High concentrations of exchangeable Mg2+ are a common characteristic of many Australian subsoils. Many subsoils have very low Ca:Mg ratios, however, the effects of Mg2+ on soil physical properties are not well understood. Many authors have investigated how Mg2+ affects the soil, and have studied: hydraulic conductivity and infiltration through leaching columns; soil strength by modulus of rupture or tensile strength of dried plates or discs of soil; and soil dispersibility by submersion of soil in distilled water, or solutions of low electrolyte concentration. While these studies have their merits, the primary focus of this study was the subsoil where there is often little opportunity for the soil to dry, and fluctuations in the soil-water content are small compared with those of the surface soil. For this reason, the effect of Mg2+ on the age-hardening of clay subsoil compared with the effects of Ca2+ and Na+ was studied. Age-hardening is the regain in soil strength with time without an appreciable decrease in the water content of the soil. It was expected that treatment with different exchangeable cations would cause different rates of age-hardening, different soil strengths and aggregate stability. Treatments consisted of moulding three clay subsoils with electrolyte solutions of NaCl, CaCl2 or MgCl2 to a pre-determined water content. The moulded soil was stored at a constant temperature and water content and allowed to age before soil strength and aggregate stability analyses were made. Experimental work showed that the composition of the electrolyte solutions did alter the rate and the overall amount of age-hardening, as determined by soil strength measured with the drop-cone penetrometer. However, the cation composition did not result in the same pattern of change in soil strength for all of the three subsoils. The responses of the three subsoils to moulding with electrolyte solutions were very different, and a mechanism to explain the observed effects was not easy to establish. The subsoils treated with MgCl2 did not consistently show greater soil strength than did the soils treated with CaCl2. Also, the subsoils treated with MgCl2 did not show, with any conviction, that Mg2+ decreases the stability of aggregates in distilled water. Results were unable to conclude whether subsoils treated with MgCl2 caused more spontaneous dispersion in distilled water, when compared with the subsoils treated with CaCl2. Age-hardening due to thixotropy and the reinforcement of particle-particle bonds was confirmed because the strength of the subsoils increased significantly over the experiment for all six treatments, without appreciable decreases in gravimetric water content. Attempts were made to isolate the mechanism behind age-hardening which lead to the conclusion that both thixotropy, and the reinforcement of particle-particle bonds were involved in age-hardening of these subsoils. Age-hardening of aggregates in either an air-dry, intermediate or moist state was not found to increase or decrease either aggregate stability or spontaneous dispersion. It may be that the methods used in this study, in particular the drop-cone penetrometer, were not sensitive enough to find any differences between Ca2+ and Mg2+ on soil physical properties. Any differences in the dispersive potential of exchangeable Ca2+ or exchangeable Mg2+ may be so small that experimental procedures need to very sensitive (and probably unrealistic) to identify these differences.