Orders in finite-dimensional algebras over number fi give rise to interesting locally symmetric spaces and algebraic varieties. Hilbert modular varieties or arithmetically defined hyperbolic 3-manifolds, compact ones as well as noncompact ones, are familiar examples. In this talk we discuss various cases related to the general linear group $GL(2)$ over orders in division algebras defined over some number field. Geometry, arithmetic, and the theory of automorphic forms are interwoven in a most fruitful way in this work. Finally we indicate a construction of non-vanishing square-integrable cohomology classes for such arithmetically defined groups.[-]

We will describe two projects. The first which is joint with Avner Ash and Paul Gunnells, concerns arithmetic subgroups $\Gamma$ of $G = SL_4(Z)$. We compute the cohomology of $\Gamma \setminus G/K$, focusing on the cuspidal degree $H^5$. We compute a range of Hecke operators on this cohomology. We fi Galois representations that appear to be attached to the Hecke eigenclasses, based on the operators we have computed. We have done this for both non-torsion and torsion classes. The second project, which is joint with Bob MacPherson, is an algorithm for computing the Hecke operators on the cohomology $H^d$ of $\Gamma$ in $SL_n(Z)$ for all $n$ and all $d$.[-]

Siegel introduced generalised theta series to study representation numbers of quadratic forms. Given an integral lattice $L$ with quadratic form $q$, Siegel's degree $n$ theta series attached to $L$ has a Fourier expansion supported on $n$-dimensional lattices, with Fourier coefficients that tells us how many times $L$ represents any given $n$-dimensional lattice. Siegel proved that this theta series is a type of automorphic form.
In this talk we explore how the theory of automorphic forms, together with the theory of quadratic forms, helps us understand these representation numbers. We reveal arithmetic relations between ”average” representation numbers (where we average over a genus), and finally we give an explicit formula for these average representation numbers in terms of the Fourier coefficients of Siegel Eisenstein series. In the case that $n = 1$ (meaning we are looking at how often $L$ represents an integer) this yields explicit numerical formulas for these average representation numbers.[-]