In igneous petrology, a wide variety of chemical variation diagrams are used to portray variations of chemical composition within and between rock suites. Unfortunately, only a small proportion of these diagrams elucidate the processes responsible for this chemical diversity. Beginning with the basic concepts of material transfer, general equations are derived which provide a theoretical basis for the development and use of chemical variation diagrams. These equations provide a means to evaluate existing chemical variation diagrams and to develop new diagrams which (i) test whether chemical compositions define a cogenetic rock suite, (ii) determine (or at least constrain) the stoichiometry of any material transfer process, or (iii) discriminate between tectono-magmatic rock suites. Specifically, equations are derived which relate changes in the composition of a system to the stoichiometry of the material transfer process affecting the system. Analytical expressions for general material transfer processes are simplified by assuming a constant process stoichiometry. The expressions for slopes and intercepts of chemical trends are expressed in terms of critical geochemical variables such as the initial system composition and the process stoichiometry. Relationships between simple material transfer processes such as crystal fractionation and the corresponding compositional trends are examined on a variety of chemical variation diagrams. This analysis demonstrates that it is possible to determine the relative changes in extensive variables caused by material transfer processes from an examination of the corresponding intensive compositional variables which are measurable. This is because ratios of intensive variables are exactly equal to the corresponding ratios of extensive variables. Where the denominator of the ratio behaves as a conserved constituent during the process, concentration data plotted as ratios (Pearce element ratios) define trends with slopes that reflect the relative changes in the corresponding extensive variables and indicate the stoichiometry of the process.