Sie sind auf Seite 1von 74

Lecture Notes on Homology Theory

Dr. Thomas Baird (illustrations by Nasser Heydari)


Winter 2014
Contents
1 Introduction 2
2 Review of Point-Set Topology 5
2.1 New spaces from old. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
2.2 Connectedness and Path-Connectedness . . . . . . . . . . . . . . . . . . . 8
2.3 Covers and Compactness . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.4 Metric spaces and the Lebesgue number lemma . . . . . . . . . . . . . . . 9
2.5 Hausdor spaces . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3 Singular Homology 10
3.1 Simplices . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.2 Chains, cycles, and boundaries . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.2.1 Homology as a functor . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.3 Homotopy Invariance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
3.3.1 Chain complexes and chain homotopy . . . . . . . . . . . . . . . . . 17
3.3.2 The prism operator . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
3.4 Relative Homology and the long exact homology sequence . . . . . . . . . 21
3.4.1 Reduced Homology . . . . . . . . . . . . . . . . . . . . . . . . . . . 24
3.5 Excision . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
3.5.1 Proof of Excision . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
3.5.2 Mapping cylinders and cones . . . . . . . . . . . . . . . . . . . . . . 31
3.6 Applications to spheres: the degree of a map . . . . . . . . . . . . . . . . . 33
3.7 Cellular homology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.7.1 Cell complexes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.7.2 Cell complex propaganda (not to be tested) . . . . . . . . . . . . . 40
3.7.3 Cellular Homology . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
3.7.4 Examples . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44
3.8 Mayer-Vietoris Sequence . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46
3.9 Homology with coecients . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
3.9.1 The Universal Coecient Theorem for Homology . . . . . . . . . . 52
3.10 Covering spaces and the transfer . . . . . . . . . . . . . . . . . . . . . . . . 54
1
4 Cohomology 56
4.1 The cup product . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60
4.2 The K unneth formula . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65
4.3 Manifolds and orientations . . . . . . . . . . . . . . . . . . . . . . . . . . . 68
4.4 The cap product and Poincare Duality . . . . . . . . . . . . . . . . . . . . 71
1 Introduction
Topology is the study of topological spaces ( e.g. subsets of R
n
) and continuous maps
between them. The basic idea of algebraic topology is to study functors F from
topological spaces to groups (or some other type of algebraic category). This means is
that for every topological space X, we assign a group F(X), and to each continuous map
f : X Y , we assign a group homomorphism F(f) : F(X) F(Y ) such that
F(f g) = F(f) F(g)
for any pair of composable maps
X
g
Y
f
Z
and also that identity maps are sent to identity maps:
F(Id
X
) = Id
F(X)
To see how this sort of thing may be useful, observe that if two spaces X and Y
are isomorphic (i.e. homeomorphic), then F(X) and F(Y ) must be isomorphic for every
functor F. It turns out that the most powerful way to prove that two spaces X, Y are
not homeomorphic is to nd a functor such that F(X) and F(Y ) are not isomorphic.
For another application, we begin with a denition. A subset A X is called a
retract, if there exists a continuous map r : X A such that f(a) = a for all a A.
1
.
Inclusion of sets denes an injective map i : A X. If A is a retract in X, then there
exists r such that
r i = Id
A
.
For any functor, this means that
F(r) F(i) = F(r i) = F(Id
A
) = Id
F(A)
.
In particular, this means that F(i) must be injective when A X is a retract (if not,
F(r) F(i) = Id
F(A)
would not be injective, a contradiction). Using this idea, we will
prove that the unit circle S
1
is not a retract inside the unit disk D
2
.
The kinds of functors we will learn about in this course are the (singular) homology
and cohomology functors. These functors come in families labelled by non-negative
integers called the degree (also called dimension): H
0
, H
1
, H
2
, ... for homology and
H
0
, H
1
, H
2
, .... for cohomology. Both homology and cohomology take values in abelian
groups, though we will also study variations that take values in vector spaces.
1
For example, the inclusion R R
2
as the x-axis is a retract using the map r(x, y) = (x, 0)
2
The historical motivation for homology theory came from vector calculus.
2
Recall that
there are various versions of the Fundamental Theorem of Calculus ( Stokes Theorem,
Greens Theorem, the Divergence Theorem) that equate an integral over a manifold (curve,
surface, solid, etc.) with an integral over its boundary (set of points, a curve, surface,
respectively). Homology emerged from eorts to understand how many independent
submanifolds there are with respect to a given domain. Roughly speaking, the 0- homology
H
0
(X) is generated by points in X, the 1-homology H
1
(X) is generated by (oriented)
closed curves in X, the 2-homology is generated by (oriented) closed surfaces, and so on.
The homology class is trivial if the curve, surface, etc. is the boundary of a surface, solid,
etc..
To see how this might work, consider the disconnected subset X R
2
pictured in
Figure 1.
Figure 1: A space X with two path components
A point p in one component cannot be joined by a continuous path to a point q in
another component. It follows that p and q determine dierent elements [p] and [q] in
H
0
(X) . We will show that there is an isomorphism H
0
(X)

= Z
n
where n is the number
of path-components of X.
Consider now a annulus A in R
2
(Figure 2).
The closed loop C represents an element in H
1
(A). It is intuitively clear that S
1
is
not the boundary of a surface in A, so C represents a non-trivial element [C] in H
1
(A).
Indeed, we will show that H
1
(A)

= Z and that C represents one of the generators (the
other generator is obtained by reversing the orientation on C). On the other hand, if
we take a union of C with a curve D that winds around the annulus in the opposite
direction, we see that together they form the boundary of a surface (Figure 3). In terms
of homology, this will mean that [C+D] = [C] +[D] = 0, or equivalently that [C] = [D].
2
The development of homology theory is usually attributed to Poincare in the 1890s, though the sub-
ject didnt really come into its own until the 1930s through the work of numerous other mathematicians.
3
Figure 2: Loop in an annulus
Figure 3: Two loop in an annulus bound a surface
4
There are many dierent kinds of homology that are dened in dierent ways. The
approach we will take in this course is called singular homology. Singular homology has
some great theoretical advantages over others (such as simplicial homology and cellular
homology), but has the draw back of being dicult to calculate directly. Indeed, it
will take some time before we establish the fact that oriented submanifolds determine
homology classes which has been the basis of todays lecture.
3
Instead, singular homology
is based on singular simplices which we will be introduced in ...
2 Review of Point-Set Topology
In this section we collect some basic facts from general topology that will be required in
this course. Proofs of these results can be found in any introductory textbook on general
topology (e.g. Munkres Topology) or in the point-set topology notes I have posted on
D2L.
Denition 1. A topological space (or simply space) (X, ) is a set X and a collection
of subsets of X, called the open sets, satisfying the following conditions:
i) and X are open,
ii) Any union of open sets is open,
iii) Any nite intersection of open sets is open.
A set is called closed if its complement is open. Usually, we will denote the topological
space (X, ) simply by X.
Example 1 (Euclidean Topology). An open ball in R
n
is a set of the form
B = B

(p) := x R
n
[ [[x p[[ <
for some p R
n
and > 0. A subset U R
n
is called open if it is a union of open
balls. Equivalently, U is open if for every p U, there exists an open ball B such that
p B U.
In the example above, we say that open balls form a basis for the Euclidean topology.
More generally, a collection of open sets B in a topological space X is called a basis if
every other open set in X is a union of sets in B.
Denition 2. A continuous map f : X Y between topological spaces is a map of
sets for which pre-images of open sets are open. I.e.
U Y is open f
1
(U) := x X[f(x) U X is open
Denition 3. A homeomorphism is a continuous bijection f : X Y such that the
inverse f
1
is also continuous. This is the notion of isomorphism for topological spaces.
3
It is possible to dene a kind of homology theory using oriented manifolds directly, called bordism.
One of the reasons that this approach is not standard in introductory courses is that the theory of
manifolds gets complicated in dimensions larger than two.
5
We will only rarely need use the abstract Denition 2 explicitly. More often, we will
make use of certain properties of continuous functions, including the following.
Proposition 2.1. Let X, Y and Z be topological spaces.
The identity map Id
X
: X X is continuous.
If f : X Y and g : Y Z are continuous, then the composition g f : X Z
is continuous.
Any constant map f : X Y is continuous.
The rst two conditions above make topological spaces + continuous maps into a
category. We will speak more about categories later.
2.1 New spaces from old.
Most of the topological spaces we encounter in this course are constructed from R
n
using
the operations below.
Denition 4. Let X be a topological space and A X a subset. The subspace topol-
ogy on A is the topology for which V A is open if and only if V = A U for some
open set U in X.
Example 2. Any subset of R
n
acquires a subspace Euclidean topology. Unless otherwise
stated, we will always assume subsets of R
n
to have this topology.
The inclusion map i : A X is continuous (with respect to the subspace topology).
In fact, we have the following special property: A map f : Y A from a topological
space Y is continuous if and only if the composition i f : Y X is continuous.
Denition 5. The product space XY of two spaces X and Y is the Cartesian product
of sets X Y , with a basis of open sets of the form U V where U X and V Y are
both open.
The above denition iterates to dene products of any nite number of spaces (innite
products require a dierent denition).
Example 3. The n-fold product RR...R is homeomorphic to R
n
with the Euclidean
topology.
The key property of product spaces is that a map
F : Z X Y
is continuous if and only if the coordinate functions F = (F
1
, F
2
) are continuous as maps
from Z to X and to Y respectively.
Denition 6. Let X

be a (possibly innite) collection of spaces indexed by . The


coproduct space or disconnected union

is the disjoint union of the sets X

with U

is open if and only if U X

is open for all .


6
The inclusions i

0
: X

0


X

are all continuous. A map F :

Y is
continuous if and only if the composition F i

: X

Y are continuous for all .


Denition 7. An equivalence relation on a set X is a relation satisfying, for all
x, y X
(i) x x
(ii) x y implies y x
(iii) x y and y z implies x z.
Given any relation R on X, we can generate the smallest equivalence relation
R
such that xRy implies x
R
y. Explicitly, we dene x
R
y if and only if there exists a
nite sequence x
i
X
n
i=0
for n 0 satisfying
x
0
= x,
x
n
= y and,
x
i
Rx
i1
or x
i1
Rx
i
for all i = 1, ..., n.
Given x X, the equivalence class of x is
[x] := y X[x y
Notice that [x] = [y] if and only if x y. The equivalence classes determine a partition
of X into disjoint sets. Let E := [x][x X be the set of equivalence classes (we will
sometimes denote E = X/ ) . There is a canonical map
Q : X E, x [x]
called the quotient map.
Denition 8. Let X be a topological space and let be an equivalence relation on the
set underlying X. The quotient topology on E is the topology for which U E is
open if and only if Q
1
(U) is open in X.
Observe that Q : X E is continuous and that a map f : E Y is continuous if
and only if f Q : X Y is continuous.
Example 4. Suppose X and Y are topological spaces, A X is a subspace, and f : A
Y is a continuous map. Dene an equivalence relation on the coproduct X

Y generated
by f(a) a for all a A. We say that the quotient space (X

Y )/ is obtained by
attaching X to Y along A using f.
Example 5. Suppose X is a topological space, and G is a group that acts on X via
homeomorphisms. Dene two points in X equivalent if they lie in the same orbit of G.
The quotient space in this case is called the orbit space and is denoted X/G.
7
2.2 Connectedness and Path-Connectedness
Let I denote the unit interval [0, 1] R with the Euclidean topology.
Denition 9. A space X is called path-connected if for any two points p, q X there
exists a continuous map : I X such that (0) = p and (1) = q.
Denition 10. A space X is called connected if there is no proper subset A X
which is both open and closed. (proper means other than X or , which are always both
open and closed).
Observe that if A X is both open and closed, then the complement A
c
is also both
open and closed, and there is a natural isomorphism A

A
c

= X. Thus spaces that are
not connected can be decomposed into a disconnected union of nonempty spaces.
Proposition 2.2. Path-connected spaces are connected.
The converse of Proposition 2.2 is not true in general. However all the connected
spaces we encounter in this course will also be path-connected.
Connectedness and path-connectedness are preserved under the following operations
A product of (path-)connected spaces is (path-)connected.
The continuous image of a (path-)connected space is (path-)connected.
Let U

be a covering of X such that each U

is (path-)connected and the inter-


section

is non-empty. Then X is (path-)connected.


2.3 Covers and Compactness
Denition 11. An open (closed) cover of a topological space X is a collection of open
(closed) sets U

such that the union

= X.
Proposition 2.3. Let U

be either an open cover or a nite closed cover of X. A map


of sets
f : X Y
between topological spaces is continuous if and only if the restrictions f[
U
: U

Y are
continuous for all (where U

has the subspace topology).


The preceding proposition will be used in two ways: to test if a map f is continuous by
considering the restrictions, and also to construct a map f by gluing together continuous
maps dened on the U

that agree on overlaps.


Denition 12. A space X is called compact if every open cover U

of X contains
a nite subcover. I.e., there exists a nite collection U
1
, ..., U
n
U

such that

n
i=1
U
i
= X.
Proposition 2.4. A subspace of R
n
is compact if and only if it is closed and bounded.
8
Compactness is preserved under the following:
A closed subspace of a compact space is compact.
A nite union of compact spaces is compact.
A product of compact spaces is compact.
If f : X Y is continuous and X is compact, then the image f(X) Y is compact.
2.4 Metric spaces and the Lebesgue number lemma
Denition 13. Let X be a set. A metric on X is a function
d : X X R
0
called the distance or metric function, satisfying
1. d(x, x

) = 0 x = x

(d separates points)
2. d(x, x

) = d(x

, x) (d is symmetric)
3. d(x, x

) d(x, x

) +d(x

, x

) (the triangle inequality)


A metric space (X, d) determines a metric topology on X, which is generated by
the basis of open balls B

(p) = x X[d(x, p) < . If A is a subset of a metric space X


then A becomes a metric space by restriction. The metric topology on A is the same as
the subspace topology on A.
The following result will come up repeatedly.
Lemma 2.5 (Lebesgue number Lemma). Let / be an open covering of a compact metric
space X. There exists > 0, called the Lebesgue number, such that for all p X, the
open ball B

(p) is contained in some U /.


2.5 Hausdor spaces
Denition 14. A space X is called Hausdor if for any pair of distinct points p, q X,
there exist open sets U, V such that p U, q V and U V = .
Proposition 2.6. Any metric space is Hausdor. In particular, any subset of R
n
is
Hausdor.
The Hausdor property is preserved under the following:
Products of Hausdor spaces are Hausdor.
Subspaces of Hausdor spaces are Hausdor.
Coproducts of Hausdor spaces are Hausdor.
9
Denition 15. A space X is locally compact, if every point p X is contained in an
open neighbourhood p U such that the closure U is compact.
For example, compact spaces are locally compact. Also, R
n
is locally compact.
Proposition 2.7. Let X be a locally compact space and let Y be a Hausdor space. If
: X Y is a continuous map that is also a bijection of sets, then is a homeomorphism.
3 Singular Homology
3.1 Simplices
The standard q-simplex
q
is the simplex spanned by the zero vector e
0
=

0 and the
standard basis vectors e
1
, ..., e
q
in R
q
(Figure 4). Thus,

q
:= (t
1
, ..., t
n
)[ 0 for all i = 1, ..., q and t
1
+... +t
q
1
Figure 4: The standard simplices
If X is a topological space, a singular q-simplex (or simply simplex)in X is a
(continuous) map
:
q
X.
A singular 0-simplex in X is simply a point in X, a singular 1-simplex in X is a
continuous path in X, etc.. We can think of singular simplices as probes used to study
the space X.
Example 6. Let v
0
, ..., v
q
be a set of q + 1-vectors in R
n
for some n. Dene
[v
0
, ..., v
q
] :
q
R
n
, (t
1
, ..., t
n
) (1 t
1
... t
q
)v
0
+t
1
v
1
+... +t
q
v
q
.
10
We call [v
0
, ..., v
q
] the ane simplex dened by v
0
, ..., v
q
. Slightly abusing notation,
dene the face maps for 0 i q, dene by
F
i
q
:
q1

q
by F
i
q
= [e
0
, ..., e
i
, ...., e
n
] where the e
i
means omit e
i
.
Figure 5: Faces of the standard 2-simplex
The ith face of a singular q-simplex :
q
X is the q 1-simplex

(i)
:
q1
X
dened be composition with the face map:

(i)
:= F
i
q
.
3.2 Chains, cycles, and boundaries
Dene S
q
(X) to be the free Abelian group generated by singular q-simplices. The elements
of S
q
(X) are called singular chains and are formal linear combinations of the form

where the coecients a

Z and the sum is over a nite number of singular q-simplices


. By convention, S
q
(X) = 0 for q < 0.
The boundary map
q
: S
q
(X) S
q1
is a homomorphism, dened on singular sim-
plices by

q
() =
q

i=0
(1)
i

(i)
and extended linearly to all of S
q
(X) by the rule

q
(

) =

q
().
We will often drop the subscript and write =
q
when it is unlikely to cause confusion.
11
Example 7. Let
1
and
2
be singular 2-simplices in X. Then 2
1
+ 3
2
S
2
(X) is a
2-chain and

2
(2
1
+ 3
2
) = 2
2
(
1
) + 3
2
(
2
)
= 2(
(0)
1

(1)
1
+
(2)
1
) + 3(
(0)
2

(1)
2
+
(2)
2
)
= 2
(0)
1
+ 2
(1)
1
2
(2)
1
+ 3
(0)
2
3
(1)
2
+ 3
(2)
2
is a 1-chain in S
1
(X).
The boundary map can be understood schematically from Figure 6, but be careful not
to confuse singular simplices (which are maps) with their images (which are sets).
Figure 6: Boundary of simplices - intuition
Proposition 3.1. The composition
q1

q
: S
q
(X) S
q2
(X) is the zero map. Drop-
ping subscripts, we write this

2
= 0.
Proof. Since S
q
(X) is generated by simplices, it suces to check that
q1

q
() = 0 for
all q-simplices .
It is an easy check that if 0 j < i q, the face maps satisfy
F
i
q
F
j
q1
= F
j
q
F
i1
q1
.
Thus
12

q1

q
() =
q1
(
q

i=0
(1)
i

(i)
)
=
q

i=0
(1)
i

q1
( F
i
q
)
=
q

i=0
(1)
i
q1

j=0
(1)
j
( F
i
q
F
j
q1
)
=
q

i=0
q1

j=0
(1)
i+j
( F
i
q
F
j
q1
)
=

0ijq1
(1)
i+j
( F
i
q
F
j
q1
) +

0j<iq
(1)
i+j
( F
i
q
F
j
q1
)
=

0ijq1
(1)
i+j
( F
i
q
F
j
q1
)

0ji1q1
(1)
i1+j
( F
j
q
F
i1
q1
)
= 0
as these two sums cancel term by term by changing the index of the rst sum by i j,
j i 1.
A more geometric illustration of = 0 is provided in Figure 7.
Figure 7:
2
= 0
The group of q-cycles Z
q
(X) is the kernel of
q
:
Z
q
(X) := S
q
(X) [ () = 0
The group of q-boundaries B
q
(X) is the image of
q+1
:
B
q
(X) := () [ S
q+1
(X).
13
by Proposition 3.1, B
q
(X) is a subgroup of Z
q
(X). The qth degree singular homology of
X is the quotient group:
H
q
(X) := Z
q
(X)/B
q
(X).
Example 8. The homology of a point. If X = pt is a single point, then there is
only one singular simplex in each degree, which is the constant map
q
:
q
pt. The
chain groups are
C
q
(pt) = Z
q

= Z.
4
The boundary map satises (for q 1)

q
(
q
) =
q

i=0
(1)
i

(i)
q
=
q

i=0
(1)
i

q1
=
_

q1
if q is even
0 if q is odd
while
0
(
0
) = 0.
It follows that for q 1
Z
q
(pt) = B
q
(pt) =
_
0 if q is even
C
q
(pt) if q is odd
while Z
0
(pt)

= Z and B
0
(pt) = 0. Thus
H
q
(pt) = Z
q
(pt)/B
q
(pt) =
_
0 if q 1
Z if q = 0
A space X for which H
q
(X)

= H
q
(pt) for all q is called acyclic meaning no cycles
that are not also boundaries.
Proposition 3.2. Let X
k
be the set of path components of a space X (indexed by k).
Then
H
q
(X) =
k
H
q
(X
k
)
for all q 0.
Proof. Because the standard q-simplex is path connected (indeed convex), the image of
a singular q-simplex :
q
X must be path connected and in particular must lie
within one of the path components of X. It follows that for all q we have a canonical
decomposition,
C
q
(X) =
q
C
q
(X
k
).
4
The case of a point is highly unusual in this respect. For most spaces Y , C
q
(Y ) has uncountable
rank.
14
Moreover, it is clear that the boundary map respects this decomposition, so that
Z
q
(X) =
k
Z
q
(X
k
)
and
B
q
(X) =
k
B
q
(X
k
)
and nally that
H
q
(X) = Z
q
(X)/B
q
(X)
= (
k
Z
q
(X
k
))/(
k
B
q
(X
k
))
=
k
(Z
q
(X
k
)/B
q
(X
k
))
=
k
H
q
(X
k
)
We denote by
0
(X) the set of path components of of a space X.
Proposition 3.3. There is a canonical isomorphism
H
0
(X) = Z
0
(X).
Thus H
0
(X)

= Z
n
where n is the number of path components of X.
Proof. By Proposition 3.2, it suces to show that if X is path connected, then there is a
canonical isomorphism
H
0
(X) = Z.
Recall that a singular 0-simplex is the same thing as a point in X. Thus
S
0
(X) = Z
0
(X) =

pX
Zp.
The standard one simplex
1
is equal to the unit interval [0, 1] R, so a singular 1-
simplex is a continuous path in : [0, 1] X. Since X is path-connected, for any two
points p, q X, there exists a path such that (0) = p and (1) = q. Consequently,
the boundary satises
() = (1) (0) = p q S
0
(X),
and thus
B
0
(X) = Span
Z
p q [ p, q X

pZ
Zp.
Observe that B
0
(X) is equal to the kernel of the homomorphism
:

pZ
Zp Z
dened on generators by (p) = 1. It follows that descends to a homomorphism
H
0
(X) = Z
0
(X)/B
0
(X)

= Z.
15
3.2.1 Homology as a functor
Suppose that f : X Y is a continuous map. If is a q simplex for X, then the
composition f is a q-simplex for Y . This denes a homomomorphism
S
q
(f) : S
q
(X) S
q
(Y ), S
q
(f)(

) =

f .
Clearly S
q
(Id
X
) = Id
Sq(X)
and S
q
(f g) = S
q
(f) S
q
(g) for composable continuous maps
f and g. Thus S
q
is functor from topological spaces to abelian groups. It allows commutes
with the boundary map.
Lemma 3.4.
q
S
q
(f) = S
q1
(f)
q
.
Proof. It is enough to check for simplices.
S
q
(f)() = (f )
=
q

i=0
(1)
i
(f )
(i)
=
q

i=0
(1)
i
f F
i
q
=
q

i=0
(1)
i
f
(i)
= S
q
(f)(
q

i=0
(1)
i

(i)
)
= S
q
(f)()
using associativity of composition.
It follows then that S
q
(f) sends Z
q
(X) to Z
q
(Y ) and B
q
(X) to B
q
(Y ) and thus induces
a homomorphism between the quotient groups
H
q
(f) : H
q
(X) H
q
(Y ).
It follows easily from the fact that S
q
is a functor that H
q
is a functor from topological
spaces to abelian groups. It is common to use short hand
f

= H
q
(f)
though we will try to avoid doing so.
16
3.3 Homotopy Invariance
Two continuous maps f, g : X Y are said to be homotopic if there exists a continuous
map
h : X I Y
where I = [0, 1] is the unit inverval and both h(x, 0) = f(x) and h(x, 1) = g(x). Intu-
itively, two maps are homotopic if one can be continuously deformed into the other.
The goal of this section is to prove the following theorem.
Theorem 3.5. Let f and g be homotopic maps from X to Y . For all q 0, the induced
maps on homology are equal: H
q
(f) = H
q
(g).
Two spaces X and Y are called homotopy equivalent if there exist maps f : X Y
and g : Y X such that f g is homotopic to Id
Y
and g f is homotopic Id
X
.
Corollary 3.6. If X and Y are homotopy equivalent, then H

q
(X)

= H

q
(Y ) for all q 0.
Proof. By Theorem 3.5 and functoriality, we have
H
q
(f) H
q
(g) = H
q
(f g) = H
q
(Id
Y
) = Id
Hq(Y )
and similarly H
q
(g) H
q
(f) = Id
Hq(X)
. The H
q
(f) and H
q
(g) are inverse isomorphisms
between H
q
(X) and H
q
(Y ).
A space is called contractible if it is homotopy equivalent to a point. Examples of
contractible spaces include all convex subspaces of R
n
(exercise). By Corollary 3.6, a
contractible space X satises H
q
(X) = 0 for q 1 and H
0
(X) = Z (i.e. contractible
spaces are acyclic).
Before proving Theorem 3.5, it will be helpful to introduce some abstract ideas about
chain complexes.
3.3.1 Chain complexes and chain homotopy
A chain complex (of abelian groups)
C := (C
q
,
q
)
qZ
is a sequence of abelian groups (C
q
)
qZ
and homomorphisms
q
: C
q
C
q1
such that

q+1
= 0 for all q Z.


q+3
C
q+2

q+2
C
q+1

q+1
C
q
q
C
q1

q1
C
q2

q2
. . .
Typically C
q
= 0 for q < 0.
Example 9. The singular chain complex S(X) = (S
q
(X),
q
)
qZ
is a chain complex.
17
We dene Z
q
(C) = ker(
q
) and B
q
(C) = im(
q+1
) and H
q
(C) = Z
q
(C)/B
q
(C), called
respectively the q-chains, q-boundaries, and q-homology groups of the chain complex. If
z Z
q
(C), denote by [z] H
q
(C) the coset represented by z.
A morphism of chain complexes f : C C

is a sequence of homomorphisms
(f
q
: C
q
C

q
)
qZ
that commutes with boundary maps: f
q1

q
=

q
f
q
for all q. In other
words, the following diagram commutes:
C
q
q

fq

C
q1
f
q1

q1
.
Example 10. Given a continuous map : X Y , the morphisms S
q
() : S
q
(X)
S
q
(Y ) determine a chain morphism S() : S(X) S(Y ).
A chain map f : C C

induces a homomorphism in homology H


q
(f) : H
q
(C)
H
q
(C

) for all q Z by the same reasoning as in 3.2.1 by the rule


H
q
(f)([z]) = [f
q
(z)].
Each H
q
is a functor from chain complexes to abelian groups.
Let f, g : C C

be two chain maps. A chain homotopy between f and g is a


sequence of homomorphism (P
q
: C
q
C

q+1
)
qZ
such that

q+1
P
q
+P
q1

q
= f
q
g
q
.
C
q+1

q+1

C
q
q

fqgq

Pq
.}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
C
q1

P
q1
.}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
}
C

q+1

q+1

q1
.
The chain maps f and g are called chain homotopic if there exists a chain homotopy
between them.
Proposition 3.7. If chain maps f, g : C C

are chain homotopic, then the induced


maps on homology are equal: H
q
(f) = H
q
(g) as maps from H
q
(C) to H
q
(C

).
Proof. Let [z] H
q
(C) be represented by z Z
q
(C). Then
f
q
(z) g
q
(z) = P
q1

q
(z) +
q+1
P
q
(z) =
q+1
P
q
(z)
is a boundary. Thus
H
q
(f)([z]) H
q
(g)([z]) = [f
q
(z)] [g
q
(z)] = [f
q
(z) g
q
(z)] = 0
so H
q
(f)([z]) = H
q
(g)([z]).
18
3.3.2 The prism operator
For t I = [0, 1], dene
i
t
: X X I, i
t
(x) = (x, t).
Lemma 3.8. The two maps i
0
, i
1
: X X I, determine chain homotopic chain mor-
phisms S(i
0
) and S(i
1
).
Proof. The rst step is to dene a decomposition of
q
I R
n+1
into (q +1)-simplices.
Denote the vertices lying in
q
0 by v
0
, ..., v
q
and those lying in
q
1 by w
0
, ..., w
q
.
For each i, the image of the ane simplex
[v
0
, ..., v
i
, w
i+1
, ...., w
n
] :
q

q
I
can be thought of as the graph of a map from
q
to I, because composing the projection

q
I
q
is the identity map. These graphs slice
q
I into the images of ane
(q + 1)-simplices
[v
0
, ..., v
i
, w
i
, ..., w
q
] :
q+1

q
I, i 0, ..., q.
Figure 8: Prism decomposition
For arbitrary X, dene the prism operator P
q
: S
q
(X) S
q+1
(X I) on simplices
by
P
q
() :=
q

i=0
(1)
i
( Id
I
) [v
0
, ..., v
i
, w
i
, ..., w
q
].
19
This will be our chain homotopy.

q+1
P
q
() =
q+1
q

i=0
(1)
i
( Id
I
) [v
0
, ..., v
i
, w
i
, ..., w
q
]
=
q

i=0

ji
(1)
i+j
( Id
I
) [v
0
, .., v
j
, ..., v
i
, w
i
, ..., w
q
]
+
q

i=0

ji
(1)
i+j+1
( Id
I
) [v
0
, ..., v
i
, w
i
, ..., w
j
..., w
q
]
On the other hand,
P
q1

q
() = P
q1
(
q

j=0
(1)
j
[e
0
, ..., e
j
, ..., e
q
])
=
q

i=0

j<i
(1)
i+j+1
( Id
I
) [v
0
, ..., v
j
, ..., v
i
, w
i
, ..., w
q
])
+
q

i=0

j>i
(1)
i+j
( Id
I
) [v
0
, .., v
i
, w
i
, ..., w
j
, ..., w
q
].
Adding together, we get

q+1
P
q
() +P
q1

q
() =
q

i=0
( Id
I
) [v
0
, ..., v
i
, w
i
, ..., w
q
]

i=0
( Id
I
) [v
0
, ..., v
i
, w
i
, ..., w
q
]
= Id
I
[w
0
, ..., w
q
] Id
I
[v
0
, ..., v
q
]
= i
1
i
0
.
It follows that

q+1
P
q
+P
q1

q
= S
q
(i
1
) S
q
(i
0
).
Proof of Theorem 3.5. A homotopy between two maps f, g : X Y is a map h : XI
Y such that f = hi
0
and g = hi
1
. By Lemma 3.8, S
q
(i
0
) and S
q
(i
1
) are chain homotopic,
so Proposition 3.7 implies H
q
(i
0
) = H
q
(i
1
). Finally we see that
H
q
(f) = H
q
(h) H
q
(i
0
) = H
q
(h) H
q
(i
0
) = H
q
(g).
In fact, it is not hard to show that S(f) and S(g) are chain homotopic via the chain
homotopy S
q+1
(H) P
q
.
20
3.4 Relative Homology and the long exact homology sequence
A topological pair (X, A) consists of a space X and a subspace A X. A pair (X, A)
gives rise to an inclusion of chain groups S
q
(A) S
q
(X) (technically the inclusion map
i : A X determines injective homomorphism S
q
(i)). Dene the relative chain group
of the pair to be the quotient group
S
q
(X, A) := S
q
(X)/S
q
(A).
The relative chain groups combine to form the relative chain complex
. . . S
q+1
(X, A)

q+1
S
q
(X, A)
q
S
q1
(X, A)

q1
. . .
where the boundary map is dened by the following commutative diagram
S
q
(X)

S
q1
(X)

S
q
(X, A)
q

S
q1
(X, A)
where the vertical arrows are quotient maps. Note that
q
is well dened because
q
sends
S
q
(A) to S
q1
(A) and that
2
= 0 because
2
= 0. It follows that we can dene relative
cycles, relative boundaries, and relative homology as described in 3.3.1, which are denoted
Z
q
(X, A), B
q
(X, A), H
q
(X, A)
respectively. Geometrically, a relative cycle in Z
q
(X, A) is represented by a chain in Z
q
(X)
whose boundary lands in S
q1
(A).
Remark 1. Observe that if A = , then S
q
(A) = 0 for all q. It follows that S
q
(X, ) =
S
q
(X) and that H
q
(X, ) = H
q
(X). Thus it is possible to think of homology as just a
special case of relative homology.
A map of topological pairs
f : (X, A) (X

, A

)
is a continuous map f : X X

such that f(A) A

. Such a map determines a


morphism of chain complexes S(f) : S(X, A) S(X

, A

) and thus also a homomorphism


on homology
H
q
(f) : H
q
(X, A) H
q
(X

, A

).
The following properties are proven similarly to their counterparts for H
q
(X).
H
q
is functor from topological pairs to Abelian groups. I.e. H
q
(f)H
q
(g) = H
q
(f g)
and H
q
(Id
(X,A)
) = Id
Hq(X,A)
.
If X
k
is the set of path components of X and A
k
= A X
k
, then there is a
canonical isomorphism H
q
(X, A) =

k
H
q
(X
k
, A
k
).
21
Let h : XI X

be a homotopy between maps map such that h


t
(a) := h(a, t) A

for all a A and t I. Then H


q
(h
0
) = H
q
(h
1
) as homomorphisms from H
q
(X, A)
to H
q
(X

, A

).
The quotient morphisms S
q
(X) S
q
(X, A) t together into a morphism of chain
complexes j : S(X) S(X, A). Combined with inclusion chain morphism i : S(A)
S(X) we get a commutative diagram
. . .
S
q+1
(A)

i

S
q
(A)

i

S
q1
(A)

i

. . .
. . .
S
q+1
(X)

j

S
q
(X)

j

S
q1
(X)

j

. . .
. . .
S
q+1
(X, A)

S
q
(X, A)

S
q1
(X, A)
. . .
. (1)
By functoriality, these chain morphisms give rise to homology homomorphisms H
q
(A)
H
q
(X) H
q
(X, A) for all q 0. The most important property of relative homology is
the existence of a connecting homomorphism
H
q
(X, A)

H
q1
(A),
(denoted by using abuse of notation). Let H
q
(X, A), be represented by a cycle
j(z) S
q
(X, A) which is the image of a chain z S
q
(X). Then (z) must land in
S
q
(A) S
q
(X), because

(j(z)) = j((z)) = 0 S
q
(X, A) = S
q
(X)/S
q
(A). Dene
([z]) = [(z)] H
q1
(A).
(We leave it as an exercise prove that this homomorphism is well dened and does not
depend on the representatives z or z.). The connecting homomorphism permits us to
extend to a long sequence of homomorphisms
H
q+1
(X, A)

H
q
(A)
Hq(i)
H
q
(X)
Hq(j)
H
q
(X, A)

H
q1
(A) . . . (2)
By convention, this sequence ends at H
0
(X) H
0
(X, A) 0.
Denition 16. A sequence of abelian groups and homomorphisms
A
f
B
g
C
is called exact at B if ker(g) = im(f).
Theorem 3.9. The sequence (3) is exact (meaning exact at all groups in the sequence).
It is called the long exact homology sequence associated to the pair (X, A).
22
Proof. We will only prove exactness at H
q
(X, A) and leave the rest as an exercise. Con-
sider
H
q
(X)
Hq(j)
H
q
(X, A)

H
q1
(A)
If im(H
q
(j)), that means we can choose a representative cycle z C
q
(X) such
that [j(z)] = . Since z is a cycle, it follows that (z) = 0, so
() = [(z)] = [0] = 0 H
q1
(A).
Therefore im(H
q
(j)) ker().
Conversely, suppose that ker(). This means that we can choose a chain z S
q
(X)
such that [j(z)] = and (z) = () for some S
q
(A). Consequently, z is a cycle
in S
q
(X) and
H
q
(j)([z ]) = [j(z )] = [j(z)] [j()] = [j(z)] = .
So ker() im(i

).
Proofs like the one above are called a diagram chase, because they amount to chasing
elements through a diagram like (1).
Denote by
0
(X, A) the set of path components of X that do not intersect A.
Proposition 3.10. There is a canonical isomorphism
H
0
(X, A)

= Z
0
(X, A).
In particular, H
0
(X, A)

= Z
m
where m is the number of path components of X that do
not intersect A.
Proof. Let i : A X denote the inclusion of A into X. Then we have an exact sequence
H
0
(A)
H
0
(i)
H
0
(X)
H
0
(j)
H
0
(X, A) 0.
Exactness implies that H
0
(X, A)

= H
0
(X)/im(H
0
(i)). We know (Prop 3.3) that H
0
(X) =
Z
0
(X) and the image of H
0
(i) is generated by those path components of X that contain
a path component of A. The result follows.
One of the great merits of the long exact homology sequence is that it is functorial
with respect to maps of pairs.
Proposition 3.11. Let f : (X, A) (Y, B) be a map of pairs. This induces homomor-
phisms on homology such that the following diagram commutes.
. . .
H
q+1
(X, A)
f

H
q
(A)

f

H
q
(X)
f

H
q
(X, A)

H
q1
(A)

f

. . .
. . .
H
q+1
(Y, B)

H
q
(B)

H
q
(Y )

H
q
(Y, B)

H
q1
(B)
. . .
Proof. We leave verication up to the reader.
23
The following lemma is helpful for calculations:
Lemma 3.12 (The Five Lemma). Consider a commutative diagram of abelian groups
A
1

A
2

A
3

A
4

A
5

B
1

B
2

B
3

B
4

B
5
where the rows are exact. If , , and are all isomorphisms, then is also an isomor-
phism.
Proof. Diagram chase.
Corollary 3.13. If f : (X, A) (Y, B) is a map of pairs such that two out of three
families of induced maps on homology
H
q
(A) H
q
(B)
qZ
,
H
q
(X) H
q
(Y )
qZ
.
H
q
(X, A) H
q
(Y, B)
qZ
,
are isomorphisms in all degree, then the remaining family is isomorphisms in all degrees.
Proof. Simply apply the Five Lemma to the diagram:
. . .
H
q+1
(X, A)
f

H
q
(A)

f

H
q
(X)
f

H
q
(X, A)

H
q1
(A)

f

. . .
. . .
H
q+1
(Y, B)

H
q
(B)

H
q
(Y )

H
q
(Y, B)

H
q1
(B)
. . .
Example 11. If a map of pairs f : (X, A) (Y, B) restricts to homotopy equivalences
between X and Y and between A and B, then f

: H
q
(X, A) H
q
(Y, B) is an isomor-
phism in all degrees.
3.4.1 Reduced Homology
It is sometimes convenient to use a modied version of singular homology called reduced homology.
For any space X, there exists a unique map to a point : X pt. Dene the reduced
homology

H
q
(X) := ker H
q
().
It is an easy consequence of (8) that if X has n path components,

H
q
(X)

=
_
H
q
(X) if q 1
Z
n1
if q = 0
24
More canonically,

H
0
(X) is the kernel of the map Z
0
(X) Z that sends each generator
to 1. For relative homology we dene

H
q
(X, A) = H
q
(X, A)
if A ,= is non-empty. Basically, reduced homology is designed so that H
q
(pt) = 0 for
all degrees without exception and this sometimes makes calculations less clumsy.
Functoriality, homotopy invariance, and the long exact sequence all work for relative
homology. In particular, if A is non-empty, then we have a long exact sequence


H
q+1
(X, A)


H
q
(A)

H
q
(X)

H
q
(X, A)


H
q1
(A) . . . (3)
Remark 2. If X is a path-connected space and p X, then the long exact sequence
denes a natural isomorphism
H
q
(X, p) =

H
q
(X, p)

=

H
q
(X)
in all degrees.
3.5 Excision
The last property we need before we can do calculations is called excision. Given an
ordered pair (X, A) we say a subspace of B A can be excised if the inclusion (X
B, A B) (X, A) induces isomorphisms
H
q
(X B, A B)

= H
q
(X, A)
in all degrees q.
Theorem 3.14. If the closure of B is contained in the interior of A, then B can be
excised.
Corollary 3.15. If V B A and
1. V can be excised, and
2. the inclusion (X B, AB) (X V, AV ) determine homotopy equivalences
X B X V and A B A V ,
then B can be excised.
Proof. We want to prove that H
q
(X B, A B) H
q
(X, A) is an isomorphism. By
functoriality, it is enough to show that the homomorphisms H
q
(XB, AB) H
q
(X
V, AV ) and H
q
(XV, AV ) H
q
(X, A) are isomorphisms. The rst is an isomorphism
by homotopy invariance (Example 11) and the second is an isomorphism because V can
be excised.
We postpone the proof so that we can (nally!) do some actual calculations.
25
Proposition 3.16. The homology groups of the unit sphere S
n
for n 1 satisfy
H
q
(S
n
) =
_
Z if q = 0 or q = n
0 otherwise
Proof. It will be more convenient to work with reduced homology, so our goal is to prove
that

H
q
(S
n
) =
_
Z if q = n
0 otherwise
.
Let E
n
+
and E
n

denote the upper and lower closed hemispheres of S


n
. Note that for n 1,
E
n
+
E
n

= S
n1
. I claim that by an excision that

H
q
(S
n
, E
n

)

=

H
q
(E
n
+
, S
n1
), q Z, n 1.
Here I am excising the interior of the lower hemisphere E
n

. This does not satisfy the


hypotheses of Theorem 3.14, but a slightly smaller open disk does and then I can apply
Corollary 3.15.
Now consider the long exact sequences (LES) associated to these pairs. Because E
n
+
is contractible, the LES of the pair (E
n
+
, S
n1
) breaks into isomorphisms
0

H
q
(E
n
+
, S
n1
)


H
q1
(S
n1
) 0
for all n 1 and all q Z. Likewise, the LES of (S
n
, E
n

) gives rise to isomorphisms


0

H
q
(S
n
)


H
q
(S
n
, E
n

) 0
for all n 1 and all q Z. Combined, we obtain isomorphisms

H
q
(S
n
)

=

H
q1
(S
n1
)
for all n 1 and all q Z.
Since S
0
is a disconnected union of two points, it follows that
H
q
(S
0
) =
_
Z if q = 0
0 otherwise
.
The result now follows by induction.
Proposition 3.16 hints at the special role that spheres play in homology. Later, we
will consider a special class of spaces built out of spheres called cell-complexes that are
particularly well suited to algebraic topology.
Theorem 3.17 (Brouwer Fixed Point Theorem). Let f : D
n
D
n
be a continuous map
from the closed n-disk D
n
to itself. There exists p D
n
such that f(p) = p.
Proof. For the sake of contradiction, suppose that no such p exists. Then f(x) ,= x for
all x D
n
and we can dene a continuous map r : D
n
S
n1
as illustrated in Figure 9
Notice that for points x S
n1
, r(x) = x. This implies that r is retract. In par-
ticular, this means r : H
n
(D
n
) H
n
(S
n1
) is surjective which contradicts the fact that
H
n1
(S
n1
)

= Z and H
n1
(D
n
) = 0.
26
Figure 9: Brouwer retraction
3.5.1 Proof of Excision
We begin with some notation. Let 1 := V
i
be an open cover of a space X. Dene
the chain group S
V
q
(X) to be the subgroup of S
q
(X) generated by simplices whose image
lands entirely within some V
i
1. The proof of excision boils down to showing that every
chain in S
q
(X) can be replaced - up to a boundary - by one in S
V
q
(X), where the relevant
open cover is X B, int(A).
5
The idea is to subdivide simplices into unions of simplices through a process called
barycentric subdivision. Iterating the process results in smaller and smaller simplices
until each one is contained in some set in the cover 1 via a Lebesgue number argument.
First we dene barycentric subdivision for ane simplices. If [v
0
, ..., v
q
] is an ane
simplex in R
n
, the barycentre is the point
b =
1
q + 1
v
0
+... +
1
q + 1
v
q
.
The (geometric) barycentric subdivision of [v
0
, ..., v
q
] is dened inductively to be the union
of simplices [b, w
0
, ..., w
q1
] where [w
0
, ..., w
q1
] arises in the barycentric subdivision of a
face [v
0
, ..., v
i
, ..., v
q
].
Recall that the diameter of a set A R
n
equals sup[x y[ [ x, y A.
Lemma 3.18. The diameter of (the image of ) any simplex arising in the barycentric
subdivision of [v
0
, ..., v
q
] is no greater than
q
q+1
times the diameter of [v
0
, ..., v
q
].
Proof. The distance between any two points v and

t
i
v
i
in (the image of) the simplex
[v
0
, ..., v
q
] satises
[v

t
i
v
i
[ = [

t
i
(v v
i
)[

t
i
[v v
i
[ max[v
i
v[.
5
In fact, since the boundary operator sends S
V
q
(X) to S
V
q1
(X), it actually forms a sub-chain complex
S
V
(X). Hatcher proves that the inclusion S
V
(X) S(X) is a chain homotopy equivalence and thus
induces an isomorphism on homology.
27
Figure 10: Barycentric subdivision
because t
i
0 and t
0
+...+t
q
= 1. Therefore the diameter of an ane simplex [v
0
, ..., v
q
] is
equal to the maximum distance between vertices max[v
i
v
j
[. Now let [b, w
0
, ..., w
q1
]
be a simplex in the subdivision of [v
0
, ..., v
q
]. By induction on q, we may assume that for
any i, j 0, ..., q 1, that
[w
i
w
j
[ diam([w
0
, ..., w
q1
])
q 1
q
diam[v
0
, ..., v
k
, ..., v
q
])
q
q + 1
diam([v
0
, ..., v
q
]).
On the other hand, if one of the vertices is b we have a bound
[b w
j
[ max[b v
i
[ [ i = 0, ..., q.
Notice that
b =
1
q + 1
v
i
+
q
q + 1
b
i
where b
i
is the barycentre of the face [v
0
, ..., v
i
, ..., v
q
]. Thus
[v
i
b[ =
q
q + 1
[v
i
b
i
[ diam([v
0
, ..., v
n
]).
completing the calculation.
28
A consequence of Lemma 3.18 is that repeatedly iterating barycentric subdivision
allows us to decompose a given simplex into simplices of arbitrarily small diameter.
We now turn this geometric construction into an algebraic one. We will dene a chain
map:
B : S(X) S(X)
which is chain homotopic to the identity map. This means B induces the identity map
on homology. For any chain c S
q
(X) we will show that B
k
(c) S
V
q
(X) for suciently
large k and the theorem will follow pretty easily from that.
To begin, assume X is a convex subset of R
n
and consider only chains lying in the
subcomplex S
aff
(X) S(X) generated by ane chains.
Let [v
0
, ..., v
q
] be an ane simplex in (a subset of) R
n
. Given b R
n
, dene
b[v
0
, ..., v
q
] = [b, v
0
, ...., v
n
]
and extend linearly to a map S
aff
q
(X) S
aff
q+1
(X).
Observe that
b[v
0
, ..., v
q
] = [v
0
, ..., v
q
] +
q

i=0
(1)
i+1
[b, ..., v
i
, ..., v
q
] = [v
0
, ..., v
q
] b[v
0
, ..., v
q
]
or more succinctly
b = b (4)
Next we dene the chain map B on S
aff
(X) S
aff
(X) by induction on degree. For
an ane simplex , denote b

its barycentre. For zero simplex B


0
() = . For q 1,
dene
B
q
() = b

B
q1
().
A little thought conrms that the ane simplices arising in the chain B
q
() are exactly
those arising in the geometric barycentric subdivision above.
We wish to show that B
q1
= B
q
to establish that B is a chain map. By induction,
suppose that B
q2
= B
q1
. For an ane q-simplex we have
B
q
() = b

B
q1

= B
q1
b

B
q1

= B
q1
b

B
q2

= B
q1

using induction, (4), and


2
= 0.
Next, we construct a chain homotopy
T
q
: S
aff
q
(X) S
aff
q+1
(X)
qZ
.
such that
T +T = Id B.
29
Figure 11: Chain homotopy geometrically
We dene T
q
inductively, by the T
1
= 0 and
T
q
= b

( T
q1
)
Assume inductively that T
q2
+T
q1
= Id B
q1
. Then
T
q
= b

( T
q1
)
= T
q1
b

( T
q1
)
= T
q1
b

( T
q1
)
= T
q1
b

(B
q1
() T
q1

2
)
= T
q1
b

B
q1
()
= T
q1
B
q

completing the induction.


Finally, we extend the denition of B and of T to the singular chain complex S(X)
of an arbitrary topological space. Let
q
:
q

q
be the identity map on the standard
simplex. For singular q-simplex dene
B = S
q
()B(
q
), T = S
q+1
()T(
q
).
Since commutes both with S
q
() and with B for ane chains, we get the formulas
B = S
q
()B(
q
) = S
q
()B(
q
) = B
and similarly
T +T = S()(T
q
+T
q
) = S()(
q
B
q
) = B.
Remark 3. The construction of B is totally natural and independent of the space X.
Thus, if A is a subspace of X, then B sends S(A) to S(A) and thus determines a chain map
B : S(X, A) S(X, A). Because T is natural the relative chain map is also homotopic
to the identity and B induces identity maps on H
q
(X, A) for any pair.
30
Lemma 3.19. Given an open cover 1 of X, and a chain c S(X), there exists k 0
such that B
k
(c) S
V
(X).
Proof. A chain is a nite formal sum of simplices and B distributes over sums. Thus it
is enough to prove that some k works for each simplex and then the largest of these will
work for the chain.
Let :
q
X be a singular simplex (a continuous map). The preimages
1
(V
i
) [ V
i

1 is an open cover of
q
. Therefore, there exists a Lebesgue number d > 0 such that
subset of
q
of diameter less than d is contained in some open set
1
(V
i
) in the cover.
Applying Lemma 3.18, by iterating barycentric subdivision a nite number k times, all
simplices in the decomposition are made to have diameter less than d. It follows that
B
k
() S
V
(X).
Proof of Theorem 3.14. We are attempting to show that the map H
q
(X B, A B)
H
q
(X, A) is an isomorphism.
To prove surjectivity we must show that every element of H
q
(X, A) is represented by
a chain S
q
(X) contained in the subgroup S
q
(X B). Let H
q
(X, A) be represented
by z S
q
(X). By Lemma 3.19, we can choose k such that B
k
(z) also represents , and
every simplex occurring in B
k
(z) maps into either X B or A. Discard those simplices
landing in A to get a chain c S
q
(X B) representing .
To prove injectivity, suppose H
q
(XB, AB) maps to 0 H
q
(X, A). This means
that is represented by z S
q
(X B) and there exists S
q+1
(X) such that
() = z +
where S
q
(A). Choose k so that the simplices of B
k
() are contained in one of X B
or A. Decompose B
k
() =
1
+
2
where
1
S(X B) and
2
S(A). Then

1
+
2
= B
k
() = B
k
= B
k
(z) + B
k
()
thus
(
1
) B
k
(z) = (
2
) + B
k
()
lies in both S
q
(XB) and S
q
(A), so must lie in S
q
(AB). It follows that B
k
(z) S
k
(XB)
represents and that B
k
(z) = (
1
) + where = ((
2
) + B
k
()) S
q
(A B), so
B
k
(z) is a relative boundary in S
q
(X B, A B), and we conclude that = 0.
3.5.2 Mapping cylinders and cones
A subspace A X is called a deformation retract if there is a homotopy h : XI A
such that h(x, 0) = x and h(x, 1) A for all x X and h(a, t) = a for all a A and t I.
The inclusion i : A X is a homotopy equivalence (this was a homework problem).
A closed subspace A X is called a neighbourhood deformation retract if there
exists an open neighbourhood A U X such that A is a deformation retract of U.
Example 12. The inclusion of S
1
R
2
is a neighbourhood deformation retract because
it includes as a deformation retract into an open annulus.
31
Proposition 3.20. Let A X be a neighbourhood deformation retract that intersects
every path component of X. Then there is a canonical isomorphism
H
q
(X, A)

=

H
q
(X/A)
where X/A is quotient space of X obtained by identifying all of A to a point.
Proof. Since A U is a homotopy equivalence, we know by Example 11 that
H
q
(X, A)

= H
q
(X, U)
for all q. By excision (Theorem 3.14)
H
q
(X A, U A)

= H
q
(X, U).
On the other hand, if we denote by A/A the point in X/A that A is collapsed to, it is not
hard to see that U/A deformation retracts onto A/A. Thus

H
q
(X/A)

= H
q
(X/A, A/A)

= H
q
((X/A) (A/A), (U/A) (A/A))

= H
q
(X A, U A),
where the rst isomorphism follows from Remark 2 since X/A is path connected.
Example 13. The quotient space D
n
/D
n
is homeomorphic to the sphere S
n
. It follows
that
H
q
(D
n
, D
n
)

=

H
q
(S
n
)
which was basically what we used in the proof of Proposition 3.16.
Example 14. If X contains a contractible neighbourhood deformation retraction A,
then H
q
(X)

= H
q
(X/A) for all q. (Indeed, one may show that X X/A is a homotopy
equivalence.
The hypotheses of Proposition 3.20 hold in many situations, but not always, so it is
convenient to have a construction that works in general. Let f : Y X be a continuous
map. The mapping cylinder associated to f is the quotient space (or adjunction)
Cyl(f) := (Y I) .
f
X = ((Y I) . X)/
where the relation is generated by (y, 1) f(y) for all y Y . The inclusion X Cyl(f)
is a homotopy equivalence with homotopy inverse Cyl(f) X x x and (y, t) f(x).
The mapping cone of f : Y X is the quotient space
Cone(f) := Cyl(f)/(Y 0)
Proposition 3.21. Given any map of spaces f : Y X such that the image of f
intersects every path component of X, we can dene a long exact sequence in homology


H
q+1
(Cone(f))

H
q
(Y )

H
q
(X)

H
q
(Cone(f))

H
q1
(Y ) ...
in case i : Y X is a subspace inclusion, this is canonically isomorphic to the long exact
sequence of the pair


H
q+1
(X, Y )

H
q
(Y )

H
q
(X)

H
q
(X, Y )

H
q1
(Y ) ...
32
Proof. The subspace Y 0 is a closed subset of Cyl(f) and is a deformation retract of
the open subset Y [0, 1), so Y 0 is a neighbourhood deformation retract in Cyl(f).
Thus by ... we have canonical isomorphisms

H
q
(Cone(f))

=

H
q
(Cyl(f), Y 0).
Since Y 0 is homotopy equivalent to Y and Cyl(f) is homotopy equivalent to X, the
long exact sequence ... can be obtained from the LES of the pair (Cyl(f), Y 0) by
replacing groups with isomorphic groups.
In case f : Y X is subspace inclusion, then the homotopy equivalence Cyl(f) X
sending (y, t) Y I to f(y) restricts to a homeomorphism from Y 0 to Y . The
resulting morphism of long exact sequences
H
q
(Y 0)

H
q
(Cyl(f))

H
q
(Cyl(f), Y 0)

H
q1
(Y 0)

H
q1
(Cyl(f))

H
q
(Y )

H
q
(X)

H
q
(X, Y )

H
q1
(Y )

H
q1
(X)
which must be an isomorphism by the Five Lemma.
Example 15. The wedge sum Let X
k
be a collection of spaces containing base points
p
k
X
k
. The wedge product is the space

k
X
k
= (.
k
X
k
)/
where we identify basepoints p
i
p
j
for all i, j. If the base points neighbourhood defor-
mation retracts then

H(
k
X
k
)

=
k

H(X
k
)
by Proposition 3.20.
3.6 Applications to spheres: the degree of a map
Recall that our calculation of H
n
(S
n
) relied on the following sequence of isomorphisms

H
n
(S
n
)

=

H
n
(S
n
, E
n

)

H
n
(E
n
+
, S
n1
)

=

H
n
(S
n1
)
We can use this to construct a cycle representing the generator of H
1
(S
1
) by the Figure
12
Indeed, the cycle we have constructed is the barycentric subdivision of a simplex
:
1
S
1
that winds once around the circle. It is not hard to show that (exercise) that
any chain of 1-simplices that wraps once around the circle also represents the generator
of H
1
(S
1
).
Recall that

H
n
(S
n
)

= Z. Given a continuous map f : S
n
S
n
, the induced map
f

:

H
n
(S
n
)

H
n
(S
n
) must be of the form f

() = d for some integer d Z. We call


d = deg(f) the degree of the map f, .
33
Figure 12: Constructing a cycle generating H
1
(S
1
)
Figure 13: Cycles generating H
1
(S
1
)
Since H
n
is a functor, we see immediately that deg(Id
S
n) = 1, that deg(f g) =
deg(f) deg(g) for two maps f, g : S
n
S
n
, and that homotopic maps are have the same
degree.
6
Proposition 3.22. A map f : S
n
S
n
that is not surjective has degree zero.
Proof. Suppose p S
n
is not in the image of f. Then f factors through the inclusion
map S
n
S
n
p S
n
so by functoriality H
q
(f) factors through H
q
(S
n
p)

= 0 and
thus must be zero.
Given a space X, dene the suspension SX := X I/ to be the quotient of X I
where collapses X 0 and X 1 to distinct points. If f : X Y is a map, dene
the suspension of f
Sf : SX SY, Sf(x, t) = (f(x), t).
This denes the suspension functor from spaces to spaces.
Lemma 3.23. The suspension of a sphere satises SS
n
= S
n+1
. Given a map f : S
n

S
n
, the suspension Sf : S
n+1
S
n+1
satises deg(f) = deg(Sf).
Proof. The homeomorphism SS
n
= S
n+1
is pretty clear; this is the picture where S
n
includes into S
n+1
as the equator. Because the long exact homology sequence is functorial
with respect to pairs and the excision isomorphism is canonical, we obtain a commutative
6
It is also true that two maps from S
n
to S
n
are homotopic if and only if they have the same degree.
The proof of this is beyond the scope of this course.
34
diagram of with horizontal arrows isomorphisms

H
n
(S
n
)

(Sf)

H
n
(S
n
, E
n

)
(Sf)


H
n
(E
n
+
, S
n1
)
(Sf)


H
n
(S
n1
)
f

H
n
(S
n
)

H
n
(S
n
, E
n

)

H
n
(E
n
+
, S
n1
)

H
n
(S
n1
)
so S(f) and f have the same degree.
Proposition 3.24. Let r
n
: S
n
S
n
be a restriction of a reection on R
n+1
R
n+1
.
Then deg(f) = 1.
Proof. For n 1, we can identify r
n
= Sr
n1
, so by induction it suces to prove the case
n = 0. In this case, S
0
= N, S is a pair of points and r
0
transposes them. The points
represent 0-simplices and

H
0
(S
0
) is generated by [N] [S]. We have
(r
0
)

([N] [S]) = [r
0
(N)] [r
0
(S)] = [S] [N] = ([N] [S])
so deg(r
0
) = 1.
We dene the antipodal map on S
n
by x x.
Proposition 3.25. If f : S
n
S
n
is a map with no xed points (i.e. there is no point
p S
n
such that f(p) = p), then f is homotopic to the antipodal map. In particular,
deg(f) = (1)
n+1
.
Proof. If f has no xed points, then the path tf(x) (1 t)x does not pass through the
origin. It follows that
h : S
n
I S
n
, h
t
(x) =
tf(x) (1 t)x
[tf(x) (1 t)x[
is a homotopy joining the antipodal map h
0
to f = h
1
. Finally, note that the antipodal
map is equal to a composition of (n+1) reections on S
n
R
n+1
so it has degree (1)
n+1
by Proposition 3.24.
Theorem 3.26 (Hairy Ball Theorem). Every continuous vector eld on an even dimen-
sional sphere has a zero.
Proof. A continuous vector eld on a S
n
is equivalent to a map V : S
n
R
n+1
such that
V (x) is orthogonal to x for all x S
n
. If a non-vanishing vector eld V exists, then we
can dene an associated map f : S
n
S
n
by f(x) = V (x)/[V (x)[ which has no xed
points. By Proposition ..., this implies that deg(f) = (1)
n+1
One the other hand, since f(x) and x are always orthogonal, we can build a homotopy
h : S
n
I S
n
, h
t
(x) = cos(t/2)x +sin(t/2)f(x)
between the identity map and f, from with we conclude that deg(f) = 1. If n is even,
this leads to a contradiction.
35
Remark 4. In contrast with Theorem 3.26, if n is odd S
n
always admits a non-vanishing
vector eld. This is because S
2m1
R
2m
= C
m
and we can use complex scalar multica-
tion to rotate each vector by 90 degrees. Explicitly, V (x
1
, y
1
, ..., x
n
, y
n
) = (y
1
, x
1
, ..., y
n
, x
n
).
Alternative approach:
For n 0, the homology group

H
n
(S
n
) is isomorphic to Z. There are two possible
isomorphisms H
n
(S
n
)

= Z depending on a choice of generator. A choice of this generator
is called an (global) orientation of S
n
.
Given a point p S
n
and an open neighbourhood p U S
n
, we have canonical
isomorphisms
H
n
(S
n
)

=
H
n
(S
n
, S
n
p)

=
H
n
(U, U p). (5)
composing the long exact sequence of the pair (S
n
, S
n
p) with excision. A choice of
orientation for H
n
(U, U p)

= Z is called a local orientation of S
n
at p. Because
the isomorphism (5) is natural, an orientation of S
n
determines local orientations at all
points p S
n
, and vice versa.
Now suppose that f : S
n
S
n
is a map and for some point p S
n
the preimage
f
1
(p) is a nite set of points q
1
, ..., q
k
S
n
.
7
Suppose further that for some open
neighbourhood p U the preimage f
1
(U) is a disjoint union of open sets V
1
... V
k
for which q
i
V
i
. For each i, the restriction of f induces homomorphism
H
n
(V
i
, V
i
q
i
) H
n
(U, U p).
Since both groups are isomorphic to Z, the homomorphism must be multiplication by an
integer d
i
which we call the local degree.
Proposition 3.27. Under the conditions above, the degree of f is the sum of the local
degrees: deg(f) =

k
i=1
d
i
.
Proof. Fix an orientation Z = H
n
(S
n
) and use this to impose local orientations at all
points. We have a commutative diagram where natural and orientation isomorphisms are
indicated by double lines.
H
n
(S
n
)
f

H
n
(S
n
)
H
n
(S
n
, S
n
f
1
(p))
f

H
n
(S
n
, S
n
p)
Z
A

k
i=1
H
n
(V
i
, V
i
q
i
)
f

H
n
(U, U p)
Z
k
B

Z
7
Such a point always exists if f is dierentiable (Sards Theorem)
36
In matrix notation, we have
A =
_
_
_
_
1
1
...
1
_
_
_
_
B =
_
d
1
d
2
... d
k
_
So the composition is d
1
+... +d
k
is the degree of f.
In the simplest case, p and U can be chosen so that f restricts to local homemorphisms
V
i
U. In this case the local degrees are all 1, so the degree is obtained by counting
points q
1
, ..., q
k
signs according to whether f is locally orientation preserving or reversing.
Example 16. We can construct a map S
n
S
n
of degree d 2 as follows. Let A S
n
be the complement of d disjoint open disks B
i
in S
n
. Let
q : S
n
S
n
= X/A

=
d
S
n
be the quotient map. The orientation on S
n
induces local orientations and hence global
orientations on each sphere in the wedge sum. Let
p :
d
S
n
S
n
map each sphere by a degree 1 homeomorphism to S
n
.
The preimage (pq)
1
(y) of a generic point y S
n
consists of a single point in each
disk B
i
each with local degree is 1 because pq is a local homeomorphism. Therefore
deg(pq) = d. By precomposing pq with a reection, we can construct a map of degree d.
Consider the map given d Z
w
d
: S
1
S
1
, w
d
(e
i
) = e
id
for d 1 we can see by Figure 14 that deg(w
d
) = d.
Figure 14: the winding map w
4
37
Note that w
d
is equal to the composition of w
d
with a reection, so deg(w
d
) = d.
By suspension, we construct maps
S
n
w
d
: S
n+1
S
n+1
of degree d for any integer.
Theorem 3.28 (Fundamental Theorem of Algebra). A complex polynomial function f(z)
of degree d 1 has a complex root.
Proof. The case d = 1 is obvious, so suppose that d 2. We assume f is monic for
simplicity so f(z) = z
d
+ O(z
d1
). Assume that f(z) has no complex roots. Then there
is a well-dened, continuous map
g : C S
1
, g(z) =
f(z)
[f(z)[
_
=
z
d
+O(z
d1
)
[z
d
+O(z
d1
)[
_
.
Dene a homotopy h : S
1
I S
1
by
h
t
(e
i
) = g(
t
1 t
e
i
).
for t < 1 and extend by continuity for t = 1. We have h
0
(e
i
) = g(0) is a constant and
thus deg(h
0
) = 0. On the other hand, for large values of z, g(z) becomes dominated by
the highest degree terms in the numerator and denominator, so in the limit t 1, we
have
h
1
(e
i
) = e
id
so deg(h
1
) = d, which contradicts degree being a homotopy invariant.
3.7 Cellular homology
3.7.1 Cell complexes
Let
D
n
:= x R
n
[[x[ 1
denote the unit disk or closed n-cell with boundary
S
n1
= D
n
:= x R
n
[[x[ = 1.
Given a topological space X and a continuous map f : S
n1
X, we may construct a
new space
Y := (X HD
n
)/
where we quotient by the equivalence relation generated by p f(p) for all p S
n1
.
We say that Y is obtained from X by attaching an n-cell; the map f is called the
attaching map. More generally, if we have a collection of maps f

: S
n1
X, then we
construct
Y = (X H(

D
n

))/
38
where p f

(p) for all p S


n1

and .
A cell complex (also called CW-complex) is a space that is constructed inductively
by attaching cells. For instance,
A 0-dimensional complex X
0
is a discrete set of points (i.e. a disconnected union of
0-cells).
A 1-dimensional cell complex X
1
is a space constructed by attaching a collection of
1-cells to X
0
.
A 2-dimensional cell complex X
2
is constructed by attaching 2-cells to X
1
.
and so on ...
In general, a cell complex X may have cells in arbitrarily high dimensions, in which case
it is called -dimensional. Each n-cell determines a characteristic map

: D
n
X.
A subset S X is open/closed if and only if
1

(S) D
n
is open/closed for all cells.
Example 17. A wedge of n-spheres
I
S
n
is constructed by attaching I many n-cells onto
a point X
0
= p by the only possible attaching map f : S
n1
p.
Example 18. The torus S
1
S
1
can be constructed by attaching a 2-cell onto a wedge
of two circles X = S
1
S
1
. If we denote by a and b the loops dened by the two circles
in X, then the attaching map f : S
1
X is the loop a b a
1
b
1
.
Example 19. More generally, the genus g surface
g
is constructed by gluing a 2-cell to
a wedge of 2g circles. If the loops dened by the circles are called a
1
, b
1
, ..., a
g
, b
g
, then
the attaching map sends S
1
to the concatenation

g
i=1
[a
i
, b
i
], where [a
i
, b
i
] = a
i
b
i
a
1
i
b
1
i
is the commutator.
A subspace A X is called a subcomplex if it is a closed union of cells (that is,
of images of characteristic maps). Given a subcomplex A X, the quotient space X/A
dened by identifying all points in A with each other, is naturally a cell complex called a
quotient complex of X.
Proposition 3.29. A subcomplex A of a cell complex X is a neighbourhood deformation
retract. Thus, H
q
(X, A)

=

H
q
(X/A) we have a long exact sequence in homology
. . .

H
q+1
(X/A) H
q
(A) H
q
(X)

H
q
(X/A) H
q1
(A) . . .
Proof. Skipped.
The subcomplex X
n
X consisting of all cells of dimension n is called the n-
skeleton of X (by convention X
1
= ). If X is innite dimensional, the topology on
X satises S X is open (resp. closed) if and only if S X
n
is open (resp. closed) in
X
n
for all n. In particular, a map f : X Y is continuous if and only if the restrictions
f
n
: X
n
Y are continuous for all n. We say X has the direct limit topology with respect
to X
n
.
39
Figure 15: orientable surfaces
Lemma 3.30. Let X be a cell complex and C X a compact subspace. Then C is
contained within nitely many cells of X.
Proof. Choose a sequence of points x
i
C lying in distinct cells. We will show that the
set S := x
i
is nite. We begin by showing S is closed.
First observe that
S X is closed S X
n
is closed in X, n
We use induction on n.
Clearly SX
0
is closed in X
0
hence in X, because every subset of X
0
is closed. Assume
by induction that S X
n1
is closed in X. Thus for any characteristic map

: D
k
X
the pre-image
1

(S X
n1
) is closed in D
k
. For k < n, the pre-image
1

(S X
n
) =

)(S X
n1
) is closed in D
k
. For k = n, the pre-image
1

(S X
n
) D
n
equals

(S X
n1
) plus at most one point, thus it is a union of two closed sets, hence is closed
in D
n
. We deduce that S X
n
is closed in X
n
hence also in X. By induction, this holds
for all n so S is closed in X.
The same argument shows that every subset of S is also closed, so S has the discrete
topology. But S is a closed subset of the compact set C, so it is compact. We conclude
that S is nite.
3.7.2 Cell complex propaganda (not to be tested)
We present some results showing that many interesting spaces are either homeomorphic
or homotopy equivalent to cell complexes. The material in this section will not be tested.
40
Denition 17. A real analytic function f : R
n
R is a innitely dierentiable
function such that at every point p R
n
, f equals its Taylor series at p on some positive
radius. A real analytic set X R
n
is the solution set nite collection of equations
f
1
(x) = ... = f
n
(x) = 0, for f
i
real analytic.
Example 20. Polynomial functions, exponential functions, trigonometric functions, etc.
are real analytic.
Theorem 3.31 (Lojasiewicz 1964). Every real analytic set X R
n
is homeomorphic to
a cell complex.
Let X and Y be two topological spaces. Let Cont(X, Y ) be the set of all continuous
maps from X to Y with the compact-open topology.
Example 21. The space LY = Cont(S
1
, Y ) is called the free loop space of Y .
Theorem 3.32 (Milnor 1959). If X and Y are cell complexes and X is compact, then
Cont(X, Y ) is homotopy equivalent to a cell complex.
Denition 18. A topological space X is called a (topological) n-manifold if it is Haus-
dor and if every point p is contained in an open neighbourhood p U X that is
homeomorphic to R
n
.
Every open set in R
n
is an n-manifold.
The sphere S
n
is an n-manifold.
Surfaces of any genus are 2-manifolds.
The product of an m-manifold and an n-manifold is an m +n-manifold.
An example of a space that is locally Euclidean but is not a manifold is constructed
by taking two copies of the real line R H R = R a, b and forming the quotient by
(t, a) (t, b) if t ,= 0. This space looks locally like R, but the points (0, a) and (0, b)
cannot be separated by open sets.
Theorem 3.33. Every compact n-manifold is homotopy equivalent to a cell complex. It
remains an open question whether or not every compact n-manifold is homeomorphic to
a cell complex.
3.7.3 Cellular Homology
Lemma 3.34. If X is a cell complex, then:
(a) H
q
(X
n
, X
n1
) is zero if q ,= n and is an free abelian group with generators corre-
sponding to the n-cells when q = n.
(b) H
q
(X
n
) = 0 for q > n. Thus H
q
(X) = 0 for q > dim(X).
41
(c) The inclusion i : X
n
X induces an isomorphism H
q
(i) : H
q
(X
n
) H
q
(X) for
q < n.
Proof. By Proposition 3.29, we have isomorphism H
q
(X
n
, X
n1
)

=

H
q
(X
n
/X
n1
) and
X
n
/X
n1
is a wedge of spheres indexed by the n-cells of X. Property (a) follows.
Property (b) is proven by induction. Clearly true for n = 0. Now suppose it has been
proven for n 1. The long exact sequence of the pair contains
H
q
(X
n1
) H
q
(X
n
) H
q
(X
n
, X
n1
)
where both H
q
(X
n1
) = H
q
(X
n
, X
n1
) = 0 for q > n by induction and property (a). Thus
H
q
(X
n
) = 0 as well.
To prove property (c), consider the exact sequence
H
q+1
(X
n+1
, X
n
) H
q
(X
n
) H
q
(X
n+1
) H
q
(X
n+1
, X
n
).
By (a), the two groups on the end vanish if q < n so H
q
(X
n
)

= H
q
(X
n+1
). Repeating this
argument, we get
H
q
(X
n
)

= H
q
(X
n+1
)

= H
q
(X
n+2
)

= ...
which suces if X is nite dimensional. To take care of the innite dimensional case,
observe that Lemma 3.30 implies that every chain in S
q
(X) must be in the image of
S
q
(X
n
) for some n (since the union of images of simplices occurring in the chain is a
compact subset of X). Thus every cycle Z
q
(X) arises as the image of a cycle in Z
q
(X
n
)
for some n, and every boundary in B
q
(X
n
) arises as the image of a boundary in B
q
(X
n
)
for some n. The result follows (details left to reader).
Dene a homomorphism d
n
: H
n
(X
n
, X
n1
) H
n1
(X
n1
, X
n2
) by the commutative
diagram
0
0

P
P
P
P
P
P
P
P
P
P
P
P
P
P
P H
n
(X
n+1
)

= H
n
(X)

j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
H
n
(X
n
)
jn

R
R
R
R
R
R
R
R
R
R
R
R
R

l
l
l
l
l
l
l
l
l
l
l
l
l
H
n+1
(X
n+1
, X
n
)
d
n+1

n+1

n
n
n
n
n
n
n
n
n
n
n
n
H
n
(X
n
, X
n1
)
dn

S
S
S
S
S
S
S
S
S
S
S
S
S
S
H
n1
(X
n1
, X
n2
)
H
n1
(X
n1
)
j
n1

l
l
l
l
l
l
l
l
l
l
l
l
l
l
0

j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
j
42
where the diagonal maps occur in the long exact sequences of pairs. Notice that d
n
d
n+1
=
0 because it factors through
n
j
n
= 0. Thus (H
n
(X
n
, X
n1
), d
n
)
nZ
forms a chain
complex, called the cellular chain complex. The homology of the cellular chain complex
is called the cellular homology.
Theorem 3.35. The cellular homology groups are naturally isomorphic to the singular
homology groups.
Proof. From diagram, we may identify H
n
(X)

= H
n
(X
n
)/im(
n+1
). Since j
n
is injective,
this is isomorphic to im(j
n
)/im(d
n+1
). By exactness, this is the same as ker(
n
)/im(d
n+1
).
Finally, because j
n1
is injective, this is equal to ker(d
n
)/im(d
n+1
).
Theorem 3.35 is very useful for calculations, because it allows us to replace the usually
uncountably innite rank S
q
(X) by the often nite rank H
n
(X
n
, X
n1
). Before getting
started with examples, we want a more direct understanding of the boundary maps d
n
.
Denote by e
n

the set of n-cells of a cell complex X, so that H


n
(X
n
, X
n1
) is the
free abelian group generated by e
n

.
Proposition 3.36. For n > 1, the cellular boundary map satises
d
n
(e
n

) =

d
,
e
n1

where d
,
is the degree of the map
S
n1

X
n1
S
n1

dened by composing the attaching map of e


n

with the quotient map X


n1
S
n1

=
X
n1
/(X
n1
e
n

).
Proof. The proof is based on the following commutative diagram
H
n
(D
n

, D
n

H
n1
(D
n

)
f

(
,
)

H
n1
(S
n1

)
H
n
(X
n
, X
n1
)
n

dn

S
S
S
S
S
S
S
S
S
S
S
S
S
S

H
n1
(X
n1
)
q

j
n1

H
n1
(X
n1
/X
n2
)

H
n1
(X
n1
, X
n2
)

=

H
n1
(X
n1
/X
n2
, X
n2
/X
n2
)
where

is the characteristic map for e


n

and f

the attaching map.


q : X
n1
X
n1
/X
n2
is the quotient map.
q

: X
n1
/X
n2
S
n1

is the quotient map obtained by collapsing everything be


the sphere to a point.
43

,
= q

q f

is the map whose degree is d


,
.
The map

sends the generator of H


n
(D
n

, D
n

) to e
n

and q

can be understood as
projection onto the -summand of

H
n1
(X
n1
/X
n2
)

= H
n1
(X
n1
, X
n2
). The result
follows by a diagram chase.
3.7.4 Examples
Example 22. Recall that a genus g surface
g
is constructed by attaching a 2-cell to
a wedge of 2g circles using an attaching map

g
i=1
[a
i
, b
i
]. The cellular complex is:
0 Z
d
2
Z
2g
d
1
Z
The boundary map d
1
is zero, because each 1-cell meets the 0-cell twice and the attaching
map S
0
pt sends

H
0
(S
0
) to zero by denition. The boundary map d
2
is also zero,
because the attaching map winds around each loop twice, but in opposite directions,
giving total degree zero. It follows that
H
q
(
g
) =
_

_
Z if q = 0, 2
Z
2g
if q = 1
0 otherwise
Example 23. The non-orientable surface N
g
of genus g 0 is constructed by attach-
ing a single 2-cell to a wedge sum of g + 1 circles a
0
, ..., a
g
by the attaching map a
2
0
...a
2
g
.
The surface N
0
in the real projective plane and N
1
is the Klein bottle. The cellular
chain complex is
Z
d
2
Z
g+1
d
1
Z.
As before, d
1
= 0. The attaching map for the 2-cell winds twice around each circle in the
same direction and thus has degree 2 for each 1-cell. Consequently, d
2
(n) = (2n, ..., 2n). If
we do a change of basis for Z
2g
using generators (1, 1, ..., 1), (0, 1, 0, ..., 0), ..., (0, 0, ..., 0, 1)
then with respect to the new basis, d
2
(n) = (2, 0, ..., 0).
H
q
(N
g
) =
_

_
Z if q = 0
Z
2
Z
g
if q = 1
0 otherwise
Example 24. A product of spheres S
m
S
n
with m, n 1 has the structure of a
cell complex with four cells, in dimensions 0, m, n, and m+n. To see this observe that
D
m+n

= D
m
D
n
and
S
m
S
n

= (D
m
D
n
)/

= (D
m
/ ) (D
n
/ )
where is generated by the relations (x, y) (x, y

) if y, y

D
n
and (x, y) (x

, y)
if x, x

D
m
. Since the relations occur only in the boundary of D
m
D
n
, this can be
understood as attaching a (m +n)-cell to the quotient of
(D
m
D
n
) = D
m
D
n
D
m
D
n
44
Figure 16: non-orientable surfaces
which is identied as the wedge sum S
m
S
n
, which is a cell complex with cells in
dimension 0, m, and n (compare the case of a torus S
1
S
1
).
Suppose that n m and n > 1. Then the (n + 1)-skeleton is equal to S
m
S
n
so by
Lemma 3.34, H
q
(S
m
S
n
) = H
q
(S
m
S
n
) for q n. It follows that the boundary map
in the cellular chain complex is trivial and that
H
q
(S
m
S
n
) =
_
Z if q = 0, m, n, or m +n
0 otherwise
Example 25. Real projective space RP
n
is the set of 1-dimensional vector subspaces
of R
n+1
. Topologically, RP
n
is the quotient space of R
n
0 by the relation v v for
R

a non-zero scalar. Equivalently, RP


n
is identied with quotient space S
n
/ by
the relation v v.
By restricting to a closed hemisphere of S
n
, we see that RP
n
can be obtained by
attaching an n-cell to RP
n1
using the quotient S
n1
RP
n
as the attaching map. By
induction, RP
n
= e
0
e
1
... e
n
is a cell complex with one cell in each dimension n.
Thus, the cellular chain complex looks like
0 Z
dn
Z
d
n1

d
1
Z 0
To understand the boundary maps, we must determine the degrees of the composition of
the attaching map and the quotient map
S
q1

RP
q1
q
RP
q1
/RP
q2

= S
q1
.
Observe that q restricts to local homeomorphisms from each of the two open hemi-
spheres S
q1
S
q2
onto S
q1
pt so by ... the degree is either 0 or 2. Notice that
these two restrictions are interchanged by precomposing with the antipodal map which
has degree (1)
q
. Thus the local degrees have the same sign if q is even and dierent
signs if q is odd. Thus d
q
is multiplication by 0 if q is odd and 2 if q is even. The chain
complex is
0 Z
0
Z
2
Z
0

2
Z
0
Z 0
45
if n is odd and
0 Z
2
Z
0
Z
2

2
Z
0
Z 0
if n is even. We obtain homology groups,
H
q
(RP
n
) =
_

_
Z if q = 0 or q = n and n is odd
Z
2
if 0 < q < (n 1) and is odd
0 otherwise
Example 26. Complex projective space CP
n
is the set of one dimensional vector
subspaces of C
n+1
. It may also be construct as the quotient of the sphere S
2n+1
by the
relation v v where S
1
is a unit scalar. Similar to the real projective spaces, we
can construct CP
n
inductively by attaching a 2n-dimensional cell to CP
n1
. To see this,
consider the embedding from D
2n
C
n
to C
n
C by
w (w,
_
1 [w[
2
).
The boundary of D
2n
is sent to the unit sphere in S
2n1
C
n
0 and there is a one-
to-one correspondence between the interior of D
n
and the one-dimensional subspaces of
C
n
C not contained in C
n
0. Thus CP
n
is obtained by attaching D
2n
to CP
n1
by
the quotient map S
2n1
CP
n1
.
Since CP
n
= e
0
e
2
...e
2n
only has cells in even dimension, this means the boundary
maps in the cellular chain complex are necessarily zero and we obtain
H
q
(CP
n
) =
_
Z if 0 q 2n and is even
0 otherwise
3.8 Mayer-Vietoris Sequence
The Mayer-Vietoris sequence is an alternative to the long exact sequence of a pair (X, A)
that sometimes more convenient to use.
Let X be a topological space and let A, B X be a pair of subspaces such that
A B = X (i.e. a covering of X).
Theorem 3.37. Let X be a topological space and let A, B X be a pair of subspaces
such that A B = X. If we have an excision isomorphism
H
q
(B, A B) = H
q
(X B
c
, A B
c
)

= H
q
(X, A)
then there is a long exact sequence in homology
H
n
(A B)
I
H
n
(A) H
n
(B)
J
H
n
(X)
D
H
n1
(A B) . . .
where I, J, and D are dened below. A similar sequence exists for reduced homology


H
n
(A B)


H
n
(A)

H
n
(B)


H
n
(X)


H
n1
(A B) . . . .
46
Example 27. The excision hypothesis of Theorem 3.37 is satisied if
int(A) int(B) = X
or if X is a cell complex and A, B are sub-cell complexes.
Proof of Theorem 3.37. consider that inclusion determines of pairs (X B
c
, A B
c
) =
(B, A B) (X, A) and thus a morphism of long exact sequences
. . .
H
q+1
(B, A B)

H
q
(A B)

H
q
(B)

H
q
(B, A B)

. . .
. . .
H
q+1
(X, A)

H
q
(A)

H
q
(X)

H
q
(X, A)
. . .
where every third map between the sequences is an isomorphism. The results follows by
the following lemma.
Lemma 3.38. Given a commuting diagram

C
i+1

i+1

h
i+1

A
i
f
i

B
i

g
i

C
i
h
i

A
i1

i1

i+1
h

i+1

i
f

i
g

i
h

i1

in which the rows are exact and the
i
are isomorphisms, then there is a long exact
sequence
A
i
I
A

i
B
i
J
B

i
D
A
i1

where
I(a) = (
i
(a), f
i
(a)),
J(a

, b) = f

i
(a)
i
(b),
D(b

) = h
i

1
i
g

i
(b

).
Proof. Diagram chase.
Corollary 3.39. For a path connected space X, there is a canonical isomorphism

H
q+1
(SX)

=

H
q
(X)
for all q Z, where SX denotes the suspension of X.
Proof. Homework exercise.
47
3.9 Homology with coecients
So far we have developed singular homology theory for integer coecients, meaning that
our chains are nite formal sums

with coecients a

Z. More generally, it is
possible (and useful) to work with coecients in any commutative ring R with identity
1 R; in particular, this means we have a canonical ring homomorphism Z R. The
most interesting cases are when R = Z
n
is the ring of integers modulo n, or when R is a
eld, such as Q, R, C or Z
p
for prime p.
This generalization is pretty straight forward. Dene S
q
(X; R) to be the free R-module
whose chains are nite sums the form

where the coecients a

are elements
of R. So far we have been studying the case S
q
(X) = S
q
(X; Z). The boundary map

q
: S
q
(X; R) S
q1
(X; R) is dened as before

q
(

) =

(a

) =

i=0
(1)
i
a

(i)
and satises
2
= 0 giving rise to a chain complex S(X; R) = (S
q
(X; R);
q
). This
determines cycles Z
q
(X; R) := ker(
q
), boundaries B
q
(X; R) := im(
q+1
) and homology
H
q
(X; R) := Z
q
(X; R)/B
q
(X; R),
all of which are now R-modules. We call H
q
(X; R) singular homology with R-coecients.
8
The proofs of all the main theorems carry over unchanged. In particular
H
q
(; R) is a functor for topological spaces to the category of R-modules. Thus for
every continuous map f : X Y , there is a homomorphism of R-modules
H
q
(f; R) = f

: H
q
(X; R) H
q
(Y ; R)
such that (f g)

= f

and (Id
X
)

= Id
Hq(X;R)
.
If f, g : X Y are homotopic, then f

= g

.
A topological pair (X, A) gives rise to a long exact sequence of R-modules
H
q
(A; R) H
q
(X; R) H
q
(X, A; R)

H
q1
(A; R) ...
where the connecting homomorphism is dened using the same diagram chase as
before.
The excision isomorphism H
q
(XB, AB : R)

= H
q
(X, A; R) hold under the same
hypotheses as before.
H
0
(X; R) is a free R-module generated by the path components of X.
8
Note that an abelian group is the same thing as a Z-module. If R is a eld, an R-module is just an
R-vector space.
48
H
0
(pt; R) = R and H
q
(pt; R) = 0 for q ,= 0. Dene

H
q
(X; R) to be the kernel of the
natural map H
q
(X; R) H
q
(pt; R).


H
n
(S
n
; R)

= R and

H
q
(S
n
; R) = 0 for q ,= n. If f : S
n
S
n
is a degree d map,
then f

is simply multiplication by d R.
If X is a cell complex, then H
n
(X
n
, X
n1
; R) is a free R-module with genera-
tors corresponding to n-cells. We can form a cellular chain complex C
n
(X; R) =
H
n
(X
n
, X
n1
; R) and boundary map dened by ... and cellular homology is natu-
rally isomorphic to singular homology.
The Mayer-Vietoris Sequence is still exact and dened in the same way.
One of the big advantages in working with coecients in a eld F is that homology
are now vector spaces. This means for instance that short exact sequences always split
and this can simplify a lot of calculations.
Example 28. Recall that the cellular chain complex C(RP
2
) (Example 25) is either of
0 Z
0
Z
2
Z
0

2
Z
0
Z 0
0 Z
2
Z
0
Z
2

2
Z
0
Z 0
depending on whether n is odd or even.
Now consider the cellular chain complex C(RP
n
; Z
2
). The chain groups become free
Z
2
-modules and the boundary maps are all zero because 0 = 2 in Z
2
0 Z
2
0=2
Z
2
0=2
Z
2
0=2

0=2
Z
2
0=2
Z
2
0.
Thus
H
q
(RP
n
; Z
2
) =
_
Z
2
for q = 0, ..., n
0 otherwise.
On the other hand, if F is a eld of characteristic other than 2, then 2 is a unit. The
cellular chain complexes becomes one of
0 F
0
F

=
F
0


=
F
0
F 0
0 F

=
F
0
F

=


=
F
0
F 0
and we get homology groups
H
q
(RP
n
; F) =
_
F if q = 0 or q = n and n is odd
0 otherwise.
49
The only isomorphism invariant of a vector space is the dimension. Given a eld F the
dimensions dim(H
q
(X; F)) are called Betti numbers of X. These are basic invariants
of a space X that are often easier to compute than the homology groups H
q
(X; Z).
Suppose that X is a space for which dim(H
q
(X); F) is nite in all degrees and is zero
in all but nitely many degrees. Dene the Euler characteristic
(X) :=

q=0
(1)
q
dim(H
q
(X; F))
to be the alternating sum of the Betti numbers.
Proposition 3.40. If X is nite cell complex, the Euler characteristic is well dened and
satises
(X) :=

n=0
(1)
n
c
n
where c
n
is the number of cells of dimension n. In particular, (X) is a homotopy invari-
ant of X and is independent of the eld F.
Proof. We will simplify notation to prove this result in a more abstract context. Suppose
we have a nite length chain complex of vector spaces
0 C
N
n
...

1
C
0

0
0
for which each C
q
is nite dimension c
q
. For each q we have short exact sequences of
vector spaces
0 Z
q
C
q
B
q1
0
and
0 B
q
Z
q
H
q
0.
Denote by z
q
, b
q
, h
q
the dimensions of Z
q
, B
q
, H
q
respectively. The rank nullity theorem
gives us equations
c
q
= z
q
+b
q1
, h
q
= z
q
b
q
for all q Z. In case C
q
= C
q
(X; F) we get
(X) =

q
(1)
q
c
q
=

q
(1)
q
(z
q
+b
q1
)
=

q
(1)
q
z
q
(1)
q1
b
q1
=

q
(1)
q
(z
q
b
q
)
=

q
(1)
q
h
q
50
A surface of genus g has Euler characteristic 1 2g + 1 = 2 2g.
CP
n
has Euler characteristic 1 + 0 + 1 + 0 +... + 1 = n + 1.
A sphere has Euler characteristic 1 + 1 = 2 if n is even and 1 1 = 0 if n is odd.
A well known example of the Euler characteristic comes up for the boundaries of convex
polytopes, like a tetrahedron or a cube. The Euler characteristic is then V E+F where
V is the number of vertices, E is the number of edges, and F is the number of faces. Since
the boundary of a convex polytope must be homeomorphic to S
2
, it follows that
V E +F = 2
Theorem 3.41 (Lefshetz Fixed Point Theorem). Let X be a nite cell complex and let
f : X X be a continuous map. The Lefshetz number for f is the number
L(f) :=

q=0
(1)
q
Tr(H
q
(f; Q))
where Tr(H
q
(f; Q)) is the trace of the linear map H
q
(f; Q) : H
q
(X; Q) H
q
(X; Q). If
L(f) ,= 0, then f has a xed point.
Some examples
(Generalization of Brower Fixed Point Theorem) If X is a nite cell complex such
that H
0
(X; Q) = Q and H
q
(X; Q) = 0 for q > 0, then L(f) = 1 for any map
f : X X. Consequently every f must have a xed point.
Note that it is a necessary condition that X is nite, because the translation map
t : R R, t(x) = x + 1 has L(t) = 1 but no xed points.
A map of spheres f : S
n
S
n
with degree d has Lefshetz number L(f) = 1+(1)
n
d
which vanishes if and only if d = (1)
n+1
, recovering ...
If f is homotopic to the identity map, then L(f) = (X). Thus if (X) ,= 0 then
any f homotopic to the identity must have a xed point.
Proof. We will only briey sketch the argument.
Suppose that the map f : X X restricts to maps f : X
k
X
k
for each skeleton.
If this happens, it is not hard to show that f determines a morphism from the cellular
chain complex of X to itself:
. . .
C
q+1
(X)

C
q+1
(f)

C
q
(X)

Cq(f)

C
q1
(X)

C
q1
(f)

. . .
. . .
C
q+1
(X)

C
q
(X)

C
q1
(X)
. . .
51
Using an argument similar to the proof of Proposition ..., one shows that
L(f) =

q=0
(1)
q
Tr(C
q
(f; Q)).
In particular, this means that if L(f) ,= 0, then Tr(C
q
(f; Q)) ,= 0 for some q. This implies
in particular that some cell e
k
X must map onto itself, i.e. e
k
f(e
k
).
To go further, we must impose more assumptions on X and f. First, we suppose that
X has the special property that the attaching maps of cells are all injective and hence
that the characteristic map D
k
X are injective. Next, we suppose that f sends each
k-cell either in to the k 1-skeleton or homeomorphically onto a k-cell. These conditions
are satised if X is a simplical complex and f is a simplicial map.
With these assumptions imposed, we deduce that L(f) ,= 0 then f sends some k-cell
to itself. By the Brouwer Fixed Point Theorem, we deduce that there must be a xed
point.
To make this argument rigorous, we would have to show that any continuous map
f : X X without xed points can be replaced up to homotopy by a simplical map
f

: X

without xed points. This is true (see 2.C of Hatcher), but the proof is
rather long and technical.
3.9.1 The Universal Coecient Theorem for Homology
The Universal Coecient Theorem for Homology allows H
q
(X; R) to be calculated from
H
q
(X; Z) and H
q1
(X; Z). We need some denitions before stating the formula.
Every Abelian group G can be t into a short exact sequence of the form
0 Z
r
A
Z
g
G 0 (6)
where Z
r
and Z
g
are (possibly innitely generated) free abelian groups on r and g gen-
erators respectively. We call (6) a free resolution of G. We think of the basis of Z
g
as
the generators of G and the basis of Z
r
as relations. The homomorphism A is dened in
terms of an g r matrix.
Example 29. For example, every nitely generated abelian group is isomorphic to
G = Z
f
Z
n
1
... Z
nt
for some choice f, n
1
,..., n
t
. The integer f is called the rank of G. This ts into a
short exact sequence
0 Z
t
A
Z
f
Z
t
G 0 (7)
where A = (0, A

) and A

is the diagonal matrix with entries n


1
, ..., n
t
.
Now let R be another abelian group. Given a free resolution (6) of G, we obtain a
map
R
r
A
R
R
g
52
where A
R
is given by the same integral matrix as A. We dene
Tor
0
Z
(G, R) := cok(A
R
) = R
g
/im(A
R
).
Tor
1
Z
(G, R) := ker(A
R
).
Equivalently, these are the homology groups of the chain complex 0 R
r
R
g
0.
The group Tor
0
Z
(G, R) is known as the tensor product of G and R and is more com-
monly denoted
G
Z
R := Tor
Z
(G, R).
If R is a ring, then both Tor
0
Z
(G, R) and Tor
1
Z
(G, R) are R-modules.
Theorem 3.42 (Universal Coecient Theorem). For any space X and coecient ring
R, there is a non-canonical isomorphism
H
q
(X; R) = Tor
0
Z
(H
q
(X), R) Tor
1
Z
(H
q1
(X), R).
Proof. Skipped. See Hatcher 3.A for a proof.
The moral of Theorem 3.42 is that homology groups with coecents in R contain no
more information than homology groups with integer coecients. Here are a couple of
easy consequences.
Corollary 3.43. If H
q
(X; Z) is free abelian in every degree, then H
q
(X; R) is a free
R-module of the same rank in each degree.
Proof. For a free abelian group G = Z
g
, we form a resolution
0 0 Z
g

=
G 0.
Thus Tor
0
Z
(Z
g
, R) = R
g
and Tor
1
Z
(Z
g
, R) = 0 for any R for any R. It follows that
H
q
(X; R) = Tor
0
Z
(H
q
(X), R) Tor
1
Z
(H
q1
(X), R) = H
q
(X)
Z
R = R
n
where n is the rank of H
q
(X).
Corollary 3.44. Suppose that H
q
(X; F) is nitely generated in all degrees. If F is a
eld, then
dim
F
(H
q
(X; F)) rank(H
q
(X; Z))
with equality if F has characteristic zero (e.g. F = Q, R, C).
Proof. For a nitely generated abelian group G, we have resolution (7) above. If F is a
eld then Tor
0
Z
(G, F) and Tor
1
Z
(G, F) are vector spaces of dimension f + t rank(A

F
)
and t rank(A

F
). Consequently
H
q
(X; F) = Tor
0
Z
(H
q
(X), F) Tor
1
Z
(H
q1
(X), F)
has dimension at least as large as the rank of H
q
(X) with equality if and only if rank(A

F
) =
t respectively. If F has characteristic zero, then equality holds.
53
3.10 Covering spaces and the transfer
A covering map f : X Y is a continuous map such that for every point p Y , there
is an open neighbourhood p U Y such that the preimage f
1
(U) is a disjoint union
of open sets V
i
X each mapping homeomorphically to U by f (we call U a trivializing
open set for the cover). We call X a covering space of Y . The V
i
are called sheets of
the cover. If Y is path connected, then the number of sheets is independent of p and U
and f : X Y is called an n-sheeted cover if the number of sheets over each point is
n.
Some examples
The map R S
1
dened by t e
2it
is a -sheeted cover.
The winding map w
d
: S
1
S
1
is a d-sheeted cover (for d > 1).
The quotient map S
n
RP
n
is a 2-sheeted cover.
Given a pair of coprime integers p, q Z, the lense space L
p,q
is the quotient space
S
3
/ where we view S
3
C
2
as the unit sphere and impose relation (z, w) =
(e
2i/p
z, e
2iq/p
z). Each equivalence class consists of a exactly p points and the
quotient map S
3
L
p,q
is a p-sheeted cover.
The basic property of covering spaces used for all applications in topology is the
following.
Lemma 3.45. Let f : X Y be a covering space and let :
q
Y be a singular
simplex. Let x X satisfy f(x) = (

0). Then there is a unique singular simplex :

q
X such that f = .
X
f

}
}
}
}
}
}
}
}
Y
Proof. This is proven in the course on homotopy theory. We sketch the argument.
If the image of lies in an trivializing open set U Y , then is obtained by composing
with the homeomorphism f
1
: U

= V
i
to the sheet V
i
containing x. This lift is clearly
unique.
For the general case, simply decompose
q
using iterated barycentric subdivision until
each subsimplex lands in a trivializing set U Y and then lift one simplex at a time.
A deck transformation of a covering space is a homeomorphism : X X such that
f = f .
X

A
A
A
A
A
A
A
X
f
.~
~
~
~
~
~
~
~
Y
54
The deck transformations form a group under composition that acts on X. The covering
space is called Galois if the deck transformation group acts freely and transitively on
bres. I.e. if for any two points x, x

f
1
(y) there is a unique deck transformation
such that (x) = x

. In this case, the preimages of points f


1
(y) are equal to the orbits
of a free group action on X.
In the examples above, the deck transformation groups Z acting by translation on R,
Z
d
acting by rotation on S
1
, Z
2
acting by the antipodal map on S
n
, and Z
p
acting on S
3
.
All three examples are Galois.
Theorem 3.46. Suppose that f : X Y is a n-sheeted covering map for nite n. Then
there exists a natural homomorphism

: H
q
(Y ; R) H
q
(X; R)
such that H
q
(f)

is multiplication by n on H
q
(Y ; R).
In particular, if R is a coecient ring containing
1
n
then

must be injective. If
furthermore the covering is Galois, then the image of

is
H
q
(Y ; R)

= H
q
(X; R)
G
with the subgroup of H
q
(X; R) of elements xed by the deck transformation group G.
Proof. The transfer map

is determined by a chain morphism : S(Y ) S(X) dened


on simplices :
q
Y to the sum
() =
n

i=1

i
where
1
, ...,
n
are the n lifts of (one for each sheet). It is not hard to see that =
so is a chain morphism. At the level of chains
f

() = f

(
n

i=1

i
) =
n

i=1
f
i
= n.
Thus H
q
(f) () is multiplication by n.
Now suppose that n is invertible in R. Then
1
n
f

is the identity on H
q
(Y ; R) so

must be injective (in fact the inclusion splits!).


Finally, suppose additionally that the covering is Galois with deck transformation
group G of order n. Dene a homomorphism

: H
q
(X; R) H
q
(X; R)
G
by

() =
1
n
n

i=1
g

()
If H
q
(X; R)
G
then

() = and thus
im(

) = H
q
(X; R)
G
.
Note also that n

. Since f

is surjective and n is a unit it follows that


im(

) = im(n

) = im(

) = im(

)

= H
q
(Y ; R).
55
4 Cohomology
Let R be a commutative ring and let M be a R-module. The dual module
M

:= Hom
R
(M, R)
is the set of R-module homomorphisms from M to R. M

is an R module under addition


and scalar multiplication of functions. There is a natural isomorphism
(
i
M
i
)

i
M

i
,
dened by the rule (
1
,
2
, ..., )(m
1
+m
2
+...) =

i
(m
i
). In particular, for free modules
we have
(
i
R)

i
R

i
R. (8)
More directly, (8) holds because a homomorphism out of a free module is specied by
listing where the free generators are sent.
Given a R-module homomorphism f : M N, dene the transpose
f

: N

which sends N

= Hom
R
(N, R) to f

() = f.
M
f

()=f

A
A
A
A
A
A
A
A
N

.~
~
~
~
~
~
~
R
Dualization is a contravariant functor from R-modules to R-modules. This means
that Id

M
= Id
M
and (g f)

= f

. The rst is obvious and the second follows from


associativity of composition:
(g f)

() = (g f) = ( g) f = f

(g

()) = (f

)()
M
f

()

B
B
B
B
B
B
B
B
N

.~
~
~
~
~
~
~
R
Given a chain complex of R-modules
... C
n+1

n+1
C
n
n
C
n1

n1
...
we form the dual chain complex (or co-chain complex)
... C
n+1

n+1
C
n
n
C
n1

n1
...
56
where C
n
= C

n
and
n
=

n
. Note that

n

n+1
=

n+1
= (
n+1

n
)

= 0

= 0
We can now dene cochains Z
n
:= ker(
n+1
), coboundaries B
n
:= im(
n
) and cohomology
H
n
:= Z
n
/B
n
.
Denition 19. The singular cohomology of a pair of spaces (X, A), denoted H
q
(X, A; R)
q 0, is the cohomology of the singular cochain complex
S
q1
(X, A; R) S
q
(X, A; R) S
q+1
(X, A; R) . . .
obtained by dualizing the singular chain complex of X.
More geometrically, we can understand a singular co-chain S
q
(X; R) as a function
that assigns a scalar to every singular simplex :
q
X. The pairing
S
q
(X; R) S
q
(X; R) R
is sometimes called integration, because it is an algebraic analogue of integrating a
dierential form over parametrized manifold.
For example,
A 0-cochain is simply a function (of sets) f : X R, since 0-simplices correspond
to points in X.
A 1-cochain assigns a scalar to every continuous path : [0, 1] X.
A 2-cochain assigns a scalar to every map :
2
X.
The analogue of Stokes Theorem follows just by denition. If S
q
(X : R) and
S
q1
(X; R) then
() = ()().
or in integral notation
_

=
_

.
For example, given a 1 simplex : [0, 1] X and 0-cochain f : X R, we have
_

f =
_

f = f((1) (0)) = f((1)) f((0)).


This example really illustrate why the boundary of a 1-simple requires signs: to recover
the Fundamental Theorem of Calculus.
An easy consequence is that this integration pairing descends to homology and coho-
mology.
57
Proposition 4.1. There is a natural pairing
H
q
(X, A; R) H
q
(X, A; R) R
dened by ([], []) () for representative (co)cycles and . This determines a
natural homomorphism
H
q
(X, A; R) H
q
(X, A; R)

. (9)
If H
q
(X, A; R) is free in all degrees, then (9) is an isomorphism.
Remark 5. More generally, singular cohomology can always be computed from integral
homology using the Universal Coecient Theorem for Cohomology. See Hatcher 3.1 for
details.
Proof. Choosing dierent representatives of the homology and cohomology classes, we get
( +)( +) = () + () + ()() + ()()
= () + ()() + () +(
2
)
= ().
establishing that (9) is a well-dened function, which is clearly a homomorphism.
We prove the second part only when R = F is a eld (the general case is an exercise).
Abbreviate S
q
(X, A; R) = S
q
. Recall that we have short exact sequences
0 Z
q
S
q
B
q1
0
0 B
q
Z
q
H
q
0.
These sequences split, giving rise to isomorphisms S
q
= Z
q
B

q1
and Z
q
= B
q
H

q
. We
have an isomorphism of chain complexes
. . .
S
q+1

S
q

S
q1

. . .
. . .
B
q+1
H

q+1
B

B
q
H

q
B

q1

B
q1
H

q1
B

q2
. . .
where

sends summands B
q
and H

q
to zero and send B

q1
isomorphically to B
q1
.
Dualizing get
. . .
S
q+1

S
q

S
q1

. . .
. . .
B

q+1
H

q+1
B

q
H

q
B

q1

q1
H

q1
B

q2

. . .
where

sends H

q
and B

q1
to zero and sends B

q
isomorphically to B

q
. Taking coho-
mology, we nd that
H
q
= H

q
= H

q
.
58
All of the main theorems about homology have versions in cohomology. The main
dierence is that arrows have to be reversed. We summarize:
H
q
() = H
q
(; R) is a contravariant functor for topological spaces to the category
of R-modules. Thus for every continuous map f : X Y , there is a homomorphism
of R-modules
H
q
(X)
H
q
(f)=f

H
q
(Y )
such that (f g)

= g

and (Id
X
)

= Id
H
q
(X;R)
.
If f, g : X Y are homotopic, then f

= g

.
A topological pair (X, A) gives rise to a long exact sequence of R-modules
H
q1
(A) H
q
(X, A) H
q
(X) H
q
(A) H
q+1
(A) ...
where the connecting homomorphism is dened using the same diagram chase as
before.
The excision isomorphism H
q
(X B, A B)

= H
q
(X, A) hold under the same
hypotheses as before.
H
0
(X) =

0
(X)
R is a product of copies of R indexed by the path components of
X.
H
0
(pt) = R and H
q
(pt) = 0 for q ,= 0. Dene

H
q
(X) to be the cokernel of the
natural map H
q
(pt) H
q
(X).


H
n
(S
n
)

= R and

H
q
(S
n
) = 0 for q ,= n. If f : S
n
S
n
is a degree d map, then f

is simply multiplication by d R.
For a cell-complex X, the cohomology of the dual of the cellular chain complex is
naturally isomorphic to the singular cohomology. In this case, where the boundary
maps are given by integer matrices, the co-boundary maps are simply the transposes
of those matrices.
Under the same hypotheses as for homology, there is a long exact Mayer-Vietoris
Sequence
H
q
(X) H
q
(A) H
q
(B) H
q
(A B) H
q+1
(X) . . . .
Example: RP
n
Recall that the cellular chain complex has the form
0 R
0
R
2
R ... R
2
R
0
R
if n is odd and
0 R
2
R
0
R ... R
2
R
0
R
if n is even. The cellular cochain complex is obtained by reversing arrows
0 R
0
R
2
R ... R
2
R
0
R
59
if n is odd and
0 R
2
R
0
R ... R
2
R
0
R
If R = Z
2
then the coboundary maps are all zero and we get
H
q
(RP
n
; Z
2
) =
_
Z
2
for q = 0, ..., n
0 otherwise.
If R = F if a eld of odd or zero characteristic, then the coboundary maps alternate
between zero and isomorphisms and we get
H
q
(RP
n
; F) =
_
F if q = 0 or q = n and n is odd
0 otherwise.
both of which agree with the homology groups (as they must by Proposition 4.1).
If R = Z then we obtain
H
q
(RP
n
; Z) =
_

_
Z if q = 0 or q = n and n is odd
Z
2
if 2 q n and is even
0 otherwise
Notice that the integer cohomology groups do not agree with the integer homology
groups in this case (for homology the Z
2
groups occur in odd degree). The general
situation for integer cohomology is as follows (see Hatcher 3.1 for proof).
Proposition 4.2. If the homology groups H
q
(X, A; Z) are nitely generated in all degrees,
then H
q
(X; Z) is isomorphic to a direct sum of the free part of H
q
(X; Z) with the torsion
part of H
q1
(X; Z)
4.1 The cup product
A graded R-algebra A

is a direct sum of R-modules


A

iZ
A
i
equipped with a multiplication A
i
A
j
A
i+j
which is distributive with respect to
addition and scalar multiplication.
(r
1
m
1
+r
2
m
2
) n = r
1
(m
1
n) +r
2
(m
2
n)
n (r
1
m
1
+r
2
m
2
) = r
1
(n m
1
) +r
2
(n m
2
).
The elements of A
i
A

are called homogeneous of degree d. We say A

is graded
commutative if the multiplication satises
m n = (1)
i+j
n m
for homogeneous elements m A
i
and n A
j
.
Examples:
60
A polynomial ring R[x] is a graded R-algebra if we give x some degree d. It is a free
module
R[x] =

i=0
Rx
i
where Rx
i
has degree di and multiplication is x
i
x
j
= x
i+j
. If d is even, then R[x] is
graded commutative.
More generally, the polynomial ring R[x
1
, ..., x
n
] with given degrees d
i
= deg(x
i
)
is a graded R-algebra. It is a free R-module with homogeneous generators, the
monomials x
i
1
1
...x
in
n
which have degree d
1
i
1
+... +d
n
i
n
. It is graded commutative if
the degrees d
i
are all even.
The exterior (or Grassmann) algebra (y
1
, ..., y
n
) is the R-algebra generated by
y
1
, ..., y
n
and identity 1 with multiplication
1. y
i
y
j
= (1)y
j
y
i
2. y
i
y
i
= 0.
It is a free R-module with 2
n
generators y
i
1
.... y
i
k
[1 i
1
< ... < i
k
n. It is
graded commutative if the y
i
all have odd degree.
For example, (y
1
, y
2
, y
3
) has free module generators
1, y
1
, y
2
, y
3
, y
1
y
2
, y
1
y
3
, y
2
y
3
, y
1
y
2
y
3
.
The goal of this section is to establish the following.
Theorem 4.3. There is a multiplication called the cup product that makes the direct sum
H

(X; R) :=

q=0
H
q
(X; R)
into a graded commutative, associative R-algebra for which H
q
(X; R) has degree q. There
is a multiplicative identity, denoted 1, which is represented by the 0-cocycle that sends
every 0-simplex to 1.
Example 30. Recall that for n > 1, the S
n
has cohomology H
q
(S
n
) = R if q = 0, n and
zero otherwise. If we select a generator x H
n
(S
n
), then x
2
H
2n
(S
n
) = 0 must be
zero. Thus there is an isomorphism of graded R-algebras
H

(S
n
; R)

= R[x]/(x
2
)
where we have taken the quotient of the polynomial ring R[x] by the ideal (x
2
) generated
by x
2
.
61
We will show later that
H

(CP
n
; R)

= R[x]/(x
n+1
)
where deg(x) = 2 and that
H

(RP
n
; Z
2
)

= Z
2
[y]/(y
n+1
)
where deg(y) = 1.
We begin by dening the cup product at the level of cochains.
S
p
(X, A; R) S
q
(X, A; R) S
p+q
(X, A; R), (, )
dened on a (p +q)-simplex by
( )() = ( [e
0
, ..., e
p
])( [e
p
, ..., e
p+q
])
where [e
0
, ..., e
p
] :
p

p+q
and [e
p
, ..., e
p+q
] :
q

p+q
are ane simplices.
The cup product is both associative and distributive with respect to addition (Home-
work).
Drawings ...
Lemma 4.4. The cup product satises a Leibnitz rule
( ) = + (1)
p

for S
p
(X) and S
q
(X; R).
Proof. For :
p+q+1
X we have
( )() = ( [e
0
, ..., e
p+1
])( [e
p+1
, ..., e
p+q+1
])
=
p+1

i=0
(1)
i
( [e
0
, ..., e
i
, ..., e
p+1
])( [e
p+1
, ..., e
p+q+1
])
in which the last term is (1)
p+1
( [e
0
, ..., e
p
])( [e
p+1
, ..., e
p+q+1
]), and
( )() =
p+q+1

i=p
(1)
ip
( [e
0
, ..., e
p
])( [e
p
, ..., e
i
, ..., e
p+q+1
])
in which the rst term is ( [e
0
, ..., e
p
])( [e
p+1
, ..., e
p+q+1
]).
Hence
( ) =
p+q+1

i=0
(1)
i
( )( [e
0
, ..., e
i
, ..., e
p+q+1
])
=
p

i=0
(1)
i
( [e
0
, ..., e
i
, ..., e
p+1
])( [e
p+1
, ..., e
p+q+1
])
+
p+q+1

i=p+1
(1)
i
( [e
0
, ..., e
p
])( [e
p
, ..., e
i
, ..., e
p+q+1
])
= + (1)
p

62
Observe that (??) implies that the product of two cocycles is a cocycle. Also, if the
product of a cocycle with a coboundary (in either order) is a coboundary, because
= ( ) = ( )
if = 0 and
= ( ) = ( )
if = 0. It follows that the cup product descends a map
: H
p
(X; R) H
q
(X; R) H
p+q
(X; R)
which is both associative and bilinear with respect to the R-module structure.
Thus the cup product makes the direct sum
H

(X; R) :=

q=0
H
q
(X; R)
into a graded, associative R-algebra. There is a multiplicative identity, denoted 1,
which is represented by the 0-cocycle that sends every 0-simplex to 1.
Theorem 4.5. Given H
p
(X; R) and H
q
(X; R), the cup product satises the
identity
= (1)
pq
. (10)
Sketch. The proof of this result is actually surprisingly dicult, because the identity (10)
does not hold for cochains. We will sketch the argument.
The idea of the argument is to construct a morphism of cochain complexes (i.e. a map

: S
q
(X) S
q
(X) such that

) such that
1. Satises

( ) = (1)
pq

()

(). for S
p
(X) and S
q
(X).
2. Determines the identity map on cohomology H(

) = Id
H

(X)
.
Given these two properties, if and are cocycles representing cohomology classes
[] and [] we have
[] [] = [ ]
= [

( )]
= [(1)
pq

()

()]
= (1)
pq
[

()] [

()]
= (1)
pq
[] []
thus proving the theorem.
63
Given a singular q-simplex :
q
X, dene
= [e
q
, e
q1
, ..., e
1
, e
0
]
by composing with the ane simplex that reverses the order of the vertices. Dene a
homomorphism

q
: S
q
(X) S
q
(X),
which sends a simplex to

q
() =
q

where
q
= (1)
q(q+1)/2
. Clearly
q
is an isomorphism, because (
q
)
2
is the identity.
Claim: The
q
combine to form a chain morphism : S

(X) S

(X).
Proof.
() = (
q
[e
n
, ..., e
0
]) =
q

i=0
(1)
i

q
[e
q
, ..., e
qi
, ..., e
0
]
while
() = (
n

q=0
(1)
i
[e
0
, ..., e
i
, ..., e
q
])
=
n

q=0
(1)
i

q1
[e
q
, ..., e
i
, ..., e
0
]).
Thus
=
and it is just a matter of checking that the sign
(1)
i
(1)
qi

q1

q
= (1)
q
(1)
q(q1)/2
(1)
q(q+1)/2
= 1.
Since is a chain morphism, the transpose

: S

(X) S

(X) is a cochain
morphism, because

= ( )

= ( )

.
Claim: The cochain map

induces the identity map on cohomology.


Proof. This is the part we will skip. Topologically this is intuitively plausible, because we
are simply reparametrizing simplices, without changing their images. The proof depends
on constructing a chain homotopy P : S
q
(X) S
q+1
(X) between and the identity map:
P +P = Id
S(X)
.
64
The construction is quite similar to the one used to prove homotopy invariance, but altered
to deal with the reordering of vertices. The transpose of P is a cochain homotopy between

and the identity, because

Id
S

(X)
= ( Id
S(X)
)

= (P +P)

= P

+P

which thus induces identity on cohomology.


Claim: For S
p
(X) and S
q
(X), we have

( ) = (1)
pq

()

().
Proof of claim. Let be a p +q simplex. Then

( )() = ( )(())
= ( )(
p+q
[e
p+q
, ..., e
0
])
=
p+q
( [e
p+q
, ..., e
p
])( [e
p
, ..., e
0
])
while
(

()

())() =

()( [e
0
, ..., e
p
])

()( [e
p
, ..., e
p+q
])
=
p

q
( [e
p
, ..., e
0
])( [e
p+q
, ..., e
p
])
=
p

q
( [e
p+q
, ..., e
p
])( [e
p
, ..., e
0
]).
It remains to show the signs work out :
p+q
=
p

q
(1)
pq
.
4.2 The K unneth formula
Given R-modules M and N, the tensor product M
R
N = M N is aa R-module
possessing the following universal property. There is a canonical bilinear map t : MN
M N such that for any R-module P and any bilinear map b : M N P, there exists
a unique homomorphism
b
such that b =
b
t:
M N
t

M
M
M
M
M
M
M
M
M
M
M
M
M
R
N

P
Of course, this works backwards as well. Given a homomorphism : M
R
N P
the composition t is a bilinear map from M N to P. Thus we have a one-to-one
correspondence
bilinear maps b : M N P
1:1
homomorphisms : M
R
N P
The tensor product of two modules always exists and is unique (up to canonical isomor-
phism). We use notation t(m, n) = mn. It is a little complicated to describe in general,
but for free modules it is very easy to understand.
65
Proposition 4.6. If M =

iI
Rx
i
and N =

jJ
Ry
j
are free modules with generators
x
i
and y
j
respectively. Then
M
R
N =

(i,j)IJ
R(x
i
y
j
)
is a free module with generators x
i
y
j
.
Proof. Dene the map
M N

(i,j)IJ
R(x
i
y
j
)
by
t(

i
r
i
x
i
,

j
r
j
y
j
) =

i,j
r
i
r
j
x
i
y
j
.
The map t is clearly bilinear. Given another bilinear map b : M N P, dene the
homomorphism

b
:

(i,j)IJ
R(x
i
y
j
) P
which sends generators

b
(x
i
y
j
) = b(x
i
, y
j
).
It is easy to see that b =
b
t and that
b
is the unique homomorphism satisfying this
property.
Now we connect this idea with topology. Given two spaces, we may form the product
space X Y which comes equipped with projection maps
X Y

G
G
G
G
G
G
G
G
G

X
.v
v
v
v
v
v
v
v
v
X Y
These can be used to dene a bilinear map
H
p
(X) H
q
(Y ) H
p+q
(X Y ), (, )

X
()

Y
().
By the universal property, this determines a homomorphism
: H
p
(X) H
q
(Y ) H
p+q
(X Y ). (11)
Theorem 4.7. If both H
p
(X; R) and H
q
(Y ; R) are free R-modules for all degrees p and
q, then there is an isomorphism
H
n
(X Y ; R)

=

p+q=n
H
p
(X; R)
R
H
q
(Y ; R)
dened by adding up the natural homomorphisms 11 .
66
Proof. A proof in the case when X and Y are cell complexes is found in Hatcher 3.2.
Example 31. Consider the 3-manifold S
1

g
where
g
is a genus g surface. Then
H
0
(S
1

g
) = H
0
(S
1
) H
0
(
g
)

= R.
H
1
(S
1

g
) = H
0
(S
1
) H
1
(
g
) +H
1
(S
1
) H
0
(
g
)

= R
2g
R = R
2g+1
H
2
(S
1

g
) = H
0
(S
1
) H
2
(
g
) +H
1
(S
1
) H
1
(
g
)

= R R
2g
= R
2g+1
H
3
(S
1

g
) = H
1
(S
1
) H
2
(
g
)

= R.
In fact, the Kunneth Theorem also determines the multiplicative structure of H

(X
Y ).
Given two graded associative R-algebras, A

and B

, the tensor algebra is (A


R
B)

=
A

R
B

is a the graded, associative algebra with graded modules


(A
R
B)
n
:=
p+q=n
A
p

R
B
q
and multiplication
(a
1
b
1
) (a
2
b
2
) = (1)
deg(a
2
) deg(b
1
)
(a
1
a
2
) (b
1
b
2
).
Observe that if A

and B

are graded commutative, then so is A

, because if a
i
A
p
i
and b
i
B
q
i
then
(a
1
b
1
) (a
2
b
2
) = (1)
p
2
q
1
(a
1
a
2
) (b
1
b
2
)
= (1)
p
2
q
1
((1)
p
1
p
2
a
2
a
1
) ((1)
q
1
q
2
b
2
b
1
)
= (1)
p
2
q
1
+p
1
p
2
+q
1
q
2
(a
2
a
1
) (b
2
b
1
)
= (1)
p
2
q
1
+p
1
p
2
+q
1
q
2
+p
1
q
2
(a
1
b
1
) (a
2
b
2
)
= (1)
(p
1
+q
1
)(p
2
+q
2
)
(a
1
b
1
) (a
2
b
2
)
Example 32. If x
1
, ..., x
n
have even degrees, then
R[x
1
] ... R[x
n
] = R[x
1
, ..., x
n
].
Example 33. If y
1
, ..., y
n
have odd degrees, then
(y
1
) .... (y
n
) = (y
1
, ..., y
n
).
.
Theorem 4.8 (Kunneth formula). The Kunneth isomorphisms described in ... combine
to form an isomorphism of graded R-algebras
H

(X Y ; R)

= H

(X; R)
R
H

(Y ; R).
67
Proof. Recall that if
i
H
p
i
(X; R) and
j
H
p
j
(Y ; R) then the Kunneth homomor-
phism : H
p
i
(X) H
q
i
(Y ) H
p
i
+q
i
(X Y ) sends
i

j
to

1
(
i
)
2
(
j
). Thus
(
1

1
) (
2

2
) =
a
(
1
)

b
(
1
)

a
(
2
)

b
(
2
)
=
a
(
1
)

a
(
2
)

b
(
1
)

b
(
2
)
=
a
(
1

2
)

b
(
1

2
)
= ((
1

2
) (
1

2
))
= ((
1

1
) (
2

2
))
This extends by linearity to an isomorphism of graded R-algebras.
Example 34. for integers k
1
, ..., k
n
Z, the product of spheres S
k
1
... S
kn
has
cohomology ring
H

(S
k
1
... S
kn
)

= (x
1
, ..., x
k
)
where the generator x
i
has degree k
i
.
4.3 Manifolds and orientations
An n-manifold (a.k.a. manifold of dimension n) is a Hausdor topological space M
which is locally Euclidean, meaning that for each point p M there exists an open
neighbourhood p U M such that U is homeomorphic to R
n
. We say that M is closed
if it is compact. We assume for simplicity that M is path connected.
Examples:
Euclidean space R
n
is an n-manifold that is not closed.
The sphere S
n
is a closed n-manifold.
The orientable and non-orientable surfaces
g
and N
g
are closed 2-manifolds.
Real projective space RP
n
is a closed n-manifold (it looks locally like R
n
).
Complex projective space CP
n
is a closed 2n-manifold (it looks locally like C
n

=
R
2n
).
The product on a m-manifold with an n-manifold is an (m+n)-manifold, by taking
products of euclidean neighbourhoods.
Using excision, we have isomorphisms
H
n
(M, M p) = H
n
(U, U p)

= H
n
(R
n
, R
n
p)

= Z.
A local orientation of M is a choice of generator
p
H
n
(M, M p).
68
Given a Euclidean neighbourhood R
n

= U M, then for any two points p, q U we
can nd a compact set B U corresponding to a closed ball in R
n
and containing p and
q. We have isomomorphisms
H
n
(U, U p)

=
H
n
(U, U B)

=
H
n
(U, U q) (12)
The local orientations at p and q are called consistent if the isomorphism (12) sends
one to the other. A (global) orientation of M is a set
p
[ p M of local orientations
for every point in M, which is consistent with respect to every Euclidean neighbourhood.
If an orientation exists, we say M is orientable. Otherwise M is non-orientable.
The concept of a global orientation can be understood more geometrically as follows.
Consider the set

M consisting of ordered pairs (p,
p
) where p M and
p
is one of the
two generators of H
n
(M, M p). We have a forgetful function (of sets)
:

M M
which sends (p,
p
) to p and the pre-image of any point p M is the two possible
orientations at that point. We topologies

M by partitioning the pre-image of a Euclidean
neighbourhood U M according to consistent orientations, an having each partition map
homeomorphically onto U. This makes :

M M into a 2-fold covering space, called
the orientation cover. Notice that

M is a manifold which is orientable because it comes
equipped with a global orientation by construction.
A global orientation for M is equivalent to a continuous map s : M

M such that
s = Id
M
. We call s section of the covering map .
If an orientation s exists, then the map
s : M

M,
which takes the opposite value at each point is also an orientation. In this case covering
space decomposes into two copies of M

M

= M HM
each mapping homeomorphically to M via .
If no orientation of M exists, then some loop in M lifts (via the simplex lifting prop-
erty) to a non-closed path in

M from which it follows that

M is path connected.
Example 35. A loop that winds once around RP
2
= N
0
lifts to a path connecting
antipodal points on S
2
=
0
.
Recall that when we studied spheres, a global orientation was dened to be a generator
of H
n
(S
n
), which determines local orientations at each point p S
n
via the homomor-
phism
H
n
(S
n
)

=
H
n
(S
n
, S
n
p)
It turns out that for closed manifolds, the same idea works.
69
Theorem 4.9. Let M be a n-manifold. Then H
q
(M; R) = 0 for q > n. If M is compact,
then
H
n
(M; Z) =
_
Z if M is orientable
0 if M is non-orientable
If M is orientable, we call a generator of H
n
(M; Z) an orientation class.
Proof. Sketch: The proof uses a version of Mayer-Vietoris that I havent covered yet. Let
A and B be closed subsets of M, then we have a short exact sequence
H
q+1
(M, M(AB)) H
q
(M, M(AB)) H
q
(M, MA)H
q
(M, MB) H
q
(M, MAB) ...
Given a closed ball B in M contained in a Euclidean neighbourhood U, then by excision
we have
H
q
(M, M B; Z)

= H
q
(B, B; Z) =
_
Z if q = n
0 if q > n
Now choose a collection of balls B
i
such that their interiors cover M. Since M is compact,
we may choose this covering to be nite B
1
, ..., B
N
. We use the Mayer-Vietoris sequence
and induction to show that
H
q
(M, M (
N
i=1
B
i
)) = H
q
(M)
_
Z if q = n
0 if q > n
This gets a bit tricky, because we have to keep track of the intersections between balls.
One must prove that the balls can be chosen to be good in the sense that any intersection
of B
i
1
... B
in
is also homeomorphic to a ball. In R
n
this holds by convexity, but for
general manifolds it takes some care.
Examples:
The spheres S
n
, the orientable surfaces
g
and CP
n
are all orientable.
The non-orientable surface N
g
have orientation cover
g
.
RP
n
is orientable when n is odd and non-orientable when n is even.
If f : M N is a continuous map between two oriented, closed n-manifolds, it
determines a group homomorphism
f

: H
n
(M; Z) H
n
(N; Z)
which by ... is a homomorphism from Z to Z and thus corresponds to multiplication by an
integer d Z. We call d the degree of f. By a similar argument as worked for spheres,
one can calculate the degree of d by adding up local degrees of a nite set of points f
1
(p)
for regular value p N.
70
More generally, one can dene local/global orientations of a manifold M with respect
to a coecient ring R; that is, a choice of generator
p
H
n
(M, Mp; R) = R. The most
interesting case is when R = Z
2
, because then there is only one choice of generator at
each point and choices have to be consistent. It follows that every manifold is orientable
with respect to Z
2
and we get the following result.
Theorem 4.10. If M is closed n-manifold, then H
n
(M; Z
2
)

= Z
2
.
For instance, we see this with H
n
(RP
n
; Z
2
) = Z
2
and H
2
(N
g
; Z
2
) = Z
2
.
4.4 The cap product and Poincare Duality
The cap product is a bilinear map that takes as input a singular n-chain and a singular
k-cochain, and outputs a n k-chain (we assume n k):
: S
n
(X; R) S
k
(X; R) S
nk
(X; R).
Given a n-simplex :
n
X and a cochain S
k
(X; R), the cup product is
dened
= ( [e
0
, ..., e
k
]) [e
k
, ..., e
n
]
which extends by linearity to any n-chain.
The cap product is dual to the cup product in the sense that if S
nk
(X; R) is a
cochain, then
( ) = ( )(),
so that the homomorphism
( ) : S
nk
(X; R) S
n
(X; R)
is the transpose of the linear map
( ) : S
n
(X; R) S
nk
(X; R).
Theorem 4.11. The cap product determines a bilinear map (also called the cap product)
: H
n
(X; R) H
k
(X; R) H
nk
(X; R)
by the rule
[] [] = [ ]. (13)
Proof. We begin by proving Leibnitz rule:
Lemma 4.12. If S
n
(X) and S
k
(X) then
( ) = (1)
k
( )
71
Proof. It is enough to consider case that = is a singular n-simplex. The result is
veried by direct calculation.
( ) = (( [e
0
, ..., e
k
]) [e
k
, ..., e
n
])
=
n

i=k
(1)
ik
(( [e
0
, ..., e
k
]) [e
k
, ..., e
i
, ..., e
n
])
() = ()( [e
0
, ..., e
k+1
]) [e
k+1
, ..., e
n
]
=
k+1

i=0
(1)
i
( [e
0
, ..., e
i
, ..., e
k+1
]) [e
k+1
, ..., e
n
]
where the last term equals, (up to sign) the rst term of the previous sum. Finally,
() = (
n

i=0
(1)
i
[e
0
, ..., e
i
, ..., e
n
])
=
k

i=0
(1)
i
( [e
0
, ..., e
i
, ..., e
k+1
]) [e
k+1
, ..., e
n
]
+
n

k+1
(1)
i
( [e
0
, ..., e
k
]) [e
k
, ..., e
i
, ..., e
n
]
From this it follows that if = 0 and = 0, then
( ) = = 0,
so the cap product appearing in (13) is a cycle that represents a homology class. To
see that it is well-dened, we use dierent representatives [+] = [] and [+] = []
and check
[( +) ( +)] = [ + + + () ()]
= [ +( ) ( ) ( )]
= [ +
_

_
]
= [ ].
We are now able to state the Poincare Duality Theorem.
Theorem 4.13. Let M be a closed, R-oriented n-manifold M with orientation class
[M] H
n
(M; R). Cap product with respect to [M] denes an isomorphism
[M] ( ) : H
k
(M; R)

=
H
nk
(M; R).
72
Proof. We unfortunately dont have time to prove this. The argument is similar to the
one constructing the orientation class in that it begins with a local result, that leads to
a global result via Mayer-Vietoris and induction with respect to a nite cover of M by
balls. To do the proof properly, would require introducing the notion of cohomology
with compact supports.
Example 36. Recall that if H
q
(X; R) is a free R-module in all degrees, then there is a
natural isomorphism
H
q
(X; R)

= H
q
(X; R).
Combined with Poincare duality, this implies for R-oriented closed n-manifold M, that
H
q
(M; R) and H
nq
(M; R) are dual modules and that H
q
(M; R) and H
nq
(M; R) are
dual modules, which in particular means that they have the same rank. This explains the
word duality in Poincare duality.
Example 37. In case the homology groups are not all free, Theorem 4.13 still works. For
odd n, the projective space RP
n
is orientable and we have
H
k
(RP
n
; Z)

= H
nk
(RP
n
; Z)

=
_
Z
2
if 0 < k < n is odd (thus n k is even)
0 0 < k < n is even (thus n k is odd)
The duality between H
k
(M; R) and H
nk
(M; R) is better understood using the cup
product.
Corollary 4.14. If M is closed and R-oriented, and H
q
(X; R) is a free R-module in all
degrees, then the bilinear pairing
H
k
(M; R) H
nk
(M; R) R
sending (, ) to ([M]) is the duality pairing and thus determines a natural isomor-
phism
H
k
(M; R) = (H
nk
(M; R))

and vice-versa.
Proof. This follows immediately from Theorem ... and Theorem 4.13 from the identity
( )([M]) = ([M] ).
Example 38. Suppose X = (S
1
)
n
is a product of n circles. The cohomology ring is
H

(X; R) = (x
1
, ..., x
n
).
In this case, the generator in top degree is x
1
... x
n
and the duality pairing is between

k
(x
1
, ..., x
n
) and
nk
(x
1
, ..., x
n
), each of which has rank
_
n
k
_
=
_
n
nk
_
.
73
Example 39. Recall H
q
(RP
n
) = Z
2
in degrees 0 q n. Denote by y the generator of
H
1
(RP; Z
2
). The sequence of inclusions
RP
1
RP
2
RP
3
...
determines a sequence of surjections of cohomology rings
H

(RP
1
; Z
2
) H

(RP
2
; Z
2
) H

(RP
3
; Z
2
) ...
and is an isomorphism in degrees q less than the dimension. For n = 2, Poincare duality
implies that the generator y H
1
(RP
2
; Z
2
) must satisfy
(y y)([RP
2
]) = 1 ,= 0
so y
2
= y y is the generator of H
2
(RP
2
; Z
2
). Since ... are isomorphisms in degree 1, y
2
must be the generator of H
2
(RP
n
; Z
2
) for all n. Similarly, when n = 3 Poincare duality
implies that y
3
:= y y
2
must be a generator of H
3
(RP
n
; Z
2
) and thus this holds for all
n. Proceeding inductively in this manner, we prove that
H

(RP
n
; Z
2
)

= Z
2
[y]/(y
n+1
).
A similar argument works for to prove that
H

(CP
n
; R)

= R[x]/(x
n+1
)
for deg(x) = 2.
Example 40. For the genus g orientable surface
g
, the cup product boils down to the
pairing
H
1
(
g
; R) H
1
(
g
; R) H
2
(
g
; R)

= R
Because the cup product is graded commutative, the pairing above is anti-symmetric. S
74

Das könnte Ihnen auch gefallen