Convergence and Continuity 2019/2020

Example 2.10
1 1 1 1 1
Let (xn )1
n=1 be the sequence given by xn = 2n , i.e. 2 , 4 , 8 , 16 , . . .. We would like to
1
compare this with the sequence yn = n (which we know converges to zero).

1
We claim that 2n  n1 , 8n 2 N.

Proof. We prove this by induction. We know that the statement is true for n = 1
(where it says that 12  1). Now assume that the claim holds for some n 2 N. Then
1 1 1 (?) 1 1 1 1
=  =  .
2n+1 22 n 2n n+n n+1
So the fact that the claim holds for n [we used this in (?)] implies that it also holds
for n + 1. Hence it holds for all n 2 N.

The claim, together with Lemma 2.3, then implies that (xn ) converges to zero.

We want to generalise this and show that the sequence given by xn = cn for some
c 2 R converges to zero if |c| < 1 (and does not converge to zero if |c| 1). To prove
this, it is useful to prove the following lemma first.

Lemma 2.11 (Bernoulli’s inequality)

If ↵ 2 R, ↵ 1, then 8n 2 N : (1 + ↵)n 1 + n↵.
Proof. We prove this again by induction. Obviously the statement is true for n = 1
(where it just says 1 + ↵ 1 + ↵). Assume now that it is true for some n 2 N. Then
(?)
(1 + ↵)n+1 = (1 + ↵)(1 + ↵)n (1 + ↵)(1 + n↵) = 1 + ↵ + n↵ + n↵2 1 + (n + 1)↵,
where in (?) we used the induction hypothesis as well as ↵ 1. Hence the claim
being true for n implies that it is true for n + 1 as well. Thus it holds for all
n 2 N.

Theorem 2.12 (“Geometric sequences”)

If c 2 R and the sequence (xn )1 n
n=1 is defined by xn = c , then

i) (xn ) converges to zero if |c| < 1.

ii) (xn ) does not converge to zero if |c| 1.

Proof. If c = 0, then xn = 0 for all n, and (xn ) converges to zero. So we may

assume c 6= 0.
1
i) If |c| < 1, then |c| = 1+↵ for some ↵ > 0. So by Lemma 2.11, we have
⇣ 1 ⌘n 1 1
|cn | = |c|n =  < .
1+↵ 1 + n↵ n↵
1
But ( n↵ ) converges to zero (by Lemma 2.5, since n1 converges to zero). Hence
the sequence (cn ) converges to zero by Lemma 2.3.

26
Convergence and Continuity 2019/2020

ii) If |c| 1 then |c| = 1 + ↵ for some ↵ 0. Hence, by Lemma 2.11

|cn | = |c|n 1 + n↵ 1,

for all n 2 N. But the constant sequence 1, 1, 1, . . . does not converge to zero
and hence (cn ) does not converge to zero by Corollary 2.4.
What about products and quotients of sequences? If (xn )1 1
n=1 and (yn )n=1 both con-
xn 1
verge to zero, what can we say about the sequences (xn · yn )1
n=1 and ( yn )n=1 ?

We cannot say much about the quotient ( xynn )1 n=1 . It will depend on whether the
nominator or denominator go to zero faster. For example, let (xn )1 n=1 be given by
xn = n1 and (yn )1
n=1 defined by y n = 1
n2
. We know that (x n ) and (y n ) both converge
to zero. But we have:
xn
• yn = n, so ( xynn ) does not converge to zero.
yn
• xn = n1 , so ( xynn ) does converge to zero.
Di↵erently from this fact, we can always say something about the product (xn yn )1
n=1 .

Theorem 2.13 (Product of sequences converging to zero)

If (xn )1 1 1
n=1 and (yn )n=1 both converge to zero then (xn · yn )n=1 converges to zero.

Proof. We follow a similar strategy to the proof of Theorem 2.8, but instead of split-
p p
ting " into 2" + 2" , we split it into " ".

We must prove that

8" > 0 9N 2 N 8n 2 N, n > N : |xn yn | < ".

To prove this, suppose we are given " > 0. Then

p
9Nx 2 N 8n 2 N, n > Nx : |xn | < "˜ = "

since (xn ) converges to zero, and

p
9Ny 2 N 8n 2 N, n > Ny : |yn | < "˜ = "

since (yn ) converges to zero. Let N = max{Nx , Ny }. Then for all n > N , we have
p p
|xn yn | = |xn ||yn | < " " = ".

Hence (xn · yn )1
n=1 converges to zero.

Example 2.14
2n 1 1
Consider xn = 3n (n+1) 2 . By Lemma 2.3, ( (n+1)2 ) converges to zero, as (n+1)2
< n1 .
2n
By Theorem 2.12, 3n = ( 23 )n converges to zero. Thus, by Theorem 2.13, (xn )
converges to zero.

27
Convergence and Continuity 2019/2020

Remark. Of course, we could have obtained this also by estimating xn < n1 and
using only earlier results, but sometimes it is difficult to obtain such estimates.

The proof of Theorem 2.13 suggests that we might be able to prove more general
results by splitting " up in other ways, e.g. " = "c · c for some real constant c > 0.
This is our next goal.

Definition 2.15 (bounded sequences)

We say that a sequence (xn )1
n=1 (of real numbers) is bounded above if 9M 2 R
such that xn  M for all n 2 N. The sequence is bounded below if 9m 2 R such
that xn m for all n 2 N. We say that (xn ) is bounded if it is bounded above
and below.

R
6
s s
s
M
s s
s

s s
-N

s
1 2 3 4 5
s s
s
m

Remark. Clearly, (xn )1 1

n=1 is bounded if and only if (|xn |)n=1 is bounded above.

Example 2.16
Every sequence (xn )1
n=1 which converges to zero is bounded.

Proof. If (xn ) converges to zero then taking " = 1 in the definition of convergence
to zero (Definition 2.1), there exists some N 2 N such that for all n > N we have
|xn | < 1. This is illustrated in the following figure.

R
6
s
s
1 s
s s s s -N
s s
3s
s
1 2 4 5
1 s

28
Convergence and Continuity 2019/2020

Take M to be max{|x1 |, |x2 |, . . . , |xN |, 1}. Then for all n 2 N, we have |xn |  M ,
i.e. M  xn  M for all n 2 N.

This shows that M is an upper bound and m = M is a lower bound for the
sequence (xn ) – as defined in Definition 2.15.

We are now ready to state and prove the more general version of Theorem 2.13.

Theorem 2.17 (Product of sequences, more general version)

If (xn )1 1 1
n=1 is bounded and (yn )n=1 converges to zero then (xn yn )n=1 converges to
zero.
Proof. As (xn )1
n=1 is bounded, there exists M 2 R such that |xn |  M for all n 2 N.
Hence
|xn yn | = |xn ||yn |  M |yn | = |M yn |, 8n 2 N.
But (yn ) converges to zero and so (M yn ) converges to zero as well (by Lemma 2.5),
so (xn yn ) converges to zero by Lemma 2.3.

Remark. Alternatively, we could prove this like we proved Theorem 2.13, splitting
" "
up " = M · M . (As (yn ) converges to zero, we will get |yn |  M for large enough n.)
2.B Convergence of sequences to non-zero elements of R

Definition 2.18 (Convergence of a sequence to x 2 R)

A sequence (xn )1
n=1 converges to x 2 R if and only if (xn x)1
n=1 converges to
1
zero. Equivalently, (xn )n=1 converges to x 2 R if and only if

8" > 0 9N 2 N 8n > N : |xn x| < ". (2.4)

We use the notation xn ! x as n ! 1 or limn!1 xn = x.

Theorem 2.19
Suppose (xn )1 1
n=1 converges to x 2 R and (yn )n=1 converges to y 2 R. Then

i) (cxn )1
n=1 converges to cx for any constant c 2 R.

ii) (xn + yn )1
n=1 converges to x + y.

iii) (xn yn )1
n=1 converges to xy.

iv) if y 6= 0 and yn 6= 0 for all n 2 N, then ( xynn )1

n=1 converges to
x
y.

Proof.
i) By the definition of convergence of (xn ) to x, we know that the sequence (xn x)
converges to zero. So by Lemma 2.5, the sequence c(xn x) converges to zero,
i.e. (cxn cx) converges to zero. But then by definition (cxn ) converges to cx.

29
Convergence and Continuity 2019/2020

ii) As xn ! x and yn ! y as n ! 1, we know that (xn x) and (yn y) are two

sequences that converge to zero. Hence by Theorem 2.8, ((xn x)+(yn y))1 n=1
converges to zero. Rewriting this, we obtain that ((xn + yn ) (x + y))1 n=1
converges to zero. But then by definition (xn + yn ) converges to x + y.

iii) We know that (xn x) and (yn y) converge to zero. We write

x n yn xy = (xn x)(yn y) + x(yn y) + (xn x)y. (2.5)

Then, we know that

• ((xn y)(yn y)) converges to zero (by Theorem 2.13).

• (x(yn y)) converges to zero (by Lemma 2.5, as (yn y) converges to zero).
• ((xn x)y) converges to zero (again by Lemma 2.5, as (xn x) converges
to zero).

So by Theorem 2.8, the right hand side of (2.5) converges to zero and hence
(xn yn xy) converges to zero. But then by definition (xn yn ) converges to xy.

iv) This is a question on Coursework Sheet 4.

Examples 2.20
i) Consider the sequence (xn ) given by xn = c for all n 2 N. This sequence
converges to c.

Proof. We must prove that (xn c) converges to zero. But xn c = 0 for all
n 2 N and we already know that the zero sequence converges to zero.

3n2 +5n+2
ii) xn = 2n2 +n+1
. This sequence converges to 32 .

Proof. We write xn in the form

5 2
3+ n + n2
xn = 1 1 .
2+ n + n2

But (3 + n5 + n22 ) converges to 3 + 0 + 0 = 3. [This uses Theorem 2.19 ii) and the
facts that the constant sequence 3 converges to 3 while ( n5 ) and ( n22 ) converge to
zero – the latter due to Lemma 2.3 and Lemma 2.5, as seen several times before.]

Similarly, (2 + n1 + n12 ) converges to 2 + 0 + 0 = 2, again by Theorem 2.19 ii)

[and the fact that ( n1 ) and ( n12 ) converge to zero].

In a last step, we then apply Theorem 2.19 iv) to conclude that (xn ) converges
to 32 .

30
Convergence and Continuity 2019/2020

Lemma 2.21 (Limits are unique)

If (xn )1
n=1 converges to x 2 R and also converges to y 2 R then x = y. That is: if
a limit exists, then it is unique.

Proof. Suppose x 6= y. Without loss of generality, assume y > x and write ↵ for the
di↵erence y x = |y x|.

Since (xn ) converges to x, we know that 9Nx 2 N such that 8n > Nx , we have
|xn x| < ↵2 .

Similarly, since (xn ) converges to y, we know that 9Ny 2 N such that 8n > Ny , we
have |xn y| < ↵2 .

↵
Now consider some xn with n max{Nx , Ny }. This xn satisfies |xn x| < 2 and
|xn y| < ↵2 . Hence, using the triangle inequality,
↵ ↵
|y x|  |y xn | + |xn x| = |xn y| + |xn x| < + = ↵.
2 2
So |y x| < ↵, contradicting our definition of ↵ as |x y|.

Lemma 2.22
Suppose we have two sequences (xn ) and (yn ) with xn ! x and yn ! y as n ! 1.
Suppose further that xn  yn for all n 2 N. Then x  y.

Proof. Since xn ! x as n ! 1, we can find for every " > 0 some Nx 2 N such that
8n > Nx : |xn x| < ", hence in particular xn > x ". Since yn ! y as n ! 1,
we can find for every " > 0 some Ny 2 N such that 8n > Ny : |yn y| < ", and
thus in particular yn < y + ". Setting N = max{Nx , Ny } we obtain for all n > N
that
x " < xn  yn < y + ",
that is x y < 2". But because this is true for every " > 0, we conclude that
x y  0, or equivalently x  y.

Examples 2.23
i) If (xn )1
n=1 converges to x 2 R and there exists a constant c 2 R such that xn  c
for all n 2 N, then x  c.

Proof. We know that (yn ) defined by yn = c for all n 2 N converges to c. We

can then apply Lemma 2.22 to (xn ) and this choice of (yn ).

Similarly, if (xn )1
n=1 converges to x 2 R and there exists a constant c 2 R such
that xn c for all n 2 N, then x c.

31