Abstract

The Late Cretaceous (ca. 75 Ma) Philipsburg Batholith in SW Montana provides an excellent setting to examine the emplacement mechanisms of large-volume silicic plutons in the Sevier fold-and-thrust belt of the western U.S. Cordillera. Magnetic fabrics, determined using the anisotropy of magnetic susceptibility from 119 sites and anisotropy of anhysteretic remanent magnetization data from a subset of these sites, suggest that the Philipsburg Batholith is a tabular body and that its emplacement involved subhorizontal magma flow controlled by local thrust faults, including the Georgetown-Princeton thrust, which served as conduits for magma ascent. Field relations and magnetic fabric data, which are typically well characterized at the site level and dominantly subhorizontal except near major thrust faults, suggest that the Philipsburg Batholith was emplaced within a fault-bend fold at the top of a ramp in the Georgetown-Princeton thrust system. New 40Ar/39Ar age spectrum data indicate that the entire Philipsburg Batholith was emplaced rapidly at ca. 74.8 ± 0.1 Ma. Paleomagnetic directional data indicate that the Philipsburg Batholith has experienced ∼9°–16° ± 8° of west-side-down tilt (tilt estimate and associated error are based on a new approach described herein) since emplacement in the Late Cretaceous. The data do not support any internal deformation of the batholith since emplacement and suggest that Sevier-age thin-skinned deformation in the Flint Creek Range had ceased by ca. 75 Ma. The modest magnitude of observed tilt could have occurred during subsequent Cenozoic extension. The inferred mode of emplacement of the Philipsburg Batholith may typify how large volumes of silicic magma are emplaced in shallow, fold-and-thrust belt settings.

INTRODUCTION

Emplacement of silicic plutons is an important component in the growth of continental crust. Most silicic intrusions are emplaced within convergent plate boundaries, where the principal stresses causing crustal shortening would seemingly prohibit emplacement of large magma bodies (Glazner, 1991; Karlstrom et al., 1993; Hutton, 1988, 1997; Paterson and Miller, 1998). This long-standing room problem is a fundamental question in our understanding of the role of plate tectonics and magma generation in the formation of continental crust. Many workers have suggested emplacement mechanisms that may explain magma ascent and intrusion during crustal shortening, and these typically require dilational space to be formed by some component of oblique faulting and nonplane strain (Tikoff and Teyssier, 1992; Castro and Fernandez, 1998; Benn et al., 1998a; Titus et al., 2005). Although deformation of the crust undoubtedly plays a major role in the ascent and emplacement of large, typically silicic, magma bodies, the actual mode of emplacement during crustal shortening is poorly understood. Knowledge of the internal architecture of granitic intrusions in relation to pre- and synemplacement structures and associated deformation of the host rocks will lead to more accurate models for the growth of continental crust during orogeny.

The Late Cretaceous to early Tertiary history of SW Montana is characterized by thin-skinned fold-and-thrust belt development (Sevier orogeny) and higher-angle Laramide-style reverse faulting involving Precambrian basement rocks (Harlan et al., 1988; Schmidt et al., 1990; O'Neill et al., 1990; Kalakay et al., 2001; Harlan et al., 2008). Contemporaneous granitic intrusions of a range of dimensions were emplaced east of the main magmatic arc (exposed in the Idaho Batholith), the largest pluton being the composite Boulder Batholith (Hamilton and Myers, 1974). This temporal and spatial overlap between magma emplacement and foreland thrust belt development offers an ideal setting for the study of plutonism associated with crustal shortening.

Field relations in SW Montana have prompted numerous studies of emplacement mechanisms for Late Cretaceous intrusions. Hamilton and Myers (1974) argued that the ca. 70–80 Ma Boulder Batholith was emplaced as a tabular sheet under the Elkhorn Mountain volcanics and thus questioned earlier interpretations by Klepper et al. (1971, 1974), who interpreted the Boulder Batholith as a steep-sided, thick intrusion. Some emplacement mechanisms involved lateral magma flow from the Idaho Batholith along major décollement surfaces beneath Sevier-style thrust plates (Hyndman et al., 1975) or emplacement into a pull-apart space formed by intersecting fault sets in the Helena Salient (Schmidt et al., 1990). More recently, shallow crustal intrusions (emplaced at 1–10 km depth) of Late Cretaceous age in SW Montana have been interpreted to have filled volumes of minimal differential stress within fault-bend folds above thrust ramps (Kalakay et al., 2001). The Kalakay et al. (2001) model (Fig. 1) proposes that ramp-flat geometry of thrust faults allows magma to fill space formed by local extension, possibly aided by back thrusting. An implicit assumption in this model is the migration of magma along the thrust surface to the top of the ramp instead of subvertical magma flow as dike-like sheets.

The Philipsburg Batholith is a Late Cretaceous intrusion in the Flint Creek Range (Fig. 2), west of the Boulder Batholith, and it lies in a similar structural position as the Boulder Batholith, as well as the three nearby intrusions studied by Kalakay et al. (2001). In this study, we use anisotropy of magnetic susceptibility (AMS) data, supplemented in some cases by anisotropy of anhysteretic remanent magnetization data, to approximate magma flow fabrics in order to evaluate the internal architecture and emplacement mode of the Philipsburg Batholith. The 40Ar/39Ar geochronologic data were obtained on biotite separates to better define the emplacement history of the batholith and to test the hypothesis that emplacement occurred by multiple pulses separated by resolvable time periods. Paleomagnetic data show that both plutonwide postemplacement deformation and any preferential localized internal deformation in the Philipsburg Batholith were minimal and validate interpretations based on the AMS data. Estimated mean paleomagnetic directions reveal no statistically resolvable component of internal deformation of the batholiths, and, hence, no corrections to AMS fabrics are required due to internal deformation. AMS data show mostly subhorizontal fabrics radiating away from local thrusts and are interpreted to suggest that emplacement was largely concentric away from at least one magma channel centered along the Georgetown-Princeton thrust. Our interpretation is fully consistent with the model of Kalakay et al. (2001) for the emplacement of the plutons in SW Montana.

Lageson et al. (2001) discussed the influence of large-volume plutons on the development of the Sevier-age orogenic wedge in SW Montana and argued that critical wedge taper was promoted by injection of large magma bodies into the allochthon of the thrust system, leading to further propagation of the thrust belt. If magma emplacement atop thrust ramps is common, then the added volume of plutons emplaced into the foreland may have similarly influenced the development of thrust belts in other orogenic belts throughout the geologic past (e.g., Mazzarini et al., 2004).

REGIONAL GEOLOGY AND SETTING OF THE PHILIPSBURG BATHOLITH

The Philipsburg Batholith is one of three large silicic intrusions in the Flint Creek Range, which lies ∼40 km west of Butte, Montana (Fig. 2). The 122 km2 surface area Philipsburg Batholith crops out to the west of the Mount Powell Batholith and Royal Stock (Fig. 3). All three bodies were intruded into metasedimentary and sedimentary rocks ranging in age from the Mesoproterozoic Belt Supergroup to the Lower Cretaceous Kootenai Formation. The Flint Creek Range is separated from the Boulder Batholith to the east by the Deer Lodge Valley, a Cenozoic normal fault–bounded valley.

The sedimentary sequence in the Flint Creek Range was intensely deformed during regional development of the Sevier fold-and-thrust belt in the Late Cretaceous (Hyndman et al., 1972; O'Neill et al., 2004). The principal fault in the western Flint Creek Range is the west-dipping Georgetown-Princeton thrust (Fig. 3), which places Belt Supergroup rocks over Paleozoic rocks as young as the Mississippian Madison Group (Hyndman et al., 1982). The fault trace runs roughly north-south, except where the fault swings to the east as it nears the contact with the Philipsburg Batholith. There is no evidence that the Georgetown-Princeton thrust cuts the Philipsburg Batholith, and therefore intrusion must have occurred during or after final movement on the Georgetown-Princeton thrust. The fault segment south of the Philipsburg Batholith ends at the center of the southern batholith margin (Ehinger, 1971; Hyndman et al., 1982; O'Connell, 2000). An eastern splay of the Georgetown-Princeton thrust is exposed south of the Philipsburg Batholith, exhibits steep dips, and has been mapped by some workers as a high-angle normal fault (e.g., Lonn et al., 2003).

At its northern margin, the Philipsburg Batholith cuts the high-angle Bungalow fault, a few kilometers west of the Georgetown-Princeton thrust (Fig. 3). The sense of displacement on the Bungalow fault is unclear due to poor exposure; both normal and reverse motion has been reported (Lonn et al., 2003; Lewis, 1998; Hyndman et al., 1982). We interpret the Bungalow fault as a back thrust accompanying Late Cretaceous displacement along the Georgetown-Princeton thrust. This movement sense is suggested by near vertical dips of the thrust plane, movement prior to intrusion, and fault contacts between Belt Supergroup rocks on the east side of the fault and Upper Paleozoic through Mesozoic strata west of the fault.

Strata in both the hanging wall and the footwall of the Georgetown-Princeton thrust system have been folded. The Philipsburg anticline trends N-S along the western margin of the Philipsburg Batholith and is cut by the intrusion. The Royal Gold Creek anticline, the Finley Basin anticline, and the Racetrack folds deform strata in the footwall of the Georgetown-Princeton thrust. The Racetrack folds include anticlines and synclines in an overall synclinorium between the Philipsburg Batholith and the Mount Powell Batholith and Royal Stock to the east. The synclinorium widens to the north. The fold axis of the Royal Gold Creek anticline changes from N-S to NW-SE closer to the Philipsburg Batholith. The Cable Mountain anticline is located between the main Georgetown-Princeton thrust and the eastern thrust splay south of the Philipsburg Batholith. The N-S fold axis of the Cable Mountain anticline is also deflected to the east closer to the Philipsburg Batholith.

The two intrusions in the eastern Flint Creek Range have been interpreted by O'Neill et al. (2004) as part of the uplifted footwall of the Anaconda metamorphic core complex, which developed during the middle Eocene. The detachment fault for the Anaconda metamorphic core complex dips shallowly to the east. The trace of the detachment is mapped on the eastern edge of the Flint Creek Range (Fig. 3) and extends south along the east side of the Anaconda Range. The Philipsburg Batholith is exposed ∼5 km west of the trace of the detachment, and no evidence of extension within the batholith was observed in the field or has been described by previous workers.

The Philipsburg Batholith ranges from granite to granodiorite/quartz monzodiorite (Fig. 4). The Philipsburg Batholith was originally mapped as a single intrusion, with some internal gradations in modal percent biotite, hornblende, and K-feldspar (Ehinger, 1971). Hyndman et al. (1982) divided the Philipsburg Batholith into the more mafic Bimetallic Stock to the west and the more felsic Dora Thorn Pluton to the east (Fig. 3). The modal percent of Fe-Mg silicate minerals varies considerably throughout the Philipsburg Batholith, especially within the Dora Thorn Pluton. The Bimetallic Stock averages 25% Fe-Mg phases, with hornblende often more abundant than biotite. The Dora Thorn Pluton averages 15% Fe-Mg silicates, but some samples have less than 5% biotite and no hornblende. Spatially restricted compositional gradations (Hyndman et al., 1982) in the pluton suggest that the Philipsburg Batholith may have been formed by multiple pulses of magmatism over a short span of time.

Several observations are consistent with emplacement and crystallization of the Philipsburg Batholith at a shallow crustal level. The typical silicate mineral crystal size in the pluton is 0.5–2 mm and is consistent throughout the field area with minimal fining toward the margins. Hyndman et al. (1982) reported some coarsening in the Bimetallic Stock near the contact with the Dora Thorn Pluton. The metamorphic aureole around the batholith is ∼2 km wide, and wall rocks are metamorphosed to the hornblende-hornfels facies (Hyndman et al., 1982). O'Connell (2000) placed the metamorphic assemblages of two samples from the eastern contact aureole in pressure-temperature (P-T) space and inferred temperatures of 500–650 °C and pressures of 1.2–2.7 kbar. Aluminum-in-hornblende geobarometry and hornblende-plagioclase geothermometry on four samples from the pluton indicate intrusion pressures between 0.9 and 2.0 kbar and temperatures of ∼720–820 °C, although no confidence limits were reported (O'Connell, 2000). Hyndman et al. (1982) suggested a maximum pressure of 0.6 kbar based on the presence of miarolitic cavities at the highest levels of the pluton. These observations suggest that the Philipsburg Batholith was emplaced at depths less than ∼6 km.

The Philipsburg Batholith has yielded K-Ar age information on hornblende and biotite (Hyndman et al., 1972). A sample from the southwest part of the intrusion yielded a hornblende date of 76.7 ± 2.5 Ma and a biotite date of 74.0 ± 2.1 Ma. A sample from the central part of the intrusion yielded a hornblende date of 72.0 ± 2.5 Ma and a biotite date of 73.4 ± 2.1 Ma. The age determinations are all within error of ca. 74 Ma. The slightly older dates in the southwest were used by Hyndman et al. (1982) to support the division of the Philipsburg Batholith into two plutons. Three apatite fission-track dates for the Philipsburg Batholith provided an average age of 65.9 Ma (Baty, 1973), ∼8 m.y. younger than K-Ar dates.

METHODS

Magnetic Anisotropy

Low-field anisotropy of magnetic susceptibility (AMS) is a quantitative method for determining the orientation and degree of development of preferred fabrics of Fe-bearing minerals and has been applied to many rock types and geologic problems (Hrouda, 1982). AMS has been extensively utilized to approximate magma flow fabrics in silicic plutons and to define emplacement models for silicic intrusions (Bouchez et al., 1990; Bouchez, 1997; Benn et al., 1998b; Aranguren et al., 2003; Titus et al., 2005; Gebelin et al., 2006; Stevenson et al., 2007).

The magnetic susceptibility, K, relates the induced magnetization of a solid to the strength of an applied magnetic field. The susceptibility of a rock combines the contributions to K from all minerals (diamagnetic, paramagnetic, and ferro/ferrimagnetic) making up the rock, and AMS is therefore the result of preferential alignment of anisotropic minerals. The anisotropy in K for a mineral or an assemblage of minerals in a sample can be expressed as a symmetric second-rank tensor with eigenvectors Kmax, Kint, and Kmin. Kmax is the AMS lineation, and Kmin is normal to the AMS foliation. The bulk susceptibility, K, for a sample is the mean of the three susceptibility eigenvalues. The degree of anisotropy, P′, and the shape parameter, T, are used to describe the relative strengths of Kmax, Kint, and Kmin. P′ ranges from 1 (isotropic) to infinity, and T ranges from −1 to 1, with negative values for prolate and positive values for oblate susceptibility ellipsoids. P′ and T are calculated following Jelinek (1978) as: 
graphic
and 
graphic
where Kmax, Kint, and Kmin are the eigenvalues of the susceptibility tensor, and η is defined as: 
graphic

In total, 1316 samples from 119 field sampling sites (6–15 samples/site) were collected in the Philipsburg Batholith. Roughly half of the AMS samples were also used for paleomagnetic study. In all cases, AMS was measured before discrete specimens were demagnetized for paleomagnetic measurements. Sites were located near logging roads/trails or within stream valleys and were spaced ∼1 km apart. Sampling density was lower in the northern and eastern parts of the batholith due to limited accessibility. Site locations were defined with a hand-held Global Positioning System (GPS) unit with an accuracy of ±10 m. Samples at each site were collected within ∼50 m2 (a single outcrop). All samples were collected using a portable gas drill with a nonmagnetic diamond drill bit and were oriented with magnetic and sun compasses. The samples were cut into 25 × 22 mm right cylinder specimens with a nonmagnetic diamond blade before measurements were taken to best approximate the volume of a sphere. At least one specimen from each sample was measured using a Kappabridge KLY-4S. Susceptibility axes Kmax > Kint > Kmin, degree of anisotropy P′, and shape parameter T were calculated using the program Anisoft (version 4.2) from Agico, Inc.

The anisotropy of anhysteretic remanent magnetization (AARM) (Jackson, 1991; Rochette et al., 1992; Raposo and Berquo, 2008) of samples can be measured to produce a second rank tensor or AARM ellipsoid, similar to the AMS ellipsoid. AARM measures only the preferred fabric of minerals capable of retaining a remanent magnetization after removal of a magnetic field, and therefore the preferred orientation of diamagnetic or paramagnetic minerals does not contribute to AARM.

The AARM values of 63 samples from eight sampling sites were measured to compare with AMS measurements from the same sites. AARM measurements were performed after specimens were alternating field (AF) demagnetized to 100 mT. The specimens for AARM measurement were cut into 15 × 13 mm right cylinders (∼2.5 cm3). An ARM was imparted along the first of 15 specific axes in a direct field of 0.1 mT superimposed on an alternating field decaying from 100 mT. The specimen was then measured and AF demagnetized to 100 mT. This process was repeated with ARM being measured for the 15 sample positions described in Tauxe (1998). The AARM ellipsoid and principal axes AARMmax > AARMint > AARMmin were then calculated, and errors were calculated using the Jelinek method of linear perturbation analysis (Tauxe, 1998).

Paleomagnetism and Rock Magnetism

Paleomagnetic data from 516 independently oriented samples from 63 sites (8–10 samples per site) were collected in order to evaluate postemplacement deformation of the Philipsburg Batholith. At least one specimen per sample was measured using a 2G Enterprises superconducting rock magnetometer at the University of New Mexico Paleomagnetism Laboratory. Specimens were progressively demagnetized in a series of alternating fields (20–30 steps) until at least 90% of the natural remanent magnetization (NRM) had been removed. Most Philipsburg Batholith specimens were 90% demagnetized in a range of peak fields from 15 to 35 mT, and few specimens required demagnetizing at fields above 60 mT. For a selected number of samples, an additional specimen was first demagnetized by low-temperature demagnetization (LTD) to 77 K using liquid nitrogen and warming in a low-field space. Three to four steps of LTD were followed by progressive alternating field (AF) demagnetization. LTD takes advantage of the change in the crystal structure of magnetite, from cubic to monoclinic at the Verwey transition (∼120 K). As the structure changes, the crystalline anisotropy of magnetite decreases, and remanent magnetization is only retained by elongate single-domain (SD) grains due to shape anisotropy (Dunlop and Özdemir, 1997). For at least half of the selected samples from the Philipsburg Batholith, over 70% of NRM was randomized using the LTD technique. The high percentage of NRM reduced in these samples suggests that the NRM is largely held by multidomain (MD) magnetite. The other population of samples subjected to LTD from the Philipsburg Batholith shows less than 10% reduction in NRM, and these samples were subsequently AF demagnetized.

The direction of characteristic remanent magnetization (ChRM) was determined using at least three sequential demagnetization steps chosen by visual inspection of orthogonal vector diagrams. ChRM was calculated using principal component analysis (Kirschvink, 1980). For eight specimens from one site, no linear demagnetization interval defined by three or more demagnetization steps could be identified. In these cases, great circles were fit to curvilinear demagnetization intervals. Site means were calculated using Fisher (1953) statistics when ChRMs were well determined for all samples, or by using a combination of stable end point and great circle analysis (McFadden and McElhinny, 1988).

Rock magnetic experiments included acquisition of an isothermal remanent magnetization (IRM) to saturation, direct field demagnetization of the IRM, AF demagnetization of anhysteretic remanent magnetization (ARM) acquired in a direct field of 0.1 mT and a peak alternating field of 98 mT, and AF demagnetization of a 98 mT isothermal remanent magnetization (IRM). These tests were performed on a small, but representative, group of samples from 14 sites in the Philipsburg Batholith. Hysteresis curves for 12 selected samples were obtained using a vibrating sample magnetometer at the Institute for Rock Magnetism, University of Minnesota.

40Ar/39Ar Geochronology

The 40Ar/39Ar method was used to acquire additional geochronologic data on the Philipsburg Batholith in order to test the proposed age difference, as suggested by K-Ar data, between the Bimetallic Stock and Dora Thorn Pluton, and to place a minimum age on movement of the Georgetown-Princeton thrust. The 40Ar/39Ar measurements were made on 13 biotite separates from the Philipsburg Batholith at the New Mexico Geochronology Research Laboratory, New Mexico Tech. Nine of the samples were from the Dora Thorn Pluton, two were from the Bimetallic Stock, and two were collected near the proposed contact between the two intrusions (Fig. 5). All samples were analyzed by incremental heating in a furnace using nine heating steps.

All 13 samples were irradiated in a single 20-hole irradiation disk with fluence monitors of Fish Canyon sanidine (28.02 Ma; Renne et al., 1998) placed in seven locations. Six crystals from each standard position were analyzed by laser fusion to calculate J-values for the irradiation tray. Because the signal to blank ratios (typically >1000 for 40Ar) were high for both the flux monitors and the biotite data, various blank determining methods did not have significant effects upon J-values or final age determinations. Analytical data and additional information about methods are provided in the GSA Data Repository.1

RESULTS

40Ar/39Ar Geochronology

Age spectra (Figs. 6A and 6B) for the 13 biotite samples are generally flat for significant gas fractions and typically show younger apparent ages in the first few steps and age plateaus defined by heating steps above 800 °C. The plateau segments generally have mean square of weighted deviates (MSWD) values that indicate a normal distribution of ages defining the plateau age. Only samples PB19, 81, 91, and 105 have plateau ages with associated MSWD values above the 95% confidence level for n – 1 degrees of freedom. For samples PB81, 91, and 105, we chose the displayed plateau age as the preferred age despite the moderately high MSWD. For sample PB19, the ‘G’ step is anomalously old, and for this sample, we suggest that the total gas age is the best age. K20 values (as determined by 39Ar) indicate that most samples are pristine and overall are not altered; however, some samples show values below 8% and thus suggest some chlorite alteration. In general, there is a correlation between lower K2O and higher MSWD; this may be expected based on the work of Lo and Onstott (1988), who showed age spectra complexity as a function of chloritization. The assigned dates for the 13 Philipsburg Batholith samples are summarized in Table 1.

Eleven of the samples yield 40Ar/39Ar biotite dates between 73.6 and 75.6 Ma, whereas samples PB17 and PB111 have younger dates of 65.4 ± 0.1 and 67.4 ± 0.1 Ma, respectively. A projection of the biotite dates onto a southeast-northwest line that bisects the Philipsburg Batholith, which is thus perpendicular to the northeast-southwest margins of the Mount Powell Batholith, shows a distinct age increase away from the Mount Powell Batholith (Fig. 7). A combination of the eight oldest biotite dates (except for PB19, which did not give a plateau) yields a weighted mean age of 74.77 ± 0.06 Ma (MSWD = 1.39) (Table 1). This normal distribution of dates indicates that there is no significant age difference among these eight samples, which indicates that the Philipsburg Batholith was emplaced as a single intrusion and that any age difference between the Bimetallic Stock and the Dora Thorn Plutons cannot be resolved by the biotite dates.

As indicated already, the four youngest dates are spatially associated with distance from the Mount Powell Batholith and systematically increase along a southeast to northwest trend. There is no textural or mineralogic evidence in the pluton at these sampling sites that would indicate a separate intrusion from the rest of the Philipsburg Batholith. It is likely that the dates obtained from these samples have been partially reset by intrusion of the Mount Powell Batholith, ∼1.5 km to the east. This pluton has also been mapped as Late Cretaceous, but the precise age of emplacement is unknown. Apatite fission-track dates (Baty, 1973) from the Mount Powell, Mount Royal, and Philipsburg Batholiths scatter between ca. 50.7 and 70.8 Ma and do not statistically allow assessment of pluton age differences. More geochronologic data from the eastern margin of the Philipsburg Batholith and from the Mount Powell Batholith are needed to better explain the young age determinations provided by samples PB17 and PB111.

Paleomagnetism

Most specimens demagnetized using AF methods yield coherent linear demagnetization trends (Fig. 8), allowing the calculation of characteristic remanent magnetizations (ChRMs) and estimates of site mean directions. Magnetizations isolated over the lowest range of peak fields, and thus carried by lowest coercivity materials, are similar to present-day field directions. For most samples from most sites, a north-northeast–directed, moderately positive inclination magnetization was isolated over a range of moderate coercivities, and is interpreted as a normal polarity ChRM for the batholith. Site means of the ChRM were estimated for 63 sites (Table 2). Six sites yielded south-seeking magnetizations of moderately negative inclination, and these are considered to be of reverse polarity. ChRMs of all accepted samples from one reverse polarity site were determined using great circle analysis. Five sites yielding reverse polarity magnetizations also had a subset (up to three) of samples that yielded normal polarity magnetizations, and these were inverted to reverse polarity data for the calculation of estimated site means. Thirty-one sites provided α95 values associated with estimated site means greater than 15°, and we used this confidence value to arbitrarily exclude these sites from further calculations. Five sites yielding reverse polarity ChRMs had α95 values greater than 15° and were excluded. The remaining 32 site mean directions are moderately well defined, with α95 values ranging from 4.7° to 14.9° and estimated precision parameter (k) values of 12–108. The dominance of normal polarity magnetizations in the Philipsburg Batholith is consistent with an emplacement age of slightly older than ca. 74.7 Ma. This approximate age lies within geomagnetic polarity Chron C33n, which was ∼5.5 m.y. in duration, between ca. 79.1 and 73.6 Ma (Cande and Kent, 1995). As further discussed later herein, the high number of rejected sites is due to the control of multidomain magnetite on the ChRM.

The 32 accepted site means were used to calculate a grand mean assuming a Fisherian distribution (Fisher, 1953). Four site mean directions were more than two angular standard deviations (59°) from the grand mean. These sites were omitted, and a grand mean for the remaining 28 sites (Fig. 9) was calculated (declination = 015.7°, inclination = 70.6°, k = 21.8, s = 17.3°, α95 = 5.7°).

Rock Magnetism and Fe-Oxide Observations

Isothermal remanent magnetization (IRM) acquisition response for all Philipsburg Batholith samples examined showed complete saturation by 150 mT, with most samples ∼80% saturated by 50 mT (Fig. 10). Backfield demagnetization of saturation IRM yielded coercivity of remanence (Hcr) values between 12 and 24 mT. The low fields required for saturation of IRM and the low Hcr values imply multidomain (MD) magnetite as the dominant magnetic phase in the Philipsburg Batholith. Modified Lowrie-Fuller tests (Johnson et al., 1975) show that for most samples, higher destructive fields were required to demagnetize 98 mT IRM than were required to demagnetize ARM, a response characteristic of dominance by MD magnetite (Fig. 11). The normalized NRM demagnetization curves are comparable to ARM demagnetization curves for about half of the specimens, indicating that ARM is a good proxy for low-field thermoremanent magnetization (TRM). The Philipsburg Batholith ChRM is assumed to be a TRM acquired in Earth's magnetic field during subsolidus cooling.

The median destructive field (MDF) is that required to demagnetize 50% of a magnetization. The ratio MDFARM/MDFSIRM plotted against bulk susceptibility of Philipsburg Batholith samples (Fig. 12) shows that Philipsburg Batholith samples are dominated by MD magnetite, with MDF ratios less than unity, over the entire range of bulk susceptibilities.

Saturation magnetization, Ms, ranges from 1.12 to 1.89 Am2/kg, saturation remanence (Mrs) ranges from 0.007 to 0.021 Am2/kg, coercivity (Hc) ranges from 0.6 to 1.5 mT, and coercivity of remanence (Hcr) ranges from 7.4 to 12.0 mT. Representative hysteresis loops for the Philipsburg Batholith (Fig. 13) show the ramp-shaped curves typical of dominance by MD magnetite grains (Dunlop and Özdemir, 1997). The Philipsburg Batholith samples have low Mrs/Ms values (0.014–0.004) and high Hcr/Hc values (16–7), which further suggest dominance by MD magnetite (Day et al., 1977).

Petrographic inspection shows that magnetite is the most abundant Fe-Ti–oxide phase. Representative reflected light photomicrographs (Fig. 4) show that magnetite grains are located within biotite, feldspar, and amphibole, as well as along grain boundaries. The approximate diameter of magnetite grains ranges from ∼4 to 200 μm, and magnetite grains often form clusters up to 2 mm. Most magnetite grains are equant, but some are about twice as wide along one axis in thin section, and some are irregularly shaped. Ilmenite lamellae in magnetite are rare. Trace amounts of pyrite are present.

AMS Fabrics in the Philipsburg Batholith

Average bulk susceptibility for all 119 sites ranges from 2 to 6 × 10−2 SI/volume (Fig. 14), indicating that the AMS signal is dominated by ferro/ferrimagnetic mineral fabrics, most likely defined by magnetite (Hrouda, 1982). The degree of anisotropy, P′, at the site level ranges from 1.02 to 1.33, with an average of 1.11 or 11% anisotropy (Table 3; Fig. 15). Hrouda (1982) suggested that magma flow usually produces P′ values <1.2, and because there is no evidence for subsolidus deformation in thin section, all Philipsburg Batholith AMS fabrics are assumed to be magmatic in character. P′ values are fairly uniform across the intrusion, although values seem to increase slightly from west to east. Map view contours of P′ (Fig. 16) reveal two areas of higher anisotropy in the SE and SW corners of the Philipsburg Batholith. For the shape parameter, T, sites with T > 0.25 are typically considered oblate, T < −0.25 are considered prolate, and −0.25 < T < 0.25 are considered triaxial. The value of T is positive for most sites in the batholith, indicating that oblate fabrics generally dominate over prolate fabrics. No relationship is apparent between T and P′ values (Table 3; Fig. 15).

Principal AMS directions for sample ellipsoids are generally well clustered at the site level (Table 3; Fig. 17), meaning that fabrics are consistent over the scale of individual outcrops. Foliation orientations for oblate and triaxial sites and lineation directions for triaxial and prolate sites (Fig. 18) allow assessment of magma flow. AMS foliations generally show shallow dips in the center and steeper dips near the eastern margin of the batholith, although there is considerable scatter in dip values across the map area. AMS foliations generally dip away from an area in the center of the Dora Thorn Pluton. Foliations in the Bimetallic Stock are, on average, steeper than those of the Dora Thorn Pluton. Lineation data are generally subparallel or oblique to the strike of foliations, with most rakes less than 50°. Lineations are rarely in the foliation dip direction. Strike-parallel lineations are especially common near the margins of the batholith.

The distribution of populations of AMS foliation dip values (Fig. 19) shows that the steepest (75°–90°) are concentrated in three areas near the estimated position of the Georgetown-Princeton thrust below the Philipsburg Batholith, and at several sites near the proposed contact between the Dora Thorn Pluton and the Bimetallic Stock. An area of shallow dips exists near the southeast margin of the batholith. The shallowest dip in this area is from one of the two sites yielding samples with relatively young 40Ar/39Ar dates.

AARM Fabrics in the Philipsburg Batholith

AARM ellipsoids from six sites in the Philipsburg Batholith have principal axes that are coaxial with AMS ellipsoid axes. One of the samples (PB12) has parallel axes, but the AARM2 axis and the K3 axis of the AMS ellipsoid are parallel. One site has an AARM fabric that is oblique to the AMS fabric. Two sites have prolate AARM ellipsoids, two are triaxial, and four sites have oblate AARM ellipsoids. All of the oblate sites have coaxial AMS and AARM ellipsoids. AARM foliations and lineations are similar to AMS fabrics (Table 4). Both AMS and AARM foliations have strikes that are roughly parallel to the margins of the Philipsburg Batholith for all sites located near the margins, except for site PB12, which has an AARM foliation that is perpendicular to the southeast margin. The AARM results from the eight measured sites generally fit the pattern of steeper, margin-parallel foliations near the batholith margins and shallower foliations in the interior of the batholith.

DISCUSSION

Timing of Emplacement and Sevier Fold-and-Thrust Belt Development

Assuming that the reported 40Ar/39Ar biotite dates older than ca. 74 Ma represent an essentially uniform rate of postemplacement cooling of the Philipsburg Batholith, their weighted mean age of 74.8 ± 0.1 Ma suggests that the Philipsburg Batholith was intruded over a duration of less than ∼0.5 m.y. These new age determinations are consistent with previous K-Ar results, and they do not rule out a division of the Philipsburg Batholith into two or more separate magma pulses, as proposed by Hyndman et al. (1982). The trend of slightly younger ages from west to east (Fig. 7) may still imply that the Bimetallic Stock was intruded before the Dora Thorn Pluton; however, the available data do not require making an age distinction. Although Philipsburg Batholith dates define a normal distribution at 74.77 ± 0.6 Ma, not all individual sample ages overlap within error. This suggests that the Philipsburg Batholith may represent an amalgamation of multiple intrusive events (Glazner et al., 2004), as supported by subtle variations in mafic mineral content throughout the batholith and by model cooling times for moderate sized plutons, which are considerably less than 2 m.y. (Glazner et al., 2004).

Although we have no direct age information on the Mt. Powell pluton, we suggest that the emplacement age of this pluton is younger than or equal to the youngest biotite (ca. 65.4 Ma) date from the Philipsburg Batholith. Based on argon diffusion kinetics of biotite, partial to complete argon loss requires temperatures in excess of 300 °C, but argon loss is also somewhat dependent upon the duration of the heating event. Simple conductive heat-flow calculations indicate that the Philipsburg Batholith would have been residing at ∼200 °C prior to emplacement of the Mt. Powell pluton in order for sampling localities in the eastern part of the Philipsburg Batholith (∼1.5 km away) to be heated to ∼300 °C. This relatively high ambient temperature places the Philipsburg Batholith at ∼6–8 km depth (depending upon choice of geothermal gradient) and is consistent with other depth estimates and the formation of a notably wide metamorphic contact aureole.

In principle, paleomagnetic data from large-volume, silicic to intermediate composition plutons characterized by a magnetic mineralogy of varied grain size and shape support the hypothesis that magnetization blocking takes place over a relatively wide range of temperatures below the maximum blocking temperature, and, thus, depending on pluton cooling rate, the data integrate geomagnetic field behavior over considerable periods of time. This feature of large-volume silicic plutons has been noted by several workers (e.g., Frei et al., 1984; Hagstrum et al., 1985; Frei, 1986; Ferranti et al., 2009), and the assumption has been made that relatively small volumes of a pluton can average geomagnetic field behavior (e.g., paleosecular variation) over considerable time periods. One consequence of this is that the within-site dispersion of paleomagnetic data may be comparable to between-site dispersion of results from an entire pluton. This is part of the reason that we have applied perhaps overly stringent acceptance criteria to our paleomagnetic data set and have such a high number of rejected sites. Another consequence is that data from selected parts of plutons may be used to evaluate the magnitude of absolute deformation or relative deformation in comparison to other parts of the pluton. To assess the possibility of internal deformation of the Philipsburg Batholith following the acquisition of remanent magnetization, we arbitrarily divided the batholith into six sections (Fig. 20). Four of these sections have a sufficient number of accepted site mean directions, and estimates of mean paleomagnetic directions for each of these, as well as the overall grand mean direction, are summarized in Table 5. The grand mean 95% confidence cone overlaps each confidence cone for the four section means, suggesting the absence of appreciable postemplacement deformation among different sections of the Philipsburg Batholith (Fig. 20).as the overall grand mean direction, are summarized in Table 5. The grand mean 95% confidence cone overlaps each confidence cone for the four section means, suggesting the absence of appreciable postemplacement deformation among different sections of the Philipsburg Batholith (Fig. 20).

We interpret the lack of field relations or paleomagnetic evidence supporting any internal, postemplacement deformation of the Philipsburg Batholith to suggest that Sevier-age deformation in the Flint Creek Range had ceased and thrust-belt development had shifted farther to the east by ca. 75 Ma. This result is consistent with previous work on the timing of folding and thrusting in the Helena Salient east of the Boulder Batholith, which generally started at ca. 80 Ma (Schmidt et al., 1990) and ended prior to ca. 50 Ma (Harlan et al., 1988). Notably, Harlan et al. (2008) demonstrated that the complete development of the Doherty Mountain fold complex, east of the Boulder Batholith, took place after 77 Ma. The Doherty Mountain fold complex formed in response to thrust displacement along the Jefferson Canyon–Lombard fault system.

To assess the possibility of whole-scale, postemplacement deformation involving rotation or tilt (or a combination of both) of the Philipsburg Batholith, the grand mean was compared to expected paleofield directions (Table 6) calculated from different paleomagnetic pole compilations for the Late Cretaceous (Besse and Courtillot, 2003; Gunderson and Sheriff, 1991; Van der Voo, 1993) for North America (Fig. 21). Rotation and flattening estimates, as well as tilt determinations (discussed later herein), are given in Table 6. Comparisons between the grand mean and each of the Late Cretaceous expected directions imply that the Philipsburg Batholith may have experienced some 23°–48° ± 19° of clockwise vertical-axis rotation and insignificant inclination shallowing since emplacement. Rotation and flattening error values were calculated following Beck (1980) with modification by Demarest (1983). Later herein, we explore whether this is a viable interpretation of the paleomagnetic data in the context of the structural history of the Philipsburg Batholith and host rocks.

All major structures, including the Georgetown-Princeton thrust, in the western Flint Creek Range are oriented roughly N-S, and thus roughly parallel with the overall regional structural grain in this part of western Montana (Schmidt et al., 1988; Lageson et al., 2001; Woodward, 1982). No strike-slip or oblique-slip motion faults of any orientation have been described in or around the Philipsburg Batholith. An inferred modest magnitude of clockwise vertical-axis rotation appears inconsistent with local geologic field relations. In addition, significant vertical-axis rotation is not supported by the paleomagnetic data reported by Eldredge and Van der Voo (1988) in their regional study of vertical-axis rotations in the Helena Salient and, more locally, by Harlan et al. (2008) in the Doherty Mountain fold complex east of the Boulder Batholith. A more plausible explanation for the discrepancy between observed paleomagnetic data for the Philipsburg Batholith and the expected latest Cretaceous paleofield directions involves crustal tilt around a horizontal axis. Utilizing a new error calculation method we describe here (Appendix), depending on the reference direction selected, the grand mean paleomagnetic direction for the batholith can be explained by a 9°–16° ± 8° west-side-down tilt around a horizontal tilt axis with an azimuth ranging between ∼351° and 015° (Fig. 21). A very modest magnitude of west-side-down tilt in the western Flint Creek Range could be explained by further crustal shortening related to eastward propagation of the Sevier thrust belt following emplacement of the Philipsburg Batholith. Tilting due to continued Sevier thrusting to the east must have occurred prior to the cessation of thrust and fold shortening at ca. 50 Ma (Harlan et al., 1988). Alternatively, the inferred modest west-down tilt of the western Flint Creek Range could have taken place subsequent to ca. 50 Ma due to flexural uplift of the footwall, including the Philipsburg Batholith, below the east-dipping Anaconda metamorphic core complex detachment (O'Neill et al., 2004). At present, we have no basis to prefer either mechanism to explain a modest amount of west-side-down tilting. Notably, both explanations are more consistent with the local and regional geology than a greater magnitude of vertical-axis rotation.

Lageson et al. (2001) suggested that intrusion of the ca. 70–80 Ma Boulder Batholith and coeval eruption of the Elkhorn Mountain volcanics into the active Late Cretaceous fold-and-thrust belt in SW Montana resulted in a supercritical orogenic wedge, which then led to increased shortening in the Helena Salient. The intrusion of the Philipsburg Batholith and other Late Cretaceous intrusions in the Flint Creek Range west of the Boulder Batholith may have also led to an overthickened orogenic wedge. Overthickening in the western part of the fold-and-thrust belt may have accentuated deformation to the east in the vicinity of the Boulder Batholith. Alternatively, overthickening of the orogenic wedge in the Flint Creek Range may have led to orogenic collapse and extension, resulting in formation of the Anaconda metamorphic core complex in the Eocene, as proposed by O'Neill et al. (2004).

Emplacement of the Philipsburg Batholith

The lack of field evidence indicating that the Georgetown-Princeton thrust offsets the Philipsburg Batholith implies that pluton emplacement followed thrust motion. The position of the fault prior to intrusion of the Philipsburg Batholith (Fig. 3) is critical in interpreting the relationship between the structures related to crustal shortening and magma emplacement. The Georgetown-Princeton thrust has an east-west trace near both the northern and southern margins of the Philipsburg Batholith, although deflection of the fault trace is more pronounced near the northern margin. We interpret the eastward deflection of the trace of the Georgetown-Princeton thrust near the Philipsburg Batholith to suggest that the footwall of the thrust was depressed during intrusion and that the magma filled space formed by both roof uplift and depression of the footwall.

The AMS fabric data obtained in this study are interpreted to reflect megascopic-scale silicate mineral fabrics acquired during supersolidus magma flow attending emplacement of the Philipsburg Batholith. High bulk susceptibilities typical of Philipsburg Batholith rocks indicate that the AMS fabrics are controlled by magnetite, rather than paramagnetic silicate minerals. All sites in the batholith have degrees of anisotropy (P′) typical of magma flow fabrics observed in other plutons (Hrouda, 1982). Because all rock magnetic data indicate that the pluton is dominated by multidomain magnetite, the observed AMS fabrics likely reflect shape anisotropy of elongate magnetite grains. Magnetite is typically concentrated in hornblende or biotite grains, or located on grain boundaries of Fe-Mg silicates. These spatial relations suggest that the magnetite-dominated AMS fabrics reflect preferred orientation of the silicate mineral network. A magmatic origin for AMS fabrics is corroborated by orientations of previously measured mafic enclaves in the southeast part of the Philipsburg Batholith, which are elongate parallel to the margin of the intrusion and roughly parallel to AMS lineations in the same area (O'Connell, 2000).

We interpret the AMS data to indicate that emplacement of the Philipsburg Batholith involved subhorizontal magma spreading from a limited number of locations. Most AMS lineations and foliations in the Philipsburg Batholith are subhorizontal, with average foliation dips that are shallow to the south-southwest. If foliation dips are corrected for ∼10° of estimated west-side-down tilt of the pluton, the average is even closer to horizontal. In the area around Fred Burr Lake in the center of the Dora Thorn Pluton, AMS foliations dip concentrically, with values less than 30°. Contours of foliation dips show an overall pattern of shallow values in the Dora Thorn Pluton, except near the estimated subsurface location of the Georgetown-Princeton thrust and near the eastern margin of the intrusion. Near intrusion margins, foliation dips range from shallow to steep, yet all foliation strikes are parallel to margins. Subhorizontal AMS lineations are roughly parallel to foliation strike and to batholith margins, and they are interpreted as stretching lineations due to magma flow impinging on wall-rock contacts. AMS fabrics reflecting magma interaction with the pluton margins are well illustrated in the southeast part of the Philipsburg Batholith, where all lineations trend parallel to the margin, are subhorizontal, and P′ values are very high compared to the rest of the intrusion. These well-developed fabrics are interpreted to have resulted from west-to-east magma flow impinging on the southeast margin of the pluton.

The Bimetallic Stock is characterized by much steeper foliation dips than the Dora Thorn Pluton. The steepest dips are along the inferred contact between the Bimetallic Stock and Dora Thorn Pluton, suggesting that the Bimetallic Stock was fed by magma conduits in the vicinity of the internal contact. The gradational internal contact between the Dora Thorn Pluton and the Bimetallic Stock is close to and parallel with the inferred subsurface location of the Bungalow fault, which we interpret as a back thrust associated with the Georgetown-Princeton thrust, implying that magma feeding the Bimetallic Stock used the Bungalow fault as a conduit during emplacement.

AMS data are interpreted to indicate that the Georgetown-Princeton thrust served as a conduit for rising magma feeding the Dora Thorn Pluton. Shallowly dipping foliations abruptly change to steeply dipping foliations above the estimated location of the Georgetown-Princeton thrust; two areas in the southeast and one area in the north of the Dora Thorn Pluton show near-vertical foliations (Fig. 19). In the same two areas of the southeast Dora Thorn Pluton, AMS lineations plunge steeper than 50° (Fig. 18B). One of the areas with near-vertical foliations and steep lineations lies between the main Georgetown-Princeton thrust and an eastern thrust splay. Magma may have risen along both these faults during emplacement of the Dora Thorn Pluton.

Steeply dipping foliations do not completely characterize the pluton above the inferred location of the Georgetown-Princeton thrust or the contact between the Bimetallic Stock and Dora Thorn Pluton. An area of steep dips in the Fred Burr Creek drainage in the south of the Bimetallic Stock is separated from steep dips near South Boulder Creek in the north of the Bimetallic Stock by an area of moderately dipping foliations. Near Fred Burr Creek, the modal percent hornblende is highest, and the hornblende has slightly coarser grain size than in the northern Bimetallic Stock. In the southern Dora Thorn Pluton, AMS foliation dips are shallow along the inferred subsurface location of the Georgetown-Princeton thrust. The limited areas of steep foliation dips along the internal intrusive contact and the Georgetown-Princeton thrust are interpreted to indicate that magma did not rise in a sheet along faults during emplacement. Instead, magma is interpreted to have flowed in conduits near the fault surfaces before spreading horizontally. We interpret the Georgetown-Princeton thrust in the vicinity of the Philipsburg Batholith as a thrust ramp. The thrust surface in this area rises to higher stratigraphic levels in the Belt Supergroup, as demonstrated by the lack of carbonate strata of the lower Belt Supergroup exposed east of the thrust system. The Philipsburg anticline, on the west side of the Philipsburg Batholith, is interpreted as a ramp-top anticline. A ramp-type geometry can explain the steep dips of the main Georgetown-Princeton thrust as well as the eastern splay, which we interpret as an imbricate thrust typical of ramp-flat settings. The Bungalow fault, again, would be considered a back thrust off the Georgetown-Princeton thrust.

Kalakay et al. (2001) interpreted the footwall synclines with fold axes lying east of and parallel to the margins of the McCartney Mountain Pluton and the Pioneer Batholith as fault-propagation structures beneath structurally higher hanging-wall anticlines. They used this inferred relationship to support their ramp-top emplacement model for the two intrusions. The Racetrack folds synclinorium, east of the Philipsburg Batholith, lies in a similar structural position to the footwall synclines discussed by Kalakay et al. (2001) and supports the ramp-flat geometry interpretation for the Georgetown-Princeton thrust near the Philipsburg Batholith.

We propose that emplacement of the Philipsburg Batholith occurred after final movement along the Georgetown-Princeton thrust system, facilitating magma transport to the level of emplacement via conduits along the thrust surface. AMS fabrics indicate that the magma utilized the Bungalow fault, the main Georgetown-Princeton thrust, and an eastern thrust splay to rise before subhorizontal lateral spreading atop the thrust ramp. The Philipsburg Batholith was emplaced as a tabular body fed from multiple sites of nearly vertical magma flow. The space required for emplacement was facilitated by roof uplift in the ramp-top setting due to previous fault-bend folding and back thrusting in the hanging wall of the Georgetown-Princeton thrust. Depression of the footwall may have aided emplacement in the ramp-top setting. This emplacement model is similar to the ramp-top emplacement model of Kalakay et al. (2001) (Fig. 1) for other intrusions in SW Montana, and AMS fabric data from the Philipsburg Batholith may support the application of a ramp-top emplacement model to other Late Cretaceous intrusions in SW Montana and other fold-and-thrust belts.

The difference in composition between the Bimetallic Stock and the Dora Thorn Pluton led Hyndman et al. (1982) to split the Philipsburg Batholith into two separate intrusions. They proposed that the two magmas differentiated at depth from a common source. The 40Ar/39Ar age data and AMS data indicate that magma of the Bimetallic Stock was transported vertically along the Bungalow fault and was emplaced within and adjacent to the Philipsburg anticline very shortly before the Dora Thorn magma rose along the Georgetown-Princeton thrust. The Dora Thorn magma spread laterally to the west of the Georgetown-Princeton thrust due to roof uplift, until it reached the still-cooling Bimetallic stock, resulting in a gradational contact. Dora Thorn magma also spread laterally to the east of the Georgetown-Princeton thrust, filling space caused by depression of the footwall. The high P′ values for AMS fabrics east of the thrust suggest a more forceful emplacement of the magma near the eastern margin of the Dora Thorn Pluton.

Comparison of AARM and AMS Fabrics

The high bulk susceptibilities of Philipsburg Batholith rocks suggest that their AMS fabrics are dominated by the preferred orientation of magnetite grains. We interpret the data from sites yielding similar AARM and AMS fabrics (Table 4) to confirm our assumption that multidomain magnetite controls the observed AMS fabrics. The nonparallel relationship between AARM and AMS fabrics for two of the eight sites examined in the Philipsburg Batholith is difficult to interpret and suggests that the AMS technique may not exactly reflect the preferred orientation of ferro/ferrimagnetic minerals at all sites. One possibility is that, although these samples are dominated by multidomain magnetite, a small fraction of single-domain magnetite may control the AMS fabrics. The different response of single-domain magnetite to induced and remanent magnetization methods could explain the difference between the AMS and AARM fabrics (Jackson, 1991). The observed differences among AARM and AMS fabrics from our select number of sites in the Philipsburg Batholith may call into question some of the underlying assumptions of the AMS technique as a proxy for magma flow fabrics of ferromagnetic-dominated granitoids. The assumptions that (1) high bulk susceptibility implies that magnetic fabrics are controlled by magnetite (Bouchez, 1997), and that (2) samples containing abundant multidomain magnetite have susceptibility fabrics that are also controlled by multidomain magnetite, rather than a small but important single-domain magnetite population, may not be valid in all cases. We assume that AMS fabrics, overall, in the Philipsburg Batholith reflect alignment of major silicate and oxide phases in the magma during emplacement. Our AARM results are based on a relatively small number of sites, and thus a robust assessment of the validity of these assumptions is precluded. Nonetheless, the similarity between AMS and AARM fabrics at six of eight measured sites gives us confidence that the assumptions are realistic. Further work on both AARM and AMS fabrics from silicic plutons should lead to a better understanding of the conditions under which the assumptions of the AMS technique are valid.

CONCLUSIONS

The emplacement and deformation history of a large-volume pluton in the Sevier fold-and-thrust belt of SW Montana, the Philipsburg Batholith, has been refined using 40Ar/39Ar age determinations on biotite separates, paleomagnetic data, field relations, and AMS/AARM fabric data. Paleomagnetic data indicate that the Philipsburg Batholith has not been internally deformed, but it may have experienced a small magnitude of west-side-down tilt since emplacement. A very modest amount of whole-scale, west-side-down tilting of the Flint Creek Range could have been accommodated by further crustal shortening related to eastward propagation of the Sevier thrust belt following emplacement of the Philipsburg Batholith or by flexural uplift after ca. 50 Ma of the footwall, including the Philipsburg Batholith, below the east-dipping Anaconda metamorphic core complex detachment. Sevier-age thin-skinned deformation in the western Flint Creek Range ceased before emplacement of the Philipsburg Batholith around 75 Ma. Emplacement of the Philipsburg Batholith may have occurred over a relatively short time span, possibly less than 1 m.y., and by amalgamation of multiple magma pulses. The spatial distribution of interpreted biotite cooling ages does not preclude the possibility that emplacement generally progressed from west to east. These results are consistent with the gradational contacts described between the more mafic Bimetallic Stock and the less mafic Dora Thorn Pluton. AMS fabrics, interpreted to reflect magma flow fabrics during emplacement, show that the Philipsburg Batholith magma utilized the Georgetown-Princeton thrust system as a conduit for magma ascent. The batholith is dominated by subhorizontal AMS fabrics, and we interpret it to have been intruded as a tabular body, with magma filling space above a ramp in the Georgetown-Princeton thrust within a fault-bend anticline, aided by back thrusting on the Bungalow fault. The ramp-top emplacement of the Philipsburg Batholith is consistent with the similar thrust ramp emplacement models proposed for other intrusions in SW Montana by Kalakay et al. (2001) using field relations, suggesting that this may be the dominant emplacement mechanism for Late Cretaceous intrusions in the fold-and-thrust belt of SW Montana.

Our study supports the hypothesis that space required for the ascent and shallow-level emplacement of large-volume magma bodies within regimes of crustal shortening is facilitated by step-overs or irregularities in the geometry of fault planes. These are sites of diminished maximum principal stress as well as differential stress, where magma can rise buoyantly. Magma flow along fault surfaces, in particular at shallow crustal levels, is more likely than ascent involving diapirism and/or stoping, especially in regions of horizontal shortening, because magma could rise as multiple, small-volume batches. Releasing bends in strike-slip systems have previously been invoked as sites of reduced compressive stress to facilitate pluton emplacement (e.g., in the Sierra Nevada Batholith; Titus et al., 2005). We expect magma batches rising along thrust faults will coalesce to form plutons at sites of diminished compressive stress, such as within fault-bend folds above thrust ramps, as shown for the Philipsburg Batholith.

Finally, we have developed a new method for calculating horizontal-axis tilt errors from paleomagnetic data (Appendix). This tilt error calculation method is useful for comparing measured paleomagnetic directions with expected directions calculated using published paleomagnetic poles, and it is more accurate for tilt calculations than the formula for vertical-axis rotation proposed by Beck (1980) and modified by Demarest (1983), or that suggested by McWilliams (1984).

APPENDIX. CALCULATING HORIZONTAL-AXIS TILT AND ASSOCIATED UNCERTAINTY

A horizontal tilt axis is the vector normal to a vertical plane that includes a small circle containing the end points of both the observed and expected paleomagnetic vectors (Fig. 22A). The tilt axis can be determined using the declination and inclination of the expected direction (Dx,Ix) and the observed direction (Dy,Iy) by first calculating the total declination difference between the two directions: 
graphic
Angles Da and Db are the angles in the horizontal plane (plane e, Figs. 22A and 22E) between Dr and Dx and Dr and Dy, respectively, where Dr is the direction of the horizontal tilt axis. The distance from the origin to the intersection with the vertical plane (plane f, Figs. 22A and 22F) along the direction of the tilt axis is ir. Equations A4 and A5 can be rearranged to solve for cos(Ir) and set equal to each other in order to solve for Da and Db, using the relationship: 
graphic
 
graphic
 
graphic
 
graphic
Equation 9 can be further simplified using the trigonometric identity for the cosine of the difference of two angles: 
graphic
 
graphic
 
graphic
 
graphic
 
graphic
Again, Da is the declination angle between the expected direction and the tilt axis, so the azimuth of the tilt axis, Dr, is: 
graphic
To calculate the magnitude of tilt about the horizontal tilt axis, we solve for the inclination, Ir, of the vector from the origin to the low point on the small circle using Equation A4 and the value calculated for Da in Equation A14. It follows from Figures 22D and 22E that: 
graphic
 
graphic
 
graphic
and similarly, 
graphic
 
graphic
The estimate of total tilt, T, between the expected and observed directions is: 
graphic
The error in the estimate of total tilt can be calculated by considering how the α95 confidence values for the expected and observed directions project onto the small circle around the tilt axis, similar to calculating ΔD for the expected and observed values when finding ΔR for vertical-axis rotation. The angular errors in the small circle of tilt, Δtx and Δty for expected and observed directions, respectively, are the projections of the α95 value on the small circle and can be found by scaling α95x and α95y by the radius, hr, of the small circle of tilt. When the observed and expected directions are of shallow inclination, and thus lie in a small circle close to the edge of the unit sphere, hr is small, and the projection of α95 on the small circle, Δt, is much larger than α95 (Fig. 23B). Conversely, when the small circle of tilt, defined by observed and expected directions of steep inclinations, is farther from the edge of the unit sphere, hr is large, and the projection of α95 on the small circle, Δt, is closer to α95 (Fig. 23C). In the case where tilt occurs along a great circle, hr = 1 and Δt = α95. Equation 16 states that hr = sin(Ir), so the individual tilt errors can be calculated by: 
graphic
 
graphic
The error in tilt, ΔT, can then be found by: 
graphic

We thank Mike Jackson for help with hysteresis measurements at the University of Minnesota Institute of Rock Magnetism. The 40Ar/39Ar geochronology was performed with the assistance of Bill McIntosh and Lisa Peters, for which we are very thankful. We are grateful to Jack Grow and Christina Carr for field assistance during the summer of 2007. This project was funded by grants to T. Naibert from the Geological Society of America, and the Office of Graduate Studies and Department of Earth and Planetary Sciences at the University of New Mexico. The efforts of two anonymous reviewers are greatly appreciated.

1GSA Data Repository Item 2010210, Table DR1, Ar/Ar biotite step-heating data, all samples, Philipsburg Batholith, is available at www.geosociety.org/pubs/ft2010.htm, or on request from editing@geosociety.org, Documents Secretary, GSA, P.O. Box 9140, Boulder, CO 80301-9140, USA.