0022-3530/89 $3.00Petrology of Biotite-Cordierite-Garnet Gneiss of theMcCullough Range, Nevada II. P-T-aHlO Pathand Growth of Cordierite During Late Stages ofLow-P Granulite-Grade Metamorphismby EDWARD D. YOUNGDepartment of Geological Sciences, University of Southern California, Los Angeles,California 90089-0740(Received 10 February 1987; revised typescripts accepted 29 July 1988)ABSTRACTThermodynamic calculations based on addition of mass balance equations to the Gibbs Method(Spear, 1986) are used to model the cordierite-producing reaction in pelitic gneiss from theMcCullough Range, southern Nevada. Calculations which treat the model paragenesis as a systemopen to transfer of H2O are consistent with textural relations. Results indicate that cordierite grew bythe continuous net-transfer reaction:076 BIO + 1-72 SILL + 355 QTZ + 0-27 PLG + 0-005 GRT+ O06Al2R2_+Si _, [BIO] ^ 1-00 CRD+1-02 KSP +076 H2O+ O30 FeMg_ ,[CRD] +0.15 FeMg_ ,[BIO] + O0005 FeMg_ ,[GRT]+ O005 CaNaAL ,Si_ ,[PLG]with decreasing P, decreasing T, and increasing aH2O. The steep retrograde dP/dTpath for these low-pressure granulites contrasts with isobaric cooling paths typical of higher pressure granulites, andsuggests uplift and erosion were active during Proterozoic granulite-grade metamorphism in this area.INTRODUCTIONThe importance of continuous net-transfer reactions, which affect changes in mineralcompositions and modal abundances during metamorphism, has been emphasized byLoomis (1975) and Thompson (1976). Characterization of these important reactions hashistorically relied upon interpretation of topologies of simplified projections. More recently,several workers have utilized quantitative approaches to study changes in mode duringmetamorphism (Rice & Ferry, 1982; Thompson, 1982; Thompson et al., 1982; Chamberlain,1986). Korzhinskii (1959) illustrated the advantages of considering both the intensive andextensive factors of state of a simple thermodynamic system. Full understanding of theeffects of continuous net-transfer reactions requires examination of both types of variables inconcert.In this paper, the evolution of independent intensive variables and changes in modeduring the later stages of granulite grade metamorphism of a pelitic gneiss are examined. Thegneiss is part of a Proterozoic low-pressure granulite complex exposed in the McCulloughRange of southern Nevada. The petrology of these rocks is summarized in a companionpaper (Part I, Young et al., 1989). The method proposed by Spear (1986) is used here to[Journal c( Petrology. VoL 30, Part 1, pp. 61-78, 1989] it} Oifon) Uravoihy Preu 1989 at University of California, Los Angeles on October 19, 2011petrology.oxfordjournals.orgDownloaded from62 EDWARD D. YOUNGdemonstrate quantitatively that cordierite in the pelitic gneiss grew by a reaction involvingconsumption of biotite, sillimanite, quartz, and plagioclase, and growth of K-feldspar. It isfurther shown that the cordierite-producing reaction progressed with a substantial decreasein pressure, a small decrease in temperature, and an increase in activity of H2O. These resultsare consistent with textural evidence and published experimental data, and illustrate theusefulness of this approach for interpreting parageneses related by continuous net-transferreactions. The derived P-T path lends a constraint on possible causes for the Proterozoiclow-pressure granulite grade metamorphic event recorded by these rocks.PELITIC GNEISS PARAGENESESThe pelitic gneisses contain the assemblage K-feldspar + plagioclase + quartz + biotite+ garnet+cordierite + sillimanite -I- ilmenite + hercynite + graphite. The petrology and con-ditions of metamorphism are described in Part I. In brief, mineral based geothermobarome-try yields temperatures of 590-750 °C and pressures of ~ 3 to 4 kb. Mineral equilibria recordlow/O2 (^QFM) and aHj0 (^0-26) during at least the later stages of metamorphism.Cores of cordierites poikilitically enclosing biotite and lesser garnet, together with thoseinclusions, comprise part of a paragenesis in these rocks distinct from that denned by othermatrix grains, including cordierite oikocryst rims. The inclusions of biotite ± garnet withincordierite are significantly more magnesian than matrix grains. Biotites enclosed incordierite also have higher Al concentrations relative to matrix biotites. Zoning incordierites devoid of ferromagnesian silicate inclusions is characterized by relatively flatcompositional gradients presumably attributable to partial homogenization by efficientvolume diffusion at high temperatures (Tracy et al, 1976; Woodsworth, 1977; Grew, 1981;Tracy & Dietsh, 1982; Indares & Martignole, 1984; Schreurs & Westra, 1986). In contrast,cordierite oikocrysts exhibit pronounced zoning characterized by significantly higherMg/(Mg + Fe) in cores near inclusions of biotite ± garnet (Fig. 1). The magnesium-richcomposition of the inclusions and adjacent host cordierite (hereafter referred to as inclusionassemblages) cannot be attributed to cation exchange in that they alone comprise the majorferromagnesian silicates. Moreover, an increase in Mg/(Mg + Fe) in all three phases isinconsistent with retrograde cation exchange (cf. Tracy et al., 1976; Schreurs & Westra,1986). These compositions must be related to cordierite rim and matrix grain compositions(hereafter referred to as matrix assemblages) by a continuous net-transfer reaction. Althoughthe shape of zoning profiles in cordierite oikocrysts may have been modified by diffusion, thechange in composition from the inclusion assemblages to the matrix assemblages is a vestigeof zoning incurred during this reaction. Inclusion assemblages reflect conditions whichprevailed during the early stages of cordierite growth. Matrix assemblages record conditionsattained after crystallization of cordierite was complete. The differences in compositions ofphases in these two parageneses are used to deduce the variations in P, T, aHlO, and modalabundances of minerals which occurred during cordierite growth.METHODTheoryThe Gibbs Method (Spear et al., 1982) is a thermodynamic formulism permitting changesin mineral compositions to be related to corresponding variations in intensive factors ofstate. The method involves solving a system of linear differential equations, yielding usefulpartial derivatives. Spear (1986) has shown that the addition of mass-balance equations to at University of California, Los Angeles on October 19,