1st December 2004

Munkres 31
Ex. 31.1 (Morten Poulsen). Let a and b be distinct points of X. Note that X is Hausdorff,
since X is regular. Thus there exists disjoint open sets A and B such that a A and b B. By
lemma 31.1(a) there exists open sets U and V such that
a U U A and b V V B.
Clearly U V = .
Ex. 31.2 (Morten Poulsen). Let A and B be disjoint closed subsets of X. Since X normal
there exists disjoint open sets U0 and U1 such that A U0 and B U1 . By lemma 31.1(b) there
exists open sets V0 and V1 such that
A V0 V0 U0 and B V1 V1 U1
Clearly U V = .
Ex. 31.3 (Morten Poulsen).
Theorem 1. Every order topology is regular.
Proof. Let X be an ordered set. Let x X and let U be a neighborhood of x, may assume
U = (a, b), a < b . Set A = (a, x) and B = (x, b). Using the criterion for regularity in
lemma 31.1(b) there are four cases:
(1)
(2)
(3)
(4)

If
If
If
If

u A and v B then x (u, v) (u, v) [u, v] (a, b).

Thus X is regular.

Ex. 31.5. The diagonal Y Y is closed as Y is Hausdorff [Ex 17.13]. The map (f, g) : X
Y Y is continuous [Thm 18.4, Thm 19.6] so
{x X | f (x) = g(x)} = (f, g)1 ()
is closed.
Ex. 31.6. Let p : X Y be closed continuous surjective map. Then X normal Y normal.
For this exercise and the next we shall use the following lemma from [Ex 26.12].
Lemma 2. Let p : X Y be a closed map.
(1) If p1 (y) U where U is an open subspace of X, then p1 (W ) U for some neighborhood
W Y of y.
(2) If p1 (B) U for some subspace B of Y and some open subspace U of X, then p1 (W )
U for some neighborhood W Y of B.
Proof. Note that




p1 (W ) U p(x) W x U x 6 U p(x) 6 W p(X U ) Y W
p(X U ) W =
(1) The point y does not belong to the closed set p(X U ). Therefore a whole neighborhood
W Y of y is disjoint from p(X U ), i.e. p1 (W ) U .
S
(2) Each point y B has a neighborhood Wy such that p1 (Wy ) U . The union W = Wy is
then a neighborhood of B with p1 (W ) U .

1

Since points are closed in X and p is closed, all points in p(X) are closed. All fibres p1 (y) X
are therefore also closed. Let y1 and y2 be two distinct points in Y . Since X is normal we
can separate the disjoint closed sets p1 (y1 ) and p1 (y1 ) by disjoint neighborhoods U1 and U2 .
Using Lemma 2.(1), choose neighborhoods W1 of y1 and W2 of y2 such that p1 (W1 ) U1 and
p1 (W2 ) U2 . Then W1 and W2 are disjoint. Thus Y is Hausdorff.
Essentially the same argument, but now using Lemma 2.(2), shows that we can separate disjoint
closed sets in Y by disjoint open sets. Thus Y is normal.
Alternatively, see [Lemma 73.3].
Example: If X is normal and A X is closed, then the quotient space X/A is normal.
Ex. 31.7. Let p : X Y be closed continuous surjective map such that p1 (y) is compact for
each y Y (a perfect map).
(a). X Hausdorff Y Hausdorff.
Let y1 and y2 be two distinct points in Y . By an upgraded version [Ex 26.5] of [Lemma 26.4] we
can separate the two disjoint compact subspaces p1 (y1 ) and p1 (y2 ) by disjoint open subspaces
U1 p1 (y1 ) and U2 p1 (y2 ) of the Hausdorff space X. Choose (Lemma 2) open sets W1 3 y1
and W2 3 y2 such that p1 (W1 ) U1 and p1 (W2 ) U2 . Then W1 and W2 are disjoint. This
shows that Y is Hausdorff as well.
(b). X regular Y regular.
Y is Hausdorff by (a). Let C Y be a closed subspace and y Y a point outside C. It is enough
to separate the compact fibre p1 (y) X and the closed set p1 (C) X by disjoint open set.
(Lemma 2 will provide open sets in Y separating y and C.) Each x p1 (y) can be separated by
disjoint open sets from p1 (C) since X is regular. Using compactness of p1 (y) we obtain (as in
the proof [Thm 26.3]) disjoint open sets U p1 (y) and V p1 (C) as required.
(c). X locally compact Y locally compact [1, 3.7.21].
Using compactness of p1 (y) and local compactness of X we construct an open subspace U X
and a compact subspace C X such that p1 (y) U C. In the process we need to know
that a finite union of compact subspaces is compact [Ex 26.3]. By Lemma 2, there is an open set
W 3 y such that p1 (y) p1 (W ) U C. Then y W p(C) where p(C) is compact [Thm
26.5]. Thus Y is locally compact.
(d). X 2nd countable Y 2nd countable.
Let {Bj }jZ+ be countable basis for X. For each finite subset J ZS
+ , let UJ X be the union of
all open sets of the form p1 (W ) with open W Y and p1 (W ) jJ Bj . There are countably
many open sets UJ . The image p(UJ ) is a union of open
S sets in Y , hence open. Let now V Y
be any open subspace. The inverse image p1 (V ) = yV p1 (y) is a union of fibres. Since each
S
fibre p1 (y) is compact, it can be covered by a finite union jJ(y) Bj of basis sets contained in
S
p1 (V ). By Lemma 2, there is an open set W Y such that p1 (y) p1 (W ) jJ(y) Bj .
S
Taking the union of all these open sets W , we get p1 (y) UJ(y) jJ(y) Bj p1 (V ). We
S
S
now have p1 (V ) = yV UJ(y) so that V = pp1 (V ) = yV p(UJ(y) ) is a union of sets from the
countable collection {p(UJ )} of open sets. Thus Y is 2nd countable.
Example: If Y is compact, then the projection map 2 : X Y Y is perfect. (Show that
2 is closed!)
Ex. 31.8. It is enough to show that p : X G\X is a perfect map [Ex 31.6, Ex 31.7]. We show
that
(1) The saturation GA of any closed subspace A X is closed. (The map p is closed.)
(2) The orbit Gx of any point x X is compact. (The fibres p1 (Gx) = Gx are compact.)
S
(1) Let y X be any point outside GA = gG gA. For any g G, g 1 y is outside the closed set
A X. By continuity of the action G X X,
Ug1 Vg X A

for open sets G Ug 3 g and X Vg 3 y. The compact space G can be covered by finitely
many of the open sets Ug , say G = U1 Un . Let V = V1 Vn be the intersection of the
corresponding neighborhoods of y. Then
[
[
G1 V =
Ui1 V
Ui1 Vi X A
i

so y V G(X A) = X GA.
(2) The orbit Gx of a point x X is compact because [Thm 26.5] it is the image of the compact
space G under the continuous map G X : g gx.
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