Mathematical Database

CONVERGENCE AND LIMITS

1. Introduction

In “Introductory Theory of Differentiation” we introduced the theory of limits of functions. In

particular we defined the meaning of the phrase
lim f ( x ) = L .
x→a

In this article we will pursue this matter further, where we will study one-sided limits (left-hand and
right-hand limits) of functions, as well as limits at infinity and limits of sequences.

2. What is a function?

To avoid putting too much into one single article, in “Introductory Theory of Differentiation”
we made no attempts to clarify what a function is. We just let the reader interpret the word
“function” in whatever way he/she likes. Now to make our discussion precise, let us first clarify
what we mean by a function. Let D be a collection of real numbers. For example, D may be the set
of all non-zero real numbers, or D may be the interval [0, 1] that consists of the collection of all real
numbers x satisfying 0 ≤ x ≤ 1 . A real-valued function f defined on D is a rule that assigns one and
only one real number to each real number in D. For example,
sin x
f ( x) =
x
is a real-valued function defined on the set of all non-zero real numbers (note that it is not defined at
the real number x = 0 because division by 0 is meaningless), while
x2 − 4
g ( x) =
x−2

is a real-valued function defined on the set of all real numbers that are not equal to 2. Also
h( x ) = x

is a real-valued function defined on the set of all non-negative real numbers (note that the square
root of a negative real number is undefined). If f is a real-valued function defined on D, then we call
D the domain of the function f.

The above definition makes rigorous what we mean by “functions”. We can now proceed to
discuss limits of functions.

Page 1 of 28
 Mathematical Database

3. More on Limits of Functions

Recall that, intuitively, by

lim f ( x ) = L ,
x→a

we mean that f (x) can be made very close to L whenever x is very close to a. More precisely, this
means that we can make f (x) arbitrarily close to L (i.e. as close to L as we want) by simply taking x
to be sufficiently close to a. To talk about whether f (x) is close enough to L or whether x is close
enough to a, it is desirable that we can introduce a quantity to measure how close two real numbers
are. To do so, let us introduce here the concept of a distance.

Definition 3.1.
Let a and b be two real numbers. Then the distance d(a, b) between a and b is defined to be the
absolute value of their difference, i.e.

d ( a, b) = a − b .

This distance has a number of useful properties:

(1) d (a, b) ≥ 0 for any real numbers a and b, with d (a, b) = 0 if and only if a = b.
(2) d (a, b) = d (b, a) for any real numbers a and b.
(3) d (a, c) ≤ d (a, b) + d (b, c) for any real numbers a, b and c.

The last property (3) is commonly known as the triangle inequality. (Compare with the one
given in Section 3 of “Introductory Theory of Differentiation”!) Can you prove these three
properties?

Now that we have a distance function, we can quantify how close two real numbers are. This
paves the way towards defining the meaning of the phrase
lim f ( x ) = L .
x→a

Just one more remark for now: If we consider the real-valued function
h( x ) = x

Page 2 of 28
 Mathematical Database

that is only defined on the set of non-negative real numbers, it certainly makes no sense to talk
about the limit of h as x tends to –1. The reason is that –1 is too far away from D. We want to take
care of this when we define limits of functions. This leads us to introduce the concept of a limit
point.

Figure 1: Graph of f ( x ) = x

Definition 3.2.
Let D be a collection of real numbers. We say a real number a is a limit point of D if for any δ > 0,
there exists a real number x in D that satisfies x ≠ a and d(x, a) < δ .

For example, if D is the set of all non-zero real numbers, then 0 is a limit point of D even
though 0 is not in D. Indeed every real number is a limit point of this D. If now D is the set of all
positive real numbers, then any non-negative real number is a limit point of D, while any negative
number is not a limit point of D. In fact the limit points of D are precisely those real numbers that
are arbitrarily close to D.

Together with the discussion at the beginning of the Section (also see the discussion in Section
3 of “Introductory Theory of Differentiation”), we are thus led to the following definition:
(Compare with the Definition 3.1 of “Introductory Theory of Differentiation”!)

Page 3 of 28
 Mathematical Database

Definition 3.3.
Let D be a collection of real numbers and let f be a real-valued function defined on D. Suppose that
a is a limit point of D and L is a real number. We say “f (x) tends to L as x tends to a”, or
equivalently

lim f ( x ) = L ,
x→a

if and only if the following holds:

For any ε > 0, there exists δ > 0 such that whenever x is a real number in D that satisfies x ≠ a and
d(x, a) < δ, we have

d ( f ( x), L) < ε .

(Note that we have required that x to be in D so that f (x) is defined.)

Does the above definition make sense for you? (It is important that it does, because then you
understand why it is such defined.)

Definition 3.4.

Let D be a collection of real numbers and let f be a real-valued function defined on D. Suppose that
a is a limit point of D. If there is a real number L such that

lim f ( x ) = L ,
x→a

then we say that

lim f ( x)
x→a

exists (and is equal to L).

Can you give an example where lim f ( x) does not exist?

x→a

Example 3.1.
Show that for any real number a,
lim x
x →a

exists and equals a .

Solution.
First, observe that the function
Page 4 of 28
 Mathematical Database

x6 x

is defined for any real number x, so the domain of this function is the set of all real numbers. In
particular, any real number a is a limit point of the domain of this function. Next, we divide our
proof into a number of cases:
Case 1: a > 0
For any ε > 0, we can take δ = min (ε, a) > 0, and then we will have, for any real number x that
satisfies d(x, a) < δ , that
x − a = x−a <δ ≤ε.

This proves, in this case, that

lim x
x →a

exists and equals a .

Case 2: a = 0
For any real number x, we have x − a = x − 0 = x , so for any ε > 0, we can take δ = ε > 0, and
then we will have, for any real number x that satisfies d(x, a) < δ , that
x − a = x <δ =ε .

As a result, in this case, we also have that

lim x
x →a

exists and equals a .

Case 3: a < 0
For any ε > 0, we can take δ = min (ε, −a) > 0, and then we will have, for any real number x that
satisfies d(x, a) < δ , that
x − a = −x + a = x − a < δ ≤ ε .

This proves, in this case, that

lim x
x →a

again exists and equals a . This completes our proof.

Alternative Solution.
Observe that for any real numbers x and a, we have
x − a ≤ x−a

Page 5 of 28
 Mathematical Database

(because x − a ≤ x − a and a − x ≤ x − a both hold by Triangle Inequality). Hence for any ε >
0, we can take δ = ε > 0, and then we will have, for any real number x that satisfies d(x, a) < δ , that
x − a ≤ x−a <δ =ε.

This also completes our proof.

Figure 2: Graph of f ( x ) = x

When we write
lim f ( x ) = L ,
x→a

we want to be certain that there is no other real number M such that

lim f ( x ) = M
x→a

also holds. This is the content of the following Proposition.

Proposition 3.1.

Let D be a collection of real numbers and let f be a real-valued function defined on D. Suppose that
a is a limit point of D and L, M are real numbers. If f (x) tends to both L and M as x tends to a, then
L = M. In other words, the limit of a function is unique (if it exists).

Proof. Suppose on the contrary that L ≠ M . Then L − M > 0 , so take

L−M
ε=
2

Page 6 of 28
 Mathematical Database

we have ε > 0, and hence (by definition that f (x) tends to both L and M as x tends to a), we have
that there exists δ 1 > 0 and δ 2 > 0 such that:
(I) whenever x is a real number in D that satisfies x ≠ a and d ( x, a ) < δ 1 , we have
d ( f ( x), L) < ε ;
(II) whenever x is a real number in D that satisfies x ≠ a and d ( x, a ) < δ 2 , we have
d ( f ( x), M ) < ε .
Now set δ = min(δ 1 , δ 2 ) , we have δ > 0, so taking x0 in D that satisfies d( x0 , a) < δ (this is
possible because a is a limit point of D), we have, by (I) and (II), that
d ( f ( x0 ), L) < ε and d ( f ( x0 ), M ) < ε .

This proves that

L − M = d ( L, M ) ≤ d ( f ( x0 ), L) + d ( f ( x0 ), M ) < ε + ε = 2ε = L − M ,

which is a contradiction. Hence L = M, as desired.

Q.E.D.

Given two real-valued functions f and g that are defined on the same domain D, we can define
their sum f + g and their product f ⋅ g by

(f + g ) ( x)  f ( x) + g ( x) and ( f ⋅ g ) ( x)  f ( x) ⋅ g ( x)

for all x in D. Then f + g and f ⋅ g are again real-valued functions defined on D. We can also define
the quotient function
f
g

by
f f ( x)
( x) 
g g ( x)
for all x in D’, where D’ is the set of all x in D that satisfies g ( x) ≠ 0 . Then the quotient function is
a real-valued function defined on D’.

In the next Proposition we are concerned with the limits of the sums, products and quotients of
two functions. Its proof is basically the same as that of Proposition 3.2 in “Introductory Theory of
Differentiation”; only minor modifications are called for. It is a good exercise that you reproduce
their proofs using mainly the properties (1) to (3) of the distance d introduced above. Try it out!

Page 7 of 28
 Mathematical Database

Proposition 3.2.

Let D be a collection of real numbers and let f, g be real-valued functions defined on D. Suppose
that a is a limit point of D. If

lim f ( x) and lim g ( x)

x→a x→a

both exists, then:

1. lim ( f + g ) ( x) exists and is equal to lim f ( x) + lim g ( x) ;
x →a x→a x→a

2.
x →a
( x →a
)(
lim ( f ⋅ g ) ( x) exists and is equal to lim f ( x) ⋅ lim g ( x) ; and
x→a
)
3. If

lim g ( x) ≠ 0 ,
x →a

then there exists δ > 0 such that g ( x) ≠ 0 for all x in D that satisfies d(x, a) < δ, so if we
denote by D’ the set of all x in D that satisfies g ( x) ≠ 0 , then a is a limit point of D’,
f
lim ( x)
x→a g

exists and is equal to

lim f ( x )
x→a
.
lim g ( x )
x→a

Recall that we have

lim x = a
x→a

and
lim c = c
x→a

for any real constant c. Thus from Proposition 3.2 we obtain:

Corollary 3.3.

For any polynomial p(x) and any real number a, we have

lim p ( x) = p (a ) .
x→a

This says that we can evaluate limits of polynomials by direct substitution. (But note that in
general limits CANNOT be evaluated by direct substitution!)

Page 8 of 28
 Mathematical Database

Next we state and prove the Sandwich Theorem, which is often so useful.

Proposition 3.4. (Sandwich Theorem)

Let D be a collection of real numbers and let f, g, h be real-valued functions defined on D. Suppose
that a is a limit point of D, and suppose that there exists δ > 0 such that
f ( x ) ≤ g ( x ) ≤ h( x )
whenever x is a real number in D that satisfies x ≠ a and d(x, a) < δ . If

lim f ( x) and lim h( x)

x→a x →a

both exists and have a common value of L, then

lim g ( x)
x→a

also exists and is equal to L.

Proof. Let ε > 0 be given. Then since

lim f ( x) and lim h( x)
x→a x →a

both exists and are equal to L, we have that there exists δ 1 > 0 and δ 2 > 0 such that:
(I) whenever x is a real number in D that satisfies x ≠ a and d ( x, a ) < δ 1 , we have f ( x) − L < ε ;
in particular for any such x we have L − ε < f ( x) .
(II) whenever x is a real number in D that satisfies x ≠ a and d ( x, a ) < δ 2 , we have h( x) − L < ε ;
in particular for any such x we have h( x) < L + ε .
Hence setting δ = min(δ 1 , δ 2 ) , we have δ > 0, and whenever x is a real number in D that satisfies
x ≠ a and d(x, a) < δ , we have, by (I) and (II), that
L − ε < f ( x ) ≤ g ( x ) ≤ h( x ) < L + ε .

In particular for any such x we have

g ( x) − L < ε .

This proves that

lim g ( x)
x→a

exists and is equal to L, and we are done.

Q.E.D.

Page 9 of 28
 Mathematical Database

Example 3.2.
For each real number x, let
⎧ m
⎪ x if x = for some positive integers m and n
f ( x) = ⎨ n
⎪⎩0 otherwise.

Show that
lim f ( x)
x →0

exists and find its value.

Solution.
Simply observe that the inequalities
0 ≤ f ( x) ≤ x

hold for any real number x. Thus from the fact that
lim 0 and lim x
x→ 0 x →0

both exist and equal 0 we conclude, using the Sandwich Theorem, that
lim f ( x)
x →0

exists and is equal to 0.

Exercise
Throughout the exercise, suppose that f is a real-valued function and a is a limit point of the domain
of definition of f. Also assume L is a real number.

1. Show that the following limits exist and find their values.
x2 + 3 x2 − 4 x−5
(a) lim (b) lim (c) lim
x→0 x−5 x→2 x−2 x →5 x 2 − 25

2. Show that if
lim f ( x)
x→a

exists and is equal to L, then

lim f ( x)
x→a

Page 10 of 28
 Mathematical Database

exists and is equal to L . Does the converse hold?

3. Show that the following statements are equivalent:

(a) lim f ( x) = 0
x→a

(b) lim f ( x) = 0
x →a

4. Assume f (a) is defined. Show that if

f ( x) − f (a)
lim = L,
x→a x−a

then
lim f ( x)
x→a

exists and equals f (a).

5. Suppose that the only property that you know of the sine function is that sin x ≤ 1 for all real
numbers x. Prove that
⎛ 1⎞
lim ⎜ x sin ⎟ = 0 .
x →0
⎝ x⎠

6. (a) Let a be a real number. If 0 ≤ a < ε holds for all positive numbers ε, show that a = 0.
What if we only know 0 ≤ a ≤ ε for all positive numbers ε ? Does the same conclusion
still hold?
(b) Let a and b be real numbers. If 0 ≤ d ( a, b) ≤ ε holds for all positive numbers ε, show that
a = b.

7. Suppose that
lim f ( x ) = L .
x→a

Show that if L > 0, then there exists δ > 0 such that

f ( x) > 0
holds for all x in the domain of definition of f that satisfies x ≠ a and d ( x, a ) < δ . Also show,
by means of a counter-example, that if we only have L ≥ 0 , then the above conclusion does
not hold even with f ( x) > 0 replaced by f ( x) ≥ 0 .

Page 11 of 28
 Mathematical Database

4. One-sided Limits

In this section we look at how left-hand limits and right-hand limits may be defined. We shall
only work on left-hand limits; it should then be clear how we can define right-hand limits.

Intuitively, by
lim f ( x) = L
x→a −

we mean that f (x) can be made arbitrarily close to L as long as we take x to be strictly smaller than
a and at the same time sufficiently close to a. This leads us to the following definition of left limit
points and left-hand limits.

Definition 4.1.
Let D be a collection of real numbers. We say a real number a is a left limit point of D if for any
δ > 0, there exists a real number x in D that satisfies a − δ < x < a .

Note that a left limit point of D is always a limit point of D, but not vice versa.

Definition 4.2.
Let D be a collection of real numbers and let f be a real-valued function defined on D. Suppose that
a is a left limit point of D and L is a real number. We say “f (x) tends to L as x tends to a−”, or
equivalently

lim f ( x) = L ,
x→a −

if and only if the following holds:

For any ε > 0, there exists δ > 0 such that whenever x is a real number in D that satisfies
a − δ < x < a , we have
d ( f ( x), L) < ε .

Just like the case for limits, left-hand limits may not always exist. We have the following
definition:

Page 12 of 28
 Mathematical Database

Definition 4.3.

Let D be a collection of real numbers and let f be a real-valued function defined on D. Suppose that
a is a left limit point of D. If there is a real number L such that

lim f ( x) = L ,
x→a −

then we say that the left-hand limit of f exists at a.

Example 4.1.
For each real number x, let [x] denote the greatest integer smaller than or equal to x. If a is a real
number, show that
lim ( x − [ x])
x→a −

exists and find its value.

Figure 3: Graph of f ( x) = x − [ x]

Solution.
We will prove that for any real number a,
⎧1 if a is an integer
lim ( x − [ x]) = ⎨
x→a −
⎩a − [a] if a is not an integer.

Case 1: a is an integer.
Then for any real number x that satisfies a − 1 ≤ x < a , we have [x] = a – 1, so for any such x we
have
d ( x − [ x], 1) = d ( x − a + 1, 1) = d ( x, a ) .

Hence given any ε > 0, we can simply take δ = min(ε , 1) > 0, and then we will have, for any real x
that satisfies a − δ < x < a , that a − 1 ≤ x < a , so d ( x − [ x], 1) = d ( x, a ) < ε . This proves that in
this case
lim ( x − [ x])
x→a −

exists and is equal to 1.

Case 2: a is not an integer.
Page 13 of 28
 Mathematical Database

Then a – [a] > 0, and for any real number x that satisfies [a ] ≤ x < a , we have [x] = [a], so for any
such x we have

d ( x − [ x], a − [a ]) = d ( x − [a ], a − [a ]) = d ( x, a ) .

Hence given any ε > 0, we can take δ = min ( ε , a − [a ]) > 0 , and then we will have, for any real x
that satisfies a − δ < x < a , that [a ] ≤ x < a , so d ( x − [ x], a − [a ]) = d ( x, a ) < ε . This proves that in
this case
lim ( x − [ x])
x→a −

also exists and is equal to a − [a].

Left-hand limits have properties similar to those of ordinary limits. In particular, Propositions
3.1, 3.3 and 3.4 hold for left-hand limits with trivial modifications. Can you work them out?

Now that we can define left limit points and left-hand limits, we can similarly define right
limit points and right-hand limits. Right-hand limits enjoy properties similar to left-hand limits. The
link between the three types of limits we introduced so far is given in the following Proposition:

Proposition 4.1.

Let D be a collection of real numbers and let f be a real-valued function defined on D. Suppose that
a is a left limit point of D as well as a right limit point of D. Then for any real number L, we have

lim f ( x ) = L
x→a

if and only if
lim f ( x ) = lim f ( x ) = L .
x→a− x→a +

Proof. Suppose that

lim f ( x ) = L .
x→a

Let ε > 0 be given. Then there exists δ > 0 such that whenever x is a real number in D that satisfies
x ≠ a and x − a < δ , we have d ( f ( x), L ) < ε . Hence if x is a real number in D that satisfies
a − δ < x < a , we have d ( f ( x), L ) < ε ; if x is a real number in D that satisfies a < x < a + δ , we
also have d ( f ( x), L ) < ε . This proves that
lim f ( x ) = L = lim f ( x) .
x→a− x→a +

Conversely, suppose that

Page 14 of 28
 Mathematical Database

lim f ( x ) = lim f ( x ) = L .
x→a− x→a +

Let ε > 0 be given. Then there exists δ 1 > 0 and δ 2 > 0 such that:
(I) whenever x is a real number in D that satisfies a − δ 1 < x < a , we have d ( f ( x), L ) < ε ; and

(II) whenever x is a real number in D that satisfies a < x < a + δ 2 , we have d ( f ( x), L ) < ε .

Hence setting δ = min(δ 1 , δ 2 ) , we have δ > 0, and whenever x is a real number in D that satisfies
x ≠ a and d(x, a) < δ , we have, by (I) and (II), that
d ( f ( x), L ) < ε .

This proves that

lim f ( x ) = L .
x→a

Q.E.D.

The above Proposition sometimes helps us determine whether a given limit

lim f ( x)
x→a

exists. This is illustrated in the following Example. (An example where the above Proposition does
not help is given in the Exercise at the end of this section. See Question 1 there.)

Example 4.2.

x
Figure 4: Graph of f ( x) =
x

Show that the limit

x
lim
x →0 x
does not exist.

Page 15 of 28
 Mathematical Database

Solution.
It is easy to see that the function
x
x6
x
is defined for all non-zero real numbers. So 0 is both a left limit point and a right limit point of the
domain of this function. Now
x ⎧1 if x > 0
=⎨
x ⎩−1 if x < 0

so
x x
lim = −1 and lim =1.
x →0− x x →0 + x
Since the left-hand and right-hand limits are not equal, we conclude that
x
lim
x →0 x
does not exist.

Note that in Proposition 4.1, it was assumed that a is a left limit point of D as well as a right
limit point of D, where D is the domain of the function being considered. What if this does not hold?
For instance, what if a is only a left limit point of D but not a right limit point of D? The answer is
actually simple: if a is only a left limit point of D but not a right limit point of D, then
lim f ( x)
x→a

exists if and only if

lim f ( x)
x→a −

exists. Can you prove this?

Exercise

1. Let g be the real-valued function defined on the set of all real numbers by
⎧ 1
⎪1 if is an integer
g ( x) = ⎨ x
⎪⎩0 otherwise.

Page 16 of 28
 Mathematical Database

Show that lim g ( x ) does not exist. Explain why it is not helpful here to consider left-hand and
x →0

right-hand limits.
(Hint: Show that no real numbers can be the limit of g(x) as x tends to zero by definition of
limit. Consideration of left-hand and right-hand limits is not helpful here because it is not the
“jump” of the function that causes the non-existence of limit. Rather it is the oscillations in the
function that leads to the non-existence of limit.)

2. Discuss the existence of the limit lim ( x − [ x]) for various values of a.
x→a

5. Limits at Infinity

Sometimes we wish to discuss the behaviour of a function f (x) as x tends to infinity. This is
done as follows.

Definition 5.1.
Let D be a collection of real numbers. We say positive infinity ( +∞ ) is a limit point of D if for any
real number A, there exists a real number x in D that satisfies x > A .

Definition 5.2.
Let D be a collection of real numbers and let f be a real-valued function defined on D. Suppose that
+∞ is a limit point of D and L is a real number. We say “f (x) tends to L as x tends to +∞ ”, or
equivalently

lim f ( x ) = L ,
x →+∞

if and only if the following holds:

For any ε > 0, there exists a real number A such that whenever x is a real number in D that satisfies
x > A , we have
d ( f ( x), L) < ε .

The point here is that as long as x is sufficiently large (i.e. sufficiently close to +∞ ), we will
have f (x) arbitrarily close to the limit L. Similarly we can discuss the limit of f (x) as x tends to

Page 17 of 28
 Mathematical Database

negative infinity. We sometimes drop the + sign in front of +∞ and simply write ∞ for positive
infinity, especially when no confusion will occur. (This will happen in Section 7.)

Note how the above definition resembles the definition of limit that we gave in Sections 3 and
4. Propositions 3.1, 3.2 and 3.4 still hold (under the evident modifications). Can you work out how?

Example 5.1.
Show that
1
lim = 0.
x →+∞ x

1
Figure 5: Graph of f ( x) =
x

Solution.
It is easy to see that the function
1
x6
x

is defined for all non-zero real numbers. So +∞ is a limit point of the domain of this function. Now
given ε > 0, we can simply take A = 1/ ε . Then for any non-zero real number x that satisfies x > A,
we have
⎛1 ⎞ 1 1
d ⎜ ,0⎟ = < = ε .
⎝x ⎠ x A

(Note we used A > 0 here.) Hence

1
lim = 0.
x →+∞ x

Page 18 of 28
 Mathematical Database

Example 5.2.
Compute
x
lim .
x →+∞ x − 1

x
Figure 6: Graph of f ( x) =
x −1

Solution.
It is easy to see that the function
x
x6
x −1
is defined for all x ≠ 1 . So +∞ is a limit point of the domain of this function. Now the crucial
observation is that
x 1 1
= ≈ =1
x −1 1 1− 0
1−
x

when x is close to +∞ . So using the estimate

x 1
−1 =
x −1 x −1
1
which holds for x > 1, we see that given any ε > 0, we can simply take A = 1 + . Then for any real
ε
number x that satisfies x > A, we have
x 1 1
−1 = < =ε .
x −1 x −1 A −1

This proves
x
lim = 1.
x →+∞ x − 1

Page 19 of 28
 Mathematical Database

Alternative Solution.
From
x 1
=
x −1 1
1−
x

which holds for x ≠ 0 we have, from a variation of Proposition 3.2 and from the previous Example,
that
x
lim
x →+∞ x −1

exists and that

x 1
lim = = 1.
x →+∞ x − 1 1− 0

6. Infinite Limits

It is sometimes useful too to be able to give

lim f ( x) = +∞
x→a

a precise meaning. We do so in the case where a is a real number (and a is a limit point of the
domain of definition of f ). The case where a is positive or negative infinity is left to you.

Definition 6.1.
Let D be a collection of real numbers and let f be a real-valued function defined on D. Suppose that
a real number a is a limit point of D. We say “f (x) tends to +∞ as x tends to a”, or equivalently

lim f ( x) = +∞ ,
x→a

if and only if the following holds:

For any M > 0, there exists δ > 0 such that whenever x is a real number in D that satisfies x ≠ a and
d(x, a) < δ, we have

f ( x) > M .

The point here is that f (x) is arbitrarily large as long as x is sufficiently close to a. Can you
work out some examples for yourself?

Page 20 of 28
 Mathematical Database

Exercise

1. Suppose that f is a real-valued function and a is a limit point of the domain of definition of f.
Suppose that
lim f ( x) = +∞ .
x→a

Show that
1
lim = 0.
x→a f ( x)

7. Limits of Sequences

Finally we consider a special type of functions, namely those whose domain of definition is the
set of positive integers.

Definition 7.1.

Let x be a real-valued function defined on the set of positive integers. Then we say x is a sequence
of real numbers. Usually we think of x(n) as the n-th term of the sequence, and we write xn instead
of x(n). The sequence x is also written as { xn } .

For example,
1
xn =
n

defines a sequence of real numbers.

It is obvious that for a function defined on the set of positive integers, the only limit point of its
domain of definition is +∞ . Thus the definition in Section 5 applies. We state it here in a way that
is equivalent to, but that looks slightly different from, the definition given in Section 5.

Definition 7.2.

Let { xn } be a sequence of real numbers and L be a real number. We say “ xn tends to L as n tends to
infinity”, or equivalently

Page 21 of 28
 Mathematical Database

lim xn = L ,
n →∞

if and only if the following holds:

For any ε > 0, there exists a positive integer N such that whenever n is a positive integer that
satisfies n > N, we have

d ( xn , L) < ε .

Definition 7.3.

Let { xn } be a sequence of real numbers. If there is a real number L such that

lim xn = L ,
n →∞

then we say that

lim xn
n →∞

exists (and is equal to L). We also say that the sequence { xn } converges (or is convergent). If a
sequence does not converge, we say that it diverges (or is divergent).

Again variants of Propositions 3.1, 3.2 and 3.4 hold. In particular, the limit of a sequence of
real numbers is unique (if it exists at all); the sum and product of two convergent sequences are
convergent; the quotient of two convergent sequences is also convergent provided that the limit of
the sequence in the denominator is non-zero; and the Sandwich Theorem holds. To emphasize these
theorems, we state them here in the apparently new setting of sequences. Can you supply their
proofs?

Proposition 7.1.

Let { xn } be a sequence of real numbers. Suppose that L and M are real numbers. If xn tends to both
L and M as n tends to infinity, then L = M. In other words, the limit of a sequence is unique (if it
exists at all).

Proposition 7.2.

Let { xn } and { yn } be two sequences of real numbers. If

lim xn and lim yn

n →∞ n →∞

Page 22 of 28
 Mathematical Database

both exists, then:

1. lim ( xn + yn ) exists and is equal to lim xn + lim yn ;

x →a n →∞ n →∞

2.
x →a
( n →∞
)(
lim ( xn yn ) exists and is equal to lim xn ⋅ lim yn ; and
n →∞
)
3. If lim yn ≠ 0 , then there exists a positive integer N such that yn ≠ 0 for all positive integers n
n →∞

xn x
> N, so is a real number whenever n > N. In this case we actually have lim n exists and is
yn n →∞ yn
equal to
lim xn
n →∞
.
lim yn
n →∞

Proposition 7.3. (Sandwich Theorem)

Let { xn } , { yn } and { zn } be three sequences of real numbers and L be a real number. Suppose that
there exists a positive integer N such that
xn ≤ yn ≤ zn
for all positive integers n > N . If

lim xn and lim zn

n →∞ n →∞

both exists and have a common value of L, then

lim yn
n →∞

also exists and is equal to L.

It is “well-known” that
1
lim = 0.
n →∞ n

This can be proved, but it involves a deep assumption on the collection of real numbers. So we will
simply assume it here and prove the following results.

Example 7.1.
Prove that for any −1 < a < 1, we have
lim a n = 0 .
n →∞

Page 23 of 28
 Mathematical Database

Solution.
We divide the proof into three steps:
Step 1: 0 < a < 1
Since in this case
1
> 1,
a

we can write
1
a=
1+ x

for some x > 0. Then for any positive integer n, we have, by binomial theorem, that
(1 + x) n ≥ 1 + nx ,

and thus
1
1 1
0 ≤ an = ≤ = n .
(1 + x) n 1 + nx 1
+x
n

By Proposition 7.2, as n tends to infinity, the right hand side tends to

0
=0.
0+ x
1
(Note we used lim = 0 here.) This proves, by Sandwich Theorem, that
n →∞ n
lim a n
n →∞

exists and is equal to 0.

Step 2: a = 0
Then a n = 0 for all positive integers n, so clearly
lim a n = 0 .
n →∞

Step 3: −1 < a < 1

Then for any positive integer n, we have
− a ≤ an ≤ a
n n

where 0 < a < 1 . Thus by Step 1, we have both

lim a = 0 = lim − a
n →∞
n

n →∞
( n
).
Page 24 of 28
 Mathematical Database

This proves, by Sandwich Theorem, that

lim a n
n →∞

exists and is equal to 0.

Example 7.2.
Prove that for any a > 0, we have
1

lim a = 1 .n
n →∞

Solution.
We divide the proof into two cases:
Case 1: a ≥ 1

Since in this case

1

an ≥ 1
for all positive integers n, we can define a non-negative sequence { xn } by
1

a = 1 + xn
n

Then for any positive integer n, we have, by binomial theorem, that

a = (1 + xn )n ≥ 1 + nxn ,

and thus
a −1
0 ≤ xn ≤ .
n
1
By Proposition 7.2, as n tends to infinity, the right hand side tends to 0. (Note we used lim =0
n →∞ n

here.) This proves, by Sandwich Theorem, that

lim xn
n →∞

exists and is equal to 0. By Proposition 7.2 again, we have that

1
n
lim a
n →∞

exists and is equal to 1 + 0 = 1. This completes our proof in this case.

Case 2: 0 < a < 1
Then

Page 25 of 28
 Mathematical Database
1
>1
a

so by Case 1, we have
1
⎛ 1 ⎞n
lim ⎜ ⎟ = 1 .
⎝ ⎠
n →∞ a

This says
1
−
lim a n
= 1,
n →∞

and from
1
1
an = 1
−
n
a

we have, by Proposition 7.2, that

1

lim a n
n →∞

1
exists and is equal to = 1.
1

Exercise
Throughout the exercise, suppose that f is a real-valued function and a is a limit point of the domain
of definition of f. Also assume L is a real number.

1. Determine whether each of the following limit exists or not. Find the limit, with justification, if
the limit exists; explain why the limit fails to exist otherwise.

( x − 1)
2
1 x−3
(a) lim (b) lim (c) lim
x →3 1 + x x →1 x −1 x →3 x2 − 9

x 2 + 3x ⎛1 1 ⎞ ⎛1 1 ⎞
(d) lim (e) lim ⎜ − ⎟ (f) lim ⎜ − ⎟
x →0 x ⎝
x →0 x x + x2 ⎠ ⎝
x →+∞ x x +1⎠

2. Show that the following sequences converge and find their limits. (You may assume that
1
lim = 0,
n →∞ n

so you may use the results in Examples 7.1 and 7.2 where appropriate.)

Page 26 of 28
 Mathematical Database

( −1)
n
2n n n2
(a) lim (b) lim (c) lim
n →∞ n + 3 n →∞ n2 − n + 2 n →∞ n !

n
3n 3n ⎛3⎞
(d) lim (Hint: For n > 3, 0 < ≤ 6 ⎜ ⎟ .)
n →∞ n ! n! ⎝4⎠

3. Let { xn } be a sequence of real numbers. Show that the following are equivalent:

(a) { xn } diverges

(b) For any real number l, there exists ε > 0 such that for any positive integer N, there exists a
positive integer n > N such that d ( xn , l ) ≥ ε .

(Make sure you can do the above negation properly even if you are not given the answer!)

4. Use Question 3 to show that {( −1) } and ⎩⎧⎨sin n2π ⎭⎫⎬ are divergent sequences.
n

5. Show that if f is a real-valued function defined on the set of all real numbers and if
lim f ( x ) = L ,
x →+∞

then for any sequence {xn } of real numbers that tends to infinity, we have
lim f ( xn ) = L .
n →∞

In particular we have
lim f (n) = L .
n →∞

6. Show that the converse to Question 5 holds, namely that: If f is a real-valued function defined
on the set of all real numbers and if
lim f ( xn ) = L
n →∞

for any sequence {xn } of real numbers that tends to infinity, then
lim f ( x ) = L .
x →+∞

7. Show that if
lim f ( x ) = L
x→a

Page 27 of 28
 Mathematical Database

then
lim f ( x − 1) = L .
x → a +1

Does the converse hold? Can you interpret this graphically?

8. Show that if
lim f ( x ) = L
x→a

then
⎛ x⎞
lim f ⎜ ⎟ = L .
x →2 a
⎝2⎠

Does the converse hold? Can you interpret this graphically? How far can you generalize the
above result?

9. Prove that if
lim xn = L
n →∞

then
lim xn +1 = L .
n →∞

(Hint: Consider { xn +1 } as a new sequence and argue by definition of limit.) Can you see any
connection between Questions 7, 8 and the above?

10. Let { xn } be a sequence of non-negative real numbers. Show that if { xn } is convergent, then

lim xn ≥ 0 .
n →∞

Show that it is NOT true that if { xn } is a convergent sequence of positive real numbers, then

lim xn > 0 .
n →∞

Can you generalize this to the setting of limit of functions?

Page 28 of 28