1

Lecture 2: Convergence of Sequences

In Lecture 1, some properties of R were established. In this lecture, we introduce the concept of
convergence of a sequence of real numbers. We will see in the subsequent lectures that this concept
is useful in

(i) establishing more properties of R which are needed;

(ii) studying continuity, differentiation and integration of functions;

(iii) approximating irrational numbers and also in finding numerical solutions to some equations.

Sequences

Intuitively, a sequence of real numbers can be thought of as a list of real numbers x1 , x2 , . . .. In

effect, we are associating a real number xn for every n ∈ N. Therefore, in a formal sense, a sequence
of real numbers can be defined as a function from N to R. Henceforth, the sequence f : N → R
defined by f (n) = xn for all n ∈ N, will be denoted by either (x1 , x2 , x3 , . . .) or (xn ). We call xn
the nth term of the sequence (xn ).

We list below some examples of sequences:

1. (n) = (1, 2, 3, . . .)
2. ( n1 ) = (1, 12 , 13 , . . .)
n
3. ( (−1) 1 −1
n ) = (−1, 2 , 3 , . . .)

4. (1 − n1 ) = (0, 1 − 12 , 1 − 13 , . . .)
5. ((−1)n ) = (−1, +1, −1, +1, . . .)

Convergence of a Sequence

Before giving the formal definition of convergence of a sequence, let us take a look at the
behaviour of the sequences in the above examples. In the case of the sequences, ( n1 ), (1 − 1/n)
and ((−1)n /n), the nth term of each sequence appears to be approaching a fixed number as n gets
larger. On the contrary, the nth term of the sequence (n) increases as n increases. In the case of
the sequence ((−1)n ), the nth term oscillates between −1 and 1, accordingly as n is odd or even.

Let us focus our attention on those sequences (xn ), where the nth term xn approaches a fixed
number as n increases. To express the statement “xn approaches a fixed number” in a formal sense,
we need the concept of neighbourhood. For a ∈ R and  > 0, the −neighbourhood of a is the
open interval (a − , a + ) = {x ∈ R : a −  < x < a + }. On visualizing geometrically, it appears
that if the terms of the sequence (xn ) come eventually inside every −neighbourhood of x0 then xn
approaches x0 as n increases. This geometric visualization leads to the following formal definition
of convergence of a sequence of real numbers.

Definition 2.1. We say that a sequence (xn ) converges to a real number x0 if for every  > 0,
there exists N ∈ N such that xn ∈ (x0 − , x0 + ) whenever n ≥ N.

Note that in Definition 2.1, the condition “xn ∈ (x0 − , x0 + ) whenever n ≥ N ” can be
replaced by “|xn − x0 | <  for all n ≥ N ”. Observe that the value of N depends on the choice of .
Please write to [email protected] if any typos/mistakes are found in these notes.
 2

In general, as the value of  becomes smaller, the value of N becomes larger. This is illustrated in
Example 2.1.

Example 2.1. 1. Consider the sequence ( √1n ). We will show that this sequence converges to 0,
using Definition 2.1. Let  be any given positive real number. We have to find N ∈ N such that
| √1n − 0| <  for all n ≥ N . Note that the inequality √1n <  is true for all n > 12 . Hence choose
1 1
any N ∈ N such that N > 2
. Then for every n ≥ N , we have n > 2
. Thus | √1n − 0| <  for all
n ≥ N . This shows that ( √1n ) converges to 0.

2. Let xn = 8 for all n (such a sequence is called constant sequence), and  > 0 be given. Then
|xn − 8| <  for all n ≥ 1. Hence xn → 8. In this case, we take N = 1 for any given .

We now see that a sequence cannot converge to more than one real number. To prove this,
assume that (xn ) converges to x0 and y0 , and x0 < y0 . Let  = (y0 − x0 )/4. By Definition 2.1,
there exists N1 ∈ N such that xn ∈ (x0 − , x0 + ) whenever n ≥ N1 and there exists N2 ∈ N such
that xn ∈ (y0 − , y0 + ) whenever n ≥ N2 . Thus xn ∈ (x0 − , x0 + ) ∩ (y0 − , y0 + ) whenever
n ≥ max{N1 , N2 }. This is not possible, because, (x0 − , x0 + ) ∩ (y0 − , y0 + ) = ∅.

If (xn ) converges to some x0 then we say that (xn ) converges or (xn ) is a convergent sequence. We
call x0 the limit of the sequence (xn ) and we write lim xn = x0 or simply xn → x0 . A sequence
n→∞
which does not converge is called a divergent sequence. For example, the sequence ((−1)n ) is a
divergent sequence (see Problem 3 of PP2).

Elementary results about convergent sequences

We can show, as shown in Example 2.1, that n1 → 0, n12 → 0 and 1 + 101n → 1. In each of these
examples, from the behaviour of the elements of the sequence, it was possible to guess the limit
and then verify it (using the definition). However, in most cases, it will not be possible to follow
this method. In some cases, it might be possible to guess the limit, but verifying the convergence
from the definition will be difficult. Clearly, there is a need for more tools to address the issues of
convergence of sequences and the computation of their limits. In this context, the Limit Theorem,
Sandwich Theorem and Ratio Test (for sequences) are presented below.

Theorem 2.1 (Limit Theorem). Suppose xn → x and yn → y. Then

(i) xn + yn → x + y

(ii) xn yn → xy
xn x
(iii) yn → y if y 6= 0 and yn 6= 0 for all n.

Example 2.2. 1. Let (xn ) be the sequence where xn = 121+1 + 221+2 + · · · + n21+n for every n ∈ N.
We can rewrite xn as xn = 1 − 12 + 12 − 13 + · · · + n1 − n+1
1 1
= 1 − n+1 . Since n1 → 0, using Theorem
2.1 (i), we see that (xn ) converges and the limit of (xn ) is 1.
8 2
n3 +8n2 +2n 1+ n +
n2
2. Let (xn ) be the sequence where xn = 3n3 +2
. Here, xn = 2
3+ 3
. Use Theorem 2.1 and
n
1
verify that xn → 3.

Theorem 2.2 (Sandwich Theorem). Let (xn ), (yn ) and (zn ) be sequences, where xn ≤ yn ≤ zn
for all n and both (xn ) and (zn ) converge to x0 . Then yn → x0 .
 3

Proof: Let  > 0 be given. Since xn → x0 and zn → x0 , there exist N1 ∈ N and N2 ∈ N such that

xn ∈ (x0 − , x0 + ) for all n ≥ N1

and
zn ∈ (x0 − , x0 + ) for all n ≥ N2 .
Choose N = max{N1 , N2 }. Since xn ≤ yn ≤ zn , we have

yn ∈ (x0 − , x0 + ) for all n ≥ N.

This proves that yn → x0 . 

−1 sinn
Example 2.3. 1. Since n ≤ n ≤ n1 , by the sandwich theorem sinn
n → 0.
n2 n2 n2 n.n2 n.n2
2. If xn = n3 +n+1
+ n3 +n+2
+ ··· + n3 +2n
,n ∈ N, then n3 +2n
≤ xn ≤ n3 +n+1
. Thus xn → 1.

3. If x ∈ R and 0 < x < 1, then we show that xn → 0 as follows. Since 0 < x < 1, there exists
1
a > 0, such that that x = 1+a . By Bernoulli’s inequality (see Problem 2 of PP1), 0 < xn =
1 1 1 n
(1+a)n ≤ 1+na < na . From the sandwich theorem, we get x → 0.

1 1
4. Let x ∈ R and x > 0. We show that x n → 1. Suppose x > 1 and x n = 1 + dn for some dn > 0.
By Bernoulli’s inequality, x = (1 + dn )n > 1 + ndn > ndn which implies that 0 < dn < nx for all
1
n ∈ N. By the sandwich theorem dn → 0 and hence x n → 1. If 0 < x < 1, let y = x1 so that
1 1
y n → 1 and hence x n → 1.
1 1
5. We show that n n → 1. Let n n = 1 + kn for some kn > 0 when n > 1. Hence n = (1 + kn )n > 1
for n > 1. If n > 1, then by the binomial theorem, n ≥ 1 + 12 n(n − 1)kn2 . Therefore, for n > 1,
q 1
n − 1 ≥ 2 n(n − 1)kn and hence kn ≤ n2 . By the sandwich theorem kn → 0 and therefore n n → 1.
1 2

We need the following definition for stating the ratio test for sequence.

Definition 2.2. We say that (xn ) diverges to ∞ and write xn → ∞ if, for every M > 0, there
exists N ∈ N such that xn > M whenever n ≥ N . We say that (xn ) diverges to −∞ and write
xn → −∞ if, for every M > 0, there exists N ∈ N such that xn < −M whenever n ≥ N .

Example 2.4. Let xn > 0 for all n ∈ N. Then xn → 0 if and only if x1n → ∞ (see Problem 13 in
PP2). Hence if r > 1, then rn → ∞. Similarly, for any p > 0, np → ∞.

Theorem 2.3 (Ratio test for sequence). Let xn > 0 for all n and lim xxn+1
n
= λ for some
n→∞
λ ≥ 0. Then

(i) if λ < 1 then xn → 0;

(ii) if λ > 1 then xn → ∞.

xn+1
Proof. (i) Since λ < 1, choose r such that λ < r < 1. As lim = λ, there exists N ∈ N such
n→∞ xn
xn+1
that xn < r for all n ≥ N . Hence, for all n ≥ N ,
xN n+1
0 < xn+1 < rxn < r2 xn−1 < · · · < rn−N +1 xN = r .
rN
As N is a fixed number, C = xrNN is a constant. Further, 0 < xn+1 < Crn+1 for all n ≥ N and
rn → 0 (as 0 < r < 1). Using the sandwich theorem, we see that xn → 0.
 4

1 yn+1 1
(ii) Let yn = xn for all n ∈ N. Then lim = < 1. By part (i), yn → 0. Hence xn → ∞.
n→∞ yn λ

n xn+1
Example 2.4. 1. Let xn = 2n , n ∈ N. Since lim = 12 , by the ratio test, xn → 0.
n→∞ xn
p
2. Let us slightly change the preceding example. For every n ∈ N, let xn = nqn for some q > 1 and
p > 0. Apply the ratio test and show that xn → 0. Notice that both (np ) and (q n ) diverge to ∞.
The fact that xn → 0 reveals that (q n ) diverges to ∞ “faster” than (np ).
xn+1
3. Let xn = ny n−1 for some y ∈ (0, 1). Since lim = y, by the ratio test, xn → 0.
n→∞ xn

4. In Theorem 2.3, if λ = 1 then we cannot conclude either the convergence or divergence of the
given sequence. For example, consider the sequences (n), ( n1 ) and (2 + n1 ). Observe that in each of
these cases λ = 1. Further investigation is necessary if λ = 1.