Present some empirical evidence on long-run Purchasing Power Parity (PPP) using a sample of annual data covering the period 1900-1987. Three exchange rates are used in the analysis: the dollar, yen and French franc against the pound. It is shown that PPP does hold in the long run and that is some evidence for the proportionality and symmetry restrictions. Evidence for exclusiveness is mixed and the results are influenced by model specification. Estimates of error-correction models show that there are substantial short-run deviation from PPP that take two to three years to correct.