Munkers Soln Selected
Munkers Soln Selected
Munkers Soln Selected
Amit Rajaraman
Last updated March 31, 2021
Contents
Let T be the topology generated by A and T 0 be a topology that contains A. Let U ∈ T . Then U = i∈I Bi
S
for some (Bi )i∈I in A. However, each Bi is also in T 0 . Since an arbitrary union of open sets is open, U ∈ T 0 as
well. Therefore, T ⊆ T 0 , proving the result. The solution for the case where A is a subbasis is very similar and
so omitted.
T = {U ∩ A : U open in X}
= {(U ∩ Y ) ∩ A : U open in X}
= {V ∩ A : V open in Y },
It is easily seen that TY0 is finer than TY . We further see that it need not be strictly finer by considering the
example X = {a, b, c}, Y = {a, b}, T = {∅, X, {a}, {b}, {a, b}}, and T 0 as the discrete topology on X.
Exercise 2.16.3. Consider Y = [−1, 1] as a subspace of R. Which of the following is open in Y ? Which are open
in R?
1
A= x: <x<1
2
1
B= x: <x≤1
2
1
C= x: ≤x<1
2
1
D= x: ≤x≤1
2
E = x : 0 < x < 1 and 1/x 6∈ Z+
Munkres Solutions 3 / 13 Amit Rajaraman
Solution
A and B are open in Y and only A is open in R. This is reasonably straightforward to prove.
C is not open in Y (and so not R either) because there is no basis element U of the order topology such that
1/2 ∈ U ⊆ C. A similar argument holds for D as well.
E is open in both R and Y because it can be written as a union of basis elements
\ 1 1
E= , .
+
n+1 n
n∈Z
Exercise 2.16.4. A map f : X → Y is said to be an open map if for every open U of X, f (U ) is open in Y . Show
that π1 : X × Y → X and π2 : X × Y → Y are open.
Solution
We shall only show that π1 is open, the other case is nearly identical. Let
[
U= Ui × Vi
i∈I
be open in X × Y for some indexing set I, where each Ui and Vi are open in X and Y respectively. Then,
!
[ [ [
π1 (U ) = π1 Ui × Vi = π1 (Ui × Vi ) = Ui
i∈I i∈I i∈I
is open in X.
Solution
for some C ∈ C. Closure under finite intersections is shown similarly. We trivially have ∅, X ∈ T because
X, ∅ ∈ C.
Solution
Let U be open in X such that Y \ A = U ∩ Y . Then, we can write A as X \ ((X \ Y ) ∪ U ). Since X \ Y and U
are open in X, A is closed in X.
Solution
Munkres Solutions 4 / 13 Amit Rajaraman
Observe that
(X × Y ) \ (A × B) = ((X \ A) × (Y \ B)) ∪ ((X \ A) × Y ) ∪ (A × (Y \ B)).
Since each of the sets on the right are open in X × Y , A × B is closed.
Exercise 2.17.4. Show that if U is open in X and A is closed in X, U \ A is open in X and A \ U is closed in Y .
Solution
Solution
Let us first show the forward direction. Let (x, y) ∈ (X × X) \ ∆. Then since X is Hausdorff, there are some
open U1 3 x, U2 3 y such that U1 ∩ U2 = ∅. Since the open set U1 × U2 3 (x, y) does not intersect ∆, ∆ is
closed (no other point is in its closure).
For the backward direction, this means that for any (x, y) ∈ X ×X with x 6= y, there exists an open set U 3 (x, y)
such that U ∩ ∆ = ∅. We may further assume that U is a basis element of the product topology and can be
written as U1 × U2 . However, (U1 × U2 ) ∩ ∆ = ∅ iff U1 ∩ U2 = ∅, so the required follows.
Exercise 2.17.14. In the finite complement topology on R, what point or points does the sequence xn = 1/n
converge to?
Solution
We claim that xn converges to any point in R. Let x ∈ R and U 3 x be open. U contains all but finitely many
xi since we are in the finite complement topology and therefore, xn converges to x.
Exercise 2.17.19. If A ⊆ X, define the boundary of A by
Bd A = A ∩ X \ A.
(a) Show that A and Bd A are disjoint, and A = A◦ ∪ Bd A.
◦
Solution
(a) Let x ∈ A \ A◦ . Then for any open U 3 x, U 6⊆ A (otherwise, A◦ ∪ U ) A◦ is open and contained in A).
That is, U ∩ (X \ A) 6= ∅. However, this implies that x ∈ X \ A, that is, A \ A◦ ⊆ X \ A. Therefore,
A \ A◦ = (A \ A) ∪ (A \ A◦ ) ⊆ X \ A
A ⊆ A◦ ∪ X \ A
= A ∩ (A◦ ∪ X \ A)
= A◦ ∪ (A ∪ X \ A) = A◦ ∪ Bd A.
Solution
Exercise 2.18.8. Let Y be an ordered set in the order topology. Let f, g : X → Y be continuous.
(a) Show that the set {x : f (x) ≤ g(x)} is closed in X.
(b) Let h : X → Y be the function
h(x) = min{f (x), g(x)}.
Show that h is continuous.
Solution
(a) It suffices to show that V = {x : f (x) > g(x)} is open in X. Let x ∈ V . Since Y is Hausdorff, there are
open sets U1 , U2 such that f (x) ∈ U1 , g(x) ∈ U2 , and U1 ∩ U2 = ∅.
We can consider a basis element B1 ⊆ U1 that contains f (x) in lieu of U1 . That is, we may suppose wlog
that U1 and U2 are disjoint basis elements. Further, we may assume that U1 is of the form (a, ∞) (if it is of
the form (a, c) instead, we can replace it with (a, ∞)). Due to the disjointness assumption, this means that
we can consider U1 to be of the form (a, ∞) and U2 of the form (−∞, b) such that for any c ∈ U1 , d ∈ U2 ,
c>d
Now, let U = f −1 (U1 ) ∩ g −1 (U2 ) 3 x. f and g are continuous so U is open. Further, for any y ∈ U ,
f (y) > g(y) (by the above assumption), that is, U ⊆ V .
It follows that V is open (for any x ∈ V , there is an open V ⊆ U such that x ∈ V ).
(b) Let U1 = {x : f (x) ≥ g(x)} and U2 = {x : f (x) ≤ g(x)}. By (a), U1 and U2 are both closed. Since g is
continuous on U1 , f is continuous on U2 , and f = g on U1 ∩ U2 , we can use the pasting lemma to conclude
that h is continuous (h(x) = g(x) on U1 and f (x) on U2 ).
Solution
Let the two topological spaces above be denoted by X and Y respectively. Consider the map f : X → Y with
Solution
We first show the backward direction. Suppose (πα (xi )) converges to πα (x). LetQ U be an open set containing x
and B ⊆ U be a basis element of the product topology containing x. Let B = α Uα where Uα 6= Xα for the
finite set {α1 , . . . , αn }. Since (παj (xi )) converges to παj (x) for each j, each Uαj contains all but finitely many
πα (xi ). As a result, B contains all but finitely many xi and therefore, (xi ) converges to x.
On the other hand, let (xi ) converge to x. Let Uα beQ an open set in Xα containing πα (x). We wish toQ show that
it contains all but finitely many πα (xi ). Let U 0 = Uβ0 be a basis element containing x and V = Vβ , where
Vβ = Uβ0 for β 6= α and Vα = Uα ∩ Uα0 . Since V is an open set containing x, it contains all but finitely many xi .
In particular, Vα contains all but finitely many πα (xi ). The required follows.
Observe that the forward proof works even if we use the box topology instead, but the backward direction breaks.
To see that the result need not hold for the box, let the product space be Rω , πn (xi ) = n/i and πn (x) = 0 for
each i, n.
Exercise 2.19.7. Let R∞ be the subset of Rω consisting of all sequences that are eventually 0 (xi 6= 0 for finitely
many i). What is the closure of R∞ in Rω in the box and product topologies?
Solution
We claim that the closure under the product topology is Rω . Let x ∈ Rω and U = Un be a basis element
Q
containing x. We wish to determine when U ∩ R∞ 6= ∅. Consider y ∈ R∞ such that yn = xn if Un 6= R and 0
otherwise. Then y ∈ U ∩ R∞ , thus proving the result.
For the box topology, we claim that R∞ is closed. Indeed, for any x ∈ Rω \ R∞ , consider the open set U = Un ,
Q
where Un = (xn /2, 3xn /2) if xn 6= 0 and R otherwise. Then U ∩ R∞ is empty, completing the proof.
(b) Let X 0 denote a space having the same underlying set as X. Show that if d : X 0 × X 0 → R is continuous, the
topology of X 0 is finer than the topology of X.
This means that if X has a metric d, the metric topology induced by it is the coarsest topology with respect to which
d is continuous.
Solution
(a) Let x = (x1 , x2 ) ∈ X × X and for some ε > 0, U = B(f (x), ε) be a basis element of R containing f (x).
Consider the open sets U1 = Bd (x1 , ε/4) and U2 = Bd (x2 , ε/4). If y1 ∈ U1 and y2 ∈ U2 , then
Solution
We claim that the closure of R∞ is the set of all sequences that converge to 0. Let x ∈ Rω .
Case 1. The sequence x does not converge to 0. There is then some ε > 0 such that for infinitely many n,
|xn | > ε. We may assume that ε < 1. Consider the open set U = Bρ (x, ε). We claim that U ∩ R∞ = ∅.
Indeed, for any y ∈ R∞ , |xn − y| > ε for infinitely many n so y 6∈ Bd (x, ε), thus proving that x 6∈ R∞ .
Case 2. The sequence x converges to x. Let Bρ (x, ε) be an arbitrary basis element containing x. Then
there exists N ∈ N such that for all n > N , |xn | < ε/2. Consider the element y ∈ R∞ such that yn = xn
for n ≤ N and yn = 0 otherwise. Then y ∈ Bρ (x, ε), thus proving that x ∈ R∞ .
Exercise 2.20.6. Let ρ be the uniform metric on Rω . Given x = (x1 , x2 , . . .) ∈ Rω and 0 < ε < 1, let
Solution
(a) Consider
x0 = (x1 + ε/2, x2 + 2ε/3, . . . , xn + nε/(n + 1), . . .).
Obviously, x0 ∈ U (x, ε). However, ρ(x, x0 ) = sup{nε/(n + 1) : n ∈ N} = ε, so x0 6∈ Bρ (x, ε).
(b) Let x0 be as in the previous part. Suppose that U (x, ε) is open in the uniform topology. Then, there is a
δ > 0 such that Bρ (x0 , δ) ⊆ U (x, ε). In particular, (x0n − δ, x0n + δ) ⊆ (x − ε, x + ε) for any n. This yields
a contradiction since it would imply that δ < ε/(n + 1) for any n and therefore, U (x, ε) is not open in the
uniform topology.
(c) Let y ∈ U (x, δ) for some δ < ε. Then ρ(x, y) ≤ δ < ε, so y ∈ Bρ (x, ε). One direction of the inclusion follows.
For the other, let y ∈ Bρ (x, ε) and δ = ρ(x, y). Since δ < ε, we can choose a δ 0 such that δ < δ 0 < ε. Then,
y ∈ U (x, δ 0 ), thus proving the result.
Solution
Let x ∈ X and V = BdY (f (x), ε) be a basis element in Y containing f (x). Then U = BdX (x, ε) is an open
set in X containing x such that f (U ) ⊆ V . Therefore, f is continuous. Showing that f is open is similarly
straight-forward.
Exercise 2.21.3. Let Xn be a metric space with metric dn for each n ∈ Z+ . Let di = min{di , 1}. Show that
Solution
Q
Let U = Ui be a basis element of the product topology, where Ui 6= Xi for i = α1 , . . . , αn . Let x ∈ U . For
each j, let Bdα (xαj , εj ) ⊆ Uαj . Let ε = min{εj /2αj }. Then x ∈ BD (x, ε) ⊆ U , so the metric topology is finer
j
than the product topology.
For the other direction, let ε > 0 and U = BD (x, ε). WeQwant to show that there is an element of the product
topology containing x that is contained in U . Let V = Vi , where Vi = Bdi (xi , iε/2) for each i. Note that V
is a basis element of the product topology since Vi = Xi for i > 2/ε. Since x ∈ V ⊆ U , the product topology is
finer than the metric topology, thus proving the result.
(a) Let p : X → Y be a continuous map. Show that if there is a continuous map f : Y → X such that p ◦ f equals
the identity map on Y , then p is a quotient map.
(b) If A ⊆ X, a retraction of X onto A is a continuous map r : X → A such that r(a) = a for all a ∈ A. Show that
a retraction is a quotient map.
Solution
(a) It is obvious that p is surjective since for any y ∈ Y , p(f (y)) = y. Let U ⊆ Y . We want to show that if
p−1 (U ) is open, then so is U . Indeed, we then have U = (p ◦ f )−1 (U ) = f −1 (p−1 (U )) is open since p−1 (U )
is open and f is continuous.
(b) It is clear that r is surjective. Let U ⊆ A. We want to show that if r−1 (U ) is open, then U is open in A.
Indeed, note that U = r−1 (U ) ∩ A, so is open in A by definition.
Exercise 2.22.3. Let π1 : R × R → R be the projection on the first coordinate. Let A be the subspace of R × R
consisting of all points x × y for which x ≥ 0 or y = 0. Let q : A → R be obtained by restricting π1 . Show that q is
a quotient map that is neither open nor closed.
Solution
Exercise
S 3.23.2. Let {An } be a sequence of connected subspaces of X such that An ∩ An+1 6= ∅ for all n. Show
that An is connected.
Solution
S
Let A = An . Suppose instead that A = U ∪ V is a separation. Since A1 is connected, we may assume without
loss of generality that A1 ⊆ U . We shall now show by induction that for any n, An ⊆ U . Indeed, if An ⊆ U
for some n ≥ 1, then since An+1 is connected, An+1 ⊆ U or An+1 ⊆ V . However, ∅ 6= An ∩ An+1 ⊆ U , so
An+1 ⊆ U as well. This contradicts the non-emptiness of V , completing the proof.
Solution
Denote the space by Y . Suppose instead that Y = U ∪ V is a separation. Since A is connected, we may assume
wlog that A ⊆ U . For any α, Aα ⊆ U or Aα ⊆ V . However, A ∩ Aα 6= ∅ so Aα ⊆ U . This contradicts the
non-emptiness of V , proving the result.
Exercise 3.23.4. Show that if X is an infinite set, it is connected in the finite complement topology.
Solution
Solution
Let S be the set of all bounded sequences. Then it is not too difficult to show that both S and Rω \ S are open,
so Rω is disconnected.
Exercise 3.23.9. Let A ( X and B ( Y . If X and Y are connected, show that (X × Y ) \ (A × B) is connected.
Solution
Exercise 3.23.11. Let p : X → Y be a quotient map. Show that if each p−1 ({y}) is connected and Y is connected,
then X is connected.
Munkres Solutions 10 / 13 Amit Rajaraman
Solution
Suppose otherwise and let X = U ∪ V be a separation. For any y, either p−1 ({y}) ⊆ U or p−1 ({y}) ⊆ Y (due
to connectedness). That is, U and V are saturated. But then, Y = p(X) = p(U ) ∪ p(V ) is a separation of Y ,
yielding a contradiction and proving the result.
Solution
Consider the continuous map g : S 1 → R given by x 7→ f (x) − f (−x). Choose an arbitrary x ∈ S1 . If g(x) = 0,
then we are done. Otherwise, let y = −x be the diametrically opposite point to x. Then g(y) = −g(x). Since
S 1 is connected and 0 lies between g(x) and g(y), it follows that there exists some x0 ∈ S 1 such that g(x0 ) = 0.
Exercise 3.24.3. Let f : X → X be continuous. Show that if X = [0, 1], there is a point x such that f (x) = x.
What happens if X = [0, 1) or (0, 1]?
Solution
Consider the continuous map g : X → R given by x 7→ f (x) − x. Observe that g(0) ≥ 0 or g(1) ≤ 0. If equality
holds at either place, we are done. Otherwise, 0 lies between them. Since X is connected, the claim follows.
The claim need not hold if X = [0, 1) or (0, 1]. Indeed, consider the functions x 7→ (x + 1)/2 and x 7→ x/2.
Solution
It is obvious that any singleton in R` is connected and path-connected. We claim that these are the only
non-empty connected (and path-connected) subspaces. Let A ⊆ R` have at least two elements. Let x, y ∈ A
with x < y. Then (−∞, y) ∩ A and [y, ∞) ∩ A forms a separation of A, so it is not connected, and thus not a
component. Therefore, the components of R` are the singletons.
Recall that any continuous maps connected subspaces to connected subspaces. In particular, any continuous
map R → R` maps R to a connected subspace of R` . However, this must be a singleton and therefore, the
continuous maps R → R` are the singletons.
Solution
(b) Show that if X is compact Hausdorff under both T and T 0 , either T and T 0 are equal or they are not comparable.
Solution
(a) If X is compact under T 0 , then it is compact under T . Indeed, any open cover under T is an open cover
under T 0 , and compactness implies the existence of a finite subcover.
The converse need not hold, as can be seen with the example of X = R, T 0 as the discrete topology, and T
as the indiscrete topology.
(b) Suppose T ⊆ T 0 . Let U ∈ T 0 . Then X \ U is closed, and thus compact under T 0 . By a method similar
to part (a), X \ U is also compact under T , and thus closed in T . This implies that U is open in T , and
therefore T 0 ⊆ T , proving the claim.
Exercise 3.26.2.
(a) Show that in the finite complement topology on R, every subspace is compact.
(b) If R has the topology consisting of all sets A such that R \ A is either countable or all of R, is [0, 1] a compact
space?
Solution
(a) Let A ⊆ R and A be an open cover of A. Let U ∈ A. Then (R \ U ) ∩ A is finite, suppose it is equal to
{x1 , . . . , xn }. For each 1 ≤ i ≤ n, let Ui ∈ A such that xi ∈ Ui . Then {U, U1 , . . . , Un } forms a finite subcover
of A.
(b) No, consider the open cover
Solution
S
Let U1 , . . . , Un be compact subspaces of X, U = 1≤i≤n Ui , and A be an open cover of U . For each 1 ≤ i ≤ n,
let Ai be aSfinite subcover of A of Ui (such a subcover exists because Ui is compact and A is an open cover of
Ui ). Then 1≤i≤n Ai forms a finite subcover of U , proving the required.
Exercise 3.26.5. Let A and B be disjoint compact subspaces of the Hausdorff subspace X. Show that there exist
disjoint open subsets U and V containing A and B respectively.
Solution
For each a ∈ A, let Ua and Va be disjoint open subsets containing {a} and B respectively.
S T {Ua : a ∈ A}
Then
forms an open cover of A, so has a finite subcover, say {Ua1 , . . . , Uan }. Then 1≤i≤n Uai and 1≤i≤n Vai are
disjoint open subsets containing A and B respecitvely, completing the proof.
Exercise 3.26.6. Show that if f : X → Y is continuous, where X is compact and Y is Hausdorff, then f is a closed
map.
Solution
Let A be a closed subspace of X. Then A is compact, and since f is continuous, so is f (A). Recall that compact
subspaces of Hausdorff spaces are closed. Therefore, f (A) is closed, so f is closed.
Munkres Solutions 12 / 13 Amit Rajaraman
Solution
T
(a) Let {x} be a one-point set in X and B a countable basis at x. We claim that U := B∈B B = {x}. Let
y ∈ X be distinct from x. Since X is T1 , there is an open V such that x ∈ U 63 y. Letting B be a basis
element such that x ∈ B ⊆ U , we see that y 6∈ B. Therefore, y 6∈ W and W = {x}.
(b)
Exercise 4.30.2. Show that if X has a countable basis (Bn ), then every basis C of X contains a countable basis for
X.
Solution
For each m, n, if possible, choose Cn,m ∈ C such that Bn ⊆ Cn,m ⊆ Bm . We claim that this (countable) subset
of C is a basis. Let x ∈ X and Bm a basis element that contains x. Since C is a basis, there is some C ∈ C such
that x ∈ C ⊆ Bm . Since (Bn ) is a basis, there exists n such that x ∈ Bn ⊆ C. Now, since Bn ⊆ C ⊆ Bm , we
have Cn,m such that x ∈ Cn,m ⊆ Bm , proving the claim.
Exercise 4.30.3. Let X have a countable basis; let A be an uncountable subset of X. Show that uncountably many
points of A are limit points of A.
Solution
Let A0 be the set of limit points of A in A and suppose instead that it is countable. For each x ∈ A \ A0 , let Bx
be a basis element such that x ∈ Bx and Bx ∩ A = {x}. It follows that for x 6= y, Bx 6= By . But since A \ A0 is
countable, this contradicts the second countability of X, proving the claim.
Exercise 4.30.4. Show that every compact metrizable space X has a countable basis.
Solution
S
Let An be a finite covering of XS by open (1/n)-balls – such a covering exists because X = x∈X B(x, 1/n).
Consider the countable set B = n∈N An . We claim that B is a basis of X. Let x ∈ X and δ > 0. We wish to
show that there is some B ∈ B such that x ∈ B ⊆ B(x, δ). Fix a N > 2/δ and let B ∈ AN contain x. Then
x ∈ B ⊆ B(x, δ). Indeed, for any y ∈ B, d(y, x) ≤ 2/N < δ, so y ∈ B(x, δ).
Exercise 4.30.5.
(a) Show that every metrizable space with a countable dense subset has a countable basis.
(b) Show that every metrizable Lindelöf space has a countable basis.
Solution
(a) Let X be a metrizable space and D a countable dense subset. Let B = {B(x, 1/n) : n ∈ N, x ∈ D}. It is not
too difficult to show that B is a countable basis similar to the solution to the previous question.
Munkres Solutions 13 / 13 Amit Rajaraman
(b) This is identical to the previous question, except that An is a countable covering of X by open (1/n)-balls.
Exercise 4.30.9. Let A be a closed subspace of X. Show that if X is Lindelöf, A is Lindelöf. Show by example
that if X has a countable dense subset, A need not have a countable dense subset.
Solution
Let A be an open cover of A. We must show that it has a countable subcover. Consider the open cover
A0 = {U ∪ (X \ A) : U ∈ A} of X (Why are these sets open?). Then A0 has a countable subcover, say S 0 , which
gives a countable subcover S = {V ∩ A : V ∈ S 0 } of A.
For the second part, consider X = R` × R` and A = {x × (−x) : x ∈ R` }.
Exercise 4.30.10. Show that if X is a countable product of spaces having countable dense subsets, then X has a
countable dense subset.
Solution
Let (Xn ) be spaces having countable dense subsets (An ). For each n, fix an arbitrary xn ∈ Xn . Consider the
subset A of X defined by
[ nY o
A= Un : Un = An for finitely many n and is {xn } otherwise .