Mush!

Evidence for emplacement as crystal-laden mushes is abundant in mafic magma systems (i.e. products of broadly basaltic magmatism), including anorthosite complexes, layered mafic intrusions and a variety of sills and dikes. Some of the best examples involve the way feldspar becomes concentrated into anorthositic rocks. Proterozoic anorthosite massifs (e.g. Nain, Rogaland, Adirondacks), whose bulk compositions are characteristically hyperfeldspathic, are best interpreted as resulting from emplacement of plagioclase-rich mushes that ascended to shallow crustal emplacement sites from deep (~Moho) staging chambers in which ~An50 crystals floated due to density relations at high pressure. Supporting evidence includes large (up to ~1 m) grain size, compositional homogeneity of plagioclase with variable Mg# (caused by trapped liquid effects), and protoclastic textures. Isotopic disequilibrium between cumulus plagioclase and post-cumulus pyroxene result from progressive contamination with continental components. This is dramatically demonstrated at Nain (Labrador), where the anorthositic crystal mushes (~1.3 Ga) were emplaced into early Archean (~3.8 Ga) country rocks. High-Al, high-pressure orthopyroxene megacrysts are commonly dragged upward in feldspathic mushes to shallow emplacement sites, where they exsolved plagioclase lamellae (Bybee & Ashwal, this meeting). Archean calcic anorthosites (e.g. Fiskenaesset, W Greenland) and related sills, dikes and flows that contain homogeneous megacrysts up to 10s of cm across of ~An80 also must have formed by mush emplacement, although probably from shallower staging chambers in oceanic rather than continental crust. Many layered mafic intrusions (e.g. Bushveld, Stillwater, Dufek, Duluth) contain thick horizons of anorthosite in which plagioclase compositions are uniform, in some cases throughout >1000 m of stratigraphy. This is best interpreted as representing repeated emplacement of plagioclase-rich mushes from one or more deeper crystallizing magma chambers. In the Bushveld Complex, where deep drill cores have allowed near-continuous measurements of mineral compositions and geophysical properties, the results reveal a subtle cyclicity, invisible in outcrops, over scales of 50 - 100 m, commonly associated with broad reversals in mineral compositional trends. Each of these can be interpreted as a blending zone involving a new addition of crystal-laden magma. Much of Bushveld stratigraphy, and that of other layered intrusions contains plagioclase:pyroxene demonstrably higher than cotectic proportions, supporting overall construction by plagioclase-rich magmas. Isotopic disequilibrium effects, similar to those described above, have been detected in Bushveld cumulates, lending further support to mush emplacement models. Large layered intrusions were probably constructed by repeated emplacement of dozens of individual magmatic entities with variable crystal:melt ratios (Marsh, 2006, Elements). Even the properties of very small (~30 km2) layered intrusions like that at Doros, Namibia (Owen-Smith & Ashwal, this meeting) show abundant evidence for multiple mush emplacement. In plutonic magma bodies arising from mafic magmatism, therefore, the involvement of crystal mushes appears to be the rule rather than the exception.