SUMMARYDisjunctive kriging is a technique for estimating values of spatially distributed random variables by combining data non‐linearly and with minimum variance. Data are transformed using Hermite polynomials to a normal distribution, and the resultant transformation is assumed to produce a stationary bivariate normal distribution for all pairs of data. Conditional probabilities that the true values exceed or are less than a specified critical threshold can then be calculated. For soil survey both the estimated values of soil properties and their associated probabilities can be mapped isarithmically as aids to decisions on land management.The technique is described and illustrated from two case studies. The pH, exchangeable potassium and available phosphorus are mapped over 77 ha of the Broom's Barn Farm, and in the eastern Border Region of Scotland the available copper and cobalt in the topsoil are mapped. The estimates are similar to those obtained by simple (linear) kriging. The conditional probabilities were estimated for values falling short of the minimum recommended thresholds, and maps of them showed where farmers would be advised to remedy or forestall deficiencies in lime and plant nutrients and to take prophylactic measures for their livestock.