Moduli of abelian varieties (Q5921631)

Let \(X\) be an abelian variety over an algebraically closed field \(k\). For any very ample line bundle \(L\) on \(X\), the corresponding embedding \(\varphi_L: X\hookrightarrow \mathbb{P}^N\), \(N=\dim_k H^0(X,L)-1\), gives \(X\) the structure of a closed subvariety of \(\mathbb{P}^N\). The pair \((X,L)\) is then called a polarized abelian variety, with polarization (bundle) \(L\), and the elements of \(H^0(X,L^{\otimes k})\) are called the \(k^{\text{th}}\)-order theta functions on \(X\) with respect to the polarization \(L\). As the embedding \(\varphi_L\) is defined by means of a basis of the space of first-order theta functions, \(H^0(X,L)\), the ideal defining \(\varphi_L(X)\) in \(\mathbb{P}^N\) may then be considered as the ideal generated by the algebraic relations among the basic theta functions of \((X,L)\) in the ring \(\bigoplus_kH^0 (X,L^{\otimes k})\). According to the classical terminology, those relations are called theta relations for \((X,L)\). In the 1960s, D. Mumford worked on the magnificent project to set up a purely algebraic theory of theta functions for polarized abelian varieties over arbitrary ground fields. Continuing the earlier fundamental work of Igusa [cf. \textit{J.-I. Igusa}, Am. J. Math. 86, 219-246 (1964; Zbl 0146.31703) and 88, 221-236 (1966; Zbl 0146.31704) and the book ``Theta functions'' (1972; Zbl 0251.14016)], \textit{D. Mumford} presented his celebrated series of papers ``On the equations defining abelian varieties. I, II, III'' [Invent. Math. 1, 287-354 (1966); 3, 75-135; 215-244 (1967; Zbl 0219.14024)], in which he gave both a very complete description of the homogeneous coordinate ring of a single polarized abelian variety and concrete projective embeddings of the moduli spaces of suitably polarized abelian varieties. However, Igusa's and Mumford's deep-going results on the equations (theta relations) defining polarized abelian varieties, or their moduli spaces, respectively, depend on the crucial assumption that the polarizing line bundle is a certain power of another bundle. This means, in particular, that their methods do not automatically apply to some special cases such as elliptic normal curves of odd degree or, from the moduli aspect, moduli spaces of elliptic curves with odd level structure. The monograph under review aims at analyzing some of those cases of polarized abelian varieties, which do not fit (so far) into the general theory of Igusa-Mumford. Basically, the authors modify (and generalize) some of Mumford's constructions to arrive at theta relations which are also satisfied by polarized abelian varieties with level structures different from those covered by Mumford's approach. Although these new equations are not shown to define the corresponding moduli spaces precisely, they are nevertheless highly interesting for their own sake, just because they define significant geometric loci containing those moduli spaces, on the one hand, and because they generalize (in most cases) relations among classical theta functions, on the other hand. These classical theta relations, which were found by \textit{Felix Klein} and \textit{Heinrich Burckhardt} in the last century, are thus shown to be of very general character, in that they also follow from Mumford's abstract approach to general theta functions. The main part of the text is devoted to a subtle analysis of the geometric loci defined by the (old and new) theta relations that the authors have derived in the first, more technical part of the book. The explicit equations make it possible to generalize theorems about moduli spaces for certain polarized abelian varieties to theorems about wider classes of algebraic varieties, and that is, apart from the enhanced framework for algebraic theta functions, another aspect which makes the authors' work particularly valuable. The structure of the book is as follows: Chapter I gives an introductory account on standard Heisenberg groups, their representation theory and their interrelations. This is used, in chapter II, to discuss the Heisenberg groups of line bundles on abelian varieties, including the Mumford functor and the new concept of strongly symmetric line bundles, which is successfully used in the sequel. In chapter III, the authors introduce the notion of a strongly symmetric theta structure and derive, in the vein of Mumford's approach, various theta formulae such as an addition formula, a duplication formula, an inversion formula, a product formula and their fundamental relation among the corresponding theta constants. Chapter IV is devoted to the study of the fundamental relations in the case of line bundles of odd level, i.e., in the case which is not covered by Mumford's techniques. Particular attention is paid to modular curves of prime level and to a generalization of Burckhardt's quartic invariant. Chapter V, entitled ``Invariant theory, arithmetic, and vector bundles'', deals with miscellaneous relations of the foregoing results to invariant theory, number theory and to automorphic forms. The results presented here have been partly obtained by the A. Adler, and they appear now in a systematic context. These five chapters are supplemented by five appendices which are reproductions of five unpublished papers of the A. Adler. Appendix 1 provides an explicit form of the Weil representation of a finite symplectic group, which is used to analyze the equations of F. Klein and H. Burckhardt in the general context. Appendix 2 deals with the construction of modular forms of fractional weight for the modular curve \(\Gamma(11)\), whereas appendix 3 describes different embeddings of \(\Gamma(11)\) into \(\mathbb{P}^4\). Appendix 4 is concerned with some aspects of the geometry of the Hessian of a cubic threefold, which generalize the methods applied to Klein's curve of level 11 and are discussed more thoroughly in the concluding appendix 5. Altogether, the present book represents a very interesting and valuable contribution towards Mumford's theory of general theta functions and moduli of polarized abelian varieties. Mumford's powerful machinery is shown to work also in some of those cases that had not been included in this framevork, up to now, and the classical theta relations of Klein and Burckhardt are here tied into the general theory. Also, it is highly gratifying that A. Adler's beautiful, so far unpublished articles have been woven into the text and, in this way, have not only become accessible to the wider mathematical community, but have simultaneously found their appropriate place in the related literature.