Digital processing for partitioning a composite population into constituent single populations of known distribution type was attempted with a microcomputer on the basis of nonlinear least squares method.
The technique of partitioning was applied to two examples of actual geochemical exploration data, each of which was inferred to be composed of two single and roughly log-normal populations.
The results show that the method was quite satisfactory for providing basic and sound information for the interpretation of composite geochemical data, when applied together with the determination of statistical frequency distribution described in the previous report (OTSU et al., 1983).
