lim
n

n 5n 2 n lim
2

lim
n

n 2 5n 2 n
n 5n 2 n . 2
n 5n 2 n
2

n 2 5n 2 n 2
n 2 5n 2 n

5n 2
n 2 5n 2 n

lim
n
lim
n

5n 2 / n

n 2 5n 2 n n

2
n
lim
n
5 2
1 2 1
n n
5

50 5
.
11 2

Theorem 2.1.4 : (Squeeze Theorem for Sequences)

Let an and bn be convergent sequences and suppose that
sequence such that an xn bn for all n . If the sequence an and
converges to L , then the sequence xn converges to L .

xn
bn

is a
both

Proof : For any 0 , lim an lim bn L implies that there exists N1 , N 2

n

such that an L for n N1 , bn L for n N 2 .

Define N maks N1 , N 2 .
Then for n N ,

an L and bn L .
For each n N ,

L an xn bn L ,
which is equivalent to xn L . This shows that xn converges to L .

n
Example 2.1.6 : Consider the sequence n . Since n 2 10 n for all positive
10
integers n , we find that
n
n 1
0 n 2 ,
10
n
n

for all positive integers n . Since the sequence

1
and both converge to
n

n
converges to 0 .
n
10

0 , the squeeze theorem asserts that the sequence

2.2: MONOTONE SEQUENCES

Definition 2.2.1: A sequence an is
(i)
(ii)
(iii)
(iv)
(v)

Increasing if an1 an for all n ;

Strictly increasing if an1 an for all n ;
Decreasing if an1 an for all n ;
Strictly decreasing if an1 an for all n ;

Monotonic if an is either increasing or decreasing.

Example 2.2.1:
(i)

1
1 1 1

The sequence , , , ..., n , .... is decreasing.

2
2 4 8

(ii)

The

sequence

: n 1, 2, .... 1, 0, 1, 0,1, 0, 1, 0, ....

sin
2

is

neither increasing nor decreasing.

(iii)

The

sequence

3
n
n

: n N , , , .... ,
, ...

2
2
2
2

is

an

increasing sequence.
(iv)

The sequence is both an increasing and decreasing sequence.

Theorem 2.2.1(Monotone Convergence Theorem): Let an be a sequence.

(i)

If an an increasing sequence that is bounded from above, then an

converges, and lim an sup an : n N .
n

(ii)

If an a decreasing sequence that is bounded from below, then an

converges, and lim an inf an : n N .
n

Proof:
(i)

Since

an

is bounded from above, the completeness axiom

guarantees that there exists a number b sup an : n N . Let 0 ,

since b is not an upper bound of the set an : n N there exists a

positive number N such that a N b , and the fact that an is

increasing, then it follows that

an a N for all n N .
Thus we have

b a N an b ,
for all n N . Then we conclude that an b for all n N , that is

lim an b sup an : n N .
n

(ii)

Define bn an , n N . Then the sequence bn is an increasing

sequence because an1 an implies that bn1 bn for all n N . Since
an is bounded from below, then there exists a number R such
that an for all n N . Then bn an for all n N and bn is
bounded from above.
From part (i), we have

lim bn u sup bn : n N .
n

lim an lim bn u

Then

inf bn : n N
inf an : n N .

Example 2.2.2 : Given a1 3 ; an1 3

2
,n 1
an

(i)
(ii)

Show that an 2 for all n N .

Show that the sequence an is decreasing.

(iii)

Does lim an exists? If yes find the limit.

n

10

Solution :
(i)

We use induction to show that an 2, n N . a1 3 2 and this is true

for n 1 . Assume that it is true for n k , that is ak 2 . Then
2
ak 1 3
3 1 2 . Therefore this is true for n k 1 . According to
ak
the induction principle, this is true for all n N .

(ii)

2
an1 an 3 an
an

an 1 an 2
3an 2 an2

an
an
0 , since an 2 .
Then an is decreasing.

(iii)

Since an is bounded from below and it is decreasing, then the limit

of an must exists. Let lim an . The fact that an 2 for all n N ,
n

implies that lim an 2 . From the above formula an1 3

n

2
, n 1 , it
an

follows that

2
lim an1 lim 3 .
n
n
an

Using the fact that lim an implies lim an1 , we then have 3
n

then 1 2 0 that is 1 or 2. Since 2 , then lim an 2 .

2
,

2.3: SUBSEQUENCES

an

Definition 2.3.1 : Let

be a sequence and let

sequence of positive integers. The sequence

an such that bk an

nk

be a strictly increasing

b is called a subsequence of
n

Note : For a general case, nk k , for all values of n . We can use stepping stone
analogy to describe subsequence. A subsequence is obtained from a sequence
by removing some of the stones and leaving the rest in their original order.
11

Example 2.3.1 : Consider the sequence

an 1 , 1 , 1 , .... , 1 ,....

the sequence
n
1 2 3
1
1 1

1 1 1 1 1 1 1

and
are both
,.......
1, , ,....,
1, , , , , , , , .......
3
5
2
n

1
3
5
18
23
24
25
26

1 1 1 1 1

subsequences of an . The sequence ,1, , , , , ..... is not a

3 5 7 9 11

subsequence of an because it violates the original order in an .

Subsequences can provide a simple way to show that a sequence does not
converge.
Theorem 2.3.1 : Let an be a sequence of real numbers. If
, then every subsequence of an converges to .

Proof: Suppose that

an . Given

a converges to
n

an

converges to

and let an k be any subsequence of

0 , there exists a positive integer N such that an for all

n N . Since nN N and when k N , we have nk nN . Then for k N , it

follows that nk nN N .
Thus, we have
ank , for k N ,

That is lim ank .

n

Corollary 2.3.1 : If a subsequence of

not converge.

an

does not converge, then

an

does

Corollary 2.3.2 : If an has two subsequences that converge to different limits,

then an does not converge.
These two corollaries are consequences of the theorem 2.3.1.

Example 2.3.2:
(i)

1
1
1
1

Let an 1, , 2 , , 3 , , 4 , ,....,... the subsequence

2
3
4
1

does not converge then according to corollary 2.3.1

converge.

1, 2 , 3 , 4 ,.......
an

does not

12

(ii)

an 1n 1,1, 1,1,..... has subsequences

1 , 1 , 1 , 1 ,... and 1,1,1,1,....... which converge to 1 and 1
respectively. Then according to corollary 2.3.2 an does not
The sequence

converge.
(iii)

The sequence 1, 2 , 3 , 4 ,....... is not bounded and it does not have any
subsequence that converges.

NOTE : Looking at (iii), we know that not all sequence contains subsequence that
converges. The theorem below provides a very elementary sufficient condition for
the existence of a convergent subsequence. This theorem also plays an
important role in our development of continuous functions. Before we consider
this Theorem (Bolzano Weierstrass) we state and prove the nested Interval
Theorem.
Theorem 2.3.2 (Nested Interval Theorem) : Suppose n an , bn , is a
sequence of closed intervals such that, for all n , n1 n . (such a sequence is

called a nested sequence of intervals). Then

n .

n 1

Proof: Since 1 2 3 .... n ...., then for each m , n N , we have

am anm bnm bn .
Thus A am : m N is bounded from above by bn , n N and the supremum of
A exists, say sup A x . Then an x bn for all n N , and this implies that x n

for all n N . Therefore x n and

n 1

n 1

Theorem 2.3.3: (Bolzano Weierstrass for sequences)

If an is a bounded sequence, then an has a convergent subsequence.
Proof: Let A be a set that contains all terms of an . If A is finite, then at least
one of the terms of an repeats itself infinitely many times. Then we can choose
a constant subsequence where the value of each term takes the value of the
repeated term of an . Then this subsequence will converge to that constant.

13

Assume A is infinite. Because an is bounded, there is a number M 0 such

that an M for all n . Let M , M a , b with A a , b . Now divide

a ,b

ab
ab
into two subintervals a ,
and
, b . A is infinite and A a , b

implies that at least one of the two subintervals contain infinitely many terms of
A . Choose one and denote it by I1 . Then we have I1 a , b , length of

ba
and A1 I1 A infinite. Divide I1 into two subintervals of equal
2
ba
length. Then since A1 is infinite, it follows that I 2 I1 , length of I 2
and
22
A2 I 2 A1 infinite. Continuing in this manner, we obtain the following :I1

(i)

A nested sequence of closed intervals n that satisfies.

I1 I 2 I 3 .... I n ....,

ba
,n N .
2n

(ii)

The length of I n

(iii)

Each set An n An1 , n N is infinite

By Theorem 2.3.2, property (i) implies that

I
n 1

will show next that there is a subsequence

A0 A

. Assume that x I n . We
n 1

a of a that converges to
nk

x.

ba
1 . Then m1 x 1, x 1 and this is
2 m1
implies that

There exist m1 N such that

because, for any y m1 , x m1

yx

ba
1 .
2 m1

Since m1 Am1 1 is infinite, it follows that

x 1, x 1 Am 1 .

such that an1 x 1, x 1 Am1 1 . There exists m2 N

Choose

an

with m2 m1 and

ba 1
. Then there exists m2 x 1 , x 1 . Since m2 Am2 1 is infinite,
2
2
2
2 m2
then there exists

an x 1 , x 1 Am 1 , n2 n1 .
2
2
2

14

Continue by induction, we then have an1 , an2 , ...., ank ,.... with n1 n2 n3 ... nk ....

and ank x 1 , x 1
of

an

mk 1

, k 1, 2 , 3 , .... Then an is a subsequence

ank x 1 , k 1, 2 , 3 , .....
k

and

. For any

Archimedean principle, there exists K N such that 1

0 , according to

. Then for k K ,

an x 1 1 .
k
K
k

Therefore

2.4

a x .

nk

CAUCHY SEQUENCE

Definition 2.4.1 : A sequence an is a Cauchy sequence if for each 0 there

exists a positive integer N such that am an for all m , n N .
Note: For a convergent sequence, the terms are eventually close to the limit
whereas for a Cauchy sequence, the terms are eventually close to each other.

Example 2.4.1 : Given an

Solution :

am an

N such that

1
, n N , show that an is a Cauchy sequence.
n

1 1
1 1
maks , . For any 0 , there exist
m n
m n

1
. Then for all m , n N ,
N

am an .
Therefore an is a Cauchy sequence.

Theorem 2.4.1: Every Cauchy sequence is bounded.

Proof: Let an be a Cauchy sequence. For 1 0 , there exists n N such
that for m , n N ,

am an 1 .

15

Particularly an a N 1 for all n N , that is an a N 1 , for n N .

Now define M maks

, a2 , .... , a N 1 , a N 1. It follows that an M , for

all n N . Therefore an is bounded.

Theorem 2.4.2: A sequence

sequence.

an

converges if and only if it is a Cauchy

Proof: Let an converges to a limit a . For any 0 there exist N such

that for n N ,

an a .
2
Then for m , n N ,

an am an a a am
an a a am

2 2

Then an is a Cauchy sequence.

Conversely, let an be a Cauchy sequence. According to theorem 2.4.1, an is
bounded. Then by Bolzano Weiertrass Theorem, an has a convergent
subsequence , an k . Let a be the limit of an k . We now show that an a .

2
is a Cauchy sequence, there exist N 2 N such that

For any 0 , lim ank a implies that there exist N 1 such that ank a
n

for all k N1 . Since an

an am for all n, m N 2 .
2

Now define N maks N1 , N 2 . Since nN N , then for n N ,

a n a a n a nN a nN a a n a nN a nN a

.
2 2

16

2.5

LIMIT SUPERIOR & LIMIT INFERIOR

Definition 2.5.1 : Let an be a sequence in . Then

(i)

lim an , if for each M 0 , there exist a positive integer N such

n

that for all n N , an M . (The sequence diverges to ).

(ii)

lim an , if for each L 0 , there exist a positive integer N such

n

that for all n N , an L . (The sequence diverges to ).

The limits in this definition are to be considered in the extended sense, that is, it
is possible for a limit of a sequence to be or .
Examples 2.5.1:
(i)

The sequence

a , a 1 ,
n

is increasing and unbounded. Thus

lim a .
n

(ii)

The sequence ln

1
1
ln
.
decreasing and unbounded, lim
n
n
n

Examples 2.5.2:
(i)
The sequence 2 n diverges to .
(ii)
The sequence n 2 diverges to .
(iii)
The sequence 1 nn does not diverge to neither nor .

Definition 2.5.2 : Let an be a sequence in the extended real number set.

(i)

The limit superior of an is defined by lim an inf sup ak , and

(ii)

The limit inferior of an is defined by lim an sup inf ak .

k n

k n

Theorem 2.5.1 : Let an be a sequence in the extended real number set. Then

lim an lim an .
Proof :
Define yn sup ak and z n inf ak for n N . Since
k n

then we have

k n

yn

decreasing and

z
n

ak : k n 1 ak : k n ,

increasing. It follows that

17

sup ak : k n 1 sup ak : k n and inf ak : k n 1 inf ak : k n . For each

n N , yn ak and z n ak for k n .
For any n , m N ,

yn sup ak inf ak zm .
k m
k n

That is yn sup z m .
m

Therefore inf yn sup z m that is

n

lim an lim an .

Theorem 2.5.2 : Let an be a sequence in the extended real number set. Then

lim an lim an if and only if lim an .

n

Proof : Let lim an lim an and let 0 . Then inf yn and

n

yn

is

decreasing implies that there exist N1 N such that yn for all n N1 .

Similarly sup z n and

z increasing implies that there exists

n

N 2 N such

that zn for all n N 2.

Define N maks N1 , N 2 . For all n N , we have z n yn since

zn an yn for n N , then for all n N

an .
Therefore lim an .
n

Conversely, let lim an . For any 0 , there exists N such that for n N ,
n

an

that is

an for n N .
2
2

Then for all n N ,

18

yn sup ak
k n

zn inf
ak
k n

.
2

Then z n yn for n N . It follows that lim yn lim z n . From

n

theorem

2.2.1

(Monotone

Convergence

Theorem),

lim yn inf yn
n

and

lim zn sup zn , that is lim an lim an .

n
n

Case for : Let lim an lim an . Since sup z n lim an , then for each
n

M 0 , there exist N such that z N M . Since

z is increasing, it follows
n

that z n M for all n N . The fact that an z n , implies an M for n N .

Therefore lim an .
n

Conversely, let lim an , for any M 0 , then there exist N such that
n

an M 1 for n N . Then for n N , z n inf ak M 1 M .

k n

Therefore lim a n sup z n . From the fact that lim an lim an , it follows that
n

lim an lim an . For case , it follows a similar fashion as in the case for

.
2.6

SERIES OF REAL NUMBERS

Definition

2.6.1

an

Let

be

sequence

of

real

numbers

and

sn a1 a2 a3 ..... an ak , n N . Then the sequence sn is called series.

k 1

number an is the n th

sn , series is sometimes written as an . The

n 1
term of the series and sn is said to be the n th partial

sum. If the sequence

Notes: Besides from the symbol

n 1

converge to s , then the series

a
n 1

is said to

converge to s , that is

n 1

k 1

an lim
ak s .
n
19

If

n 1

does not converge, then

a
n 1

does not converge as well.

Example 2.6.1 : Show that the series

k (k 1)

converges and find the sum.

k 1

Solution :

1
1
1

,Then
k (k 1) k k 1
n
1 1 1 1 1
1 1
1
1
1


....

k 1 1 2 2 3
k 1 k
n 1 n n n 1
1
1
.
n 1
So the lim s n 1 .
Since

This series is called a Telescoping series.

Example 2.6.2 : Show that the series

diverges. This series is a

k 1

Harmonic series.
Solution :

We prove this using contradiction. Now suppose that

converges.

k 1

1
. By assumption, the limit exists, say lim s n . Thus for 1,
k 1 k
There exist integer N 0 such that for each n N , 1 s n 1 .
1 1
1
s 2 1 ... m
2 3
2
n

Let s n

1 1 1 1 1 1 1 1 1
1
1
1
1
... ... m 1
m 1
... m
2 3 4 5 6 7 8 9 10
16
2
2 1 2 2

1 1 1 1 1 1 1 1 1
1
1
1
1
... ... m m ... m
2 4 4 8 8 8 8 16 16
16
2
2
2

20

1 1 1 1
1
...
2 2 2 2
2
m
1 .
2
1

Choose m big enough so that 2 N and 1

m

s2 1 .

m
1, then
2

This gives a contradiction and therefore

diverges.

k 1

Example 2.6.3: Show that

k
k 1

converges.

Solution:

1 1
1
1
= 1 2 2 ... 2
2
2 3
n
k 1 k
n

sn

1
1
1

...
1 2 2 3
(n 1)n
n 1
1
1
k 1 k ( k 1)

1
1 1
n
1
2
n
2.
So s n is bounded above. Since s n is increasing then lim s n exists, that is
n

k
k 1

converges.

21

Theorem 2.6.1 : If the series

a
n 1

converge, then lim an 0 .

Proof:

s n s n 1 a n ,

lim s n lim s n 1 lim an .

n

This Theorem gives rise to the following corollary as a test for convergence of
series.
Corollary 2.6.1 : If lim an does not exist or lim an 0 then
n

a
n 1

does not

converge.

Lemma 2.6.1:

a
k 1

that

k m 1

converges, if and only if for any 0 there is an N such

for all n m N .
n

Proof: For each positive integer n , let s n a k . If m and n are positive

k 1

integers with n m, then we have

n

k m 1

k 1

k 1

ak ak sn sm .

The result follows from the fact that s n converges if and only if it is a Cauchy
sequence.

Definition 2.6.2 : We say that the series

diverges, then we say that

n 1

converges. If

a
n 1

converges but

a
n 1

an converges absolutely if
n

n 1

n 1

converges conditionally.

22

Theorem 2.6.2 : If the series

a
n 1

converges, then so does

a
n 1

Proof :
This Theorem is a consequences of Lemma 2.6.1 and the inequality
n

k m 1

| a

k m 1

|.

Note : If a given series converge absolutely, rearranging the terms in the series
gives us a new series that converge absolutely.
Note : It is not necessarily true that when lim an 0 , then
n

a
n 1

converges. For

1
1
1
1
and
have lim 0 and lim 2 0 respectively, but

2
n n
n n
n 1 n
n 1 n

1
diverges and 2 converges.
n 1 n

example,

Theorem 2.6.3 : Let

an and
n 1

(i)

(ii)
(iii)

n
n 1

b
n 1

be two convergent series. Then

n 1

n 1

n 1

n 1

an bn also converges and an bn an bn

For any ,

a
n 1

converges and

a
n 1

an .
n 1

A finite number of terms can be added or removed from any series

without affecting convergence ( or divergence).

23

Proof :
(i)

k 1

k 1

(ak bk ) lim
(ak bk )
n
n a n b
lim
k k
n
k 1
k 1

lim
ak lim
bk
n
n
k 1

k 1

k 1

k 1

ak bk .
(ii)

k 1

lim a k
n

k 1

lim
ak
n
k 1
n

lim
ak
n
k 1

ak .
k 1

(iii) Assume now that m terms a n 1 , a n 2 ,...., a n m have been removed from the

series

a . Then we have
k 1

Tn s n m (a n 1 a n 2 ... a n m ) ,

with Tn is the n-th partial sum of the series formed from

a
k 1

, with the m terms

removed. Since s n m converge (because s n converges) and has a limit, then Tn

will have a limit. Thus the new series converges if the old one converges.

When adding a finite number of real numbers, the order in which the terms are
added together makes no difference in the sum. This is a consequences of
commutative and associative properties of addition. However, this property of
addition does not extend to infinite sums of real numbers.
24

Consider the two series

1 1 + 1 1 + 1 1.

and

1 + 1 -1 + 1 + 1 - 1 + 1 + 1,

which consist of the same numbers, but in different orders. The first series has
bounded partial sums and the second series has unbounded partial sums.
It is clear that the order in which the terms appear in an infinite series may affect
the sum.
Definition 2.6.3 : A rearrangement of a series

(i)

a
n 1

a1 a2 a3 .... an .... rearranges the order of terms in the

series

by

a
n 1

f (n)

rearrangement

mapping

to

a f (1) a f ( 2 ) a f ( 3) .... a f ( n ) ....

(ii)

Let

b
n 1

be a series such that

bn a f (n ) , for n 1, 2 , 3 , ....

Then

b
n 1

is said to be a rearrangement of

The function f is a permutation function from Z Z which is one to one

and onto. In other words permutation is a sequence of positive integers such that
each positive integer appears exactly once. The following example gives us an
idea of a permutation of positive integers consisting of one odd integer followed
by two even integers, with the order of the odd and even integers preserved.

1, 2, 4, 3, 6, 8, 5, 10, 12, 7, 14, 16, 9, 18, 20,

Theorem 2.6.4 : Let

a
n 1

be an absolute convergent series having sum s . Then

every rearrangement of

a
n 1

also converges absolutely and has sum s .

Proof : Let bn be defined as in Definition 2.6.3 (ii). Then

b1 .... bn a f (1) .... a f ( n ) ak ,

n 1

25

so

n 1

bn

has bounded partial sums. Hence

n 1

show that

b
n 1

converges absolutely. To

s , let t n b1 .... bn , sn a1 .... an . Given 0 ,

choose N so that s N s and so that

2

k 1

aN k

Then
tn s tn sN sN s tn sN

Choose M so that

.
2

.
2

1, 2 , ... , N f (1), f (2), ... , f ( M ) . Then

n M implies

f (n) N , and for such n we have

tn s N b1 ... bn a1 ... aN
a f (1) ... a f ( n ) a1 ... a N

a N 1 a N 2 ...

,
2

since all the terms a1 , ..., aN cancel out in the subtraction. Hence, n M implies

tn s and this means

s.

26