David Cox
John Little
Hal Schenck
DEPARTMENT OF MATHEMATICS, AMHERST COLLEGE, AMHERST, MA
01002
Email address: dac@math.amherst.edu
DEPARTMENT OF MATHEMATICS AND COMPUTER SCIENCE, COLLEGE OF
THE HOLY CROSS, WORCESTER, MA 01610
Email address: little@mathcs.holycross.edu
DEPARTMENT OF MATHEMATICS, UNIVERSITY OF ILLINOIS AT URBANA
CHAMPAIGN, URBANA, IL 61801
Email address: schenck@math.uiuc.edu
c _2010, David Cox, John Little and Hal Schenck
Contents
Preface vii
Notation xi
Part I. Basic Theory of Toric Varieties 1
Chapter 1. Afﬁne Toric Varieties 3
§1.0. Background: Afﬁne Varieties 3
§1.1. Introduction to Afﬁne Toric Varieties 10
§1.2. Cones and Afﬁne Toric Varieties 23
§1.3. Properties of Afﬁne Toric Varieties 35
Appendix: Tensor Products of Coordinate Rings 48
Chapter 2. Projective Toric Varieties 49
§2.0. Background: Projective Varieties 49
§2.1. Lattice Points and Projective Toric Varieties 54
§2.2. Lattice Points and Polytopes 62
§2.3. Polytopes and Projective Toric Varieties 74
§2.4. Properties of Projective Toric Varieties 85
Chapter 3. Normal Toric Varieties 93
§3.0. Background: Abstract Varieties 93
§3.1. Fans and Normal Toric Varieties 105
§3.2. The OrbitCone Correspondence 114
§3.3. Toric Morphisms 125
§3.4. Complete and Proper 139
iii
iv Contents
Appendix: Nonnormal Toric Varieties 149
Chapter 4. Divisors on Toric Varieties 153
§4.0. Background: Valuations, Divisors and Sheaves 153
§4.1. Weil Divisors on Toric Varieties 168
§4.2. Cartier Divisors on Toric Varieties 174
§4.3. The Sheaf of a TorusInvariant Divisor 187
Chapter 5. Homogeneous Coordinates on Toric Varieties 193
§5.0. Background: Quotients in Algebraic Geometry 193
§5.1. Quotient Constructions of Toric Varieties 203
§5.2. The Total Coordinate Ring 217
§5.3. Sheaves on Toric Varieties 224
§5.4. Homogenization and Polytopes 230
Chapter 6. Line Bundles on Toric Varieties 243
§6.0. Background: Sheaves and Line Bundles 243
§6.1. Ample Divisors on Complete Toric Varieties 260
§6.2. The Nef and Mori Cones 279
§6.3. The Simplicial Case 290
Appendix: Quasicoherent Sheaves on Toric Varieties 301
Chapter 7. Projective Toric Morphisms 305
§7.0. Background: Quasiprojective Varieties and Projective Morphisms 305
§7.1. Polyhedra and Toric Varieties 309
§7.2. Projective Morphisms and Toric Varieties 318
§7.3. Projective Bundles and Toric Varieties 324
Appendix: More on Projective Morphisms 334
Chapter 8. The Canonical Divisor of a Toric Variety 339
§8.0. Background: Reﬂexive Sheaves and Differential Forms 339
§8.1. OneForms on Toric Varieties 350
§8.2. Differential Forms on Toric Varieties 357
§8.3. Fano Toric Varieties 371
Chapter 9. Sheaf Cohomology of Toric Varieties 379
§9.0. Background: Cohomology 379
§9.1. Cohomology of Toric Divisors 390
§9.2. Vanishing Theorems I 401
Contents v
§9.3. Vanishing Theorems II 413
§9.4. Applications to Lattice Polytopes 422
§9.5. Local Cohomology and the Total Coordinate Ring 435
Appendix: Introduction to Spectral Sequences 448
Part II. Topics in Toric Geometry 451
Chapter 10. Toric Surfaces 453
§10.1. Singularities of Toric Surfaces and Their Resolutions 453
§10.2. Continued Fractions and Toric Surfaces 461
§10.3. Gr¨ obner Fans and M
c
Kay Correspondences 479
§10.4. Smooth Toric Surfaces 489
§10.5. RiemannRoch and Lattice Polygons 496
Chapter 11. Toric Resolutions and Toric Singularities 507
§11.1. Resolution of Singularities 507
§11.2. Other Types of Resolutions 519
§11.3. Rees Algebras and Multiplier Ideals 528
§11.4. Toric Singularities 540
Chapter 12. The Topology of Toric Varieties 553
Chapter 13. The RiemannRoch Theorem 555
Chapter 14. GIT and MMP for Toric Varieties 557
§14.1. Toric Geometric Invariant Theory 557
§14.2. The Secondary Fan 557
§14.3. Toric Minimal Model Program 557
Chapter 15. The Toric Minimal Model Program 559
Bibliography 561
Chapter 8
The Canonical Divisor of a
Toric Variety
§8.0. Background: Reﬂexive Sheaves and Differential Forms
This chapter will study the canonical divisor of a toric variety. The theory devel
oped in Chapters 6 and 7 dealt with Cartier divisors and line bundles. As we will
see, the canonical divisor of a normal toric variety is a Weil divisor that is not
necessarily Cartier. We will also study the sheaves associated to Weil divisors.
Reﬂexive Sheaves. A Weil divisor D on a normal variety X gives the sheaf O
X
(D)
deﬁned by
Γ(U, O
X
(D)) =¦f ∈ C(X)
∗
[ (div( f ) +D)[
U
≥0¦∪¦0¦.
Our ﬁrst task is to characterize these sheaves.
Recall that the dual of a sheaf of O
X
modules F is F
∨
= Hom
O
X
(F, O
X
).
We say that F is reﬂexive if the natural map
F −→F
∨∨
is an isomorphism. It is easy to see that locally free sheaves are reﬂexive. Here are
some properties of reﬂexive sheaves.
Proposition 8.0.1. Let F be a coherent sheaf on a normal variety X and consider
the inclusion j : U
0
֒→X where U
0
is open with codim(X ¸U
0
) ≥2. Then:
(a) F
∨
and hence F
∨∨
are reﬂexive.
(b) If F is reﬂexive, then F ≃ j
∗
(F[
U
0
).
(c) If F[
U
0
is locally free, then F
∨∨
≃ j
∗
(F[
U
0
).
339
340 Chapter 8. The Canonical Divisor of a Toric Variety
Proof. Recall from §4.0 that the direct image j
∗
G of a sheaf G on U
0
is deﬁned by
Γ(U, j
∗
G) =Γ(U ∩U
0
, G) for U ⊆X open.
Parts (a) and (b) of the proposition are proved in [79, Cor. 1.2 and Prop. 1.6].
For part (c), we ﬁrst observe that restriction is compatible with taking the dual, i.e.,
(G
∨
)[
U
0
=(G[
U
0
)
∨
for any coherent sheaf G on X. Then
F
∨∨
≃ j
∗
((F
∨∨
)[
U
0
) = j
∗
((F[
U
0
)
∨∨
) ≃ j
∗
(F[
U
0
),
where the ﬁrst isomorphism follows from parts (a) and (b), and the last follows
since F[
U
0
is locally free and hence reﬂexive.
Later in the section we will study the sheaf Ω
p
X
of pforms on X. This sheaf is
locally free when X is smooth. For X normal, however, Ω
p
X
may be badly behaved,
though it is locally free on the smooth locus of X. Hence we can use part (c) of
Proposition 8.0.1 to create a reﬂexive version of Ω
p
X
.
For more on reﬂexive sheaves, the reader should consult [79] and [148].
Reﬂexive Sheaves of Rank One. We ﬁrst deﬁne the rank of a coherent sheaf on an
irreducible variety X. Recall that K
X
is the constant sheaf on X given by C(X).
Deﬁnition 8.0.2. Given a F coherent sheaf on irreducible variety X, the global
sections of F ⊗
O
X
K
X
form a ﬁnitedimensional vector space over C(X) whose
dimension is the rank of F.
For a locally free sheaf, the rank is just the rank of the associated vector bundle.
Other properties of the rank will be studied in Exercise 8.0.1.
In the smooth case, reﬂexive sheaves of rank 1 are easy to understand.
Proposition 8.0.3. On a smooth variety, a coherent sheaf of rank 1 is reﬂexive if
and only if it is a line bundle.
This is proved in [79, Prop. 1.9]. We now have all of the tools needed to
characterize which coherent sheaves on a normal variety come from Weil divisors.
Theorem 8.0.4. Let L be a coherent sheaf on a normal variety X. Then the
following are equivalent.
(a) L is reﬂexive of rank 1.
(b) There is an open subset j : U
0
֒→X such that codim(X ¸U
0
) ≥ 2, L[
U
0
is a
line bundle on U
0
, and L ≃ j
∗
(L[
U
0
).
(c) L ≃O
X
(D) for some Weil divisor D on X.
Proof. (a) ⇒(b) Since X is normal, its singular locus Y = Sing(X) has codimen
sion at least two in X by Proposition 4.0.17. Then U
0
= X ¸Y is smooth, which
implies that L[
U
0
is a line bundle by Proposition 8.0.3. Hence L ≃ j
∗
(L[
U
0
) by
Proposition 8.0.1.
§8.0. Background: Reﬂexive Sheaves and Differential Forms 341
(b) ⇒ (c) The line bundle L[
U
0
can be written as O
U
0
(E) for some Cartier
divisor E =
i
a
i
E
i
on U
0
. Consider the Weil divisor D=
i
a
i
D
i
, where D
i
is the
Zariski closure of E
i
in X. Given f ∈ C(X)
∗
, note that
div( f ) +D ≥0 ⇐⇒ (div( f ) +D)[
U
0
≥0
since codim(X ¸U
0
) ≥ 2, and the same holds over any open set of X. Combining
this with E =D[
U
0
, we obtain
O
X
(D) ≃ j
∗
O
U
0
(E) = j
∗
(L[
U
0
) ≃L.
(c) ⇒ (a) The proof of (b) ⇒ (c) shows that O
X
(D) ≃ j
∗
(O
X
(D)[
U
0
). But
codim(X ¸U
0
) ≥ 2, and O
X
(D)[
U
0
=O
U
0
(D[
U
0
) is locally free since U
0
is smooth.
Thus j
∗
(O
X
(D)[
U
0
) ≃ O
X
(D)
∨∨
by Proposition 8.0.1, so O
X
(D) ≃ O
X
(D)
∨∨
is
reﬂexive and has rank 1 since it is a line bundle on U
0
.
Tensor Products and Duals. Given Weil divisors D, E on a normal variety X, the
map f ⊗g → f g deﬁnes a sheaf homomorphism
(8.0.1) O
X
(D) ⊗
O
X
O
X
(E) −→O
X
(D+E).
This is an isomorphism when D or E is Cartier but may fail to be an isomorphism
in general.
Example 8.0.5. Consider the afﬁne quadric cone X = V(y
2
−xz) ⊆ C
3
. From
examples in previous chapters, we know that this is a normal toric surface. The
line L = V(y, z) gives a Weil divisor that is not Cartier, though 2L is Cartier (this
follows from Example 4.2.3). The coordinate ring of X is R = C[x, y, z]¸y
2
−xz).
Let x, y, z denote the images of the variables in R. In Exercise 8.0.2 you will show
the following:
• Γ(X, O
X
(−L)) is the ideal ¸y, z) ⊆R.
• Γ(X, O
X
(−2L)) is the ideal ¸z) ⊆R (principal since −2L is Cartier).
• On global sections, the image of the map
O
X
(−L) ⊗
O
X
O
X
(−L) −→O
X
(−2L)
is ¸y, z)
2
, which is a proper subset of Γ(X, O
X
(−2L)) =¸z).
It follows that O
X
(−L) ⊗
O
X
O
X
(−L) ,≃O
X
(−2L). ♦
If we apply (8.0.1) when E =−D, we get a map
O
X
(D) ⊗
O
X
O
X
(−D) −→O
X
,
which in turn induces a map
(8.0.2) O
X
(−D) −→O
X
(D)
∨
.
As noted in §6.0, this is an isomorphism when D is Cartier. In general, we have the
following result about the maps (8.0.1) and (8.0.2).
342 Chapter 8. The Canonical Divisor of a Toric Variety
Proposition 8.0.6. Let D, E be Weil divisors on a normal variety X. Then (8.0.1)
induces an isomorphism
(O
X
(D) ⊗
O
X
O
X
(E))
∨∨
≃O
X
(D+E).
Furthermore, (8.0.2) already is an isomorphism, i.e.,
O
X
(−D) ≃O
X
(D)
∨
.
Proof. The ﬁrst isomorphism follows from Proposition 8.0.1 since O
X
(D+E) is
reﬂexive and (8.0.1) is an isomorphism on the smooth locus of X. The second
isomorphism follows similarly since both sheaves are reﬂexive and (8.0.2) is an
isomorphism on the smooth locus.
Divisor Classes. Recall that Weil divisors D and E on X are linearly equivalent,
written D ∼E, if D=E +div( f ) for some f ∈ C(X)
∗
.
Proposition 8.0.7. Let X be a normal variety.
(a) If D and E are Weil divisors on X, then
O
X
(D) ≃O
X
(E) ⇐⇒D ∼E.
(b) If D is a Weil divisor on X, then
D is Cartier ⇐⇒O
X
(D) is a line bundle.
Proof. Part (a) for Cartier divisors was proved in Proposition 6.0.22. The only
place that we used Cartier in that proof was to show
O
X
(D) ≃O
X
(E) =⇒O
X
(D−E) ≃O
X
.
When D, E are Weil divisors, we prove this as follows: O
X
(D) ≃O
X
(E) implies
O
X
(D) ⊗
O
X
O
X
(−E) ≃O
X
(E) ⊗
O
X
O
X
(−E).
Taking the double dual and using Proposition 8.0.6, we get O
X
(D−E) ≃O
X
.
One direction of part (b) was proved in Chapter 6 (see Proposition 6.0.17 and
Theorem 6.0.18). Conversely, if O
X
(D) is a line bundle on X, then Theorem 6.0.20
shows that O
X
(D) ≃ O
X
(E) for some Cartier divisor E. Thus D ∼ E by part (a).
Then we are done since any divisor linearly equivalent to a Cartier divisor is Cartier
by Exercise 4.0.5.
In Chapter 4 we deﬁned the class group Cl(X) and Picard group Pic(X) in
terms of Weil and Cartier divisors. Then, in Chapter 6, we reinterpreted Pic(X)
in terms of isomorphism classes of line bundles, where the group operation was
tensor product and the inverse was the dual. We can now reinterpret Cl(X) in terms
of isomorphism classes of reﬂexive sheaves of rank 1, where the group operation
is the double dual of the tensor product and the inverse is the dual. This follows
immediately from Propositions 8.0.6 and 8.0.7.
§8.0. Background: Reﬂexive Sheaves and Differential Forms 343
K¨ ahler Differentials. In order to give an algebraic deﬁnition of differential form
on a variety, we begin with the case of a Calgebra.
Deﬁnition 8.0.8. Let R be a Calgebra. The module of K¨ ahler differentials of R
over C, denoted Ω
R/C
, is the Rmodule generated by the formal symbols df for
f ∈ R, modulo the relations
(a) d(c f +g) =cdf +dg for all c ∈ C, f , g ∈ R.
(b) d( f g) = f dg+gdf for all f , g ∈ R.
Example 8.0.9. If R =C[x
1
, . . . , x
n
], then
Ω
R/C
≃
n
i=1
Rdx
i
.
This follows because the relations deﬁning Ω
R/C
imply df =
n
i=1
∂ f
∂x
i
dx
i
for all
f ∈ R (Exercise 8.0.3). ♦
A Calgebra homomorphism R →S induces a natural homomorphism
Ω
R/C
−→Ω
S/C
.
When we regard Ω
S/C
as an Rmodule, we obtain a homomorphism of Smodules
S⊗
R
Ω
R/C
−→Ω
S/C
.
Here is a case when this map is easy to understand. See [120, Thm. 25.2] for a
proof.
Proposition 8.0.10. Let R →S be a surjection of Calgebras with kernel I. Then
there is an exact sequence of Smodules
I/I
2
−→S⊗
R
Ω
R/C
−→Ω
S/C
−→0,
where [ f ] ∈ I/I
2
maps to 1⊗df ∈ S⊗
R
Ω
R/C
.
Example 8.0.11. Let R = C[x
1
, . . . , x
n
] and S = R/I, where I = ¸ f
1
, . . . , f
s
). The
generators of I give a surjection R
s
→I and hence a surjection S
s
→I/I
2
. Combin
ing this with Proposition 8.0.10 and Example 8.0.9, we obtain an exact sequence
S
s
α
−→S
n
−→Ω
S/C
−→0,
where α is given by the reduction of the ns Jacobian matrix
(8.0.3)
_
_
_
∂ f
1
∂x
1
∂ fs
∂x
1
.
.
.
.
.
.
∂ f
1
∂xn
∂ fs
∂xn
_
_
_
modulo the ideal I (Exercise 8.0.4). This presentation of Ω
S/C
is very useful for
computing examples. ♦
344 Chapter 8. The Canonical Divisor of a Toric Variety
K¨ ahler differentials also behave well under localization, as you will prove in
Exercise 8.0.5.
Proposition 8.0.12. Let R
f
be the localization of R at a nonnilpotent element
f ∈ R. Then Ω
R
f
/C
≃Ω
R/C
⊗R
f
.
Cotangent and Tangent Sheaves. Now we globalize Deﬁnition 8.0.8.
Deﬁnition 8.0.13. Let X be a variety. The cotangent sheaf Ω
1
X
is the sheaf of
O
X
modules deﬁned via
Ω
1
X
(U) =Ω
O
X
(U)/C
on afﬁne open sets U. The tangent sheaf T
X
is the dual sheaf
T
X
=(Ω
1
X
)
∨
=Hom
O
X
(Ω
1
X
, O
X
).
The reason for the superscript in the notation for the cotangent sheaf will be
come clear later in this chapter. In Exercise 8.0.6 you will use Example 8.0.11 and
Proposition 8.0.12 to show that Ω
1
X
is a coherent sheaf. See [78, II.8] for a slightly
different approach to deﬁning the sheaf Ω
1
X
, and [78, II.8, Comment 8.9.2] for the
connection between these methods.
When U = Spec(R) is an afﬁne open of X, the deﬁnition of the tangent sheaf
implies that
T
X
(U) =Hom
R
(Ω
R/C
, R).
This can also be described in terms of derivations—see Exercises 8.0.7–8.0.8.
When X is smooth, these sheaves are nicely behaved, as shown by the follow
ing result from [78, Thm. II.8.15].
Theorem 8.0.14. A variety X is smooth if and only if Ω
1
X
is locally free. When this
happens, Ω
1
X
and T
X
are locally free sheaves of rank n, n =dimX.
In the smooth toric case, it is easy to see that the cotangent sheaf is locally free.
Example 8.0.15. A smooth cone σ ⊆N
R
≃R
n
of dimension r gives the afﬁne toric
variety
U
σ
≃C
r
(C
∗
)
n−r
⊆C
n
.
Then Example 8.0.9 and Proposition 8.0.12 imply that Ω
1
Uσ
is locally free of rank n.
It follows immediately that Ω
1
X
Σ
is locally free for any smooth toric variety X
Σ
. ♦
We know from Chapter 6 that a locally free sheaf is the sheaf of sections of a
vector bundle. When X is smooth, the vector bundles corresponding to Ω
1
X
and T
X
are called the cotangent bundle and tangent bundle respectively.
§8.0. Background: Reﬂexive Sheaves and Differential Forms 345
Example 8.0.16. We construct the cotangent bundle for P
2
. Recall that P
2
has a a
cover by the afﬁne open sets
U
σ
0
≃Spec(C[x, y])
U
σ
1
≃Spec(C[yx
−1
, x
−1
])
U
σ
2
≃Spec(C[xy
−1
, y
−1
]).
where σ
0
, σ
1
, σ
2
are the maximal cones in the usual fan for P
2
.
Let C
2
=Spec(R) for R =C[x, y]. The module Ω
R/C
is a free Rmodule of rank
2 with generators dx, dy by Example 8.0.9. Thus a 1form on C
2
may be written
uniquely as f
1
dx + f
2
dy, where f
i
∈ R. To generalize this to P
2
, we require that
after changing coordinates, dx and dy transform via the Jacobian matrix described
in Example 8.0.11. More precisely, the matrix for the transition function φ
i j
will
be the Jacobian of the map U
σ
j
→U
σ
i
.
On U
σ
2
, the coordinates (a
1
, a
2
) are represented in terms of the (x, y) coordi
nates on U
σ
0
as (xy
−1
, y
−1
), yielding
φ
20
=
_
1/y −x/y
2
0 −1/y
2
_
.
Next, we compute φ
12
. Things get messy if we keep everything in (x, y) coordi
nates, so we ﬁrst translate to coordinates (a
1
, a
2
) on U
σ
2
, and then translate back.
On U
σ
2
we identify (a
1
, a
2
) with (xy
−1
, y
−1
). Then U
σ
1
has coordinates
(yx
−1
, x
−1
) =
_
1
a
1
,
a
2
a
1
_
.
So in terms of (a
1
, a
2
), we have
φ
12
=
_
−1/a
2
1
0
−a
2
/a
2
1
1/a
1
_
.
Rewriting this in terms of (x, y) yields
φ
12
=
_
−y
2
/x
2
0
−y/x
2
y/x
_
.
Finally, computing φ
10
directly, we obtain
φ
10
=
_
−y/x
2
1/x
−1/x
2
0
_
.
A check shows that φ
10
= φ
12
◦φ
20
. Similar computations show that the compati
bility criteria are satisﬁed for all i, j, k, i.e.,
φ
ik
=φ
i j
◦φ
jk
.
Since det(φ
i j
) is invertible on U
σ
i
∩U
σ
j
, the same is true for φ
i j
. Hence we obtain
a rank 2 vector bundle on P
2
whose sheaf of sections is Ω
1
P
2
. ♦
346 Chapter 8. The Canonical Divisor of a Toric Variety
Relation with the Zariski Tangent Space. The deﬁnition of the tangent sheaf
T
X
seems far removed from the deﬁnition of the Zariski tangent space T
p
(X) =
Hom
C
(m
X,p
/m
2
X,p
, C) given in Chapter 1. Here we explain (without proof) the
connection.
The stalk (T
X
)
p
of the tangent sheaf at p ∈ X can be described as follows. The
stalk of Ω
1
X
at p is the module of K¨ ahler differentials
(Ω
1
X
)
p
=Ω
O
X,p
/C
,
where O
X,p
is the local ring of X at p. Since O
X,p
/m
X,p
≃ C and Ω
C/C
= 0 (easy
to check), the exact sequence of Proposition 8.0.10 gives a surjection
m
X,p
/m
2
X,p
−→Ω
O
X,p
/C
⊗
O
X,p
C
which is an isomorphism of vector spaces over C by [78, Prop. II.8.7]. Since T
X
is dual to Ω
1
X
, taking the dual of the above isomorphism gives
(8.0.4) (T
X
)
p
⊗
O
X,p
C ≃Hom
C
(m
X,p
/m
2
X,p
, C) =T
p
(X).
This omits many details but should help you understand why T
X
is the correct
deﬁnition of tangent sheaf.
Example 8.0.17. Let V ⊆C
n
be deﬁned by I =I(V) =¸ f
1
, . . . , f
s
) ⊆C[x
1
, . . . , x
n
].
The coordinate ring of V is S = C[x
1
, . . . , x
n
]/I, so that Example 8.0.11 gives the
exact sequence
S
s
−→S
n
−→Ω
S/C
−→0.
Now take p ∈V and tensor with O
V,p
to obtain the exact sequence
O
s
V,p
−→O
n
V,p
−→Ω
O
V,p
/C
−→0
(Exercise 8.0.9). If we tensor this with C and dualize, (8.0.4) and the isomorphism
Ω
O
X,p
/C
⊗
O
X,p
C ≃m
X,p
/m
2
X,p
give the exact sequence
0 −→T
p
(V) −→C
n
δ
−→C
s
,
where δ comes from the s n Jacobian matrix
_
∂ f
i
∂x
j
(p)
_
(Exercise 8.0.9). This
explains the description of T
p
(V) given in Lemma 1.0.6. ♦
Conormal and Normal Sheaves. Given a closed subvariety i : Y ֒→X, it is natural
to ask how their cotangent sheaves relate. We begin with the exact sequence
0 −→I
Y
−→O
X
−→i
∗
O
Y
−→0,
which we write more informally as
0 −→I
Y
−→O
X
−→O
Y
−→0.
The quotient sheaf I
Y
/I
2
Y
has a natural structure as sheaf of O
Y
modules, as does
F ⊗
O
X
O
Y
for any sheaf F of O
X
modules. The following basic result is proved
in [78, Prop. II.8.12 and Thm. II.8.17].
§8.0. Background: Reﬂexive Sheaves and Differential Forms 347
Theorem 8.0.18. Let Y be a closed subvariety of a variety X. Then:
(a) There is an exact sequence of O
Y
modules
I
Y
/I
2
Y
−→Ω
1
X
⊗
O
X
O
Y
−→Ω
1
Y
−→0.
(b) If X and Y are smooth, then this sequence is also exact on the left and I
Y
/I
2
Y
is locally free of rank equal to the codimension of Y.
Note that part (a) of this theorem is a global version of Proposition 8.0.10. We
call I
Y
/I
2
Y
the conormal sheaf of Y in X and call its dual
N
Y/X
=(I
Y
/I
2
Y
)
∨
=Hom
O
Y
(I
Y
/I
2
Y
, O
Y
)
the normal sheaf of Y in X. When X andY are smooth, we can dualize the sequence
appearing in Theorem 8.0.18 to obtain the exact sequence
0 −→T
Y
−→T
X
⊗
O
X
O
Y
−→N
Y/X
−→0.
The vector bundle associated to N
Y/X
is the normal bundle of Y in X. Then the
above sequence says that when the tangent bundle of X is restricted to the subvari
ety Y, it contains the tangent bundle of Y with quotient given by the normal bundle.
This is the algebraic analog of what happens in differential geometry, where the
normal bundle is the orthogonal complement of the tangent bundle of Y.
Differential Forms. We call Ω
1
X
the sheaf of 1forms, and we deﬁne the sheaf of
pforms to be the wedge product
Ω
p
X
=
_
p
Ω
1
X
.
For any sheaf F of O
X
modules, the exterior power
_
p
F is the sheaf associated
to the presheaf which to each open set U assigns the O
X
(U)module
_
p
F(U).
Example 8.0.19. For C
n
, Ω
1
C
n is the sheaf associated to the free Rmodule Ω
R/C
=
n
i=1
Rdx
i
, R =C[x
1
, . . . , x
n
]. Then Ω
p
C
n is the sheaf associated to
_
p
Ω
R/C
=
1≤i
1
<<ip≤n
Rdx
i
1
∧ ∧dx
in
.
It follows that Ω
p
C
n is free of rank
_
n
p
_
. ♦
More generally, Theorem 8.0.14 implies that Ω
p
X
is locally free of rank
_
n
p
_
when X is smooth of dimension n. In particular, Ω
n
X
is a line bundle in this case.
Zariski pForms and the Canonical Sheaf . For a normal variety X, the sheaf of
pforms Ω
p
X
may fail to be locally free. However, this sheaf is locally free on
the smooth locus of X, and the complement of the smooth locus has codimension
≥ 2 since X is normal. Hence we can use Proposition 8.0.1 to deﬁne the sheaf!of
Zariski pforms
(8.0.5)
´
Ω
p
X
=(Ω
p
X
)
∨∨
= j
∗
Ω
p
U
0
,
348 Chapter 8. The Canonical Divisor of a Toric Variety
where j : U
0
֒→X is the inclusion of the smooth locus of X. It follows that
´
Ω
p
X
is a
reﬂexive sheaf of rank
_
n
p
_
, where n =dimX.
For later purposes, we note that by Proposition 8.0.1, (8.0.5) is valid for any
smooth open subset U
0
⊆X whose complement has codimension ≥2.
The case p =n is especially important.
Deﬁnition 8.0.20. The canonical sheaf of a normal variety X is
ω
X
=
´
Ω
n
X
,
where n is the dimension of X. This is a reﬂexive sheaf of rank 1, so that
ω
X
≃O
X
(D)
for some Weil divisor D on X. We call this divisor a canonical divisor of X, often
denoted K
X
.
Proposition 8.0.7 shows that the canonical divisor K
X
is welldeﬁned up to
linear equivalence and hence gives a unique divisor class in Cl(X), known as the
canonical class of X. In the toric case, we will see later in the chapter that there is
a natural choice for the canonical divisor.
When X is smooth, we call ω
X
the canonical bundle since it is a line bundle. In
this case, the canonical divisor is Cartier. There are also singular varieties whose
canonical divisors are Cartier—these are the Gorenstein varieties to be studied later
in the chapter.
While it often sufﬁces to know ω
X
up to isomorphism, there are situations
where a unique model of ω
X
is required. One such construction uses
_
n
Ω
C(X)/C
,
where C(X) is the ﬁeld of rational functions on X. We can regard
_
n
Ω
C(X)/C
as
the constant sheaf of rational nforms on X, similar to way that C(X) gives the
constant sheaf K
X
of rational functions on X. There is an obvious sheaf map
Ω
n
X
−→
_
n
Ω
C(X)/C
.
You will prove the following result in Exercise 8.0.10.
Proposition 8.0.21. When X is normal, image of the map Ω
n
X
→
_
n
Ω
C(X)/C
is
ω
X
⊆
_
n
Ω
C(X)/C
.
The canonical sheaf can be deﬁned for any irreducible variety X as a subsheaf
of
_
n
Ω
C(X)/C
, though the deﬁnition is more sophisticated (see [112, §9]). When
X is projective, another approach is given in [78, III.7], where ω
X
is called the
dualizing sheaf. We will see in Chapter 9 that ω
X
plays a key role in Serre duality.
Exercises for §8.0.
8.0.1. The rank of a coherent sheaf on an irreducible variety was deﬁned in Deﬁnition 8.0.2.
Here are some properties of the rank.
§8.0. Background: Reﬂexive Sheaves and Differential Forms 349
(a) Let R be an integral domain with ﬁeld of fractions K, and let M be a ﬁnitely gener
ated Rmodule. Show that M⊗
R
K is a ﬁnitedimensional vector space over K whose
dimension equals the rank of the coherent sheaf
¯
M on Spec(R).
(b) Let F be a coherent sheaf on X let U ⊆X be an nonempty open subset. Prove that F
and F[
U
have the same rank.
(c) Let 0 →F →G →H →be an exact sequence of sheaves on X. Prove that rank(G) =
rank(F) +rank(H ).
8.0.2. Prove the claims made in Example 8.0.5.
8.0.3. Let R =C[x
1
, . . . , x
n
]. In Example 8.0.9 we claimed that df =
n
i=1
∂ f
∂xi
dx
i
in Ω
R/C
for all f ∈ R. Prove this.
8.0.4. Prove that the map α in the exact sequence from Example 8.0.11 comes from the
Jacobian matrix (8.0.3).
8.0.5. Prove Proposition 8.0.12.
8.0.6. Prove that the cotangent sheaf Ω
1
X
deﬁned in Deﬁnition 8.0.13 is a coherent sheaf.
8.0.7. Given a Calgebra R and an Rmodule M, a Cderivation δ : R →M is a Clinear
map that satisﬁes the Leibniz rule, i.e., δ( f g) = f δ(g) +gδ( f ) for all f , g ∈ R.
(a) Show that f →df deﬁnes a Cderivation d : R →Ω
R/C
.
(b) More generally, show that if φ : Ω
R/C
→ M is an Rmodule homomorphism, then
φ◦ d : R →M is a Cderivation.
8.0.8. Continuing Exercise 8.0.7, we let Der
C
(R, M) denote the set of all Cderivations
δ : R →M. This is an Rmodule where (rδ)( f ) =rδ( f ).
(a) Use part (b) of Exercise 8.0.7 to construct an Rmodule isomorphism Der
C
(R, M) ≃
Hom
R
(Ω
R/C
, M). Explain why d : R →Ω
R/C
is called the universal derivation.
(b) Let T
X
be the tangent sheaf of a variety X and let U = Spec(R) be an afﬁne open
subset of X. Prove that T
X
(U) =Der
C
(R, R).
8.0.9. Fill in the details omitted in Example 8.0.17.
8.0.10. Prove Proposition 8.0.21. Hint: Ω
1
X
is locally free when restricted to the smooth
locus of X. Exercise 8.0.11 below will be useful.
8.0.11. Let j : U ֒→X be the inclusion of a nonempty open subset of a variety X.
(a) Show that there is a sheaf map F → j
∗
(F[
U
) for any sheaf F on X.
(b) Show that the map of part (a) is an isomorphism when X is irreducible and F is a
constant sheaf.
8.0.12. Let 1 ∈ T
N
be the identity element of the torus T
N
and let m ⊆ C[M] be the corre
sponding maximal ideal. Set N
C
=N⊗
Z
C and M
C
=M⊗
Z
C.
(a) Let f =
m
c
m
χ
m
∈ C[M]. Show that f ∈ m
2
if and only if
m
c
m
= 0 in C and
m
c
m
m = 0 in M
C
. Hint: Pick a basis e
1
, . . . , e
n
of M and set t
i
= χ
ei
, so that f is
a Laurent monomial in t
1
, . . . , t
n
. Then show that f ∈ m
2
if and only if f ∈ m and
∂ f
∂ti
(1) =0 for all i.
(b) Use part (a) to construct an isomorphism m/m
2
≃ M
C
, and conclude that the Zariski
tangent space of T
N
at the identity is naturally isomorphic to N
C
.
350 Chapter 8. The Canonical Divisor of a Toric Variety
8.0.13. Let F be a free module of rank n over a ring R and ﬁx 0 ≤ p ≤n. Recall that wedge
product induces an isomorphism
_
n−p
F ≃Hom
R
(
_
p
F,
_
n
F).
(a) If X is smooth of dimension n, then show that Ω
n−p
X
≃Hom
OX
(Ω
p
X
, Ω
n
X
).
(b) Show that if X is a normal variety of dimension n, then there is an isomorphism
´
Ω
n−p
X
≃Hom
OX
(
´
Ω
p
X
, ω
X
). Hint: If you get stuck, see [42, Prop. 4.7].
(c) In a similar vein, show that the tangent sheaf T
X
of a normal variety X satisﬁes T
X
≃
Hom
OX
(
´
Ω
1
X
, O
X
).
§8.1. OneForms on Toric Varieties
In this section we will describe two interesting exact sequences that involve the
sheaves Ω
1
X
Σ
and
´
Ω
1
X
Σ
on a normal toric variety X
Σ
.
The Torus. The coordinate ring of the torus T
N
is the semigroup algebra C[M].
Then the map
Ω
C[M]/C
−→M⊗
Z
C[M]
deﬁned by dχ
m
→m⊗χ
m
is easily seen to be an isomorphism. It follows that
(8.1.1) Ω
1
T
N
≃M⊗
Z
O
T
N
,
and dualizing, we obtain
T
T
N
≃N⊗
Z
O
T
N
=N
C
⊗
C
O
T
N
.
This makes intuitive sense since T
N
= N ⊗
Z
C
∗
as a complex Lie group. Thus its
tangent space at the identity is N⊗
Z
C =N
C
via the exponential map. This is also
true algebraically, as shown in Exercise 8.0.12. The group action transports the
tangent space N
C
over the whole torus, which explains the above trivialization of
the tangent bundle T
T
N
.
As a consequence, the 1form
dχ
m
χ
m
is a global section of Ω
1
T
N
that maps to m⊗1
in (8.1.1) and hence is invariant under the action of T
N
. See Exercise 8.1.1 for more
on invariant 1forms on the torus.
The First Exact Sequence. Now consider the toric variety X
Σ
of the fan Σ. For
ρ ∈ Σ(1), the inclusion i : D
ρ
֒→X
Σ
gives the sheaf i
∗
O
Dρ
on X
Σ
, which following
§8.0 we write as O
Dρ
. Using the map M →Z given by m→¸m, u
ρ
), we obtain the
composition
M⊗
Z
O
X
Σ
−→Z⊗
Z
O
X
Σ
=O
X
Σ
−→O
Dρ
.
This gives a natural map
(8.1.2) β : M⊗
Z
O
X
Σ
−→
ρ
O
Dρ
,
where the direct sum is over all ρ ∈ Σ(1). We also have a canonical map
(8.1.3) α : Ω
1
X
Σ
−→M⊗
Z
O
X
Σ
§8.1. OneForms on Toric Varieties 351
constructed as follows. On the afﬁne piece U
σ
=Spec(C[σ
∨
∩M]), α is deﬁned by
dχ
m
∈ Ω
C[σ
∨
∩M]/C
−→m⊗χ
m
∈ M⊗
Z
C[σ
∨
∩M].
These C[σ
∨
∩M]module homomorphisms Ω
C[σ
∨
∩M]/C
→M⊗
Z
C[σ
∨
∩M] patch
to give the desired map (8.1.3) (Exercise 8.1.2). Note that over the torus T
N
, the
map α of (8.1.3) reduces to the isomorphism (8.1.1).
Theorem 8.1.1. For a smooth toric variety X
Σ
, the sequence
0 −→Ω
1
X
Σ
α
−→M⊗
Z
O
X
Σ
β
−→
ρ
O
Dρ
−→0
formed using (8.1.2) and (8.1.3) is exact.
Proof. We ﬁrst verify that β ◦α is the zero map. On the afﬁne piece U
σ
⊆X
Σ
, the
subvariety D
ρ
∩U
σ
⊆U
σ
is deﬁned by the ideal I
ρ
=I(D
ρ
∩U
σ
) ⊆C[σ
∨
∩M]. By
Propositions 4.0.28 and 4.3.2, we have
(8.1.4) I
ρ
=Γ(U
σ
, O
X
Σ
(−D
ρ
)) =
div(χ
m
)[
Uσ
≥Dρ[
Uσ
C χ
m
=
m∈σ
∨
∩M,¸m,uρ)>0
C χ
m
.
Over U
σ
, the composition Ω
1
X
Σ
→M⊗
Z
O
X
Σ
→O
Dρ
takes a 1form dχ
m
, m ∈
σ
∨
∩M, to ¸m, u
ρ
)χ
m
∈ C[σ
∨
∩M]/I
ρ
. This is obviously zero if ¸m, u
ρ
) = 0, and
if ¸m, u
ρ
) , =0, it vanishes since χ
m
∈ I
ρ
in this case.
We now verify that the sequence is exact over U
σ
. Since σ is smooth, we
may assume σ = Cone(e
1
, . . . , e
r
), where r ≤ n and e
1
, . . . , e
n
is a basis of N.
Then U
σ
= C
r
(C
∗
)
n−r
. Let x
1
, . . . , x
n
denote the characters of the correspond
ing dual basis of M, also denoted e
1
, . . . , e
n
. The coordinate ring of U
σ
is R =
C[x
1
, . . . , x
r
, x
±1
r+1
, . . . , x
±1
n
], and the 1forms on U
σ
form the free Rmodule Ω
R/C
=
n
i=1
Rdx
i
by Example 8.0.9 and Proposition 8.0.12. Since α takes dx
i
to e
i
⊗x
i
,
we see that α can be regarded as the map
Ω
R/C
=
n
i=1
Rdx
i
−→M⊗
R
R =
n
i=1
R
that sends
n
i=1
f
i
dx
i
to ( f
1
x
1
, . . . , f
n
x
n
). This gives the exact sequence
0 −→Ω
R/C
−→
n
i=1
R −→
r
i=1
R/¸x
i
) −→0
since x
r+1
, . . . , x
n
are units in R, and the theorem follows.
Logarithmic Forms. The exact sequence of Theorem 8.1.1 has a lovely interpreta
tion in terms of residues of logarithmic 1forms. The idea is that M⊗
Z
O
X
Σ
can be
thought of as the sheaf Ω
1
X
Σ
(logD) of 1forms on X
Σ
with logarithmic poles along
D=
ρ
D
ρ
. We begin with an example.
352 Chapter 8. The Canonical Divisor of a Toric Variety
Example 8.1.2. The coordinate ring of C
n
is R =C[x
1
, . . . , x
n
], and the divisor D is
the sum of the coordinate hyperplanes D
i
= V(x
i
). As above, Ω
R/C
=
n
i=1
Rdx
i
.
Now introduce some denominators: a rational 1form ω has logarithmic poles
along D if
ω =
n
i=1
f
i
dx
i
x
i
, f
i
∈ R.
These form the free Rmodule
n
i=1
R
dx
i
x
i
, and the corresponding sheaf is deﬁned
to be Ω
1
C
n (logD). The formal calculation
dχ
m
χ
m
=
n
i=1
¸m, e
i
)
dx
i
x
i
shows that the map
dχ
m
χ
m
→m⊗1 induces an isomorphism of sheaves
Ω
1
C
n (logD) ≃M⊗
C
n
O
C
n
such that the map α : Ω
1
C
n →M⊗
C
n
O
C
n
of (8.1.3) is induced by the inclusion of
1forms Ω
1
C
n ֒→Ω
1
C
n (logD). ♦
This construction works for any smooth afﬁne toric variety U
σ
, and the sheaves
of logarithmic 1forms on U
σ
patch to give the sheaf Ω
1
X
Σ
(logD) for any smooth
toric variety X
Σ
. Furthermore, we have a canonical isomorphism
(8.1.5) Ω
1
X
Σ
(logD) ≃M⊗
Z
O
X
Σ
such that the map α of (8.1.3) comes from the inclusion of 1forms.
The construction of Ω
1
X
Σ
(logD) can be done more generally. Let X be a smooth
variety. A divisor D =
i
D
i
on X has smooth normal crossings if every D
i
is
smooth and irreducible, and for every p ∈ X, the divisors containing p meet nicely.
More precisely, if I
p
= ¦i [ p ∈ D
i
¦, we require that the tangent spaces T
p
(D
i
) ⊆
T
p
(X) meet transversely, i.e.,
codim
_
i∈Ip
T
p
(D
i
)
_
=[I
p
[.
For example, the divisor D =
ρ
D
ρ
is a smooth normal crossing divisor on any
smooth toric variety X
Σ
. A nice discussion of Ω
1
X
(logD) for complex manifolds
can be found in [72, p. 449].
The Poincar´ e Residue Map. Let f (z) be an analytic (also called holomorphic)
function deﬁned in a punctured neighborhood of a point p ∈ C. Take a counter
clockwise loop C around p on X such that f (z) has no other poles inside C. Then
the residue of the 1form ω = f (z)dz at p is the contour integral
res
p
(ω) =
1
2πi
_
C
ω.
§8.1. OneForms on Toric Varieties 353
In particular, if p = 0 and f (z) =
g(z)
z
, with g(z) analytic at zero, then the residue
theorem tells us that res
p
(ω) is the coefﬁcient of
1
z
in the Laurent series for f (z) at
0, and is equal to g(0). Note that the 1form ω = f (z)dz =g(z)
dz
z
has a logarithmic
pole at p =0.
When there are several variables, we can do the same construction by working
one variable at a time. Here is an example.
Example 8.1.3. Given f = f (x
1
, . . . , x
n
) ∈ R = C[x
1
, . . . , x
n
], we get the logarith
mic 1form ω = f
dx
1
x
1
. In terms of the above discussion of residues, we can regard
f (0, x
2
, . . . , x
n
) as the “residue” of ω at V(x
1
). Note also that f (0, x
2
, . . . , x
n
) repre
sents the class of f in R/¸x
1
). Doing this for every variable shows that the map
Ω
1
C
n (logD) ≃M⊗
Z
O
C
n
β
−→
n
i=1
O
D
i
can be interpreted as a sum of “residue” maps. ♦
More generally, if X is a smooth variety and D=
i
D
i
a smooth normal cross
ing divisor, one can deﬁne the Poincar´ e residue map
P
r
: Ω
1
X
(logD) −→
i
O
D
i
(see [144, p. 254]) such that we have an exact sequence
(8.1.6) 0 −→Ω
1
X
−→Ω
1
X
(logD)
Pr
−→
i
O
D
i
−→0.
When applied to a smooth toric variety X
Σ
and the divisor D =
ρ
D
ρ
, this gives
the exact sequence of Theorem 8.1.1 via the isomorphism (8.1.5).
The Normal Case. When X
Σ
is normal, we get an analog of Theorem 8.1.1 that
uses the sheaf
´
Ω
1
X
Σ
of Zariski 1forms in place of Ω
1
X
Σ
. Since M⊗
Z
O
X
Σ
is reﬂexive,
taking the double dual of (8.1.3) gives a map
´
Ω
1
X
Σ
−→M⊗
Z
O
X
Σ
.
Theorem 8.1.4. Let X
Σ
be a normal toric variety. Then:
(a) The sequence
0 −→
´
Ω
1
X
Σ
−→M⊗
Z
O
X
Σ
−→
ρ
O
Dρ
is exact.
(b) If X
Σ
is simplicial, then the map on the right is surjective, so that
0 −→
´
Ω
1
X
Σ
−→M⊗
Z
O
X
Σ
−→
ρ
O
Dρ
−→0
is exact.
354 Chapter 8. The Canonical Divisor of a Toric Variety
Proof. Let j : U
0
⊆ X
Σ
be the inclusion map for U
0
=
ρ
U
ρ
. Note that U
0
is
a smooth toric variety whose fan has the same 1dimensional cones as Σ, and
codim(X ¸U
0
) ≥ 2 by the OrbitCone Correspondence. By Theorem 8.1.1, we
have an exact sequence
0 −→Ω
1
U
0
−→M⊗
Z
O
U
0
−→
ρ
O
Dρ∩U
0
−→0,
so that applying j
∗
gives the exact sequence
0 −→ j
∗
Ω
1
U
0
−→ j
∗
(M⊗
Z
O
U
0
) −→
ρ
j
∗
O
Dρ∩U
0
since j
∗
is left exact (Exercise 8.1.3). However, j
∗
Ω
1
U
0
=
´
Ω
1
X
Σ
by the remarks
following (8.0.5), and j
∗
(M⊗
Z
O
U
0
) =M⊗
Z
O
X
Σ
by Proposition 8.0.1. Hence we
get an exact sequence
0 −→
´
Ω
1
X
Σ
−→M⊗
Z
O
X
Σ
−→
ρ
j
∗
O
Dρ∩U
0
.
In Exercise 8.1.4 you will show that the maps M⊗
Z
O
X
Σ
→ j
∗
O
Dρ∩U
0
factor
M⊗
Z
O
X
Σ
−→O
Dρ
−→ j
∗
O
Dρ∩U
0
,
where O
Dρ
→ j
∗
O
Dρ∩U
0
is injective. The exact sequence of part (a) then follows
immediately.
It remains to show that M⊗
Z
O
X
Σ
→
ρ
O
Dρ
is surjective when X
Σ
is simpli
cial. Given σ ∈ Σ, we need to show that
M⊗
Z
O
Uσ
−→
ρ∈σ(1)
O
Dρ∩Uσ
is surjective. Fix ρ ∈ σ(1) and pick m ∈ M such that ¸m, u
ρ
) , = 0 and ¸m, u
ρ
′ ) = 0
for all ρ
′
,= ρ in σ(1). Such an m exists since σ is simplicial. Then m⊗1 maps
to a nonzero constant function on O
Dρ∩Uσ
and to the zero function on O
D
ρ
′ ∩Uσ
for
ρ
′
,=ρ. The desired surjectivity now follows easily.
When X
Σ
has no torus factors, we learned in §5.3 that graded modules over the
total coordinate ring S = C[x
ρ
[ ρ ∈ Σ(1)] give quasicoherent sheaves on X
Σ
. It is
easy to describe a graded Smodule that gives
´
Ω
1
X
Σ
. We begin with
M⊗
Z
S −→Z⊗
Z
S =S −→S/¸x
ρ
),
where the ﬁrst map comes from m →¸m, u
ρ
) and the second map is obvious. This
gives a homomorphism M⊗
Z
S →
ρ
S/¸x
ρ
), and we deﬁne
´
Ω
1
S
to be the kernel
of this map. Hence we have an exact sequence of graded Smodules
(8.1.7) 0 −→
´
Ω
1
S
−→M⊗S −→
ρ
S/¸x
ρ
).
Using Example 6.0.10 and Theorem 8.1.4, we obtain the following result.
Corollary 8.1.5. When X
Σ
has no torus factors,
´
Ω
1
X
Σ
is the sheaf associated to the
graded Smodule
´
Ω
1
S
.
§8.1. OneForms on Toric Varieties 355
The Euler Sequence. In [78, Thm. II.8.13], Hartshorne constructs an exact se
quence
(8.1.8) 0 −→Ω
1
P
n −→O
P
n (−1)
n+1
−→O
P
n −→0,
called the Euler sequence of P
n
. He goes on to say “This is a fundamental result,
upon which we will base all future calculations involving differentials on projective
varieties.” Of course, P
n
is toric, and there is a toric generalization of this result,
due to Jaczewski [99] in the smooth case and Batyrev and Cox [10, Thm. 12.1] in
the simplicial case.
Theorem 8.1.6. Let X
Σ
be a simplicial toric variety with no torus factors, i.e.,
¦u
ρ
[ ρ ∈ Σ(1)¦ spans N
R
. Then there is an exact sequence
0 −→
´
Ω
1
X
Σ
−→
ρ
O
X
Σ
(−D
ρ
) −→Cl(X
Σ
) ⊗
Z
O
X
Σ
−→0.
Furthermore, if X
Σ
is smooth, then the sequence can be written
0 −→Ω
1
X
Σ
−→
ρ
O
X
Σ
(−D
ρ
) −→Pic(X
Σ
) ⊗
Z
O
X
Σ
−→0.
Proof. Consider the following diagram:
0
0
0
0
// ´
Ω
1
X
Σ
//
M⊗
Z
O
X
Σ
//
ρ
O
Dρ
//
0
0
//
ρ
O
X
Σ
(−D
ρ
)
//
ρ
O
X
Σ
//
ρ
O
Dρ
//
0
0
//
Cl(X
Σ
) ⊗
Z
O
X
Σ
//
Cl(X
Σ
) ⊗
Z
O
X
Σ
//
0
//
0
0 0 0
The top row is from Theorem 8.1.4 and is exact since X
Σ
is simplicial. Also, by
Proposition 4.0.28, each ρ ∈ Σ(1) gives an exact sequence
0 −→O
X
Σ
(−D
ρ
) −→O
X
Σ
−→O
Dρ
−→0,
and the middle row is the direct sum of these exact sequences. The third row is the
obvious exact sequence that uses the identity map on Cl(X
Σ
) ⊗
Z
O
X
Σ
.
Since X
Σ
has no torus factors, we have the exact sequence
0 −→M −→
ρ
Z −→Cl(X
Σ
) −→0
from Theorem 4.1.3, and tensoring this with O
X
Σ
gives the middle column. The
column on the right is the another obvious exact sequence, and one can check
without difﬁculty that the solid arrows in the diagram commute (Exercise 8.1.5).
356 Chapter 8. The Canonical Divisor of a Toric Variety
Then commutativity and exactness imply the existence of the dotted arrows in the
diagram, which give the desired exact sequence by a standard diagram chase.
The exact sequence of sheaves in Theorem 8.1.6 is the (generalized) Euler
sequence of the toric variety X
Σ
. We will use it in the next section to determine the
canonical sheaf of X
Σ
. The Euler sequence also encodes relations generalizing the
classical Euler relation for homogeneous polynomials (see Exercise 8.1.8).
Exercises for §8.1.
8.1.1. We will study invariant 1forms and derivations on the torus. Since the torus T
N
=
Spec(C[M]) is afﬁne, we know from Exercise 8.0.8 that the derivations Der
C
(C[M], C[M])
give the global sections of the tangent sheaf T
TN
.
(a) For u ∈ N, deﬁne ∂
u
: C[M] →C[M] by
∂
u
(χ
m
) =¸m, u)χ
m
.
Prove that ∂
u
∈ Der
C
(C[M], C[M])
(b) Let x
1
, . . . , x
n
be the characters corresponding to the elements of M dual to some par
ticular basis e
1
, . . . , e
n
of N. Thus C[M] =C[x
±1
1
, . . . , x
±1
n
]. Prove that ∂
ei
=x
i
∂
∂xi
and
that Γ(T
N
, T
TN
) is the free C[M]module generated by x
1
∂
∂x1
, . . . , x
n
∂
∂xn
.
(c) Dualizing, conclude that Γ(T
N
, Ω
1
TN
) is the free C[M]module generated by the T
N

invariant differentials
dx1
x1
, . . . ,
dxn
xn
.
8.1.2. Consider the afﬁne toric variety U
σ
=Spec(C[σ
∨
∩M]).
(a) Prove that the map
dχ
m
∈ Ω
C[σ
∨
∩M]/C
−→m⊗χ
m
∈ M⊗
Z
C[σ
∨
∩M]
deﬁnes a C[σ
∨
∩M]module homomorphism.
(b) For a toric variety X
Σ
, prove that these homomorphisms patch together to give the map
α : Ω
1
XΣ
−→M⊗
Z
O
XΣ
in (8.1.3).
8.1.3. Let 0 →F →G →H →0 be an exact sequence of sheaves on X and let f : X →Y
be a morphism.
(a) Prove that 0 → f
∗
F → f
∗
G → f
∗
H is exact on Y.
(b) Suppose that Y = ¦pt¦ and f : X →Y is the obvious map. Use part (a) to give a new
proof of Proposition 6.0.8.
8.1.4. In the proof of Theorem 8.1.4, show that the map M⊗
Z
O
XΣ
→ j
∗
O
Dρ∩U0
factors
M⊗
Z
O
XΣ
−→O
Dρ
−→ j
∗
O
Dρ∩U0
,
where O
Dρ
→ j
∗
O
Dρ∩U0
is injective.
8.1.5. The proof of Theorem 8.1.6 contains the square
M⊗
Z
O
XΣ
//
ρ
O
Dρ
ρ
O
XΣ
//
ρ
O
Dρ
.
Describe the maps in this square carefully and prove that it commutes.
§8.2. Differential Forms on Toric Varieties 357
8.1.6. Show that the Euler sequence from Theorem 8.1.6 reduces to (8.1.8) when X =P
n
.
8.1.7. Sometimes the name Euler sequence is used to refer to an exact sequence for the
tangent sheaf T
XΣ
of a smooth toric variety.
(a) Show that for P
n
, we have an exact sequence
0 −→O
P
n −→O
P
n (1)
n+1
−→T
P
n −→0.
Hint: Use (8.1.8).
(b) What is the corresponding sequence for a general smooth toric variety X
Σ
for Σ as in
Theorem 8.1.6?
8.1.8. Let f be a homogeneous polynomial of degree d in C[x
1
, . . . , x
n
]. The classical Euler
relation is the equation
(8.1.9)
n
i=1
x
i
∂ f
∂x
i
=d f .
In this exercise, you will prove this relation and consider generalizations encoded by the
generalized Euler sequence from a toric variety.
(a) Prove (8.1.9). Hint: Differentiate the equation
f (tx
1
, . . . , tx
d
) =t
d
f (x
1
, . . . , x
n
)
with respect to t.
(b) To see how the classical Euler relation generalizes, recall from Chapter 5 that given
a toric variety X
Σ
with no torus factors (i.e., ¦u
ρ
[ ρ ∈ Σ(1)¦ spans N
R
), we have the
total coordinate ring
S = C[x
ρ
[ ρ ∈ Σ(1)],
graded by Cl(X
Σ
). The graded pieces S
β
for β ∈ Cl(X
Σ
) consist of homogeneous
polynomials as described by (5.2.1) from Chapter 5. If φ ∈ Hom
Z
(Cl(X
Σ
), Z) and
f ∈ S
β
show that we have a generalized Euler relation
ρ∈Σ(1)
φ([D
ρ
])x
ρ
∂ f
∂x
ρ
=φ(β) f .
Hint: Follow what you did for part a, which is the case X =P
n−1
.
(c) When Cl(X
Σ
) has rank greater than 1, there will be several distinct generalized Euler
relations on homogeneous elements of S. Compute the Euler relations on X =P
1
P
1
.
§8.2. Differential Forms on Toric Varieties
We now study the sheaves of pforms Ω
p
X
Σ
and Zariski pforms
´
Ω
p
X
Σ
of a toric
variety X
Σ
. Recall from §8.0 that the canonical sheaf is ω
X
Σ
=
´
Ω
n
X
Σ
, n =dimX
Σ
.
Properties of Wedge Products. We will need the following properties of wedge
products of free Rmodules.
358 Chapter 8. The Canonical Divisor of a Toric Variety
Proposition 8.2.1. Let F, G, H be free Rmodules of ﬁnite rank.
(a) An Rmodule homomorphism φ : F →G induces a homomorphism
_
p
φ :
_
p
F −→
_
p
G.
(b) Let 0 →F →G→H →0 be an exact sequence with rankF =m and rankH =
n. Then rankG=m+n and there is a natural isomorphism
_
m+n
G≃
_
m
F ⊗
R
_
n
H.
Proof. Part (a) is straightforward (see Exercise 8.2.1 for an explicit description
of
_
p
φ), as is the rank assertion in part (b). For the isomorphism of part (b),
we assume n > 0 and deﬁne a map
_
m
F ⊗
R
_
n
H →
_
m+n
G as follows. If the
maps in the exact sequence are α : F →G and β : G →H, then one checks that
_
m
α :
_
m
F →
_
m
G is injective and
_
n
β :
_
n
G →
_
n
H is surjective. Then
µ⊗ν ∈
_
m
F ⊗
R
_
n
H maps to
_
m
α(µ) ∧ν
′
∈
_
m+n
G, where
_
n
ν
′
= ν. This
map is welldeﬁned and gives the desired isomorphism (Exercise 8.2.1).
A corollary of this proposition is that if 0 →F →G →H →0 is an exact
sequence of locally free sheaves on a variety X with rankF =m and rankH =m,
then rankG =m+n and there is a natural isomorphism
(8.2.1)
_
m+n
G ≃
_
m
F ⊗
O
X
_
n
H .
Example 8.2.2. Suppose that Y ⊆ X is a smooth subvariety of a smooth variety,
and let n =dimX, m =dimY. Then we have the exact sequence
0 −→I
Y
/I
2
Y
−→Ω
1
X
⊗
O
X
O
Y
−→Ω
1
Y
−→0
from Theorem 8.0.18, where I
Y
⊆O
X
is the ideal sheaf of Y. By (8.2.1), we obtain
_
n
(Ω
1
X
⊗
O
X
O
Y
) ≃
_
n−m
(I
Y
/I
2
Y
) ⊗
O
Y
_
m
Ω
1
Y
.
One can check that
_
n
(Ω
1
X
⊗
O
X
O
Y
) ≃
_
n
(Ω
1
X
) ⊗
O
X
O
Y
. Now recall that ω
X
=
_
n
Ω
1
X
and ω
Y
=
_
m
Ω
1
Y
and that the normal sheaf of Y ⊆X is N
Y/X
=(I
Y
/I
2
Y
)
∨
.
Hence the above isomorphism implies
ω
Y
≃ω
X
⊗
O
X
_
n−m
N
Y/X
.
This isomorphism is called the adjunction formula. ♦
The Canonical Sheaf of a Toric Variety. Our ﬁrst major result gives a formula for
the canonical sheaf of a toric variety.
Theorem 8.2.3. For a toric variety X
Σ
, the canonical sheaf ω
X
Σ
is given by
ω
X
Σ
≃O
X
Σ
_
−
ρ
D
ρ
_
.
Thus K
X
Σ
=−
ρ
D
ρ
is a torusinvariant canonical divisor on X
Σ
.
§8.2. Differential Forms on Toric Varieties 359
Proof. We ﬁrst assume that X
Σ
is smooth with no torus factors. Then we have the
Euler sequence
0 −→Ω
1
X
Σ
−→
ρ
O
X
Σ
(−D
ρ
) −→Pic(X
Σ
) ⊗
Z
O
X
Σ
−→0
from Theorem 8.1.6. Each O
X
Σ
(−D
ρ
) is a line bundle since X
Σ
is smooth, and if
we set r = [Σ(1)[, then one sees easily that Pic(X
Σ
) ⊗
Z
O
X
Σ
≃ O
r−n
X
Σ
. Hence we
can apply part (b) of Proposition 8.2.1 to obtain
(8.2.2)
_
n
Ω
1
X
Σ
⊗
O
X
Σ
_
r−n
O
r−n
X
Σ
≃
_
r
_
ρ
O
X
Σ
(−D
ρ
)
_
.
It follows by induction from Proposition 8.2.1 that the righthand side of (8.2.2) is
isomorphic to
ρ
O
X
Σ
(−D
ρ
) ≃O
X
Σ
_
−
ρ
D
ρ
_
.
Turning to the lefthand side of (8.2.2), note that
_
r−n
O
r−n
X
Σ
≃ O
X
Σ
, so that the
lefthand side is isomorphic to
_
n
Ω
1
X
Σ
=Ω
n
X
Σ
=ω
X
Σ
since X
Σ
is smooth. This proves the result when X
Σ
is smooth without torus factors.
In Exercise 8.2.2 you will deduce the result for an arbitrary smooth toric variety.
Now suppose that X
Σ
is normal but not necessarily smooth. Let j : U
0
⊆ X
Σ
be the inclusion map for U
0
=
ρ
U
ρ
. We saw in the proof of Theorem 8.1.4 that
U
0
is a smooth toric variety satsifying codim(X ¸U
0
) ≥ 2. Now consider ω
X
Σ
and
O
X
Σ
(−
ρ
D
ρ
). Since the fans for U
0
and X
Σ
have the same 1dimensional cones,
these sheaves become isomorphic over U
0
by the smooth case. Since these sheaves
are reﬂexive and codim(X ¸U
0
) ≥ 2, we conclude that ω
X
Σ
≃ O
X
Σ
(−
ρ
D
ρ
) by
Proposition 8.0.1.
Here are some examples.
Example 8.2.4. Theorem 8.2.3 implies that the canonical bundle of P
n
is
ω
P
n ≃O
P
n (−n−1)
for all n ≥1 since Cl(P
n
) ≃Z and D
0
∼D
1
∼ ∼D
n
. In Exercise 8.2.3, you will
see another way to understand and derive this isomorphism. ♦
Example 8.2.5. The previous example shows that ω
P
2 ≃ O
P
2 (−3). We will com
pute this directly using
ω
P
2 =Ω
2
P
2
=
_
2
Ω
1
P
2
and the description of Ω
1
P
2
as a rank 2 vector bundle given in Example 8.0.16.
Recall that the transition functions for this bundle are given by:
φ
20
=
_
1/y −x/y
2
0 −1/y
2
_
, φ
12
=
_
−y
2
/x
2
0
−y/x
2
y/x
_
, φ
10
=
_
−y/x
2
1/x
−1/x
2
0
_
.
360 Chapter 8. The Canonical Divisor of a Toric Variety
By Exercise 8.2.1, the corresponding maps on
_
2
are given by the determinants of
these 22 matrices:
_
2
φ
20
=
−1
y
3
,
_
2
φ
12
=
−y
3
x
3
,
_
2
φ
10
=
1
x
3
.
Note that each is a cube. It is also evident that
_
2
φ
10
=
_
2
φ
12
_
2
φ
20
,
so that these give the transition functions for a line bundle on P
2
.
On the other hand,
_
2
Ω
1
P
2
≃ O
P
2 (−3) says the canonical bundle of P
2
is
the third tensor power of the tautological bundle described in Examples 6.0.19
and 6.0.21. To see this directly, we ﬁrst need to calibrate the coordinate systems.
Example 8.0.16 used coordinates x, y fromU
σ
0
, and Example 6.0.19 used homoge
neous coordinates x
0
, x
1
, x
2
for P
2
, with the standard open cover U
i
=P
2
¸V(x
i
).
Letting x = x
0
/x
2
and y = x
1
/x
2
gives an isomorphism U
σ
0
=U
2
. Translating
coordinates for U
σ
1
, we have
_
1/x, y/x
_
=
_
1/(x
0
/x
2
), (x
1
/x
2
)/(x
0
/x
2
)
_
=
_
x
2
/x
0
, x
1
/x
0
_
,
hence U
σ
1
=U
0
. A similar computation shows U
σ
2
=U
1
. We are now set for the
ﬁnal calculation. Keep in mind that the φ
i j
are in the coordinate system with charts
U
σ
i
. We will use θ
i j
to denote the same transition function, but using the U
i
charts.
Thus we have
_
2
φ
20
= −1/y
3
= (−x
2
/x
1
)
3
= θ
12
_
2
φ
12
= −y
3
/x
3
= (−x
1
/x
0
)
3
= θ
01
_
2
φ
10
= 1/x
3
= (x
2
/x
0
)
3
= θ
02
.
Up to a sign, these are indeed the cubes of the transition functions that we computed
for the tautological bundle O
P
n (−1) in Example 6.0.19. In Exercise 8.2.4, you will
work through the deﬁnition of the canonical bundle directly to ﬁnd the transition
functions given in Example 8.0.16. ♦
Example 8.2.6. When we computed the class group of the Hirzebruch surface H
r
in Example 4.1.8, we wrote the divisors D
ρ
as D
1
, D
2
, D
3
, D
4
and showed that
D
3
∼D
1
D
4
∼rD
1
+D
2
.
Thus Cl(H
r
) = Pic(H
r
) ≃ Z
2
is freely generated by the classes of D
1
and D
2
. It
follows that the canonical bundle can be written
ω
Hr
≃O
Hr
(−D
1
−D
2
−D
3
−D
4
) ≃O
Hr
(−(r +2)D
1
−2D
2
)
by Theorem 8.2.3. ♦
§8.2. Differential Forms on Toric Varieties 361
The Canonical Module. For a toric variety X
Σ
without torus factors, the canonical
sheaf ω
X
Σ
comes from a graded Smodule, where S = C[x
ρ
[ ρ ∈ Σ(1)] is the total
coordinate ring of X
Σ
. This module is easy to describe explicitly.
Each variable x
ρ
∈ S has degree deg(x
ρ
) =[D
ρ
] ∈ Cl(X
Σ
). Deﬁne
β
0
=deg
_
ρ
x
ρ
_
=
_
ρ
D
ρ
¸
∈ Cl(X
Σ
).
Then S(−β
0
) is the graded Smodule where S(−β
0
)
α
=S
α−β
0
for α ∈ Cl(X
Σ
). As
in §5.3, the coherent sheaf associated to S(−β
0
) is denoted O
X
Σ
(−β
0
). We have
the following result.
Proposition 8.2.7. O
X
Σ
(−β
0
) ≃ω
X
Σ
.
Proof. According to Proposition 5.3.7, O
X
Σ
(−β
0
) ≃O
X
Σ
(D) for any Weil divisor
with −β
0
= [D] ∈ Cl(X
Σ
). The deﬁnition of β
0
allows us to pick D = −
ρ
D
ρ
.
Then Theorem 8.2.3 implies
O
X
Σ
(−β
0
) ≃O
X
Σ
_
−
ρ
D
ρ
_
≃ω
X
Σ
.
We call S(−β
0
) the canonical module of S.
Corollary 8.2.8. For any normal toric variety X
Σ
we have an exact sequence
0 −→ω
X
Σ
−→O
X
Σ
−→
ρ
O
Dρ
.
Proof. First suppose that X
Σ
has no torus factors. Multiplication by
ρ
x
ρ
induces
an exact sequence of graded Smodules
0 −→S(−β
0
) −→S −→
ρ
S/¸x
ρ
)
since β
0
=deg(
ρ
x
ρ
). The sheaf associated to S/¸x
ρ
) is O
Dρ
(Exercise 8.2.5), and
then we are done by Example 6.0.10 and Proposition 8.2.7. When X
Σ
has a torus
factor, the result follows using the strategy described in Exercise 8.2.2.
We can also describe ω
X
Σ
in terms of nforms in the x
ρ
. Fix a basis e
1
, . . . , e
n
of
M. For each nelement subset I =¦ρ
1
, . . . , ρ
n
¦ ⊆Σ(1), we get the nn determinant
det(u
I
) =det(¸e
i
, u
ρ
j
)).
This depends on the ordering of the ρ
i
, as does the nform dx
ρ
1
∧ ∧dx
ρn
, though
the product
det(u
I
)dx
ρ
1
∧ ∧dx
ρn
depends only on e
1
, . . . , e
n
. It follows that the nform
(8.2.3) Ω
0
=
[I[=n
det(u
I
)
_
ρ/ ∈I
x
ρ
_
dx
ρ
1
∧ ∧dx
ρn
is welldeﬁned up to ±1. If we set deg(dx
ρ
) = deg(x
ρ
), then Ω
0
∈
_
n
Ω
S/C
is
homogeneous of degree β
0
. This gives the submodule
SΩ
0
⊆
_
n
Ω
S/C
,
362 Chapter 8. The Canonical Divisor of a Toric Variety
which is isomorphic to S(−β
0
) since deg(Ω
0
) =β
0
.
To see where Ω
0
comes from, let L =C(x
ρ
[ ρ ∈ Σ(1)) be the ﬁeld of fractions
of S. Then L is the function ﬁeld of C
Σ(1)
, and the surjective toric morphism
π : C
Σ(1)
¸Z(Σ) −→X
Σ
from Proposition 5.1.9 induces an injection on function ﬁelds
π
∗
: C(X
Σ
) −→L.
Furthermore, the proof of Theorem 5.1.10 implies that for any m∈M, the character
χ
m
∈ C(X
Σ
) maps to
π
∗
(χ
m
) =
ρ
x
¸m,uρ)
ρ
.
We regard C(X
Σ
) as a subﬁeld of L via π
∗
, so that C(X
Σ
) ⊆L and χ
m
=
ρ
x
¸m,uρ)
ρ
.
This induces an inclusion of K¨ ahler nforms
(8.2.4)
_
n
Ω
C(X
Σ
)/C
⊆
_
n
Ω
L/C
.
A basis e
1
, . . . , e
n
of M gives coordinates t
i
=χ
e
i
for the torus T
N
, and we know
from §8.1 that
dt
i
t
i
is a T
N
invariant section of Ω
1
T
N
. Then
dt
1
t
1
∧ ∧
dt
n
t
n
is a T
N
invariant section of Ω
n
T
N
which pulls back to a rational nform in
_
n
Ω
L/C
via (8.2.4). Note that
dt
i
t
i
=
ρ
¸e
i
, u
ρ
)
dx
ρ
x
ρ
since t
i
=
ρ
x
¸e
i
,uρ)
ρ
. When we mutiply by
ρ
x
ρ
to clear denominators, we obtain
ρ
x
ρ
dt
1
t
1
∧ ∧
dt
n
t
n
=
ρ
x
ρ
_
ρ
¸e
1
, u
ρ
)
dx
ρ
x
ρ
_
∧ ∧
_
ρ
¸e
n
, u
ρ
)
dx
ρ
x
ρ
_
=
[I[=n
det(u
I
)
_
ρ/ ∈I
x
ρ
_
dx
ρ
1
∧ ∧dx
ρn
=Ω
0
.
Hence Ω
0
arises in a completely natural way.
This can be interpreted in terms of sheaves as follows. We regard (8.2.4) as an
inclusion of constant sheaves on X
Σ
. Then the inclusion
ω
X
Σ
⊆
_
n
Ω
C(X
Σ
)/C
from Proposition 8.0.21 induces an inclusion
ω
X
Σ
⊆
_
n
Ω
L/C
.
On the other hand, it is easy to see that SΩ
0
⊆
_
n
Ω
S/C
induces an inclusion
¯
SΩ
0
⊆
_
n
Ω
L/C
.
§8.2. Differential Forms on Toric Varieties 363
Using the above derviation of Ω
0
, one can prove that
¯
SΩ
0
=ω
X
Σ
as subsheaves of the constant sheaf
_
n
Ω
L/C
—see [112, Prop. 14.14]. Note that
this is an equality of sheaves, not just an isomorphism. This shows that from the
point of view of differential forms, SΩ
0
deserves to be called the canonical module.
It is isomorphic to the earlier version S(−β
0
) of the canonical module via the map
that takes Ω
0
∈ SΩ
0
to 1 ∈ S(−β
0
).
The Afﬁne Case. The canonical sheaf ω
Uσ
of a normal afﬁne toric variety U
σ
is
determined by its module of global sections, the canonical module. When we think
of ω
Uσ
as the ideal sheaf O
Uσ
(−
ρ
D
ρ
), we get the ideal
Γ(U
σ
, ω
Uσ
) ⊆Γ(U
σ
, O
Uσ
) =C[σ
∨
∩M]
Proposition 8.2.9. Γ(U
σ
, ω
Uσ
)⊆C[σ
∨
∩M] is the ideal generated by the characters
χ
m
for all m ∈ M in the interior of σ
∨
.
Proof. Corollary 8.2.8 gives the exact sequence
0 −→Γ(U
σ
, ω
Uσ
) −→C[σ
∨
∩M] −→
ρ
C[σ
∨
∩M]/I
ρ
,
where I
ρ
=I(D
ρ
∩U
σ
) is the ideal of D
ρ
∩U
σ
⊂U
σ
. Since
I
ρ
=
m∈σ
∨
∩M,¸m,uρ)>0
C χ
m
⊆
m∈σ
∨
∩M
C χ
m
=C[σ
∨
∩M]
by (8.1.4), it follows that Γ(U
σ
, ω
X
Σ
) is the direct sum of C χ
m
for all m∈ M such
that ¸m, u
ρ
) > 0 for all ρ. Since the u
ρ
generate σ, this is equivalent to saying that
m is in the interior of σ
∨
.
The Projective Case. A full dimensional lattice polytope P ⊆M
R
gives two inter
ested graded rings, each with its own canonical module. For the ﬁrst, we use an
auxiliary variable t so that for every k ∈ N, a lattice point m ∈ (kP) ∩M gives a
character χ
m
t
k
on T
N
C
∗
. These characters span the algebra
(8.2.5) S
P
=
m∈(kP)∩M
C χ
m
t
k
.
This ring graded by setting deg(χ
m
t
k
) =k. We can also describe S
P
in terms of the
cone over P deﬁned by
C(P) =Cone(P¦1¦) ⊆M
R
R.
The “slice” of C(P) at height k is kP (see Figure 4 is §2.2), which implies that S
P
is the semigroup algebra
S
P
=C[C(P) ∩(MZ)]
364 Chapter 8. The Canonical Divisor of a Toric Variety
from the proof of Theorem 5.4.8. Then Proposition 8.2.9 tells us that the canonical
sheaf of the associated toric variety comes from the ideal
I
P
=
m∈Int(kP)∩M
C χ
m
t
k
⊆S
P
.
This is the canonical module of S
P
.
The second ring associated to P uses the total coordinate ring S of the projective
toric variety X
P
and the corresponding ample divisor D
P
. Let α = [D
P
] ∈ Cl(X
P
)
be the divisor class of D
P
. The graded pieces S
kα
⊆S, k ∈ N, form the graded ring
S
•α
=
∞
k=0
S
kα
.
The canonical module of S is S(−β
0
), which is isomorphic to the ideal
¸
ρ
x
ρ
_
⊆S
via multiplication by
ρ
x
ρ
since β
0
= deg
_
ρ
x
ρ
). When restricted to S
•α
, this
gives the ideal
I
•α
=
∞
k=0
I
kα
⊆S
•α
,
where I
kα
⊂ S
kα
consisting of monomials of degree α in which every variable ap
pears to a positive power.
It follows that the polytope P gives graded rings S
P
and S
•α
, each of which has
an ideal representing the canonical module. These are related as follows.
Theorem 8.2.10. The graded rings S
P
and S
•α
are naturally isomorphic via an
isomorphism that takes I
P
to I
•α
.
Proof. The proof of Theorem 5.4.8 used homogenization to construct an isomor
phism S
P
≃S
•α
. It is straightforward to see that this isomorphism carries the ideal
I
P
⊆S
P
to the ideal I
•α
⊆S
•α
(Exercise 8.2.6).
For readers familiar with the Proj construction (see the appendix to Chapter 7),
we note that X
P
≃ Proj(S
P
) by Theorem 7.A.1. Furthemore, graded S
P
modules
give quasicoherent sheaves on Proj(S
P
), as described in [78, II.5] (this is similar to
the construction given in §5.3). Then the following is true (Exercise 8.2.7).
Proposition 8.2.11. The sheaf on X
P
associated to the ideal I
P
⊆ S
P
by the Proj
construction is the canonical sheaf of X
P
.
The situation becomes even nicer when P is a normal polytope. The ample
divisor D
P
is very ample and hence embeds X
P
into a projective space, which gives
the homogeneous coordinate ring C[X
P
]. As we learned in §2.0, C[X
P
] is also the
ordinary coordinate ring of its afﬁne cone
´
X
P
, i.e.,
C[X
P
] =C[
´
X
P
].
§8.2. Differential Forms on Toric Varieties 365
Furthermore, since P is normal, Theorem 5.4.8 gives isomorphisms
S
P
≃S
•α
≃C[X
P
]
and implies that
´
X
P
is the normal afﬁne toric variety given by
´
X
P
=Spec(S
P
) =Spec(S
•α
).
It follows that the canonical sheaf of the afﬁne cone of X
P
comes from the ideal
I
P
⊆S
P
, and when we use the grading on S
P
and I
P
, they give X
P
and its canonical
sheaf via X
P
≃Proj(S
P
). Everything ﬁts together very nicely.
There is a more general notion of canonical module that applies to any graded
CohenMacaulay ring S
•
=
∞
k=0
S
k
where S
0
is a ﬁeld—see [29, Sec. 3.6]. The
canonical modules of S
P
≃S
•α
constructed above are canonical in this sense.
When the Canonical Divisor is Cartier. For a toric variety X
Σ
, the Weil divisor
K
X
Σ
=−
ρ
D
ρ
is called the canonical divisor. Note that K
X
Σ
need not be a Cartier
divisor if X
X
Σ
is not smooth. In fact, from Theorem 4.2.8, we have the following
characterization of the cases when the canonical divisor is Cartier.
Proposition 8.2.12. Let X
Σ
be a normal toric variety. Then K
X
Σ
is Cartier if and
only if for each maximal cone σ ∈ Σ, there exists m
σ
∈ M such that
¸m
σ
, u
ρ
) =1 for all ρ ∈ σ(1).
Similarly, K is QCartier if and only if for each maximal σ ∈ Σ, there exists
m
σ
∈ M
Q
such that ¸m
σ
, u
ρ
) =1 for all ρ ∈ σ(1).
Example 8.2.13. Let σ = Cone(de
1
−e
2
, e
2
) ⊆ R
2
. The afﬁne toric variety U
σ
is the rational normal cone
´
C
d
. We computed Cl(
´
C
d
) ≃ Z/dZ in Example 4.1.4,
where the Weil divisors D
1
, D
2
coming from the rays satisfy D
2
∼D
1
, dD
1
∼0.
The canonical divisor K
b
C
d
= −D
1
−D
2
has divisor class corresponding to
[−2] ∈ Z/dZ. Since the Picard group of a normal afﬁne toric variety is trivial
(Proposition 4.2.2), it follows that K
b
C
d
is Cartier if and only if d ≤2.
Another way to see this is via Proposition 8.2.12, where one easily computes
that m
σ
=(2/d)e
1
+e
2
satisﬁes ¸m
σ
, de
1
−e
2
) =¸m
σ
, e
2
) =1. This lies in M=Z
2
if and only if d ≤2. ♦
We will use the following terminology.
Deﬁnition 8.2.14. Let K
X
be a canonical divisor on a normal variety X. We say
that X is Gorenstein if K
X
is a Cartier divisor.
Since the canonical sheaf ω
X
=O
X
(K
X
) is reﬂexive, we have
X is Gorenstein ⇐⇒ K
X
is Cartier ⇐⇒ ω
X
is a line bundle
by Proposition 8.0.7. All smooth varieties are Gorenstein, of course. We will study
further examples of singular Gorenstein varieties in the next section of the chapter.
366 Chapter 8. The Canonical Divisor of a Toric Variety
Reﬁnements. A reﬁnement Σ
′
of a fan Σ induces a proper birational toric mor
phism φ : X
Σ
′ →X
Σ
. The canonical sheaves of X
Σ
′ and X
Σ
are related as follows.
Theorem 8.2.15. Let φ : X
Σ
′ →X
Σ
be the toric morphism induced by a reﬁnement
Σ
′
of a fan Σ in N
R
≃R
n
. Then
φ
∗
ω
X
Σ
′
≃ω
X
Σ
.
In particular, if Σ
′
is a smooth reﬁnement of Σ, then
φ
∗
Ω
n
X
Σ
′
≃ω
X
Σ
.
Proof. First assume X
Σ
=U
σ
, so that Σ
′
reﬁnes σ. Since K
Σ
′ = −
ρ
′
∈Σ
′
(1)
D
ρ
′ ,
the description of global sections given in Proposition 4.3.3 implies that
Γ(X
Σ
′
, ω
X
Σ
′
) =
m∈P
′
∩M
C χ
m
,
where
P
′
=¦m ∈ M
R
[ ¸m, u
ρ
′ ) ≥1 for all ρ
′
∈ Σ
′
(1)¦.
We clearly have P
′
∩M⊆Int(σ
∨
) ∩M since σ(1) ⊆Σ
′
(1). The opposite inclusion
also holds, as we now prove. Given m ∈ Int(σ
∨
) ∩M, we have ¸m, u) > 0 for all
u ,= 0 in σ. In particular, ¸m, u
ρ
′ ) > 0 for all ρ
′
∈ Σ
′
(1). This is an integer, so that
¸m, u
ρ
′ ) ≥1, which implies m ∈ P
′
∩M, as desired. Since
Γ(U
σ
, ω
Uσ
) =
m∈Int(σ
∨
)∩M
C χ
m
by Proposition 8.2.9, we conclude that Γ(X
Σ
′ , ω
X
Σ
′
) = Γ(U
σ
, ω
Uσ
). Then we have
φ
∗
ω
X
Σ
′
≃ω
Uσ
since U
σ
is afﬁne.
In the general case, X
Σ
is covered by afﬁne open subsets U
σ
for σ ∈ Σ, and
φ
−1
(U
σ
) is the toric variety of the reﬁnement of σ induced by Σ
′
. The above
paragraph gives isomorphisms
φ
∗
ω
X
Σ
′ [
Uσ
≃ω
X
Σ
[
Uσ
which are compatible with the inclusion U
τ
⊆U
σ
when τ is a face of σ. Hence
these isomorphisms patch to give the desired isomorphism φ
∗
ω
X
Σ
′
≃ω
X
Σ
.
In Chapter 11, we will prove the existence of smooth reﬁnements. It follows
that for any ndimensional toric variety X
Σ
, there are two ways of constructing the
canonical sheaf ω
X
Σ
from the sheaf of nforms of a smooth toric variety:
• (Internal) ω
X
Σ
= j
∗
Ω
n
U
0
, where j : U
0
⊆X
Σ
is the inclusion of the smooth locus
of X
Σ
.
• (External) ω
X
Σ
= φ
∗
Ω
n
X
Σ
′
, where φ : X
Σ
′ →X
Σ
is the toric morphism coming
from a smooth reﬁnement Σ
′
of Σ.
§8.2. Differential Forms on Toric Varieties 367
Sheaves of pforms. The sheaves
´
Ω
p
X
Σ
for 1 < p <n =dim(X
Σ
) are also important
for the geometry of X
Σ
. We construct a sequence
(8.2.6) 0 −→
´
Ω
p
X
Σ
αp
−→
_
p
M⊗
Z
O
X
Σ
βp
−→
ρ
_
p−1
(ρ
⊥
∩M) ⊗
Z
O
Dρ
as follows. The map α : Ω
1
X
Σ
→M⊗
Z
O
X
Σ
from (8.1.3) induces
_
p
α : Ω
p
X
Σ
=
_
p
Ω
1
X
Σ
−→
_
p
M⊗
Z
O
X
Σ
,
and then α
p
=
__
p
α
_
∨∨
. To deﬁne β
p
, recall that for u ∈ N, the contraction map
i
u
:
_
p
M →
_
p−1
M has the property that
i
u
(m
1
∧ ∧m
p
) =
p
i=1
(−1)
i−1
¸m
i
, u)m
1
∧ ∧ ´ m
i
∧ ∧m
p
when m
1
, . . . , m
p
∈ M. Note also that im(i
u
) ⊆
_
p−1
(u
⊥
∩M) (Exercise 8.2.8).
An element ρ ∈ Σ(1) gives u
ρ
as usual. Using i
uρ
:
_
p
M →
_
p−1
(ρ
⊥
∩M)
and O
X
Σ
→O
Dρ
, we obtain the composition
_
p
M⊗
Z
O
X
Σ
−→
_
p−1
(ρ
⊥
∩M) ⊗
Z
O
X
Σ
−→
_
p−1
(ρ
⊥
∩M) ⊗
Z
O
Dρ
.
This gives a natural map
β
p
:
_
p
M⊗
Z
O
X
Σ
−→
ρ
_
p−1
(ρ
⊥
∩M) ⊗
Z
O
Dρ
.
Theorem 8.2.16. The sequence (8.2.6) is exact for any normal toric variety X
Σ
.
Proof. We begin with the smooth case. Since exactness of a sheaf sequence is
local, we can work over an afﬁne toric variety U
σ
= C
r
(C
∗
)
n−r
, where σ is
generated by the ﬁrst r elements of a basis of e
1
, . . . , e
n
of N. By abuse of notation,
e
1
, . . . , e
n
will denote the corresponding dual basis of M. Setting x
i
=χ
e
i
, we get
U
σ
=Spec(R), R =C[x
1
, . . . , x
r
, x
±1
r+1
, . . . , x
±1
n
].
Then (8.2.6) comes from the exact sequence of Rmodules
0 −→
_
p
Ω
R/C
αp
−→
_
p
M⊗
Z
R
βp
−→
r
i=1
_
p−1
(e
⊥
i
∩M) ⊗
Z
R/¸x
i
),
where α
p
maps dx
i
1
∧ ∧dx
ip
to e
i
1
∧ ∧e
ip
⊗x
i
1
x
ip
by the description of α
given in (8.1.3).
It is thus obvious that β
p
◦α
p
= 0. To prove exactness, we regard
_
p
M⊗
Z
R
as
_
p
F, where F is the free R module with basis e
1
, . . . , e
n
. Now suppose that
β
p
(ω) =0 for some ω ∈
_
p
F. We can write ω uniquely as
ω =
i
1
<<ip
f
i
1
...ip
e
i
1
∧ ∧e
ip
.
Exactness will follow once we prove x
i
1
x
ip
divides f
i
1
...ip
for all i
1
< < i
p
.
368 Chapter 8. The Canonical Divisor of a Toric Variety
If an index i > r appears in f
i
1
...ip
, then x
i
automatically divides f
i
1
...ip
since x
i
is invertible in R. Now suppose that an index i ≤ r appears in f
i
1
...ip
. We can write
ω =e
i
∧ω
1
+ω
2
, where e
i
does not appear in ω
2
and all f
i
1
...ip
involving the index
i appear in ω
1
. Since i
e
1
(ω) =ω
1
, the only way for β
p
(ω) to vanish is for f
i
1
...ip
to
be zero in R/¸x
i
), i.e., for x
i
to divide f
i
1
...ip
. This completes the proof of exactness
in the smooth case.
The proof for the general case follows from the smooth case by an argument
similar to what we did in the proof of Theorem 8.1.4.
Note that when p = 1, the exact sequence (8.2.6) reduces to the sequence ap
pearing in Theorem 8.1.4, and when p =n, we get the sequence in Corollary 8.2.8
since ω
X
Σ
=
´
Ω
n
X
Σ
.
When X
Σ
has no torus factors, it is easy to ﬁnd a graded Smodule whose
associated sheaf is
´
Ω
p
X
Σ
. Adapting the deﬁnition of β
p
in (8.2.6) to the module
case, we get a homomorphism
_
p
M⊗
Z
S −→
ρ
_
p−1
(ρ
⊥
∩M) ⊗
Z
S/¸x
ρ
)
of graded Smodules whose kernel we denote
´
Ω
p
S
. This gives an exact sequence of
graded Smodules
0 −→
´
Ω
p
S
−→
_
p
M⊗
Z
S −→
ρ
_
p−1
(ρ
⊥
∩M) ⊗
Z
S/¸x
ρ
).
Using Example 6.0.10, we obtain the following corollary of Theorem 8.2.16.
Corollary 8.2.17. If X
Σ
has no torus factors, then
´
Ω
p
X
Σ
is the sheaf associated to
the graded Smodule
´
Ω
p
S
.
The Afﬁne Case. For an afﬁne toric variety U
σ
, the sheaf
´
Ω
p
Uσ
is determined by its
global sections, which can be described using Theorem 8.2.16. We will need the
following notation. If m ∈ M and σ ∈ Σ, let
V
σ
(m) =Span
C
(m
0
∈ M [ m
0
∈ the minimal face of σ
∨
containing m) ⊆M
C
,
where M
C
=M⊗
Z
C. Here is a result due to Danilov [42, Prop. 4.3].
Proposition 8.2.18. Let U
σ
be the toric variety of the cone σ. Then we have an
isomorphism
Γ(U
σ
,
´
Ω
p
Uσ
) ≃
m∈σ
∨
∩M
_
p
V
σ
(m) χ
m
.
Proof. Using the inclusion ρ
⊥
∩M⊆M and the exact sequence of Theorem 8.2.16
over U
σ
, we obtain the exact sequence
0 −→Γ(U
σ
,
´
Ω
p
Uσ
) −→
_
p
M⊗
Z
Γ(U
σ
, O
Uσ
)
βp
−→
ρ
_
p−1
M⊗
Z
Γ(U
σ
, O
Dρ
)
§8.2. Differential Forms on Toric Varieties 369
of modules over Γ(U
σ
, O
X
Σ
) =C[σ
∨
∩M] =
m∈σ
∨
∩M
C χ
m
. Also note that
(8.2.7) Γ(U
σ
, O
Dρ
) =
m∈σ
∨
∩M,¸m,uρ)=0
C χ
m
by the analysis given in the proof of Theorem 8.1.1. It follows that we have a direct
sum decomposition Γ(U
σ
,
´
Ω
p
Uσ
) =
m∈σ
∨
∩M
Γ(U
σ
,
´
Ω
p
Uσ
)
m
, where
(8.2.8) 0 −→Γ(U
σ
,
´
Ω
p
Uσ
)
m
−→
_
p
M
C
βp
−→
¸m,uρ)=0
_
p−1
M
C
is exact for m ∈ σ
∨
∩M and β
p
is the sum of the contraction maps i
uρ
for all ρ
satisfying ¸m, u
ρ
) =0.
Thus ω ∈
_
p
M
C
is in the kernel of β
p
if and only if i
uρ
(ω) = 0 for all ρ
with ¸m, u
ρ
) = 0. You will show in Exercise 8.2.8 that i
uρ
(ω) = 0 if and only if
ω ∈
_
p
(ρ
⊥
)
C
. It follows that the kernel of β
p
in (8.2.8) is the intersection
(8.2.9)
¸m,uρ)=0
_
p
(ρ
⊥
)
C
=
_
p
_
¸m,uρ)=0
(ρ
⊥
)
C
_
.
However, the intersection
F =
¸m,uρ)=0
ρ
⊥
∩σ
∨
is the minimal face of σ
∨
containing m, and one sees easily that
Span
R
(m
0
∈ M [ m
0
∈ F) =
¸m,uρ)=0
ρ
⊥
.
It follows that V
σ
(m) =
¸m,uρ)=0
(ρ
⊥
)
C
. This plus (8.2.9) imply that the kernel of
β
p
in (8.2.8) is
_
p
V
σ
(m), as claimed.
The Simplicial Case. When X
Σ
is simplicial, the sequence (8.2.6) is exact on the
right when p = 1 by part (b) of Theorem 8.1.4. For p > 1, similar though longer
exact sequences exist in the simplicial case, as we now describe without proof.
Given a fan Σ, consider the sheaf on X
Σ
deﬁned by
K
j
(Σ, p) =
σ∈Σ( j)
_
p−j
(σ
⊥
∩M) ⊗
Z
O
V(σ)
,
where V(σ) =O(σ) ⊆X
Σ
is the orbit closure corresponding to σ ∈ Σ. Thus
K
0
(Σ, p) =
_
p
M⊗
Z
O
X
Σ
K
1
(Σ, p) =
ρ∈Σ(1)
_
p−1
(ρ
⊥
∩M) ⊗
Z
O
V(ρ)
=
ρ∈Σ(1)
_
p−1
(ρ
⊥
∩M) ⊗
Z
O
Dρ
,
and the exact sequence (8.2.6) can be written
0 −→
´
Ω
p
X
Σ
−→K
0
(Σ, p) −→K
1
(Σ, p).
When X
Σ
is simplicial, one can extend this exact sequence as follows.
370 Chapter 8. The Canonical Divisor of a Toric Variety
Theorem 8.2.19. When X
Σ
is simiplicial, there is an exact sequence
0 −→
´
Ω
p
X
Σ
−→K
0
(Σ, p) −→K
1
(Σ, p) −→ −→K
p
(Σ, p) −→0.
A discussion and proof of this result can be found in [137, Sec. 3.2]. We
will use this exact sequence in Chapter 9 to prove a vanishing theorem for sheaf
cohomology on simpicial toric varieties.
Exercises for §8.2.
8.2.1. Given a map of free Rmodules φ : F →G, we can pick bases of F, G and represent
φ by a n m matrix with entries in R, where m = rank(F), n = rank(G). These bases
induce bases of the wedge products
_
p
F and
_
p
G. Then prove that the induced map
_
p
φ :
_
p
F →
_
p
G is given by the p p minors of φ (with appropriate signs).
8.2.2. Complete the proof of Theorem 8.2.3 in the smooth case by showing how to reduce
to the case when X
Σ
is smooth with no torus factors. Hint: First prove that any smooth
toric variety is equivariantly isomorphic to a product X
Σ
(C
∗
)
r
, where X
Σ
is a smooth
toric variety with no torus factors. Then consider X
Σ
C
r
.
8.2.3. Let T
N
be the torus of P
n
. If x
0
, . . . , x
n
are the usual homogeneous coordinates, then
y
i
= x
i
/x
0
for i = 1, . . . , n are afﬁne coordinates for the open subset where x
0
,= 0. The
differential nform ω =
dy1
y1
∧ ∧
dyn
yn
spans Γ(T
N
, Ω
n
TN
) = Γ(T
N
,
_
n
Ω
1
TN
) and has poles of
order 1 along each D
i
= V(x
i
) for i = 1, . . . , n. Show that if we write z
j
= x
j
/x
i
for the
afﬁne coordinates on the complement of V(x
i
), and change coordinates to the z
j
, j ,= i,
then we can see ω also has a pole of order 1 along V(x
0
). Hence on P
n
, ω deﬁnes a section
of O
P
n (−
n
i=0
D
i
).
8.2.4. Using the open subsets U
σi
⊆P
2
(see Figure 2 in Example 3.1.9), compute Ω
1
P
2
(U
σi
),
and write down the transition functions on U
σi
∩U
σj
. Compare to Example 8.0.16. If the
result of your computation differs, describe how you can explain this via a change of basis.
8.2.5. Let S be the total coordinate ring of a toric variety X
Σ
without torus factors. Prove
that O
Dρ
is the sheaf associated to the graded Smodule S/¸x
ρ
). Hint: Consider the exact
sequence 0 →¸x
ρ
) →S →S/¸x
ρ
) →0 and apply Proposition 5.3.7 and Example 6.0.10.
8.2.6. Prove Theorem 8.2.10.
8.2.7. (For readers familiar with the Proj construction) Prove that the sheaf associated to
the ideal I
P
⊆S
P
from Theorem 8.2.10 is the canonical sheaf of X
P
=Proj(S
P
). Hint: Let S
be the total coordinate ring of X
P
. We saw in §5.3 that a graded Smodule such as S(−β
0
)
gives a sheaf on X
P
. Compare this to how I
•α
gives a sheaf on Proj(S
•α
). See the proof of
Theorem 7.A.1.
8.2.8. Let F be a free module of ﬁnite rank over a domain R. Given u ∈ F
∨
, we get the
kernel u
⊥
⊆F and the contraction map i
u
:
_
p
F →
_
p−1
F. Assume u =re
1
where r ∈ R
and e
1
, . . . , e
n
is a basis of F
∨
.
(a) Prove that the image of i
u
is contained in
_
p−1
u
⊥
.
(b) Given ω ∈
_
p
F, prove that i
u
(ω) =0 if and only if ω ∈
_
p
u
⊥
.
§8.3. Fano Toric Varieties 371
8.2.9. Assume that X
Σ
has no torus factors. For β ∈ Cl(X
Σ
), deﬁne
´
Ω
p
XΣ
(β) to be the sheaf
associated to the graded Smodule
´
Ω
p
S
(β). Prove that Γ(X
Σ
,
´
Ω
p
XΣ
(β)) ≃ (
´
Ω
p
S
)
β
. Hint: If
you get stuck, see [10, Prop. 8.5].
8.2.10. Compute the nformΩ
0
deﬁned in (8.2.3) for P
n
, P(q
0
, . . . , q
n
), and P
1
P
1
.
8.2.11. Prove that our favorite afﬁne toric variety V = V(xy −zw) ⊆ C
4
is Gorenstein.
Hint: Example 1.2.19.
8.2.12. Prove that a product of two Gorenstein toric varieties is again Gorenstein. Hint:
Use Propositions 3.1.14 and 8.2.12.
8.2.13. We will see in Chapter 10 that every 2dimensional rational cone in R
2
is lattice
equivalent to a cone of the form σ =Cone(e
2
, de
1
−ke
2
) for integers d > 0 and 0 ≤k < d
with gcd(d, k) =1. Prove that U
σ
is Gorenstein if and only if d =k +1.
8.2.14. A strongly convex rational polyhedral cone is Gorenstein if its associated afﬁne
toric variety is Gorenstein. In the discussion of the algebra (8.2.5), we saw that a full
dimensional lattice polytope P ⊆M
R
gives the cone
C(P) =Cone(P¦1¦) ⊆M
R
R.
(a) Prove that C(P) is Gorenstein.
(b) Prove that the dual cone C(P)
∨
⊆N
R
R is Gorenstein if and only if P is reﬂexive.
In general, a cone is reﬂexive Gorenstein is both the cone and its dual are Gorenstein.
Reﬂexive Gorenstein cones play an important role in mirror symmetry—see [11, 126].
8.2.15. Explain why S
•α−β0
=
∞
k=0
S
kα−β0
gives another model for the canonical module
of the graded ring S
•α
discussed in the text.
§8.3. Fano Toric Varieties
We ﬁnish this chapter with a discussion of an interesting class of projective toric
varieties and their corresponding polytopes.
Deﬁnition 8.3.1. A complete normal variety X is said to be a Gorenstein Fano
variety if the anticanonical divisor −K
X
is Cartier and ample.
Thus Gorenstein Fano varieties are projective. When X is smooth, we will
simply say that X is Fano.
Example 8.3.2. Example 8.2.5 shows that O
P
2 (−K
P
2) ≃O
P
2 (3) is ample, so P
2
is
a Fano variety. We continue this example to introduce the key ideas that will lead
to a classiﬁcation of 2dimensional Gorenstein Fano toric varieties.
The standard fan for P
2
= X
Σ
has minimal generators u
0
= −e
1
−e
2
, u
1
= e
1
and u
2
=e
2
. The polytope corresponding to the anticanonical divisor of P
2
is
P =¦m ∈ R
2
[ ¸m, u
i
) ≥−1, i =0, 1, 2¦.
In Exercise 8.3.1 you will check that
P =Conv(−e
1
−e
2
, 2e
1
−e
2
, −e
1
+e
2
).
372 Chapter 8. The Canonical Divisor of a Toric Variety
t
t
t
t
t
d
t
t
t
t
d
d
d
d
d
d
P
t
d
t
t
¨
¨
¨
¨
¡
¡
¡
¡d
d
P
◦
Figure 1. The anticanonical polytope P and its dual P
◦
for P
2
This lattice polygon is shown in Figure 1. The open circle in the ﬁgure represents
the origin, which is the unique interior lattice point of P. Also, by Exercise 8.3.1,
the dual polytope P
◦
=¦u ∈ N
R
[ ¸m, u) ≥−1 for all m ∈ P¦ is given by
P
◦
=Conv(e
1
, e
2
, −e
1
−e
2
),
which is the polytope generated by the ray generators of the fan of P
2
. ♦
Example 8.3.3. In Exercise 8.3.2, you will generalize Example 8.3.2 by showing
that the weighted projective space P(q
0
, . . . , q
n
) is Gorenstein Fano if and only if
q
i
[ q
0
+ +q
n
for all 0 ≤i ≤n. ♦
The special features of Example 8.3.2 involve an unexpected relation between
Fano toric varieties and the reﬂexive polytopes introduced in Chapter 2.
Fano Toric Varieties and Reﬂexive Polytopes. Recall from Deﬁnition 2.3.11 that
a lattice polytope in M
R
is reﬂexive if its facet presentation is
(8.3.1) P =¦m ∈ M
R
[ ¸m, u
F
) ≥−1 for all facets F¦.
It follows that if P is reﬂexive, the origin is the unique interior lattice point of P
(Exercise 2.3.5). Since a
F
=1 for all facets F, the dual polytope is
(8.3.2) P
◦
=Conv(u
F
[ F is a facet of P)
(Exercise 2.2.1). Finally, P
◦
is a lattice polytope and is reﬂexive (Exercise 2.3.5).
The polytopes pictured in Figure 1 are reﬂexive.
The following result gives the connection between projective Gorenstein Fano
toric varieties and reﬂexive polytopes generalizing what we saw above for P
2
.
Theorem 8.3.4. Let X
Σ
be a normal toric variety. If X
Σ
is a projective Gorenstein
Fano variety, then the polytope associated to the anticanonical divisor −K
X
Σ
=
−
ρ
D
ρ
is reﬂexive. Conversely, if X
P
is the projective toric variety associated to
a reﬂexive polytope P, then X
P
is a Gorenstein Fano variety.
§8.3. Fano Toric Varieties 373
Proof. If X
Σ
is Gorenstein Fano, then −K
X
Σ
is Cartier and ample. This implies
that polytope associated to −K
X
Σ
=
ρ
D
ρ
has facet presentation
P =¦m ∈ M
R
[ ¸m, u
ρ
) ≥−1 for all ρ ∈ Σ(1)¦.
and hence is reﬂexive by (8.3.1). Conversely, let P be a reﬂexive polytope in M
R
.
The normal fan Σ
P
of P deﬁnes the variety X
P
=X
Σ
P
. The facet presentation (8.3.1)
of P has a
F
=1 for every facet F of P. Hence the Cartier divisor corresponding to
P is D
P
=
F
D
F
= −K
X
P
. We proved that D
P
is ample in Proposition 6.1.4, so
that −K
X
P
is ample. Hence X
P
is Gorenstein Fano.
Classiﬁcation. By Theorem 8.3.4, classifying toric Gorenstein Fano varieties is
equivalent to classifying reﬂexive polytopes P in M
R
. Since reﬂexive polytopes
contain the origin as an interior point, “classify” means up to invertible linear maps
of M
R
induced by isomorphisms of M. This is called lattice equivalence. The
ﬁrst interesting case is in dimension two, where toric Gorenstein Fano surfaces
correspond to 16 equivalences classes of reﬂexive polygons.
We begin with some general facts about a reﬂexive polytope P. First note that
the lattice points of P are the origin and the lattice points lying on the boundary.
Furthermore, any boundary lattice point is primitive. We will also need two less
obvious results from [134].
The ﬁrst result concerns projections of reﬂexive polytopes. Given a reﬂexive
polygon P and a primitive element m ∈ M, there is a projection map
π
m
: M
R
−→M
R
/Rm
whose image is a polytope whose vertices lie in M/Zm.
Lemma 8.3.5. Let P be a reﬂexive polytope in M
R
≃R
n
and let m be a lattice point
m in the boundary of P. Then π
m
(P) is a in lattice polytope in M
R
/Rm containing
the origin as an interior lattice point, and
π
m
(P) =π
m
_
m∈F facet of P
F
_
.
Proof. For each p ∈ π
m
(P), π
−1
m
(p) is a line parallel to the line Rm. By taking the
point in P that maps to p and is farthest along this line in the direction of m, we see
that the points in π
m
(P) are in bijective correspondence with the points in
U =¦x ∈ P [ x +λm / ∈ P for all λ > 0¦.
If F is an facet of P containing m, one easily sees that F ⊆U. Conversely, let x ∈U.
One can show without difﬁculty there exists some facet F of P, not parallel to m,
containing x and such that ¸u
F
, m) < 0. Since P is reﬂexive and m ∈ M, ¸u
F
, m) is
an integer ≥ −1. Hence ¸u
F
, m) = −1, so that x, m ∈ F. Thus U ⊆
m∈F
F, and
from here the lemma follows easily.
374 Chapter 8. The Canonical Divisor of a Toric Variety
The second result is the following fact about pairs of lattice points on the
boundary of an reﬂexive polytope. The vertices are primitive vectors in M, but
we can have other lattice points on the boundary of P as well.
Lemma 8.3.6. Let m, m
′
be distinct lattice points on the boundary of a reﬂexive
polytope P. Then exactly one of the following holds:
(a) m and m
′
lie in a common edge of P,
(b) m+m
′
=0, or
(c) m+m
′
is also on the boundary of P.
The proof is left to the reader as Exercise 8.3.3.
The 2Dimensional Case. The following theorem classiﬁes reﬂexive polygons in
the plane M
R
≃ R
2
, up to lattice equivalence. Figure 2 on the next page shows 16
lattice polygons, where the open circle in the center of each polygon is the origin
and the solid circles are the lattice points on the boundary. The numbers in the
labels give the number of boundary lattice points. Polygons 3 and 9 are the dual
pair from Example 8.3.2.
Theorem 8.3.7. There are exactly 16 equivalence classes of reﬂexive polygons in
the plane, shown in Figure 2.
Proof. We will sketch the proof following [135, Proposition 4.1], and leave the
details for the reader to verify (Exercise 8.3.4). We consider several cases.
Case A. First assume that P is a reﬂexive polygon such that each edge contains
exactly two lattice points, and ﬁx one such pair. These lattice points must form a
basis for M since the triangle formed by these two vertices and the origin has lattice
points only at the vertices (part (a) of Exercise 8.3.4). Hence we can use a lattice
equivalence to place the two of them at e
1
and e
2
.
Subcase A.1. If P has exactly three vertices, the third is located at ae
1
+be
2
for some a, b ∈ Z. Since 0 is the only lattice point in the interior, we must have
a =b =−1 (part (b) of Exercise 8.3.4). This gives the polygon of type 3.
Subcase A.2. Next, still in Case A, assume that there are three distinct vertices
m, m
′
, m
′′
such that m+m
′
= m
′′
. There must be edges of P containing the pairs
m
′
, m
′′
and m
′
, m
′′
(part (c) of Exercise 8.3.4). Hence each pair must yield a basis
for M and we can place m = e
1
, m
2
=−e
1
+e
2
, and then m
′′
= e
2
. Project P from
m
′′
=e
2
using Lemma 8.3.5. It follows that P can only contain points −e
1
, 0, e
1
on
the line y =0 and −e
2
, e
1
−e
2
on the line y =−1. Moreover, P cannot contain any
points with y ≤−2. It follows that the only other possible polygons in this case are
4b, 5a, 6a (part (d) of Exercise 8.3.4) up to lattice equivalence.
Subcase A.3. Finally, in Case A, if we are not in subcases A.1 or A.2, then by
Lemma 8.3.6, if m and m
′
are not in the same edge, we must have m+m
′
=0. The
only possibility here is clearly polygon 4a.
§8.3. Fano Toric Varieties 375
t
d
t
t
¡
¡
¡
¡
¨
¨
¨
¨
d
d
3
t
t
d
t
t
d
d
d
d
4a
t t
d
t
t
e
e
e
e
d
d
4b
t t t
d
t
e
e
e
e¡
¡
¡
¡
4c
t t
t d t
t
d
d
d
d
5a
t t t
t d
t
d
d¡
¡
¡
¡
5b
t t
t d t
t t
d
d
d
d
6a
t t t
t d t
t
d
d
6b
t t t
t d t
t¨
¨
¨
¨
6c
t t t
t d
t
t
6d
t t t
t d t
t t
7a
t t t
t d
t t
t
¡
¡
¡
¡
7b
t
t
t
t
d
t
t
t
t
8a
t
t
t
t
t
d
t
t
t
8b
t
t
t
t
t
t
d
t
t
¡
¡
¡
¡
¡
¡
¡
¡
8c
t
t
t
t
t
d
t
t
t
t
d
d
d
d
d
d
9
Figure 2. The 16 equivalence classes of reﬂexive lattice polygons in R
2
Case B. Assume there are exactly three lattice points on some edge of P. By an
isomorphism of M, we can place these at −e
1
+e
2
, e
2
, e
1
+e
2
. Projecting from m=
e
2
and using Lemma 8.3.5, we see that P must be contained in the strip −1 ≤x ≤1.
Moreover, P cannot contain any points with y < −3, or else −e
2
,= 0 would be an
interior lattice point. Up to lattice equivalence, the possibilities are 4c, 5b, 6b, 6c,
6d, 7a, 7b, 8a, 8b, and 8c (part (e) of Exercise 8.3.4).
Case C. Assume no edge of P has exactly three lattice points and some edge has
four or more. Place the vertex of this edge at −e
1
−e
2
so that −e
1
, −e
1
+e
2
and
−e
1
+2e
2
also lie on this edge. Projecting from −e
1
via Lemma 8.3.5 shows that
P lies above the line y =−1, and since the origin is an interior point, there must be
a lattice point m =ae
1
+be
2
with a > 0 and b ≥−1.
Subcase C.1 If b = −1, then 2e
1
−e
2
must lie in P. If −e
1
+2e
2
and 2e
1
−e
2
do not lie on an edge of P, their sum e
1
+e
2
would lie in P by Lemma 8.3.6 and
force e
1
, e
2
to be interior points. This is impossible, so −e
1
+2e
2
and 2e
1
−e
2
lie
on an edge of P. Hence P has type 9.
376 Chapter 8. The Canonical Divisor of a Toric Variety
Subcase C.2 If b =0, then m=e
1
, for otherwise e
1
would be an interior point.
Applying Lemma 8.3.6 to e
1
and −e
1
+e
2
shows that e
2
∈ P. If no edge connects
e
1
, e
2
, then Lemma 8.3.6 would imply e
1
+e
2
∈ P. This point and −e
1
+2e
2
would
force e
2
to be interior, again impossible. Hence e
1
, e
2
lie on an edge, which forces P
to be contained in the polygon of type 9. Then our hypotheses on P force equality.
Subcase C.3 If b ≥1, then e
2
becomes an interior point of P (part (f) of Exer
cise 8.3.4). Hence this subcase cannot occur.
There are many other proofs of classiﬁcation given in Theorem 8.3.7, including
[39, 134, 145]. References to further proofs are given in [39, Thm. 6.10].
The collection of 2dimensional reﬂexive polygons also exhibits some very
interesting symmetries and regularities. For instance, we know that if P is reﬂexive,
then its dual is also reﬂexive. In Exercise 8.3.5 you will determine which are the
dual pairs in the list above. There is also a very interesting relation between the
numbers of boundary lattice points in P and P
◦
: in all cases, we have
(8.3.3) [∂P∩M[ +[∂P
◦
∩N[ =12.
We will see in Chapter 10 that there is an explanation for this coincidence coming
from the cohomology theory of sheaves on surfaces.
Higher Dimensions. Reﬂexive polytopes of dimension greater than two and their
associated toric varieties are currently being actively studied. See for instance [135]
and the references therein. One reason for the interest in these varieties is the rela
tion with mirror symmetry in physics. It is known that there are a ﬁnite number of
equivalence classes of reﬂexive polytopes in all dimensions. Using their computer
program PALP, Kreuzer and Skarke [111] determined that there are 4319 classes
of 3dimensional reﬂexive polytopes and 473800776 classes of 4dimensional re
ﬂexive polytopes. Since these numbers grow so quickly, most more recent work
has focused on subclasses, for instance the polytopes giving smooth Fano toric
varieties. These polytopes have a deceptively simple description (Exercise 8.3.6).
There are 5 types of such polytopes in dimension 2 (Exercise 8.3.7), and Batyrev
[8] and Sato [155] have shown that there are 18 types in dimension 3 and 124 types
in dimension 4. See [136] for a classiﬁcation algorithm and further references.
Exercises for §8.3.
8.3.1. Verify the claims about the polygons P and P
◦
deﬁned in Example 8.3.2.
8.3.2. Prove the claim about P(q
0
, . . . , q
n
) made in Example 8.3.3. Hint: Exercise 4.2.11.
8.3.3. Prove Lemma 8.3.6. Hint: Assume that (a) and (b) do not hold. Show that for any
facet F of P, ¸u
F
, m) +¸u
F
, m
′
) >−2, and use this to conclude that (c) must hold.
8.3.4. In this exercise, you will supply the details for the proof of Theorem 8.3.7.
(a) Let m, m
′
∈ M ≃ Z
2
and assume that Conv(0, m, m
′
) has no lattice points other than
the vertices. Prove that m, m
′
form a basis of M.
§8.3. Fano Toric Varieties 377
(b) Show that if P is a reﬂexive triangle with vertices at e
1
, e
2
, then the third vertex must
be −e
1
−e
2
. Hint: One way to show this succinctly is to use the projections from the
vertices e
1
and e
2
as in Lemma 8.3.5.
(c) In the case that each facet contains exactly two vertices, show that if there are vertices
m, m
′
, m
′′
with m+m
′
=m
′′
, then the pairs m, m
′′
and m
′
, m
′′
must lie in edges.
(d) Complete the proof that the polygons of types 4b, 5a, 6a are the only possibilities in
Subcase A.2. Hint: Show that Conv(±e
2
, ±(e
2
−e
1
), e
1
) is lattice equivalent to 5a.
(e) Show that every reﬂexive polygon in Case B is equivalent to one of type 4c, 5b, 6b,
6c, 6d, 7a, 7b, 8a, 8b, or 8c.
(f) Prove the claim made in Subcase C.3.
8.3.5. Consider the 16 reﬂexive polygons in Figure 2. Since each polygon in the ﬁgure is
reﬂexive, its dual must also appear in the ﬁgure, up to lattice equivalence.
(a) For each polygon in the Figure 2, determine its dual polygon and where the dual ﬁts
in the classiﬁcation. In some cases, you will need to ﬁnd an isomorphism of M that
takes the dual to a polygon in the ﬁgure. Also, some of the polygon are selfdual, up
to lattice equivalence.
(b) Show that the relation (8.3.3) holds for each polar pair.
8.3.6. A lattice polytope is called Fano if the origin is an interior point and the vertices of
every facet form a basis of the lattice. Let Q ⊆N
R
be a Fano polytope.
(a) Prove that Q is reﬂexive.
(b) Part (a) implies that Q
◦
⊆ M
R
is a lattice polytope. Prove that X
Q
◦ is a smooth Fano
toric variety whose normal fan is formed by taking cones over proper faces of Q.
(c) Prove that every smooth Fano toric variety arises this way.
8.3.7. Of the 16 Gorenstein Fano varieties classiﬁed in Theorem 8.3.7, exactly ﬁve ar
smooth. Find them. Hint: See Exercise 8.3.8.
8.3.8. Some of the polygons in Figure 2 give wellknown toric varieties. For example,
type 9 gives P
2
and type 8a gives P
1
P
1
. This follows easily by computing the normal
fan. Here you will describe the toric surfaces coming from some of the other polygons in
Figure 2. This exercise is based on [39, Rem. 6.11].
(a) Show that type 8b corresponds to the blowup of P
2
at one of the torus ﬁxed points.
Also show that this is the Hirzebruch surface H
1
. Hint: Remember the description of
blowing up given in §2.3.
(b) Similarly show that type 7a (resp. 6a) corresponds to the blowup of P
2
at two (resp.
three) torus ﬁxed points.
(c) Showthat type 3 corresponds to a quotient P
2
/(Z/3Z) and is isomorphic to the surface
in P
3
deﬁned by w
3
= xyz. Hint: Compute the normal fan and show that its minimal
generators span a sublattice of index 3 in Z
2
. Proposition 3.3.7 will be helpful. For
the ﬁnal assertion, use the characters coming from the lattice points of the polygon.
(d) Show that type 8c gives the weighted projective space P(1, 1, 2) and type 4c gives the
quotient P(1, 1, 2)/(Z/2Z). Hint: See Example 3.1.17.
(e) Show that type 4a gives the quotient P
1
P
1
/(Z/2Z).
(f) Show that type 6d gives the weighted projective space P(1, 2, 3).
378 Chapter 8. The Canonical Divisor of a Toric Variety
(g) Show that type 7b (resp. 6b) corresponds to the blowup of P(1, 1, 2) at one (resp. two)
smooth torus ﬁxed points.
(h) Showthat type 5b gives the blowup of P(1, 2, 3) at its unique smooth torus ﬁxed point.
8.3.9. In this exercise you will consider a toric surface that is not quite Fano. The lattice
polygon P = Conv(±3e
1
, ±3e
2
, ±2e
1
±2e
2
) ⊆ M
R
= R
2
gives a toric surface X
P
. Let
D =
ρ
D
ρ
be the anticanonical divisor of X
P
.
(a) Prove that the ample Cartier divisor D
P
associated to P is given by D
P
=6D. Conclude
that D is not Cartier and that 6D is the smallest integer multiple of D that is Cartier.
(b) Show that the normal fan of P has minimal generators ±e
1
±2e
2
, ±2e
1
±e
2
and that
the minimal generators are the vertices of Q =Conv(±e
1
±2e
2
, ±2e
1
±e
2
) ⊆N
R
.
(c) Show that the dual of Q is Q
◦
=
1
6
P and conclude that 6 is the smallest integer multiple
of Q
◦
that is a lattice polytope.
8.3.10. A complete toric surface X
Σ
is log del Pezzo if some integer multiple of its anti
canonical divisor −K
XΣ
is an ample Cartier divisor. The index of X
Σ
is the smallest positive
integer ℓ such that such that −ℓK
XΣ
is Cartier. This exercise and the next will consider this
interesting class of toric surfaces.
(a) Prove that a complete toric surface is log del Pezzo of index 1 if and only if it is
Gorenstein Fano.
(b) Prove that the toric surface of Exercise 8.3.9 is log del Pezzo of index 6.
8.3.11. A lattice polygon Q ⊆ N
R
is called LDP if the origin is an interior point of Q and
the vertices of Q are primitive vectors in N. The dual Q
◦
⊆M
R
of a LDP polygon contains
the origin as an interior point but may fail to be a lattice polygon. The index of Q is the
smallest positive integer ℓ such that such that ℓQ
◦
is a lattice polygon. This exercise will
explore the relation betweek LDP polygons and toric log del Pezzo surfaces.
(a) Show that the polygon Q of Exercise 8.3.9 is LDP of index 6.
(b) A LDP polygon Q ⊆ N
R
gives a fan Σ in N
R
by taking the cones over the faces of Q.
Show that the minimal generators of Σ are the vertices of Q and that the toric surface
X
Σ
is log del Pezzo of index equal to the index of Q.
(c) Conversely, let X
Σ
be a toric log del Pezzo surface and let Q be the convex hull of
the minimal generators of Σ. Note that Q is LDP and that Σ is the fan obtained by
taking the cones over the faces of Q. Hint: The key point is to show that every minimal
generator of Σ is a vertex of Q. This can be proved using the strict convexity of the
support function of −ℓK
XΣ
.
This exercise shows that classifying toric log del Pezzo surfaces up to isomorphism is
equivalent to classifying LDP polygons up to lattice equivalence. This is an active area of
research—see [101].
8.3.12. Let X
Σ
be a smooth projective toric variety. As in Deﬁnition 6.3.9, a primitive
colection P =¦ρ
1
, . . . , ρ
k
¦ ⊆Σ(1) gives a primitive relation
u
ρ1
+ +u
ρ
k
=
ρ∈γ(1)
c
ρ
u
ρ
, c
ρ
∈ Z
>0
,
where γ ∈ Σ is the minimal cone containing
k
i=1
u
ρi
. Following [8], the degree of P is
∆(P) = k −
ρ∈γ(1)
c
ρ
. Prove that X
Σ
is Fano if and only if ∆(P) > 0 for all primitive
collections P. This is [8, Prop. 2.3.6]. Hint: Use (6.3.10) and Theorem 6.3.8.
Chapter 9
Sheaf Cohomology of
Toric Varieties
§9.0. Background: Cohomology
By Proposition 6.0.8, a short exact sequence of sheaves on X
(9.0.1) 0 →F −→G −→H −→0
gives rise to an exact sequence of global sections
(9.0.2) 0 −→Γ(X, F) −→Γ(X, G) −→Γ(X, H ).
The failure of Γ(X, G) →Γ(X, H ) to be surjective is measured by a sheaf coho
mology group. The main goal of this chapter is to understand sheaf cohomology
on a toric variety.
Sheaves and Cohomology. Asheaf F on a variety X has sheaf cohomology groups
H
p
(X, F). The abstract deﬁnition uses an exact sequence of sheaves
0 −→F −→I
0
d
0
−−→I
1
d
1
−−→I
2
d
2
−−→ ,
where I
0
, I
1
, . . . are injective. This term is from homological algebra: a sheaf I
is injective if given a sheaf homomorphismH
α
−→I and an injection β : H →G,
there exists a sheaf homomorphism θ making the diagram below commute:
I
0
//
H
α
OO
β
//
G.
θ
aa
379
380 Chapter 9. Sheaf Cohomology of Toric Varieties
We say that I
•
is an injective resolution of F. From I
•
we get the complex of
global sections
Γ(X, I
•
) : Γ(X, I
0
)
d
0
−−→Γ(X, I
1
)
d
1
−−→Γ(X, I
2
)
d
2
−−→ .
The term complex refers to the fact that d
p+1
◦d
p
= 0 for all p ≥ 0. Then the pth
sheaf cohomology group of F is deﬁned to be
(9.0.3) H
p
(X, F) =H
p
(Γ(X, I
•
)) =ker(d
p
)/im(d
p−1
),
where for p =0, we deﬁne d
−1
to be the zero map 0 →Γ(X, I
0
). One can prove
that injective resolutions always exist and that two different injective resolutions of
F give the same sheaf cohomology groups.
This deﬁnition of H
p
(X, F) has some very nice properties, including:
• H
0
(X, F) =Γ(X, F).
• A sheaf homomorphism F →G induces a homomorphism of cohomology
groups H
i
(X, F) →H
i
(X, G) that is compatible with composition and takes
the identity to the identity, i.e., F →H
i
(X, F) is a functor.
• A short exact sequence of sheaves (9.0.1) gives a long exact sequence
0 −→H
0
(X, F) −→H
0
(X, G) −→H
0
(X, H )
∂
0
−→
H
1
(X, F) −→H
1
(X, G) −→H
1
(X, H)
∂
1
−→
∂
p−1
−→
H
p
(X, F) −→H
p
(X, G) −→H
p
(X, H)
∂p
−→ .
We call ∂
p
: H
p
(X, H ) →H
p+1
(X, F) a connecting homomorphism.
You will prove the ﬁrst bullet in Exercise 9.0.1. For the second bullet, the key step
is to show that given a sheaf homomorphism α : F →G and injective resolutions
F →A
•
, G →B
•
, there are sheaf homomorphisms α
p
: A
p
→B
p
such that the
diagram
F
//
α
A
0
d
0
//
α
0
A
1
d
1
//
α
1
G
//
A
0
d
0
//
A
1
d
1
//
commutes. We say that α
•
: A
•
→B
•
is a map of complexes. Then α
p
induces the
desired map H
p
(X, F) →H
p
(X, G). Finally, for the last bullet, an exact sequence
(9.0.1) lifts to an exact sequence of injective resolutions
0
//
A
• //
B
• //
C
• //
0
0
//
F
//
OO
G
//
OO
H
OO
//
0,
and one can show that taking global sections gives an exact sequence of complexes
(9.0.4) 0 −→Γ(X, A
•
) −→Γ(X, B
•
) −→Γ(X, C
•
) −→0.
§9.0. Background: Cohomology 381
It is a general fact in homological algebra that any exact sequence of complexes
0 −→A
•
−→B
•
−→C
•
−→0
gives a long exact sequence
(9.0.5)
0 −→H
0
(A
•
) −→H
0
(B
•
) −→H
0
(C
•
)
∂
0
−→
H
1
(A
•
) −→H
1
(B
•
) −→H
1
(C
•
)
∂
1
−→
∂
p−1
−→
H
p
(A
•
) −→H
p
(B
•
) −→H
p
(C
•
)
∂p
−→ .
Applied to (9.0.4), we get the desired long exact sequence in sheaf cohomology.
In the language of homological algebra, Γ is left exact since (9.0.2) is exact.
Then the sheaf cohomology groups are the derived functors of Γ. The texts [72,
78, 175] discuss sheaf cohomology and homological algebra in more detail. We
especially recommend Appendix 3 of [48] and Chapters 2 and 3 of [97].
ˇ
Cech Cohomology. While the abstract deﬁnition of sheaf cohomology has nice
properties, it is not useful for explicit computations. Fortunately, there is a down
toearth way of viewing sheaf cohomology, in terms of the
ˇ
Cech complex, which
we now describe.
Let U = ¦U
i
¦
ℓ
i=1
be an open cover of X. The deﬁnition of a sheaf F on X
shows that H
0
(X, F) =Γ(X, F) is the kernel of the map
(9.0.6)
1≤i≤ℓ
F(U
i
)
d
0
−→
1≤i<j≤ℓ
F(U
i
∩U
j
),
where d
0
is deﬁned as follows: if α=(α
i
) ∈
1≤i≤ℓ
F(U
i
), then the component of
d
0
(α) in F(U
i
∩U
j
) for i < j is given by α
j [
U
i
∩U
j
−α
i[
U
i
∩U
j
. Here, α
j
→α
j[
U
i
∩U
j
is the restriction map F(U
j
) →F(U
i
∩U
j
), and similarly for α
i
→α
i[
U
i
∩U
j
.
To get “higher” information about how sections of F ﬁt together, we extend
(9.0.6) to the
ˇ
Cech complex. We will use the following notation. Let [ℓ] =¦1, . . . , ℓ¦
be the index set for the open cover and let [ℓ]
p
denote the set of all (p+1)tuples
(i
0
, . . . , i
p
) of elements of I satisfying i
0
< < i
p
.
Deﬁnition 9.0.1. The group of pth
ˇ
Cech cochains is
ˇ
C
p
(U , F) =
(i
0
,...,ip)∈[ℓ]p
F(U
i
0
∩ ∩U
ip
).
One can think of an element of
ˇ
C
p
(U , F) is as a function α that assigns an el
ement of F(U
i
0
∩ ∩U
ip
) to each (i
0
, . . . , i
p
) ∈ [ℓ]
p
. Then we deﬁne a differential
ˇ
C
p
(U , F)
d
p
−→
ˇ
C
p+1
(U , F)
382 Chapter 9. Sheaf Cohomology of Toric Varieties
by describing how d
p
(α) operates on elements of [ℓ]
p+1
:
d
p
(α)(i
0
, . . . , i
p+1
) =
p+1
k=0
(−1)
k
α(i
0
, . . . ,
´
i
k
, . . . , i
p+1
)[
U
i
0
∩∩U
i
p+1
.
As above, α(i
0
, . . . ,
´
i
k
, . . . , i
p+1
)[
U
i
0
∩∩U
i
p+1
is obtained by applying the restriction
map
F(U
i
0
∩ ∩
´
U
i
k
∩ ∩U
i
p+1
) −→F(U
i
0
∩ ∩U
i
p+1
)
to α(i
0
, . . . ,
´
i
k
, . . . , i
p+1
). In Exercise 9.0.2 you will verify that d
p
◦d
p−1
=0.
Deﬁnition 9.0.2. Given a sheaf F on X and an open cover U =¦U
i
¦
i∈I
of X, the
ˇ
Cech complex is
ˇ
C
•
(U , F) : 0 −→
ˇ
C
0
(U , F)
d
0
−→
ˇ
C
1
(U , F)
d
1
−→
ˇ
C
2
(U , F)
d
2
−→ ,
and the pth
ˇ
Cech cohomology group is
ˇ
H
p
(U , F) =H
p
(
ˇ
C
•
(U , F)) =ker(d
p
)/im(d
p−1
).
Notice that
ˇ
H
0
(U , F) =H
0
(X, F) =Γ(X, F) since F is a sheaf. However,
ˇ
H
p
(U , F) need not equal H
p
(X, F) for p > 0. We will soon see that there is a
nice case where equality occurs for all p.
Cohomology of a Quasicoherent Sheaf . To compute the cohomology of a qua
sicoherent sheaf F on a variety X using
ˇ
Cech cohomology, we need to ﬁnd open
covers U of X such that
ˇ
H
p
(U , F) equals H
p
(X, F) for all p. The following
vanishing theorem of Serre is very useful in this regard. Proofs can be found in
[72, 78, 175].
Theorem 9.0.3 (Serre Vanishing for Afﬁne Varieties). Let F be a quasicoherent
sheaf on an afﬁne variety U. Then H
p
(U, F) =0 for all p > 0.
By Theorem 9.0.3, we can compute the cohomology of a quasicoherent sheaf
F using any afﬁne open cover. Here is the rough intuition:
• Consider an arbitrary open cover U =¦U
i
¦ of X. Since we can construct X by
“gluing together” the U
i
, the cohomology of X should be obtained from the co
homologies of the U
i
and their various intersections. In other words, H
•
(X, F)
should be determined by the cohomology groups H
•
(U
i
0
∩ ∩U
i
p+1
, F) as we
vary over all p.
• Now suppose that U = ¦U
i
¦ is an afﬁne open cover. Then U
i
0
∩ ∩U
i
p+1
is
afﬁne and hence has vanishing higher cohomology by Serre’s theorem. So all
that is left is H
0
(U
i
0
∩ ∩U
i
p+1
, F). This gives the
ˇ
Cech complex, which
thus computes the sheaf cohomology of F.
This suggests the following result.
§9.0. Background: Cohomology 383
Theorem 9.0.4. Let U = ¦U
i
¦ be an afﬁne open cover of a variety X and let F
be a quasicoherent sheaf on X. Then there are natural isomorphisms
ˇ
H
p
(U , F) ≃H
p
(X, F)
for all p ≥0.
The proof will be given later in the section after we introduce a spectral se
quence that makes the above intuition rigorous. A more elementary proof that does
not use spectral sequences can be found in [78, Theorem III.4.5].
Here is an application of Theorem 9.0.4.
Example 9.0.5. We compute the cohomology of O
P
1 , O
P
1 (−1), and Ω
1
P
1
on P
1
.
Consider the afﬁne open cover U =¦U
0
,U
1
¦ of P
1
, where
U
0
=Spec(C[x]) and U
1
=Spec(C[x
−1
]).
Note also that
U
0
∩U
1
=Spec(C[x, x
−1
]).
For any sheaf F on P
1
, the
ˇ
Cech complex is
F(U
0
) ⊕F(U
1
)
d
0
−→F(U
0
∩U
1
).
It follows that H
p
(P
1
, F) =0 for p ≥2. Hence we need only consider H
0
(P
1
, F)
and H
1
(P
1
, F). For simplicity, we write these sheaf cohomology groups as H
0
(F)
and H
1
(F) respectively.
For F =O
P
1, the
ˇ
Cech complex becomes
(9.0.7) C[x] ⊕C[x
−1
]
d
0
−→C[x, x
−1
]
where d
0
( f (x), g(x
−1
)) = f (x) −g(x
−1
). Then
H
0
(O
P
1 ) =ker(d
0
) =C
H
1
(O
P
1 ) =coker(d
0
) =0.
The assertion for H
0
is clear since f (x) −g(x
−1
) = 0 implies that f and g are the
same constant, and the assertion for H
1
follows since an element of C[x, x
−1
] can
be written
a
−m
x
−m
+ +a
−1
x
−1
. ¸¸ .
−g(x
−1
)
+a
0
+a
1
x + +a
n
x
n
. ¸¸ .
f (x)
.
We can represent O
P
1 (−1) as the line bundle O
P
1 (−D), where the divisor D is
one of the ﬁxed points of the torus action on P
1
. It follows that we have an exact
sequence of sheaves
0 −→O
P
1(−1) −→O
P
1 −→O
D
−→0.
The long exact sequence in sheaf cohomology gives
0 →H
0
(O
P
1 (−1)) →H
0
(O
P
1 ) →H
0
(O
D
) →H
1
(O
P
1 (−1)) →H
1
(O
P
1 ) → .
384 Chapter 9. Sheaf Cohomology of Toric Varieties
The map H
0
(O
P
1 ) →H
0
(O
D
) is the isomorphism that sends a constant function of
P
1
to the same constant function on D. Since H
1
(O
P
1 ) =0, the long exact sequence
implies that
(9.0.8) H
0
(O
P
1 (−1)) =H
1
(O
P
1 (−1)) =0.
Finally, for Ω
1
P
1
, we use the Euler sequence
0 −→Ω
1
P
1
−→O
P
1(−1) ⊕O
P
1 (−1) −→O
P
1 −→0
from (8.1.8). Since H
p
(X, F ⊕G) ≃ H
p
(X, F) ⊕H
p
(X, G) (Exercise 9.0.3), the
vanishing (9.0.8) and the long exact sequence for cohomology imply that
H
0
(Ω
1
P
1
) =0
H
1
(Ω
1
P
1
) ≃H
0
(O
P
1 ) =C.
Earlier, in Example 6.0.5, we showed that the surjective sheaf homomorphism
O
P
1 (−1) ⊕O
P
1 (−1) →O
P
1
is not surjective on global sections. This is what forces H
1
(Ω
1
P
1
) to be nonzero. ♦
Akey part of Example 9.0.5 was the surjectivity of (9.0.7). This is the algebraic
analog of a Cousin problem in complex analysis. A discussion of Cousin problems
and their relation to sheaves and cohomology can be found in [109, Ch. 13].
Serre Vanishing for Projective Varieties. The Serre vanishing theorem for afﬁne
varieties (Theorem 9.0.3) has a projective version.
Theorem 9.0.6 (Serre Vanishing for Projective Varieties). Let L be an ample line
bundle on a projective variety X. Then for any coherent sheaf F on X, we have
H
p
(X, F ⊗
X
L
⊗ℓ
) =0
for all p > 0 and ℓ ≫0.
A proof can be found in [78, Prop. II.5.3]. We will use this result later in the
chapter to prove some interesting vanishing theorems for toric varieties.
Higher Direct Images. Given a morphism f : X →Y of varieties and a sheaf F
of O
X
modules on X, the direct image is the sheaf f
∗
F on Y deﬁned by
U −→F( f
−1
(U))
for U ⊆Y open. We noted in Example 4.0.24 that f
∗
F is a sheaf of O
Y
modules.
The deﬁnition of f
∗
F implies in particular that
H
0
(Y, f
∗
F) =H
0
(X, F)
since f
−1
(Y) =X. More generally, there are homomorphisms
(9.0.9) H
p
(Y, f
∗
F) −→H
p
(X, F),
§9.0. Background: Cohomology 385
which need not be isomorphisms for p > 0.
In this situation, we also get the higher direct image R
p
f
∗
F, which is the sheaf
on Y associated to the presheaf deﬁned by
U −→H
p
( f
−1
(U), F).
Proposition 9.0.7. Let f : X →Y be a morphism and F be a quasicoherent sheaf
on X. Then:
(a) The higher direct images R
p
f
∗
F are quasicoherent sheaves on Y.
(b) If U ⊆Y is afﬁne, then R
p
f
∗
F[
U
is the sheaf associated to the O
Y
(U)module
H
p
( f
−1
(U), F).
A proof can be found in [78, Propositions II.5.8 and III.8.5]. One especially
nice case is when the higher direct images R
p
f
∗
F vanish for p > 0. We will
see below that the maps (9.0.9) are isomorphisms when this happens. The proof
involves our next topic, spectral sequences.
Spectral Sequences. Readers not familar with spectral sequences should glance
at the appendix to this chapter before proceeding further. Here we discuss two
spectral sequences relevant to this section.
First suppose that f : X →Y is a morphism and F is a quasicoherent sheaf
on X. As above, we get the higher direct images R
p
f
∗
F, which are sheaves on
Y. Proposition 9.0.7 shows that these sheaves compute the cohomology of F
over certain open subsets of X. So when we “put these together,” i.e., compute
H
p
(Y, R
q
f
∗
F), we should get the cohomology of F on all of X. The precise form
of this intuition is the Leray spectral sequence:
E
p,q
2
=H
p
(Y, R
q
f
∗
F) ⇒H
p+q
(X, F).
Furthermore, the map H
p
(Y, f
∗
F) →H
p
(X, F) from (9.0.9) is the edge homo
morphism E
p,0
2
→H
p
(X, F) (see Deﬁnition 9.A.4). This spectral sequence is dis
cussed in [64, II.4.17] and [72, p. 463]
Now assume R
q
f
∗
F =0 for q > 0. Then
E
p,q
2
=
_
H
p
(Y, f
∗
F) q =0
0 q > 0.
Then Proposition 9.A.5 implies that (9.0.9) is an isomorphism. Thus we have
proved the following.
Proposition 9.0.8. Suppose f : X →Y is a morphism and F is a quasicoherent
sheaf on X such that R
q
f
∗
F =0 for q >0. Then the map (9.0.9) is an isomorphism
H
p
(Y, f
∗
F) ≃H
p
(X, F).
For our second spectral sequence, let U =¦U
i
¦ be an open cover of X and F
be a sheaf on X. In the discussion leading up to Theorem 9.0.4, we asserted that
386 Chapter 9. Sheaf Cohomology of Toric Varieties
H
•
(X, F) is determined by H
•
(U
i
0
∩ ∩U
i
p+1
, F) as we vary over all p. The
precise meaning is given by the E
1
spectral sequence
(9.0.10) E
p,q
1
=
(i
0
,...,ip)∈[ℓ]p
H
q
(U
i
0
∩ ∩U
ip
, F) ⇒H
p+q
(X, F),
where the differential d
p,q
1
: E
p,q
1
→E
p+1,q
1
is induced by inclusion, with signs sim
ilar to the differential in the
ˇ
Cech complex. This spectral sequence is constructed
in [64, II.5.4].
We now have the tools needed to prove Theorem 9.0.4.
Proof of Theorem 9.0.4. We are assuming that U =¦U
i
¦ is an afﬁne open cover
of X. First observe that the q =0 terms of (9.0.10) are given by
E
p,0
1
=
(i
0
,...,ip)∈[ℓ]p
H
0
(U
i
0
∩ ∩U
ip
, F) =
ˇ
C
p
(U , F),
and the differentials d
p,0
1
are the differentials in the
ˇ
Cech complex. Hence
E
p,0
2
=ker(d
1
: E
p,0
1
→E
p+1,0
1
)
_
im(d
1
: E
p−1,0
1
→E
p,0
1
)
=H
p
(
ˇ
C
•
(U , F)) =
ˇ
H
p
(U , F).
Since F is quasicoherent and the intersections U
i
0
∩ ∩U
ip
are afﬁne for all p ≥0,
it follows from Theorem 9.0.3 that E
p,q
1
=0 for q >0. This implies that E
p,q
2
=0 for
q > 0. Using Proposition 9.A.5 again, we conclude that the edge homomorphism
E
p,0
2
=
ˇ
H
p
(U , F) −→H
p
(X, F)
is an isomorphism for all p ≥0.
CohenMacaulay Varieties. We next discuss a class of varieties that play a crucial
role in duality theory and are interesting in their own right.
We ﬁrst deﬁne what it means for a local ring (R, m) to be CohenMacaulay.
Elements f
1
, . . . , f
s
∈ m form a regular sequence if f
i
is not a zero divisor in
R/¸ f
1
, . . . , f
i−1
) for all i, and the depth of R is the maximal length of a regular
sequence. Then R is CohenMacaulay if its depth equals its dimension. Examples
include regular local rings. A nice discussion of what CohenMacaulay means can
be found in [157, 10.2]. See also [108, pp. 153–155].
A variety X is CohenMacaulay if its local rings (O
X,p
, m
X,p
) are Cohen
Macaulay for all p ∈ X. Thus smooth varieties are CohenMacaulay. Later in
the chapter we will prove that normal toric varieties are CohenMacaulay.
Serre Duality. CohenMacaulay varieties provide the natural setting for a basic
duality theorem of Serre. Here is the simplest version.
§9.0. Background: Cohomology 387
Theorem 9.0.9 (Serre Duality I). Let ω
X
be the canonical sheaf of a complete
normal CohenMacaulay variety X of dimension n. Then for every locally free
sheaf F of ﬁnite rank on X, there are natural isomorphisms
H
p
(X, F)
∨
≃H
n−p
(X, ω
X
⊗
O
X
F
∨
).
In particular, when D is a Cartier divisor on X and K
X
is a canonical divisor, we
have isomorphisms
H
p
(X, O
X
(D))
∨
≃H
n−p
(X, O
X
(K
X
−D)).
A proof for the projective case can be found in [78, Thm. III.7.6]. The assertion
for divisors follows from
ω
X
⊗
O
X
O
X
(D)
∨
≃O
X
(K
X
) ⊗
O
X
O
X
(−D) ≃O
X
(K
X
−D),
where the last isomorphism holds since D is Cartier.
There is also a more general version of Serre duality that applies when F is
coherent but not necessarily locally free. The cohomology groups H
p
(X
Σ
, F) are
the derived functors of the global section functor F →Γ(X
Σ
, F), and in the same
way, the Ext groups Ext
p
O
X
Σ
(G, F) are the derived functors of the Hom functor
F →Hom
O
X
Σ
(G, F) for ﬁxed G. Then we have the following result.
Theorem 9.0.10 (Serre Duality II). Let ω
X
be the canonical sheaf of a complete
normal CohenMacaulay variety X of dimension n. Then for every coherent sheaf
F on X, there are natural isomorphisms
H
p
(X, F)
∨
≃Ext
n−p
O
X
(F, ω
X
).
When X is projective, this is proved in [78, Thm. III.7.6]. We will give a
version of Serre duality especially adapted to the toric case in §9.2.
When X fails to be CohenMacaulay, there is a more general duality theorem
where canonical sheaf ω
X
is replaced with the dualizing complex ω
•
X
. A discussion
of this version of duality can be found in [137, §3.2].
Singular Cohomology. Our discussion of the sheaf cohomology of a toric variety
in §9.1 will use some algebraic topology. Here we review the topological invariants
we will need, beginning with the singular cohomology groups
H
p
(Z, R),
where Z is a topological space and R is a commutative ring, usually Z, R or C.
These are deﬁned using continuous maps
γ : ∆
n
−→Z,
where ∆
n
is the standard nsimplex. There are several good introductions to sin
gular cohomology, including [71], [81] and [131].
Here are some important properties of singular cohomology:
388 Chapter 9. Sheaf Cohomology of Toric Varieties
• A continuous map f : Z →W induces f
∗
: H
p
(W, R) →H
p
(Z, R) such that
homotopic maps induce the same map on cohomology and the identity map
induces the identity on cohomology.
• If i : A ֒→Z is a deformation retract (i.e., there is a continuous map r : Z →A
such that r ◦i =1
A
and i ◦r is homotopic to 1
Z
), then i
∗
: H
p
(Z, R) →H
p
(A, R)
is an isomorphism.
• If Z is contractible (i.e, there is z ∈ Z such that ¦z¦ ֒→ Z is a deformation
retract), then
H
p
(Z, R) =
_
R p =0
0 otherwise.
• For n > 1, the singular cohomology of the (n−1)sphere S
n−1
⊆R
n
is
H
p
(S
n−1
, R) =
_
R p =0, n−1
0 otherwise.
We will always assume that Z is a locally contractible metric space. This allows
us to interpret the singular cohomology of Z in terms of sheaf cohomology as
follows. The ring R deﬁnes a presheaf on Z where R is the group of sections over
every nonempty open U ⊂ Z. The corresponding sheaf is the constant sheaf of R.
By [22, III.1], the sheaf cohomology of the constant sheaf of R on Z is the singular
cohomology H
p
(Z, R).
The Cohomology Ring. For a commutative ring R, H
•
(Z, R) is a graded Ralgebra
with multiplication given by cup product. A continuous map induces a ring homo
morphism on cohomology, so in particular, a deformation retract i : A ֒→Z induces
a ring isomorphism i
∗
: H
•
(Z, R) ≃H
•
(A, R).
Here are some examples.
Example 9.0.11. The real torus (S
1
)
n
has cohomology ring
H
•
((S
1
)
n
, R) =R[α
1
, . . . , α
n
]
where the α
i
lie in H
1
((S
1
)
n
, R) and satisfy the relations α
i
α
j
=−α
j
α
i
and α
2
i
=0
(see Examples 3.11 and 3.15 of [81]). Thus
H
•
((S
1
)
n
, R) ≃
_
•
R
n
,
i.e., the cohomology ring is the exterior algebra of a free Rmodule. ♦
Example 9.0.12. The torus (C
∗
)
n
contains the real torus (S
1
)
n
as a deformation
retract via (t
1
, . . . , t
n
) →(t
1
/[t
1
[, . . . , t
n
/[t
n
[). Hence
H
•
((C
∗
)
n
, R) ≃H
•
((S
1
)
n
, R) ≃
_
•
R
n
.
More canonically, the torus T
N
=N⊗C
∗
has cohomology ring
H
•
(T
N
, R) ≃
_
•
M⊗
Z
R,
§9.0. Background: Cohomology 389
where M is the dual of N. Thus a lattice homomorphism N →N
′
gives a map
T
N
→T
N
′ of tori, and the induced map H
•
(T
N
′ , R) →H
•
(T
N
, R) is the map
_
•
M
′
⊗
Z
R −→
_
•
M⊗
Z
R
determined by the dual homomorphism M
′
→M. ♦
Example 9.0.13. The cohomology ring of P
n
over Q is
H
•
(P
n
, Q) ≃Q[α]/¸α
n+1
),
where α ∈ H
2
(P
n
, Q) (see [81, Theorem 3.12]). Later in the book we will give
a similar quotient description of the cohomology ring of any complete simplicial
toric variety. ♦
Reduced Cohomology. Given a topological space Z, we have the canonical map
Z →¦pt¦ that sends elements of Z to pt. This induces the map
R =H
0
(¦pt¦, R) −→H
0
(Z, R)
whose cokernel is denoted
¯
H
0
(Z, R). Thus, assuming R ,=0,
¯
H
0
(Z, R) =0 ⇐⇒ Z is path connected.
The reduced cohomology of Z with coefﬁcients in R is deﬁned for p ≥0 by
¯
H
p
(Z, R) =
_
¯
H
0
(Z, R) p =0
H
p
(Z, R) p > 0,
and for p =−1 by
¯
H
−1
(Z, R) =
_
0 Z ,=∅
R Z =∅.
The deﬁnition of
¯
H
−1
may look strange but means that the sequence
(9.0.11) 0 −→
¯
H
−1
(Z, R) −→R −→H
0
(Z, R) −→
¯
H
0
(Z, R) −→0
is exact for all Z, including Z = ∅ (Exercise 9.0.4). We will use
¯
H
−1
and (9.0.11)
in §9.1 when we compute sheaf cohomology on a toric variety.
Exercises for §9.0.
9.0.1. Use (9.0.3) to show that H
0
(X, F) =Γ(X, F).
9.0.2. Check that the map d
p
deﬁned on the
ˇ
Cech cochains satisﬁes d
p
◦ d
p−1
=0.
9.0.3. Let F, G be sheaves on X. Prove H
p
(X, F ⊕G) ≃H
p
(X, F) ⊕H
p
(X, G).
9.0.4. Prove the exactness of (9.0.11).
9.0.5. A morphism f : X →Y is afﬁne if Y has an afﬁne open cover ¦U
i
¦ such that f
−1
(U
i
)
is afﬁne for all i. Nowassume that f : X →Y is afﬁne and let F be a quasicoherent sheaf on
X. Use Theorem 9.0.3, Proposition 9.0.7 and Theorem 9.0.8 to prove that H
p
(Y, f
∗
F) ≃
H
p
(X, F) for all p ≥0.
390 Chapter 9. Sheaf Cohomology of Toric Varieties
9.0.6. Use Exercise 9.0.5 to prove the following isomorphisms in cohomology:
(a) H
p
(X, i
∗
F) ≃H
p
(Y, F) when i : Y ֒→X is closed in X and F is quasicoherent.
(b) H
p
(X, π
∗
F) ≃H
p
(V, F) when π : V →X is a vector bundle and F is quasicoherent.
9.0.7. Let X be a variety that is a union of afﬁne open subsets X =U
1
∪ ∪U
s
.
(a) Let F be quasicoherent on X. Prove that H
p
(X, F) =0 for p > s −1.
(b) Let F be quasicoherent on P
n
. Prove that H
p
(P
n
, F) =0 for p > n.
9.0.8. Consider the (n −1)sphere S
n−1
⊆R
n
. Show that
¯
H
p
(S
n−1
, R) ≃
_
R p =n −1
0 otherwise.
Hint: Remember that S
0
consists of two points.
§9.1. Cohomology of Toric Divisors
Demazure [45] gave a concrete description of the sheaf cohomology of the sheaf
O
X
Σ
(D) of a torusinvariant Cartier divisor D on a toric variety X
Σ
. We generalize
this to torusinvariant QCartier divisors, inspired by papers of Eisenbud, Mustat ¸ˇ a
and Stillman [50], Hering, K¨ uronya and Payne [85], and Perling [143].
The Toric
ˇ
Cech Complex. When we compute sheaf cohomology using
ˇ
Cech co
homology, the obvious choice of open cover is
U =¦U
σ
¦
σ∈Σmax
,
where Σ
max
is the set of maximal cones in Σ. We write these as σ
i
and order them
according to their indices.
Given a torusinvariant Cartier divisor D=
ρ
a
ρ
D
ρ
on X
Σ
, the
ˇ
Cech complex
is given by
ˇ
C
p
(U , O
X
Σ
(D)) =
γ=(i
0
,...,ip)∈[ℓ]p
H
0
(U
σ
i
0
∩ ∩U
σ
ip
, O
X
Σ
(D)).
If we set σ
γ
=σ
i
0
∩ ∩σ
ip
∈ Σ for γ =(i
0
, . . . , i
p
) ∈ [ℓ]
p
, then we rewrite this as
(9.1.1)
ˇ
C
p
(U , O
X
Σ
(D)) =
γ∈[ℓ]p
H
0
(U
σγ
, O
X
Σ
(D)).
The Grading on Cohomology. For an afﬁne open subset U
σ
, Proposition 4.3.3
implies that the sections of O
X
Σ
(D) over U
σ
can be written
H
0
(U
σ
, O
X
Σ
(D)) =
m
C χ
m
,
where the direct sum is over all m ∈ M such that ¸m, u
ρ
) ≥ −a
ρ
for all ρ ∈ σ(1).
We can write this as
H
0
(U
σ
, O
X
Σ
(D)) =
m∈M
H
0
(U
σ
, O
X
Σ
(D))
m
,
§9.1. Cohomology of Toric Divisors 391
where for m ∈ M,
(9.1.2) H
0
(U
σ
, O
X
Σ
(D))
m
=
_
C χ
m
¸m, u
ρ
) ≥−a
ρ
for all ρ ∈ σ(1)
0 otherwise.
This induces a grading of the
ˇ
Cech complex (9.1.1). The
ˇ
Cech differential
is built from the restriction maps and hence respects the grading of the complex.
Since H
p
(X
Σ
, O
X
(D)) =
ˇ
H
p
(U , O
X
(D)), we obtain a natural decomposition of
sheaf cohomology
H
p
(X
Σ
, O
X
(D)) =
m∈M
H
p
(X
Σ
, O
X
(D))
m
.
When decomposing cohomology this way, we often refer to the m∈ M as weights.
Here is an example of how weights can be used to compute sheaf cohomology.
Example 9.1.1. On P
2
, label the rays as usual: u
0
= −e
1
−e
2
, u
1
= e
1
, u
2
= e
2
,
and maximal cones σ
i
, starting with σ
0
in the ﬁrst quadrant and going counter
clockwise. We will compute H
p
(P
2
, O
P
2 (a)) for a ∈ Z, where O
P
2(a) =O
P
2 (aD
0
)
for the divisor D
0
corresponding to u
0
.
Let U
i
be the afﬁne open corresponding to σ
i
. Then U
i j
is U
i
∩U
j
and U
012
is
the triple intersection. This allows us to write the
ˇ
Cech complex as
(9.1.3) 0 −→
ˇ
C
0
(a)
0
B
B
@
1 −1 0
1 0 −1
0 1 −1
1
C
C
A
−−−−−−−−−−−→
ˇ
C
1
(a)
“
1 −1 1
”
−−−−−−−−−→
ˇ
C
2
(a) −→0,
where
ˇ
C
0
(a) =
3
i=1
H
0
(U
i
, O
P
2 (aD
0
))
ˇ
C
1
(a) =
i<j
H
0
(U
i j
, O
P
2 (aD
0
))
ˇ
C
2
(a) =H
0
(U
012
, O
P
2 (aD
0
)).
We will compute the cohomology of this complex using the graded pieces for
m ∈ M = Z
2
. For this purpose, let l
0
, l
1
, l
2
be the lines in M
R
= R
2
deﬁned by
¸m, u
0
) = −a, ¸m, u
1
) = 0, ¸m, u
2
) = 0. These lines divide the plane into various
regions called chambers. To get a disjoint decomposition, we deﬁne
C
−++
=¦m ∈ M
R
[ ¸m, u
0
) <−a, ¸m, u
1
) ≥0, ¸m, u
2
) ≥0¦.
Note that a minus sign corresponds to strict inequality (< 0) while a plus sign
corresponds to a weak inequality (≥0). The other chambers C
+−+
, C
+−−
, etc. are
deﬁned similarly.
392 Chapter 9. Sheaf Cohomology of Toric Varieties
We ﬁrst consider the case when a < 0. The corresponding chamber decompo
sition of M
R
is shown in Figure 1. The labels l
i
are placed on the plus side of the
lines, i.e., where ¸m, u
0
) ≥ −a, ¸m, u
1
) ≥ 0, ¸m, u
2
) ≥ 0. Each chamber is labeled
with its sign pattern, and the shading indicates which chamber the points on the
lines belong to.
l
1
l
2
l
0
− + +
− − −
+ + −
+ − +
− − +
+ − − − + +
0
a
a
Figure 1. The chamber decomposition for a < 0
The cone σ
i
is generated by u
j
, u
k
, where ¦i, j, k¦ =¦0, 1, 2¦. Thus
(9.1.4)
H
0
(U
0
, O
P
2 (a))
m
,=0 ⇐⇒ m∈C
−++
∩M
H
0
(U
1
, O
P
2 (a))
m
,=0 ⇐⇒ m∈C
+−+
∩M
H
0
(U
2
, O
P
2 (a))
m
,=0 ⇐⇒ m∈C
++−
∩M.
This follows easily from (9.1.2) (Exercise 9.1.1). Hence we can determine when
ˇ
C
0
(a)
m
,=0.
The cone σ
1
∩σ
2
is generated by u
0
, so its dual is the halfplane of M
R
where
¸m, u
0
) ≥a, i.e., on the plus side of l
0
. Using Figure 1 and (9.1.2) again, we get
H
0
(U
12
, O
P
2 (a))
m
,=0 ⇐⇒ m ∈ (C
+−+
∪C
+−−
∪C
++−
) ∩M.
Similar results hold for U
01
and U
02
, so we know when
ˇ
C
1
(a)
m
,=0. Finally, since
U
012
is the torus of P
2
, we have
ˇ
C
2
(a)
m
=H
0
(U
012
, O
P
2 (a))
m
,=0 for all m ∈ M.
Putting everything together, we get Table 1 on the next page, which shows the
dimension of
ˇ
C
p
(a) for m ∈ M and a < 0. For example, dim
ˇ
C
1
(a)
m
= 2 when
m ∈ C
−++
since both U
01
and U
02
contribute 1dimensional subspaces. Be sure
you understand this.
§9.1. Cohomology of Toric Divisors 393
m∈ M is in dim
ˇ
C
0
(a)
m
dim
ˇ
C
1
(a)
m
dim
ˇ
C
2
(a)
m
C
−++
∪C
+−+
∪C
++−
1 2 1
C
−−+
∪C
−+−
∪C
+−−
0 1 1
C
−−−
0 0 1
Table 1. Dimension of
ˇ
C
p
(a)m for a < 0
When a < 0, it is now easy to understand the
ˇ
Cech complex
(9.1.5) 0 −→
ˇ
C
0
(a)
m
−→
ˇ
C
1
(a)
m
−→
ˇ
C
2
(a)
m
−→0.
The ﬁrst line of Table 1 corresponds to m∈C
−++
∪C
+−+
∪C
++−
. One can check
without difﬁculty that for these m’s, the complex (9.1.5) exact, and the same is true
for the second row as well.
These remarks imply that if m is a lattice point which is not in the interior of
the triangle
a∆
2
=Conv¦(0, 0), (a, 0), (0, a)¦,
then
H
p
(P
2
, O
P
2 (a))
m
=0 for all p.
However, if m is a lattice point in the interior of a∆
2
, then H
2
(P
2
, O
P
2 (a))
m
is
nonzero. Summing up, we have:
(9.1.6) h
p
(a) =
_
0 p ,=2
¸
¸
Int(a∆
2
) ∩M
¸
¸
=
_
−a−1
2
_
p =2
when a < 0. Here, h
p
(a) =dimH
p
(P
2
, O
P
2 (a)).
In Exercise 9.1.2, you will will adapt the above methods to show that
(9.1.7) h
p
(a) =
_
¸
¸
a∆
2
∩M
¸
¸
=
_
a+2
2
_
p =0
0 p > 0
when a ≥ 0. Thus, for any line bundle on P
2
, we have completely determined the
dimensions of the cohomology groups H
p
(P
2
, O
P
2 (a)). The values for −7 ≤a ≤4
are depicted in Table 2 on the next page. Note the symmetry in the table. In
Exercise 9.1.3 you will explore how this symmetry relates to Serre duality. ♦
Computing Cohomology. Given a divisor D=
ρ
a
ρ
D
ρ
on X
Σ
and m∈ M, deﬁne
two subsets of [Σ[:
• (When Σ is simplicial) V
simp
D,m
=
σ∈Σ
D,m
Conv(u
ρ
[ ρ ∈ σ(1)), where
Σ
D,m
=¦σ ∈ Σ [ ¸m, u
ρ
) <−a
ρ
for all ρ ∈ σ(1)¦.
• (When D is QCartier) V
supp
D,m
= ¦u ∈ [Σ[ [ ¸m, u) < ϕ
D
(u)¦, where ϕ
D
is the
support function of D (see Exercise 9.1.4).
Figure 2 in Example 9.1.3 below gives a picture of these sets.
394 Chapter 9. Sheaf Cohomology of Toric Varieties
a h
0
(a) h
1
(a) h
2
(a)
−7 0 0 15
−6 0 0 10
−5 0 0 6
−4 0 0 3
−3 0 0 1
−2 0 0 0
−1 0 0 0
0 1 0 0
1 3 0 0
2 6 0 0
3 10 0 0
4 15 0 0
Table 2. Sheaf cohomology of O
P
2 (a) on P
2
Theorem 9.1.2. Let D=
ρ
a
ρ
D
ρ
be a Weil divisor on X
Σ
. Fix m∈ M and p ≥0.
(a) If Σ is simplicial, then H
p
(X
Σ
, O
X
Σ
(D))
m
≃
¯
H
p−1
(V
simp
D,m
, C).
(b) If D is QCartier, then H
p
(X
Σ
, O
X
Σ
(D))
m
≃
¯
H
p−1
(V
supp
D,m
, C).
Proof. For part (a), ﬁrst note that if σ ∈ Σ, then ¦ρ ∈ σ(1) [ ¸m, u
ρ
) <−a
ρ
¦ deter
mines a (possibly trivial) face τ _σ since Σ is simplicial. It follows easily that
(9.1.8) V
simp
D,m
∩σ =Conv(u
ρ
[ ρ ∈ τ(1)).
Hence, for all σ ∈ Σ, we have
(9.1.9)
H
0
(U
σ
, O
X
Σ
(D))
m
,=¦0¦ ⇐⇒ ¸m, u
ρ
) ≥−a
ρ
for all ρ ∈ σ(1)
⇐⇒ V
simp
D,m
∩σ =∅,
where the ﬁrst equivalence uses (9.1.2) and the second follows from (9.1.8).
The equation (9.1.8) shows that V
simp
D,m
∩ σ is convex and hence connected
when it is nonempty. Thus H
0
(V
D,m
∩σ, C) = C when V
simp
D,m
∩ σ ,= ∅. Since
H
0
(U
σ
, O
X
Σ
(D))
m
= C χ
m
when it is nonzero, the equivalences (9.1.9) give a
canonical exact sequence
(9.1.10) 0 −→H
0
(U
σ
, O
X
Σ
(D))
m
−→C −→H
0
(V
simp
D,m
∩σ, C) −→0
for all σ ∈ Σ. It follows that for every p ≥0, we get an exact sequence
0 −→
γ∈[ℓ]p
H
0
(U
σγ
, O
X
Σ
(D))
m
−→
γ∈[ℓ]p
C −→
γ∈[ℓ]p
H
0
(V
simp
D,m
∩σ
γ
, C) −→0,
which we write as
0 −→
ˇ
C
p
(U , O
X
Σ
(D))
m
−→B
p
−→C
p
−→0.
§9.1. Cohomology of Toric Divisors 395
The formula for the differential in the
ˇ
Cech complex can be used to deﬁne differ
entials B
p
→B
p+1
and C
p
→C
p+1
. Then we get an exact sequence of complexes
0 −→
ˇ
C
•
(U , O
X
Σ
(D))
m
−→B
•
−→C
•
−→0.
Since
ˇ
H
p
(U , O
X
Σ
(D)) ≃ H
p
(X
Σ
, O
X
Σ
(D)), the long exact sequence from (9.0.5)
becomes
0 →H
1
(X
Σ
, O
X
Σ
(D))
m
→H
0
(B
•
) →H
0
(C
•
) →H
1
(X
Σ
, O
X
Σ
(D))
m
→ .
In Exercises 9.1.5 and 9.1.6 you will use the theory of Koszul complexes to show
that the complex B
•
has very simple cohomology:
(9.1.11) H
p
(B
•
) =
_
C p =0
0 p > 0.
Thus our long exact sequence breaks up into an exact sequence
0 →H
0
(X
Σ
, O
X
Σ
(D))
m
→C →H
0
(C
•
) →H
1
(X
Σ
, O
X
Σ
(D))
m
→0
and isomorphisms
H
p−1
(C
•
) ≃H
p
(X
Σ
, O
X
Σ
(D))
m
, p ≥2.
We will show below that
(9.1.12) H
p
(C
•
) ≃H
p
(V
simp
D,m
, C), p ≥0.
For p ≥2, this gives the desired isomorphism
¯
H
p−1
(V
simp
D,m
, C) =H
p−1
(V
simp
D,m
, C) ≃H
p
(X
Σ
, O
X
Σ
(D))
m
.
When p =1, we obtain the exact sequence
0 →H
0
(X
Σ
, O
X
Σ
(D))
m
→C →H
0
(V
simp
D,m
, C) →H
1
(X
Σ
, O
X
Σ
(D))
m
→0.
Since H
0
(X
Σ
, O
X
Σ
(D))
m
,=0 implies V
simp
D,m
=∅ (Exercise 9.1.7), we obtain
(9.1.13)
¯
H
0
(V
simp
D,m
, C) ≃H
1
(X
Σ
, O
X
Σ
(D))
m
¯
H
−1
(V
simp
D,m
, C) ≃H
0
(X
Σ
, O
X
Σ
(D))
m
,
where the last line uses the exact sequence (9.0.11).
It remains to prove (9.1.12). Since [Σ[ is the union of the maximal cones and
V
simp
D,m
⊆ [Σ[, we get the closed cover C = ¦V
simp
D,m
∩σ
i
¦ of V
simp
D,m
, where the σ
i
are
the maximal cones of Σ. Furthermore, C
•
is the “
ˇ
Cech” complex
ˇ
C
•
(C, C) for the
constant sheaf C on V
simp
D,m
, where we use quotation marks since C is a closed cover
rather than an open cover.
There is an E
1
spectral sequence
E
p,q
1
=
γ∈[ℓ]p
H
q
(V
simp
D,m
∩σ
γ
, C) ⇒H
p+q
(V
simp
D,m
, C)
396 Chapter 9. Sheaf Cohomology of Toric Varieties
that applies to this situation, were the differential d
p,q
1
is determined by the
ˇ
Cech
differential (see [64, Theorem II.5.4.2]). In particular,
E
p,0
1
=
ˇ
C
p
(C, C) =C
p
.
Also, we noted earlier in the proof that each V
D,m
∩σ
γ
is convex. It follows that
V
D,m
∩σ
γ
is contractible, so that
E
p,q
1
=
γ∈[ℓ]p
H
q
(V
simp
D,m
∩σ
γ
, C) =0, q > 0.
As in the proof of Theorem 9.0.4, this implies that
E
p,q
2
=
_
H
p
(
ˇ
C
•
(C, C)) =H
p
(C
•
) q =0
0 q > 0.
Following the proof of Theorem 9.0.4, we invoke Proposition 9.A.3 to conclude
that the edge homomorphisms are isomorphisms, i.e.,
H
p
(C
•
) ≃H
p
(V
simp
D,m
, C).
The proves (9.1.12) and completes the proof of part (a).
For part (b), we assume that D is QCartier with support function ϕ
D
. Then
one can modify the proof of (9.1.9) to obtain the equivalence
H
0
(U
σ
, O
X
Σ
(D))
m
,=¦0¦ ⇐⇒ V
supp
D,m
∩σ =∅
(Exercise 9.1.8). This allows us to replace (9.1.10) with the exact sequence
0 −→H
0
(U
σ
, O
X
Σ
(D))
m
−→C −→H
0
(V
supp
D,m
∩σ, C) −→0.
From here, the proof is identical to what we did in part (a) (Exercise 9.1.8).
Example 9.1.3. Let X
Σ
be the toric surface obtained by blowing up P
2
at the three
ﬁxed points of the torus action. The minimal generators u
1
, . . . , u
6
of the fan Σ
give divisors D
1
, . . . , D
6
on X
Σ
. Consider the divisor D =−D
3
+2D
5
−D
6
. Let us
compute H
1
(X
Σ
, O
X
Σ
(D))
m
for m =e
1
. Since
¸m, u
1
) ≥0, ¸m, u
2
) ≥0, ¸m, u
3
) < 1, ¸m, u
4
) < 0, ¸m, u
5
) ≥−2, ¸m, u
6
) < 1,
we get the sets V
simp
D,m
and V
supp
D,m
shown in Figure 2 on the next page (Exercise 9.1.9).
The open circle at the origin is a reminder that V
supp
D,m
does not contain the origin.
Since both sets have two connected components, Theorem 9.1.2 implies that
H
1
(X
Σ
, O
X
Σ
(D))
m
≃
¯
H
0
(V
simp
D,m
, C) ≃
¯
H
0
(V
supp
D,m
, C) ≃C. ♦
Each representation H
p
(X
Σ
, O
X
Σ
(D))
m
in Theorem 9.1.2 has its advantages.
For example, the vanishing theorems in §9.2 use V
supp
D,m
because of its relation to
convexity, while for surfaces, we will see below that V
simp
D,m
is easier to work with.
§9.1. Cohomology of Toric Divisors 397
u
1
u
4
u
2
u
5
u
3
u
6
V
D,m
simp
→
←V
D,m
simp
u
1
u
4
u
2
u
5
u
3
u
6
V
D,m
supp
V
D,m
supp
Figure 2. V
simp
D,m
and V
supp
D,m
for D =−D3 +2D5 −D6 and m =e1
A limitation of Theorem 9.1.2 is that it assumes that Σ is simplicial or D is
QCartier. Exercise 9.1.10 will discuss a version of the theorem that holds for
any fan Σ and any torusinvariant Weil divisor D on X
Σ
. Other treatments of
H
p
(X
Σ
, O
X
Σ
(D))
m
can be found in [50], [85] and [143].
Also note that Theorem 9.1.2 can be stated using relative cohomology instead
of reduced cohomology. Consider, for example, the case when D is QCartier.
Then the inclusion V
supp
D,m
⊆[Σ[ gives the long exact sequence
−→H
p
([Σ[,V
supp
D,m
, C) −→H
p
([Σ[, C) −→H
p
(V
supp
D,m
, C) −→ .
Since [Σ[ is contractible (it is a cone), this sequence and Theorem 9.1.2 imply
(9.1.14) H
p
(X
Σ
, O
X
Σ
(D))
m
≃H
p
([Σ[,V
supp
D,m
, C), p ≥0
(Exercise 9.1.11). Since V
supp
D,m
is open in [Σ[, algebraic geometers often write this
relative cohomology group as H
p
Z
D,m
([Σ[, C), where Z
D,m
is the closed set
Z
D,m
=[Σ[ ¸V
D,m
=¦u ∈ [Σ[ [ ¸m, u) ≥ϕ
D
(u)¦.
Then part (b) of Theorem 9.1.2 can be restated as
H
p
(X
Σ
, O
X
Σ
(D))
m
≃H
p
Z
D,m
([Σ[, C).
This is the version that appears in [59, p. 74].
The Surface Case. A toric surface is simplicial, which means that we can compute
H
p
(X
Σ
, O
X
Σ
(D))
m
using V
simp
D,m
. The topology of V
simp
D,m
is especially simple and can
be encoded by a sign sequence. We illustrate this with an example.
398 Chapter 9. Sheaf Cohomology of Toric Varieties
u
1
u
4
u
2
u
5
u
3
u
6
V
D,m
simp
→
←V
D,m
simp
+ −
−
−
+
+
Figure 3. The sign pattern for D =−D3 +2D5 −D6 and m =e1
Example 9.1.4. We return to the situation of Example 9.1.3. Write the divisor
D = −D
3
+2D
5
−D
6
as D =
i
a
i
D
i
, and label the ray generated by u
i
with + if
¸m, u
i
) ≥−a
i
and with − if ¸m, u
i
) <−a
i
. Then the picture of V
simp
D,m
from Figure 2
gives the sign pattern shown in Figure 3.
If we start at u
1
and go counterclockwise around the origin, we get the sign
pattern ++−−+−, which we regard as cyclic. The positions of the −’s determine
V
simp
D,m
, so that connected components of V
simp
D,m
correspond to strings of consecutive
−’s in the sign pattern. ♦
These observations holds in general. Let X
Σ
be a complete toric surface with
minimal generators u
1
, . . . , u
r
arranged counterclockwise around the origin. Given
a torusinvariant Weil divisor D=
i
a
i
D
i
on X
Σ
and m∈M≃Z
2
, the sign pattern
sign
D
(m) is the string of length r whose ith entry is + if ¸m, u
i
) ≥ −a
i
and − if
¸m, u
i
) < −a
i
. We regard sign
D
(m) as cyclic. Thus, for example, −−++++−
has only one string of consecutive −’s.
We have the following result from [86] and [94] (Exercise 9.1.12).
Proposition 9.1.5. Given a Weil divisor D =
i
a
i
D
i
on a complete toric surface
X
Σ
, the dimension of H
p
(X
Σ
, O
X
Σ
(D))
m
is given by
dimH
0
(X
Σ
, O
X
Σ
(D))
m
=
_
1 if sign
D
(m) =+ +(⇔V
simp
D,m
=∅)
0 otherwise
dimH
1
(X
Σ
, O
X
Σ
(D))
m
=max(0, #connected components of V
simp
D,m
−1)
=max(0, #strings of consecutive −’s in sign
D
(m) −1)
dimH
2
(X
Σ
, O
X
Σ
(D))
m
=
_
1 if sign
D
(m) =− −(⇔V
simp
D,m
is a cycle)
0 otherwise.
§9.1. Cohomology of Toric Divisors 399
Example 9.1.6. Let us revisit Example 9.1.1. For each m ∈ M = Z
2
, the minimal
generators u
0
, u
1
, u
2
give a sign pattern for the line bundle O
P
2 (a) = O
P
2 (aD
0
).
When a < 0, all possible sign patterns are recorded in Figure 1 of Example 9.1.1.
Using Proposition 9.1.5, we obtain:
• H
0
(P
2
, O
P
2 (a)) =0 since +++ does not appear.
• H
1
(P
2
, O
P
2 (a)) =0 since all patterns have one string of consecutive −’s.
• H
2
(P
2
, O
P
2 (a)) =
m∈Int(a∆
2
)
C χ
m
since −−− labels the interior of a∆
2
.
Thus the computation of Example 9.1.1 follows immediately from Figure 1. ♦
Example 9.1.7. Consider the surface X
Σ
and divisor D = −D
3
+2D
5
−D
6
from
Example 9.1.3. The sign patterns sign
D
(m) for all m ∈ M ≃ Z
2
give the chamber
decomposition shown in Figure 4 (Exercise 9.1.9).
− − − + + −
− − − + + + + − − − + + + + − − + +
− + + + + −
+ + − − + −
+ + + − − −
+ + − − − −
+ + + − + −
+ + − − − +
− − + + + −
− + + + − −
− + − + + −
+−−−+−
l
6
l
2
l
5
l
3
l
4
= l
1
e
1 0
2e
1
←
+ + + + − −
←
+ + − + + −
←
+ − − + + +
+ − − + + −
→
+ + + + + −
→
Figure 4. The chamber decompostion for D =−D3 +2D5 −D6
As in Figure 1, we have lines l
i
deﬁned by ¸m, u
i
) = −a
i
, where D =
i
a
i
D
i
.
We put the label l
i
on the side where ¸m, u
i
) ≥−a
i
, and the shading indicates where
the boundary points lie. Each chamber is labeled with its sign pattern (some labels
are rather small), and we get ﬁve 1dimensional chambers since l
1
=l
4
.
We now compute cohomology using Proposition 9.1.5. First,
H
0
(X
Σ
, O
X
Σ
(D)) =H
2
(X
Σ
, O
X
Σ
(D)) =0
since neither ++++++nor −−−−−−appear anywhere. Furthermore, of the
six chambers with sign patterns having more than one string of consecutive −’s,
400 Chapter 9. Sheaf Cohomology of Toric Varieties
four of them (+++−+−, −+−++−, +−−−+−, plus the 1dimensional
+−−++−) have no lattice points. Of the remaining two, the lattice points are:
• m =0 from ++−++−(1dimensional).
• m =e
1
, 2e
1
from ++−−+−.
These lattice points are indicated with white dots in Figure 4. Hence
H
1
(X
Σ
, O
X
Σ
(D)) ≃C χ
0
⊕C χ
e
1
⊕C χ
2e
1
.
In particular, dimH
1
(X
Σ
, O
X
Σ
(D)) =3. ♦
The paper [86] applies these methods to classify line bundles with vanishing
higher cohomology on the toric surface coming from three consecutive blowups of
the Hirzebruch surface H
2
.
Exercises for §9.1.
9.1.1. In Example 9.1.1, verify the sign patterns in Figure 1.
9.1.2. Use the methods of Example 9.1.1 to prove (9.1.7).
9.1.3. The canonical bundle of P
2
is ω
P
2 =O
P
2 (−3). Use this and Serre duality to explain
the symmetry in Table 2 in Example 9.1.1.
9.1.4. Recall that a Weil divisor D =
ρ
a
ρ
D
ρ
on X
Σ
is QCartier if some positive integer
multiple of ℓD is Cartier. Let ϕ
ℓD
be the support function of ℓD as in Theorem 4.2.12.
(a) Show that (1/ℓ)ϕ
ℓD
: [Σ[ →R is a support function (Deﬁnition 4.2.11) and depends
only on D. We deﬁne the support function of D to be ϕ
D
=(1/ℓ)ϕ
ℓD
.
(b) Show that D is QCartier if and only if for every σ ∈ Σ, there is m
σ
∈ M
Q
such that
¸m
σ
, u
ρ
) =−a
ρ
for all ρ ∈ σ(1). When this is satisﬁed, show that ϕ
D
(u) =¸m
σ
, u) for
all u ∈ σ.
9.1.5. Given a ring R and elements f
1
, . . . , f
ℓ
∈ R, deﬁne
d
p
:
_
p
R
ℓ
−→
_
p+1
R
ℓ
by d
p
(α) = (
ℓ
i=1
f
i
e
i
) ∧α, where e
1
, . . . , e
ℓ
are the standard basis of R
ℓ
. Setting K
p
=
_
p
R
ℓ
, we get the complex
K
•
: K
0
d
0
−→K
1
d
1
−→
d
ℓ−2
−→K
ℓ−1
d
ℓ−1
−→K
ℓ
,
and for an Rmodule M, we get the Koszul complex
K
•
( f
1
, . . . , f
ℓ
; M) =K
•
⊗
R
M.
Thus K
•
=K
•
( f
1
, . . . , f
ℓ
; R).
(a) Given ϕ : R
ℓ
→R, show that there are maps s
p
: K
p
→K
p−1
such that s
1
= ϕ and
s
p+q
(α∧β) =s
p
(α) ∧β +(−1)
p
α∧s
q
(β) for all α ∈ K
p
and β ∈ K
q
.
(b) Show that for all α ∈ K
p
, the maps s
p
satisfy
d
p−1
(s
p
(α)) +s
p+1
(d
p
(α)) =ϕ(
ℓ
i=1
f
i
e
i
)α.
(c) Now assume that ¸ f
1
, . . . , f
ℓ
) = R. Prove that K
•
( f
1
, . . . , f
ℓ
; M) is exact for every R
module M. Hint: Deﬁne ϕ(e
i
) =g
i
where
ℓ
i=1
g
i
f
i
=1.
§9.2. Vanishing Theorems I 401
9.1.6. When f
1
= = f
ℓ
=1 and Mis an Rmodule, the Koszul complex K
•
(1, . . . , 1; M)
of Exercise 9.1.5 becomes
(9.1.15) M−→
1≤i≤ℓ
M−→
1≤i<j≤ℓ
M−→
1≤i<j<k≤ℓ
M−→ ,
where the entries of the matrices representing the differentials are 0 or ±1.
(a) Use the previous exercise to prove that (9.1.15) is exact.
(b) Prove that the complex B
•
deﬁned in the proof of Theorem 9.1.2 satisﬁes (9.1.11).
Hint: Show B
•
is the Koszul complex (9.1.15) with M=C, minus its ﬁrst term.
9.1.7. Prove that H
0
(X
Σ
, O
XΣ
(D))
m
,=0 if and only if V
simp
m,D
=∅. Then prove (9.1.13).
9.1.8. Complete the proof of part (b) of Theorem 9.1.2.
9.1.9. Verify Figure 2 in Example 9.1.3 and Figure 4 in Example 9.1.7.
9.1.10. Let D =
ρ
a
ρ
D
ρ
be a Weil divisor on X
Σ
. Inspired by [143], we describe a
“simplicial” computation of H
p
(X
Σ
, O
XΣ
(D))
m
that applies even when Σ is not simplicial.
Let R
Σ(1)
be a real vector space with basis e
ρ
, ρ ∈ Σ(1). Each σ ∈ Σ gives the simplex
∆
σ
= Conv(e
ρ
[ ρ ∈ σ(1)¦. For m ∈ M, let Σ
D,m
= ¦σ ∈ Σ [ ¸m, u
ρ
) < −a
ρ
, ρ ∈ σ(1)¦ as
in the text, and set ∆
D,m
=
σ∈ΣD,m
∆
σ
.
(a) For σ ∈ Σ, show that H
0
(U
σ
, O
XΣ
(D))
m
,=0 if and only if ∆
D,m
∩∆
σ
=∅.
(b) Adapt the proof of Theorem 9.1.2 to show that H
p
(X
Σ
, O
XΣ
(D))
m
≃
¯
H
p−1
(∆
D,m
, C).
Readers familiar with simplicial complexes should note that ∆
D,m
has a natural structure
as a simplicial complex, so
¯
H
p−1
(∆
D,m
, C) can be computed using simplicial cohomology.
9.1.11. Prove (9.1.14).
9.1.12. Prove Proposition 9.1.5.
9.1.13. Prove that H
0
(X
Σ
, O
XΣ
(D))
m
,=0 implies that H
p
(X
Σ
, O
XΣ
(D))
m
=0 for all p >0.
Hint: Exercise 9.1.7.
9.1.14. Assume Σis simplicial and let V
D,m
=
_
σ∈ΣD,m
σ
_
¸¦0¦, where Σ
D,m
is as deﬁned
in the text.
(a) DrawV
D,m
in the situation of Example 9.1.3.
(b) Show that V
simp
D,m
⊆V
D,m
is a deformation retract.
(c) Conclude that H
p
(X
Σ
, O
XΣ
(D))
m
≃
¯
H
p−1
(V
D,m
, C).
§9.2. Vanishing Theorems I
In this section we prove two basic vanishing theorems for QCartier divisors on a
toric variety X
Σ
. We then apply these results to show that normal toric varieties are
CohenMacaulay and hence satisfy Serre duality. We also give a version of Serre
duality that is special to the toric case.
402 Chapter 9. Sheaf Cohomology of Toric Varieties
Nef QCartier Divisors. A Weil divisor D on a normal variety X is QCartier if
some positive integer multiple is Cartier. Also recall that the intersection product
D C is deﬁned whenever D is QCartier and C is a complete irreducible curve in
X. Then D is nef if D C ≥ 0 for all complete irreducible curves C ⊆ X. This
generalizes the deﬁnition of nef Cartier divisor given in §6.2.
In the toric case, we characterize QCartier nef divisors as follows.
Lemma 9.2.1. Let D =
ρ
a
ρ
D
ρ
be a QCartier divisor on X
Σ
. If Σ has convex
support, then the following are equivalent:
(a) D is nef.
(b) For some integer ℓ > 0, ℓD is Cartier and basepoint free, i.e., O
X
Σ
(ℓD) is a
line bundle generated by its global sections.
(c) The support function ϕ
D
: [Σ[ →R is convex.
Proof. The proof combines ideas from Theorems 6.1.10, 6.2.12 and 7.2.2. We
leave the details to the reader (Exercise 9.2.1).
In the Cartier case, we know that D is nef if and only if O
X
Σ
(D) is generated
by global sections (Theorems 6.1.10 and6.2.12). This fails in the QCartier case,
as shown by the following example.
Example 9.2.2. Let X
Σ
=P(2, 3, 5) and let D
0
, D
1
, D
2
be the divisors given by the
minimal generators u
0
, u
1
, u
2
. Note that 2u
0
+3u
1
+5u
2
= 0. Then Cl(X
Σ
) ≃ Z,
where the classes of D
0
, D
1
, D
2
map to 2, 3, 5 respectively. Let D = D
1
−D
0
and
note that D
0
∼ 2D, D
1
∼ 3D, D
2
∼ 5D. It also easy to see that D is QCartier and
nef. However, you will check in Exercise 9.2.2 that H
0
(X
Σ
, O
X
Σ
(D)) = 0. Thus
O
X
Σ
(D) is not generated by its global sections. ♦
Demazure Vanishing. The most basic vanishing theorem for toric varieties is the
following result, ﬁrst proved by Demazure [45] for Cartier divisors.
Theorem 9.2.3 (Demazure Vanishing). Let D be a QCartier divisor on X
Σ
. If [Σ[
is convex and D is nef, then
H
p
(X
Σ
, O
X
Σ
(D)) =0 for all p > 0.
Proof. The support function ϕ
D
: [Σ[ →R is convex by Lemma 9.2.1. Fix m ∈ M
and let V
supp
D,m
= ¦u ∈ [Σ[ [ ¸m, u) < ϕ
D
(u)¦ as in Theorem 9.1.2. Let u, v ∈V
supp
D,m
,
so that ¸m, u) < ϕ
D
(u) and ¸m, v) < ϕ
D
(u). Then, for 0 ≤t ≤1, we have
¸m, (1−t)u+tv) =(1−t)¸m, u) +t¸m, v)
<(1−t)ϕ
D
(u) +tϕ
D
(v)
≤ϕ
D
((1−t)u+tv).
§9.2. Vanishing Theorems I 403
The last line follows since ϕ
D
is convex. This implies that V
supp
D,m
is convex and
hence contractible. Combining this with Theorem 9.1.2, we obtain
H
p
(X
Σ
, O
X
Σ
(D))
m
≃
¯
H
p−1
(V
supp
D,m
, C) =0 for all p > 0.
In particular, if D is a basepoint free Cartier divisor on X
Σ
, then D is nef, so
that Theorem 9.2.3 implies
H
p
(X
Σ
, O
X
Σ
(D)) =0 for all p > 0.
Example 9.2.4. For P
n
, aD
0
is basepoint free for a ≥ 0. Hence the higher co
homology of O
P
n (a) = O
P
n (aD
0
) vanishes by Theorem 9.2.3, which agrees with
what we found in Example 9.1.1 when n =2. ♦
Vanishing of Higher Direct Images. Here is an easy application of Demazure
vanishing.
Theorem 9.2.5. Let φ : X
Σ
1
→X
Σ
2
be a proper toric morphism and let L be a line
bundle on X
Σ
2
Then:
(a) R
p
φ
∗
φ
∗
L =0 for all p > 0.
(b) If the map φ: N
1
→N
2
on the lattices of oneparameter subgroups is surjective,
then φ
∗
φ
∗
L ≃L.
Proof. First assume L = O
X
Σ
2
, so that φ
∗
L = O
X
Σ
1
. Let U
σ
⊆X
Σ
2
be the afﬁne
open corresponding to σ ∈ Σ
2
. By Proposition 9.0.7, R
p
φ
∗
O
X
Σ
1
[
Uσ
is the sheaf
associated to the module H
p
(φ
−1
(U
σ
), O
X
Σ
1
). Since φ is proper, Theorem 3.4.7
tells us that φ
−1
(U
σ
) is a toric variety whose fan is supported on the convex set
φ
−1
R
(σ). Thus H
p
(φ
−1
(U
σ
), O
X
Σ
1
) =0 for p > 0 by Theorem 9.2.3. It follows that
R
p
φ
∗
O
X
Σ
1
=0 for p > 0 by Proposition 9.0.7.
The morphism φ induces a homomorphism of sheaves O
X
Σ
2
→φ
∗
O
X
Σ
1
on X
Σ
2
(Example 4.0.25). It sufﬁces to show that this map is an isomorphism on each afﬁne
open U
σ
⊆ X
Σ
2
, σ ∈ Σ
2
. Note also that the restriction φ
−1
(U
σ
) →U
σ
is again a
proper toric morphism. Hence we may assume that X
Σ
2
=U
σ
. Then [Σ
1
[ =φ
−1
R
(σ)
is a convex cone, so that
H
0
(X
Σ
1
, O
X
Σ
1
) =
m∈[Σ
1
[
∨
∩M
1
C χ
m
by Exercise 4.3.4. We analyze [Σ
1
[
∨
∩M
1
as follows.
The surjection N
1
→N
2
gives an injection M
2
⊆M
1
on character lattices. Then
σ ⊆ (N
2
)
R
has dual σ
∨
⊆ (M
2
)
R
⊆ (M
1
)
R
and [Σ
1
[ ⊆ (N
1
)
R
has dual [Σ
1
[
∨
⊆
(M
1
)
R
. Also, [Σ
1
[ =φ
−1
R
(σ) implies σ
∨
=[Σ
1
[
∨
. It follows that
[Σ
1
[
∨
∩M
1
=σ
∨
∩M
1
=σ
∨
∩(M
2
)
R
∩M
1
=σ
∨
∩M
2
,
404 Chapter 9. Sheaf Cohomology of Toric Varieties
where the last equality follows since M
1
is saturated in M
2
by the surjectivity of
N
1
→N
2
. Hence
H
0
(X
Σ
1
, O
X
Σ
1
) =
m∈σ
∨
∩M
2
C χ
m
=H
0
(U
σ
, O
Uσ
).
This shows that O
Uσ
→φ
∗
O
X
Σ
1
induces an isomorphism on global sections, which
gives the desired sheaf isomorphism since we are working with quasicoherent
sheaves over an afﬁne variety.
For a line bundle L on X
Σ
2
, take U
σ
⊆X
Σ
2
. Then L[
Uσ
≃O
X
Σ
2
[
Uσ
, so that
R
p
φ
∗
φ
∗
L[
Uσ
≃R
p
φ
∗
φ
∗
O
X
Σ
2
[
Uσ
≃R
p
φ
∗
O
X
Σ
1
[
Uσ
=0
for p > 0. Furthermore, when φ : N
1
→N
2
is surjective, we have
φ
∗
φ
∗
L ≃φ
∗
O
X
Σ
1
⊗
O
X
Σ
2
L ≃L,
where the ﬁrst isomorphism is part (a) of Exercise 9.2.3 and the second follows
from φ
∗
O
X
Σ
1
≃O
X
Σ
2
.
The Injectivity Lemma. We introduce a method that will be used several times in
this section and the next. The general framework uses a morphism f : Y →X and
coherent sheaves G on Y and F on X such that:
• f is an afﬁne morphism, meaning f
−1
(U) ⊆X is afﬁne when U ⊆Y is afﬁne.
• There is a split injection i : F → f
∗
G of O
X
modules. This means that there is
a homomorphism r : f
∗
G →F such that r ◦i =1
F
.
By functorality, a split injection i : F → f
∗
G induces a split injection
H
p
(X, F) −→H
p
(X, f
∗
G).
However, since f is afﬁne, we also have
H
p
(X, f
∗
G) ≃H
p
(Y, G)
by Exercise 9.0.5. Hence we get an injection
H
p
(X, F) ֒→H
p
(Y, G).
In particular, H
p
(Y, G) = 0 implies H
p
(X, F) = 0. This method is used in the
proofs of many vanishing theorems—see [114, Ch. 4].
In the toric context, Fujino [54] identiﬁed the best choice for the morphism f .
Given a fan Σ in N
R
and a positive integer ℓ, let φ
ℓ
: N →N be multiplication by
ℓ. This maps Σ to itself and hence induces a toric morphism φ
ℓ
: X
Σ
→X
Σ
. The
dual of φ
ℓ
is multiplication by ℓ on M, and the restriction of φ
ℓ
to T
N
= M⊗
Z
C
∗
is the group homomorphism given by raising to the ℓth power. Furthermore, given
any σ ∈ Σ, we have φ
−1
ℓ
(U
σ
) =U
σ
since (φ
ℓ
)
−1
R
(σ) =σ. This shows that φ
ℓ
is an
afﬁne morphism.
Here is our ﬁrst injectivity lemma.
§9.2. Vanishing Theorems I 405
Lemma 9.2.6. Let D be a Weil divisor on X
Σ
and let ℓ be a positive integer. Then
there is an injection
H
p
(X
Σ
, O
X
Σ
(D)) ֒→H
p
(X
Σ
, O
X
Σ
(ℓD)) for all p ≥0.
Proof. We will construct a split injection O
X
Σ
(D) ֒→φ
ℓ∗
O
X
Σ
(ℓD). The lemma
then follows immediately from the above discussion.
For D=
ρ
a
ρ
D
ρ
and σ ∈Σ, let P=¦m∈M
R
[ ¸m, u) ≥−a
ρ
for all ρ ∈σ(1)¦.
Over U
σ
, the sheaves O
X
Σ
(D) and O
X
Σ
(ℓD) come from the C[σ
∨
∩M]modules
m∈P∩M
C χ
m
and
m∈(ℓP)∩M
C χ
m
.
Since φ
−1
ℓ
(U
σ
) =U
σ
, φ
ℓ∗
O
X
Σ
(ℓD) is determined by
m∈(ℓP)∩M
C χ
m
with the
module structure given by χ
m
a =χ
ℓm
a. This implies that the map
(9.2.1)
m∈P∩M
C χ
m
−→
m∈(ℓP)∩M
C χ
m
that sends χ
m
to χ
ℓm
is a homomorphism of C[σ
∨
∩M]modules (Exercise 9.2.4).
Furthermore, one can show that the map given by
(9.2.2) r(χ
m
) =
_
0 m / ∈ ℓM
χ
m
′
m =ℓm
′
, m
′
∈ M
deﬁnes a C[σ
∨
∩M]module splitting of (9.2.1) (Exercise 9.2.4).
The map (9.2.1) and the splitting r are easily seen to be compatible with the
inclusion U
τ
⊆U
σ
when τ is a face of σ. It follows that the maps (9.2.1) patch to
give the split injection O
X
Σ
(D) ֒→φ
ℓ∗
O
X
Σ
(ℓD).
BatyrevBorisov Vanishing. In Example 9.1.1, we saw that O
P
2 (a) has nontrivial
H
2
when a < 0. If we write a =−b for b > 0, we can rewrite (9.1.6) as
dimH
p
(P
2
, O
P
2 (−b)) =
_
0 p ,=2
¸
¸
Int(b∆
2
) ∩M
¸
¸
p =2.
This has been generalized by Batyrev and Borisov [11]. Recall that a Weil divisor
D=
ρ
a
ρ
D
ρ
gives the polyhedron
P
D
=¦m ∈ M
R
[ ¸m, u
ρ
) ≥−a
ρ
for all ρ ∈ Σ(1)¦,
which is a polytope when Σ is complete (Proposition 4.3.8).
Theorem 9.2.7 (BatyrevBorisov Vanishing). Let D =
ρ
a
ρ
D
ρ
be a QCartier
divisor on a complete toric variety X
Σ
. If D is nef, then
(a) H
p
(X
Σ
, O
X
Σ
(−D)) =0 for all p ,=dimP
D
.
(b) When p =dimP
D
, H
p
(X
Σ
, O
X
Σ
(−D)) ≃
m∈Relint(P
D
)∩M
C χ
−m
.
406 Chapter 9. Sheaf Cohomology of Toric Varieties
Proof. First assume that D is Cartier and dimP
D
= dimN
R
= dimX
Σ
. Then D is
basepoint free since it is nef (Theorem 6.2.12). By Proposition 6.1.21 and Theo
rem 6.1.22, combining cones of Σ that share the same linear functional relative to
ϕ
D
gives a fan Σ
D
such that Σ reﬁnes Σ
D
and ϕ
D
is strictly convex relative to Σ
D
.
For m ∈ M, H
p
(X
Σ
, O
X
Σ
(−D))
m
≃
¯
H
p−1
(V
supp
−D,m
, C) by Theorem 9.1.2. Set
Z
−D,m
=N
R
¸V
supp
−D,m
=¦u ∈ N
R
[ ¸m, u) ≥ϕ
−D
(u)¦.
Since ϕ
D
is strictly convex relative to Σ
D
, the proof of Theorem 9.2.3 implies that
Z
−D,m
is convex. In fact, it is strongly convex. To see why, note that for any nonzero
u ∈ N
R
, the strict convexity of ϕ
D
implies 0 = ϕ
D
(u+(−u)) > ϕ
D
(u) +ϕ
D
(−u)
since u and −u do not lie in the same cone of Σ
D
. Thus
ϕ
−D
(−u) =−ϕ
D
(−u) > ϕ
D
(u) for all u ,=0 in N
R
.
Now assume u ∈ Z
−D,m
¸¦0¦. Then ¸m, u) ≥ϕ
−D
(u), so that
ϕ
D
(u) ≥¸m, −u).
Combining the displayed inequalities gives ϕ
−D
(−u) > ¸m, −u), which implies
that −u ∈V
supp
−D,m
=N
R
¸Z
−D,m
. Hence Z
−D,m
is strongly convex.
We next prove that
(9.2.3) Z
−D,m
=¦0¦ ⇐⇒ m ∈ −Int(P
D
) ∩M.
If m ∈ −Int(P
D
) ∩M, then −m ∈ Int(P
D
), so that
¸−m, u
ρ
) >−a
ρ
=ϕ
D
(u
ρ
) for all ρ ∈ Σ(1).
Thus
(9.2.4) ¸m, u
ρ
) < ϕ
−D
(u
ρ
) for all ρ ∈ Σ(1),
which easily implies that Z
−D,m
= ¦0¦ (Exercise 9.2.5). On the other hand, if we
have m / ∈ −Int(P
D
) ∩M, then one of the inequalities of (9.2.4) must fail, i.e.,
¸m, u
ρ
0
) ≥ϕ
−D
(u
ρ
0
) for some ρ
0
∈ Σ(1).
Then u
ρ
0
∈ Z
−D,m
, so that Z
−D,m
,=¦0¦. This proves (9.2.3).
We are now ready to compute H
p
(X
Σ
, O
X
Σ
(−D))
m
. If m∈ −Int(P
D
) ∩M, then
V
supp
−D,m
=N
R
¸¦0¦ is homotopic to S
n−1
. Hence
H
p
(X
Σ
, O
X
Σ
(−D))
m
≃
¯
H
p−1
(V
supp
−D,m
, C) ≃
¯
H
p−1
(S
n−1
, C) =
_
0 p ,=n
C p =n.
If m / ∈ −Int(P
D
) ∩M, then V
supp
−D,m
= R
n
¸Z
−D,m
, where Z
−D,m
is a closed strongly
convex cone of positive dimension. If u ,= 0 in Z
−D,m
, we showed above that
−u ∈V
supp
−D,m
. It is easy to see that V
supp
D,m
contracts to −u (Exercise 9.2.6). Hence
H
p
(X
Σ
, O
X
Σ
(−D))
m
≃
¯
H
p−1
(V
supp
−D,m
, C) =0 for all p ≥0.
This completes the proof when D is Cartier and P
D
has dimension n.
§9.2. Vanishing Theorems I 407
Now suppose D is Cartier but dimP
D
< n. We will use the Proposition 6.1.21
and Theorem 6.1.22. After translation, P
D
spans (M
D
)
R
⊆ M
R
, where M
D
⊆ M is
dual to a surjection φ : N →N
D
, and the normal fan of P
D
lies in (N
D
)
R
. If X
D
is
the toric variety of P
D
⊆(M
D
)
R
, then Theorem 6.1.22 shows that φ induces a toric
morphism φ : X
Σ
→X
D
such that D is the pullback of the ample divisor D
′
on X
D
determined by P
D
. In particular, O
X
Σ
(−D) ≃φ
∗
O
X
D
(−D
′
).
Since φ is proper and φ : N →N
D
is surjective, we have
R
p
φ
∗
O
X
Σ
(−D) =0, p > 0
φ
∗
O
X
Σ
(−D) =O
X
D
(−D
′
)
by Theorem 9.2.5. Then Proposition 9.0.8 implies that
(9.2.5) H
p
(X
D
, O
X
D
(−D
′
)) ≃H
p
(X
Σ
, O
X
Σ
(−D)).
This proves the theorem, as we now explain. Since M
D
is saturated in M by the
surjectivity of N →N
D
, we have
P
D
∩M =P
D
∩(M
D
)
R
∩M =P
D
′ ∩M
D
,
Since P
D
= P
D
′ is full dimensional in (M
D
)
R
, we are done by (9.2.5) and the full
dimensional case considered above.
Finally, we need to consider what happens when Dis QCartier. Pick an integer
ℓ > 0 such that ℓD is Cartier. By Lemma 9.2.6, we have an injection
H
p
(X
Σ
, O
X
Σ
(−D)) ֒→H
p
(X
Σ
, O
X
Σ
(−ℓD)).
Since ℓD is Cartier and nef with P
ℓD
=ℓP
D
, the Cartier case proved above implies
that H
p
(X
Σ
, O
X
Σ
(−D)) =0 for p ,=dimP
D
. Furthermore, when p =dimP
D
,
H
p
(X
Σ
, O
X
Σ
(−ℓD)) =
m∈M
¯
H
p−1
(V
supp
−ℓD,m
, C) χ
m
=
m∈Relint(ℓP
D
)∩M
C χ
−m
,
from which we conclude
¯
H
p−1
(V
supp
−ℓD,m
, C) =
_
C m ∈ −Relint(ℓP
D
) ∩M
0 otherwise.
Since V
supp
−D,m
=V
supp
−ℓD,ℓm
and ℓm ∈ −Relint(ℓP
D
) ∩M ⇔m ∈ −Relint(P
D
) ∩M,
H
p
(X
Σ
, O
X
Σ
(−D)) =
m∈M
¯
H
p−1
(V
supp
−D,m
, C) χ
m
=
m∈Relint(P
D
)∩M
C χ
−m
.
Example 9.2.8. We saw in Example 9.2.4 that the higher cohomology of O
P
n (a)
vanishes when a ≥ 0. When a < 0, O
P
n (a) = O
P
n (−[a[D
0
). Since [a[D
0
is ample
with polytope [a[∆
n
, Theorem 9.2.7 implies
dimH
p
(P
n
, O
P
n (a)) =
_
0 p ,=n
¸
¸
Int(a∆
n
) ∩M
¸
¸
p =n
when a < 0. This generalizes (9.1.6) from Example 9.1.1. ♦
408 Chapter 9. Sheaf Cohomology of Toric Varieties
The CohenMacaulay Property. Here is a surprising consequence of Batyrev
Borisov vanishing.
Theorem 9.2.9. A normal toric variety is CohenMacaulay.
Proof. First suppose that X
Σ
is projective and let D be a torusinvariant ample
divisor on X
Σ
. Corollary 5.72 of [108, p. 182] implies that X
Σ
is CohenMacaulay
if and only if
(9.2.6) H
p
(X
Σ
, O
X
Σ
(−ℓD)) =0 for all ℓ ≫0, p < n.
Now take any integer ℓ > 0. Since D is ample, the divisor ℓD is nef and its
polytope P
ℓD
has dimension n =dimX
Σ
. Then (9.2.6) follows from Theorem 9.2.7.
We conclude that X
Σ
is CohenMacaulay.
Next consider an afﬁne toric variety U
σ
. We can ﬁnd a projective toric variety
that contains U
σ
as an afﬁne open subset (Exercise 9.2.7). Since being Cohen
Macaulay is a local property, it follows that U
σ
is CohenMacaulay, and then any
normal toric variety X
Σ
is CohenMacaulay.
This result was originally proved by Hochster [89]. Another proof can be found
in [42, Thm. 3.4]. A projective variety is called arithmetically CohenMacaulay
(aCM for short) if its afﬁne cone is CohenMacaulay. In Exercise 9.2.8 you will
show that normal lattice polytopes give aCM toric varieties.
Serre Duality. Theorem 9.2.9 shows that Serre duality holds for any normal toric
variety. However, version stated in Theorem 9.0.9 only applies to locally free
sheaves, while the more general version Theorem 9.0.10 uses Ext groups. Many
of the sheaves we deal with, such as O
X
Σ
(D) and
´
Ω
p
X
Σ
, fail to be locally free when
X
Σ
is not smooth. Fortunately, there is an “Extfree” version of Serre duality that
holds for these sheaves.
Theorem 9.2.10 (Toric Serre Duality). Assume that X
Σ
is a complete toric variety
of dimension n.
(a) If D is a QCartier divisor and K
X
Σ
is the canonical divisor, then we have
natural isomorphisms
H
p
(X
Σ
, O
X
Σ
(D))
∨
≃H
n−p
(X
Σ
, O
X
Σ
(K
X
Σ
−D)).
(b) If X
Σ
is simplicial and F is locally free, then there are natural isomorphisms
H
p
(X
Σ
,
´
Ω
q
X
Σ
⊗
O
X
Σ
F)
∨
≃H
n−p
(X
Σ
,
´
Ω
n−q
X
Σ
⊗
O
X
Σ
F
∨
).
Remark 9.2.11. Here are two interesting aspects of Theorem 9.2.10:
(a) O
X
Σ
(K
X
Σ
−D) need not be isomorphic to
ω
X
Σ
⊗
O
X
Σ
O
X
Σ
(D)
∨
=O
X
Σ
(K
X
Σ
) ⊗
O
X
Σ
O
X
Σ
(−D)
when D is QCartier. So Theorem 9.2.10 does not follow from Theorem 9.0.9.
§9.2. Vanishing Theorems I 409
(b) Every Weil divisor is QCartier on a simplicial toric variety. Theorem 9.2.10
holds for all Weil divisors in this case.
Before beginning the proof of Theorem 9.2.10, we introduce some tools from
commutative algebra. This is typical of algebraic geometry—once a variety ceases
to be smooth, the theory becomes more technically demanding.
In §9.0 we deﬁned the depth of a local ring (R, m). More generally, the depth of
a ﬁnitely generated Rmodule F is the maximal length of a sequence f
1
, . . . , f
s
∈ m
such that f
i
is not a zero divisor in F/¸ f
1
, . . . , f
i−1
)F for all i, and the dimension of
F is dimR/Ann(F), where Ann(F) =¦f ∈ R [ f F =0¦.
Then we deﬁne a coherent sheaf F on a variety X to be maximal Cohen
Macaulay (MCM for short) if for every point x ∈ X, the stalk F
x
is an O
X,x

module whose depth and dimension both equal dimX. For example, if X is Cohen
Macaulay, then every locally free sheaf on X is MCM.
The following result from [143, Prop. 4.24] shows that MCM sheaves satisfy a
nice version of Serre duality. See also [108, Thm. 5.71].
Theorem 9.2.12 (Serre Duality III). Let F be a MCM sheaf on a complete normal
CohenMacaulay variety X of dimension n. There there are natural isomorphisms
H
p
(X, F)
∨
≃H
n−p
(X, Hom
O
X
(F, ω
X
)).
Proof. We will use the sheaf version of Ext, where the sheaves Ext
q
O
X
(F, ω
X
) are
the derived functors of F →Hom
O
X
(F, ω
X
). The stalk at x ∈ X is
(9.2.7) Ext
q
O
X
(F, ω
X
)
x
=Ext
q
O
X,x
(F
x
, (ω
X
)
x
),
and since
Hom
O
X
(F, −) =Γ(X, Hom
O
X
(F, −)),
we can apply the Grothendieck spectral sequence for the composition of functors
[177, Thm. 5.8.3] to obtain the spectral sequence
E
p,q
2
=H
p
(X, Ext
q
O
X
(F, ω
X
)) ⇒Ext
p+q
O
X
(F, ω
X
).
Since F is MCM, [29, Cor. 3.5.11] implies that Ext
q
O
X,x
(F
x
, (ω
X
)
x
) = 0 for
q > 0 and x ∈ X. By (9.2.7), we obtain Ext
q
O
X
(F, ω
X
) =0 for q > 0. Hence
H
p
(X, Hom
O
X
(F, ω
X
)) ≃Ext
p
O
X
(F, ω
X
)
by Proposition 9.A.5. Since H
p
(X, F)
∨
≃Ext
n−p
O
X
(F, ω
X
) by the version of Serre
duality given in Theorem 9.0.10, we are done.
We can now prove the toric version of Serre duality.
Proof of Theorem 9.2.10. We ﬁrst show that O
X
Σ
(D) is MCM when D is Q
Cartier. Pick ℓ > 0 such that ℓD is Cartier. We will use splitting methods, though
410 Chapter 9. Sheaf Cohomology of Toric Varieties
for clarity we replace φ
ℓ
: N →N with the sublattice ℓN ⊆ N. The dual lattice
ℓ
−1
M ⊇M gives semigroup algebras S
σ,N
=C[σ
∨
∩M] ⊆S
σ,ℓN
=C[σ
∨
∩ℓ
−1
M].
Over an afﬁne open subset U
σ
, we can pick m
σ
∈ ℓ
−1
M such that ϕ
D
(u) =
¸m
σ
, u) for u ∈ σ. Now consider the S
σ,N
module A = Γ(U
σ
, O
X
Σ
(D)). If we set
P =m
σ
+σ
∨
, then one easily sees that
(9.2.8) A =
m∈P∩M
C χ
m
⊆ B =
m∈P∩ℓ
−1
M
C χ
m
= χ
mσ
S
σ,ℓN
,
where the last equality uses m
σ
∈ ℓ
−1
M. Note that B =χ
mσ
S
σ,ℓN
is free over S
σ,ℓN
and hence is MCM over S
σ,ℓN
. However, S
σ,ℓN
is ﬁnitely generated as a S
σ,N

module (Exercise 9.2.9), so that B is MCM over S
σ,N
by [29, Ex. 1.2.26].
Another property of MCM modules is that any nontrivial direct summand of an
MCM module is again MCM. This follows from [29, Thm. 3.5.7]. Hence it sufﬁces
to split (9.2.8). In this situation, we use the map r : B →A that sends χ
m
∈ B to
r(χ
m
) =
_
1 if χ
m
∈ A
0 otherwise.
Similar to what we did in Proposition 9.2.6, r is a homomorphism of S
σ,N
modules.
From here, it follows easily that O
X
Σ
(D) is MCM when D is QCartier.
Now we prove duality. Since O
X
Σ
(D) is MCM, Theorem 9.2.12 implies that
H
p
(X
Σ
, O
X
Σ
(D))
∨
≃H
n−p
(X
Σ
, Hom
O
X
Σ
(O
X
Σ
(D), ω
X
Σ
))
=H
n−p
(X
Σ
, Hom
O
X
Σ
(O
X
Σ
(D), O
X
Σ
(K
X
Σ
))).
However, Hom
O
X
(O
X
(D), O
X
(E)) ≃ O
X
(E −D) for any Weil divisors D, E on a
normal variety X (Exercise 9.2.10). This gives the desired duality for O
X
Σ
(D).
For part(b), we ﬁrst show that
´
Ω
q
X
Σ
is MCM when X
Σ
is simplicial. Since Σ is
complete, we can work locally over U
σ
, dimσ = n. The minimal generators of σ
form a basis of a sublattice N
′
⊂N of ﬁnite index such that σ is smooth relative to
N
′
. Let M
′
be the dual of N
′
. As above, we get semigroup algebras S
σ,N
⊆S
σ,N
′ .
By Proposition 8.2.18, the restriction of
´
Ω
q
X
Σ
to U
σ
=U
σ,N
is determined by
the S
σ,N
module
A =
m∈σ
∨
∩M
_
q
V
σ
(m) χ
m
,
where V
σ
(m) =Span
C
(m
0
∈ M[ m
0
∈ the minimal face of σ
∨
containing m). Now
consider the larger S
σ,N
module deﬁned by
B =
m∈σ
∨
∩M
′
_
q
V
σ
(m) χ
m
.
Since V
σ
(m) is unaffected when M is replaced by M
′
, we see that as a S
σ,N
′ module,
B =Γ(U
σ,N
′ ,
´
Ω
q
U
σ,N
′
).
§9.2. Vanishing Theorems I 411
However,
´
Ω
q
U
σ,N
′
is locally free since U
σ,N
′ is smooth by our choice of N
′
. Hence
B is MCM over S
σ,N
′ . Arguing as in part (a), B is MCM over S
σ,N
and we have a
splitting map r deﬁned by
r(χ
m
) =
_
χ
m
if m∈ σ
∨
∩M
0 otherwise.
Thus
´
Ω
q
X
Σ
is MCM, and then
´
Ω
q
X
Σ
⊗
O
X
Σ
F is MCM since F is locally free.
The proof is now easy to ﬁnish, since Theorem 9.2.12 implies
H
p
(X
Σ
,
´
Ω
q
X
Σ
⊗
O
X
Σ
F)
∨
≃H
n−p
(X
Σ
, Hom
O
X
Σ
(
´
Ω
q
X
Σ
⊗
O
X
Σ
F, ω
X
Σ
))
However, using the local freeness of F and Exercise 8.0.13, we have
Hom
O
X
Σ
(
´
Ω
q
X
Σ
⊗
O
X
Σ
F, ω
X
Σ
) ≃Hom
O
X
Σ
(
´
Ω
q
X
Σ
, ω
X
Σ
) ⊗
O
X
Σ
F
∨
≃
´
Ω
n−q
X
Σ
⊗
O
X
Σ
F
∨
.
From here, the theorem follows easily.
The proof of part (a) of the theorem was inspired by [42, Lem. 3.4.2]. Other
proofs that QCartier divisors give MCM sheaves can be found in [27, Cor. 4.2.2]
and [143, Prop. 4.22]. In part (b) we followed [42, Prop. 4.8].
We should also mention that besides QCartier divisors, there can be other
Weil divisors that give MCM sheaves. For example, ω
X
is MCM on any normal
CohenMacaulay variety (see [29, Def. 3.3.1]), so that ω
X
Σ
is MCM on any toric
variety X
Σ
, even when the canonical class K
X
Σ
fails to be QCartier. You will give
a toric proof of this in Exercise 9.2.11.
In Exercise 9.2.12 you will use Alexander duality to give a purely toric proof
of Serre duality for simplicial toric varieties.
Exercises for §9.2.
9.2.1. Prove Proposition 9.2.1.
9.2.2. Verify the claims made in Example 9.2.2.
9.2.3. We saw in Example 4.0.25 a morphism of varieties f : X →Y induces a homomor
phism O
Y
→ f
∗
O
X
of O
Y
modules.
(a) Let L be a line bundle on Y. Construct an isomorphism f
∗
f
∗
L ≃ ( f
∗
O
X
) ⊗
OY
L
of O
Y
modules. Hint: Construct a homomorphism and then study the homomorphism
over open subsets of Y where L is trivial.
(b) Generalize part (b) by showing that for any sheaf G of O
X
modules on X and any line
bundle L on Y, there is an isomorphismf
∗
(G ⊗
OX
f
∗
L) ≃ f
∗
G ⊗
OY
L. This is the
projection formula.
9.2.4. Consider the map
m∈P∩M
C χ
m
−→
m∈(ℓP)∩M
C χ
m
from (9.2.1), where the
C[σ
∨
∩M]module structure on
m∈(ℓP)∩M
C χ
m
is given by χ
m
a = χ
ℓm
a. Prove that
(9.2.1) and (9.2.2) are C[σ
∨
∩M]module homomorphisms.
9.2.5. Prove that (9.2.4) implies Z
−D,m
=¦0¦, as claimed in the proof of Theorem 9.2.7.
412 Chapter 9. Sheaf Cohomology of Toric Varieties
9.2.6. As in the proof of Theorem 9.2.7, assume that Z
−D,m
is strongly convex and that u ∈
Z
−D,m
is nonzero. Thus −u ∈ V
supp
−D,m
. Prove that for every v ∈V
supp
−D,m
, the line segment uv
is contained in V
supp
−D,m
(this means that V
supp
−D,m
is star shaped with respect to −u). Conclude
that the constant map γ : V
supp
−D,m
→¦−u¦ is a contraction.
9.2.7. Prove that every afﬁne toric variety U
σ
is contained in a projective toric variety X
Σ
.
9.2.8. The lattice points of a normal lattice polytope P give a projective embedding of the
toric variety X
P
. Prove that X
P
is aCM in this embedding.
9.2.9. Assume that N
1
⊆ N
2
has ﬁnite index, with dual M
2
⊆ M
1
, and let σ be a cone in
(N
1
)
R
=(N
2
)
R
. This gives semigroup algebras S
σ,N2
=C[σ
∨
∩M
2
] ⊆S
σ,N1
=C[σ
∨
∩M
1
].
Prove that S
σ,N1
is ﬁnitely generated as a module over S
σ,N2
. Hint: Let m
1
, . . . , m
r
∈ M
2
generate σ
∨
and consider ¦
r
i=1
δ
i
m
i
[ 0 ≤δ
i
< 1¦ ∩M
1
.
9.2.10. Let D, E be Weil divisors on a normal variety X and let U ⊆X be the smooth locus.
Then Γ(X, O
X
(D)) ≃Γ(U, O
X
(D)[
U
), and the same holds for E.
(a) Construct a natural map Φ
X
: Γ(X, O
X
(E −D)) →Hom
OX
(O
X
(D), O
X
(E)) and prove
that Φ
X
is an isomorphism when X is smooth.
(b) Prove that Hom
OX
(O
X
(D), O
X
(E)) is reﬂexive. Hint: Let U is the smooth locus and
study the restriction map Hom
OX
(O
X
(D), O
X
(E)) →Hom
OU
(O
X
(D)[
U
, O
X
(D)[
U
).
(c) Show that Φ
X
is an isomorphism and that O
X
(E −D) ≃Hom
OX
(O
X
(D), O
X
(E)).
9.2.11. Following [42, Cor. 3.5], you will show that ω
XΣ
is MCM on a normal toric variety.
Fix a cone σ ⊆N
R
and pick a basis e
1
, . . . , e
n
of M such that e
n
is in the interior of σ
∨
. Let
ℓ =lcm(¸e
n
, u
ρ
) [ ρ ∈σ(1)) and let M be the lattice with basis e
1
, . . . , e
n−1
, ℓ
−1
e
n
, with dual
N ⊆N. By Proposition 8.2.9, ω
Uσ
comes fromthe ideal
m∈Int(σ
∨
)∩M
C χ
m
⊆C[σ
∨
∩M].
(a) Prove that u
ρ
= (ℓ/¸e
n
, u
ρ
))u
ρ
lies in N. Then use ¸ℓ
−1
e
n
, u
ρ
) = 1 to show that u
ρ
is
the minimal generator of ρ with respect to N.
(b) Conclude that the canonical divisor of U
σ,N
is Cartier.
(c) Construct a splitting of
m∈Int(σ
∨
)∩M
C χ
m
⊆
m∈Int(σ
∨
)∩M
C χ
m
and conclude that
the canonical sheaf of U
σ,N
is MCM.
9.2.12. Alexander duality [131, §71] states that if A ⊆S
n−1
is a closed subset such that the
pair (S
n−1
, A) is triangulable (see [131, p. 150]), then
¯
H
p−1
(A, C)
∨
≃
¯
H
n−p−1
(S
n−1
¸ A, C).
You will prove Serre duality when D=
ρ
a
ρ
D
ρ
on a complete simplicial toric variety X
Σ
of dimension n. Let K =K
XΣ
be the canonical divisor of X
Σ
. Theorem 9.1.2 implies
H
p
(X
Σ
, O
XΣ
(D))
m
≃
¯
H
p−1
(V
simp
D,m
, C), m ∈ M
Set A
D,m
=V
simp
D,m
and ∆
σ
=Conv(u
ρ
[ ρ ∈ σ(1)¦. Your goal is to prove that
(9.2.9) H
p
(X
Σ
, O
XΣ
(D))
∨
m
≃H
n−p
(X
Σ
, O
XΣ
(K −D))
−m
.
(a) Explain why S =
σ∈Σ
∆
σ
is homeomorphic to S
n−1
. Note that A
D,m
, A
K−D,−m
⊆S.
(b) Prove that A
D,m
∩A
K−D,−m
= ∅ and that for σ ∈ Σ, σ is contained in neither A
D,m
nor A
K−D,−m
if and only if there are ρ
1
, ρ
2
∈ σ(1) such that ¸m, u
ρ1
) < −a
ρ1
and
¸m, u
ρ2
) ≥−a
ρ2
. We say that σ is intermediate when this happens.
§9.3. Vanishing Theorems II 413
(c) Fix an intermediate cone σ ∈ Σ. Let ∆
−
σ
be the face of ∆
σ
generated by u
ρ
’s with
¸m, u
ρ
) < −a
ρ
and ∆
+
σ
be the face generated by u
ρ
’s with ¸m, u
ρ
) ≥ −a
ρ
. Show
that every u ∈ ∆
σ
can be written uniquely as u = (1 −t)u
+
+tu
−
where u
+
∈ ∆
+
σ
,
u
−
∈ ∆
−
σ
and 0 ≤t ≤1. Then show that ∆
+
σ
⊆∆
σ
¸ ∆
−
σ
is a deformation retract.
(d) Prove that A
K−D,−m
⊆S ¸ A
D,m
is a deformation retract.
(e) Finally, prove (9.2.9) by applying Alexander duality to A
D,m
⊆S.
This exercise was inspired by [42, Prop. 7.7.1] and [59, Sec. 4.4].
§9.3. Vanishing Theorems II
Vanishing theorems play an important role in algebraic geometry. A glance at the
index of Lazarsfeld’s twovolume treatise [114] lists vanishing theorems due to
Bogomolov, Demailly, Fujita, GrauertRiemenschneider,
Grifﬁths, KawamataViehweg, Kodaira, Koll´ ar, Le Potier,
Manivel, Miyoka, Nadel, Nakano, and Serre.
We will explore toric versions of several of these results. In some cases, the toric
version is stronger, which is to be expected since toric varieties are so special.
Twisting. If D is a Cartier divisor and F a coherent sheaf on a normal variety X,
then the twist of F by D is the sheaf
F(D) =F ⊗
O
X
O
X
(D).
For example, if K
X
is a canonical divisor on X, then
ω
X
(D) =ω
X
⊗
O
X
O
X
(D) ≃O
X
(K
X
) ⊗
O
X
O
X
(D) ≃O
X
(K
X
+D).
This notation will be used some of the vanishing theorems stated below. However,
when we start working with nonCartier divisors, we will drop the twist notation.
Kodaira and Nakano. Two of the earliest vanishing theorems are due to Kodaira
and Nakano. For an ample divisor D on a smooth projective variety X of dimension
n, Kodaira vanishing asserts that
H
p
(X, ω
X
(D)) =0, p > 0,
and Nakano vanishing states that
H
p
(X, Ω
q
X
(D)) =0, p+q > n.
Nakano’s theorem generalizes Kodaira’s since ω
X
=Ω
n
X
in the smooth case.
In the toric case, we get more vanishing, in what has become known as the
BottSteenbrinkDanilov vanishing theorem.
Theorem 9.3.1 (BottSteenbrinkDanilov Vanishing). Let D be an ample divisor
on a projective toric variety X
Σ
. Then for all p > 0 and q ≥0, we have
H
p
(X
Σ
,
´
Ω
q
X
Σ
(D)) =0.
414 Chapter 9. Sheaf Cohomology of Toric Varieties
Proof. We give a proof due to Fujino [54] that uses the morphism φ
ℓ
: X
Σ
→X
Σ
from Lemma 9.2.6. We can assume that D is torusinvariant with support function
ϕ
D
: N
R
→R. Let L =O
X
Σ
(D) be the associated line bundle.
The morphism ϕ
ℓ
has two key properties. The ﬁrst concerns the pullback
φ
∗
ℓ
O
X
Σ
(D). Proposition 6.1.20 implies that φ
∗
ℓ
O
X
Σ
(D) comes from a divisor whose
support function is ϕ
D
◦φ
ℓ
. Since φ
ℓ
is multiplication by ℓ, ϕ
D
◦φ
ℓ
is the support
function of ℓD. Hence
φ
∗
ℓ
L =φ
∗
ℓ
O
X
Σ
(D) ≃O
X
Σ
(ℓD) ≃O
X
Σ
(D)
⊗ℓ
=L
⊗ℓ
.
The second key property of φ
ℓ
is that there is a split injection
(9.3.1)
´
Ω
q
X
Σ
֒→φ
ℓ∗
´
Ω
q
X
Σ
.
We will assume this for now.
Given these properties of φ
ℓ
, the theorem follows easily. Tensoring (9.3.1) with
L gives a split injection
´
Ω
q
X
Σ
⊗
X
Σ
L ֒→φ
ℓ∗
´
Ω
q
X
Σ
⊗
X
Σ
L.
Since φ
ℓ∗
´
Ω
q
X
Σ
⊗
X
Σ
L ≃ φ
ℓ∗
(
´
Ω
q
X
Σ
⊗
X
Σ
φ
∗
ℓ
L) (this is the projection formula from
Exercise 9.2.3) and φ
∗
ℓ
L ≃L
⊗ℓ
, we obtain a split injection
´
Ω
q
X
Σ
⊗
X
Σ
L ֒→φ
ℓ∗
(
´
Ω
q
X
Σ
⊗
X
Σ
L
⊗ℓ
).
As in the discussion leading up to Lemma 9.2.6, this gives the injection
H
p
(X
Σ
,
´
Ω
q
X
Σ
⊗
X
Σ
L) ֒→H
p
(X
Σ
,
´
Ω
q
X
Σ
⊗
X
Σ
L
⊗ℓ
).
When p >0, the righthand side vanishes for ℓ sufﬁciently large by Serre vanishing
(Theorem 9.0.6). Hence the lefthand side also vanishes for p > 0, which is what
we want.
It remains to prove (9.3.1). If we set D = 0 in the proof of Lemma 9.2.6, we
get a split injection i : O
X
Σ
→φ
ℓ∗
O
X
Σ
. Recall that locally,
• φ
ℓ∗
O
X
Σ
looks like O
X
Σ
, with module structure is given by χ
m
a =χ
ℓm
a.
• i sends χ
m
to χ
ℓm
.
• The splitting r : φ
ℓ∗
O
X
Σ
→O
X
Σ
is deﬁned by
r(χ
m
) =
_
0 m / ∈ ℓM
χ
m
′
m =ℓm
′
, m
′
∈ M.
If we tensor i : O
X
Σ
→φ
ℓ∗
O
X
Σ
with
_
q
M, we get a sheaf homomorphism
(9.3.2)
_
q
M⊗
Z
O
X
Σ
֒→φ
ℓ∗
__
q
M⊗
Z
O
X
Σ
_
together with a splitting map
(9.3.3) φ
ℓ∗
__
q
M⊗
Z
O
X
Σ
_
−→
_
q
M⊗
Z
O
X
Σ
.
Thus (9.3.2) is a split injection.
§9.3. Vanishing Theorems II 415
By Theorem 8.2.16,
´
Ω
q
X
Σ
sits inside
_
q
M⊗
Z
O
X
Σ
, so that φ
ℓ∗
´
Ω
q
X
Σ
is a subsheaf
of φ
ℓ∗
__
q
M⊗
Z
O
X
Σ
_
. These subsheaves relate to (9.3.2) and (9.3.3) as follows.
Over U
σ
, Theorem 8.2.18 implies
Γ(U
σ
,
´
Ω
q
X
Σ
) ≃
m∈σ
∨
∩M
_
q
V
σ
(m) χ
m
⊆
m∈σ
∨
∩M
_
q
M
C
χ
m
=Γ(U
σ
,
_
q
M⊗
Z
O
X
Σ
),
where V
σ
(m) ⊆M
C
is spanned by the lattice points lying in the same minimal face
of σ
∨
as m. Then over U
σ
, we have:
• φ
ℓ∗
´
Ω
q
X
Σ
looks like
´
Ω
q
X
Σ
with module structure χ
m
a =χ
ℓm
a.
• The map (9.3.2) takes
_
q
V
σ
(m) χ
m
to
_
q
V
σ
(m) χ
ℓm
=
_
q
V
σ
(ℓm) χ
ℓm
⊆Γ(U
σ
,
´
Ω
q
X
Σ
)
since V
σ
(m) =V
σ
(ℓm). Thus (9.3.2) induces a map (9.3.1).
• The map (9.3.2) takes
_
q
V
σ
(m) χ
m
to 0 or, when m =ℓm
′
, to
_
q
V
σ
(m) χ
m
′
=
_
q
V
σ
(m
′
) χ
ℓm
′
⊆Γ(U
σ
,
´
Ω
q
X
Σ
)
since V
σ
(m) =V
σ
(m
′
). Thus (9.3.3) induces a map φ
ℓ∗
´
Ω
q
X
Σ
→
´
Ω
q
X
Σ
.
This gives the desired split injection (9.3.1).
Although Theorem 9.3.1 generalizes the vanishing theorems of Kodaira and
Nakano, its name “BottSteenbrinkDanilov” reﬂects the more special vanishing
that happens for projective spaces (Bott [20]), weighted projective spaces (Steen
brink [167]), and projective toric varieties (Danilov [42], stated without proof).
Theorem 9.3.1 was ﬁrst proved by Batyrev and Cox [10] in the simplicial case and
by Buch, Thomsen, Lauritzen and Mehta [31] in general. Further proofs have been
given by Fujino [54] (noted above) and Mustat ¸ˇ a [133].
The Simplicial Case. When X
Σ
is simplicial, there is another vanishing theorem
involving the sheaves
´
Ω
q
X
Σ
.
Theorem 9.3.2. If X
Σ
is a complete simplicial toric variety, then
H
p
(X
Σ
,
´
Ω
q
X
Σ
) =0 for all p ,=q.
Proof. Since X
Σ
is simplicial, we have the exact sequence
(9.3.4) 0 −→
´
Ω
q
X
Σ
−→K
0
(Σ, q) −→K
1
(Σ, q) −→ −→K
q
(Σ, q) −→0
from Theorem 8.2.19, where
K
j
(Σ, q) =
σ∈Σ( j)
_
q−j
(σ
⊥
∩M) ⊗
Z
O
V(σ)
and V(σ) =O(σ) ⊆X
Σ
is the orbit closure corresponding to σ ∈ Σ. Since V(σ) is
a toric variety, its structure sheaf has vanishing higher cohomology by Demazure
vanishing, so that H
p
(X
Σ
, K
j
(Σ, q)) = 0 for p > 0. Since the above sequence has
416 Chapter 9. Sheaf Cohomology of Toric Varieties
q +1 terms with vanishing higher cohomology, an easy argument using the long
exact sequence in cohomology implies H
p
(X
Σ
,
´
Ω
q
X
Σ
) =0 for p >q (Exercise 9.3.1).
Now suppose p < q. Since X
Σ
is simplicial, the version of Serre duality given
in Theorem 9.2.10 implies
H
p
(X
Σ
,
´
Ω
q
X
Σ
)
∨
≃H
n−p
(X
Σ
,
´
Ω
n−q
X
Σ
)
The righthand side vanishes since n−p > n−q, and the result follows.
When X
Σ
complete but not necessarily simplicial, Danilov [42, Cor. 12.7]
proves that H
p
(X
Σ
,
´
Ω
q
X
Σ
) =0 when q > p.
In [121], Mavlyutov discusses a vanishing theorem for
´
Ω
q
X
Σ
(D), where D is a
nef Cartier divisor. His result goes as follows.
Theorem 9.3.3. Let X
Σ
be a complete simplicial toric variety. If D is a nef Cartier
divisor on X
Σ
, then
H
p
(X
Σ
,
´
Ω
q
X
Σ
(D)) =0
whenever p > q or q > p+dimP
D
.
The proof for p >q is relatively easy (Exercise 9.3.2). Note that Theorem 9.3.2
is the case D = 0 of Theorem 9.3.3. The paper [121] computes H
p
(X
Σ
,
´
Ω
q
X
Σ
(D))
explicitly for all p, q.
QWeil Divisors. AQWeil divisor or D on a normal variety X is a formal Qlinear
combination of prime divisors. Thus a positive integer multiple of D is an ordinary
Weil divisor, often called integral in this context. A QWeil divisor is QCartier is
some positive multiple is integral and Cartier. In the literature (see [114, 1.1.4]),
QCartier QWeil divisors are often called Qdivisors. These divisors and their
close cousins, Rdivisors, are essential tools in modern algebraic geometry.
For x ∈ R, ⌊x⌋ is the greatest integer ≤x and ⌈x⌉ is the least integer ≥x. Then,
given a QWeil divisor D =
i
a
i
D
i
, we get integral Weil divisors
⌊D⌋ =
i
⌊a
i
⌋D
i
(the “round down” of D)
⌈D⌉ =
i
⌈a
i
⌉D
i
(the “round up” of D).
We now prove an injectivity lemma for QWeil divisors due to Fujino [54].
Lemma 9.3.4. Let D be a QWeil divisor on a toric variety X
Σ
and let ℓ > 0 be an
integer such that ℓD is integral. Then for all p ≥0 there is an injection
H
p
(X
Σ
, O
X
Σ
(⌊D⌋)) ֒→H
p
(X
Σ
, O
X
Σ
(ℓD)).
Proof. We adapt the proof of Lemma 9.2.6 to our situation. Write D =
ρ
a
ρ
D
ρ
,
where a
ρ
= b
ρ
+ε
ρ
for b
ρ
∈ Z and 0 ≤ ε
ρ
< 1. Thus ⌊D⌋ =
ρ
b
ρ
D
ρ
. With ℓ as
above, we claim that φ
ℓ
: X
Σ
→X
Σ
from Lemma 9.2.6 gives a split injection
(9.3.5) O
X
Σ
(⌊D⌋) ֒→φ
ℓ∗
O
X
Σ
(ℓD).
§9.3. Vanishing Theorems II 417
Assuming (9.3.5), lemma follows from the discussion leading up to Lemma 9.2.6.
It remains to prove the existence of a split injection (9.3.5). Take an afﬁne open
subset U
σ
⊆X
Σ
and consider
P
ℓ
=¦m ∈ M
R
[ ¸m, u
ρ
) ≥−ℓa
ρ
for all ρ ∈ σ(1)¦.
Then
Γ(U
σ
, O
X
Σ
(⌊D⌋)) =
m∈P
1
∩M
C χ
m
, Γ(U
σ
, O
X
Σ
(ℓD)) =
m∈P
ℓ
∩M
C χ
m
,
where the ﬁrst equality uses a
ρ
=b
ρ
+ε
ρ
, 0 ≤ε
ρ
< 1. Since P
ℓ
∩ℓM =ℓ(P
1
∩M),
the map m →ℓm induces an inclusion
m∈P
1
∩M
C χ
m
֒→
m∈P
ℓ
∩M
C χ
m
.
This is a C[σ
∨
∩M]module homomorphism, provided that the righthand side has
module structure χ
m
a = χ
ℓm
a, and the usual formula for the splitting map r is
also a C[σ
∨
∩M]module homomorphism. From here, we get the required split
injection (9.3.5) without difﬁculty.
Here are Qdivisor versions of Demazure and BatyrevBorisov vanishing.
Theorem 9.3.5. Let D be a QCartier QWeil divisor on a toric variety X
Σ
.
(a) If [Σ[ is convex and D is nef, then
H
p
(X
Σ
, O
X
Σ
(⌊D⌋)) =0 for all p > 0.
(b) If Σ is complete and D is nef, then
H
p
(X
Σ
, O
X
Σ
(−⌈D⌉)) =0 for all p ,=dimP
D
.
Proof. Pick ℓ > 0 with ℓD Cartier. For part (a), H
p
(X
Σ
, O
X
Σ
(ℓD)) = 0 for p > 0
by Theorem 9.2.3. The desired vanishing follows immediately from Lemma 9.3.4.
For part (b), replacing D with −D in Lemma 9.3.4 gives
H
p
(X
Σ
, O
X
Σ
(−⌈D⌉)) =H
p
(X
Σ
, O
X
Σ
(⌊−D⌋)) ֒→H
p
(X
Σ
, O
X
Σ
(−ℓD)),
and then we are done by Theorem 9.2.7.
Here is an example that show how part (a) of Theorem 9.3.5 extends Demazure
vanishing beyond nef QCartier divisors.
Example 9.3.6. The fan for the Hirzebruch surface H
r
has minimal generators
u
1
= (−1, r), u
2
= (0, 1), u
3
= (1, 0), u
4
= (0, −1), giving divisors D
1
, . . . , D
4
. By
Example 6.1.17,
Pic(H
r
)
R
≃¦aD
3
+bD
4
[ a, b ∈ R¦
and the nef cone is generated by D
3
and D
4
. Since D
1
∼D
3
and D
2
∼D
4
−rD
3
, it
follows that a QWeil divisor D =a
1
D
1
+ +a
4
D
4
is nef if and only
(9.3.6) a
1
+a
3
≥ra
2
and a
2
+a
4
≥0.
418 Chapter 9. Sheaf Cohomology of Toric Varieties
We will show that the divisors aD
3
+bD
4
, a, b ≥ −1, have vanishing higher
cohomology when r > 0. Given such a divisor, pick a rational number 0 < ε <
1
2
and consider the QWeil divisor
D =2εD
1
+
ε
r
D
2
+(a+1−ε)D
3
+(b+1−
ε
r
)D
4
.
This satisﬁes (9.3.6) and hence is nef. Then ⌊D⌋ =aD
3
+bD
4
has vanishing higher
cohomology by Theorem 9.3.5. Taking a =−1 or b =−1, we get nonnef divisors
whose higher cohomology vanishes. ♦
Lemma 9.3.4 also leads to the following result due to Mustat ¸ˇ a [133].
Theorem 9.3.7. Let X
Σ
be a projective toric variety and let ρ
1
, . . . , ρ
r
∈ Σ(1) be
distinct. Then for p > 0 and any ample Cartier divisor D on X
Σ
, we have
H
p
(X
Σ
, O
X
Σ
(D−D
ρ
1
− −D
ρr
)) =0.
Proof. Let B =D
ρ
1
+ +D
ρr
. Since D is ample, Serre vanishing (Theorem 9.0.6)
implies that we can ﬁnd ℓ > 0 such that H
p
(X
Σ
, X
Σ
, O
X
Σ
(ℓD−B)) = 0 for p > 0.
Now let E =D−ℓ
−1
B. Since ⌊E⌋ =D−B, the inclusion
H
p
(X
Σ
, O
X
Σ
(⌊E⌋)) ֒→H
p
(X
Σ
, O
X
Σ
(ℓE))
from Lemma 9.3.4 implies
H
p
(X
Σ
, O
X
Σ
(D−B)) ֒→H
p
(X
Σ
, O
X
Σ
(ℓD−B)) =0, p > 0.
Here is a nonQCartier divisor with vanishing higher cohomology.
Example 9.3.8. Consider the complete fan Σ in R
3
shown in Figure 5. It has
z
y
x
ρ
0
ρ
1
ρ
3
ρ
4
ρ
2
Figure 5. A complete fan in R
3
divisors D
0
, . . . , D
4
corresponding to ρ
0
, . . . , ρ
4
. In Exercise 9.3.3 you will show
that for a QWeil divisor D =a
0
D
0
+ +a
4
D
4
, we have:
§9.3. Vanishing Theorems II 419
• D is QCartier if and only a
1
+a
3
=a
2
+a
4
.
• D is QCartier and nef if and only a
1
+a
3
=a
2
+a
4
≥0.
In particular, D = D
3
+D
4
is QCartier and nef, so that D
4
= D−D
3
has
vanishing higher cohomology by Theorem 9.3.7. Yet D
4
is not QCartier. ♦
Iitaka Dimension and Big Divisors. Given a nef Cartier divisor D on a complete
toric variety X
Σ
and an integer ℓ > 0, the global sections W
ℓ
= H
0
(X
Σ
, O
X
Σ
(ℓD))
give a morphism φ
W
ℓ
: X
Σ
→P(W
∨
ℓ
) as in Lemma 6.0.28. Also recall that since D
is nef, its polytope P
D
is a lattice polytope. The map φ
W
ℓ
and the polytope ℓP
D
are
related as follows.
Lemma 9.3.9. For ℓ ≫0, the image of φ
W
ℓ
is isomorphic to the toric variety of
ℓP
D
. In particular, dimφ
W
ℓ
(X
Σ
) =dimP
D
for ℓ ≫0.
Proof. This follows from Proposition 6.1.22 (Exercise 9.3.4).
This situation is a special case of the deﬁnition of the Iitaka dimension κ(X, D)
(see [114, Def. 2.1.3]) of a Cartier divisor D on a complete irreducible variety X. In
this terminology, Lemma 9.3.9 implies that a nef Cartier divisor D on a complete
toric variety X
Σ
has Iitaka dimension
κ(X
Σ
, D) =dimP
D
.
In general, a Cartier divisor is big if it has maximal Iitaka dimension, which for a
nef Cartier divisor D on a complete toric variety X
Σ
means
D is big ⇐⇒ dimP
D
=dimX
Σ
.
It should be clear what it means for a QCartier QWeil divisor to be big and nef.
KawamataViehweg. The classic version of Kodaira vanishing can be stated as
H
p
(X, O
X
(K
X
+D)) =0 for all p > 0
when D is an ample line bundle on a smooth projective variety X. In the 1982,
Kawamata and Viehweg independently weakened the hypotheses on D. Here is the
toric version of their result.
Theorem 9.3.10 (Toric KawamataViehweg). Let X
Σ
be a complete toric variety
and let D be a QCartier QWeil divisor on X
Σ
that is big and nef. Then
H
p
(X
Σ
, O
X
Σ
(K
X
Σ
+⌈D⌉)) =0 for all p > 0.
Proof. Let n =dimX
Σ
. When D is Cartier, Serre duality implies
H
p
(X
Σ
, O
X
Σ
(K
X
Σ
+D))
∨
≃H
n−p
(X
Σ
, O
X
Σ
(−D)).
Since D is nef, the latter vanishes for n−p ,=dimP
D
by Theorem 9.2.7, and since
D is big, the condition on p becomes n−p ,=n, i.e, p ,=0.
420 Chapter 9. Sheaf Cohomology of Toric Varieties
In general, write D=
ρ
a
ρ
D
ρ
where a
ρ
=b
ρ
−ε
ρ
for b
ρ
∈ Z and 0 ≤ε
ρ
< 1.
Pick ℓ such that ℓD is Cartier and 0 < ε
ρ
+ℓ
−1
< 1 for all ρ. Let E =D+ℓ
−1
K
X
Σ
.
Since ⌊b
ρ
−ε
ρ
−ℓ
−1
⌋ =b
ρ
−1, we have ⌊E⌋ =⌈D⌉ +K
X
Σ
. Thus the inclusion
H
p
(X
Σ
, O
X
Σ
(⌊E⌋)) ֒→H
p
(X
Σ
, O
X
Σ
(ℓE))
from Lemma 9.3.4 implies
H
p
(X
Σ
, O
X
Σ
(K
X
Σ
+⌈D⌉)) ֒→H
p
(X
Σ
, O
X
Σ
(K
X
Σ
+ℓD)).
Then we are done by the Cartier case already proved.
Another approach to Theorem 9.3.10 is to prove a version of Serre duality that
implies H
p
(X
Σ
, O
X
Σ
(K
X
Σ
+⌈D⌉))
∨
≃H
n−p
(X
Σ
, O
X
Σ
(−⌈D⌉)) (Exercise 9.3.5).
GrauertRiemenschneider. Our next vanishing result features the higher direct
images of the canonical sheaf. We will need the following preliminary result.
Proposition 9.3.11. Let Σbe a simplicial fan whose support [Σ[ is strongly convex.
Then H
p
(X
Σ
, ω
X
Σ
) =0 for all p > 0.
Proof. Fix p ≥1 and m ∈ M. Theorem 9.1.2 with D =K
X
Σ
=−
ρ
D
ρ
implies
H
p
(X
Σ
, ω
X
Σ
)
m
=H
p
(X
Σ
, O
X
Σ
(D))
m
≃
¯
H
p−1
(V
simp
D,m
, C),
where V
simp
D,m
=
σ∈Σ
D,m
Conv(u
ρ
[ ρ ∈ σ(1)) and Σ
D,m
= ¦σ ∈ Σ [ ¸m, u
ρ
) < 1 for
all ρ ∈ σ(1)¦. Consider the set
W =¦u ∈ [Σ[ ¸¦0¦ [ ¸m, u) ≤0¦ =([Σ[ ¸¦0¦) ∩¦u ∈ N
R
[ ¸m, u) ≤0¦,
and observe that W is convex since [Σ[ is strongly convex. Note also that Σ
D,m
=
¦σ ∈ Σ [ ¸m, u
ρ
) ≤0 for ρ ∈ σ(1)¦ since ¸m, u
ρ
) ∈ Z. Thus
(9.3.7) V
simp
D,m
⊆W.
The proposition will follow once we prove that (9.3.7) is a deformation retract.
Given σ ∈ Σ with W ∩σ ,=∅, we construct r
σ
: W ∩σ →V
simp
D,m
∩σ as follows.
The assumption W ∩σ ,=∅ implies that
A =¦ρ ∈ σ(1) [ ¸m, u
ρ
) ≤0¦ , =∅.
Since σ is simplicial, u ∈σ can be uniquely written u =
ρ∈σ(1)
λ
ρ
u
ρ
, λ
ρ
≥0, and
when u ∈ W ∩σ, deﬁne r
σ
(u) = (
ρ∈A
λ
ρ
)
−1
ρ∈A
λ
ρ
u
ρ
. The sum
ρ∈A
λ
ρ
is
nonzero since u ∈W ∩σ, so that r
σ
(u) ∈ Conv(u
ρ
[ ρ ∈ A) ⊆V
simp
D,m
. It is also easy
to see that r
σ
is compatible with r
τ
whenever τ is a face of σ. Thus we get a map
r : W →V
simp
D,m
, which is the desired retraction by Exercise 9.3.6.
Here is a toric version of GrauertRiemenschneider vanishing.
§9.3. Vanishing Theorems II 421
Theorem 9.3.12 (Toric GrauertRiemenschneider Vanishing). Let φ : X
Σ
→X
Σ
′ be
a surjective proper toric morphism between toric varieties of the same dimension.
If X
Σ
is simplicial, then
R
p
φ
∗
ω
X
Σ
=0 for all p > 0.
If in addition φ is birational, then φ
∗
ω
X
Σ
≃ω
X
Σ
′
.
Proof. Our hypothesis on φ implies that the associated lattice map φ : N →N
′
induces an isomorphism φ
R
: N
R
≃ N
′
R
such that Σ is the inverse image of Σ
′
.
Thus, given σ ∈ Σ
′
, we see that φ
−1
R
(σ) is strongly convex. This is the support of
the fan of φ
−1
(U
σ
), so that
H
p
(φ
−1
(U
σ
), ω
X
Σ
) =0
for p >0 by Proposition 9.3.11. Then R
p
φ
∗
ω
X
Σ
=0 for p >0 by Proposition 9.0.7.
The ﬁnal assertion of the theorem is an easy consequence of Theorem 8.2.15
(Exercise 9.3.7).
Most versions of Theorem 9.3.12 in the literature assume that X
Σ
is smooth.
Other Vanishing Theorems. There are many more toric vanishing theorems. Both
Fujino [54] and Mustat ¸ˇ a [133] state general vanishing theorems that imply versions
of Theorems 9.3.1, 9.3.5, 9.3.7 and 9.3.10. Further vanishing results can be found
in [58] and [143].
Exercises for §9.3.
9.3.1. A sheaf G on a variety X is acyclic if H
p
(X, G) =0 for all p >0. Now suppose that
we have an exact sequence of sheaves
0 −→F −→G
0
−→G
1
−→ −→G
r
−→0,
where G
0
, . . . , G
r
are acyclic. Prove H
p
(X, F) =0 for all p > r by induction on r.
9.3.2. Use (9.3.4) and the previous exercise to prove Theorem 9.3.3 when p > q.
9.3.3. Let D
0
, . . . , D
4
be the divisors from Example 9.3.8 and consider a QWeil divisor
D =a
0
D
0
+ +a
4
D
4
.
(a) Show that D
0
∼2D
3
+2D
4
, D
1
∼D
3
, D
2
∼D
4
.
(b) Show that D is QCartier if and only a
1
+a
3
= a
2
+a
4
.
(c) Show that D is QCartier and nef if and only a
1
+a
3
=a
2
+a
4
≥0.
9.3.4. Complete the proof of Lemma 9.3.9.
9.3.5. Let D be a QCartier QWeil divisor on a complete toric variety X
Σ
.
(a) Pick ℓ > 0 such ℓD is Cartier and consider O
XΣ
(⌊D⌋) ֒→φ
ℓ∗
O
XΣ
(ℓD) from (9.3.5).
Use this to prove that O
XΣ
(⌊D⌋) is MCM. Hint: Replace φ
ℓ
: N →N with ℓN ⊆ N as
in the proof of Theorem 9.2.10.
422 Chapter 9. Sheaf Cohomology of Toric Varieties
(b) Adapt the proof of Theorem 9.2.10 to show that
H
p
(X
Σ
, O
XΣ
(⌊D⌋))
∨
≃H
n−p
(X
Σ
, O
XΣ
(K
XΣ
−⌊D⌋)).
(c) Prove H
p
(X
Σ
, O
XΣ
(−⌈D⌉))
∨
≃H
n−p
(X
Σ
, O
XΣ
(K
XΣ
+⌈D⌉)). Hint: ⌊−D⌋ =−⌈D⌉.
(d) Prove Theorem 9.3.10 using part (c) and Theorem 9.3.5.
9.3.6. Consider the map r
σ
: W∩σ →V
simp
D,m
∩σ deﬁned in the proof of Proposition 9.3.11.
(a) Prove that these maps patch to give a retraction r : W →V
simp
D,m
.
(b) Prove that when regraded as a map fromW∩σ to itself, r
σ
is homotopic to the identity.
Then formulate and prove a similar result for r.
9.3.7. Let φ : X
Σ
→X
Σ
′ be a proper birational toric morphism.
(a) Prove that φ : N →N
′
is an isomorphism.
(b) Let Σ
0
be the fan in N
R
consisting of the cones φ
−1
R
(σ) for σ ∈ Σ
′
. Prove that Σ is a
reﬁnement of Σ
0
.
(c) Complete the proof of Theorem 9.3.12.
9.3.8. Given a toric variety X
Σ
, let D be a Weil divisor and E =
ρ
a
ρ
D
ρ
a QWeil divisor
such that 0 ≤a
ρ
≤1 for all ρ and ℓE is integral for some integer ℓ > 0. Prove that there is
an injection
H
p
(X
Σ
, O
XΣ
(D)) ֒→H
p
(X
Σ
, O
XΣ
(D+ℓ(D+E))) for all p ≥0.
This is a strengthened version of Theorem 0.1 from [133]. Hint: As suggested by [54],
apply Lemma 9.3.4 to the QWeil divisor D+
ℓ
ℓ+1
E.
§9.4. Applications to Lattice Polytopes
In this section we use vanishing theorems to study lattice polytopes.
The Euler Characteristic of a Sheaf . Let F be a coherent sheaf on a complete
variety X. Its Euler characteristic χ(F) is deﬁned to be the alternating sum
χ(F) =
p≥0
(−1)
p
dimH
p
(X, F).
Our hypotheses on X and F guarantee that dimH
p
(X, F) < ∞ for all p, and
H
p
(X, F) =0 for p > dimX by [78, Thm. III.2.7]. Hence χ(F) is a welldeﬁned
integer. The Euler characteristic satisﬁes
(9.4.1) χ(G) =χ(F) +χ(H )
whenever we have an exact sequence 0 →F →G →H →0 of coherent sheaves.
This follows from the long exact sequence in cohomology (Exercise 9.4.1).
§9.4. Applications to Lattice Polytopes 423
Given a line bundle L on X, deﬁne
L
⊗ℓ
=
_
¸
¸
¸
_
¸
¸
¸
_
L ⊗
O
X
⊗
O
X
L
. ¸¸ .
ℓ times
ℓ > 0
(L
∨
)
⊗(−ℓ)
ℓ < 0
O
X
ℓ =0,
where as usual L
∨
=Hom
O
X
(L, O
X
). A basic result [105] states that
χ(F ⊗
O
X
L
⊗ℓ
), ℓ ∈ Z,
is a polynomial in ℓ, called the Hilbert polynomial. Exercises 9.4.2 and 9.4.3 sketch
a proof in the ample case. When L = O(D) for a Cartier divisor D, the Hilbert
polynomial is written
χ(F(ℓD)), ℓ ∈ Z,
via the twisting convention introduced in §9.2.
Example 9.4.1. We will compute χ(O
P
n (ℓ)). When ℓ ≥0, Example 9.2.4 implies
that H
p
(P
n
, O
P
n (ℓ)) =0 for p > 0, so that
(9.4.2) χ(O
P
n (ℓ)) =dim H
0
(P
n
, O
P
n (ℓ)) =[ℓ∆
n
∩Z
n
[ =
_
ℓ+n
n
_
.
The second equality follows from Proposition 4.3.3 since the polytope associated
to ℓD
0
is ℓ∆
n
, where ∆
n
is the standard nsimplex from Example 4.3.6. You will
prove the last equality in Exercise 9.4.4. When ℓ < 0, Example 9.2.8 implies
(9.4.3)
χ(O
P
n (ℓ)) =(−1)
n
dim H
n
(P
n
, O
P
n (ℓ)) =(−1)
n
[Int(ℓ∆
n
) ∩Z
n
[
=(−1)
n
_
−ℓ−1
n
_
,
where the last equality uses Exercise 9.4.4.
Now consider the polynomial
p(x) =
(x +n)(x +n−1) (x +1)
n!
∈ Q[x]
and observe that p(ℓ) =
_
ℓ+n
n
_
when ℓ ∈ N.
We claim that p(ℓ) = χ(O
P
n (ℓ)) for all ℓ ∈ Z. For ℓ ≥ 0 this follows easily
from (9.4.2). When ℓ < 0, note that
p(ℓ) =(−1)
n
_
−ℓ−1
n
_
(Exercise 9.4.4). Then p(ℓ) =χ(O
P
n (ℓ)) for ℓ < 0 by (9.4.3).
Note also that when ℓ > 0, (9.4.2) and (9.4.3) imply that
p(ℓ) =[ℓ∆
n
∩Z
n
[
p(−ℓ) =(−1)
n
[Int(ℓ∆
n
) ∩Z
n
[.
We will see below that this is a special case of Ehrhart reciprocity. ♦
424 Chapter 9. Sheaf Cohomology of Toric Varieties
The Ehrhart Polynomial. Given a full dimensional lattice polytope P ⊆ M
R
, the
functions
L(P) =[P∩M[
L
∗
(P) =[Int(P) ∩M[
count lattice points in P or in its interior. For example, if ℓ > 0 is an integer, then
(9.4.2) and (9.4.3) imply that L(ℓ∆
n
) =
_
ℓ+n
n
_
and L
∗
(ℓ∆
n
) =
_
ℓ−1
n
_
.
Theorem 9.4.2 (Ehrhart Reciprocity). Let P ⊆ M
R
≃ R
n
be a full dimensional
lattice polytope. There is a polynomial Ehr
P
(x) ∈ Q[x] such that if ℓ ∈ N, then
Ehr
P
(ℓ) =L(ℓP).
Furthermore, if ℓ ∈ N is positive, then
Ehr
P
(−ℓ) =(−1)
n
L
∗
(ℓP).
Proof. Let X
P
be the toric variety of P and D
P
the associated ample divisor. We
will show that the Hilbert polynomial
Ehr
P
(ℓ) =χ(O
X
P
(ℓD
P
))
has the desired properties. The case when ℓ ≥0 is easy since ℓD
P
is basepoint free
by Proposition 6.1.4. Thus
χ(O
X
P
(ℓD
P
)) =dimH
0
(X
P
, O
X
P
(ℓD
P
)) =[ℓP∩M[ =L(ℓP)
by Demazure vanishing (Theorem 9.2.3) and Example 4.3.7.
Now assume ℓ > 0. Then P
ℓD
P
=ℓP is full dimensional, hence
χ(O
X
P
(−ℓD
P
)) =(−1)
n
dimH
n
(X
P
, O
X
P
(−ℓD
P
)) =(−1)
n
[Int(ℓP) ∩M[
=(−1)
n
L
∗
(ℓP)
by BatyrevBorisov vanishing (Theorem 9.2.7).
The polynomial Ehr
P
in Theorem 9.4.2 is called the Ehrhart polynomial of P.
An elementary approach to Ehrhart polynomials and Ehrhart Reciprocity can be
found in [12], and the program Normaliz [30] computes Ehrhart polynomials.
When P is very ample, the Ehrhart polynomial has a nice interpretation. The
very ample divisor D
P
on X
P
gives a projective embedding i : X
P
֒→P
s−1
, where
s = [P∩M[. Its homogeneous ideal I(X
P
) ⊆ C[x
1
, . . . , x
s
] gives the homogeneous
coordinate ring
C[X
P
] =C[x
1
, . . . , x
s
]/I(X
P
).
This graded ring has a Hilbert function
ℓ −→dimC[X
P
]
ℓ
, ℓ ≥0,
which is a polynomial for ℓ ≫0 (see [37, Ch. 6, §4]). This is the Hilbert polynomial
of X
P
⊆P
s−1
,
§9.4. Applications to Lattice Polytopes 425
Proposition 9.4.3. If P is a very ample, then the Ehrhart polynomial Ehr
P
equals
the Hilbert polynomial of the toric variety X
P
under the projective embedding given
by the very ample divisor D
P
.
Proof. Consider the exact sequence
0 −→I
X
P
−→O
P
s−1 −→O
X
P
−→0.
Since O
X
P
(D
P
) is the restriction of O
P
s−1 (1) to X
P
, tensoring with O
P
s−1 (ℓ) gives
an exact sequence
0 −→I
X
P
(ℓ) −→O
P
s−1 (ℓ) −→O
X
P
(ℓD
P
) −→0.
In Exercise 9.4.5 you will show that the resulting long exact sequence is
0 →I(X
P
)
ℓ
→C[x
1
, . . . , x
s
]
ℓ
→H
0
(X
P
, O
X
P
(ℓD
P
)) →H
1
(P
s−1
, I
X
P
(ℓ)) → .
The H
1
term vanishes for ℓ ≫0 by Serre vanishing (Theorem 9.0.6). Hence for ℓ
large, we get an isomorphism
C[X
P
]
ℓ
≃H
0
(X
P
, O
X
P
(ℓD
P
)).
This implies that the Hilbert polynomial of X
P
is the Ehrhart polynomial of P.
We can describe the degree and leading coefﬁcient of the Ehrhart polynomial.
For the leading coefﬁcient, we use the normalized volume in M
R
. Let e
1
, . . . , e
n
be a basis of M and consider the simplex ∆
n
= Conv(0, e
1
, . . . , e
n
) ⊆ M
R
. Then
the normalized volume is the usual ndimensional Lebesgue measure, scaled so
that ∆
n
has volume equal to 1. Thus the “unit cube” ¦
n
i=1
λ
i
e
i
[ 0 ≤ λ
i
≤ 1¦ has
normalized volume n!.
Given a full dimensional lattice polytope P ⊆ M
R
, its normalized volume is
denoted Vol(P) and is computed by the limit
(9.4.4)
Vol(P)
n!
= lim
ℓ→∞
L(ℓP)
ℓ
n
.
This is proved in many places, including [12, Lem. 3.19]. Since L(ℓP) is given
by the polynomial Ehr
P
(ℓ), it follows easily that Ehr
P
has degree n and its leading
coefﬁcient Vol(P)/n! (Exercise 9.4.6).
Here is a classic application of these ideas.
Example 9.4.4. Let P ⊂ R
2
be a lattice polygon with Ehrhart polynomial Ehr
P
.
The leading coefﬁcient of Ehr
P
is
1
2
Vol(P), which is the usual Euclidean area
Area(P) since the normalized volume of the unit square is 2. The constant term is
also easy to compute, since Ehr
P
(0) =L(0 P) =1 by Theorem 9.4.2. Thus
Ehr
P
(x) =Area(P)x
2
+
1
2
Bx +1,
where B is yet to be determined. The reason for the
1
2
will soon become clear.
426 Chapter 9. Sheaf Cohomology of Toric Varieties
By Ehrhart reciprocity, we have
(9.4.5)
Area(P) +
1
2
B+1 =Ehr
P
(1) =L(P)
Area(P) −
1
2
B+1 =Ehr
P
(−1) =(−1)
2
L
∗
(P) =L
∗
(P).
Solving for B gives B =L(P) −L
∗
(P) =[∂P∩M[, where ∂P is the boundary of P.
Thus the Ehrhart polynomial of P is
Ehr
P
(x) =Area(P)x
2
+
1
2
[∂P∩M[ x +1.
Furthermore, solving the bottom equation of (9.4.5) for the area gives
Area(P) =L
∗
(P) +
1
2
[∂P∩M[ −1.
This equation, called Pick’s formula, shows how to compute the area of a lattice
polygon in terms of its lattice points. ♦
In Example 9.4.4, we noted that Ehr
P
(0) = L(0 P) = 1. The latter equality
is obvious since 0 P = ¦0¦. However, the equality Ehr
P
(0) = L(0 P) (proved in
Theorem 9.4.2) is more subtle since it can fail when we replace the polytope P with
a polytopal complex (a collection of polytopes where the intersection of any two is
a face of each). Here is a simple example.
Example 9.4.5. Consider the boundary ∂∆
2
of the 2simplex ∆
2
⊆R
2
. Thus ∂∆
2
consists of three line segments. One easily computes that
L(ℓ∂∆
2
) =[(ℓ∂∆
2
) ∩Z
2
[ =3ℓ
when ℓ >0 is an integer. This gives the “Ehrhart polynomial” Ehr
∂∆
2
(x) =3x with
the property that Ehr
∂∆
2
(ℓ) =L(ℓ∂∆
2
) for integers ℓ > 0. However,
Ehr
∂∆
2
(0) =3 0 =0 ,=L(0 ∂∆
2
) =L(¦0¦) =1.
Further examples (and an explanation) will be given in Execise 9.4.7. ♦
The pEhrhart Polynomials. Following Materov [118], we use the sheaves
´
Ω
p
X
P
to generalize the Ehrhart polynomial when the lattice polytope P is simple. Recall
that a full dimensional lattice polytope P ⊆M
R
≃R
n
is simple if every vertex is the
intersection of exactly n facets. Hence each vertex has n facet normals that gener
ate the corresponding cone in the normal fan Σ
P
. It follows that Σ
P
is simplicial
whenever P is simple. Hence the toric variety X
P
is simplicial.
We proved earlier that the Ehrhart polynomial Ehr
P
(x) of P satisﬁes
Ehr
P
(ℓ) =χ(O
X
P
(ℓD
P
)), ℓ ∈ Z,
where D
P
is the ample divisor coming from P. More generally, given an integer
0 ≤ p ≤ n, the Euler characteristic χ(
´
Ω
p
X
P
(ℓD
P
)) is a polynomial function of ℓ.
Then the pEhrhart polynomial of P, denoted Ehr
p
P
(x), is the unique polynomial in
Q[x] that satisﬁes
Ehr
p
P
(ℓ) =χ(
´
Ω
p
X
P
(ℓD
P
)), ℓ ∈ Z.
§9.4. Applications to Lattice Polytopes 427
Note that Ehr
0
P
(x) is the ordinary Ehrhart polynomial Ehr
P
(x). To state the
properties of Ehr
p
P
(x), we will use the following notation:
P(i) =¦Q [ Q is an idimensional face of P¦
f
i
=[P(i)[ =#idimensional faces of P
h
p
=
n
i=p
(−1)
i−p
_
i
p
_
f
i
.
The f
i
are the face numbers of P. Furthermore, if Q is a face of P, we deﬁne
L(Q) =[Q∩M[
L
∗
(Q) =[Relint(Q) ∩M[.
These invariants are related to the pEhrhart polynomials by the following result of
Materov [118].
Theorem 9.4.6. Let X
P
be the toric variety of a full dimensional simple lattice
polytope P ⊆M
R
≃R
n
.
(a) If ℓ > 0 is an integer, then
Ehr
p
P
(ℓ) =
n
i=p
_
i
p
_
Q∈P(i)
L
∗
(ℓQ).
(b) If ℓ ≥0 is an integer, then
Ehr
p
P
(ℓ) =
p
i=0
(−1)
i
_
n−i
n−p
_
Q∈P(n−i)
L(ℓQ).
(c) If 0 ≤ p ≤n, then
Ehr
p
P
(−x) =(−1)
n
Ehr
n−p
P
(x).
(d) If 0 ≤ p ≤n, then
Ehr
p
P
(0) =(−1)
p
h
p
.
We will defer the proof for now. Theorem 9.4.6 has some nice consequences.
First, setting x =0 in part (c) and using part (d), we see that
(−1)
p
h
p
=(−1)
n
(−1)
n−p
h
n−p
.
This proves that
(9.4.6) h
p
=h
n−p
for 0 ≤ p ≤n. These are called the DehnSommerville equations.
428 Chapter 9. Sheaf Cohomology of Toric Varieties
Also, if we write Ehr
p
P
(ℓ) using part (b) of the theorem and Ehr
n−p
P
(ℓ) using
part (a), then part (c) gives the following formulas for ℓ > 0:
Ehr
p
P
(ℓ) =
p
i=0
(−1)
i
_
n−i
n−p
_
Q∈P(n−i)
L(ℓQ),
Ehr
p
P
(−ℓ) =(−1)
n
Ehr
n−p
P
(ℓ) =(−1)
n
n
i=n−p
_
i
n−p
_
Q∈P(i)
L
∗
(ℓQ).
For p =0 and ℓ > 0, these equations reduce to Ehrhart reciprocity:
Ehr
0
P
(ℓ) =L(ℓP)
Ehr
0
P
(−ℓ) =(−1)
n
L
∗
(ℓP).
Thus Theorem 9.4.6 simultaneously generalizes the DehnSommerville equations
and Ehrhart reciprocity. The proof of the theorem will use the following lemma.
Lemma 9.4.7. Let P ⊆ M
R
≃ R
n
be a full dimensional lattice polytope with toric
variety X
P
and ample divisor D
P
. Given an integer ℓ > 0 and m ∈ (ℓP) ∩M, let
V
ℓP
(m) be the subspace of M
C
such that V
ℓP
(m) +m is the smallest afﬁne subspace
of M
C
containing the minimal face of ℓP containing m. Then
H
0
(X
P
,
´
Ω
p
X
P
(ℓD
P
)) =
m∈(ℓP)∩M
p
V
ℓP
(m) χ
m
.
Proof. We adapt the strategy used in the proof of Proposition 8.2.18. To simplify
notation, set D = ℓD
P
. Using ρ
⊥
∩M ⊆M and tensoring (8.2.6) with O
X
P
(D), we
obtain the exact sequence
0 −→
´
Ω
p
X
P
(D) −→
_
p
M⊗
Z
O
X
P
(D) −→
ρ
_
p−1
M⊗
Z
O
Dρ
(D).
Take global sections over X
P
and consider the graded piece for m ∈ M. Since
H
0
(X
P
, O
X
P
(D)) =
m∈(ℓP)∩M
C χ
m
, we get the exact sequence
0 −→H
0
(X
P
,
´
Ω
p
X
P
(D))
m
−→
_
p
M
C
−→
ρ
_
p−1
M⊗
Z
H
0
(X
P
, O
Dρ
(D))
m
when m∈(ℓP)∩M. To compute H
0
(X
P
, O
Dρ
(D))
m
, recall from Proposition 4.0.28
that we have an exact sequence
0 −→O
X
P
(−D
ρ
) −→O
X
P
−→O
Dρ
−→0.
Tensor this with O
X
P
(D) and take global sections to obtain
0 −→H
0
(X
P
, O
X
P
(D−D
ρ
)) −→H
0
(X
P
, O
X
P
(D)) −→H
0
(X
P
, O
Dρ
(D)) −→0,
where the exactness on right follows from H
1
(X
P
, O
X
P
(D−D
ρ
)) = 0 courtesy of
Theorem 9.3.7. Comparing the polytopes of D−D
ρ
and D =
ρ
a
ρ
D
ρ
shows that
H
0
(X
P
, O
Dρ
(D))
m
=
_
C ¸m, u
ρ
) =−ℓa
ρ
0 otherwise.
§9.4. Applications to Lattice Polytopes 429
Let F
ρ
be the facet of ℓP corresponding to ρ. For m ∈ ℓP, ¸m, u
ρ
) = −ℓa
ρ
if and
only if m ∈ F
ρ
. Hence we get an exact sequence
0 −→H
0
(X
P
,
´
Ω
p
X
P
(D))
m
−→
_
p
M
C
βp
−→
m∈Fρ
_
p−1
M
C
,
where β
p
is a sum of contraction maps i
uρ
deﬁned in the discussion leading up to
Theorem 8.2.16.
Thus ω ∈
_
p
M
C
is in the kernel of β
p
if and only if i
uρ
(ω) = 0 for all ρ with
m∈ F
ρ
. We know from Exercise 8.2.8 that i
uρ
(ω) =0 if and only if ω ∈
_
p
(ρ
⊥
)
C
.
It follows that the kernel of β
p
in (8.2.8) is the intersection
(9.4.7)
m∈Fρ
_
p
(ρ
⊥
)
C
=
_
p
_
m∈Fρ
(ρ
⊥
)
C
_
.
Since F =
m∈Fρ
F
ρ
is the minimal face of ℓP containing m, one sees easily that
Span
R
(m
0
−m∈ M [ m
0
∈ F) =
m∈Fρ
ρ
⊥
.
Thus V
ℓP
(m) =
m∈Fρ
(ρ
⊥
)
C
. This and (9.4.7) imply ker(β
p
) =
_
p
V
ℓP
(m).
We are now ready to prove our main result.
Proof of Theorem 9.4.6. We begin with part (a). Lemma 9.4.7 implies that
dim H
0
(X
P
,
´
Ω
p
X
P
(ℓD
P
)) =
m∈(ℓP)∩M
_
dimV
ℓP
(m)
p
_
when ℓ > 0. Given a face Q of P and m ∈ (ℓP) ∩M, note that
m ∈ Relint(ℓQ) ∩M ⇐⇒ Q is the minimal face of P containing m.
When this happens, we have dim Q = dimV
ℓP
(m). Since the idimensional faces
of ℓP are ℓQ for Q ∈ P(i), we obtain
dim H
0
(X
P
,
´
Ω
p
X
P
(ℓD
P
)) =
i≥0
_
i
p
_
Q∈P(i)
L
∗
(ℓQ) =
n
i=p
_
i
p
_
Q∈P(i)
L
∗
(ℓQ).
We also have H
q
(X
P
,
´
Ω
p
X
P
(ℓD
P
)) =0 for q > 0 by Theorem 9.3.1. Hence
Ehr
p
P
(ℓ) =χ(
´
Ω
p
X
P
(ℓD
P
)) =dim H
0
(X
P
,
´
Ω
p
X
P
(ℓD
P
)) =
n
i=p
_
i
p
_
Q∈P(i)
L
∗
(ℓQ),
which proves part (a).
Now consider Ehr
p
P
(0). If ℓ > 0 and Q ∈ P(i), then L
∗
(ℓQ) =(−1)
i
Ehr
Q
(−ℓ)
by Ehrhart reciprocity. Since the previous display holds for all ℓ >0, we obtain the
polynomial identity
(9.4.8) Ehr
p
P
(x) =
n
i=p
_
i
p
_
Q∈P(i)
(−1)
i
Ehr
Q
(−x).
430 Chapter 9. Sheaf Cohomology of Toric Varieties
Setting x =0 gives
Ehr
p
P
(0) =
n
i=p
(−1)
i
_
i
p
_
Q∈P(i)
Ehr
Q
(0) =(−1)
p
h
p
by the deﬁnition of h
p
, and part (d) follows.
For part (c), we use Serre duality as given in Theorem 9.2.10, which implies
(9.4.9) H
q
(X
Σ
,
´
Ω
p
X
P
(ℓD
P
))
∨
≃H
n−q
(X
Σ
,
´
Ω
n−p
X
P
(−ℓD
P
))
since ℓD
P
is Cartier for any ℓ ∈ Z. This easily implies
χ(
´
Ω
p
X
P
(ℓD
P
)) =(−1)
n
χ(
´
Ω
n−p
X
P
(−ℓD
P
))
(Exercise 9.4.8). Thus Ehr
p
P
(ℓ) =(−1)
n
Ehr
n−p
P
(−ℓ) for ℓ ∈ Z, proving part (d).
Finally, for part (b), take ℓ ≥0 and consider
Ehr
p
P
(ℓ) =(−1)
n
Ehr
n−p
P
(−ℓ) =(−1)
n
n
i=n−p
_
i
n−p
_
Q∈P(i)
(−1)
i
Ehr
Q
(ℓ),
where the ﬁrst equality uses part (d) and the second uses (9.4.8) with p replaced by
n−p. Since ℓ ≥0 implies Ehr
Q
(ℓQ) =L(ℓQ), we obtain
Ehr
p
P
(ℓ) =
n
i=n−p
(−1)
n−i
_
i
n−p
_
Q∈P(i)
L(ℓQ) =
p
j=0
(−1)
j
_
n− j
n−p
_
Q∈P(i)
L(ℓQ),
where the last equality follows by setting j =n−i. This completes the proof.
Identities equivalent to Theorem 9.4.6 were discovered by McMullen in 1977
in a very different context. See [12, Ch. 5] for details and references, including a
nontoric version of Theorem 9.4.6 in [12, Ex. 5.8–5.10].
Examples. We ﬁrst give a classic example of the DehnSommerville equations.
Example 9.4.8. Let P be a simple 3dimensional lattice polytope in R
3
. The Dehn
Sommerville equations can be written
h
3
=h
0
, so f
3
= f
0
− f
1
+ f
2
− f
3
h
2
=h
1
, so f
2
−3f
3
= f
1
−2f
2
+3f
3
.
Since f
3
=1 (P is the only 3dimensional face of itself), we obtain
f
0
− f
1
+ f
2
=2
f
1
=3f
2
−6.
The ﬁrst equation Euler’s celebrated formula V −E +F = 2, which holds for any
3dimensional polytope. The second seems more mysterious, but when combined
with the ﬁrst reduces to 2f
1
=3f
0
. This holds because is P is simple—every vertex
meets three edges and every edge meets two vertices. ♦
§9.4. Applications to Lattice Polytopes 431
There is a version of the DehnSommerville equations for the dual polytope
P
◦
⊆ N
R
, which is simplicial since P is simple (Exercise 9.4.9). More on the
DehnSommerville equations and toric varieties can be found in [59, Sec. 5.6]. We
also recommend [12, Ch. 5] and [179, Sec. 8.3].
We next give an application of Theorem 9.4.6.
Example 9.4.9. The standard nsimplex ∆
n
has P
n
as its associated toric variety.
Given ℓ > 0, note that
Q∈∆n(i)
L
∗
(ℓ∆
n
) =
_
n+1
n−i
__
ℓ −1
i
_
because
• An nsimplex has
_
n+1
i+1
_
=
_
n+1
n−i
_
faces of dimension i.
• Each Q ∈ ∆
n
(i) is lattice isomorphic to the standard isimplex and hence has
_
ℓ−1
i
_
interior lattice points by (9.4.3).
Then part (a) of Theorem 9.4.6 gives the formula
Ehr
p
∆n
(ℓ) =
n
i=p
_
i
p
__
n+1
n−i
__
ℓ −1
i
_
.
Since
_
i
p
__
ℓ−1
i
_
=
_
ℓ−1
p
__
ℓ−p−1
i−p
_
, this becomes
Ehr
p
∆n
(ℓ) =
_
ℓ −1
p
_
i≥0
_
n+1
n−i
__
ℓ −p−1
i −p
_
=
_
ℓ −1
p
__
n+ℓ −p
ℓ
_
,
where the last equality uses the Vandermonde identity discussed in Exercise 9.4.10.
Hence
dim H
0
(P
n
, Ω
p
P
n
(ℓ)) =
_
ℓ −1
p
__
n+ℓ −p
ℓ
_
.
This formula was ﬁrst proved by Bott [20] in 1957. ♦
Cohomology of pForms. Given a simple polytope P as above, the sheaf
´
Ω
p
X
P
has
Euler characteristic
(9.4.10) χ(
´
Ω
p
X
P
) =Ehr
p
P
(0) =(−1)
p
h
p
by Theorem 9.4.6. The factor of (−1)
p
is explained as follows.
Theorem 9.4.10. Let X
P
be the toric variety of a full dimensional simple lattice
polytope in P ⊆M
R
≃R
n
. Then
dim H
q
(X
P
,
´
Ω
p
X
P
) =
_
h
p
q = p
0 q ,= p.
432 Chapter 9. Sheaf Cohomology of Toric Varieties
Proof. The toric variety X
P
is simplicial, so that Theorem 9.3.2 applies to X
P
. This
shows that H
q
(X
P
,
´
Ω
p
X
P
) =0 for q ,= p. It follows that
χ(
´
Ω
p
X
P
) =(−1)
p
dim H
p
(X
P
,
´
Ω
p
X
P
).
The theorem follows by comparing this with (9.4.10).
A different proof of dim H
p
(X
P
,
´
Ω
p
X
P
) =h
p
will be sketched in Exercise 9.4.11.
In Chapter 12, the sum
h
p
=
n
i=p
(−1)
i−p
_
i
p
_
f
i
will arise naturally when we compute the singular cohomology of a simplicial toric
variety. More precisely, we will show that
h
p
=dim H
2p
(X
P
, C).
The link between this and Theorem 9.4.10 is the Hodge decomposition
H
2p
(X
P
, C) ≃
i+j=2p
H
i
(X
P
,
´
Ω
j
X
P
).
When X
P
is smooth, this is a classical fact—see, for example, [72, p. 116]. In the
simplicial case, this is covered in [168].
Exercises for §9.4.
9.4.1. Prove (9.4.1).
9.4.2. Here are some cases where it is easy to show that Euler characteristics give Hilbert
polynomials.
(a) Let F be a ﬁnitely generated graded module over the polynomial ring S =C[x
0
, . . . , x
n
].
By [37, Ch. 6, Thm. (3.8)], there is an exact sequence
0 −→F
r
−→ −→F
1
−→F
0
−→F −→0,
where each F
i
is a ﬁnite direct sum of modules for the form S(a). Now let F be a
coherent sheaf on P
n
. Prove that there is an exact sequence of sheaves
0 −→F
r
−→ −→F
1
−→F
0
−→F −→0,
where each F
i
is a ﬁnite direct sum of sheaves of the form O
P
n (a). Hint: Apply
Example 6.0.10 with X
Σ
=P
n
.
(b) Use part (a) together with (9.4.1) and Example 9.4.1 to show that χ(F(ℓ)) is a poly
nomial in ℓ ∈ Z, where F(ℓ) =F ⊗
O
P
n
O
P
n (ℓ).
(c) Let F be a coherent sheaf on a complete variety X and let D be a very ample divisor
on X. Use part (b) and Exercise 9.0.6 to show that χ(F(ℓD)) is a polynomial in ℓ ∈Z.
9.4.3. Let F be a coherent sheaf on a projective variety X, and let D be an ample divisor
on X. Thus there is k
0
> 0 such that kD is very ample for all k ≥k
0
.
(a) Let a, k be integers with k ≥k
0
. Use Exercise 9.4.2 to show that there is a polynomial
p
a,k
(x) ∈ Q[x] such that p
a,k
(ℓ) =χ(F((a +ℓk)D)) for all ℓ ∈ Z.
§9.4. Applications to Lattice Polytopes 433
(b) Show that the polynomials p
a,k
(x/k) and p
a,m
(x/m) are equal when k, m ≥ k
0
. Con
clude that there is a polynomial p
a
(x) such that p
a
(ℓ) =χ(F((a+ℓ)D)) for all ℓ ≥k
0
.
(c) Show that the polynomials p
a
(x −a) and p
b
(x −b) are equal for any a, b ∈ Z and
conclude that there is a polynomial p(x) ∈ Q[x] such that p(ℓ) = χ(F(ℓD)) for all
ℓ ∈ Z.
9.4.4. In this exercise ℓ will always denote an integer.
(a) Prove that [(ℓ∆
n
) ∩Z
n
[ =
_
ℓ+n
n
_
for ℓ nonnegative. Hint: Lattice points in ℓ∆
n
give
monomials of degree ℓ in x
0
, . . . , x
n
by Exercise 4.3.6. Here is a combinatorial proof.
Begin with a list of 1’s of length ℓ +n. In n of the positions, convert the 1 to a 0. This
divides the remaining 1’s into n +1 groups. The number of elements in each group
gives the exponents of x
0
, . . . , x
n
. See also [36, Ex. 13 of Ch. 9, §3].
(b) Prove that [Int(ℓ∆
n
) ∩Z
n
[ =
_
ℓ−1
n
_
for ℓ positive. Hint: Show that shifting the interior
lattice points by (1, . . . , 1) gives the lattice points in (ℓ −n −1)∆
n
.
(c) Prove that p(x) = (x +n)(x +n −1) (x +1)/n! satisﬁes p(ℓ) = (−1)
n
_
−ℓ−1
n
_
for ℓ
negative.
9.4.5. A projective variety X ⊆P
n
gives 0 →I
X
→O
P
n →O
X
→0. After tensoring with
O
P
n (ℓ), we have an exact sequence 0 →I
X
(ℓ) →O
P
n (ℓ) →O
X
(ℓ) →0 whose long exact
sequence in cohomology begins
0 →H
0
(P
n
, I
X
(ℓ)) →H
0
(P
n
, O
P
n (ℓ)) →H
0
(P
n
, O
X
(ℓ)) →H
1
(P
n
, I
X
(ℓ)) → .
We know from Example 4.3.6 that H
0
(P
n
, O
P
n (ℓ)) = S
ℓ
, where S = C[x
0
, . . . , x
n
] with the
standard grading. Also recall that as a sheaf on P
n
, O
X
(ℓ) stands for i
∗
O
X
(ℓ), where
i : X ֒→P
n
is the inclusion map. It follows that H
0
(P
n
, O
X
(ℓ)) =H
0
(X, O
X
(ℓ)).
(a) Show that H
0
(P
n
, I
X
(ℓ)) =I(X)
ℓ
, where I(X) ⊆S is the ideal of X.
(b) Show that the Hilbert polynomial of the coordinate ring C[X] = S/I(X) is the Euler
characteristic χ(O
X
(ℓ)).
9.4.6. Let p(x) be a polynomial such that lim
ℓ→∞
p(ℓ)/ℓ
n
exists and is nonzero.
(a) Prove that n =deg(p(x)) and that the above limit is the leading coefﬁcient of p(x).
(b) Use part (a) to prove the assertion made following (9.4.4) concerning the degree and
leading coefﬁcient of the Ehrhart polynomial.
9.4.7. Here we compute the “Ehrhart polynomials” of some simple polytopal complexes.
(a) As in Example 9.4.5, let ∂∆
2
be the boundary of the standard simplex ∆
2
⊆R
2
. Prove
that [(ℓ ∂∆
2
) ∩Z
2
[ =3ℓ, so that Ehr
∂∆2
(x) =3x, with constant term 0.
(b) Consider the “butterﬂy” B =Conv(0, e
1
±e
2
) ∪Conv(0, −e
1
±e
2
) with boundary ∂B.
Prove that [(ℓ ∂B) ∩Z
2
[ =8ℓ −1, so that Ehr
∂B
(x) =8x −1, with constant term −1.
(c) The Euler characteristic of a wellbehaved topological space Z is deﬁned to be e(Z) =
p
(−1)
p
rankH
p
(Z, Z). Show that e(∂∆
2
) =0 and e(∂B) =−1.
The answer to part (c) is not a coincidence—given any polytopal complex P whose ver
tices are lattice points, the constant term of its Ehrhart polynomial Ehr
P
is e(P) (see
[116]). Since a polytope P is contractible, this gives a different way of seeing that the
constant term of Ehr
P
is 1.
9.4.8. Show that (9.4.9) to prove that χ(
´
Ω
p
XP
(ℓD
P
)) =(−1)
n
χ(
´
Ω
n−p
XP
(−ℓD
P
)).
434 Chapter 9. Sheaf Cohomology of Toric Varieties
9.4.9. Let P ⊆M
R
≃R
n
be a full dimensional simplicial lattice polytope and deﬁne
h
simp
p
=
n
i=p
(−1)
i−p
_
i
p
_
f
n−i−1
.
Rescaling and translating P does not affect the face numbers f
i
. So we may that assume the
origin is an interior point of P. Let P
◦
⊆N
R
be the dual polytope of P. Use Exercise 2.3.4
and Proposition 2.3.6 to prove the following:
(a) P
◦
is simple.
(b) The face numbers f
P
i
of P and f
P
◦
i
of P
◦
are related by f
P
n−i−1
= f
P
◦
i
.
(c) h
simp
p
=h
simp
n−p
.
9.4.10. The goal of this exercise is to prove the identity used in Example 9.4.9.
(a) Prove the Vandermonde identity, which states that
i+j=k
_
a
i
__
b
j
_
=
_
a+b
k
_
for a, b ∈N.
Hint: (1 +x)
a
(1 +x)
b
=(1 +x)
a+b
.
(b) Use part (a) to show that
i≥0
_
n+1
n−i
__
ℓ−p−1
i−p
_
=
_
n+ℓ−p
ℓ
_
, as claimed in Example 9.4.9.
9.4.11. For a full dimensional simple polytope P ⊆M
R
≃R
n
, Theorem 8.2.19 gives
0 −→
´
Ω
p
XP
−→K
0
(Σ
P
, p) −→K
1
(Σ
P
, p) −→ −→K
p
(Σ
P
, p) −→0,
where K
i
(Σ
P
, p) =
σ∈ΣP(i)
_
p−i
(σ
⊥
∩M) ⊗
Z
O
V(σ)
and V(σ) ⊆X
P
is the closure of the
orbit corresponding to σ ∈ Σ
P
. You will use this to compute dim H
p
(X
P
,
´
Ω
p
XP
).
(a) Show that χ(
´
Ω
p
XP
) =
p
j=0
(−1)
j
χ(K
j
(Σ
P
, p)).
(b) Use Theorem 9.2.3 to show that χ(K
j
(Σ
P
, p)) =dim H
0
(X
P
, K
j
(Σ
P
, p)).
(c) Use Proposition 2.3.6 to show that dim H
0
(X
P
, K
j
(Σ
P
, p)) =
_
n−j
n−p
_
f
n−j
. Hint: σ
⊥
∩M
has rank n − j for σ ∈ Σ
P
( j).
(d) Use Theorem 9.3.2 to conclude that dim H
p
(X
P
,
´
Ω
p
XP
) =h
n−p
. Hint: Set j =n −i.
9.4.12. When a polytope P has rational but not integral coordinates, the counting function
L(ℓP) is almost a polynomial. Here you will study P =Conv(0, e
1
,
1
2
e
2
) ⊆R
2
.
(a) Given ℓ ∈ N, prove that
L(ℓP) =
_
1
4
ℓ
2
+ℓ +1 ℓ even
1
4
ℓ
2
+ℓ +
3
4
ℓ odd.
This is an example of a quasipolynomial. See [12, Sec. 3.7] for a discussion of the
Ehrhart quasipolynomial of a rational polytope.
(b) The weighted projective plane P(1, 1, 2) is given by the fan in R
2
with minimal gen
erators u
0
= −e
1
−2e
2
, u
1
= e
1
, u
2
= e
2
. Show that X
2P
= P(1, 1, 2) with D
2P
= 2D
0
,
where D
0
is the divisor corresponding to u
0
. Also show that
χ(O
P(1,1,2)
(ℓD
0
)) =L(ℓP).
The Euler characteristic is not a polynomial in ℓ. The reason is that D
0
is not Cartier.
9.4.13. Let P ⊆ M
R
≃ R
n
be a full dimensional lattice poltyope, not necessarily simple.
Let Ehr
p
P
(x) be the unique polynomial satisfying Ehr
p
P
(ℓ) =χ(
´
Ω
p
XP
(ℓD
P
)) for all ℓ ∈ Z.
§9.5. Local Cohomology and the Total Coordinate Ring 435
(a) Prove that if ℓ > 0 in an integer, then
Ehr
p
P
(ℓ) =
n
i=p
_
i
p
_
Q∈P(i)
L
∗
(ℓQ).
Hint: Look at the hypotheses of Theorem 9.3.1 and Lemma 9.4.7.
(b) Prove that Ehr
0
P
(−ℓ) =(−1)
n
Ehr
n
P
(ℓ). Hint: Serre duality.
(c) Prove that Ehr
p
P
(0) =(−1)
p
h
p
. Hint: Follow the proof of Theorem 9.4.6.
This shows that some parts of Theorem 9.4.6 hold for arbitrary lattice polytopes. In the
next exercise you will see that other parts can fail.
9.4.14. This exercise will show how things can go wrong for a nonsimple polytope. Let
P =Conv(±e
1
±e
2
, e
3
) ⊆R
3
. Note that P is not simple, so that X
P
is not simplicial.
(a) Show that h
1
,=h
2
.
(b) Conclude that for this polytope, (9.4.9) cannot hold for all p, q, ℓ. So the version of
Serre duality stated in part (b) of Theorem 9.2.10 can fail for nonsimplicial toric
varieties. Hint: The DehnSommerville equations follow from Theorem 9.4.6.
§9.5. Local Cohomology and the Total Coordinate Ring
In this section, we use local cohomology and Ext to compute the cohomology of a
coherent sheaf on a toric variety X
Σ
. Our treatment is based on [50].
We ﬁrst review the basics of local cohomology.
Local Cohomology. Let R be a ﬁnitely generated Calgebra, I ⊆ R an ideal, and
Man Rmodule. First deﬁne
Γ
I
(M) =¦a ∈ M[ I
k
a =0 for some k ∈ N¦.
This is the Itorsion submodule of M. One checks easily that M→Γ
I
(M) is left
exact. Hence, just as we did for global sections of sheaves in §9.0, we get the
derived functors M→H
p
I
(M) such that
• H
0
I
(M) =Γ
I
(M).
• A short exact sequence 0 →M→N →P →0 gives a long exact sequence
0 →H
0
I
(M) →H
0
I
(N) →H
0
I
(P) →H
1
I
(M) →H
1
I
(N) →H
1
I
(P) → .
The Local
ˇ
Cech Complex. As with sheaf cohomology, there is a
ˇ
Cech complex
for local cohomology. The sheaf case used an afﬁne open cover; here we use
generators of the ideal. More precisely, if I = ¸ f
1
, . . . , f
ℓ
), let [ℓ] = ¦1, . . . , ℓ¦ be
the index set and as in §9.0, let [ℓ]
p
denote the set of all (p+1)tuples (i
0
, . . . , i
p
)
of elements of I satisfying i
0
< < i
p
. Also set f =( f
1
, . . . , f
ℓ
)
Given an Rmodule M, deﬁne
ˇ
C
p
(f, M) =
(i
0
,...,i
p−1
)∈[ℓ]
p−1
M
f
i
0
f
i
p−1
,
436 Chapter 9. Sheaf Cohomology of Toric Varieties
where M
f
i
0
f
i
p−1
is the localization of M at f
i
0
f
i
p−1
. An element of
ˇ
C
p
(f, M) is
a function α that assigns an element of M
f
i
0
f
i
p−1
to each (i
0
, . . . , i
p−1
) ∈ [ℓ]
p−1
.
Then deﬁne a differential
d
p
:
ˇ
C
p
(f, M) −→
ˇ
C
p+1
(f, M)
by
d
p
(α)(i
0
, . . . , i
p
) =
p
k=0
(−1)
k
α(i
0
, . . . ,
´
i
k
, . . . , i
p
),
where we regard α(i
0
, . . . ,
´
i
k
, . . . , i
p
) as an element of M
f
i
0
f
ip
via the map
M
f
i
0
c
f
i
k
f
ip
→M
f
i
0
f
ip
given by localization at f
i
k
. Similar to Exercise 9.0.2, we have d
p
◦d
p−1
=0.
Deﬁnition 9.5.1. Given an Rmodule M and generators f = ( f
1
, . . . , f
ℓ
) of I, the
local
ˇ
Cech complex is
ˇ
C
•
(f, M) : 0 −→
ˇ
C
0
(f, M)
d
0
−→
ˇ
C
1
(f, M)
d
1
−→
ˇ
C
2
(f, M)
d
2
−→ .
Just as the
ˇ
Cech complex computes sheaf cohomology, the local
ˇ
Cech complex
computes local cohomology. See [97, Thm. 7.13] for a proof of the following.
Theorem 9.5.2. Given an Rmodule M and generators f = ( f
1
, . . . , f
ℓ
) of I, there
are natural isomorphisms
H
p
I
(M) ≃H
p
(
ˇ
C
•
(f, M))
for all p ≥0.
Example 9.5.3. For I =¸x, y) ⊆S =C[x, y], the local
ˇ
Cech complex is
0 −→S −→S
x
⊕S
y
−→S
xy
−→0.
Theorem 9.5.2 implies that H
2
I
(S) is the cokernel of S
x
⊕S
y
→ S
xy
. Consider
x
−1
y
−1
C[x
−1
, y
−1
] =Span
C
(x
−a
y
−b
[ a, b > 0) with Smodule structure given by
x
i
y
j
x
−a
y
−b
=
_
x
−(a−i)
y
−(b−j)
i < a, j < b
0 otherwise.
Then the map S
xy
→x
−1
y
−1
C[x
−1
, y
−1
] deﬁned by f (x, y)/(xy)
k
→ f (x, y) x
−k
y
−k
gives an exact sequence
(9.5.1) S
x
⊕S
y
−→S
xy
−→x
−1
y
−1
C[x
−1
, y
−1
] −→0
(Exercise 9.5.1). Thus H
2
I
(S) ≃x
−1
y
−1
C[x
−1
, y
−1
].
It follows that if x and y have degree 1, then H
2
I
(S) is graded with
H
2
I
(S) = ⊕H
2
I
(S)
−6
⊕H
2
I
(S)
−4
⊕H
2
I
(S)
−2
⊕0⊕
and dim H
2
I
(S)
−ℓ
=ℓ −1 for ℓ > 0 (Exercise 9.5.1). ♦
§9.5. Local Cohomology and the Total Coordinate Ring 437
The number ℓ−1 just computed is also the dimension of H
1
(P
1
, O
P
1 (−ℓ)) (see
Example 9.4.1). This is no accident, since we will prove below that
H
2
I
(S)
a
≃H
1
(P
1
, O
P
1 (a))
for all a ∈ Z.
Relation with Ext. For us, an especially useful aspect of local cohomology is its
relation to Ext. In §9.0 we introduced Ext in the context of sheaves. There is
also a module version, where for a ﬁxed Rmodule N, the derived functors of
M →Hom
R
(N, M) are denoted Ext
p
R
(N, M). They have the expected properties,
and more importantly, are easy to compute by computer algebra systems such as
Macaulay 2.
To see the relation between Ext and local cohomology, we begin with the ob
servation that an Rmodule homomorphism φ : R/I
k
→M is determined by φ(1),
and choosing any a ∈ Mgives a homomorphism, provided that I
k
a =0. It follows
that Hom
R
(R/I
k
, M) consists of those elements of Mannihilated by I
k
. Comparing
this to the deﬁnition of Γ
I
(M), we obtain
Γ
I
(M) =lim
−→
k
Hom
R
(R/I
k
, M).
Since direct limit is an exact functor (Exercise 9.5.2), it follows without difﬁculty
that the derived functors are related the same way (see [97, Thm. 7.8] for a proof).
Theorem 9.5.4. H
p
I
(M) =lim
−→
k
Ext
p
R
(R/I
k
, M).
We will need a variant of this result. For an ideal I =¸ f
1
, . . . , f
ℓ
), let
I
[k]
=¸ f
k
1
, . . . , f
k
ℓ
).
By [97, Rem. 7.9], we have the following result.
Theorem 9.5.5. H
p
I
(M) =lim
−→
k
Ext
p
R
(R/I
[k]
, M).
Here is a simple example.
Example 9.5.6. Let I = ¸x, y) ⊆ S = C[x, y]. We know by Example 9.5.3 that
H
2
I
(S)
−5
has dimension 4. To compute this using Ext, let A = S/I
[k]
= S/¸x
k
, y
k
).
Since A and S are graded Smodules, the Ext group Ext
2
S
(A, S) is also graded.
Let us compute the graded piece Ext
2
S
(A, S)
−5
for various values of k. In
Macaulay 2, we set the ring S =Q[x, y] by the command
S = QQ[x,y]
and then A for k =3 is given by
A = coker matrix{{xˆ3,yˆ3}}
The dimension of Ext
2
S
(A, S)
−5
is computed by the command
438 Chapter 9. Sheaf Cohomology of Toric Varieties
hilbertFunction(5,Extˆ2(A,S))
which gives the answer 2. Since dim H
2
I
(S)
−5
= 4, we see that k = 3 is not big
enough. If we switch to k ≥4, then the answer stabilizes at the number 4. ♦
In this example, Ext
2
S
(S/I
[k]
, S)
−5
= H
2
I
(S)
−5
for k sufﬁciently large. We will
prove below that for any degree a ∈ Z, we have
Ext
2
S
(S/I
[k]
, S)
a
=H
2
I
(S)
a
provided that k is sufﬁciently large. The problem is that Ext
2
S
(S/I
[k]
, S) is ﬁnitely
generated over S but H
2
I
(S) is not (Exercise 9.5.4). So we cannot use the same k
for all degrees a. Explicit bounds on k in terms of a are needed in order to turn the
method of Example 9.5.6 into an algorithm.
This concludes our overview of local cohomology; for more, see [48, App. 4]
or [97]. Our next task to is apply local cohomology to toric varieties.
The Toric Case. The total coordinate ring S = C[x
ρ
[ ρ ∈ Σ(1)] of a toric variety
X
Σ
is graded by the class group Cl(X
Σ
), where a monomial
ρ
x
aρ
ρ
has degree
[
ρ
a
ρ
D
ρ
] ∈ Cl(X
Σ
). We also have the irrelevant ideal
B(Σ) =¸x
ˆ σ
[ σ ∈ Σ
max
), x
ˆ σ
=
ρ,∈σ(1)
x
ρ
,
introduced in Chapter 5. We assume that X
Σ
has no torus factors.
We proved in Chapter 5 that a ﬁnitely generated graded Smodule M and a
divisor class α ∈ Cl(X
Σ
) give the following:
• The coherent sheaf F =
¯
M on X
Σ
. Every coherent sheaf on X
Σ
arises in this
way (Proposition 5.3.9).
• The shifted module M(α), where M(α)
β
=M
α+β
for β ∈ Cl(X
Σ
).
The sheaf associated to S(α) is denoted O
X
Σ
(α), and Proposition 5.3.7 tells us that
O
X
Σ
(α) ≃ O
X
Σ
(D) when D is a Weil divisor with divisor class α. More generally,
if F =
¯
M, then F(α) will denote the sheaf associated to M(α). When α is the
class of a Cartier divisor, one can prove that
(9.5.2) F(α) ≃F ⊗
X
Σ
O
X
Σ
(α)
(Exercise 9.5.3). Our ﬁrst main result is that the local cohomology for the irrelevant
ideal B(Σ) computes the sheaf cohomology of all twists a coherent sheaf on X
Σ
.
Theorem 9.5.7. Let M be a ﬁnitely generated graded Smodule with associated
coherent sheaf F =
¯
Mon X
Σ
. If p ≥2, then
H
p
B(Σ)
(M) ≃
α∈Cl(X
Σ
)
H
p−1
(X
Σ
, F(α)).
§9.5. Local Cohomology and the Total Coordinate Ring 439
Furthermore, we have an exact sequence
0 −→H
0
B(Σ)
(M) −→M−→
α∈Cl(X
Σ
)
H
0
(X
Σ
, F(α)) −→H
1
B(Σ)
(M) −→0.
Proof. Let Σ
max
=¦σ
1
, . . . , σ
ℓ
¦, so that B(Σ) is generated by the monomials f
i
=
x
ˆ σ
i
, i ∈ [ℓ] = ¦1, . . . , ℓ¦. Then the terms of the local
ˇ
Cech complex
ˇ
C
•
(f, M) are
direct sums of localizations M at products of various f
i
’s. These localizations are
Cl(X
Σ
)graded since M is graded and the f
i
are monomials. The differentials also
preserve the grading, which by Theorem 9.5.2 implies that H
p
B(Σ)
(M) has a natural
Cl(X
Σ
)grading such that for all α ∈ Cl(X
Σ
), we have
H
p
B(Σ)
(M)
α
=H
p
(
ˇ
C
•
(f, M)
α
).
We will relate
ˇ
C
p
(f, M)
α
to the
ˇ
Cech complex for F(α) given in (9.1.1).
To compute
ˇ
C
p
(f, M)
α
, ﬁrst observe that
ˇ
C
p
(f, M)
α
=
(i
0
,...,i
p−1
)∈[ℓ]
p−1
_
M
f
i
0
f
i
p−1
_
α
=
(i
0
,...,i
p−1
)∈[ℓ]
p−1
_
M(α)
f
i
0
f
i
p−1
_
0
since shifting commutes with localization. For γ =(i
0
, . . . , i
p−1
) ∈[ℓ]
p−1
, note that
(9.5.3) f
i
0
f
i
p−1
=x
ˆ σ
i
0
x
ˆ σ
i
p−1
=
ρ/ ∈σ(1)
x
aρ
ρ
,
where a
ρ
= [¦i
j
[ ρ / ∈ σ
i
j
¦[ > 0. If we set σ
γ
= σ
i
0
∩ ∩σ
i
p−1
, then x
ˆ σγ
gives
the same localization as (9.5.3). Since F(α) is the sheaf associated to M(α),
Proposition 5.3.3 implies that
_
M(α)
f
i
0
f
i
p−1
_
0
=
_
M(α)
x
ˆ σγ
_
0
≃Γ(U
σγ
, F(α)).
Comparing this with the
ˇ
Cech complex (9.1.1) for the open cover U = ¦U
σ
i
¦
i∈[ℓ]
,
we obtain
ˇ
C
p
(f, M)
α
=
γ∈[ℓ]
p−1
Γ(U
σγ
, F(α)) ≃
ˇ
C
p−1
(U , F(α)).
This isomorphism is compatible with the differentials. It follows that the local
ˇ
Cech
complex
ˇ
C
•
(f, M)
α
is obtained from the
ˇ
Cech complex
ˇ
C
•
(U , F(α)) by deleting
the ﬁrst term and shifting the remaining terms. When p ≥2, this implies
H
p
B(Σ)
(M)
α
≃H
p−1
(X
Σ
, F(α)),
and with a little work (Exercise 9.5.5) we also get an exact sequence
0 −→H
0
B(Σ)
(M)
α
−→M
α
−→H
0
(X
Σ
, F(α)) −→H
1
B(Σ)
(M)
α
−→0.
440 Chapter 9. Sheaf Cohomology of Toric Varieties
Theorems 9.5.5 and 9.5.7 imply that when p ≥1,
H
p
(X
Σ
, F(α)) ≃lim
−→
k
Ext
p+1
S
(S/B(Σ)
[k]
, M)
α
when F =
¯
M and α ∈ Cl(X
Σ
). We can compute Ext
p+1
S
(S/B(Σ)
[k]
, M)
α
by the
methods of Example 9.5.6. The problem is the direct limit. We tackle this next.
Stabilization of Ext. In the toric case, Ext has a nice relation to local cohomology.
Lemma 9.5.8. Let M be a ﬁnitely generated Cl(X
Σ
)graded Smodule and ﬁx α ∈
Cl(X
Σ
). If Σ is a complete fan, then there exists k
0
∈ N such that for all k ≥k
0
, the
natural map S/B(Σ)
[k+1]
→S/B(Σ)
[k]
induces an isomorphism
Ext
p
S
(S/B(Σ)
[k]
, M)
α
≃Ext
p
S
(S/B(Σ)
[k+1]
, M)
α
.
In particular, k ≥k
0
implies
Ext
p
S
(S/B(Σ)
[k]
, M)
α
≃H
p
B(Σ)
(M)
α
.
Proof. We give a proof only for M=S following Mustat ¸ˇ a [132, Thm. 1.1]. For the
general case, one replaces M with a free resolution and uses a spectral sequence
argument. See [50, Prop. 4.1] for the details.
Earlier we described Ext using the derived functors of M→Hom
R
(N, M) for
ﬁxed N. Ext also comes from the derived functors of N→Hom
R
(N, M) for ﬁxed M,
where one uses a projective resolution of N instead of an injective resolution of M
(see, for example, [48, A3.11]). In particular, we can compute Ext
p
S
(S/B(Σ)
[k]
, S)
using any free resolution of S/B(Σ)
[k]
. An especially nice resolution is given by
the Taylor resolution, which is described as follows.
We begin with S/B(Σ). The minimal generators of B(Σ) are f
i
=x
ˆ σ
i
for i ∈[ℓ].
Let F
s
be a free S module with basis e
I
for all I ⊆ [ℓ] with [I[ = s. To deﬁne the
differential d
s
: F
s
→F
s−1
, take I, J ⊆[ℓ] with [I[ =[J[ +1 =s and list the elements
of I as i
1
< < i
s
. Then deﬁne
c
IJ
=
_
0 J ,⊆I
(−1)
r
f
I
/f
J
I =J ∪¦i
r
¦,
where f
I
=lcm( f
i
[ i ∈ I) and similarly for f
J
. Then
d
s
(e
I
) =
J
c
IJ
e
J
.
Since F
0
=S, we have an obvious map F
0
→S/B(Σ). One can prove that
0 −→F
ℓ
d
ℓ
−→
d
2
−→F
1
d
1
−→F
0
−→S/B(Σ) −→0
is exact (see [48, Ex. 17.11]). This is the Taylor resolution (F
•
, d
•
) of S/B(Σ).
This construction applies to any monomial ideal. In particular, it works for
B(Σ)
[k]
= ¸ f
k
1
, . . . , f
k
ℓ
). Let (F
[k]
•
, d
[k]
•
) denote the Taylor resolution of S/B(Σ)
[k]
.
It has the same modules F
[k]
s
= F
s
, and since the f
i
are squarefree, we have f
k
I
=
§9.5. Local Cohomology and the Total Coordinate Ring 441
lcm( f
k
i
[ i ∈ I). Thus the differentials d
[k]
s
in the Taylor resolution of S/B(Σ)
[k]
are
given by
d
[k]
s
(e
I
) =
J
c
k
IJ
e
J
, c
IJ
as above.
We now compare the Taylor resolutions of S/B(Σ)
[k]
and S/B(Σ)
[k+1]
. The
maps φ
s
: F
[k+1]
s
→F
[k]
s
deﬁned by φ
s
(e
I
) = f
I
e
I
induces a commutative diagram
F
[k+1]
s
φs
//
d
[k+1]
s
F
[k]
s
d
[k]
s
F
[k+1]
s−1
φ
s−1
//
F
[k]
s−1
.
These maps are compatible with the surjection S/B(Σ)
[k+1]
→S/B(Σ)
[k]
. Hence
(9.5.4) H
p
(Hom
S
(F
[k+1]
•
, S)) →H
p
(Hom
S
(F
[k]
•
, S))
induced by the φ
s
can be identiﬁed with the canonical map
Ext
p
S
(S/B(Σ)
[k]
, S) →Ext
p
S
(S/B(Σ)
[k+1]
, S).
The next key idea involves the use of a ﬁner grading than the Cl(X
Σ
)grading
used so far. Recall that this grading is induced by the map
(9.5.5) Z
Σ(1)
−→Cl(X
Σ
)
where x
a
=
ρ
x
aρ
ρ
has degree [
ρ
a
ρ
D
ρ
]. The ring S also has a Z
Σ(1)
grading
where deg(x
a
) =a. Since B(Σ)
[k]
is a monomial ideal, the quotient ring S/B(Σ)
[k]
is Z
Σ(1)
graded. Then Ext
i
S
(S/B(Σ)
[k]
, S) inherits a natural Z
Σ(1)
grading.
Nowgrade the Taylor resolution (F
[k]
•
, d
[k]
•
) by setting deg(e
I
) =kdeg( f
I
). This
guarantees two things:
• The differential d
[k]
s
has degree 0, so the isomorphism Ext
p
S
(S/B(Σ)
[k]
, S) ≃
H
p
(Hom
S
(F
[k]
•
, S) is Z
Σ(1)
graded.
• The map φ
s
has degree 0, so (9.5.4) is Z
Σ(1)
graded.
It follows that (F
[k]
s
)
∨
=Hom
S
(F
[k]
s
, S) has dual basis e
∗
I
with deg(e
∗
I
) =−kdeg( f
I
).
Given a, b ∈ Z
Σ(1)
, deﬁne
a ≥b if and only if a
ρ
≥b
ρ
for all ρ ∈ Σ(1).
We claim that if a ≥(−k, . . . , −k), then
(9.5.6) (φ
∨
s
)
a
: (F
[k]
s
)
∨
a
−→(F
[k+1]
s
)
∨
a
is an isomorphism for all s.
To prove this, ﬁrst observe that φ
∨
s
(e
∗
I
) = f
I
e
∗
I
. It follows that (φ
∨
s
)
a
is injective.
Now take x
b
e
∗
I
∈ (F
[k+1]
s
)
∨
a
. Then b−(k +1)deg( f
I
) =a, so
b =(k +1)deg( f
I
) +a ≥(k +1)deg( f
I
) +(−k, . . . , −k).
442 Chapter 9. Sheaf Cohomology of Toric Varieties
Since f
I
is squarefree, (k +1)deg( f
I
) is a vector with whose ρth entry is (k +1) if
x
ρ
divides f
I
and 0 otherwise. Then the above inequality implies that f
I
divides x
b
,
so that x
b
e
∗
I
is in the image of (φ
∨
s
)
a
. It follows that (9.5.6) is an isomorphism.
The local cohomology H
p
B(Σ)
(S) has a Z
Σ(1)
grading which is compatible its
Cl(X
Σ
)grading via (9.5.5). This follows easily from Theorem 9.5.5 and the Taylor
resolution. If α ∈ Cl(X
Σ
) and p ≥2, then
H
p
B(Σ)
(S)
α
≃H
p−1
(X
Σ
, O
X
Σ
(α)).
by Theorem 9.5.7. The righthand side is ﬁnitedimensional since Σ is complete.
This implies that when we decompose H
p
B(Σ)
(S) into its nonzero Z
Σ(1)
graded
pieces H
p
B(Σ)
(S)
a
, only ﬁnitely many can appear in H
p
B(Σ)
(S)
α
. For these ﬁnitely
many a’s, pick k
0
such that they all satisfy a ≥ (−k
0
, . . . , −k
0
). This k
0
has the
required properties. The argument for p =0, 1 is covered in Exercise 9.5.6.
Theorem 9.5.7 and Lemma 9.5.8 imply that for p ≥1 and k ≫0,
H
p
(X
Σ
, O
X
Σ
(α)) ≃H
p+1
B(Σ)
(S)
α
≃Ext
p+1
S
(S/B(Σ)
[k]
, S)
Notice how these isomorphisms generalize Examples 9.5.3 and 9.5.6. Our ﬁnal
task is give an explicit method for deciding when k is big enough.
Bounds. For a complete fan Σ, graded Smodule M, and divisor class α ∈ Cl(X
Σ
),
the paper [50] gives an explict value for the number k
0
appearing in Lemma 9.5.8.
We will state the results of [50] without proof, though the following example sug
gests some of the ideas involved.
Example 9.5.9. Let Σ be a complete fan in M
R
≃ R
2
with minimal generators
u
1
, . . . , u
r
arranged counterclockwise around the origin. Suppose we have a divisor
D =
i
a
i
D
i
on X
Σ
and m ∈ M such that H
1
(X
Σ
, O
X
Σ
(D))
m
,= 0. It follows from
Proposition 9.1.5 that the sign pattern of m (determined by ¸m, u
i
) +a
i
≥0 or < 0)
has at least two strings of consecutive −’s. The set I = ¦i ∈ [r] [ ¸m, u
i
) +a
i
< 0¦
records the locations of the −’s in the sign pattern of m.
Since H
1
(X
Σ
, O
X
Σ
(D))
m
,=0 is ﬁnitedimensional, there are only ﬁnitely many
m’s with the same sign pattern. In other words, the inequalities
¸m, u
i
) +a
i
< 0, i ∈ I
¸m, u
i
) +a
i
≥0, i ∈ [r] ¸I
have only ﬁnitely many integer solutions, which means that the region of the plane
deﬁned by these inequalities is bounded. You can see and example of this in Fig
ure 4 from Example 9.1.7. Since this region determined by the u
i
and a
i
, it should
be possible to bound the size of the m’s that appear in terms of the u
i
and a
i
.
To relate this to Lemma 9.5.8, let b = (¸m, u
1
) +a
1
, . . . , ¸m, u
r
) +a
r
). Then it
is easy to see that
H
1
(X
Σ
, O
X
Σ
(D))
m
≃H
2
B(Σ)
(S)
b
.
§9.5. Local Cohomology and the Total Coordinate Ring 443
Thus bounding the m’s with H
1
(X
Σ
, O
X
Σ
(D))
m
,= 0 is equivalent to bounding the
b’s with H
2
B(Σ)
(S)
b
,= 0. And once we ﬁnd this bound, we know how to pick k
0
so
that b ≥(−k
0
, . . . , −k
0
) for all such b’s. The proof of Lemma 9.5.8 shows that this
k
0
works. ♦
In the general case, we proceed as follows. Pick a basis e
1
, . . . , e
n
of M and
let A be the matrix whose rows give the coefﬁcients of the u
ρ
’s with respect to the
chosen basis. Then A is an r n matrix, where r =[Σ(1)[. For this matrix, deﬁne
(9.5.7)
q
n
=min([nonzero nn minors of A[)
Q
1
=max([entries of A[)
Q
n−1
=max([(n−1) (n−1) minors of A[)
The following bound is proved in [50, Cor. 3.3].
Theorem 9.5.10. Given a complete fan Σ in N
R
≃ R
n
and a divisor class α =
[
ρ
a
ρ
D
ρ
] ∈ Cl(X
Σ
), we have
Ext
p
S
(S/B(Σ)
[k]
, S)
α
≃H
p
B(Σ)
(S)
α
for all k ≥k
0
, where
k
0
=n
2
max
ρ
([a
ρ
[)
Q
1
Q
n−1
q
n
.
For a ﬁnitely generated graded Smodule M, the formula for k
0
involves the
minimal free resolution of M. Write the resolution as
−→F
2
−→F
1
−→F
0
−→M−→0
and write
F
j
=
β∈Cl(X
Σ
)
S(−β)
r
j,β
.
Finally, for each β with r
j,β
,=0, write β =[
ρ
a
β,ρ
D
ρ
] ∈ Cl(X
Σ
). Then [50, Prop.
4.1] gives the following bound.
Theorem 9.5.11. Let Σ and α be as in Theorem 9.5.10, and let M be a ﬁnitely
generated graded Smodule. Then
Ext
p
S
(S/B(Σ)
[k]
, M)
α
≃H
p
B(Σ)
(M)
α
for all k ≥k
0
, where
k
0
=n
2
max
ρ, j,r
j,β
,=0
([a
ρ
−a
β,ρ
[)
Q
1
Q
n−1
q
n
.
444 Chapter 9. Sheaf Cohomology of Toric Varieties
The Cotangent Bundle. We end this section by using our methods to calculate the
cohomology of the cotangent bundle of a smooth complete toric variety X
Σ
. The
ﬁrst step is to ﬁnd a graded Smodule Mwith Ω
1
X
Σ
≃
¯
M.
In (8.1.7) we constructed an exact sequence
0 −→Ω
1
S
−→M⊗
Z
S →
ρ
S/¸x
ρ
)
where M⊗
Z
S →
ρ
S/¸x
ρ
) is deﬁned by m⊗ f →
ρ
¸m, u
ρ
)[ f ], and in Corol
lary 8.1.5 we showed that Ω
1
X
Σ
is the sheaf of Ω
1
S
. However, we need a description
of Ω
1
S
that is easier to implement on a computer. We do this as follows.
Pick bases of M and Pic(X
Σ
) and order the elements of Σ(1) as ρ
1
, . . . , ρ
r
.
Then the basic exact sequence
0 −→M −→Z
Σ(1)
−→Pic(X
Σ
) −→0
can be written
(9.5.8) 0 −→Z
n
A
−→Z
r
B
−→Z
r−n
−→0,
where A is an r n matrix whose ith row consists of the coefﬁcients of u
ρ
i
in
the basis of M. This is the same matrix A that appears in Theorem 9.5.10. The
(r −n) r matrix B is called the Gale dual of A.
Lemma 9.5.12. The (r −n) r matrix
θ =B
_
_
_
_
_
x
ρ
1
0 0
0 x
ρ
2
0
.
.
.
.
.
.
.
.
.
.
.
.
0 0 x
ρr
_
_
_
_
_
induces a graded homomorphism θ :
r
i=1
S(−deg(x
ρ
i
)) →S
r−n
of degree 0 such
that Ω
1
X
Σ
is the sheaf associated to ker(θ).
Proof. Let x be the r r diagonal matrix of variables appearing in the statement of
the lemma. Similar to the proof of Theorem 8.1.6, we have a commutative diagram
0
0
0
0
//
Ω
1
S
//
Z
n
⊗
Z
S
//
A
r
i=1
S/¸x
ρ
i
)
0
//
r
i=1
S(−deg(x
ρ
i
))
x
//
Z
r
⊗
Z
S
//
B
r
i=1
S/¸x
ρ
i
)
//
0
0
//
Z
r−n
⊗
Z
S
//
Z
r−n
⊗
Z
S
//
0
//
0
0 0
§9.5. Local Cohomology and the Total Coordinate Ring 445
The rows are exact, as are the center and right columns. By the diagram chase from
the proof of Theorem 8.1.6, the dotted arrows are exact, and by commutivity, the
dotted arrow from
r
i=1
S(−deg(x
ρ
i
)) to Z
r−n
⊗
Z
S is given by θ =B x.
Example 9.5.13. For P
2
, the matrices A and B of (9.5.8) are given by
A =
_
_
1 0
0 1
−1 −1
_
_
, B =
_
1 1 1
_
.
To create the module Ω
1
S
in Macaulay 2, we use the commands
S = QQ[x,y,z]
OM = ker matrix{{x,y,z}}
We also need a free resolution of Ω
1
S
, which can be computed by Macaulay 2:
F = res OM
F.dd
The ﬁrst command computes the resolution; the second prints out the differentials.
The result is
0 −→S(−3)
0
@
x
y
z
1
A
−−−→S(−2)
3
−→Ω
1
S
−→0.
The numbers from (9.5.7) are q
2
= Q
1
= 1, so that for a ∈ Z, the formula for k
0
from Theorem 9.5.11 is
k
0
=4max([a−2[, [a−3[).
For a in the range −4 ≤a ≤4, we can use k
0
=28. This implies that for p ≥1,
H
p
(P
2
, Ω
1
P
2
(a)) ≃Ext
p
S
(S/¸x
28
, y
28
, z
28
), Ω
1
S
)
a
when −4 ≤ a ≤ 4. This can be computed by the methods of Example 9.5.6. We
can also compute H
0
(P
2
, Ω
1
P
2
(a)) directly fromΩ
1
S
(Exercise 9.5.7). The results are
shown in Table 3 on the next page. In Exercise 9.5.8 you will calculate this table
by hand. Notice also that the symmetry of the table comes from Serre duality. ♦
Here is a slightly more complicated example.
Example 9.5.14. Let us compute the ﬁrst cohomology of various twists of the
cotangent sheaf of the Hirzebruch surface H
2
. We will use the notation of Exam
ple 9.3.6. Using the basis of Pic(H
2
) given by the classes of D
3
, D
4
, the matrices
of (9.5.8) are
A =
_
_
_
_
−1 2
0 1
1 0
0 −1
_
_
_
_
, B =
_
1 −2 1 0
0 1 0 1
_
.
446 Chapter 9. Sheaf Cohomology of Toric Varieties
Table 3. dim H
p
(P
2
, Ω
1
P
2
(a)) for −4 ≤a ≤4
a¸p 0 1 2
−4 0 0 15
−3 0 0 8
−2 0 0 3
−1 0 0 0
0 0 1 0
1 0 0 0
2 3 0 0
3 8 0 0
4 15 0 0
The total coordinate ring is S = C[x
1
, x
2
, x
3
, x
4
], where the degrees of the variables
are given by the columns of B, i.e.,
deg(x
1
) =(1, 0), deg(x
2
) =(−2, 1), deg(x
3
) =(1, 0), deg(x
4
) =(0, 1).
Then we create Ω
1
S
within Macaulay 2 by the commands
S = QQ[x1,x2,x3,x4,Degrees=>{{1,0},{2,1},{1,0},{0,1}}]
OM = ker matrix{{x1,2
*
x2,x3,0},{0,x2,0,x4}}
Be sure you understand how this uses the matrix θ from Lemma 9.5.12.
Using Macaulay 2, we also compute the free resolution
0 −→S(0, −2) −→S(1, −2)
2
⊕S(−2, 0) −→Ω
1
S
−→0.
The numbers from (9.5.7) are q
2
= 1 and Q
1
= 2, so that for α = [aD
3
+bD
4
] in
Pic(H
2
), the formula for k
0
from Theorem 9.5.11 is
k
0
=16max([a−2[, [b[, [a+1[, [b−2[, [a[, [b−2[).
We can use our methods to calculate the cohomology groups of Ω
1
H
2
(a, b) =
Ω
1
H
2
(aD
3
+bD
4
) for all a, b. Table 4 records some computations of H
1
. There are
Table 4. dimH
1
(H2, Ω
1
H
2
(a, b)) for −2 ≤a, b ≤2
a¸b –2 –1 0 1 2
–2 0 0 3 4 3
–1 0 0 2 2 2
0 1 1 2 1 1
1 2 2 2 0 0
2 3 4 3 0 0
many things we can see from this table, including:
§9.5. Local Cohomology and the Total Coordinate Ring 447
• Theorem 9.4.10 implies that dim H
1
(H
2
, Ω
1
H
2
) =h
1
= f
1
−2f
2
=4−2 1 =2
since H
2
comes from a quadrilateral (see Examples 2.3.15 and 3.1.16). This
explains the 2 when (a, b) =(0, 0).
• Serre duality implies that H
1
(H
2
, Ω
1
H
2
(a, b))
∨
≃H
1
(H
2
, Ω
1
H
2
(−a, −b)). This
explains the symmetry in the table.
• BottDanilovSteenbrink vanishing implies that H
1
(H
2
, Ω
1
H
2
(a, b)) = 0 for
a, b > 0. This explains the 0s in the lower right corner of the table.
You will verify the details of this example in Exercise 9.5.9. One suggestion is
that after computing an Ext group E =Ext
p
S
(A, Ω
1
S
), use the command
rank source basis({a,b},E)
to compute dim E
(a,b)
rather than
hilbertFunction({a,b},E)
since the former is faster than the latter and gives the same answer. ♦
An algorithmic approach to these calculations is described in [50, Sec. 5]
Exercises for §9.5.
9.5.1. Prove the exactness of (9.5.1).
9.5.2. Suppose that (A
i
, φ
i
), (A
′
i
, φ
′
i
), (A
′′
i
, φ
′′
i
) are directed systems, and that for each i,
there exist d
i
and δ
i
commuting with the φ’s, such that we have an exact sequence
0 −→A
i
di
−→A
′
i
δi
−→A
′′
i
−→0.
Prove the exactness of the sequence
0 −→lim
−→
i
A
i
−→lim
−→
i
A
′
i
−→lim
−→
i
A
′′
i
−→0.
9.5.3. Prove (9.5.2) when α ∈Cl(X
Σ
) is the class of a Cartier divisor. Hint: Look carefully
at Proposition 5.3.3 and remember that the restriction of O
XΣ
(α) to U
σ
is trivial.
9.5.4. For I =¸x, y) ⊆S =C[x, y], prove that H
2
I
(S) is not ﬁnitely generated as an Smodule.
Hint: Use Example 9.5.3 to show that H
2
I
(S)
a
,=0 for all a ≤−2.
9.5.5. Complete the proof of Theorem 9.5.7 by proving that there is an exact sequence
0 −→H
0
B(Σ)
(M)
α
−→M
α
−→H
0
(X
Σ
, F(α)) −→H
1
B(Σ)
(M)
α
−→0
for all α ∈ Cl(X
Σ
).
9.5.6. Prove that H
0
B(Σ)
(S) =H
1
B(Σ)
(S) =0. Hint: Proposition 5.3.7.
9.5.7. Here are some details to check from Example 9.5.13
(a) Use the exact sequence from Lemma 9.5.12 to show that (Ω
1
S
)
α
≃ H
0
(X
Σ
, Ω
1
XΣ
(α))
for all α ∈ Pic(X
Σ
) when X
Σ
is smooth and complete.
(b) Verify the ﬁrst column of Table 3.
(c) Show that the ﬁrst column agrees with the Bott formula from Example 9.4.9.
448 Chapter 9. Sheaf Cohomology of Toric Varieties
9.5.8. We can check Table 3 of Example 9.5.13 with a “barehanded” approach. Observe
that we have exact sequences involving Ω
1
P
2
:
0 −→Ω
1
P
2 −→O
3
P
2 (−1) −→O
P
2 −→0
0 −→O
P
2 (−3) −→O
3
P
2 (−2) −→Ω
1
P
2 −→0.
The ﬁrst comes from Lemma 9.5.12, and the second comes by sheaﬁfying the free resolu
tion of Ω
1
S
computed in Example 9.5.13. Twist these sequences by a ∈ Z and consider the
resulting long exact sequences in cohomology. Using the vanishing theorems we know,
conclude that the nonzero values for H
p
(P
2
, Ω
1
P
2
(a)) are:
dimH
0
(P
2
, Ω
1
P
2 (a)) =a
2
−1 if a ≥2
dimH
1
(P
2
, Ω
1
P
2 (a)) =1 if a =1
dimH
2
(P
2
, Ω
1
P
2 (a)) =a
2
−1 if a ≤−2.
These formulas give the numbers in Table 3.
9.5.9. This exercise concerns Example 9.5.14.
(a) Use Macaulay 2 to compute the free resolution of Ω
1
S
.
(b) Show how the formula given for k
0
follows from Theorem 9.5.11.
(c) Compute the entries in Table 4.
Appendix: Introduction to Spectral Sequences
Spectral sequences are used several times in this chapter. Here we give the necessary
background, though our discussion is far from complete. We refer the reader to [72] or
[177] for a more complete treatment of spectral sequences.
Deﬁnition and Intuition. At ﬁrst glance, spectral sequences may seemrather complicated.
Fortunately, most of the ones encountered in this book will be fairly simple to understand.
Deﬁnition 9.A.1. A spectral sequence over C consists of vector spaces E
p,q
r
and linear
maps
d
p,q
r
: E
p,q
r
−→E
p+r,q−r+1
r
such that
(a) d
p+r,q−r+1
r
◦ d
p,q
r
=0 for all p, q, r.
(b) E
p,q
r+1
is the cohomology of (E
p,q
r
, d
p,q
r
), i.e.,
E
p,q
r+1
=ker(d
p,q
r
: E
p,q
r
→E
p+r,q−r+1
r
)
_
im(d
p−r,q+r−1
r
: E
p−r,q+r−1
r
→E
p,q
r
)
We will always work with ﬁrst quadrant spectral sequences, which means that E
p,q
r
=0
when p < 0 or q < 0. Thus the nonvanishing terms lie in the quadrant where p, q ≥0. The
minimum value of r is usually 1 or 2, in which case we say that (E
p,q
r
, d
p,q
r
) is an E
1
or E
2
spectral sequence respectively.
When working with an E
1
spectral sequence, we often know the differentials d
p,q
1
explicitly. In general, however, the differentials d
p,q
r
for r ≥ 2 are much more difﬁcult to
describe.
Appendix: Introduction to Spectral Sequences 449
For ﬁxed p, q, the differentials mapping to and from E
p,q
r
vanish when r is large by our
ﬁrst quadrant assumption. It follows that
E
p,q
r
=E
p,q
r+1
=E
p,q
r+2
=
for r ≫0. This common value is deﬁned to be E
p,q
∞
.
Deﬁnition 9.A.2. A spectral sequence (E
p,q
r
, d
p,q
r
) converges to vector spaces H
k
, k ≥ 0,
if there are subspaces
0 =F
k+1
H
k
⊆F
k
H
k
⊆F
k−1
H
k
⊆ ⊆F
1
H
k
⊆ F
0
H
k
=H
k
such that
E
p,q
∞
≃F
p
H
p+q
/F
p+1
H
p+q
For an E
1
or E
2
spectral sequence, we write this as
E
p,q
1
⇒H
p+q
or E
p,q
2
⇒H
p+q
respectively.
The intuition behind a convergent E
2
spectral sequence is that we often know some
thing about the E
2
terms and are interested in what the spectral sequence converges to.
One can think of E
p,q
2
as a ﬁrst approximation to the convergent, and then E
p,q
3
, E
p,q
4
, . . . as
better and better approximations. A nice introduction to this way of thinking can be found
in the ﬁrst chapter of [123].
Here is an example of how to work with a spectral sequence.
Proposition 9.A.3. Suppose E
p,q
2
⇒H
p+q
is an E
2
spectral sequence with the property
that E
p,q
2
=0 for all q > 0. Then
E
p,0
2
≃H
p
for all p ≥0.
Proof. First observe that all differentials d
p,q
r
vanish for r ≥ 2 since E
p,q
2
= 0 for q > 0. It
follows that E
p,q
2
=E
p,q
∞
for all p, q. Then we have
E
p,0
2
≃F
p
H
p
/F
p+1
H
p
=F
p
H
p
and 0 =E
p−1,1
2
=E
p−2,2
2
= =E
0,p
2
implies
0 =F
p−1
H
p
/F
p
H
p
=F
p−2
H
p
/F
p−1
H
p
= =F
0
H
p
/F
1
H
p
,
so that F
p
H
p
= F
p−1
H
p
= =F
1
H
p
=F
0
H
p
=H
p
. Thus E
p,0
2
≃H
p
, as claimed.
The hypothesis of Proposition 9.A.3 implies that the differentials d
p,q
2
vanish for all
p, q. In general, we say that a spectral sequence degenerates at E
r
if the differentials d
p,q
r
vanish for all p, q. Note that degeneration at E
r
implies E
p,q
r
≃E
p,q
∞
.
Edge Homomorphisms. Another feature of a convergent E
2
spectral sequence is that all
differentials starting from E
p,0
r
vanish, so that the spectral sequence gives maps
E
p,0
2
−→E
p,0
3
−→ −→E
p,0
∞
≃F
p
H
p
/F
p+1
H
p
=F
p
H
p
⊆H
p
.
Deﬁnition 9.A.4. The map E
p,0
2
→H
p
is called an edge homomorphism.
This leads to a more precise version of Proposition 9.A.3.
450 Chapter 9. Sheaf Cohomology of Toric Varieties
Proposition 9.A.5. Suppose E
p,q
2
⇒H
p+q
is an E
2
spectral sequence with the property
that E
p,q
2
=0 for all q > 0. Then the edge homomorphism
E
p,0
2
−→H
p
is an isomorphism for all p ≥ 0.
For many E
2
spectral sequences, we often know the edge homomorphisms E
p,0
2
→H
p
explicitly. This is true in the two applications of Proposition 9.A.5 given in §9.0.
Chapter 10
Toric Surfaces
§10.1. Singularities of Toric Surfaces and Their Resolutions
In this chapter, we will apply the theory developed so far to give a detailed treat
ment of the structure of 2dimensional normal toric varieties (toric surfaces).
Singular Points of Toric Surfaces. If X
Σ
is the toric surface of a fan Σin N
R
≃ R
2
,
then minimal generators of the rays ρ ∈ Σ(1) are primitive and hence extend to a
basis of N. Then Theorem 3.1.19 implies that the toric surface obtained by re
moving the ﬁxed points of the torus action (i.e., the points corresponding to the
2dimensional cones under the OrbitCone Correspondence) is smooth. There are
only ﬁnitely many such points, so X
Σ
has at most ﬁnitely many singular points.
Moreover, 2dimensional cones are always simplicial, so from Example 1.3.20,
each of these singular points is a ﬁnite abelian quotient singularity (isomorphic to
the image of the origin in the quotient C
2
/G where G is a ﬁnite abelian group).
All cones are assumed to be rational and polyhedral. A 2dimensional strongly
convex cone in N
R
≃R
2
has the following normal form that will facilitate our study
of the singularities of toric surfaces.
Proposition 10.1.1. Let σ ⊆ N
R
≃ R
2
be a 2dimensional strongly convex cone.
Then there exists a basis e
1
, e
2
for N such that
σ =Cone(e
2
, de
1
−ke
2
),
where d > 0, 0 ≤k < d, and gcd(d, k) =1.
Proof. We will need the following modiﬁed division algorithm here and at several
other points in this chapter (Exercise 10.1.1):
(10.1.1)
Given integers l and d > 0, there are unique integers
s and k such that l =sd −k and 0 ≤k < d.
453
454 Chapter 10. Toric Surfaces
Say σ =Cone(u
1
, u
2
), where u
i
are primitive vectors. Since u
1
is primitive, we
can take it as part of a basis of N, and we let e
2
= u
1
. Since σ is strongly convex,
for any basis e
′
1
, e
2
for N, it will be true that
u
2
=de
′
1
+le
2
for some d ,= 0. By replacing e
′
1
by −e
′
1
if necessary, we can assume d > 0. By
(10.1.1), there are integers s, k such that l = sd −k, where 0 ≤ k < d. Using this
integer s, let e
1
=e
′
1
+se
2
. Then e
1
, e
2
is also a basis for N and
u
2
=de
1
+(l −sd)e
2
=de
1
−ke
2
.
Hence σ = Cone(e
2
, de
1
−ke
2
) as claimed, and gcd(d, k) = 1 follows since u
2
is
primitive.
We will call the integers d, k in this statement the parameters of the cone σ,
and ¦e
1
, e
2
¦ is called a normalized basis for N relative to σ. The uniqueness of d, k
will be studied in Proposition 10.1.3 below.
Using the normal form, we next describe the local structure of the point p
σ
in the afﬁne toric variety U
σ
. Recall from Example 1.3.20 that if N
′
⊆ N is the
sublattice generated by the ray generators of σ, then U
σ
≃C
2
/G, where G=N/N
′
.
In our situation, N =Ze
1
⊕Ze
2
, and
N
′
=Ze
2
⊕Z(de
1
−ke
2
) =dZe
1
⊕Ze
2
,
so it follows easily that
(10.1.2) G=N/N
′
≃Z/dZ.
In particular, for singularities of toric surfaces, the ﬁnite group G is always cyclic.
The action of G on C
2
is determined by the integers d, k as follows. We write
µ
d
=¦ζ ∈ C [ ζ
d
=1¦
for the group of dth roots of unity in C. Then a choice of a primitive dth root of
unity deﬁnes an isomorphism of groups µ
d
≃Z/dZ.
Proposition 10.1.2. Let M
′
be the dual lattice of N
′
and let m
1
, m
2
∈ M
′
be dual
to u
1
, u
2
in N
′
. Using the coordinates x = χ
m
1
and y = χ
m
2
of C
2
, the action of
ζ ∈ µ
d
≃N/N
′
on C
2
is given by
ζ (x, y) =(ζx, ζ
k
y).
Furthermore, U
σ
≃C
2
/µ
d
with respect to this action.
Proof. The general discussion in §1.3 shows that the quotient N/N
′
≃ Z/dZ acts
on the coordinate ring of C
2
via
(10.1.3) (u+N
′
) χ
m
′
=e
2πi¸m
′
,u)
χ
m
′
,
where m
′
∈ σ
∨
∩M
′
and u = je
1
for 0 ≤ j ≤d −1.
§10.1. Singularities of Toric Surfaces and Their Resolutions 455
An easy calculation shows that ¸m
1
, e
1
) = 1/d and ¸m
2
, e
1
) = k/d. Hence if
we set up the isomorphism µ
d
≃N/N
′
by mapping e
2πi j/d
→ je
1
+N
′
, then for all
ζ =e
2πi j/d
∈ µ
d
, we have
ζ (x, y) =(e
2πi j/d
x, e
2πi jk/d
y) =(ζx, ζ
k
y)
by (10.1.3). This what we wanted to show.
We next describe the slight but manageable ambiguity in the normal form for
2dimensional cones. Two cones are lattice equivalent if there is a bijective Z
linear mapping ϕ : N →N taking one cone to the other. After choice of basis for
N, such mappings are deﬁned by matrices in GL(2, Z).
Proposition 10.1.3. Let σ = Cone(e
2
, de
1
−ke
2
) and ¯ σ = Cone(e
′
2
,
¯
de
′
1
−
¯
ke
′
2
) be
cones in normal form that are lattice equivalent. Then
¯
d = d and either
¯
k = k or
k
¯
k ≡1 mod d.
Proof. Since the cones are lattice equivalent, writing N
′
and
¯
N
′
for the sublattices
as in (10.1.2), there is a bijective Zlinear mapping ϕ: N →N such that ϕ(N
′
) =
¯
N
′
.
Hence N/
¯
N
′
≃N/N
′
, so
¯
d =d. The statement about k and
¯
k is left to the reader in
Exercise 10.1.2.
Here are two examples to illustrate Proposition 10.1.2.
Example 10.1.4. First consider a cone
σ =Cone(e
2
, de
1
−e
2
)
with parameters d > 1 (so the cone is not smooth) and k = 1. This is precisely
the cone considered in Example 1.2.21. The corresponding toric surface U
σ
is
the rational normal cone
´
C
d
⊆ C
d+1
. The quotient
´
C
d
≃ C
2
/µ
d
was studied in
the special case d = 2 in Example 1.3.19, and the general case was described in
Exercise 1.3.11. With the notation of Proposition 10.1.2, ζ ∈ µ
d
acts on (x, y) ∈ C
2
via ζ (x, y) =(ζx, ζy) and the ring of invariants is
C[x, y]
µ
d
=C[x
d
, x
d−1
y, . . . , xy
d−1
, y
d
],
so
U
σ
≃C
2
/µ
d
≃Spec(C[x
d
, x
d−1
y, . . . , xy
d−1
, y
d
]).
On the other hand, from Example 1.2.21, we also have the description
U
σ
≃Spec(C[s, st, st
2
, . . . , st
d
]).
Exercise 10.1.3 studies the relation between these representations of the coordinate
ring of U
σ
. ♦
456 Chapter 10. Toric Surfaces
Example 10.1.5. Next consider a cone σ with parameters d and k = d −1, so
d =k +1. We will express everything in terms of the parameter k in the following.
Unlike the previous example, this is a case we have not encountered previously.
Note that k ≡−1 mod d. Hence by Proposition 10.1.2, the action of G=N/N
′
on
C
2
is given by
ζ (x, y) =(ζx, ζ
−1
y).
It is easy to check that the ring of invariants here is
C[x, y]
µ
k+1
=C[x
k+1
, y
k+1
, xy].
Moreover we have an isomorphism of rings
ϕ : C[X,Y, Z]/¸Z
k+1
−XY) ≃C[x
k+1
, y
k+1
, xy]
X →x
k+1
Y →y
k+1
Z →xy,
so we may identify the toric surface U
σ
with the variety V(Z
k+1
−XY) ⊆C
3
. ♦
The origin is unique singular point of the afﬁne variety of Example 10.1.5 and
is called a rational double point (or Du Val singularity) of type A
k
. Another stan
dard form of these singularities is given in Exercise 10.1.4. They are called double
points because the lowest degree nonzero term in the deﬁning equation has degree
two (i.e., the multiplicity of the singularity is two). The rational double points
are the simplest singularities from a certain point of view. The exact deﬁnition,
which we will give in §10.4, depends on the notion of a resolution of singularities,
which will be introduced shortly. All rational double points appear as singularities
of quotient surfaces C
2
/G where G is a ﬁnite subgroup of SU(2, C). There is a
complete classiﬁcation of such points in terms of the Dynkin diagrams of types
A
k
, D
k
, E
6
, E
7
, and E
8
. The groups corresponding to the diagrams D
k
, E
6
, E
7
, E
8
are
not abelian, so by the comment after (10.1.2), such points do not appear on toric
surfaces. We will see one way that the Dynkin diagram A
k
appears from the geom
etry of the toric surface U
σ
in Exercise 10.1.5, and we will return to this example
in §10.4. More details on these singularities can be found in [156, Ch. VI] and in
the article [47].
Here is another interesting aspect of Example 10.1.5. Recall that a normal
variety X is Gorenstein if its canonical divisor is Cartier (Deﬁnition 8.2.14). The
following result was proved in Exercise 8.2.13 of Chapter 8.
Proposition 10.1.6. For a cone σ =Cone(e
2
, de
1
−ke
2
) in normal form, the afﬁne
toric surface U
σ
is Gorenstein if and only if k =d −1.
§10.1. Singularities of Toric Surfaces and Their Resolutions 457
Toric Resolution of Singularities. Let X be a normal toric surface, and denote by
X
sing
the ﬁnite set of singular points of X (possibly empty).
Deﬁnition 10.1.7. A proper morphism ϕ : Y →X is a resolution of singularities
of X if Y is a smooth surface and ϕ induces an isomorphism of varieties
(10.1.4) Y ¸ ϕ
−1
(X
sing
) ≃X ¸ X
sing
.
Such a mapping modiﬁes X to produce a smooth variety without changing the
smooth locus X ¸ X
sing
. One of the most appealing aspects of toric varieties is the
way that many questions that are difﬁcult for general varieties admit simple and
concrete solutions in the toric case. The problem of ﬁnding resolutions of singu
larities is a perfect example. We illustrate this by constructing explicit resolutions
of singularities of the toric surfaces from Examples 10.1.4 and 10.1.5.
Example 10.1.8. Consider the rational normal cone of degree d, the afﬁne toric
surface U
σ
for σ =Cone(e
2
, de
1
−e
2
) studied in Example 10.1.4. Let Σ be the fan
in Figure 1 obtained by inserting a new ray τ = Cone(e
1
) subdividing σ into two
2dimensional cones:
σ
1
=Cone(e
2
, e
1
)
σ
2
=Cone(e
1
, de
1
−e
2
).
σ σ
1
σ
2
← τ
Figure 1. The cone σ and the reﬁnement given by σ1, σ2, τ
We now use some results from Chapter 3. The identity mapping on the lattice
N is compatible with the fans Σ and σ as in Deﬁnition 3.3.1. By Theorem 3.3.4,
we have a corresponding toric blowup morphism
(10.1.5) φ : X
Σ
−→U
σ
.
Note that both σ
1
and σ
2
(as well as all of their faces) are smooth cones. Hence
Theorem 3.1.19 implies that X
Σ
is a smooth surface. In addition, the toric mor
phism φ is proper by Theorem 3.4.7 since Σ is a reﬁnement of σ. Finally, we claim
that φ satisﬁes (10.1.4). This follows from the OrbitCone Correspondence on the
458 Chapter 10. Toric Surfaces
two surfaces: if p
σ
is the distinguished point corresponding to the 2dimensional
cone σ in (the singular point of U
σ
at the origin), then φ restricts to an isomorphism
X
Σ
¸φ
−1
(p
σ
) ≃U
σ
¸¦p
σ
¦ =(U
σ
)
smooth
.
The inverse image E = φ
−1
(p
σ
) is the curve on X
Σ
given by the closure of the
T
N
orbit O(τ) corresponding to the ray τ. That is, the singular point “blows up”
to E ≃ P
1
on the smooth surface. It follows that X
Σ
and the morphism (10.1.5)
give a toric resolution of singularities of the rational normal cone. We call E the
exceptional divisor on the smooth surface. We will say more about how E sits
inside the surface X
Σ
in §10.4. ♦
Example 10.1.9. We consider the case d = 4 of Example 10.1.5, for which the
surface U
σ
has a rational double point of type A
3
. We will leave the details, as well
as the generalization to all d ≥2, to the reader (Exercise 10.1.5). It is easy to ﬁnd
subdivisions of
σ =Cone(e
2
, 4e
1
−3e
2
)
yielding collections of smooth cones. The most economical way to do this is to
insert three new rays ρ
1
= Cone(e
1
), ρ
2
= Cone(2e
1
−e
2
), ρ
3
= Cone(3e
1
−2e
2
)
to obtain a fan Σ consisting of four 2dimensional cones and their faces.
The fan produced by this subdivision is somewhat easier to visualize if we draw
the cones relative to a different basis u
1
, u
2
for N. For u
1
=e
2
and u
2
=e
1
−e
2
, the
cone σ =Cone(u
1
, u
1
+4u
2
) and the fan Σ with maximal cones
(10.1.6)
σ
1
=Cone(u
1
, u
1
+u
2
)
σ
2
=Cone(u
1
+u
2
, u
1
+2u
2
)
σ
3
=Cone(u
1
+2u
2
, u
1
+3u
2
)
σ
4
=Cone(u
1
+3u
2
, u
1
+4u
2
)
appear in Figure 2.
σ σ
1
σ
2
σ
3
σ
4
Figure 2. The cone σ and the reﬁnement Σ
§10.1. Singularities of Toric Surfaces and Their Resolutions 459
You will check that each of these cones is smooth. Hence X
Σ
is a smooth
surface. Since Σ is a reﬁnement of σ, we have a proper toric morphism
φ : X
Σ
→U
σ
.
As in the previous example, φ restricts to an isomorphism from X
Σ
¸ φ
−1
(p
σ
) to
X
Σ
¸¦p
σ
¦. In this case, the exceptional divisor E =φ
−1
(p
σ
) is the union
E =V(τ
1
) ∪V(τ
2
) ∪V(τ
3
)
on X
Σ
. The curves V(τ
i
) are isomorphic to P
1
. The ﬁrst two intersect transversely
at the ﬁxed point of the T
N
action on X
Σ
corresponding to the cone σ
2
, while the
second two intersect transversely at the ﬁxed point corresponding to σ
3
. ♦
In these examples, we constructed toric resolutions of afﬁne toric surfaces with
just one singular point. The same techniques can be applied to any normal toric
surface X
Σ
.
Theorem 10.1.10. Let X
Σ
be a normal toric surface. There exists a smooth fan
Σ
′
reﬁning Σ such that the associated toric morphism φ : X
Σ
′
→X
Σ
is a toric
resolution of singularities.
Proof. It sufﬁces to show the existence of the smooth fan Σ
′
reﬁning Σ. The
reasoning given in Example 10.1.8 applies to show that the corresponding toric
morphism φ is proper and birational, hence a resolution of singularities of X
Σ
.
We will prove this by induction on an integer invariant of fans that measures
the complexity of the singularities on the corresponding surfaces. Let σ
1
, . . . , σ
ℓ
denote the 2dimensional cones in a fan Σ. For each i, we will write N
i
for the
sublattice of N generated by the ray generators of σ
i
. Then we deﬁne
s(Σ) =
ℓ
i=1
(mult(σ
i
) −1) ,
where mult(σ
i
) = [N : N
i
] as in §6.3. If s(Σ) = 0, then ℓ = 0 or mult(σ
i
) = 1 for
all i. It is easy to see that this implies that Σ is a smooth fan. Hence X
Σ
is already
smooth and we take this as the base case for our induction.
For the induction step, we assume that the existence of smooth reﬁnements has
been established for all fans Σ with s(Σ) < s, and consider a fan Σ with s(Σ) =s.
If s ≥ 1, then there exists some nonsmooth cone σ
i
in Σ. By Proposition 10.1.1,
there is a basis e
1
, e
2
for N such that σ
i
= Cone(e
2
, de
1
−ke
2
) with parameters
d > 0, 0 ≤k < d, and gcd(d, k) = 1. Consider the reﬁnement Σ
′
of Σ obtained by
subdividing the cone σ
i
into two new cones
σ
′
i
=Cone(e
2
, e
1
)
σ
′′
i
=Cone(e
1
, de
1
−ke
2
)
460 Chapter 10. Toric Surfaces
with a new 1dimensional cone ρ =Cone(e
1
). We must show that s(Σ
′
) < s(Σ) to
invoke the induction hypothesis and conclude the proof.
In s(Σ), the terms corresponding to the other cones σ
j
for j ,=i are unchanged.
The cone σ
′
i
is smooth since e
1
, e
2
is the normalized basis of N relative to σ
i
. So it
contributes a zero term in s(Σ
′
). Now consider the cone σ
′′
i
. In order to compute
its contribution to s(Σ
′
), we must determine the parameters of σ
′′
i
.
In terms of the basis e
1
, e
2
for N, the Zlinear mapping deﬁned by the matrix
A =
_
0 −1
1 0
_
(a “90degree rotation”) takes σ
′′
i
to Cone(e
2
, ke
1
+de
2
). Since A ∈ GL(2, Z), it
deﬁnes an automorphism of N, and hence σ
′′
i
will have the same parameters as
Cone(e
2
, ke
1
+de
2
). But now we apply (10.1.1) to write
(10.1.7) d =sk −l
where 0 ≤ l < k. Since gcd(d, k) = 1, we have gcd(k, l) = 1 as well. Hence the
cone σ
′′
i
has parameters k and l obtained from (10.1.7). Since k < d, if N
′′
i
is the
sublattice generated by the ray generators of σ
′′
i
, then by (10.1.2),
[N : N
′′
i
] =k <[N : N
i
] =d.
It follows that s(Σ
′
) < s(Σ), and the proof is complete by induction.
We will see in the next section that in the afﬁne case, the reﬁnement that gives
the resolution of singularities of U
σ
has a very nice description. As a preview,
notice that in Examples 10.1.8 and 10.1.9, the reﬁnement of the given cone σ was
produced by subdividing along the rays through the Hilbert basis (the irreducible
elements) of the semigroup σ∩N.
A resolution of a nonnormal toric surface singularity can be constructed by
ﬁrst saturating the associated semigroup as in Theorem 1.3.5, then applying the
results of this section. Toric resolutions of singularities for toric varieties of di
mension three and larger also exist. However, we postpone the higherdimensional
case until Chapter 11.
Exercises for §10.1.
10.1.1. Adapt the usual proof of the integer division algorithm to prove (10.1.1).
10.1.2. In this exercise, you will develop further properties of the parameters d, k in the
normal form for cones from Proposition 10.1.1 and prove part of Proposition 10.1.3.
(a) Show that if ¯ σ is obtained from a cone σ by parameters d, k by a Zlinear mapping of
N deﬁned by a matrix in GL(2, Z), then the parameter
¯
k of ¯ σ satisﬁes either
¯
k = k, or
k
¯
k ≡ 1 mod d. Hint: There is a choice of orientation to be made in the normalization
process. Recall that gcd(d, k) =1, so there are integers
¯
d,
¯
k such that d
¯
d +k
¯
k =1.
§10.2. Continued Fractions and Toric Surfaces 461
(b) Show that if σ is a cone with parameters d, k, then the dual cone σ
∨
⊆M
R
has param
eters d, d −k. Hint: Use the normal form for σ, write down σ
∨
in the corresponding
dual basis in M, then change bases in M to normalize σ
∨
.
10.1.3. With the notation in Example 10.1.4, show that
C[s, st, st
2
, . . . , st
d
] ≃C[x
d
, x
d−1
y, . . . , xy
d−1
, y
d
]
under s →x
d
and t →y/x, and use Proposition 10.1.2 to explain where these identiﬁcations
come from in terms of the semigroup S
σ
. Hint: We have s =χ
m1
and t =χ
m2
where e
1
, e
2
is the normalized basis for N and m
1
, m
2
is the dual basis for M.
10.1.4. In Example 10.1.5, we gave one form of the rational double point of type A
k
,
namely the singular point at (0, 0, 0) on the surface V = V(Z
k+1
−XY) ⊆ C
3
. Another
commonly used normal form for this type of singularity is the singular point at (0, 0, 0) on
the surface W =V(X
k+1
+Y
2
+Z
2
). Show that V and W are isomorphic as afﬁne varieties,
hence the singularities at the origin are analytically equivalent. Hint: There is a linear
change of coordinates in C
3
that does this.
10.1.5. In this exercise, you will check the claims made in Example 10.1.9 and show how
to extend the results there to the case σ =Cone(e
2
, de
1
−(d −1)e
2
) for general d.
(a) Check that each of the four cones in (10.1.6) is smooth, so that the toric surface X
Σ
is
smooth by Theorem 3.1.19.
(b) For general d, show how to insert new rays ρ
i
to subdivide σ and obtain a fan Σ whose
associated toric surface is smooth. Try to do this with as few new rays as possible.
Hence we obtain toric resolutions of singularities φ : X
Σ
→U
σ
for all d.
(c) Identify the inverse image C = φ
−1
(p
σ
) in general. For instance, how many irre
ducible components does C have? How are they connected? Hint: One way to repre
sent the structure is to drawa graph with vertices corresponding to the components and
connect two vertices by an edge if and only if the components intersect on X
Σ
. Do you
notice a relation between this graph and the Dynkin diagram A
k
= A
d−1
mentioned
before? We will discuss the relation in detail in §10.4.
§10.2. Continued Fractions and Toric Surfaces
To relate continued fractions to toric surfaces, we begin with the afﬁne toric surface
U
σ
of a cone σ ⊆ N
R
≃ R
2
in normal form with parameters d, k. We will always
assume d > k > 0, so that U
σ
has a unique singular point.
HirzebruchJung Continued Fractions. When we construct a resolution of singu
larities of U
σ
by following the proof of Theorem 10.1.10, the ﬁrst step is to reﬁne
the cone σ =Cone(e
2
, de
1
−ke
2
) to a fan containing the 2dimensional cones
σ
′
=Cone(e
2
, e
1
) and σ
′′
=Cone(e
1
, de
1
−ke
2
).
The ﬁrst is smooth, but the second may not be. However, we saw in the proof of
Theorem 10.1.10 that the cone σ
′′
has parameters k, k
1
satsifying
d =b
1
k −k
1
,
462 Chapter 10. Toric Surfaces
where b
1
≥2, 0 ≤k
1
< k as in (10.1.1). We used slightly different notation before,
writing s rather than b
1
and l rather than k
1
; the new notation will help us keep
track of what happens as we continue the process and reﬁne the cone σ
′′
.
Using the normalized basis for N relative to σ
′′
, we insert a new ray and obtain
a new smooth cone and a second, possibly nonsmooth cone with parameters k
1
, k
2
,
where
k =b
2
k
1
−k
2
using (10.1.1). Doing this repeatedly yields a modiﬁed Euclidean algorithm
(10.2.1)
d =b
1
k −k
1
k =b
2
k
1
−k
2
.
.
.
k
r−3
=b
r−1
k
r−2
−k
r−1
k
r−2
=b
r
k
r−1
that computes the parameters of the new cones produced as we successively sub
divide to produce the fan giving the resolution of singularities. The process termi
nates with k
r
= 0 for some r as shown, since as in the usual Euclidean algorithm,
the k
i
are a strictly decreasing sequence of nonnegative numbers. Also, by (10.1.1),
we have b
i
≥2 for all i.
The equations (10.2.1) can be rearranged:
(10.2.2)
d/k =b
1
−k
1
/k
k/k
1
=b
2
−k
2
/k
1
.
.
.
k
r−3
/k
r−2
=b
r−1
−k
r−1
/k
r−2
k
r−2
/k
r−1
=b
r
and spliced together to give a type of continued fraction expansion for the rational
number d/k, with minus signs:
(10.2.3) d/k =b
1
−
1
b
2
−
1
−
1
b
r
.
This is the HirzebruchJung continued fraction expansion of d/k. For obvious
typographical reasons, it is desirable to have a more compact way to represent
these expressions. We will use the notation
d/k =[[b
1
, b
2
, . . . , b
r
]].
§10.2. Continued Fractions and Toric Surfaces 463
The integers b
i
are the partial quotients of the HirzebruchJung continued fraction,
and the truncated HirzebruchJung continued fractions
[[b
1
, b
2
, . . . , b
i
]], 1 ≤i ≤r,
are the convergents.
Example 10.2.1. Consider the rational number 17/11. The HirzebruchJung con
tinued fraction expansion is
17/11 =[[2, 3, 2, 2, 2, 2]],
as may be veriﬁed directly using the modiﬁed Euclidean algorithm (10.2.1). ♦
Proposition 10.2.2. Let d > k > 0 be integers with gcd(d, k) = 1 and let d/k =
[[b
1
, . . . , b
r
]]. Deﬁne sequences P
i
and Q
i
recursively as follows. Set
(10.2.4)
P
0
=1, Q
0
=0
P
1
=b
1
, Q
1
=1,
and for all 2 ≤i ≤r, let
(10.2.5)
P
i
=b
i
P
i−1
−P
i−2
Q
i
=b
i
Q
i−1
−Q
i−2
.
Then the P
i
, Q
i
satisfy:
(a) The P
i
and Q
i
are increasing sequences of integers.
(b) [[b
1
, . . . , b
i
]] =P
i
/Q
i
for all 1 ≤i ≤r.
(c) P
i−1
Q
i
−P
i
Q
i−1
=1 for all 1 ≤i ≤r.
(d) The convergents form a strictly decreasing sequence:
d
k
=
P
r
Q
r
<
P
r−1
Q
r−1
< <
P
1
Q
1
.
Proof. The proof of part (a) is left to the reader (Exercise 10.2.1).
To prove part (b), ﬁrst observe that the expression on the right side of (10.2.3)
makes sense when the b
j
are any rational numbers (not just integers) such that all
denominators in (10.2.3) are nonzero. We will show that the sequences deﬁned by
(10.2.5) satisfy
[[b
1
, . . . , b
s
]] =
P
s
Q
s
for all such lists b
1
, . . . , b
s
. The proof is by induction on the length s of the list.
When s =1, we have [[b
1
]] =b
1
=
P
1
Q
1
by (10.2.4). Now assume that the result has
been proved for all lists of of length t and consider the expression
[[b
1
, . . . , b
t+1
]] =[[b
1
, . . . , b
t
−
1
b
t+1
]],
464 Chapter 10. Toric Surfaces
where the right side comes from a list of length t. By the induction hypothesis, this
equals
_
b
t
−
1
b
t+1
_
P
t−1
−P
t−2
_
b
t
−
1
b
t+1
_
Q
t−1
−Q
t−2
.
By the recurrences (10.2.5), this equals
P
t
−
1
b
t+1
P
t−1
Q
t
−
1
b
t+1
Q
t−1
=
b
t+1
P
t
−P
t−1
b
t+1
Q
t
−Q
t−1
=
P
t+1
Q
t+1
,
which is what we wanted to show.
Part (c) will be proved by induction on i. The base case i = 1 follows directly
from (10.2.4). Now assume that the result has been proved for i ≤ s, and consider
i =s +1. Using the recurrences (10.2.5), we have
P
s
Q
s+1
−P
s+1
Q
s
=P
s
(b
s
Q
s
−Q
s−1
) −(b
s
P
s
−P
s−1
)Q
s
=P
s−1
Q
s
−P
s
Q
s−1
=1
by the induction hypothesis.
Finally, from part (b), for each 1 ≤i ≤r −1, we have
P
i−1
Q
i−1
=
P
i
Q
i
+
1
Q
i−1
Q
i
.
Hence
P
i
Q
i
<
P
i−1
Q
i−1
since Q
i−1
Q
i
> 0 by part (a). Hence part (d) follows.
HirzebruchJung Continued Fractions and Resolutions. When σ is a cone with
parameters d > k > 0, the process of computing the HirzebruchJung continued
fraction of d/k yields a convenient method for ﬁnding a reﬁnement Σ of σ such
that φ : X
Σ
→U
σ
is a toric resolution of singularities.
Theorem 10.2.3. Let σ =Cone(e
2
, de
1
−ke
2
) be in normal form. Let u
0
=e
2
and
use the integers P
i
and Q
i
from Proposition 10.2.2 to construct vectors
u
i
=P
i−1
e
1
−Q
i−1
e
2
, 1 ≤i ≤r +1.
Then the cones
σ
i
=Cone(u
i−1
, u
i
), 1 ≤i ≤r +1,
have the following properties:
(a) Each σ
i
is a smooth cone and u
i−1
, u
i
are its ray generators.
(b) For each i, σ
i+1
∩σ
i
=Cone(u
i
).
§10.2. Continued Fractions and Toric Surfaces 465
(c) σ
1
∪ ∪σ
r+1
= σ, so the fan Σ consisting of the σ
i
and their faces gives a
smooth reﬁnement of σ.
(d) The toric morphism φ : X
Σ
→U
σ
is a resolution of singularities.
Proof. Both statements in part (a) follow easily from part (c) of Proposition 10.2.2.
For part (b), we note that the ratio −Q
i−1
/P
i−1
represents the slope of the line
through u
i
in the coordinate system relative to the normalized basis e
1
, e
2
for σ. By
part (d) of Proposition 10.2.2, these slopes form a strictly decreasing sequence for
i ≥0, which implies the statement in part (b).
Part (c) follows from part (b) by noting that u
0
= e
2
and P
r
/Q
r
= d/k, so
u
r+1
=de
1
−ke
2
. Hence the cones σ
i
ﬁll out σ.
Part (d) now follows by the reasoning used in Examples 10.1.8 and 10.1.5.
Example 10.2.4. Consider the cone σ = Cone(e
2
, 7e
1
−5e
2
) in normal form. To
construct the resolution of singularities of the afﬁne toric surface U
σ
, we simply
compute the HirzebruchJung continued fraction expansion of the rational number
d/k =7/5 using the modiﬁed Euclidean algorithm:
7 =2 5−3
5 =2 3−1
3 =3 1.
Hence b
0
=b
1
=2, b
2
=3, and
(10.2.6) 7/5 =[[2, 2, 3]].
Then from Proposition 10.2.2 we have
P
0
=1, Q
0
=0
P
1
=2, Q
1
=1
P
2
=b
2
P
1
−P
0
=3, Q
2
=b
2
Q
1
−Q
0
=2
P
3
=b
3
P
2
−P
1
=7, Q
3
=b
3
Q
2
−Q
1
=5.
Theorem 10.2.3 gives the vectors
u
0
=e
2
, u
1
=e
1
, u
2
=2e
1
−e
2
, u
3
=3e
1
−2e
2
, u
4
=7e
1
−5e
2
and the cones
(10.2.7)
σ
1
=Cone(e
2
, e
1
)
σ
2
=Cone(e
1
, 2e
1
−e
2
)
σ
3
=Cone(2e
1
−e
2
, 3e
1
−2e
2
)
σ
4
=Cone(3e
1
−2e
2
, 7e
1
−5e
2
)
shown in Figure 3 on the next page. The cones σ
i
give a smooth reﬁnement of σ.
You will see another example of this process in Exercise 10.2.2. ♦
466 Chapter 10. Toric Surfaces
σ
1
σ
2
σ
3
σ
4
Figure 3. The reﬁnement Σ and open circles showing ui =Pi−1e1 −Qi−1e2
Next we show that the vectors u
i
from Theorem 10.2.3 determine the partial
quotients in the HirzebruchJung continued fraction expansion of d/k.
Theorem 10.2.5. Let σ =Cone(e
2
, de
1
−ke
2
) be in normal form, and let
d/k =[[b
1
, b
2
, . . . , b
r
]].
Then the vectors u
0
, u
1
, . . . , u
r+1
constructed in Theorem 10.2.3 satisfy
(10.2.8) u
i−1
+u
i+1
=b
i
u
i
, b
i
≥2,
for 1 ≤i ≤r.
Proof. By the recurrences (10.2.5),
u
i−1
+u
i+1
=(P
i−2
e
1
−Q
i−2
e
2
) +(P
i
e
1
−Q
i
e
2
)
=(P
i−2
+P
i
)e
1
−(Q
i−2
+Q
i
)e
2
=b
i
(P
i−1
e
1
−Q
i−1
e
2
) =b
i
u
i
.
Later in this chapter we will see several important consequences of (10.2.8)
connected with the geometry of smooth toric surfaces.
The nonuniqueness of Proposition 10.1.3 has a nice relation to Theorem 10.2.3.
For instance, Example 10.2.4 used the HirzebruchJung expansion 7/5 =[[2, 2, 3]].
Since 5 3 ≡1 mod 7, the cone of Example 10.2.4 also has parameters d =7, k =3.
We leave it to the reader to check that
7/3 =[[3, 2, 2]],
with the partial quotients the same as those in (10.2.6), but listed in reverse order.
This pattern holds for all HirzebruchJung continued fractions. We give a proof
that uses the properties of the associated toric surfaces.
§10.2. Continued Fractions and Toric Surfaces 467
Proposition 10.2.6. Let 0 <k,
¯
k <d and assume k
¯
k ≡1 mod d. If the Hirzebruch
Jung continued fraction expansion of d/k is
d/k =[[b
1
, b
2
, . . . , b
r
]],
then the HirzebruchJung continued fraction expansion of d/
¯
k is
d/
¯
k =[[b
r
, b
r−1
, . . . , b
1
]].
Proof. Let σ = Cone(e
2
, de
1
−ke
2
) and ¯ σ = Cone(e
2
, de
1
−
¯
ke
2
) be the corre
sponding cones in normal form. Since k
¯
k ≡ 1 mod m, there is an integer
¯
d such
that d
¯
d +k
¯
k =1. The Zlinear mapping ϕ: N →N deﬁned with respect to the basis
e
1
, e
2
by the matrix
A =
_
¯
k d
¯
d −k
_
is bijective, maps ¯ σ to σ, and is orientationreversing. Thus ϕ(de
1
−
¯
ke
2
) =e
2
and
ϕ(e
2
) = de
1
−ke
2
. If we apply Theorem 10.2.3 to σ, then we obtain vectors u
i
satisfying the equations
u
i−1
+u
i+1
=b
i
u
i
for all 1 ≤ i ≤ r. We claim that when we apply the mapping ϕ
−1
deﬁned by the
inverse of the matrix A above, then the vectors u
i
are taken to corresponding vec
tors ¯ u
i
for the cone ¯ σ. But since ϕ and ϕ
−1
are orientationreversing, the partial
quotients in the HirzebruchJung continued fraction will be listed in the opposite
order. You will complete the proof of this assertion in Exercise 10.2.3.
Hilbert Bases and Convex Hulls. Our next result gives two alternative ways to
understand the vectors u
i
in Theorem 10.2.3. The idea is that σ gives two objects:
• The semigroup σ ∩N. Since σ is strongly convex, its irreducible elements
form the unique minimal generating set called the Hilbert basis of σ∩N. (See
Proposition 1.2.22.)
• The convex hull Θ
σ
= Conv(σ ∩(N ¸¦0¦)). This is an unbounded polygon in
the plane whose bounded edges contain ﬁnitely many lattice points.
Example 10.2.7. Consider the cone σ =Cone(e
2
, 7e
1
−5e
2
) from Example 10.2.4.
Figure 4 on the next page shows the convex hull Θ
σ
, where the white circles rep
resent the lattice points on the bounded edges. We will see below that these lattice
points give the Hilbert basis of σ ∩N. ♦
Here is the general result suggested by Example 10.2.7.
Theorem 10.2.8. Let σ =Cone(e
2
, de
1
−ke
2
) be in normal form and let
S =¦u
0
, u
1
, . . . , u
r+1
¦
be the set of vectors constructed in Theorem 10.2.3. Then:
468 Chapter 10. Toric Surfaces
Θ
σ
Figure 4. The convex hull Θσ and the lattice points on its bounded edges
(a) S is the Hilbert basis of the semigroup σ ∩N.
(b) S is the set of lattice points on the bounded edges of Θ
σ
.
Proof. For part (a), we use the notation of Theorem 10.2.3, where the cone σ
i
is
generated by u
i−1
, u
i
. Then σ
i
∩N is generated as a semigroup by u
i−1
, u
i
since σ
i
is smooth. Using σ =σ
1
∪ ∪σ
r+1
, one sees easily that S generates σ ∩N.
We claim next that all the u
i
are irreducible elements of σ∩N. This is clear for
u
0
=e
2
and u
r+1
=de
1
−ke
2
since they are the ray generators for σ. If 1 ≤i ≤r and
u
i
is not irreducible, then u
i
would have to be a linear combination of the vectors
in S¸¦u
i
¦ with nonnegative integer coefﬁcients, i.e.,
u
i
=P
i−1
e
1
−Q
i−1
e
2
=
j,=i
c
j
u
j
=
_
j,=i
c
j
P
j−1
_
e
1
−
_
j,=i
c
j
Q
j−1
_
e
2
with c
j
≥0 in Z. Hence
P
i−1
=
j,=i
c
j
P
j−1
, Q
i−1
=
j,=i
c
j
Q
j−1
.
Since the P
i
and Q
i
are strictly increasing by part (a) of Proposition 10.2.2, we must
have c
j
=0 for all j > i. But this would imply that u
i
is a linear combination with
nonnegative integer coefﬁcients of the vectors in ¦u
0
, . . . , u
i−1
¦. This contradicts
the observation made in the proof of Theorem 10.2.3 that the slopes of the u
i
are
strictly decreasing. It follows that the u
i
are irreducible elements of σ ∩N.
Finally, we must show that there are no other irreducible elements in σ ∩N.
But this follows from what we have already said. Since σ = σ
1
∪ ∪σ
r+1
, if u
is irreducible, then u ∈ σ
i
∩N for some i. But then u = c
i−1
u
i−1
+c
i
u
i
for some
c
i−1
, c
i
≥0 in Z. Thus u is irreducible only if u =u
i−1
or u
i
.
§10.2. Continued Fractions and Toric Surfaces 469
For part (b), ﬁrst observe that by Proposition 10.2.2, we have
P
i−2
Q
i
−P
i
Q
i−2
=P
i−2
(b
i
Q
i−1
−Q
i−2
) −(b
i
P
i−1
−P
i−2
)Q
i−2
=b
i
(P
i−2
Q
i−1
−P
i−1
Q
i−2
) =b
i
≥2.
Combining this with part (c) of Proposition 10.2.2, one obtains the inequality
−(Q
i−1
−Q
i−2
)
P
i−1
−P
i−2
≤
−(Q
i
−Q
i−1
)
P
i
−P
i−1
.
Since u
i
= P
i−1
e
1
−Q
i−1
e
2
, inequality tells us that the slopes of the line segments
u
i−1
u
i
and u
i
u
i+1
are related by
slope of u
i−1
u
i
≤slope of u
i
u
i+1
.
This implies that these line segements lie on boundary of Θ
σ
. From here, it is easy
to see that the u
i
are the lattice points of the bounded edges of Θ
σ
.
Ordinary Continued Fractions. The HirzebruchJung continued fractions studied
above are less familiar than ordinary continued fraction expansions in which the
minus signs are replaced by plus signs. If d > k > 0 are integers, then the or
dinary continued fraction expansion of d/k may be obtained by performing the
same sequence of integer divisions used in the usual Euclidean algorithm for the
gcd. Starting with k
−1
= d and k
0
= k, we write a
i
for the quotient and k
i
for the
remainder at each step, so that the ith division is given by
(10.2.9) k
i−2
=a
i
k
i−1
+k
i
,
where 0 ≤k
i
< k
i−1
.
Let k
s−1
be the ﬁnal nonzero remainder (which equals gcd(d, k)). The resulting
equations splice together to form the ordinary continued fraction
d/k =a
1
+
1
a
2
+
1
+
1
a
s
.
To distinguish these from HirzebruchJung continued fractions, we will use the
notation
(10.2.10) d/k =[a
1
, a
2
, . . . , a
s
]
for the ordinary continued fraction. The a
i
are the (ordinary) partial quotients of
d/k, and the truncated continued fractions
[a
1
, a
2
, . . . , a
i
], 1 ≤i ≤s,
are the (ordinary) convergents of d/k.
470 Chapter 10. Toric Surfaces
Example 10.2.9. By Example 10.2.1, the HirzebruchJung continued fraction of
17/11 is
17/11 =[[2, 3, 2, 2, 2, 2]],
and the ordinary continued fraction is
17/11 =[1, 1, 1, 5].
The partial quotients and the lengths are different. However, each expansion deter
mines the other, and there are methods for computing the HirzebruchJung partial
quotients b
j
in terms of the ordinary partial quotients a
i
and vice versa. See [41,
Prop. 3.6], [88, p. 257], or [146, Prop. 2.3]. See also Exercise 10.2.4. ♦
The following result is mostly parallel to Proposition 10.2.2, but shows that
ordinary continued fractions are slightly more complicated than HirzebruchJung
continued fractions. The proof is left to the reader (Exercise 10.2.5).
Proposition 10.2.10. Let d > k > 0 be integers with gcd(d, k) = 1, and let d/k =
[a
1
, . . . , a
s
]. Deﬁne sequences p
i
and q
i
recursively as follows. First set
(10.2.11)
p
0
=1, q
0
=0
p
1
=a
1
, q
1
=1,
and for all 2 ≤i ≤s, let
(10.2.12)
p
i
=a
i
p
i−1
+p
i−2
q
i
=a
i
q
i−1
+q
i−2
.
Then the p
i
, q
i
satisfy:
(a) [a
1
, . . . , a
i
] = p
i
/q
i
for all 1 ≤i ≤s.
(b) p
i
q
i−1
−p
i−1
q
i
=(−1)
i
for all 1 ≤i ≤s.
(c) The convergents converge to d/k, but in an oscillating fashion:
p
1
q
1
<
p
3
q
3
< ≤
d
k
≤ <
p
4
q
4
<
p
2
q
2
.
Ordinary Continued Fractions and Convex Hulls. Felix Klein discovered a lovely
geometric interpretation of ordinary continued fractions. Given a basis u
o
−1
, u
o
0
of
N ≃ Z
2
and relatively prime integers d > k > 0, compute the continued fraction
d/k =[a
1
, . . . , a
s
] and set
(10.2.13) u
o
i
=q
i
u
o
−1
+p
i
u
o
0
, 1 ≤i ≤s.
In this notation, the superscript “o” stands for “ordinary.” Then u
o
s
= ku
o
−1
+du
o
0
,
and part (c) of Proposition 10.2.10 implies that the u
o
i
lie on one side of the ray
Cone(u
o
s
) for even indices and on the other side for odd indices. To give a careful
description of what is happening, we introduce the cones
σ
−1
=Cone(u
o
−1
, u
o
s
), σ
0
=Cone(u
o
0
, u
o
s
)
and associated convex hulls Θ
i
=Θ
σ
i
=Conv(σ
i
∩(N¸¦0¦)), i =−1, 0.
§10.2. Continued Fractions and Toric Surfaces 471
Θ
0
Θ
−1
σ
0
σ
−1
u
0
o
u
−1
o
u
1
o
u
3
o
u
2
o
Figure 5. The cones σ−1, σ0, the convex hulls Θ−1, Θ0, and their vertices
Example 10.2.11. For d =7, k =5, the expansion 7/5 =[1, 2, 2] gives the vectors
u
o
−1
, . . . , u
o
3
shown in Figure 5. In this ﬁgure, it is clear that the u
o
i
are the vertices
of the convex hulls Θ
−1
and Θ
0
. ♦
This example is a special case of the following general result.
Theorem 10.2.12. For u
o
−1
, . . . , u
o
s
and Θ
−1
, Θ
0
as above, we have:
(a) Θ
−1
has vertex set ¦u
o
2j−1
[ 1 ≤ j ≤⌊s/2⌋¦∪¦u
o
s
¦.
(b) Θ
0
has vertex set ¦u
o
2j
[ 0 ≤ j ≤⌊s/2⌋¦∪¦u
o
s
¦.
(c) For 1 ≤ i ≤ s, u
o
i−2
u
o
i
is an edge of Θ
−1
(resp. Θ
0
) for i odd (resp. even) with
a
i
+1 lattice points.
Proof. First note that by Proposition 10.2.10, the vectors u
o
i
satisfy the recursion
(10.2.14) u
o
i
=a
i
u
o
i−1
+u
o
i−2
, 1 ≤i ≤s.
Since u
o
i−1
is primitive (Proposition 10.2.10), part (c) follows from parts (a) and (b).
We now prove the theorem using induction on the length s of the continued
fraction expansion. Consider Figure 6 on the next page, which shows the ﬁrst quad
rant determined by u
o
−1
, u
o
0
, together with the vector u
o
s
. In the picture, σ
0
(darker)
lies above the ray determined by u
o
s
and σ
−1
(lighter) lies below. The ﬁgure also
shows u
o
1
and the smaller cone σ
1
=Cone(u
o
1
, u
o
s
) ⊆σ
−1
=Cone(u
o
−1
, u
o
s
).
Since u
o
s
= ku
o
−1
+du
o
0
, the ray starting from u
o
−1
through u
o
1
passes through
the upper edge of σ
−1
at a point between u
o
−1
+⌊d/k⌋u
o
1
and u
o
−1
+(⌊d/k⌋+1)u
o
1
.
But by the computation of the ordinary continued fraction, ⌊d/k⌋ = a
1
. Therefore
the segment from u
o
−1
to u
o
1
is the ﬁrst bounded edge of Θ
−1
. It follows that
Θ
−1
∩Cone(u
o
−1
, u
o
1
)
472 Chapter 10. Toric Surfaces
σ
0
σ
1
σ
−1
u
0
o
u
−1
o
u
1
o
= a
1
u
0
o
+ u
−1
o
u
s
o
Figure 6. The cones σ0, σ1 ⊆σ−1 and vectors u
o
−1
, u
o
0
, u
o
1
, u
o
s
has vertices u
o
−1
, u
o
1
. It remains to understand Θ
0
and Θ
−1
∩σ
1
. This is where
induction comes into play.
Apply the above construction to the new basis u
o
0
, u
o
1
and the continued fraction
k
d −a
1
k
=
1
d
k
−a
1
=[a
2
, . . . , a
s
]
of length s −1. If we start numbering at 0 rather than at −1, then the vectors
u
o
0
, u
o
1
, . . . , u
o
s
are the same as before by the recursion (10.2.14). This gives the
cones σ
0
, σ
1
shown in Figure 6. By induction, the vectors u
o
0
, u
o
1
, . . . , u
o
s
give
the vertices of the corresponding convex hulls Θ
0
= Conv(σ
0
∩(N ¸ ¦0¦)) and
Θ
1
= Conv(σ
1
∩(N ¸¦0¦)). It follows easily that the theorem holds for continued
fraction expansions of length s.
A discussion of Klein’s formulation of Theorem 10.2.12 can be found in [146].
In Exercise 10.2.6 you will apply the theorem to the resolution of pairs of singular
points on certain toric surfaces. Geometric pictures similar to Figure 5 have also
appeared in recent work of McDuff on symplectic embeddings of 4dimensional
ellipsoids (see [125]).
Ordinary Continued Fractions and the Supplementary Cone. To relate the above
theorem to toric geometry, we follows the approach of [146]. Given a cone σ =
Cone(e
2
, de
1
−ke
2
) ⊆ N
R
≃ R
2
in normal form, its supplement is the cone σ
′
=
Cone(−e
2
, de
1
−ke
2
). Thus σ ∪σ
′
is the right halfplane, and the cones σ, σ
′
give
the convex hulls
Θ
σ
=Conv(σ ∩(N ¸¦0¦))
Θ
σ
′
=Conv(σ
′
∩(N¸¦0¦)).
Example 10.2.13. When d =7, k =5, Figure 7 on the next page shows the cones
σ, σ
′
and the convex hulls Θ
σ
, Θ
σ
′ . The open circles in the ﬁgure are the vertices
§10.2. Continued Fractions and Toric Surfaces 473
Θ
σ
Θ
σ′
σ′
σ
7e
1
− 5e
2
e
2
e
1
Figure 7. The cones σ, σ
′
, the convex hulls Θσ, Θ
σ
′ , and their vertices
of Θ
σ
and Θ
σ
′ . Also observe that the fourth quadrant portion of Figure 7 becomes
Figure 5 after a 90
◦
counterclockwise rotation. ♦
For σ =Cone(e
2
, de
1
−ke
2
), the ordinary continued fraction expansion d/k =
[a
1
, a
2
, . . . , a
s
] gives the sequences p
i
, q
i
, 0 ≤ i ≤ s, deﬁned in Proposition 10.2.10.
Then deﬁne the vectors
(10.2.15) u
o
−1
=−e
2
, u
o
i
= p
i
e
1
−q
i
e
2
, 0 ≤i ≤s.
These vectors enable us to describe the vertices of Θ
σ
, Θ
σ
′
as follows.
Theorem 10.2.14. Let σ = Cone(e
2
, de
1
−ke
2
) be a cone in normal form with
supplement σ
′
. Also let u
o
i
, −1 ≤i ≤s, be as deﬁned in (10.2.15). Then:
(a) Then set of vertices of Θ
σ
′ is
¦u
o
2j−1
[ 1 ≤ j ≤⌊s/2⌋¦∪¦u
o
s
¦.
(b) If a
1
=1, then the set of vertices of Θ
σ
is
¦e
2
¦∪¦u
o
2j
[ 1 ≤ j ≤⌊s/2⌋¦∪¦u
o
s
¦.
(c) If a
1
> 1, then the set of vertices of Θ
σ
is
¦e
2
¦∪¦u
o
2j
[ 0 ≤ j ≤⌊s/2⌋¦∪¦u
o
s
¦.
Proof. First note that since u
o
−1
=−e
2
and u
o
0
=e
1
, we can rewrite (10.2.15) as
u
o
i
= p
i
e
1
−q
i
e
2
=q
i
u
−1
+p
i
u
0
, 0 ≤i ≤s.
Thus we are in the situation of Theorem 10.2.12, where we have the cones σ
−1
=
Cone(u
o
−1
, u
o
s
) and σ
0
= Cone(u
o
0
, u
o
s
) and associated convex hulls Θ
−1
= Θ
σ
−1
and Θ
0
=Θ
σ
0
. Then Theorem 10.2.12 implies that the vectors u
o
−1
, u
o
0
, . . . , u
o
s
give
the vertices of Θ
−1
and Θ
0
.
474 Chapter 10. Toric Surfaces
However, σ = Cone(e
1
, e
2
) ∪σ
0
and σ
′
= σ
−1
. In particular, Θ
σ
′ = Θ
−1
, so
that part (a) of the theorem follows immediately. For parts (b) and (c), note that
Θ
σ
=(Θ
σ
∩Cone(e
1
, e
2
)) ∪Θ
0
.
The intersection Θ
σ
∩Cone(e
1
, e
2
) has vertices e
1
= u
o
0
, e
2
, while Θ
0
has vertices
u
o
0
, u
o
2
, . . . , u
o
s
. If a
1
= 1, then e
2
, u
o
0
, u
o
2
are collinear, so that u
o
0
is not a vertex.
This proves part (b) of the theorem. Finally, if a
1
> 1, then one can prove without
difﬁculty that u
o
0
is a vertex (Exercise 10.2.7), and part (c) follows.
Example 10.2.15. Figure 7 above illustrates Theorem 10.2.12 for the cone σ =
Cone(e
2
, 7e
1
−5e
2
). Since 7/5 =[1, 2, 2], we use part (b) of the theorem. ♦
Ordinary Continued Fractions and the Dual Cone. For a cone σ in normal form,
one surprise is that its supplement σ
′
is essentially the dual of σ.
Lemma 10.2.16. Given a cone σ = Cone(e
2
, de
1
−ke
2
) ⊆ N
R
in normal form, its
supplementary cone σ
′
⊆N
R
is isomorphic to σ
∨
⊆M
R
.
Proof. Let e
∗
1
, e
∗
2
be the basis of M dual to the basis e
1
, e
2
of N. Then the isomor
phism deﬁned by e
1
→e
∗
2
and e
2
→−e
∗
1
takes σ
′
=Cone(−e
2
, de
1
−ke
2
) to
Cone(−(−e
∗
1
), d(e
∗
2
) −k(−e
∗
1
)) =Cone(e
∗
1
, ke
∗
1
+de
∗
2
) =σ
∨
.
The isomorphism σ
′
≃ σ
∨
from this lemma leads to some nice results about
σ
∨
and the associated convex hull Θ
σ
∨ =Conv(σ
∨
∩(M¸¦0¦)). Speciﬁcally, this
isomorphism takes the vectors
u
o
−1
=−e
2
, u
o
i
= p
i
e
1
−q
i
e
2
, 0 ≤i ≤s
from (10.2.15) to the dual vectors
m
−1
=e
∗
1
, m
i
=q
i
e
∗
1
+p
i
e
∗
2
, 0 ≤i ≤s.
In particular, m
s
= ke
∗
1
+de
∗
2
, so that σ
∨
= Cone(m
−1
, m
s
) in this notation. Then
Theorem 10.2.14 implies that the vertex set of the convex hull Θ
σ
∨ is
(10.2.16) ¦m
2j−1
[ 1 ≤ j ≤⌊s/2⌋¦∪¦m
s
¦.
Hence, in the language of §7.1, Θ
σ
∨ ⊆M
R
is a full dimensional lattice polyhedron.
We can describe its recession cone and normal fan as follows.
Proposition 10.2.17. Let σ = Cone(e
2
, de
1
−ke
2
) ⊆ N
R
≃ R
2
be in normal form
and set P =Θ
σ
∨ =Conv(σ
∨
∩(M¸¦0¦)). Then:
(a) P ⊆M
R
is a full dimensional lattice polyhedron with recession cone σ
∨
.
(b) The normal fan Σ
P
of P is the reﬁnement of σ obtained by adding the minimal
generators
u
o
0
, u
o
2
, u
o
4
, . . . , u
o
s−1
(s odd)
u
o
0
, u
o
2
, u
o
4
, . . . , u
o
s−2
, u
o
s
−u
o
s−1
(s even).
§10.2. Continued Fractions and Toric Surfaces 475
(c) The toric morphism X
P
→U
σ
is projective and X
P
is Gorenstein with at worst
rational double points.
Remark 10.2.18. When s is even, we can explain u
o
s
−u
o
s−1
as follows. By Theo
rem 10.2.14, u
o
s−2
and u
o
s
give an edge Θ
σ
. The vector determined by this edge is
u
o
s
−u
o
s−2
=a
s
u
o
s−1
. Hence the edge has a
s
+1 lattice points since u
o
s−1
is primitive.
Thus, if we start from u
o
s
, the next lattice point along the edge is
u
o
s
−u
o
s−1
.
Since d > k > 0 and d/k =[a
1
, . . . , a
s
] is computed using the Euclidean algorithm,
we have a
s
≥2 (Exercise 10.2.8). It follows that u
o
s
−u
o
s−1
not a vertex of Θ
σ
.
Proof. Part (a) is straightforward (Exercise 10.2.9). For part (b), we will assume
that s is even and leave the case when s is odd to the reader (Exercise 10.2.9).
If we write s =2ℓ, then the vertices (10.2.16) of P are m
−1
, m
1
, . . . , m
2ℓ−1
, m
2ℓ
.
Thus the bounded edges of P =Θ
σ
∨ are
m
−1
m
1
, m
1
m
3
, . . . , m
2ℓ−3
m
2ℓ−1
, m
2ℓ−1
m
2ℓ
.
The inwardpointing normals of these edges give the rays that reﬁne σ in the normal
fan of P.
The m
i
’s satisfy the same recursion m
i
=a
i
m
i−1
+m
i−2
, 1 ≤ i ≤ s, as the u
o
i
’s
in (10.2.14). Hence, for the edges m
2j−1
m
2j+1
, 1 ≤ j ≤⌊s/2⌋, we have
m
2j+1
−m
2j−1
=a
2j+1
m
2j
=a
2j+1
(q
2j
e
∗
1
+p
2j
e
∗
2
).
Since u
o
2j
= p
2j
e
1
−q
2j
e
2
, one easily computes that ¸m
2j+1
−m
2j−1
, u
o
2j
) = 0. It
follows that u
o
2j
is the inwardpointing normal of this edge since it lies in σ and is
primitive. This takes care of all of the bounded edges except for m
2ℓ−1
m
2ℓ
. Here,
we compute
m
2ℓ
−m
2ℓ−1
=(q
2ℓ
−q
2ℓ−1
)e
∗
1
+(p
2ℓ
−p
2ℓ−1
)e
∗
2
.
This is clearly normal to u
o
2ℓ
−u
o
2ℓ−1
=(p
2ℓ
−p
2ℓ−1
)e
1
−(q
2ℓ
−q
2ℓ−1
)e
2
. The latter
vector is easily seen to be primitive by part (b) of Proposition 10.2.10. Furthermore,
u
o
2ℓ
−u
o
2ℓ−1
∈ σ by Remark 10.2.18. Hence this is the inwardpointing normal of
the ﬁnal bounded edge m
2ℓ−1
m
2ℓ
.
For part (c), note that X
P
→U
σ
is projective by Theorem 7.1.10. To complete
the proof, we need to show that each maximal cone of Σ
P
gives a Gorenstein afﬁne
toric variety. For simplicity, we assume that s is odd (see Exercise 10.2.9 for the
even case). Since σ = Cone(e
2
, de
1
−ke
2
) = Cone(e
2
, u
o
s
), the maximal cones of
Σ
P
consist of two “boundary cones” Cone(e
1
, u
o
0
) and Cone(u
o
s−1
, u
o
s
), plus the
“interior cones” Cone(u
2j−2
, u
2j
) ∈ Σ
P
, 1 ≤ j ≤ ⌊s/2⌋. The boundary cones are
easily seen to be smooth and hence Gorenstein (Exercise 10.2.9). For an interior
476 Chapter 10. Toric Surfaces
cone Cone(u
2j−2
, u
2j
), we use part (b) of Proposition 10.2.10 to compute
¸m
2j−1
, u
o
2j−2
) =¸q
2j−1
e
∗
1
+p
2j−1
e
∗
2
, p
2j−2
e
1
−q
2j−2
e
2
)
=−(p
2j−1
q
2j−2
−p
2j−2
q
2j−1
) =−(−1)
2j−1
=1,
and a similar computation gives ¸m
2j−1
, u
o
2j
) = p
2j
q
2j−1
−p
2j−1
q
2j
=(−1)
2j
=1.
By Proposition 8.2.12, we conclude that the corresponding afﬁne toric variety is
Gorenstein, as desired. Then the singular points of X
P
are rational double points
by Example 10.1.5 and Proposition 10.1.6.
In §11.3 we will revisit this result, where we will learn that the morphism
X
P
→U
σ
from Proposition 10.2.17 is the blowup of the singular point of U
σ
.
Just as the morphism X
P
→U
σ
is projective, one can show more generally
that the resolution of singularities X
Σ
→U
σ
from Theorem 10.2.3 is a projective
morphism. This requires ﬁnding a lattice polyhedron with the correct normal fan.
You will explore one way of doing this in Exercise 10.2.10.
We next consider the Hilbert basis H of σ
∨
∩M. Recall from Lemma 1.3.10
that [H [ is the dimension of the Zariski tangent space at the singular point of U
σ
and is the dimension of the most efﬁcient embedding of U
σ
into afﬁne space.
Theorem 10.2.8 tells us that the Hilbert basis of σ ∩N is computed using the
HirzebruchJung continued fraction expansion of d/k. Since σ
∨
has parameters
d, d −k (part (b) of Exercise 10.1.2), it follows that we need the HirzebruchJung
continued fraction expansion of d/(d −k) to get the Hilbert basis of σ
∨
∩M. By
Exercise 10.2.4, the ordinary continued fraction
d/k =[a
1
, . . . , a
s
]
gives the HirzebruchJung continued fraction
d/(d −k) =
_
[[(2)
a
1
−1
, a
2
+2, (2)
a
3
−1
, a
4
+2, . . . , (2)
a
s−1
−1
, a
s
+2]] s even
[[(2)
a
1
−1
, a
2
+2, (2)
a
3
−1
, a
4
+2, . . . , a
s−1
+2, (2)
as−1
]] s odd.
Theorem 10.2.8, applied to this expansion, gives the Hilbert basis of σ
∨
∩M. To see
the underlying geometry, we need the following observation (Exercise 10.2.11):
(10.2.17)
In Theorem 10.2.5, three consecutive lattice points u
i−1
, u
i
, u
i+1
are collinear if and only if u
i−1
+u
i+1
=2u
i
, i.e., b
i
=2.
This means that a string of consecutive 2’s in a HirzebruchJung continued fraction
of d/(d −k) gives a string of lattice points in the relative interior of a bounded
edge of the convex hull Θ
σ
∨. For the vertices m
−1
, m
1
, . . . of Θ
σ
∨, this gives two
ways to think about lattice points on an edge connecting two adjacent vertices. For
example, the edge m
−1
m
1
has a
1
−1 lattice points in its relative interior because:
• The HirzebruchJung expansion of d/(d−k) starts with a
1
−1 consecutive 2’s.
• m
1
−m
−1
=a
1
u
0
, u
0
primitive, gives a
1
+1 lattice points on the edge.
This pattern continues for the other bounded edges of Θ
σ
∨.
§10.2. Continued Fractions and Toric Surfaces 477
The connection between continued fractions and toric surfaces is surprisingly
rich and varied and is one of the reasons why toric varieties are so much fun to
study. See [41] and [146] for a further discussion of this wonderful topic.
Exercises for §10.2.
10.2.1. Prove part (a) of Proposition 10.2.2: Show that the sequences P
i
and Q
i
from
(10.2.5) are increasing sequences of nonnegative numbers. Hint: Use b
i
≥2 for all i.
10.2.2. In §10.1, we constructed several resolutions of singularities in a rather ad hoc way.
In this exercise, we will see that the resolutions given by Theorem 10.2.3 are the same as
what we saw before.
(a) When σ has parameters d, 1, showthat the HirzebruchJung continued fraction method
gives the same resolution of U
σ
as the one given in Example 10.1.8.
(b) Do the same for Example 10.1.9. Hint: First show that
d
d −1
=[[2, 2, . . . , 2]],
where there are d −1 2’s.
10.2.3. Verify the last claim in the proof of Proposition 10.2.6.
10.2.4. This exercise will consider some relations between ordinary continued fraction
expansions and HirzebruchJung continued fraction expansions.
(a) Given integers a
1
, a
2
> 0 and a variable x, prove that
[a
1
, a
2
, x] =[[a
1
+1, (2)
a2−1
, x +1]],
where for any l ≥0, (2)
l
denotes a string of l 2’s. Hint: Argue by induction on a
1
.
(b) Use part (a) to prove the equality [1, 1, 1, 5] =[[2, 3, 2, 2, 2, 2]] from Example 10.2.9.
(c) Given d/k = [a
1
, . . . , a
s
], prove that d/(d −k) has the HirzebruchJung expansion
given in the discussion leading up to (10.2.17). Hint: If you get stuck, see [146].
10.2.5. In this exercise we will consider Proposition 10.2.10.
(a) Prove the proposition. Hint: For part (a), argue by induction on the length s ≥ 1 of
the expansion. The expression [a
1
, a
2
, . . . , a
i
] is welldeﬁned when the a
i
are positive
rational numbers, so we can write
[a
1
, a
2
, . . . , a
i−1
, a
i
] =[a
1
, a
2
, . . . , a
i−1
+1/a
i
].
Then use the recurrences in (10.2.12). Then follow the reasoning from the proof of
Proposition 10.2.2.
(b) Suppose we modify the initialization and the recurrences (10.2.12) as follows. Let
r
−1
=1, s
−1
=0
r
0
=0, s
0
=1.
Then, for all 1 ≤i ≤s, compute
r
i
=r
i−2
−a
i
r
i−1
s
i
=s
i−2
−a
i
s
i−1
(note the change in sign!). What is true about r
i
d +s
i
k for all i? What do we get with
i =s? Hint: This fact is the basis for the extended Euclidean algorithm.
478 Chapter 10. Toric Surfaces
10.2.6. In this exercise, you will show that the ordinary continued fraction expansion of a
rational number d/k can be used to construct simultaneous resolutions of pairs of singular
ities of certain toric surfaces. Let 1 < k < d be relatively prime integers, and let P be the
triangle Conv(0, de
1
, ke
2
) in R
2
.
(a) Draw the normal fan Σ
P
and show that X
P
has exactly two singular points. Note that
X
P
is isomorphic to the weighted projective plane P(1, k, d).
(b) Adapt Theorem 10.2.12 to produce a resolution of singularities of X
P
from the or
dinary continued fraction expansion of d/k. Hint: First reﬁne Σ
P
by introducing 1
dimensional cones Cone(−e
1
) and Cone(−e
2
). Then apply Theorem 10.2.12 to the
third quadrant of your drawing.
10.2.7. Complete the proof of Theorem 10.2.14 by showing that u
o
0
is a vertex of Θ
σ
if
and only if a
1
> 1.
10.2.8. For relative prime integers d >k >0, we used the Euclidean algorithmto construct
d/k =[a
1
, . . . , a
s
]. Prove that a
s
≥2.
10.2.9. Prove part (a) of Proposition 10.2.17. Also prove part (b) for s odd and part (c) for
s even.
10.2.10. In this exercise, given a cone σ in normal form with parameters d, k, you will
see how to construct an unbounded polyhedron P in M
R
≃ R
2
whose recession cone is
σ
∨
= Cone(e
1
, ke
1
+de
2
), and whose normal fan deﬁnes the resolution of U
σ
from The
orem 10.2.3. Let P
i
, Q
i
be the sequences constructed in Proposition 10.2.2. Let m be the
smallest positive integer such that md >P
1
+ +P
r−1
and mk >Q
1
+ +Q
r−1
. Starting
from the point A
r
=(mk, md) on the upper boundary ray of σ
∨
, construct the points
A
r−1
=(mk −Q
r−1
, md −P
r−1
)
A
r−2
=(mk −Q
r−1
−Q
r−2
, md −P
r−1
−P
r−2
)
.
.
.
A
1
=(mk −
r−1
i=1
Q
i
, md −
r−1
i=1
P
i
)
A
0
=(mk −
r−1
i=1
Q
i
, 0).
Then let P be the polyhedron with one edge along the positive xaxis starting at A
0
, edges
A
i
A
i+1
for i =0, . . . , r −1, and one edge along the upper edge of σ
∨
starting from A
r
.
(a) Drawthe polyhedron P for the cone with parameters (d, k) =(7, 5). What is the integer
m in this case?
(b) Show that σ
∨
is the recession cone of P.
(c) Show that the normal fan of P is the fan giving the resolution of U
σ
constructed in
Theorem 10.2.3.
10.2.11. Prove (10.2.11).
10.2.12. Let d/k be a rational number in lowest terms with 0 < k < d, and let
d/k =[[b
1
, . . . , b
r
]] =[a
1
, . . . , a
s
]
be its continued fraction expansions. You will show that the sequences P
i
, Q
i
and p
i
, q
i
considered in this section can be expressed in matrix form.
§10.3. Gr¨ obner Fans and M
c
Kay Correspondences 479
(a) Let M
−
(b) =
_
b −1
1 0
_
. Show that for all 1 ≤i ≤r,
_
P
i
−P
i−1
Q
i
−Q
i−1
_
=M
−
(b
1
)M
−
(b
2
) M
−
(b
i
).
(b) Let M
+
(a) =
_
a 1
1 0
_
. Show that for all 1 ≤i ≤s,
_
p
i
p
i−1
q
i
q
i−1
_
=M
+
(a
1
)M
+
(a
2
) M
+
(a
i
).
10.2.13. In this exercise, you will apply the results of this section to the weighted projective
plane P(q
0
, q
1
, q
2
) from §2.0 and Example 3.1.17.
(a) Construct a resolution of singularities for any P(1, 1, q
2
), where q
2
≥2. What smooth
toric surface is obtained in this way? A complete classiﬁcation of the smooth complete
toric surfaces will be developed in §10.4.
(b) Do the same for P(1, q
1
, q
2
) in general. Hint: Exercise 10.2.6.
§10.3. Gr¨ obner Fans and M
c
Kay Correspondences
The fans obtained by resolving the singularities of the afﬁne toric toric surfaces U
σ
have unexpected descriptions that involve Gr¨ obner bases and representation theory.
In this section we will present these ideas, following [95] and [96].
Gr¨ obner Bases and Gr¨ obner Fans. We assume the reader knows about Gr¨ obner
bases (see [36]) and Gr¨ obner fans (see [37, Ch. 8, §4] or [170]). A nonzero ideal
I ⊂ C[x
1
, . . . , x
n
] has a unique reduced Gr¨ obner basis with respect to each mono
mial order > on the polynomial ring. However, the set of distinct reduced marked
Gr¨ obner bases for I (i.e., reduced Gr¨ obner bases with marked leading terms in each
polynomial) is ﬁnite. This implies that the ideal I has a ﬁnite universal Gr¨ obner
basis, i.e., a ﬁnite subset U ⊂ I that is a Gr¨ obner basis for all monomial orders
simultaneously.
Let w∈ R
n
≥0
be a weight vector in the positive orthant (so w could be taken as
the ﬁrst row of a weight matrix deﬁning a monomial order). Let
G =¦g
1
, . . . , g
t
¦
be one of the reduced marked Gr¨ obner bases for I, where
g
i
=x
α(i)
+
β
c
iβ
x
β
,
and x
α(i)
is marked as the leading term of g
i
. If w α(i) > w β whenever c
iβ
,=0,
then I will have Gr¨ obner basis G with respect to any monomial order deﬁned by a
weight matrix with ﬁrst row w. The set
(10.3.1) C
G
=¦w ∈ R
n
≥0
[ w α(i) ≥w β whenever c
iβ
,=0¦
is the intersection of a ﬁnite collection of halfspaces, hence has the structure of a
closed convex polyhedral cone in R
n
≥0
. The cones C
G
as G runs over all distinct
480 Chapter 10. Toric Surfaces
marked Gr¨ obner bases of I, together with all of their faces, have the structure of a
fan in R
n
≥0
called the Gr¨ obner fan of I. In particular, for each pair G, G
′
of marked
Gr¨ obner bases, the cones C
G
and C
G
′ intersect along a common face where the
wweights of terms in some polynomials in G (and in G
′
) coincide.
A First Example. Let σ be a cone in normal form with parameters d, k, and recall
gcd(d, k) =1 by hypothesis. By Proposition 10.1.2, the group
G
d,k
=¦(ζ, ζ
k
) ∈ (C
∗
)
2
[ ζ
d
=1, 0 ≤k ≤d −1¦ ≃µ
d
acts on C
2
by componentwise multiplication
(10.3.2) (ζ, ζ
k
) (x, y) =(ζx, ζ
k
y),
with quotient C
2
/G
d,k
≃U
σ
.
Let I(G
d,k
) be the ideal deﬁning G
d,k
as a variety in C
2
. In the next extended
example, we will introduce the ﬁrst main result of this section.
Example 10.3.1. Let d =7, k =5 and I =I(G
7,5
). It is easy to check that
I =¸x
7
−1, y −x
5
).
Moreover, for lexicographic order with y > x, the set
G
(1)
=¦x
7
−1, y −x
5
¦
is the reduced marked Gr¨ obner basis for I, where the underlines indicate the leading
terms. The corresponding cone in the Gr¨ obner fan of I is
C
G
(1) =¦w =(a, b) ∈ R
2
≥0
[ b ≥5a¦ =Cone(e
2
, e
1
+5e
2
).
There are three other marked reduced Gr¨ obner bases of I:
G
(2)
=¦x
5
−y, x
2
y −1, y
2
−x
3
¦,
G
(3)
=¦x
3
−y
2
, x
2
y −1, y
3
−x¦,
G
(4)
=¦y
7
−1, x −y
3
¦.
It is easy to check that each of these sets is a Gr¨ obner basis for I using Buchberger’s
criterion. The corresponding cones are
C
G
(2) =¦(a, b) ∈ R
2
≥0
[ b ≤5a, 2b ≥3a¦ =Cone(e
1
+5e
2
, 2e
1
+3e
2
),
C
G
(3) =¦(a, b) ∈ R
2
≥0
[ 2b ≤3a, 3b ≥a¦ =Cone(2e
1
+3e
2
, 3e
1
+e
2
),
C
G
(4) =¦(a, b) ∈ R
2
≥0
[ 3b ≤a¦ =Cone(3e
1
+e
2
, e
1
).
Since these three cones ﬁll out rest of the ﬁrst quadrant in R
2
, the Gr¨ obner fan of
I consists of the four cones C
G
(i) and their faces, as shown in Figure 8 on the next
page. We will denote this fan by Γ in the following.
Next, let us consider the resolution of singularities of
X
Σ
−→U
σ
§10.3. Gr¨ obner Fans and M
c
Kay Correspondences 481
(3,1)
(2,3)
(1,5)
Figure 8. The Gr¨ obner fan Γ
for the cone σ with parameters d = 7, k = 5 computed in Example 10.2.4 in the
last section. The reader can check that the linear transformation T : N
R
→N
R
with
matrix relative to the basis e
1
, e
2
given by
(10.3.3) A =
_
7 0
−5 1
_
maps the cones C
G
(i) in the Gr¨ obner fan Γ to the corresponding cones σ
i
in the fan
Σ. The matrix in (10.3.3) is invertible, but its inverse is not an integer matrix. The
image of the lattice N = Z
2
under T is the proper sublattice 7Ze
1
⊕Ze
2
, and T
−1
maps N to the lattice
N
′
=¦(a/7, b/7) [ a, b ∈ Z, b ≡5a mod 7¦ =N +Z
_
1
7
e
1
+
5
7
e
2
_
.
There is an exact sequence
0 −→N −→N
′
−→G −→0
induced by the map N
′
→(C
∗
)
2
deﬁned by (a/7, b/7) →(e
2πia/7
, e
2πib/7
). Letting
τ =Cone(e
1
, e
2
), the corresponding toric morphism U
τ,N
−→U
τ,N
′ is the quotient
mapping C
2
−→C
2
/G.
It is easy to check that w
0
= (0, 1) and w
1
= (1/7, 5/7) form a basis of the
lattice N
′
. In Figure 9 on the next page, the fan deﬁned by the cones with ray
generators on the left is the same as in Figure 8 above, and the fan on the right is
the same as in Figure 3 above. Note that T(w
i
) =u
i
for 0 ≤i ≤4 in Figure 9.
With respect to N
′
, the cones in the Gr¨ obner fan Γ are smooth cones, and it
follows from the discussion of toric morphisms in §3.3 that the toric surfaces X
Γ,N
′
and X
Σ
are isomorphic, In other words, the Gr¨ obner fan Γ of the ideal I encodes
the structure of the resolution of singularities of U
σ
. ♦
482 Chapter 10. Toric Surfaces
(1,0) = w
4
= 7w
1
 5w
0
(3/7,1/7) = w
3
= 3w
1
 2w
0
(2/7,3/7) = w
2
= 2w
1
 w
0
(1/7,5/7) = w
1
(0,1) = w
0
u
0
u
1
u
2
u
3
u
4
σ
Figure 9. The toric variety X
Γ,N
′ is the resolution of Uσ
A Tale of Two Fans. We next show that the last observation in Example 10.3.1
holds in general. As in the example, consider the ideal I = I(G
d,k
) and the action
of G
d,k
given in (10.3.2). Each monomial in C[x, y] is equivalent modulo I to one
of the monomials x
j
, j = 0, . . . , d −1. This may be seen, for instance, from the
remainders on division by the lexicographic Gr¨ obner basis ¦x
d
−1, y −x
k
¦. As
a result, we have a direct sum decomposition of the coordinate ring of the variety
G
d,k
as a Cvector space:
(10.3.4) C[x, y]/I ≃
d−1
j=0
V
j
,
where V
j
is the 1dimensional subspace spanned by x
j
mod I.
The following result establishes a ﬁrst connection between the ideal I and the
resolution of singularities of U
σ
described in Theorem 10.2.3.
Proposition 10.3.2. Let I =I(G
d,k
) where 0 < k < d and gcd(d, k) =1. Consider
u
0
=e
2
u
i
=P
i−1
e
1
−Q
i−1
e
2
, i =1, . . . , r +1,
from Theorem 10.2.3. Let T : N
R
→N
R
be the linear transformation with matrix
A =
_
d 0
−k 1
_
and let w
i
=T
−1
(u
i
) for i =0, . . . , r +1. Write w
i
=
1
d
(a
i
e
1
+b
i
e
2
) and deﬁne
g
i
=x
b
i
−y
a
i
, 0 ≤i ≤r +1.
(a) The polynomials g
0
, . . . , g
r+1
are contained in the ideal I.
(b) S =¦ae
1
+be
2
∈ N [ a, b ≥0, x
b
−y
a
∈ I¦ ⊆N is an additive semigroup.
(c) ¦dw
i
[ i =0, . . . , r +1¦ is the Hilbert basis of the semigroup S of part (b).
§10.3. Gr¨ obner Fans and M
c
Kay Correspondences 483
Proof. It is an easy calculation to show T
−1
maps σ = Cone(e
2
, de
1
−ke
2
) to the
ﬁrst quadrant R
2
≥0
. Moreover, w
0
=e
2
, and for i =1, . . . , r +1,
w
i
=
1
d
_
P
i−1
e
1
+(kP
i−1
−dQ
i−1
)e
2
_
.
Therefore, g
0
=x
d
−1 and
g
i
=x
kP
i−1
−dQ
i−1
−y
P
i−1
for i =1, . . . , r +1. Since ζ
d
=1, these polynomials clearly vanish at (ζ, ζ
k
) ∈G
d,k
.
Therefore g
i
∈ I(G
d,k
) =I for all i.
The proof of part (b) is left to the reader as Exercise 10.3.3. For part (c), it
follows from parts (a) and (b) that dw
i
is contained in S. On the other hand, let
ae
1
+be
2
∈ S. Then x
b
−y
a
∈ I, which implies that b ≡ak mod d. Hence
T
_
1
d
(ae
1
+be
2
)
_
=ae
1
+
b−ak
d
e
2
must be an element of σ ∩N. Since the u
i
are the Hilbert basis for the semigroup
σ ∩N by Theorem 10.2.8, this vector is a nonnegative integer combination of the
u
i
. Hence ae
1
+be
2
is a nonnegative integer combination of the dw
i
. It follows that
S is generated by the dw
i
. The dw
i
are irreducible in S because the corresponding
u
i
=T(w
i
) are irreducible in the semigroup σ∩N.
We also have a ﬁrst result about reduced Gr¨ obner bases of the ideal I.
Lemma 10.3.3. Every element of a reduced Gr¨ obner basis of I =I(G
d,k
) is either
of the form x
b
−y
a
or of the form x
s
y
t
−1 for s, t > 0.
Proof. Since I is generated by x
d
−1, y −x
k
, the Buchberger algorithm implies
that a reduced Gr¨ obner basis G of I consists of binomials. By taking out common
factors, every g ∈ G can be written g =x
i
y
j
h, where h =x
b
−y
a
or x
s
y
t
−1. Then
h ∈ I since it vanishes on G
d,k
. Its leading term divisible by the leading term of an
element of G, which is impossible in a reduced Gr¨ obner basis unless g =h.
The Gr¨ obner bases in Example 10.3.1 give a nice illustration of Lemma 10.3.3.
Our next lemma relates the polynomials g
i
= x
b
i
−y
a
i
from Proposition 10.3.2 to
the reduced Gr¨ obner bases of I.
Lemma 10.3.4. Let g
i
= x
b
i
−y
a
i
be as in Proposition 10.3.2 and ﬁx a monomial
order > on C[x, y].
(a) The a
i
are increasing and the b
i
are decreasing with i.
(b) There is some index i =i
0
(depending on >) such that
LT
>
(g
i
) =x
b
i
for all i ≤i
0
, and LT
>
(g
i
) =y
a
i
for all i > i
0
.
(c) If i =i
0
is the index from part (b), then g
i
0
and g
i
0
+1
are elements of the reduced
Gr¨ obner basis of I =I(G
d,k
) with respect to >.
484 Chapter 10. Toric Surfaces
Proof. You will prove parts (a) and (b) in Exercise 10.3.4. For part (c), let G
be the reduced Gr¨ obner basis of I with respect to >. Since g
i
0
∈ I by part (a) of
Proposition 10.3.2, there is g ∈ G whose leading term divides LT
>
(g
i
0
) =x
b
i
0
. By
Lemma 10.3.3, it follows that g =x
b
−y
a
with LT
>
(g) = x
b
. In particular, b ≤b
i
0
and ae
1
+be
2
∈ S, where S is the semigroup from part (b) of Proposition 10.3.2.
Then part (c) of the same proposition implies that ae
1
+be
2
must be a nonnegative
integer combination
(10.3.5) ae
1
+be
2
=
r+1
i=0
ℓ
i
dw
i
=
r+1
i=0
ℓ
i
(a
i
e
1
+b
i
e
2
), ℓ
i
∈ N.
Since b ≤ b
i
0
and the b
i
decrease with i, (10.3.5) can include only the dw
i
with
i ≥i
0
. Suppose that dw
i
appears in (10.3.5) with i > i
0
. Then a ≥a
i
and y
a
i
> x
b
i
,
so that LT
>
(g
i
) = y
a
i
divides y
a
. Since g
i
∈ I, LT
>
(g
i
) is divisible by the leading
term of some h ∈ G. Hence LT
>
(h) divides y
a
, which is a term of g =x
b
−y
a
∈ G.
This is impossible in a reduced Gr¨ obner basis, hence i >i
0
cannot occur in (10.3.5).
From here, it follows easily that g =g
i
0
, giving g
i
0
∈ G as desired.
The statement for g
i
0
+1
follows by an argument parallel to the one above. The
details are left to the reader (Exercise 10.3.4).
We are now ready for the ﬁrst major result of this section.
Theorem 10.3.5. Let Γ be the fan in R
2
with maximal cones
γ
i
=Cone(a
i−1
e
1
+b
i−1
e
2
, a
i
e
1
+b
i
e
2
), 1 ≤i ≤r +1,
for a
i
, b
i
as in Proposition 10.3.2. Then Γ is the Gr¨ obner fan of the ideal I(G
d,k
).
Proof. The cones γ
i
ﬁll out the ﬁrst quadrant R
2
≥0
. Take any w = ae
1
+be
2
lying
in the interior of some γ
i
and let > be a monomial order deﬁned by a weight matrix
with w as ﬁrst row. This gives a reduced Gr¨ obner basis G. The theorem will follow
once we prove that γ
i
is the Gr¨ obner cone C
G
of G.
First observe that for >, we must have must have i
0
= i −1 in part (b) of
Lemma 10.3.4. It follows from part (c) of the lemma that g
i−1
and g
i
are elements
of G. Since a reduced Gr¨ obner basis only one element with leading term a power
of x and only one with leading term a power of y, all other elements of G will have
the form x
s
y
t
−1, s, t > 0, by Lemma 10.3.3. Therefore, the Gr¨ obner cone C
G
is
exactly γ
i
and we are done.
As in Example 10.3.1, the following statement is an immediate consequence.
Corollary 10.3.6. Let N
′
be the lattice
N
′
=¦(a/d, b/d) [ a, b ∈ Z, b ≡k mod d¦ =Z
_
1
d
e
1
+
k
d
e
2
_
⊕Ze
2
.
The toric surfaces X
Σ
and X
Γ,N
′ are isomorphic.
In other words, the Gr¨ obner fan of I(G
d,k
) can be used to construct a resolution
of singularities of the afﬁne toric surface U
σ
when σ has parameters d, k.
§10.3. Gr¨ obner Fans and M
c
Kay Correspondences 485
Connections with Representation Theory. We now consider the above results
from a different point of view. We assume the reader is familiar with the beginnings
of representation theory for ﬁnite abelian groups.
The group G =G
d,k
from (10.3.2) acts on V =C
2
by the 2dimensional linear
representation of the group µ
d
of dth roots of unity deﬁned by
(10.3.6)
ρ : µ
d
−→GL(V) =GL(2, C)
ζ −→
_
ζ 0
0 ζ
k
_
.
Since µ
d
is abelian, its irreducible representations are 1dimensional over C, and
hence each is deﬁned by a character
χ
j
: µ
d
−→C
∗
ζ −→ζ
−j
for j =0, . . . , d −1. The reason for the minus sign will soon become clear.
Via (10.3.2), we get the induced action of µ
d
on the polynomial ring C[x, y] by
ζ x =ζ
−1
x, ζ y =ζ
−j
y,
as explained in §5.0. Each monomial x
a
y
b
spans an invariant subspace where
the action of µ
d
is given by the irreducible representation with character χ
−j
for
j ≡ a +kb mod d. We call a +kb mod d the weight of the monomial x
a
y
b
with
respect to this action of µ
d
. Since the ideal of the group G⊂(C
∗
)
2
is invariant, the
action descends to the quotient C[x, y]/I(G), and we have a representation of µ
d
on C[x, y]/I(G). The direct sum decomposition (10.3.4) shows that the irreducible
representation with character χ
−j
appears exactly once in this representation, as
the subspace V
j
in (10.3.4). This means that the representation on C[x, y]/I(G) is
isomorphic to the regular representation of µ
d
(Exercise 10.3.5).
A 2dimensional M
c
Kay Correspondence. In 1979, M
c
Kay pointed out that there
is a onetoone correspondence between the irreducible representations of µ
d
and
the components of the exceptional divisor in the resolution φ : X
Σ
−→U
σ
when σ
was a cone in normal form with parameters k =d −1, as in Example 10.1.5. In this
case, the singular point of U
σ
is a rational double point, and the image of the rep
resentation ρ from (10.3.6) lies in SL(2, C). A great deal of research was devoted
to explaining the original M
c
Kay correspondence in representationtheoretic and
geometric terms (work of GonzalezSprinberg, Artin, and Verdier). However, for
1 < k < d −1, ρ(µ
d
) is a subgroup of GL(2, C), not SL(2, C), and there are more
irreducible representations of µ
d
than components of the exceptional divisor. The
M
c
Kay correspondence can be extended to these cases by identifying certain spe
cial representations that correspond to the components of the exceptional divisor
(work of Wunram, Esnault, Ito and Nakamura, Kidoh, and others).
486 Chapter 10. Toric Surfaces
We will describe a generalized M
c
Kay correspondence that applies for all d, k.
Writing G=G
d,k
as before, consider the ring of invariants C[x, y]
G
. You will prove
the following in Exercise 10.3.6.
Lemma 10.3.7. Let V
j
be an irreducible representations of µ
d
with character χ
−j
,
and consider the action of G = G
d,k
≃ µ
d
on C[x, y] ⊗
C
V
j
. Then the subspace of
invariants (C[x, y] ⊗
C
V
j
)
G
has the structure of a module over the ring C[x, y]
G
.
Example 10.3.8. Let G = G
7,5
as in Example 10.3.1. The ring of invariants is
C[x, y]
G
=C[x
7
, x
2
y, xy
4
, y
7
] in this case. If v
j
is the basis of the representation V
j
,
then it is easy to check that x
a
y
b
⊗v
j
is invariant under G if and only if a+kb− j ≡
0 mod 7, or in other words if and only if x
a
y
b
has weight j under this action of µ
7
.
First consider the case j =1 in Lemma 10.3.7. The monomials in the comple
ment of the monomial ideal
M=¸x
7
, x
2
y, xy
4
, y
7
)
that have weight 1 are x and y
3
. Then x ⊗v
1
and y
3
⊗v
1
generate the module
(C[x, y] ⊗V
1
)
G
. Since x and y
3
have the same weight with respect to this action
of µ
7
, the difference x −y
3
is an element of the ideal I(G), and this is one of the
polynomials g
i
as in the proof of Proposition 10.3.2.
On the other hand, if j = 2, then there are three monomials with weight 2 in
the complement of M, and these give three generators of (C[x, y] ⊗V
2
)
G
, namely
x
2
⊗v
2
, xy
3
⊗v
2
, and y
6
⊗v
2
. It is still true that x
2
−xy
3
, x
2
−y
6
, and xy
3
−y
6
are
elements of I(G), but these polynomials cannot appear in a reduced Gr¨ obner basis
for I(G). Moreover, no proper subset of the three generators generates the whole
module (C[x, y] ⊗V
2
)
G
. ♦
Deﬁnition 10.3.9. Let G = G
d,k
≃ µ
d
as above. We say that the representation V
j
is special with respect to k if (C[x, y] ⊗V
j
)
G
is minimally generated as a module
over the invariant ring C[x, y]
G
by two elements.
Hence, in Example 10.3.8, V
1
is special while V
2
is not. According to our
deﬁnition, the trivial representation V
0
is never special, since (C[x, y] ⊗V
0
)
G
is
generated by the single monomial 1 over the invariant ring. Our next theorem
gives a rudimentary form of a M
c
Kay correspondence for the group G
d,k
.
Theorem 10.3.10 (M
c
Kay Correspondence). Let σ be a cone with parameters
d, k, where 0 < k < d and gcd(d, k) = 1. Then there is a onetoone correspon
dence between the representations of µ
d
that are special with respect to k and the
components of the exceptional divisor for the minimal resolution φ : X
Σ
→U
σ
.
Proof. Write G = G
d,k
as above and consider the set B of monomials in the com
plement of the ideal Mgenerated by the Ginvariant monomials. This set contains
L =¦1, x, x
2
, . . . , x
d−1
, y, y
2
, . . . , y
d−1
¦.
§10.3. Gr¨ obner Fans and M
c
Kay Correspondences 487
Since gcd(d, k) =1, for each 1 ≤ j ≤d −1, there is an integer 1 ≤a
j
≤d −1 such
that x
j
and y
a
j
have equal weight (equal to j) for the action of G. The representa
tion V
j
is special with respect to k if and only if these are the only two monomials
of weight j in the set B, and nonspecial if and only if there is some monomial
x
a
y
b
with a, b > 0 in B which also has weight j. Since x
j
−y
a
j
∈ I(G), saying
V
j
is special with respect to k is in turn equivalent to saying that the corresponding
vector a
j
e
1
+ je
2
is an irreducible element in the semigroup from part (b) of Propo
sition 10.3.2 (Exercise 10.3.8). By Theorem 10.3.5, Cone(a
j
e
1
+ je
2
) is one of the
1dimensional cones of the Gr¨ obner fan of I(G). Then V
j
corresponds to one of the
1dimensional cones in the fan Σ and hence to one of the irreducible components
of the exceptional divisor.
The original M
c
Kay correspondence is the following special case.
Corollary 10.3.11. When k = d −1, there is a onetoone correspondence be
tween the set of all irreducible representations of µ
d
and the components of the
exceptional divisor of the minimal resolution φ : X
Σ
−→U
σ
.
Proof. In this case, the invariant ring is C[x
d
, xy, y
d
], so the sets L and B in the
proof of the theorem coincide.
There has also been much work devoted to extend the M
c
Kay correspondence
to ﬁnite abelian subgroups G ⊂ GL(n, C) for n ≥ 3, and several other ways to
understand these constructions have also been developed, including the theory of
GHilbert schemes. See Exercise 10.3.10 for the beginnings of this.
Exercises for §10.3.
10.3.1. In this exercise, you will verify the claims made in Example 10.3.1, and extend
some of the observations there.
(a) Show that each of the G
(i)
is a Gr¨ obner basis of I(G
7,5
).
(b) Show that G =¦x
5
−1, y −x
3
, x
2
y −1, x −y
2
, y
5
−1¦ is a universal Gr¨ obner basis for
I(G
7,5
).
(c) Determine the cones C
G
(i) using (10.3.1).
(d) Verify the ﬁnal claim that linear transformation deﬁned by the matrix A from (10.3.3)
maps the Gr¨ obner cones C
G
(i) to the σ
i
for i =1, 2, 3.
10.3.2. Verify the conclusions of Proposition 10.3.2 and Theorem 10.3.5 for the case d =
17, k =11.
10.3.3. Show that the set S deﬁned in part (b) of Proposition 10.3.2 is an additive semi
group. Hint: A direct proof starts from two general elements ae
1
+be
2
and a
′
e
1
+b
′
e
2
in
S. Consider (x
b
−y
a
)(x
b
′
+y
a
′
) and (x
b
+y
a
)(x
b
′
−y
a
′
).
10.3.4. In this exercise you will complete the proof of Lemma 10.3.4.
(a) Show that for each i, kP
i
−dQ
i
= k
i
, where the k
i
are produced by the modiﬁed Eu
clidean algorithm from (10.2.1).
488 Chapter 10. Toric Surfaces
(b) Prove part (a) of Lemma 10.3.4.
(c) Verify that there is an index i
0
as in part (b) of Lemma 10.3.4.
(d) Verify that if i
0
is as in part (c), then g
i0+1
is contained in the reduced Gr¨ obner basis.
10.3.5. If G is any ﬁnite group, the (left) regular representation of G is deﬁned as follows.
Let W be a vector space over C of dimension [G[ with a basis ¦e
h
[ h ∈ G¦ indexed by the
elements of G. For each g ∈ G let ρ(g) : W −→W be deﬁned by ρ(g)(e
h
) =e
gh
.
(a) Show that g →ρ(g) is a group homomorphismfrom G to GL(W).
(b) Now let G be the cyclic group µ
d
of order d. Show that W is the direct sum of 1
dimensional invariant subspaces W
j
, j =0, . . . , d −1 on which G acts by the character
χ
j
deﬁned in the text, so that W decomposes as W ≃⊕
d−1
j=0
V
j
.
10.3.6. In this exercise you will consider the module structures from Lemma 10.3.7.
(a) Prove Lemma 10.3.7.
(b) Verify the claims made in Example 10.3.8.
10.3.7. In this exercise you will prove an alternate characterization of the special represen
tations with respect to k from Deﬁnition 10.3.9. We write G =G
d,k
≃µ
d
as usual.
(a) Let Ω
2
C
2
=¦ f dx ∧dy [ f ∈ C[x, y]¦. Show that ζ x
a
y
b
dx ∧dy =ζ
a+b+1+k
x
a
y
b
dx ∧dy
deﬁnes an action of G on Ω
2
C
2
.
(b) Show that the spaces of Ginvariants (Ω
2
C
2
)
G
and (Ω
2
C
2
⊗V
j
)
G
have the structure of
modules over the invariant ring C[x, y]
G
.
(c) Show that V
j
is special with respect to k if and only if the “multiplication map”
(Ω
2
C
2 )
G
⊗(C[x, y] ⊗V
j
)
G
−→(Ω
2
C
2 ⊗V
j
)
G
is surjective.
10.3.8. In the proof of the M
c
Kay correspondence, show that a
j
e
1
+ je
2
is irreducible in
the semigroup S from Proposition 10.3.2 if and only if the representation V
j
is special with
respect to k.
10.3.9. Let g
i
, 0 ≤i ≤r +1, be the binomials constructed in Proposition 10.3.2. Show that
U =¦g
1
, . . . , g
r
¦ ∪¦x
a
y
b
−1 [ x
a
y
b
is G
d,k
invariant¦
is a universal Gr¨ obner basis for I(G
d,k
) (not always minimal, however).
10.3.10. Let G=G
d,k
act on C
2
as in (10.3.2). As a point set, G can be viewed as the orbit
of the point (1, 1) under this action. The ideal I = I(G) is invariant under the action of G
on C[x, y] and as we have seen, the corresponding representation on C[x, y]/I is isomorphic
to the regular representation of G.
(a) Showthat if p =(ξ, η) is any point in C
2
other than the origin, the ideal I of the orbit of
p is another Ginvariant ideal and the corresponding representation of G on C[x, y]/I
is also isomorphic to the regular representation of G.
(b) The GHilbert scheme can be deﬁned as the set of all Ginvariant ideals in C[x, y] such
that the representation of G on C[x, y]/I is isomorphic to the regular representation of
G. Show that every such ideal has a set of generators of the form
¦x
a
−αy
c
, y
b
−βx
d
, x
a−d
y
b−c
−αβ¦
§10.4. Smooth Toric Surfaces 489
for some α, β ∈ C and where x
a
and y
c
(resp. y
b
and x
d
) have equal weights for the
action of G. It can be seen fromthis result that the GHilbert scheme is also isomorphic
to the minimal resolution of singularities of U
σ
.
§10.4. Smooth Toric Surfaces
This section will use §10.1 and §10.2 to classify smooth complete toric surfaces
and study the relation between continued fractions and intersection products of
divisors on the resulting resolutions of singularities.
Classiﬁcation of Smooth Toric Surfaces. We will show that smooth complete
toric surfaces are all obtained by toric blowups from either P
2
, P
1
P
1
, or one
of the Hirzebruch surfaces H
r
with r ≥2 from Example 3.1.16. The proof will be
based on the following facts.
First, Proposition 3.3.15 implies that if σ =Cone(u
1
, u
2
) is a smooth cone and
we reﬁne σ by inserting the new 1dimensional cone τ = Cone(u
1
+u
2
), then on
the resulting toric surface, the smooth point p
σ
is blown up to a copy of P
1
.
For the second ingredient of the proof, we introduce the following notation. If
Σ is a smooth complete fan, then list the ray generators of the 2dimensional cones
in Σ as u
0
, u
1
, . . . , u
r−1
in clockwise order around the origin in N
R
, and we will
consider the indices as integers modulo r, so u
r
= u
0
. Then we have the following
statement parallel to (10.2.8).
Lemma 10.4.1. Let u
0
, . . . , u
r
be the ray generators for a smooth complete fan Σ
in N
R
≃R
2
. For all i, there exist integers b
i
, i =0, . . . , r −1, such that
(10.4.1) u
i−1
+u
i+1
=b
i
u
i
.
Proof. This is a special case of the wall relation (6.3.5).
We also have the following result.
Lemma 10.4.2. Let Σ be a smooth fan that reﬁnes a smooth cone σ. Then Σ is
obtained from σ by a sequence of star subdivisions as in Deﬁnition 3.3.13.
Proof. Suppose Σ has r 2dimensional cones with ray generators u
0
, . . . , u
r
, listed
clockwise starting from u
0
. We argue by induction on r. If r =1, there is only one
cone in Σ and there is nothing to prove. Assume the result has been proved for
all Σ with r 2dimensional cones, and consider the case of r +1 cones. There are
r “interior” rays that by (6.3.5) give wall relations u
i−1
+u
i+1
= b
i
u
i
, b
i
∈ Z, for
1 ≤ i ≤ r. Note that σ is strongly convex so b
i
> 0 for all i. We claim that there
exists some i such that b
i
= 1. If not, i.e., if b
i
≥ 2 for all i, then as in §10.2, the
HirzebruchJung continued fraction
[[b
1
, b
2
, . . . , b
r
]]
490 Chapter 10. Toric Surfaces
represents a rational number d/k and the cone σ has parameters d, k with d >k >0.
But then d ≥2, which would contradict the assumption that σ is a smooth cone.
Hence there exists an i, 1 ≤i ≤r, such that
u
i−1
+u
i+1
=u
i
.
In this situation, Cone(u
i−1
, u
i+1
) is also smooth (Exercise 10.4.1). Moreover,
Cone(u
i−1
, u
i
) and Cone(u
i
, u
i+1
) are precisely the cones in the star subdivision
of Cone(u
i−1
, u
i+1
). Then we are done by induction.
We are now ready to state our classiﬁcation theorem.
Theorem 10.4.3. Every smooth complete toric surface X
Σ
is obtained from either
P
2
, P
1
P
1
, or H
r
, r ≥2
by a ﬁnite sequence of blowups at ﬁxed points of the torus action.
Proof. We follow the notation of Lemma 10.4.1. As in the proof of Lemma 10.4.2,
if b
i
= 1 in (10.4.1) for some i, then our surface is a blowup of the smooth surface
corresponding to the fan where u
i
is removed. Hence we only need to consider the
case in (10.4.1) where b
i
,=1 for all i.
Suppose ﬁrst that u
j
= −u
i
for some i < j. We relabel the vertices to make
u
j
= −u
1
for some j. Note that j > 2 since the cones must be strongly convex.
Then from (10.4.1),
u
0
=−u
2
+b
1
u
1
.
Using the basis u
1
, u
2
of N, we get the picture shown in Figure 10. Comparing this
u
0
u
1
u
2
u
j
Figure 10. The ray generators u0, u1, u2, uj when uj =−u1
with the fans for P
2
, P
1
P
1
and H
a
(Figures 2, 3 and 4 from §3.1), we see that
Σ is a reﬁnement of the fan of H
r
if r =b
1
> 2. The same follows if b
1
<−2 and
r =[b
1
[ (see Exercise 10.4.2). Since b
1
,=1, the remaining possibilities are b
1
=0
or −1, where we get a reﬁnement of the fan of P
1
P
1
or P
2
respectively. Then
the theorem follows from Lemma 10.4.2 in this case.
§10.4. Smooth Toric Surfaces 491
We will complete the proof using primitive collections (Deﬁnition 5.1.5). The
ﬁrst step is to show that X
Σ
is projective (Exercise 10.4.3). This will allow us to use
Proposition 7.3.6, which asserts that Σ has a primitive collection whose minimal
generators sum to 0.
If Σ(1) has only three elements, it is easy to see that X
Σ
= P
2
since Σ is
smooth and complete (Exercise 10.4.4). If [Σ(1)[ > 3, every primitive collection
of Σhas exactly two elements (Exercise 10.4.4). By Proposition 7.3.6, one of these
primitive collections must have minimal generators u
i
, u
j
that satisfy u
i
+u
j
= 0.
Hence u
j
=−u
i
, and we are done by the earlier part of the proof.
Since there is also a Hirzebruch surface H
1
, the statement of this theorem
might seem puzzling. The reason that H
1
is not included is that this surface is
actually a blowup of P
2
(Exercise 10.4.5).
The problem of classifying smooth compete toric varieties of higher dimension
is much more difﬁcult. We did this when rankPic(X
Σ
) = 2 in Theorem 7.3.7. See
[7] for the case when rankPic(X
Σ
) =3.
Intersection Products on Smooth Surfaces. A fundamental feature of the theory
of smooth surfaces is the intersection product on divisors. In §6.2, we deﬁned DC
when D is a Cartier divisor and C is complete irreducible curve. On a smooth
complete surface, this means that the intersection product D C is deﬁned for all
divisors D and C. In particular, taking D=C gives the selfintersection D D=D
2
.
Here is a useful result about intersection numbers on a smooth toric surface.
Theorem 10.4.4. Let D
ρ
be the divisor on a smooth toric surface X
Σ
corresponding
to ρ = Cone(u) which is the intersection of 2dimensional cones Cone(u, u
1
) and
Cone(u, u
2
) in Σ. Then:
(a) D
ρ
D
ρ
=−b, where u
1
+u
2
=bu as in (10.4.1).
(b) For a divisor D
ρ
′ ,=D
ρ
, we have
D
ρ
′ D
ρ
=
_
1 ρ
′
=Cone(u
i
), i =1, 2
0 otherwise.
Proof. Since X
Σ
is smoooth, part (a) follows from Lemma 6.3.4 once you compare
(10.4.1) to (6.3.5). Part (b) follows from Corollary 6.3.3 and Lemma 6.3.4.
Example 10.4.5. Let σ =Cone(u
1
, u
2
) be a cone in a smooth fan Σ, and consider
the star subdivision, in which ρ =Cone(u
1
+u
2
) is inserted to subdivide σ into two
cones. Call the reﬁned fan Σ
′
. Then the exceptional divisor E =D
ρ
of the blowup
φ : X
Σ
′ →X
Σ
satisﬁes E E =−1 on X
Σ
′ . ♦
492 Chapter 10. Toric Surfaces
Complete curves with selfintersection number −1 on a smooth surface are
called exceptional curves of the ﬁrst kind. They can always be contracted to a
smooth point on a birationally equivalent surface, as in the above example.
One of the foundational results in the theory of general algebraic surfaces is
that every smooth complete surface S has at least one relatively minimal model.
This means that there is a birational morphism S →S, where S is a smooth surface
with the property that if φ : S →S
′
is a birational morphism to another smooth
surface S
′
, then φ is necessarily an isomorphism. This is proved in [78, V.5.8].
Interestingly, the possible relatively minimal models for rational surfaces are pre
cisely the surfaces P
2
, P
1
P
1
, and H
r
, r ≥2, from Theorem 10.4.3.
On a smooth complete toric surface X
Σ
, the intersection product can be re
garded as a Zvalued symmetric bilinear form on Pic(X
Σ
). Here is an example.
Example 10.4.6. Consider the Hirzebruch surface H
r
. Using the fan shown in
Figure 3 of Example 4.1.8, we get divisors D
1
, . . . , D
4
corresponding to minimal
generators u
1
= −e
1
+re
2
, u
2
= e
2
, u
3
= e
1
, u
4
= −e
2
. By Theorem 10.4.4, we
have the selfintersections
D
1
D
1
=D
3
D
3
=0, D
2
D
2
=−r, D
4
D
4
=r.
The Picard group Pic(H
r
) is generated by the classes of D
3
and D
4
. Note also that
D
3
D
4
=D
4
D
3
=1
by Theorem 10.4.4. The intersection product is described by the matrix
_
D
3
D
3
D
3
D
4
D
4
D
3
D
4
D
4
_
=
_
0 1
1 r
_
.
If D ∼aD
3
+bD
4
and E ∼cD
3
+dD
4
are any two divisors on the surface, then
(10.4.2) D E =
_
a b
_
_
0 1
1 r
__
c
d
_
=bc +ad +rbd.
For instance, with D =E =D
2
∼−rD
3
+D
4
, we obtain
D
2
D
2
=1 (−r) +(−r) 1+r 1 1 =−r.
The selfintersection numbers D
1
D
1
= D
3
D
3
= 0 reﬂect the ﬁbration structure
on H
r
studied in Example 3.3.20. The divisors D
1
and D
3
are ﬁbers of the map
ping H
r
→P
1
. Such curves always have selfintersection equal to zero. You will
compute several other intersection products on H
r
in Exercise 10.4.6. ♦
Resolution of Singularities Reconsidered. Another interesting class of smooth
toric surfaces are those that arise from a resolution of singularities of the afﬁne
toric surface U
σ
of a 2dimensional cone σ. Here is a simple example.
Example 10.4.7. Let σ = Cone(e
2
, de
1
−e
2
) have parameters d, 1 where d > 1.
The resolution of singularities X
Σ
→U
σ
constructed in Example 10.1.8 uses the
§10.4. Smooth Toric Surfaces 493
smooth reﬁnement of σ obtained by adding Cone(e
1
). This gives the exceptional
divisor E on X
Σ
. Since
e
2
+(de
1
−e
2
) =de
1
,
we see that
E E =−d
is the selfintersection number of E. ♦
More generally, suppose that the smooth toric surface X
Σ
is obtained via a
resolution of singularities of U
σ
, where the 2dimensional cone σ has parameters
d, k with d > 1. Let the HirzebruchJung continued fraction expansion of d/k be
d/k =[[b
1
, b
2
, . . . , b
r
]].
Recall from Theorem 10.2.3 that Σ is obtained from σ =Cone(u
0
, u
r+1
) by adding
rays generated by u
1
, . . . , u
r
, and by Theorem 10.2.5, we have
u
i−1
+u
i+1
=b
i
u
i
, 1 ≤i ≤r.
It follows that D
1
, . . . , D
r
are complete curves in X
Σ
. They are the irreducible
components of the exceptional ﬁber, with selfintersections
D
i
D
i
=−b
i
, 1 ≤i ≤r,
by Theorem 10.4.4. Then the intersection matrix (D
i
D
j
)
1≤i, j≤r
is given by
(10.4.3) D
i
D
j
=
_
¸
_
¸
_
−b
i
if j =i
1 if [i − j[ =1
0 otherwise.
In Exercise 10.4.7 you will show that the associated quadratic form is negative
deﬁnite. This condition is necessary for the contractibility of a complete curve C
on a smooth surface S, i.e., the existence of a proper birational morphism π : S →S,
where π(C) is a (possibly singular) point on S.
The resolutions described here have another important property.
Deﬁnition 10.4.8. A resolution of singularities φ : Y →X is said to be minimal if
for every resolution of singularities ψ : Z →X
Y
φ
Z
ρ
??
ψ
//
X
there exists a morphism ρ : Z →Y such that ψ =φ◦ρ.
It is easy to see that if a minimal resolution of X exists, then it is unique up to
isomorphism. If X has a unique singular point p, then using the theory of birational
morphisms of surfaces it is not difﬁcult to show that a resolution of singularities
494 Chapter 10. Toric Surfaces
φ : Y →X is minimal if the exceptional ﬁber contains no irreducible components E
with E E =−1 (see Exercise 10.4.8). By Theorem 10.4.4 and the fact that b
i
≥2
in HirzebruchJung continued fractions, this holds for the resolutions constructed
in Theorem 10.2.3. Hence we have the following.
Corollary 10.4.9. The resolution of singularities of the afﬁne toric surface U
σ
constructed in Theorem 10.2.3 is minimal.
Rational Double Points Reconsidered. If σ has parameters d, d −1, then from
Exercise 10.2.2, the HirzebruchJung continued fraction expansion of d/(d −1) is
given by
d/(d −1) =[[2, 2, . . . , 2]],
with d −1 terms. Hence b
i
= 2 for all i, and (10.4.3) gives the (d −1) (d −1)
matrix
_
_
_
_
_
_
_
−2 1 0 0
1 −2 1 0
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
0 1 −2 1
0 0 1 −2
_
_
_
_
_
_
_
representing the intersection product on the subgroup of Pic(X
Σ
) generated by the
components of the exceptional divisor for the resolution of a rational double point
of type A
d−1
. We can now fully explain the terminology for these singularities.
The problem of classifying lattices
Ze
1
⊕ ⊕Ze
s
with negative deﬁnite bilinear forms B satisfying B(e
i
, e
i
) = −2 for all i arises
in many areas within mathematics, most notably in the classiﬁcation of complex
simple Lie algebras via root systems. The matrix above is the (negative of) the
Cartan matrix for the root system of type A
d−1
, which is often represented by the
Dynkin diagram:
t t t t t q q q
with d −1 vertices. The vertices represent the lattice basis vectors. The edges
connect the pairs with B(e
i
, e
j
) ,= 0 and B(e
i
, e
i
) = −2 for all i as above. In our
case, the vertices represent the components D
i
of the exceptional divisor, and the
bilinear form is the intersection product.
A precise deﬁnition of a surface rational double point follows.
Deﬁnition 10.4.10. A singular point p of a normal surface X is a rational double
point or Du Val singularity if X has a minimal resolution of singularities φ : Y →X
such that if K
Y
is a canonical divisor on Y, then every irreducible component E
i
of
the exceptional divisor E over p satisﬁes
K
Y
E
i
=0.
§10.4. Smooth Toric Surfaces 495
We can relate these concepts to the toric case as follows.
Proposition 10.4.11. Assume σ has parameters d > k > 0 and let φ : X
Σ
→U
σ
be the resolution of singularities constructed in Theorem 10.2.3. Then the singular
point of U
σ
is a rational double point if and only if k =d −1.
Proof. The canonical divisor of X
Σ
is K
X
Σ
=−
r+1
i=0
D
i
, and one computes that
K
X
Σ
D
i
=b
i
−2, 1 ≤i ≤r.
Thus the singular point is a rational double point if and only if b
i
=2 for all i. This
easily implies k =d −1. You will verify these claims in Exercise 10.4.9.
There is much more to say about rational double points. For example, one can
show that E = E
1
+ +E
r
satisﬁes E E = −2 (you will prove this in the toric
case in Exercise 10.4.9). From a more sophisticated point of view, E E = −2
implies that the canonical sheaf on Y is the pullback of the canonical sheaf on X
under φ, and this is one way to deﬁne rational singularities more generally. See
[47] for more on rational double points.
Exercises for §10.4.
10.4.1. Here you will verify several statements made in the proof of Lemma 10.4.2.
(a) Show that if the cone σ is strictly convex, then the integers b
i
in (10.4.1) must be
strictly positive.
(b) Show that if u
i−1
+u
i+1
=u
i
, then Cone(u
i−1
, u
i+1
) must also be smooth.
10.4.2. In the proof of Theorem 10.4.3, verify that if u
j
= −u
1
and u
0
= −u
2
+b
1
u
1
with
b
1
<−2, then Σ is a reﬁnement of a fan Σ
′
with X
Σ
′ ≃H
r
, where r =[b
1
[.
10.4.3. Let Σ be a complete fan in N
R
≃ R
2
. Prove that X
Σ
is projective. Hint: Create a
(probably nonrational) polygon whose vertices lie on the rays of Σ. This gives the support
function of an Rdivisor that lies in the interior of Nef (X
Σ
).
10.4.4. In this exercise, you will prove some facts used in the proof of Theorem 10.4.3.
Let Σ be a smooth complete fan in N
R
≃R
2
.
(a) If [Σ(1)[ =3, prove that X
Σ
≃P
2
.
(b) If [Σ(1)[ > 3, prove that every primitive collection of Σ(1) has two elements.
10.4.5. In the statement of Theorem 10.4.3, you might have noticed the absence of the
Hirzebruch surface H
1
. Show that this surface is isomorphic to the blowup of P
2
at one of
its torusﬁxed points. See Exercise 3.3.6 for more details.
10.4.6. This exercise studies several further examples of the intersection product on H
r
.
(a) Compute D
1
D
1
using (10.4.2) and also directly from Theorem 10.4.4.
(b) Compute K
2
=K K on H
r
, where K =K
Hr
is the canonical divisor.
10.4.7. Show that the matrix deﬁned by (10.4.3) has a negativedeﬁnite associated qua
dratic form. Hint: Recall that if B(x, y) is a bilinear form, the associated quadratic form is
Q(x) =B(x, x).
496 Chapter 10. Toric Surfaces
10.4.8. Let X have a unique singular point p and let φ : Y →X be a resolution of singu
larities such that no component E of the exceptional ﬁber φ
−1
(p) has E E = −1. In this
exercise, you will show that Y is a minimal resolution of X according to Deﬁnition 10.4.8.
Let ψ : Z →X be another resolution of singularities and consider the possibly singular
surface S = Z
X
Y. Let R be a resolution of S. Then we have a commutative diagram of
morphisms
R
α
//
β
Y
ϕ
Z
ψ
//
X
(a) Explain why it sufﬁces to show that β must be an isomorphism.
(b) If not, apply [78, V.5.3] to show that β factors as a sequence of blowups of points.
Hence R must contain curves L with L L =−1 in the exceptional ﬁber over p.
(c) Let L be an irreducible curve on R with L L = −1 and show that E = α(L) satisﬁes
E E =−1.
(d) Deduce that β is an isomorphism, hence φ : Y →X is a minimal resolution.
10.4.9. This exercise deals with the proof of Proposition 10.4.11.
(a) Show that K
XΣ
D
i
=b
i
−2 for 1 ≤i ≤r.
(b) Show that d/k =[[2, . . . , 2]] if and only if k =d −1. Hint: Exercise 10.2.2.
(c) Show that E =D
1
+ +D
r
satisﬁes E E =−2.
10.4.10. Let σ have parmeters d, d −1, so that the singular point of U
σ
is a rational double
point. By Proposition 10.1.6, U
σ
is Gorenstein, so that its canonical sheaf ω
Uσ
is a line
bundle. Let φ : X
Σ
→U
σ
be the resolution constructed in Theorem 10.2.3. Prove that
φ
∗
ω
Uσ
is the canonical sheaf of X
Σ
.
10.4.11. Another interesting numerical fact about the integers b
i
from (10.4.1) is the fol
lowing. Suppose a smooth fan Σ has 1dimensional cones labeled as in Lemma 10.4.1.
Then
(10.4.4) b
0
+b
1
+ +b
r−1
=3r −12.
This exercise will sketch a proof of (10.4.4).
(a) Show that (10.4.4) holds for the standard fans of P
2
, P
1
P
1
and H
r
, r ≥2.
(b) Show that if (10.4.4) holds for a smooth fan Σ, then it holds for the fan obtained by
performing a star subdivision on one of the 2dimensional cones of Σ.
(c) Deduce that (10.4.4) holds for all smooth fans using Theorem 10.4.3.
§10.5. RiemannRoch and Lattice Polygons
RiemannRoch theorems are a class of results about the dimensions of sheaf co
homology groups. The original statement along these lines was the theorem of
Riemann and Roch concerning sections of line bundles on algebraic curves. This
result and its generalizations to higherdimensional varieties can be formulated
most conveniently in terms of the Euler characteristic of a sheaf, deﬁned in §9.4.
§10.5. RiemannRoch and Lattice Polygons 497
RiemannRoch for Curves. A modern form of the RiemannRoch theorem for
curves states that if D is a divisor on a smooth projective curve C, then
(10.5.1) χ(O
C
(D)) =deg(D) +χ(O
C
),
where the degree deg(D) is deﬁned in Deﬁnition 6.2.2. This equality can be rewrit
ten using Serre duality as follows. Namely, if K
C
is a canonical divisor on C, then
we have
H
1
(C, O
C
(D)) ≃H
0
(C, O
C
(K
C
−D))
∨
H
1
(C, O
C
) ≃H
0
(C, O
C
(K
C
))
∨
.
The integer g =dimH
0
(C, O
C
(K
C
)) is the genus of the curve C. Then (10.5.1) can
be rewritten in the form commonly used in the theory of curves:
(10.5.2) dimH
0
(C, O
C
(D)) −dimH
0
(C, O
C
(K
C
−D)) =deg(D) +1−g.
A proof of this theorem and a number of its applications are given in [78, Ch. IV].
Also see Exercise 10.5.1 below. As a ﬁrst consequence, note that if D = K
C
is a
canonical divisor, then
(10.5.3) deg(K
C
) =2g−2.
We will need to use (10.5.2) most often in the simple case X ≃P
1
. Then g =0
and the RiemannRoch theorem for P
1
is the statement for all divisors D on P
1
,
(10.5.4) χ(O
P
1 (D)) =deg(D) +1.
The Adjunction Formula. For a smooth curve C contained in a smooth surface X,
the canonical sheaves ω
C
of the curve and ω
X
of the surface are related by
(10.5.5) ω
C
≃ω
X
(C) ⊗
O
X
O
C
.
This follows without difﬁculty from Example 8.2.2 (Exercise 10.5.2) and has the
following consequence for the intersection product on X.
Theorem 10.5.1 (Adjunction Formula). Let C be a smooth curve contained in a
smooth complete surface X. Then
K
X
C+C C =2g−2,
where g is the genus of the curve C.
Proof. Let i : C ֒→X be the inclusion map. Then
ω
C
≃ω
X
(C) ⊗
O
X
O
C
=i
∗
ω
X
(C) =i
∗
O
X
(K
X
+C),
so that
2g−2 =deg(ω
C
) =deg(i
∗
O
X
(K
X
+C)) =(K
X
+C) C,
where the ﬁrst equality is (10.5.3) and the last is the deﬁnition of (K
X
+C)C given
in §6.2.
498 Chapter 10. Toric Surfaces
RiemannRoch for Surfaces. The statement for surfaces corresponding to (10.5.1)
is given next.
Theorem 10.5.2 (RiemannRoch for Surfaces). Let D be a divisor on a smooth
projective surface X with canonical divisor K
X
. Then
χ(O
X
(D)) =
D D−D K
X
2
+χ(O
X
).
We will only prove this for X a smooth complete toric surface; there is a simple
and concrete proof in this case.
Proof. The theorem certainly holds for D=0 since O
X
(D) =O
X
in this case. Our
proof will use the special properties of smooth complete toric surfaces. Recall that
if X = X
Σ
, then Pic(X) is generated by the classes of the divisors D
i
, i = 1, . . . , r,
corresponding to the 1dimensional cones in Σ. Hence, to prove the theorem, it
sufﬁces to show that if the theorem holds for a divisor D, then it also holds for
D+D
i
and D−D
i
for all i.
Assume the theorem holds for D. By Proposition 4.0.28, the sequence
0 −→O
X
(−D
i
) −→O
X
−→O
D
i
−→0
is exact. Tensoring this with O
X
(D+D
i
) gives the exact sequence
0 −→O
X
(D) −→O
X
(D+D
i
) −→O
D
i
(D+D
i
) −→0.
By (9.4.1), it follows that
χ(O
X
(D+D
i
)) =χ(O
X
(D)) +χ(O
D
i
(D+D
i
)).
By the induction hypothesis,
(10.5.6) χ(O
X
(D)) =
D D−D K
X
2
+χ(O
X
).
For χ(O
D
i
(D+D
i
)), recall that D
i
≃ P
1
. Hence, by the RiemannRoch theorem
for P
1
given in (10.5.4), we have
(10.5.7) χ(O
D
i
(D+D
i
)) =D D
i
+D
i
D
i
+1.
Combining (10.5.6) and (10.5.7), we obtain
χ(O
X
(D+D
i
)) =
D D−D K
X
2
+D D
i
+D
i
D
i
+1+χ(O
X
)
=
(D+D
i
) (D+D
i
) +D
i
D
i
−D K
X
+2
2
+χ(O
X
)
However, using K
X
=−(D
1
+ +D
r
) and Theorem 10.4.4, one computes that
(10.5.8) D
i
K
X
=−D
i
D
i
−2.
§10.5. RiemannRoch and Lattice Polygons 499
Substituting this into the above expression for χ(O
X
(D+D
i
)) and simplifying, we
obtain
χ(O
X
(D+D
i
)) =
(D+D
i
) (D+D
i
) −(D+D
i
) K
X
2
+χ(O
X
),
which shows that the theorem holds for D+D
i
The proof for D−D
i
is similar and is left to the reader (Exercise 10.5.3).
The following statement is sometimes considered as the topological part of the
RiemannRoch theorem for surfaces.
Theorem 10.5.3 (Noether’s Theorem). Let X be a smooth projective surface with
canonical divisor K
X
. Then
χ(O
X
) =
K
X
K
X
+e(X)
12
,
where e(X) is the topological Euler characteristic of X deﬁned by
e(X) =
4
i=0
(−1)
i
dimH
i
(X, C).
As before, we will give a proof only for a smooth complete toric surface. We
will also use the Hodge decomposition
H
p
(X, C) ≃
i+j=p
H
i
(X, Ω
j
X
)
mentioned at the end of §9.4.
Proof. Demazure vanishing (Theorem 9.2.3) implies that for a smooth complete
toric surface X =X
Σ
, we have
(10.5.9)
χ(O
X
) =dimH
0
(X, O
X
) −dimH
1
(X, O
X
) +dimH
2
(X, O
X
)
=1−0+0 =1.
Thus Noether’s theorem for a smooth complete toric surface is equivalent to
(10.5.10) K
X
K
X
+e(X) =12.
We prove this as follows. Set r =[Σ(1)[ and let the minimal generators of the rays
be u
0
, . . . , u
r−1
as in Lemma 10.4.1. Since K
X
=−
r−1
i=0
D
i
, (10.5.8) implies
K
X
K
X
=−
r−1
i=0
D
i
K
X
=−
r−1
i=0
(−D
i
D
i
−2) =2r +
r−1
i=0
D
i
D
i
.
If u
i−1
+u
u+i
=b
i
u
i
as in (10.4.1), then D
i
D
i
=−b
i
by Theorem 10.4.4. Hence
K
X
K
X
=2r −
r−1
i=0
b
i
=2r −(3r −12) =12−r,
500 Chapter 10. Toric Surfaces
where the equality
r−1
i=0
b
i
=3r −12 is from Exercise 10.4.11.
We next compute e(X). Exercise 10.4.3 shows that Σ is the normal fan of
a polygon with r = [Σ(1)[ sides. Then the formula for dimH
q
(X, Ω
p
X
) given in
Theorem 9.4.10 implies
dimH
0
(X, C) =dimH
0
(X, O
X
) =1
dimH
1
(X, C) =dimH
0
(X, Ω
1
X
) +dimH
1
(X, O
X
) =0+0 =0
dimH
2
(X, C) =dimH
0
(X, Ω
2
X
) +dimH
1
(X, Ω
1
X
) +dimH
2
(X, O
X
)
=0+ f
1
−2f
2
=r −2
dimH
3
(X, C) =dimH
1
(X, Ω
2
X
) +dimH
2
(X, Ω
1
X
) =0+0 =0
dimH
4
(X, C) =dimH
2
(X, Ω
2
X
) =1.
where f
1
=r and f
2
=1 are the face numbers of a polygon with r sides. It follows
that e(X) = 1 +(r −2) +1 = r. Then (10.5.10) follows easily from the above
computation of K
X
K
X
.
We will give a topological proof of e(X) =r in Chapter 12, and in Chapter 13,
we will interpret e(X) in terms of the Chern classes of the tangent bundle.
The RiemannRoch theorems for curves and surfaces have been vastly gen
eralized by results of Hirzebruch and Grothendieck, and the precise relation of
Noether’s theorem to the RiemannRoch theorem for surfaces is a special case of
their approach. We will discuss RiemannRoch theorems higherdimensional toric
varieties in Chapter 13.
Lattice Polygons. For the remainder of this section, we will explore the relation
between toric surfaces and the geometry and combinatorics of lattice polygons. We
will see that the results from §9.4 for lattice polytopes have an especially nice form
for lattice polygons.
Let X =X
Σ
be a smooth complete toric surface. Since χ(O
X
) =1 by (10.5.9),
RiemannRoch for a divisor D on X becomes
(10.5.11) χ(O
X
(D)) =
D D−D K
X
2
+1,
Thus, for any ℓ ∈ Z,
(10.5.12) χ(O
X
(ℓD)) =
ℓD ℓD−ℓD K
X
2
+1 =
1
2
(D D)ℓ
2
−
1
2
(D K
X
)ℓ +1.
The theory developed in §9.4 guarantees that χ(O
X
(ℓD)) is a polynomial in ℓ; the
above computation gives explicit formulas for the coefﬁcients in terms of intersec
tion products.
Here is an example of how this formula works.
§10.5. RiemannRoch and Lattice Polygons 501
Example 10.5.4. Let X be the Hirzebruch surface H
2
. We will use the notation of
Example 10.4.6. The divisor D = D
1
+D
2
is clearly effective, but the inequalities
(9.3.6) deﬁning the nef cone show that D is not nef. Using K
X
= −D
1
− −D
4
and Example 10.4.6, one computes that
D D=0, D K
X
=−2.
Then (10.5.12) implies that
χ(O
X
(ℓD)) =ℓ +1.
An easy application of Serre duality gives H
2
(X, O
X
(ℓD)) = 0 (Exercise 10.5.4).
Thus
dimH
0
(X, O
X
(ℓD)) −dimH
1
(X, O
X
(ℓD)) =ℓ +1.
Things get more surprising when we compute dimH
0
(X, O
X
(ℓD)). Using the ray
generators u
1
, . . . , u
4
from Example 10.4.6, the polygon P
D
corresponding to D =
D
1
+D
2
is deﬁned by the inequalities
¸m, u
1
) ≥−1, ¸m, u
2
) ≥−1, ¸m, u
3
) ≥0, ¸m, u
4
) ≥0.
The polygon P
D
is shown in Figure 11.
P
D
u
2
u
4
u
3
u
1
Figure 11. The polygon of the divisor D and the fan of H2
Even though P
D
is not a lattice polytope, Proposition 4.3.3 still applies. Thus
dimH
0
(X, O
X
(ℓD)) =[P
ℓD
∩M[ =[ℓP
D
∩M[ =
_
1
4
ℓ
2
+ℓ +1 ℓ even
1
4
ℓ
2
+ℓ +
3
4
ℓ odd,
where the ﬁnal equality follows from Exercise 9.4.12. Combining this with the
above computation of χ(O
X
(ℓD)), we obtain
dimH
1
(X, O
X
(ℓD)) =
_
1
4
ℓ
2
ℓ even
1
4
ℓ
2
−
1
4
ℓ odd.
This is a vivid example how the Euler characteristic smooths out the complicated
behavior of the individual cohomology groups. ♦
502 Chapter 10. Toric Surfaces
On the other hand, if D is nef, the higher cohomology is trivial by Demazure
vanishing, so that the Euler characteristic reduces to dimH
0
. We exploit this as
follows. Suppose that P ⊆ M
R
≃ R
2
is a lattice polygon. By Theorem 9.4.2 and
Example 9.4.4, the Ehrhart polynomial Ehr
P
(x) ∈ Q[x] of P satisﬁes
(10.5.13) Ehr
P
(ℓ) =[(ℓP) ∩M[ =Area(P)ℓ
2
+
1
2
[∂P∩M[ ℓ +1
for ℓ ∈ N. We next describe this polynomial in terms of intersection products.
By the results of §2.3, we get the projective toric surface X
P
coming from the
normal fan Σ
P
of P. In general X
P
will not be smooth, so we compute a minimal
resolution of singularities
φ : X
Σ
−→X
P
using the methods of this chapter. Recall that X
P
has the ample divisor D
P
whose
associated polygon is P.
Proposition 10.5.5. There is unique torusinvariant nef divisor D on X
Σ
such that
(a) The support function of D equals the support function of D
P
.
(b) χ(O
X
Σ
(ℓD)) is the Ehrhart polynomial of P.
Proof. Proposition 6.1.20 implies that there is a divisor D on X
Σ
such that satisﬁes
part (a). As in §6.1, we call D the pullback of D
P
. Since D
P
has a convex support
function, the same is true for D, so that D is nef. Furthermore, P is the polytope
associated to D
P
and hence is the polytope associated to D since the polytope of a
nef divisor is determined by its support function (Theorem 6.1.10).
It follows that dimH
0
(X
Σ
, O
X
Σ
(ℓD)) = [P
ℓD
∩M[ = [(ℓP) ∩M[ when ℓ ≥ 0,
so that H
0
(X
Σ
, O
X
Σ
(ℓD)) equals the Ehrhart polynomial of P when ℓ ≥ 0. How
ever, ℓD is nef when ℓ ≥ 0 and hence has trivial higher cohomology by Demazure
vanishing (Theorem 9.2.3). Thus ℓ ≥0 implies
χ(O
X
Σ
(ℓD)) =dimH
0
(X
Σ
, O
X
Σ
(ℓD)) =[(ℓP) ∩M[.
Since χ(O
X
Σ
(ℓD)) is a polynomial in ℓ, it must be the Ehrhart polynomial of P.
Theorem 10.5.5 and (10.5.12) imply that the Ehrhart polynomial of P is
Ehr
P
(ℓ) =
1
2
(D D)ℓ
2
−
1
2
(D K
X
Σ
)ℓ +1.
Comparing this to the formula (10.5.13) for Ehr
P
(ℓ), we get the following result.
Proposition 10.5.6. Let P be a lattice polygon and let D be the pullback of D
P
constructed in Proposition 10.5.5. Then
D D=2Area(P)
−D K
X
Σ
=[∂P∩M[.
§10.5. RiemannRoch and Lattice Polygons 503
Example 10.5.7. Take the fan of H
2
shown in Figure 11 from Example 10.5.4
and combine the two 2dimensional cones containing u
2
into a single cone. The
resulting fan has minimal generators u
1
, u
3
, u
4
that satisfy u
1
+u
3
+2u
4
=0, so the
resulting toric variety is P(1, 1, 2).
Let P = Conv(0, 2e
1
, −e
2
) ⊆ M
R
, which is the double of the polytope shown
in Figure 11. The normal fan of P is the fan of P(1, 1, 2). The minimal generators
u
1
, u
3
, u
4
of this fan give divisors D
′
1
, D
′
3
, D
′
4
on P(1, 1, 2), and the divisor D
P
is
easily seen to be the ample divisor 2D
′
1
.
Since the fan of H
2
reﬁnes the fan of P(1, 1, 2), the resulting toric morphism
H
2
→P(1, 1, 2) is a resolution of singularities. By considering the support function
of D
P
, we ﬁnd that the pullback of D
P
= 2D
′
1
is D = 2D
1
+2D
2
. We leave it as
Exercise 10.5.5 to compute D D and D K
H
2
and verify that they give the numbers
predicted by Proposition 10.5.6. ♦
Sectional Genus. The divisor D
P
on X
P
is very ample since dimP = 2. Hence
it gives a projective embedding X
P
֒→P
s
such that O
P
s (1) restricts to O
X
Σ
(D
P
).
In geometric terms, this means that hyperplanes H ⊆ P
s
give curves X
P
∩H ⊆ X
P
that are linearly equivalent to D
P
. For some hyperplanes, the intersection X
P
∩H
can be complicated. Since X
P
has only ﬁnitely many singular points, the Bertini
theorem [78, II 8.18 and III 7.9.1] guarantees that when H is generic, C = X
P
∩H
is a smooth connected curve contained in the smooth locus of X
P
. The genus g of
C is called the sectional genus of the surface X
P
.
We will compute g in terms of the geometry of P using the adjunction formula.
Since we need a smooth surface for this, we use a resolution φ : X
Σ
→X
P
and note
that C can be regarded as a curve in X
Σ
since φ is an isomorphism away from the
singular points of X
P
. Since C ∼ D
P
on X
P
, we have C ∼ D on X
Σ
, where D is the
pullback of D
P
. Then the adjunction formula (Theorem 10.5.1) implies
2g−2 =K
X
Σ
C+C C =K
X
Σ
D+D D,
so that
(10.5.14) g =
1
2
D (K
X
Σ
+D) +1.
Then we have the following result.
Proposition 10.5.8. The sectional genus of X
P
is g =[Int(P) ∩M[.
Proof. Pick’s formula from Example 9.4.4 can be written as
[Int(P) ∩M[ =Area(P) −
1
2
[∂P∩M[ +1,
which by Proposition 10.5.6 becomes
[Int(P) ∩M[ =
1
2
D D+
1
2
D K
X
Σ
+1.
The righthand side is g by (10.5.14), completing the proof.
504 Chapter 10. Toric Surfaces
Example 10.5.9. Let P = d∆
2
= Conv(0, de
1
, de
2
). Then X
P
is the projective
plane P
2
in its dth Veronese embedding, and D
P
∼dL, where L ⊆P
2
is a line. The
hyperplane sections are the curves of degree d in P
2
, and the smooth ones have
genus
g =[Int(d∆
2
) ∩M[ =
(d −1)(d −2)
2
.
You will check this assertion and another example in Exercise 10.5.6. ♦
The curves C ⊆ X
Σ
studied here can be generalized to the study of hypersur
faces in projective toric varieties coming from sections of a nef line bundle. The
geometry and topology of these hypersurfaces have been studied in many papers,
including [9, 10, 43, 122].
Reﬂexive Polygons and The Number 12. Our ﬁnal topic gives a way to understand
a somewhat mysterious formula we noted in the last section of Chapter 8. Recall
from Theorem 8.3.7 that there are exactly 16 equivalence classes of reﬂexive lattice
polytopes in R
2
, shown in Figure 3 of §8.3. The article [145] gives four different
proofs of the following result.
Theorem 10.5.10. Let P be a reﬂexive lattice polygon in M
R
≃R
2
. Then
[∂P∩M[ +[∂P
◦
∩N[ =12.
One proof consists of a casebycase veriﬁcation of the statement for each of
16 equivalence classes. You proved the theorem this way in Exercise 8.3.5. The
argument was straightforward but not very enlightening! Here we will give another
proof using Noether’s theorem.
Proof. Since Noether’s theorem requires a smooth surface, we need to reﬁne the
normal fan Σ
P
of P ⊆ M
R
. Since P is reﬂexive, we can do this using the dual
polygon P
◦
⊆ N
R
. We know from (8.3.2) that the vertices of P
◦
are the minimal
generators of Σ
P
. Let Σ be the reﬁnement of Σ
P
whose 1dimensional cones are
generated by the rays through the lattice points on the boundary of P
◦
. This is
illustrated in Figure 12.
s
s
s
s
s
s
s
s
s
s
d
d
d
d
d
d
P
◦
s
s
s
s
¨
¨
¨
¨
¡
¡
¡
¡d
d
P
e
e
e
e
r
r
r
r
Figure 12. A reﬂexive polygon P and its dual P
◦
§10.5. RiemannRoch and Lattice Polygons 505
The fan Σ has the following properties:
• For each cone of Σ, its minimal generators and the origin form a triangle whose
only lattice points are the vertices. Thus Σ is smooth by Exercise 8.3.4.
• The minimal generators of Σ are the lattice points of P
◦
lying on the boundary.
Thus [Σ(1)[ =[∂P
◦
∩N[.
From the ﬁrst bullet, we get a resolution φ : X
Σ
→X
P
. Recall that D
P
=−K
X
P
since
P is reﬂexive. The wonderful fact is that its pullback via φ is again anticanonical,
i.e., D=−K
X
Σ
. To prove this, recall that D and D
P
have the same support function
ϕ, which takes the value 1 at the vertices of P
◦
since D
P
= −K
X
P
. It follows that
ϕ =1 on the boundary of P
◦
. Then D = −K
X
Σ
because the minimal generators of
Σ all lie on the boundary.
Now apply Noether’s theorem to the toric surface X
Σ
. By (10.5.10), we have
K
X
Σ
K
X
Σ
+e(X
Σ
) =12.
We analyze each term on the left as follows. First, D=−K
X
Σ
implies
K
X
Σ
K
X
Σ
=−D K
X
Σ
=[∂P∩M[,
where the last equality follows from Proposition 10.5.6. Second, e(X
Σ
) is the num
ber of minimal generators of Σ by the proof of Theorem 10.5.3. In other words,
e(X
Σ
) =[Σ(1)[ =[∂P
◦
∩N[,
where the second equality follows from the above analysis of Σ. Hence the theorem
is an immediate consequence of Noether’s theorem.
A key step in the above proof was showing that the pullback of the canonical
divisor on X
P
was the canonical divisor on X
P
. This may fail for a general resolution
of singularities. We will say more about this when we study crepant resolutions in
Chapter 11.
Exercises for §10.5.
10.5.1. The RiemannRoch theorem for curves, in the form (10.5.1), can be proved by
much the same method as used in the proof of Theorem 10.5.2. Namely, show that if
(10.5.1) holds for a divisor D then it also holds for the divisors D+P and D−P, where P
is an arbitrary point on the curve.
10.5.2. Prove the adjunction formula (Theorem 10.5.1) using (10.5.3) and (10.5.5).
10.5.3. Complete the proof of Theorem 10.5.2 by showing that if the theorem holds for
D, then it also holds for D−D
i
where D
i
is any one of the divisors corresponding to the
1dimensional cones in Σ.
10.5.4. Let D=
ρ
a
ρ
D
ρ
be an effective QCartier Weil divisor on a complete toric variety
X
Σ
of dimension n. Use Serre duality (Theorem9.2.10) to prove that H
n
(X
Σ
, O
XΣ
(D)) =0.
10.5.5. In Example 10.5.7, compute D D and D K
H2
and check that they agree with the
numbers given by Proposition 10.5.6.
506 Chapter 10. Toric Surfaces
10.5.6. This exercise studies the sectional genus of toric surfaces.
(a) Verify the formula given in Example 10.5.9 for the sectional genus of P
2
in its dth
Veronese embedding.
(b) Let P =Conv(0, ae
1
, be
2
, ae
1
+be
2
). What is the smooth toric surface X
P
in this case?
Show that its sectional genus is (a −1)(b −1).
10.5.7. Let P be a reﬂexive polygon.
(a) Prove that the singularities (if any) of the toric surface X
P
are rational double points.
Hint: Proposition 10.1.6.
(b) Prove that X
P
has sectional genus g =1. This means that smooth anticanonical curves
in X
P
are all elliptic curves.
(c) Explain how part (b) relates to Exercise 10.5.6.
10.5.8. According to Theorem 10.4.3, every smooth toric surface is a blowup of either P
2
,
P
1
P
1
, or H
r
for r ≥ 2. For each of the 16 reﬂexive polygons in Figure 3 of §8.3, the
process described in the proof of Theorem 10.5.10 produces a smooth toric surface X
Σ
.
Where does X
Σ
ﬁt in this classiﬁcation in each case? (This gives a classiﬁcation of smooth
toric Del Pezzo surfaces.)
10.5.9. As in §9.3, the pEhrhart polynomials of a lattice polygon P ⊂M
R
are deﬁned by
Ehr
p
P
(ℓ) =χ(
´
Ω
p
XP
(ℓD
P
)), p =0, 1, 2.
We know that Ehr
0
P
is the usual Ehrhart polynomial Ehr
P
, and then Ehr
2
P
(x) = Ehr
P
(−x)
by Theorem 9.4.6. The remaining case is Ehr
1
P
. Prove that
Ehr
1
P
(x) =2Area(P)x
2
+ f
1
−2,
where f
1
is the number of edges of P. Hint: Use Theorem 9.4.10 for the constant term and
part (c) of Theorem 9.4.6 for the coefﬁcient of x. For the leading coefﬁcient, tensor the
exact sequence of Theorem 8.1.6 with O
XP
(ℓD
P
)), take the Euler characteristic, and then
let ℓ →∞.
Chapter 11
Toric Resolutions and
Toric Singularities
This chapter will study the singularities of a normal toric variety. We begin in
§11.1 with the existence of toric resolutions of singularities. In §11.2 we consider
more special resolutions, including simple normal crossing resolutions, crepant
resolutions, log resolutions, and embedded resolutions. There are also relations
with ideal sheaves, Rees algebras, and multiplier ideals, to be studied in §11.3, and
ﬁnally, in §11.4 we consider some important classes of toric singularities.
§11.1. Resolution of Singularities
The singular locus of an irreducible variety X, denoted X
sing
, is the set of all sin
gular points of X. One can prove that X
sing
is a proper closed subvariety of X (see
[78, Thm. I.5.3]). We call the complement X ¸X
sing
the smooth locus of X.
Deﬁnition 11.1.1. Given an irreducible variety X, a resolution of singularities of
X is a morphism f : X
′
→X such that:
(a) X
′
is smooth and irreducible.
(b) f is proper.
(c) f induces an isomorphism of varieties f
−1
(X ¸X
sing
) ≃X ¸X
sing
.
Furthermore, f : X
′
→X is a projective resolution if f is a projective morphism.
A resolution of singularities modiﬁes X to make it smooth without changing
the part that is already smooth. In particular, a resolution of singularities is a bira
tional morphism. Hironaka [87] proved the existence of a resolution of singularities
over an algebraically closed ﬁeld of characteristic 0. An introduction to Hironaka’s
proof and more recent developments can be found in [82].
507
508 Chapter 11. Toric Resolutions and Toric Singularities
Fortunately, the toric case is much simpler, partly because toric varieties cannot
have arbitrarily bad singularities. Hence we will be able to give a complete proof
of resolution of singularities for toric varieties. The key idea is that given a fan Σ
in N
R
, the resolution will be the toric morphism
φ : X
Σ
′ −→X
Σ
coming from a suitable reﬁnement Σ
′
of Σ.
The Singular Locus. The ﬁrst step is to determine the singular locus of X
Σ
. This
has the following purely toric description.
Proposition 11.1.2. Let X
Σ
be the toric variety of the fan Σ. Then:
(X
Σ
)
sing
=
_
σ not smooth
V(σ)
X
Σ
¸(X
Σ
)
sing
=
_
σ smooth
U
σ
,
where V(σ) =O(τ) is the closure of the T
N
orbit corresponding to σ.
Proof. Recall that σ is smooth if its minimal generators can be extended to a basis
of N and that σ is smooth if and only if U
σ
is smooth. Also observe that
• if σ is smooth, then so is every face of σ, hence
• if σ is not smooth, then so is every cone of Σ containing σ.
The ﬁrst bullet tells us that the smooth cones of Σ form a fan whose toric variety is
σ smooth
U
σ
. This open set of X
Σ
is clearly smooth, and its complement in X
Σ
is
σ not smooth
O(σ) by the OrbitCone correspondence.
We also have
σ not smooth
V(σ) =
σ not smooth
O(σ) by the second bullet and
the OrbitCone correspondence. Hence the proof will be complete once we show
that every point of O(σ) is singular in X
Σ
when σ is not smooth. It sufﬁces to work
in the afﬁne open set U
σ
. Let N
σ
= Span(σ) ∩N and pick a complement N
2
⊆ N
such that N =N
σ
⊕N
2
. By (1.3.2), we have
U
σ,N
≃U
σ,Nσ
T
N
2
.
Since dimσ =rankN
σ
, the orbit O
Nσ
(σ) ⊆U
σ,Nσ
consists of the unique ﬁxed point
of the action of T
Nσ
. Since σ is not smooth with respect to N
σ
, the proof of Theo
rem 1.3.12 shows that this ﬁxed point is singular in U
σ,Nσ
. Then (1.3.3) shows that
every point of O
Nσ
(σ) T
N
2
is singular in U
σ,Nσ
T
N
2
. Since the above product
decomposition induces
O
N
(σ) ≃O
Nσ
(σ) T
N
2
(Exercise 11.1.1), we see that all points of O(σ) =O
N
(σ) are singular in X
Σ
.
It follows that when constructing reﬁnements of a fan Σ, we need to be sure to
leave the smooth cones of Σ unchanged.
§11.1. Resolution of Singularities 509
Star Subdivisions. We will construct reﬁnements using a generalization of the star
subdivision Σ
∗
(σ) deﬁned in §3.3. Given a fan Σ in N
R
and a primitive element
v ∈ [Σ[ ∩N¸¦0¦, let Σ
∗
(v) be the set of the following cones:
(a) σ, where v / ∈ σ ∈ Σ.
(b) Cone(τ, v), where v / ∈ τ ∈ Σ and ¦v¦∪τ ⊆σ ∈ Σ.
We call Σ
∗
(v) the star subdivision of Σ at v. This terminology is justiﬁed by the
following result.
Lemma 11.1.3. Σ
∗
(v) is a reﬁnement of Σ.
Proof. Let Cone(τ, v) be a cone as in (b) above. Then observe that v / ∈ Span(τ)
since v / ∈ τ and τ =σ ∩Span(τ). It follows that dimCone(τ, v) =dimτ +1.
To show that Σ
∗
(v) is a fan, ﬁrst consider a face of a cone in Σ
∗
(v). If the
cone arises from (a) above, then the face clearly lies in Σ
∗
(v). If the cone arises
from (b), then we have a face γ of Cone(τ, v). Write γ = H
m
∩Cone(τ, v) for
m∈ Cone(τ, v)
∨
and observe that m ∈ τ
∨
and ¸m, v) ≥0. There are two cases:
• ¸m, v) =0, in which case γ =Cone(H
m
∩τ, v) ∈ Σ
∗
(v).
• ¸m, v) > 0, in which case γ =H
m
∩τ ∈ Σ
∗
(v).
You will provide the details of these cases in Exercise 11.1.2.
Next consider two cones in Σ
∗
(v). If neither contains v, then the intersection
is clearly in Σ
∗
(v). Now suppose that one does not contain v and the other does.
This gives cones v / ∈ σ and Cone(τ, v) as in (a) and (b) above. We claim that
(11.1.1) σ∩Cone(τ, v) =σ ∩τ,
which will imply that σ∩Cone(τ, v) ∈Σ
∗
(v). One inclusion of (11.1.1) is obvious;
for the other inclusion, let u ∈ σ ∩Cone(τ, v) and write u =u
0
+λv for u
0
∈ τ and
λ ≥0. Also let σ
′
∈ Σ be a cone containing ¦v¦∪τ. Then
u =u
0
+λv ∈ σ ∩Cone(τ, v) ⊆σ ∩σ
′
.
Since u
0
, λv ∈σ
′
and σ∩σ
′
is a face of σ
′
, it follows that u
0
, λv ∈σ∩σ
′
. But v / ∈σ,
so that λ =0 and hence u =u
0
∈ τ. Thus u ∈ σ∩τ, as desired.
The remaining case to consider is when both cones contain v. If the cones are
Cone(τ
1
, v) and Cone(τ
2
, v), then we claim that
(11.1.2) Cone(τ
1
, v) ∩Cone(τ
2
, v) =Cone(τ
1
∩τ
2
, v),
which will imply that Cone(τ
1
, v) ∩Cone(τ
2
, v) ∈ Σ
∗
(v). As above, one inclusion
is obvious, and for the other, take u ∈ Cone(τ
1
, v) ∩Cone(τ
2
, v). Then
u =u
1
+λ
1
v =u
2
+λ
2
v,
where u
i
∈ τ
i
and λ
i
≥ 0. We may assume that λ
1
≥ λ
2
, in which case u
2
= u
1
+
(λ
2
−λ
1
)ν ∈ Cone(τ
1
, v). Thus u
2
∈ τ
2
∩Cone(τ
1
, v), which by (11.1.1) implies
u
2
∈ τ
2
∩Cone(τ
1
, v) =τ
1
∩τ
2
.
510 Chapter 11. Toric Resolutions and Toric Singularities
Thus u =u
2
+λ
2
v ∈ Cone(τ
1
∩τ
2
, v), as desired.
Finally, we need to show that Σ
∗
(v) is a reﬁnement of Σ. Since every cone of
Σ
∗
(v) is clearly contained in a cone of Σ, we need only show that each σ ∈ Σ is
a union of cones in Σ
∗
(v). When v / ∈ σ, this is obvious since σ ∈ Σ
∗
(v). On the
other hand, when v ∈ σ, pick cone generators m
i
∈ σ
∨
, i = 1, . . . , r, and note that
v ∈ σ implies ¸m
i
, v) ≥0 for all i. Then set
(11.1.3) λ = min
¸m
i
,v)>0
¸m
i
, u)
¸m
i
, v)
.
We claim that u
0
= u −λv lies in σ. To prove this, take m =
r
i=1
µ
i
m
i
∈ σ
∨
,
µ
i
≥0, and compute
¸m, u
0
) =
¸
r
i=1
µ
i
m
i
, u−λv
_
=
¸m
i
,v)=0
µ
i
¸m
i
, u) +
¸m
i
,v)>0
µ
i
¸m
i
, v)
_
¸m
i
, u)
¸m
i
, v)
−λ
_
.
The ﬁrst sum is ≥0 since u ∈σ, and the second is ≥0 by the deﬁnition of λ. Hence
u
0
∈ (σ
∨
)
∨
=σ, as claimed.
The deﬁnition (11.1.3) implies that we can pick i with ¸m
i
, v) > 0 and λ =
¸m
i
,u)
¸m
i
,v)
. Then ¸m
i
, u
0
) =0, so that u
0
lies in the face τ =H
m
i
∩σ of σ. This face does
not contain v since ¸m
i
, v) > 0, hence u ∈ Cone(τ, v) ∈ Σ
∗
(v).
Example 11.1.4. Let σ = Cone(u
1
, . . . , u
r
) ∈ Σ be a smooth cone in N
R
≃ R
n
.
In §3.3, we deﬁned the star subdivision Σ
∗
(σ) using u
0
= u
1
+ +u
r
. It is easy
to check that this is the star subdivision Σ
∗
(u
0
) since u
0
∈ σ ∩N is primitive. See
Figure 9 in §3.3 for a picture when n =3. ♦
Example 11.1.5. Figure 1 shows a fan Σ in R
3
consisting of two 3dimensional
cones (one simplicial, the other not) that meet along a 2dimensional face τ. The
v
τ
↓
Figure 1. A fan Σ and its star subdivision Σ
∗
(v)
§11.1. Resolution of Singularities 511
ﬁgure also shows the star subdivision induced by a primitive vector v ∈ Relint(τ).
The fan Σ
∗
(v) has ﬁve 3dimensional cones. ♦
The deﬁnition of Σ
∗
(v) implies that v generates a 1dimensional cone of Σ
∗
(v)
and hence gives a divisor on X
Σ
∗
(v)
. Recall that a Weil divisor is QCartier if some
positive integer multiple is Cartier.
Proposition 11.1.6. The star subdivision Σ
∗
(v) has the following properties:
(a) The 1dimensional cones of Σ
∗
(v) consist of the 1dimesional cones of Σ plus
the ray ρ
v
=Cone(v) generated by v, i.e.,
Σ
∗
(v)(1) =Σ(1) ∪¦ρ
v
¦.
(b) The torusinvariant divisor prime D
ρv
is QCartier on X
Σ
∗
(v)
.
(c) The induced toric morphism φ : X
Σ
∗
(v)
→X
Σ
is projective.
Proof. The ﬁrst assertion is Exercise 11.1.3. To show that D
ρv
is QCartier, we
construct a support function ϕ : [Σ
∗
(v)[ →R as follows. For Cone(τ, v) ∈ Σ
∗
(v),
consider the linear function deﬁned on Span(Cone(τ, v)) =Span(τ)+Rv that maps
Span(τ) to 0 and v to 1. This is welldeﬁned since v / ∈ Span(τ). Then deﬁne
ϕ[
Cone(τ,v)
to be the restriction of this linear map to Cone(τ, v). The remaining
cones in Σ
∗
(v) are the cones σ ∈ Σ not containing v. We set ϕ[
σ
= 0 for each of
these. By (11.1.1) and (11.1.2), we get a well deﬁned map ϕ : [Σ
∗
(v)[ →R that is
linear on each cone of Σ
∗
(v). Furthermore, ϕ is clearly deﬁned over N
Q
since τ
is rational and v ∈ N. It follows that some positive integer multiple kϕ is integral
with respect to N, i.e., kϕ ∈ SF(Σ
∗
(v), N). Then by Theorem 4.2.12,
D
kϕ
=−
ρ∈Σ
∗
(v)(1)
kϕ(u
ρ
)D
ρ
is a Cartier divisor on X
Σ
∗
(v)
. However, v is primitive and hence is the ray generator
of ρ
v
. Since ϕ vanishes on all other ray generators, we have D
kϕ
=−kϕ(v)D
ρv
=
−kD
ρv
. This proves that −D
ρv
and hence D
ρv
are QCartier.
For the ﬁnal assertion, we use the projectivity criterion from Theorem 7.2.9.
Since Σ
∗
(v) reﬁnes Σ, it sufﬁces to ﬁnd a torusinvariant Cartier divisor on X
Σ
∗
(v)
whose support function is strictly convex with respect to the fan
Σ
∗
(v)
σ
=¦σ
′
∈ Σ
∗
(v) [ σ
′
⊆σ¦
for every σ ∈Σ. The required Cartier divisor is easy to ﬁnd: it is the divisor −kD
ρv
,
where k > 0 is from the previous paragraph, with support function kϕ. We can
ignore the factor of k when thinking about strict convexity. Hence the proof will be
complete once we prove that ϕ has the desired strict convexity.
When v / ∈ σ, this is obvious since σ ∈ Σ
∗
(v) in this case. So it remains to con
sider what happens when v ∈ σ. By restricting to Span(σ), we can assume that σ is
512 Chapter 11. Toric Resolutions and Toric Singularities
full dimensional. Then, as noted in the discussion leading up to Proposition 7.2.3,
the strict convexity criteria from Lemma 6.1.13 apply to this situation.
The analysis of Σ
∗
(v) given in the proof of Lemma 11.1.3 shows that σ =
τ
Cone(τ, v), where the union is over all facets τ of σ not containing v. Now ﬁx
such a facet τ and consider the cone σ
′
= Cone(τ, v) ∈ Σ
∗
(v). Pick m ∈ σ
∨
∩M
such that H
m
∩σ =τ. Since v ∈ σ ¸τ, we have
(11.1.4) ¸m, u
ρ
) =0, ρ ∈ τ(1) and ¸m, v) > 0.
We claim that after rescaling m by a positive constant, we have ϕ(u) = ¸m, u) for
all u ∈σ
′
. To see why, note that ϕ is linear on σ
′
, vanishes on τ, and takes the value
1 on v. Comparing this to (11.1.4), our claim follows immediately.
Given any ρ ∈ σ(1) ¸τ(1), the way we picked m implies that
¸m, u
ρ
) > 0.
However, m represents ϕ on σ
′
, and
σ(1) ¸τ(1) =Σ
∗
(v)
σ
(1) ¸σ
′
(1)
by part (a) and the deﬁnitions of Σ
∗
(v) and Σ
∗
(v)
σ
. By (f) ⇒(a) of Lemma 6.1.13,
it follows that ϕ[
σ
is strictly convex with respect to Σ
∗
(v)
σ
.
Simplicialization. As an application of star subdivisions, we prove that every fan
has an efﬁcient simplicial reﬁnement. Here is the precise result.
Proposition 11.1.7. Every fan Σhas a reﬁnement Σ
′
with the following properties:
(a) Σ
′
is simplicial.
(b) Σ
′
(1) =Σ(1).
(c) Σ
′
contains every simplicial cone of Σ.
(d) Σ
′
is obtained from Σ by a sequence of star subdivisions.
(e) The induced toric morphism X
Σ
′ →X
Σ
is projective.
Proof. First observe that part (c) follows from part (b). To see why, suppose that
Σ
′
is a reﬁnement of Σ satisfying Σ
′
(1) = Σ(1) and let σ ∈ Σ be simplicial. If
σ
′
∈ Σ
′
lies in σ, then Σ
′
(1) = Σ(1) implies σ
′
(1) ⊆ σ(1), and hence σ
′
is a face
of σ because the latter is simplicial. Since σ =
σ
′
∈Σ
′
,σ
′
⊆σ
σ
′
, it follows that some
σ
′
in this union must equal σ, i.e., σ ∈ Σ
′
.
Note also that part (e) is a consequence of part (d). This follows because, ﬁrst,
star subdivisions give projective morphisms by Proposition 11.1.6, and second,
compositions of projective morphisms are projective by Proposition 7.0.5.
Hence it sufﬁces to ﬁnd a sequence of star subdivisions that lead to a simpli
cial fan with the same 1dimensional cones as Σ. Our proof, based on an idea of
Thompson [174], uses complete induction on
r =[¦ρ ∈ Σ(1) [ D
ρ
is not QCartier on X
Σ
¦[.
§11.1. Resolution of Singularities 513
If r = 0, then every D
ρ
is QCartier. Since the D
ρ
generate Cl(X
Σ
), every Weil
divisor on X
Σ
is QCartier. By Proposition 4.2.7, we conclude that Σ is simplicial.
If r > 0, then we can pick ρ ∈ Σ(1) such that D
ρ
is not QCartier. Consider
the star subdivision Σ
∗
(u
ρ
), where u
ρ
is the ray generator of ρ. Proposition 11.1.6
implies that
Σ
∗
(u
ρ
)(1) =Σ(1) ∪¦ρ¦ =Σ(1)
and that D
ρ
is QCartier on X
Σ
∗
(uρ)
. Note also that if D
ρ
′ is QCartier on X
Σ
, then
it remains QCartier on X
Σ
∗
(uρ)
(think support functions). It follows easily that
¦ρ
′
∈ Σ
∗
(u
ρ
)(1) [ D
ρ
′
is not QCartier on X
Σ
∗
(uρ)
¦
has strictly fewer than r elements. Our inductive hypothesis gives a reﬁnement Σ
′
of Σ
∗
(u
ρ
) which is easily seen to be the desired reﬁnement of Σ.
In Exercise 11.1.4 you will show that the number of star subdivisions needed
to create the simplicial reﬁnement Σ
′
of Σ described in Proposition 11.1.7 is at
most the rank of Cl(X
Σ
)/Pic(X
Σ
). A different proof of Proposition 11.1.7 can be
found in [55]. See also [51, Thm. V.4.2].
The Multiplicity of a Simplicial Cone. When we have a fan that is simplicial but
not smooth, we need to subdivide the nonsmooth cones to make them smooth.
The elegant approach used in Chapter 10 does not generalize to higher dimensions.
Instead, we will use the multiplicity of a simplicial cone from §6.3 and show how a
carefully chosen star subdivision of a nonsmooth cone can lower its multiplicity.
Let us recall the deﬁnition of multiplicity. Given a simplicial cone σ ⊂N
R
with
generators u
1
, . . . , u
d
, let N
σ
=Span(σ)∩N and note that Zu
1
+ +Zu
d
⊆N
σ
has
ﬁnite index. Then the multiplicity or index of σ is the index
(11.1.5) mult(σ) =[N
σ
: Zu
1
+ +Zu
d
].
The multiplicity of a simplicial cone has the following properties.
Proposition 11.1.8. Let σ ⊆ N
R
be a simplicial cone and let u
1
, . . . , u
d
and N
σ
be
as in (11.1.5). Then:
(a) σ is smooth if and only if mult(σ) =1.
(b) mult(σ) is the number of points in P
σ
∩N, where
P
σ
=
_
d
i=1
λ
i
u
i
: 0 ≤λ
i
< 1¦.
(c) Let e
1
, . . . , e
d
be a basis of N
σ
and write u
i
=
d
j=1
a
i j
e
j
. Then
mult(σ) =[det(a
i j
)[.
(d) mult(τ) ≤mult(σ) whenever τ _σ.
514 Chapter 11. Toric Resolutions and Toric Singularities
Proof. The proof of part (a) is straightforward, and part (b) is equally easy since
the composition
P
σ
∩N ֒→N
σ
−→N
σ
/(Zu
1
+ +Zu
d
)
is a bijection (Exercise 11.1.5). Part (c) is the standard fact that [det(a
i j
)[ is the
index of Zu
1
+ +Zu
d
in N
σ
(see [154, Corollary 9.63]). Finally, if τ _ σ, then
τ(1) ⊆σ(1). This implies P
τ
⊆P
σ
, and then part (d) follows from part (b).
Some further properties of mult(σ) are discussed in Exercise 11.1.5.
The Resolution. We can now construct a fan that resolves the singularities of a
normal toric variety X
Σ
.
Theorem 11.1.9. Every fan Σ has a reﬁnement Σ
′
with the following properties:
(a) Σ
′
is smooth
(b) Σ
′
contains every smooth cone of Σ.
(c) Σ
′
is obtained from Σ by a sequence of star subdivisions.
(d) The toric morphism φ : X
Σ
′ →X
Σ
is a projective resolution of singularities.
Proof. We ﬁrst observe that part (d) follows from parts (a), (b) and (c). Since
the reﬁnement Σ
′
comes from a sequence of star subdivisions, φ : X
Σ
′ →X
Σ
is
projective by the proof of Proposition 11.1.7. Furthermore, φ is the identity on
the smooth locus of X
Σ
by Proposition 11.1.2 and X
Σ
′
is smooth Theorem 3.1.19.
Hence φ is a resolution of singularities.
To produce a reﬁnement that satisﬁes (a), (b) and (c), we begin with Proposi
tion 11.1.7, which uses star subdivisions to construct a simplicial reﬁnement with
the same simplicial cones as Σ. This reﬁnement does not change the smooth cones
since smooth cones are simplicial.
Hence we may assume that Σis simplicial. To measure how close Σis to being
smooth, we deﬁne
mult(Σ) =max
σ∈Σ
mult(σ).
Note that mult(Σ) =1 if and only if Σis a smooth fan. We will show that whenever
mult(Σ) > 1, we can ﬁnd a star subdivision Σ
∗
(v) that does not change the smooth
cones of Σ and satisﬁes either
(11.1.6)
mult(Σ
∗
(v)) <mult(Σ), or
mult(Σ
∗
(v)) =mult(Σ) and Σ
∗
(v) has fewer cones of this multiplicity.
Once this is done, the theorem will follow immediately.
Suppose mult(Σ) >1 and let σ
0
∈ Σ have maximal multiplicity. By part (b) of
Proposition 11.1.8, there is v ∈ P
σ
0
∩N ¸¦0¦. The star subdivision Σ
∗
(v) consists
§11.1. Resolution of Singularities 515
of the cones of Σ not containing v together with the cones Cone(τ, v) where v / ∈ τ
and τ is a face of a cone σ ∈ Σ containing v. We claim that
(11.1.7) mult(Cone(τ, v)) < mult(σ).
Assume (11.1.7) for the moment. Part (b) of Proposition 11.1.8 shows that v
lies in no smooth cone of Σ, so that Σ
∗
(v) contains all smooth cones of Σ. Since all
cones of Σ
∗
(v) either lie in Σ or satsify the inequality (11.1.7), it follows that when
we replace Σto Σ
∗
(v), we lose at least one cone of maximum multiplicity (namely,
σ
0
) and create no new cones of this multiplicity. Thus Σ
∗
(v) satisﬁes (11.1.6).
It remains to prove (11.1.7). Given v ∈ P
σ
0
∩N ∈ ¦0¦ as above, we can write
v =λ
1
u
1
+ +λ
d
u
d
, 0 < λ
i
< 1,
where u
1
, . . . , u
d
are the ray generators of the minimal face τ
0
of σ
0
containing v.
We also have Cone(τ, v) ⊆ σ. Since v ∈ σ implies that τ
0
is a face of σ, we can
write the ray generators of σ as
u
1
, . . . , u
d
, u
d+1
, . . . , u
s
, s =dimσ.
Also, v / ∈ τ implies that there is i ∈ ¦1, . . . , d¦ such that u
i
/ ∈ τ. Then
mult(Cone(τ, v)) ≤mult(Cone(u
1
, . . . , ´ u
i
, . . . , u
d
, u
d+1
, . . . , u
s
, v)
=[det(u
1
, . . . , ´ u
i
, . . . , u
d
, u
d+1
, . . . , u
s
, v)[
=[det(u
1
, . . . , ´ u
i
, . . . , u
d
, u
d+1
, . . . , u
s
, λ
i
u
i
)[
=λ
i
[det(u
1
, . . . , u
s
)[ =λ
i
mult(σ) < mult(σ).
The ﬁrst line follows from part (d) of Proposition 11.1.8 and the second follows
from part (c). Here, det(w
1
, . . . , w
d
) = det(b
i j
) where w
i
=
d
j=1
b
i j
e
j
. The last
two lines follow from 0 < λ
i
< 1 and standard properties of determinants. This
completes the proof of (11.1.7), and the theorem follows.
The Exceptional Locus. The resolution φ : X
Σ
′ →X
Σ
from Theorem 11.1.9 is an
isomorphism above the smooth locus of X
Σ
. It remains to consider what happens
above the singular locus. The exceptional locus of φ is Exc(φ) =φ
−1
((X
Σ
)
sing
).
Proposition 11.1.2 implies that the irreducible components of (X
Σ
)
sing
are given
by the orbit closures
(X
Σ
)
sing
=V(σ
1
) ∪ ∪V(σ
s
)
where σ
1
, . . . , σ
s
are the minimal nonsmooth cones of Σ. Hence, to understand the
exceptional locus Exc(φ), it sufﬁces to describe φ
−1
(V(σ
i
)) for 1 ≤i ≤s. We will
use the following more general result.
Proposition 11.1.10. Let Σ
′
be a reﬁnement of Σ with induced toric morphism
φ : X
Σ
′ →X
Σ
, and ﬁx σ ∈ Σ. Then the irreducible components of φ
−1
(V(σ)) are
φ
−1
(V(σ)) =V(σ
′
1
) ∪ ∪V(σ
′
r
),
where σ
′
1
, . . . , σ
′
r
are the minimal cones of Σ
′
that meet the relative interior of σ.
516 Chapter 11. Toric Resolutions and Toric Singularities
Proof. Since φ is equivariant, it maps T
N
′ orbits into T
N
orbits. More precisely, we
know from Lemma 3.3.21 that for σ ∈ Σ, a cone σ
′
∈Σ
′
satisﬁes φ(O(σ
′
)) ⊆O(σ)
if and only if σ is the minimal cone of Σ containing σ
′
. Furthermore, the latter
happens if and only if σ
′
intersects the relative interior of σ. From here, the proof
is an easy application of the OrbitCone correspondence (Exercise 11.1.6).
Example 11.1.11. Our ﬁrst example of a toric resolution was in Example 10.1.9,
where we considered the reﬁnement Σ of the cone σ shown in Figure 2. Here, U
σ
σ Σ
Figure 2. A cone σ with smooth reﬁnement Σ
has a unique singular point, and the minimal cones of Σ that meet the interior of σ
are the interior rays of Σ. Hence the exceptional locus of X
Σ
→U
σ
is a divisor. ♦
Example 11.1.12. A favorite cone is σ = Cone(e
1
, e
2
, e
1
+e
3
, e
2
+e
3
) ⊆ R
3
. In
Figure 3 on the next page, we show σ and three smooth reﬁnements Σ
1
, Σ
2
, Σ
3
that
give toric morphisms
X
Σ
3
y
y
y
y
""
E
E
E
E
X
Σ
1
""
F
F
F
F
X
Σ
2
x
x
x
x
U
σ
For Σ
1
, Σ
2
, the smallest cone meeting the interior of σ is 2dimensional, so the
exceptional locus is P
1
in each case. On the other hand, the ray τ in Figure 3 is the
smallest cone of Σ
3
meeting the interior of σ. Hence the exceptional locus is V(τ),
which by Example 3.2.8 is P
1
P
1
.
The exceptional loci in the above resolutions are related by the diagram
(11.1.8)
P
1
P
1
zzu
u
u
u
u
$$
I
I
I
I
I
P
1
$$
I
I
I
I
I
// _ _ _ _ _ _ _
P
1
zzu
u
u
u
u
¦pt¦
where the nontrivial maps in (11.1.8) are the projections onto the factors of P
1
P
1
(Exercise 11.1.7). ♦
§11.1. Resolution of Singularities 517
z
y
x
σ
z
y
x
τ
Σ
3
z
y
x
Σ
1
z
y
x
Σ
2
Figure 3. The cone σ with smooth reﬁnements Σ1, Σ2, Σ3
518 Chapter 11. Toric Resolutions and Toric Singularities
This example was ﬁrst described by Atiyah [2] in 1958. The rational map
X
1
X
2
in (11.1.8) is an isomorphism outside of a set of codimension 2. In the
minimal model program, such rational maps are called ﬂips or ﬂops, depending on
the context. We will say more about this in Chapter 15.
In §10.4, we noted that when X
Σ
is a toric surface, there is a unique minimal
toric resolution of singularities X
Σ
′ →X
Σ
. Example 11.1.12 shows that unique
minimal resolutions do not exist in dimensions ≥3.
Exercises for §11.1.
11.1.1. The proof of Proposition 11.1.2 constructed an isomorphism U
σ,N
≃U
σ,Nσ
T
N2
and claimed that it induced an isomorphism of orbits
O
N
(σ) ≃O
Nσ
(σ) T
N2
.
Prove this.
11.1.2. Let Cone(τ, v) be cone of the star subdivision Σ
∗
(v) containing v, and let γ be the
face of Cone(τ, v) determined by m∈ Cone(τ, v)
∨
. Thus γ =H
m
∩Cone(τ, v). As noted in
the text, we have m ∈ σ
∨
and ¸m, v) ≥0.
(a) If ¸m, v) =0, prove that γ =Cone(H
m
∩τ, v) ∈ Σ
∗
(v).
(b) If ¸m, v) > 0, in which case γ =H
m
∩τ ∈ Σ
∗
(v).
11.1.3. Let Σ
∗
(v) be the star subdivision determined by v and let ρ
v
=Cone(v). Prove that
Σ
∗
(v)(1) =Σ(1) ∪¦ρ
v
¦.
11.1.4. Recall that if G is a ﬁnitely generated abelian group and G
tor
is the subgroup of
elements of ﬁnite order, then G/G
tor
≃ Z
r
, where r is called the rank of G. Prove that the
number to star subdivisions need to create the simplicial reﬁnement Σ
′
of Σ described in
Proposition 11.1.7 is at most the rank of Cl(X
Σ
)/Pic(X
Σ
).
11.1.5. This exercise will study the multiplicity mult(σ) of a simplicial cone σ.
(a) Prove part (a) of Proposition 11.1.8.
(b) Prove part (b) of Proposition 11.1.8.
(c) Prove that mult(σ) is the normalized volume of P
σ
⊆ (N
σ
)
R
= Span(σ), where “nor
malized” means that the parallelepiped determined by a basis of N
σ
has volume 1.
Hint: Use part (c) of Proposition 11.1.8 and remember how to compute the volume of
a parallelepiped (see [169, Theorem 15.2.1]).
(d) Let τ be a face of σ and let u
1
, . . . , u
d
be the minimal generators of σ. Prove that
mult(σ) =mult(τ)[N
σ
: N
τ
+Zu
1
+ +Zu
d
]
and use this to give another proof of part (d) of Proposition 11.1.8.
11.1.6. Complete the proof of Proposition 11.1.10.
11.1.7. Verify the claims made in Example 11.1.12.
11.1.8. Let Σ be the fan in R
n+1
consisting of Cone(e
0
, . . . , e
n
) ⊆ R
n+1
and its faces. Let
v =
n
i=0
q
i
e
i
, where the q
i
are positive integers satisfying gcd(q
0
, . . . , q
n
) = 1. The star
subdivision Σ
∗
(v) give a toric morphism φ : X
Σ
∗
(v)
→X
Σ
= C
n+1
. Prove that φ
−1
(0) ≃
P(q
0
, . . . , q
n
).
§11.2. Other Types of Resolutions 519
11.1.9. Consider the complete fan Σ in R
3
with six 3dimensional cones and minimal
generators (±1, ±1, ±1). Thus Σ consists of the cones over the faces of a cube centered at
the origin in R
3
.
(a) Show that the simplicialization process described in Theorem 11.1.7 requires exactly
two star subdivisions to produce a simplicial reﬁnement Σ
′
of Σ.
(b) Show that Σ
′
is smooth, so that X
Σ
′ →X
Σ
is a projective resolution of singularities.
(c) Find a smooth reﬁnement Σ
′′
of Σ such that Σ
′′
(1) = Σ(1) but X
Σ
′′ is not projective.
Conclude that X
Σ
′′ →X
Σ
is not a projective resolution of singularities.
11.1.10. The simplicial fan constructed in Proposition 11.1.7 is efﬁcient (no new rays)
but noncanonical (no canonical choice for the star divisions used in construction). Here
you will explore a different method for simplicialization that is canonical but introduces
many new rays. The barycenter v
σ
of a nonzero cone σ ⊆ N
R
is the minimal generator
of Cone(
ρ∈σ(1)
u
ρ
) ∩N. Now list the nonzero cones of the fan Σ as σ
1
, . . . , σ
r
, where
dimσ
1
≤ ≤dimσ
r
. Then the barycentric subdivision of Σ is the fan
β(Σ) =
_
_
Σ
∗
(v
σr
)
_
∗
(v
σr−1
)
_
∗
(v
σ1
).
Thus β(Σ) is obtained fromΣby a sequence of star subdivisions where we use the barycen
ters of the cones, starting with the biggest cones and working down.
(a) Consider the fan Σ in R
3
from Exercise 11.1.9. Draw two pictures to illustrate the
construction of β(Σ). The ﬁrst picture, drawn on the surface of the cube, should
illustrate the intermediate fan obtained by taking star subdivisions at the barycenters
of the 3dimensional cones. The second picture should show how this gets further
subdivided to obtain β(Σ).
(b) When the cones are listed by increasing dimension, cones of the same dimension can
appear in any order. Hence the list is not unique. However, they all give the same
barycentric subdivision β(Σ). Prove this.
(c) Prove that β(Σ) is a simplicial fan.
Part (a) of this exercise was inspired by [51, p. 74].
§11.2. Other Types of Resolutions
We next consider simple normal crossing, crepant, log, and embedded resolutions,
which are important reﬁnements of what it means to be a resolution of singularities.
SNC Resolutions. Example 11.1.12 illustrates that the exceptional locus of a res
olution of singularities need not always be a divisor. Yet in many situations in
algebraic geometry, one wants a divisor, and in fact, one often wants a smooth
normal crossing divisor, as mentioned in §8.1. Recall that a divisor D =
i
D
i
on a smooth variety X has smooth normal crossings (SNC for short) if every D
i
is
smooth and irreducible, and for all p ∈X, the divisors containing p meet as follows:
if I
p
=¦i [ p ∈ D
i
¦, then the tangent spaces T
p
(D
i
) ⊆T
p
(X) meet transversely, i.e.,
codim
_
i∈Ip
T
p
(D
i
)
_
=[I
p
[.
520 Chapter 11. Toric Resolutions and Toric Singularities
For example, given a smooth toric variety and a nonempty subset A ⊆ Σ(1), the
divisor D=
ρ∈A
D
ρ
is SNC.
Deﬁnition 11.2.1. A resolution of singularities f : X
′
→X is called SNC if the
exceptional locus Exc( f ) is a divisor with simple normal crossings.
Theorem 11.2.2. Every fan Σ has a reﬁnement Σ
′
as in Theorem 11.1.9 such that
the toric morphism φ : X
Σ
′ →X
Σ
is a projective SNC resolution of singularities.
Proof. By Theorem 11.1.9, Σ has a smooth reﬁnement Σ
′
that gives a resolution
of singularities. Let V(σ
i
) be an irreducible component of (X
Σ
)
sing
and let σ ∈ Σ
′
be a minimal cone that meets the relative interior of σ
i
. If σ is not a ray, then
consider the star subdivision Σ
′∗
(v
σ
), where v
σ
=
ρ∈σ(1)
u
ρ
. Since Σ
′
is smooth,
this is the fan denoted Σ
′∗
(σ) in Deﬁnition 3.3.17. In the discussion that followed
the deﬁnition in §3.3, we saw that corresponding toric morphism is the blowup
Bl
V(σ)
(X
Σ
′
) →X
Σ
′
. Since v
σ
lies in the relative interior of σ, the inverse image of
V(σ) is the divisor corresponding to the ray of Σ
′∗
(v
σ
) generated by v
σ
.
The composed map Bl
V(σ)
(X
Σ
′
) →X
Σ
is a resolution of singularities whose
exceptional locus has one less nondivisorial component. Repeating this ﬁnitely
many times gives a projective resolution (star subdivisions give projective mor
phisms) whose exceptional locus is a divisor, automatically SNC by the remark
following Deﬁnition 11.2.1.
Example 11.2.3. The resolution X
Σ
1
→U
σ
constructed in Example 11.1.12 is not
SNC. Applying the process described in the proof of Theorem 11.2.2 gives the
resolution X
Σ
3
→U
σ
, which is SNC. ♦
Log Resolutions. There are several types of log resolutions. Here we discuss the
one for Qdivisors.
Deﬁnition 11.2.4. Let D =
i
d
i
D
i
be a Qdivisor on a normal variety X. Then
a projective resolution of singularities f : X
′
→ X is a log resolution of D if
i
f
−1
(D
i
) ∪Exc( f ) is a SNC divisor, where as usual Exc( f ) is the exceptional
locus of f .
For a toric variety X
Σ
, the natural choice for D is a torusinvariant Qdivisor
D =
ρ
d
ρ
D
ρ
, a
ρ
∈ Q. In this case, a projective SNC resolution X
Σ
′ →X
Σ
is au
tomatically a log resolution of (X, D) since any collection of torusinvariant prime
divisors on X
Σ
′ is SNC.
In §11.3 we will discuss a different type of log resolution, where the given data
consists of a normal variety X and a sheaf of ideals a ⊆ O
X
. See [114, 9.1.B] for
more on log resolutions.
Crepant Resolutions. Given the importance of the canonical sheaf ω
X
, it is natural
to ask how ω
X
changes in a resolution of singularities f : X
′
→X. The nicest case is
§11.2. Other Types of Resolutions 521
when X is normal and QGorenstein, i.e., ω
X
=O
X
(K
X
) for a QCartier canonical
divisor K
X
. (Note that if one canonical divisor is QCartier, then they all are.)
Suppose that ℓK
X
is Cartier, where ℓ > 0 is an integer. It follows that f
∗
(ℓK
X
)
is a Cartier divisior, and then f
∗
K
X
is deﬁned to be the Qdivisor
f
∗
K
X
=
1
ℓ
f
∗
(ℓK
X
).
Example 11.2.7 below will show that f
∗
K
X
need not be integral.
Deﬁnition 11.2.5. Let X be a normal QGorenstein variety. Then a resolution of
singularities f : X
′
→X is crepant if f
∗
(K
X
) is integral and K
X
′
∼ f
∗
(K
X
).
Since K
X
′ is only unique up to linear equivalence, the deﬁnition of crepant is
often stated as K
X
′
= f
∗
K
X
. The term “crepant” is due to Reid [148] and signals
that there is no discrepancy between K
X
′
and f
∗
K
X
.
Example 11.2.6. The toric variety U
σ
of σ = Cone(e
1
, e
2
, e
1
+e
3
, e
2
+e
3
) ⊆ R
3
is Gorenstein since the minimal generators of σ lie in the plane ¸m, u) = −1, for
m=−e
1
−e
2
. Thus the canonical divisor is K
Uσ
=div(χ
m
).
For the fan Σ
1
from Figure 3 of Example 11.1.12, the resolution X
Σ
1
→U
σ
is
crepant since
φ
∗
(K
Uσ
) =φ
∗
(div
Uσ
(χ
m
)) =div
X
Σ
(χ
m
) =K
X
Σ
.
The subscripts indicate where the divisor is being computed. You will verify this
computation in Exercise 11.2.1, where you will also show that the SNC resolution
X
Σ
3
→U
σ
from Example 11.1.12 is not crepant. ♦
Example 11.2.7. Consider σ =Cone(e
1
, e
2
, e
3
) ⊆N
R
=R
3
relative to the lattice
N =¦(a, b, c) ∈ Z
3
[ a+b+c ≡0 mod 2¦.
The minimal generators are u
i
=2e
i
, 1 ≤i ≤3, and the class group of U
σ
is Z/2Z,
where the torusinvariant divisors D
i
all map to the generator. It follows that the
canonical divisor K
Uσ
= −D
1
−D
2
−D
3
is not Cartier, but 2K
Uσ
= div(χ
m
) for
m=−e
1
−e
2
−e
3
∈ M
R
(Exercise 11.2.2).
The star subdivision Σwith respect to u
0
=e
1
+e
2
+e
3
∈N
R
gives a resolution
of singularities φ : X
Σ
→U
σ
. The four minimal generators of Σ give divisors
denoted D
0
, D
1
, D
2
, D
3
by abuse of notation. Then
φ
∗
(2K
Uσ
) =φ
∗
(div
U
σ,N
(χ
m
)) =div
X
Σ
(χ
m
)
=−3D
0
−2D
1
−2D
2
−2D
3
=2K
X
Σ
−D
0
,
so that K
X
Σ
= φ
∗
(K
Uσ
) +
1
2
D
0
(Exercise 11.2.2). Thus φ is not crepant. Note also
that φ
∗
(K
Uσ
) is not an integral divisor even though K
Uσ
is. ♦
For a toric surface, we have the minimal resolution X
Σ
′ →X
Σ
constructed in
§10.2. It is easy to characterize when this resolution is crepant.
522 Chapter 11. Toric Resolutions and Toric Singularities
Proposition 11.2.8. Let X
Σ
be a toric surface and consider the minimal resolution
of singularities φ : X
Σ
′ →X
Σ
from §10.2. Then the following are equivalent:
(a) φ is crepant.
(b) X
Σ
is Gorenstein.
(c) The singularties of X
Σ
are at worst rational double points.
Proof. The equivalence (b) ⇔ (c) follows from Proposition 10.1.6. For (a) ⇔
(c), one can reduce to the case of an afﬁne surface U
σ
with ray generators u
1
, u
2
.
Here, the key observation is that (a) is equivalent to saying that the minimal smooth
reﬁnement of σ has ray generators lying on the line segment connecting u
1
and u
2
,
u
2
u
1
σ
Figure 4. Cone σ with reﬁnement ray generators on a line segement
as illustrated in Figure 4. Using the methods of §10.2, one sees that the rational
double points are the only toric surface singularities where is happens. We leave
the details as Exercise 11.2.3
Here is an example that we have already seen.
Example 11.2.9. Let P be a reﬂexive polygon in R
2
. In Theorem 10.5.10, we
proved that the minimal resolution X
Σ
→X
P
comes from the fan given by the lattice
points on the boundary of P
◦
. Figure 5 makes it clear that X
Σ
→X
P
is crepant, a
fact we used in the proof of Theorem 10.5.10. ♦
s
s
s
s
s
s
s
s
s
s
d
d
d
d
d
d
P
◦
s
s
s
s
¨
¨
¨
¨
¡
¡
¡
¡d
d
P
e
e
e
e
r
r
r
r
Figure 5. A reﬂexive polygon P and its dual P
◦
§11.2. Other Types of Resolutions 523
More generally, crepant resolutions of toric varieties coming from reﬂexive
polytopes are important in mirror symmetry because they give rise to CalabiYau
varieties. A smooth projective variety Y is CalabiYau if ω
Y
is trivial and
(11.2.1) H
1
(Y, O
X
) = =H
d−1
(Y, O
Y
) =0, d =dimY.
Nowﬁx a reﬂexive polyope P⊆M
R
of dimension ≥2. We know from §8.3 that
X
P
is a Gorenstein Fano toric variety, so that −K
X
P
is Cartier and ample. Assume
also that φ : X
Σ
→X
P
is a crepant resolution. Then K
X
Σ
= φ
∗
(K
X
P
) is basepoint
free. By the Bertini theorem [78, Cor. III.10.9 and Ex. III.11.3], a generic section of
O
X
Σ
(−K
X
Σ
) gives a smooth connected hypersurface Y ⊆X
Σ
such that Y ∼−K
X
Σ
.
We call Y an anticanonical hypersurface.
Proposition 11.2.10. With the above hypotheses, Y is CalabiYau.
Proof. Since I
Y
= O
X
Σ
(−Y) by Proposition 4.0.28, we have I
Y
≃ O
X
Σ
(K
X
Σ
) ≃
Ω
n
X
Σ
, n =dimX
Σ
. Then the desired vanishing (11.2.1) follows from the long exact
cohomology sequence associated to
0 −→I
Y
−→O
X
Σ
−→O
Y
−→0
plus the vanishing of H
p
(X
Σ
, Ω
n
X
Σ
) for p < n from Theorem 9.3.2.
To compute the canonical bundle of Y, we use the adjunction formula from
Example 8.2.2:
ω
Y
≃ω
X
Σ
⊗
O
X
Σ
(I
Y
/I
2
Y
)
∨
≃ω
X
Σ
(Y) ⊗
O
X
Σ
O
Y
,
where we use I
Y
/I
2
Y
≃I
Y
⊗
O
X
Σ
(O
X
Σ
/I
Y
) =O
X
Σ
(−Y) ⊗
O
X
Σ
O
Y
. Since
ω
X
Σ
=O
X
Σ
(K
X
Σ
) ≃O
X
Σ
(−Y),
one sees immediately that ω
Y
=Ω
n
Y
is trivial.
Example 11.2.11. Continuing with Example 11.2.9, let P be a reﬂexive polygon
with minimal resolution X
Σ
→X
P
. In this case, a generic section of O
X
Σ
(−K
X
Σ
) is
a smooth connected curve C ⊆X
Σ
. We have two ways to compute its genus g:
(a) Recall from §10.5 that g is the sectional genus of X
P
, which equals [Int(P)∩M[
by Proposition 10.5.8. Since P is reﬂexive, the origin is its unique interior
point. Hence g =1.
(b) The canonical bundle of C is trivial by Proposition 11.2.10. The canonical
divisor has degree 0, so 0 =2g−2 by (10.5.3), hence g =1.
Thus C is an elliptic curve, which is a CalabiYau variety of dimension 1. ♦
Unfortunately, for a reﬂexive polytope of dimension ≥ 3, X
P
need not have a
crepant resolution. Here, the best one can do is the maximal projective crepant
partial desingularization described by Batyrev in [6]. The resulting anticanonical
hypersurfaces are singular CalabiYau varieties (see [35, Def. 1.4.1]).
We will study toric singularities in more detail in §11.4.
524 Chapter 11. Toric Resolutions and Toric Singularities
Embedded Resolutions. Modern treatments of the resolution of singularities use
embedded resolutions. For an embedded resolution, we have an irreducible variety
X contained in a smooth variety W, which we think of as the “ambient space.” The
idea is to repeatedly blowup the ambient space along smooth subvarieties (thereby
changing the ambient space) until the proper transform of X becomes smooth.
Here is a more careful description. An irreducible smooth subvariety Z ⊆W
gives the blowup f : Bl
Z
(W) →W, where Bl
Z
(W) is smooth and the exceptional
locus Exc( f ) (the subset of Bl
Z
(W) where f is not an isomorphism) is the divisor
given by the projective bundle of the normal bundle of Z ⊆W. We call Z the center
of the blowup. Then X ⊂ W has a proper transform in Bl
Z
(W) deﬁned as the
Zariski closure
f
−1
(X ¸Z) ⊆Bl
Z
(W).
Then one seeks a series of smooth centers Z
i
⊆W
i
, 0 ≤ i ≤ ℓ, such that W
0
=W,
W
i+1
= Bl
Z
i
(W
i
), and Z
i+1
is transverse to the exceptional locus of W
i+1
→W.
Furthermore, the ﬁnal map f
ℓ
: W
ℓ
→W needs to have the following properties:
• The induced map f [
X
ℓ
: X
ℓ
→X is a resolution of singularities, where X
ℓ
is the
proper transform of X in W
ℓ
.
• X
ℓ
is transverse to the exceptional locus of f
ℓ
.
The last bullet implies that f [
X
ℓ
: X
ℓ
→ X is a SNC resolution of singularities,
and we will see in §11.2 that f [
X
ℓ
is always a projective morphism. A careful
explanation of what this means can be found in [82], along with a discussion of
further properties one can impose on an embedded resolution.
The toric version of this, we have an equivariant embedding X ֒→W of toric
varieties, with W smooth, and at each stage of the above process, W
i
is toric, and
the center Z
i
⊆W
i
is an orbit closure. Then Z
i
is smooth, W
i+1
=Bl
Z
i
(W
i
) is toric,
and W
i+1
→W
i
is a toric morphism. When such an embedded resolution exists, the
result is a toric SNC resolution of singularities X
ℓ
→X achieved entirely through
blowups of torusinvariant centers in smooth toric ambient spaces. The papers [14],
[16] and [65] prove the existence of embedded toric resolution, and more impor
tantly, they give algorithms for ﬁnding the centers Z
0
, Z
1
, . . . needed at each stage
of the resolution process. See also [173]. In contrast, the resolution constructed in
Theorem 11.1.9 depended on many choices and is far from unique.
The proof of toric embedded resolution, though much simpler than the general
case, is still not easy. Hence we will conﬁne ourselves to some examples, together
with some remarks about nonnormal toric varieties.
Example 11.2.12. The toric variety V(xy −zw) ⊆C
4
is the afﬁne toric variety U
σ
of the cone σ from Example 11.1.12. We can resolve the singularity of U
σ
by
blowing up the origin in C
4
. The proper transform of V(xy−zw) in Bl
0
(C
4
) is the
toric variety X
Σ
3
from Example 11.1.12. There are two ways to see this.
§11.2. Other Types of Resolutions 525
First, the isomorphism U
σ
≃ V(xy −zw) induces the isomorphism (C
∗
)
3
≃
V(xy−zw) ∩(C
∗
)
4
given by (t
1
, t
2
, t
3
) →(t
1
, t
2
, t
3
, t
1
t
2
t
−1
3
) (see Example 1.1.5). At
the level of oneparameter subgroups, this gives the embedding
φ : Z
3
−→Z
4
, (a, b, c) −→(a, b, c, a+b−c).
Since σ =φ
−1
R
(R
4
≥0
) and R
4
≥0
is the cone for C
4
, the corresponding toric morphism
is the embedding φ : U
σ
→C
4
with image V(xy −zw).
Let Σ be the star subdivision of R
4
≥0
at (1, 1, 1, 1). This gives the toric variety
X
Σ
= Bl
0
(C
4
). Via φ
R
, Σ induces the reﬁnement Σ
3
of σ described in Exam
ple 11.1.12 (Exercise 11.2.4). This gives a toric morphism φ : X
Σ
3
→Bl
0
(C
4
)
whose image is the proper transform (we will say more about this below).
Second, the proper transform can studied by the method of Example 5.2.11.
The blowup has quotient representation
Bl
0
(C
4
) =(C
5
¸C¦(0, 0, 0, 0)¦)/C
∗
,
where λ (t, x, y, z, w) =(λ
−1
t, λx, λy, λz, λw), and the blowup map Bl
0
(C
4
) →C
4
is given by
(t, x, y, z, w) −→(tx, ty, tz, tw).
Via this map, 0 = xy −zw becomes 0 = (tx)(ty) −(tz)(tw) =t
2
(xy −zw), so that
the proper transform is deﬁned by xy −zw =0 in Bl
0
(C
4
). There are two cases:
• If t ,= 0, then t (t, x, y, z, w) = (1, tx, ty, tz, tw), so that this part of the proper
transform is isomorphic to V(xy −zw) ¸¦0¦ ⊆C
4
.
• If t =0, then this part of the propert transform is the ﬁber over 0 ∈ V(xy−zw)
and is the subvariety of P
3
deﬁned by xy = zw. This quadric surface is the
Segre embedding of P
1
P
1
.
One can also show that the proper transform is smooth by checking locally on the
afﬁne pieces of Bl
0
(C
4
), similar to Example 5.2.11 (Exercise 11.2.4). ♦
The general situation has an interesting wrinkle. Given a smooth toric variety
X
Σ
0
with torus T
N
0
, any subtorus T ⊆T
N
0
gives the toric variety
(11.2.2) X =T ⊆X
Σ
0
with torus T. However, X need not be normal, as shown by the simple example
where T =¦(t
2
, t
3
) [ t ∈ C
∗
¦ ⊆(C
∗
)
2
and X =T =V(x
3
−y
2
) ⊆C
2
.
For this reason, the embedded resolution of toric varieties described in [14],
[16] and [65] works with an equivariant embedding X ⊆W of toric varieties where
W is smooth but X is not assumed to be normal. These papers use cones and fans
for the ambient space W and its blowups, but use binomial equations for X and its
proper transforms, just as we did in the second half of Example 11.2.12.
526 Chapter 11. Toric Resolutions and Toric Singularities
Example 11.2.13. In Example 5.2.11 we showed that the proper transform of the
nonnormal toric variety V(x
3
−y
2
) ⊆C
2
is smooth in Bl
0
(C
2
). This not an embed
ded resolution since the proper transform is not transverse to the exceptional locus.
Two more blowups are needed to get an embedded resolution (Exercise 11.2.5). ♦
While the toric variety X from (11.2.2) need not be normal, its normalization is
easy to construct. Suppose in general that X
Σ
0
is a normal toric variety, where Σ
0
is
a fan in (N
0
)
R
. Then subtori of T
N
0
correspond bijectively to sublattices of N
0
with
torsionfree quotient (Exercise 11.2.6). Let T
N
be the subtorus coming from N ⊆N
0
and let X = T
N
⊆ X
Σ
0
be its Zariski closure. Using the subspace N
R
⊆ (N
0
)
R
, the
fan of the normalization of X is described as follows.
Proposition 11.2.14. Let X
Σ
0
be the toric variety of a fan Σ
0
in (N
0
)
R
. Fix a
subtorus T
N
⊆ T
N
0
and let X = T
N
⊆ X
Σ
0
. Then X is a (possibly nonnormal) toric
variety with torus T
N
whose normalization is the toric variety X
Σ
of the fan
Σ =¦σ ∩N
R
[ σ ∈ Σ
0
¦
in N
R
. In particular, if X is normal, then X ≃X
Σ
.
Proof. The proof is similar to the proof of Theorem 3.A.5.
Example 11.2.15. As in Example 11.2.12, consider U
σ
≃V(xy−zw) ⊆C
4
, and let
X be the proper transform of V(xy−zw) in Bl
0
(C
4
). Then X is the Zariski closure
of its torus in Bl
0
(C
4
), and we also know that X is smooth by the second half of
Example 11.2.12. Using Proposition 11.2.14, it is now easy to show that X is the
toric variety X
Σ
3
, as claimed in Example 11.2.12 (Exercise 11.2.4). ♦
Exercises for §11.2.
11.2.1. In Exercise 11.2.6,
(a) Prove that φ
∗
(div
Uσ
(χ
m
)) =div
XΣ
(χ
m
) for any m ∈ M. Hint: Use Proposition 6.1.20.
(b) Show that the resolution X
Σ3
→U
σ
from Example 11.1.12 is not crepant.
11.2.2. Fill in the details omitted in Example 11.2.7.
11.2.3. Complete the proof of Proposition 11.2.8. Hint: If σ has parameters d, k, then the
results of §10.2 give a smooth reﬁnement Σ of σ with minimal generators u
0
, . . . , u
r+1
such
that u
0
=e
2
, u
1
=e
1
, u
r+1
=de
1
−ke
2
. When are u
0
, u
1
, u
r+1
collinear?
11.2.4. This exercise concerns Example 11.2.12.
(a) In the ﬁrst half of the example, prove the claim that Σ
3
is the reﬁnement of σ induced
by the fan for Bl
0
(C
4
) via the map φ
R
.
(b) Fill in the details omitted in the second half of the example.
(c) Use parts (a) and (b) together Proposition 11.2.14 to show that X
Σ3
is isomorphic to
the proper transform of V(xy −zw) in Bl
0
(C
4
).
11.2.5. Explain why the blowup Bl
0
(C
2
) →C
2
does not give an embedded resolution of
singularities of V(x
3
−y
2
) ⊆C
2
. Then construct a blowup of Bl
0
(C
2
) that does.
§11.2. Other Types of Resolutions 527
11.2.6. Fix a lattice N
0
. Prove that there is a bijection between sublattices of N
0
with
torisionfree quotient and subtori of T
N0
.
11.2.7. Consider the complete fan Σ in R
3
consisting of the cones over the faces of the
cube with vertices (±1, ±1, ±1). In Exercise 11.1.9 you constructed a reﬁnement Σ
′
giving
a projective resolution of singularities.
(a) Explain why X
Σ
′ →X
Σ
is not SNC.
(b) Find a nice reﬁnement of Σ
′
that gives a projective SNC resolution of singularities.
11.2.8. The goal of this exercise and the next is to construct an embedded resolution of the
rational double point singularity A
k
fromExample 10.1.5. We will change notation slightly
and consider V(xz−y
k+1
) ⊆C
3
, k ≥1. The blowup Bl
0
(C
3
) has homogeneous coordinates
t, x, y, z in the sense of Chapter 5 such that Bl
0
(C
3
) →C
3
is given by (t, x, y, z) →(tx, ty, tz).
(a) Show that the proper transform of V(xz −y
k+1
) is deﬁned by xz =t
k−1
z
k+1
.
(b) Bl
0
(C
3
) is covered by three afﬁne open subsets isomorphic to C
3
. By the discussion
of local coordinates in §5.2, these afﬁne charts are given by setting one of x, y, z equal
to 1. Show that the proper transform is smooth in the x = 1 and z = 1 charts and is
deﬁned by xz =t
k−1
in the y =1 chart. Conclude that the proper transform is normal.
(c) Show that the embedded resolution of V(xz −y
k+1
) ⊆ C
3
can be done using ⌈k/2⌉
blowups at smooth points of the ambient space. Hint: Use part (b) and recursion.
11.2.9. We continue the study of V(xz −y
k+1
) ⊆ C
3
begun in Exercise 11.2.8, this time
focusing on the fan. First observe that V(xz −y
k+1
) is the Zariski closure of the subtorus
¦(t
1
, t
2
, t
−1
1
t
k+1
2
) [ t
1
, t
2
∈ C
∗
¦ ⊆(C
∗
)
3
corresponding to the sublattice
N =Z(1, 0, −1) +Z(0, 1, k +1) ⊆Z
3
.
Since V(xz −y
k+1
) is normal (Example 10.1.5), Proposition 11.2.14 implies that it is the
afﬁne toric variety of the cone σ =N
R
∩R
3
≥0
since R
3
≥0
is the cone for C
3
.
(a) Show that σ =Cone((k +1, 1, 0), (0, 1, k +1)).
(b) Bl
0
(C
3
) is the toric variety of the star subdivision of R
3
≥0
at ν =(1, 1, 1). This induces
a reﬁnement Σ of σ, shown for k = 2 in Figure 6 on the next page. Prove that for
k ≥1, Σ is the fan of the proper transform. Hint: Part (b) of the previous exercise.
(c) When k = 1, show that Σ is smooth with exactly two 2dimensional cones. Draw a
picture similar to Figure 4 and explain why it is simpler.
(d) When k ≥ 2, show that the fan Σ has three 2dimensional cones, where the outer two
cones are smooth and the middle cone gives A
k−1
.
The methods of §10.2 resolve U
σ
by subdividing σ into k +1 smooth cones (see Exer
cise 10.2.2). When k ≥ 2, the fan Σ constructed in part (b) has the outermost cones of the
smooth reﬁnement plus a middle cone that combines the remaining cones of the smooth
reﬁnement. When k ≥3, blowing up Bl
0
(C
3
) reﬁnes Σ by adding two more cones fromthe
smooth reﬁnement, and each successive blowup adds in two more cones from the smooth
reﬁnement until the process is complete.
528 Chapter 11. Toric Resolutions and Toric Singularities
← σ ← induced
refinement of σ
Figure 6. σ for k =2 and the reﬁnement induced by blowing up 0 ∈ C
3
§11.3. Rees Algebras and Multiplier Ideals
So far, we have constructed resolution of singularities using repeated blowups. By
using the blowup of a sheaf of ideals, it is often possible to construct the resolu
tion in a single blowup. From an abstract point of view, this is a standard fact.
Given any birational projective morphism f : X →Y of irreducible varieties with Y
quasiprojective, there is an ideal sheaf I ⊆ O
Y
such that f : X →Y is the blowup
map π : Bl
I
(Y) →Y. This is proved in [78, Thm. II.7.17]. In the toric case, we
will see that the blowup has an especially nice interpretation.
Blowups and Polyhedra. We begin with a full dimensional lattice polyhedron P
in M
R
. Thus P = Q+C, where Q is a lattice polytope and C is strongly convex.
In order to get a blowup, we assume in addition that C is full dimensional, so that
σ =C
∨
⊆ N
R
is also full dimensional and strongly convex. Then P =Q+σ
∨
and
the normal fan Σ
P
is a reﬁnement of σ. Hence we have a birational toric morphism
φ : X
P
−→U
σ
,
which is projective by Theorem 7.1.10. Our goal is to represent this morphism as
the blowup of an ideal a ⊆C[σ
∨
∩M].
Translating P by an element of M has no effect on X
P
. For m ∈ Int(σ
∨
) ∩M,
one sees easily that P+λm ⊆ σ
∨
for integers λ ≫0. Hence we may assume that
P ⊆σ
∨
. This gives the ideal
(11.3.1) a =
m∈P∩M
C χ
m
⊆R =C[σ
∨
∩M] =
m∈σ
∨
∩M
C χ
m
,
and from a ⊆R, we get the Rees algebra
R[a] =
∞
k=0
a
k
t
k
,
§11.3. Rees Algebras and Multiplier Ideals 529
where t is a dummy variable that keeps track of the grading. This graded Ralgebra
is generated by its elements of degree 1 as an Ralgebra and satisﬁes R[a]
0
=R.
Using the Proj construction described in the appendix to Chapter 7, we get a
projective morphism
(11.3.2) X =Proj(R[a]) −→Spec(R) =U
σ
.
This is the blowup of Spec(R) with respect to a, denoted Bl
a
(Spec(R)). For us, a
key properties of this blowup involves the ideal sheaf aO
X
, deﬁned as follows. On
an afﬁne open subset U =Spec(A) ⊆X, the map U →Spec(R) comes from a ring
homomorphism R →A, and aO
U
is the sheaf associated to the ideal aA ⊆ A. For
the blowup morphism (11.3.2) the sheaf aO
X
is a line bundle—this is [78, Prop.
II.7.13]. See [78, II.7] for a discussion of blowing up ideals and ideal sheaves.
Besides the Rees algebra, the polyhedron P gives a second graded algebra.
Recall from §7.1 that from P ⊆ M
R
we get the cone C(P) ⊆ M
R
R, where the
slice at height k > 0 is kP and the slice at height 0 is σ
∨
(see Figure 1 from §7.1).
The associated semigroup algebra S
P
= C[C(P) ∩(MZ)] is graded by height,
and we showed in the proof of Theorem 7.A.1 that (S
P
)
0
=R.
Proposition 11.3.1. With the above setup, there is an inclusion R[a] ֒→S
P
that
makes S
P
the integral closure of R[a]. Furthermore, this inclusion is an equality if
and only P is normal.
Proof. First observe that R[a] is a semigroup algebra. Let
A =
_
(σ
∨
∩M) ¦0¦
_
∪
_
(P∩M) ¦1¦
_
⊆MZ.
We leave it to reader to show that R[a] =C[NA] (Exercise 11.3.1). Also note:
• We have ZA =MZ since σ
∨
is full dimensional and P ⊆σ
∨
.
• Cone(A) =C(P).
Then Proposition 1.3.8 implies that S
P
=C[C(P) ∩(MZ)] the integral closure of
R[a] =C[NA].
Since R[a] and S
P
are equal in degrees ≤1, the ﬁnal assertion of the proposition
is an easy consequence of the deﬁnition of normal polyhedron.
To translate Proposition 11.3.1 into geometric terms, we will use the Cartier
divisor D
P
of the polyhedron P. Recall that the Cartier data of D
P
is given by the
vertices v ∈ P. This means that the support function of D
P
satisﬁes ϕ(u) = ¸v, u)
when u lies in the cone σ
v
∈Σ
P
corresponding to v. Thus ϕ(u) ≥0 since v ∈P⊆σ
∨
and u ∈ σ
v
⊆ σ. It follows that D =
ρ
ϕ(u
ρ
)D
ρ
is an effective divisor. Since
D
P
=−
ρ
ϕ(u
ρ
)D
ρ
, we conclude that D
P
=−D, where D ≥0.
In particular, O
X
P
(D
P
) = O
X
P
(−D) is an ideal sheaf of O
X
P
. This compares to
aO
X
P
as follows.
530 Chapter 11. Toric Resolutions and Toric Singularities
Corollary 11.3.2. With the same hypotheses as Proposition 11.3.1, we have:
(a) The map X
P
→U
σ
factors
X
P
−→Bl
a
(U
σ
) −→U
σ
,
where the ﬁrst morphism is normalization and the second is the blowup of a.
(b) aO
X
P
=O
X
P
(D
P
), where D
P
is the Cartier divisor associated to P
Proof. For part (a), recall from Theorem 7.A.1 that X
P
=Proj(S
P
), where S
P
is the
integral closure of R[a] by Proposition 11.3.1. You will prove in Exercise 11.3.2
that the inclusion R[a] ֒→S
P
induces the normalization map
X
P
=Proj(S
P
) −→Proj(R[a]).
For part (b), take a vertex v ∈ P. This gives the cone σ
v
∈ Σ
P
, and since
P =Conv(m [ χ
m
∈ a), one sees easily that σ
∨
v
=Cone(m−v [ χ
m
∈ a). Then any
χ
m
∈ a can be written
χ
m
=χ
v
χ
m−v
∈ χ
v
C[σ
∨
v
∩M],
from which we conclude that aC[σ
∨
v
∩M] =χ
v
C[σ
∨
v
∩M] since χ
v
∈ a. Since X
P
is covered by the afﬁne open subsets U
σv
, it follows that aC[σ
∨
v
∩M] is the line
bundle of the Cartier divisor whose Cartier data is given by the vertices of P. As
noted above, this divisor is D
P
, which completes the proof.
This corollary shows that the map X
P
→U
σ
is a single blowup followed by a
normalization. Moreover, replacing P with a sufﬁciently high multiple, we may
assume that P is normal (Theorem 7.1.9). This does not change the normal fan, so
that in this case, the map X
P
→U
σ
is the blowup of the ideal a.
By Theorem 7.2.4, it follows that any projective toric resolution X
Σ
→U
σ
is
the blowup of a suitable ideal a ⊆ C[σ
∨
∩M]. For the special case of blowing up
the origin 0 ∈ C
n
or more generally a coordinate subspace ¦0¦C
n−r
⊆ C
n
, the
ideal a is easy to describe (Exercise 11.3.3).
More on the Ideal. If we ﬁx the cone σ ⊆N
R
, then we can give a purely algebraic
description of the ideals that arise from (11.3.1). We ﬁrst need a deﬁnition.
Deﬁnition 11.3.3. An ideal a in a ring R is integrally closed if whenever an element
r ∈ R satisﬁes an equation
r
k
+a
1
r
k−1
+ +a
k−1
r +a
k
=0
with a
i
∈ a
i
for 1 ≤i ≤k, we have r ∈ a.
Then we have the following result.
§11.3. Rees Algebras and Multiplier Ideals 531
Proposition 11.3.4. Fix a full dimensional strongly convex cone σ ⊆N
R
. Then the
maps
P −→ a =
m∈P∩M
C χ
m
a −→P =Conv(m [ χ
m
∈ a)
induce a bijection
_
lattice polyhedra P ⊆σ
∨
with recession cone σ
∨
_
←→
_
_
_
nonzero integrally closed
ideals a ⊆C[σ
∨
∩M]
generated by characters
_
_
_
.
Proof. Given a lattice polyhedron P ⊆ σ
∨
with recession cone σ
∨
, we need to
show that a is integrally closed. Standard arguments (see [48, p. 137] or [172, Cor.
1.3.1]) imply that
a =¦r ∈ C[σ
∨
∩M] [ r
k
+a
1
r
k−1
+ +a
k
=0 for some a
i
∈ a
i
¦
is an integrally closed ideal containing a. Hence it sufﬁces to prove a =a.
Since a is generated by characters, it is invariant under the T
N
action on R,
which easily implies that a is T
N
invariant. Then Lemma 1.1.16 implies that a is
also generated by characters (see also [172, Prop. 1.4.2]). Thus, to prove a =a, we
need to show that if a character χ
m
∈ a satisﬁes
χ
km
+a
1
χ
(k−1)m
+ +a
k−1
χ
m
+a
k
=0
with a
i
∈ a
i
, then χ
m
∈ a. Solving for the ﬁrst term of the above equation gives
χ
km
=−a
1
χ
(k−1)m
− −a
k−1
χ
m
−a
k
.
Writing a
i
as a linear combination of ifold products of characters in a, it follows
that for some 1 ≤i ≤k, we must have an equality
χ
km
=χ
m
1
χ
m
i
χ
(k−i)m
with χ
m
1
, . . . , χ
m
i
∈ a, i.e., m
1
, . . . , m
i
∈ P. This implies
m=
1
i
(m
1
+ +m
i
),
which lies in P since P is convex. Hence χ
m
∈a, proving that a is integrally closed.
In the other direction, a nonzero ideal a = ¸χ
m
1
, . . . , χ
ms
) ⊆ C[σ
∨
∩M] gives
the convex set P = Conv(m [ χ
m
∈ I) ⊆ σ
∨
. In Exercise 7.1.10 you showed that
P = Conv(m
1
, . . . , m
s
) +σ
∨
. Hence P is a lattice polyhedron contained in I with
recession cone σ
∨
.
It remains to show that these maps are inverses of each other. One direction is
easy, for if P→a, then a →Conv(m[ m∈P∩M), which equals P since Pis a lattice
polyhedron. The other direction takes more thought. A nonzero integrally closed
ideal a gives P = Conv(m [ χ
m
∈ a). We need to show that m ∈ P∩M implies
χ
m
∈ a. To prove this, ﬁrst note that m ∈ P∩M implies m ∈ Conv(m
1
, . . . , m
s
)
where χ
m
1
, . . . , χ
ms
∈ a. By Carath´ eodory’s theorem (see [179, Prop. 1.15]), we
532 Chapter 11. Toric Resolutions and Toric Singularities
can assume that m
1
, . . . , m
s
are afﬁnely independent. Since m, m
1
, . . . , m
s
∈ M, it
follows that
m =
s
i=1
λ
i
m
i
, where λ
i
≥0,
s
i=1
λ
i
=1, λ
i
∈ Q.
Alfter clearing denominators, we get km =
s
i=1
µ
i
m
i
where k, µ
i
, . . . , µ
s
∈ N and
s
i=1
µ
i
=k > 0. Then r =χ
m
satisﬁes the equation
r
k
−
s
i=1
(χ
m
i
)
µ
i
=0.
Since
s
i=1
(χ
m
i
)
µ
i
∈ a
k
and a is integrally closed, we must have r =χ
m
∈ a.
Example 11.3.5. For the cone σ ⊆ R
3
of Example 11.1.12, Figure 7 shows the
polyhedron P⊆σ
∨
obtained by pushing one of the facets of σ
∨
one unit to the right.
translate
this facet →
one unit
to the
right
P
σ
Figure 7. The polyhedron P and the dual cone σ
∨
The corresponding ideal a is generated by the vertices of P (a rare occurrence, but
true here), so that
a =¸χ
e
2
, χ
e
1
+e
2
−e
3
) ⊆C[σ
∨
∩M].
Looking at P, one sees that its normal fan is the fan Σ
2
from Figure 3 in Exam
ple 11.1.12. Since P is normal (Exercise 11.3.4), it follows that X
Σ
2
=Bl
a
(U
σ
).
We can also think about this from the point of view of the simplicialization
process described in the proof of Proposition 11.1.7. Since σ is not simplicial,
we can take ρ ∈ σ(1) such that D
ρ
is not Cartier, say ρ = Cone(e
2
). The star
subdivision at e
2
gives the fan Σ
2
such that D
ρ
is becomes Cartier on X
Σ
2
. Then
one can check in that P is the polyhedron of −D
ρ
, so that
aO
X
Σ
2
=O
X
Σ
2
(−D
ρ
)
(Exercise 11.3.4). ♦
§11.3. Rees Algebras and Multiplier Ideals 533
Example 11.3.6. The maximal ideal m =
m∈σ
∨
∩(M\¡0¦)
C χ
m
⊆ C[σ
∨
∩M] is
clearly integrally closed. The polyhedron associated to m is the convex hull P
m
=
Conv(σ
∨
∩(M¸¦0¦)).
When σ has dimension 2, this convex hull was denoted Θ
σ
∨ in §10.2, i.e.,
P
m
= Θ
σ
∨. In Proposition 10.2.17, we described the properties of the associated
projective toric morphism X
Pm
→U
σ
. Since lattice polyhedra are normal in dimen
sion 2, Proposition 11.3.1 and Corollary 11.3.2 imply that this morphism is the
blow up of the ideal m, i.e., X
Pm
=Bl
m
(U
σ
). ♦
Example 11.3.7. Consider the graph G in Figure 8, where the vertices are labeled
with variables x, y, z. This is gives the edge ideal a =¸xy, yz, zx) ⊆C[x, y, z] whose
generators correspond to the edges of the graph. Note that a is radical since it
Graph G
x
z
y
⇒
Ideal
⇓
〈
xy, yz, zx
〉
Fan Σ
e
3
e
1
e
2
x
z
y
Polyhedron P
⇐
a
Figure 8. Graph, ideal, polyhedron, and fan
is generated by squarefree monomials and hence is integrally closed since radical
ideals are automatically integrally closed (Exercise 11.3.5).
The ideal a gives the polyhedron P shown in Figure 8. Since P sits inside the
ﬁrst octant in R
3
, the shaded facets form the “back” of P from our point of view.
The vertices of P correspond to the generators of a, though in general, the ideal of
a polyhedron usually has more generators than just the characters of the vertices.
534 Chapter 11. Toric Resolutions and Toric Singularities
One can show that P is normal (Exercise 11.3.5), so that
X
Σ
=X
P
=Bl
a
(C
3
) =Proj(R[a]), R =C[x, y, z],
by Proposition 11.3.1 and Corollary 11.3.2. Another proof of normality comes
from [160, Thm. 6.3], which implies that R[a] is normal since G is an odd cycle.
Hence P is normal Proposition 11.3.1. ♦
When a ⊆ C[σ
∨
∩M] is generated by characters but not integrally closed, we
have the following description of its integral closure.
Proposition 11.3.8. Let a ⊆ C[σ
∨
∩M] be an ideal generated by characters and
set P =Conv(m [ χ
m
∈ a). Then:
(a) a =
m∈P∩M
C χ
m
, so P ↔a in the correspondence of Proposition 11.3.4.
(b) aO
X
P
=aO
X
P
=O
X
P
(D
P
).
Proof. We leave the proof of part (a) as Exercise 11.3.6. For part (b), applying
Corollary 11.3.2 to P and a gives aO
X
P
= O
X
P
(D
P
). Furthermore, the proof of the
corollary shows that
aC[σ
∨
v
∩M] =χ
v
C[σ
∨
v
∩M]
when v is a vertex of P. However, P =Conv(m [ χ
m
∈ a) implies that the vertices
come from a, i.e., χ
v
∈ a. Hence aC[σ
∨
v
∩M] =aC[σ
∨
v
∩M], and part (b) follows
immediately.
Part (a) of the proposition is wellknown when a ⊆C[x
1
, . . . , x
n
] is a monomial
ideal. See, for example, [172, Prop. 1.4.6]. Another nice result of “monomial
commutative algebra” is the Monomial Brianc¸onSkoda Theorem, which asserts
that if a ⊆C[σ
∨
∩M] is an ideal generated by characters, then
a
n+ℓ−1
⊆a
ℓ
where n =rankM and ℓ ≥1 is arbitrary. A proof can be found in [173, §4]. Mono
mial ideals are an active area of research in commutative algebra (see [176] and the
references therein).
Global Aspect. Given a projective toric resolution of singularities X
Σ
′
→X
Σ
, it is
natural to ask if X
Σ
′ can be represented as the blowup Bl
I
(X
Σ
) for some torus
invariant sheaf of ideals I ⊆ O
X
Σ
. When X
Σ
is quasiprojective, this can be done
by adapting the technique used in the proof of [78, Thm. II.7.17]. In general, it
might not be possible to ﬁnd a sheaf of ideals, but one can always ﬁnd a sheaf of
fractional ideals that works. This is explained in [104, Ch. I, Thms. 10 and 11].
§11.3. Rees Algebras and Multiplier Ideals 535
Log Resolutions. Suppose that we have a sheaf of ideals I ⊆ O
X
, where X is
smooth. Then IO
Y
is a line bundle on Y = Bl
I
(Y), though Y may be rather
singular. A log resolution for I asks for a smooth version of this. Here is the
precise deﬁnition.
Deﬁnition 11.3.9. Let I ⊆ O
X
be an ideal sheaf on a smooth variety X. Then a
projective birational morphism f : X
′
→X is a log resolution of I if X
′
is smooth,
IO
X
′ =O
X
′ (−D) for a divisor D on X
′
, and Supp(D) ∪Exc( f ) is a SNC divisor.
This type of log resolution is often called a principalization. In the toric case,
we have the following principalization result of Howald [90] (see also [173]).
Theorem 11.3.10. Let a ⊆C[x
1
, . . . , x
n
] be a monomial ideal. Then there is a toric
morphism φ : X
Σ
→C
n
that is a log resolution of a.
Proof. Let P = Conv(m [ χ
m
∈ a) as in Proposition 11.3.8. This gives X
P
→C
n
with aO
X
P
= O
X
P
(D
P
). Since X
P
may be singular, we take a projective resolution
of singularities X
Σ
→X
P
. The composed map φ : X
P
→C
n
is projective, toric, and
birational, though its exceptional locus Exc(φ) need not be a divisor. This is easily
ﬁxed by adapting the strategy used in the proof of Theorem 11.2.2. Hence we may
assume that Exc(φ) is a divisor.
Then let −D be the pullback of D
P
via X
Σ
→X
P
. One easily checks that
aO
X
Σ
= O
X
Σ
(−D), and Supp(−D) ∪Exc(φ) is SNC since it is a union of torus
invariant prime divisors on a smooth toric variety.
A stronger principalization result, proved by Goward in [69], asserts that the
log resolution φ : X
Σ
→C
n
can be chosen to be a composition of blowups with
smooth torusinvariant centers. This is analogous to embedded resolution of a
toric variety, which also uses blowups with smooth torusinvariant centers. See
Exercise 11.3.7 for an example.
Multiplier Ideals. Given a proper birational morphism f : X
′
→X where X
′
and
X are smooth, we have the relative canonical divisor
K
X
′
/X
=K
X
′ − f
∗
K
X
,
As noted in [114, 9.1.B], this divisor is supported on the exceptional locus Exc( f )
and hence satisﬁes
(11.3.3) f
∗
O
X
′ (K
X
′
/X
) ≃O
X
.
In the toric context, these assertions are easy to understand and are covered in
Exercise 11.3.8.
One consequence of (11.3.3) is that if E is any effective divisor on X
′
, then
O
X
′ (K
X
′
/X
−E) ⊆O
X
′ (K
X
′
/X
),
536 Chapter 11. Toric Resolutions and Toric Singularities
so that f
∗
O
X
′ (K
X
′
/X
−E) ⊆ f
∗
O
X
′ (K
X
′
/X
) ≃ O
X
is an ideal sheaf. When applied
to the log resolution of an ideal sheaf, this gives the following deﬁnition.
Deﬁnition 11.3.11. Let I ⊆ O
X
be a nonzero ideal sheaf on a smooth variety X
and let f : X
′
→X be a log resolution of I such that IO
X
′
= O
X
′
(−D) for an
effective divisor D on X
′
. Given any rational number c > 0, the multiplier ideal
¸(c I) is the ideal sheaf deﬁned by
¸(c I) = f
∗
O
X
′
(K
X
′
/X
−⌊cD⌋).
One can show that this deﬁnition is independent of the log resolution [114,
Thm. 9.2.18]. See [18] for an introduction to multiplier ideals and [114, Ch. 9–10]
for a thorough discussion and many applications.
In the toric context, Howald [90] showed how to compute the multiplier ideal
¸(c a) of a monomial ideal a ⊆C[x
1
, . . . , x
n
].
Theorem 11.3.12. Let P = Conv(m [ χ
m
∈ a) be the polyhedron associated to a
monomial ideal a ⊆C[x
1
, . . . , x
n
]. Then, for any c > 0 in Q, we have
¸(c a) =
m+e
0
∈Int(cP)
C χ
m
.
were e
0
=(1, . . . , 1) ∈ Z
n
.
Proof. We use the log resolution φ : X
Σ
→C
n
constructed in Theorem 11.3.10.
Thus aO
X
Σ
= O
X
Σ
(−D), where −D is the pullback of D
P
on X
P
, and it follows
that the polyhedron of −D is P
−D
=P. If we write D =
ρ
a
ρ
D
ρ
, a
ρ
≥0, then
P =¦m ∈ M
R
[ ¸m, u
ρ
) ≥a
ρ
for all ρ ∈ Σ(1)¦,
hence
cP =¦m ∈ M
R
[ ¸m, u
ρ
) ≥ca
ρ
for all ρ ∈ Σ(1)¦. (11.3.4)
Also note that div
C
n
(χ
−e
0
) =K
C
n
, so that φ
∗
K
C
n
=div
X
Σ
(χ
−e
0
). Hence
K
X
Σ
/C
n −⌊cD⌋ =K
X
Σ
−φ
∗
K
C
n
−⌊cD⌋ =
ρ
(−1+¸e
0
, u
ρ
) −⌊ca
ρ
⌋)D
ρ
.
The multiplier ideal ¸(c a) is given by the global sections of the sheaf of this
divisor. By Proposition 4.3.3, we obtain
χ
m
∈ ¸(c a) ⇐⇒ ¸m, u
ρ
) ≥1−¸e
0
, u
ρ
) +⌊ca
ρ
⌋ for all ρ ∈ Σ(1)
⇐⇒ ¸m+e
0
, u
ρ
) ≥1+⌊ca
ρ
⌋ for all ρ ∈ Σ(1)
⇐⇒ ¸m+e
0
, u
ρ
) > ca
ρ
for all ρ ∈ Σ(1).
By (11.3.4), we conclude that χ
m
∈ ¸(c a) if and only if m+e
0
lies in the interior
of cP. This completes the proof.
§11.3. Rees Algebras and Multiplier Ideals 537
Theorem 11.3.12 shows that the characters in the multiplier ideal ¸(c a) come
from translates by −e
0
=(−1, . . . , −1) of interior points of c times the polyhedron.
This result is also discussed in [114, Thm. 9.3.27].
Example 11.3.13. Consider the monomial ideal a = ¸x
4
, x
2
y, y
3
) ⊆ C[x, y]. The
multiplier ideal ¸(1 a) =¸(a) is given by
¸(a) =¸x
2
, xy, y
2
).
This follows from Figure 9, which shows the polyhedron P (light shading), its
P
←interior
points of P
←translated interior
points of P
Figure 9. The polyhedron P, its interior lattice points, and their translates
interior lattice points (in the dark shading), and the result of translating these lattice
points by (−1, −1) (bounded by the dashed line). ♦
Example 11.3.14. For the ideal a =¸xy, yz, zx) ⊆C[x, y, z] of Example 11.3.7, one
sees easily that (1, 1, 1) lies in the interior of P. It follows immediately that the
multiplier ideal ¸(a) is trivial. ♦
Exercises for §11.3.
11.3.1. Prove the claim made in the proof of Proposition 11.3.1 that R[a] =C[NA].
11.3.2. In Proposition 11.3.1 we showed that S
P
=C[C(P) ∩(MZ)] the integral closure
of R[a] =C[NA]. Prove that Proj(S
P
) →Proj(R[a]) is the normalization map. Hint: Look
at the proof of Theorem 7.A.1. A character χ
m
∈ a gives an element χ
m
t of degree 1 in
R[a] and S
P
, which in turn gives open sets
U =Spec(R[a]
(χ
m
t)
) and U
′
=Spec((S
P
)
(χ
m
t)
)
of Proj(R[a]) and Proj(S
P
) respectively. Show that these open sets cover Proj(R[a]) and
Proj(S
P
) and that U
′
is the normalization of U.
11.3.3. For the blowup of the origin 0 ∈C
n
, ﬁnd an explicit ideal a ⊆R=C[x
1
, . . . , x
n
] such
that Bl
0
(C
n
) = Proj(R[a]). Then do the same for the blowup of the coordinate subspace
¦0¦ C
n−r
⊆C
n
.
11.3.4. Fill in the details omitted in the discussion of Example 11.3.5.
11.3.5. This exercise is concerned with Example 11.3.7.
538 Chapter 11. Toric Resolutions and Toric Singularities
(a) Prove that an ideal in C[x
1
, . . . , x
n
] generated by squarefree monomials is radical.
(b) Prove that every radical ideal is integrally closed.
(c) Prove that the polytope P pictured in Figure 8 of Example 11.3.7 is normal.
11.3.6. Prove Proposition 11.3.8.
11.3.7. Let a =¸x
3
, y
2
) ⊆C[x, y].
(a) Construct a log resolution φ : X
Σ
→C
2
of a using the method described in the proof
of Theorem 11.3.10.
(b) Interpret the fan X
Σ
constructed in part (a) as a sequence of three star subdivisions,
each of which blows up a ﬁxed point of the torus action.
(c) Let E
i
be the proper transform in X
Σ
exceptional locus of the ith blowup. Show that
aO
XΣ
=O
XΣ
(−2E
1
−3E
2
−6E
3
).
This exercise is taken from [18].
11.3.8. Let φ : X
Σ
′ →X
Σ
be the toric morphism coming from a smooth reﬁnement Σ
′
of a
smooth fan Σ.
(a) Prove that K
X
Σ
′
−φ
∗
K
XΣ
is supported on the exceptional locus Exc(φ) and is ≥0.
(b) Let D be an effective divisor supported on Exc(φ). Prove that φ
∗
O
X
′
Σ
(D) ≃O
XΣ
. Hint:
Reduce to the afﬁne case.
11.3.9. Compute the multiplier ideal ¸(a) of the following monomial ideals:
(a) a =¸x
8
, y
6
) ⊆C[x, y] (see [90, Ex. 2]).
(b) a =¸x
7
, x
3
y, xy
2
, y
6
) ⊆C[x, y] (see [114, Ex. 9.3.28]).
11.3.10. Let a =¸x
a1
1
, . . . , x
an
n
) ⊆C[x
1
, . . . , x
n
], where a
1
, . . . , a
n
are positive integers. Prove
that ¸(c a) is trivial if and only if c <
n
i=1
1
ai
(see [90, Ex. 3]).
11.3.11. Consider the monomial ideal a = ¸xy, yz, zx) ⊆ C[x, y, z] and polyhedron P from
Figure 8 in Example 11.3.7. The toric morphism X
P
→C
3
satisﬁes aO
XP
= O
XP
(D
P
). To
get a log resolution of a as in Theorem11.3.10, we need a smooth reﬁnement of the normal
fan Σ of P. Figure 10 shows Σ and two smooth reﬁnements Σ
1
and Σ
2
.
Σ
e
3
e
1
e
2
Σ
1
e
3
e
1
e
2
Σ
2
e
3
e
1
e
2
Figure 10. The fan Σ and smooth reﬁnements Σ1 and Σ2
(a) Explain howΣ
1
and Σ
2
can be obtained fromΣ by a series of star subdivisions. Hence
X
Σ1
→C
3
and X
Σ2
→C
3
are birational projective toric morphisms.
§11.3. Rees Algebras and Multiplier Ideals 539
(b) Show that X
Σ1
→C
3
is a composition of blowups at smooth torusinvariant centers.
(c) Showthat X
Σ2
→C
3
is not a composition of blowups at smooth torusinvariant centers.
11.3.12. In this exercise and the next we follow [17] and explore the toric case of two
variants of multiplier ideals. Let U
σ
= Spec(C[σ
∨
∩M]) be an afﬁne toric variety. Fix an
effective torusinvariant Qdivisor D and an ideal a ⊂C[σ
∨
∩M] generated by characters.
Also assume that K
Uσ
+D is QCartier. Thus there are m
0
∈ M and r > 0 in Z with
div(χ
−m0
) =r(K
Uσ
+D). Also let P ⊆σ
∨
be the polyhedron of a.
(a) Prove that there is a toric morphism φ : X
Σ
→U
σ
which is a log resolution for both D
and a. Write aO
XΣ
=O
XΣ
(−A), where A is an effective divisor.
(b) Given c > 0 in Q, prove that we have a natural inclusion
φ
∗
O
XΣ
(K
XΣ
−⌊φ
∗
(K
Uσ
+D) +cA⌋) ⊆O
Uσ
.
This gives a multiplier ideal ¸(D, c a) ⊆C[σ
∨
∩M].
(c) Adapt the proof of Theorem 11.3.12 to show that
¸(D, c a) =
m+
1
r
m0∈Int(cP)
C χ
m
.
Part (c) is [17, Thm. 1]. Note that Theorem 11.3.12 is a special case of this result. Earlier,
we deﬁned the multiplier ideal ¸(c a) for C
n
. This deﬁnition extends to QGorenstein
afﬁne toric varieties. However, when U
σ
fails to be QGorenstein, then we use the log
version, since K
Uσ
+D could be QCartier for some Qdivisor D. See [114, 9.3.G].
11.3.13. To deﬁne a multiplier ideal that works for any afﬁne toric variety U
σ
, we proceed
as follows. Let a ⊂ C[σ
∨
∩M] be generated by characters and let φ : X
Σ
→U
σ
be a log
resolution such that aO
XΣ
=O
XΣ
(−A), A ≥0. Following [17], the multiplier module is
φ
∗
O
XΣ
(K
XΣ
−⌊cA⌋) ⊆ω
Uσ
, c > 0 in Q.
By Proposition 8.2.9, we can regard ω
Uσ
as an ideal of C[σ
∨
∩M] generated by characters
coming from lattice points in the interior of σ
∨
. Then the multiplier module gives an ideal
of C[σ
∨
∩M] denoted ¸
ω
(c a). Prove that this ideal is given by
¸
ω
(c a) =
m∈Int(cP)
C χ
m
,
where as usual P is the polyhedron of a. This is [17, Thm. 2].
11.3.14. Let a ⊆R=C[x
1
, . . . , x
n
] be a monomial ideal. In the text, we used a log resolution
to deﬁne the multiplier ideal ¸(c a). This exercise will present a different construction of
¸(c a). Let S be the integral closure of the Rees algebra R[a]. This gives the projective
morphism φ : X = Proj(S) →C
n
. Prove that
¸(c a) =φ
∗
O
X
(K
X
−φ
∗
K
C
n −⌊cD⌋),
where aO
X
= O
X
(−D). Hint: Use Proposition 11.3.1. (For experts, this works because X
has rational singualarities.)
11.3.15. This exercise is for readers who know about symbolic powers of ideal sheaves
(see, for example, [114, Def. 9.3.4]). Let D=
i
D
i
be a sum of distinct prime divisors on
a normal variety X. Then O
X
(−D) ⊆O
X
is a sheaf of radical ideals. For an integer n ≥1,
show that O
X
(−nD) equals the nth symbolic power O
X
(−D)
n
. Hint: Exercise 4.0.14.
540 Chapter 11. Toric Resolutions and Toric Singularities
§11.4. Toric Singularities
We next discuss the singularities of normal toric varieties. This is an active area of
research, and our treatment will omit many important topics.
General Properties of Toric Singularities. Although a normal toric variety need
not be smooth, it is at least CohenMacaulay. Furthermore, its singularities are of
the following special type.
Deﬁnition 11.4.1. A normal variety X has rational singularities if for every reso
lution of singularities f : X
′
→X, we have
R
p
f
∗
O
X
′ =0 for all p > 0.
One can show that if the above vanishing holds for one resolution of singular
ities, then it holds for all [108, Thm. 5.10]. Hence, in the toric case, we can use a
toric resolution of singularities.
Theorem 11.4.2. A normal toric variety X
Σ
has rational singularities.
Proof. Let φ : X
Σ
′ →X
Σ
be a toric resolution of singularites. In particular, φ is
proper, so that R
p
φ
∗
O
X
Σ
′
= 0 for p > 0 by Theorem 9.2.5. Then we are done by
the above remark.
The reader should consult [108, Sec. 5.1] for a careful discussion of rational
singularities.
The Simplicial Case. We next give several characterizations of the singularities of
a simplicial toric variety. The ﬁrst is topological, based on the following idea. If X
is a smooth irreducible variety of dimension n, then any x ∈ X has a neighborhood
x ∈U in the classical topology that is homeomorphic to an open ball in C
n
≃R
2n
.
By excision and the long exact sequence for relative cohomology, one obtains
H
p
(X, X ¸¦x¦, Z) ≃H
p
(U,U ¸¦x¦, Z) ≃
_
0 p ,=2n
Z p =2n.
When we weaken the coefﬁcients from Z to Q, we get the following deﬁnition.
Deﬁnition 11.4.3. An irreducible variety X of dimension n is rationally smooth if
for every x ∈ X, we have
H
p
(X, X ¸¦x¦, Q) ≃
_
0 p ,=2n
Q p =2n.
Smooth varieties are obviously rationally smooth. Here is a example that will
be useful below.
Example 11.4.4. If G⊆GL(n, C) is a ﬁnite group, then C
n
/G is rationally smooth
by [24, Prop. A.1]. ♦
§11.4. Toric Singularities 541
Before giving our second characterization, we need some terminology. The
structure sheaf O
X
of a variety X is a sheaf in the Zariski topology. When we
switch to the classical topology, the corresponding sheaf is the sheaf of analytic
functions on X, denoted O
an
X
(see [78, App. B] for a brief description). Then any
classical open subset U ⊆X gives the analytic variety (U, O
an
U
) =(U, O
an
X
[
U
), and
varieties X
1
and X
2
are locally analytically equivalent at p
1
∈X
1
and p
2
∈X
2
if there
are classical neighborhoods p
1
∈U
1
⊆ X
1
and p
2
∈U
2
⊆ X
2
such that U
1
≃U
2
as
analytic varieties, where the isomorphism take p
1
to p
2
.
Deﬁnition 11.4.5. Let X be an irreducible variety of dimension n.
(a) A point p ∈ X is a ﬁnite quotient singularity if there is a ﬁnite subgroup G ⊆
GL(n, C) such that p ∈ X is locally analytically equivalent to 0 ∈ C
n
/G.
(b) X is an orbifold or is quasismooth or has ﬁnite quotient singularities if every
point of X is a ﬁnite quotient singularity.
(c) X has abelian ﬁnite quotient singularities if every point of X is a ﬁnite quotient
singularity such that the ﬁnite group G in part (a) is abelian.
Here is a lemma whose proof is almost obvious.
Lemma 11.4.6. Let G⊆GL(n, C) be a ﬁnite subgroup.
(a) C
n
/G is an orbifold.
(b) If G is abelian, then C
n
/G has abelian ﬁnite quotient singularities.
Proof. Deﬁnition 11.4.5 guarantees that 0 ∈ C
n
/G is a ﬁnite quotient singularity.
But what about the other points of C
n
/G? For v ∈ C
n
, let G
v
=¦g ∈ G [ g v = v¦
be its isotropy subgroup. We will show that 0 ∈ C
n
/G
v
is locally analytically
equivalent to v ∈ C
n
/G.
Note that w →w+v is equivariant with respect to G
v
and hence induces a
local analytic equivalence from 0 ∈ C
n
/G
v
to v ∈ C
n
/G
v
. To complete the proof,
we need to ﬁnd a local analytic equivalence from v ∈ C
n
/G
v
to v ∈ C
n
/G.
Pick coset representatives so that G=
i
g
i
G
v
. The points g
i
v are distinct in
C
n
/G
v
, so there is a classical neighborhood v ∈U such that the g
i
U are disjoint,
and replacing U with
g∈Gν
gU, we can assume that U is G
v
invariant. Then V =
i
g
i
U is a Ginvariant neighborhood of v, and one sees easily that U/G
v
≃V/G.
This gives the desired local analytic equivalence.
On C
n
/G, the coordinates x
1
, . . . , x
n
on C
n
become “local coordinates” on the
quotient. The intuition is that anything Ginvariant in the x
i
descends to C
n
/G. For
example, one can show that Γ(C
n
/G,
´
Ω
p
C
n
/G
) ≃Γ(C
n
, Ω
p
C
n )
G
, so that
´
Ω
p
C
n
/G
comes
from Ginvariant pforms in the local coordinates.
542 Chapter 11. Toric Resolutions and Toric Singularities
Example 11.4.7. If G ⊆ SL(n, C), then dx
1
∧ ∧dx
n
is Ginvariant. It follows
that ω
C
n
/G
≃
´
Ω
n
C
n
/G
is trivial, i.e., C
n
/G has trivial canonical divisor. This is why
ﬁnite subgroups of SL(n, C) are so important. ♦
In the analytic category, the existence of local coordinates enables one to do
analysis on orbifolds. See [35, App. A.3] for further discussion and references.
We now characterize simplicial toric varieties in terms of their singularities.
Theorem 11.4.8. Given a normal toric variety X
Σ
, the following are equivalent:
(a) Σ is simplicial.
(b) X
Σ
has abelian ﬁnite quotient singularities.
(c) X
Σ
is an orbifold.
(d) X
Σ
is rationally smooth.
Proof. First assume that Σ is simplicial and take σ ∈ Σ. If N
′
is the sublattice of
N ≃ Z
n
generated by the minimal generators of σ, then U
σ,N
′ ≃ C
k
, where k =
dimσ. Then Proposition 3.3.11 implies that there is a ﬁnite abelian group G with
U
σ
=U
σ,N
≃(U
σ,N
′
(C
∗
)
n−k
)/G ≃(C
k
(C
∗
)
n−k
)/G ⊆C
n
/G.
Since G is a subgroup of (C
∗
)
n
⊆ GL(n, C), Lemma11.4.6 implies that U
σ
and
hence X
Σ
have abelian ﬁnite quotient singularities.
This proves (a) ⇒ (b). Then (b) ⇒ (c) is obvious, and (c) ⇒ (d) follows
from Example 11.4.4. It remains to prove (d) ⇒(a). For this, we can assume that
σ ⊆ N
R
≃ R
n
is a cone such that U
σ
is rationally smooth. A face τ _ σ gives the
following objects:
• The sublattice N
τ
=Span(τ) ∩N and quotient lattice N(τ) =N/N
τ
.
• The exact sequence 1 →T
Nτ
→T
N
→T
N(τ)
→1 of tori.
• The orbit closure V(τ) =O(τ) ⊆U
σ
of dimension n−dimτ.
Since O(τ) ≃T
N(τ)
by Theorem 3.2.6, it follows easily that V(τ) is the ﬁxed point
set of the action of T
Nτ
on U
σ
(Exercise 11.4.1). In the standard notation for ﬁxed
point sets, this means U
T
Nτ
σ
=V(τ). Hence
(11.4.1) dimU
T
Nτ
σ
=n−dimτ.
Now let τ
1
, . . . , τ
r
be the facets of σ. Then each T
i
= T
Nτ
i
is a subtorus of
T =T
Nσ
of codimension 1 (Exercise 11.4.1). Furthermore, for any p ∈V(σ),
(11.4.2)
dim
p
U
σ
−dim
p
U
T
σ
=n−(n−dimσ) =dimσ
dim
p
U
T
i
σ
−dim
p
U
T
σ
=(n−dimτ
i
) −(n−dimσ) =1
§11.4. Toric Singularities 543
by (11.4.1). Since U
σ
is rationally smooth at p, a result of Brion [24, Thm., p. 130]
implies that
dim
p
U
σ
−dim
p
U
T
σ
=
T
′
_
dim
p
U
T
′
σ
−dim
p
U
T
σ
_
,
where the sum is over all subtori T
′
⊆ T of codimension 1. Using (11.4.2), we
obtain
dimσ ≥
r
i=1
_
dim
p
U
T
i
σ
−dim
p
U
T
σ
_
=r.
For any cone, the number of facets is at least its dimension, with equality if and
only if the cone is simplicial (Exercise 11.4.1). Hence the above inequality implies
that σ is simplicial, and the proof is complete.
We note one other important characterization of simplicial toric varieties: a
normal toric variety is simplicial if and only if it is Qfactorial, meaning that every
Weil divisor is QCartier. We proved this in Proposition 4.2.7.
Terminal and Canonical Singularities. Given a normal variety X, recall from
§11.2 that a resolution of singularities f : X
′
→X is crepant if X is QGorenstein
and K
X
′ = f
∗
K
X
. Having a crepant resolution is rare, so for most varieties, we
have K
X
′ ,= f
∗
K
X
no matter which resolution we use. Measuring the difference
between these divisors—their discrepancy—leads to some interesting classes of
singularities.
In general, let K
X
be a canonical divisor on a normal QGorenstein variety X.
Given a proper birational morphism f : X
′
→X with X
′
smooth, one can ﬁnd a
canonical divisor K
X
′ on X
′
such that
(11.4.3) K
X
′ = f
∗
K
X
+
i
a
i
E
i
,
where a
i
∈ Q and the E
i
are the irreducible divisors lying in the exceptional locus
Exc( f ). This is proved in [119, Rem. 412].
Deﬁnition 11.4.9. Let X be a normal QGorenstein variety.
(a) X has terminal singularities if there is a proper birational f : X
′
→X with X
′
smooth, such that the coefﬁcients a
i
in (11.4.3) satisfy a
i
> 0 for all i.
(b) X has canonical singularities if there is a proper birational f : X
′
→X with X
′
smooth, such that the coefﬁcients a
i
in (11.4.3) satisfy a
i
≥0 for all i.
If X has terminal (resp. canonical) singularities, then the inequalities a
i
> 0
(resp. a
i
≥0) hold for all proper birational morphisms f : X
′
→X with X
′
smooth
[119, Rem. 412 and 422]. Note that f : X
′
→X may fail to be a resolution of
singularities since (for instance) some smooth points of X may get blown up by f .
However, it is often useful to be able to work with these more general morphisms.
544 Chapter 11. Toric Resolutions and Toric Singularities
If X is not smooth and has terminal singularities, then X cannot have a SNC
crepant resolution. At a deeper level, terminal singularities are central to the mini
mal model program, since a minimal model is a Qfactorial projective variety with
terminal singularities. Canonical singularities are also connected with the minimal
model program, since (roughly speaking) canonical singularites are the those that
appear on canonical models of varieties of general type. We refer the reader to
[119, Ch. 4] for a general discussion of terminal and canonical singularities.
Let us apply these ideas in the toric context. Given a normal QGorenstein
toric variety X
Σ
, we can ﬁnd a toric resolution of singularities φ : X
Σ
′ →X
Σ
. Here,
we have the toric canonical divisors K
X
Σ
′
and K
X
Σ
, and the relation (11.4.3) can be
described explicitly as follows.
Lemma 11.4.10. Let ϕ be the support function of the QCartier divisor K
X
Σ
and
let φ : X
Σ
′ →X
Σ
be the toric morphism coming from a reﬁnement Σ
′
of Σ. Then:
K
X
Σ
′
=φ
∗
K
X
Σ
+
ρ∈Σ
′
(1)\Σ(1)
(ϕ(u
ρ
) −1)D
ρ
.
Proof. The key observation is that K
X
Σ
and its pullback φ
∗
K
X
Σ
have the same
support function. Thus φ
∗
K
X
Σ
= −
ρ∈Σ
′ ϕ(u
ρ
)D
ρ
. Then the desired equation
follows immediately since ϕ(u
ρ
) =1 for all ρ ∈ Σ(1).
Here is an easy consequence of this lemma.
Proposition 11.4.11. Let X
Σ
be a normal Gorenstein toric variety. Then X
Σ
has
canonical singularities.
Proof. The support function ϕ of K
X
Σ
is integral with respect to N since X
Σ
is
Gorenstein. Now let Σ
′
be a smooth reﬁnement of Σ. Given u
ρ
∈ Σ
′
(1), we have:
• ϕ(u
ρ
) ∈ Z since u
ρ
∈ [Σ[ ∩N.
• ϕ(u
ρ
) >0 since u
ρ
lies in a cone σ ∈ Σ and ϕ >0 on σ since ϕ takes the value
1 on every minimal generator of σ.
It follows that ϕ(u
ρ
) ≥1, and we are done by Lemma 11.4.10.
We can determine when an afﬁne toric variety U
σ
has terminal or canonical
singularities by studying the lattice points of the polytope
P
σ
=Conv(0, u
ρ
[ ρ ∈ σ(1)).
Here is the result.
Proposition 11.4.12. Let U
σ
and P
σ
be as above.
(a) The following are equivalent:
(i) U
σ
is QGorenstein
(ii) There is m ∈ M
Q
such that ¸m, u
ρ
) =1 for all ρ ∈ σ(1).
(iii) P
σ
has a unique facet not containing the origin.
§11.4. Toric Singularities 545
(b) If U
σ
is QGorenstein, then:
(i) U
σ
has terminal singularities ⇔ the only lattice points of P
σ
are given
by its vertices.
(ii) U
σ
has canonical singularities ⇔ the only nonzero lattice points of P
σ
lie in the facet not containing the origin.
Proof. We begin with part (a). On an afﬁne toric variety, the canonical divisor is
QCartier if and only some multiple is the divisor of a character. This easily gives
the equivalence of (i) and (ii). Furthermore, given m∈ M
Q
as in (ii), P
σ
lies on one
side of the hyperplane deﬁned by ¸m, u) = 1. The face cut out by this hyperplane
contains all of the u
ρ
and hence is the unique facet of P
σ
not containing the origin,
proving (iii). We leave (iii) ⇒(ii) as Exercise 11.4.2.
For part (b), ﬁrst note that the support function of K
Uσ
is ϕ(u) = ¸m, u) for
m∈ M
Q
is as in the previous paragraph. Then
P
σ
=¦u ∈ σ [ ¸m, u) ≤1¦.
Now take any primitive vector v ∈ σ∩N different from the u
ρ
. By taking a smooth
reﬁnement Σ of the star subdivision given by v, we get a proper birational mor
phism φ : X
Σ
→U
σ
such that v is the minimal generator of ρ = Cone(v) ∈ Σ(1).
By Lemma 11.4.10, the coefﬁcient of D
ρ
in K
X
Σ
−φ
∗
K
Uσ
is
(11.4.4) ϕ(v) −1 =¸m, v) −1.
If U
σ
has terminal singularities, then (11.4.4) is positive. Hence ¸m, v) > 1, so
that v / ∈ P
σ
. It follows that the only lattice points of P
σ
are its vertices. Conversely,
suppose that P
σ
satisﬁes this condition and let Σ be a smooth reﬁnement of σ.
Given any u
ρ
∈ Σ(1) ¸σ(1), we have u
ρ
/ ∈ P
σ
, so that (11.4.4) is positive. Hence
U
σ
has terminal singularities.
The proof for canonical singularities is similar. The only difference is that the
coefﬁcient (11.4.4) is allowed to be zero, which happens only for lattice points
lying on the facet of P
σ
not containing the origin. You will supply the details in
Exercise 11.4.2.
Example 11.4.13. Consider the lattice N = ¦(a, b, c) ∈ Z
3
[ a+b+c ≡ 0 mod 2¦
and cone σ =Cone(e
1
, e
2
, e
3
) from Example 11.2.7. There are two ways to see that
U
σ
has terminal singularities:
• The resolution φ : X
Σ
→U
σ
constructed in Example 11.2.7 satisﬁes K
X
Σ
=
φ
∗
K
Uσ
+
1
2
D
0
. Since
1
2
> 0 and D
0
is the exceptional locus, U
σ
has terminal
singularities by Deﬁnition 11.4.9.
• The minimal generators of σ with respect to N are 2e
1
, 2e
2
, 2e
3
, so that P
σ
=
Conv(0, 2e
1
, 2e
2
, 2e
3
). One easily checks that P
σ
∩N consists only of vertices,
so that U
σ
has terminal singularities by Proposition 11.4.12. ♦
In the surface case, terminal and canonical singularities are easy to understand.
546 Chapter 11. Toric Resolutions and Toric Singularities
Theorem 11.4.14. Let X
Σ
be a toric surface. Then:
(a) X
Σ
has terminal singularities ⇔X
Σ
is smooth.
(b) X
Σ
has canonical singularities ⇔X
Σ
is Gorenstein ⇔X
Σ
has at worst rational
double points.
Proof. Part (a) follows because a 2dimensional cone σ = Cone(u
1
, u
2
) such that
P
σ
∩N =Conv(0, u
1
, u
2
) ∩N =¦0, u
1
, u
2
¦ is smooth (Exercise 8.3.4).
For part (b), Proposition 11.2.8 implies that being Gorenstein is equivalent to
having rational double points by, and if X
Σ
is Gorenstein, then it has canonical
singularities by Proposition 11.4.11. To complete the proof, we will show that an
afﬁne toric surface U
σ
with canonical singularities has a rational double point. Let
d, k be the parameters of σ, so that σ has minimal generators e
2
, de
1
−ke
2
. Then
P
σ
=¦u ∈ σ [ ¸m, u) ≤1¦, m =
k+1
d
e
1
+e
2
.
If k < d −1, then ¸m, e
1
) =
k+1
d
< 1, so that e
1
is an interior point of P
σ
. This
is impossible since U
σ
has canonical singularities. The only remaining case is
k =d −1, which gives a rational double point by Example 10.1.5.
Reduction to Canonical and Terminal Singularities. We next discuss how any
toric singularity can be made ﬁrst canonical and then terminal. We begin with the
canonical case.
Proposition 11.4.15. A strongly convex cone σ ⊆N
R
has a canonically determined
reﬁnement Σ
can
such that X
Σcan
has canonical singularities and the induced toric
morphism φ : X
Σcan
→U
σ
is projective.
Proof. Let Θ
σ
= Conv(σ ∩(N ¸¦0¦)). In Exercise 11.4.3 you will show that Θ
σ
is a lattice polyhedron with σ as recession cone. You will also show that taking
cones over bounded faces of Θ
σ
gives a fan Σ
can
in N
R
such that the morphism
φ : Σ
can
→X
Σ
is projective.
Take a cone σ
′
∈ Σ
can
and suppose that σ
′
is the cone over the bounded face F
of Θ
σ
. Then F is the unique facet of P
σ
′ not containing the origin, so that U
σ
′ is Q
Gorenstein by Proposition 11.4.12. Since F is a facet of Θ
σ
=Conv(σ∩(N¸¦0¦)),
the only nonzero lattice points of P
σ
′ lie in F, so that U
σ
′ has canonical singularities
by Proposition 11.4.12.
Example 11.4.16. When σ is 2dimensional, the reﬁnement Σ
can
has an elegant
description as follows. Recall from Proposition 8.2.9 that the canonical sheaf ω
Uσ
comes from the ideal
Γ(U
σ
, ω
Uσ
) =
m∈Int(σ
∨
)∩M
C χ
m
.
This ideal is integrally closed with associated polyhedron
P
ω
=Conv(Int(σ) ∩N)
§11.4. Toric Singularities 547
as in Proposition 11.3.4. The reasoning in Example 11.3.6 shows that X
Pω
→U
σ
is
the blowup of this ideal. We claim that
Σ
can
is the normal fan of P
ω
,
i.e., X
Σcan
= X
Pω
. This will imply that X
Pω
has at worst rational double points by
Proposition 11.4.14 and Theorem 11.4.15.
In dimension 2, a reﬁnement of σ is determined uniquely by its ray generators.
Thus it sufﬁces to prove that
(11.4.5) the vertices of Θ
σ
are the ray generators of the normal fan of P
ω
.
We show this as follows. Recall that by Proposition 10.2.17, the ray generators of
the normal fan of P
m
= Θ
σ
∨ = Conv(σ
∨
∩(M¸ ¦0¦)) are almost the vertices of
Θ
σ
. The difference arises from slight complications that can occur at the edges of
σ, as can be seen by looking at Theorem 10.2.12 and Proposition 10.2.17.
Let us compare P
m
and P
ω
. Since the latter uses only interior lattice points,
these polyhedra differ along the edges of σ
∨
. To see how this affects their normal
fans, consider Figure 11, which shows what can happen at an edge τ _σ
∨
. Let a
1
be the “lattice length” (#number of lattice points−1) of the bounded edge of P
m
a
1
> 1
τ
a
1
= 1
τ
Figure 11. Two examples of Pm (outlined in bold) and Pω (shaded)
that touches τ. As you can see, P
m
and P
ω
have the same inner normal vectors near
τ when a
1
>1, but P
ω
has one less inner normal when a
1
=1. The same thing hap
pens at the other edge of σ
∨
. Then (11.4.5) follows by comparing Theorem 10.2.12
and Proposition 10.2.17 to Figure 11 (Exercise 11.4.4). ♦
Before we can take a canonical singularity and make it terminal, we need
to extend the deﬁnition of crepant, which was originally deﬁned only for reso
lutions of singularities.crepant morphism Given normal varieties X,Y where Y is
QGorenstein, a morphism f : X →Y is crepant if K
X
= f
∗
K
Y
. Then we have the
following result.
548 Chapter 11. Toric Resolutions and Toric Singularities
Proposition 11.4.17. If X
Σ
has canonical singularities, then we can ﬁnd a sim
plicial reﬁnement Σ
′
of Σ such that X
Σ
′ has terminal singularities and the induced
toric morphism φ: X
Σ
′ →X
Σ
is projective and crepant. Furthermore, X
Σ
′ is Goren
stein if and only if X
Σ
is Gorenstein.
Proof. Suppose that X
Σ
has canonical singularities. To measure how far X
Σ
is
from being terminal, set
t(Σ) =
σ∈Σmax
[(P
σ
∩N) ¸¦vertices¦[.
By Proposition 11.4.12, X
Σ
has terminal singularities if and only if t(Σ) =0.
If t(Σ) > 0, pick a nonvertex ν ∈ P
σ
∩N for some σ ∈ Σ
max
. Then the star
subdivision Σ
′
=Σ
∗
(ν) satisﬁes the following:
• X
Σ
′ has canonical singularities with t(Σ
′
) =t(Σ) −1.
• X
Σ
′ is Gorenstein if and only if X
Σ
is.
• X
Σ
′ →X
Σ
is crepant.
You will verify these properties in Exercise 11.4.5. Figure 12 shows what happens
for the afﬁne toric variety of a cone σ ⊆ R
3
with t(σ) = 2. Here, ν is one of the
P
σ ν
Figure 12. The polyedron Pσ and the star subdivision of σ coming from ν
two nonvertex lattice points of P
σ
and the star subdivision Σ
′
has t(Σ
′
) =1.
By induction on t(Σ), a sequence of these star subdivisions gives a reﬁnement
Σ
′
of Σ such that X
Σ
′ has terminal singularities, and X
Σ
′ is Gorenstein if and only
if X
Σ
is. If Σ
′
is simplicial, then we are done. If not, then the simplicial reﬁnement
constructed in Proposition 11.1.7 is easily seen to have the desired properties.
Terminal Singularities. Propositions 11.4.15 and 11.4.17 allow us to focus on
toric varieties with terminal singularities. We begin with the afﬁne case.
§11.4. Toric Singularities 549
Gorenstein afﬁne toric varieties with terminal singularities can be classiﬁed in
terms of empty lattice polytopes, which are lattice polytopes whose only lattice
points are their vertices. Here is the precise result.
Proposition 11.4.18. Classifying Gorenstein afﬁne toric varieties with terminal
singularities that come from ndimensional cones σ ⊆ N
R
≃ R
n
is equivalent to
classifying (n−1)dimensional empty lattice polytopes.
Proof. First assume that U
σ
is Gorenstein with terminal singularities. Then there
is m ∈ M such that ¸m, u
ρ
) = 1 for all ρ ∈ σ(1). Extend m to a basis of M and
consider the dual basis of N
R
≃R
n
. Since dim σ =n, it follows that σ =Cone(F),
where F ⊆ R
n−1
¦1¦ is the facet of P
σ
not containing the origin. Then F is a
lattice polytope of dimension n−1 and is empty since U
σ
has terminal singularities
(Proposition 11.4.12).
Conversely, given an empty lattice polytope P ⊆ R
n−1
of dimension n−1, we
get the ndimensional cone σ = Cone(P¦1¦) ⊆ R
n
. In Exercise 11.4.6 you will
show that U
σ
is Gorenstein and has terminal singularities since P is empty.
Proposition 11.4.18 has some nice consequences in the 3dimensional case.
Proposition 11.4.19. Let X
Σ
be a 3dimensional Gorenstein toric variety. Then:
(a) If X
Σ
is simplicial, then X
Σ
has terminal singularities ⇔X
Σ
is smooth.
(b) X
Σ
has a resolution of singularities φ : X
Σ
′
→X
Σ
such that φ is projective and
crepant.
Proof. For part (a), assume that X
Σ
is simplicial with terminal singularities and
take σ ∈ Σ(3). Since 2dimensional empty lattice simplices are lattice equivalent
to the standard 2simplex (Exercise 8.3.4), Proposition 11.4.18 implies that σ is
smooth. When σ ∈ Σ(2), Proposition 11.4.12 implies that P
σ
is a 2dimensional
empty lattice simplex, which as already noted is smooth. Hence X
Σ
is smooth.
For part (b), recall that X
Σ
has canonical singularities by Proposition 11.4.11.
Then Proposition 11.4.17 gives a simplicial reﬁnement Σ
′
such that φ: X
Σ
′ →X
Σ
is
projective and crepant, X
Σ
′ has terminal singularities, and X
Σ
′ is Gorenstein since
X
Σ
is. Then X
Σ
′ is smooth by part (a). One can also check that the reﬁnement
constructed in Proposition 11.4.17 does not affect the smooth cones of Σ (Exer
cise 11.4.7). Hence φ is the desired resolution of singularities.
Corollary 11.4.20. The toric variety of a 3dimensional reﬂexive polytope has a
projective crepant resolution.
Proof. This follows immediately from Proposition 11.4.19 since the toric variety
of a reﬂexive polytope is a Gorenstein Fano variety by Theorem 8.3.4.
In the 3dimensional case, there are also results about terminal singularities
that do not assume that the variety is Gorenstein. To state we result, we deﬁne the
550 Chapter 11. Toric Resolutions and Toric Singularities
index of a QGorenstein variety X to be the smallest positive integer r such that
rK
X
is Cartier. The X is Gorenstein if and only if it is QGorenstein of index 1.
Then we can classify all 3dimensional cones that give terminal singularities
as follows.
Theorem 11.4.21. Let σ ⊆ N
R
≃ R
3
be a 3dimensional cone whose afﬁne toric
variety U
σ
has terminal singularities.
(a) If σ is simplicial, then N has a basis u
1
, u
2
, u
3
such that
σ =Cone(u
1
, u
2
, u
1
+pu
2
+qu
3
),
where 0 ≤ p < q are relatively prime. Furthermore, U
σ
is QGorenstein of
index q, and
U
σ
≃C
3
/µ
q
,
where µ
q
=¦ζ ∈ C
∗
[ ζ
q
=1¦ acts on C
3
via ζ (x, y, z) =(ζ
−1
x, ζ
−p
y, ζz).
(b) If σ is not simplicial, then N has a basis u
1
, u
2
, u
3
such that
σ =Cone(u
1
, u
2
, u
1
+u
3
, u
2
+u
3
).
Furthermore, U
σ
is Gorenstein and U
σ
≃V(xy −zw) ⊆C
4
.
Proof. For part (a), ﬁrst note that P
σ
is a 3dimensional empty lattice simplex.
By the Terminal Lemma [137, §1.6], N has a basis such that the vertices of P
σ
are 0, u
1
, u
2
, u
1
+ pu
2
+qu
3
, where 0 ≤ p < q are relatively prime. This gives the
desired formula for σ, and the quotient construction from Chapter 5 easily gives
the representation U
σ
≃ C
3
/µ
q
in the statement of the theorem (Exercise 11.4.8).
To compute the index, let m =(1, 1, −p/q) relative to the dual basis of M. Then
¸m, u
1
) =¸m, u
2
) =¸m, u
3
) =1.
Since gcd(p, q) =1, qm is the smallest positive multiple of m that is a lattice point,
hence the index is q. See [57, Thm. 2.2] for part (b).
In [137, §1.6], Oda discusses the Terminal Lemma, with references. In the
literature, the action of µ
q
on C
3
is often written differently. Given the action
ζ (x, y, z) =(ζ
−1
x, ζ
−p
y, ζz) as in the theorem, let 0 ≤a < q be the multiplicative
inverse of −p modulo q. Using the automorphism of µ
q
given by ζ →ζ
a
and
changing coordinates in C
3
, the action becomes ζ (x, y, z) = (ζ
a
x, ζ
−a
y, ζz). See
[137, §1.6] for references to the classiﬁcation of 3dimensional cones whose toric
varieties have canonical singularities. Also, [5] and the references therein discuss
the classiﬁcation of empty lattice simplices in dimensions 3 and 4.
Part (b) of Theorem 11.4.21 is especially nice, since it shows that one of our
favorite examples, V(xy −zw) ⊆C
4
, is even more interesting than we realized.
The singularities of a normal variety have codimension ≥2. In Exercise 11.4.9
you will prove that the following result, which tells us that the codimension is
higher when the singularities are terminal.
§11.4. Toric Singularities 551
Proposition 11.4.22. Assume that X
Σ
has only terminal singularities.
(a) The singular locus (X
Σ
)
sing
of X
Σ
has codimension codim(X
Σ
)
sing
≥3.
(b) If in addition X
Σ
is Gorenstein and simplicial, then codim(X
Σ
)
sing
≥4.
Log Singularities. Besides terminal and canonical singularities, there are other
classes of singularities that are important for the minimal model program. We
will consider two, log canonical and Kawamata log terminal. The latter is usually
abbreviated klt. These singularities are deﬁned for a pair (X, D), where X is a
normal variety and D is a Qdivisor with coefﬁcients in the interval [0, 1]. See
[108, p. 98] or [119, Ch. 11] for a nice discussion of why pairs are useful. A more
recent reference is [15]. Pairs are sometimes called log pairs.
Let (X, D), D =
i
d
i
D
i
, be a pair. Also ﬁx a proper birational morphism
f : X
′
→X such that X
′
is smooth and Exc( f ) ∪
i
f
−1
(D
i
) is a SNC divisor. Note
that f may fail be a log resolution of D. We deﬁne the birational transform D
′
of
D to be the divisor on X
′
deﬁned by D
′
=
i
d
i
D
′
i
, where D
′
i
= f
−1
(D
i
∩U) and
U ⊆X is the largest open subset where f
−1
is a morphism.
Let (X, D) and f : X
′
→X be as above, and assume that K
X
+D is QCartier.
By [108, Sec. 2.3], we can pick a canonical divisor K
X
′ such that
(11.4.6) K
X
′ +D
′
= f
∗
(K
X
+D) +
i
a
i
E
i
,
where a
i
∈ Q and the E
i
are the irreducible divisors lying in the exceptional locus.
This is the log version of (11.4.3).
Deﬁnition 11.4.23. Let (X, D) be a pair such that K
X
+D is QCartier, and write
D=
i
d
i
D
i
with d
i
∈ [0, 1] ∩Q.
(a) (X, D) has log canonical singularites if there is f : X
′
→X as above such that
the coefﬁcients a
i
in (11.4.6) satisfy a
i
≥−1 for all i.
(b) (X, D) has klt singularites if d
i
∈[0, 1) for all i and there is f : X
′
→X as above
such that the coefﬁcients a
i
in (11.4.6) satisfy a
i
>−1 for all i.
One can show that this deﬁnition is independent of the morphism f : X
′
→X.
In practice, one often says “(X, D) is log canonical” instead of “(X, D) has log
canonical singularities,” and “(X, D) is klt” has a similar meaning.
In the toric case, we consider pairs (X
Σ
, D) where D is torusinvariant.
Proposition 11.4.24. Let X
Σ
be a normal toric variety, and let D =
ρ
d
ρ
D
ρ
,
where d
ρ
∈ [0, 1] ∩Q. If K
X
Σ
+D is QCartier, then:
(a) (X
Σ
, D) is log canonical.
(b) If in addition d
ρ
∈ [0, 1) for all ρ ∈ Σ(1), then (X
Σ
, D) is klt.
Proof. Let ϕ be the support function of K
X
Σ
+D, so that ϕ(u
ρ
) = 1 −d
ρ
for all
ρ ∈ Σ(1). Let φ : X
Σ
′ →X
Σ
be a toric log resolution. For the rest of the proof,
552 Chapter 11. Toric Resolutions and Toric Singularities
D
ρ
will denote the divisor on X
Σ
′ corresponding to ρ ∈ Σ
′
(1). Thus the birational
transform of D is D
′
=
ρ∈Σ(1)
d
ρ
D
ρ
.
To simplify notation, let Σ
′
(1) = A∪B, where A = Σ(1) and B = Σ
′
(1) ¸ A.
Since ϕ is the support function of φ
∗
(K
X
Σ
+D), we obtain
φ
∗
(K
X
Σ
+D) =−
ρ∈A∪B
ϕ(u
ρ
)D
ρ
=
ρ∈A
(−1+d
ρ
)D
ρ
−
ρ∈B
ϕ(u
ρ
)D
ρ
=−
ρ∈A∪B
D
ρ
+
ρ∈A
d
ρ
D
ρ
+
ρ∈B
(1−ϕ(u
ρ
))D
ρ
= K
X
Σ
′
+ D
′
+
ρ∈B
(1−ϕ(u
ρ
))D
ρ
.
Hence
(11.4.7) K
X
Σ
′
+D
′
=φ
∗
(K
X
Σ
+D) +
ρ∈B
(ϕ(u
ρ
) −1)D
ρ
.
This is the log version of Lemma 11.4.10.
To analyze the coefﬁcients, take ρ ∈ B =Σ
′
(1) ¸Σ(1). Since Σ
′
reﬁnes Σ, we
have u
ρ
∈ σ for some σ ∈ Σ. Thus u
ρ
=
γ∈σ(1)
λ
γ
u
γ
, where λ
γ
≥0. Hence
(11.4.8) ϕ(u
ρ
) =
γ∈σ(1)
λ
γ
ϕ(u
γ
) =
γ∈σ(1)
λ
γ
(1−d
γ
).
We have d
γ
≤ 1 by assumption, so that 1−d
γ
≥ 0. It follows that ϕ(u
ρ
) ≥ 0, and
thus the coefﬁcients ϕ(u
ρ
) −1 in (11.4.7) are all ≥−1. This proves that (X
Σ
, D) is
log canonical.
Now assume in addition that d
γ
< 1 for all γ ∈ Σ(1). Then 1 −d
γ
> 0 in
(11.4.8), and since the λ
γ
are not all 0 (u
ρ
,=0), we conclude that ϕ(u
ρ
) > 0 for all
ρ ∈ B. This shows that the coefﬁcients ϕ(u
ρ
) −1 in (11.4.7) are all > −1. Hence
(X
Σ
, D) is klt, as claimed.
Here is an easy corollary that you will prove in Exercise 11.4.10.
Corollary 11.4.25. Let X
Σ
be a normal toric variety. Then:
(a) (X
Σ
,
ρ
D
ρ
) is log canonical.
(b) If X
Σ
is QGorenstein, then (X
Σ
, 0) is klt.
Here is an example based on an observation of Chen and Shokurov [33].
Example 11.4.26. Recall from §8.3 that a projective normal variety X is Goren
stein Fano if −K
X
is ample. The log version says that a normal projective variety
X is of Fano type if there is Qdivisor D such that −(K
X
+D) is Qample (meaning
that some positive multiple in Cartier and ample) and (X, D) is klt. Let us show
that every projective toric variety X
Σ
is of Fano type.
§11.4. Toric Singularities 553
Since X
Σ
is projective, there is a lattice polytope P that gives an ample divisor
D
P
on X
Σ
. Multiplying P by a suitably large integer, we can assume that P has
an interior lattice point m, and translating by m (which gives a linearly equivalent
divisor), we can assume that 0 is an interior point. Since D
P
=
ρ
a
ρ
D
ρ
and
P =¦m ∈ M
R
[ ¸m, u
ρ
) ≥−a
ρ
¦,
it follows that a
ρ
> 0 for all ρ. Then set
D =
ρ
(1−εa
ρ
)D
ρ
,
where ε ∈ Q is positive and satisﬁes εa
ρ
< 1 for all ρ. Then the coefﬁcients of D
lie in [0, 1) for all ρ.
By construction, D satisﬁes K
X
Σ
+D = −εD
P
, so that K
X
Σ
+D is QCartier.
Then (X
Σ
, D) is klt by Proposition 11.4.24. It follows that X
Σ
is of Fano type since
−(K
X
Σ
+D) =εD
P
is Qample. ♦
In the minimal model program, many results assume that (X, D) is a klt pair.
See [15, Sec. 4.3] for some examples.
Exercises for §11.4.
11.4.1. In this exercise you will supply some details omitted from the proof of (d) ⇒(a)
in Theorem 11.4.8.
(a) Prove that V(τ) is the unique ﬁxed point of the action of T
Nτ
on U
σ
.
(b) Explain why T
Nτ
is a subtorus of T
N
of codimension 1 when τ is a facet of σ.
(c) Prove that σ is simplicial if and only if its dimension equals the number of its facets.
11.4.2. This exercise is concerned with the proof of Proposition 11.4.12.
(a) Prove (iii) ⇒(ii) in part (a).
(b) Prove the characterization of canonical singularities given in part (b).
11.4.3. As in Proposition 11.4.15, let Θ
σ
= Conv(σ ∩N ¸ ¦0¦).
(a) Prove that Θ
σ
is a lattice polyhedron with σ as recession cone.
(b) Prove that taking cones over bounded faces of Θ
σ
gives a fan Σ
can
reﬁning σ.
(c) Prove that Σ
can
→U
σ
is projective. Hint: Suppose σ ∈ Σ
can
comes from a facet of
Θ
σ
. If the facet is deﬁned by ¸m, −) = a, then consider the support function ϕ on
Θ
σ
=[Σ
can
[ whose restriction to σ is given by
1
a
m.
11.4.4. Complete the proof of (11.4.5) sketched in Example 11.4.16.
11.4.5. Prove the properties of Σ
′
=Σ
∗
(ν) stated in the three bullets in the proof of Propo
sition 11.4.17.
11.4.6. In the proof Proposition 11.4.18, we sawthat an elementary lattice polytope P gives
a cone σ of one dimension higher. Prove that U
σ
is Gorenstein with terminal singularities.
Hint: P is the facet of P
σ
lying at “height 1.”
554 Chapter 11. Toric Resolutions and Toric Singularities
11.4.7. Let φ : X
Σ
′ → X
Σ
be the “terminalization” constructed in Proposition 11.4.17.
Prove that all terminal simplices of Σ lie in Σ
′
and explain why this implies that φ is an
isomorphism above the Qfactorial terminal locus of X
Σ
.
11.4.8. Complete the proof of Theorem 11.4.21 by using the methods of Chapter 5 to
describe the quotient representation of U
σ
. Hint: Lemma 5.1.1 will be useful.
11.4.9. Prove Proposition 11.4.22. Hint: Study the cones σ ∈ Σ of dimension ≤ 3 using
the methods of the proof of Proposition 11.4.19.
11.4.10. Prove Corollary 11.4.25.
11.4.11. Here is a partial converse to Theorem 11.4.17. Assume that X
Σ
is QGorenstein.
Prove that if Σ has a reﬁnement Σ
′
such that X
Σ
′ has terminal singularities and X
Σ
′ →X
Σ
is crepant, then X
Σ
has canonical singularities.
11.4.12. Explain how Propositions 11.4.15 and 11.4.17 give the minimal resolution of
singularities of a toric surface X
Σ
. Draw some pictures to illustrate what is happening.
11.4.13. Let X
Σ
be a QGorenstein toric variety with terminal singularities.
(a) Give an example where X
Σ
has a crepant resolution of singularities. Hint: Try one of
our favorite examples.
(b) If X
Σ
is simplicial but not smooth, then prove that all resolutions of singularities are
not crepant.
(c) If X
Σ
is not smooth, then prove that all SNC resolutions of singularities are not crepant.
Bibliography
[1] O. Aichholzer, Extremal properties of 0/1polytopes of dimension 5, in Polytopes – Combi
natorics and Computation, (G. Kalai and G. Ziegler, eds.), DMV Seminar 29, Birkh¨ auser,
Boston, 2000, 111–130.
[2] M. Atiyah, On analytic surfaces with double points, Proc. Roy. Soc. London Ser. A 247 (1958)
237244.
[3] and I. G. MacDonald, Introduction to Commutative Algebra, AddisonWesley, Read
ing, MA, 1969.
[4] M. Barile, On simplicial toric varieties of codimension 2, Rend. Istit. Mat. Univ. Trieste 39
(2007), 9–42
[5] , D. Bernardi, A. Borisov and J.M. Kantor, On empty lattice simplicies in dimension
4, preprint, 2009. Available electronically at arXiv:0912.5310 [math.AG].
[6] V. Batyrev, Dual polyhedra and mirror symmetry for CalabiYau hypersurfaces in toric vari
eties, J. Algebraic Geom. 3 (1994), 493–535.
[7] , On the classiﬁcation of smooth projective toric varieties, Tohoku Math. J. 43 (1991),
569–585.
[8] , On the classiﬁcation of toric Fano 4folds, J. Math. Sci. (New York) 94 (1999),
1021–1050.
[9] , Variations of the mixed Hodge structure of afﬁne hypersurfaces in algebraic tori,
Duke Math. J. 69 (1993), 349–409.
[10] and D. Cox, On the Hodge stucture of projective hypersurfaces in toric varieties,
Duke Math. J. 75 (1994), 293–338.
[11] and L. Borisov, On CalabiYau complete intersections in toric varieties, in Higher
dimensional complex varieties (Trento, 1994), (M. Andreatta and T. Peternell, eds.), de
Gruyter, Berlin, 1996, 39–65.
[12] M. Beck and S. Robins, Computing the continuous discretely, Springer, New York, 2007.
[13] A. A. Beilinson, J. Bernstein and P. Deligne, Faisceaux pervers, in Analysis and topology on
singular spaces, I (Luminy, 1981), Astrisque 100 (1982) 5–171.
[14] E. Bierstone and P. Milman, Desingularization of toric and binomial varieties, J. Algebraic
Geom. 15 (2006), 443–486.
561
562 Bibliography
[15] C. Birkar, Birational Geometry, lecture notes for a course taught at Cambridge University,
Winter, 2007. Available electronically at arXiv:0706.1794 [math.AG].
[16] R. Blanco and S. Encinas, Embedded desingularization of toric varieties, preprint, 2009.
Available electronically at arXiv:0901.2211 [math.AG].
[17] M. Blickle, Multiplier ideals and modules on toric varieties, Math. Z. 248 (2004), 113–121.
[18] and R. Lazarsfeld, An informal introduction to multiplier ideals, in Trends in Com
mutative Algebra, (L. Avramov, M. Green and C. Huneke, eds.), Math. Sci. Res. Inst. Publ.
51, Cambridge Univ. Press, Cambridge, 2004, 87–114.
[19] A. Borel, et al., Seminar on transformation groups, Annals of Math. Studies 46, Princeton
Univ. Press, Princeton, 1960.
[20] R. Bott, Homogeneous vectors bundles, Ann. Math. 66 (1957), 203–248.
[21] T. Braden, Remarks on the combinatorial intersection cohomology of fans, Pure Appl. Math.
Q. 2 (2006), 1149–1186.
[22] G. Bredon, Sheaf Theory, second edition, Springer, New York, 1997.
[23] M. Brion, Piecewise polynomial functions, convex polytopes and enumerative geometry, in
Parameter spaces (Warsaw, 1994), Banach Center Publ. 36, Polish Acad. Sci., Warsaw, 1996,
25–44.
[24] , Rational smoothness and ﬁxed points of torus actions, Transform. Groups 4 (1999),
127–156.
[25] A. Brønsted, An Introduction to Convex Polytopes, Springer, 1983.
[26] W. Bruns and J. Gubeladze, Polytopal linear groups, J. Algebra 218 (1999), 715–737.
[27] and , Semigroup algebras and discrete geometry, in Geometry of toric vari
eties, (L. Bonavero and M. Brion, eds.), S´ emin. Congr. 6, Soc. Math. France, Paris, 2002,
43–127.
[28] , and N. V. Trung, Normal polytopes, triangulations, and Koszul algebras, J.
Reine Angew. Math. 485 (1997), 123–160.
[29] and J. Herzog, CohenMacaulay Rings, Cambridge Univ. Press, Cambridge, 1993.
[30] and B. Ichim, NORMALIZ: Computing normalizations of afﬁne semigroups, with
contributions by C. S¨ oger, available electronically at www.math.uos.de/normaliz/,
2009.
[31] A. Buch, J. F. Thomsen, N. Lauritzen and V. Mehta, The Frobenius morphism on a toric
variety, Tohoku Math. J. 49 (1997), 355–366.
[32] E. Cattani, D. Cox and A. Dickenstein, Residues in toric varieties, Compositio Math. 108
(1997), 35–76.
[33] Y. Chen and V. Shokurov, Strong rational connectedness of toric varieties, preprint, 2009.
Available electronically at arXiv:0905.1430 [math.AG].
[34] D. Cox, The homogeneous coordinate ring of a toric variety, J. Algebraic Geom. 4 (1995),
15–50.
[35] and S. Katz, Mirror Symmetry and Algebraic Geometry, Amer. Math. Soc., Provi
dence, RI, 1999.
[36] , J. Little and D. O’Shea, Ideals, Varieties and Algorithms, third edition, Undergradu
ate Texts in Math., Springer, New York, 2007.
[37] , and , Using Algebraic Geometry, second edition, Graduate Texts in
Math. 185, Springer, New York, 2005.
[38] and C. von Renesse, Primitive collections and toric varieties, Tohoku Math. J. 61
(2009), 309–332.
Bibliography 563
[39] D. Dais, Geometric combinatorics in the study of compact toric surfaces, in Algebraic and
Geometric Combinatorics, (C. Athanasiadis, V. Batyrev, D. Dais, M. Henk and F. Santos,
eds.), Contemp. Math. 423, Amer. Math. Soc., Providence, RI, 2006, 71–123.
[40] , Resolving 3dimensional toric singularities, in Geometry of toric varieties, (L.
Bonavero and M. Brion, eds.), S´ emin. Congr. 6, Soc. Math. France, Paris, 2002, 213–225
[41] , U.U. Haus and M. Henk, On crepant resolutions of 2parameter series of Gorenstein
cyclic quotient singularities, Results Math. 33 (1998), 208–265.
[42] V. Danilov, The geometry of toric varieties, Russian Math. Surveys 33 (1978), 97–154; Us
pekhi Mat. Nauk. 33 (1978), 85–134.
[43] and A. Khovanski
˘
i, Newton polyhedra and an algorithm for calculating Hodge
Deligne numbers (Russian), Izv. Akad. Nauk SSSR Ser. Mat. 50 (1986), 925–945. English
translation Math. USSRIzv. 29 (1987), 279–298.
[44] D. Delzant, Hamiltoniens p´ eriodiques et images convexes de l’application moment, Bull. Soc.
Math. France 116 (1988), 315–339.
[45] M. Demazure, Sousgroupes alg´ ebriques de rang maximum du groupe de Cremona, Ann. Sci.
´
Ecole Norm. Sup. 3 (1970), 507–588.
[46] I. Dolgachev, Lectures on Invariant Theory, Cambridge Univ. Press, Cambridge, 2003.
[47] A. Durfee, Fifteen characterizations of rational double points and simple critical points, En
seign. Math. 25 (1979), 131–163.
[48] D. Eisenbud, Commutative Algebra with a View Toward Algebraic Geometry, Graduate Texts
in Math. 150, Springer, New York, 1995.
[49] and J. Harris, The Geometry of Schemes, Graduate Texts in Math. 197, Springer, New
York, 2000.
[50] , M. Mustat ¸ˇ a, Mircea and M. Stillman, Cohomology on toric varieties and local coho
mology with monomial supports, J. Symbolic Comput. 29 (2000), 583–600.
[51] G. Ewald, Combinatorial Convexity and Algebraic Geometry, Graduate Texts in Math. 168,
Springer, New York, 1996.
[52] and U. Wessels, On the ampleness of invertible sheaves in complete projective vari
eties, Results Math. 19 (1991), 275–278.
[53] E. M. Feichtner, Rational versus real cohomology algebras of lowdimensional toric varieties,
Proc. Amer. Math. Soc. 131 (2003), 1695–1704.
[54] O. Fujino, Multiplication maps and vanishing theorems for toric varieties, Math. Z. 257
(2007), 631–641.
[55] , Notes on toric varieties from Mori theoretic viewpoint, Tohoku Math. J. 55 (2003),
551–564.
[56] , On the KleimanMori cone, Proc. Japan Acad. Ser. A Math. Sci. 81 (2005), 80–84.
[57] , Threedimensional terminal toric ﬂips, Cent. Eur. J. Math. 7 (2009), 46–53.
[58] , Vanishing theorems for toric polyhedra, in Higher dimensional algebraic varieties
and vector bundles, (S. Mukai, ed.), RIMS Lecture Notes B9, Res. Inst. Math. Sci. (RIMS),
Kyoto, 2008, 81–95.
[59] W. Fulton, Introduction to Toric Varieties, Annals of Math. Studies 131, Princeton Univ. Press,
Princeton, NJ, 1993.
[60] , Intersection theory, second edition, Ergebnisse der Mathematik und ihrer Grenzge
biete, 3. Folge, Band 2, Springer, Berlin, 1998.
[61] , Equivariant Cohomology in Algebraic Geometry, Eilenberg lectures, Columbia Uni
versity, Spring 2007, notes by D. Anderson. Available electronically at www.math.lsa.
umich.edu/
˜
dandersn/eilenberg.
564 Bibliography
[62] and R. Lazarsfeld, Connectivity and its applications in algebraic geometry, in Alge
braic Geometry: Proceedings University of Illinois at Chicago Circle, 1980, (A. Libgober and
P. Wagreich, eds), Lecture Notes in Math. 862, Springer, Berlin, 1981, 26–92.
[63] I. M. Gelfand, M. M. Kapranov and A. V. Zelevinsky, Discriminants, Resultants and Multidi
mensional Determinants, Birkh¨ auser, Boston, 1994.
[64] R. Godement, Th´ eorie ds Faisceaux, Hermann, Paris, 1964.
[65] P. Gonz´ alez P´ erez and B. Teissier, Embedded resolutions of non necessarily normal afﬁne
toric varieties, C. R. Math. Acad. Sci. Paris 334 (2002), 379–382.
[66] M. Goresky, R. Kottwitz and R. MacPherson, Equivariant cohomology, Koszul duality, and
the localization theorem, Invent. Math. 131 (1998), 25–83.
[67] and R. MacPherson, Intersection homology theory, Topology 19 (1980), 135–162.
[68] and , Intersection homology theory II, Invent. Math. 72 (1983), 77–129.
[69] R. Goward, A simple algorithm for principalization of monomial ideals, Trans. Amer. Math.
Soc. 357 (2005), 4805–4812.
[70] H. Grauert and R. Remmert, Coherent Analytic Sheaves, Grundlehren der mathematischen
Wissenschaften 265, Springer, Berlin, 1984.
[71] M. Greenberg and J. Harper, Algebraic Topology: A First Course, Benjamin/Cummings, Read
ing, MA, 1981. Reprinted by Westview Press, Boulder, CO.
[72] P. Grifﬁths and J. Harris, Principles of Algebraic Geometry, Wiley, New York, 1978
[73] A. Grothendieck, Revˆ etements ´ etales et groupe fondamental (SGA 1), S´ eminaire de g´ eom´ etrie
alg´ ebrique du Bois Marie 1960–61 (with two papers by M. Raynaud), Lecture Notes in
Math. 224, Springer, Berlin, 1971. Updated and annotated reprint, Documents Math´ ematiques
(Paris) 3, Soci´ et´ e Math´ ematique de France, Paris, 2003.
[74] ,
´
El´ ements de g´ eom´ etrie alg´ ebrique (r´ edig´ es avec la collaboration de Jean
Dieudonn´ e), II.
´
Etude globale ´ el´ ementaire de quelques classes de morphismes, Inst. Hautes
tudes Sci. Publ. Math. 8 (1961), 5–222.
[75] B. Gr¨ unbaum, Convex Polytopes, Wiley Interscience, New York, 1967. Second edition pre
pared by V. Kaibel, V. Klee and G. Ziegler, Springer, 2003.
[76] C. Haase and I. Melnikov, The reﬂexive dimension of a lattice polytope, Ann. Comb. 10 (2006),
211–217.
[77] J. Harris, Algebraic Geometry: A First Course, Graduate Texts in Math. 133, Springer, New
York, 1992.
[78] R. Hartshorne, Algebraic Geometry, Graduate Texts in Math. 52, Springer, New York, 1977.
[79] , Stable reﬂexive sheaves, Math. Ann. 254 (1980), 121–176.
[80] B. Hassett, Introduction to Algebraic Geometry, Cambridge Univ. Press, Cambridge, 2007.
[81] A. Hatcher, Algebraic Topology, Cambridge Univ. Press, Cambridge, 2002.
[82] H. Hauser, The Hironaka Theorem on resolution of singularities (Or: a proof that we always
wanted to understand), Bull. Amer. Math. Soc. 40 (2003), 323–403.
[83] D. Hensley, Continued Fractions, World Scientiﬁc, Singapore, 2006.
[84] R. Hemmecke, R. Hemmecke and P. Malkin, 4ti2 Version 1.2—Computation of Hilbert bases,
Graver bases, toric Gr¨ obner bases, and more, available electronically at www.4ti2.de,
2005.
[85] M. Hering, A. K¨ uronya and S. Payne, Asymptotic cohomological functions of toric divisors,
Adv. Math. 207 (2006), 634–645.
[86] L. Hille and M. Perling, A counterexample to King’s conjecture, Compos. Math. 142 (2006),
1507–1521.
Bibliography 565
[87] H. Hironaka, Resolution of singularities of an algebraic variety over a ﬁeld of characteristic
zero III, Ann. Math. 79 (1964), 109–326.
[88] F. Hirzebruch, Hilbert Modular Surfaces, Ens. Math. 19 (1973), 183–281.
[89] M. Hochster, Rings of invariants of tori, CohenMacaulay rings generated by monomials, and
polytopes, Ann. Math. 96 (1972), 318–337.
[90] J. Howald, Multiplier ideals of monomial ideals, Trans. Amer. Math. Soc. 353 (2001), 2665–
2671.
[91] S. Hu, Semistable degeneration of toric varieties and their hypersurfaces, Comm. Anal.
Geom. 14 (2006), 59–89.
[92] K. Hulek, Elementary Algebraic Geometry, Amer. Math. Soc., Providence, RI, 2003.
[93] J. Humphreys, Linear Algebraic Groups, Springer, Graduate Texts in Math. 21, New York,
1975.
[94] N. Ilten, Deformations of smooth toric surfaces, preprint, 2009. Available electronically at
arXiv:0902.0529 [math.AG].
[95] Y. Ito, Minimal resolution via Gr¨ obner basis, in Algebraic Geometry in East Asia, (A.
Ohbuchi, K. Konno, S. Usui, A. Moriwaki and N. Nakayama, eds.), World Sci. Publ., River
Edge, NJ, 2002, 165–174.
[96] , Special M
c
Kay Correspondence, in Geometry of toric varieties, (L. Bonavero and
M. Brion, eds.), S´ emin. Congr. 6, Soc. Math. France, Paris, 2002, 213–225
[97] S. Iyengar, G. Leuschke, A. Leykin, C. Miller, E. Miller, A. Singh and U. Walther, Twenty
four Hours of Local Cohomology, Grad. Stud. Math. 87, Amer. Math. Soc., Providence, RI,
2007.
[98] N. Jacobson, Basic Algebra II, Freeman, San Francisco, 1980.
[99] K. Jaczewski, Generalized Euler sequences and toric varieties, in Classiﬁcation of Algebraic
Varieties (L’Aquila, 1992), (C. Ciliberto, E. L. Livorni and A. Sommese, eds.), Contemp.
Math. 162, Amer. Math. Soc., Providence, RI, 1994, 227–247.
[100] A. Jordan, Homology and Cohomology of Toric Varieties, Ph.D. thesis, University of Kon
stanz, 1998. Available electronically at kops.ub.unikonstanz.de/volltexte/
2006/2037/pdf/preprint\_057.pdf.
[101] A. Kasprzyk, M. Kreuzer and B. Nill, On the combinatorial classiﬁcation of toric log del
Pezzo surfaces, preprint, 2008. Available electronically at arXiv:0810.2207 [math.AG].
[102] A. Katsabekis and A. Thoma, Toric sets and orbits on toric varieties, J. Pure Appl. Algebra
181 (2003), 75–83.
[103] K. Kaveh, Note on cohomology rings of spherical varieties and volume polynomial, preprint,
2010. Available electronically at arXiv:math/0312503 [math.AG].
[104] G. Kempf, F. Knudsen, D. Mumford and B. SaintDonat, Toroidal Embeddings I, Lecture
Notes in Math. 339, Springer, New York, 1973.
[105] S. Kleiman, Toward a numerical theory of ampleness, Ann. of Math. 84 (1966), 293–344.
[106] P. Kleinschmidt, A classiﬁcation of toric varieties with few generators, Aequationes Math. 35
(1988), 254–266.
[107] A. Klyachko, Equivariant vector bundles over toric varieties, Math. USSRIzv. 35 (1990),
337–375.
[108] J. Koll´ ar and S. Mori, Birational Geometry of Algebraic Varieties, Cambridge Univ. Press,
Cambridge, 1998.
[109] S. Krantz, Geometric Function Theory: Explorations in Complex Analysis, Birkh¨ auser,
Boston, 2005.
566 Bibliography
[110] A. Kresch, GromovWitten invariants of a class of toric varieties, Michigan Math. J. 48 (2000),
369–391.
[111] M. Kreuzer and H. Skarke, PALP: A package for analyzing lattice polytopes with applications
to toric geometry, Comput. Phys. Comm. 157 (2004), 87–106.
[112] E. Kunz (with contributions by D. Cox and A. Dickenstein), Residues and Duality for Projec
tive Algebraic Varieties, Amer. Math. Soc., Providence, RI, 2008.
[113] R. Laterveer, Linear systems on toric varieties, Tohoku Math. J. 48 (1996), 451–458.
[114] R. Lazarsfeld, Positivity in Algebraic Geometry, Vols. I and II, Springer, New York, 2004.
[115] Y. Liu, L. E. Trotter and G. Ziegler, On the height of the minimal Hilbert basis, Results Math.
23 (1993), 374–376.
[116] I. G. Macdonald, The volume of a lattice polyhedron, Proc. Camb. Phil. Soc. 59 (1963), 719–
726.
[117] P. Mani, Spheres with few vertices, Journal of Comb. Theory Ser. A 13 (1992), 346–352.
[118] E. Materov, The Bott formula for toric varieties, Mosc. Math. J. 2 (2002), 161–182, 200.
[119] K. Matsuki, Introduction to the Mori Program, Universitext, Springer, New York, 2002.
[120] H. Matsumura, Commutative Ring Theory, Cambridge Univ. Press, Cambridge, 1986.
[121] A. Mavlyutov, Cohomology of rational forms and a vanishing theorem on toric varieties, J.
Reine Angew. Math. 615 (2008), 45–58.
[122] , Semiample hypersurfaces in toric varieties Duke Math. J. 101 (2000), 85–116.
[123] J. McCleary, A User’s Guide to Spectral Sequences, Cambridge Univ. Press, Cambridge, 2001.
[124] M. McConnell, The rational homology of toric varieties is not an combinatorial invariant,
Proc. Amer. Math. Soc. 104 (1989), 986991.
[125] D. McDuff, Symplectic embeddings of 4dimensional ellipsoids, J. Topol. 2 (2009), 1–22.
[126] E. Miller and B. Sturmfels, Combinatorial Commutative Algebra, Graduate Texts in Math.
227, Springer, New York, 2005.
[127] T. Morita, Hilbert schemes of ﬁnite abelian group orbits and Gr¨ obner fans, preprint, 2007.
Available electronically at arXiv:0712.2892 [math.AG].
[128] D. Mumford, Algebraic Geometry I: Complex Projective Varieties, Springer, Berlin, 1976.
[129] , The Red Book of Varieties and Schemes, second expanded edition, Lecture Notes in
Math. 1358, Springer, Berlin, 1999.
[130] , J. Fogarty, J. and F. Kirwan, Geometric Invariant Theory, third edition, Ergebnisse
der Mathematik und ihrer Grenzgebiete 34, Springer, Berlin, 1994.
[131] J. Munkres, Elements of Algebraic Topology, AddisonWesley, Reading, MA, 1984. Reprinted
by Westview Press, Boulder, CO.
[132] M. Mustat ¸ˇ a, Local cohomology at monomial ideals, J. Symbolic Comput. 29 (2000), 709–720.
[133] , Vanishing theorems on toric varieties, Tohoku Math. J. 54 (2002), 451–470.
[134] B. Nill, Gorenstein toric Fano varieties, Manuscr. Math. 116 (2005), 183–210.
[135] , Volume and lattice points of reﬂexive simplices, Discr. Comp. Geom. 37 (2007), 301–
320.
[136] M. Øbro, An algorithm for the classiﬁcation of smooth Fano polytopes, preprint, 2007. Avail
able electronically at arXiv:0704.0049 [math.CO].
[137] T. Oda, Convex Bodies and Algebraic Geometry, Springer, New York, 1988.
[138] , Torus Embeddings and Applications. Based on joint work with Katsuya Miyake, Tata
Institute of Fundamental Research, Bombay; Springer, Berlin, 1978.
Bibliography 567
[139] , Simple convex polytopes and the strong Lefschetz theorem, J. Pure Appl. Algebra 71
(1991), 265–286.
[140] S. Ogata and K. Nakagawa, On generators of ideals deﬁning projective toric varieties,
manuscripta math. 108 (2002), 3342.
[141] S. Payne, Moduli of toric vector bundles, Compos. Math. 144 (2008), 1199–1213.
[142] , Equivariant Chow cohomology of toric varieties, Math. Res. Lett. 13 (2006), 29–41.
[143] M. Perling, Divisorial cohomology vanishing on toric varieties, preprint, 2009, to appear.
[144] C. Peters and J. Steenbrink, Mixed Hodge Structures, Springer, New York, 2008.
[145] B. Poonen and F. RodriguezVillegas, Lattice polygons and the number 12, Am. Math.
Monthly 107 (2000), 238–250.
[146] P. PopescuPampu, The geometry of continued fractions and the topology of surface singular
ities, in Singularities in geometry and topology 2004, (J.P. Brasselet and T. Suwa, eds.), Adv.
Stud. Pure Math., 46, Math. Soc. Japan, Tokyo, 2007, 119–195.
[147] A. V. Pukhlikov and A. G. Khovanskii, The RiemannRoch theorem for integrals and sums of
quasipolynomials on virtual polyhedra, Algebra i Analiz 4 (1992), 188–216, translation in St.
Petersburg Math J. 4 (1993), 789–812.
[148] M. Reid, Canonical 3folds, in Journ´ ees de G´ eometrie Alg´ ebrique d’Angers, Juillet 1979/Al
gebraic Geometry, Angers, 1979, (A. Beauville, ed.), Sijthoff & Noordhoff, Alphen aan den
Rijn, 1980, 273–310.
[149] , Decomposition of Toric Morphisms, in Arithmetic and Geometry, (M. Artin and J.
Tate, eds.), Birkh¨ auser, Boston Basel Berlin, 1983, 395–418.
[150] , Undergraduate Algebraic Geometry, Cambridge Univ. Press, Cambridge, 1989.
[151] E. Reyes, R. Villarreal and L. Z´ arate, A note on afﬁne toric varieties, Linear Algebra Appl.
318 (2000), 173–179.
[152] J. RichterGebert, B. Sturmfels and T. Theobald, First steps in tropical geometry, in Idem
potent Mathematics and Mathematical Physics, (G. Litvinov and V. Maslov, eds.), Contemp.
Math. 377, Amer. Math. Soc., Providence, RI, 2003.
[153] R. T. Rockafellar, Convex Analysis, Princeton Univ. Press, Princeton, NJ, 1970.
[154] J. J. Rotman, Advanced Modern Algebra, PrenticeHall, Upper Saddle River, NJ, 2002.
[155] H. Sato, Toward the classiﬁcation of higherdimensional toric Fano varieties, Tohoku Math.
J. 52 (2000), 383–413.
[156] I. Shafarevich, Basic Algebraic Geometry, Volumes 1 and 2, Springer, New York, 1994 and
1996.
[157] H. Schenck, Computational Algebraic Geometry, Cambridge Univ. Press, Cambridge, 2003.
[158] J.P. Serre, Faisceaux alg´ ebriques coh´ erents, Ann. Math. 61, (1955), 197–278.
[159] , G´ eom´ etrie alg´ ebrique et g´ eom´ etrie analytique, Ann. Inst. Fourier (Grenoble) 6
(1956), 1–42.
[160] A. Simis, W. Vasconcelos and R. Villarreal, On the ideal theory of graphs, J. Algebra 167
(1994), 389–416.
[161] N. J. A. Sloane, Online Encyclopedia of Integers Sequences, 2010. Published electronically
at the web site oeis.org.
[162] K. Smith, L. Kahanp¨ a¨ a, P. Kek¨ al¨ ainen and W. Traves, An Invitation to Algebraic Geometry,
Springer, New York, 2000.
[163] F. Sottile, Toric ideals, real toric varieties, and the algebraic moment map, in Topics in Al
gebraic Geometry and Geometric Modeling (Proceedings of AGGM, Vilnius, Lithuania.), (R.
Goldman and R
˙
Krasuaskas eds), Contemp. Math. 334, Amer. Math. Soc., Providence, RI,
2003, 225–240. Corrected version available electronically at arXiv:math/0212044 [math.AG].
568 Bibliography
[164] E. Spanier, Algebraic Topology, McGrawHill, New York, 1966. Reprinted by Springer Ver
lag, New York, 1994.
[165] R. Stanley, Generalized Hvectors, intersection cohomology of toric varieties, and related
results, in Commutative algebra and combinatorics (Kyoto 1985), (M. Nagata and H. Mat
sumura, eds.), Adv. Stud. Pure Math. 11, NorthHolland, Amsterdam, 1987, 187–213.
[166] H. Stark, An Introduction to Number Theory, MIT Press, Cambridge, MA, 1978.
[167] J. Steenbrink, Intersection form for quasihomogeneous singularities, Compositio Math. 34
(1977), 211–223.
[168] , Mixed Hodge structure on the vanishing cohomology, in Real and complex singu
larities (Proc. Ninth Nordic Summer School/NAVF Sympos. Math., Oslo, 1976), Sijthoff and
Noordhoff, Alphen aan den Rijn, 1977, 525–563.
[169] R. S. Strichartz, The Way of Analysis, Jones and Bartlett, Boston, MA, 1995.
[170] B. Sturmfels, Gr¨ obner Bases and Convex Polytopes, Amer. Math. Soc., Providence, RI, 1996.
[171] H. Sumihiro, Equivariant completion, I, II, J. Math. Kyoto Univ. 14 (1974), 1–28; 15 (1975),
573–605.
[172] I. Swanson and C. Huneke, Integral Closure of Ideals, Rings, and Modules, LMS Lecture
Note Series 336, Cambridge Univ. Press, Cambridge, 2006.
[173] B. Tessier, Monomial ideals, binomial ideals, polynomial ideals, in Trends in Commutative
Algebra, (L. Avramov, M. Green and C. Huneke, eds.), Math. Sci. Res. Inst. Publ. 51, Cam
bridge Univ. Press, Cambridge, 2004, 211–246.
[174] H. Thompson, Comments on toric varieties, preprint, 2003. Available electronically at
arXiv:math/0310336 [math.AG].
[175] K. Ueno, Algebraic Geometry 2, Amer. Math. Soc., Providence, RI, 2001.
[176] R. Villarreal, Monomial Algebras, Marcel Dekker, Inc., New York, 2001.
[177] C. Weibel, An Introduction to Homological Algebra, Cambridge Univ. Press, Cambridge,
1994.
[178] O. Zariski and P. Samuel, Commutative Algebra, Volume II, Van Nostrand, Princeton, 1960.
Corrected reprint, Springer, 1976.
[179] G. Ziegler, Lectures on Polytopes, Graduate Texts in Math. 152, Springer, 1995.
Index
adjunction formula, 358, 497
afﬁne morphism, 389
Alexander duality, 412
analytic variety, 541
annihilator, 409
arithmetically CohenMacaulay (aCM), 408
barycentric subdivision, 519
Bertini theorem, 503
birational transform of a divisor, 551
blowup
of an ideal, 529
Bott formula, 431, 447
CalabiYau variety, 523
canonical bundle, 348
canonical class, 348
canonical divisor, 348, 365
of a toric variety, 358
relative, 535
canonical module
of a projective toric variety, 364
of an afﬁne toric variety, 363
of the total coordinate ring, 361
canonical sheaf, 348, 357
of a toric variety, 358
canonical singularities, 543
Carath´ eodory’s theorem, 531
Cartan matrix, 494
Cartier divisor
big, 419
ˇ
Cech cochains, 381
ˇ
Cech cohomology, 381–383
ˇ
Cech Complex
toric, 390
ˇ
Cech complex, 381, 382
ceiling ⌈ ⌉, 416
class group, 342
CohenMacaulay variety, 386
toric, 408
compex
map of, 380
complex, 380
exact sequence of, 380, 381
Koszul, 395, 400
cone
Gorenstein, 371
reﬂexive Gorenstein, 371
connecting homomorphism, 380
conormal sheaf, 347
constant sheaf, 388
continued fraction
Felix Klein, 470
HirzebruchJung, 462, 463, 467
ordinary, 469, 470
contractible, 388
convergent, 469
convergents, 463
coordinate ring
homogeneous, 364
cotangent bundle, 344
toric, 444
cotangent sheaf, 344
of a toric variety, 351
Cousin problem, 384
crepant resolution, 521, 523
deformation retract, 388
DehnSommerville equations, 427, 428, 430, 431
depth, 386, 409
derivation, 349
torusinvariant, 356
569
570 Index
universal, 349
derived functor, 381, 435, 437
dimension
of a module, 409
direct image, 384
higher, 385
left exact, 356
discrepancy, 543
division algorithm, 469
modiﬁed, 453, 461
divisor
anticanonical, 371
toric, 372
linear equivalence of, 342
double dual, 339, 342
double point, 456
Du Val singularity, 456, 494
dual sheaf, 339, 342
dualizing complex, 387
dualizing sheaf, 348
Dynkin diagram, 456, 461, 494
edge ideal of a graph, 533
Ehrhart polynomial, 424, 425, 502
of a polygon, 425, 426
of a polytopal complex, 426, 433
Ehrhart quasipolynomial, 434
Ehrhart reciprocity, 424, 428
Euclidean algorithm, 462, 469
modifed, 462
Euler characteristic
of a sheaf, 422, 432
of a topological space, 433
Euler relation, 357
generalized, 357
Euler sequence, 355, 357
generalized, 356
toric, 355
exact sequence
long, 381
of complexes, 380
exceptional curve of the ﬁrst kind, 492
exceptional divisor, 458
exceptional locus, 515
Ext, 387, 437
face numbers, 427
Fano type, 552
toric, 552
ﬁnite quotient singularity, 541
ﬂoor ⌊ ⌋, 416
fractional ideal
sheaf of, 534
functor, 380
derived, 381
GHilbert scheme, 487, 489
Gale duality, 444
Gr¨ obner basis, 479
marked, 479
reduced, 479
universal, 479
Gr¨ obner fan, 480
resolution of singularities, 484
higher direct image, 385
Hilbert basis, 467, 468
Hilbert polynomial
of a projective variety, 424, 425
of a sheaf, 423
Hilbert polynomial of a sheaf, 432
Hirzebruch surface, 360, 400, 417, 445, 492, 503
Iitaka dimension, 419
index of a QGorenstein toric variety, 549
index of a simplicial cone, 513
injective resolution, 379, 380
injective sheaf, 379
injectivity lemma, 404, 405, 416, 422
integral closure
of an ideal, 531, 534
integrally closed ideal, 530
intersection product
on a toric surface, 491
Jacobian matrix, 343
K¨ ahler differentials, 343, 344, 346
Kawamata log terminal singularities, 551
klt singularities, 551
Koszul complex, 395, 400
lattice equivalence, 373, 455
lattice polytope
empty, 549
LDP polygon, 378
index of, 378
left exact, 381
Leray spectral sequence, 385
line bundle, 340
local
ˇ
Cech complex, 435, 436
local analytic equivalence, 541
local cohomology, 435–438
log canonical singualarities, 551
log del Pezzo surface, 378
index of, 378
log pair, 551
log resolution, 520, 535
logarithmic 1forms, 351, 352
logarithmic poles, 352
long exact sequence, 435
Macaulay 2, 437, 445, 446
maximal CohenMacaulay (MCM), 409
Index 571
maximal projective crepant partial
desingularization, 523
mirror symmetry, 376
Monomial Brianc¸onSkoda theorem, 534
monomial ideal, 534
multiplicity of a simplicial cone, 459, 513, 514
properties of, 513
multiplier ideal, 536, 539
toric, 536, 539
multiplier module, 539
M
c
Kay correspondence, 485–487
Noether’s theorem, 499, 505
normal bundle, 347
normal sheaf, 347, 358
Normaliz, 424
normalized basis, 454
orbifold, 541
pair (X, D), 551
parameters of a cone, 454
partial quotient, 469
partial quotients, 463
pEhrhart polynomial, 426, 427, 506
Picard group, 342
Pick’s formula, 426
Poincar´ e residue, 352, 353
polytopal complex, 426
polytope
normal, 364
principalization, 535
Proj, 364
QCartier divisor
nef, 402
Qdivisor, 416
QWeil divisor, 416
round down of, 416
round up of, 416
rank
of a coherent sheaf, 340
of a module, 358
rank of a ﬁnitely generated abelian group, 518
rational double point, 456, 458, 461, 494
rational normal cone, 455, 457
rational singularities, 540
rational singularity
toric, 540
rationally smooth, 540
Rdivisor, 416
reduced cohomology, 389
Rees algebra, 528
reﬁnement of a fan, 366
reﬂexive polygon, 374, 504
reﬂexive polytope, 372, 373, 523
classiﬁcation, 376
reﬂexive sheaf, 339
of rank 1, 340
regular representation, 485
regular sequence, 386
relative cohomology, 397
representations of ﬁnite groups, 485
special, 485, 486
resolution of singualities
toric surface, 464
resolution of singularities, 457, 507
crepant, 505, 521
embedded, 524
Gr¨ obner fan, 484
Hironaka, 507
log, 520
minimal, 493
projective, 507
SNC, 520
toric, 514
toric surface, 459, 464
RiemannRoch
surfaces, 498
topological part, 499
RiemannRoch theorem
curves, 497
sectional genus, 503
of a toric surface, 503
Serre duality, 387, 412
MCM sheaves, 409
toric, 408
sheaf
of 1forms, 344
of Zariski pforms, 347
sheaf cohomology, 379, 380, 383
long exact sequence, 381
toric, 394, 401, 438
toric surface, 398
sheaf of pforms, 347
simple polytope, 426
simplicialization, 512, 514
singular cohomology, 387, 388
reduced, 389
relative, 397
ring, 388
singular locus, 507
toric, 508
singular point
of a toric surface, 453
smooth Fano polytope, 377
smooth locus, 507
smooth normal crossing divisor, 519
smooth normal crossings, 352
SNC divisor, 519
spectral sequence, 385, 448
convergent, 449
572 Index
degenerates, 449
differentials, 448
E
1
, 448
E
2
, 448
edge homomorphism, 449, 450
ﬁrst quadrant, 448
Leray, 385
of a closed cover, 395
of an open cover, 386
star subdivision, 509, 512, 514
properties of, 509, 511
structure sheaf
analytic, 541
supplementary cone, 472, 473
isomorphic to dual cone, 474
support function
of a QCartier divisor, 400
of a Cartier divisor
strictly convex, 512
symbolic powers, 539
tangent bundle, 344
tangent sheaf, 344
Taylor resolution, 440, 441
tensor product
of sheaves, 342
Terminal lemma, 550
terminal singularities, 543
toric singularities
canonical, 544–546, 548
klt, 551
log canonical, 551
terminal, 545, 548, 549
toric singularity
ﬁnite quotient, 542
simplicial, 542
toric surface
resolution of singularities, 459, 464
singular points, 453
smooth classiﬁcation, 489, 490
torsion submodule, 435
total coordinate ring, 354, 438
twist of a sheaf, 413
Vandermonde identity, 431, 434
vanishing theorem
BatyrevBorisov, 405, 408, 417
BottSteenbrinkDanilov, 413, 415
Demazure, 402, 417
GrauertRiemenschneider, 420, 421
higher direct images, 403
KawamataViehweg, 419
Kodaira, 413, 415
Mavlyutov, 416
Mustat ¸ˇ a, 418
Nakano, 413, 415
Serre afﬁne, 382
Serre projective, 384
Zariski pforms, 415, 431
variety
Gorenstein, 365
Gorenstein Fano, 371
toric, 372
wedge product, 347, 358
Weil divisor, 340
Zariski 1forms
of a toric variety, 353
on a toric variety, 355
Zariski pforms, 347, 408, 413, 415, 416, 426, 431
on a simplicial toric variety, 370
on a toric variety, 367
on an afﬁne toric variety, 368
Zariski tangent space, 346
c 2010, David Cox, John Little and Hal Schenck
Contents
Preface Notation Part I. Basic Theory of Toric Varieties Chapter 1. §1.0. §1.1. §1.2. §1.3. Afﬁne Toric Varieties
vii xi 1 3 3 10 23 35 48 49 49 54 62 74 85 93 93 105 114 125 139 iii
Background: Afﬁne Varieties Introduction to Afﬁne Toric Varieties Cones and Afﬁne Toric Varieties Properties of Afﬁne Toric Varieties
Appendix: Tensor Products of Coordinate Rings Chapter 2. §2.0. §2.1. §2.2. §2.3. §2.4. Projective Toric Varieties
Background: Projective Varieties Lattice Points and Projective Toric Varieties Lattice Points and Polytopes Polytopes and Projective Toric Varieties Properties of Projective Toric Varieties Normal Toric Varieties
Chapter 3. §3.0. §3.1. §3.2. §3.3. §3.4.
Background: Abstract Varieties Fans and Normal Toric Varieties The OrbitCone Correspondence Toric Morphisms Complete and Proper
iv
Contents
Appendix: Nonnormal Toric Varieties Chapter 4. §4.0. §4.1. §4.2. §4.3. Divisors on Toric Varieties
149 153 153 168 174 187 193 193 203 217 224 230 243 243 260 279 290 301 305 305 309 318 324 334 339 339 350 357 371 379 379 390 401
Background: Valuations, Divisors and Sheaves Weil Divisors on Toric Varieties Cartier Divisors on Toric Varieties The Sheaf of a TorusInvariant Divisor Homogeneous Coordinates on Toric Varieties
Chapter 5. §5.0. §5.1. §5.2. §5.3. §5.4.
Background: Quotients in Algebraic Geometry Quotient Constructions of Toric Varieties The Total Coordinate Ring Sheaves on Toric Varieties Homogenization and Polytopes Line Bundles on Toric Varieties
Chapter 6. §6.0. §6.1. §6.2. §6.3.
Background: Sheaves and Line Bundles Ample Divisors on Complete Toric Varieties The Nef and Mori Cones The Simplicial Case
Appendix: Quasicoherent Sheaves on Toric Varieties Chapter 7. §7.0. §7.1. §7.2. §7.3. Projective Toric Morphisms
Background: Quasiprojective Varieties and Projective Morphisms Polyhedra and Toric Varieties Projective Morphisms and Toric Varieties Projective Bundles and Toric Varieties
Appendix: More on Projective Morphisms Chapter 8. §8.0. §8.1. §8.2. §8.3. The Canonical Divisor of a Toric Variety
Background: Reﬂexive Sheaves and Differential Forms OneForms on Toric Varieties Differential Forms on Toric Varieties Fano Toric Varieties Sheaf Cohomology of Toric Varieties
Chapter 9. §9.0. §9.1. §9.2.
Background: Cohomology Cohomology of Toric Divisors Vanishing Theorems I
1. Chapter 13. §11. §11. §10.3.4. §10.4. §9. §11. Chapter 14. §11.2. §14.4. Vanishing Theorems II Applications to Lattice Polytopes Local Cohomology and the Total Coordinate Ring 413 422 435 448 451 453 453 461 479 489 496 507 507 519 528 540 553 555 557 557 557 557 559 561 Appendix: Introduction to Spectral Sequences Part II.5. Toric Surfaces Singularities of Toric Surfaces and Their Resolutions Continued Fractions and Toric Surfaces Gr¨ bner Fans and McKay Correspondences o Smooth Toric Surfaces RiemannRoch and Lattice Polygons Toric Resolutions and Toric Singularities Chapter 11.1.3. Bibliography .3.2. §10. Resolution of Singularities Other Types of Resolutions Rees Algebras and Multiplier Ideals Toric Singularities The Topology of Toric Varieties The RiemannRoch Theorem GIT and MMP for Toric Varieties Chapter 12. §9.1. Toric Geometric Invariant Theory The Secondary Fan Toric Minimal Model Program The Toric Minimal Model Program Chapter 15. §14.3.2. Topics in Toric Geometry Chapter 10.5. §10. §14.Contents v §9. §10.
Then: (a) F ∨ and hence F ∨∨ are reﬂexive. 339 . OX ).1. A Weil divisor D on a normal variety X gives the sheaf OX (D) deﬁned by Γ(U . It is easy to see that locally free sheaves are reﬂexive. Recall that the dual of a sheaf of OX modules F is F ∨ = HomOX (F . Here are some properties of reﬂexive sheaves.0. Let F be a coherent sheaf on a normal variety X and consider the inclusion j : U0 ֒→ X where U0 is open with codim(X \U0 ) ≥ 2. We say that F is reﬂexive if the natural map F −→ F ∨∨ is an isomorphism.0. (c) If F U0 is locally free. the canonical divisor of a normal toric variety is a Weil divisor that is not necessarily Cartier. Reﬂexive Sheaves. Background: Reﬂexive Sheaves and Differential Forms This chapter will study the canonical divisor of a toric variety.Chapter 8 The Canonical Divisor of a Toric Variety §8. Our ﬁrst task is to characterize these sheaves. The theory developed in Chapters 6 and 7 dealt with Cartier divisors and line bundles. OX (D)) = { f ∈ C(X )∗  (div( f ) + D)U ≥ 0} ∪ {0}. We will also study the sheaves associated to Weil divisors. (b) If F is reﬂexive. then F ∨∨ ≃ j∗ (F U0 ). then F ≃ j∗ (F U0 ). As we will see. Proposition 8.
Theorem 8.2 and Prop. Other properties of the rank will be studied in Exercise 8. Given a F coherent sheaf on irreducible variety X . Recall from §4. a coherent sheaf of rank 1 is reﬂexive if and only if it is a line bundle.6]. This sheaf is p locally free when X is smooth. j∗ G ) = Γ(U ∩U0 . and L ≃ j∗ (L U0 ). This is proved in [79. 1. however. We ﬁrst deﬁne the rank of a coherent sheaf on an irreducible variety X .0 that the direct image j∗ G of a sheaf G on U0 is deﬁned by Γ(U . though it is locally free on the smooth locus of X .1 to create a reﬂexive version of ΩX . Then the following are equivalent. . Then U0 = X \ Y is smooth. the global sections of F ⊗OX KX form a ﬁnitedimensional vector space over C(X ) whose dimension is the rank of F .0. Then F ∨∨ ≃ j∗ ((F ∨∨ )U0 ) = j∗ ((F U0 )∨∨ ) ≃ j∗ (F U0 ). 1.e. reﬂexive sheaves of rank 1 are easy to understand. its singular locus Y = Sing(X ) has codimension at least two in X by Proposition 4. ΩX may be badly behaved. For more on reﬂexive sheaves.1. (G ∨ )U0 = (G U0 )∨ for any coherent sheaf G on X . we ﬁrst observe that restriction is compatible with taking the dual. We now have all of the tools needed to characterize which coherent sheaves on a normal variety come from Weil divisors. (b) There is an open subset j : U0 ֒→ X such that codim(X \ U0 ) ≥ 2.. For part (c). L U0 is a line bundle on U0 . 1. Parts (a) and (b) of the proposition are proved in [79. the rank is just the rank of the associated vector bundle.9].0.0. For X normal. For a locally free sheaf.0.0. the reader should consult [79] and [148]. (c) L ≃ OX (D) for some Weil divisor D on X . i.3.1. Deﬁnition 8. Recall that KX is the constant sheaf on X given by C(X ).4. Reﬂexive Sheaves of Rank One. In the smooth case.0. and the last follows since F U0 is locally free and hence reﬂexive.0. Prop. which implies that L U0 is a line bundle by Proposition 8.0. (a) ⇒ (b) Since X is normal.2. G ) for U ⊆ X open. Hence L ≃ j∗ (L U0 ) by Proposition 8. (a) L is reﬂexive of rank 1. Proof. The Canonical Divisor of a Toric Variety Proof. On a smooth variety. Cor. Let L be a coherent sheaf on a normal variety X . p Later in the section we will study the sheaf ΩX of pforms on X .3.17. Proposition 8. Hence we can use part (c) of p Proposition 8.340 Chapter 8. where the ﬁrst isomorphism follows from parts (a) and (b).
E on a normal variety X . which in turn induces a map (8. This is an isomorphism when D or E is Cartier but may fail to be an isomorphism in general. From examples in previous chapters. we have the following result about the maps (8. Thus j∗ (OX (D)U0 ) ≃ OX (D)∨∨ by Proposition 8.0. In general. note that div( f ) + D ≥ 0 ⇐⇒ (div( f ) + D)U0 ≥ 0 since codim(X \U0 ) ≥ 2. Given Weil divisors D. we get a map OX (D) ⊗OX OX (−D) −→ OX . Let x. we obtain OX (D) ≃ j∗ OU0 (E) = j∗ (L U0 ) ≃ L . Given f ∈ C(X )∗ . In Exercise 8. (c) ⇒ (a) The proof of (b) ⇒ (c) shows that OX (D) ≃ j∗ (OX (D)U0 ). this is an isomorphism when D is Cartier.3). where Di is the Zariski closure of Ei in X .0.0. and the same holds over any open set of X . The line L = V(y.0. z 2 . z] y 2 − xz .0. which is a proper subset of Γ(X . and OX (D)U0 = OU0 (DU0 ) is locally free since U0 is smooth. z ⊆ R. Combining this with E = DU0 . the map f ⊗ g → f g deﬁnes a sheaf homomorphism (8.0. y. . Example 8. Consider the afﬁne quadric cone X = V(y 2 − xz) ⊆ C3 . the image of the map OX (−L) ⊗OX OX (−L) −→ OX (−2L) is y. • Γ(X .2. ♦ As noted in §6. z denote the images of the variables in R.1. OX (−2L)) is the ideal z ⊆ R (principal since −2L is Cartier).0.5.1) OX (D) ⊗OX OX (E) −→ OX (D + E). y.2). The coordinate ring of X is R = C[x.1) and (8. Tensor Products and Duals. Consider the Weil divisor D = i ai Di .§8. OX (−2L)) = z .0. Background: Reﬂexive Sheaves and Differential Forms 341 (b) ⇒ (c) The line bundle L U0 can be written as OU0 (E) for some Cartier divisor E = i ai Ei on U0 .0. It follows that OX (−L) ⊗OX OX (−L) ≃ OX (−2L).2 you will show the following: • Γ(X . OX (−L)) is the ideal y. But codim(X \U0 ) ≥ 2.0. though 2L is Cartier (this follows from Example 4.1) when E = −D.2) OX (−D) −→ OX (D)∨ . so OX (D) ≃ OX (D)∨∨ is reﬂexive and has rank 1 since it is a line bundle on U0 . we know that this is a normal toric surface. z) gives a Weil divisor that is not Cartier. • On global sections. If we apply (8.
22.1 since OX (D + E) is reﬂexive and (8. written D ∼ E. Recall that Weil divisors D and E on X are linearly equivalent. Part (a) for Cartier divisors was proved in Proposition 6. The Canonical Divisor of a Toric Variety Proposition 8.0.0.342 Chapter 8. Then (8. Taking the double dual and using Proposition 8. OX (−D) ≃ OX (D)∨ . Then. (8. Divisor Classes.0. (b) If D is a Weil divisor on X .0. Furthermore. We can now reinterpret Cl(X ) in terms of isomorphism classes of reﬂexive sheaves of rank 1.7.0.0.6 and 8.1) is an isomorphism on the smooth locus of X . Let D. we prove this as follows: OX (D) ≃ OX (E) implies OX (D) ⊗OX OX (−E) ≃ OX (E) ⊗OX OX (−E). if OX (D) is a line bundle on X .0. we get OX (D − E) ≃ OX .6.1) induces an isomorphism (OX (D) ⊗OX OX (E))∨∨ ≃ OX (D + E). (a) If D and E are Weil divisors on X .0. then OX (D) ≃ OX (E) ⇐⇒ D ∼ E.0. where the group operation was tensor product and the inverse was the dual. E are Weil divisors.0.0. .18). we reinterpreted Pic(X ) in terms of isomorphism classes of line bundles. In Chapter 4 we deﬁned the class group Cl(X ) and Picard group Pic(X ) in terms of Weil and Cartier divisors. i. Proof.0. This follows immediately from Propositions 8.e.0. where the group operation is the double dual of the tensor product and the inverse is the dual. then D is Cartier ⇐⇒ OX (D) is a line bundle.0. The ﬁrst isomorphism follows from Proposition 8. in Chapter 6. then Theorem 6. Let X be a normal variety. Proof.0.2) is an isomorphism on the smooth locus. Proposition 8..5. Then we are done since any divisor linearly equivalent to a Cartier divisor is Cartier by Exercise 4.20 shows that OX (D) ≃ OX (E) for some Cartier divisor E. The second isomorphism follows similarly since both sheaves are reﬂexive and (8. One direction of part (b) was proved in Chapter 6 (see Proposition 6.7. Conversely. E be Weil divisors on a normal variety X . Thus D ∼ E by part (a). The only place that we used Cartier in that proof was to show OX (D) ≃ OX (E) =⇒ OX (D − E) ≃ OX . if D = E + div( f ) for some f ∈ C(X )∗ . When D.2) already is an isomorphism.6.17 and Theorem 6.
3). ∂ f1 ∂xn α ··· ∂ fs ∂xn modulo the ideal I (Exercise 8. (8. The generators of I give a surjection R s → I and hence a surjection S s → I/I 2 . . .0. is the Rmodule generated by the formal symbols df for f ∈ R. (b) d( f g) = f dg + g df for all f . we obtain an exact sequence S s −→ S n −→ ΩS/C −→ 0. we begin with the case of a Calgebra. . Let R = C[x1 .4). .0. where [ f ] ∈ I/I 2 maps to 1 ⊗ df ∈ S ⊗R ΩR/C . f . . dxi for all ♦ A Calgebra homomorphism R → S induces a natural homomorphism ΩR/C −→ ΩS/C . Let R be a Calgebra.0.0. xn ] and S = R/I. g ∈ R. fs . When we regard ΩS/C as an Rmodule.11.0. . This presentation of ΩS/C is very useful for computing examples. If R = C[x1 . . n ∂f i=1 ∂xi This follows because the relations deﬁning ΩR/C imply df = f ∈ R (Exercise 8. The module of K¨ hler differentials of R a over C. Example 8. . Here is a case when this map is easy to understand. . . modulo the relations (a) d(c f + g) = c df + dg for all c ∈ C. then n ΩR/C ≃ i=1 R dxi . we obtain a homomorphism of Smodules S ⊗R ΩR/C −→ ΩS/C . Proposition 8. See [120. Thm.8. . Let R → S be a surjection of Calgebras with kernel I.0.2] for a proof. In order to give an algebraic deﬁnition of differential form a on a variety.9. Background: Reﬂexive Sheaves and Differential Forms 343 K¨ hler Differentials.0.0. Then there is an exact sequence of Smodules I/I 2 −→ S ⊗R ΩR/C −→ ΩS/C −→ 0. where α is given by the reduction of the n × s Jacobian matrix ∂ f1 ∂ fs ∂x1 · · · ∂x1 . where I = f1 . . Combining this with Proposition 8. xn ].10. . denoted ΩR/C . Example 8. .0.§8. .9. Deﬁnition 8. 25. .0.3) . ♦ . g ∈ R.10 and Example 8.
the vector bundles corresponding to Ω1 and TX X are called the cotangent bundle and tangent bundle respectively.0.0.8. Theorem 8. II.14. The Canonical Divisor of a Toric Variety K¨ hler differentials also behave well under localization. these sheaves are nicely behaved. Proposition 8. A smooth cone σ ⊆ NR ≃ Rn of dimension r gives the afﬁne toric variety Uσ ≃ Cr × (C∗ )n−r ⊆ Cn .8.11 and Proposition 8.0. Thm. Then ΩR f /C ≃ ΩR/C ⊗ R f .0. In Exercise 8.0. II. II. ♦ It follows immediately that ΩXΣ Σ We know from Chapter 6 that a locally free sheaf is the sheaf of sections of a vector bundle. n = dim X .9 and Proposition 8. as you will prove in a Exercise 8. 1 is locally free for any smooth toric variety X . When X is smooth.12 imply that ΩUσ is locally free of rank n. Deﬁnition 8. A variety X is smooth if and only if Ω1 is locally free. When U = Spec(R) is an afﬁne open of X .8. This can also be described in terms of derivations—see Exercises 8.0. When this X happens.8] for a slightly X different approach to deﬁning the sheaf Ω1 . X X The reason for the superscript in the notation for the cotangent sheaf will become clear later in this chapter. See [78.15.0.6 you will use Example 8. Ω1 and TX are locally free sheaves of rank n. it is easy to see that the cotangent sheaf is locally free. Cotangent and Tangent Sheaves. and [78. the deﬁnition of the tangent sheaf implies that TX (U ) = HomR (ΩR/C . Let R f be the localization of R at a nonnilpotent element f ∈ R.8. Let X be a variety. The cotangent sheaf Ω1 is the sheaf of X OX modules deﬁned via Ω1 (U ) = ΩOX (U)/C X on afﬁne open sets U .0.0.0.7–8. X In the smooth toric case.0.5.12 to show that Ω1 is a coherent sheaf.9.12. R).0.13. Now we globalize Deﬁnition 8. OX ). Comment 8. Example 8.0.15].2] for the X connection between these methods. When X is smooth. as shown by the following result from [78. 1 Then Example 8.344 Chapter 8. . The tangent sheaf TX is the dual sheaf TX = (Ω1 )∨ = HomOX (Ω1 .
y) coordinates. y−1 ]). Recall that P 2 has a a cover by the afﬁne open sets Uσ0 ≃ Spec(C[x. where σ0 .§8. the coordinates (a1 . 2 1/a −a2 /a1 1 1 a2 . σ2 are the maximal cones in the usual fan for P 2 . we obtain φ10 = A check shows that φ10 = φ12 ◦ φ20 . 0 −1/y 2 Next. −y/x 2 y/x −y/x 2 1/x . i. yielding φ20 = 1/y −x/y 2 . a2 ) are represented in terms of the (x. Thus a 1form on C2 may be written uniquely as f1 dx + f2 dy. On Uσ2 we identify (a1 . Things get messy if we keep everything in (x. More precisely.16.e. The module ΩR/C is a free Rmodule of rank 2 with generators dx. dx and dy transform via the Jacobian matrix described in Example 8. y−1 ). and then translate back. Background: Reﬂexive Sheaves and Differential Forms 345 Example 8. computing φ10 directly. φik = φi j ◦ φ jk . so we ﬁrst translate to coordinates (a1 .0. Since det(φi j ) is invertible on Uσi ∩Uσ j . x−1 ) = So in terms of (a1 . y]) Uσ1 ≃ Spec(C[yx−1 . Then Uσ1 has coordinates (yx−1 . We construct the cotangent bundle for P 2 . x−1 ]) Uσ2 ≃ Spec(C[xy−1 . To generalize this to P 2 . a2 ) on Uσ2 . the matrix for the transition function φi j will be the Jacobian of the map Uσ j → Uσi . y) yields φ12 = −y 2 /x 2 0 . a2 ). a2 ) with (xy−1 . we have φ12 = 2 −1/a1 0 . −1/x 2 0 Finally. j.0. P .0. where fi ∈ R. Hence we obtain ♦ a rank 2 vector bundle on P 2 whose sheaf of sections is Ω1 2 . y]. dy by Example 8. y−1 ). Similar computations show that the compatibility criteria are satisﬁed for all i. σ1 . the same is true for φi j .11. .9.0. Let C2 = Spec(R) for R = C[x. y) coordinates on Uσ0 as (xy−1 . a1 a1 Rewriting this in terms of (x. On Uσ2 . we require that after changing coordinates.. we compute φ12 . k.
p /C .0.17. which we write more informally as 0 −→ IY −→ OX −→ OY −→ 0.0. The following basic result is proved in [78.p C X. Prop. This explains the description of Tp (V ) given in Lemma 1.p /m2 . so that Example 8. Now take p ∈ V and tensor with OV . . II.17]. . . ♦ δ Conormal and Normal Sheaves.p This omits many details but should help you understand why TX is the correct deﬁnition of tangent sheaf. C) = Tp (X ).p /mX.p /m2 −→ ΩOX.8.9).4) and the isomorphism ΩOX.0. the exact sequence of Proposition 8. X.p is the local ring of X at p.p −→ ΩOV . Given a closed subvariety i : Y ֒→ X .p to obtain the exact sequence s n OV .p /C ⊗OX.p /C −→ 0 (Exercise 8. The stalk (TX ) p of the tangent sheaf at p ∈ X can be described as follows.p −→ OV . Let V ⊆ Cn be deﬁned by I = I(V ) = f1 . II.0. . it is natural to ask how their cotangent sheaves relate.11 gives the exact sequence S s −→ S n −→ ΩS/C −→ 0. xn ]. . If we tensor this with C and dualize. . taking the dual of the above isomorphism gives X (8.6. .p C ≃ HomC (mX. We begin with the exact sequence 0 −→ IY −→ OX −→ i∗ OY −→ 0. Example 8. as does F ⊗OX OY for any sheaf F of OX modules.p .0.p /mX. C) given in Chapter 1. The Canonical Divisor of a Toric Variety Relation with the Zariski Tangent Space.p /C ⊗OX. .9). Since TX is dual to Ω1 .p C ≃ mX.p 0 −→ Tp (V ) −→ Cn −→ Cs . ∂ where δ comes from the s × n Jacobian matrix ∂xfij (p) (Exercise 8.346 Chapter 8. .4) (TX ) p ⊗OX. II.10 gives a surjection mX. X where OX. fs ⊆ C[x1 . The coordinate ring of V is S = C[x1 .0. . . xn ]/I.0.12 and Thm. . Since OX.8.8. The a stalk of Ω1 at p is the module of K¨ hler differentials X (Ω1 ) p = ΩOX. . The deﬁnition of the tangent sheaf TX seems far removed from the deﬁnition of the Zariski tangent space Tp (X ) = 2 HomC (mX. (8. The quotient sheaf IY /IY2 has a natural structure as sheaf of OY modules.p /m2 give the exact sequence X.p which is an isomorphism of vector spaces over C by [78.0. Prop.7]. Here we explain (without proof) the connection.p ≃ C and ΩC/C = 0 (easy to check).
X For any sheaf F of OX modules. this sheaf is locally free on the smooth locus of X .0. . In particular. ♦ It follows that p ΩC n is free of rank p More generally. and we deﬁne the sheaf of X pforms to be the wedge product p ΩX = p Ω1 . Hence we can use Proposition 8. the sheaf of p pforms ΩX may fail to be locally free. Ωn is a line bundle in this case.5) p p p ΩX = (ΩX )∨∨ = j∗ ΩU0 . .14 implies that ΩX is locally free of rank n p when X is smooth of dimension n. it contains the tangent bundle of Y with quotient given by the normal bundle. Let Y be a closed subvariety of a variety X . Then: (a) There is an exact sequence of OY modules 1 IY /IY2 −→ Ω1 ⊗OX OY −→ ΩY −→ 0. Theorem 8. This is the algebraic analog of what happens in differential geometry. Note that part (a) of this theorem is a global version of Proposition 8. Then ΩCn is the sheaf associated to i=1 p ΩR/C = 1≤i1 <···<i p ≤n n p R dxi1 ∧ · · · ∧ dxin . For a normal variety X . R = C[x1 .19. The vector bundle associated to NY /X is the normal bundle of Y in X . . X (b) If X and Y are smooth.0. where the normal bundle is the orthogonal complement of the tangent bundle of Y . Differential Forms.18. We call Ω1 the sheaf of 1forms. the exterior power p F is the sheaf associated to the presheaf which to each open set U assigns the OX (U )module p F (U ).1 to deﬁne the sheaf!of Zariski pforms (8.18 to obtain the exact sequence 0 −→ TY −→ TX ⊗OX OY −→ NY /X −→ 0.0. X Zariski pForms and the Canonical Sheaf . . and the complement of the smooth locus has codimension ≥ 2 since X is normal. then this sequence is also exact on the left and IY /IY2 is locally free of rank equal to the codimension of Y . Background: Reﬂexive Sheaves and Differential Forms 347 Theorem 8. Then the above sequence says that when the tangent bundle of X is restricted to the subvariety Y . OY ) the normal sheaf of Y in X .§8. Ω1 n is the sheaf associated to the free Rmodule ΩR/C = C p n R dxi . Example 8.0. .0. When X and Y are smooth. xn ].0. For Cn . .0. we can dualize the sequence appearing in Theorem 8.10. However. We call IY /IY2 the conormal sheaf of Y in X and call its dual NY /X = (IY /IY2 )∨ = HomOY (IY /IY2 .0.
When X is projective. We call this divisor a canonical divisor of X . There are also singular varieties whose canonical divisors are Cartier—these are the Gorenstein varieties to be studied later in the chapter. another approach is given in [78. similar to way that C(X ) gives the constant sheaf KX of rational functions on X . The rank of a coherent sheaf on an irreducible variety was deﬁned in Deﬁnition 8.0. This is a reﬂexive sheaf of rank 1.0. It follows that ΩX is a n reﬂexive sheaf of rank p . we call ωX the canonical bundle since it is a line bundle. where n = dim X . We can regard n ΩC(X)/C as the constant sheaf of rational nforms on X .0.5) is valid for any smooth open subset U0 ⊆ X whose complement has codimension ≥ 2.2.0. The canonical sheaf of a normal variety X is ωX = Ωn . Proposition 8. In this case. the canonical divisor is Cartier. In the toric case.0. The Canonical Divisor of a Toric Variety p where j : U0 ֒→ X is the inclusion of the smooth locus of X . where C(X ) is the ﬁeld of rational functions on X . 8. so that ωX ≃ OX (D) for some Weil divisor D on X . Exercises for §8. n You will prove the following result in Exercise 8. (8. though the deﬁnition is more sophisticated (see [112. One such construction uses n ΩC(X)/C . Here are some properties of the rank. For later purposes.348 Chapter 8.0. While it often sufﬁces to know ωX up to isomorphism. image of the map Ωn → X ωX ⊆ n ΩC(X)/C is ΩC(X)/C .0. There is an obvious sheaf map Ωn −→ X n ΩC(X)/C . there are situations where a unique model of ωX is required.0. Proposition 8. known as the canonical class of X .7 shows that the canonical divisor KX is welldeﬁned up to linear equivalence and hence gives a unique divisor class in Cl(X ). X where n is the dimension of X .7].20. When X is normal. .21. often denoted KX .1.0. where ωX is called the dualizing sheaf. The case p = n is especially important. The canonical sheaf can be deﬁned for any irreducible variety X as a subsheaf of n ΩC(X)/C .10. III. When X is smooth.1. we note that by Proposition 8. We will see in Chapter 9 that ωX plays a key role in Serre duality. we will see later in the chapter that there is a natural choice for the canonical divisor. §9]). Deﬁnition 8.
Then show that f ∈ m if and only if f ∈ m and ∂f ∂ti (1) = 0 for all i.0.0. Hint: ΩX is locally free when restricted to the smooth locus of X. Hint: Pick a basis e1 . Continuing Exercise 8.9 we claimed that df = for all f ∈ R. M). . Prove Proposition 8. . 8.7 to construct an Rmodule isomorphism DerC (R. .2. Given a Calgebra R and an Rmodule M. .0.5. we let DerC (R. so that f is 2 a Laurent monomial in t1 . (b) Show that the map of part (a) is an isomorphism when X is irreducible and F is a constant sheaf. Explain why d : R → ΩR/C is called the universal derivation. Show that f ∈ m2 if and only if m cm = 0 in C and ei m cm m = 0 in MC . This is an Rmodule where (rδ)( f ) = rδ( f ).4. In Example 8.7.0. then φ ◦ d : R → M is a Cderivation. Background: Reﬂexive Sheaves and Differential Forms 349 (a) Let R be an integral domain with ﬁeld of fractions K. (b) Let F be a coherent sheaf on X let U ⊆ X be an nonempty open subset. Let j : U ֒→ X be the inclusion of a nonempty open subset of a variety X.9. Prove that F and F U have the same rank.13 is a coherent sheaf. (b) More generally. (a) Show that there is a sheaf map F → j∗ (F U ) for any sheaf F on X. a Cderivation δ : R → M is a Clinear map that satisﬁes the Leibniz rule. g ∈ R.0.11 comes from the Jacobian matrix (8.11 below will be useful. (b) Let TX be the tangent sheaf of a variety X and let U = Spec(R) be an afﬁne open subset of X. 8. R). Exercise 8.11. 8.8. n ∂f i=1 ∂xi dxi in ΩR/C 8. (a) Let f = m cm χ m ∈ C[M].0. 8.0.0.0.0.6.0. Prove that the cotangent sheaf Ω1 deﬁned in Deﬁnition 8. Show that M ⊗R K is a ﬁnitedimensional vector space over K whose dimension equals the rank of the coherent sheaf M on Spec(R).17. Let 1 ∈ TN be the identity element of the torus TN and let m ⊆ C[M] be the corresponding maximal ideal.e. . Prove this. . Prove Proposition 8. Fill in the details omitted in Example 8. Prove that the map α in the exact sequence from Example 8. 8. δ( f g) = f δ(g) + gδ( f ) for all f .0.0. i. and conclude that the Zariski tangent space of TN at the identity is naturally isomorphic to NC . . Prove the claims made in Example 8. Prove that TX (U) = DerC (R. . 8. and let M be a ﬁnitely generated Rmodule.5.0.0..0.10. Set NC = N ⊗Z C and MC = M ⊗Z C. .0.0.12. .0.0. 8. M) ≃ HomR (ΩR/C . . xn ].7.§8.tn . 1 8.0. (c) Let 0 → F → G → H → be an exact sequence of sheaves on X.3. (a) Show that f → df deﬁnes a Cderivation d : R → ΩR/C .0. show that if φ : ΩR/C → M is an Rmodule homomorphism.3). Let R = C[x1 . M) denote the set of all Cderivations δ : R → M.0. . X 8. en of M and set ti = χ . . (a) Use part (b) of Exercise 8.12. Prove that rank(G ) = rank(F ) + rank(H ).21. 8. (b) Use part (a) to construct an isomorphism m/m2 ≃ MC .
It follows that (8. uρ .1.1) and dualizing. Ωn ).2) β : M ⊗Z OXΣ −→ ρ ODρ . See Exercise 8. F). We also have a canonical map (8. Prop.7]. m Ω1N ≃ M ⊗Z OTN .1. Let F be a free module of rank n over a ring R and ﬁx 0 ≤ p ≤ n.12.0 we write as ODρ . the inclusion i : Dρ ֒→ XΣ gives the sheaf i∗ ODρ on XΣ .1. OX ). X (b) Show that if X is a normal variety of dimension n. T The First Exact Sequence. The Canonical Divisor of a Toric Variety 8. X (c) In a similar vein. the 1form dχm is a global section of ΩTN that maps to m ⊗ 1 χ in (8. 1 As a consequence. The group action transports the tangent space NC over the whole torus. Using the map M → Z given by m → m.1.1) and hence is invariant under the action of TN .0.3) α : Ω1 Σ −→ M ⊗Z OXΣ X . we obtain TTN ≃ N ⊗Z OTN = NC ⊗C OTN . Now consider the toric variety XΣ of the fan Σ. X X The Torus. where the direct sum is over all ρ ∈ Σ(1). Then the map ΩC[M]/C −→ M ⊗Z C[M] deﬁned by dχ m → m ⊗ χ m is easily seen to be an isomorphism. ωX ). Thus its tangent space at the identity is N ⊗Z C = NC via the exponential map. This makes intuitive sense since TN = N ⊗Z C∗ as a complex Lie group.1. OneForms on Toric Varieties In this section we will describe two interesting exact sequences that involve the sheaves Ω1 Σ and Ω1 Σ on a normal toric variety XΣ . then show that ΩX ≃ H omOX (ΩX .0.1 for more on invariant 1forms on the torus. The coordinate ring of the torus TN is the semigroup algebra C[M].1. we obtain the composition M ⊗Z OXΣ −→ Z ⊗Z OXΣ = OXΣ −→ ODρ . X §8.13. p n−p (a) If X is smooth of dimension n. For ρ ∈ Σ(1). which explains the above trivialization of the tangent bundle TTN . This gives a natural map (8. which following §8. 4. Hint: If you get stuck. see [42. then there is an isomorphism p Ωn−p ≃ H omOX (ΩX . as shown in Exercise 8. show that the tangent sheaf TX of a normal variety X satisﬁes TX ≃ H omOX (Ω1 . Recall that wedge n−p p n product induces an isomorphism F ≃ HomR ( F.350 Chapter 8. This is also true algebraically.
α is deﬁned by dχ m ∈ ΩC[σ∨ ∩M]/C −→ m ⊗ χ m ∈ M ⊗Z C[σ ∨ ∩ M].2. .0.2) and (8.1. xr+1 . .3) (Exercise 8.1. The exact sequence of Theorem 8. xn are units in R.3) is exact. . xn ]. div(χ m )U ≥Dρ U σ σ α β ρ ODρ −→ 0 m∈σ∨ ∩M. and the 1forms on Uσ form the free Rmodule ΩR/C = n i=1 R dxi by Example 8. . where r ≤ n and e1 . OXΣ (−Dρ )) = C · χm = C · χ m. These C[σ ∨ ∩ M]module homomorphisms ΩC[σ∨ ∩M]/C → M ⊗Z C[σ ∨ ∩ M] patch to give the desired map (8.1). we see that α can be regarded as the map n n ΩR/C = i=1 R dxi −→ M ⊗R R = i=1 R that sends n i=1 fi dxi to ( f1 x1 . the sequence 0 −→ Ω1 Σ −→ M ⊗Z OXΣ −→ X formed using (8. . .1. Since σ is smooth. .4) Iρ = Γ(Uσ . This gives the exact sequence n r 0 −→ ΩR/C −→ i=1 R −→ i=1 R/ xi −→ 0 since xr+1 .0. xn denote the characters of the corresponding dual basis of M. Since α takes dxi to ei ⊗ xi . we have (8. . . Then Uσ = Cr × (C∗ )n−r .2). . OneForms on Toric Varieties 351 constructed as follows.28 and 4. the map α of (8.3) reduces to the isomorphism (8.§8.1.1. Note that over the torus TN . Proof. we may assume σ = Cone(e1 . .uρ >0 Over Uσ .1. . en . Theorem 8. . to m.1. . it vanishes since χ m ∈ Iρ in this case. . . . uρ χ m ∈ C[σ ∨ ∩ M]/Iρ . Logarithmic Forms. . . On the afﬁne piece Uσ = Spec(C[σ ∨ ∩ M]). . By Propositions 4. . and the theorem follows. xr .1. Let x1 . .12. er ). . . .3.1 has a lovely interpretation in terms of residues of logarithmic 1forms. For a smooth toric variety XΣ . The idea is that M ⊗Z OXΣ can be thought of as the sheaf Ω1 Σ (log D) of 1forms on XΣ with logarithmic poles along X D = ρ Dρ . and if m.9 and Proposition 8. . .0. We now verify that the sequence is exact over Uσ . . uρ = 0. the composition Ω1 Σ → M ⊗Z OXΣ → ODρ takes a 1form dχ m . the subvariety Dρ ∩Uσ ⊆ Uσ is deﬁned by the ideal Iρ = I(Dρ ∩Uσ ) ⊆ C[σ ∨ ∩ M]. m. We ﬁrst verify that β ◦ α is the zero map. . . The coordinate ring of Uσ is R = ±1 ±1 C[x1 . . . This is obviously zero if m. m ∈ X σ ∨ ∩ M. . On the afﬁne piece Uσ ⊆ XΣ . We begin with an example.1. also denoted e1 . en is a basis of N. fn xn ).1. uρ = 0.1.
p. . . The coordinate ring of Cn is R = C[x1 . More precisely. m. A divisor D = i Di on X has smooth normal crossings if every Di is smooth and irreducible. C C This construction works for any smooth afﬁne toric variety Uσ .3) comes from the inclusion of 1forms. if I p = {i  p ∈ Di }.5) Ω1 Σ (log D) ≃ M ⊗Z OXΣ X such that the map α of (8.352 Chapter 8. The formal calculation to be ΩCn dχ m = χm shows that the map dχ m χm n i=1 n dxi i=1 R xi . Let f (z) be an analytic (also called holomorphic) e function deﬁned in a punctured neighborhood of a point p ∈ C. The Canonical Divisor of a Toric Variety Example 8. Now introduce some denominators: a rational 1form ω has logarithmic poles along D if n dxi fi ω= . fi ∈ R. ΩR/C = i=1 R dxi .1.2. A nice discussion of Ω1 (log D) for complex manifolds X can be found in [72. 449]. codim i∈Ip Tp (Di ) = I p . we have a canonical isomorphism (8.e.1. i. Let X be a smooth X variety. . Furthermore. xn ]. ei i=1 dxi xi → m ⊗ 1 induces an isomorphism of sheaves Ω1 n (log D) ≃ M ⊗Cn OCn C such that the map α : Ω1 n → M ⊗Cn OCn of (8. C .1. and the divisor D is n the sum of the coordinate hyperplanes Di = V(xi ).. and the sheaves of logarithmic 1forms on Uσ patch to give the sheaf Ω1 Σ (log D) for any smooth X toric variety XΣ .1. and for every p ∈ X . xi These form the free Rmodule and the corresponding sheaf is deﬁned 1 (log D).3) is induced by the inclusion of C ♦ 1forms Ω1 n ֒→ Ω1 n (log D). As above. The Poincar´ Residue Map. we require that the tangent spaces Tp (Di ) ⊆ Tp (X ) meet transversely. The construction of Ω1 Σ (log D) can be done more generally. the divisor D = ρ Dρ is a smooth normal crossing divisor on any smooth toric variety XΣ . Take a counterclockwise loop C around p on X such that f (z) has no other poles inside C. For example. . the divisors containing p meet nicely. Then the residue of the 1form ω = f (z) dz at p is the contour integral res p (ω) = 1 2πi ω.
Then: (a) The sequence 0 −→ Ω1 Σ −→ M ⊗Z OXΣ −→ X is exact. . xn ) represents the class of f in R/ x1 . Let XΣ be a normal toric variety. xn ) as the “residue” of ω at V(x1 ). Since M ⊗Z OXΣ is reﬂexive. x2 . Doing this for every variable shows that the map Ω1 n (log D) ≃ M ⊗Z OCn −→ C i=1 β n ODi ♦ i Di can be interpreted as a sum of “residue” maps. and is equal to g(0). ρ ODρ ρ ODρ −→ 0 . we get the logarithmic 1form ω = f dx1 . this gives The Normal Case. p. i ρ Dρ .1. . . . . .1. so that 0 −→ Ω1 Σ −→ M ⊗Z OXΣ −→ X is exact.1. if X is a smooth variety and D = ing divisor.5). x2 .1 that uses the sheaf Ω1 Σ of Zariski 1forms in place of Ω1 Σ . .3. . we get an analog of Theorem 8. When there are several variables.1.4. . xn ) ∈ R = C[x1 . . When applied to a smooth toric variety XΣ and the divisor D = the exact sequence of Theorem 8. In terms of the above discussion of residues. one can deﬁne the Poincar´ residue map e Pr : Ω1 (log D) −→ X i a smooth normal cross ODi (see [144. then the residue z theorem tells us that res p (ω) is the coefﬁcient of 1 in the Laurent series for f (z) at z 0. with g(z) analytic at zero. . Example 8. we can regard x1 f (0. Given f = f (x1 . .§8.1.1 via the isomorphism (8. (b) If XΣ is simplicial. . Note also that f (0. if p = 0 and f (z) = g(z) . we can do the same construction by working one variable at a time. X X taking the double dual of (8. Here is an example. More generally. 254]) such that we have an exact sequence (8.3) gives a map Ω1 Σ −→ M ⊗Z OXΣ . When XΣ is normal. then the map on the right is surjective. . xn ].6) r 0 −→ Ω1 −→ Ω1 (log D) −→ X X P ODi −→ 0. . Note that the 1form ω = f (z) dz = g(z) dz has a logarithmic z pole at p = 0. X Theorem 8. OneForms on Toric Varieties 353 In particular.1.1.1. .
so that applying j∗ gives the exact sequence 1 0 −→ j∗ ΩU0 −→ j∗ (M ⊗Z OU0 ) −→ ρ j∗ ODρ ∩U0 1 since j∗ is left exact (Exercise 8. The desired surjectivity now follows easily. When XΣ has no torus factors. uρ′ = 0 for all ρ′ = ρ in σ(1).1.1. S . uρ = 0 and m.1. However. where the ﬁrst map comes from m → m.3 that graded modules over the total coordinate ring S = C[xρ  ρ ∈ Σ(1)] give quasicoherent sheaves on XΣ . Note that U0 is a smooth toric variety whose fan has the same 1dimensional cones as Σ.4. we obtain the following result. Hence we have an exact sequence of graded Smodules (8. In Exercise 8. This gives a homomorphism M ⊗Z S → ρ S/ xρ .4 you will show that the maps M ⊗Z OXΣ → j∗ ODρ ∩U0 factor M ⊗Z OXΣ −→ ODρ −→ j∗ ODρ ∩U0 . Hence we get an exact sequence 0 −→ Ω1 Σ −→ M ⊗Z OXΣ −→ X ρ j∗ ODρ ∩U0 .0. we learned in §5. It remains to show that M ⊗Z OXΣ → cial.0.1.7) 0 −→ Ω1 −→ M ⊗ S −→ S ρ S/ xρ . It is easy to describe a graded Smodule that gives Ω1 Σ . Fix ρ ∈ σ(1) and pick m ∈ M such that m. j∗ ΩU0 = Ω1 Σ by the remarks X following (8. The Canonical Divisor of a Toric Variety Proof. When XΣ has no torus factors. Given σ ∈ Σ. Then m ⊗ 1 maps to a nonzero constant function on ODρ ∩Uσ and to the zero function on ODρ′ ∩Uσ for ρ′ = ρ.1. where ODρ → j∗ ODρ ∩U0 is injective. uρ and the second map is obvious.3). We begin with X M ⊗Z S −→ Z ⊗Z S = S −→ S/ xρ . and j∗ (M ⊗Z OU0 ) = M ⊗Z OXΣ by Proposition 8.5). we have an exact sequence 1 0 −→ ΩU0 −→ M ⊗Z OU0 −→ ρ ODρ ∩U0 −→ 0. Such an m exists since σ is simplicial. and we deﬁne Ω1 to be the kernel S of this map. Let j : U0 ⊆ XΣ be the inclusion map for U0 = ρ Uρ .354 Chapter 8.5.1. By Theorem 8. Using Example 6. and codim(X \ U0 ) ≥ 2 by the OrbitCone Correspondence.1.10 and Theorem 8. ΩXΣ is the sheaf associated to the graded Smodule Ω1 . The exact sequence of part (a) then follows immediately.1. 1 Corollary 8. we need to show that M ⊗Z OUσ −→ ρ ODρ is surjective when XΣ is simpli ρ∈σ(1) ODρ ∩Uσ is surjective.0.
4 and is exact since XΣ is simplicial. and tensoring this with OXΣ gives the middle column.3. if XΣ is smooth. The third row is the obvious exact sequence that uses the identity map on Cl(XΣ ) ⊗Z OXΣ .6.1. and the middle row is the direct sum of these exact sequences.28. Thm. {uρ  ρ ∈ Σ(1)} spans NR . due to Jaczewski [99] in the smooth case and Batyrev and Cox [10. Also. upon which we will base all future calculations involving differentials on projective varieties.” Of course. Hartshorne constructs an exact sequence (8. P called the Euler sequence of P n . i. by Proposition 4. He goes on to say “This is a fundamental result. Consider the following diagram: 0 0 0 / / Ω1 XΣ ρ OXΣ (−Dρ ) 0 / M ⊗Z OX Σ / / ρ OXΣ 0 ρ ODρ /0 / ρ ODρ /0 /0 0 / Cl(XΣ ) ⊗Z OX Σ / Cl(XΣ ) ⊗Z OX Σ /0 0 0 0 The top row is from Theorem 8. we have the exact sequence 0 −→ M −→ ρ Z −→ Cl(XΣ ) −→ 0 from Theorem 4. 12.1. P n is toric. Then there is an exact sequence 0 −→ Ω1 Σ −→ X ρ OXΣ (−Dρ ) −→ Cl(XΣ ) ⊗Z OXΣ −→ 0. Let XΣ be a simplicial toric variety with no torus factors. Furthermore. each ρ ∈ Σ(1) gives an exact sequence 0 −→ OXΣ (−Dρ ) −→ OXΣ −→ ODρ −→ 0. Since XΣ has no torus factors. Proof.8.0.. OneForms on Toric Varieties 355 The Euler Sequence.1. Thm.1. and there is a toric generalization of this result. II. Theorem 8.e.1. and one can check without difﬁculty that the solid arrows in the diagram commute (Exercise 8. In [78. .§8. then the sequence can be written 0 −→ Ω1 Σ −→ X ρ OXΣ (−Dρ ) −→ Pic(XΣ ) ⊗Z OXΣ −→ 0.5).1.1] in the simplicial case.13].8) 0 −→ Ω1 n −→ OPn (−1)n+1 −→ OPn −→ 0. The column on the right is the another obvious exact sequence.
deﬁne ∂u : C[M] → C[M] by ∂u (χ m ) = m.8.1.1. C[M]) (b) Let x1 .1. . In the proof of Theorem 8. X 8.3. x x 8. .1. Use part (a) to give a new proof of Proposition 6. We will study invariant 1forms and derivations on the torus. . . We will use it in the next section to determine the canonical sheaf of XΣ . (c) Dualizing. .1. .1.8).6 contains the square M ⊗Z OXΣ / ρ OXΣ ρ ODρ / ρ ODρ .1. xn be the characters corresponding to the elements of M dual to some par±1 ∂ ±1 ticular basis e1 . .1. Prove that ∂u ∈ DerC (C[M]. 8.6 is the (generalized) Euler sequence of the toric variety XΣ . (a) For u ∈ N. C[M]) give the global sections of the tangent sheaf TTN .0. .2. . (a) Prove that 0 → f∗ F → f∗ G → f∗ H is exact on Y . .8 that the derivations DerC (C[M]. en of N. Consider the afﬁne toric variety Uσ = Spec(C[σ ∨ ∩ M]). . Describe the maps in this square carefully and prove that it commutes.0.3). xn ∂xn . prove that these homomorphisms patch together to give the map α : Ω1 Σ −→ M ⊗Z OXΣ in (8. xn ]. 8. where ODρ → j∗ ODρ ∩U0 is injective. Let 0 → F → G → H → 0 be an exact sequence of sheaves on X and let f : X → Y be a morphism. . dxnn . .5. conclude that Γ(TN . Thus C[M] = C[x1 . The Euler sequence also encodes relations generalizing the classical Euler relation for homogeneous polynomials (see Exercise 8. we know from Exercise 8.4. Since the torus TN = Spec(C[M]) is afﬁne. .1. The Canonical Divisor of a Toric Variety Then commutativity and exactness imply the existence of the dotted arrows in the diagram. . . . . Prove that ∂ei = xi ∂xi and ∂ ∂ that Γ(TN . Exercises for §8. (b) For a toric variety XΣ . . The exact sequence of sheaves in Theorem 8.356 Chapter 8. . (b) Suppose that Y = {pt} and f : X → Y is the obvious map. Ω1N ) is the free C[M]module generated by the TN T invariant differentials dx11 . The proof of Theorem 8. .1. which give the desired exact sequence by a standard diagram chase. (a) Prove that the map dχ m ∈ ΩC[σ∨ ∩M]/C −→ m ⊗ χ m ∈ M ⊗Z C[σ ∨ ∩ M] deﬁnes a C[σ ∨ ∩ M]module homomorphism. show that the map M ⊗Z OXΣ → j∗ ODρ ∩U0 factors M ⊗Z OXΣ −→ ODρ −→ j∗ ODρ ∩U0 .4.1. u χ m .1. TTN ) is the free C[M]module generated by x1 ∂x1 . 8.
txd ) = t d f (x1 .1. . Sometimes the name Euler sequence is used to refer to an exact sequence for the tangent sheaf TXΣ of a smooth toric variety. (a) Prove (8. ∂xρ Hint: Follow what you did for part a. you will prove this relation and consider generalizations encoded by the generalized Euler sequence from a toric variety. {uρ  ρ ∈ Σ(1)} spans NR ). recall from Chapter 5 that given a toric variety XΣ with no torus factors (i.8). . (b) To see how the classical Euler relation generalizes. xn ) with respect to t. The classical Euler relation is the equation n (8. Hint: Differentiate the equation f (tx1 . .1. Differential Forms on Toric Varieties p p We now study the sheaves of pforms ΩXΣ and Zariski pforms ΩXΣ of a toric variety XΣ .1.2.8.1. graded by Cl(XΣ ). (a) Show that for P n .2. Let f be a homogeneous polynomial of degree d in C[x1 .9) i=1 xi ∂f = d· f. . . (b) What is the corresponding sequence for a general smooth toric variety XΣ for Σ as in Theorem 8. X Properties of Wedge Products. .1. Differential Forms on Toric Varieties 357 8.6? 8. The graded pieces Sβ for β ∈ Cl(XΣ ) consist of homogeneous polynomials as described by (5. which is the case X = P n−1 . we have an exact sequence 0 −→ OPn −→ OPn (1)n+1 −→ TPn −→ 0. . n = dim XΣ .1.1) from Chapter 5.8) when X = P n . Z) and f ∈ Sβ show that we have a generalized Euler relation φ([Dρ ]) xρ ρ∈Σ(1) ∂f = φ(β) · f .1. there will be several distinct generalized Euler relations on homogeneous elements of S. Hint: Use (8.7. .1. xn ]. (c) When Cl(XΣ ) has rank greater than 1. .1. Recall from §8.0 that the canonical sheaf is ωXΣ = Ωn Σ . ∂xi In this exercise. . we have the total coordinate ring S = C[xρ  ρ ∈ Σ(1)]. . If φ ∈ HomZ (Cl(XΣ ). Compute the Euler relations on X = P 1 × P 1 .§8. . ..2.6.e. We will need the following properties of wedge products of free Rmodules. §8. 8.9). Show that the Euler sequence from Theorem 8.6 reduces to (8.
Now recall that ωX = X X n 1 1 ΩX and ωY = m ΩY and that the normal sheaf of Y ⊆ X is NY /X = (IY /IY2 )∨ .2. Example 8. For the isomorphism of part (b).2. Let F.1). then one checks that m α : m F → m G is injective and n β : n G → n H is surjective. Theorem 8. This map is welldeﬁned and gives the desired isomorphism (Exercise 8. G. H be free Rmodules of ﬁnite rank.2.2. the canonical sheaf ωXΣ is given by ωXΣ ≃ OXΣ − Thus KXΣ = − ρ Dρ ρ Dρ . we obtain n (Ω1 ⊗OX OY ) ≃ X n−m (IY /IY2 ) ⊗OY m 1 ΩY .2. then rank G = m + n and there is a natural isomorphism (8. Then µ ⊗ ν ∈ m F ⊗R n H maps to m α(µ) ∧ ν ′ ∈ m+n G. is a torusinvariant canonical divisor on XΣ . Then rank G = m + n and there is a natural isomorphism m+n G≃ m F ⊗R n H. and let n = dim X . Then we have the exact sequence 1 0 −→ IY /IY2 −→ Ω1 ⊗OX OY −→ ΩY −→ 0 X from Theorem 8. ♦ This isomorphism is called the adjunction formula. Hence the above isomorphism implies ωY ≃ ωX ⊗OX n−m NY /X . (b) Let 0 → F → G → H → 0 be an exact sequence with rank F = m and rank H = n.3.358 Chapter 8.1). where IY ⊆ OX is the ideal sheaf of Y . Our ﬁrst major result gives a formula for the canonical sheaf of a toric variety.1. as is the rank assertion in part (b). By (8. The Canonical Divisor of a Toric Variety Proposition 8. . For a toric variety XΣ .1 for an explicit description of p φ). If the maps in the exact sequence are α : F → G and β : G → H. Suppose that Y ⊆ X is a smooth subvariety of a smooth variety. (a) An Rmodule homomorphism φ : F → G induces a homomorphism p φ: p F −→ p G. Part (a) is straightforward (see Exercise 8.0.2.2. Proof. we assume n > 0 and deﬁne a map m F ⊗R n H → m+n G as follows.1) m+n G≃ m F ⊗OX n H. The Canonical Sheaf of a Toric Variety. where n ν ′ = ν. One can check that n (Ω1 ⊗OX OY ) ≃ n (Ω1 ) ⊗OX OY . A corollary of this proposition is that if 0 → F → G → H → 0 is an exact sequence of locally free sheaves on a variety X with rank F = m and rank H = m.2. m = dimY .18.
We saw in the proof of Theorem 8. . We ﬁrst assume that XΣ is smooth with no torus factors. then one sees easily that Pic(XΣ ) ⊗Z OXΣ ≃ OXΣ . P Recall that the transition functions for this bundle are given by: φ20 = −y/x 2 1/x 1/y −x/y 2 −y 2 /x 2 0 .2.2. and if r−n we set r = Σ(1). In Exercise 8.2.2.2.1.2. Differential Forms on Toric Varieties 359 Proof. we conclude that ωXΣ ≃ OXΣ (− ρ Dρ ) by Proposition 8. It follows by induction from Proposition 8.6. so that the Ω1 Σ = Ωn Σ = ωXΣ X X since XΣ is smooth. ρ OXΣ (−Dρ ) ≃ OXΣ − Turning to the lefthand side of (8.4 that U0 is a smooth toric variety satsifying codim(X \U0 ) ≥ 2.2.1. Now suppose that XΣ is normal but not necessarily smooth.2. Hence we can apply part (b) of Proposition 8. Here are some examples.§8. Let j : U0 ⊆ XΣ be the inclusion map for U0 = ρ Uρ .1 that the righthand side of (8. Theorem 8.1 to obtain (8.2). Each OXΣ (−Dρ ) is a line bundle since XΣ is smooth.0. In Exercise 8. φ10 = 2 2 . The previous example shows that ωP2 ≃ OP2 (−3). you will see another way to understand and derive this isomorphism.3 implies that the canonical bundle of P n is ωPn ≃ OPn (−n − 1) for all n ≥ 1 since Cl(P n ) ≃ Z and D0 ∼ D1 ∼ · · · ∼ Dn . This proves the result when XΣ is smooth without torus factors. Since the fans for U0 and XΣ have the same 1dimensional cones. Since these sheaves are reﬂexive and codim(X \ U0 ) ≥ 2.4.1.0.5. these sheaves become isomorphic over U0 by the smooth case.2.2) is isomorphic to ρ Dρ . Now consider ωXΣ and OXΣ (− ρ Dρ ).3. φ12 = −1/x 2 0 −y/x y/x 0 −1/y . Example 8. Then we have the Euler sequence 0 −→ Ω1 Σ −→ X ρ OXΣ (−Dρ ) −→ Pic(XΣ ) ⊗Z OXΣ −→ 0 from Theorem 8. We will compute this directly using ω P2 = Ω2 2 = 2 Ω1 2 P P and the description of Ω1 2 as a rank 2 vector bundle given in Example 8.2. ♦ Example 8.2) n Ω1 Σ ⊗OXΣ X r−n r−n OXΣ ≃ r ρ OXΣ (−Dρ ) .2. note that lefthand side is isomorphic to n r−n r−n OXΣ ≃ OXΣ .16.2 you will deduce the result for an arbitrary smooth toric variety.
19 used homogeneous coordinates x0 .0. 2 Ω1 2 ≃ OP2 (−3) says the canonical bundle of P 2 is P the third tensor power of the tautological bundle described in Examples 6. A similar computation shows Uσ2 = U1 .1.4. but using the Ui charts. y/x = 1/(x0 /x2 ).0.8. Thus we have 2 φ20 = −1/y 3 = (−x2 /x1 )3 = θ12 2 φ12 = −y 3 /x 3 = (−x1 /x0 )3 = θ01 2 φ10 = 1/x 3 = (x2 /x0 )3 = θ02 .1. x3 2 φ10 = Note that each is a cube. The Canonical Divisor of a Toric Variety By Exercise 8.16 used coordinates x.360 Chapter 8. Thus Cl(Hr ) = Pic(Hr ) ≃ Z2 is freely generated by the classes of D1 and D2 .2.16. Keep in mind that the φi j are in the coordinate system with charts Uσi .0. Up to a sign. we ﬁrst need to calibrate the coordinate systems. so that these give the transition functions for a line bundle on P 2 . When we computed the class group of the Hirzebruch surface Hr in Example 4.19 and 6.2. y from Uσ0 . hence Uσ1 = U0 .21.2. we wrote the divisors Dρ as D1 . D3 . To see this directly.19. these are indeed the cubes of the transition functions that we computed for the tautological bundle OPn (−1) in Example 6. On the other hand.2. with the standard open cover Ui = P 2 \ V(xi ). x1 /x0 .3. x3 φ20 = −1 . the corresponding maps on these 2 × 2 matrices: 2 2 are given by the determinants of 1 . y3 2 φ12 = −y 3 . We are now set for the ﬁnal calculation.0. D2 . x1 . ♦ Example 8.0. you will work through the deﬁnition of the canonical bundle directly to ﬁnd the transition functions given in Example 8. In Exercise 8. we have 1/x. x2 for P 2 .6.0. (x1 /x2 )/(x0 /x2 ) = x2 /x0 . Example 8. Letting x = x0 /x2 and y = x1 /x2 gives an isomorphism Uσ0 = U2 . ♦ . We will use θi j to denote the same transition function. It follows that the canonical bundle can be written ωHr ≃ OHr (−D1 − D2 − D3 − D4 ) ≃ OHr (−(r + 2)D1 − 2D2 ) by Theorem 8. and Example 6. D4 and showed that D3 ∼ D1 D4 ∼ rD1 + D2 . It is also evident that 2 φ10 = 2 φ12 · 2 φ20 . Translating coordinates for Uσ1 .
.2. This depends on the ordering of the ρi . Proposition 8. . where S = C[xρ  ρ ∈ Σ(1)] is the total coordinate ring of XΣ . .7. though the product det(uI ) dxρ1 ∧ · · · ∧ dxρn depends only on e1 .3 implies OXΣ (−β0 ) ≃ OXΣ − ρ Dρ ≃ ωXΣ . en .10 and Proposition 8.8. the coherent sheaf associated to S(−β0 ) is denoted OXΣ (−β0 ). Then S(−β0 ) is the graded Smodule where S(−β0 )α = Sα−β0 for α ∈ Cl(XΣ ). the canonical sheaf ωXΣ comes from a graded Smodule.5). First suppose that XΣ has no torus factors.3) Ω0 = I=n det(uI ) ρ∈I xρ / dxρ1 ∧ · · · ∧ dxρn n is welldeﬁned up to ±1. OXΣ (−β0 ) ≃ ωXΣ . .7. Proof. then Ω0 ∈ homogeneous of degree β0 . The deﬁnition of β0 allows us to pick D = − ρ Dρ . It follows that the nform (8. ρ xρ Proof. We call S(−β0 ) the canonical module of S.2. For a toric variety XΣ without torus factors. .0. we get the n×n determinant det(uI ) = det( ei .2. Differential Forms on Toric Varieties 361 The Canonical Module. and then we are done by Example 6. uρ j ). Each variable xρ ∈ S has degree deg(xρ ) = [Dρ ] ∈ Cl(XΣ ). Corollary 8. the result follows using the strategy described in Exercise 8. as does the nform dxρ1 ∧ · · · ∧ dxρn . Then Theorem 8. .7. For any normal toric variety XΣ we have an exact sequence 0 −→ ωXΣ −→ OXΣ −→ ρ ODρ .2. en of M.§8. ρn } ⊆ Σ(1). This gives the submodule S Ω0 ⊆ n ΩS/C is ΩS/C . For each nelement subset I = {ρ1 .2. According to Proposition 5. We can also describe ωXΣ in terms of nforms in the xρ . . We have the following result.2.3. This module is easy to describe explicitly.2. .2. The sheaf associated to S/ xρ is ODρ (Exercise 8. OXΣ (−β0 ) ≃ OXΣ (D) for any Weil divisor with −β0 = [D] ∈ Cl(XΣ ).3. . . .2. If we set deg(dxρ ) = deg(xρ ). As in §5. Fix a basis e1 . When XΣ has a torus factor. Multiplication by an exact sequence of graded Smodules 0 −→ S(−β0 ) −→ S −→ ρ S/ induces xρ since β0 = deg( ρ xρ ). Deﬁne β0 = deg ρ xρ = ρ Dρ ∈ Cl(XΣ ). . .
n On the other hand.uρ ρ xρ We regard C(XΣ ) as a subﬁeld of L via π ∗ . . Then the inclusion ωXΣ ⊆ ωXΣ ⊆ n ΩC(XΣ )/C n from Proposition 8.4).0.2.1. en of M gives coordinates ti = χ ei for the torus TN .uρ ρ xρ n ΩL/C .4) n . the character χ m ∈ C(XΣ ) maps to π ∗ (χ m ) = m.uρ ρ xρ . The Canonical Divisor of a Toric Variety which is isomorphic to S(−β0 ) since deg(Ω0 ) = β0 . Hence Ω0 arises in a completely natural way.1.2. . Note that dxρ dti = ei .4) as an inclusion of constant sheaves on XΣ .10 implies that for any m ∈ M.2. uρ ti xρ ρ since ti = ρ xρ ei . When we mutiply by ρ xρ ρ ρ xρ to clear denominators. we obtain en . uρ ρ dt1 dtn ∧ ··· ∧ = t1 tn = e1 .362 Chapter 8. . it is easy to see that S Ω0 ⊆ S Ω0 ⊆ n ΩS/C induces an inclusion ΩL/C . ΩC(XΣ )/C ⊆ n ΩL/C . the proof of Theorem 5. and we know from §8. . m. Then L is the function ﬁeld of CΣ(1) . so that C(XΣ ) ⊆ L and χ m = This induces an inclusion of K¨ hler nforms a (8. We regard (8. uρ dxρ ∧ ···∧ xρ dxρ xρ det(uI ) I=n ρ∈I xρ / dxρ1 ∧ · · · ∧ dxρn = Ω0 . This can be interpreted in terms of sheaves as follows. To see where Ω0 comes from. let L = C(xρ  ρ ∈ Σ(1)) be the ﬁeld of fractions of S.21 induces an inclusion ΩL/C . Then T ti dtn dt1 ∧ ··· ∧ t1 tn is a TN invariant section of ΩnN which pulls back to a rational nform in T via (8. and the surjective toric morphism π : CΣ(1) \ Z(Σ) −→ XΣ from Proposition 5.9 induces an injection on function ﬁelds π ∗ : C(XΣ ) −→ L. Furthermore. . A basis e1 .1 that dti is a TN invariant section of Ω1N .
Differential Forms on Toric Varieties 363 Using the above derviation of Ω0 . not just an isomorphism. This ring graded by setting deg(χ mt k ) = k. The “slice” of C(P) at height k is k P (see Figure 4 is §2. a lattice point m ∈ (k P) ∩ M gives a character χ mt k on TN × C∗ .14]. Since the uρ generate σ. This shows that from the point of view of differential forms. OUσ ) = C[σ ∨ ∩ M] Proposition 8. one can prove that S Ω0 = ωXΣ as subsheaves of the constant sheaf n ΩL/C —see [112. ωXΣ ) is the direct sum of C · χ m for all m ∈ M such that m. the canonical module. Γ(Uσ . Prop. ωUσ ) −→ C[σ ∨ ∩ M] −→ ρ C[σ ∨ ∩ M]/Iρ . Since Iρ = m∈σ∨ ∩M. Note that this is an equality of sheaves. S Ω0 deserves to be called the canonical module.9. We can also describe SP in terms of the cone over P deﬁned by C(P) = Cone(P × {1}) ⊆ MR × R.1. It is isomorphic to the earlier version S(−β0 ) of the canonical module via the map that takes Ω0 ∈ S Ω0 to 1 ∈ S(−β0 ).5) SP = m∈(k P)∩M C · χ mt k . uρ > 0 for all ρ. A full dimensional lattice polytope P ⊆ MR gives two interested graded rings. These characters span the algebra (8. it follows that Γ(Uσ . each with its own canonical module. The Afﬁne Case. Corollary 8. where Iρ = I(Dρ ∩Uσ ) is the ideal of Dρ ∩Uσ ⊂ Uσ . we get the ideal Γ(Uσ . this is equivalent to saying that m is in the interior of σ ∨ . which implies that SP is the semigroup algebra SP = C[C(P) ∩ (M × Z)] .8 gives the exact sequence 0 −→ Γ(Uσ .2).2. Proof. The Projective Case. When we think of ωUσ as the ideal sheaf OUσ (− ρ Dρ ). For the ﬁrst.§8.2. ωUσ)⊆C[σ ∨ ∩M] is the ideal generated by the characters χ m for all m ∈ M in the interior of σ ∨ . we use an auxiliary variable t so that for every k ∈ N.2. 14. ωUσ ) ⊆ Γ(Uσ . The canonical sheaf ωUσ of a normal afﬁne toric variety Uσ is determined by its module of global sections.uρ >0 C · χm ⊆ m∈σ∨ ∩M C · χ m = C[σ ∨ ∩ M] by (8.4). m.2.
Furthemore. The second ring associated to P uses the total coordinate ring S of the projective toric variety XP and the corresponding ample divisor DP . graded SP modules give quasicoherent sheaves on Proj(SP ). which gives the homogeneous coordinate ring C[XP ]. as described in [78.364 Chapter 8.9 tells us that the canonical sheaf of the associated toric variety comes from the ideal IP = m∈Int(k P)∩M C · χ mt k ⊆ SP .2. each of which has an ideal representing the canonical module.3). C[XP ] = C[XP ].A..7). It follows that the polytope P gives graded rings SP and S•α . . II.4.2. Let α = [DP ] ∈ Cl(XP ) be the divisor class of DP . Proof. k ∈ N. The canonical module of S is S(−β0 ). Proposition 8. The proof of Theorem 5.2. The graded rings SP and S•α are naturally isomorphic via an isomorphism that takes IP to I•α . When restricted to S•α .2.6).8. where Ikα ⊂ Skα consisting of monomials of degree α in which every variable appears to a positive power.e. form the graded ring ∞ S•α = k=0 Skα . For readers familiar with the Proj construction (see the appendix to Chapter 7). The graded pieces Skα ⊆ S. The sheaf on XP associated to the ideal IP ⊆ SP by the Proj construction is the canonical sheaf of XP . This is the canonical module of SP . i. this gives the ideal ∞ I•α = k=0 Ikα ⊆ S•α .4. Then the following is true (Exercise 8.2.8 used homogenization to construct an isomorphism SP ≃ S•α .11. These are related as follows. Then Proposition 8. The Canonical Divisor of a Toric Variety from the proof of Theorem 5.10. As we learned in §2.1. The situation becomes even nicer when P is a normal polytope. which is isomorphic to the ideal ρ xρ ⊆ S via multiplication by ρ xρ since β0 = deg ρ xρ ).5] (this is similar to the construction given in §5. It is straightforward to see that this isomorphism carries the ideal IP ⊆ SP to the ideal I•α ⊆ S•α (Exercise 8. C[XP ] is also the ordinary coordinate ring of its afﬁne cone XP .0. The ample divisor DP is very ample and hence embeds XP into a projective space. we note that XP ≃ Proj(SP ) by Theorem 7. Theorem 8.
Differential Forms on Toric Varieties 365 Furthermore. Then KXΣ is Cartier if and only if for each maximal cone σ ∈ Σ. Since the Picard group of a normal afﬁne toric variety is trivial (Proposition 4.2. where the Weil divisors D1 .2. since P is normal. When the Canonical Divisor is Cartier. and when we use the grading on SP and IP . All smooth varieties are Gorenstein. For a toric variety XΣ .2.2). ♦ We will use the following terminology. we have X is Gorenstein ⇐⇒ KX is Cartier ⇐⇒ ωX is a line bundle by Proposition 8. the Weil divisor KXΣ = − ρ Dρ is called the canonical divisor. Everything ﬁts together very nicely.2. There is a more general notion of canonical module that applies to any graded CohenMacaulay ring S• = ∞ Sk where S0 is a ﬁeld—see [29. We will study further examples of singular Gorenstein varieties in the next section of the chapter.1. there exists mσ ∈ MQ such that mσ .12. Let XΣ be a normal toric variety. we have the following characterization of the cases when the canonical divisor is Cartier.2. K is QCartier if and only if for each maximal σ ∈ Σ. Similarly. where one easily computes that mσ = (2/d)e1 + e2 satisﬁes mσ . In fact.7.4.2. Sec.§8. from Theorem 4.14.12.4.2. Let KX be a canonical divisor on a normal variety X . Theorem 5. The k=0 canonical modules of SP ≃ S•α constructed above are canonical in this sense.6]. We say that X is Gorenstein if KX is a Cartier divisor. The canonical divisor KCd = −D1 − D2 has divisor class corresponding to b [−2] ∈ Z/dZ. D2 coming from the rays satisfy D2 ∼ D1 .8 gives isomorphisms SP ≃ S•α ≃ C[XP ] and implies that XP is the normal afﬁne toric variety given by XP = Spec(SP ) = Spec(S•α ). dD1 ∼ 0. The afﬁne toric variety Uσ is the rational normal cone Cd . This lies in M = Z2 if and only if d ≤ 2. It follows that the canonical sheaf of the afﬁne cone of XP comes from the ideal IP ⊆ SP . de1 − e2 = mσ . it follows that KCd is Cartier if and only if d ≤ 2. Proposition 8. b Another way to see this is via Proposition 8. they give XP and its canonical sheaf via XP ≃ Proj(SP ). e2 = 1.0. Let σ = Cone(de1 − e2 .8. e2 ) ⊆ R2 . . uρ = 1 for all ρ ∈ σ(1). Deﬁnition 8. uρ = 1 for all ρ ∈ σ(1). there exists mσ ∈ M such that mσ . Since the canonical sheaf ωX = OX (KX ) is reﬂexive. We computed Cl(Cd ) ≃ Z/dZ in Example 4. 3. of course.13. Note that KXΣ need not be a Cartier divisor if XXΣ is not smooth. Example 8.
It follows that for any ndimensional toric variety XΣ . Then φ∗ ωXΣ′ ≃ ωXΣ . First assume XΣ = Uσ . In the general case. we have m. ωUσ ) = m∈Int(σ∨ )∩M C · χm by Proposition 8.366 Chapter 8.15.2. In Chapter 11.3 implies that Γ(XΣ′ . • (External) ωXΣ = φ∗ Ωn Σ′ . We clearly have P′ ∩ M ⊆ Int(σ ∨ ) ∩ M since σ(1) ⊆ Σ′ (1). In particular. ωUσ ). uρ′ > 0 for all ρ′ ∈ Σ′ (1). where φ : XΣ′ → XΣ is the toric morphism coming X from a smooth reﬁnement Σ′ of Σ. if Σ′ is a smooth reﬁnement of Σ. the description of global sections given in Proposition 4. . uρ′ ≥ 1 for all ρ′ ∈ Σ′ (1)}. ωXΣ′ ) = m∈P′ ∩M C · χ m. Given m ∈ Int(σ ∨ ) ∩ M. XΣ is covered by afﬁne open subsets Uσ for σ ∈ Σ. as desired. and φ−1 (Uσ ) is the toric variety of the reﬁnement of σ induced by Σ′ . A reﬁnement Σ′ of a fan Σ induces a proper birational toric morphism φ : XΣ′ → XΣ . m. In particular. The opposite inclusion also holds. Then we have φ∗ ωXΣ′ ≃ ωUσ since Uσ is afﬁne. uρ′ ≥ 1. so that Σ′ reﬁnes σ. so that m.3. This is an integer. we will prove the existence of smooth reﬁnements. Proof. Since Γ(Uσ . ωXΣ′ ) = Γ(Uσ . there are two ways of constructing the canonical sheaf ωXΣ from the sheaf of nforms of a smooth toric variety: n • (Internal) ωXΣ = j∗ ΩU0 . where j : U0 ⊆ XΣ is the inclusion of the smooth locus of XΣ . The above paragraph gives isomorphisms φ∗ ωXΣ′ Uσ ≃ ωXΣ Uσ which are compatible with the inclusion Uτ ⊆ Uσ when τ is a face of σ. Let φ : XΣ′ → XΣ be the toric morphism induced by a reﬁnement Σ′ of a fan Σ in NR ≃ Rn . u > 0 for all u = 0 in σ. The canonical sheaves of XΣ′ and XΣ are related as follows. Hence these isomorphisms patch to give the desired isomorphism φ∗ ωXΣ′ ≃ ωXΣ . then n φ∗ ΩXΣ′ ≃ ωXΣ . which implies m ∈ P′ ∩ M.9. Theorem 8. as we now prove. The Canonical Divisor of a Toric Variety Reﬁnements. we conclude that Γ(XΣ′ . where P′ = {m ∈ MR  m.2. Since KΣ′ = − ρ′ ∈Σ′ (1) Dρ′ .
en will denote the corresponding dual basis of M.§8. . The sequence (8.. . Then (8. . .. we regard p M ⊗Z R as p F. This gives a natural map M ⊗Z OXΣ −→ (ρ⊥ ∩ M) ⊗Z ODρ . . Theorem 8.3) induces p p α : ΩXΣ = 1 ΩXΣ −→ M ⊗Z OXΣ .. It is thus obvious that β p ◦ α p = 0. . The sheaves ΩXΣ for 1 < p < n = dim (XΣ ) are also important for the geometry of XΣ .2. ±1 ±1 R = C[x1 .6) is exact for any normal toric variety XΣ . We can write ω uniquely as ω= i1 <···<i p fi1 . xr+1 . where σ is generated by the ﬁrst r elements of a basis of e1 . xn ]. . xr .. . . e1 . To prove exactness.2. We construct a sequence (8.i p for all i1 < · · · < i p .2. Note also that im(iu ) ⊆ (u⊥ ∩ M) (Exercise 8. we obtain the composition p M→ p−1 (ρ⊥ ∩ M) M ⊗Z OXΣ −→ βp : p p−1 (ρ⊥ ∩ M) ⊗Z OXΣ −→ p−1 ρ p−1 (ρ⊥ ∩ M) ⊗Z ODρ . p α . . p An element ρ ∈ Σ(1) gives uρ as usual. Setting xi = χ ei . u m1 ∧ · · · ∧ mi ∧ · · · ∧ m p p−1 when m1 . Differential Forms on Toric Varieties 367 p Sheaves of pforms. . Now suppose that β p (ω) = 0 for some ω ∈ p F. . we can work over an afﬁne toric variety Uσ = Cr × (C∗ )n−r . . . we get Uσ = Spec(R).6) comes from the exact sequence of Rmodules 0 −→ p ΩR/C −→ αp p M ⊗Z R −→ i=1 βp r p−1 (e⊥ ∩ M) ⊗Z R/ xi . . recall that for u ∈ N.2. the contraction map and then α p = p−1 p M has the property that M→ iu : p ∨∨ iu (m1 ∧ · · · ∧ m p ) = i=1 (−1)i−1 mi . en of N. . . To deﬁne β p . We begin with the smooth case. Exactness will follow once we prove xi1 · · · xi p divides fi1 . By abuse of notation. .2. where F is the free R module with basis e1 .8). i where α p maps dxi1 ∧ · · · ∧ dxi p to ei1 ∧ · · · ∧ ei p ⊗ xi1 · · · xi p by the description of α given in (8. . .6) p 0 −→ ΩXΣ −→ αp p M ⊗Z OXΣ −→ p p βp p−1 ρ (ρ⊥ ∩ M) ⊗Z ODρ 1 as follows. The map α : ΩXΣ → M ⊗Z OXΣ from (8.2. . m p ∈ M.1. .1. Proof. en . Since exactness of a sheaf sequence is local.i p ei1 ∧ · · · ∧ ei p . . .16.3). Using iuρ : and OXΣ → ODρ . .
6) to the module case.16 over Uσ . The Canonical Divisor of a Toric Variety If an index i > r appears in fi1 .8 n since ωXΣ = ΩXΣ .368 Chapter 8..2.3]. If XΣ has no torus factors.4..2. where ei does not appear in ω2 and all fi1 .. let Vσ (m) = SpanC (m0 ∈ M  m0 ∈ the minimal face of σ ∨ containing m) ⊆ MC . The proof for the general case follows from the smooth case by an argument similar to what we did in the proof of Theorem 8. Since ie1 (ω) = ω1 .i p . then xi automatically divides fi1 .i p . p The Afﬁne Case. which can be described using Theorem 8. Adapting the deﬁnition of β p in (8. Here is a result due to Danilov [42.16. we get a homomorphism p M ⊗Z S −→ p−1 ρ (ρ⊥ ∩ M) ⊗Z S/ xρ p of graded Smodules whose kernel we denote ΩS . ODρ ) ... where MC = M ⊗Z C. Prop. and when p = n. 4.i p involving the index i appear in ω1 ..e.10. ΩUσ ) ≃ m∈σ∨ ∩M p Vσ (m) · χ m . For an afﬁne toric variety Uσ . This completes the proof of exactness in the smooth case.17. If m ∈ M and σ ∈ Σ. p Corollary 8. This gives an exact sequence of graded Smodules p 0 −→ ΩS −→ p M ⊗Z S −→ p−1 ρ (ρ⊥ ∩ M) ⊗Z S/ xρ . the sheaf ΩUσ is determined by its global sections. for xi to divide fi1 . Using the inclusion ρ⊥ ∩ M ⊆ M and the exact sequence of Theorem 8.1.2.0. Using Example 6.16.. We can write ω = ei ∧ ω1 + ω2 .2.18. it is easy to ﬁnd a graded Smodule whose p associated sheaf is ΩXΣ . the exact sequence (8..4. Let Uσ be the toric variety of the cone σ..6) reduces to the sequence appearing in Theorem 8. i.2. we obtain the exact sequence p 0 −→ Γ(Uσ . OUσ ) −→ βp p−1 ρ M ⊗Z Γ(Uσ . then ΩXΣ is the sheaf associated to p the graded Smodule ΩS .i p . ΩUσ ) −→ p M ⊗Z Γ(Uσ .i p to be zero in R/ xi . Now suppose that an index i ≤ r appears in fi1 ..1. Proposition 8.i p since xi is invertible in R. Note that when p = 1. Then we have an isomorphism p Γ(Uσ . Proof.2.2. we obtain the following corollary of Theorem 8.. the only way for β p (ω) to vanish is for fi1 . we get the sequence in Corollary 8. We will need the following notation..2.. When XΣ has no torus factors.
uρ =0 p (ρ⊥ )C = p m. where V (σ) = O(σ) ⊆ XΣ is the orbit closure corresponding to σ ∈ Σ. p) = ρ∈Σ(1) p M ⊗Z OXΣ p−1 (ρ⊥ ∩ M) ⊗Z OV (ρ) = ρ∈Σ(1) p−1 (ρ⊥ ∩ M) ⊗Z ODρ . uρ = 0. Given a fan Σ. ΩUσ )m −→ p MC −→ βp p−1 m.uρ =0 ρ .2. consider the sheaf on XΣ deﬁned by K j (Σ.1. the intersection F= m. one can extend this exact sequence as follows. When XΣ is simplicial. p) = K 1 (Σ.2. Thus ω ∈ p MC is in the kernel of β p if and only if iuρ (ω) = 0 for all ρ with m. uρ = 0.uρ =0 by the analysis given in the proof of Theorem 8. The Simplicial Case.2.uρ =0 (ρ⊥ )C .2.8) is p Vσ (m). SpanR (m0 ∈ M  m0 ∈ F) = It follows that Vσ (m) = m.9) m. p).9) imply that the kernel of β p in (8.uρ =0 (ρ ⊥ )C . as claimed. as we now describe without proof.7) Γ(Uσ .2. It follows that we have a direct p p sum decomposition Γ(Uσ . ODρ ) = m∈σ∨ ∩M C · χ m. the sequence (8.1.1.§8. Thus K 0 (Σ.uρ =0 MC is exact for m ∈ σ ∨ ∩ M and β p is the sum of the contraction maps iuρ for all ρ satisfying m. Differential Forms on Toric Varieties 369 of modules over Γ(Uσ . Also note that C · χm m∈σ∨ ∩M.uρ =0 ρ⊥ ∩ σ ∨ is the minimal face of σ∨ containing m. ΩUσ )m . This plus (8. m. However.2.2. It follows that the kernel of β p in (8. and the exact sequence (8. For p > 1.8) is the intersection (8.6) can be written p 0 −→ ΩXΣ −→ K 0 (Σ. and one sees easily that ⊥ m.6) is exact on the right when p = 1 by part (b) of Theorem 8. where (8. p) = σ∈Σ( j) p− j (σ ⊥ ∩ M) ⊗Z OV (σ) . ΩUσ ) = m∈σ∨ ∩M Γ(Uσ . p) −→ K 1 (Σ. When XΣ is simplicial.2.8) p 0 −→ Γ(Uσ .2.2. You will show in Exercise 8.4. . OXΣ ) = C[σ ∨ ∩ M] = (8. similar though longer exact sequences exist in the simplicial case.8 that iuρ (ω) = 0 if and only if ω ∈ p (ρ⊥ )C .
Given u ∈ F ∨ . p) −→ K 1 (Σ.10 is the canonical sheaf of XP = Proj(SP ). .2. Hint: Consider the exact sequence 0 → xρ → S → S/ xρ → 0 and apply Proposition 5.2. there is an exact sequence p 0 −→ ΩXΣ −→ K 0 (Σ.2].9).2.2. ΩnN ) = Γ(TN . (a) Prove that the image of iu is contained in (b) Given ω ∈ p p−1 ⊥ u .10.16. Hence on P n . 8. then yi = xi /x0 for i = 1.2. . 8. j = i.2. Assume u = re1 where r ∈ R ∨ and e1 . Show that if we write z j = x j /xi for the afﬁne coordinates on the complement of V(xi ). p ⊥ F. xn are the usual homogeneous coordinates. 3.3. Then prove that the induced map p p p φ: F→ G is given by the p × p minors of φ (with appropriate signs). .0. n = rank(G). A discussion and proof of this result can be found in [137. Prove Theorem 8.6. n are afﬁne coordinates for the open subset where x0 = 0. we can pick bases of F. Given a map of free Rmodules φ : F → G.8. (For readers familiar with the Proj construction) Prove that the sheaf associated to the ideal IP ⊆ SP from Theorem 8. We will use this exact sequence in Chapter 9 to prove a vanishing theorem for sheaf cohomology on simpicial toric varieties.1. Hint: Let S be the total coordinate ring of XP . Then consider XΣ × Cr .2. When XΣ is simiplicial. We saw in §5. Let F be a free module of ﬁnite rank over a domain R. Let S be the total coordinate ring of a toric variety XΣ without torus factors.5. The Canonical Divisor of a Toric Variety Theorem 8. . 8. then we can see ω also has a pole of order 1 along V(x0 ).2. . . . describe how you can explain this via a change of basis. Let TN be the torus of P n .2. The n differential nform ω = dy11 ∧ · · · ∧ dynn spans Γ(TN . 1 8. prove that iu (ω) = 0 if and only if ω ∈ u . Exercises for §8. we get the p p−1 kernel u⊥ ⊆ F and the contraction map iu : F→ F. . Complete the proof of Theorem 8. . .0.2. Compare to Example 8.3 in the smooth case by showing how to reduce to the case when XΣ is smooth with no torus factors. .3 that a graded Smodule such as S(−β0 ) gives a sheaf on XP . where XΣ is a smooth toric variety with no torus factors. 8. If the result of your computation differs. and change coordinates to the z j . .2.7. These bases p p induce bases of the wedge products F and G. n.2. .1.7 and Example 6. . p) −→ · · · −→ K p (Σ. Using the open subsets Uσi ⊆ P 2 (see Figure 2 in Example 3. compute ΩP2 (Uσi ).2.370 Chapter 8. Hint: First prove that any smooth toric variety is equivariantly isomorphic to a product XΣ × (C∗ )r . Compare this to how I•α gives a sheaf on Proj(S•α ). en is a basis of F .4. 8.A.19. . where m = rank(F).1. .10. G and represent φ by a n × m matrix with entries in R. See the proof of Theorem 7. Sec. Ω1N ) and has poles of T T y y order 1 along each Di = V(xi ) for i = 1.3. and write down the transition functions on Uσi ∩ Uσ j . . 8. ω deﬁnes a section n of OPn (− i=0 Di ). 8.2. p) −→ 0. If x0 . Prove that ODρ is the sheaf associated to the graded Smodule S/ xρ .
.2.5 shows that OP2 (−KP2 ) ≃ OP2 (3) is ample. (a) Prove that C(P) is Gorenstein.§8. Example 8. Fano Toric Varieties 371 p 8. In Exercise 8. Example 8. i = 0. 8.2.2. The standard fan for P 2 = XΣ has minimal generators u0 = −e1 − e2 . . When X is smooth. 8. 8. 2e1 − e2 .2. . 8. 8.13. de1 − ke2) for integers d > 0 and 0 ≤ k < d with gcd(d. we saw that a full dimensional lattice polytope P ⊆ MR gives the cone C(P) = Cone(P × {1}) ⊆ MR × R. In general. 8. P(q0 . A strongly convex rational polyhedral cone is Gorenstein if its associated afﬁne toric variety is Gorenstein. Hint: If you get stuck. a cone is reﬂexive Gorenstein is both the cone and its dual are Gorenstein.3.15. The polytope corresponding to the anticanonical divisor of P 2 is P = {m ∈ R2  m. We continue this example to introduce the key ideas that will lead to a classiﬁcation of 2dimensional Gorenstein Fano toric varieties.2.1. Prove that Uσ is Gorenstein if and only if d = k + 1. . ui ≥ −1.14 and 8.3.5). Hint: Example 1. ΩXΣ (β)) ≃ (ΩS )β . Prop. (b) Prove that the dual cone C(P)∨ ⊆ NR × R is Gorenstein if and only if P is reﬂexive.2. We will see in Chapter 10 that every 2dimensional rational cone in R2 is lattice equivalent to a cone of the form σ = Cone(e2 .11. Fano Toric Varieties We ﬁnish this chapter with a discussion of an interesting class of projective toric varieties and their corresponding polytopes.19. −e1 + e2 ).2.1. Explain why S•α−β0 = k=0 Skα−β0 gives another model for the canonical module of the graded ring S•α discussed in the text.3) for P n . see [10.2. . For β ∈ Cl(XΣ ). Prove that a product of two Gorenstein toric varieties is again Gorenstein. ∞ §8. qn ). 8.9.14. deﬁne ΩXΣ (β) to be the sheaf p p p associated to the graded Smodule ΩS (β). k) = 1. Hint: Use Propositions 3. Prove that our favorite afﬁne toric variety V = V(xy − zw) ⊆ C4 is Gorenstein. we will simply say that X is Fano. 2}.12. Prove that Γ(XΣ . Assume that XΣ has no torus factors.3. so P 2 is a Fano variety. and P1 × P1 .1 you will check that P = Conv(−e1 − e2 . 126].2. Deﬁnition 8.3.10.2.2. In the discussion of the algebra (8. 1.5].3. Reﬂexive Gorenstein cones play an important role in mirror symmetry—see [11.2.12. A complete normal variety X is said to be a Gorenstein Fano variety if the anticanonical divisor −KX is Cartier and ample. Thus Gorenstein Fano varieties are projective. u1 = e1 and u2 = e2 .2. Compute the nform Ω0 deﬁned in (8.
the dual polytope P◦ = {u ∈ NR  m. which is the polytope generated by the ray generators of the fan of P 2 . If XΣ is a projective Gorenstein Fano variety.3.1) P = {m ∈ MR  m. e2 .2. −e1 − e2 ).1). qn ) is Gorenstein Fano if and only if qi q0 + · · · + qn for all 0 ≤ i ≤ n.3. you will generalize Example 8.3. The polytopes pictured in Figure 1 are reﬂexive.2 by showing that the weighted projective space P(q0 .3.372 Chapter 8.2 involve an unexpected relation between Fano toric varieties and the reﬂexive polytopes introduced in Chapter 2.5). Finally. ♦ Example 8. by Exercise 8. . the dual polytope is (8. then XP is a Gorenstein Fano variety. Also. P◦ is a lattice polytope and is reﬂexive (Exercise 2. which is the unique interior lattice point of P.3.3. if XP is the projective toric variety associated to a reﬂexive polytope P. ♦ The special features of Example 8. Fano Toric Varieties and Reﬂexive Polytopes. then the polytope associated to the anticanonical divisor −KXΣ = − ρ Dρ is reﬂexive. u ≥ −1 for all m ∈ P} is given by P◦ = Conv(e1 .11 that a lattice polytope in MR is reﬂexive if its facet presentation is (8.2) P◦ = Conv(uF  F is a facet of P) (Exercise 2.3. Since aF = 1 for all facets F. . Let XΣ be a normal toric variety.5).3. The anticanonical polytope P and its dual P◦ for P 2 This lattice polygon is shown in Figure 1. It follows that if P is reﬂexive. Theorem 8.4. the origin is the unique interior lattice point of P (Exercise 2. uF ≥ −1 for all facets F}. Recall from Deﬁnition 2. In Exercise 8. Conversely. The following result gives the connection between projective Gorenstein Fano toric varieties and reﬂexive polytopes generalizing what we saw above for P 2 .3. The open circle in the ﬁgure represents the origin.1. . .3. The Canonical Divisor of a Toric Variety t d t dt d t d dt d t t t dt t ¡d ¡ d dt ¨ ¡¨¨ ¨ t ¡ P P◦ Figure 1. .3.3.2.
3. containing x and such that uF . By Theorem 8. one easily sees that F ⊆ U . m is an integer ≥ −1. The normal fan ΣP of P deﬁnes the variety XP = XΣP . Lemma 8. πm (p) is a line parallel to the line Rm. We begin with some general facts about a reﬂexive polytope P. Since reﬂexive polytopes contain the origin as an interior point. One can show without difﬁculty there exists some facet F of P. This is called lattice equivalence. there is a projection map πm : MR −→ MR /Rm whose image is a polytope whose vertices lie in M/Zm.§8.3. m < 0.3. Since P is reﬂexive and m ∈ M. “classify” means up to invertible linear maps of MR induced by isomorphisms of M. let P be a reﬂexive polytope in MR . Conversely. −1 Proof. so that −KXP is ample. / If F is an facet of P containing m. The ﬁrst interesting case is in dimension two. The facet presentation (8.4. The ﬁrst result concerns projections of reﬂexive polytopes. not parallel to m. Classiﬁcation. Given a reﬂexive polygon P and a primitive element m ∈ M.4.5. . First note that the lattice points of P are the origin and the lattice points lying on the boundary. This implies that polytope associated to −KXΣ = ρ Dρ has facet presentation P = {m ∈ MR  m. Then πm (P) is a in lattice polytope in MR /Rm containing the origin as an interior lattice point. then −KXΣ is Cartier and ample. and from here the lemma follows easily. so that x. By taking the point in P that maps to p and is farthest along this line in the direction of m. If XΣ is Gorenstein Fano. Let P be a reﬂexive polytope in MR ≃ Rn and let m be a lattice point m in the boundary of P. m = −1. classifying toric Gorenstein Fano varieties is equivalent to classifying reﬂexive polytopes P in MR . Conversely.1). uρ ≥ −1 for all ρ ∈ Σ(1)}. Hence uF . where toric Gorenstein Fano surfaces correspond to 16 equivalences classes of reﬂexive polygons. Thus U ⊆ m∈F F. and hence is reﬂexive by (8.3. Furthermore. and πm (P) = πm m∈F facet of P F . uF . we see that the points in πm (P) are in bijective correspondence with the points in U = {x ∈ P  x + λm ∈ P for all λ > 0}. m ∈ F.1. Hence the Cartier divisor corresponding to P is DP = F DF = −KXP . Hence XP is Gorenstein Fano. We will also need two less obvious results from [134]. any boundary lattice point is primitive. We proved that DP is ample in Proposition 6.3.1) of P has aF = 1 for every facet F of P. For each p ∈ πm (P). let x ∈ U . Fano Toric Varieties 373 Proof.
The 2Dimensional Case. Hence each pair must yield a basis for M and we can place m = e1 . assume that there are three distinct vertices m. and leave the details for the reader to verify (Exercise 8. we must have a = b = −1 (part (b) of Exercise 8. The following theorem classiﬁes reﬂexive polygons in the plane MR ≃ R2 . The vertices are primitive vectors in M.1]. Subcase A. Proof.4). where the open circle in the center of each polygon is the origin and the solid circles are the lattice points on the boundary. if we are not in subcases A. Then exactly one of the following holds: (a) m and m′ lie in a common edge of P. We will sketch the proof following [135.3.3.3.1. Proposition 4. we must have m + m′ = 0.4). b ∈ Z.3. Hence we can use a lattice equivalence to place the two of them at e1 and e2 . Let m. Project P from m′′ = e2 using Lemma 8.3. m′ .3. P cannot contain any points with y ≤ −2. It follows that the only other possible polygons in this case are 4b. or (c) m + m′ is also on the boundary of P. if m and m′ are not in the same edge. First assume that P is a reﬂexive polygon such that each edge contains exactly two lattice points. The Canonical Divisor of a Toric Variety The second result is the following fact about pairs of lattice points on the boundary of an reﬂexive polytope. e1 − e2 on the line y = −1.3. but we can have other lattice points on the boundary of P as well. . There are exactly 16 equivalence classes of reﬂexive polygons in the plane. Finally. shown in Figure 2. still in Case A. m2 = −e1 + e2 . We consider several cases. (b) m + m′ = 0.3. in Case A.2. the third is located at ae1 + be2 for some a.4) up to lattice equivalence. Subcase A. e1 on the line y = 0 and −e2 . up to lattice equivalence.3. 0. There must be edges of P containing the pairs m′ .3. and then m′′ = e2 .4). m′′ and m′ .1 or A.2.5.374 Chapter 8.3. Case A.6. m′′ such that m + m′ = m′′ . Polygons 3 and 9 are the dual pair from Example 8. Theorem 8. This gives the polygon of type 3. Figure 2 on the next page shows 16 lattice polygons.4). These lattice points must form a basis for M since the triangle formed by these two vertices and the origin has lattice points only at the vertices (part (a) of Exercise 8. The only possibility here is clearly polygon 4a. then by Lemma 8.6. Since 0 is the only lattice point in the interior. If P has exactly three vertices.3.7. Lemma 8. Next. m′ be distinct lattice points on the boundary of a reﬂexive polytope P. 5a. and ﬁx one such pair. Subcase A.2. It follows that P can only contain points −e1 . The proof is left to the reader as Exercise 8. The numbers in the labels give the number of boundary lattice points. 6a (part (d) of Exercise 8.3. m′′ (part (c) of Exercise 8. Moreover.
their sum e1 + e2 would lie in P by Lemma 8. e2 to be interior points. 5b. so −e1 + 2e2 and 2e1 − e2 lie on an edge of P. The 16 equivalence classes of reﬂexive lattice polygons in R2 Case B. or else −e2 = 0 would be an interior lattice point. . Case C. Assume there are exactly three lattice points on some edge of P.5.3. and since the origin is an interior point. −e1 + e2 and −e1 + 2e2 also lie on this edge. we can place these at −e1 + e2 . Place the vertex of this edge at −e1 − e2 so that −e1 .3.1 If b = −1. there must be a lattice point m = ae1 + be2 with a > 0 and b ≥ −1. Moreover. If −e1 + 2e2 and 2e1 − e2 do not lie on an edge of P. This is impossible. then 2e1 − e2 must lie in P. 6b. 6c. the possibilities are 4c. and 8c (part (e) of Exercise 8. Up to lattice equivalence.3. Projecting from m = e2 and using Lemma 8. e2 . Hence P has type 9. we see that P must be contained in the strip −1 ≤ x ≤ 1. P cannot contain any points with y < −3.6 and force e1 . 6d.3. 7b. 8a.§8. Projecting from −e1 via Lemma 8. Assume no edge of P has exactly three lattice points and some edge has four or more. Subcase C. 8b.5 shows that P lies above the line y = −1. e1 + e2 . 7a. By an isomorphism of M.3. Fano Toric Varieties 375 t ¡d ¡ d dt ¨ ¡¨¨ ¡ t ¨ t d t d dt d dt t e e t d d dt e et t e e ¡ d ¡ e ¡ t e¡ t t 3 t t d t d dt d dt t 4a t t d ¡ d ¡ t d¡ t ¡ t 4b t d d dt t t t 4c t t t d dt t t t t t d t d t d d t 5a t t t t t 5b t t d t t 6d t t t d t t 6a t 6b ¡ t d ¡ ¡ t ¡ t t 7b t t d t ¨¨ t ¨ ¨ t t 6c 7a t ¡ d ¡ ¡ t ¡ t t t t t d t t t t t t 8a t t t 8b t ¡ t ¡ ¡ 8c t ¡ t 9 d t dt d t d dt d t t t dt Figure 2.4).
qn ) made in Example 8. . . m′ ∈ M ≃ Z2 and assume that Conv(0. you will supply the details for the proof of Theorem 8.3.3. 134. Exercises for §8. m′ form a basis of M. Hint: Assume that (a) and (b) do not hold. 8. The collection of 2dimensional reﬂexive polygons also exhibits some very interesting symmetries and regularities.7). uF . m. Subcase C. Prove the claim about P(q0 .3.3. we have (8. We will see in Chapter 10 that there is an explanation for this coincidence coming from the cohomology theory of sheaves on surfaces. For instance. e2 . Prove Lemma 8.3. Hence e1 .2. See for instance [135] and the references therein.376 Chapter 8. Kreuzer and Skarke [111] determined that there are 4319 classes of 3dimensional reﬂexive polytopes and 473800776 classes of 4dimensional reﬂexive polytopes. then its dual is also reﬂexive. 8.3. Using their computer program PALP.2 If b = 0. Prove that m.3.7. including [39. . for otherwise e1 would be an interior point. m′ ) has no lattice points other than the vertices. (a) Let m. m + uF . Since these numbers grow so quickly.11. Reﬂexive polytopes of dimension greater than two and their associated toric varieties are currently being actively studied. There are many other proofs of classiﬁcation given in Theorem 8.4). m′ > −2.2.6 would imply e1 + e2 ∈ P.3.10].6. Hint: Exercise 4. again impossible. and Batyrev [8] and Sato [155] have shown that there are 18 types in dimension 3 and 124 types in dimension 4.4. One reason for the interest in these varieties is the relation with mirror symmetry in physics. It is known that there are a ﬁnite number of equivalence classes of reﬂexive polytopes in all dimensions. most more recent work has focused on subclasses. then m = e1 .3.6 to e1 and −e1 + e2 shows that e2 ∈ P. 8. Higher Dimensions. e2 lie on an edge. then Lemma 8. .3.3. 145]. Show that for any facet F of P.3. In this exercise. we know that if P is reﬂexive.3.3. The Canonical Divisor of a Toric Variety Subcase C. . References to further proofs are given in [39.3) ∂P ∩ M + ∂P◦ ∩ N = 12.3. If no edge connects e1 . 6. Verify the claims about the polygons P and P◦ deﬁned in Example 8. then e2 becomes an interior point of P (part (f) of Exercise 8.2. See [136] for a classiﬁcation algorithm and further references. Hence this subcase cannot occur.7. These polytopes have a deceptively simple description (Exercise 8. which forces P to be contained in the polygon of type 9.1. for instance the polytopes giving smooth Fano toric varieties. Thm. There are 5 types of such polytopes in dimension 2 (Exercise 8. 8.3. and use this to conclude that (c) must hold. In Exercise 8.3 If b ≥ 1. This point and −e1 + 2e2 would force e2 to be interior.3. Then our hypotheses on P force equality. There is also a very interesting relation between the numbers of boundary lattice points in P and P◦ : in all cases.3.5 you will determine which are the dual pairs in the list above.6). Applying Lemma 8.3.
§8.3. Fano Toric Varieties
377
(b) Show that if P is a reﬂexive triangle with vertices at e1 , e2 , then the third vertex must be −e1 − e2 . Hint: One way to show this succinctly is to use the projections from the vertices e1 and e2 as in Lemma 8.3.5. (c) In the case that each facet contains exactly two vertices, show that if there are vertices m, m′ , m′′ with m + m′ = m′′ , then the pairs m, m′′ and m′ , m′′ must lie in edges. (d) Complete the proof that the polygons of types 4b, 5a, 6a are the only possibilities in Subcase A.2. Hint: Show that Conv(±e2 , ±(e2 − e1 ), e1 ) is lattice equivalent to 5a. (e) Show that every reﬂexive polygon in Case B is equivalent to one of type 4c, 5b, 6b, 6c, 6d, 7a, 7b, 8a, 8b, or 8c. (f) Prove the claim made in Subcase C.3. 8.3.5. Consider the 16 reﬂexive polygons in Figure 2. Since each polygon in the ﬁgure is reﬂexive, its dual must also appear in the ﬁgure, up to lattice equivalence. (a) For each polygon in the Figure 2, determine its dual polygon and where the dual ﬁts in the classiﬁcation. In some cases, you will need to ﬁnd an isomorphism of M that takes the dual to a polygon in the ﬁgure. Also, some of the polygon are selfdual, up to lattice equivalence. (b) Show that the relation (8.3.3) holds for each polar pair. 8.3.6. A lattice polytope is called Fano if the origin is an interior point and the vertices of every facet form a basis of the lattice. Let Q ⊆ NR be a Fano polytope. (a) Prove that Q is reﬂexive. (b) Part (a) implies that Q◦ ⊆ MR is a lattice polytope. Prove that XQ◦ is a smooth Fano toric variety whose normal fan is formed by taking cones over proper faces of Q. (c) Prove that every smooth Fano toric variety arises this way. 8.3.7. Of the 16 Gorenstein Fano varieties classiﬁed in Theorem 8.3.7, exactly ﬁve ar smooth. Find them. Hint: See Exercise 8.3.8. 8.3.8. Some of the polygons in Figure 2 give wellknown toric varieties. For example, type 9 gives P 2 and type 8a gives P 1 × P 1 . This follows easily by computing the normal fan. Here you will describe the toric surfaces coming from some of the other polygons in Figure 2. This exercise is based on [39, Rem. 6.11]. (a) Show that type 8b corresponds to the blowup of P 2 at one of the torus ﬁxed points. Also show that this is the Hirzebruch surface H1 . Hint: Remember the description of blowing up given in §2.3. (b) Similarly show that type 7a (resp. 6a) corresponds to the blowup of P 2 at two (resp. three) torus ﬁxed points. (c) Show that type 3 corresponds to a quotient P 2 /(Z/3Z) and is isomorphic to the surface in P 3 deﬁned by w3 = xyz. Hint: Compute the normal fan and show that its minimal generators span a sublattice of index 3 in Z2 . Proposition 3.3.7 will be helpful. For the ﬁnal assertion, use the characters coming from the lattice points of the polygon. (d) Show that type 8c gives the weighted projective space P(1, 1, 2) and type 4c gives the quotient P(1, 1, 2)/(Z/2Z). Hint: See Example 3.1.17. (e) Show that type 4a gives the quotient P1 × P1 /(Z/2Z). (f) Show that type 6d gives the weighted projective space P(1, 2, 3).
378
Chapter 8. The Canonical Divisor of a Toric Variety
(g) Show that type 7b (resp. 6b) corresponds to the blowup of P(1, 1, 2) at one (resp. two) smooth torus ﬁxed points. (h) Show that type 5b gives the blowup of P(1, 2, 3) at its unique smooth torus ﬁxed point. 8.3.9. In this exercise you will consider a toric surface that is not quite Fano. The lattice polygon P = Conv(±3e1 , ±3e2 , ±2e1 ± 2e2 ) ⊆ MR = R2 gives a toric surface XP . Let D = ρ Dρ be the anticanonical divisor of XP . (a) Prove that the ample Cartier divisor DP associated to P is given by DP = 6D. Conclude that D is not Cartier and that 6D is the smallest integer multiple of D that is Cartier. (b) Show that the normal fan of P has minimal generators ±e1 ± 2e2, ±2e1 ± e2 and that the minimal generators are the vertices of Q = Conv(±e1 ± 2e2, ±2e1 ± e2 ) ⊆ NR .
1 (c) Show that the dual of Q is Q◦ = 6 P and conclude that 6 is the smallest integer multiple ◦ of Q that is a lattice polytope.
8.3.10. A complete toric surface XΣ is log del Pezzo if some integer multiple of its anticanonical divisor −KXΣ is an ample Cartier divisor. The index of XΣ is the smallest positive integer ℓ such that such that −ℓKXΣ is Cartier. This exercise and the next will consider this interesting class of toric surfaces. (a) Prove that a complete toric surface is log del Pezzo of index 1 if and only if it is Gorenstein Fano. (b) Prove that the toric surface of Exercise 8.3.9 is log del Pezzo of index 6. 8.3.11. A lattice polygon Q ⊆ NR is called LDP if the origin is an interior point of Q and the vertices of Q are primitive vectors in N. The dual Q◦ ⊆ MR of a LDP polygon contains the origin as an interior point but may fail to be a lattice polygon. The index of Q is the smallest positive integer ℓ such that such that ℓQ◦ is a lattice polygon. This exercise will explore the relation betweek LDP polygons and toric log del Pezzo surfaces. (a) Show that the polygon Q of Exercise 8.3.9 is LDP of index 6. (b) A LDP polygon Q ⊆ NR gives a fan Σ in NR by taking the cones over the faces of Q. Show that the minimal generators of Σ are the vertices of Q and that the toric surface XΣ is log del Pezzo of index equal to the index of Q. (c) Conversely, let XΣ be a toric log del Pezzo surface and let Q be the convex hull of the minimal generators of Σ. Note that Q is LDP and that Σ is the fan obtained by taking the cones over the faces of Q. Hint: The key point is to show that every minimal generator of Σ is a vertex of Q. This can be proved using the strict convexity of the support function of −ℓKXΣ . This exercise shows that classifying toric log del Pezzo surfaces up to isomorphism is equivalent to classifying LDP polygons up to lattice equivalence. This is an active area of research—see [101]. 8.3.12. Let XΣ be a smooth projective toric variety. As in Deﬁnition 6.3.9, a primitive colection P = {ρ1 , . . . , ρk } ⊆ Σ(1) gives a primitive relation uρ1 + · · · + uρk =
ρ∈γ(1) cρ uρ , k i=1 uρi .
cρ ∈ Z>0 ,
Following [8], the degree of P is where γ ∈ Σ is the minimal cone containing ∆(P) = k − ρ∈γ(1) cρ . Prove that XΣ is Fano if and only if ∆(P) > 0 for all primitive collections P. This is [8, Prop. 2.3.6]. Hint: Use (6.3.10) and Theorem 6.3.8.
Chapter 9
Sheaf Cohomology of Toric Varieties
§9.0. Background: Cohomology
By Proposition 6.0.8, a short exact sequence of sheaves on X (9.0.1) 0 → F −→ G −→ H −→ 0
gives rise to an exact sequence of global sections (9.0.2) 0 −→ Γ(X , F ) −→ Γ(X , G ) −→ Γ(X , H ).
The failure of Γ(X , G ) → Γ(X , H ) to be surjective is measured by a sheaf cohomology group. The main goal of this chapter is to understand sheaf cohomology on a toric variety. Sheaves and Cohomology. A sheaf F on a variety X has sheaf cohomology groups H p (X , F ). The abstract deﬁnition uses an exact sequence of sheaves 0 −→ F −→ I 0 − → I 1 − → I 2 − → · · · , − − − where I 0 , I 1 , . . . are injective. This term is from homological algebra: a sheaf I α is injective if given a sheaf homomorphism H −→ I and an injection β : H → G , there exists a sheaf homomorphism θ making the diagram below commute: I a O
α θ β d0 d1 d2
0
/H
/ G.
379
380
Chapter 9. Sheaf Cohomology of Toric Varieties
We say that I • is an injective resolution of F . From I • we get the complex of global sections Γ(X , I • ) : Γ(X , I 0 ) − → Γ(X , I 1 ) − → Γ(X , I 2 ) − → · · · . − − − The term complex refers to the fact that d p+1 ◦ d p = 0 for all p ≥ 0. Then the pth sheaf cohomology group of F is deﬁned to be (9.0.3) H p (X , F ) = H p (Γ(X , I • )) = ker(d p )/im(d p−1 ), where for p = 0, we deﬁne d −1 to be the zero map 0 → Γ(X , I 0 ). One can prove that injective resolutions always exist and that two different injective resolutions of F give the same sheaf cohomology groups. This deﬁnition of H p (X , F ) has some very nice properties, including: • H 0 (X , F ) = Γ(X , F ). • A sheaf homomorphism F → G induces a homomorphism of cohomology groups H i (X , F ) → H i(X , G ) that is compatible with composition and takes the identity to the identity, i.e., F → H i(X , F ) is a functor. • A short exact sequence of sheaves (9.0.1) gives a long exact sequence
0 0 −→ H 0 (X , F ) −→ H 0 (X , G ) −→ H 0 (X , H ) −→ 1 H 1 (X , F ) −→ H 1 (X , G ) −→ H 1 (X , H ) −→ · · · −→
d0
d1
d2
∂
∂
∂ p−1
H p (X , F ) −→ H p (X , G ) −→ H p (X , H ) −→ · · · . We call ∂ p : H p (X , H ) → H p+1 (X , F ) a connecting homomorphism. You will prove the ﬁrst bullet in Exercise 9.0.1. For the second bullet, the key step is to show that given a sheaf homomorphism α : F → G and injective resolutions F → A • , G → B • , there are sheaf homomorphisms α p : A p → B p such that the diagram F
α
∂p
/A0
α0
d0
/ A1
α1
d1
/ ··· / ···
G
/A0
d0
/ A1
d1
commutes. We say that α• : A • → B • is a map of complexes. Then α p induces the desired map H p (X , F ) → H p (X , G ). Finally, for the last bullet, an exact sequence (9.0.1) lifts to an exact sequence of injective resolutions 0 0
/ A• O /F / B• O /G / C• O /H /0 / 0,
and one can show that taking global sections gives an exact sequence of complexes (9.0.4) 0 −→ Γ(X , A • ) −→ Γ(X , B • ) −→ Γ(X , C • ) −→ 0.
i p )∈[ℓ] p F (Ui0 ∩ · · · ∩Ui p )...0. While the abstract deﬁnition of sheaf cohomology has nice properties. it is not useful for explicit computations. . Let [ℓ] = {1. F ) p .§9.4). Then the sheaf cohomology groups are the derived functors of Γ. . there is a downˇ toearth way of viewing sheaf cohomology.0. i p ) ∈ [ℓ] p .1. 78. ˇ One can think of an element of C p (U . . Here. Applied to (9. in terms of the Cech complex. F ) is the kernel of the map (9. F ) = (i0 .. Fortunately. Then we deﬁne a differential d ˇ ˇ C p (U . ˇ Cech Cohomology. . Let U = {Ui }ℓ be an open cover of X . The group of pth Cech cochains is ˇ C p (U . ˇ Deﬁnition 9. 175] discuss sheaf cohomology and homological algebra in more detail. i p ) of elements of I satisfying i0 < · · · < i p . In the language of homological algebra.6) to the Cech complex. Background: Cohomology 381 It is a general fact in homological algebra that any exact sequence of complexes 0 −→ A• −→ B • −→ C • −→ 0 gives a long exact sequence 0 0 −→ H 0 (A• ) −→ H 0 (B • ) −→ H 0 (C • ) −→ ∂ (9. . The texts [72.0. F ) −→ C p+1 (U . . which we now describe. We especially recommend Appendix 3 of [48] and Chapters 2 and 3 of [97].0. The deﬁnition of a sheaf F on X i=1 shows that H 0 (X . . and similarly for αi → αi Ui ∩U j . we extend ˇ (9. α j → α j Ui ∩U j is the restriction map F (U j ) → F (Ui ∩U j ). F ) is as a function α that assigns an element of F (Ui0 ∩ · · · ∩Ui p ) to each (i0 . then the component of d 0 (α) in F (Ui ∩U j ) for i < j is given by α j Ui ∩U j − αi Ui ∩U j .0.0. . .0. To get “higher” information about how sections of F ﬁt together. where d 0 is deﬁned as follows: if α = (αi ) ∈ 1≤i≤ℓ F (Ui ). ℓ} be the index set for the open cover and let [ℓ] p denote the set of all (p + 1)tuples (i0 .. . We will use the following notation.5) 1 H 1 (A• ) −→ H 1 (B • ) −→ H 1 (C • ) −→ · · · −→ ∂ ∂ p−1 H p (A• ) −→ H p (B • ) −→ H p (C • ) −→ · · · . F ) = Γ(X .6) 1≤i≤ℓ ∂p F (Ui ) −→ 1≤i< j≤ℓ d0 F (Ui ∩U j ).2) is exact. Γ is left exact since (9. we get the desired long exact sequence in sheaf cohomology. . .
. Since we can construct X by “gluing together” the Ui . . . F ) for p > 0. Cohomology of a Quasicoherent Sheaf . However. i p+1 )Ui ∩···∩Ui is obtained by applying the restriction 0 p+1 map F (Ui0 ∩ · · · ∩ Uik ∩ · · · ∩Ui p+1 ) −→ F (Ui0 ∩ · · · ∩Ui p+1 ) to α(i0 . . Given a sheaf F on X and an open cover U = {Ui }i∈I of X . Deﬁnition 9.0. i p+1 )Ui 0 ∩···∩Ui p+1 .382 Chapter 9. . α(i0 . Proofs can be found in [72. ik . By Theorem 9. i p+1 ) = k=0 p (−1)k α(i0 . . Here is the rough intuition: • Consider an arbitrary open cover U = {Ui } of X . We will soon see that there is a nice case where equality occurs for all p.0. F ) −→ C1 (U . 175]. H • (X . The following covers U of X such that H vanishing theorem of Serre is very useful in this regard. F ) = H 0 (X . . . ik . This gives the Cech complex. F ) for all p. . the cohomology of X should be obtained from the cohomologies of the Ui and their various intersections. . . F ) equals H p (X . . . the ˇ Cech complex is d d d ˇ ˇ ˇ ˇ C • (U .0. which thus computes the sheaf cohomology of F . Let F be a quasicoherent sheaf on an afﬁne variety U . To compute the cohomology of a quaˇ sicoherent sheaf F on a variety X using Cech cohomology. 0 1 2 ˇ and the pth Cech cohomology group is ˇ ˇ H p (U . . . . F )) = ker(d p )/im(d p−1 ). i p+1 ). As above. 78. In other words. So all ˇ that is left is H 0 (Ui0 ∩ · · · ∩ Ui p+1 . F ) −→ · · · . Theorem 9. In Exercise 9. F ) = Γ(X . . F ) should be determined by the cohomology groups H • (Ui0 ∩· · ·∩Ui p+1 .2 you will verify that d p ◦ d p−1 = 0.2. . This suggests the following result. F ) −→ C 2 (U . . • Now suppose that U = {Ui } is an afﬁne open cover. . .3. ˇ H p (U ˇ Notice that H 0 (U . F ) = H p (C • (U . . F ) = 0 for all p > 0. . . Sheaf Cohomology of Toric Varieties by describing how d p (α) operates on elements of [ℓ] p+1 : p+1 d (α)(i0 . we can compute the cohomology of a quasicoherent sheaf F using any afﬁne open cover. Then Ui0 ∩ · · · ∩ Ui p+1 is afﬁne and hence has vanishing higher cohomology by Serre’s theorem. . . .0. . Then H p (U . F ) need not equal H p (X . F ) since F is a sheaf. F ) : 0 −→ C 0 (U . F ) as we vary over all p. F ). .3 (Serre Vanishing for Afﬁne Varieties). ik . we need to ﬁnd open ˇ p (U . .
and the assertion for H 1 follows since an element of C[x. It follows that we have an exact sequence of sheaves 0 −→ OP1 (−1) −→ OP1 −→ OD −→ 0. Hence we need only consider H 0 (P1 .4. . x−1 ] can be written a−m x−m + · · · + a−1 x−1 + a0 + a1 x + · · · + an x n . ˇ For any sheaf F on P1 . The assertion for H 0 is clear since f (x) − g(x−1 ) = 0 implies that f and g are the same constant.U1 } of P1 . Example 9. Theorem III. OP1 (−1).0. where the divisor D is one of the ﬁxed points of the torus action on P1 . x−1 ] H 0 (OP1 ) = ker(d 0 ) = C H 1 (OP1 ) = coker(d 0 ) = 0.0. where d 0 ( f (x). the Cech complex becomes (9. For simplicity. It follows that H p (P1 . F ) and H 1 (P1 . Background: Cohomology 383 Theorem 9. and Ω1 1 on P1 .5. Let U = {Ui } be an afﬁne open cover of a variety X and let F be a quasicoherent sheaf on X .7) C[x] ⊕ C[x−1 ] −→ C[x.0.5].§9. where U0 = Spec(C[x]) Note also that U0 ∩U1 = Spec(C[x. The long exact sequence in sheaf cohomology gives 0 → H 0 (OP1 (−1)) → H 0 (OP1 ) → H 0 (OD ) → H 1 (OP1 (−1)) → H 1 (OP1 ) → · · · . F ) for all p ≥ 0. g(x−1 )) = f (x) − g(x−1 ). F ) = 0 for p ≥ 2.0. x−1 ]). A more elementary proof that does not use spectral sequences can be found in [78. the Cech complex is F (U0 ) ⊕ F (U1 ) −→ F (U0 ∩U1 ). −g(x−1 ) f (x) d0 d0 and U1 = Spec(C[x−1 ]). P Consider the afﬁne open cover U = {U0 . Here is an application of Theorem 9. We compute the cohomology of OP1 . F ) ≃ H p (X .0. Then there are natural isomorphisms ˇ H p (U . The proof will be given later in the section after we introduce a spectral sequence that makes the above intuition rigorous. ˇ For F = OP1 .4. F ). Then We can represent OP1 (−1) as the line bundle OP1 (−D). we write these sheaf cohomology groups as H 0 (F ) and H 1 (F ) respectively.4.
The Serre vanishing theorem for afﬁne varieties (Theorem 9.0.0.0.8) H 0 (OP1 (−1)) = H 1 (OP1 (−1)) = 0.0. the direct image is the sheaf f∗ F on Y deﬁned by U −→ F ( f −1 (U )) for U ⊆ Y open. More generally. G ) (Exercise 9.0.8) and the long exact sequence for cohomology imply that H 0 (Ω1 1 ) = 0 P 1 H 1 (ΩP1 ) ≃ H 0 (OP1 ) = C. A proof can be found in [78.3) has a projective version.0. We noted in Example 4.0. F ). Sheaf Cohomology of Toric Varieties The map H 0 (OP1 ) → H 0 (OD ) is the isomorphism that sends a constant function of P1 to the same constant function on D.3). Then for any coherent sheaf F on X . Given a morphism f : X → Y of varieties and a sheaf F of OX modules on X . We will use this result later in the chapter to prove some interesting vanishing theorems for toric varieties. Theorem 9. f∗ F ) = H 0 (X . Finally. This is what forces H 1 (Ω1 1 ) to be nonzero. Higher Direct Images.24 that f∗ F is a sheaf of OY modules. Serre Vanishing for Projective Varieties.5. the long exact sequence implies that (9. This is the algebraic analog of a Cousin problem in complex analysis. The deﬁnition of f∗ F implies in particular that H 0 (Y .0. II.384 Chapter 9. . P ♦ A key part of Example 9. 13]. the vanishing (9. Let L be an ample line bundle on a projective variety X . Ch.3]. Earlier. for Ω1 1 . F ⊕ G ) ≃ H p (X .0. there are homomorphisms (9. Prop.6 (Serre Vanishing for Projective Varieties).0.5 was the surjectivity of (9. we use the Euler sequence P 0 −→ Ω1 1 −→ OP1 (−1) ⊕ OP1 (−1) −→ OP1 −→ 0 P from (8.5. Since H 1 (OP1 ) = 0.1. A discussion of Cousin problems and their relation to sheaves and cohomology can be found in [109.8). we showed that the surjective sheaf homomorphism OP1 (−1) ⊕ OP1 (−1) → OP1 is not surjective on global sections. in Example 6. f∗ F ) −→ H p (X . F ) since f −1 (Y ) = X . Since H p (X .9) H p (Y .7). F ⊗X L ⊗ℓ ) = 0 for all p > 0 and ℓ ≫ 0. we have H p (X . F ) ⊕ H p (X .
7 shows that these sheaves compute the cohomology of F over certain open subsets of X . spectral sequences. we asserted that . One especially nice case is when the higher direct images R p f∗ F vanish for p > 0. 463] Now assume R q f∗ F = 0 for q > 0. Propositions II.9) are isomorphisms when this happens. Let f : X → Y be a morphism and F be a quasicoherent sheaf on X .0. f∗ F ) ≃ H p (X . Then: (a) The higher direct images R p f∗ F are quasicoherent sheaves on Y .5. First suppose that f : X → Y is a morphism and F is a quasicoherent sheaf on X . R q f∗ F ) ⇒ H p+q (X . we should get the cohomology of F on all of X . Proposition 9. F ) from (9. F ). (b) If U ⊆ Y is afﬁne.4).8.q E2 = H p (Y . F ). compute H p (Y . For our second spectral sequence.5 implies that (9. R q f∗ F ).0. then R p f∗ F U is the sheaf associated to the OY (U )module H p ( f −1 (U ).e.0.5]. we also get the higher direct image R p f∗ F .4.0.0. let U = {Ui } be an open cover of X and F be a sheaf on X .0. F ).4. So when we “put these together. Then the map (9. Then Proposition 9.0 morphism E2 → H p (X . which is the sheaf on Y associated to the presheaf deﬁned by U −→ H p ( f −1 (U ).. Readers not familar with spectral sequences should glance at the appendix to this chapter before proceeding further. Suppose f : X → Y is a morphism and F is a quasicoherent sheaf on X such that R q f∗ F = 0 for q > 0. f∗ F ) → H p (X . Background: Cohomology 385 which need not be isomorphisms for p > 0. F ). II.8 and III. A proof can be found in [78. Proposition 9. we get the higher direct images R p f∗ F .q E2 = H p (Y .A.§9. which are sheaves on Y .9) is an isomorphism H p (Y . Here we discuss two spectral sequences relevant to this section.17] and [72.0.7. Proposition 9. In the discussion leading up to Theorem 9. p. This spectral sequence is discussed in [64.A. the map H p (Y .0. The precise form of this intuition is the Leray spectral sequence: p. As above. We will see below that the maps (9. In this situation. Spectral Sequences. The proof involves our next topic.9) is an isomorphism. Then p. f∗ F ) 0 q=0 q > 0. F ) (see Deﬁnition 9. Thus we have proved the following.9) is the edge homop.0.” i. Furthermore.8.
.0. Then R is CohenMacaulay if its depth equals its dimension.3 that E1 = 0 for q > 0.0 ˇ E2 = H p (U . A variety X is CohenMacaulay if its local rings (OX. F ) is an isomorphism for all p ≥ 0.4. . We ﬁrst deﬁne what it means for a local ring (R. Using Proposition 9. ˇ H 0 (Ui0 ∩ · · · ∩Ui p . We now have the tools needed to prove Theorem 9.q it follows from Theorem 9. Later in the chapter we will prove that normal toric varieties are CohenMacaulay.p ) are CohenMacaulay for all p ∈ X . Serre Duality. . F ) = C p (U ..0 p+1. F ) −→ H p (X . Hence p. . Examples include regular local rings. Elements f1 .q p.0 E2 = ker(d1 : E1 → E1 ) im(d1 : E1 → E1 ) ˇ ˇ = H p (C • (U . 153–155].4]. F ). The precise meaning is given by the E1 spectral sequence (9.0.p . fi−1 for all i.. CohenMacaulay Varieties.0. and the depth of R is the maximal length of a regular sequence. CohenMacaulay varieties provide the natural setting for a basic duality theorem of Serre. .0 p−1. A nice discussion of what CohenMacaulay means can be found in [157.. 10.. we conclude that the edge homomorphism p.10) p. fs ∈ m form a regular sequence if fi is not a zero divisor in R/ f1 .10) are given by p. .q E1 = (i0 . pp. We are assuming that U = {Ui } is an afﬁne open cover of X . F ). See also [108.q where the differential d1 : E1 → E1 is induced by inclusion.0. F ) ⇒ H p+q (X .2]. II.i p )∈[ℓ] p p. mX.5. F ) is determined by H • (Ui0 ∩ · · · ∩ Ui p+1 . We next discuss a class of varieties that play a crucial role in duality theory and are interesting in their own right. .A. Here is the simplest version.. First observe that the q = 0 terms of (9. Thus smooth varieties are CohenMacaulay. F ). .. This implies that E2 = 0 for q > 0.0 E1 = (i0 . m) to be CohenMacaulay. H q (Ui0 ∩ · · · ∩Ui p .0 p. F )) = H p (U .5 again.. Since F is quasicoherent and the intersections Ui0 ∩· · ·∩Ui p are afﬁne for all p ≥ 0.0 p.4. Proof of Theorem 9.0 ˇ and the differentials d1 are the differentials in the Cech complex.q p+1. . Sheaf Cohomology of Toric Varieties H • (X .0.i p )∈[ℓ] p p.386 Chapter 9. F ) as we vary over all p. p.. This spectral sequence is constructed in [64.q p. with signs simˇ ilar to the differential in the Cech complex.
Singular Cohomology.§9.0. These are deﬁned using continuous maps γ : ∆n −→ Z. Background: Cohomology 387 Theorem 9. there is a more general duality theorem • where canonical sheaf ωX is replaced with the dualizing complex ωX . III. the Ext groups ExtOX (G . III. we have isomorphisms H p (X . The cohomology groups H p (XΣ . In particular. A discussion of this version of duality can be found in [137.1 will use some algebraic topology. Our discussion of the sheaf cohomology of a toric variety in §9. F ) are the derived functors of the Hom functor Σ F → HomOXΣ(G . R). F ) for ﬁxed G . beginning with the singular cohomology groups H p (Z. Here we review the topological invariants we will need. Then we have the following result. where Z is a topological space and R is a commutative ring. When X fails to be CohenMacaulay. Theorem 9.0. F )∨ ≃ ExtOX (F . There are several good introductions to singular cohomology.7.7. this is proved in [78. The assertion for divisors follows from ωX ⊗OX OX (D)∨ ≃ OX (KX ) ⊗OX OX (−D) ≃ OX (KX − D). usually Z.2]. Here are some important properties of singular cohomology: .6]. OX (D))∨ ≃ H n−p (X . F ) are the derived functors of the global section functor F → Γ(XΣ . R or C. including [71]. A proof for the projective case can be found in [78. F )∨ ≃ H n−p (X . ωX ⊗OX F ∨ ). where the last isomorphism holds since D is Cartier. ωX ). Let ωX be the canonical sheaf of a complete normal CohenMacaulay variety X of dimension n. and in the same p way. When X is projective.6]. when D is a Cartier divisor on X and KX is a canonical divisor. there are natural isomorphisms H p (X .9 (Serre Duality I).0. OX (KX − D)).2. §3. Then for every coherent sheaf F on X . Let ωX be the canonical sheaf of a complete normal CohenMacaulay variety X of dimension n. We will give a version of Serre duality especially adapted to the toric case in §9. Thm. There is also a more general version of Serre duality that applies when F is coherent but not necessarily locally free. there are natural isomorphisms n−p H p (X .10 (Serre Duality II). Thm. Then for every locally free sheaf F of ﬁnite rank on X . F ). [81] and [131]. where ∆n is the standard nsimplex.
R).11 and 3. the torus TN = N ⊗ C∗ has cohomology ring M ⊗Z R. . so in particular. R) ≃ H • (TN . R) is a graded Ralgebra with multiplication given by cup product. Example 9. The ring R deﬁnes a presheaf on Z where R is the group of sections over every nonempty open U ⊂ Z. . R) = 0 otherwise. . . • For n > 1. then R p=0 H p (Z. R) is an isomorphism. More canonically. . the singular cohomology of the (n − 1)sphere S n−1 ⊆ Rn is H p (S n−1 . The torus (C∗ )n contains the real torus (S1 )n as a deformation retract via (t1 . By [22. R) = R 0 p = 0. the sheaf cohomology of the constant sheaf of R on Z is the singular cohomology H p (Z.. ♦ i. For a commutative ring R. The corresponding sheaf is the constant sheaf of R.e. αn ] where the αi lie in H 1 ((S1 )n . . . R).0. R) → H p (A.e. H • (Z. a deformation retract i : A ֒→ Z induces a ring isomorphism i∗ : H • (Z. . A continuous map induces a ring homomorphism on cohomology. The real torus (S1 )n has cohomology ring H • ((S1 )n . Sheaf Cohomology of Toric Varieties • A continuous map f : Z → W induces f ∗ : H p (W . then i∗ : H p (Z. • If i : A ֒→ Z is a deformation retract (i. . . R) and satisfy the relations αi α j = −α j αi and α2 = 0 i (see Examples 3.11.. .tn /tn ). n − 1 otherwise. Hence H • ((C∗ )n .e.12. R) ≃ H • (A. R) such that homotopic maps induce the same map on cohomology and the identity map induces the identity on cohomology. Here are some examples. the cohomology ring is the exterior algebra of a free Rmodule. III. R) ≃ H • ((S1 )n .1]. We will always assume that Z is a locally contractible metric space. there is a continuous map r : Z → A such that r ◦ i = 1A and i ◦ r is homotopic to 1Z ). The Cohomology Ring.tn ) → (t1 /t1 . R) ≃ • • n R .388 Chapter 9. R) = R[α1 . . R) ≃ • n R . there is z ∈ Z such that {z} ֒→ Z is a deformation retract). This allows us to interpret the singular cohomology of Z in terms of sheaf cohomology as follows.15 of [81]). • If Z is contractible (i. Example 9.0. Thus H • ((S1 )n . R) → H p (Z. .
0. Theorem 3.4). H 0 (Z. where α ∈ H 2 (P n . ˇ 9. G ). Use Theorem 9. A morphism f : X → Y is afﬁne if Y has an afﬁne open cover {Ui } such that f −1 (Ui ) is afﬁne for all i. The deﬁnition of H −1 may look strange but means that the sequence (9.0.3) to show that H 0 (X. we have the canonical map Z → {pt} that sends elements of Z to pt.11) in §9. The cohomology ring of P n over Q is H • (P n .0. R) −→ H 0 (Z.1. Q) ≃ Q[α]/ αn+1 .7 and Theorem 9. We will use H −1 and (9. R) whose cokernel is denoted H 0 (Z.1 when we compute sheaf cohomology on a toric variety. and the induced map H • (TN ′ .0.5. R) = 0 Z=∅ R Z = ∅. 9. Later in the book we will give a similar quotient description of the cohomology ring of any complete simplicial toric variety.0. ♦ Reduced Cohomology. Q) (see [81.13. 9. Now assume that f : X → Y is afﬁne and let F be a quasicoherent sheaf on X. assuming R = 0. R) −→ 0 is exact for all Z. R) H p (Z. F ) = Γ(X.8 to prove that H p (Y . The reduced cohomology of Z with coefﬁcients in R is deﬁned for p ≥ 0 by H p (Z.3.0.0. Let F .0. F ). 9. R) → H • (TN . Given a topological space Z. F ) for all p ≥ 0. f∗ F ) ≃ H p (X.11).12]). Thus.0. R).0.0. This induces the map R = H 0 ({pt}.0. 9. Use (9. F ) ⊕ H p (X.2. Thus a lattice homomorphism N → N ′ gives a map TN → TN ′ of tori.11) 0 −→ H −1 (Z. including Z = ∅ (Exercise 9. Background: Cohomology 389 where M is the dual of N. R) is the map • M ′ ⊗Z R −→ • M ⊗Z R ♦ determined by the dual homomorphism M ′ → M. Check that the map d p deﬁned on the Cech cochains satisﬁes d p ◦ d p−1 = 0.3.0. F ⊕ G ) ≃ H p (X.0.0.§9. H 0 (Z. Example 9. Proposition 9. Prove H p (X. R) −→ R −→ H 0 (Z. Exercises for §9. R) = and for p = −1 by H −1 (Z.4. G be sheaves on X. R) −→ H 0 (Z. R) p=0 p > 0. . R) = 0 ⇐⇒ Z is path connected. Prove the exactness of (9.0.
F ) when i : Y ֒→ X is closed in X and F is quasicoherent.7. uρ ≥ −aρ for all ρ ∈ σ(1). Prove that H p (X. 9.0. OXΣ (D)) = m∈M H 0 (Uσ . OXΣ (D)). where the direct sum is over all m ∈ M such that m. OXΣ (D)) = γ∈[ℓ] p H 0 (Uσγ . §9.0. inspired by papers of Eisenbud. When we compute sheaf cohomology using Cech cohomology. Let X be a variety that is a union of afﬁne open subsets X = U1 ∪ · · · ∪Us . . Prove that H p (P n . 9. OXΣ (D)) = γ=(i0 . Show that H p (S n−1 . F ) when π : V → X is a vector bundle and F is quasicoherent.1. F ) = 0 for p > s − 1. . (a) Let F be quasicoherent on X. then we rewrite this as (9.0.6. Proposition 4. R) ≃ R 0 p = n−1 otherwise. . Hint: Remember that S 0 consists of two points. Mustata ¸ˇ and Stillman [50].5 to prove the following isomorphisms in cohomology: (a) H p (X. K¨ ronya and Payne [85]. Consider the (n − 1)sphere S n−1 ⊆ Rn . OXΣ (D))m . We write these as σi and order them according to their indices. Cohomology of Toric Divisors Demazure [45] gave a concrete description of the sheaf cohomology of the sheaf OXΣ (D) of a torusinvariant Cartier divisor D on a toric variety XΣ .3 implies that the sections of OXΣ (D) over Uσ can be written H 0 (Uσ . and Perling [143]. Sheaf Cohomology of Toric Varieties 9.3.8. the Cech complex is given by ˇ C p (U . i∗ F ) ≃ H p (Y . i p ) ∈ [ℓ] p . (b) H p (X.. where Σmax is the set of maximal cones in Σ.0. OXΣ (D)). . π∗ F ) ≃ H p (V . OXΣ (D)) = m C · χ m. the obvious choice of open cover is U = {Uσ }σ∈Σmax . For an afﬁne open subset Uσ . Use Exercise 9. (b) Let F be quasicoherent on P n . u ˇ ˇ The Toric Cech Complex. If we set σγ = σi0 ∩ · · · ∩ σi p ∈ Σ for γ = (i0 . Hering.1) ˇ C p (U . We generalize this to torusinvariant QCartier divisors.1. ... The Grading on Cohomology..i p )∈[ℓ] p H 0 (Uσi0 ∩ · · · ∩Uσi p . We can write this as H 0 (Uσ . ˇ Given a torusinvariant Cartier divisor D = ρ aρ Dρ on XΣ .390 Chapter 9. F ) = 0 for p > n.
This allows us to write the Cech complex as 1 −1 0 C B B1 0 −1C A @ “ ” 1 −1 1 0 1 −1 1 0 ˇ − − − − −→ ˇ − − − −→ ˇ 0 −→ C (a) − − − − − − C (a) − − − − − C 2 (a) −→ 0.1. The Cech differential is built from the restriction maps and hence respects the grading of the complex. Example 9.§9. The other chambers C+−+ . Note that a minus sign corresponds to strict inequality (< 0) while a plus sign corresponds to a weak inequality (≥ 0). To get a disjoint decomposition. (9. OXΣ (D))m = C · χm 0 m. u2 = 0. We will compute H p (P 2 . starting with σ0 in the ﬁrst quadrant and going counterclockwise. OP2 (a)) for a ∈ Z. l1 . 3 0 1 (9. C+−− . OP2 (aD0 )) H 0 (Ui j . When decomposing cohomology this way.1. u0 = −a. u1 = 0. For this purpose. . u2 ≥ 0}. On P 2 . ˇ ˇ This induces a grading of the Cech complex (9. OX (D))m . OX (D)). Cohomology of Toric Divisors 391 where for m ∈ M.1. u0 < −a. label the rays as usual: u0 = −e1 − e2 . Then Ui j is Ui ∩ U j and U012 is ˇ the triple intersection. l2 be the lines in MR = R2 deﬁned by m.3) where ˇ C 0 (a) = i=1 H 0 (Ui .1.1). where OP2 (a) = OP2 (aD0 ) for the divisor D0 corresponding to u0 . uρ ≥ −aρ for all ρ ∈ σ(1) otherwise. etc. we obtain a natural decomposition of sheaf cohomology H p (XΣ .1. m. These lines divide the plane into various regions called chambers. OP2 (aD0 )). We will compute the cohomology of this complex using the graded pieces for m ∈ M = Z2 . we often refer to the m ∈ M as weights. let l0 .2) H 0 (Uσ . are deﬁned similarly. OP2 (aD0 )) i< j ˇ C1 (a) = ˇ2 C (a) = H 0 (U012 .1. u1 = e1 . OX (D)) = H p (U . Let Ui be the afﬁne open corresponding to σi . u2 = e2 . ˇ Since H p (XΣ . OX (D)) = m∈M H p (XΣ . m. Here is an example of how weights can be used to compute sheaf cohomology. we deﬁne C−++ = {m ∈ MR  m. m. u1 ≥ 0. and maximal cones σi . m.
The cone σ1 ∩ σ2 is generated by u0 . and the shading indicates which chamber the points on the lines belong to. l1 −−+ l0 +−+ a −−− 0 −++ l2 +−− a ++− −++ Figure 1. For example. The chamber decomposition for a < 0 The cone σi is generated by u j .1. i.2) again.1. Putting everything together. so its dual is the halfplane of MR where m. OP2 (a))m = 0 ⇐⇒ m ∈ C++− ∩ M. Each chamber is labeled with its sign pattern. we get Table 1 on the next page. OP2 (a))m = 0 ⇐⇒ m ∈ C−++ ∩ M (9. dim C1 (a)m = 2 when m ∈ C−++ since both U01 and U02 contribute 1dimensional subspaces.392 Chapter 9..4) H 0 (U1 . m. 2}. OP2 (a))m = 0 ⇐⇒ m ∈ C+−+ ∩ M H 0 (U2 . u0 ≥ a. OP2 (a))m = 0 ⇐⇒ m ∈ (C+−+ ∪C+−− ∪C++− ) ∩ M. so we know when C1 (a)m = 0. 1.. j. which shows the ˇ ˇ dimension of C p (a) for m ∈ M and a < 0. Sheaf Cohomology of Toric Varieties We ﬁrst consider the case when a < 0. since 2 .1. where {i. Using Figure 1 and (9. u0 ≥ −a. ˇ Similar results hold for U01 and U02 .e. we get H 0 (U12 . u1 ≥ 0. m. The corresponding chamber decomposition of MR is shown in Figure 1.1. on the plus side of l0 . OP2 (a))m = 0 for all m ∈ M. where m. Be sure you understand this. Hence we can determine when ˇ C 0 (a)m = 0. This follows easily from (9. i. we have U012 is the torus of P ˇ C 2 (a)m = H 0 (U012 . . k} = {0. Finally. The labels li are placed on the plus side of the lines. uk .2) (Exercise 9. u2 ≥ 0.e.1). Thus H 0 (U0 .
3 below gives a picture of these sets. you will will adapt the above methods to show that (9. if m is a lattice point in the interior of a∆2 .m = {σ ∈ Σ  m. Cohomology of Toric Divisors 393 ˇ ˇ ˇ m ∈ M is in dim C 0 (a)m dim C1 (a)m dim C 2 (a)m C−++ ∪C+−+ ∪C++− 1 2 1 C−−+ ∪C−+− ∪C+−− 0 1 1 C−−− 0 0 1 ˇ Table 1. Given a divisor D = two subsets of Σ: • (When Σ is simplicial) VD.6) h p (a) = 0 Int(a∆2 ) ∩ M = −a−1 2 p=2 p=2 when a < 0. (0. then H p (P 2 . Dimension of C p (a)m for a < 0 ˇ When a < 0. In Exercise 9. OP2 (a)). where ϕD is the support function of D (see Exercise 9. OP2 (a)). One can check without difﬁculty that for these m’s. However. it is now easy to understand the Cech complex ˇ ˇ ˇ (9.5) 0 −→ C 0 (a)m −→ C1 (a)m −→ C 2 (a)m −→ 0.§9.1. ♦ Computing Cohomology. h p (a) = dim H p (P 2 . Figure 2 in Example 9.1. for any line bundle on P 2 . uρ < −aρ for all ρ ∈ σ(1)}. In Exercise 9.5) exact. the complex (9. u < ϕD (u)}. and the same is true for the second row as well.m = supp simp ρ aρ Dρ on XΣ and m ∈ M. 0). . where ΣD. • (When D is QCartier) VD. OP2 (a))m is nonzero. then H 2 (P 2 .1. These remarks imply that if m is a lattice point which is not in the interior of the triangle a∆2 = Conv{(0. Note the symmetry in the table. Here.1.2.1. Thus.1.1.m Conv(uρ  ρ ∈ σ(1)). 0).1. Summing up. The ﬁrst line of Table 1 corresponds to m ∈ C−++ ∪C+−+ ∪C++− .7) h p (a) = a∆2 ∩ M = 0 a+2 2 p=0 p>0 when a ≥ 0. we have completely determined the dimensions of the cohomology groups H p (P 2 . a)}. we have: (9.4). (a.3 you will explore how this symmetry relates to Serre duality.1. deﬁne σ∈ΣD.m = {u ∈ Σ  m. OP2 (a))m = 0 for all p. The values for −7 ≤ a ≤ 4 are depicted in Table 2 on the next page.
we get an exact sequence 0 −→ C −→ γ∈[ℓ] p simp H 0 (VD.m ∩ σ = Conv(uρ  ρ ∈ τ (1)). Let D = (a) If Σ is simplicial.8). OXΣ (D))m −→ γ∈[ℓ] p γ∈[ℓ] p simp simp for all σ ∈ Σ. H 0 (Uσ . Fix m ∈ M and p ≥ 0.8) shows that VD.2.9) give a canonical exact sequence (9.8) VD.1. C) −→ 0 H 0 (Uσγ . It follows easily that (9. . ≃ H p−1 (VD. It follows that for every p ≥ 0. then H p (XΣ . OXΣ (D))m = {0} ⇐⇒ m.m ∩ σ = ∅.1.m ∩ σ = ∅. then ρ aρ Dρ be a Weil divisor on XΣ .1. Since H 0 (Uσ .394 Chapter 9. supp simp H p (XΣ . OXΣ (D))m −→ C −→ H 0 (VD.1.2) and the second follows from (9. C) = C when VD. Sheaf cohomology of OP2 (a) on P 2 Theorem 9. the equivalences (9. OXΣ (D))m (b) If D is QCartier. C).1.9) where the ﬁrst equivalence uses (9. OXΣ (D))m ≃ H p−1 (VD. OXΣ (D))m −→ B p −→ C p −→ 0.m .1. simp simp simp Hence. OXΣ (D))m = C · χ m when it is nonzero.m ∩ σγ . C). uρ ≥ −aρ for all ρ ∈ σ(1) ⇐⇒ VD.m ∩ σ. Thus H 0 (VD. uρ < −aρ } determines a (possibly trivial) face τ σ since Σ is simplicial.m .1. Sheaf Cohomology of Toric Varieties a h0 (a) h1 (a) h2 (a) −7 0 0 15 −6 0 0 10 −5 0 0 6 −4 0 0 3 −3 0 0 1 −2 0 0 0 −1 0 0 0 0 1 0 0 1 3 0 0 2 6 0 0 3 10 0 0 4 15 0 0 Table 2.1. The equation (9. which we write as ˇ 0 −→ C p (U .m ∩ σ is convex and hence connected when it is nonempty. we have (9. For part (a). C) −→ 0.10) 0 −→ H 0 (Uσ . then {ρ ∈ σ(1)  m. ﬁrst note that if σ ∈ Σ. Proof.m ∩ σ. for all σ ∈ Σ.
this gives the desired isomorphism H p−1 (VD.5) becomes 0 → H 1 (XΣ . Furthermore. Since H 0 (XΣ . C) simp simp . We will show below that (9.1. we obtain (9. OXΣ (D))m −→ B • −→ C • −→ 0.m = ∅ (Exercise 9.11). p ≥ 0. OXΣ (D))m .0. simp simp simp simp simp simp where the last line uses the exact sequence (9.q E1 = γ∈[ℓ] p H q (VD. For p ≥ 2.m . ˇ Since H p (U . OXΣ (D))m → · · · . OXΣ (D)) ≃ H p (XΣ . C) ≃ H 1 (XΣ . There is an E1 spectral sequence p.m .1. OXΣ (D))m → C → H 0 (VD.m . OXΣ (D))m H −1 (VD.m . C • is the “Cech” complex C • (C .m ∩ σγ .m .13) H 0 (VD. p ≥ 2. C) ≃ H p (XΣ . OXΣ (D))m → H 0 (B • ) → H 0 (C • ) → H 1 (XΣ .m . C) = H p−1 (VD. Cohomology of Toric Divisors 395 ˇ The formula for the differential in the Cech complex can be used to deﬁne differp → B p+1 and C p → C p+1 .1.5 and 9.12) simp H p (C • ) ≃ H p (VD.m ∩ σi } of VD.7). we obtain the exact sequence 0 → H 0 (XΣ . OXΣ (D))m . where the σi are ˇ ˇ the maximal cones of Σ. C).1.m .m . OXΣ (D))m .12).1. When p = 1.1. OXΣ (D))m = 0 implies VD.§9. C) ⇒ H p+q (VD. OXΣ (D))m → 0. we get the closed cover C = {VD. where we use quotation marks since C is a closed cover rather than an open cover. OXΣ (D))m → 0 and isomorphisms H p−1 (C • ) ≃ H p (XΣ . Since Σ is the union of the maximal cones and simp simp ⊆ Σ.1. OXΣ (D)).m It remains to prove (9. C) → H 1 (XΣ . simp VD.6 you will use the theory of Koszul complexes to show that the complex B • has very simple cohomology: (9. C) ≃ H 0 (XΣ . OXΣ (D))m → C → H 0 (C • ) → H 1 (XΣ . C) for the simp constant sheaf C on VD. In Exercises 9.1.m . the long exact sequence from (9.11) H p (B • ) = C 0 p=0 p > 0. Then we get an exact sequence of complexes entials B ˇ 0 −→ C • (U .0. Thus our long exact sequence breaks up into an exact sequence 0 → H 0 (XΣ .
0. we assume that D is QCartier with support function ϕD . we invoke Proposition 9. . u5 ≥ −2.12) and completes the proof of part (a). u6 < 1. we noted earlier in the proof that each VD.m ∩ σ = ∅ (Exercise 9.1.1.8).9) to obtain the equivalence H 0 (Uσ . m. The proves (9. . Let XΣ be the toric surface obtained by blowing up P 2 at the three ﬁxed points of the torus action. simp supp ♦ Each representation H p (XΣ .3 to conclude that the edge homomorphisms are isomorphisms. simp .4. Following the proof of Theorem 9. supp From here. Let us compute H 1 (XΣ . Sheaf Cohomology of Toric Varieties p. C). m.1. It follows that VD. Example 9. .2 has its advantages. . this implies that p. C) ≃ C.1.10) with the exact sequence supp 0 −→ H 0 (Uσ . simp q > 0.m ∩ σγ is convex.e. p. u2 ≥ 0. OXΣ (D))m for m = e1 . . u3 < 1. Since m. supp For example. . u1 ≥ 0.2 use VD.q E1 = γ∈[ℓ] p H q (VD. Consider the divisor D = −D3 + 2D5 − D6 . Also. simp supp we get the sets VD. u4 < 0.0 ˇ E1 = C p (C .4.8).m shown in Figure 2 on the next page (Exercise 9. were the differential d1 is determined by the Cech differential (see [64.q E2 = ˇ H p (C • (C .1.2]). supp The open circle at the origin is a reminder that VD.m . OXΣ (D))m = {0} ⇐⇒ VD.396 Chapter 9.5. This allows us to replace (9. C) ≃ H 0 (VD.1. As in the proof of Theorem 9. simp H p (C • ) ≃ H p (VD.1. D6 on XΣ . . OXΣ (D))m in Theorem 9. . C)) = H p (C • ) q = 0 0 q > 0.m does not contain the origin. The minimal generators u1 . For part (b).m ∩ σγ . u6 of the fan Σ give divisors D1 . Theorem II.4.1.9). the vanishing theorems in §9.m . m. so that p. while for surfaces. Theorem 9. the proof is identical to what we did in part (a) (Exercise 9. C) = 0. C) −→ 0. OXΣ (D))m −→ C −→ H 0 (VD.m and VD. In particular.2 implies that H 1 (XΣ . Then one can modify the proof of (9.3. Since both sets have two connected components. we will see below that VD.0.q ˇ that applies to this situation. C) = C p . m.A..m because of its relation to convexity. m.m is easier to work with.1.m ∩ σγ is contractible.m . OXΣ (D))m ≃ H 0 (VD.m ∩ σ. i.
algebraic geometers often write this p relative cohomology group as HZD.11).m = {u ∈ Σ  m.VD. OXΣ (D))m can be found in [50].m .m is open in Σ. 74].2 is that it assumes that Σ is simplicial or D is QCartier. A toric surface is simplicial. where ZD.m is the closed set ZD. OXΣ (D))m using VD.2 can be stated using relative cohomology instead of reduced cohomology. The Surface Case. Consider.VD.m u3 u2 u1 u4 u5 u4 u5 u6 simp u6 ← VD. Other treatments of H p (XΣ .m supp VD.1. C). This is the version that appears in [59. C).m is especially simple and can be encoded by a sign sequence. Since Σ is contractible (it is a cone). Exercise 9.14) supp H p (XΣ . u ≥ ϕD (u)}.m . We illustrate this with an example. which means that we can compute simp simp H p (XΣ . the case when D is QCartier.§9.10 will discuss a version of the theorem that holds for any fan Σ and any torusinvariant Weil divisor D on XΣ .m simp supp Figure 2. C) −→ H p (Σ.1. supp Then the inclusion VD. C) −→ H p (VD.1.2 can be restated as p H p (XΣ .1. [85] and [143]. .m . p.m = Σ \VD. for example.m for D = −D3 + 2D5 − D6 and m = e1 A limitation of Theorem 9.1.m (Σ.m . this sequence and Theorem 9.m ⊆ Σ gives the long exact sequence · · · −→ H p (Σ.2 imply (9. Since VD. OXΣ (D))m ≃ H p (Σ. Then part (b) of Theorem 9. C) −→ · · · . VD. OXΣ (D))m ≃ HZD. Also note that Theorem 9.m and VD.1. The topology of VD. C).1.1. supp supp p≥0 supp (Exercise 9. Cohomology of Toric Divisors 397 u3 simp VD.m → u2 u1 supp VD.m (Σ.
m − 1) = max(0. ♦ These observations holds in general. If we start at u1 and go counterclockwise around the origin. . the sign pattern signD (m) is the string of length r whose ith entry is + if m. Proposition 9. . We have the following result from [86] and [94] (Exercise 9. We return to the situation of Example 9. for example.1. we get the sign pattern + + − − +−.m is a cycle) otherwise. We regard signD (m) as cyclic.m from Figure 2 gives the sign pattern shown in Figure 3.m + − Figure 3. OXΣ (D))m = 1 0 simp if signD (m) = − · · · − (⇔ VD.1. − − + + + + − has only one string of consecutive −’s. Given a torusinvariant Weil divisor D = i ai Di on XΣ and m ∈ M ≃ Z2 . Thus.m = ∅) otherwise simp dim H 1 (XΣ .4. OXΣ (D))m is given by dim H 0 (XΣ . ui ≥ −ai and − if m. Write the divisor D = −D3 + 2D5 − D6 as D = i ai Di .1. .5. so that connected components of VD.3. Then the picture of VD. Given a Weil divisor D = i ai Di on a complete toric surface XΣ .1. The positions of the −’s determine simp simp VD. #connected components of VD.m correspond to strings of consecutive −’s in the sign pattern. . OXΣ (D))m = max(0.12).398 Chapter 9. and label the ray generated by ui with + if simp m. OXΣ (D))m = 1 0 simp if signD (m) = + · · · + (⇔ VD. which we regard as cyclic. the dimension of H p (XΣ .m . Sheaf Cohomology of Toric Varieties − + u3 simp VD. ui < −ai . ur arranged counterclockwise around the origin. . #strings of consecutive −’s in signD (m) − 1) dim H 2 (XΣ . ui ≥ −ai and with − if m. The sign pattern for D = −D3 + 2D5 − D6 and m = e1 Example 9. Let XΣ be a complete toric surface with minimal generators u1 .m → u2 u1 − u4 u5 + simp u6 ← VD. ui < −ai .
We put the label li on the side where m.1.1. Each chamber is labeled with its sign pattern (some labels are rather small).1. When a < 0. The sign patterns signD (m) for all m ∈ M ≃ Z2 give the chamber decomposition shown in Figure 4 (Exercise 9.7. and the shading indicates where the boundary points lie. of the six chambers with sign patterns having more than one string of consecutive −’s. First.1.§9.1. OP2 (a)) = m∈Int(a∆2 ) C · χ m since − − − labels the interior of a∆2 .6.1. −+++−− ←++++−− −++++− −−+++− +++++−→ −+−++− +++−−− +++−+− l3 ←++−++− 0 −−−++− l6 +−−++−→ e1 2e1 ++−−−− ++−−−+ ++−−+− +−−−+− −−−+++ +−−−++ ←+−−+++ l4 = l1 ++−−++ l2 l5 Figure 4. Example 9. the minimal generators u0 . H 0 (XΣ . Consider the surface XΣ and divisor D = −D3 + 2D5 − D6 from Example 9.1. we have lines li deﬁned by m. we obtain: • H 0 (P 2 .1. • H 1 (P 2 . The chamber decompostion for D = −D3 + 2D5 − D6 As in Figure 1.1. OXΣ (D)) = 0 since neither + + + + ++ nor − − − − −− appear anywhere.1. • H 2 (P 2 . Furthermore.5.3.5. For each m ∈ M = Z2 .1 follows immediately from Figure 1.1. ♦ Thus the computation of Example 9. OP2 (a)) = 0 since all patterns have one string of consecutive −’s.1. where D = i ai Di . and we get ﬁve 1dimensional chambers since l1 = l4 . Cohomology of Toric Divisors 399 Example 9. u1 . ui = −ai . We now compute cohomology using Proposition 9. Using Proposition 9. ui ≥ −ai . Let us revisit Example 9. . OP2 (a)) = 0 since + + + does not appear. u2 give a sign pattern for the line bundle OP2 (a) = OP2 (aD0 ). all possible sign patterns are recorded in Figure 1 of Example 9.9). OXΣ (D)) = H 2 (XΣ .
3. . there is mσ ∈ MQ such that mσ . deﬁne dp : by d (α) = ( p ℓ R . When this is satisﬁed. Setting K p = d0 d1 d ℓ−2 d ℓ−1 K • : K 0 −→ K 1 −→ · · · −→ K ℓ−1 −→ K ℓ .1. Prove that K • ( f1 . we get the Koszul complex K • ( f1 .1.1. In Example 9. . Of the remaining two. (b) Show that D is QCartier if and only if for every σ ∈ Σ. . . fℓ . . we get the complex p ℓ i=1 f i ei ) ∧ α. 9. Hint: Deﬁne ϕ(ei ) = gi where i=1 gi fi = 1. (c) Now assume that f1 . the maps s p satisfy d p−1 (s p (α)) + s p+1 (d p (α)) = ϕ( ℓ i=1 f i ei ) α. .1.5. p ℓ R −→ p+1 ℓ R where e1 . 9. . OXΣ (D)) ≃ C · χ 0 ⊕ C · χ e1 ⊕ C · χ 2e1 . (a) Given ϕ : Rℓ → R.1. . M) is exact for every Rℓ module M. In particular. Exercises for §9.12. and for an Rmodule M. uρ = −aρ for all ρ ∈ σ(1). . These lattice points are indicated with white dots in Figure 4. fℓ .11) and depends only on D. 9. + − − − +−. Let ϕℓD be the support function of ℓD as in Theorem 4. Use the methods of Example 9. We deﬁne the support function of D to be ϕD = (1/ℓ)ϕℓD . • m = e1 . The canonical bundle of P 2 is ωP2 = OP2 (−3).1. Given a ring R and elements f1 . . 9. 9. plus the 1dimensional + − − + +−) have no lattice points. Sheaf Cohomology of Toric Varieties four of them (+ + + − +−. . fℓ = R. 2e1 from + + − − +−. ♦ The paper [86] applies these methods to classify line bundles with vanishing higher cohomology on the toric surface coming from three consecutive blowups of the Hirzebruch surface H2 . OXΣ (D)) = 3. . fℓ ∈ R. . . Hence H 1 (XΣ .1. dim H 1 (XΣ . . u for all u ∈ σ. M) = K • ⊗R M.4. .2.1. . fℓ . the lattice points are: • m = 0 from + + − + +− (1dimensional). (a) Show that (1/ℓ)ϕℓD : Σ → R is a support function (Deﬁnition 4. .2. Recall that a Weil divisor D = ρ aρ Dρ on XΣ is QCartier if some positive integer multiple of ℓD is Cartier. . . R).400 Chapter 9. . show that there are maps s p : K p → K p−1 such that s1 = ϕ and s p+q (α ∧ β) = s p (α) ∧ β + (−1) p α ∧ s q (β) for all α ∈ Kp and β ∈ Kq .2.1.1.7). show that ϕD (u) = mσ . . .1. Use this and Serre duality to explain the symmetry in Table 2 in Example 9. . eℓ are the standard basis of Rℓ . Thus K • = K • ( f1 .1.1. (b) Show that for all α ∈ K p . − + − + +−. . verify the sign patterns in Figure 1.1 to prove (9.
1. 9.15) M −→ 1≤i≤ℓ M −→ M −→ M −→ · · · . OXΣ (D))m ≃ H p−1 (VD. Vanishing Theorems I In this section we prove two basic vanishing theorems for QCartier divisors on a toric variety XΣ .1. OXΣ (D))m = 0 if and only if ∆D. where ΣD. (a) Draw VD.5 becomes (9.8.1.1. 9. 9.m = σ∈ΣD. When f1 = · · · = fℓ = 1 and M is an Rmodule. Prove Proposition 9. . Let RΣ(1) be a real vector space with basis eρ . so H p−1 (∆D. simp 9. (a) For σ ∈ Σ.1.1.1. Prove that H 0 (XΣ .m = {σ ∈ Σ  m. C) can be computed using simplicial cohomology.m . uρ < −aρ . minus its ﬁrst term. 9. we describe a “simplicial” computation of H p (XΣ .m is as deﬁned in the text. We also give a version of Serre duality that is special to the toric case.13).1.1. show that H 0 (Uσ . 1≤i< j≤ℓ 1≤i< j<k≤ℓ where the entries of the matrices representing the differentials are 0 or ±1. Each σ ∈ Σ gives the simplex ∆σ = Conv(eρ  ρ ∈ σ(1)}. OXΣ (D))m = 0 if and only if Vm. .1.1. OXΣ (D))m that applies even when Σ is not simplicial.9.m σ \ {0}. OXΣ (D))m = 0 implies that H p (XΣ .14).13.m .1.11. Hint: Exercise 9. (b) Adapt the proof of Theorem 9. §9.1. let ΣD. the Koszul complex K • (1.2.1.10.2 to show that H p (XΣ . Prove that H 0 (XΣ .1. 1. (b) Prove that the complex B • deﬁned in the proof of Theorem 9.1.1. Then prove (9.m .m ∩ ∆σ = ∅. Inspired by [143].1.m is a deformation retract.6.11). simp (b) Show that VD. (c) Conclude that H p (XΣ . Assume Σ is simplicial and let VD.1.2 satisﬁes (9. . OXΣ (D))m = 0 for all p > 0.15) with M = C.m has a natural structure as a simplicial complex.m in the situation of Example 9.1. Verify Figure 2 in Example 9.7.15) is exact.1.14. OXΣ (D))m ≃ H p−1 (∆D.5.2. M) of Exercise 9.3.m ∆σ . 9. ρ ∈ σ(1)} as in the text.3 and Figure 4 in Example 9. .1. . C).2.7. ρ ∈ Σ(1).m ⊆ VD. C).1.D = ∅. Hint: Show B • is the Koszul complex (9. Vanishing Theorems I 401 9.§9. We then apply these results to show that normal toric varieties are CohenMacaulay and hence satisfy Serre duality. Prove (9.12. 9. and set ∆D. Complete the proof of part (b) of Theorem 9.m = σ∈ΣD.1. (a) Use the previous exercise to prove that (9. For m ∈ M.7. Readers familiar with simplicial complexes should note that ∆D. 9. Let D = ρ aρ Dρ be a Weil divisor on XΣ .
Also recall that the intersection product D ·C is deﬁned whenever D is QCartier and C is a complete irreducible curve in X . (b) For some integer ℓ > 0.10. D2 map to 2.2. A Weil divisor D on a normal variety X is QCartier if some positive integer multiple is Cartier. OXΣ (ℓD) is a line bundle generated by its global sections. Let D = D1 − D0 and note that D0 ∼ 2D. we know that D is nef if and only if OXΣ (D) is generated by global sections (Theorems 6.m .10 and6.2. If Σ is convex and D is nef. Theorem 9. The proof combines ideas from Theorems 6.1. Proof. D2 ∼ 5D. Then Cl(XΣ ) ≃ Z. u < ϕD (u) and m.. Example 9. Then D is nef if D · C ≥ 0 for all complete irreducible curves C ⊆ X .12 and 7. we characterize QCartier nef divisors as follows. so that m. Let XΣ = P(2.m = {u ∈ Σ  m. If Σ has convex support. OXΣ (D)) = 0 for all p > 0.e. Let u. v < ϕD (u). D2 be the divisors given by the minimal generators u0 .1. Let D be a QCartier divisor on XΣ .1. 3. D1 . 5 respectively. u < ϕD (u)} as in Theorem 9. ♦ Demazure Vanishing. 3. then H p (XΣ . Lemma 9. However. OXΣ (D)) = 0. ℓD is Cartier and basepoint free.2. Then. Proof. (c) The support function ϕD : Σ → R is convex. This generalizes the deﬁnition of nef Cartier divisor given in §6. D1 ∼ 3D.2.12).402 Chapter 9. 5) and let D0 . In the Cartier case.3 (Demazure Vanishing). 6. we have m.2.2. This fails in the QCartier case. Thus OXΣ (D) is not generated by its global sections.2 that H 0 (XΣ . Fix m ∈ M supp supp and let VD. then the following are equivalent: (a) D is nef. u2 . v < (1 − t)ϕD (u) + tϕD (v) ≤ ϕD ((1 − t)u + tv). Let D = ρ aρ Dρ be a QCartier divisor on XΣ .1. u1 . In the toric case. ﬁrst proved by Demazure [45] for Cartier divisors. Note that 2u0 + 3u1 + 5u2 = 0.2. (1 − t)u + tv = (1 − t) m. v ∈ VD.2. The support function ϕD : Σ → R is convex by Lemma 9. you will check in Exercise 9. u + t m. The most basic vanishing theorem for toric varieties is the following result. i. . where the classes of D0 . Sheaf Cohomology of Toric Varieties Nef QCartier Divisors.2.2.2. as shown by the following example. for 0 ≤ t ≤ 1.2. It also easy to see that D is QCartier and nef.1).1. D1 .2. We leave the details to the reader (Exercise 9.
3.7. Hence we may assume that XΣ2 = Uσ . aD0 is basepoint free for a ≥ 0.3.4. (b) If the map φ : N1 → N2 on the lattices of oneparameter subgroups is surjective. which agrees with what we found in Example 9. Example 9.3 implies H p (XΣ .4. Proof. OXΣ (D)) = 0 for all p > 0.0. Then σ ⊆ (N2 )R has dual σ ∨ ⊆ (M2 )R ⊆ (M1 )R and Σ1  ⊆ (N1 )R has dual Σ1 ∨ ⊆ −1 (M1 )R . It follows that R p φ∗ OXΣ1 = 0 for p > 0 by Proposition 9. then φ∗ φ∗ L ≃ L . The surjection N1 → N2 gives an injection M2 ⊆ M1 on character lattices. so that Theorem 9. supp In particular.2. Let Uσ ⊆ XΣ2 be the afﬁne open corresponding to σ ∈ Σ2 .5. OXΣ1 ).0.1.7 tells us that φ−1 (Uσ ) is a toric variety whose fan is supported on the convex set −1 φR (σ). Thus H p (φ−1 (Uσ ).m . we obtain supp H p (XΣ . Note also that the restriction φ−1 (Uσ ) → Uσ is again a −1 proper toric morphism.2. First assume L = OXΣ2 . then D is nef. Then Σ1  = φR (σ) is a convex cone. Theorem 9. Also. Let φ : XΣ1 → XΣ2 be a proper toric morphism and let L be a line bundle on XΣ2 Then: (a) R p φ∗ φ∗ L = 0 for all p > 0. For P n . σ ∈ Σ2 . OXΣ (D))m ≃ H p−1 (VD.§9.4. It sufﬁces to show that this map is an isomorphism on each afﬁne open Uσ ⊆ XΣ2 . Theorem 3. so that H 0 (XΣ1 .m is convex and hence contractible. R p φ∗ OXΣ1 Uσ is the sheaf associated to the module H p (φ−1 (Uσ ). if D is a basepoint free Cartier divisor on XΣ . Vanishing Theorems I 403 The last line follows since ϕD is convex. so that φ∗ L = OXΣ1 . ♦ Vanishing of Higher Direct Images. C) = 0 for all p > 0.3.0.2. This implies that VD.1 when n = 2. OXΣ1 ) = m∈Σ1 ∨ ∩M1 C · χm by Exercise 4. Here is an easy application of Demazure vanishing. Combining this with Theorem 9. The morphism φ induces a homomorphism of sheaves OXΣ2 → φ∗ OXΣ1 on XΣ2 (Example 4. OXΣ1 ) = 0 for p > 0 by Theorem 9.2.2.2. Σ1  = φR (σ) implies σ ∨ = Σ1 ∨ . By Proposition 9. Since φ is proper.25). Hence the higher cohomology of OPn (a) = OPn (aD0 ) vanishes by Theorem 9. We analyze Σ1 ∨ ∩ M1 as follows.7.1. It follows that Σ1 ∨ ∩ M1 = σ ∨ ∩ M1 = σ ∨ ∩ (M2 )R ∩ M1 = σ ∨ ∩ M2 .2. .
we have φ−1 (Uσ ) = Uσ since (φℓ )−1 (σ) = σ. and the restriction of φℓ to TN = M ⊗Z C∗ is the group homomorphism given by raising to the ℓth power.3 and the second follows from φ∗ OXΣ1 ≃ OXΣ2 . meaning f −1 (U ) ⊆ X is afﬁne when U ⊆ Y is afﬁne. Furthermore. Hence H 0 (XΣ1 . This maps Σ to itself and hence induces a toric morphism φℓ : XΣ → XΣ .0. Given a fan Σ in NR and a positive integer ℓ. This shows that φℓ is an R ℓ afﬁne morphism. 4]. This method is used in the proofs of many vanishing theorems—see [114. • There is a split injection i : F → f∗ G of OX modules.5. we also have H p (X . G ). The general framework uses a morphism f : Y → X and coherent sheaves G on Y and F on X such that: • f is an afﬁne morphism.404 Chapter 9. Hence we get an injection H p (X . F ) ֒→ H p (Y . take Uσ ⊆ XΣ2 . G ) by Exercise 9. G ) = 0 implies H p (X .2. Sheaf Cohomology of Toric Varieties where the last equality follows since M1 is saturated in M2 by the surjectivity of N1 → N2 . F ) −→ H p (X . For a line bundle L on XΣ2 . This means that there is a homomorphism r : f∗ G → F such that r ◦ i = 1F . OXΣ1 ) = m∈σ∨ ∩M2 C · χ m = H 0 (Uσ . OUσ ). since f is afﬁne. Here is our ﬁrst injectivity lemma. given any σ ∈ Σ. This shows that OUσ → φ∗ OXΣ1 induces an isomorphism on global sections. so that R p φ∗ φ∗ L Uσ ≃ R p φ∗ φ∗ OXΣ2 Uσ ≃ R p φ∗ OXΣ1 Uσ = 0 for p > 0. 2 where the ﬁrst isomorphism is part (a) of Exercise 9. F ) = 0. . However. In the toric context. Fujino [54] identiﬁed the best choice for the morphism f . Furthermore. By functorality. when φ : N1 → N2 is surjective. The Injectivity Lemma. We introduce a method that will be used several times in this section and the next. The dual of φℓ is multiplication by ℓ on M. f∗ G ). H p (Y . Then L Uσ ≃ OXΣ2 Uσ . f∗ G ) ≃ H p (Y . which gives the desired sheaf isomorphism since we are working with quasicoherent sheaves over an afﬁne variety. In particular. we have φ∗ φ∗ L ≃ φ∗ OXΣ1 ⊗OXΣ L ≃ L . let φℓ : N → N be multiplication by ℓ. Ch. a split injection i : F → f∗ G induces a split injection H p (X .
the sheaves OXΣ (D) and OXΣ (ℓD) come from the C[σ ∨ ∩ M]modules C · χm m∈P∩M and m∈(ℓP)∩M C · χ m. which is a polytope when Σ is complete (Proposition 4.2. uρ ≥ −aρ for all ρ ∈ Σ(1)}.1.3.4). OXΣ (−D)) ≃ m∈Relint(PD )∩M C · χ −m . OXΣ (D)) ֒→ H p (XΣ .1) and the splitting r are easily seen to be compatible with the inclusion Uτ ⊆ Uσ when τ is a face of σ.2. It follows that the maps (9. Let D = divisor on a complete toric variety XΣ . Vanishing Theorems I 405 Lemma 9. OP2 (−b)) = 0 Int(b∆2 ) ∩ M p=2 p = 2.1) (Exercise 9. one can show that the map given by (9. BatyrevBorisov Vanishing. Proof.8). H p (XΣ . We will construct a split injection OXΣ (D) ֒→ φℓ∗ OXΣ (ℓD).6.4). ρ aρ Dρ be a QCartier . Then there is an injection H p (XΣ .§9. This has been generalized by Batyrev and Borisov [11]. The lemma then follows immediately from the above discussion. Over Uσ . (b) When p = dim PD .2. In Example 9.2. u ≥ −aρ for all ρ ∈ σ(1)}. The map (9.2.1) patch to give the split injection OXΣ (D) ֒→ φℓ∗ OXΣ (ℓD). This implies that the map (9. OXΣ (−D)) = 0 for all p = dim PD . φℓ∗ OXΣ (ℓD) is determined by m∈(ℓP)∩M C · χ m with the ℓ module structure given by χ m · a = χ ℓm a. If D is nef. m′ ∈ M deﬁnes a C[σ ∨ ∩ M]module splitting of (9. let P = {m ∈ MR  m. then (a) H p (XΣ . For D = ρ aρ Dρ and σ ∈ Σ.1.2. Theorem 9.2. OXΣ (ℓD)) for all p ≥ 0. Recall that a Weil divisor D = ρ aρ Dρ gives the polyhedron PD = {m ∈ MR  m. Furthermore.7 (BatyrevBorisov Vanishing).1.2.2) r(χ m ) = 0 ′ χm m ∈ ℓM / m = ℓm′ . Since φ−1 (Uσ ) = Uσ .2. we saw that OP2 (a) has nontrivial H 2 when a < 0. If we write a = −b for b > 0.2.6) as dim H p (P 2 . we can rewrite (9. Let D be a Weil divisor on XΣ and let ℓ be a positive integer.1) m∈P∩M C · χ m −→ m∈(ℓP)∩M C · χm that sends χ m to χ ℓm is a homomorphism of C[σ ∨ ∩ M]modules (Exercise 9.
2. which easily implies that Z−D. Thus (9. −u . Hence supp H p (XΣ . C) by Theorem 9.21 and Theorem 6. then V−D.1.3) Z−D. uρ0 ≥ ϕ−D (uρ0 ) for some ρ0 ∈ Σ(1).m ..m \ {0}.2. so that ϕD (u) ≥ m. If u = 0 in Z−D. supp V−D. . u ≥ ϕ−D (u)}.m = NR \V−D. First assume that D is Cartier and dim PD = dim NR = dim XΣ .e.2. Hence Z−D. so that Z−D.2.5).2.m = Rn \ Z−D.3). OXΣ (−D))m . so that We are now ready to compute H p (XΣ .1. C) = supp 0 C p=n p = n.406 Chapter 9.4) must fail. If m ∈ −Int(PD ) ∩ M.m = {u ∈ NR  m.22. On the other hand.m = {0}. if we have m ∈ −Int(PD ) ∩ M.m .12). C) = 0 supp for all p ≥ 0.6). −m. uρ > −aρ = ϕD (uρ ) for all ρ ∈ Σ(1).1. To see why.m .m contracts to −u (Exercise 9. then = NR \ {0} is homotopic to S n−1 .2.3 implies that Z−D. This proves (9. uρ < ϕ−D (uρ ) for all ρ ∈ Σ(1). C) ≃ H p−1 (S n−1 .4) m.2. the proof of Theorem 9. It is easy to see that VD. Since ϕD is strictly convex relative to ΣD .m = {0} (Exercise 9. Then D is basepoint free since it is nef (Theorem 6.m .2. u ≥ ϕ−D (u). H p (XΣ .m is a closed strongly / convex cone of positive dimension. By Proposition 6. This completes the proof when D is Cartier and PD has dimension n. Now assume u ∈ Z−D. the strict convexity of ϕD implies 0 = ϕD (u + (−u)) > ϕD (u) + ϕD (−u) since u and −u do not lie in the same cone of ΣD .m . it is strongly convex. For m ∈ M. where Z−D. We next prove that (9.m = {0} ⇐⇒ m ∈ −Int(PD ) ∩ M.m is strongly convex. OXΣ (−D))m ≃ H p−1 (V−D.m supp supp If m ∈ −Int(PD ) ∩ M. i.m . we showed above that supp supp −u ∈ V−D. Sheaf Cohomology of Toric Varieties Proof. Hence H p (XΣ . If m ∈ −Int(PD ) ∩ M. Thus ϕ−D (−u) = −ϕD (−u) > ϕD (u) for all u = 0 in NR . Then m.2.m is convex. Combining the displayed inequalities gives ϕ−D (−u) > m. −u . note that for any nonzero u ∈ NR . Set Z−D. / m. OXΣ (−D))m ≃ H p−1 (V−D.m . In fact. then −m ∈ Int(PD ). combining cones of Σ that share the same linear functional relative to ϕD gives a fan ΣD such that Σ reﬁnes ΣD and ϕD is strictly convex relative to ΣD .m . then one of the inequalities of (9. Then uρ0 ∈ Z−D. which implies supp that −u ∈ V−D. OXΣ (−D))m ≃ H p−1 (V−D.m = NR \ Z−D.
OXΣ (−D)) = 0 for p = dim PD .2. We saw in Example 9. we need to consider what happens when D is QCartier.2. Since ℓD is Cartier and nef with PℓD = ℓPD . Theorem 9.8. C) = supp supp supp C 0 m ∈ −Relint(ℓPD ) ∩ M otherwise. H p (XΣ .5) and the full dimensional case considered above.5) H p (XD . OPn (a)) = 0 Int(a∆n ) ∩ M p=n p=n ♦ when a < 0. OXΣ (−D)). OXΣ (−D)) ֒→ H p (XΣ . In particular. C) · χ m = m∈Relint(ℓPD )∩M supp C · χ −m . Vanishing Theorems I 407 Now suppose D is Cartier but dim PD < n. the Cartier case proved above implies that H p (XΣ . from which we conclude H p−1 (V−ℓD.4 that the higher cohomology of OPn (a) vanishes when a ≥ 0. OXΣ (−ℓD)). PD spans (MD )R ⊆ MR . we have R p φ∗ OXΣ (−D) = 0. OPn (a) = OPn (−aD0 ).2.2.22. When a < 0. Since MD is saturated in M by the surjectivity of N → ND .2.5. Since V−D.§9. Pick an integer ℓ > 0 such that ℓD is Cartier. Furthermore.2. Example 9. Then Proposition 9.1. Since PD = PD′ is full dimensional in (MD )R . when p = dim PD .1. OXΣ (−ℓD)) = m∈M H p−1 (V−ℓD.m = V−ℓD.22 shows that φ induces a toric morphism φ : XΣ → XD such that D is the pullback of the ample divisor D′ on XD determined by PD . We will use the Proposition 6. Finally. By Lemma 9. we have PD ∩ M = PD ∩ (MD )R ∩ M = PD′ ∩ MD . This generalizes (9. p > 0 φ∗ OXΣ (−D) = OXD (−D′ ) by Theorem 9.21 and Theorem 6.8 implies that (9.6) from Example 9. H p (XΣ .1. where MD ⊆ M is dual to a surjection φ : N → ND . Since aD0 is ample with polytope a∆n .6.m .1.1.0.m .2. C) · χ m = m∈Relint(PD )∩M supp C · χ −m . If XD is the toric variety of PD ⊆ (MD )R . . then Theorem 6. This proves the theorem.1. we are done by (9. OXΣ (−D) ≃ φ∗ OXD (−D′ ). we have an injection H p (XΣ .m . Since φ is proper and φ : N → ND is surjective. OXΣ (−D)) = m∈M H p−1 (V−D. OXD (−D′ )) ≃ H p (XΣ . as we now explain.7 implies dim H p (P n . and the normal fan of PD lies in (ND )R . After translation.2.ℓm and ℓm ∈ −Relint(ℓPD ) ∩ M ⇔ m ∈ −Relint(PD ) ∩ M.
9 shows that Serre duality holds for any normal toric variety. We conclude that XΣ is CohenMacaulay. 3. version stated in Theorem 9. We can ﬁnd a projective toric variety that contains Uσ as an afﬁne open subset (Exercise 9. Here are two interesting aspects of Theorem 9. Corollary 5. Thm.7). Fortunately.9. Then (9. Theorem 9. p. In Exercise 9. However. Sheaf Cohomology of Toric Varieties The CohenMacaulay Property. ΩXΣ ⊗OXΣ F ∨ ). Another proof can be found in [42. Now take any integer ℓ > 0.4].2. Proof. Many p of the sheaves we deal with. Next consider an afﬁne toric variety Uσ .2.10: (a) OXΣ (KXΣ − D) need not be isomorphic to ωXΣ ⊗OXΣ OXΣ (D)∨ = OXΣ (KXΣ ) ⊗OXΣ OXΣ (−D) when D is QCartier. Theorem 9. First suppose that XΣ is projective and let D be a torusinvariant ample divisor on XΣ . Serre Duality. such as OXΣ (D) and ΩXΣ . then we have natural isomorphisms H p (XΣ .9 only applies to locally free sheaves. (a) If D is a QCartier divisor and KXΣ is the canonical divisor. it follows that Uσ is CohenMacaulay. (b) If XΣ is simplicial and F is locally free.6) H p (XΣ . Ωq Σ ⊗OXΣ F )∨ ≃ H n−p (XΣ . Here is a surprising consequence of BatyrevBorisov vanishing. there is an “Extfree” version of Serre duality that holds for these sheaves.2. Since being CohenMacaulay is a local property. A normal toric variety is CohenMacaulay.2.10 uses Ext groups. then there are natural isomorphisms n−q H p (XΣ . X Remark 9.0. Assume that XΣ is a complete toric variety of dimension n. OXΣ (−ℓD)) = 0 for all ℓ ≫ 0. 182] implies that XΣ is CohenMacaulay if and only if (9. the divisor ℓD is nef and its polytope PℓD has dimension n = dim XΣ . OXΣ (D))∨ ≃ H n−p (XΣ . p < n.0.11. Theorem 9.6) follows from Theorem 9. fail to be locally free when XΣ is not smooth. OXΣ (KXΣ − D)).2. .10 (Toric Serre Duality).2.2. So Theorem 9.72 of [108.2. Since D is ample.0.408 Chapter 9. This result was originally proved by Hochster [89].7.2. A projective variety is called arithmetically CohenMacaulay (aCM for short) if its afﬁne cone is CohenMacaulay.10 does not follow from Theorem 9.8 you will show that normal lattice polytopes give aCM toric varieties.9. while the more general version Theorem 9. and then any normal toric variety XΣ is CohenMacaulay.2.2.
x module whose depth and dimension both equal dim X .10 holds for all Weil divisors in this case. E xt OX (F .2. H omOX (F . where Ann(F) = { f ∈ R  f · F = 0}.x (Fx . fi−1 F for all i.0. In §9. ωX ) are the derived functors of F → H omOX (F .2. 3. Proof. . ωX )x = ExtOX. Vanishing Theorems I 409 (b) Every Weil divisor is QCartier on a simplicial toric variety.7) and since HomOX (F . if X is CohenMacaulay.q q p+q E2 = H p (X . By (9. Let F be a MCM sheaf on a complete normal CohenMacaulay variety X of dimension n. . −) = Γ(X .2.2. See also [108. the stalk Fx is an OX. . Theorem 9.5. We will use splitting methods.§9. fs ∈ m such that fi is not a zero divisor in F/ f1 . and the dimension of F is dim R/Ann(F). Cor. Since H p (X . ωX ) by the version of Serre duality given in Theorem 9.2.0 we deﬁned the depth of a local ring (R. . though .10.11] implies that ExtOX. F )∨ ≃ H n−p (X .24] shows that MCM sheaves satisfy a nice version of Serre duality. q E xtOX (F . Prop. ωX ). The stalk at x ∈ X is (9.x (Fx . We ﬁrst show that OXΣ (D) is MCM when D is QCartier. There there are natural isomorphisms H p (X . For example. we are done. Then we deﬁne a coherent sheaf F on a variety X to be maximal CohenMacaulay (MCM for short) if for every point x ∈ X . the depth of a ﬁnitely generated Rmodule F is the maximal length of a sequence f1 . .2. This is typical of algebraic geometry—once a variety ceases to be smooth. . then every locally free sheaf on X is MCM.10. we introduce some tools from commutative algebra. (ωX )x ). we can apply the Grothendieck spectral sequence for the composition of functors [177. More generally. . F )∨ ≃ ExtOX (F . Theorem 9. .71]. Proof of Theorem 9. ωX )). [29. H omOX (F .3] to obtain the spectral sequence p.2. Hence p H p (X . −)). Thm.5. The following result from [143.8. (ωX )x ) = 0 for q q > 0 and x ∈ X . ωX ) = 0 for q > 0. ωX ) n−p by Proposition 9. ωX ). 5. Before beginning the proof of Theorem 9. ωX )) ⇒ ExtOX (F .10. 4. Thm.7). where the sheaves E xtOX (F . H omOX (F . q q We can now prove the toric version of Serre duality. We will use the sheaf version of Ext. q Since F is MCM.A. Pick ℓ > 0 such that ℓD is Cartier. ωX )) ≃ ExtOX (F . the theory becomes more technically demanding. 5.12 (Serre Duality III). m). we obtain E xt OX (F .
8). so that B is MCM over Sσ. r is a homomorphism of Sσ. Sheaf Cohomology of Toric Varieties for clarity we replace φℓ : N → N with the sublattice ℓN ⊆ N.N ′ module. Sσ.2.12 implies that H p (XΣ .N module (Exercise 9.8) A = m∈P∩M C · χm ⊆ B = m∈P∩ℓ−1 M C · χ m = χ mσ Sσ. H omOXΣ (OXΣ (D). The dual lattice ℓ−1 M ⊇ M gives semigroup algebras Sσ. OX (E)) ≃ OX (E − D) for any Weil divisors D. OXΣ (KXΣ ))). the restriction of Ωq Σ to Uσ = Uσ. Ex.N ′ . Thm. ΩU σ. we can work locally over Uσ . Since Vσ (m) is unaffected when M is replaced by M ′ .5.2.9). This gives the desired duality for OXΣ (D).ℓN .N ′ . The minimal generators of σ form a basis of a sublattice N ′ ⊂ N of ﬁnite index such that σ is smooth relative to N ′ . u for u ∈ σ. Similar to what we did in Proposition 9.N modules.2. where the last equality uses mσ ∈ ℓ−1 M. From here. H omOXΣ (OXΣ (D).18. By Proposition 8. Let M ′ be the dual of N ′ . we can pick mσ ∈ ℓ−1 M such that ϕD (u) = mσ .2. we see that as a Sσ. we use the map r : B → A that sends χ m ∈ B to r(χ m ) = 1 0 if χ m ∈ A otherwise. where Vσ (m) = SpanC (m0 ∈ M  m0 ∈ the minimal face of σ ∨ containing m). Since OXΣ (D) is MCM. 3.N ′ ). we get semigroup algebras Sσ.2.ℓN is ﬁnitely generated as a Sσ. Now consider the larger Sσ. q B = Γ(Uσ. Hence it sufﬁces to split (9. .26].ℓN is free over Sσ. ωXΣ )) = H n−p (XΣ . However. dim σ = n. In this situation.6.N = C[σ ∨ ∩ M] ⊆ Sσ.2. Another property of MCM modules is that any nontrivial direct summand of an MCM module is again MCM.7]. OXΣ (D)). Theorem 9.ℓN = C[σ ∨ ∩ ℓ−1 M]. E on a normal variety X (Exercise 9. Now consider the Sσ.N by [29.2. Since Σ is complete. For part(b). Over an afﬁne open subset Uσ . Note that B = χ mσ Sσ. we ﬁrst show that ΩXΣ is MCM when XΣ is simplicial.N module A = Γ(Uσ . H omOX (OX (D).N module A= m∈σ∨ ∩M q q Vσ (m) · χ m .N module deﬁned by B= m∈σ∨ ∩M′ q Vσ (m) · χ m . This follows from [29. Now we prove duality.ℓN and hence is MCM over Sσ. OXΣ (D))∨ ≃ H n−p (XΣ . 1.ℓN .N is determined by X the Sσ. it follows easily that OXΣ (D) is MCM when D is QCartier.2.410 Chapter 9.N ⊆ Sσ. If we set P = mσ + σ ∨ . As above. However. then one easily sees that (9.10).
using the local freeness of F and Exercise 8. Arguing as in part (a). Cor.2.1. Consider the map m∈P∩M C · χ m −→ m∈(ℓP)∩M C · χ m from (9.2.12 you will use Alexander duality to give a purely toric proof of Serre duality for simplicial toric varieties.2. Verify the claims made in Example 9. ωX is MCM on any normal CohenMacaulay variety (see [29. ΩU ′ is locally free since Uσ.2.N and we have a splitting map r deﬁned by r(χ m ) = q q χm 0 if m ∈ σ ∨ ∩ M otherwise. 4. where the C[σ ∨ ∩ M]module structure on m∈(ℓP)∩M C · χ m is given by χ m · a = χ ℓm a. We should also mention that besides QCartier divisors. In part (b) we followed [42. 3. ωXΣ ) ≃ H omOXΣ(ΩXΣ .4) implies Z−D.0.2.2. even when the canonical class KXΣ fails to be QCartier. Prove that (9.2. Construct an isomorphism f∗ f ∗ L ≃ ( f∗ OX ) ⊗OY L of OY modules.2.1]). Exercises for §9. For example. From here.7.2. The proof of part (a) of the theorem was inspired by [42. Vanishing Theorems I 411 q However.2.2.2. and then ΩXΣ ⊗OXΣ F is MCM since F is locally free.4. so that ωXΣ is MCM on any toric variety XΣ . We saw in Example 4. there is an isomorphism f∗(G ⊗OX f ∗ L ) ≃ f∗ G ⊗OY L . You will give a toric proof of this in Exercise 9.2) are C[σ ∨ ∩ M]module homomorphisms.2. This is the projection formula.1).11. (a) Let L be a line bundle on Y . Hence σ. In Exercise 9. Ωq Σ ⊗OXΣ F )∨ ≃ H n−p (XΣ .N ′ . the theorem follows easily. H omOXΣ(Ωq Σ ⊗OXΣ F . as claimed in the proof of Theorem 9. B is MCM over Sσ. 3.2.5.0. Thus ΩXΣ is MCM.§9.25 a morphism of varieties f : X → Y induces a homomorphism OY → f∗ OX of OY modules. there can be other Weil divisors that give MCM sheaves. since Theorem 9.m = {0}.1.2].1) and (9.3. (b) Generalize part (b) by showing that for any sheaf G of OX modules on X and any line bundle L on Y . Hint: Construct a homomorphism and then study the homomorphism over open subsets of Y where L is trivial. 9.2. we have q q n−q H omOXΣ(ΩXΣ ⊗OXΣ F . 9.2. Prove Proposition 9. Other proofs that QCartier divisors give MCM sheaves can be found in [27.N ′ is smooth by our choice of N ′ .8].3. ωXΣ )) X X However. 9. 9. 4. Def. Prove that (9. ωXΣ ) ⊗OXΣ F ∨ ≃ ΩXΣ ⊗OXΣ F ∨ .2.2.2] and [143.2.12 implies H p (XΣ . Lem.2. .13. 9. 4. Prop.N B is MCM over Sσ. The proof is now easy to ﬁnish. Prop.4.22].
Prove that every afﬁne toric variety Uσ is contained in a projective toric variety XΣ .2. C). Alexander duality [131.2.m (this means that V−D. mr ∈ M2 generate σ ∨ and consider { r δi mi  0 ≤ δi < 1} ∩ M1 .2.2. Let K = KXΣ be the canonical divisor of XΣ . (c) Construct a splitting of m∈Int(σ∨ )∩M C ·χ m ⊆ the canonical sheaf of Uσ.N is MCM.2.9) H p (XΣ . p. m σ∈Σ ∆σ (a) Explain why S = is homeomorphic to S n−1 . . We say that σ is intermediate when this happens. (b) Prove that H omOX (OX (D).m nor AK−D.9. ωUσ comes from the ideal m∈Int(σ∨ )∩M C ·χ m ⊆ C[σ ∨ ∩M]. A) is triangulable (see [131.10. en−1 . (c) Show that ΦX is an isomorphism and that OX (E − D) ≃ H omOX (OX (D).8. OX (E)). Following [42. (b) Prove that AD. the line segment uv supp supp is contained in V−D. uρ )uρ lies in N. (a) Prove that uρ = (ℓ/ en . 9.7. uρ1 < −aρ1 and m.2. Sheaf Cohomology of Toric Varieties 9.−m if and only if there are ρ1 . OX (E)) is reﬂexive. Then Γ(X. The lattice points of a normal lattice polytope P give a projective embedding of the toric variety XP . 9. 150]).2 implies simp H p (XΣ .N1 = C[σ ∨ ∩ M1 ]. (a) Construct a natural map ΦX : Γ(X.m is nonzero.m . ρ2 ∈ σ(1) such that m. Hint: Let m1 .N2 = C[σ ∨ ∩ M2 ] ⊆ Sσ. OXΣ (K − D))−m . OX (E)) and prove that ΦX is an isomorphism when X is smooth.N1 is ﬁnitely generated as a module over Sσ. OX (D)U ). E be Weil divisors on a normal variety X and let U ⊆ X be the smooth locus.m ∩ AK−D.12. Cor.−m ⊆ S. Prove that for every v ∈ V−D. . You will prove Serre duality when D = ρ aρ Dρ on a complete simplicial toric variety XΣ of dimension n. Thus −u ∈ V−D.9.−m = ∅ and that for σ ∈ Σ. As in the proof of Theorem 9. Let D.2. OX (D)U ). .m . OX (E − D)) → HomOX (OX (D). C). Theorem 9. (b) Conclude that the canonical divisor of Uσ. m ∈ M simp Set AD.m is star shaped with respect to −u). σ is contained in neither AD. . . Note that AD.2. 3. Fix a cone σ ⊆ NR and pick a basis e1 .m is strongly convex and that u ∈ supp supp Z−D. uρ = 1 to show that uρ is the minimal generator of ρ with respect to N. .2.11. AK−D. .N2 . . 9.m .7. and let σ be a cone in (N1 )R = (N2 )R . This gives semigroup algebras Sσ. Conclude supp that the constant map γ : V−D.5]. 9. Your goal is to prove that (9. and the same holds for E. ℓ−1 en . uρ2 ≥ −aρ2 . . §71] states that if A ⊆ S n−1 is a closed subset such that the pair (S n−1 .m .2. . Assume that N1 ⊆ N2 has ﬁnite index. OX (D)) ≃ Γ(U. assume that Z−D. m∈Int(σ∨ )∩M C·χ m and conclude that 9. C)∨ ≃ H n−p−1 (S n−1 \ A. Hint: Let U is the smooth locus and study the restriction map HomOX (OX (D). you will show that ωXΣ is MCM on a normal toric variety. Prove that XP is aCM in this embedding.m and ∆σ = Conv(uρ  ρ ∈ σ(1)}. OX (E)) → HomOU(OX (D)U . . Then use ℓ−1 en . uρ  ρ ∈ σ(1)) and let M be the lattice with basis e1 .412 Chapter 9. Let ℓ = lcm( en . OXΣ (D))m ≃ H p−1 (VD. then H p−1 (A. Prove that Sσ.6. .m → {−u} is a contraction. en of M such that en is in the interior of σ ∨ . OXΣ (D))∨ ≃ H n−p (XΣ .N is Cartier. with dual N ⊆ N.m = VD. . with dual M2 ⊆ M1 .1. i=1 9. By Proposition 8.
we will drop the twist notation. which is to be expected since toric varieties are so special. We will explore toric versions of several of these results. uρ ≥ −aρ . Prop. σ σ (d) Prove that AK−D. Grifﬁths. p > 0. However. A glance at the index of Lazarsfeld’s twovolume treatise [114] lists vanishing theorems due to Bogomolov. σ − u− ∈ ∆− and 0 ≤ t ≤ 1.9) by applying Alexander duality to AD. the toric version is stronger. we have H p (XΣ .1 (BottSteenbrinkDanilov Vanishing). and Serre. Then show that ∆+ ⊆ ∆σ \ ∆σ is a deformation retract. p + q > n. Sec. 7. prove (9. if KX is a canonical divisor on X . ωX (D)) = 0. Miyoka. KawamataViehweg. In some cases. If D is a Cartier divisor and F a coherent sheaf on a normal variety X .3. uρ < −aρ and ∆σ be the face generated by uρ ’s with m. Demailly. Koll´ r. This notation will be used some of the vanishing theorems stated below. This exercise was inspired by [42. For example. GrauertRiemenschneider.1] and [59. Let D be an ample divisor on a projective toric variety XΣ . Kodaira vanishing asserts that H p (X . Twisting. Show that every u ∈ ∆σ can be written uniquely as u = (1 − t)u+ + tu− where u+ ∈ ∆+ .m ⊆ S. Nakano’s theorem generalizes Kodaira’s since ωX = Ωn in the smooth case.§9. ΩX (D)) = 0. when we start working with nonCartier divisors. Kodaira and Nakano.4]. (e) Finally.2.m is a deformation retract. Then for all p > 0 and q ≥ 0.7. ΩXΣ (D)) = 0. then the twist of F by D is the sheaf F (D) = F ⊗OX OX (D). then ωX (D) = ωX ⊗OX OX (D) ≃ OX (KX ) ⊗OX OX (D) ≃ OX (KX + D). 4. Kodaira. Two of the earliest vanishing theorems are due to Kodaira and Nakano. §9. and Nakano vanishing states that H p (X . For an ample divisor D on a smooth projective variety X of dimension n. X In the toric case. Let ∆σ be the face of ∆σ generated by uρ ’s with + m. Vanishing Theorems II 413 − (c) Fix an intermediate cone σ ∈ Σ. we get more vanishing.3. in what has become known as the BottSteenbrinkDanilov vanishing theorem. Nakano. Theorem 9.−m ⊆ S \ AD. Le Potier. a Manivel. q q . Vanishing Theorems II Vanishing theorems play an important role in algebraic geometry.3. Fujita. Nadel.
• φℓ∗ OXΣ looks like OXΣ . If we set D = 0 in the proof of Lemma 9.2) (9. Hence the lefthand side also vanishes for p > 0. When p > 0. • i sends χ m to χ ℓm .2. • The splitting r : φℓ∗ OXΣ → OXΣ is deﬁned by r(χ m ) = 0 ′ χm q q q ΩXΣ ֒→ φℓ∗ ΩXΣ .1) with L gives a split injection Ωq Σ ⊗XΣ L ֒→ φℓ∗ Ωq Σ ⊗XΣ L .2. We give a proof due to Fujino [54] that uses the morphism φℓ : XΣ → XΣ from Lemma 9. ϕD ◦ φℓ is the support function of ℓD. ΩXΣ ⊗XΣ L ) ֒→ H p (XΣ .20 implies that φ∗ OXΣ (D) comes from a divisor whose ℓ ℓ support function is ϕD ◦ φℓ . the righthand side vanishes for ℓ sufﬁciently large by Serre vanishing (Theorem 9. . we get a sheaf homomorphism q If we tensor i : OXΣ → φℓ∗ OXΣ with (9. ΩXΣ ⊗XΣ L ⊗ℓ ). X X Since φℓ∗ Ωq Σ ⊗XΣ L ≃ φℓ∗ (Ωq Σ ⊗XΣ φ∗ L ) (this is the projection formula from X X ℓ Exercise 9. Sheaf Cohomology of Toric Varieties Proof. Thus (9.3. ℓ ℓ The second key property of φℓ is that there is a split injection (9. Hence φ∗ L = φ∗ OXΣ (D) ≃ OXΣ (ℓD) ≃ OXΣ (D)⊗ℓ = L ⊗ℓ . X X As in the discussion leading up to Lemma 9.6.1.1).6). Tensoring (9.3. Proposition 6. Let L = OXΣ (D) be the associated line bundle. m′ ∈ M.414 Chapter 9.3.3) q M ⊗Z OXΣ ֒→ φℓ∗ q M ⊗Z OXΣ q together with a splitting map φℓ∗ M ⊗Z OXΣ −→ M ⊗Z OXΣ .2.3.2. this gives the injection H p (XΣ . which is what we want. Given these properties of φℓ .2) is a split injection. we obtain a split injection ℓ Ωq Σ ⊗XΣ L ֒→ φℓ∗ (Ωq Σ ⊗XΣ L ⊗ℓ ).3. The morphism ϕℓ has two key properties. with module structure is given by χ m · a = χ ℓm a.0. we get a split injection i : OXΣ → φℓ∗ OXΣ . We can assume that D is torusinvariant with support function ϕD : NR → R.6. M.3) and φ∗ L ≃ L ⊗ℓ . The ﬁrst concerns the pullback φ∗ OXΣ (D).1) We will assume this for now. Since φℓ is multiplication by ℓ. Recall that locally. It remains to prove (9.3. the theorem follows easily.6. q q m ∈ ℓM / m = ℓm′ .
3. q)) = 0 for p > 0.3) induces a map φℓ∗ Ωq Σ → ΩXΣ . Ωq Σ ) ≃ X m∈σ∨ ∩M q Vσ (m) · χ m ⊆ m∈σ∨ ∩M q MC · χ m = Γ(Uσ .3.2) takes q q q q Vσ (m) · χ m to q Vσ (m) · χ ℓm = q Vσ (ℓm) · χ ℓm ⊆ Γ(Uσ . where (σ ⊥ ∩ M) ⊗Z OV (σ) and V (σ) = O(σ) ⊆ XΣ is the orbit closure corresponding to σ ∈ Σ. ΩXΣ ) ′ q since Vσ (m) = Vσ (m′ ).2) and (9.3.3. These subsheaves relate to (9. stated without proof).1). X Proof. Thus (9.2. If XΣ is a complete simplicial toric variety.2) induces a map (9.2.§9. Further proofs have been given by Fujino [54] (noted above) and Mustata [133]. • The map (9. Since V (σ) is a toric variety. q M ⊗Z OXΣ ). Ωq Σ ) = 0 for all p = q. so that φℓ∗ ΩXΣ is a subsheaf X of φℓ∗ q M ⊗Z OXΣ . Theorem 9. When XΣ is simplicial. we have the exact sequence (9. q) = σ∈Σ( j) q− j from Theorem 8.3.1).3. Ωq Σ sits inside q M ⊗Z OXΣ .3) as follows. when m = ℓm′ . Then over Uσ . Thomsen. there is another vanishing theorem involving the sheaves Ωq Σ .3. to ′ Vσ (m) · χ m = q q Vσ (m′ ) · χ ℓm ⊆ Γ(Uσ .3. Since the above sequence has . q) −→ 0 X K j (Σ. then H p (XΣ .19. X This gives the desired split injection (9. we have: • φℓ∗ ΩXΣ looks like ΩXΣ with module structure χ m · a = χ ℓm a. ¸ˇ The Simplicial Case.3. X Theorem 9.1 generalizes the vanishing theorems of Kodaira and Nakano.16. Lauritzen and Mehta [31] in general.2. Since XΣ is simplicial. ΩXΣ ) q since Vσ (m) = Vσ (ℓm). its structure sheaf has vanishing higher cohomology by Demazure vanishing.2. • The map (9. K j (Σ.3. Theorem 8.3. where Vσ (m) ⊆ MC is spanned by the lattice points lying in the same minimal face of σ ∨ as m. Thus (9.2) takes q Vσ (m) · χ m to 0 or.4) 0 −→ Ωq Σ −→ K 0 (Σ. Vanishing Theorems II 415 q By Theorem 8. q) −→ · · · −→ K q (Σ.18 implies Γ(Uσ . Over Uσ . Although Theorem 9. q) −→ K 1 (Σ. and projective toric varieties (Danilov [42].1 was ﬁrst proved by Batyrev and Cox [10] in the simplicial case and by Buch.3. so that H p (XΣ . its name “BottSteenbrinkDanilov” reﬂects the more special vanishing that happens for projective spaces (Bott [20]).3. weighted projective spaces (Steenbrink [167]).
2. Write D = ρ aρ Dρ .2. For x ∈ R.4]).3. Note that Theorem 9. often called integral in this context. we get integral Weil divisors ⌊D⌋ = ⌈D⌉ = i ⌊ai ⌋Di i ⌈ai ⌉Di (the “round down” of D) (the “round up” of D). OXΣ (⌊D⌋)) ֒→ H p (XΣ . 1. ΩXΣ (D)) = 0 whenever p > q or q > p + dim PD .2). we claim that φℓ : XΣ → XΣ from Lemma 9. X q In [121]. A QWeil divisor is QCartier is some positive multiple is integral and Cartier. Ωq Σ ) = 0 when q > p. Then. Then for all p ≥ 0 there is an injection H p (XΣ . Let D be a QWeil divisor on a toric variety XΣ and let ℓ > 0 be an integer such that ℓD is integral.2. Since XΣ is simplicial.3. His result goes as follows. In the literature (see [114.3. When XΣ complete but not necessarily simplicial. the version of Serre duality given in Theorem 9. ⌊x⌋ is the greatest integer ≤ x and ⌈x⌉ is the least integer ≥ x.4. . Rdivisors. With ℓ as above.3. Theorem 9. Thus a positive integer multiple of D is an ordinary Weil divisor. The proof for p > q is relatively easy (Exercise 9. QCartier QWeil divisors are often called Qdivisors. ΩXΣ ) The righthand side vanishes since n − p > n − q.7] proves that H p (XΣ .6 gives a split injection (9. Danilov [42. an easy argument using the long exact sequence in cohomology implies H p (XΣ .3.1). Sheaf Cohomology of Toric Varieties q + 1 terms with vanishing higher cohomology. where D is a nef Cartier divisor. ΩXΣ )∨ ≃ H n−p (XΣ .2 q is the case D = 0 of Theorem 9. Let XΣ be a complete simplicial toric variety. We now prove an injectivity lemma for QWeil divisors due to Fujino [54]. then q H p (XΣ .3. Proof.10 implies n−q q H p (XΣ . Cor. These divisors and their close cousins.3. q.3.6 to our situation.5) OXΣ (⌊D⌋) ֒→ φℓ∗ OXΣ (ℓD). OXΣ (ℓD)). If D is a nef Cartier divisor on XΣ .3. The paper [121] computes H p (XΣ . QWeil Divisors. given a QWeil divisor D = i ai Di . Ωq Σ ) = 0 for p > q (Exercise 9. A QWeil divisor or D on a normal variety X is a formal Qlinear combination of prime divisors. where aρ = bρ + ερ for bρ ∈ Z and 0 ≤ ερ < 1. Lemma 9. ΩXΣ (D)) explicitly for all p. 12.1.416 Chapter 9. Thus ⌊D⌋ = ρ bρ Dρ . are essential tools in modern algebraic geometry. Mavlyutov discusses a vanishing theorem for ΩXΣ (D). X Now suppose p < q. and the result follows. We adapt the proof of Lemma 9.
3. (b) If Σ is complete and D is nef.2.5 extends Demazure vanishing beyond nef QCartier divisors.6. we get the required split injection (9.6.2. Since D1 ∼ D3 and D2 ∼ D4 − rD3 . lemma follows from the discussion leading up to Lemma 9. By Example 6.6) a1 + a3 ≥ ra2 and a2 + a4 ≥ 0.17. 1). OXΣ (⌊D⌋)) = 0 for all p > 0. Pick ℓ > 0 with ℓD Cartier. OXΣ (⌊D⌋)) = m∈P1 ∩M C · χ m .1. then H p (XΣ . OXΣ (−⌈D⌉)) = H p (XΣ .5.3.§9. C[σ ∨ ∩ M]module . and then we are done by Theorem 9. . (a) If Σ is convex and D is nef. the map m → ℓm induces an inclusion C · χ m ֒→ m∈P1 ∩M m∈Pℓ ∩M C · χ m. Let D be a QCartier QWeil divisor on a toric variety XΣ . u2 = (0. provided that the righthand side has module structure χ m · a = χ ℓm a. u3 = (1. −1).3. u4 = (0.4 gives H p (XΣ . it follows that a QWeil divisor D = a1 D1 + · · · + a4 D4 is nef if and only (9. . For part (a).5) without difﬁculty. Then Γ(Uσ .3. The fan for the Hirzebruch surface Hr has minimal generators u1 = (−1. For part (b). and the usual formula for the splitting map r is also a C[σ ∨ ∩ M]module homomorphism.3. OXΣ (−⌈D⌉)) = 0 for all p = dim PD . D4 .3. Pic(Hr )R ≃ {aD3 + bD4  a.2. H p (XΣ . Here is an example that show how part (a) of Theorem 9. OXΣ (−ℓD)). OXΣ (ℓD)) = m∈Pℓ ∩M C · χ m.4. Example 9.3. replacing D with −D in Lemma 9.3. Take an afﬁne open subset Uσ ⊆ XΣ and consider Pℓ = {m ∈ MR  m. It remains to prove the existence of a split injection (9.7. uρ ≥ −ℓaρ for all ρ ∈ σ(1)}. 0 ≤ ερ < 1.5). Proof. . Here are Qdivisor versions of Demazure and BatyrevBorisov vanishing. From here.5). The desired vanishing follows immediately from Lemma 9.3. Vanishing Theorems II 417 Assuming (9. OXΣ (ℓD)) = 0 for p > 0 by Theorem 9.3. Theorem 9. 0). . giving divisors D1 .3. b ∈ R} and the nef cone is generated by D3 and D4 . OXΣ (⌊−D⌋)) ֒→ H p (XΣ . Since Pℓ ∩ ℓM = ℓ(P1 ∩ M). Γ(Uσ . where the ﬁrst equality uses aρ = bρ + ερ . then H p (XΣ . This is a homomorphism. r).
4 also leads to the following result due to Mustata [133].3. . OXΣ (⌊E⌋)) ֒→ H p (XΣ . . . OXΣ (D − B)) ֒→ H p (XΣ .3. we have: . ¸ˇ Theorem 9.3. Since D is ample. OXΣ (ℓE)) from Lemma 9. . .8. Here is a nonQCartier divisor with vanishing higher cohomology. .6) implies that we can ﬁnd ℓ > 0 such that H p (XΣ . In Exercise 9. Sheaf Cohomology of Toric Varieties We will show that the divisors aD3 + bD4 . D4 corresponding to ρ0 . the inclusion H p (XΣ .7. OXΣ (ℓD − B)) = 0. OXΣ (ℓD − B)) = 0 for p > 0. have vanishing higher 1 cohomology when r > 0. ρ4 .6) and hence is nef. Consider the complete fan Σ in R3 shown in Figure 5. b ≥ −1. Serre vanishing (Theorem 9. . Given such a divisor.4 implies H p (XΣ .5. . . Then ⌊D⌋ = aD3 + bD4 has vanishing higher cohomology by Theorem 9. . pick a rational number 0 < ε < 2 and consider the QWeil divisor D = 2εD1 + ε D2 + (a + 1 − ε)D3 + (b + 1 − ε )D4 . Let B = Dρ1 + · · · + Dρr . A complete fan in R3 divisors D0 . OXΣ (D − Dρ1 − · · · − Dρr )) = 0. r r This satisﬁes (9. XΣ .3 you will show that for a QWeil divisor D = a0 D0 + · · · + a4 D4 .0. a. Example 9. Then for p > 0 and any ample Cartier divisor D on XΣ .418 Chapter 9. Let XΣ be a projective toric variety and let ρ1 .3. we get nonnef divisors whose higher cohomology vanishes. Taking a = −1 or b = −1. Since ⌊E⌋ = D − B. It has z ρ1 ρ2 ρ4 ρ3 y x ρ0 Figure 5.3. we have H p (XΣ . Now let E = D − ℓ−1B. p > 0. ρr ∈ Σ(1) be distinct.3. .3. ♦ Lemma 9. Proof. .
Lemma 9.28. Proof. Since D is nef. Vanishing Theorems II 419 • D is QCartier if and only a1 + a3 = a2 + a4 . its polytope PD is a lattice polytope. Given a nef Cartier divisor D on a complete toric variety XΣ and an integer ℓ > 0. The map φWℓ and the polytope ℓPD are related as follows. In particular.1.9 implies that a nef Cartier divisor D on a complete toric variety XΣ has Iitaka dimension κ(XΣ .3. Def. Also recall that since D is nef. For ℓ ≫ 0. Here is the toric version of their result.2.22 (Exercise 9. When D is Cartier. OX (KX + D)) = 0 for all p > 0 when D is an ample line bundle on a smooth projective variety X . ♦ Iitaka Dimension and Big Divisors. the image of φWℓ is isomorphic to the toric variety of ℓPD . In general. 2. . This follows from Proposition 6.10 (Toric KawamataViehweg).4). which for a nef Cartier divisor D on a complete toric variety XΣ means D is big ⇐⇒ dim PD = dim XΣ . OXΣ (KXΣ + D))∨ ≃ H n−p (XΣ . D) = dim PD . Proof. Serre duality implies H p (XΣ . dim φWℓ (XΣ ) = dim PD for ℓ ≫ 0. KawamataViehweg. the global sections Wℓ = H 0 (XΣ .1. In the 1982. In this terminology.3. In particular. The classic version of Kodaira vanishing can be stated as H p (X . D) (see [114. the condition on p becomes n − p = n. and since D is big. Kawamata and Viehweg independently weakened the hypotheses on D. Yet D4 is not QCartier. Let n = dim XΣ .7.0.3. Lemma 9.3. p = 0.§9. It should be clear what it means for a QCartier QWeil divisor to be big and nef.7. i. This situation is a special case of the deﬁnition of the Iitaka dimension κ(X . Theorem 9. OXΣ (ℓD)) give a morphism φWℓ : XΣ → P(Wℓ∨ ) as in Lemma 6. OXΣ (−D)).3]) of a Cartier divisor D on a complete irreducible variety X . • D is QCartier and nef if and only a1 + a3 = a2 + a4 ≥ 0.e.9.3. Let XΣ be a complete toric variety and let D be a QCartier QWeil divisor on XΣ that is big and nef. D = D3 + D4 is QCartier and nef. the latter vanishes for n − p = dim PD by Theorem 9. a Cartier divisor is big if it has maximal Iitaka dimension. Then H p (XΣ . OXΣ (KXΣ + ⌈D⌉)) = 0 for all p > 0. so that D4 = D − D3 has vanishing higher cohomology by Theorem 9.3.
uρ ≤ 0 for ρ ∈ σ(1)} since m. simp . OXΣ (KXΣ + ⌈D⌉))∨ ≃ H n−p (XΣ . It is also easy to see that rσ is compatible with rτ whenever τ is a face of σ. Proposition 9. Thus we get a map simp r : W → VD. Here is a toric version of GrauertRiemenschneider vanishing. The sum ρ∈A λρ is nonzero since u ∈ W ∩ σ. Thus (9. OXΣ (⌊E⌋)) ֒→ H p (XΣ . Then H p (XΣ .m . Thus the inclusion H p (XΣ .m = {σ ∈ Σ  m.11.5). Fix p ≥ 1 and m ∈ M. Theorem 9.420 Chapter 9.3. so that rσ (u) ∈ Conv(uρ  ρ ∈ A) ⊆ VD. The assumption W ∩ σ = ∅ implies that A = {ρ ∈ σ(1)  m. Sheaf Cohomology of Toric Varieties In general. OXΣ (−⌈D⌉)) (Exercise 9. and when u ∈ W ∩ σ. deﬁne rσ (u) = ( ρ∈A λρ )−1 ρ∈A λρ uρ . Proof. u ≤ 0}. we construct rσ : W ∩ σ → VD. λρ ≥ 0.7) VD.3. OXΣ (KXΣ + ⌈D⌉)) ֒→ H p (XΣ . Our next vanishing result features the higher direct images of the canonical sheaf.m ⊆ W . Note also that ΣD.m ∩ σ as follows. Consider the set W = {u ∈ Σ \ {0}  m.10 is to prove a version of Serre duality that implies H p (XΣ .m . C). Let E = D + ℓ−1 KXΣ . Given σ ∈ Σ with W ∩ σ = ∅. OXΣ (KXΣ + ℓD)).3. uρ ∈ Z. GrauertRiemenschneider. simp simp simp The proposition will follow once we prove that (9.1. OXΣ (ℓE)) from Lemma 9. We will need the following preliminary result.m .6. where VD.4 implies H p (XΣ .3. u ≤ 0} = (Σ \ {0}) ∩ {u ∈ NR  m.m Conv(uρ  ρ ∈ σ(1)) and ΣD. write D = ρ aρ Dρ where aρ = bρ − ερ for bρ ∈ Z and 0 ≤ ερ < 1. ωXΣ ) = 0 for all p > 0.m = σ∈ΣD. which is the desired retraction by Exercise 9.2 with D = KXΣ = − simp ρ Dρ implies H p (XΣ . uρ < 1 for all ρ ∈ σ(1)}. Another approach to Theorem 9. Since σ is simplicial. ωXΣ )m = H p (XΣ . u ∈ σ can be uniquely written u = ρ∈σ(1) λρ uρ .3.7) is a deformation retract. Then we are done by the Cartier case already proved. Since ⌊bρ − ερ − ℓ−1 ⌋ = bρ − 1. Pick ℓ such that ℓD is Cartier and 0 < ερ + ℓ−1 < 1 for all ρ. uρ ≤ 0} = ∅. we have ⌊E⌋ = ⌈D⌉ + KXΣ . Let Σ be a simplicial fan whose support Σ is strongly convex.3.3. OXΣ (D))m ≃ H p−1 (VD. and observe that W is convex since Σ is strongly convex.m = {σ ∈ Σ  m.
Use (9.3. Prove H p (X.3. then φ∗ ωXΣ ≃ ωXΣ′ . so that H p (φ−1 (Uσ ). Our hypothesis on φ implies that the associated lattice map φ : N → N ′ ′ induces an isomorphism φR : NR ≃ NR such that Σ is the inverse image of Σ′ . (a) Show that D0 ∼ 2D3 + 2D4 .3.3. .12 (Toric GrauertRiemenschneider Vanishing).3 when p > q.3. Hint: Replace φℓ : N → N with ℓN ⊆ N as in the proof of Theorem 9.10. given σ ∈ Σ′ . Exercises for §9.1.5.2.3.7). A sheaf G on a variety X is acyclic if H p (X. (c) Show that D is QCartier and nef if and only a1 + a3 = a2 + a4 ≥ 0. Most versions of Theorem 9.3. we see that φR (σ) is strongly convex. . 9. . then R p φ∗ ωXΣ = 0 for all p > 0. Both ¸ˇ Fujino [54] and Mustata [133] state general vanishing theorems that imply versions of Theorems 9.1.3.15 (Exercise 9.§9.8 and consider a QWeil divisor D = a 0 D0 + · · · + a 4 D4 . . .7 and 9.3.3.3.2. Proof. Vanishing Theorems II 421 Theorem 9. F ) = 0 for all p > r by induction on r.10. There are many more toric vanishing theorems.3. (a) Pick ℓ > 0 such ℓD is Cartier and consider OXΣ (⌊D⌋) ֒→ φℓ∗ OXΣ (ℓD) from (9. ωXΣ ) = 0 for p > 0 by Proposition 9.5).3. Then R p φ∗ ωXΣ = 0 for p > 0 by Proposition 9. . G ) = 0 for all p > 0. Use this to prove that OXΣ (⌊D⌋) is MCM. Let φ : XΣ → XΣ′ be a surjective proper toric morphism between toric varieties of the same dimension. Further vanishing results can be found in [58] and [143].3.3.12 in the literature assume that XΣ is smooth.3. .11. 9. Gr are acyclic.0. 9. D2 ∼ D4 . Now suppose that we have an exact sequence of sheaves 0 −→ F −→ G0 −→ G1 −→ · · · −→ Gr −→ 0. The ﬁnal assertion of the theorem is an easy consequence of Theorem 8. . D4 be the divisors from Example 9.3. Let D be a QCartier QWeil divisor on a complete toric variety XΣ . This is the support of the fan of φ−1 (Uσ ). Complete the proof of Lemma 9. Let D0 . 9. If XΣ is simplicial. (b) Show that D is QCartier if and only a1 + a3 = a2 + a4 .3. .9.2.7. If in addition φ is birational.5.3.3. Other Vanishing Theorems. where G0 . 9.3. 9. −1 Thus. 9.4. D1 ∼ D3 .4) and the previous exercise to prove Theorem 9.
12.8. OXΣ (D + ℓ(D + E))) for all p ≥ 0. This follows from the long exact sequence in cohomology (Exercise 9.3.10 using part (c) and Theorem 9. OXΣ (KXΣ − ⌊D⌋)). let D be a Weil divisor and E = ρ aρ Dρ a QWeil divisor such that 0 ≤ aρ ≤ 1 for all ρ and ℓE is integral for some integer ℓ > 0. Prove that Σ is a reﬁnement of Σ0 .4.1 from [133]. F ) < ∞ for all p. Sheaf Cohomology of Toric Varieties (b) Adapt the proof of Theorem 9. OXΣ (KXΣ + ⌈D⌉)). OXΣ (⌊D⌋))∨ ≃ H n−p (XΣ . Hence χ(F ) is a welldeﬁned integer. 9.2.3.10 to show that H p (XΣ . Consider the map rσ : W ∩ σ → VD. F ).422 Chapter 9.3. Hint: As suggested by [54]. F ) = 0 for p > dim X by [78. Our hypotheses on X and F guarantee that dim H p (X . Given a toric variety XΣ . Then formulate and prove a similar result for r.3.4. rσ is homotopic to the identity.m .6. Let φ : XΣ → XΣ′ be a proper birational toric morphism.4. −1 simp §9.11. .3. This is a strengthened version of Theorem 0. (b) Prove that when regraded as a map from W ∩σ to itself. The Euler Characteristic of a Sheaf . Hint: ⌊−D⌋ = −⌈D⌉. (c) Prove H p (XΣ . ℓ apply Lemma 9. (d) Prove Theorem 9. (a) Prove that φ : N → N ′ is an isomorphism. Prove that there is an injection H p (XΣ . Its Euler characteristic χ(F ) is deﬁned to be the alternating sum χ(F ) = p≥0 (−1) p dim H p (X .3.7. Let F be a coherent sheaf on a complete variety X . and H p (X .1). III. (b) Let Σ0 be the fan in NR consisting of the cones φR (σ) for σ ∈ Σ′ .3.4 to the QWeil divisor D + ℓ+1 E.7]. Thm. simp 9.5.3. OXΣ (D)) ֒→ H p (XΣ .m ∩ σ deﬁned in the proof of Proposition 9. The Euler characteristic satisﬁes (9. OXΣ (−⌈D⌉))∨ ≃ H n−p (XΣ . Applications to Lattice Polytopes In this section we use vanishing theorems to study lattice polytopes.2. (a) Prove that these maps patch to give a retraction r : W → VD.1) χ(G ) = χ(F ) + χ(H ) whenever we have an exact sequence 0 → F → G → H → 0 of coherent sheaves. 9. (c) Complete the proof of Theorem 9.
Applications to Lattice Polytopes 423 where as usual L ∨ = HomOX (L .2.4. OPn (ℓ)) = 0 for p > 0.2.2 and 9. (9. ℓ ∈ Z.4.4.4. note that p(ℓ) = (−1)n −ℓ−1 n (Exercise 9. We will see below that this is a special case of Ehrhart reciprocity.4. is a polynomial in ℓ.3 since the polytope associated to ℓD0 is ℓ∆n . where the last equality uses Exercise 9.4. so that (9.1.6. ℓ ∈ Z. the Hilbert polynomial is written χ(F (ℓD)). via the twisting convention introduced in §9.4.4. Note also that when ℓ > 0.4.2) χ(OPn (ℓ)) = dim H 0 (P n .8 implies (9.3).3) χ(OPn (ℓ)) = (−1)n dim H n (P n .3) imply that p(ℓ) = ℓ∆n ∩ Zn  p(−ℓ) = (−1)n Int(ℓ∆n ) ∩ Zn . When ℓ < 0. Example 9.2. Then p(ℓ) = χ(OPn (ℓ)) for ℓ < 0 by (9. OPn (ℓ)) = ℓ∆n ∩ Zn  = ℓ+n n .4 implies that H p (P n . A basic result [105] states that χ(F ⊗OX L ⊗ℓ ). When ℓ < 0.2). deﬁne L ⊗OX · · · ⊗OX L ℓ times ⊗ℓ L = (L ∨ )⊗(−ℓ) OX ℓ>0 ℓ<0 ℓ = 0.3.3. Example 9. ♦ .4. OPn (ℓ)) = (−1)n Int(ℓ∆n ) ∩ Zn  = (−1)n −ℓ−1 n . You will prove the last equality in Exercise 9. We will compute χ(OPn (ℓ)).4. where ∆n is the standard nsimplex from Example 4. Example 9. OX ).4. ℓ+n n We claim that p(ℓ) = χ(OPn (ℓ)) for all ℓ ∈ Z.§9. Now consider the polynomial p(x) = and observe that p(ℓ) = (x + n)(x + n − 1) · · · (x + 1) ∈ Q[x] n! when ℓ ∈ N.4). called the Hilbert polynomial. When L = O(D) for a Cartier divisor D.4. For ℓ ≥ 0 this follows easily from (9.2) and (9. When ℓ ≥ 0. Exercises 9. The second equality follows from Proposition 4.4. Given a line bundle L on X .4.3 sketch a proof in the ample case.
and the program Normaliz [30] computes Ehrhart polynomials. which is a polynomial for ℓ ≫ 0 (see [37.4. ℓ ≥ 0. There is a polynomial EhrP (x) ∈ Q[x] such that if ℓ ∈ N. An elementary approach to Ehrhart polynomials and Ehrhart Reciprocity can be found in [12].2. then (9. Furthermore. 6.3) and Example 4. . Ch. xs ] gives the homogeneous coordinate ring C[XP ] = C[x1 . Let P ⊆ MR ≃ Rn be a full dimensional lattice polytope. . Its homogeneous ideal I(XP ) ⊆ C[x1 .2 (Ehrhart Reciprocity). if ℓ ∈ N is positive. §4]). n n Theorem 9. xs ]/I(XP ). . . The very ample divisor DP on XP gives a projective embedding i : XP ֒→ P s−1 . This is the Hilbert polynomial of XP ⊆ P s−1 . For example. Given a full dimensional lattice polytope P ⊆ MR .2. The case when ℓ ≥ 0 is easy since ℓDP is basepoint free by Proposition 6. where s = P ∩ M. OXP (−ℓDP )) = (−1)n Int(ℓP) ∩ M = (−1)n L∗ (ℓP) by BatyrevBorisov vanishing (Theorem 9. then EhrP (ℓ) = L(ℓP). When P is very ample.424 Chapter 9. the functions L(P) = P ∩ M L∗ (P) = Int(P) ∩ M count lattice points in P or in its interior. the Ehrhart polynomial has a nice interpretation.3.4. if ℓ > 0 is an integer. We will show that the Hilbert polynomial EhrP (ℓ) = χ(OXP (ℓDP )) has the desired properties. Sheaf Cohomology of Toric Varieties The Ehrhart Polynomial. This graded ring has a Hilbert function ℓ −→ dim C[XP ]ℓ . then EhrP (−ℓ) = (−1)n L∗ (ℓP). The polynomial EhrP in Theorem 9.3) imply that L(ℓ∆n ) = ℓ+n and L∗ (ℓ∆n ) = ℓ−1 .4. Thus χ(OXP (ℓDP )) = dim H 0 (XP .4.2) and (9. hence χ(OXP (−ℓDP )) = (−1)n dim H n (XP . . Let XP be the toric variety of P and DP the associated ample divisor. .1. .2 is called the Ehrhart polynomial of P. Then PℓDP = ℓP is full dimensional.4.7. OXP (ℓDP )) = ℓP ∩ M = L(ℓP) by Demazure vanishing (Theorem 9. Now assume ℓ > 0. . Proof. .7).
4. Lem. e1 . Thus EhrP (x) = Area(P) x2 + 1 B x + 1. OXP (ℓDP )) → H 1 (P s−1 . . The reason for the 1 2 will soon become clear. . tensoring with OPs−1 (ℓ) gives an exact sequence 0 −→ IXP (ℓ) −→ OPs−1 (ℓ) −→ OXP (ℓDP ) −→ 0. . 1 The leading coefﬁcient of EhrP is 2 Vol(P). . scaled so that ∆n has volume equal to 1. including [12. en ) ⊆ MR . Consider the exact sequence 0 −→ IXP −→ OPs−1 −→ OXP −→ 0. we use the normalized volume in MR .4. since EhrP (0) = L(0 · P) = 1 by Theorem 9. . ℓ→∞ ℓn n! This is proved in many places.2. We can describe the degree and leading coefﬁcient of the Ehrhart polynomial. . IXP (ℓ)) → · · · . . The H 1 term vanishes for ℓ ≫ 0 by Serre vanishing (Theorem 9. . which is the usual Euclidean area Area(P) since the normalized volume of the unit square is 2.4. If P is a very ample. 2 where B is yet to be determined.4) Vol(P) L(ℓP) = lim . Here is a classic application of these ideas. Since L(ℓP) is given by the polynomial EhrP (ℓ). . .4.6).6).4. Proof. en be a basis of M and consider the simplex ∆n = Conv(0. xs ]ℓ → H 0 (XP . . Hence for ℓ large. This implies that the Hilbert polynomial of XP is the Ehrhart polynomial of P. The constant term is also easy to compute.4.3.4. then the Ehrhart polynomial EhrP equals the Hilbert polynomial of the toric variety XP under the projective embedding given by the very ample divisor DP . For the leading coefﬁcient. . Let P ⊂ R2 be a lattice polygon with Ehrhart polynomial EhrP . . Applications to Lattice Polytopes 425 Proposition 9. its normalized volume is denoted Vol(P) and is computed by the limit (9. Given a full dimensional lattice polytope P ⊆ MR .0. Thus the “unit cube” { n λi ei  0 ≤ λi ≤ 1} has i=1 normalized volume n!. Since OXP (DP ) is the restriction of OPs−1 (1) to XP .4. OXP (ℓDP )). Example 9. we get an isomorphism C[XP ]ℓ ≃ H 0 (XP . In Exercise 9.§9.5 you will show that the resulting long exact sequence is 0 → I(XP )ℓ → C[x1 . Then the normalized volume is the usual ndimensional Lebesgue measure.19]. 3. Let e1 . it follows easily that EhrP has degree n and its leading coefﬁcient Vol(P)/n! (Exercise 9.
we use the sheaves ΩXP to generalize the Ehrhart polynomial when the lattice polytope P is simple.426 Chapter 9.4. It follows that ΣP is simplicial whenever P is simple. is the unique polynomial in Q[x] that satisﬁes p p EhrP (ℓ) = χ(ΩXP (ℓDP )). The latter equality is obvious since 0 · P = {0}.2) is more subtle since it can fail when we replace the polytope P with a polytopal complex (a collection of polytopes where the intersection of any two is a face of each). 2 This equation.4.4.4. Sheaf Cohomology of Toric Varieties By Ehrhart reciprocity. Following Materov [118]. the Euler characteristic χ(ΩXP (ℓDP )) is a polynomial function of ℓ. called Pick’s formula. we have (9. shows how to compute the area of a lattice polygon in terms of its lattice points. Hence each vertex has n facet normals that generate the corresponding cone in the normal fan ΣP .7. we noted that EhrP (0) = L(0 · P) = 1. Recall that a full dimensional lattice polytope P ⊆ MR ≃ Rn is simple if every vertex is the intersection of exactly n facets.5. Thus ∂∆2 consists of three line segments. More generally. ♦ In Example 9. Consider the boundary ∂∆2 of the 2simplex ∆2 ⊆ R2 . ℓ ∈ Z. Example 9. Further examples (and an explanation) will be given in Execise 9. p Then the pEhrhart polynomial of P.5) for the area gives Area(P) = L∗ (P) + 1 ∂P ∩ M − 1. Hence the toric variety XP is simplicial. Here is a simple example. ℓ ∈ Z. We proved earlier that the Ehrhart polynomial EhrP (x) of P satisﬁes EhrP (ℓ) = χ(OXP (ℓDP )). . However. Furthermore. 2 Solving for B gives B = L(P) − L∗ (P) = ∂P ∩ M.4. denoted EhrP (x). where DP is the ample divisor coming from P. Thus the Ehrhart polynomial of P is 1 EhrP (x) = Area(P) x2 + 2 ∂P ∩ M x + 1.5) Area(P) + 1 B + 1 = EhrP (1) = L(P) 2 Area(P) − 1 B + 1 = EhrP (−1) = (−1)2 L∗ (P) = L∗ (P). This gives the “Ehrhart polynomial” Ehr∂∆2 (x) = 3x with the property that Ehr∂∆2 (ℓ) = L(ℓ ∂∆2 ) for integers ℓ > 0. One easily computes that L(ℓ ∂∆2 ) = (ℓ ∂∆2 ) ∩ Z2  = 3ℓ when ℓ > 0 is an integer. However. given an integer p 0 ≤ p ≤ n. the equality EhrP (0) = L(0 · P) (proved in Theorem 9. solving the bottom equation of (9.4. ♦ p The pEhrhart Polynomials. where ∂P is the boundary of P. Ehr∂∆2 (0) = 3 · 0 = 0 = L(0 · ∂∆2 ) = L({0}) = 1.4.
§9. we deﬁne L(Q) = Q ∩ M L∗ (Q) = Relint(Q) ∩ M.4. This proves that (9. if Q is a face of P. (d) If 0 ≤ p ≤ n.4. These invariants are related to the pEhrhart polynomials by the following result of Materov [118]. then n p EhrP (ℓ) = i=p i p L∗ (ℓQ). Q∈P(n−i) (c) If 0 ≤ p ≤ n. we see that (−1) p h p = (−1)n (−1)n−p hn−p . Furthermore.6) h p = hn−p for 0 ≤ p ≤ n. Applications to Lattice Polytopes 427 0 Note that EhrP (x) is the ordinary Ehrhart polynomial EhrP (x). Theorem 9. First. then p n−p EhrP (−x) = (−1)n EhrP (x). (a) If ℓ > 0 is an integer. then p p EhrP (ℓ) = i=0 (−1)i n−i n− p L(ℓQ).4. then p EhrP (0) = (−1) p h p . Theorem 9. p The fi are the face numbers of P.6. We will defer the proof for now. setting x = 0 in part (c) and using part (d). Q∈P(i) (b) If ℓ ≥ 0 is an integer. These are called the DehnSommerville equations. To state the p properties of EhrP (x).4. . Let XP be the toric variety of a full dimensional simple lattice polytope P ⊆ MR ≃ Rn . we will use the following notation: P(i) = {Q  Q is an idimensional face of P} fi = P(i) = # idimensional faces of P n hp = i=p (−1)i−p i fi .6 has some nice consequences.
2. set D = ℓDP . . Given an integer ℓ > 0 and m ∈ (ℓP) ∩ M. let VℓP (m) be the subspace of MC such that VℓP (m) + m is the smallest afﬁne subspace of MC containing the minimal face of ℓP containing m. To compute H 0 (XP . Comparing the polytopes of D − Dρ and D = ρ aρ Dρ shows that H 0 (XP .7.6 simultaneously generalizes the DehnSommerville equations and Ehrhart reciprocity. Let P ⊆ MR ≃ Rn be a full dimensional lattice polytope with toric variety XP and ample divisor DP .4. we get the exact sequence p 0 −→ H 0 (XP .6) with OXP (D). uρ = −ℓaρ otherwise. Lemma 9. Using ρ⊥ ∩ M ⊆ M and tensoring (8.0. OXP (D − Dρ)) = 0 courtesy of Theorem 9. Since H 0 (XP . these equations reduce to Ehrhart reciprocity: 0 EhrP (ℓ) = L(ℓP) 0 EhrP (−ℓ) = (−1)n L∗ (ℓP). we obtain the exact sequence p 0 −→ ΩXP (D) −→ p M ⊗Z OXP (D) −→ p−1 ρ M ⊗Z ODρ (D). Proof. OXP (D)) = m∈(ℓP)∩M C · χ m . recall from Proposition 4. if we write EhrP (ℓ) using part (b) of the theorem and EhrP (ℓ) using part (a). We adapt the strategy used in the proof of Proposition 8. Take global sections over XP and consider the graded piece for m ∈ M. To simplify notation.7. ODρ (D)) −→ 0. ΩXP (D))m −→ p MC −→ p−1 ρ M ⊗Z H 0 (XP . Then p p H 0 (XP . Tensor this with OXP (D) and take global sections to obtain 0 −→ H 0 (XP . OXP (D)) −→ H 0 (XP .3.18. ODρ (D))m when m ∈ (ℓP) ∩ M. OXP (D − Dρ )) −→ H 0 (XP . Sheaf Cohomology of Toric Varieties p n−p Also. where the exactness on right follows from H 1 (XP .2. Q∈P(n−i) n p n−p EhrP (−ℓ) = (−1)n EhrP (ℓ) = (−1)n i=n−p i n− p L∗ (ℓQ). ΩXP (ℓDP )) = m∈(ℓP)∩M VℓP (m) · χ m . The proof of the theorem will use the following lemma. Q∈P(i) For p = 0 and ℓ > 0.428 Chapter 9. ODρ (D))m . ODρ (D))m = C 0 m. Thus Theorem 9.4. then part (c) gives the following formulas for ℓ > 0: p p EhrP (ℓ) = i=0 (−1)i n−i n− p L(ℓQ).28 that we have an exact sequence 0 −→ OXP (−Dρ ) −→ OXP −→ ODρ −→ 0.
16. Applications to Lattice Polytopes 429 Let Fρ be the facet of ℓP corresponding to ρ. where β p is a sum of contraction maps iuρ deﬁned in the discussion leading up to Theorem 8. ΩXP (ℓDP )) = 0 for q > 0 by Theorem 9. then L∗ (ℓQ) = (−1)i EhrQ (−ℓ) by Ehrhart reciprocity.4. one sees easily that ⊥ m∈Fρ ρ . ΩXP (ℓDP )) = i≥0 i p n L (ℓQ) = Q∈P(i) i=p ∗ i p L∗ (ℓQ). m. we obtain the polynomial identity n (9. When this happens.3. note that m ∈ Relint(ℓQ) ∩ M ⇐⇒ Q is the minimal face of P containing m. Q∈P(i) . Thus ω ∈ p MC is in the kernel of β p if and only if iuρ (ω) = 0 for all ρ with m ∈ Fρ .6. Lemma 9.§9.8 that iuρ (ω) = 0 if and only if ω ∈ p (ρ⊥ )C . We begin with part (a).7 implies that p dim H 0 (XP . C This and (9. Since F = m∈Fρ Fρ is the minimal face of ℓP containing m.1. Q∈P(i) p We also have H q (XP . For m ∈ ℓP. we obtain dim H 0 p (XP . Proof of Theorem 9.2. ΩXP (ℓDP )) = i=p i p L∗ (ℓQ).4. We are now ready to prove our main result.2. If ℓ > 0 and Q ∈ P(i). p SpanR (m0 − m ∈ M  m0 ∈ F) = Thus VℓP (m) = m∈Fρ (ρ ⊥) . We know from Exercise 8. ΩXP (ℓDP )) = m∈(ℓP)∩M dim VℓP (m) p when ℓ > 0.7) m∈Fρ p (ρ⊥ )C = p m∈Fρ (ρ ⊥ )C . p Now consider EhrP (0). Q∈P(i) which proves part (a).4. Hence n p p p EhrP (ℓ) = χ(ΩXP (ℓDP )) = dim H 0 (XP . Given a face Q of P and m ∈ (ℓP) ∩ M.8) is the intersection (9. Hence we get an exact sequence p 0 −→ H 0 (XP .4.4.7) imply ker(β p ) = VℓP (m).4.2. ΩXP (D))m −→ p MC −→ βp p−1 m∈Fρ MC . we have dim Q = dim VℓP (m).8) p EhrP (x) = i=p i p (−1)i EhrQ (−x). Since the previous display holds for all ℓ > 0. uρ = −ℓaρ if and only if m ∈ Fρ . Since the idimensional faces of ℓP are ℓQ for Q ∈ P(i). It follows that the kernel of β p in (8.
Sheaf Cohomology of Toric Varieties Setting x = 0 gives n p EhrP (0) = i=p (−1)i i p EhrQ (0) = (−1) p h p Q∈P(i) by the deﬁnition of h p .8. This completes the proof.4. Let P be a simple 3dimensional lattice polytope in R3 . The DehnSommerville equations can be written h3 = h0 .8–5. so f3 = f0 − f1 + f2 − f3 h2 = h1 . 5.8).4. proving part (d). ΩXP (ℓDP ))∨ ≃ H n−q (XΣ . Examples. take ℓ ≥ 0 and consider n p n−p EhrP (ℓ) = (−1)n EhrP (−ℓ) = (−1)n i=n−p i n− p (−1)i EhrQ (ℓ). Q∈P(i) where the last equality follows by setting j = n − i. The ﬁrst equation Euler’s celebrated formula V − E + F = 2.4. We ﬁrst give a classic example of the DehnSommerville equations. Since ℓ ≥ 0 implies EhrQ (ℓQ) = L(ℓQ). which holds for any 3dimensional polytope. Ch. which implies (9. 5] for details and references.4. Thus EhrP (ℓ) = (−1)n EhrP (−ℓ) for ℓ ∈ Z. ΩXP (−ℓDP )) since ℓDP is Cartier for any ℓ ∈ Z. but when combined with the ﬁrst reduces to 2 f1 = 3 f0 . for part (b). For part (c). so f2 − 3 f3 = f1 − 2 f2 + 3 f3 .10]. This easily implies p n−p χ(ΩXP (ℓDP )) = (−1)n χ(ΩXP (−ℓDP )) p n−p (Exercise 9. Ex. Identities equivalent to Theorem 9.8) with p replaced by n − p. This holds because is P is simple—every vertex meets three edges and every edge meets two vertices.4.430 Chapter 9. we obtain f0 − f1 + f2 = 2 f1 = 3 f2 − 6. See [12.6 were discovered by McMullen in 1977 in a very different context. we obtain n p EhrP (ℓ) = (−1)n−i i=n−p i n− p p L(ℓQ) = Q∈P(i) j=0 (−1) j n− j n− p L(ℓQ). Q∈P(i) where the ﬁrst equality uses part (d) and the second uses (9. including a nontoric version of Theorem 9. Finally.4. The second seems more mysterious. and part (d) follows. we use Serre duality as given in Theorem 9.9) p n−p H q (XΣ . Example 9.2.10. ♦ . Since f3 = 1 (P is the only 3dimensional face of itself).6 in [12.
which is simplicial since P is simple (Exercise 9.4. Given ℓ > 0. ℓ p Ehr∆n (ℓ) = ℓ−1 p i≥0 where the last equality uses the Vandermonde identity discussed in Exercise 9.9).§9. note that L∗ (ℓ∆n ) = Q∈∆n (i) n+1 n−i ℓ−1 i because • An nsimplex has n+1 i+1 = n+1 n−i faces of dimension i. Sec.6. . 5.3]. ΩPn (ℓ)) = .10. Example 9. Theorem 9. p ℓ This formula was ﬁrst proved by Bott [20] in 1957. We next give an application of Theorem 9.6]. 8.4.6 gives the formula n p Ehr∆n (ℓ) = i=p i p ℓ−1 i ℓ−1 p ℓ−p−1 i−p i p n+1 n−i ℓ−1 .9. this becomes n+1 n−i ℓ− p−1 ℓ−1 = i− p p n+ℓ− p . Let XP be the toric variety of a full dimensional simple lattice polytope in P ⊆ MR ≃ Rn . Ch. Given a simple polytope P as above.6. More on the DehnSommerville equations and toric varieties can be found in [59. the sheaf ΩXP has Euler characteristic (9. 5] and [179. i Since = .4.3).4. Then p dim H q (XP . We also recommend [12.4.4.4. The factor of (−1) p is explained as follows.10) p p χ(ΩXP ) = EhrP (0) = (−1) p h p by Theorem 9. The standard nsimplex ∆n has P n as its associated toric variety. ♦ p Cohomology of pForms.10. • Each Q ∈ ∆n (i) is lattice isomorphic to the standard isimplex and hence has ℓ−1 interior lattice points by (9. i Then part (a) of Theorem 9. Applications to Lattice Polytopes 431 There is a version of the DehnSommerville equations for the dual polytope P◦ ⊆ NR . Hence ℓ−1 n+ℓ− p p dim H 0 (P n .4. Sec.4.4. ΩXP ) = hp 0 q= p q = p.
In Chapter 12. 9. (a) Let a.10 with XΣ = P n .10). there is an exact sequence 0 −→ Fr −→ · · · −→ F1 −→ F0 −→ F −→ 0. It follows that p p χ(ΩXP ) = (−1) p dim H p (XP . j When XP is smooth. where F (ℓ) = F ⊗OPn OPn (ℓ). xn ]. . The toric variety XP is simplicial. (3. we will show that h p = dim H 2p (XP .4.0.11. 6.4. so that Theorem 9.k (x) ∈ Q[x] such that pa. (b) Use part (a) together with (9.4.2 applies to XP . 9.1. Exercises for §9. ΩXP ). (c) Let F be a coherent sheaf on a complete variety X and let D be a very ample divisor on X. . k be integers with k ≥ k0 . Prove (9.4. (a) Let F be a ﬁnitely generated graded module over the polynomial ring S = C[x0 . Now let F be a coherent sheaf on P n .3.0. Thm.2 to show that there is a polynomial pa. [72.1) and Example 9. . Use Exercise 9. C).1).4. Prove that there is an exact sequence of sheaves 0 −→ Fr −→ · · · −→ F1 −→ F0 −→ F −→ 0.8)]. C) ≃ i+ j=2p H i (XP . . In the simplicial case. for example. Thus there is k0 > 0 such that kD is very ample for all k ≥ k0 . Ch. By [37. this is covered in [168]. 116]. The theorem follows by comparing this with (9.4. More precisely. Use part (b) and Exercise 9.4.1 to show that χ(F (ℓ)) is a polynomial in ℓ ∈ Z. p A different proof of dim H p (XP .2. where each Fi is a ﬁnite direct sum of sheaves of the form OPn (a). ΩXP ). ΩXP ) = 0 for q = p. Here are some cases where it is easy to show that Euler characteristics give Hilbert polynomials. ΩXP ) = h p will be sketched in Exercise 9.10 is the Hodge decomposition H 2p (XP .4. where each Fi is a ﬁnite direct sum of modules for the form S(a). Let F be a coherent sheaf on a projective variety X. Sheaf Cohomology of Toric Varieties Proof.k (ℓ) = χ(F ((a + ℓk)D)) for all ℓ ∈ Z.4.6 to show that χ(F (ℓD)) is a polynomial in ℓ ∈ Z. the sum n i (−1)i−p hp = fi p i=p will arise naturally when we compute the singular cohomology of a simplicial toric variety. and let D be an ample divisor on X. 9.432 Chapter 9. The link between this and Theorem 9. this is a classical fact—see. p.3. This p shows that H q (XP .4.4. Hint: Apply Example 6. .
m (x/m) are equal when k. .5.5. (b) Show that the Hilbert polynomial of the coordinate ring C[X] = S/I(X) is the Euler characteristic χ(OX (ℓ)). xn by Exercise 4. This divides the remaining 1’s into n + 1 groups.6. Hint: Show that shifting the interior n lattice points by (1.4. The number of elements in each group gives the exponents of x0 .4. Prove that (ℓ · ∂B) ∩ Z2  = 8ℓ − 1. so that Ehr∂∆2 (x) = 3x.7. (b) Use part (a) to prove the assertion made following (9.§9. Applications to Lattice Polytopes 433 (b) Show that the polynomials pa. 13 of Ch. 1) gives the lattice points in (ℓ − n − 1)∆n. Here is a combinatorial proof. this gives a different way of seeing that the constant term of EhrP is 1.4. (c) Show that the polynomials pa (x − a) and pb (x − b) are equal for any a. (a) Prove that n = deg(p(x)) and that the above limit is the leading coefﬁcient of p(x). We know from Example 4. .4) concerning the degree and leading coefﬁcient of the Ehrhart polynomial. . OPn (ℓ)) → H 0 (P n . the constant term of its Ehrhart polynomial EhrP is e(P) (see [116]). 9. so that Ehr∂B (x) = 8x − 1.4. See also [36. where S = C[x0 . b ∈ Z and conclude that there is a polynomial p(x) ∈ Q[x] such that p(ℓ) = χ(F (ℓD)) for all ℓ ∈ Z. where i : X ֒→ P n is the inclusion map. . IX (ℓ)) → H 0 (P n .4. Show that (9. where I(X) ⊆ S is the ideal of X. Conclude that there is a polynomial pa (x) such that pa (ℓ) = χ(F ((a + ℓ)D)) for all ℓ ≥ k0 . e1 ± e2 ) ∪ Conv(0. OPn (ℓ)) = Sℓ .4. It follows that H 0 (P n . (a) As in Example 9. . Ex. The answer to part (c) is not a coincidence—given any polytopal complex P whose vertices are lattice points. let ∂∆2 be the boundary of the standard simplex ∆2 ⊆ R2 . p n−p 9.8.4. §3]. In this exercise ℓ will always denote an integer. . Here we compute the “Ehrhart polynomials” of some simple polytopal complexes. xn ] with the standard grading. Let p(x) be a polynomial such that limℓ→∞ p(ℓ)/ℓn exists and is nonzero. Prove that (ℓ · ∂∆2 ) ∩ Z2  = 3ℓ. IX (ℓ)) → · · · . Z). . OX (ℓ) stands for i∗ OX (ℓ). OX (ℓ)) = H 0 (X. . OX (ℓ)). .k (x/k) and pa. . m ≥ k0 . with constant term 0. with constant term −1. 9. Since a polytope P is contractible. (b) Consider the “butterﬂy” B = Conv(0.9) to prove that χ(ΩXP (ℓDP )) = (−1)n χ(ΩXP (−ℓDP )). (a) Prove that (ℓ∆n ) ∩ Zn  = ℓ+n for ℓ nonnegative. .4. . convert the 1 to a 0. −e1 ± e2 ) with boundary ∂B. After tensoring with OPn (ℓ). . (b) Prove that Int(ℓ∆n ) ∩ Zn  = ℓ−1 for ℓ positive. xn . In n of the positions. Hint: Lattice points in ℓ∆n give n monomials of degree ℓ in x0 .3. . (c) Prove that p(x) = (x + n)(x + n − 1) · · ·(x + 1)/n! satisﬁes p(ℓ) = (−1)n negative. . −ℓ−1 n for ℓ 9. .4. Begin with a list of 1’s of length ℓ + n. Also recall that as a sheaf on P n . Show that e(∂∆2 ) = 0 and e(∂B) = −1. OX (ℓ)) → H 1 (P n .6. we have an exact sequence 0 → IX (ℓ) → OPn (ℓ) → OX (ℓ) → 0 whose long exact sequence in cohomology begins 0 → H 0 (P n . IX (ℓ)) = I(X)ℓ . A projective variety X ⊆ P n gives 0 → IX → OPn → OX → 0. 9. 9. (a) Show that H 0 (P n .3.6 that H 0 (P n .4. (c) The Euler characteristic of a wellbehaved topological space Z is deﬁned to be e(Z) = p p p (−1) rankH (Z. .
where D0 is the divisor corresponding to u0 . Here you will study P = Conv(0. (a) Prove the Vandermonde identity.11. = n+ℓ−p ℓ . p)).13.4.3. 3.9. K j (ΣP . 2) is given by the fan in R2 with minimal generators u0 = −e1 − 2e2 .3. where K i (ΣP . Hint: σ ⊥ ∩ M p (d) Use Theorem 9. The Euler characteristic is not a polynomial in ℓ. e1 .4. a i b j simp 9. (b) Use Theorem 9. p) −→ K 1 (ΣP .9. n− j n−p fn− j . as claimed in Example 9. . ΩXP ). p) −→ 0. 1.3.1. Sec. (a) Given ℓ ∈ N.9. Let P◦ ⊆ NR be the dual polytope of P. p)) = has rank n − j for σ ∈ ΣP ( j).10.2) (ℓD0 )) = L(ℓP). 1. Also show that χ(OP(1.2. p Rescaling and translating P does not affect the face numbers fi . This is an example of a quasipolynomial.4. u1 = e1 . p (a) Show that χ(ΩXP ) = p j j j=0 (−1) χ(K (ΣP . (c) Use Proposition 2. ΩXP ) = hn−p . Sheaf Cohomology of Toric Varieties 9. The goal of this exercise is to prove the identity used in Example 9. Let P ⊆ MR ≃ Rn be a full dimensional simplicial lattice polytope and deﬁne n h simp = p i=p (−1)i−p i fn−i−1 . You will use this to compute dim H p (XP .4.4. When a polytope P has rational but not integral coordinates. 2) with D2P = 2D0 . p p p Let EhrP (x) be the unique polynomial satisfying EhrP (ℓ) = χ(ΩXP (ℓDP )) for all ℓ ∈ Z. 9. Use Exercise 2. which states that Hint: (1 + x)a (1 + x)b = (1 + x)a+b . Show that X2P = P(1. The reason is that D0 is not Cartier. For a full dimensional simple polytope P ⊆ MR ≃ Rn .4 and Proposition 2. Hint: Set j = n − i. b ∈ N. p) = p−i σ∈ΣP (i) (σ ⊥ ∩ M) ⊗Z OV (σ) and V (σ) ⊆ XP is the closure of the p orbit corresponding to σ ∈ ΣP . the counting function 1 L(ℓP) is almost a polynomial. (b) The weighted projective plane P(1. u2 = e2 .4.7] for a discussion of the Ehrhart quasipolynomial of a rational polytope.12. See [12.3 to show that χ(K j (ΣP . 2 e2 ) ⊆ R2 . p) −→ · · · −→ K p (ΣP . ◦ ◦ (c) h p simp = hn−p . 9.434 Chapter 9. 9. not necessarily simple.6 to prove the following: (a) P◦ is simple.19 gives p 0 −→ ΩXP −→ K 0 (ΣP . p)). Theorem 8. So we may that assume the origin is an interior point of P. P (b) The face numbers fiP of P and fiP of P◦ are related by fn−i−1 = fiP . Let P ⊆ MR ≃ Rn be a full dimensional lattice poltyope.2.4.3. p)) = dim H 0 (XP . (b) Use part (a) to show that n+1 i≥0 n−i ℓ−p−1 i−p i+ j=k = a+b k for a. K j (ΣP .2 to conclude that dim H p (XP . prove that L(ℓP) = 1 2 4ℓ +ℓ+1 3 1 2 4ℓ +ℓ+ 4 ℓ even ℓ odd.6 to show that dim H 0 (XP .
5. fℓ ) Given an Rmodule M. As with sheaf cohomology. . i p ) of elements of I satisfying i0 < · · · < i p . then n p EhrP (ℓ) = i=p i p L∗ (ℓQ). so that XP is not simplicial. This exercise will show how things can go wrong for a nonsimple polytope.. . This shows that some parts of Theorem 9. .6. . Hint: The DehnSommerville equations follow from Theorem 9.0. fℓ . . ℓ. (b) Conclude that for this polytope. deﬁne ˇ C p (f. there is a Cech complex for local cohomology. Let P = Conv(±e1 ± e2 .9) cannot hold for all p. Local Cohomology.i p−1 )∈[ℓ] p−1 M fi0 ··· fi p−1 .1 and Lemma 9. .. ˇ ˇ The Local Cech Complex. Let R be a ﬁnitely generated Calgebra. .2.4. Hint: Serre duality. . . One checks easily that M → ΓI (M) is left exact.4. (a) Show that h1 = h2 . let [ℓ] p denote the set of all (p + 1)tuples (i0 . . 0 n (b) Prove that EhrP (−ℓ) = (−1)n EhrP (ℓ). we get the derived functors M → HIp (M) such that • HI0 (M) = ΓI (M). • A short exact sequence 0 → M → N → P → 0 gives a long exact sequence 0 → HI0 (M) → HI0 (N) → HI0 (P) → HI1 (M) → HI1 (N) → HI1 (P) → · · · . Also set f = ( f1 .4.§9. Note that P is not simple. This is the Itorsion submodule of M. . I ⊆ R an ideal.5. Our treatment is based on [50].6. Hence.3.4. So the version of Serre duality stated in part (b) of Theorem 9. (9. . here we use generators of the ideal. 9. The sheaf case used an afﬁne open cover.4. Local Cohomology and the Total Coordinate Ring 435 (a) Prove that if ℓ > 0 in an integer. just as we did for global sections of sheaves in §9. We ﬁrst review the basics of local cohomology. q. Hint: Follow the proof of Theorem 9. In the next exercise you will see that other parts can fail. p (c) Prove that EhrP (0) = (−1) p h p .6 hold for arbitrary lattice polytopes. ℓ} be the index set and as in §9. e3 ) ⊆ R3 . . we use local cohomology and Ext to compute the cohomology of a coherent sheaf on a toric variety XΣ .. M) = (i0 .0.7. . . . and M an Rmodule. if I = f1 . Local Cohomology and the Total Coordinate Ring In this section.10 can fail for nonsimplicial toric varieties. let [ℓ] = {1.4. First deﬁne ΓI (M) = {a ∈ M  I k · a = 0 for some k ∈ N}. §9.14.. More precisely. . Q∈P(i) Hint: Look at the hypotheses of Theorem 9.
i p ).5. . ˇ Example 9.5. y−1 ] = SpanC (x−a y−b  a. . the local Cech complex computes local cohomology. . M) −→ · · · . M) −→ C1 (f.5. 7. the local Cech complex is 0 −→ S −→ Sx ⊕ Sy −→ Sxy −→ 0. Then the map Sxy → x−1 y−1 C[x−1 . ♦ .2. Theorem 9. Similar to Exercise 9. M) by p d p (α)(i0 . . y) · x−k y−k gives an exact sequence (9. See [97. we have d p ◦ d p−1 = 0. Given an Rmodule M and generators f = ( f1 .3. .436 Chapter 9. i p ) as an element of M fi0 ··· fi p via the map M fi 0 given by localization at fik . . M) is a function α that assigns an element of M fi0 ··· fi p−1 to each (i0 . Thm. j < b otherwise. the ˇ local Cech complex is d d d ˇ ˇ ˇ ˇ C • (f.1). An element of C p (f. Consider x−1 y−1 C[x−1 . y−1 ].1). .2 implies that HI2 (S) is the cokernel of Sx ⊕ Sy → Sxy . y)/(xy)k → f (x. . . . . M) −→ C p+1 (f. . . . ik .5.1. y]. y−1 ] −→ 0 (Exercise 9. ik . Theorem 9. .2. . fℓ ) of I. . fℓ ) of I. . i p ) = k=0 (−1)k α(i0 . It follows that if x and y have degree 1. Deﬁnition 9. . y ⊆ S = C[x. M) : 0 −→ C 0 (f. . .5.1) Sx ⊕ Sy −→ Sxy −→ x−1 y−1 C[x−1 . . y−1 ] deﬁned by f (x. . . .13] for a proof of the following. . Thus HI2 (S) ≃ x−1 y−1 C[x−1 .5. . Sheaf Cohomology of Toric Varieties ˇ where M fi0 ··· fi p−1 is the localization of M at fi0 · · · fi p−1 . M)) I for all p ≥ 0. b > 0) with Smodule structure given by x i y j · x−a y−b = x−(a−i) y−(b− j) 0 i < a. where we regard α(i0 . For I = x. Given an Rmodule M and generators f = ( f1 . .0. Then deﬁne a differential ˇ ˇ d p : C p (f. .5. . then HI2 (S) is graded with HI2 (S) = · · · ⊕ HI2 (S)−6 ⊕ HI2 (S)−4 ⊕ HI2 (S)−2 ⊕ 0 ⊕ · · · and dim HI2 (S)−ℓ = ℓ − 1 for ℓ > 0 (Exercise 9. there are natural isomorphisms ˇ H p (M) ≃ H p (C • (f. . . i p−1 ) ∈ [ℓ] p−1 . M) −→ C 2 (f. 0 1 2 ···c ··· fi p f ik → M fi0 ··· fi p ˇ ˇ Just as the Cech complex computes sheaf cohomology.
Rem. . This is no accident.y] and then A for k = 3 is given by A = coker matrix{{xˆ3. y ⊆ S = C[x.5. . .0 we introduced Ext in the context of sheaves. y]. M) consists of those elements of M annihilated by I k . we have the following result. It follows that HomR (R/I k . let A = S/I [k] = S/ x k .4. we obtain ΓI (M) = lim HomR (R/I k .5.6.5. In §9. We know by Example 9. M). Example 9. S)−5 for various values of k. an especially useful aspect of local cohomology is its relation to Ext. p Theorem 9. OP1 (−ℓ)) (see Example 9. M). and choosing any a ∈ M gives a homomorphism. Local Cohomology and the Total Coordinate Ring 437 The number ℓ− 1 just computed is also the dimension of H 1 (P1 . 7. There is also a module version. − → k We will need a variant of this result.1). . we set the ring S = Q[x. we begin with the observation that an Rmodule homomorphism φ : R/I k → M is determined by φ(1). let I [k] = f1k . M). y k .5. provided that I k · a = 0. Let I = x. Since A and S are graded Smodules. M) are denoted ExtR (N.8] for a proof). HIp (M) = lim ExtR (R/I k . For us. − → k Here is a simple example. Thm. By [97. fℓ . .5. since we will prove below that HI2 (S)a ≃ H 1 (P1 . S Let us compute the graded piece Ext2 (A. are easy to compute by computer algebra systems such as Macaulay 2. fℓk . They have the expected properties.yˆ3}} The dimension of Ext2 (A. . . the derived functors of p M → HomR (N. and more importantly. it follows without difﬁculty that the derived functors are related the same way (see [97. M).3 that HI2 (S)−5 has dimension 4. OP1 (a)) for all a ∈ Z. where for a ﬁxed Rmodule N. − → k Since direct limit is an exact functor (Exercise 9. Relation with Ext.5. y] by the command S = QQ[x.2). To compute this using Ext.5. S)−5 is computed by the command S . HIp (M) = lim ExtR (R/I [k] . . p Theorem 9.4.9]. In S Macaulay 2. S) is also graded. the Ext group Ext2 (A. 7. Comparing this to the deﬁnition of ΓI (M).§9. To see the relation between Ext and local cohomology. For an ideal I = f1 .
Our next task to is apply local cohomology to toric varieties.S)) which gives the answer 2. α∈Cl(XΣ ) . then F (α) will denote the sheaf associated to M(α). one can prove that (9. for more. This concludes our overview of local cohomology.Extˆ2(A. The sheaf associated to S(α) is denoted OXΣ (α).3). see [48. Explicit bounds on k in terms of a are needed in order to turn the method of Example 9. Sheaf Cohomology of Toric Varieties hilbertFunction(5. 4] or [97].4). introduced in Chapter 5. The problem is that ExtS (S/I [k] .5. S)−5 = HI2 (S)−5 for k sufﬁciently large. Our ﬁrst main result is that the local cohomology for the irrelevant ideal B(Σ) computes the sheaf cohomology of all twists a coherent sheaf on XΣ . Every coherent sheaf on XΣ arises in this way (Proposition 5. if F = M. ExtS (S/I [k] . then the answer stabilizes at the number 4.5. Theorem 9. More generally. App.3. So we cannot use the same k for all degrees a. We proved in Chapter 5 that a ﬁnitely generated graded Smodule M and a divisor class α ∈ Cl(XΣ ) give the following: • The coherent sheaf F = M on XΣ . where M(α)β = Mα+β for β ∈ Cl(XΣ ). We assume that XΣ has no torus factors. We also have the irrelevant ideal ˆ B(Σ) = x σ  σ ∈ Σmax .9). where a monomial ρ xρ ρ has degree [ ρ aρ Dρ ] ∈ Cl(XΣ ). ˆ xσ = ρ∈σ(1) xρ .6 into an algorithm. • The shifted module M(α).7 tells us that OXΣ (α) ≃ OXΣ (D) when D is a Weil divisor with divisor class α.3. S)a = HI2 (S)a 2 provided that k is sufﬁciently large.2) F (α) ≃ F ⊗XΣ OXΣ (α) (Exercise 9. When α is the class of a Cartier divisor.438 Chapter 9.5. If p ≥ 2. If we switch to k ≥ 4. we see that k = 3 is not big ♦ enough. and Proposition 5.7.5. The Toric Case. we have 2 ExtS (S/I [k] . Let M be a ﬁnitely generated graded Smodule with associated coherent sheaf F = M on XΣ . F (α)). The total coordinate ring S = C[xρ  ρ ∈ Σ(1)] of a toric variety a XΣ is graded by the class group Cl(XΣ ). 2 In this example. then p HB(Σ) (M) ≃ H p−1 (XΣ .5. Since dim HI2 (S)−5 = 4. We will prove below that for any degree a ∈ Z. S) is ﬁnitely generated over S but HI2 (S) is not (Exercise 9.
so that B(Σ) is generated by the monomials fi = ˆ ˇ ˇ x σi . α∈Cl(XΣ ) Proof. M)α to the Cech complex for F (α) given in (9. F (α)) ≃ C p−1 (U .. F (α)) by deleting ˇ ˇ complex C the ﬁrst term and shifting the remaining terms.. we have an exact sequence 0 0 −→ HB(Σ) (M) −→ M −→ 1 H 0 (XΣ . M)α is obtained from the Cech complex C • (U ..1). . . ˇ This isomorphism is compatible with the differentials. ˇ Comparing this with the Cech complex (9. . then x σγ gives / the same localization as (9. M) are direct sums of localizations M at products of various fi ’s. . ˇ To compute C p (f.5.. σℓ }.3) ˆ fi0 · · · fi p−1 = x σi0 · · · x σi p−1 ˆ = ρ∈σ(1) / xρρ . ˇ ˇ We will relate C p (f. F (α)).1) for the open cover U = {Uσi }i∈[ℓ] .1.2 implies that HB(Σ) (M) has a natural Cl(XΣ )grading such that for all α ∈ Cl(XΣ ).5. . For γ = (i0 . M)α . Then the terms of the local Cech complex C • (f. and with a little work (Exercise 9. F (α)). F (α)) −→ HB(Σ) (M)α −→ 0. M)α = γ∈[ℓ] p−1 ˇ Γ(Uσγ . M)α = (i0 .. . M)α ).i p−1 )∈[ℓ] p−1 M(α) fi0 ··· fi p−1 0 since shifting commutes with localization. i ∈ [ℓ] = {1. . ℓ}.1. i p−1 ) ∈ [ℓ] p−1 . . F (α)).5.3).3.i p−1 )∈[ℓ] p−1 M fi0 ··· fi p−1 α = (i0 . These localizations are Cl(XΣ )graded since M is graded and the fi are monomials. we obtain ˇ C p (f.5) we also get an exact sequence 0 1 0 −→ HB(Σ)(M)α −→ Mα −→ H 0 (XΣ .3 implies that M(α) fi0 ··· fi p−1 0 = M(α)x σγ ˆ 0 ≃ Γ(Uσγ . this implies p HB(Σ) (M)α ≃ H p−1 (XΣ . Proposition 5. which by Theorem 9. Local Cohomology and the Total Coordinate Ring 439 Furthermore. When p ≥ 2. . ﬁrst observe that ˇ C p (f. Since F (α) is the sheaf associated to M(α). a ˆ where aρ = {i j  ρ ∈ σi j } > 0. note that (9. If we set σγ = σi0 ∩ · · · ∩ σi p−1 .. we have p ˇ HB(Σ) (M)α = H p (C • (f. . The differentials also p preserve the grading. It follows that the local Cech ˇ• (f. F (α)) −→ HB(Σ) (M) −→ 0.5. .§9.. . Let Σmax = {σ1 . .5..
1.6. ˆ We begin with S/B(Σ). which is described as follows.1]. one replaces M with a free resolution and uses a spectral sequence argument. Earlier we described Ext using the derived functors of M → HomR (N. In particular. we have an obvious map F0 → S/B(Σ). F (α)) ≃ lim ExtS (S/B(Σ)[k] . This construction applies to any monomial ideal. M)α by the methods of Example 9. .440 Chapter 9.5. We can compute ExtS (S/B(Σ)[k] . the natural map S/B(Σ)[k+1] → S/B(Σ)[k] induces an isomorphism p p ExtS (S/B(Σ)[k] .1] for the details. Let (F• . For the ¸ˇ general case. M)α − → k p+1 when F = M and α ∈ Cl(XΣ ). See [50. Prop. Ext has a nice relation to local cohomology. An especially nice resolution is given by the Taylor resolution.5. Proof. We tackle this next. . 17. .7 imply that when p ≥ 1. for example. we have fIk = .5. d• ) denote the Taylor resolution of S/B(Σ)[k] . it works for [k] [k] B(Σ)[k] = f1k . take I. Sheaf Cohomology of Toric Varieties Theorems 9.5 and 9. Then deﬁne cIJ = 0 (−1)r fI / fJ ds (eI ) = J⊆I I = J ∪ {ir }. J ⊆ [ℓ] with I = J + 1 = s and list the elements of I as i1 < · · · < is . then there exists k0 ∈ N such that for all k ≥ k0 . Stabilization of Ext. Ext also comes from the derived functors of N → HomR (N. Let Fs be a free S module with basis eI for all I ⊆ [ℓ] with I = s. M)α ≃ HB(Σ) (M)α . To deﬁne the differential ds : Fs → Fs−1 . M)α ≃ ExtS (S/B(Σ)[k+1] . p+1 H p (XΣ . In the toric case. we can compute ExtS (S/B(Σ)[k] . We give a proof only for M = S following Mustata [132. Then Since F0 = S. M) for ﬁxed N.11]). Ex. A3. The problem is the direct limit.11]). where fI = lcm( fi  i ∈ I) and similarly for fJ . J cIJ eJ . [k] It has the same modules Fs = Fs . In particular.8. 4. This is the Taylor resolution (F• . Lemma 9. One can prove that ℓ 2 1 0 −→ Fℓ −→ · · · −→ F1 −→ F0 −→ S/B(Σ) −→ 0 d d d is exact (see [48. M)α . where one uses a projective resolution of N instead of an injective resolution of M p (see. k ≥ k0 implies p p ExtS (S/B(Σ)[k] . In particular. fℓk . [48. Let M be a ﬁnitely generated Cl(XΣ )graded Smodule and ﬁx α ∈ Cl(XΣ ).5. The minimal generators of B(Σ) are fi = x σi for i ∈ [ℓ]. Thm. d• ) of S/B(Σ). S) using any free resolution of S/B(Σ)[k] . If Σ is a complete fan. M) for ﬁxed M. and since the fi are squarefree. .
S) inherits a natural ZΣ(1) grading. This guarantees two things: p • The differential ds has degree 0. d• ) by setting deg(eI ) = k deg( fI ).4) is ZΣ(1) graded. S) has dual basis e∗ with deg(e∗ ) = −k deg( fI ). It follows that (Fs )∨ = HomS (Fs .5. Recall that this grading is induced by the map (9. We now compare the Taylor resolutions of S/B(Σ)[k] and S/B(Σ)[k+1] . the quotient ring S/B(Σ)[k] is ZΣ(1) graded. . . .5. b ∈ ZΣ(1) . s s I I [k+1] ∨ )a .4) H p (HomS (F• .§9. We claim that if a ≥ (−k. −k). It follows that (φ∨ )a is injective. S). Since B(Σ)[k] is a monomial ideal. . S)) → H p (HomS (F• . Then Exti (S/B(Σ)[k] . S)) [k] induced by the φs can be identiﬁed with the canonical map p p ExtS (S/B(Σ)[k] .6) (φ∨ )a : (Fs )∨ −→ (Fs s a [k] [k+1] ∨ )a [k] [k] is an isomorphism for all s. I I Given a. deﬁne a ≥ b if and only if aρ ≥ bρ for all ρ ∈ Σ(1). S) → ExtS (S/B(Σ)[k+1] . s−1 These maps are compatible with the surjection S/B(Σ)[k+1] → S/B(Σ)[k] . S Now grade the Taylor resolution (F• .5. The ring S also has a ZΣ(1) grading where deg(x a ) = a. so Now take x b e∗ ∈ (Fs I b = (k + 1)deg( fI ) + a ≥ (k + 1)deg( fI ) + (−k.5. The next key idea involves the use of a ﬁner grading than the Cl(XΣ )grading used so far. . so (9. so the isomorphism ExtS (S/B(Σ)[k] . The [k] [k+1] → Fs deﬁned by φs (eI ) = fI eI induces a commutative diagram maps φs : Fs Fs ds [k+1] φs [k] / Fs [k] ds [k] [k+1] Fs−1 [k+1] [k+1] φs−1 / F [k] . ﬁrst observe that φ∨ (e∗ ) = fI e∗ . Then b − (k + 1)deg( fI ) = a.5) a ZΣ(1) −→ Cl(XΣ ) where x a = ρ xρ ρ has degree [ ρ aρ Dρ ]. . −k). cIJ as above. S) ≃ [k] H p (HomS (F• .5. [k] [k] [k] • The map φs has degree 0. . . Hence (9. Thus the differentials ds in the Taylor resolution of S/B(Σ)[k] are given by [k] k ds (eI ) = J cIJ eJ . To prove this. S) is ZΣ(1) graded. then (9. . Local Cohomology and the Total Coordinate Ring 441 lcm( fik  i ∈ I).
5.5. . This follows easily from Theorem 9. Our ﬁnal task is give an explicit method for deciding when k is big enough.5. by Theorem 9. p+1 p+1 H p (XΣ .1. OXΣ (D))m = 0. The set I = {i ∈ [r]  m.5.6. The righthand side is ﬁnitedimensional since Σ is complete.8.442 Chapter 9. For a complete fan Σ. For these ﬁnitely many a’s.5. so that x b e∗ is in the image of (φ∨ )a . (k + 1)deg( fI ) is a vector with whose ρth entry is (k + 1) if xρ divides fI and 0 otherwise.8 imply that for p ≥ 1 and k ≫ 0. it should be possible to bound the size of the m’s that appear in terms of the ui and ai .5. m. pick k0 such that they all satisfy a ≥ (−k0 . The argument for p = 0.5). 1 is covered in Exercise 9. let b = ( m. s I p The local cohomology HB(Σ) (S) has a ZΣ(1) grading which is compatible its Cl(XΣ )grading via (9.5. .5 that the sign pattern of m (determined by m. .6) is an isomorphism. To relate this to Lemma 9. Since this region determined by the ui and ai . −k0 ). the inequalities m. . . .7 and Lemma 9. Suppose we have a divisor D = i ai Di on XΣ and m ∈ M such that H 1 (XΣ . though the following example suggests some of the ideas involved. which means that the region of the plane deﬁned by these inequalities is bounded. OXΣ (D))m = 0 is ﬁnitedimensional. ur + ar ). This k0 has the required properties.9. Since H 1 (XΣ . ui + ai < 0. You can see and example of this in Figure 4 from Example 9. . . S) Notice how these isomorphisms generalize Examples 9. ui + ai ≥ 0. Bounds. Let Σ be a complete fan in MR ≃ R2 with minimal generators u1 . i ∈ I m. .5. i ∈ [r] \ I have only ﬁnitely many integer solutions. ui + ai < 0} records the locations of the −’s in the sign pattern of m. the paper [50] gives an explict value for the number k0 appearing in Lemma 9.5. Sheaf Cohomology of Toric Varieties Since fI is squarefree. Then it is easy to see that 2 H 1 (XΣ . ui + ai ≥ 0 or < 0) has at least two strings of consecutive −’s. .1.7. Then the above inequality implies that fI divides x b . there are only ﬁnitely many m’s with the same sign pattern. and divisor class α ∈ Cl(XΣ ). Example 9. . . u1 + a1 . then p HB(Σ) (S)α ≃ H p−1 (XΣ . graded Smodule M. only ﬁnitely many can appear in HB(Σ) (S)α . ur arranged counterclockwise around the origin. p This implies that when we decompose HB(Σ) (S) into its nonzero ZΣ(1) graded p p pieces HB(Σ) (S)a .8. Theorem 9. .6.5. We will state the results of [50] without proof.5. It follows from Proposition 9. If α ∈ Cl(XΣ ) and p ≥ 2.7. OXΣ (D))m ≃ HB(Σ) (S)b . It follows that (9.5. OXΣ (α)) ≃ HB(Σ) (S)α ≃ ExtS (S/B(Σ)[k] .5 and the Taylor resolution. In other words. OXΣ (α)).3 and 9.
Then A is an r × n matrix. . we proceed as follows. .7) Q1 = max(entries of A) Qn−1 = max((n − 1) × (n − 1) minors of A) The following bound is proved in [50. en of M and let A be the matrix whose rows give the coefﬁcients of the uρ ’s with respect to the chosen basis. Prop. Then p p ExtS (S/B(Σ)[k] . S)α ≃ HB(Σ) (S)α for all k ≥ k0 . ρ aβ.5. Theorem 9. Theorem 9.5. Given a complete fan Σ in NR ≃ Rn and a divisor class α = [ ρ aρ Dρ ] ∈ Cl(XΣ ).β = 0.β =0 (aρ − aβ.ρ Dρ ] ∈ Cl(XΣ ). for each β with r j. .11. . Then [50. OXΣ (D))m = 0 is equivalent to bounding the 2 b’s with HB(Σ) (S)b = 0. 3. Write the resolution as · · · −→ F2 −→ F1 −→ F0 −→ M −→ 0 and write Fj = β∈Cl(XΣ ) S(−β)r j.10. −k0 ) for all such b’s. Pick a basis e1 .§9.5.5. . write β = [ 4. where k0 = n2 maxρ (aρ ) Q1 Qn−1 . ♦ In the general case. we know how to pick k0 so that b ≥ (−k0 . the formula for k0 involves the minimal free resolution of M. qn For a ﬁnitely generated graded Smodule M.5.8 shows that this k0 works. . where k0 = n2 maxρ. Let Σ and α be as in Theorem 9.β . Finally. For this matrix. And once we ﬁnd this bound. qn . M)α ≃ HB(Σ) (M)α for all k ≥ k0 .3]. . . Cor. The proof of Lemma 9.5. deﬁne qn = min(nonzero n × n minors of A) (9. we have p p ExtS (S/B(Σ)[k] . j.ρ ) Q1 Qn−1 . Local Cohomology and the Total Coordinate Ring 443 Thus bounding the m’s with H 1 (XΣ . and let M be a ﬁnitely generated graded Smodule.r j.1] gives the following bound.10. where r = Σ(1).
Sheaf Cohomology of Toric Varieties The Cotangent Bundle. .12. 0 0 ··· ··· . . . The ﬁrst step is to ﬁnd a graded Smodule M with Ω1 Σ ≃ M. uρ [ f ].10. .7) we constructed an exact sequence 1 0 −→ ΩS −→ M ⊗Z S → ρ S/ xρ where M ⊗Z S → ρ S/ xρ is deﬁned by m ⊗ f → ρ m. Then the basic exact sequence 0 −→ M −→ ZΣ(1) −→ Pic(XΣ ) −→ 0 can be written (9. Similar to the proof of Theorem 8. However. X In (8.5.1. We end this section by using our methods to calculate the cohomology of the cotangent bundle of a smooth complete toric variety XΣ . . A B where A is an r × n matrix whose ith row consists of the coefﬁcients of uρi in the basis of M. The (r − n) × r matrix B is called the Gale dual of A..5 we showed that Ω1 Σ is the sheaf of Ω1 .6. . We do this as follows. we have a commutative diagram 0 0 0 / / Ω1 S r i=1 S(−deg(xρi )) 0 / Z n ⊗Z S / A x 0 r i=1 S/ r i=1 S/ xρi xρi /0 /0 / Zr ⊗ ZS / B 0 / Zr−n ⊗ S Z / Zr−n ⊗ S Z /0 0 0 . ··· 0 0 .5. xρr induces a graded homomorphism θ : r S(−deg(xρi )) → S r−n of degree 0 such i=1 that Ω1 Σ is the sheaf associated to ker(θ). Let x be the r × r diagonal matrix of variables appearing in the statement of the lemma.444 Chapter 9. X Proof. . This is the same matrix A that appears in Theorem 9.8) 0 −→ Zn −→ Zr −→ Zr−n −→ 0. The (r − n) × r matrix xρ1 0 0 xρ 2 θ = B· . Lemma 9. . S Pick bases of M and Pic(XΣ ) and order the elements of Σ(1) as ρ1 .5. . we need a description X S of Ω1 that is easier to implement on a computer.1. ρr . and in Corollary 8. . .1. .
Ω1 2 (a)) ≃ ExtS (S/ x 28 .14. A= 1 0 1 0 1 0 0 −1 .5.8) are −1 2 0 1 . B = 1 1 1 . ♦ Here is a slightly more complicated example. as are the center and right columns.5. the second prints out the differentials.dd The ﬁrst command computes the resolution. For P 2 .3. This can be computed by the methods of Example 9. which can be computed by Macaulay 2: S F = res OM F. −→ S The numbers from (9.z] OM = ker matrix{{x. the dotted arrow from r S(−deg(xρi )) to Zr−n ⊗Z S is given by θ = B · x. the formula for k0 from Theorem 9.5. By the diagram chase from the proof of Theorem 8. y 28 . Local Cohomology and the Total Coordinate Ring 445 The rows are exact.1. This implies that for p ≥ 1. Ω1 2 (a)) directly from Ω1 (Exercise 9.z}} We also need a free resolution of Ω1 . Notice also that the symmetry of the table comes from Serre duality. so that for a ∈ Z. The results are S P shown in Table 3 on the next page. We can also compute H 0 (P 2 . the dotted arrows are exact.11 is k0 = 4 max(a − 2. ΩS )a P x @yA z To create the module Ω1 in Macaulay 2. the matrices of (9. i=1 Example 9.5.7). In Exercise 9. a − 3).6. We will use the notation of Example 9.5.§9.8) are given by 1 0 1 . D4 .7) are q2 = Q1 = 1. z 28 . For a in the range −4 ≤ a ≤ 4. Example 9. we use the commands S when −4 ≤ a ≤ 4. and by commutivity. The result is 0 1 0 −→ S(−3) − − S(−2)3 −→ Ω1 −→ 0.y.13.5. the matrices A and B of (9. p 1 H p (P 2 .6. Let us compute the ﬁrst cohomology of various twists of the cotangent sheaf of the Hirzebruch surface H2 .5.8 you will calculate this table by hand.5. A= 0 −1 −1 S = QQ[x.6.5.y. B = 1 −2 1 0 . we can use k0 = 28.5. Using the basis of Pic(H2 ) given by the classes of D3 .
where the degrees of the variables are given by the columns of B. b ≤ 2 H a\b –2 –1 0 1 2 –2 0 0 3 4 3 –1 0 0 2 2 2 0 1 1 2 1 1 1 2 2 2 0 0 2 3 4 3 0 0 many things we can see from this table.e.x4}} Be sure you understand how this uses the matrix θ from Lemma 9. 0).0}. Ω1 2 (a)) for −4 ≤ a ≤ 4 P a\p 0 −4 0 −3 0 −2 0 −1 0 0 0 1 0 2 3 3 8 4 15 1 2 0 15 0 8 0 3 0 0 1 0 0 0 0 0 0 0 0 0 The total coordinate ring is S = C[x1 . deg(x3 ) = (1.12.2*x2. so that for α = [aD3 + bD4 ] in Pic(H2 ). Then we create Ω1 within Macaulay 2 by the commands S S = QQ[x1. deg(x2 ) = (−2.5. 1). Sheaf Cohomology of Toric Varieties Table 3. −2)2 ⊕ S(−2. b. b. b)) for −2 ≤ a. There are H Table 4. we also compute the free resolution 0 −→ S(0. 0) −→ Ω1 −→ 0. b − 2). b − 2.7) are q2 = 1 and Q1 = 2. deg(x4 ) = (0. i.0}.x2. a.x4.446 Chapter 9. 1).{2. including: . Using Macaulay 2. deg(x1 ) = (1. We can use our methods to calculate the cohomology groups of Ω1 2 (a.1}}] OM = ker matrix{{x1.{1. the formula for k0 from Theorem 9. 0). Ω1 2 (a.11 is k0 = 16 max(a − 2. dim H p (P 2 . x2 ..5.x3. dim H 1 (H2 .{0.0}. a + 1. −2) −→ S(1. x3 .1}.x3.0. b) = H Ω1 2 (aD3 + bD4 ) for all a. x4 ].5.Degrees=>{{1.{0.x2. Table 4 records some computations of H 1 . S The numbers from (9.
5. 5] Exercises for §9. F (α)) −→ HB(Σ) (M)α −→ 0 for all α ∈ Cl(XΣ ). . Hint: Use Example 9. use the command S rank source basis({a.5.2. Prove that HB(Σ) (S) = HB(Σ) (S) = 0.b) rather than hilbertFunction({a. (c) Show that the ﬁrst column agrees with the Bott formula from Example 9.5.5.E) to compute dim E(a.9.7 by proving that there is an exact sequence 1 0 0 −→ HB(Σ) (M)α −→ Mα −→ H 0 (XΣ .5.5. and that for each i. 9.5. b) = (0. Suppose that (Ai . Ω1 2 (a.5. Ω1 2 (a.10 implies that dim H 1 (H2 .15 and 3.6.13 1 (a) Use the exact sequence from Lemma 9. φ′′ ) are directed systems.5.3. i i i i there exist di and δi commuting with the φ’s. • BottDanilovSteenbrink vanishing implies that H 1 (H2 .5.5. −b)). (b) Verify the ﬁrst column of Table 3. ΩH2 (−a. (A′ . Here are some details to check from Example 9. Hint: Look carefully at Proposition 5. b))∨ ≃ H 1 (H2 . i i ♦ d δ Prove the exactness of the sequence 0 −→ lim Ai −→ lim A′ −→ lim A′′ −→ 0.2) when α ∈ Cl(XΣ ) is the class of a Cartier divisor. You will verify the details of this example in Exercise 9. This H explains the symmetry in the table. 9. Complete the proof of Theorem 9.7. One suggestion is p that after computing an Ext group E = ExtS (A.5. An algorithmic approach to these calculations is described in [50. b > 0.1. i i − → − → − → i i i 9.3 and remember that the restriction of OXΣ (α) to Uσ is trivial. 9.3.3. This explains the 0s in the lower right corner of the table.12 to show that (Ω1 )α ≃ H 0 (XΣ .5.9.b}. 0).7. y ⊆ S = C[x. b)) = 0 for H a. such that we have an exact sequence i i 0 −→ Ai −→ A′ −→ A′′ −→ 0. 9.4. Ω1 2 ) = h1 = f1 − 2 f2 = 4 − 2 · 1 = 2 H since H2 comes from a quadrilateral (see Examples 2. Ω1 ).1.16). 1 • Serre duality implies that H 1 (H2 .E) since the former is faster than the latter and gives the same answer. φi ).5. Prove (9.5. Local Cohomology and the Total Coordinate Ring 447 • Theorem 9.5. prove that HI2 (S) is not ﬁnitely generated as an Smodule. y]. 9.4.3. (A′′ .b}. Prove the exactness of (9. φ′ ). Hint: Proposition 5. Sec.5. This explains the 2 when (a. For I = x.3 to show that HI2 (S)a = 0 for all a ≤ −2.§9. ΩXΣ (α)) S for all α ∈ Pic(XΣ ) when XΣ is smooth and complete. 1 0 9.4.1).
q −→ Erp+r.13. . In general. p. We refer the reader to [72] or [177] for a more complete treatment of spectral sequences. in which case we say that (Erp. At ﬁrst glance. and the second comes by sheaﬁfying the free resolution of Ω1 computed in Example 9. Ω1 2 (a)) = 1 P These formulas give the numbers in Table 3.e.q → Erp+r. (a) Use Macaulay 2 to compute the free resolution of Ω1 .5.q When working with an E1 spectral sequence. Fortunately. most of the ones encountered in this book will be fairly simple to understand.5.448 Chapter 9.q = 0 for all p.q .13 with a “barehanded” approach. Thus the nonvanishing terms lie in the quadrant where p. Ω1 2 (a)) = a2 − 1 if a ≥ 2 P dim H 1 (P 2 . drp. Observe that we have exact sequences involving Ω1 2 : P 1 3 0 −→ ΩP2 −→ OP2 (−1) −→ OP2 −→ 0 3 0 −→ OP2 (−3) −→ OP2 (−2) −→ Ω1 2 −→ 0. though our discussion is far from complete. P The ﬁrst comes from Lemma 9. Here we give the necessary background.q : Erp. q ≥ 0. We can check Table 3 of Example 9.12. P Appendix: Introduction to Spectral Sequences Spectral sequences are used several times in this chapter.q ) We will always work with ﬁrst quadrant spectral sequences. we often know the differentials d1 p.q−r+1 ◦ drp.A.9.q+r−1 : Erp−r.5.q Er+1 = ker(drp. A spectral sequence over C consists of vector spaces Erp.q (b) Er+1 is the cohomology of (Erp. Deﬁnition and Intuition.14. drp.11. Deﬁnition 9. This exercise concerns Example 9.q .q ).q−r+1 ) im(drp−r. if a = 1 dim H 2 (P 2 .5. spectral sequences may seem rather complicated.5.5. Sheaf Cohomology of Toric Varieties 9.q explicitly.q−r+1 such that (a) drp+r. Twist these sequences by a ∈ Z and consider the S resulting long exact sequences in cohomology. ΩP2 (a)) are: dim H 0 (P 2 . however. Ω1 2 (a)) = a2 − 1 if a ≤ −2.q = 0 when p < 0 or q < 0. 1 conclude that the nonzero values for H p (P 2 . i. (c) Compute the entries in Table 4. r.q and linear maps drp. q.q+r−1 → Erp.1. Using the vanishing theorems we know.. S (b) Show how the formula given for k0 follows from Theorem 9.q ) is an E1 or E2 spectral sequence respectively. p. the differentials dr for r ≥ 2 are much more difﬁcult to describe. which means that Erp.8.5. The minimum value of r is usually 1 or 2. p.q : Erp. 9.
q One can think of E2 as a ﬁrst approximation to the convergent. p.q Proposition 9.q ≃ E∞ . The map E2 → H p is called an edge homomorphism.3. p.0 vanish. Note that degeneration at Er implies Erp. p.q . as better and better approximations. This common value is deﬁned to be E∞ .1 p−2.A. as claimed.q vanish for r ≥ 2 since E2 = 0 for q > 0.q p.A. The intuition behind a convergent E2 spectral sequence is that we often know something about the E2 terms and are interested in what the spectral sequence converges to.q Erp.0 p. Deﬁnition 9. and then E3 . q.q E∞ ≃ F p H p+q /F p+1 H p+q For an E1 or E2 spectral sequence.q Proof. In general. Edge Homomorphisms. A nice introduction to this way of thinking can be found in the ﬁrst chapter of [123].0 E2 ≃ F p H p /F p+1 H p = F p H p p−1. the differentials mapping to and from Erp. Then p. q.0 E2 −→ E3 −→ · · · −→ E∞ ≃ F p H p /F p+1 H p = F p H p ⊆ H p . q. Here is an example of how to work with a spectral sequence.2. It follows that p.3. . Then we have p. .3 implies that the differentials d2 vanish for all p. k ≥ 0. Another feature of a convergent E2 spectral sequence is that all differentials starting from Erp. It p. .q p. drp.q E1 ⇒ H p+q or p.q p. p. we say that a spectral sequence degenerates at Er if the differentials drp.q for r ≫ 0.0 Deﬁnition 9. so that the spectral sequence gives maps p.0 E2 ≃ H p for all p ≥ 0. . E4 .q p. p. Suppose E2 ⇒ H p+q is an E2 spectral sequence with the property p. q.2 0.A. This leads to a more precise version of Proposition 9.q vanish when r is large by our ﬁrst quadrant assumption.0 p.4.q follows that E2 = E∞ for all p. p.A. First observe that all differentials drp.q p. we write this as p.q E2 ⇒ H p+q respectively. A spectral sequence (Erp.0 so that F p H p = F p−1 H p = · · · = F 1 H p = F 0 H p = H p .Appendix: Introduction to Spectral Sequences 449 For ﬁxed p.q that E2 = 0 for all q > 0.A.q The hypothesis of Proposition 9.q vanish for all p. Thus E2 ≃ H p .q ) converges to vector spaces H k . if there are subspaces 0 = F k+1 H k ⊆ F k H k ⊆ F k−1 H k ⊆ · · · ⊆ F 1 H k ⊆ F 0 H k = H k such that p.p and 0 = E2 = E2 = · · · = E2 implies 0 = F p−1 H p /F p H p = F p−2 H p /F p−1 H p = · · · = F 0 H p /F 1 H p .q = Er+1 = Er+2 = · · · p.
q Proposition 9.0.5.0 E2 −→ H p is an isomorphism for all p ≥ 0. Suppose E2 ⇒ H p+q is an E2 spectral sequence with the property p. This is true in the two applications of Proposition 9.A. p.A. we often know the edge homomorphisms E2 → H p explicitly.0 For many E2 spectral sequences.450 Chapter 9.5 given in §9.q that E2 = 0 for all q > 0. Then the edge homomorphism p. Sheaf Cohomology of Toric Varieties p. .
Chapter 10
Toric Surfaces
§10.1. Singularities of Toric Surfaces and Their Resolutions
In this chapter, we will apply the theory developed so far to give a detailed treatment of the structure of 2dimensional normal toric varieties (toric surfaces). Singular Points of Toric Surfaces. If XΣ is the toric surface of a fan Σ in NR ≃ R2 , then minimal generators of the rays ρ ∈ Σ(1) are primitive and hence extend to a basis of N. Then Theorem 3.1.19 implies that the toric surface obtained by removing the ﬁxed points of the torus action (i.e., the points corresponding to the 2dimensional cones under the OrbitCone Correspondence) is smooth. There are only ﬁnitely many such points, so XΣ has at most ﬁnitely many singular points. Moreover, 2dimensional cones are always simplicial, so from Example 1.3.20, each of these singular points is a ﬁnite abelian quotient singularity (isomorphic to the image of the origin in the quotient C2 /G where G is a ﬁnite abelian group). All cones are assumed to be rational and polyhedral. A 2dimensional strongly convex cone in NR ≃ R2 has the following normal form that will facilitate our study of the singularities of toric surfaces. Proposition 10.1.1. Let σ ⊆ NR ≃ R2 be a 2dimensional strongly convex cone. Then there exists a basis e1 , e2 for N such that σ = Cone(e2 , de1 − ke2 ), where d > 0, 0 ≤ k < d, and gcd(d, k) = 1. Proof. We will need the following modiﬁed division algorithm here and at several other points in this chapter (Exercise 10.1.1): (10.1.1) Given integers l and d > 0, there are unique integers s and k such that l = sd − k and 0 ≤ k < d. 453
454
Chapter 10. Toric Surfaces
Say σ = Cone(u1 , u2 ), where ui are primitive vectors. Since u1 is primitive, we can take it as part of a basis of N, and we let e2 = u1 . Since σ is strongly convex, for any basis e′ , e2 for N, it will be true that 1 u2 = de′ + le2 1 for some d = 0. By replacing e′ by −e′ if necessary, we can assume d > 0. By 1 1 (10.1.1), there are integers s, k such that l = sd − k, where 0 ≤ k < d. Using this integer s, let e1 = e′ + se2 . Then e1 , e2 is also a basis for N and 1 u2 = de1 + (l − sd)e2 = de1 − ke2 . Hence σ = Cone(e2 , de1 − ke2 ) as claimed, and gcd(d, k) = 1 follows since u2 is primitive. We will call the integers d, k in this statement the parameters of the cone σ, and {e1 , e2 } is called a normalized basis for N relative to σ. The uniqueness of d, k will be studied in Proposition 10.1.3 below. Using the normal form, we next describe the local structure of the point pσ in the afﬁne toric variety Uσ . Recall from Example 1.3.20 that if N ′ ⊆ N is the sublattice generated by the ray generators of σ, then Uσ ≃ C2 /G, where G = N/N ′ . In our situation, N = Ze1 ⊕ Ze2 , and N ′ = Ze2 ⊕ Z(de1 − ke2 ) = dZe1 ⊕ Ze2 , so it follows easily that (10.1.2) G = N/N ′ ≃ Z/dZ.
In particular, for singularities of toric surfaces, the ﬁnite group G is always cyclic. The action of G on C2 is determined by the integers d, k as follows. We write µd = {ζ ∈ C  ζ d = 1} for the group of dth roots of unity in C. Then a choice of a primitive dth root of unity deﬁnes an isomorphism of groups µd ≃ Z/dZ. Proposition 10.1.2. Let M ′ be the dual lattice of N ′ and let m1 , m2 ∈ M ′ be dual to u1 , u2 in N ′ . Using the coordinates x = χ m1 and y = χ m2 of C2 , the action of ζ ∈ µd ≃ N/N ′ on C2 is given by ζ · (x, y) = (ζ x, ζ k y). Furthermore, Uσ ≃ C2 /µd with respect to this action. Proof. The general discussion in §1.3 shows that the quotient N/N ′ ≃ Z/dZ acts on the coordinate ring of C2 via (10.1.3) (u + N ′ ) · χ m = e2πi m ,u χ m ,
′ ′ ′
where m′ ∈ σ ∨ ∩ M ′ and u = je1 for 0 ≤ j ≤ d − 1.
§10.1. Singularities of Toric Surfaces and Their Resolutions
455
An easy calculation shows that m1 , e1 = 1/d and m2 , e1 = k/d. Hence if we set up the isomorphism µd ≃ N/N ′ by mapping e2πi j/d → je1 + N ′ , then for all ζ = e2πi j/d ∈ µd , we have ζ · (x, y) = (e2πi j/d x, e2πi jk/d y) = (ζ x, ζ k y) by (10.1.3). This what we wanted to show. We next describe the slight but manageable ambiguity in the normal form for 2dimensional cones. Two cones are lattice equivalent if there is a bijective Zlinear mapping ϕ : N → N taking one cone to the other. After choice of basis for N, such mappings are deﬁned by matrices in GL(2, Z). Proposition 10.1.3. Let σ = Cone(e2 , de1 − ke2 ) and σ = Cone(e′ , de′ − ke′ ) be 2 1 2 cones in normal form that are lattice equivalent. Then d = d and either k = k or kk ≡ 1 mod d. Proof. Since the cones are lattice equivalent, writing N ′ and N ′ for the sublattices as in (10.1.2), there is a bijective Zlinear mapping ϕ : N → N such that ϕ(N ′ ) = N ′ . Hence N/N ′ ≃ N/N ′ , so d = d. The statement about k and k is left to the reader in Exercise 10.1.2. Here are two examples to illustrate Proposition 10.1.2. Example 10.1.4. First consider a cone σ = Cone(e2 , de1 − e2 ) with parameters d > 1 (so the cone is not smooth) and k = 1. This is precisely the cone considered in Example 1.2.21. The corresponding toric surface Uσ is the rational normal cone Cd ⊆ Cd+1 . The quotient Cd ≃ C2 /µd was studied in the special case d = 2 in Example 1.3.19, and the general case was described in Exercise 1.3.11. With the notation of Proposition 10.1.2, ζ ∈ µd acts on (x, y) ∈ C2 via ζ · (x, y) = (ζx, ζy) and the ring of invariants is C[x, y]µd = C[x d , x d−1 y, . . . , xy d−1 , y d ], so Uσ ≃ C2 /µd ≃ Spec(C[x d , x d−1 y, . . . , xy d−1 , y d ]). On the other hand, from Example 1.2.21, we also have the description Uσ ≃ Spec(C[s, st, st 2 , . . . , st d ]). Exercise 10.1.3 studies the relation between these representations of the coordinate ring of Uσ . ♦
456 Chapter 10. E6 . so by the comment after (10.13 of Chapter 8. Moreover we have an isomorphism of rings ϕ : C[X . There is a complete classiﬁcation of such points in terms of the Dynkin diagrams of types Ak . E7 .2). y k+1 .1. Unlike the previous example.2. C).1. xy]. y k+1 . More details on these singularities can be found in [156.4. Dk .6. the afﬁne toric surface Uσ is Gorenstein if and only if k = d − 1. so d = k + 1.5. We will express everything in terms of the parameter k in the following. y]µk+1 = C[x k+1 .2.1. It is easy to check that the ring of invariants here is C[x. and E8 . such points do not appear on toric surfaces. For a cone σ = Cone(e2 . The groups corresponding to the diagrams Dk .5.4.5 and is called a rational double point (or Du Val singularity) of type Ak .4. . y) = (ζ x. which will be introduced shortly. Z]/ Z k+1 − XY ≃ C[x k+1 . the action of G = N/N ′ on C2 is given by ζ · (x. Ch. E8 are not abelian.1. Another standard form of these singularities is given in Exercise 10. The rational double points are the simplest singularities from a certain point of view.e. Hence by Proposition 10. The following result was proved in Exercise 8. The exact deﬁnition. which we will give in §10. All rational double points appear as singularities of quotient surfaces C2 /G where G is a ﬁnite subgroup of SU(2. xy] X → x k+1 Y → y k+1 Z → xy. so we may identify the toric surface Uσ with the variety V(Z k+1 − XY ) ⊆ C3 . depends on the notion of a resolution of singularities. Toric Surfaces Example 10. We will see one way that the Dynkin diagram Ak appears from the geometry of the toric surface Uσ in Exercise 10.1.1.1.Y . and we will return to this example in §10. ζ −1 y). E7 .2.5. de1 − ke2 ) in normal form. Proposition 10. Note that k ≡ −1 mod d..1. ♦ The origin is unique singular point of the afﬁne variety of Example 10. Next consider a cone σ with parameters d and k = d − 1. the multiplicity of the singularity is two). this is a case we have not encountered previously.14). E6 . Recall that a normal variety X is Gorenstein if its canonical divisor is Cartier (Deﬁnition 8. Here is another interesting aspect of Example 10. VI] and in the article [47]. They are called double points because the lowest degree nonzero term in the deﬁning equation has degree two (i.
de1 − e2 ) studied in Example 10. The identity mapping on the lattice N is compatible with the fans Σ and σ as in Deﬁnition 3.4.1.1. σ σ1 ←τ σ2 Figure 1.1.7 since Σ is a reﬁnement of σ. Consider the rational normal cone of degree d. de1 − e2 ). we claim that φ satisﬁes (10. Hence Theorem 3.3. One of the most appealing aspects of toric varieties is the way that many questions that are difﬁcult for general varieties admit simple and concrete solutions in the toric case. the afﬁne toric surface Uσ for σ = Cone(e2 .4). In addition. e1 ) σ2 = Cone(e1 . Deﬁnition 10. Note that both σ1 and σ2 (as well as all of their faces) are smooth cones.5) φ : XΣ −→ Uσ .4) Y \ ϕ−1 (Xsing ) ≃ X \ Xsing . By Theorem 3.1.19 implies that XΣ is a smooth surface.1.3. Singularities of Toric Surfaces and Their Resolutions 457 Toric Resolution of Singularities.5. The problem of ﬁnding resolutions of singularities is a perfect example. The cone σ and the reﬁnement given by σ1 . Let Σ be the fan in Figure 1 obtained by inserting a new ray τ = Cone(e1 ) subdividing σ into two 2dimensional cones: σ1 = Cone(e2 .4 and 10. Let X be a normal toric surface.§10. Finally. A proper morphism ϕ : Y → X is a resolution of singularities of X if Y is a smooth surface and ϕ induces an isomorphism of varieties (10. We illustrate this by constructing explicit resolutions of singularities of the toric surfaces from Examples 10. we have a corresponding toric blowup morphism (10.1.1.1.1.4. Example 10. σ2 . the toric morphism φ is proper by Theorem 3.7.1.1.4. and denote by Xsing the ﬁnite set of singular points of X (possibly empty).8. This follows from the OrbitCone Correspondence on the . Such a mapping modiﬁes X to produce a smooth variety without changing the smooth locus X \ Xsing . τ We now use some results from Chapter 3.
1. For u1 = e2 and u2 = e1 − e2 . The fan produced by this subdivision is somewhat easier to visualize if we draw the cones relative to a different basis u1 .5. We consider the case d = 4 of Example 10. σ4 σ3 σ2 σ σ1 Figure 2. u1 + 2u2 ) σ3 = Cone(u1 + 2u2 .1.1.1. u1 + 3u2 ) σ4 = Cone(u1 + 3u2 . Toric Surfaces two surfaces: if pσ is the distinguished point corresponding to the 2dimensional cone σ in (the singular point of Uσ at the origin). It follows that XΣ and the morphism (10. ♦ Example 10. u1 + 4u2 ) appear in Figure 2. to the reader (Exercise 10. The cone σ and the reﬁnement Σ . 4e1 − 3e2 ) yielding collections of smooth cones. That is. ρ2 = Cone(2e1 − e2 ).6) σ2 = Cone(u1 + u2 .5) give a toric resolution of singularities of the rational normal cone. We will leave the details. The inverse image E = φ−1 (pσ ) is the curve on XΣ given by the closure of the TN orbit O(τ ) corresponding to the ray τ . It is easy to ﬁnd subdivisions of σ = Cone(e2 . u1 + u2 ) (10. the singular point “blows up” to E ≃ P1 on the smooth surface.5). as well as the generalization to all d ≥ 2. We call E the exceptional divisor on the smooth surface. then φ restricts to an isomorphism XΣ \ φ−1 (pσ ) ≃ Uσ \ {pσ } = (Uσ )smooth . We will say more about how E sits inside the surface XΣ in §10. for which the surface Uσ has a rational double point of type A3 . The most economical way to do this is to insert three new rays ρ1 = Cone(e1 ). the cone σ = Cone(u1 .458 Chapter 10.4.1. u1 + 4u2 ) and the fan Σ with maximal cones σ1 = Cone(u1 . ρ3 = Cone(3e1 − 2e2 ) to obtain a fan Σ consisting of four 2dimensional cones and their faces.9. u2 for N.
1. Singularities of Toric Surfaces and Their Resolutions 459 You will check that each of these cones is smooth. de1 − ke2 ) with parameters d > 0. Let XΣ be a normal toric surface. e1 ) σi′′ = Cone(e1 . de1 − ke2 ) . The curves V (τi ) are isomorphic to P1 . we assume that the existence of smooth reﬁnements has been established for all fans Σ with s(Σ) < s. the exceptional divisor E = φ−1 (pσ ) is the union E = V (τ1 ) ∪V (τ2 ) ∪V (τ3 ) on XΣ . The same techniques can be applied to any normal toric surface XΣ . In this case. hence a resolution of singularities of XΣ . Let σ1 . and consider a fan Σ with s(Σ) = s. k) = 1. Then we deﬁne ℓ s(Σ) = i=1 (mult(σi ) − 1) . . Hence XΣ is already smooth and we take this as the base case for our induction.10. The ﬁrst two intersect transversely at the ﬁxed point of the TN action on XΣ corresponding to the cone σ2 . . It is easy to see that this implies that Σ is a smooth fan. As in the previous example. The reasoning given in Example 10. Consider the reﬁnement Σ′ of Σ obtained by subdividing the cone σi into two new cones σi′ = Cone(e2 . we constructed toric resolutions of afﬁne toric surfaces with just one singular point. and gcd(d. ♦ In these examples.8 applies to show that the corresponding toric morphism φ is proper and birational.1. Since Σ is a reﬁnement of σ. then ℓ = 0 or mult(σi ) = 1 for all i. then there exists some nonsmooth cone σi in Σ. while the second two intersect transversely at the ﬁxed point corresponding to σ3 . If s(Σ) = 0.§10. We will prove this by induction on an integer invariant of fans that measures the complexity of the singularities on the corresponding surfaces. Proof. There exists a smooth fan Σ′ reﬁning Σ such that the associated toric morphism φ : XΣ′ → XΣ is a toric resolution of singularities. φ restricts to an isomorphism from XΣ \ φ−1 (pσ ) to XΣ \ {pσ }. By Proposition 10. For each i.1. For the induction step. there is a basis e1 .1. If s ≥ 1. 0 ≤ k < d. e2 for N such that σi = Cone(e2 .3. . σℓ denote the 2dimensional cones in a fan Σ.1. where mult(σi ) = [N : Ni ] as in §6. we have a proper toric morphism φ : XΣ → Uσ . . It sufﬁces to show the existence of the smooth fan Σ′ reﬁning Σ. Hence XΣ is a smooth surface. Theorem 10. we will write Ni for the sublattice of N generated by the ray generators of σi .
[N : Ni′′ ] = k < [N : Ni ] = d. then applying the results of this section.1.5.1. Toric Surfaces with a new 1dimensional cone ρ = Cone(e1 ). We must show that s(Σ′ ) < s(Σ) to invoke the induction hypothesis and conclude the proof. However.1. Adapt the usual proof of the integer division algorithm to prove (10. Hence the cone σi′′ has parameters k and l obtained from (10. and the proof is complete by induction. It follows that s(Σ′ ) < s(Σ). then the parameter k of σ satisﬁes either k = k. In this exercise. Since k < d. Now consider the cone σi′′ . k) = 1.1. k such that d d + kk = 1.1). . so there are integers d.9.2. if Ni′′ is the sublattice generated by the ray generators of σi′′ . Since A ∈ GL(2. (a) Show that if σ is obtained from a cone σ by parameters d.2).1. Exercises for §10.460 Chapter 10.3.1. The cone σi′ is smooth since e1 . the reﬁnement of the given cone σ was produced by subdividing along the rays through the Hilbert basis (the irreducible elements) of the semigroup σ ∩ N. we have gcd(k. Hint: There is a choice of orientation to be made in the normalization process. In terms of the basis e1 . In s(Σ). and hence σi′′ will have the same parameters as Cone(e2 .1 and prove part of Proposition 10. k in the normal form for cones from Proposition 10. l) = 1 as well.1) to write (10. ke1 + de2 ).1. 10.7) d = sk − l where 0 ≤ l < k. then by (10. As a preview. k) = 1. you will develop further properties of the parameters d. the terms corresponding to the other cones σ j for j = i are unchanged. notice that in Examples 10.7).1. it deﬁnes an automorphism of N.1. A resolution of a nonnormal toric surface singularity can be constructed by ﬁrst saturating the associated semigroup as in Theorem 1. Toric resolutions of singularities for toric varieties of dimension three and larger also exist. ke1 + de2 ).1. We will see in the next section that in the afﬁne case. Since gcd(d. the reﬁnement that gives the resolution of singularities of Uσ has a very nice description. we must determine the parameters of σi′′ .1. or kk ≡ 1 mod d. But now we apply (10. k by a Zlinear mapping of N deﬁned by a matrix in GL(2. Recall that gcd(d. e2 is the normalized basis of N relative to σi . 10. Z). e2 for N. So it contributes a zero term in s(Σ′ ).8 and 10. In order to compute its contribution to s(Σ′ ).3. Z).1. the Zlinear mapping deﬁned by the matrix A= 0 −1 1 0 (a “90degree rotation”) takes σi′′ to Cone(e2 . we postpone the higherdimensional case until Chapter 11.1.
5. 10. Continued Fractions and Toric Surfaces To relate continued fractions to toric surfaces.1. We will always assume d > k > 0. xy d−1 .9 and show how to extend the results there to the case σ = Cone(e2 . hence the singularities at the origin are analytically equivalent. Another commonly used normal form for this type of singularity is the singular point at (0. namely the singular point at (0. HirzebruchJung Continued Fractions. then change bases in M to normalize σ ∨ . .1. st. In this exercise. In Example 10.1. 0) on the surface W = V(X k+1 +Y 2 + Z 2 ). Show that V and W are isomorphic as afﬁne varieties. 10. show how to insert new rays ρi to subdivide σ and obtain a fan Σ whose associated toric surface is smooth. (a) Check that each of the four cones in (10.10. st 2 .1.2. Try to do this with as few new rays as possible.1. write down σ ∨ in the corresponding dual basis in M. we begin with the afﬁne toric surface Uσ of a cone σ ⊆ NR ≃ R2 in normal form with parameters d. §10.§10. The ﬁrst is smooth. Hint: Use the normal form for σ. so that Uσ has a unique singular point. we gave one form of the rational double point of type Ak . 0. show that C[s.19.4.2.5. . de1 − ke2 ) to a fan containing the 2dimensional cones σ ′ = Cone(e2 . 0. st d ] ≃ C[x d . (b) For general d.1. Hence we obtain toric resolutions of singularities φ : XΣ → Uσ for all d. Do you notice a relation between this graph and the Dynkin diagram Ak = Ad−1 mentioned before? We will discuss the relation in detail in §10.1. so that the toric surface XΣ is smooth by Theorem 3.10 that the cone σ ′′ has parameters k. y d ] under s → x d and t → y/x. . then the dual cone σ ∨ ⊆ MR has parameters d. but the second may not be. k.3. we saw in the proof of Theorem 10.1. 10.2 to explain where these identiﬁcations come from in terms of the semigroup Sσ .1. e2 is the normalized basis for N and m1 . Hint: We have s = χ m1 and t = χ m2 where e1 .4. de1 − (d − 1)e2 ) for general d. .1.6) is smooth.4.1. 0) on the surface V = V(Z k+1 − XY ) ⊆ C3 . de1 − ke2 ). x d−1 y. . However. (c) Identify the inverse image C = φ−1 (pσ ) in general. . For instance. how many irreducible components does C have? How are they connected? Hint: One way to represent the structure is to draw a graph with vertices corresponding to the components and connect two vertices by an edge if and only if the components intersect on XΣ . the ﬁrst step is to reﬁne the cone σ = Cone(e2 . and use Proposition 10. Hint: There is a linear change of coordinates in C3 that does this. k1 satsifying d = b1 k − k1 . . k. With the notation in Example 10. . you will check the claims made in Example 10. . m2 is the dual basis for M. When we construct a resolution of singularities of Uσ by following the proof of Theorem 10. d − k. Continued Fractions and Toric Surfaces 461 (b) Show that if σ is a cone with parameters d. e1 ) and σ ′′ = Cone(e1 .
since as in the usual Euclidean algorithm. We used slightly different notation before. where k = b2 k1 − k2 using (10. we have bi ≥ 2 for all i. kr−3 = br−1 kr−2 − kr−1 kr−2 = br kr−1 that computes the parameters of the new cones produced as we successively subdivide to produce the fan giving the resolution of singularities. Also. the ki are a strictly decreasing sequence of nonnegative numbers.2. . we insert a new ray and obtain a new smooth cone and a second. . it is desirable to have a more compact way to represent these expressions. possibly nonsmooth cone with parameters k1 .1. Toric Surfaces where b1 ≥ 2. the new notation will help us keep track of what happens as we continue the process and reﬁne the cone σ ′′ . . by (10. Doing this repeatedly yields a modiﬁed Euclidean algorithm d = b1 k − k1 k = b2 k1 − k2 . . Using the normalized basis for N relative to σ ′′ .1.1. . .1) (10.2. 0 ≤ k1 < k as in (10. . . The equations (10. (10.2. b2 .2.3) d/k = b1 − b2 − 1 1 ··· − 1 br . br ]].1) can be rearranged: d/k = b1 − k1 /k k/k1 = b2 − k2 /k1 . writing s rather than b1 and l rather than k1 . We will use the notation d/k = [[b1 . kr−3 /kr−2 = br−1 − kr−1 /kr−2 kr−2 /kr−1 = br and spliced together to give a type of continued fraction expansion for the rational number d/k.1).462 Chapter 10. k2 . .1). The process terminates with kr = 0 for some r as shown.2) This is the HirzebruchJung continued fraction expansion of d/k. with minus signs: (10. For obvious typographical reasons.1).
Set (10.2. bs . The HirzebruchJung continued fraction expansion is 17/11 = [[2. bt+1 ]] = [[b1 . . . (b) [[b1 . P1 = b1 . are the convergents. . Example 10.3) makes sense when the b j are any rational numbers (not just integers) such that all denominators in (10. . 3. Continued Fractions and Toric Surfaces 463 The integers bi are the partial quotients of the HirzebruchJung continued fraction. br ]]. . We will show that the sequences deﬁned by (10. Qi satisfy: (a) The Pi and Qi are increasing sequences of integers.3) are nonzero. (d) The convergents form a strictly decreasing sequence: d Pr Pr−1 P1 = < < ··· < .2.4). . . Consider the rational number 17/11. . .§10.1). The proof of part (a) is left to the reader (Exercise 10.2. P When s = 1. .2. bi ]]. b2 . and the truncated HirzebruchJung continued fractions [[b1 . Q0 = 0 Q1 = 1. 2.2.4) and for all 2 ≤ i ≤ r. Pi = bi Pi−1 − Pi−2 Qi = bi Qi−1 − Qi−2 . as may be veriﬁed directly using the modiﬁed Euclidean algorithm (10. k Qr Qr−1 Q1 Proof. .1).2. . Let d > k > 0 be integers with gcd(d.2. . 2]]. P0 = 1. .2. . .5) satisfy Ps [[b1 . bs ]] = Qs for all such lists b1 . 2. ﬁrst observe that the expression on the right side of (10. To prove part (b). . . . Proposition 10. . . we have [[b1 ]] = b1 = Q11 by (10. .2. . .2. ♦ 1 ≤ i ≤ r. 2. (c) Pi−1 Qi − PiQi−1 = 1 for all 1 ≤ i ≤ r. . bt − 1 bt+1 ]]. let (10. .5) Then the Pi . k) = 1 and let d/k = [[b1 . The proof is by induction on the length s of the list. Now assume that the result has been proved for all lists of of length t and consider the expression [[b1 .1. . . . bi ]] = Pi /Qi for all 1 ≤ i ≤ r. Deﬁne sequences Pi and Qi recursively as follows.2.2.
bt+1 Qt − Qt−1 Qt+1 Part (c) will be proved by induction on i. Let u0 = e2 and use the integers Pi and Qi from Proposition 10. have the following properties: (a) Each σi is a smooth cone and ui−1 . Now assume that the result has been proved for i ≤ s. Qi−1 Qi Qi−1 Qi Hence Pi Pi−1 < Qi Qi−1 since Qi−1 Qi > 0 by part (a).2. (b) For each i. and consider i = s + 1. 1 ≤ i ≤ r + 1.2 to construct vectors ui = Pi−1 e1 − Qi−1 e2 . Hence part (d) follows. By the recurrences (10.5). 1 ≤ i ≤ r + 1. Theorem 10. this equals 1 Pt − bt+1 Pt−1 1 Qt − bt+1 Qt−1 which is what we wanted to show.5). ui are its ray generators.2. = Pt+1 bt+1 Pt − Pt−1 = . Let σ = Cone(e2 . the process of computing the HirzebruchJung continued fraction of d/k yields a convenient method for ﬁnding a reﬁnement Σ of σ such that φ : XΣ → Uσ is a toric resolution of singularities.2. for each 1 ≤ i ≤ r − 1.3.2. Using the recurrences (10. By the induction hypothesis. de1 − ke2 ) be in normal form. we have Pi−1 Pi 1 = + . . Then the cones σi = Cone(ui−1 . this equals 1 bt − bt+1 Pt−1 − Pt−2 1 bt − bt+1 Qt−1 − Qt−2 . HirzebruchJung Continued Fractions and Resolutions. we have Ps Qs+1 − Ps+1 Qs = Ps (bs Qs − Qs−1 ) − (bs Ps − Ps−1)Qs = Ps−1 Qs − Ps Qs−1 =1 by the induction hypothesis. Finally. from part (b). Toric Surfaces where the right side comes from a list of length t. When σ is a cone with parameters d > k > 0.2.464 Chapter 10.4). σi+1 ∩ σi = Cone(ui ). ui ). The base case i = 1 follows directly from (10.
§10. we simply compute the HirzebruchJung continued fraction expansion of the rational number d/k = 7/5 using the modiﬁed Euclidean algorithm: 7 = 2·5−3 5 = 2·3−1 3 = 3 · 1.2. e2 for σ.4. 2e1 − e2 ) σ3 = Cone(2e1 − e2 . e1 ) (10.2. 7e1 − 5e2 ) shown in Figure 3 on the next page. u3 = 3e1 − 2e2 . Example 10. You will see another example of this process in Exercise 10. (d) The toric morphism φ : XΣ → Uσ is a resolution of singularities. u1 = e1 . 3]].5. 7e1 − 5e2 ) in normal form. For part (b). ♦ Q0 = 0 Q1 = 1 Q2 = b2 Q1 − Q0 = 2 Q3 = b3 Q2 − Q1 = 5. Consider the cone σ = Cone(e2 .2. these slopes form a strictly decreasing sequence for i ≥ 0.7) σ2 = Cone(e1 . we note that the ratio −Qi−1 /Pi−1 represents the slope of the line through ui in the coordinate system relative to the normalized basis e1 . By part (d) of Proposition 10.2.2.8 and 10. Part (c) follows from part (b) by noting that u0 = e2 and Pr /Qr = d/k. so ur+1 = de1 − ke2 .2. b2 = 3. Theorem 10.2. The cones σi give a smooth reﬁnement of σ. so the fan Σ consisting of the σi and their faces gives a smooth reﬁnement of σ. P1 = 2. Hence the cones σi ﬁll out σ.2. Then from Proposition 10.2.1. u2 = 2e1 − e2 . P2 = b2 P1 − P0 = 3. Continued Fractions and Toric Surfaces 465 (c) σ1 ∪ · · · ∪ σr+1 = σ. Both statements in part (a) follow easily from part (c) of Proposition 10.3 gives the vectors u0 = e2 . To construct the resolution of singularities of the afﬁne toric surface Uσ . Proof. P3 = b3 P2 − P1 = 7. 3e1 − 2e2 ) σ4 = Cone(3e1 − 2e2 .2. u4 = 7e1 − 5e2 and the cones σ1 = Cone(e2 . Part (d) now follows by the reasoning used in Examples 10. 2. . Hence b0 = b1 = 2. and (10.6) 7/5 = [[2.1.2.2.2 we have P0 = 1. which implies the statement in part (b).
. k = 3. 2. We leave it to the reader to check that 7/3 = [[3. 2]].2.2. . with the partial quotients the same as those in (10. For instance. . the cone of Example 10.1. Toric Surfaces σ1 σ2 σ3 σ4 Figure 3.2.4 used the HirzebruchJung expansion 7/5 = [[2.6).2.2. br ]].2. .8) for 1 ≤ i ≤ r. Theorem 10. ui−1 + ui+1 = bi ui . Let σ = Cone(e2 . bi ≥ 2. ui−1 + ui+1 = (Pi−2 e1 − Qi−2 e2 ) + (Pi e1 − Qi e2 ) = (Pi−2 + Pi)e1 − (Qi−2 + Qi)e2 = bi (Pi−1 e1 − Qi−1 e2 ) = bi ui . Since 5·3 ≡ 1 mod 7. Later in this chapter we will see several important consequences of (10.2.2. This pattern holds for all HirzebruchJung continued fractions. Example 10.2. .2.5. . 2.4 also has parameters d = 7.3 has a nice relation to Theorem 10. u1 .3.466 Chapter 10.3 determine the partial quotients in the HirzebruchJung continued fraction expansion of d/k. The nonuniqueness of Proposition 10.8) connected with the geometry of smooth toric surfaces. . We give a proof that uses the properties of the associated toric surfaces. .3 satisfy (10. 3]].5). The reﬁnement Σ and open circles showing ui = Pi−1 e1 − Qi−1 e2 Next we show that the vectors ui from Theorem 10. and let d/k = [[b1 . b2 . ur+1 constructed in Theorem 10. By the recurrences (10. . but listed in reverse order. Then the vectors u0 . de1 − ke2 ) be in normal form. Proof.
Then: . de1 − ke2 ) be the corresponding cones in normal form. Since σ is strongly convex.7. Continued Fractions and Toric Surfaces 467 Proposition 10. .) • The convex hull Θσ = Conv(σ ∩ (N \ {0})).6. The Zlinear mapping ϕ : N → N deﬁned with respect to the basis e1 . u1 .2.2. (See Proposition 1.2. then the vectors ui are taken to corresponding vectors ui for the cone σ. where the white circles represent the lattice points on the bounded edges. there is an integer d such that d d + kk = 1.2. br−1 . . . The idea is that σ gives two objects: • The semigroup σ ∩ N. But since ϕ and ϕ−1 are orientationreversing. . Let 0 < k. . b1 ]]. We claim that when we apply the mapping ϕ−1 deﬁned by the inverse of the matrix A above. then we obtain vectors ui satisfying the equations ui−1 + ui+1 = bi ui for all 1 ≤ i ≤ r. . If we apply Theorem 10.2. Since kk ≡ 1 mod m. We will see below that these lattice points give the Hilbert basis of σ ∩ N.2.2. ur+1 } be the set of vectors constructed in Theorem 10.2.3. the partial quotients in the HirzebruchJung continued fraction will be listed in the opposite order.3 to σ.7. . de1 − ke2 ) be in normal form and let S = {u0 .§10.3. Thus ϕ(de1 − ke2 ) = e2 and ϕ(e2 ) = de1 − ke2 .2.22. k < d and assume kk ≡ 1 mod d.4. Let σ = Cone(e2 . This is an unbounded polygon in the plane whose bounded edges contain ﬁnitely many lattice points. de1 − ke2 ) and σ = Cone(e2 . e2 by the matrix A= k d d −k is bijective. its irreducible elements form the unique minimal generating set called the Hilbert basis of σ ∩ N. and is orientationreversing. Figure 4 on the next page shows the convex hull Θσ . You will complete the proof of this assertion in Exercise 10. . ♦ Here is the general result suggested by Example 10. If the HirzebruchJung continued fraction expansion of d/k is d/k = [[b1 . Our next result gives two alternative ways to understand the vectors ui in Theorem 10. Hilbert Bases and Convex Hulls. b2 . . Example 10. Consider the cone σ = Cone(e2 .2. . then the HirzebruchJung continued fraction expansion of d/ k is d/ k = [[br . . maps σ to σ.3.2. br ]]. Let σ = Cone(e2 . . 7e1 − 5e2 ) from Example 10.8. Theorem 10. Proof.
.e.3. then ui would have to be a linear combination of the vectors in S \ {ui } with nonnegative integer coefﬁcients. We claim next that all the ui are irreducible elements of σ ∩ N. But this follows from what we have already said.3 that the slopes of the ui are strictly decreasing. ui = Pi−1 e1 − Qi−1 e2 = j=i c ju j = j=i c j Pj−1 e1 − j=i c j Q j−1 e2 with c j ≥ 0 in Z.2. . ui . ci ≥ 0 in Z. Finally. Qi−1 = j=i c j Q j−1 . This contradicts the observation made in the proof of Theorem 10. we use the notation of Theorem 10. Using σ = σ1 ∪ · · · ∪ σr+1. Since the Pi and Qi are strictly increasing by part (a) of Proposition 10. The convex hull Θσ and the lattice points on its bounded edges (a) S is the Hilbert basis of the semigroup σ ∩ N. (b) S is the set of lattice points on the bounded edges of Θσ . . ui−1 }. Since σ = σ1 ∪ · · · ∪ σr+1 . If 1 ≤ i ≤ r and ui is not irreducible. . Toric Surfaces Θσ Figure 4. we must show that there are no other irreducible elements in σ ∩ N. But this would imply that ui is a linear combination with nonnegative integer coefﬁcients of the vectors in {u0 . one sees easily that S generates σ ∩ N. where the cone σi is generated by ui−1 . Proof. This is clear for u0 = e2 and ur+1 = de1 −ke2 since they are the ray generators for σ. i.468 Chapter 10. then u ∈ σi ∩ N for some i. if u is irreducible.2. For part (a). Then σi ∩ N is generated as a semigroup by ui−1 . we must have c j = 0 for all j > i.. Hence Pi−1 = j=i c j Pj−1 .2. . Thus u is irreducible only if u = ui−1 or ui . But then u = ci−1 ui−1 + ci ui for some ci−1 . ui since σi is smooth.2. It follows that the ui are irreducible elements of σ ∩ N.
we have Pi−2 Qi − PiQi−2 = Pi−2 (bi Qi−1 − Qi−2 ) − (bi Pi−1 − Pi−2 )Qi−2 = bi (Pi−2 Qi−1 − Pi−1 Qi−2 ) = bi ≥ 2. . The ai are the (ordinary) partial quotients of d/k. so that the ith division is given by (10. To distinguish these from HirzebruchJung continued fractions. Ordinary Continued Fractions.2. and the truncated continued fractions [a1 . .§10.2. inequality tells us that the slopes of the line segments ui−1 ui and ui ui+1 are related by slope of ui−1 ui ≤ slope of ui ui+1 . a2 . This implies that these line segements lie on boundary of Θσ . we will use the notation (10. Starting with k−1 = d and k0 = k. . ki−2 = ai ki−1 + ki . Combining this with part (c) of Proposition 10. .2. one obtains the inequality −(Qi−1 − Qi−2 ) −(Qi − Qi−1 ) ≤ . 1 ≤ i ≤ s. . ai ]. a2 . Pi−1 − Pi−2 Pi − Pi−1 Since ui = Pi−1 e1 − Qi−1 e2 .2. as ] for the ordinary continued fraction. it is easy to see that the ui are the lattice points of the bounded edges of Θσ . The HirzebruchJung continued fractions studied above are less familiar than ordinary continued fraction expansions in which the minus signs are replaced by plus signs. are the (ordinary) convergents of d/k.2. If d > k > 0 are integers. . k)). .9) where 0 ≤ ki < ki−1 . . we write ai for the quotient and ki for the remainder at each step. Let ks−1 be the ﬁnal nonzero remainder (which equals gcd(d. The resulting equations splice together to form the ordinary continued fraction d/k = a1 + a2 + 1 1 ··· + 1 as . Continued Fractions and Toric Surfaces 469 For part (b). From here.2.2.10) d/k = [a1 . ﬁrst observe that by Proposition 10. then the ordinary continued fraction expansion of d/k may be obtained by performing the same sequence of integer divisions used in the usual Euclidean algorithm for the gcd. .
470
Chapter 10. Toric Surfaces
Example 10.2.9. By Example 10.2.1, the HirzebruchJung continued fraction of 17/11 is 17/11 = [[2, 3, 2, 2, 2, 2]], and the ordinary continued fraction is 17/11 = [1, 1, 1, 5]. The partial quotients and the lengths are different. However, each expansion determines the other, and there are methods for computing the HirzebruchJung partial quotients b j in terms of the ordinary partial quotients ai and vice versa. See [41, Prop. 3.6], [88, p. 257], or [146, Prop. 2.3]. See also Exercise 10.2.4. ♦ The following result is mostly parallel to Proposition 10.2.2, but shows that ordinary continued fractions are slightly more complicated than HirzebruchJung continued fractions. The proof is left to the reader (Exer