Small property size is often cited as one of the major causes of rangeland degradation in Australia. However, there is some conjecture as to the importance of this effect and the process by which small property sizes lead to rangeland degradation. Relatively little empirical analysis of these issues has been undertaken, especially in a dynamic context which is all important in the case of rangeland degradation. Regression and dynamic programming techniques are employed in this study to investigate and measure the impact of property sizes on the use and state of one of Australia's most important rangelands, the Queensland mulga rangeland. Regression analysis of cross sectional data reveals significant correlations between property size, stocking rate and degradation. These correlations are confirmed in a normative stochastic dynamic programming model which demonstrates that it is economically optimal for graziers managing smaller properties to adopt higher stocking rates. For these graziers, the longterm costs of land degradation are exceeded by short-term financial benefits of heavier stocking. Thus government policy aimed at arresting the serious degradation occurring in the mulga rangelands should focus on measures to facilitate property build-up..