OBJECTIVE This article examines trends in hip and knee replacement surgery between 1981/82 and 1998/99, focussing on procedures involving seniors. It also presents 1998/99 data on readmissions within 30 days. METHODS Data on hip and knee replacement are from the Hospital Morbidity Database for 1981/82 through 1998/99. The Person-oriented Information Database is used to examine readmissions in 1998/99. Supplementary data on arthritis are from the 1998/99 National Population Health Survey. METHODS Hospitalization rates were calculated by dividing the number of hip and knee surgery separations by the population estimates for the relevant age/sex group and multiplying by 100,000. Population estimates for 1998 were used to calculate age-adjusted hospitalization rates. RESULTS Between 1981/82 and 1998/99, the numbers and rates of hip and knee replacement increased substantially, while length of stay for both procedures declined. By 1998/99, knee replacements outnumbered those for hip. Both procedures had relatively low in-hospital mortality and post-surgery complication rates.