Abstract

Paulsens is a mesothermal orogenic gold deposit located in the Wyloo Inlier on the southern margin of the Pilbara craton of Western Australia. Gold occurs in quartz-sulfide veins hosted within a folded and faulted gabbro dike, from which baddeleyite yields a U-Pb crystallization age of 2701 ± 11 Ma. Monazite and xenotime in the veins and from hydrothermally altered country rocks yield three distinct U-Pb dates of ca. 2400, 1730, and 1680 Ma. Textural relationships between euhedral xenotime and pyrite with rounded native gold inclusions from within the quartz-sulfide veins show that the primary gold mineralization was synchronous with xenotime crystallization at 2403 ± 5 Ma, and coeval with pervasive alteration of the host rocks, which yield monazite ages of 2398 ± 37 and 2403 ± 38 Ma. Regional-scale hydrothermal events at ca. 1730 and 1680 Ma are linked to the growth of monazite within phyllitic rocks at 1730 ± 28 and 1721 ± 32 Ma, carbonate veining at 1655 ± 37 Ma, and gold remobilization or introduction of new gold at 1680 ± 9 Ma. The ca. 2400 Ma age for mineralization and hydrothermal alteration does not correspond with any known deformation event in the region, indicating a significantly different and more complicated low-temperature tectonothermal evolution for the southern Pilbara region than previously recognized. The in situ secondary ion mass spectrometry dating of monazite and xenotime employed here will lead to better targeting of orogenic gold deposits in the northern Capricorn Orogen, and these techniques can be utilized for orogenic gold exploration worldwide.

Introduction

Exploration targeting of gold deposits can be significantly improved by understanding metallogenic events in both space and time (Hronsky et al., 2012). By knowing the ages of hydrothermal mineralization, host rocks, and regional tectonothermal events, the search space can be minimized, and the financial risk to explorers greatly reduced (Rasmussen et al., 2006; Hronsky and Groves, 2008). However, many chronometers either are scarce in orogenic gold deposits or are susceptible to isotopic resetting during subsequent metamorphism and deformation (Kerrich and Cassidy, 1994; Chesley, 1999), so that the ages of many gold deposits worldwide are poorly constrained (Chesley, 1999). Nevertheless, many orogenic gold deposits contain trace amounts of the rare earth element (REE)-bearing phosphate minerals monazite ((Ce,La,Nd,Th)PO4) or xenotime (YPO4) intergrown with ore minerals (Vielreicher et al., 2003; Carpenter et al., 2005; Zhang et al., 2014). These phosphate minerals are robust chronometers that are resistant to diffusive Pb loss at temperatures up to 750°C (Harrison et al., 2002; Cherniak, 2010). Instead, monazite and xenotime undergo dissolution and reprecipitation reactions at temperatures ¼400°C (Townsend et al., 2000; Rasmussen and Muhling, 2007), leading to crystals that have multiple, but discrete, age domains from which precise dates can be obtained.

The Paulsens gold deposit is located in the northern part of the Capricorn Orogen, on the southern margin of the Pilbara craton in Western Australia, and is hosted in low-grade metasedimentary and metavolcanic rocks of the ca. 2775 to 2629 Ma Fortescue Group in the Wyloo Inlier (Thorne and Trendall, 2001; Thorne et al., 2011; Fig. 1). Paulsens has an endowment of 1,114,000 ounces of gold, comprising 854,000 ounces mined between 2005 and 2016 (Northern Star Resources Limited, 2015a, 2016), and a remaining resource of 943,000 t at 8.58 g/t of gold for a total of 260,000 ounces contained gold as of June 30, 2016 (Northern Star Resources Limited, 2016). Historically, Paulsens was known as the Melrose mine, which was active in the 1930s with reports of 916 ounces of gold recovered from 2,955 t of ore (Forman, 1938; Blight, 1985; Northern Star Resources Limited, 2015b). The geological evolution of the area around Paulsens is not well known, and information is restricted to exploration company reports and short summaries in reports of the Geological Survey of Western Australia. Gold mineralization is contained within auriferous quartz-sulfide veins hosted in a folded and faulted gabbro dike (Fielding and Stokes, 2014; Northern Star Resources Limited, 2015b).

In this study, a combination of field mapping, petrography, and multimineral U-Th-Pb sensitive high-resolution ion microprobe (SHRIMP) geochronology has been used to address the absolute timing of mineralization at Paulsens and its relationship to known tectonothermal events in the region. Samples were taken from the Paulsens East and Gabbro Offset prospects that are located adjacent to the main Paulsens deposit (Fig. 2). These prospects share many similarities with Paulsens, but have not been subject to the same degree of deformation, thereby making it easier to establish relationships between deformation and ore formation. Results from this study provide U-Pb ages for (1) the emplacement of the gabbroic host rocks and maximum depositional ages for the Hardey Formation and (2) the timing of punctuated hydrothermal alteration and gold mineralization.

Geologic Setting

The Proterozoic Capricorn Orogen in Western Australia is a major zone of deformation, metamorphism, and magmatism located between the Yilgarn and Pilbara cratons (Tyler and Thorne, 1990; Cawood and Tyler, 2004; Sheppard et al., 2010; Thorne et al., 2011). It is the product of at least seven tectonic events, with intracratonic reworking and basin formation spanning more than 1.6 billion years (Martin and Morris, 2010; Sheppard et al., 2010; Johnson et al., 2011, 2013). Tectonothermal events include the 2215 to 2145 Ma Ophthalmia Orogeny, which is thought to reflect collision of the Pilbara craton with the Glenburgh terrane (Occhipinti et al., 2004; Johnson et al., 2011), and the 2005 to 1950 Ma Glenburgh Orogeny, which records collision of the Yilgarn craton with the combined Pilbara craton-Glenburgh terrane and marks the assembly of the West Australian craton (Johnson et al.,2011). This was followed by intracontinental reworking during the 1820 to 1770 Ma Capricorn Orogeny (Cawood and Tyler, 2004; Sheppard et al., 2010), the 1680 to 1620 Ma Mangaroon Orogeny (Sheppard et al., 2005), the 1320 to 1170 Ma Mutherbukin Tectonic Event (Korhonen et al., 2017), the 1030 to 955 Ma Edmundian Orogeny (Martin and Thorne, 2004; Sheppard et al., 2007), and the ca. 570 Ma Mulka Tectonic Event (Johnson et al., 2013). The northern part of the Capricorn Orogen includes Archean rocks of the Pilbara craton, which are overlain by Archean to Paleoproterozoic rocks of, in ascending order, the Fortescue Group, Hamersley Group, Turee Creek Group, Shingle Creek Group (formerly the lower Wyloo Group), Wyloo Group (formerly the upper Wyloo Group), and Capricorn Group (Fig. 3A; Thorne and Trendall, 2001).

The Fortescue Group was deposited during protracted rifting of the Pilbara craton between ca. 2775 and 2629 Ma. In the South Pilbara sub-basin, deposition was controlled by E-to SE-trending extensional faults (Blake, 1993; Thorne and Trendall, 2001; Trendall et al., 2004; Thorne et al., 2011). Here the group is composed of a 6.5-km-thick succession of metasedimentary and metavolcanic rocks deposited unconformably on granite-greenstone rocks of the Pilbara craton and comprises, in ascending order, the Bellary Formation, Mt. Roe Basalt, Hardey Formation, Boongal Formation, Pyradie Formation, Bunjinah Formation, and Jeerinah Formation (Thorne and Trendall, 2001; Thorne et al., 2011; Fig. 3B). Thorne and Trendall (2001) divided the stratigraphy into five major tectonostratigraphic sequences with different depositional settings. Sequences 1 and 2 were deposited in fault-bounded subbasins in a coastal to shallow-marine environment, and comprise, respectively, sedimentary rocks, basaltic lavas, and volcaniclastic rocks of the Bellary Formation and Mt. Roe Basalt, and mixed sedimentary and volcanic rocks of the Hardey Formation. Merging of the subbasins, subsidence, and regional tilting resulted in deposition of sequences 3 and 4, which consist of subaqueous basaltic to komatiitic lavas of the Boongal Formation and the Bunjinah and Pyradie formations, respectively, which were deposited in a deep shelf setting. Further subsidence resulted in a major marine transgression and the deposition of sequence 5, consisting of interbedded deep-marine sedimentary and volcaniclastic rocks, and basaltic lavas of the Jeerinah Formation.

A deep-crustal seismic reflection survey across the Capricorn Orogen imaged the crustal architecture of the orogen and the cratonic margins, identifying at least five major mantle-tapping structures (Johnson et al., 2013). In the northern Capricorn Orogen, the Nanjilgardy, Baring Downs, and Blair faults are of particular interest because they are spatially associated with gold and base metal occurrences, including the Paulsens and Mount Olympus deposits (Johnson et al., 2013; Fig. 1). Deformation, metamorphism, and hydrothermal activity in the Wyloo Inlier have been interpreted to be related to either the 2215 to 2145 Ma Ophthalmia Orogeny (Rasmussen et al., 2005; Martin and Morris, 2010) or the 1820 to 1770 Ma Capricorn Orogeny (Thorne and Trendall, 2001) or both, although no geochronological data are available from around Paulsens to verify these suggestions.

Local Geology

The Paulsens gold mine is situated at the northwestern end of the Wyloo Inlier, within metasedimentary and metavolcanic rocks of the Hardey Formation, near the base of the Fortescue Group (Owen, 2000; Thorne and Trendall, 2001). The mine geologists at Paulsens divided the Hardey Formation into five informal members—from oldest to youngest, the Horsewell sandstone, Melrose argillite, Madang tuff, Tin Hut basalt, and Beaghy sandstone (Fig. 2; Owen, 2000). These strata are cut at a low angle by a ~50-m-thick, folded and faulted, medium- to coarse-grained mafic dike, known as the Paulsens gabbro, which, over short distances, follows the contact between fine-grained sandstone and laminated carbonaceous shale of the Melrose argillite (Fielding and Stokes, 2014). Regional-scale epidote-actinolite greenschist facies metamorphism has affected the rocks throughout the Wyloo Inlier (White et al., 2014a).

Structural observations across the Wyloo Inlier, as well as from underground drives and diamond drill core, show that the rocks were subjected to at least three low-grade deformation events (Owen, 2000). The first two events were responsible for the production of two dominant regional-scale cleavages. The S1 cleavage is axially planar to regional-scale, tight to upright folds that define the long axis of the Wyloo Inlier. The S1 fabric is a penetrative, spaced cleavage with an average orientation of 140/85SW, and is tentatively correlated with the D3oph event of the Ophthalmia Orogeny (Tyler and Thorne, 1990; Tyler, 1991). The D2 event is also marked by a spaced cleavage (S2) that is subparallel to the regional S1 fabric, with an average orientation of 125/85SW. D2 caused the tightening of F1 folds and attenuation of the northern limbs to form narrow, intense shear zones up to 50 m wide. The D2 event may be equivalent to the D2ash event of the Capricorn Orogeny, during which preexisting structures were either reactivated or tightened (Thorne and Seymour, 1991). Due to the near-coaxial nature of the two cleavages, distinguishing between S1 and S2 in the field can be difficult. The third event (D3) is characterized by strike-slip faulting, possibly through the reactivation of preexisting faults, such as the Hardey fault to the northeast of the Paulsens mine (Fig. 2). The Hardey fault is a splay of the crustal-scale Nanjilgardy fault, which may have been a conduit for both hydrothermal and mineralized fluids during reactivation over multiple deformation events (Johnson et al., 2013).

Mineralization and Alteration

The Paulsens orebody is hosted by a 40-m-thick auriferous quartz-sulfide vein within the Paulsens gabbro, where the gabbro crosscuts fine-grained carbonaceous sandstone and siltstone of the Melrose argillite. Rheological contrasts between the gabbro and surrounding sedimentary rocks resulted in brittle fracturing of the gabbro during regional-scale F1 folding, allowing for deposition of the auriferous quartz-sulfide vein (Fielding and Stokes, 2014; Fig. 4). Mineralization is located at the margins of the vein and is referred to as Paulsens Upper zone and Paulsens Lower zone mineralization (Northern Star Resources Limited, 2015b). Upper zone mineralization is defined by brecciated massive pyrite (Fig. 5A; Owen, 2000). Gold is rarely observed in hand specimen in the Upper zone, but petrographic studies show that early gold forms rounded inclusions within pyrite crystals and displays simple monocrystalline twinned microstructures with silver contents of 8.0 to 8.5 wt %. A second phase of gold formed along fractures and at grain boundaries where the pyrite is brecciated. This style of gold is typically associated with chalcopyrite and pyrrhotite, and has a lower silver content, ranging from 6.6 to 7.2 wt %, with simple monocrystalline microstructure and some twin planes (Hancock and Thorne, 2016). Lower zone mineralization is hosted within the finely laminated basal part of the quartz-sulfide vein, which contains abundant carbonaceous stylolites and highly altered wall-rock inclusions of argillite that are pervasively altered to muscovite-ankerite ± chlorite (Fig. 5B). The stylolites and wall-rock inclusions are associated with abundant free gold (Fielding and Stokes, 2014), with silver contents between 6.8 and 7.6 wt % and simple polycrystalline microstructure with polysynthetic and incoherent twins similar to the second style of gold in the Upper zone mineralization (Hancock and Thorne, 2016).

Proximal to the Paulsens mineralization, regional-scale greenschist facies metamorphic assemblages are overprinted by a hydrothermal alteration assemblage that is associated with gold mineralization. Primary mineral assemblages within the Paulsens gabbro are almost entirely replaced, with plagioclase altered to muscovite, and pyroxene replaced mainly by ankerite but with patches of muscovite-quartz-ankerite. Iron oxide minerals, possibly originally ilmenite and titanomagnetite, have been totally altered to leucoxene. Sedimentary rocks of the Melrose argillite member are now composed almost entirely of muscovite with minor quartz-ankerite ± chlorite. Alteration distal to the orebody is difficult to differentiate from the regional metamorphic background; however, a change from ankerite to calcite along with an increase in

chlorite and decrease in muscovite content occurs progressively away from the orebody.

Gabbro Offset and Paulsens East are two smaller prospects adjacent to the main Paulsens orebody (Fig. 2). They share many characteristics of Paulsens, but have not been subject to the same degree of deformation, which allows more straightforward determination of timing relationships between ore formation and the regional geologic history. Gabbro Offset is located 350 m southwest of the main Paulsens orebody and consists of a series of parallel quartz-carbonate-sulfide veins within the Paulsens gabbro. Mineralized veins are typically narrow, usually between 0.1 and 10 m wide, with visible gold along their margins or within massive pyrite similar to Paulsens Upper zone mineralization. Petrographic examination of pyrite from samples (GSWA 209907) from Gabbro Offset reveals two styles of native gold inclusions within the pyrite. Pyrite crystals show subtle zoning with an inclusion-rich core surrounded by a thin rim (<1 mm) of solid pyrite. Rounded blebs of native gold and chalcopyrite within the cores are interpreted to be primary inclusions (Fig. 6A). These inclusions are up to 200 μm in diameter and form the majority of visible gold in the samples. The pyrite is locally brecciated, and chalcopyrite, pyrrhotite, and gold occur along fractures and pyrite grain boundaries (Fig. 6B). Although there is some evidence for gold remobilization (from the rounded gold inclusions in the cores of pyrites) at a local scale (Fig. 6C), it is possible that this secondary gold represents the introduction of new gold from an additional hydrothermal event.

Paulsens East mineralization is located approximately 1 km east of the main Paulsens orebody (Fig. 2) and, historically, gold was extracted from a series of shallow workings. Mineralization is hosted within parallel, steeply NE dipping, 1- to 5-m-wide veins comprising quartz, carbonate, and oxidized sulfides that are hosted within the Melrose argillite on the northern side of the Hardey fault. Sedimentary rocks in contact with the mineralized veins at Paulsens East have been deformed and pervasively altered to an assemblage containing muscovite-quartz ± ankerite ± chlorite ± monazite (Fig. 6D), similar to that of the main Paulsens orebody.

Style of Gold Mineralization

The Paulsens deposit exhibits many of the characteristics of a typical orogenic gold deposit. Orogenic gold deposits form in the upper to middle crust in compressional tectonic settings related to accretionary or collisional orogenesis and tend to have a close spatial relationship to transcrustal structures which mark the boundaries between continental blocks (Groves et al., 1998; Goldfarb et al., 2001; Hronsky et al., 2012). Host rocks are varied, with Archean deposits commonly formed in volcanic-dominated sequences and Paleoproterozoic and Phanerozoic deposits formed in siliciclastic sequences (Goldfarb and Groves, 2015). Most deposits are hosted by rocks that have been metamorphosed to greenschist facies and commonly have alteration assemblages of carbonate-iron sulfide ± white mica ± chlorite (Goldfarb et al., 2001). Gold mineralization is characteristically high grade (5–30 g/t Au) and associated with quartz ± carbonate veins with ≤3 to 5% sulfides (Groves et al., 1998). Unlike other styles of mineralization (e.g., epithermal and intrusion related), orogenic gold deposits cannot be related to individual intrusions and do not display zoned alteration halos (Sillitoe, 1997; Goldfarb and Groves, 2015). Paulsens has many of these traits: a spatial relationship to the mantle-tapping Nanjilgardy fault, siliciclastic sedimentary host rocks that have been metamorphosed to greenschist facies, the high-grade nature of gold mineralization (average 8.59 g/t Au), gold hosted in quartz-carbonate-sulfide veins surrounded by muscovite-ankerite ± chlorite alteration, and an absence of coeval igneous intrusions.

Samples and Analytical Methods

Samples from Gabbro Offset were collected from three drill cores supplied by Northern Star Resources (PDU2153, PLDD015W1, and PDU2217); samples from Paulsens East were collected from outcrops where relationships between mineralization and deformation could be ascertained. A list of dated samples is provided in Table 1.

Monazite, xenotime, zircon, and baddeleyite were analyzed for U, Th, and Pb isotopes using the SHRIMP II instrument at the John de Laeter Centre at Curtin University in Perth, Western Australia. Appendix 1 describes the methodology in detail.

Detrital zircons were separated from sedimentary host rocks using standard magnetic and density techniques. Zircons, together with zircon reference materials (BR266 and OGC), were cast in 25-mm epoxy mounts and polished to expose the interiors of the crystals. Each mount was gold coated and characterized using transmitted light, reflected light, and cathodoluminescence (CL) images (Fig. 7).

U-Th-Pb analysis of monazite, xenotime, and baddeleyite was conducted in situ in order to preserve textural context. Grains for analysis were identified using a scanning electron microscope (SEM) in polished thin sections, and 2- or 3-mm-diameter plugs were extracted from the polished thin sections with a hollow-core rotary drill and mounted in 25-mm-diameter epoxy discs. The mounts were cleaned and gold coated before each analytical session. Reference materials were set into separate mounts and gold coated simultaneously with sample mounts. Standard and sample mounts were loaded together into the SHRIMP for concurrent analysis during each analytical session.

Geochronology Results and Interpretation

For all samples, age estimates are derived from 204-corrected 207Pb*/206Pb* ratios. Results for individual analyses are quoted with 1σ uncertainties; weighted mean 207Pb*/206Pb* dates are quoted with 95% confidence intervals.

Detrital zircon

Detrital zircons from two Hardey Formation sandstone samples (Table 1) were dated to provide a maximum depositional age for the supracrustal rocks into which the Paulsens gabbro was emplaced.

GSWA 209903—quartz sandstone, Hardey Formation: Medium-grained, strongly cleaved, quartz sandstone of the Melrose argillite member of the Hardey Formation was collected from 109.3- to 109.7-m depth in drill core PDU2153 in the hanging wall of the Gabbro Offset deposit.

Zircons are generally colorless, up to 150 μm in length, and euhedral, although rounded terminations and pitting are

indicative of detrital transport (Fig. 7A). Sixty-five analyses were made of 64 zircons, of which two analyses >5% discordant were excluded from the age analysis. The zircons range in age from ca. 3456 to 2923 Ma (n = 63). The youngest detrital zircon age component consists of 15 analyses that yield a weighted mean 207pb*/206pb* date of 2941 ± 5 Ma (MSWD =1.4; Table 2, Fig. 8A, C), providing a maximum age for deposition of the quartz sandstone.

GSWA 209911—quartz sandstone, Hardey Formation:Medium- and fine-grained quartz sandstone, with very fine grained matrix, from the Melrose argillite member was collected from the hanging wall to the Gabbro Offset deposit from drill core PLDD015W1 between 473.6- and 473.9-m depth.

Zircons are euhedral, up to 150 μm long, and mainly colorless, and have rounded terminations and pitting characteristic of detrital transport (Fig. 7B). Sixty-two analyses were made on 62 grains, of which 20 analyses >5% discordant were excluded from the age analysis. The zircons range in age from ca. 3450 to 2700 Ma. The youngest detrital zircon age component, consisting of nine analyses, yields a weighted mean 207Pb*/206Pb* date of 2750 ± 10 Ma (MSWD = 1.5; Table 3, Fig. 8B, D), providing a maximum age for deposition of the sandstone.

Baddeleyite

GSWA 209905—altered dolerite, Paulsens gabbro: A finegrained, dark-colored portion of the Paulsens gabbro that is host to the Gabbro Offset deposit was collected from 346.0- to 346.2-m depth in drill core PDU2153 (Table 1). At this locality, the gabbro is highly altered and primary plagioclase and pyroxene are replaced by muscovite and ankerite ± muscovite ± chlorite, respectively, and iron oxide minerals have been altered to leucoxene. However, euhedral baddeleyite crystals up to 100 μm long are preserved throughout the sample. The outer surfaces of baddeleyite grains are commonly altered to very fine grained aggregates of zircon (Fig. 9A).

Seventeen analyses were made of 11 baddeleyite grains. Two analyses with elevated common Pb and two analyses >5% discordant were excluded from the age analysis. A single analysis yielded a concordant 207Pb*/206Pb* date significantly younger than the main cluster, and is interpreted to reflect ancient Pb loss during alteration. The remaining 12 analyses yield a weighted mean 207Pb*/206Pb* date of 2701 ± 11 Ma (MSWD = 1.4; Table 4, Fig. 10A), interpreted as the time of igneous crystallization of the Paulsens gabbro, the host rock to gold mineralization.

Xenotime

GSWA 209907—auriferous quartz-sulfide vein: In the lower part of drill core PDU2153 between 394.4 and 395.5 m, a series of 1-cm to 1- to 2-m-wide massive sulfide and quartz-sulfide veins are abundant. These veins are part of the Gabbro Offset deposit and share similarities to Paulsens Upper zone mineralization. A massive sulfide vein was sampled for geochronology (Table 1), the lower 10 cm of which contained xenotime. The sample is associated with gold grades of 21 ppm, with gold formed as rounded inclusions within pyrite crystals (Fig. 6A) and along grain boundaries and fractures in areas where the pyrite is brecciated (Fig. 6B). The sample comprises an interlocking assemblage of quartz and pyrite, and contains xenotime in a number of textural settings: as euhedral grains enclosed entirely within pyrite (Fig. 9C), as subhedral grains within the quartz-pyrite matrix commonly intergrown along the margins of brecciated pyrite (Fig. 9D), and as euhedral to subhedral grains within the quartz matrix. Commonly, the matrix xenotime grains are chemically zoned, showing core and rim growth structures (Fig. 9D).

Eighty analyses were made on 37 grains, with multiple analyses performed within or across single grains. Irrespective of age or site of analysis, uranium and thorium contents range from 104 to 928 ppm and 2 to 695 ppm, respectively. Six analyses >5% discordant and 12 analyses that represent mixed ages, or parts of grains that have undergone Pb loss, are not considered geologically significant and are excluded from the age analysis. Twenty-six analyses of euhedral xenotime grains entirely enclosed within pyrite grains or as core zones of matrix xenotime present along brecciated pyrite grain boundaries (Fig. 9D) yield a weighted mean 207Pb*/206Pb* date of 2403 ± 5 Ma (MSWD = 1.1; Table 5, Fig. 10B). Thirty-six analyses of xenotime rims and discrete grains within the quartz matrix yield a weighted mean 207Pb*/206Pb* date of 1680 ± 9 Ma (MSWD = 1.4). We interpret these two dates to represent times of xenotime growth and subsequent dissolution- reprecipitation during two discrete periods of hydrothermal activity. The first event at ca. 2403 Ma is interpreted to date the formation of the quartz-sulfide vein, because grains of this age are contained entirely within the pyrite crystals. The second date at ca. 1680 Ma is interpreted to represent a subsequent period of hydrothermal activity, responsible for the brecciation and fracturing of pyrite grains and the dissolution-reprecipitation of preexisting ca. 2403 Ma xenotime.

GSWA 209909bankerite-quartz vein: Minor, 5- to 10-cm-wide carbonate veins within the fine-grained portion of the Paulsens gabbro (GSWA 209905) at the Gabbro Offset deposit were sampled from drill core PDU2153 between 345.1 and 345.3 m (Table 1). Assays from depths between 345 and 346 m returned up to 48 ppm Au. Xenotime occurs in the carbonate vein as irregular grains up to 50 μm across, interstitial to and included within other minerals.

Seven analyses were made on seven xenotime grains. Uranium and thorium contents vary from 200 to 550 ppm and 250 to 1,350 ppm, respectively. Two analyses >5% discordant were not used in the age analysis. One grain is significantly older than the others, with a date of 2415 ± 11 Ma (1σ). There are no textural or chemical differences between this grain and the others analyzed. The remaining four analyses of four grains yield a weighted mean 207Pb*/206Pb* date of 1655 ± 37 Ma (MSWD = 1.3; Table 6, Fig. 10C). The older xenotime date corresponds with the age of monazite in the altered host gabbro (GSWA 209909a) and is interpreted to be inherited from the hydrothermally altered host rocks. The younger age of xenotime growth at ca. 1655 Ma, obtained from multiple grains, is attributed to a second, discrete hydrothermal event, probably dating emplacement of the ankerite-quartz veins, and is within uncertainty of xenotime rims in the auriferous quartz-sulfide vein (GSWA 209907).

Monazite

GSWA 209902—carbonaceous phyllite, Hardey Formation:Very fine grained, black, carbonaceous, quartz-muscovite-chlorite phyllite of the Melrose argillite member was sampled from drill core PDU2153 between 164.2 and 164.3 m in the hanging wall to the Gabbro Offset deposit (Table 1). The phyllite has a very strong foliation that has disrupted and deformed 3- to 20-mm-thick quartz ± pyrite veins. The sample contains randomly distributed, 1-mm-long, elongate porphyroblasts consisting of intergrown monazite and apatite, in equal amounts, with lesser chlorite and quartz. It is possible that the apatite has replaced monazite, a common reaction during greenschist to amphibolite facies metamorphism and hydrothermal alteration (Finger et al., 1998; Rasmussen and Muhling, 2009; White et al., 2014b). However, it is not clear whether the monazites are primary metamorphic or hydrothermal porphyroblasts or whether they replaced former regional-scale peak-metamorphic porphyroblasts such as andalusite (?) that grew at the same time as the foliation.

Monazite within two porphyroblasts was analyzed, with five spots on each. Uranium and thorium contents are 200 to 300 ppm and 2,000 to 10,000 ppm, respectively. Of the 10 analyses, one analysis is >5% discordant, and one is significantly younger than other analyses from the same grain and probably reflects partial Pb loss. These two analyses are not considered to be geologically significant and were not included in the age analysis. The remaining eight analyses define a weighted mean 207Pb*/206Pb* date of 2401 ± 14 Ma (MSWD = 1.5; Table 7, Fig. 10D), interpreted as the age of monazite growth during hydrothermal alteration of the carbonaceous phyllite. This date is indistinguishable from that obtained on xenotime interpreted to date the timing of emplacement of the auriferous quartz-sulfide vein (GSWA 209907).

GSWA 209909a—altered dolerite, Paulsens gabbro: A finegrained, highly altered portion of the Paulsens gabbro (that is host to the Gabbro Offset deposit) containing a thin ankerite-quartz vein (GSWA 209909b) was collected from drill core PDU2153 between 345.1 and 345.3 m (Table 1). This sample is located 1 m above gabbro sample GSWA 209905, which yielded a baddeleyite crystallization age of ca. 2701 Ma. Locally, the sampled dolerite is pervasively altered by mus- covite-ankerite-leucoxene ± chlorite, crosscut by ankerite-quartz veins, and contains up to 48 ppm Au. Monazite in the sample occurs as irregular grains up to 100 μm across, intergrown with and containing inclusions of leucoxene after iron oxide minerals.

Eight analyses were made of five monazite grains. Uranium and thorium contents are 30 to 140 ppm and 15,000 to 55,000 ppm, respectively. Among the eight analyses, one indicates high common Pb and is excluded. The remaining seven analyses yield consistent 207Pb*/206Pb* dates, but 238U/206Pb* dates that vary broadly with Th content, which is probably a residual effect from the Th-correlated interference correction on 204Pb or from the matrix correction. Therefore, a less rigorous (10%) discordance cutoff was applied to these analyses. The seven analyses yield a weighted mean 207Pb*/206Pb* date of 2398 ± 37 Ma (MSWD = 1.4; Table 8, Fig. 10E), interpreted as the age of monazite growth during hydrothermal alteration of the dolerite. This age is indistinguishable from that obtained from hydrothermal monazite in sample GSWA 209902 and the age of xenotime cores in the auriferous quartz-sulfide vein (GSWA 209907).

GSWA 209912—carbonaceous phyllite, Hardey Formation:A fine-grained, strongly foliated, black carbonaceous muscovite phyllite of the Melrose argillite member was sampled from drill core PDU2217 between 103.8 and 104.0 m in the footwall to the Gabbro Offset deposit (Table 1). The sample is located in a D2/D3 high-strain zone and contains lenses and stringers of quartz veins and coarse-grained pyrite. The sample also contains elongate porphyroblasts up to 3 mm long comprising intergrowths of monazite and florencite in subequal amounts with lesser quartz and chlorite. In this sample, it appears that the monazite is partially replaced by florencite, which is common during greenschist facies metamorphism (Rasmussen and Muhling, 2009). However, it is unclear if the growth of monazite is related to low-grade metamorphism and deformation associated with fault reactivation or if it has locally replaced former regional-scale peak metamorphic porphyroblasts (such as andalusite [?]) during hydrothermal alteration.

Twenty-two analyses were made on three monazite grains within the porphyroblasts. Uranium and thorium contents vary from 50 to 120 ppm and 3,400 to 14,000 ppm, respectively. The largest grain yielded 12 analyses <5% discordant; the remaining data vary in their 238U/206Pb* ages, partly due to the low precision of low-U analyses. The 12 concordant analyses define a weighted mean 207Pb*/206Pb* date of 1730 ± 28 Ma (MSWD = 1.1; Table 9, Fig. 10F). This date is interpreted as the age of monazite growth during hydrothermal alteration of the carbonaceous phyllite, possibly during fault reactivation.

GSWA 219512, 219513, and 219517—micaceous phyllites, Hardey Formation: Fine-grained and slightly weathered phyllites were collected from shallow adits at Paulsens East (Table 1) in a 0.5- to 1.0-m-wide hydrothermal alteration halo surrounding 1- to 2-m-wide mineralized quartz-sulfide veins. Samples 219513 and 219517 contain large monazite grains up to 3 mm long, intergrown with minerals that define the main hydrothermal alteration assemblage (muscovite, quartz, and pyrite; Fig. 6D). The quartz-sulfide veins yielded rock-chip assays up to 6.65 ppm Au (Owen, 2000).

Sixteen analyses were made on seven monazite grains. Uranium and thorium contents of the samples varied from 30 to 135 ppm and 590 to 5,280 ppm, respectively. Four analyses >5% discordant and three with elevated common Pb were excluded from the age analysis. The remaining nine analyses provide concordant data that define a weighted mean 207Pb*/206Pb* date of 2403 ± 38 Ma (MSWD = 0.43; Table 10, Fig. 10G). This date is interpreted to represent the timing of hydrothermal alteration associated with the emplacement of the auriferous quartz-sulfide veins at Paulsens East and is within uncertainty of ages of hydrothermal monazite from altered samples at the Gabbro Offset deposit (GSWA 209902 and 209909a) as well as ages of xenotime cores from the auriferous quartz-sulfide vein at Gabbro Offset (GSWA 209907).

Samples GSWA 219512 and 219513 also contain abundant 1- to 3-mm-long, elongate porphyroblasts of interlocking monazite-florencite (Fig. 9B), together with minor chlorite and quartz, similar to the monazite-florencite porphyroblasts in sample GSWA 209912. Fifteen analyses were made on three monazite porphyroblasts. Uranium and thorium contents vary from 170 to 1,210 ppm and 380 to 8,945 ppm, respectively. Seven analyses >5% discordant and one analysis with elevated common Pb were excluded from the age analysis. The remaining seven concordant analyses define a weighted mean 207Pb»/206Pb» date of 1721 ± 32 Ma (MSWD = 0.58; Table 11, Fig. 10H). Because sample GSWA 219513 also contains abundant, coarse-grained ca. 2403 Ma monazite intergrown with the main hydrothermal alteration assemblage (i.e., not detrital), the date of ca. 1721 Ma for the monazite porphyroblasts must represent the age of a secondary hydrothermal event that affected the phyllites. This age is within uncertainty of that of monazite porphyroblasts (GSWA 209912) in a D2/D3 shear zone at the Gabbro Offset deposit.

Discussion

SHRIMP U-Th-Pb geochronology has been used to define the age of the host rocks to gold mineralization at the Gabbro Offset deposit, as well as the timing of regional-scale hydrothermal activity associated with the emplacement of the auriferous quartz-sulfide veins and subsequent regional-scale hydrothermal events (Fig. 11, Table 12).

Age of the host rocks

Sedimentary rocks of the Hardey Formation were deposited between ca. 2763 and 2745 Ma (Trendall et al., 2004). The detrital zircon data presented here from two quartz sandstone samples from near the middle of the Hardey Formation (Melrose argillite) have relatively similar detrital age modes (Fig. 8), although sample GSWA 209911 contains a significant proportion of detritus older than ca. 3400 Ma and younger than ca. 2900 Ma compared with sample GSWA 209903. Despite these differences, the similarity in age modes at ca. 3450, 3240 to 3230, 3020, and 2944 Ma indicate that they both shared a common source region. The age of 2750 ± 10 Ma for the youngest detrital zircon age component in sandstone sample GSWA 209911 (Fig. 8) therefore provides a maximum depositional age for the Hardey Formation in the South Pilbara sub-basin. A minimum age constraint is provided by the baddeleyite crystallization age of 2701 ± 11 Ma for the Paulsens gabbro (Fig. 10A; GSWA 209905), which intrudes these sedimentary rocks. The crystallization age of the Paulsens gabbro suggests it may have been a feeder to mafic volcanic horizons in the 2715 to 2629 Ma Jeerinah Formation (Trendall et al., 2004).

Timing of hydrothermal activity and gold mineralization

Euhedral xenotime within and interlocking with pyrite grains is dated at 2403 ± 5 Ma (GSWA 209907, Fig. 9C), providing a direct crystallization age for the auriferous quartz-sulfide veins at the Gabbro Offset deposit, which are peripheral to the main Paulsens lode. Additionally, hydrothermal monazite within a pervasively altered part of the Paulsens gabbro (2398 ± 37 Ma; GSWA 209909a) and in the altered wall-rock margins to quartz-sulfide veins at Paulsens East (2403 ± 38 Ma; GSWA 219513 and 219517) also records a period of regional- to localscale hydrothermal activity associated with the emplacement of auriferous quartz-sulfide veins at ca. 2400 Ma. Within the main quartz-sulfide vein at Gabbro Offset, xenotime crystals were also dated at 1680 ± 9 Ma. These grains only occur within the quartz matrix or at the margins of brecciated sulfide grains, suggesting that this date represents a second period of hydrothermal alteration during which ca. 2400 Ma xenotime grains underwent dissolution and reprecipitation.

Gold mineralization at Gabbro Offset is present in two distinct forms: (1) as rounded inclusions within euhedral pyrite grains (Fig. 6A) and (2) as free gold located along cracks and fractures of brecciated pyrite grains (Fig. 6B). Because the euhedral pyrite is intergrown with and contains euhedral grains of xenotime dated at ca. 2400 Ma, the rounded, primary gold inclusions in this pyrite are also interpreted to have formed at this time. However, where the pyrite is locally fractured or brecciated, containing flakes of free gold (Fig. 6B), chalcopyrite, and pyrrhotite, xenotime in the matrix and along pyrite grain boundaries has undergone dissolution and reprecipitation reactions at ca. 1680 Ma. This suggests that these rocks were subjected to a secondary hydrothermal event responsible for the brittle deformation of pyrite grains, xenotime dissolution, and reprecipitation, and either the local remobilization (Fig. 6C) or introduction of new gold, or both, at ca. 1680 Ma. Additionally, xenotime from ankerite-quartz veins emplaced within the Paulsens gabbro, which are associated with up to 48 ppm Au, has been dated at 1655 ± 37 Ma (GSWA 209909b). This indicates that this second hydrothermal event was also responsible for the emplacement of gold-bearing ankerite-quartz veins. A study on the mineralogy and chemistry of gold at the main Paulsens lode indicates that the youngest phase of gold mineralization is also associated with free gold with elevated Hg and Ag which formed in pyrite fractures and along stylolites (Hancock and Thorne, 2016), indicating a commonality between the youngest gold mineralizing event at the main Paulsens lode and the secondary event at the Gabbro Offset deposit.

Relationships to regional orogenic events

Our in situ dating of monazite and xenotime at the Paulsens deposit has identified three discrete hydrothermal events at ca. 2400, 1730, and 1680 Ma. The first hydrothermal event at ca. 2400 Ma is marked by the emplacement of auriferous quartz-sulfide veins at Gabbro Offset and Paulsens East, resulting in pervasive alteration of gabbroic and phyllitic host rocks as well as the growth of monazite within carbonaceous phyllites, either as primary porphyroblasts or as pseudomorphs of former metamorphic porphyroblasts. The timing of the mineralizing hydrothermal event (ca. 2400 Ma) does not correlate with any known hydrothermal or deformational event in the northern Capricorn Orogen. However, cryptic events, including the growth of monazite in phyllitic rocks across the Pilbara region from Whim Creek in the north to Mt. Tom Price in the south (Rasmussen et al., 2005) and resetting of high-U zircons in tuffaceous mudstones of the Hamersley Group (Pickard, 2002), are recorded between ca. 2430 and 2399 Ma. The age of monazite growth decreases progressively southward (Rasmussen et al., 2005) and the data presented here are within uncertainty of monazite growth at 2399 ± 6 Ma at Mt. Tom Price, located 150 km to the east of the Paulsens deposit (Rasmussen et al., 2005).

In the Pilbara region, the cause of this cryptic event is currently unknown, but indirect evidence for a tectonic event at ca. 2400 Ma is present in the stratigraphic record of Proterozoic sedimentary rocks along the southern Pilbara margin. A study of detrital zircons from the matrix of the Meteorite Bore Member of the Kungarra Formation in the 2445 to 2208 Ma Turee Creek Group indicates that they were sourced, together with abundant rhyolite clasts, from the 2449 to 2445 Ma Woongarra Rhyolite (Blake and Barley, 1992; Takahara et al., 2010; Simonson et al., 2014). In the Hardey syncline, from where the samples were taken, the Woongarra Rhyolite is situated 1,500 m stratigraphically below the Meteorite Bore Member, implying at least 500 to 1,000 m of uplift and erosion of sedimentary rocks along the southern Pilbara margin between ca. 2445 and 2208 Ma (Takahara et al., 2010). Our data show that this event was associated with the emplacement of quartz-sulfide veins during extensive hydrothermal activity and gold mineralization at Paulsens (Fig. 11).

The second hydrothermal event is marked by the growth of monazite within carbonaceous phyllites of the Hardey Formation at 1730 ± 27 and 1721 ± 32 Ma. It is unclear if the monazite in these samples grew during low-grade metamorphism and deformation associated with fault reactivation (GSWA 209912 is located in a D2/D3 shear zone at the Gabbro Offset deposit) or if it has locally altered former regional-scale peak metamorphic porphyroblasts during subsequent hydrothermal activity. However, we note that this event is within uncertainty of dextral strike-slip faulting and hydrothermal gold mineralization at the Mt. Olympus deposit, dated at 1738 ± 5 Ma (Young et al., 2003; Şener et al., 2005), about 150 km to the southeast of Paulsens.

The youngest hydrothermal event is recorded at the Gabbro Offset deposit by the dissolution and reprecipitation of xenotime rims, dated at 1680 ± 9 Ma, on older ca. 2400 Ma xenotime cores within the quartz matrix of the auriferous quartz-sulfide vein, and by xenotime grains, dated at 1655 ± 37 Ma, within discrete ankerite-quartz veins. This hydrothermal event appears to have been responsible for the local brecciation and fracturing of pyrite grains in the quartz-sulfide vein and for the local remobilization of gold (Fig. 6C). However, it is possible that this event was also responsible for the introduction of new gold precipitated as free gold in pyrite fractures and within stylolites at the main Paulsens lode (Hancock and Thorne, 2016). This event is coincident with medium- to high-grade metamorphism, deformation, and magmatism during the 1680 to 1620 Ma Mangaroon Orogeny (Sheppard et al., 2005), which affected rocks of the Gascoyne Province farther south. Rasmussen et al. (2007a) also noted hydrothermal monazite growth at ca. 1650 Ma in the Soansville Group in the Pilbara craton and attributed it to reactivation along preexisting N- to NE-trending, craton-scale structures that acted as conduits for hydrothermal fluid flow during the Mangaroon Orogeny. Rasmussen et al. (2007b) also documented the presence of ca. 1670 Ma hydrothermal xenotime at the Mt. Tom Price iron ore mine. The structural effect of this event at Paulsens is not known, but may have resulted in the reactivation of preexisting faults, such as the Nanjilgardy and Hardey faults.

Importantly, we find no evidence for hydrothermal activity related to either the 2215 to 2145 Ma Ophthalmia Orogeny or the 1820 to 1770 Ma Capricorn Orogeny. This is surprising because these events played a significant role in the tectonic evolution of the region elsewhere in the southern Pilbara craton and northern Capricorn Orogen (Young et al., 2003; Rasmussen et al., 2005, 2006; Şener et al., 2005; Thorne et al., 2011).

Conclusions

Our results reveal a significantly different and more complicated low-temperature tectonothermal evolution for the southern Pilbara region than had been thought. Previous studies (e.g., Thorne and Seymour, 1991) suggested that gold mineralization in the Wyloo Inlier is associated with post-Wyloo Group quartz veins, linking the timing of orogenic gold mineralization to the 1820 to 1770 Ma Capricorn Orogeny. Results from this study show that two gold mineralizing events occurred at Paulsens. Primary gold mineralization at Paulsens was much older, occurring at ca. 2400 Ma, and does not correlate with a known orogenic event in the northern Capricorn Orogen. However, uplift, erosion, and hydrothermal activity throughout the Pilbara craton may reflect a previously unrecognized orogenic event at ca. 2400 Ma. The second gold event, at ca. 1680 Ma, may be linked to reactivation of the Hardey fault, which forms a splay off the lithospheric-scale Nanjilgardy fault (Fig. 1).

Recognition of a new orogenic event has important implications for exploration targeting of orogenic gold deposits in the northern Capricorn Orogen. Previously, rocks affected by the Capricorn Orogeny, up to and including the Ashburton Formation, were considered the most prospective for orogenic gold mineralization similar to that at Paulsens. However, we have demonstrated that the primary mineralization at Paulsens, which shares many of the characteristics of orogenic gold deposits, occurred at ca. 2400 Ma, about 570 million years prior to deposition of Wyloo Group sedimentary rocks. Therefore, older rocks of the Fortescue and Hamersley groups are more likely candidates to host significant gold mineralization similar to that at Paulsens. The second gold event occurring at ca. 1680 Ma is possibly related to the reactivation of major lithospheric fault structures and their splays. This event, at least locally, caused the remobilization of preexisting gold, but may also have been responsible for the introduction of new gold along these mantle-tapping structures. Nevertheless, the age constraints for mineralization at Paulsens differ from those proposed by Young et al. (2003) and Sener et al. (2005) for the timing of gold mineralization at Mount Olympus (1738 ± 5 Ma) in the Wyloo Group, suggesting the potential for several different gold mineralizing events across the northern Capricorn Orogen associated with major structures such as the Nanjilgardy fault.

For exploration of orogenic gold, it is critical to have an understanding of the distribution of mineralization through both space and time. In situ geochronology of monazite and xenotime has important implications for exploration targeting by providing absolute ages for orogenic gold mineralization and allowing mineralization to be linked to specific orogenic events. In turn, this allows explorers to minimize their exploration search area by targeting the most prospective tectono-thermal events and host stratigraphy.

Acknowledgments

This research is part of a Ph.D. project by I. Fielding, funded through an ARC Linkage grant (LP130100922), an industry scholarship by Northern Star Resources (NSR), and the WA Government Exploration Incentive Scheme. Geochronology was carried out using a SHRIMP II ion microprobe at the John de Laeter Centre at Curtin University. SPJ and MTDW publish with permission of the director of the Geological Survey of Western Australia. We thank NSR for samples and for permission to publish, and A. White, D. Huston, and an anonymous reviewer for their helpful comments.

Analytical Methods

U-Pb zircon geochronology

Analytical methods for U-Pb zircon geochronology are described in detail by Wingate and Lu (2016), and only a summary is provided here. During all analytical sessions, an O2- primary beam with a spot size of 20 to 30 μm was used with beam intensity of 1.5 to 3.5 nA. The secondary ion beam was focused through a 100-μm collector slit onto an electron multiplier to produce mass peaks with flat tops and a mass resolution (1% peak height) of 5,100 to 5,460.

Data were collected in sets of six scans, with reference standards analyzed after every five sample analyses. Count times per scan for Pb isotopes 204, background positions 204.1, 206, 207, and 208, were 10, 10, 10, 30, and 10 s, respectively. U-Th-Pb ratios and absolute abundances were determined relative to the BR266 standard zircon (559 Ma, 903 ppm U; Stern, 2001), analyses of which were interspersed with those of unknown zircons. Instrumental mass fractionation (IMF) in 207Pb/206Pb ratios was monitored during each session by repeated analysis of the 3465 Ma OGC zircon standard (OG1 of Stern et al., 2009). No IMF correction was required because the measured values of OGC were in agreement with the reference value within 2σ uncertainties. Raw zircon data were processed using the SQUID 2 add-in (v. 2.50.12.03.08) for Excel 2003 (Ludwig, 2009) and plotted using the ISOPLOT add-in (v. 3.76.12.02.24; Ludwig, 2003). Measured compositions were corrected for common Pb using nonradiogenic 204Pb and contemporaneous Pb composition according to the terrestrial Pb evolution model of Stacey and Kramers (1975). Mean ages are quoted with 95% confidence levels.

In situ U-Pb baddeleyite, monazite, and xenotime geochronology

For in situ baddeleyite, monazite, and xenotime analysis, a primary beam of O2- ions was focused through a 50-μm Kohler aperture to produce an oval 10-μm-wide spot on the sample surface with a current of 0.2 to 0.4 nA. The secondary ion system was focused through a 100-μm collector slit onto an electron multiplier to produce mass peaks with flat tops and a mass resolution (1% peak height) of >5,200 in all sessions. Background counts from scattered ions were reduced using a flight retardation lens, which is known to cause slight sessiondependent IMF of Pb isotopes. IMF corrections were applied to all analyses using the monazite standard Z2908 as an IMF monitor (see below).

Data were collected in sets of eight scans, with monazite, xenotime, or baddeleyite reference material analyzed every four to six sample analyses. Count times per scan for Pb isotopes 204, background position 204.045, 206, 207, and 208, were 10, 10, 10, 30, and 10 s, respectively.

Baddeleyite was analyzed using the conventional zircon 9-peak run table, calibrated against baddeleyite reference material Phalaborwa (Heaman and LeCheminant, 1993). Monazite was analyzed with a 13-peak run table as defined in Fletcher et al. (2010), which includes mass stations for the estimation of La, Ce, and Nd ((REEPO2+)) and Y (YCeO+). Measurements of monazite standards FRENCH, Z2234, and Z2908 (see Fletcher et al., 2010, for details) were done concurrently for Pb/U and Pb/Th calibration (FRENCH), IMF corrections (Z2908), and matrix corrections required for variable U, Th, Y, and Nd contents (Z2234 and Z2908).

Xenotime was analyzed with a 9-peak run table following analytical protocols in Fletcher et al. (2004) and Fletcher et al. (2010). Pb/U calibrations and matrix corrections for U and Th contents were based on concurrent measurements of the standards MG-1 (Fletcher et al., 2004) and Z6413 (“Xeno1” Stern and Rayner, 2003). Pb/Th was determined indirectly, using a fixed Th/U calibration (Fletcher et al., 2004). Matrix corrections for REEs assumed the samples have REE abundances similar to Xeno1.

Raw data from analyses on baddeleyite, monazite, and xenotime were processed using the SQUID 2 add-in (v. 2.50.12.03.08) for Excel 2003 (Ludwig, 2009) and plotted using the ISOPLOT add-in (v. 3.76.12.02.24; Ludwig, 2003). Common Pb corrections were based on measured 204Pb/206Pb ratios and contemporaneous Pb composition, according to the terrestrial Pb evolution model of Stacey and Kramers (1975). Matrix effect corrections were made for all monazite and xenotime data using procedures described by Fletcher et al. (2010) and Fletcher et al. (2004), respectively. Pooled ages are quoted with 95% confidence levels, whereas individual analyses are presented with 1σ errors.