Are you sure?
This action might not be possible to undo. Are you sure you want to continue?
FOR ADVANCED STUDIES
Trieste
U. Bruzzo
INTRODUCTION TO
ALGEBRAIC TOPOLOGY AND
ALGEBRAIC GEOMETRY
Notes of a course delivered during the academic year 2002/2003
La ﬁlosoﬁa `e scritta in questo grandissimo libro che continuamente
ci sta aperto innanzi a gli occhi (io dico l’universo), ma non si pu`o
intendere se prima non si impara a intender la lingua, e conoscer i
caratteri, ne’ quali `e scritto. Egli `e scritto in lingua matematica, e
i caratteri son triangoli, cerchi, ed altre ﬁgure geometriche, senza
i quali mezi `e impossibile a intenderne umanamente parola; senza
questi `e un aggirarsi vanamente per un oscuro laberinto.
Galileo Galilei (from “Il Saggiatore”)
i
Preface
These notes assemble the contents of the introductory courses I have been giving at
SISSA since 1995/96. Originally the course was intended as introduction to (complex)
algebraic geometry for students with an education in theoretical physics, to help them to
master the basic algebraic geometric tools necessary for doing research in algebraically
integrable systems and in the geometry of quantum ﬁeld theory and string theory. This
motivation still transpires from the chapters in the second part of these notes.
The ﬁrst part on the contrary is a brief but rather systematic introduction to two
topics, singular homology (Chapter 2) and sheaf theory, including their cohomology
(Chapter 3). Chapter 1 assembles some basics fact in homological algebra and develops
the ﬁrst rudiments of de Rham cohomology, with the aim of providing an example to
the various abstract constructions.
Chapter 4 is an introduction to spectral sequences, a rather intricate but very power
ful computation tool. The examples provided here are from sheaf theory but this com
putational techniques is also very useful in algebraic topology.
I thank all my colleagues and students, in Trieste and Genova and other locations,
who have helped me to clarify some issues related to these notes, or have pointed out
mistakes. In this connection special thanks are due to Fabio Pioli. Most of Chapter 3 is
an adaptation of material taken from [2]. I thank my friends and collaborators Claudio
Bartocci and Daniel Hern´andez Ruip´erez for granting permission to use that material.
I thank Lothar G¨ottsche for useful suggestions and for pointing out an error and the
students of the 2002/2003 course for their interest and constant feedback.
Genova, 4 December 2004
Contents
Part 1. Algebraic Topology 1
Chapter 1. Introductory material 3
1. Elements of homological algebra 3
2. De Rham cohomology 7
3. MayerVietoris sequence in de Rham cohomology 10
4. Elementary homotopy theory 11
Chapter 2. Singular homology theory 17
1. Singular homology 17
2. Relative homology 25
3. The MayerVietoris sequence 28
4. Excision 32
Chapter 3. Introduction to sheaves and their cohomology 37
1. Presheaves and sheaves 37
2. Cohomology of sheaves 43
Chapter 4. Spectral sequences 53
1. Filtered complexes 53
2. The spectral sequence of a ﬁltered complex 54
3. The bidegree and the ﬁveterm sequence 58
4. The spectral sequences associated with a double complex 59
5. Some applications 62
Part 2. Introduction to algebraic geometry 67
Chapter 5. Complex manifolds and vector bundles 69
1. Basic deﬁnitions and examples 69
2. Some properties of complex manifolds 72
3. Dolbeault cohomology 73
4. Holomorphic vector bundles 73
5. Chern class of line bundles 77
6. Chern classes of vector bundles 79
7. KodairaSerre duality 81
8. Connections 82
iii
iv CONTENTS
Chapter 6. Divisors 87
1. Divisors on Riemann surfaces 87
2. Divisors on higherdimensional manifolds 94
3. Linear systems 95
4. The adjunction formula 97
Chapter 7. Algebraic curves I 101
1. The Kodaira embedding 101
2. RiemannRoch theorem 104
3. Some general results about algebraic curves 105
Chapter 8. Algebraic curves II 111
1. The Jacobian variety 111
2. Elliptic curves 116
3. Nodal curves 120
Bibliography 125
Part 1
Algebraic Topology
CHAPTER 1
Introductory material
The aim of the ﬁrst part of these notes is to introduce the student to the basics of
algebraic topology, especially the singular homology of topological spaces. The future
developments we have in mind are the applications to algebraic geometry, but also
students interested in modern theoretical physics may ﬁnd here useful material (e.g.,
the theory of spectral sequences).
As its name suggests, the basic idea in algebraic topology is to translate problems
in topology into algebraic ones, hopefully easier to deal with.
In this chapter we give some very basic notions in homological algebra and then
introduce the fundamental group of a topological space. De Rham cohomology is in
troduced as a ﬁrst example of a cohomology theory, and is homotopic invariance is
proved.
1. Elements of homological algebra
1.1. Exact sequences of modules. Let R be a ring, and let M, M
, M
be
Rmodules. We say that two Rmodule morphisms i : M
→ M, p: M → M
form an
exact sequence of Rmodules, and write
0 → M
j
−−→M
j
−−→M
→ 0 ,
if i is injective, p is surjective, and ker p = Imi.
A morphism of exact sequences is a commutative diagram
0 −−−−→ M
−−−−→ M −−−−→ M
−−−−→ 0
0 −−−−→ N
−−−−→ N −−−−→ N
−−−−→ 0
of Rmodule morphisms whose rows are exact.
1.2. Diﬀerential complexes. Let R be a ring, and M an Rmodule.
Definition 1.1. A diﬀerential on M is a morphism d: M → M of Rmodules such
that d
2
≡ d ◦ d = 0. The pair (M, d) is called a diﬀerential module.
The elements of the spaces M, Z(M, d) ≡ ker d and B(M, d) ≡ Imd are called
cochains, cocycles and coboundaries of (M, d), respectively. The condition d
2
= 0 implies
3
4 1. INTRODUCTORY MATERIAL
that B(M, d) ⊂ Z(M, d), and the Rmodule
H(M, d) = Z(M, d)/B(M, d)
is called the cohomology group of the diﬀerential module (M, d). We shall often write
Z(M), B(M) and H(M), omitting the diﬀerential d when there is no risk of confusion.
Let (M, d) and (M
, d
) be diﬀerential Rmodules.
Definition 1.2. A morphism of diﬀerential modules is a morphism f : M → M
of
Rmodules which commutes with the diﬀerentials, f ◦ d
= d ◦ f.
A morphism of diﬀerential modules maps cocycles to cocycles and coboundaries to
coboundaries, thus inducing a morphism H(f): H(M) → H(M
).
Proposition 1.3. Let 0 → M
j
−−→M
j
−−→M
→ 0 be an exact sequence of dif
ferential Rmodules. There exists a morphism δ : H(M
) → H(M
) (called connecting
morphism) and an exact triangle of cohomology
H(M)
1(j)
H(M
)
δ
.t
t
t
t
t
t
t
t
t
H(M
)
1(j)
Proof. The construction of δ is as follows: let ξ
∈ H(M
) and let m
be a
cocycle whose class is ξ
. If m is an element of M such that p(m) = m
, we have
p(d(m)) = d(m
) = 0 and then d(m) = i(m
) for some m
∈ M
which is a cocycle.
Now, the cocycle m
deﬁnes a cohomology class δ(ξ
) in H(M
), which is independent of
the choices we have made, thus deﬁning a morphism δ : H(M
) → H(M
). One proves
by direct computation that the triangle is exact.
The above results can be translated to the setting of complexes of Rmodules.
Definition 1.4. A complex of Rmodules is a diﬀerential Rmodule (M
•
, d) which
is Zgraded, M
•
=
¸
a∈Z
M
a
, and whose diﬀerential fulﬁlls d(M
a
) ⊂ M
a+1
for every
n ∈ Z.
We shall usually write a complex of Rmodules in the more pictorial form
. . .
o
n−2
−−→M
a−1
o
n−1
−−→M
a
o
n
−−→M
a+1
o
n+1
−−→ . . .
For a complex M
•
the cocycle and coboundary modules and the cohomology group
split as direct sums of terms Z
a
(M
•
) = ker d
a
, B
a
(M
•
) = Imd
a−1
and H
a
(M
•
) =
Z
a
(M
•
)/B
a
(M
•
) respectively. The groups H
a
(M
•
) are called the cohomology groups
of the complex M
•
.
1. HOMOLOGICAL ALGEBRA 5
Definition 1.5. A morphism of complexes of Rmodules f : N
•
→ M
•
is a collec
tion of morphisms ¦f
a
: N
a
→ M
a
[ n ∈ Z¦, such that the following diagram commutes:
M
a
;
n
−−−−→ N
a
o
o
M
a+1
;
n+1
−−−−→ N
a+1
.
For complexes, Proposition 1.3 takes the following form:
Proposition 1.6. Let 0 → N
•
j
−−→M
•
j
−−→P
•
→ 0 be an exact sequence of com
plexes of Rmodules. There exist connecting morphisms δ
a
: H
a
(P
•
) → H
a+1
(N
•
) and
a long exact sequence of cohomology
. . .
δ
n−1
−−→H
a
(N
•
)
1(j)
−−→H
a
(M
•
)
1(j)
−−→H
a
(P
•
)
δ
n
−−→
δ
n
−−→H
a+1
(N
•
)
1(j)
−−→H
a+1
(M
•
)
1(j)
−−→H
a+1
(P
•
)
δ
n+1
−−→ . . .
Proof. The connecting morphism δ : H
•
(P
•
) → H
•
(N
•
) deﬁned in Proposition
1.3 splits into morphisms δ
a
: H
a
(P
•
) → H
a+1
(N
•
) (indeed the connecting morphism
increases the degree by one) and the long exact sequence of the statement is obtained
by developing the exact triangle of cohomology introduced in Proposition 1.3.
1.3. Homotopies. Diﬀerent (i.e., nonisomorphic) complexes may nevertheless
have isomorphic cohomologies. A suﬃcient conditions for this to hold is that the two
complexes are homotopic. While this condition is not necessary, in practice the (by far)
commonest way to prove the isomorphism between two cohomologies is to exhibit a
homototopy between the corresponding complexes.
Definition 1.7. Given two complexes of Rmodules, (M
•
, d) and (N
•
, d
), and two
morphisms of complexes, f, g : M
•
→ N
•
, a homotopy between f and g is a morphism
K: N
•
→ M
•−1
(i.e., for every k, a morphism K: N
I
→ M
I−1
) such that d
◦ K +K◦
d = f −g.
The situation is depicted in the following commutative diagram.
. . .
M
I−1
·
o
M
I
;
j
·
o
1
.w
w
w
w
w
w
w
w
M
I+1
·
1
.w
w
w
w
w
w
w
w
. . .
. . .
N
I−1
o
N
I
o
N
I+1 . . .
Proposition 1.8. If there is a homotopy between f and g, then H(f) = H(g),
namely, homotopic morphisms induce the same morphism in cohomology.
6 1. INTRODUCTORY MATERIAL
Proof. Let ξ = [m] ∈ H
I
(M
•
, d). Then
H(f)(ξ) = [f(m)] = [g(m)] + [d
(K(m))] + [K(dm)] = [g(m)] = H(g)(ξ)
since dm = 0, [d
(K(m))] = 0.
Definition 1.9. Two complexes of Rmodules, (M
•
, d) and (N
•
, d
), are said to
be homotopically equivalent (or homotopic) if there exist morphisms f : M
•
→ N
•
,
g : N
•
→ M
•
, such that:
f ◦ g : N
•
→ N
•
is homotopic to the identity map id
.
;
g ◦ f : M
•
→ M
•
is homotopic to the identity map id
A
.
Corollary 1.10. Two homotopic complexes have isomorphic cohomologies.
Proof. We use the notation of the previous Deﬁnition. One has
H(f) ◦ H(g) = H(f ◦ g) = H(id
.
) = id
1
(
.)
H(g) ◦ H(f) = H(g ◦ f) = H(id
A
) = id
1
(
A)
so that both H(f) and H(g) are isomorphism.
Definition 1.11. A homotopy of a complex of Rmodules (M
•
, d) is a homotopy
between the identity morphism on M, and the zero morphism; more explicitly, it is a
morphism K: M
•
→ M
•−1
such that d ◦ K +K ◦ d = id
A
.
Proposition 1.12. If a complex of Rmodules (M
•
, d) admits a homotopy, then it is
exact (i.e., all its cohomology groups vanish; one also says that the complex is acyclic).
Proof. One could use the previous deﬁnitions and results to yield a proof, but it
is easier to note that if m ∈ M
I
is a cocycle (so that dm = 0), then
d(K(m)) = m−K(dm) = m
so that m is also a coboundary.
Remark 1.13. More generally, one can state that if a homotopy K: M
I
→ M
I−1
exists for k ≥ k
0
, then H
I
(M, d) = 0 for k ≥ k
0
. In the case of complexes bounded
below zero (i.e., M = ⊕
I∈N
M
I
) often a homotopy is deﬁned only for k ≥ 1, and it
may happen that H
0
(M, d) = 0. Examples of such situations will be given later in this
chapter.
Remark 1.14. One might as well deﬁne a homotopy by requiring d
◦K−K◦d = . . . ;
the reader may easily check that this change of sign is immaterial.
2. DE RHAM COHOMOLOGY 7
2. De Rham cohomology
As an example of a cohomology theory we may consider the de Rham cohomology
of a diﬀerentiable manifold X. Let Ω
I
(X) be the vector space of diﬀerential kforms
on X, and let d: Ω
I
(X) → Ω
I+1
(X) be the exterior diﬀerential. Then (Ω
•
(X), d) is
a diﬀerential complex of Rvector spaces (the de Rham complex), whose cohomology
groups are denoted H
I
o1
(X) and are called the de Rham cohomology groups of X. Since
Ω
I
(X) = 0 for k > n and k < 0, the groups H
I
o1
(X) vanish for k > n and k < 0.
Moreover, since ker[d: Ω
0
(X) → Ω
1
(X)] is formed by the locally constant functions on
X, we have H
0
o1
(X) = R
C
, where C is the number of connected components of X.
If f : X → Y is a smooth morphism of diﬀerentiable manifolds, the pullback morph
ism f
∗
: Ω
I
(Y ) → Ω
I
(X) commutes with the exterior diﬀerential, thus giving rise to a
morphism of diﬀerential complexes (Ω
•
(Y ), d) → (Ω
•
(X), d)); the corresponding morph
ism H(f): H
•
o1
(Y ) → H
•
o1
(X) is usually denoted f

.
We may easily compute the cohomology of the Euclidean spaces R
a
. Of course one
has H
0
o1
(R
a
) = ker[d : C
∞
(R
a
) → Ω
1
(R
a
)] = R.
Proposition 1.1. (Poincar´e lemma) H
I
o1
(R
a
) = 0 for k > 0.
Proof. We deﬁne a linear operator K: Ω
I
(R
a
) → Ω
I−1
(R
a
) by letting, for any
kform ω ∈ Ω
I
(R
a
), k ≥ 1, and all x ∈ R
a
,
(Kω)(x) = k
¸
1
0
t
I−1
ω
j
1
j
2
...j
k
(tx) dt
x
j
1
dx
j
2
∧ ∧ dx
j
k
.
One easily shows that dK+Kd = Id; this means that K is a homotopy of the de Rham
complex of R
a
deﬁned for k ≥ 1, so that, according to Proposition 1.12 and Remark
1.13, all cohomology groups vanish in positive degree. Explicitly, if ω is closed, we have
ω = dKω, so that ω is exact.
Exercise 1.2. Realize the circle S
1
as the unit circle in R
2
. Show that the in
tegration of 1forms on S
1
yields an isomorphism H
1
o1
(S
1
) · R. This argument can
be quite easily generalized to show that, if X is a connected, compact and orientable
ndimensional manifold, then H
a
o1
(X) · R.
If a manifold is a cartesian product, X = X
1
X
2
, there is a way to compute the
de Rham cohomology of X out of the de Rham cohomology of X
1
and X
2
(K¨ unneth
theorem, cf. [3]). For later use, we prove here a very particular case. This will serve
also as an example of the notion of homotopy between complexes.
Proposition 1.3. If X is a diﬀerentiable manifold, then H
I
o1
(X R)
· H
I
o1
(X) for all k ≥ 0.
8 1. INTRODUCTORY MATERIAL
Proof. Let t a coordinate on R. Denoting by p
1
, p
2
the projections of X R onto
its two factors, every kform ω on X R can be written as
(1.1) ω = f p
∗
1
ω
1
+g p
∗
1
ω
2
∧ p
∗
2
dt
where ω
1
∈ Ω
I
(X), ω
2
∈ Ω
I−1
(X), and f, g are functions on XR.
1
Let s: X → XR
be the section s(x) = (x, 0). One has p
1
◦s = id
A
(i.e., s is indeed a section of p
1
), hence
s
∗
◦p
∗
1
: Ω
•
(X) → Ω
•
(X) is the identity. We also have a morphism p
∗
1
◦s
∗
: Ω
•
(XR) →
Ω
•
(XR). This is not the identity (as a matter of fact one, has p
∗
1
◦s
∗
(ω) = f(x, 0) p
∗
1
ω
1
).
However, this morphism is homotopic to id
Ω
•
(A×R)
, while id
Ω
•
(A)
is deﬁnitely homotopic
to itself, so that the complexes Ω
•
(X) and Ω
•
(X R) are homotopic, thus proving our
claim as a consequence of Corollary 1.10. So we only need to exhibit a homotopy
between p
∗
1
◦ s
∗
and id
Ω
•
(A×R)
.
This homotopy K: Ω
•
(X R) → Ω
•−1
(X R) is deﬁned as (with reference to
equation (1.1))
K(ω) = (−1)
I
¸

0
g(x, s) ds
p
∗
2
ω
2
.
The proof that K is a homotopy is an elementary direct computation,
2
after which one
gets
d ◦ K +K ◦ d = id
Ω
•
(A×R)
−p
∗
1
◦ s
∗
.
In particular we obtain that the morphisms
p

1
: H
•
o1
(X) → H
•
o1
(X R), s

: H
•
o1
(X R) → H
•
o1
(X)
are isomorphisms.
Remark 1.4. If we take X = R
a
and make induction on n we get another proof of
Poincar´e lemma.
Exercise 1.5. By a similar argument one proves that for all k > 0
H
I
o1
(X S
1
) · H
I
o1
(X) ⊕H
I−1
o1
(X).
Now we give an example of a long cohomology exact sequence within de Rham’s the
ory. Let X be a diﬀerentiable manifold, and Y a closed submanifold. Let r
I
: Ω
I
(X) →
Ω
I
(Y ) be the restriction morphism; this is surjective. Since the exterior diﬀerential com
mutes with the restriction, after letting Ω
I
(X, Y ) = ker r
I
a diﬀerential d
: Ω
I
(X, Y ) →
1
In intrinsic notation this means that
Ω
k
(X ×R) C
∞
(X ×R) ⊗
C
∞
(X)
[Ω
k
(X) ⊕Ω
k−1
(X)].
2
The reader may consult e.g. [3], §I.4.
2. DE RHAM COHOMOLOGY 9
Ω
I+1
(X, Y ) is deﬁned. We have therefore an exact sequence of diﬀerential modules, in
a such a way that the diagram
0
Ω
I−1
(X, Y )
o
Ω
I−1
(X)
o
·
k−1
Ω
I−1
(Y )
o
0
0
Ω
I
(X, Y )
Ω
I
(X)
·
k
Ω
I
(Y )
0
commutes. The complex (Ω
•
(X, Y ), d
) is called the relative de Rham complex,
3
and its
cohomology groups by H
I
o1
(X, Y ) are called the relative de Rham cohomology groups.
One has a long cohomology exact sequence
0 → H
0
o1
(X, Y ) → H
0
o1
(X) → H
0
o1
(Y )
δ
→ H
1
o1
(X, Y )
→ H
1
o1
(X) → H
1
o1
(Y )
δ
→ H
2
o1
(X, Y ) → . . .
Exercise 1.6. 1. Prove that the space ker d
: Ω
I
(X, Y ) → Ω
I+1
(X, Y ) is for all
k ≥ 0 the kernel of r
I
restricted to Z
I
(X), i.e., is the space of closed kforms on X
which vanish on Y . As a consequence H
0
o1
(X, Y ) = 0 whenever X and Y are connected.
2. Let n = dimX and dimY ≤ n − 1. Prove that H
a
o1
(X, Y ) → H
a
o1
(X) surjects,
and that H
I
o1
(X, Y ) = 0 for k ≥ n + 1. Make an example where dimX = dimY and
check if the previous facts still hold true.
Example 1.7. Given the standard embedding of S
1
into R
2
, we compute the relative
cohomology H
•
o1
(R
2
, S
1
). We have the long exact sequence
0 → H
0
o1
(R
2
, S
1
) → H
0
o1
(R
2
) → H
0
o1
(S
1
)
δ
→ H
1
o1
(R
2
, S
1
)
→ H
1
o1
(R
2
) → H
1
o1
(S
1
)
δ
→ H
2
o1
(R
2
, S
1
) → H
2
o1
(R
2
) → 0 .
As in the previous exercise, we have H
I
o1
(R
2
, S
1
) = 0 for k ≥ 3. Since H
0
o1
(R
2
) · R,
H
1
o1
(R
2
) = H
2
o1
(R
2
) = 0, H
0
o1
(S
1
) · H
1
o1
(S
1
) · R, we obtain the exact sequences
0 → H
0
o1
(R
2
, S
1
) → R
·
→ R → H
1
o1
(R
2
, S
1
) → 0
0 → R → H
2
o1
(R
2
, S
1
) → 0
where the morphism r is an isomorphism. Therefore from the ﬁrst sequence we get
H
0
o1
(R
2
, S
1
) = 0 (as we already noticed) and H
1
o1
(R
2
, S
1
) = 0. From the second we
obtain H
2
o1
(R
2
, S
1
) · R.
From this example we may abstract the fact that whenever X and Y are connected,
then H
0
o1
(X, Y ) = 0.
Exercise 1.8. Consider a submanifold Y of R
2
formed by two disjoint embedded
copies of S
1
. Compute H
•
o1
(R
2
, Y ).
3
Sometimes this term is used for another cohomology complex, cf. [3].
10 1. INTRODUCTORY MATERIAL
3. MayerVietoris sequence in de Rham cohomology
The MayerVietoris sequence is another example of long cohomology exact sequence
associated with de Rham cohomology, and is very useful for making computations.
Assume that a diﬀerentiable manifold X is the union of two open subset U, V . For
every k, 0 ≤ k ≤ n = dimX we have the sequence of morphisms
(1.2) 0 → Ω
I
(X)
j
→ Ω
I
(U) ⊕Ω
I
(V )
j
→ Ω
I
(U ∩ V ) → 0
where
i(ω) = (ω
l
, ω
\
), p((ω
1
, ω
2
)) = ω
1l∩\
−ω
2l∩\
.
One easily checks that i is injective and that ker p = Imi. The surjectivity of p is
somehow less trivial, and to prove it we need a partition of unity argument. From
elementary diﬀerential geometry we recall that a partition of unity subordinated to the
cover ¦U, V ¦ of X is a pair of smooth functions f
1
, f
2
: X → R such that
supp(f
1
) ⊂ U, supp(f
2
) ⊂ V, f
1
+f
2
= 1.
Given τ ∈ Ω
I
(U ∩ V ), let
ω
1
= f
2
τ, ω
2
= −f
1
τ.
These kform are deﬁned on U and V , respectively. Then p((ω
1
, ω
2
)) = τ. Thus the
sequence (1.2) is exact. Since the exterior diﬀerential d commutes with restrictions, we
obtain a long cohomology exact sequence
(1.3) 0 → H
0
o1
(X) → H
0
o1
(U) ⊕H
0
o1
(V ) → H
0
o1
(U ∩ V )
δ
→ H
1
o1
(X) →
→ H
1
o1
(U) ⊕H
1
o1
(V ) → H
1
o1
(U ∩ V )
δ
→ H
2
o1
(X) → . . .
This is the MayerVietoris sequence. The argument may be generalized to a union
of several open sets.
4
Exercise 1.1. Use the MayerVietoris sequence (1.3) to compute the de Rham
cohomology of the circle S
1
.
Example 1.2. We use the MayerVietoris sequence (1.3) to compute the de Rham
cohomology of the sphere S
2
(as a matter of fact we already know the 0th and 2nd
group, but not the ﬁrst). Using suitable stereographic projections, we can assume that
U and V are diﬀeomorphic to R
2
, while U ∩V · S
1
R. Since S
1
R is homotopic to
S
1
, it has the same de Rham cohomology. Hence the sequence (1.3) becomes
0 → H
0
o1
(S
2
) → R ⊕R → R → H
1
o1
(S
2
) → 0
0 → R → H
2
o1
(S
2
) → 0.
From the ﬁrst sequence, since H
0
o1
(S
2
) · R, the map H
0
o1
(S
2
) → R ⊕ R is injective,
and then we get H
1
o1
(S
2
) = 0; from the second sequence, H
2
o1
(S
2
) · R.
4
The MayerVietoris sequence foreshadows the
ˇ
Cech cohomology we shall study in Chapter 3.
4. HOMOTOPY THEORY 11
Exercise 1.3. Use induction to show that if n ≥ 3, then H
I
o1
(S
a
) · R for k = 0, n,
H
I
o1
(S
a
) = 0 otherwise.
Exercise 1.4. Consider X = S
2
and Y = S
1
, embedded as an equator in S
2
.
Compute the relative de Rham cohomology H
•
o1
(S
2
, S
1
).
4. Elementary homotopy theory
4.1. Homotopy of paths. Let X be a topological space. We denote by I the
closed interval [0, 1]. A path in X is a continuous map γ : I → X. We say that X
is pathwise connected if given any two points x
1
, x
2
∈ X there is a path γ such that
γ(0) = x
1
, γ(1) = x
2
.
A homotopy Γ between two paths γ
1
, γ
2
is a continuous map Γ: I I → X such
that
Γ(t, 0) = γ
1
(t), Γ(t, 1) = γ
2
(t).
If the two paths have the same end points (i.e. γ
1
(0) = γ
2
(0) = x
1
, γ
1
(1) = γ
2
(1) = x
2
),
we may introduce the stronger notion of homotopy with ﬁxed end points by requiring
additionally that Γ(0, s) = x
1
, Γ(1, s) = x
2
for all s ∈ I.
Let us ﬁx a base point x
0
∈ X. A loop based at x
0
is a path such that γ(0) = γ(1) =
x
0
. Let us denote L(x
0
) th set of loops based at x
0
. One can deﬁne a composition
between elements of L(x
0
) by letting
(γ
2
γ
1
)(t) =
γ
1
(2t), 0 ≤ t ≤
1
2
γ
2
(2t −1),
1
2
≤ t ≤ 1.
This does not make L(x
0
) into a group, since the composition is not associative (com
posing in a diﬀerent order yields diﬀerent parametrizations).
Proposition 1.1. If x
1
, x
2
∈ X and there is a path connecting x
1
with x
2
, then
L(x
1
) · L(x
2
).
Proof. Let c be such a path, and let γ
1
∈ L(x
1
). We deﬁne γ
2
∈ L(x
2
) by letting
γ
2
(t) =
c(1 −3t), 0 ≤ t ≤
1
3
γ
1
(3t −1),
1
3
≤ t ≤
2
3
c(3t −2),
2
3
≤ t ≤ 1.
This establishes the isomorphism.
4.2. The fundamental group. Again with reference with a base point x
0
, we
consider in L(x
0
) an equivalence relation by decreeing that γ
1
∼ γ
2
if there is a homotopy
with ﬁxed end points between γ
1
and γ
2
. The composition law in L
a
0
descends to a
group structure in the quotient
π
1
(X, x
0
) = L(x
0
)/ ∼ .
12 1. INTRODUCTORY MATERIAL
π
1
(X, x
0
) is the fundamental group of X with base point x
0
; in general it is nonabelian,
as we shall see in examples. As a consequence of Proposition 1.1, if x
1
, x
2
∈ X and
there is a path connecting x
1
with x
2
, then π
1
(X, x
1
) · π
1
(X, x
2
). In particular, if
X is pathwise connected the fundamental group π
1
(X, x
0
) is independent of x
0
up to
isomorphism; in this situation, one uses the notation π
1
(X).
Definition 1.2. X is said to be simply connected if it is pathwise connected and
π
1
(X) = ¦e¦.
The simplest example of a simply connected space is the onepoint space ¦∗¦.
Since the deﬁnition of the fundamental group involves the choice of a base point, to
describe the behaviour of the fundamental group we need to introduce a notion of map
which takes the base point into account. Thus, we say that a pointed space (X, x
0
) is a
pair formed by a topological space X with a chosen point x
0
. A map of pointed spaces
f : (X, x
0
) → (Y, y
0
) is a continuous map f : X → Y such that f(x
0
) = y
0
. It is easy
to show that a map of pointed spaces induces a group homomorphism f
∗
: π(X, x
0
) →
π
1
(Y, y
0
).
4.3. Homotopy of maps. Given two topological spaces X, Y , a homotopy betwe
en two continuous maps f, g : X → Y is a map F : XI → Y such that F(x, 0) = f(x),
F(x, 1) = g(x) for all x ∈ X. One then says that f and g are homotopic.
Definition 1.3. One says that two topological spaces X, Y are homotopically equi
valent if there are continuous maps f : X → Y , g : Y → X such that g ◦ f is homotopic
to id
A
, and f ◦ g is homotopic to id
Y
. The map f, g are said to be homotopical equi
valences,.
Of course, homeomorphic spaces are homotopically equivalent.
Example 1.4. For any manifold X, take Y = XR, f(x) = (x, 0), g the projection
onto X. Then F : X I → X, F(x, t) = x is a homotopy between g ◦ f and id
A
, while
G: X RI → X R, G(x, s, t) = (x, st) is a homotopy between f ◦ g and id
Y
. So X
and X R are homotopically equivalent. The reader should be able to concoct many
similar examples.
Given two pointed spaces (X, x
0
), (Y, y
0
), we say they are homotopically equivalent
if there exist maps of pointed spaces f : (X, x
0
) → (Y, y
0
), g : (Y, y
0
) → (X, x
0
) that
make the topological spaces X, Y homotopically equivalent.
Proposition 1.5. Let f : (X, x
0
) → (Y, y
0
) be a homotopical equivalence. Then
f
∗
: π
∗
(X, x
0
) → (Y, y
0
) is an isomorphism.
Proof. Let g : (Y, y
0
) → (X, x
0
) be a map that realizes the homotopical equival
ence, and denote by F a homotopy between g ◦ f and id
A
. Let γ be a loop based at x
0
.
4. HOMOTOPY THEORY 13
Then g ◦ f ◦ γ is again a loop based at x
0
, and the map
Γ: I I → X, Γ(s, t) = F(γ(s), t)
is a homotopy between γ and g ◦ f ◦ γ, so that γ = g ◦ f ◦ γ in π
1
(X, x
0
). Hence,
g
∗
◦ f
∗
= id
π
1
(A.a
0
)
. In the same way one proves that f
∗
◦ g
∗
= id
π
1
(Y.j
0
)
, so that the
claim follows.
Corollary 1.6. If two pathwise connected spaces X and Y are homotopic, then
their fundamental groups are isomorphic.
Definition 1.7. A topological space is said to be contractible if it is homotopically
equivalent to the onepoint space ¦∗¦.
A contractible space is simply connected.
Exercise 1.8. 1. Show that R
a
is contractible, hence simply connected. 2. Com
pute the fundamental groups of the following spaces: the punctured plane (R
2
minus a
point); R
3
minus a line; R
a
minus a (n −2)plane (for n ≥ 3).
4.4. Homotopic invariance of de Rham cohomology. We may now prove the
invariance of de Rham cohomology under homotopy.
Lemma 1.9. Let X, Y be diﬀerentiable manifolds, and let f, g : X → Y be two
homotopic smooth maps. Then the morphisms they induce in cohomology coincide,
f

= g

.
Proof. We choose a homotopy between f and g in the form of a smooth
5
map
F : X R → Y such that
F(x, t) = f(x) if t ≤ 0, F(x, t) = g(x) if t ≥ 1 .
We deﬁne sections s
0
, s
1
: X → X R by letting s
0
(x) = (x, 0), s
1
(x) = (x, 1). Then
f = F ◦ s
0
, g = F ◦ s
1
, so f

= s

0
◦ F

and g

= s

1
◦ F

. Let p
1
: X R → X,
p
2
: X R → R be the projections. Then s

0
◦ p

1
= s

1
◦ p

1
= Id. By Proposition 1.3 p

1
is an isomorphism. Then s

0
= s

1
, and f

= F

= g

.
Proposition 1.10. Let X and Y be homotopic diﬀerentiable manifolds. Then
H
I
o1
(X) · H
I
o1
(Y ) for all k ≥ 0.
Proof. If f, g are two smooth maps realizing the homotopy, then f

◦g

= g

◦f

=
Id, so that both f

and g

are isomorphisms.
5
For the fact that F can be taken smooth cf. [3].
14 1. INTRODUCTORY MATERIAL
4.5. The van Kampen theorem. The computation of the fundamental group
of a topological space is often unsuspectedly complicated. An important tool for such
computations is the van Kampen theorem, which we state without proof. This theorem
allows one, under some conditions, to compute the fundamental group of an union U∪V
if one knows the fundamental groups of U, V and U ∩ V . As a prerequisite we need
the notion of amalgamated product of two groups. Let G, G
1
, G
2
be groups, with ﬁxed
morphisms f
1
: G → G
1
, f
2
: G → G
2
. Let F the free group generated by G
1
H G
2
and
denote by the product in this group.
6
Let R be the normal subgroup generated by
elements of the form
7
(xy) y
−1
x
−1
with x, y both in G
1
or G
2
f
1
(g) f
2
(g)
−1
for g ∈ G.
Then one deﬁnes the amalgamated product G
1
∗
G
G
2
as F/R. There are natural maps
g
1
: G
1
→ G
1
∗
G
G
2
, g
2
: G
2
→ G
1
∗
G
G
2
obtained by composing the inclusions with
the projection F → F/R, and one has g
1
◦ f
1
= g
2
◦ f
2
. Intuitively, one could say that
G
1
∗
G
G
2
is the smallest subgroup generated by G
1
and G
2
with the identiﬁcation of
f
1
(g) and f
2
(g) for all g ∈ G.
Exercise 1.11. (1) Prove that if G
1
= G
2
= ¦e¦, and G is any group, then
G
1
∗
G
G
2
= ¦e¦.
(2) Let Gbe the group with three generators a, b, c, satisfying the relation ab = cba.
Let Z → G be the homomorphism induced by 1 → c. Prove that G ∗
Z
G is
isomorphic to a group with four generators m, n, p, q, satisfying the relation
mnm
−1
n
−1
p q p
−1
q
−1
= e.
Suppose now that a pathwise connected space X is the union of two pathwise con
nected open subsets U, V , and that U ∩V is pathwise connected. There are morphisms
π
1
(U ∩ V ) → π
1
(U), π
1
(U ∩ V ) → π
1
(V ) induced by the inclusions.
Proposition 1.12. π
1
(X) · π
1
(U) ∗
π
1
(l∩\ )
π
1
(V ).
This is a simpliﬁed form of van Kampen’s theorem, for a full statement see [6].
Example 1.13. We compute the fundamental group of a ﬁgure 8. Think of the ﬁgure
8 as the union of two circles X in R
2
which touch in one pount. Let p
1
, p
2
be points
in the two respective circles, diﬀerent from the common point, and take U = X −¦p
1
¦,
V = X −¦p
2
¦. Then π
1
(U) · π
1
(V ) · Z, while U ∩ V is simply connected. It follows
that π
1
(X) is a free group with two generators. The two generators do not commute;
this can also be checked “experimentally” if you think of winding a string along the
6
F is the group whose elements are words x
1
1
x
2
. . . x
n
or the empty word, where the letters x
i
are
either in G
1
or G
2
,
i
= ±1, and the product is given by juxtaposition.
7
The ﬁrst relation tells that the product of letters in the words of F are the product either in G
1
or G
2
, when this makes sense. The second relation kind of “glues” G
1
and G
2
along the images of G.
4. HOMOTOPY THEORY 15
ﬁgure 8 in a proper way... More generally, the fundamental group of the corolla with n
petals (n copies of S
1
all touching in a single point) is a free group with n generators.
Exercise 1.14. Prove that for any n ≥ 2 the sphere S
a
is simply connected. Deduce
that for n ≥ 3, R
a
minus a point is simply connected.
Exercise 1.15. Compute the fundamental group of R
2
with n punctures.
4.6. Other ways to compute fundamental groups. Again, we state some res
ults without proof.
Proposition 1.16. If G is a simply connected topological group, and H is a normal
discrete subgroup, then π
1
(G/H) · H.
Since S
1
· R/Z, we have thus proved that
π
1
(S
1
) · Z.
In the same way we compute the fundamental group of the ndimensional torus
T
a
= S
1
S
1
(n times) · R
a
/Z
a
,
obtaining π
1
(T
a
) · Z
a
.
Exercise 1.17. Compute the fundamental group of a 2dimensional punctured torus
(a torus minus a point). Use van Kampen’s theorem to compute the fundamental
group of a Riemann surface of genus 2 (a compact, orientable, connected 2dimensional
diﬀerentiable manifold of genus 2, i.e., “with two handles”). Generalize your result to
any genus.
Exercise 1.18. Prove that, given two pointed topological spaces (X, x
0
), (Y, y
0
),
then
π
1
(X Y, (x
0
, y
0
)) · π
1
(X, x
0
) π
1
(Y, y
0
).
This gives us another way to compute the fundamental group of the ndimensional
torus T
a
(once we know π
1
(S
1
)).
Exercise 1.19. Prove that the manifolds S
3
and S
2
S
1
are not homeomorphic.
Exercise 1.20. Let X be the space obtained by removing a line from R
2
, and a
circle linking the line. Prove that π
1
(X) · Z ⊕ Z. Prove the stronger result that X is
homotopic to the 2torus.
CHAPTER 2
Singular homology theory
1. Singular homology
In this Chapter we develop some elements of the homology theory of topological
spaces. There are many diﬀerent homology theories (simplicial, cellular, singular,
ˇ
Cech
Alexander, ...) even though these theories coincide when the topological space they
are applied to is reasonably wellbehaved. Singular homology has the disadvantage of
appearing quite abstract at a ﬁrst contact, but in exchange of this we have the fact that
it applies to any topological space, its functorial properties are evident, it requires very
little combinatorial arguments, it relates to homotopy in a clear way, and once the basic
properties of the theory have been proved, the computation of the homology groups is
not diﬃcult.
1.1. Deﬁnitions. The basic blocks of singular homology are the continuous maps
from standard subspaces of Euclidean spaces to the topological space one considers. We
shall denote by P
0
, P
1
, . . . , P
a
the points in R
a
P
0
= 0, P
j
= (0, 0, . . . , 0, 1, 0, . . . , 0) (with just one 1 in the ith position).
The convex hull of these points is denoted by ∆
a
and is called the standard nsimplex.
Alternatively, one can describe ∆
I
as the set of points in R
a
such that
x
j
≥ 0, i = 1, . . . , n,
a
¸
j=1
x
j
≤ 1.
The boundary of ∆
a
is formed by n + 1 faces F
j
a
(i = 0, 1, . . . , n) which are images of
the standard (n − 1)simplex by aﬃne maps R
a−1
→ R
a
. These faces may be labelled
by the vertex of the simplex which is opposite to them: so, F
j
a
is the face opposite to
P
j
.
Given a topological space X, a singular nsimplex in X is a continuous map σ: ∆
a
→
X. The restriction of σ to any of the faces of ∆
a
deﬁnes a singular (n − 1)simplex
σ
j
= σ
1
i
n
(or σ ◦ F
j
a
if we regard F
j
a
as a singular (n −1)simplex).
If Q
0
, . . . , Q
I
are k +1 points in R
a
, there is a unique aﬃne map R
I
→ R
a
mapping
P
0
, . . . , P
I
to the Q’s. This aﬃne map yields a singular ksimplex in R
a
that we denote
< Q
0
, . . . , Q
I
>. If Q
j
= P
j
for 0 ≤ i ≤ k, then the aﬃne map is the identity on R
I
, and
we denote the resulting singular ksimplex by δ
I
. The standard nsimplex ∆
a
may so
17
18 2. HOMOLOGY THEORY
also be denoted < P
0
, . . . , P
a
>, and the face F
j
a
of ∆
a
is the singular (n − 1)simplex
< P
0
, . . . ,
ˆ
P
j
, . . . , P
a
>, where the hat denotes omission.
Choose now a commutative unital ring R. We denote by S
I
(X, R) the free group
generated over R by the singular ksimplexes in X. So an element in S
I
(X, R) is a
“formal” ﬁnite linear combination (called a singular chain)
σ =
¸
;
a
;
σ
;
with a
;
∈ R, and the σ
;
are singular ksimplexes. Thus, S
I
(X, R) is an Rmodule, and,
via the inclusion Z → R given by the identity in R, an abelian group. For k ≥ 1 we
deﬁne a morphism ∂ : S
I
(X, R) → S
I−1
(X, R) by letting
∂σ =
I
¸
j=0
(−1)
j
σ ◦ F
j
I
for a singular ksimplex σ and exteding by Rlinearity. For k = 0 we deﬁne ∂σ = 0.
Example 2.1. If Q
0
, . . . , Q
I
are k + 1 points in R
a
, one has
∂ < Q
0
, . . . , Q
I
>=
I
¸
j=0
(−1)
j
< Q
0
, . . . ,
ˆ
Q
j
, . . . , Q
I
> .
Proposition 2.2. ∂
2
= 0.
Proof. Let σ be a singular ksimplex.
∂
2
σ =
I
¸
j=0
(−1)
j
∂(σ ◦ F
j
I
) =
I
¸
j=0
(−1)
j
I−1
¸
;=0
(−1)
;
σ ◦ F
j
I
◦ F
;
I−1
=
I
¸
;<j=1
(−1)
j+;
σ ◦ F
;
I
◦ F
j−1
I−1
+
I−1
¸
0=j≤;
(−1)
j+;
σ ◦ F
j
I
◦ F
;
I−1
Resumming the ﬁrst sum by letting i = j, j = i −1 the last two terms cancel.
So (S
•
(X, R), ∂) is a (homology) graded diﬀerential module. Its homology groups
H
I
(X, R) are the singular homology groups of X with coeﬃcients in R. We shall use
the following notation and terminology:
Z
I
(X, R) = ker ∂ : S
I
(X, R) → S
I−1
(X, R) (the module of kcycles);
B
I
(X, R) = Im∂ : S
I+1
(X, R) → S
I
(X, R) (the module of kboundaries);
therefore, H
I
(X, R) = Z
I
(X, R)/B
I
(X, R). Notice that Z
0
(X, R) ≡ S
0
(X, R).
1.2. Basic properties.
Proposition 2.3. If X is the union of pathwise connected components X
;
, then
H
I
(X, R) · ⊕
;
H
I
(X
;
, R) for all k ≥ 0.
1. SINGULAR HOMOLOGY 19
Proof. Any singular ksimplex must map ∆
I
inside a pathwise connected compon
ents (if two points of ∆
I
would map to points lying in diﬀerent components, that would
yield path connecting the two points).
Proposition 2.4. If X is pathwise connected, then H
0
(X, R) · R.
Proof. This follows from the fact that a 0cycle c =
¸
;
a
;
x
;
is a boundary if and
only if
¸
;
a
;
= 0. Indeed, if c is a boundary, then c = ∂(
¸
;
b
;
γ
;
) for some paths γ
;
, so
that c =
¸
;
b
;
(γ
;
(1) −γ
;
(0)), and the coeﬃcients sum up to zero. On the other hand,
if
¸
;
a
;
= 0, choose a base point x
0
∈ X. Then one can write
c =
¸
;
a
;
x
;
=
¸
;
a
;
x
;
−(
¸
;
a
;
)x
0
=
¸
;
a
;
(x
;
−x
0
) = ∂
¸
;
a
;
γ
;
if γ
;
is a path joining x
0
to x
;
.
This means that B
0
(X, R) is the kernel of the surjective map Z
0
(X, R) = S
0
(X, R) →
R given by
¸
;
a
;
x
;
→
¸
;
a
;
, so that H
0
(X, R) = Z
0
(X, R)/B
0
(X, R) · R.
Let f : X → Y be a continuous map of topological spaces. If σ is a singular ksimplex
in X, then f ◦σ is a singular ksimplex in Y . This yields a morphism S
I
(f): S
I
(X, R) →
S
I
(Y, R) for every k ≥ 0. It is immediate to prove that S
I
(f) ◦ ∂ = ∂ ◦ S
I+1
(f):
S
I
(f)(∂σ) = f ◦
I+1
¸
j=0
(−1)
j
σ ◦ F
j
I+1
= ∂(f ◦ σ) = ∂(S
I
(f)(σ)) .
This implies that f induces a morphism H
I
(X, R) → H
I
(Y, R), that we denote f
.
. It
is also easy to check that, if g : Y → W is another continous map, then S
I
(g ◦ f) =
S
I
(g) ◦ S
I
(f), and (g ◦ f)
.
= g
.
◦ f
.
.
1.3. Homotopic invariance.
Proposition 2.5. If f, g : X → Y are homotopic map, the induced maps in homo
logy coincide.
It should be by now clear that this yields as an immediate consequence the homotopic
invariance of the singular homology.
Corollary 2.6. If two topological spaces are homotopically equivalent, their singu
lar homologies are isomorphic.
To prove Proposition 2.5 we build, for every k ≥ 0 and any topological space X, a
morphism (called the prism operator) P : S
I
(X) → S
I+1
(X I) (here I denotes again
the unit closed interval in R). We deﬁne the morphism P in two steps.
Step 1. The ﬁrst step consists in deﬁnining a singular (k + 1)chain π
I+1
in the
topological space ∆
I
I by subdiving the polyhedron ∆
I
I ⊂ R
I+1
(a “prysm”
20 2. HOMOLOGY THEORY
A
0
A
1
B
0
B
1
Figure 1. The prism π
2
over ∆
1
over the standard symplesx ∆
I
) into a number of singular (k + 1)simplexes, and sum
ming them with suitable signs. The polyhedron ∆
I
I ⊂ R
I+1
has 2(k + 1) vertices
A
0
, . . . , A
I
, B
0
, . . . , B
I
, given by A
j
= (P
j
, 0), B
j
= (P
j
, 1). We deﬁne
π
I+1
=
I
¸
j=0
(−1)
j
< A
0
, . . . , A
j
, B
j
, . . . , B
I
> .
For instance, for k = 1 we have
π
2
=< A
0
, B
0
, B
1
> − < A
0
, A
1
, B
1
> .
Step 2. If σ is a singular ksimplex in a topological space X, then σid is a continous
map ∆
I
I → X I. Therefore it makes sense to deﬁne the singular (k + 1)chain
P(σ) in X as
(2.1) P(σ) = S
I+1
(σ id)(π
I+1
).
The deﬁnition of the prism operator implies its functoriality:
Proposition 2.7. If f : X → Y is a continuous map, the diagram
S
I
(X)
1
S
k
(;)
S
I+1
(X I)
S
k+1
(;×id)
S
I
(Y )
1
S
I+1
(Y I)
commutes.
Proof. It is just a matter of computation.
S
I+1
(f id) ◦ P(σ) = S
I+1
(f id) ◦ S
I+1
(σ id)(π
I+1
)
= S
I+1
(f ◦ σ id)(π
I+1
) = P(S
I
(f)) .
The relevant property of the prism operator is proved in the next Lemma.
1. SINGULAR HOMOLOGY 21
Lemma 2.8. Let λ
0
, λ
j
: X → X I be the maps λ
0
(x) = (x, 0), λ
1
(x) = (x, 1).
Then
(2.2) ∂ ◦ P +P ◦ ∂ = S
I
(λ
1
) −S
I
(λ
0
)
as maps S
I
(X) → S
I
(X I).
Proof. Let δ
I
: ∆
I
→ ∆
I
be the identity map regarded as singular ksimplex in
∆
I
. Notice that P(δ
I
) = π
I+1
.
We ﬁrst check the identity (2.2) for X = ∆
I
, applying both sides of (2.2) to δ
I
. The
right side yields
< B
0
, . . . , B
I
> − < A
0
, . . . A
I
> .
We compute now the action of the left side of (2.2) on δ
I
.
∂P(δ
I
) =
I
¸
j=0
(−1)
j
∂ < A
0
, . . . , A
j
, B
j
, . . . , B
I
>
=
I
¸
;≤j=0
(−1)
j+;
< A
0
, . . . ,
ˆ
A
;
, . . . A
j
, B
j
, . . . , B
I
>
+
I
¸
j≤;=0
(−1)
j+;+1
< A
0
, . . . A
j
, B
j
, . . . ,
ˆ
B
;
, . . . B
I
> .
All terms with i = j cancel with the exception of < B
0
, . . . , B
I
> − < A
0
, . . . A
I
>. So
we have
∂P(δ
I
) = < B
0
, . . . , B
I
> − < A
0
, . . . A
I
>
+
I
¸
;<j=1
(−1)
j+;
< A
0
, . . . ,
ˆ
A
;
, . . . A
j
, B
j
, . . . , B
I
>
−
I
¸
j<;=1
(−1)
j+;
< A
0
, . . . A
j
, B
j
, . . . ,
ˆ
B
;
, . . . B
I
> .
On the other hand, one has
∂δ
I
=
I
¸
;=0
(−1)
;
< P
0
, . . . ,
ˆ
P
;
, . . . , P
I
> .
Since
P(< P
0
, . . . ,
ˆ
P
;
, . . . , P
I
>) =
¸
j<;
(−1)
j
< A
0
, . . . , A
j
, B
j
, . . . ,
ˆ
B
;
, . . . , B
I
>
−
¸
j;
(−1)
j
< A
0
, . . . ,
ˆ
A
;
, . . . , A
j
, B
j
, . . . , B
I
>
we obtain the equation (2.2) (note that exchanging the indices i, j changes the sign).
22 2. HOMOLOGY THEORY
We must now prove that if equation (2.2) holds when both sides are applied to δ
I
,
then it holds in general. One has indeed
∂P(σ) = ∂S
I+1
(σ id)(P(δ
I
)) = S
I
(σ id)(∂P(δ
I
))
P(∂σ) = P∂(S
I
(σ)(δ
I
))
= P(S
I−1
(σ)(∂δ
I
)) = S
I
(σ id)(P(∂δ
I
))
so that
∂P(σ) +P(∂σ) = S
I+1
(σ id)(∂P(δ
I
)) +P(∂δ
I
))
= S
I+1
(σ id)(S
I
(
¯
λ
1
) −S
I
(
¯
λ
0
)) = S
I
(λ
1
) −S
I
(λ
0
)
where
¯
λ
0
,
¯
λ
1
are the obvious maps ∆
I
→ ∆
I
I.
Equation (2.2) states that P is a hotomopy (in the sense of homological algebra)
between the maps λ
0
and λ
1
, so that one has (λ
1
)
.
= (λ
2
)
.
in homology.
Proof of Proposition 2.5. Let F be a hotomopy between the maps f and g. Then,
f = F ◦ λ
0
, g = F ◦ λ
1
, so that
f
.
= (F ◦ λ
0
)
.
= F
.
◦ (λ
0
)
.
= F
.
◦ (λ
1
)
.
= (F ◦ λ
1
)
.
= g
.
.
Corollary 2.9. If X is a contractible space then
H
0
(X, R) · R, H
I
(X, R) = 0 for k > 0.
1.4. Relation between the ﬁrst fundamental group and homology. A loop
γ in X may be regarded as a closed singular 1simplex. If we ﬁx a point x
0
∈ X, we
have a settheoretic map χ: L(x
0
) → S
1
(X, Z). The following result tells us that χ
descends to a group homomorphism χ: π
1
(X, x
0
) → H
1
(X, Z).
Proposition 2.10. If two loops γ
1
, γ
2
are homotopic, then they are homologous
as singular 1simplexes. Moreover, given two loops at x
0
, γ
1
, γ
2
, then χ(γ
2
◦ γ
1
) =
χ(γ
1
) +χ(γ
2
) in H
1
(X, Z).
Proof. Choose a homotopy with ﬁxed endpoints between γ
1
and γ
2
, i.e., a map
Γ: I I → X such that
Γ(t, 0) = γ
1
(t), Γ(t, 1) = γ
2
(t), Γ(0, s) = Γ(1, s) = x
0
for all s ∈ I.
Deﬁne the loops γ
3
(t) = Γ(1, t), γ
4
(t) = Γ(0, t), γ
5
(t) = Γ(t, t). Both loops γ
3
and
γ
4
are actually the constant loop at x
0
. Consider the points P
0
, P
1
, P
2
, Q = (1, 1) in
R
2
, and deﬁne the singular 2simplex
σ = Γ◦ < P
0
, P
1
, Q > −Γ◦ < P
0
, P
2
, Q >
1. SINGULAR HOMOLOGY 23
∧ γ
4
∧ γ
3
>
γ
2
>
γ
1
γ
5
P
0
P
1
Q P
2
Figure 2
(cf. Figure 2). We then have
∂σ = Γ◦ < P
1
, Q > −Γ◦ < P
0
, Q > +Γ◦ < P
0
, P
1
>
− Γ◦ < P
2
, Q > +Γ◦ < P
0
, Q > −Γ◦ < P
0
, P
2
>
= γ
3
−γ
5
+γ
1
−γ
2
+γ
5
+γ
4
= γ
1
−γ
2
.
This proves that χ(γ
1
) and χ(γ
2
) are homologous. To prove the second claim we need
to deﬁne a singular 2simplex σ such that
∂σ = γ
1
+γ
2
−γ
2
γ
1
.
Consider the point T = (0,
1
2
) in the standard 2simplex ∆
2
and the segment Σ
joining T with P
1
(cf. Figure 3). If Q ∈ ∆
2
lies on or below Σ, consider the line joining
P
0
with Q, parametrize it with a parameter t such that t = 0 in P
0
and t = 1 in the
intersection of the line with Σ, and set σ(Q) = γ
1
(t). Analogously, if Q lies above or
on Σ, consider the line joining P
2
with Q, parametrize it with a parameter t such that
t = 1 in P
2
and t = 0 in the intersection of the line with Σ, and set σ(Q) = γ
2
(t). This
deﬁnes a singular 2simplex σ: ∆
2
→X, and one has
∂σ = σ◦ < P
1
, P
2
> −σ◦ < P
0
, P
2
> +σ◦ < P
0
, P
1
>
= γ
2
−γ
2
γ
1
+γ
1
.
We recall from basic group theory the notion of commutator subgroup. Let G be
any group, and let C(G) be the subgroup generated by elements of the form ghg
−1
h
−1
,
g, h ∈ G. The subgroup C(G) is obviously normal in G; the quotient group G/C(G) is
abelian. We call it the abelianization of G. It turns out that the ﬁrst homology group
of a space with integer coeﬃcients is the abelianization of the fundamental group.
Proposition 2.11. If X is pathwise connected, the morphism χ: π
1
(X, x
0
) →
H
1
(X, Z) is surjective, and its kernel is the commutator subgroup of π
1
(X, x
0
).
24 2. HOMOLOGY THEORY
>
∧
∧
γ
1
γ
2
γ
1
γ
2
d
d
d
d
d
d
d
d
d
d
d
d
r
r
r
r
r
r
r
r
r
r
r
r ¨
¨
¨
¨
¨
¨
e
e
e
e
e
e
e
e
•
Q
•Q
P
0
P
2
P
1
T
Σ
Figure 3
Proof. Let c =
¸
;
a
;
σ
;
be a 1cycle. So we have
0 = ∂c =
¸
;
a
j
(σ
;
(1) −σ
;
(0)).
In this linear combination of points with coeﬃcients in Z some of the points may coin
cide; the sum of the coeﬃcients corresponding to the same point must vanish. Choose a
base point x
0
∈ X and for every j choose a path α
;
from x
0
to σ
;
(0) and a path β
;
from
x
0
to σ
;
(1), in such a way that they depend on the endpoints and not on the indexing
(e.g, if σ
;
(0) = σ
I
(0), choose α
;
= α
I
). Then we have
¸
;
a
;
(β
;
−α
;
) = 0.
Now if we set ¯ σ
;
= α
;
+ σ
;
− β
;
we have c =
¸
;
a
;
¯ σ
;
. Let γ
;
be the loop β
−1
σ
;
α;
then,
χ(
Π
;
γ
o
j
;
) = [c]
so that χ is surjective.
To prove the second claim we need to show that the commutator subgroup of
π
1
(X, x
0
) coincides with ker χ. We ﬁrst notice that since H
1
(X, Z) is abelian, the
commutator subgroup is necessarily contained in ker χ. To prove the opposite inclusion,
let γ be a loop that in homology is a 1boundary, i.e., γ = ∂
¸
;
a
;
σ
;
. So we may write
(2.3) σ
;
= γ
0;
−γ
1;
+γ
2;
for some paths γ
I;
, k = 0, 1, 2. Choose paths (cf. Figure 4)
α
0;
from x
0
to γ
1;
(0) = γ
2;
(0) = P
0
α
1;
from x
0
to γ
2;
(1) = γ
0;
(0) = P
1
α
2;
from x
0
to γ
1;
(1) = γ
0;
(1) = P
2
and consider the loops
β
0;
= α
−1
0;
γ
−1
1;
α
2;
, β
1;
= α
−1
2;
γ
0;
α
1;
, β
2;
= α
−1
1;
γ
2;
α
0;
.
2. RELATIVE HOMOLOGY 25
α
1;
α
0;
α
2;
γ
2;
γ
0;
γ
1;
P
1
P
0
P
2
x
0
Figure 4
Note that the loops
β
;
= β
0;
β
1;
β
2;
= α
−1
0;
γ
−1
1;
γ
0;
γ
2;
α
0;
are homotopic to the constant loop at x
0
(since the image of a singular 2simplex is
contractible). As a consequence one has the equality in π
1
(X, x
0
)
Π
;
[β
;
]
o
j
= e.
This implies that the image of Π
;
[β
;
]
o
j
in π
1
(X, x
0
)/C(π
1
(X, x
0
)) is the identity. On the
other hand from (2.3) we see that γ coincides, up to reordering of terms, with Π
;
β
o
j
;
, so
that the image of the class of γ in π
1
(X, x
0
)/C(π
1
(X, x
0
)) is the identity as well. This
means that γ lies in the commutator subgroup.
So whenever in the examples in Chapter 1 the fundamental groups we computed
turned out to be abelian, we were also computing the group H
1
(X, Z). In particular,
Corollary 2.12. H
1
(X, Z) = 0 if X is simply connected.
Exercise 2.13. Compute H
1
(X, Z) when X is: 1. the corolla with n petals, 2. R
a
minus a point, 3. the circle S
1
, 4. the torus T
2
, 5. a punctured torus, 6. a Riemann
surface of genus g.
2. Relative homology
2.1. The relative homology complex. Given a topological space X, let A be
any subspace (that we consider with the relative topology). We ﬁx a coeﬃcient ring R
which for the sake of conciseness shall be dropped from the notation. For every k ≥ 0
there is a natural inclusion (injective morphism of Rmodules) S
I
(A) ⊂ S
I
(X); the ho
mology operators of the complexes S
•
(A), S
•
(X) deﬁne a morphism δ : S
I
(X)/S
I
(A) →
S
I−1
(X)/S
I−1
(A) which squares to zero. If we deﬁne
Z
I
(X, A) = ker ∂ :
S
I
(X)
S
I
(A)
→
S
I−1
(X)
S
I−1
(A)
26 2. HOMOLOGY THEORY
B
I
(X, A) = Im∂ :
S
I+1
(X)
S
I+1
(A)
→
S
I
(X)
S
I
(A)
we have B
I
(X, A) ⊂ Z
I
(X, A).
Definition 2.1. The homology groups of X relative to A are the Rmodules
H
I
(X, A) = Z
I
(X, A)/B
I
(X, A). When we want to emphasize the choice of the ring R
we write S
I
(X, A; R).
The relative homology is more conveniently deﬁned in a slightly diﬀerent way, which
makes clearer its geometrical meaning. It will be useful to consider the following diagram
0
Z
I
(X)
o
k
Z
I
(X, A)
S
I
(A)
∂
S
I
(X)
o
k
∂
S
I
(X)/S
I
(A)
∂
0
0
B
I−1
(A)
B
I−1
(X)
o
k−1
B
I−1
(X, A)
0
Let
Z
I
(X, A) = ¦c ∈ S
I
(X) [ ∂c ∈ S
I−1
(A)¦
B
I
(X, A) = ¦c ∈ S
I
(X) [ c = ∂b +c
with b ∈ S
I+1
(X), c
∈ S
I
(A)¦ .
Thus, Z
I
(X, A) is formed by the chains whose boundary is in A, and B
I
(A) by the
chains that are boundaries up to chains in A.
Lemma 2.2. Z
I
(X, A) is the preimage of Z
I
(X, A) under the quotient homomorph
ism q
I
; that is, an element c ∈ S
I
(X) is in Z
I
(X, A) if and only if q
I
(c) ∈ Z
I
(X, A).
Proof. If q
I
(c) ∈ Z
I
(X, A) then 0 = ∂ ◦ q
I
(c) = q
I−1
◦ ∂(c) so that c ∈ Z
I
(X, A).
If c ∈ Z
I
(X, A) then q
I−1
◦ ∂(c) = 0 so that q
I
(c) ∈ Z
I
(X, A).
Lemma 2.3. c ∈ S
I
(X) is in B
I
(X, A) if and only if q
I
(c) ∈ B
I
(X, A).
Proof. If c = ∂b + c
with b ∈ S
I+1
(X) and c
∈ S
I
(A) then q
I
(c) = q
I
◦ ∂b =
∂ ◦ q
I+1
(b) ∈ B
I
(X, A). Conversely, if q
I
(c) ∈ B
I
(X, A) then q
I
(c) = ∂ ◦ q
I+1
(b) for
some b ∈ S
I+1
(X), then c −∂b ∈ ker q
I−1
so that c = ∂b +c
with c
∈ S
I
(A).
Proposition 2.4. For all k ≥ 0, H
I
(X, A) · Z
I
(X, A)/B
I
(X, A).
2. RELATIVE HOMOLOGY 27
Proof. What we should do is to prove the commutativity and the exactness of the
rows of the diagram
0
S
I
(A)
∼
B
I
(X, A)
o
k
B
I
(X, A)
0
0
S
I
(A)
Z
I
(X, A)
o
k
Z
I
(X, A)
0
Commutativity is obvious. For the exactness of the ﬁrst row, it is obvious that S
I
(A) ⊂
B
I
(X, A) and that q
I
(c) = 0 if c ∈ S
I
(A). On the other hand if c ∈ B
I
(X, A) we have
c = ∂b + c
with b ∈ S
I+1
(X) and c
∈ S
I
(A), so that q
I
(c) = 0 implies 0 = q
I
◦ ∂b =
∂ ◦ q
I+1
(b), which in turn implies c ∈ S
I
(A). To prove the surjectivity of q
I
, just notice
that by deﬁnition an element in B
I
(X, A) may be represented as ∂b with b ∈ S
I+1
(X).
As for the second row, we have S
I
(A) ⊂ Z
I
(X, A) from the deﬁnition of Z
I
(X, A).
If c ∈ S
I
(A) then q
I
(c) = 0. If c ∈ Z
I
(X, A) and q
I
(c) = 0 then c ∈ S
I
(A) by the
deﬁnition of Z
I
(X, A). Moreover q
I
is surjective by Lemma 2.2.
2.2. Main properties of relative homology. We list here the main properties
of the cohomology groups H
I
(X, A). If a proof is not given the reader should provide
one by her/himself.
• If A is empty, H
I
(X, A) · H
I
(X).
• The relative cohomology groups are functorial in the following sense. Given to
pological spaces X, Y with subsets A ⊂ X, B ⊂ Y , a continous map of pairs is a
continuous map f : X → Y such that f(A) ⊂ B. Such a map induces in natural way
a morphisms of Rmodules f
.
: H
•
(X, A) → H
•
(Y, B). If we consider the inclusion of
pairs (X, ∅) → (X, A) we obtain a morphism H
•
(X) →
•
H(X, A).
• The inclusion map i : A → X induces a morphism H
•
(A) → H
•
(X) and the
composition H
•
(A) → H
•
(X) → H
•
(X, A) vanishes (since Z
I
(A) ⊂ B
I
(X, A)).
• If X = ∪
;
X
;
is a union of pathwise connected components, then H
I
(X, A) ·
⊕
;
H
I
(X
;
, A
;
) where A
;
= A∩ X
;
.
Proposition 2.5. If X is pathwise connected and A is nonempty, then H
0
(X, A)
= 0.
Proof. If c =
¸
;
a
;
x
;
∈ S
0
(X) and γ
;
is a path from x
0
∈ A to x
;
, then
∂(
¸
;
a
;
x
;
) = c −(
¸
;
a
;
)x
0
so that c ∈ B
0
(X, A).
Corollary 2.6. H
0
(X, A) is a free Rmodule generated by the components of X
that do not meet A.
Indeed H
;
(X
;
, A
;
) = 0 if A
;
is empty.
Proposition 2.7. If A = ¦x
0
¦ is a point, H
I
(X, A) · H
I
(X) for k > 0.
28 2. HOMOLOGY THEORY
Proof.
Z
I
(X, A) = ¦c ∈ S
I
(X) [ ∂c ∈ S
I−1
(A)¦ = Z
I
(X) when k > 0
B
I
(X, A) = ¦c ∈ S
I
(X) [ c = ∂b +c
with b ∈ S
I+1
(X), c
∈ S
I
(A)¦
= B
I
(X) when k > 0.
2.3. The long exact sequence of relative homology. By deﬁnition the relative
homology of X with respect to A is the homology of the quotient complex S
•
(X)/S
•
(A).
By Proposition 1.6, adapted to homology by reversing the arrows, one obtains a long
exact cohomology sequence
→ H
2
(A) → H
2
(X) → H
2
(X, A)
→ H
1
(A) → H
1
(X) → H
1
(X, A)
→ H
0
(A) → H
0
(X) → H
0
(X, A) → 0
Exercise 2.8. Assume to know that H
1
(S
1
, R) · R and H
I
(S
1
, R) = 0 for k >
1. Use the long relative homology sequence to compute the relative homology groups
H
•
(R
2
, S
1
; R).
3. The MayerVietoris sequence
The MayerVietoris sequence (in its simplest form, that we are going to consider
here) allows one to compute the homology of a union X = U ∪ V from the knowledge
of the homology of U, V and U ∩V . This is quite similar to what happens in de Rham
cohomology, but in the case of homology there is a subtlety. Let us denote A = U ∩ V .
One would think that there is an exact sequence
0 → S
I
(A)
j
→ S
I
(U) ⊕S
I
(V )
j
→ S
I
(X) → 0
where i is the morphism induced by the inclusions A → U, A → V , and p is given by
p(σ
1
, σ
2
) = σ
1
− σ
2
(again using the inclusions U → X, V → X). However, it is not
possible to prove that p is surjective (if σ is a singular ksimplex whose image is not
contained in U or V , it is not in general possible to write it as a diﬀerence of standard
ksimplexes in U, V ). The trick to circumvent this diﬃculty consists in replacing S
•
(X)
with a diﬀerent complex that however has the same homology.
Let U = ¦U
α
¦ be an open cover of X.
Definition 2.1. A singular kchain σ =
¸
;
a
;
σ
;
is Usmall if every singular k
simplex σ
;
maps into an open set U
α
∈ U for some α. Moreover we deﬁne S
U
•
(X) as
the subcomplex of S
•
(X) formed by Usmall chains.
1
The homology diﬀerential ∂ restricts to S
U
•
(X), so that one has a homology H
U
•
(X).
1
Again, we understand the choice of a coeﬃcient ring R.
3. THE MAYERVIETORIS SEQUENCE 29
r
r
r
r
r
r
£
£
£
£
£
£
£
£
£
£
£
£
B
E
0
E
1
Figure 5. The join B(< E
0
, E
1
>)
Proposition 2.2. H
U
•
(X) · H
•
(X).
To prove this isomorphism we shall build a homotopy between the complexes S
U
•
(X)
and S
•
(X). This will be done in several steps.
Given a singular ksimplex < Q
0
, . . . , Q
I
> in R
a
and a point B ∈ R
a
we consider
the singular simplex < B, Q
0
, . . . , Q
I
>, called the join of B with < Q
0
, . . . , Q
I
>. This
operator B is then extended to singular chains in R
a
by linearity. The following Lemma
is easily proved.
Lemma 2.3. ∂ ◦ B +B ◦ ∂ = Id on S
I
(R
a
) if k > 0, while ∂ ◦ B(σ) = σ −(
¸
;
a
;
)B
if σ =
¸
;
a
;
x
;
∈ S
0
(R
a
).
Next we deﬁne operators Σ: S
I
(X) → S
I
(X) and T : S
I
(X) → S
I+1
(X). The
operator Σ is called the subdivision operator and its eﬀect is that of subdividing a
singular simplex into a linear combination of “smaller” simplexes. The operators Σ
and T, analogously to what we did for the prism operator, will be deﬁned for X = ∆
I
(the space consisting of the standard ksimplex) and for the “identity” singular simplex
δ
I
: ∆
I
→ ∆
I
, and then extended by functoriality. This should be done for all k. One
deﬁnes
Σ(δ
0
) = δ
0
, T(δ
0
) = 0.
and then extends recursively to positive k:
Σ(δ
I
) = B
I
(Σ(∂δ
I
)), T(δ
I
) = B
I
(δ
I
−Σ(δ
I
) −T(∂δ
I
))
where the point B
I
is the barycenter of the standard ksimplex ∆
I
,
B
I
=
1
k + 1
I
¸
;=0
P
;
.
Example 2.4. For k = 1 one gets Σ(δ
1
) =< B
1
P
1
> − < B
1
P
0
>; for k = 2, the
action of Σ splits ∆
2
into smaller simplexes as shown in Figure 6.
30 2. HOMOLOGY THEORY
d
d
d
d
d
d
d
d
d
d
d
d
r
r
r
r
r
r
r
r
r
r
r
r
e
e
e
e
e
e
e
e
e
e
e
e
B
2
P
0
P
1
P
2
M
1
M
2
M
0
Figure 6. The subdivision operator Σ splits ∆
2
into the chain
< B
2
, M
0
, P
2
> − < B
2
, M
0
, P
1
> − < B
2
, M
1
, P
2
> + < B
2
, M
1
, P
0
>
+ < B
2
, M
2
, P
1
> − < B
2
, M
2
, P
0
>
The deﬁnition of Σ and T for every topological space and every singular ksimplex
σ in X is
Σ(σ) = S
I
(σ)(Σ(δ
I
)), T(σ) = S
I+1
(σ)(T(δ
I
)).
Lemma 2.5. One has the identities
∂ ◦ Σ = Σ ◦ ∂, ∂ ◦ T +T ◦ ∂ = Id −Σ.
Proof. These identities are proved by direct computation (it is enough to consider
the case X = ∆
I
).
The ﬁrst identity tells us that Σ is a morphism of diﬀerential complexes, and the
second that T is a homotopy between Σ and Id, so that the morphism Σ
.
induced in
homology by Σ is an isomorphism.
The diameter of a singular ksimplex σ in R
a
is the maximum of the lengths of
the segments contained in σ. The proof of the following Lemma is an elementary
computation.
Lemma 2.6. Let σ =< E
0
, . . . , E
I
>, with E
0
, . . . , E
I
∈ R
a
. The diameter of every
simplex in the singular chain Σ(σ) ∈ S
I
(R
a
) is at most k/k + 1 times the diameter of
σ.
Proposition 2.7. Let X be a topological space, U = ¦U
α
¦ an open cover, and σ
a singular ksimplex in X. There is a natural number r > 0 such that every singular
simplex in Σ
·
(σ) is contained in a open set U
α
.
Proof. As ∆
I
is compact there is a real positive number such that σ maps a
neighbourhood of radius of every point of ∆
I
into some U
α
. Since
lim
·→+∞
k
·
(k + 1)
·
= 0
3. THE MAYERVIETORIS SEQUENCE 31
there is an r > 0 such that Σ
·
(δ
I
) is a linear combination of simplexes whose diameter
is less than . But as Σ
·
(σ) = S
I
(σ)(Σ
·
(δ
I
)) we are done.
This completes the proof of Proposition 2.2. We may now prove the exactness of
the MayerVietoris sequence in the following sense. If X = U ∪ V (union of two open
subsets), let U = ¦U, V ¦ and A = U ∩ V .
Proposition 2.8. For every k there is an exact sequence of Rmodules
0 → S
I
(A)
j
→ S
I
(U) ⊕S
I
(V )
j
→ S
U
I
(X) → 0 .
Proof. One has a diagram of inclusions
U
;
U
@
@
@
@
@
@
@
@
A
/
U
~
~
~
~
~
~
~
/
V
@
@
@
@
@
@
@
X
V
;
V
~
~
~
~
~
~
~
~
Deﬁning i(σ) = (
l
◦ σ, −
\
◦ σ) and p(σ
1
, σ
2
) = j
l
◦ σ
1
+j
\
◦ σ
2
, the exactness of the
MayerVietoris sequence is easily proved.
The morphisms i and p commute with the homology operator ∂, so that one obtains
a long homology exact sequence involving the homologies H
•
(A), H
•
(V ) ⊕ H
•
(V ) and
H
U
•
(X). But in view of Proposition 2.2 we may replace H
U
•
(X) with the homology
H
•
(X), so that we obtain the exact sequence
→ H
2
(A) → H
2
(U) ⊕H
2
(V ) → H
2
(X)
→ H
1
(A) → H
1
(U) ⊕H
1
(V ) → H
1
(X)
→ H
0
(A) → H
0
(U) ⊕H
0
(V ) → H
0
(X) → 0
Exercise 2.9. Prove that for any ring R the homology of the sphere S
a
with
coeﬃcients in R, n ≥ 2, is
H
I
(S
a
, R) =
R for k = 0 and k = n
0 for 0 < k < n and k > n.
Exercise 2.10. Show that the relative homology of S
2
mod S
1
with coeﬃcients in
Z is concentrated in degree 2, and H
2
(S
2
, S
1
) · Z ⊕Z.
Exercise 2.11. Use the MayerVietoris sequence to compute the homology of a
cylinder S
1
R minus a point with coeﬃcients in Z. (Hint: since the cylinder is
homotopic to S
1
, it has the same homology). The result is (calling X the space)
H
0
(X, Z) · Z, H
1
(X, Z) · Z ⊕Z, H
2
(X, Z) = 0 .
Compare this with the homology of S
2
minus three points.
32 2. HOMOLOGY THEORY
4. Excision
If a space X is the union of subspaces, the MayerVietoris suquence allows one to
compute the homology of X from the homology of the subspaces and of their intersec
tions. The operation of excision in some sense gives us information about the reverse
operation, i.e., it tells us what happen to the homology of a space if we “excise” a sub
pace out of it. Let us recall that given a map f : (X, A) → (Y, B) (i.e., a map f : X → Y
such that f(A) ⊂ B) there is natural morphism f
.
: H
•
(X, A) → H
•
(Y, B).
Definition 2.1. Given nested subspaces U ⊂ A ⊂ X, the inclusion map (X−U, A−
U) → (X, A) is said to be an excision if the induced morphism H
I
(X − U, A − U) →
H
I
(X, A) is an isomorphism for all k.
If (X −U, A−U) → (X, A) is an excision, we say that U “can be excised.”
To state the main theorem about excision we need some deﬁnitions from topology.
Definition 2.2. 1. Let i : A → X be an inclusion of topological spaces. A map
r : X → A is a retraction of i if r ◦ i = Id
¹
.
2. A subspace A ⊂ X is a deformation retract of X if Id
A
is homotopically equivalent
to i ◦ r, where r : X → A is a retraction.
If r : X → A is a retraction of i : A → X, then r
.
◦i
.
= Id
1
•
(¹)
, so that i
.
: H
•
(A) →
H
•
(X) is injective. Moreover, if A is a deformation retract of X, then H
•
(A) · H
•
(X).
The same notion can be given for inclusions of pairs, (A, B) → (X, Y ); if such a map is
a deformation retract, then H
•
(A, B) · H
•
(X, Y ).
Exercise 2.3. Show that no retraction S
a
→ S
a−1
can exist.
Theorem 2.4. If the closure U of U lies in the interior int(A) of A, then U can be
excised.
Proof. We consider the cover U = ¦X − U, int(A)¦ of X. Let c =
¸
;
a
;
σ
;
∈
Z
I
(X, A), so that ∂c ∈ S
I−1
(A). In view of Proposition 2.2 we may assume that c is U
small. If we cancel from σ those singular simplexes σ
;
taking values in int(A), the class
[c] ∈ H
I
(X, A) is unchanged. Therefore, after the removal, we can regard c as a relative
cycle in X−U mod A−U; this implies that the morphism H
I
(X−U, A−U) → H
I
(X, A)
is surjective.
To prove that it is injective, let [c] ∈ H
I
(X −U, A−U) be such that, regarding c as
a cycle in X mod A, it is a boundary, i.e., c ∈ B
I
(X, A). This means that
c = ∂b +c
with b ∈ S
I+1
(X), c
∈ S
I
(A) .
We apply the operator Σ
·
to both sides of this inequality, and split Σ
·
(b) into b
1
+ b
2
,
where b
1
maps into X −
¯
U and b
2
into int(A). We have
Σ
·
(c) −∂b
1
= Σ
·
(c
) +∂b
2
.
4. EXCISION 33
The chain in the left side is in X−U while the chain in the right side is in A; therefore,
both chains are in (X −U) ∩ A = A−U . Now we have
Σ
·
(c) = Σ
·
(c
) +∂b
2
+∂b
1
with Σ
·
(c
)+∂b
2
∈ S
I
(A−U) and ∂b
1
∈ S
I+1
(X−U) so that Σ
·
(c) ∈ B
I
(X−U, A−U),
which implies [c] = 0 (in H
I
(X −U, A−U)).
Exercise 2.5. Let B an open band around the equator of S
2
, and x
0
∈ B. Compute
the relative homology H
•
(S
2
−x
0
, B −x
0
; Z).
To describe some more applications of excision we need the notion of augmented
homology modules. Given a topological space X and a ring R, let us deﬁne
∂

: S
0
(X, R) → R
¸
;
a
;
σ
;
→
¸
;
a
;
.
We deﬁne the augmented homology modules
H

0
(X, R) = ker ∂

/B
0
(X, R) , H

I
(X, R) = H
I
(X, R) for k > 0 .
If A ⊂ X, one deﬁnes the augmented relative homology modules H

I
(X, A; R) in a
similar way, i.e.,
H

I
(X, A; R) = H
I
(X, A; R) if A = ∅, H

I
(X, A; R) = H
I
(X, R) if A = ∅ .
Exercise 2.6. Prove that there is a long exact sequence for the augmented relative
homology modules.
Exercise 2.7. Let B
a
be the closed unit ball in R
a+1
, S
a
its boundary, and let E
±
a
be the two closed (northern, southern) emispheres in S
a
.
1. Use the long exact sequence for the augmented relative homology modules to
prove that H

I
(S
a
) · H

I
(S
a
, E
−
a
) and H

I−1
(S
a−1
) · H

I
(B
a
, S
a−1
). So we have
H

I
(B
a
, S
a−1
) = 0 for k < n, H

a
(B
a
, S
a−1
) · R
2. Use excision to show that H

I
(S
a
, E
−
a
) · H

I
(B
a
, S
a−1
).
3. Deduce that H

I
(S
a
) · H

I−1
(S
a−1
).
Exercise 2.8. Let S
a
be the sphere realized as the unit sphere in R
a+1
, and let
r : S
a
→ S
a
→ S
a
be the reﬂection
r(x
0
, x
1
, . . . , x
a
) = (−x
0
, x
1
, . . . , x
a
).
34 2. HOMOLOGY THEORY
Prove that r
.
: H
a
(S
a
) → H
a
(S
a
) is the multiplication by −1. (Hint: this is trivial for
n = 0, and can be extended by induction using the commutativity of the diagram
H
a
(S
a
)
∼
·
H

a−1
(S
a−1
)
·
H
a
(S
a
)
∼
H

a−1
(S
a−1
)
which follows from Exercise 2.7.
Exercise 2.9. 1. The rotation group O(n + 1) acts on S
a
. Show that for any
M ∈ O(n + 1) the induced morphism M
.
: H
a
(S
a
) → H
a
(S
a
) is the multiplication by
det M = ±1.
2. Let a: S
a
→ S
a
be the antipodal map, a(x) = −x. Show that a
.
: H
a
(S
a
) →
H
a
(S
a
) is the multiplication by (−1)
a+1
.
Example 2.10. We show that the inclusion map (E
+
a
, S
a−1
) → (S
a
, E
−
a
) is an
excision. (Here we are excising the open southern emisphere, i.e., with reference to the
general theory, X = S
a
, U = the open southern emisphere, A = E
−
a
.)
The hypotheses of Theorem 2.4 are not satisﬁed. However it is enough to consider
the subspace
V =
¸
x ∈ S
a
[ x
0
> −
1
2
¸
.
V can be excised from (S
a
, E
−
a
). But (E
+
a
, S
a−1
) is a deformation retract of (S
a
−
V, E
−
a
−V ) so that we are done.
We end with a standard application of algebraic topology. Let us deﬁne a vector
ﬁeld on S
a
as a continous map v : S
a
→ R
a+1
such that v(x) x = 0 for all x ∈ S
a
(the
product is the standard scalar product in R
a+1
).
Proposition 2.11. A nowhere vanishing vector ﬁeld v on S
a
exists if and only if
n is odd.
Proof. If n = 2m+ 1 a nowhere vanishing vector ﬁeld is given by
v(x
0
, . . . , x
2n+1
) = (−x
1
, x
0
, −x
3
, x
2
, . . . , −x
2n+1
, x
2n
) .
Conversely, assume that such a vector ﬁeld exists. Deﬁne
w(x) =
v(x)
v(x)
;
this is a map S
a
→ S
a
, with w(x) x = 0 for all x ∈ S
a
. Deﬁne
F : S
a
I → S
a
F(x, t) = x cos tπ +w(x) sin tπ.
Since
F(x, 0) = x, F(x,
1
2
) = w(x), F(x, 1) = −x
4. EXCISION 35
the three maps Id, w, a are homotopic. But as a consequence of Exercise 2.9, n must
be odd.
CHAPTER 3
Introduction to sheaves and their cohomology
1. Presheaves and sheaves
Let X be a topological space.
Definition 3.1. A presheaf of Abelian groups on X is a rule
1
{ which assigns an
Abelian group {(U) to each open subset U of X and a morphism (called restriction map)
ϕ
l.\
: {(U) → {(V ) to each pair V ⊂ U of open subsets, so as to verify the following
requirements:
(1) {(∅) = ¦0¦;
(2) ϕ
l.l
is the identity map;
(3) if W ⊂ V ⊂ U are open sets, then ϕ
l.W
= ϕ
\.W
◦ ϕ
l.\
.
The elements s ∈ {(U) are called sections of the presheaf { on U. If s ∈ {(U) is
a section of { on U and V ⊂ U, we shall write s
\
instead of ϕ
l.\
(s). The restriction
{
l
of { to an open subset U is deﬁned in the obvious way.
Presheaves of rings are deﬁned in the same way, by requiring that the restriction
maps are ring morphisms. If { is a presheaf of rings on X, a presheaf ´ of Abelian
groups on X is called a presheaf of modules over { (or an {module) if, for each open
subset U, ´(U) is an {(U)module and for each pair V ⊂ U the restriction map
ϕ
l.\
: ´(U) → ´(V ) is a morphism of {(U)modules (where ´(V ) is regarded as
an {(U)module via the restriction morphism {(U) → {(V )). The deﬁnitions in this
Section are stated for the case of presheaves of Abelian groups, but analogous deﬁnitions
and properties hold for presheaves of rings and modules.
Definition 3.2. A morphism f : { → O of presheaves over X is a family of morph
isms of Abelian groups f
l
: {(U) → O(U) for each open U ⊂ X, commuting with the
1
This rather naive terminology can be made more precise by saying that a presheaf on X is a
contravariant functor from the category O
X
of open subsets of X to the category of Abelian groups.
O
X
is deﬁned as the category whose objects are the open subsets of X while the morphisms are the
inclusions of open sets.
37
38 3. SHEAVES AND THEIR COHOMOLOGY
restriction morphisms; i.e., the following diagram commutes:
{(U)
;
U
−−−−→ O(U)
ϕ
U,V
ϕ
U,V
{(V )
;
V
−−−−→ O(V )
Definition 3.3. The stalk of a presheaf { at a point x ∈ X is the Abelian group
{
a
= lim
−→
l
{(U)
where U ranges over all open neighbourhoods of x, directed by inclusion.
Remark 3.4. We recall here the notion of direct limit. A directed set I is a partially
ordered set such that for each pair of elements i, j ∈ I there is a third element k such
that i < k and j < k. If I is a directed set, a directed system of Abelian groups is
a family ¦G
j
¦
j∈1
of Abelian groups, such that for all i < j there is a group morphism
f
j;
: G
j
→ G
;
, with f
jj
= id and f
j;
◦ f
;I
= f
jI
. On the set G =
¸
j∈1
G
j
, where
¸
denotes disjoint union, we put the following equivalence relation: g ∼ h, with g ∈ G
j
and h ∈ G
;
, if there exists a k ∈ I such that f
jI
(g) = f
;I
(h). The direct limit l of the
system ¦G
j
¦
j∈1
, denoted l = lim
−→j∈1
G
j
, is the quotient G/ ∼. Heuristically, two elements
in G represent the same element in the direct limit if they are ‘eventually equal.’ From
this deﬁnition one naturally obtains the existence of canonical morphisms G
j
→ l. The
following discussion should make this notion clearer; for more detail, the reader may
consult [12].
If x ∈ U and s ∈ {(U), the image s
a
of s in {
a
via the canonical projection
{(U) → {
a
(see footnote) is called the germ of s at x. From the very deﬁnition of direct
limit we see that two elements s ∈ {(U), s
∈ {(V ), U, V being open neighbourhoods
of x, deﬁne the same germ at x, i.e. s
a
= s
a
, if and only if there exists an open
neighbourhood W ⊂ U ∩ V of x such that s and s
coincide on W, s
W
= s
W
.
Definition 3.5. A sheaf on a topological space X is a presheaf T on X which fulﬁlls
the following axioms for any open subset U of X and any cover ¦U
j
¦ of U.
S1) If two sections s ∈ T(U), ¯ s ∈ T(U) coincide when restricted to any U
j
, s
l
i
=
¯ s
l
i
, they are equal, s = ¯ s.
S2) Given sections s
j
∈ T(U
j
) which coincide on the intersections, s
jl
i
∩l
j
=
s
;
l
i
∩l
j
for every i, j, there exists a section s ∈ T(U) whose restriction to
each U
j
equals s
j
, i.e. s
l
i
= s
j
.
Thus, roughly speaking, sheaves are presheaves deﬁned by local conditions. The
stalk of a sheaf is deﬁned as in the case of a presheaf.
1. PRESHEAVES AND SHEAVES 39
Example 3.6. If T is a sheaf, and T
a
= ¦0¦ for all x ∈ X, then T is the zero sheaf,
T(U) = ¦0¦ for all open sets U ⊂ X. Indeed, if s ∈ T(U), since s
a
= 0 for all x ∈ U,
there is for each x ∈ U an open neighbourhood U
a
such that s
l
x
= 0. The ﬁrst sheaf
axiom then implies s = 0. This is not true for a presheaf, cf. Example 3.14 below.
A morphism of sheaves is just a morphism of presheaves. If f : T → ( is a morphism
of sheaves on X, for every x ∈ X the morphismf induces a morphism between the stalks,
f
a
: T
a
→ (
a
, in the following way: since the stalk T
a
is the direct limit of the groups
T(U) over all open U containing x, any g ∈ T
a
is of the form g = s
a
for some open
U ÷ x and some s ∈ T(U); then set f
a
(g) = (f
l
(s))
a
.
A sequence of morphisms of sheaves 0 → T
→ T → T
→ 0 is exact if for every
point x ∈ X, the sequence of morphisms between the stalks 0 → T
a
→ T
a
→ T
a
→ 0 is
exact. If 0 → T
→ T → T
→ 0 is an exact sequence of sheaves, for every open subset
U ⊂ X the sequence of groups 0 → T
(U) → T(U) → T
(U) is exact, but the last
arrow may fail to be surjective. An instance of this situation is contained in Example
3.11 below.
Exercise 3.7. Let 0 → T
→ T → T
→ 0 be an exact sequence of sheaves. Show
that 0 → T
→ T → T
is an exact sequence of presheaves.
Example 3.8. Let G be an Abelian group. Deﬁning {(U) ≡ G for every open
subset U and taking the identity maps as restriction morphisms, we obtain a presheaf,
called the constant presheaf
˜
G
A
. All stalks (
˜
G
A
)
a
of
˜
G
A
are isomorphic to the group
G. The presheaf
˜
G
A
is not a sheaf: if V
1
and V
2
are disjoint open subsets of X, and
U = V
1
∪V
2
, the sections g
1
∈
˜
G
A
(V
1
) = G, g
2
∈
˜
G
A
(V
2
) = G, with g
1
= g
2
, satisfy the
hypothesis of the second sheaf axiom S2) (since V
1
∩V
2
= ∅ there is nothing to satisfy),
but there is no section g ∈
˜
G
A
(U) = G which restricts to g
1
on V
1
and to g
2
on V
2
.
Example 3.9. Let (
A
(U) be the ring of realvalued continuous functions on an open
set U of X. Then (
A
is a sheaf (with the obvious restriction morphisms), the sheaf of
continuous functions on X. The stalk (
a
≡ ((
A
)
a
at x is the ring of germs of continuous
functions at x.
Example 3.10. In the same way one can deﬁne the following sheaves:
The sheaf (
∞
A
of diﬀerentiable functions on a diﬀerentiable manifold X.
The sheaves Ω
j
A
of diﬀerential pforms, and all the sheaves of tensor ﬁelds on a
diﬀerentiable manifold X.
The sheaf of holomorphic functions on a complex manifold and the sheaves of holo
morphic pforms on it.
The sheaves of forms of type (p, q) on a complex manifold X.
Example 3.11. Let X be a diﬀerentiable manifold, and let d: Ω
•
A
→ Ω
•
A
be the
exterior diﬀerential. We can deﬁne the presheaves Z
j
A
of closed diﬀerential pforms, and
40 3. SHEAVES AND THEIR COHOMOLOGY
B
j
A
of exact pdiﬀerential forms,
Z
j
A
(U) = ¦ω ∈ Ω
j
A
(U) [ dω = 0¦,
B
j
A
(U) = ¦ω ∈ Ω
j
A
(U) [ ω = dτ for some τ ∈ Ω
j−1
A
(U)¦.
Z
j
A
is a sheaf, since the condition of being closed is local: a diﬀerential form is closed if
and only if it is closed in a neighbourhood of each point of X. On the contrary, B
j
A
is
not a sheaf. In fact, if X = R
2
, the presheaf B
1
A
of exact diﬀerential 1forms does not
fulﬁll the second sheaf axiom: consider the form
ω =
xdy −ydx
x
2
+y
2
deﬁned on the open subset U = X − ¦(0, 0)¦. Since ω is closed on U, there is an
open cover ¦U
j
¦ of U by open subsets where ω is an exact form, ω
l
i
∈ B
1
A
(U
j
) (this is
Poincar´e’s lemma). But ω is not an exact form on U because its integral along the unit
circle is diﬀerent from 0.
This means that, while the sequence of sheaf morphisms 0 → R → (
∞
A
o
−−→Z
1
A
→ 0
is exact (Poincar´e lemma), the morphism (
∞
A
(U)
o
−−→Z
1
A
(U) may fail to be surjective.
1.1.
´
Etal´e space. We wish now to describe how, given a presheaf, one can natur
ally associate with it a sheaf having the same stalks. As a ﬁrst step we consider the case
of a constant presheaf
˜
G
A
on a topological space X, where G is an Abelian group. We
can deﬁne another presheaf G
A
on X by putting G
A
(U) = ¦locally constant functions
f : U → G¦,
2
where
˜
G
A
(U) = G is included as the constant functions. It is clear that
(G
A
)
a
= G
a
= G at each point x ∈ X and that G
A
is a sheaf, called the constant sheaf
with stalk G. Notice that the functions f : U → G are the sections of the projection
π:
¸
a∈A
G
a
→ X and the locally constant functions correspond to those sections which
locally coincide with the sections produced by the elements of G.
Now, let { be an arbitrary presheaf on X. Consider the disjoint union of the stalks
{ =
¸
a∈A
{
a
and the natural projection π: { → X. The sections s ∈ {(U) of the
presheaf { on an open subset U produce sections s: U → { of π, deﬁned by s(x) = s
a
,
and we can deﬁne a new presheaf {
r
by taking {
r
(U) as the group of those sections
σ: U → { of π such that for every point x ∈ U there is an open neighbourhood V ⊂ U
of x which satisﬁes σ
\
= s for some s ∈ {(V ).
That is, {
r
is the presheaf of all sections that locally coincide with sections of {. It
can be described in another way by the following construction.
Definition 3.12. The set {, endowed with the topology whose base of open subsets
consists of the sets s(U) for U open in X and s ∈ {(U), is called the ´etal´e space of the
presheaf {.
2
A function is locally constant on U if it is constant on any connected component of U.
1. PRESHEAVES AND SHEAVES 41
Exercise 3.13. (1) Show that π: { → X is a local homeomorphism, i.e., every
point u ∈ { has an open neighbourhood U such that π: U → π(U) is a
homeomorphism.
(2) Show that for every open set U ⊂ X and every s ∈ {(U), the section s: U → {
is continuous.
(3) Prove that {
r
is the sheaf of continuous sections of π: { → X.
(4) Prove that for all x ∈ X the stalks of { and {
r
at x are isomorphic.
(5) Show that there is a presheaf morphism φ: { → {
r
.
(6) Show that φ is an isomorphism if and only if { is a sheaf.
{
r
is called the sheaf associated with the presheaf {. In general, the morphism
φ: { → {
r
is neither injective nor surjective: for instance, the morphism between the
constant presheaf
˜
G
A
and its associated sheaf G
A
is injective but nor surjective.
Example 3.14. As a second example we study the sheaf associated with the presheaf
B
I
A
of exact kforms on a diﬀerentiable manifold X. For any open set U we have an
exact sequence of Abelian groups (actually of Rvector spaces)
0 → B
I
A
(U) → Z
I
A
(U) → H
I
A
(U) → 0
where H
I
A
is the presheaf that with any open set U associates its kth de Rham cohomo
logy group, H
I
A
(U) = H
I
11
(U). Now, the open neighbourhoods of any point x ∈ X
which are diﬀeomorphic to R
a
(where n = dimX) are coﬁnal
3
in the family of all open
neighbourhoods of x, so that (H
I
A
)
a
= 0 by the Poincar´e lemma. In accordance with
Example 3.6 this means that (H
I
A
)
r
= 0, which is tantamount to (B
I
A
)
r
· Z
I
A
.
In this case the natural morhism H
I
A
→ (H
I
A
)
r
is of course surjective but not
injective. On the other hand, B
I
A
→ (B
I
A
)
r
= Z
I
A
is injective but not surjective.
Definition 3.15. Given a sheaf T on a topological space X and a subset (not
necessarily open) S ⊂ X, the sections of the sheaf T on S are the continuous sections
σ: S → T of π: T → X. The group of such sections is denoted by Γ(S, T).
Definition 3.16. Let {, O be presheaves on a topological space X.
4
(1) The direct sum of { and O is the presheaf { ⊕ O given, for every open subset
U ⊂ X, by ({ ⊕O)(U) = {(U) ⊕O(U) with the obvious restriction morphisms.
3
Let I be a directed set. A subset J of I is said to be coﬁnal if for any i ∈ I there is a j ∈ J
such that i < j. By the deﬁnition of direct limit we see that, given a directed family of Abelian groups
{G
i
}
i∈I
, if {G
j
}
j∈J
is the subfamily indexed by J, then
lim
−→
i∈I
G
i
lim
−→
j∈J
G
j
;
that is, direct limits can be taken over coﬁnal subsets of the index set.
4
Since we are dealing with Abelian groups, i.e. with Zmodules, the Hom modules and tensor
products are taken over Z.
42 3. SHEAVES AND THEIR COHOMOLOGY
(2) For any open set U ⊂ X, let us denote by Hom({
l
, O
l
) the space of morph
isms between the restricted presheaves {
l
and O
l
; this is an Abelian group in a nat
ural manner. The presheaf of homomorphisms is the presheaf Hom({, O) given by
Hom({, O)(U) = Hom({
l
, O
l
) with the natural restriction morphisms.
(3) The tensor product of { and O is the presheaf ({ ⊗O)(U) = {(U) ⊗O(U).
If T and ( are sheaves, then the presheaves T ⊕ ( and Hom(T, () are sheaves.
On the contrary, the tensor product of T and ( previously deﬁned may not be a sheaf.
Indeed one deﬁnes the tensor product of the sheaves T and ( as the sheaf associated
with the presheaf U → T(U) ⊗((U).
It should be noticed that in general Hom(T, ()(U) · Hom(T(U), ((U)) and
Hom(T, ()
a
· Hom(T
a
, (
a
).
1.2. Direct and inverse images of presheaves and sheaves. Here we study
the behaviour of presheaves and sheaves under change of base space. Let f : X → Y be
a continuous map.
Definition 3.17. The direct image by f of a presheaf { on X is the presheaf f
∗
{
on Y deﬁned by (f
∗
{)(V ) = {(f
−1
(V )) for every open subset V ⊂ Y . If T is a sheaf
on X, then f
∗
T turns out to be a sheaf.
Let { be a presheaf on Y .
Definition 3.18. The inverse image of { by f is the presheaf on X deﬁned by
U → lim
−→
l⊂;
−1
(\ )
{(V ).
The inverse image sheaf of a sheaf T on Y is the sheaf f
−1
T associated with the inverse
image presheaf of T.
The stalk of the inverse image presheaf at a point x ∈ X is isomorphic to {
;(a)
.
It follows that if 0 → T
→ T → T
→ 0 is an exact sequence of sheaves on Y , the
induced sequence
0 → f
−1
T
→ f
−1
T → f
−1
T
→ 0
of sheaves on X, is also exact (that is, the inverse image functor for sheaves of Abelian
groups is exact).
The ´etal´e space f
−1
T of the inverse image sheaf is the ﬁbred product
5
Y
A
T. It
follows easily that the inverse image of the constant sheaf G
A
on X with stalk G is the
constant sheaf G
Y
with stalk G, f
−1
G
A
= G
Y
.
5
For a deﬁnition of ﬁbred product see e.g. [15].
2. COHOMOLOGY OF SHEAVES 43
2. Cohomology of sheaves
We wish now to describe a cohomology theory which associates cohomology groups
to a sheaf on a topological space X.
2.1.
ˇ
Cech cohomology. We start by considering a presheaf { on X and an open
cover U of X. We assume that U is labelled by a totally ordered set I, and deﬁne
U
j
0
...j
p
= U
j
0
∩ ∩ U
j
p
.
We deﬁne the
ˇ
Cech complex of U with coeﬃcients in { as the complex whose pth term
is the Abelian group
ˇ
C
j
(U, {) =
¸
j
0
<···<j
p
{(U
j
0
...j
p
) .
Thus a pcochain α is a collection ¦α
j
0
...j
p
¦ of sections of {, each one belonging to the
space of sections over the intersection of p + 1 open sets in U. Since the indexes of the
open sets are taken in strictly increasing order, each intersection is counted only once.
The
ˇ
Cech diﬀerential δ :
ˇ
C
j
(U, {) →
ˇ
C
j+1
(U, {) is deﬁned as follows: if α = ¦α
j
0
...j
p
¦
∈
ˇ
C
j
(U, {), then
¦(δα)
j
0
...j
p+1
¦ =
j+1
¸
I=0
(−1)
I
α
j
0
...bı
k
...j
p+1
l
i
0
...i
p+1
.
Here a caret denotes omission of the index. For instance, if p = 0 we have α = ¦α
j
¦ and
(3.1) (δα)
jI
= α
Il
i
∩l
k
−α
jl
i
∩l
k
.
It is an easy exercise to check that δ
2
= 0. Thus we obtain a cohomology theory. We
denote the corresponding cohomology groups by H
I
(U, {).
Lemma 3.1. If T is a sheaf, one has an isomorphism H
0
(U, T) · T(X)
Proof. We have H
0
(U, T) = ker δ :
ˇ
C
0
(U, {) →
ˇ
C
1
(U, {). So if α ∈ H
0
(U, T) by
(3.1) we see that
α
Il
i
∩l
k
= α
jl
i
∩l
k
.
By the second sheaf axiom this implies that there is a global section ˜ α ∈ T(X) such
that ˜ α
l
i
= α
j
. This yields a morphism H
0
(U, T) → T(X), which is evidently surjective
and is injective because of the ﬁrst sheaf axiom.
Example 3.2. We consider an open cover U of the circle S
1
formed by three sets
which intersect only pairwise. We compute the
ˇ
Cech cohomology of U with coeﬃcients
in the constant sheaf R. We have C
0
(U, R) = C
1
(U, R) = R ⊕ R ⊕ R, C
I
(U, R) =
0 for k > 1 because there are no triple intersections. The only nonzero diﬀerential
d
0
: C
0
(U, R) → C
1
(U, R) is given by
d
0
(x
0
, x
1
, x
2
) = (x
1
−x
2
, x
2
−x
0
, x
0
−x
1
).
44 3. SHEAVES AND THEIR COHOMOLOGY
Hence
H
0
(U, R) = ker d
0
· R
H
1
(U, R) = C
1
(U, R)/ Imd
0
· R.
It is possible to deﬁne
ˇ
Cech cohomology groups depending only on the pair (X, T),
and not on a cover, by letting
H
I
(X, T) = lim
−→
U
ˇ
H
I
(U, T).
The direct limit is taken over a coﬁnal subset of the directed set of all covers of X (the
order is of course the reﬁnement of covers: a cover V = ¦V
;
¦
;∈J
is a reﬁnement of U if
there is a map f : I → J such that V
;(j)
⊂ U
j
for every i ∈ I). The order must be ﬁxed
at the outset, since a cover may be regarded as a reﬁnement of another in many ways.
As diﬀerent coﬁnal families give rise to the same inductive limit, the groups H
I
(X, T)
are well deﬁned.
2.2. Fine sheaves.
ˇ
Cech cohomology is wellbehaved when the base space X is
paracompact. (It is indeed the bad behaviour of
ˇ
Cech cohomology on nonparacompact
spaces which motivated the introduction of another cohomology theory for sheaves,
usually called sheaf cohomology; cf. [5].) In this and in the following sections we consider
some properties of
ˇ
Cech cohomology that hold in that case.
Definition 3.3. A sheaf of rings { on a topological space X is ﬁne if, for any
locally ﬁnite oper cover U = ¦U
j
¦
j∈1
of X, there is a family ¦s
j
¦
j∈1
of global sections of
{ such that:
(1)
¸
j∈1
s
j
= 1;
(2) for every i ∈ I there is a closed subset S
j
⊂ U
j
such that (s
j
)
a
= 0 whenever
x / ∈ S
j
.
The family ¦s
j
¦ is called a partition of unity subordinated to the cover U. For
instance, the sheaf of continuous functions on a paracompact topological space as well
as the sheaf of smooth functions on a diﬀerentiable manifold are ﬁne, while sheaves of
complex or real analytic functions are not.
Definition 3.4. A sheaf T of Abelian groups on a topological space X is said to be
acyclic if H
I
(X, T) = 0 for k > 0.
Proposition 3.5. Let { be a ﬁne sheaf of rings on a paracompact space X. Every
sheaf ´ of {modules is acyclic.
2. COHOMOLOGY OF SHEAVES 45
Proof. Let U = ¦U
j
¦
j∈1
be a locally ﬁnite open cover of X, and let ¦ρ
j
¦ be a
partition of unity of { subordinated to U. For any α ∈
ˇ
C
o
(U, ´) with q > 0 we set
(Kα)
j
0
...j
q−1
=
¸
;∈1
;<j
0
ρ
;
a
;j
0
...j
q−1
−
¸
;∈1
j
0
<;<j
1
ρ
;
a
j
0
;j
1
...j
q−1
+. . .
=
o
¸
I=0
(−1)
I
¸
;∈1
j
k−1
<;<j
k
ρ
;
a
j
0
...j
k−1
;j
k
...j
q−1
.
This deﬁnes a morphism K:
ˇ
C
I
(U, ´) →
ˇ
C
I−1
(U, ´) such that δK + Kδ = id (i.e.,
K is a homotopy operator); then α = δKα if δα = 0, so that H
I
(U, ´) = 0 for k > 0.
Since on a paracompact space the locally ﬁnite open covers are coﬁnal in the family
of all covers, we can take direct limit on such covers, thus getting H
I
(X, ´) = 0 for
k > 0.
Example 3.6. Using this result we may recast the proof of the exactness of the
MayerVietoris sequence for de Rham cohomology in a slightly diﬀerent form. Given a
diﬀerentiable manifold X, let U be the open cover formed by two sets U and V . Since
ˇ
C
2
(U, Ω
I
) = 0 (there are no triple intersections) we have an exact sequence
0 → H
0
(U, Ω
I
) →
ˇ
C
0
(U, Ω
I
)
δ
→
ˇ
C
1
(U, Ω
I
) → 0 .
which in principle is exact everywhere but at C
1
(U, Ω
I
). However since the sheaves Ω
I
are acyclic by Proposition 3.5, one has H
1
(U, Ω
I
) = 0, which means that δ is surjective,
and the sequence is exact at that place as well. We have the identiﬁcations
H
0
(U, Ω
I
) = Ω
I
(X), C
0
(U, Ω
I
) = Ω
I
(U) ⊕Ω
I
(V ), C
1
(U, Ω
I
) = Ω
I
(U ∩ V )
so that we obtain the exactness of the MayerVietoris sequence.
2.3. Long exact sequences in
ˇ
Cech cohomology. We wish to show that when
X is paracompact, any exact sequence of sheaves induces a corresponding long exact
sequence in
ˇ
Cech cohomology.
Lemma 3.7. Let X be any topological space, and let
(3.2) 0 → {
→ { → {
→ 0
be an exact sequence of presheaves on X. Then one has a long exact sequence
0 → H
0
(X, {
) → H
0
(X, {) → H
0
(X, {
) → H
1
(X, {
) → . . .
→ H
I
(X, {
) → H
I
(X, {) → H
I
(X, {
) → H
I+1
(X, {
) → . . .
46 3. SHEAVES AND THEIR COHOMOLOGY
Proof. For any open cover U the exact sequence (3.2) induces an exact sequence
of diﬀerential complexes
0 →
ˇ
C
•
(U, {
) →
ˇ
C
•
(U, {) →
ˇ
C
•
(U, {
) → 0
which induces the long cohomology sequence
0 → H
0
(U, {
) → H
0
(U, {) → H
0
(U, {
) → H
1
(U, {
) → . . .
→ H
I
(U, {
) → H
I
(U, {) → H
I
(U, {
) → H
I+1
(U, {
) → . . .
Since the direct limit of a family of exact sequences yields an exact sequence, by taking
the direct limit over the open covers of X one obtains the required exact sequence.
Lemma 3.8. Let X be a paracompact topological space, { a presheaf on X whose
associated sheaf is the zero sheaf, let U be an open cover of X, and let α ∈
ˇ
C
I
(U, {).
There is a reﬁnement W of U such that τ(α) = 0, where τ :
ˇ
C
I
(U, {) →
ˇ
C
I
(W, {) is
the morphism induced by restriction.
Proof. The proof relies on a standard paracompactness argument. See [13] '2.9.
Proposition 3.9. Let { be a presheaf on a paracompact space X, and let {
r
be the
associated sheaf. For all k ≥ 0, the natural morphism H
I
(X, {) → H
I
(X, {
r
) is an
isomorphism.
Proof. One has an exact sequence of presheaves
0 → O
1
→ { → {
r
→ O
2
→ 0
with
(3.3) O
r
1
= O
r
2
= 0 .
This gives rise to
(3.4) 0 → O
1
→ { → T → 0 , 0 → T → {
r
→ O
2
→ 0
where T is the quotient presheaf {/O
1
, i.e. the presheaf U → {(U)/O
1
(U). By Lemma
3.8 the isomorphisms (3.3) yield H
I
(X, O
1
) = H
I
(X, O
2
) = 0. Then by taking the long
exact sequences of cohomology from the exact sequences (3.4) we obtain the desired
isomorphism.
Using these results we may eventually prove that on paracompact spaces one has
long exact sequences in
ˇ
Cech cohomology.
Theorem 3.10. Let
0 → T
→ T → T
→ 0
2. COHOMOLOGY OF SHEAVES 47
be an exact sequence of sheaves on a paracompact space X. There is a long exact
sequence of
ˇ
Cech cohomology groups
0 → H
0
(X, T
) → H
0
(X, T) → H
0
(X, T
) → H
1
(X, T
) → . . .
→ H
I
(X, T
) → H
I
(X, T) → H
I
(X, T
) → H
I+1
(X, T
) → . . .
Proof. Let { be the quotient presheaf T/T
; then {
r
· T
. One has an exact
sequence of presheaves
0 → T
→ T → { → 0 .
By taking the associated long exact sequence in cohomology (cf. Lemma 3.7) and using
the isomorphism H
I
(X, {) = H
I
(X, T
) one obtains the required exact sequence.
2.4. Abstract de Rham theorem. We describe now a very useful way of com
puting cohomology groups; this result is sometimes called “abstract de Rham theorem.”
As a particular case it yields one form of the socalled de Rham theorem, which states
that the de Rham cohomology of a diﬀerentiable manifold and the
ˇ
Cech cohomology of
the constant sheaf R are isomorphic.
Definition 3.11. Let T be a sheaf of abelian groups on X. A resolution of T is a
collection of sheaves of abelian groups ¦L
I
¦
I∈N
with morphisms i : T → L
0
, d
I
: L
I
→
L
I+1
such that the sequence
0 → T
j
→ L
0
o
0
→ L
1
o
1
→ . . .
is exact. If the sheaves L
•
are acyclic (ﬁne) the resolution is said to be acyclic (ﬁne).
Lemma 3.12. If 0 → T → L
•
is a resolution, the morphism i
A
: T(X) → L
0
(X) is
injective.
Proof. Let O be the quotient L
0
/T. Then the sequence of sheaves
0 → T → L
0
→ O → 0
is exact. By Exercise 3.7, the sequence of abelian groups
0 → T(X) → L
0
(X) → O(X)
is exact. This implies the claim.
However the sequence of abelian groups
0 → L
0
(X)
o
0
−−→L
1
(X)
o
1
−−→ . . .
is not exact. We shall consider its cohomology H
•
(L
•
(X), d). By the previous Lemma
we have H
0
(L
•
(X), d) · H
0
(X, T).
Theorem 3.13. If 0 → T → L
•
is an acyclic resolution there is an isomorphism
H
I
(X, T) · H
I
(L
•
(X), d) for all k ≥ 0.
48 3. SHEAVES AND THEIR COHOMOLOGY
Proof. Deﬁne O
I
= ker d
I
: L
I
→ L
I+1
. The resolution may be split into
0 → T → L
0
A
→ O
1
→ 0 , 0 → O
I
→ L
I
→ O
I+1
→ 0 . k ≥ 1
Since the sheaves L
I
are acyclic by taking the long exact sequences of cohomology we
obtain a chain of isomorphisms
H
I
(X, T) · H
I−1
(X, O
1
) · · H
1
(X, O
I−1
) ·
H
0
(X, O
I
)
ImH
0
(X, L
I−1
)
By Exercise 3.7 H
0
(X, O
I
) = O
I
(X) is the kernel of d
I
: L
I
(X) → L
I+1
so that the
claim is proved.
Corollary 3.14. (de Rham theorem.) Let X be a diﬀerentiable manifold. For all
k ≥ 0 the cohomology groups H
I
11
(X) and H
I
(X, R) are isomorphic.
Proof. Let n = dimX. The sequence
(3.5) 0 → R → Ω
0
A
o
−−→Ω
1
A
o
−−→ → Ω
a
A
→ 0
(where Ω
0
A
≡ (
∞
A
) is exact (this is Poincar´e’s lemma). Moreover the sheaves Ω
•
A
are
modules over the ﬁne sheaf of rings (
∞
A
, hence are acyclic. The claim then follows for
the previous theorem.
Corollary 3.15. Let U be a subset of a diﬀerentiable manifold X which is diﬀeo
morphic to R
a
. Then H
I
(U, R) = 0 for k > 0.
2.5. Soft sheaves. For later use we also introduce and study the notion of soft
sheaf. However, we do not give the proofs of most claims, for which the reader is referred
to [2, 5, 22]. The contents of this subsection will only be used in Section 4.5.
Definition 3.16. Let T be a sheaf a on a topological space X, and let U ⊂ X be
a closed subset of X. The space T(U) (called “the space of sections of T over U”) is
deﬁned as
T(U) = lim
−→
\ ⊃l
T(V )
where the direct limit is taken over all open neighbourhoods V of U.
A consequence of this deﬁnition is the existence of a natural restriction morphism
T(X) → T(U).
Definition 3.17. A sheaf T is said to be soft if for every closed subset U ⊂ X the
restriction morphism T(X) → T(U) is surjective.
Proposition 3.18. If 0 → T
→ T → T
→ 0 is an exact sequence of soft sheaves
on a paracompact space X, for every open subset U ⊂ X the sequence of groups
0 → T
(U) → T(U) → T
(U) → 0
is exact.
2. COHOMOLOGY OF SHEAVES 49
Proof. One can e.g. adapt the proof of Proposition II.1.1 in [2].
Corollary 3.19. The quotient of two soft sheaves on a paracompact space is soft.
Proposition 3.20. Any soft sheaf of rings { on a paracompact space is ﬁne.
Proof. Cf. Lemma II.3.4 in [2].
Proposition 3.21. Every sheaf T on a paracompact space admits soft resolutions.
Proof. Let o
0
(T) be the sheaf of discontinuous sections of T (i.e., the sheaf of
all sections of the sheaf space T). The sheaf o
0
(T) is obviously soft. Now we have an
exact sequence 0 → T → o
0
(T) → T
1
→ 0. The sheaf T
1
is not soft in general, but it
may embedded into the soft sheaf o
0
(T
1
), and we have an exact sequence 0 → T
1
→
o
0
(T
1
) → T
2
→ 0. Upon iteration we have exact sequences
0 → T
I
j
k
−−→o
I
(T)
j
k
−−→T
I+1
→ 0
where o
I
(T) = o
0
(T
I
). One can check that the sequence of sheaves
0 → T → o
0
(T)
;
0
−−→o
1
(T)
;
1
−−→ . . .
(where f
I
= i
I+1
◦ p
I
) is exact.
Proposition 3.22. If T is a sheaf on a paracompact space, the sheaf o
0
(T) is
acyclic.
Proof. The endomorphism sheaf cnd(o
0
(T)) is soft, hence ﬁne by Proposition
3.20. Since o
0
(T) is an cnd(o
0
(T))module, it is acyclic.
6
Proposition 3.23. On a paracompact space soft sheaves are acyclic.
Proof. If T is a soft sheaf, the sequence 0 → T(X) → o
0
T(X) → T
1
(X) → 0
obtained from 0 → T → o
0
T → T
1
→ 0 is exact (Proposition 3.18). Since T and
o
0
T are soft, so is T
1
by Corollary 3.19, and the sequence 0 → T
1
(X) → o
1
T(X) →
T
2
(X) → 0 is also exact. With this procedure we can show that the complex o
•
(T)(X)
is exact. But since all sheaves o
•
(T) are acyclic by the previous Proposition, by the
abstract de Rham theorem the claim is proved.
Note that in this way we have shown that for any sheaf T on a paracompact space
there is a canonical soft resolution.
6
We are cheating a little bit, since the sheaf of rings End(S
0
(F)) is not commutative. However a
closer inspection of the proof would show that it works anyways.
50 3. SHEAVES AND THEIR COHOMOLOGY
2.6. Leray’s theorem for
ˇ
Cech cohomology. If an open cover U of a topolo
gical space X is suitably chosen, the
ˇ
Cech cohomologies H
•
(U, T) and H
•
(X, T) are
isomorphic. Leray’s theorem establishes a suﬃcient condition for such an isomorphism
to hold. Since the cohomology H
•
(U, T) is in generally much easier to compute, this
turns out to be a very useful tool in the computation of
ˇ
Cech cohomology groups.
We say that an open cover U = ¦U
j
¦
j∈1
of a topological space X is acyclic for a sheaf
T if H
I
(U
j
0
...j
p
, T) = 0 for all k > 0 and all nonvoid intersections U
j
0
...j
p
= U
j
0
∩ ∩U
j
p
,
i
0
. . . i
j
∈ I.
Theorem 3.24. (Leray’s theorem) Let T be a sheaf on a paracompact space X, and
let U be an open cover of X which is acyclic for T and is indexed by an ordered set.
Then, for all k ≥ 0, the cohomology groups H
I
(U, T) and H
I
(X, T) are isomorphic.
To prove this theorem we need to construct the socalled
ˇ
Cech sheaf complex. For
every nonvoid intersection U
j
0
...j
p
let j
j
0
...j
p
: U
j
0
...j
p
→ X be the inclusion. For every p
deﬁne the sheaf
ˇ
(
j
(U, T) =
¸
j
0
<···<j
p
(j
j
0
...j
p
)
∗
T
l
i
0
...i
p
(every factor (j
j
0
...j
p
)
∗
T
l
i
0
...i
p
is the sheaf T ﬁrst restricted to U
j
0
...j
p
and the exteded by
zero to the whole of X). The
ˇ
Cech diﬀerential induces sheaf morphisms δ :
ˇ
(
j
(U, T) →
ˇ
(
j+1
(U, T). From the deﬁnition, we get isomorphisms
(3.6)
ˇ
(
j
(U, T)(X) ·
ˇ
C
j
(U, T) ,
i.e., by taking global sections of the
ˇ
Cech sheaf complex we get the
ˇ
Cech cochain group
complex. Moreover we have:
Lemma 3.25. For all p and k,
H
I
(X,
ˇ
(
j
(U, T)) ·
¸
j
0
<···<j
p
H
I
(U
j
0
...j
p
, T) .
We may now prove Leray’s theorem. It is not diﬃcult to prove that the complex
ˇ
(
•
(U, T) is a resolution of T (cf. [2], Prop. II.3.3). Under the hypothesis of Leray’s
theorem, by Lemma 3.25 this resolution is acyclic. By the abstract de Rham theorem,
the cohomology of the global sections of the resolution is isomorphic to the cohomology
of T. But, due to the isomorphisms (3.6), the cohomology of the global sections of the
resolution is the cohomology H
•
(U, T).
2.7. Good covers. By means of Leray’s theorem we may reduce the problem of
computing the
ˇ
Cech cohomology of a diﬀerentiable manifold with coeﬃcients in the
constant sheaf R (which, via de Rham theorem, amounts to computing its de Rham
cohomology) to the computation of the cohomology of a cover with coeﬃcients in R;
thus a problem which in principle would need the solution of diﬀerential equations on
2. COHOMOLOGY OF SHEAVES 51
topologically nontrivial manifolds is reduced to a simpler problem which only involves
the intersection pattern of the open sets of a cover.
Definition 3.26. A locally ﬁnite open cover U of a diﬀerentiable manifold is good
if all nonempty intersections of its members are diﬀeomorphic to R
a
.
Good covers exist on any diﬀerentiable manifold (cf. [17]). Due to Corollary 3.15,
good covers are acyclic for the constant sheaf R. We have therefore
Proposition 3.27. For any good cover U of a diﬀerentiable manifold X one has
isomorphisms
H
I
(U, R) · H
I
(X, R) , k ≥ 0 .
The cover of Example 3.2 was good, so we computed there the de Rham cohomology
of the circle S
1
.
2.8. Flabby sheaves. Another kind of sheaves which can be introduced is that of
ﬂabby sheaves (also called “ﬂasque”). A sheaf T on a topological space X is said to
be ﬂabby if for every open subset U ⊂ X the restriction morphism T(X) → T(U) is
surjective. It is easy to prove that ﬂabby sheaves are soft: if U ⊂ X is a closed subset,
by deﬁnition of direct limit, for every s ∈ T(U) there is an open neighbourhood V of
U and a section s
∈ T(V ) which restricts to s. Since T is ﬂabby, s
can be extended
to the whole of X. So on a paracompact space, ﬂabby sheaves are acyclic, and by the
abstract de Rham theorem ﬂabby resolution can be used to compute cohomology. We
should also notice that the canonical soft resolution o
•
(T) we constructed in Section
2.5 is ﬂabby, as one can easily check by the deﬁnition itself.
One can further pursue this line and use ﬂabby resolutions (for instance, the canon
ical ﬂabby resolution of Section 2.5) to deﬁne cohomology. That is, for every sheaf T,
its cohomology is by deﬁnition the cohomology of the global sections of its canonical
ﬂabby resolution (it then turns out that cohomology can be computed with any acyclic
resolution). This has the advantage of producing a cohomology theory (called sheaf
cohomology) which is bellbehaved (e.g., it has long exact sequences in cohomology) on
every topological space, not just on paracompact ones. In this connection the reader
may refer to [5, 4, 2], or to [20] where a diﬀerent and more general approach to sheaf
cohomology (using injective resolutions) is pursued; also the original paper by Grothen
dieck [8] can be fruitfully read. It follows from our treatment that on a paracompact
topological space the sheaf and
ˇ
Cech cohomology coincide, but in general they do not
(cf. [11], especially the exercise section, for a discussion of the comparison between the
two cohomologies).
52 3. SHEAVES AND THEIR COHOMOLOGY
2.9. Comparison with other cohomologies. In algebraic topology one attaches
to a topological space X several cohomologies with coeﬃcients in an abelian group
G. Loosely speaking, whenever X is paracompact and locally Euclidean, all these
cohomologies coincide with the
ˇ
Cech cohomology of X with coeﬃcients in the constant
sheaf G. In particular, we have the following result:
Proposition 3.28. Let X be a paracompact locally Euclidean topological space, and
let G be an abelian group. The singular cohomology of X with coeﬃcients in G is
isomorphic to the
ˇ
Cech cohomology of X with coeﬃcients in the constant sheaf G.
CHAPTER 4
Spectral sequences
Spectral sequences are a powerful tool for computing homology, cohomology and
homotopy groups. Often they allow one to trade a diﬃcult computation for an easier
one. Examples that we shall consider are another proof of the
ˇ
Cechde Rham theorem,
the Leray spectral sequence, and the K¨ unneth theorem.
Spectral sequences are a diﬃcult topic, basically because the theory is quite intrin
cate and the notation is correspondingly cumbersome. Therefore we have chosen what
seems to us to be the simplest approach, due to Massey [18]. Our treatment basically
follows [3].
1. Filtered complexes
Let (K, d) be a graded diﬀerential module, i.e.,
K =
a∈Z
K
a
, d: K
a
→ K
a+1
, d
2
= 0 .
A graded submodule of (K, d) is a graded subgroup K
⊂ K such that dK
⊂ K
, i.e.,
K
=
a∈Z
K
a
, K
a
⊂ K
a
, d: K
a
→ K
a+1
.
A sequence of nested graded submodules
K = K
0
⊃ K
1
⊃ K
2
⊃ . . .
is a ﬁltration of (K, d). We then say that (K, d) is ﬁltered, and associate with it the
graded complex
1
Gr(K) =
j∈Z
K
j
/K
j+1
, K
j
= K if p ≤ 0 .
Note that by assumption (since every K
j+1
is a graded subgroup of K
j
) the ﬁltration
is compatible with the grading, i.e., if we deﬁne K
j
j
= K
j
∩ K
j
, then
(4.1) K
a
= K
a
0
⊃ K
a
1
⊃ K
a
2
⊃ . . .
is a ﬁltration of K
j
, and moreover dK
a
j
⊂ K
a+1
j
.
1
The choice of having K
p
= K for p ≤ 0 is due to notational convenience.
53
54 4. SPECTRAL SEQUENCES
Example 4.1. A double complex is a collection of abelian groups K
j.o
, with p, q ≥
0,
2
and morphisms δ
1
: K
j.o
→ K
j+1.o
, δ
2
: K
j.o
→ K
j.o+1
such that
δ
1
2
= δ
2
2
= 0 , δ
1
δ
2
+δ
2
δ
1
= 0 .
Let (T, d) be the associated total complex:
T
j
=
j+o=j
K
j.o
, d: T
j
→ T
j+1
deﬁned by d = δ
1
+δ
2
(note that the deﬁnition of d implies d
2
= 0). Then letting
T
j
=
j≥j. o≥0
K
j.o
we obtain a ﬁltration of (T, d). This satiﬁes T
j
· T for p ≤ 0. The successive quotients
of the ﬁltration are
T
j
/T
j+1
=
o∈N
K
j.o
.
Definition 4.2. A ﬁltration K
•
of (K, d) is said to be regular if for every i ≥ 0
the ﬁltration (4.1) is ﬁnite; in other words, for every i there is a number (i) such that
K
j
j
= 0 for p > (i).
For instance, the ﬁltration in Example 4.1 is regular since T
j
j
= 0 for p > i, and
indeed
T
j
j
= T
j
∩ T
j
=
j−j
;=0
K
j−;.;
.
2. The spectral sequence of a ﬁltered complex
At ﬁrst we shall not consider the grading. Let K
•
be a ﬁltration of a diﬀerential
module (K, d), and let
G =
j∈Z
K
j
.
The inclusions K
j+1
→ K
j
induce a morphism i : G → G (“the shift by the ﬁltering
degree”), and one has an exact sequence
(4.2) 0 → G
j
−−→G
;
−−→E → 0
2
This assumption is made here for simplicity but one could let p, q range over the integers; however
some of the results we are going to give would be no longer valid.
2. THE SPECTRAL SEQUENCE OF A FILTERED COMPLEX 55
with E · Gr(K). The diﬀerential d induces diﬀerentials in G and E, so that from (4.2)
one gets an exact triangle in cohomology (cf. Section 1.1)
(4.3) H(G)
j
H(G)
;
.v
v
v
v
v
v
v
v
v
H(E)
I
H
H
H
H
H
H
H
H
H
where k is the connecting morphism.
Let us now assume that the ﬁltration K
•
has ﬁnite length, i.e., K
j
= 0 for p greater
than some (called the length of the ﬁltration).
Since dK
j
⊂ K
j
for every p, we may consider the cohomology groups H(K
j
). The
morphism i induces morphisms i : H(K
j+1
) → H(K
j
). Deﬁne G
1
to be the direct sum
of the terms on the sequence (which is not exact)
0 → H(K
/
)
j
−−→H(K
/−1
)
j
−−→ . . .
j
−−→H(K
1
)
j
−−→H(K)
∼
−−→H(K
−1
)
∼
−−→ . . . ,
i.e., G
1
=
¸
j∈Z
H(K
j
) · H(G). Next we deﬁne G
2
as the sum of the terms of the
sequence
0 → i(H(K
/
))) → i(H(K
/−1
)) → . . .
→ i(H(K
1
)) → H(K)
∼
−−→H(K
−1
)
∼
−−→ . . .
Note that the morphism i(H(K
1
)) → H(K) is injective, since it is the inclusion of the
image of i : H(K
1
) → H(K) into H(K). This procedure is then iterated: G
3
is the sum
of the terms in the sequence
0 → i(i(H(K
/
)))) → i(i(H(K
/−1
))) → i(i(H(K
2
))
→ i(H(K
1
)) → H(K)
∼
−−→H(K
−1
)
∼
−−→ . . .
and now the morphisms i(i(H(K
2
)) → i(H(K
1
)) and i(H(K
1
)) → H(K) are injective.
When we reach the step , all the morphisms in the sequence
0 → i
/
(H(K
/
))) → i
/−1
(H(K
/−1
)) → . . .
→ i(H(K
1
)) → H(K)
∼
−−→H(K
−1
)
∼
−−→ . . .
are injective, so that G
/+2
· G
/+1
, and the procedure stabilizes: G
·
· G
·+1
for r ≥ +1.
We deﬁne G
∞
= G
/+1
; we have
G
∞
·
j∈Z
F
j
where F
j
= i
j
(H(K
j
)), i.e., F
j
is the image of H(K
j
) into H(K). The groups F
j
provide a ﬁltration of H(K),
(4.4) H(K) = F
0
⊃ F
1
⊃ ⊃ F
/
⊃ F
/+1
= 0 .
56 4. SPECTRAL SEQUENCES
We come now to the construction of the spectral sequence. Recall that since dK
j
⊂
K
j
, and E =
¸
j
K
j
/K
j+1
, the diﬀerential d acts on E, and one has a cohomology
group H(E) wich splits into a direct sum
H(E) ·
j∈Z
H(K
j
/K
j+1
, d) .
The cohomology group H(E) ﬁts into the exact triangle (4.3), that we rewrite as
(4.5) G
1
j
1
G
1
;
1
.·







E
1
I
1
B
B
B
B
B
B
B
B
where E
1
= H(E). We deﬁne d
1
: E
1
→ E
1
by letting d
1
= j
1
◦ k
1
; then d
2
1
= 0 since
the triangle is exact. Let E
2
= H(E
1
, d
1
) and recall that G
2
is the image of G
1
under
i : G
1
→ G
1
. We have morphisms
i
2
: G
2
→ G
2
, , j
2
: G
2
→ E
2
, k
2
: E
2
→ G
2
where
(i) i
2
is induced by i
1
by letting i
2
(i
1
(x)) = i
1
(i
1
(x)) for x ∈ G
1
;
(ii) j
2
is induced by j
1
by letting j
2
(i
1
(x)) = [j
1
(x)] for x ∈ G
1
, where [ ] denotes
taking the homology class in E
2
= H(E
1
, d
1
).
(iii) k
2
is induced by k
1
by letting k
2
([e]) = i
1
(k
1
(e)).
Exercise 4.1. Show that the morphisms j
2
and k
2
are well deﬁned, and that the
triangle
(4.6) G
2
j
2
G
2
;
2
.·







E
2
I
2
B
B
B
B
B
B
B
B
is exact.
We call (4.6) the derived triangle of (4.5). The procedure leading from (4.5) to the
triangle (4.6) can be iterated, and we get a sequence of exact triangles
G
·
j
r
G
·
;
r
.·







E
·
I
r
B
B
B
B
B
B
B
B
where each group E
·
is the cohomology group of the diﬀerential module (E
·−1
, d
·−1
),
with d
·−1
= j
·−1
◦ k
·−1
.
As we have already noticed, due to the assumption that the ﬁltration K
•
has ﬁnite
length , the groups G
·
stabilize when r ≥ + 1, and the morphisms i
·
: G
·
→ G
·
2. THE SPECTRAL SEQUENCE OF A FILTERED COMPLEX 57
become injective. Thus all morphisms k
·
: E
·
→ G
·
vanish in that range, which implies
d
·
= 0, so that the groups E
·
stabilize as well: E
·+1
· E
·
for r ≥ + 1. We denote by
E
∞
= E
/+1
the stable value.
Thus, the sequence
0 → G
∞
j
∞
−−→G
∞
→ E
∞
→ 0
is exact, which implies that E
∞
is the associated graded module of the ﬁltration (4.4)
of H(K):
E
∞
·
j≤/
F
j
/F
j+1
.
Definition 4.2. A sequence of diﬀerential modules ¦(E
·
, d
·
)¦ such that H(E
·
, d
·
)
· E
·+1
is said to be a spectral sequence. If the groups E
·
eventually become stationary,
we denote the stationary value by E
∞
. If E
∞
is isomorphic to the associated graded
module of some ﬁltered group H, we say that the spectral sequence converges to H.
So what we have seen so far in this section is that if (K, d) is a diﬀerential module
with a ﬁltration of ﬁnite length, one can build a spectral sequence which converges to
H(K).
Remark 4.3. It may happen in special cases that the groups E
·
stabilize before
getting the value r = + 1. That happens if and only if d
·
= 0 for some value r = r
0
.
This implies that d
·
= 0 also for r > r
0
, and E
·+1
· E
·
for all r ≥ r
0
. When this
happens we say that the spectral sequence degenerates at step r
0
.
Now we consider the presence of a grading.
Theorem 4.4. Let (K, d) be a graded diﬀerential module, and K
•
a regular ﬁltration.
There is a spectral sequence ¦(E
·
, d
·
)¦, where each E
·
is graded, which converges to the
graded group H
•
(K, d).
Note that the ﬁltration need not be of ﬁnite length: the length (i) of the ﬁltration
of K
j
is ﬁnite for every i, but may increase with i.
Proof. For every n and p we have d(K
a
j
) ⊂ K
a+1
j
, therefore we have cohomology
groups H
a
(K
j
). As a consequence, the groups G
·
are graded:
G
·
·
a∈Z
F
a
·
=
a.j∈Z
i
·−1
(H
a
(K
j
))
and the groups E
·
are accordingly graded. We may construct the derived triangles as
before, but now we should pay attention to the grading: the morphisms i and j have
degree zero, but k has degree one (just check the deﬁnition: k is basically a connecting
morphism).
Fix a natural number n, and let r ≥ (n + 1) + 1; for every p the morphisms
i
·
: F
a+1
·
→ F
a+1
·
58 4. SPECTRAL SEQUENCES
are injective, and the morphisms
k
·
: E
a
·
→ F
a+1
·
are zero. These are the same statements as in the ungraded case. Therefore, as it
happened in the ungraded case, the groups E
a
·
become stationary for r big enough.
Note that G
a
∞
= ⊕
j∈Z
F
a
j
, where F
a+1
j+1
= i
/(a+1)
(H
a+1
(K
j+1
)), and that the morphism
i
∞
sends F
a
j+1
injectively into F
a
j
for every n, and there is an exact sequence
0 → G
a
∞
j
∞
−−→G
a
∞
→ E
a
∞
→ 0 .
This implies that E
·
is the graded module associated with the graded complex H
•
(K, d).
The last statement in the proof means that for each n, F
a
•
is a ﬁltration of H
a
(K, d),
and E
a
∞
·
¸
j∈Z
F
a
j
/F
a
j+1
.
3. The bidegree and the ﬁveterm sequence
The terms E
·
of the spectral sequence are actually bigraded; for instance, since the
ﬁltration and the degree of K are compatible, we have
K
j
/K
j+1
·
o∈Z
K
o
j
/K
o
j+1
·
o∈Z
K
j+o
j
/K
j+o
j+1
and E
0
= E is bigraded by
E
0
=
j.o∈Z
E
j.o
0
with E
j.o
0
= K
j+o
j
/K
j+o
j+1
.
Note that the total complex associated with this bidegree yields the gradation of E.
Let us go to next step. Since d: K
j+o
j
→ K
j+o+1
j
, i.e., d: E
j.o
0
→ E
j.o+1
0
, and
E
1
= H(E, d), if we set
E
j.o
1
= H
o
(E
j.•
0
, d) · H
j+o
(K
j
/K
j+1
)
we have E
1
·
¸
j.o∈Z
E
j.o
1
.
If we go one step further we can show that
d
1
: E
j.o
1
→ E
j+1.o
1
.
Indeed if x ∈ E
j.o
1
· H
j+o
(K
j
/K
j+1
) we write x as x = [e] where e ∈ K
j+o
j
/K
j+o
j+1
so
that k
1
(x) = i
1
(k(e)) ∈ H
j+o+1
(K
j+1
) and
d
1
(x) = j
1
(k
1
(x)) = j
1
(k(e)) ∈ H
j+o+1
(K
j+1
/K
j+2
) · E
j+1.o
1
.
As a result we have E
2
·
¸
j.o∈Z
E
j.o
2
with
E
j.o
2
· H
j
(E
•.o
1
, d
1
) .
The same analysis shows that in general E
·
·
¸
j.o∈Z
E
j.o
·
with
d
·
: E
j.o
·
→ E
j+·.o−·+1
·
4. THE SPECTRAL SEQUENCES ASSOCIATED WITH A DOUBLE COMPLEX 59
and moreover we have
E
j.o
∞
· F
j+o
j
/F
j+o
j+1
.
The next two Lemmas establish the existence of the morphisms that we shall use to
introduce the socalled ﬁveterm sequence, and will anyway be useful in the following.
Lemma 4.1. There are canonical morphisms H
o
(K) → E
0.o
·
.
Proof. Since K
j
· K for p ≤ 0 we have F
a
j
· H
a
(K
j
) = H
a
(K) for p ≤ 0,
hence E
j.o
∞
= 0 for p < 0 and E
0.o
∞
· F
o
0
/F
o
1
· H
o
(K)/F
o
1
, so that there is a surjective
morphism H
o
(K) → E
0.o
∞
.
Note now that a nonzero class in E
0.o
·
cannot be a boundary, since then it should
come from E
−·.o+·−1
·
= 0. So cohomology classes are cycles. Since cohomology classes
are elements in E
0.o
·+1
, we have inclusions E
0.o
·+1
⊂ E
0.o
·
(E
0.o
·+1
is the subgroup of cycles
in E
0.o
·
). This yields an inclusion E
0.o
∞
⊂ E
0.o
·
for all r.
Combining the two arguments we obtain morphisms H
o
(K) → E
0.o
·
.
Lemma 4.2. Assume that K
a
j
= 0 if p > n (so, in particular, the ﬁltration is
regular). Then for every r ≥ 2 there is a morphism E
j.0
·
→ H
j
(K).
Proof. The hypothesis of the Lemma implies that E
j.o
·
= 0 for q < 0 (indeed,
F
j+o
j
= i
·
(H
j+o
(K
j
)) for r big enough, so that F
j+o
o
= 0 if q < 0 since then K
j+1
j
= 0).
As a consequence, for r ≥ 2 the diﬀerential d
·
: E
j.0
·
→ E
j+·.1−·
·
maps to zero, i.e., all
elements in E
j.0
·
are cycles, and determine cohomology classes in E
j.0
·+1
. This means we
have a morphism E
j.0
·
→ E
j.0
·+1
, and composing, morphisms E
j.0
·
→ E
j.0
∞
.
Since F
a
j
= 0 for p > n we have E
j.0
∞
· F
j
j
/F
j
j+1
· F
j
j
so that one has an injective
morphism E
j.0
∞
→ H
j
(K). Composing we have a morphism E
j.0
·
→ H
j
(K).
Proposition 4.3. (The ﬁveterm sequence). Assume that K
a
j
= 0 if p > n. There
is an exact sequence
0 → E
1.0
2
→ H
1
(K) → E
0.1
2
o
2
−−→E
2.0
2
→ H
2
(K) .
Proof. The morphisms involved in the sequence in addition to d
2
have been deﬁned
in the previous two Lemmas. We shall not prove the exactness of the sequence here, for
a proof cf. e.g. [5].
4. The spectral sequences associated with a double complex
In this Section we consider a double complex as we have deﬁned in Example 4.1.
Due to the presence of the bidegree, the result in Theorem 4.4 may be somehow reﬁned.
We shall use the notation in Example 4.1. The group
G =
j∈Z
T
j
=
j∈Z
a≥j. o∈N
K
j.o
60 4. SPECTRAL SEQUENCES
has natural gradation G = ⊕
a∈Z
G
a
given by
(4.7) G
a
=
j∈Z
T
a
j
·
j∈Z
a−j
;=0
K
a−;.;
but it also bigraded, with bidegree
G
j.o
= T
j+o
o
.
Notice that if we form the total complex
¸
j+o=a
G
j.o
we obtain the complex (4.7) back:
j+o=a
G
j.o
·
j+o=a
o
;=0
K
j+o−;.;
=
a−j
;=0
K
a−;.;
= G
a
.
The operators δ
1
, δ
2
and d = δ
1
+δ
2
act on G:
δ
1
: G
a.o
→ G
a+1.o
, δ
2
= G
a.o
→ G
a.o+1
, d: G
I
→ G
I+1
.
We analyze the spectral sequence associated with these data. The ﬁrst three terms
are easily described. One has
E
j.o
0
· T
j+o
j
/T
j+o
j+1
· K
j.o
so that the diﬀerential d
0
: E
j.o
0
→ E
j.o+1
0
coincides with δ
2
: K
j.o
→ K
j.o+1
, and one
has
(4.8) E
j.o
1
· H
o
(K
j.•
, δ
2
) .
At next step we have d
1
: E
j.o
1
→ E
j+1.o
1
with E
j.o
1
· H
j+o
(T
j
/T
j+1
) and T
j
/T
j+1
·
¸
o∈Z
K
j.o
. Hence the diﬀerential
d
1
: H
j+o
(
a∈Z
K
j.a
) → H
j+o+1
(
a∈Z
K
j+1.a
)
is identiﬁed with δ
1
, and
(4.9) E
j.o
2
· H
j
(E
•.o
1
, δ
1
) .
One should notice that by exchanging the two degrees in K (i.e., considering another
double complex
K such that
K
j.o
= K
o.j
), we obtain another spectral sequence, that
we denote by ¦
E
·
,
d
·
¦. Both sequences converge to the same graded group, i.e., the
cohomology of the total complex (but the corresponding ﬁltrations are in general diﬀer
ent), and this often provides interesting information. For the second spectral sequence
we get
E
o.j
1
· H
j
(K
•.o
, δ
1
) (4.10)
E
o.j
2
· H
o
(
E
j.•
1
, δ
2
) . (4.11)
4. THE SPECTRAL SEQUENCES ASSOCIATED WITH A DOUBLE COMPLEX 61
Example 4.1. A simple application of the two spectral sequences associated with
a double complex provides another proof of the
ˇ
Cechde Rham theorem, i.e., the iso
morphism H
•
(X, R) · H
•
11
(X) for a diﬀerentiable manifold X. Let U = ¦U
j
¦ be a
good cover of X, and deﬁne the double complex
K
j.o
=
ˇ
C
j
(U, Ω
o
) ,
i.e., K
•.o
is the complex of
ˇ
Cech cochains of U with coeﬃcients in the sheaf of diﬀerential
qforms. The ﬁrst diﬀerential δ
1
is basically the
ˇ
Cech diﬀerential δ, while δ
2
is the
exterior diﬀerential d.
3
Actually δ and d commute rather than anticommute, but this
is easily settled by deﬁning the action of δ
1
on K
j.o
as δ
1
= (−1)
o
δ (this of course
leaves the spaces of boundaries and cycles unchanged). We start analyzing the spectral
sequences from the terms E
1
. For the ﬁrst, we have
E
j.o
1
· H
o
(K
j.•
, d) ·
¸
j
0
<···<j
p
H
o
11
(U
j
0
...j
p
) .
Since all U
j
0
...j
p
are contractible we have
E
j.0
1
·
ˇ
C
j
(U, R)
E
j.o
1
= 0 for q = 0 .
As a consequence we have E
j.o
2
= 0 for q = 0, while
E
j.0
2
· H
j
(
ˇ
C
•
(U, R), δ) = H
j
(U, R) .
This implies that d
2
= 0, so that the spectral sequence degenerates at the second step,
and E
j.o
∞
= 0 for q = 0 and E
j.0
∞
· H
j
(U, R). The resulting ﬁltration of H
j
(T, D) has
only one nonzero quotient, so that H
j
(T, D) · H
j
(U, R).
Let us now consider the second spectral sequence. We have
E
j.o
1
· H
o
(K
•.j
, δ) = H
o
(
ˇ
C
•
(U, Ω
j
), δ) = H
o
(U, Ω
j
) .
Since the sheaves Ω
j
are acyclic, we have
E
j.0
1
· H
0
(U, Ω
j
) · Ω
j
(X)
E
j.o
1
= 0 for q = 0 .
At next step we have therefore
E
j.o
2
= 0 for q = 0, and
E
j.0
2
· H
j
(Ω
•
(X), d)) · H
j
11
(X) .
Again the spectral sequence degenerates at the second step, and we have H
j
(T, D) ·
H
j
11
(X). Comparing with what we got from the ﬁrst sequence, we obtain H
j
11
(X) ·
H
j
(U, R). Taking a direct limit on good covers, we obtain H
j
(X, R) · H
j
11
(X).
3
Here a notational conﬂict arises, so that we shall denote by D the diﬀerential of the total complex
T.
62 4. SPECTRAL SEQUENCES
Remark 4.2. From this example we may get the general result that if at step r,
with r ≥ 1, we have E
j.o
·
= 0 for q = 0 (or for p = 0) then the sequence degenerates at
step r, and E
j.0
·
· H
j
(T, d) (or E
0.o
·
· H
o
(T, d)).
5. Some applications
5.1. The spectral sequence of a resolution. In this section we extend Example
4.1 to a much general situation. Let (L
•
, f) be a complex of sheaves on a paracompact
topological space X, and let U be an open cover of X. We introduce the double complex
K
j.o
=
ˇ
C
j
(U, L
o
). We shall denote by H
o
(L
•
) the cohomology sheaves of the complex
L
•
. These are the sheaves associated with the quotient presheaves
˜
H
o
(U) =
ker f : L
o
(U) → L
o+1
(U)
Imf : L
o−1
(U) → L
o
(U)
.
The E
1
term of the ﬁrst spectral sequence is
E
j.o
1
· H
o
(K
j.•
, δ
2
) = H
o
(
ˇ
C
j
(U, L
•
), f)) ·
ˇ
C
j
(U,
˜
H
o
(L
•
)) .
The second term of the sequence is
E
j.o
2
· H
j
(E
•.o
1
, δ
1
) · H
j
(
ˇ
C
•
(U,
˜
H
o
(L
•
)), δ) · H
j
(U, H
o
(L
•
))
where, since X is paracompact, we have replaced the presheaves
˜
H
•
with the corres
ponding sheaves H
•
(possibly replacing the cover U by a suitable reﬁnement).
For the second spectral sequence we have
E
j.o
1
· H
o
(K
•.j
, δ
1
) · H
o
(
ˇ
C
•
(U, L
j
), δ
1
) · H
j
(U, L
o
)
E
o.j
2
· H
j
(
E
o.•
1
, δ
2
) · H
j
(H
o
(U, L
•
), f) .
Let assume now that L
•
is a resolution of a sheaf T; then H
o
(L
•
) = 0 for q = 0,
and H
0
(L
•
) · T. The ﬁrst spectral sequence degenerates at the second step, and we
have E
j.o
2
= 0 for q = 0 and E
j.0
2
· H
j
(U, T). The second spectral sequence does not
degenerate, but we may say that it converges to the graded group H
•
(U, T) (since the
same does the ﬁrst sequence). By taking direct limit over the cover U, we have:
Proposition 4.1. Given a resolution L
•
of a sheaf T on a paracompact space X,
there is a spectral sequence E
•
whose second term is E
j.o
2
= H
o
(H
j
(X, L
•
), f), which
converges to the graded group H
•
(X, T).
The canonical ﬁltrations of a double complex always satisfy the hypothesis of Lemma
4.2. So, considering the ﬁrst spectral sequence, we obtain morphisms (again taking a
direct limit)
H
o
(L
•
(X), f) → H
o
(X, T) .
5. SOME APPLICATIONS 63
In general these are not isomorphisms. The same morphisms could be obtained by
breaking the exact sequence 0 → T → L
•
into short exact sequences, taking the asso
ciated long exact cohomology sequences and suitably composing the morphisms, as in
the proof of the abstract de Rham theorem 3.13.
A further specialization is obtained if the resolution L
•
is acyclic; then the second
spectral sequence degenerates at the second step as well, and we get isomorphisms
H
j
(X, T) · H
j
(L
•
(X), f), i.e., we have another proof of the abstract de Rham theorem
3.13.
5.2. The spectral sequence of a ﬁbred space. Let T be a sheaf on a para
compact space X and π: X → Y a continuous map, where Y is a second paracompact
space. We shall use the fact that every sheaf of abelian groups on space admits ﬂabby
resolutions (cf. Sections 3.2.5 and 3.2.8). We shall associate a spectral sequence to these
data. We consider the complex
(4.12) 0 → π
∗
T → π
∗
L
0
;
−−→π
∗
L
1
;
−−→ . . .
where (L
•
, f) is a ﬂabby resolution of T. The morphism π
∗
T → π
∗
L
0
is injective, but
otherwise the complex (4.12) is no longer exact. However, the sheaves π
∗
L
•
are ﬂabby.
We denote by R
I
π
∗
T the cohomology sheaves H
I
(π
∗
L
•
).These sheaves are called the
higher direct images of T. Note that R
0
π
∗
T · π
∗
T.
Proposition 4.2. The sheaf R
I
π
∗
T is isomorphic to the sheaf associated with the
presheaf {
I
on Y deﬁned by {
I
(U) = H
I
(π
−1
(U), T).
This implies that the sheaves R
I
π
∗
T do not depend, up to isomorphism, on the
choice of the resolution.
Proof. R
I
π
∗
T is by deﬁnition the sheaf associated with the presheaf
U
ker f : L
I
(π
−1
(U)) → L
I+1
(π
−1
(U))
Imf : L
I−1
(π
−1
(U)) → L
I
(π
−1
(U))
= H
I
(L
•
(π
−1
(U), f) .
Since the restriction of a ﬂabby sheaf to an open subset is ﬂabby, by the abstract de
Rham theorem we have isomorphisms
H
I
(L
•
(π
−1
(U), f) · H
I
(π
−1
(U), T) ,
whence the claim follows.
Let us consider the double complex
ˇ
C
j
(U, π
∗
L
o
), where U is a locally ﬁnite open
cover of Y . The two spectral sequences we have previously studied yield at the second
term
E
j.o
2
· H
j
(U, R
o
π
∗
T)
E
j.o
2
· H
o
(H
j
(U, π
∗
L
•
), f)
64 4. SPECTRAL SEQUENCES
Since the sheaves π
∗
L
•
are soft (hence acyclic) the second spectral sequence degenerates,
and one has
E
j.o
∞
= 0 for p = 0, and
E
0.o
∞
·
E
0.o
2
· H
o
(H
0
(Y, π
∗
L
•
), f)
· H
o
(L
•
(X), f) · H
o
(X, T) .
Again after taking a direct limit, we have:
Proposition 4.3. Given a continuous map of paracompact spaces π: X → Y and a
sheaf T on X, there is a spectral sequence E
•
whose second term is E
j.o
2
= H
j
(Y, R
o
π
∗
T),
which converges to the graded group H
•
(X, T).
We describe without proof the relation between the stalks of the sheaf R
I
π
∗
T
at points y ∈ Y and the cohomology groups H
I
(π
−1
(y), T); here T is to be con
sidered as restricted to π
−1
, i.e., more precisely we should write H
I
(π
−1
(y), i
−1
j
T) where
i
j
: π
−1
(y) → X is the inclusion. Since
(R
I
π
∗
F)
j
= lim
−→
j∈l
(R
I
π
∗
T)(U) · lim
−→
j∈l
H
I
(π
−1
(U), T) ,
while H
I
(π
−1
(y), T) is the direct limit of the groups H
I
(V, T) where V ranges over all
open neighbourhoods of π
−1
(y), there is a natural map
(4.13) (R
I
π
∗
T)
j
→ H
I
(π
−1
(y), T) .
This is an isomorphism under some conditions, e.g., if Y is locally compact and π is
proper (cf. [5]). This happens for instance when both X and Y are compact.
As a simple Corollary to Proposition 4.3 one obtains Leray’s theorem:
Corollary 4.4. If every point y ∈ Y has a system of neighbourhoods whose pre
images are acyclic for T, then H
I
(X, T) · H
I
(Y, π
∗
T) for all k ≥ 0.
Proof. The hypothesis of the Corollary means that every y ∈ Y has a system of
neighbourhoods ¦U¦ such that H
I
(π
−1
(U), T) = 0 for all k > 0. This implies that
R
I
π
∗
T = 0 for k > 0, so that the only nonzero terms in the spectral sequence E
2
are
E
j.0
2
· H
j
(Y, π
∗
T). The sequence degenerates and the claim follows.
5.3. The K¨ unneth theorem. Let X, Y be topological spaces, and G an abelian
group. We shall denote by the same symbol G the corresponding constant sheaves on
the spaces X, Y and X Y . The K¨ unneth theorem computes the cohomology groups
H
•
(X Y, G) in terms of the groups H
•
(X, Z) and H
•
(Y, G).
We shall need the following version of the universal coeﬃcient theorem.
Proposition 4.5. If X is a paracompact topological space and G a torsionfree
group, then H
I
(X, G) · H
I
(X, Z) ⊗
Z
G for all k ≥ 0.
Proof. Cf. [19].
5. SOME APPLICATIONS 65
Proposition 4.6. Assume that the groups H
•
(Y, G) have no torsion over Z, and
that X and Y are compact Hausdorﬀ and locally Euclidean. Then,
H
I
(X Y, G) ·
j+o=I
H
j
(X, Z) ⊗H
o
(Y, G) .
Proof. Let π: X Y → X be the projection onto the ﬁrst factor. If U is a
contractible open set in U, then by the homotopic invariance of the cohomology with
coeﬃcients in a constant sheaf (which follows e.g. from its isomorphism with singular
cohomology) we have H
•
(U Y, G) · H
•
(Y, G). If V ⊂ U, the morphism H
•
(U
Y, G) → H
•
(V Y, G) corresponds to the identity of H
•
(Y, G). Under the present
hypotheses the morphism (4.13) is an isomorphism. These facts imply that R
j
π
∗
G is the
constant sheaf on X with stalk H
j
(Y, G). The second term of the spectral sequence of
Proposition 4.3 becomes E
j.o
2
· H
j
(X, H
o
(Y, G)). By the universal coeﬃcient theorem,
since the groups H
o
(Y, G) have no torsion over Z, we have E
j.o
2
· H
j
(X, Z)⊗
Z
H
o
(Y, G).
Part 2
Introduction to algebraic geometry
CHAPTER 5
Complex manifolds and vector bundles
In this chapter we give a sketchy introduction to complex manifolds. The reader is
assumed to be acquainted with the rudiments of the theory of diﬀerentiable manifolds.
1. Basic deﬁnitions and examples
1.1. Holomorphic functions. Let U ⊂ C be an open subset. We say that a
function f : U → C is holomorphic if it is C
1
and for all x ∈ U its diﬀerential Df
a
: C →
C is not only Rlinear but also Clinear. If elements in C are written z = x + iy, and
we set f(x, y) = α(x, y) +iβ(x, y), then this condition can be written as
(5.1) α
a
= β
j
, α
j
= −β
a
(these are the CauchyRiemann conditions). If we use z, ¯ z as variables, the Cauchy
Riemann conditions read f
¯ :
= 0, i.e. the holomorphic functions are the C
1
function of
the variable z. Moreover, one can show that holomorphic functions are analytic.
The same deﬁnition can be given for holomorphic functions of several variables.
Definition 5.1. Two open subsets U, V of C
a
are said to biholomorphic if there
exists a bijective holomorphic map f : U → V whose inverse is holomorphic. The map
f itself is then said to be biholomorphic.
1.2. Complex manifolds. Complex manifolds are deﬁned as diﬀerentiable mani
folds, but requiring that the local model is C
a
, and that the transition functions are
biholomorphic.
Definition 5.2. An ndimensional complex manifold is a second countable Haus
dorﬀ topological space X together with an open cover ¦U
j
¦ and maps ψ
j
: U
j
→ C
a
which
are homeomorphisms onto their images, and are such that all transition functions
ψ
j
◦ ψ
−1
;
: ψ
;
(U
j
∩ U
;
) → ψ
j
(U
j
∩ U
;
)
are biholomorphisms.
Example 5.3. (The Riemann sphere) Consider the sphere in R
3
centered at the
origin and having radius
1
2
, and identify the tangent planes at (0, 0,
1
2
) and (0, 0, −
1
2
) with
C. The stereographic projections give local complex coordinates z
1
, z
2
; the transition
function z
2
= 1/z
1
is deﬁned in C

= C −¦0¦ and is biholomorphic.
69
70 5. COMPLEX MANIFOLDS AND VECTOR BUNDLES
1dimensional complex manifolds are called Riemann surfaces. Compact Riemann
surfaces play a distinguished role in algebraic geometry; they are all algebraic (i.e. they
are sets of zeroes of systems of homogeneous polynomials), as we shall see in Chapter
7.
Example 5.4. (Projective spaces) We deﬁne the ndimensional complex projective
space as the space of complex lines through the origin of C
a+1
, i.e.
P
a
=
C
a+1
−¦0¦
C
∗
.
By standard topological arguments P
a
with the quotient topology is a Hausdorﬀ second
countable space.
Let π: C
a+1
− ¦0¦ → P
a
be the projection, If w = (w
0
, . . . , w
a
) ∈ C
a+1
we shall
denote π(w) = [w
0
, . . . , w
a
]. The numbers (w
0
, . . . , w
a
) are said to be the homogeneous
coordinates of the point π(w). If (u
0
, . . . , u
a
) is another set of homogeneous coordinates
for π(w), then u
j
= λw
j
, with λ ∈ C
∗
(i = 0, . . . , n).
Denote by
˜
U
j
⊂ C
a+1
the open set where w
j
= 0, let U
j
= π(
˜
U
j
), and deﬁne a map
ψ
j
: U
j
→ C
a
, ψ([w
0
, . . . , w
a
]) =
w
0
w
j
, . . . ,
w
j−1
w
j
,
w
j+1
w
j
, . . . ,
w
a
w
j
.
The sets U
j
cover P
a
, the maps ψ
j
are homeomorphisms, and their transition functions
ψ
j
◦ ψ
−1
;
: ψ
;
(U
;
) → ψ
j
(U
j
),
ψ
j
◦ ψ
−1
;
(z
1
, . . . , z
a
) =
z
1
z
j
, . . . ,
z
j−1
z
j
,
z
j+1
z
j
, . . . ,
1
z
j
, . . .
z
a
z
j
,
↑
jth argument
are biholomorphic, so that P
a
is a complex manifold (we have assumed that i < j). The
map π restricted to the unit sphere in C
a+1
is surjective, so that P
a
is compact. The
previous formula for n = 1 shows that P
1
is biholomorphic to the Riemann sphere.
The coordinates deﬁned by the maps ψ
j
, usually denoted (z
1
, . . . , z
a
), are called
aﬃne or Euclidean coordinates.
Example 5.5. (The general linear complex group). Let
M
I.a
= ¦k n matrices with complex entries, k ≤ n¦
ˆ
M
I.a
= ¦matrices in M
I.a
of rank k¦, i.e.
ˆ
M
I.a
=
/
¸
j=1
¦A ∈ M
I.a
such that det A
j
= 0¦
where A
j
, . . . , A
/
are the k k minors of A. M
I.a
is a complex manifold of dimension
kn;
ˆ
M
I.a
is an open subset in M
I.a
, as its second description shows, so it is a complex
manifold of dimension kn as well. In particular, the general linear group Gl(n, C) =
ˆ
M
a.a
is a complex manifold of dimension n
2
. Here are some of its relevant subgroups:
1. BASIC DEFINITIONS AND EXAMPLES 71
(i) U(n) = ¦A ∈ Gl(n, C) such that AA
†
= I¦;
(ii) SU(n) = ¦A ∈ U(n) such that det A = 1¦;
these two groups are real (not complex!) manifolds, and dim
R
U(n) = n
2
, dim
R
SU(n) =
n
2
−1.
(iii) the group Gl(k, n; C) formed by invertible complex matrices having a block
form
(5.2) M =
A 0
B C
where the matrices A, B, C are k k, (n −k) k, and (n −k) (n −k), respectively.
Gl(k, n; C) is a complex manifold of dimension k
2
+n
2
−nk. Since a matrix of the form
(5.2) is invertible if and only if A and C are, while B can be any matrix, Gl(k, n; C) is
biholomorphic to the product manifold Gl(k, C) Gl(n −k, C) M
I.a
.
1.3. Submanifolds. Given a complex manifold X, a submanifold of X is a pair
(Y, ι), where Y is a complex manifold, and ι : Y → X is an injective holomorphic map
whose jacobian matrix has rank equal to the dimension of Y at any point of Y (of course
Y can be thought of as a subset of X).
Example 5.6. Gl(k, n; C) is a submanifold of Gl(n, C).
Example 5.7. For any k < n the inclusion of C
I+1
into C
a+1
obtained by setting
to zero the last n −k coordinates in C
a+1
yields a map P
I
→ P
a
; the reader may check
that this realizes P
I
as a submanifold of P
a
.
Example 5.8. (Grassmann varieties) Let
G
I.a
= ¦space of kdimensional planes in C
a
¦
(so G
1.a
≡ P
a
−1). This is the Grassmann variety of kplanes in C
a
. Given a kplane,
the action of Gl(n, C) on it yields another plane (possibly coinciding with the previous
one). The subgroup of Gl(n, C) which leaves the given kplane ﬁxed is isomorphic to
Gl(k, n; C), so that
G
I.a
·
Gl(n, C)
Gl(k, n; C)
.
As the reader may check, this representation gives G
I.a
the structure of a complex
manifold of dimension k(n−k). Since in the previous reasoning Gl(n, C) can be replaced
by U(n), and since Gl(k, n; C)∩U(n) = U(k)U(n−k), we also have the representation
G
I.a
·
U(n)
U(k) U(n −k)
showing that G
I.a
is compact.
An element in G
I.a
singles out (up to a complex factor) a decomposable element in
Λ
I
C
a
,
λ = v
1
∧ ∧ v
I
72 5. COMPLEX MANIFOLDS AND VECTOR BUNDLES
where the v
j
are a basis of tangent vectors to the given kplane. So G
I.a
imbeds into
P(Λ
I
C
a
) = P
.
, where N =
a
I
− 1 (this is called the Pl¨ ucker embedding. If a basis
¦v
1
, . . . , v
a
¦ is ﬁxed in C
a
, one has a representation
λ =
a
¸
j
1
.....j
k
=1
P
j
1
...j
k
v
j
1
∧ ∧ v
j
k
;
the numbers P
j
1
...j
k
are the Pl¨ ucker coordinates on the Grassmann variety.
2. Some properties of complex manifolds
2.1. Orientation. All complex manifolds are oriented. Consider for simplicity the
1dimensional case; the jacobian matrix of a transition function z
= f(z) = α(x, y) +
iβ(x, y) is (by the CauchyRiemann conditions)
J =
α
a
α
j
β
a
β
j
=
α
a
α
j
−α
j
α
a
so that det J = α
2
a
+α
2
j
> 0, and the manifold is oriented.
Notice that we may always conjugate the complex structure, considering (e.g. in the
1dimensional case) the coordinate change z → ¯ z; in this case we have J =
1 0
0 −1
,
so that the orientation gets reversed.
2.2. Forms of type (p, q). Let X be an ndimensional complex manifold; by the
identiﬁcation C
a
· R
2a
, and since a biholomorphic map is a C
∞
diﬀeomorphism, X
has an underlying structure of 2ndimensional real manifold. Let TX be the smooth
tangent bundle (i.e. the collection of all ordinary tangent spaces to X). If (z
1
, . . . , z
a
) is
a set of local complex coordinates around a point x ∈ X, then the complexiﬁed tangent
space T
a
X ⊗
R
C admits the basis
∂
∂z
1
a
, . . . ,
∂
∂z
a
a
,
∂
∂¯ z
1
a
, . . . ,
∂
∂¯ z
a
a
.
This yields a decomposition
TX ⊗C = T
X ⊕T
X
which is intrinsic because X has a complex structure, so that the transition functions
are holomorphic and do not mix the vectors
∂
∂:
i
with the
∂
∂¯ :
i
. As a consequence one has
a decomposition
Λ
j
T
∗
X ⊗C =
j+o=j
Ω
j.o
X where Ω
j.o
X = Λ
j
(T
X)
∗
⊗Λ
o
(T
X)
∗
.
The elements in Ω
j.o
X are called diﬀerential forms of type (p, q), and can locally be
written as
η = η
j
1
...j
p
.;
1
...;
q
(z, ¯ z) dz
j
1
∧ ∧ dz
j
p
∧ d¯ z
;
1
∧ ∧ d¯ z
;
q
.
4. HOLOMORPHIC VECTOR BUNDLES 73
The compositions
Ω
j+1.o
X
Ω
j.o
X
o
Λ
j+o+1
T
∗
X
∂
o
o
o
o
o
o
o
o
o
o
o
¯
∂
O
O
O
O
O
O
O
O
O
O
O
Ω
j.o+1
X
deﬁne diﬀerential operators ∂,
¯
∂ such that
∂
2
=
¯
∂
2
= ∂
¯
∂ +
¯
∂∂ = 0
(notice that the CauchyRiemann condition can be written as
¯
∂f = 0).
3. Dolbeault cohomology
Another interesting cohomology theory one can consider is the Dolbeault cohomology
associated with a complex manifold X. Let Ω
j.o
denote the sheaf of forms of type (p, q)
on X. The Dolbeault (or CauchyRiemann) operator
¯
∂ : Ω
j.o
→ Ω
j.o+1
squares to zero.
Therefore, the pair (Ω
j.•
(X),
¯
∂) is for any p ≥ 0 a cohomology complex. Its cohomology
groups are denoted by H
j.o
¯
∂
(X), and are called the Dolbeault cohomology groups of X.
We have for this theory an analogue of the Poincar´e Lemma, which is sometimes
called the
¯
∂Poincar´e Lemma (or Dolbeault or Grothendieck Lemma).
Proposition 5.1. Let ∆ be a polycylinder in C
a
(that is, the cartesian product of
disks in C). Then H
j.o
¯
∂
(∆) = 0 for q ≥ 1.
Proof. Cf. [9].
Moreover, the kernel of the morphism
¯
∂ : Ω
j.0
→ Ω
j.1
is the sheaf of holomorphic
pforms Ω
j
. Therefore, the Dolbeault complex of sheaves Ω
j.•
is a resolution of Ω
j
,
i.e. for all p = 0, . . . , n (where n = dim
C
X) the sheaf sequence
0 → Ω
j
→ Ω
j.0
¯
∂
−−→Ω
j.1
¯
∂
−−→ . . .
¯
∂
−−→Ω
j.1
→ 0
is exact. Moreover, the sheaves Ω
j.o
are ﬁne (they are (
∞
A
modules). Then, exactly as
one proves the de Rham theorem (Theorem 3.3.14), one obtains the Dolbeault theorem:
Proposition 5.2. Let X be a complex manifold. For all p, q ≥ 0, the cohomology
groups H
j.o
¯
∂
(X) and H
o
(X, Ω
j
) are isomorphic.
4. Holomorphic vector bundles
4.1. Basic deﬁnitions. Holomorphic vector bundles on a complex manifold X are
deﬁned in the same way than smooth complex vector bundles, but requiring that all the
maps involved are holomorphic.
74 5. COMPLEX MANIFOLDS AND VECTOR BUNDLES
Definition 5.1. A complex manifold E is a rank n holomorphic vector bundle on
X if there are
(i) an open cover ¦U
α
¦ of X
(ii) a holomorphic map π: E → X
(iii) holomorphic maps ψ
α
: π
−1
(U
α
) → U
α
C
a
such that
(i) π = pr
1
◦ψ
α
, where = pr
1
is the projection onto the ﬁrst factor of U
α
C
a
;
(ii) for all p ∈ U
α
∩ U
β
, the map
pr
2
◦ψ
β
◦ ψ
−1
α
(p, •): C
a
→ C
a
is a linear isomorphism.
Vector bundles of rank 1 are called line bundles.
With the data that deﬁne a holomorphic vector bundle we may construct holo
morphic maps
g
αβ
: U
α
∩ U
β
→ Gl(n, C)
given by
g
αβ
(p) x = pr
2
◦ψ
α
◦ ψ
−1
β
(ψ, x) .
These maps satisfy the cocycle condition
g
αβ
g
βγ
g
γα
= Id on U
α
∩ U
β
∩ U
γ
.
The collection ¦U
α
, ψ
α
¦ is a trivialization of E.
For every x ∈ X, the subset E
a
= π
−1
(x) ⊂ E is called the ﬁbre of E over x. By
means of a trivialization around x, E
a
is given the structure of a vector space, which is
actually independent of the trivialization.
A morphism between two vector bundles E, F over X is a holomorphic map f : E →
F such that for every x ∈ X one has f(E
a
) ⊂ F
a
, and such that the resulting map
f
a
: E
a
→ F
a
is linear. If f is a biholomorphism, it is said to be an isomorphism of
vector bundles, and E and F are said to be isomorphic.
A holomorphic section of E over an open subset U ⊂ X is a holomorphic map
s: U → E such that π ◦ s = Id. With reference to the notation previously introduced,
the maps
s
(α)j
: U
α
→ E, s
(α)j
(x) = ψ
−1
α
(x, e
j
), i = 1, . . . , n
where ¦e
j
¦ is the canonical basis of C
a
, are sections of E over U
α
. Let E(U
α
) denote
the set of sections of E over U
α
; it is a free module over the ring O(U
α
) of holomorphic
functions on U
α
, and its subset ¦s
(α)j
¦
j=1.....a
is a basis. On an intersection U
α
∩U
β
one
has the relation
s
(α)j
=
a
¸
I=1
(g
αβ
)
jI
s
(β)I
.
4. HOLOMORPHIC VECTOR BUNDLES 75
Exercise 5.2. Show that two trivializations are equivalent (i.e. describe isomorphic
bundles) if there exist holomorphic maps λ
α
: U
α
→ Gl(n, C) such that
(5.3) g
αβ
= λ
α
g
αβ
λ
−1
β
Exercise 5.3. Show that the rule that to any open subset U ⊂ X assigns the
O
∞
A
(U)module of sections of a holomorphic vector bundle E deﬁnes a sheaf c (which
actually is a sheaf of O
A
modules).
If E is a holomorphic (or smooth complex) vector bundle, with transition functions
g
αβ
, then the maps
(5.4) g
αβ
= (g
T
αβ
)
−1
(where T denotes transposition) deﬁne another vector bundle, called the dual vector
bundle to E, and denoted by E
∗
. Sections of E
∗
can be paired with (or act on) sections
of E, yielding holomorphic (smooth complexvalued) functions on (open sets of) X.
Example 5.4. The space E = X C
a
, with the projection onto the ﬁrst factor, is
obviously a holomorphic vector bundle, called the trivial vector bundle of rank n. We
shall denote such a bundle by C
a
(in particular, C denotes the trivial line bundle). A
holomorphic vector bundle is said to be trivial when it is isomorphic to C
a
.
Every holomorphic vector bundle has an obvious structure of smooth complex vector
bundle. A holomorphic vector bundle may be trivial as a smooth bundle while not being
trivial as a holomorphic bundle. (In the next sections we shall learn some homological
techniques that can be used to handle such situations).
Example 5.5. (The tangent and cotangent bundles) If X is a complex manifold,
the “holomorphic part” T
X of the complexiﬁed tangent bundle is a holomorphic vector
bundle, whose rank equals the complex dimension of X. Given a holomorphic atlas
for X, the locally deﬁned holomorphic vector ﬁelds
∂
∂:
1
. . . ,
∂
∂:
n
provide a holomorphic
trivialization of X, such that the transition functions of T
X are the jacobian matrices
of the transition functions of X. The dual of T
X is the holomorphic cotangent bundle
of X.
Example 5.6. (The tautological bundle) Let (w
1
, . . . , w
a+1
) be homogeneous co
ordinates in P
a
. If to any p ∈ P
a
(which is a line in C
a+1
) we associate that line we
obtain a line bundle, the tautological line bundle L of P
a
. To be more concrete, let us
exhibit a trivialization for L and the related transition functions. If ¦U
j
¦ is the standard
cover of P
a
, and p ∈ U
j
, then w
j
can be used to parametrize the points in the line p. So
if p has homogeneous coordinates (w
0
, . . . , w
a
), we may deﬁne ψ
j
: π
−1
(U
j
) → U
j
C
as ψ
j
(u) = (p, w
j
) if p = π(u). The transition function is then g
jI
= w
j
/w
I
. The dual
bundle H = L
∗
acts on L, so that its ﬁbre at p = π(u), u ∈ C
a+1
can be regarded as
76 5. COMPLEX MANIFOLDS AND VECTOR BUNDLES
the space of linear functionals on the line Cu ≡ L
j
, i.e. as hyperplanes in C
a+1
. Hence
H is called the hyperplane bundle. Often L is denoted O(−1), and H is denoted O(1)
— the reason of this notation will be clear in Chapter 6.
In the same way one deﬁnes a tautological bundle on the Grassmann variety G
I.a
;
it has rank k.
Exercise 5.7. Show that that the elements of a basis of the vector space of global
sections of L can be identiﬁed homogeneous coordinates, so that dimH
0
(P
a
, L) = n +
1. Show that the global sections of H can be identiﬁed with the linear polynomials
in the homogeneous coordinates. Hence, the global sections of H
·
are homogeneous
polynomials of order r in the homogeneous coordinates.
4.2. More constructions. Additional operations that one can perform on vector
bundles are again easily described in terms of transition functions.
(1) Given two vector bundles E
1
and E
2
, of rank r
1
and r
2
, their direct sum E
1
⊕E
2
is the vector bundle of rank r
1
+ r
2
whose transition functions have the block matrix
form
g
(1)
αβ
0
0 g
(2)
αβ
(2) We may also deﬁne the tensor product E
1
⊗ E
2
, which has rank r
1
r
2
and has
transition functions g
(1)
αβ
g
(2)
αβ
. This means the following: assume that E
1
and E
2
trivialize
over the same cover ¦U
α
¦, a condition we may always meet, and that in the given
trivializations, E
1
and E
2
have local bases of sections ¦s
(α)j
¦ and ¦t
(α)I
¦. Then E
1
⊗E
2
has local bases of sections ¦s
(α)j
⊗t
(α)I
¦ and the corresponding transition functions are
given by
s
(α)j
⊗t
(α)I
=
·
1
¸
n=1
·
2
¸
a=1
(g
(1)
αβ
)
jn
(g
(2)
αβ
)
Ia
s
(β)n
⊗t
(β)a
.
In particular the tensor product of line bundles is a line bundle. If L is a line
bundle, one writes L
a
for L ⊗ ⊗ L (n factors). If L is the tautological line bundle
on a projective space, one often writes L
a
= O(−n), and similarly H
a
= O(n) (notice
that O(−n)
∗
= O(n)).
(3) If E is a vector bundle with transition functions g
αβ
, we deﬁne its determinant
det E as the line bundle whose transition functions are the functions det g
αβ
. The
determinant bundle of the holomorphic tangent bundle to a complex manifold is called
the canonical bundle K.
Exercise 5.8. Show that the canonical bundle of the projective space P
a
is iso
morphic to O(−n −1).
Example 5.9. Let π: C
a+1
− ¦0¦ → P
a
be the usual projection, and let (w
1
, . . . ,
w
a+1
) be homogeneous coordinates in P
a
. The tangent spaces to P
a
are generated by
5. CHERN CLASSES 77
the vectors π
∗
∂
∂&
i
, and these are subject to the relation
a+1
¸
j=1
w
j
π
∗
∂
∂w
j
= 0 .
If is a linear functional on C
a+1
the vector ﬁeld
v(w) = (w)
∂
∂w
j
(i is ﬁxed) satisﬁes v(λw) = λv(w) and therefore descends to P
a
. One can then deﬁne
a map
E: H
⊕(a+1)
→ TP
a
(σ
1
, . . . , σ
a+1
) →
a+1
¸
j=1
σ
j
(w)
∂
∂w
j
(recall that the sections of H can be regarded as linear functionals on the homogeneous
coordinates). The map E is apparently surjective. Its kernel is generated by the section
σ
j
(w) = w
j
, i = 1, . . . , n + 1; notice that this is the image of the map
C → H
⊕(a+1)
, 1 → (w
1
, . . . , w
a+1
) .
The morphism H
⊕(a+1)
→ TP
a
may be regarded as a sheaf morphism O
P
n
(1)
⊕(a+1)
→ TP
a
, the second sheaf being the tangent sheaf of P
a
, i.e., the sheaf of germs of
holomorphic vector ﬁelds on P
a
, and one has an exact sequence
0 → O
P
n
→ O
P
n
(1)
⊕(a+1)
→ TP
a
→ 0
called the Euler sequence.
5. Chern class of line bundles
5.1. Chern classes of holomorphic line bundles. Let X a complex manifold.
We deﬁne Pic(X) (the Picard group of X) as the set of holomorphic line bundles on X
modulo isomorphism. The group structure of Pic(X) is induced by the tensor product
of line bundles L⊗L
; in particular one has L⊗L
∗
· C (think of it in terms of transition
functions — here C denotes the trivial line bundle, whose class [C] is the identity in
Pic(X)), so that the class [L
∗
] is the inverse in Pic(X) of the class [L].
Let O denote the sheaf of holomorphic functions on X, and O
∗
the subsheaf of
nowhere vanishing holomorphic funtions. If L · L
then the transition functions g
αβ
,
g
αβ
of the two bundles with respect to a cover ¦U
α
¦ of X are 2cocycles O
∗
, and satisfy
g
αβ
= g
αβ
λ
α
λ
β
with λ
α
∈ O
∗
(U
α
),
so that one has an identiﬁcation Pic(X) · H
1
(X, O
∗
). The long cohomology sequence
associated with the exact sequence
0 → Z → O
exp
−−→O
∗
→ 0
78 5. COMPLEX MANIFOLDS AND VECTOR BUNDLES
(where exp f = e
2πj;
) contains the segment
H
1
(X, Z) → H
1
(X, O) → H
1
(X, O
∗
)
δ
−−→H
2
(X, Z) → H
2
(X, O)
where δ is the connecting morphism. Given a line bundle L, the element
c
1
(L) = δ([L]) ∈ H
2
(X, Z)
is the ﬁrst Chern class
1
of L. The fact that δ is a group morphism means that
c
1
(L ⊗L
) = c
1
(L) +c
1
(L
) .
In general, the morphism δ is neither injective nor surjective, so that
(i) the ﬁrst Chern class does not classify the holomorphic line bundles on X; the
group
Pic
0
(X) = ker δ · H
1
(X, O)/ ImH
1
(X, Z)
classiﬁes the line bundles having the same ﬁrst Chern class.
(ii) not every element in H
2
(X, Z) is the ﬁrst Chern class of a holomorphic line
bundle.
The image of c
1
is a subgroup NS(X) of H
2
(X, Z), called the N´eronSeveri group of X.
Exercise 5.1. Show that all line bundles on C
a
are trivial.
Exercise 5.2. Show that there exist nontrivial holomorphic line bundles which are
trivial as smooth complex line bundles.
Notice that when X is compact the sequence
0 → H
0
(X, Z) → H
0
(X, O) → H
0
(X, O
∗
) → 0
is exact, so that Pic
0
(X) = H
1
(X, O)/H
1
(X, Z). If in addition dimX = 1 we have
H
2
(X, O) = 0, so that every element in H
2
(X, Z) is the ﬁrst Chern class of a holomorphic
line bundle.
2
From the deﬁnition of connecting morphism we can deduce an explicit formula for
a
ˇ
Cech cocycle representing c
1
(L) with respect to the cover ¦U
α
¦:
¦c
1
(L)¦
αβγ
=
1
2πj
(log g
αβ
+ log g
βγ
+ log g
γα
) .
From this one can easily prove that, if f : X → Y is a holomorphic map, and L is a line
bundle on Y , then
c
1
(f
∗
L) = f

(c
1
(L)) .
1
This allows us also to deﬁne the ﬁrst Chern class of a vector bundle E of any rank by letting
c
1
(E) = c
1
(det E).
2
Here we use the fact that if X is a complex manifold of dimension n, then H
k
(X, O) = 0 for all
k > n.
6. CHERN CLASSES 79
5.2. Smooth line bundles. The ﬁrst Chern class can equally well be deﬁned
for smooth complex line bundles. In this case we consider the sheaf ( of complex
valued smooth functions on a diﬀerentiable manifold X, and the subsheaf (
∗
of nowhere
vanishing functions of such type. The set of isomorphism classes of smooth complex line
bundles is identiﬁed with the cohomology group H
1
(X, (
∗
). However now the sheaf (
is acyclic, so that the obstruction morphism δ establishes an isomorphism H
1
(X, (
∗
) ·
H
2
(X, Z). The ﬁrst Chern class of a line bundle L is again deﬁned as c
1
(L) = δ([L]),
but now c
1
(L) classiﬁes the bundle (i.e. L · L
if and only if c
1
(L) = c
1
(L
)).
Exercise 5.3. (A rather pedantic one, to be honest...) Show that if X is a complex
manifold, and L is a holomorphic line bundle on it, the ﬁrst Chern classes of L regarded
as a holomorphic or smooth complex line bundle coincide. (Hint: start from the inclusion
O → (, write from it a diagram of exact sequences, and take it to cohomology ...)
6. Chern classes of vector bundles
In this section we deﬁne higher Chern classes for complex vector bundles of any rank.
Since the Chern classes of a vector bundle will depend only on its smooth structure, we
may consider a smooth complex vector bundle E on a diﬀerentiable manifold X. We
are already able to deﬁne the ﬁrst Chern class c
1
(L) of a line bundle L, and we know
that c
1
(L) ∈ H
2
(X, Z). We proceed in two steps:
(1) we ﬁrst deﬁne Chern classes of vector bundles that are direct sums of line bundles;
(2) and then show that by means of an operation called cohomology base change we
can always reduce the computation of Chern classes to the previous situation.
Step 1. Let σ
j
, i = 1 . . . k, denote the symmetric function of order i in k arguments.
3
.
Since these functions are polynomials with integer coeﬃcients, they can be regarded as
functions on the cohomology ring H
•
(X, Z). In particular, if α
1
, . . . , α
I
are classes in
H
2
(X, Z), we have σ
j
(α
1
, . . . , α
I
) ∈ H
2j
(X, Z).
If E = L
1
⊕ ⊕L
I
, where the L
j
’s are line bundles, for i = 1...k we deﬁne the ith
Chern class of E as
c
j
(E) = σ
j
(c
1
(L
1
), . . . , c
1
(L
I
)) ∈ H
2j
(X, Z) .
3
The symmetric functions are deﬁned as
σ
i
(x
1
, . . . , x
k
) =
X
1≤j
1
<···<j
i
≤n
x
j
1
· · · · · x
j
i
.
Thus, for instance,
σ
1
(x
1
, . . . , x
k
) = x
1
+· · · +x
k
σ
2
(x
1
, . . . , x
k
) = x
1
x
2
+x
1
x
3
+· · · +x
k−1
x
k
. . .
σ
k
(x
1
, . . . , x
k
) = x
1
· · · · · x
k
.
As a ﬁrst reference for symmetric functions see e.g. [21].
80 5. COMPLEX MANIFOLDS AND VECTOR BUNDLES
We also set c
0
(E) = 1; identifying H
0
(X, Z) with Z (assuming that X is connected) we
may think that c
0
(E) ∈ H
0
(X, Z).
Step 2 relies on the following result (sometimes called the splitting principle), which
we do not prove here.
Proposition 5.1. Let E be a complex vector bundle on a diﬀerentiable manifold
X. There exists a diﬀerentiable map f : Y → X, where Y is a diﬀerentiable manifold,
such that
(1) the pullback bundle f
∗
E is a direct sum of line bundles;
(2) the morphism f

: H
•
(X, Z) → H
•
(Y, Z) is injective;
(3) the Chern classes c
j
(f
∗
E) lie in the image of the morphism f

.
Definition 5.2. The ith Chern class c
j
(E) of E is the unique class in H
2j
(X, Z)
such that f

(c
j
(E)) = c
j
(f
∗
E).
We also deﬁne the total Chern class of E as
c(E) =
I
¸
j=0
c
j
(E) ∈ H
•
(X, Z) .
The main property of the Chern classes are the following.
(1) If two vector bundles on X are isomorphic, their Chern classes coincide.
(2) Functoriality: if f : Y → X is a diﬀerentiable map of diﬀerentiable manifolds,
and E is a complex vector bundle on X, then
f

(c
j
(E)) = c
j
(f
∗
E) .
(3) Whitney product formula: if E, F are complex vector bundles on X, then
c(E ⊕F) = c(E) ∪ c(F) .
(4) Normalization: identify the cohomology group H
2
(P
a
, Z) with Z by identifying
the class of the hyperplane H with 1 ∈ Z. Then c
1
(H) = 1.
These properties characterize uniquely the Chern classes (cf. e.g. [13]). Notice that,
in view of the splitting principle, it is enough to prove the properties (1), (2), (3) when
E and F are line bundles. Then (1) and (2) are already known, and (3) follows from
elementary properties of the symmetric functions.
The reader can easily check that all Chern classes (but for c
0
, obviously) of a trivial
vector bundle vanish. Thus, Chern classes in some sense measure the twisting of a
bundle. It should be noted that, even in smooth case, Chern classes do not in general
classify vector bundles, even as smooth bundles (i.e., generally speaking, c(E) = c(F)
does not imply E · F). However, in some speciﬁc instances this may happen.
Exercise 5.3. Prove that for any vector bundle E one has c
1
(E) = c
1
(det E).
7. KODAIRASERRE DUALITY 81
7. KodairaSerre duality
In this section we introduce KodairaSerre duality, which will be one of the main
tools in our study of algebraic curves. To start with a simple situation, let us study
the analogous result in de Rham theory. Let X be a diﬀerentiable manifold. Since the
exterior product of two closed forms is a closed form, one can deﬁne a bilinear map
H
j
11
(X) ⊗H
;
11
(X) → H
j+;
11
(X), [τ] ⊗[ω] → [τ ∧ ω].
As we already know, via the
ˇ
Cechde Rham isomorphism this product can be identiﬁed
with the cup product. If X is compact and oriented, by composition with the map
4
∫
A
: H
a
11
(X) → R, ∫
A
[ω] =
A
ω
where n = dimX, we obtain a pairing
H
j
11
(X) ⊗H
a−j
11
(X) → R, [τ] ⊗[ω] → ∫
A
[τ ∧ ω]
which is quite easily seen to be nondegenerate. Thus one has an isomorphism
H
j
11
(X)
∗
· H
a−j
11
(X)
(this is a form of Poincar´e duality).
If X is an ndimensional compact complex manifold, in the same way we obtain a
nondegenerate pairing between Dolbeault cohomology groups
(5.5) H
j.o
¯
∂
(X) ⊗H
a−j.a−o
¯
∂
(X) → C,
and a duality
H
j.o
¯
∂
(X)
∗
· H
a−j.a−o
¯
∂
(X).
Exercise 5.1. (1) Let E be a holomorphic vector bundle on a complex manifold
X, denote by c the sheaf of its holomorphic sections, and by c
∞
the sheaf of its smooth
sections. Show (using a local trivialization and proving that the result is independent
of the trivialization) that one can deﬁne a Clinear sheaf morphism
(5.6)
¯
∂
1
: c
∞
→ Ω
0.1
⊗c
∞
which obeys a Leibniz rule
¯
∂
1
(fs) = f
¯
∂
1
s +
¯
∂f ⊗s
for s ∈ c
∞
(U), f ∈ C
∞
(U).
(2) Show that
¯
∂
1
deﬁnes an exact sequence of sheaves
(5.7) 0 → Ω
j
⊗c → Ω
j.0
⊗c
∞
¯
∂
E
−−→Ω
j.1
⊗c
∞
¯
∂
E
−−→ . . .
¯
∂
E
−−→Ω
j.a
⊗c
∞
→ 0.
4
This map is well deﬁned because diﬀerent representatives of [ω] diﬀer by an exact form, whose
integral over X vanishes.
82 5. COMPLEX MANIFOLDS AND VECTOR BUNDLES
Here Ω
j
is the sheaf of holomorphic pforms. In particular, c = ker(
¯
∂
1
: c
∞
→ Ω
0.1
⊗
c
∞
).
(3) By taking global sections in (5.7), and taking coholomology from the resulting (in
general) nonexact sequence, one deﬁnes Dolbeault cohomology groups with coeﬃcients
in E, denoted H
j.o
¯
∂
(X, E). Use the same argument as in the proof of de Rham’s theorem
to prove an isomorphism
(5.8) H
j.o
¯
∂
(X, E) · H
o
(X, Ω
j
⊗c).
By combining the pairing (5.5) with the action of the sections of E
∗
on the sections
of E we obtain a nondegenerate pairing
H
j.o
¯
∂
(X, E) ⊗H
a−j.a−o
¯
∂
(X, E
∗
) → C
and therefore a duality
H
j.o
¯
∂
(X, E)
∗
· H
a−j.a−o
¯
∂
(X, E
∗
).
Using the isomorphism (5.8) we can express this duality in the form
H
j
(X, Ω
o
⊗c)
∗
· H
a−j
(X, Ω
a−o
⊗c
∗
).
This is the KodairaSerre duality. In particular for q = 0 we get (denoting K = Ω
a
=
det T
∗
X, the canonical bundle of X)
H
j
(X, c)
∗
· H
a−j
(X, K ⊗c
∗
).
This is usually called Serre duality.
8. Connections
In this section we give the basic deﬁnitions and sketch the main properties of con
nections. The concept of connection provides the correct notion of diﬀerential operator
to diﬀerentiate the sections of a vector bundle.
8.1. Basic deﬁnitions. Let E a complex, in general smooth, vector bundle on a
diﬀerentiable manifold X. We shall denote by c the sheaf of sections of E, and by Ω
1
A
the sheaf of diﬀerential 1forms on X. A connection is a sheaf morphism
∇: c → Ω
1
A
⊗c
satisfying a Leibniz rule
∇(fs) = f∇(s) +df ⊗s
for every section s of E and every function f on X (or on an open subset). The Leibniz
rule also shows that ∇ is Clinear. The connection ∇ can be made to act on all sheaves
Ω
I
A
⊗c, thus getting a morphism
∇: Ω
I
A
⊗c → Ω
I+1
A
⊗c ,
8. CONNECTIONS 83
by letting
∇(ω ⊗s) = dω ⊗s + (−1)
I
ω ⊗∇(s).
If ¦U
α
¦ is a cover of X over which E trivializes, we may choose on any U
α
a set
¦s
α
¦ of basis sections of c(U
α
) (notice that this is a set of r sections, with r = rk E).
Over these bases the connection ∇ is locally represented by matrixvalued diﬀerential
1forms ω
α
:
∇(s
α
) = ω
α
⊗s
α
.
Every ω
α
is as an r r matrix of 1forms. The ω
α
’s are called connection 1forms.
Exercise 5.1. Prove that if g
αβ
denotes the transition functions of E with respect
to the chosen local basis sections (i.e., s
α
= g
αβ
s
β
), the transformation formula for the
connection 1forms is
(5.9) ω
α
= g
αβ
ω
β
g
−1
αβ
+dg
αβ
g
−1
αβ
.
The connection is not a tensorial morphism, but rather satiﬁes a Leibniz rule; as a
consequence, the transformation properties of the connection 1forms are inhomogeneous
and contain an aﬃne term.
Exercise 5.2. Prove that if E and F are vector bundles, with connections ∇
1
and
∇
2
, then the rule
∇(s ⊗t) = ∇
1
(s) ⊗t +s ⊗∇
2
(t)
(minimal coupling) deﬁnes a connection on the bundle E ⊗F (here s and t are sections
of E and F, respectively).
Exercise 5.3. Prove that is E is a vector bundle with a connection ∇, the rule
< ∇
∗
(τ), s >= d < τ, s > − < τ, ∇(s) >
deﬁnes a connection on the dual bundle E
∗
(here τ, s are sections of E
∗
and E, respect
ively, and <, > denotes the pairing between sections of E
∗
and E).
It is an easy exercise, which we leave to the reader, to check that the square of the
connection
∇
2
: Ω
I
A
⊗c → Ω
I+2
A
⊗c
is flinear, i.e., it satisﬁes the property
∇
2
(fs) = f∇
2
(s)
for every function f on X. In other terms, ∇
2
is an endomorphism of the bundle E
with coeﬃcients in 2forms, namely, a global section of the bundle Ω
2
A
⊗End(E). It is
called the curvature of the connection ∇, and we shall denote it by Θ. On local basis
sections s
α
it is represented by the curvature 2forms Θ
α
deﬁned by
Θ(s
α
) = Θ
α
⊗s
α
.
84 5. COMPLEX MANIFOLDS AND VECTOR BUNDLES
Exercise 5.4. Prove that the curvature 2forms may be expressed in terms of the
connection 1forms by the equation (Cartan’s structure equation)
(5.10) Θ
α
= dω
α
−ω
α
∧ ω
α
.
Exercise 5.5. Prove that the transformation formula for the curvature 2forms is
Θ
α
= g
αβ
Θ
β
g
−1
αβ
.
Due to the tensorial nature of the curvature morphism, the curvature 2forms obey a
homogeneous transformation rule, without aﬃne term.
Since we are able to induce connections on tensor products of vector bundles (and
also on direct sums, in the obvious way), and on the dual of a bundle, we can induce
connections on a variety of bundles associated to given vector bundles with connec
tions, and thus diﬀerentiate their sections. The result of such a diﬀerentiation is called
the covariant diﬀerential of the section. In particular, given a vector bundle E with
connection ∇, we may diﬀerentiate its curvature as a section of Ω
2
A
⊗End(E).
Proposition 5.6. (Bianchi identity) The covariant diﬀerential of the curvature of
a connection is zero, ∇Θ = 0.
Proof. A simple computation shows that locally ∇Θ is represented by the matrix
valued 3forms
dΘ
α
+ω
α
∧ Θ
α
−Θ
α
∧ ω
α
.
By plugging in the structure equation (5.10) we obtain ∇Θ = 0.
8.2. Connections and holomorphic structures. If X is a complex manifold,
and E a C
∞
complex vector bundle on it with a connection ∇, we may split the latter
into its (1,0) and (0,1) parts, ∇
and ∇
, according to the splitting Ω
1
A
⊗C = Ω
1.0
A
⊕Ω
0.1
A
.
Analogously, the curvature splits into its (2,0), (1,1) and (0,2) parts,
Θ = Θ
2.0
+ Θ
1.1
+ Θ
0.2
.
Obviously we have
Θ
2.0
= (∇
)
2
, Θ
1.1
= ∇
◦ ∇
+∇
◦ ∇
, Θ
0.2
= (∇
)
2
.
In particular ∇
is a morphism Ω
j.o
A
⊗ c → Ω
j.o+1
A
⊗ c. If Θ
0.2
= 0, then ∇
is a
diﬀerential for the complex Ω
j.•
A
⊗c. The same condition implies that the kernel of the
map
(5.11) ∇
: c → Ω
0.1
A
⊗c
has enough sections to be the sheaf of sections of a holomorphic vector bundle.
8. CONNECTIONS 85
Proposition 5.7. If Θ
0.2
= 0, then the C
∞
vector bundle E admits a unique
holomorphic structure, such that the corresponding sheaf of holomorphic sections is iso
morphic to the kernel of the operator (5.11). Moreover, under this isomorphism the
operator (5.11) concides with the operator
¯
δ
1
deﬁned in Exercise 5.1.
Proof. Cf. [16], p. 9.
Conversely, if E is a holomophic vector bundle, a connection ∇ on E is said to be
compatible with the holomorphic structure of E if ∇
= ∂
1
.
8.3. Hermitian bundles. A Hermitian metric h of a complex vector bundle E is
a global section of E ⊗E
∗
which when restricted to the ﬁbres yields a Hermitian form
on them (more informally, it is a smoothly varying assignation of Hermitian structures
on the ﬁbres of E). On a local basis of sections ¦s
α
¦, of E, h is represented by matrices
h
α
of functions on U
α
which, when evaluated at any point of U
α
, are Hermitian and
positive deﬁnite. The local basis is said to be unitary if the corresponding matrix h is
the identity matrix.
A pair (E, h) formed by a holomorphic vector bundle with a hermitian metric is
called a hermitian bundle. A connection ∇ on E is said to be metric with respect to h
if for every pair s, t of sections of E one has
dh(s, t) = h(∇s, t) +h(s, ∇t) .
In terms of connection forms and matrices representing h this condition reads
(5.12) dh
α
= ˜ ω
α
h
α
+h
α
¯ ω
α
where˜denotes transposition and¯denotes complex conjugation (but no transposition,
i.e., it is not the hermitian conjugation). This equation implies right away that on a
unitary frame, the connection forms are skewhermitian matrices.
Proposition 5.8. Given a hermitian bundle (E, h), there is a unique connection ∇
on E which is metric with respect to h and is compatible with the holomorphic structure
of E.
Proof. If we use holomophic local bases of sections, the connection forms are of
type (1,0). Then equation (5.12) yields
(5.13) ˜ ω
α
= ∂h
α
h
−1
α
and this equations shows the uniqueness. As for the existence, one can easily check
that the connection forms as deﬁned by equation (5.13) satisfy the condition (5.9) and
therefore deﬁne a connection on E. This is metric w.r.t. h and compatible with the
holomorphic structure of E by construction.
86 5. COMPLEX MANIFOLDS AND VECTOR BUNDLES
Example 5.9. (Chern classes and Maxwell theory) The Chern classes of a complex
vector bundle E can be calculated in terms of a connection on E via the socalled Chern
Weil representation theorem. Let us discuss a simple situation. Let L be a complex line
bundle on a smooth 2dimensional manifold X, endowed with a connection, and let F
be the curvature of the connection. F can be regarded as a 2form on X. In this case
the ChernWeil theorem states that
(5.14) c
1
(L) =
i
2π
A
F
where we regard c
1
(L) as an integer number via the isomorphism H
2
(X, Z) · Z given
by integration over X. Notice that the Chern class of F is independent of the connection
we have chosen, as it must be. Alternatively, we notice that the complexvalued form F
is closed (Bianchi identity) and therefore singles out a class [F] in the complexiﬁed de
Rham group H
2
11
(X) ⊗
R
C · H
2
(X, C); the class
j
2π
[F] is actually real, and one has
the equality
c
1
(L) =
i
2π
[F]
in H
2
11
(X). If we consider a static spherically symmetric magnetic ﬁeld in R
3
, by
solving the Maxwell equations we ﬁnd a solution which is singular at the origin. If we
do not consider the dependence from the radius the vector potential deﬁnes a connection
on a bundle L deﬁned on an S
2
which is spanned by the angular spherical coordinates.
The fact that the Chern class of L as given by (5.14) can take only integer values is
known in physics as the quantization of the Dirac monopole.
CHAPTER 6
Divisors
Divisors are a powerful tool to study complex manifolds. We shall start with the one
dimensional case. The notion will be later generalized to higher dimensional manifolds.
1. Divisors on Riemann surfaces
Let S be a Riemann surface (a complex manifold of dimension 1). A divisor D on
S is a locally ﬁnite formal linear combinations of points of S with integer coeﬃcients,
D =
¸
a
j
p
j
, a
j
∈ Z, p
j
∈ S,
where “locally ﬁnite” means that every point p in S has a neighbourhood which contains
only a ﬁnite number of p
j
’s. If S is compact, this means that the number of points is
ﬁnite. We say that the divisor D is eﬀective if a
j
≥ 0 for all i. We shall then write
D ≥ 0.
The set of all divisors of S forms an abelian group, denoted by Div(S).
Let f a holomorphic function deﬁned in a neighbourhood of p, and let z be a local
coordinate around p. There exists a unique nonnegative integer a and a holomorphic
function h such that
f(z) = (z −z(p))
o
h(z)
and h(p) = 0. We deﬁne
ord
j
f = a.
Notice that
(6.1) ord
j
fg = ord
j
f + ord
j
g.
If f is a meromorphic function which in a neighbourhood of p can be written as f = g/h,
with g and h holomorphic, we deﬁne
ord
j
f = ord
j
g −ord
j
h.
We say that f has a zero of order a at p if ord
j
f = a > 0 (then f is holomorphic in a
neighbourhood of p), and that it has a pole of order a if ord
j
f = −a < 0.
With each meromorphic function f we may associate the divisor
(f) =
¸
j∈S
ord
j
f p;
if f = g/h with g and h relatively prime, then (f) = (g) −(h).
87
88 6. DIVISORS
1.1. Sheaftheoretic description of divisors. The group of divisors Div can
be described in sheaftheoretic terms as follows. Let ´
∗
be the sheaf of meromorphic
functions that are not identically zero. We have an exact sequence
0 → O
∗
→ ´
∗
→ ´
∗
/O
∗
→ 0
of sheaves of abelian groups (notice that the group structure is multiplicative).
Proposition 6.1. There is a group isomorphism Div(S) · H
0
(S, ´
∗
/O
∗
).
Proof. Given a cover U = ¦U
α
¦ of X, one has a commutative diagram of exact
sequences
0
H
0
(S, ´
∗
/O
∗
)
C
0
(U, ´
∗
) −−−−→ C
0
(U, ´
∗
/O
∗
) −−−−→
¸
α
H
1
(U
α
, O
∗
) = 0
δ
δ
C
1
(U, O
∗
) −−−−→ C
1
(U, ´
∗
) −−−−→ C
1
(U, ´
∗
/O
∗
)
where H
1
(U
α
, O
∗
) = 0 because U
α
· C holomorphically (here δ denotes the
ˇ
Cech
cohomology operator). This diagram shows that a global section s ∈ H
0
(S, ´
∗
/O
∗
)
can be represented by a 0cochain ¦f
α
∈ ´
∗
(U
α
)¦ ∈
ˇ
C
0
(U, ´
∗
) subject to the condition
f
α
/f
β
∈ O
∗
(U
α
∩ U
β
), so that ord
j
f
α
does not depend on α, and the quantity ord
j
s is
well deﬁned. We set D =
¸
j
ord
j
s p.
Conversely, given D =
¸
a
j
p
j
, we may choose an open cover ¦U
α
¦ such that each
U
α
contains at most one p
j
, and functions g
jα
∈ O(U
α
) such that that g
jα
has a zero of
order one at p
j
if p
j
∈ U
α
. We set
f
α
=
¸
j
g
o
i
jα
.
Then f
α
/f
β
∈ O
∗
(U
α
∩ U
β
), so that ¦f
α
¦ determines a global section of ´
∗
/O
∗
.
The two constructions are one the inverse of the other, so that they establish an
isomorphism of sets. The fact that this is also a group homomorphism follows from the
formula (6.1), which holds also for meromorphic functions.
1.2. Correspondence between divisors and line bundles. Let D ∈ Div(S),
and let ¦U
α
¦ be an open cover of S with meromorphic functions ¦f
α
¦ which deﬁne the
divisor, according to Proposition 6.1. Then the functions
g
αβ
=
f
α
f
β
∈ O
∗
(U
α
∩ U
β
)
1. DIVISORS ON RIEMANN SURFACES 89
obviously satisfy the cocycle condition, and deﬁne a line bundle, which we denote by [D].
The line bundle [D] in independent, up to isomorphism, of the set of functions deﬁning
D; if ¦f
α
¦ is another set, then ord
j
i
f
α
= ord
j
i
f
α
, so that the functions h
α
= f
α
/f
α
are
holomorphic and nowhere vanishing, and
g
αβ
=
f
α
f
β
=
f
α
f
β
h
β
h
α
= g
αβ
h
β
h
α
,
so that the transition functions g
αβ
deﬁne an isomorphic line bundle.
If D = D
(1)
+D
(2)
then f
α
= f
(1)
α
f
(2)
α
by eq. (6.1), so that [D
(1)
+D
(2)
] = [D
(1)
] ⊗
[D
(2)
], and one has a homomorphism Div(S) → Pic(S).
We oﬀer now a sheaftheoretic description of this homomorphism. Let f = ¦f
α
¦ ∈
H
0
(S, ´
∗
); let us set f
α
= g
α
/h
α
, with g
α
,h
α
∈ O(U
α
) relatively prime. We have
(f) = (g) −(h), with (g) and (h) eﬀective divisors. The line bundle [(f)] has transition
functions
g
αβ
=
g
α
g
β
h
β
h
α
=
f
α
f
β
= 1
(since f is a
ˇ
Cech cocycle) so that [(f)] = C, i.e. [(f)] is the trivial line bundle.
Conversely, let D be a divisor such that [D] = C; then the transition functions of
[D] have the form
g
αβ
=
h
α
h
β
with h
α
∈ O
∗
(U
α
).
Let ¦f
α
¦ be meromorphic functions which deﬁne D, so that one also has g
αβ
=
;
α
;
β
, and
f
α
h
α
= g
αβ
f
β
h
α
=
f
β
h
β
;
the quotients
;
α
I
α
therefore determine a global nonzero meromorphic function, namely:
Proposition 6.2. The line bundle associated with a divisor D is trivial if and only
if D is the divisor of a global meromorphic function.
In view of the identiﬁcations Div(S) · H
0
(S, ´
∗
/O
∗
) and Pic(S) · H
1
(S, O
∗
) this
statement is equivalent to the exactness of the sequence
H
0
(S, ´
∗
) → H
0
(S, ´
∗
/O
∗
) → H
1
(S, O
∗
).
Definition 6.3. Two divisors D, D
∈ Div(S) are linearly equivalent if D
= D+(f)
for some f ∈ H
0
(S, ´).
Quite evidently, D and D
are linearly equivalent if and only if [D] · [D
], so that
there is an injective group homomorphism
Div(S)/¦linear equivalence¦ → Pic(S).
90 6. DIVISORS
1.3. Holomorphic and meromorphic sections of line bundles. If L is a line
bundle on S, we denote by O(L) the sheaf of its holomorphic sections, and by ´(L) the
sheaf of its meromorphic sections, the latter being deﬁned as ´(L) = O(L) ⊗
O
´. If L
has transition functions g
αβ
with respect to a cover ¦U
α
¦ of S, then a global holomorphic
section s ∈ H
0
(S, O(L)) of L corresponds to a collection of functions ¦s
α
∈ O(U
α
)¦
such that s
α
= g
αβ
s
β
on U
α
∩ U
β
. The same holds for meromorphic sections. A ﬁrst
consequence of this is that, if s, s
∈ H
0
(S, ´(L)), we have
s
α
s
α
=
g
αβ
s
β
g
αβ
s
β
=
s
β
s
β
on U
α
∩ U
β
,
so that the quotient of s and s
is a welldeﬁned global meromorphic function on S.
Let s ∈ H
0
(S, ´(L)); we have
s
α
s
β
= g
αβ
∈ O
∗
(U
α
∩ U
β
)
so that
ord
j
s
α
= ord
j
s
β
for all p ∈ U
α
∩ U
β
;
the quantity ord
j
s is well deﬁned, and we may associate with s the divisor
(s) =
¸
j∈S
ord
j
s p.
By construction we have [(s)] · L. Obviously, s is holomorphic if and only if (s) is
eﬀective.
So we have
Proposition 6.4. A line bundle L is associated with a divisor D (i.e. L = [D] for
some D ∈ Div(S)) if and only if it has a global nontrivial meromorphic section. L is the
line bundle associated with an eﬀective divisor if and only if it has a global nontrivial
holomorphic section.
Proof. The “if” part has already been proven. For the “only if” part, let L = [D]
with D a divisor with local equations f
α
= 0. Then f
α
= g
αβ
f
β
, where the functions
g
αβ
are transition functions for L; the functions f
α
glue to yield a global meromorphic
section s of L. If D is eﬀective the functions f
α
are holomorphic so that s is holomorphic
as well.
Corollary 6.5. The line bundle [p] trivializes over the cover ¦U
1
, U
2
¦, where U
1
=
S −¦p¦ and U
2
is a neighbourhood of p, biholomorphic to a disc in C.
Proof. Since [p] is eﬀective it has a global holomorphic section which vanishes only
at p, so that [p] is trivial on U
1
. Of course it is trivial on U
2
as well.
So the same happens for the line bundles [kp], k ∈ Z.
1. DIVISORS ON RIEMANN SURFACES 91
For the remainder of this section we assume that S is compact. Let us deﬁne the
degree of a divisor D =
¸
a
j
p
j
as the integer
deg D =
¸
a
j
.
For simplicity we shall write O(D) for O([D]).
Corollary 6.6. If deg D < 0, then H
0
(S, O(D)) = 0.
If L is a line bundle we denote by
S
c
1
(L) the number obtained by integrating over S
a diﬀerential 2form which via de Rham isomorphism represents
1
the
ˇ
Cech cohomology
class c
1
(L) regarded as an element in H
2
(S, R).
Proposition 6.7. For any D ∈ Div(S) one has
S
c
1
(D) = deg D.
Before proving this result we need some preliminaries. We deﬁne a hermitian metric
on a line bundle L as an assignment of a hermitian scalar product in each L
j
which is
C
∞
in p; thus a hermitian metric is a C
∞
section h of the line bundle L
∗
⊗L
∗
such that
each h(p) is a hermitian scalar product in L
j
. In terms of a local trivialization over an
open cover ¦U
α
¦ a hermitian metric is represented by nonvanishing real functions h
α
on
U
α
. On U
α
∩U
β
one has h
α
= [g
αβ
[
2
h
β
, so that the 2form
j
2π
¯
∂∂ log h
α
does not depend
on α, and deﬁnes a global closed 2form on S, which we denote by Θ.
Lemma 6.8. The class of Θ is the image in H
2
11
(S) of c
1
(L).
Proof. We need the explicit form of the de Rham correspondence. One has exact
sequences
(6.2) 0 → R → (
∞
→ Z
1
→ 0, 0 → Z
1
→ Ω
1
→ Z
2
→ 0 .
(Here Ω
1
is the sheaf of smooth realvalued 1forms.) From the long exact cohomology
sequences of the second sequence we get
H
0
(S, Ω
1
) → H
0
(S, Z
2
) → H
1
(S, Z
1
) → 0
so that the connecting morphism H
0
(S, Z
2
) → H
1
(S, Z
1
) induces an isomorphism
H
2
11
(S) → H
1
(S, Z
1
). Since we may write Θ =
j
2π
d∂ log h
α
a cocycle representing
the image of [Θ] in H
1
(S, Z
1
) is ¦θ
α
−θ
β
¦, with
θ
α
=
j
2π
∂ log h
α
.
Notice that
θ
α
−θ
β
=
j
2π
∂ (log h
α
−log h
β
) =
j
2π
d log g
αβ
so that d(θ
α
−θ
β
) = 0.
1
The reader should check that the integral does not depend on the choice of the representative.
92 6. DIVISORS
If we consider now the ﬁrst of the sequences (6.2) we obtain from its long cohomology
exact sequence a segment
0 → H
1
(S, Z
1
) → H
2
(S, R) → 0
so that the connecting morphism is now an isomorphism. If we apply it to the 1cocycle
¦θ
α
−θ
β
¦ we get the 2cocycle of R
1
2πj
log g
αβ
+
1
2πj
log g
βγ
+
1
2πj
log g
γα
= (c
1
(L))
αβγ
.
Proof of Proposition 6.7: Since c
1
and deg are both group homomorphisms, it is
enough to consider the case D = [p]. Consider the open cover ¦U
1
, U
2
¦, where U
1
=
S −¦p¦, and U
2
is a small patch around p. Then
S
c
1
(D) =
S
Θ =
j
2π
lim
c→0
S−1(c)
d∂ log h
1
where B() is the disc [z[ < , with z a local coordinate around p, and z(p) = 0. Since
¯
∂∂ =
1
2
d(∂ −
¯
∂), and assuming that h
1l
2
−1(c)
= [z[
2
, which can always be arranged, we
have
S
c
1
(D) =
1
2πj
lim
c→0
∂1(c)
∂ log z¯ z =
1
2πj
∂1(c)
dz
z
= 1
having used Stokes’ theorem and the residue theorem (note a change of sign due to a
reversal of the orientation of ∂B()).
This result suggests to set
deg L =
S
c
1
(L)
for all line bundles on S.
Corollary 6.9. If deg L < 0, then H
0
(S, O(L)) = 0.
Proof. If there is a nonzero s ∈ H
0
(S, O(L)), then L = [D] with D = (s). Since
deg D < 0 by the previous Proposition, this contradicts Corollary 6.6.
Corollary 6.10. A global meromorphic function on a compact Riemann surface
has the same number of zeroes and poles (both counted with their multiplicities).
Proof. If f global meromorphic function, we must show that deg(f) = 0. But f
is a global meromorphic section of the trivial line bundle C, whence
deg(f) =
S
c
1
(C) = 0 .
1. DIVISORS ON RIEMANN SURFACES 93
1.4. The fundamental exact sequence of an eﬀective divisor. Let us ﬁrst
deﬁne for all p ∈ S the sheaf k
j
as the 1dimensional skyscraper sheaf concentrated at
p, namely, the sheaf
k
j
(U) = C if p ∈ U, k
j
(U) = 0 if p / ∈ U.
k
j
has stalk C at p and stalk 0 elsewhere.
Let D =
¸
a
j
p
j
be an eﬀective divisor. Then the line bundle L = [D] has at least
one section s; this allows one to deﬁne a morphism O → O(D) by letting f → f s
l
for
every f ∈ O(U). We also deﬁne the skyscraper sheaf k
1
=
¸
j
(k
j
i
)
o
i
concentrated on
D.
Proposition 6.11. The sequence
(6.3) 0 → O → O(D) → k
1
→ 0
is exact.
Proof. We shall actually prove the exactness of the sequence
(6.4) 0 → O(−D) → O → k
1
→ 0
from which the previous sequence is obtained by tensoring by O(D).
2
Notice also that
k
1
⊗
O
O(D) · k
1
because in a neighbourhood of every point p
j
the sheaf O(D) is
isomorphic to O.
The exactness of the sequence (6.4) follows from the fact the any local holomorphic
function can be represented around p
j
in the form (Taylor polynomial)
f(z) = f(z
0
) +
o
i
−1
¸
I=1
1
k!
f
(I)
(z
0
) (z −z
0
)
I
+ (z −z
0
)
o
i
g(z)
where z
0
= z(p), and g is a holomorphic function. The term (z −z
0
)
o
i
g(z) is a section
of O(−D), while the ﬁrst two terms on the right single out a section of k
1
.
The sheaf O(−D) can be regarded as the sheaf of holomorphic functions which at
p
j
have a zero of order at least a
j
. Since O(D) · O(−D)
∗
, the O(D) may be identiﬁed
with the sheaf of meromorphic functions which at p
j
have a pole of order at most a
j
.
In particular one may write
0 → O(−2p) → O → k
j
⊕T
∗
j
S → 0
where T
∗
j
S is considered as a skyscraper sheaf concentrated at p (indeed the quantity
f
(z
0
) determines an element in T
∗
j
S).
If E is a holomorphic vector bundle on S, let us denote E(D) = E ⊗[D]. Then by
tensoring the exact sequence (6.4) by O(E) we get
2
Here we use the fact that tensoring all elements of an exact sequence by the sheaf of sections of a
vector bundle preserves exactness. This is quite obvious because by the local triviality of E the stalk of
O(E) at p is O
k
p
, with k the rank of E.
94 6. DIVISORS
0 → O(E(−D)) → O(E) → E
1
→ 0
where E
1
= ⊕
j
E
⊕o
i
j
i
is a skyscraper sheaf concentrated on D.
2. Divisors on higherdimensional manifolds
We start with some preparatory material.
Definition 6.1. An analytic subvariety V of a complex manifold X is a subset of
X which is locally deﬁned as the zero set of a ﬁnite collection of holomorphic functions.
An analytic subvariety V is said to be reducible if V = V
1
∪ V
2
with V
1
and V
2
properly contained in V . V is said to be irreducible if it is not reducible.
A point p ∈ V is a smooth point of V if around p the subvariety V is a submanifold,
namely, it can be written as f
1
(z
1
, . . . , z
a
) = . . . f
I
(z
1
, . . . , z
a
) = 0 with rank J = k,
where ¦z
1
, . . . , z
a
¦ is a local coordinate system for X around p, and J is the jacobian
matrix of the functions f
1
, . . . f
I
. The set of smooth points of V is denoted by V
∗
; the
set V

= V − V
∗
is the singular locus of V . The dimension of V is by deﬁnition the
dimension of V
∗
.
If dimV = dimX −1, V will be called an analytic hypersurface.
Proposition 6.2. Any analytic subvariety V can be expressed around a point p ∈ V
as the union of a ﬁnite number of analytic subvarieties V
j
which are irreducible around
p, and are such that V
j
⊂ V
;
.
Proof. This follows from the fact that the stalk O
j
is a unique factorization domain
([9] page 12).
3
Let us sketch the proof for hypersurfaces. In a neighbourhood of p the
hypersurface V is given by f = 0. Denoting by the same letter the germ of f in p,
since O
j
(where O is the sheaf of holomorphic functions on X) is a unique factorization
domain we have
f = f
1
f
n
,
where the f
j
’s are irreducible in O
j
, and are deﬁned up to multiplication by invertible
elements in O
j
; if V
j
is the locus of zeroes of f
j
, then V = ∪
j
V
j
. Since f
j
irreducible, V
j
is irreducible as well; since it is not true that f
;
= gf
j
for some g ∈ O
j
which vanishes
at p, we also have V
j
⊂ V
;
.
We may now give the general deﬁnition of divisor:
3
Let us recall this notion: one says that a ring R is an integral domain if uv = 0 implies that either
u = 0 or v = 0. An element u ∈ R in an integral domain is said to be irreducible if u = vw implies
that v or w is a unit; R is a unique factorization domain if any element u can be written as a product
u = u
1
· · · . . . u
m
, where the u
i
are irreducible and unique up to multiplication by units.
3. LINEAR SYSTEMS 95
Definition 6.3. A divisor D on a complex manifold X is a locally ﬁnite formal
linear combination with integer coeﬃcients D =
¸
a
j
V
j
, where the V
j
’s are irreducible
analytic hypersurfaces in X.
If V ⊂ X is an analytic irreducible hypersurface, and p ∈ V , we may choose around
p a coordinate system ¦w, z
2
, . . . , z
a
¦ such that V is given around p by w = 0. Given a
function f deﬁned in a neighbourhood of p, let a be the greatest integer such that
f(w, z
2
, . . . , z
a
) = w
o
h(w, z
2
, . . . , z
a
)
with h(p) = 0. The function f has the same representation in all nearby points of V ,
so that a is constant on the connected components of V , namely, it is constant on V ,
so that we may deﬁne
ord
\
f = a.
With this proviso all the theory previously developed applies to this situation; the
only deﬁnition which no longer applies is that of degree of a line bundle, in that c
1
(L)
is still represented by a 2form, while the quantities that can be integrated on X are
the 2nforms if dim
C
X = n. Proposition 6.7 must now be reformulated as follows. Let
D =
¸
a
j
V
j
be a divisor, and let V
∗
j
be the smooth locus of V
j
. We then have:
Proposition 6.4. For any divisor D ∈ Div(X) and any (2n −2)form φ on X,
A
c
1
(D) ∧ φ =
¸
j
a
j
\
∗
i
φ.
Proof. The proof is basically the same as in Proposition 6.7 (cf. [9] page 141).
3. Linear systems
In this section we consider a compact complex manifold X of arbitrary dimension.
Let D =
¸
a
j
V
j
∈ Div(X), and deﬁne [D[ as the set of all eﬀective divisors linearly
equivalent to D. We start by showing that there is an isomorphism
λ: PH
0
(X, O(D)) → [D[ .
We ﬁx a global meromorphic section s
0
of [D], and set
(6.5) s ∈ H
0
(X, O(D)) →
s
s
0
+D ∈ [D[ ;
one should notice that ord
j
i


0
≥ −a
j
if p
j
∈ V
j
so that


0
+D is indeed eﬀective.
If s
= αs with α ∈ C
∗
then


0
=


0
so that equation (6.5) does deﬁne a map
PH
0
(X, O(D)) → [D[. This map is
(i) injective because if λ(s
1
) = λ(s
2
) then s
1
/s
2
is a global nonvanishing holomorphic
function, i.e. s
1
= αs
2
with α ∈ C
∗
.
96 6. DIVISORS
(ii) Surjective because if D
1
∈ [D[ then D
1
= D + (f) for a global meromorphic
function f with ord
j
i
(f) ≥ −a
j
if p
j
∈ V
j
. So fs
0
is a global holomorphic section of [D].
Definition 6.1. A linear system is the set of divisors corresponding to a linear
subspace of PH
0
(X, O(D)). A linear system is said to be complete if it corresponds to
the whole of PH
0
(X, O(D)).
So a linear system is of the form E = ¦D
λ
¦
λ∈P
m
for some m. The number m is
called the dimension of E. A onedimensional linear system is called a pencil, a two
dimensional one a net, and a threedimensional one a web. Since all divisors in a linear
system have the same degree, one can associate a degree to a linear system.
Remark 6.2. If the elements λ
0
, . . . , λ
n
are independent in P
n
(which means that
they are images of linearly independent elements in C
n+1
), and E = ¦D
λ
¦
λ∈P
m
is a
linear system, then
D
λ
0
∩ ∩ D
λ
m
=
¸
λ∈P
m
D
λ
.
For instance, if m = 1, and D
λ
0
and D
λ
1
have local equations f = 0 and g = 0, then D
λ
has local equation c
0
f + c
1
g = 0 if λ = c
0
λ
0
+ c
1
λ
1
. So D
λ
0
∩ D
λ
1
⊂ ∩
λ∈P
1
D
λ
, which
implies D
λ
0
∩ D
λ
1
= ∩
λ∈P
1
D
λ
.
Definition 6.3. If E = ¦D
λ
¦
λ∈P
m
is a linear system, we deﬁne its base locus as
B(E) = ∩
λ∈P
m
D
λ
.
Example 6.4. If E = ¦D
λ
¦
λ∈P
1
is a pencil, every p ∈ X − B(E) lies on a unique
D
λ
, so that there is a welldeﬁned map X −B(E) → P
1
. This map is holomorphic. We
may indeed write a local equation for D
λ
in the form
(6.6) f(z
1
, . . . , z
a
) +λg(z
1
, . . . , z
a
) = 0
where f and f are local deﬁning functions for D
0
and D
∞
(holomorphic because the
divisors in E are eﬀective). f and g do not vanish simultaneously on X − B(E),
so that they do not vanish separately either. Then the above map is given by λ =
−f(z
1
, . . . , z
a
)/g(z
1
, . . . , z
a
).
Example 6.5. Since H
1
(P
a
, O) = H
2
(P
a
, O) = 0, the line bundles on P
a
are
classiﬁed by H
2
(P
a
, Z) · Z. Moreover, since c
1
(H) = 1 under this identiﬁcation
(i.e. deg H = 1), all divisors are linearly equivalent to multiples of H; in other terms,
on P
a
the only complete linear system of degree d is [dH[.
Notice that [H[ is basepoint free, i.e. B([H[) = ∅.
A fundamental result in the theory of linear systems is the following.
Proposition 6.6. (Bertini’s theorem) The generic element of a linear system is
smooth away from the base locus.
4. THE ADJUNCTION FORMULA 97
By this we mean that the set of divisors in a linear system E which have singular
points outside the base locus form a subvariety of E of dimension strictly smaller than
that of E.
Proof. If E is linear system, and D ∈ E has singularities outside B(E), Bertini’s
theorem would be violated by all pencils containing D. It is therefore suﬃcient to
prove the theorem for pencils; in this case genericity means that the divisors having
singularities out of the base locus are ﬁnite in number.
So let E = ¦D
λ
¦
λ∈P
1
be a pencil, locally described by eq. (6.6), where the coordinates
¦z
1
, . . . , z
a
¦ can be deﬁned on an open subset ∆ ⊂ X whose image in C
a
is a polydisc.
Let p
λ
be a singular point of D
λ
which is not contained in the base locus. We have the
conditions
(6.7) f(p
λ
) +λg(p
λ
) = 0
(6.8)
∂f
∂z
j
(p
λ
) +λ
∂f
∂z
j
(p
λ
) = 0, i = 1, . . . , n
f(p
λ
), g(p
λ
) = 0.
We then have λ = −f(p
λ
)/g(p
λ
), so that
∂f
∂z
j
−
f
g
∂g
∂z
j
= 0 in p
λ
,
and
(6.9)
∂
∂z
j
f
g
= 0 in p
λ
.
Let Y be the locus in ∆P
1
cut out by the conditions (6.7) and (6.8); Y is an analytic
variety, so the same holds true for its image V in ∆. Actually V is nothing but the locus
of all singular points of the divisors D
λ
. Equation (6.9) shows that f/g is constant on
the connected components of V −B, that is, every connected component of V −B meets
only one divisor of the pencil. Since the connected components of V − B are ﬁnitely
many by Proposition 6.2, the divisors which have singularities outside B(E) are ﬁnite
in number.
4. The adjunction formula
If V is a smooth analytic hypersurface in a complex manifold X, we may relate
the canonical bundles K
\
and K
A
. We shall denote by ι
\
: V → X the inclusion; one
has an injective morphism TV → ι
∗
\
TX of bundles on V . If we choose around p ∈ V
a coordinate system (z
1
, . . . , z
a
) for X such that z
1
= 0 locally describes V , then the
vector ﬁeld
∂
∂z
1
restricted to V locally generates the quotient sheaf N
\
= ι
∗
\
TX/TV ,
so that N
\
is the sheaf of sections of a line bundle, which is called the normal bundle
to V .
98 6. DIVISORS
The dual N
∗
\
, the conormal bundle to V , is the subbundle of ι
∗
\
T
∗
X whose sections
are holomorphic 1forms which are zero on vectors tangent to V .
We ﬁrst prove the isomorphism
(6.10) N
∗
\
· ι
∗
\
[−V ].
We consider the exact sequence of vector bundles on V
0 → N
∗
\
→ ι
∗
\
T
∗
X → T
∗
V → 0
whence we get
4
(6.11) ι
∗
\
K
A
· K
\
⊗N
∗
\
.
If, relative to an open cover ¦U
α
¦ of X, the divisor V is locally given by functions
f
α
∈ O(U
α
), the line bundle [V ] has transition functions g
αβ
= f
α
/f
β
. The 1form
df
α
\ ∩l
α
is a section of N
∗
\ \ ∩l
α
, which never vanishes because V is smooth. On
U
α
∩ U
β
we have
df
α
= d(g
αβ
f
β
) = dg
αβ
f
β
+g
αβ
df
β
= g
αβ
df
β
the last equality holding on V ∩ U
α
∩ U
β
. This equation shows that the 1forms df
α
do not glue to a global section of N
∗
\
, but rather to a global section of the line bundle
N
∗
\
⊗ι
∗
\
[V ], so that this bundle is trivial, and the isomorphism (6.10) holds.
By combining the formula (6.10) with the isomorphism (6.11) we obtain the adjunc
tion formula:
(6.12) K
\
· ι
∗
\
(K
A
⊗[V ]).
Sometimes an additive notation is used, and then the adjunction formula reads
K
\
= K
A\
+ [V ]
\
.
Example 6.1. Let V be the divisor cut out from P
3
by the quartic equation
(6.13) w
4
0
+w
4
1
+w
4
2
+w
4
3
= 0
where the w
j
’s are homogeneous coordinates in P
3
. It is easily shown the V is smooth,
and it is of course compact: so it is a 2dimensional compact complex manifold, called
the Fermat surface. By a nontrivial result, known as Lefschetz hyperplane theorem
([9] p. 156) one has H
1
(V, R) = 0, so that H
1
(V, O
\
) = 0. Then the group Pic
0
(V ),
which classiﬁes the line bundles whose ﬁrst Chern classes vanishes, is trivial: if a line
bundle L on V is such that c
1
(L) = 0, then it is trivial, and every line bundle is fully
classiﬁed by its ﬁrst Chern class. (The same happens on P
3
, since H
1
(P
3
, O
P
3
) = 0).
4
We use the fact that whenever
0 → E → F → G → 0
is an exact sequence of vector bundles, then det F det E⊗det G, as one can prove by using transition
functions.
4. THE ADJUNCTION FORMULA 99
We also know that K
P
3
= O
P
3
(−4H), where H is any hyperplane in P
3
. Therefore
ι
∗
\
K
A
· O
\
(−4H
\
), where H
\
= H ∩ V is a divisor in V .
Let us compute c
1
([V ]
\
) = ι
∗
\
c
1
([V ]). We use the following fact: if D
1
, D
2
, D
3
are
irreducible divisors in P
3
, then we can move the divisors inside their linear equivalence
classes in such a way that they intersects at a ﬁnite number of points. This number is
computed by the integral
P
3
c
1
([D
1
]) ∧ c
1
([D
2
]) ∧ c
1
([D
3
])
where one considers the Chern classes c
1
([D
j
]) as de Rham cohomology classes. If we
take D
1
= V , D
2
= D
3
= H the number of intersection points is 4, because such is the
degree of the algebraic system formed by the equation (6.13) and by the equations of
two (diﬀerent) hyperplanes. Since the class h, where h = c
1
([H]), generates H
2
(P
3
, Z),
we have c
1
([V ]) = 4h, that is, V ∼ 4H. Then [V ]
\
· O
\
(4H
\
).
From the adjunction formula we get K
\
· C: the canonical bundle of V is trivial.
Since we also have H
1
11
(V ) = 0, V is an example of a K3 surface.
CHAPTER 7
Algebraic curves I
The main purpose of this chapter is to show that compact Riemann surfaces can be
imbedded into projective space (i.e. they are algebraic curves), and to study some of
their basic properties.
1. The Kodaira embedding
We start by showing that any compact Riemann surface can be embedded as a
smooth subvariety in a projective space P
.
; this is special instance of the socalled
Kodaira’s embedding theorem. Together with Chow’s Lemma this implies that every
compact Riemann surface is algebraic.
We recall that, given two complex manifolds X and Y , we say that (Y, ι) is a
submanifold of X is ι is an injective holomorphic map Y → X whose diﬀerential ι
∗j
:
T
j
Y → T
ι(j)
X is of maximal rank (given by the dimension of Y ) at all p ∈ Y . In other
terms, ι maps isomorphically Y onto a smooth subvariety of X.
Proposition 7.1. Any compact Riemann surface can be realized as a submanifold
of P
.
for some N.
Proof. Pick up a line bundle L on S such that deg L > deg K + 2 (choose an
eﬀective divisor D with enough points, and let L = [D]). By Serre duality we have
(7.1) H
1
(S, O(L −2p)) · H
0
(S, O(L −2p)
−1
⊗K)
∗
= 0
for any p ∈ S, since deg(K −L+2p) < 0 (here L−2p = L⊗[−2p]). Consider now the
exact sequence
0 → O(L −2p) → O(L)
o
p
⊕ev
p
−−→ T
∗
j
S ⊕L
j
→ 0
(the morphism d
j
is Cartan’s diﬀerential followed by evaluation at p, while ev
j
is the
evaluation of sections at p). Due to (7.1) we get
0 → H
0
(S, O(L −2p)) → H
0
(S, O(L))
o
p
⊕ev
p
−−→ T
∗
j
S ⊕L
j
→ 0
so that dim[D[ ≥ 1. Let N = dim[D[, and let ¦s
0
, . . . , s
.
¦ be a basis of [D[. If U is an
open neighbourhood of p, and φ: L
l
→ U C is a local trivialization of L, the quantity
(7.2) [φ ◦ s
0
, . . . , φ ◦ s
.
] ∈ P
.
101
102 7. ALGEBRAIC CURVES I
does not depend on the trivialization φ; we have therefore established a (holomorphic)
map ι
1
: S → P
.
.
1
We must prove that (1) ι
1
is injective, and (2) the diﬀerential (ι
1
)
∗
never vanishes. (1) It is enough to prove that, given any two points p, q ∈ S, there is a
section s ∈ H
0
(S, O(L)) such that s(p) = λs(q) for all λ ∈ C
∗
; this in turn implied by
the surjectivity of the map
H
0
(S, O(L))
·
p,q
−−→L
j
⊕L
o
, s → s(p) +s(q).
To show this we start from the exact sequence
0 → O(L −p −q) → O(L)
·
p,q
−−→L
j
⊕L
o
→ 0
and note that in coholomology we have
H
0
(S, O(L −p −q))
·
p,q
−−→L
j
⊕L
o
→ H
1
(S, O(L −p −q)) = 0
since
H
1
(S, O(L −p −q)) · H
0
(S, O(L −p −q)
−1
⊗K)
∗
= 0
because deg(L −p −q)
−1
⊗K = deg K −deg L + 2 < 0.
(2) We shall actually show that the adjoint map (ι
1
)
∗
: T
∗
ι
L
(j)
P
.
→ T
∗
j
S is surject
ive. The cotangent space T
∗
j
S can be realized as the space of equivalence classes of
holomorphic functions which have the same value at p (e.g
˙
, the zero value) and have
a ﬁrstorder contact (i.e. they have the same diﬀerential at p). Let φ be a trivializing
map for L around p; we must ﬁnd a section s ∈ H
0
(S, O(L)) such that φ ◦ s(p) = 0
(i.e. s(p) = 0) and (φ ◦ s)
∗
is surjective at p. This is equivalent to showing that the
map H
0
(S, O(L−p))
o
p
−−→T
∗
j
S is surjective, since O(L−p) is the sheaf of holomorphic
sections of L vanishing at p. We consider the exact sheaf sequence
0 → O(L −2p) → O(L −p)
o
p
−−→T
∗
j
S → 0;
by Serre duality,
H
1
(S, O(L −2p))
∗
· H
0
(S, O(−L + 2p +K)) = 0
so that H
0
(S, O(L −p))
o
p
−−→T
∗
j
S is surjective.
Given any complex manifold X, one says that a line bundle L on X is very ample
if the construction (7.2) deﬁnes an imbedding of X into PH
0
(X, O(L)). A line bundle
L is said to be ample if L
a
is very ample for some natural n. A suﬃcient condition for
a line bundle to be ample may be stated as follows (cf. [9]).
Definition 7.2. A (1,1) form ω on a complex manifold is said to be positive if it
can be locally written in the form
ω = i ω
j;
dz
j
∧ d¯ z
;
1
This map actually depends on the choice of a basis of D; however, diﬀerent choices correspond
to an action of the group PGl(N + 1, C) on P
N
and therefore produce isomorphic subvarieties of P
N
.
1. THE KODAIRA EMBEDDING 103
with ω
j;
a positive deﬁnite hermitian matrix.
Proposition 7.3. If the ﬁrst Chern class of a line bundle L on a complex manifold
can be represented by a positive 2form, then L is ample.
While we have seen that any compact Riemann surface carries plenty of very ample
line bundles, this in general is not the case: there are indeed complex manifolds which
cannot be imbedded into any projective space.
A ﬁrst consequence of the imbedding theorem expressed by Proposition 7.1 is that
any line bundle on a compact Riemann surface comes from a divisor, i.e. Div(S)/linear
equivalence · Pic(S).
Proposition 7.4. If M is a smooth 1dimensional
2
analytic submanifold of pro
jective space P
a
(i.e. M is the imbedding of a compact Riemann surface into P
a
), and
L is a line bundle on M, there is a divisor D on M such that L = [D].
Proof. We must ﬁnd a global meromorphic section of L. Let H
A
be the restriction
to M of the hyperplane bundle H of P
a
, and let V be the intersection of M with a
hyperplane in P
a
(so [V ] · H
A
, and since V is eﬀective, H
A
has global holomorphic
sections). We shall show that for a big enough integer m the line bundle L + mH
A
(= L ⊗ H
n
A
) has a global holomorphic section s; if t is a holomorphic section of H
A
,
the required meromorphic section of L is s/t
n
.
We have an exact sequence
0 → O
A
(−H
A
)

−−→O
A
→ k
\
→ 0
so that after tensoring by L +mH
A
,
(7.3) 0 → O
A
(L + (m−1)H
A
)

−−→O
A
(L +mH
A
) → k
\
→ 0.
(Here

−−→ denotes the morphism given by multiplication by s). The associated long
cohomology exact sequence contains the segment
H
0
(M, O
A
(L +mH
A
))
·
−−→C
.
→ H
1
(M, O
A
(L + (m−1)H
A
))
where N = deg V . But
H
1
(M, O
A
(L + (m−1)H
A
)) · H
0
(M, K
A
⊗O(−L −(m−1)H
A
))
∗
= 0
by Serre duality and the vanishing theorem (if m is big enough, deg K
A
⊗ O(−L −
(m−1)H
A
) < 0). Therefore the morphism r in (7.3) is surjective, and H
0
(M, O
A
(L+
mH
A
)) = 0.
We shall now proceed to identify compact Riemann surfaces with (smooth) algebraic
curves. Given a homogeneous polynomial F on C
a+1
the zero locus of F in P
a
is by
deﬁnition the projection to P
a
of the zero locus of F in C
a+1
.
2
This result is actually true whatever is the dimension of M, cf. [9].
104 7. ALGEBRAIC CURVES I
Definition 7.5. A (projective) algebraic variety is a subvariety of P
a
which is the
zero locus of a ﬁnite collection of homogeneous polynomials. We shall say that an
algebraic variety is smooth if it is so as an analytic subvariety of P
a
.
The dimension of an algebraic variety is its dimension as an analytic subvariety of
P
a
. A onedimensional algebraic variety is called an algebraic curve.
The following fundamental result, called Chow’s lemma, it is not hard to prove; we
shall anyway omit its proof for the sake of brevity (cf. [9] page 167).
Proposition 7.6. (Chow’s lemma) Any analytic subvariety of P
a
is algebraic.
Exercise 7.7. Use Chow’s lemma to show that H
0
(P
a
, H
o
) — where H is the
hyperplane line bundle — can be identiﬁed with the space of homogeneous polynomials
of degree d on C
a+1
.
Using Chow’s lemma together with the imbedding theorem (Proposition 7.1) we
obtain
Corollary 7.8. Any compact Riemann surface is a smooth algebraic curve.
We switch from the terminology “compact Riemann surface” to “algebraic curve”,
understanding that we shall only consider smooth algebraic curves.
3
We shall usually denote an algebraic curve by the letter C.
2. RiemannRoch theorem
A fundamental result in the study of algebraic curves in the RiemannRoch theorem.
Let C be an algebraic curve, and denote by K its canonical bundle.
4
We denote g =
h
0
(K), and call it the arithmetic genus of C (this number will be shortly identiﬁed with
the topological genus of C).
Proposition 7.1. (RiemannRoch theorem) For any line bundle L on C one has
h
0
(L) −h
1
(L) = deg L −g + 1.
Proof. If L = C is the trivial line bundle, the result holds obviously (notice that
H
1
(C, O)
∗
· H
0
(C, K) by Serre duality). Exploiting the fact that L = [D] for some
divisor D, it is enough to prove that if the results hold for L = [D], then it also holds
for L
= [D +p] and L
= [D −p].
In the ﬁrst case we start from the exact sequence
0 → O(D) → O(D +p) → k
j
→ 0
3
Strictly speaking an algebraic curve consists of more data than a compact Riemann surface S,
since the former requires an imbedding of S into a projective space, i.e. the choice of an ample line
bundle.
4
We introduce the following notation: if E is a sheaf of O
C
modules, then h
i
(E) = dimH
i
(C, E).
3. GENERAL RESULTS 105
which gives (since H
1
(C, k
j
) = 0)
0 → H
0
(S, O(D)) → H
0
(S, O(D +p)) → C → H
1
(S, O(D)) → H
1
(S, O(D +p)) → 0
whence
h
0
(L
) −h
1
(L
) = h
0
(L) −h
1
(L) + 1 = deg L −g + 2 = deg L
−g + 1.
Analogously for L
.
By using the RiemannRoch theorem and Serre duality we may compute the degree
of K, obtaining
deg K = 2g −2.
This is called the RiemannHurwitz formula. It allows us to identify g with the to
pological genus g
top
of C regarded as a compact oriented 2dimensional real manifold
S. To this end we may use the GaussBonnet theorem, which states that the integ
ral of the Euler class of the real tangent bundle to S is the Euler characteristic of S,
χ(S) = 2−2g
top
. On the other hand the complex structure of C makes the real tangent
bundle into a complex holomorphic line bundle, isomorphic to the holomorphic tangent
bundle TC, and under this identiﬁcation the Euler class corresponds to the ﬁrst Chern
class of TC. Therefore we get deg K = 2g
top
−2, namely,
5
g = g
top
.
3. Some general results about algebraic curves
Let us ﬁx some notations and give some deﬁnitions.
3.1. The degree of a map. Let C be an algebraic curve, and ω a smooth 2
form on C, such that
C
ω = 1; the de Rham cohomology class [ω] may be regarded
as an element in H
2
(C, Z), and actually provides a basis of that space, allowing an
identiﬁcation H
2
(C, Z) · Z. If f : C
→ C is a nonconstant holomorphic map between
two algebraic curves, then f

[ω] is a nonzero element in H
2
(C
, Z), and there is a well
deﬁned integer n such that
f

[ω] = n[ω
],
where ω
is a smooth 2form on C
such that
C
ω
= 1. If p ∈ C we have
deg f
∗
(p) =
C
c
1
(f
∗
[p]) =
C
f

c
1
([p]) = n
C
c
1
([p]) = n,
so that the map f takes the value p exactly n times, including multiplicities in the sense
of divisors; we may say that f covers C n times.
6
The integer n is called the degree if f.
5
This need not be true if the algebraic curve C is singular. However the RiemannRoch theorem is
still true (provided we know what a line bundle on a singular curve is!) with g the arithmetic genus.
6
Since two holomorphic functions of one variable which agree on a nondiscrete set are identical,
and since C
is compact, the number of points in f
−1
(p) is always ﬁnite.
106 7. ALGEBRAIC CURVES I
3.2. Branch points. Given again a nonconstant holomorphic map f : C
→ C, we
may ﬁnd a coordinate z around any q ∈ C
and a coordinate w around f(q) such that
locally f is described as
(7.4) w = z
·
.
The number r−1 is called the ramiﬁcation index of f at q (or at p = f(q)), and p = f(q)
is said to be a branch point if r(p) > 1. The branch locus of f is the divisor in C
B
=
¸
o∈C
(r(q) −1) q
or its image in C
B =
¸
o∈C
(r(q) −1) f(q).
For any p ∈ C we have
f
∗
(p) =
¸
o∈;
−1
(j)
r(q) q
deg f
∗
(p) =
¸
o∈;
−1
(j)
r(q) = n.
From these formulae we may draw the following picture. If p ∈ C
does not lie in
the branch locus, then exactly n distinct points of C
are mapped to f(p), which means
that f : C
−B
→ C −B is a covering map.
7
It p ∈ C
is a branch point of ramiﬁcation
index r −1, at p exactly r sheets of the covering join together.
There is a relation between the canonical divisors of C
and C and the branch locus.
Let η be a meromorphic 1form on C, which can locally be written as
η =
g(w)
h(w)
dw.
From (7.4) we get
f
∗
η =
g(z
·
)
h(z
·
)
dz
·
= rz
·−1
g(z
·
)
h(z
·
)
dz
so that
ord
j
f
∗
η = ord
;(j)
η +r −1.
This implies the relation between divisors
(f
∗
η) = f
∗
(η) +
¸
j∈C
(r(p) −1) p.
On the other hand the divisor (η) is just the canonical divisor of C, so that
K
C
= f
∗
K
C
+B
.
7
A (holomorphic) covering map f : X → Y , with X connected, is a map such that each p ∈ Y has a
connected neighbourhood U such that f
−1
(U) = ∪
α
U
α
is the disjoint union of open subsets of X which
are biholomorphic to U via f.
3. GENERAL RESULTS 107
From this formula we may draw an interesting result. By taking degree we get
deg K
C
= ndeg K
C
+
¸
j∈C
(r(p) −1);
by using the RiemannHurwitz formula we obtain
(7.5) g(C
) = n(g(C) −1) + 1 +
1
2
¸
j∈C
(r(p) −1) .
Exercise 7.1. Prove that if f : C
→ C is nonconstant, then f

: H
0
(C, K
C
) →
H
0
(C
, K
C
) is injective. (Hint: a nonzero element ω ∈ H
0
(C, K
C
) is a global holo
morphic 1form on C which is diﬀerent from zero at all points in an open dense subset
of C. Write an explicit formula for f
∗
ω....)
Both equation (7.5) and the previous Exercise imply
g(C
) ≥ g(C).
3.3. The genus formula for plane curves. An algebraic curve C is said to be
plane if it can be imbedded into P
2
. Its image in P
2
is the zero locus of a homogeneous
polynomial; the degree d of this polynomial is by deﬁnition the degree of C. As a
divisor, C is linearly equivalent to dH (indeed, since Pic(P
2
) · Z, any divisor D on P
2
is linearly equivalent to mH for some m; if D is eﬀective, m is the number of intersection
points between D and a generic hyperplane in P
2
, and this is given by the degree of the
polynomial cutting D).
8
We want to show that for smooth plane curves the following relation between genus
and degree holds:
(7.6) g(C) =
1
2
(d −1)(d −2).
(For singular plane curves this formula must be modiﬁed.) We may prove this equa
tion by using the adjunction formula: C is imbedded into P
2
as a smooth analytic
hypersurface, so that
K
C
= ι
∗
(K
P
2
+C),
where ι : C → P
2
. Recalling that K
P
2
= −3H we then have K
C
= (d −3)ι
∗
H.
8
We are actually using here a piece of intersection theory. The fact is that any kdimensional
analytic subvariety V of an ndimensional complex manifold X determines a homology class [V ] in the
homology group H
2k
(X, Z). Assume that X is compact, and let W be an (n −k)dimensional analytic
subvariety of X; the homology cap product H
2k
(X, Z) ∩ H
2n−2k
(X, Z) → Z, which is dual to the cup
product in cohomology, associates the integer number [V ] ∩ [W] with the two subvarieties. One may
pick up diﬀerent representatives V
and W
of [V ] and [W] such that V
and W
meet transversally,
i.e. they meet at a ﬁnite number of points; then the the number [V ] ∩[W] counts the intersection points
(cf. [9] page 49).
In our case the number of intersection points is given by the number of solutions to an algebraic
system, given by the equation of C in P
2
(which has degree d) and the linear equation of a hyperplane.
For a generic choice of the hyperplane, the number of solutions is d.
108 7. ALGEBRAIC CURVES I
To carry on the computation, we notice that, as a divisor on C, ι
∗
H = C ∩ H, so
that
deg ι
∗
H = d,
and
deg K
C
= d(d −3) = 2g −2
whence the formula (7.6).
Example 7.2. Consider the aﬃne curve in C
2
having equation
y
2
= x
6
−1 .
By writing this equation in homogeneous coordinates one obtain a curve in P
2
which
is a double covering of P
1
branched at 6 points. By the RiemannHurwitz formula we
may compute the genus, obtaining g = 2. Thus the formula (7.6), which would yield
g = 10, fails in this case. This happens because the curve is singular at the point at
inﬁnity.
3.4. The residue formula. A meromorphic 1form on an algebraic curve C is a
meromorphic section of the canonical bundle K. Given a point p ∈ C, and a local
holomorphic coordinate z such that z(p) = 0, a meromorphic 1form ϕ is locally written
around p in the form ϕ = f dz, where f is a meromorphic function. Let a be coeﬃcient
of the z
−1
term in the Laurent expansion of f around p, and let B a small disc around
p; by the Cauchy formula we have
a =
∂1
ϕ
so that the number a does not depend on the representation of ϕ. We call it the residue
of ϕ at p, and denote it by Res
j
(ϕ).
Given a meromorphic 1form ϕ its polar divisor is D =
¸
j
p
j
, where the p
j
’s are the
points where the local representatives of ϕ have poles of order 1.
Proposition 7.3. Let D =
¸
j
p
j
be the polar divisor of a meromorphic 1form ϕ.
Then
¸
j
Res
j
i
(ϕ) = 0.
Proof. Choose a small disc B
j
around each point p
j
. Then
¸
j
Res
j
i
(ϕ) =
∂∪
i
1
i
ϕ = −
C−∪
i
1
i
dϕ = 0 .
3. GENERAL RESULTS 109
3.5. The g = 0 case. We shall now show that all algebraic curves of genus zero
are isomorphic to the Riemann sphere P
1
. Pick a point p ∈ C; the line bundle [p] is
trivial on C −¦p¦, and has a holomorphic section s
0
which is nonzero on C −¦p¦ and
has a simple zero at p (this means of course that (s
0
) = p). On the other hand, since by
Serre duality h
1
(O) = h
0
(K) = 0, by taking the cohomology exact sequence associated
with the sequence
0 → O → O(p) → k
j
→ 0
we obtain the existence of a global section s of [p] which does not vanish at p. Of course
s vanishes at some other point s
0
. Then the quotient f = s/s
0
is a global meromorphic
function, with a simple pole at p and a zero at p
0
.
9
By considering ∞ as the value of f
at p, we may think of f as a holomorphic nonconstant map f : C → P
1
; this map takes
the value ∞ only once. Suppose that f takes the same value α at two distinct points
of C; then then function f − α has two zeroes and only one simple pole, which is not
possible. Thus f is injective. The following Lemma implies that f is surjective as well,
so that it is an isomorphism.
Lemma 7.4. Any holomorphic map between compact complex manifolds of the same
dimension whose Jacobian determinant is not everywhere zero is surjective.
Proof. Let f : X → Y be such a map, and let n = dimX = dimY . Let ω
be a volume form on Y ; since the Jacobian determinant of f is not everywhere zero,
and where it is not zero is positive, we have
A
f
∗
ω > 0. Assume q = Imf. Since
H
2a
(Y − ¦q¦, R) = 0 (prove it by using a MayerVietoris argument), we have ω = dη
on Y −¦q¦. But then
A
f
∗
ω =
∂A
f
∗
η = 0,
a contradiction.
9
Otherwise one can directly identify the sections of L with meromorphic functions having (only) a
single pole at p, since such functions can be developed around p in the form
f(z) =
a
z
+g(z) ,
where g is a holomorphic function. a ∈ C should be indentiﬁed with the projection of f onto k
p
. (Here
z is a local complex coordinate such that z(p) = 0.)
CHAPTER 8
Algebraic curves II
In this chapter we further study the geometry of algebraic curves. Topics covered
include the Jacobian variety of an algebraic curve, some theory of elliptic curves, and
the desingularization of nodal plane singular curves (this will involve the introduction
of the notion of blowup of a complex surface at a point).
1. The Jacobian variety
A fundamental tool for the study of an algebraic curve C is its Jacobian variety
J(C), which we proceed now to deﬁne. Let V be an mdimensional complex vector
space, and think of it as an abelian group. A lattice Λ in V is a subgroup of V of the
form
(8.1) Λ =
2n
¸
j=1
n
j
v
j
, n
j
∈ Z
¸
where ¦v
j
¦
j=1.....2n
is a basis of V as a real vector space. The quotient space T = V/Λ
has a natural structure of complex manifold, and one of abelian group, and the two
structures are compatible, i.e. T is a compact abelian complex Lie group. We shall
call T a complex torus. Notice that by varying the lattice Λ one gets another complex
torus which may not be isomorphic to the previous one (the complex structure may be
diﬀerent), even though the two tori are obviously diﬀeomorphic as real manifolds.
Example 8.1. If C is an algebraic curve of genus g, the group Pic
0
(C), classifying
the line bundles on C with vanishing ﬁrst Chern class, has a structure of complex torus
of dimension g, since it can be represented as H
1
(C, O)/H
1
(C, Z), and H
1
(C, Z) is a
lattice in H
1
(C, O). This is the Jacobian variety of C. In what follows we shall construct
this variety in a more explicit way.
Consider now a smooth algebraic curve C of genus g ≥ 1. We shall call abelian
diﬀerentials the global sections of K (i.e. the global holomorphic 1forms). If ω in
abelian diﬀerential, we have dω = 0 and ω ∧ ω = 0; this means that ω singles out a
cohomology class [ω] in H
1
(C, C), and that
(8.2)
C
ω ∧ ω = 0.
111
112 8. ALGEBRAIC CURVES II
Moreover, since locally ω = f(z) dz, we have
(8.3) i
C
ω ∧ ¯ ω > 0 if ω = 0.
If γ is a smooth loop in C, and ω ∈ H
0
(C, K), the number
γ
ω depends only on
the homology class of γ and the cohomology class of ω, and expresses the pairing <, >
between the Poincar´e dual spaces H
1
(C, C) = H
1
(C, Z) ⊗
Z
C and H
1
(C, C).
Pick up a basis ¦[γ
1
], . . . , [γ
2j
]¦ of the 2gdimensional free Zmodule H
1
(C, Z), where
the γ
j
’s are smooth loops in C, and a basis ¦ω
1
, . . . , ω
j
¦ of H
0
(C, K). We associate with
these data the g 2g matrix Ω whose entries are the numbers
Ω
j;
=
γ
j
ω
j
.
This is called the period matrix. Its columns Ω
;
are linearly independent over R: if for
all i = 1, . . . g
0 =
2j
¸
;=1
λ
;
Ω
j;
=
2j
¸
;=1
λ
;
γ
j
ω
j
then also
¸
2j
;=1
λ
;
γ
j
¯ ω
j
= 0. Since ¦ω
j
, ¯ ω
j
¦ is a basis for H
1
(C, C), this implies
¸
2j
;=1
λ
;
[γ
;
] = 0, that is, λ
;
= 0. So the columns of the period matrix generate a
lattice Λ in C
j
. The quotient complex torus J(C) = C
j
/Λ is the Jacobian variety of C.
Deﬁne now the intersection matrix Q by letting Q
−1
j;
= [γ
;
] ∩ [γ
j
] (this is the Z
valued “cap” or “intersection” product in homology). Notice that Q is antisymmetric.
Intrinsically, Q is an element in Hom
Z
(H
1
(C, Z), H
1
(C, Z)). Since the cup product in
cohomology is Poincar´e dual to the cap product in homology, for any abelian diﬀerentials
ω, τ we have
[ω] ∪ [τ] =< Q[ω], [τ] > .
The relations (8.2), (8.3) can then be written in the form
ΩQ
˜
Ω = 0, i ΩQΩ
†
> 0
(here ˜ denotes transposition, and
†
hermitian conjugation). In this form they are called
Riemann bilinear relations.
A way to check that the construction of the Jacobi variety does not depend on the
choices we have made is to restate it invariantly. Integration over cycles deﬁnes a map
i : H
1
(C, Z) → H
0
(C, K)
∗
, i([γ])(ω) =
γ
ω.
This map is injective: if i([γ])(ω) = 0 for a given γ and all ω then γ is homologous to
the constant loop. Then we have the representation J(C) = H
0
(C, K)
∗
/H
1
(C, Z).
Exercise 8.2. By regarding J(C) as H
0
(C, K)
∗
/H
1
(C, Z), show that Serre and
Poincar´e dualities establish an isomorphism J(C) · Pic
0
(C).
1. THE JACOBIAN VARIETY 113
1.1. The Abel map. After ﬁxing a point p
0
in C and a basis ¦ω
1
, . . . , ω
j
¦ in
H
0
(C, K) we deﬁne a map
(8.4) µ: C → J(C)
by letting
µ(p) =
j
j
0
ω
1
, . . . ,
j
j
0
ω
j
.
Actually the value of µ(p) in C
j
will depend on the choice of the path from p
0
to p;
however, if δ
1
and δ
2
are two paths, the oriented sum δ
1
− δ
2
will deﬁne a cycle in
homology, the two values will diﬀer by an element in the lattice, and µ(p) is a well
deﬁned point in J(C).
From (8.4) we may get a group homomorphism
(8.5) µ: Div(C) → J(C)
by letting
µ(D) =
¸
j
µ(p
j
) −
¸
;
µ(q
;
) if D =
¸
j
p
j
−
¸
;
q
;
.
All of this depends on the choice of the base point p
0
, note however that if deg D = 0
then the choice of p
0
is immaterial.
Proposition 8.3. (Abel’s theorem) Two divisors D, D
∈ Div(C) are linearly equi
valent if and only if µ(D) = µ(D
).
Proof. For a proof see [9] page 232.
Corollary 8.4. The Abel map µ: C → J(C) is injective.
Proof. If µ(p) = µ(q) by the previous Proposition p ∼ q as divisors, but since
g(C) ≥ 1 this implies p = q (this follows from considerations analogous to those in
subsection 7.3.5).
Abel’s theorem may be stated in a fancier language as follows. Let Div
o
(C) be
the subset of Div(C) formed by the divisors of degree d, and let Pic
o
(C) be the set of
line bundles of degree d.
1
One has a surjective map : = Div
o
(C) → Pic
o
(C) whose
kernel is isomorphic to H
0
(C, ´
∗
)/H
0
(C, O
∗
). Then µ ﬁlters through a morphism
ν : Pic
o
(C) → J(C), and one has a commutative diagram
Div
o
(C)
/
j
K
K
K
K
K
K
K
K
K
K
Pic
o
(C)
ν
J(C)
;
1
Notice that Pic
d
(C) Pic
d
(C) as sets for all values of d and d
.
114 8. ALGEBRAIC CURVES II
moreover, the morphism ν is injective (if ν(L) = 0, set L = (D) (i.e. L = [D]); then
µ(L) = 0, that is, L is trivial).
We can actually say more about the morphism ν, namely, that it is a bijection. It
is enough to prove that ν is surjective for a ﬁxed value of d (cf. previous footnote).
Let C
o
be the dfold cartesian product of C with itself. The symmetric group S
o
of
order d acts on C
o
; we call the quotient Sym
o
(C) = C
o
/S
o
the dfold symmetric product
of C. Sym
o
(C) can be identiﬁed with the set of eﬀective divisors of C of degree d. The
map µ deﬁnes a map µ
o
: Sym
o
(C) → J(C).
Any local coordinate z on C yields a local coordinate system ¦z
1
, . . . , z
o
¦ on C
o
,
z
j
(p
1
, . . . , p
o
) = z(p
j
),
and the elementary symmetric functions of the coordinates z
j
yield a local coordin
ate system for Sym
o
(C). Therefore the latter is a ddimensional complex manifold.
Moreover, the holomorphic map
C
o
→ J(C), (p
1
, . . . , p
o
) → µ(p
1
) + +µ(p
o
)
is S
o
invariant, hence it descends to a map Sym
o
(C) → J(C), which coincides with µ
o
.
So the latter is holomorphic.
Exercise 8.5. Prove that Sym
o
(P
1
) · P
o
. (Hint: write explicitly a morphism in
homogeneous coordinates.)
The surjectivity of ν follows from the following fact, usually called Jacobi inversion
theorem.
Proposition 8.6. The map µ
j
: Sym
j
(C) → J(C) is surjective.
Proof. Let D =
¸
p
j
∈ Sym
j
(C), with all the p
j
’s distinct, and let z
j
be a local
coordinate centred in p
j
; then ¦z
1
, . . . , z
j
¦ is a local coordinate system around D. If D
is near D we have
(8.6)
∂
∂z
j
(µ
j
(D
))
;
=
∂
∂z
j
j
i
j
0
ω
;
= h
;j
where h
;j
is the component of ω
;
on dz
j
.
Consider now the matrix
(8.7)
¸
¸
ω
1
(p
1
) . . . ω
1
(p
j
)
. . . . . . . . .
ω
j
(p
1
) . . . ω
j
(p
j
)
¸
We may choose p
1
so that ω
1
(p
1
) = 0, and then subtracting a suitable multiple of ω
1
from ω
2
, . . . , ω
j
we may arrange that ω
2
(p
1
) = = ω
j
(p
1
) = 0. We next choose p
2
so
that ω
2
(p
2
) = 0, and arrange that ω
3
(p
2
) = = ω
j
(p
3
) = 0, and so on. In this way the
matrix (8.7) is upper triangular. With these choices of the abelian diﬀerentials ω
j
and of
1. THE JACOBIAN VARIETY 115
the points p
j
the Jacobian matrix ¦h
;j
¦ is upper triangular as well, and since ω
j
(p
j
) = 0,
its diagonal elements h
jj
are nonzero at D, so that at the point D corresponding to our
choices the Jacobian determinant is nonzero. This means that the determinant is not
everywhere zero, and by Lemma 7.4 one concludes.
Proposition 8.7. The map µ
j
is generically onetoone.
Proof. Let u ∈ J(C), and choose a divisor D ∈ µ
−1
j
(u). By Abel’s theorem the
ﬁbre µ
−1
j
(u) is formed by all eﬀective divisors linearly equivalent to D, hence it is a
projective space. But since dimJ(C) = dimSym
o
(C) the ﬁbre of µ
j
is generically
0dimensional, so that generically it is a point.
This means that µ
j
establishes a biholomorphic correspondence between a dense
subset of Sym
o
(C) and a dense subset of J(C); such maps are called birational.
Corollary 8.8. Every divisor of degree ≥ g on an algebraic curve of genus g is
linearly equivalent to an eﬀective divisor.
Proof. Let D ∈ Div
o
(C) with d ≥ g. We may write D = D
+D
with deg D
= g
and D
≥ 0. By mapping D
to J(C) by Abel’s map and taking a counterimage in
Sym
j
(C) we obtain an eﬀective divisor E linearly equivalent to D
. Then E + D
is
eﬀective and linearly equivalent to D.
Corollary 8.9. Every elliptic smooth algebraic curve (i.e. every smooth algebraic
curve of genus 1) is of the form C/Λ for some lattice Λ ⊂ C.
Proof. We have J(C) = C/Λ, and the map µ
1
concides with µ,
µ(p) =
j
j
0
ω.
By Abel’s theorem, µ(p) = µ(q) if and only if there is on C a meromorphic function
f such that (f) = p − q; but on C there are no meromorphic functions with a single
pole, so that µ is injective. µ is also surjective by Lemma 7.4 (this is a particular case
of Jacobi inversion theorem), hence it is bijective.
Corollary 8.10. The canonical bundle of any elliptic curve is trivial.
Proof. We represent an elliptic curve C as a quotient C/Λ. The (trivial) tangent
bundle to C is invariant under the action of Λ, therefore the tangent bundle to C is
trivial as well.
Another consequence is that if C is an elliptic algebraic curve and one chooses a
point p ∈ C, the curve has a structure of abelian group, with p playing the role of the
identity element.
116 8. ALGEBRAIC CURVES II
1.2. Jacobian varieties are algebraic. According to our previous discussion, any
1dimensional complex torus is algebraic. This is no longer true for higher dimensional
tori. However, the Jacobian variety of an algebraic curve is always algebraic.
Let Λ be a lattice in C
a
. Any point in the lattice singles out univoquely a cell in the
lattice, and two opposite sides of the cell determine after identiﬁcation a closed smooth
loop in the quotient torus T = C
a
/Λ. This provides an identiﬁcation Λ · H
1
(T, Z).
Let now ξ be a skewsymmetric Zbilinear form on H
1
(T, Z). Since Hom
Z
(Λ
2
H
1
(T,
Z), Z) · H
2
(T, Z) canonically (check this isomorphism as an exercise), ξ may be re
garded as a smooth complexvalued diﬀerential 2form on T.
Proposition 8.11. The 2form ξ which on the basis ¦e
;
¦ is represented by the
intersection matrix Q
−1
is a positive (1,1) form.
Proof. If ¦e
;
, j = 1 . . . 2n¦ are the real basis vectors in C
a
generating the lattice,
they can be regarded as basis in H
1
(T, Z). They also generate 2n real vector ﬁelds on
T (after identifying C
a
with its tangent space at 0 the e
;
yield tangent vectors to T at
the point corresponding to 0; by transporting them in all points of T by left transport
one gets 2n vector ﬁelds, which we still denote by e
;
). Let ¦z
1
, . . . , z
a
¦ be the natural
local complex coordinates in T; the period matrix may be described as
Ω
j;
=
c
j
dz
j
.
After writing ξ on the basis ¦dz
j
, d¯ z
;
¦ one can check that the stated properties of ξ are
equivalent to the Riemann bilinear relations.
2
There exists on J(C) a (in principle smooth) line bundle L whose ﬁrst Chern class
is the cohomology class of ξ. This line bundle has a connection whose curvature is
(cohomologous to)
2π
j
ξ; since this form is of type (1,1), L may be given a holomorphic
structure. With this structure, it is ample by Proposition 7.3.
3
This deﬁnes a projective
imbedding of J(C), so that the latter is algebraic.
2. Elliptic curves
Consider the curve C
in C
2
given by an equation
(8.8) y
2
= P(x),
2
So we are not only proving that the Jacobian variety of an algebraic curve is algebraic, but, more
generally, that any complex torus satisfying the Riemann bilinear relations is algebraic.
3
We are using the fact that if a smooth complex vector bundle E on a complex manifold X has a
connection whose curvature has no (0,2) part, then the complex structure of X can be “lifted” to E.
Cf. [17]. Otherwise, we may use the fact that the image of the map c
1
in H
2
(J(C), Z) (the N´eronSeveri
group of J(C), cf. subsection 5.5.1) may be represented as H
2
(J(C), Z) ∩ H
1,1
(J(C), Z), i.e., as the
group of integral 2classes that are of Hodge type (1,1). The class of ξ is clearly of this type.
2. ELLIPTIC CURVES 117
where x, y are the standard coordinates in C
2
, and P(x) is a polynomial of degree 3.
By writing the equation (8.8) in homogeneous coordinates, C
may be completed to an
algebraic curve C imbedded in P
2
— a cubic curve in P
2
. Let us assume that C is
smooth. By the genus formula we see that C is an elliptic curve.
Exercise 8.1. Show that ω = dx/y is a nowhere vanishing abelian diﬀerential on
C. After proving that all elliptic curves may be written in the form (8.8), this provides
another proof of the triviality of the canonical bundle of an elliptic curve. (Hint: around
each branch point, z =
P(x) is a good local coordinate...)
The equation (8.8) moreover exhibits C as a cover of P
1
, which is branched of order
2 at the points where y = 0 and at the point at inﬁnity. One also checks that the point
at inﬁnity is a smooth point. We want to show that every smooth elliptic curve can be
realized in this way.
So let C be a smooth elliptic curve. If we ﬁx a point p in C and consider the exact
sequence of sheaves on C
0 → O(p) → O(2p) → k
j
→ 0 ,
proceeding as usual (Serre duality and vanishing theorem) one shows that H
0
(C, O(2p))
is nonzero. A nontrivial section f can be regarded as a global meromorphic function
holomorphic in C−¦p¦, having a double pole at p. Moreover we ﬁx a nowhere vanishing
holomorphic 1form ω (which exists because K is trivial). We have
Res
j
(fω) = 0 .
We realize C as C/Λ; these singles out a complex coordinate z on the open subset of C
corresponding to the fundamental cell of the lattice Λ. Then we may choose ω = dz,
and f may be chosen in such a way that
f(z) =
1
z
2
+O(z) .
On the other hand, the meromorphic function df/ω is holomorphic outside p, and has
a triple pole at p. We may choose constants a, b, c such that
˜
f = a
df
ω
+bf +c =
1
z
3
+O(z) .
The line bundle O(3p) is very ample, i.e., its complete linear system realizes the Kodaira
imbedding of C into projective space. By RiemannRoch and the vanishing theorem we
have h
0
(3p) = 3, so that C is imbedded into P
2
. To realize explicitly the imbedding we
may choose three global sections corresponding to the meromorphic functions 1, f,
˜
f.
We shall see that these are related by a polynomial identity, which then expresses the
equation cutting out C in P
2
.
We indeed have, for suitable constants α, β, γ,
˜
f
2
=
1
z
6
+
α
z
2
+O(
1
z
), f
3
=
1
z
6
+
β
z
3
+
γ
z
2
+O(
1
z
) ,
118 8. ALGEBRAIC CURVES II
so that, setting δ = α −β,
˜
f
2
+β
˜
f −f
3
+δf = O(
1
z
) .
So the meromorphic function in the lefthand side is holomorphic away from p, and has
at p a simple pole. Such a function must be constant, otherwise it would provide an
isomorphism of C with the Riemann sphere.
Thus C may be described as a locus in P
2
whose equation in aﬃne coordinates is
(8.9) y
2
+βy = x
3
−δx +
for a suitable constant . By a linear transformation on y we may set β = 0, and then
by a linear transformation of x we may set the two roots of the polynomial in the right
hand side of (8.9) to 0 and 1. So we express the elliptic curve C in the standard form
(Weierstraß representation)
4
(8.10) y
2
= x(x −1)(x −λ) .
Exercise 8.2. Determine for what values of the parameter λ the curve (8.10) is
smooth.
We want to elaborate on this construction. Having ﬁxed the complex coordinate
z, the function f is basically ﬁxed as well. We call it the Weierstraß {function. Its
derivative is {
= −2
˜
f. Notice that { cannot contain terms of odd degree in its Laurent
expansion, otherwise {(z) − {(−z) would be a nonconstant holomorphic function on
C. So
{(z) =
1
z
2
+az
2
+bz
4
+O(z
6
)
{
(z) = −
2
z
3
+ 2az + 4bz
3
+O(z
5
)
({(z))
3
=
1
z
6
+
3a
z
2
+ 3b +O(z
2
)
({
(z))
2
=
4
z
6
−
8a
z
2
−16b +O(z)
for suitable constants a, b. From this we see that { satisﬁes the condition
({
)
2
−4{
3
+ 20 a = constant
one usually writes g
2
for 20 a and g
3
for the constant in the righthand side.
In terms of this representation we may introduce a map j : ´
1
→ C, where ´
1
is
the set of isomorphism classes of smooth elliptic curves (the moduli space of genus one
4
Even though the Weierstraß representation only provides the equation of the aﬃne part of an
elliptic curve, the latter is nevertheless completely characterized. It is indeed true that any aﬃne plane
curve can be uniquely extended to a compact curve by adding points at inﬁnity, as one can check by
elementary considerations.
2. ELLIPTIC CURVES 119
curves)
5
j(C) =
1728 g
3
2
g
3
2
−27 g
2
3
.
One shows that this map is bijective; in particular ´
1
gets a structure of complex
manifold. The number j(C) is called the jinvariant of the curve C. We may therefore
say that the moduli space ´
1
is isomorphic to C.
6
Exercise 8.3. Write the jinvariant as a function of the parameter λ in equation
(8.10). Do you think that λ is a good coordinate on the moduli space ´
1
?
The holomorphic map
ψ: C → P
2
, z → [1, {(z), {
(z)]
imbeds C into P
2
as the cubic curve cut out by the polynomial
F = y
2
−4x
3
+g
2
x +g
3
(we use the same aﬃne coordinates as in the previous representation). Since
˜
f = df/ω
we have
ω =
dx
y
and the inverse of ψ is the Abel map,
7
ψ
−1
(p) =
j
j
0
dx
y
mod Λ
having chosen p
0
at the point at inﬁnity, p
0
= ψ(0) = [0, 0, 1].
In terms of this construction we may give an elementary geometric visualization of
the group law in an elliptic curve. Let us choose p
0
as the identity element in C. We
shall denote by ¯ p the element p ∈ C regarded as a group element (so ¯ p
0
= 0). By Abel’s
theorem, Proposition 8.3, we have that
¯ p
1
+ ¯ p
2
+ ¯ p
3
= 0 if and and only if p
1
+p
2
+p
3
∼ 3 p
0
(indeed one may think that ¯ p = µ(p), and one has µ(p
1
+p
2
+p
3
−3 p
0
) = 0).
Let M(x, y) = mx + ny + q be the equation of the line in P
2
through the points
p
1
, p
2
, and let p
4
be the further intersection of this line with C ⊂ P
2
. The function
M(z) = M({(z), {
(z)) on C vanishes (of order one) only at the points p
1
, p
2
, p
4
, and
has a pole at p
0
. This pole must be of order three, so that the divisor of M(z) is
p
1
+p
2
+p
4
−3 p
0
, i.e
˙
, p
1
+p
2
+p
4
−3 p
0
∼ 0.
5
The fancy coeﬃcient 1728 comes from arithmetic geometry, where the theory is tailored to work
also for ﬁelds of characteristic 2 and 3.
6
By uniformization theory one can also realize this moduli space as a quotient H/Sl(2, Z), where H is
the upper half complex plane. This is not contradictory in that the quotient H/Sl(2, Z) is biholomorphic
to C! (Notice that on the contrary, H and C are not biholomorphic). Cf. [10].
7
One should bear in mind that we have identiﬁed C with a quotient C/Λ.
120 8. ALGEBRAIC CURVES II
If p
1
+ p
2
+ p
3
∼ 3 p
0
, then p
3
∼ p
4
, so that p
3
= p
4
, and p
1
, p
2
, p
3
are collinear.
Vice versa, if p
1
, p
2
, p
3
are collinear, p
1
+p
2
+p
3
−3 p
0
is the divisor of the meromorphic
function M, so that p
1
+p
2
+p
3
−3 p
0
∼ 0. We have therefore shown that ¯ p
1
+¯ p
2
+¯ p
3
= 0
if and only if p
1
, p
2
, p
3
are collinear points in P
2
.
Example 8.4. Let C be an elliptic curve having a Weierstraß representation y
2
=
x
3
−1. C is a double cover of P
1
, branched at the three points
p
1
= (1, 0), p
2
= (α, 0), p
3
= (α
2
, 0)
(where α = e
2πj/3
) and at the point at inﬁnity p
0
. The points p
1
, p
2
, p
3
are collinear, so
that ¯ p
1
+ ¯ p
2
+ ¯ p
3
= 0.
The two points q
1
= (0, i), q
2
= (0, −i) lie on C. The line through q
1
, q
2
intersects
C at the point at inﬁnity, as one may check in homogeneous coordinates. So in this case
the elements ¯ q
1
, ¯ q
2
are one the inverse of the other, and q
1
+q
2
∼ 2 p
0
. More generally,
if q ∈ C is such that ¯ q = −¯ p, then p + q ∼ 2 p
0
, and q is the further intersection of C
with the line going through p, p
0
; if p = (a, b), then q = (a, −b). So the branch points
p
j
are 2torsion elements in the group, 2 ¯ p
j
= 0.
3. Nodal curves
In this section we show how (plane) curve singularities may be resolved by a pro
cedure called blowup.
3.1. Blowup. Blowing up a point in a variety
8
means replacing the point with all
possible directions along which one can approach it while moving in the variety. We
shall at ﬁrst consider the blowup of C
2
at the origin; since this space is 2dimensional,
the set of all possible directions is a copy of P
1
. Let x, y be the standard coordinates in
C
2
, and w
0
, w
1
homogeneous coordinates in P
1
. The blowup of C
2
at the origin is the
subvariety Γ of C
2
P
1
deﬁned by the equation
xw
1
−y w
0
= 0 .
To show that Γ is a complex manifold we cover C
2
P
1
with two coordinate charts,
V
0
= C
2
U
0
and V
1
= C
2
U
1
, where U
0
, U
1
are the standard aﬃne charts in P
1
,
with coordinates (x, y, t
0
= w
1
/w
0
) and (x, y, t
1
= w
0
/w
1
). Γ is a smooth hypersurface
in C
2
P
1
, hence it is a complex surface. On the other hand if we put homogeneous
coordinates (v
0
, v
1
, v
2
) in C
2
, then Γ can be regarded as a open subset of the quadric in
P
2
P
1
having equation v
1
w
1
−v
2
w
0
= 0, so that Γ is actually algebraic.
8
Our treatment of the blowup of an algebraic variety is basically taken from [1].
3. NODAL CURVES 121
Since Γ is a subset of C
2
P
1
there are two projections
(8.11)
Γ
π
σ
P
1
C
2
which are holomorphic. If p ∈ C
2
−¦0¦ then σ
−1
(p) is a point (which means that there
is a unique line through p and 0), so that
σ: Γ −σ
−1
(0) → C
2
−¦0¦
is a biholomorphism.
9
On the contrary σ
−1
(0) · P
1
is the set of lines through the origin
in C
2
.
The ﬁbre of π over a point (w
0
, w
1
) ∈ P
1
is the line xw
1
−y w
0
= 0, so that π makes
Γ into the total space of a line bundle over P
1
. This bundle trivializes over the cover
¦U
0
, U
1
¦, and the transition function g : U
0
∩ U
1
→ C
∗
is g(w
0
, w
1
) = w
0
/w
1
, so that
the line bundle is actually the tautological bundle O
P
1
(−1).
This construction is local in nature and therefore can be applied to any complex
surface X (twodimensional complex manifold) at any point p. Let U be a chart around
p, with complex coordinates (x, y). By repeating the same construction we get a complex
manifold U
with projections
U
π
−−−−→ P
1
σ
U
and
σ: U
−σ
−1
(p) → U −¦p¦
is a biholomorphism, so that one can replace U by U
inside X, and get a complex
manifold X
with a projection σ: X
→ X which is a biholomorphism outside σ
−1
(p).
The manifold X
is the blowup of X at p. The inverse image E = σ
−1
(p) is a divisor
in X
, called the exceptional divisor, and is isomorphic to P
1
. The construction of the
blowup Γ shows that X
is algebraic if X is.
Example 8.1. The blowup of P
2
at a point is an algebraic surface X
1
(an example
of a Del Pezzo surface); the manifold Γ, obtained by blowing up C
2
at the origin, is
biholomorphic to X
1
minus a projective line (so X
1
is a compactiﬁcation of Γ).
3.2. Transforms of a curve. Let C be a curve in C
2
containing the origin. We
denote as before Γ the blowup of C
2
at the origin and make reference to the diagram
(8.11). Notice that the inverse image σ
−1
(C) ⊂ Γ contains the exceptional divisor E,
and that σ
−1
(C) ` E is isomorphic to C −¦0¦.
9
So, according to a terminology we have introduce in a previous chapter, the map σ is a birational
morphism.
122 8. ALGEBRAIC CURVES II
Definition 8.2. The curve σ
−1
(C) ⊂ Γ is the total transform of C. The curve
obtained by taking the topological closure of σ
−1
(C) ` E in Γ is the strict transform of
C.
We want to check what points are added to σ
−1
(C) ` E when taking the topological
closure. To this end we must understand what are the sequences in C
2
which converge to
0 that are lifted by σ to convergent sequences. Let ¦p
I
= (x
I
, y
I
)¦
I∈N
be a sequence of
points in C
2
converging to 0; then σ
−1
(x
I
, y
I
) is the point (x
I
, y
I
, w
0
, w
1
) with x
I
w
1
−
y
I
w
0
= 0. Assume that for k big enough one has w
0
= 0 (otherwise we would assume
w
1
= 0 and would make a similar argument). Then w
1
/w
0
= y
I
/x
I
, and ¦σ
−1
(p
I
)¦
converges if and only if ¦y
I
/x
I
¦ has a limit, say h; in that case ¦σ
−1
(p
I
)¦ converges to
the point (0, 0, 1, h) of E. This means that the lines r
I
joining 0 to p
I
approach the
limit line r having equation y = hk. So a sequence ¦p
I
= (x
I
, y
I
)¦ convergent to 0 lifts
to a convergent sequence in Γ if and only if the lines r
I
admit a limit line r; in that
case, the lifted sequence converges to the point of E representing the line r.
The strict transform C
of C meets the exceptional divisor in as many points as
are the directions along which one can approach 0 on C, namely, as are the tangents
at C at 0. So, if C is smooth at 0, its strict transform meets E at one point. Every
intersection point must be counted with its multiplicity: if at the point 0 the curve C
has m coinciding tangents, then the strict transform meets the exceptional divisor at a
point of multiplicity m.
Definition 8.3. Let the (aﬃne plane) curve C be given by the equation f(x, y) = 0.
We say that C has multiplicity m at 0 if the Taylor expansion of f at 0 starts at degree
m.
This means that the curve has m tangents at the point 0 (but some of them might
coincide). By choosing suitable coordinates one can apply this notion to any point of a
plane curve.
Example 8.4. A curve is smooth at 0 if and only if its multiplicity at 0 is 1. The
curves xy = 0, y
2
= x
2
and y
2
= x
3
have multiplicity 2 at 0. The ﬁrst two have two
distinct tangents at 0, the third has a double tangent.
If the curve C has multiplicity m at 0 than it has m tangents at 0, and its strict
transform meets the exceptional divisor of Γ at m points (notice however that these
points are all distinct only if the m tangents are distincts).
Definition 8.5. A singular point of a plane curve C is said to be nodal if at that
point C has multiplicity 2, and the two tangents to the curve at that point are distinct.
Exercise 8.6. With reference to equation (8.10), determine for what values of λ
the curve has a nodal singularity.
3. NODAL CURVES 123
Exercise 8.7. Show that around a nodal singularity a curve is isomorphic to an
open neighbourhood of the origin of the curve xy = 0 in C
2
.
Example 8.8. (Blowing up a nodal singularity.) We consider the curve C ⊂ C
2
having equation x
3
+ x
2
− y
2
= 0. This curve has multiplicity 2 at the origin, and its
two tangents at the origin have equations y = ±x. So C has a nodal singularity at the
origin. We recall that Γ is described as the locus
¦(u, v, w
0
, w
1
) ∈ C
2
P
1
[ uw
0
= v w
1
¦ .
The projection σ is described as
(8.12)
x = u
y = uw
0
/w
1
x = v w
1
/w
0
y = v
in Γ∩V
1
and Γ∩V
0
, respectively. By substituting the ﬁrst of the representations (8.12)
into the equation of C we obtain the equation of the restriction of the total transform
to Γ ∩ U
1
:
u
2
(u + 1 −t
2
) = 0
where t = w
0
/w
1
. u
2
= 0 is the equation of the exceptional divisor, so that the
equation of the strict transform is u + 1 − t
2
= 0. By letting u = 0 we obtain the
points (0, 0, 1, 1) and (0, 0, 1, −1) as intersection points of the strict transform with the
exceptional divisor. By substituting the second representation in eq. (8.12) we obtain
the equation of the total transform in Γ ∩ U
0
; the strict transform now has equation
t
3
v +t
2
−1, yielding the same intersection points.
The total transform is a reducible curve, with two irreducible components which
meet at two points.
Exercise 8.9. Repeat the previous calculations for the nodal curve xy = 0. In
particular show that the total transform is a reducible curve, consisting of the excep
tional divisor and two more genus zero components, each of which meets the exceptional
divisor at a point.
Example 8.10. (The cusp) Let C be curve with equation y
2
= x
3
. This curve
has multiplicity 2 at the origin where it has a double tangent.
10
Proceeding as in the
previous example we get the equation v t
3
= 1 for C
in Γ∩V
0
, so that C
does not meet
E in this chart. In the other chart the equation of C
is t
2
= u, so that C
meets E at
the point (0, 0, 0, 1); we have one intersection point because the two tangents to C at
the origin coincide.
The strict transform is an irreducible curve, and the total transform is a reducible
curve with two components meeting at a (double) point.
10
Indeed this curve can be regarded as the limit for α → 0 of the family of nodal curves x
3
+α
2
x
2
−
y
2
= 0, which at the origin are tangent to the two lines y = ±αx.
124 8. ALGEBRAIC CURVES II
3.3. Normalization of a nodal plane curve. It is clear from the previous ex
amples that the strict transform of a plane nodal curve C (i.e., a plane curve with only
nodal singularities) is again a nodal curve, with one less singular point. Therefore after
a ﬁnite number of blowups we obtain a smooth curve N, together with a birational
morphism π: N → C. N is called the normalization of C.
Example 8.11. Let us consider the smooth curve C
0
in C
2
having equation y
2
=
x
4
−1. Projection onto the xaxis makes C
0
into a double cover of C, branched at the
points (±1, 0) and (±i, 0). The curve C
0
can be completed to a projective curve simply
by writing its equation in homogeneous coordinates (w
0
, w
1
, w
2
) and considering it as
a curve C in P
2
; we are thus compactifying C
0
by adding a point at inﬁnity, which in
this case is not a branch point. The equation of C is
w
2
0
w
2
2
−w
4
1
+w
4
0
= 0 .
This curve has genus 1 and is singular at inﬁnity (as one could have alredy guessed since
the genus formula for smooth plane curves does not work); indeed, after introducing
aﬃne coordinates ξ = w
0
/w
2
, η = w
1
/w
2
(in this coordinates the point at inﬁnity on
the xaxis is η = ξ = 0) we have the equation
ξ
2
= η
4
−ξ
4
showing that C is indeed singular at inﬁnity. One can redeﬁne the coordinates ξ, η so
that C has equation
(ξ −η
2
)(ξ +η
2
) = 0
showing that C is a nodal curve. Then it can be desingularized as in Example 8.8.
A genus formula. We give here, without proof, a formula which can be used to
compute the genus of the normalization N of a nodal curve C. Assume that N has t
irreducible components N
1
, . . . , N

, and that C has δ singular points. Then:
g(C) =

¸
1
g(N
j
) + 1 −t +δ.
For instance, by applying this formula to Example 8.8, we obtain that the normalization
is a projective line.
Bibliography
[1] F. Acquistapace, F. Broglia & F. Lazzeri, Topologia delle superﬁcie algebriche in
P
3
(C), Quaderni dell’Unione Matematica Italiana 13, Pitagora, Bologna 1979.
[2] C. Bartocci, U. Bruzzo & D. Hern´andez Ruip´erez, The Geometry of Supermani
folds, Kluwer Acad. Publishers, Dordrecht 1991.
[3] R. Bott & L. Tu, Diﬀerential Forms in Algebraic Topology, SpringerVerlag, New
York 1982.
[4] G.E. Bredon, Sheaf theory, McGrawHill, New York 1967.
[5] R. Godement, Th´eorie des Faisceaux, Hermann, Paris 1964.
[6] B. Gray, Homotopy theory, Academic Press, New York 1975.
[7] M.J. Greenberg, Lectures on Algebraic Topology, Benjamin, New York 1967.
[8] A. Grothendieck, Sur quelques points d’alg`ebre homologique, Tˆohoku Math. J. 9
(1957) 119221.
[9] P. A. Griﬃths & J. Harris, Principles of Algebraic Geometry, Wiley, New York
1994.
[10] R. Hain, Moduli of Riemann surfaces, transcendental aspects. ICTP Lectures Notes,
School on Algebraic Geometry, L. G¨ottsche ed., ICTP, Trieste 2000 (available from
http://www.ictp.trieste.it/ pub off/lectures/).
[11] R. Hartshorne, Algebraic geometry, SpringerVerlag, New York 1977.
[12] P. J. Hilton & U. Stammbach, A Course in Homological Algebra, SpringerVerlag,
New York 1971.
[13] F. Hirzebruch, Topological Methods in Algebraic Geometry, SpringerVerlag, Berlin
1966.
[14] S.T. Hu, Homotopy theory, Academic Press, New York 1959.
[15] D. Husemoller, Fibre Bundles, McGrawHill, New York 1966.
[16] S. Kobayashi, Diﬀerential Geometry of Complex Vector Bundles, Princeton Uni
versity Press, Princeton 1987.
[17] S. Kobayashi & K. Nomizu, Foundations of Diﬀerential Geometry, Vol. I, Inter
science Publ., New York 1963.
125
126 BIBLIOGRAPHY
[18] W.S. Massey, Exact couples in algebraic topology, I, II, Ann. Math. 56 (1952),
3363396; 57 (1953), 248286.
[19] E.H. Spanier,Algebraic topology, Corrected reprint, SpringerVerlag, New York
Berlin 1981.
[20] B.R. Tennison, Sheaf theory, London Math. Soc. Lect. Notes Ser. 20, Cambridge
Univ. Press, Cambridge 1975.
[21] B.L. Van der Waerden, Algebra, Frederick Ungar Publ. Co., New York 1970.
[22] R.O. Wells Jr., Diﬀerential analysis on complex manifolds, SpringerVerlag, New
YorkBerlin 1980.
La ﬁlosoﬁa ` scritta in questo grandissimo libro che continuamente e ci sta aperto innanzi a gli occhi (io dico l’universo), ma non si pu` o intendere se prima non si impara a intender la lingua, e conoscer i caratteri, ne’ quali ` scritto. Egli ` scritto in lingua matematica, e e e i caratteri son triangoli, cerchi, ed altre ﬁgure geometriche, senza i quali mezi ` impossibile a intenderne umanamente parola; senza e questi ` un aggirarsi vanamente per un oscuro laberinto. e Galileo Galilei (from “Il Saggiatore”)
i
Preface
These notes assemble the contents of the introductory courses I have been giving at SISSA since 1995/96. Originally the course was intended as introduction to (complex) algebraic geometry for students with an education in theoretical physics, to help them to master the basic algebraic geometric tools necessary for doing research in algebraically integrable systems and in the geometry of quantum ﬁeld theory and string theory. This motivation still transpires from the chapters in the second part of these notes. The ﬁrst part on the contrary is a brief but rather systematic introduction to two topics, singular homology (Chapter 2) and sheaf theory, including their cohomology (Chapter 3). Chapter 1 assembles some basics fact in homological algebra and develops the ﬁrst rudiments of de Rham cohomology, with the aim of providing an example to the various abstract constructions. Chapter 4 is an introduction to spectral sequences, a rather intricate but very powerful computation tool. The examples provided here are from sheaf theory but this computational techniques is also very useful in algebraic topology. I thank all my colleagues and students, in Trieste and Genova and other locations, who have helped me to clarify some issues related to these notes, or have pointed out mistakes. In this connection special thanks are due to Fabio Pioli. Most of Chapter 3 is an adaptation of material taken from [2]. I thank my friends and collaborators Claudio Bartocci and Daniel Hern´ndez Ruip´rez for granting permission to use that material. a e I thank Lothar G¨ttsche for useful suggestions and for pointing out an error and the o students of the 2002/2003 course for their interest and constant feedback. Genova, 4 December 2004
.
Cohomology of sheaves Chapter 4. Introduction to sheaves and their cohomology 1. Relative homology 3. The bidegree and the ﬁveterm sequence 4. Excision Chapter 3. Elements of homological algebra 2. KodairaSerre duality 8. Some applications Part 2. Chern classes of vector bundles 7. Basic deﬁnitions and examples 2. Introductory material 1. Connections iii . Elementary homotopy theory Chapter 2.Contents Part 1. Complex manifolds and vector bundles 1. Chern class of line bundles 6. The spectral sequence of a ﬁltered complex 3. MayerVietoris sequence in de Rham cohomology 4. Singular homology 2. De Rham cohomology 3. Introduction to algebraic geometry Chapter 5. Spectral sequences 1. Some properties of complex manifolds 3. Algebraic Topology 1 3 3 7 10 11 17 17 25 28 32 37 37 43 53 53 54 58 59 62 67 69 69 72 73 73 77 79 81 82 Chapter 1. Filtered complexes 2. Singular homology theory 1. Presheaves and sheaves 2. The spectral sequences associated with a double complex 5. Dolbeault cohomology 4. The MayerVietoris sequence 4. Holomorphic vector bundles 5.
Some general results about algebraic curves Chapter 8. Divisors 1. Divisors on Riemann surfaces 2. The Jacobian variety 2. RiemannRoch theorem 3. Elliptic curves 3.iv CONTENTS Chapter 6. Divisors on higherdimensional manifolds 3. Nodal curves Bibliography 87 87 94 95 97 101 101 104 105 111 111 116 120 125 . Algebraic curves I 1. The adjunction formula Chapter 7. Algebraic curves II 1. Linear systems 4. The Kodaira embedding 2.
Part 1 Algebraic Topology .
.
As its name suggests. A diﬀerential on M is a morphism d : M → M of Rmodules such that d2 ≡ d ◦ d = 0.1.g.1. Let R be a ring. the basic idea in algebraic topology is to translate problems in topology into algebraic ones. The future developments we have in mind are the applications to algebraic geometry. p is surjective. and is homotopic invariance is proved. the theory of spectral sequences). We say that two Rmodule morphisms i : M → M . d). M be Rmodules. The condition d2 = 0 implies 3 .2. hopefully easier to deal with. Z(M. −−→ 0 −− i p 1. and ker p = Im i. Elements of homological algebra 1. De Rham cohomology is introduced as a ﬁrst example of a cohomology theory.. Exact sequences of modules. respectively. and let M . d) ≡ Im d are called cochains. but also students interested in modern theoretical physics may ﬁnd here useful material (e. Diﬀerential complexes. In this chapter we give some very basic notions in homological algebra and then introduce the fundamental group of a topological space. d) is called a diﬀerential module. A morphism of exact sequences is a commutative diagram 0 −−→ M −−→ M −−→ M −−→ 0 −− −− −− −− 0 −−→ N −−→ N −−→ N −− −− −− of Rmodule morphisms whose rows are exact. cocycles and coboundaries of (M. and M an Rmodule. if i is injective. Let R be a ring. p : M → M form an exact sequence of Rmodules. The elements of the spaces M . M . especially the singular homology of topological spaces. The pair (M. d) ≡ ker d and B(M. and write 0 → M −−→ M −−→ M → 0 . Definition 1. 1.CHAPTER 1 Introductory material The aim of the ﬁrst part of these notes is to introduce the student to the basics of algebraic topology.
A morphism of diﬀerential modules maps cocycles to cocycles and coboundaries to coboundaries. omitting the diﬀerential d when there is no risk of confusion. d) is called the cohomology group of the diﬀerential module (M. thus deﬁning a morphism δ : H(M ) → H(M ). . A complex of Rmodules is a diﬀerential Rmodule (M • .4. The groups H n (M • ) are called the cohomology groups of the complex M • . Now. d) ⊂ Z(M. d). One proves by direct computation that the triangle is exact. d)/B(M. −−→ M n−1 −−→ M n −−→ M n+1 −−→ . Definition 1. If m is an element of M such that p(m) = m . Proposition 1. The above results can be translated to the setting of complexes of Rmodules. dn−2 dn−1 d dn+1 For a complex M • the cocycle and coboundary modules and the cohomology group split as direct sums of terms Z n (M • ) = ker dn . which is independent of the choices we have made. There exists a morphism δ : H(M ) → H(M ) (called connecting morphism) and an exact triangle of cohomology H(M ) O H(i) i p / H(M ) tt tt tt t ytt δ H(p) H(M ) Proof. Let (M. We shall often write Z(M ). INTRODUCTORY MATERIAL that B(M. thus inducing a morphism H(f ) : H(M ) → H(M ). d) = Z(M. d). d) which is Zgraded. and the Rmodule H(M. . We shall usually write a complex of Rmodules in the more pictorial form n . . A morphism of diﬀerential modules is a morphism f : M → M of Rmodules which commutes with the diﬀerentials. B n (M • ) = Im dn−1 and H n (M • ) = Z n (M • )/B n (M • ) respectively. the cocycle m deﬁnes a cohomology class δ(ξ ) in H(M ). Let 0 → M −−→ M −−→ M → 0 be an exact sequence of differential Rmodules. and whose diﬀerential fulﬁlls d(M n ) ⊂ M n+1 for every n ∈ Z. Definition 1. B(M ) and H(M ).4 1. The construction of δ is as follows: let ξ ∈ H(M ) and let m be a cocycle whose class is ξ . . d) and (M . we have p(d(m)) = d(m ) = 0 and then d(m) = i(m ) for some m ∈ M which is a cocycle.2. . f ◦ d = d ◦ f .3. M • = n∈Z M n . d ) be diﬀerential Rmodules.
for every k. then H(f ) = H(g). nonisomorphic) complexes may nevertheless have isomorphic cohomologies.3. While this condition is not necessary. . fn+1 i p δn−1 H(i) H(p) δ δ H(i) H(p) δn+1 Proof. Proposition 1.. f.5.e.3 splits into morphisms δn : H n (P • ) → H n+1 (N • ) (indeed the connecting morphism increases the degree by one) and the long exact sequence of the statement is obtained by developing the exact triangle of cohomology introduced in Proposition 1. A suﬃcient conditions for this to hold is that the two complexes are homotopic. .. The situation is depicted in the following commutative diagram.1. (M • . . If there is a homotopy between f and g. A morphism of complexes of Rmodules f : N • → M • is a collection of morphisms {fn : N n → M n  n ∈ Z}. −−→ H n (N • ) −−→ H n (M • ) −−→ H n (P • ) −−→ n −−→ H n+1 (N • ) −−→ H n+1 (M • ) −−→ H n+1 (P • ) −−→ . in practice the (by far) commonest way to prove the isomorphism between two cohomologies is to exhibit a homototopy between the corresponding complexes. Given two complexes of Rmodules.7.3 takes the following form: Proposition 1.8. . g : M • → N • ... HOMOLOGICAL ALGEBRA 5 Definition 1. homotopic morphisms induce the same morphism in cohomology. d) and (N • . The connecting morphism δ : H • (P • ) → H • (N • ) deﬁned in Proposition 1.. and two morphisms of complexes. d ). Let 0 → N • −−→ M • −−→ P • → 0 be an exact sequence of complexes of Rmodules. M n+1 − − → N n+1 −− For complexes. Proposition 1.. a morphism K : N k → M k−1 ) such that d ◦ K + K ◦ d = f − g.. / .e.6. namely.. Definition 1. a homotopy between f and g is a morphism K : N • → M •−1 (i..3. 1. . Homotopies. . . There exist connecting morphisms δn : H n (P • ) → H n+1 (N • ) and a long exact sequence of cohomology n .. Diﬀerent (i. / M k−1 d / / Mk M k+1 ww ww K ww K ww g w f w ww ww {w {w / Nk / N k+1 / N k−1 d d d / . such that the following diagram commutes: Mn d −−→ −− fn Nn d .
In the case of complexes bounded below zero (i. and the zero morphism. d ).13. such that: f ◦ g : N • → N • is homotopic to the identity map idN .12. INTRODUCTORY MATERIAL Proof. and it may happen that H 0 (M.e. the reader may easily check that this change of sign is immaterial. Then H(f )(ξ) = [f (m)] = [g(m)] + [d (K(m))] + [K(dm)] = [g(m)] = H(g)(ξ) since dm = 0.. one can state that if a homotopy K : M k → M k−1 exists for k ≥ k0 .10.6 1. all its cohomology groups vanish. Proof. d) = 0 for k ≥ k0 . g : N • → M • .14. Remark 1. more explicitly.e. Two homotopic complexes have isomorphic cohomologies. it is a morphism K : M • → M •−1 such that d ◦ K + K ◦ d = idM . (M • . A homotopy of a complex of Rmodules (M • . One has H(f ) ◦ H(g) = H(f ◦ g) = H(idN ) = idH ( N ) H(g) ◦ H(f ) = H(g ◦ f ) = H(idM ) = idH ( M ) so that both H(f ) and H(g) are isomorphism. One could use the previous deﬁnitions and results to yield a proof. . One might as well deﬁne a homotopy by requiring d ◦K −K ◦d = . then H k (M. Proposition 1. d) and (N • . then d(K(m)) = m − K(dm) = m so that m is also a coboundary. .11.9. are said to be homotopically equivalent (or homotopic) if there exist morphisms f : M • → N • . d) admits a homotopy. one also says that the complex is acyclic). Corollary 1. Definition 1. then it is exact (i. .. Examples of such situations will be given later in this chapter. d). Definition 1. We use the notation of the previous Deﬁnition. M = ⊕k∈N M k ) often a homotopy is deﬁned only for k ≥ 1. d) is a homotopy between the identity morphism on M . d) = 0. Proof. g ◦ f : M • → M • is homotopic to the identity map idM . but it is easier to note that if m ∈ M k is a cocycle (so that dm = 0). If a complex of Rmodules (M • . [d (K(m))] = 0. More generally. Let ξ = [m] ∈ H k (M • . Remark 1. . Two complexes of Rmodules.
12 and Remark 1. d)). We may easily compute the cohomology of the Euclidean spaces Rn . Let Ωk (X) be the vector space of diﬀerential kforms on X. we have HdR (X) = RC . there is a way to compute the de Rham cohomology of X out of the de Rham cohomology of X1 and X2 (K¨nneth u theorem. if ω is closed. For later use. Since k Ωk (X) = 0 for k > n and k < 0.3. compact and orientable n ndimensional manifold. then HdR (X) R. Show that the in1 tegration of 1forms on S 1 yields an isomorphism HdR (S 1 ) R. and let d : Ωk (X) → Ωk+1 (X) be the exterior diﬀerential. If f : X → Y is a smooth morphism of diﬀerentiable manifolds. so that ω is exact. if X is a connected. DE RHAM COHOMOLOGY 7 2. e Proof.ik (tx) dt xi1 dxi2 ∧ · · · ∧ dxik . 1 (Kω)(x) = k 0 tk−1 ωi1 i2 . the pullback morphism f ∗ : Ωk (Y ) → Ωk (X) commutes with the exterior diﬀerential. d) → (Ω• (X). Realize the circle S 1 as the unit circle in R2 . (Poincar´ lemma) HdR (Rn ) = 0 for k > 0.1. we have ω = dKω.. and all x ∈ Rn . Explicitly. De Rham cohomology As an example of a cohomology theory we may consider the de Rham cohomology of a diﬀerentiable manifold X. X = X1 × X2 . so that. since ker[d : Ω0 (X) → Ω1 (X)] is formed by the locally constant functions on 0 X. cf. where C is the number of connected components of X. Proposition 1. k then HdR (X × R) . whose cohomology k groups are denoted HdR (X) and are called the de Rham cohomology groups of X. k Proposition 1. for any kform ω ∈ Ωk (Rn ). Exercise 1. If a manifold is a cartesian product. Moreover.13. One easily shows that dK + Kd = Id. This will serve also as an example of the notion of homotopy between complexes. this means that K is a homotopy of the de Rham complex of Rn deﬁned for k ≥ 1. the corresponding morph• • ism H(f ) : HdR (Y ) → HdR (X) is usually denoted f . we prove here a very particular case. k HdR (X) for all k ≥ 0. If X is a diﬀerentiable manifold. k ≥ 1. [3]).. Then (Ω• (X). all cohomology groups vanish in positive degree. We deﬁne a linear operator K : Ωk (Rn ) → Ωk−1 (Rn ) by letting. thus giving rise to a morphism of diﬀerential complexes (Ω• (Y ).2. the groups HdR (X) vanish for k > n and k < 0. Of course one 0 has HdR (Rn ) = ker[d : C ∞ (Rn ) → Ω1 (Rn )] = R.2. according to Proposition 1. d) is a diﬀerential complex of Rvector spaces (the de Rham complex). This argument can be quite easily generalized to show that.
every kform ω on X × R can be written as (1. so that the complexes Ω• (X) and Ω• (X × R) are homotopic. By a similar argument one proves that for all k > 0 k HdR (X × S 1 ) k−1 k HdR (X) ⊕ HdR (X). g are functions on X ×R. ..2 after which one gets d ◦ K + K ◦ d = idΩ• (X×R) − p∗ ◦ s∗ . 2The reader may consult e.5. Let rk : Ωk (X) → Ωk (Y ) be the restriction morphism. this morphism is homotopic to idΩ• (X×R) .1)) t K(ω) = (−1)k 0 g(x. e Exercise 1. p2 the projections of X × R onto its two factors.g.4. 1 This homotopy K : Ω• (X × R) → Ω•−1 (X × R) is deﬁned as (with reference to equation (1. thus proving our claim as a consequence of Corollary 1. 2 The proof that K is a homotopy is an elementary direct computation. 1 In particular we obtain that the morphisms • • p1 : HdR (X) → HdR (X × R). 1 1 However. Y ) → 1In intrinsic notation this means that Ωk (X × R) C ∞ (X × R) ⊗C ∞ (X) [Ωk (X) ⊕ Ωk−1 (X)]. after letting Ωk (X. Denoting by p1 . §I.1 Let s : X → X ×R be the section s(x) = (x. 0) p∗ ω1 ). Since the exterior diﬀerential commutes with the restriction. Let t a coordinate on R. Now we give an example of a long cohomology exact sequence within de Rham’s theory. Let X be a diﬀerentiable manifold. This is not the identity (as a matter of fact one. has p∗ ◦s∗ (ω) = f (x. ω2 ∈ Ωk−1 (X). One has p1 ◦ s = idX (i. Y ) = ker rk a diﬀerential d : Ωk (X. • • s : HdR (X × R) → HdR (X×) are isomorphisms. and Y a closed submanifold. hence s∗ ◦ p∗ : Ω• (X) → Ω• (X) is the identity.1) ω = f p∗ ω1 + g p∗ ω2 ∧ p∗ dt 1 1 2 where ω1 ∈ Ωk (X). INTRODUCTORY MATERIAL Proof. s) ds p∗ ω2 .4. while idΩ• (X) is deﬁnitely homotopic to itself. [3]. and f . s is indeed a section of p1 ).8 1. So we only need to exhibit a homotopy between p∗ ◦ s∗ and idΩ• (X×R) . Remark 1. 0). We also have a morphism p∗ ◦ s∗ : Ω• (X × R) → 1 1 Ω• (X×R).10.e. this is surjective. If we take X = Rn and make induction on n we get another proof of Poincar´ lemma.
S 1 ) → 0 r δ δ where the morphism r is an isomorphism. d ) is called the relative de Rham complex. we have HdR (R2 . Y ) 1 1 2 → HdR (X) → HdR (Y ) → HdR (X.. Y ) → . HdR (S 1 ) HdR (S 1 ) R. Y ) = 0 for k ≥ n + 1. S 1 ) → HdR (R2 ) → HdR (S 1 ) → HdR (R2 . Compute HdR (R2 . Given the standard embedding of S 1 into R2 . DE RHAM COHOMOLOGY 9 Ωk+1 (X. 0 then HdR (X. From this example we may abstract the fact that whenever X and Y are connected. Y ) d / Ωk−1 (X) d rk−1 / Ωk−1 (Y ) d /0 0 / Ωk (X. 3 and its k cohomology groups by HdR (X. S 1 ) → HdR (R2 ) → 0 . Y ) is deﬁned. We have the long exact sequence 1 0 0 0 0 → HdR (R2 . . Y ) is for all k ≥ 0 the kernel of rk restricted to Z k (X). Since HdR (R2 ) R. One has a long cohomology exact sequence 0 0 0 1 0 → HdR (X. S 1 ) → 0 2 0 → R → HdR (R2 . Y ) = 0.7. Y ). We have therefore an exact sequence of diﬀerential modules. cf. Example 1. δ δ Exercise 1.8. S 1 ) R. k and that HdR (X. . k 0 As in the previous exercise. As a consequence HdR (X. S 1 ) = 0 for k ≥ 3. S 1 ) = 0 (as we already noticed) and HdR (R2 . S 1 ) = 0. we obtain the exact sequences 0 1 0 → HdR (R2 .2. Y ) → HdR (X) → HdR (Y ) → HdR (X. S 1 ) → R → R → HdR (R2 . S 1 ). 3Sometimes this term is used for another cohomology complex. 1 0 1 2 HdR (R2 ) = HdR (R2 ) = 0. The complex (Ω• (X. we compute the relative • cohomology HdR (R2 . Therefore from the ﬁrst sequence we get 0 1 HdR (R2 . Prove that HdR (X. . is the space of closed kforms on X 0 which vanish on Y . From the second we 2 obtain HdR (R2 .6.e. S 1 ) 2 2 1 1 → HdR (R2 ) → HdR (S 1 ) → HdR (R2 . [3]. Make an example where dim X = dim Y and check if the previous facts still hold true. 1. Y ) → Ωk+1 (X. i. n n 2. Y ). Exercise 1. Y ) are called the relative de Rham cohomology groups. Y ) / Ωk (X) rk / Ωk (Y ) /0 commutes. Consider a submanifold Y of R2 formed by two disjoint embedded • copies of S 1 . Prove that the space ker d : Ωk (X. Let n = dim X and dim Y ≤ n − 1. Y ) → HdR (X) surjects. in a such a way that the diagram 0 / Ωk−1 (X. Y ) = 0 whenever X and Y are connected.
from the second sequence.4 Exercise 1. From elementary diﬀerential geometry we recall that a partition of unity subordinated to the cover {U. and to prove it we need a partition of unity argument. One easily checks that i is injective and that ker p = Im i. δ δ 0 → Ωk (X) → Ωk (U ) ⊕ Ωk (V ) → Ωk (U ∩ V ) → 0 i p supp(f2 ) ⊂ V.1. ω2 )) = τ . 1 2 and then we get HdR (S 2 ) = 0.3) to compute the de Rham cohomology of the circle S 1 . 0 0 From the ﬁrst sequence.2) is exact.3) 0 → HdR (X) → HdR (U ) ⊕ HdR (V ) → HdR (U ∩ V ) → HdR (X) → 1 1 1 2 → HdR (U ) ⊕ HdR (V ) → HdR (U ∩ V ) → HdR (X) → . Thus the sequence (1. MayerVietoris sequence in de Rham cohomology The MayerVietoris sequence is another example of long cohomology exact sequence associated with de Rham cohomology. HdR (S 2 ) R.2. V . we can assume that U and V are diﬀeomorphic to R2 . Since S 1 × R is homotopic to S 1 . . respectively. V } of X is a pair of smooth functions f1 . The argument may be generalized to a union of several open sets. Use the MayerVietoris sequence (1. We use the MayerVietoris sequence (1. The surjectivity of p is somehow less trivial. ω2 )) = ω1U ∩V − ω2U ∩V . let ω1 = f2 τ.2) where i(ω) = (ωU . Assume that a diﬀerentiable manifold X is the union of two open subset U . . while U ∩ V S 1 × R. ωV ). since HdR (S 2 ) R. and is very useful for making computations. but not the ﬁrst). p((ω1 . For every k. we obtain a long cohomology exact sequence 0 0 0 0 1 (1. This is the MayerVietoris sequence. Then p((ω1 . it has the same de Rham cohomology.3) to compute the de Rham cohomology of the sphere S 2 (as a matter of fact we already know the 0th and 2nd group. Hence the sequence (1. 4The MayerVietoris sequence foreshadows the Cech cohomology we shall study in Chapter 3. f1 + f2 = 1. f2 : X → R such that supp(f1 ) ⊂ U. the map HdR (S 2 ) → R ⊕ R is injective. ˇ . Using suitable stereographic projections. 0 ≤ k ≤ n = dim X we have the sequence of morphisms (1. ω2 = −f1 τ.3) becomes 0 1 0 → HdR (S 2 ) → R ⊕ R → R → HdR (S 2 ) → 0 2 0 → R → HdR (S 2 ) → 0. Example 1. Since the exterior diﬀerential d commutes with restrictions. INTRODUCTORY MATERIAL 3.10 1. These kform are deﬁned on U and V . Given τ ∈ Ωk (U ∩ V ).
4. 2 ≤ t ≤ 1. x0 ) = L(x0 )/ ∼ . Let X be a topological space. S 1 ). The composition law in Lx0 descends to a group structure in the quotient π1 (X. A path in X is a continuous map γ : I → X. 4. 1) = γ2 (t). we consider in L(x0 ) an equivalence relation by decreeing that γ1 ∼ γ2 if there is a homotopy with ﬁxed end points between γ1 and γ2 .1. since the composition is not associative (composing in a diﬀerent order yields diﬀerent parametrizations). If the two paths have the same end points (i. then L(x1 ) L(x2 ). Consider X = S 2 and Y = S 1 . We deﬁne γ2 ∈ L(x2 ) by letting 1 c(1 − 3t). Let us ﬁx a base point x0 ∈ X. Again with reference with a base point x0 . A loop based at x0 is a path such that γ(0) = γ(1) = x0 .4. Γ(1. HOMOTOPY THEORY k Exercise 1. n. s) = x2 for all s ∈ I. 3 This establishes the isomorphism. Let us denote L(x0 ) th set of loops based at x0 . Homotopy of paths. A homotopy Γ between two paths γ1 . x2 ∈ X and there is a path connecting x1 with x2 . 11 R for k = 0. Exercise 1. We say that X is pathwise connected if given any two points x1 . 0) = γ1 (t). 0≤t≤ 1 2 1 2 ≤ t ≤ 1. Proposition 1. Let c be such a path. γ2 is a continuous map Γ : I × I → X such that Γ(t. 1]. The fundamental group. γ1 (0) = γ2 (0) = x1 . Elementary homotopy theory 4. We denote by I the closed interval [0. 4. Use induction to show that if n ≥ 3. • Compute the relative de Rham cohomology HdR (S 2 . Γ(t.3. we may introduce the stronger notion of homotopy with ﬁxed end points by requiring additionally that Γ(0. . 0 ≤ t ≤ 3 2 γ2 (t) = γ (3t − 1). and let γ1 ∈ L(x1 ). One can deﬁne a composition between elements of L(x0 ) by letting (γ2 · γ1 )(t) = γ1 (2t).e. If x1 .1. x2 ∈ X there is a path γ such that γ(0) = x1 . γ(1) = x2 .2. then HdR (S n ) k HdR (S n ) = 0 otherwise. Proof. This does not make L(x0 ) into a group. γ2 (2t − 1). s) = x1 . 1 ≤ t ≤ 3 3 1 c(3t − 2). embedded as an equator in S 2 . γ1 (1) = γ2 (1) = x2 ).
if X is pathwise connected the fundamental group π1 (X. In particular. st) is a homotopy between f ◦ g and idY . g the projection onto X.12 1. a homotopy between two continuous maps f. x1 ) π1 (X.1. One then says that f and g are homotopic. x0 ) → (Y. in this situation. one uses the notation π1 (X). Y homotopically equivalent. as we shall see in examples. For any manifold X. y0 ) → (X. Definition 1. f (x) = (x. Let f : (X. t) = (x. Proof. x0 ). 0). y0 ) be a homotopical equivalence. Thus. Y . in general it is nonabelian. take Y = X × R. Definition 1. 4. So X and X × R are homotopically equivalent. x2 ∈ X and there is a path connecting x1 with x2 . x0 ) → (Y. g are said to be homotopical equivalences. g : Y → X such that g ◦ f is homotopic to idX . The map f . y0 ) is a continuous map f : X → Y such that f (x0 ) = y0 . It is easy to show that a map of pointed spaces induces a group homomorphism f∗ : π(X. y0 ).5. As a consequence of Proposition 1. x0 ) → π1 (Y. Proposition 1. t) = x is a homotopy between g ◦ f and idX . x0 ) is a pair formed by a topological space X with a chosen point x0 . Of course. g : (Y. F (x. we say they are homotopically equivalent if there exist maps of pointed spaces f : (X. g : X → Y is a map F : X × I → Y such that F (x. while G : X × R × I → X × R.3. y0 ). F (x. Let γ be a loop based at x0 . we say that a pointed space (X. G(x.2. x0 ) that make the topological spaces X. Given two pointed spaces (X. 0) = f (x). y0 ) is an isomorphism.. Example 1. X is said to be simply connected if it is pathwise connected and π1 (X) = {e}.3. x0 ) be a map that realizes the homotopical equivalence. Y are homotopically equivalent if there are continuous maps f : X → Y . x0 ) → (Y. homeomorphic spaces are homotopically equivalent. A map of pointed spaces f : (X. x0 ) is the fundamental group of X with base point x0 . Homotopy of maps. The simplest example of a simply connected space is the onepoint space {∗}. (Y.4. x0 ) → (Y. s. Then F : X × I → X. to describe the behaviour of the fundamental group we need to introduce a notion of map which takes the base point into account. then π1 (X. . y0 ) → (X. Then f∗ : π∗ (X. and f ◦ g is homotopic to idY . and denote by F a homotopy between g ◦ f and idX . Let g : (Y. Since the deﬁnition of the fundamental group involves the choice of a base point. INTRODUCTORY MATERIAL π1 (X. One says that two topological spaces X. if x1 . x0 ) is independent of x0 up to isomorphism. Given two topological spaces X. y0 ). The reader should be able to concoct many similar examples. 1) = g(x) for all x ∈ X. x2 ).
p2 : X × R → R be the projections. so f = s0 ◦ F and g = s1 ◦ F .4. Then the morphisms they induce in cohomology coincide. s1 (x) = (x. Then f = F ◦ s0 . Homotopic invariance of de Rham cohomology. then f ◦ g = g ◦ f = Id. Then s0 ◦ p1 = s1 ◦ p1 = Id. 0). 4. g = F ◦ s1 . Proposition 1. HOMOTOPY THEORY 13 Then g ◦ f ◦ γ is again a loop based at x0 . 2. Then k HdR (Y ) for all k ≥ 0. Let X. 1. Definition 1. hence simply connected. so that the claim follows. g : X → Y be two homotopic smooth maps. Proof.4. A contractible space is simply connected. Hence. Lemma 1.y0 ) . We choose a homotopy between f and g in the form of a smooth 5 map F : X × R → Y such that F (x. k HdR (X) Proof. s1 : X → X × R by letting s0 (x) = (x. g are two smooth maps realizing the homotopy. then their fundamental groups are isomorphic. 1). g∗ ◦ f∗ = idπ1 (X. Corollary 1. x0 ). Rn minus a (n − 2)plane (for n ≥ 3). and let f. In the same way one proves that f∗ ◦ g∗ = idπ1 (Y. By Proposition 1. Compute the fundamental groups of the following spaces: the punctured plane (R2 minus a point). t) = f (x) if t ≤ 0.3 p1 is an isomorphism. Γ(s. Let p1 : X × R → X. If f . t) = F (γ(s). Exercise 1. . so that γ = g ◦ f ◦ γ in π1 (X. and f = F = g . 5For the fact that F can be taken smooth cf.9. f =g . and the map Γ : I × I → X. t) = g(x) if t ≥ 1 . Let X and Y be homotopic diﬀerentiable manifolds. F (x. Then s0 = s1 .6. A topological space is said to be contractible if it is homotopically equivalent to the onepoint space {∗}. so that both f and g are isomorphisms. [3]. R3 minus a line.10. t) is a homotopy between γ and g ◦ f ◦ γ.x0 ) . If two pathwise connected spaces X and Y are homotopic. Y be diﬀerentiable manifolds. We deﬁne sections s0 . We may now prove the invariance of de Rham cohomology under homotopy. Show that Rn is contractible.7.8.
13. f2 : G → G2 . g2 : G2 → G1 ∗G G2 obtained by composing the inclusions with the projection F → F/R. under some conditions. The two generators do not commute. then G1 ∗G G2 = {e}. As a prerequisite we need the notion of amalgamated product of two groups. x or the empty word. Think of the ﬁgure 8 as the union of two circles X in R2 which touch in one pount. An important tool for such computations is the van Kampen theorem. (1) Prove that if G1 = G2 = {e}. G2 be groups. The second relation kind of “glues” G1 and G2 along the images of G. . q. p. This theorem allows one. V and U ∩ V . satisfying the relation ab = cba. and that U ∩ V is pathwise connected. while U ∩ V is simply connected. p2 be points in the two respective circles. and the product is given by juxtaposition. diﬀerent from the common point. where the letters x are 2 i n 1 either in G1 or G2 . This is a simpliﬁed form of van Kampen’s theorem. 7The ﬁrst relation tells that the product of letters in the words of F are the product either in G 1 or G2 . this can also be checked “experimentally” if you think of winding a string along the 6F is the group whose elements are words x 1 x . Proposition 1. one could say that G1 ∗G G2 is the smallest subgroup generated by G1 and G2 with the identiﬁcation of f1 (g) and f2 (g) for all g ∈ G. Intuitively. (2) Let G be the group with three generators a. We compute the fundamental group of a ﬁgure 8. satisfying the relation m n m−1 n−1 p q p−1 q −1 = e. There are morphisms π1 (U ∩ V ) → π1 (U ). and G is any group. INTRODUCTORY MATERIAL 4. Then π1 (U ) π1 (V ) Z. Let F the free group generated by G1 G2 and denote by · the product in this group. . π1 (X) π1 (U ) ∗π1 (U ∩V ) π1 (V ). n. Prove that G ∗Z G is isomorphic to a group with four generators m. i = ±1. b.12. Let G. V = X − {p2 }. The van Kampen theorem. The computation of the fundamental group of a topological space is often unsuspectedly complicated. f1 (g) · f2 (g)−1 Then one deﬁnes the amalgamated product G1 ∗G G2 as F/R. π1 (U ∩ V ) → π1 (V ) induced by the inclusions. Suppose now that a pathwise connected space X is the union of two pathwise connected open subsets U . y both in G1 or G2 for g ∈ G. V . which we state without proof. Example 1. to compute the fundamental group of an union U ∪V if one knows the fundamental groups of U . c. . when this makes sense. for a full statement see [6]. There are natural maps g1 : G1 → G1 ∗G G2 .14 1.6 Let R be the normal subgroup generated by elements of the form7 (xy) · y −1 · x−1 with x. Let Z → G be the homomorphism induced by 1 → c. with ﬁxed morphisms f1 : G → G1 . Exercise 1. and one has g1 ◦ f1 = g2 ◦ f2 . Let p1 . G1 .11. and take U = X − {p1 }. It follows that π1 (X) is a free group with two generators.5.
. If G is a simply connected topological group. This gives us another way to compute the fundamental group of the ndimensional torus T n (once we know π1 (S 1 )).15. Exercise 1.16. we have thus proved that π1 (S 1 ) Z. we state some results without proof. Exercise 1. Prove that for any n ≥ 2 the sphere S n is simply connected. y0 )) π1 (X.. Use van Kampen’s theorem to compute the fundamental group of a Riemann surface of genus 2 (a compact. Prove the stronger result that X is homotopic to the 2torus. . (Y.20. Proposition 1. connected 2dimensional diﬀerentiable manifold of genus 2. Since S 1 R/Z. then π1 (X × Y. Compute the fundamental group of a 2dimensional punctured torus (a torus minus a point). Exercise 1. (x0 . Generalize your result to any genus. y0 ).. Again. and H is a normal discrete subgroup.e. x0 ) × π1 (Y. Prove that π1 (X) Z ⊕ Z. Rn /Zn . Exercise 1. Deduce that for n ≥ 3. i. More generally. Prove that the manifolds S 3 and S 2 × S 1 are not homeomorphic.18.4. and a circle linking the line.17. x0 ). the fundamental group of the corolla with n petals (n copies of S 1 all touching in a single point) is a free group with n generators. then π1 (G/H) H. Let X be the space obtained by removing a line from R2 .19. “with two handles”). given two pointed topological spaces (X. Exercise 1.14.6. Rn minus a point is simply connected. y0 ). orientable. Prove that. 4. Compute the fundamental group of R2 with n punctures. Exercise 1. Other ways to compute fundamental groups. HOMOTOPY THEORY 15 ﬁgure 8 in a proper way. In the same way we compute the fundamental group of the ndimensional torus T n = S 1 × · · · × S 1 (n times) obtaining π1 (T n ) Zn .
.
. . . i The boundary of ∆n is formed by n + 1 faces Fn (i = 0. . 0.. Pn the points in Rn P0 = 0. . The basic blocks of singular homology are the continuous maps from standard subspaces of Euclidean spaces to the topological space one considers. Singular homology In this Chapter we develop some elements of the homology theory of topological ˇ spaces. it relates to homotopy in a clear way. . there is a unique aﬃne map Rk → Rn mapping P0 .. . Fn is the face opposite to Pi .1. Pi = (0. Pk to the Q’s. The standard nsimplex ∆n may so 17 . and once the basic properties of the theory have been proved. . These faces may be labelled i by the vertex of the simplex which is opposite to them: so. If Qi = Pi for 0 ≤ i ≤ k. We shall denote by P0 . cellular. it requires very little combinatorial arguments. then the aﬃne map is the identity on Rk . . . . . 0. P1 . Singular homology has the disadvantage of appearing quite abstract at a ﬁrst contact. Alternatively. . . the computation of the homology groups is not diﬃcult. . This aﬃne map yields a singular ksimplex in Rn that we denote < Q0 . . . 0) (with just one 1 in the ith position). 0. . There are many diﬀerent homology theories (simplicial. Qk >. . one can describe ∆k as the set of points in Rn such that n xi ≥ 0. Deﬁnitions. i If Q0 . its functorial properties are evident. . . The restriction of σ to any of the faces of ∆n deﬁnes a singular (n − 1)simplex i i σi = σFn (or σ ◦ Fn if we regard Fn as a singular (n − 1)simplex). singular. . . a singular nsimplex in X is a continuous map σ : ∆n → X. n. . but in exchange of this we have the fact that it applies to any topological space. . 1. . and we denote the resulting singular ksimplex by δk . . . The convex hull of these points is denoted by ∆n and is called the standard nsimplex. .) even though these theories coincide when the topological space they are applied to is reasonably wellbehaved. 1. . .CHAPTER 2 Singular homology theory 1. . i=1 xi ≤ 1. n) which are images of the standard (n − 1)simplex by aﬃne maps Rn−1 → Rn . . 1. CechAlexander. Qk are k + 1 points in Rn . Given a topological space X. i = 1.
. . . and. R). R). . R) = Im ∂ : Sk+1 (X. We denote by Sk (X. R) → Sk (X. R) (the module of kboundaries). . via the inclusion Z → R given by the identity in R. therefore. Qk >= i=0 ˆ (−1)i < Q0 . . .1. Notice that Z0 (X. R) by letting k ∂σ = i=0 i (−1)i σ ◦ Fk for a singular ksimplex σ and exteding by Rlinearity. . Qi . ∂) is a (homology) graded diﬀerential module. Sk (X. If Q0 . . Proof. R)/Bk (X. Example 2. . and the face Fn of ∆n is the singular (n − 1)simplex ˆ < P0 . R) are the singular homology groups of X with coeﬃcients in R. R). an abelian group. So an element in Sk (X. . . R) the free group generated over R by the singular ksimplexes in X. . Let σ be a singular ksimplex. Proposition 2. Qk are k + 1 points in Rn . Proposition 2. Pn >. R) = ker ∂ : Sk (X. Choose now a commutative unital ring R. one has k ∂ < Q0 . Its homology groups Hk (X. R) (the module of kcycles). . Qk > .2. R) → Sk−1 (X. . . . R) ≡ S0 (X. So (S• (X. . . R) → Sk−1 (X. R) ⊕j Hk (Xj . . For k = 0 we deﬁne ∂σ = 0. . . We shall use the following notation and terminology: Zk (X. Hk (X. j = i − 1 the last two terms cancel. R) is a “formal” ﬁnite linear combination (called a singular chain) σ= j aj σj with aj ∈ R. . then Hk (X. . Thus. ∂ 2 = 0. R) is an Rmodule. . . Pn >. . Bk (X. R) for all k ≥ 0. R) = Zk (X. and the σj are singular ksimplexes. Basic properties.18 2. If X is the union of pathwise connected components Xj .3. .2. where the hat denotes omission. Pi . 1. For k ≥ 1 we deﬁne a morphism ∂ : Sk (X. k k k−1 ∂ σ = i=0 k 2 (−1) ∂(σ ◦ i i Fk ) = i=0 (−1) i j=0 k−1 j i (−1)j σ ◦ Fk ◦ Fk−1 = j<i=1 j i−1 (−1)i+j σ ◦ Fk ◦ Fk−1 + 0=i≤j j i (−1)i+j σ ◦ Fk ◦ Fk−1 Resumming the ﬁrst sum by letting i = j. . HOMOLOGY THEORY i also be denoted < P0 .
5. If f. then Sk (g ◦ f ) = Sk (g) ◦ Sk (f ). R)/B0 (X. so that c = j bj (γj (1) − γj (0)). then H0 (X.3. and the coeﬃcients sum up to zero. It is also easy to check that. If X is pathwise connected. It is immediate to prove that Sk (f ) ◦ ∂ = ∂ ◦ Sk+1 (f ): k+1 Sk (f )(∂σ) = f ◦ i=0 i (−1)i σ ◦ Fk+1 = ∂(f ◦ σ) = ∂(Sk (f )(σ)) . R) R. If σ is a singular ksimplex in X. their singular homologies are isomorphic.1. and (g ◦ f ) = g ◦ f . 1. Any singular ksimplex must map ∆k inside a pathwise connected components (if two points of ∆k would map to points lying in diﬀerent components. that would yield path connecting the two points). It should be by now clear that this yields as an immediate consequence the homotopic invariance of the singular homology. SINGULAR HOMOLOGY 19 Proof. R) = S0 (X. R) → R given by j aj xj → j aj . We deﬁne the morphism P in two steps. R) for every k ≥ 0. if c is a boundary. R) is the kernel of the surjective map Z0 (X. Indeed. This yields a morphism Sk (f ) : Sk (X. so that H0 (X. g : X → Y are homotopic map. then c = ∂( j bj γj ) for some paths γj . choose a base point x0 ∈ X. To prove Proposition 2. On the other hand. This follows from the fact that a 0cycle c = j aj xj is a boundary if and only if j aj = 0.6. This means that B0 (X. Let f : X → Y be a continuous map of topological spaces. This implies that f induces a morphism Hk (X. R) = Z0 (X. R) R. Then one can write c= j aj xj = j aj xj − ( j aj )x0 = j aj (xj − x0 ) = ∂ j aj γj if γj is a path joining x0 to xj . R) → Hk (Y. if g : Y → W is another continous map. Homotopic invariance. The ﬁrst step consists in deﬁnining a singular (k + 1)chain πk+1 in the topological space ∆k × I by subdiving the polyhedron ∆k × I ⊂ Rk+1 (a “prysm” . R) → Sk (Y. that we denote f . Proof. if j aj = 0. Corollary 2. then f ◦σ is a singular ksimplex in Y . If two topological spaces are homotopically equivalent.4. Proposition 2. Step 1. for every k ≥ 0 and any topological space X.5 we build. the induced maps in homology coincide. Proposition 2. R). a morphism (called the prism operator ) P : Sk (X) → Sk+1 (X × I) (here I denotes again the unit closed interval in R).
.1) P (σ) = Sk+1 (σ × id)(πk+1 ). . . B0 . HOMOLOGY THEORY B0 B1 A0 A1 Figure 1. Sk+1 (f × id) ◦ P (σ) = Sk+1 (f × id) ◦ Sk+1 (σ × id)(πk+1 ) = Sk+1 (f ◦ σ × id)(πk+1 ) = P (Sk (f )) . . The relevant property of the prism operator is proved in the next Lemma. . . B1 > . Ai . for k = 1 we have π2 =< A0 . Step 2. . then σ×id is a continous map ∆k × I → X × I. . The prism π2 over ∆1 over the standard symplesx ∆k ) into a number of singular (k + 1)simplexes. . Bk > . . The deﬁnition of the prism operator implies its functoriality: Proposition 2. Bi = (Pi . 1). Ak . Bi . the diagram Sk (X) Sk (f ) P / Sk+1 (X × I) Sk+1 (f ×id) Sk (Y ) commutes. given by Ai = (Pi . Therefore it makes sense to deﬁne the singular (k + 1)chain P (σ) in X as (2. It is just a matter of computation. P / Sk+1 (Y × I) Proof. Bk . 0). . . . If σ is a singular ksimplex in a topological space X. We deﬁne k πk+1 = i=0 (−1)i < A0 . . . B0 . For instance.7. If f : X → Y is a continuous map.20 2. The polyhedron ∆k × I ⊂ Rk+1 has 2(k + 1) vertices A0 . and summing them with suitable signs. A1 . . . B1 > − < A0 .
Bi . Bj . . . . . Bi . Bk > ˆ (−1)i < A0 . λ1 (x) = (x.2) ∂ ◦ P + P ◦ ∂ = Sk (λ1 ) − Sk (λ0 ) as maps Sk (X) → Sk (X × I). . . Since ˆ P (< P0 . . Ak > k + j<i=1 k ˆ (−1)i+j < A0 . . . Bk > − < A0 . . Aj .2) (note that exchanging the indices i. Bj . . . . Aj . . . Bi . Let λ0 .1. . Pj . . 1). Bk > . . Notice that P (δk ) = πk+1 . . k ∂P (δk ) = i=0 k (−1)i ∂ < A0 . Ak > . . . . . . .8. . Pk >) = i<j ˆ (−1)i < A0 . . . . Bi . . . Ai . . . . All terms with i = j cancel with the exception of < B0 . . . . Then (2. We compute now the action of the left side of (2. . j changes the sign). Bi . Bk > ˆ (−1)i+j < A0 . . 0). . . Bk > − < A0 . . . . . .2) to δk . . . . . . Ai . . . . . . . So we have ∂P (δk ) = < B0 . The right side yields < B0 . . i<j=1 − On the other hand. . . . . . . . Ai . . Bk > − < A0 . . . Ai . Aj . . Pj . Bj . . . Let δk : ∆k → ∆k be the identity map regarded as singular ksimplex in ∆k . Bk > . . . . Bk > j≤i=0 k = + i≤j=0 ˆ (−1)i+j+1 < A0 . . . . . . . . . Proof. . . . . . . Bi . .2) for X = ∆k . . . . . Ak >. . . . . Pk > . Ai . one has k ∂δk = j=0 ˆ (−1)j < P0 . . λi : X → X × I be the maps λ0 (x) = (x. Bk > ˆ (−1)i+j < A0 . . . . . . SINGULAR HOMOLOGY 21 Lemma 2. . . . . . . applying both sides of (2. .2) on δk . Ai . . We ﬁrst check the identity (2. Bi . Bk > i>j − we obtain the equation (2. . Ai .
2) holds when both sides are applied to δk . x0 ) → H1 (X. If we ﬁx a point x0 ∈ X. 1. The following result tells us that χ descends to a group homomorphism χ : π1 (X. given two loops at x0 . t). Moreover. γ5 (t) = Γ(t. Then. s) = x0 for all s ∈ I. Z). R) R.2) states that P is a hotomopy (in the sense of homological algebra) between the maps λ0 and λ1 .9.. we have a settheoretic map χ : L(x0 ) → S1 (X. Γ(t. A loop γ in X may be regarded as a closed singular 1simplex.e. Proof. Corollary 2. t). If two loops γ1 . Q > −Γ◦ < P0 . Consider the points P0 . Proof of Proposition 2. 1) = γ2 (t).4. P1 . a map Γ : I × I → X such that Γ(t. Equation (2. If X is a contractible space then H0 (X. then it holds in general. R) = 0 for k > 0. Q > . 1) in R2 .10. so that f = (F ◦ λ0 ) = F ◦ (λ0 ) = F ◦ (λ1 ) = (F ◦ λ1 ) = g . Relation between the ﬁrst fundamental group and homology. Deﬁne the loops γ3 (t) = Γ(1. γ2 . s) = Γ(1. then they are homologous as singular 1simplexes. Choose a homotopy with ﬁxed endpoints between γ1 and γ2 . g = F ◦ λ1 . γ1 . Let F be a hotomopy between the maps f and g. Q = (1. Proposition 2. and deﬁne the singular 2simplex σ = Γ◦ < P0 . γ2 are homotopic. then χ(γ2 ◦ γ1 ) = χ(γ1 ) + χ(γ2 ) in H1 (X. Z). P2 . P2 . so that one has (λ1 ) = (λ2 ) in homology. i. f = F ◦ λ0 .5. λ1 are the obvious maps ∆k → ∆k × I. 0) = γ1 (t). Hk (X. Both loops γ3 and γ4 are actually the constant loop at x0 . t). HOMOLOGY THEORY We must now prove that if equation (2. One has indeed ∂P (σ) = ∂Sk+1 (σ × id)(P (δk )) = Sk (σ × id)(∂P (δk )) P (∂σ) = P ∂(Sk (σ)(δk )) = P (Sk−1 (σ)(∂δk )) = Sk (σ × id)(P (∂δk )) so that ∂P (σ) + P (∂σ) = Sk+1 (σ × id)(∂P (δk )) + P (∂δk )) ¯ ¯ = Sk+1 (σ × id)(Sk (λ1 ) − Sk (λ0 )) = Sk (λ1 ) − Sk (λ0 ) ¯ ¯ where λ0 .22 2. Γ(0. Z). γ4 (t) = Γ(0. P1 .
and set σ(Q) = γ2 (t). If X is pathwise connected. Consider the point T = (0. consider the line joining P2 with Q. consider the line joining P0 with Q. Q > +Γ◦ < P0 . This proves that χ(γ1 ) and χ(γ2 ) are homologous. parametrize it with a parameter t such that t = 0 in P0 and t = 1 in the intersection of the line with Σ. Analogously. Figure 3). parametrize it with a parameter t such that t = 1 in P2 and t = 0 in the intersection of the line with Σ. This deﬁnes a singular 2simplex σ : ∆2 →X. Figure 2). It turns out that the ﬁrst homology group of a space with integer coeﬃcients is the abelianization of the fundamental group. and its kernel is the commutator subgroup of π1 (X. g. Z) is surjective. SINGULAR HOMOLOGY 23 P2 γ2 > Q γ4 ∧ γ5 ∧ γ3 P0 > γ1 Figure 2 P1 (cf. and one has ∂σ = σ◦ < P1 . Proposition 2. 1 ) in the standard 2simplex ∆2 and the segment Σ 2 joining T with P1 (cf. the morphism χ : π1 (X. Q > −Γ◦ < P0 . P1 > = γ2 − γ2 · γ1 + γ1 . . h ∈ G. The subgroup C(G) is obviously normal in G.11. We call it the abelianization of G. If Q ∈ ∆2 lies on or below Σ. Let G be any group. Q > +Γ◦ < P0 . To prove the second claim we need to deﬁne a singular 2simplex σ such that ∂σ = γ1 + γ2 − γ2 · γ1 . We recall from basic group theory the notion of commutator subgroup.1. and let C(G) be the subgroup generated by elements of the form ghg −1 h−1 . Q > −Γ◦ < P0 . the quotient group G/C(G) is abelian. P2 > = γ3 − γ5 + γ1 − γ2 + γ5 + γ4 = γ1 − γ2 . P1 > − Γ◦ < P2 . P2 > +σ◦ < P0 . We then have ∂σ = Γ◦ < P1 . and set σ(Q) = γ1 (t). x0 ). P2 > −σ◦ < P0 . if Q lies above or on Σ. x0 ) → H1 (X.
24 2. To prove the opposite inclusion. HOMOLOGY THEORY P2 e d ed γ2 ∧ e d e d •Q e d γ T r e d 2 rre d e rr d γ1 ∧ r ¨r d ¨ rrd • ¨¨ Σ rd Q ¨ r ¨ r d > P0 γ1 P1 Figure 3 Proof.g. −1 β2j = α1j · γ2j · α0j . in such a way that they depend on the endpoints and not on the indexing (e. let γ be a loop that in homology is a 1boundary. 1.3) σj = γ0j − γ1j + γ2j for some paths γkj . x0 ) coincides with ker χ. j Now if we set σj = αj + σj − βj we have c = ¯ then. So we may write (2. if σj (0) = σk (0). from x0 to γ1j (0) = γ2j (0) = P0 from x0 to γ2j (1) = γ0j (0) = P1 from x0 to γ1j (1) = γ0j (1) = P2 . So we have 0 = ∂c = j ai (σj (1) − σj (0)). To prove the second claim we need to show that the commutator subgroup of π1 (X. Figure 4) α0j α1j α2j and consider the loops −1 −1 β0j = α0j · γ1j · α2j . Then we have aj (βj − αj ) = 0.. Choose paths (cf. 2. ¯ χ( Πj γj j ) = [c] so that χ is surjective. We ﬁrst notice that since H1 (X. Let γj be the loop β −1 · σj · α. Let c = j aj σj be a 1cycle.e. the sum of the coeﬃcients corresponding to the same point must vanish. Choose a base point x0 ∈ X and for every j choose a path αj from x0 to σj (0) and a path βj from x0 to σj (1). −1 β1j = α2j · γ0j · α1j . Z) is abelian. k = 0. γ = ∂ j aj σj . In this linear combination of points with coeﬃcients in Z some of the points may coincide. choose αj = αk ). i. the commutator subgroup is necessarily contained in ker χ. a j aj σj .
so that the image of the class of γ in π1 (X. x0 )/C(π1 (X. the circle S 1 . Corollary 2.1. a punctured torus. the torus T 2 . x0 )/C(π1 (X. 2. For every k ≥ 0 there is a natural inclusion (injective morphism of Rmodules) Sk (A) ⊂ Sk (X). let A be any subspace (that we consider with the relative topology). Given a topological space X. 2. Exercise 2. S• (X) deﬁne a morphism δ : Sk (X)/Sk (A) → Sk−1 (X)/Sk−1 (A) which squares to zero.12. 6. RELATIVE HOMOLOGY 25 P2 γ0j P1 γ1j α2j α0j γ2j P0 α1j Figure 4 Note that the loops −1 −1 βj = β0j · β1j · β2j = α0j · γ1j · γ0j · γ2j · α0j x0 are homotopic to the constant loop at x0 (since the image of a singular 2simplex is contractible). The relative homology complex. If we deﬁne Zk (X. A) = ker ∂ : Sk (X) Sk−1 (X) → Sk (A) Sk−1 (A) . x0 )) is the identity. So whenever in the examples in Chapter 1 the fundamental groups we computed turned out to be abelian.2. Z) when X is: 1. 5. As a consequence one has the equality in π1 (X. This implies that the image of Πj [βj ]aj in π1 (X. Relative homology 2. We ﬁx a coeﬃcient ring R which for the sake of conciseness shall be dropped from the notation. In particular. up to reordering of terms. 4. 3. x0 )) is the identity as well. H1 (X. we were also computing the group H1 (X. a Riemann surface of genus g. Z). the corolla with n petals. the homology operators of the complexes S• (A).13. x0 ) Πj [βj ]aj = e. This means that γ lies in the commutator subgroup. Z) = 0 if X is simply connected. with Πj βj j . Rn minus a point.3) we see that γ coincides. Compute H1 (X. On the a other hand from (2.
If qk (c) ∈ Zk (X. A) k / Sk (X)/Sk (A) ∂ Sk (A) ∂ qk /0 0 Let / Bk−1 (A) / Bk−1 (X) qk−1 /B k−1 (X. A) under the quotient homomorphism qk . A). A) then qk (c) = ∂ ◦ qk+1 (b) for some b ∈ Sk+1 (X). R). A) if and only if qk (c) ∈ Bk (X. Proof. A) is formed by the chains whose boundary is in A. and Bk (A) by the chains that are boundaries up to chains in A. Sk+1 (X) Sk (X) → Sk+1 (A) Sk (A) Definition 2. If c ∈ Zk (X. Lemma 2.2.26 2. A). A) = {c ∈ Sk (X)  c = ∂b + c with b ∈ Sk+1 (X). A. c ∈ Sk (A)} . HOMOLOGY THEORY Bk (X. A). Proposition 2. A) if and only if qk (c) ∈ Zk (X. Thus. When we want to emphasize the choice of the ring R we write Sk (X. Zk (X.3. A). If c = ∂b + c with b ∈ Sk+1 (X) and c ∈ Sk (A) then qk (c) = qk ◦ ∂b = ∂ ◦ qk+1 (b) ∈ Bk (X. A). Proof. A). A) Zk (X. Zk (X. . if qk (c) ∈ Bk (X. c ∈ Sk (X) is in Bk (X. A) ⊂ Zk (X. A) = Im ∂ : we have Bk (X. A) is the preimage of Zk (X. which makes clearer its geometrical meaning. A). Conversely. For all k ≥ 0. A)/Bk (X. A) = {c ∈ Sk (X)  ∂c ∈ Sk−1 (A)} Bk (X. A) /0 Zk (X. A) = Zk (X. It will be useful to consider the following diagram 0 qk Zk (X) / Sk (X) ∂ / Z (X. that is. The homology groups of X relative to A are the Rmodules Hk (X. an element c ∈ Sk (X) is in Zk (X.1. then c − ∂b ∈ ker qk−1 so that c = ∂b + c with c ∈ Sk (A).4. Lemma 2. A) then qk−1 ◦ ∂(c) = 0 so that qk (c) ∈ Zk (X. A) then 0 = ∂ ◦ qk (c) = qk−1 ◦ ∂(c) so that c ∈ Zk (X. A). The relative homology is more conveniently deﬁned in a slightly diﬀerent way. Hk (X. A)/Bk (X.
A). If A = {x0 } is a point. A). A).2. A) k / Z (X. Hk (X. then Hk (X. If c = ∂( j aj xj ) = c − ( aj xj ∈ S0 (X) and γj is a path from x0 ∈ A to xj . j Corollary 2. What we should do is to prove the commutativity and the exactness of the rows of the diagram 0 / Sk (A) ∼ / Bk (X. Proposition 2. For the exactness of the ﬁrst row.2. • If X = ∪j Xj is a union of pathwise connected components. • The inclusion map i : A → X induces a morphism H• (A) → H• (X) and the composition H• (A) → H• (X) → H• (X. If c ∈ Sk (A) then qk (c) = 0. it is obvious that Sk (A) ⊂ Bk (X. On the other hand if c ∈ Bk (X. Given topological spaces X.7. so that qk (c) = 0 implies 0 = qk ◦ ∂b = ∂ ◦ qk+1 (b). A) is a free Rmodule generated by the components of X that do not meet A.2. A) and that qk (c) = 0 if c ∈ Sk (A). A) = 0. A). A) and qk (c) = 0 then c ∈ Sk (A) by the deﬁnition of Zk (X. A) k /0 0 / Sk (A) qk /0 Commutativity is obvious. To prove the surjectivity of qk . If a proof is not given the reader should provide one by her/himself. Indeed Hj (Xj . Aj ) = 0 if Aj is empty. 2. A) Hk (X) for k > 0. Such a map induces in natural way a morphisms of Rmodules f : H• (X. A) / Zk (X. a continous map of pairs is a continuous map f : X → Y such that f (A) ⊂ B. If we consider the inclusion of pairs (X. A) we have c = ∂b + c with b ∈ Sk+1 (X) and c ∈ Sk (A). then H0 (X. • The relative cohomology groups are functorial in the following sense. just notice that by deﬁnition an element in Bk (X. H0 (X. B). Y with subsets A ⊂ X. A) may be represented as ∂b with b ∈ Sk+1 (X). Proposition 2. . ∅) → (X. Aj ) where Aj = A ∩ Xj . RELATIVE HOMOLOGY 27 Proof. Moreover qk is surjective by Lemma 2. A) vanishes (since Zk (A) ⊂ Bk (X. If c ∈ Zk (X. A) Hk (X). A) → H• (Y. then j aj )x0 so that c ∈ B0 (X. A) ⊕j Hk (Xj . A). If X is pathwise connected and A is nonempty. • If A is empty. A) qk / B (X. we have Sk (A) ⊂ Zk (X.6. Proof.5. As for the second row. Hk (X. A) we obtain a morphism H• (X) →• H(X. We list here the main properties of the cohomology groups Hk (X. Main properties of relative homology. which in turn implies c ∈ Sk (A). A)). A) from the deﬁnition of Zk (X. B ⊂ Y .
it is not in general possible to write it as a diﬀerence of standard ksimplexes in U . i p . The trick to circumvent this diﬃculty consists in replacing S• (X) with a diﬀerent complex that however has the same homology. A) → 0 Exercise 2. However. one obtains a long exact cohomology sequence · · · → H2 (A) → H2 (X) → H2 (X. By deﬁnition the relative homology of X with respect to A is the homology of the quotient complex S• (X)/S• (A). but in the case of homology there is a subtlety.3. we understand the choice of a coeﬃcient ring R. A → V . By Proposition 1. R) R and Hk (S 1 . A) → H0 (A) → H0 (X) → H0 (X.8. One would think that there is an exact sequence 0 → Sk (A) → Sk (U ) ⊕ Sk (V ) → Sk (X) → 0 where i is the morphism induced by the inclusions A → U . R) = 0 for k > 1. This is quite similar to what happens in de Rham cohomology. Use the long relative homology sequence to compute the relative homology groups H• (R2 . it is not possible to prove that p is surjective (if σ is a singular ksimplex whose image is not contained in U or V . The long exact sequence of relative homology. and p is given by p(σ1 . 1Again. Zk (X. Assume to know that H1 (S 1 . R). A) → H1 (A) → H1 (X) → H1 (X. A) = {c ∈ Sk (X)  c = ∂b + c with b ∈ Sk+1 (X).1. that we are going to consider here) allows one to compute the homology of a union X = U ∪ V from the knowledge of the homology of U . HOMOLOGY THEORY Proof. V ). σ2 ) = σ1 − σ2 (again using the inclusions U → X. c ∈ Sk (A)} = Bk (X) when k > 0. Let us denote A = U ∩ V . 3. 2. The MayerVietoris sequence The MayerVietoris sequence (in its simplest form. Definition 2. A) = {c ∈ Sk (X)  ∂c ∈ Sk−1 (A)} = Zk (X) when k > 0 Bk (X.28 2. V and U ∩ V . Moreover we deﬁne S• (X) as the subcomplex of S• (X) formed by Usmall chains. Let U = {Uα } be an open cover of X.1 U U The homology diﬀerential ∂ restricts to S• (X). V → X). A singular kchain σ = j aj σj is Usmall if every singular kU simplex σj maps into an open set Uα ∈ U for some α. so that one has a homology H• (X). S1 . adapted to homology by reversing the arrows.6.
. called the join of B with < Q0 . j=0 Example 2. T (δk ) = Bk (δk − Σ(δk ) − T (∂δk )) T (δ0 ) = 0. j aj )B Next we deﬁne operators Σ : Sk (X) → Sk (X) and T : Sk (X) → Sk+1 (X). Lemma 2. The operators Σ and T . . . .2. . the action of Σ splits ∆2 into smaller simplexes as shown in Figure 6. The join B(< E0 . THE MAYERVIETORIS SEQUENCE 29 E0 £ £ £ £ £ £ £ £ £ r £ B rr rr £ r£ E1 Figure 5.4. . The operator Σ is called the subdivision operator and its eﬀect is that of subdividing a singular simplex into a linear combination of “smaller” simplexes. where the point Bk is the barycenter of the standard ksimplex ∆k . . For k = 1 one gets Σ(δ1 ) =< B1 P1 > − < B1 P0 >. . and then extended by functoriality. will be deﬁned for X = ∆k (the space consisting of the standard ksimplex) and for the “identity” singular simplex δk : ∆k → ∆k . E1 >) U Proposition 2. .3.3. Qk >. H• (X) H• (X). ∂ ◦ B + B ◦ ∂ = Id on Sk (Rn ) if k > 0. for k = 2. while ∂ ◦ B(σ) = σ − ( if σ = j aj xj ∈ S0 (Rn ). . Qk >. . One deﬁnes Σ(δ0 ) = δ0 . This will be done in several steps. . Qk > in Rn and a point B ∈ Rn we consider the singular simplex < B. and then extends recursively to positive k: Σ(δk ) = Bk (Σ(∂δk )). Q0 . This should be done for all k. Given a singular ksimplex < Q0 . 1 Bk = k+1 k Pj . The following Lemma is easily proved. analogously to what we did for the prism operator. . This operator B is then extended to singular chains in Rn by linearity. U To prove this isomorphism we shall build a homotopy between the complexes S• (X) and S• (X).
∂ ◦ T + T ◦ ∂ = Id −Σ. M0 . M0 . with E0 . T (σ) = Sk+1 (σ)(T (δk )). There is a natural number r > 0 such that every singular simplex in Σr (σ) is contained in a open set Uα . M2 . P1 > − < B2 . so that the morphism Σ induced in homology by Σ is an isomorphism. P0 > + < B2 . .30 2. and the second that T is a homotopy between Σ and Id. The proof of the following Lemma is an elementary computation. As ∆k is compact there is a real positive number such that σ maps a neighbourhood of radius of every point of ∆k into some Uα . P2 > − < B2 . M2 . . and σ a singular ksimplex in X.5. HOMOLOGY THEORY P2 e d ed e d e d e d M M1 r e d 0 rre d rr e r d e r d B2 rrd e rd e r e r d P1 M2 P0 Figure 6. The subdivision operator Σ splits ∆2 into the chain < B2 . One has the identities ∂ ◦ Σ = Σ ◦ ∂. Proof. Lemma 2. The ﬁrst identity tells us that Σ is a morphism of diﬀerential complexes. M1 . Proposition 2. Let σ =< E0 . . . . Ek ∈ Rn . P2 > + < B2 . U = {Uα } an open cover. Since kr =0 r→+∞ (k + 1)r lim . P1 > − < B2 .7. Proof. . Let X be a topological space. The diameter of every simplex in the singular chain Σ(σ) ∈ Sk (Rn ) is at most k/k + 1 times the diameter of σ. . The diameter of a singular ksimplex σ in Rn is the maximum of the lengths of the segments contained in σ. M1 . Ek >. P0 > The deﬁnition of Σ and T for every topological space and every singular ksimplex σ in X is Σ(σ) = Sk (σ)(Σ(δk )). These identities are proved by direct computation (it is enough to consider the case X = ∆k ). Lemma 2. .6.
But as Σr (σ) = Sk (σ)(Σr (δk )) we are done. H2 (X. R) = R for k = 0 and k = n 0 for 0 < k < n and k > n . n ≥ 2. For every k there is an exact sequence of Rmodules U 0 → Sk (A) → Sk (U ) ⊕ Sk (V ) → Sk (X) → 0 .2 we may replace H• (X) with the homology H• (X).11.9. so that one obtains a long homology exact sequence involving the homologies H• (A).8. V } and A = U ∩ V . We may now prove the exactness of the MayerVietoris sequence in the following sense. Exercise 2.2. Exercise 2. Use the MayerVietoris sequence to compute the homology of a cylinder S 1 × R minus a point with coeﬃcients in Z. Compare this with the homology of S 2 minus three points. let U = {U. the exactness of the MayerVietoris sequence is easily proved. Z) = 0 .3. is Hk (S n . so that we obtain the exact sequence · · · → H2 (A) → H2 (U ) ⊕ H2 (V ) → H2 (X) → H1 (A) → H1 (U ) ⊕ H1 (V ) → H1 (X) → H0 (A) → H0 (U ) ⊕ H0 (V ) → H0 (X) → 0 Exercise 2. The morphisms i and p commute with the homology operator ∂. Z) Z ⊕ Z. If X = U ∪ V (union of two open subsets).10. Proposition 2. This completes the proof of Proposition 2. σ2 ) = jU ◦ σ1 + jV ◦ σ2 . Z) Z. (Hint: since the cylinder is homotopic to S 1 . it has the same homology). and H2 (S 2 . i p Proof. One has a diagram of inclusions U @ @@ jU ~? @@ ~~ ~ @@ @ ~~ ~ U A@ @ @@ @@ @ V >X ~~ ~~ ~~ ~~ jV V Deﬁning i(σ) = ( U ◦ σ. H• (V ) ⊕ H• (V ) and U U H• (X). The result is (calling X the space) H0 (X. THE MAYERVIETORIS SEQUENCE 31 there is an r > 0 such that Σr (δk ) is a linear combination of simplexes whose diameter is less than . Show that the relative homology of S 2 mod S 1 with coeﬃcients in Z is concentrated in degree 2. But in view of Proposition 2. . Prove that for any ring R the homology of the sphere S n with coeﬃcients in R. S 1 ) Z ⊕ Z. H1 (X. − V ◦ σ) and p(σ1 .
” To state the main theorem about excision we need some deﬁnitions from topology. This means that c = ∂b + c with b ∈ Sk+1 (X). c ∈ Bk (X. then H• (A. A−U ) → Hk (X. A − U ) be such that. int(A)} of X. we can regard c as a relative cycle in X −U mod A−U . Exercise 2. Theorem 2. Moreover. then U can be excised. A) → (Y. Excision If a space X is the union of subspaces. A) is unchanged. i. B) (i. Y ). so that ∂c ∈ Sk−1 (A). Therefore. we say that U “can be excised. ¯ where b1 maps into X − U and b2 into int(A).. let [c] ∈ Hk (X − U. B) → (X.4. after the removal. Proof. If we cancel from σ those singular simplexes σj taking values in int(A). regarding c as a cycle in X mod A. then r ◦ i = IdH• (A) .3. .e. We consider the cover U = {X − U . if A is a deformation retract of X. Definition 2. and split Σr (b) into b1 + b2 . Let i : A → X be an inclusion of topological spaces. A map r : X → A is a retraction of i if r ◦ i = IdA . then H• (A) H• (X). this implies that the morphism Hk (X −U. it is a boundary. If (X − U.32 2. Show that no retraction S n → S n−1 can exist.1. The operation of excision in some sense gives us information about the reverse operation. (A. it tells us what happen to the homology of a space if we “excise” a subpace out of it.e. B). A).2. B) H• (X. A). A) is surjective.. i. A− U ) → (X. We apply the operator Σr to both sides of this inequality. Given nested subspaces U ⊂ A ⊂ X. a map f : X → Y such that f (A) ⊂ B) there is natural morphism f : H• (X. A) → H• (Y. A) is an excision. c ∈ Sk (A) . HOMOLOGY THEORY 4. 2. Let c = j aj σj ∈ Zk (X. the inclusion map (X −U. To prove that it is injective. The same notion can be given for inclusions of pairs.. In view of Proposition 2. A − U ) → (X. A subspace A ⊂ X is a deformation retract of X if IdX is homotopically equivalent to i ◦ r.2 we may assume that c is Usmall. Let us recall that given a map f : (X. if such a map is a deformation retract. A) is said to be an excision if the induced morphism Hk (X − U. the class [c] ∈ Hk (X. Definition 2. If the closure U of U lies in the interior int(A) of A.e. the MayerVietoris suquence allows one to compute the homology of X from the homology of the subspaces and of their intersections. where r : X → A is a retraction. so that i : H• (A) → H• (X) is injective. A) is an isomorphism for all k. If r : X → A is a retraction of i : A → X. We have Σr (c) − ∂b1 = Σr (c ) + ∂b2 . 1. Y ). A − U ) → Hk (X.
A−U ). Deduce that Hk (S n ) Hk−1 (S n−1 ). Given a topological space X and a ring R. Let B an open band around the equator of S 2 . Prove that there is a long exact sequence for the augmented relative homology modules.8. . Exercise 2.4. . S n−1 ). Hn (B n . xn ) = (−x0 . 3. To describe some more applications of excision we need the notion of augmented homology modules. A. . Use the long exact sequence for the augmented relative homology modules to − prove that Hk (S n ) Hk (S n . let us deﬁne ∂ : S0 (X. Let B n be the closed unit ball in Rn+1 . Use excision to show that Hk (S n . A. R) if A = ∅. Hk (X. . Z). . Compute the relative homology H• (S 2 − x0 . .5. Now we have Σr (c) = Σr (c ) + ∂b2 + ∂b1 with Σr (c )+∂b2 ∈ Sk (A−U ) and ∂b1 ∈ Sk+1 (X −U ) so that Σr (c) ∈ Bk (X −U. xn ). S n−1 ) = 0 for k < n. R) . En ) and Hk−1 (S n−1 ) Hk (B n . B − x0 . therefore. A − U )). S n its boundary. x1 .7. 1. x1 . R) = ker ∂ /B0 (X. S n−1 ). and let En be the two closed (northern.e. We deﬁne the augmented homology modules H0 (X. . A. which implies [c] = 0 (in Hk (X − U. . R) for k > 0 . Hk (X. and x0 ∈ B. A. . R) → R aj σj j → j aj . R) if A = ∅ . both chains are in (X − U ) ∩ A = A − U . southern) emispheres in S n . S n−1 ) R − 2. one deﬁnes the augmented relative homology modules Hk (X. EXCISION 33 The chain in the left side is in X − U while the chain in the right side is in A. Hk (X.6. Exercise 2. R) = Hk (X. So we have Hk (B n . ± Exercise 2. R) = Hk (X. If A ⊂ X. and let r : S n → S n → S n be the reﬂection r(x0 .. Let S n be the sphere realized as the unit sphere in Rn+1 . En ) Hk (B n . R) in a similar way. Exercise 2. i. R) = Hk (X.
i. with reference to the − general theory. Deﬁne w(x) = v(x) . 1 ) = w(x). X = S n .. .34 2. Show that a : Hn (S n ) → Hn (S n ) is the multiplication by (−1)n+1 . En ). (Here we are excising the open southern emisphere.10. Proof. x2 . Let us deﬁne a vector ﬁeld on S n as a continous map v : S n → Rn+1 such that v(x) · x = 0 for all x ∈ S n (the product is the standard scalar product in Rn+1 ). Proposition 2. We end with a standard application of algebraic topology. . 0) = x. The rotation group O(n + 1) acts on S n . Exercise 2. Deﬁne F : Sn × I → Sn F (x.7. (Hint: this is trivial for n = 0. En ) is an excision. But (En . with w(x) · x = 0 for all x ∈ S n . assume that such a vector ﬁeld exists. 2. A = En .11. However it is enough to consider the subspace 1 V = x ∈ S n  x0 > − 2 . x0 . . . 2 F (x. Conversely. Let a : S n → S n be the antipodal map. x2m ) . . 1. x2m+1 ) = (−x1 . En − V ) so that we are done. 1) = −x = x cos tπ + w(x) sin tπ. a(x) = −x. HOMOLOGY THEORY Prove that r : Hn (S n ) → Hn (S n ) is the multiplication by −1. S n−1 ) is a deformation retract of (S n − − V.) The hypotheses of Theorem 2. −x3 . Show that for any M ∈ O(n + 1) the induced morphism M : Hn (S n ) → Hn (S n ) is the multiplication by det M = ±1. v(x) this is a map S n → S n .9. A nowhere vanishing vector ﬁeld v on S n exists if and only if n is odd.e.4 are not satisﬁed. S n−1 ) → (S n . − + V can be excised from (S n . We show that the inclusion map (En . −x2m+1 . If n = 2m + 1 a nowhere vanishing vector ﬁeld is given by v(x0 . F (x. . t) Since F (x. + − Example 2. . . . U = the open southern emisphere. and can be extended by induction using the commutativity of the diagram Hn (S n ) r ∼ n−1 ) /H n−1 (S r ∼ n−1 ) /H n−1 (S Hn (S n ) which follows from Exercise 2.
n must be odd.4. But as a consequence of Exercise 2. w.9. . a are homotopic. EXCISION 35 the three maps Id.
.
2. then ϕU. M(U ) is an R(U )module and for each pair V ⊂ U the restriction map ϕU. Definition 3. we shall write sV instead of ϕU.V : P(U ) → P(V ) to each pair V ⊂ U of open subsets. OX is deﬁned as the category whose objects are the open subsets of X while the morphisms are the inclusions of open sets. (2) ϕU.1. If R is a presheaf of rings on X. A morphism f : P → Q of presheaves over X is a family of morphisms of Abelian groups fU : P(U ) → Q(U ) for each open U ⊂ X. for each open subset U .W = ϕV. A presheaf of Abelian groups on X is a rule1 P which assigns an Abelian group P(U ) to each open subset U of X and a morphism (called restriction map) ϕU. by requiring that the restriction maps are ring morphisms.V : M(U ) → M(V ) is a morphism of R(U )modules (where M(V ) is regarded as an R(U )module via the restriction morphism R(U ) → R(V )).CHAPTER 3 Introduction to sheaves and their cohomology 1. Presheaves and sheaves Let X be a topological space.V (s). (3) if W ⊂ V ⊂ U are open sets. If s ∈ P(U ) is a section of P on U and V ⊂ U . 37 . The deﬁnitions in this Section are stated for the case of presheaves of Abelian groups.U is the identity map. Presheaves of rings are deﬁned in the same way. but analogous deﬁnitions and properties hold for presheaves of rings and modules.W ◦ ϕU. so as to verify the following requirements: (1) P(∅) = {0}. The restriction PU of P to an open subset U is deﬁned in the obvious way.V . The elements s ∈ P(U ) are called sections of the presheaf P on U . a presheaf M of Abelian groups on X is called a presheaf of modules over R (or an Rmodule) if. commuting with the 1This rather naive terminology can be made more precise by saying that a presheaf on X is a contravariant functor from the category OX of open subsets of X to the category of Abelian groups. Definition 3.
V being open neighbourhoods of x. sW = s W . On the set G = i∈I Gi . we put the following equivalence relation: g ∼ h. sUi = si . SHEAVES AND THEIR COHOMOLOGY restriction morphisms. sx = sx .e. with fii = id and fij ◦ fjk = fik . ¯ ¯ S2) Given sections si ∈ F(Ui ) which coincide on the intersections. Heuristically.V ϕU.V P(V ) − − → Q(V ) −V − Definition 3. si Ui ∩Uj = sj Ui ∩Uj for every i. Definition 3. The following discussion should make this notion clearer. A directed set I is a partially ordered set such that for each pair of elements i. two elements − → in G represent the same element in the direct limit if they are ‘eventually equal.. is the quotient G/ ∼.38 3. If I is a directed set. deﬁne the same germ at x. . s = s.’ From this deﬁnition one naturally obtains the existence of canonical morphisms Gi → l. j ∈ I there is a third element k such that i < k and j < k. the reader may consult [12]. there exists a section s ∈ F(U ) whose restriction to each Ui equals si . The stalk of a presheaf P at a point x ∈ X is the Abelian group Px = lim P(U ) − → U f f where U ranges over all open neighbourhoods of x. a directed system of Abelian groups is a family {Gi }i∈I of Abelian groups. From the very deﬁnition of direct limit we see that two elements s ∈ P(U ). directed by inclusion. j. such that for all i < j there is a group morphism fij : Gi → Gj . If x ∈ U and s ∈ P(U ). S1) If two sections s ∈ F(U ). s ∈ F(U ) coincide when restricted to any Ui . denoted l = limi∈I Gi . sheaves are presheaves deﬁned by local conditions.e. i.3.4. i. Remark 3.e. for more detail.5. The stalk of a sheaf is deﬁned as in the case of a presheaf. s ∈ P(V ). A sheaf on a topological space X is a presheaf F on X which fulﬁlls the following axioms for any open subset U of X and any cover {Ui } of U . sUi = ¯ sUi . i. if there exists a k ∈ I such that fik (g) = fjk (h). the following diagram commutes: P(U ) − − → Q(U ) −U − ϕU. if and only if there exists an open neighbourhood W ⊂ U ∩ V of x such that s and s coincide on W . U . they are equal. We recall here the notion of direct limit. The direct limit l of the system {Gi }i∈I . roughly speaking. the image sx of s in Px via the canonical projection P(U ) → Px (see footnote) is called the germ of s at x. Thus. where denotes disjoint union. with g ∈ Gi and h ∈ Gj .
any g ∈ Fx is of the form g = sx for some open U x and some s ∈ F(U ). we obtain a presheaf. in the following way: since the stalk Fx is the direct limit of the groups F(U ) over all open U containing x. If f : F → G is a morphism of sheaves on X. If F is a sheaf. Show that 0 → F → F → F is an exact sequence of presheaves. and ˜ ˜ U = V1 ∪ V2 . The presheaf GX is not a sheaf: if V1 and V2 are disjoint open subsets of X.11. there is for each x ∈ U an open neighbourhood Ux such that sUx = 0. All stalks (GX )x of GX are isomorphic to the group ˜ G.7. and let d : Ω• → Ω• be the X X p exterior diﬀerential. If 0 → F → F → F → 0 is an exact sequence of sheaves. Example 3. In the same way one can deﬁne the following sheaves: ∞ The sheaf CX of diﬀerentiable functions on a diﬀerentiable manifold X. We can deﬁne the presheaves ZX of closed diﬀerential pforms. ˜ but there is no section g ∈ GX (U ) = G which restricts to g1 on V1 and to g2 on V2 . Let CX (U ) be the ring of realvalued continuous functions on an open set U of X. Let G be an Abelian group. but the last arrow may fail to be surjective. A morphism of sheaves is just a morphism of presheaves. The sheaves Ωp of diﬀerential pforms.9. Then CX is a sheaf (with the obvious restriction morphisms). and . satisfy the hypothesis of the second sheaf axiom S2) (since V1 ∩ V2 = ∅ there is nothing to satisfy). PRESHEAVES AND SHEAVES 39 Example 3. This is not true for a presheaf. The ﬁrst sheaf axiom then implies s = 0. if s ∈ F(U ). and all the sheaves of tensor ﬁelds on a X diﬀerentiable manifold X. since sx = 0 for all x ∈ U . Let X be a diﬀerentiable manifold.1. A sequence of morphisms of sheaves 0 → F → F → F → 0 is exact if for every point x ∈ X. Example 3.6. The sheaves of forms of type (p. the sections g1 ∈ GX (V1 ) = G. The sheaf of holomorphic functions on a complex manifold and the sheaves of holomorphic pforms on it. q) on a complex manifold X. F(U ) = {0} for all open sets U ⊂ X.10. Example 3.14 below. Example 3. the sequence of morphisms between the stalks 0 → Fx → Fx → Fx → 0 is exact.11 below. The stalk Cx ≡ (CX )x at x is the ring of germs of continuous functions at x. Indeed. fx : Fx → Gx .8. Let 0 → F → F → F → 0 be an exact sequence of sheaves. then F is the zero sheaf. Example 3. the sheaf of continuous functions on X. cf. for every open subset U ⊂ X the sequence of groups 0 → F (U ) → F(U ) → F (U ) is exact. g2 ∈ GX (V2 ) = G. ˜ ˜ ˜ called the constant presheaf GX . Exercise 3. Deﬁning P(U ) ≡ G for every open subset U and taking the identity maps as restriction morphisms. for every x ∈ X the morphism f induces a morphism between the stalks. An instance of this situation is contained in Example 3. with g1 = g2 . and Fx = {0} for all x ∈ X. then set fx (g) = (fU (s))x .
That is. X p BX (U ) = {ω ∈ Ωp (U )  ω = dτ X for some τ ∈ Ωp−1 (U )}. the morphism CX (U ) −−→ ZX (U ) may fail to be surjective. 0)}. As a ﬁrst step we consider the case ˜ of a constant presheaf GX on a topological space X. Definition 3. where G is an Abelian group. there is an 1 open cover {Ui } of U by open subsets where ω is an exact form. The sections s ∈ P(U ) of the presheaf P on an open subset U produce sections s : U → P of π. The set P. given a presheaf. e d d ´ e 1. It is clear that (GX )x = Gx = G at each point x ∈ X and that GX is a sheaf. endowed with the topology whose base of open subsets consists of the sets s(U ) for U open in X and s ∈ P(U ).12. 2 where GX (U ) = G is included as the constant functions. In fact. while the sequence of sheaf morphisms 0 → R → CX −−→ ZX → 0 ∞ 1 is exact (Poincar´ lemma). ∞ 1 This means that. the presheaf BX of exact diﬀerential 1forms does not fulﬁll the second sheaf axiom: consider the form xdy − ydx ω= x2 + y 2 deﬁned on the open subset U = X − {(0. Now. SHEAVES AND THEIR COHOMOLOGY p BX of exact pdiﬀerential forms. Consider the disjoint union of the stalks P = x∈X Px and the natural projection π : P → X. 2A function is locally constant on U if it is constant on any connected component of U . P is the presheaf of all sections that locally coincide with sections of P. if X = R2 . deﬁned by s(x) = sx . since the condition of being closed is local: a diﬀerential form is closed if p and only if it is closed in a neighbourhood of each point of X.40 3. Etal´ space. ωUi ∈ BX (Ui ) (this is Poincar´’s lemma). X p ZX is a sheaf. . is called the ´tal´ space of the e e presheaf P. called the constant sheaf with stalk G. We wish now to describe how. one can naturally associate with it a sheaf having the same stalks. It can be described in another way by the following construction. and we can deﬁne a new presheaf P by taking P (U ) as the group of those sections σ : U → P of π such that for every point x ∈ U there is an open neighbourhood V ⊂ U of x which satisﬁes σV = s for some s ∈ P(V ). Notice that the functions f : U → G are the sections of the projection π : x∈X Gx → X and the locally constant functions correspond to those sections which locally coincide with the sections produced by the elements of G. let P be an arbitrary presheaf on X. But ω is not an exact form on U because its integral along the unit e circle is diﬀerent from 0. Since ω is closed on U .1. We can deﬁne another presheaf GX on X by putting GX (U ) = {locally constant functions ˜ f : U → G}. On the contrary. BX is 1 not a sheaf. p ZX (U ) = {ω ∈ Ωp (U )  dω = 0}.
. For any open set U we have an exact sequence of Abelian groups (actually of Rvector spaces) k k k 0 → BX (U ) → ZX (U ) → HX (U ) → 0 k where HX is the presheaf that with any open set U associates its kth de Rham cohomok k logy group. . injective. (6) Show that φ is an isomorphism if and only if P is a sheaf. 3Let I be a directed set. with Zmodules. the Hom modules and tensor products are taken over Z. (1) Show that π : P → X is a local homeomorphism. − → that is. On the other hand. Given a sheaf F on a topological space X and a subset (not necessarily open) S ⊂ X. A subset J of I is said to be coﬁnal if for any i ∈ I there is a j ∈ J such that i < j. the open neighbourhoods of any point x ∈ X which are diﬀeomorphic to Rn (where n = dim X) are coﬁnal3 in the family of all open k neighbourhoods of x. BX X X Definition 3.e. X k k In this case the natural morhism HX → (HX ) is of course surjective but not k → (B k ) = Z k is injective but not surjective. then lim Gi − → i∈I j∈J lim Gj . for every open subset U ⊂ X. (2) Show that for every open set U ⊂ X and every s ∈ P(U ).e. which is tantamount to (B k ) k Example 3.16.6 this means that (HX ZX . P is called the sheaf associated with the presheaf P. every point u ∈ P has an open neighbourhood U such that π : U → π(U ) is a homeomorphism. (5) Show that there is a presheaf morphism φ : P → P . the morphism φ : P → P is neither injective nor surjective: for instance. the section s : U → P is continuous. i. Q be presheaves on a topological space X. The group of such sections is denoted by Γ(S. direct limits can be taken over coﬁnal subsets of the index set. the sections of the sheaf F on S are the continuous sections σ : S → F of π : F → X. (4) Prove that for all x ∈ X the stalks of P and P at x are isomorphic. HX (U ) = HDR (U ). By the deﬁnition of direct limit we see that. 4Since we are dealing with Abelian groups. In general.1. the morphism between the ˜ constant presheaf GX and its associated sheaf GX is injective but nor surjective. if {Gj }j∈J is the subfamily indexed by J. PRESHEAVES AND SHEAVES 41 Exercise 3.13.15. by (P ⊕ Q)(U ) = P(U ) ⊕ Q(U ) with the obvious restriction morphisms. Let P. (3) Prove that P is the sheaf of continuous sections of π : P → X. Now. As a second example we study the sheaf associated with the presheaf of exact kforms on a diﬀerentiable manifold X. Definition 3. 4 (1) The direct sum of P and Q is the presheaf P ⊕ Q given.14. given a directed family of Abelian groups {Gi }i∈I . In accordance with e k ) = 0. F). i. k BX Example 3. so that (HX )x = 0 by the Poincar´ lemma.
let us denote by Hom(PU . the inverse image functor for sheaves of Abelian groups is exact). SHEAVES AND THEIR COHOMOLOGY (2) For any open set U ⊂ X. this is an Abelian group in a natural manner. Indeed one deﬁnes the tensor product of the sheaves F and G as the sheaf associated with the presheaf U → F(U ) ⊗ G(U ). G) are sheaves. The inverse image of P by f is the presheaf on X deﬁned by U→ lim − → −1 P(V ). It follows that if 0 → F → F → F → 0 is an exact sequence of sheaves on Y . The ´tal´ space f −1 F of the inverse image sheaf is the ﬁbred product 5 Y ×X F. Definition 3. the induced sequence 0 → f −1 F → f −1 F → f −1 F → 0 of sheaves on X. G(U )) and 1. Q)(U ) = Hom(PU . the tensor product of F and G previously deﬁned may not be a sheaf. Here we study the behaviour of presheaves and sheaves under change of base space. The presheaf of homomorphisms is the presheaf Hom(P. . (3) The tensor product of P and Q is the presheaf (P ⊗ Q)(U ) = P(U ) ⊗ Q(U ). The stalk of the inverse image presheaf at a point x ∈ X is isomorphic to Pf (x) . It e e follows easily that the inverse image of the constant sheaf GX on X with stalk G is the constant sheaf GY with stalk G. G)x Hom(Fx .g. Gx ). QU ) with the natural restriction morphisms. is also exact (that is.18. On the contrary. Direct and inverse images of presheaves and sheaves. It should be noticed that in general Hom(F. then f∗ F turns out to be a sheaf.42 3.2. The direct image by f of a presheaf P on X is the presheaf f∗ P on Y deﬁned by (f∗ P)(V ) = P(f −1 (V )) for every open subset V ⊂ Y . Q) given by Hom(P. then the presheaves F ⊕ G and Hom(F. Definition 3. QU ) the space of morphisms between the restricted presheaves PU and QU . Let P be a presheaf on Y . If F is a sheaf on X. Hom(F(U ). 5For a deﬁnition of ﬁbred product see e. G)(U ) Hom(F. U ⊂f (V ) The inverse image sheaf of a sheaf F on Y is the sheaf f −1 F associated with the inverse image presheaf of F. If F and G are sheaves. [15]. Let f : X → Y be a continuous map.17. f −1 GX = GY .
Thus we obtain a cohomology theory. Example 3.ip+1 . By the second sheaf axiom this implies that there is a global section α ∈ F(X) such ˜ 0 (U. which is evidently surjective that αUi = αi . It is an easy exercise to check that δ 2 = 0.. x1 . P) → C 1 (U.....2. ..1) (δα)ik = αkUi ∩Uk − αiUi ∩Uk . Cohomology of sheaves We wish now to describe a cohomology theory which associates cohomology groups to a sheaf on a topological space X.. We compute the Cech cohomology of U with coeﬃcients in the constant sheaf R.ip+1 Ui0 . F) by (3. This yields a morphism H ˜ and is injective because of the ﬁrst sheaf axiom. Since the indexes of the open sets are taken in strictly increasing order. 0 i0 <···<ip Thus a pcochain α is a collection {αi0 .1. R) = 0 for k > 1 because there are no triple intersections. We have C 0 (U. x2 − x0 ..ip+1 } = k=0 (−1)k αi0 . ı Here a caret denotes omission of the index. each one belonging to the space of sections over the intersection of p + 1 open sets in U.. P). R) = R ⊕ R ⊕ R. Cech cohomology. P) → C p+1 (U.1) we see that αkUi ∩Uk = αiUi ∩Uk .bk .. P) is deﬁned as follows: if α = {αi0 . We consider an open cover U of the circle S 1 formed by three sets ˇ which intersect only pairwise.ip = Ui0 ∩ · · · ∩ Uip . F) = ker δ : C 0 (U. R) is given by d0 (x0 . then ∈C p+1 {(δα)i0 . We start by considering a presheaf P on X and an open cover U of X.ip } of sections of P.. ˇ We deﬁne the Cech complex of U with coeﬃcients in P as the complex whose pth term is the Abelian group ˇ C p (U.ip } ˇ p (U. R) = C 1 (U.2. and deﬁne Ui0 . We have H 0 (U. We denote the corresponding cohomology groups by H k (U. R) → C 1 (U.. The only nonzero diﬀerential d0 : C 0 (U. ˇ ˇ ˇ The Cech diﬀerential δ : C p (U. x2 ) = (x1 − x2 . one has an isomorphism H 0 (U. F) F(X) ˇ ˇ Proof... For instance. If F is a sheaf. F) → F(X).1. x0 − x1 ). P).. So if α ∈ H 0 (U.. C k (U. We assume that U is labelled by a totally ordered set I. each intersection is counted only once.ip ) . COHOMOLOGY OF SHEAVES 43 2. P). ˇ 2. Lemma 3. P) = P(Ui . if p = 0 we have α = {αi } and (3.
SHEAVES AND THEIR COHOMOLOGY Hence H 0 (U. F) are well deﬁned. [5]. A sheaf of rings R on a topological space X is ﬁne if. Every sheaf M of Rmodules is acyclic. R) = C 1 (U. The order must be ﬁxed at the outset.) In this and in the following sections we consider ˇ some properties of Cech cohomology that hold in that case.5. R)/ Im d0 ˇ It is possible to deﬁne Cech cohomology groups depending only on the pair (X. cf. (It is indeed the bad behaviour of Cech cohomology on nonparacompact spaces which motivated the introduction of another cohomology theory for sheaves. / The family {si } is called a partition of unity subordinated to the cover U. (2) for every i ∈ I there is a closed subset Si ⊂ Ui such that (si )x = 0 whenever x ∈ Si . . ˇ 2. A sheaf F of Abelian groups on a topological space X is said to be acyclic if H k (X. Let R be a ﬁne sheaf of rings on a paracompact space X. the groups H k (X. and not on a cover. H 1 (U. For instance.44 3. for any locally ﬁnite oper cover U = {Ui }i∈I of X. R) = ker d0 R R.2. As diﬀerent coﬁnal families give rise to the same inductive limit.4. F) = lim H k (U. F). − → U The direct limit is taken over a coﬁnal subset of the directed set of all covers of X (the order is of course the reﬁnement of covers: a cover V = {Vj }j∈J is a reﬁnement of U if there is a map f : I → J such that Vf (i) ⊂ Ui for every i ∈ I). there is a family {si }i∈I of global sections of R such that: (1) i∈I si = 1. Definition 3. since a cover may be regarded as a reﬁnement of another in many ways. Proposition 3. F). Fine sheaves. Cech cohomology is wellbehaved when the base space X is ˇ paracompact. Definition 3.3. while sheaves of complex or real analytic functions are not. the sheaf of continuous functions on a paracompact topological space as well as the sheaf of smooth functions on a diﬀerentiable manifold are ﬁne. F) = 0 for k > 0. by letting ˇ H k (X. usually called sheaf cohomology.
Example 3.ik−1 jik . Ωk ) = Ωk (X). Let X be any topological space. so that H k (U. P ) → H k (X. Using this result we may recast the proof of the exactness of the MayerVietoris sequence for de Rham cohomology in a slightly diﬀerent form. Lemma 3.. Ωk ) = Ωk (U ) ⊕ Ωk (V ). . Since ˇ C 2 (U.2.. . M) such that δK + Kδ = id (i. any exact sequence of sheaves induces a corresponding long exact ˇ sequence in Cech cohomology. which means that δ is surjective. Given a diﬀerentiable manifold X. For any α ∈ C q (U. COHOMOLOGY OF SHEAVES 45 Proof. then α = δKα if δα = 0. M) = 0 for k > 0. P) → H k (X. . P) → H 0 (X. .. We wish to show that when X is paracompact. ˇ 2. P ) → . C 0 (U.iq−1 = j∈I j<i0 q ρj aji0 .2) 0→P →P→P →0 be an exact sequence of presheaves on X. one has H 1 (U. K is a homotopy operator). Ωk ). Then one has a long exact sequence 0 → H 0 (X.7. C 1 (U.6. Ωk ) = Ωk (U ∩ V ) so that we obtain the exactness of the MayerVietoris sequence. P ) → H k+1 (X. ˇ ˇ This deﬁnes a morphism K : C k (U. which in principle is exact everywhere but at C 1 (U. P ) → H 0 (X. M) → C k−1 (U.. → H k (X.iq−1 + . Let U = {Ui }i∈I be a locally ﬁnite open cover of X.3.. However since the sheaves Ωk are acyclic by Proposition 3. .iq−1 − j∈I i0 <j<i1 ρj ai0 ji1 . .5. Ωk ) → 0 .. Long exact sequences in Cech cohomology.. Ωk ) → C 1 (U. .. and the sequence is exact at that place as well. We have the identiﬁcations H 0 (U. thus getting H k (X. P ) → H 1 (X.. Ωk ) → C 0 (U. = k=0 (−1)k j∈I ik−1 <j<ik ρj ai0 . P ) → . let U be the open cover formed by two sets U and V . Ωk ) = 0. M) = 0 for k > 0.e.iq−1 . we can take direct limit on such covers... Since on a paracompact space the locally ﬁnite open covers are coﬁnal in the family of all covers. and let {ρi } be a ˇ partition of unity of R subordinated to U. Ωk ) = 0 (there are no triple intersections) we have an exact sequence δ ˇ ˇ 0 → H 0 (U. M) with q > 0 we set (Kα)i0 . and let (3.
3) This gives rise to (3. For all k ≥ 0. Proposition 3. One has an exact sequence of presheaves 0 → Q1 → P → P → Q2 → 0 with (3. → H k (U. where τ : C the morphism induced by restriction. .10. P ) → H 0 (U. 0 → T → P → Q2 → 0 Q1 = Q2 = 0 . Theorem 3. k (U. i. Q2 ) = 0. Proof. . where T is the quotient presheaf P/Q1 .46 3. P ) → H 1 (U.e.4) 0 → Q1 → P → T → 0 .9.4) we obtain the desired isomorphism. P ) → 0 which induces the long cohomology sequence 0 → H 0 (U. P ) is an isomorphism.8. Let 0→F →F →F →0 . Q1 ) = H k (X. P ) → H k+1 (U. Proof. the natural morphism H k (X. By Lemma 3. P ) → C • (U. See [13] §2. and let P be the associated sheaf. Using these results we may eventually prove that on paracompact spaces one has ˇ long exact sequences in Cech cohomology. P) → C • (U. . and let α ∈ C k (U. P a presheaf on X whose ˇ associated sheaf is the zero sheaf. P). The proof relies on a standard paracompactness argument. Lemma 3. P) → H k (X. the presheaf U → P(U )/Q1 (U ). P) → C k (W.2) induces an exact sequence of diﬀerential complexes ˇ ˇ ˇ 0 → C • (U. P) is ˇ ˇ There is a reﬁnement W of U such that τ (α) = 0. P ) → .3) yield H k (X. P) → H k (U. Let P be a presheaf on a paracompact space X. P ) → H k (U. Since the direct limit of a family of exact sequences yields an exact sequence.9. let U be an open cover of X. . P ) → . For any open cover U the exact sequence (3. Let X be a paracompact topological space. by taking the direct limit over the open covers of X one obtains the required exact sequence.8 the isomorphisms (3. Then by taking the long exact sequences of cohomology from the exact sequences (3. SHEAVES AND THEIR COHOMOLOGY Proof. P) → H 0 (U.
. i d d is exact. . . If the sheaves L• are acyclic (ﬁne) the resolution is said to be acyclic (ﬁne). Proof. F ) → H 1 (X. Lemma 3. By the previous Lemma we have H 0 (L• (X). F . then P sequence of presheaves 0 → F → F → P → 0. If 0 → F → L• is a resolution. d) H 0 (X.2. We shall consider its cohomology H • (L• (X).4. d). F ) → H k+1 (X. We describe now a very useful way of computing cohomology groups.7) and using the isomorphism H k (X. . F) → H 0 (X. F ) one obtains the required exact sequence. One has an exact By taking the associated long exact sequence in cohomology (cf. This implies the claim. . Let P be the quotient presheaf F/F . F ) → . Let F be a sheaf of abelian groups on X. COHOMOLOGY OF SHEAVES 47 be an exact sequence of sheaves on a paracompact space X. Theorem 3. F ) → H 0 (X. F). the morphism iX : F(X) → L0 (X) is injective. . Let Q be the quotient L0 /F. . . which states ˇ that the de Rham cohomology of a diﬀerentiable manifold and the Cech cohomology of the constant sheaf R are isomorphic.” As a particular case it yields one form of the socalled de Rham theorem.12. Abstract de Rham theorem. F ) → . F) → H k (X. However the sequence of abelian groups 0 1 0 → L0 (X) −−→ L1 (X) −−→ . A resolution of F is a collection of sheaves of abelian groups {Lk }k∈N with morphisms i : F → L0 . Then the sequence of sheaves 0 → F → L0 → Q → 0 is exact. . d d is not exact. F ) → H k (X.11. d) for all k ≥ 0. → H k (X. the sequence of abelian groups 0 → F(X) → L0 (X) → Q(X) is exact. F) H k (L• (X). this result is sometimes called “abstract de Rham theorem. There is a long exact ˇ sequence of Cech cohomology groups 0 → H 0 (X. Proof. By Exercise 3.13.7. Definition 3. 2. If 0 → F → L• is an acyclic resolution there is an isomorphism H k (X. Lemma 3. dk : Lk → Lk+1 such that the sequence 0 1 0 → F → L0 → L1 → . P) = H k (X.
R) = 0 for k > 0. Definition 3. (de Rham theorem. However. The claim then follows for the previous theorem.18.16. Then H k (U. Soft sheaves. Lk−1 ) By Exercise 3. Qk ) Im H 0 (X. The space F(U ) (called “the space of sections of F over U ”) is deﬁned as F(U ) = lim F(V ) − → V ⊃U where the direct limit is taken over all open neighbourhoods V of U . The sequence (3. The contents of this subsection will only be used in Section 4. Let F be a sheaf a on a topological space X. for which the reader is referred to [2.15. 22]. Qk−1 ) H 0 (X. hence are acyclic. A sheaf F is said to be soft if for every closed subset U ⊂ X the restriction morphism F(X) → F(U ) is surjective.17.) Let X be a diﬀerentiable manifold. and let U ⊂ X be a closed subset of X. A consequence of this deﬁnition is the existence of a natural restriction morphism F(X) → F(U ). Qk ) = Qk (X) is the kernel of dk : Lk (X) → Lk+1 so that the claim is proved. Deﬁne Qk = ker dk : Lk → Lk+1 . Let n = dim X. Q1 ) ··· H 1 (X. we do not give the proofs of most claims. For all k k ≥ 0 the cohomology groups HDR (X) and H k (X. k ≥ 1 Since the sheaves Lk are acyclic by taking the long exact sequences of cohomology we obtain a chain of isomorphisms H k (X.5.5. 5. Corollary 3. Proof. Corollary 3.5) 0 → R → Ω0 −−→ Ω1 −−→ · · · → Ωn → 0 X X X d d ∞ e (where Ω0 ≡ CX ) is exact (this is Poincar´’s lemma). Definition 3.48 3. . SHEAVES AND THEIR COHOMOLOGY Proof.7 H 0 (X. Let U be a subset of a diﬀerentiable manifold X which is diﬀeomorphic to Rn . Moreover the sheaves Ω• are X X ∞ modules over the ﬁne sheaf of rings CX . for every open subset U ⊂ X the sequence of groups 0 → F (U ) → F(U ) → F (U ) → 0 is exact. For later use we also introduce and study the notion of soft sheaf. If 0 → F → F → F → 0 is an exact sequence of soft sheaves on a paracompact space X. F) H k−1 (X. Proposition 3. R) are isomorphic. The resolution may be split into 0 → F → L0 → Q1 → 0 . 2. X 0 → Qk → Lk → Qk+1 → 0 .14.
. it is acyclic.23.1.1 in [2]. . Now we have an exact sequence 0 → F → S 0 (F) → F1 → 0. Lemma II. Proposition 3. 6We are cheating a little bit.g. Corollary 3. the sheaf of all sections of the sheaf space F).22. the sequence 0 → F(X) → S 0 F(X) → F1 (X) → 0 obtained from 0 → F → S 0 F → F1 → 0 is exact (Proposition 3. With this procedure we can show that the complex S • (F)(X) is exact.19.18). One can check that the sequence of sheaves 0 → F → S 0 (F) −−→ S 1 (F) −−→ . Let S 0 (F) be the sheaf of discontinuous sections of F (i. Upon iteration we have exact sequences k k 0 → Fk −−→ S k (F) −−→ Fk+1 → 0 i p where S k (F) = S 0 (Fk ). Any soft sheaf of rings R on a paracompact space is ﬁne. and we have an exact sequence 0 → F1 → S 0 (F1 ) → F2 → 0. by the abstract de Rham theorem the claim is proved. The quotient of two soft sheaves on a paracompact space is soft. . (where fk = ik+1 ◦ pk ) is exact. Proposition 3. However a f0 f1 closer inspection of the proof would show that it works anyways.19. The sheaf S 0 (F) is obviously soft. If F is a sheaf on a paracompact space.3. Note that in this way we have shown that for any sheaf F on a paracompact space there is a canonical soft resolution.20.6 Proposition 3.e.4 in [2]. so is F1 by Corollary 3. Proof. Proof. COHOMOLOGY OF SHEAVES 49 Proof. since the sheaf of rings End(S 0 (F )) is not commutative. Since S 0 (F) is an End(S 0 (F))module.20. Proposition 3. On a paracompact space soft sheaves are acyclic. but it may embedded into the soft sheaf S 0 (F1 ).21. If F is a soft sheaf. and the sequence 0 → F1 (X) → S 1 F(X) → F2 (X) → 0 is also exact. adapt the proof of Proposition II. Proof.2. But since all sheaves S • (F) are acyclic by the previous Proposition.. One can e. Cf. the sheaf S 0 (F) is acyclic. The endomorphism sheaf End(S 0 (F)) is soft. Every sheaf F on a paracompact space admits soft resolutions. The sheaf F1 is not soft in general. Since F and S 0 F are soft. hence ﬁne by Proposition 3. Proof.
. C p (U. ˇ To prove this theorem we need to construct the socalled Cech sheaf complex. F)(X) ˇ C p (U. the Cech cohomologies H • (U.ip and the exteded by ˇ ˇ zero to the whole of X).e. F) is a resolution of F (cf. the cohomology of the global sections of the resolution is the cohomology H • (U. For every p deﬁne the sheaf ˇ C p (U. Moreover we have: Lemma 3. by Lemma 3..ip : Ui0 .ip i0 <···<ip (every factor (ji0 . due to the isomorphisms (3.ip = Ui0 ∩· · ·∩Uip . We say that an open cover U = {Ui }i∈I of a topological space X is acyclic for a sheaf F if H k (Ui0 .ip is the sheaf F ﬁrst restricted to Ui0 . F) = 0 for all k > 0 and all nonvoid intersections Ui0 .24. F) are isomorphic. F)) i0 <···<ip H k (Ui0 .. It is not diﬃcult to prove that the complex ˇ• (U.ip → X be the inclusion.25 this resolution is acyclic.... F) → p+1 (U. Since the cohomology H • (U. Leray’s theorem establishes a suﬃcient condition for such an isomorphism to hold. i0 . we get isomorphisms ˇ C (3. the cohomology groups H k (U. the cohomology of the global sections of the resolution is isomorphic to the cohomology of F. From the deﬁnition..ip .. SHEAVES AND THEIR COHOMOLOGY ˇ 2. ˇ H k (X. F) . By the abstract de Rham theorem... II. for all k ≥ 0.7. The Cech diﬀerential induces sheaf morphisms δ : C p (U. ˇ ˇ i.. . [2].25. F) . For every nonvoid intersection Ui0 .. this ˇ turns out to be a very useful tool in the computation of Cech cohomology groups. Theorem 3. F)... Under the hypothesis of Leray’s C theorem.. We may now prove Leray’s theorem... For all p and k.3). (Leray’s theorem) Let F be a sheaf on a paracompact space X. . amounts to computing its de Rham cohomology) to the computation of the cohomology of a cover with coeﬃcients in R.ip )∗ FUi0 .6) ˇ C p (U. 2. thus a problem which in principle would need the solution of diﬀerential equations on . By means of Leray’s theorem we may reduce the problem of ˇ computing the Cech cohomology of a diﬀerentiable manifold with coeﬃcients in the constant sheaf R (which. by taking global sections of the Cech sheaf complex we get the Cech cochain group complex. Prop. F). Then. and let U be an open cover of X which is acyclic for F and is indexed by an ordered set.ip )∗ FUi0 .ip let ji0 .3.. F) = (ji0 . But.ip . Good covers. F) is in generally much easier to compute. via de Rham theorem... ip ∈ I.6..6).50 3. F) are isomorphic.. F) and H • (X.. F) and H k (X. If an open cover U of a topoloˇ gical space X is suitably chosen. Leray’s theorem for Cech cohomology.
Definition 3. for every s ∈ F(U ) there is an open neighbourhood V of U and a section s ∈ F(V ) which restricts to s. A sheaf F on a topological space X is said to be ﬂabby if for every open subset U ⊂ X the restriction morphism F(X) → F(U ) is surjective. For any good cover U of a diﬀerentiable manifold X one has isomorphisms H k (U. A locally ﬁnite open cover U of a diﬀerentiable manifold is good if all nonempty intersections of its members are diﬀeomorphic to Rn . This has the advantage of producing a cohomology theory (called sheaf cohomology) which is bellbehaved (e.g.8. also the original paper by Grothendieck [8] can be fruitfully read. Another kind of sheaves which can be introduced is that of ﬂabby sheaves (also called “ﬂasque”). especially the exercise section. ﬂabby sheaves are acyclic.2. [17]). COHOMOLOGY OF SHEAVES 51 topologically nontrivial manifolds is reduced to a simpler problem which only involves the intersection pattern of the open sets of a cover. It is easy to prove that ﬂabby sheaves are soft: if U ⊂ X is a closed subset. for a discussion of the comparison between the two cohomologies). In this connection the reader may refer to [5. 2. and by the abstract de Rham theorem ﬂabby resolution can be used to compute cohomology. k ≥ 0. 4. So on a paracompact space. by deﬁnition of direct limit. or to [20] where a diﬀerent and more general approach to sheaf cohomology (using injective resolutions) is pursued.. Good covers exist on any diﬀerentiable manifold (cf. for every sheaf F.26.2 was good. That is. We have therefore Proposition 3. not just on paracompact ones. it has long exact sequences in cohomology) on every topological space. 2].27. but in general they do not (cf. R) . [11]. good covers are acyclic for the constant sheaf R. Since F is ﬂabby. Due to Corollary 3. so we computed there the de Rham cohomology of the circle S 1 . One can further pursue this line and use ﬂabby resolutions (for instance. its cohomology is by deﬁnition the cohomology of the global sections of its canonical ﬂabby resolution (it then turns out that cohomology can be computed with any acyclic resolution). The cover of Example 3.5) to deﬁne cohomology. . Flabby sheaves. It follows from our treatment that on a paracompact ˇ topological space the sheaf and Cech cohomology coincide.15. as one can easily check by the deﬁnition itself. the canonical ﬂabby resolution of Section 2.5 is ﬂabby. We should also notice that the canonical soft resolution S • (F) we constructed in Section 2. R) H k (X. s can be extended to the whole of X.
all these ˇ cohomologies coincide with the Cech cohomology of X with coeﬃcients in the constant sheaf G. In particular. In algebraic topology one attaches to a topological space X several cohomologies with coeﬃcients in an abelian group G. SHEAVES AND THEIR COHOMOLOGY 2. whenever X is paracompact and locally Euclidean.52 3.28.9. . Let X be a paracompact locally Euclidean topological space. Comparison with other cohomologies. The singular cohomology of X with coeﬃcients in G is ˇ isomorphic to the Cech cohomology of X with coeﬃcients in the constant sheaf G. and let G be an abelian group. we have the following result: Proposition 3. Loosely speaking.
i. d). and the K¨nneth theorem.. K= n∈Z Kn ..e. d) is ﬁltered. . d) be a graded diﬀerential module. i. i. d2 = 0 . 1The choice of having K = K for p ≤ 0 is due to notational convenience. d: K n →K n+1 . Kp = K if p ≤ 0 .CHAPTER 4 Spectral sequences Spectral sequences are a powerful tool for computing homology. d : K n → K n+1 . K n ⊂ Kn . . A graded submodule of (K. 1. . u Spectral sequences are a diﬃcult topic. Note that by assumption (since every Kp+1 is a graded subgroup of Kp ) the ﬁltration i is compatible with the grading. cohomology and homotopy groups. is a ﬁltration of (K. Examples that we shall consider are another proof of the Cechde Rham theorem. Our treatment basically follows [3].e. the Leray spectral sequence. due to Massey [18]. Filtered complexes Let (K. A sequence of nested graded submodules K = K0 ⊃ K1 ⊃ K2 ⊃ .1) n n n K n = K0 ⊃ K1 ⊃ K2 ⊃ . Often they allow one to trade a diﬃcult computation for an easier ˇ one. p 53 . if we deﬁne Kp = K i ∩ Kp .e. n n+1 is a ﬁltration of K i . and moreover dKp ⊂ Kp . Therefore we have chosen what seems to us to be the simplest approach. . We then say that (K. basically because the theory is quite intrincate and the notation is correspondingly cumbersome. then (4.. and associate with it the graded complex1 Gr(K) = p∈Z Kp /Kp+1 . d) is a graded subgroup K ⊂ K such that dK ⊂ K . K = n∈Z K n .
1. q≥0 K i. A double complex is a collection of abelian groups K p. d) be the associated total complex : Ti = p+q=i δ 1 δ2 + δ2 δ1 = 0 .2) 0 → G −−→ G −−→ E → 0 i j 2This assumption is made here for simplicity but one could let p.1 is regular since Tp = 0 for p > i.q . q range over the integers.1) is ﬁnite. d).2. however some of the results we are going to give would be no longer valid.q+1 such that δ1 2 = δ2 2 = 0 .q . K p.54 4. and one has an exact sequence (4. The successive quotients K p. Let (T. A ﬁltration K• of (K. SPECTRAL SEQUENCES Example 4. and indeed i−p i Tp = T ∩ Tp = j=0 i K i−j.q → K p+1. d : T i → T i+1 deﬁned by d = δ1 + δ2 (note that the deﬁnition of d implies d2 = 0). Then letting Tp = i≥p. δ2 : K p. Let K• be a ﬁltration of a diﬀerential module (K.q → K p.q .q . i For instance. the ﬁltration in Example 4. . The inclusions Kp+1 → Kp induce a morphism i : G → G (“the shift by the ﬁltering degree”). in other words. This satiﬁes Tp of the ﬁltration are Tp /Tp+1 = Definition 4. for every i there is a number (i) such that i Kp = 0 for p > (i). and let G= p∈Z Kp .2 and morphisms δ1 : K p. q∈N we obtain a ﬁltration of (T. The spectral sequence of a ﬁltered complex At ﬁrst we shall not consider the grading.q T for p ≤ 0. q ≥ 0. with p.j . 2. d) is said to be regular if for every i ≥ 0 the ﬁltration (4. d).
The morphism i induces morphisms i : H(Kp+1 ) → H(Kp ). where Fp = p )).2) one gets an exact triangle in cohomology (cf. G1 = sequence p∈Z H(Kp ) H(G). The diﬀerential d induces diﬀerentials in G and E. Let us now assume that the ﬁltration K• has ﬁnite length...e. since it is the inclusion of the image of i : H(K1 ) → H(K) into H(K). −−→ H(K1 ) −−→ H(K) −−→ H(K−1 ) −−→ . Fp is the image of H(Kp ) into H(K).. i i ∼ ∼ i. . . so that from (4. . and the procedure stabilizes: Gr We deﬁne G∞ = G +1 . so that G +2 G +1 . we have G∞ Fp p∈Z Gr+1 for r ≥ +1. This procedure is then iterated: G3 is the sum of the terms in the sequence 0 → i(i(H(K )))) → i(i(H(K −1 ))) → i(i(H(K2 )) → i(H(K1 )) → H(K) −−→ H(K−1 ) −−→ .1) (4. Next we deﬁne G2 as the sum of the terms of the 0 → i(H(K ))) → i(H(K −1 )) → . . .2. Since dKp ⊂ Kp for every p. ∼ ∼ and now the morphisms i(i(H(K2 )) → i(H(K1 )) and i(H(K1 )) → H(K) are injective. all the morphisms in the sequence 0 → i (H(K ))) → i −1 (H(K −1 )) → . → i(H(K1 )) → H(K) −−→ H(K−1 ) −−→ . i. Section 1. → i(H(K1 )) → H(K) −−→ H(K−1 ) −−→ . we may consider the cohomology groups H(Kp ). .. THE SPECTRAL SEQUENCE OF A FILTERED COMPLEX 55 with E Gr(K).4) H(K) = F0 ⊃ F1 ⊃ · · · ⊃ F ⊃ F +1 ip (H(K = 0. .. ... . The groups Fp provide a ﬁltration of H(K). i. (4. ∼ ∼ are injective. ∼ ∼ Note that the morphism i(H(K1 )) → H(K) is injective..e.3) H(G) i cHH HH HH H k HH / H(G) vv vv vv v {vv j H(E) where k is the connecting morphism.e. Deﬁne G1 to be the direct sum of the terms on the sequence (which is not exact) 0 → H(K ) −−→ H(K i i −1 ) −−→ . When we reach the step . .. Kp = 0 for p greater than some (called the length of the ﬁltration).
56
4. SPECTRAL SEQUENCES
We come now to the construction of the spectral sequence. Recall that since dKp ⊂ Kp , and E = p Kp /Kp+1 , the diﬀerential d acts on E, and one has a cohomology group H(E) wich splits into a direct sum H(E)
p∈Z
H(Kp /Kp+1 , d) .
The cohomology group H(E) ﬁts into the exact triangle (4.3), that we rewrite as (4.5) G1
i1
`BB BB BB B k1 B
/ G1     j1 ~
E1 where E1 = H(E). We deﬁne d1 : E1 → E1 by letting d1 = j1 ◦ k1 ; then d2 = 0 since 1 the triangle is exact. Let E2 = H(E1 , d1 ) and recall that G2 is the image of G1 under i : G1 → G1 . We have morphisms i2 : G2 → G2 , , where (i) i2 is induced by i1 by letting i2 (i1 (x)) = i1 (i1 (x)) for x ∈ G1 ; (ii) j2 is induced by j1 by letting j2 (i1 (x)) = [j1 (x)] for x ∈ G1 , where [ ] denotes taking the homology class in E2 = H(E1 , d1 ). (iii) k2 is induced by k1 by letting k2 ([e]) = i1 (k1 (e)). Exercise 4.1. Show that the morphisms j2 and k2 are well deﬁned, and that the triangle (4.6) G2
i2
j2 : G2 → E2 ,
k2 : E 2 → G 2
`BB BB BB B k2 B
/ G2     j2 ~
E2 is exact. We call (4.6) the derived triangle of (4.5). The procedure leading from (4.5) to the triangle (4.6) can be iterated, and we get a sequence of exact triangles Gr
ir
`BB BB BB B kr B
/ Gr     jr ~
Er where each group Er is the cohomology group of the diﬀerential module (Er−1 , dr−1 ), with dr−1 = jr−1 ◦ kr−1 . As we have already noticed, due to the assumption that the ﬁltration K• has ﬁnite length , the groups Gr stabilize when r ≥ + 1, and the morphisms ir : Gr → Gr
2. THE SPECTRAL SEQUENCE OF A FILTERED COMPLEX
57
become injective. Thus all morphisms kr : Er → Gr vanish in that range, which implies dr = 0, so that the groups Er stabilize as well: Er+1 Er for r ≥ + 1. We denote by E∞ = E +1 the stable value. Thus, the sequence
∞ 0 → G∞ −−→ G∞ → E∞ → 0
i
is exact, which implies that E∞ is the associated graded module of the ﬁltration (4.4) of H(K): E∞ Fp /Fp+1 .
p≤
Definition 4.2. A sequence of diﬀerential modules {(Er , dr )} such that H(Er , dr ) Er+1 is said to be a spectral sequence. If the groups Er eventually become stationary, we denote the stationary value by E∞ . If E∞ is isomorphic to the associated graded module of some ﬁltered group H, we say that the spectral sequence converges to H. So what we have seen so far in this section is that if (K, d) is a diﬀerential module with a ﬁltration of ﬁnite length, one can build a spectral sequence which converges to H(K). Remark 4.3. It may happen in special cases that the groups Er stabilize before getting the value r = + 1. That happens if and only if dr = 0 for some value r = r0 . This implies that dr = 0 also for r > r0 , and Er+1 Er for all r ≥ r0 . When this happens we say that the spectral sequence degenerates at step r0 . Now we consider the presence of a grading. Theorem 4.4. Let (K, d) be a graded diﬀerential module, and K• a regular ﬁltration. There is a spectral sequence {(Er , dr )}, where each Er is graded, which converges to the graded group H • (K, d). Note that the ﬁltration need not be of ﬁnite length: the length (i) of the ﬁltration of K i is ﬁnite for every i, but may increase with i.
n n+1 Proof. For every n and p we have d(Kp ) ⊂ Kp , therefore we have cohomology groups H n (Kp ). As a consequence, the groups Gr are graded:
Gr
n∈Z
n Fr = n,p∈Z
ir−1 (H n (Kp ))
and the groups Er are accordingly graded. We may construct the derived triangles as before, but now we should pay attention to the grading: the morphisms i and j have degree zero, but k has degree one (just check the deﬁnition: k is basically a connecting morphism). Fix a natural number n, and let r ≥ (n + 1) + 1; for every p the morphisms
n+1 n+1 ir : Fr → Fr
58
4. SPECTRAL SEQUENCES
are injective, and the morphisms
n n+1 kr : Er → Fr
are zero. These are the same statements as in the ungraded case. Therefore, as it n happened in the ungraded case, the groups Er become stationary for r big enough. n+1 n Note that Gn = ⊕p∈Z Fp , where Fp+1 = i (n+1) (H n+1 (Kp+1 )), and that the morphism ∞ n n i∞ sends Fp+1 injectively into Fp for every n, and there is an exact sequence
∞ n 0 → Gn −−→ Gn → E∞ → 0 . ∞ ∞
i
This implies that Er is the graded module associated with the graded complex H • (K, d).
n The last statement in the proof means that for each n, F• is a ﬁltration of H n (K, d), n n n and E∞ p∈Z Fp /Fp+1 .
3. The bidegree and the ﬁveterm sequence The terms Er of the spectral sequence are actually bigraded; for instance, since the ﬁltration and the degree of K are compatible, we have Kp /Kp+1
q∈Z q q Kp /Kp+1 q∈Z p+q p+q Kp /Kp+1
and E0 = E is bigraded by E0 =
p,q∈Z p,q E0 p,q p+q p+q with E0 = Kp /Kp+1 .
Note that the total complex associated with this bidegree yields the gradation of E.
p,q+1 p+q p+q+1 p,q , and Let us go to next step. Since d : Kp → Kp , i.e., d : E0 → E0 E1 = H(E, d), if we set p,q p,• E1 = H q (E0 , d)
H p+q (Kp /Kp+1 )
we have E1
p,q p,q∈Z E1 .
If we go one step further we can show that
p,q p+1,q d 1 : E1 → E1 . p,q p+q p+q Indeed if x ∈ E1 H p+q (Kp /Kp+1 ) we write x as x = [e] where e ∈ Kp /Kp+1 so that k1 (x) = i1 (k(e)) ∈ H p+q+1 (Kp+1 ) and
d1 (x) = j1 (k1 (x)) = j1 (k(e)) ∈ H p+q+1 (Kp+1 /Kp+2 ) As a result we have E2
p,q p,q∈Z E2 p,q E2
p+1,q E1 .
with
•,q H p (E1 , d1 ) . p,q p,q∈Z Er
The same analysis shows that in general Er
with
p,q p+r,q−r+1 d r : Er → Er
1−r As a consequence. There are canonical morphisms H q (K) → Er . morphisms Er → E∞ . p.q q q q hence E∞ = 0 for p < 0 and E∞ F0 /F1 H q (K)/F1 .0 p p p n Since Fp = 0 for p > n we have E∞ Fp /Fp+1 Fp so that one has an injective p.4. all p. the result in Theorem 4. e. The hypothesis of the Lemma implies that Er = 0 for q < 0 (indeed. [5]. n Lemma 4.q Proof. we have inclusions Er+1 ⊂ Er (Er+1 is the subgroup of cycles 0.q 0. and determine cohomology classes in Er+1 .q in Er ). and will anyway be useful in the following. So cohomology classes are cycles. so that there is a surjective 0. n Proof. p+q Fp p. Assume that Kp = 0 if p > n (so. The spectral sequences associated with a double complex In this Section we consider a double complex as we have deﬁned in Example 4.1. since then it should −r. THE SPECTRAL SEQUENCES ASSOCIATED WITH A DOUBLE COMPLEX 59 and moreover we have p. This means we p.0 p. p.0 have a morphism Er → Er+1 .0 p. 0. This yields an inclusion E∞ ⊂ Er for all r.0 p. d Proof.0 morphism E∞ → H p (K).e. in particular. for a proof cf.0 p.q Lemma 4. and composing. 0.0 0. i. so that Fq = 0 if q < 0 since then Kp = 0).3.0 2 0 → E2 → H 1 (K) → E2 −−→ E2 → H 2 (K) . Due to the presence of the bidegree. n Proposition 4. The next two Lemmas establish the existence of the morphisms that we shall use to introduce the socalled ﬁveterm sequence.q are elements in Er+1 . the ﬁltration is p. The group G= p∈Z Tp = p∈Z n≥p. Assume that Kp = 0 if p > n.1.4 may be somehow reﬁned.0 regular).q morphism H q (K) → E∞ . Since cohomology classes 0. q∈N K i.q E∞ p+q p+q Fp /Fp+1 .g. (The ﬁveterm sequence). Since Kp K for p ≤ 0 we have Fp H n (Kp ) = H n (K) for p ≤ 0.. Composing we have a morphism Er → H p (K). 0.q 0. We shall not prove the exactness of the sequence here.1 2. p. We shall use the notation in Example 4. 4.q 0. Then for every r ≥ 2 there is a morphism Er → H p (K).q Note now that a nonzero class in Er cannot be a boundary.0 elements in Er are cycles.2.0 p+r.q 0.q .1. The morphisms involved in the sequence in addition to d2 have been deﬁned in the previous two Lemmas.q+r−1 come from Er = 0.0 p.q Combining the two arguments we obtain morphisms H q (K) → Er . There is an exact sequence 1.q 0.q 0. p+1 p+q = ir (H p+q (Kp )) for r big enough. for r ≥ 2 the diﬀerential dr : Er → Er maps to zero.
q we obtain the complex (4.• . The ﬁrst three terms are easily described.n ) is identiﬁed with δ1 . i. One has p. d : Gk → Gk+1 .q+1 .q .• H q ( E1 .p ).n ) → H p+q+1 ( n∈Z K p+1. Hence the diﬀerential q∈Z K d1 : H p+q ( n∈Z K p. δ1 ) p. H p+q (Tp /Tp+1 ) and Tp /Tp+1 p.p E2 H p (K •. One should notice that by exchanging the two degrees in K (i.q p+q=n j=0 K = j=0 The operators δ1 .e.j p. δ1 ) .10) (4.p E1 q. δ 2 ) .11) q. and (4. Both sequences converge to the same graded group.q E1 H q (K p.7) Gn = p∈Z n Tp p∈Z j=0 K n−j.8) p.q p+1.q p. considering another double complex K such that K p. and this often provides interesting information. δ2 = Gn.q with E1 At next step we have d1 : E1 → E1 p. the cohomology of the total complex (but the corresponding ﬁltrations are in general diﬀerent). we obtain another spectral sequence.j but it also bigraded. dr }..q → Gn. For the second spectral sequence we get (4.j = Gn .q coincides with δ2 : K p. n−p G p+q=n p.q .q = Tq .q p.60 4. δ2 ) . Notice that if we form the total complex q p+q=n G p+q−j. with bidegree p+q Gp.q = K q.q → K p.q H p (E1 .9) p. .q+1 .. We analyze the spectral sequence associated with these data. and one so that the diﬀerential d0 : E0 → E0 has (4.7) back: K n−j. that we denote by { Er . δ2 and d = δ1 + δ2 act on G: δ1 : Gn.q E2 •.q → Gn+1.e.q E0 p+q p+q Tp /Tp+1 K p.q+1 p.q . SPECTRAL SEQUENCES has natural gradation G = ⊕n∈Z Gn given by n−p (4.
. 3Here a notational conﬂict arises.q E1 ˇ H q (K •. and deﬁne the double complex ˇ K p. Actually δ and d commute rather than anticommute. so that the spectral sequence degenerates at the second step. R). so that H p (T. This implies that d2 = 0. D) p p HDR (X). Ωp ) . Let us now consider the second spectral sequence. Comparing with what we got from the ﬁrst sequence.0 E1 H 0 (U. We start analyzing the spectral sequences from the terms E1 . and we have H p (T. D) H p (U. i.p . d)) p HDR (X) . Ωq ) . R) HDR (X). Since all Ui0 . we have p. p.. δ) = H q (C • (U.0 E2 H p (Ω• (X). R).0 E1 ˇ C p (U. Taking a direct limit on good covers. we obtain HDR (X) p H p (U. The ﬁrst diﬀerential δ1 is basically the Cech diﬀerential δ. d) i0 <···<ip q HDR (Ui0 . A simple application of the two spectral sequences associated with ˇ a double complex provides another proof of the Cechde Rham theorem. p.q E1 H q (K p. R). we obtain H p (X.ip ) .ip are contractible we have p. and p. For the ﬁrst. ˇ i. Ωp ). The resulting ﬁltration of H p (T. . so that we shall denote by D the diﬀerential of the total complex T.1..• . K •. R).q E1 = 0 for q = 0 . D) has and E∞ = 0 for q = 0 and E∞ only one nonzero quotient. while p. Ωp ) Ωp (X) p. R) . δ) = H p (U.. THE SPECTRAL SEQUENCES ASSOCIATED WITH A DOUBLE COMPLEX 61 Example 4.q At next step we have therefore E2 = 0 for q = 0.q H p (U.0 E2 ˇ H p (C • (U.q is the complex of Cech cochains of U with coeﬃcients in the sheaf of diﬀerential ˇ qforms. R) p. we have p.e.q = C p (U.q as δ1 = (−1)q δ (this of course leaves the spaces of boundaries and cycles unchanged). Again the spectral sequence degenerates at the second step.q As a consequence we have E2 = 0 for q = 0.4. the iso• morphism H • (X.e.q E1 = 0 for q = 0 .. We have p..0 p. while δ2 is the 3 exterior diﬀerential d. δ) = H q (U. p. Let U = {Ui } be a good cover of X. Since the sheaves Ωp are acyclic. R) HDR (X) for a diﬀerentiable manifold X. but this is easily settled by deﬁning the action of δ1 on K p.
and let U be an open cover of X. Im f : Lq−1 (U ) → Lq (U ) The E1 term of the ﬁrst spectral sequence is p.p . and Er H p (T. p. f ) . considering the ﬁrst spectral sequence. So. Hq (L• )) ˜ where. Let (L• .p E2 ˇ H q (C • (U.62 4. F) .2. f )) ˇ ˜ C p (U.• H p ( E1 .0 0. The second term of the sequence is p. since X is paracompact. Lq ) q. there is a spectral sequence E• whose second term is Ep.q E1 H q (K •.1 to a much general situation. 5. F) (since the same does the ﬁrst sequence).q p. δ2 ) = H q (C p (U. d) (or Er H q (T. we have: Proposition 4. and we p.1. We introduce the double complex ˇ K p. δ1 ) H p (U. we have replaced the presheaves H• with the corresponding sheaves H• (possibly replacing the cover U by a suitable reﬁnement).q with r ≥ 1.q = C p (U. L• ). δ 2 ) H p (H q (U.0 have E2 = 0 for q = 0 and E2 H p (U. Hq (L• )). The ﬁrst spectral sequence degenerates at the second step. By taking direct limit over the cover U. then Hq (L• ) = 0 for q = 0. f ) be a complex of sheaves on a paracompact topological space X. δ) H p (U.2. The spectral sequence of a resolution. The canonical ﬁltrations of a double complex always satisfy the hypothesis of Lemma 4. δ1 ) ˇ ˜ H p (C • (U. L• ). which 2 converges to the graded group H • (X. The second spectral sequence does not degenerate.1. SPECTRAL SEQUENCES Remark 4. we have Er = 0 for q = 0 (or for p = 0) then the sequence degenerates at p. Given a resolution L• of a sheaf F on a paracompact space X.q step r. δ1 ) q. F).q H p (E1 . Hq (L• )) . L• ).• . Let assume now that L• is a resolution of a sheaf F. d)).q E2 •. and H0 (L• ) F. F). These are the sheaves associated with the quotient presheaves ker f : Lq (U ) → Lq+1 (U ) ˜ Hq (U ) = . we obtain morphisms (again taking a direct limit) H q (L• (X). Lq ). In this section we extend Example 4. Lp ). but we may say that it converges to the graded group H • (U. Some applications 5. .q = H q (H p (X. f ). For the second spectral sequence we have p. f ) → H q (X. We shall denote by Hq (L• ) the cohomology sheaves of the complex L• . From this example we may get the general result that if at step r.q E1 ˇ H q (K p.
. The sheaf Rk π∗ F is isomorphic to the sheaf associated with the presheaf P k on Y deﬁned by P k (U ) = H k (π −1 (U ).2.12) is no longer exact. f ) is a ﬂabby resolution of F. f ) . we have another proof of the abstract de Rham theorem 3. by the abstract de Rham theorem we have isomorphisms H k (L• (π −1 (U ). i. where U is a locally ﬁnite open cover of Y . π∗ Lq ). Sections 3. The spectral sequence of a ﬁbred space. taking the associated long exact cohomology sequences and suitably composing the morphisms. However. where Y is a second paracompact space.2. then the second spectral sequence degenerates at the second step as well. f ). The same morphisms could be obtained by breaking the exact sequence 0 → F → L• into short exact sequences. The two spectral sequences we have previously studied yield at the second term p.2.13.13. and we get isomorphisms H p (X. on the choice of the resolution. . as in the proof of the abstract de Rham theorem 3. F) . We consider the complex (4.12) 0 → π∗ F → π∗ L0 −−→ π∗ L1 −−→ . SOME APPLICATIONS 63 In general these are not isomorphisms. H p (U. f f where (L• . Rq π∗ F) H q (H p (U. Proof.5. Im f : Lk−1 (π −1 (U )) → Lk (π −1 (U )) Since the restriction of a ﬂabby sheaf to an open subset is ﬂabby. but otherwise the complex (4. F).8). Proposition 4.e. We shall associate a spectral sequence to these data. f ) . A further specialization is obtained if the resolution L• is acyclic.q E2 p. Note that R0 π∗ F π∗ F. Let F be a sheaf on a paracompact space X and π : X → Y a continuous map.These sheaves are called the higher direct images of F. Rk π∗ F is by deﬁnition the sheaf associated with the presheaf U ker f : Lk (π −1 (U )) → Lk+1 (π −1 (U )) = H k (L• (π −1 (U ). This implies that the sheaves Rk π∗ F do not depend. F) H p (L• (X). We shall use the fact that every sheaf of abelian groups on space admits ﬂabby resolutions (cf. the sheaves π∗ L• are ﬂabby.2. up to isomorphism.. The morphism π∗ F → π∗ L0 is injective.q E2 H k (π −1 (U ).5 and 3. π∗ L• ). f ) whence the claim follows. 5. ˇ Let us consider the double complex C p (U. We denote by Rk π∗ F the cohomology sheaves Hk (π∗ L• ).
H We shall need the following version of the universal coeﬃcient theorem. This is an isomorphism under some conditions. e. so that the only nonzero terms in the spectral sequence E2 are Ep. here F is to be considered as restricted to π −1 . H q (L• (X). Y be topological spaces. The K¨ nneth theorem.13) (Rk π∗ F)y → H k (π −1 (y). then H k (X. 2 which converges to the graded group H • (X. if Y is locally compact and π is proper (cf. Z) and H • (Y.. i−1 F) where y iy : π −1 (y) → X is the inclusion.4. F). π∗ F).q and one has E∞ = 0 for p = 0. Z) ⊗Z G for all k ≥ 0. [5]). This implies that Rk π∗ F = 0 for k > 0. F) . SPECTRAL SEQUENCES Since the sheaves π∗ L• are soft (hence acyclic) the second spectral sequence degenerates.64 4. Cf.. If every point y ∈ Y has a system of neighbourhoods whose preimages are acyclic for F.e. Given a continuous map of paracompact spaces π : X → Y and a sheaf F on X. p. Rq π∗ F). F). We shall denote by the same symbol G the corresponding constant sheaves on the spaces X. [19]. F) = 0 for all k > 0. there is a natural map (4.0 H p (Y. . i. G). there is a spectral sequence E• whose second term is Ep.g. We describe without proof the relation between the stalks of the sheaf Rk π∗ F at points y ∈ Y and the cohomology groups H k (π −1 (y). − → while H k (π −1 (y). Proof. Proposition 4.q E∞ 0. F) is the direct limit of the groups H k (V.3 one obtains Leray’s theorem: Corollary 4. Let X.q = H p (Y.q E2 H q (H 0 (Y.5.3. F) . F) where V ranges over all open neighbourhoods of π −1 (y). then H k (X. Y and X × Y . Since (Rk π∗ F )y = lim (Rk π∗ F)(U ) − → y∈U y∈U lim H k (π −1 (U ). F) . we have: Proposition 4. more precisely we should write H k (π −1 (y). This happens for instance when both X and Y are compact. f ) H q (X. As a simple Corollary to Proposition 4. If X is a paracompact topological space and G a torsionfree group. 2 5. G) H k (X.3. The sequence degenerates and the claim follows. and 0. π∗ F) for all k ≥ 0. F) H k (Y. π∗ L• ). and G an abelian u group. Proof. G) in terms of the groups H • (X. The K¨nneth theorem computes the cohomology groups u • (X × Y. The hypothesis of the Corollary means that every y ∈ Y has a system of neighbourhoods {U } such that H k (π −1 (U ). f ) Again after taking a direct limit.
SOME APPLICATIONS 65 Proposition 4. Proof. If U is a contractible open set in U . G) . If V ⊂ U . G) corresponds to the identity of H • (Y. G).5. G).6. and that X and Y are compact Hausdorﬀ and locally Euclidean.13) is an isomorphism. G) → H • (V × Y. Assume that the groups H • (Y. G) p+q=k H p (X. from its isomorphism with singular cohomology) we have H • (U × Y. G) H • (Y. These facts imply that Rp π∗ G is the constant sheaf on X with stalk H p (Y.g. Z)⊗Z H q (Y. H q (Y. G) have no torsion over Z.q H p (X. then by the homotopic invariance of the cohomology with coeﬃcients in a constant sheaf (which follows e. the morphism H • (U × Y. Z) ⊗ H q (Y. Under the present hypotheses the morphism (4. G)). G).q H p (X. Let π : X × Y → X be the projection onto the ﬁrst factor. 2 . The second term of the spectral sequence of Proposition 4. Then. we have Ep.3 becomes Ep. G). H k (X × Y. By the universal coeﬃcient theorem. 2 since the groups H q (Y. G) have no torsion over Z.
.
Part 2 Introduction to algebraic geometry .
.
the Cauchy¯ Riemann conditions read fz = 0. Let U ⊂ C be an open subset. and that the transition functions are biholomorphic. one can show that holomorphic functions are analytic.e. y) + iβ(x. Basic deﬁnitions and examples 1. If we use z. Complex manifolds are deﬁned as diﬀerentiable manifolds. y) = α(x. The map f itself is then said to be biholomorphic. Definition 5.3. z2 . 1 ) and (0. The reader is assumed to be acquainted with the rudiments of the theory of diﬀerentiable manifolds. but requiring that the local model is Cn . V of Cn are said to biholomorphic if there exists a bijective holomorphic map f : U → V whose inverse is holomorphic. The same deﬁnition can be given for holomorphic functions of several variables.2. The stereographic projections give local complex coordinates z1 . 0. 1. the transition function z2 = 1/z1 is deﬁned in C = C − {0} and is biholomorphic. Example 5. and we set f (x. and identify the tangent planes at (0. Two open subsets U . 69 . Complex manifolds.1. z as variables. Definition 5.1) αx = βy . (The Riemann sphere) Consider the sphere in R3 centered at the origin and having radius 1 .1. αy = −βx (these are the CauchyRiemann conditions). − 1 ) with 2 2 2 C. Moreover.2. the holomorphic functions are the C 1 function of ¯ the variable z. If elements in C are written z = x + iy. then this condition can be written as (5. y). i. Holomorphic functions.CHAPTER 5 Complex manifolds and vector bundles In this chapter we give a sketchy introduction to complex manifolds. and are such that all transition functions −1 ψi ◦ ψj : ψj (Ui ∩ Uj ) → ψi (Ui ∩ Uj ) are biholomorphisms. An ndimensional complex manifold is a second countable Hausdorﬀ topological space X together with an open cover {Ui } and maps ψi : Ui → Cn which are homeomorphisms onto their images. We say that a function f : U → C is holomorphic if it is C 1 and for all x ∈ U its diﬀerential Dfx : C → C is not only Rlinear but also Clinear. 0. 1.
. . The coordinates deﬁned by the maps ψi . .4. ˆ Mk. . . n).n is a complex manifold of dimension n2 . ..n is a complex manifold of dimension ˆ kn. .e. usually denoted (z 1 . .e. z1 z i−1 z i+1 1 zn ... i . and their transition functions −1 ψi ◦ ψj : −1 ψi ◦ ψj (z 1 . Here are some of its relevant subgroups: ..n = {matrices in Mk. so that Pn is a complex manifold (we have assumed that i < j). . .. . Example 5. i. . If w = (w0 . wn ) are said to be the homogeneous coordinates of the point π(w).. . are biholomorphic. .n = {k × n matrices with complex entries.. . and deﬁne a map ψi : Ui → Cn . The previous formula for n = 1 shows that P1 is biholomorphic to the Riemann sphere.. (Projective spaces) We deﬁne the ndimensional complex projective space as the space of complex lines through the origin of Cn+1 . so it is a complex manifold of dimension kn as well. . Compact Riemann surfaces play a distinguished role in algebraic geometry. z n ). let Ui = π(Ui ). un ) is another set of homogeneous coordinates for π(w). . In particular. (The general linear complex group). i. . Let Mk. i zi z z z z ↑ jth argument . . are called aﬃne or Euclidean coordinates. wn ]) = wi−1 wi+1 wn w0 . k ≤ n} ˆ Mk.n = i=1 i. . . . The map π restricted to the unit sphere in Cn+1 is surjective. {A ∈ Mk.. . i wi wi wi w . ˜ ˜ Denote by Ui ⊂ Cn+1 the open set where wi = 0. If (u0 . The sets Ui cover Pn . . . C∗ By standard topological arguments Pn with the quotient topology is a Hausdorﬀ secondcountable space. with λ ∈ C∗ (i = 0. as its second description shows. . . .n of rank k}. i .. the maps ψi are homeomorphisms. so that Pn is compact. they are sets of zeroes of systems of homogeneous polynomials).n such that det Ai = 0} where Ai . . Example 5. . Pn = Let π : Cn+1 − {0} → Pn be the projection. .n . Mk. The numbers (w0 .. wn ) ∈ Cn+1 we shall denote π(w) = [w0 . .70 5. they are all algebraic (i. ψ([w0 . . the general linear group Gl(n. .e.. wn ].. COMPLEX MANIFOLDS AND VECTOR BUNDLES 1dimensional complex manifolds are called Riemann surfaces... . z n ) = ψj (Uj ) → ψi (Ui ).. as we shall see in Chapter 7. then ui = λwi .. Mk.. A are the k × k minors of A.5. . . . Cn+1 − {0} .n is an open subset in Mk. . C) = ˆ Mn.
C) can be replaced by U (n). where Y is a complex manifold.6. Submanifolds. For any k < n the inclusion of Ck+1 into Cn+1 obtained by setting to zero the last n − k coordinates in Cn+1 yields a map Pk → Pn . Example 5.2) M= A 0 B C where the matrices A. Given a kplane. this representation gives Gk. C) is a submanifold of Gl(n. C) × Mk. λ = v1 ∧ · · · ∧ vk U (n) U (k) × U (n − k) . n. This is the Grassmann variety of kplanes in Cn . C) As the reader may check.n . C) Gk. a submanifold of X is a pair (Y. C) formed by invertible complex matrices having a block form (5. Gl(k. the action of Gl(n. while B can be any matrix. C) on it yields another plane (possibly coinciding with the previous one). and since Gl(k. C) is a complex manifold of dimension k 2 + n2 − nk. (n − k) × k. 1.n . n. (ii) SU (n) = {A ∈ U (n) such that det A = 1}.n ≡ Pn − 1).n singles out (up to a complex factor) a decomposable element in Λ k Cn . ι). Since in the previous reasoning Gl(n. C are k × k.1. B. An element in Gk.n showing that Gk. Gl(k. Since a matrix of the form (5. (Grassmann varieties) Let Gk. (iii) the group Gl(k. n. dimR SU (n) = n2 − 1. and (n − k) × (n − k). Example 5.n = {space of kdimensional planes in Cn } (so G1.n is compact. BASIC DEFINITIONS AND EXAMPLES 71 (i) U (n) = {A ∈ Gl(n. the reader may check that this realizes Pk as a submanifold of Pn .7. Gl(k. and ι : Y → X is an injective holomorphic map whose jacobian matrix has rank equal to the dimension of Y at any point of Y (of course Y can be thought of as a subset of X). C) × Gl(n − k. respectively. and dimR U (n) = n2 . Example 5. C) which leaves the given kplane ﬁxed is isomorphic to Gl(k.2) is invertible if and only if A and C are. Given a complex manifold X.n the structure of a complex manifold of dimension k(n−k).3. C). Gl(k. we also have the representation Gk. n. C)∩U (n) = U (k)×U (n−k). C) is biholomorphic to the product manifold Gl(k. n. so that Gl(n. n. C) such that AA† = I}. The subgroup of Gl(n.8. these two groups are real (not complex!) manifolds. C). n.
. by the identiﬁcation Cn R2n ... and since a biholomorphic map is a C ∞ diﬀeomorphism. 2. The elements in Ωp. Forms of type (p.j1 .q X where Ωp... If a basis k n ..... . X has an underlying structure of 2ndimensional real manifold. in this case we have J = ¯ 0 −1 so that the orientation gets reversed. . one has a representation {v1 .n imbeds into u P(Λk Cn ) = PN . the collection of all ordinary tangent spaces to X)..g... and can locally be written as η = ηi1 . in the 1 0 . x ∂ ∂z n . then the complexiﬁed tangent space Tx X ⊗R C admits the basis ∂ ∂z 1 . q).72 5..e.jq (z. . the jacobian matrix of a transition function z = f (z) = α(x..ik vi1 ∧ · · · ∧ vik . COMPLEX MANIFOLDS AND VECTOR BUNDLES where the vi are a basis of tangent vectors to the given kplane. 2.q X are called diﬀerential forms of type (p. . Orientation.. . so that the transition functions ∂ are holomorphic and do not mix the vectors ∂z i with the ∂∂ i . u the numbers Pi1 . y) is (by the CauchyRiemann conditions) J= αx αy βx βy = αx αy −αy αx 2 2 so that det J = αx + αy > 0. z n ) is a set of local complex coordinates around a point x ∈ X. 1dimensional case) the coordinate change z → z . All complex manifolds are oriented. Let T X be the smooth tangent bundle (i.q X = Λp (T X)∗ ⊗ Λq (T X)∗ .. . As a consequence one has z ¯ a decomposition Λi T ∗ X ⊗ C = p+q=i Ωp.ik are the Pl¨cker coordinates on the Grassmann variety... x ∂ ∂z1 ¯ . considering (e. y) + iβ(x. .. Notice that we may always conjugate the complex structure. Some properties of complex manifolds 2.1. Let X be an ndimensional complex manifold. So Gk. vn } is ﬁxed in C n λ= i1 . Consider for simplicity the 1dimensional case. . where N = n − 1 (this is called the Pl¨cker embedding.ip . and the manifold is oriented.2.ik =1 Pi1 .. ¯ z z . x This yields a decomposition TX ⊗ C = T X ⊕ T X which is intrinsic because X has a complex structure. If (z 1 . z ) dz i1 ∧ · · · ∧ dz ip ∧ d¯j1 ∧ · · · ∧ d¯jq . q). x ∂ ∂zn ¯ .
1 is the sheaf of holomorphic pforms Ωp . ¯ Therefore.q disks in C). Cf.q denote the sheaf of forms of type (p. HOLOMORPHIC VECTOR BUNDLES 73 The compositions Ω o7 ∂ oooo ooo ooo p+1.• is a resolution of Ωp . .1 → 0 ∞ is exact. n (where n = dimC X) the sheaf sequence 0 → Ωp → Ωp. . Therefore. but requiring that all the maps involved are holomorphic. ∂ such that ¯ ¯ ¯ ∂ 2 = ∂ 2 = ∂ ∂ + ∂∂ = 0 ¯ (notice that the CauchyRiemann condition can be written as ∂f = 0). Then. . the pair (Ωp.q X Ωp. 3.14). one obtains the Dolbeault theorem: ¯ ∂ ¯ ∂ ¯ ∂ Proposition 5. and are called the Dolbeault cohomology groups of X. Let Ωp. e Proposition 5. ¯ Proof. i.q groups H∂ (X) and H q (X. . Holomorphic vector bundles on a complex manifold X are deﬁned in the same way than smooth complex vector bundles. −−→ Ωp. Let ∆ be a polycylinder in Cn (that is. Basic deﬁnitions. Holomorphic vector bundles 4. For all p. ¯ 4.3. the sheaves Ωp.1 −−→ . ∂) is for any p ≥ 0 a cohomology complex.q are ﬁne (they are CX modules). exactly as one proves the de Rham theorem (Theorem 3. [9]. . Then H∂ (∆) = 0 for q ≥ 1.0 → Ωp. the kernel of the morphism ∂ : Ωp. q ≥ 0. which is sometimes e ¯ called the ∂Poincar´ Lemma (or Dolbeault or Grothendieck Lemma). the cartesian product of p.2.q groups are denoted by H∂ (X).1. Its cohomology p.• (X). the cohomology p.e. Dolbeault cohomology Another interesting cohomology theory one can consider is the Dolbeault cohomology associated with a complex manifold X. for all p = 0.q X d / Λp+q+1 T ∗ X OOO OOO OO ¯ OOO' ∂ Ωp. The Dolbeault (or CauchyRiemann) operator ∂ : Ωp. ¯ We have for this theory an analogue of the Poincar´ Lemma.q+1 squares to zero. Moreover. . Let X be a complex manifold.q → Ωp. the Dolbeault complex of sheaves Ωp. Ωp ) are isomorphic.4.q+1 X ¯ deﬁne diﬀerential operators ∂.1. ¯ Moreover.0 −−→ Ωp. q) ¯ on X. .
x) .74 5. The collection {Uα . A complex manifold E is a rank n holomorphic vector bundle on X if there are (i) an open cover {Uα } of X (ii) a holomorphic map π : E → X (iii) holomorphic maps ψα : π −1 (Uα ) → Uα × Cn such that (i) π = pr1 ◦ψα . If f is a biholomorphism. the subset Ex = π −1 (x) ⊂ E is called the ﬁbre of E over x. Let E(Uα ) denote the set of sections of E over Uα . F over X is a holomorphic map f : E → F such that for every x ∈ X one has f (Ex ) ⊂ Fx . . the map −1 pr2 ◦ψβ ◦ ψα (p. ei ). i = 1. are sections of E over Uα . which is actually independent of the trivialization. COMPLEX MANIFOLDS AND VECTOR BUNDLES Definition 5. For every x ∈ X. .1. C) given by −1 gαβ (p) · x = pr2 ◦ψα ◦ ψβ (ψ. it is a free module over the ring O(Uα ) of holomorphic functions on Uα . and E and F are said to be isomorphic. it is said to be an isomorphism of vector bundles. the maps −1 s(α)i : Uα → E.. n where {ei } is the canonical basis of Cn . . where = pr1 is the projection onto the ﬁrst factor of Uα × Cn . and such that the resulting map fx : Ex → Fx is linear. A morphism between two vector bundles E.. ψα } is a trivialization of E. . and its subset {s(α)i }i=1.n is a basis. A holomorphic section of E over an open subset U ⊂ X is a holomorphic map s : U → E such that π ◦ s = Id. . These maps satisfy the cocycle condition gαβ gβγ gγα = Id on Uα ∩ Uβ ∩ Uγ .. By means of a trivialization around x.. Ex is given the structure of a vector space. On an intersection Uα ∩ Uβ one has the relation n s(α)i = k=1 (gαβ )ik s(β)k . •) : Cn → Cn is a linear isomorphism. (ii) for all p ∈ Uα ∩ Uβ . With the data that deﬁne a holomorphic vector bundle we may construct holomorphic maps gαβ : Uα ∩ Uβ → Gl(n. s(α)i (x) = ψα (x. With reference to the notation previously introduced. Vector bundles of rank 1 are called line bundles.
wn+1 ) be homogeneous coordinates in Pn . So if p has homogeneous coordinates (w0 .e. the locally deﬁned holomorphic vector ﬁelds ∂z 1 . If {Ui } is the standard cover of Pn . Example 5. Show that the rule that to any open subset U ⊂ X assigns the of sections of a holomorphic vector bundle E deﬁnes a sheaf E (which actually is a sheaf of OX modules). the tautological line bundle L of Pn . with the projection onto the ﬁrst factor. (In the next sections we shall learn some homological techniques that can be used to handle such situations). . HOLOMORPHIC VECTOR BUNDLES 75 Exercise 5. . . . Sections of E ∗ can be paired with (or act on) sections of E. . A holomorphic vector bundle is said to be trivial when it is isomorphic to Cn . and p ∈ Ui . so that its ﬁbre at p = π(u). The dual bundle H = L∗ acts on L. wn ). . then wi can be used to parametrize the points in the line p. . then the maps (5.3. We shall denote such a bundle by Cn (in particular. describe isomorphic bundles) if there exist holomorphic maps λα : Uα → Gl(n. with transition functions gαβ . u ∈ Cn+1 can be regarded as . .2. The space E = X × Cn . C denotes the trivial line bundle). . Example 5. called the dual vector bundle to E. the “holomorphic part” T X of the complexiﬁed tangent bundle is a holomorphic vector bundle. . is obviously a holomorphic vector bundle.4.3) gαβ = λα gαβ λ−1 β Exercise 5. we may deﬁne ψi : π −1 (Ui ) → Ui × C as ψi (u) = (p. Example 5. (The tangent and cotangent bundles) If X is a complex manifold. let us exhibit a trivialization for L and the related transition functions.5. Given a holomorphic atlas ∂ ∂ for X. To be more concrete. whose rank equals the complex dimension of X.4) T gαβ = (gαβ )−1 (where T denotes transposition) deﬁne another vector bundle. yielding holomorphic (smooth complexvalued) functions on (open sets of) X. Show that two trivializations are equivalent (i. called the trivial vector bundle of rank n. If to any p ∈ Pn (which is a line in Cn+1 ) we associate that line we obtain a line bundle. A holomorphic vector bundle may be trivial as a smooth bundle while not being trivial as a holomorphic bundle. (The tautological bundle) Let (w1 . The transition function is then gik = wi /wk . ∂z n provide a holomorphic trivialization of X.4. C) such that (5. The dual of T X is the holomorphic cotangent bundle of X.6. such that the transition functions of T X are the jacobian matrices of the transition functions of X. . wi ) if p = π(u). Every holomorphic vector bundle has an obvious structure of smooth complex vector bundle. and denoted by E ∗ . ∞ OX (U )module If E is a holomorphic (or smooth complex) vector bundle.
the global sections of H r are homogeneous polynomials of order r in the homogeneous coordinates. The determinant bundle of the holomorphic tangent bundle to a complex manifold is called the canonical bundle K. . (3) If E is a vector bundle with transition functions gαβ . wn+1 ) be homogeneous coordinates in Pn . a condition we may always meet. Show that the global sections of H can be identiﬁed with the linear polynomials in the homogeneous coordinates. their direct sum E1 ⊕ E2 is the vector bundle of rank r1 + r2 whose transition functions have the block matrix form (1) gαβ 0 0 gαβ (2) (2) We may also deﬁne the tensor product E1 ⊗ E2 . . Exercise 5. In the same way one deﬁnes a tautological bundle on the Grassmann variety Gk. i. . Additional operations that one can perform on vector bundles are again easily described in terms of transition functions.2. one writes Ln for L ⊗ · · · ⊗ L (n factors). Often L is denoted O(−1). Then E1 ⊗ E2 has local bases of sections {s(α)i ⊗ t(α)k } and the corresponding transition functions are given by r1 r2 s(α)i ⊗ t(α)k = m=1 n=1 (gαβ )im (gαβ )kn s(β)m ⊗ t(β)n . and similarly H n = O(n) (notice that O(−n)∗ = O(n)). we deﬁne its determinant det E as the line bundle whose transition functions are the functions det gαβ . 4. of rank r1 and r2 .9. as hyperplanes in Cn+1 . The tangent spaces to Pn are generated by . L) = n + 1. which has rank r1 r2 and has (1) (2) transition functions gαβ gαβ . one often writes Ln = O(−n). E1 and E2 have local bases of sections {s(α)i } and {t(α)k }. Exercise 5.8. This means the following: assume that E1 and E2 trivialize over the same cover {Uα }. . COMPLEX MANIFOLDS AND VECTOR BUNDLES the space of linear functionals on the line Cu ≡ Lp . Show that that the elements of a basis of the vector space of global sections of L can be identiﬁed homogeneous coordinates. so that dim H 0 (Pn . it has rank k. and H is denoted O(1) — the reason of this notation will be clear in Chapter 6. (1) (2) In particular the tensor product of line bundles is a line bundle.76 5. Let π : Cn+1 − {0} → Pn be the usual projection. Show that the canonical bundle of the projective space Pn is isomorphic to O(−n − 1).e. (1) Given two vector bundles E1 and E2 . and let (w1 .7. If L is the tautological line bundle on a projective space. and that in the given trivializations.n . Example 5. Hence H is called the hyperplane bundle. If L is a line bundle. More constructions. Hence.
σn+1 ) → i=1 σi (w) ∂ ∂wi (recall that the sections of H can be regarded as linear functionals on the homogeneous coordinates). . gαβ of the two bundles with respect to a cover {Uα } of X are 2cocycles O∗ . and satisfy gαβ = gαβ λα λβ with λα ∈ O∗ (Uα ). Its kernel is generated by the section σi (w) = wi . notice that this is the image of the map C → H ⊕(n+1) . The morphism H ⊕(n+1) → T Pn may be regarded as a sheaf morphism OPn (1)⊕(n+1) → T Pn . in particular one has L⊗L∗ C (think of it in terms of transition functions — here C denotes the trivial line bundle. If L L then the transition functions gαβ . the sheaf of germs of holomorphic vector ﬁelds on Pn . One can then deﬁne a map E : H ⊕(n+1) → T Pn n+1 is a linear functional on Cn+1 (σ1 . n + 1. .5.e. the second sheaf being the tangent sheaf of Pn . whose class [C] is the identity in Pic(X)). i = 1. The map E is apparently surjective. 5. CHERN CLASSES ∂ the vectors π∗ ∂wi . Chern classes of holomorphic line bundles. We deﬁne Pic(X) (the Picard group of X) as the set of holomorphic line bundles on X modulo isomorphism. so that the class [L∗ ] is the inverse in Pic(X) of the class [L]. Chern class of line bundles 5. Let O denote the sheaf of holomorphic functions on X. . . . Let X a complex manifold. .. . The group structure of Pic(X) is induced by the tensor product of line bundles L⊗L . i. . ∂wi If the vector ﬁeld ∂ v(w) = (w) ∂wi (i is ﬁxed) satisﬁes v(λ w) = λ v(w) and therefore descends to Pn . . wn+1 ) .1. 1 → (w1 . . . The long cohomology sequence exp 0 → Z → O −−→ O∗ → 0 . O∗ ). so that one has an identiﬁcation Pic(X) associated with the exact sequence H 1 (X. and these are subject to the relation n+1 77 wi π∗ i=1 ∂ = 0. and one has an exact sequence 0 → OPn → OPn (1)⊕(n+1) → T Pn → 0 called the Euler sequence. . and O∗ the subsheaf of nowhere vanishing holomorphic funtions.
Z) δ classiﬁes the line bundles having the same ﬁrst Chern class. If in addition dim X = 1 we have H 2 (X. so that every element in H 2 (X.1. the morphism δ is neither injective nor surjective.2. Exercise 5. Z) is the ﬁrst Chern class of a holomorphic line bundle. the group Pic0 (X) = ker δ H 1 (X. COMPLEX MANIFOLDS AND VECTOR BUNDLES (where exp f = e2πif ) contains the segment H 1 (X. Z) → H 0 (X. O∗ ) −−→ H 2 (X. then c1 (f ∗ L) = f (c1 (L)) . . From this one can easily prove that. Z). The image of c1 is a subgroup NS(X) of H 2 (X.2 From the deﬁnition of connecting morphism we can deduce an explicit formula for ˇ a Cech cocycle representing c1 (L) with respect to the cover {Uα }: {c1 (L)}αβγ = 1 2πi (log gαβ + log gβγ + log gγα ) . The fact that δ is a group morphism means that c1 (L ⊗ L ) = c1 (L) + c1 (L ) . Show that there exist nontrivial holomorphic line bundles which are trivial as smooth complex line bundles. 2Here we use the fact that if X is a complex manifold of dimension n. (ii) not every element in H 2 (X. Notice that when X is compact the sequence 0 → H 0 (X. and L is a line bundle on Y . so that (i) the ﬁrst Chern class does not classify the holomorphic line bundles on X. called the N´ronSeveri group of X. e Exercise 5. Z) is the ﬁrst Chern class1 of L. if f : X → Y is a holomorphic map. O) = 0 for all k > n. O) = 0. Z) → H 2 (X. O)/ Im H 1 (X. Z) → H 1 (X.78 5. Given a line bundle L. Z). O∗ ) → 0 is exact. In general. the element c1 (L) = δ([L]) ∈ H 2 (X. so that Pic0 (X) = H 1 (X. O) → H 1 (X. O) → H 0 (X. O) where δ is the connecting morphism. Show that all line bundles on Cn are trivial. 1This allows us also to deﬁne the ﬁrst Chern class of a vector bundle E of any rank by letting c1 (E) = c1 (det E). Z) is the ﬁrst Chern class of a holomorphic line bundle. then H k (X. O)/H 1 (X.
6. CHERN CLASSES
79
5.2. Smooth line bundles. The ﬁrst Chern class can equally well be deﬁned for smooth complex line bundles. In this case we consider the sheaf C of complexvalued smooth functions on a diﬀerentiable manifold X, and the subsheaf C ∗ of nowhere vanishing functions of such type. The set of isomorphism classes of smooth complex line bundles is identiﬁed with the cohomology group H 1 (X, C ∗ ). However now the sheaf C is acyclic, so that the obstruction morphism δ establishes an isomorphism H 1 (X, C ∗ ) H 2 (X, Z). The ﬁrst Chern class of a line bundle L is again deﬁned as c1 (L) = δ([L]), but now c1 (L) classiﬁes the bundle (i.e. L L if and only if c1 (L) = c1 (L )). Exercise 5.3. (A rather pedantic one, to be honest...) Show that if X is a complex manifold, and L is a holomorphic line bundle on it, the ﬁrst Chern classes of L regarded as a holomorphic or smooth complex line bundle coincide. (Hint: start from the inclusion O → C, write from it a diagram of exact sequences, and take it to cohomology ...) 6. Chern classes of vector bundles In this section we deﬁne higher Chern classes for complex vector bundles of any rank. Since the Chern classes of a vector bundle will depend only on its smooth structure, we may consider a smooth complex vector bundle E on a diﬀerentiable manifold X. We are already able to deﬁne the ﬁrst Chern class c1 (L) of a line bundle L, and we know that c1 (L) ∈ H 2 (X, Z). We proceed in two steps: (1) we ﬁrst deﬁne Chern classes of vector bundles that are direct sums of line bundles; (2) and then show that by means of an operation called cohomology base change we can always reduce the computation of Chern classes to the previous situation. Step 1. Let σi , i = 1 . . . k, denote the symmetric function of order i in k arguments.3. Since these functions are polynomials with integer coeﬃcients, they can be regarded as functions on the cohomology ring H • (X, Z). In particular, if α1 , . . . , αk are classes in H 2 (X, Z), we have σi (α1 , . . . , αk ) ∈ H 2i (X, Z). If E = L1 ⊕ · · · ⊕ Lk , where the Li ’s are line bundles, for i = 1...k we deﬁne the ith Chern class of E as ci (E) = σi (c1 (L1 ), . . . , c1 (Lk )) ∈ H 2i (X, Z) .
3The symmetric functions are deﬁned as
σi (x1 , . . . , xk ) = Thus, for instance,
X
1≤j1 <···<ji ≤n
xj1 · · · · · xji .
σ1 (x1 , . . . , xk ) = x1 + · · · + xk σ2 (x1 , . . . , xk ) = x1 x2 + x1 x3 + · · · + xk−1 xk ... σk (x1 , . . . , xk ) = x1 · · · · · xk . As a ﬁrst reference for symmetric functions see e.g. [21].
80
5. COMPLEX MANIFOLDS AND VECTOR BUNDLES
We also set c0 (E) = 1; identifying H 0 (X, Z) with Z (assuming that X is connected) we may think that c0 (E) ∈ H 0 (X, Z). Step 2 relies on the following result (sometimes called the splitting principle), which we do not prove here. Proposition 5.1. Let E be a complex vector bundle on a diﬀerentiable manifold X. There exists a diﬀerentiable map f : Y → X, where Y is a diﬀerentiable manifold, such that (1) the pullback bundle f ∗ E is a direct sum of line bundles; (2) the morphism f : H • (X, Z) → H • (Y, Z) is injective; (3) the Chern classes ci (f ∗ E) lie in the image of the morphism f . Definition 5.2. The ith Chern class ci (E) of E is the unique class in H 2i (X, Z) such that f (ci (E)) = ci (f ∗ E). We also deﬁne the total Chern class of E as
k
c(E) =
i=0
ci (E) ∈ H • (X, Z) .
The main property of the Chern classes are the following. (1) If two vector bundles on X are isomorphic, their Chern classes coincide. (2) Functoriality: if f : Y → X is a diﬀerentiable map of diﬀerentiable manifolds, and E is a complex vector bundle on X, then f (ci (E)) = ci (f ∗ E) . (3) Whitney product formula: if E, F are complex vector bundles on X, then c(E ⊕ F ) = c(E) ∪ c(F ) . (4) Normalization: identify the cohomology group H 2 (Pn , Z) with Z by identifying the class of the hyperplane H with 1 ∈ Z. Then c1 (H) = 1. These properties characterize uniquely the Chern classes (cf. e.g. [13]). Notice that, in view of the splitting principle, it is enough to prove the properties (1), (2), (3) when E and F are line bundles. Then (1) and (2) are already known, and (3) follows from elementary properties of the symmetric functions. The reader can easily check that all Chern classes (but for c0 , obviously) of a trivial vector bundle vanish. Thus, Chern classes in some sense measure the twisting of a bundle. It should be noted that, even in smooth case, Chern classes do not in general classify vector bundles, even as smooth bundles (i.e., generally speaking, c(E) = c(F ) does not imply E F ). However, in some speciﬁc instances this may happen. Exercise 5.3. Prove that for any vector bundle E one has c1 (E) = c1 (det E).
7. KODAIRASERRE DUALITY
81
7. KodairaSerre duality In this section we introduce KodairaSerre duality, which will be one of the main tools in our study of algebraic curves. To start with a simple situation, let us study the analogous result in de Rham theory. Let X be a diﬀerentiable manifold. Since the exterior product of two closed forms is a closed form, one can deﬁne a bilinear map
j i+j i HDR (X) ⊗ HDR (X) → HDR (X),
[τ ] ⊗ [ω] → [τ ∧ ω].
ˇ As we already know, via the Cechde Rham isomorphism this product can be identiﬁed with the cup product. If X is compact and oriented, by composition with the map4
X n ∫ : HDR (X) → R,
∫ [ω] =
X X
ω
where n = dim X, we obtain a pairing
n−i i HDR (X) ⊗ HDR (X) → R,
[τ ] ⊗ [ω] → ∫ [τ ∧ ω]
X
which is quite easily seen to be nondegenerate. Thus one has an isomorphism
i HDR (X)∗ n−i HDR (X)
(this is a form of Poincar´ duality). e If X is an ndimensional compact complex manifold, in the same way we obtain a nondegenerate pairing between Dolbeault cohomology groups (5.5) and a duality
p,q H∂ (X)∗ ¯ n−p,n−q (X). H∂ ¯ p,q n−p,n−q H∂ (X) ⊗ H∂ (X) → C, ¯ ¯
Exercise 5.1. (1) Let E be a holomorphic vector bundle on a complex manifold X, denote by E the sheaf of its holomorphic sections, and by E ∞ the sheaf of its smooth sections. Show (using a local trivialization and proving that the result is independent of the trivialization) that one can deﬁne a Clinear sheaf morphism (5.6) which obeys a Leibniz rule ¯ ¯ ¯ ∂E (f s) = f ∂E s + ∂f ⊗ s for s ∈ E ∞ (U ), f ∈ C ∞ (U ). ¯ (2) Show that ∂E deﬁnes an exact sequence of sheaves (5.7)
E E E 0 → Ωp ⊗ E → Ωp,0 ⊗ E ∞ −−→ Ωp,1 ⊗ E ∞ −−→ . . . −−→ Ωp,n ⊗ E ∞ → 0.
¯ ∂E : E ∞ → Ω0,1 ⊗ E ∞
¯ ∂
¯ ∂
¯ ∂
4This map is well deﬁned because diﬀerent representatives of [ω] diﬀer by an exact form, whose
integral over X vanishes.
Ωp ⊗ E). vector bundle on a diﬀerentiable manifold X. one deﬁnes Dolbeault cohomology groups with coeﬃcients p. and by Ω1 X the sheaf of diﬀerential 1forms on X.1. X X H n−p (X. By combining the pairing (5. E = ker(∂E : E ∞ → Ω0. and taking coholomology from the resulting (in general) nonexact sequence.82 5. E ∗ ) → C H∂ (X. COMPLEX MANIFOLDS AND VECTOR BUNDLES ¯ Here Ωp is the sheaf of holomorphic pforms. We shall denote by E the sheaf of sections of E. E). in general smooth. In particular. E)∗ This is usually called Serre duality. . 8. E (3) By taking global sections in (5. Ωn−q ⊗ E ∗ ).q H∂ (X. denoted H∂ (X.5) with the action of the sections of E ∗ on the sections of E we obtain a nondegenerate pairing n−p. Let E a complex. Connections In this section we give the basic deﬁnitions and sketch the main properties of connections.7). This is the KodairaSerre duality. ¯ Using the isomorphism (5.q in E.q H∂ (X.8) p. E ∗ ).q (X.8) we can express this duality in the form H p (X. Use the same argument as in the proof of de Rham’s theorem ¯ to prove an isomorphism (5. A connection is a sheaf morphism : E → Ω1 ⊗ E X satisfying a Leibniz rule (f s) = f (s) + df ⊗ s for every section s of E and every function f on X (or on an open subset). K ⊗ E ∗ ). The Leibniz rule also shows that is Clinear. Basic deﬁnitions. E)∗ ¯ n−p. 8. Ωq ⊗ E)∗ H n−p (X. In particular for q = 0 we get (denoting K = Ωn = det T ∗ X. The concept of connection provides the correct notion of diﬀerential operator to diﬀerentiate the sections of a vector bundle. E) ⊗ H∂ ¯ ¯ and therefore a duality p. thus getting a morphism X : Ωk ⊗ E → Ωk+1 ⊗ E . the canonical bundle of X) H p (X. E) ¯ H q (X.n−q H∂ (X.n−q p.1 ⊗ ∞ ). The connection can be made to act on all sheaves Ωk ⊗ E.
2 is an endomorphism of the bundle E with coeﬃcients in 2forms. sα = gαβ sβ ). CONNECTIONS 83 by letting (ω ⊗ s) = dω ⊗ s + (−1)k ω ⊗ (s). it satisﬁes the property (f s) = f (s) for every function f on X. respectively. the transformation properties of the connection 1forms are inhomogeneous and contain an aﬃne term. we may choose on any Uα a set {sα } of basis sections of E(Uα ) (notice that this is a set of r sections. Prove that if E and F are vector bundles. Exercise 5.. the transformation formula for the connection 1forms is (5. Every ωα is as an r × r matrix of 1forms. .2. Exercise 5.1. Prove that if gαβ denotes the transition functions of E with respect to the chosen local basis sections (i. i. the rule (τ ). The ωα ’s are called connection 1forms. The connection is not a tensorial morphism. Prove that is E is a vector bundle with a connection < ∗ .e. s >= d < τ. If {Uα } is a cover of X over which E trivializes. It is X called the curvature of the connection . then the rule (s ⊗ t) = 1 (s) 1 and ⊗t+s⊗ 2 (t) (minimal coupling) deﬁnes a connection on the bundle E ⊗ F (here s and t are sections of E and F .8. (s) > deﬁnes a connection on the dual bundle E ∗ (here τ . > denotes the pairing between sections of E ∗ and E). s > − < τ. In other terms. s are sections of E ∗ and E. to check that the square of the connection 2 : Ωk ⊗ E → Ωk+2 ⊗ E X X 2 2 is f linear. as a consequence. and we shall denote it by Θ.3. respectively).9) −1 −1 ωα = gαβ ωβ gαβ + dgαβ gαβ . a global section of the bundle Ω2 ⊗ End(E). Exercise 5. with connections 2 .e. and < . which we leave to the reader. It is an easy exercise. namely. with r = rk E). but rather satiﬁes a Leibniz rule. Over these bases the connection is locally represented by matrixvalued diﬀerential 1forms ωα : (sα ) = ωα ⊗ sα .. On local basis sections sα it is represented by the curvature 2forms Θα deﬁned by Θ(sα ) = Θα ⊗ sα .
Connections and holomorphic structures. then is a X X p. The same condition implies that the kernel of the map (5.• diﬀerential for the complex ΩX ⊗ E.10) Θα = dωα − ωα ∧ ωα . and on the dual of a bundle.1) parts.6.5. 8. X Proposition 5. In particular is a morphism Ωp.0 = ( )2 . A simple computation shows that locally valued 3forms Θ is represented by the matrix dΘα + ωα ∧ Θα − Θα ∧ ωα . given a vector bundle E with connection . Θ = Θ2. Due to the tensorial nature of the curvature morphism. the curvature 2forms obey a homogeneous transformation rule.2) parts. we can induce connections on a variety of bundles associated to given vector bundles with connections. we may diﬀerentiate its curvature as a section of Ω2 ⊗ End(E).1 + Θ0. Θ0.0).q ⊗ E → Ωp. according to the splitting Ω1 ⊗C = Ω1.4.0 + Θ1. X X X Analogously. the curvature splits into its (2. If Θ0. Θ = 0.q+1 ⊗ E. Exercise 5. Prove that the curvature 2forms may be expressed in terms of the connection 1forms by the equation (Cartan’s structure equation) (5. and thus diﬀerentiate their sections. Since we are able to induce connections on tensor products of vector bundles (and also on direct sums.2 = ( )2 . In particular. (1. we may split the latter into its (1.1 = ◦ + ◦ . Proof. (Bianchi identity) The covariant diﬀerential of the curvature of a connection is zero. and E a C ∞ complex vector bundle on it with a connection . Θ1.1 . COMPLEX MANIFOLDS AND VECTOR BUNDLES Exercise 5.1) and (0.84 5. The result of such a diﬀerentiation is called the covariant diﬀerential of the section. .11) : E → Ω0.0 ⊕Ω0. and .2.10) we obtain Θ = 0.2 = 0. in the obvious way). Obviously we have Θ2. If X is a complex manifold. without aﬃne term. By plugging in the structure equation (5.0) and (0.2 . Prove that the transformation formula for the curvature 2forms is −1 Θα = gαβ Θβ gαβ .1 ⊗ E X has enough sections to be the sheaf of sections of a holomorphic vector bundle.
Then equation (5.13) ωα = ∂hα h−1 ˜ α and this equations shows the uniqueness. [16]. h). t) + h(s.. On a local basis of sections {sα }. A pair (E. t of sections of E one has dh(s. The local basis is said to be unitary if the corresponding matrix h is the identity matrix. This equation implies right away that on a unitary frame.t.12) dhα = ωα hα + hα ωα ˜ ¯ where ˜ denotes transposition and ¯ denotes complex conjugation (but no transposition. In terms of connection forms and matrices representing h this condition reads (5. t) = h( s.8. Given a hermitian bundle (E. are Hermitian and positive deﬁnite. it is a smoothly varying assignation of Hermitian structures on the ﬁbres of E). one can easily check that the connection forms as deﬁned by equation (5. then the C ∞ vector bundle E admits a unique holomorphic structure. Proof.e.3. Conversely.1. If Θ0. 9. of E.r. . such that the corresponding sheaf of holomorphic sections is isomorphic to the kernel of the operator (5.13) satisfy the condition (5. Cf. a connection compatible with the holomorphic structure of E if = ∂E . when evaluated at any point of Uα . the connection forms are skewhermitian matrices. CONNECTIONS 85 Proposition 5.7. A Hermitian metric h of a complex vector bundle E is a global section of E ⊗ E ∗ which when restricted to the ﬁbres yields a Hermitian form on them (more informally.0). A connection on E is said to be metric with respect to h if for every pair s. This is metric w. If we use holomophic local bases of sections. h and compatible with the holomorphic structure of E by construction. h) formed by a holomorphic vector bundle with a hermitian metric is called a hermitian bundle. there is a unique connection on E which is metric with respect to h and is compatible with the holomorphic structure of E. Proposition 5. i.11). it is not the hermitian conjugation).2 = 0. if E is a holomophic vector bundle. the connection forms are of type (1. Hermitian bundles. Moreover.9) and therefore deﬁne a connection on E. on E is said to be 8. h is represented by matrices hα of functions on Uα which. As for the existence.8. Proof. p.11) concides with the operator δE deﬁned in Exercise 5. under this isomorphism the ¯ operator (5.12) yields (5. t) .
F can be regarded as a 2form on X. Let L be a complex line bundle on a smooth 2dimensional manifold X. endowed with a connection. as it must be. . and let F be the curvature of the connection. If we consider a static spherically symmetric magnetic ﬁeld in R3 . The fact that the Chern class of L as given by (5. Z) Z given by integration over X. and one has the equality i [F ] c1 (L) = 2π 2 in HDR (X).86 5.14) can take only integer values is known in physics as the quantization of the Dirac monopole. Let us discuss a simple situation. If we do not consider the dependence from the radius the vector potential deﬁnes a connection on a bundle L deﬁned on an S 2 which is spanned by the angular spherical coordinates. In this case the ChernWeil theorem states that i (5. C). Alternatively. COMPLEX MANIFOLDS AND VECTOR BUNDLES Example 5. we notice that the complexvalued form F is closed (Bianchi identity) and therefore singles out a class [F ] in the complexiﬁed de i 2 Rham group HDR (X) ⊗R C H 2 (X.14) c1 (L) = F 2π X where we regard c1 (L) as an integer number via the isomorphism H 2 (X. (Chern classes and Maxwell theory) The Chern classes of a complex vector bundle E can be calculated in terms of a connection on E via the socalled ChernWeil representation theorem. Notice that the Chern class of F is independent of the connection we have chosen. the class 2π [F ] is actually real. by solving the Maxwell equations we ﬁnd a solution which is singular at the origin.9.
pi ∈ S. The notion will be later generalized to higher dimensional manifolds. The set of all divisors of S forms an abelian group. and that it has a pole of order a if ordp f = −a < 0. There exists a unique nonnegative integer a and a holomorphic function h such that f (z) = (z − z(p))a h(z) and h(p) = 0. we deﬁne ordp f = ordp g − ordp h. We shall start with the onedimensional case. where “locally ﬁnite” means that every point p in S has a neighbourhood which contains only a ﬁnite number of pi ’s. with g and h holomorphic. We shall then write D ≥ 0. If S is compact. With each meromorphic function f we may associate the divisor (f ) = p∈S ordp f · p. this means that the number of points is ﬁnite. D= ai pi . denoted by Div(S). A divisor D on S is a locally ﬁnite formal linear combinations of points of S with integer coeﬃcients. We deﬁne ordp f = a. If f is a meromorphic function which in a neighbourhood of p can be written as f = g/h. Let f a holomorphic function deﬁned in a neighbourhood of p. then (f ) = (g) − (h). 1. 87 . We say that the divisor D is eﬀective if ai ≥ 0 for all i. and let z be a local coordinate around p.1) ordp f g = ordp f + ordp g. Notice that (6. ai ∈ Z. if f = g/h with g and h relatively prime.CHAPTER 6 Divisors Divisors are a powerful tool to study complex manifolds. Divisors on Riemann surfaces Let S be a Riemann surface (a complex manifold of dimension 1). We say that f has a zero of order a at p if ordp f = a > 0 (then f is holomorphic in a neighbourhood of p).
M∗ ) subject to the condition fα /fβ ∈ O∗ (Uα ∩ Uβ ). O∗ ) = 0 because Uα C holomorphically (here δ denotes the Cech cohomology operator). Then fα /fβ ∈ O∗ (Uα ∩ Uβ ). we may choose an open cover {Uα } such that each Uα contains at most one pi .1. given D = ai pi . We set fα = i ai giα . M∗ ) − − → C 0 (U. Proof.2. one has a commutative diagram of exact sequences 0 H 0 (S.1. There is a group isomorphism Div(S) H 0 (S. according to Proposition 6. and functions giα ∈ O(Uα ) such that that giα has a zero of order one at pi if pi ∈ Uα . The two constructions are one the inverse of the other. so that {fα } determines a global section of M∗ /O∗ . and the quantity ordp s is well deﬁned. 1. O∗ ) − − → C 1 (U. We set D = p ordp s · p. Correspondence between divisors and line bundles. The fact that this is also a group homomorphism follows from the formula (6.1. M∗ /O∗ ) ˇ can be represented by a 0cochain {fα ∈ M∗ (Uα )} ∈ C 0 (U. Sheaftheoretic description of divisors. which holds also for meromorphic functions. and let {Uα } be an open cover of S with meromorphic functions {fα } which deﬁne the divisor. O ∗ ) α =0 C 1 (U. Let D ∈ Div(S). Given a cover U = {Uα } of X. M∗ /O∗ ) − − → −− −− δ δ αH 1 (U . Let M∗ be the sheaf of meromorphic functions that are not identically zero.1). so that they establish an isomorphism of sets. M∗ /O∗ ). DIVISORS 1.88 6. We have an exact sequence 0 → O∗ → M∗ → M∗ /O∗ → 0 of sheaves of abelian groups (notice that the group structure is multiplicative). This diagram shows that a global section s ∈ H 0 (S. so that ordp fα does not depend on α. Proposition 6. Conversely. The group of divisors Div can be described in sheaftheoretic terms as follows. M∗ /O∗ ) −− −− ˇ where H 1 (Uα . M∗ ) − − → C 1 (U. Then the functions gαβ = fα ∈ O∗ (Uα ∩ Uβ ) fβ . M∗ /O∗ ) C 0 (U.
so that H 1 (S. If D = D(1) + D(2) then fα = fα fα by eq. [(f )] is the trivial line bundle.3. then ordpi fα = ordpi fα . and therefore determine a global nonzero meromorphic function. The line bundle [D] in independent. and one has a homomorphism Div(S) → Pic(S). up to isomorphism. so that one also has gαβ = fβ fβ fα . M∗ ). i. and deﬁne a line bundle. M). let D be a divisor such that [D] = C. gαβ = hβ Let {fα } be meromorphic functions which deﬁne D. M∗ /O∗ ) → H 1 (S.1. The line bundle [(f )] has transition functions gα hβ fα gαβ = = =1 gβ h α fβ (1) (2) ˇ (since f is a Cech cocycle) so that [(f )] = C. with gα . Conversely. O∗ ). so that [D(1) + D(2) ] = [D(1) ] ⊗ [D(2) ]. The line bundle associated with a divisor D is trivial if and only if D is the divisor of a global meromorphic function. Let f = {fα } ∈ let us set fα = gα /hα . which we denote by [D]. so that the functions hα = fα /fα are holomorphic and nowhere vanishing. Definition 6. of the set of functions deﬁning D. = fβ fβ hα hα so that the transition functions gαβ deﬁne an isomorphic line bundle.hα ∈ O(Uα ) relatively prime. = gαβ = hα hα hβ the quotients fα hα fα fβ . (6. We oﬀer now a sheaftheoretic description of this homomorphism. M∗ /O∗ ) and Pic(S) statement is equivalent to the exactness of the sequence H 0 (S. We have (f ) = (g) − (h). and gαβ = hβ fα fα hβ = gαβ . M∗ ) → H 0 (S. namely: Proposition 6. if {fα } is another set. D and D are linearly equivalent if and only if [D] there is an injective group homomorphism Div(S)/{linear equivalence} → Pic(S).e. In view of the identiﬁcations Div(S) H 0 (S. Quite evidently. DIVISORS ON RIEMANN SURFACES 89 obviously satisfy the cocycle condition. D ∈ Div(S) are linearly equivalent if D = D+(f ) for some f ∈ H 0 (S. then the transition functions of [D] have the form hα with hα ∈ O∗ (Uα ).2. with (g) and (h) eﬀective divisors. O∗ ) this .1). [D ]. H 0 (S. Two divisors D.
s ∈ H 0 (S. we have gαβ sβ sβ sα = = sα gαβ sβ sβ Let s ∈ H 0 (S. Then fα = gαβ fβ . Holomorphic and meromorphic sections of line bundles. and we may associate with s the divisor (s) = p∈S ordp s · p. and by M(L) the sheaf of its meromorphic sections.90 6. Obviously. Proof. s is holomorphic if and only if (s) is Proposition 6.e. Since [p] is eﬀective it has a global holomorphic section which vanishes only at p. A ﬁrst consequence of this is that. A line bundle L is associated with a divisor D (i. For the “only if” part.3.4. If L has transition functions gαβ with respect to a cover {Uα } of S. L is the line bundle associated with an eﬀective divisor if and only if it has a global nontrivial holomorphic section. By construction we have [(s)] eﬀective. L = [D] for some D ∈ Div(S)) if and only if it has a global nontrivial meromorphic section. we have sα = gαβ ∈ O∗ (Uα ∩ Uβ ) sβ so that ordp sα = ordp sβ for all p ∈ Uα ∩ Uβ . Of course it is trivial on U2 as well. the functions fα glue to yield a global meromorphic section s of L. let L = [D] with D a divisor with local equations fα = 0. So the same happens for the line bundles [kp]. where U1 = S − {p} and U2 is a neighbourhood of p. Proof. So we have L. if s. then a global holomorphic section s ∈ H 0 (S. the latter being deﬁned as M(L) = O(L) ⊗O M. k ∈ Z.5. on Uα ∩ Uβ . biholomorphic to a disc in C. The “if” part has already been proven. so that [p] is trivial on U1 . U2 }. the quantity ordp s is well deﬁned. . If L is a line bundle on S. If D is eﬀective the functions fα are holomorphic so that s is holomorphic as well. M(L)). O(L)) of L corresponds to a collection of functions {sα ∈ O(Uα )} such that sα = gαβ sβ on Uα ∩ Uβ . we denote by O(L) the sheaf of its holomorphic sections. so that the quotient of s and s is a welldeﬁned global meromorphic function on S. DIVISORS 1. The same holds for meromorphic sections. Corollary 6. The line bundle [p] trivializes over the cover {U1 . M(L)). where the functions gαβ are transition functions for L.
One has exact sequences (6. which we denote by Θ.2) 0 → R → C ∞ → Z 1 → 0. DIVISORS ON RIEMANN SURFACES 91 For the remainder of this section we assume that S is compact. If deg D < 0. then H 0 (S. Z 1 ) induces an isomorphism i 2 HDR (S) → H 1 (S. with θα = Notice that θα − θβ = so that d(θα − θβ ) = 0. Corollary 6. S ai . Since we may write Θ = 2π d∂ log hα a cocycle representing the image of [Θ] in H 1 (S. Before proving this result we need some preliminaries. Z 2 ) → H 1 (S. 0 → Z 1 → Ω1 → Z 2 → 0 . In terms of a local trivialization over an open cover {Uα } a hermitian metric is represented by nonvanishing real functions hα on i ¯ Uα . O(D)) = 0. i 2π (log hα − log hβ ) = d log gαβ . so that the 2form 2π ∂∂ log hα does not depend on α. If L is a line bundle we denote by S c1 (L) the number obtained by integrating over S ˇ a diﬀerential 2form which via de Rham isomorphism represents1 the Cech cohomology class c1 (L) regarded as an element in H 2 (S. We deﬁne a hermitian metric on a line bundle L as an assignment of a hermitian scalar product in each Lp which is C ∞ in p. On Uα ∩ Uβ one has hα = gαβ 2 hβ . and deﬁnes a global closed 2form on S. Proposition 6.8. i 2π ∂ i 2π ∂ log hα .) From the long exact cohomology sequences of the second sequence we get H 0 (S. Z 2 ) → H 1 (S. The class of Θ is the image in HDR (S) of c1 (L). Ω1 ) → H 0 (S. 2 Lemma 6. For any D ∈ Div(S) one has c1 (D) = deg D. 1The reader should check that the integral does not depend on the choice of the representative. thus a hermitian metric is a C ∞ section h of the line bundle L∗ ⊗ L∗ such that each h(p) is a hermitian scalar product in Lp . R).6. Proof. Z 1 ) → 0 so that the connecting morphism H 0 (S. (Here Ω1 is the sheaf of smooth realvalued 1forms. Z 1 ) is {θα − θβ }. We need the explicit form of the de Rham correspondence.1. Z 1 ). Let us deﬁne the degree of a divisor D = ai pi as the integer deg D = For simplicity we shall write O(D) for O([D]).7.
it is enough to consider the case D = [p]. and assuming that h1U2 −B( ) = z2 .7: Since c1 and deg are both group homomorphisms. .6. Corollary 6. Then c1 (D) = S S Θ= i d∂ 2π lim →0 S−B( ) log h1 where B( ) is the disc z < . which can always be arranged. then L = [D] with D = (s). R) → 0 so that the connecting morphism is now an isomorphism. If deg L < 0.10. Corollary 6. A global meromorphic function on a compact Riemann surface has the same number of zeroes and poles (both counted with their multiplicities).92 6. Proof. But f is a global meromorphic section of the trivial line bundle C.9. whence deg(f ) = S c1 (C) = 0 .2) we obtain from its long cohomology exact sequence a segment 0 → H 1 (S. this contradicts Corollary 6. Consider the open cover {U1 . Proof of Proposition 6. then H 0 (S. we 2 have dz 1 1 c1 (D) = 2πi lim ∂ log z z = 2πi ¯ =1 →0 ∂B( ) S ∂B( ) z having used Stokes’ theorem and the residue theorem (note a change of sign due to a reversal of the orientation of ∂B( )). U2 }. DIVISORS If we consider now the ﬁrst of the sequences (6. O(L)) = 0. Since ¯ ¯ ∂∂ = 1 d(∂ − ∂). with z a local coordinate around p. O(L)). we must show that deg(f ) = 0. and z(p) = 0. This result suggests to set deg L = S c1 (L) for all line bundles on S. Z 1 ) → H 2 (S. If f global meromorphic function. Proof. and U2 is a small patch around p. Since deg D < 0 by the previous Proposition. where U1 = S − {p}. If there is a nonzero s ∈ H 0 (S. If we apply it to the 1cocycle {θα − θβ } we get the 2cocycle of R 1 2πi log gαβ + 1 2πi log gβγ + 1 2πi log gγα = (c1 (L))αβγ .
let us denote E(D) = E ⊗ [D].1. This is quite obvious because by the local triviality of E the stalk of k O(E) at p is Op . We shall actually prove the exactness of the sequence (6.4) by O(E) we get 2Here we use the fact that tensoring all elements of an exact sequence by the sheaf of sections of a vector bundle preserves exactness. If E is a holomorphic vector bundle on S. Then by tensoring the exact sequence (6.3) is exact.4) follows from the fact the any local holomorphic function can be represented around pi in the form (Taylor polynomial) ai −1 f (z) = f (z0 ) + k=1 1 (k) f (z0 ) (z − z0 )k + (z − z0 )ai g(z) k! where z0 = z(p). the O(D) may be identiﬁed with the sheaf of meromorphic functions which at pi have a pole of order at most ai . . The sequence (6. Let D = ai pi be an eﬀective divisor. The term (z − z0 )ai g(z) is a section of O(−D). Then the line bundle L = [D] has at least one section s. The fundamental exact sequence of an eﬀective divisor. Let us ﬁrst deﬁne for all p ∈ S the sheaf kp as the 1dimensional skyscraper sheaf concentrated at p. / kp has stalk C at p and stalk 0 elsewhere. The exactness of the sequence (6.4. the sheaf kp (U ) = C if p ∈ U.11. Proof.2 Notice also that kD ⊗O O(D) kD because in a neighbourhood of every point pi the sheaf O(D) is isomorphic to O. with k the rank of E. The sheaf O(−D) can be regarded as the sheaf of holomorphic functions which at pi have a zero of order at least ai . this allows one to deﬁne a morphism O → O(D) by letting f → f sU for every f ∈ O(U ). In particular one may write ∗ 0 → O(−2p) → O → kp ⊕ Tp S → 0 ∗ where Tp S is considered as a skyscraper sheaf concentrated at p (indeed the quantity ∗ f (z0 ) determines an element in Tp S). while the ﬁrst two terms on the right single out a section of kD .4) 0 → O(−D) → O → kD → 0 0 → O → O(D) → kD → 0 from which the previous sequence is obtained by tensoring by O(D). Since O(D) O(−D)∗ . kp (U ) = 0 if p ∈ U. DIVISORS ON RIEMANN SURFACES 93 1. We also deﬁne the skyscraper sheaf kD = i (kpi )ai concentrated on D. namely. Proposition 6. and g is a holomorphic function.
. . R is a unique factorization domain if any element u can be written as a product u = u1 · · · . An analytic subvariety V is said to be reducible if V = V1 ∪ V2 with V1 and V2 properly contained in V . An analytic subvariety V of a complex manifold X is a subset of X which is locally deﬁned as the zero set of a ﬁnite collection of holomorphic functions. A point p ∈ V is a smooth point of V if around p the subvariety V is a submanifold. where the fi ’s are irreducible in Op . if Vi is the locus of zeroes of fi . fk (z 1 . If dim V = dim X − 1. um . . then V = ∪i Vi . and are deﬁned up to multiplication by invertible elements in Op . . We may now give the general deﬁnition of divisor: 3Let us recall this notion: one says that a ring R is an integral domain if uv = 0 implies that either u = 0 or v = 0. V is said to be irreducible if it is not reducible. 2. This follows from the fact that the stalk Op is a unique factorization domain ([9] page 12). and J is the jacobian matrix of the functions f1 . Definition 6. Proposition 6. . Divisors on higherdimensional manifolds We start with some preparatory material. In a neighbourhood of p the hypersurface V is given by f = 0. . where the ui are irreducible and unique up to multiplication by units. . .1. since it is not true that fj = gfi for some g ∈ Op which vanishes at p. . . . fk . the set Vs = V − V ∗ is the singular locus of V . we also have Vi ⊂ Vj .3 Let us sketch the proof for hypersurfaces. z n } is a local coordinate system for X around p. The set of smooth points of V is denoted by V ∗ . . . An element u ∈ R in an integral domain is said to be irreducible if u = vw implies that v or w is a unit. Denoting by the same letter the germ of f in p. Vi is irreducible as well. DIVISORS 0 → O(E(−D)) → O(E) → ED → 0 ⊕a where ED = ⊕i Epi i is a skyscraper sheaf concentrated on D.94 6. . Any analytic subvariety V can be expressed around a point p ∈ V as the union of a ﬁnite number of analytic subvarieties Vi which are irreducible around p. namely. . . z n ) = .2. . V will be called an analytic hypersurface. . Proof. The dimension of V is by deﬁnition the dimension of V ∗ . where {z 1 . since Op (where O is the sheaf of holomorphic functions on X) is a unique factorization domain we have f = f1 · · · · · fm . Since fi irreducible. it can be written as f1 (z 1 . z n ) = 0 with rank J = k. . . and are such that Vi ⊂ Vj .
3. LINEAR SYSTEMS
95
Definition 6.3. A divisor D on a complex manifold X is a locally ﬁnite formal linear combination with integer coeﬃcients D = ai Vi , where the Vi ’s are irreducible analytic hypersurfaces in X. If V ⊂ X is an analytic irreducible hypersurface, and p ∈ V , we may choose around p a coordinate system {w, z 2 , . . . , z n } such that V is given around p by w = 0. Given a function f deﬁned in a neighbourhood of p, let a be the greatest integer such that f (w, z 2 , . . . , z n ) = wa h(w, z 2 , . . . , z n ) with h(p) = 0. The function f has the same representation in all nearby points of V , so that a is constant on the connected components of V , namely, it is constant on V , so that we may deﬁne ordV f = a. With this proviso all the theory previously developed applies to this situation; the only deﬁnition which no longer applies is that of degree of a line bundle, in that c1 (L) is still represented by a 2form, while the quantities that can be integrated on X are the 2nforms if dimC X = n. Proposition 6.7 must now be reformulated as follows. Let D = ai Vi be a divisor, and let Vi∗ be the smooth locus of Vi . We then have: Proposition 6.4. For any divisor D ∈ Div(X) and any (2n − 2)form φ on X, c1 (D) ∧ φ =
X i
ai
φ.
Vi∗
Proof. The proof is basically the same as in Proposition 6.7 (cf. [9] page 141). 3. Linear systems In this section we consider a compact complex manifold X of arbitrary dimension. Let D = ai Vi ∈ Div(X), and deﬁne D as the set of all eﬀective divisors linearly equivalent to D. We start by showing that there is an isomorphism λ : PH 0 (X, O(D)) → D . We ﬁx a global meromorphic section s0 of [D], and set (6.5) s ∈ H 0 (X, O(D)) →
s s0 s s0
s s0
+ D ∈ D ;
s s0
one should notice that ordpi If s = α s with α ∈ C∗ then
≥ −ai if pi ∈ Vi so that =
s s0
+ D is indeed eﬀective.
so that equation (6.5) does deﬁne a map
PH 0 (X, O(D)) → D. This map is (i) injective because if λ(s1 ) = λ(s2 ) then s1 /s2 is a global nonvanishing holomorphic function, i.e. s1 = α s2 with α ∈ C∗ .
96
6. DIVISORS
(ii) Surjective because if D1 ∈ D then D1 = D + (f ) for a global meromorphic function f with ordpi (f ) ≥ −ai if pi ∈ Vi . So f s0 is a global holomorphic section of [D]. Definition 6.1. A linear system is the set of divisors corresponding to a linear subspace of PH 0 (X, O(D)). A linear system is said to be complete if it corresponds to the whole of PH 0 (X, O(D)). So a linear system is of the form E = {Dλ }λ∈Pm for some m. The number m is called the dimension of E. A onedimensional linear system is called a pencil, a twodimensional one a net, and a threedimensional one a web. Since all divisors in a linear system have the same degree, one can associate a degree to a linear system. Remark 6.2. If the elements λ0 , . . . , λm are independent in Pm (which means that they are images of linearly independent elements in Cm+1 ), and E = {Dλ }λ∈Pm is a linear system, then Dλ0 ∩ · · · ∩ Dλm =
λ∈Pm
Dλ .
For instance, if m = 1, and Dλ0 and Dλ1 have local equations f = 0 and g = 0, then Dλ has local equation c0 f + c1 g = 0 if λ = c0 λ0 + c1 λ1 . So Dλ0 ∩ Dλ1 ⊂ ∩λ∈P1 Dλ , which implies Dλ0 ∩ Dλ1 = ∩λ∈P1 Dλ . Definition 6.3. If E = {Dλ }λ∈Pm is a linear system, we deﬁne its base locus as B(E) = ∩λ∈Pm Dλ . Example 6.4. If E = {Dλ }λ∈P1 is a pencil, every p ∈ X − B(E) lies on a unique Dλ , so that there is a welldeﬁned map X − B(E) → P1 . This map is holomorphic. We may indeed write a local equation for Dλ in the form (6.6) f (z 1 , . . . , z n ) + λg(z 1 , . . . , z n ) = 0
where f and f are local deﬁning functions for D0 and D∞ (holomorphic because the divisors in E are eﬀective). f and g do not vanish simultaneously on X − B(E), so that they do not vanish separately either. Then the above map is given by λ = −f (z 1 , . . . , z n )/g(z 1 , . . . , z n ). Example 6.5. Since H 1 (Pn , O) = H 2 (Pn , O) = 0, the line bundles on Pn are classiﬁed by H 2 (Pn , Z) Z. Moreover, since c1 (H) = 1 under this identiﬁcation (i.e. deg H = 1), all divisors are linearly equivalent to multiples of H; in other terms, on Pn the only complete linear system of degree d is dH. Notice that H is basepoint free, i.e. B(H) = ∅. A fundamental result in the theory of linear systems is the following. Proposition 6.6. (Bertini’s theorem) The generic element of a linear system is smooth away from the base locus.
4. THE ADJUNCTION FORMULA
97
By this we mean that the set of divisors in a linear system E which have singular points outside the base locus form a subvariety of E of dimension strictly smaller than that of E. Proof. If E is linear system, and D ∈ E has singularities outside B(E), Bertini’s theorem would be violated by all pencils containing D. It is therefore suﬃcient to prove the theorem for pencils; in this case genericity means that the divisors having singularities out of the base locus are ﬁnite in number. So let E = {Dλ }λ∈P1 be a pencil, locally described by eq. (6.6), where the coordinates z n } can be deﬁned on an open subset ∆ ⊂ X whose image in Cn is a polydisc. Let pλ be a singular point of Dλ which is not contained in the base locus. We have the conditions {z 1 , . . . , (6.7) (6.8) f (pλ ) + λg(pλ ) = 0 ∂f ∂f (pλ ) + λ i (pλ ) = 0, ∂z i ∂z i = 1, . . . , n
f (pλ ), g(pλ ) = 0. We then have λ = −f (pλ )/g(pλ ), so that f ∂g ∂f − =0 ∂z i g ∂z i and (6.9) ∂ ∂z i f g =0 in pλ . in pλ ,
Let Y be the locus in ∆ × P1 cut out by the conditions (6.7) and (6.8); Y is an analytic variety, so the same holds true for its image V in ∆. Actually V is nothing but the locus of all singular points of the divisors Dλ . Equation (6.9) shows that f /g is constant on the connected components of V −B, that is, every connected component of V −B meets only one divisor of the pencil. Since the connected components of V − B are ﬁnitely many by Proposition 6.2, the divisors which have singularities outside B(E) are ﬁnite in number. 4. The adjunction formula If V is a smooth analytic hypersurface in a complex manifold X, we may relate the canonical bundles KV and KX . We shall denote by ιV : V → X the inclusion; one has an injective morphism T V → ι∗ T X of bundles on V . If we choose around p ∈ V V a coordinate system (z 1 , . . . , z n ) for X such that z 1 = 0 locally describes V , then the ∂ vector ﬁeld restricted to V locally generates the quotient sheaf NV = ι∗ T X/T V , V ∂z 1 so that NV is the sheaf of sections of a line bundle, which is called the normal bundle to V .
R) = 0. It is easily shown the V is smooth.1.11) we obtain the adjunction formula: (6. so that H 1 (V. We ﬁrst prove the isomorphism (6. as one can prove by using transition . the conormal bundle to V . and every line bundle is fully classiﬁed by its ﬁrst Chern class. then det F functions. and the isomorphism (6. the line bundle [V ] has transition functions gαβ = fα /fβ . On Uα ∩ Uβ we have dfα = d(gαβ fβ ) = dgαβ fβ + gαβ dfβ = gαβ dfβ the last equality holding on V ∩ Uα ∩ Uβ . The 1form ∗ dfα V ∩Uα is a section of NV V ∩Uα .10) with the isomorphism (6. since H 1 (P3 . which never vanishes because V is smooth. so that this bundle is trivial. Example 6. DIVISORS ∗ The dual NV . V Sometimes an additive notation is used. but rather to a global section of the line bundle ∗ NV ⊗ ι∗ [V ].11) ι∗ K X V ∗ KV ⊗ N V . 156) one has H 1 (V. the divisor V is locally given by functions fα ∈ O(Uα ). is trivial: if a line bundle L on V is such that c1 (L) = 0. OV ) = 0. Then the group Pic0 (V ). det E ⊗ det G. called the Fermat surface. Let V be the divisor cut out from P3 by the quartic equation (6. is the subbundle of ι∗ T ∗ X whose sections V are holomorphic 1forms which are zero on vectors tangent to V .13) 4 4 4 4 w0 + w1 + w2 + w3 = 0 where the wi ’s are homogeneous coordinates in P3 . then it is trivial. V By combining the formula (6. By a nontrivial result.98 6. (The same happens on P3 . and then the adjunction formula reads KV = KX V + [V ]V . If. relative to an open cover {Uα } of X. known as Lefschetz hyperplane theorem ([9] p. and it is of course compact: so it is a 2dimensional compact complex manifold. OP3 ) = 0).10) holds.10) ∗ NV ι∗ [−V ]. 4We use the fact that whenever 0→E→F →G→0 is an exact sequence of vector bundles. V We consider the exact sequence of vector bundles on V ∗ 0 → NV → ι∗ T ∗ X → T ∗ V → 0 V whence we get4 (6. This equation shows that the 1forms dfα ∗ do not glue to a global section of NV .12) KV ι∗ (KX ⊗ [V ]). which classiﬁes the line bundles whose ﬁrst Chern classes vanishes.
because such is the degree of the algebraic system formed by the equation (6. V ∼ 4H. We use the following fact: if D1 . D2 . 1 (V ) = 0. generates H 2 (P3 . Then [V ]V OV (4HV ). If we take D1 = V . THE ADJUNCTION FORMULA 99 We also know that KP3 = OP3 (−4H). D3 are V irreducible divisors in P3 .4. Since we also have HDR . D2 = D3 = H the number of intersection points is 4. we have c1 ([V ]) = 4h. Z). From the adjunction formula we get KV C: the canonical bundle of V is trivial. V is an example of a K3 surface. where H is any hyperplane in P3 . where HV = H ∩ V is a divisor in V . Therefore ι∗ KX OV (−4HV ). then we can move the divisors inside their linear equivalence classes in such a way that they intersects at a ﬁnite number of points.13) and by the equations of two (diﬀerent) hyperplanes. where h = c1 ([H]). V Let us compute c1 ([V ]V ) = ι∗ c1 ([V ]). that is. Since the class h. This number is computed by the integral c1 ([D1 ]) ∧ c1 ([D2 ]) ∧ c1 ([D3 ]) P3 where one considers the Chern classes c1 ([Di ]) as de Rham cohomology classes.
.
1. and let {s0 . the quantity (7.1) H 1 (S. given two complex manifolds X and Y . Consider now the exact sequence ∗ 0 → O(L − 2p) → O(L) −−→ Tp S ⊕ Lp → 0 dp ⊕evp (the morphism dp is Cartan’s diﬀerential followed by evaluation at p. .1) we get ∗ 0 → H 0 (S. We recall that. ι maps isomorphically Y onto a smooth subvariety of X. Any compact Riemann surface can be realized as a submanifold of PN for some N . . . and φ : LU → U × C is a local trivialization of L. Proposition 7. . ι) is a submanifold of X is ι is an injective holomorphic map Y → X whose diﬀerential ι∗p : Tp Y → Tι(p) X is of maximal rank (given by the dimension of Y ) at all p ∈ Y . we say that (Y.2) [φ ◦ s0 . If U is an open neighbourhood of p. O(L − 2p)−1 ⊗ K)∗ = 0 for any p ∈ S. . O(L)) −−→ Tp S ⊕ Lp → 0 dp ⊕evp so that dim D ≥ 1.e. they are algebraic curves). φ ◦ sN ] ∈ PN 101 . since deg(K − L + 2p) < 0 (here L − 2p = L ⊗ [−2p]). Proof. By Serre duality we have (7. . and let L = [D]). while evp is the evaluation of sections at p). and to study some of their basic properties. this is special instance of the socalled Kodaira’s embedding theorem. In other terms. Let N = dim D.CHAPTER 7 Algebraic curves I The main purpose of this chapter is to show that compact Riemann surfaces can be imbedded into projective space (i. . sN } be a basis of D. 1. The Kodaira embedding We start by showing that any compact Riemann surface can be embedded as a smooth subvariety in a projective space PN . . Pick up a line bundle L on S such that deg L > deg K + 2 (choose an eﬀective divisor D with enough points. Together with Chow’s Lemma this implies that every compact Riemann surface is algebraic. O(L − 2p)) H 0 (S. O(L − 2p)) → H 0 (S. Due to (7.
1 We must prove that (1) ιL is injective. H 1 (S. O(L)). O(L − p)) −−→ Tp S is surjective. This is equivalent to showing that the ∗ map H 0 (S. O(L − p − q)−1 ⊗ K)∗ = 0 because deg(L − p − q)−1 ⊗ K = deg K − deg L + 2 < 0. given any two points p. s(p) = 0) and (φ ◦ s)∗ is surjective at p.q rp. dp dp rp. O(L)) such that φ ◦ s(p) = 0 (i. and (2) the diﬀerential (ιL )∗ never vanishes. since O(L − p) is the sheaf of holomorphic sections of L vanishing at p. O(L − p)) −−→ Tp S is surjective. ALGEBRAIC CURVES I does not depend on the trivialization φ. A suﬃcient condition for a line bundle to be ample may be stated as follows (cf. the zero value) and have ˙ a ﬁrstorder contact (i. O(L)) −−→ Lp ⊕ Lq . O(L − p − q)) H 0 (S. O(L − 2p))∗ dp H 0 (S. O(L − p − q)) = 0 since H 1 (S. q ∈ S. A (1.e. Given any complex manifold X. however. The cotangent space Tp S can be realized as the space of equivalence classes of holomorphic functions which have the same value at p (e. we have therefore established a (holomorphic) map ιL : S → PN . this in turn implied by the surjectivity of the map H 0 (S.2.q rp. Definition 7.102 7. A line bundle L is said to be ample if Ln is very ample for some natural n. [9]). one says that a line bundle L on X is very ample if the construction (7. they have the same diﬀerential at p). O(L − p − q)) −−→ Lp ⊕ Lq → H 1 (S. O(−L + 2p + K)) = 0 ∗ so that H 0 (S. diﬀerent choices correspond to an action of the group PGl(N + 1. there is a section s ∈ H 0 (S. To show this we start from the exact sequence 0 → O(L − p − q) → O(L) −−→ Lp ⊕ Lq → 0 and note that in coholomology we have H 0 (S. We consider the exact sheaf sequence ∗ 0 → O(L − 2p) → O(L − p) −−→ Tp S → 0. C) on PN and therefore produce isomorphic subvarieties of PN . Let φ be a trivializing map for L around p. (1) It is enough to prove that.e.g. . we must ﬁnd a section s ∈ H 0 (S.2) deﬁnes an imbedding of X into PH 0 (X. O(L)) such that s(p) = λs(q) for all λ ∈ C∗ .q s → s(p) + s(q). ∗ (2) We shall actually show that the adjoint map (ιL )∗ : Tι∗ (p) PN → Tp S is surjectL ∗ ive. by Serre duality.1) form ω on a complex manifold is said to be positive if it can be locally written in the form ω = i ωij dz i ∧ d¯j z 1This map actually depends on the choice of a basis of D.
A ﬁrst consequence of the imbedding theorem expressed by Proposition 7. Proof. and since V is eﬀective. OM (L + mHM )) = 0. i. We shall show that for a big enough integer m the line bundle L + mHM m (= L ⊗ HM ) has a global holomorphic section s. if t is a holomorphic section of HM .3) is surjective. Therefore the morphism r in (7. [9]. If the ﬁrst Chern class of a line bundle L on a complex manifold can be represented by a positive 2form. M is the imbedding of a compact Riemann surface into Pn ). s (Here −−→ denotes the morphism given by multiplication by s). We have an exact sequence 0 → OM (−HM ) −−→ OM → kV → 0 so that after tensoring by L + mHM . OM (L + (m − 1)HM )) where N = deg V . We must ﬁnd a global meromorphic section of L.1 is that any line bundle on a compact Riemann surface comes from a divisor. there is a divisor D on M such that L = [D]. deg KM ⊗ O(−L − (m − 1)HM ) < 0). We shall now proceed to identify compact Riemann surfaces with (smooth) algebraic curves.4.3. Given a homogeneous polynomial F on Cn+1 the zero locus of F in Pn is by deﬁnition the projection to Pn of the zero locus of F in Cn+1 . and L is a line bundle on M . then L is ample.e. and H 0 (M. the required meromorphic section of L is s/tm . (7. . and let V be the intersection of M with a hyperplane in Pn (so [V ] HM .1. Proposition 7. While we have seen that any compact Riemann surface carries plenty of very ample line bundles. Let HM be the restriction to M of the hyperplane bundle H of Pn . this in general is not the case: there are indeed complex manifolds which cannot be imbedded into any projective space. If M is a smooth 1dimensional 2 analytic submanifold of projective space Pn (i. KM ⊗ O(−L − (m − 1)HM ))∗ = 0 r by Serre duality and the vanishing theorem (if m is big enough. HM has global holomorphic sections).3) s s 0 → OM (L + (m − 1)HM ) −−→ OM (L + mHM ) → kV → 0. cf. OM (L + mHM )) −−→ CN → H 1 (M. 2This result is actually true whatever is the dimension of M . Proposition 7. Div(S)/linear equivalence Pic(S). But H 1 (M. OM (L + (m − 1)HM )) H 0 (M. The associated long cohomology exact sequence contains the segment H 0 (M. THE KODAIRA EMBEDDING 103 with ωij a positive deﬁnite hermitian matrix.e.
The following fundamental result. the choice of an ample line bundle.104 7.7. then hi (E) = dim H i (C. A (projective) algebraic variety is a subvariety of Pn which is the zero locus of a ﬁnite collection of homogeneous polynomials. RiemannRoch theorem A fundamental result in the study of algebraic curves in the RiemannRoch theorem. A onedimensional algebraic variety is called an algebraic curve. We shall say that an algebraic variety is smooth if it is so as an analytic subvariety of Pn . Exercise 7. it is not hard to prove. understanding that we shall only consider smooth algebraic curves.6. We switch from the terminology “compact Riemann surface” to “algebraic curve”. then it also holds for L = [D + p] and L = [D − p].4 We denote g = h0 (K). In the ﬁrst case we start from the exact sequence 0 → O(D) → O(D + p) → kp → 0 3Strictly speaking an algebraic curve consists of more data than a compact Riemann surface S. the result holds obviously (notice that H 0 (C. (Chow’s lemma) Any analytic subvariety of Pn is algebraic. O)∗ since the former requires an imbedding of S into a projective space. Any compact Riemann surface is a smooth algebraic curve.3 We shall usually denote an algebraic curve by the letter C. Exploiting the fact that L = [D] for some divisor D. Using Chow’s lemma together with the imbedding theorem (Proposition 7. Proof. we shall anyway omit its proof for the sake of brevity (cf. and call it the arithmetic genus of C (this number will be shortly identiﬁed with the topological genus of C). [9] page 167). Use Chow’s lemma to show that H 0 (Pn .e. it is enough to prove that if the results hold for L = [D]. The dimension of an algebraic variety is its dimension as an analytic subvariety of Pn . ALGEBRAIC CURVES I Definition 7. H 1 (C. (RiemannRoch theorem) For any line bundle L on C one has h0 (L) − h1 (L) = deg L − g + 1. i. and denote by K its canonical bundle.8. C . 2. Proposition 7. H d ) — where H is the hyperplane line bundle — can be identiﬁed with the space of homogeneous polynomials of degree d on Cn+1 .1) we obtain Corollary 7. 4We introduce the following notation: if E is a sheaf of O modules. Let C be an algebraic curve. If L = C is the trivial line bundle. called Chow’s lemma. Proposition 7. E).5.1. K) by Serre duality).
which states that the integral of the Euler class of the real tangent bundle to S is the Euler characteristic of S. and since C is compact.1. such that C ω = 1. where ω is a smooth 2form on C such that deg f ∗ (p) = C C ω = 1. To this end we may use the GaussBonnet theorem. 5This need not be true if the algebraic curve C is singular. 3. 3. By using the RiemannRoch theorem and Serre duality we may compute the degree of K. Z) Z. Z). Z). and actually provides a basis of that space. O(D)) → H 0 (S. O(D)) → H 1 (S. allowing an identiﬁcation H 2 (C. However the RiemannRoch theorem is still true (provided we know what a line bundle on a singular curve is!) with g the arithmetic genus. C c1 (f ∗ [p]) = C f c1 ([p]) = n so that the map f takes the value p exactly n times. obtaining deg K = 2g − 2. If f : C → C is a nonconstant holomorphic map between two algebraic curves. Analogously for L . If p ∈ C we have c1 ([p]) = n. the de Rham cohomology class [ω] may be regarded as an element in H 2 (C. including multiplicities in the sense of divisors. On the other hand the complex structure of C makes the real tangent bundle into a complex holomorphic line bundle. .5 g = gtop . then f [ω] is a nonzero element in H 2 (C . Some general results about algebraic curves Let us ﬁx some notations and give some deﬁnitions. we may say that f covers C n times. Therefore we get deg K = 2gtop − 2. isomorphic to the holomorphic tangent bundle T C. and there is a well deﬁned integer n such that f [ω] = n[ω ]. Let C be an algebraic curve. the number of points in f −1 (p) is always ﬁnite. This is called the RiemannHurwitz formula. namely.6 The integer n is called the degree if f . GENERAL RESULTS 105 which gives (since H 1 (C. and under this identiﬁcation the Euler class corresponds to the ﬁrst Chern class of T C. It allows us to identify g with the topological genus gtop of C regarded as a compact oriented 2dimensional real manifold S. O(D + p)) → 0 whence h0 (L ) − h1 (L ) = h0 (L) − h1 (L) + 1 = deg L − g + 2 = deg L − g + 1. χ(S) = 2 − 2gtop . O(D + p)) → C → H 1 (S.3. kp ) = 0) 0 → H 0 (S. The degree of a map. 6Since two holomorphic functions of one variable which agree on a nondiscrete set are identical. and ω a smooth 2form on C.
q∈f −1 (p) deg f ∗ (p) = From these formulae we may draw the following picture. we may ﬁnd a coordinate z around any q ∈ C and a coordinate w around f (q) such that locally f is described as (7. h(w) g(z r ) r g(z r ) dz = rz r−1 dz h(z r ) h(z r ) (r(p) − 1) · p. so that KC = f ∗ KC + B . Let η be a meromorphic 1form on C. .106 7. The branch locus of f is the divisor in C B = q∈C (r(q) − 1) · q or its image in C B= q∈C (r(q) − 1) · f (q). There is a relation between the canonical divisors of C and C and the branch locus. 7A (holomorphic) covering map f : X → Y . On the other hand the divisor (η) is just the canonical divisor of C. For any p ∈ C we have f ∗ (p) = q∈f −1 (p) r(q) · q r(q) = n. Given again a nonconstant holomorphic map f : C → C. which means that f : C − B → C − B is a covering map.4) w = zr . and p = f (q) is said to be a branch point if r(p) > 1. The number r −1 is called the ramiﬁcation index of f at q (or at p = f (q)). at p exactly r sheets of the covering join together.7 It p ∈ C is a branch point of ramiﬁcation index r − 1. This implies the relation between divisors (f ∗ η) = f ∗ (η) + p∈C g(w) dw. Branch points. with X connected. which can locally be written as η= From (7. is a map such that each p ∈ Y has a connected neighbourhood U such that f −1 (U ) = ∪α Uα is the disjoint union of open subsets of X which are biholomorphic to U via f .2. then exactly n distinct points of C are mapped to f (p).4) we get f ∗η = so that ordp f ∗ η = ordf (p) η + r − 1. If p ∈ C does not lie in the branch locus. ALGEBRAIC CURVES I 3.
As a divisor. if D is eﬀective. KC ) is injective. KC ) → H 0 (C .. i.) Both equation (7. Write an explicit formula for f ∗ ω. The fact is that any kdimensional analytic subvariety V of an ndimensional complex manifold X determines a homology class [V ] in the homology group H2k (X. and this is given by the degree of the polynomial cutting D).3. the number of solutions is d. by using the RiemannHurwitz formula we obtain (7. so that KC = ι∗ (KP2 + C). they meet at a ﬁnite number of points. For a generic choice of the hyperplane. associates the integer number [V ] ∩ [W ] with the two subvarieties. In our case the number of intersection points is given by the number of solutions to an algebraic system.) We may prove this equation by using the adjunction formula: C is imbedded into P2 as a smooth analytic hypersurface. 8 We want to show that for smooth plane curves the following relation between genus and degree holds: (7. C is linearly equivalent to dH (indeed. By taking degree we get deg KC = n deg KC + p∈C (r(p) − 1). then the the number [V ] ∩ [W ] counts the intersection points (cf. Assume that X is compact. the homology cap product H2k (X.6) 1 g(C) = 2 (d − 1)(d − 2). Its image in P2 is the zero locus of a homogeneous polynomial. Recalling that KP2 = −3H we then have KC = (d − 3)ι∗ H. Z) → Z.5) g(C ) = n(g(C) − 1) + 1 + 1 2 p∈C (r(p) − 1) . GENERAL RESULTS 107 From this formula we may draw an interesting result. An algebraic curve C is said to be plane if it can be imbedded into P2 . any divisor D on P2 is linearly equivalent to mH for some m. then f : H 0 (C. KC ) is a global holomorphic 1form on C which is diﬀerent from zero at all points in an open dense subset of C.5) and the previous Exercise imply g(C ) ≥ g(C). Prove that if f : C → C is nonconstant. m is the number of intersection points between D and a generic hyperplane in P2 . The genus formula for plane curves. Exercise 7.3. [9] page 49). One may pick up diﬀerent representatives V and W of [V ] and [W ] such that V and W meet transversally.. Z). (Hint: a nonzero element ω ∈ H 0 (C. . given by the equation of C in P2 (which has degree d) and the linear equation of a hyperplane.1. and let W be an (n − k)dimensional analytic subvariety of X. 8We are actually using here a piece of intersection theory. (For singular plane curves this formula must be modiﬁed. which is dual to the cup product in cohomology. since Pic(P2 ) Z.e. the degree d of this polynomial is by deﬁnition the degree of C.. Z) ∩ H2n−2k (X. where ι : C → P2 . 3.
By writing this equation in homogeneous coordinates one obtain a curve in P2 which is a double covering of P1 branched at 6 points.6). A meromorphic 1form on an algebraic curve C is a meromorphic section of the canonical bundle K. and denote it by Resp (ϕ). Let a be coeﬃcient of the z −1 term in the Laurent expansion of f around p. . Thus the formula (7. a meromorphic 1form ϕ is locally written around p in the form ϕ = f dz. Proof. ALGEBRAIC CURVES I To carry on the computation. Consider the aﬃne curve in C2 having equation y 2 = x6 − 1 .4.2. Given a meromorphic 1form ϕ its polar divisor is D = i pi .6). so that deg ι∗ H = d.108 7. where the pi ’s are the points where the local representatives of ϕ have poles of order 1. i pi be the polar divisor of a meromorphic 1form ϕ. and let B a small disc around p. By the RiemannHurwitz formula we may compute the genus. This happens because the curve is singular at the point at inﬁnity. We call it the residue of ϕ at p. by the Cauchy formula we have a= ∂B ϕ so that the number a does not depend on the representation of ϕ. and a local holomorphic coordinate z such that z(p) = 0. where f is a meromorphic function. Example 7. Proposition 7. Let D = Then i Respi (ϕ) = 0. Then Respi (ϕ) = i ∂∪i Bi ϕ=− C−∪i Bi dϕ = 0 . ι∗ H = C ∩ H. fails in this case.3. and deg KC = d(d − 3) = 2g − 2 whence the formula (7. Choose a small disc Bi around each point pi . obtaining g = 2. The residue formula. which would yield g = 10. we notice that. Given a point p ∈ C. 3. as a divisor on C.
so that it is an isomorphism. Of course s vanishes at some other point s0 . Lemma 7. since the Jacobian determinant of f is not everywhere zero. Let f : X → Y be such a map. The g = 0 case. which is not possible. then then function f − α has two zeroes and only one simple pole. On the other hand. this map takes the value ∞ only once.4. by taking the cohomology exact sequence associated with the sequence 0 → O → O(p) → kp → 0 we obtain the existence of a global section s of [p] which does not vanish at p. Let ω be a volume form on Y . we have X f ∗ ω > 0. R) = 0 (prove it by using a MayerVietoris argument). 9Otherwise one can directly identify the sections of L with meromorphic functions having (only) a single pole at p.9 By considering ∞ as the value of f at p. with a simple pole at p and a zero at p0 . z where g is a holomorphic function. Any holomorphic map between compact complex manifolds of the same dimension whose Jacobian determinant is not everywhere zero is surjective. and where it is not zero is positive. (Here z is a local complex coordinate such that z(p) = 0. the line bundle [p] is trivial on C − {p}. and has a holomorphic section s0 which is nonzero on C − {p} and has a simple zero at p (this means of course that (s0 ) = p). since such functions can be developed around p in the form a f (z) = + g(z) .5. Assume q = Im f . Pick a point p ∈ C. But then f ∗ω = X ∂X f ∗ η = 0. a contradiction. and let n = dim X = dim Y . we may think of f as a holomorphic nonconstant map f : C → P1 . since by Serre duality h1 (O) = h0 (K) = 0.3. Thus f is injective. we have ω = dη on Y − {q}. GENERAL RESULTS 109 3. The following Lemma implies that f is surjective as well. Since H 2n (Y − {q}.) . Suppose that f takes the same value α at two distinct points of C. Then the quotient f = s/s0 is a global meromorphic function. Proof. a ∈ C should be indentiﬁed with the projection of f onto kp . We shall now show that all algebraic curves of genus zero are isomorphic to the Riemann sphere P1 .
.
has a structure of complex torus of dimension g. we have dω = 0 and ω ∧ ω = 0. since it can be represented as H 1 (C. We shall call T a complex torus. If ω in abelian diﬀerential.1) Λ= i=1 ni vi .e. T is a compact abelian complex Lie group.. Z) is a lattice in H 1 (C. . O)/H 1 (C.e. Consider now a smooth algebraic curve C of genus g ≥ 1.1. and one of abelian group. and the desingularization of nodal plane singular curves (this will involve the introduction of the notion of blowup of a complex surface at a point).CHAPTER 8 Algebraic curves II In this chapter we further study the geometry of algebraic curves. ni ∈ Z where {vi }i=1. O). the global holomorphic 1forms). 1. Example 8. This is the Jacobian variety of C. some theory of elliptic curves. Z). Topics covered include the Jacobian variety of an algebraic curve. and think of it as an abelian group. even though the two tori are obviously diﬀeomorphic as real manifolds. and that (8. and the two structures are compatible.. this means that ω singles out a cohomology class [ω] in H 1 (C. which we proceed now to deﬁne. If C is an algebraic curve of genus g. classifying the line bundles on C with vanishing ﬁrst Chern class. A lattice Λ in V is a subgroup of V of the form 2m (8. C).2) C 111 ω ∧ ω = 0. i.. Let V be an mdimensional complex vector space. and H 1 (C.. the group Pic0 (C). The quotient space T = V /Λ has a natural structure of complex manifold. In what follows we shall construct this variety in a more explicit way. The Jacobian variety A fundamental tool for the study of an algebraic curve C is its Jacobian variety J(C).2m is a basis of V as a real vector space. Notice that by varying the lattice Λ one gets another complex torus which may not be isomorphic to the previous one (the complex structure may be diﬀerent). We shall call abelian diﬀerentials the global sections of K (i.
Intrinsically. ALGEBRAIC CURVES II Moreover. for any abelian diﬀerentials e ω. This is called the period matrix. and † hermitian conjugation). that is. Z). Z) → H 0 (C. . . the number γ ω depends only on the homology class of γ and the cohomology class of ω. C). A way to check that the construction of the Jacobi variety does not depend on the choices we have made is to restate it invariantly. [τ ] > . i([γ])(ω) = γ ω. So the columns of the period matrix generate a lattice Λ in Cg . where the γi ’s are smooth loops in C. e . If γ is a smooth loop in C. . and expresses the pairing < . Z)). Exercise 8. We associate with these data the g × 2g matrix Ω whose entries are the numbers Ωij = γj ωi . λj = 0. Z). [γ2g ]} of the 2gdimensional free Zmodule H1 (C. τ we have [ω] ∪ [τ ] =< Q[ω]. Its columns Ωj are linearly independent over R: if for all i = 1. since locally ω = f (z) dz. . K)∗ . This map is injective: if i([γ])(ω) = 0 for a given γ and all ω then γ is homologous to the constant loop. K)∗ /H1 (C.2). . and ω ∈ H 0 (C. (8. show that Serre and Poincar´ dualities establish an isomorphism J(C) Pic0 (C). > between the Poincar´ dual spaces H1 (C. ¯ Since {ωi . K). Notice that Q is antisymmetric. Z). e Pick up a basis {[γ1 ]. i Ω Q Ω† > 0 (here ˜ denotes transposition. g 2g 2g 0= j=1 λj Ωij = j=1 λj γj ωi then also 2g j=1 λj [γj ] 2g j=1 λj γj ωi = 0. ωg } of H 0 (C. Deﬁne now the intersection matrix Q by letting Q−1 = [γj ] ∩ [γi ] (this is the Zij valued “cap” or “intersection” product in homology). H1 (C. . this implies ¯ = 0. Then we have the representation J(C) = H 0 (C.3) i C ω∧ω >0 ¯ if ω = 0. The quotient complex torus J(C) = Cg /Λ is the Jacobian variety of C. In this form they are called Riemann bilinear relations.112 8. K)∗ /H1 (C. Z) ⊗Z C and H 1 (C. Q is an element in HomZ (H 1 (C. and a basis {ω1 . By regarding J(C) as H 0 (C. Since the cup product in cohomology is Poincar´ dual to the cap product in homology. Z).3) can then be written in the form ˜ Ω Q Ω = 0. K). we have (8. . The relations (8. Integration over cycles deﬁnes a map i : H1 (C. . . C) = H1 (C. C). .2. . ωi } is a basis for H 1 (C.
the oriented sum δ1 − δ2 will deﬁne a cycle in homology. . . and let Picd (C) be the set of line bundles of degree d. . D ∈ Div(C) are linearly equivalent if and only if µ(D) = µ(D ). Proof.1. The Abel map µ : C → J(C) is injective. J(C) 1Notice that Picd (C) Picd (C) as sets for all values of d and d . . O∗ ).3.5) by letting µ(D) = i µ : Div(C) → J(C) µ(pi ) − j µ(qj ) if D= i pi − j qj .1.4) 1. Proof. All of this depends on the choice of the base point p0 . The Abel map. (Abel’s theorem) Two divisors D. If µ(p) = µ(q) by the previous Proposition p ∼ q as divisors. however. note however that if deg D = 0 then the choice of p0 is immaterial. Actually the value of µ(p) in Cg will depend on the choice of the path from p0 to p. if δ1 and δ2 are two paths.4) we may get a group homomorphism (8. but since g(C) ≥ 1 this implies p = q (this follows from considerations analogous to those in subsection 7. p0 ωg . . and µ(p) is a welldeﬁned point in J(C). Abel’s theorem may be stated in a fancier language as follows. the two values will diﬀer by an element in the lattice. . Then µ ﬁlters through a morphism ν : Picd (C) → J(C). From (8. and one has a commutative diagram Divd (C) / Picd (C) KKK KKK ν K µ KKK % . M∗ )/H 0 (C.5). K) (8. .4.1 One has a surjective map : = Divd (C) → Picd (C) whose kernel is isomorphic to H 0 (C. .3. For a proof see [9] page 232. Corollary 8. ωg } in we deﬁne a map µ : C → J(C) p p by letting µ(p) = p0 ω1 . After ﬁxing a point p0 in C and a basis {ω1 . . Let Divd (C) be the subset of Div(C) formed by the divisors of degree d. Proposition 8. THE JACOBIAN VARIETY 113 H 0 (C.
z d } on C d . Prove that Symd (P1 ) homogeneous coordinates. then {z 1 . set L = (D) (i. . L = [D]). . ALGEBRAIC CURVES II moreover.. We next choose p2 so that ω2 (p2 ) = 0. . . . ωg (p1 ) . So the latter is holomorphic. (p1 . . we call the quotient Symd (C) = C d /Sd the dfold symmetric product of C. the holomorphic map C d → J(C). ω1 (pg ) . . and the elementary symmetric functions of the coordinates z i yield a local coordinate system for Symd (C). that is. and so on. the morphism ν is injective (if ν(L) = 0. With these choices of the abelian diﬀerentials ωi and of . usually called Jacobi inversion theorem. Proof. ωg (pg ) We may choose p1 so that ω1 (p1 ) = 0. .6) ∂ ∂ (µg (D ))j = i ∂z i ∂z pi ωj = hji p0 where hji is the component of ωj on dz i . . .. . . ωg we may arrange that ω2 (p1 ) = · · · = ωg (p1 ) = 0. z i (p1 . previous footnote). which coincides with µd . . In this way the matrix (8.7) ω1 (p1 ) . and arrange that ω3 (p2 ) = · · · = ωg (p3 ) = 0. . Any local coordinate z on C yields a local coordinate system {z 1 . The map µg : Symg (C) → J(C) is surjective. (Hint: write explicitly a morphism in The surjectivity of ν follows from the following fact.. Moreover. . The map µ deﬁnes a map µd : Symd (C) → J(C). . namely. . We can actually say more about the morphism ν. and then subtracting a suitable multiple of ω1 from ω2 . that it is a bijection.. Let C d be the dfold cartesian product of C with itself. . . and let z i be a local coordinate centred in pi . Symd (C) can be identiﬁed with the set of eﬀective divisors of C of degree d. Consider now the matrix (8. pd ) = z(pi ).5. Exercise 8.7) is upper triangular. .e.. . Let D = pi ∈ Symg (C). Therefore the latter is a ddimensional complex manifold. . hence it descends to a map Symd (C) → J(C).) Pd . with all the pi ’s distinct. z g } is a local coordinate system around D.. then µ(L) = 0. . .114 8. Proposition 8. It is enough to prove that ν is surjective for a ﬁxed value of d (cf. The symmetric group Sd of order d acts on C d . pd ) → µ(p1 ) + · · · + µ(pd ) is Sd invariant. If D is near D we have (8.6. . . L is trivial).
8. . Corollary 8. Proof. but on C there are no meromorphic functions with a single pole. Every divisor of degree ≥ g on an algebraic curve of genus g is linearly equivalent to an eﬀective divisor. This means that the determinant is not everywhere zero. We may write D = D + D with deg D = g and D ≥ 0. Another consequence is that if C is an elliptic algebraic curve and one chooses a point p ∈ C.e. and choose a divisor D ∈ µ−1 (u). and since ωi (pi ) = 0. hence it is a g projective space.7. By Abel’s theorem the g ﬁbre µ−1 (u) is formed by all eﬀective divisors linearly equivalent to D. Corollary 8.4 one concludes. The (trivial) tangent bundle to C is invariant under the action of Λ. We represent an elliptic curve C as a quotient C/Λ. with p playing the role of the identity element.4 (this is a particular case of Jacobi inversion theorem). and by Lemma 7. so that at the point D corresponding to our choices the Jacobian determinant is nonzero. Proof. Proposition 8. By Abel’s theorem. µ(p) = µ(q) if and only if there is on C a meromorphic function f such that (f ) = p − q.1.9. We have J(C) = C/Λ. and the map µ1 concides with µ. Let u ∈ J(C). so that µ is injective. its diagonal elements hii are nonzero at D.10. therefore the tangent bundle to C is trivial as well. This means that µg establishes a biholomorphic correspondence between a dense subset of Symd (C) and a dense subset of J(C). The canonical bundle of any elliptic curve is trivial. such maps are called birational. By mapping D to J(C) by Abel’s map and taking a counterimage in Symg (C) we obtain an eﬀective divisor E linearly equivalent to D . so that generically it is a point. every smooth algebraic curve of genus 1) is of the form C/Λ for some lattice Λ ⊂ C. the curve has a structure of abelian group. Proof. p µ(p) = p0 ω. Every elliptic smooth algebraic curve (i. Then E + D is eﬀective and linearly equivalent to D. But since dim J(C) = dim Symd (C) the ﬁbre of µg is generically 0dimensional. Let D ∈ Divd (C) with d ≥ g. Proof. The map µg is generically onetoone. THE JACOBIAN VARIETY 115 the points pi the Jacobian matrix {hji } is upper triangular as well. hence it is bijective. Corollary 8. µ is also surjective by Lemma 7.
Since HomZ (Λ2 H1 (T.3 This deﬁnes a projective imbedding of J(C). With this structure. Any point in the lattice singles out univoquely a cell in the lattice. This line bundle has a connection whose curvature is (cohomologous to) 2π ξ. the Jacobian variety of an algebraic curve is always algebraic. Proof. we may use the fact that the image of the map c1 in H 2 (J(C). 2n} are the real basis vectors in Cn generating the lattice. . 3We are using the fact that if a smooth complex vector bundle E on a complex manifold X has a connection whose curvature has no (0. . 2.2.1). Z) H 2 (T. Let {z 1 . Z). that any complex torus satisfying the Riemann bilinear relations is algebraic. which we still denote by ej ). Z) (the N´ronSeveri e group of J(C). j = 1 . Z). .11. After writing ξ on the basis {dz i . Elliptic curves Consider the curve C in C2 given by an equation (8. by transporting them in all points of T by left transport one gets 2n vector ﬁelds. . Jacobian varieties are algebraic. L may be given a holomorphic i structure.2) part. They also generate 2n real vector ﬁelds on T (after identifying Cn with its tangent space at 0 the ej yield tangent vectors to T at the point corresponding to 0. This is no longer true for higher dimensional tori. but. then the complex structure of X can be “lifted” to E.1).2 There exists on J(C) a (in principle smooth) line bundle L whose ﬁrst Chern class is the cohomology class of ξ. . Z).5. Z) ∩ H 1. subsection 5. 2So we are not only proving that the Jacobian variety of an algebraic curve is algebraic. Otherwise. If {ej . Let Λ be a lattice in Cn . z n } be the natural local complex coordinates in T . it is ample by Proposition 7. . The class of ξ is clearly of this type.3. the period matrix may be described as Ωij = ej dz i . any 1dimensional complex torus is algebraic.8) y 2 = P (x).e. Z) canonically (check this isomorphism as an exercise). Cf.. However. Z). ALGEBRAIC CURVES II 1. d¯j } one can check that the stated properties of ξ are z equivalent to the Riemann bilinear relations. i. This provides an identiﬁcation Λ H1 (T. ξ may be regarded as a smooth complexvalued diﬀerential 2form on T .116 8.1 (J(C). Proposition 8. cf. they can be regarded as basis in H1 (T.1) form. [17]. and two opposite sides of the cell determine after identiﬁcation a closed smooth loop in the quotient torus T = Cn /Λ. Z). . so that the latter is algebraic. more generally. as the group of integral 2classes that are of Hodge type (1. Let now ξ be a skewsymmetric Zbilinear form on H1 (T.1) may be represented as H 2 (J(C). since this form is of type (1. The 2form ξ which on the basis {ej } is represented by the intersection matrix Q−1 is a positive (1. According to our previous discussion.
.8). (Hint: around each branch point.2. C may be completed to an algebraic curve C imbedded in P2 — a cubic curve in P2 . which then expresses the equation cutting out C in P2 .1. γ.8) moreover exhibits C as a cover of P1 .. ω z The line bundle O(3p) is very ample. We have Resp (f ω) = 0 . the meromorphic function df /ω is holomorphic outside p. We indeed have. We shall see that these are related by a polynomial identity. 6 z z z z . z z z f3 = 1 β γ 1 + 3 + 2 + O( ) . We may choose constants a. To realize explicitly the imbedding we ˜ may choose three global sections corresponding to the meromorphic functions 1. and f may be chosen in such a way that 1 f (z) = 2 + O(z) . these singles out a complex coordinate z on the open subset of C corresponding to the fundamental cell of the lattice Λ. having a double pole at p. One also checks that the point at inﬁnity is a smooth point.. We want to show that every smooth elliptic curve can be realized in this way. 1 α 1 ˜ f 2 = 6 + 2 + O( ). After proving that all elliptic curves may be written in the form (8. which is branched of order 2 at the points where y = 0 and at the point at inﬁnity. its complete linear system realizes the Kodaira imbedding of C into projective space. so that C is imbedded into P2 . i. We realize C as C/Λ. Exercise 8. and has a triple pole at p. z On the other hand. β. Moreover we ﬁx a nowhere vanishing holomorphic 1form ω (which exists because K is trivial). By RiemannRoch and the vanishing theorem we have h0 (3p) = 3. b. O(2p)) is nonzero.8) in homogeneous coordinates. f . z = P (x) is a good local coordinate. So let C be a smooth elliptic curve. By writing the equation (8. this provides another proof of the triviality of the canonical bundle of an elliptic curve. Show that ω = dx/y is a nowhere vanishing abelian diﬀerential on C. By the genus formula we see that C is an elliptic curve. and P (x) is a polynomial of degree 3. f . A nontrivial section f can be regarded as a global meromorphic function holomorphic in C −{p}.) The equation (8. y are the standard coordinates in C2 . for suitable constants α. Let us assume that C is smooth. If we ﬁx a point p in C and consider the exact sequence of sheaves on C 0 → O(p) → O(2p) → kp → 0 .e. Then we may choose ω = dz. ELLIPTIC CURVES 117 where x. proceeding as usual (Serre duality and vanishing theorem) one shows that H 0 (C. c such that df 1 ˜ f = a + bf + c = 3 + O(z) .
So we express the elliptic curve C in the standard form (Weierstraß representation)4 (8. Exercise 8. Having ﬁxed the complex coordinate z. Determine for what values of the parameter λ the curve (8.10) y 2 = x(x − 1)(x − λ) . Notice that P cannot contain terms of odd degree in its Laurent expansion. the latter is nevertheless completely characterized. We call it the Weierstraß Pfunction. and has at p a simple pole.9) y 2 + βy = x3 − δx + for a suitable constant . where M1 is the set of isomorphism classes of smooth elliptic curves (the moduli space of genus one 4Even though the Weierstraß representation only provides the equation of the aﬃne part of an elliptic curve. as one can check by elementary considerations.10) is smooth. So 1 P(z) = 2 + az 2 + bz 4 + O(z 6 ) z 2 P (z) = − 3 + 2az + 4bz 3 + O(z 5 ) z 1 3a (P(z))3 = 6 + 2 + 3b + O(z 2 ) z z 8a 4 2 (P (z)) = 6 − 2 − 16b + O(z) z z for suitable constants a. By a linear transformation on y we may set β = 0. Such a function must be constant. It is indeed true that any aﬃne plane curve can be uniquely extended to a compact curve by adding points at inﬁnity. From this we see that P satisﬁes the condition (P )2 − 4P 3 + 20 a = constant one usually writes g2 for 20 a and g3 for the constant in the righthand side. In terms of this representation we may introduce a map j : M1 → C. .118 8. We want to elaborate on this construction. 1 ˜ ˜ f 2 + β f − f 3 + δf = O( ) . b. Thus C may be described as a locus in P2 whose equation in aﬃne coordinates is (8. otherwise P(z) − P(−z) would be a nonconstant holomorphic function on C. setting δ = α − β.9) to 0 and 1.2. z So the meromorphic function in the lefthand side is holomorphic away from p. the function f is basically ﬁxed as well. and then by a linear transformation of x we may set the two roots of the polynomial in the righthand side of (8. otherwise it would provide an isomorphism of C with the Riemann sphere. ALGEBRAIC CURVES II so that. Its ˜ derivative is P = −2f .
p2 . and let p4 be the further intersection of this line with C ⊂ P2 . This pole must be of order three. p0 = ψ(0) = [0.e. We shall denote by p the element p ∈ C regarded as a group element (so p0 = 0). P(z).3. p4 . so that the divisor of M (z) is p1 + p2 + p4 − 3 p0 . Z) is biholomorphic to C! (Notice that on the contrary. 6 j(C) = Exercise 8. This is not contradictory in that the quotient H/Sl(2. P (z)) on C vanishes (of order one) only at the points p1 . . where the theory is tailored to work also for ﬁelds of characteristic 2 and 3. Let us choose p0 as the identity element in C. 6By uniformization theory one can also realize this moduli space as a quotient H/Sl(2. in particular M1 gets a structure of complex manifold. ˙ 5The fancy coeﬃcient 1728 comes from arithmetic geometry.10). H and C are not biholomorphic).7 ψ −1 (p) = p p0 dx y mod Λ having chosen p0 at the point at inﬁnity. Proposition 8. P (z)] imbeds C into P2 as the cubic curve cut out by the polynomial F = y 2 − 4x3 + g2 x + g3 ˜ (we use the same aﬃne coordinates as in the previous representation).3. In terms of this construction we may give an elementary geometric visualization of the group law in an elliptic curve. [10]. 3 g2 − 27 g3 One shows that this map is bijective. and has a pole at p0 . ELLIPTIC CURVES 119 curves) 5 3 1728 g2 2 . The function M (z) = M (P(z). p1 + p2 + p4 − 3 p0 ∼ 0. where H is the upper half complex plane. Since f = df /ω we have dx ω= y and the inverse of ψ is the Abel map. Z). 0. ¯ Let M (x.2. The number j(C) is called the jinvariant of the curve C. Cf. y) = mx + ny + q be the equation of the line in P2 through the points p1 . Write the jinvariant as a function of the parameter λ in equation (8. p2 . 7One should bear in mind that we have identiﬁed C with a quotient C/Λ. 1]. z → [1. We may therefore say that the moduli space M1 is isomorphic to C. and one has µ(p1 + p2 + p3 − 3 p0 ) = 0). By Abel’s ¯ ¯ theorem. we have that p1 + p2 + p 3 = 0 ¯ ¯ ¯ if and and only if p1 + p2 + p 3 ∼ 3 p 0 (indeed one may think that p = µ(p). i. Do you think that λ is a good coordinate on the moduli space M1 ? The holomorphic map ψ : C → P2 .
Vice versa. We have therefore shown that p1 + p2 + p3 = 0 ¯ ¯ ¯ if and only if p1 . The points p1 .4. 3. Blowing up a point in a variety8 means replacing the point with all possible directions along which one can approach it while moving in the variety. the set of all possible directions is a copy of P1 . ¯ 3. p2 . 2 pi = 0. ALGEBRAIC CURVES II If p1 + p2 + p3 ∼ 3 p0 .1. and p1 . then p3 ∼ p4 . v2 ) in C2 . since this space is 2dimensional. and w0 . C is a double cover of P1 . with coordinates (x. then q = (a. p2 . Example 8.120 8. b). Γ is a smooth hypersurface in C2 × P1 . We shall at ﬁrst consider the blowup of C2 at the origin. as one may check in homogeneous coordinates. ¯ ¯ if q ∈ C is such that q = −¯. Blowup. v1 . p3 = (α2 . q2 are one the inverse of the other. p0 . then p + q ∼ 2 p0 . where U0 . q2 intersects C at the point at inﬁnity. p3 are collinear points in P2 . V0 = C2 × U0 and V1 = C2 × U1 . then Γ can be regarded as a open subset of the quadric in P2 × P1 having equation v1 w1 − v2 w0 = 0. −b). q2 = (0. The blowup of C2 at the origin is the subvariety Γ of C2 × P1 deﬁned by the equation x w1 − y w0 = 0 . The line through q1 . To show that Γ is a complex manifold we cover C2 × P1 with two coordinate charts. if p = (a. and q1 + q2 ∼ 2 p0 . Let x. i). U1 are the standard aﬃne charts in P1 . so that p3 = p4 . 0). t0 = w1 /w0 ) and (x. p1 + p2 + p3 − 3 p0 is the divisor of the meromorphic function M . So the branch points pi are 2torsion elements in the group. −i) lie on C. t1 = w0 /w1 ). p3 are collinear. p2 . Let C be an elliptic curve having a Weierstraß representation y 2 = x3 − 1. so that Γ is actually algebraic. p2 = (α. So in this case the elements q1 . w1 homogeneous coordinates in P1 . so that p1 +p2 +p3 −3 p0 ∼ 0. y be the standard coordinates in C2 . Nodal curves In this section we show how (plane) curve singularities may be resolved by a procedure called blowup. branched at the three points p1 = (1. p3 are collinear. y. and q is the further intersection of C ¯ p with the line going through p. 0) (where α = e2πi/3 ) and at the point at inﬁnity p0 . 8Our treatment of the blowup of an algebraic variety is basically taken from [1]. 0). p2 . p3 are collinear. More generally. so that p1 + p2 + p3 = 0. ¯ ¯ ¯ The two points q1 = (0. hence it is a complex surface. On the other hand if we put homogeneous coordinates (v0 . if p1 . y. .
The manifold X is the blowup of X at p. so that the line bundle is actually the tautological bundle OP1 (−1). The blowup of P2 at a point is an algebraic surface X1 (an example of a Del Pezzo surface). and the transition function g : U0 ∩ U1 → C∗ is g(w0 . so that π makes Γ into the total space of a line bundle over P1 . Let U be a chart around p. y). 3. Example 8. .11). This construction is local in nature and therefore can be applied to any complex surface X (twodimensional complex manifold) at any point p. and get a complex manifold X with a projection σ : X → X which is a biholomorphism outside σ −1 (p). If p ∈ C2 − {0} then σ −1 (p) is a point (which means that there is a unique line through p and 0). NODAL CURVES 121 Since Γ is a subset of C2 × P1 there are two projections (8. This bundle trivializes over the cover {U0 . The inverse image E = σ −1 (p) is a divisor in X .9 On the contrary σ −1 (0) in C2 . and is isomorphic to P1 . and that σ −1 (C) \ E is isomorphic to C − {0}.2. 9So. w1 ) = w0 /w1 . with complex coordinates (x. w1 ) ∈ P1 is the line x w1 − y w0 = 0. Let C be a curve in C2 containing the origin.11) σ Γ π / P1 C2 which are holomorphic. called the exceptional divisor. is biholomorphic to X1 minus a projective line (so X1 is a compactiﬁcation of Γ).1. so that σ : Γ − σ −1 (0) → C2 − {0} is a biholomorphism.3. the map σ is a birational morphism. Transforms of a curve. The construction of the blowup Γ shows that X is algebraic if X is. We denote as before Γ the blowup of C2 at the origin and make reference to the diagram (8. U1 }. according to a terminology we have introduce in a previous chapter. so that one can replace U by U inside X. P1 is the set of lines through the origin The ﬁbre of π over a point (w0 . By repeating the same construction we get a complex manifold U with projections π U − − → P1 −− σ U and σ : U − σ −1 (p) → U − {p} is a biholomorphism. Notice that the inverse image σ −1 (C) ⊂ Γ contains the exceptional divisor E. the manifold Γ. obtained by blowing up C2 at the origin.
w1 ) with xk w1 − yk w0 = 0.2. ALGEBRAIC CURVES II Definition 8. With reference to equation (8. This means that the curve has m tangents at the point 0 (but some of them might coincide). in that case. then the strict transform meets the exceptional divisor at a point of multiplicity m. as are the tangents at C at 0. Then w1 /w0 = yk /xk . Definition 8. y 2 = x2 and y 2 = x3 have multiplicity 2 at 0. namely. yk . The ﬁrst two have two distinct tangents at 0. Let the (aﬃne plane) curve C be given by the equation f (x. the third has a double tangent. Definition 8. 0. If the curve C has multiplicity m at 0 than it has m tangents at 0. then σ −1 (xk . The strict transform C of C meets the exceptional divisor in as many points as are the directions along which one can approach 0 on C. if C is smooth at 0. w0 . in that case {σ −1 (pk )} converges to the point (0. its strict transform meets E at one point. and the two tangents to the curve at that point are distinct.5. So a sequence {pk = (xk . Assume that for k big enough one has w0 = 0 (otherwise we would assume w1 = 0 and would make a similar argument). yk )} convergent to 0 lifts to a convergent sequence in Γ if and only if the lines rk admit a limit line r. h) of E. This means that the lines rk joining 0 to pk approach the limit line r having equation y = hk. and {σ −1 (pk )} converges if and only if {yk /xk } has a limit. We say that C has multiplicity m at 0 if the Taylor expansion of f at 0 starts at degree m.4. . The curve σ −1 (C) ⊂ Γ is the total transform of C. To this end we must understand what are the sequences in C2 which converge to 0 that are lifted by σ to convergent sequences. yk )}k∈N be a sequence of points in C2 converging to 0.122 8.10). say h. yk ) is the point (xk . 1. the lifted sequence converges to the point of E representing the line r. A curve is smooth at 0 if and only if its multiplicity at 0 is 1. We want to check what points are added to σ −1 (C) \ E when taking the topological closure. Example 8. determine for what values of λ the curve has a nodal singularity. The curve obtained by taking the topological closure of σ −1 (C) \ E in Γ is the strict transform of C. A singular point of a plane curve C is said to be nodal if at that point C has multiplicity 2. y) = 0. Let {pk = (xk .6. By choosing suitable coordinates one can apply this notion to any point of a plane curve. Exercise 8. So. The curves xy = 0. Every intersection point must be counted with its multiplicity: if at the point 0 the curve C has m coinciding tangents.3. and its strict transform meets the exceptional divisor of Γ at m points (notice however that these points are all distinct only if the m tangents are distincts).
0. w0 . This curve has multiplicity 2 at the origin where it has a double tangent. Exercise 8. In the other chart the equation of C is t2 = u.12) we obtain the equation of the total transform in Γ ∩ U0 .12) y = u w0 /w1 x = v w /w 1 0 y = v in Γ ∩ V1 and Γ ∩ V0 . The projection σ is described as x = u (8. consisting of the exceptional divisor and two more genus zero components. 0. we have one intersection point because the two tangents to C at the origin coincide. −1) as intersection points of the strict transform with the exceptional divisor. (The cusp) Let C be curve with equation y 2 = x3 . and its two tangents at the origin have equations y = ±x. (8. respectively.10 Proceeding as in the previous example we get the equation v t3 = 1 for C in Γ ∩ V0 . Show that around a nodal singularity a curve is isomorphic to an open neighbourhood of the origin of the curve xy = 0 in C2 . So C has a nodal singularity at the origin. v.8. and the total transform is a reducible curve with two components meeting at a (double) point. each of which meets the exceptional divisor at a point.3. This curve has multiplicity 2 at the origin. with two irreducible components which meet at two points. which at the origin are tangent to the two lines y = ±α x. so that C meets E at the point (0. the strict transform now has equation t3 v + t2 − 1.7. 2 . NODAL CURVES 123 Exercise 8. Repeat the previous calculations for the nodal curve xy = 0. 1) and (0. (Blowing up a nodal singularity. 10Indeed this curve can be regarded as the limit for α → 0 of the family of nodal curves x3 + α2 x2 − y = 0. The strict transform is an irreducible curve. Example 8. By letting u = 0 we obtain the points (0. u2 = 0 is the equation of the exceptional divisor. 1). By substituting the ﬁrst of the representations (8. w1 ) ∈ C2 × P1  u w0 = v w1 } . In particular show that the total transform is a reducible curve. 0. so that the equation of the strict transform is u + 1 − t2 = 0. yielding the same intersection points.12) into the equation of C we obtain the equation of the restriction of the total transform to Γ ∩ U1 : u2 (u + 1 − t2 ) = 0 where t = w0 /w1 . The total transform is a reducible curve. so that C does not meet E in this chart. 1. 1. Example 8. 0.10. By substituting the second representation in eq. We recall that Γ is described as the locus {(u.) We consider the curve C ⊂ C2 having equation x3 + x2 − y 2 = 0.9.
Assume that N has t irreducible components N1 . ALGEBRAIC CURVES II 3. Normalization of a nodal plane curve. Projection onto the xaxis makes C0 into a double cover of C. w2 ) and considering it as a curve C in P2 . w1 . This curve has genus 1 and is singular at inﬁnity (as one could have alredy guessed since the genus formula for smooth plane curves does not work). Then: t g(C) = 1 g(Ni ) + 1 − t + δ. η = w1 /w2 (in this coordinates the point at inﬁnity on the xaxis is η = ξ = 0) we have the equation ξ2 = η4 − ξ4 showing that C is indeed singular at inﬁnity.124 8. Then it can be desingularized as in Example 8.3. We give here. with one less singular point. N is called the normalization of C. Example 8. we obtain that the normalization is a projective line. It is clear from the previous examples that the strict transform of a plane nodal curve C (i. The curve C0 can be completed to a projective curve simply by writing its equation in homogeneous coordinates (w0 .11. Therefore after a ﬁnite number of blowups we obtain a smooth curve N . after introducing aﬃne coordinates ξ = w0 /w2 . . Let us consider the smooth curve C0 in C2 having equation y 2 = x4 − 1. One can redeﬁne the coordinates ξ.8. by applying this formula to Example 8. 0) and (±i. η so that C has equation (ξ − η 2 )(ξ + η 2 ) = 0 showing that C is a nodal curve. The equation of C is 2 2 4 4 w0 w2 − w1 + w0 = 0 . indeed.e.8. and that C has δ singular points. A genus formula. a formula which can be used to compute the genus of the normalization N of a nodal curve C. Nt . . without proof. a plane curve with only nodal singularities) is again a nodal curve. which in this case is not a branch point. together with a birational morphism π : N → C. . we are thus compactifying C0 by adding a point at inﬁnity. branched at the points (±1. 0).. For instance. . .
Wiley. [17] S. Foundations of Diﬀerential Geometry. Kobayashi. Fibre Bundles. J. The Geometry of Supermania e folds. Th´orie des Faisceaux. Homotopy theory. Princeton 1987.. Pitagora. Diﬀerential Forms in Algebraic Topology.it/ pub off/lectures/). I. L.. Hartshorne. SpringerVerlag. SpringerVerlag. New York 1994. McGrawHill. New York 1975. Benjamin. Homotopy theory. Publishers. Stammbach. U. Sheaf theory. Bott & L. Lectures on Algebraic Topology. Hain. [8] A. New York 1977.Bibliography [1] F.trieste. Interscience Publ. Trieste 2000 (available from o http://www.ictp. Husemoller. Princeton University Press. [4] G. A. Kobayashi & K. Principles of Algebraic Geometry. Godement. e [6] B. [10] R. Hermann.E. New York 1963. Academic Press. Bruzzo & D. Dordrecht 1991. [12] P. Gray. Topological Methods in Algebraic Geometry. Algebraic geometry. Diﬀerential Geometry of Complex Vector Bundles. New York 1966. New York 1959. transcendental aspects. SpringerVerlag. 125 . Hilton & U. Lazzeri. [15] D. Broglia & F. [2] C. Nomizu. Quaderni dell’Unione Matematica Italiana 13. Kluwer Acad. SpringerVerlag.T. [13] F. Berlin 1966.J. Griﬃths & J. Bartocci. 9 e o (1957) 119221. [16] S. [14] S. Topologia delle superﬁcie algebriche in P3 (C). F. Grothendieck. Paris 1964. Tˆhoku Math. Hirzebruch. Bredon. Hern´ndez Ruip´rez. [11] R. New York 1971. Greenberg. Harris. New York 1982. Sur quelques points d’alg`bre homologique. ICTP Lectures Notes. [9] P. New York 1967. Bologna 1979. Tu. Acquistapace. G¨ttsche ed. A Course in Homological Algebra. [3] R. Academic Press. [5] R. Vol. Hu. School on Algebraic Geometry. ICTP. J. McGrawHill. [7] M. Moduli of Riemann surfaces. New York 1967.
SpringerVerlag. [21] B.H. Spanier. Exact couples in algebraic topology. Cambridge Univ. Co. 57 (1953). [19] E. 56 (1952).S.126 BIBLIOGRAPHY [18] W. II.Algebraic topology. Wells Jr.. Frederick Ungar Publ. 20. New York 1970. Corrected reprint. SpringerVerlag. Diﬀerential analysis on complex manifolds. Massey. [20] B. London Math. Ann. Press.O. 248286. New YorkBerlin 1980. 3363396. Soc. Notes Ser. . [22] R. I. New YorkBerlin 1981. Lect.L.R. Van der Waerden. Math.. Sheaf theory. Cambridge 1975. Algebra. Tennison.
This action might not be possible to undo. Are you sure you want to continue?
Use one of your book credits to continue reading from where you left off, or restart the preview.