Abstract

The Wooley Creek batholith is a tilted, zoned, calc-alkaline plutonic complex in the Klamath Mountains, northern California, USA. It consists of three main compositional-temporal zones. The lower zone consists of gabbro through tonalite. Textural heterogeneities on the scale of tens to hundreds of meters combined with bulk-rock data suggest that it was assembled from numerous magma batches that did not interact extensively with one another despite the lack of sharp contacts and identical ages of two lower zone samples (U-Pb [zircon] chemical abrasion–isotope dilution–thermal ionization mass spectrometry ages of 158.99 ± 0.17 and 159.22 ± 0.10 Ma). The upper zone is slightly younger, with 3 samples yielding ages from 158.25 ± 0.46 to 158.21 ± 0.17 Ma, and is upwardly zoned from tonalite to granite. This zoning can be explained by crystal-liquid separation and is related to upward increases in the proportions of interstitial K-feldspar and quartz. Porphyritic dacitic to rhyodacitic roof dikes have compositions coincident with evolved samples of the upper zone. These data indicate that the upper zone was an eruptible mush that crystallized from a nearly homogeneous parental magma that evolved primarily by upward percolation of interstitial melt. The central zone is a recharge area with variably disrupted synplutonic dikes and swarms of mafic enclaves. Central zone ages (159.01 ± 0.20 to 158.30 ± 0.16 Ma) are similar to both lower and upper zones crystallization ages. In the main part of the Wooley Creek batholith, age data constrain magmatism to a short period of time (<1.3 m.y.). However, age data cannot be used to identify distinct magma chambers within the batholith; such information must be extracted from a combination of field observations and the chemical compositions of the rocks and their constituent minerals.

INTRODUCTION

Suprasubduction zone magmatism is responsible for the generation of large volumes of magma expressed as batholiths and volcanic arcs. Because arcs develop episodically, a great deal of recent research has focused on the tempo of intrusive and volcanic activity, on the temporal and petrological relationships between plutonic and volcanic magmas, and on the conditions necessary for large volumes of magma to exist in subvolcanic magma chambers. A number of workers have used high-precision U-Pb (zircon) ages to show that some large batholiths are assembled over several million years (e.g., Coleman et al., 2004; Glazner et al., 2004; Matzel et al., 2006; Grunder et al., 2008; Schaltegger et al., 2009). In the case of large zoned intrusions such as the Tuolumne batholith (California, USA), these results have been interpreted to indicate that large intrusions result from emplacement and amalgamation of many distinct magma batches that were not able to interact extensively with each other (Coleman et al., 2004; Glazner et al., 2004; Bartley et al., 2008). Nevertheless, detailed studies of magmatic fabrics and zoning patterns in parts of the Tuolumne batholith indicate that distinct intrusive contacts do not always separate batches of different ages (e.g., Memeti et al., 2010).

Models in which individual magma batches crystallize completely, preventing extensive interaction with one another, are contradicted by the presence of voluminous ignimbrites that are erupted during single events (e.g. Bacon and Druitt, 1988; Bachmann et al., 2002; Hildreth, 2004; Christiansen, 2005; Lipman, 2007). These eruptions require storage of large volumes of magma in the crust, at least for short periods of geologic time. Discrepancies between models for development of ignimbrite magma systems and models for magma batches unable to interact with one another have led some to suggest that large-volume ignimbrites are unrelated to typical batholith-scale plutons (Mills and Coleman, 2010; Tappa et al., 2011).

It is clear that there is a need for field and laboratory tests to determine whether magma batches emplaced incrementally remain as compositionally isolated entities or whether collections of magma batches are able to interact and perhaps homogenize in large crustal magma chambers (e.g., Tepley et al., 2000; Ohba et al., 2007; Turnbull et al., 2010; Ruprecht et al., 2012). The first logical test involves high-precision dating, commonly U-Pb dating of zircon. In cases where emplacement ages span a range of time longer than the time in which a magma batch should solidify, incremental emplacement with little to no communication between the batches is supported. However, it is possible that emplacement of batches unable to interact with one another occurred in such a short period of time that even high-precision dating cannot distinguish batches; in such cases, field, petrographic, and geochemical data must be used. For example, batch-wise emplacement with little or no communication between batches might be expected to leave mineral and geochemical evidence of isolated magma units unrelated by in situ differentiation processes such as fractional crystallization, magma mixing, or assimilation. We apply such tests to the Wooley Creek batholith (WCb), a large (326 km2), tilted plutonic complex in the Klamath Mountain province of northern California (USA; Fig. 1). Detailed mapping and petrographic data are combined with high-precision U-Pb dating of zircon and bulk-rock geochemical data to characterize zoning of the batholith, determine the physical and temporal relationships between magma zones, and develop a model for batholith assembly.

GEOLOGIC SETTING

Klamath Mountains

The Klamath Mountains geologic province, northwestern California and southwestern Oregon, consists of a sequence of tectonostratigraphic terranes accreted to North America. The province is subdivided into four belts, each of which consists of a number of terranes. The belts are separated by east-dipping shear zones with reverse sense of motion. In general, the overall age of the belts becomes younger to the west, structurally downward (see Irwin, 1960; Davis, 1968; Saleeby et al., 1982; Snoke and Barnes, 2006). Most of the component terranes of the province have oceanic affinities (ophiolites, mid-oceanic ridge basalt–related mafic assemblages, subduction mélanges, arc sequences, flysch deposits) and many formed in suprasubduction zone settings.

The WCb and Slinkard pluton (Fig. 1) were emplaced into three lithologically, chemically, and isotopically distinct tectonostratigraphic terranes of the western Paleozoic and Triassic belt (Fig. 1; Hotz, 1971; Irwin, 1994; Snoke and Barnes, 2006). The structurally lowest host unit is the Triassic–Jurassic Rattlesnake Creek terrane, an ophiolitic mélange that contains metaserpentinite, metagabbro, metabasite, hemipelagic metasedimentary rocks, and marble (Donato et al., 1982) metamorphosed ca. 168 Ma (Garlick et al., 2009). The Middle Jurassic western Hayfork terrane overlies the Rattlesnake Creek terrane and consists of an arc-related clastic, volcaniclastic, and hemipelagic sequence (Fig. 1; Donato et al., 1982, 1996; Wright and Fahan, 1988). The volcanogenic rocks of the western Hayfork terrane range from basaltic to andesitic (Barnes et al., 1995). The structurally highest host-rock unit is the Triassic eastern Hayfork terrane, which is a chert-argillite mélange and broken formation with local olistostromal units (e.g., Wright, 1982; Donato et al., 1982; Ernst et al., 2008). Mélange blocks include metasandstone, marble, and serpentinized peridotite. Contractional deformation beginning ca. 152 Ma associated with the Nevadan orogeny placed these terranes along with the WCb in thrust contact above low-grade metasedimentary rocks of the Galice Formation and the Condrey Mountain Schist (Fig. 1; Irwin, 1972; Jachens et al., 1986). The Condrey Mountain Schist is exposed in a domal window northeast of the WCb (Fig. 1). The metamorphic grade of hanging-wall rocks above the Condrey Mountain dome increases toward the dome from greenschist (at ∼300 MPa) to granulite (at ∼800 MPa) facies (Donato, 1987, 1989; Petersen, 1982; Lieberman and Rice, 1986; Garlick et al., 2009), indicating that doming caused radial tilting of the hanging wall to the southwest, away from the dome (current geographical coordinates). Effects of this tilting were observed in the contact metamorphic assemblages in the WCb aureole, where mineral assemblages give pressure estimates of ∼300 ± 150 MPa along the southern contact, farthest from the dome, to ∼650 ± 150 MPa along the northeastern contact, nearest the dome (Barnes et al., 1986b). These pressure differences show that regional tilting of ∼15° and subsequent erosion have exposed ∼9 km of structural relief from north to south across the WCb.

The intrusive system consists of two plutons (Fig. 1), the WCb and the Slinkard pluton, plus basaltic, andesitic, and dacitic dikes that crop out along the southwestern margin of the WCb. The Slinkard pluton, which crops out northeast of the WCb (Fig. 1), is structurally, compositionally, and temporally linked to the WCb, and therefore extends the exposed vertical extent of the system to at least 12 km (Barnes et al., 1986b, 1990). The original thermal ionization U-Pb (multicrystal zircon) dating of the WCb and Slinkard pluton yielded an age of ca. 161 Ma +4, −2 Ma (Late Jurassic; Barnes et al., 1986a). In this study we focus on the WCb, because of the availability of high-precision U-Pb (zircon ages) and the fact that many samples of the Slinkard pluton have undergone deuteric alteration.

BATHOLITH ZONATION

The WCb is broadly zoned from more mafic rocks in the structurally deeper northern and eastern parts of the pluton to the most felsic rocks in the structurally shallower southwestern area (Barnes, 1983). This zoning is shown in detail in Figure 2, which illustrates the variation in rock type and in varietal minerals across the batholith. On the basis of rock type and mineral assemblage, lower and upper zones can be mapped. These zones are locally separated by a 2–3-km-wide region that contains numerous synplutonic mafic dikes, microgranitoid enclaves and enclave swarms, and a variety of host-rock xenoliths in a sheeted tonalitic to quartz dioritic matrix. Locally intense hypersolidus deformation is observed. We refer to this zone as the central zone, and discuss its relationship to the upper and lower zones in the following. Along the western, southwestern, and southern margins, discontinuous mafic selvages of pyroxene-bearing quartz diorite and gabbro crop out; these zones reach 1 km in width (Fig. 2).

Lower Zone

At the outcrop scale, the lower zone shows varying degrees of heterogeneity. In some locations, for example in the Pigeon Roost area (Fig. 2), rock types vary on the scale of meters to tens of meters, whereas in other parts local modal layering is observed (Fig. 3A).

The lower zone consists of gabbro through tonalite. Lower zone samples are generally characterized by the assemblage augite + orthopyroxene + biotite ± Ca-amphibole (hereafter hornblende) (Figs. 4A, 4B). Samples that lack pyroxene contain clusters of hornblende or more commonly actinolitic amphibole rimmed by hornblende. These clusters are interpreted to be relict pyroxene surrounded by magmatic (nonperitectic) hornblende. As with the variation in rock type, there is little or no systematic large-scale spatial variation of the mafic mineral assemblage; however, texture varies at scales of tens to hundreds of meters. Oxide minerals (magnetite and ilmenite) are in low abundance in most samples.

Variably shaped masses of pyroxenite, melagabbro, and gabbro are locally abundant in the Pigeon Roost region (location in Fig. 2; Figs. 3B, 3C). Grain size in such bodies is typically coarser than in the more common quartz diorite–tonalite. Contact relationships between quartz diorite–tonalite host and these coarser bodies are variable but are consistent with the coarse-grained pyroxenites-melagabbro-gabbro bodies being intrusive into and synmagmatic with their hosts. Evidence for this relationship consists of (1) coarse pyroxene grains from the pyroxenite-melagabbro as xenocrysts in the surrounding rocks, (2) the presence of melagabbro dikes into the host diorite (Fig. 3B) with local back-veining and/or disruption of pyroxenite-melagabbro by the surrounding rocks, and (3) rare enclaves of the host quartz diorite–tonalite enclosed in pyroxenites-melagabbro; in such instances a 1–3-cm-thick rind of hornblende separates the two rock types. Compositional banding that consists of centimeter to tens of meters variations in the proportions of mafic minerals is also present in the lower zone (Fig. 3A). Elsewhere, for example at Medicine Mountain (Fig. 2), large areas are underlain by quartz diorite–tonalite with scant variations in color index and rare bodies of melagabbro-pyroxenite. Many parts of the lower zone are cut by planar, commonly net-veined mafic dikes. These dikes typically are parallel walled and fine grained and range from gabbroic to quartz dioritic, with hornblende ± biotite as the mafic phases.

Upper Zone

The upper zone ranges from medium- to coarse-grained tonalite to granite and appears fairly homogeneous at the scale of the outcrop (Fig. 3D). Figure 2 shows that this compositional zonation, while variable, is from more mafic to more felsic rocks toward the west-southwest, structurally upward. The texture of nearly all upper zone samples is hypidiomorphic granular, with sparse centimeter-scale hornblende and plagioclase phenocrysts. The groundmass is composed of a seriate distribution of hornblende, plagioclase, biotite, quartz, and K-feldspar from ∼5 mm to ∼0.2 mm in diameter. Hornblende and plagioclase are euhedral to subhedral and are weakly oriented in samples with magmatic foliation. Biotite is euhedral to anhedral and in most samples is not oriented in the foliation plane. Quartz varies from interstitial in tonalitic rocks to euhedral and/or subhedral in granite. Euhedral quartz is a common inclusion in poikilitic K-feldspar in the granites, whereas K-feldspar is everywhere interstitial to poikilitic (Fig. 4C). Broken plagioclase crystals form inclusions in K-feldspar in granodioritic and granitic samples (Fig. 4C). In summary, the mineral assemblages and overall textures of upper zone rocks are essentially identical from one sample to another; samples differ only in mineral proportions and quartz habits. Accessory phases in the upper zone are apatite, zircon, allanite, and epidote.

Felsic dikes occur sparsely in the upper zone, and in the southern part of the pluton a small, medium-grained granitic intrusion with associated pegmatites, referred to as the late granite, cuts the upper zone rocks (Fig. 2). Mafic dikes are also uncommon; however, enclave swarms and accumulations of mafic enclaves (pillows) are common, particularly in the structurally lowest part of the zone (see following). It is common that the color index of tonalitic and granodioritic rocks near swarms of mafic enclaves is higher than the color index of rocks distal from enclave swarms.

Central Zone

The central zone (Fig. 2) represents the most heterogeneous part of the batholith in terms of lithology, mineralogy, structure, and intrusive relationships. In areas of excellent exposure, such as the Cuddihy Lakes basin (Fig. 2), the central zone is underlain by sheets of medium- to coarse-grained quartz diorite and tonalite (Fig. 3F). Some sheets contain abundant mafic magmatic enclaves and others are essentially enclave free (Fig. 3F). Central zone samples contain hornblende and plagioclase with the same habit as in the upper zone (Fig. 4D). Biotite varies from euhedral to interstitial and quartz and K-feldspar are interstitial. Most samples lack pyroxene but some contain clusters of amphibole similar to those from the lower zone that are interpreted to be relict pyroxene. Accessory minerals are apatite, zircon, and scant allanite.

The central zone is also characterized by abundant synplutonic dikes (Fig. 3E). These mafic-intermediate dikes are similar to those in the lower zone, fine to medium grained and consisting of hornblende + plagioclase ± biotite and accessory minerals. Net veining of these dikes is pervasive and with increasing structural height, ductile to brittle deformation and disruption of dikes becomes common (see following) (Fig. 3E; Barnes et al., 1986a, 1990).

Appinitic dikes and masses crop out mainly in the central zone, but are also sparsely present in the lower and upper zones. In nearly all cases, these dikes are back veined by the host. The appinites contain euhedral, centimeter-scale blocky hornblende set in an interstitial to poikilitic groundmass of calcic plagioclase ± quartz ± K-feldspar. It is common for euhedral, millimeter-scale pyroxene crystals to be enclosed in poikilitic plagioclase.

BATHOLITH–HOST ROCK CONTACTS

Contact relationships were described in Barnes (1983) and are reviewed here, along with new data and descriptions from the southwestern contact. The contact between the lower zone and host rocks along the northeastern contact is best described as a contact zone in which plutonic and metamorphic rocks are interleaved over a 30–100 m distance. This part of the contact currently dips shallowly to the north, whereas the eastern and western contacts between the lower zone and the host rocks are steep to vertical. Along the eastern contact of the lower zone, foliations in the batholith and host rock are subparallel to the contact, but along the northwestern margin of the pluton such foliations are discordant to the contact (Fig. 5A). In this area, the northwestern contact is locally defined by a younger, brittle, high-angle fault that overprints and deforms the contact. However, south of the fault, the foliation remains discordant to the contact (Fig. 2).

The western, southwestern, and southeastern contacts are primarily between rocks of the upper zone and host rocks. This contact is generally steep and sharp, but is locally gradational over <10 m of interleaved plutonic rock and host rocks. The discontinuous mafic selvages along the western contact (Fig. 2) are underlain by pyroxene-bearing mafic and/or intermediate rocks (mafic selvages; Barnes, 1983) that separate upper zone rocks from host rocks of the western Hayfork terrane. Contacts between the selvages and host rocks are sharp, but the internal contacts between the selvages and the upper zone rocks are gradational (Barnes, 1983; Coint, 2012). Other such contact zones may exist because the color index of upper zone rocks is locally higher adjacent to the upper contact than it is 100 m inside the pluton. However, the quality of exposures does not permit mapping of such zones. The southeastern contact exposed in the Salmon River (Fig. 6) dips moderately to the west.

The southernmost contact with the host rocks is a plexus of mafic to intermediate rocks that intrude chert-argillite breccia and calc-silicate rocks of the eastern Hayfork terrane (Fig. 6). This region was mapped as a mafic selvage similar to that observed on the western margin (e.g., Barnes et al., 1986a); however, the southern contact zone is distinct from the western contact zones in at least three ways. First, the southern selvage consists in part of coarse- to fine-grained gabbro with variable crystal size and texture. Some gabbro samples contain sparse plagioclase and pyroxene phenocrysts with relict olivine, others have granoblastic textures, and still others contain abundant glomerocrysts of subhedral to euhedral orthopyroxene and augite with interstitial oxide minerals set in a matrix of euhedral, aligned plagioclase in which the foliation wraps around the glomerocrysts. Second, the contact between the southern selvage and the upper zone granodiorite is sharp, but lobate. Within the gabbroic rocks the proportion of plagioclase phenocrysts decreases away from the contact with granodiorite (Fig. 7B), suggesting that these crystals were entrained from the granodiorite. In some places, mafic lobes detached from the gabbro form gabbro enclaves within the granodiorite (Fig. 7B). Therefore, the southernmost selvage is interpreted to be comagmatic with granodiorite of the upper zone. A xenolith found in the granodiorite, preserving the contact between the gabbro and the eastern Hayfork terrane (Fig. 7A), indicates that part of the southern selvage was emplaced before the granodiorite. Third, the southern selvage is cut by late andesite and granodioritic dikes; the latter are rich in magmatic enclaves. These dikes are generally orthogonal to one other, with the andesitic dikes oriented approximately east-west and the granodioritic dikes approximately north-south (Fig. 7A).

The contact aureole of the WCb is variable in width and degree of emplacement-related metamorphism and deformation. At the map scale (e.g., Fig. 2), host-rock structures are weakly to strongly discordant except for regions within the contact aureole where significant ductile deformation and emplacement-related dynamothermal metamorphism occurred. Preexisting foliations in the host terranes along the margins of the intrusion are locally deflected, subparallel to the contact within a contact aureole from tens of meters to ∼1 km wide. In Barnes (1983), it was noted that intense deformation and isoclinal folding occurred within 50 m of the intrusion along the northwestern and northeastern margins. However, in many localities preexisting foliations and other structures are clearly discordant with the intrusive contact.

ROOF-ZONE DIKES

The southwestern contact of the batholith is the highest structural level exposed and is referred to here as the roof zone (Fig. 2; Barnes et al., 1986b). The intrusive contact is steeply dipping and does not restore to a subhorizontal roof; however, we retain the usage so as to avoid confusion with the earlier papers of Barnes et al. (1986a, 1986b). The roof zone is characterized by porphyritic dikes (to 3 m wide) (Fig. 3D) that range from basaltic to rhyodacitic in composition (Figs. 4E, 4F; Barnes et al., 1986a, 1990). A few of the mafic dikes are hornblende pyroxene microgabbro. However, most mafic dikes show evidence for static (contact) metamorphism or contain elongate to acicular Ca-amphibole in a matrix of epidotized plagioclase, suggestive of crystallization from an H2O-saturated melt. Intermediate-composition dikes are predominantly andesitic with phenocrysts of plagioclase, augite, and enstatite (Fig. 4E). These dikes intrude rocks of the upper zone within 2 km of the roof zone as well as the host rocks structurally above the batholith (Fig. 2). A smaller number of intermediate dikes are pyroxene-hornblende andesite.

Dacitic to rhyodacitic roof dikes contain phenocrysts of plagioclase + hornblende + quartz ± biotite ± rare augite (Fig. 4F). Dike groundmass textures vary from granophyric through hypidiomorphic granular, and it is common to find decimeter-scale rounded mafic magmatic enclaves in these dikes. Some dacitic dikes intrude andesitic dikes, some form the center of composite andesite + dacite dikes, and others are cut by andesitic dikes (Barnes et al., 1986a).

MAGMATIC ENCLAVES

Magmatic enclaves are widespread in the WCb, but their proportions vary. In general, these enclaves have a higher color index than their host rocks, but a few have equivalent or lower color index and differ from the host mainly in texture. The mafic magmatic enclaves vary from fine- to medium-grained and equigranular masses to porphyritic, with phenocrysts of plagioclase and hornblende, and rarely biotite. Groundmass phases are hornblende, plagioclase, and biotite ± quartz ± K-feldspar, with accessory apatite and scant Fe-Ti oxides ± zircon ± allanite.

In the lower zone, mafic magmatic enclaves are generally oblate ellipsoids and occur as isolated, 10-cm-scale bodies. In contrast, both isolated enclaves and enclave swarms are common in the central zone and the lower part of the upper zone. The enclave swarms may consist of rounded to tabular enclaves oriented parallel to magmatic foliation or of collections of rounded enclaves similar to so-called pillow swarms (Barnes, 1983; cf. Wiebe et al., 2002; Wiebe and Collins, 1998); in the former case, enclaves tend to vary in texture and phenocryst proportions, whereas in the latter case they tend to be texturally similar to one another.

XENOLITHS

Xenoliths are widespread in parts of the WCb but vary significantly in size, rock type, and abundance (Fig. 5D). In the lower zone, xenoliths are common within 500 m of the contact but are sparse in the rest of the zone. They consist primarily of amphibolite, biotite-rich schist, migmatitic quartzofeldspathic and calc-silicate gneiss, plus scant metaperidotite and rare garnet metagabbro.

A discontinuous zone across the center of the pluton that broadly overlaps with the central zone (e.g., Fig. 5D) is particularly rich in xenoliths, which range in size from centimeter scale to at least 50 × 10 m. The largest elongate blocks are oriented subparallel to the contact. Some xenoliths are enclosed by gabbro, whereas others are enclosed by tonalite, indicating that the blocks were isolated during emplacement of multiple magma batches. Rock types include migmatitic quartzofeldspathic and calc-silicate gneiss, metaquartzite and/or metachert, and rare skarn. Xenoliths in the upper zone occur almost exclusively within 200–500 m of the western-southwestern contact. In the Ten Bear Mountain area (Figs. 2 and 5B), swarms of migmatitic quartzofeldspathic and calc-silicate gneiss xenoliths crop out in zones as wide as 500 m. Along the southwestern contact in exposures in the Salmon River, xenoliths from the eastern Hayfork terrane are abundant, range from centimeter to meter scale, and form a ghost stratigraphy within the southern selvage (Fig. 6). These xenoliths consist of calc-silicate, metachert, and meta-argillitic rock types, and some are migmatitic. However, most xenoliths in the southern selvage also preserve preemplacement metamorphic fabrics, in contrast to the gneissic xenoliths that crop out further from intrusive contacts.

INTERNAL DEFORMATION AND FABRIC DEVELOPMENT

Structures within the batholith can be divided into three domains based on orientation, degree of development, and inferred conditions of formation. These domains correspond to the lithologic zonation described here. Primary structures within the lower zone include variably developed magmatic foliations defined by the alignment of plagioclase laths (Fig. 4A), elongate pyroxene prisms, biotite [001] faces, and sparse ellipsoidal magmatic enclaves. Within the lower zone near the northern contact this fabric is locally overprinted by recrystallized biotite and to a lesser extent hornblende (Fig. 4B). Plagioclase displays undulose extinction and minor subgrain development. However, in some samples interstitial quartz crystals are undeformed, indicating that quartz grew in the absence of significant strain before final solidification of the magma. Tonalitic rocks along the northern (structurally lowest) contact contain a crystal-plastic texture defined by recrystallized biotite and quartz aligned with, and overprinting, the magmatic foliation (Fig. 4B). The foliation is subparallel to the northern contact and varies in intensity along strike (Fig. 5). Away from host-rock contacts, magmatic foliations within the lower zone are generally north striking and steeply dipping, although a weak great circle with a shallowly north plunging B-axis may be defined (Fig. 5B). However, regional fold patters have not been observed in the host rocks; therefore, it is unlikely that the entire batholith–host rock system was folded while still partially molten. Magmatic lineations are difficult to discern, but where observed in the field are within the foliation plane and plunge shallowly north-south (Fig. 5B).

Foliation in the upper zone is primarily magmatic and is defined by the weak to moderate alignment of hornblende and plagioclase. In contrast to the lower zone, euhedral to subhedral biotite is randomly oriented and does not define the foliation. Despite an overall subhedral shape, quartz crystals display evidence of subgrain formation and minor recrystallization in some samples. Poles to foliations within the upper zone are widely distributed but generally define a great circle with a shallowly southwest plunging B axis (Fig. 5C). Magmatic lineations within the upper zone plunge shallowly with a variety of trends (Fig. 5C).

Fabric development in the central zone is dominantly magmatic and includes foliations, lineations, folds, and boudinage in mafic to intermediate dikes. Compared to the lower and upper zones, foliations are generally more strongly developed, are north-striking and steeply dipping, and are axial planar to several generations of open to isoclinally folded, synplutonic dikes (e.g., fig. 4B in Barnes et al., 1986a). Magmatic enclaves occur as ellipsoids and as angular blocks aligned in the foliation plane. Mafic dikes display various states of disruption, pinch-swell features, and boudinage (Fig. 3E). Late crosscutting basaltic to basaltic-andesite dikes are gently folded about the magmatic foliation and contain an axial-planar foliation defined by hornblende and plagioclase phenocrysts. The axial-planar foliation is parallel to the fabric in the host quartz-diorite to tonalite. Other finely crystalline synplutonic dikes show evidence for early brecciation followed by folding and boudinage.

Figure 5B displays an interpretive magmatic foliation trend line map, based on structural measurements throughout the batholith (Fig. 5A). Foliations in the lower zone are generally north trending, whereas in the upper zone foliations define overlapping onion-skin patterns. From outcrop to map scale, magmatic foliations are slightly to strongly discordant to internal, gradational lithologic contacts, and therefore must postdate lithologic zonation in the area where discordance is observed.

GEOCHEMICAL DATA

Geochronology

Multicrystal Isotope Dilution (ID)–Thermal Ionization Mass Spectrometry Analysis

The original mapping and geochronologic studies of the WCb and Slinkard pluton indicated that contacts between the zones of the pluton were gradational and that all magmatic units were coeval within the +4, −2 m.y. uncertainties of multicrystal thermal ionization mass spectrometry (TIMS) dating of zircon (Barnes, 1983; Barnes et al., 1986a). Our recent work has attempted to test these conclusions on the basis of further detailed mapping and single-crystal chemical abrasion (CA) TIMS dating of zircon.

Field work from 1984 to 2012 resulted in multiple traverses across boundaries between the lower and central zones and between the central and upper zones, as well as boundaries between western mafic selvages and the upper zone. It has not been possible to identify mappable contacts, or even contact zones, in these areas. Instead, the boundaries have been identified in terms of mineral assemblages and textures. Moreover, field evidence for magma mixing is widespread in the central zone (Barnes et al., 1986a). In addition, the discussion herein (and see Fig. 2C) shows that the trajectories of magmatic foliation crosscut zone boundaries without deflection. All of these field data indicate that the lower, central, and upper zones of the batholith were in a magmatic state at the same time, albeit with varying magma viscosities and crystal proportions.

New Single-Crystal ID-TIMS Data

Three samples from the lower zone, two from the central zone, and five from the upper zone were dated by CA-TIMS at the University of Wyoming (see Table 1).

The analytical method is in Table 1. Crystals were selected to avoid xenocrysts and inherited grains; therefore, the ages obtained are considered crystallization ages. Antecrysts as defined in Miller et al. (2007) were excluded from the calculations as they represent inherited zircon crystals that grew earlier within the same igneous system.

All but one of the ages reported here result from the average of three to five individual single-crystal ages (Table 1), whereas the age of the late-stage granite represents a single dated zircon grain.

Two samples from the lower zone were dated: a two-pyroxene diorite and a biotite hornblende tonalite. Zircon grains in these samples are large (300–600 µm) euhedral crystals that are broadly zoned in cathodoluminescence (CL) images. The ages are identical within the analytical uncertainty, 159.22 ± 0.10 Ma and 158.99 ± 0.17 Ma.

In the upper zone, dated samples were collected from structurally lower to higher levels and are representative of the compositional range, from biotite hornblende tonalite to biotite hornblende granite. In all samples from the upper zone, zircon is elongated to equant and varies in size between 100 and 200 µm. Truncated CL zoning is visible in some CL images, suggesting that some zircon cores may be inherited. Ages from structurally lower to higher samples are 158.22 ± 0.29 Ma, 158.25 ± 0.46 Ma, and 158.21 ± 0.17 Ma.

Samples from the central zone were collected in Cuddihy Lakes basin (Fig. 2). Sample WCB-4909 is a biotite hornblende quartz diorite collected in the zone of sheets described previously (Fig. 3E), whereas Z5 is a tonalite that was interpreted to be the youngest unit in the area because it cuts older sheets. Zircon in central zone samples is equant (100 and 300 µm) and displays sharply oscillatory zoning. As in the upper zone, truncation of CL zones in some zircon suggests that some cores might be inherited or that Zr concentrations in the magma fluctuated, resulting in partial dissolution of already grown crystals. Sample Z5 is 159.01 ± 0.20 Ma and sample WCB-4909 is 158.30 ± 0.16 Ma; these data belie the field observation that Z5 is the youngest part of the central zone. Moreover, overlap of the age of Z5 with ages from the lower zone and overlap of the age of WCB-4909 with ages from the upper zone strongly suggest that the central zone contains magma batches from both zones of the batholith.

Two samples from mafic selvages were dated. Zircon in these rocks is prismatic and displays simple CL zones. Sample WCB-2408, from the southern selvage, yielded an age of 158.32 ± 0.32 Ma and sample WCB-10510, from the western selvage, gave an age of 159.28 ± 0.17 Ma; these ages overlap the upper and lower zone ages, respectively.

Zircon from the late granite in the southern part of the batholith (MMB-377) is prismatic to equant. CL zones in the late granite crystals are sharp and narrow and vary from 100 to 250 µm. The late granite yielded a zircon age of 155.60 ± 1.19 Ma based on a single zircon analysis.

Bulk-Rock Major and Trace Elements

The following discussion builds on previous work (Barnes, 1983; Barnes et al., 1986a, 1990) with the addition of many new major and trace element data. The complete data set is available in the Supplemental Table1.

Samples from the lower zone of the WCb (excluding felsic dikes) range from 46 to 56 wt% SiO2 (Figs. 8A−8F). In contrast, upper zone samples range from 52 to 74 wt% SiO2 and central zone samples are, with one exception, in the 51–54 wt% range (Figs. 8A−8F). Lower zone samples are distinct in having higher MgO contents and lower TiO2, K2O, P2O5 Sr, Zr, Hf, and rare earth element (REE) contents than upper zone samples with similar silica contents. Moreover, lower zone samples show crudely increasing TiO2 (Fig. 8D) and P2O5 (Fig. 8L) with increasing SiO2, whereas these oxide abundances decrease in upper zone samples. Pyroxenite-melagabbro blocks and intrusions in the lower zone have higher MgO (Fig. 8A) and CaO (not shown) and lower Al2O3 and Sr (Figs. 8C, 8F) than the majority of lower zone samples, an indication that most pyroxenite and melagabbro samples are pyroxene cumulates. In most major element plots, upper zone samples show smooth, nearly linear variation as a function of SiO2 contents (Figs. 8A−8D). In contrast, several trace element trends show more complicated variation with SiO2. For example, the abundances of Zr and Hf broadly increase until ∼60 wt% SiO2 content (Fig. 8K; Hf not shown), and Ba increases to ∼64 wt% SiO2 content (Fig. 8E). These increases are followed by decreasing concentrations at higher SiO2 contents. Lanthanum abundances increase broadly to 62–63 wt% SiO2 content, and then remain constant or decrease (Fig. 8G); however, Sm, Yb, and Y abundances decrease throughout the silica range of upper zone samples (Figs. 8H–8J). The late-stage granitic intrusion in the southern part of the upper zone is distinct in having higher Sr, Zr, Hf, Ba, and Sm than most samples from the upper zone.

Compositions of central zone rocks generally overlap with the low SiO2 end of the upper zone and in several cases extend the upper zone trend to lower SiO2 contents. Central zone rocks are overall slightly enriched in P2O5, Sm, and Yb compared to upper zone samples at the same SiO2 content.

Basaltic roof-zone dikes overlap in composition with the lower zone and are most similar to high-MgO lower zone samples, with some overlap with pyroxenite and melagabbro blocks (Fig. 8). In contrast, the andesitic, dacitic, and rhyodacitic roof-zone dikes overlap with samples from the central and upper zones. Barnes et al. (1986a, 1990) interpreted the andesitic roof-zone dikes to be equivalent to lower zone rocks on the basis of similar mineral assemblages. However, this interpretation is contradicted by the broad overlap of andesitic roof-zone dike compositions with central and upper zone compositions and the lack of correlation is confirmed by trace element analysis of augite from the dikes and lower zone (Coint, 2012).

DISCUSSION

Temporal Relationships Between Zones of the Batholith

Geochronology has played an important role in our understanding of the timing and pace of pluton assembly (Coleman et al., 2004; Glazner et al., 2004; Miller et al., 2007; Schaltegger et al., 2009; Memeti et al., 2010; Paterson et al., 2011; Schoene et al., 2012). Such data have been used to assess whether multiple magma batches were emplaced into the middle to upper crust and solidified to form large, cohesive plutonic masses (e.g., Coleman et al., 2004; Matzel et al., 2006). Several assumptions are made in regard to these data. The first is that the youngest age provides the minimum crystallization age for a given sample, which is then assumed to represent a specific magma batch (e.g., Coleman et al., 2004). Older ages are interpreted as resulting from inheritance, either from the same magmatic system (antecrysts) or from the host rocks (xenocrysts) (Miller et al., 2007). The second assumption is related to whether intrusions assembled by multiple batch emplacement can remain partially molten for long durations. Intrusions of similar spatial scale and emplacement depth as the WCb are inferred to cool through the solidus in <0.5 m.y. to >1 m.y. (e.g., Paterson et al., 2011). Geochronology studies of large intermediate to silicic ignimbrites also indicate a span of ages of ∼100–300 k.y. for the respective magma chambers (e.g., Brown and Fletcher, 1999; Vazquez and Reid, 2002; Bachmann et al., 2007a, 2007b). Therefore, if ages of nonxenocrystic zircons span more than 1–2 m.y., they are commonly interpreted as antecrysts inherited from older magma batches that were reworked during pluton assembly.

As discussed here and elsewhere (Barnes, 1983; Barnes et al., 1986a), field observations indicate that the three zones of the WCb have gradational contacts with one another, thereby suggesting that the three zones were comagmatic. The contact between rocks of the upper zone and the mafic selvage along the western margin of the pluton is also gradational. Therefore, it is expected that the ages of all of these units should be within 100–300 k.y. In all zones, antecrysts were identified and excluded from the calculations (black rectangles in Fig. 9B). The available data (Table 1; Fig. 9) show that when individual sample dates are considered, overlap exists between ages of samples from the lower, one central zone sample (Z5), plus the western mafic selvage. These overlapping ages are distinct from ages of upper zone, the other central zone sample (WCB-4909), and the southern mafic selvage. The lower and upper zone ages do not overlap, the latter being at least 110 k.y. younger at the 95% confidence level, probably ∼800 k.y. younger. Clearly, the lower and upper zone magmas were emplaced relatively quickly compared to many plutons for which high-precision age data are available (e.g., North Cascades plutons: Matzel et al., 2006; Adamello: Schoene et al., 2012; Tuolumne: Memeti et al., 2010). For example, assembly of the upper zone occurred over as few as 100 k.y. (to 500 k.y. maximum if errors are taken into account). It is therefore not possible to use the age data alone to determine the extent of interactions between the different magma batches and whether two or more batches joined to form a single connected magma body.

Ages of the two samples from the central zone are distinct (159.01 ± 0.20 Ma and 158.30 ± 0.16 Ma); the first is coeval with the lower zone samples, whereas the second is coeval with the upper zone. The difference in ages of the two central zone samples is consistent with the sheeted nature of the zone (Fig. 3F). Evidently, despite the intense mingling and mixing that occurred in this part of the pluton, geochronologic evidence for batch-wise emplacement is preserved in the field and in the U-Pb ages. The lobate contacts between sheets (e.g., Fig. 3F), abundant mutually intrusive contacts between sheets suggests, and the intensity of magmatic deformation in much of the central zone indicate that some or all of the magma batches in this zone remained at magmatic conditions for time intervals to ∼300 k.y.

In contrast to the gradational contact relations between the main zones of the pluton, the late-stage granite has sharp, clear intrusive contacts with upper zone rocks and its U-Pb (zircon) age is distinctly younger, 155.60 ± 1.19 Ma.

In summary, the age data indicate that (1) the upper and lower zones of the WCb have distinct ages, but that these distinctions are blurred in the central zone, the latter evidently formed from magmas from both upper and lower zones; (2) gradational contacts between all zones, and the presence of temporally distinct sheets in the central zone, attest either to long-lived interstitial melts throughout the batholith or to reheating (defrosting) of lower zone and mafic selvage rocks as upper zone magmas were emplaced.

Internal Fabric Formation and Batholith Assembly

Host Rock–Pluton Relationship and Pluton Emplacement

Host-rock displacement and emplacement of magmas into the growing WCb occurred by multiple mechanisms that varied in magnitude and significance over the duration of batholith assembly. The presence or absence of emplacement-related deformation within the contact aureole may be due to the lithologically diverse and structurally complex host rocks (see discussion of Geologic Setting) as well as the nature of heat transfer in the aureole. Because of the discontinuous and/or disrupted nature of the host-rock terranes, it is difficult to quantify deformation within the aureole. However, detailed mapping within the aureole indicates that emplacement-related deformation along the northeastern contact includes a component of distributed ductile flow in a zone as much as 1 km wide (Donato et al., 1982; Barnes, 1983). In contrast, along the southern and southwestern margins, dynamothermal metamorphism is commonly absent, although a thermal aureole is well developed and reaches at least 200 m in width (e.g., Fig. 6); in these regions, host-rock structures are discordant to the intrusion contact.

In some areas, particularly along the northeastern and eastern margins, lit-par-lit dike–host rock contact zones several meters to hundreds of meters wide are consistent with magma emplacement by diking. These zones are broadly concordant with the elongate eastern and western margins of the batholith (Fig. 5) and with the overall elongation of the zones within the batholith. Thus, it is possible that the overall assembly of the WCb was facilitated by emplacement of north-trending (current geographic reference frame) elongate batches of magma. Exposures of the southern WCb contacts along the Salmon River (Fig. 6) display a variety of features that are consistent with this interpretation, including elongate outcrop- to map-scale north-trending intrusive sheets and host-rock screens. Contacts between diorite and granodiorite intrusive units in this region, also north trending, are locally lobate and indicative of synmagmatic recharge of new magmas into preexisting crystal-rich mushes (e.g., Fig. 7).

The presence of xenoliths adjacent to host-rock contacts indicates that stoping occurred during assembly of the batholith. However, xenoliths are also present in the interior of the batholith, some in dense swarms (Fig. 5D; Barnes, 1983). Work in progress demonstrates that these xenoliths are can be related to the three host-rock terranes on the basis of their geochemical signatures (Barnes et al., 2011). Because these xenolith swarms occur in the interior of the batholith as well as near the margins, it is evident that host rocks were incorporated into the growing batholith throughout much of its assembly.

Relationship Between Magma Batches and Magmatic Foliation

Foliation within the intrusion is of magmatic origin except along the northeastern contact, where near solidus and subsolidus deformation overprinted a magmatic fabric within a few hundred meters of the contact (Fig. 5). These protomylonitic fabrics are attributed to local mechanical coupling between the crystallizing lower zone and the host rocks (i.e., ballooning). Along the western, eastern, and southern margins of the intrusion the trends of magmatic foliation are variably discordant to the contact.

Magmatic foliations in the central zone and most of the lower zone strike broadly north-south, parallel to the elongation of the batholith. In contrast, magmatic foliations in the upper zone display a variable orientation with a weakly defined concentric pattern (Fig. 5). Moreover, in the northwestern part of the pluton, magmatic foliations in the upper and adjacent lower zone rocks strike approximately east-west, locally transect gradational compositional boundaries, and are at high angles to the batholith–host rock contact (Fig. 5). In general, foliation in all parts of the batholith is formed by alignment of plagioclase and either pyroxenes (lower zone) or hornblende (central and upper zones). In contrast, biotite is rarely oriented parallel to magmatic foliation in the upper and central zones, and in the lower zone only thin, poikilitic biotite crystals are in the foliation. These observations lead to the interpretation of magmatic fabric formation in the lower zone by flow and displacement of preexisting magmas during emplacement of new magmas into the growing batholith. Within the upper zone, the weakly concentric foliation pattern is attributed to formation in a mushy state prior to crystallization of biotite. Moreover, the single trend defined by bulk-rock data (Fig. 8) and the textural and mineralogical homogeneity of the upper zone indicates chemical connection over the entire exposed area of the upper zone. Therefore, we hypothesize that the anastomosing and weakly concentric magmatic fabric pattern is the result of convective overturn within the upper zone crystal mush prior to biotite crystallization and final solidification of the WCb.

Bulk-Rock Data

Lower Zone: Crystal Accumulation and Incremental Assembly

Compositional variation among lower zone samples is most readily described as an array of data points in which clear-cut trends are difficult to identify (Fig. 8). Overall, the expected decrease in MgO and increase in K2O with increasing SiO2 content is observed; however, with the exception of the cluster of tonalitic compositions, concentrations of elements such as TiO2, P2O5, Sr, Zr, and the REEs show broad scatter rather than distinct trends (Fig. 8). This lack of a well-defined compositional trend in the lower zone could be explained as the result of variable proportions of cumulate minerals from one sample to the next. It could also be related to emplacement of multiple batches of intermediate magmas (basaltic andesite to andesite) in which the batches were unrelated by in situ magmatic processes. Field evidence clearly shows that the lower zone formed from several magma batches, thus part of the scatter observed in the geochemical data must be related to the presence of compositionally distinct magma batches. Nevertheless, textural data (Fig. 4A) and the high CaO contents of many lower zone samples are consistent with accumulation of pyroxene ± plagioclase. In any case, both of these conclusions differ from those in Barnes (1983), wherein it was suggested that the lower zone differentiated from a large magma batch that was episodically recharged and mixed.

Evolution of the Upper and Central Zones by Fractional Crystallization

Upper zone samples show somewhat less scatter than is seen in the lower zone. In addition, for most elements the dacitic-rhyodacitic roof dikes have similar compositions and plot on the same trends as evolved samples of the upper zone (i.e., from the structurally highest levels of the pluton). As is the case for the lower zone, the range of upper zone compositions and the scatter observed in the bulk-rock data could result from emplacement of multiple magma batches of different composition or could result from differentiation of a more or less homogeneous upper zone magma. In the latter case, compositional scatter at any given SiO2 content would result from variable amounts of cumulate phases (e.g., Deering and Bachmann, 2010). The consistency of hornblende compositions and zoning patterns demonstrates the presence of a large, homogeneous upper zone magma body in which upward tonalite to granite zoning developed as this magma cooled and crystallized, mainly by upward separation of evolved melt-rich magma (Coint, 2012). This interpretation would explain the scatter observed within upper zone samples as the effect of variable proportions of cumulate minerals. It would also explain the central zone as a transition zone where lower zone magmas provided a mushy structural base for the upper zone magma. The structural setting would be an ideal trap for injections of mafic magmas after which they were variably deformed (synplutonic dikes) or disrupted (enclave swarms). Mafic magmas that rose through the central zone formed clusters of mafic pillows or enclave swarms in the lowest parts of the upper zone. The dacitic-rhyodacitic roof dikes are explained as leaks from the upper part of the magma body during differentiation.

If the upper zone was once a large homogeneous magma batch, then the variations of elements such as Zr, Hf, Sr, Ba, and the REEs have implications for the parental magma composition and the nature of the differentiation process. The trends for Zr, Hf, Ba, and La show changes in slope when plotted against SiO2, and the SiO2 value of these slope changes varies from one element to the next (Fig. 8). Such changes in slope are characteristic of fractional crystallization (Bowen, 1928). Moreover, the changes may be used to approximate the composition of the parental magma. For example, in the plot of Zr versus SiO2, the change in slope occurs at a SiO2 value of ∼60 wt% (Fig. 8). This silica value restricts parental compositions to ≤60 wt% SiO2. If the simplest case, i.e., that the upper zone parent had 60 wt% SiO2 and ∼125 ppm Zr, is assumed, then the upper and central zone samples with lower SiO2 and Zr would represent cumulates (e.g., hornblende + plagioclase ± biotite, without cumulate zircon) from the parental magma. The roof dikes and the two central zone samples with SiO2 > 60% and Zr > ∼125 ppm would be differentiates related to those cumulates (Fig. 8K). In contrast, upper and central zone samples with SiO2 < 60 wt% and Zr > 125 ppm would represent zircon + hornblende + plagioclase ± biotite cumulates and the samples that trend toward higher SiO2 and lower Zr contents would be differentiates.

Following the same logic, the upper zone magma was saturated in apatite and a Ti-bearing phase at the time of emplacement (decreasing P2O5 and TiO2; Figs. 8D and 8L, respectively) and became saturated in a Ba-rich phase (biotite or K-feldspar) at ∼64 wt% SiO2. We suggest that biotite is the probable Ba-rich phase because there is no commensurate decrease in K2O at ∼64 wt% SiO2 and because textural evidence shows K-feldspar to be the last phase to crystallize in all upper zone and roof-dike samples. Variation in La (Fig. 8G) suggests that a light REE–rich phase became stable at ∼63 wt% SiO2 (allanite), but that the middle and heavy REEs were compatible throughout crystallization of the upper zone magma. These trends would suggest that hornblende + apatite fractionation controlled middle and heavy REE variations and are consistent with the steady decrease in P2O5 with increasing SiO2 among the central and upper zone samples (Fig. 8L). If zircon had controlled the bulk-rock heavy REE budget, Yb and Y would show the same behavior as Zr, which is not the case (Fig. 8).

CONCLUSIONS

Contrary to earlier interpretations (Barnes, 1983; Barnes et al., 1986a, 1990), the WCb did not develop from a single big tank magma body. Rather, the batholith consists of two petrologically and temporally distinct zones: a lower zone that formed from multiple batches of gabbroic through tonalitic magmas at 159.22 ± 0.10 Ma and an upper zone of tonalite to granite emplaced at 158.99 ± 0.17 Ma. Compared to the Tuolumne batholith (Coleman et al., 2004; Paterson et al., 2011), the assembly of the WCb was rapid, the bulk of the batholith being emplaced in no more than 1.5 m.y. Moreover, despite the distinct ages of upper versus lower zones, their gradational mutual contacts and evidence for reworking of older into younger magmas (Coint, 2012) suggest that melt was present in the oldest parts of the system as upper zone magmas were emplaced. In the upper zone, the upward compositional zoning from tonalite to granite and the lack of internal intrusive contacts indicate that differentiation of this zone occurred within a single large magma batch.

The central zone yielded ages suggestive of interleaving and/or mixing of lower and upper zone rocks and magmas. Injection of basaltic magmas, now synplutonic dikes and enclave swarms, into the central and upper zone magmas provided heat to sustain a long-lived magma body. Moreover, the similarities in mineral assemblage and composition between the upper zone and dacitic roof dikes indicate that the upper zone was eruptible. The high crystallinity of these dikes (23%–54% phenocrysts) indicates that eruption was possible when the upper zone magma was a crystal mush.

The bulk compositions of many central zone rocks indicate a petrologic affinity to the upper zone, yet the ages of central zone rocks indicate affinities to both upper and lower zone magmatism. The age data indicate that the central zone is truly a transition zone that contains vestiges of structurally high levels of the lower zone into which numerous sheets of upper zone–like magmas were emplaced. The central zone may therefore be interpreted as part of the feeder system to the upper zone; if so, perhaps all of the upper zone magmas were emplaced in small increments, but these small incrementally emplaced magmas were homogenized by convective mixing prior to development of upward zoning. The thermal energy to drive convection is thought to come from repeated intrusions of mafic and intermediate magma that are locally preserved as synplutonic dikes and enclave swarms.

The WCb is an excellent example of a magmatic system that was emplaced incrementally. In the lower and central zones individual batches are preserved, whereas in the upper zone they homogenized to form a large volume of intermediate composition eruptible mush. Therefore, timing of emplacement is not the only controlling factor on whether a large interconnected magma chamber may form. In the case of the WCb, homogenization of upper zone magmas and possibly triggering of roof-zone dike emplacement probably required the added heat provided by emplacement of mafic magmas into the center of the batholith.

We thank Monika Leopold, Samantha Buck, and Brendan Hargrove for assistance in the field, and Dave Atwood and Glenna Atwood for their hospitality. We appreciate the helpful comments and advice of Olivier Bachmann, Scott Paterson, an anonymous reviewer, and A.E. Mike Williams. This work was supported by National Science Foundation grant EAR-0838342 to Yoshinobu and C. Barnes, grant EAR-0838546 to Chamberlain, a 2009 Geological Society of America Penrose grant to Coint, and the Texas Tech University Department of Geosciences.

1Supplemental Table. Bulk-rock geochemistry. If you are viewing the PDF of this paper or reading it offline, please visit http://dx.doi.org/10.1130/GES00930.S1 or the full-text article on www.gsapubs.org to view the Supplemental Table.