Zircon growth in UHT leucosome: constraints from zircon garnet rare earth elements (REE) relations in Napier Complex, East Antarctica

Transcription

1 180 Journal of Mineralogical and Petrological T. Hokada Sciences, and S.L. Volume Harley99, page , 2004 Special Issue Zircon growth in UHT leucosome: constraints from zircon garnet rare earth elements (REE) relations in Napier Complex, East Antarctica Tomokazu HOKADA *,** and Simon L. HARLEY *** * National Institute of Polar Research, Itabashi, Tokyo , Japan ** Department of Polar Science, The Graduate University for Advanced Studies, Itabashi, Tokyo , Japan *** Department of Geology & Geophysics, University of Edinburgh, Edinburgh EH9 3JW, UK Feldspathic leucosomes occur in an ultrahigh temperature (UHT) metamorphosed garnet bearing paragneiss/ quartzite in the Napier Complex, East Antarctica. Coarse ( μm) zircon grains occurring in the leucosomes display 3 textural domains: (I) dark CL (cathodoluminescence) structured inner core, (II) bright CL structured outer core, and (III) dark CL structureless rim. Chemical compositions, especially chondrite normalized REE patterns obtained by SIMS analysis, correlate with these three domains: the inner core (domain I) shows HREE enrichment with Yb(n)/Gd(n) = 3.3, whereas the outer core (II) and rim (III) have flat to relatively depleted HREE patterns with Yb(n)/Gd(n) = Th/U ratios decrease from 3.2 in the zircon inner core (I) to 1.2 in outer core (II) and to 0.3 in the rim (III). Garnets near such zircon grains display two trace element compositional features. Firstly, high core Zr contents (300 ppm) decrease to 100 ppm within μm of grain rims. Secondly, the HREE distribution between zircon inner core (I) and garnet core is 2 at Gd ( Gd D Zrn/Grt = 2), rising to 8 at Lu ( Lu D Zrn/Grt = 8), whereas those defined from zircon outer core (II) or rim (III) and garnet core or rim are much lower and generally below 1 for Gd thruogh to Lu ( Gd D Zrn/Grt = ; Lu D Zrn/Grt = ). This marked change in the HREE distribution between zircon and garnet must reflect either a change in the minerals with which the zircon was growing or being modified, a change in the physical and chemical conditions of zircon growth, or a combination of the two. Based on comparisons with recent estimates of equilibrium zircon/garnet HREE distribution coefficients we infer that the inner core (I) did not grow with the garnet that occurs in the paragneiss but grew within a garnet absent melt that was then injected into the gneiss. The resulting leucosome then underwent wall rock reaction with the enclosing garnet bearing gneiss, causing a decrease in garnet to Zr contents to values approaching equilibrium with melt, and precipitating the zircon outer core (II). Finally, the zircon rim (III) and later monazite formed in a HREE depleted environment. Melt injection, reaction and crystallization of the leucosomes took place within the time interval Ma at the end of the UHT history of the Napier Complex. Introduction Zircon U Pb chronology is now considered as the most convincing age determination method for high and ultrahigh temperature (HT UHT) metamorphic rocks. High spatial and mass resolution analysis using the ion microprobe makes it possible to determine the U Pb ages of zircon micro domains smaller than μm in diameter. It is, however, not always possible to unambiguously interpret the zircon ages based on zircon crystal morphology or internal zonal structures alone. Zircon trace element T. Hokada, hokada@nipr.ac.jp Corresponding author geochemistry has recently been used to investigate the processes or conditions of zircon crystallization. Application of rare earth element (REE) partitioning between zircon and garnet has been applied to interpret the event significance of zircon ages (Harley et al., 2001; Harley, 2002; Rubatto, 2002; Whitehouse and Platt, 2003). Both zircon and garnet favour heavy rare earth elements (HREEs) relative to light to middle rare earth elements (L MREEs). Recent studies have proposed equilibrium REE D zircon/garnet values that are consistent and near unity for MREE, but different by as much as an order of magnitude for HREE. Rubatto (2002) proposed that HREE are more favorably incorporated into zircon than garnet, whereas

2 Zircon growth in UHT leucosome 181 Harley et al. (2001) and Whitehouse and Platt (2003) suggested that the equilibrium REE D zircon/garnet values for HREE are less than unity ( Yb D zircon/garnet 0.7), indicating that the HREE are preferred in garnet over zircon for the pelite garnet compositions and high temperature conditions considered in their studies. The Napier Complex is recognised as the classic example of UHT metamorphism. It has experienced regional >1100 C peak metamorphism at GPa, characterized by the occurrence of diagnostic mineral parageneses including sapphirine + quartz, osumilite, high Al orthopyroxene (up to 12.5 wt% Al 2 O 3 ), ternary mesoperthitic feldspar, and inverted metamorphic pigeonite (e.g., Harley, 1998; Harley and Motoyoshi, 2000; and see references therein). In spite of such high metamorphic temperatures, evidence of melting and migmatisation is less clear (e.g., Sheraton et al., 1987). However, localized patchy vein leucosomes that may represent partial melts formed during the UHT metamorphic history do occur, and yield feldspar temperatures of >950 C consistent with this interpretation (Hokada et al., 2004). In this work we discuss zircon crystallization in an example of such a UHT feldspathic leucosome from Mt. Riiser Larsen (Hokada, 2001; Hokada et al., 2004), using trace element and REE ion microprobe analyses to evaluate the timing of zircon growth in relation to garnet in the leucosome host and hence the meaning of the measured zircon U Pb ages in terms of the age of UHT metamorphism in this complex terrane. Geological Setting Archaean granulite facies UHT metamorphic rocks comprise the Napier Complex in East Antarctica (Fig. 1; e.g., Sheraton et al., 1987; Harley and Black, 1997). The Napier Complex is exposed as mountains or nunataks within the ice sheet in a coastal area of Antarctica from longitude 46 E to 57 E, and occupies an area of about 400 km 200 km (i.e km 2 ). Mt. Riiser Larsen, the study area, is located on the northeast coast of Amundsen Bay within the UHT region (Fig. 1). The area is composed of orthopyroxene bearing felsic orthogneiss, garnet bearing felsic gneiss and two pyroxene bearing mafic granulite with subordinate pelitic, psammitic, siliceous, aluminous and ferruginous paragneisses, pyroxenite and ultramafic granulite (Ishizuka et al., 1998). These Archaean gneisses are intensely deformed and display layer parallel foliations with tight, intrafolial folds. A relatively massive and thick tonalitic orthogneiss unit overlies the layered gneisses in the eastern part of Mt. Riiser Larsen with apparent concordance. Mineral assemblages are nearly anhydrous, consistent with recrystallization under UHT metamorphic conditions, and confirmed by the occasional presence of high F phlogopite inclusions in the main silicate minerals (Motoyoshi and Hensen, 2001). The gneissic layering is locally cut by dolerite dykes which probably intruded in Figure 1. Geological outline of the Napier Complex and the surrounding area in Antarctica.

3 182 T. Hokada and S.L. Harley Figure 2. (a) Photograph of the garnet orthopyroxene bearing paragneiss sample (R2301B) including the UHT feldspathic leucosome. Qtz, quartz; Grt, garnet; Opx, orthopyroxene; Fsp, perthitic feldspar. (b) Sketch of the leucosome sample. Z, zircon; M, monazite. SIMS analysis spots on garnet are given as g1 g25. the Proterozoic age ( 1.2 Ga: e.g., Sheraton and Black, 1981). Petrography The leucosome bearing sample R2301B occurs in a garnet orthopyroxene bearing paragneiss/quartzite layer (20 30 m thick) within the layered paragneisses. The host garnet orthopyroxene gneiss is coarse grained and is composed of quartzo feldspathic, siliceous and aluminous layers that locally contain sapphirine and osumilite. The major constituents of the quartzo feldspathic layers are quartz, antiperthitic plagioclase (an 25 ab 46 or 29 ), garnet (core: alm 58 prp 36 grs 5 sps 1 rim: alm 59 prp 34 grs 6 sps 1 ) and orthopyroxene (Mg/(Mg + Fe) = 0.63). Ilmenite, biotite and zircon are present as accessory minerals. The siliceous layer is composed mostly of quartz and garnet (core: alm 49 prp 48 grs 3 rim: alm 53 prp 45 grs 2 ) with a subordinate amount of orthopyroxene (Mg/(Mg + Fe) = ). Quartz in this layer is xenoblastic to rounded (< 5 mm in diameter), whilst garnet forms commonly rounded porphyroblasts up to 1 cm in diameter, or occasionally occurs as xenomorphic crystals. Opaque minerals (e.g., ilmenite in the host gneiss) are absent in the siliceous layer. Rutile is an accessory. Patch and vein leucosomes can be identified by the occurrence of perthitic alkali feldspar (an 7 ab 28 or 65 ; Fig. 2a, b), which is not observed in the host paragneiss/quartzite. The leucosome feldspar occurs interstitially among rounded quartz grains and has concave or embayed grain boundaries, suggestive of direct crystallization from melt. Minimum T estimated using integrated pre exsolution perthite compositions are >950 C at 0.8 GPa (Hokada et al., 2004). Garnet near these leucosomes is occasionally lobate and interstitial to quartz grains, again suggestive of mineral melt interaction. Idiomorphic or sub rounded coarse grained zircon grains (up to 400 μm diameter) occur in or near the leucosome feldspars. These zircons occasionally show oscillatory zoned and sector zoned internal structures. Coarse grained and irregular shaped monazite (up to 500 μm), which is not present in the host paragneiss, occurs in the leucosome domains. Monazite

4 Zircon growth in UHT leucosome 183 is commonly associated with zircon, and is interstitial to or surrounding zircon grains and hence clearly post dates zircon growth. Table 1. Representative electron microprobe analyses of garnet and perthitic feldspar in sample R2301B Analytical method Trace elements and REE chemistry Major elements of garnet and feldspar were analyzed using electron microprobe (JEOL JXA 8800) at the National Institute of Polar Research, and are given in Table 1. Rare earth elements (REEs) and trace elements within zircon (Table 2), garnet (Table 3) and monazite (Table 4) were analyzed using ion microprobe (CAMECA ims 4f) at the University of Edinburgh. Analytical conditions and correction procedures essentially follow those of Hinton and Upton (1991). Analyses were conducted using a 6nA O primary beam focused to a μm spot. Secondary ions extracted at a potential of 4500V were measured at a 120eV high energy offset for zircon and monazite, and 75eV energy offset for garnet, with a 19eV energy window. The NIST 610 glass standard was used for calibration, and ion yields corrected by reference to zircon SL13 (for zircons) and the DDI garnet standard (for garnets). Zircon Zircon grains typically display 3 textural domains (Fig. 3A D): structured inner cores that have a low CL response (inner core I); structured outer cores that are bright in their CL response (outer core II); and relatively structureless outer rims (rim III). All three zircon textural domains commonly show positive Ce and negative Eu anomalies on chondrite normalized REE diagrams (Fig. 4). The zircon inner cores (I) have high REE concentrations, with HREE enrichments of over 1000 times chondrite at Dy, are characterized by Yb(n)/Gd(n) 3.3, and have high Th/U ratios of 3.2. Zircon outer cores (II) have lower REE concentrations and relatively constant M HREE patterns of about times chondrite at Dy, Yb(n)/Gd(n) of 0.7, and Th/U ratios of 1.2. Zircon rims (III) have REE concentrations and chondrite normalized patterns similar to the zircon outer cores, at about times chondrite at Dy and with Yb(n)/ Gd(n) of 0.8, but Th/U ratios are considerably lower (0.3). LREE concentrations in the zircon rims (III) are slightly lower than in the structured higher Th/U cores (I) and (II). Garnet * Total Fe as FeO. ** X Mg : Mg/(Mg+Fe2+). Garnet (core: alm 49 prp 48 grs 03 rim: alm 53 prp 45 grs 2 ) displays two important compositional features in terms of trace elements: Zr contents are in the range 300 ppm for the interior parts of the one large (4 mm diameter) grain analyzed. The rim and near rim areas of this grain (i.e. within 200 μm of rims) show a marked drop in Zr to ca. 100 ppm, accompanied by a marked decrease in Hf (from ca. 18 ppm to 9 ppm) and minor decreases in Y (from 540 ppm to 460 ppm, Fig. 5). Chondrite normalized high Zr garnet cores exhibit REE patterns characterized by low LREE with high Sm/Nd, negative Eu anomalies, and flat M HREE with Yb(n)/Gd(n) near Low Zr garnet rims are depleted in Dy, Tb and particularly Gd relative to the cores, and show consistent decreases in both Sm and

5 184 T. Hokada and S.L. Harley Table 2. Zircon REE, P, Y, Hf, Th and U concentrations (ppm) in R2301B leucosome * Chondrite normalized. Nd. Monazite Monazite exhibits HREE depleted chondrite normalized patterns with negative Eu anomalies. Compositional variations are observed for M HREE with respect to the irregular shaped internal zoning visible under BEI (Fig. 3G, 3H). REE contents in monazite are higher than in garnet and zircon rims except at Yb and Lu, where the contents are similar. Discussion The origin of the patch vein leucosomes was discussed in Hokada et al. (2004). It was suggested that these leucosomes were possibly formed through partial melting of the nearby host gneiss during the UHT metamorphism. However, the process of such UHT melt formation is too complicated to be examined in detail here. We, therefore, discuss only the leucosome crystallization process using the REE chemical data in this paper. Rare earth element (REE) distribution coefficients calculated between zircon and garnet ( REE D zircon/garnet ) in the R2301B leucosome sample range from 0.5 to 8 for M HREE (Sm Lu; Table 5). Distribution coefficients greater than 5, and up to 1000 at La, characterize the LREE (La, Ce, Pr and Nd) although these are relatively scattered due to the low LREE concentrations in both zircon and garnet. In more detail, key differences in the apparent REE D zircon/garnet are observed when the REE data linked with the texturally defined zircon domains are considered separately. The high REE zircon inner cores (I) yield REE D zircon/garnet values that range from 2.3 at Gd to 8.3 at Lu when coupled with the 300 ppm Zr garnet cores, and from 3.4 at Gd to 9.3 at Lu when coupled with the 100 ppm Zr garnet rims. The progressive increase in REE D zircon/garnet from Gd to Lu so calculated for zircon inner cores (I) is consistent in a general sense with previous interpretations of natural samples that infer greater incorporation of the heavier REE into zircon compared with garnet (e.g. Rubatto, 2002) when in equilibrium and in the presence of melt. If these zircon inner cores (I) grew with the garnets that contain 300 ppm Zr, then they could have formed at C provided that the distribution of Zr between garnet and melt, Zr D Garnet/Liquid, is in the range

6 Zircon growth in UHT leucosome 185 Table 3. Garnet REE, P, Ca, Ti, Y, Zr and Hf concentrations (ppm) in R2301B leucosome * Chondrite normalized. Table 4. Monazite REE, Y, F, Al, Si, P and Ca concentrations (ppm) in R2301B leucosome

7 186 T. Hokada and S.L. Harley Figure 3. SIMS analysis spots on zircon and monazite. (a), (c) and (e) Cathodoluminescence images (CLI) of zircon. Italic numbers give the U Pb age (Ma) obtained by SHRIMP analysis (Hokada et al., 2004). (b), (d) and (f) Sketch of the zircon grains as same as in Figure 3 (a, c and e). (g) (h) Backscattered electron image (BEI) of monazite. (Green, 1994). This would yield calculated Zr contents of ppm in the melt, attainable at these UHT conditions (Watson and Harrison, 1983; Watson, 1996). If this is the case the REE D zircon/garnet could be inferred to reflect equilibrium under UHT conditions, in the presence of melt.

8 Zircon growth in UHT leucosome 187 Figure 4. Chondrite normalized REE pattern of zircon, garnet and monazite in R2301B analyzed by SIMS. Abundances of REE in C1 chondrites are cited from Anders and Grevesse (1989). (a) Zircon. (b) Garnet. (c) Monazite. (d) Zircon/garnet partitioning. The low REE zircon outer cores (II) or rims (III)/ garnet cores or rims, in contrast, define REE D zircon/garnet values that range from near unity ( ) at Gd down to at Yb and Lu. These REE D zircon/garnet values are similar to the D values obtained by Harley et al. (2001) and Harley (2002) and interpreted to reflect HT UHT REE equilibrium between zircon and garnet. They are also consistent with D values obtained by Whitehouse and Platt (2003) for zircon interpreted to have grown in the presence of garnet. In the specific case of this leucosome, the marked change in the distribution of REE, particularly HREE, between zircon and garnet must reflect either a change in the minerals with which the zircon was growing or being modified and/or a change in the physical and chemical conditions attending zircon formation. Given the REE distribution systematics described above, there are two different scenarios for zircon growth and modification in this rock, each with very different consequences for the interpretation of the zircon age data. Scenario 1: In this interpretation the structured zircon inner cores (I) grew with garnet cores at equilibrium under UHT conditions, and the relevant zircon/garnet REE distribution pattern under these conditions is of increasing REE D zircon/garnet values towards heavier REE. In this interpretation, the flat to negatively sloping zircon REE patterns relating to zircon (II) and (III) would have to reflect growth or modification of zircon with either another mineral (e.g., monazite) or under different physical and chemical conditions in which REE D zircon/garnet changed markedly without any significant changes in the major element chemistries of the silicate minerals present in the rock. Scenario 2: In this interpretation the structured zircon inner cores (I) did not grow with the garnet that occurs in the rock but was instead introduced into the garnet bearing paragneiss with melt to form the veining leucosome (Fig. 6). This melt/leucosome then underwent wall rock reaction with the enclosing garnet bearing paragneiss, causing the garnet to decrease its Zr and Hf contents (to values approaching equilibrium with the melt) and precipitating the final zircon outer cores (II) or rims (III) that have flat to negatively sloping HREE patterns and yield REE D zircon/garnet of 1.0 to 0.7 for the HREE from Gd to Lu. In this case the zircon inner core (I)/garnet D values do not represent equilibrium, and the apparent preferential distribution of HREE into zircon over garnet is a misleading artifact of coupling two mineral compositions that are not in equilibrium during their growth. On the other hand, zircon/garnet REE D values obtained for the zircon outer cores (II) or rims (III) and garnet rims would approximate equilibrium. Monazite formed last, interstitial

9 188 T. Hokada and S.L. Harley Figure 5. Compositional variation of Yb(n)/Gd(n) ratios and Zr Hf and Y contents in garnet grain traverse from the rim to core. Table 5. REE distribution between zircon and garnet ( REE D zircon/garnet ) in R2301B leucosome to zircon and garnet, in a HREE depleted environment mediated by the prior growth and modification of zircon and garnet. In scenario 1, the flat to negatively sloping zircon HREE group (II and III) would have to reflect growth or modification subsequent to garnet crystallization. However, there is no evidence in this rock for garnet breakdown, and no evidence for the growth of new competitor minerals for HREE. Additionally, this model does not provide a reasonable explanation for the decrease in Zr in garnet from 300 to 100 ppm in the presence of melt, and cannot account for this without invoking unlikely factors such as major changes in the Zr D between the minerals and/or local temperature variations. In contrast, scenario 2 does not require a special process to be invoked to explain the final flat zircon patterns. Moreover, scenario 2 is consistent with the observed crystallization sequences of garnet, zircon and monazite and compatible with physical processes of melt migration and injection in the crust. For these reasons, the scenario 2 is the preferred interpretation of the textural and REE features of zircon, garnet and monazite in this UHT leucosome/paragneiss. In this case only zircon domains (II) and (III) preserve HREE patterns indicative of formation in the presence of garnet, which in this sample was present prior to melt injection. The resultant zircon (II/III)/garnet HREE D of are consistent with values reported by Harley et al. (2001) for other Napier Complex leucosomes and interpreted on textural ground to reflect equilibrium or near equilibrium conditions. Hokada et al. (2004) have reported SHRIMP U Pb ages of Ma from the structured inner cores (I) and outer cores (II) of the zircons in R2301B. In scenario 2 it is possible that the ages of these chemically distinct zircon domains, both high REE and low REE, could be similar and indeed the same within error, if initial zircon crystallization in the leucosome took place just prior to injection of the melt into neighboring wall rocks. Melt wall rock reaction and further crystallization of zircon within the leucosomes then continued over a narrow time interval at the end of the UHT history of the Napier Complex. In this interpretation all the zircon types (I, II and III) preserved in the leucosome would have formed during cooling from UHT conditions.

10 Zircon growth in UHT leucosome 189 Figure 6. Schematic illustration of the inferred crystallization process of the zircon, garnet and monazite in R2301B leucosome sample. Conclusions Zircon garnet REE relations in the UHT leucosome from the Napier Complex suggest the following: 1. Chemical compositions, especially chondrite normalized REE patterns analyzed by SIMS, correlate with three textural domains of zircon grains: the structured inner cores (I) show HREE enrichment whereas the outer cores (II) and rims (III) have flat to relatively depleted HREE patterns. REE in garnet show less variation than those in zircon, and are typified by almost flat M HREE with slight MREE depletions on garnet rims, which may reflect the crystallization of late monazite. 2. When existing zircon granet REE distribution data are considered it is possible that either the zircon inner cores (I) or outer rims (III) are in equilibrium with garnet. However, from the textural observations and changes in REE chemistries it is more likely that the zircon outer rims (III) are in REE equilibrium with garnet, with apparent REE D zircon/garnet values that range from Gd D zircon/garnet = 1.0 to Yb D zircon/garnet = 0.7, consistent with those reported by Harley et al. (2001). 3. The zircon inner core (I) crystallized on cooling in the UHT leucosome but did not grow with the garnet present in the rock. The zircon outer cores (II) and rims (III) grew with garnet, and later monazite formed in a HREE depleted environment mediated by the presence of zircon and garnet rims. 4. Ages obtained on the zircons from this leucosome do not define the peak of UHT metamorphism but instead relate to zircon crystallization from UHT melt during cooling at the end of the UHT history of the Napier Complex. Acknowledgments Sample R2301B was collected by TH during Japanese Antarctic Research Expedition (JARE), and the members of JARE particularly Y. Osanai, T. Toyoshima, M. Owada, T. Tsunogae and W.A. Crowe are thanked for their support during the field work. TH acknowledges the National Institute of Polar Research (NIPR), especially K. Shiraishi and Y. Motoyoshi, for enabling him to participate in JARE. The authors wish to acknowledge careful review by Y. Hiroi and T. Tsujimori, and editorial assistance by M. Santosh. We are grateful to R.W. Hinton and N.M. Kelly for technical support and assistance in ion microprobe analysis. This work was carried out whilst TH was visiting the University of Edinburgh supported by a Grant in Aid for the Young Scientists from the Japan Society for the Promotion of Science to TH and by a Royal Society grant to SLH. References Anders, E. and Grevesse, N. (1989) Abundances of the elements: Meteoritic and solar. Geochimica et Cosmochimica Acta, 53, Green, T.H. (1994) Experimental studies of trace element partitioning applicable to igneous petrogenesis Sedona 16 years later. Chemical Geology, 117, Harley, S.L. (1998) On the occurrence and characterisation of ultrahigh temperature crustal metamorphism. In What Drives Metamorphism and Metamorphic Reactions? (Treloar, PJ. and O Brien, P. Eds.). Geological Society of London Special Publication 138, London, Harley, S.L. (2002) Zircon garnet REE distribution patterns and

11 190 T. Hokada and S.L. Harley the behaviour of zircon during UHT metamorphism. Programme with Abstracts, International Mineralogical Association 18, Edinburgh, 236. 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 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, Harley, S.L., Kinny, P.D., Snape, I. and Black, L.P. (2001) Zircon chemistry and the definition of events in Archaean granulite terrains. In Extended Absracts of 4th International Archaean Symposium (Cassidy, K.F., Dunphy, J.M. and Van Kranendonk, M.J. Eds.). AGSO Geoscience Australia Record 2001/37, Canberra, Hinton, R.W. and Upton, B.G.J. (1991) The chemistry of zircon: variations within and between large crystals from syenite and alkali basalt xenoliths. Geochimica et Cosmochimica Acta, 55, Hokada, T. (2001) Feldspar thermometry in ultrahigh temperature metamorphic rocks: evidence of crustal metamorphism attaining 1100 C in the Archean Napier Complex, East Antarctica. American Mineralogist, 86, Hokada, T., Misawa, K., Yokoyama, K., Shiraishi, K. and Yamaguchi, A., (2004). SHRIMP and electron microprobe chronology of UHT metamorphism in the Napier Complex, East Antarctica: implications for zircon growth at >1000 C. Contributions to Mineralogy and Petrology, 147, 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, Motoyoshi, Y. and Hensen, B.J. (2002) F rich phlogopite stability in ultra high temperature metapelites from the Napier Complex, East Antarctica. American Mineralogist, 86, Rubatto, D. (2002) Zircon trace element geochemistry: partitioning with garnet and the link between U Pb ages and metamorphism. Chemical Geology, 184, 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. Bulletins of Australia Bureau of Mineralogical Resources, 223, 1 51 Watson, E.B. (1996) Dissolution, growth and survival of zircons during crustal fusion: kinetic principles, geological models and implications for isotopic inheritence. Transactions of Royal Society of Edinburgh: Earth Sciences, 87, Watson, E.B. and Harrison, T.M., (1984) Zircon saturation revisited: temperature and composition effects in a variety of crustal magma types. Earth and Planetary Science Letters, 64, Whitehouse, M.J. and Platt, J.P., (2003) Dating high grade metamorphism constraints from rare earth elements in zircon and garnet. Contributions to Mineralogy and Petrology, 145, (Manuscript recieved;11 December, 2003) (Manuscript accepted; 3 may 2004)