Chapter 5: Continuous Functions

5.1. Definitions and examples. Intuitively, a continuous function is one such that its graph is
an unbroken curve.

continuous

not continuous

Some reasons why a curve is broken:

Definition f is said to be continuous at a if the following conditions are satisfied:

: (i) f is defined in a neighborhood Vh (a) of a.
: (ii) lim f (x) exists.
xa
: (iii) lim f (x) = f (a).
xa

definition of continuity: f is continuous at a if > 0, > 0 such that

| f (x) f (a)| <

whenever |x a| < .

Definition If f is continuous at every point in a set S , then we say f is continuous on S .

Example Polynomial. Recall that if p(x) = c0 + c1 x + c2 x2 + + cn xn is a polynomial, then

for any a R,
lim p(x) = p(a).
xa

Thus p is continuous at every point a in R, that is, p is continuous on R.

1
Some specific examples: 3x2 + 4x + 5, 4x7 + x 3 and 2x + 3 are continuous on R.
2

Example Absolute-value function. Let f (x) = |x|.

f(x)=|x|

For all a R,
lim f (x) = lim|x| = |a| = f (a).
xa

xa

So f is continuous everywhere.

Example Square-root function. Let f (x) = x.

Then for all a > 0,

lim f (x) = lim x = a = f (a).

xa

xa

So f is continuous on (0, ).
Example sin x and cos x are continuous on R.
Proof: Use two facts (see section 8.4 of text):
xy
x+y
sin x sin y = 2 sin
cos
for all x, y R.
2
2
| sin x| |x| and | cos x| 1 for all x R.
Let a R. Then



xa
x + a 



| sin x sin a| = 2 sin
cos
2
2 

Theorem 5.1.1. (Sequential Criterion for Continuity) f is continuous at x = a if and only

if for every sequence (xn ) in the domain of f such that xn a, we have f (xn ) f (a).

Proof: This follows from the sequential criterion for limit of functions and the definition of
continuity. 

Definition If f is not continuous at a, then we say f is discontinuous at a .

Example Let f (x) = x/(x 1). Where is f continuous?

For any a , 1,
x
lim xa x
a
lim f (x) = lim
=
=
= f (a).
xa
xa x 1
lim xa (x 1) a 1
So f is continuous at every a , 1, i.e. f is continuous on the set R\{1}.
On the other hand,
f (1) is not defined.
1
lim
does not exist.
x1 x 1
So f is discontinuous at x = 1.
Example Rational functions. If f (x) = p(x)/q(x) where p and q are polynomials and
q(a) , 0, then
p(x)
p(x) lim xa p(x) p(a)
lim f (x) = lim
= lim
=
=
= f (a).
xa
xa q(x)
xa q(x)
lim xa q(x) q(a)
So f is continuous everywhere except at the zeros of q.

Example Let f (x) = [x].

For any n Z, lim[x] does not exist. So f is discontinuous at all the integral points.
xn
f is continuous everywhere else. (Why?)
Sometimes we can save a function which is discontinuous at a point:
If lim f (x) = L exists, but f (a) is not defined, then we can simply define f (a) = L. The
xa
resulting function will be continuous at a.
1
Example Let f (x) = x sin , x , 0.
x
f (0) is not defined, so f is discontinuous at x = 0.
However, by the Squeeze theorem,
lim x sin
x0

1
= 0.
x

So we define f (0) = 0, i.e. the new f is

x,0
x sin

x
f (x) =

0
x = 0.

The new f is continuous at x = 0, so it is continuous on R.

Some functions are hopeless!

If lim f (x) does not exist, then there is no way to save the function.
xa

Example Let f (x) = x/(x 1), x , 1.

Since lim f (x) does not exist, we cannot define f (1) in such a way that f is continuous at x = 1.
x1

Example A very bad function: Let

1 xQ

f (x) =

0 x < Q.
We have proved before that lim f (x) does not exist for any a R. So f is not continuous
xa
anywhere.
Exercise A slightly better function: Let

x xQ

f (x) =

0 x < Q.
Where is f continuous?

Example Thomaes function.

Let f : (0, 1) R be defined by

0 if x is irrational

f (x) =

p
1

if x = , p, q N and gcd(p, q) = 1.

q
q

So f (1/ 2) = 0, f (2/3) = 1/3, f (0.6) = f (6/10) = f (3/5) = 1/5.

Where is f continuous?
Solution. Claim: f is not continuous at all rational points.

In fact, if a Q and (xn ) is a irrational sequence in (0, 1) such that xn a, then f (xn ) = 0
0 , f (a).
Is f continuous at the irrational points?
Equivalently, let a be an irrational point in (0, 1). Is lim f (x) = f (a) = 0?
xa

Let > 0. We need to find > 0 such that f (x) < for all x (a , a + ).
Observation 1: The irrational xs do not cause any trouble.
Observation 2: f (p/q) is small if q is large.
Observation 3: There are only finitely many rational numbers in (0, 1) with small denominators, i.e. p/q with q 1/.
Observation 4: We can choose > 0 so small that the interval (a , a + ) will miss all the
rational numbers with small denominators.
Observation 5: is the we are looking for. Now convince yourself that all x (a , a + )
are such that f (x) < .
5.2. Combinations of continuous functions. .
Theorem 5.2.1. Suppose that f and g are continuous at x = a.
: (a) f g, f g and c f are also continuous at x = a, where c is a constant.
: (b) If g(a) , 0, then f /g is also continuous at x = a.

Proof: Follows from the limit theorem for functions. 

Example Let f (x) = tan x. Where is f continuous?

Solution: We have
sin x
.
f (x) =
cos x
sin x and cos x are continuous everywhere.

1
cos x = 0 if and only if x = (n + ) for some n Z.
2
1
Thus f is continuous on R\{(n + ) : n Z}.
2

Exercise Where is cot x continuous?

Definition Composite Functions.

Suppose that f : A R, g : B R and
f (A) B.
We define the composite function g f : A R by
(g f )(x) = g[ f (x)],

x A.

Theorem 5.2.2. Suppose that f and g are such that g f is defined. If f is continuous at a,
and g is continuous at f (a), then g f is continuous at a.
Proof: Use sequential criterion.
Suppose xn a.
Since f is continuous at a, f (xn ) f (a).
Since g is continuous at f (a), g[ f (xn )] g[ f (a)], i.e. (g f )(xn ) (g f )(a). So g f is
continuous at a. 

Theorem 5.2.3. Suppose that f : A R, g : B R and f (A) B, so that g f is defined.

If f is continuous on A, and g is continuous on B, then g f is continuous on A.

Example Let g(x) = |x|. If f : A R, then

(g f )(x) = g[ f (x)] = | f (x)| = | f |(x),

x A.

If f is continuous at a, then by Theorem 5.2.2, | f | is continuous at a.

If f is continuous on A, then by Theorem 5.2.3, | f | is continuous on A.

Some specific examples:
 3

 x 2x + 5 
: (a) f (x) = 
is continuous on R\{1}.
 x 1 
: (b) g(x) = | sin x| is continuous everywhere.

Exercise Is it true that

| f | continuous = f continuous?

Example Let g(x) =

x. If f : A R+ , i.e f (x) > 0 for each x A, then

(g f )(x) = g[ f (x)] =

f (x) =

f (x),
x A.
p
If f is continuous at a, then by Theorem 5.2.2, pf is continuous at a.
If f is continuous on A, then by Theorem 5.2.3, f is continuous on A.
Some special cases:

: (a) f (x) = x2 + x + 1 is continuous everywhere.

: (b) g(x) = sin x is continuous on (0, ).

Exercise Where is
f (x) =

x sin(1/x2 ) + cos(x3 + 5|x|)

.

(x 1)(x 2)

continuous?
5.3. Continuous functions on intervals. .
Definition If f is defined on the closed interval [a, b], then we say that f is continuous on
[a, b] if
: (i) f is continuous on (a, b) in the usual sense, ie. lim f (x) = f (c) for all c (a, b);
xc
: (ii) lim+ f (x) = f (a) and lim f (x) = f (b).
xa

xb

Definition A function f : A R is said to be bounded on A if there exists M > 0 such that

| f (x)| M,

x A.

So in this case, the set f (A) is bounded.

Definition A function f : A R is said to be unbounded on A if it is not bounded on A.

Example Is the function 1/x bounded on [2, )?

Solution: If x [2, ), then x 2, so that
 
 1  = 1 1 .
 x x 2
So can take M = 1/2.

Theorem 5.3.1. If f is continuous on [a, b], then f is bounded on [a, b].

Proof: Suppose f is not bounded. Then for each n N, xn [a, b] such that
| f (xn )| > n.
Since (xn ) is bounded, by the Bolzano-Weierstrass theorem, (xn ) has a convergent subsequence
(xnk ).
Let c = lim xnk . Since a xnk b for all k N, a c b.
k

Since f is continuous at c, f (xnk ) f (c).

On the other hand, | f (xnk )| > nk k , i.e ( f (xnk )) is unbounded. So ( f (xnk )) is divergent.
But this contradicts the fact f (xnk ) f (c).
So f is bounded on [a, b]. 

10

Definition Let f : A R.
: (i) We say f has an absolute maximum on A if there exists x0 A such that
f (x0 ) f (x),

x A.

So in this case, f (x0 ) = sup f (A).

x0 is call an absolute maximum point for f on A.
: (ii) We say f has an absolute minimum on A if there exists x1 A such that
f (x1 ) f (x),

x A.

So in this case, f (x1 ) = inf f (A).

x1 is call an absolute minimum point for f on A.
Example Consider the function f (x) = 1/x, x A.
: (i) If A = (0, ), then f (A) is not bounded above and inf f (A) = 0. In this case, f is not
bounded on A and it has no absolute maximum nor absolute minimum.
: (ii) If A = (1, 2), then
1
inf f (A) = .
2
So f is bounded on A but it has no absolute maximum nor absolute minimum.
sup f (A) = 1,

: (iii) If A = [1, 2], then

absolute maximum: f (1) = sup f (A) = 1,
1
absolute minimum: f (2) = inf f (A) = .
2

Extreme-value Theorem. If f is continuous on [a, b], then f has an absolute maximum and
an absolute minimum on [a, b].
Proof: By Theorem 5.3.1, f is bounded on [a, b]. Let M = sup f ([a, b]).
Suppose f (x) < M for all x [a, b]. Let
1
,
x [a, b].
M f (x)
Since M f (x) , 0 for all x [a, b], by Theorem 5.2.1, g is continuous on [a, b]. By Theorem
5.3.1 again, there exists L > 0 such that
1
g(x) =
< L,
x [a, b].
M f (x)
g(x) =

11

It follows that

1
,
x [a, b].
L
Thus M 1/L is an upper bound of f ([a, b]), and this contradicts M = sup f ([a, b]). So there
exists x0 [a, b] such that f (x0 ) = M.
f (x) < M

Similar arguments shows that there exists x1 [a, b] such that f (x1 ) = inf f ([a, b]). 
Remark See page 131 of the textbook for an alternative proof of the above theorem.
Lemma 5.3.2. Suppose that f is continuous on [a, b].
(a) If lim+ f (x) = f (a) < 0, then 1 > 0 such that
xa

f (x) < 0,

x [a, a + 1 ).

(b) If lim f (x) = f (b) > 0, then 2 > 0 such that

xb

f (x) > 0,

x (b 2 , b].

Location of Roots Theorem. If f is continuous on [a, b], f (a) < 0 < f (b), then there exists
a point c in (a, b) such that f (c) = 0.
Proof: Let A = {u [a, b] : f (x) < 0 x [a, u]}.
Since a A, A , .
Since A is bounded, c = sup A exists.
By the previous lemma, 1 , 2 > 0 such that
f (x) < 0,

x [a, a + 1 ),

f (x) > 0,
x (b 2 , b].
2
1
A and b
< A, so that
Hence a +
2
2
1
2
a<a+
c<b
< b.
2
2
In particular, a < c < b.
Claim: f (c) = 0.

12

Suppose f (c) , 0. Then either f (c) < 0 and f (c) > 0.

Case 1: f (c) < 0
Then > 0 such that
f (x) < 0,
Now since c = sup A, x0 A such that

x (c , c + ).
c < x0 c.

So
f (x) < 0,

x [a, x0 ]

()

Next choose x1 (c, c + ). Then since [x0 , x1 ] (c , c + ),

f (x) < 0,

x [x0 , x1 ].

This together with () give

f (x) < 0,
x [a, x1 ].
So x1 A. But x1 > c = sup A. So we obtain a contradiction if we assume f (c) < 0.
Case 2: f (c) > 0
Again there exists > 0 such that
f (x) > 0,

x (c , c + ).

Also, since c = sup A, there exists x0 A such that

c < x0 c.
Since x0 A, f (x0 ) < 0. But x0 (c , c + ), so we also have f (x0 ) > 0. So we obtain a
contradiction by assuming f (c) > 0.
Since both case 1 and case 2 cannot occur, we must have f (c) = 0 

Exercise Show that the equation

x3 x + 2 = 0
has a solution between 2 and 1.
Intermediate Value Theorem. If f is continuous on [a, b], and k is between f (a) and f (b),
then there exists a point c in (a, b) such that f (c) = k.
Proof: Assume that f (a) < k < f (b). Let g(x) = f (x) k. Then g is continuous on [a, b],
g(a) = f (a) k < 0 and g(b) = f (b) k > 0. By the Location of Roots Theorem, c (a, b)
such that
g(c) = f (c) k = 0.

13

So f (c) = k. 

Exercise Let f be a continuous function on [a, b] and suppose that f (a) , f (b). Prove that
there is a number c in (a, b) such that
f (c) =
Hint:

1
4
f (a) + f (b).
5
5

1 4
+ = 1.
5 5

Theorem 5.3.3. If f is continuous on [a, b], then

f ([a, b]) = [m, M],
where m = inf f ([a, b]) and M = sup f ([a, b]).
Proof: By the Extreme-value Theorem, there exist x0 and x1 in [a, b] such that f (x0 ) = m and
f (x1 ) = M.
Suppose that k [m, M]. Then
m = f (x0 ) k M = f (x1 ).
By the Intermediate Value Theorem, there exists a point c between x0 and x1 such that f (c) = k.
This shows that k f ([a, b]). Since k [m, M] is arbitrary, f ([a, b]) = [m, M]. 
Theorem 5.3.3 states that the image of a closed bounded interval under a continuous function
is also a closed bounded interval.
In general, if I is an interval and f is continuous on I, then f (I) is also an interval. However, I
and f (I) may not be of the same type.
Example Let f (x) = 1/(x2 + 1), I1 = (1, 1) and I2 = [0, ). Then f (I1 ) = (1/2, 1] and
f (I2 ) = (0, 1].
5.4. Monotone and inverse functions. .
Definition Let A R and f : A R.
(a) f is increasing on A if
x1 , x2 A and x1 x2 = f (x1 ) f (x2 ).
(b) f is strictly increasing on A if
x1 , x2 A and x1 < x2 = f (x1 ) < f (x2 ).

14

(c) f is decreasing on A if
x1 , x2 A and x1 x2 = f (x1 ) f (x2 ).
(d) f is strictly decreasing on A if
x1 , x2 A and x1 < x2 = f (x1 ) > f (x2 ).
(e) f is monotone if it is either increasing or decreasing.
(f) f is strictly monotone if it either strictly increasing or strictly decreasing.
It turns out that a monotone function defined on an interval always has one-sided limits.
Theorem 5.4.1. Let I R be an interval and f : I R be an increasing function. If c I
is not an end point of I, then lim f (x) and lim+ f (x) exit and they are given by
xc

xc

lim f (x) = sup{ f (x) : x I, x < c} and lim+ f (x) = inf{ f (x) : x I, x > c}.

xc

xc

Proof: Let S = { f (x) : x I, x < c}. If x I and x < c, then f (x) f (c). Thus f (c) is
an upper bound of S . By the Supremum property of R, L = sup S exists. We shall prove that
lim f (x) = L.
xc

Let > 0. Then since L is not an upper bound of S , there exists x I such that x < c
and
L < f (x ) L.
()
Let = c x . If c < x < c, then x > x , so that f (x) f (x ). This together with () gives
L < f (x ) f (x) L.
Hence
c < x < c = | f (x) L| < .
This proves lim f (x) = L. The proof for the other formula is similar. 
xc

Remark By the above theorem, if f is strictly increasing and is discontinuous at c, then

lim f (x) < lim+ f (x).

xc

xc

The difference
j f (c) = lim+ f (x) lim f (x)
xc

xc

is called the jump of f at c.

Let I R be an interval and f : I R be a strictly monotone function. Then f is injective.
Let J = f (I). Then we can define a function g : J R as follows: For each y f (I), there is a
unique x I such that f (x) = y. We set
g(y) = x.
In other words, g is the inverse function of f . It is denoted by f 1 .

15

Continuous Inverse Theorem. Let I R be an interval and f : I R be a strictly

monotone function. If f is continuous on I and J = f (I), then its inverse function f 1 : J R
is strictly monotone and continuous on J.
Proof: We shall assume that f is strictly increasing. The other case is similar.
Since f is continuous on I and I is an interval, J = f (I) is also an interval. Let f 1 : J R
be the inverse function of f .
Claim: f 1 is strictly increasing on J.
In fact, let y1 < y2 be two points in J. Then there exist x1 , x2 I such that f (x1 ) = y1 and
f (x2 ) = y2 . By the definition of f 1 , f 1 (y1 ) = x1 and f 1 (y2 ) = x2 . If x1 x2 , then since f is
strictly increasing, y1 = f (x1 ) f (x2 ) = y2 , which contradicts the fact that y1 < y2 . So we must
have x1 < x2 , that is, f 1 (y1 ) < f 1 (y2 ). This proves the claim.
Next we shall prove that f 1 is continuous on J. Assume that f 1 is discontinuous at c J
and c is not an end point of J (the case when c is an end point is left as an exercise). Then

lim f 1 (y) < lim+ f 1 (y).

yc

yc

Choose any point x0 such that x0 , f 1 (c) and

lim f 1 (y) < x0 < lim+ f 1 (y).

yc

Then x0 I. (Why?)

yc

16

We now let y0 J. If y0 < c, then

f 1 (y0 ) < lim f 1 (y)
yc

so that f 1 (y0 ) , x0 . On the other hand, if y0 > c, then

f 1 (y0 ) > lim+ f 1 (y),
yc

and again f 1 (y0 ) , x0 . We have also chosen x0 such that x0 , f 1 (c). Hence x0 , f 1 (y0 ) for
all y0 J, i.e. f (x0 ) < J. Since x0 I, this contradicts f (I) = J. 
The nth root function:
We shall use the Continuous Inverse Theorem to show that the nth root
n
function y y is continuous on [0, ). Let f : [0, ) R be given by
f (x) = xn ,

x [0, ).

If 0 < x1 < x2 , then

f (x1 ) = x1n < x2n = f (x2 ).
So f is strictly increasing on [0, ).
(To see this we use induction: If we assume x1k < x2k , then
x1k+1 = x1 x1k < x1 x2k < x2 x2k = x2k+1 .)

17

Moreover, it is continuous there and its range is also [0, ). By the Continuous Inverse Theorem, its inverse function f 1 is strictly increasing and continuous on [0, ). If y [0, ) and
x = f 1 (y), then
y = ( f f 1 )(y) = f [ f 1 (y)] = f (x) = xn .
Thus x is the nth root of y. So we can write

f 1 (y) = y1/n
or
f 1 (y) = n y.

18

5.5. Uniform continuity. Consider the functions f : R R, f (x) = 2x and g : (0, ) R,

g(x) = 1/x. Both function are continuous on their domains.
Let > 0.
(i) If a R and = /2, then
|x a| < = | f (x) f (a)| < .
Note that depends only on , and is independent of the point a.
(ii) If a (0, ), then



1 1  |x a|

|g(x) g(a)| =   =
.
x a
a|x|
If |x a| < a/2, then x > a/2 and
|x a|
2
|g(x) g(a)| =
< 2 |x a|.
a|x|
a
!
a 1
If we take = min , a2 , then
2 2
|x a| < = |g(x) g(a)| < .

Note that depends on both the point a and .

We say that f is uniformly continuous on R, but g is not uniformly continuous on (0, ).
Definition Let I R be an interval and f : I R. We say that f is uniformly continuous on
I if for every > 0, there exists > 0 such that
x, a I, |x a| < = | f (x) f (a)| < .
(This says that for a given > 0, a can be chosen such that it works for all the points in I.)
Some important facts about uniform continuity. See Section 5.4 of the textbook for their
proofs.
(1) If f is continuous on [a, b], then it is uniformly continuous on [a, b].
(2) If I is an interval and f : I R satisfies the Lipschitz condition on I, that is, there is a
K > 0 such that
| f (x) f (a)| K|x a|,

x, a I,

then f is uniformly continuous on I.

(3) If f : I R is uniformly continuous on I and (xn ) is a Cauchy sequence in I, then
( f (xn )) is a Cauchy sequence in R.
(4) A function f : (a, b) R is uniformly continuous on (a, b) if and only if f (a) and f (b)
can be defined so that the extended function is continuous on [a, b].