A method for the experimental determination of the amount of clay dispersed from soil into water is described. The method was evaluated using soil samples from agricultural fields in 18 locations in Poland. Soil particle size distributions, contents of organic matter and exchangeable cations were measured by standard methods. Sub-samples were placed in distilled water and were subjected to four different energy inputs obtained by different numbers of inversions (end-over-end movements). The amounts of clay that dispersed into suspension were measured by light scattering (turbidimetry). An empirical equation was developed that provided an approximate fit to the experimental data for turbidity as a function of number of inversions. It is suggested that extrapolation of the fitted equation to zero inversions enables the amount of spontaneously-dispersed clay to be estimated. This method introduces the possibility of replacing the existing subjective, qualitative method of determining spontaneously-dispersed clay with a quantitative, objective method. Even though the dispersed clay is measured under saturated conditions, soil samples retain a ‘memory’ of the water contents at which they have been stored.