Arc-basaltic melts of typical hydration state (≥ 2 wt% H2O) are buoyant relative to all common crustal igneous rock types and may pass adiabatically from a mantle source through continental crust directly to shallow depths or volcanic vents without significant crystallisation in transit. Basalts are the most common eruptive rocks on continents. So why is the average bulk composition of continental crust “andesitic”, and why is continental crust vertically zoned with “basaltic” compositions at the bottom? In non-compressive segments of continent-margin magmatic belts, the frequency mode for eruptive and hypabyssal (aphanitic groundmass) rocks is ~50 wt% SiO2, and whole-rock Sr/Y falls with rising SiO2. In orogenically deforming segments of continent-margin magmatic belts, the frequency mode of eruptive and hypabyssal rocks is ~60 wt% SiO2, and Sr/Y rises with rising SiO2. Horizontal compression inhibits ascent of buoyant magmas by dyke propagation. Typical values of orogenic horizontal deviatoric stress (~10–30 MPa) can exceed the magma buoyancy force driving dyke propagation and can trap kilometres-thick sills of buoyant magma at all crustal depths. As horizontal compressive stress increases during orogeny, resistance to dyke propagation increases, so more advanced chemical differentiation is achieved at deeper crustal levels. The temperature of fluid-undersaturated country rock at the arc Moho typically is higher than the magma's wet solidus temperature, so residual felsic melts cannot freeze by conductive heat loss to country rocks. Thermally “immortal” residual granitoid magmas may remain stored in Moho-vicinity stress traps as long as strong orogenic compressive stress lasts—5-10 million years or more—and experience multiple replenishments by basaltic melt from the mantle. Ultramafic crystal cumulates produced near the Moho are seismically indistinguishable from regional mantle and are included in the mantle composition inventory, leaving a biased composition inventory of continental crust above the seismic Moho. The volcanic-to-plutonic mass ratio in arc segments is a sensitive function of the tectonic stress regime. Rates of trans-crustal magma transmission should not be misinterpreted as arc magma production rates. “High-flux” pulses of granitoid magmatism in the upper crust tend to substantially lag plate-kinematic indicators of magma production rates in the mantle. Episodes of high-flux granitoid magmatism in the upper crust occur as orogenic compressive stress wanes from peak values, permitting escape of buoyant magmas from long-term storage in stress traps in the deeper crust (or uppermost mantle). Orogenic collapse in “over-thickened” orogens begins at the laterally unconfined highest elevations, and propagates slowly downward, successively tapping stress-trapped magma reservoirs having higher Sr/Y. Whole-rock Sr/Y at 60–70 wt% SiO2 is a better proxy for tectonic stress than for crustal thickness. During fractional crystallisation in intermittently replenished stress traps near the Moho, thermally immortal residual granitoid melts evolve to high (“adakitic”) Sr/Y and exceptional inheritances of dissolved H2O, SO3, and Cl, which endow them with magmatic-hydrothermal copper-ore-forming capability when such melts escape to shallow depths as orogenic stress begins to wane from peak values at Moho-vicinity depths. Improved strategies in the search for porphyry copper ore deposits emerge from understanding the igneous petrogenesis.