Distribution modulo one and Ratner's theorem

Abstract

Measure rigidity is a branch of ergodic theory that has recently contributed to the solution of some fundamental problems in number theory and mathematical physics. Examples are proofs of quantitative versions of the Oppenheim conjecture, related questions on the spacings between the values of quadratic forms, a proof of quantum unique ergodicity for certain classes of hyperbolic surfaces, and an approach to the Littlewood conjecture on the nonexistence of multiplicatively badly approximable numbers. In these lectures we discuss a few simple applications of one of the central results in measure rigidity: Ratner's theorem. We shall investigate the statistical properties of certain number theoretic sequences, specifically the fractional parts of $mα$, $m=1,2,3,...$, (a classical, well understood problem) and of $\sqrt{mα}$ (as recently studied by Elkies and McMullen). By exploiting equidistribution results on a certain homogeneous space, we will show that the statistical properties of these sequences can exhibit significant deviations from those of independent random variables. The ``randomness'' of other, more generic sequences such as $m^2α$ and $2^m α$ mod 1 has been studied extensively. These notes are based on lectures presented at the Institute Henri Poincare Paris, June 2005, and at the summer school `Equidistribution in number theory', CRM Montreal, July 2005. The author gratefully acknowledges support by an EPSRC Advanced Research Fellowship.