HYPERCOVERINGS

Contents
1. Introduction 1
2. Hypercoverings 2
3. Acyclicity 4
4. Covering hypercoverings 7
5. Adding simplices 9
6. Homotopies 10
7. Cech cohomology associated to hypercoverings 12
8. Cohomology and hypercoverings 14
9. Hypercoverings of spaces 16
10. Other chapters 19
References 20
1. Introduction
Let C be a site, see Sites, Denition 6.2. Let X be an object of C. Given an abelian
sheaf F on C we would like to compute its cohomology groups
H
i
(X, F).
According to our general denitions (insert future reference here) this cohomology
group is computed by choosing an injective resolution
0 F I
0
I
1
. . .
and setting
H
i
(X, F) = H
i
((X, I
0
) (X, I
1
) (X, I
2
) . . .)
We will have to do quite a bit of work to prove that we may also compute these
cohomology groups without choosing an injective resolution. Also, we will only do
this in case the site C has bre products.
A hypercovering in a site is a generalization of a covering. See [MA71, Expose V,
Sec. 7]. A hypercovering is a special case of a simplicial augmentation where one
has cohomological descent, see [MA71, Expose Vbis]. A nice manuscript on coho-
mological descent is the text by Brian Conrad, see http://math.stanford.edu/
~
conrad/papers/hypercover.pdf. Brians text follows the exposition in [MA71,
Expose Vbis], and in particular discusses a more general kind of hypercoverings,
such as proper hypercoverings of schemes used to compute etale cohomology for
example. A proper hypercovering can be seen as a hypercovering in the category of
schemes endowed with a dierent topology than the etale topology, but still they
can be used to compute the etale cohomology.
This is a chapter of the Stacks Project, version 27b740c, compiled on Aug 20, 2011.
1
2 HYPERCOVERINGS
2. Hypercoverings
In order to start we make the following denition. The letters SR stand for
Semi-Representable.
Denition 2.1. Let C be a site with bre products. Let X Ob(C) be an object
of C. We denote SR(C, X) the category of semi-representable objects dened as
follows
(1) objects are families of morphisms {U
i
X}
iI
, and
(2) morphisms {U
i
X}
iI
{V
j
X}
jJ
are given by a map : I J
and for each i I a morphism f
i
: U
i
V
(i)
over X.
This denition is dierent from the one in [MA71, Expose V, Sec. 7], but it seems
exible enough to do all the required arguments. Note that this is a big category.
We will later bound the size of the index sets I that we need and we can then
redene SR(C, X) to become a category.
Denition 2.2. Let C be a site with bre products. Let X Ob(C) be an object
of C. We denote F the functor which associates a sheaf to a semi-representable
object. In a formula
F : SR(C, X) PSh(C)
{U
i
X}
iI

iI
h
Ui
where h
U
denotes the representable presheaf associated to the object U.
Given a morphism U X we obtain a morphism h
U
h
X
of representable
presheaves. Thus it makes more sense to think of F as a functor into the category
of presheaves of sets over h
X
, namely PSh(C)/h
X
.
Lemma 2.3. Let C be a site with bre products. Let X Ob(C) be an object of
C. The category SR(C, X) has coproducts and nite limits. Moreover, the functor
F commutes with coproducts and bre products, and transforms products into bre
products over h
X
. In other words, it commutes with nite limits as a functor into
PSh(C)/h
X
.
Proof. It is clear that the coproduct of {U
i
X}
iI
and {V
j
X}
jJ
is
{U
i
X}
iI
{V
j
X}
jJ
and similarly for coproducts of families of families
of morphisms with target X. The object {X X} is a nal object of SR(C, X).
Suppose given a morphism (, f
i
) : {U
i
X}
iI
{V
j
X}
jJ
and a morphism
(, g
k
) : {W
k
X}
kK
{V
j
X}
jJ
. The bred product of these morphisms
is given by
{U
i

fi,Vj,g
k
W
k
X}
(i,j,k)IJK such that k=(i)=(j)
The bre products exist by the assumption that C has bre products. Thus
SR(C, X) has nite limits, see Categories, Lemma 16.4. The statements on the
functor F are clear from the constructions above.
Denition 2.4. Let C be a site with bred products. Let X be an object of C.
Let f = (, f
i
) : {U
i
X}
iI
{V
j
X}
jJ
be a morphism in the category
SR(C, X). We say that f is a covering if for every j J the family of morphisms
{U
i
V
j
}
iI,(i)=j
is a covering for the site C.
Lemma 2.5. Let C be a site with bred products. Let X Ob(C).
HYPERCOVERINGS 3
(1) A composition of coverings in SR(C, X) is a covering.
(2) A base change of coverings is a covering.
(3) If A B and K L are coverings, then AK B L is a covering.
Proof. Immediate from the axioms of a site. (Number (3) is the composition
AK BK BL and hence a composition of basechanges of coverings.)
According to the results in the chapter on simplicial methods the coskelet of a
truncated simplicial object of SR(C, X) exists. Hence the following denition makes
sense.
Denition 2.6. Let C be a site. Let X Ob(C) be an object of C. A hypercovering
of X is a simplicial object K in the category SR(C, X) such that
(1) The object K
0
is a covering of X for the site C.
(2) For every n 0 the canonical morphism
K
n+1
(cosk
n
sk
n
K)
n+1
is a covering in the sense dened above.
Condition (1) makes sense since each object of SR(C, X) is after all a family of
morphisms with target X. It could also be formulated as saying that the morphism
of K
0
to the nal object of SR(C, X) is a covering.
Example 2.7. Let {U
i
X}
iI
be a covering of the site C. Set K
0
= {U
i

X}
iI
. Then K
0
is a 0-truncated simplicial object of SR(C, X). Hence we may
form
K = cosk
0
K
0
.
Clearly K passes condition (1) of Denition 2.6. Since all the morphisms K
n+1

(cosk
n
sk
n
K)
n+1
are isomorphisms it also passes condition (2). Note that the terms
K
n
are the usual
K
n
= {U
i0

X
U
i1

X
. . .
X
U
in
X}
(i0,i1,...,in)I
n+1
Lemma 2.8. Let C be a site with bre products. Let X Ob(C) be an object of C.
The collection of all hypercoverings of X forms a set.
Proof. Since C is a site, the set of all coverings of S forms a set. Thus we see
that the collection of possible K
0
forms a set. Suppose we have shown that the
collection of all possible K
0
, . . . , K
n
form a set. Then it is enough to show that given
K
0
, . . . , K
n
the collection of all possible K
n+1
forms a set. And this is clearly true
since we have to choose K
n+1
among all possible coverings of (cosk
n
sk
n
K)
n+1
.
Remark 2.9. The lemma does not just say that there is a conal system of choices
of hypercoverings that is a set, but that really the hypercoverings form a set.
The category of presheaves on C has nite (co)limits. Hence the functors cosk
n
exists for presheaves of sets.
Lemma 2.10. Let C be a site with bre products. Let X Ob(C) be an object of
C. Let K be a hypercovering of X. Consider the simplicial object F(K) of PSh(C),
endowed with its augmentation to the constant simplicial presheaf h
X
.
(1) The morphism of presheaves F(K)
0
h
X
becomes a surjection after shea-
cation.
4 HYPERCOVERINGS
(2) The morphism
(d
1
0
, d
1
1
) : F(K)
1
F(K)
0

h
X
F(K)
0
becomes a surjection after sheacation.
(3) For every n 1 the morphism
F(K)
n+1
(cosk
n
sk
n
F(K))
n+1
turns into a surjection after sheacation.
Proof. We will use the fact that if {U
i
U}
iI
is a covering of the site C, then
the morphism

iI
h
Ui
h
U
becomes surjective after sheacation, see Sites, Lemma 12.5. Thus the rst asser-
tion follows immediately.
For the second assertion, note that according to Simplicial, Example 16.2 the sim-
plicial object cosk
0
sk
0
K has terms K
0
. . . K
0
. Thus according to the denition
of a hypercovering we see that (d
1
0
, d
1
1
) : K
1
K
0
K
0
is a covering. Hence (2)
follows from the claim above and the fact that F transforms products into bred
products over h
X
.
For the third, we claim that cosk
n
sk
n
F(K) = F(cosk
n
sk
n
K) for n 1. To prove
this, denote temporarily F

the functor SR(C, X) PSh(C)/h

X
. By Lemma 2.3
the functor F

commutes with nite limits. By our description of the cosk

n
functor
in Simplicial, Section 16 we see that cosk
n
sk
n
F

(K) = F

(cosk
n
sk
n
K). Recall
that the category used in the description of (cosk
n
U)
m
in Simplicial, Lemma 16.3
is the category (/[m])
opp
n
. It is an amusing exercise to show that (/[m])
n
is
a nonempty connected category (see Categories, Denition 15.1) as soon as n
1. Hence, Categories, Lemma 15.2 shows that cosk
n
sk
n
F

(K) = cosk
n
sk
n
F(K).
Whence the claim. Property (2) follows from this, because now we see that the
morphism in (2) is the result of applying the functor F to a covering as in Denition
2.4, and the result follows from the rst fact mentioned in this proof.
3. Acyclicity
Let C be a site. For a presheaf of sets F we denote Z
F
the presheaf of abelian
groups dened by the rule
Z
F
(U) = free abelian group on F(U).
We will sometimes call this the free abelian presheaf on F. Of course the con-
struction F Z
F
is a functor and it is left adjoint to the forgetful functor
PAb(C) PSh(C). Of course the sheacation Z
#
F
is a sheaf of abelian groups,
and the functor F Z
#
F
is a left adjoint as well. We sometimes call Z
#
F
the free
abelian sheaf on F.
For an object X of the site C we denote Z
X
the free abelian presheaf on h
X
, and
we denote Z
#
X
its sheacation.
Denition 3.1. Let C be a site. Let K be a simplicial object of PSh(C). By the
above we get a simplicial object Z
#
K
of Ab(C). We can take its associated complex
of abelian presheaves s(Z
#
K
), see Simplicial, Section 20. The homology of K is the
homology of the complex of abelian sheaves s(Z
#
K
).
HYPERCOVERINGS 5
In other words, the ith homology H
i
(K) of K is the sheaf of abelian groups
H
i
(K) = H
i
(s(Z
#
K
)). In this section we worry about the homology in case K
is a hypercovering of an object X of C.
Lemma 3.2. Let C be a site. Let F G be a morphism of presheaves of sets.
Denote K the simplicial object of PSh(C) whose nth term is the (n + 1)st bre
product of F over G, see Simplicial, Example 3.5. Then, if F G is surjective
after sheacation, we have
H
i
(K) =
_
0 if i > 0
Z
#
G
if i = 0
The isomorphism in degree 0 is given by the morphsm H
0
(K) Z
#
G
coming from
the map (Z
#
K
)
0
= Z
#
F
Z
#
G
.
Proof. Let G

G be the image of the morphism F G. Let U Ob(C). Set

A = F(U) and B = G

(U). Then the simplicial set K(U) is equal to the simplicial

set with n-simplices given by
A
B
A
B
. . .
B
A (n + 1 factors).
By Simplicial, Lemma 27.4 the morphism K(U) B is a homotopy equivalence.
Hence applying the functor free abelian group on to this we deduce that
Z
K
(U) Z
B
is a homotopy equivalence. Note that s(Z
B
) is the complex
. . .

bB
Z
0

bB
Z
1

bB
Z
0

bB
Z 0
see Simplicial, Lemma 20.3. Thus we see that H
i
(s(Z
K
(U))) = 0 for i > 0, and
H
0
(s(Z
K
(U))) =

bB
Z =

sG

(U)
Z. These identications are compatible with
restriction maps.
We conclude that H
i
(s(Z
K
)) = 0 for i > 0 and H
0
(s(Z
K
)) = Z
G
, where here we
compute homology groups in PAb(C). Since sheacation is an exact functor we
deduce the result of the lemma. Namely, the exactness implies that H
0
(s(Z
K
))
#
=
H
0
(s(Z
#
K
)), and similarly for other indices.
Lemma 3.3. Let C be a site. Let f : L K be a morphism of simplicial objects
of PSh(C). Let n 0 be an integer. Assume that
(1) For i < n the morphism L
i
K
i
is an isomorphism.
(2) The morphism L
n
K
n
is surjective after sheacation.
(3) The canonical map L cosk
n
sk
n
L is an isomorphism.
(4) The canonical map K cosk
n
sk
n
K is an isomorphism.
Then H
i
(f) : H
i
(L) H
i
(K) is an isomorphism.
Proof. This proof is exactly the same as the proof of Lemma 3.2 above. Namely,
we rst let K

n
K
n
be the sub presheaf which is the image of the map L
n
K
n
.
Assumption (2) means that the sheacation of K

n
is equal to the sheacation
of K
n
. Moreover, since L
i
= K
i
for all i < n we see that get an n-truncated
simplicial presheaf U by taking U
0
= L
0
= K
0
, . . . , U
n1
= L
n1
= K
n1
, U
n
=
K

n
. Denote K

= cosk
n
U, a simplicial presheaf. Because we can construct K

m
as
a nite limit, and since sheacation is exact, we see that (K

m
)
#
= K
m
. In other
words, (K

)
#
= K
#
. We conclude, by exactness of sheacation once more, that
6 HYPERCOVERINGS
H
i
(K) = H
i
(K

). Thus it suces to prove the lemma for the morphism L K

, in
other words, we may assume that L
n
K
n
is a surjective morphism of presheaves!
In this case, for any object U of C we see that the morphism of simplicial sets
L(U) K(U)
satises all the assumptions of Simplicial, Lemma 27.3. Hence it is a homotopy
equivalence, and thus
Z
L
(U) Z
K
(U)
is a homotopy equivalence too. This for all U. The result follows.
Lemma 3.4. Let C be a site. Let K be a simplicial presheaf. Let G be a presheaf.
Let K G be an augmentation of K towards G. Assume that
(1) The morphism of presheaves K
0
G becomes a surjection after sheaca-
tion.
(2) The morphism
(d
1
0
, d
1
1
) : K
1
K
0

G
K
0
becomes a surjection after sheacation.
(3) For every n 1 the morphism
K
n+1
(cosk
n
sk
n
K)
n+1
turns into a surjection after sheacation.
Then H
i
(K) = 0 for i > 0 and H
0
(K) = Z
#
G
.
Proof. Denote K
n
= cosk
n
sk
n
K for n 1. Dene K
0
as the simplicial object
with terms (K
0
)
n
equal to the (n + 1)-fold bred product K
0

G
. . .
G
K
0
, see
Simplicial, Example 3.5. We have morphisms
K . . . K
n
K
n1
. . . K
1
K
0
.
The morphisms K K
i
, K
j
K
i
for j i 1 come from the universal properties
of the cosk
n
functors. The morphism K
1
K
0
is the canonical morphism from
Simplicial, Remark 17.4. We also recall that K
0
cosk
1
sk
1
K
0
is an isomorphism,
see Simplicial, Lemma 17.3.
By Lemma 3.2 we see that H
i
(K
0
) = 0 for i > 0 and H
0
(K
0
) = Z
#
G
.
Pick n 1. Consider the morphism K
n
K
n1
. It is an isomorphism on terms
of degree < n. Note that K
n
cosk
n
sk
n
K
n
and K
n1
cosk
n
sk
n
K
n1
are
isomorphisms. Note that (K
n
)
n
= K
n
and that (K
n1
)
n
= (cosk
n1
sk
n1
K)
n
.
Hence by assumption, we have that (K
n
)
n
(K
n1
)
n
is a morphism of presheaves
which becomes surjective after sheacation. By Lemma 3.3 we conclude that
H
i
(K
n
) = H
i
(K
n1
). Combined with the above this proves the lemma.
Lemma 3.5. Let C be a site with bre products. Let X be an object of of C. Let
K be a hypercovering of X. The homology of the simplicial presheaf F(K) is 0 in
degrees > 0 and equal to Z
#
X
in degree 0.
Proof. Combine Lemmas 3.4 and 2.10.
HYPERCOVERINGS 7
4. Covering hypercoverings
Here are some ways to construct hypercoverings. We note that since the category
SR(C, X) has bre products the category of simplicial objects of SR(C, X) has bre
products as well, see Simplicial, Lemma 7.2.
Lemma 4.1. Let C be a site with bre products. Let X be an object of C. Let
K, L, M be simplicial objects of SR(C, X). Let a : K L, b : M L be morphisms.
Assume
(1) K is a hypercovering of X,
(2) the morphism M
0
L
0
is a covering, and
(3) for all n 0 in the diagram
M
n+1

(cosk
n
sk
n
M)
n+1

L
n+1

(cosknsknL)n+1
(cosk
n
sk
n
M)
n+1

L
n+1

(cosk
n
sk
n
L)
n+1
the arrow is a covering.
Then the bre product K
L
M is a hypercovering of X.
Proof. The morphism (K
L
M)
0
= K
0

L0
M
0
K
0
is a base change of a
covering by (2), hence a covering, see Lemma 2.5. And K
0
{X X} is a
covering by (1). Thus (K
L
M)
0
{X X} is a covering by Lemma 2.5. Hence
K
L
M satises the rst condition of Denition 2.6.
We still have to check that
K
n+1

Ln+1
M
n+1
= (K
L
M)
n+1
(cosk
n
sk
n
(K
L
M))
n+1
is a covering for all n 0. We abbreviate as follows: A = (cosk
n
sk
n
K)
n+1
,
B = (cosk
n
sk
n
L)
n+1
, and C = (cosk
n
sk
n
M)
n+1
. The functor cosk
n
sk
n
commutes
with bre products, see Simplicial, Lemma 16.13. Thus the right hand side above
is equal to A
B
C. Consider the following commutative diagram
K
n+1

Ln+1
M
n+1

M
n+1

K
n+1

L
n+1

L
n+1

B
C

C

A

B
This diagram shows that
K
n+1

Ln+1
M
n+1
= (K
n+1

B
C)
(Ln+1
B
C),
M
n+1
Now, K
n+1

B
C A
B
C is a base change of the covering K
n+1
A via the
morphism A
B
C A, hence is a covering. By assumption (3) the morphism is
a covering. Hence the morphism
(K
n+1

B
C)
(Ln+1
B
C),
M
n+1
K
n+1

B
C
8 HYPERCOVERINGS
is a covering as a base change of a covering. The lemma follows as a composition
of coverings is a covering.
Lemma 4.2. Let C be a site with bre products. Let X be an object of C. If K, L
are hypercoverings of X, then K L is a hypercovering of X.
Proof. You can either verify this directly, or use Lemma 4.1 above and check that
L {X X} has property (3).
Let C be a site with bre products. Let X be an object of C. Since the category
SR(C, X) has coproducts and nite limits, it is permissible to speak about the
objects U K and Hom(U, K) for certain simplicial sets U (for example those with
nitely many nondegenerate simplices) and any simplicial object K of SR(C, X).
See Simplicial, Sections 12 and 14.
Lemma 4.3. Let C be a site with bre products. Let X be an object of C. Let K
be a hypercovering of X. Let k 0 be an integer. Let u : Z K
k
be a covering in
in SR(C, X). Then there exists a morphism of hypercoverings f : L K such that
L
k
K
k
factors through u.
Proof. Denote Y = K
k
. There is a canonical morphism K Hom([k], Y )
corresponding to id
Y
via Simplicial, Lemma 14.5. We will use the description
of Hom([k], Y ) and Hom([k], Z) given in that lemma. In particular there is a
morphism Hom([k], Y ) Hom([k], Z) which on degree n terms is the morphism

:[k][n]
Y

:[k][n]
Z.
Set
L = K
Hom([n],Y )
Hom([n], Z).
The morphism L
k
K
k
sits in to a commutative diagram
L
k

:[k][n]
Y
pr
id
[k]

K
k

:[k][n]
Z
pr
id
[k]

Z
Since the composition of the two bottom arrows is the identity we conclude that
we have the desired factorization.
We still have to show that L is a hypercovering of X. To see this we will use Lemma
4.1. Condition (1) is satised by assumption. For (2), the morphism
Hom([k], Y )
0
Hom([k], Z)
0
is a covering because it is a product of coverings, see Lemma 2.5. For (3) suppose
rst that n 1. In this case by Simplicial, Lemma 18.12 we have Hom([k], Y ) =
cosk
n
sk
n
Hom([k], Y ) and similarly for Z. Thus condition (3) for n > 0 is clear.
For n = 0, the diagram of condition (3) of Lemma 4.1 is, according to Simplicial,
Lemma 18.13, the diagram

:[k][1]
Z

Z Z

:[k][1]
Y

Y Y
HYPERCOVERINGS 9
with obvious horizontal arrows. Thus the morphism is the morphism

:[k][1]
Z

:[k][1] not onto

Z

:[k][1] onto
Y
which is a product of coverings and hence a covering according to Lemma 4.1 once
again.
Lemma 4.4. Let C be a site with bre products. Let X be an object of C. Let
K be a hypercovering of X. Let n 0 be an integer. Let u : F F(K
n
) be
a morphism of presheaves which becomes surjective on sheacation. Then there
exists a morphism of hypercoverings f : L K such that F(f
n
) : F(L
n
) F(K
n
)
factors through u.
Proof. Write K
n
= {U
i
X}
iI
. Thus the map u is a morphism of presheaves of
sets u : F h
ui
. The assumption on u means that for every i I there exists a
covering {U
ij
U
i
}
jIi
of the site C and a morphism of presheaves t
ij
: h
Uij
F
such that u t
ij
is the map h
Uij
h
Ui
coming from the morphism U
ij
U
i
. Set
J =
iI
I
i
, and let : J I be the obvious map. For j J denote V
j
= U
(j)j
.
Set Z = {V
j
X}
jJ
. Finally, consider the morphism u

: Z K
n
given by
: J I and the morphisms V
j
= U
(j)j
U
(j)
above. Clearly, this is a
covering in the category SR(C, X), and by construction F(u

) : F(Z) F(K
n
)
factors through u. Thus the result follows from Lemma 4.3 above.
5. Adding simplices
In this section we prove some technical lemmas which we will need later. Let C be
a site with bre products. Let X be an object of C. As we pointed out in Section
4 above, the objects U K and Hom(U, K) for certain simplicial sets U and any
simplicial object K of SR(C, X) are dened. See Simplicial, Sections 12 and 14.
Lemma 5.1. Let C be a site with bre products. Let X be an object of C. Let K
be a hypercovering of X. Let U V be simplicial sets, with U
n
, V
n
nite nonempty
for all n. Assume that U has nitely many nondegenerate simplices. Suppose n 0
and x V
n
, x U
n
are such that
(1) V
i
= U
i
for i < n,
(2) V
n
= U
n
{x},
(3) any z V
j
, z U
j
for j > n is degenerate.
Then the morphism
Hom(V, K)
0
Hom(U, K)
0
of SR(C, X) is a covering.
Proof. If n = 0, then it follows easily that V = U [0] (see below). In this
case Hom(V, K)
0
= Hom(U, K)
0
K
0
. The result, in this case, then follows from
Lemma 2.5.
Let a : [n] V be the morphism associated to x as in Simplicial, Lemma 11.3.
Let us write [n] = i
(n1)!
sk
n1
[n] for the (n 1)-skeleton of [n]. Let b :
[n] U be the restriction of a to the (n 1) skeleton of [n]. By Simplicial,
10 HYPERCOVERINGS
Lemma 18.7 we have V = U
[n]
[n]. By Simplicial, Lemma 14.6 we get that
Hom(V, K)
0

Hom(U, K)
0

Hom([n], K)
0

Hom([n], K)
0
is a bre product square. Thus it suces to show that the bottom horizontal arrow
is a covering. By Simplicial, Lemma 18.11 this arrow is identied with
K
n
(cosk
n1
sk
n1
K)
n
and hence is a covering by denition of a hypercovering.
Lemma 5.2. Let C be a site with bre products. Let X be an object of C. Let K
be a hypercovering of X. Let U V be simplicial sets, with U
n
, V
n
nite nonempty
for all n. Assume that U and V have nitely many nondegenerate simplices. Then
the morphism
Hom(V, K)
0
Hom(U, K)
0
of SR(C, X) is a covering.
Proof. By Lemma 5.1 above, it suces to prove a simple lemma about inclusions
of simplicial sets U V as in the lemma. And this is exactly the result of Simplicial,
Lemma 18.8.
6. Homotopies
Let C be a site with bre products. Let X be an object of C. Let L be a simplicial
object of SR(C, X). According to Simplicial, Lemma 14.4 there exists an object
Hom([1], L) in the category Simp(SR(C, X)) which represents the functor
T Mor
Simp(SR(C,X))
([1] T, L)
There is a canonical morphism
Hom([1], L) L L
coming from e
i
: [0] [1] and the identication Hom([0], L) = L.
Lemma 6.1. Let C be a site with bre products. Let X be an object of C. Let L be
a simplicial object of SR(C, X). Let n 0. Consider the commutative diagram
(6.1.1) Hom([1], L)
n+1

(cosk
n
sk
n
Hom([1], L))
n+1

(L L)
n+1

(cosk
n
sk
n
(L L))
n+1
coming from the morphism dened above. We can identify the terms in this diagram
as follows, where [n+1] = i
n!
sk
n
[n+1] is the n-skeleton of the (n+1)-simplex:
Hom([1], L)
n+1
= Hom([1] [n + 1], L)
0
(cosk
n
sk
n
Hom([1], L))
n+1
= Hom([1] [n + 1], L)
0
(L L)
n+1
= Hom(([n + 1] [n + 1], L)
0
(cosk
n
sk
n
(L L))
n+1
= Hom([n + 1] [n + 1], L)
0
HYPERCOVERINGS 11
and the morphism between these objects of SR(C, X) come from the commutative
diagram of simplicial sets
(6.1.2) [1] [n + 1] [1] [n + 1]

[n + 1] [n + 1]

[n + 1] [n + 1]

Moreover the bre product of the bottom arrow and the right arrow in (6.1.1) is
equal to
Hom(U, L)
0
where U [1] [n + 1] is the smallest simplicial subset such that both [n +
1] [n + 1] and [1] [n + 1] map into it.
Proof. The rst and third equalities are Simplicial, Lemma 14.4. The second and
fourth follow from the cited lemma combined with Simplicial, Lemma 18.11. The
last assertion follows from the fact that U is the push-out of the bottom and right
arrow of the diagram (6.1.2), via Simplicial, Lemma 14.6. To see that U is equal
to this push-out it suces to see that the intersection of [n + 1] [n + 1] and
[1] [n+1] in [1] [n+1] is equal to [n+1] [n+1]. This we leave
to the reader.
Lemma 6.2. Let C be a site with bre products. Let X be an object of C. Let K, L
be hypercoverings of X. Let a, b : K L be morphisms of hypercoverings. There
exists a morphism of hypercoverings c : K

K such that a c is homotopic to

b c.
Proof. Consider the following commutative diagram
K

def
c

K
(LL)
Hom([1], L)

Hom([1], L)

K
(a,b)

L L
By the functorial property of Hom([1], L) the composition of the horizontal mor-
phisms corresponds to a morphism K

[1] L which denes a homotopy between

c a and c b. Thus if we can show that K

is a hypercovering of X, then we
obtain the lemma. To see this we will apply Lemma 4.1 to the pair of morphisms
K L L and Hom([1], L) L L. Condition (1) of Lemma 4.1 is statis-
ed. Condition (2) of Lemma 4.1 is true because Hom([1], L)
0
= L
1
, and the
morphism (d
1
0
, d
1
1
) : L
1
L
0
L
0
is a covering of SR(C, X) by our assumption
that L is a hypercovering. To prove condition (3) of Lemma 4.1 we use Lemma 6.1
above. According to this lemma the morphism of condition (3) of Lemma 4.1 is
the morphism
Hom([1] [n + 1], L)
0
Hom(U, L)
0
where U [1] [n + 1]. According to Lemma 5.2 this is a covering and hence
the claim has been proven.
Remark 6.3. Note that the crux of the proof is to use Lemma 5.2. This lemma
is completely general and does not care about the exact shape of the simplicial
sets (as long as they have only nitely many nondegenerate simplices). It seems
12 HYPERCOVERINGS
altogether reasonable to expect a result of the following kind: Given any mor-
phism a : K [k] L, with K and L hypercoverings, there exists a morphism
of hypercoverings c : K

K and a morphism g : K

[k] L such that

g|
K

[k]
= a (c id
[k]
). In other words, the category of hypercoverings is in
a suitable sense contractible.
7. Cech cohomology associated to hypercoverings
Let C be a site with bre products. Let X be an object of C. Consider a presheaf
of abelian groups F on the site C. It denes a functor
F : SR(C, X)
opp
Ab
{U
i
X}
iI

iI
F(U
i
)
Thus a simplicial object K of SR(C, X) is turned into a cosimplicial object F(K)
of Ab. In this situation we dene

H
i
(K, F) = H
i
(s(F(K))).
Recall that s(F(K)) is the cochain complex associated to the cosimplicial abelian
group F(K), see Simplicial, Section 22. In this section we prove analogues of
some of the results for Cech cohomology of open coverings proved in Cohomology,
Sections 9, 10 and 11.
Lemma 7.1. Let C be a site with bre products. Let X be an object of C. Let K be
a hypercovering of X. Let F be a sheaf of abelian groups on C. Then

H
0
(K, F) =
F(X).
Proof. We have

H
0
(K, F) = Ker(F(K
0
) F(K
1
))
Write K
0
= {U
i
X}. It is a covering in the site C. As well, we have that K
1

K
0
K
0
is a covering in SR(C, X). Hence we may write K
1
=
i0,i1I
{V
i0i1j
X}
so that the morphism K
1
K
0
K
0
is given by coverings {V
i0i1j
U
i0

X
U
i1
}
of the site C. Thus we can further identify

H
0
(K, F) = Ker(

i
F(U
i
)

i0i1j
F(V
i0i1j
))
with obvious map. The sheaf property of F implies that

H
0
(K, F) = H
0
(X, F).
In fact this property characterizes the abelian sheaves among all abelian presheaves
on C of course. The analogue of Cohomology, Lemma 7.2 in this case is the following.
Lemma 7.2. Let C be a site with bre products. Let X be an object of C. Let K
be a hypercovering of X. Let I be an injective sheaf of abelian groups on C. Then

H
p
(K, I) =
_
I(X) if p = 0
0 if p > 0
HYPERCOVERINGS 13
Proof. Observe that for any object Z = {U
i
X} of SR(C, X) and any abelian
sheaf F on C we have
F(Z) =

F(U
i
)
=

Mor
PSh(C)
(h
Ui
, F)
= Mor
PSh(C)
(F(Z), F)
= Mor
PAb(C)
(Z
F(Z)
, F)
= Mor
Ab(C)
(Z
#
F(Z)
, F)
Thus we see, for any simplicial object K of SR(C, X) that we have
(7.2.1) s(F(K)) = Hom
Ab(C)
(s(Z
#
K
), F)
see Denition 3.1 for notation. Now, we know that s(Z
#
K
) is quasi-isomorphic to
Z
#
X
if K is a hypercovering, see Lemma 3.5. We conclude that if I is an injective
abelian sheaf, and K a hypercovering, then the complex s(I(K)) is acyclic except
possibly in degree 0. In other words, we have

H
i
(K, I) = 0
for i > 0. Combined with Lemma 7.1 the lemma is proved.
Next we come to the analogue of Cohomology, Lemma 7.3. To state it we need to
introduce a little more notation. Let C be a site with bre products. Let F be a
sheaf of abelian groups on C. The symbol H
i
(F) indicates the presheaf of abelian
groups on C which is dened by the rule
H
i
(F) : U H
i
(U, F)
where U ranges over the objects of C.
Lemma 7.3. Let C be a site with bre products. Let X be an object of C. Let K
be a hypercovering of X. Let F be a sheaf of abelian groups on C. There is a map
s(F(K)) R(X, F)
in D
+
(Ab) functorial in F, which induces natural transformations

H
i
(K, ) H
i
(X, )
as functors Ab(C) Ab. Moreover, there is a spectral sequence (E
r
, d
r
)
r0
with
E
p,q
2
=

H
p
(K, H
q
(F))
converging to H
p+q
(X, F). This spectral sequence is functorial in F and in the
hypercovering K.
Proof. We could prove this by the same method as employed in the corresponding
lemma in the chapter on cohomology. Instead let us prove this by a double complex
argument.
Choose an injective resolution F I

in the category of abelian sheaves on C.

Consider the double complex A
,
with terms
A
p,q
= I
q
(K
p
)
where the dierential d
p,q
1
: A
p,q
A
p+1,q
is the one coming from the dierential
I
p
I
p+1
and the dierential d
p,q
2
: A
p,q
A
p,q+1
is the one coming from the
dierential on the complex s(I
p
(K)) associated to the cosimplicial abelian group
14 HYPERCOVERINGS
I
p
(K) as explained above. As usual we denote sA

the simple complex associated

to the double complex A
,
. We will use the two spectral sequences (

E
r
,

d
r
) and
(

E
r
,

d
r
) associated to this double complex, see Homology, Section 19.
By Lemma 7.2 the complexes s(I
p
(K)) are acyclic in positive degrees and have
H
0
equal to I
p
(X). Hence by Homology, Lemma 19.6 and its proof the spectral
sequence (

E
r
,

d
r
) degenerates, and the natural map
I

(X) sA

is a quasi-isomorphism of complexes of abelian groups. In particular we conclude

that H
n
(sA

) = H
n
(X, F).
The map s(F(K)) R(X, F) of the lemma is the composition of the natural
map s(F(K)) sA

followed by the inverse of the displayed quasi-isomorphism

above. This works because I

(X) is a representative of R(X, F).

Consider the spectral sequence (

E
r
,

d
r
)
r0
. By Homology, Lemma 19.3 we see
that

E
p,q
2
= H
p
II
(H
q
I
(A
,
))
In other words, we rst take cohomology with respect to d
1
which gives the groups

E
p,q
1
= H
p
(F)(K
q
). Hence it is indeed the case (by the description of the dier-
ential

d
1
) that

E
p,q
2
=

H
p
(K, H
q
(F)). And by the other spectral sequence above
we see that this one converges to H
n
(X, F) as desired.
We omit the proof of the statements regarding the functoriality of the above con-
structions in the abelian sheaf F and the hypercovering K.
8. Cohomology and hypercoverings
Let C be a site with bre products. Let X be an object of C. Let F be a sheaf of
abelian groups on C. Let K, L be hypercoverings of X. If a, b : K L are homo-
topic maps, then F(a), F(b) : F(K) F(L) are homotopic maps, see Simplicial,
Lemma 25.4. Hence have the same eect on cohomology groups of the associated
cochain complexes, see Simplicial, Lemma 25.6. We are going to use this to dene
the colimit over all hypercoverings.
Let us temporarily denote HC(C, X) the category of hypercoverings of X. We have
seen that this is a category and not a big category, see Lemma 2.8. This will
be the index category for our diagram, see Categories, Section 13 for notation.
Consider the diagram

H
i
(, F) : HC(C, X) Ab.
By Lemma 4.2 and Lemma 6.2, and the remark on homotopies above, this diagram
is directed, see Categories, Denition 17.1. Thus the colimit

H
i
HC
(X, F) = colim
KHC(C,X)

H
i
(K, F)
has a particularly simple discription (see location cited).
Theorem 8.1. Let C be a site with bre products. Let X be an object of C. Let
i 0. The functors
Ab(C) Ab
F H
i
(X, F)
F

H
i
HC
(X, F)
HYPERCOVERINGS 15
are canonically isomorphic.
Proof using spectral sequences. Suppose that H
p
(X, F) for some p 0.
Let us show that is in the image of the map

H
p
(X, F) H
p
(X, F) of Lemma
7.3 for some hypercovering K of X.
This is true if p = 0 by Lemma 7.1. If p = 1, choose a Cech hypercovering K of
X as in Example 2.7 starting with a covering K
0
= {U
i
X} in the site C such
that |
Ui
= 0, see Cohomology on Sites, Lemma 8.3. It follows immediately from
the spectral sequence in Lemma 7.3 that comes from an element of

H
1
(K, F) in
this case. In general, choose any hypercovering K of X such that maps to zero in
H
p
(F)(K
0
) (using Example 2.7 and Cohomology on Sites, Lemma 8.3 again). By
the spectral sequence of Lemma 7.3 the obstruction for to come from an element of

H
p
(K, F) is a sequence of elements
1
, . . . ,
p1
with
q


H
pq
(K, H
q
(F)) (more
precisely the images of the
q
in certain subquotients of these groups).
We can inductively replace the hypercovering K by renements such that the ob-
structions
1
, . . . ,
p1
restrict to zero (and not just the images in the subquotients
so no subtlety here). Indeed, suppose we have already managed to reach the sit-
uation where
q+1
, . . . ,
p1
are zero. Note that
q


H
pq
(K, H
q
(F)) is the class
of some element

q
H
q
(F)(K
pq
) =

H
q
(U
i
, F)
if K
pq
= {U
i
X}
iI
. Let
q,i
be the component of

q
in H
q
(U
i
, F). As
q 1 we can use Cohomology on Sites, Lemma 8.3 yet again to choose coverings
{U
i,j
U
i
} of the site such that each restriction
q,i
|
Ui,j
= 0. Consider the object
Z = {U
i,j
X} of the category SR(C, X) and its obvious morphism u : Z K
pq
.
It is clear that u is a covering, see Denition 2.4. By Lemma 4.3 there exists a
morphism L K of hypercoverings of X such that L
pq
K
pq
factors through
u. Then clearly the image of
q
in H
q
(F)(L
pq
). is zero. Since the spectral
sequence of Lemma 7.3 is functorial this means that after replacing K by L we
reach the situation where
q
, . . . ,
p1
are all zero. Continuing like this we end up
with a hypercovering where they are all zero and hence is in the image of the map

H
p
(X, F) H
p
(X, F).
Suppose that K is a hypercovering of X, that

H
p
(K, F) and that the image of
under the map

H
p
(X, F) H
p
(X, F) of Lemma 7.3 is zero. To nish the proof
of the theorem we have to show that there exists a morphism of hypercoverings
L K such that restricts to zero in

H
p
(L, F). By the spectral sequence of
Lemma 7.3 the vanishing of the image of in H
p
(X, F) means that there exist
elements
1
, . . . ,
p2
with
q


H
p1q
(K, H
q
(F)) (more precisely the images of
these in certain subquotients) such that the images d
p1q,q
q+1

q
(in the spectral
sequence) add up to . Hence by exacly the same mechanism as above we can nd
a morphism of hypercoverings L K such that the restrictions of the elements

q
, q = 1, . . . , p 2 in

H
p1q
(L, H
q
(F)) are zero. Then it follows that is zero
since the morphism L K induces a morphism of spectral sequences according to
Lemma 7.3.
Proof without using spectral sequences. We have seen the result for i = 0,
see Lemma 7.1. We know that the functors H
i
(X, ) form a universal -functor,
see Derived Categories, Lemma 19.4. In order to prove the theorem it suces to
show that the sequence of functors

H
i
HC
(X, ) forms a -functor. Namely we know
16 HYPERCOVERINGS
that Cech cohomology is zero on injective sheaves (Lemma 7.2) and then we can
apply Homology, Lemma 9.4.
Let
0 F G H 0
be a short exact sequence of abelian sheaves on C. Let

H
p
HC
(X, H). Choose
a hypercovering K of X and an element H(K
p
) representing in cohomology.
There is a corresponding exact sequence of complexes
0 s(F(K)) s(G(K)) s(H(K))
but we are not assured that there is a zero on the right also and this is the only
thing that prevents us from dening () by a simple application of the snake
lemma. Recall that
H(K
p
) =

H(U
i
)
if K
p
= {U
i
X}. Let =

i
with
i
H(U
i
). Since G H is a surjection
of sheaves we see that there exist coverings {U
i,j
U
i
} such that
i
|
Ui,j
is the
image of some element
i,j
G(U
i,j
). Consider the object Z = {U
i,j
X} of the
category SR(C, X) and its obvious morphism u : Z K
p
. It is clear that u is a
covering, see Denition 2.4. By Lemma 4.3 there exists a morphism L K of
hypercoverings of X such that L
p
K
p
factors through u. After replacing K by L
we may therefore assume that is the image of an element G(K
p
). Note that
d() = 0, but not necessarily d() = 0. Thus d() F(K
p+1
) is a cocycle. In this
situation we dene () as the class of the cocycle d() in

H
p+1
HC
(X, F).
At this point there are several things to verify: (a) () does not depend on the
choice of , (b) () does not depend on the choice of the hypercovering L K such
that lifts, and (c) () does not depend on the initial hypercovering and chosen
to represent . We omit the verication of (a), (b), and (c); the independence of
the choices of the hypercoverings really comes down to Lemmas 4.2 and 6.2. We
also omit the verication that is functorial with respect to morphisms of short
exact sequences of abelian sheaves on C.
Finally, we have to verify that with this denition of our short exact sequence of
abelian sheaves above leads to a long exact sequence of Cech cohomology groups.
First we show that if () = 0 (with as above) then is the image of some
element

H
p
HC
(X, G). Namely, if () = 0, then, with notation as above, we
see that the class of d() is zero in

H
p+1
HC
(X, F). Hence there exists a morphism of
hypercoverings L K such that the restriction of d() to an element of F(L
p+1
)
is equal to d() for some F(L
p
). This implies that |
Lp
+ form a cocycle, and
determine a class

H
p
(L, G) which maps to as desired.
We omit the proof that if

H
p+1
HC
(X, F) maps to zero in

H
p+1
HC
(X, G), then it is
equal to () for some

H
p
HC
(X, H).
9. Hypercoverings of spaces
The theory above is mildly interesting even in the case of topological spaces. In
this case we can work out what is a hypercovering and see what the result actually
says.
HYPERCOVERINGS 17
Let X be a topological space. Consider the site T
X
of Sites, Example 6.4. Recall
that an object of T
X
is simply an open of X and that morphisms of T
X
correspond
simply to inclusions. So what is a hypercovering of X for the site T
X
?
Let us rst unwind Denition 2.1. An object of SR(C, X) is simply given by a set
I and for each i I an open U
i
X. Let us denote this by {U
i
}
iI
since there can
be no confusion about the morphism U
i
X. A morphism {U
i
}
iI
{V
j
}
jJ
between two such objects is given by a map of sets : I J such that U
i
V
(i)
for all i I. When is such a morphism a covering? This is the case if and only if
for every j J we have V
j
=

iI, (i)=j
U
i
(and is a covering in the site T
X
).
Using the above we get the following description of a hypercovering in the site T
X
.
A hypercovering of X in T
X
is given by the following data
(1) a simplicial set I (see Simplicial, Section 11), and
(2) for each n 0 and every i I
n
an open set U
i
X.
We will denote such a collection of data by the notation (I, {U
i
}). In order for this
to be a hypercovering of X we require the following properties
for i I
n
and 0 a n + 1 we have U
i
U
d
n
a
(i)
,
for i I
n
and 0 a n we have U
i
= U
s
n
a
(i)
,
we have
(9.0.1) X =
_
iI0
U
i
,
for every i
0
, i
1
I
0
, we have
(9.0.2) U
i0
U
i1
=
_
iI1, d
1
0
(i)=i0, d
1
1
(i)=i1
U
i
,
for every n 1 and every (i
0
, . . . , i
n+1
) (I
n
)
n+2
such that d
n
b1
(i
a
) =
d
n
a
(i
b
) for all 0 a < b n + 1 we have
(9.0.3) U
i0
. . . U
in+1
=
_
iIn+1, d
n+1
a
(i)=ia, a=0,...,n+1
U
i
,
each of the open coverings (9.0.1), (9.0.2), and (9.0.3) is an element of
Cov(T
X
) (this is a set theoretic condition, bounding the size of the index
sets of the coverings).
Condititions (9.0.1) and (9.0.2) should be familiar from the chapter on sheaves on
spaces for example, and condition (9.0.3) is the natural generalization.
Remark 9.1. One feature of this description is that if one of the multiple in-
tersections U
i0
. . . U
in+1
is empty then the covering on the right hand side
may be the empty covering. Thus it is not automatically the case that the maps
I
n+1
(cosk
n
sk
n
I)
n+1
are surjective. This means that the geometric realization
of I may be an interesting (non-contractible) space.
In fact, let I

n
I
n
be the subset consisting of those simplices i I
n
such that
U
i
= . It is easy to see that I

I is a subsimplicial set, and that (I

, {U
i
}) is
a hypercovering. Hence we can always rene a hypercovering to a hypercovering
where none of the opens U
i
is empty.
Remark 9.2. Let us repackage this information in yet another way. Namely,
suppose that (I, {U
i
}) is a hypercovering of the topological space X. Given this
data we can construct a simplicial toplogical space U

by setting
U
n
=

iIn
U
i
,
18 HYPERCOVERINGS
and where for given : [n] [m] we let morphisms U() : U
n
U
m
be the mor-
phism coming from the inclusions U
i
U
(i)
for i I
n
. This simplicial topological
space comes with an augmentation : U

X. With this morphism the simplicial

space U

becomes a hypercovering of X along which one has cohomological descent

in the sense of [MA71, Expose Vbis]. In other words, H
n
(U

F) = H
n
(X, F).
(Insert future reference here to cohomology over simplicial spaces and cohomologi-
cal descent formulated in those terms.) Suppose that F is an abelian sheaf on X.
In this case the spectral sequence of Lemma 7.3 becomes the spectral sequence with
E
1
-term
E
p,q
1
= H
q
(U
p
,

q
F) H
p+q
(U

F) = H
p+q
(X, F)
comparing the total cohomology of

F to the cohomology groups of F over the

pieces of U

. (Insert future reference to this spectral sequence here.)

In topology we often want to nd hypercoverings of X which have the property that
all the U
i
come from a given basis for the topology of X and that all the coverings
(9.0.2) and (9.0.3) are from a given conal collection of coverings. Here are two
example lemmas.
Lemma 9.3. Let X be a topological space. Let B be a basis for the topology of X.
There exists a hypercovering (I, {U
i
}) of X such that each U
i
is an element of B.
Proof. Let n 0. Let us say that an n-truncated hypercovering of X is given by
an n-truncated simplicial set I and for each i I
a
, 0 a n an open U
i
of X such
that the conditions dening a hypercovering hold whenever they make sense. In
other words we require the inclusion relations and covering conditions only when
all simplices that occur in them are a-simplices with a n. The lemma follows if
we can prove that given a n-truncated hypercovering (I, {U
i
}) with all U
i
B we
can extend it to an (n+1)-truncated hypercovering without adding any a-simplices
for a n. This we do as follows. First we consider the (n+1)-truncated simplicial
set I

dened by I

= sk
n+1
(cosk
n
I). Recall that
I

n+1
=
_
(i
0
, . . . , i
n+1
) (I
n
)
n+2
such that
d
n
b1
(i
a
) = d
n
a
(i
b
) for all 0 a < b n + 1
_
If i

n+1
is degenerate, say i

= s
n
a
(i) then we set U
i
= U
i
(this is forced on us
anyway by the second condition). We also set J
i
= {i

} in this case. If i

n+1
is
nondegerate, say i

= (i
0
, . . . , i
n+1
), then we choose a set J
i
and an open covering
(9.3.1) U
i0
. . . U
in+1
=
_
iJ
i

U
i
,
with U
i
B for i J
i
. Set
I
n+1
=

n+1
J
i

There is a canonical map : I

n+1
I

n+1
which is a bijection over the set of
degenerate simplices in I

n+1
by construction. For i I
n+1
we dene d
n+1
a
(i) =
d
n+1
a
((i)). For i I
n
we dene s
n
a
(i) I
n+1
as the unique simplex lying over
the degenerate simplex s
n
a
(i) I

n+1
. We omit the verication that this denes an
(n + 1)-truncated hypercovering of X.
Lemma 9.4. Let X be a topological space. Let B be a basis for the topology of X.
Assume that
(1) X is quasi-compact,
HYPERCOVERINGS 19
(2) each U B is quasi-compact open, and
(3) the intersection of any two quasi-compact opens in X is quasi-compact.
Then there exists a hypercovering (I, {U
i
}) of X with the following properties
(1) each U
i
is an element of the basis B,
(2) each of the I
n
is a nite set, and in particular
(3) each of the coverings (9.0.1), (9.0.2), and (9.0.3) is nite.
Proof. This follows directly from the construction in the proof of Lemma 9.3 if we
choose nite coverings by elements of B in (9.3.1). Details omitted.
10. Other chapters
(1) Introduction
(2) Conventions
(3) Set Theory
(4) Categories
(5) Topology
(6) Sheaves on Spaces
(7) Commutative Algebra
(8) Sites and Sheaves
(9) Homological Algebra
(10) Derived Categories
(11) More on Algebra
(12) Simplicial Methods
(13) Sheaves of Modules
(14) Modules on Sites
(15) Injectives
(16) Cohomology of Sheaves
(17) Cohomology on Sites
(18) Hypercoverings
(19) Schemes
(20) Constructions of Schemes
(21) Properties of Schemes
(22) Morphisms of Schemes
(23) Coherent Cohomology
(24) Divisors
(25) Limits of Schemes
(26) Varieties
(27) Chow Homology
(28) Topologies on Schemes
(29) Descent
(30) Adequate Modules
(31) More on Morphisms
(32) More on Flatness
(33) Groupoid Schemes
(34) More on Groupoid Schemes
(35)

Etale Morphisms of Schemes
(36)

Etale Cohomology
(37) Algebraic Spaces
(38) Properties of Algebraic Spaces
(39) Morphisms of Algebraic Spaces
(40) Decent Algebraic Spaces
(41) Topologies on Algebraic Spaces
(42) Descent and Algebraic Spaces
(43) More on Morphisms of Spaces
(44) Quot and Hilbert Spaces
(45) Spaces over Fields
(46) Cohomology of Algebraic Spaces
(47) Stacks
(48) Formal Deformation Theory
(49) Groupoids in Algebraic Spaces
(50) More on Groupoids in Spaces
(51) Bootstrap
(52) Examples of Stacks
(53) Quotients of Groupoids
(54) Algebraic Stacks
(55) Sheaves on Algebraic Stacks
(56) Criteria for Representability
(57) Properties of Algebraic Stacks
(58) Morphisms of Algebraic Stacks
(59) Introducing Algebraic Stacks
(60) Examples
(61) Exercises
(62) Guide to Literature
(63) Desirables
(64) Coding Style
(65) GNU Free Documentation Li-
cense
(66) Auto Generated Index
20 HYPERCOVERINGS
References
[MA71] J.L. Verdier M. Artin, A. Grothendieck, Theorie de topos et cohomologie etale des
schemas i, ii, iii, Lecture Notes in Mathematics, vol. 269, 270, 305, Springer, 1971.