The fissure eruptions of Miocene basalt in the Columbia River region and the even more voluminous Eocene basalts of the Olympic Mountains—Oregon Coast Range occupy two different tectonic settings. Each of these great piles of “tholeiitic” lavas shows almost no differentiation, and rocks of intermediate (andesitic) composition are generally absent. In the adjacent “first cycle” Cascade Mountains, about 75 per cent of the volcanic rocks are andesites, though they range from olivine basalts to rhyolites. Violent steam explosions characterized their extrusion. The rocks are notably higher in CaO and Al2O3 than those of the adjacent basaltic provinces. Rejuvenated or “second cycle” mountains,—the Rocky Mountains of Colorado—show a highly diversified volcanic sequence, including rhyolites, trachytes, latites, andesites, basalts, and also feldspathoidal lavas. Larsen’s studies of the San Juan area prove the coexistence and comingling of magmas of widely different compositions in spongy subterranean chambers. Tertiary igneous sequences occur in the elevated continental plates on both sides of the Rockies. The laccolithic complexes of the Colorado Plateau and the diorite-syenite-shonkinite (or lamprophyre) complexes found east of the Rockies resemble the San Juan sequence but are more alkalic. Still more alkalic are the highly feldspathoidal and ferromagnesian extrusives of the Navajo-Hopi country, which in their general features resemble the highly alkalic and ultramafic volcanic rocks of Uganda. Such alkalic mafic lavas commonly appear in small volume on the Precambrian shields. Common associates are intrusive complexes of alnoite, kimberlite, and mica peridotite. Also present on some shields are voluminous piles of tholeiitic basalt. The igneous cycle has ended; a new one begins. The widely held theory that all volcanic rocks originate by differentiation of basalt does not seem to fit this complex picture. The following hypothesis, admittedly oversimplified, is tentatively offered: Tholeiitic magma arises in a deep universal earth shell, but the fissure eruptions do not tap a great molten pool undergoing crystal differentiation. Differentiation occurs only after irruption to high levels. When a submarine trough is downbuckled into a tectogene, water plus other easily removable constituents are distilled from the metamorphosing geosynclinal sediments. These mix with the primitive tholeiite, modifying it to highly explosive steam-rich andesitic magma. Continued metamorphism of the tectogene root ultimately results in partial melting, causing the rise of trondjhemitic and granodioritic batholiths which may invade the slightly older extrusive andesites. Successive periods of anatexis will sweat upward more and more of the granitic fractions of the tectogene. Left behind in the dwindling root, the refractory greenstones, limestones, and serpentines have meanwhile been dehydrated by increasing metamorphism to amphibolites, skarns, and uralitized or biotized ultramafics. Locally these refractory hornblende- and biotite-rich rocks may be partially melted and erupted to produce the highly alkalic mafic lavas that characterize shields and continental plates.