Elliptic Surface of Kodaira Dimension 1

Selecta Mathematica (2022) 28:25 Selecta Mathematica
https://doi.org/10.1007/s00029-021-00747-x New Series
Transversality of sections on elliptic surfaces with

applications to elliptic divisibility sequences and
geography of surfaces
Douglas Ulmer1 · Giancarlo Urzúa2
Accepted: 18 November 2021

© The Author(s), under exclusive licence to Springer Nature Switzerland AG 2021
Abstract
We consider elliptic surfaces E over a field k equipped with zero section O and another
section P of infinite order. If k has characteristic zero, we show there are only finitely
many points where O is tangent to a multiple of P. Equivalently, there is a finite list
of integers such that if n is not divisible by any of them, then n P is not tangent to
O. Such tangencies can be interpreted as unlikely intersections. If k has characteristic
zero or p > 3 and E is very general, then we show there are no tangencies between
O and n P. We apply these results to square-freeness of elliptic divisibility sequences
and to geography of surfaces. In particular, we construct mildly singular surfaces of
arbitrary fixed geometric genus with K ample and K 2 unbounded.
Keywords Elliptic surfaces · Unlikely intersections · Elliptic divisibility sequences ·

Stable surfaces · Geography of surfaces
Mathematics Subject Classification Primary 14J27; Secondary 11B39 · 14J29
1 Introduction
Our aim in this paper is to study transversality properties of sections of elliptic sur-
faces and to deduce consequences for elliptic divisibility sequences and geography of
surfaces.
B Douglas Ulmer
[email protected]
Giancarlo Urzúa
[email protected]
1 Department of Mathematics, University of Arizona, Tucson, AZ 85721, USA

2 Facultad de Matemáticas, Pontificia Universidad Católica de Chile, Santiago, Chile
0123456789().: V,-vol
25 Page 2 of 36 D. Ulmer, G. Urzúa
To state the first result, let k be a field of characteristic zero and let C be a smooth,
projective, geometrically irreducible curve over k. Let π : E → C be a relatively
minimal Jacobian elliptic surface over k (i.e., a smooth, projective elliptic surface
with a section O which will play the role of zero section), and let P be another
section. We write n P for the section induced by multiplication by n in the group law
of the fibers of E → C. Assume that P has infinite order, i.e., n P = O for all n = 0.
As we will see below, except in degenerate situations the intersection number (n P).O
grows like a constant times n 2 . Our first result says that the intersections are usually
transverse.
Theorem 1.1 The set

T = {t ∈ C | n P is tangent to O over t }
n=0
is finite.
Here and in the rest of the paper, we conflate the sections O : C → E and P : C → E
with their images O(C) ⊂ E and P(C) ⊂ E. Thus we say “P is tangent to O” rather
than “the image of P is tangent to the image of O.”
In [32], we give an explicit upper bound on the cardinality of T .
Remark 1.2 We note that a tangency between n P and O can be regarded as an “unlikely
intersection” as follows: Let TE be the tangent bundle of E and let PTE be the associated
projective bundle. Thus PTE → E is a P1 -bundle, and the total space PTE is a smooth,
projective threefold. If C ⊂ E is a smooth curve, then there is a canonical lift of C to
C̃ ⊂ PTE defined by sending a point t ∈ C to the class of its tangent line TC,t ⊂ TE ,t
in PTE . Two curves C1 and C2 in E that meet at y ∈ E are tangent there if and only
if their lifts meet at a point of PTE over y. Thus a tangency between C1 and C2 is
equivalent to the “unlikely” intersection of the two curves C̃1 and C̃2 in the threefold
PTE . We refer to [33] for a comprehensive account of work on unlikely intersections
up to 2012.
Remarks 1.3 (1) A result very similar to our Theorem 1.1 was communicated to us
by Corvaja, Demeio, Masser, and Zannier after we posted the first version of this
paper. Their methods are rather different, see [11]. They show more generally
that finiteness holds when the cyclic group {n P|n ∈ Z} is replaced by a finitely
generated, torsion-free group of sections.
(2) On the other hand, our methods lead to non-trivial results in families, and in partic-
ular we show that for “generic” data, the set T above is empty. (See Theorems 1.7
and 1.8 below.) This is crucial for our application to geography of surfaces.
(3) In the first version of this paper, we used a trivialization essentially equivalent to
the Betti foliation discussed in Sect. 3 of this version. Later, we learned of the
“Betti” terminology used by several authors, including in [12], and adopted it in
this paper.
We next reformulate Theorem 1.1 in analogy with the “elliptic divisibility sequence”
associated to an elliptic curve and a point. (See [28, Exers. 3.34-36, 9.4, 9.12] for
Transversality of sections on elliptic surfaces Page 3 of 36 25
definitions and examples, and [18] for more on the function field case.) Define a
sequence of effective divisors on C for n ≥ 1 by
Dn := O ∗ (n P),
i.e., Dn is the pull-back along the zero section of the divisor n P on E. (We will give
several other equivalent definitions in Sect. 2.)
The sequence Dn is a natural analogue of an elliptic divisibility sequence. In par-
ticular, we will see below that if m divides n, then Dm divides Dn (i.e., Dn − Dm is
effective), and that Möbius inversion gives a sequence of effective divisors Dm such
that

Dn = Dm .
m|n
We say that a divisor on C is reduced if it has the form

D= ti
i
where the ti are distinct closed points of C (i.e., each non-zero coefficient of D equals
1). This is an analogue of an integer being square-free.
Theorem 1.4 Given E and P as above, there is a finite set of integers M =
{m 1 , . . . , m k } such that
(1) O and n P intersect transversally if and only if n is not divisible by any element
of M.
(2) Dn is reduced if and only if n is not divisible by any element of M.
(3) Dm is reduced if and only if m ∈/ M.
Remark 1.5 Thereom 1.4 is much stronger than what one might predict from standard
conjectures. For simplicity, assume that C = P1 , let F = C(C), and let E/F be the
generic fiber of E → C. Choosing a coordinate t on P1 so that none of the Dn involve
the place at infinity, we may identify each Dn with a monic polynomial f n in t, and
to say that Dn is reduced is to say that gcd( f n , d f n /dt) = 1. Arguments similar to
those in [27] applied to a certain Buium jet space of E/F together with a function
field analogue of Vojta’s conjecture suggest that
?
deg gcd( f n , d f n /dt) = o(n 2 ) = o(h(n P))
where h(n P) is the canonical height of n P. (See Sect. 2.7 for definitions.) But Theo-
rem 1.4 shows that deg gcd( f n , d f n /dt) is in fact bounded!
Remark 1.6 We have no reason to believe that the analogues of Theorems 1.1 and 1.4
(with n restricted to be prime to the characteristic) are false in positive characteristic.
However, our proof uses analytic techniques and does not obviously carry over to the
arithmetic situation.
The next two results hold for k a field of characteristic zero or sufficiently large p.
As before, C is a smooth, projective, geometrically irreducible curve over k. The next
result says roughly that if E → C is a very general Jacobian elliptic surface with an
additional section P, there are no tangencies between n P and O for n = 0. Recall
that a line bundle L on C is said to be globally generated (or base point free) if for
every t ∈ C, there is a global section of L which does not vanish at t.
Theorem 1.7 Suppose k is a field of characteristic zero or of characteristic p > 3,

and let C be as above. Let L be a globally generated line bundle on C of degree d and
set
V = H 0 (L 2 ⊕ L 3 ⊕ L 4 ).
Then for a very general a = (a2 , a3 , a4 ) ∈ V , the elliptic surface E → C associated

to
E : y 2 + a3 y = x 3 + a2 x 2 + a4 x
equipped with the usual zero section O and the section P = (0, 0) has the following
properties:
(1) P has infinite order.

(2) The singular fibers of E → C are nodal cubics (i.e., Kodaira type I1 ).
(3) P meets each singular fiber in a non-torsion point.
(4) If n is not a multiple of the characteristic of k, then n P meets O transversally in
d(n 2 − 1) points.
(Here and elsewhere in the paper, L i means L ⊗i and not L ⊕i .) We will explain the
construction of the elliptic surface attached to a in Sect. 5.4 and the meaning of “very
general” in Sect. 6.
As with many results about “very general” points, Theorem 1.7 does not allow one
to deduce the existence of examples over “small” (countable) fields such as number
fields or global function fields. However, after relaxing condition (2) above, we can
write down such examples explicitly, at least when L is the square of a globally
generated line bundle.
Theorem 1.8 Let k be a field of characteristic 0 or a field of characteristic p > 2 which

is not algebraic over the prime field F p . Let C be a smooth, projective, geometrically
irreducible curve over k with a non-trivial line bundle L which is the square of a
globally generated line bundle F. Then there exist infinitely many pairs (E, P) where
E is a Jacobian elliptic surface E → C equipped with a section P such that:

(2) The singular fibers of E → C are of Kodaira type I0∗ .

⎧ 2
⎪
⎪ n −1
⎨ d if n is odd,
2
⎪
⎩ n − 4 d if n is even
2
⎪
2
points, where deg L = 2d.
(5) O ∗ (1E /C ) ∼
= L.
The starting point for our collaboration was a remarkable application of Theo-
rem 1.7 to the geography of surfaces due to the second-named author. We give some
background before stating the result: There has been a great deal of interest in volumes
of stable surfaces, i.e., in the set of values of K X2 where X runs through stable surfaces
in the sense of Kollár, Shepherd-Barron, and Alexeev (KSBA). A key ingredient in the
construction of the KSBA compactifications of moduli spaces of surfaces of general
type (see [21]) is a descending chain condition on the set of K X2 , namely that it admits
no strictly decreasing sequences. The famous article [4] establishes this condition.
(See also [5].) As this is a set of positive rational numbers, it must have a minimum,
whose value is still unknown. Aleexev and Liu have recently found various other spe-
cial properties around accumulation points and bounds [1–3,22]. For example, [22]
shows that when the geometric genus is positive then the volume can be optimally
bounded from below. This naturally raises the question: Are there upper bounds for
K 2 of stable surfaces with a fixed geometric genus?
Recall [8, I.5.5, VI.1] that a smooth, projective, non-ruled surface X over a field
of characteristic zero has c2 (X ) ≥ 0, so Noether’s formula shows that K X2 ≤ 12(1 +
pg ), i.e., the self-intersection of the canonical bundle K X is bounded in terms of the
geometric genus pg . The question is whether such bounds continue to hold for mildly
singular surfaces.
Any geometric approach to a negative answer to this question via surfaces with
rational singularities requires a family of minimal resolution surfaces for any given
pg > 0 with an unbounded number of special rational curves. Our work supplies such
families and moreover allows for good control on the singularities involved:
Theorem 1.9 Given integers g ≥ 0 and N , there exists a normal projective surface X
over C with the following properties:
(1) X has geometric genus pg = g.
(2) X has only one singular point, which is log-terminal.
(3) K X is Q-Cartier and ample.
(4) K X2 > N .
1.10 Plan of the paper
In Sect. 2 we present foundational material on torsion points and intersections on

elliptic surfaces, including a discussion of basic properties of our elliptic divisibility
sequences. We then reformulate Theorems 1.1 and 1.4 as Theorem 2.5. We prove
Theorem 2.5 in Sect. 3. Section 4 discusses two moduli spaces which play a key role
in the proof of Theorem 1.7. Section 5 discusses a construction of elliptic surfaces
equipped with extra structure associated to elements in certain Riemann-Roch spaces.
We then prove Theorem 1.7 in Sect. 6. In Sect. 7, we give an explicit construction
of surfaces satisfying the requirements of Theorem 1.8. Finally, in Sect. 8, we prove
Theorem 1.9.
2 Preliminaries on torsion and intersections
In this section we gather various foundational results on torsion, intersections, heights,

and elliptic divisibility sequences. Some of this material also appears in [18], although
our point of view is more geometric. Throughout, π : E → C will be a relatively
minimal Jacobian elliptic surface over a field k with zero section O.
2.1 Multiplication by n
Let E sm denote the locus where π is smooth (i.e., the complement of the singular
points in the bad fibers). Then by [15, Prop. II.2.7], E sm is a commutative group
scheme over C. Let n be an integer not divisible by the characteristic of k. Consider
the homomorphism of group schemes given multiplication by n: [n] : E sm → E sm .
Clearly, [n] fixes the zero section O pointwise. If x ∈ O, the tangent space to E sm at
x splits canonically into the sum of two lines, namely the tangent space to O at x and
the tangent space to the fiber of π through x. Since [n] fixes O, [n] acts as the identity
on the former. A calculation in the formal group of E sm [28, Ch. IV] shows that [n]
acts as multiplication by n on the tangent space to the fiber of π through x. It follows
that [n] is étale at every point of O, and since [n] is a group scheme homomorphism
it is étale everywhere.
The morphism [n] is also quasi-finite: it has degree n 2 on the smooth geometric
fibers of π , and degree dividing n 2 on all geometric fibers [28, Ch. III and §VII.6]. It
is not in general finite if π has singular fibers.
If P : C → E is a section of π , P necessarily lands in the smooth locus E sm and
we may define a new section n P as the composition [n] ◦ P. This is the meaning of
the notation n P used in the introduction.
2.2 Torsion
With E as above and n > 0 and relatively prime to the characteristic of k, we define
E[n] = [n]−1 (O),
i.e, E[n] is the inverse image of the zero section under [n]. Since [n] is étale, E[n] is
a reduced, closed subscheme of E sm of dimension 1, and in particular, locally closed
in E. Since [n] is quasi-finite, E[n] is étale and quasi-finite over C of generic degree
n 2 . It is in general not finite over C.
The fiber of E[n] over a geometric point t of C consists of the points of π −1 (t) with
order divisible by n. We define
E[n] ⊂ E[n]
to be the subscheme whose fiber over t consists of the points of π −1 (t) of order exactly
n. If m divides n, then E[m] is a closed subscheme of E[n], and we have a disjoint
union
E[n] = ∪m|n E[m] (2.1)
where m runs over the positive divisors of n. Each E[m] is a union of irreducible
components of E[n] and is étale and quasi-finite over C. Note that E[1] = E[1] = O.
We refer to the unions of irreducible components of E[n] as “torsion multisections”.
2.3 Divisibility sequences
In the introduction, we defined divisors Dn for n ≥ 1 by
Dn = O ∗ (n P).
In this section we examine alternative definitions and properties of these divisors

always assuming that n is relatively prime to the characteristic of k.
For two smooth curves C1 and C2 on E with no irreducible components in com-
mon, write C1 ∩ C2 for the intersection zero-cycle. This is a zero-dimensional closed
subscheme of E. With this notation,
Dn = π∗ (n P ∩ O) = π∗ (n P ∩ E[1]) .
Note that n P meets O = E[1] over t if and only if P meets E[n] over t, and since
[n] is étale, the intersection multiplicity of n P and E[1] over t equals the intersection
multiplicity of P and E[n] over t. In other words, we have
Dn = π∗ (P ∩ E[n]) . (2.2)
Define

Dn = π∗ P ∩ E[n] .
Then the disjoint union (2.1) yields a decomposition of Dn into effective divisors:

Dn = Dm (2.3)
m|n
where the sum is over positive divisors of n.

Note in particular that if t is a closed point of C and P(t) is a torsion point, say of
order exactly m, then t appears in Dn if and only if m divides n, and the multiplicity
of t in such Dn is equal to the multiplicity of t in Dm .
Remark 2.4 A section P can meet at most one torsion point over a given t ∈ C. This
and D have disjoint support. In particular, as soon
implies that if m 1 = m 2 , then Dm 1 m2

as Dn = 0, Dn has “primitive divisors’ i.e., points in its support which are not in the
support of Dm for m < n. The existence of primitive divisors for all sufficiently large
n is established in [18, §5] by showing that Dn = 0 for all sufficiently large n. The
key idea is an estimation of intersection numbers using heights as in Sect. 2.7 below.
Another simple proof of the existence of primitive divisors (suggested by a referee)
can be given using the fact that the “Betti coordinates” (as in Sect. 3) of a non-torsion
section P give rise to a locally defined, open map from C to R2 . The existence and
openness of this map can also be viewed as the key point in the simplest case (I0 ) of
the proof of Theorem 2.5.
We now state a result which implies Theorems 1.1 and 1.4:
Theorem 2.5 Let E → C be a relatively minimal Jacobian elliptic surface over the
complex numbers C, with zero section O and another section P which is not torsion.
Then the set

Ttor := {t ∈ C | P is tangent to E[n] over t }
n=0
is finite.
2.6 Proof that Theorem 2.5 implies Theorems 1.1 and 1.4
First we note that the general cases of Theorems 1.1 and 1.4 follow from the case
k = C. Indeed, since the hypotheses and conclusion of Theorems 1.1 and 1.4 are
insensitive to the ground field, we may replace k with a subfield k which is finitely
generated over Q (take the field generated by the coefficients defining C, E, π , and
P), then embed k in C. Thus it suffices to treat the case k = C.
Next, by the definition of Dn , to say that n P is tangent to O over t is to say that
t appears in Dn with multiplicity greater than 1. By the equality (2.2), to say that t
appears in Dn with multiplicity greater than 1 is to say that P is tangent to E[n] over t.
Thus the set Ttor of Theorem 2.5 is equal to the set T of Theorem 1.1, and Theorem 2.5
is equivalent to the case k = C of Theorem 1.1.
To finish, we show that Theorems 1.1 and 1.4 are equivalent. First note that points
(1) and (2) of Theorem 1.4 are equivalent by the definition of Dn . Moreover, Dm is
non-reduced if and only if Dn is non-reduced for all multiples n of m. Thus point

(3) of Theorem 1.4 implies points (1) and (2), and Theorem 1.4 is equivalent to the
statement that the set of m such that Dm is non-reduced is finite.
Now consider the “incidence correspondence”
I := (t, m) m > 0 and P is tangent to E[m] over t ⊂ C × Z>0 .

The set T of Theorem 1.1 is the image of the projection I → C and the set M of
Theorem 1.4 is the image of the projection I → Z>0 . The fibers of I → C are finite
(and in fact empty or singletons) because P meets E[m] for at most one value of m
and a fortiori can be tangent to at most one E[m] . The fibers of I → Z>0 are finite
because for a fixed m, P meets E[m] at only finitely many points, so a fortiori can
be tangent to E[m] at only finitely many points. This establishes that Theorem 1.1
and Theorem 1.4 are equivalent, and it completes the proof that Theorem 2.5 implies
Theorems 1.1 and 1.4

We will prove Theorem 2.5 in Sect. 3. First, we review material on heights used
later in the paper.
2.7 Heights
We refer to [13] or [25] or [26] or [31] for the basic assertions on heights in this section.
As usual, π : E → C is a relatively minimal Jacobian elliptic surface over a field k.
Given a section P of π , there is a unique Q-divisor C P supported on the non-
identity components of the fibers of π with the property that P − O + C P has zero
intersection multiplicity with every irreducible component of every fiber of π . There
is a simple recipe for C P that depends only on the components of the reducible fibers
met by P, and in particular, for a fixed E, there are only finitely many possibilities for
C P as P varies over all sections. If π has irreducible fibers, or more generally, if P
passes through the identity component of every fiber, then C P = 0.
There is a canonical Q-valued symmetric bilinear form on the group of sections of
E defined by
P, Q := −(P − O + C P ).(Q − O) (2.4)
where the dot refers to the intersection number on E. If E → C is non-constant (i.e.,

is not isomorphic over k to a product E 0 × C), then this pairing is non-degenerate
modulo torsion and positive definite. We define ht(P) = P, P. (Note that this is
twice the height considered in [18].)
Lemma 2.8 For E and P as above,
ht(P) 2
(n P).O = n + O(1).
2
If π has irreducible fibers and d = deg O ∗ (1E /C ), then
ht(P) 2
(n P).O = n − d.
2
In the first display, O(1) depends only on E and P, not on n.
Proof It follows from the canonical bundle formula for elliptic surfaces, adjunction,
and the definition of d that every section P of π satisfies P 2 = −d. From the height
formula, we find
n 2 ht(P) = ht(n P) = −(n P − O + Cn P ).(n P − O),
and so
ht(P) 2 ht(P) 2
(n P).O = n − d + Cn P .(n P − O) = n + O(1).
2 2
If π has irreducible fibers, then C Q = 0 for all Q and we deduce the stated exact
formula.

In the following lemma we use square brackets to indicate the class of a curve in
NS(E), the Néron-Severi group of E. This allows us to distinguish between n[P] (n
times the class of P) and [n P] (the class of n P).
Lemma 2.9 Suppose that π : E → C has irreducible fibers, d = deg O ∗ (1E /C ), and
P is a section of π which does not meet O. Let F be a fiber of π . Then we have an
equality
[n P] = n[P] + (1 − n)[O] + d(n 2 − n)[F]
in NS(E).
Proof We have an equality [n P] − [O] = n([P] − [O]) in the Picard group of the
generic fiber of E, so there is an equality of the form
[n P] = n[P] + (1 − n)[O] + c[F]
in NS(E), and we just need to determine the coefficient of [F]. We do this by intersect-
ing with [O]. By assumption [P].[O] = 0, so ht(P) = 2d. By the previous lemma,
[n P].[O] = d(n 2 − 1) and solving for c yields c = d(n 2 − n).

3 Proof of Theorem 2.5
We first note that Theorem 2.5 is a statement about intersections on an elliptic surface
over the complex numbers. To prove it, we may replace E and C with the corresponding
complex manifolds and make use of the classical topology, i.e., the topology induced
by the metric topology on C. For the rest of this section, we make this replacement,
although we will not change the notation.
Our strategy will be to show that the subset Ttor ⊂ C is closed and discrete (in the
classical topology). This implies Theorem 2.5 since C is compact. We note in passing
that the set of points of intersection of P and E[n] for varying n is usually everywhere
classically dense in P, so the discreteness that lies at the heart of the theorem is not
evident.
In fact, will consider a more general set of tangencies and prove that a certain subset
TBetti ⊂ C is closed and discrete (and thus finite) and contains all points of Ttor over
which E has good reduction. Since the set of points of bad reduction is finite, this will
establish that Ttor is also finite.
We will establish the desired discreteness by using the complex analytic description
of E → C given by Kodaira in [20, §8]. Let C 0 ⊂ C be the maximal open subset over
which E has good reduction, and let E 0 = π −1 (C 0 ).
Let t ∈ C 0 . Write H for the upper half plane. Then there is a neighborhood U of t
biholomorphic to a disk and holomorphic functions τ : → H and w : → C
such that π −1 (U ) → U sits in a diagram
π −1 (U ) ( × C)/(Zτ + Z)
P|U [w] (3.1)
U
where the horizontal maps are biholomorphic, and ( × C)/(Zτ + Z) means the
quotient of × C by Z2 acting as
(a, b)(z, w) = (z, w + aτ (z) + b) .
For z ∈ , corresponding to u ∈ U , P(u) corresponds to [w](z), which is the class

of w(z) in {z} × C/(Zτ (z) + Z). We also assume t ∈ U corresponds to 0 ∈ .
Next, we consider a trivialization of π −1 (U ) → U as a real analytic manifold.
Introduce real coordinates as follows: z = x + i y on the base , τ = ρ + iσ on
the upper half plane H, and w = u + iv in the C which uniformizes the fibers of
π −1 (U ) → U . Let (r , s) be coordinates on R2 , and note that w = r τ + s if and only
if r = v/σ and s = (uσ − vρ)/σ .
Consider the diagram
×C × R2
( × C)/(Zτ + Z) × (R/Z)2 (3.2)
where the upper horizontal map is

v uσ − vρ
(z, w) = (x + i y, u + iv) → (z, r , s) = z, , ,
σ σ
with inverse
(z, w) = (z, r τ (z) + s) ← (z, r , s).

The two vertical maps are the natural quotients, the middle horizontal map is induced
by the upper horizontal map, and the diagonal maps are the projections to the first
factor.
The top horizontal map is a real-analytic isomorphism which is R-linear on each
fiber of the projection to . The choice of this map is motivated by the fact that torsion
sections of E over U correspond the surfaces × (r , s) ⊂ × (R/Z)2 where r and
s are rational numbers. In other words, we have changed coordinates so that every
torsion section becomes a constant section. It will be of interest to consider all the
horizontal sections × (r , s) for arbitrary real numbers r and s.
Definition 3.1 With notation as above and (r , s) ∈ R2 , define the local Betti leaf
Lr ,s ⊂ π −1 (U ) as the image of the map U → ( × C)/(Zτ + Z) ∼= π −1 (U ) sending
z to the class of (z, r τ (z) + s).
This terminology is inspired by [12], where r and s are called “Betti coordinates”.
Clearly the assignment (r , s) → Lr ,s factors through (R/Z)2 . Each Lr ,s is a closed
holomorphic submanifold of π −1 (U ), and the set of Lr ,s as (r , s) runs through (R/Z)2
is a foliation of π −1 (U ). Although the indexing of Lr ,s by (r , s) depends on the choice
of period map τ , the submanifolds Lr ,s themselves are intrinsic (i.e., independent of
τ ). There is a corresponding global foliation of E 0 which we will not consider in this
paper, except implicitly in the following remark: If for some non-empty open U ⊂ C 0
and some (r , s), P(U ) = Lr ,s (i.e., P lands in a leaf of the local foliation), then by
analytic continuation, the same holds over every open. In this case, we say “P lies in
the Betti foliation”.
Note that if (r , s) ∈ Q2 , then each point of Lr ,s is a torsion point in its fiber. More
precisely, if n is the smallest positive integer such that (nr , ns) ∈ Z2 , then Lr ,s is a
connected component of E[n] ∩ π −1 (U ). On the other hand, if (r , s) ∈ R2 \ Q2 , then
Lr ,s is disjoint from every E[n]. Thus, if P lies in the Betti foliation, it is a torsion
section (which we have ruled out by hypothesis) or it meets no torsion sections over
C 0 and so is not tangent to any torsion section over C 0 . In the latter case, Theorem 2.5
is obviously true, so we may assume from now on that P does not lie in the Betti
foliation.
We now define

TBetti := t ∈ C 0 |P is tangent to some Lr ,s over t .
The preceding paragraph shows that Ttor ∩ C 0 ⊂ TBetti and so, as explained above,
to prove Theorem 2.5 it will suffice to prove that TBetti is a closed, discrete subset of
C. More formally:
Claim 3.2 For every t ∈ C, there is a classical open neighborhood Ut of t in C such

that (Ut \ {t}) ∩ TBetti = ∅. In other words, for every (r , s) ∈ R2 , P is not tangent to
Lr ,s over Ut \ {t}.
To establish Claim 3.2, we will consider cases according to the reduction type of E
at t. We use the standard Kodaira notation (In , In∗ ,…) to index the cases.
The case of I0 reduction: Using diagrams (3.1) and (3.2), we identify the section
P over U with the graph of a function φ : → R2 . Write (r0 , s0 ) = φ(0) for the
image of P over t. Since P is assumed not to be contained in Lr0 ,s0 , we may shrink
U so that P meets Lr0 ,s0 only over t, in other words, so that the only value of z with
φ(z) = (r0 , s0 ) is z = 0.
It is clear that P is tangent to some Lr ,s over t if and only if the derivative of
(x, y) → (r , s) (as a map of 2-manifolds) vanishes at the z corresponding to t .
To finish, we claim that after possibly shrinking U and , the derivative of φ does
not vanish away from 0 ∈ . To see this, apply the Lojasiewicz gradient inequality
([10,24]) to the components of (φ1 , φ2 ) of φ: That result says that after shrinking ,
there are constants C > 0 and 0 < θ < 1 such that
|∇φi (z)| ≥ C |φi (z) − φi (0)|θ
for all z ∈ . But if z ∈ \ {0}, φ(z) = φ(0) so one of the φi (z) = φi (0) which
implies that ∇φi (z) = 0 and so the derivative of φ is also non-zero.
This establishes Claim 3.2 at points of good reduction: if t is such a point, there
is an open neighborhood Ut of t in C such that P is not tangent to any Lr ,s over any
t ∈ Ut \ {t}. To finish the proof, we deal with tangencies near places of bad reduction.
The case of I1 reduction: We next consider the case of multiplicative reduction with
an irreducible special fiber. I.e., assume that E has reduction type I1 over t ∈ C. Let
→ C be a holomorphic parameterization of a neighborhood of t, where is the
unit disk and 0 ∈ maps to t. Again, over the course of the proof we will reduce the
radius of but not change the notation. Let X → be the pull-back of E → C to
, let = \ {0}, and let X → be the restriction of X → to . Note that
the special fiber
X \ X = nodal cubic ∼
= C× ∪ {q}
where q is the node of the cubic.

According to Kodaira [20, pp. 596ff], we may shrink and choose → C so that
X has the form

X ∼
= × C× /Z
where the action of Z on × C× is
m · (z, w) = (z, z m w).
Moreover, as explained starting in the last paragraph of [20, p. 597], there is a

holomorphic map
φ : × C× → X
such that {0} × C× maps biholomorphically to the complement of q in the special

fiber, and × C× → X ⊂ X is the natural quotient map. We may thus identify the
section P with a holomorphic map f : → C× .
Now let V ⊂ be a non-empty, connected and simply connected subset, and
choose a branch of the logarithm log : V → C. Then the local Betti leaf Lr ,s over
V (computed with respect to the period τ (z) = (1/2πi) log z) is the image of the
map V → X which sends z to the class of (z, er log z e2πis ). The local Betti leaf Lr ,s
through (z, w) has

log |w| 1 w
r= and s= log .
log |z| 2πi |w|
(Note that the logarithms appearing in the expression for r are evaluated at real num-
bers, so we use the standard real logarithm, and the class of s in R/Z is independent
of the choice of logarithm.)
Then we calculate that P is tangent to Lr ,s at z ∈ V if and only if
f (z) log | f (z)|

f (z) = z → er log z e2πis (z) = . (3.3)
z log |z|
Note that the expression on the right is well defined independently of the choice of V
and the logarithm.
Now we assume that there is a sequence of tangencies accumulating at t and derive
a contradiction. More precisely, assume that there is a sequence z i ∈ tending to 0
such that for all i,
f (z i ) log | f (z i )|
f (z i ) = . (3.4)
zi log |z i |
We will show that there is no holomorphic function satisfying such equalities.

If f is identically zero, then f is a non-zero constant. Equation (3.3) shows that
the constant value of f must have absolute value 1, so f (z) = e2πis for some real s,
and we find that P lies in a local Betti leaf L0,s , in contradiction to our assumption.
Now assume that f is not identically zero, so f takes its value at z = 0 to finite
order N = ord z=0 ( f (z) − f (0)) ≥ 1. Then g(z) = z f (z)/ f (z) is holomorphic on
, and Eq. (3.4) says that g(z i ) = (log | f (z i )|)/(log |z i |) for all i. Shrinking if
necessary, we have estimates
B1 |z| N < |g(z)| < B2 |z| N (3.5)
for some positive constants B1 and B2 and all z ∈ . Shrinking again if necessary,
we may write f (z) = w0 (1 + h(z)) with
C1 |z| N < |h(z)| < C2 |z| N

for some positive constants C1 and C2 and all z ∈ . We have
log |1 + h(z)| ≤ log(1 + |h(z)|) ≤ |h(z)| < C2 |z| N
for z ∈ . If |w0 | = 1 (so log |w0 | = 0), we have
log | f (z)| log |1 + h(z)| C2 |z| N

= ≤ .
log |z| log |z| log |z|
Taking z i close to zero and noting that g(z i ) = (log | f (z i )|)/(log |z i |) we get a
contradiction to the lower bound in Equation (3.5).
To finish, assume that |w0 | = 1. Then
log | f (z)| log |w0 | log |1 + h(z)| log |w0 | C2 |z| N

≥ − ≥ − .
log |z| log |z| log |z| log |z| log |z|
Taking z i close to zero and noting that g(z i ) = (log | f (z i )|)/(log |z i |) we get a
contradiction to the upper bound in Equation (3.5).
This establishes that there is no accumulation of tangencies between P and local
Betti leaves at t when E has reduction of type I1 at t.
The case of Ib reduction: Now consider the case of multiplicative reduction of type Ib
over t ∈ C. This case is very similar to the I1 case, with some notational complications.
Let → C be a holomorphic parameterization of a neighborhood of t, where is
the unit disk and 0 ∈ maps to t. Again, over the course of the proof we will reduce
the radius of but not change the notation. Let X → be the pull-back of E → C
to , let = \ {0}, and let X → be the restriction of X → to . Then the
special fiber X \ X has the form

X \ X = chain of b copies of P1 ∼
= Ci× ∪ {qi }
i∈Z/bZ
where the qi are the nodes of the chain. Let
X sm = X \ {q1 , . . . , qb }
be the smooth locus of X → .

Kodaira [20, pp. 599ff], gives a covering of X sm by b open sets as follows: for
i ∈ Z/bZ, let

Wi = Wi ∪ Ci× , Wi = × C× /Z
where the action of Z on × C× is
m · (z, w) = (z, z bm w).

For z ∈ and w ∈ C× , write (z, w)i for the class of (z, w) in Wi . Then X sm is
obtained by glueing the Wi according to the rule
(z, w)i = (z, z j−i w) j
for all z ∈ , w ∈ C× , and i, j ∈ Z/bZ. Thus Xz , the fiber of X sm → over z = 0,

is the elliptic curve C× /z bZ and the fiber over z = 0 is a disjoint union of b copies of
C× , one appearing in each open set Wi .
Now assume that the section P meets the special fiber at w0 ∈ Ci× . Then we may
choose a small disk D around w0 in Ci× and shrink so that the image of
× D → Wi → X
contains the image of P over . We may then identify P with a function f : → D,

and the conditions on f for P to be tangent to a local Betti leaf are the same as they
are in the I1 case. Thus the rest of the argument is essentially identical to that in the
I1 case, and we will omit the rest of the details.
The case of Ib∗ reduction: Now consider the case where E has reduction of type Ib∗ at
t. Choose as usual a parameterization → C of a neighborhood of t and let X →
be the pull-back of E → C. Let ˜ → be a double cover ramified at 0 ∈ ˜ and let
˜ ˜ ˜
= \ {0}, so that → is an unramified double cover. Then it is well known
that X̃ , the pull-back of X → to ˜ has an extension to X̃ → ˜ whose fiber over 0
is of type I2b . Moreover, the section P of X → induces a section P̃ of X̃ → . ˜ We
apply the argument of the previous section to conclude that after shrinking , there ˜
are no points of ˜ over which P̃ is tangent to a local Betti leaf. Since X̃ → X is
étale, the same must be true after shrinking , i.e., P is not tangent to a local Betti
leaf over . (It is clear from the definition that local Betti leaves are preserved under
an étale base change.) This proves the desired discreteness near a point where E has
Ib∗ reduction.
The cases of I I , I I ∗ , I I I , I I I ∗ , I V , and I V ∗ reduction: Finally, consider the cases
where E has additive and potentially good reduction. Then by an argument parallel to
that of the previous case, we may focus attention on a disk near t, pull-back to a
ramified cover ˜ → of order 2, 3, 4, or 6, and reduce to the case of good reduction.
We leave the details as an exercise for the reader.
This completes the proof that the set of points t ∈ C 0 over which P is tangent
to a local Betti leaf is discrete and therefore finite, and it concludes the proof of
Theorem 2.5.

Remark 3.3 This proof gives no bounds on the cardinality of Ttor or TBetti , i.e., on
the number tangencies. In [32], we give an explicit upper bound on the number of
tangencies. In fact, by exploiting the global Betti foliation, we give an exact formula
for tangencies counted with multipicities and taking into account the behavior of P at
the bad fibers. The resulting upper bound depends only on topological properties of E.
Remark 3.4 A referee points out that case I0 of the argment above also follows directly
from Lemma 6.4 in [9].
4 Interlude on moduli of elliptic curves with a diﬀerential and a point
In this section, we discuss certain moduli spaces of elliptic curves with additional
structure. These spaces will be useful when we consider families of elliptic surfaces
in the following section. We work in more generality than needed in this paper, and
readers who are so inclined may replace the base ring R below with a field k of
characteristic = 2, 3 or even with C.
We begin by noting that there is a standard model for an elliptic curve E equipped
with a non-zero differential ω and a non-trivial point P: Given the data, choose a
Weierstrass model of E
y 2 + a1 x y + a3 y = x 3 + a2 x 2 + a4 x + a6
such that ω = d x /(2y + a1 x + a3 ). Then there is a unique change of coordinates
x = x + r , y = y + sx + t such that P has coordinates (x, y) = (0, 0) and a1 = 0.
Thus there is a unique triple (a2 , a3 , a4 ) such that E is the elliptic curve defined by
y 2 + a3 y = x 3 + a2 x 2 + a4 x,
the differential is ω = d x/(2y + a3 ), and the point is P = (0, 0).

We want to formalize this observation. Following Deligne [14], we say that a curve
of genus 1 over a base scheme S is a proper, flat, finitely presented morphism
π :W→S
whose geometric fibers are reduced and irreducible curves of arithmetic genus 1
equipped with a section O : S → W whose image is contained in the locus where π
is smooth.
Let R = Z[1/6] and consider the stack M over Spec R whose value on an R-scheme
S is the set of triples (W → S, ω, P) where W → S is a curve of genus 1 over S as
defined above, ω is a nowhere vanishing section of O ∗ (1W /S ), and P : S → W is a
section disjoint from O. Two such triples (W → S, ω, P) and (W → S, ω , P ) are
isomorphic if there exists an S-isomorphism W → W carrying ω to ω and P to P .
Proposition 4.1 The stack M is represented by the affine scheme Spec R[a2 , a3 , a4 ].
The universal object over M is the projective family of plane cubics W →
Spec R[a2 , a3 , a4 ] defined by
y 2 + a3 y = x 3 + a2 x 2 + a4 x
equipped with the differential ω = d x/(2y+a3 ) and the section P given by x = y = 0.

The substack of M where the curve W → S has smooth fibers is represented by the
open subscheme where = 0 and the substack where the fibers of W → S are
either smooth or nodal is represented by the open subscheme where either = 0 or
24 a22 − 24 3a4 = 0.
More formally, “the projective family of plane cubics defined by y 2 + a3 y =

x3 + a2 x 2 + a4 x” is defined as follows: Let R be the graded R[a2 , a3 , a4 ]-algebra
R = R[a2 , a3 , a4 ][x, y, z]/(y 2 z + a3 yz 2 − x 3 − a2 x 2 z − a4 x z 2 )
where x, y, and z have weight 1. Then W = ProjSpec R[a2 ,a3 ,a4 ] (R).
Here and later in the paper, whenever we have elements a2 , a3 , a4 in some ring, we
set
c4 (a2 , a3 , a4 ) = 16a22 − 48a4

= 24 a22 − 24 3a4
c6 (a2 , a3 , a4 ) = 288a2 a4 − 64a23 − 216a32
= 25 32 a2 a4 − 26 a23 − 23 33 a32
(a2 , a3 , a4 ) = −16a23 a32 + 16a22 a42 + 72a2 a32 a4 − 27a34 − 64a43
= −24 a23 a32 + 24 a22 a42 + 23 32 a2 a32 a4 − 33 a34 − 26 a43 .
We often omit the ai and simply write c4 , c6 , or . However, in the proof just below, we
do not omit the ai , i.e., we distinguish between the elements c4 , c6 generating a two-
variable polynomial ring R[c4 , c6 ] and the elements c4 (a2 , a3 , a4 ) and c6 (a2 , a4 , a6 )
in the ring R[a2 , a3 , a4 ].
Proof of Proposition 4.1 By [14, Prop. 2.5], the stack of pairs (W → S, ω) as above
is represented by the affine scheme Spec R[c4 , c6 ] with universal curve
c4 c6
y 2 = x 3 − x − 5 3
24 3 2 3
and universal differential d x /2y . (Deligne uses the more traditional coordinates
g2 = c4 /(22 3) and g3 = c6 /(23 33 ), but this is immaterial since 1/6 ∈ R.) Define a
morphism Spec R[a2 , a3 , a4 ] → Spec R[c4 , c6 ] by sending
c4 → c4 (a2 , a3 , a4 ) and c6 → c6 (a2 , a3 , a4 ).
Then pulling back the universal curve over Spec R[c4 , c6 ] to Spec R[a2 , a3 , a4 ] and
making the change of coordinates x = x + a2 /3, y = y + a3 /2 yields the curve and
differential mentioned in the statement of the theorem.
To finish the proof, one checks that the fibers of Spec R[a2 , a3 , a4 ] → Spec R[c4 , c6 ]
are the affine plane curves
c4 c6
y 2 = x 3 − 4
x − 5 3,
2 3 2 3
i.e., Spec R[a2 , a3 , a4 ] is the universal curve over Spec R[c4 , c6 ] minus its zero section.
Indeed, the fiber over (c4 , c6 ) is
c4 = 16a22 − 48a4
c6 = 288a2 a4 − 64a23 − 216a32 .
Eliminating a4 and dividing by 25 33 , we find
a32 a23 c4 a2 c6
2
= 3
− 4 2 − 5 3.
2 3 2 3 2 3
Thus setting a3 = 2y and a2 = 3x yields the stated fiber.

This means that to give a morphism to Spec R[a2 , a3 , a4 ] is to give a morphism
to Spec R[c4 , c6 ] (i.e., a family of curves and a differential) together with a non-zero
point in each fiber. This completes the proof that Spec R[a2 , a3 , a4 ] represents M.
The assertions about the locus where W has good or nodal fibers follows from [14,
Prop. 5.1], and this completes the proof of the proposition.

In light of the proposition, from now we change notation and let M be defined as
the scheme Spec R[a2 , a3 , a4 ]. Also, we write Msm for the locus where = 0 and
Mn for the locus where = 0 and c4 = 0. Similarly, let N = Spec R[c4 , c6 ], N sm
the locus where c43 − c62 = 0, and N n the locus where c43 − c62 = 0 and c4 = 0.
4.2 Torsion
Let π : W → M be the universal curve. Then the smooth locus of π is a commutative

group scheme over M and we may speak of points of finite order in the fibers. For each
n > 1, let M[n] be the locus where P has order dividing n, let M[n] be the locus where
P has order exactly n, and let Msm [n] = Msm ∩M[n] and Msm [n] = Msm ∩M[n] .
Let n > 1 and let k be a field of characteristic zero or prime to 6n. For R-schemes,
write − ⊗ k for the base change along the unique morphism Spec k → Spec R. Then
it follows from [15, I.6 and II.1.18-20] that M[n] ⊗ k is locally closed in M ⊗ k,
everywhere regular and of codimension 1, and that Msm [n] ⊗ k is a divisor in Msm
which is étale and finite of degree n 2 over N sm .
In fact, there are explicit recursive equations for divisors Dn ⊂ M such that
Msm [n] = Msm ∩Dn , namely the “division polynomials” evaluated at P [28, Ex. 3.7].
More precisely, for each n > 1, there is a homogenous polynomial ψn in a2 , a3 , a4
(where ai has weight i) of degree n 2 − 1 such that Dn is defined by ψn . We have
ψ2 = a3 ,
ψ3 = a2 a32 − a42 ,
ψ4 = 2a2 a33 a4 − 2a3 a43 − a35 ,
and the higher ψn are defined recursively by
ψ2m+1 = ψm+2 ψm3 − ψm−1 ψm+1

3
m ≥ 2,
ψ2 ψ2m = ψm−1
2
ψm ψm+2 − ψm−2 ψm ψm+1
2
m ≥ 3.
4.3 Nodal cubics with a point
Let k be a field of characteristic zero or p > 3, and let a = (a2 , a3 , a4 ) be a k-valued

point of Mn , i.e., such that (a) = 0 and c4 (a) = 0. Then by Proposition 4.1, the
plane cubic
E a : y 2 + a3 y = x 3 + a2 x 2 + a4 x
over k is nodal. We further assume that (a3 , a4 ) = (0, 0) so that P = (0, 0) and the
node, call it Q, are distinct. Let Gm be the multiplicative group over k. Then, possibly
after extending k quadratically, there is a group isomorphism
E a \ {Q} → Gm
which is unique up to pre-composing with inversion. We want to write down an explicit

expression for the image of P under such an isomorphism.
This is a straightforward calculation: The node is defined by the vanishing of 2y+a3
and 3x 2 + 2a2 x + a4 , and one finds that its coordinates are

18a32 − 8a2 a4 −a3
Q= ,
c4 2
where as usual c4 = 16a22 − 48a4 . Changing coordinates
18a32 − 8a2 a4 −a3

x = x + , y = y +
c4 2
brings E a into the form
−c6 2
y 2 = x 3 + x
4c4
where as usual c6 = 288a2 a4 −64a23 −216a32 . Letting γ be a square root of −c6 /(4c4 ),
the map to Gm is
y − γ x
(x , y ) →
y + γ x
and we find that P maps to
a3 c4 − γ (16a2 a4 − 36a32 )
(4.1)
a3 c4 + γ (16a2 a4 − 36a32 )
which (not surprisingly) is an algebraic expression in the original a2 , a3 , a4 .
5 From E/K to E → C
We remind the reader how to go from an elliptic curve over a function field to an elliptic
surface. Although this is not strictly necessary for our main purposes, it suggests a
fruitful point of view on finite-dimensional families of elliptic surfaces parameterized
by certain Riemann-Roch spaces.
5.1 General construction
Let k be a field of characteristic 0 or p > 3, let C be a smooth, projective, absolutely

irreducible curve over k, and let K = k(C). Let E be an elliptic curve over K equipped
with a non-zero rational point P ∈ E(K ).
Choose a non-zero differential ω on E. Then by Proposition 4.1, there is a unique
triple a = (a2 , a3 , a4 ) of elements of K such that E is isomorphic to
y 2 + a3 y = x 3 + a2 x 2 + a4 x,
P is (0, 0) and ω = d x/(2y + a3 ). Let D be the smallest divisor on C such that

div(ai ) + i D is effective for i = 2, 3, 4. (Here “smallest” is with respect to the usual
partial ordering: D1 ≥ D2 if D1 − D2 is effective.) Let L = OC (D) so that we may
regard ai as a global section of L ⊗i .
If U ⊂ C is a non-empty Zariski open subset and φ is a trivialization of L over U
(i.e., a nowhere vanishing section of L), then over U we may regard the ai as functions,
and we get a morphism U → M. Pulling back the universal curve gives a family
WU → U
of curves of genus 1 (in the sense used before Proposition 4.1) with a section PU
disjoint from O, and the general fiber of WU → U is E/K equipped with P. If {U j }
is an open cover with trivializations φ j of L |U j , there is a unique way to glue over
the intersections compatible with the identification of the generic fiber of WU j → U j
with E/K , and the result is a global family W → C of curves of genus 1 equipped
with a section which we again denote by P. Writing P for the P2 bundle over C given
by

P = PC L 2 ⊕ L 3 ⊕ OC
(with coordinates [x, y, z] on the fibers), we see that W is the closed subset of P
defined by the equation
y 2 + a3 y = x 3 + a2 x 2 + a4 x
and P is the section [0, 0, 1].

The surface W may have isolated singularities, and if so, we resolve them and then
blow down any remaining (−1)-curves in the fibers of the map to C, thus obtaining a
smooth, relatively minimal elliptic surface E → C with a section again denoted by P.
5.2 A geometric subtlety
There is a subtle point hiding in the last step of this construction: The section P of
W → C is disjoint from O, yet a section of E → C may very well meet O. Therefore,
there may be some blowing down in the last step to force such an intersection. We
make a few more comments about this situation and then give an example.
The underlying issue is that the local models WU → U are in a sense minimal
with respect to pairs “elliptic fibration + nowhere zero section,” but they may not be
minimal if we forget the section. We can quantify this as follows: Given E/K and P,
choosing ω leads to coefficients ai ∈ K and to invariants
c4 = 24 (a22 − 3a4 ) and c6 = 25 32 a2 a4 − 26 a23 .
Recall that D was defined as the smallest divisor on C such that div(ai ) + i D ≥ 0 for
i = 2, 3, 4. Similarly, let D be the smallest divisor on C such that div(c j ) + j D ≥ 0
for j = 2, 4. Then it is clear that D ≥ D and the points entering into D − D are
exactly those where the model W → C is not minimal (in the sense of [28, p. 816]).
Moreover, while W sits naturally as a divisor in

P = PC L 2 ⊕ L 3 ⊕ OC ,
the minimal Weierstrass family associated to W → C is naturally a divisor in

P = PC L 2 ⊕ L 3 ⊕ OC
where L = OC (D ). The choice of ω defines (possibly rational) sections of L and L

with divisors D and D respectively. Since O ∗ (1E /C ) = L , in some sense L is more
natural than L.
5.3 An example
Let C = P1 and K = k(t), and let E/K be defined by
w2 = z 3 + t 2 z − 1
with point P = (t −2 , t −3 ) and differential ω = dz/2w. The standard model coming

from Proposition 4.1 for this data is
y 2 + 2t −3 y = x 3 + 3t −2 x 2 + (3t −4 + t 2 )x
with P = (0, 0) and ω = d x/(2y + 2t −3 ). Also, c4 = −48t 2 and c6 = 864 and we

find that
D = 0 + ∞ and D = ∞.
The local model WA1 → A1 is given by
y 2 + 2y = x 3 + 3x 2 + (3 + t 6 )x.
The fiber over t = 0 is a cubic with cusp at t = 0, x = y = −1, and the surface
WA1 is singular at this point. Resolving the singularity requires blowing up once and
normalizing, and a further blow down removes a (−1)-curve in the fiber. This last
blow down brings the section P into contact with the zero section O.
5.4 Starting with the line bundle
We take the following point of view on constructing elliptic surfaces over C: Start with
a line bundle L on C. Then for each
a = (a2 , a3 , a4 ) ∈ H 0 (C, L 2 ⊕ L 3 ⊕ L 4 )
with (a2 , a3 , a4 ) = 0, we get W → C defined by the vanishing of
y 2 z + a3 yz 2 = x 3 + a2 x 2 z + a4 x z 2
in

P = PC L 2 ⊕ L 3 ⊕ OC .
For “most” choices of a, W → C is already minimal and L = L. This holds if

(a2 , a3 , a4 ) has order of vanishing < 12 (as a section of L 12 ) at each place of C.
If has only simple zeroes, then W → C is minimal and W is regular, so E = W.
In this way, we get flat families of elliptic surfaces parameterized by open subsets of
certain Riemann-Roch spaces. We will justify the claim “most” in the next section.
6 Very general elliptic surfaces with two sections
In this section, k is a field of characteristic zero or p > 3 and C is a smooth, projective,

absolutely irreducible curve over k. Let L be a line bundle on C which is globally
generated and write d for the degree of L.
Let a = (a2 , a3 , a4 ) be an element of V = H 0 (C, L 2 ⊕ L 3 ⊕ L 4 ) with (a) = 0.

Then as explained in Sect. 5.4 we get a family Wa → C of curves of genus 1 and a
relatively minimal elliptic surface Ea → C equipped with a section P. Our aim is to
show that for a very general choice of a, P is transverse to Ea [n] for all n and enjoys
other desirable properties.
We first consider the case where d = 0, so L is trivial and the ai are constants. In
this case, it is clear that P is transverse to all torsion sections if and only if it is disjoint
from all torsion sections, if and only if it is of infinite order. This happens for very
general choices of a, but not on a Zariski open. That suggests what to expect in the
general case.
We restate Theorem 1.7 (in the case where L is non-trivial) with an additional claim:
Theorem 6.1 Let L be a globally generated line bundle on C of degree d > 0, and set
V = H 0 (L 2 ⊕ L 3 ⊕ L 4 ).
Then for a very general a = (a2 , a3 , a4 ) ∈ V , the elliptic surface Ea → C associated

to
E: y 2 + a3 y = x 3 + a2 x 2 + a4 x
equipped with the section P = (0, 0) has the following properties:

(1) P has infinite order
(2) The singular fibers of Ea → C are nodal cubics (i.e., Kodaira type I1 ).
(4) If n is not a multiple of the characteristic of k, then P is transverse to Ea [n].
d(n 2 − 1) points.
Here, as usual, “for a very general a” means that there is a countable collection of
non-empty, Zariski open subsets of V such that if a lies in their intersection, then the
assertion holds for a. We will prove several lemmas, each asserting that some Zariski
open subset is non-empty, and then put them together to prove the theorem at the end
of this section. It is no loss of generality to assume that k is algebraically closed, so
for convenience we assume this for the rest of the section.
Recall that “L is globally generated” means that for all t ∈ C, there is a global
section of L not vanishing at t. It is a standard exercise to show that L is globally
generated and has positive degree if and only if there is a non-constant morphism
f : C → P1 such that L = f ∗ OP1 (1). Moreover, if L is globally generated, then the
set of global sections of L with reduced divisors (i.e., distinct zeroes) is non-empty
and Zariski open.
Lemma 6.2 The subset V ⊂ V consisting of a such that (a) has 12d distinct zeroes
(as a section of L 12 ) is Zariski open and not empty. There are a ∈ V whose zeroes
are disjoint from any given finite subset of points of C.
Proof It is clear that the locus of a ∈ V where (a) has distinct zeroes is Zariski open.
To prove the lemma, we need to check that V is not empty. We do this constructively.
First assume C = P1 and L = OP1 (1). Set a2 = 0, a3 = c ∈ k, and a4 = t 4 . Then
= −27c4 − 64t 12 which has distinct zeroes as a section of OP1 (12) if c = 0.
Moreover, varying c, we can arrange for the zeros to avoid any finite subset of P1 .
In the general case, choose a morphism f : C → P1 such that L = f ∗ (OP1 (1)).
Let S ⊂ P1 be the branch locus of f . Then setting a2 = 0, a3 = c, and a4 = f ∗ (t 4 ),
where c is chosen so that the zeroes of −27c4 − 64t 12 are disjoint from S, yields an
explicit a with the required properties. Varying c allows us to avoid any finite subset
of C.

As noted in Sect. 5.4, if a ∈ V , then the corresponding elliptic surface Wa is

smooth (so no resolution of singularities is needed), Wa → C is relatively minimal
(so we may set Ea = Wa ), and L = O ∗ (1Ea /C ). Moreover, the bad fibers of Ea → C
are all of type I1 . From now on we always choose a from V .
Lemma 6.3 For every n ≥ 1, there is a non-empty, Zariski open subset Vn of V such
that if a ∈ Vn , then the section P of Ea → C does not intersect any singular fiber in
a point of order exactly n.
Proof It is clear that the locus of a where P has the stated property is open, and our task
is to show it is non-empty. Since the bad fibers are all of type I1 , if k has characteristic
p > 0 and n is divisible by p, there are no points of order exactly n in the fiber, so we
may take Vn = V .
Now assume that n is not divisible by the characteristic of k. We check constructively
that there is a non-empty set as described in the statement. As in the previous lemma,
we may reduce to the case C = P1 and L = OP1 (1). Take a2 = 0, a3 = c, a4 = t 4 .
Then the bad fibers are at the roots of t 12 = (−27/64)c4 and at each such root,
the coordinate in Gm of P was given at (4.1). For the data we are considering, the
coordinate is
4t 4 − 3cγ
where γ = (−9c3 /2)1/2 t −2 .
4t 4 + 3cγ
Then for each n, there are only finitely many values of c such that for some root t of
t 12 = (−27/64)c4 , the displayed quantity is an n-th root of unity. This proves that Vn
is non-empty for each n.

Remark 6.4 Over an uncountable field, intersecting the opens in the theorem gives a
non-empty set. We can do a bit better over C: There is an everywhere dense classical
open set in V such that P meets each singular fiber away from the unit circle S 1 ⊂ C× .
Lemma 6.5 If the characteristic of k is p > 3, then for all a ∈ V and any n divisible
by p, P does not have order exactly n.
Proof It will suffice to show that when a ∈ V , Ea → C has no non-trivial p-torsion

sections. First note that since a ∈ V , the zeroes of c4 are disjoint from those of .
This implies that j = c43 / has simple poles, so it is not a constant (implying that
Ea → C is non-isotrivial) and not a p-th power. Then [30, Prop I.7.3] implies that
Ea → C has no p-torsion. (In [30], the ground field is finite, but the argument there
works over any field of positive characteristic.)

Proposition 6.6 For every n not divisible by the characteristic of k, there is a non-
empty, Zariski open subset Wn ⊂ V such that if a ∈ Wn , then n P = 0 and P is
transverse to Ea [n].
Proof Again, it is clear that the set of a with the desired properties is open. Unfortu-
nately, it seems hopeless to give a constructive proof that it is non-empty, so we have
to do something more sophisticated.
Recall the moduli space M of Sect. 4. We write Mk for
M ⊗Z[1/6] k = Spec k[a2 , a3 , a4 ]
and Mk [n] for the locally closed, smooth, codimension 1 locus parameterizing triples
(E, ω, P) where P has order n.
Recall also that V = H 0 (C, L 2 ⊕ L 3 ⊕ L 4 ) and V is the open subset consisting of
a such that (a) has distinct zeroes. Choose an open subset U ⊂ C and a trivialization
of L over U . Then for a = (a2 , a3 , a4 ) the ai may be regarded as functions on U ,
and we get a morphism f a : U → Mk . To say that n P = O is to say that f a (U ) is
contained in Mk [n]. To say that P is tangent to Ea [n] over x ∈ U is to say that f a (U )
is tangent to Mk [n] at f a (x). We will show that these conditions do not hold for most
a.
Consider the morphism
F : V × U → Mk (a, t) → F(a, t) := f a (t)
and let
Dn := F −1 (Mk [n]) ∩ (V × U ) .
We will use the global generation of L to show that Dn is a smooth, locally closed subset
of codimension 1 in V × U , and that there is a non-empty open subset WU ,n ⊂ V
such that the projection Dn → V is étale over WU ,n . This means that if a ∈ WU ,n ,
then {a} × U is transverse to Dn , i.e., that P meets the n-torsion multisection of Ea
transversally over U . Taking a finite cover {U j } of C and setting Wn = ∩ j WU j ,n will
complete the proof.
Since L is globally generated, so are its powers L i for i = 2, 3, 4. This means that
for every t ∈ U , there are global sections a2 , a3 , a4 not vanishing at t, and for all but
finitely many t there are global sections s2 , s3 , s4 which vanish to order 1 at t. (Since
L i is globally generated, there are sections of L i inducing a morphism C → P1 . If t
is not in the ramification locus, a section si as above can be obtained by pulling back
a section of OP1 (1) vanishing simply at the point of P1 under t.)
For each t ∈ U , the restriction
Ft : V × {t} → Mk
is a linear map, and since L is globally generated, it is surjective. Thus the fibers are
all affine spaces of dimension h − 3 where h = dim V . Therefore, F is surjective and
smooth (smooth because it is submersive, i.e., it has a surjective differential at every
point). Moreover, the fibers of F are Ah−3 -bundles over U , and in particular, they are
all irreducible of dimension h − 2. It follows that each irreducible component Dn,i of
Dn is smooth and locally closed in V × U of codimension 1 and has the form
Dn,i = F −1 (Mk [n]i ) ∩ (V × U )
where Mk [n]i is an irreducible component of Mk [n].

Consider an irreducible component Dn,i of Dn . We are going to produce a point
of Dn,i at which the projection Dn,i → V is étale. Start by choosing any point
(a, t) ∈ Dn,i and let m = F(a, t). The fiber of F over m is an Ah−3 bundle over U ,
and F −1 (m) ∩ (V × U ) is a non-empty open subset of this bundle, so it projects to
a non-empty open subset of U . This means that we may find another point (a , t ) =
(a2 , a3 , a4 , t ) in Dn,i such that L i admits global sections si vanishing simply at t for
i = 2, 3, 4.
For all triples (α2 , α3 , α4 ) ∈ k 3 , we have
F(a2 + α2 s2 , a3 + α3 s3 , a4 + α4 s4 , t) = m,
and for a non-empty open subset of triples (αi ) ∈ k 3 , we have that
a := (a2 + α2 s2 , a3 + α3 s3 , a4 + α4 s4 ) ∈ V .
By a suitable choice of the αi we may arrange for the differential of F restricted to

{a } × U to carry the tangent space of U at t to a line in the tangent space of Mk at m
not contained in TM[n],m . For such a choice, we conclude that {a } × U is transverse
to Dn,i at (a , t). This proves that the projection Dn,i → V is étale at (a , t).
n,i such that Dn,i → V
It follows that there is a Zariski open subset Dn,i o of D o
is étale. The image of Dn,i \ Dn,i in V is contained in a proper closed subset, and
o
removing these subsets for all i yields an open subset WU ,n over which Dn → V is
étale. Covering C with finitely many U j and setting Wn = ∩ j WU j ,n yields an open
subset of V such that if a ∈ Wn , then P does not have order n and is transverse to
Ea [n]. This completes the proof of the proposition.

Proof of Theorem 6.1 Consider the intersection

⎛ ⎞ ⎛ ⎞

V = ⎝ Vn ⎠ ⎝ Wn ⎠ ⊂ V .
n≥1 p | n
The preceding lemmas show that if a ∈ V , then the corresponding Ea has the proper-
ties asserted in the Theorem. Indeed, since a ∈ V , (a) as 12d distinct zeroes, and so
Ea has 12d bad fibers of type I1 and no other bad fibers. Since a ∈ ∩n Vn , Lemma 6.3
shows that P does not meet a bad fiber in a torsion point. Since a ∈ ∩n Wn , Lemma 6.5
and Proposition 6.6 show that P has infinite order, and Proposition 6.6 shows that if
n is prime to the characteristic, then P is transverse to Ea [n]. This establishes points
(1) through (4) of the Theorem.
The transversality in point (5) is equivalent to that in (4), so to finish we just need
to calculate the intersection multiplicity (n P).O. For this, we first note that P.O = 0
by construction, and as explained in the proof of Lemma 2.8, O 2 = P 2 = −d. Thus
ht(P) = 2d and Lemma 2.8 implies that (n P).O = d(n 2 − 1), as required.
This completes the proof of the theorem.

Remark 6.7 When C = P1 , every line bundle of non-negative degree is globally gener-
ated. Thus, starting from data (E, P) over P1 , we can find a deformation (E , P ) with
the same base C and bundle L such that P is transverse to all torsion multisections.
For a general C, if we do not assume any positivity for L = O ∗ (1E /C ), it may be
impossible to produce deformations with fixed C and L. Here are two alternatives:
First, we may embed L → L where L is globally generated, and deform a non-
minimal model of E (lying in PC (L 2 ⊕ L 3 ⊕ OC )). Second, it seems likely that the
ideas of Moishezon [23], as explained in [16, Thm. I.4.8] would allow one to find a
deformation of E where the base curve is also allowed to vary (i.e., deform to E → C
and section P ) with the desired transversality.
Remark 6.8 Suppose that k has characteristic zero and that π : E → C and P satisfy
the conclusions of Theorem 6.1. If n 1 and n 2 are two distinct integers, then n 1 P ∪ n 2 P
is a normal crossings divisor on E. More generally, if N ⊂ Z is a non-empty finite set,
then

D= nP
n∈N
is a curve on E with only ordinary multiple points. Indeed, it is a union of smooth

components which meet pairwise transversally. This is clear from the facts that O and
n P meet transversally for all n = 0 and that O ∪ (n 2 − n 1 )P is carried isomorphically
to n 1 P ∪ n 2 P under translation by n 1 P.
7 Explicit examples with even height over small ﬁelds
In this section, we show by explicit construction that there are pairs (E, P) with P
transverse to torsion multisections over fields k such as number fields and global func-
tion fields. The precise statement is Theorem 1.8 in the introduction. For simplicity,
we assume throughout that the characteristic of k is not 2. We begin by constructing
examples of height 2 over P1 .
Proposition 7.1 Let k be a field of characteristic = 2. Then there exist Jacobian elliptic
surfaces E → P1 over k equipped with a section P such that
(2) The singular fibers of E → P1 are of Kodaira type I0∗ .

⎧ 2
⎪
⎪ n −1
⎨ if n is odd,
2
⎪
⎩ n − 4 if n is even,
2
⎪
2
points.
(5) The height of E is 2, i.e., O ∗ (1E /P1 ) ∼
= OP1 (2).
Proof We will construct one such E → P1 for every elliptic curve E over k. Suppose
that f ∈ k[x] is a monic polynomial of degree 3 such that E is defined by y 2 = f (x).
Form the product E ×k E, and let {±1} ⊂ Aut(E) act diagonally. The quotient
(E ×k E)/(±1) is a singular (Kummer) surface, and projection to the first factor
induces a morphism
(E ×k E)/(±1) → E/(±1) ∼
= P1 .
Let E → P1 be the regular minimal model of (E ×k E)/(±1) → P1 . Thus E is

obtained from (E ×k E)/(±1) by blowing up the 16 fixed points of ±1 on E ×k E,
and the bad fibers of E → P1 are of type I0∗ and lie over t = ∞ and the roots of f (t).
Let n ⊂ E ×k E be the graph of multiplication by n, which we may regard as
the image of a section to E ×k E → E. Then n is preserved by ±1 and maps with
degree 2 to a section of (E ×k E)/(±1) → P1 , and this section lifts to a section n P
of E → P1 . (The notation is consistent in that n P is n times P in the group law of E.)
The following diagram summarizes the data:
E ×k E (E ×k E)/(±1) E
n nP
E E/(±1) ∼
= P1 P1 .
It will be convenient to have a Weierstrass equation for E → P1 . If f (x) =

x3 + ax 2 + bx + c, then E is the Néron model of the elliptic curve
y 2 = x 3 + a f (t)x 2 + b f 2 (t)x + c f 3 (t)
over k(t), and the point P has coordinates (x, y) = (t f (t), f 2 (t)). Indeed, if the two
factors of E × E are v 2 = f (u) and s 2 = f (r ), then the field of invariants of ±1 is
generated by u, r , and z = vs, and these satisfy the equation
z 2 = f (u) f (r ).
Setting u = t and z = y/ f (u), and r = x/ f (u) yields the equation and point above.
We now verify the cases n = 1 and n = 2 of the proposition. Since P has polynomial
coefficients, it does not meet O over any finite value of t, and since its x and y
coordinates have degrees 4 and 6, and E has height 2, P also does not meet O over
t = ∞. In summary, P meets O nowhere, as claimed. For later use, we note that at the
roots of f , P specializes to (0, 0), i.e., to a singular point of the fiber of (E ×k E)/(±1),
so P lands on a non-identity component of the fiber of E. At t = ∞, P specializes to
(1, 1), a non-singular, finite point of the fiber (i.e., a point not on O).
A tedious but straightforward calculation (or an algebra package …) shows that 2P
has coordinates ((1/4)t 4 + · · · , (1/8)t 6 + · · · ) where · · · indicates terms of lower
degree in t. The argument of the previous paragraph shows that 2P meets O nowhere,
as claimed. For later use, we note that 2P passes through a finite point of the identity
component in each of the bad fibers.
Now consider n > 2. It is clear that n meets E × {0} exactly at the points of E of
order n, and each of these intersections is transverse. If ( p, 0) is such a point which is
not of order 2, then the quotient map
E ×k E → (E ×k E)/(±1)
is étale in a neighborhood of ( p, 0) and it sends n 2-to-1 to a curve that meets O

transversally. Moreover, the map
E → (E ×k E)/(±1)
is an isomorphism in a neighborhood of such a point. This proves that n P meets O

transversally over the values of t such that there is a point (t, v) with v 2 = f (t) which
is n-torsion and not 2-torsion. There are
⎧ 2
⎨ n −1 if n is odd
2
⎩ n 2 −4 if n is even
2
such values of t.
It remains to consider what happens over the roots of f (t) and t = ∞. But we
checked above that P meets a non-trivial point of the identity component at t = ∞
and such a point is either of infinite order or of order p when k has characteristic p. So,
for n prime to the characteristic of k, n P does not meet O over t = ∞. Similarly, over
the roots of f (t), P passes through the non-identity component and 2P passes through
a non-trivial point of the identity component, so n P does not meet O when n is prime
to the characteristic. We have thus identified all points where n P and O intersect,
the intersections are transverse, and their number is as stated in the proposition. This
completes the proof of the proposition.

Remark 7.2 As a check, we compute the intersection number (n P).O using heights
as in Lemma 2.8. We have O 2 = P 2 = −2 and P.O = 0. Since P passes through
a non-identity component of the fibers over roots of f (t) and through the identity
component at t = ∞, the “correction term” is −C P .(P − O) = −3. (See table 1.19
in [13].) Using the formula (2.4) for the height pairing yields ht(P) = 1.
Similarly, for any odd n, −Cn P .(n P − O) = −3 and using that ht(n P) = n 2 and
calculating as in Lemma 2.8 we find (n P).O = (n 2 − 1)/2.
On the other hand, for even n, n P passes through the identity component in all bad
fibers, so −Cn P .(n P − O) = 0 and we find that (n P).O = (n 2 − 4)/2.
This confirms that the intersections we saw above are all transverse.
Proof of Theorem 1.8 Proposition 7.1 implies the case of the Theorem where C = P1
and L = OP1 (2), and we get infinitely many examples because k is infinite. Indeed,
for each j ∈ k, there is an elliptic curve E with j-invariant j, and elliptic curves with
distinct j-invariants give rise to non-isomorphic E → P1 since the non-singular fibers
are twists of the chosen E.
We deduce the general case by a pull-back construction. Write E → P1 for one of
the surfaces constructed in Proposition 7.1. Let f : C → P1 be a non-constant mor-
phism defined by sections of the globally generated line bundle F, so F = f ∗ OP1 (1)
and L = f ∗ OP1 (2). The conclusions of the theorem will hold for E := f ∗ E → C if
the branch locus of f is disjoint from the set of points of P1 over which E has bad
reduction or n P meets O. From the construction of E , we see that the set to be avoided
is precisely the set of x coordinates of torsion points of the elliptic curve y 2 = f (x)
used to construct E . Although this set is infinite, we will see that it is sparse in k.
We divide into two cases according to the characteristic of k, starting with the case
of characteristic zero. Choose an elliptic curve E over Q, and an auxiliary prime such
that equations defining E are -integral and E has good reduction modulo . Then
[28, VIII.7.1] implies that the x-coordinate of a torsion point Q (defined over some
number field K and taken with respect to an -integral model) is “almost integral,” i.e.,
it satisfies 2 x(Q) is integral at all primes of K over . Construct E → P1Q using E as
in Proposition 7.1. Then choose any non-constant morphism f : C → P1k defined by
sections of F. Composing φ with a linear fractional transformation, we may arrange
that the branch locus of f consists of points with finite, non-zero coordinates, and
that any of those coordinates which lie in a number field have large denominators at
primes over . They are thus distinct from the x-coordinates of torsion points of E,
and E = f ∗ E satisfies the requirements of the theorem.
When k has characteristic p > 2, the argument is similar, but simpler: Choose
an embedding F p (t) → k, an elliptic curve E over F p (t), and a place v of F p (t)
where E has good reduction. Then by [29, §4], the coordinates of any torsion point
Q of E (defined over some algebraic extension K of F p (t) and taken with respect
to an integral model) are integral at places of K over v. Use E to construct E as in
Proposition 7.1. Then choose any non-constant morphism f : C → P1k defined by
sections of F. Composing φ with a linear fractional transformation, we may arrange
that the branch locus of f consists of points with finite, non-zero coordinates, and that
any of those coordinates which are algebraic over F p (t) are not integral at places over
v. They are thus distinct from the x-coordinates of torsion points of E, and E = f ∗ E
satisfies the requirements of the theorem.

Remark 7.3 It seems likely that when k is a number field or a global function field, the
construction in Proposition 7.1 gives rise to elliptic divisibility sequences Dn whose
“new parts” Dn are often irreducible, i.e., prime divisors.
8 Application to geography of surfaces
In this section, we will prove Theorem 1.9. Let k = C, C = P1 , and L = OP1 (d) where
d = g +1, which by assumption satisfies d ≥ 1. Theorem 1.7 guarantees the existence
of an elliptic surface π : E → P1 of height d (i.e., such that O ∗ (1E /P1 ) = L) with a
section P such that for all n, n P meets O transversally in d(n 2 − 1) points. Moreover,
π has irreducible fibers. Let F be the class of a fiber of π . We have O 2 = P 2 = −d,
F 2 = 0, and the canonical divisor of E is
K E = (d − 2)F.
Thus the geometric genus of E is d − 1 = g.

Fix an integer n > 1. Later in the proof, we will need to assume that n is sufficiently
large. Let h : Y → E be the result of blowing up all but one of the points of intersection
of O and n P, let E i (i = 1, . . . , d(n 2 − 1) − 1) be the exceptional divisors, and let
C j be the strict transform the section j P.
Write F̃ for the strict transform of a general fiber of π in Y . We have

C02 = Cn2 = −dn 2 + 1, C0 .Cn = 1, and K Y = (d − 2) F̃ + Ei .
i
It is a simple exercise to check that the intersection pairing on Y is negative definite

on the lattice spanned by C0 and Cn , and that pa (Z ) ≤ 0 for all effective divisors
supported on C0 ∪ Cn . Thus by Artin’s contractibility theorem [6, Thm. 2.3] or [7,
Thm 3.9], we may contract C0 ∪ Cn . In other words, there is a proper birational
morphism f : Y → X where X is a normal, projective surface, f (C0 ∪ Cn ) = {x},
and f induces an isomorphism
Y \ (C0 ∪ Cn ) ∼
= X \ {x}.
Proof that X satisfies the conditions of Theorem 1.9

We have already observed that X is normal and projective. Since the geometric genus
is a birational invariant, and E has geometric genus g, so does X .
It is evident that X has exactly one singular point, namely x, and the minimal
resolution of x is the union of two smooth rational curves (C0 and Cn ) meeting at one
point and having self-intersection −a := −dn 2 + 1. Such a singularity is analytically
equivalent to a cyclic quotient singularity of type 1/(a 2 − 1)(1, a), as one sees by
considering the Hirzebruch-Jung continued fraction
1 a2 − 1
a− = .
a a
(See [8, Ch. 3].) In particular, it follows that X is Q-Gorenstein and K X is Q-Cartier.
We next compute the discrepancy of x (as defined for example in [19]) and verify
that x is log-terminal. Since C j is smooth and rational with self-intersection −a, we
have C j .K Y = a − 2. Define coefficients α0 , αn ∈ Q by
K Y = f ∗ K X + α0 C0 + αn Cn
(an equality in Pic(Y ) ⊗ Q). Then
0 = ( f ∗ C0 ).K X
= C0 . f ∗ K Y
= (a − 2) + α0 a − αn
and similarly,
0 = (a − 2) − α0 + αn a.
We find that
a−2
α0 = αn = − > −1.
a−1
This confirms that x is a log-terminal singularity, and we have
a−2
f ∗ K X = (d − 2) F̃ + Ei + (C0 + Cn ) .
a−1
i
Next, write b := (a − 2)/(a − 1) and compute
K X2 = ( f ∗ K X )2
2

= (d − 2) F̃ + E i + b(C0 + Cn )
i
= dn 2 (4b − 2b2 − 1) + d + 1 + 4b2 − 12b.
As n → ∞, a → ∞ and b → 1, so K X2 grows like dn 2 and in particular is unbounded

as n varies.
To finish the proof, it remains to check that K X is ample, which we do using the
Nakai-Moishezon criterion [7, Thm. 1.22]. We have already seen that K X2 > 0, so it
will suffice to check that for every irreducible curve C on X , C.K X > 0. For any such
curve C,
f ∗ C = D + m 0 C0 + m n Cn
where D is an irreducible curve not equal to C0 or Cn , and m j ≥ 0 for j = 0, n. It

will thus suffice to prove that D. f ∗ K X > 0 for all irreducible curves not equal to C0
or Cn and that C j . f ∗ K X = 0 for j = 0, n. For the latter assertion, one computes that
C j . f ∗ K X = f ∗ (C j ).K X = 0
for j = 0, n.
For the former assertion, we make a case by case analysis of the possibilities for
D. They are:
• the strict transform F̃ of a general fiber of π , for which we have
F̃. f ∗ K X = 2(a − 2)/(a − 1) > 0;
• one of the exceptional curves E i , for which we have
E i . f ∗ K X = −1 + 2(a − 2)/(a − 1),
which is > 0 if d > 1 or n > 2;

• the strict transform G̃ i = F̃ − E i of a fiber of π passing though an intersection
point of O and n P, for which we have G̃ i . f ∗ K X = 1;
• and the strict transform Q̃ of a multisection Q of π not equal to O or n P. Let e be
the degree of π|Q : Q → P1 , assume that n > 2 so that b > 1/2, and recall that

f ∗ K X = (d − 2) F̃ + E i + b (C0 + Cn ) .
i
If d > 2, we have
Q̃. f ∗ K X ≥ (d − 2)e > 0.

If Q̃. i E i > e or Q̃.C0 > 2e, then again it is clear that Q̃. f ∗ K X > 0 as

required. To finish, assume that d ≤ 2, Q̃. i E i ≤ e, and Q̃.C0 ≤ 2e. Applying
h ∗ to the equality in Lemma 2.9 implies that

Cn = nC1 + (1 − n)C0 − n E i + d(n 2 − n) F̃
i
in NS(Y ). If Q = P, we find that
Q̃.Cn ≥ (1 − n)2e − ne + d(n 2 − n)e,
which is > e for all n ≥ 4, and this shows that Q̃. f ∗ K X > 0. If Q = P, then
Q̃.Cn ≥ −nd − n + d(n 2 − n),
and we find that Q̃. f ∗ K X > 0 for all n ≥ 5. (When Q = P,we can also calculate
directly that Q̃. f ∗ K X = d − 2 + b(dn 2 − 2dn) which goes to infinity with n.)
This completes the check that Q̃. f ∗ K X > 0 for all irreducible multisections Q̃
not equal to C0 or Cn .
The itemized list completes the verification that K X is ample, and this finishes the
proof of the theorem.

The first-named author thanks Seoyoung Kim, Nicole Looper, and Joe Silverman for
helpful conversations at the 2019 AMS Mathematics Research Community meeting
in Whispering Pines, Rhode Island, and for pointing out [17] and its antecedents. He
also thanks the Simons Foundation for partial support in the form of Collaboration
Grant 359573. The second-named author thanks FONDECYT for support from grant
1190066. Both authors thank Matthias Schütt for comments and corrections, and they
thank Pietro Corvaja, Brian Lawrence, and Umberto Zannier for their comments on an
earlier version of this paper and their pointers to related literature, notably the preprint
[11].
References
1. Alekseev, V.A., Liu, V.: On accumulation points of volumes of log surfaces. Izv. Ross. Akad. Nauk
Ser. Mat. 83, 5–25 (2019)
2. Alexeev, V., Liu, W.: Log surfaces of Picard rank one from four lines in the plane. Eur. J. Math. 5,
622–639 (2019)
3. Alexeev, V., Liu, W.: Open surfaces of small volume. Algebr. Geom. 6, 312–327 (2019)
4. Alexeev, V.: Boundedness and K 2 for log surfaces. Int. J. Math. 5, 779–810 (1994)
5. Alexeev, V., Mori, S.: Bounding singular surfaces of general type, Algebra, arithmetic and geometry
with applications (West Lafayette, IN, 2000), 2004, pp 143–174
6. Artin, M.: Some numerical criteria for contractability of curves on algebraic surfaces. Am. J. Math.
84, 485–496 (1962)
7. Bădescu, L.: Algebraic surfaces, Universitext. Springer, New York (2001)
8. Barth, W.P., Hulek, K., Peters, C.A.M., Van de Ven, A.: Compact complex surfaces, Second, Ergebnisse
der Mathematik und ihrer Grenzgebiete. 3. Folge. A Series of Modern Surveys in Mathematics, vol.
4, Springer, Berlin (2004)
9. Bedford, E., Lyubich, M., Smillie, J.: Polynomial diffeomorphisms of C 2 . IV. The measure of maximal
entropy and laminar currents. Invent. Math. 112, 77–125 (1993)
10. Bierstone, E., Milman, P.D.: Semianalytic and subanalytic sets. Inst. Hautes Études Sci. Publ. Math.
67, 5–42 (1988)
11. Corvaja, P., Demeio, J., Masser, D., Zannier, U.: On the torsion values for sections of an elliptic scheme,
2019. Preprint arxiv:1909.01253
12. Corvaja, P., Masser, D., Zannier, U.: Torsion hypersurfaces on abelian schemes and Betti coordinates.
Math. Ann. 371, 1013–1045 (2018)
13. Cox, D.A., Zucker, S.: Intersection numbers of sections of elliptic surfaces. Invent. Math. 53, 1–44
(1979)
14. Deligne, P.: Courbes elliptiques: formulaire d’après J. Tate, Modular functions of one variable, IV
(Proc. Internat. Summer School, Univ. Antwerp, Antwerp, 1972), 1975, pp. 53–73. Lecture Notes
inMath., Vol. 476
15. Deligne, P., Rapoport, M.: Les schémas de modules de courbes elliptiques (1973), 143–316. Lecture
Notes in Math., Vol. 349
16. Friedman, R., Morgan, J.W.: Smooth Four-Manifolds and Complex Surfaces, Ergebnisse derMathe-
matik und ihrer Grenzgebiete (3) [Results in Mathematics and Related Areas (3)], vol. 27. Springer,
Berlin (1994)
17. Ghioca, D., Hsia, L.-C., Tucker, T.: A variant of a theorem by Ailon-Rudnick for elliptic curves. Pacific
J. Math. 295, 1–15 (2018)
18. Ingram, P., Mahé, V., Silverman, J.H., Stange, K.E., Streng, M.: Algebraic divisibility sequences over
function fields. J. Aust. Math. Soc. 92, 99–126 (2012)
19. Kawamata, Y., Matsuda, K., Matsuki, K.: Introduction to the minimal model problem. Algeb. Geomet.
Sendai 1987, 283–360 (1985)
20. Kodaira, K.: On compact analytic surfaces II. Ann. Math. 77, 563–626 (1963)
21. Kollár, J., Shepherd-Barron, N.I.: Threefolds and deformations of surface singularities. Invent. Math.
91, 299–338 (1988)
22. Liu, W.: The minimal volume of log surfaces of general type with positive geometric genus. Preprint
arxiv:1706.03716
23. Moishezon, B.: Complex surfaces and connected sums of complex projective planes, Lecture Notes
inMathematics, Vol. 603, Springer, Berlin, 1977.With an appendix by R. Livne
24. Łojasiewicz, S.: Ensembles semi-analytiques, 1964. IHES Preprint
25. Shioda, T.: On the Mordell-Weil lattices. Comment. Math. Univ. St. Paul. 39, 211–240 (1990)
26. Shioda, T.: Mordell-Weil lattices for higher genus fibration over a curve, New trends in algebraic
geometry (Warwick, 1996), 1999, pp. 359–373
27. Silverman, J.H.: Generalized greatest common divisors, divisibility sequences, and Vojtas conjecture
for blowups. Monatsh. Math. 145, 333–350 (2005)
28. Silverman, J.H.: The Arithmetic of Elliptic Curves, Second, Graduate Texts in Mathematics, vol. 106.
Springer, Dordrecht (2009)
29. Tate, J.T.: Algorithm for determining the type of a singular fiber in an elliptic pencil, Modular functions
of one variable, iv (Proc. Internat. Summer School, Univ. Antwerp, Antwerp, 1972), 1975, pp. 33–52.
Lecture Notes in Math., Vol. 476
30. Ulmer, D.: Elliptic curves over function fields, Arithmetic of L-functions (Park City, UT, 2009), 2011,
pp. 211–280
31. Ulmer, D.: On Mordell-Weil groups of Jacobians over function fields. J. Inst. Math. Jussieu 12, 1–29
(2013)
32. Ulmer, D., Urzúa, G.: Bounding tangencies of sections on elliptic surfaces. Int. Math. Res. Not. IMRN
6, 4768–4802 (2021)
33. Zannier, U.: Some problems of unlikely intersections in arithmetic and geometry, Annals of Mathe-
matics Studies, vol. 181, Princeton University Press, Princeton, NJ, 2012.With appendixes by David
Masser
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.

Elliptic Surface of Kodaira Dimension 1

Uploaded by

Copyright:

Available Formats

Elliptic Surface of Kodaira Dimension 1

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Elliptic Surface of Kodaira Dimension 1

Uploaded by

Copyright:

Available Formats

Selecta Mathematica (2022) 28:25 Selecta Mathematica

https://doi.org/10.1007/s00029-021-00747-x New Series

Transversality of sections on elliptic surfaces with

Douglas Ulmer1 · Giancarlo Urzúa2

Accepted: 18 November 2021

Keywords Elliptic surfaces · Unlikely intersections · Elliptic divisibility sequences ·

Mathematics Subject Classification Primary 14J27; Secondary 11B39 · 14J29

1 Department of Mathematics, University of Arizona, Tucson, AZ 85721, USA

We say that a divisor on C is reduced if it has the form

Theorem 1.7 Suppose k is a field of characteristic zero or of characteristic p > 3,

Then for a very general a = (a2 , a3 , a4 ) ∈ V , the elliptic surface E → C associated

(1) P has infinite order.

Theorem 1.8 Let k be a field of characteristic 0 or a field of characteristic p > 2 which

(1) P has infinite order.

(4) If n is not a multiple of the characteristic of k, then n P meets O transversally in

1.10 Plan of the paper

In Sect. 2 we present foundational material on torsion points and intersections on

2 Preliminaries on torsion and intersections

In this section we gather various foundational results on torsion, intersections, heights,

E[n] = [n]−1 (O),

E[n] = ∪m|n E[m] (2.1)

2.3 Divisibility sequences

In the introduction, we defined divisors Dn for n ≥ 1 by

In this section we examine alternative definitions and properties of these divisors

where the sum is over positive divisors of n.

non-reduced if and only if Dn is non-reduced for all multiples n of m. Thus point

Now consider the “incidence correspondence”

I := (t, m) m > 0 and P is tangent to E[m] over t ⊂ C × Z>0 .

P, Q := −(P − O + C P ).(Q − O) (2.4)

where the dot refers to the intersection number on E. If E → C is non-constant (i.e.,

Lemma 2.8 For E and P as above,

If π has irreducible fibers and d = deg O ∗ (1E /C ), then

In the first display, O(1) depends only on E and P, not on n.

n 2 ht(P) = ht(n P) = −(n P − O + Cn P ).(n P − O),

[n P] = n[P] + (1 − n)[O] + d(n 2 − n)[F]

[n P] = n[P] + (1 − n)[O] + c[F]

3 Proof of Theorem 2.5

(a, b)(z, w) = (z, w + aτ (z) + b) .

For z ∈ , corresponding to u ∈ U , P(u) corresponds to [w](z), which is the class

( × C)/(Zτ + Z) × (R/Z)2 (3.2)

where the upper horizontal map is

(z, w) = (z, r τ (z) + s) ← (z, r , s).

Claim 3.2 For every t ∈ C, there is a classical open neighborhood Ut of t in C such

|∇φi (z)| ≥ C |φi (z) − φi (0)|θ

where q is the node of the cubic.

where the action of Z on  × C× is

m · (z, w) = (z, z m w).

Moreover, as explained starting in the last paragraph of [20, p. 597], there is a

such that {0} × C× maps biholomorphically to the complement of q in the special

  f (z) log | f (z)|

We will show that there is no holomorphic function satisfying such equalities.

B1 |z| N < |g(z)| < B2 |z| N (3.5)

C1 |z| N < |h(z)| < C2 |z| N

for some positive constants C1 and C2 and all z ∈  . We have

log |1 + h(z)| ≤ log(1 + |h(z)|) ≤ |h(z)| < C2 |z| N

for z ∈  . If |w0 | = 1 (so log |w0 | = 0), we have

log | f (z)| log |1 + h(z)| C2 |z| N

log | f (z)| log |w0 | log |1 + h(z)| log |w0 | C2 |z| N

E[n] = ∪m|n E[m] (2.1)

I := (t, m) m > 0 and P is tangent to E[m] over t ⊂ C × Z>0 .

P, Q := −(P − O + C P ).(Q − O) (2.4)

where the action of Z on × C× is

f (z) log | f (z)|

for some positive constants C1 and C2 and all z ∈ . We have

for z ∈ . If |w0 | = 1 (so log |w0 | = 0), we have

where the action of Z on × C× is

for all z ∈ , w ∈ C× , and i, j ∈ Z/bZ. Thus Xz , the fiber of X sm → over z = 0,

y 2 + a1 x y + a3 y = x 3 + a2 x 2 + a4 x + a6

c4 → c4 (a2 , a3 , a4 ) and c6 → c6 (a2 , a3 , a4 ).

Thus setting a3 = 2y and a2 = 3x yields the stated fiber.

where L = OC (D ). The choice of ω defines (possibly rational) sections of L and L

For “most” choices of a, W → C is already minimal and L = L. This holds if

F : V × U → Mk (a, t) → F(a, t) := f a (t)

F(a2 + α2 s2 , a3 + α3 s3 , a4 + α4 s4 , t) = m,

a := (a2 + α2 s2 , a3 + α3 s3 , a4 + α4 s4 ) ∈ V .

is étale in a neighborhood of ( p, 0) and it sends n 2-to-1 to a curve that meets O