Abstract. In 1967, Shioda [23] determined the ring of invariants of binary octavics and their syzygies using the symbolic method. We discover that the syzygies determined in [23] are incorrect. In this paper, we compute the correct equations among the invariants of the binary octavics and give necessary and sufficient conditions for two genus 3 hyperelliptic curves to be isomorphic over an algebraically closed field k, char k = 2, 3, 5, 7. For the first time, an explicit equation of the hyperelliptic moduli for genus 3 is computed in terms of absolute invariants.
IntroductionLet k be an algebraically closed field. A binary form of degree d is a homogeneous polynomial f (X, Y ) of degree d in two variables over k. Let V d be the k-vector space of binary forms of degree d. The group GL 2 (k) of invertible 2 × 2 matrices over k acts on V d by coordinate change. Many problems in algebra involve properties of binary forms which are invariant under these coordinate changes. In particular, any hyperelliptic genus g curve over k has a projective equation of the form Z 2 Y 2g = f (X, Y ), where f is a binary form of degree d = 2g + 2 and non-zero discriminant. Two such curves are isomorphic if and only if the corresponding binary forms are conjugate under GL 2 (k). Therefore the moduli space H g of hyperelliptic genus g curves is the affine variety whose coordinate ring is the ring of GL 2 (k)-invariants in the coordinate ring of the set of elements of V d with non-zero discriminant. It is well known that the moduli spaces H g of hyperelliptic curves of genus g, g = 4, are all rational varieties, i.e. isomorphic to a purely transcendental extension field k(t 1 , . . . , t r ); see Igusa [9], Katsylo [10].Generators for this and similar invariant rings in lower degree were constructed by Clebsch, Bolza and others in the last century using complicated symbolic calculations. For the case of sextics, Igusa [9] extended this to algebraically closed fields of any characteristic using difficult techniques of algebraic geometry. For a modern treatment of the degree six case see [11].The case of binary octavics has been first studied during the 19th century by von Gall [24] and Alagna [1,2]. Shioda in his thesis [23] determined the structure of the ring of invariants R 8 , which turns out to be generated by nine SL(2, k)-invariants J 2 , · · · , J 10 satisfying five algebraic relations. He computed explicitly these five syzygies, and determined the corresponding syzygy-sequence and therefore the structure of the ring R 8 ; see Shioda [23].This paper started as a project to implement an algorithm which determines if two genus 3 hyperelliptic curves are isomorphic over C. According to Shioda [23, Thm. 5]; two genus 3 hyperelliptic curves are isomorphic if and only if the corresponding 9-tuples (J 2 , . . . , J 10 ) are equivalent, satisfying five syzygies R i (J 2 , . . . , J 10 ) = 0, for i = 1, . . . , 5 and non-zero discriminant ∆ = 0. While trying to implement the syzygies R i (J 2 , . . . , J 10 ) = 0, for i = 1, . . . , 5 w...