Grad Alg Top Mooremethod

Algebraic Topology
M382C
Michael Starbird
Fall 2007
Contents
1 Introduction
1.1 Basic Examples . . . . . . . . . . . . . . .
1.2 Simplices . . . . . . . . . . . . . . . . . .
1.3 Simplicial Complexes . . . . . . . . . . . .
1.4 2-manifolds . . . . . . . . . . . . . . . . .
1.4.1 2-manifolds as simplicial complexes
1.4.2 2-manifolds as quotient spaces . .
1.5 Questions . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
5
5
6
7
9
11
14
17
2 2-manifolds
2.1 Classification of 2-manifolds .
2.1.1 Classification Proof I .
2.1.2 Classification Proof II
2.2 PL Homeomorphism . . . . .
2.3 Invariants . . . . . . . . . . .
2.3.1 Euler characteristic . .
2.3.2 Orientability . . . . .
2.4 CW complexes . . . . . . . .
2.5 2-manifolds with boundary .
2.6 *Non-compact surfaces . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
19
19
20
23
27
28
28
29
35
38
42
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
3 Fundamental group and covering spaces

3.1 Fundamental group . . . . . . . . . . . .
3.1.1 Cartesian products . . . . . . . .
3.1.2 Induced homomorphisms . . . .
3.2 Retractions and fixed points . . . . . . .
3.3 Van Kampens Theorem, I . . . . . . . .
3.3.1 Van Kampens Theorem: simply
tion case . . . . . . . . . . . . . .
3
.
.
.
.
.
.
.
.
.
.
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
connected
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
. . . . . .
intersec. . . . . .
49
50
56
57
58
60
61
CONTENTS
3.3.2 Van Kampens Theorem: simply connected pieces case
3.4 Fundamental groups of surfaces . . . . . . . . . . . . . . . . .
3.5 Van Kampens Theorem, II . . . . . . . . . . . . . . . . . . .
3.6 3-manifolds . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.1 Lens spaces . . . . . . . . . . . . . . . . . . . . . . . .
3.6.2 Knots in S3 . . . . . . . . . . . . . . . . . . . . . . . .
3.7 Homotopy equivalence of spaces . . . . . . . . . . . . . . . . .
3.8 Higher homotopy groups . . . . . . . . . . . . . . . . . . . . .
3.9 Covering spaces . . . . . . . . . . . . . . . . . . . . . . . . . .
3.10 Theorems about groups . . . . . . . . . . . . . . . . . . . . .
63
63
66
67
67
69
72
73
75
82
4 Homology
4.1 Z2 homology . . . . . . . . . . . . . . . . . . . . . . . . . . .
4.1.1 Simplicial Z2 homology . . . . . . . . . . . . . . . . .
4.1.2 CW Z2 -homology . . . . . . . . . . . . . . . . . . . . .
4.2 Homology from parts, special cases . . . . . . . . . . . . . . .
4.3 Chain groups and induced homomorphisms . . . . . . . . . .
4.4 Applications of Z2 homology . . . . . . . . . . . . . . . . . .
4.5 Z2 Mayer-Vietoris Theorem . . . . . . . . . . . . . . . . . . .
4.6 Introduction to simplicial Z-homology . . . . . . . . . . . . .
4.6.1 Chains, boundaries, and definition of simplicial Zhomology . . . . . . . . . . . . . . . . . . . . . . . . .
4.7 Chain groups and induced homomorphisms . . . . . . . . . .
4.8 Relationship between fundamental group and first homology .
4.9 Mayer-Vietoris Theorem . . . . . . . . . . . . . . . . . . . . .
83
84
85
86
91
93
96
97
100
A Review of Point-Set Topology
111
B Review of Group Theory
117
C Review of Graph Theory
125
D The Jordan Curve Theorem
127
100
105
106
107
Chapter 1
Introduction
Abstracting and generalizing essential features of familiar objects often lead
to the development of important mathematical ideas. One goal of geometrical analysis is to describe the relationships and features that make up the
essential qualities of what we perceive as our physical world. The strategy
is to find ideas that we view as central and then to generalize those ideas
and to explore those more abstract extensions of what we perceive directly.
Much of topology is aimed at exploring abstract versions of geometrical
objects in our world. The concept of geometrical abstraction dates back
at least to the time of Euclid (c. 225 B.C.E.) The most famous and basic
spaces are named for him, the Euclidean spaces. All of the objects that we
will study in this course will be subsets of the Euclidean spaces.
1.1
Basic Examples
Definition ( Rn ). We define real or Euclidean n-space, denoted by Rn , as

the set
Rn := {(x1 , x2 , . . . , xn )|xi R for i = 1, . . . , n}.
We begin by looking at some basic subspaces of Rn .
Definition (standard n-disk). The n-dimensional disk, denoted Dn is defined as
Dn := {(x1 , . . . , xn ) Rn |0 xi 1 for i = 1, . . . , n }
n times
z
}|
{
[0, 1] [0, 1] [0, 1] Rn .
5
CHAPTER 1. INTRODUCTION
For example, D1 = [0, 1]. D1 is also called the unit interval, sometimes
denoted by I.
Definition (standard n-ball, standard n-cell). The n-dimensional ball or
cell, denoted Bn , is defined as:
Bn := {(x1 , . . . , xn ) Rn |x21 + . . . + x2n 1}.
Fact 1.1. The standard n-ball and the standard n-disk are compact and
homeomorphic.
Definition (standard n-sphere). The n-dimensional sphere, denoted Sn , is
defined as
Sn := {(x0 , . . . , xn ) Rn+1 |x20 + . . . + x2n = 1}.
Note. Bd Bn+1 = Sn
As usual, the term n-sphere will apply to any space homeomorphic to
the standard n-sphere.
Question 1.2. Describe S0 , S1 , and S2 . Are they homeomorphic? If not,
are there any properties that would help you distinguish between them?
1.2
Simplices
One class of spaces in Rn we will be studying will be manifolds or kmanifolds, which are made up of pieces that locally look like Rk , put together in a nice way. In particular, we will be studying manifolds that
use triangles (or their higher-dimensional equivalents) as the basic building
blocks.
Since k-dimensional triangles in Rn (called simplices) are the basic
building blocks we will be using, we begin by giving a vector description of
them.
Definition (1-simplex). Let v0 , v1 be two points in Rn . If we consider v0
and v1 as vectors from the origin, then 1 = {v1 + (1 )v0 | 0 1}
is the straight line segment between v0 and v1 . 1 can be denoted by {v0 v1 }
or {v1 v0 } (the order the vertices are listed in doesnt matter). The set 1 is
called a 1-simplex or edge with vertices (or 0-simplices) v0 and v1 .
1.3. SIMPLICIAL COMPLEXES
Definition (2-simplex). Let v0 , v1 , and v2 be three non-collinear points in

Rn . Then
2 = {0 v0 + 1 v1 + 2 v2 | 0 + 1 + 2 = 1 and 0 i 1i = 0, 1, 2}
is a triangle with edges {v0 v1 }, {v1 v2 }, {v0 v2 } and vertices v0 , v1 , and v2 .
The set 2 is a 2-simplex with vertices v0 , v1 , and v2 and edges {v0 v1 },
{v1 v2 }, and {v0 v2 }. {v0 v2 v2 } denotes the 2-simplex 2 (where the order the
vertices are listed in doesnt matter).
Note that the plural of simplex is simplices.
Definition (n-simplex and face of a simplex). Let {v0 , v2 , . . . , vn } be a set
affine independent points in RN . Then an n-simplex n (of dimension n),
denoted {v0 v1 v2 . . . vn }, is defined to be the following subset of RN :

(
)
n
X

n
= 0 v0 + 1 v1 + ... + n vn
i = 1 ; 0 i 1, i = 0, 1, 2, . . . , n .

i=0
An i-simplex whose vertices are any subset of i + 1 of the vertices of n is an

(i-dimensional) face of n . The face obtained by deleting the vm vertex from
the list of vertices of n is often denoted by {v0 v1 v2 . . . vc
m . . . vn }. (Note that
it is an (n 1)-simplex.)
Exercise 1.3. Show that the faces of a simplex are indeed simplices.
Fact 1.4. The standard n-ball, standard n-disk and the standard n-simplex
are compact and homeomorphic.
We will use the terms n-disk, n-cell, n-ball interchangeably to refer to
any topological space homeomorphic to the standard n-ball.
1.3
Simplicial Complexes
Simplices can be assembled to create polyhedral subsets of Rn known as

complexes. These simplicial complexes are the principal objects of study for
this course.
Definition (finite simplicial complex). Let T be a finite collection of simplices in Rn such that for every simplex ij in T , each face of ij is also a
simplex in T and any two simplices in T are either disjoint or their interS
section is a face of each. Then the subset K of Rn defined by K = ij
running over all simplices ij in T is a finite simplicial complex with triangulation T , denoted (K, T ). The set K is often called the underlying space
of the simplicial complex. If n is the maximum dimension of all simplices
in T , then we say (K, T ) is of dimension n.
Example 1. Consider (K, T ) to be the simplicial complex in the plane where
T
={
{(0, 0)(0, 1)(1, 0)}, {(0, 0)(0, 1)}, {(0, 1)(1, 0)},
{(0, 0)(0, 1)}, {(0, 1)(1, 0)}, {(1, 0)(0, 0)},
{(0, 0)}, {(0, 1)}, {(1, 0)}, {(0, 1)}} .
So K is a filled in triangle and a hollow triangle as pictured.

(0,1)
(0,0)
(1,0)
(0,1)
Exercise 1.5. Draw a space made of triangles that is not a simplicial complex, and explain why it is not a simplicial complex.
We have started by making spaces using simplices as building blocks.
But what if we have a space, and we want to break it up into simplices? If J
is a topological space homeomorphic to K where K is a the underlying space
of a simplicial complex (K, T ) in Rm , then we say that J is triangulable.
Exercise 1.6. Show that the following space is triangulable:
by giving a triangulation of the space.
1.4. 2-MANIFOLDS
Definition (subdivision). Let (K, T ) be a finite simplicial complex. Then

T 0 is a subdivision of T if (K, T 0 ) is a finite simplicial complex, and each
simplex in T 0 is a subset of a simplex in T.
Example 2. The following picture illustrates a finite simplicial complex and
a subdivision of it.
(K,T)
(K,T)
There is a standard subdivision of a triangulation that later will be

useful:
Definition (derived subdivision).
1. Let 2 be a 2-simplex with vertices v0 , v1 , and v2 . Then p =
1
1
2
3 v1 + 3 v2 is the barycenter of .
1
3 v0
2. Let T be a triangulation of a simplicial 2-complex with 2-simplices

{i }ki=1 . The first derived subdivision of T , denoted T 0 , is the union
of all vertices of T with the collection of 2-simplices obtained from
T by breaking each i in T into six pieces as shown, together with
their edges and vertices, and finally the edges and vertices obtained by
breaking each edge that is not a face of a 2-simplex into two edges.
Notice that the new vertices are the barycenter of each i in T and the
center of each edge in T . The second derived subdivision, denoted T 00 ,
is (T 0 )0 , the first derived subdivision of T 0 , and so on.(See Figure 1.1)
Example 3. Figure 1.2 illustrates a finite simplicial complex and the second
derived subdivision of it.
1.4
2-manifolds
The concept of the real line and the Euclidean spaces produced from the
real line are fundamental to a large part of mathematics. So it is natural to
10
Figure 1.1: Barycentric subdivision of a 2-simplex
Figure 1.2: Second barycentric subdivision of a 2-simplex
be particularly interested in topological spaces that share features with the

Euclidean spaces. Perhaps the most studied spaces considered in topology
are those that look locally like the Euclidean spaces. The most familiar such
space is the 2-sphere since it is modelled by the surface of Earth, particularly
in flat places like Kansas or the middle of the ocean. If you are on a ship
in the middle of the Pacific Ocean, the surrounding terrain looks like the
surrounding terrain if you were living on a plane, which is Euclidean 2space or R2 . The concept of a space being locally homeomorphic to R2
is sufficiently important that it has a name, in fact, two names. A space
locally homeomorphic to R2 is called a surface or 2-manifold. The 2-sphere
is a surface as is the torus (which looks like an inner-tube or the surface of
a doughnut).
Definition (2-manifold or surface). A 2-manifold or surface is a separable,

metric space 2 such that for each p 2 , there is a neighborhood U of p
that is homeomorphic to R2 .
1.4. 2-MANIFOLDS
1.4.1
11
2-manifolds as simplicial complexes
For now, we will restrict ourselves to 2-manifolds that are subspaces of Rn

and that are triangulated.
Definition (triangulated 2-manifold). A triangulated compact 2-manifold
is a space homeomorphic to a subset M 2 of Rn such that M 2 is the underlying
space of a simplicial complex (M 2 , T ).
Example 4. The tetrahedral surface below, with triangulation
T
= {{v0 v1 v2 }, {v0 v1 v3 }, {v0 v2 v3 }, {v1 v2 v3 },

{v0 v1 }, {v0 v2 }, {v0 v3 }, {v1 v2 }, {v1 v3 }, {v2 v3 },
{v0 }, {v1 }, {v2 }, {v3 }}
is a triangulated 2-manifold (homeomorphic to S2 ).
V3
V2
V0
V1
Figure 1.3: Tetrahedral surface
The following theorem asserts that every compact 2-manifold is triangulable, but its proof entails some technicalities that would take us too far
afield. So we will analyze triangulated 2-manifolds and simply note here
without proof that our results about triangulated 2-manifolds actually hold
in the topological category as well.
Theorem 1.7. A compact, 2-manifold is homeomorphic to a compact, triangulated 2-manifold, in other words, all compact 2-manifolds are triangulable.
Definitions (1-skeleton and dual 1-skeleton).
1. The 1-skeleton of a triangulation T equals
T } and is denoted T (1) .
{j | j is a 1-simplex in
12
S
2. The dual 1-skeleton of a triangulation T equals {j | j is an edge of
a 2-simplex in T 0 and neither vertex of j is a vertex of a 2-simplex of
T }. An edge in the dual 1-skeleton has each of its ends at the barycenters of 2-simplices of the original triangulation, that is, physically each
edge in the dual 1-skeleton is composed of two segments, each running
from the barycenter of a 2-simplex to the middle of the edge they share
in the original triangulation. So an edge in the dual 1-skeleton is the
union of two 1-simplices in T 0 .
Examples 5. The following are triangulable 2-manifolds:

a. S2
b. T2 := S1 S1 R4 or any other space homeomorphic to the boundary

of a doughnut, the torus.
c. Double torus:(See Figure 1.4)

The following example cannot be embedded in R3 ; however, it can be embedded in R4 .
d. The Klein bottle, denoted K2 :(See Figure 1.5)
There is another 2-manifold that cannot be embedded in R3 that we will
study, which requires the use of the quotient or identification topology (see
Appendix A):
1.4. 2-MANIFOLDS
13
Figure 1.4: The double torus or surface of genus 2
Figure 1.5: The Klein Bottle
e. The projective plane, denoted RP2 , := space of all lines through 0 in

R3 where the basis for the topology is the collection of open cones with
the cone point at the origin.
Exercise 1.8.
1. Show RP2
= S2 /hx xi, that is, the 2-sphere with diametrically
opposite points identified.
2. Show that RP2 is also homeomorphic to a disk with two edges on its
boundary (called a bigon), identified as indicated in Figure 1.6.
3. Show that RP2
obius band with a disk attached to its boundary
= M
(See Figure 1.7).
Exercise 1.9. Show that T2 as defined above is homeomorphic to the surface
in R3 parametrized by:

1
1
(, 1 + cos , sin |0 2, 0 2
2
2
in cylindrical coordinates.
14
Figure 1.6: RP2
Figure 1.7: The Mobius band
There is a way of obtaining more 2-manifolds by connecting two or

more together. For instance, the double torus looks like two tori that have
been joined together.
Definition (connected sum). Let M12 and M22 be two compact, connected,
triangulated 2-manifolds and let D1 and D2 be 2-simplices in the triangulations of M1 and M2 respectively. Paste M12 Int D1 and M22 Int D2 along
the boundaries of Bd D1 and Bd D2 . The resulting manifold is called the
connected sum of M12 and M22 , and denoted by M12 # M22 . Similarly, define
the connected sum of n 2-manifolds recursively.
This definition of connected sum can in fact be generalized to the connected sum of any two n-manifolds. Can you see how to do it?
Exercise 1.10. Show that RP2 # RP2 is homeomorphic to the Klein bottle.
Exercise 1.11. Show that T2 # RP2 , where T2 is the torus, is homeomorphic K 2 # RP2 , where K2 is the Klein bottle.
1.4.2
2-manifolds as quotient spaces
There is another way of thinking of 2-manifolds, as the abstract spaces

obtained from a particular kind of quotients (see Appendix A for a review
of quotient spaces).
1.4. 2-MANIFOLDS
15
The process of identifying all elements of an equivalence class to a single

one is often called a gluing when the equivalence classes are mostly small,
having 1 or 2 or a finite number of points in each.
In our case, we will be looking at the quotient spaces obtained from
polygonal disks, where all points of the interior of the disk are in their own
equivalence class, the points on the interior of the edges are in two-point
equivalence classes, and the vertices of the polygonal disks are in equivalence
classes with any number of other vertices. We think of obtaining the 2manifold by gluing the edges of the polygonal disk to each other pairwise,
in some particular pattern.
Examples 6. In these examples the kind of arrow indicates which edges are
glued together, while the orientations of the arrows indicate how to glue the
two edges together. You should convince yourself that any two gluing maps
that agree with the given orientations will yield homeomorphic spaces.
1. (torus)
b
b
a
a
b
2. (sphere) (See Figure 1.8)
b
b
a
a
Figure 1.8: The sphere
3. (sphere) (See Figure 1.9)

4. (double torus) (See Figure 1.10)
16
Figure 1.9: Another way to see the sphere

a
b
c
a
d
d
c
Figure 1.10: The double torus
5. (Klein bottle) (See Figure 1.11)
b
a
a
b
Figure 1.11: The Klein bottle
6. (projective plane) (See Figure 1.12)

7. (projective plane) (See Figure 1.13)
You should check to see that alternative presentations of the same space are
homeomorphic. You should also check that these spaces are homoeomorphic
to the triangulable 2-manifolds described in the previous subsection.
1.5. QUESTIONS
17
Figure 1.12: The projective plane
b
a
a
b
Figure 1.13: Another version of the projective plane
The following theorem will be put off to chapter 4 (and stated in a

slightly different but equivalent way). Surprisingly, it is highly non-trivial
to prove but not surprisingly it is incredibly useful.
Theorem 1.12 (Jordan Curve Theorem). Let h : [0, 1] D2 be a topological embedding where h(0), h(1) Bd(D2 ). Then h([0, 1]) separates D2 into
exactly two pieces.
Theorem 1.13. Any polygonal disk with edges identified in pairs is homeomorphic to a compact, connected, triangulated 2-manifold.
Theorem 1.14. Any compact, connected, triangulated 2-manifold is homeomorphic to a polygonal disk with edges identified in pairs.
1.5
Questions
The most fundamental questions in topology are:

Question 1.15. How are spaces similar and different? Particularly, which
are homeomorphic? Which arent?
Showing two spaces are homeomorphic means we must construct a homeomorphism between them. But how do we show two spaces are not homeomorphic? When we are confronted with the task of trying to explore one
18
space or to specify what is different about two spaces, we must examine the
spaces looking for features or properties that are of topological significance.
Question 1.16. What features of the examples studied are interesting either
in their own right or for the purpose of distinguishing one from another?
Chapter 2
2-manifolds
2.1
Classification of compact 2-manifolds
A surface, or 2-manifold, is locally homeomorphic to R2 , so we know how

these spaces look locally. But what are the possibilities for the global character of these spaces? We have seen several examples (the 2-sphere, the
torus, the Klein bottle, RP2 ). Now we seek to organize our understanding
of the collection of all surfaces, that is, to recognize, describe, and classify
each surface as one from a simple list of possible homeomorphism classes. So
we need to use the local Euclidean feature of 2-manifolds to help us describe
the overall structure of these surfaces.
In working with these compact 2-manifolds, we want to think of them as
physical objects made of simple building blocks, namely, triangles. In fact,
we will begin by considering just 2-manifolds that reside in Rn and are made
of triangles. This investigation of these simple compact 2-manifolds actually
is comprehensive since every compact 2-manifold is homeomorphic to one
made of finitely many triangles which is embedded in Rn . The advantage of
working with objects made from a finite number of triangles is that we can
use inductive procedures moving from triangle to triangle.
The main thing to have in mind at this point is that we should view
2-manifolds as concrete, physical objects that are constructed from a finite
number of flat triangles (simplices) that fit together as specified: they overlap, if at all, only along a shared edge or at a vertex of each. This physical
view of 2-manifolds will allow us to understand them so clearly that we
can describe an effective method for determining the global structure of the
object by knowing the local structure.
The goal of the following two sections is to prove (in two different ways)
19
20
CHAPTER 2. 2-MANIFOLDS
that every compact, triangulated 2-manifold can be constructed by taking

the connected sum of simple 2-manifolds, namely the sphere, torus, and
projective plane.
In the following section we proceed with a sequence of theorems that
show us that after removing one disk, any compact, triangulated 2-manifold
is just a disk with strips attached in particularly simple ways. We continually
use the local structure of the triangulated 2-manifold to see how the whole
thing fits together.
The second proof of the classification theorem views each compact, triangulated 2-manifold as the quotient space of a polygonal disk with its edges
identified in pairs.
Definition (regular neighborhood). Let M 2 be a 2-manifold with triangu2 , T ) . The regular
lation T = {i }ki=1 . Let A be a subcomplex
S 00of (M
00
neighborhood of A, denoted N (A), equals {j | j T 00 and j00 A 6= }.
Exercise 2.1. The boundary of a tetrahedron is naturally triangulated with
a triangulation T consisting of four 2-simplexes together with their six edges
and four vertices. On the boundary of a tetrahedron locate the first and second derived subdivisions of T , the 1-skeleton of T , the regular neighborhood
of the 1-skeleton of T , the regular neighborhoods of a vertex and an edge of
T , and the dual 1-skeleton of T .
Exercise 2.2. On the accompanying pictures of the second derived subdivisions of triangulations of the torus and the Klein bottle, find regular
neighborhoods of subsets of the 1-skeleton.
Exercise 2.3. Characterize graphs in the 1-skeleton of T for the triangulations of the sphere, torus, and projective plane whose regular neighborhoods
are homeomorphic to a disk.
2.1.1
Classification of compact, connected 2-manifolds, I
The basic idea of this proof is to show that removing an open disk from
a compact triangulated 2-manifold gives us a space homeomorphic to a
(closed) disk with some number of bands attached to its boundary in a
specified way. The number of bands, and how they are attached then gives
us the classification of the surface.
Theorem 2.4. Let M 2 be a compact, triangulated 2-manifold with triangulation T . Let S be a tree whose edges are 1-simplices in the 1-skeleton of T .
Then N (S), the regular neighborhood of S, is homeomorphic to D2 .
2.1. CLASSIFICATION OF 2-MANIFOLDS
21
Theorem 2.5. Let M 2 be a compact, triangulated 2-manifold with triangulation T . Let S be a tree equal to a union of edges in the dual 1-skeleton of
T . Then {j00 | j00 T 00 and j00 S 6= } is homeomorphic to D2 .
Theorem 2.6. Let M 2 be a connected, compact, triangulated 2-manifold
with triangulation T . Let S be a tree in the 1-skeleton of T . Let S 0 be the
subgraph of the dual 1-skeleton of T whose edges do not intersect S. Then
S 0 is connected.
The following two theorems state that M 2 can be divided into two pieces,
one a disk D0 , and the other a disk (D1 ) with bands (the Hi s) attached to
it.
Theorem 2.7. Let M 2Sbe a connected,
compact, triangulated 2-manifold.

Then M 2 = D0 D1 ki=1 Hi where D0 , D1 , and each Hi is homeomorS
phic to D2 , Int D0 D1 = , the Hi s are disjoint, ki=1 Int Hi (D0 D1 ) =
, and for each i, Hi D1 equals 2 disjoint arcs each arc on the boundary
of each of Hi and D1 .
Theorem 2.8. Let M 2 be a connected, compact, triangulated 2-manifold.
Then:
1. There is a disk D0 in M 2 such that M 2 (Int D0 ) is homeomorphic to
the following subset of R3 : a disk D1 with a finite number of disjoint
strips, Hi for i {1, . . . n}, attached to boundary of D1 where each
strip has no twist or 1/2 twist. (See Figure 2.1.)

Sk
2. Furthermore, the boundary of the disk with strips, D1
i=1 Hi , is
connected.
Exercise 2.9. In the set-up in the previous theorem, any strip Hi divides
the boundary of D0 into two edges e1i and e2i , where Hi is not attached.
Show that if a strip Hj is attached to D0 with no twists, then there must be
a strip Hk that is attached to both e1j and e2j .
Then there is a disk D0 in M 2 such that M 2 Int D0 is homeomorphic to a
disk D1 with strips attached as follows: first come a finite number of strips
with 1/2 twist each whose attaching arcs are consecutive along Bd D1 , next
come a finite number of pairs of untwisted strips, each pair with attaching
arcs entwined as pictured with the four arcs from each pair consecutive along
Bd D1 .
22
Figure 2.1: A disk with four handles attached.
Figure 2.2: Twisted strips and entwined strips
Theorem 2.11. Let X be a disk D0 with one strip attached with a 1/2 twist
with its attaching arcs consecutive along Bd D0 and one pair of untwisted
strips with attaching arcs entwined as pictured with the four arcs consecutive
along Bd D0 . Let Y be a disk D1 with three strips with a 1/2 twist each whose
attaching arcs are consecutive along Bd D1 . Then X is homeomorphic to Y .
Then there is a disk D0 in M 2 such that M 2 Int D0 is homeomorphic to
one of the following:
a) a disk D1 ,
b) a disk D1 with k 12 -twisted strips with consecutive attaching arcs, or
c) a disk D1 with k pairs of untwisted strips, each pair in entwining position with the four attaching arcs from each pair consecutive.
23
Figure 2.3: These spaces are homeomorphic.
Figure 2.4: Entwining pair of strips
Theorem 2.13 (Classification of compact, connected 2-manifolds). Any

connected, compact, triangulated 2-manifold is homeomorphic to the 2-sphere
S2 , a connected sum of tori, or a connected sum of projective planes.
Notice that at this point we have shown that any compact, connected,
triangulated 2-manifold is a sphere, the connected sum of n tori, or the
connected sum of n projective planes; however, we have not yet established
that those possibilities are all topologically distinct. The classification of
2-manifolds requires us to prove our suspicions that any two different connected sums are indeed not homeomorphic. Before we develop tools for
confirming those suspicions, we digress to develop another proof of this first
part of the classification theorem.
2.1.2
Classification of compact, connected 2-manifolds, II
We now outline a different approach to proving that any compact, connected, triangulated 2-manifold is a sphere, the connected sum of tori, or
the connected sum of projective planes. This approach uses the quotient or
identification topology described in the previous chapter.
Suppose that we are gluing the edges of a polygonal disk to create a
2-manifold. If we assign a unique letter to each pair of edges that are glued
24
together, and we read the letters as we follow the edges along the boundary
of the disk (starting at a certain edge) going clockwise, we get a word
made up of these letters. However, to specify the gluing we need to know
not only which edges are glued together, but in what orientation. To keep
track of that, we will write the letter alone if the orientation given on the
edge agrees with the direction were reading the edges in, and the letter to
the 1 power if it disagrees. For example, abca1 dcb1 d represents a gluing
of the octagon as indicated, so that the orientations of two identified edges
agree:
a
b
c
a
d
d
c
Figure 2.5: The genus two surface
Definition (gluing of a 2n-gon with edges identified in pairs). An expression

( word) of n letters, such as abca1 dcb1 d, where each letter appears exactly
twice, represents the 2-manifold obtained by gluing the edges of a 2n-gon in
pairs as indicated by the sequence of letters. Notice that a pair of edges with
the same letter really has two different possible gluings. To determine which
gluing, we need to look at the superscript or lack of subscript of each letter.
A letter without a subscript is viewed as oriented clockwise around the 2ngon, while a superscript 1, as in a1 , indicates that that edge is oriented
counterclockwise. Then the identification of the pair of edges respects those
directions. So the equivalence classes of the disk specified by such a 2n length
string of n letters consist of every singleton in the interior of the 2n-gon,
pairs of points one from each interior of the edges with the same label, and
then equivalence classes of vertices as come together when the edges are
identified as specified. The equivalence classes among vertices might have
any number of vertices in them, depending on the string of letters.
Theorem 2.14.
1. The bigon with edges identified by aa1 is homeomorphic to S2 .
2. The bigon with edges identified by bb is homeomorphic to RP2 .
25
3. The square with edges identified by cdc1 d1 is homeomorphic to T2 .

Theorem 2.15 (connected sum relation). The gluing of a square given by
ccdd is homeomorphic to RP2 #RP2 and the gluing of an octagon given by
aba1 b1 cdc1 d1 is homeomorphic to T2 #T2 .
Question 2.16. Generalize the above to the connected sum of any two surfaces.
The next sequence of theorems will show us how to take a 2n-gon with
edges identified in pairs and modify the gluing prescription to find a canonical representation of the same 2-manifold.
Theorem 2.17. Let Abb1 C be a string of 2n letters where each letter
occurs twice, with or without a superscript (so A and C should each be
construed as being comprised of many letters). Then the 2-manifold obtained
by identifying a 2n-gon following the gluing Abb1 C is homeomorphic to the
2-manifold which is obtained by identifying a (2n 2)-gon following the
gluing given by AC.
Theorem 2.18. Suppose a 2-manifold M 2
6 S2 is represented by a 2n=
gon with edges identified in pairs. Then a homeomorphic 2-manifold can be
represented by a 2k-gon with edges identified in pairs where all the vertices
are in the same equivalence class, that is, all the vertices are identified to
each other.
S2 is represented by a 2n-gon
Theorem 2.19. Suppose a 2-manifold M 2 6=
with edges identified in pairs. Then a homeomorphic 2-manifold can be represented by a 2k-gon with edges identified in pairs where all the vertices are
identified and every pair of edges with the same orientation are consecutive.
Theorem 2.20. Suppose a 2-manifold M 2 6
= S2 is represented by a 2ngon with edges identified in pairs. Then a homeomorphic 2-manifold can be
represented by a 2k-gon with edges identified in pairs where all the vertices
are identified, every pair of edges with the same orientation are consecutive,
and all other edges are grouped in disjoint sets of two intertwined pairs
following the pattern aba1 b1 .
Theorem 2.21. The 2-manifold represented by aba1 b1 cc is homeomorphic to the 2-manifold represented by ddeef f .
Question 2.22. Re-state the above theorem in terms of connected sum.
26
Theorem 2.23. Any compact, connected, triangulated 2-manifold is homeomorphic to a 2n-gon with edges identified in pairs as specified in one of
the three following ways: aa1 , or a0 a0 a1 a1 . . . an an (where n 0) or
1
1 1
a0 a1 a1
0 a1 . . . an1 an an1 an (where n 1 is odd ).
Theorem 2.24 (Classification of compact, connected 2-manifolds). Any
connected, compact, triangulated 2-manifold is homeomorphic to the 2-sphere
S2 , a connected sum of tori, or a connected sum of projective planes.
2.2. PL HOMEOMORPHISM
2.2
27
PL Homeomorphism
Our goal is to organize connected, compact, triangulated 2-manifolds by

homeomorphism type. The concept of topological homeomorphism does
not reflect the triangulated structure we have associated with these objects,
so here we present a natural way of equating two triangulated 2-manifolds
that includes the simplicial structure of them as well as the topological type.
The basic strategy is first to define an equivalence between two triangulated 2-manifolds with triangulations T1 and T2 if the simplices of T1
correspond to the simplices of T2 in a straightforward 1-1 fashion. Then we
describe another idea of equivalence if the two 2-manifolds can be subdivided
to find new triangulations that have this 1-1 correspondence.
Definition (simplicial homeomorphism). We will say that M12 with triangulation T1 is simplicially homeomorphic to M22 with triangulation T2 if and
only if there exists a homeomorphism from M12 to M22 that gives a one-to-one
correspondence between T1 and T2 in the following way: the homeomorphism
maps each simplex in T1 linearly to a single simplex in T2 . So the vertices of
T1 go to the vertices of T2 and the rest of the homeomorphism is determined
by extending the map on the vertices linearly over each simplex.
Of course, we have seen that a space can have many different triangulations. Therefore, the concept of a simplicial homeomorphism is too restrictive. An underlying space with a triangulation and the same space with its
second derived subdivision triangulation are not simplicially isomorphic. So
we can give a broader concept of equivalence:
Definition (PL homeomorphism). M12 with triangulation T1 is PL homeomorphic to M22 with triangulation T2 if and only if there exist subdivisions
T10 and T20 of T1 and T2 respectively such that (M12 , T10 ) is simplicially isomorphic to (M22 , T20 ).
The letters PL come from piecewise linear, as the correspondence described above gives a homeomorphism between M12 and M22 that can be
realized as a map that is linear when restricted to each simplex of T10 .
For 2-manifolds you may assume without proof that a homeomorphism
between two manifolds induces a PL-homeomorphism. This however is not
true for general n-manifolds.
28
2.3
Invariants
One of our goals in studying topological spaces is to be able to distinguish

non-homeomorphic spaces from one another. A fundamental strategy to tell
the difference between two topological spaces is to find some feature of one
space that, on the one hand, is preserved under homeomorphism and, on
the other hand, is not shared by the other space. In distinguishing spaces in
a general topology course, we might look at topological properties such as
being normal, compact, or connected. However, since we are now trying to
distinguish among spaces all of which are compact, metric spaces, we need to
look for different types of features that are invariant under homeomorphisms.
We use the word invariant to refer to any property of a space that is shared
by any homeomorphic space. That is, it is a property that is preserved
by homeomorphisms. So compactness, normality, and connectedness are all
invariants. The diameter of a 2-manifold embedded in R3 , on the other
hand, is not an invariant.
The crux of the whole course is to define and use invariants that are useful
for distinguishing one space from another, especially invariants that can help
us distinguish rather nice subsets of Rn that might be constructed from a
finite number of simplices. We begin now by defining an invariant that will
help us distinguish one compact, connected, triangulated 2-manifold from
some others.
2.3.1
Euler characteristic
Definition (Euler characteristic). Let M 2 be a 2-manifold with triangulation T . Let

v =
number of vertices in T
e =
number of 1-simplices in T
number of 2-simplices in T
and define the Euler characteristic, (M 2 ), of M 2 by (M 2 ) = v e + f .

Theorem 2.25. Let M 2 be a connected, compact, triangulated 2-manifold
with triangulation T . Let T 0 be a subdivision of T . Then (M 2 , T ) =
(M 2 , T 0 ).
In other words, for a triangulated, compact 2-manifold, the Euler characteristic is preserved under subdivision.
2.3. INVARIANTS
29
Theorem 2.26. Let M12 and M22 be connected, compact, triangulated 2manifolds. If M12 is PL-homeomorphic to M22 , then (M12 ) = (M22 ).
Since PL-homeomorphic manifolds must have the same Euler characteristic, Euler characteristic helps to distinguish between 2-manifolds that are
not PL-homeomorphic.
Theorem 2.27.
1. (S2 ) = 2.
2. (T2 ) = 0.
3. (RP2 ) = 1.
4. (K2 ) = 0.
Theorem 2.28. Let M12 and M22 be two connected, compact, triangulated
2-manifolds. Then (M12 # M22 ) = (M12 ) + (M22 ) 2.
Theorem 2.29. Let T2i be the torus for i = 1, . . . , n. Then

#
i=1
T2i

= 2 2n.
Definition (genus). The genus of S2 = 0. The genus of a 2-manifold

n
= # T2 is n.
i=1
Theorem 2.30. Let RP2i be the projective plane for i = 1, . . . , n. Then

n
# RP2 = 2 n.
i=1
2.3.2
Orientability
Euler characteristic is a useful invariant, in that it helps to distinguish 2manifolds. However, it does not distinguish between the torus and the Klein
bottle, for example. In fact, for each even number 0 there are two nonhomeomorphic compact, connected, triangulated 2-manifolds of that Euler
characteristic, one a connected sum of tori, and one a connected sum of projective planes. So although Euler characteristic is useful for distinguishing
non-homeomorphic surfaces, it does not differentiate all different surfaces.
30
There is a second invariant which, when combined with Euler characteristic, will allow us to distinguish between any two non-homeomorphic,
compact, connected 2-manifolds. This invariant is orientability.
A surface is orientable if we can choose an ordered basis for the local
Euclidean structure at each point of the surface in such a way that the
bases change smoothly as the point moves along a path in the surface.
Note that orientability on its own is a very coarse invariant: a 2-manifold
is either orientable or non-orientable. In other words, orientability divides
the set of all 2-manifolds into two classes. It turns out that the combination
of orientability and Euler characteristic is enough to differentiate any two
compact, connected, triangulated 2-manifolds.
We can explore the concept of orientability in triangulated surfaces by
considering orderings of the vertices of each simplex.
First let us see what we mean by an orientation of a 0-, 1-, and 2-simplex.
Definitions (oriented simplices). Let 2 be the 2-simplex {v0 v1 v2 }, 1 be
the 1-simplex {w0 w1 }, and 0 be the 0-simplex {u0 }.
1. Two orderings of the vertices v0 , v1 , . . . vn of an n-simplex n are
said to be equivalent if they differ by an even permutation. Thus
{v0 , v1 , v2 } {v1 , v2 , v0 }. However, {v0 , v1 , v2 } 6 {v1 , v0 , v2 } since
they differ by a single 2-cycle, which is an odd permutation. Note that
this equivalence relation produces precisely two equivalence classes of
orderings of vertices of an n-simplex for n 1. An equivalence class
will be denoted by [v0 v1 . . . vn ], where {v0 , v1 , . . . , vn } is an element of
the equivalence class.
2. An orientation of the 2-simplex 2 is a one-to-one and onto function
o from the two equivalence classes of the orderings of the vertices of
2 to {1, 1}. Note that there are two possible such orientations for
2 . Any vertex ordering that lies in the equivalence class whose image
is +1 will be called positively oriented or will be said to have a positive
orientation, orderings in the other class will be said to be negatively
oriented or have a negative orientation. We can indicate the chosen
positive orientation for 2 by denoting 2 as [v0 v1 v2 ], where [v0 v1 v2 ]
is in the positive equivalence class. Note that o[v0 v1 v2 ] = o[v1 v0 v2 ].
You can draw a circular arrow inside the 2-simplex in the direction
indicated by any of the positively oriented orderings. That circular
arrow (which will be either clockwise or counterclockwise on the page)
indicates the choice of (positive) orientation for that 2-simplex.
2.3. INVARIANTS
31
3. An orientation of the 1-simplex 1 is a one-to-one and onto function

o from the two orderings of the vertices of 1 to{1, 1}. The ordering whose image is +1 has the positive orientation, the other has a
negative orientation. As with 2 , note that there are only two possible
orientations for 1 , and that o[w0 w1 ] = o[w1 w0 ]. We think of [w0 w1 ]
as being the orientation that points from w0 to w1 .
4. Since 0 has a single equivalence class of orderings of its vertex, we
have a slightly different definition of orientation for a 0-simplex. An
orientation of a 0-simplex is a function o from {[u0 ]} to {1, 1}.
Definition (induced orientation on an edge). If we choose an orientation
of a 2-simplex, then there are associated orientations on each of the three
edges called the induced orientations on the edges. If [v0 v1 v2 ] is the positive
orientation of a 2-simplex, then the orientations induced on the edges are
a. [v1 v2 ].
b. [v0 v2 ] = [v2 v0 ].
c. [v0 v1 ].
respectively.
Note. The definition of induced orientation is a natural one, since the induced orientations on the edges give a directed cycle of edges (v0 to v1 to v2
and back to v0 ) which follow the selected positive ordering of the vertices of
the 2-simplex.
Exercise 2.31. Show that the induced orientation on an edge of a 2-simplex
is well defined; in other words, that it is independent of the choice of positive
equivalence class representative.
Definition (induced orientation on a vertex). The orientations induced on
the vertices of 1 = [v0 v1 ] are
a. [v0 ].
b. [v1 ].
respectively.
32
We can now define what we mean by an orientable, triangulated 2manifold. Intuitively, a triangulated 2-manifold is orientable if it is possible
to select orientations for each 2-simplex in such a way that neighboring 2simplices have compatible orientations. The concept of compatible comes
from the following observation. If you draw two triangles in the plane that
share an edge and orient them both in a counterclockwise ordering, say,
then the shared edge has induced orientations from the two triangles that
are opposite to each other. In other words, when the orientations on both
triangles are the same, then the induced orientations on a shared edge are
opposite. This observation gives rise to the definition of orientability.
Definition (orientablity). A triangulated 2-manifold M 2 is orientable if and
only if an orientation can be assigned to each 2-simplex in the triangulation
such that given any 1-simplex e 1 2 , the orientation induced on e by 1 is
opposite to the orientation induced by 2 . Otherwise, M 2 is non-orientable.
A choice of orientations of the 2-simplices of a triangulation of M 2 satisfying the condition stated above is called an orientation of M 2 .
Note.
Theorem 2.32. Suppose (M 2 , T ) is a 2-manifold with triangulation T and
T 0 is a subdivision of T . Then if (M 2 , T ) is orientable, so is (M 2 , T 0 ).
Theorem 2.33. Orientability is preserved under PL homeomorphism.
Theorem 2.34. M 2 is orientable if and only if it contains no M
obius band.
Theorem 2.35. Let M = M1 # . . . # Mn . Then M is orientable if and only
if Mi is orientable for each i {1, . . . , n}.
Compact, connected, triangulated 2-manifolds are determined by orientability and Euler characteristic.
Theorem 2.36 (Classification of compact, connected 2-manifolds). If M 2
is a connected, compact, triangulated 2-manifold then:
(a) if (M 2 ) = 2, then M 2
= S2 .
(b) if M 2 is orientable and (M 2 ) = 2 2n, for n 1, then
M
=
2
#
i=1
Ti2

.
2.3. INVARIANTS
33
(c) if M 2 is non-orientable and (M 2 ) = 2 n, for n 1, then

n
2
2
M = # RPi .
i=1
Notice that orientable connected, compact, triangulated 2-manifolds must

have even Euler Characteristic.
34
Problem 2.37. Identify the following 2-manifolds as a sphere, or a connected sum of n tori (specifying n), or a connected sum of n projective planes
(specifying n).
a. T#RP
b. K#RP
c. RP#T#K#RP
d. K#T#T#RP#K#T
2.4. CW COMPLEXES
2.4
35
CW complexes
Triangulating a surface in order to calculate the Euler Characteristic can

be quite tedious and time-consuming. However, we dont need to divide a
surface into such small pieces in order to compute the Euler Characteristic.
We can instead write the 2-manifold as the union of much larger cells that fit
together appropriately and use them to compute the Euler Characteristic.
Our strategy for discovering an appropriate generalization of triangulation
is to start with a triangulated surface and systematically enlarge the triangles and edges to produce other cell decompositions of the surface that will
continue to reveal the Euler Characteristic. In a way, this process of enlargement and amalgamation is the opposite of the subdivision process that we
saw earlier preserved the Euler Characteristic. We begin by amalgamating
two adjacent 2-simplexes of a triangulation.
Theorem 2.38. Let (M 2 , T ) be a triangulated 2-manifold. Suppose =
{uvw} and 0 = {uvw0 } are two distinct 2-simplexes in T that share the
edge e = {uv}. Then we can create a new structure for M 2 alternative to
T , namely, P where P = T { } {, 0 ,e}, where = 0 is the polygon
formed by the union of the two 2-simplices along their shared edge. If v 0 ,
e0 , f 0 are the numbers of vertices, edges, and polygons in P , then the Euler
Characteristic (M 2 , T ) = v 0 e0 + f 0 (see Figure 2.6).
u
w
w
v
w
v
Figure 2.6: The basic idea of CW complexes
The previous theorem amalgamated two triangles together; however, we

can continue in that vein by amalgamating polygonal disks together that we
may have created.
Theorem 2.39. Let (M 2 , T ) be compact, triangulated 2-manifold with Euler
characteristic (M 2 , T ). Suppose we create a polygonal structure P on M 2
inductively as follows. Let P0 = T . Suppose we have created Pi . Suppose
36
two 2-dimensional objects and 0 in Pi share a connected path of edges

in the boundary of each from vertex u to w (v 6= w). We create Pi+1 by
removing and 0 from Pi , removing all the edges in the path from vertex
u to w, removing all vertices of the edges in that path except for u and w,
and putting in the single two dimensional object 0 . Then if v, e, f
are the numbers of vertices, edges, and 2-dimensional objects in Pi+1 , then
(M 2 , T ) = v e + f (see Figure 2.7).
Figure 2.7: Removing a path from a CW complex

Notice that a 2-dimensional object in Pi may no longer be homeomorphic
to a disk, but the interior of each is homeomorphic to an open disk. We can
continue our inductive definition of our new structure on M 2 by similarly
reducing the number of 1-dimensional objects.
Theorem 2.40. Let (M 2 , T ) be compact, triangulated 2-manifold with a
polygonal structure P as defined inductively in the previous theorem. Suppose we substitute P with a new structure obtained inductively as follows.
Let P = P0 . If Pi has an edge e with a free vertex v, that is, v is not the
boundary of any other edge in Pi , then remove v and e from Pi to create
Pi+1 . If Pi has a vertex v that is one end of an edge e in Pi and one end of
an edge f in Pi and v is not on the end of any other edge, then remove v, e,
and f from Pi and put in the new 1-dimensional object e f to create Pi+1 .
Then if v 0 , e0 , f 0 are the numbers of vertices, 1-dimensional objects, and 2dimensional objects in an inductively defined P , then (M 2 , T ) = v 0 e0 +f 0 .
Exercise 2.41. Start with a triangulation of S2 and carry out the preceding
process as far as possible. What structure do you get? Confirm that you
get the right Euler Characteristic.
Exercise 2.42. Start with a triangulation of T2 and carry out the preceding
process as far as possible. What structure do you get? Confirm that you
get the right Euler characteristic.
2.4. CW COMPLEXES
37
u3
u2
v
g2
u1
g
u4
g1
u0
u3
g4
u2
g3
g2
u1
u4
g1
u0
Figure 2.8: Removing edges, vertices, and faces from a CW complex
We will now formalize what we have observed by defining a CW decomposition of a 2-manifold.

Definition (interior of a 0, 1, and 2-simplex).
1. For each 2-simplex
2 = {v0 v1 v2 } let Int 2 = {0 v0 + 1 v1 + 0 v2 |0 < i < 1}.
2. For each 1-simplex 1 = {w0 w1 } let Int 1 = {0 w0 + 1 w1 |0 < i <
1}.
3. For each 0-simplex 0 = {u0 } let Int 0 = 0 .
Theorem 2.43. Let (M 2 , T ) be a compact, triangulated 2-manifold with
triangulation T . Then M 2 equals the disjoint union of the Int i where
i T .
Definition (open n-cell from T ). Let (M 2 , T ) S
be a compact, triangulated
2-manifold with triangulation T . Suppose C = {Int i |i T } is homeomorphic to an open k-ball (k {0, 1, 2}). Then C is an open k-cell from
T.
Definition (cellular decomposition). Let (M 2 , T ) be a compact, triangulated
2-manifold with triangulation T . If M 2 is the disjoint union of Cik (k =
0, 1, 2 and i = 1, . . . , nk ), where each Cik is an open k-cell from T , then
S = {Cik } is a cellular decomposition of M 2 .
These decompositions are called cellular decompositions or CW decompositions because the space can be viewed as constructed from the images of
first vertices then 1-cells with their interiors mapped homeomorphically and
their boundaries mapped onto 0-cells (points), and then 2-cells with their
interiors mapped homeomorphically and their boundaries mapped to the set
of images of the lower dimensional cells.
Theorem 2.44. Let S be a cellular decomposition of a compact, triangulated
2-manifold (M 2 , T ). If v, e, and f are the number of 0, 1 and 2 cells in S,
then the Euler Characteristic (M 2 , T ) = v e + f .
38
Problem 2.45. Identify the following surfaces:

a. The surface obtained by identifying the edges of the octagon as indicated:
a
b
c
a
d
d
c
Figure 2.9: The genus two surface
b. The surface obtained by identifying the edges of the decagon as indicated (See Figure 2.10):
b
c
e
c
Figure 2.10: The decagon with edges indentified in pairs
2.5
2-manifolds with boundary
Exercise 2.46. What should be the definition of a connected, compact, triangulated 2-manifold-with-boundary?
2.5. 2-MANIFOLDS WITH BOUNDARY
39
Your definition should be general enough to include the following examples of 2-manifolds-with-boundary:
1. D2
2. A2 := {(x1 , x2 ) R2 | 12 x21 + x22 1}, the annulus (See Figure 2.11)
Figure 2.11: The annulus
3. Pair of pants:
Figure 2.12: The pair of pants
4. A disk with two intertwined handles attached, as shown in Figure 2.13.

5. M
obius band, see Figure 2.14
Exercise 2.47. Formulate the necessary definitions and theorem statements
that classify compact, connected, triangulated 2-manifolds-with-boundary.
Prove your theorems.
40
Figure 2.13:
Figure 2.14:
Once you have the above work done, you should be able to completely
classify and identify all connected, compact, triangulated 2-manifolds, with
and without boundary.
To distinguish a compact manifold with no boundary from one with topological boundary or to emphasize that a compact manifold has no boundary
the term closed manifold is often used. Beware: this term does not mean
topologically closed, as in the complement of an open set, but rather it
means a manifold that is compact without boundary. Both closed manifolds and compact manifolds-with-boundary are in fact closed subsets (in
the topological sense) of Rn and non-compact manifolds might be embedded as topologically closed subsets of Rn . This unfortunate terminology is
one of many examples of the use of a single word to signify several different
meanings. Context usually makes the meaning clear.
Problem 2.48. Identify the following surfaces made by two disks joined by
bands as indicated (See Figures 2.15 and 2.16):
Exercise 2.49. Fill out the following table, using the connected sum decom-
2.5. 2-MANIFOLDS WITH BOUNDARY
41
Figure 2.15: a. n twisted bands
Figure 2.16: b. 1 untwisted band and n 1 twisted bands
position. The number of boundary components is denoted by ||.

||
2
1
0
1
2
3
4
5
0
orient.
1
non-or.
orient.
2
non-or.
orient.
3
non-or.
orient.
non-or.
42
Figure 2.17: Examples of non-compact surfaces with infinite genus
2.6
*Non-compact surfaces
The surfaces studied so far in this chapter are all compact and connected
2-manifolds, with or without boundary. We can also consider non-compact
2-manifolds, but we will not do so in this class. An interesting question to
ask yourself is: how do you extend all the concepts learned about compact
spaces to non-compact ones?
For example, can you formulate and prove a classification theorem for
non-compact, connected, triangulated 2-manifolds? One of the difficulties
that arises in the non-compact case is that we no longer have a finite set of
simplices in the triangulation.
The following exercises illustrate what some of the complications of classifying non-compact 2-manifolds may be, even when we restrict to the orientable case:
Exercise 2.50. Below are some non-compact 2-manifolds. Are any of these
spaces are homeomorphic? (Beware! It may be harder than you think!) Can
you prove whether they are or are not homeomorphic? (See Figure 2.17)
Exercise 2.51. Let M be the non-compact 2-manifold made by taking the
two parallel planes {(x, y, 1)|x, y R} and {(x, y, 0)|x, y R}, removing
2.6. *NON-COMPACT SURFACES
43
disks {(x, y, 1)|(x a)2 + (y b)< 14 , a, b N} and {(x, y, 0)|(x a)2 + (y

b)< 14 , a, b N}, and finally gluing annuli {(x, y, z)|(x a)2 + (y b)= 14 , a, b
N, 0 z 1}. Is this space homeomorphic to any of the examples shown
above?
44
v0
v1
v2
v3
v0
v3
b
v4
v0
v4
v1
v2
Figure 2.18: The torus triangulated
v0
v0
v1
45
v2
v3
v0
v3
b
v4
v0
v4
v1
v2
Figure 2.19: The torus triangulated by its second baricentric subdivision
v0
46
v0
v1
v2
v3
v0
v3
b
v4
v0
v4
v1
v2
Figure 2.20: The Klein bottle triangulated
v0
v0
v1
47
v2
v3
v0
v3
b
v4
v0
v4
v1
v2
Figure 2.21: The Klein bottle triangled by its second baricentric subdivision
v0
48
Chapter 3
Fundamental group and

covering spaces
We do not need to calculate the Euler characteristic of a torus and a sphere
to intuit that they are not homeomorphic. The difference between them (or
between them and the surface of a two-holed doughnut) is associated with
something we describe as a holes. We need to make this concept precise.
What do we mean by a hole?
There are two ways to make this concept definite, both developed by
Henri Poincar`e at the turn of the 20th century. The first of these methods to
capture the intuitive idea of holes in a space is called the fundamental group
of a space. Unlike the Euler characteristic, which is a numerical invariant,
and orientation, which is a parity (+ or ) invariant, the fundamental group
is an algebraic group associated to the space. One would expect this more
complex invariant to carry more information about the space, and indeed it
often does.
Intuitively, the basic goal of the fundamental group is to recognize holes
in a space. If we think about a racetrack, we will have a good idea of how
the fundamental group works. From a starting point, a car measures its
progress by counting laps. Going once around is different from going twice
around. The Indy 500 involves going many times around the track. Going
backward around the track is frowned upon in competitive races, but if a car
did that, we would know how to count such a feat using negative numbers.
Counting number of times around the track is the most basic feature of the
fundamental group. But another feature of the racetrack also suggests a
basic idea of the fundamental group, namely, when a car has completed a
lap, the exact path of the car is not important as long as it stays on the
49
50 CHAPTER 3. FUNDAMENTAL GROUP AND COVERING SPACES
Figure 3.1: The Annulus
track. To make the analogy exact, we will insist that the car does return to
the exact point where it started.
Lets become a little more mathematical in our description of the fundamental group of the racetrack. A racetrack is known mathematically as an
annulus, which we could describe as A2 = {(x, y) R2 |1/2 x2 + y 2 1}
(See Figure 3.1).
We choose a point x0 on the annulus to be the base point. Then consider any continuous function f from a simple closed curve (
= S1 ) into the
1
annulus. We will choose a point on S to start, call it z, and require that
f (z) = x0 . Intuitively, that map f of S1 into the annulus goes around the
annulus some number of times. Physically, we can think of f as laying a
rubber band around the annulus with the point z placed on top of x0 . If we
think about the intuitive concept of how many times the map goes around
the annulus, we soon see that, in the rubber band model, if we could distort
the rubber band to a new position without lifting it out of the annulus,
then it would still go around the same number of times. Sliding around and
distorting a map are really giving a continuous family of maps of S1 into
the annulus. Such a continuous family of maps is called a homotopy. The
fundamental group just makes mathematically precise the idea of putting
maps of a circle into a space into equivalence classes via the idea of homotopy. To get to the idea of fundamental group, well first develop the idea
of a homotopy between maps from any space to any other space and then
specialize that idea to define the fundamental group.
3.1
Fundamental group
We all have a familiarity with the idea of a homotopy, because we have all
watched movies. Lets think of the physical objects that are being projected
onto the screen as the domain and think of the screen as the plane R2 . So,
3.1. FUNDAMENTAL GROUP
51
our domain might be two people, a dog, and a table. We will assume that all
of these objects are projected at every moment of our movie and that we are
thinking of an idealized projection system that projects that image at every
instant of time. So, this movie has uncountably many frames per second.
We are also thinking of the camera as being fixed throughout. When the
scene opens at time 0, the people, dog, and table are shown posed in some
no doubt interesting tableau. Every point in each person, dog, and table is
mapped to a point on the screen. Notice that this map is not 11. Points on
the inside of each object certainly are mapped to places where other points
map as well. At time 0, the action commences. The people move about,
gesticulating animatedly. The dog barks and wags its tail. The table just
sits there. At every instant of the film, each point in the people, dog, and
table are mapped via projection to a point in R2 , also known as the silver
screen. At time 1, the movie is over. It ends with the points of the people,
dog, and table being mapped to points on the screen. The first scene of the
move, which was a function from the domain (people, dog, table) into the
range (R2 ), was transformed via a continuous family of maps (the scenes at
each moment) into the last scene of the movie, which was another function
from the domain (people, dog, table) into the range (R2 ). The beginning
scene and final scene of our movie illustrate the idea of homotopic maps,
which we now define formally.
Definition (homotopic maps). Let f , g : X Y be two continuous functions. f and g are said to be homotopic if there is a continuous map
F : X [0, 1] Y such that F (x, 0) = f (x) and F (x, 1) = g(x) for all
x X. We denote that maps f and g are homotopic by writing f ' g. The
map F is called a homotopy between f and g.
If g is a constant map (mapping all points in X to a single point in Y )
and f ' g, then we say f is null homotopic.
Theorem 3.1. Given topological spaces X and Y , ' is an equivalence relation on the set of all continuous functions from X to Y .
We can think of a homotopy of two maps as a continuous 1-parameter
family of maps Ft from X to Y deforming f into g (i.e., F0 = f and
F1 = g).
Notice that in the gripping movie that we described earlier, the table
remained fixed throughout. So at each moment of the film, the function
when restricted to the table was the same for the whole duration of the
movie. That consistency on a subset of the domain gives rise to the idea of
a relative homotopy.

Definition (relative homotopy). Given topological spaces X and Y with
S X, then two continuous functions f , g : X Y are homotopic relative
to S if and only if there is a continuous function H : X [0, 1] Y such
that
H(x, 0) = f (x)
x X
H(x, 1) = g(x)
x X
H(x, t) = f (x) = g(x) x S & t [0, 1]
In other words, H is a 1-parameter family of maps Ht : X Y (t [0, 1])
which continuously deforms f into g while keeping the images of points in
the set S fixed.
Theorem 3.2. Given topological spaces X and Y with S X, being homotopic relative to S is an equivalence relation on the set of all continuous
functions from X to Y .
One of the motivations for studying homotopic maps is to capture the
idea of holes in a space. Our method of recognizing a hole is to think about
going around the hole. So this focus on marching around holes gives special
interest to the paths that we follow in going from place to place in the space
and the loops that may go around holes. These paths can be thought of as
maps from the interval into a space, so such maps are given a special name.
Definition (paths and loops). A continuous function : [0, 1] X is a
path. If (0) = (1) = x0 , then is a loop (or closed path) based at x0 .
Notice that a path is a function rather than a subset of the topological
space in which the image of the path sits. Two paths from the same starting
point to the same ending point are equivalent if, keeping the end points fixed
at all times, we can find a continuous family of paths that morph one path
into the other. The morphing is formalized as a homotopy relative to the
endpoints of [0, 1].
Definition (path equivalence). Two paths , are equivalent, denoted
, if and only if and are homotopic relative to {0, 1}. Denote the
equivalence class of paths equivalent to by [] (See Figure 3.2).
The concept of path equivalence applies to loops as wellsince loops are
paths. We will actually be most concerned with equivalences of loops, so we
give them special attention. Technically, a loop is a path, that is, a function
from [0, 1], whose endpoints are mapped to the same place, but intuitively,
a loop is a map from S1 into the space. That intuition is formalized using
the following wrapping map.
53
(0)
(0)
(1)
(1)
Figure 3.2: Path equivalence
Definition (standard wrapping map). The map : R1 S1 R2 defined

by t 7 (cos 2t, sin 2t) is called the standard wrapping map of R1 to S1 .
Theorem 3.3. Let be a loop into the topological space X. Then =
where is the standard wrapping map and is a continuous function from
S1 into X.
The above theorem allows us to think of a loop as a map from a circle
when it is useful for us to do so. This description also allows us to state a
useful characterization of triviality of a loop.
Definition (homotopically trivial loop). Let X be a topological space. A
loop is homotopically trivial or is a trivial loop if is equivalent to the
constant path e(0) where e(0) takes [0, 1] to (0).
Theorem 3.4. Let X be a topological space and let p be a point in X.
Then a loop = (where is the standard wrapping map and is a
continuous function from S1 into X) is homotopically trivial if and only if
can be extended to a continuous function from B2 into X.
We return now to the exploration of paths. The physical idea of walking
from point a to point b and then proceeding from there to point c yields the
natural idea of how to combine paths.
Definition (path product). Let , be paths with (1) = (0). Then their
product, denoted , is the path that first moves along , followed by
moving along . is defined by:

(2t),
0 t 12
(t) =
1
(2t 1),
2 <t1
(0)
(1)
(0)
(1)
Figure 3.3: Path Product
Notice the need to speed up in order to accomplish both the paths and
during the prescribed 1 unit of time allotted for a path.
Theorem 3.5. If 0 and 0 , then 0 0 .
Thus products of paths can be extended to products of equivalence
classes by defining [] [] := [ ]. Products of paths and products of
equivalence classes of paths enjoy the associative property.
Theorem 3.6. Given , , and , then ( ) ( ) and ([]
[]) [] [] ([] []).
If we think of a path as taking us from (0) to (1), then traversing
that same trail in reverse is the inverse path.
Definition (path inverse). Let be a path, then its path inverse 1 is the
path defined by 1 (t) = (1 t).
If we take a path and then take its inverse, that combined path is equivalent to not moving at all.
Theorem 3.7. Let be a path with (0) = x0 , then 1 ex0 , where
ex0 is the constant path ex0 : [0, 1] x0 . Stated differently, if is a path,
then 1 is homotopically trivial.
We now have all the ingredients to associate a group with a topological
space. This group has been designed to try to capture the idea of holes in
the space.
Definition (fundamental group). Let x0 X, a topological space. Then the
set of equivalence classes of loops based at x0 with binary operation [][] =
[ ] is a called the fundamental group of X based at x0 and is denoted
1 (X, x0 ). The point x0 is called the base point of the fundamental group.
Theorem 3.8. The fundamental group 1 (X, x0 ) is a group. The identity
element is the class of homotopically trivial loops based at x0 .
55
The fundamental group is defined for a space X with a specified base

point selected. However, for many spaces the choice of base point is not
significant, because the fundamental group computed using one base point
is isomorphic to the fundamental group using any other point. In particular,
path connected spaces enjoy this independence of base points.
Theorem 3.9. If X is path connected, then 1 (X, p)
= 1 (X, q) for any
points p, q X.
Since in path connected spaces the fundamental group is independent of
the base point (up to isomorphism), for such spaces X we sometimes just
write 1 (X) for the fundamental group without specifying the base point.
(Note. A corollary is a theorem whose truth is an immediate consequence
of the statement of a preceding theorem. A scholium is a theorem whose
truth is an immediate consequence of the proof of a preceding theorem, but
does not follow immediately from the statement of the preceding theorem.)
Scholium. Suppose X is a topological space and p, q X lie in the same
path component. Then 1 (X, p) is isomorphic to 1 (X, q).
We have now defined the fundamental group of a space, so lets find
the fundamental groups of some spaces. We begin with several example of
spaces that have trivial fundamental groups.
Exercise 3.10. We will use 1 to denote the trivial group:
1. 1 ([0, 1])
= 1.
2. 1 (S0 , 1)
= 1 where S0 is the zero-dimensional sphere {1, 1} R1 .
3. 1 (convex set)
= 1.
4. 1 (cone)
= 1.
5. 1 (cone over Hawaiian earring)
= 1.
6. 1 (Rn )
= 1 for n for n 1.
7. 1 (S2 )
= 1.
A space whose fundamental group is trivial is called simply connected.
Definition (simply connected). A path-connected topological space with trivial fundamental group is said to be simply connected or 1-connected.

Of course, the fundamental group would not serve a useful purpose if all
spaces were simply connected. The first example we will consider of a space
with non-trivial fundamental group is the circle. The following theorem will
require some significant work to prove.
Theorem 3.11. The fundamental group of the circle S1 is infinite cyclic,
that is, 1 (S1 )
= Z.
3.1.1
Cartesian products
To add to the spaces whose fundamental groups we can compute, let us

now look at the Cartesian products of spaces and observe that the fundamental group of a product of topological spaces is just the product of the
fundamental groups of the factors .
Theorem 3.12. Let (X, x0 ), (Y, y0 ) be path connected spaces. Then
1 (X Y, (x0 , y0 ))
= 1 (X, x0 ) 1 (Y, y0 ).
Exercise 3.13. Find:
1. 1 (T2
= S1 S1 );
2. 1 (D2 S1 );
3. 1 (S2 S1 );
4. 1 (S2 S2 );
5. 1 (S2 S2 S2 );
6. 1 (Sp1 . . . Spk ) where pj 2 for 1 j k.
The last 3 sections of the previous exercise present us with many examples of simply-connected spaces. We will see later that Cartesian products
of different collections of spheres yield topologically different spaces; however, at this point in the course, it is not obvious how we are going to detect
the differences in these spaces. These examples raise the question of what
additional ideas beyond the fundamental group we will need to show that
locally homeomorphic spaces arent actually homeomorphic. For now, we
leave this tantalizing question, but we will return to it in a later chapter.
For the moment we will be content with having established the fundamental
groups of quite a few spaces.
3.1.2
57
Induced homomorphisms
One of the standard techniques of mathematics is to explore the question of

how structure on one mathematical object is transported to another mathematical object via a map. We have now defined the fundamental group
for a space. Topological spaces are mapped to one another via continuous
functions. So we can ask how the fundamental group of one space is carried
to a target space via a continuous function.
Definition (induced homomorphism). Let f : X Y be a continuous
function. Then f : 1 (X, x0 ) 1 (Y, f (x0 )) defined by f ([]) = [f ] is
called the induced homomorphism on fundamental groups.
Definition (well-defined function). Let f : X Y and suppose that X
has a partition X of equivalence classes with equivalence relation . Then
f : X Y given by f ([x]) = f (x) is a well-defined function if and only
if f (x0 ) = f (x1 ) for all x0 x1 .
In other words a function defined on a set of equivalence classes is welldefined if its image is independent of the choice of representative of the
equivalence class.
Exercise 3.14. Check that for a continuous function f : X Y , the
induced homomorphism on the fundamental group f is well-defined.
Theorem 3.15. If g : (X, x0 ) (Y, y0 ), f : (Y, y0 ) (Z, z0 ) are continuous functions, then (f g) = f g .
Theorem 3.16. If f , g : (X, x0 ) (Y, y0 ) are continuous functions and f
is homotopic to g relative to x0 , then f := g .
Theorem 3.17. If h : X Y is a homeomorphism then h : 1 (X, x0 )
1 (Y, f (x0 )) is a group isomorphism. So homeomorphic spaces have naturally isomorphic fundamental groups.
The above theorem tells us that the fundamental group of a path connected space is a topological invariant, and hence we can establish that
two path connected spaces are not homeomorphic if we can show that they
have different (non-isomorphic) fundamental groups. Thus, the fundamental group helps to distinguish among spaces, but it does not always detect
the differences between spaces, as we have already seen among our examples
of spaces with trivial fundamental groups.
The correspondence between a topological space and an algebraic group
is extremely useful, because we can use algebra to answer topological questions (e.g., are two spaces not homeomorphic?) and, as we shall see later,
we can use topology to answer algebraic questions.
3.2
Retractions and fixed points
A retraction is a map that maps a space X into a subset A, keeping the

points in A fixed.
Definition (retraction, retract). Let A X. A continuous function r :
X A is a retraction if and only if for every a A, r(a) = a. If r : X A
is a retraction, then A is a retract of X.
Theorem 3.18. Let A be a retract of X and let i : A , X be the inclusion
map. Then i : 1 (A) 1 (X) is injective.
Question 3.19. Give an example to show why the conclusion of the previous
theorem does not follow merely from the assumption that A is a subset of
X.
Theorem 3.20. Let A be a retract of X, a0 A and r : A , X the
retraction. Then r : 1 (X, a0 ) 1 (A, a0 ) is surjective.
Theorem 3.21 (No Retraction Theorem for D2 ). There is no retraction
from D2 to its boundary.
Theorem 3.22. The identity map i : S1 S1 is not null homotopic.
Theorem 3.23. The inclusion map j : S1 R2 0 is not null homotopic.
A fixed point free map on a disk would imply the existence of retraction
of a disk to its boundary, which we know to be impossible. Thus the No
Retraction Theorem for D2 can be used to deduce the Brouwer Fixed Point
Theorem for D2 .
Theorem 3.24 (Brouwer Fixed-Point Theorem for D2 ). Let f : D2 D2
be a continuous map, then there is some x D2 for which f (x) = x.
Definition (strong deformation retract). Let A X. A continuous function r : X A is a strong deformation retraction if and only if there is a
homotopy R : X [0, 1] X such that
R(x, 0) = x
x X
R(x, 1) = r(x)
x X
R(a, t) = a
a A & t [0, 1].
If r : X A is a strong deformation retraction, then A is a strong deformation retract of X.
3.2. RETRACTIONS AND FIXED POINTS
59
We can re-word the definition in terms of homotopies: a strong deformation retraction r : X A is a retraction that is homotopic to the identity
on X relative to A.
Exercise 3.25. Show that R2 {2 points} strong deformation retracts onto
the wedge of two circles. In addition, show that R2 {2 points} strong
deformation retracts onto a theta curve. Are the wedge of two circles and
the theta curve homeomorphic? (See Figure 3.4)
Figure 3.4: The wedge of two circles (left) and theta curve right
Theorem 3.26. If r : X A is a strong deformation retraction and a A,

then 1 (X, a)
= 1 (A, a).
Exercise 3.27. Compute the fundamental groups of the following spaces:
1. 1 (solid torus
= D2 S1 )
=
2. 1 (R2 0)
=
3. 1 (house with 2 rooms)
= (See Figure 3.5.)
4. 1 (dunces hat)
= (See Figure 3.6)
Definition (contractible space). The topological space X is contractible if
and only if the identity map on X is null homotopic.
Theorem 3.28. A contractible space is simply connected.
Theorem 3.29. A retract of a contractible space is contractible.
Theorem 3.30. The House with Two Rooms is contractible.
Theorem 3.31. The Dunces Hat is contractible.
Figure 3.5: The House with Two Rooms
a
a
Figure 3.6: The Dunces Hat
3.3
Van Kampens Theorem, I
At this stage, S1 is the only source of examples of spaces with non-trivial fundamental groups. It seems a lot of work for not much payoff! So the question
now becomes: How can we compute the fundamental group of more complex
spaces, in particular, spaces with 1 (X) 6= 1. So far, we only have the ability
to compute the fundamental group of spaces two ways: taking the Cartesian
product of spaces whose fundamental groups we already know, and taking
strong deformation retracts involving one space whose fundamental group
we already know. The first allowed us to compute the fundamental group of
products of Sn , and, in particular, the fundamental group of the torus and
the annulus, and the second allowed us to calculate the fundamental group
of the annulus a second way, and also the fundamental group of other spaces
such as the dunces hat and the house with two rooms. But at this point we
cant compute the fundamental groups of any surfaces other than the torus,
let alone more general spaces.
To expand the number of spaces whose fundamental groups we can compute, our strategy will be to think about breaking up spaces into pieces
whose fundamental groups we know and analyzing how the fundamental
3.3. VAN KAMPENS THEOREM, I
61
group of the whole relates to the fundamental groups of the pieces.

Theorem 3.32. Let X = U V , where U and V are open and U V is
path connected, and let p U V . Then any element of 1 (X, p) has a
representative 1 1 n n , where each i is a loop in U based at p and
each i is a loop in V based at p.
In its full generality, Van Kampens Theorem will allow us to compute
the fundamental group of any space that, as in the theorem above, is broken up into two open and path connected subspaces with non-empty path
connected intersection. One of the challenges of computing 1 (X) is that
the representative 1 1 n n is not necessarily unique, even allowing for
the choice of representatives of i in 1 (U ) and i in 1 (V ).
Therefore, before we see the theorem in full generality, let us look at
some special cases of the theorem which are not only useful, but will allow
us to understand the underlying ideas used in the general version. In the
case where U V is simply connected, the representative 1 1 n n is in
a certain way unique, making the calculation of 1 (X) quite simple.
3.3.1
Van Kampens Theorem: simply connected intersection case
There are many ways of creating new groups from given groups. One of
them, which is useful in the context of fundamental groups, is the free
product. Here is how to think about the free product of two groups G
and H. Each element of the free product G H is just a finite string of
letters from G and H juxtaposed. So elements (called words) of G H
are of the form g1 h1 g2 h2 . . . gn hn , h1 g2 h2 . . . gn hn , g1 h1 g2 h2 . . . hn1 gn or
h1 g1 h2 g2 . . . gn1 hn . To multiply two such strings, we just juxtapose them.
If there are two gs next to each other, we multiply them in G and replace
them by the answer. We do the same procedure with two hs next to each
other. If any g or h is the identity in its respective group, then we erase
it. This means that the identities are really considered the same identity,
which is the identity of G H. And thats it.
Here is a more formal definition using group presentations:
Definition (free product of groups). Let G and H be two groups. The free
product of G and H, denoted GH is the group generated by all the elements
of G and H, subject only to the relations for G and H. In other words, if
G = hg1 , . . . , gn |r1 , . . . , ru i

and
H = hh1 , . . . , hm |t1 , . . . , tw i,
then
G H = hg1 , . . . , gn , h1 , . . . , hm |r1 , . . . , ru , t1 , . . . , tw i
The following version of Van Kampens Theorem tells us that when X
can be decomposed into two open subspaces that are path connected and
whose intersection is path connected and simply connected, then the fundamental group of the entire space is the free product of the fundamental
group of the two subspaces.
Theorem 3.33 (Van Kampens Theorem, simply connected intersection
case). Let X = U V , where U, V are open, path connected subsets of
X, U V is path connected and simply connected, and x U V . Then
1 (X, x)
= 1 (U, x) 1 (V, x).
Definition (wedge of two spaces). Let A and B be two disjoint spaces, with
points p A and q B. Then the wedge of A and B, denoted A B, is
defined as the quotient space A B/p q. In other words, we glue p to q.
Corollary 3.34. Let denote the wedge of two circles. Then 1 ()
=
Z Z.
Question 3.35. Let X be the wedge of n circles. What is 1 (X)?
Theorem 3.36. If A and B are each connected, then A B is connected.
If A and B are each path connected, then A B is path connected.
The following example shows the necessity of the hypotheses that U and
V are open in the above theorem.
Theorem 3.37. Let X be the wedge of two cones over two Hawaiian earrings, where they are identified at the points of tangency of the circles of
each Hawaiian earring, as in the figure below. Then 1 (X)
6 1.
=
Wedge of cones over Hawaiian earrings

Question 3.38. State conditions that suffice to ensure that 1 (A B)
=
1 (A) 1 (B).
3.4. FUNDAMENTAL GROUPS OF SURFACES
63
[We state the following theorem about the fundamental group of the
Hawaiian earring just for interest, but it does not fit into the flow of the
current discussion.
*Theorem 3.39. Show that 1 (Hawaiian earring) is not finitely generated,
in fact, 1 (Hawaiian earring) is not countably generated.]
3.3.2
Van Kampens Theorem: simply connected pieces case
Sometimes, one can break up a space into two pieces that are each simply
connected.
Theorem 3.40 (Van Kampens Theorem, simply connected pieces case).
Let X = U V where U and V are open, path connected, and simply connected subsets of X and U V is path connected. Then X is simply connected.
Exercise 3.41.
1. 1 (S2 ) = 1
2. 1 (Sn ) = 1
A good strategy when presented with a theorem is to ask yourself if the
hypotheses can be further weakened. A good habit when learning a new
theorem is to try to generate counterexamples that illustrate the limits of
theorems and the necessity of each hypothesis.
Question 3.42. Can you find an example where U and V are simply connected, but X = U V is not simply connected?
3.4
Fundamental groups of surfaces
Throughout this course, surfaces have been our motivating examples. It is

thus natural that we would like to find the fundamental group of all compact,
connected surfaces. In particular, we would like to know if the fundamental
group is an invariant that distinguishes them. Of course, we will eventually
also want to be able to calculate 1 for other spaces.
We have seen that the strong deformation retract of a space has the same
fundamental group as the space. Our strategy is to show that a punctured
surface strong deformation retracts to a space whose fundamental group we
know. We will then find 1 of closed surfaces if we can understand what
effect capping off the puncture with a disk has on the fundamental group.

A second strategy will be to find the fundamental group of the connected
sum of two surfaces given that we know each of their fundamental groups.
Exercise 3.43. Describe a strong deformation retract, together with its
fundamental group, of a once-punctured compact, connected, triangulated
2-manifold.
The answer to the above exercise gives us the fundamental group of a
once-punctured surface.
Exercise 3.44. Let M 2 be a compact, connected, triangulated 2-manifold,
and assume we write M 2 as U D2 , where U and D2 are open subspaces
of M 2 , D2 is an open disk, U D2 ' A2 is an open annulus, and p A2 .
Describe the non-trivial elements of 1 (U, p) that are trivial in 1 (M 2 , p).
Exercise 3.45. State and prove a theorem that allows you to calculate
1 (M 2 ) for any compact, connected, triangulated 2-manifold M 2 .
Exercise 3.46.
k
1. Describe a group presentation of # T2i .

i=1
k
2. Describe a group presentations of # RP2i .

i=1
There is a problem with group presentations, namely, it can be extremely

difficult to see whether two presentations represent the same group or not.
So, how do we know whether the groups we find above are or are not isomorphic? Well, one strategy for detecting differences among the fundamental
groups of surfaces is to abelianize the groups and see what we get. If the
abelianizations are not isomorphic, then neither were the original groups.
Exercise 3.47. Explicitly determine, using the Classification of Finitely
Generated Abelian Groups, what the abelianizations of the fundamental groups
found in the previous exercises are. What, if anything, distinguishes orientable from non-orientable surfaces? Are any of these abelianized groups
isomorphic? Is this invariant (the abelianized fundamental group) a complete invariant for closed surfacesi.e., is it sufficient to distinguish between
any two surfaces?
Thus far, we have succeeded in finding a representation of the fundamental group of a surface by viewing the surface a a punctured surface union a
disk. But our previous analysis of surfaces involved connected sums, so it is
3.4. FUNDAMENTAL GROUPS OF SURFACES
65
natural for us to think about how the fundamental group of a connected sum
is related to the fundamental groups of the two punctured surfaces whose
union makes it up.
Exercise 3.48. Suppose that M 2 = T1 #T2 where T1 and T2 are tori and
M 2 = U V where U is an open set of T1 homeomorphic to T1 (a disk), V is
an open set of T2 homeomorphic to T2 (a disk), and U V is homeomorphic
to an open annulus. Let p U V . We know from a previous exercise that
1 (U, p) is generated by two loops and . Likewise, 1 (V, p) is generated by
two loops and . Consider the loop that generates 1 (U V, p). Represent
in terms of the generators of 1 (U, p). Now represent in terms of the
generators of 1 (V, p). So the single loop is equivalent to two different loops
in M 2 . 1 (M 2 , p) is generated by {, , , }. What relations exist among
these generators? Give a presentation of 1 (M 2 , p) whose generators are
{, , , }.
The exercise above gives the basic insight into how we can deduce the
fundamental group of a union of two pieces from the fundamental groups of
the two pieces.
3.5
Van Kampens Theorem, II
Van Kampens Theorem says that if we can split a space X into two parts U
and V satisfying certain conditions, and we know what 1 (U ) and 1 (V ) are,
then we can find 1 (X) (up to isomorphism, as always). We have already
seen two special cases of this result. Let us now see Van Kampens Theorem
in full generality.
Group presentations are useful, concrete ways of representing 1 (although not completely trouble-free, as it is hard to tell when two presentations represent the same group). Van Kampens Theorem can be stated in
the language of group presentations.
Theorem 3.49 (Van Kampens Theorem; group presentations version).
Let X = U V , where U, V are open and path connected and U V is path
connected and non-empty. Let x U V .
Let 1 (U, x) = hg1 , . . . , gn |r1 , . . . , rm i, 1 (V, x) = hh1 , . . . , ht |s1 , . . . , su i
and 1 (U V, x) = hk1 , . . . , kv |t1 , . . . , tw i then
1 (X, x) = hg1 , . . . , gn , h1 , . . . , ht | r1 , . . . , rm , s1 , . . . , su ,
i (k1 ) = j (k1 ), . . . , i (kv ) = j (kv )i
where i, j are the inclusion maps of U V into U and V respectively.
Without the language of group presentations, Van Kampens Theorem
is stated as follows:
Theorem 3.50 (Van Kampens Theorem). Let X = U V where U, V are
open and path connected and U V is path connected and non-empty. Let
x U V . Then
1 (U, x) 1 (V, x)
1 (X, x)
=
N
where N is the smallest normal subgroup containing {i ()j (1 )}1 (U V,x)
and i, j are the inclusion maps of U V in U and V respectively. Note that
N is the set of products of conjugates of i ()j (1 ).
Exercise 3.51. Use Van Kampens theorem to explicitly calculate the group
presentation of the double torus T2 # T2 .
The following two exercises probably should have occurred in the fundamental group section.
Exercise 3.52. Let K be the space shown in Figure 3.7. What is 1 (K)?
Exercise 3.53. For any finitely presented group G, describe a 2-complex K
such that 1 (K) = G.
3.6. 3-MANIFOLDS
67
v
a
b
c
a
Figure 3.7: A six sided two cycle
3.6
3-manifolds
In the preceding chapter and in this chapter we have studied 2-manifolds

in some detail. In point of fact, we only need Euler characteristic and orientability to completely classify compact 2-manifolds. But what happens
when we look at higher-dimensional manifolds? There, is no longer a
useful invariant (orientability is always of interest). For any closed, compact, oriented 3-manifold M 3 , (M 3 ) = 0 (its zero for any odd-dimensional
closed, compact, oriented manifold). 3-manifolds are not as well understood
as 2-manifolds, but the fundamental group can be a useful invariant in the
study of 3-manifolds.
Let us look at two specific examples: lens spaces, which are compact
3-manifolds with no boundary, and knot exteriors, which are compact 3manifolds with boundary.
3.6.1
Lens spaces
Every compact, connected 3-manifold can be constructed by taking familiar

3-manifolds with boundaries (such as solid tori or solid double tori, or, in
general, solid n-tori) and gluing a pair of them together along their boundaries. The 3-manifolds that can be obtained by gluing a pair of solid tori
together along their boundaries are called lens spaces. These relatively simple 3-manifolds are completely classified.
Let p and q be relatively prime natural numbers. A (p, q)-lens space,
denoted L(p, q), indexed by p/q Q, can be defined in several different
ways. L(p, q) was first defined as an identification space that started with a
3-ball drawn in the shape of a lens (hence the name). The top and bottom
hemispheres of this lens are each divided into p triangle-like wedges. Each
triangle from the top hemisphere is identified with a triangle in the bottom
hemisphere that is a certain specified number (relatively prime to p) of
11111111111111
00000000000000
0000
1111
000000000000001111
11111111111111
0000
00000000000000
11111111111111
0000
1111
000000000000001111
11111111111111
0000
00000000000000
11111111111111
0000
1111
00000000
11111111
0000000000000011111111
11111111111111
0000
1111
00000000
00000000
11111111
00000000
11111111
00000000
11111111
00000000
11111111
Figure 3.8: Lens space as a quotient of a lens
J
Figure 3.9: Solid torus with meridian
triangles around the equator. The resulting quotient space is a lens space
(See Figure 3.8).
Definition (isotopy). A homotopy Ht : X Y (t I) is an isotopy if and
only if for every t in I, Ht is an embedding.
Exercise 3.54. Show that isotopies form an equivalence relation on the set
of all embeddings of X into Y .
Definition (meridian). Let V
= D2 S1 be a solid torus. A simple closed
curve J on Bd V is a meridian if and only if it bounds a disk in V (See
Figure 3.9).
Definition (longitude). Let V
= D2 S1 . A simple closed curve K on Bd V
is a longitude or longitudinal curve if and only if K represents the generator
3.6. 3-MANIFOLDS
69
of 1 (V ), that is, K goes once around V . A longitude can be isotoped to

intersect a meridianal curve once.
The meridian of a solid torus is unique up to isotopy, but the longitude
is not, since a longitude can spiral around the torus a number of full turns
as it goes around. A choice of longitude is called a f raming.
Lemma 3.55. Two simple closed curves and in T2
= S1 S1 are
homotopic if and only if they are isotopic.
Lemma 3.56. Given a meridian and longitude , {[], []} forms a basis
for 1 (Bd(D2 S1 )).
Lemma 3.57. Let {[], []} be a basis for 1 (Bd(D2 S1 ))
= Z Z. Then
q[] + p[] has a simple closed curve representative if and only if p and q are
relatively prime.
We can therefore use q[] + p[] (where p and q are relatively prime)
to mean a simple closed curve representative of that class. Two different
representatives that are simple will be isotopic.
Definition (lens space). Let V1 and V2 be two solid tori with meridians
i and chosen longitudes i respectively. Let h : Bd(V1 ) Bd(V2 ) be a
homeomorphism such that h(1 ) goes to a curve in the isotopy class q2 +p2
(where p,q are coprime). Then the quotient space V1 h V2 is the (p, q)-lens
space L(p, q).
Lens spaces can also be defined other ways, for example, as quotient
spaces of S3 under certain group actions. Beware that some authors use
L(p, q) to mean L(q, p).
Exercise 3.58. Use Van Kampens Theorem to explicitly calculate a group
presentation of 1 (L(p, q)).
3.6.2
Knots in S3
Definition (knot). Let i : D2 S1 S3 be a PL embedding, that is, an

injective continuous function that is linear on each simplex of a triangulation
of D2 S1 . Then K = i(0 S1 )
= S1 is a knot in S3 , and N (K) =
(Int(D2 S1 )) is an open regular neighborhood of K. The knot complement
or knot exterior of K is MK = S3 N (K).
Figure 3.10: The trefoil knot
If we look at a knot from above, we see a curve with crossings where it

goes over or under itself. For example, this picture is a picture of a trefoil
knot (See Figure 3.10):
If we are given a picture of a projection of a knot K into R2 where gaps
indicate where undercrossings occur and where all crossings are transverse
crossings of two arcs, then we can use the pictures, along Van Kampens
Theorem to produce a presentation of 1 (MK ). Roughly speaking, each arc
on the picture gives a generator and each crossing represents a relation.
For each arc in a knot projection, draw a labeled perpendicular arrow as
shown (See Figure 3.11):
The arrow ai , for example, represents the loop in MK obtained by starting well above the knot (at the base point chosen for 1 (MK )), going straight
down to the tail of ai , then going along ai under the knot, and finally returning to the starting point going straight from the head of ai .
Lemma 3.59. Every loop in MK is homotopic in MK to a product of ai s.
In other words, the loops {ai } generate 1 (MK ).
Lemma 3.60. At every crossing, such as that illustrated in Figure 3.12, the
following relation holds: acb1 = c or acb1 c1 = 1.
Theorem 3.61. Let K be a knot in S3 and let {ai } be the set of loops consisting of one loop for each arc in a knot projection of K as described above.
1
Then 1 (MK ) = {a1 , a2 , . . . , an |ai aj a1
where there is one relation of
k aj
1 1
the form ai aj ak aj for each crossing in the knot projection}.
3.6. 3-MANIFOLDS
71
a1
a3
a2
Figure 3.11: The arrows for the arcs of a trefoil knot
b
c
a
Figure 3.12: The arrows around a crossing
Figure 3.13: The unknot
Figure 3.14: The figure-8 knot
Exercise 3.62. Find the fundamental group of the complement of the unknot (See Figure 3.13).
Exercise 3.63. Find the fundamental group of the complement of the trefoil
knot.
Exercise 3.64. Find the fundamental group of the complement of the figure8 knot (See Figure 3.14).
3.7
Homotopy equivalence of spaces
Definition (homotopy equivalence). Two spaces X and Y are said to be

homotopy equivalent or to have the same homotopy type if there exist maps
f : X Y and g : Y X such that g f ' iX and f g ' iY , where
iX denotes the identity on X and iY denotes the identity on Y . We write
X Y to mean that X and Y are homotopy equivalent.
3.8. HIGHER HOMOTOPY GROUPS
73
Theorem 3.65. If X is a strong deformation retract of Y , then X and Y

are homotopy equivalent.
Theorem 3.66. If X Y then 1 (X)
= 1 (Y ).
In other words 1 doesnt distinguish spaces that have the same homotopy type.
3.8
Higher homotopy groups
The fundamental group of a space X is the set of homotopy equivalence

classes of maps of S1 into X with certain constraints. This idea can be
generalized to maps of Sn into X, giving the higher homotopy groups.
Recall that a loop can be thought of as either a map from [0, 1] = D1
to X where (Bd D1 ) = x0 for some point x0 X, or as a map S1 X,
and in fact we used the two ways interchangeably. We will use the second
way of looking at the higher-dimensional analogues of paths:
Definition (product of homotopy classes). Let X be a topological space and
x0 X. Let f , g : (Dn , Dn ) (X, x0 ), that is f and g are continuous
maps that take Dn 7 x0 . Let [f ] and [g] denote the respective homotopy
classes of these maps rel Dn . Then we define [f ] [g] to be the homotopy
class of:

(2x1 , x2 , . . . , xn ),
0 t 12
f g(x1 , x2 , . . . , xn ) =
1
(2x1 1, x2 , . . . , xn ),
2 <t1
where (x1 , x2 , . . . , xn ) Dn
Exercise 3.67. The set of homotopy classes of maps f : (Dn , Dn )

(X, x0 ), where Dn 7 x0 , with the product defined above, forms a group.
Definition (higher homotopy groups). The above mentioned group is called
the nth homotopy group of X based at x0 and is denoted n (X, x0 ). The
point x0 is called the base point of the homotopy group.
Theorem 3.68. Homotopy equivalent spaces have the same homotopy groups.
Figure 3.15: Two spheres with the same basepoint
Homotopy groups are generally hard to compute (even for Sn ). In the

next chapter we will develop the study of homology groups, which turn
out to be easier to compute and, hence, are generally more useful than the
higher homotopy groups in distinguishing higher dimensional topological
spaces from one another.
3.9. COVERING SPACES
3.9
75
Covering spaces
Theorem 3.69.
1. Any loop : I S1 can be written =
e, where
e : I R1 ,
e(0) = 0 and is the standard wrapping map.

2.
e(1) is an integer.
3. 1 and 2 are equivalent loops in S1 if and only if
f1 (1) =
f2 (1).
4. 1 (S1 ) = Z.
The preceding theorem outlines a method to compute 1 (S1 ). The preimage of any small open arc in S1 under the standard wrapping map is a
collection of open intervals in R1 and is a homeomorphism when restricted
to any single one of these open intervals. This map and its use in the above
theorem can be generalized to create the concept that is the subject of this
section. From the idea that the real line is covering the simple closed curved
via the wrapping map, the term covering space is used to refer to these
generalized wrapping maps. Just as the wrapping map was useful in our
method of computing the fundamental group of the circle, covering spaces
in general are useful for understanding the structure of the fundamental
groups of spaces.
e be connected, locally path connected
Definition (covering space). Let X, X
e X be a continuous function. Then the pair (X,
e p) is
spaces and let p : X
a covering space of X if and only if for each x X there exists an open
set U containing x such that p restricted to each component of p1 (U ) is a
e p) is a covering space of X, we refer to
homeomorphism onto U . When (X,
e
the space X as a cover of X and p as a covering map.
e = R1 , and p : R1 S1 be defined by p(t) =
Example 1. Let X = S1 , X
(cos t, sin t). Then (R1 , p) is a covering space of S1 .
e = S1 , and p : S1 S1 be defined by p(z) = z n ,
Example 2. Let X = S1 , X
where z C is a complex number with |z| = 1. Then (S1 , p) is a covering
space of S1 .
e are as Figure 3.16, and
Example 3. If X
= the wedge of two circles, Xs
e p) is a covering space of
ps are the maps indicated, then in each case, (X,
X.
Figure 3.16: Several coverings of the wedge of two circles
Figure 3.17:
77
Example 4. Let 2 be a PL non-separating, two-sided, properly embedded

surface in a connected 3-manifold M 3 . Gluing two copies of M 3 N (2 )
together gives a covering space of M 3 . See Figure 3.17.
e p) be a covering space of X. If x, y X, then
Theorem 3.70. Let (X,
1
1
|p (x)| = |p (y)|.
e p) is a covering space of a space X and
Definition (n-fold covering). If (X,
1
e p) is called an n-fold covering of X.
n = |p (x)| for some x X, then (X,
e is a cover of degree n.
We also say X
Example 5. The example above, namely, p : S1 S1 defined by p(z) = z n ,
where z C, is an n-fold covering of S1 by itself.
Exercise 3.71.
1. Describe two non-homeomorphic 2-fold covers of the Klein bottle.
2. Describe all non-homeomorphic 2-fold covers of the wedge of two circles.
3. Describe all non-homeomorphic 3-fold covers of the wedge of two circles.
There is a quick way of eliminating many possibilities from the potential
n-fold coverings of a surface.
Theorem 3.72. Let F be a compact connected surface and pn : Fe F be
an n-fold covering of F . Then F is a compact surface and (Fe) = n(F ).
Theorem 3.73. Let F be a compact connected orientable manifold and
p : Fe F be an n-fold covering of F . Then Fe is orientable.
Exercise 3.74.
1. Describe all non-homeomorphic 3-fold covers of the Klein bottle.
2. Describe all non-homeomorphic 2-fold covers of T2 # T2 .
3. Describe all non-homeomorphic 3-fold covers of T2 # T2 # T2 .
4. Describe all non-homeomorphic 3-fold covers of RP2 .
Exercise 3.75. Let F, G be compact, connected, orientable surfaces. Provide necessary and sufficient conditions for F to cover G.

Exercise 3.76. Let F, G be compact, connected, surfaces and let G be nonorientable. Provide necessary and sufficient conditions for F to cover G.
Exercise 3.77. Let F, G be compact, connected, surfaces. Provide necessary
and sufficient conditions for F to cover G.
e p) of X and a
Definition (lift of a function). Given a covering space (X,
e is
continuous function f : Y X, then a continuous function fe : Y X
e
called a lift of f if p f = f .
X
?

p

Y f /X
f
Theorem 3.78. Let (R1 , ) be the standard wrapping map covering of S1 .

Then any path f : [0, 1] S1 has a lift.
e p) is a cover of X, Y is connected, and f , g : Y X
e
Theorem 3.79. If (X,
are continuous functions such that p f = p g, then {y | f (y) = g(y)} is
empty or all of Y .
e p) be a cover of X and let f be a path in X. Then
e such that p(x0 ) = f (0), there exists a unique lift fe of f
for each x0 X
e
satisfying f (0) = x0 .
Question 3.81. Let p be a k-fold covering of S1 by itself and a loop in
S1 which when lifted to R1 by the standard lift has
e(0) = 0 and
e(1) = n.
For which integers n does lift to a loop in the k-fold covering?
e p) be a cover of X and
Theorem 3.82 (Homotopy Lifting Lemma). Let (X,
e
e
, be two paths in X. If
e, are lifts of , satisfying
e(0) = (0),
then
e e if and only if .
e p) is a cover of X, then p is a monomorphism (i.e.,
e into 1 (X).
11 or injective) from 1 (X)
The previous theorem implies that the fundamental group of a cover of
X is isomorphic to a subgroup of the fundamental group of the space X.
e p) be a cover of X, a loop in X, and x
e
e0 X
such that p(e
x0 ) = (0). Then lifts to a loop based at x
e0 if and only if
e
[] p (1 (X, x
e0 )).
79
Figure 3.18: Consider this picture
Exercise 3.85. Restate the proof of the fact that 1 (S1 )

= Z in terms of
covering spaces.
e p) be a covering space of X and let x0 X. Fix
1
e x
x
e0 p (x0 ). Then a subgroup H of 1 (X, x0 ) is in {p (1 (X,
e))}p(ex)=x0
e x
if and only if H is a conjugate of p (1 (X,
e0 )).
e p) be a covering space of X. Choose x X, then
1
e
|p (x)| = [1 (X) : p (1 (X))].
So, the index of the subgroup of 1 (X) corresponding to a finite covering
e
(X, p) equals the degree of the covering.
e p) of S1 and the subgroup p (1 (X))
e
Exercise 3.88. Describe a 3-fold cover (X,
1
of 1 (S ).
e p) be a covering space of X and x
e x0 X
e0 X,
with p(e
x0 ) = x0 . Also let f : Y X be continuous where Y is connected
and locally path connected and y0 Y such that f (y0 ) = x0 . Then there
e such that p fe = f and f (y0 ) = x
is a lift fe : Y X
e0 if and only if
e x
f (1 (Y, y0 )) p (1 (X,
e0 )). Furthermore, fe is unique.
e = R, (X,
e ) be the covering space of X
Exercise 3.90. Let X = S1 , X
given by the standard wrapping map, and Y as in Figure 3.19. When does
a map f : Y X not have a lift? Why is this example here?
e1 , p1 ) and (X
e2 , p2 ) be covering spaces
Definition (cover isomorphism). Let (X
e
e
of X. Then a map f : X1 X2 such that f is a homeomorphism and
p2 f = p1 is called a cover isomorphism.
Figure 3.19: A covering of the figure 8 or wedge of two circles
e1 , p1 ) and (X
e2 , p2 ) be covering spaces of X. Let x
Theorem 3.91. Let (X
e1
e
e
X1 and x
e2 X2 such that p1 (e
x1 ) = p2 (e
x2 ). Then there is cover isomorphism
e1 X
e2 with f (e
e1 , x
e2 , x
f :X
x1 ) = x
e2 if and only if p (1 (X
e1 )) = p (1 (X
e2 )).
e p) be a covering space. Then
Definition (covering transformation). Let (X,
e
a cover isomorphism from X to itself is called a covering transformation.
e p), is a group where the
The set of covering transformations, denoted C(X,
group operation is composition.
e p) for the covering space of the figure eight
Exercise 3.92. What is C(X,
shown in Figure 3.19.
e p) is a covering space of X and f C(X,
e p), then
f = IdXe if and only if f has a fixed point.
e p) be a covering space of X. If p (1 (X))/
e
Definition (regular cover). Let (X,
e
1 (X), then (X, p) is a regular covering space.
Question 3.94. Consider the second three-fold covering space of the figure
e which, when conjugated,
eight in Example 3. Find an element of p (1 (X))
e
is not in p (1 (X)).
e p) is a regular covering space of X and x1 , x2 X
e
e p) such that
such that p(x1 ) = p(x2 ), then there exists a unique h C(X,
h(x1 ) = x2 .
81
Question 3.96. The preceding theorem tells us that for a regular covering
space, there is a (unique) covering transformation carrying any point in the
set p1 (x) to any other point in the same set. Is this true of an irregular
covering space?
Theorem 3.97. A covering space is regular if and only if for every loop
either all its lifts are loops or all its lifts are paths that are not loops.
Exercise 3.98.
1. Describe all regular 3-fold covering spaces of a figure eight.
2. Describe all irregular 3-fold covering spaces of a figure eight.
3. Describe all regular 4-fold covering spaces of a figure eight.
4. Describe all irregular 4-fold covering spaces of a figure eight.
5. Describe all regular 3-fold covering spaces of a wedge of 3 circles.
6. Describe all regular 4-fold covering spaces of a wedge of 3 circles.
There is an important correspondence between the covering transformations of regular covers of X and the normal subgroups of 1 (X).
e p) be a regular covering space of X. Then C(X,
e p)
=
e
e
e
1 (X)/p (1 (X)). In particular, C(X, p) = 1 (X) if X is simply connected.
Exercise 3.100. Observe that the standard wrap map is a regular covering
map of S1 by R1 . Describe the covering transformations for this covering
space. Describe the covering map that maps R2 to the torus T2 and describe
the covering transformations for this covering space.
Definition (semi-locally simply connected). A space X is called semi-locally
simply connected if and only if every x X is contained in an open set U
such that every loop in U based at x is homotopically trivial in X.
Note that U need not be simply connected itself.
Theorem 3.101 (Existence of covering spaces). Let X be connected, locally path connected, and semi-locally simply connected. Then for every
e p) of X and x
e such that
G < 1 (X, x0 ) there is a covering space (X,
e0 X
e
e
p (1 (X, x
e0 )) = G. Furthermore, (X, p) is unique up to isomorphism.
Definition (universal cover). A connected, locally path connected cover is
called universal if and only if its fundamental group is trivial.

Corollary 3.102. Every connected, locally path connected, semi-locally simply connected space has a universal covering space.
Exercise 3.103. Find universal covers for the Klein bottle, torus, and proe p)
jective plane. Then show explicitly that C(X,
= 1 (X).
3.10
Theorems about groups
The algebra of a fundamental group tells us something about the topology

of its corresponding space. It is also possible to use topology to study
algebraic groups. One can construct a complex that represents any finitely
generated group, and use the fundamental group and covering spaces to
deduce properties of the group.
Theorem 3.104. Every tree is simply connected.
Theorem 3.105. Let G be a graph, and T be a maximal tree in G. Then if
{e1 , . . . en } is the set of edges that are not in T , 1 (G) = Fn , the free group
on n generators; and there is a system of generators that are in one-to-one
correspondence with the edges {e1 , . . . en }.
Corollary 3.106. A subgroup H of a free group Fn is always a free group.
e of a wedge of n-circles.
Question 3.107. Describe a regular k-fold cover X
e given k and n? What does this
What is the number of generators of 1 (X),
tell us about the normal subgroups of finite index of the free group with n
generators?
Exercise 3.108.
1. Let F be a free group on n letters. Let G F be of finite index k and
contain 7 free generators. What can the value of n be?
Exercise 3.109. Let Sk be the set of all k-fold covers of K, the Klein Bottle.
1. Describe Sk .
2. Describe all subgroups of 1 (K) of finite index.
Exercise 3.110. Let F be the closed orientable surface of genus 2, and
G = 1 (F ). Show that all subgroups of G of finite index k are isomorphic.
Chapter 4
Homology
We have seen that the fundamental group of a space uses loops that are not
null-homotopic to understand the holeyness of a space. For distinguishing
spaces, the fundamental group is a valuable tool, but has some challenges
associated with it. For one thing, the fundamental group is in general not
an abelian group. This feature makes it difficult to determine in general
whether two fundamental groups might be the same if they have different
presentations. The Classification of Abelian Groups allows us to determine
which abelian groups are isomorphic and which are not, so it would be nice to
associate meaningful abelian groups with spaces. Of course, the fundamental group of a space may be giving us more refined information. However,
in many cases, we dont need that much information to capture some of the
holeyness of a space. For example, recall that when we used the fundamental group to distinguish surfaces from one another, the abelianizations
of the fundamental groups were sufficient to make the distinctionsthe whole
fundamental group was not used.
The second shortcoming of the fundamental group for measuring holeyness of spaces is that it captures only holes that are surrounded by loops.
It does not measure holes surrounded by spheres, for example. A 3-ball
with a sub-ball removed still has trivial fundamental group. In this chapter, then, we will introduce the concept of homology, which associates with
spaces abelian groups that measure holes in all dimensions and are often
easy to compute.
We will study the homology of simplicial complexes. The ideas we develop here can be generalized to apply to more comprehensive classes of
spaces, but we will not study those generalizations in this course.
83
84
CHAPTER 4. HOMOLOGY
v2
e2
v1
e6
e3
v5
e4
1
e1
v3
e7
e5
v4
Figure 4.1: A two complex
4.1
Z2 homology
In the fundamental group, we concerned ourselves with loops. In homology

we will be concerned with cycles, which will help us quantify in some way the
number of holes in a surface or space. Cycles will come in all dimensions.
Intuitively, a 1-cycle is a sum of edges in a simplicial complex that makes
a loop or loops. Consider the space made up of two 2-simplexes (1 , 2 ),
seven edges (ei , i = 1, 2, ..., 7), and five vertices (vi , i = 1, 2, ..., 5) as shown.
Then there are several cycles (loops of edges) in this example. For notational
consistency that will become clear in our later development of Z homology,
we will write each loop as a sum of edges, but you should think of the
loop as a physical loop sitting inside our space and should be wondering,
Why dont we just use a union symbol? e1 + e2 + e4 + e5 is one loop,
e1 + e2 + e6 + e7 + e5 is another loop, and e6 + e7 + e4 is a third loop.
Which ones represent holes in our space? Well, e1 + e2 + e4 + e5 bounds
a filled in piece comprised of the two 2-simplices, so it doesnt represent a
hole, and it should be considered equivalent to 0. e1 + e2 + e6 + e7 + e5 and
e6 + e7 + e4 both go around the same hole, so they should be viewed as the
same. In fact, we know that these loops are homotopic to each other by
a homotopy that takes the edges e5 + e1 + e2 to e3 , so if we were looking
at 1 then wed have two equivalent loops in the sense of the fundamental
group. However, we want to keep things as simple as possible, so we will
give a simple criterion for declaring that these two cycles are Z2 equivalent.
Consider the edges that the two cycles e1 + e2 + e6 + e7 + e5 and e6 + e7 + e4
differ by, namely, e1 + e2 + e4 + e5 . That difference is a cycle that bounds
an object composed of two 2-simplices. So the intuitive concept that we will
formalize below is that two cycles will be declared to be equivalent if their
difference is the boundary of some collection of simplices in our space.
4.1. Z2 HOMOLOGY
4.1.1
85
Simplicial Z2 homology
Recall the definitions of a simplex, a face of a simplex, and a finite simplicial

complex. We now define the objects used to define homologychains, cycles,
and equivalence classes of cycles. When dealing with Z2 homology, we could
use slightly simpler notation; however, instead, we will use notation in this
section that will later be valuable when we deal with Z homology where
orientation comes into play.
P
Definition (Z2 n-chain). The Z2 -sum of k n-simplices ki=1 in (where the
sum is mod 2) is called a (Z2 ) n-chain.
Definition (Z2 -boundary of a simplex). The Z2 -boundary of any n-simplex
n is
n
X
n
( ) =
in1
i=0
where the
in1 s
are the (n 1)-dimensional faces of n .
Definition
(Z2 -boundary of an n-chain). The Z2 -boundary of the n-chain
Pk
n
i=1 i is
k
k
X
X
n
(
i ) =
(in )
i=1
i=1
where the sum is mod 2; i.e., if a simplex appears an even number of times
it cancels.
For example, 1 + 2 + 2 + 3 + 3 + 3 = 1 + 3 .
Notice that boundary of an n-chain is an (n 1)-chain.
Definition (Z2 n-cycle). A Z2 n-cycle is an n-chain whose Z2 -boundary is
zero.
Theorem 4.1. For any Z2 n-chain C, ((C)) = 0, that is, the boundary
of any n-chain is an (n 1)-cycle.
Definition (Z2 -equivalence of cycles). In a complex (K, T ) n-cycles An and
B n are Z2 -equivalent if and only if An B n (= An + B n ) = (C n+1 ), where
C n+1 is an (n + 1)-chain. The equivalence class that An is a member of will
be denoted by [An ].
[Note: We use the minus sign in An B n in anticipation of Z homology
to come.]
86
CHAPTER 4. HOMOLOGY
Question 4.2. List all the equivalence classes of 0-, 1- and 2-cycles in
a triangulated sphere, torus, projective plane, Klein bottle, and then each
compact, connected, triangulated 2-manifold.
Definition (Z2 nth -simplicial homology). The Z2 nth -homology of a finite simplicial complex (K, T ), denoted Hn ((K, T ); Z2 ) is the additive group
whose elements are equivalence classes of cycles under the Z2 -equivalence
defined above, where [An ] + [B n ] := [An + B n ].
Exercise 4.3. Show that the addition for Hn ((K, T ); Z2 ) defined above is
well-defined.
Theorem 4.4. Let (K, T ) be a finite simplicial complex with triangulation
T , then Hn ((K, T ); Z2 ) is an abelian group.
Theorem 4.5. Let K be connected and let (K, T ) be a finite simplicial
complex with triangulation T , then H0 ((K, T ); Z2 ) is Z2 .
Homology would be an uninteresting concept if it depended on the particular triangulation selected; however, in fact, the homology groups of a
space are independent of which triangulation is used. For now, lets accept
the following two theorems, which are the basis of that fact, and later well
look at an outline of how they are proved. The first of these theorems simply
states that subdividing the triangulation of a complex does not change the
homology that we compute.
Theorem 4.6. Let (K, T ) be a finite simplicial complex and let T 0 be a
subdivision of T . Then Hn ((K, T ); Z2 ) is isomorphic to Hn ((K, T 0 ); Z2 ) .
Theorem 4.7. Let K be a subset of Rn and let T and T 0 be triangulations
of K that make (K, T ) and (K, T 0 ) finite simplicial complexes. Then there
is a triangulation T 00 that is a subdivision of both T and T 0 .
Corollary 4.8. Let K be the underlying subset of a finite simplicial complex.
Then Hn ((K, T ); Z2 )
= Hn ((K, T 0 ); Z2 ) for any triangulations T and T 0 of
K.
This corollary means that the homology of a set is independent of the
particular triangulation chosen for it.
4.1.2
CW Z2 -homology
The difficulty with simplicial complexes is that they frequently have a lot of
simplices. To compute Hn , we would need to consider the boundaries of all
4.1. Z2 HOMOLOGY
87
the possible subsets of simplices in the triangulation, which is a laborious

undertaking. But, on the bright side, we saw in the previous section that
Hn ((K, T ); Z2 ) is the same for any triangulation T of K. We want to further
simplify the task of computing homology groups by considering CW decompositions, which we encountered earlier when we discussed simpler ways to
compute the Euler Characteristic.
As we did before, lets look at the triangulation subdivision result a bit
backwards. We can view the subdivision result as saying that grouping simplices into bigger simplices does not affect the Z2 -homology of the simplicial
complex. Lets push that strategy even further by showing that we can compute the same homology groups by breaking K up into cells that may or
may not be simplices rather than sticking to triangulations. Our definition
of CW complex of any dimension captures the idea of cellulating K rather
than triangulating K, as we saw before for 2-complexes.
First, lets note that a simplicial complex of any dimension can be written
as the disjoint union of the interiors of the simplices in its triangulation.
(Recall that the boundary of a vertex is the empty set.)
Definition (interior of a simplex). For each simplex , let the interior of
, denoted by Int() or , be = .
Theorem 4.9. Let (K, T ) be a finite simplicial complex where T = {i }i=1,..k .
Then K
k
G
K=
i ,
1
where t denotes the disjoint union.

To simplify our computation of homology groups, it is often useful to
view K as a union of open cells other than interiors of simplices, because
then we can use many fewer cells. We accomplish such a decomposition of
K by grouping simplices to create larger cells.
Definition (open cell complex, open cell decomposition, and CW decomposition). Let (K, T ) be a finite simplicial complex. Suppose we write K as the
disjoint union of Ak00 , Ak11 , Ak22 , . . . , Akn3 such that each Aki i is homeomorphic
to the interior of a ki -cell, each Aki i is the disjoint union of interiors of simk
plices in T , and the closure of each Aki i minus Aki i is the union of Aj j (each
of dimension less than ki ). Let S be the set of Aki i s. Then (K, S) will be
called an open cell complex and S will be called an open cell decomposition
of K. S is equivalent to what is sometimes called a CW decomposition of
K.
88
CHAPTER 4. HOMOLOGY
Examples 1.
1. Let (K, T ) be a simple closed curve with triangulation shown (a square).
Then one vertex and one open cell would form a CW decomposition of
K.
Figure 4.2: A 1-cycle with four edges
2. Let (K, T ) be a 3-simplex with triangulation shown (a tetrahedron).

Then one vertex, one open 2-cell, and one open 3-cell would form a
CW decomposition of K. This example shows that it is not necessary
to have every dimension represented.
Figure 4.3: A 2-cycle with four faces
3. Let (K, T ) be the 2-complex created by starting with a triangle and

attaching a disk whose boundary goes around it twice. We could look
at this complex as a quotient space of a 6-sided triangulated disk with
opposite edges identified with arrows all going clockwise around the
hexagon. It could exist in R4 . Then a CW decomposition of K could
consist of one open 2-cell, one open 1-cell, and one vertex, as shown.
We can define Z2 -homology for a CW complex exactly as we defined it
for a triangulated complex. As expected we will soon see that the homology
groups are exactly the same as they are when we compute the Z2 -homology
groups using the triangulation rather than the open cell decomposition. Here
then are the relevant definitions.
4.1. Z2 HOMOLOGY
89
v
a
b
c
a
Figure 4.4: A six sided two cycle
Definition (boundary of open cell). Let (K, T ) be a simplicial complex. Let

A be an element of an open cell decomposition (K, S) of (K, T ). So A is
homeomorphic to the interior of a k-cell and A = union of ik where each
ik is a k-simplexPin T . Then A := the sum of all (k 1)-cells in S that
are contained in
ik .
P
Definition (boundary of an n-chain). Let B =
Ai where each Ai is an
open n-cell
P in a CW complex (K, S). Then B is an n-chain and the boundary
of B is
Ai .
Recall that the sum is done mod 2.
Definition (n-cycle). An n-cycle in an open cell complex (K, S) is a sum
of n-cells in S, that is, an n-chain, with empty boundary.
Definition (equivalence of cycles). A k-cycle Ak is Z2 -equivalent to a kcycle B k if and only if there exists a (k + 1)-chain C k+1 such that C k+1 =
Ak B k (= Ak + B k because we are summing mod 2).
Definition (Z2 nth -CW homology). The Z2 nth homology group of a CW
complex (K, S), denoted Hn ((K, S); Z2 ), is the group whose elements are
the Z2 -equivalence classes of n-cycles with addition defined by [An ] + [B n ] =
[An + B n ].
Finally, we confirm our goal that the CW homology is the same as simplicial homology, and once again we will accept this theorem and its corollary
as true and later outline proofs of them.
Theorem 4.10. Let (K, S) be a CW decomposition of the finite simplicial
complex (K, T ). Then for each n, Hn ((K, S); Z2 ) = Hn ((K, T ); Z2 ).
Corollary 4.11. Hn ((K, T ); Z2 ) does not depend on the triangulation or
CW decomposition of K used to compute Hn .
90
CHAPTER 4. HOMOLOGY
Since any triangulation or CW decomposition of K yields the same homology groups, we often suppress the triangulation or CW decomposition
in the notation and just refer to Hn (K; Z2 ).
Question 4.12. For each space K below, describe a CW decomposition of
it and describe Hn (K; Z2 ) for n = 0, 1, 2, 3, . . .:
1. the sphere.
2. the torus.
3. the projective plane.
4. the Klein bottle.
5. the double torus.
6. any compact, connected, triangulated 2-manifold.
7. the M
obius band.
8. the annulus.
9. Two (hollow) triangles joined at a vertex.
4.2. HOMOLOGY FROM PARTS, SPECIAL CASES
91
Question 4.13. What is Hn (K; Z2 ) if dim(K)< n?

Question 4.14. What are Hn (G; Z2 ) for n = 0, 1, 2, . . . for a graph G?
Question 4.15. What are Hn (Sk ; Z2 ) for n = 0, 1, 2, . . . and k = 0, 1, 2, . . .?
Question 4.16. For any n, what is Hn (M n ; Z2 ) where M n is a connected
n-manifold?
Question 4.17. For any n, what is Hn (M n ; Z2 ) where M n is a connected
n-manifold with non-empty boundary?
Question 4.18. What are Hn (T ; Z2 ) for n = 0, 1, 2, . . . for a solid torus T ?
4.2
Homology from parts, special cases
One strategy for computing homology is to divide the space into parts and
see how the homology groups of the whole are related to the homology
groups of its parts. The next several theorems should remind the reader
of special cases of Van Kampens theorem from the theory of fundamental
groups.
Theorem 4.19. Let K L denote the wedge of finite simplicial complexes
K and L. Then Hn (K L; Z2 )
= Hn (K; Z2 ) Hn (L; Z2 ) for n > 0.
Theorem 4.20. Suppose M is a finite simplicial complex with subcomplexes
K and L such that K L = M . If Hn (K L; Z2 ) = 0, then Hn (M ; Z2 )
=
Hn (K; Z2 ) Hn (L; Z2 ).
Exercise 4.21. Compute Hn (K; Z2 ) (n = 0, 1, 2, . . .) for each complex K
below.
1. A wedge of k circles.
2. A wedge of a 2-sphere and a circle.
3. A 2-sphere union its equatorial disk.
4. A double solid torus.
Theorem 4.22. Let M be a finite simplicial complex with subcomplexes K
and L such that K L = M . Suppose n 2 and that for every n-cycle Z in
K L, Z Z2 0 in K and Z Z2 0 in L, and Hn1 (K L; Z2 ) = 0. Then
Hn (M ; Z2 )
= Hn (K; Z2 ) Hn (L; Z2 ).
The case n = 1 is slightly different, namely, suppose for every 1-cycle Z
in K L, Z Z2 0 in K and Z Z2 0 in L, and H0 (K L; Z2 ) = Z2 . Then
H1 (M ; Z2 )
= H1 (K; Z2 ) H1 (L; Z2 ).
92
CHAPTER 4. HOMOLOGY
Again, the previous theorem might remind the reader of a a special case
of Van Kampens Theorem.
Question 4.23. State and prove a theorem giving the Z2 -homology groups
for connected compact 2-manifolds, using the theorem above and the connected sum decomposition of each 2-manifold.
4.3. CHAIN GROUPS AND INDUCED HOMOMORPHISMS
4.3
93
Chain groups and induced homomorphisms
Definition (chain group). Let K be a finite simplicial complex. The chain

group Cn (K; Z2 ) is the group whose elements are the n-chains of K where
the group operation is Z2 chain addition. So each n-simplex of K is a
generator for the abelian group Cn (K; Z2 ), which is therefore the direct sum
of k copies of Z2 where k is the number of n-simplexes in K.
We will now investigate how simplicial maps from one complex to another
induce homomorphisms on homology groups.
Definition. Let (K, T ) and (L, S) be finite simplicial complexes. A map
g : (K, T ) (L, S) is a simplicial map if and only if for any n =
{v0 v1 v2 . . . vn } T , there is a simplex = {w0 w1 w2 . . . wk } S such
that for each i = 0, 1,
. . . n, g(vi ) = wj for some j P
{0, 1, 2, . . . , k} and
P2,
n
n
for eachPpoint x =
i=0 i = 1, then
i=0 i vi where 0 i 1 and
g(x) = ni=0 i g(vi ). Note that g(vi ) is not necessarily distinct from g(vk ),
and that g is linear on each simplex of T .
Question 4.24. Let (K, T ) and (L, S) be finite simplicial complexes. Let
f : (K, T ) (L, S) be a simplicial map. Find a natural definition for
the induced map f# : Cn (K; Z2 ) Cn (L; Z2 ). Notice that the map f# is a
homomorphism of n-chains of K to n-chains of L.
Recall that the boundary of an n-chain is an (n 1)-cycle.
Theorem 4.25. The map : Cn (K; Z2 ) Cn1 (K; Z2 ) is a group homomorphism.
Theorem 4.26. Let f : K L be a simplicial map, and let f# be the
induced map f# : Cn (K; Z2 ) Cn (L; Z2 ). Then for any C Cn (K; Z2 ),
(f# (C)) = f# ((C)). In other words, the diagram:
Cn (K; Z2 ) Cn1 (K; Z2 )
f
f# y
y#
Cn (L; Z2 ) Cn1 (L; Z2 )

commutes.
Definition. Let (K, T ) and (L, S) be finite simplicial complexes. Let f :
K L be a simplicial map. The induced homomorphism f : Hn (K; Z2 )
Hn (L; Z2 ) is defined by f ([A]) = [f# (A)].
94
CHAPTER 4. HOMOLOGY
Note that at this point, it is not clear that f takes cycles to cycles. The
following theorem asserts that all is well.
Theorem 4.27. Let (K, T ) and (L, S) be finite simplicial complexes. Let
f : K L be a simplicial map. Then the induced homomorphism f :
Hn (K; Z2 ) Hn (L; Z2 ) is a well-defined homomorphism.
In order to deal with continuous functions rather than simply with simplicial maps, we sketch a proof below that any continuous function can be
approximated by a simplicial map that is homotopic to it.
Theorem 4.28. Let (K, T ) and (L, S) be finite simplicial complexes and
f : K L be a continuous function. Then there exists a subdivision T 0 of
T and a simplicial map g : (K, T 0 ) (L, S) such that f is homotopic to g
and g is a simplicial map.
The proof of this theorem uses the following.
Definition (star of a vertex). Let (K, T ) be a simplicial complex. Then
the star of v, where v is a vertex of T , is defined as St(v, T ) = {
T |v is a vertex of }.
Sketch of Proof. Subdivide T to obtain T 0 such that for every vertex v of
T 0 , f (St(v, T 0 )) Int(St(w, S)) for some vertex w of S. For each v define
g(v) = w where w is any vertex in S for which f (St(v, T 0 )) Int(St(w, S)).
Extend g linearly over each simplex. Show that g is a simplicial map and g
is homotopic to f .
2.
Theorem 4.29. Let T and T 0 be two different triangulations of a finite
simplicial complex K and let (L, S) be a finite simplicial complex. Suppose
f : (K, T ) (L, S) and g : (K, T 0 ) (L, S) are simplicial maps such that
f is homotopic to g. Then the induced homomorphism f : Hn (K; Z2 )
Hn (L; Z2 ) is the same homomorphism as g .
Sketch of Proof. Find a simplicial approximation of the homotopy H
between f and g that is linear with respect to a common subdivision T of
T and T 0 . For any n-cycle A in (K, T ), show that [f# (A)] = [g# (A)] by
using the (n + 1)-chain H# (A [0, 1]).
2.
These results allow us to extend our definition of the induced homomorphism f to apply to continuous functions f : K L (rather than just
applying to simplicial maps) and to know that f is well-defined.
Theorem 4.30. Let K and L be finite simplicial complexes such that K is
a strong deformation retract of L. Then Hn (K; Z2 )
= Hn (L; Z2 ).
95
Theorem 4.31. If A, B are finite simplicial complexes and A is homotopy

equivalent to B, then Hn (A; Z2 )
= Hn (B; Z2 ).
In other words homology does not distinguish between a space and a
strong deformation retract of it, or between homotopy equivalent spaces in
general. So, once again, homology is not a complete invariant. It is however,
very useful, and, in general, easier to compute than homotopy.
Theorem 4.32. Hn (Dunces hat)
= 0 for all n 6= 0.
Theorem 4.33. Hn (House with 2 rooms; Z2 )
= 0 for all n 6= 0.
Theorem 4.34. If f : K L is a homeomorphism between finite simplicial
complexes K and L, then f : Hn (K; Z2 ) Hn (L; Z2 ) is an isomorphism.
In other words, homology groups are topological invariants.
96
4.4
CHAPTER 4. HOMOLOGY
Applications of Z2 homology
Recall the definition of a retraction.

Definition (retraction). Let K L, then a continuous function r : L K
is a retraction if and only if r(x) = x for all x K.
Theorem 4.35 (No Retraction Theorem). Let M n be an n-manifold with
M n 6= . Then there is no retraction r : M n M n .
Theorem 4.36 (n-dimensional Brouwer Fixed Point Theorem). For every
continuous function f : Dn Dn there exists a point x Dn such that
f (x) = x.
We will develop a proof of the Borsuk-Ulam Theorem below through a
sequence of preliminary theorems.
Lemma 4.37. Let M n be a finite, triangulated, connected n-manifold without boundary. Let f : M n M n be a simplicial map. Then f : Hn (M n ; Z2 )
Hn (M n ; Z2 ) is onto if and only if f# (M n ) = M n .
Theorem 4.38. Let f : S1 S1 be an antipode preserving map (that is,
for every x S1 , f (x) = f (x)). Then f : H1 (S1 ; Z2 ) H1 (S1 ; Z2 ) is
onto.
Theorem 4.39 (Borsuk-Ulam for S2 ). Let f : S2 R2 be a continuous
map. Then there is an x S2 such that f (x) = f (x).
Hint. If there existed a counterexample to the Borsuk-Ulam Theorem,
then there would be a continuous function from the equator to itself that
is antipode preserving where the induced homomorphism on the Z2 first
homology is not onto.
The above strategy for proving the Borsuk-Ulam Theorem in dimension
2 can be extended to work in all dimensions.
Theorem 4.40. Let f : Sn Sn be an antipode preserving map (that is,
for every x Sn , f (x) = f (x)). Then f : Hn (Sn ; Z2 ) Hn (Sn ; Z2 ) is
onto.
Theorem 4.41 (Borsuk-Ulam). Let f : Sn Rn be a continuous function.
Then there is an x Sn such that f (x) = f (x).
The following lemmas will allow us to prove the n-dimensional Jordan
Curve Theorem.
4.5. Z2 MAYER-VIETORIS THEOREM
97
Lemma 4.42. Let f : Sn1 Sn be a simplicial map. Then there exist Z2

n-chains An and C n such that (An ) = (C n ) = f# (Sn1 ) and An C n =
Sn .
Lemma 4.43. Let h : Sn1 Sn be a topological embedding. Then there
exists an > 0 such that if f : Sn1 Sn is a simplicial map such that
d(f (x), h(x)) < for all x Sn1 , then f# (Sn1 ) does not bound an n-chain
in the -neighborhood of h(Sn1 ).
The proof of the following Theorem is in the appendix.
Theorem 4.44 (n-dimensional Jordan Curve Theorem). Let h : Sn1 Sn
be a topological embedding. Then h(Sn1 ) separates Sn into two components
and is the boundary of each.
Theorem 4.45. A topologically embedded compact, triangulated, connected
(n 1)-dimensional manifold separates Rn .
Theorem 4.46 (Invariance of Domain or Invariance of Dimension Theorem). n-manifolds are not homeomorphic to m-manifolds if n 6= m.
These are just some of the fundamental theorems of topology that we
can prove using Z2 homology.
4.5
Z2 Mayer-Vietoris Theorem
The goal of this section is to describe how the homology groups of a complex
are related to the homology groups of pieces of the complex. We begin with
some theorems that relate cycles in the whole complex to cycles and chains
in its parts.
Theorem 4.47. Let K be a finite simplicial complex and K0 and K1 be
subcomplexes such that K = K0 K1 . If A0 , A1 are (n 1)-cycles in K0
and K1 respectively and if A0 Z2 A1 in K, then there is a (n 1)-cycle C
in K0 K1 such that A0 Z2 C in K0 and A1 Z2 C in K1 .
subcomplexes such that K = K0 K1 . Let Z be a Z2 n-cycle on K. Then
there exist Z2 n-chains W0 and W1 in K0 and K1 respectively such that:
1. Z = W0 + W1 and
2. (W0 ) = (W1 ) is an (n 1)-cycle C in K0 K1 .
98
CHAPTER 4. HOMOLOGY
Furthermore, if Z = W00 + W10 where Wi0 is an n-chain in Ki , and C 0 =

(W00 ) = (W10 ) is an (n1)-cycle, then C 0 is Z2 -equivalent to C in K0 K1 .
Question 4.49. Let K be a simplicial complex and K0 and K1 be subcomplexes such that K = K0 K1 . Describe the natural homomorphisms below,
and verify that they are homomorphisms:
1. : Hn (K0 K1 ; Z2 ) Hn (K0 ; Z2 ) Hn (K1 ; Z2 ).
2. : Hn (K0 ; Z2 ) Hn (K1 ; Z2 ) Hn (K; Z2 )
3. : Hn (K; Z2 ) Hn1 (K0 K1 ; Z2 )
Van Kampens Theorem was phrased in terms of quotient groups of free
products. The analogous theorem in homology couches its result in terms
of exact sequences.
Definition (exact sequences).
1. Given a sequence (finite or infinite) of groups and homomorphisms:
1
. . . G1 G2 G3 . . .
then the sequence is exact at G2 if and only if im 1 = ker 2 .
2. The sequence is called an exact sequence if it is everywhere exact (except at the first and last groups if they exist).
Theorem 4.50.
1. 0 A B is exact at A if and only if is one-to-one.
2. B C 0 is exact at B if and only if is onto.
3. 0 A B 0 is exact if and only if is an isomorphism.
4. 0 A B C 0 is exact if and only if C

= B/(A).
Theorem 4.51 (Z2 Mayer-Vietoris). Let K be a finite simplicial complex

and K0 and K1 be subcomplexes such that K = K0 K1 . The sequence
Hn (K0 K1 ; Z2 ) Hn (K0 ; Z2 ) Hn (K1 ; Z2 ) Hn (K; Z2 ) Hn1 (K0
K1 ; Z2 ) . . . using the natural homomorphisms above, is exact.
Question 4.52. Use the Mayer-Vietoris Theorem to compute Hn (M ; Z2 )
for every compact, triangulated 2-manifold M 2 .
4.5. Z2 MAYER-VIETORIS THEOREM
99
Figure 4.5: Here are two interesting spaces
Question 4.53. What compact, triangulated 2-manifolds are not distinguished from one another by Z2 -homology?
Question 4.54. Use the Mayer-Vietoris Theorem to compute Hn (K; Z2 )
for the complexes K pictured in Figure 4.5.
Exercise 4.55. Use the Mayer-Vietoris Theorem to find the Z2 homology
of the following:
1. Sn .
2. a cone over a finite simplicial complex (K, T ) (that is, the finite simplicial complex (K St v, T St v) created by adding a vertex v in a higher
dimension and creating from each simplex in T , a new simplex with
one more vertex, v.).
3. a suspension over a finite simplical complex (K, T ) (that is, the finite
simplicial complex created by gluing two cones over K along K).
4. RPn (= Sn with antipodal points identified).
Question 4.56. What are Hn (L(p, q); Z2 ) for n = 0, 1, 2, . . .?
100
4.6
CHAPTER 4. HOMOLOGY
Introduction to simplicial Z-homology
In this section, we will refine the concept of homology to include the idea
of the orientation of simplices. In Z2 -homology, orientation of simplices was
not mentioned. For example, Z2 -homology made no distinction between the
edge [v0 v1 ] and the edge [v1 v0 ]. That edge was either present or absent.
Failing to consider the orientation of simplices caused Z2 -homology to fail
to detect distinctions among some 2-manifolds. Recall that Z2 -homology
does not distinguish between T2 , the torus, and K2 , the Klein bottle, for
example. Note that orientability is a basic difference between those two
spaces.
So our strategy now is to repeat the same development that we used for
Z2 -homology, but now taking orientation into account. The effect of this
shift in perspective is that now our basic objects will be oriented simplices
(in effect, the order of the vertices matters) and they will come with an
integer coefficient. These coefficients make the objects of Z-homology become abstractions of the directly geometrical interpretation available when
we were talking about Z2 -homology. What does five times a cycle mean geometrically? What does negative two times a simplex mean geometrically?
From one point of view, these are just abstract generalizations, but in another sense, the coefficients retain some geometrical meaning. In particular,
the negative of a simplex means the same simplex with the opposite orientation. Including these coefficients will allow us to draw finer distinctions
among spaces. For example, we will find that T2 and K2 have different second homology groups (that is, second Z-homology groups), and, in fact, we
will see that the Z-homology groups will successfully distinguish all compact,
connected surfaces from one another.
Z-homology is the most common type of homology used, so when people
refer to the homology of a space without specifying what type of homology,
they are referring to Z-homology.
4.6.1
Chains, boundaries, and definition of simplicial Z-homology
The definition of homology groups (that is, Z-homology groups) of a finite

simplicial complex (K, T ) involves chains, boundaries, cycles, and the concept of when cycles are equivalent. The building blocks for these ideas are
the simplices of (K, T ) just as in Z2 -homology; however, in this case the ordering of the vertices makes a difference, so we will be working with ordered
vertices of each simplex.
4.6. INTRODUCTION TO SIMPLICIAL Z-HOMOLOGY
101
Definition (oriented simplex). Let (K, T ) be a simplicial complex. An oriented n-simplex n = [v0 v1 . . . vn ] is an n-simplex {v0 v1 . . . vn } in T along
with a particular ordering of its vertices up to even permutation. That is,
the orientation of an n-simplex is the choice of an equivalence class of all
possible ordering of its vertices, where two orderings are equivalent if and
only if they differ by an even permutation.
In other words, the oriented n-simplex with underlying (unoriented) nsimplex {v0 v1 . . . vn } and whose chosen ordering of the vertices differs from
[v0 v1 vn ] by an even permutation represents the same oriented n-simplex
n . Any ordering of those same vertices that differs from [v0 v1 vn ] by an
odd permutation is the negative of that oriented n-simplex, and is denoted
by n .
Definition (n-chain group). In order to define the n-chain group Cn (K, T ),
we need to say what the elements are and what the addition operation is. The
n-chain group Cn (K, T ) is the free abelian group whose set of generators
consists of one oriented n-simplex for each n-simplex in (K, T ).
Example 2. Let (K, T ) be the simplicial complex in the plane where
T
= { {(0, 0)(0, 1)(1, 0)}, {(0, 0)(0, 1)}, {(0, 1)(1, 0)},
{(0, 0)(0, 1)}, {(0, 1)(1, 0)}, {(1, 0)(0, 0)},
{(0, 0)}, {(0, 1)}, {(1, 0)}, {(0, 1)}} .
So K is a filled in triangle and a hollow triangle as pictured in Figure 4.6.
Figure 4.6: A simplicial complex
Then C2 (K, T ) has one generator which is an oriented 2-simplex, perhaps, [(0, 1)(0, 0)(1, 0)]. So C2 (K, T ) is isomorphic to Z. We are free to
choose any ordering of the vertices in selecting this generator of C2 (K, T ).
C1 (K, T ) is a free abelian group on 5 generators, which could be the
oriented 1-simplices [(0, 0)(0, 1)], [(1, 0)(0, 1)], [(0, 1)(0, 0)], [(0, 1)(1, 0)],
102
CHAPTER 4. HOMOLOGY
and [(1, 0)(0, 0)]. We are free to choose any ordering of the vertices that we
wish for each 1-simplex.
C0 (K, T ) is a free abelian group on 4 generators, namely, [(0, 0)], [(0, 1)],
[(1, 0)], and [(0, 1)]. Of course, in the case of vertices, there is no choice
of order involved.
Definition (Z n-chain). Each element of Cn (K, T ) is called an n-chain.
It
Pk
n
is the formal sum of oriented n-simplices with coefficients in Z, i=1 ci i .
The definition of Cn (K, T ) already tells us what the sums of n-chains
are, but lets just record it here.
P
Definition (Z n-chain sum). The Z-sum of two n-chains C = ki=1 ci in
P
P
Pm
n
and D = j=1 di jn is given by C + D = ki=1 ci in + m
j=1 di j .
For example, 31n +2n +2n 53n = 31n +22n 53n and 1n +(1n ) = 0,
the empty n-chain.
Definition (induced orientation on a subsimplex). A non-oriented (n 1)subsimplex {v0 . . . vbj . . . vn } of an oriented n-simplex = [v0 . . . vn ] will have
the orientation induced by if it is oriented as j = (1)j [v0 . . . vbj . . . vn ].
Definition (boundary of a simplex). The Z-boundary of an oriented nsimplex n = [v0 . . . vn ] is
( n ) =
n
X
kn1 =
n
X
(1)k [v0 . . . vbk . . . vn ]
k=0
k=0
kn1 s
are oriented (n 1)-dimensional boundary simplices of

That is, the
n with their orientations induced from n .
Exercise 4.57. Find the boundary of the oriented 2-simplex = [v0 v1 v2 ]
and the boundary of the oriented 3-simplex = [w0 w1 w2 w3 ].
Repeat the procedure for and . What is the relationship between
the boundary of and the boundary of ? What is the relationship between
the boundary of and the boundary of ?
Theorem 4.58. For any n-simplex
() = ().
Definition (Z-boundary
of an n-chain). The boundary (implicitly Z-boundary)
P
of the n-chain ki=1 ci in is
!
k
k
X
X
n
ci i =
ci (in ).
i=1
i=1
4.6. INTRODUCTION TO SIMPLICIAL Z-HOMOLOGY
103
Note that boundary of an n-chain is an (n 1)-chain.

Definition (Z n-cycle). An n-cycle (implicitly a Z-n-cycle) is an n-chain
whose boundary is zero.
Theorem 4.59. For any n-chain C, ((C)) = 0, that is, the boundary of
any n-chain is an (n 1)-cycle.
Definition (equivalence of cycles). In a complex (K, T ), n-cycles An and
B n are equivalent if and only if An B n = (C n+1 ), where C n+1 is an
(n + 1)-chain. The equivalence class that An is a member of will be denoted
by [An ].
Definition (nth -simplicial homology). The nth -homology of a finite simplicial complex (K, T ), denoted Hn ((K, T ); Z) or just Hn (K), is the additive
group whose elements are equivalence classes of cycles under the equivalence
defined above where [An ] + [B n ] := [An + B n ].
Exercise 4.60. Show that the addition for Hn ((K, T ); Z) defined above is
well-defined.
Theorem 4.61. Let (K, T ) be a finite simplicial complex with triangulation
T , then Hn ((K, T ); Z) is an abelian group.
Theorem 4.62. Let K be connected and let (K, T ) be a finite simplicial
complex with triangulation T , then H0 ((K, T ); Z) is Z. If K has r components, then H0 ((K, T ); Z) is a free abelian group whose rank is r.
Exercise 4.63. Give an oriented triangulation of a M
obius band such that
the central circle forms a 1-cycle Z. Show that the boundary 1-cycle Z 0 is
equivalent to either 2Z or 2Z.
Definition ( operator). Let (K, T ) be a finite simplicial complex. The
boundary operator restricted to the n-chain group Cn (K, T ) gives a homomorphism : Cn (K, T ) Cn1 (K, T ). We sometimes will write n to
emphasize what n we are restricting ourselves to.
Theorem 4.64. 2 = 0. In other words, for n 2 the composition of
: Cn (K, T ) Cn1 (K, T ) and : Cn1 (K, T ) Cn2 (K, T ) is the
trivial homomorphism. To extend the statement to n = 1, we can define
C1 (K, T ) = 0
Elements of the kernel of the boundary operator are the cycles.
104
CHAPTER 4. HOMOLOGY
Definition (Zn ). Let (K, T ) be a finite simplicial complex. Then the group
ker : Cn (K, T ) Cn1 (K, T ) is denoted by Zn (K, T ). Its elements are
called cycles.
Elements of the image of the boundary operator are the cycles that
bound.
Definition (Bn ). Let (K, T ) be a finite simplicial complex. Then the group
Im : Cn+1 (K, T ) Cn (K, T ) is denoted by Bn (K, T ). Its elements are
called bounding cycles.
These allow us to give an algebraic definition of the homology groups.
Definition (algebraic definition of Hn ). Let (K, T ) be a finite simplicial
complex. Then Hn (K; Z)
= Zn (K; Z)/Bn (K; Z).
We refer to this fact as homology is cycles mod boundaries. You should
verify that this definition is equivalent to the first definition of homology
given.
As with Z2 homology, integer homology would be an uninteresting concept if it depended on the particular triangulation selected. Fortunately, as
in the Z2 case, one can show that integer homology is independent of triangulation. So we will write Hn (K; Z) or Hn (K) instead of Hn ((K, T ); Z)
to denote the nth -homology group of the complex K. As with Z2 -homology,
Z-homology can be computed from a CW-decomposition. Also, just as in
the case of Z2 -homology, Z-homology does not distinguish between homotopy equivalent spaces. Feel free to use these facts in doing the following
exercises.
Question 4.65. For each space K below, describe Hn (K; Z) for n = 0, 1,
2, 3, . . .:
1. the sphere.
2. the torus.
3. the projective plane.
4. the Klein bottle.
5. the double torus.
6. any compact, connected, triangulated 2-manifold.
7. the M
obius band.
105
8. the annulus.
9. Two (hollow) triangles joined at a vertex.
Question 4.66. What is Hn (K) if dim(K)< n?
Question 4.67. What are Hn (G) for n = 0, 1, 2, . . . for a graph G?
Question 4.68. What are Hn (Sk ) for n = 0, 1, 2, . . . and k = 0, 1, 2, . . .?
Question 4.69. For any n, what is Hn (M n ) where M n is a connected nmanifold?
Question 4.70. For any n, what is Hn (M n ) where M n is a connected nmanifold with non-empty boundary?
Question 4.71. What are Hn (T ) for n = 0, 1, 2, . . . for a solid torus T ?
4.7
Chain groups and induced homomorphisms
The ideas in this section are small variations of those that occurred in the
Z2 -homology section.
We will now investigate how simplicial maps from one complex to another
induce homomorphisms on homology groups.
Definition. Let (K, T ) and (L, S) be finite simplicial complexes.
Pk Letnf :
(K, T ) (L, S) be a simplicial map. Suppose that C =
i=1 ci i
Cn (K; Z). To get a natural definition for the induced map f# : Cn (K; Z)
Cn (L; Z) we first define f# ([v0 . . . vn ]) = 0 if the images of the vi s are not
distinct, and otherwise f# ([v0 . . . vn ]) = [f (v0 ) . . . f (vn )]. We can then exP
tend f# to all elements of Cn (K; Z) by f# (C) = ki=1 ci f# (in ). Notice that
the map f# is a homomorphism of n-chains of K to n-chains of L.
Theorem 4.72. Let f : K L be a simplicial map, and let f# be the
induced map f# : Cn (K; Z) Cn (L; Z). Then for any C Cn (K; Z),
(f# (C)) = f# ((C)). In other words, the diagram:
Cn (K; Z) Cn1 (K; Z)
f
f# y
y#
Cn (L; Z) Cn1 (L; Z)

commutes.
106
CHAPTER 4. HOMOLOGY
Definition. Let (K, T ) and (L, S) be finite simplicial complexes. Let f :

K L be a simplicial map. The induced homomorphism f : Hn (K)
Hn (L) is defined by f ([A]) = [f# (A)].
As before, the induced map can be defined for continuous functions by
using the simplicial approximation theorems we proved before. Specifically
the induced map for a continuous function f is defined to be the induced
map for any simplicial approximation that is homotopic to f .
f : K L be a continuous function. Then the induced map f : Hn (K; Z)
Hn (L; Z) is a well-defined homomorphism.
The essence of the fact that homotopic continuous functions induce the
same homomorphism on homology was used in proving the preceding theorem, but we record the fact below for clarity.
f, g : K L be homotopic continuous functions. Then f = g .
Theorem 4.75. Let K and L be finite simplicial complexes such that K is
a strong deformation retract of L. Then Hn (K; Z)
= Hn (L; Z).
Theorem 4.76. If A, B are finite simplicial complexes and A is homotopy
equivalent to B, then Hn (A)
= Hn (B).
Theorem 4.77. If f : K L is a homeomorphism between finite simplicial
complexes K and L, then f : Hn (K; Z) Hn (L; Z) is an isomorphism.
In other words, homology groups are topological invariants.
4.8
Relationship between fundamental group and

first homology
There is a close connection between the fundamental group of a space and

its first homology group.
Theorem 4.78. Suppose that K is a finite, connected simplicial complex.
Then H1 (K; Z) ' (1 (K))/[1 (K), 1 (K)], that is, the first homology group
of K is isomorphic to the abelianization of the fundamental group of K.
4.9. MAYER-VIETORIS THEOREM
107
Sketch of proof. Let : 1 (K) H1 ((K, T ); Z) be the map that takes

an element [] of 1 (K) to the element [# (S1 )] of H1 ((K, T ); Z) where
is understood to be a simplicial map from S1 into (K, T ). First note that
is a well-defined, surjective homomorphism.
It remains to show that Ker() is the commutator
subgroup of 1 (K).
P
For this purpose, let [] Ker(). Let C 2 = ki=1 ci i2 be a 2-chain such
that (C 2 ) = # (S1 ). Since (C 2 ) = # (S1 ), for each edge of any i2 that is
not in # (S1 ), that edge must be cancelled out when computing (C 2 ). So
we can create an abstract 2-manifold whose 2-simplexes are the i2 s where
we take several copies of a simplex depending on its coefficient. Using the
Classification Theorem of oriented 2-manifolds, we can recognize that is
in the commutator subgroup of 1 (K).
4.9
Mayer-Vietoris Theorem
The goal of this section is to describe how the homology groups of a complex
are related to the homology groups of pieces of the complex. We begin with
some theorems that relate cycles in the whole complex to cycles and chains
in its parts.
Theorem 4.79. Let M be a finite simplicial complex with subcomplexes K
and L such that K L = M . Suppose n 2 and that for every n-cycle Z
in K L, Z Z 0 in K and Z Z 0 in L, and Hn1 (K L) = 0. Then
Hn (M )
= Hn (K) Hn (L).
The case n = 1 is slightly different, namely, suppose for every 1-cycle
Z in K L, Z Z 0 in K and Z Z 0 in L, and H0 (K L) = Z. Then
H1 (M )
= H1 (K) H1 (L).
subcomplexes such that K = K0 K1 . If A0 , A1 are (n 1)-cycles in K0
and K1 respectively and if A0 A1 in K, then there is an (n 1)-cycle C
in K0 K1 such that A0 C in K0 and A1 C in K1 .
subcomplexes such that K = K0 K1 . Let Z be an n-cycle on K. Then
there exist n-chains W0 and W1 in K0 and K1 respectively such that:
1. Z = W0 W1 and
2. (W0 ) = (W1 ), and (W0 ) is an (n 1)-cycle C in K0 K1 .
108
CHAPTER 4. HOMOLOGY
Furthermore, if Z = W00 W10 where Wi0 is an n-chain in Ki , and C 0 =

(W00 ) = (W10 ) is an (n 1)-cycle, then C 0 is equivalent to C in K0 K1 .
Question 4.82. Let K be a simplicial complex and K0 and K1 be subcomplexes such that K = K0 K1 . Describe the natural homomorphisms below,
and verify that they are homomorphisms:
1. : Hn (K0 K1 ) Hn (K0 ) Hn (K1 ).
2. : Hn (K0 ) Hn (K1 ) Hn (K)
3. : Hn (K) Hn1 (K0 K1 )
Theorem 4.83 (Mayer-Vietoris). Let K be a finite simplicial complex and
K0 and K1 be subcomplexes such that K = K0 K1 . The sequence
Hn (K0 K1 ) Hn (K0 ) Hn (K1 ) Hn (K) Hn1 (K0 K1 ) . . .
using the natural homomorphisms above, is exact.
Question 4.84. Use the Mayer-Vietoris Theorem to compute Hn (M 2 ) for
every compact, triangulated 2-manifold M 2 .
Question 4.85. Use the Mayer-Vietoris Theorem to compute Hn (K) for
the complexes K pictured below.
Figure 4.7: Here are two interesting spaces with interesting homology
Exercise 4.86. Use the Mayer-Vietoris Theorem to find the homology of

the following:
1. Sn .
2. a cone over a finite simplicial complex (K, T ) (that is, the finite simplicial complex (v K, v T ) created by adding a vertex v in a higher
dimension and creating from each simplex in T , a new simplex with
one more vertex, v.).
4.9. MAYER-VIETORIS THEOREM
109
3. a suspension over a finite simplical complex (K, T ) (that is, the finite
simplicial complex created by gluing two cones over K along K).
4. RPn (= Sn with antipodal points identified).
Question 4.87. What are Hn (L(p, q)) for n = 0, 1, 2, . . .?
110
CHAPTER 4. HOMOLOGY
Appendix A
Review of Point-Set
Topology
Definition (topology). A topology on a set X is a non-empty collection T
of subsets of X such that:
1. T
2. X T
[
3.
U T for any {U }J T
4.
J
n
\
Ui T for any finite collection{U1 , . . . , Un } T .
The set X with a given topology T is called a topological space, and

is denoted by (X, T ) or simply X if the topology is implicitly understood.
Elements of X are called points of X.
Definition (open set, neighborhood). Given a topological space X with a
topology T , the sets in T are called open sets. If x contains X and U is an
open set containing x, we say U is a neighborhood of x.
Note that the definition of neighborhood includes the fact that it is open.
Not all authors use that definition, some say a set N 3 x is a neighborhood
of x if there is an open set U such that x U N . In this case N need not
be open.
Example 1.
1. For any set X, T = {, X} is called the trivial topology on X.
111
112
APPENDIX A. REVIEW OF POINT-SET TOPOLOGY
2. R with
T = { U = iJ (ai , bi )|ai < bi , ai , bi R} {} ,
called the standard topology on R. All non-empty open subsets of R
consist of unions of open intervals.
The second example above suggests that we can define a topology by
giving a subcollection of open sets that generate the toplogy:
Definition (basis for a topology). A basis of a topology is a sub-collection
B T such that for every U T , U can be written as the union of an
arbitrary collection of sets in B.
Equivalently, B T is a basis for T if for every U T , and every x U ,
there is some B B such that x B B.
Definition (limit point). Given a toplogical space (X, T ), and A X,
x X is a limit point of A if any neighborhood of x contains a point in A
different from x. The set of all limit points of A is denoted by A0 .
Note that if x A0 , then x may or may not be in A itself. Also, if x A,
it may or may not be in A0 .
Definition (closed set). Given a toplogical space (X, T ), then if C X
contains all of its limit points, C is called a closed set.
Theorem A.1. Given a toplogical space (X, T ), then if C X closed, then
C = X U for some U T .
Note that there can be (and often are) sets that are neither open nor
closed, and sets that are both open and closed.
Exercise A.2. Define a topology in terms of the closed sets. What conditions must a collection of closed sets possess so that their complements form
a topology?
Definition (closure of a set). Let (X, T ) be a topological space. Then the
closure of A X, denoted by A or Cl A, is A A0 .
Theorem A.3. Let (X, T ) be a topological space. Then the closure of
A X, A is the smallest closed set containing A, in other words, it is the
intersection of all closed sets of (X, T ) that contain A.
113
Definition (interior of a set). Let (X, T ) be a set X endowed with the
topology T . Then the interior of A X denoted by A or Int A, is the largest

open set contained in A, in other words, the union of all open sets of T that
are contained in A.
Definition (boundary of a set). Let (X, T ) be a set X endowed with the
topology T . Then the boundary of A X denoted by Bd A or A, is Cl A
Int A.
Definition (subspace topology). Let A X, a topological space with topology T . Then A inherits a topology from X in a natural way: TA = {U
A|U T }. A with this topology is called a (topological) subspace of X.
Definition (continuity). Let X and Y be two sets with their respective
topologies, and f : X Y a function such that for each open set V in
Y , f 1 (V ) is open in X. Then f is called a continuous function from X to
Y.
In other words f : X Y is continuous if and only if the preimage of
every open set in Y is open in X. Note that the image of an open set under
a continuous function need not be open.
Definition (quotient topology). Let X be a topological space and Y be a
set and let f : X Y be a surjective (onto) function. Then the quotient
topology on Y is defined by saying that a subset U is open in Y if and only
if f 1 (U ) is open in X.
Theorem A.4. Let X be a topological space, f : X Y a surjective function onto the space Y that has the quotient topology. Then f is continuous
and, furthermore, if the topology on Y included any more sets in addition
to those in the quotient topology, then f would fail to be continuous.
Definition (identification or quotient space). Let X be a topological space
and X be a partition of X. Let f : X X be the map taking each point
x X to [x], the element of X containing x ([x] is called the equivalence
class of x). Then X with the quotient topology is called the identification
space of X under the partition X . The quotient topology applied in this
manner to a partition of X is sometimes called the identification topology.
Definition (homeomorphism). Let X and Y be two sets with their respective
topologies, and f : X Y a one-to-one and onto (or surjective) function
which is continuous, and such that its inverse function f 1 is continuous.
Then f is called a homeomorphism, and we say that X and Y are homeomorphic.
114
Definition (Hausdorff). A topological space X is Hausdorff if for all x 6= y

points in X, there are disjoint open sets U and V such that x U and
y V.
Definition (cover). Let A be a subset of X, a topological space. A cover of
A is a collection U of subsets of X such that A is contained (as a subset) in
their union. An open cover is a cover in which all the sets of the cover are
open. A subcover is a subcollection of U that is a cover of A.
Definition (compactness). Let A be a subset of X, a topological space. A
is compact if every open cover of A have a finite subcover.
Definition (connectedness). Let X be a topological space. Then X is connected if whenever X is decomposed into the union of two disjoint open sets
U and V (so X = U V and U V = ), either U or V must be the empty
set.
Definition (path). Let p : I X be a continuous function, where I = [0, 1]
with the subspace topology inherited from the standard topology on R and X
a topological space. Suppose p(0) = x and p(1) = y. Then p is called a path
in X from x to y.
Definition (path connectedness). A space X is path connected if for any
two points x and y in X there is a path from x to y
Definition (local path connectedness). A space X is locally path connected
if for every x X and neighborhood U of x, there is a path connected open
set V containing x such that V U .
The following example is a space with infinitely many circles all tangent
at a point. This Hawaiian earring sometimes is useful in seeing the limits
of generality of some of our theorems
Example 2 (Hawaiian earring).

(
(
))

[
1 2
1

2
2
Sn R Sn = (x, y) x
+y = 2

2n
4n
n=1
Definition (metric). Let X be a set and d : X X R0,+ , a function to

the non-negative real numbers, satisfying, for all x, y, and z in X:
1. d(x, y) = 0 if and only if x = y
2. d(x, y) = d(y, x)
115
3. d(x, z) d(x, y) + d(y, z) (triangle inequality)
is called a d-metric on X.
Definition (metric space). Let X be a space with a metric d. Then B =
{Bd (x, r) = {y X|d(x, y) < r}} is a basis for a topology on X, called the
metric topology. Bd (x, r) is called the open ball of radius r centered at x.
116
Appendix B
Review of Group Theory

Definition (group). A group is a set G along with a binary operation G
G G, denoted by (multiplicative notation) or + (additive notation)
satisfying the following three conditions:
1. There exists an element 1 G, called the identity element such that
g 1 = 1 g for all g G. When we are dealing with multiplicative
groups we will write 1G to denote the identity of G.
2. For every g G there exists an element g 1 G, called the inverse of
G, such that g g 1 = g 1 g = 1.
3. For all g1 , g2 , g3 G we have (g1 g2 ) g3 = g1 (g2 g3 ) (associativity)
In additive notation the identity element is denoted by 0: g+0 = 0+g = g
g G, the inverse by g so that we write g+(g) = (g)+g = 0. Additive
notation is usually reserved for commutative or abelian groups:
Definition (abelian or commutative group). A group G is commutative or
abelian if g1 g2 = g2 g1 for all g1 , g2 G.
Often instead of using to denote the group operation, we use juxtaposition. In other words, xy = x y. We often use the verb to multiply to
indicate the group operation.
Definition (trivial group). The trivial group is the group that contains
only one element, namely the identity. In other words G = {1} or = {0}
(depending on which notation is being used).
Definition (permutation). Let A be a set of n elements. Then a permutaion
is a bijective function from A to itself. Usually we use positive integers to
117
118
APPENDIX B. REVIEW OF GROUP THEORY
describe A, that is A = {1, . . . , n}. Let {a1 , . . . , am } A, then we use

(a1 a2 . . . am ) to represent the function that takes ai to ai+1 for 1 i m1
and am to a1 . Such a permutation is called an m-cycle. A 2-cycle is called
a transposition.
Exercise B.1.
1. Show that the set of all permutations on n elements forms a group with
the group operation of function composition.
2. Show that any permutation can be written as a composition of disjoint
cycles.
3. Show that any m-cycle can be written as a composition of transpositions.
Definition (order of a group). Let G be a finite group. Then the groups
cardinality |G| is called the order of G.
Definition (symmetric group). The group of all permutations on the first
n positive integers is called the symmetric group, denoted by Sn .
Question B.2. What is the order of Sn ?
Note that Sn is not an abelian group for n 3.
Definition (even and odd permutations). A permutation is even if it can
be written as the composition of an even number of transpositions and odd
otherwise.
Exercise B.3.
1. Show that an n-cycle can be written as the composition of n 1 transpositions. Thus a 3-cycle is an even permutation and a 4-cycle is an
odd permutation!
2. Show that the group of even permutations is a subgroup of Sn .
Definition (alternating group). The group of even permutations is called
the alternating group, denoted by An .
Question B.4. What is the order of An ?
Definition (dihedral group). The symmetry group of a regular n-sided polygon (under composition) is called the dihedral group.
119
Exercise B.5. Show that if we let a represent a reflection along a line
passing through the polygons center and a vertex, and b a rotation of 2/n
around its center, then the elements of Dn are the set {1, b, . . . , bn1 , ab, . . . , abn1 }
Exercise B.6. Show that in Dn as above, we have ab = bn1 a, and thus
Dn is not abelian for n > 2.
Definition (subgroup). A subgroup H of a group G is a subset of G such
that H is a group with the binary operation of G.
Exercise B.7. Show that Dn is isomorphic to a proper subgroup of Sn .
Question B.8. Under what conditions, if ever, is Dn is isomorphic to a
subgroup of An ?
Since the symmetries of a polygon induce permutations on its vertices,
it is easy to see that Dn
= H Sn , and H 6= Sn .
Definition (left coset). Let g G, a group, and H be a subgroup of G.
Then the left coset of H by g is
gH := {gh|h H}.
We can define the right coset Hg similarly.
Exercise B.9. Let g, g 0 G. Then either gH = g 0 H or gH g 0 H = .
Definition (index of a subgroup). Let H be a subgroup of G, then the index
of H in G, denoted [G : H], is the number of left cosets of H in G.
Theorem B.10 (Lagranges Theorem). Let G be a finite group, and H a
subgroup. Then the cardinality |H| of H divides the cardinality |G|of G and
[G : H] =
|G|
|H|
Definition (normal subgroup). A subgroup H of G is called a normal subgroup of G (denoted H G) if gHg 1 = H, where aHb := {ahb|h H}.
Multiplying a group or an element on the left by one element and on the
right by its inverse is called conjugation, so a normal subgroup is one which
is unchanged (set-wise) by conjugation.
Theorem B.11. Let H G be a normal subgroup. Then its left and right
cosets coincide for all g G, in other words gH = Hg for all g G.
120
Definition (direct product, direct sum). The direct product G H of two

groups G and H is the set G H with the group operation defined by (g, h)
(g 0 , h0 ) = (gg 0 , hh0 ). When the groups are additive we call this direct sum
and write G H.
Definition (homomorphism). A function f : G H is a (group) homomorphism if f (g g 0 ) = f (g) f (g 0 ) for all g, g 0 G.
In other words f preserves the group structure in the image of G.
Definition (isomorphism). A bijective homomorphism f : G H is an
isomorphism, in which case we say G is isomorphic to H and write G
= H.
But what about when the homomorphism is not bijective?
Definition (kernel). The kernel of a homomorphism f : G H is
Ker f := {g G|f (g) = 1H }
Theorem B.12. An onto homomorphism f : G H is an isomorphism if
and only if Ker f = {1G }.
Theorem B.13. Let f : G H be a homomorphism from a group G to a
group H, then Ker f G.
Definition (quotient group). A normal subgroup N G has its left cosets
equal its right cosets: gN = N g. Therefore the set G/N := {gN |g G} of
all left cosets of N is a group with the group operation
(gN ) (g 0 N ) := gg 0 N
This group is called the quotient group of G by N.
Definition (normalizer of a subgroup).
of G. Then
Let H be a subgroup

the normalizer of H in G is N (H) = g G|gHg 1 = H .
Note that N (H) is a subgroup of G, H N (H), and that it is the
largest subgroup of G in which H is normal, meaning that any subgroup of
G containing H in which H is normal must be contained in N (H).
Theorem B.14 (First isomorphism theorem). Let f : G H be an onto
(or surjective) homomorphism with Ker f = N . Then H
= G/N .
121
Definition (cyclic subgroup). Let g G. Then hgi the cyclic subgroup
generated by g is the subgroup formed by all powers of g:
hgi := {g n |n Z}
n times
n times
z
}|
{
z }| {
n
where g = g g g if n > 0, g 0 = 1, and g n = g 1 g 1 g 1 for n N.
Note that with additive notation hgi = {ng|g G, n Z}, where ng =

n times
n times
}|
{
z
}|
{
z
g + g + + g for n N, g 0 = 0, and ng = g + g + + g for n N.
Definition (cyclic group). If G = hgi for some g G we say G is a cyclic
group with generator g.
Note that cyclic groups are abelian.
Definition (finite cyclic group of order n). If G = hgi and there exists
n Z such that g n = 1, then there exists a least n N such that g n = 1. G
is said to have order n, |G| = n.
Theorem B.15. A cyclic group that is non-finite must be isomorphic to Z.
Theorem B.16. A finite cyclic group of order n is isomorphic to Zn , the
integers with addition mod n.
n times
z
}|
{
Definition (free abelian group of rank n). A group G
= Z Z . . . Z is
called the free abelian group of rank n. G has a generating set of n elements
of infinite order, one for each Z factor.
Definition (generators). Let G be a group and S G. Then the smallest
subgroup H of G containing S is called the subgroup generated by S. If
H = G then we say G is generated by S, or that S generates G.
Note that the set of generators of a group is by no means necessarily
unique. We can view the subgroup H generated by S as the set of all
possible products g1 g2 . . . gn where gi S or gi1 S. We can also view H
as the intersection of all subgroups of G that contain S.
Exercise B.17.
1. Verify that the dihedral group Dn = {1, b, . . . , bn1 , ab, . . . , abn1 } is
generated by {a, b}.
122
2. Show that the symmetric group Sn , for n 2, is generated by the set

of 2-cycles: {(12), (23), . . . , (n 1, n)}.
3. Show that the symmetric group Sn , for n 2, is generated by the pair
of cycles (12) and (12 . . . n).
Definition (finitely generated group). A group is finitely generated if there
exists a finite subset S of G that generates G.
Theorem B.18 (Classification of Finitely Generated Abelian Groups). Let
G be a finitely generated abelian group. Then G is isomorphic to:
H0 H1 . . . Hm
where H0 is a free abelian group, and Hi
= Zpi (i = 1, . . . , n) where pi is a
prime. The rank of H0 is unique, and the orders p1 , . . . , pm are unique up
to reordering.
Definition (commutator, commutator subgroup). A commutator in a group
G is an element of the form ghg 1 h1 . The commutator subgroup G0 is the
subgroup generated by the commutators of G.
Theorem B.19. G0 G, and is the smallest subgroup for which G/G0 is
abelian. In other words, if there is a subgroup N G such that G/N is
abelian, then G0 N .
There is a useful notation for groups that, roughly speaking, uses the
fact that if we know a set of generators and the rules (called relations)
to tell when two elements are the same, then the group (up to isomorphism)
is determined by a list of these generators and relations. What follows is a
very non-technical description of the generator-relation notation for groups.
For example, in Dn (as described above) it is enough to know that
there are two generators a and b, of order 2 and n respectively, and that
they satisfy ab = bn1 a. These facts determine a complete list of elements
1, b, . . . , bn1 , ab, . . . , abn1 . The expressioin ab = bn1 a can be written as
aba1 b = 1, and the word aba1 b is called a relation. By a and bs order, we
also know that a and b satisfy a2 = 1 and bn = 1. The two letters a and b,
together with the above relations completely determine the group Dn , and
thus we write:
Dn = ha, b|a2 , bn , aba1 bi
Similarly, we can write the cyclic group of order n as:
Cn = ha|an i.
123
We can write the infinite cyclic group as:
C = ha|
i.
We can denote the free abelian group of rank n as:

1
Fnab = ha1 , . . . , an |ai aj a1
i aj , i 6= j {1, . . . , n}i
Exercise B.20. Confirm that the lists of generators and relations given
above completely determine the groups.
We should note that since the relations g g 1 = 1, g 1 = g and 1 g = g
hold for any g G, as they are implicit in the definition of a group, such
relations are not included in the list of relations. In general a group G can be
written as G = hgenerators|relationsi. This is called a group presentation G.
This notation is very useful, especially when dealing with fundamental group
and Van Kampens theorem. The problem with this notation, however, is
that it is very difficult, in general, given two groups with this notation, to
tell if the groups are isomorphic or not, or even if two words represent the
same group element.
Exercise B.21. What is a group presentation for an arbitrary finitely generated abelian group? for the symmetric group?
124
Appendix C
Review of Graph Theory

Although a graph is an abstract object (a pair (V, E), where E = {v, w},
and unordered pair of elements of V , we will look at embedded graphs formed
of 1-simplices in Rn .
Definition (graph). A graph G is the union of 1-simplices {i }ki=1 in Rn
such that for i 6= j, i j is empty or an endpoint of each of i and j .
The i s are the edges of G.
Definition (walk). A walk is a finite sequence of oriented edges, with the
vertices starting at a vertex v0 and ending at a vertex vn :
[v0 v1 ], [v1 v2 ], . . . , [vi vi+1 ], . . . , [vn1 vn ].
Note that any two successive edges share a vertex. We say the walk is a
walk from v0 to vn .
Definition (trail). A trail is a path from v0 to vn where no edge is repeated,
in other words it is finite sequence of edges vertices starting at vertex v0 and
ending at vertex vn :
[v0 v1 ], [v1 v2 ], . . . , [vi vi+1 ], . . . , [vn1 vn ],
and [vi vi+1 ] 6= [vj vj+1 ] for i 6= j and 0 i, j n 1.
Definition (circuit). A circuit is a trail from v0 back to v0 (also called a
closed trail), in other words it is finite sequence of edges vertices starting
and ending at a vertex v0 :
[v0 v1 ], [v1 v2 ], . . . , [vi vi+1 ], . . . , [vn1 v0 ],
and no edge is repeated, that is: [vi vi+1 ] 6= [vj vj+1 ] for i 6= j and 1 i, j
n 1 (where the addition in the indices is done modulo n).
125
126
APPENDIX C. REVIEW OF GRAPH THEORY
Definition (path). A path is a walk from v0 to vn :

[v0 v1 ], [v1 v2 ], . . . , [vi vi+1 ], . . . , [vn1 vn ].
where vi 6= vj when i 6= j. We say the walk is a path from v0 to vn .
Definition (cycle). A cycle is a path from v0 to v0 where vi 6= vj whenever
i 6= j, for 0 i, j n, in other words, it is finite sequence of edges vertices
starting and ending at a vertex v0 :
[v0 v1 ], [v1 v2 ], . . . , [vi vi+1 ], . . . , [vn1 v0 ],
where no vertex (other than the starting vertex) is repeated, except for the
ending vertex and beginning vertex respectively of two successive edges.
Definition (tree). A tree is a connected graph with no circuits.
Definition (maximal tree). Given a connected graph G with edges {i }ki=1 ,
a subgraph T of G is a maximal tree if and only if T is a tree and for any
edge e of G not in T , T e has a circuit.
Theorem C.1. Let G be a connected graph. Then G contains a maximal
tree and every maximal tree for G contains every vertex of G.
127
Michael Starbird
Department of Mathematics
RLM 8.100
The University of Texas at Austin
Austin, TX 78712
starbird@math.utexas.edu

Grad Alg Top Mooremethod

Uploaded by

Grad Alg Top Mooremethod

Uploaded by

Algebraic Topology

3 Fundamental group and covering spaces

A Review of Point-Set Topology

B Review of Group Theory

C Review of Graph Theory

D The Jordan Curve Theorem

Definition ( Rn ). We define real or Euclidean n-space, denoted by Rn , as

1.3. SIMPLICIAL COMPLEXES

Definition (2-simplex). Let v0 , v1 , and v2 be three non-collinear points in

An i-simplex whose vertices are any subset of i + 1 of the vertices of n is an

Simplices can be assembled to create polyhedral subsets of Rn known as

So K is a filled in triangle and a hollow triangle as pictured.

by giving a triangulation of the space.

Definition (subdivision). Let (K, T ) be a finite simplicial complex. Then

There is a standard subdivision of a triangulation that later will be

2. Let T be a triangulation of a simplicial 2-complex with 2-simplices

Figure 1.1: Barycentric subdivision of a 2-simplex

Figure 1.2: Second barycentric subdivision of a 2-simplex

be particularly interested in topological spaces that share features with the

Definition (2-manifold or surface). A 2-manifold or surface is a separable,

2-manifolds as simplicial complexes

For now, we will restrict ourselves to 2-manifolds that are subspaces of Rn

= {{v0 v1 v2 }, {v0 v1 v3 }, {v0 v2 v3 }, {v1 v2 v3 },

is a triangulated 2-manifold (homeomorphic to S2 ).

Examples 5. The following are triangulable 2-manifolds:

b. T2 := S1 S1 R4 or any other space homeomorphic to the boundary

c. Double torus:(See Figure 1.4)

Figure 1.4: The double torus or surface of genus 2

Figure 1.5: The Klein Bottle

e. The projective plane, denoted RP2 , := space of all lines through 0 in

Figure 1.6: RP2

Figure 1.7: The Mobius band

There is a way of obtaining more 2-manifolds by connecting two or

2-manifolds as quotient spaces

There is another way of thinking of 2-manifolds, as the abstract spaces

The process of identifying all elements of an equivalence class to a single

2. (sphere) (See Figure 1.8)

Figure 1.8: The sphere

3. (sphere) (See Figure 1.9)

Figure 1.9: Another way to see the sphere

Figure 1.10: The double torus

5. (Klein bottle) (See Figure 1.11)

Figure 1.11: The Klein bottle

6. (projective plane) (See Figure 1.12)

Figure 1.12: The projective plane

Figure 1.13: Another version of the projective plane

The following theorem will be put off to chapter 4 (and stated in a

The most fundamental questions in topology are:

Classification of compact 2-manifolds

A surface, or 2-manifold, is locally homeomorphic to R2 , so we know how

that every compact, triangulated 2-manifold can be constructed by taking

Classification of compact, connected 2-manifolds, I

2.1. CLASSIFICATION OF 2-MANIFOLDS

Figure 2.1: A disk with four handles attached.

Figure 2.2: Twisted strips and entwined strips

2.1. CLASSIFICATION OF 2-MANIFOLDS

Figure 2.3: These spaces are homeomorphic.

Figure 2.4: Entwining pair of strips

Theorem 2.13 (Classification of compact, connected 2-manifolds). Any

Classification of compact, connected 2-manifolds, II

Figure 2.5: The genus two surface

Definition (gluing of a 2n-gon with edges identified in pairs). An expression

2.1. CLASSIFICATION OF 2-MANIFOLDS

3. The square with edges identified by cdc1 d1 is homeomorphic to T2 .

Our goal is to organize connected, compact, triangulated 2-manifolds by

One of our goals in studying topological spaces is to be able to distinguish

Definition (Euler characteristic). Let M 2 be a 2-manifold with triangulation T . Let

and define the Euler characteristic, (M 2 ), of M 2 by (M 2 ) = v e + f .

Definition (genus). The genus of S2 = 0. The genus of a 2-manifold

Theorem 2.30. Let RP2i be the projective plane for i = 1, . . . , n. Then

3. An orientation of the 1-simplex 1 is a one-to-one and onto function

(c) if M 2 is non-orientable and (M 2 ) = 2 n, for n 1, then

Notice that orientable connected, compact, triangulated 2-manifolds must

Triangulating a surface in order to calculate the Euler Characteristic can

Figure 2.6: The basic idea of CW complexes

The previous theorem amalgamated two triangles together; however, we