Decomposable Jacobians
Angel Carocca, Herbert Lange, Rub\'i E. Rodr\'iguez

TL;DR
This paper presents examples of smooth projective curves with Jacobians that are isogenous to a product of many Jacobians, highlighting complex decompositions in algebraic geometry.
Contribution
It provides explicit examples of curves with highly decomposable Jacobians, expanding understanding of Jacobian structures in algebraic geometry.
Findings
Constructed curves with Jacobians isogenous to large products
Demonstrated the existence of complex Jacobian decompositions
Extended known examples of Jacobian decompositions
Abstract
In this paper we give examples of smooth projective curves whose Jacobians are isogenus to a product of an arbitrarily high number of Jacobians
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsAdvanced Differential Equations and Dynamical Systems · Geometric and Algebraic Topology · Algebraic Geometry and Number Theory
Decomposable Jacobians
Angel Carocca, Herbert Lange and Rubí E. Rodríguez
Departamento de Matemática y Estadística, Universidad de La Frontera, Avenida Francisco Salazar 01145, Casilla 54-D, Temuco, Chile.
Department Mathematik, Universität Erlangen, Cauerstrasse 11, 91058 Erlangen, Germany.
Departamento de Matemática y Estadística, Universidad de La Frontera, Avenida Francisco Salazar 01145, Casilla 54-D, Temuco, Chile.
Abstract.
In this paper we give examples of smooth projective curves whose Jacobians are isogenus to a product of an arbitrarily high number of Jacobians.
Key words and phrases:
Jacobian, Prym variety, Coverings
1991 Mathematics Subject Classification:
14H40, 14H30
The authors were partially supported by Grants Fondecyt 1190991 and CONICYT PAI Atracción de Capital Humano Avanzado del Extranjero PAI80160004
1. Introduction
In [5], Ekedahl and Serre gave examples of curves whose Jacobian is completely decomposable, i.e. isogenous to a product of elliptic curves. The highest genus of their examples is . They asked among other things whether the genus of a curve with a completely decomposable Jacobian is bounded above. Although in the meantime many other examples of such Jacobians have been given (see e.g. [6]), no example of genus bigger than 1297 seems to be known.
In this paper we consider an easier question, namely: can a Jacobian be isogenous to the product of arbitrary many Jacobians of the same genus (not necessarily equal to one)? In [4] we gave examples of Jacobians which are isogenous to an arbitrary number of Prym varieties of the same dimension. The main result of this paper is the following theorem (see Corollary 5.3).
Theorem 1.1**.**
Given any positive integer , there exists a smooth projective curve whose Jacobian is isogenous to the product of Jacobian varieties of the same dimension.
To be more precise, in Corollary 5.2, for any positive integer and any integer we give examples of curves of genus whose Jacobians are isogenous to the product of Jacobians of curves of genus The idea is to consider curves of genus with an action of the group of signature . Then the quotient curves will have the asserted properties.
The Jacobians are Jacobians of subcovers of the curve of genus and they are in fact isomorphic to Prym varieties of étale double covers of other subcovers. Using these isomorphisms gives a second proof of the theorem, which actually gives a bit more, namely an estimate of the degree of the isogeny.
In Section 2 we investigate the representations of the group . In Section 3 we recall some results on curves with -action. In Section 4 we study the diagram of subcovers of the curve . Section 5 contains the first proof of the above mentioned theorem. In Section 6 we give the proof using the Prym varieties and the trigonal construction. Finally in the last section we compute the group algebra decomposition of the Jacobian of the curve .
2. The group
Consider the group the alternating group of degree 4 with the Klein group and , and for any positive integer let
[TABLE]
where the subgroups for each and for . Obviously is of order
[TABLE]
Also is a -Sylow subgroup of and
In this section we will show that all subgroups of of index are maximal in (there is no subgroup strictly between it and ); if we consider only such containing (there is exactly one in each conjugacy class of subgroups of index four), they correspond bijectively to the subgroups of index four in that are normal in , and also to the set of irreducible representations of of degree three. An example of such a subgroup is with . We need some preliminaries.
Lemma 2.1**.**
* has no subgroups of index *
Proof.
Suppose such that Then and by the Sylow Theorem In this way, we have This implies
[TABLE]
and
Hence for some . Since we have
[TABLE]
The last equation follows from the fact that and . Then and hence
[TABLE]
a contradiction. ∎
Now let be any subgroup such that As we saw above, there are such subgroups. According to Lemma 2.1, is maximal in . Moreover, we have,
Lemma 2.2**.**
* ; and *
Proof.
Since is a maximal subgroup of and we have
Since we have that Also, since is abelian. Therefore, Also ∎
According to a result of elementary number theory, the number
[TABLE]
is an integer for all positive integers .
Lemma 2.3**.**
There are complex irreducible representations of degree and complex irreducible representations of degree of ( the trivial character). These are all the irreducible complex representations of .
Proof.
For any irreducible character of and any element we denote by the conjugate character of defined by , The stabilizer of the trivial character of is the group , whereas the stabilizer of any non-trivial character of is trivial. Hence there are exactly orbits for the action of on the set of all irreducible characters of . Let be a system of representatives of these orbits and the irreducible characters of Then, according to [9, Proposition 25] the irreducible representations of are and the induced representation for each ∎
This implies immediately,
Corollary 2.4**.**
The irreducible rational representations of are exactly the trivial representation , the representations of degree for , and the representation of degree .
Lemma 2.5**.**
Let be a non-trivial irreducible character of and the corresponding irreducible -dimensional representation of . Then
- (i)
** 2. (ii)
* ; and *
Proof.
(i): Since is isomorphic to a finite cyclic subgroup of and we have
(ii): Since and we have
Hence Therefore
We have So
[TABLE]
It is well known that . Hence
[TABLE]
∎
Conversely, we have
Lemma 2.6**.**
Let such that Then
- (i)
There is a non-trivial character* of such that is an irreducible representation of with * 2. (ii)
* and *
Proof.
(i): We have that is isomorphic to the Klein group of order Consider such that and Then since
There is a non-trivial character of such that In this way, for the induced representation we have by the previous lemma that
(ii): Since with , we have and . ∎
Combining these lemmas we get
Proposition 2.7**.**
- (i)
There are canonical bijections between the following sets
- •
**
- •
**
- •
** 2. (ii)
[TABLE]
Proof.
(i): The bijections are given by
[TABLE]
If denotes the representation of induced by the trivial representation of then for each such ,
[TABLE]
So is of degree 3.
(ii): There are non-trivial irreducible characters of . The group acts non-trivially on them. So Lemma 2.6 and part (i) of the proposition imply the assertion. ∎
Remark 2.8**.**
Let such that (the corresponding maximal subgroup), and as in the Proposition 2.7.
Consider such that and Observe that for each there are three such , forming a conjugacy class of subgroups of . Also, consider , the irreducible character of with Then
- •
for So ;
- •
for ;
- •
;
- •
for ;
- •
3. Action of Hecke algebras of an abelian variety
Let be any finite group acting on an abelian variety over the field of complex numbers and let be a subgroup. In this section we recall a result of [4] together with the notation needed for it.
The element
[TABLE]
is an idempotent of the group algebra . The Hecke algebra for in is defined to be the subalgebra
[TABLE]
of . The action of on induces an algebra homomorphism
[TABLE]
in a natural way. We denote the elements of and their images by the same letter. For any element we define its image in by
[TABLE]
where is any positive integer such that is in . It is an abelian subvariety which does not depend on the chosen integer . Consider the abelian subvariety of given by
[TABLE]
Restricting (3.1) to gives an algebra homomorphism
Let denote the irreducible rational representations of . To any there corresponds an irreducible complex representation , uniquely determined up to an element of the Galois group of over , where is the field obtained by adjoining to the values of the character of .
To each we can associate a central idempotent of by
[TABLE]
Let denote the representation of induced by the trivial representation of . It decomposes as
[TABLE]
with and the Schur index of . Renumbering if necessary, let denote the set of all irreducible rational representations of such that . Then there is a bijection from this set to the set of all irreducible rational representations of the algebra . An analogous statement holds for the complex irreducible representations of and of . Let denote the representation of associated to the complex irreducible representation of and Galois associated to The dimension of is given by
[TABLE]
where denotes the field of definition of the representation . Recall that is the Schur index of .
For consider the central idempotents of given by
[TABLE]
Then decomposes as
[TABLE]
Defining for the abelian subvarieties
[TABLE]
one obtains the following isogeny decomposition of , given by the addition map
[TABLE]
It is uniquely determined by and the action of and called the isotypical decomposition of . So in order to describe the abelian subvariety , it suffices to describe the . This is done by [4, Theorem 4.3]. For it we need one more notation. Consider the decomposition of into double cosets of in ,
[TABLE]
with . Then a basis for the Hecke algebra is given by the elements
[TABLE]
for . If denotes the character of the representation , then [4, Theorem 4.3] says,
Theorem 3.1**.**
Suppose and . Then
[TABLE]
where the index [math] means the connected component containing [math].
4. The diagram of subcovers of
We start this section recalling some basic properties on groups action on smooth projective curves. Let be a smooth projective curve of genus a finite group acting on and be the quotient projection. This cover may be partially characterized by a vector of numbers where is the genus of the integer is the number of branch points of the cover and the integers are the orders of the cyclic subgroups of which fix points on . We call the branching data of on . These numbers satisfy the Riemann-Hurwitz formula
[TABLE]
A tuple of elements of is called a *generating vector of type * if
[TABLE]
where
Riemann’s Existence Theorem gives us the following theorem (see [2, Proposition 2.1])
Theorem 4.1**.**
The group acts on a smooth projective curve of genus with branching data if and only if has a generating vector of type satisfying the Riemann-Hurwitz formula (4.1).
From now on let be the group of Section 2. Recall that
[TABLE]
where for each the subgroups are isomorphic to Klein group and the subgroups are isomorphic to the alternating group of degree four.
Remark 4.2**.**
For we have that is isomorphic to the alternating group of degree four. It is easy to check that acts on a smooth projective curve of genus with branching data and
In this case, if denotes the Prym variety of then, by the trigonal construction (see [7]), we have is isomorphic to as principally polarized abelian variety.
Lemma 4.3**.**
For the group acts on a smooth projective curve of genus with branching data for
Proof.
We illustrate by giving a generating vector for
For consider the elements and
Since we have Also, it is clear that In this way is a generating vector of type for
For is it clear that is a generating vector of type for
Now, for we can apply a similar procedure as the one described above in both cases even or odd.
∎
Let denote a smooth projective curve with an action of with signature The action induces an action of on the Jacobian of . In this section we will study the subcovers of which will be important for us.
For any subgroup of let denote the quotient curve
[TABLE]
According to Proposition 2.7 the subgroup has exactly subgroups of index which are normal in . Let be these subgroups and denote the corresponding maximal subgroups of For any we choose a subgroup containing with . According to Remark 2.8, the conjugacy class of is uniquely determined. With these notations we have the following diagram of covers of curves.
[TABLE]
Lemma 4.4**.**
Let be Galois with group and action with signature . Then
- (i)
* is totally ramified and is étale;* 2. (ii)
Over each of the branch points of the map admits ramification points of ramification index and point which is étale over .
Proof.
Let be a branch point of .
(i): Since is cyclic of degree 3, it is either étale over or totally ramified. Suppose it is étale. Since is a branch point of , it is also a branch point of the Galois cover of type (i.e. all points of over are of ramification degree 2 with respect to ). This gives a contradiction, since is not divisible by 3. This completes the proof of (i).
(ii): Let denote the branch points of and be a point in the fibre over . Then either is of ramification degree 2 over or étale over . In the first case the map is étale over and in the second case totally ramified over .
Let be the number of points in the fibre in over at which the map is ramified. According to the Riemann-Hurwitz (4.1) formula we get
[TABLE]
where the inequality follows from the inequalities . So we have
[TABLE]
On the other hand, [8, Corollary 3.4] gives a method to compute the genus of Let be the stabilizer of the branch point with Then
[TABLE]
Hence the equivalence (4.3) gives the assertion. ∎
From the Riemann-Hurwitz (4.1) formula we immediately get from Lemma 4.4,
Lemma 4.5**.**
Under these assumptions we have
- •
;
- •
;
- •
;
- •
;
- •
;
- •
.
Corollary 4.6**.**
If denotes the Prym variety of , we have
[TABLE]
Proof.
. So Lemma 4.5 implies the assertion. ∎
Corollary 4.6 suggests that there is a relation between these Pryms and which we investigate in the next section. The second equality is explained by the following proposition.
Proposition 4.7**.**
For in there is a canonical isomorphism of principally polarized abelian varieties
[TABLE]
Proof.
Observe that for each in , acts on the corresponding curve (of genus ) and the involutions in act without fixed points (since is étale). It then follows from the trigonal construction (see [7]) that is isomorphic to as principally polarized abelian varieties. ∎
5. The isogeny decomposition of the Jacobian of
Let the notations be as above. For denote by the covering given by . The pull back homomorphisms
[TABLE]
are isogenies onto their images. Considering the composition of these isogenies with the addition map we get a canonical homomorphism
[TABLE]
The main result of this section is the following theorem.
Theorem 5.1**.**
The homomorphism is an isogeny.
Proof.
Recall that for any subgroup of , denotes the representation of induced by the trivial representation of . Moreover, denotes the trivial representation and the irreducible rational represetations of degree 3 of . It is easy to see that
[TABLE]
Let and denote the corresponding central idempotents of the Hecke algebra . Then according to (3.4), the idempotent decomposes as
[TABLE]
Considering the idempotents as elements of ,
[TABLE]
since and
[TABLE]
since according to Remark 2.8. Hence according to (3.5), the addition map gives an isogeny
[TABLE]
Combining this with the isogenies we get the isogeny as claimed. ∎
Combing several results, we can say,
Corollary 5.2**.**
For any positive integer consider the integer . Then there exist curves of genus for any integer whose Jacobian is isogenous to the product of Jacobians.
Proof.
According to Lemma 4.3 there exist curves of genus with an action of the group of Section 2 and with . So Theorem 5.1 gives the assertion. ∎
Corollary 5.3**.**
Given any positive integer , there exist smooth projective curves whose Jacobian is isogenous to the product of Jacobian varieties of the same dimension.
Proof.
Choose a positive integer such that . This is equivalent to . According to the previous corollary there exist curves whose Jacobian is isogenous to the product of Jacobians. ∎
Remark 5.4**.**
The isogeny is the isotypical decomposition of with respect to the Hecke algebra action of on .
In order to describe the isotypical components of by equations we will apply Theorem 3.1. There are conjugacy classes of involutions in , and double cosets in ,
[TABLE]
with . All other have representatives in exactly one of the conjugacy classes of involutions. So
[TABLE]
where is an involution of . According to Section 3, a basis of given by
[TABLE]
Lemma 5.5**.**
- (i)
* for ;* 2. (ii)
for all we have for all and ,
[TABLE]
Proof.
(i): Clearly . Hence which implies the assertion.
(ii): The first equation is obvious. For the second observe that for all elements of order three in , and that each is composed by the three elements in the conjugacy class of and six elements of order three. This implies the assertion. ∎
The following theorem gives the description of the isotypical components of .
Theorem 5.6**.**
[TABLE]
Proof.
According to (5.1) we have . Since moreover is defined over , we can apply Theorem 3.1 to give
[TABLE]
where the last equation follows from Lemma 5.5(ii). ∎
6. The degree of the isogeny
In Section 5 we used the right hand side of the diagram (4.2) to decompose the Jacobian of the curve . By Proposition 4.7 we know that for all there is a canonical isomorphism of principally polarized abelian varieties . In this section we will see that one can also use the left hand side of the diagram to decompose the Jacobian , here as a product of Prym varieties. In view of Proposition 4.7, this is the same decomposition as above, however has the advantage that it gives in addition something about the degree of the isogeny.
Let the notation be as in the last section. For we denote
[TABLE]
the maps of diagram (4.2). The maps and are the induced homomorphisms of the associated Jacobians. Then the addition map gives a homomorphism
[TABLE]
According to Corollary 4.6, and are of the same dimension. To be more precise, we have
Theorem 6.1**.**
* is an isogeny with kernel contained the the -division points.*
Applying the isomorphism of Proposition 4.7, the homomorphism coincides with the isogeny of Theorem 5.1. The new result is the estimate of its degree, namely
[TABLE]
which follows from the fact that the group of 2-division points of is of order . The proof is analogous to the proof of [4, Theorem 3.1] (also see [3, Proposition 3.2]). For the convenience of the reader we give full details. We need the following proposition.
Proposition 6.2**.**
Let be a Galois cover of smooth projective curves with Galois group and a subgroup. Denote by and the corresponding covers. If is a complete set of representatives of , then we have
[TABLE]
Proof.
This is a special case of [4, Corollary 3.3] ∎
Denote for ,
[TABLE]
and let
[TABLE]
Recall that generates the group . Then we have the following commutative diagram,
[TABLE]
with .
For consider the following subdiagram
[TABLE]
with and .
Proposition 6.3**.**
For the map
[TABLE]
is multiplication by .
For the proof of the proposition we need the following lemma.
Lemma 6.4**.**
For any two subgroups of of index we have
[TABLE]
Proof.
Since , we have which implies the assertion. ∎
Proof of Proposition 6.3.
Since is an isogeny, it suffices to show that the composition
[TABLE]
is multiplication by .
Now from Proposition 6.2 we deduce
[TABLE]
For the proof of (6.3) we apply Proposition 6.2 with . Let be any element of and be any element of ; then is a complete set of representatives of . So Proposition 6.2 gives
[TABLE]
But implies that . This gives
[TABLE]
But is a double cover, which implies . Since and are arbitrary elements of and respectively, this gives the assertion.
Now for any ,
[TABLE]
By equation (6.3) and Lemma 6.4 we have
[TABLE]
and for and 2,
[TABLE]
since and by Lemma 6.4 half of the elements of the subgroup belong to , hence fix , and the other half belongs to and hence sends to . Together they complete the proof of the proposition. ∎
Proof of Theorem 6.1.
Since
[TABLE]
Proposition 6.3 implies that is multiplication by . In particular has finite kernel. But according to Corollary 4.6, and have the same dimension. So is an isogeny. ∎
Remark 6.5**.**
As we noticed already above, the isogenies and are compatible with respect to the isomorphism of Proposition 4.7. Hence Theorem 6.1 is also the isotypical decomposition of .
7. The group algebra decomposition of
Let as act on the curve as above. For the trivial subgroup the Hecke algebra coincides with the whole group algebra and we have and for all . According to (3.5) and [4, equation (2.3)] the isotypical decomposition of is
[TABLE]
where we write for simplicity for each irreducible rational representation of , for which we refer to Corollary 2.4. Note that , since .
This decomposition can be decomposed further. In fact, as outlined in [4, Section 2], for each there is a not uniquely determined abelian subvariety such that . On the other hand, at least by these methods cannot be decomposed further. Hence we get a decomposition
[TABLE]
which is called the group algebra decomposition of (see [1, Section 13.6]).
Lemma 7.1**.**
- (i)
; 2. (ii)
* for .*
Corollary 7.2**.**
A group algebra decomposition of is
[TABLE]
Proof of the Lemma 7.1.
(i): The covering is Galois with group of order 3. There are 2 irreducible rational representation of this group, namely and the trivial representation . To corresponds the Prym variety and to the trivial representation . This implies and thus , since the isotypical decomposition is uniquely determined.
(ii): According to Theorem 6.1, contained in and thus in , since is ramified. So clearly it is contained in . Since one can conclude from Lemma 4.5 that , this implies that there is an isogeny .
The last isogeny is a consequence of Proposition 4.7. ∎
The reference list from the paper itself. Each links out to its DOI / PubMed record.
- 1[1] Ch. Birkenhake, H. Lange, Complex abelian varieties . 2nd edition, Grundl. Math. Wiss. vol 302, Springer (2004).
- 2[2] A. Broughton, Classifying finite group actions on surfaces of low genus , J. Pure Appl. Algebra 69, (1991), no. 3, 233–270.
- 3[3] A. Carocca, H. Lange, R. E. Rodríguez, Étale double covers of cyclic p 𝑝 p -gonal curves . To appear in J. Algebra. ar Xiv:math/1904.02684 (2019)
- 4[4] A. Carocca, H. Lange, R. E. Rodríguez, Equations for abelian subvarieties . Preprint (2019).
- 5[5] T. Ekedahl, J.-P. Serre: Examples de courbes algébriques à jacobienne complètement décomposable . C. R. Acad. Sci. 317, Sér. I, 509-513 (1993).
- 6[6] J. Paulus, A. M. Rojas, Completely decomposable Jacobian varieties in new genera . Exp. Math. 26, no. 4, 430–445, (2017).
- 7[7] S. Recillas , R. E. Rodríguez: Prym varieties and fourfold covers . ar Xiv:math/0303155 (2003).
- 8[8] A. M. Rojas, Group actions on Jacobian varieties . Rev. Mat. Iberoam. 23, 397-420 (2007).
