American Mineralogist, Volume 86, pages , 2001 LETTER

Transcription

1 American Mineralogist, Volume 86, pages , 2001 LETTER Feldspar thermometry in ultrahigh-temperature metamorphic rocks: Evidence of crustal metamorphism attaining ~1100 C in the Archean Napier Complex, East Antarctica TOMOKAZU HOKADA* National Institute of Polar Research, Kaga, Itabashi-ku, Tokyo , Japan ABSTRACT Ultrahigh metamorphic temperatures attained in the mid- to lower-crust have been assessed by examining the mineral chemistry of ternary feldspars with relatively coarse exsolution lamellae from the Archean Napier Complex, East Antarctica. Chemical compositions of re-integrated perthitic, mesoperthitic and antiperthitic feldspars are calculated from the modal proportions and the chemical analyses of host and lamellar domains formed through exsolution. Based on ternary feldspar solvus models, re-integrated compositions of feldspars from a variety of rock types yield the minimum equilibrium temperatures ranging from 1000 to 1110 C (0.8 GPa). These data confirm the suggestion that the regional thermal conditions of the Napier Complex reached or exceeded 1100 C. As feldspar is one of the common constituents of the crustal rocks, this approach could be applicable to a wide variety of rocks in which feldspar represents exsolution textures. INTRODUCTION Peak conditions of high-temperature metamorphism are commonly estimated by using geothermobarometers or by examining mineral parageneses. Ultrahigh-temperature (UHT) metamorphism, which is defined as crustal metamorphism that occurred at peak conditions in excess of 900 C at mid- to deepcrustal levels (Harley 1998), of the Archean Napier Complex in East Antarctica is indicated by the mineral parageneses sapphirine+quartz and/or osumilite (Dallwitz 1968; Ellis et al. 1980; Grew 1980, 1982; Motoyoshi and Hensen 1989) in Mgrich rocks and inverted metamorphic pigeonite (Grew 1982; Harley 1987; Sandiford and Powell 1986, 1988; Harley 1998) in Fe-rich rocks. However, these diagnostic mineral parageneses are not common and, in addition, they do not give the actual temperatures attained, but only indicate that temperatures exceeded the lower temperature stability limits of the parageneses. Fe-Mg cation exchange geothermometers also have been applied to UHT metamorphic rocks (Harley 1983, 1985, 1998), but they give only closure temperatures (<1000 C) of intracrystalline diffusion. Harley and Motoyoshi (2000) estimated the peak temperatures of a distinctive sapphirine quartzite from the Napier Complex as evidence for >1120 C based on orthopyroxene containing up to 12 wt% Al 2 O 3. Nonetheless, they noted that it remains to be demonstrated that these extremely high metamorphic temperatures were achieved over a regional scale in the Napier Complex. Given the relative scarcity of critical assemblages, information constraining the maximum temperatures attained dur- * Present address: Department of Geology, National Science Museum, Hyakunin-cho, Shinjuku-ku, Tokyo , Japan. hokada@kahaku.go.jp ing the UHT metamorphism from a wide variety of common rock types is needed. Feldspars are ubiquitous in UHT metamorphic rocks, not only orthogneisses but also paragneisses of the Napier Complex and they commonly show perthitic, antiperthitic, or mesoperthitic exsolution textures (e.g., Harley 1985; Sheraton et al. 1987). Moreover, some of mesoperthitic feldspars are characterized by high anorthite contents (up to 17 mol%; Sandiford 1985) suggesting extremely high equilibrium temperatures. Similar compositions of mesoperthitic feldspars representing >1000 C also have been reported from high-temperature granulite xenoliths in central Mexico (Hayob et al. 1989). The present study evaluates feldspar geothermometers that are applicable to UHT conditions and calculates equilibrium temperatures recorded in the hypersolvus ternary feldspars of the Napier Complex. GEOLOGICAL SETTING The Napier Complex occupies an area of 400 km 200 km in Enderby Land, East Antarctica (Fig. 1). It consists of granulite-facies metamorphic rocks formed by multiple thermal events, including Late-Archean UHT metamorphism (Harley and Black 1997), and was subsequently intruded by unmetamorphosed diabase dikes at 1.2 Ga (Sheraton and Black 1981). Analyzed samples were collected from Mt. Riiser-Larsen and Tonagh Island, both located in the highest-grade region characterized by the occurrence of sapphirine+quartz paragenesis in Mg-rich rocks (Fig. 1). Orthopyroxene-bearing felsic gneiss, garnet-bearing felsic gneiss, and two-pyroxene-bearing mafic granulite are dominant at both localities, whereas metapelitic and ultramafic gneisses are minor (Ishizuka et al. 1998; Osanai et al. 1999) X/01/ $

2 HOKADA: FELDSPAR THERMOMETRY IN UHT METAMORPHIC ROCK 933 FIGURE 1. Geologic sketch map of the Napier Complex and surrounding area in East Antarctica. The estimated boundary between the Napier Complex and the Rayner Complex is adapted from Sheraton et al. (1987). The Sapphirine+quartz (Spr+Qtz) in isograd is taken from Harley and Hensen (1990). ANALYTICAL METHOD Two-feldspar geothermometers are based on the temperature dependence of the solvus composition between coexisting alkali feldspar and plagioclase. The thermometers can be applied not only to two-feldspar pairs, but also to single hypersolvus feldspars, the compositions of which give minimum temperatures. Recent two-feldspar geothermometers (Fuhrman and Lindsley 1988; Lindsley and Nekvasil 1989; Elkins and Grove 1990) depend on thermodynamic modeling of feldspar in the ternary (CaAl 2 Si 2 O 8 -NaAlSi 3 O 8 -KAlSi 3 O 8 ) system. Kroll et al. (1993) reviewed the experimental and thermodynamic basis of the two-feldspar geothermometer and summarized the problems that occur when applying it to natural rocks. Area scan electron microprobe analyses of fine-scale exsolution textures have usually been applied to obtain preexsolution original compositions of feldspars (e.g., Harley 1985; Sandiford 1985; Sheraton et al. 1987). However, such a method may involve systematic matrix-effects error; the different compositions during the area scan require different absorption, fluorescence, and atomic number corrections (Bohlen and Essene 1977; Raase 1998). Instead, compositions of re-integrated perthitic, mesoperthitic, and antiperthitic feldspars from the Napier Complex were calculated from chemical analyses of homogeneous host and lamellae domains and their areal proportions. The analyses were obtained with a JEOL JXA-8800M wavelength-dispersive electron microprobe at the National Institute of Polar Research. Si, Ti, Al, Cr, Fe, Mn, Mg, Ca, Na, K, and Ba were measured with a focused <1 µm beam at 15kV and a beam current of 8 na, conditions under which Na is not evaporated. Analyses with the highest Ca and K contents were chosen for the chemical composition of plagioclase and K-feldspar domains of the exsolution, respectively, to avoid analyses affected by beam overlap between the host and the lamellae domains. No significant compositional heterogeneity of the analyzed domains was found. Areal proportions of host and lamellae were estimated by computer image analysis of backscattered electron images (Fig. 2). Measurements were made on thin section surfaces and it was assumed that the measured areas of lamellae and host corresponded to their volumes. A maximum of 4 domains ( µm 2 each) were selected in each feldspar grain, and the numbers of pixels in the lamellae and host were calculated. The volume proportions of lamellae and host domains were converted to weight percentages using densities (2.67 g/cm 3 for plagioclase and 2.57 g/cm 3 for alkali feldspar domains, respectively; Smith 1974). These densities were used for each host and lamellae domain of plagiocase and alkali feldspar to calculate re-integrated weight composition of feldspar grain. The bulk molar compositions of feldspar were obtained from the weight percentages and the chemical analyses of each domain. In some cases, very fine-scale lamellae of secondary albite are formed in the K-feldspar domain (Kroll et al. 1993; Raase

3 934 HOKADA: FELDSPAR THERMOMETRY IN UHT METAMORPHIC ROCK FIGURE 2. Back-scattered electron images of feldspars in UHT gneisses. Mineral abbreviations are as follows: Grt = garnet, Opx = orthopyroxene, Pl = plagioclase, Qtz = quartz, P = perthite, M = mesoperthite, and A = antiperthite. (a) Antiperthitic plagioclase containing exsolved rods of K-feldspar in orthopyroxene felsic gneiss (sample 10807). (b) Perthitic alkali feldspar with strings of plagioclase lamellae in garnet felsic gneiss (sample 20516). (c) Mesoperthitic alkali feldspars and plagioclase (no exsolution lamellae) from garnet-porphyroblastbearing felsic gneiss (sample 12006). (d) Sapphirine-bearing aluminous gneiss (A1106A) from Tonagh Island contains mesoperthite and recrystallized alkali feldspar with perthitic lamellae and plagioclase with or without antiperthitic lamellae. TABLE 1. Representative electron microprobe analyses of the host and lamelae domains of feldspar and estimations of the re-integrated feldspar compositions Sample R2301B A1106A Texture antiperthitic feldspar perthitic feldspar mesoperthitic feldspar Domain host lamellae int. host lamellae int. * * int. wt% (Pl) (Kfs) (Kfs) (Pl) (Pl) (Kfs) SiO Al 2 O Fe 2 O CaO Na 2 O K 2 O BaO Total Formula proportions of cations based on 8 O atoms Si Al Fe Ca Na K Ba Total An Ab Or Note: int. = re-integrated composition. * Each domain occurs in almost equal proportions.

4 HOKADA: FELDSPAR THERMOMETRY IN UHT METAMORPHIC ROCK 935 TABLE 2. Re-integrated compositions of feldspars along with the areal proportions and chemical compositions of exsolution lamellae and host domains Areal Host and lamellae compositions Re-integrated Equilibrium Sample Texture proportions (%) Pl domain (mol%) Kfs domain (mol%) composition (mol%) temperature ( C) Pl Kfs An Ab Or An Ab Or An Ab Or T(FL) T(LN) T(EG) Mt. Riiser-Larsen Orthopyroxene felsic gneiss avg A A A Garnet felsic gneiss avg P P P P P P Garnet porphyroblast-bearing felsic gneiss avg M M M M M Garnet-orthopyroxene-bearing quartzo-feldspathic gneiss avg R2301A A A A A Garnet-orthopyroxene-bearing siliceous gneiss avg R2301B P P Tonagh Island Sapphirine-garnet-orthopyroxene gneiss avg.* A1106A M M M M M M M A A P P Note: A = antiperthitic, M = mesoperthitic, P = perthitic. Temperature estimations at 0.8 Ga are based on the following models; T(FL) = Fuhrman and Lindsley (1988), T(LN) = Lindsley and Nekvasil (1989), T(EG) = Elkins and Grove (1990). * Average temperature excluding antiperthite and perthite. 1998), but these were not observed in the analyzed samples with optical and electron microscopes. SAMPLES AND RESULTS Analyzed samples and results are summarized in Tables 1 and 2 and in Figure 3. Temperatures were calculated for P = 0.8 GPa, which is the pressure estimated for the study areas (Harley and Hensen 1990). Celsian component (BaAl 2 Si 2 O 8 ) in the re-integrated feldspars is <0.5 mol%, and was disregarded for the temperature estimations. Ferric iron is also present in the feldspars but it is negligible in amount (Fe 2 O 3 <0.1 wt%) and was ignored in the calulations. The samples used in this study include orthogneisses (orthopyroxene-bearing felsic gneiss and garnet felsic gneisses) and paragneisses (garnet-orthopyroxene gneisses and sapphirine-bearing gneiss, which may be a sort of metasomatic rock); however, it is not easy to estimate the protolith of the highgrade metamorphic rocks. Not only paragneisses but also orthogneisses were extensively recrystallized to anhydrous mineral assemblages. Moreover, feldspars in both orthogneisses and paragneisses yield similar equilibrium temperatures, as will be described below, and there is, therefore, no doubt that the feldspars discussed in this study being metamorphic phases, not inherited igneous or metaig-neous phases as was suggested by Harley (1998). Detailed sample descriptions are as follows: (1) Orthopyroxene-bearing felsic gneiss (TH , hereafter shortened to 10807) from Mt. Riiser-Larsen is the

5 936 HOKADA: FELDSPAR THERMOMETRY IN UHT METAMORPHIC ROCK FIGURE 3. (a) Ternary plots of re-integrated feldspar compositions for all analyzed samples along with the solvus curves calculated at 0.8 GPa using the model of Fuhrman and Lindsey (1988) for C, and those of Lindsley and Nekvasil (1989) and Elkins and Grove (1990) for 1100 C. (b) Exsolved feldspar compositions from the Mt. Riiser-Larsen plotted on stacked ternary diagrams. Symbols correspond to sample numbers as shown. Isotherm and the tie-lines for 700 C are calculated using the model of Fuhrman and Lindsley (1988). (c) Re-integrated feldspar compositions from the Mt. Riiser-Larsen. Isotherms are shown at 0.8 GPa for C. (d) Exsolved feldspar compositions from Tonagh Island plotted as in b. (e) Re-integrated feldspar compositions from Tonagh Island plotted as in c. The tie-lines for 900 C are also calculated using Fuhrman and Lindsley (1988). Both re-integrated and further back-calculated perthite-antiperthite pair compositions are shown (see text for details).

6 HOKADA: FELDSPAR THERMOMETRY IN UHT METAMORPHIC ROCK 937 most common rock type in the Napier Complex. Major constituent minerals are quartz, plagioclase (antiperthite, Fig. 2a), and orthopyroxene. Antiperthitic plagioclase is the only feldspar phase in the rock. Exsolution lamellae are restricted to the core of the feldspar grains, and the K component in the rim portion is interpreted to have migrated to the rim of the crystal by retrograde diffusion. Therefore, the re-integration was applied only for the core of feldspar grains, and the re-integrated compositions show minimum equilibrium temperatures ranging from 1000 to 1040 C [Fig. 3c, T(FL) in Table 2 is used hereafter, as will be discussed in the next section]. (2) Garnet-bearing felsic gneiss (TH , TH , TH , hereafter shortened to 20515, 20516, 20518) from Mt. Riiser-Larsen, which is also a widespread rock type, is composed mainly of quartz, alkali feldspar (perthite, Fig. 2b), and garnet. Perthite occurs as the only feldspar phase, and it contains thin film-like, Na-rich exsolution lamellae. Re-integrated chemical compositions of perthitic feldspar show minimum equilibrium temperatures of C (Fig. 3c). (3) TH (hereafter shortened to 12006) is the garnetporphyroblast-bearing portion of a garnet-bearing felsic gneiss from Mt. Riiser-Larsen. This sample is composed of garnet porphyroblasts up to 5 cm across, plagioclase (no exsolution lamellae were found, as will be discussed below), and alkali feldspar (mesoperthite, Fig. 2c) with a trace amount of quartz. The mode of occurrence and texture of mesoperthitic alkali feldspar is heterogeneous, and the re-integrated chemical compositions suggest minimum equilibrium temperatures of C (Fig 3c). No significant compositional difference is found between the chemical compositions of lamellae-free plagioclase and those of exsolved plagioclase lamellae in mesoperthite. (4) Garnet-orthopyroxene-bearing gneisses from Mt. Riiser- Larsen are well-layered and quite heterogeneous. A quartzofeldspathic layer (R A, hereafter shortened to R2301A) is composed of quartz, plagioclase (antiperthite), garnet, and orthopyroxene, whereas another siliceous sample (R B, hereafter shortened to R2301B) is composed of quartz and garnet with minor amounts of orthopyroxene and alkali feldspar (perthite). Re-integrated chemical compositions of antiperthitic feldspar in R2301A indicate C and those of perthitic feldspar in R2301B yield C. (5) Sapphirine-garnet-orthopyroxene gneiss (A A, hereafter shortened to A1106A) forms a layer several centimeters to a few meters thick around an ultramafic lens on Tonagh Island. The major constituents are sapphirine, garnet, orthopyroxene, plagioclase (antiperthite), and alkali feldspar (mesoperthite and perthite, Fig. 2d). Re-integrated compositions of mesoperthitic feldspar suggest C (open circles in Fig. 3e), and those of antiperthitic and perthitic feldspars yield C (open squares in Fig. 3e). These results imply that the antiperthite-perthite pair recrystallized from the breakdown of mesoperthite during a retrograde stage of metamorphism. However, antiperthite-perthite compositions (open squares in Fig. 3e) do not agree with the calculated tielines. This is interpereted to result from retrograde interdomain K- Na exchange after the formation of exsolution lamellae (Kroll et al. 1993), and original equilibrium compositions of the lamellae and host domains are inferred (small closed squares in Fig. 3e). FELDSPAR GEOTHERMOMETERS FOR UHT CONDITIONS Temperature estimates from the re-integrated feldspars are slightly different when using the geothermometers of Fuhrman and Lindsley (1988), Lindsley and Nekvasil (1989), and Elkins and Grove (1990); nevertheless, average temperatures estimated for each rock type fall within the range of C (Table 2). Although it is not easy to estimate which geothermometer is better to use for the UHT conditions, the model of Fuhrman and Lindsley is preferred for this study based on the following reasons. The Ca-(Na,K) exchange along with Al-Si in the feldspar crystal will be inhibited (or very slow if it occurs), whereas K-Na exchange will be much easier (e.g., Grove et al. 1984). Therefore, a trace amount of the KAlSi 3 O 8 component in plagioclase will diffuse away from the crystal before the exsolution lamellae precipitates during cooling from the peak conditions. To put it another way, the precipitation of antiperthitic lamellae in plagioclase will be postponed to slightly after the cooling from the peak metamorphism, and the re-integrated compositions of plagioclase probably yield slightly lower temperatures than the peak conditions. This is evidenced by the occurrence of lamellae-free grains or domains (rim portion) of plagioclase (Figs. 2a and 2c). In contrast, even though a high amount of the CaAl 2 Si 2 O 8 component cannot remain dissolved in alkali feldspar, it doesn t diffuse outward to grain boundaries but precipitates as exsolution lamellae just after the peak conditions. For these reasons, perthitic to mesoperthitic feldspar should yield higher temperature than antiperthitic feldspar. The Fuhrman and Lindsley model yields a lower temperature for antiperthitic feldspars (average C) than for perthitic to mesoperthitic feldspars (average C), and this model will be preferable. Whereas the models of Lindsley and Nekvasil and of Elkins and Grove yield the higher temperatures for antiperthitic feldspars (average and C, respectively) than for alkali feldspars ( C and C). THERMAL PEAK CONDITIONS OF THE NAPIER COMPLEX Re-integrated compositions of feldspars in the garnetporphyroblast-bearing gneiss (12006) and the garnetorthopyroxene gneiss (R2301A) from the Mt. Riiser-Larsen, and the sapphirine-bearing gneiss (A1106A) from Tonagh Island yield equilibrium temperatures ranging from 1050 to 1110 C, suggesting that the thermal peak conditions of both Mt. Riiser-Larsen and Tonagh Island were near 1100 C. These estimated values are the closure temperatures of intracrystalline diffusion in the feldspar grain and, hence, a minimum temperature estimate for the thermal peak conditions. Temperatures obtained here are similar to the recent estimate of the peak metamorphic conditions of >1120 C for Alrich orthopyroxene in a sapphirine-quartzite from Mt. Riiser-Larsen (Harley and Motoyoshi 2000). However, their estimate was based on a single sample. Harley (1998) summarized the other quantitative temperature estimations for the Napier Complex. Specifically, metamorphic pigeonite and subcalcic pyroxene from meta-ironstones imply C and garnet-orthopyroxene geothermometry yields C. These temperatures are consistent with the mineral assemblage constraints such as sapphirine+orthopyroxene+quartz,

7 938 HOKADA: FELDSPAR THERMOMETRY IN UHT METAMORPHIC ROCK garnet+orthopyroxene+sillimanite+quartz, spinel+quartz, and osumilite+garnet. The present study reports the evidence of regional metamorphic conditions in which temperature reached 1100 C from rocks of different composition from two widely spaced localities (the Mt. Riiser-Larsen and Tonagh Island are about 50 km away from each other) from the highest-grade region of the Napier Complex. Moreover, the results presented here suggest that feldspar compositions could be a useful indicator for evaluating the thermal peak conditions of UHT metamorphism. The recrystallized perthite-antiperthite pairs from Tonagh island yield approximately 940 C, and this temperature is interpreted as the conditions of retrograde local recrystallization event. Exsolved compositions suggests that the final equilibrium conditions were below 700 C (Figs. 3b and 3d). ACKNOWLEDGMENTS The rocks described in this paper were collected during and Japanese Antarctic Research Expedition (JARE). The members of JARE, particularly H. Ishizuka, M. Ishikawa, S. Suzuki, Y. Osanai, T. Toyoshima, M. Owada, T. Tsunogae, and W.A. Crowe, are thanked for their support during the field work. The earlier draft was reviewed and improved by Y. Motoyoshi and K. Shiraishi. The constructive reviews by E. Essene, E.S. Grew, and R.F. Dymek improved the paper considerably. This work was partly supported by a Grant-in-Aid from the Ministry of Science, Sports, and Culture, Japan, to K. Shiraishi (no ). Thanks are also due to them. REFERENCES CITED Bohlen, S.R. and Essene, E.J. (1977) Feldspar and oxide thermometry of granulites in the Adirondack Highlands. Contributions to Mineralogy and Petrology, 62, Dallwitz, W.B. (1968) Coexisting sapphirine and quartz in granulites from Enderby Land, Antarctica. Nature, 219, Elkins, L.T. and Grove, T.L. (1990) Ternary feldspar experiments and thermodynamic models. American Mineralogist, 75, Ellis, D.J., Sheraton, J.W., England, R.N., and Dallwitz, W.B. (1980) Osumilitesapphirine-quartz granulites from Enderby Land, Antarctica: mineral assemblages and reactions. Contributions to Mineralogy and Petrology, 72, Fuhrman, M.L. and Lindsley, D.H. (1988) Ternary-feldspar modeling and thermometry. American Mineralogist, 73, Grew, E.S. (1980) Sapphirine + quartz association from Archaean rocks in Enderby Land, Antarctica. American Mineralogist, 65, (1982) Osumilite in the sapphirine-quartz terrane of Enderby Land, Antarctica: implications for osumilite petrogenesis in the granulite facies. American Mineralogist, 67, Grove, T.L., Baker, M.B., and Kinzler, R.J. (1984) Coupled CaAl-NaSi diffusion in plagioclase feldspar: experiments and applications to cooling rate speedometry. Geochimica et Cosmochimica Acta, 48, Hayob, J.L., Essene, E.J., Ruiz, J., Ortega-Gutiérrez, F., and Aranda-Gómez, J.J. (1989) Young high-temperature granulites from the base of the crust in central Mexico. Nature, 342, Harley, S.L. (1983) Regional geobarometry-geothermometry and metamorphic evolution of Enderby Land, Antarctica. In R.L. Oliver, P.R. James, and J.B. Jago, Eds., Antarctic Earth Science, Australian Academy of Science, Canberra. (1985) Garnet-orthopyroxene bearing granulites from Enderby Land, Antarctica: metamorphic pressure-temperature-time evolution of the Archaean Napier Complex. Journal of Petrology, 26, (1987) A pyroxene-bearing metaironstone and other pyroxene-granulites from Tonagh Island, Enderby Land, Antarctica: further evidence for very high temperature (> 980 C) Archaean regional metamorphism in the Napier Complex. Journal of Metamorphic Geology, 5, (1998) On the occurrence and characterisation of ultrahigh-temperature crustal metamorphism. In P.J. Treloar, and P. O Brien, Eds., What Drives Metamorphism and Metamorphic Reactions? Geological Society of London, Special Publication, p Harley, S.L. and Black, L.P. (1997) A revised Archaean chronology for the Napier Complex, Enderby Land, from SHRIMP ion-microprobe studies. Antarctic Science, 9, Harley, S.L. and Hensen, B.J. (1990) Archaean and Proterozoic high-grade terranes of East Antarctica (40-80 E): a case study of diversity in granulite facies metamorphism. In J.R. Ashworth, and M. Brown, Eds. High-temperature Metamorphism and Crustal Anatexis, p Unwin Hyman, London. Harley, S.L. and Motoyoshi, Y. (2000) Al zoning in orthopyroxene in a sapphirine quartzite: evidence for >1120 C UHT metamorphism in the Napier Complex, Antarctica, and implications for the entropy of sapphirine. Contributions to Mineralogy and Petrology, 138, Ishizuka, H., Ishikawa, M., Hokada, T., and Suzuki, S. (1998) Geology of the Mt. Riiser-Larsen area of the Napier Complex, Enderby Land, East Antarctica. Polar Geoscience, 11, Kroll, H., Evangelakakis, C., and Voll, G. (1993) Two-feldspar geothermometry: a review and revision for slowly cooled rocks. Contributions to Mineralogy and Petrology, 114, Lindsley, D.H. and Nekvasil, H. (1989) A ternary feldspar model for all reasons. Eos, 70, 506. Motoyoshi, Y. and Hensen, B.J. (1989) Sapphirine-quartz-orthopyroxene symplectites after cordierite in the Archaean Napier Complex, Antarctica: evidence for a counterclockwise P-T path? European Journal of Mineralogy, 1, Osanai, Y., Toyoshima, T., Owada, M., Tsunogae, T., Hokada, T., and Crowe, W.A. (1999) Geology of ultrahigh-temperature metamorphic rocks from Tonagh Island in the Napier Complex, East Antarctica. Polar Geoscience, 12, Raase, P. (1998) Feldspar thermometry: a valuable tool for deciphering the thermal history of granulite-facies rocks, as illustrated with metapelites from Sri Lanka. The Canadian Mineralogist, 36, Sandiford, M. (1985) The metamorphic evolution of granulites at Fyfe Hills: implications for Archaean crustal thickness in Enderby Land, Antarctica. Journal of Metamorphic Geology, 3, Sandiford, M. and Powell, R. (1986) Pyroxene exsolution in granulites from Fyfe Hills, Enderby Land, Antarctica: evidence for 1000 C metamorphic temperatures in Archaean continental crust. American Mineralogist, 71, (1988) Pyroxene exsolution in granulites from Fyfe Hills, Enderby Land, Antarctica: evidence for 1000 C metamorphic temperatures in Archaean continental crust Reply. American Mineralogist, 73, Sheraton, J.W. and Black, L.P. (1981) Geochemistry and geochronology of Proterozoic tholeiite dykes of East Antarctica: evidence for mantle metasomatism. Contributions to Mineralogy and Petrology, 78, Sheraton, J.W., Tingey, R.J., Black, L.P., Offe, L.A., and Ellis, D.J. (1987) Geology of Enderby Land and western Kemp Land, Antarctica. 51 p. Australian Government Publishing Service, Canberra. Smith, J.V. (1974) Feldspar Minerals. Volume 1. Crystal structure and physical properties, 625 p. Springer-Verlag, Berlin. MANUSCRIPT RECEIVED OCTOBER 30, 2000 MANUSCRIPT ACCEPTED MARCH 10, 2001 MANUSCRIPT HANDLED BY ROBERT F. DYMEK