Presentation and interpretation of chemical data for igneous rocks

Arguments are made in favor of using variation diagrams to plot analyses of igneous rocks and their derivatives and modeling differentiation processes by least-squares mixing procedures. These methods permit study of magmatic differentiation and related processes in terms ofall of the chemical data available. Data are presented as they are reported by the chemist and specific processes may be modeled and either quantitatively described or rejected as inappropriate or too simple.

Examples are given of the differing interpretations that can arise when data are plotted on an AEM ternary vs. the same data on a full set of MgO variation diagrams. Mixing procedures are illustrated with reference to basaltic lavas from the Columbia Plateau.