Tendency to crust is a potentially useful index for assessing soil degradation and for assisting land use planning in South Africa. In this study, the influence of land use, geology and vegetation cover on the tendency of soil to form a surface crust was investigated in six vegetation types. Crusting at all sites was greater in exposed soils than soils under vegetation, as determined by infiltration rate, water dispersible clay and modulus of rupture. In Renosterveld, crusting was markedly greater in exposed soil than vegetation covered soil (mean infiltration 16 vs 44 mm h(-1); dispersible clay 2.6 vs 2.2%; modulus of rupture 121 vs 64 kPa). Greater crusting in exposed soil was attributed to lower soluble salt and labile carbon (C) contents and an associated increase in the dispersion of clay. In Karoo, crusting of exposed, shale-derived soils was greater than that of exposed, dolerite-derived soils (infiltration 40 vs 83 mm h(-1); dispersible clay 2 vs 1.2%), and a similar pattern was evident in Tall Grassveld (infiltration 18 vs 36 mm h(-1); dispersible clay 1.2 vs 0.9%; modulus of rupture 31 vs 21 kPa). In Upland Grassland, cultivation of maize and rye enhanced crusting. In Thicket, crusting was greater in soils from open, degraded vegetation than intact, densely wooded sites (infiltration 19 vs 51 mm h(-1); modulus of rupture 16 vs 34 kPa), probably due to lower content of soil C. In Bushveld, crusting was greater in annually burnt plots than unburnt plots (infiltration 109 vs 163 mm h(-1); dispersible clay 0.9 vs 0.6% on granite-derived soils; and infiltration 56 vs 72 mm h(-1); dispersible clay 1.5 vs 1.3% on basalt-derived soils). Greater crusting of soil from burnt plots was ascribed to a reduction in soil C and soluble salts as well as a greater exchangeable sodium percentage.