The combination of multi-element neutron activation analysis and multivariate statistics for characterisation in geochemistry

Discriminant analysis, a statistical method of pattern recognition, was applied to the study of mineralisation in granites, to the classification of diamonds, to the identification of sedimentary units from the Witwatersrand and to the classification of coals from the Witbank Coalfield in South Africa. The results show that the methods can be used to identify and map the mineralised phase of the granite. In the case of the diamonds trace element signatures were found which were characteristic of the sources and these were used to classify very pure unknown samples with a high rate of success. In both the coal and particularly in the case of the Witwatersrand sediments these methods showed themselves to be a powerful tool for the identification of sedimentary units. (author)