The late Miocene Cordillera Blanca Batholith lies directly over thick (50 km) crust, inboard of the older Cretaceous Coastal Batholith. Its peraluminous ‘S’ type mineralogy and its position suggest recycling of continental crust, which is commonly thought to be an increasingly important component in magmas inboard of continental margins. However, the peraluminous, apparent ‘S’ type character of the batholith is an artefact of deformation and uplift along a major crustal lineament. The batholith is a metaluminous ‘I’ type and the dominant high-silica rocks (>70%) are Na rich with many of the characteristics of subducted oceanic slab melts. However, the position of the batholith and age of the oceanic crust at the trench during the Miocene preclude slab melting. Instead, partial melting of newly underplated Miocene crust is proposed. In this dynamic model newly underplated basaltic material is melted to produce high-Na, low HREE, high-Al ‘trondhjemitic’ type melts with residues of garnet, clinopyroxene and amphibole. Such Na-rich magmas are characteristic of thick Andean crust; they are significantly different from typical cole-alkaline, tonalite-grano-diorite magmas, and their presence along the spine of the Andes provokes questions about models of trondhjemite genesis by melting of subducted oceanic crust, as well as any generalized, circum-Pacific model involving consistent isotopic or chemical changes inboard from the trench.

▪ Abstract Considerable progress has been made over the past decade in understanding the static rheological properties of granitic magmas in the continental crust. Changes in H₂O content, CO₂ content, and oxidation state of the interstitial melt phase have been identified as important compositional factors governing the rheodynamic behavior of the solid/fluid mixture. Although the strengths of granitic magmas over the crystallization interval are still poorly constrained, theoretical investigations suggest that during magma ascent, yield strengths of the order of 9 kPa are required to completely retard the upward flow in meter-wide conduits. In low Bagnold number magma suspensions with moderate crystal contents (solidosities 0.1 ≤ φ ≤ 0.3), viscous fluctuations may lead to flow differentiation by shear-enhanced diffusion. AMS and microstructural studies support the idea that granite plutons are intruded as crystal-poor liquids (φ ≤ 50%), with fabric and foliation development restricted to the final stages of emplacement. If so, then these fabrics contain no information on the ascent (vertical transport) history of the magma. Deformation of a magmatic mush during pluton emplacement can enhance significantly the pressure gradient in the melt, resulting in a range of local macroscopic flow structures, including layering, crystal alignment, and other mechanical instabilities such as shear zones. As the suspension viscosity varies with stress rate, it is not clear how the timing of proposed rheological transitions formulated from simple equations for static magma suspensions applies to mixtures undergoing shear. New theories of magmas as multiphase flows are required if the full complexity of granitic magma rheology is to be resolved.