American Mineralogist, Volume 90, pages , 2005

Transcription

1 American Mineralogist, Volume 90, pages , 2005 Electron probe (Ultrachron) microchronometry of metamorphic monazite: Unraveling the timing of polyphase thermotectonism in the easternmost Wyoming Craton (Black Hills, South Dakota) PETER S. DAHL, 1, * MICHAEL P. TERRY, 2 MICHAEL J. JERCINOVIC, 3 MICHAEL L. WILLIAMS, 3 MICHAEL A. HAMILTON, 4, KENNETH A. FOLAND, 5 SUSANNE M. CLEMENT, 6, AND LAVERNE M. FRIBERG 6 1 Department of Geology, Kent State University, Kent, Ohio 44242, U.S.A. 2 Bayerisches Geoinstitut, Universität Bayreuth, D Bayreuth, Germany 3 Department of Geosciences, University of Massachusetts, Amherst, Massachusetts 01003, U.S.A. 4 J.C. Roddick Ion Microprobe (SHRIMP II) Laboratory, Geological Survey of Canada, 601 Booth St., Ottawa, Ontario, K1A 0E8, Canada 5 Department of Geological Sciences, The Ohio State University, Columbus, Ohio 43210, U.S.A. 6 Department of Geology, The University of Akron, Akron, Ohio 44325, U.S.A. ABSTRACT A metapelite from the easternmost Wyoming craton (Black Hills, South Dakota) has been analyzed by microstructural methods to unravel polyphase deformational history associated with Ma assembly of southern Laurentia. Three deformational fabrics are recognized in oriented thin sections: an ENE-trending S 1 fabric, preserved as oblique inclusion trails in garnet porphyroblasts; a NNWtrending S 2 fabric, preserved as microlithons in the rock matrix; and a ßattening fabric, S 3, which transposed S 1 -S 2 and dominates the matrix. A complex monazite porphyroblast has been analyzed in situ with the electron microprobe (Ultrachron) to constrain the timing of S 1 -S 3 fabric formation associated with monazite growth. The core of this grain uniquely preserves the S 1 -S 2 fabrics as sigmoidal inclusion trails. The mean total-pb age of this domain is 1750 ± 10 Ma (all dates reported at 95% conþdence; n = 39 spots), which is equivalent to the published 207 Pb/ 206 Pb age for the same domain. These results validate the total-pb dating method in general and the Ultrachron in particular, for reliable age determination in low-th monazite, and are interpreted as 1750 Ma minimum ages for the S 1 -S 2 fabrics and sequential, D 1 -D 2 collisional events that imposed them (~N-directed arc accretion and ~E-W continental collision, respectively). A higher-th,y rim of this same Rosetta grain truncates the S 1 -S 2 sigmoid, and is associated with resorption textures in garnet porphyroblasts, coupled release of Y, and an S 3 fabric that pervasively overprinted S 1 -S 2 in the rock matrix. The mean Ultrachron date of this domain is 1692 ± 5 Ma (n = 17 spots), which is slightly younger that the published isotopic age for all monazite rims combined. These results support a ~ Ma timeframe for localized doming (D 3 ) related to granite magmatism, the onset of which has been dated independently at 1715 ± 3 Ma. The timing of post-d 3 cooling through 350 and 300 C is constrained by 40 Ar/ 39 Ar dates of ~1610 and ~1480 Ma obtained for separates of D 3 matrix muscovite and biotite, respectively, which are interpreted as closure ages. This study shows that fabrics in poly-deformed rocks can be dated by linking monazite spot ages to key microtextures. Further, the results of this micrometer-scale study enhance previous knowledge of local thermotectonism (Black Hills) and regional terrane assembly (Laurentia). INTRODUCTION * pdahl@kent.edu Current address: Jack Satterly Geochronology Laboratory, Department of Geology, Earth Sciences Centre, University of Toronto, Toronto, Ontario, Canada M5S 3B1. Current address: Department of Geology, Kent State University, Kent, OH Monazite is a widespread phosphate mineral that occurs as an accessory in diverse crystalline rocks. In poly-deformed rocks, it commonly preserves discrete age-composition domains within single thin sections, individual grains, and/or speciþc microtextural environments. As such, occurrences of polygenetic monazite are utilized increasingly for deciphering complex histories of grain growth, recrystallization, dissolution, and regrowth (e.g., Lanzirotti and Hanson 1996; Vry et al. 1996; Ayers et al. 1999; Williams et al. 1999; Pyle and Spear 1999; Vavra and Schaltegger 1999; Wing et al. 2003). Moreover, the record of such grain-scale processes in monazite can be linked to regional thermotectonic histories when integrated with a full complement of petrofabric, petrologic, and geochronologic data (e.g., Zhu and OʼNions 1999a, 1999b; Montel et al. 2000; Terry et al. 2000a, 2000b; Pyle and Spear 2003; Williams and Jercinovic 2002; Foster and Parrish 2003; Shaw et al. 2001; Gibson et al. 2004). For constraining absolute ages of monazite at the micrometer scale, chemical (henceforth, total-pb) dating with the electron microprobe has matured into a viable complement to isotopic X/05/ $05.00/DOI: /am

2 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE 1713 dating with the ion microprobe (e.g., Dahl et al. 2005). In addition, whereas traditional electron microprobes (e.g., the Cameca SX-50) have been optimized for total-pb dating of monazite (e.g., Jercinovic and Williams 2005), a new trace-element microprobe (the Cameca Ultrachron) has been speciþcally designed for dating monazite grains that are relatively small, especially low in Th, and/or young. Dahl et al. (2005) analyzed 34 grains of monazite both isotopically (SHRIMP) and chemically (Cameca SX-50) in six polymetamorphosed Paleoproterozoic rocks from the Black Hills, South Dakota (Fig. 1), and then compared these data at various scales of observation to provide a cross check of the total Pb method. They further documented at least four discrete age generations of monazite growth in the Black Hills between ~ Ma, which they correlated with four deformational fabrics (S 1 S 4 ; Fig. 1) distinguished by Redden et al. (1990). However, these correlations (Fig. 2f) were advanced without beneþt of direct microtextural linkage between monazite ages and deformational fabrics. One of the rocks, metapelite PR-1 (Fig. 1), is especially suited for studying agefabric linkage because it preserves the S 1, S 2, and S 3 fabrics in unambiguous microstructural interrelationships. Monazite in this rock is abundant, complexly zoned, and low in Th (Dahl et al. 2005); moreover, one grain among ~100 examined in three thin sections also contains sigmoidal inclusion trails indicative of a relict fabric. The primary objective of this paper was to link Ultrachron dates obtained for this single, unique monazite grain to key structural elements revealed in detailed petrofabric study of the host rock. Integrating the monazite dates and fabric orientations with independent knowledge of major-mineral petrogenesis further permitted a secondary objective to be addressed, which was to FIGURE 1. Geologic map of the Black Hills crystalline core, South Dakota, showing distribution of Neoarchean basement granitoids and Paleoproterozoic cover rocks, and orientations of deformational fabrics (S 1 S 4 ). Metapelite PR-1 (locality highlighted) is the main focus of study (this paper); sample locations for the companion study (Dahl et al. 2005) are also shown. BFG = Biotite-feldspar gneiss; BMG = granite at Bear Mountain; BMD and HPD = Bear Mountain and Harney Peak domes; LEG = Little Elk granite; GJF = Grand Junction fault. Lower left inset: regional distribution of metamorphic zones and fabric orientations. BT = biotite; GRT = garnet; ST = staurolite; KY = kyanite; SIL = sillimanite; KFS = K feldspar. Lower right inset: generalized lithotectonic map showing location of the Black Hills relative to Archean cratons and Paleoproterozoic orogens. WC and SC = Wyoming and Superior cratons; DB = Dakota block (Baird et al. 1996); BHO, CPO, and THO = Black Hills (Dahl and Frei 1998), Central Plains (Sims and Peterman 1986), and Trans-Hudson (Sims et al. 1991) orogens; CB and HR = Cheyenne and Hartville-Rawhide fold-thrust belts (teeth on upper plates; Sims 1995); MT = Montana; ND and SD = North and South Dakota; NE = Nebraska; WY = Wyoming. The CPO is considered equivalent to the eastern Yavapai (YAV) arc terrane, following Karlstrom et al. (2002). Bold arrows depict arccontinent (D 1 ) and continent-continent (D 2 ) collisions that produced S 1 and S 2 fabrics in the Black Hills. Maps modiþed after DeWitt et al. (1989), Redden et al. (1990), Klasner and King (1990), Sims (1995), Dahl et al. (1999), and Stock (2004).

3 1714 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE FIGURE 2. Normalized total-pb age-probability plots of 54 monazite domains based on 354 electron microprobe analyses (EMP, Cameca SX-50) in six Precambrian rocks from the Black Hills, whose locations are shown in Figure 1. Total-Pb monazite dates and preliminary tectonic interpretations are from Dahl et al. (2005). (a-d) Metapelites PR-1, ST-112, SC-9a, and EG-56. Mixed isotopic dates of monazite in garnet determined by Pb stepwise leaching (PbSL) of garnet are shown for comparison. (e) Granitic gneisses LEG-8 and BM-20. (f) Summary age-probability plot based on all EMP data combined, with 54 monazite dates provisionally assigned to four discrete events in the Black Hills D 1 -D 4. Superimposed for comparison (in gray-scale) is a probability plot of 45 independent 207 Pb/ 206 Pb dates of monazite, zircon, and allanite in metapelites (12 samples) and one sample each of iron formation and Harney Peak granite. Of these, 22 were produced by ion microprobe (IMP; Dahl et al and in prep.), 21 by Pb stepwise leaching of garnet (PbSL; Schaller et al. 1997, Dahl and Frei 1998, Dahl et al. 1998, 2005), and two by thermal ionization mass spectrometry (TIMS; Redden et al. 1990). The total-pb and isotopic dates broadly agree, although mixing of the PbSL dates has obscured the overall isotopic distribution. THO = Trans-Hudson orogeny (initial Wyoming-Superior convergence); YAV = docking of Yavapai arc terrane and Wyoming craton (D 1 ); BHO = Black Hills orogeny (Wyoming-Superior collision, D 2 ); HPM = Harney Peak magmatic event, with local doming (D 3 ); and MAZ = Mazatzal terrane accretion from the south at Ma (Magnani et al. 2004), which may have affected Black Hills (D 4?).

4 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE 1715 constrain the histories of polyphase thermotectonism of the Black Hills and of ~ Ma terrane assembly in southern Laurentia. By permitting these objectives to be addressed, metapelite PR-1 is regarded as a geological Rosetta stone. PALEOPROTEROZOIC SETTING AND PREVIOUS GEOCHRONOLOGY The Laramide Black Hills uplift, located on the eastern edge of the Archean Wyoming craton, exposes a crystalline core of Neoarchean basement granitic gneisses overlain by two Paleoproterozoic successions dominated by rift-related, metasupracrustal rocks of the southern Trans-Hudson orogen (THO) (Redden et al. 1990; designated as Black Hills orogen in Fig. 1). The older succession (~ Ma; Dahl et al. 2003) is conþned to the Nemo area (Fig. 1), whereas the younger and areally predominant succession (mostly ~ Ma; Bekker et al. 2003; Dahl et al. 2005) hosts the sample studied here. The crystalline core was intruded by the Harney Peak leucogranite (HPG), an early sill of which is dated at 1715 ± 3 Ma (U-Pb, monazite; Redden et al. 1990). East of the Black Hills uplift, Precambrian rocks are buried under Phanerozoic cover but are known from drill data. Here, the southern THO is dominated by a N-S-trending belt of ~ Ma arc-related rocks (Sims et al. 1991). Farther to the east, the Dakota block (North Dakota; Baird et al. 1996) is a buried Archean terrane sandwiched between the Wyoming and Superior cratons, which may extend into present-day South Dakota (Klasner and King 1990; Nabelek et al. 2001; Stock 2004). These relationships are shown in Figure 1 lower right inset. The Black Hills crystalline core preserves a complex record of Paleoproterozoic thermotectonism, as depicted in Figure 1. The earliest structural elements consist of ENE-trending fold nappes/thrusts (F 1 ) that were refolded into NNW-trending upright folds (F 2 ) with a steep, locally penetrative, axial planar foliation, S 2 (Redden et al. 1990). The nappe/thrusts originated during an episode of NW-vergent, thin-skinned thrusting (D 1 ) that deformed bedding (S 0 ). However, neither axial-planar foliation nor metamorphic assemblages related to this event have been recognized in the Black Hills, suggesting that present exposures largely represent the upper nappe regime (Redden et al. 1990). The one exception is represented by structurally deep exposures restricted to the Bear Mountain gneiss dome (Fig. 1), where metapelite PR-1 preserves a relict, ENE-trending, and S-dipping fabric (S 1 ) in garnet porphyroblasts as documented in this study. This relict fabric probably represents medium-p metamorphism in the lower nappe regime. The D 1 event and the related S 1 fabric are attributed to NW-directed accretion of Central Plains (or Yavapai) island-arcs (Fig. 1, lower right inset), which occurred as early as ~1785 Ma (Dahl et al. 1999, 2005). Younger F 2 folds, axial planar S 2 foliation (Fig. 1), and lower-p regional metamorphic assemblages originated during a major episode of thermotectonism (D 2 ) associated with ~E-W crustal shortening (DeWitt et al. 1986, 1989; Redden et al. 1990; Terry and Friberg 1990). This major event is widely attributed to collision of the Wyoming craton either with the Superior craton (Dahl and Frei 1998; Dahl et al. 1999) or with the intervening Dakota block (Nabelek et al. 2001; Chamberlain et al. 2002). The onset of D 2 collisional convergence in the Black Hills portion of the THO (designated as Black Hills orogen, BHO, in Fig. 1) is thought to have occurred at Ma (Dahl et al. 2005). Latesyn-collisional unrooþng of the Black Hills (Holm et al. 1997; Holm 1999; Nabelek et al. 2001) culminated in localized granite magmatism (HPG, Fig. 1), high-t low-p metamorphism (Helms and Labotka 1991), and associated doming (D 3 ) of the country rocks beginning at ~1715 Ma (Ratté and Zartman 1970; Ratté 1986; Redden et al. 1990; Duke et al. 1990; Dahl et al. 1998). The resultant concentric fabric (S 3 ) adjacent to the main HPG pluton is shown in Figure 1. Post-D 3 events include development of a weak NE-trending fabric (S 4 ) throughout the Black Hills, during a deformational event (D 4 ) of uncertain age and origin (Redden et al. 1990), and a later period of renewed regional unrooþng beginning at ~1500 Ma (Holm et al. 1997). SAMPLES AND METHODS Metapelite PR-1 was collected within the Bear Mountain gneiss dome from an outcrop of kyanite-zone cover rocks lying ~1 km east of the contact with the Archean granite core (Fig. 1). Deposited within the ~ Ma sedimentary succession (Vandelehr formation, Redden et al. 1990), this rock subsequently experienced multiple episodes of thermotectonism spanning a ~ Ma timeframe (Dahl and Frei 1998; Dahl et al. 1999, 2005). Lithologically, the rock is a medium-grained schist dominated by quartz, plagioclase, muscovite, biotite, and sillimanite, with local ~ cm thick horizons of abundant inclusionrich porphyroblasts of garnet and staurolite as well as biotite that is randomly oriented and relatively coarse-grained. Metapelite PR-1 is from the same outcrop as, and is lithologically equivalent to, sample MT-7 of Terry and Friberg (1990) and Terry (1990). Three oriented thin sections were cut to include its garnet-staurolite-rich horizons, with views targeted on selected foliation surfaces and lineation directions. Detailed interpretive sketches were made of microtextural relationships among key phases in these sections, with special emphasis on a key grain of matrix monazite containing a relict sigmoidal inclusion trail. In situ analyses of this monazite grain ( grain 1 ) were made in polished thin section using the electron microprobe (EMP, Cameca Ultrachron and Cameca SX- 50) facilities at the University of Massachusetts. The SX-50 was used for analysis of Ca, P, Si, Th, U, Y, and selected REE of this grain, whereas the Ultrachron was used primarily to establish its total-pb date. Details of analytical methodology for the SX-50 (and carried over to the Ultrachron) are primarily as described by Williams and Jercinovic (2002) with several notable improvements as noted in Jercinovic and Williams (2005) and Dahl et al. (2005). With the SX-50, high-resolution X-ray images were collected of the full polished thin section (for Mg and Ce), of a garnet porphyroblast (for Ca, Fe, Mg, and Mn), and of monazite grain 1 (for Y, Th, U, and Pb or Ca). With the Ultrachron, accelerating potential was 15 kv with a focused electron beam, estimated to be 0.7 µm in diameter for trace-element analyses at 200 na. These conditions result in a general excitation volume of 0.73 µm 3 for spot analysis of monazite if the targeted region is modeled as a sphere. Repeated spot analyses within each such domain were used to infer a mean date, according to the formulation of Montel et al. (1996). Precision of the Ultrachron monazite dates is reported herein as standard error of the mean date obtained within a single compositional domain, whereas accuracy was monitored by analysis of a consistency standard before and after each analytical session. Williams et al. (2005) discussed uncertainty components associated with electron microprobe dating and noted that, due primarily to the effects of background modeling and estimation, the uncertainty associated with accuracy of the microprobe date could be on the order of twice the precision. All errors in EMP data are reported, depicted, and discussed in this paper at the 2σ level (95% conþdence limits). Detailed methods for the SHRIMP U-Th-Pb isotopic analysis of the monazite discussed below and in Figure 3 were described in Dahl et al. (2005). During the same analytical session, a reconnaissance effort was made to investigate the in situ isotopic compositions of two large porphyroblasts of garnet and staurolite, imaged in detail using a scanning electron microscope before and after analysis, and referenced against an unradiogenic plagioclase grain in the same sample. Mass resolution as determined on zircon measured concurrently in this SHRIMP session was ~5200. Secondary ions were generated from a roughly µm target spot by using an O primary beam whose strength averaged 13.8 na. Count times were maximized at 40 s for several mass stations including 204 Pb, 206 Pb, 207 Pb, 208 Pb,

5 1716 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE FIGURE 3. Microtextures in composite metapelite samples MT-7 and PR-1 (Fig. 1), with views showing cuts in three orientations. Foliations S 1, S 2, and S 3 are related to N-vergent nappe/thrusting event, upright folding associated with ~E-W continental collision, and localized doming related to subjacent intrusion of Harney Peak granite, respectively. Mineral abbreviations: Bt = biotite; Chl = chlorite; Fsp = feldspar; Gt = garnet; Ms = muscovite; Mnz = monazite; Qz = quartz; and St = staurolite. (a) Sketch view on S 3 foliation surface, the orientation of which is N31 W and 42 NE. Older foliation S 1 is preserved in garnet porphyroblasts, whereas wrapping foliation S 3b (earlier S 2?) is present only adjacent to staurolite porphyroblast. Approximate orientations of S 1 and S 2 are N84 E 62 SE and N62 W 75 SW, respectively. Lineation L 3b, deþned by the intersection of S 2 and S 3, plunges S62 E (118 ). (b) Sketch view on NE-SW plane down the plunge of L 3, along the direction S62 E (118 ). Garnet contains foliations S 1 /S 2 and S 2 /S 3a, whereas matrix exhibits foliations S 2 and S 3b. S 3b is axial planar to folded (i.e., ßattened) S 2, and L 3 is a true stretching lineation (S 2 -S 3 intersection). Bold arrows indicate shear couple induced by doming (D 3 ). (c) False-color Mg X-ray map of view on NNW-SSE plane approximating the S 2 foliation surface (which also contains lineation L 3, not labeled). Partially resorbed garnet contains foliation S 1 and predates matrix foliation S 3, whereas late staurolite postdates S 3. White circles depict locations of monazite crystals; box locates the only monazite (grain 1, Table 1) that preserves foliation S 1 among ~100 grains observed in three polished thin sections. (d) Back-scattered electron image of monazite grain 1 (BSE-bright), boxed in Figure 3c. Sigmoidal inclusion trails (BSE-dark, deþned mostly by muscovite and quartz but also by aluminosilicate) are deþned by foliations S 1 and S 2 ; S 3 is shown in matrix (BSE-dark). Y-enriched rim on texturally older core is related to resorption of adjacent garnet (Fig. 3c). 248 ThO, and 254 UO, as well as background (0.1 amu above 204 Pb). Data for each spot were acquired over 5 scans through the mass sequence, and a typical spot analysis time typically lasted min. Incremental heating, 40 Ar/ 39 Ar analyses were performed on biotite and muscovite separated from PR-1 using standard techniques. The measurements were performed at Ohio State University using procedures described previously (Foland et al. 1993), except that Ar measurements were made using a new mass spectrometer and associated extraction lines. Aliquots were irradiated in the Ford Nuclear Reactor (University of Michigan), and then step-heated in a high-vacuum, low-blank furnace to successively higher temperatures, for ~30 min at each. RESULTS Petrofabric analysis of metapelite PR-1 Exposures at the PR-1 locality (Fig. 1) exhibit a dominant matrix foliation (S 3 ) that trends NW-SE with a moderate NE dip, whereas oriented inclusion trails observed in the garnet porphyroblasts preserve a relict ENE-trending foliation (S 1 ). The megascopic observation of consistent orientations for this

6 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE 1717 FIGURE 4. X-ray chemical maps of Ca (a) and Fe (b) in PR-1 garnet, highlighting resorption textures. Grt 1 and Grt 2 represent D 2 and D 3 garnet, respectively. Diameter of porphytoblast is ~5 3 mm. earliest foliation (S 1 ) within the porphyroblasts implies that they did not rotate during growth or subsequent deformation (cf., Evins 2005). The three oriented thin sections (see Fig. 3) were cut: (1) on the S 3 foliation surface (strike = 149, dip = 42 NE); (2) along the stretching lineation (trend = 118, plunge = 25 ); and (3) on the NNW-SSE-trending S 2 foliation surface, parallel to lineation and perpendicular to S 3 foliation. The S 2 foliation is the most prominent structural element in the Precambrian core of the Black Hills, and is characterized by a NS to NNW strike and steep dip. At the PR-1 locality, however, it is strongly overprinted and transposed by the S 3 doming fabric, although is locally preserved in microlithons between the layers of S 3 foliation (Fig. 3b). An S 1 orientation of S, calculated from apparent dips and true thin-section orientations (Figs. 3a and 3c), agrees with an orientation of S measured in outcrop from the inclusion trails in the porphyroblasts. From these calculations, the apparent dip predicted for S 1 is 44º (Fig. 3b), which is a lower value than that inferred from inclusion trails in garnet (~60º). This discrepancy may result from the partial rotation of S 1 into S 2, which implies that this particular garnet may not be cut through its center. Hence, this fabric is assigned as S 1 /S 2 in Figure 3b. The ENE-trending S 1 fabric, preserved as obliquely oriented inclusion trails in garnet porphyroblasts (Fig. 3c), is also found in a unique grain of matrix monazite (grain 1), as shown in Figures 3c and 3d. In the garnet, included S 1 phases consist primarily of quartz, chlorite, and biotite (Terry and Friberg 1990), whereas in the monazite grain, muscovite, quartz, and minor aluminosilicate (kyanite?) deþne the S 1 inclusion assemblage (BSE-dark areas in Fig. 3d). The different fabric orientations and the Z-shape preserved in Figure 3b are consistent with left-lateral shearing with top-ne (in thin section) or west-side up (regionally), and the apparent rota- FIGURE 5. X-ray chemical maps of Th (a) and Y (b) in PR-1 monazite (grain 1; cf., Figs. 3c,d), showing locations of spots analyzed by EMP (Cameca Ultrachron). Ultrachron dates of core and rim domains are also indicated (mean ± twice the standard error).

7 1718 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE tion of S 1 and possibly S 2 in garnet is referred to as S 2/3a. S 2 is also evident toward the ends of S 1 -S 2 sigmoidal inclusion trails hosted by the core domain of monazite grain 1 (Fig. 3d), where it forms shallow intersections with the plane of the polished section. This apparent transposition of S 1 into S 2 is preserved only rarely in monazite (Fig. 3d) but not at all in the matrix, where both fabrics are largely overprinted by S 3 (Fig. 3c). The S 3 matrix foliation is deþned principally by muscovite, biotite, and plagioclase. This assemblage is overgrown by randomly oriented (i.e., post-d 3 ) porphyroblasts of prograde staurolite (Fig. 3c). Foliations S 2 and S 3 form an intersection lineation that is parallel to the stretching lineation seen on S 3 foliation surfaces (L 3 ; see Figs. 3a and 3b). The stretching lineation is also deþned by quartz in strain shadows around garnet, which are visible on S 3 foliation surfaces (but not shown in Fig. 3). These strain shadows are locally asymmetric and, consistent with the Z-shape deþned by the S 2 -S 3 interference patterns (Fig. 3b), also indicate left-lateral shearing that is synchronous with S 3. Although not shown here, kinematic indicators in sections parallel to Figures 3c and 3d are top-nw or left-lateral in the Þeld. The presence of these indicators in thin section and their consistent sense of shear are compatible with a triaxial strain regime that accords well with Þeld indications of sinistral oblique convergence (transpression of Caddey et al. 1991) and simultaneous dome formation (Kuhl 1982; Gosselin et al. 1988). Two growth generations of garnet, staurolite, and sillimanite (Terry and Friberg 1990; Terry 1990), and at least four of monazite (Dahl et al. 2005), have been recognized in PR-1. Initial sillimanite formed during decompression from the kyanite to sillimanite stability Þelds, and prograde sillimanite formed later as sub-oriented needles after muscovite. The early and late staurolite generations are syn-s 2 (Terry and Friberg 1990) and post-s 3 (Fig. 3c), respectively. The two generations of garnet growth are manifested as coarse-grained porphyroblasts (see Fig. 3c), which exhibit early, graphite-rich interiors (dusty-looking patches) that are overgrown and intergrown by late, graphite-free material (false-colored reddish-brown). X-ray compositional maps (Fig. 4, see previous page) further reveal that the (almandine) garnet interior is Caenriched (X Ca = ) and Fe-depleted relative to the later overgrowths and intergrowths. Also, the sharp boundary between these compositional zones (Fig. 4a) and occurrence of this garnet in a GASP assemblage suggest that garnet growth was episodic, occurring Þrst at medium-high P (Ca-enriched zones) and later at lower P (Ca-depleted zones). The older D 2 garnet population postdates the included S 1 fabric (Fig. 3c) but predates the matrix S 3 fabric, whereas the younger D 3 garnet is probably associated with S 3 (Terry and Friberg 1990; Friberg et al. 1996). Two episodes of garnet breakdown are also evident, judging from the resorption rims shown in Figure 4. The Þrst resorption event is represented by the irregular inner boundary, preserved between the Ca-enriched and Ca-depleted garnet (Fig. 4a), whereas the second is represented by deep embayments observed at the outer garnet boundary (Figs. 4a and 4b). Moreover, the outer resorption rim truncates the inner rim, as best shown along the bottom garnet edge (Fig. 4a), thereby establishing their relative ages. Other garnet crystals in this rock exhibit atoll textures, in which garnet cores are replaced by randomly oriented micas and quartz (Terry 1990). The preservation of these resorption textures indicates that discrete episodes of garnet disequilibrium punctuated the thermotectonism. Age-composition data for PR-1 monazite (grain 1) Two generations of growth are also evident in monazite grain 1, as represented by a core domain that contains the inclusion trails and a younger, inclusion-free rim domain that truncates the sigmoid (Fig. 3d). The Y and Th abundances of these domains are highlighted in X-ray images (Figs. 5a and 5b, see previous page) and depicted in a scatter-plot (Fig. 6a), showing that the core domain is depleted in Y relative to the rim domain and mostly depleted in Th as well. In addition, the core itself exhibits modest bimodal variation in Y abundance that correlates inversely with Th abundance (Fig. 6a). Considered alone, the pattern of Th distribution suggests that the core was both overgrown and intergrown by the Th-enriched material that dominates in the rim (yellow pattern in Fig. 5b). However, this material differs markedly in Y content from core to rim (dark-blue and yellow patterns, respectively, in Fig. 5a), eliminating the possibility of intergrowth. Instead, the core consists of two domains of differing Y and Th abundances (Fig. 5), both of which are texturally older than the rim domain. Elemental abundances in the core and rim domains of monazite grain 1, determined by microprobe analysis, are presented in Tables 1 and 2, and spot locations are shown in Figure 5. As shown in Figure 6a, the 39 core domain analyses yield a date of 1750 ± 10 Ma, with no apparent age differences related to variable Th and Y abundance; 17 analyses of the Y-enriched rim yield a younger date of 1692 ± 5 Ma (uncertainties reported as twice the standard error of the mean). Figure 6a (inset) is a relative probability plot of monazite dates for the core and rim domains (95% conþdence), which assumes Gaussian distributions. As indicated in Table 2 (SX-50 data), the young rim of monazite grain 1: (1) is relatively enriched in Si, Ca, P, Th, U, Th/U, Y, and Gd; (2) has similar Pr and Sm abundances; and (3) is depleted in La, Ce, and Nd compared to the older core. It appears that the heavier the REE, the more it is incorporated preferentially into the monazite rim. The core-to-rim Th and Y abundances of grain 1 are compared to those of other PR-1 monazite grains in Figure 6b (SX-50 data of Dahl et al. 2005). The negative slopes and general subparallelism of tie lines in Figure 6b indicate that Th-depleted rims of grains typically mimic their relatively Th-enriched cores, with the core of grain 1 exhibiting the lowest Th abundance within the entire monazite data set. All grains show similar spreads in Y abundance, however, with Y-enriched rims having formed on older, relatively Y-depleted cores at either ~ or ~ Ma. Hence, the diachronous growth of these monazite rims may reßect the two garnet resorption events described above (see Fig. 4). U-Th-Pb isotopic data for PR-1 garnet, staurolite, and plagioclase SHRIMP analyses of the garnet and staurolite porphyroblasts in Figure 3c (Table 3) reveal that these phases are essentially nonradiogenic inasmuch as: (1) 208 Pb/ 206 Pb and 207 Pb/ 206 Pb ratios of 2.11 ± 0.09 and 0.85 ± 0.03, respectively, equate to the common Pb signature found in coexisting plagioclase; and (2)

8 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE 1719 concentrations of both U and Th are <1 ppb, which, assuming the 1760 Ma age, equate to only ~0.2 ppb of radiogenic Pb. Massbalance calculations based upon these results and relative modal abundances indicate furthermore that, in the Pb stepwise leaching (PbSL) experiments of Dahl and Frei (1998), only ~ % of the total radiogenic Pb was actually leached from the garnet and staurolite hosts per se, whereas ~ % was leached from the included monazite. These garnet results are consistent with those reported by DeWolf et al. (1996). 40 Ar/ 39 Ar dates of PR-1 muscovite and biotite Argon incremental-release data for PR-1 micas are illustrated in Figure 7, and full data are available from the repository. Both spectra show some discordance. For muscovite (Fig. 7a), 13 increments yield a plateau 40 Ar/ 39 Ar date of 1608 ± 8 Ma (2σ) for the middle ~50% of the Ar. Indistinguishable values for the total-gas (1606 Ma) and plateau dates suggest that the discor- TABLE 1. Ultrachron microprobe analyses of PR-1 monazite grain 1 Domain Y Th Pb U Th/U Th/Y Y/U Chemical &Spot (ppm) (ppm) (ppm) (ppm) date (Ma) Core Domain Average SStd *StdErr FIGURE 6. Apparent age-th-y relationships in monazite from metapelite PR-1. (a) EMP (Ultrachron) age-th-y plot for grain 1, highlighting differences between core and rim domains (cf., Fig. 5). Inset shows age-probability plot for grain 1 and results of an age-calibration cross check using monazite age standard GSC8153. (b) EMP (SX-50) Th-Y plot for all analyzed grains, showing sub-parallelism of core-rim tielines in Th-Y space. Rim Domain Average SStd *StdErr Notes: Spots are listed in descending order of apparent age. Locations of spots (n = 56) are shown in Figure 5. Abbreviations: SStd = standard deviation; StdErr = standard error of mean (ages reported at 95% conf.).

9 1720 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE dance may reßect recoil redistribution of 39 Ar. The biotite (Fig. 7b) yields a total-gas date of 1414 Ma, but the spectrum shows great discordance at low temperature. Fractions between ~40 and 90% Ar release show only minor discordance and deþne a pseudoplateau date of ~1480 Ma. These ~ Ma 40 Ar/ 39 Ar dates of mica are substantially younger than the ~ Ma total-pb dates (henceforth interpreted as growth ages) obtained for monazite growth in the same rock and, as such, may be interpreted as cooling ages that were set following the Harney Peak magmatic event. low-th monazite (mean Th concentration = ~5700 ppm). The mean Ultrachron age obtained for the grain 1 core is also similar to the mean SHRIMP and PbSL dates obtained for the cores of other monazite grains in metapelite PR-1. Figure 8 is a U-Pb concordia plot of 18 SHRIMP analyses of apparent DISCUSSION AND CONCLUSIONS Accuracy and precision of PR-1 Ultrachron dates The total-pb age, 1750 ± 10 Ma, obtained for the grain 1 core (Ultrachron; Th/U = ) agrees with the weighted-mean 207 Pb/ 206 Pb isotopic age, 1747 ± 14 Ma, inferred by Dahl et al. (2005) for this same domain (SHRIMP; Th/U = ; Fig. 8, inset). The SHRIMP data set represents a single age population (MSWD = 0.49; criteria of Wendt and Carl 1991), which is probably also true for the Ultrachron data. This comparison demonstrates that the Ultrachron is capable of yielding highly accurate and precise ages, rivaling the SHRIMP, even for relatively low-th monazite. In contrast, the total-pb age of 1777 ± 17 Ma previously reported for this domain (SX-50; Dahl et al. 2005) is barely within 2σ error of the Ultrachron and SHRIMP ages, suggesting that these data did not pinpoint the age of this TABLE 2. Analytical summary of PR-1 monazite grain 1 Parameter Core domain n Rim domain n Data from SX-50 Si (ppm) 610 ± ± 16 2 P ± ± Ca 1825 ± ± 25 2 Th 5594 ± ± La ± ± 59 2 Ce ± ± Pr ± ± Nd ± ± Sm ± ± Gd ± ± Y 4871 ± ± Age (Ma) * 1777 ± ± Data from Ultrachron Th (ppm) 5737 ± ± U 2250 ± ± Pb 1131 ± ± Y 4050 ± ± Th/U 2.6 ± ± Age (Ma) 1750 ± ± 5 17 Data from SHRIMP Th/U * 2.5 ± ± Age (Ma) * 1747 ± ± 18 2 Notes: Uncertainties = 95% conf.; n = number of spot analyses. * Data from Dahl et al. (2005). All other data are from this study. FIGURE Ar/ 39 Ar incremental-release spectra for muscovite (a) and biotite (b), based on separates obtained from metapelite PR-1. K/Cl, K/Ca, and apparent age are shown for each thermal increment. Muscovite exhibits a plateau date of 1608 ± 4 Ma, whereas biotite exhibits a nearplateau date of ~1480 Ma, both interpreted as cooling ages (see text). Supporting data are available as a repository item. TABLE 3. Isotopic and chemical data for garnet, staurolite, and plagioclase in metapelite PR-1 (GSC SHRIMP II) (U) (Th) Th/U 206Pb/ ±1σ 207Pb/ ±1σ 208Pb/ ±1σ 207Pb/ ±1σ 208Pb/ ±1σ Phase ppb ppb 204Pb 204Pb 204Pb 206Pb 206Pb Garnet Staurolite 0.7 nd na Plagioclase Notes: Accuracy of U and Th abundances estimated at ±20% relative. Abbreviations: nd = not detected; na = not analyzed.

10 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE Pb/ 238 U Metapelite PR-1 monazite (n = 18) Intercept age = 1752 ± 6 Ma MSWD = 0.99 (95% conf.) data-point error ellipses are 68.3% conf Metapelite PR-1 cores of grains 1, 9, 15, 16, 19, 27, & 34 monazite grain 1 box heights are 95% conf. (cores of grains 12 & 14 excluded) Pb/ 235 U FIGURE 8. U-Pb concordia plot representing cores of seven monazite grains analyzed in situ by ion microprobe (SHRIMP) in metapelite PR- 1 (data from Dahl et al. 2005). Upper-intercept 207 Pb/ 206 Pb age is 1752 ± 6 Ma. Inset shows weighted-mean age plot of six SHRIMP analyses of the core domain of monazite grain 1. Mean 207 Pb/ 206 Pb age of this domain is 1747 ± 14 Ma, which agrees with mean total-pb age of 1750 ± 10 Ma (Fig. 6a). Only isotopically concordant monazite analyses (i.e., 95%) are included in the plots, which were made using Isoplot/Ex (ver. 3; Ludwig 2004). Error ellipses and error bars represent 1σ and 2σ uncertainties, respectively. ~ Ma ages representing grains 1, 9, 15, 16, 19, 27, and 34 (cf., Fig. 4b of Dahl et al. 2005). This data set yields a 207 Pb/ 206 Pb upper-intercept age of 1752 ± 6 Ma (2σ; MSWD = 0.99) and represents a single age population (criteria of Wendt and Carl 1991). Only (relatively discordant) grains 12 and 14 from the SHRIMP data set were excluded in Figure 8 because their cores appear to represent part of an older age population or populations ( 1785 Ma; see Figs. 3a, 6a, and 6b of Dahl et al. 2005). Thus, the mean PbSL date of 1760 ± 7 Ma obtained by Dahl and Frei (1998) for low-th monazite inclusions in PR- 1 garnet and staurolite porphyroblasts represents an analytical mix of the subordinate ~1785 Ma age and predominant 1750 Ma populations, as illustrated in Figure 2a (likewise for metapelites ST-112 and SC-9a; Figs. 2b and 2c). The Y-enriched rim of PR-1 monazite grain 1 (Fig. 5a) yields younger total-pb ages of 1680 ± 16 Ma (SX-50; Dahl et al. 2005) and 1692 ± 5 Ma (Ultrachron; Tables 1 and 2). These ages agree within analytical (2σ) error, although the Ultrachron age is apparently more precise by a factor of three, for a similar number of spot analyses. In contrast, the most precise 207 Pb/ 206 Pb ages of this grain rim (1723 ± 18 Ma; n = 1) and of all monazite rims (1714 ± 13 Ma; n = 6), as determined by SHRIMP (see Figs. 4c and 7a of Dahl et al. 2005), appear to be slightly older than the grain 1 total-pb age. Perhaps the SHRIMP spots incorporated a minor component of older core material; otherwise, there is no obvious explanation for this age discrepancy. Given its relatively low Th abundance (see Fig. 6b) and the fact that the Ultrachron is designed to optimize trace-element analysis, the actual rim age of monazite grain 1 is interpreted to be 1692 ± 5 Ma, which is intermediate within the previously inferred ~ Ma range. 207 Pb/ 206 Pb age (Ma) 1820 Mean age = 1747 ± 14 Ma MSWD = 0.49 (95% conf.) Age-fabric-composition relationships in PR-1 monazite and associated phases Establishing a 1750 ± 10 Ma age for the low-th,y core domain of grain 1 is critical to constraining the thermotectonic history, because this single grain encloses the regional S 1 and S 2 fabrics. Because both the ENE-trending S 1 and NNW-trending S 2 fabrics in the BMD (Fig. 1) are 1750 Ma or older, the ~NWdirected thrusting (D 1 ) and onset of ~E-W convergence (D 2 ) that respectively produced these fabrics are bracketed between ~1790 Ma (oldest CPO rocks, Fig. 1; Premo and van Schmus 1989) and 1750 Ma. The best estimates for the absolute age of D 1 (S 1 ) are 1774 ± 10 and 1776 ± 10 Ma, based on precise EMP (SX-50) and concordant (for Pb/U) SHRIMP ages of high-th (8 10 wt%; Fig. 6b) monazite grains 9 and 34, respectively (2σ; see Tables 2 and 3 of Dahl et al. 2005). In contrast, the younger D 2 event was probably ongoing either as a single pulse or as multiple pulses for at least ~35 Myr from 1750 Ma to 1715 Ma, when it was terminated by post-tectonic unrooþng and granite emplacement. Thus, growth of the (~1775 and ~1750 Ma) monazite populations is believed to have accompanied the development of S 1 and S 2 fabrics independently constrained to this timeframe (Dahl and Frei 1998; Dahl et al. 1999). Any further support for these age-fabric assignments is obscured by the fact that elongate monazite grains (9 and 34, etc.) were reoriented during D 3 doming to conform to the S 3 fabric. The higher-th,y rim domain of grain 1 truncates the S 1 -S 2 sigmoid, so it is younger microtexturally, consistent with its younger Ultrachron age of 1692 ± 5 Ma. Taken together, these relationships suggest a 1692 ± 5 Ma age for the overprinting S 3 fabric (cf., Figs. 3c and 6a), for which SHRIMP data for all Y-enriched monazite rims combined suggest a 1714 ± 13 Ma age (upper-concordia intercept; see Fig. 4c of Dahl et al. 2005). Collectively, these age-fabric data support a ~ Ma timeframe for localized doming (D 3 ) related to late-syn- to postcollisional magmatism (Harney Peak granite), the onset of which is independently dated at 1715 ± 3 Ma (Redden et al. 1990). Growth ages of the garnet and staurolite porphyroblasts also can be inferred from the monazite-derived age-fabric relationships. For example, the graphite- and Ca-enriched garnet interiors enclosing the S 1 fabric (cf., Figs. 3c and 4a) probably formed at 1750 ± 10 Ma, because they are microtexturally equivalent to the dated core of monazite grain 1 (Figs. 3c, 3d, and 6a). Likewise, the graphite- and Ca-depleted garnet rims/overgrowths (cf., Figs. 3c and 4a) probably began forming at ~1715 Ma, with the onset of D 3 doming and Harney Peak-related reheating of the country rocks (Friberg et al. 1996). Randomly oriented staurolite fully encloses the S 3 fabric (Fig. 3c), texturally postdates the D 3 garnet, and thus formed at or just after 1715 Ma. Origin of the Y-enriched monazite rims on older, Y-depleted cores Monazite grains in PR-1 typically exhibit Y-enriched rims (see Fig. 6b), with ~ Ma ages potentially produced by reactions in which garnet (Gibson et al. 2004), allanite (Wing et al. 2003), xenotime (Pyle and Spear 1999), and/or apatite (Yang and Pattison 2004) was consumed. Neither allanite nor xenotime have been recognized, whereas garnet exhibits prominent resorption textures (Fig. 4) indicating consumption of this major phase.

11 1722 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE Garnet is a well-known Y sink (Pyle and Spear 1999; Gibson et al. 2004), breakdown of which may release substantial Y for incorporation into monazite rims/overgrowths forming on older cores (e.g., see Finger and Helmy 1998; Kohn and Malloy 2004). Moreover, the garnetiferous horizon in metapelite PR-1 not only contains the highest modal abundance of garnet among the six Black Hills rocks analyzed but also exhibits the highest coreto-rim spreads in Y abundance of monazite grains (Fig. 6b; cf., Table 2 of Dahl et al. 2005). Therefore, it is proposed that the Y-enriched signatures of monazite rims mirror the history of garnet breakdown in this rock. In particular, PR-1 garnet porphyroblasts indicate two episodes of consumption, as represented by the inner and outer resorption rims documented in Figure 4a. Correspondingly, matrix monazite grains exhibit two generations of Y-enriched rims (Fig. 6b) that gave total-pb ages of 1692 ± 5 Ma (Fig. 6a) and 1753 ± 4 Ma (Dahl et al. 2005). Thus, it is considered that coupled garnet resorption and monazite overgrowth on older monazite cores occurred episodically at these times. As a corollary, the older, low-y cores of matrix monazite (grains 1, 9, 12, 14, 27, and 34; Fig. 6b) probably formed during and after initial growth of garnet, which sequestered the bulk of available Y in the garnetiferous horizons (e.g., see Zhu and OʼNions 1999b; Foster and Parrish 2003). Unlike Y, the Th and U contained in PR-1 monazite rims cannot have been derived from breakdown of garnet, which is devoid of both Th and U (<1 ppb) and contains no radiogenic Pb, as shown in Table 3. The array of subparallel Th-Y tie lines (Fig. 6b) suggests that Th in the monazite rims was relatively immobile and derived locally, possibly from partial recrystallization of adjacent monazite cores. Moreover, the negative slopes exhibited by most tie lines in Figure 6b (grain 1 is the exception) suggest that Th abundances in the monazite rims were diluted by the inßux of other elements (including Y). If, instead, another phase (such as allanite) were supplying Th to the monazite rims, without a contribution from the associated monazite cores, a uniform Th abundance of rims unrelated to core abundances would be expected rather than the subparallelism of core-rim tie lines actually observed. Pressure-temperature-deformation-time (PTDt) path for the Bear Mountain dome (BMD) The linkage of microstructure, petrology, and microchronometry permits reconstruction of a pressure-temperature-deformation time (PTDt) path for metapelite PR-1 and the Bear Mountain domal area. This path (Fig. 9) is modiþed from published sources (Terry and Friberg 1990; DeYoreo et al. 1991; Williams and Karlstrom 1996; Holm et al. 1997; Dahl and Frei 1998) and superimposed on a phase diagram for the KFMASH system (Spear et al. 1999). Two main clockwise paths are depicted for the interval between ~1780 and ~1480 Ma. The Þrst is a looping path (cf., Williams and Karlstrom 1996) that delineates a period of conductive and/or shear heating of metasediments (e.g., England and Thompson 1986; Nabelek et al. 2001) associated with tectonic burial to the mid-lower crust and regional, medium-p metamorphism (Terry and Friberg 1990). Based on monazite age-fabric relationships, these events occurred at ~ Ma. According to Holm et al. (1997), these events were followed by an episode of regional uplift/unrooþng and resultant mid-crustal cooling, shown in Figure 9 as having occurred some time between ~ Ma (as discussed below). The second clockwise path represents a transient reheating event (e.g., see Lux et al. 1986; DeYoreo et al. 1991; Sandiford et al. 1991) caused by ascent of Harney Peak granitic magmas into the midcrust (Redden et al. 1990; Holm et al. 1997) at ~ Ma. This event caused HTLP regional-contact metamorphism and localized D 3 doming of the country rocks, and was followed by a ~ Ma interval of slow, near-isobaric cooling of the mid-crust through ~ C. The following discussion links the age-composition-microtexture relationships established above for PR-1 monazite (and related phases, Figs. 3 7; see also Dahl et al. 2005) and the corresponding PTDt path proposed for the BMD (Fig. 9). Indicated breaks in the path reßect lack of information regarding details of P-T evolution between the designated thermotectonic events (D 1, D 2, and D 3 ). The timing and temperature for the onset of ~NW-directed D 1 nappe/thrusts in the Black Hills that produced the S 1 fabric at the PR-1 locality are depicted in Figure 9 at 1775 ± 10 Ma and ~400 C. The basis for the 1775 Ma estimate was discussed above (i.e., oldest consistent monazite core ages). A more conservative FIGURE 9. Pressure-temperature-deformation-time (PTDt) path for the Bear Mountain dome (BMD, metapelite PR-1), inferred from phase petrology, microfabric analysis, and monazite geochronometry (this study; modiþed after Terry and Friberg 1990; Holm et al. 1997; Dahl and Frei 1998). Thermotectonic events and their inferred (monazite) ages are: D 1 (docking of CPO, at ~1775 Ma); D 2 (Black Hills orogeny, at ~ Ma); and D 3 (regional unrooþng, localized doming, and granitic magmatism, at ~ Ma). Growth ages of Y-rich monazite rims on older Y-depleted cores are schematically related to reactions that may have occurred at 1753, 1714, and 1692 Ma. Post-D 3 40 Ar/ 39 Ar cooling ages of hornblende (~1690 Ma; Dahl et al. 1999) and micas (~ Ma; this study) are also shown. Mineral abbreviations: And = andalusite; Ky = kyanite; Sil = sillimanite; Bt = biotite; Chl = chlorite; Cld = chloritoid; Grt = garnet; Kfs = K-feldspar; Ms = muscovite; Prl = pyrophyllite; Qz = quartz; St = staurolite; V = vapor (H 2 O). Petrogenetic grid is modiþed after Spear et al. (1999), and dashed isopleths for the reaction of Bt-Qz to form Grt were calculated using TWEEQU (ver. 2.02; Berman 1991). See text for discussion.

12 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE 1723 interval of Ma for D 1 is indicated from the Ultrachron age of the monazite grain enclosing the S 1 fabric (Figs. 3d and 6) and the oldest known rocks in the accreted Central Plains terrane (Premo and van Schmus 1989; Dahl et al. 1999). The basis for the ~400 C temperature estimate is that the S 1 inclusion trails in garnet (Fig. 3c) contain biotite and kyanite (Terry and Friberg 1990), neither of which can form below ~400 C (Spear et al. 1999). The onset of D 1 thermotectonism is thus represented in Figure 9 by the kyanite-in reaction: pyrophyllite = kyanite + quartz + H 2 O (1) (Spear et al. 1999) at 1750 ± 10 Ma (~1775 Ma?) and ~400 C. Initially, the ~ Ma kyanite isograd may have been subhorizontal (Dahl and Frei 1998) but was subsequently deformed during D 3 doming (Dahl et al. 1999) to account for its present-day, oval map pattern (Fig. 1, left inset). The regional ~E-W shortening and tectonic burial (D 2 ) that produced F 2 folds and S 2 axial-planar foliation in the Black Hills was accompanied by garnet growth. Because this garnet overgrew the S 1 fabric in metapelite PR-1 (Fig. 3c) and contains kyanite inclusions, it is proposed that the model garnet-in reaction: chloritoid + biotite = garnet + chlorite + H 2 O (2) (Spear et al. 1999) was crossed, and at higher temperature than the kyanite-in reaction. Published thermobarometry for the BMD and environs indicates that growth of garnet cores (e.g., D 2 garnet Grt 1 in Figs. 4a and 9) began at kyanite-zone conditions of ~5 kbar and ~450 C and continued to a maximum pressure of ~7 kbar (Terry and Friberg 1990). These P-T relationships, shown in Figure 9, indicate that metapelite PR-1 underwent tectonic burial to a depth of ~25 km. Microtextural equivalence of the garnet cores and dated grain 1, both of which enclose the S 1 fabric (cf., Figs. 3c and 3d), imply an age of 1750 ± 10 Ma (Fig. 6a) for the onset of D 2 garnet growth. This garnet contains monazite inclusions of relatively low Th/U ratio (~ ), whose mean age of 1760 ± 7 Ma (Dahl and Frei 1998) represents a mix of ~1775 and ~1750 Ma age populations that also typify the rock matrix (Fig. 2a; see also Dahl et al. 2005). Subsequently, the Þrst of two recognized generations of prograde staurolite (Terry and Friberg 1990) may have formed by discrete episodes of prograde consumption of D 2 garnet (contributing Y to monazite) according to the model staurolite-in reaction: garnet + chlorite + muscovite = staurolite + biotite + H 2 O (3) (Spear et al. 1999). The D 2 garnet cores (Grt 1 ) preserve clear microtextural evidence of initial breakdown in the form of highly irregular boundaries between the Ca-enriched (i.e., medium-high-p) cores and Ca-depleted (i.e., lower-p) rims. Interrelationship among garnet breakdown, development of Y-enriched monazite rims/overgrowths, and staurolite growth has been suggested for metapelites elsewhere (e.g., Finger and Helmy 1998; Zhu and OʼNions 1999b; Spear and Pyle 2002; Gibson et al. 2004; Kohn and Malloy 2004). In PR-1, the oldest Y-enriched rims/overgrowths on matrix monazite grains are 1753 ± 4 Ma (rims of grains 12 and 14). The accuracy of this rim age is somewhat questionable, however, in light of U-Pb discordance concerns raised above (cf., Figs. 3b, 4a, 4b, and 6b of Dahl et al. 2005), as well as lack of microtextural evidence that staurolite formed directly from garnet. Thus, early prograde staurolite growth is only suggested as having occurred at 1753 ± 4 Ma and ~575 C (see Fig. 9). At the PR-1 locality, the BMD was unroofed from ~25 km below the paleosurface to ~16 km, with initial sillimanite growth reßecting decompression from the kyanite to sillimanite stability Þelds (Terry and Friberg 1990). This unrooþng (~2.5 kbars, or ~9 km) was probably responsible for post-d 2, pre- D 3 cooling of the BMD mid-crust (Holm et al. 1997), both of which are depicted in Figure 9. UnrooÞng/cooling of the PR-1 locality probably occurred some time between ~ Ma, although monazite ages of ~ Ma unique to adjacent rocks (granite BM-20 and metapelite ST-112; Figs. 1 and 2) suggest that local unrooþng was underway by this time. Further, the ST-112 locality sits astride the Grand Junction fault (GJF, Fig. 1), where structurally deep BMD rocks to the west (like PR-1) were vertically juxtaposed against structurally shallower rocks being folded immediately to the east (Friberg et al. 1996). Brecciated garnet indicative of faulting is common in the GJF (Friberg et al. 1996), and metamorphic ßuids necessary to trigger monazite growth could have been focused in the fault zone. Core and rim domains of ST-112 monazite formed at 1732 ± 9 and 1710 ± 11 Ma (Dahl et al. 2005). From these relationships it is proposed that the GJF accommodated differential uplift/unrooþng of the Bear Mountain dome (west) and the lower-p synformal rocks (east) to the same mid-crustal level beginning at 1732 ± 9 Ma. These events and resultant mid-crustal cooling in the Black Hills were interrupted by the onset of Harney Peak magmatism and associated reheating of metamorphic country rocks, the onset of which has been dated precisely at 1715 Ma (Redden et al. 1990; Schaller et al. 1997; Dahl and Frei 1998; Dahl et al. 2005). In the Bear Mountain dome itself, exposed granitic gneisses have yielded a ~2600 Ma Neoarchean age (McCombs et al. 2004). However, local gravity data have been interpreted as indicating the presence of subsurface Harney Peak granite (Duke et al. 1990), an interpretation that is supported by occurrences of concentric S 3 doming fabric (Fig. 1) and associated monazite dated at ~ Ma (Fig. 6; and Fig. 4 in Dahl et al. 2005). By comparison, muscovite in a fracture Þlling associated with doming in the eastern BMD near the GJF yielded an Rb Sr date of 1680 ± 25 Ma (R.E. Zartman in Ratté 1986). The most precise 207 Pb/ 206 Pb ages of monazite available for the southern Black Hills potentially bracket the time interval of Harney Peak magmatism, doming, and metamorphism between 1715 ± 3 Ma (Redden et al. 1990) and 1692 ± 5 Ma (this study). Additionally, precise 207 Pb/ 206 Pb ages of large apatite grains appear to date an intervening episode of pegmatite intrusion at 1702 ± 4 Ma (Krogstad and Walker 1994). Considered together, these results suggest that Harney Peak magmatism and associated midcrustal events in the Black Hills spanned a ~23 ± 8 m.y. interval, which is similar to that documented in the Tuolumne intrusive suite, California (Coleman et al. 2004).

13 1724 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE Both heat-ßow (Holm et al. 1997) and petrologic evidence (sequential growth of D 3 garnet/mica/sillimanite and post-d 3 staurolite) support the near-isobaric, prograde reheating path shown in Figure 9. This path is contrained by published thermobarometry indicating peak P-T conditions of ~ kbar and ~600 C in the BMD and environs (Terry and Friberg 1990; Helms and Labotka 1991; Friberg et al. 1996). The path is further constrained by the growth of prograde staurolite at sillimanitezone conditions (Figs. 3c and 9), which implies a minimum pressure of ~4 kbar (assuming Eq. 3 and the aluminosilicate triple point of Spear et al. 1999). Finally, if the growth of Y-enriched monazite rims (Fig. 5a) was coupled with garnet breakdown as argued above, then renewed crossing of Equation 3 to produce the post-d 3 staurolite occurred at 1714 ± 13 Ma and/or 1692 ± 5 Ma (monazite rim ages). There is no microtextural evidence that PR-1 garnet and staurolite actually formed by prograde crossings of Equations 2 and 3, respectively, as depicted schematically in Figure 9. Rather, staurolite (Fig. 3c) appears to have formed more at the expense of matrix micas than of garnet. Likewise, there is no evidence for garnet resorption (Fig. 4a) or atoll textures having resulted from retrograde crossings of Equation 2. Instead, the embayed and atoll garnets associated with randomly oriented mica in metapelite PR-1 preserve evidence of post-d 3, partial replacement by mica and quartz (Fig. 4a; see also Terry 1990). Also, post-d 3 pseudomorphs of coarse biotite ± quartz after garnet are commonplace in spotted schist SC-9a (Redden and Duke 1996; Clement 1995; Friberg and Dahl 1999). Considered together, these textural relationships suggest that a ßuid-driven ionic reaction such as: almandine + 14 H 2 O + 2 K Fe 2+ = 2 annite + quartz + 24 H + (4) (Carmichael 1969) was responsible for both garnet consumption and coupled growth of Y-enriched monazite rims in metapelites SC-9a and PR-1. A thermal (i.e., non-ionic) reaction that may closely monitor the behavior of Equation 4 is: almandine + muscovite + 2 albite + 2 H 2 O = annite + 3 quartz + 2 paragonite. (5) The P-T conditions for this retrograde reaction were determined by inputting SC-9a mineral compositions (Clement 1995) into the TWEEQU program of Berman (ver 2.02, 1991) and varying the activity of H 2 O, as plotted in Figure 9. According to Equation 5, biotite is favored over garnet by decreased T and/or increased H 2 O activity, whereas garnet is favored over biotite by increased T and/or decreased H 2 O activity. Indeed, Friberg and Dahl (1999) observed that neocrystallization of garnet was localized in the biotite-quartz spots of metapelite SC-9a, which they considered as evidence that the reverse of Equations 4 and 5 had produced a second generation of prograde garnet in that rock. Applying these observations to PR-1, it is concluded that destabilization of garnet Grt 2 could have been prograde (Eq. 3; coupled with growth of monazite and staurolite; Fig. 4c of Dahl et al. 2005) and/or retrograde (Eq. 5; coupled with growth of monazite and biotite; Figure 6a, this study). These garnet breakdown reactions are schematically indicated in Figure 9 as occurring at 1714 ± 13 and/or 1692 ± 5 Ma, respectively. Finally, a near-isobaric, ~ Ma cooling path through the andalusite stability Þeld is constrained for the BMD (Fig. 9) by 40 Ar/ 39 Ar cooling ages (Dahl et al. 1999; this study) and by occurrences of retrograde andalusite enclosing the S 3 fabric (northern edge of Harney Peak dome; Redden and Duke 1996). Hornblende from an amphibolite adjacent to metapelite PR-1 yielded the precise plateau age of 1691 ± 10 Ma (Dahl et al. 1999) shown in Figure 9, which probably reßects the time both of Ar closure as the rock cooled through ~ C and of late andalusite growth in the Harney Peak dome. This cooling age equates to the 1692 ± 5 Ma age of Y-enriched monazite rim growth in metapelite PR-1 (Fig. 6a). Cooling in the midcrust continued from ~ C at 1691 Ma (hornblende closure) to ~350 C at 1608 ( 40 Ar/ 39 Ar plateau age of muscovite), which nominally corresponds to a time-integrated cooling rate of ~2 3 C/m.y. However, continued cooling through ~300 C (biotite closure) was delayed until ~1480 Ma (pseudo-plateau age of biotite) suggesting that, by ~ Ma, cooling had slowed to ~0.4 C/m.y. This apparent reduction of cooling rate may signify that by ~ Ma a normal continental geotherm (~25 C/km) had been reestablished following the ~ Ma thermal perturbation (Harney Peak magmatic event). Black Hills thermotectonism and assembly of southern Laurentia Proximity of the Black Hills to adjacent, largely buried Precambrian terranes (see Figs. 1 and 10) has made this well-exposed Laramide uplift a logical focal point for reconstructing the details of Paleoproterozoic supercontinent assembly in south-central Laurentia. Figure 10 illustrates spatial and temporal details of terrane assembly as discussed below beginning at ~1850 Ma. Various workers (Dahl et al. 1999; Nabelek et al. 2001; Chamberlain et al. 2002) have recognized that terminal collision in the northern segment of the THO (~ Ma; Hearne, Sask, and Superior cratons; Bickford et al. 1990) predated its counterpart in the southern segment (~ Ma; Wyoming, Dakota, and Superior cratons; Baird et al. 1996). Based upon this diachroneity, Dahl and Frei (1998) distinguished a ~ Ma Black Hills orogen (BHO, Fig. 1) from the ~ Ma THO immediately to the east, and Dahl et al. (1999) proposed that Wyoming was a separate microcontinent that docked with Laurentia at ~ Ma. Regional truncation of ~ Ma arc-related rocks of the southern THO by the younger ~ Ma Central Plains orogen (CPO) (Figs. 1 and 10) has been inferred from geophysical trends (Sims and Peterman 1986; Sims et al. 1991; Sims 1995). However, geological relationships between the diachronous THO and BHO have not been fully established. In the Black Hills, evidence of Ma metamorphism consistent with this truncation is provided by ~ Ma total-pb ages of monazite in metapelites EG-56 and PR-1 (Figs. 2a and 2d). These ages cannot signify a detrital origin, inasmuch as depositional ages of the host rocks are within ~ and ~ Ma timeframes, respectively (Dahl and Mc- Combs 2005); instead, they indicate in-situ monazite growth at ~ Ma (Dahl et al. 2005). Because pre-s 1, Ma deformational fabric has not been recognized in the Black

14 DAHL ET AL.: ELECTRON PROBE (ULTRACHRON) MICROCHRONOMETRY OF MONAZITE 1725 FIGURE 10. Schematic spatial and temporal relationships of Ma terrane assembly in S-central Laurentia and resultant timing of deformational fabric formation in the Black Hills, as inferred in this study (cf., Fig. 1, lower right inset). Position of the Wyoming craton is arbitrarily Þxed. Teeth denote upper plates of subduction zones. BH = Black Hills; CB = Cheyenne belt (Chamberlain 1998); D = Dakota block (Baird et al. 1996); S and W = Superior and Wyoming cratons; T = Trans- Hudson arc terrane (Sims and Peterman 1986); and Y = northern Yavapai arc terrane (Karlstrom et al. 2002). NNW directions of Yavapai and Superior terrane transport, relative to the Wyoming craton, are inferred from Day et al. (1999) and Klasner and King (1990), respectively. Hills, however, further insight regarding the tectonic setting of ~ Ma monazite growth is gained only by considering regional relationships with the buried THO, Dakota block, and Superior craton to the east (Fig. 1). Baird et al. (1996) inferred from geophysical data (E-W seismic transect of the Williston basin, ND) that the Dakota block is an intervening suspect terrane bounded by W- and E-dipping subduction zones that face the Wyoming and Superior cratons, respectively (Fig. 10), and that faulted crustal wedges overlie both subduction zones. Following this model (see also Klasner and King 1990), the western crustal wedge is dominated by faulted metasupracrustal rocks of the Black Hills (BHO; west) and arc-related rocks (THO; east; Fig. 1). These spatial relationships, coupled with known ~ Ma protolith ages of the arc-related rocks (Sims and Peterman 1986; Sims et al. 1991), suggest that the ~ Ma monazite ages in the Black Hills indicate the timing of (contact?) metamorphism associated with (continental-arc?) magmatism in the adjacent, buried THO. As implied in Figure 10, both processes may have resulted from initial Ma convergence of the Wyoming and Superior cratons (Wyoming- Dakota?), distinct from terminal Wyoming-(Dakota)-Superior collision(s) that occurred between 1800 and 1700 Ma. The remainder of this discussion focuses on the main post Ma deformational fabrics recognized in the Black Hills ENE-trending S 1, NNW-trending S 2, and concentric S 3 (Figs. 1 and 2f) and their regional tectonic signiþcance. This paper and the companion study (Dahl et al. 2005) have established ages of 1775 ± 10 Ma, 1750 ± 10 Ma, and 1714 ± 13 to 1692 ± 5 Ma for these fabrics, related to N-NW-directed arc accretion, E-Wdirected continental collision, and local doming, respectively. Henceforth, following Karlstrom et al. (2002), the CPO (Sims and Peterman 1986) and Colorado province (Chamberlain 1998) are considered as part of a more extensive, northern Yavapai island-arc terrane (designated as YAV and Y in Figs. 1 and 10, respectively). The timing of Yavapai-Wyoming arc-continent suturing along the well-exposed Cheyenne fold/thrust belt of SE Wyoming (CB; Figs. 1 and 10) and its buried eastern extension, south of the Black Hills (SD), is broadly constrained between ~ Ma (Chamberlain 1998; Chamberlain et al. 2002; Dahl and Frei 1998; Dahl et al. 1999). However, the most precise fabric-age relationships now suggest that the arc-continent accretion was time-transgressive along ~250 km of the CB and its eastern extension, as schematically shown in Figure 10. That is, docking of the Yavapai terrane with the southern Wyoming craton occurred at 1775 ± 10 Ma south of the Black Hills (U-Th-Pb monazite; this study), 1763 ± 6 Ma to the west (Laramie Mountains, U-Pb sphene; Resor et al. 1996), and 1753 ± 3 to 1748 ± 6 Ma even farther to the west (Medicine Bow Mountains, U-Pb zircon and sphene; Strickland et al. 2004). This apparent age progression further implies that Yavapai-Wyoming collision was oblique and N-NW-directed, with the leading edge docking south of the Black Hills at 1775 ± 10 Ma (monazite data, this study) and the trailing edge docking in the Medicine Bow Mountains at 1749 ± 7 Ma (combining the data of Strickland et al. 2004), as shown in Figure 10. Accordingly, within 2σ uncertainties, the time interval of Yavapai docking could have been as brief as ~9 m.y. or as long as 43 m.y. along the ~250 km distance between