Theorems on Continuity

Aditya Ghosh

Exercises/Problems

1. Let f : I → R be a continuous function. If I is a closed bounded interval, then we

know that f must be bounded. Show that the result fails to hold in each of the
following cases: (a) I is bounded interval but not closed. (b) I is not bounded.
(c) I is not an interval.

2. Let f : [a, b] → R be a continuous function with the property that for every
x ∈ [a, b], there exists y ∈ [a, b] such that |f (y)| ≤ 21 |f (x)|. Show that there exists
c ∈ [a, b] such that f (c) = 0.

3. Suppose that f, g : [a, b] → R are continuous and such that f (a) < g(a) and
f (b) > g(b). Show that there exists c ∈ (a, b) such that f (c) = g(c).

4. Suppose that f : [a, b] → R is continuous. Let x1 , x2 , . . . , xn be any n points in

(a, b). Show that there exists x0 ∈ (a, b) such that

1
f (x0 ) = (f (x1 ) + f (x2 ) + · · · + f (xn )) .
n

5. Prove that the equation (1 − x) cos x = sin x has at least one solution in (0, 1).

6. If f : [0, 1] → [0, 1] is a continuous function, then show that there exists c ∈ [0, 1]
such that f (c) = c2 .

7. Suppose that f : [0, 1] → [0, 1] is a continuous function with f (0) = 0 and

f (1) = 1. Show that there exists c ∈ (0, 1) such that c2 + (f (c))2 = 1.

8. Suppose that f : [0, 2] → R is continuous and f (0) = f (2). Prove that there exists
a, b ∈ [0, 2] such that b − a = 1 and f (b) = f (a).

9. An athlete runs a distance of 6 km in 30 minutes. Prove that somewhere during

the run he covered a distance of 1 km in exactly 5 minutes.

10. Consider f : R\{0} → R defined by f (x) = 1/x. Is f continuous? Note that

f (−1) < 0 < f (1) and f (x) 6= 0 for any x. Does it contradict the intermediate
value theorem?

1
11. Let f : R → R be continuous and periodic with period T > 0. Prove that there
exists x0 such that f (x0 + T /2) = f (x0 ).

12. Let f : R → R be continuous and periodic with period T > 0. Prove that there
exists x0 such that f (x0 + π) = f (x0 ). Convince yourself that the same result
holds even if we replace π with any other number.

13. Suppose that f and g have the intermediate value property on some closed
bounded interval I. Is it necessary that f + g also has the intermediate value
property on I?

14. Suppose that f : [a, b] → R is a continuous function that takes rational values
only. What can you say about f ?

15. Let f : R → R be a continuous function that satisfies f (q + 1/n) = f (q) for every
q ∈ Q and for every n ∈ N. Show that f must be a constant function.

16. Suppose f : R → R is continuous and injective. Show that f must be strictly

monotonic.

17. Let f : [a, b] → [c, d] be a strictly increasing function, where c = f (a), d = f (b).
Is it necessary that f −1 exists?

18. Let f : [a, b] → [c, d] be a continuous and strictly increasing function, where
c = f (a), d = f (b). Show that f −1 exists and is strictly increasing on the interval
[c, d]. Furthermore, show that f −1 is continuous on [c, d].

19. Suppose x1 = tan−1 2 > x2 > x3 > · · · are positive real numbers, satisfying

sin(xn+1 − xn ) + 2−(n+1) sin xn sin xn+1 = 0 for every n ≥ 1.

π
Find an expression for cot xn . Hence show that lim xn = .
n→∞ 4
20. Let f (x) = an xn + an−1 xn−1 + · · · + a1 x + a0 be a polynomial with real coefficients,
where n > 0 is even. If an > 0 and a0 < 0, then show that the equation f (x) = 0
has at least two real roots.

21. Show that there exists a set of 100 consecutive integers of which exactly 19 are
primes. (Hint: Do you know that there is a set of 100 consecutive integers which
does not contain any prime?)

2
 Solutions

1. (a) Take I = (0, 1] and f (x) = 1/x. (b) Take I = (0, ∞) and f (x) = x. (c) Take
I = [0, π/2) ∪ (π/2, π] and f (x) = tan x.

2. The function |f | is continuous on [a, b], hence attains a minimum value, say at c.
But the given property with x = c produces some y such that |f (y)| ≤ 12 |f (c)|. Since
|f (c)| ≤ |f (y)| ≤ 21 |f (c)|, it follows that f (c) = 0.
Alternate solution: Start with some x1 . Let x2 be the number such that |f (x2 )| ≤
1
2
|f (x1 )|. Next, x3 be the number such that |f (x3 )| ≤ 12 |f (x2 )|, and so on. Thus, we
get a sequence xn such that |f (xn+1 )| ≤ 21 |f (xn )| holds for every n ≥ 1. It follows
that f (xn ) → 0 as n → ∞. Now, xn might not converge; but since it is bounded it
has a convergent subsequence xnk . Suppose xnk → c as k → ∞. Then f (xnk ) → f (c)
as k → ∞. But earlier we had f (xnk ) → 0. Therefore, f (c) = 0.

3. Define h(x) = f (x) − g(x). Since h(a) < 0, h(b) > 0 and h is continuous on [a, b],
the conclusion follows.

4. Let m = min{f (x1 ), f (x2 ), . . . , f (xn )} and M = max{f (x1 ), f (x2 ), . . . , f (xn )}. Clearly,
m = f (xi ) and M = f (xj ) for some 1 ≤ i, j ≤ n. Now, the average y = (f (x1 ) +
f (x2 ) + · · · + f (xn ))/n lies between m = f (xi ) and M = f (xj ). Therefore, there
exists x0 between xi and xj such that f (x0 ) = y.

5. Consider f (x) = (1 − x) cos x − sin x. Observe that f is continuous, f (0) > 0 and
f (1) < 0.

6. Consider g(x) = f (x) − x2 . Observe that g is continuous, g(0) ≥ 0 and g(1) ≤ 0.

7. Consider g(x) = x2 + (f (x))2 − 1.

8. Consider g(x) = f (x + 1) − f (x). Observe that g is continuous, g(0) and g(1) are
either zero, or have opposite signs. Thus, we get c ∈ [0, 1] such that g(c) = 0 =⇒
f (c + 1) = f (c). Set b = c + 1 and a = c.

9. Let x be the distance (in km) along the path. For x ∈ [0, 5], let f (x) denote the
time elapsed (in minutes) for running from the point x to x + 1. The function f is
continuous, and f (0) + f (1) + · · · + f (5) = 30. Now use problem 4.

3
10. First, f is continuous on its domain. Second, it does not contradict the intermediate
value theorem. Because in that theorem we assume that domain of f is an interval.

11. Consider g(x) = f (x + T /2) − f (x). Observe that g is continuous, g(0) and g(T /2)
are either zero, or have opposite signs. Thus, we get x0 ∈ [0, T /2] such that g(x0 ) =
0 =⇒ f (x0 + T /2) = f (x0 ).

12. Since f : [0, T ] → R is continuous, there exists a, b ∈ [0, T ] such that f (a) ≤ f (x) ≤
f (b) holds for every x ∈ [0, T ]. But f is periodic with period T > 0. Hence it follows
that f (a) ≤ f (x) ≤ f (b) holds for every x ∈ R. Then we get f (a) ≤ f (a + π)
and f (b + π) ≤ f (b). Therefore for g(x) = f (x + π) − f (x), we have g(a) ≥ 0 and
g(b) ≤ 0. The rest follows from the continuity of g.

13. No, f +g need not have IVP. Here is a counter-example: Take I = [−1, 1] and define
 
 sin(1/x) if x 6= 0  − sin(1/x) if x 6= 0
f (x) = , g(x) = .
 0 if x = 0  1 if x = 0

Show that f, g have IVP but f + g does not have IVP.

14. We can show that f must be a constant function. Because if f takes at least two
distinct values, say a < b, then the function has to take each value between a and b.
But there is an irrational number between a and b, which can’t belong to the range
f.

15. Take a rational number r = m/n. If m > 0, then f ( m

n
) = f ( m−1
n
) = · · · = f (0).
If m < 0, then f ( m
n
) = f ( m+1
n
) = · · · = f (0). Thus, for every r ∈ Q, we have
f (r) = f (0). For any irrational x, take a sequence of rationals converging to x and
use the continuity of f.

16. Fix any two numbers a, b, say a < b. Assume w.l.o.g. that f (a) < f (b). We shall
show that for any x, y ∈ R, x < y ⇐⇒ f (x) < f (y). Take any x, y. Pick M > 0
sufficiently large such that a, b, x, y ∈ [−M, M ]. Consider f on the interval [−M, M ].
Since f is continuous and injective, f must be monotone on this interval. Since
−M ≤ a < b ≤ M and f (a) < f (b), therefore f must be increasing on [−M, M ].
Hence, x < y ⇐⇒ f (x) < f (y).

4
 
 x if 0 ≤ x < 1
17. No. Consider f : [0, 2] → [0, 3] defined by f (x) = . How-
 x+1 if 1 ≤ x ≤ 2
ever, the result holds if f is given to be continuous.

18. Take any k ∈ (c, d). Since f (a) < k < f (b) and f is continuous on [a, b], there exists
c ∈ (a, b) such that f (c) = k. And since f is one-one, this c will be unique. Thus,
f −1 exists and it is easy to show that it is strictly increasing. To show that f −1 is
continuous, first take y0 = f (x0 ). Our goal is to show that f −1 is continuous at y0 .
Let me show the case y0 ∈ (c, d) and leave the boundary cases for you. Fix an ε > 0,
small enough so that x0 − ε, x0 + ε are within [a, b]. Since f is increasing, we have
f (x0 − ε) < f (x) < f (x0 + ε) for x0 − ε < x < x0 + ε. Take δ to be the minimum
of the two positive numbers f (x0 + ε) − f (x0 ) and f (x0 ) − f (x0 − ε). Note that for
y ∈ (y0 − δ, y0 + δ), we have f (x0 − ε) ≤ f (x0 ) − δ < y < f (x0 ) + δ ≤ f (x0 + ε).
(Draw a picture here!) Now, using the fact that f −1 is increasing, we get x0 − ε <
f −1 (y) < x0 + ε. Thus, we have shown that for any (small enough) ε > 0, there
exists δ > 0 such that |f −1 (y) − f −1 (y0 )| < ε holds whenever |y − y0 | < δ, which
lets us to conclude that f −1 (y) is continuous at y = y0 . (The boundary cases are
similar: you just need to use one side instead of two.)

Remark. Note that if we exclude the boundary points, then the above argument
also applies when c = −∞ or d = ∞. Here c = f (a) should be taken in the sense
that c = limx→a+ f (x) (of course we have not discussed such limits yet, we will
do that soon!), and similarly d = limx→b− f (x). For example, f (x) = tan(x) is a
strictly increasing and continuous function from (−π/2, π/2) to (−∞, ∞), and we
know that limx→ π − tan(x) = ∞ and limx→ −π + tan(x) = −∞. Therefore, its inverse
2 2
−1
g(y) = tan (y) is continuous on (−∞, ∞).

Corollary. Suppose that f : I → J is bijective and continuous, where I, J ⊆ R are

intervals. Then f −1 is also continuous on its domain, J.

(Since f is continuous and injective, it must be strictly monotone. Now consider two
cases based on whether f is increasing or decreasing, and apply the above result.)

Another remark. If the function is given to be continuous at just one point, then
the version of the above result that we need is as follows: Suppose f is continuous at

5
 x0 and strictly monotone in an open neighborhood of x0 . Then we can say that f −1
will be continuous at y0 = f (x0 ). (Do you see that the above proof goes smoothly
even in this setup?)

19. First note that sin xn 6= 0 for each n ≥ 1. Divide both sides by sin xn sin xn+1 and
1
show that cot xn+1 − cot xn = 2n+1
holds for every n ≥ 1. Using this relation, along
with the given fact that cot x1 = 1/2, deduce that for every n ≥ 1,

1 1 1 1 1 1 1
cot xn = cot xn−1 + n
= cot xn−2 + n−1 + n = · · · = + 2 + · · · + n = 1 − n
2 2 2 2 2 2 2

Hence, cot xn = yn → 1 as n → ∞. Now, from the last problem we know that cot−1
is continuous at y = 1 (in fact, it is continuous on R). Therefore, from yn → 1, we
can say that xn = cot−1 (yn ) → cot−1 (1) = π/4.

Remark. Do you see why cot−1 is continuous on R? To apply the result in the last
problem, you need to take (0, π) as the domain of cot(x), not (−π/2, π/2)\{0}.

20. We proceed similar to what we did for showing that every polynomial of odd degree
has at least one real root. First we observe/recall that

f (x) f (x)
lim n
= an = lim .
x→∞ x x→−∞ xn

Since an > 0 and n > 0 is even, we can say that for some sufficiently large a > 0,
the numbers P (−a) and P (a) are both positive. And we have P (0) < 0. Invoking
the intermediate value property, we conclude that the polynomial has at least one
zero in each of the intervals (−a, 0) and (0, a).

21. Observe that the set S = {1, 2, . . . , 100} contains 25 primes. And the set T =
{101! + 2, 101! + 3, . . . , 101! + 101} does not contain any prime. Think of shifting
the block of 100 numbers gradually so that we reach T from S. In each step, we
include a new number and delete an old one. The number of primes in the block
either increases by 1, or decreases by 1, or remains unchanged. Since it decreases to
0 from 25, there must be some stage when the block has exactly 19 primes. (Note
that this argument is a discrete analog of the intermediate value property.)