Institution: | Michael Ed. Hohn, Organic Geochemistry Unit, School of Chemistry, University of Bristol, Bristol, England La Verne M. Friberg, Department of Geology, University of Akron, Akron, Ohio 44325, U.S.A. |
Abstract: | Principal components analysis of elemental distributions among minerals elucidate chemical, thermal, and equilibrium effects. Previous approaches in applying this multivariate statistical method to petrologic data are united into a single model, so-called three-mode principal components analysis. The major advantage of the model lies in the separate analysis of each of the three modes: element, mineral, and sample. The three sets of principal components are related through a “core matrix”. The model accommodates the ordination of pair-wise combination of the modes, such as samples and minerals in visualization of multidimensional tie-lines. Because of its generality, the model permits great flexibility in the study of petrologic data. |