1304 2520

a
r
X
i
v
:
1
3
0
4
.
2
5
2
0
v
1

[
m
a
t
h
.
A
G
]

9

A
p
r

2
0
1
3
A FUNCTORIAL FORMALISM FOR QUASI-COHERENT SHEAVES
ON A GEOMETRIC STACK
LEOVIGILDO ALONSO TARRO, ANA JEREMAS LPEZ, MARTA PREZ RODRGUEZ,
AND MARA J. VALE GONSALVES
ABSTRACT. A geometric stack is a quasi-compact and semi-separated al-
gebraic stack. We prove that the quasi-coherent sheaves on the small at
topology, Cartesian presheaves on the underlying category, and comodules
over a Hopf algebroid associated to a presentation of a geometric stack
are equivalent categories. As a consequence, we show that the category
of quasi-coherent sheaves on a geometric stack is a Grothendieck category.
We associate, in a 2-functorial way, to a 1-morphism of geometric stacks
f : XY, an adjunction f
for the corresponding categories of quasi-

coherent sheaves that agrees with the classical one dened for schemes.
This construction is described both geometrically in terms of the small at
site and algebraically in terms of the Hopf algebroid.
CONTENTS
Introduction 1
1. Sheaves on ringed sites 4
2. Cartesian presheaves on afne schemes 8
3. Quasi-coherent sheaves on geometric stacks 11
4. Descent of quasi-coherent sheaves on geometric stacks 17
5. Representing quasi-coherent sheaves by comodules 22
6. Properties and functoriality of quasi-coherent sheaves 29
7. Describing functoriality via comodules 36
8. Deligne-Mumford stacks and functoriality for the tale topology 40
References 45
INTRODUCTION
If one looks at the literature for the notion of quasi-coherent sheaf on an
algebraic stack, one nds a somewhat confusing situation. For starters, the
Date: April 10, 2013.
2000 Mathematics Subject Classication. 14A20 (primary); 14F05, 18F20 (secondary).
This work has been partially supported by Spains MEC and E.U.s FEDER re-
search project MTM2008-03465 and MTM2011-26088, together with Xunta de Galicias
PGIDIT10PXIB207144PR .
1
2 L. ALONSO, A. JEREMAS, M. PREZ, AND M. J. VALE
very notion of quasi-coherent sheaf does not refer to a ringed site but is of-
ten mistook with a related notion, that of Cartesian presheaf. After [AJPV],
we became interested in the properties of its derived category. This is our
approach to develop a sound basis for this study.
As a starting point, the denition of quasi-coherent sheaf should be given
in the setting of a ringed site by an extension of the original denition given
in [EGA I, (5.1.3)]. After our research took off, we discovered that the Stacks
Project [SP] uses this kind of denition. The main technical problem is the
lack of functoriality of the lisse-tale site. The Stacks Project solution is to
work over the big at sites. We do not follow this path because the category
of sheaves of modules over a big (ringed) site does not possess a generator.
This paper was developed independently of [SP], so we have tried to refer
mainly to [LMB] for the basics on stacks.
We assume always that the stacks we consider are semi-separated, i.e. the
diagonal is an afne morphism. On schemes this hypothesis is associated to a
nice cohomological behavior. The stacks that are furthermore quasi-compact
are called geometric stacks in [L] and this will be our standing assumption.
A useful feature of geometric stacks is that they may be represented by a
purely algebraic object, a Hopf algebroid. For such a stack X we consider a
certain variant of the small at site, that we will denote by X
fppf
(see 3.5).
Its associated topos is not functorial, but still, it has enough functoriality
properties to allow us to develop a useful formalism. Notice that it is ner
than the classical lisse-tale site, but has the advantage that a presentation
yields a covering for every object in the small at topology. This allows for
simpler proofs and the category of quasi-coherent sheaves does not change.
In the afne case there are three ways of understanding a quasi-coherent
sheaf, namely,
as a sheaf of modules such that locally admits a presentation,
as a cartesian presheaf
1
, and
as a module over the ring of global sections.
These three aspects are, in some sense, present also in the global case. In this
paper we show that they arise quite naturally when we consider a geometric
stack. For the small at site, cartesian presheaves of modules are sheaves (in
fact, for any topology coarser than the fpqc) and moreover they are the same
as quasi-coherent sheaves. This is proved in Theorem 3.11. A geometric stack
may be described by a Hopf algebroid, as the stackication of its associated
afne groupoid scheme. We prove that there is an equivalence of categories
between comodules over a Hopf algebroid and quasi-coherent sheaves on its
associated geometric stack.
This equivalence is developed in several steps. In section 4, we see that
quasi-coherent sheaves on a geometric stack correspond to descent data for
quasi-coherent sheaves on an afne cover (Theorem 4.6). Incidentally, this is
1
This property corresponds essentially to the fact that a localization of a module in an
element may be computed as a tensor product.
QC-SHEAVES ON A GEOMETRIC STACK 3
the denition of quasi-coherent sheaf considered in the classic literature, e.g.
[V1]. Once we are in the afne setting, we see that a descent datum corre-
sponds to certain supplementary structure on its module of global sections.
By an algebraic procedure, in section 5, we show that this structure can be
transformed into the structure of a comodule, establishing a further equiva-
lence of categories. All these equivalences combined provide the equivalence
between quasi-coherent sheaves on the geometric stack and comodules over
the Hopf algebroid (Corollary 5.9). An analogous result is due by Hovey un-
der a different setting, as we will discuss later. An important consequence of
this equivalence is that the category of quasi-coherent sheaves on a geometric
stack is Grothendieck, i.e. abelian, with exact ltered direct limits and a set
of generators.
The motivation for the results in the present paper is to have a convenient
formalism for doing cohomology. So, it is crucial to have a good functoriality.
We devote section 6 to this issue. Specically, we show that a map between
geometric stacks induce a pair of adjoint functors between the corresponding
categories of quasi-coherent sheaves. This construction is not straightfor-
ward because, as we remarked before, there is no underlying map of toposes,
so the denition of the direct image has extra complications. Moreover, we
show that these constructions are 2-functorial. One needs to check in ad-
dition that 2-cells induce a natural transformation between their associated
functors, a feature of stacks.
In section 7, we express the adjunction in a purely algebraic way. The
benet of this result is to put ourselves in the most convenient setting for
approaching cohomological questions. It is noteworthy that for a 1-morphism
of geometric stacks f : XY, the functor f
has a nicer geometric description

while its adjoint f
is easier to describe algebraically.

In the nal section we compare our functoriality formalism with the clas-
sical one. Let us recall that the category quasi-coherent sheaves on a scheme
does not change if we replace the Zariski by the tale topology. Moreover,
the tale topology gives a reasonable topos for a geometric Deligne-Mumford
stack. So in this setting there is a couple of adjoint functors that restrict to
the classical ones in the scheme case and agree with the ones just dened in
the general case of a geometric stack. This is shown in Propositions 8.13, 8.14
and Corollary 8.15.
In future work, we plan to apply the results obtained here to study the
cohomology of geometric stacks.
Finally, let us compare our approach with previously published results.
The classical reference [LMB] denes quasi-coherent sheaf as a mixture
of Cartesian sheaf plus a local condition of quasi-coherence. Besides, it over-
looks the lack of functoriality of the lisse-tale site. Olsson provides a remedy
[O] but his pull-back functor has a complicated description through simpli-
cial sites. Another approach is given by Hovey [Ho1], [Ho2]. His denition of
quasi-coherent sheaf is in fact the notion of Cartesian presheaf, as he consid-
ers mostly the discrete topology. So his theorem on the equivalence of quasi-
coherent sheaves and comodules over a Hopf algebroid is related to ours, but
posed in a different framework. This point of view is also embraced by Goerss
in [Go] who denes quasi-coherent sheaves trough the Cartesian condition.
We have to mention also Pribbles Ph. D. thesis [P]. We have used several
times the exact sequence associated to the monad (or triple) corresponding to
a presentation. This has allowed us to simplify some arguments.
In [SP], they put very few restrictions on the diagonal, which makes the
situation more complicated. In the general setting a sheaf is quasi-coherent
if, and only if, is Cartesian and locally quasi-coherent. As we have already re-
marked, they work over big sites for which functoriality is easy but sacrices
other useful features for the usual constructions in cohomology.
Acknowledgements. We thank M. Olsson for the useful exchange concern-
ing the algebraic structure of inverse images from which the consideration of
Lemma 1.8 arose.
1. SHEAVES ON RINGED SITES
1.0. In this paper we will use as underlying axiomatics that of von Neu-
mann, Gdel and Bernays using sets and classes. A category C has a class
of objects and a class of arrows but the class of maps between two objets
A, B C, denoted Hom
C
(A, B), is always assumed to be a set. For readers
fond of Bourbaki-Grothendieck universes (and for comparison with results in
[SGA 4
1
]) x two universes U V and identify elements of U with sets and
elements of V with classes.
1.1. Ringed categories. Let C be a small
2
category. We will denote by
Pre(C) :=Fun(C
o
, Ab) the category of contravariant functors from C to abelian
groups. We will usually refer to objects in this category as presheaves on C.
We will say that (C, O) is a ringed category if O is a ring object in Pre(C). In
other words O : C
o
Ring is a presheaf, or else O is a presheaf together with
a couple of maps : O O O and 1: Z
C
O, with Z
C
denoting the constant
presheaf with value Z; this data makes commutative the usual diagrams
expressing the associativity and commutativity of the product and the fact
that 1 is a unit.
1.2. Modules. Let (C, O) be a ringed category we will denote by Pre(C, O)-Mod
or simply by Pre-O-Mod, if no confusion arises, the category of presheaves
M Pre(C) equipped with an action of O, O M M, that makes commu-
tative the diagrams expressing the fact that the sections of M are modules
over the sections of O and the restriction maps are linear. We will say that
M is an O-Module.
2
Its class of objects is a set, not a proper class.
1.3. Cartesian presheaves. Let (C, O) be a ringed category. We will say
that an O-module M is a Cartesian presheaf if for every morphism f : X Y
in C the O(X)-linear map
M( f )
a
: O(X)
O(Y)
M(Y) M(X)
adjoint to the O(Y)-linear map M( f ) : M(Y) M(X) is an isomorphism. We
will denote the category of Cartesian presheaves on (C, O) by Crt(C, O).
1.4. Ringed sites and topos. As usual we will call a small category C a site
if it is equipped with a Grothendieck topology [SGA 4
1
, Exp. II 1]. If it is
necessary to make it explicit, we will denote the site by (C, ), otherwise we
will keep the notation C. A site has associated a topos, its category of sheaves
of sets. For a site (C, ) we will denote its associated topos by C
. We may
also consider sheaves with additional algebraic structures, in particular, a
sheaf of rings O. We will call the pair (C, O) a ringed site and (C
, O) a ringed
topos. We will always refer with the notation (C
, O)-Mod or simply O-Mod to

the category of sheaves of modules over the sheaf of rings O, and not to the
biggest category of presheaves of modules.
1.5. Let C be a site and X and object in C. We will consider the category
whose objects are morphisms u : U X, and morphisms are commutative
triangles.
3
We will denote such an object by (U, u), or simply by U if no con-
fusion arises. The category C/X is a site. The coverings of the corresponding
topology of an object U are the families of morphisms { f
i
: U
i
U} in C/X
that are coverings in C. If (C, O) is a ringed site, then (C/X, O|
X
) is a ringed
site with the induced sheaf of rings O|
X
given by O|
X
(U, u) =O(U). If M is a
sheaf of O-Modules on the site (C, O) we dene the sheaf restriction of M to
C/X, denoted M|
X
, as the sheaf of O|
X
-Modules on the site (C/X, O|
X
) given
by M|
X
(U, u) =M(U).
1.6. Quasi-coherent sheaves. Let (C, O) be a ringed site with a topology .
We will say that a sheaf of O-Modules M is quasi-coherent if for every object
X C there is a covering {p
i
: X
i
X}
iL
such that for the restriction of M to
each C/X
i
, there is a presentation
O|
(I)
X
i
O|
(J)
X
i
M|
X
i
0,
where by F
(I)
we denote a coproduct of copies of a sheaf F indexed by some
set I. If C is the site of open sets of a classical topology, this denition agrees
with [EGA I, 0, (5.1.3)]. We will denote the category of quasi-coherent sheaves
on (C, O) by Qco(C
, O).
1.7. Let f : (C, ) (D, ) be a continuous functor of sites [SGA 4
1
, III, (1.1)]
such that f preserves bered products. In view of [SGA 4
1
, III, (1.6)], the
continuity of f is equivalent to the fact that it takes coverings in (C, ) to
coverings in (D, ).
3
Sometimes called the comma category.
Whenever C possesses colimits, this morphism induces a pair of adjoint
functors between the associated topos
D
f
1
.
Let us recall rst the denition of f
. If G D
and V C, then
(f
G)(V) =G(f(V)).
Let now F C
, f
1
F is dened as the sheaf associated to the presheaf that
assigns to each U D the colimit
lim
I
f
U
F(V),
where I
f
U
is the category whose objects are the pairs (V, g) with V C and
g: U f(V) is a morphism in D, and whose morphisms
h: (V, g) (V
, g
)
are dened as morphisms
h: V V
such that f(h) g = g
(see [SGA 4
1
, I, 5.1] and [SGA 4
1
, III, 1.2.1] where f
1
is denoted f
s
). Notice
that, in general, f
1
might not be exact. If f
1
is exact, we say that the pair
(f
, f
1
) is a morphism of topos from D
to C
according to [SGA 4
1
, IV 3.1].
We want to extend this formalism to the case of ringed sites. To have
the possibility of transporting algebraic structures through the inverse image
functor, we need the following lemma, communicated to us by Olsson.
Lemma 1.8. Let I be a category with nite products and let F
1
, . . ., F
r
a nite
family of functors F
i
: I
o
Set, i {1. . . r}. The natural map
lim
I
(F
1
F
r
)
_
_
lim
I
F
1
_
_

_
_
lim
I
F
r
_
_
is an isomorphism.
Proof. By using induction, we may assume that r =2. Notice that
lim
I
F
1
lim
I
F
2
is identical to
lim
II
F
1
F
2
where F
1
F
2
: I
o
I
o
Set is the functor that sends a pair (U, V) I
o
I
o
to
F
1
(U) F
2
(V). The diagonal functor : I
o
I
o
I
o
induces a map
z: lim
I
F
1
F
2
lim
II
F
1
F
2
= lim
I
F
1
lim
I
F
2
We will check rst that z is surjective using the fact that I has products.
Indeed, for any two objects U, V I and u F
1
(U), v F
2
(V) the element
[(u, v)] lim
II
F
1
F
2
equals the class of (F
1
(p
1
)(u), F
2
(p
2
)(v)) F
1
(UV)F
2
(UV) via the natural
maps p
1
: UV U, p
2
: UV V, so any element lies in the image of z.
Now, we prove that z is injective. Consider the functor : I
o
I
o
I
o
that
sends (U, V) to U V (notice that the product is taken in I). It gives the
following natural transformation
F
1
F
2
(F
1
F
2
)
dened as
F
1
(p
1
) F
2
(p
2
) : F
1
(U) F
2
(V) F
1
(UV) F
2
(UV)
on (U, V) I
o
I
o
. The functor induces a map
z
: lim
II
F
1
F
2
lim
I
F
1
F
2
.
To see that z is injective is enough to check that z
z =id. The map z
z sends
the class of (u
1
, u
2
) F
1
(U) F
2
(U) to the class of (F
1
(p
1
)(u
1
), F
2
(p
2
)(u
2
))
F
1
(U U) F
2
(UU). Let : U UU be the diagonal morphism. As we
claimed, the class of (F
1
(p
1
)(u
1
), F
2
(p
2
)(u
2
)) equals the class of the element
(F
1
(p
1
)(u
1
), F
2
(p
2
)(u
2
)) =(u
1
, u
2
).
1.9. Let f : (C, ) (D, ) be a continuous functor of sites as in 1.7. Let O be
a sheaf of rings on (C, ) and P a sheaf of rings on (D, ). Let us be given a
morphism of sheaves of rings f
#
: O f
P or, equivalently, its adjoint map

f
#
: f
1
O P.
Consider rst the case in which (f
, f
1
) is a morphism of topos. In this case
we say that ((f
, f
1
), f
#
) is a morphism of ringed topos [SGA 4
1
, IV, 13.1] and
provides an adjunction
P-Mod
f
O-Mod.
The functor f
: O-Mod P-Mod is dened by

f
F :=P
f
1
O
f
1
F
for a O-module F. Observe that for a P-module G, f
G inherits a structure
of O-module via f
#
and we have the claimed adjunction f
.
In general f
1
may not be exact, and the previous discussion does not ap-
ply. However, if we assume that the category I
f
U
has nite products, then by
Lemma 1.8, the functor that denes the inverse image for presheaves com-
mutes with nite products. The sheacation functor commutes also with
nite products since it is exact. It follows that f
1
also commutes with nite
products. Thus, f
1
O has canonically the structure of a sheaf of rings and
also f
#
is a homomorphism. By the same reason, if F is an O-module then
f
1
F is an f
1
O-module. The functor f
: O-Mod P-Mod given by

f
F :=P
f
1
O
f
1
F
provides an adjunction f
as in the previous situation.

Notice that we use here f
1
for general sheaves and reserve f
for modules
over ringed topos, contradicting the usage of [SGA 4
1
, IV].
Remark. The perspicuous reader will notice that Lemma 1.8 is also implicitly
used in [O, Def. 3.11].
2. CARTESIAN PRESHEAVES ON AFFINE SCHEMES
2.1. Let X = Spec(A) be an afne scheme. We will work with the category
A/X of at nitely presented afne schemes over X, in other words the
opposite category to the category of at nite presentation A-algebras. Note
that this is an essentially small category. We will endow it with the topology
whose coverings are nite families {p
i
: U
i
U}
iL
such that p
i
are at nite
presentation maps such that
iL
Im(p
i
) = U. We will denote this site by
A
fppf
/X and by X
fppf
its associated topos. Due to the fact that we limit
the possible objects to those at of nite presentation, this site should not
be regarded as a big site and sometimes we will refer to it informally as the
small at site.
Let

A or O
X
denote the ring valued sheaf dened by

A(Spec(B)) =B. Note
that

A is a sheaf on A
fppf
/X as follows from faithfully at descent [SGA 4
2
,
VII, 2c], see also [SGA 1, VIII, 1.6] and [SGA 3
1
, IV, 6.3.1]. We will denote
the corresponding ringed topos as (X
fppf
,

A) or (X
fppf
, O
X
).
Let Crt(X) :=Crt(A/X,

A). We recall the somewhat classical
Proposition 2.2. With the previous notation, there is an equivalence of cate-
gories
Crt(X)
= A-Mod.
Proof. Consider the functors (X, ) : Crt(X) A-Mod and ()
crt
: A-Mod
Crt(X) dened by (X, M) := M(X, id
X
) and M
crt
(Spec(R)) := R
A
M. It is
clear that they are mutually quasi-inverse.
Since each Cartesian presheaf in Crt(X) is isomorphic to a presheaf of the
form M
crt
, then it is a sheaf [SGA 4
2
, VII, 2c]. The category Crt(X) is a full
subcategory of the category of sheaves of

A-Modules closed under the forma-
tion of kernels, cokernels and coproducts, and it is an abelian category.
2.3. Let f : X Y be a morphism of afne schemes with X = Spec(B) and
Y =Spec(A). We have a continuous functor
f
A
: A
fppf
/Y A
fppf
/X
given on objects by f
A
(V, v) :=(V
Y
X, p
2
) with p
2
denoting the projection
and on maps by the functoriality of pull-backs. It is clear that f
A
preserves
bered products, therefore by 1.7, induces a pair of adjoint functors
X
fppf
f
1
Y
fppf
,
where we are taking the usual shortcuts f
1
:=( f
A
)
1
and f
:=( f
A
)
.
Moreover, we have the morphism of sheaves of rings
f
A #
:

A f
B
given by f
A #
(V, v) : R R
A
B, f
A #
(V, v)(r) = r 1, for V =Spec(R). Since
the category I
f
A
(U,u)
has nite products, by 1.9 we have the adjunction
B-Mod
f
A-Mod
where f
:= ( f
A
)
. Notice that if f is a at nitely presented morphism,

then f
F(U, u) =F(U, f u).

Remark. We will abuse notation slightly and write G(U) for G(U, u) for any
(U, u) A/X and G X
fppf
, if the morphism u is understood.
Proposition 2.4. It holds that
(i) If F Crt(Y) then f
F Crt(X).
(ii) If G Crt(X) then f
G Crt(Y).
Proof. (i) Let (U, u) A
fppf
/X with U =Spec(S). Denote I
f
A
(U,u)
by I
U
and let
A
U
:= lim
I
U
A(V).
The sheaf f
F is the sheaf associated to the presheaf P given by

P(U, u) =S
A
U
lim
I
U
F(V).
We have for every S-module N the following isomorphisms
Hom
S
(S
A
U
lim
I
U
F(V), N) Hom
A
U
( lim
I
U
F(V), N)
lim
I
U
Hom
A(V)
(F(V), N)
lim
I
U
Hom
S
(S
A(V)
F(V), N)
Hom
S
( lim
I
U
(S
A(V)
F(V)), N).
By Yonedas lemma, it follows that f
F is isomorphic to the sheaf associated

to the presheaf P
such that
P
(U, u) = lim
I
U
S
A(V)
F(V).
Let h: ((V, v), g) ((V
, v
), g
) be a morphism in I
U
, with V = Spec(R), V
=
Spec(R
) and h =Spec(). We have the commutative diagram

S
R
F(V
) S
R
F(V)
S
R
R
A
F(Y) S
R
R
A
F(Y)
S
A
F(Y) S
A
F(Y)
idF(h)
idF(v
)
a
idF(v)
a
idid
where the top horizontal map is the transition morphism and the vertical
maps are isomorphisms by the denition of Cartesian presheaf and the prop-
erties of the tensor product. Thus, P
(U, u) =S
A
F(Y) and therefore f
F is
isomorphic to (B
A
F(Y))
crt
.
(ii) Let h: (V, v) (V
, v
) be a morphism in A
fppf
/Y, V = Spec(R), V
=
Spec(R
) and h =Spec(). We have to prove that the morphism

f
G(h)
a
: R
R
f
G(V
) f
G(V)
is an isomorphism (see 1.3). We have that f
G(V) =G(V
Y
X) and similarly
for f
G(V
). Notice that f
G(h) =G(h
Y
id
X
), thus the morphism
G(h
Y
id
X
)
a
: (R
A
B)
R
A
B
G(V
Y
X) G(V
Y
X)
is an isomorphismbecause G is Cartesian. The result follows because f
G(h)
a
is identied with the composition of the following isomorphisms
R
R
G(V
Y
X)

(R
A
B)
R
A
B
G(V
Y
X)
G(hid)
a
G(V
Y
X).
Notice that f
(G) is isomorphic to (
A
G(X))
crt
, where
A
G(X) means G(X) con-
sidered as an A-module.
Corollary 2.5. A morphism of afne schemes f : X Y induces a pair of
adjoint functors
Crt(X)
f
Crt(Y).
Remark. We leave as an exercise for the interested reader to recognize that
the previous adjunction agrees with the classical one
Qco(X
fppf
, O
X
)
f
Qco(Y
fppf
, O
Y
),
via the fact that in both cases the (pre)sheaves can be represented by its
module of global sections. Notice that we consider the fppf topology instead
of the usual Zariski topology but this does not change essentially the category
of quasi-coherent sheaves. We will see later (in Theorem 3.11) the agreement
of quasi-coherent sheaves with Cartesian presheaves in a much more general
setting that includes all quasi-compact and semi-separated schemes.
3. QUASI-COHERENT SHEAVES ON GEOMETRIC STACKS
3.1. We will follow the conventions of [LMB], specially for the denition of
algebraic stack. But for our purposes we will restrict to what we call geomet-
ric stacks after Lurie (see [L]). So, from now on, a geometric stack X will be a
stack on the tale topology of (afne) schemes over a base scheme (that will
not play any role in what follows)
4
such that:
(i) the stack X is semi-separated, i.e. the diagonal morphism : X
XX is representable by afne schemes;
(ii) the stack X is algebraic and quasi-compact, this amounts to the exis-
tence of a smooth and surjective morphism p: X X from an afne
scheme X.
We will refer to the map p: X X as a presentation of X.
A map representable by afne schemes is usually denominated an afne
morphism [LMB, 3.10]. Observe that every morphism u: U X from an
afne scheme U is afne. Indeed, if v: V X is another morphism with V
an afne scheme, we have that U
X
V X
XX
(UV) is 1-isomorphic to an
afne scheme and this 1-isomorphism is an isomorphism because U
X
V is
in fact a category bered in sets. We will identify it with the spectrum of its
global sections. We recall that the pair (X, X
X
X) with the obvious structure
morphisms is a scheme in groupoids whose associated stack is X, i.e.
X=[(X, X
X
X)]
with the notation as in [LMB, (3.4.3)].
Lemma 3.2. Let f : XY a morphism of geometric stacks. The relative di-
agonal
X/Y
: XX
Y
X is an afne morphism.
Proof. Consider the following composition
X
f
XY
p
2
Y.
Notice that f p
2
f
. Semi-separatedness on schemes is preserved by base
change [AJPV, Proposition 2.2 (iii)], so by [LMB, Remarque (4.14.1)] it is
also preserved by base change on algebraic stacks, this guarantees the semi-
separatedness of p
2
. The semi-separatedness of
f
follows from the fact that
is a quasi-section, so its diagonal map is an equivalence.
Proposition 3.3. Let
X
X Y
f
g
f
be a 2-Cartesian square of algebraic stacks. If X, Y and Y
are geometric
stacks, then so is X
.
4
In fact, it can be shown that the fppf topology yields the same category of stacks.
Proof. First, X
is an algebraic stack by [LMB, (4.5)(i)].

Let us check that the diagonal morphism
X
: X
is afne. Factor
it as
X
/X
X
X
X
h
X
where h is the morphism in the following 2-Cartesian square

X
X
X
X XX
h
g
The morphism h is afne because it is obtained by base change from

X
. The
morphism
X
/X
/Y
X
X
, so by Lemma 3.2, it is afne. It follows that

X
is afne as wanted.
Finally, given presentations p: X X and q: Y
we have an induced
morphism
p
Y
q: X
Y
Y
Notice that X
Y
Y
is afne over X because is the base change of the map

q that is afne, therefore X
Y
Y
is an afne scheme. The morphism p

Y
q
is smooth and surjective because both conditions are stable by base change
applying [LMB, Prop. (4.13) et Rem. (4.14.1)] in view of [EGA I, (1.3.9)]. This
makes of p
Y
q a presentation and the proof is complete.
3.4. For X a geometric stack, we dene the category A/X as follows. Ob-
jects are pairs (U, u) with U an afne scheme and u: U X a at nitely
presented 1-morphism of stacks. Morphisms ( f , ) : (U, u) (V, v) are com-
mutative diagrams of 1-morphisms of stacks. Let us spell this out. To give
such an ( f , ) amounts to give a diagram
U V
X
f
u v
where f : U V is a morphism of afne schemes making the diagram 2-

commutative, through the 2-cell : u vf . The composition of morphisms is
given by
(g, ) ( f , ) =(gf , f ).
Notice that A/X is an essentially small category. In fact, let (U, u) A/X
with U =Spec(B), by the semi-separateness of X, the scheme U
X
X is afne.
This afne scheme is isomorphic to the spectrum of a nitely presented alge-
bra B
over the ring of global sections of the structure sheaf of X. By faithful

atness B is isomorphic to a subring of B
. This makes sense of what follows.

3.5. The category A/Xis ringed by the presheaf O : A/XRing, dened by
O(U, u) =B with U =Spec(B). We dene Cartesian presheaves on a geometric
stack X as Cartesian presheaves over this ringed category,
Crt(X) :=Crt(A/X, O).
Now, on A/X we may dene a topology by declaring as covering the nite
families {( f
i
,
i
) : (U
i
, u
i
) (V, v)}
iI
, with every f
i
a at nitely presented
map, that are jointly surjective. It becomes a site that we denote A
fppf
/X.
The associated topos, i.e.the category of sheaves of sets on A
fppf
/Xis denoted
simply X
fppf
. The site A
fppf
/Xis a ringed site by the sheaf of rings associated
to the presheaf O that we keep denoting the same.
5
We dene
Qco(X) :=Qco(X
fppf
, O).
Remark.
(i) Every smooth morphism is at of nite presentation, therefore our
site is ner than the usual lisse-tale topology, see [LMB, (12.1)].
(ii) The topos X
fppf
has enough points. The argument is the same as for
the lisse-tale topos, see [LMB, (12.2.2)].
In what follows we will abuse notation slightly and write F(U) for F(U, u)
and analogously for any object in A/X if the morphism u: U X is obvious.
Lemma 3.6. A Cartesian presheaf F Crt(X) is actually a sheaf of O-Modu-
les.
Proof. Let (V, v) A/X with V =Spec(C). Let {( f
i
,
i
) : (U
i
, u
i
) (V, v)}
iI
be
a covering with U
i
= Spec(B
i
). Let p
i j
1
and p
i j
2
be the two canonical projec-
tions from U
i
V
U
j
to U
i
and U
j
, respectively. Consider (U
i
V
U
j
, vf
i
p
i j
1
)
A/X (notice that vf
i
p
i j
1
=vf
j
p
i j
2
). We have the diagram
F(V)

iI
F(U
i
)

i, jI
F(U
i
V
U
j
)
1
l
where l =(F( f
i
,
i
))
iI
, for (x
i
)
iI
F(U
i
), we denote by
1
(x
i
) the element
whose component in F(U
i
V
U
j
) is F(p
i j
1
,
1
i
p
i j
1
)(x
i
) and by
2
(x
i
) the ele-
ment whose component in F(U
i
V
U
j
) is F(p
i j
2
,
1
j
p
i j
2
)(x
j
). Let us check that
this diagram is an equalizer. Being F Cartesian, this diagram is isomorphic
to the diagram
F(V)

iI
B
i
C
F(V)

i, jI
(B
i
C
B
j
)
C
F(V)
1
l
where l
(x) = (1x)
iI
, for x F(V), and for (b
i
y
i
)
iI

B
i
C
F(V), the
elements
1
((b
i
y
i
)
iI
) and
2
((b
i
y
i
)
iI
) are those whose components in
(B
i
C
B
j
)
C
F(V) are b
i
1y
i
, and 1b
j
y
j
, respectively. Let B:=
iI
B
i
5
The next lemma will show that there is no ambiguity.
and the morphism : C B induced by the covering { f
i
: U
i
V}. Since I is
nite, it sufces to prove that the diagram
F(V) B
C
F(V) (B
C
B)
C
F(V)
j
2
id
j
1
id
l
is an equalizer, where l
(x) =1x, j
1
(b) = b1 and j
2
(b) =1b, for x F(V)
and b B. As B is faithfully at over C, the diagram
C

B
j
2
j
1
B
C
B
is an equalizer and stays so after tensoring by F(V) [SGA1, VIII 1.5]. The
result follows.
Remark.
(i) The proof works for any topology coarser than fpqc, we have chosen
the fppf topology because it gives a reasonable small site.
(ii) The category Crt(X) is a priori a full subcategory of Pre-O
X
-Mod. The
previous lemma implies that it is a full subcategory of O
X
-Mod, as
Qco(X) is.
Proposition 3.7. The category Crt(X) is a full subcategory of O
X
-Mod closed
under the formation of kernel, cokernels and coproducts, and therefore it is an
abelian category.
Proof. Let K be the kernel of the morphism M M
with M, M
Crt(X)
and ( f , ) : (U, u) (V, v) be a morphism in A/X with U =Spec(B) and V =
Spec(C). If f is at, we have the following commutative diagram
0 B
C
K (V) B
C
M(V) B
C
M
(V)
0 K (U) M(U) M
(U)

where the rows are exact and the midle and right vertical arrows are iso-
morphisms because M and M
are Cartesian sheaves. It follows that the

remaining vertical arrow K ( f , )
a
: B
C
K (V) K (U) is an isomorphism.
If f is not at, let p: X X be an afne presentation with X = Spec A
0
.
If p
1
and p
1
, and p
2
and p
2
are the projections from U
X
X and V
X
X to
U and V, and X, respectively, and : up
1
pp
2
and
: vp
1
pp
2
are the
canonical 2-morphisms, we have a commutative diagram of afne schemes
U
X
X U
X V
X
X V
p
1
p
2
f
f
p
2
p
1
where the projections p
1
, p
2
, p
1
and p
2
are at maps and
=p
1
. Put
Spec(B
) =U
X
X and Spec(C
) =V
X
X. Since the projection p
1
is faithfully
at, to prove that the morphism K( f , )
a
is an isomorphism it sufces to
verify that the morphism
id
B

B
K( f , )
a
: B
B
(B
C
K (V)) B
B
K (U)
is an isomorphism. But this morphism is the composition of the following
chain of isomorphisms
B
B
(B
C
K(V)) B
C
C
C
K(V)
B
C
K (V
X
X, pp
2
) (via K (p
1
, (
)
1
))
B
C
(C
A
0
K(X)) (via K (p
2
, id))
B
A
0
K (X)
K(U
X
X, pp
2
) (via K (p
2
, id))
B
B
K (U) (via K (p
1
,
1
))
The presheaf cokernel of M M
is a Cartesian presheaf and then it is

the cokernel in the category of sheaves of O-modules. Similarly, the coproduct
of a set of Cartesian presheaves is a Cartesian presheaf and then it is the
coproduct in the category of sheaves of O-modules.
Let X be a geometric stack and p: X X a presentation with X an afne
scheme. Then p induces a continuous morphism of sites
p
A
: A
fppf
/XA
fppf
/X
that is given on objects by p
A
(U, u) :=(U
X
X, p
2
), with p
2
: U
X
X X the
projection, and on maps by the functoriality of pull-backs.
As discussed in 1.7 this morphism induces a pair of adjoint functors
X
fppf
p
1
X
fppf
.
Notice that p is a at nite presentation 1-morphism, so (p
1
F)(V, v) =
F(V, pv), for F X
fppf
. In particular, p
1
O
X
= O
X
, therefore p
1
induces
a functor between the categories of sheaves of O-Modules, i.e. p
1
agrees
with the functor usually denoted p
. So we have an adjunction as in 1.9

O
X
-Mod
p
O
X
-Mod, (3.7.1)
with a particularly simple description of the functors involved, due to the fact
that p induces a restriction functor between the corresponding sites.
Proposition 3.8. It holds that
(i) If F Crt(X) then p
F Crt(X).
(ii) If G Crt(X) then p
G Crt(X).
Proof. We prove (i). Let h: (U, u) (V, v) be a morphism in A
fppf
/X with
U =Spec(B) and V =Spec(C). We have to check that the morphism
(p
F(h))
a
: B
C
p
F(V, v) p
F(U, u)
is an isomorphism. But p
F(V, v) = F(V, pv) and analogously for (U, u).

Therefore the previous morphism becomes
F(h, id
pu
)
a
: B
C
F(V, pv) F(U, pu)
where (h, id
pu
) : (U, pu) (V, pv) is a morphism in A
fppf
/X. But this map is
an isomorphism because F is Cartesian.
The proof of (ii) is similar to the proof given in Proposition 2.4 (ii).
Corollary 3.9. In this setting, the adjunction (3.7.1) restricts to
Crt(X)
p
Crt(X).
Let us now prove a converse to Proposition 3.8 (i).
Lemma 3.10. Let F be an O
X
-module. If p
F Crt(X), then F Crt(X).

Proof. Set T:= p
. Since F is a sheaf, we have the equalizer of O

X
-modules
F

F
TF
TF
T
F
T
2
F
where is the unit of the adjunction p
. As a consequence of 3.8 TF,

T
2
F Crt(X), it follows that F Crt(X).
Theorem 3.11. Let X be a geometric stack. The subcategories Crt(X) and
Qco(X) of O
X
-Mod agree.
Proof. Let us prove rst that Crt(X) Qco(X).
Consider a presentation p: X X as before and F Crt(X). Let X =
Spec(A
0
). First, notice that the Cartesian sheaf p
F sits in an exact sequence

O
(I)
X
O
(J)
X
p
F 0 (3.11.1)
that comes from a presentation of M:=(X, p
F) as A
0
-module applying the
functor ()
crt
of Proposition 2.2. Let (U, u) A
fppf
/X, and put V =U
X
X, p
1
and p
2
being the projections and : up
1
pp
2
the canonical 2-morphism of
the pullback. We have the covering {(p
1
,
1
) : (V, pp
2
) (U, u)} in A
fppf
/X
with p
2
: V X a at nite presentation map. Applying p
2
to the resolution
(3.11.1) we obtain the exact sequence of sheaves of O
V
-modules
O
(I)
V
O
(J)
V
p
2
p
F 0.
The result follows from the fact that the category O
V
-Mod is equivalent to
the category of sheaves of O|
(V, pp
2
)
-modules over the site (A
fppf
/X)/(V, pp
2
)
by the comparison result [SGA 4
1
, III, Thorme 4.1].
Let us prove now that Qco(X) Crt(X).
Since F Qco(X), there exists a covering {( f
i
,
i
) : (U
i
, u
i
) (X, p)}
iI
such
that for the restriction of F to each (A
fppf
/X)/(U
i
, u
i
) there is a presentation
O|
(I)
(U
i
,u
i
)
O|
(J)
(U
i
,u
i
)
F|
(U
i
,u
i
)
0
i.e. a presentation of O
U
i
-modules
O
(I)
U
i
O
(J)
U
i
u
i
F 0
u
i
F being the sheaf of O
U
i
-modules given by u
i
F(W, w) = F(W, u
i
w) for
(W, w) A
fppf
/U
i
. The 2-morphism
i
: u
i
pf
i
induces an isomorphism be-
tween the sheaves u
i
F and f
i
p
F. Since O
U
i
Crt(U
i
), then u
i
F Crt(U
i
)
and thus f
i
p
F Crt(U
i
).
Because p
F is a sheaf and { f
i
: U
i
X}
iI
is a covering in A
fppf
/X we
obtain the equalizer of O
X
-modules
p
iI
f
i
f
i
p
i, jI
f
i j
f
i j
p
F,
where f
i j
: U
i
X
U
j
X is the composition of the rst projection with f
i
, or,
what amounts to the same, the composition of the second projection with f
j
.
Since f
i
f
i
p
F, f
i j
f
i j
p
F Crt(X), then p
F Crt(X). The result follows

from Lemma 3.10.
Remark. The consequence of the previous theorem is that not only all quasi-
coherent sheaves are Cartesian (presheaves) but that this fact characterizes
them on a geometric stack. From now on, we will use freely this identica-
tion when dealing with quasi-coherent sheaves and use the most convenient
characterization for the issue at hand.
4. DESCENT OF QUASI-COHERENT SHEAVES ON GEOMETRIC STACKS
We proceed in our goal of describing quasi-coherent sheaves through al-
gebraic data on an afne groupoid scheme whose stackication is the initial
geometric stack. As a rst step, we will develop a descent result for quasi-
coherent sheaves with respect to presentations of stacks.
4.1. The category of descent data. Let X be a geometric stack and p: X
Xa presentation with X an afne scheme and G an O
X
-Module. As before, for
an object (U, u) A
fppf
/X, we will abbreviate G(U, u) by G(U) if no confusion
arises. Analogously, for a morphism (h, ) we will shorten G(h, ) by G(h).
Consider the basic pull-back diagram
X
X
X X
X X
p
2
p
1
p
p
(4.1.1)
A descent datum on G is an isomorphism t : p
2
G p
1
G of O
X
X
X
-Modules
verifying the cocycle condition
p
12
t p
23
t = p
13
t
on A
fppf
/X
X
X
X
X, where as usual p
i j
: X
X
X
X
X X
X
X (i < j
{1, 2, 3}) denotes the several projections obtained omitting the factor not de-
picted. We dene the category Desc(X/X) by taking as its objects the pairs
(G, t) with G an O
X
-Module and t : p
2
G p
1
G a descent datum on G; and
its morphisms h: (G, t) (G
, t
) with h: G G
a morphism of O
X
-Modules
such that the diagram
p
2
G p
1
G
p
2
G
1
G
t
p
2
h p
1
h
t
commutes.
4.2. Descent datum associated to a sheaf of modules. The 2-morphism
: pp
1
pp
2
induces an isomorphism
: p
2
p
1
p
given by
(
F
)(U, u) =F(id
U
, u), for F O
X
-Mod and (U, u) A
fppf
/X
X
X.
Notice that for any O
X
-Module F, it holds that (p
F,
F
) Desc(X/X). In-
deed, consider the pullback
X
X
X
X
X X
X
X
X X
p
3
p
12
p
p p
2
(4.2.1)
With the previous notations, we have that p
13
=
p
12
and p
23
=
and
thus, p
23
p
12
=p
13
. Now for (U, u) A
fppf
/X
X
X
X
X we have
(p
12
F
p
23
F
)(U, u) =F(id
U
, p
12
u) F(id
U
, p
23
u)
=F(id
U
, p
13
u)
=(p
13
F
)(U, u).
This construction denes a functor
D: O
X
-Mod Desc(X/X)
by D(F) :=(p
F,
F
).
4.3. From descent data to sheaves of modules. We dene a functor
G: Desc(X/X) O
X
-Mod
assigning to (G, t) Desc(X/X) the equalizer of the pair of morphisms
p
G
a
G
b
(G,t)
p
p
2
p
1
G
where
a
G
:=((
)
p
1
G
) (p
1G
)
b
(G,t)
:=(p
p
2
t) (p
2G
)
with
: p
p
1
p
p
2
denoting the functorial isomorphism and
i
the unit
of the adjunction p
i
p
i
for i {1, 2}. We denote this kernel by G(G, t). This
construction is clearly functorial.
Proposition 4.4. The previously dened functors are adjoint, DG.
Proof. Let i
G
: G(G, t) p
G be the canonical morphism. For each morphism

of O
X
-modules g: F G(G, t), the adjoint to i
G
g: F p
G through p
provides the claimed morphism (p
F,
F
) (G, t) of descent data.
Reciprocally, if f : D(F) (G, t) is a morphism in Desc(X/X), the morphism
F p
G given by the adjunction p
factors through G(G, t) giving a

morphismF G(G, t), as desired.
To establish the next main result, Theorem 4.6, we will need a few techni-
cal preparations. Let (V, v) A
fppf
/X and consider the following pull-backs
V
p
2
(X
X
X) X
X
X
V X
v
2
v
1
p
2
v
V
X
X X
V X
w
2
w
1
p
pv
Denote W=V
p
2
(X
X
X). Set h: V
X
X X
X
X the morphism such that
p
1
h =w
2
, p
2
h =vw
1
and h =
1
. The morphism h
: V
X
X W verifying
v
1
h
=w
1
and v
2
h
= h is an isomorphism. Let us dene

: p
2
p
1
p
the isomorphism of functors given by (

G
)(V, v) =G(h
) for each G O
X
-Mod.
Set
c
(G,t)
:=
G
(p
2
t)
2G
.
Lemma 4.5. With the previous notations, the following identity holds
p
a
G
c
(G,t)
= p
b
(G,t)
c
(G,t)
.
Proof. Consider the pull-backs
W
1
X
X
X
V
X
X X
1
p
1
w
2
W
2
X
X
X
V
X
X X
1
p
2
w
2
Let l : W
2
V
X
X be the morphism such that w
1
l = w
1
1
, w
2
l = p
1
2
and
l =
1
1
. Put l
: W
2
W
1
for the morphism verifying that
1
l
= l
and
2
l
2
. Notice that l
is an isomorphism and p
((
)
p
1
G
)(V, v) =G(l
).
We have
(p
((
)
p
1
G
p
1G
) c
(G,t)
))(V, v) =G(l
)G(
1
)G(h
)t(W, v
2
)G(v
1
)
=G(h
l)t(W, v
2
)G(v
1
)
= t(W
2
, hl)G(h
l)G(v
1
)
= t(W
2
, hl)G(w
1
l).
On the other hand, we have
(p
(p
p
2
t p
2G
) c
(G,t)
)(V, v) = t(W
2
,
2
)G(
1
)G(h
)t(W, v
2
)G(v
1
)
= t(W
2
,
2
)t(W
2
, h
1
)G(h
1
)G(v
1
)
= t(W
2
,
2
)t(W
2
, h
1
)G(w
1
l).
Thus, it is sufcient to prove that t(W
2
,
2
)t(W
2
, h
1
) = t(W
2
, hl). For that,
consider the 2-pull-back
(X
X
X)
pp
2
X
V V
X
X
X X
c
2
c
1
pv
pp
2
Let k: W
2
(X
X
X)
pp
2
X
V be the isomorphism given by c
1
k =
2
, c
2
k =
w
1
1
, and k =
1
1
. Consider the morphism
id
X
v : (X
X
X)
pp
2
X
V X
X
X
X
X
satisfying p
12
(id
X
v) = c
1
, p
3
(id
X
v) = vc
2
, and
(id
X
v) = . The mor-
phism g = (id
X
v)k is a nitely presented at map. Applying the cocycle
condition of t to (W
2
, g), we obtain that
t(W
2
, p
12
g) t(W
2
, p
23
g) = t(W
2
, p
13
g)
and the result follows because p
12
g =
2
, p
23
g = h
1
and p
13
g = hl.
Theorem 4.6. The functor D: O
X
-Mod Desc(X/X) (and, therefore its ad-
joint G) is an equivalence of categories.
Proof. We will see rst that , the unit of the adjunction DG, is an isomor-
phism. Let F O
X
-Mod and consider the following commutative diagram
F p
F p
F
G(p
F,
F
) p
F p
p
2
p
1
p
2
a
p
F
b
(p
F,
F
)
via
1
F
F
i
p
F
where
F
is the unit of the adjunction p
applied to F. The rst row is

an equalizer because F is a sheaf and the claim follows.
Let us prove now that , the counit of the adjunction D G, is an iso-
morphism. Let (G, t) Desc(X/X). We have that
(G,t)
=
G
p
i
G
, where
G
denotes the counit of the adjunction p
.
p
G(G, t) p
G p
p
2
p
1
G
G
p
a
G
p
b
(G,t)

(G,t)
p
i
G
c
(G,t)
Since p
preserve nite limits, the horizontal row is exact and so by Lemma

4.5 there exists a morphism r : G p
G(G, t) with p
i
G
r = c
(G,t)
. It holds
that c
(G,t)

(G,t)
= p
i
G
. Indeed,
c
(G,t)

(G,)
=
G
p
2
t
2G
G
p
i
G
=
p
G
p
(
G
p
2
t
2G
) p
i
G
=
p
G
p
G
p
((
)
p
1
G
) p
1G
p
i
G
.
and after an straightforward calculation we see that
G
p
G
p
((
)
p
1
G
) p
1G
=id
p
G
.
We see now that r is the inverse of
(G,t)
. Indeed,
p
i
G
r
(G,t)
= c
(G,t)

(G,t)
= p
i
G
and p
i
G
is an monomorphism of sheaves, therefore r
(G,t)
=id
G
. Since G is
a sheaf,
2G
is injective and thus c
(G,t)
is an monomorphism of sheaves. From
c
(G,t)

(G,t)
r = p
i
G
r = c
(G,t)
it follows that
(G,t)
r =id
p
G(G,t)
.
We denote by Desc
qc
(X/X) the full subcategory of the category Desc(X/X)
whose objects are the pairs (G, t) with G Qco(X).
Corollary 4.7. The categories Qco(X) and Desc
qc
(X/X) are equivalent.
Proof. The result follows because for F Qco(X) we have D(F) Desc
qc
(X/X)
and, conversely, for (G, t) Desc
qc
(X/X) it holds that G(G, t) Qco(X) as fol-
lows from Propositions 3.7, 3.8 and 4.6, having in mind Theorem 3.11.
Remark. Our method of proof is not too distant from Olssons [O, 4]. Notice,
however, that he is working with Cartesian sheaves on a simplicial site while
our result follows from the more general Theorem 4.6 that applies not only to
quasi-coherent Modules.
5. REPRESENTING QUASI-COHERENT SHEAVES BY COMODULES
5.1. Hopf algebroids as afne models of geometric stacks. Let X be
a geometric stack and p: X X a presentation with X afne. The bered
product X
X
X is an afne scheme because X is geometric. As we recalled in
3, the pair (X, X
X
X) is a scheme in groupoids whose associated stack is X.
Let A
0
and A
1
be the rings dened by X = Spec(A
0
) and X
1
:= X
X
X =
Spec(A
1
). Denote the pair A
:= (A
0
, A
1
). The dual structure of a scheme
in groupoids is called a Hopf algebroid. Let us spell out the corresponding
structure of A
, there are:
two homomorphisms
L
,
R
: A
0
A
1
,
corresponding respectively to the projections p
1
, p
2
: X
X
X X;
the counit homomorphism
: A
1
A
0
,
corresponding to the diagonal : X X
X
X;
the conjugation homomorphism
: A
1
A
1
,
corresponding to the map interchanging the factors in X
X
X;
and the comultiplication
: A
1
A
1
R
L
A
1
,
corresponding, via the isomorphism
X
X
X
X
X

(X
X
X)
p
1
p
2
(X
X
X)
induced by the projections p
23
and p
12
, to the projection
p
13
: X
X
X
X
X X
X
X,
expressing the composition of arrows in the scheme in groupoids
(X, X
X
X).
These data is subject to the following identities:
(i)
L
=id
A
0
=
R
(source and target of identity).
(ii) If j
1
, j
2
: A
1
A
1
R
L
A
1
are the maps given by j
1
(b) = b1, j
2
(b) =
1b, then
L
= j
1
L
and
R
= j
2
R
(source and target of a com-
position).
(iii)
L
=
R
and
R
=
L
(source and target of inverse).
(iv) (id
A
1
)=id
A
1
=(id
A
1
) (identity).
(v) (id
A
1
)=(id
A
1
) (associativity of composition).
(vi) If is the multiplication on A
1
, then (id
A
1
)=
R
and
(id
A
1
)=
L
(inverse).
(vii) =id
A
1
(inverting twice).
Compare with [R, A1.1.1.] and [LMB, (2.4.3)].
The geometric stack dened by a scheme in groupoids (Spec(A
0
), Spec(A
1
))
is denoted
Stck(A
) :=[(Spec(A
0
), Spec(A
1
))]
(with the right hand side notation as in [LMB, (3.4.3)]). So, with the notation
at the beginning of this section, X=Stck(A
).
From now on, every geometric stack we consider will be the stack associ-
ated to a Hopf algebroid. From a geometric point of view, the data of a Hopf
algebroid is more rigid than the underlying algebraic stack alone. It is equiv-
alent to the datum of an algebraic stack together with a smooth presentation
by an afne scheme. In any case, the algebroid determines the stack and
we will use this fact in what follows. Note that is an isomorphism (being
inverse to itself) and that
L
and
R
are smooth morphisms.
5.2. Comodules over a Hopf algebroid. Let A
:= (A
0
, A
1
) be a Hopf al-
gebroid. A (left) A
-comodule (M,
M
) is an A
0
-module M together with an
A
0
-linear map
M
: M A
1
R
A
0
M, the target being a A
0
-module via
L
,
and verifying the following identities:
(i) (id
M
)
M
=(id
A
1
M
)
M
(coassociativity of ).
(ii) (id
M
)
M
=id
M
(
M
is counitary).
The map
M
is called the comodule structure map of M. If M and M
are
(left) A
-comodules, a map of A
-comodules is a A
0
-linear map : M M
such that the following square is commutative

M A
1
R
A
0
M
M
A
1
R
A
0
M
M
id
We denote by A
-coMod the category of (left) A
-comodules. All comodules

considered in this paper will be on the left so we will speak simply of comod-
ules.
If M is an A
0
-module one can always dene a structure of A
-comodule
on A
1
R
A
0
M by dening
A
1
R
A
0
M
:= id. This structure is called the
extended comodule structure on A
1
R
A
0
M.
Theorem5.3. If A
1
is at as an A
0
-module, then the category A
-coMod is a
Grothendieck category.
Proof. In [R, A1.1.3.] it is proved that A
-coMod is an abelian category if A

1
is a at A
0
-module. Therefore what is left to prove is that this category has
exact direct limits and a set of generators.
Since tensor product commutes with colimits, the A
0
-module colimit of a
diagram in A
-coMod is an A
-comodule. Direct limits of A
-comodules are
exact, because direct limits are exact in the category A
0
-Mod. Since A
1
is at
as an A
0
-module, the kernel as an A
0
-module of a A
-comodule morphism
inherits the structure of A
-comodule.
Finally we show that the set S of A
-subcomodules of A
n
1
for n N (viewed
as A
-comodules via ) is a set of generators of A
-coMod. Indeed, let M be a

left A
-comodule, M=0. The structure morphism =

M
: MA
1
R
A
0
M
is injective as a morphism of A
-comodules, considering the structure of ex-

tended comodule on A
1
R
A
0
M. Denote by q: A
1
R
A
0
MN the cokernel
of and let p: A
(I)
0
M be a free presentation of M as an A
0
-module. Since
the morphism idp: A
1
R
A
0
A
(I)
0
A
1
R
A
0
M is a morphism of comodules
and A
(I)
1

= A
1
R
A
0
A
(I)
0
as comodules, we have a surjective morphism of co-
modules p: A
(I)
1
A
1
R
A
0
M. Given x M, x = 0, there are n N, y A
n
1
such that p j(y) = (x) where j : A
n
1
A
(I)
1
is the corresponding canonical
morphism. If K is the kernel of q pj : A
n
1
N, K S and y K, then ( pj)|
K
factors through a non-zero map, i.e. Hom
A
-coMod
(K, M) =0.
Remark. We remind the reader that if A
is a Hopf algebroid arising from a

geometric stack, A
1
is smooth over A
0
and, therefore, at.
5.4. The category of descent data in A
. Let A
be a Hopf algebroid as
before, and M an A
0
-module. Write A
2
= A
1
R
L
A
1
. A descent datum on M
is an isomorphism : A
1
L
A
0
MA
1
R
A
0
M of A
1
-modules veryfying that
the diagram
A
2 j
1
A
0
M A
2 j
1
A
1
A
1
L
A
0
M A
2 j
1
A
1
A
1
R
A
0
M
A
2
A
1
A
1
L
A
0
M
A
2
A
1
A
1
R
A
0
M A
2 j
2
A
1
A
1
L
A
0
M
A
2
R
A
0
M A
2 j
2
A
1
A
1
R
A
0
M
id
j
1
R
= j
2
L
id
id
(5.4.1)
commutes.
We dene the category Desc(A
) by taking as its objects the pairs (M, )

with M an A
0
-module and a descent datum on M, and as its morphisms
f : (M
1
,
1
) (M
2
,
2
) those homomorphism of A
0
-modules f : M
1
M
2
such
that the diagram
A
1
L
A
0
M
1
A
1
R
A
0
M
1
A
1
L
A
0
M
2
A
1
R
A
0
M
2
1
idf idf
2
commutes.
For future reference, note that commutativity of diagram (5.4.1) is equiva-
lent to the following identity in A
1
R
L
A
1
R
A
0
M
n
i=1
ac
i
b(1x
i
) =
n
i=1
(ab)(c
i
) x
i
(5.4.2)
where a, b A
1
and x M, with (1x) =
n
i=1
c
i
x
i
.
Proposition 5.5. The categories Desc(A
) and A
-coMod are isomorphic.

Proof. Let (M, ) Desc(A
). To give a structure of A
-comodule on M one
denes =
M
: MA
1
R
A
0
M as the following composition
M

L
id
A
1
L
A
0
M

A
1
R
A
0
M.
Let us see that is counitary. Let x M, with notation as in 5.4.2 we have
(id) (x) =
n
i=1
(c
i
) x
i
.
To see that

n
i=1
(c
i
) x
i
= x, we apply the homomorphism idid to both
sides of the identity (5.4.2) for a =b =1 to get
n
i=1
(
L
(c
i
) x
i
) =
n
i=1
c
i
x
i
,
and since is an isomorphism we deduce that
n
i=1
L
(c
i
) x
i
=1x.
Thus
n
i=1
(c
i
)x
i
=(id)
n
i=1
L
(c
i
) x
i
=(id)(1x) = x.
The coassociativity of follows from the identity (5.4.2) for a = b =1. Now, if
f : (M
1
,
1
) (M
2
,
2
) is a morphim in Desc(A
), then
(idf )
1
=(idf )
1
(
L
id) =
2
(idf )(
L
id) =
2
(
L
id) f =
2
f
so f : (M
1
,
1
) (M
2
,
2
) is a morphism of A
-comodules. It follows that this

denes a functor C: Desc(A
) A
-coMod.
Conversely, let (M, ) be an A
-comodule. We dene
: A
1
L
A
0
MA
1
R
A
0
M
as the composition
A
1
L
A
0
M
id
A
1
L
L
A
1
R
A
0
M
id
A
1
R
A
0
M.
It easy to prove that is an homomorphism of A
1
-modules. We show that
is an isomorphism by constructing its inverse. We dene
as the following
composition
A
1
R
A
0
M
id
A
1
R
L
A
1
R
A
0
M
idid
A
1
R
R
A
1
L
A
0
M
id
A
1
L
A
0
M.
Let us see that and
are mutually inverse. Consider the commutative

diagramin which all tensor products are taken over A
0
(with most subindices
omitted)
A
1
R
M A
1
A
1
M A
1
A
1
M A
1
L
M
A
1
A
1
M A
1
A
1
A
1
M A
1
A
1
A
1
M A
1
A
1
M
A
1
A
1
M A
1
R
M
id
idid
id
idid idid
idid
id
id idid idid id
idid id
then we have
=(id)(idid)(idid)(idid)(id)
=(id)(id(id)id)(id)
=(id)(id
R
id)(id) =id.
A similar computation shows that
=id.
Finally, it remains to prove that veries (5.4.2). Since is coassociative,
we have for each x M
n
i=1
c
i
(x
i
) =
n
i=1
(c
i
) x
i
(5.5.1)
where (x) =
n
i=1
c
i
x
i
. But (1x) = (x), for each x M, and the result
follows multiplying by ab A
2
in both sides of (5.5.1).
Now, if f : (M
1
,
1
) (M
2
,
2
) is a morphism of A
-comodules, and
1
and
2
are their corresponding descent data, then
(idf )
1
=(id)(ididf )(id
1
) =
2
(idf )
and so f : (M
1
,
1
) (M
2
,
2
) is a morphism in Desc(A
). It follows that this

dene a functor D
: A
-coMod Desc(A
).
Note that the functors C and D
are inverse to each other. In fact, let

(M, ) Desc(A
) and : M A
1
R

A
0
M be its associated A
-comodule
structure. The descent data associated to the comodule (M, ) is the com-
position (id)(id) = . Conversely, let (M, ) be an A
-comodule. If is
the corresponding descent data on M, the structure map of the A
-comodule
associated to (M, ) is the morphism (
L
1) =.
5.6. Let X = Spec(A) be an afne scheme. Recall that by Theorem 3.11 the
categories Qco(X) and Crt(X) are the same. Correspondingly, the functors

()
and ()
crt
are identied, too. We will stick to the former notation according to
the usage of [EGA I, (1.3)].
Proposition 5.7. The couple of functors
(X, ) : Qco(X) A
0
-Mod and

() : A
0
-Mod Qco(X)
induces an equivalence of categories between Desc
qc
(X/X) and Desc(A
), that
we will keep denoting (X, ) and

().
Proof. If (G, t) Desc
qc
(X/X), we will prove that the morphism
:=(G(p
2
)
a
)
1
t
1
(X
1
, id
X
1
) G(p
1
)
a
is a descent datum on G(X).
Consider the Cartesian square
X
2
X
1
X
1
X
q
2
q
1
p
2
p
1
where X
2
= Spec(A
1
R
L
A
1
), q
1
= Spec( j
2
) and q
2
= Spec( j
1
). The multi-
plication of the scheme in groupoids (X, X
1
) is the morphism m: X
2
X
1
verifying p
1
m= p
1
q
2
, p
2
m= p
2
q
1
and m=q
1
q
2
, with notations as in
(4.1.1). Let
w: X
2
X
X
X
X
X (5.7.1)
be the isomorphism given by p
12
w = q
2
, p
3
w = p
2
q
1
and
w = q
1
with
as in 4.2.1. Its inverse w

1
, already dened on 5.1, is the unique morphism
verifying q
1
w
1
= p
23
and q
2
w
1
= p
12
. As a consequence
p
12
w= q
2
p
23
w= q
1
and p
13
w= m.
Thus, the cocycle condition of t on G reads
q
2
t q
1
t =m
t
or, equivalently,
t(X
2
, q
2
)t(X
2
, q
1
) = t(X
2
, m).
This last identity and the fact that t is a morphism of sheaves yields the
commutativity of (5.4.1).
Conversely, if (M, ) Desc(A
), using (5.4.1) it is easy to prove that the

following composition
p
M

(A
1
R
A
0
M)

(A
1
L
A
0
M) p
M
is a descent datum on

M.
Theorem 5.8. The categories Desc
qc
(X/X) and A
-coMod are equivalent.

Proof. By Proposition 5.7 the category Desc
qc
(X/X) is equivalent to Desc(A
),
which, in turn is equivalent to A
-coMod by Proposition 5.5.

Remark. This result is similar to [Ho1, Theorem 2.2]. Notice however that
Hoveys quasi-coherent sheaves over X are sheaves F for the (big) discrete
topology while we work in the fppf topology. The ultimate reason of the agree-
ment of both approaches lies in Lemma 3.6 that expresses the property of
descent of Cartesian presheaves for this topology which makes them auto-
matically sheaves. Notice also that the present proof is simpler once we have
made explicit in Proposition 5.7 that all the data involved refers to modules
on an afne scheme and, therefore, it has an essentially algebraic nature.
Corollary 5.9. For X = Stck(A
), the categories Qco(X) and A
-coMod are
equivalent.
Proof. Combine Corollary 4.7 and Theorem 5.8.
We denote by
X
p
: Qco(X) A
-coMod the equivalence of categories de-

ned as
X
p
:=C(X, ) D, with D as in 4.2 and C as in Proposition 5.5. A
quasi-inverse is
X
p
=G
() D
.
Corollary 5.10. For X=Stck(A
), Qco(X) is a Grothendieck category.

Proof. Combine Theorem 5.3 with Corollary 5.9.
Proposition 5.11. We have
(i) If (M, ) is an A
-comodule and F is its associated O

X
-module, then
the

A
0
-module p
F is

M.
(ii) If G Qco(X) then the comodule
X
p
(p
G) is isomorphic to the ex-

tended comodule A
1
R
A
0
G(X).
Proof. By Theorem 5.8 (M, ) corresponds to (
M, t) Desc
qc
(X/X). By Corol-
lary 5.9 p
F = p
G(
M, t)
M. This gives (i).

For (ii), we have by 2.2 that

G(X) and G are isomorphic

A
0
-modules and
then the comodules
X
p
(p
G(X)) = A
1
R
A
0
G(X) and
X
p
(p
G) = G(X
1
, p
2
)
are isomorphic via G(p
2
)
a
: A
1
R
A
0
G(X) G(X
1
, p
2
). To prove (ii) is suf-
cient to show that the comodule structure on A
1
R
A
0
G(X) given by 4.2,
Proposition 5.7 and Proposition 5.5, is id. Set M = G(X). The descent
datum on A
1
R
A
0
M is dened by
=(p
M(p
2
, id)
a
)
1
p
M(id,
1
) (p
M(p
1
, id))
a
.
The structure map is the composition
A
1
R
A
0
M

L
id
A
1
L
L
A
1
R
A
0
M

A
1
R
L
A
1
R
A
0
M.
Consider the isomorphism w: X
2
X
X
X
X
X in (5.7.1). We obtain
=(p
M(p
2
, id)
a
)
1
p
M(id,
1
)p
M(p
1
, id)
=(p
M(p
2
, id)
a
)
1
p
M(p
1
,
1
)
=(p
M(p
2
, id)
a
)
1
M(w)
1
M(w)
M(p
13
)
=
M(m) =id.
the last equalities hold because

M(w)p
M(p
2
, id)
a
=id and m=Spec().
6. PROPERTIES AND FUNCTORIALITY OF QUASI-COHERENT SHEAVES
We have reached our rst main objective: representing quasi-coherent
sheaves on a geometric stack. It is time to move forward to the next one,
namely, describing a nice functoriality formalism for these kind of a sheaves
for an arbitrary 1-morphism of stacks. Let then f : XY be a 1-morphism of
geometric stacks.
Unfortunately, f does not induce a (continuous) functor from A
fppf
/Y to
A
fppf
/X, so we can not apply directly the methods from 1.7 and 1.9. Of
course, it could be done if f is an afne 1-morphism, but this condition is
clearly too restrictive.
Besides, there is no hope to dene a topos morphism from X
fppf
to Y
fppf
because the topology is too ne, however we will construct an adjunction that
will induce the corresponding pair of adjoint functors between the categories
of quasi-coherent sheaves.
6.1. Direct image. For xed (V, v) A
fppf
/Y we denote by J(v, f ) the cate-
gory whose objects are the 2-commutative squares
U V
X Y
h
u v
f
(6.1.1)
where (U, u) A
fppf
/X. We denote this object by (U, u, h, ). A morphism
from (U, u, h, ) to (U
, u
, h
) is just a morphism (l, ) : (U, u) (U
, u
) in
A
fppf
/X verifying h
l = h and =
l f . If G Pre(A
fppf
/X) we dene the
presheaf f
G by:
( f
G)(V, v) = lim
J(v, f )
G(U, u).
It is a presheaf. Let (V
1
, v
1
), (V
2
, v
2
) A
fppf
/Y and ( j, ) : (V
1
, v
1
) (V
2
, v
2
).
In this situation we dene a functor
( j, )
: J(v
1
, f ) J(v
2
, f )
by ( j, )
(U, u, h, ) :=(U, u, jh, h). This induces the restriction morphism

( f
G)(V
2
, v
2
) ( f
G)(V
1
, v
1
).
Notice that if f is an afne morphism, then f
G(V, v) =G(X
Y
V, p
1
) be-
cause (X
Y
, p
1
, p
2
,
0
), with
0
the canonical 2-cell in the 2-pullback, is an
nal object of J(v, f ).
6.2. Inverse image. Let (U, u) A
fppf
/X, we denote by I(u, f ) the category
whose objects are the 2-commutative squares as 6.1.1 with (V, v) A
fppf
/Y.
We denote these objects now as (V, v, h, ). A morphism from (V, v, h, ) to
(V
, v
, h
) is a morphism ( j, ) : (V, v) (V
, v
) in A
fppf
/Y verifying jh =h
and
=h. For F Pre(A

fppf
/Y), the presheaf f
p
F is given by
( f
p
F)(U, u) = lim
I(u, f )
F(V, v)
for each (U, u) A
fppf
/X. It is also a presheaf. Consider (U
1
, u
1
), (U
2
, u
2
)
A
fppf
/X together with (l, ) : (U
1
, u
1
) (U
2
, u
2
). In this situation we dene
a functor
(l, )
: I(u
2
, f ) I(u
1
, f )
by (l, )
(V, v, h, ) :=(V, v, hl, l f ). This induces the restriction morphism

( f
p
F)(U
2
, u
2
) ( f
p
F)(U
1
, u
1
).
If f is a at nitely presented morphism, then, f
p
F(U, u) = F(U, f u) be-
cause (U, f u, id
U
, id) is an initial object of I(u, f ).
Lemma 6.3. Let f
1
, f
2
: X Y be 1-morphisms of geometric stacks. In the
previous setting, any 2-morphism: f
1
f
2
induces equivalences of categories
(i)
J
: J(v, f
1
) J(v, f
2
),
(ii)
I
: I(u, f
2
) I(u, f
1
).
Proof. For (i), dene, using the notation as in 6.1,
J
(U, u, h, ) =(U, u, h,
1
u)
and extend in an obvious way to maps. We see that these data denes a
functor and a quasi-inverse is given by (
1
)
J
.
Analogously, for (ii), dene, using the notation as in 6.2,
I
(V, v, h, ) =(V, v, h, u)
As before, this denes a functor whose quasi-inverse is (
1
)
I
.
Proposition 6.4. Let f
1
, f
2
: XY be 1-morphisms of geometric stacks. With
the previous notation, given a 2-morphism : f
1
f
2
we have the following
isomorphisms
(i)
: f
1
f
2
,
(ii)
p
: f
p
2
f
p
1
.
Proof. To prove (i), apply the equivalence
J
: J(v, f
1
) J(v, f
2
) to the sets
G(U, u) with (U, u) J(v, f
1
), this gives a map
lim
J(v, f
1
)
G(U, u) lim
J(v, f
2
)
G(U, u).
This isomorphism is natural, therefore induces a natural isomorphism
f
1
G f
2
G.
For (ii) use a similar argument with
I
in place of
J
.
Proposition 6.5. The previous constructions dene a pair of adjoint functors
Pre(A
fppf
/X)
f
p
Pre(A
fppf
/Y).
Proof. We have to prove, for F Pre(A
fppf
/Y) and G Pre(A
fppf
/X) that
there is an isomorphism
Hom
X
( f
p
F, G) Hom
Y
(F, f
G)
A map f
p
F G induces for every (U, u) A
fppf
/X and (V, v) I(u, f ) mor-
phisms
F(V, v) lim
I(u, f )
F(V, v) G(U, u)
Now xing (V, v) and varying (U, u) we get a map
F(V, v) lim
J(v, f )
G(U, u).
This construction provides the adjunction map. By a similar procedure, we
obtain a map in the opposite direction. It is not difcult to check that both
maps are mutually inverse.
6.6. The continuity of our construction is expressed by the relation between
the coverings. Let us spell it out.
Take a covering {(g
i
,
i
) : (V
i
, v
i
) (V, v)}
iI
in the site A
fppf
/Y. Given
(U, u, h, ) J(v, f ), consider for each i I the following Cartesian square of
afne schemes
U
V
V
i
V
i
U V
h
i
g
i
g
i
h
Set U
i
:=U
V
V
i
and
i
: f ug
i
v
i
h
i
, given by
i
:=
1
i
h
i
g
i
. This denes
a functor
g
i
: J(v, f ) J(v
i
, f ) (6.6.1)
by g
i
(U, u, h, ) :=(U
i
, ug
i
, h
i
,
i
).
A consequence of this observation is the following important statement.
Proposition 6.7. If G is a sheaf on A
fppf
/X, then f
G is a sheaf.
Proof. Let {(g
i
,
i
) : (V
i
, v
i
) (V, v)}
iI
be a covering in the site A
fppf
/Y.
With the previous notation, given (U, u, h, ) J(v, f ), the family
{(g
i
, id) : (U
i
, ug
i
) (U, u)}
iI
is a covering in the site A
fppf
/X. Associated to the sheaf G we have an
equalizer diagram
G(U)

iI
G(U
i
)

i, jI
G(U
i
U
U
j
).
1
Taking limits, we obtain another equalizer diagram (in the following we will
keep the notation on the maps to avoid the clutter)
lim
J(v, f )
G(U) lim
J(v, f )
iI
G(U
i
) lim
J(v, f )
i, jI
G(U
i
U
U
j
).
1
That, by commutation of limits [Bo, Prop. 2.12.1], it may be rewritten as
lim
J(v, f )
G(U)

iI
lim
J(v, f )
G(U
i
)

i, jI
lim
J(v, f )
G(U
i
U
U
j
).
1
Let us denote by J
i
(v, f ) the essential image of J(v, f ) in J(v
i
, f ) through the
functor g
i
(6.6.1). Note that
lim
J(v, f )
G(U
i
) = lim
J
i
(v, f )
G(U
i
)
All of this applies verbatim to J
i j
(v, f ).
Consider now the following commutative diagram
lim
J(v, f )
G(U)

iI
lim
J
i
(v, f )
G(U
i
)

i, jI
lim
J
i j
(v, f )
G(U
i
U
U
j
)
f
G(V)

iI
f
G(V
i
)

i, jI
f
G(V
i
V
V
j
)
1
can can
In a natural way, for every U =(U, u, h, ) J(v
i
, f ) we have (U, u, g
1
h, h)
J(v, f ). Then we get
U
i
= g
i
(U, u, g
1
h, h) J
i
(v, f )
jointly with maps in J(v
i
, f ), U U
i
and U
i
U whose composition is the
identity. This implies that the canonical map
f
G(V
i
)
can
lim
J
i
(v, f )
G(U
i
)
is a monomorphism. Similarly the rightmost vertical map is an monomor-
phism, too. The top row is an equalizer and this implies that the bottom row
is, as a consequence, f
G a sheaf.
Corollary 6.8. Let F Y
fppf
. We denote by f
1
F the sheaf associated to the
presheaf f
p
F. There is a pair of adjoint functors
X
fppf
f
1
Y
fppf
.
Proof. Combine the previous Propositions 6.5 and 6.7, having in mind the
adjoint property of sheacation.
Remark. Notice, however, that, in general, the pair ( f
1
, f
) does not dene a

morphism of topoi, since f
1
need not be left exact. For an explicit example
see [Be, 4.4.2]. In the case in which f is a at nitely presented morphism,
then f
1
is exact.
Proposition 6.9. Let f : XY and g: Y Z be 1-morphisms of geometric
stacks. It holds that (gf )
= g
Proof. Let F X
fppf
. For each (W, w) A
fppf
/Z, we have that
(gf )
F(W, w) = lim
J(w, gf )
F(U, u)
where J(w, gf ) is as in 6.1 for (W, w). On the other hand
g
F(W, w) = lim
J(w, g)
f
F(V, v)
We have a family of maps compatible with morphisms in J(w, g)
(gf )
F(W, w) f
F(V, v)
dening a natural map
(gf )
F(W, w) g
F(W, w). (6.9.1)

Let us show that this map is an isomorphism.
Next, for each (U, u, h, ) J(w, gf ) we will construct a map
g
F(W, w) F(U, u)
compatible with morphisms in J(w, gf ). Consider the 2-bered product stack
Y
:= Y
Z
W, its canonical projections q
1
and q
2
, and
0
: gq
1
wq
2
its
associated 2-morphism. Y
is a geometric stack by Proposition 3.3. There is

a morphism f
: U Y
given by the 2-isomorphism : gf u wh, such that

q
1
f
= f u, h = q
2
f
and
0
f
=. Let q: V
be a presentation. Consider
the 2-ber squares
U
U Y
q
f
2
p
1
q
q
Denote by v
: V
Y the map v
= q
1
q. It is clear that (V
, v
, q
2
q,
0
q)
J(w, g). Take v
:= v
1
. In J(w, g), we have the projections (p
1
, id) and
(p
2
, q
1
) from (V
, v
, q
2
qp
1
,
0
qp
1
) to (V
, v
, q
2
q,
0
q). We have a diagram
g
F(W, w) f
F(V
, v
) f
F(V
, v
)
F(U, u) F(U
, u
) F(U
U
U
, u
)
f
F(p
1
, id)
f
F(p
2
, q
1
F(p
1
, id)
F(p
2
, id)
F(q
, id)
can can
where p
1
, p
2
: U
U
U
are the canonical projections, u
=uq
, u
= u
p
1
and the vertical maps labelled can are induced by the limit diagrams in-
dexed by J(v
, f ) and J(v
, f ). The map is the canonical map corresponding

to the limit of the diagram indexed by J(w, g).
Notice that q is smooth and surjective and the same holds for q
therefore,
{q
: U
U} is a (at) covering in A
fppf
/X. As F is a sheaf, the lower row is
an equalizer diagram. We have that f
F(p
1
, id)= f
F(p
2
, q
1
), therefore
there is a canonical vertical dashed map. Passing to the limit along J(w, gf )
we obtain a natural map
g
F(W, w) (gf )
F(W, w)
which is inverse to 6.9.1.
Corollary 6.10. In the previous situation (gf )
1
f
1
g
1
.
Proof. It is is a consequence of the previous proposition by adjunction.
6.11. In general, for f : XY, there is no map between ringed sites, however,
we consider the morphism of sheaves of rings f
#
: O
Y
f
O
X
, given for each
(V, v) A
fppf
/Y by the induced ring homomorphism
f
#
(V,v)
: O
Y
(V, v) lim
J(v, f )
O
X
(U, u).
We want to extend the previous adjunction to the categories of sheaves of
modules. As we do not have neither a topos morphism neither a continuous
map of sites, this extension is not straightforward.
The functor f
takes O
X
-Modules to f
O
X
-Modules and, by the forgetful
functor associated to f
#
, to O
Y
-Modules. Note that the formula in Proposition
6.9 keeps holding by the transitivity of the forgetful functors.
We will apply Lemma 1.8 to guarantee that the functor f
p
commutes with
nite products and so does f
1
. It will follow that if F is an O
Y
-module,
f
1
F will be an f
1
O
Y
-module. Therefore we have to check the hypothesis
of lemma 1.8, so we are reduced to prove the following.
Lemma 6.12. The category I(u, f ) has nite products.
Proof. We only need to prove that any two objects in I(u, f ) admit a product.
For that, let (V
1
, v
1
, h
1
,
1
) and (V
2
, v
2
, h
2
,
2
) be objects in I(u, f ). For i
{1, 2}, let p
i
be the projection from V
1
Y
V
2
to V
i
, and : v
1
p
1
v
2
p
2
be
the canonical 2-morphism. If h: U V
1
Y
V
2
is the morphism verifying
p
1
h = h
1
, p
2
h = h
2
and h =
2

1
1
, then the object (V
1
Y
V
2
, v
1
p
1
, h,
1
)
together with the morphisms (p
1
, id) and (p
2
, ) is a product of (V
1
, v
1
, h
1
,
1
)
and (V
2
, v
2
, h
2
,
2
) in I(u, f ).
Corollary 6.13. The functor f
: O
Y
-Mod O
X
-Mod given by
f
F :=O
X
f
1
O
Y
f
1
F
is left adjoint to the functor f
: O
X
-Mod O
Y
-Mod.
Proof. Combine Corollary 6.8 with the previous discussion.
Corollary 6.14. Let f : XY and let g: YZ be 1-morphisms of geometric
stacks. It holds that (gf )

= f
.
Proof. It is a consequence of Proposition 6.9 by adjunction.
Corollary 6.15. Let f
1
, f
2
: XY 1-morphisms of geometric stacks. With the
previous notation, given a 2-morphism : f
1
f
2
we have isomorphisms
(i)
: f
1
f
2
(ii)
: f
2
f
1
as functors dened on O
X
-Mod and O
Y
-Mod, respectively.
Proof. Immediate from 6.4 and the previous discussion.
Remark. Notice that the isomorphism
in 4.2 is a particular case of (ii).

Proposition 6.16. If F Qco(Y) then f
F Qco(X).
Proof. Let p: X X and q : Y Y be afne presentations such that we have
a 2-commutative square
X Y
X Y
f
0
p q
f
It follows from the previous Corollary that p
F

= f
0
q
F, via
using
pseudofunctoriality. By Proposition 3.8 q
F Crt(Y) and by Proposition 2.4,

f
0
q
F Crt(X). We conclude applying Lemma 3.10 and Theorem 3.11.

Proposition 6.17. If G Qco(X) then f
G Qco(Y).
Proof. If f is an afne morphism the proof is similar to the given in Propo-
sition 2.4. If f is not afne, take an afne presentation p: X X. The
morphism f p is afne. Let T := p
as in Lemma 3.10 and consider the

following equalizer of quasi-coherent sheaves
G

G
TG
T
G
TG
T
2
G.
Since f
is left exact we have the equalizer of O

Y
-modules
f
G
f
G
f
TG
f
T
G
TG
f
T
2
G.
From Proposition 3.8 and Theorem 3.11 by using Proposition 6.9 it follows
that f
TG and f
T
2
G are quasi-coherent sheaves. Then, f
G is a quasi-
coherent Module since it is the kernel of a morphism of quasi-coherent Mod-
ules.
Corollary 6.18. There is a pair of adjoint functors
Qco(X)
f
Qco(Y).
Proof. Combine the adjunction of Corollary 6.13 with Propositions 6.16 and
6.17.
7. DESCRIBING FUNCTORIALITY VIA COMODULES
Let : A
be a homomorphism of Hopf algebroids. Following [Ho2],

we describe a pair of adjoint functors
B
-coMod A
-coMod.
B
0
A
0
We will see that for the corresponding 1-morphism of stacks f :=Stck()

f : Stck(B
) Stck(A
),
this adjunction corresponds to the previously considered f
between
sheaves of quasi-coherent modules on the small at site.
7.1. A homomorphism of Hopf algebroids : A
is a pair of ring homo-

morphisms
i
: A
i
B
i
with i {0, 1} that respects the structure of a Hopf
algebroid, namely, the following equalities hold
L
=
L
0

1
=
0
R
=
R
0

1
=
1
(
1
1
)=
1
.
7.2. The homomorphism : A
induces a functor
B
0
A
0
: A
-coMod B
-coMod,
sending an A
-comodule (M, ) to the B
-comodule (B
0
A
0
M,
) where
is
given by the composition
B
0
A
0
M
id
B
0
L
A
1
R
A
0
M
id
B
1
R
A
0
M B
1
R
B
0
(B
0
A
0
M)
where (b
0
a
1
) :=
L
(b
0
)
1
(a
1
), for a
1
A
1
and b
0
B
0
.
7.3. From [Ho2, Proposition 1.2.3], we will describe
U
: B
-coMod A
-coMod,
a right adjoint for the functor B
0
A
0
. This will be done in two steps.
Consider rst an extended comodule, that is, given N B
0
-Mod, let us
consider the comodule B
1
B
0
N with the induced structure. We dene
U
(B
1
B
0
N) := A
1
A
0
N
with the induced structure. For maps : B
1
B
0
N B
1
B
0
N
with N, N
B
0
-Mod, then U
() is dened as the following composition

A
1
R
A
0
N
id
A
1
R
L
A
1
R
A
0
N
id
1
id
A
1
R
0
B
1
R
B
0
N
id
A
1
R
0
B
1
R
B
0
N
idid
A
1
R
A
0
N
.
Now, let (N, ) be an arbitrary B
-comodule. Let
: B
1
B
0
N N
be the
cokernel of as B
0
-modules. Let
be the induced comodule structure on N
and consider the following A
-comodule homomorphism:
A
1
R
A
0
N
U
)
A
1
R
A
0
N
.
Dene
U
(N) :=ker(U
))
The extension to morphisms is straightforward by the naturally of kernels.
Proposition 7.4. Given a homomorphism : A
of Hopf algebroids,
there is an associated pair of adjoint functors
B
-coMod A
-coMod.
B
0
A
0
Proof. Let (P, ) B
-coMod and (M,
) A
-coMod. We have to construct

an isomorphism
Hom
B
-coMod
(B
0
A
0
M, P) Hom
A
-coMod
(M, U
(P))
First, if P is extended, i.e. P = B
1
B
0
N for an B
0
-module N, then we have
that
Hom
B
-coMod
(B
0
A
0
M, P) =Hom
B
-coMod
(B
0
A
0
M, B
1
B
0
N)
=Hom
B
0
-Mod
(B
0
A
0
M, N)
=Hom
A
0
-Mod
(M, N)
=Hom
A
-coMod
(M, A
1
A
0
N)
=Hom
A
-coMod
(M, U
(P))
In the general case, P may be obtained as a kernel of a morphism between
extended B
-comodules, in other words, we have an exact sequence

0 P P
with P
and P
extended B
-comodules. Now we have a diagram

Hom
B
(B
0
A
0
M, P) Hom
B
(B
0
A
0
M, P
) Hom
B
(B
0
A
0
M, P
)
Hom
A
(M, U
(P)) Hom
A
(M, U
(P
)) Hom
A
(M, U
(P
))

in which the last two vertical maps are isomorphisms because P
and P
are
extended comodules, therefore they induce the dashed vertical map which is
also an isomorphism as desired.
Remark. The previous result is due to Hovey [Ho2]. We have included a proof
for the readers convenience.
7.5. Our next task is to identify the functor U
with the direct image of

sheaves on the geometric stacks corresponding to the Hopf algebroids.
Let f : XY be a morphism of geometric stacks, p: X X and q: Y Y
afne presentations and f
0
: X Y be a morphism such that the following
square
X Y
X Y
f
0
p q
f
is 2-commutative. Set X=Stck(B
) and Y=Stck(A
). Denote the ring homo-

morphisms associated to f as
0
: A
0
B
0
and
1
: A
1
B
1
. We remind the
reader that =(
0
,
1
) is a homomorphism of Hopf algebroids (see 7.1).
Proposition 7.6. There is a natural isomorphism of functors from Qco(X) to
A
-coMod
:
Y
q
f
X
p
Proof. We will start considering an object in the source that lies in the es-
sential image of the functor T = p
. Let F Qco(X) with p
F =

M, for
M B
0
-Mod. By Proposition 5.11(ii),
X
p
TF =(B
1
B
0
M, id). By 7.3
U
X
p
TF =(A
1
A
0
M, id)
On the other hand, by 2-functoriality of direct images there is a canonical
isomorphism via
1
TF = f
F q
f
0
M= q
M
[
0
]
.
Therefore there is an isomorphism
Y
q
f
TF
Y
q
q
M
[
0
]
and we have
Y
q
q
M
[
0
]
=(A
1
A
0
M, id),
by Proposition 5.11(ii), again. This denes
TF
:
Y
q
f
TF U
X
p
TF. No-
tice that
T()
is a natural transformation.
Finally we extend this to all quasi-coherent sheaves on X and all mor-
phisms. Consider the canonical equalizer of quasi-coherent O
X
-modules
F

F
TF
TF
T
F
T
2
F.
We may thus present every object in Qco(X) as a kernel of objects that lie
in the image of the functor T. Consider the following commutative equalizer
diagram
Y
q
f
F
Y
q
f
TF
Y
q
f
T
2
F
U
X
p
F U
X
p
TF U
X
p
T
2
F.
Y
q
f
TF
Y
q
f
T
F
U
X
p
TF
U
X
p
T
F
TF

T
2
F
Y
q
f
F
U
X
p
F
Notice that U
X
p
T
F
TF
=
T
2
F

Y
q
f
T
F
being
T()
a natural transfor-
mation. Let us check that
U
X
p
TF
TF
=
T
2
F
Y
q
f
TF
.
Applying
X
p
to the map
TF
: TF T
2
F, we get
F(p
13
, id) : F(X
1
, pp
2
) F(X
1
X
X, pp
3
).
In other words, having in mind that p
13
w= m where w: X
2
X
1
X
X is the
isomorphism in 5.7.1 and m= Spec(), taking M := F(X, p), this morphism
corresponds to:
id: B
1
R
A
0
MB
1
R
L
B
1
R
A
0
M.
Now, we apply U
and we obtain the following composition

A
1
R
A
0
M
id
A
1
R
L
A
1
R
A
0
M
id
1
id
A
1
R
L
B
1
R
B
0
M.
On the other hand, consider the 2-cartesian squares
Y
Y
X X
Y Y
1 f p
q
1
(Y
Y
X)
X
X X
Y
Y
X X
1
p
p
2
2
and the morphism l : (Y
Y
X)
X
X Y
Y
X. satisfying that
1
l =
1
1
,
2
l =
2
and
1
l = f
2
1
. Applying
Y
q
f
to
TF
, we have the following
commutative diagram
Y
q
f
TF
Y
q
f
T
2
F
F(Y
Y
X, p
2
) F((Y
Y
X)
X
X, p
2
)
Y
q
f
TF
via
via
F(l, id)
Consider now the diagram
(Y
Y
X)
X
X Y
Y
X
Y
1 p
2
f
0
p
1
X
1
Y
1 p
2
Y
X
l
c
c
l
where c is the composition of the following isomorphisms

Y
1 p
2
Y
X =(Y
Y
Y)
p
2
Y
X Y
q
qf
0
X Y
Y
X,
c
is the isomorphism given by

Y
1 p
2
f
0
p
1
X
1
(Y
1 p
2
f
0
X)
X
X
cid
(Y
Y
X)
X
X
and l
= c
1
l c
. A computation shows that the following composition

Y
1 p
2
f
0
p
1
X
1
Y
1 p
2
f
0
p
1
X
1 p
2
X
X
idf
1
id
Y
1 p
2
p
1
Y
1 p
2
Y
X
s
Y
X
(Y
1 p
1
p
2
Y
1
)
Y
X
mid
Y
1 p
2
Y
X
agrees with l
, where s is the isomorphism that interchanges the factors

of the ber product. But then, considering the isomorphisms Y
1
Y
X =
Spec(A
1
R
A
0
B
0
) and Y
1
Y
X
1
=Spec(A
1
R
L
B
1
), l
is identied the spec-

trum of the composition
A
1
R
A
0
B
0
id
A
1
R
L
A
1
R
A
0
B
0
id
1
id
A
1
R
L
B
1
R
B
0
B
0
.
Finally, as the map F(l
, id) corresponds to the last map tensored with M =

F(X, p) this shows the agreement of both morphisms.
The fact that the rows are equalizers denes the promised dashed natural
isomorphism
F
.
Corollary 7.7. There is a natural isomorphism of functors
X
p
f
B
0
A
0
Y
q
.
Proof. Denote by
Y
q
a quasi-inverse of
Y
q
. The functor
X
p
f
Y
q
is left ad-
joint to U
, so it has to agree with B

0
A
0
.
8. DELIGNE-MUMFORD STACKS AND FUNCTORIALITY FOR THE TALE
TOPOLOGY
In section 6 we have shown how to obtain an adjoint functoriality of the
category of quasi-coherent sheaves on a geometric stack overcoming the dif-
culty of the lack of functoriality of the fppf topos. But one can (and should)
wonder how this construction is related to the case where a functorial topos
exists. This is the case for the tale topos on a Deligne-Mumford stack, as
follows from Theorem 8.7 below. We recall that the corresponding theory of
quasi-coherent sheaves is equivalent to the one in the Zariski topology when-
ever the stacks are equivalent to schemes [SGA 4
2
, VII, 4.3]. The answer is
that our construction agrees with it. Let us see how.
8.1. Let X be a geometric Deligne-Mumford stack and consider the category
A
et
/X whose objects are pairs (U, u) with U an afne scheme and u: U X
an tale morphism, morphisms are as in A/X. Note that all morphisms are
tale. We dene a topology in this category having as coverings nite families
of tale morphisms that are jointly surjective. We denote also by A
et
/X
this site and by X
et
the associated topos of sheaves. The category A
et
/X is
ringed by the sheaf of rings O : A
et
/X Ring, dened by O(U, u) = B with
U =Spec(B). We dene
Qco
et
(X) :=Qco(A
et
/X, O).
As in Theorem 3.11, Qco
et
(X) agrees with the category of Cartesian sheaves,
and we will use this fact freely.
Let f : XY be a 1-morphism of geometric Deligne-Mumford stacks.
8.2. Direct image for the tale topology. Let (V, v) A
et
/Y. Denote by
J
et
(v, f ) the category whose objects are the 2-commutative squares
U V
X Y
h
u v
f
(8.2.1)
where (U, u) A
et
/X. As in 6.1 (with appropriate changes) we dene for
G Pre(A
et
/X),
( f
et
G)(V, v) = lim
J
et
(v, f )
G(U, u).
Again, if f is an afne morphism, then ( f
et
G)(V, v) =G(V
Y
X, p
2
).
8.3. Inverse image for the tale topology. Now, x (U, u) A
et
/Y and
denote by I
et
(u, f ) the category whose objects are the 2-commutative squares
as in diagram (8.2.1), with (V, v) A
et
/Y. By a similar procedure to 6.2
dene for F Pre(A
et
/Y) and (U, u) A
et
/X
( f
p
F)(U, u) = lim
I
et
(u, f )
F(V, v).
Notice that if f is tale, then ( f
p
F)(U, u) =F(U, f u).
Proposition 8.4. The previous constructions dene a pair of adjoint functors
Pre(A
et
/X)
f
p
et
f
et
Pre(A
et
/Y).
Proof. The proof is completely similar to the one in Proposition 6.5.
Proposition 8.5. The functor f
p
: Pre(A
et
/Y) Pre(A
et
/X) is exact.
Proof. The functor f
p
is right exact, because it has a right adjoint. To prove
that f
p
is left exact it is enough to show that I
et
(u, f ) is a coltered category
for (U, u) A
et
/Y.
By [SGA 4
1
, I, 2.7] we have to prove that the category I
et
(u, f ) is connected
and satises PS 1) and PS 2) in loc. cit. The category I
et
(u, f ) is connected
since it has nite products (analogous to the fppf case, Lemma 6.12).
To prove PS 1), for i {1, 2} take ( j
i
,
i
) : (V
i
, v
i
, h
i
,
i
) (V, v, h, ) mor-
phisms in I
et
(u, f ), the projections p
i
: V
1
V
V
2
V
i
and h
: U V
1
V
V
2
the morphism verifying p
i
h
= h
i
. We have that (V
1
V
V
2
, v
1
p
1
, h
,
1
) is in
I
et
(u, f ) and the morphisms (p
1
, id) and (p
2
,
1
2
p
2
1
p
1
) satisfy
( j
1
,
1
) (p
1
, id) =( j
2
,
2
) (p
2
,
1
2
p
2
1
p
1
).
Finally, we prove PS 2): Let (g, ), (g
) : (V, v, h, ) (V
, v
, h
) be mor-
phisms in I
et
(u, f ) and consider the pull-back
V
Y
V
Y
p
2
p
1 v
We have the morphism r : V V
Y
V
verifying p
1
r = g, p
2
r = g
and r =
1
. Consider the Cartesian square
Z V
V V
Y
V
q
2
q
1
r
and the morphism l : U Z such that q
1
l = h and q
2
l = h
. Notice that
q
1
is tale because is an open embedding and therefore, tale. The object
(Z, vq
1
, l, ) I
et
(u, f ) and the morphism (q
1
, id) : (Z, vq
1
, l, ) (V, v, h, ) is
the equalizer of (g, ) and (g
).
Remark. Notice that this result, specically PS 2), is false for ner topologies.
Proposition 8.6. If F is a sheaf on A
et
/X, then f
et
F is a sheaf.
Proof. It can be proven similarly to Proposition 6.7.
Theorem 8.7. Let f : XY be a 1-morphism of geometric Deligne-Mumford
stacks, there is an associated morphism of topos
f
et
: X
et
Y
et
Proof. Let G Y
et
, we denote by f
1
et
G the sheaf associated to the presheaf
f
p
G. This, together with Proposition 8.6 provides an adjunction
X
et
f
1
et
f
et
Y
et
The pair ( f
et
, f
1
et
) denes a morphism of topoi, since f
1
et
is exact because so
is f
p
and the sheacation functor.
Proposition 8.8. Let f : XY and g: Y Z be 1-morphisms of geometric
Deligne-Mumford stacks. It holds that (gf )
et
= g
et
f
et
Proof. The proof is mutatis mutandis the same as Proposition 6.9.

Corollary 8.9. In the previous situation (gf )
1
et
f
1
et
g
1
et
.
Proof. It is is a consequence of the previous proposition by adjunction.
8.10. Let f : XY be a 1-morphism of geometric Deligne-Mumford stacks.
The ring homomorphism f
#
et
: O
Y
f
O
X
induces a canonical morphism of
ringed topoi
( f
et
, f
et
#
) : (X
et
, O
X
) (Y
et
, O
Y
)
where f
et
#
: f
1
O
Y
O
X
is the morphism adjoint to f
#
et
. Now we can dene
the functor f
et
: O
Y
-Mod O
X
-Mod by the usual formula
f
et
F :=O
X
f
1
et
O
Y
f
1
et
F.
It is left adjoint to f
et
: O
X
-Mod O
Y
-Mod.
Proposition 8.11. In the previous situation, we have
(i) if G Qco
et
(Y) then f
et
G Qco
et
(X), and
(ii) if F Qco
et
(X) then f
et
F Qco
et
(Y).
Proof. This can be shown in a similar way to Propositions 6.16 and 6.17.
Corollary 8.12. There is a pair of adjoint functors
Qco
et
(X)
f
et
f
et
Qco
et
(Y).
Proof. Combine the previous proposition with 8.10.
For X a geometric Deligne-Mumford stack we denote by r
X
: X
fppf
X
et
the
restriction functor.
Proposition 8.13. Let f : XY be a 1-morphismof geometric Deligne-Mum-
ford stacks. There is an isomorphismof functors fromX
fppf
to Y
et
r
Y
f
= f
et
r
X
.
Proof. As p is an afne morphism, then r
X
p
= p
et
r
X
. In general, let F
X
fppf
. Recall that we have an equalizer in X
fppf
F

F
TF
T
F
TF
T
2
F, (8.13.1)
with T= p
. By Proposition 6.9 and Proposition 8.8, using also that f p is

an afne morphism we have that
r
Y
f
=r
Y
( f p)
=( f p)
et
r
X

= f
et
p
et
r
X
From this it follows that r
Y
f
TF = f
et
r
X
TF and we construct the diagram
whose rows are equalizers
r
Y
f
F r
Y
f
TF r
Y
f
T
2
F
f
et
r
X
F f
et
r
X
TF f
et
r
X
T
2
F

that shows existence of the dashed isomorphism. Notice that this isomor-
phism is clearly compatible with morphisms.
Notice that if G Qco(X), then r
X
G Qco
et
(X), an obvious fact if we con-
sider the Cartesian characterization. We also denote by r
X
the induced func-
tor between the categories of quasi-coherent sheaves.
Proposition 8.14. The functor r
X
: Qco(X) Qco
et
(X) is an equivalence of
categories.
Proof. Let Desc
et
qc
(X/X) be the category whose objects are the pairs (G, t),
where G Qco
et
(X) and t : p
2, et
G p
1, et
G is an isomorphism verifying the
usual cocycle condition. An argument similar to the one that proves Proposi-
tion 4.6 yields an equivalence of categories
D
et
: Qco
et
(X) Desc
et
qc
(X/X).
Indeed, set X
1
:= X
X
X. We have natural isomomorphisms of functors
w
i
: p
i
_
G(X)
_
i
G(X
1
), i {1, 2}
dened, for (U, u) A
fppf
/X, with U =Spec(B) and u =Spec() by
w
i
(U, u): B
A
0
G(X) B
A
1
A
1
L
A
0
G(X)
idG(p
i
)
a
B
A
1
G(X
1
, p
i
)
with i {1, 2}. The induced functor
r
X/X
: Desc
qc
(X/X) Desc
et
qc
(X/X)
given by r
X/X
(G, t) = (r
X
G, r
X
1
t) is also an equivalence of categories; a quasi-
inverse of r
X/X
is the functor W
X/X
: Desc
et
qc
(X/X) Desc
qc
(X/X) given by
W
X/X
(G, t) =(
G(X), w
1
1

(t(X
1
, id
X
1
))w
2
)
The result follows from the fact that the following diagram
Qco(X) Qco
et
(X)
Desc
qc
(X/X) Desc
et
qc
(X/X)
r
X
D D
et
r
X/X
commutes because r
X
p
G = p
et
r
X
G.
Remark. The same strategy used in Propositions 8.13 and 8.14 may be used
to prove that on an ordinary scheme the quasi-coherent sheaves over the
Zariski topology agree with those dened like here on the tale topology. We
leave the details to the interested readers.
Corollary 8.15. In the previous setting, there is a natural isomorphism of
functors from Qco(Y) to Qco
et
(X)
r
X
f
et
r
Y
.
Proof. If the functor s
Y
: Qco
et
(Y) Qco(Y) is a quasi-inverse of r
Y
, then the
functor r
X
f
s
Y
is left adjoint to f
et
.
Remark. The previous discussion expresses the fact that the formalism of
functoriality that we established in 6 agrees with the usual formalism de-
rived from the tale topos map induced from a 1-morphism of Deligne-Mum-
ford stacks on the category of quasi coherent sheaves.
REFERENCES
[]AJPV Alonso Tarro, L.; Jeremas Lpez, A.; Prez Rodrguez, M.; Vale Gonsalves M. J.:
The derived category of quasi-coherent sheaves and axiomatic stable homotopy.
Adv. Math. 218 (2008) 12241252.
[]A Artin, M.: Grothendieck Topologies, Seminar Notes. Harvard Univ. Dept. of Math.,
Spring 1962.
[]SGA 4
1
Artin, M.; Grothendieck, A.; Verdier, J. L. (avec la collaboration de Bourbaki,
N.; Deligne, P.; Saint-Donat, B.): Thorie des topos et cohomologie tale des sch-
mas. Tome 1: Thorie des topos. (French) Sminaire de Gomtrie Algbrique du
Bois-Marie 19631964 (SGA 4). Lecture Notes in Mathematics, Vol. 269. Springer-
Verlag, Berlin-New York, 1972.
[]SGA 4
2
Artin, M.; Grothendieck, A.; Verdier, J. L. (avec la collaboration de Bourbaki, N.;
Deligne, P.; Saint-Donat, B.): Thorie des topos et cohomologie tale des schmas.
Tome 2. (French) Sminaire de Gomtrie Algbrique du Bois-Marie 19631964
(SGA 4). Lecture Notes in Mathematics, Vol. 270. Springer-Verlag, Berlin-New
York, 1972.
[]SGA 4
3
Artin, M.; Grothendieck, A.; Verdier, J. L. (avec la collaboration de Deligne, P.;
Saint-Donat, B.): Thorie des topos et cohomologie tale des schmas. Tome 3.
(French) Sminaire de Gomtrie Algbrique du Bois-Marie 19631964 (SGA 4).
Lecture Notes in Mathematics, Vol. 305. Springer-Verlag, Berlin-New York, 1973.
[]Be Behrend, K. A.: Derived -adic categories for algebraic stacks. Mem. Amer. Math.
Soc. 163 (2003), no. 774.
[]Bo Borceux, F.: Handbook of categorical algebra. 1. Basic category theory. Encyclope-
dia of Mathematics and its Applications, 50. Cambridge University Press, Cam-
bridge, 1994.
[]SGA 3
1
Demazure, M.; Grothendieck, A.: Schmas en groupes. I: Proprits gnrales
des schmas en groupes. Sminaire de Gomtrie Algbrique 196264 (SGA 3).
Lecture Notes in Mathematics, Vol. 151. Berlin-Heidelberg-New York: Springer-
Verlag (1970).
[]Go Goerss, P.: Quasi-coherent sheaves on the moduli stack of formal groups, preprint,
available at
http://www.math.northwestern.edu/pgoerss/papers/modfg.pdf.
[]SGA 1 Grothendieck, A.: Revetements tales et Groupe Fondamental. Sminaire de
Gomtrie Algbrique du Bois Marie 1960/61. Lecture Notes in Mathematics, Vol.
224. Berlin-Heidelberg-New York: Springer-Verlag, 1971
[]EGA I Grothendieck, A.; Dieudonn, J. A.: Elments de Gomtrie Algbrique I, Grund-
lehren der math. Wiss. 166, SpringerVerlag, Heidelberg, 1971.
[]Ho1 Hovey, M.: Morita theory for Hopf algebroids and presheaves of groupoids. Amer.
J. Math. 124 (2002), no. 6, 12891318.
[]Ho2 Hovey, M.: Homotopy theory of comodules over a Hopf algebroid, in Homotopy the-
ory: relations with algebraic geometry, group cohomology, and algebraic K-theory,
pp. 261304, Contemp. Math., 346, Amer. Math. Soc., Providence, RI, 2004.
[]LMB Laumon, G.; Moret-Bailly, L.: Champs algbriques. Ergebnisse der Mathematik
und ihrer Grenzgebiete, 3. Folge, 39, Springer-Verlag, Berlin, 2000.
[]L Lurie, J.: Tannaka Duality for Geometric Stacks, preprint, available at
http://www.math.harvard.edu/~lurie/papers/Tannaka.pdf.
[]O Olsson, M.: Sheaves on Artin stacks. J. Reine Angew. Math. 603 (2007), 55112.
[]P Pribble, E.: Algebraic stacks for stable homotopy theory and the algebraic chro-
matic convergence theorem, PhD thesis, Evanston, IL., June 2004
[]R Ravenel, D. C.: Complex cobordism and stable homotopy groups of spheres. Pure
and Applied Mathematics, 121. Academic Press, Inc., Orlando, FL, 1986.
[]SP Various authors: Stacks Project, http://stacks.math.columbia.edu
[]V1 Vistoli, A.: Intersection theory on algebraic stacks and on their moduli spaces.
Invent. Math. 97 (1989), no. 3, 613670.
[]V2 Vistoli, A.: Grothendieck topologies, bered categories and descent theory. Funda-
mental algebraic geometry, 1104, Math. Surveys Monogr., 123, Amer. Math. Soc.,
Providence, RI, 2005.
(L. A. T.) DEPARTAMENTO DE LXEBRA, FACULTADE DE MATEMTICAS, UNIVERSIDADE
DE SANTIAGO DE COMPOSTELA, E-15782 SANTIAGO DE COMPOSTELA, SPAIN
E-mail address: leo.alonso@usc.es
(A. J. L.) DEPARTAMENTO DE LXEBRA, FACULTADE DE MATEMTICAS, UNIVERSIDADE
E-mail address: ana.jeremias@usc.es
(M. P. R.) DEPARTAMENTO DE MATEMTICAS, ESC. SUP. DE ENX. INFORMTICA, CAM-
PUS DE OURENSE, UNIVERSIDADE DE VIGO, E-32004 OURENSE, SPAIN
E-mail address: martapr@uvigo.es
(M. J. V.) DEPARTAMENTO DE LXEBRA, FACULTADE DE MATEMTICAS, UNIVERSIDADE
E-mail address: mj.vale@usc.es

1304 2520

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

1304 2520

Uploaded by

Copyright:

Available Formats

a

for the corresponding categories of quasi-

has a nicer geometric description

is easier to describe algebraically.

, O)-Mod or simply O-Mod to

such that f(h) g = g

z =id. The map z

P or, equivalently, its adjoint map

: O-Mod P-Mod is dened by

: O-Mod P-Mod given by

as in the previous situation.

. Notice that if f is a at nitely presented morphism,

F(U, u) =F(U, f u).

F is the sheaf associated to the presheaf P given by

F is isomorphic to the sheaf associated

) and h =Spec(). We have the commutative diagram

) and h =Spec(). We have to prove that the morphism

be a 2-Cartesian square of algebraic stacks. If X, Y and Y

is an algebraic stack by [LMB, (4.5)(i)].

where h is the morphism in the following 2-Cartesian square

The morphism h is afne because it is obtained by base change from

, so by Lemma 3.2, it is afne. It follows that

is afne over X because is the base change of the map

is an afne scheme. The morphism p

where f : U V is a morphism of afne schemes making the diagram 2-

over the ring of global sections of the structure sheaf of X. By faithful

. This makes sense of what follows.

are Cartesian sheaves. It follows that the

is a Cartesian presheaf and then it is

. So we have an adjunction as in 1.9

F(V, v) = F(V, pv) and analogously for (U, u).

F Crt(X), then F Crt(X).

. Since F is a sheaf, we have the equalizer of O

. As a consequence of 3.8 TF,

F sits in an exact sequence

F Crt(X). The result follows

G be the canonical morphism. For each morphism

provides the claimed morphism (p

G given by the adjunction p

factors through G(G, t) giving a

= h is an isomorphism. Let us dene

the isomorphism of functors given by (

applied to F. The rst row is

preserve nite limits, the horizontal row is exact and so by Lemma

such that the following square is commutative

-coMod the category of (left) A

-comodules. All comodules

-coMod is an abelian category if A

-comodule. Direct limits of A

-comodules via ) is a set of generators of A

-coMod. Indeed, let M be a

-comodule, M=0. The structure morphism =

-comodules, considering the structure of ex-

is a Hopf algebroid arising from a

) by taking as its objects the pairs (M, )

-coMod are isomorphic.

-comodules. It follows that this

are mutually inverse. Consider the commutative

). It follows that this

are inverse to each other. In fact, let

as in 4.2.1. Its inverse w

), using (5.4.1) it is easy to prove that the