General Topology

A Solution Manual for Willard (2004)

Jianfei Shen
School of Economics, The University of New South Wales
Sydney, Australia October 15, 2011
Preface
Sydney, Jianfei Shen
October 15, 2011
v
Acknowledgements
vii
Contents
Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
1 Set Theory and Metric Spaces. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Set Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Metric Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2 Topological Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.1 Fundamental Concepts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
2.2 Neighborhoods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
2.3 Bases and Subbases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3 New Spaces from Old. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.1 Subspaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.2 Continuous Functions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.3 Product Spaces, Weak Topologies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.4 Quotient Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28
4 Convergence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.1 Inadequacy of Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
4.2 Nets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
4.3 Filters . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 34
5 Separation and Countability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.1 The Separation Axioms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37
5.2 Regularity and Complete Regularity . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
5.3 Normal Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
5.4 Countability Properties . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
6 Compactness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
6.1 Compact Spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45
ix
Acronyms
R the set of real numbers
I 0. 1|
P R Q
xi
1
SET THEORY AND METRIC SPACES
1.1 Set Theory
1A. Russells Paradox
I Exercise 1. The phenomenon to be presented here was rst exhibited by
Russell in 1901, and consequently is known as Russells Paradox.
Suppose we allow as sets things for which . Let P be the set of all
sets. Then P can be divided into two nonempty subsets, P
1
=

P :
_
and P
2
= { P : ]. Show that this results in the contradiction: P
1

P
1
== P
1
P
1
. Does our (naive) restriction on sets given in 1.1 eliminate the
contradiction?
Proof. If P
1
P
1
, then P
1
P
2
, i.e., P
1
P
1
. But if P
1
P
1
, then P
1
P
1
. A
contradiction. |L
1B. De Morgans laws and the distributive laws
I Exercise 2. a.
__
22/
T
2
_
=
_
22/
( T
2
).
b. T L
__
22/
T
2
_
=
_
22/
(T L T
2
).
c. If
nn
is a subset of for n = 1. 2. . . . and m = 1. 2. . . ., is it necessarily true
that
1
_
nD1
_
_
1
_
nD1

nn
_
_
=
1
_
nD1
_
_
1
_
nD1

nn
_
_

Proof. (a) If .
__
22/
T
2
_
, then . and .
_
22/
T
2
; thus, .
and . T
2
for some z, so . ( T
2
) for some z; hence .
_
22/
( T
2
).
On the other hand, if .
_
22/
( T
2
), then . T
2
for some z ,
i.e., . and . T
2
for some z . Thus, . and .
_
22/
T
2
; that is,
.
__
22/
T
2
_
.
1
2 CHAPTER 1 SET THEORY AND METRIC SPACES
(b) If . T L
__
22/
T
2
_
, then . T
2
for all z, then . (T L T
2
) for all z, i.e.,
.
_
22/
(T L T
2
). On the other hand, if .
_
22/
(T L T
2
), then . (T L T
2
)
for all z, i.e., . T or . T
2
for all z; that is, . T L
__
22/
T
2
_
.
(c) They are one and the same set. |L
1C. Ordered pairs
I Exercise 3. Show that, if (.
1
. .
2
) is dened to be

{.
1
]. {.
1
. .
2
]
_
, then
(.
1
. .
2
) = (,
1
. ,
2
) i .
1
= ,
1
and .
2
= ,
2
.
Proof. If .
1
= ,
1
and .
2
= ,
2
, then, clearly, (.
1
. .
2
) =

{.
1
]. {.
1
. .
2
]
_
=

{,
1
]. {,
1
. ,
2
]
_
= (,
1
. ,
2
). Now assume that

{.
1
]. {.
1
. .
2
]
_
=

{,
1
]. {,
1
. ,
2
]
_
.
If .
1
= .
2
, then {.
1
] = {,
1
] and {.
1
. .
2
] = {,
1
. ,
2
]. So, rst, .
1
= ,
1
and then
{.
1
. .
2
] = {,
1
. ,
2
] implies that .
2
= ,
2
. If .
1
= .
2
, then

{.
1
]. {.
1
. .
1
]
_
=

{.
1
]
_
.
So {,
1
] = {,
1
. ,
2
] = {.
1
], and we get ,
1
= ,
2
= .
1
, so .
1
= ,
1
and .
2
= ,
2
holds in this case, too. |L
1D. Cartesian products
IExercise 4. Provide an inductive denition of the ordered n-tuple (.
1
. . . . . .
n
)
of elements .
1
. . . . . .
n
of a set so that (.
1
. . . . . .
n
) and (,
1
. . . . . ,
n
) are equal i
their coordinates are equal in order, i.e., i .
1
= ,
1
. . . . . .
n
= ,
n
.
Proof. Dene (.
1
. . . . . .
n
) = {(1. .
1
). . . . . (n. .
n
)] as a nite sequence. |L
I Exercise 5. Given sets X
1
. . . . . X
n
dene the Cartesian product X
1
X
n
a. by using the denition of ordered n-tuple you gave in Exercise 4,
b. inductively from the denition of the Cartesian product of two sets,
and show that the two approaches are the same.
Proof. (a) X
1
X
n
= { (
_
n
iD1
X
i
)
n
: (i ) X
i
].
(b) From the denition of the Cartesian product of two sets, X
1
X
n
=
{(.
1
. . . . . .
n
): .
i
X
i
], where (.
1
. . . . . .
n
) = ((.
1
. . . . . .
n1
). .
n
).
These two denitions are equal essentially since there is a bijection between
them. |L
IExercise 6. Given sets X
1
. . . . . X
n
let X = X
1
X
n
and let X

be the set of
all functions from {1. . . . . n] into
_
n
kD1
X
k
having the property that (k) X
k
for each k = 1. . . . . n. Show that X

is the same set as X.

Proof. Each function can be written as {(1. .
1
). . . . . (n. .
n
)]. So dene J : X

X as J( ) = (.
1
. . . . . .
n
). |L
SECTION 1.2 METRIC SPACES 3
I Exercise 7. Use what you learned in Exercise 6 to dene the Cartesian prod-
uct X
1
X
2
of denumerably many sets as a collection of certain functions
with domain N.
Proof. X
1
X
2
consists of functions : N
_
1
nD1
X
n
such that (n) X
n
for all n N. |L
1.2 Metric Spaces
2A. Metrics on R
n
I Exercise 8. Verify that each of the following is a metric on R
n
:
a. j(x. y) =
p
n

iD1
(.
i
,
i
)
2
.
b. j
1
(x. y) =

n
iD1
[.
i
,
i
[.
c. j
2
(x. y) = max{[.
1
,
1
[. . . . . [.
n
,
n
[].
Proof. Clearly, it suces to verify the triangle inequalities for all of the three
functions. Pick arbitrary x. y. z R
n
.
(a) By Minkowskis Inequality, we have
j(x. z) =
p
n

iD1
(.
i
:
i
)
2
=
p
n

iD1
(.
i
,
i
) (,
i
:
i
)|
2
6
p
n

iD1
(.
i
,
i
)
2

p
n

iD1
(,
i
:
i
)
2
= j(x. y) j(y. z).
(b) We have
j
1
(x. z) =
n

iD1
[.
i
:
i
[ =
n

iD1
([.
i
,
i
[ [,
i
:
i
[) = j
1
(x. y) j
1
(y. z).
(c) We have
j
2
(x. z) = max{[.
1
:
1
[. . . . . [.
n
:
n
[]
6 max{[.
1
,
1
[ [,
1
:
1
[. . . . . [.
n
,
n
[ [,
n
:
n
[]
6 max{[.
1
,
1
[. . . . . [.
n
,
n
[] max{[,
1
:
1
[. . . . . [,
n
:
n
[]
= j
2
(x. y) j
2
(y. z). |L
4 CHAPTER 1 SET THEORY AND METRIC SPACES
2B. Metrics on C(I)
I Exercise 9. Let C(I) denote the set of all continuous real-valued functions on
the unit interval I and let .
0
be a xed point of I.
a. j(. g) = sup
x2I
[(.) g(.)[ is a metric on C(I).
b. o(. g) =
_
1
0
[(.) g(.)[ d. is a metric on C(I).
c. j(. g) = [(.
0
) g(.
0
)[ is a pseudometric on C(I).
Proof. Let . g. h C(I). It is clear that j, o, and j are positive, symmetric; it
is also clear that j and o satisfy M-b.
(a) We have
j(. h) = sup
x2I
[(.) h(.)[ 6 sup
x2I
([(.) g(.)[ [g(.) h(.)[)
6 sup
x2I
[(.) g(.)[ sup
x2I
[g(.) h(.)[
= j(. g) j(g. h).
(b) We have
o(. h) =
_
1
0
[(.) h(.)[ 6
_
1
0
[(.) g(.)[
_
1
0
[g(.) h(.)[
= o(. g) o(g. h).
(c) For arbitrary . g C(I) with (.
0
) = g(.
0
) we have j(. g) = 0, so j(. g) =
0 does not imply that = g. Further, j(. h) = [(.
0
)h(.
0
)[ 6 [(.
0
)g(.
0
)[
[g(.
0
) h(.
0
)[ = j(. g) j(g. h). |L
2C. Pseudometrics
I Exercise 10. Let (M. j) be a pseudometric space. Dene a relation - on M
by . - , i j(.. ,) = 0. Then - is an equivalence relation.
Proof. (i) . - . since j(.. .) = 0 for all . M. (ii) . - , i j(.. ,) = 0 i
j(,. .) = 0 i , - .. (iii) Suppose . - , and , - :. Then j(.. :) 6 j(.. ,)
j(,. :) = 0; that is, j(.. :) = 0. So . - :. |L
I Exercise 11. If M

is he set of equivalence classes in M under the equiva-

lence relation - and if j

is dened on M

by j

(.|. ,|) = j(.. ,), then j

is a
well-dened metric on M

.
Proof. j

is well-dened since it does not dependent on the representative of

.|: let .
0
.| and ,
0
,|. Then
j(.
0
. ,
0
) 6 j(.
0
. .) j(.. ,) j(,. ,
0
) = j(.. ,).
SECTION 1.2 METRIC SPACES 5
Symmetrically, j(.. ,) 6 j(.
0
. ,
0
). To verify j

is a metric on M

, it suces to
show that j

satises the triangle inequality. Let .|. ,|. :| M

. Then
j

(.|. :|) = j(.. :) 6 j(.. ,) j(,. :) = j

(.|. ,|) j

(,|. :|). |L
I Exercise 12. If h: M M

is the mapping h(.) = .|, then a set in M is

closed (open) i h() is closed (open) in M

.
Proof. Let be open in M and h(.) = .| h() for some . . Since is
open, there exist an c-disk U
p
(.. c) contained in . For each , U
p
(.. c), we
have h(,) = ,| h(), and j

(.|. ,|) = j(.. ,) 6 c. Hence, for each .| h(),

there exists an c-disk U
p
(.|. c) = h(U
p
(.. c)) contained in h(); that is, h() is
open in M

. Since h is surjective, it is now easy to see that h() is closed in

M

whenever is closed in M. |L
I Exercise 13. If is any real-valued function on a set M, then the distance
function j
(
(.. ,) = [(.) (,)[ is a pseudometric on M.
Proof. Easy. |L
I Exercise 14. If (M. j) is any pseudometric space, then a function : M R
is continuous i each set open in (M. j
(
) is open in (M. j).
Proof. Suppose that is continuous and G is open in (M. j
(
). For each . G,
there is an c > 0 such that if [(,)(.)[ < c then , G. The continuity of at
. implies that there exists > 0 such that if j(,. .) < then [(,) (.)[ < c,
and so , U. We thus proved that for each . U there exists a -disk U
p
(.. j)
contained in G; that is, G is open in (M. j).
Conversely, suppose that each set is open in (M. j) whenever it is open in
(M. j
(
). For each . (M. j
(
), there is an c-disk U
p
f
(.. c) contained in M since
M is open under j
(
; then U
p
f
(.. c) is open in (M. j) since U
p
f
(.. c) is open in
(M. j
(
). Hence, there is an -disk U
p
(.. ) such that U
p
(.. ) U
p
f
(.. c); that is,
if j(,. .) < , then [(,) (.)[ < c. So is continuous on M. |L
2D. Disks Are Open
I Exercise 15. For any subset of a metric space M and any c > 0, the set
U(. c) is open.
Proof. Let M and c > 0. Take an arbitrary point . U(. c); take an
arbitrary point , such that j(.. ,) < c. Observe that every c-disk U(,. c) is
contained in U(. c). Since . U(,. c) and U(,. c) is open, there exists an -disk
U(.. ) contained in U(,. c). Therefore, U(. c) is open. |L
6 CHAPTER 1 SET THEORY AND METRIC SPACES
2E. Bounded Metrics
I Exercise 16. If j is any metric on M, the distance function j

(.. ,) =
min{j(.. ,. ). 1] is a metric also and is bounded.
Proof. To see j

is a metric, it suces to show the triangle inequality. Let

.. ,. : M. Then
j

(.. :) = min{j(.. :). 1] 6 min{j(.. ,) j(,. :). 1]

6 min{j(.. ,). 1] min{j(,. :). 1]
= j

(.. ,) j

(,. :).
It is clear that j

is bounded above by 1. |L
I Exercise 17. A function is continuous on (M. j) i it is continuous on
(M. j

).
Proof. It suces to show that j and j

are equivalent. If G is open in (M. j),

then for each . G there is an c-disk U
p
(.. c) G. Since U
p
(.. c) U
p
(.. c),
we know G is open in (M. j

). Similarly, we can show that G is open in (M. j

)
whenever it is open in (M. j). |L
2F. The Hausdor Metric
Let j be a bounded metric on M; that is, for some constant , j(.. ,) 6 for
all . and , in M.
I Exercise 18. Show that the elevation of j to the power set P(M) as dened
in 2.4 is not necessarily a pseudometric on P(M).
Proof. Let M : {(.
1
. .
2
) R
2
: .
2
1
.
2
2
6 1], and let j be the usual metric.
Then j is a bounded metric on M. We show that the function j

: (1. J)
inf
x2T,,2T
j(.. ,), for all 1. J P(M), is not a pseudometric on P(M) by
showing that the triangle inequality fails. Let 1. J. G P(M), where 1 =
U
p
((1,4. 0). 1,4), G = U
p
((1,4. 0). 1,4), and J meets both 1 and G. Then
j

(1. G) > 0, but j

(1. J) = j

(J. G) = 0. |L
I Exercise 19. Let F (M) be all nonempty closed subsets of M and for . T
F (M) dene
J

(T) = sup{j(. .) : . T]
J(. T) = max{J

(T). J
B
()].
Then J is a metric on F (M) with the property that J({.]. {,]) = j(.. ,). It is
called the Hausdor metric on F (M).
SECTION 1.2 METRIC SPACES 7
Proof. Clearly, J is nonnegative and symmetric. If J(. T) = 0, then J

(T) =
J
B
() = 0, i.e., sup
,2B
j(. ,) = sup
x2
j(T. .) = 0. But then j(. ,) = 0 for all
, T and j(T. .) = 0 for all . . Since is closed, we have , for all
, T; that is, T . Similarly, T. Hence, = T.
We next show the triangle inequality of J. Let . T. C F (M). For an ar-
bitrary point a , take a point b C such that j(a. b) = j(T. a) (since T is
closed, such a point exists). Then
j(a. b) 6 sup
x2
j(T. .) = J
B
() 6 J(. T).
For this b T, we take a point c C such that j(b. c) 6 J(T. C). Therefore,
j(a. c) 6 j(a. b) j(b. c) 6 J(. T) J(T. C).
We thus proved that for every a , there exists c C (depends on a), such
that j(a. c) 6 J(. T) J(T. C). In particular, we have
j(a. C) = inf
z2C
j(a. :) 6 J(. T) J(T. C).
Since the above inequality holds for all a , we obtain
J
C
() = sup
x2
j(a. C) 6 J(. T) J(T. C). (1.1)
Similarly, for each c C there exists b T with j(c. b) 6 J(T. C); for this b,
there exists a with j(a. b) 6 J(. T). Hence j(a. c) 6 J(. T) J(T. C) for
all c C. The same argument shows that
J

(C) 6 J(. T) J(T. C). (1.2)

Combining (1.1) and (1.2) we get the desired result.
Finally, notice that J
fxg
({,]) = J
f,g
({.]) = j(.. ,); hence, J({.]. {,]) =
j(.. ,). |L
I Exercise 20. Prove that closed sets and T are close in the Hausdor
metric i they are uniformly close; that is, J(. T) < c i U
p
(T. c) and
T U
p
(. c).
Proof. If J(. T) < c, then sup
,2B
j(. ,) = j

(T) < c; that is, j(. ,) < c for

all , T, so T U
p
(. c). Similarly, U
p
(T. c).
Conversely, if U
p
(T. c), then j(T. .) < c for all . . Since is closed,
we have J
B
() < c; similarly, T U
p
(. c) implies that J

(T) < c. Hence,

J(. T) < c. |L
8 CHAPTER 1 SET THEORY AND METRIC SPACES
2G. Isometry
Metric spaces (M. j) and (N. o) are isometric i there is a one-one function
from M onto N such that j(.. ,) = o((.). (,)) for all . and , in M; is
called an isometry.
I Exercise 21. If is an isometry from M to N, then both and
1
are
continuous functions.
Proof. By denition, is (uniformly) continuous on M: for every c > 0, let
= c; then j(.. ,) < implies that o((.). (,)) = j(.. ,) < c.
On the other hand, for every c > 0 and , N, pick the unique
1
(,)
M (since is bijective). For each : N with o(,. :) < c, we must have
j(
1
(,).
1
(:)) = o((
1
(,)). (
1
(:))) = o(,. :) < c; that is,
1
is con-
tinuous. |L
I Exercise 22. R is not isometric to R
2
(each with its usual metric).
Proof. Consider S
1
= {(.. ,) R
2
: .
2
,
2
= 1]. Notice that there are only
two points around
1
(0. 0) with distance 1. |L
I Exercise 23. I is isometric to any other closed interval in R of the same
length.
Proof. Consider the function : I a. a 1| dened by (.) = a . for all
. I. |L
2
TOPOLOGICAL SPACES
2.1 Fundamental Concepts
3A. Examples of Topologies
I Exercise 24. If F is the collection of all closed, bounded subset of R (in its
usual topology), together with R itself, then F is the family of closed sets for a
topology on R strictly weaker than the usual topology.
Proof. It is easy to see that F is a topology. Further, for instance, (o. 0| is a
closed set of R, but it is not in F . |L
I Exercise 25. If X, show that the family of all subsets of X which contain
, together with the empty set , is a topology on X. Describe the closure and
interior operations. What topology results when = ? when = X?
Proof. Let
E = {1 X : 1] L {] .
Now suppose that 1
2
E for each z . Then
_
2
1
2
X and so
_
1
2
E. The other postulates are easy to check.
For any set T X, if T, then T E and so T
B
= T; if not, then T
B
= .
If = , then E is the discrete topology; if = X, then E = {. X]. |L
3D. Regularly Open and Regularly Closed Sets
An open subset G in a topological space is regular open i G is the interior of
its closure. A closed subset is regularly closed i it is the closure of its interior.
I Exercise 26. The complement of a regularly open set is regularly closed and
vice versa.
Proof. Suppose G is regular open; that is, G = (

G)

. Then
9
10 CHAPTER 2 TOPOLOGICAL SPACES
X G = X (

G)

= X

G = (X G)

.
Hence, X G is regularly closed. If J is regular closed, i.e., J = J

, then
X J = X J

= (X J

= (X J)

:
that is, X J is regularly open. |L
I Exercise 27. There are open sets in R which are not regularly open.
Proof. Consider Q. We have (

Q)

= R

= R = Q. So Q is not regularly open.

|L
I Exercise 28. If is any subset of a topological space, then int(cl()) is reg-
ularly open.
Proof. Let be a subset of a topological space X. We then have
int(cl()) cl(int(cl())) ==int(cl()) = int(int(cl())) int(cl(int(cl()))).
and
int(cl()) cl() ==cl(int(cl())) cl(cl()) = cl()
==int(cl(int(cl()))) int(cl()).
Therefore, int(cl()) = int(cl(int(cl()))); that is, int(cl()) is regularly open.
|L
I Exercise 29. The intersection, but not necessarily the union, of two regularly
open sets is regularly open.
Proof. Let and T be two regularly open sets in a topological space X. Then
( T)

(



T)

= (

)

T)

= T.
and
(



T)

= (

)

T)

= T T
== T = (



T)

=
_
(



T)

( T)

.
Hence, T = ( T)

.
To see that the union of two regularly open sets is not necessarily regularly
open, consider = (0. 1) and T = (1. 2) in R with its usual topology. Then
( L T)

= 0. 2|

= (0. 2) = L T. |L
SECTION 2.1 FUNDAMENTAL CONCEPTS 11
3E. Metrizable Spaces
Let X be a metrizable space whose topology is generated by a metric j.
I Exercise 30. The metric 2j dened by 2j(.. ,) = 2 j(.. ,) generates the
same topology on X.
Proof. Let O
p
be the collection of open sets in (X. j), and let O
2p
be the col-
lection of open sets in (X. 2j). If O O
p
, then for every . O, there exists an
open ball B
p
(.. c) _ O; but then B
2p
(.. c,2) O. Hence, O O
2p
. Similarly, we
can show that O
2p
O
p
. In fact, j and 2j are equivalent metrics. |L
IExercise 31. The closure of a set 1 X is given by

1 = {, X : j(1. ,) = 0].
Proof. Denote

1 : {, X : j(1. ,) = 0]. We rst show that

1 is closed
(see Denition 2.5, p. 17). Take an arbitrary . X such that for every n N,
there exists ,
n


1 with j(.. ,
n
) < 1,2n. For each ,
n


1, take :
n
1 with
j(,
n
. :
n
) < 1,2n. Then
j(.. :
n
) 6 j(.. ,
n
) j(,
n
. :
n
) < 1,n. for all n N.
Thus, j(.. 1) = 0, i.e., .

1. Therefore,

1 is closed. It is clear that 1 _

1, and
so

1

1.
We next show that

1 _

1. Take an arbitrary .

1 and a closed set 1
containing 1. If . X1, then j(.. 1) > 0 (see Exercise 35). But then j(.. 1) >
0 since 1 1 and so
inf
,2T
j(.. ,) > inf
z21
j(.. :).
Hence,

1

1. |L
I Exercise 32. The closed disk U(.. c) = {, : j(.. ,) 6 c] is closed in X, but
may not be the closure of the open disk U(.. c).
Proof. Fix . X. We show that the function j(.. ): X R is (uniformly)
continuous. For any ,. : X, the triangle inequality yields
[j(.. ,) j(.. :)[ 6 j(,. :).
Hence, for every c > 0, take = c, and j(.. ) satises the c- criterion. There-
fore, U(.. c) is closed since U(.. c) = j
1
(.. 0. c|) and 0. c| is closed in R.
To see it is not necessary that U(.. c) = U(.. c), consider c = 1 and the usual
metric on
_
(.. ,) R
2
: .
2
,
2
= 1
_
L
_
(.. 0) R
2
: 0 6 . 6 1
_
:
see Figure 2.1. Observe that (0. 0) U(.. 1), but (0. 0) U(..

1). It follows from

Exercise 31 that (0. 0) U(.. 1). |L
12 CHAPTER 2 TOPOLOGICAL SPACES
0
.
Figure 2.1. U(x,
x
1) U(x, 1).
3H. G

and J
c
Sets
I Exercise 33. The complement of a G

is an J
c
, and vice versa.
Proof. If is a G

set, then there exists a sequence of open sets {U

n
] such
that =
_
1
nD1
U
n
. Then
c
=
_
1
nD1
U
c
n
is J
c
. Vice versa. |L
I Exercise 34. An J
c
can be written as the union of an increasing sequence
J
1
J
2
of closed sets.
Proof. Let T =
_
1
nD1
1
n
, where 1
n
is closed for all n N. Dene J
1
= 1
1
and J
n
=
_
n
iD1
1
i
for n > 2. Then each J
n
is closed, J
1
J
2
, and
_
1
nD1
J
n
=
_
1
nD1
= T. |L
I Exercise 35. A closed set in a metric space is a G

.
Proof. For an arbitrary set X and a point . X, dene
j(.. ) = inf
,2
{j(.. ,)].
We rst show that j(. ): X R is (uniformly) continuous by showing
[j(.. ) j(,. )[ 6 j(.. ,). for all .. , X. (2.1)
For an arbitrary : , we have
j(.. ) 6 j(.. :) 6 j(.. ,) j(,. :).
Take the inmum over : and we get
j(.. ) 6 j(.. ,) j(,. ). (2.2)
SECTION 2.2 NEIGHBORHOODS 13
Symmetrically, we have
j(,. ) 6 j(.. ,) j(.. ). (2.3)
Hence, (2.1) follows from (2.2) and (2.3). We next show that if is closed, then
j(.. ) = 0 i . . The if part is trivial, so we do the only if part. If
j(.. ) = 0, then for every n N, there exists ,
n
such that j(.. ,
n
) < 1,n;
that is, ,
n
.. Since {,
n
] and is closed, we must have . .
Therefore,
=
1
_
nD1
{. X : j(.. ) < 1,n] .
The continuity of j(. ) implies that {. X : j(.. ) < 1,n] is open for all n.
Thus, is a G

set. |L
I Exercise 36. The rationals are an J
c
in R.
Proof. Q is countable, and every singleton set in R is closed; hence, Q is an
J
c
. |L
3I. Borel Sets
2.2 Neighborhoods
4A. The Sorgenfrey Line
I Exercise 37. Verify that the set .. :), for : > ., do form a nhood base at .
for a topology on the real line.
Proof. We need only check that for each . R, the family B
x
:{.. :) : : > .]
satises V-a, V-b, and V-c in Theorem 4.5. V-a is trivial. If .. :
1
) B
x
and
.. :
2
) B
x
, then .. :
1
) .. :
2
) = .. :
1
. :
2
) B
x
and is in .. :
1
) .. :
2
).
For V-c, let .. :) B
x
. Let :
0
(.. :|. Then
_
.. :
0
_
B
x
, and if ,
_
.. :
0
_
, the
right-open interval
_
,. :
0
_
B
,
and
_
,. :
0
_
.. :).
Then, dene open sets using V-d: G R is open if and only if G contains a
set .. :) of each of its points .. |L
I Exercise 38. Which intervals on the real line are open sets in the Sorgenfrey
topology?
Solution.
v Sets of the form (o. .), .. :), or .. o) are both open and closed.
v Sets of the form (.. :) or (.. o) are open in R, since
(.. :) =
_
{,. :) : . < , < :]. |L
14 CHAPTER 2 TOPOLOGICAL SPACES
I Exercise 39. Describe the closure of each of the following subset of the Sor-
genfrey line: the rationals Q, the set {1,n: n = 1. 2. . . . . ], the set {1,n: n = 1. 2. . . .],
the integers Z.
Solution. Recall that, by Theorem 4.7, for each 1 R, we have

1 =

. R: each basic nhood of . meets 1

_
.
Then

Q = R since for any . R, we have .. :) Q = for : > .. Similarly,
{1,n: n = 1. 2. . . .] = {1,n: n = 1. 2. . . .], and

Z = Z. |L
4B. The Moore Plane
I Exercise 40. Verify that this gives a topology on 1 .
Proof. Verify (V-a)(V-c). It is easy. |L
4E. Topologies from nhoods
I Exercise 41. Show that if each point . in a set X has assigned a collection
U
x
of subsets of X satisfying N-a through N-d of 4.2, then the collection
t =

G X : for each . in G, . U G for some U U

x
_
is a topology for X, in which the nhood system at each . is just U
x
.
Proof. We need to check G1G3 in Denition 3.1. Since G1 and G3 are evi-
dent, we focus on G2. Let 1
1
. 1
2
t. Take any . 1
1
1
2
. Then there exist
some U
1
. U
2
U
x
such that . U
1
1
1
and . U
2
1
2
. By N-b, we know
that U
1
U
2
U
x
. Hence,
. U
1
U
2
1
1
1
2
.
and so 1
1
1
2
t. The induction principle then means that t is closed under
nite intersections. |L
4F. Spaces of Functions
I Exercise 42. For each R
I
, each nite subset J of I and each positive ,
let
U(. J. ) =
_
g R
I
: [g(.) (.)[ < , for each . J
_
.
Show that the sets U(. J. ) form a nhood base at , making R
I
a topological
space.
Proof. Denote
SECTION 2.2 NEIGHBORHOODS 15
B
(
=

U(. J. ) : J I. [J[ < o. > 0

_
.
(V -a) For each U(. J. ) B
(
, we have [(.) (.)[ = 0 < for all . J;
hence, U(. J. ).
(V -b) Let U(. J
1
.
1
). U(. J
2
.
2
) B
(
. Dene U(. J
3
.
3
) by letting
J
3
= J
1
L J
2
. and
3
= min{
1
.
2
].
Clearly, U(. J
3
.
3
) B
(
. If g U(. J
3
.
3
), then
[g(.) (.)[ < min{
1
.
2
]. for all . J
1
L J
2
.
Hence, [g(.) (.)[ <
1
for all . J
1
and [g(.) (.)[ <
2
for all . J
2
; that
is, g U(. J
1
.
1
) U(. J
2
.
2
). Hence, there exists U(. J
3
.
3
) B
(
such that
U(. J
3
.
3
) U(. J
1
.
1
) U(. J
2
.
2
).
(V -c) Pick U(. J. ) B
(
. We must show that there exists some U(. J
0
.
0
)
B
(
such that if g U(. J
0
.
0
), then there is some U(g. J
0
.
0
) B

with
U(g. J
0
.
0
) U(. J. ).
Let J
0
= J, and
0
= ,2. Then U(. J. ,2) B
(
. For every g U(. J. ,2),
we have
[g(.) (.)[ < ,2. for all . J.
Let U(g. J
0
.
0
) = U(g. J. ,2). If h U(g. J. ,2), then
[h(.) (.)[ < ,2. for all . J.
Triangle inequality implies that
[h(.) (.)[ 6 [h(.) g(.)[ [g(.) (.)[ < ,2 ,2 = . for all . J:
that is, h U(. J. ). Hence, U(g. J. ,2) U(. J. ).
Now, G R
I
is open i G is contains a U(. J. ) of each G. This denes
a topology on R
I
. |L
I Exercise 43. For each R
I
, the closure of the one-point set { ] is just { ].
Proof. For every g R
I
{ ], pick . I with g(.) = (.). Dene U(g. J. )
with J = {.] and < [g(.) (.)[. Then U(g. {.]. ); that is, U(g. {.]. )
R
I
{ ]. Hence, R
I
{ ] is open, and so { ] is closed. This proves that { ] =
{ ]. |L
I Exercise 44. For R
I
and c > 0, let
V(. c) =
_
g R
I
: [g(.) (.)[ < c. for each . I
_
.
Verify that the sets V(. c) form a nhood base at , making R
I
a topological
space.
16 CHAPTER 2 TOPOLOGICAL SPACES
Proof. Denote V
(
= {V(. c) : c > 0]. We verify the following properties.
(V -a) If V(. c) V
(
, then [(.) (.)[ = 0 < c; that is, V(. c).
(V -b) Let V(. c
1
). V(. c
2
) V
(
. Let c
3
= min{c
1
. c
2
]. If g V(. c
3
), then
[g(.) (.)[ < c
3
= min{c
1
. c
2
]. for all . I.
Hence, V(. c
3
) V(. c
1
) V(. c
2
).
(V -c) For an arbitrary V(. c) V
(
, pick V(. c,2) V
(
. For each g V(. c,2),
pick V(g. c,2) V

. If h V(g. c,2), then [h(.) g(.)[ < c,2 for all . I. Hence
[h(.) (.)[ 6 [h(.) g(.)[ [g(.) (.)[ < c:
that is, V(g. c,2) V(. c). |L
I Exercise 45. Compare the topologies dened in 1 and 3.
Proof. It is evident that for every U(. J. ) B
(
, there exists V(. ) V
(
such that V(. ) U(. J. ). Hence, the topology in 1 is weaker than in 3 by
Hausdor criterion. |L
2.3 Bases and Subbases
5D. No Axioms for Subbase
I Exercise 46. Any family of subsets of a set X is a subbase for some topology
on X and the topology which results is the smallest topology containing the
given collection of sets.
Proof. Let be a family of subsets of X. Let t() be the intersection of all
topologies containing . Such topologies exist, since 2
A
is one such. Also t()
is a topology. It evidently satises the requirements unique and smallest.
The topology t() can be described as follows: It consists of , X, all nite
intersections of the -sets, and all arbitrary unions of these nite intersections.
To verify this, note that since t(), then t() must contain all the sets
listed. Conversely, because
_
distributes over
_
, the sets listed actually do
from a topology containing , and which therefore contains t(). |L
5E. Bases for the Closed Sets
IExercise 47. F is a base for the closed sets in X i the family of complements
of members of F is a base for the open sets.
SECTION 2.3 BASES AND SUBBASES 17
Proof. Let G be an open set in X. Then G = X 1 for some closed subset 1.
Since 1 =
_
T2GF
J, we obtain
G = X
_
_
_
T2GF
J
_
_
=
_
T2GF
J
c
.
Thus, {J
c
: J F ] forms a base for the open sets. The converse direction is
similar. |L
I Exercise 48. F is a base for the closed sets for some topology on X i (a)
whenever J
1
and J
2
belong to F , J
1
L J
2
is an intersection of elements of F ,
and (b)
_
T2F
J = .
Proof. If F is a base for the closed sets for some topology on X, then (a)
and (b) are clear. Suppose, on the other hand, X is a set and F a collection
of subsets of X with (a) and (b). Let T be all intersections of subcollections
from F . Then any intersection of members of T certainly belongs to T , so T
satises (F-a). Moreover, if F
1
F and F
2
F , so that
_
T2F
1
1 and
_
T2F
2
J
are elements of T , then
_
_
_
T2F
1
1
_
_
L
_
_
_
T2F
2
J
_
_
=
_
T2F
1
_
T2F
2
(1 L J).
But by property (a), the union of two elements of F is an intersection of el-
ements of F , so (
_
T2F
1
1) L (
_
T2F
2
J) is an intersection of elements of F ,
and hence belongs to T . Thus T satises (F-b). Finally, T by (b) and X T
since X is the intersection of the empty subcollection from F . Hence T sat-
ises (F-c). This completes the proof that T is the collection of closed sets of
X. |L
3
NEW SPACES FROM OLD
3.1 Subspaces
3.2 Continuous Functions
7A. Characterization of Spaces Using Functions
I Exercise 49. The characteristic function of is continuous i is both open
and closed in X.
Proof. Let 1

: X R be the characteristic function of , which is dened by

1

(.) =

1 if .
0 if . .
First suppose that 1

is continuous. Then, say, 1

1

_
_
1,2. 2
_
_
= is open,
and 1
1

_
_
1. 1,2
_
_
= X is open. Hence, is both open and closed in X.
Conversely, suppose that is both open and closed in X. For any open set
U R, we have
1
1

(U) =

if 1 U and 0 U
X if 1 U and 0 U
if 1 U and 0 U
X if 1 U and 0 U.
Then 1

is continuous. |L
IExercise 50. X has the discrete topology i whenever Y is a topological space
and : X Y , then is continuous.
Proof. Let Y be a topological space and : X Y . It is easy to see that
is continuous if X has the discrete topology, so we focus on the suciency
19
20 CHAPTER 3 NEW SPACES FROM OLD
direction. For any X, let Y = R and = 1

. Then by Exercise 49 is
open. |L
7C. Functions Agreeing on A Dense Subset
I Exercise 51. If and g are continuous functions from X to R, the set of
points . for which (.) = g(.) is a closed subset of X. Thus two continuous
maps on X to R which agree on a dense subset must agree on all of X.
Proof. Denote = {. X : (.) = g(.)]. Take a point , such that
(,) > g(,) (if it is not true then let g(,) > (,)). Take an c > 0 such that
(,) c > g(,) c. Since and g are continuous, there exist nhoods U
1
and U
2
of , such that U
1
| (c (,). c (,)) and gU
2
| (c g(,). c g(,)).
Let U = U
1
U
2
. Then U is a nhood of . and for every : U we have
(:) g(:) > (.) c| g(.) c| > 0.
Hence, U ; that is, U is open, and so {. X : (.) = g(.)] = X U is
closed.
Now suppose that D : {. X : (.) = g(.)] is dense. Take an arbitrary
. X. Since and g are continuous, for each n N, there exist nhoods V
(
and
V

such that [(,) (.)[ < 1,n for all , V

(
and [g(,) g(.)[ < 1,n for all
, V

. Let V
n
= V
(
V

. Then there exists .

n
V
n
D with [(.
n
)(.)[ < 1,2n
and [g(.
n
) g(.)[ < 1,2n. Since (.
n
) = g(.
n
), we have
[(.) g(.)[ 6 [(.) (.
n
)[ [(.
n
) g(.)[ = [(.) (.
n
)[ [g(.
n
) g(.)[
< 1,n.
Therefore, (.) = g(.). |L
7E. Range Immaterial
I Exercise 52. If Y 7 and : X Y , then is continuous as a map from
X to Y i is continuous as a map from X to 7.
Proof. Let : X 7 be continuous. Let U be open in Y . Then U = Y V for
some V which is open in 7. Therefore,

1
(U) =
1
(Y V ) =
1
(Y )
1
(V ) = X
1
(V ) =
1
(V )
is open in X, and so is continuous as a map from X to Y .
Conversely, let : X Y be continuous and V be open in 7. Then

1
(V ) =
1
(Y V ). Since Y V is open in Y and is continuous from
X to Y , the set
1
(Y V ) is open in X and so is continuous as a map from
X to 7. |L
SECTION 3.2 CONTINUOUS FUNCTIONS 21
7G. Homeomorphisms within the Line
I Exercise 53. Show that all open intervals in R are homeomorphic.
Proof. We have
v (a. b) - (0. 1) by
1
(.) = (. a),(b a).
v (a. o) - (1. o) by
2
(.) = . a 1.
v (1. o) - (0. 1) by
3
(.) = 1,..
v (o. a) - (a. o) by
4
(.) = ..
v (o. o) - (,2. ,2) by
5
(.) = arctan..
Therefore, by compositing, every open interval is homeomorphic to (0. 1). |L
I Exercise 54. All bounded closed intervals in R are homeomorphic.
Proof. a. b| - 0. 1| by (.) = (. a),(b a). |L
I Exercise 55. The property that every real-valued continuous function on X
assumes its maximum is a topological property. Thus, I : 0. 1| is not homeo-
morphic to R.
Proof. Every continuous function assumes its maximum on 0. 1|; however, .
2
has no maximum on R. Therefore, I ,- R. |L
7K. Semicontinuous Functions
I Exercise 56. If

is a lower semicontinuous real-valued function on X for

each , and if sup

(.) exists at each . X, then the function (.) =

sup

(.) is lower semicontinuous on X.

Proof. For an arbitrary a R, we have (.) 6 a i

(.) 6 a for all .

Hence,
{. X : (.) 6 ] =
_
2
{. X :

(.) 6 a] .
and so
1
(o. a| is closed; that is, is lower semicontinuous. |L
IExercise 57. Every continuous function from X to R is lower semicontinuous.
Thus the supremum of a family of continuous functions, if it exists, is lower
semicontinuous. Show by an example that lower semicontinuous cannot be
replaced by continuous in the previous sentence.
Proof. Suppose that : X R is continuous. Since (o. .| is closed in R, the
set
1
(o. .| is closed in X; that is, is lower semicontinuous.
To construct an example, let : 0. o) R be dened as follows:
22 CHAPTER 3 NEW SPACES FROM OLD

n
(.) =

n. if 0 6 . 6 1,n
1 if . > 1,n.
Then
(.) = sup
n

n
(.) =

0 if . = 0
1 if . > 0.
and is not continuous. |L
I Exercise 58. The characteristic function of a set in X is lower semicontin-
uous i is open, upper semicontinuous i is closed.
Proof. Observe that
1
1

(o. a| =

if a < 0
X if 0 6 a < 1
X if a > 1.
Therefore, 1

is LSC i is open. Similarly for the USC case. |L

I Exercise 59. If X is metrizable and is a lower semicontinuous function
from X to I, then is the supremum of an increasing sequence of continuous
functions on X to I.
Proof. Let J be the metric on X. First assume is nonnegative. Dene

n
(.) = inf
z2A
{(:) nJ(.. :)] .
If .. , X, then (:) nJ(.. :) 6 (:) nJ(,. :) nJ(.. ,). Take the inf over :
(rst on the left side, then on the right side) to obtain
n
(.) 6
n
(,) nJ(.. ,).
By symmetry,
[
n
(.)
n
(,)[ 6 nJ(.. ,):
hence,
n
is uniformly continuous on X. Furthermore, since > 0, we have
0 6
n
(.) 6 (.) nJ(.. .) = (.). By denition,
n
increases with n; we must
show that lim
n

n
is actually .
Given c > 0, by denition of
n
(.) there is a point :
n
X such that

n
(.) c > (:
n
) nJ(.. :
n
) > nJ(.. :
n
) (3.1)
since > 0. But
n
(.) c 6 (.) c; hence J(.. :
n
) 0. Since is LSC, we
have liminf
n
(:
n
) > (.) (Ash, 2009, Theorem 8.4.2); hence
(:
n
) > (.) c ev. (3.2)
By (3.1) and (3.2),

n
(.) > (:
n
) c nJ(.. :
n
) > (:
n
) c > (.) 2c
SECTION 3.2 CONTINUOUS FUNCTIONS 23
for all suciently large n. Thus,
n
(.) (.).
If [ [ 6 M < o, then M is LSC, nite-valued, and nonnegative. If 0 6
g
n
] ( M), then
n
= (g
n
M) ] and [
n
[ > M. |L
7M. C(X) and C

(X)
I Exercise 60. If and g belong to C(X), then so do g, g and a , for
a R. If, in addition, and g are bounded, then so are g, g and a .
Proof. We rst do g. Since . g C(X), for each . X and each c > 0,
there exist nhoods U
1
and U
2
of . such that U
1
| (c,2 (.). c,2 (.))
and gU
2
| (c,2 g(.). c,2 g(.)). Let U = U
1
U
2
. Then U is a nhood of .,
and for every , U, we have
[(,) g(,)| (.) g(.)|[ 6 [(,) (.)[ [g(,) g(.)[ < c:
that is, g is continuous.
We then do a . We suppose that a > 0 (all other cases are similar). For
each . X and c > 0, there exists a nhood U of . such that U| (c,a
(.). c,a (.)). Then (a )U| (c a (.). c a (.)). So a C(X).
Finally, to do g, we rst show that
2
C(X) whenever C(X). For
each . X and c > 0, there is a nhood U of . such that U| (
_
c
(.).
_
c (.)). Then
2
U| (c
2
(.). c
2
(.)), i.e.,
2
C(X). Since
(.) g(.) =
1
4
_
_
(.) g(.)
_
2

_
(.) g(.)
_
2
_
.
we know that g C(X) from the previous arguments. |L
I Exercise 61. C(X) and C

(X) are algebras over the real numbers.

Proof. It follows from the previous exercise that C(X) is a vector space on R.
So everything is easy now. |L
I Exercise 62. C

(X) is a normed linear space with the operations of addition

and scalar multiplication given above and the norm [ [ = sup
x2A
[(.)[.
Proof. It is easy to see that C

(X) is a linear space. So it suces to show that

[ [ is a norm on C

(X). We focus on the triangle inequality. Let . g C

(X).
Then for every . X, we have [(.) g(.)[ 6 [(.)[ [g(.)[ 6 [ [ [g[;
hence, [ g[ 6 [ [ [g[. |L
24 CHAPTER 3 NEW SPACES FROM OLD
3.3 Product Spaces, Weak Topologies
8A. Projection Maps
I Exercise 63. The th projection map

is continuous and open. The projec-

tion
1
: R
2
R is not closed.
Proof. Let U

be open in X

. Then
1

(U

) is a subbasis open set of the

Tychono topology on

, and so is open. Hence,

is continuous.
Take an arbitrary basis open set U in the Tychono topology. Denote 1 :
{1. . . . . n]. Then
U =

.
where U

is open in X

for every , and U

j
= X

j
for all 1. Hence,

(U) =

if =
i
for some i 1
X

otherwise.
That is,

(U) is open in X

in both case. Since any open set is a union of

basis open sets, and since functions preserve unions, the image of any open
set under

is open.
0
J
Figure 3.1. t
1
(T) D (0, 1)
Finally, let J = epi(1,.). Then J is closed in R
2
, but
1
(J) = (0. o) is open
in R; that is,
1
is not closed. See Figure 3.1. |L
I Exercise 64. Show that the projection of I R onto R is a closed map.
Proof. Let : I R R be the projection. Suppose I R is closed, and
suppose ,
0
R |. For every . I, since (.. ,
0
) and is closed,
we nd a basis open subset U(.) V(.) of I R that contains (.. ,
0
), and
U(.) V(.)| = . The collection {U(.) : . I] covers I, so nitely many
of them cover I by compactness, say U(.
1
). . . . . U(.
n
) do. Now dene V =
SECTION 3.3 PRODUCT SPACES, WEAK TOPOLOGIES 25
_
n
iD1
V(.
i
), and note that V is an open nhood of ,
0
, and V | = . So |
is closed; that is, is closed. See Lee (2011, Lemma 4.35, p. 95) for the Tube
Lemma.
Generally, if : X Y X is a projection may where Y is compact, then
is a closed map. |L
8B. Separating Points from Closed Sets
I Exercise 65. If

is a map (continuous function) of X to X

for each ,
then {

: ] separates points from closed sets in X i {

1

V | :
. V open in X

] is a base for the topology on X.

Proof. Suppose that {
1

V | : . V open in X

] consists of a base for the

topology on X. Let T be closed in X and . T. Then . X T and X T is
open in X. Hence there exists
1

V | such that .
1

V | X T; that is,

(.) V . Since V

T| = , i.e.,

T| X

V , and X

V is closed, we
get

T| X

V . Thus,

(.)

T|.
Next assume that {

: ] separates points from closed sets in X. Take

an arbitrary open subset U X and . U. Then T : X U is closed in X,
and hence there exists such that

(.)

T|. Then

(.) X

T|
and, since X

T| is open in X

, there exists an open set V of X

such that

(.) V X

T|. Therefore,
.
1

V |
1

_
X

T|
_
= X
1

T|
_
X
1

T||
X T
= U.
Hence, {
1

V | : . V open in X

] is a base for the topology on X. |L

8D. Closure and Interior in Products
Let X and Y be topological spaces containing subsets and T, respectively. In
the product space X Y :
I Exercise 66. ( T)
B
=
B
T
B
.
Proof. Since
B
is open in and T
B
T is open in T, the set
B
T
B

T is open in T; hence,
B
T
B
( T)
B
.
For the converse inclusion, let x = (a. b) ( T)
B
. Then there is an basis
open set U
1
U
2
such that x U
1
U
2
T, where U
1
is open in and U
2
is open in T. Hence, a U
1
and b U
2
T; that is, a
B
and b T
B
.
Then x
B
T
B
. |L
26 CHAPTER 3 NEW SPACES FROM OLD
I Exercise 67. T =



T.
Proof. See Exercise 68. |L
I Exercise 68. Part 2 can be extended to innite products, while part 1 can be
extended only to nite products.
Proof. Assume that y =
_
,

; we show that ,

for each ; that

is, y

. Let ,

, where U

is open in Y

; since y
1

(U

), we must
have
=
1

(U

= (U

)
_

_
.
and so U

= . This proves ,

. The converse inclusion is established

by reversing these steps: If y

, then for any open nhood

T :U

1
U

n

_

: =
1
. . . . .
n
_
_
.
each U

i
= so that T

= . |L
I Exercise 69. Fr( T) =

Fr(T)| L Fr()

T|.
Proof. We have
Fr( T) = T (X Y ) ( T)
= (



T)
_
(X Y ) (

)
_
= (



T)
_
_
X (Y T

)
_
L
_
(X

) Y
_
_
=

Fr(T)| L Fr()

T|. |L
I Exercise 70. If X

is a nonempty topological space and

, for each
, then

is dense in

is dense in X

, for each .
Proof. It follows from Exercise 68 that

:
that is,

is dense in

is dense in X

, for each . |L
8E. Miscellaneous Facts about Product Spaces
Let X

be a nonempty topological space for each , and let X =

.
I Exercise 71. If V is a nonempty open set in X, then

(V ) = X

for all but

nitely many .
Proof. Let T

be the topology on X

for each . Let V be an arbitrary open

set in X. Then V =
_
k21
T
k
, where for each k 1 we have T
k
=
2
1
k
,
SECTION 3.3 PRODUCT SPACES, WEAK TOPOLOGIES 27
and for each we have 1
k
T

while

k
:{ : 1
k
= X

]
is nite. Then
_
k21

k
is nite. If
0

_
k21

k
, then there exists k
0
1 such
that 1

0
k
0
= X

0
. Then

0
(T
k
0
) =
1

0
_

2
1
k
0
_
= X

0
.
and so X

0
=
1

0
(T
k
0
)
1

0
(V ) implies that
1

0
(V ) = X

0
. |L
I Exercise 72. If b

is a xed point in X

, for each , then X

0

0
= {. X :
.

= b

whenever =
0
] is homeomorphic to X

0
.
Proof. Write an element in X
0

0
as (.

0
. b

0
). Then consider the mapping
(.

0
. b

0
) .

0
. |L
8G. The Box Topology
Let X

be a topological space for each .

I Exercise 73. In

, the sets of the form

, where U

is open in X

for each , form a base for a topology.

Proof. Let B :

: . U

open in X

_
. Then it is clear that

B since X

is open for each . Now take any T

1
. T
2
B, with T
1
=

U
1

and T
2
=

U
2

. Let
=
_

1
.
2
. . . .
_
T
1
T
2
=

_
U
1

U
2

_
.
Then

U
1

U
2

, and so there exists an open set T

such that

U
1

U
2

. Hence,

B and T T
1
T
2
. |L
8H. Weak Topologies on Subspaces
Let X have the weak topology induced by a collection of maps

: X X

,
for .
I Exercise 74. If each X

has the weak topology given by a collection of maps

g
2
: X

Y
2
, for z

, then X has the weak topology given by the maps

g
2

: X Y
2
for and z

.
Proof. A subbase for the weak topology on X

induced by {g
2
: z

] is
_
g
1
2
(U
2
) : z

. U
2
open in Y
2
_
.
28 CHAPTER 3 NEW SPACES FROM OLD
Then a subbasic open set in X for the weak topology on X induced by
{

: ] is
_

g
1
2
(U
2
)| : . z

. U
2
open in Y
2
_
.
Since
1

(g
1
2
(U
2
)) = (g
2

)
1
(U
2
), we get the result. |L
I Exercise 75. Any T X has the weak topology induced by the maps

T.
Proof. As a subspace of X, the subbase on T is
_
T
1

(U

) : . U

open in X

_
.
On the other hand, (

T)
1
(U

) = T
1

(U

) for every and U

open
in X

. Hence, the above set is also the subbase for the weak topology induced
by {

T : ]. |L
3.4 Quotient Spaces
9B. Quotients versus Decompositions
I Exercise 76. The process given in 9.5 for forming the topology on a decom-
position space does dene a topology.
Proof. Let (X. T ) be a topological space; let D be a decomposition of X. Dene
F D is open in D ==
_
{J : J F ] is open in X. (3.3)
Let T be the collection of open sets dened by (3.3). We show that (D. T) is a
topological space.
v Take an arbitrary collection {F
i
]
i2J
T; then
_
{J : J F
i
] is open in X
for each i 1. Hence,
_
i2J
F
i
T since
_
T2
_
i2I
F
i
J =
_
i2J
_
_
_
T2F
i
J
_
_
is open in X.
v Let F
1
. F
2
T; then
_
T2F
1
1 and
_
T2F
2
J are open in X. Therefore, F
1

F
2
T since
_
T2F
1
\F
2
J =
_
_
_
T2F
1
1
_
_

_
_
_
T2F
2
J
_
_
is open in X.
v T since
_
= is open in X; nally, D T since
_
D = X. |L
SECTION 3.4 QUOTIENT SPACES 29
I Exercise 77. The topology on a decomposition space D of X is the quotient
topology induced by the natural map 1 : X D. (See 9.6.)
Proof. Let T be the decomposition topology of D, and let T
1
be the quotient
topology induced by 1. Take an open set F T; then
_
T2F
J is open in X.
Hence,
1
1
(F ) = 1
1
_
_
_
T2F
J
_
_
=
_
T2F
1
1
(J) =
_
T2F
J
is open in X, and so F T
1
. We thus proved that T T
1
.
Next take an arbitrary F T
1
. By denition, we have 1
1
(F ) =
_
T2F
J is
open in X. But then F T.
We nally prove Theorem 9.7 (McCleary, 2006, Theorem 4.18): Suppose
: X Y is a quotient map. Suppose - is the equivalence relation dened
on X by . - .
0
if (.) = (.
0
). Then the quotient space X, - is homeomorphic
to Y .
By the denition of the equivalence relation, we have the diagram:
X
X, -
Y
Y
h

1
=

1
h
Dene h: X, - Y by letting h(.|) = (.). It is well-dened. Notice that
h 1 = since for each . X we obtain
(h 1)(.) = h(1(.)) = h(.|) = (.).
Both and 1 are quotient maps so h is continuous by Theorem 9.4. We show
that h is injective, subjective and h
1
is continuous, which implies that h is a
homeomorphism. If h(.|) = h(.
0
|), then (.) = (.
0
) and so . - .
0
; that is,
.| = .
0
|, and h is injective. If , Y , then , = (.) since is surjective and
h(.|) = (.) = , so h is surjective. To see that h
1
is continuous, observe that
since is a quotient map and 1 is a quotient map, this shows 1 = h
1
and
Theorem 9.4 implies that h
1
is continuous. |L
4
CONVERGENCE
4.1 Inadequacy of Sequences
10B. Sequential Convergence and Continuity
I Exercise 78. Find spaces X and Y and a function J : X Y which is not
continuous, but which has the property that J(.
n
) J(.) in Y whenever .
n

. in X.
Proof. Let X = R
R
and Y = R. Dene J : R
R
R by letting J( ) =
sup
x2R
[(.)[. Then J is not continuous: Let
1 =
_
R
R
: (.) = 0 or 1 and (.) = 0 only nitely often
_
.
and let g R
R
be the function which is 0 everywhere. Then g

1. However,
0 J

1| since J(g) = 0, and J1| = {1]. |L

10C. Topology of First-Countable Spaces
Let X and Y be rst-countable spaces.
I Exercise 79. U X is open i whenever .
n
. U, then (.
n
) is eventually
in U.
Proof. If U is open and .
n
. U, then . has a nhood V such that . V
U. By denition of convergence, there is some positive integer n
0
such that
n > n
0
implies .
n
V U; hence, (.
n
) is eventually in U.
Conversely, suppose that whenever .
n
. U, then (.
n
) is eventually in
U. If U is not open, then there exists . U such that for every nhood of V of
. we have V (X U) = . Since X is rst-countable, we can pick a countable
nhood base {V
n
: n N] at .. Replacing V
n
=
_
n
iD1
V
i
where necessary, we
may assume that V
1
V
2
. Now V
n
(X U) = for each n, so we
can pick .
n
V
n
(X U). The result is a sequence (.
n
) contained in X U
31
32 CHAPTER 4 CONVERGENCE
which converges to . U; that is, .
n
. but (.
n
) is not eventually in U. A
contradiction. |L
IExercise 80. J X is closed i whenever (.
n
) is contained in J and .
n
.,
then . J.
Proof. Let J be closed; let (.
n
) be contained in J and .
n
.. Then .

J =
J.
Conversely, assume that whenever (.
n
) is contained in J and .
n
., then
. J. It follows from Theorem 10.4 that .

J with the hypothesis; therefore,

J J, i.e.,

J = J and so J is closed. |L
I Exercise 81. : X Y is continuous i whenever .
n
. in X, then
(.
n
) (.) in Y .
Proof. Suppose is continuous and .
n
.. Since is continuous at .,
for every nhood V of (.) in Y , there exists a nhood U of . in X such that
(U) V . Since .
n
., there exists n
0
such that n > n
0
implies that .
n
U.
Hence, for every nhood V of (.), there exists n
0
such that n > n
0
implies that
(.
n
) V ; that is, (.
n
) (.).
Conversely, let the criterion hold. Suppose that is not continuous. Then
there exists . X and a nhood V of (.), such that for every nhood base
U
n
, n N, of ., there is .
n
U
n
with (.
n
) V . By letting U
1
U
2
,
we have .
n
. and so (.
n
) (.); that is, eventually, (.
n
) is in V . A
contradiction. |L
4.2 Nets
11A. Examples of Net Converence
I Exercise 82. In R
R
, let
1 =
_
R
R
: (.) = 0 or 1. and (.) = 0 only nitely often
_
.
and g be the function in R
R
which is identically 0. Then, in the product topology
on R
R
, g

1. Find a net (
2
) in 1 which converges to g.
Proof. Let U

= {U(g. J. c) : c > 0. J R a nite set] be the nhood base of g.

Order U

as follows:
U(g. J
1
. c
1
) 6 U(g. J
2
. c
2
) == U(g. J
2
. c
2
) U(g. J. c
2
)
== J
1
J
2
and c
2
6 c
1
.
Then U

is a directed set. So we have a net (

T,t
) converging to g. |L
SECTION 4.2 NETS 33
11B. Subnets and Cluster Points
I Exercise 83. Every subnet of an ultranet is an ultranet.
Proof. Take an arbitrary subset 1 X. Let (.
2
) be an ultranet in X, and
suppose that (.
2
) is residually in 1, i.e., there exists some z
0
such that
z > z
0
implies that .
2
1. If (.
2

) is a subnet of (.
2
), then there exists some
j
0
such that z

0
> z
0
. Then for every j > j
0
, we have z

> z
0
, and so j > j
0
implies that .
2

1; that is, (.
2

) is residually in 1. |L
I Exercise 84. Every net has a subnet which is an ultranet.
Proof. See Adamson (1996, Exercise 127, p. 40). |L
I Exercise 85. If an ultranet has . as a cluster point, then it converges to ..
Proof. Let (.
2
) be an ultranet, and . be a cluster point of (.
2
). Let U be a
nhood of .. Then (.
2
) lies in U eventually since for any z
0
there exists z > z
0
such that .
2
U. |L
11D. Nets Describe Topologies
I Exercise 86. Nets have the following four properties:
a. if .
2
= . for each z , then .
2
.,
b. if .
2
., then every subnet of (.
2
) converges to .,
c. if every subnet of (.
2
) has a subnet converging to ., then (.
2
) converges to
.,
d. (Diagonal principal) if .
2
. and, for each z , a net (.
2

)
u2

converges
to .
2
, then there is a diagonal net converging to .; i.e., the net (.
2

)
22/,2

,
ordered lexicographically by , then by M
2
, has a subnet which converges to
..
Proof. (a) If the net (.
2
) is trivial, then for each nhood U of ., we have .
2
U
for all z . Hence, .
2
..
(b) Let (.
()
)
2
be a subnet of (.
2
). Take any nhood U of .. Then there
exists z
0
such that z > z
0
implies that .
2
U since .
2
.. Since is
conal in , there exists j
0
M such that (j
0
) > z
0
; since is increasing,
j > j
0
implies that (j) > (j
0
) > z
0
. Hence, there exists j
0
M such that
j > j
0
implies that .
()
U; that is, .
()
..
(c) Suppose by way of contradiction that (.
2
) does not converge to .. Then
there exists a nhood U of . such that for any z , there exists some (z) > z
with .
(2)
U. Then (.
(2)
) is a subnet of (.
2
), but which has no converging
subnets.
34 CHAPTER 4 CONVERGENCE
(d) Order {(z. j) : z . j M
2
] as follows:
(z
1
. j
1
) 6 (z
2
. j
2
) == z
1
6 z
2
, or z
1
= z
2
and j
1
6 j
2
.
Let U be the nhood system of . which is ordered by U
1
6 U
2
i U
2
U
1
for all
U
1
. U
2
U. Dene
1 =
_
(z. U) : z . U U such that .
2
U
_
.
Order 1 as follows: (z
1
. U
1
) 6 (z
2
. U
2
) i z
1
6 z
2
and U
2
U
1
. For each
(z. U) 1 pick j
2
M
2
so that .
2

U for all j > j

2
(such a j
2
exists since
.
2

.
2
and .
2
U). Dene : (z. U) .
2

for all (z. U) 1 . It now easy to

see that this subnet converges to .. |L
4.3 Filters
12A. Examples of Filter Convergence
I Exercise 87. Show that if a lter in a metric space converges, it must con-
verge to a unique point.
Proof. Suppose a lter F in a metric space (X. J) converges to .. , X. If
. = ,, then there exists r > 0 such that B(.. r) B(,. r) = . But since F .
and F ,, we must have B(.. r) F and B(,. r) F . This contradicts the
fact that the intersection of every two elements in a lter is nonempty. Thus,
. = ,. |L
12C. Ultralters: Uniqueness
IExercise 88. If a lter F is contained in a unique ultralter F
0
, then F = F
0
.
Proof. We rst show: Every lter F on a non-empty set X is the intersection of
the family of ultralters which include F .
Let 1 be a set which does not belong to F . Then for each set J F we
cannot have J 1 and hence we must have J 1
c
= . So F L{1
c
] generates
a lter on X, which is included in some ultralter F
T
. Since 1
c
F
T
we must
have 1 F
T
. Thus 1 does not belong to the intersection of the set of all
ultralters which include F . Hence this intersection is just the lter F itself.
Now, if F is contained in a unique ultralter F
0
, we must have F = F
0
. |L
SECTION 4.3 FILTERS 35
12D. Nets and Filters: The Translation Process
I Exercise 89. A net (.
2
) has . as a cluster point i the lter generated by
(.
2
) has . as a cluster point.
Proof. Suppose . is a cluster point of the net (.
2
). Then for every nhood U of
., we have .
2
U i. o. But then U meets every T
2
0
: {.
2
: z > z
0
], the lter
base of the lter F generated by (.
2
); that is, . is a cluster point of F . The
converse implication is obvious. |L
I Exercise 90. A lter F has . as a cluster point i the net based on F has .
as a cluster point.
Proof. Suppose . is a cluster point of F . If U is a nhood of ., then U meets
every J F . Then for an arbitrary (. J)
F
, pick q J U so that
(q. J)
F
, (q. J) > (. J), and 1(. J) = U; that is, . is a cluster point
of the net based on F .
Conversely, suppose the net based on F has . as a cluster point. Let U be
a nhood of .. Then for every (
0
. J
0
)
F
, there exists (. J) > (
0
. J
0
) such
that U. Then J
0
U = , and so . is a cluster point of F . |L
I Exercise 91. If (.
2

) is a subnet of (.
2
), then the lter generated by (.
2

) is
ner than the lter generated by (.
2
).
Proof. Suppose (.
2

) is a subnet of (.
2
). Let F
2

is the lter generated by

(.
2

), and F
2
be the lter generated by (.
2
). Then the base generating F
2

is the sets T
2

0
= {.
2

: j > j
0
], and the base generating F
2
is the sets
T
2
0
= {.
2
: z > z
0
]. For each such a T
2
0
, there exists j
0
such that z

0
> z
0
;
that is, T
2

0
T
2
0
. Therefore, F
2
F
2

. |L
I Exercise 92. The net based on an ultralter is an ultranet and the lter
generated by an ultranet is an ultralter.
Proof. Suppose F is an ultralter. Let 1 X and we assume that 1 F . Pick
1. If (q. J) > (. 1), then q 1; that is, 1(. J) 1 ev. Hence, the net
based on F is an ultranet.
Conversely, suppose (.
2
) is an ultranet. Let 1 X and we assume that there
exists z
0
such that .
2
1 for all z > z
0
. Then T
2
0
= {.
2
: z > z
0
] 1 and so
1 F , where F is the lter generated by (.
2
). Hence, F is an ultralter. |L
I Exercise 93. The net based on a free ultralter is a nontrivial ultranet.
Hence, assuming the axiom of choice, there are nontrivial ultranets.
Proof. Let F be a free ultralter, and (.
2
) be the net based on F . It follows
from the previous exercise that (.
2
) is an ultranet. If (.
2
) is trivial, i.e., .
2
= .
for some . X and all z
F
, then for all J F , we must have J = {.]. But
then
_
F = {.] = ; that is, F is xed. A contradiction.
36 CHAPTER 4 CONVERGENCE
Now, for instance, the Frechet lter F on R is contained in some free ultra-
lter G by Example (b) when the Axiom of Choice is assumed. Hence, the net
based on G is a nontrivial ultranet. |L
5
SEPARATION AND COUNTABILITY
5.1 The Separation Axioms
13B. T
0
- and T
1
-Spaces
I Exercise 94. Any subspace of a T
0
- or T
1
-space is, respectively, T
0
or T
1
.
Proof. Let X be a T
0
-space, and X. Let . and , be distinct points in .
Then, say, there exists an open nhood U of . such that , U. Then U
is relatively open in , contains ., and , U. The T
1
case can be proved
similarly. |L
I Exercise 95. Any nonempty product space is T
0
or T
1
i each factor space
is, respectively, T
0
or T
1
.
Proof. If X

is a T
0
-space, for each , and . = , in

, then for some

coordinate we have .

= ,

, so there exists an open set U

containing, say,
.

but not ,

. Now
1

(U

) is an open set in

containing . but not ,.

Thus,

is T
0
.
Conversely, if

is a nonempty T
0
-space, pick a xed point b

, for
each . Then the subspace T

: {.

: .

= b

unless = ] is T
0
,
by Exercise 94, and is homeomorphic to X

under the restriction to T

of the
projection map. Thus X

is T
0
, for each . The T
1
case is similar. |L
13C. The T
0
-Identication
For any topological space X, dene - by . - , i {.] = {,].
I Exercise 96. - is an equivalence relation on X.
Proof. Straightforward. |L
I Exercise 97. The resulting quotient space X, -=

X is T
0
.
37
38 CHAPTER 5 SEPARATION AND COUNTABILITY
Proof. We rst show that X is T
0
i whenever . = , then {.] = {,]. If X is
T
0
and . = ,, then there exists an open nhood U of . such that , U; then
, {.]. Since , {,], we have {.] = {,]. Conversely, suppose that . = ,
implies that {.] = {,]. Take any . = , in X and we show that there exists an
open nhood of one of the two points such that the other point is not in U. If
not, then , {.]; since {.] is closed, we have {,] {.]; similarly, {.] {,]. A
contradiction.
Now take any {.] = {,] in X, -. Then {.] = {.] = {,] = {,]. Hence, X, - is
T
0
. |L
13D. The Zariski Topology
For a polynomial 1 in n real variables, let 7(1) = {(.
1
. . . . . .
n
) R
n
:
1(.
1
. . . . . .
n
) = 0]. Let P be the collection of all such polynomials.
I Exercise 98. {7(1) : 1 P] is a base for the closed sets of a topology (the
Zariski topology) on R
n
.
Proof. Denote Z : {7(1) : 1 P]. If 7(1
1
) and 7(1
2
) belong to Z, then
7(1
1
) L 7(1
2
) = 7(1
1
1
2
) Z since 1
1
1
2
P. Further,
_
12P
7(1) =
since there are 1 P with 7(1) = (for instance, 1 = 1 X
2
1
X
2
n
).
It follows from Exercise 48 that Z is a base for the closed sets of the Zariski
topology on R
n
. |L
I Exercise 99. The Zariski topology on R
n
is T
1
but not T
2
.
Proof. To verify that the Zariski topology is T
1
, we show that every single-
ton set in R
n
is closed (by Theorem 13.4). For each (.
1
. . . . . .
n
) R
n
, dene a
polynomial 1 P as follows:
1 = (X
1
.
1
)
2
(X
n
.
n
)
2
.
Then 7(1) = {(.
1
. . . . . .
n
)]; that is, {(.
1
. . . . . .
n
)] is closed.
To see the Zariski topology is not T
2
, consider the R case. In R, the Zariski
topology coincides with the conite topology (see Exercise 100). It is well know
that the conite topology is not Hausdor (Example 13.5(a)). |L
I Exercise 100. On R, the Zariski topology coincides with the conite topology;
in R
n
, n > 1, they are dierent.
Proof. On R, every 7(1) is nite. So on R every closed set in the Zariski topol-
ogy is nite since every closed set is an intersection of some subfamily of Z.
However, if n > 1, then 7(1) can be innite: for example, consider the polyno-
mial X
1
X
2
(let X
1
= 0, then all X
2
R is a solution). |L
SECTION 5.2 REGULARITY AND COMPLETE REGULARITY 39
13H. Open Images of Hausdor Spaces
I Exercise 101. Given any set X, there is a Hausdor space Y which is the
union of a collection {Y
x
: . X] of disjoint subsets, each dense in Y .
Proof. |L
5.2 Regularity and Complete Regularity
Theorem 5.1 (Dugundji 1966). a. Let 1 : X Y be a closed map. Given any
subset S Y and any open U containing 1
1
(S), there exists an open V S
such that 1
1
(V ) U.
b. Let 1 : X Y be an open map. Given any subset S Y , and any closed
containing 1
1
S, there exists a closed T S such that 1
1
(T) .
Proof. It is enough to prove (a). Let V = Y 1(X U). Then
1
1
(S) U ==X U X 1
1
(S) = 1
1
(Y S)
==1(X U) 11
1
(Y S)|
==Y 11
1
(Y S)| V.
Since 11
1
(Y S)| Y S, we obtain
S = Y (Y S) Y 11
1
(Y S)| V :
that is, S V . Because 1 is closed, V is open in Y . Observing that
1
1
(V ) = X 1
1
1(X U)| X (X U) = U
completes the proof. |L
Theorem 5.2 (Theorem 14.6). If X is T
3
and is a continuous, open and closed
map of X onto Y , then Y is T
2
.
Proof. By Theorem 13.11, it is sucient to show that the set
:{(.
1
. .
2
) X X : (.
1
) = (.
2
)]
is closed in X X. If (.
1
. .
2
) , then .
1

1
(.
2
)|. Since a T
3
-space is
T
1
, the singleton set {.
2
] is closed in X; since is closed, {(.
2
)] is closed in
Y ; since is continuous,
1
(.
2
)| is closed in X. Because X is T
3
, there are
disjoint open sets U and V with
.
1
U. and
1
(.
2
)| V.
40 CHAPTER 5 SEPARATION AND COUNTABILITY
Since is closed, it follows from Theorem 5.1 that there exists open set W Y
such that {(.
2
)] W, and
1
(W) V ; that is,

1
(.
2
)|
1
(W) V.
Then U
1
(W) is a nhood of (.
1
. .
2
). We nally show that U
1
(W)|
= . If there exists (,
1
. ,
2
) such that (,
1
. ,
2
) U
1
(W), then ,
1

1
(,
2
)|
1
(W); that is, ,
1
U
1
(W). However, U V = and

1
(W) V imply that U
1
(W) = . A contradiction. |L
Definition 5.3. If X is a space and X, then X, denotes the quotient
space obtained via the equivalence relation whose equivalence classes are
and the single point sets {.], . X .
Theorem 5.4. If X is T
3
and Y is obtained from X by identifying a single closed
set in X with a point, then Y is T
2
.
Proof. Let be a closed subset of a T
3
-space X. Then X is an open subset
in both X and X, and its two subspace topologies agree. Thus, points in
X X, are dierent from | and have disjoint nhoods as X is Hausdor.
Finally, for . X , there exist disjoint open nhoods V(.) and W(). Their
images, (V ) and (W), are disjoint open nhoods of . and | in X,, because
V =
1
(V )| and W =
1
(W)| are disjoint open sets in X. |L
5.3 Normal Spaces
15B. Completely Normal Spaces
I Exercise 102. X is completely normal i whenever and T are subsets of
X with

T =

T = , then there are disjoint open sets U and V T.
Proof. Suppose that whenever and T are subsets of X with

T =

T =
, then there are disjoint open sets U and V T. Let Y X, and C. D Y
be disjoint closed subsets of Y . Hence,
= cl
Y
(C) cl
Y
(D) =

C Y |

D Y | =

C

D Y |.
Since D cl
Y
(D), we have

C D = . Similarly, C

D = . Hence there are
disjoint open sets U
0
and V
0
in X such that C U
0
and D V
0
. Let U = U
0
Y
and V = V
0
Y . Then U and V are open in Y , C U, and D V ; that is, Y is
normal, and so X is completely normal.
Now suppose that X is completely normal and consider the subspace Y :
X (

T). We rst show that . T Y . If Y , then there exists . with
. Y ; that is, .



T. But then .

T. A contradiction. Similarly for T.
In the normal space Y , we have
SECTION 5.4 COUNTABILITY PROPERTIES 41
cl
Y
() cl
Y
(T) =

Y |

T Y | = (



T) X (



T)| = .
Therefore, there exist disjoint open sets U cl
Y
() and V cl
Y
(T). Since
cl
Y
() and T cl
Y
(T), we get the desired result. |L
I Exercise 103. Why cant the method used to show every subspace of a reg-
ular space is regular be carried over to give a proof that every subspace of a
normal space is normal?
Proof. In the rst proof, if Y X is closed in Y and . Y , then there
must exists closed set T in X such that . T. This property is not applied if
{.] is replaced a general closed set T in Y . |L
I Exercise 104. Every metric space is completely normal.
Proof. Every subspace of a metric space is a metric space; every metric space
is normal Royden and Fitzpatrick (2010, Proposition 11.7). |L
5.4 Countability Properties
16A. First Countable Spaces
I Exercise 105. Every subspace of a rst-countable space is rst countable.
Proof. Let X. If . , then V is a nhood of . in i V = U , where
U is a nhood of . X (Theorem 6.3(d)). |L
I Exercise 106. A product

of rst-countable spaces is rst countable

i each X

is rst countable, and all but countably many of the X

are trivial
spaces.
Proof. If

is rst-countable, then each X

is rst countable since it is

homeomorphic to a subspace of

. If the number of the family of untrivial

sets {X

] is uncountable, then for .

the number of nhood bases is

uncountable. |L
I Exercise 107. The continuous image of a rst-countable space need not be
rst countable; but the continuous open image of a rst-countable space is rst
countable.
Proof. Let X be a discrete topological space. Then any function dened on X
is continuous.
Now suppose that X is rst countable, and is a continuous open map of
X onto Y . Pick an arbitrary , Y . Let .
1
(,), and U
x
be a countable
nhood base of .. If W is a nhood of ,, then there is a nhood V of . such that
42 CHAPTER 5 SEPARATION AND COUNTABILITY
(V ) W since is continuous. So there exists U U
x
with (U) W. This
proves that {(U) : U U
x
] is a nhood base of ,. Since {(U) : U U
x
] is |L
6
COMPACTNESS
6.1 Compact Spaces
17B. Compact Subsets
IExercise 108. A subset 1 of X is compact i every cover of 1 by open subsets
of X has a nite subcover.
Remark (Lee 2011, p. 94). To say that a subset of a topological space is com-
pact is to say that it is a compact space when endowed with the subspace
topology. In this situation, it is often useful to extend our terminology in the
following way. If X is a topological space and X, a collection of subsets of
X whose union contains is also called a cover of ; if the subsets are open
in X we sometimes call it an open cover of . We try to make clear in each
specic situation which kind of open cover of is meant: a collection of open
subsets of whose union is , or a collection of open subsets of X whose
union contains .
Proof. The only if part is trivial. So we focus on the if part. Let U be an
open cover of 1, i.e., U =
_
{U : U U]. For every U U, there exists an open
set V
U
in X such that U = V
U
1. Then {V
U
: U U] is an open cover of 1,
i.e., U
_
{V
U
: U U]. Then there exists a nite subcover, say V
U
1
. . . . . V
U
n
of {V
U
: U U], such that 1
_
n
iD1
V
U
i
. Hence, 1 =
_
n
iD1
(V
U
i
1); that is, 1
is compact. |L
I Exercise 109. The union of a nite collection of compact subsets of X is
compact.
Proof. Let and T be compact, and U be a family of open subsets of X
which covers L T. Then U covers and there is a nite subcover, say,
U

1
. . . . . U

n
of ; similarly, there is a nite subcover, say, U
B
1
. . . . . U
B
n
of T.
But then {U

1
. . . . . U

n
. U
B
1
. . . . . U
B
n
] is an open subcover of L T, so L T is
compact. |L
43
References
[1] Adamson, Iain T. (1996) A General Topology Workbook, Boston:
Birkhuser. [33]
[2] Ash, Robert B. (2009) Real Variables with Basic Metric Space Topology,
New York: Dover Publications, Inc. [22]
[3] Dugundji, James (1966) Topology, Boston: Allyn and Bacon, Inc. [39]
[4] Lee, John M. (2011) Introduction to Topological Manifolds, 202 of Grad-
uate Texts in Mathematics, New York: Springer-Verlag, 2nd edition. [25,
43]
[5] McCleary, John (2006) A First Course in Topology: Continuity and Di-
mension, 31 of Student Mathematical Library, Providence, Rhode Island:
American Mathematical Society. [29]
[6] Royden, Halsey and Patrick Fitzpatrick (2010) Real Analysis, New
Jersey: Prentice Hall, 4th edition. [41]
[7] Willard, Stephen (2004) General Topology, New York: Dover Publica-
tions, Inc. [i]
45