Conventional census-based measures of population mobility are conceptually abstruse, ignore multiple moves and obscure the diversity of human migration experience. This paper explores three alternatives and outlines their strengths and limitations. Application of life table techniques to convert transition rates to migration expectancies generates measures that are more readily understood, automatically standardizes for age and enables the timing of mobility to be analysed methodically. Data on movement frequencies and residence duration provide new perspectives indicating substantial chronic mobility and significant differences between frequent movers and long term stayers. A number of simple summary statistics are proposed to supplement transition rates.The Australian census, like that in most countries, measures migration as a change in place of usual residence, or transition, between two points in time. For example, according to the 1986 Census, 16.7 per cent of Australians had changed address since 1985 and 41.2 per cent had moved since 1981 (Bell 1995). Transition rates of this type have a number of advantages: they offer a straightforward, unambiguous definition of migration; they permit comparison of mobility rates among different segments of the population; and they facilitate monitoring of the way mobility changes over time.Offsetting these benefits, however, are several limitations. One deficiency, widely recognized, is that repeat and return moves occurring within the observation interval are disregarded. The longer the observation period, the greater the number of moves that are missed and conventional five year mobility rates radically understate the true extent of mobility. Similarly, moves that took place before the start of the period are not recorded. These omissions are symptomatic of the fact that transition rates measured over a fixed interval reduce to a simple dichotomy (moved/did not move) a repetitive process in which most individuals participate at some time in their lives. The mobility continuum is partitioned into two discrete elements, movers and stayers, which are treated as homogeneous. This in turn obscures the considerable diversity of human population movement.A related deficiency is that the summary statistics commonly reported from fixed interval measures are conceptually abstruse. The proportion of people who change residence in a one or five year period has little meaning in terms of individual experience and this makes it difficult to explain clearly to non-specialists how often people move. The problems of interpretation are compounded when the observed rates for selected sub-populations or time periods are standardized to control for the effects of differing age structures. Yet without such standardization, apparent variations in mobility may be quite misleading.This paper explores three alternative approaches to measuring mobility that assist in overcoming these deficiencies. The first involves applying life table techniques to census data to calculate migration expectan...