Monodromy of Codimension-One Sub-Families of Universal Curves

# Monodromy of Codimension-One Sub-Families of Universal Curves

Richard Hain Department of Mathematics
Duke University
Durham, NC 27708-0320
July 27, 2019
###### Abstract.

Suppose that , that and that . The main result is that if is a smooth variety that dominates a codimension subvariety of , the moduli space of -pointed, genus , smooth, projective curves with a level structure, then the closure of the image of the monodromy representation has finite index in . A similar result is proved for codimension families of principally polarized abelian varieties.

###### 1991 Mathematics Subject Classification:
Primary 14D05, 14H15; Secondary 14F35, 14F45
Supported in part by grants DMS-0706955 and DMS-0103667 from the National Science Foundation, and by MSRI

## 1. Introduction

Suppose that are integers satisfying , and . Denote the moduli space of smooth complex projective curves of genus with a level structure by and the th power of the universal curve over it by . The moduli space of smooth -pointed smooth projective curves of genus with a level structure is a Zariski open subset of . These will all be regarded as orbifolds. There is a natural monodromy representation

 ρ:π1(Cng[ℓ],xo)→Spg(Z)

whose image is the level -congruence subgroup of .

The profinite completion of a discrete group will be denoted by . Denote the profinite completion of the integers by . A homomorphism induces a homomorphism .

###### Theorem 1.

Suppose that and . If is a dominant morphism from a smooth variety to an irreducible divisor in , then the image of the monodromy representation

 π1(E,eo)∧→π1(D,do)∧→Spg(ˆZ)

has finite index in .

The statement is false when as will be explained in Section LABEL:sec:density. We will prove a stronger version of this theorem (Theorem LABEL:thm:density_curves), in which is replaced by a “suitably generic linear section” of dimension of it. We also prove similar result for abelian varieties (Theorem LABEL:thm:density_ppavs).

Each rational representation of determines a local system over . The theorem implies that when does not contain the trivial representation, the low dimensional cohomology of with coefficients in changes very little when is replaced by a Zariski open subset of or by its generic point.

###### Corollary 2.

Suppose that , and . If is a Zariski open subset of , then for all non-trivial, irreducible representations of , the map

 Hj(Cng[ℓ],V)→Hj(U,V)

induced by the inclusion is an isomorphism when and an injection when .

The groups are known for all when and ; a canonical subspace of it is known [hain:torelli] when when . The resulting computation of the Galois cohomology groups of the absolute Galois group of the function field , when is a number field, is an essential ingredient of the author’s investigation [hain:sec_conj] of rational points of the universal curve over the function field of .

The proof of Theorem 1 easily reduces to the case . Putman [putman:pic] has proved that the Picard group of has rank when . A standard argument using the that fact that is quasi-projective then implies that every irreducible divisor in is ample. A Lefschetz type theorem due to Goresky and MacPherson [smt] implies that when is generic, is an isomorphism. In this case, the result is immediate. The principal difficulty occurs when has self-intersections.

In this case the image of the homomorphism induced by its normalization may have infinite index in , as it does in Figure 1. This issue is addressed by a technical result, Theorem LABEL:thm:strictness, from which Theorem 1 follows directly. It is a hybrid of a “non-abelian strictness theorem” and a Lefschetz-type theorem.

The term non-abelian strictness theorem needs further explanation. Input for one such type of theorem is a diagram

 (1)
You are adding the first comment!
How to quickly get a good reply:
• Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
• Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
• Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
The feedback must be of minimum 40 characters and the title a minimum of 5 characters