The problem of predicting the possible topologies of a geochemical phase diagram, based on the chemical formula of the phases involved, is shown to be intimately connected with and aided by well-studied notions in discrete geometry: Gale diagrams, triangulations, secondary fans, and oriented matroids.
