Introduction to Metric and Topological Spaces

by Wilson Sutherland
Unofficial Solutions Manual

M OHAMMAD E HTISHAM A KHTAR

I MPERIAL C OLLEGE L ONDON
http://akhtarmath.wordpress.com
Dedicated to my Parents
ii

Preface
This is an ongoing Solutions Manual for Introduction to Metric and Topological Spaces by Wilson
Sutherland [1]. The main reason for taking up such a project is to have an electronic backup
of my own handwritten solutions.
Mathematics cannot be done without actually doing it. However at the undergraduate
level many students are put off attempting problems unless they have access to written so-
lutions. Thus I am making my work publicly available in the hope that it will encourage
undergraduates (or even dedicated high school students) to attempt the exercises and gain
confidence in their own problem-solving ability.
I am aware that questions from textbooks are often set as assessed homework for stu-
dents. Thus in making available these solutions there arises the danger of plagiarism. In
order to address this issue I have attempted to write the solutions in a manner which con-
veys the general idea, but leaves it to the reader to fill in the details.
At the time of writing this work is far from complete. While I will do my best to add
additional solutions whenever possible, I can not guarantee that any one solution will be
available at a given time. Updates will be made whenever I am free to do so.
I should point out that my solutions are not the only ways to tackle the questions. It is
possible that many ‘better’ solutions exist for any given problem. Additionally my work
has not been peer reviewed, so it is not guaranteed to be free of errors. Anyone using these
solutions does so at their own risk.
I also wish to emphasize that this is an unofficial work, in that it has nothing to do with
the original author or publisher. However, in respect of their copyright, I have chosen to
omit statements of all the questions. Indeed it should be quite impossible for one to read this
work without having a copy of the book [1] present.
I hope that the reader will find this work useful and wish him the best of luck in his
Mathematical studies.

M OHAMMAD E HTISHAM A KHTAR

I MPERIAL C OLLEGE L ONDON
Project started on 15 March 2008

The end of a solution is indicated by . Any reference such as ‘Proposition 2.3.13’, ‘Defi-
nition 3.8.1’, ‘Question 10.3.16’ refers to the relevant numbered item in Sutherland’s book
[1]. This work has been prepared using LATEX.

The latest version of this file can be found at : http://akhtarmath.wordpress.com/

Cite this file as follows :

Akhtar, M.E. Unofficial Solutions Manual for Introduction to Meric and Topological Spaces by Wil-
son Sutherland. Online book available at : http://akhtarmath.wordpress.com/
Contents

Preface ii

Quick Reference 2

1 Review of some real analysis 3

2 Continuity generalized: metric spaces 5

4 The Hausdorff condition 7

5 Compact spaces 9

7 Compactness again: convergence in metric spaces 11

Bibliography 13

Conditions of Use
This work is Copyright c 2008 Mohammad Ehtisham Akhtar. You are permitted to use and
distribute this work freely for non-profit purposes as long as you do not modify it in any
way. If you use this work or quote parts of it then you must give a proper citation. The use
and distribution of this work for commercial purposes is prohibited.
2 Contents

Quick Reference

Chapter 1: Review of some real analysis

1.5.1 , 1.5.2 ,

Chapter 2: Continuity generalized: metric spaces

2.6.1 , 2.6.2 , 2.6.3 , 2.6.4 , 2.6.5 , 2.6.10 , 2.6.12 , 2.6.13 ,

Chapter 4: The Hausdorff condition

4.3.1 , 4.3.4 ,

Chapter 5: Compact spaces

5.10.1 , 5.10.2 , 5.10.3 , 5.10.11 , 5.10.14 , 5.10.18 , 5.10.19 ,

Chapter 7: Compactness again: convergence in metric spaces

7.3.1 , 7.3.2 ,
Chapter 1
Review of some real analysis

1.5.1) We assume that A and B are nonempty. Since B is bounded above, sup B exists and is
finite. Since A ⊆ B : a ∈ A ⇒ a ∈ B ⇒ a ≤ sup B. So A is bounded above and sup B is an
upper bound for A. Therefore, sup A exists and sup A ≤ sup B.
1.5.2) We assume that A and B are nonempty. Since A, B ⊂ R are bounded above, sup A and
sup B both exist and are finite. In particular Max {sup A, sup B} exists and is finite. Now
if x ∈ A ∪ B then either x ∈ A ⇒ x ≤ sup A ≤ Max {sup A, sup B} or x ∈ B ⇒ x ≤
sup B ≤ Max {sup A, sup B}. In either case we see that A ∪ B is bounded above and that
Max {sup A, sup B} is an upper bound for A ∪ B. Let U be any upper bound for A ∪ B.
Then sup A ∪ B ≤ U . Since A ⊆ A ∪ B ⊆ R it follows from Question 1.5.1 that sup A ≤
sup A ∪ B ≤ U . Similarly sup B ≤ U . Thus Max {sup A, sup B} ≤ U . We conclude that
Max {sup A, sup B} = sup A ∪ B.
Chapter 2
Continuity generalized: metric spaces

2.6.1) If there exists p ∈ Br (x) ∩ Br (y) then d(p, x) < r and d(p, y) < r so that 2r = d(x, y) ≤
d(p, x) + d(p, y) < 2r, which is a contradiction.

2.6.2) |d(x, z) − d(y, z)| ≤ d(x, y) ⇐⇒ −d(x, y) ≤ d(x, z) − d(y, z) and d(x, z) − d(y, z) ≤
d(x, y) ⇐⇒ d(y, z) ≤ d(x, y)+d(x, z) and d(x, z) ≤ d(x, y)+d(y, z). The last two inequalities
are true by the triangle inequality so the result follows.

2.6.3) Proceed exactly as in Question 2.6.2. d(x, z) + d(y, t) ≥ |d(x, y) − d(z, t)| ⇐⇒ d(x, z) +
d(y, t) ≥ d(x, y) − d(z, t) and −d(x, z) − d(y, t) ≤ d(x, y) − d(z, t) ⇐⇒ d(x, z) + d(y, t) +
d(z, t) ≥ d(x, y) and d(z, t) ≤ d(x, y) + d(x, z) + d(y, t). The last two inequalities are true by
repeated application of the triangle inequality, so the result follows.

2.6.4) Let X := {d(s, t) | s, t ∈ C} and Y := {d(p, q) | p, q ∈ B}. Then diam B = sup Y and
diam C = sup X. Also C ⊆ B, which implies that X ⊆ Y (⊆ R). A subset of a metric space
is bounded if and only if its diameter is finite. Since B is bounded, diam B = sup Y is finite.
In particular, Y is bounded above and X ⊆ Y ⊆ R, so it follows1 that sup X ≤ sup Y < ∞.
The fact that diam C = sup X < ∞ tells us that C is bounded, and the result sup X ≤ sup Y
is equivalent to saying that diam C ≤ diam B.

2.6.5) Let A := {d(x, y) | x, y ∈ B ∪ C} so that sup A = diam (B ∪ C). Note that A = S ∪T ∪V ,

where S := {d(m, n) | m, n ∈ B}, T := {d(p, q) | p, q ∈ C}, V := {d(h, k) | h ∈ B\C, k ∈ C\B}.
Furthermore, diam B = sup S, diam C = sup T and diam B, diam C ≥ 0. Select any a ∈ A.
There are three possible cases. If a ∈ S then a ≤ sup S = diam B ≤ diam B + diam C.
Similarly, if a ∈ T then a ≤ diam B + diam C. Finally, if a ∈ V then a = d(h, k) for some
h ∈ B\C and k ∈ C\B. Choose2 any x ∈ B ∩ C. Then a = d(h, k) ≤ d(h, x) + d(x, k) with
d(h, x) ∈ S and d(x, k) ∈ T . It follows once again that a ≤ sup S + sup T = diam B + diam C.
In all cases we see that diam B + diam C is an upper bound for A so that diam (B ∪ C) =
sup A ≤ diam B + diam C as required.

2.6.10) Let S be a subset of a metric space. Any open ball in a metric space is an open set in
that space. The union of any family of open sets in a metric space is also open. Thus if S is a
union of open balls then S must be open. Conversely suppose that S is open. Then for any
S
x ∈ S there exists a real number r(x) > 0 such that Br(x) (x) ⊆ S. In fact S = x∈S Br(x) (x)
i.e. S is a union of open balls.

1
Using Question 1.5.1
2
We can do this because B ∩ C 6= ∅.
6 2. Continuity generalized: metric spaces

2.6.12) No. To see this let (M, d) be any metric space with at least two distinct points which
we will call x and y. Since x 6= y, d(x, y) 6= 0. Say d(x, y) = 2r > 0. We claim that neither
of the open sets Br (x), Br (y) ⊆ M are contained in {∅, M }. The fact that x ∈ Br (x) and y ∈
Br (y) guarantees that Br (x), Br (y) 6∈ {∅}. Also3 Br (x) ∩ Br (y) = ∅ so that Br (x), Br (y) 6∈
{M }. Therefore if M is any metric space containing at least two distinct points then M
contains at least two open sets other than ∅ and M .
2.6.13) Set ε = f (a) > 0. Since f : M → R is continuous at a ∈ M , there exists a δ > 0 such
that x ∈ Bδ (a) ⇒ |f (x) − f (a)| < ε = f (a) ⇒ f (x) > 0.

3
Using Question 2.6.1
Chapter 4
The Hausdorff condition

4.3.1) As in Example 3.1.7 [1, p.47], consider R with the Zariski topology. Suppose the topol-
ogy is Hausdorff and select x, y ∈ R with x 6= y. Then there exist Zariski open sets U, V such
that x ∈ U , y ∈ V and U ∩ V = ∅. But then U c ∪ V c = R. This is a contradiction since both
U c and V c are finite while R is infinite. So the Zariski Topology is not Hausdorff.
4.3.4) (a) Let S be the intersection of all the open subsets of T that contain x. Then {x} ⊆ S.
Suppose S 6⊆ {x}. Then there exists y ∈ S is such that y 6∈ {x} i.e. y 6= x. Since T is
Hausdorff, there exist open sets A, B such that x ∈ A, y ∈ B and A ∩ B = ∅. In particular,
y 6∈ A. But y ∈ S ⊆ A. Contradiction. So S ⊆ {x}. We conclude that {x} = S.
(b) We need to give an example of a non-Hausdorff space W in which the following holds
true : For any x ∈ W , the intersection of all open subsets of W that contain x is equal to {x}.
Consider R with the Zariski topology. This is a non-Hausdorff space1 . Now select any x ∈ R.
Let S be the intersection of all the Zariski open subsets of R that contain x. Then {x} ⊆ S.
Suppose that S 6⊆ {x}. Then there exists y ∈ S such that y 6∈ {x}. Now y 6= x belongs to
every Zariski open subset of R that contains x. It follows that y ∈ R\{y}. Contradiction.
Therefore, S ⊆ {x}. It follows that {x} = S.

1
By Question 4.3.1
Chapter 5
Compact spaces

5.10.1) Fix any n ∈ N and let X := {x1 , . . . , xn } be a topological space. For some indexing
set I let {Ci }i∈I be an open cover of X. If k ∈ {1, . . . , n} then xk ∈ X and we know that
S
X = i∈I Ci . So there exists Cik ∈ {Ci }i∈I such that xk ∈ Cik . It follows that {Ci1 , . . . , Cin }
is a finite subcover of {Ci }i∈I . Therefore X is compact.
5.10.2) Let X be a discrete space that is compact. Suppose that X is infinite. Then {{p}}p∈X
is an open cover of X with no finite subcover. This contradicts the assumption that X is
compact. Therefore X must be finite.
S
5.10.3) Let C be a collection of open sets in T that covers H ∪ K so that H ∪ K ⊆ V ∈C V .
Since H ⊆ H ∪ K, C is also an open cover for H. Since H is compact, there exists a finite
sub-collection {H1 , . . . , Hn } ⊆ C such that H ⊆ ni=1 Hi . Similarly, there exists a finite sub-
S

collection {K1 , . . . , Kl } ⊆ C such that K ⊆ lj=1 Kj . Then {H1 , . . . , Hn , K1 , . . . , Kl } ⊆ C is a

S

finite subcover of C for H ∪ K.

5.10.11) A subset of R is compact if and only if it is closed and bounded1 . If A ⊆ R is not
bounded then the continuous function f : A → R such that f (a) = a is not bounded. On the
other hand, if A is not closed then A 6= Cl(A). So there must exist at least one limit point l of
A such that l 6∈ A. Define g : A → R by g(a) = 1/(l − a). Then the function g is continuous
but not unbounded.
5.10.14) Consider the function g : T → R defined by g(x) = d(f (x), x). First we we will
show that g is continuous. Select any a ∈ T and fix E > 0. We require a D > 0 such that
d(x, a) < D ⇒ |g(x) − g(a)| = |d(f (x), x) − d(f (a), a)| < E. Now2 |d(f (x), x) − d(f (a), a)| ≤
d(f (x), f (a)) + d(x, a). Furthermore, since f : T → T is continuous at a ∈ T , there exists
δ > 0 such that d(y, a) < δ ⇒ d(f (y), f (a)) < E/2. Set D = Min {δ, E/2}. Then d(x, a) <
D ⇒ |g(x) − g(a)| = |d(f (x), x) − d(f (a), a)| ≤ d(f (x), f (a)) + d(x, a) < (E/2) + (E/2) = E.
This establishes the continuity of g. Next observe that g(T ) = {d(f (x), x) | x ∈ T }. Since
g : T → R is continuous, and T is compact, g(T ) is bounded and g attains its bounds on
T . In particular, there exists t ∈ T such that g(t) = inf g(T ). Additionally, since f (t) 6= t,
g(t) = d(f (t), t) > 0. Set ε = g(t) > 0. Then ε = inf g(T ) ≤ g(x) = d(f (x), x) for all x ∈ T .
5.10.18) Let di denote he metric on Mi (i = 1, 2, 3). Fix ε > 0. There exists a δg > 0 such
that for all a, b ∈ M2 : d2 (a, b) < δg ⇒ d3 (g(a), g(b)) < ε. Also there exists δf > 0 such
that for all x, y ∈ M1 : d1 (x, y) < δf ⇒ d2 (f (x), f (y)) < δg . Set δ = δf . Then for all
x, y ∈ M1 : d1 (x, y) < δ = δf ⇒ d2 (f (x), f (y)) < δg ⇒ d3 ((g ◦ f )(x), (g ◦ f )(y)) < ε.
1
A special case of the corresponding statement for Rn .
2
Using Question 2.6.3
10 5. Compact spaces

Therefore, g ◦ f : M1 → M3 is uniformly continuous.

5.10.19) Let (M1 , d1 ), (M2 , d2 ) be metric spaces and f : M1 → M2 be a given function. f is
not uniformly continuous on M1 if and only if there exists an ε > 0 such that for all δ > 0,
there exist x, y ∈ M1 satisfying both d1 (x, y) < δ and d2 (f (x), f (y)) ≥ ε. In order to show
that f : (0, 1) → R defined by f (x) = 1/x is not uniformly continuous on (0, 1), set ε = 1 and
choose any δ > 0. Let m, n be positive integers such that x := 1/m < δ/2 and y := 1/n < δ/2.
Then x, y ∈ (0, 1) are such that : |x − y| ≤ |x| + |y| = (1/m) + (1/n) < (δ/2) + (δ/2) = δ and
|(1/x) − (1/y)| = |m − n| ≥ 1 = ε. So f is not uniformly continuous on (0, 1).
Chapter 7
Compactness again: convergence in metric spaces

7.3.1) Let (an ) be a convergent sequence in the metric space (X, d) with limit1 l ∈ X. Fix
ε > 0. There exists N (ε) ∈ N such that n ≥ N (ε) ⇒ d(an , l) < 2ε . Therefore m, n ≥ N (ε) ⇒
d(am , an ) ≤ d(am , l) + d(an , l) < 2ε + 2ε = ε which implies that (an ) is a Cauchy sequence.
7.3.2) Let (zn ) be a Cauchy sequence in the metric space (X, d). Fix ε = 1 > 0. Then there
exists N such that m, n ≥ N ⇒ d(zm , zn ) < 1. Fix m = N +1 > N . Then n ≥ N ⇒ d(z1 , zn ) ≤
d(z1 , zm ) + d(zm , zn ) < d(z1 , zm ) + 1. Set M = max{d(z1 , z1 ), d(z1 , z2 ), . . . , d(z1 , zN −1 )}. Then
d(z1 , zn ) ≤ max{M, d(z1 , zm ) + 1} for all n. So the Cauchy sequence (zn ) is bounded.

1
Since every metric space is Hausdorff, the limit is unique.
Bibliography

[1] Sutherland, W. Introduction to Metric and Topological Spaces, 1975. Oxford Science Publications.