The shape, arrangement, and great differences in age of the intrusive complexes of the White Mountain magma series (of Jurassic to Cretaceous age) indicate an origin from discrete magma chambers, probably formed within the upper mantle. Initially basaltic, the melt of these closed chamb ers soon acquired a capping of granitic magma, formed by selective melting, as the chambers ascended through the crust. W h en the polymagmatic chambers reached the upper part of the crust, the granitic cap was fully developed and the underlying melt had become syenitic through assimilation and differentiation. As cumulate material settled, the chamber floor was built upward, and the volume of melt was greatly reduced. Ring dikes and stocks formed a complicated sequence after the chambers reached roughly their diametral distance from the Earth’s surface. Some ring dikes may have formed along paraboloidal fractures, but the steep dips of the ring dikes and evidence of repeated subsidence imply formation along cylindrical fractures. A new theory of the mechanism of ring dike formation, involving cone fracturing followed by stoping and subsidence, is explained. It accounts for the later interior stock so common in ring complexes. The model of a floored chamber with coexisting syenitic and granitic melts explains why the subsiding block stops, why mafic ring dikes are rare, why granite is found more commonly in stocks and syenite more commonly in ring dikes, why quartz syenite cuts granitic rock, why rhyolite appears at the surface before quartz syenite forms in associated ring dikes, and why rhyolite dominates the associated volcanic rocks. The model should remind us that petrographic similarity is not an infallible criterion of age, and it suggests the probability that certain rocks of the White Mountain magma series are products of mixed magmas.