Two-dimensional frequency distribution of geochemical data was determined by computation of two-dimensional cumulative frequency and its partial differentiation with respect to two variables without any assumption on the type of statistical distribution of data.
The technique was applied to actual geochemical exploration data, which proved the superiority of twodimensional distribution to one-dimensional distribution for defining geochemical anomalies and partitioning a composite population into constituting single populations.
Characteristics of two-dimensional normal distribution and apparent correlation in a composite population were also briefly discussed.
