The tensile strengths of remoulded samples of five Australian soils with differing clay type, texture and shrink-swell potential were measured as a function of exchangeable cations (Na, Ca and Mg) and exchangeable sodium percentage (ESP). Spontaneously and mechanically dispersible clays were also determined as a function of ESP. The tensile strength changed with the nature of the exchangeable cation, clay content and amounts of spontaneously and mechanically dispersible clay. In Ca-soils, the tensile strength was highly correlated with clay content and CEC. Regression analyses of data for soils containing various amounts of exchangeable sodium showed that mechanically and spontaneously dispersible clay were individually correlated with the tensile strength of remoulded soils. However, multiple regression analyses of these data indicated that spontaneously dispersible clay alone was a major predictor of the tensile strength of remoulded sodic soils. This suggests that measurement of spontaneously dispersible clay adequately accounts for the differences in tensile strengths of dry remoulded soils as influenced by ESP values. Analysis of variance of data for all the soils with varying ESP values showed that spontaneously dispersible clay was strongly correlated with clay content. Analyses of data for individual soil type showed that spontaneously dispersible clay was highly correlated with ESP. For each soil studied, increase in ESP resulted in increase of dispersible clay and hence in tensile strength. Although tensile strength increased with ESP, the rate of change of strength with ESP was different for each soil. Soil with the highest clay content gave rise to the greatest rate of change. The effect of exchangeable magnesium on tensile strength was similar to calcium. However, in two of the soils, exchangeable magnesium, in the presence of sodium, increased the strength slightly more than calcium, confirming the ionic radius effect of these elements.