April 21, 2006

Munkres 29
Ex. 29.1. Closed intervals [a, b] Q in Q are not compact for they are not even sequentially
compact [Thm 28.2]. It follows that all compact subsets of Q have empty interior (are nowhere
dense) so Q can not be locally compact.
To see that compact subsets of Q are nowhere dense we may argue as follows: If C Q is
compact and C has an interior point then there is a whole open interval (a, b) Q C and also
[a, b] Q C for C is closed (as a compact subset of a Hausdor space [Thm 26.3]). The closed
subspace [a, b] Q of C is compact [Thm 26.2]. This contradicts that no closed intervals of Q are
compact.
Ex. 29.2.
(a). Assume that the product

X

is locally compact. Projections are continuous and open [Ex

16.4], so X

is locally compact for all [Ex 29.3]. Furthermore, there are subspaces U C such
that U is nonempty and open and C is compact. Since

(U) = X

for all but nitely many ,

also

(C) = X

for all but nitely many . But C is compact so also

(C) is compact.
(b). We have

X

= X
1
X
2
where X
1
is a nite product of locally compact spaces and X
2
is
a product of compact spaces. It is clear that nite products of locally compact spaces are locally
compact for nite products of open sets are open and all products of compact spaces are compact
by Tychono. So X
1
is locally compact. X
2
is compact, hence locally compact. Thus the product
of X
1
and X
2
is locally compact.
Conclusion:

X

is locally compact if and only if X

is locally compact for all and compact

for all but nitely many .
Example: R

and Z

+
are not locally compact.
Ex. 29.3. Local compactness is not preserved under continuous maps. For an example, let
S R
2
be the graph of sin(1/x), x (0, 1]. The space {(0, 0)}S is not locally compact at (0, 0):
Any neighborhood U of (0, 0) contains an innite subset without limit points, the intersection of
S and a horizontal straight line, so U can not [Thm 28.1] be contained in any compact subset of S.
On the other hand, {(0, 0)} S is the image of a continuous map dened on the locally compact
Hausdor space {1} (0, 1] [Thm 29.2].
Local compactness is clearly preserved under open continuous maps as open continuous maps
preserve both compactness and openness.
Ex. 29.4 (Morten Poulsen). Let d denote the uniform metric. Suppose [0, 1]

is locally
compact at 0. Then 0 U C, where U is open and C is compact. There exists > 0 such that
B
d
(0, ) U. Note that A = {0, /3}

B
d
(0, ), hence A C. By theorem 28.2 A has a limit
point in C, contradicting Ex. 28.1.
Ex. 29.5 (Morten Poulsen).
Lemma 1. A homeomorphism between locally compact Hausdor spaces extends to a homeomor-
phism between the one-point compactications. In other words, homeomorphic locally compact
Hausdor spaces have homeomorphic one-point compactications.
Proof. Let f : X
1
X
2
be a homeomorphism between locally compact Hausdor spaces. Fur-
thermore let X
1
= X
1
{
1
} and X
2
= X
2
{
2
} denote the one-point compactications.
Dene

f : X
1
X
2
by

f(x) =

f(x), x X
1

2
, x =
1
.
Note that

f is bijective. Recall that for a locally compact Hausdor space X the topology on
the one-point compactication, X, is the collection
{ U | U X open} { X C |C X compact },
c.f. the proof of theorem 29.1.
1
2
If U X
2
is open then

f
1
(U) = f
1
(U) is open in X
1
. If C X
2
is compact then

f
1
(X
2
C) =

f
1
(X
2
)

f
1
(C) = X
1
f
1
(C) is open in X
1
, since f
1
(C) X
1
is
compact. It follows that

f is continuous, hence a homeomorphism, by theorem 26.6.
Finally note that the converse statement does not hold: If X
1
= [0, 1/2)(1/2, 1] and X
2
= [0, 1)
then X
1
= [0, 1] = X
2
. But X
1
and X
2
are not homeomorphic, since X
1
is not connected and
X
2
is connected.
Ex. 29.6 (Morten Poulsen). Let S
n
denote the unit sphere in R
n+1
. Let p denote the point
(0, . . . , 0, 1) R
n+1
.
Lemma 2. The punctured sphere S
n
p is homeomorphic to R
n
.
Proof. Dene f : (S
n
p) R
n
by
f(x) = f(x
1
, . . . , x
n+1
) =
1
1 x
n+1
(x
1
, . . . , x
n
).
The map f is also known as stereographic projection. It is straightforward to check that the
map g : R
n
(S
n
p) dened by
g(y) = g(y
1
, . . . , y
n
) = (t(y) y
1
, . . . , t(y) y
n
, 1 t(y)),
where t(y) = 2/(1 +y
2
), is a right and left inverse for f.
Theorem 3. The one-point compactication of R
n
is homeomorphic to S
n
.
Proof. By the preceding lemma R
n
is homeomorphic to S
n
p. The one-point compactication
of S
n
p is clearly S
n
. Now the result follows from Ex. 29.5.
Ex. 29.7. Let X be any linearly ordered space with the least upper bound property. As [a, b] =
[a, b) {b} is compact Hausdor [Thm 27.1, Thm 17.11], the right half-open interval [a, b) is
locally compact Hausdor and its Alexandro compactication is [a, b] [Thm 29.1]. Apply this
to S

= [1, ) S

= [1, ]. (Apply also to Z

+
= [1, ) Z
+
Z
+
where = 2 1 for (an
alternative answer to) [Ex 29.8])
Also X itself is locally compact Hausdor [Thm 17.11] as all closed and bounded intervals in
X are compact [Thm 27.1]. Is the one-point compactication of X a linearly ordered space?
Ex. 29.9. This follows from Ex 29.3 for the quotient map G G/H is open [SupplEx 22.5.(c)].
Ex. 29.11. It is not always true that the product of two quotient maps is a quotient map
[Example 7, p. 143] but here is a case where it is true.
Lemma 4 (Whitehead Theorem). [1, 3.3.17] Let p: X Y be a quotient map and Z a locally
compact space. Then
p 1: X Z Y Z
is a quotient map.
Proof. Let A X Z. We must show: (p 1)
1
(A) is open A is open. This means that for
any point (x, y) (p1)
1
(A) we must nd a saturated neighborhood U of x and a neighborhood
V of y such that U V (p 1)
1
(A).
Since (p 1)
1
(A) is open in the product topology there is a neighborhood U
1
of x and a
neighborhood V of y such that U
1
V (p 1)
1
(A). Since Y is locally compact Hausdor
we may assume [Thm 29.2] that V is compact and U
1
V (p 1)
1
(A). Note that also
p
1
(pU
1
) V is contained in (p 1)
1
(A). The tube lemma [Lemma 26.8] says that each point
of p
1
(pU
1
) has a neighborhood such that the product of this neighborhood with V is contained
in the open set (p 1)
1
(A). Let U
2
be the union of these neighborhoods. Then p
1
(pU
1
) U
2
and U
2
V (p 1)
1
(A). Continuing inductively we nd open sets U
1
U
2
U
i

U
i+1
such that p
1
(pU
i
) U
i+1
and U
i+1
V (p 1)
1
(A). The open set U =

U
i
is
saturated because U p
1
(pU) =

p
1
(pU
i
)

U
i+1
= U. Thus also U V is saturated and
U V

U
i
V (p 1)
1
(A).
3
Example: If p: X Z is a quotient map, then also p id: X [0, 1] Z [0, 1] is a quotient
map. This fact is important for homotopy theory.
Theorem 5. Let p: A B and q : C D be quotient maps. If B and C are locally compact
Hausdor spaces then p q : AC B D is a quotient map.
Proof. The map p q is the composition
AC
p1
//
B C
1q
//
B D
of two quotient maps and therefore itself a quotient map [p. 141].
References
[1] Ryszard Engelking, General topology, second ed., Sigma Series in Pure Mathematics, vol. 6, Heldermann Verlag,
Berlin, 1989, Translated from the Polish by the author. MR 91c:54001