1. metamorphic Ceol, 1989, 7, 261-275 Evidence for deformation-induced K-feldspar replacement by myrmekite C. SIMPSON Department of Earth and Planetary Sciences, The johns Hopkins University, Baltimore, M D 21218, USA R. P. WINTSCH Department of Geology, Indiana University, Bloomington, IN 47405, USA ABSTRACT Several examples of deformation-inducedmyrmekite have been found in two amphibolite facies mylonites derived from granitic protoliths, namely a muscovite-poor S-C mylonite and a single foliation, muscovite-poor mylonitic gneiss. Back-scattered SEM and conventional optical microscopy show that in both rock types, syntectonic myrmekitic intergrowths of oligoclase and quartz formed on the two sides of &feldspar grains that faced the local inferred incremental shortening direction for the mylonite. Myrmekite does not occur on the two ends of the grain that faced the incremental stretching direction. The replacement of K-feldspar by plagioclase and quartz results in a volume decrease and is favoured on high normal stress sites around the grains. We suggest that the ambient temperature, pressure and chemical activities were such that the replacement reaction was favoured, but the addition of extra strain energy along the high-pressure sides of the grains localized the reaction at these sites. This energy could arise from elastic strain, or strain associated with tangled dislocations or twin boundaries. The relative roles of stress and strain energy concentrations in driving the replacement reaction are not known, but both were probably important. Key w o d x Deformation; mylonite; myrmekite; strainenhanced reaction. INTRODUCTION Our understanding of the s i p f i a n c e of the interactions between deformational and metamorphic processes has been growing over the last few years with the recognition that reactions can affect rock strength (White & Knipe, 1978; Rubie, 1983), and that deformation may drive replacement reactions (Wintsch & Knipe, 1983). The role of deformation in the replacement of K-feldspar by mytinekite, a plagioclase-vermicular quartz symplectite, was considered first by Futterer (1894) who noted a greater development of mynnekite in deformed plutonic rocks than in their undeformed equivalents. Eskola (1914), Sederholm (1916) and Spencer (1945) all remarked on the tendency for myrmekite to occur in zones of ‘ ‘ ~ h e d )or ’ “granulated” rock. Myrmekite growth in positions of high stress at grain boundaries has been suggested by Schreyer (1958), Sarma & Raja (1959), Shelley (1964). Binns (1966), Parslow (1971), Bhattacharyya (1971) and Phillips & Carr (1973) among others. Summaries of these early works are given in Ashworth (1972), Phillips (1974) and Smith (1974). More recently, a correlation between mynnekite growth and crystal-plastic deformation has been suggested by several workers including Hanmer (1982), Tullis (1983), Watts & Williams (1983), Vernon, W i l l i a m s & D’Arcy (1983), S i p n (1983; 1985), Hibbard (1987), LaTour (1987) and LaTour & Barnett (1987). . . Despite this considerable attention to the problem, a single model for the relationship between myrmekite formation and the assoCiated deformation event has proved elusive. Shelley (1964) considered the quark vermicules in myrmekite to be recrystallized grains, inherited from an earlier “cataclastic” event that became incorporated into recrystalliPng feldspar. Hibbard (1987) ascribed all myrmekite (including “strained mynnekite units”) in orthogneisses to recrystallization in the presence of a late-stage, water-saturated melt during the deformation of an incompletely crystallized magma. Hanmer (1982), in his study of naturally deformed granites in Newfoundland, contended that myrmekite formation was post-deformational on the grounds that the delicate vermicular textures could not survive appreciable strain. However, a geometrical correlation between mynnekite sites on inequant K-feldspar porphyroclasts and the local k i t e strain axes was noted by Simpson (1985) in mylonitic granodiorite of the Eastern Peninsular Ranges mylonite zone, southern California. Progressive rotation of myrmekitic intergrowths into millimetre-scale shear bands was taken to illustrate syndeformational myrmekite formation, for which a solid-state difhsion-exsolution model was proposed (Simpson, 1985). Recently, several other mylonitic rocks from very different geological settings have also been found to contain K-feldspar porphyroclasts in which the position of m 262 C. SIMPSON & R . P. WINTSCH mynnekite exhibits a geometrical relationship to the finite and inferred incremental strain axes. These myrmekites d o not readily fit into a solid solution model. The best examples occur in two-foliation mylonites that are Types I and I1 S-C mylonites of Lister & Snoke (19&Q), where S = schistosity and C = cisaillement or shear (BerthC, Choukroune & Jegouzo, 1979). Similar phenomena have also been observed in mylonitic rocks that contain a single foliation (e.g. Debat, Soula, Kubin & Vidal, 1978, fig. lb; Vidal, Kubin, Debat & Soula, 1980, fig. la; Simpson, 1981). Si@cant modification of the stability fields of minerals in regions of high strain was first suggested by Harker (1932) and again by Helgeson, Delany, Nesbitt & Bird (1978) and Wintsch (1985) but few systematic studies of the phenomenon have been made until quite recently, and then only on selected minerals. Nucleation of new grains may lower the free energy of mica by a combination of reductions in strain and chemical energy (Etheridge & Hobbs, 1974; Knipe, 1981). In greenschist facies mylonites, the relative stability of feldspar with respect to mica may be shifted by either strain or stress concentrations (Knipe & Wintsch, 1985). Under very high-grade conditions, sillimanite may precipitate, and feldspar and biotite dissolve, at sites of high strain and/or stress along feldspar grain boundaries (Vernon, 1987; Wintsch & Andrews, 1988). Quantitative estimates of the effect of strain energy associated with dislocations on the relative stabilities of the A l S i 0 5 polymorphs (Kemck, 1986), on the aqueous solubility of quartz (Wintsch & Dunning, 1985), and of calcite (Bianchetti & Reeder, 1985). indicate that dislocation densities of at least 10’ocm-2 are necessary to perturb a strain-free chemical equilibrium significantly. Data from recent experiments on the high-pressure breakdown of plagioclase suggest that the presence of a high dislocation density may not in itself promote nucleation and growth of reaction products, but an enhanced reaction rate can be related to dynamic recrystallization mechanisms (Yund & Tullis, 1986). In this paper we describe the geometry and geochemistry of myrmekite rims on K-feldspar porphyroclasts in two mylonites: the Santa Rosa S C mylonite of southeastern California and the lineated Ponaganset gneiss of the Hope Valley shear zone, eastern Connecticut. The term “porphyroclast” is used to describe K-feldspar megacrysts, the size of which is reduced by syndeformational recrystallization or by myrmekite replacement and not by a fracture mechanism. A model is presented for the syntectonic replacement of K-feldspar by myrmekite as a result of increasing lattice strain in the K-feldspar host grain. No attempt is made here to discuss the origin of all myrmekite occurrences. We limit our discussion to mynnekite that shows a definite textural asymmetry that correlates with the surrounding tectonic fabric. METHODS Oriented thin sections were cut parallel to the mineral elongation lineation and perpendicular to the S foliation in the rocks. All mineral analyses were carried out on an ARL electron microprobe at Virginia Polytechnic Institute and State University, and on a Cameca electron microprobe at the University of Michigan, using natural standards and ZAF reduction programs. Back-scattered scanning electron photomicrographs were made on an Hitachi S570 electron microscope at the University of Michigan. Activity diagrams were calculated using the program DIAGRAM and the hydrolysis equilibria summarized in Bowers, Jackson & Helgeson (1984). Determination of K-feldspar structural states by X-ray difkaction would be important information for this project, but impractical without microstructural control on the relative positions of the grains within the mylonitic foliation. OCCURRENCES K-feldspar porphyroclasts that show geometrical features similar to those described here have been found by us in the following rocks, among others: Santa Catalina metamorphic core complex mylonites, south-western Arizona, in an extensional tectonic regime (Keith, Damon, Reynolds, Shafiqulla, Livington & Pushkar, 1980); Maggia mylonitic gneiss, Lepontine Alps, Switzerland, in a compressional tectonic regime (Simpson, 1981; 1982); mylonitic rocks from the Honey Hill and Tatnic faults, eastern Connecticut, in a transpressional regime (Wintsch & Sutter, 1986); Santa Rosa mylonites of the Eastern Peninsular Ranges mylonite zone (EPRMZ), south-eastern California, in a compressional tectonic regime (Sharp 1966; 1979; Simpson, 1984; 1985); and lineated Ponaganset gneiss, Hope Valley Shear Zone, eastern Connecticut, in a transpressional tectonic regime (O’Hara & Gromet, 1985). For the sake of brevity we will discuss samples from the last two of these localities only. Sang Rosa Mylonite The Santa Rosa mylonite sample is from a mylonitic granodiorite section of the EPRMZ (Fig. la); see also Sharp (1966; 1979) and Sipson (1984; 1985). Theodore (1970). Anderson (1983), Simpson (1984; 1985) and Erskine & We& (1985) have provided geological, petrologicaland petrographical details of the mylonitic rocks of the EPRMZ. Our specimens, especially SR656. were taken from the Late Cretaceous, west-directed thrust zone rnylonites, well away from the younger low-angle detachment faults that occur throughout the region (Engel & Schultejahn, 1984; Schultejahn, 1984; Erskine & Wenk, 1985). Pressuretemperature conditions during deformation are estimated to have been in the middle amphibolite facies or in the range 450-550°C and 4-5 kbar by Anderson (1983) and Simpson (1985). The total finite strain in the sample is thought to be low because the angle between C- and S-surfaces is approximately 43” and the offset along individual C-planes does not usually exceed 1.Ocm. The finite shortening direction is therefore assumed to be close to the incremental shortening direction for the rock, i.e. subperpendicular to the S-surfaces (Berth6 CI al., 1979). In thin section, S surfaces are defined by quartz ribbonscontaining 150-250 prn-sized grains, aligned biotite grains (aspect ratios 50 X loo0 pm in sections perpendicular to (001) plants) and minor recrystallized muscovite. Tabular morphology and the absence of undulatory extinction suggcsts that most of the biotite K-FELDSPAR R E P L A C E M E N T B Y M Y R M E K I T E N m Flg. 1. General locality maps for myrmekite-bearing mylonites discussed in text. Stars mark sample locations;general trend of foliations and lineationsindicated by strike lines and arrows. (a) Santa Rosa mylonite zone, Palm Canyon, southeastern California (sample SR656). modified from Erskine & We& (1985); g = Peninsular Ranges granite; mg = mylonitic Peninsular Ranges granite; ms = mylonitic mctasedimentary rocks; gd = granodiorite; teeth on thrust fault and double ticks on detachment fault both on upper plates. (b) Hope Valley shear zone, Connecticut (sample B e ) , after Gromet & O'Hara (1985). LCF= Lake Char Fault; HVSZ = Hope Valley shear zone; HVT (stippled) = Hope Valley subterrane; E-DT = Esmond-Dedham subterrane (of Avalon terrane). grains are completely recrystallized. Recystalliizcd, 50-100pmsized plagioclase grains of composition An,,, to An, generally form a polygonal mosaic between quartz ribbons, although occasional relict porphyroclasts, 2 9 pm to 1.0 mm in diameter, are present. Elongate K-feldspar megacrysts, 200 pm to 2.0 mm in diameter, also occur; representative chemical analyses of the K-feldspar are given in Table 1, columns A and B. Although difIiculty was experienced in obtaining high quality analyses of these p i n s , the overall trends were reproducible. The compositlons of oligoclase grains in the myrmekitic rims of the K-feldspar megacrysts are within a fairly tight range of An, to Ass. Muscovite occurs with or without d a t e within plagioclase porphyrodasts and constitutes <0.5% of the total rock. Where muscovite occurs at the edge of or within plagioclase augen. it generally has (001) planes aligned subparallel to the adjacent foliation planes. Within large plagioclase porphyroclasts, muscovite occurs as undeformed. anhedral grains, commonly with their (001) planes at a high angle to the surrounding foliation. Muscovite is never seen in close spatial association with myrmekite, either in the matrix or around augen. Subhedral danite grains show no deformation features. Where elongate, the allanite crystals are aligned subparallel to the adjacent foliation planes. Biotite and/or epidote commonly form a narrow rim around the allanitc grains; the biotite grains are parallel to and help define the foliation. Very minor retrogression of biotite to chlorite occurs along some C-surfaces. Epidote occurs in isolated stringers of 30pm-sized grains between quartz ribbons. S i c pre-tectonic garnet grains were observed in two out of ten thin sections. The garnet grains are fractured, with one main fracture at a high angle to the S-surfaces. Quartz is the major precipitate within the fractures. Titanite occurs as a minor accessory mineral in very small, isolated, subhedral grains; it shows no evidence of fracturing or other deformation features. The foliations in this rock form a conspicuous S-C fabric (Fig. 2); also see Simpson (1985. Fig. 3). Ssurfaces are deflected into C-bands with a consistent sense of curvature, concomitant with a decrease in rtcrystallized grain-size, and both wrap around the large K-feldspar augen. The C-surfaces contain nibons of Iine-grained (30-40pm-sized) polygonal quartz, plagiodase and K-feldspar, with very line-grained biotite and m i w r muscovite (aspect ratios 10 x 20-50 pm). Back-scattered electroo imagery (Fig. 3c), shows nearty monominerallic layers of quartz, 50-2OOpm wide, separating nearly biminerallic K-feldspar+ oligoclase layers of the same thicloess. Back-scattered electron imagery allows the precise location of myrmekite around the K-feldspar margins and in the recrystalked S and C-bands to be determined. Myrmekitic intergrowths of oligoclase (c. &;Table 1, cdumn~C and D) and quartz generally occur on the two sides of the K-feldspar augen that parallel the S-surfaces and face the incremental shortening direction (Figs 2a, 3a, 4a). Myrmekite also oaufs along some augen margins that are adjacent to C surfaces, at about 45" to the incremental shortening direction (Fig. 3a). However, myrmekite has never been found on the ends of the augen that face the incremental stretching direction. In these 'tail' regions there are subgrains and equigranular recrystallized K-feldspar grains (Fig. 4a; Table 1, cohunns E and F). Optical examination with a 264 C. SIMPSON & R. - P. WINTSCH T.bk 1. Representative electron microprobe analyses of feldspars from sample SR656. Rim analyses taken within the first 100 pm of 1 myrmekite at porphyroclast margins. Rim Core orthoclase A SiO, MZO, FeO CaO Na,O K,O Total Si Al Fe ca Na K B oligoclase in myrmekite C D E F 14.57 66.74 18.76 0.06 0.04 1 14.52 61.31 25.16 0.03 5.93 7.94 0.19 63.26 24.57 0.01 5.72 8.27 0.12 64.31 20.42 0.04 0.08 1.12 113.89 64.56 20 0.04 0.05 1.09 14.02 99.83 101.14 100.56 101.94 99.82 99.73 2.943 1.101 0.001 2.958 1.08 0.001 0.004 0.002 0.097 65.74 18.53 0.08 0.03 0.88 3.014 1.001 0.003 0.001 0.079 0.852 Cations for 8 oxygens 3.015 2.705 2.748 0.999 1.308 1.257 0.002 0.001 0 0.002 0.28 0.266 0.088 0.679 0.6% 0.837 0.011 0.007 0.099 0.811 0.819 (Ca+ Na + K) 0.932 cationsum 0.927 0.97 0.%9 0.914 0.918 0.15 8.43 91.42 0.19 9.46 90.35 28.89 70.00 1.11 27.46 71.86 0.68 0.43 10.87 88.70 0.26 10.54 89.20 An Ab Or Bi Tails recrystallized microcline universal stage indicates that the recrystallized K-feldspar grains immediately adjacent to the host crystal have not greater than 15" lattice misorientations with their host grain. The K-feldspar host grain interface with myrmekite is always sharply defined (Fig. 3b) and may form rounded, bulbous or polygonal outlines that are generally not parallel to rational crystallographic planes in the host grain. The boundary is also chemically abrupt at the scale of analysis (<5pm electron beam diameter). Oligoclase grains in the myrmekitic intergrowths, are 50 pm-sizcd at the K-feldspar contact, and decrease to c. 30 pm in the adjacent polygonally recrystallized material (Fig. 3b). The quartz vermicules commonly change from roughly cylindroidal, <5 pm-sized grains near to the K-feldspar host grain, to polygonal 10pm-sizcd grains in the matrix, which suggests that the quartz grains farthest from the porphyroclast have undergone the most grain growth by grain boundary migration whereas oligoclase grains in this position have undergone the greatest grain-size reduction. Comma-shaped myrmekite grains occasionally occur at the K-feldspar margin and give a deflection sense consistent with the sense of shear across the sample (Figs.2b and 4b). Within the recrystallized material of the S and C-bands there is commonly a dimensionally preferred orientation of quartz and oligoclase grains, even where biotite grains are absent. The resultant secondary foliation is consistent in orientation with the independently interpreted sense of shear across the sample (Simpson & Schmid, 1983; Lister & Snoke, 1984; Passchier & Simpson, 1986). Along C- and S-surfaces where K-feldspar porphyroclasts are absent, vermicular intergrowths of oligoclase and quartz may also occur, particularly where small K-feldspar grains are interspersed. Myrmekite also occurs along the edges of the larger K-feldspar 'tails' of augen (Fig. 3a). In Fig. 3c, a 0.3 m m diameter oligoclase porphyroclast between ribbons of quartz contains irregular zones filled with K-feldspar and a concentration of 10 to 20pm-sized recrystallized K-feldspar grains that have apparently grown in the low-pressure regions at the side of the porphyroclast. The structural state of these K-feldspar grains could not be Myrmekite 0.2mm K - Feld/d/Olig Approx. scale (b) K - Feldspar Q/Olig Q L Q/Olig 0.05mm Approx. scale Q. 2. Thin section sketches from sample SR656 of Santa Rosa mylonite. (a) K-feldspar (orthoclase) porphyroclast between Sand C- surfaces. adapted from Simpson (1985). Myrmekite only along K-feldspar grain margins that are parallel to S;recrystallized K-feldspar in 'tail' regions. (b) Detail of myrmekitic margin to K-feldspar porphyroclast in (a). Myrmekite morphology is consistent with independently determined sinistral shear across the mylonite. Q = quartz; Bi = biotite; Olig = oligoclase. determined; their small size also inhibits accurate refractive index measurements so that optically they are easily mistaken for recrystallized oligoclase grains. The mineralogical differences are readily apparent, however, in the back-scattered mode of the SEM (Fig. 3c). Hope Valley Shear Zone Sample B4-6 (Fig. lb) is strongly linear, weakly planar, Ponaganset biotite augen gneiss occurring at stop 6 of Gromet & O'Hara (1985). c. 2 km east of the dextral Hope Valley shear zone (O'Hara & Gromet, 1985). The lineation, plunging at 12" towards 012". is representative of the region (Gromet & O'Hara, 1985). K-feldspar augen are commonly >5cm long, tabular or ovoid in section, and show welldeveloped 'tails' of recrystallized K-feldspar that extend parallel to the mineral elongation Lineation. Many of the augen contain stable cracks subperpendicular to the mineral elongation lineation. O'Hara & Gromet (1985) suggested deformation temperatures in the range 400-500 "Cfrom the prevalence of evidence of prism ( c ) slip in K-FELDSPAR R E P L A C E M E N T BY M Y R M E K I T E lf5 Fig. 3. Back-scatted SEM photomicrographs of sample SR656.Bright white (b) = biotite; dull white (kf) = K-feldspar; medium grey ( 0 ) = oligoclase;dark grey (q) = quartz; my = myrmekite. (a) K-feldspar megacryst between S and C-surfaccs. Note absence of mymekite at short ends of megacryst. i.e. on sides that face incremental stretch direction. (b) Detail of mynnekitic margin in (a) shows gradual loss of vermicular quark texture and formation of polygonal recrystallized oligoclase grains with distance from mymekite. (c) Oligodase megacryst between quartz ribbons has 'tail' regions and irregular hctures filled with K-feldspar. 266 C. SIMPSON 81 R. P. WINTSCH K - F E L D S P A R R E P L A C E M E N T B Y M Y R M E K I T E 267 TIUC 2. Representative electron microprobe analyses of feldspars from simple M. Rim analyscs taken within first 100pm of the porphyrodast margin. core microdine A SiOz B C core Rim albitic IameUae digodase in D E mymekite F I J 65.61 64.65 64.57 22.52 18.95 18.65 0 0.04 0.05 2.8 0.03 0.04 9.77 0.47 0.57 0.12 15.51 15.35 Total 99.42 99.33 99.63 99.83 99.97 99.99 100.36 100.82 99.62 99.23 Si Al Fe ca Na K Cations for 8 oxygens 2.983 3.002 2.989 2.944 2.996 2.857 2.835 1.032 1.005 1.024 1.074 1.015 1.153 1.171 0 0 0.003 0.001 0.002 0.002 0.002 0 0.033 0.007 0.122 0.165 0.002 0 0.055 0.031 0.m 0.916 0.958 0.847 0.806 0.906 0.933 0.889 0.018 0.002 0.016 0.008 (Ca + Na + K) cationsum 0.963 0 . W 0.961 0.969 0.%7 An Ab or 64.59 22.64 0.01 3.51 9.47 0.14 H KZO NazO FeO CaO 64.89 22.23 0.07 2.58 9.92 0.29 G 64.46 18.93 0.02 0.05 0.61 15.35 4 0 3 64.75 64.83 67.16 68.7 18.39 18.84 20.79 19.76 0.06 0.04 0.05 0 0.01 0 0.7 0.14 0.34 0.81 10.78 11.33 15.78 15.11 0.35 0.04 Tails recrystallized microclillc 2.984 1.031 0.002 0.001 0.827 0.042 0.007 0.913 2.992 1.018 0.96 2.859 1.156 0 0.131 0.002 0.002 0.051 0.907 0.985 0.979 0.965 0.956 0.26 0.05 0.00 3.40 0.68 12.38 5.68 3.17 7.53 94.69 99.09 85.99 94.06 96.78 92.47 1.91 0.23 1.63 16.86 82.34 0.80 13.58 0.15 0.21 85.73 4.40 5.33 0.69 95.45 94.46 quartz and from the presence of recrystallized feldspar and hornblende; the presence of prism ( c ) slip has, however, been taken to indicate deformation temperatures as high as 650°C (Mainprice, Bouchez, Blumenfield & Tubia 1986). In thin section, the h a t e d Ponaganset gneiss contains narrow quartz ribbons, Type 3 of Boullier & Boucbez (1978). that contain 350pm to 1.25mm-sizcd grains, rtcrystaUized biotite, and elongate aggregates of polygonal, 100pm-sized recrystalked K-feldspar and plagiodase (& to AI& grains, all of which define the lineation. "he majority of the K-feldspar porphyradasts in the rock are dose to An,,,Ab,.,Or,., in composition Table 2, columns A to C). Plagiodase lamellae (Table 2, columns D and E) are scattered throughout augen cores and no systematic relationship could be detected between the specific location within the host grain and the lamella composition. In most samples the boundary between the grain and the myrmekite (Table 2,columns F to H)is chemically abrupt at the scale of analysis. However, a few oligodase grains in the myrmekite were seen on back-scattered SEM images to have a narrow (<5 pm) border of more aibitic plagioclase between them and the host K-feldspar (arrowed on Fig. 5a). Biotite and muscovite are present in roughly equal proportions and together constitute about 15% of the matrix to the K-feldspar augen. All biotite grains are SUbparaUel to the foliation plane. Alteration of biotite to chlorite and titanite, with or without Fe-xidcs, is patchy; in some areas alteration is almost nil, whereas in other areas of the same thin &on, alteration is almost complete. This alteration is probably post-tectonic, as tbere is a complete.abscncc of pressure-shadow phenomena around tbc euhedral to anhedral, 750pm to 1.0mm-sized oxide grains. Muscovite occu~sas unoriented grains within plagioclase grain aggngatts and as oriented grains intergrown with biotite. Myrmekite and muscovite are rarely found in close association. Large. fntchmd, originally igneous(?) titanite grains have small, pakr titanite grains around their margins. Rare allanite grains are rimmed by epidote. ?be margins of a typical tabular K-feldspar augen are illustrated in Fig. 4c and d. Myrmekite (An,,+ quartz) oczm in quantity on tbe two long sides of the augen that are parallel to the weak phnar fabric and the strong mineral elongation lineation (Fig. 4c). No myrmekite has been found on the two sides of the grain that face the extension direction (Fig. a), and only minor, sporadic occuf~cnceshave been recorded from the two faces that are perpendicular to the weak foliation and parallel to the Lineation. Thus, the maximum development of myrmekite is along those faces that are subperpendicular to the incremental shortening direction. Figure 5a illustrates the morphology of the myrmekite in sample B4-6. As in sample SR656, quartz vermicules 25-50pm in diameter tend to be subrounded in outline near to the K-feldspar host crystal (Fig. 5b) but become more polygonal and slightly larger with distance from the host grain. The K-feldspar crystal contains numerous inclusions of quartz and white mica, and exsolution p a t h up to 50pm in diameter. Several of the K-feldspar augen contain stable cracks at a high angle to the elongation lineation. These cracks are commonly filled with indusion-free. polygonal, apparently strain-free K-feldspar or quartz grains up to 150pm in diameter. Polygonal K-feldspar grains constitute the major part of the augen 'tails', with quartz and biotite as secondary components. Small grains of K-feldspar occur at the boundary between the host crystal and its 'tail'; thcsc are clear and optically strain-free, and no myrmekite occurs along them (Fig. 4d). DISCUSSION In the previous sections we have shown that quartzplagioclase symplectite (myrmekite) in mylonitic rocks is highly concentrated along asymmetric K-feldspar megacryst faces that are parallel to the foliation planes in the rock. In the following sections we discuss those aspects of the microstructure that a re relevant to interpreting the driving mechanism of the reaction and its non-random distribution. The geometry of the microstructure associated with these augen suggests that the faces that show myrmekite 268 C. SIMPSON & R . P. WINTSCH Flg. 5. Back-scattered SEM photomicrographs of sample B4-6. White (kt) = mcrccme; medium grey (0)= oligoclase; dark grey (q) = quartz; black = holes. (a) Bulbous morphology of myrmekite selvage leaves islands of microcline surrounded by mymekite. (b) Detail of (a) to show rounded to polygonal outlines of quartz vermicules in oligoclasc and narrow rind of more albitic plagioclase (mowed) between microcline and oligoclase. development experienced relatively high normal stress. This is especially safe to infer in the S-C mylonitic microstructure. In the usual interpretation of S-C structure, at relatively low finite strains, where the angle between S- and C-surfaces is approximately 45", the Ssurfaces face the maximum incremental shortening direction and the C-planes are parallel to the bulk rock plane for simple shear (Berth6 et al., 1979). With increasing shear strain the S-surfaces become rotated toward the flow plane and away from the plane of maximum incremental shortening; i.e. the angle between S and C decreases. In sample SR656, the average angle between S - and C-planes is 43", ignoring the small portions of the S-planes that curve asymptotically into the C-planes, and thus the main portions of the S-surfaces are still approximately subperpendicular to the inferred incremental shortening direction. In this orientation the S-planes would experience a normal compressive stress close to the maximum and only a very small component of shear stress. There is thus a very strong positive correlation between predicted high normal stress sites and the Occurrence of myrmekite. Subgrains and recrystallized grains in the 'tail' regions of K-feldspar augen have lattice orientations very similar to that of the adjacent host grain, which suggests that the tails developed by progressive subgrain rotation recrystallization (Porier & Nicolas, 1975). Minor grain-boundary migration was necessary to produce the observed polygonal grain shapes. These observations imply the operation of dislocation glide and climb mechanisms during the deformation of the K-feldspar megacrysts. The plastic deformation of K-feldspar requires at least low amphibolite facies conditions to have prevailed during the deformation (VoU, 1976; Tullis, 1983). Compositional information on the feldspars presented here allows estimates of the temperature of myrmekite development to be made, provided that K-feldspar and plagioclase were in exchange equilibrium during the replacement reactions, and that their compositions have not been modified since their growth. Except for very thin albite rims on some grains, neither feldspar is systematically zoned. Thus, in spite of the obvious replacement textures, exchange equilibrium may have been obtained. Minimum and maximum temperatures of replacement were calculated using the analytical expressions of Stormer (1975) and Whitney & Stormer (1977), which treat maximum disorder and maximum order, respectively. Average minimum and maximum temperatures calculated for the Santa Rosa rocks are 495 "C and 540 "C, in good agreement with the estimates of Anderson (1983). The temperature range calculated for the Hope Valley shear zone samples is 440-445"C, which is consistent with the estimates of O'Hara & Gromet (1985). The two main hypotheses for myrmekite development are those of replacement reactions (Becke, 1908) and solid-state exsolution (Schwantke, 1909). These ideas and K-FELDSPAR R E P L A C E M E N T B Y M Y R M E K I T E 2c9 variations thereon have been summarized and discussed in &tail by Sturt (1970), Widenfalk (1972), Smith (1974) and Phillips (1974; 1980). In Schwantke's lattice diffusion model, a change in solid composition m u m in response to elastic strain gradients, but requires a relatively long time to accomplish (see also Bayly, 1987). The development of subgrains and recrystallized grains provides strong evidence that dislocation glide and climb mechanisms were operating during myrmekite development, so that elastic strain gradients would have been continually created but also destroyed during the plastic deformation. In the solid-state ditrusion model of Simpson (1985), a concentration of Na' ions is built up along the high-pressure sides of K-feldspar porphyroclasts. Na+ has such a small ionic radius that it may diffuse through the K-feldspar lattice either by a vacancy drag effect (e.g. Hanneman & Anthony, 1969; Petrovit, 1973) or by direct interstitial diffusion (Lin & Yund, 1972; Foland, 1974). However, this model does not adequately account for the calcium component of oligoclase/quartz myrmekite and, moreover, a compositional change from alkali to plagioclase feldspar accomplished solely by a lattice diffusion mechanism requires the diffusion of all four major cations, which is unlkely to have occurred at these relatively low temperatures. We therefore rule out the lattice diffusion models of Schwantke (1909) and Simpson (1985) as responsible for the examples described here. The Becke (1908) replacement of K-feldspar by myrmekite can be written in the form: 1.25KA1Si308+ 0.75Na' ( 1 3 7 mole-') ~ + 0.25Caz+ = Na,,7sCa,, UAll.,Si2.7s08 ( l a k c mole-') . + 1.25K+ + SiO,. (1) ( 2 k c mole-') This reaction is fully consistent with the textures of sample SR656. The reaction direction can be inferred both by textural and surface area considerations. Mushroomshaped lobes of mynnekite embay the margins of the K-feldspar grains and the ratio of plagioclase to quartz is approximately 4:1, as predicted by reaction (1). The replacement of myrmekite by K-feldspar, were it to occur, would occur selectively along quartz-plagioclase interfaces, where the local bulk AljSi ratio is closest to that of the K-feldspar. This has not been observed. O n the contrary. the lobes of myrmekite are K-feldspar-free and have the lowest possible surface area, whereas the surface area of the K-feldspar grain is much increased by the replacement. Exactly the same arguments hold for the sample from the Hope Valley shear zone, but in this rock, reaction (1) must be modified for the appropriate composition of plagioclase, c. An,,. The plagioclase to quartz ratio is then 6.7:l which is consistent with observed ratios of about 8:l in thin sections. Reaction (1) thus describes the texture in Fig. 5a very well, when the reaction proceeds to the right. The Phillips (1980) modification of the Becke model requires the introduction of H,O into the reactions, so that muscovite forms at the expense of the existing K-feldspar and plagioclase, and K,O is removed from the local environment. The Santa Rosa mylonites (sample SR656) were almost certainly deformed in the presence of an aqueous phase (Anderson, 1983). The alteration of biotite to chlorite in the Ponaganset gneiss indicates that it too had access to an aqueous fluid, but this CouId have occurred at a later stage of metamorphism. In order to investigate the possible effects of changing the activities of potassium, sodium and hydrogen ions on the system K-feldspar-muscovite-albite, we illustrate solid-fluid equilibria for this system at 400 "C and 450 "C, 300 MPa (Fig. 6). Reaction (2) in Fig. 6 can be represented, using ideal formulae, by: Musc Qtz + 2K' = 3Ksp + 2H' (2) + and reaction (3) by: Musc + Qtz + 3Na' = 3Ab + K+ + 2H'. (3) Figure 6 shows that with increasing temperature, the stability field of albite increases with respect to K-feldspar and muscovite, and the K-feldspar stability field increases with respect to muscovite. Increasing pressure enlarges the stability field of muscovite relative to both feldspars, and enlarges the plagioclase field relative to K-feldspar. The diagram thus shows that plagioclase is stabilized relative to K-feldspar by increasing temperature or pressure, either of which could theoretically lead to the development of albite replacement. , The production of myrmekite with muscovite during retrograde metamorphism (Phillips, Ransom & Vernon, 1972) is also elucidated by Fig. 6. Consider a Kfeldspar-plagioclase-quartz-bearing granite at 450 "C in equilibrium with a fluid (point X in Fig. 6). With cooling, the muscovite stability field grows and by 400°C totally encompasses point X, whereupon muscovite would precipitate from the fluid. Such precipitation would lower the local activity of aqueous K+, thereby destroying the equilibrium of reaction (1). K-feldspar would then dissolve and be replaced by plagioclase plus quartz. This reaction is predicted for all rocks, but the modal significance is directly proportional to the concentration of ions in the fluid and to the amount of fluid present in the rock. The amount of mynnekite formed during cooling would thus be directly proportional to porosity. Under these conditions a close textural relationship would be expected between muscovite and myrmekite. Although a small quantity of muscovite may have been produced during my-nnekite formation in mylonitic rocks of the Hope Valley Shear zone, the general lack of association between myrmekite and muscovite in both sample studied here, and the sparsity of muscovite in the Santa Rosa samples, argue against reactions (2) and (3) and in favour of reaction (1). The equilibrium constant, log K1,for activity balance of reaction (1) is given by: + 1.25 (a,+) + log (sP) log K,= log (a& - 1.25 log (aw) - 0.75 log (a,,+) 0.25 log (+.z+). - (4) The forward reaction (1) can therefore be driven by an 4.70. 3 4.80. 0" 4.90 5.00 5.1 0 -2A MUSCOVITE I I I I I 4:4 '/ QTZ SAT. 300 MPa 1 / / h 3.8 4.80 4.90 5.0C MUSCOVITE /3 ALBITE 300 MPa + QTZ I I unstrained K-feldspar 4.4 Fig. 7. Calculated effect of adding 100 cal(418 J) to the free energy of K-feldspar on decreasing the size of the stability field of K-feldspar at 450 "C, 300 MPa at quartz saturation. Solid lines show stability fields for strain-free crystals; dashed line shows reduced size of stability field of strained K-feldspar. For further discussion, see text. pb. 6. Calculated effect of decreasing temperature from 450 "C to 400"Con the stability fields of muscovite, albite, and 'equilibrium K-feldspar' (Helgeson ei ul., 1978) at 300 MPa (3 kbar) with quartz saturation, in t e r n of the log aqueous activity ratios Na+/H+ and K+/H*. Note that with decreasing temperature, the K-feldspar field expands relative to albite, and the muscovite field expands relative to both feldspars. Similar changes occur for calcium-bearing plagioclase. For further discussion see text. Numbers 1,2 and 3 refer to equations (l), (2), (3) in text. w 0 - a. Z t 5.1( K-FELDSPAR R E P L A C E M E N T BY M Y R M E K I T E increase in the activities of K-feldspar or of aqueous Na+ and G2+. Assuming that the coupled diffusion of and silicon is relatively slow in comparison with that of aqueous Na' and Ca2', the volume change Of the solid phases during the reaction of K-feldspar to form one mole of the plagioclase product from reaction (1) is 14 cc mole-' plagioclase, a volume decrease of more than 10%. The forward reaction (1) will thus also be favoured by an increase in the normal stress on the K-feldspar grah and is more likely to occur on the edges of the K-feldspar grains that faced the incremental shortening direction. The reverse reaction (1) will be favoured in the lower pressure, 'W regions of the augen. The effect of an increase in applied differential stress on the K-feldspar grains during steady-state deformation would be to increase the dislocation density and decrease the rotation-recrystaked grain-size (Twiss, 1977) especially on the sides of the augen that faced the incremental shortening direction. The addition of strain energy associated with dislocations could be sufficient to shift the K-feldspar stability field boundaries on Fig. 6, as shown on Fig. 7. Material close to the boundary of reaction (1) would be stable as K-feldspar at low values of stored strain energy, but would form plagioclase (i.e. reaction (1) would proceed to the right) at higher energy values. Point H on the plagioclase-K-feldspar stability field boundary (Fig. 7) describes the composition of a fluid in equilibrium with unstrained solids under a hydrostatic stress state, i.e. the composition of the fluid buffered by the rock prior to deformation. Tbis equilibrium strain-free state is compared to a deformed state (dashed m e on Fig. 7), calculated by raising the free energy of K-feldspar by 100 cal (418 J); the energy could be associated with tangled dislocations. The composition of the fluid H is unchanged by the deformation, but it is no longer in equilibrium with the deformed K-feldspar grains; such a fluid would have a composition along the dashed curve. Fluid H is now entirely within the plagioclase stability field, and to change its composition to be in local equilibrium with the strained K-feldspar, the activity of aqueous Na' must drop, and that of aqueous K' must rise. This is readily accomplished by the replacement of the strained regions of K-feldspar by plagioclase. During high-temperature steady-state deformation by recovery-accommodated dislocation creep, much of the internal strain energy of the lattice is reduced by the climb of dislocations to form low-energy tilt walls (e.g. TuUs & Yund, 1985). However, at the lower temperatures of deformation in the Santa Rosa and Hope Valley mylonites, regions close to the grain boundaries of large K-feldspar augen probably contained a high density of tangled dislocations, particularly along the edges of the grains that faced the incremental shortening direction, and a dislocation density of 1010cu-2 is readily obtained in such tangles (Knipe & Wintsch, 1985; Wintsch & Dunning, 1985). To determine whether or not dislocation densities reached such high values during the deformation of the K-feldspar rims, a transmission electron microscope study M of the myrmekite margins of the augen would be of great interest. Unfortunately, neither the present dislocation density nor the recrystallized grain structure in the augen is necessarily the one that was present prior to myrmekite formation. Our examples contain clear evidence for syntectonic myrmekite development and most of the myrmekitic aggregates have undergone deformation. A higher dislocation density now present in the vicinity of the myrmekite could reflect post-formation strain, whereas a lower dislocation density in that region could simply indicate that grain boundary migration had occurred, or that reaction (1) had led to the replacement of all the strained feldspar. In addition, the original size of rotation-recrystallized new grains would be modified by grain boundary migration. Etheridge & Hobbs (1974), Knipe (1981) and Lee, Peacor, Lewis & Wintsch (1986) noted that chemical changes accompany recrystallization in micas; this is to be expected if the pressure and temperature conditions of &formation are different from those at the time of formation of the protolith. In their examples, however, the free energy drop associated solely with the chemical metastability of the micas was insufficient to cause recrystallization of micas of a different composition; strain energy must have been added. The mica examples parallel the mymekite discussed here, with the important differences that our samples require a net change in the mineral assemblage, and that the reaction mechanism was one of ion exchange. THE MODEL From the foregoing discussion, a conceptual model for the myrmekite replacement reaction can be developed and this is presented in Fig. 8. The model requires only a pre-existing K-feldspar megacryst in a matrix of quartz, plagioclase and K-feldspar. A non-hydrostatic stress state develops in the granitic rock and the large K-feldspar augen act as stress risers, such that normal stresses are highest along those margins facing the incremental shortening direction (open arrows on Fig. 8). Under the low amphibolite faaes conditions of our samples, the imposed stress causes concentrations of elastic strain energy, and strain hardening (Taylor, 1934), associated with relatively immobile dislocation tangles, is likely to occur along the K-feldspar grain boundaries (e.g. T u b & Yund, 19?7; 1987). The composition of the ambient aqueous fluid present would be buffered in the unstrained grain boundaries by the K-feldspar-plagioclase-quartz assemblage (Fig. 6). However, along the sfrained margins of the K-feldspar, the equilibrium is destroyed and a replacement reaction ensues (Fig. 7). This replacement is accompanied by a 10% volume loss, so that the high imposed stresses on the grain margins facing the incremental shortening direction are relieved. The replacement reaction is thus a form of incongruent pressure solution (Beach, 1982). Reaction (1) is a replacement reaction and so sources and sinks of aqueous components must be considered. The 272 C . SIMPSON & R. P. WINTSCH plag l a y e r qtz 1a y e r (Ca++, .Na+> /' (Ca++. Na+> augen 1ayer Ksp. replacement by plag. gc s poss b l e plag. rep acement by Ksp. 'VQ" Fig. 8. Conceptual model for development of deformation-induced myrmekite. High normal stress at some K-feldspar megacryst boundaries favours replacement of K-feldspar by plagioclase + quartz. Sources of plagioclase components and sinks for K+ may be the reverse reaction in the 'tails'. Open arrows indicate sites of high normal stress. For further discussion, see text. replacement of K-feldspar occurred at approximately constant Al and Si and can be modelled as a cation exchange reaction. Aqueous K' is large relative to Na+ and Ca2' (Helgeson & Kirkham, 1976) and is thus more likely to be concentrated in the lower normal stress 'tail' regions of the augen. In fact, K-feldspar is concentrated in the augen 'tails' almost at the expense of plagioclase, and it also occurs in fractures, unlike plagioclase. 'The migration of K' to the lower normal stress regions would increase the K+/Na' activity ratio, such that the fluid precipitates K-feldspar. Such an interpretation is required by the filling of fractures in plagiwlase by K-feldspar (Fig. 3c). Potassium may thus be conserved on the scale of a thin d o n . The possibility exists that muscovite may also act as a sink for some of the K+ released during myrmekite formation. However, the paucity of muscovite in the Santa Rosa samples argues against it acting as the predominant potassium sink in these rocks. The sodium-calcium budget is more difficult to deal with, because calcium sources are less obvious; the replacement of K-feldspar clearly offers an important sink. One answer to the source problem may again be found in Fig. 2.In the 'tail' region labelled kf, K-feldspar is Seen to. surround and embay plagioclase grains - a texture that may be evidence of plagioclase replacement by K-feldspar. Another possible local source of calcium could be provided by the dynamic recrystallization of plagioclase to a lower temperature, more sodic composition. It is, therefore, quite possible that the strainenhanced myrmekitic replacement of K-feldspar may occur with complete elemental conservation, i.e. a closed system on the scale of centimetres. The possibility of a totally open system, in which the composition of the fluid remains fixed by equilibria in an adjacent rock, may be discounted for our examples because such a fluid would invade all grain boundaries, not just the highly stressed ones. Rather, the K-FELDSPAR R E P L A C E M E N T B Y M Y R M E K I T E 273 concentration of the reaction at selected microstructurd sites demands gradients in fluid composition on the scale of mm, controlled by reactions in the local (cm-scale) environment. We do not, however, concur with the contention of Hibbud (1979; 1987) that textures consistent with 'mobile feldspar' necessarily indicate deformation in the presence of H,O-bearing magmatic fluids, i.e. deformation of an incompletely crystallized magma. Hibbard (1987) considered that the presence of myrmekite in mylonitic granite gneisses reflects relocation of late-stage magmatic fluids and is not the result of deformation of a completely crystallized rock. The examples that we have discussed here contain myrmekitic intergrowths that are clearly on the high-strain sides of the K-feldspars, in direct contrast to the low-pressure sites that are required by Hibbard's hypothesis. The high temperatures (>450°C) that are apparently necessary for the dynamic recrystallization of all feldspars at natural strain rates (e.g. Voll, 1976; Tullis, 1983) certainly favour the production of a mylonite gneiss if deformation follows immediately upon, or at a very late stage in, the crystallization of a pluton, but this is surely not the only way to produce amphibolite facies metamorphism in a tectonic environment. Deformation of a pluton buried more than about 15km beneath thrust nappes could .result in the formation of mylonitic gneisses at that depth, regardless of the age gap between pluton crystallization and deformation. Indeed, the myrmekitebearing gneisses of the Maggia Nappe in Switzerland formed during Alpine deformation of Hercynian-aged plutonic rocks (Simpson, 1983). Similarly, deformation of plutonic rocks of the lowermost upper crust (c. 15 km and below) during large-scale extension, as in the metamorphic cure complexes of Nevada and Arizona, can produce mylonitic gneisses that contain strain-related myrmekites, e.g. the Santa Catalina mylonites in southern Arizona. In addition, myrmekite is also found in retrograde mylonitic rocks of the Willyama Supergroup, Broken Hill, Australia (Phillips et d.,1972) and in metamorphosed sedimentary rocks (e.g. Nold, 1984). We suggest, therefore, that a significant number of myrmekite-bearing mylonitic gneisses may have formed some considerable time after crystallization of the magma was complete and that the presence of myrmekite, or of apparently mobile K-feldspar, should not be taken a priori to indicate deformation before crystallization was complete. Finally, the 'late-stage magmatic' argument is unsubstantiated for the samples we describe because the temperatures of formation of the rnyrmekite are entirely subsolidus. As Ashworth (1972). Smith (1974) and Phillips (1980) pointed out, there may be many different geneses for myrmekitic intergrowths. We have discussed here only myrmekite that is clearly related to the strain state in the rock. Other Occurrences of mynnekite that might bear closer examination include its widespread development in undeformed granites. Although we do not rule out the possibility of myrmekite formation in tectonically undeformed granites in the absence of a local differential stress field, nevertheless most granites contain quartz that shows evidence of lattice strain in the form of undulatory extinction and subgrain formation. This lattice strain presumably occurs in response to differential volume changes that result from Merences in the thermal conductivity and elastic anisotropy of adjacent minerals in the cooling rock. Kirby & Etheridge (1981) suggested that deformation-aided exsolution of Ca-clinopyroxene from orthopyroxene could occur due to thermal conductivity differences among adjacent minerals in a cooling rock. We suggest that myrmekite could also form due to local stress differences on K-feldspar grain boundaries during cooling of a completely crystallized granitic rock. CONCLUSIONS A geometric relationship exists in many granitic mylonites between myrmekite development on K-feldspar augen and the finite, and inferred incremental, strain state of the rock. A magmatic origin for this mynnekite is ruled out by its geometry and the subsolidus temperatures of formation. The absence of an association between muscovite and myrmekite in the rocks that we examined suggests to us that replacement reaction (1) predominated. Reaction (1) involves a net volume decrease and is thus favoured at sites of high normal stress on K-feldspar grain boundaries. In the samples we have discussed, myrmekite is syntectonic and found only on the high-pressure sides of the augen; its presence cannot be attributed solely to chemical processes. The presence of small, dynamically recrystallized new grain boundaries in the augen 'tails' demonstrates that in spite of the relatively low temperatures, deformation by crystal plastic mechanisms did OCCUT in these K-feldspar megacrysts. Elastic strain energy or energy associated with dislocations localized the replacement reactions to sites of high normal stress, causing the observed asymmetrical distribution of mynnekite. ACKNOWLEDGEMENTS Financial support was provided in part by NSF grants EAR-8121438, EAR-8305846 and EAR-8606120 to CS and EAR-8313807 and EAR-8618305 to RPW. Helpful discussions with B. Bayly, G. W. Faher, B. E. Hobbs, J. A. Tullis and R. A. Yund are gratefully acknowledged. The comments of R. H. Vernon and an anonymous reviewer helped us to clarify parts of the text. I. D. Fitzgerald, T. E. LaTour and R. A. Yund are thanked for their incisive comments on a very early (1985) version by CS. Assistance with electron microprobe and SEM analyses was ably provided by T. N. Solberg and C. Henderson. REFERENCES Anderson, J. R., 1983. Petrology of a portion of the Eastern Peninsular Ranges mylonite zone, southern California. Contributions to Mineralogy and Petrology, 84,253-271. Ashworth, J. R., 1972. Myrmekites of exsolution and replacement origins. Ceobgical Magazine, 189,45-62. n4 C . SIMPSON 81 R. P. WINTSCH Bayly, B., 1987. Nonhydrostatic thermodynamics in deforming rocks. CMOdian J o d Of Earth sciences, 24,572-579. Beach, A., 1982. Deformation mechanisms in some cover thrust sheets from the external French Alps. J o d of Sturcwal Geology, 4, 137-150. Becke, F., 1908. Ueber Mymekite. Schweizerische Miwalogbche und Petrogrophische Minedungen, 27,377-390. Berth&,D., Choukroune, P. & Jegoum, P., 1979. Orthognciss, mylonite and non-coaxial deformation of granites: the example of the South Armorican shear zone. J o d of Structural Geology, 1,3142. Bhattacharyya, C.. 1971. Mynnekite from the charnockitic rocks of the Eastern Ghats, India. Geological Magazine, 108, 433-438. Bianchetti, S. P. & Reeder, R. J., 1985. Variable dissolution rates of deformed and undeformed calcite. Geological Society of Anuricn Abstracts wilh Programs, 17,7, 522. Binns, R. A., 1%. Granitic intrusions and regional metamorphic rocks of Permian age fmm the Wongwibinda district, northeastern New South Wales. Proceedings of the Royal Society of New South Wales, 99,5-36. Boullier, A. M.& Bouchez, JCL., 1W8. Le quartz en rubans dans les mylonites. Bulktin Societie Gtologique de France, 7, 253-262. Bowers, T. S., Jackson, K. J. & Helgeson, H. C., 1984. Equilibrium Activity Dipgrams for Coexbfing Minerah and Aqueous Solutions at Pressures and Temperatures to 5 kb and 600 "C.Springer-Verlag, New York. Debat, P., Soula. J. C., Kubin, L. & Vidal, J. L., 1978. Optical studies of natural deformation microstructures in feldspars (gneisscf and pegmatites from Occitanie, southern France). L a , ll,133-146. Engel, A. E. J. & Schultejahn, P. A., 1984. Late Mesozoic and Cenozoic tectonic history of south central California. Tectonics, 3, 659-676. Erskhe, B. G. & W e d , H.-R., 1985. Evidence for Late Cretaceous crustal thinning in the Santa Rosa mylonite zone, southern California. Geology, W,274-277. Eskola. P., 1914. On the petrology of the Orijarvi region in southwestern Fiiand. Bulletin de la Conunisswn G6ologque de Finfande, 40,285 pp. Etheridge, M. A. & Hobbs, B. E., 1974. Chemical and deformational controls on recrystallization of mica. Contributions to Mineralogy and Petrology, 43, 111-124. Foland, K. A.. 1974. Alkali diffusion in orthoclase. In: Geochemical Transport and Kinetics (eds Hoffman, A. W. et al.), pp. 77-98, Carnegie Institute, Washington. Futterer, K., 1894. Ueber granitporphyr von der Greisscharte in den Zillerthaler Alpen. News Johrbuch jZr Geologic und P&ntologie, 9,509-553. Gromet, L. P. & O'Hara, K., 1985. The Hope Valley shear zone - a major late Paleozoic ductile shear zone in southeastern New England. In: NEIGC 77th. Annual Meeting, Yale Univeniry. Connecticut, Guidebook for Fwldtrips, 6,277-295. Hanmer, S. K., 1982. Microstructure and geochemistry of plagiodase and microcline in naturally deformed granite. Journal of Structural Geology, 4, 197-214. Hanneman, R. E. & Anthony, T. R.. 1969. Effects of non-cquilibrium segregation on near-surface diihsion. Acta Metalhgica, 17, 1133-1140. Harker, A., 1932. Metamorphism. Methuen, London. Helgeson, H.C. & Kirkham, D. H.. 1976. Theoretical prediction of the thermodynamic behaviour of aqueous electrolytes at high pressures and temperatures. HI.Equations of state for aqueous species at infinite dilution. American Journal of Science, 276, 97-240. Helgeson, H. C., Delany, J. M.,Nesbitt, H. W. & Bud, D. K., 1978. Summary and critique of the thermodynamic properties of rock forming minerals. American Journal of Science, 278A, 1-229. Hibbard, M. J., 1979. Mymekite as a marker between preaqueous and postaqueous phase saturation in granitic systems. Geological Society of AIWI~CM BJlehir, 90, 1047-1062. Hibbard, M. J., 1987. Deformation of incompletely magma systems: granitic gneisses and their tectonic implications. J o d of Geology, 95,543-561. Keith, S. B.. Damon, P. E., Reynolds, S. J., Shafiqulla, M., Livingston, D. E. & Pushkar, P. D.. 1980. Evidence for multiple intrusion and deformation within the Santa Catalina&con-Tortolita metamorphic core complex. In: Cordillrrm Metamorphic Core Complexes (eds Crittcnden, M. D., Jr., Coney, P. J., & Davis, G . H.), Geological Society of Americo Memoir, l53,217-267. Kemck, D. M.,1986. Dislocation strain energy in the A g i o , polymorphs. Physics and Chemistry of Minerals, W, 221-226. Kirby, S. H. & Etheridge, M. A., 1981. Exsolution of Ca-clinopyroxene from orthopyroxene aided by deformation. PhySicr Md Chemistry Of Miner&. 7,105-109. Knipe, R. J., 1981. The interaction of deformation and metamorphism in slates. Tectonophysics, 78,249-272. Knipe, R. J. & Wintsch, R. P. 1985. Heterogeneous deformation, foliation development, and metamorphic processes in a polyphase mylonite. In: Advane- in Physical Geochenuitry, 4, Metamorphic Reactiotw (eds Thompson, A. B. & Rubie, D. C.), pp. 180-210, Springer-Verlag, New York. LaTour, T. E., 1981. Geochemicd model for the symplectic formation of mynnekite during amphibolite grade progressive mylonitintion of granite. Geological Society of America Abstracts wifh Programs,19,741. LaTour, T. E. & Barnett, R. L., 1987. Mineralogical changes accompanying mylonitintion in the Bitterroot dome of the Idaho batholith: implications for timing of deformation. Geological Society of Amcrico Bulletin, 98,356-363. Lee, J. H., Peacor, D. R., Lewis, D. D. & Wmtsch. R. P. 1986. Evidence for syntectonic crystallization for the mudstone to slate transition at Lehigh Gap, Pennsylvania. Journal of Structural Geology, 8,767-780. Lin, T.-H & Yund. R. A., 1972. Potassium and sodium selfdif€usion in alkali feldspar. Coruriburiotw to Mineralogy and Petrology, 34,177-184. Lister, G . S . & Snoke, A. W.,1984. S-C mylonites. Journal of Structural Geology, 6, 617-638. Mainprice, D., Bouchez, J.-L., Blumenfeld, P. & Tubia, J. M., 1986. Dominant c slip in naturally deformed quartz: implications for dramatic plastic softening at high temperature. Geology, 14,819-822. Nold, J. L., 1984. Myrmekite in Belt Supergroup metasedimentary rocks-northeast border zone of the Idaho batholith. American Minrralogist, 69,1050-1052. O'Hara, K. & Gromet. L. P., 1985. Two distinct Precambrian (Avalonian) terranes in southeastern New England and their late Paleozoic juxtaposition. American Journal of Science, 285, 673-709. Parslow, G. R., 1971. Variations in mineralogy and major elements in the Caimsmore of Fleet granite, SW Scotland. Lithos, 4,43-55. Passchier, C. W. & Simpson, C., 1986. Porphyroclast systems as kinematic indicators. J o d of Smccnual Geology, 8,831-843. PetroviC, R., 1973. The effect of coherency stress on the mechanism of the reaction albite + K+ = K-feldspar + Na' and on the mechanical state of the resulting feldspar. Contributions to Mineralogy and Petrology, 41, 151-170. Phillips. E. R., 1974. Mymekite-one hundred years later. Litlux, 7, 181-194. Phillips, E. R., 1980. On polygenetic myrmekite. Geologml Magazine, 17,29-36. Phillips, E. R. & Carr, G. R., 1973. Myrmekite associated with alkali feldspar megacrysts in felsic rocks from New South Wales. L W , 6,245-260. Phillips, E. R., Ransom, D. M. & Vernon, R. H., 1972. Myrmekite and muscovite developed by retrograde metamorphism at Broken Hill, New South Wales. Minerologiccrl Magazine, 38,570-578. K-FELDSPAR R E P L A C E M E N T BY M Y R M E K I T E Poirier, J. P. & Nicolas, A., 1975. Deformation-induced rtcrystallization due to progressive misorientation of subgrains with special reference to mantle peridotites. Journal of Geology, 83,707-720. Rubie, D. C. 1983. Reaction enhanced ductility: The role of solid-solid univariant reactions in deformation of the crust and mantle. Tectotwphysics, 96,331-352. Sarma, S. R. & Raja, N., 1959. On myrmetite. Quarterly Journal of the Geolorical, M i n e r a l o u and Metalluraical Society of I&, 31, Schrever. W..1958. Die Quarz-FeldsDat-GeNge der miamatischin desteh von Vilshofen an der Donau. N & a Jahrbkh f i r Mineralogie, 92,147-170. SchultejaJm, P. A., 1984. The Yaqui Ridge antiform and detachment fault: Mid-Cenozoic extensional terrane west of the San Andreas fault. TectoniCr, 3, 677-691. Schwantke, A., 1909. Die beimischung von Ca in Kalifeldspat und die Mymekitbildung. Zentralblati fiir Geologic und P & ~ l o g i e , 311-316. Sederholm, J. J., 1916. On synantetic minerals and related phenomena (reactions rims, corona minerals, kelyphite, myrmekite, etc.). Bullerin de la Commission Gkologque de Ehlkande,48,146 pp. Sharp, R. V., 1966. Ancient mylonite zone and fault displacements in the Peninsular Ranges of southern California. Geological Society of America Special Paper, 101, 333. Sharp, R. V., 1979. Some characteristics of the Eastern Peninsular Ranges mylonite zone. United States Geological Survey Open F i i Report, 79-1239,258-267. Shelley, D., 1964. On myrmekite. American Mineralogist, 49, 41-52. Simpson, C., 1981. Ductile shear zones: a mechanism of rock deformation in the orthogoekes of the Maggia Nappe, Ticino, Switzerland. Unpubl. PhD Thesis, ETH-Zuirch, Switzerland. Simpson, c . , 1982. The structure of the northern lobe of the Maggia Nappe, Ticino, Switzerland. Eclogae Geologicm Helveripc, 75,495-516. Simpson, C., 1983. Strain and shape fabric variations associated with ductile shear zones. Jountal of Smctural Geology, 5, 61-72. Simpson, C., 1984. Borrego S p M g - Santa Rosa mylonite zone: a late Cretaceous west-directed thrust in southern California. Geology, l2,8-11. Simpson, C., 1985. Deformation of granitic racks a c r m the brittle-dude transition. Journal of Smr~hual Geology, 7 , 503-511. Simpson, C. & Schmid, S. M.,1983. An evaluation of criteria to deduce the sense of movement in sheared rocks. Geological society of America Bullcrin, 94,1281-1288. Smith, J. V., 1974. Feldspar Minerals 2: Chemical and texhvol properties. Springer-Verlag. New York. Spencer, E., 1945. Myrmekite in graphic granite and in vein perthite. Mineralogical Magazine, 27,79-98. Stormer, J. C., Jr.. 1975. A practical two-feldspar geothermometer. American Mineralogist, 60,661-674. Shut, B. A., 1970. Exsolution during metamomhism with particular reference to feldspar solid &lutions. Mineralogical Magazine, 37,815-832. Taylor, G. I., 1934. The mechanism of plastic deformation of crystals: Part I - Theoretical. Proceedings of rhe Royal Society of London, Serics A , 145,362-387. l k o d o r e , T. G., 1970. Petrogenesis of mylonites of high metamorphic grade in the Peninsular Ranges of southern California. Geological Society of America BuUetin, 81,435-450, I ~ - - . Zn Tullis, J. A., 1983. Deformation of feldspars. In: Feldspar Mineralogy (ed. by Ribbe, P. H.),Mineralogical Society of America short c o m e notes, 2 (2nd. edn), 297-323. TuUs, J. A. & Yund, R. A., 1977. Experimental deformation of dry Westerly granite. Journal of Geophysical R u e d , 62, 5707-5718. Tullis, J. A. & Yund, R. A., 1385. D y n d c recrys&z&on of feldspar: a mechanism for ductile shear zone formation. Geology, U,238-241. T W , J. A. & Yund, R. A., 1987. Transition from catadastic h w to dislocation creep of feldspar: Mechanisms and microstructures. Geology, 15,606409. Twiss, R. J., 1977. Theory and applicability of recrystallized grain size paleopiewmeter. R u e and Applied Geophysics, 115, 227-244. Vernon, R. H.,1987. Oriented growth of sillimanite in andahsite, PLaatas-Juan Tabo area, New Mexico, U.S.A. CPMdian Journal of Earth ScicnCe, 24,58%590. Vernon, R. H., Williams, v. A. & D'Arcy, W. F., 1983. Grain-size reduction and foliation development in a deformed granitoid batholith. Tectotwphysics, 92, 123-145. Vidal, J-L., Kubin, L., Debat, P. & Soula, J. C., 1980. Deformation and dynamic recrystallization of K-feldspar augen in orthogneiss from Montagne Noire, Occitania, southern France. Lithos, U,241-255. VoU, G., 1976. Recrystallization of quartz, biotite, and feldspars from Erstfeld to the Leventina Nappe, Swiss Alps. and its geological significance. Schweizerische MincrrJogirchc und Petrographische Miaeilungen, 56,641-647. Watts, M. J. & Williams, G. D., 1983. Strain geometry, microstructure and mineral chemistry in metagabbro shear zones: a study of softening mechanisms during progressive mylonitization. Journal of Structural Geology, 5,507-518. White, S . & Kaipe, R. J., 1978. Transformation and reaction e n h d ductility in rocks. Journal of the Geological Society of London, WS, 513-516. Whitney, J. A. & Stormer, J. C., Jr., 1977. The distribution of NaAISiO, between coexisting microcline and plagioclase and its effect on the geothermometric calculations. AmcricM Mineralogist, 62,689-891. Widenfalk, L., 1972. Mymekite-Like intergrowths in larvikite feldspars. L k ,5,255-267. Wintsch, R. P., 1985. The possible effects of deformation on chemical proccssts in metamorphic fault zones. In: Adwnccs in physical geochemistry, 4, Metamorphic Reactions (eds Thomp son, A. B. & Rubie. D. C.). pp. 251-268, Springer-Verlag, New York. Wmtsch, R. P. & Andrews, M. S., 1988. Deformation-induced growth of sillimanite: stress minerals revisited. J o d of Geology (in prtss). Wintsch, R. P. & Dunning, J., 1985. 'Ihe effect of dislocation density on the aqueous solubility of quartz and some geologic implications. Journal of Geophysical Research, W,3649-3657. Wintsch, R. P. & Knipe, R. J. 1983. Growth of a zoned plagioclase porphyroblast in a mylonite. Geology, 11,360-363. Wintsch, R. P. & Sutter, J. F.,1986. A tectonic model for the Late Paleozoic of southeastern New England. Journal of Geology, 94,459-472. Yund, R. A. & Tullis, I. A., 1986. The effect of ductile deformation on plagioclase breakdown kinetics. Geological Society of A m e h Abstracts with Programr, 18,799. Received 25 January 1988; revision accepted 27Jwu 1988