Mantle-derived amphibole within inclusions in alkalic-basaltic lavas

Amphibole in mantle-derived inclusions in alkalic-basaltic rocks occurs in four textural modes: veins and interstitial grains in chromian-spinel peridotite, poikilitic grains in igneous textured inclusions, and megacrysts. Over 90 available chemical analyses demonstrate that compositions vary widely, from Fe-Ti-rich kaersutite to richterite-pargasite solid solutions containing high concentrations of Na, Cr, and Mg. No particular composition occurs more frequently than any other. Amphiboles in the uppermost mantle are produced by precipitation of hydrous fluids that migrated upward from deeper hotter regions of the mantle where they originated by partial melting in the low-velocity zone and possibly by other but poorly understood processes of ‘dewatering’ the mantle. These fluids are conceived to have a wide range of composition, depending upon the degree of initial partial melting and the amount of mantle peridotite through which they traveled and scavenged incompatible elements. Further diversification could occur in the zone of precipitation in the uppermost mantle owing to complex reactions with the chromianspinel peridotite wall rock. Vein and poikilitic amphiboles crystallize from fluids of probable nephelinitic to basanitic composition arrested during ascent into cooler parts of the uppermost mantle. Relatively larger masses in the same composition range crystallize as varied assemblages of pyroxenes, olivine, and Cr-poor spinel enclosed by late-formed amphibole. Fractionation in these masses provides another mechanism for producing diverse compositions in potential amphibole-forming fluids. Megacrysts represent fragmented veins and regions without anhydrous crystal accumulation in possible cumulate bodies. In contrast to vein and poikilitic occurrences, interstitial amphiboles appear to be in chemical equilibrium with their surrounding host assemblages of anhydrous phases. But, again, chemical compositions vary, correlating somewhat with the particular host phase assemblage and suggesting interaction between wall rocks and small volumes of hydrous, perhaps even aqueous, fluid. Certain textural relations suggest that reactions such as spinel + diopside + aqueous fluid → amphibole may have operated. In a dynamic (flowing) mantle, superposition of plastic deformation, recrystallization, partial melting, and invasion by potential amphibole-forming fluids will make the history of a mantle-derived peridotite sample very complex. These factors, combined with the variety of variables influencing the compositions and manner of crystallization of amphibole, make the meaning of statements on‘primary’ amphibole in the uppermost mantle difficult to evaluate.