Real Sequences

By

Dr. Shubhashish Das

Assistant Professor
Department of Mathematics
Bharat Sevak Samaj College, Supaul
Supaul, Bihar-852131, India
Email ID.: [email protected]

s
In this class we shall study a special class of functions whose domain is the set

Da
N and the range is the set R− the real sequences.

Definition 1 (Sequences). A mapping f : N → R is said to be a sequence in

ish
R or a real sequence.

The f -image f (1), f (2), f (3), . . . , . . . are real numbers. The image of the n-th
sh
element, f (n), is said to be the n-th element or the n-th term of the real sequence.
We shall be mainly concerned here with real sequences and we shall use the
ha
term ‘sequence’ to mean a ‘real sequence’.
The symbols like (an ), (bn ), (un ), (xn ) etc. shall be used to denote a sequence.
b

Example.
hu

1. Let f : N → R be defined by f (n) = n, n ∈ N. Then the sequence is

denoted by (xn ), where xn = f (n) = n. It is also denoted by (1, 2, 3, 4, . . .).
.S

2. Let f : N → R be defined by f (n) = n2 , n ∈ N. Then the sequence is de-

noted by (xn ), where xn = f (n) = n2 . It is also denoted by (1, 4, 9, 16, . . .).
n
3. Let f : N → R be defined by f (n) = , n ∈ N. Then the sequence is
Dr

n+1
n
denoted by (xn ), where xn = f (n) = n+1
. It is also denoted by ( 21 , 23 , 34 , . . .).
4. Let f : N → R be defined by f (n) = (−1)n , n ∈ N. Then the sequence
is denoted by (xn ), where xn = f (n) = (−1)n . It is also denoted by
(−1, 1, −1, 1, . . .).
5. Let f : N → R be defined by f (n) = sin nπ
2
, n ∈ N. Then the sequence
is denoted by (xn ), where xn = f (n) = sin nπ
2
. It is also denoted by
(1, 0, −1, 0, 1, 0, −1, . . .).
6. Let f : N → R be defined by f (n) = 2, n ∈ N. Then the sequence is
denoted by (xn ), where xn = f (n) = 2. It is also denoted by (2, 2, 2, 2, . . .).
1
 1
7. Let f : N → R be defined by f (n) = n
, n ∈ N. Then the sequence is
denoted by (xn ), where xn = f (n) = n1 . It is also denoted by (1, 21 , 31 , . . .).

n
8. Let f : N → R be defined by f (n) = 1 + n1 , n ∈ N. Then the sequence

n
is denoted by (xn ), where xn = f (n) = 1 + n1 . It is also denoted by
(2, 94 , 64
27
, . . .).
(−1)n
9. Let f : N → R be defined by f (n) = n
, n ∈ N. Then the sequence
n
is denoted by (xn ), where xn = f (n) = (−1) n
. It is also denoted by
(−1, 12 , − 13 , . . .).

s
Definition 2 (Bounds of a sequence). A real sequence (xn ) is said to be

Da
bounded above if there exists a real number M such that xn ≤ M for all n ∈ N.
The real number M is said to be an upper bound of the sequence.
A real sequence (xn ) is said to be bounded below if there exists a real number

ish
m such that m ≤ xn for all n ∈ N. The real number m is said to be a lower
bound of the sequence.
sh
A real sequence (xn ) is said to be bounded if there exist real numbers m, M
such that m ≤ xn ≤ M for all n ∈ N.
ha
The least upper bound of a real sequence (xn ) is a real number M satisfying
the following conditions:
b

(i) xn ≤ M for all n ∈ N,

hu

(ii) for each pre-assigned positive ε, there exists a natural number k such that
M − ε < xk ≤ M.
.S

The greatest lower bound of a real sequence (xn ) is a real number m satisfying
the following conditions:
(i) m ≤ xn for all n ∈ N,
Dr

(ii) for each pre-assigned positive ε, there exists a natural number k such that
m ≤ xk < m + ε.
Example.
1


10. The sequence n
is a bounded sequence. 0 is the greatest lower bound
and 1 is the least upper bound of the sequence.
11. The sequence ((−1)n ) is bounded sequence. −1 is the greatest lower bound
and 1 is the least upper bound of the sequence.
12. The sequence (n) is bounded below and unbounded above. Here 1 is the
greatest lower bound and sup(xn ) = ∞.
2
 13. The sequence (−n) is unbounded below and bounded above. Here −1 is
the least upper bound and inf(xn ) = −∞.

Definition 3 (Convergence of sequences). Let (xn ) be a sequence. A real

number ` is said to be a limit of the sequence (xn ) if corresponding to a pre-
assigned positive ε, there exists a natural number k (depending on ε) such that

|xn − `| < ε for all n ≥ k

i.e., ` − ε < xn ≤ ` + ε for all n ≥ k.

s
Da
To be explicit, a real number ` is said to be a limit of the sequence (xn )
if corresponding to a pre-assigned positive ε, there exists a natural number k
(depending on ε) such that all elements of the sequence, excepting the first k − 1
at most, lie in the ε-neighbourhood of `.

ish
A real sequence (xn ) is said to be a convergent sequence if it has a limit ` ∈ R.
In this case the sequence is said to be converge to `.
We write it as lim xn = ` or lim xn = `.
n→∞
sh
A sequence which is not a convergent sequence, is call divergent.
ha
Theorem 1. Every convergent sequence is bounded.
b

Proof. Let (xn ) be a convergent sequence converges to `.

Then for chosen any ε > 0, there exists a natural number k such that
hu

` − ε < xn < ` + ε for all n ≥ k.

.S

Let M = max{x1 , x2 , . . . , xk−1 , ` + ε} and m = min{x1 , x2 , . . . , xk−1 , ` − ε}.

Then m ≤ xn ≤ M for all n ∈ N.
This proves that the sequence (xn ) is a bounded sequence. 
Dr

Remark 1. An unbounded sequence is not convergent.

Remark 2. A bounded sequence need not be convergent. For example, the se-
quence (1, −1, 1, −1, . . .) is a bounded sequence but the sequence does not converge
to a limit.

Theorem 2. A sequence cannot converge to more than one limit

Proof. If possible, let a sequence (xn ) have two different limits `1 and `2 , where
`1 < `2 .
3
 Let ε = 13 (`2 − `1 ).
Since `1 is a limit of the sequence (xn ), for the same chosen ε, there exists a
natural number k1 such that

|xn − `1 | < ε for all n ≥ k1 . (1)

Again, since `2 is a limit of the sequence (xn ), for the same chosen ε, there
exists a natural number k2 such that

s
Da
|xn − `2 | < ε for all n ≥ k2 (2)

Let k = max{k1 , k2 }. Then from (1) and (2), we obtain

ish
|xn − `1 | < ε & |xn − `2 | < ε for all n ≥ k. (3)
sh
Now |`2 − `1 | ≤ |xn − `1 | + |xn − `2 | < ε + ε = 2ε = 32 |`2 − `1 | which is a
contradiction.
ha
Hence `1 = `2 and this proves the theorem. 
b

Example.
hu

1


14. The sequence n
converges to 0.
.S

Proof. Let us choose a positive ε, whatever small it may be.

By Archimedean property of R, there exists a natural number k such
that
Dr

1
<k
ε
1
⇒ 0< <ε
k
1 1
⇒ 0 < ≤ < ε for all n ≥ k.
n k

1
This shows that 0 − n
< ε for all n ≥ k. This proves lim n1 = 0. 

 
n2
15. The sequence n2 +1
converges to 1.
4
 Proof. Let us choose a positive ε. Now
n2
−1 <ε
n2 + 1
1
⇒ <ε
n2 + 1
1
⇒ 2 <ε
n +1
1
⇒ < n2 + 1
ε
r
1

s
⇒ n> − 1.
ε

Da
hq i
1
Let k = ε
− 1 + 1.
Therefore, for the given ε > 0, there exists a natural number k =
hq i 2
1
− 1 + 1, such that n2n+1 − 1 < ε for all n ≥ k.

ish
ε
2
This proves lim n2n+1 = 1. 
 √ 
16. The sequence 3+2 √
n
n
converges to 2. sh
Proof. Let us choose a positive ε. Now
√
ha
3+2 n
√ −2 <ε
n
3
⇒ √ <ε
b

n
3
hu

⇒ √ <ε
n
3 √
⇒ < n
.S

ε
9
⇒ n > 2.
ε
9
Dr

Let k = ε2
+ 1.
9
Therefore, for the given ε > 0, there exists a natural number k = ε2
+1,
√
3+2 n
such that √
n
− 2 < ε for all n ≥ k.
√
This proves lim 3+2
√
n
n
= 2. 

17. The sequence (xn ) where xn = c for all n, converges to c.

Proof. Let us choose a positive ε.

Now 0 = |c − c| = |xn − c| < ε holds for all n ∈ N, i.e., |xn − c| < ε holds
for all n ≥ 1.
This proves lim xn = c. 
5
 3+2n


18. The sequence 1+n
converges to 2.

Proof. Let us choose a positive ε. Now

3 + 2n
−2 <ε
1+n
1
⇒ <ε
1+n
1
⇒ <ε
1+n
1
⇒ <n+1

s
ε
1

Da
⇒ n > − 1.
ε
1 
Let k = ε
− 1 + 1.
Therefore, for the given ε > 0, there exists a natural number k =

ish
1
− 1 + 1, such that 3+2n

ε 1+n
− 2 < ε for all n ≥ k.
This proves lim 3+2n
1+n
= 2. 
 2
 sh
19. The sequence 3+2n
1+n2
converges to 2.

Proof left to the reader as an exercise. 

ha
Theorem 3. Every convergent sequence is bounded and has a unique limit.
b

Proof left to the reader as an exercise. 

hu

Theorem 4 (Algebra of sequences). If (an ), (bn ) be two sequences such that

lim an = a, lim bn = b, then
n→∞
.S

(i) lim (an ± bn ) = lim an ± lim bn = a ± b,

n→∞ n→∞ n→∞
(ii) if c ∈ R, lim (can ) = ca,
n→∞   
(iii) lim (an bn ) = lim an lim bn = a b,
Dr

n→∞ n→∞  n→∞  

(iv) lim (an /bn ) = lim an / lim bn = a/b, provided b 6= 0, bn 6= 0 ∀n.
n→∞ n→∞ n→∞

Proof of (i). To show that lim (an ± bn ) = a ± b, we need to establish that for a
n→∞
preassigned positive ε, there exists k ∈ N such that

|(an ± bn ) − (a ± b)| < ε for all n ≥ k.

Using triangle inequality, we have

|(an ± bn ) − (a ± b)| = |(an − a) ± (bn − b)| ≤ |(an − a)| + |(bn − b)|.

Let ε be a preassigned positive real number.

6
 Since the sequence (an ) converges to a, then for the given ε there exists k1 ∈ N
such that

ε
|an − a| < for all n ≥ k1 . (1)
2

Again since the sequence (bn ) converges to b, then for the given ε there exists
k2 ∈ N such that

ε
|bn − b| < for all n ≥ k2 . (2)
2

s
ε ε
Let k = max{k1 , k2 }. Then |an − a| < 2
and |bn − b| < 2
for all n ≥ k. Now

Da
from (1) and (2), we get

|(an ± bn ) − (a ± b)| =|(an − a) ± (bn − b)|

ish
≤|(an − a)| + |(bn − b)|
ε ε
< + = ε for all n ≥ k.
2 2

This proves lim (an ± bn ) = a ± b.

n→∞
sh 

Proof of (ii). If c = 0, the theorem is obvious. So we assume c 6= 0.

ha
To show that lim (can ) = ca, we need to establish that for a preassigned
n→∞
positive ε, there exists k ∈ N such that
b
hu

|(can ) − ca| < ε for all n ≥ k.

Now we have
.S

|(can ) − ca| = |c||an − a|.

Let ε be a preassigned positive real number.

Dr

Since the sequence (an ) converges to a, then for the given ε there exists k ∈ N
such that

ε
|an − a| < for all n ≥ k. (1)
|c|

Now, form (1) we obtain for the given ε, there exists the same natural number
k such that

ε
|can − ca| = |c||an − a| < |c| = ε for all n ≥ k.
|c|

This proves lim (can ) = ca. 

n→∞
7
Proof of (iii). To show that lim (an bn ) = a b, we need to establish that for a
n→∞
preassigned positive ε, there exists k ∈ N such that

|an bn − a b| < ε for all n ≥ k.

Using triangle inequality, we have

|an bn − a b| =|an bn − an b + an b − a b|

≤|an bn − an b| + |an b − a b|

s
≤|an | |bn − b| + |b| |an − a|.

Da
Since the sequence (an ) is convergent, it is bounded. Therefore, there exists a
positive number M1 such that an ≤ M1 .
Let M = max{M1 , b}.

ish
Let ε be a preassigned positive real number.
Since the sequence (an ) converges to a, then for the given ε there exists k1 ∈ N
such that
sh
ε
|an − a| < for all n ≥ k1 . (1)
ha
M
Again since the sequence (bn ) converges to b, then for the given ε there exists
k2 ∈ N such that
b

ε
hu

|bn − b| < for all n ≥ k2 . (2)

M
ε ε
Let k = max{k1 , k2 }. Then |an − a| < M
and |bn − b| < M
for all n ≥ k. Now
.S

from (1) and (2), we get

|an bn − a b| =|an bn − an b + an b − a b|
Dr

≤|an bn − an b| + |an b − a b|

≤|an | |bn − b| + |b| |an − a|

ε ε
<M +M = ε for all n ≥ k.
M M
This proves lim (an bn ) = a b. 
n→∞

Proof of (iv). 

Theorem 5. If (an ) be a convergent sequence, converges to a real number `.

Then the (|an |) be a convergent sequence, converges to a real number |`|.
8
Proof. We know

||an | − |`|| ≤ |an − `|. (1)

Let ε be a preassigned positive real number.

Since the sequence (an ) converges to `, then for the given ε there exists k ∈ N
such that

|an − `| < ε for all n ≥ k. (2)

Combining (1) and (2), it follows that

s
Da
||an | − |`|| < ε for all n ≥ k.

This proves lim |an | = |`|. 

n→∞

Remark 3. If (|an |) is a convergent sequence, it does not necessarily imply that

ish
(an ) is a convergent sequence. For example, let an = (−1)n . Then the sequence
(|an |) = (1) is a constant sequence and converges to 1 but (an ) is not a convergent
sequence.
sh
Theorem 6. If an ≥ 0 for all n and lim an = a, then a ≥ 0.
ha
n→∞

Proof. If possible let a < 0.

Let us choose an ε > 0 such that
b

a + ε < 0. (1)
hu

Since lim an = a, then for the given ε there exists k ∈ N such that
n→∞
.S

a − ε < an < a + ε for all n ≥ k. (2)

Combining (1) and (2), we get an < a + ε < 0 for all n ≥ k, which contradicts
Dr

the hypothesis.
Hence lim an = a ≥ 0. 
n→∞

Theorem 7. If (an ), (bn ) be two sequences such that an ≤ bn for all n and
lim an = a, lim bn = b, then a ≤ b.

Proof. Let un = bn − an for all n. Then by the previous theorem (un ) is a

convergent sequence converges to b − a.
Also, we have un = bn − an ≥ 0.
Hence by the previous theorem we get lim un = b − a ≥ 0 and therefore,
n→∞
lim an = a ≤ b = lim bn . 
9
Theorem 8 (Sandwich theorem). If (an ), (bn ) and (cn ) be three sequences
such that

(i) an ≤ bn ≤ cn for all n ≥ m, where m ∈ N

(ii) lim an = lim cn = `, then lim bn = `.
n→∞ n→∞

Proof. Let ε > 0 be a preassigned real number.

Since lim an = `, then for the given ε there exists k1 ∈ N such that
n→∞

` − ε < an < ` + ε for all n ≥ k1 . (1)

s
Da
Again since lim bn = `, then for the given ε there exists k2 ∈ N such that
n→∞

` − ε < an < ` + ε for all n ≥ k2 . (2)

ish
By the given hypothesis, we have

an ≤ bn ≤ cn for all n ≥ m, where m ∈ N. (3)

Let k = max{k1 , k2 , m}.

sh
Then combining (1), (2) and (3), we get for the given ε > 0,
ha
` − ε < an ≤ bn ≤ cn < ` + ε for all n ≥ k
b

⇒ ` − ε < bn < ` + ε for all n ≥ k.

hu

This proves lim bn = `. 

n→∞
.S

Example.
√ √
20. Prove that lim ( n + 1 − n) = 0.
n→∞
Dr

Proof. We have
√ √ √ √
√ √ ( n + 1 − n)( n + 1 + n)
0≤ n+1− n= √ √
n+1+ n
1 1
=√ √ ≤ √ .
n+1+ n 2 n
√ √ 1
Let an = 0, bn = n+1 − n, cn = √
2 n
. Then both (an ) and (cn )
converges to 0.
√ √
By the Sandwich theorem lim ( n + 1 − n) = 0. 
n→∞
 
21. Prove that lim √ 1 + √ 1 + ··· + √ 1 = 1.
n→∞ n2 +1 n2 +2 n2 +n
10
 Proof. Let bn = √ 1 + √ 1 + ··· + √ 1 .
n2 +1 n2 +2 n2 +n
We have

1 1
√ =√
n2 + 1 n2 + 1
1 1
√ <√
n2 + 2 n2 + 1
1 1
√ <√
2
n +3 2
n +1
··················

s
1 1

Da
√ <√ .
n2 +n 2
n +1

Therefore, bn < √ n for all n ≥ 2.

n2 +1
Again we have

ish
1 1
√ >√
n2
+1 +n n2
1 1
√
n2 + 2
>√
sh
n2 + n
1 1
√ >√
ha
n2 + 3 n2 + n
··················
b

1 1
√ =√ .
n2 + n n2 + n
hu

Therefore, bn > √ n for all n ≥ 2.

n2 +n
Thus, √ n < bn < √ n for all n ≥ 2. Also we have lim √ n =1=
.S

n2 +n n2 +1 n→∞ n2 +1
lim √ n .
n→∞ n2 +n
 
By the Sandwich theorem lim √ 1 + √ 1 + ··· + √ 1 = 1. 
n→∞ n2 +1 n2 +2 n2 +n
Dr

Definition 4 (Null sequence). A sequence (un ) is said to be a null sequence

if lim un = 0.

un+1
Theorem 9. Let (un ) be a sequence such that lim un
= `. If 0 ≤ ` < 1, then
n→∞
lim un = 0.

Example.
1


22. The sequence n
is a null sequence.
n!
23. The sequence (xn ) is a null sequence, where xn = nn
.
11
 n!
Proof. Here xn = nn
. Now
(n+1)!
xn+1 (n+1)n+1
lim = lim n!
n→∞ xn n→∞
nn
n
n × (n + 1)!
= lim
n→∞ n! × (n + 1)(n+1)

nn 1
= lim n
= < 1.
n→∞ (n + 1) e
n!
Hence lim n = 0. 
n→∞ n

s
1 1

24. The sequence (xn ) is a null sequence, where xn = (n+1)2
+ (n+2)2
+ ···+

Da
1

(n+n)2
.
43n
25. The sequence (xn ) is a null sequence, where xn = 34n
.
bn
26. The sequence (xn ) is a null sequence, where xn = n!
, b > 0.

ish
Some important limits.

27. If |r| < 1, then lim rn = 0.

Proof.
sh 

28. If lim xn = 0 and a > 0, then lim axn = 1.

ha
Proof. 
b

29. Behavior of the sequence (rn ) for different real values of r.

hu

Proof. 

Theorem 10. A sequence (un ) is defined by un+2 = α un+1 + (1 − α)un for n ≥ 1

.S

1
and u1 > 0, u2 > 0. Then the sequence (un ) converges to 2−α
{u2 + (1 − α)u1 }.

Example.
Dr

un+1 +un
30. Prove that the sequence (un ) is defined by un+2 = 2
for n ≥ 1 and
u1 +2u2
u1 > 0, u2 > 0, converges to 3
.

1
Proof. Here α = 2
. Then by the previous theorem, (un ) converges to
1 1 2u2 +u1
2−α
{u2 + (1 − α)u1 } = 2− 12
{u2 + (1 − 12 )u1 } = 3
. 
2un+1 +un
31. Prove that the sequence (un ) is defined by un+2 = 3
for n ≥ 1 and
u1 +3u2
u1 > 0, u2 > 0, converges to 4
.
un+1 +2un
32. Prove that the sequence (un ) is defined by un+2 = 3
for n ≥ 1 and
2u1 +3u2
u1 > 0, u2 > 0, converges to 5
.
12
 2un+1 +un
33. Prove that the sequence (un ) is defined by un+2 = 3
for n ≥ 1 and
u1 +3u1
u1 > 0, u2 > 0, converges to 4
.
√
34. Prove that the sequence (un ) is defined by un+2 = un+1 un for n ≥ 1 and
p
u1 > 0, u2 > 0, converges to 3 u1 u22 .
2 1 1
35. Prove that the sequence (un ) is defined by un+2
= un+1
+ un
for n ≥ 1 and
3
u1 > 0, u2 > 0, converges to 1
+ u2
.
u1 2

Definition 5 (Cauchy sequence). A sequence (un ) is said to be a Cauchy

sequence if for a preassigned positive ε there exists a natural number k such that

s
Da
|un+p − un | < ε for all n ≥ k and for p = 1, 2, 3, . . . .

Theorem 11 (Cauchy’s general principle of convergence). A necessary

ish
and sufficient condition for the convergence of a sequence (un ) is that for a pre-
assigned positive ε there exists a natural number k such that

sh
|un+p − un | < ε for all n ≥ k and for p = 1, 2, 3, . . . .

Proof. Let (un ) be a convergent sequence and lim un = `. Then for a preassigned
ha
ε there exists a natural number k such that

ε
b

|un − `| < for all n ≥ k.

2
hu

ε
Therefore, |un+p − `| < 2
for all n ≥ k and for p = 1, 2, 3, . . . .
Now,
.S

|un+p − un | = |un+p − ` + ` − un |

≤ |un+p − `| + |` − un |
Dr

ε ε
< + = ε for all n ≥ k and for p = 1, 2, 3, . . . .
2 2

This proves that the condition is necessary.

Conversely, We now prove that the sequence (un ) is convergent as a consequence
of the given condition.
First we prove that the sequence (un ) is bounded.
Let ε = 1. Then by the given condition for ε = 1, there exists a natural number
k such that

|un+p − un | < 1 for all n ≥ k and for p = 1, 2, 3, . . . .

13
 Therefore, for p = 1, 2, 3, . . ., we have

|uk+p − uk | < 1

⇒ uk + 1 < uk+p < uk + 1.

Let B = max{u1 , u2 , u3 , . . . , uk , uk + 1} and b = min{u1 , u2 , u3 , . . . , uk , uk − 1}.

Then b ≤ un ≤ B for all n ∈ N.
This proves that (un ) is a bounded sequence.
By Bolzano-Weierstrass theorem, the sequence (un ) has a convergent subse-

s
quence. Let ` be the limit of that subsequence (unk ) of (un ).

Da
Let ε > 0 be a preassigned arbitrary real number.
Then by the given condition, there exists a natural number m such that

ish
|un+p − un | < for all n ≥ m and for p = 1, 2, 3, . . . .
3

In particular, sh
ε
|um+p − um | < for p = 1, 2, 3, . . . . (1)
3
ha
Since (unk ) converges to `, we have

ε
b

|unk − `| < for all nk > m. (2)

3
hu

Since nk > m, it follows form (1) that

ε
|unk − um | < . (3)
.S

Now,
Dr

|um+p − `| ≤ |um+p − um | + |um − unk | + |unk − `|

ε ε ε
< + + = ε for p = 1, 2, 3, . . . .
3 3 3

Therefore, for that given ε, we have |un − `| < ε for all n ≥ m + 1.

This proves that the sequence (un ) is convergent. 

Remark 4. Cauchy’s general principle of convergence states that “a real sequence

is convergent iff it is Cauchy”.

Theorem 12. Every convergent sequence is Cauchy.

14
Proof. Let (un ) be a convergent sequence and lim un = `. Then for a preassigned
ε there exists a natural number k such that

ε
|un − `| < for all n ≥ k.
2
ε
Therefore, |un+p − `| < 2
for all n ≥ k and for p = 1, 2, 3, . . . .
Now,

|un+p − un | = |un+p − ` + ` − un |

s
≤ |un+p − `| + |` − un |

Da
ε ε
< + = ε for all n ≥ k and for p = 1, 2, 3, . . . .
2 2

This proves the theorem. 

ish
Theorem 13. Every Cauchy sequence is convergent.

Proof. We now prove that the sequence (un ) is convergent as a consequence of

sh
the given condition.
First we prove that the sequence (un ) is bounded.
ha
Let ε = 1. Then by the given condition for ε = 1, there exists a natural number
k such that
b

|un+p − un | < 1 for all n ≥ k and for p = 1, 2, 3, . . . .

hu

Therefore, for p = 1, 2, 3, . . ., we have

.S

|uk+p − uk | < 1

⇒ uk + 1 < uk+p < uk + 1.

Dr

Let B = max{u1 , u2 , u3 , . . . , uk , uk + 1} and b = min{u1 , u2 , u3 , . . . , uk , uk − 1}.

Then b ≤ un ≤ B for all n ∈ N.
This proves that (un ) is a bounded sequence.
By Bolzano-Weierstrass theorem, the sequence (un ) has a convergent subse-
quence. Let ` be the limit of that subsequence (unk ) of (un ).
Let ε > 0 be a preassigned arbitrary real number.
Then by the given condition, there exists a natural number m such that

ε
|un+p − un | < for all n ≥ m and for p = 1, 2, 3, . . . .
3
15
 In particular,
ε
|um+p − um | < for p = 1, 2, 3, . . . . (1)
3
Since (unk ) converges to `, we have
ε
|unk − `| < for all nk > m. (2)
3
Since nk > m, it follows form (1) that
ε
|unk − um | < . (3)
3

s
Now,

Da
|um+p − `| ≤ |um+p − um | + |um − unk | + |unk − `|
ε ε ε
< + + = ε for p = 1, 2, 3, . . . .
3 3 3

ish
Therefore, for that given ε, we have |un − `| < ε for all n ≥ m + 1.
This proves that the sequence (un ) is convergent. 

Example.
sh
36. Show that the sequence (un ), where un = 1+ 12 + 13 +· · ·+ n1 cannot converge.
ha
Proof. Suppose, if possible, the sequence (un ) is convergent.
1
Let us take ε = and n = m and p = m in Cauchy’s general principle
b

of convergence, so that |u2m − um | < 12 .

hu

But,
1 1 1
u2m − um = + + ··· +
.S

m+1 m+2 m+m

1 1 1 1
> + + ··· + = ,
2m 2m 2m 2
i.e., |u2m − um | > 12 , which contradicts Cauchy’s general principle of con-
Dr

vergence.
Hence the sequence cannot converge. 
1 1 1
37. Show that the sequence (un ), where un = 1 + 3
+ 5
+ ··· + 2n−1
cannot
converge.

Theorem 14 (Cauchy’s first theorem on limits). If lim an = `, then

n→∞
 
a1 + a2 + · · · + an
lim = `.
n→∞ n
Proof. 
16
Corollary 1. If (an ) is a positive term sequence and lim an = `, then
n→∞
1
lim (a1 a2 · · · an ) = `. n
n→∞

Proof. 

Example.
1 + 12 + 31 + · · · + 1
n
36. Prove that lim = 0.
n→∞ n
Proof. Let ak = k1 , k = 1, 2, . . . , n.

s
1
Therefore, lim an = lim = 0. Thus by Cauchy’s first theorem on
n→∞ n→∞ n

Da
1
1+ 2
+ 3 + · · · + n1
1
limits, we have lim = 0. 
n→∞ n
1 1 1
1 + 22 + 33 + · · · + nn
37. Show that lim = 1.
n→∞ n

ish
1
Proof. Let ak = k k , k = 1, 2, . . . , n.
1
Therefore, lim an = lim n n = 1. Thus by Cauchy’s first theorem on
n→∞ n→∞
1 1 1
1 + 22 + 33 + · · · + nn
limits, we have lim
n→∞

n
= 1.
sh 


1 1 1
38. Show that lim √ +√ + ··· + √
ha
= 1.
n→∞ n2 + 1 n2 + 2 n2 + n
Proof. The given limit can be written as
b

 
1 1 1
lim  q + q + ··· + p
hu


n
n→∞
n 1 + n12 n 1 + n22 n 1+ n2
 1 
q
1
+ q 1 2 + · · · + √1+1
n
 1+ n2 1+ 2 n2 
.S

n
= lim 
n

n→∞

Let ak = q 1 , k = 1, 2, . . . , n.
k
1+
Dr

n2
Therefore, lim an = lim √1+
1
n = lim
√1 1
= 1. Thus by Cauchy’s
n→∞ n→∞ n2 n→∞ 1+ n
1 1
first theorem on limits, we have + · · · + √n12 +n = 1. 
lim √ 2 + √n2 +2
n→∞ n +1
 
1 1 1
39. Show that lim + + ··· + = 0.
n→∞ n2 (n + 1)2 (2n)2
Proof. 

Theorem 15 (Cauchy’s second theorem on limit). If (an ) is a positive term

an
sequence and lim = `, then
n→∞ an+1
1
lim (an ) n = `.
n→∞
17
Proof. 

Example.
1
36. If a > 0, then lim a n = 1.

Proof. 
1
37. Prove that lim n n = 1.

Proof. 
1
38. Prove that lim (n!)n
n
= 1e .

s
Da
Proof. 
1
39. Prove that lim {(n+1)(n+2)···(2n)}
n
n
= 4e .

Proof. 

ish
Theorem 16 (Monotone convergence theorem). A monotone increasing
bounded above sequence converges to its least upper bound.
sh
Proof. 
ha
Theorem 17. Proof. 

References
b

[1] R. G. Bartle and D. R. Sherbert, Introduction to Real Analysis, John Wiley & Sons, 1982.
hu

[2] S. C. Malik and S. Arora, Mathematical Analysis, New Age Int.(P) Ltd., New Delhi, 2008.
[3] S. K. Mapa, Real Analysis, Sarat Book Dist., Kolkata, 2006.
.S
Dr

18