Two standard tests for soil stability in water involve assessment of the readily dispersible clay (RDC) content of predried soil samples. It is shown that this predrying limits the applicability of the tests to very dry soils. Ways to avoid this limitation are developed. Undisturbed soil samples were collected from the 10- to 15-cm depth of arable topsoils at Villamblain and Faux Perche and from the 0- to 5-cm depth from plowed and direct-drilled plots at Boigneville. These soils in north-central France were collected from the field when moist (near field capacity). A range of water contents and corresponding pore water potentials (or suctions h) was produced by slow drying in air. For moist soils at field capacity (h = 100 hPa) the amount of RDC was large. Readily dispersible clay decreased progressively with increasing antecedent suction, h For soil samples dried to beyond h > 10 MPa, the RDC was reduced to small values. With such intensely dried soils, rewetting for up to 2 wk did not reverse the effect on RDC. At h = 4500 hPa, re-equilibration to h = 100 hPa reversed the effect of this moderate drying on RDC. Two standard conditions (moist and dry) are proposed for measurement of clay dispersion from soil into water. The moist condition is relevant to soil at the end of a long, wet period (e.g., at the end of the winter in temperate climates), whereas the dry condition is relevant to dry surface crusts and dry topsoils. We conjecture that the soil “resets” during long, wet periods to values of stability and RDC at equilibrium with field capacity. Experimental procedures used to determine the content of RDC in moist and dry soils are described.