Bibliography: leaves 211-217 / vii, 217 leaves ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.)--University of Adelaide, Dept. of Pure Mathematics, 1988

This thesis concerns the hyperkähler eightfold constructed by Lehn, Lehn, Sorgen, and van Straten, built from twisted cubics on a cubic fourfold. We study its period, its birational properties and we describe some geometric features.

This thesis analyses the singularities of toroidal compactifications. Motivated by a result of Shepherd-Barron about the first Voronoi compactification of the moduli space of principally polarised abelian varieties, the object taken into consideration consists of the perfect cone (also known as first Voroni) compactification of arithmetic quotients of type IV domains. These are of importance in the context of algebraic geometry because they are used to construct moduli spaces of polarised K3 surfaces and are strongly related to moduli spaces of hyperkähler varieties of higher dimension. The local analysis of singularities of a toroidal compactification reduces to that of finite quotients of toric varieties. The main result of this thesis gives a description of the singularities of the perfect cone compactification of the moduli space of pseudo-polarised K3 surfaces of square-free degree.

Bibliography: leaves 185-195 / 195 leaves ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.)--University of Adelaide, Dept. of Pure Mathematics, 1992

Bibliography: leaves 225-227 / v, 227, 37 leaves : ill ; 30 cm. / Title page, contents and abstract only. The complete thesis in print form is available from the University Library. / Thesis (Ph.D.)--University of Adelaide, Dept. of Pure Mathematics. 1985

Dans cette thèse, on considère des revêtements séparables à deux ouverts de la droite projective sur un corps fini k de caractéristique p>0 et on donne un calcul explicite de la matrice du Frobenius divisé sur le premier espace de cohomologie de Rham de X_k, fibre spéciale du revêtement X étudié. On fournit également un procédé algorithmique permettant d'obtenir la décomposition de Jordan-Hölder du φ-module filtré associé à cette matrice. / We consider X some separable covering with two open set of the projective line on a finite field k of caracteristic p>0 and we give an explicit computation of the matrix of the divided Frobenius on the first de Rham cohomology space of X_k the special fiber of X. We also explain an algorithmic process to get the Jordan-Hölder decomposition of the φ-module associated to this matrix.

Cette thèse est consacrée à l'étude des variétés projectives strictement convexes de volume fini. Une telle variété est le quotient G\U d'un ouvert proprement convexe U de l'espace projectif réel RP^(n-1) par un sous-groupe discret sans torsion G de SLn(R) qui préserve U. Dans un premier temps, on étudie l'adhérence de Zariski des holonomies de variétés projectives strictement convexes de volume fini. Pour une telle variété G\U, on montre que, soit G est Zariski-dense dans SLn(R), soit l'adhérence de Zariski de G est conjuguée à SO(1,n-1). On s'intéresse ensuite à l'espace des modules des structures projectives strictement convexes de volume fini. On montre en particulier que cet espace des modules est un fermé de l'espace des représentations. / In this thesis, we study strictly convex projective manifolds of finite volume. Such a manifold is the quotient G\U of a properly convex open subset U of the real projective space RP^(n-1) by a discrete torsionfree subgroup G of SLn(R) preserving U. We study the Zariski closure of holonomies of convex projective manifolds of finite volume. For such manifolds G\U, we show that either the Zariski closure of G is SLn(R) or it is a conjugate of SO(1,n-1).We also focuss on the moduli space of strictly convex projective structures of finite volume. We show that this moduli space is a closed set of the representation space.