A theory of the limiting behaviour of estimators and functions of estimators, their distributions and moments, as the sample size approaches infinity. The results of asymptotic theory are used as approximations in finite sample inference, when the true finite sample properties are unknown.