The Pareto distribution has long been a source of fascination to economists, and the Pareto coefficient is widely used, in theoretical and empirical studies, as a summary of the degree of concentration of top incomes. This paper examines the empirical evidence from income tax data concerning top incomes in the UK, contrasting the dramatic changes that took place in the twentieth century, after 1918, with the much more modest changes in the preceding nineteenth century. Probing beneath the surface, the paper identifies a number of features of the evolution of the UK income inequality that warrant closer attention. These include the changing shape of the upper tail, where there is a link with Pareto's theory of elites, the need for a richer functional form to describe top incomes, and the limited evidence at the top of the distribution for a Kuznets curve in nineteenth century Britain.