This paper essentially introduces a new statistical technique of data analysis together with a practical application. The technique is Correspondence Analysis and it is designed to extract the maximum information from a two dimensional array of positive numbers, such as the concentrations of various elements in several samples. It combines the advantages of R– and Q–mode analysis and also considerably reduces the problems encountered in the practical application of these techniques: Scaling problems and computer costs. The diagrams, which are produced by projecting at the same time samples and variables on the same plane with proper scaling, aid the interpretation of geological processes. The example that is discussed concerns the Bourlamaque batholith, where igneous differentiation and metamorphism have been clearly identified as the major factors controlling its geochemistry. The complete mathematical formulation is given since it has apparently never been published in the geological literature of North America.