Early Gondwanan connection for the Argentine Precordillera terrane

Early Gondwanan connection forthe Argentine Precordillera terrane

Stan Finney a;�, James Gleason b, George Gehrels c, Silvio Peralta d,Guillermo Acen‹olaza e

a Department of Geological Sciences, California State University at Long Beach, Long Beach, CA 90840, USAb Department of Geological Sciences, University of Michigan, Ann Arbor, MI 48109, USA

c Department of Geosciences, University of Arizona, Tucson, AZ 85721, USAd CONICET, Universidad Nacional de San Juan, 5400 Rivadavia, San Juan, Argentina

e CONICET, Universidad Nacional de Tucuma¤n, Tucuma¤n, Argentina

Received 13 August 2002; received in revised form 15 October 2002; accepted 27 October 2002

Abstract

The Precordillera of Argentina is widely accepted as an exotic terrane of Laurentian (North American) affinity.Newly acquired U/Pb ages on individual detrital zircons from Lower Cambrian and Upper Ordovician quartzsandstone beds in the Argentine Precordillera indicate a Gondwanan provenance not associated with any known partof Laurentia. Accordingly, the Precordillera terrane is likely underlain by basement rock of Gondwanan affinity. Inaddition, detrital zircons from the Upper Ordovician sandstone bed provide no evidence for a Mid Ordovicianposition against the inboard Famatina arc. These results demand critical re-evaluation of widely held assumptionsregarding the paleogeography of the Argentine Precordillera.8 2002 Elsevier Science B.V. All rights reserved.

Keywords: Argentina; Precordillera; paleogeography; geochronology; Cambrian; Ordovician

1. Introduction

Evidence of a Laurentian (North American) af-¢nity for the Argentine Precordillera terrane hasstimulated extensive research over the last decaderesulting in much discussion and many papers [1^

12]. The potential for a large exotic continentalfragment of North American a⁄nity residing inSouth America is intriguing in its own right; how-ever, models for the tectonic transfer of this con-tinental fragment have also had a large in£uenceon interpretations of global paleogeography in theearly Paleozoic [4]. In the most widely acceptedmodel, the so-called ‘funeral ship’ model of Tho-mas and Astini [3], the Precordillera terrane riftedfrom the southern margin of Laurentia in theEarly Cambrian, drifted across the Iapetus Oceanas a microcontinent, and docked with the proto-Andean margin of Gondwana in the Mid Ordo-

0012-821X / 02 / $ ^ see front matter 8 2002 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 0 2 ) 0 1 0 6 3 - 4

* Corresponding author. Tel. : +1-562-985-8637.E-mail addresses: sc¢[email protected] (S. Finney),

[email protected] (J. Gleason), [email protected](G. Gehrels), [email protected] (S. Peralta),[email protected] (G. Acen‹olaza).

EPSL 6476 3-1-03

Earth and Planetary Science Letters 205 (2003) 349^359

www.elsevier.com/locate/epsl

vician. In contrast, Dalziel [4] argued for a con-tinent^continent transfer. In his ‘calling card’model, the Texas Plateau (a hypothetical promon-tory of Laurentia) collided with the proto-Andeanmargin of Gondwana during the early Mid Ordo-vician. The Precordillera terrane was then de-tached from the Texas Plateau as Gondwanarifted away from Laurentia in the early Late Or-dovician. In both models, the Early to Mid Ordo-vician Famatinian orogeny is thought to mark theapproach and initial encounter of the Precordil-lera continental fragment with the proto-Andeanmargin of Gondwana. In an alternative scenario[6,10], the Precordillera was detached from Lau-rentia during the Mid to Late Ordovician, and didnot dock with Gondwana until Silurian^Devoniantime. An earlier model suggesting more direct col-lision between Laurentia and Gondwana (Northand South America) during the Ordovician [13,14]has since been discarded based mainly on paleo-magnetic and faunal evidence to the contrary [15].A much earlier model in which the Precordilleraterrane was always part of Gondwana [16] has notreceived any serious recent consideration.

In spite of these di¡erent interpretations, theLaurentian a⁄nity of the Precordillera terraneis rarely questioned because it possesses a thickAppalachian-type Cambro^Ordovician carbonateplatform succession that contains a ‘Paci¢c’benthic olenellid trilobite fauna, otherwise knownonly from Laurentia [17^19], as well as similarwarm-water conodont and brachiopod faunas[20,21] and reef and reef-mound facies [10]. Inaddition, limited data suggest that the Precordil-lera terrane is underlain by Grenvillian-age base-ment rocks of Laurentian a⁄nity [22].

In this paper, we challenge this widely held in-terpretation. Recently acquired U/Pb ages on in-dividual detrital zircons from Lower Cambrianand Upper Ordovician quartz sandstone beds inthe Argentine Precordillera and from a MiddleCambrian(?) to lowest Ordovician quartz sand-stone bed in the Sierras Pampeanas indicate adistinctive Gondwanan provenance not associatedwith any known part of Laurentia. These datastrongly support an early Gondwanan connectionfor the Precordillera terrane. The compelling Lau-rentian paleobiogeographic a⁄nity of the Precor-

dillera terrane must then be explained. Two ex-planations present themselves: either the Pre-cordillera (1) occupied a northern, low-latitudeposition in West Gondwana (early Paleozoic co-ordinates), rather than Laurentia, or (2) had amuch more complicated odyssey involving separa-tion from Gondwana and transfer to Laurentiaduring the breakup of the Pannotia superconti-nent, followed by return to Gondwana in themid Paleozoic. In addition, our new detrital zir-con data argue against the widely held interpreta-tion [1,3,8] that the Famatina system, a subduc-tion-related magmatic arc, developed in responseto the convergence, and Mid Ordovician docking,of the Precordillera. Instead, our data supportinterpretations that the Precordillera did not ar-rive at its present location relative to the Famati-na system until sometime between late Ordovicianand Devonian [6,10,16].

2. Analytical methods

Zircons were analyzed with a Micromass Iso-probe Excimer laser-ablation multicollectorICPMS equipped with nine Faraday collectors,an axial Daly detector, and four ion-countingchannels. Analyses were conducted in staticmode with 204Pb measured in an ion-countingchannel while 206Pb, 207Pb, 208Pb, 232Th, and238U were measured with Faraday detectors. Theanalyses were conducted with 50 micron spots,with a total measurement time of V90 s per anal-ysis. Inter-element fractionation was monitoredby analyzing fragments of a large concordant zir-con crystal with a known (ID-TIMS) age of564S 4 Ma (2c error) (G.E. Gehrels, unpublisheddata). Common Pb corrections were made withmeasured 204Pb, assuming an initial Pb composi-tion according to Stacey and Kramers [23]. Re-ported ages are based on 206Pb/238U for ages lessthan 1.0 Ga, and on 206Pb/207Pb for ages s 1.0Ga. Total uncertainty for these ages is V2% (1c).U/Pb ages shown in the plots of the La Ce¤bila,Cerro Totora, and Las Vacas samples are mostlyconcordant, with some up to 15% discordant.Grains with U/Pb ages that were more than 15%discordant, or that have large uncertainties, were

EPSL 6476 3-1-03

S. Finney et al. / Earth and Planetary Science Letters 205 (2003) 349^359350

excluded from the data set (approximately 10% ofall grains).

3. Terminology

Ordovician chronostratigraphic terminology isconfusing. The widely used British Series werenot well de¢ned, and their de¢nitions have beenrevised often. Because of marked paleobiogeo-graphic di¡erentiation of Ordovician faunas, inde-pendent chronostratigraphic subdivisions havebeen established for almost every paleo-plate.The geologic history of the Argentine Precordil-lera is usually expressed in terms of British Series,yet British Ordovician faunas are endemic, thusmaking correlations to the Precordillera succes-sions inaccurate. We have chosen to use the newglobal Ordovician Series that are being re-de¢nedby the International Subcommission on Ordovi-cian Stratigraphy (ICS) [24], and which are shownin the Geological Time Scale posted on the ICSwebsite (http://www.micropress.org/stratigraphy/).As recently recommended by ICS, Middle Ordo-vician and Mid Ordovician are used for the seriesand its correlative epoch, respectively. We use thegeochronologic ages that are in the ICS Geolog-ical Time Scale and those presented in Cooper[25]. We note that the integration of geochrono-logic ages with the global Ordovician chronostra-tigraphic units di¡ers from that used in severalearlier publications on the Precordillera.

By referring to the Pannotia landmass in thispaper, we mean the short-lived, latest Precambri-an supercontinent that existed immediately fol-lowing the opening of the Paci¢c Ocean basin,after the breakout of East Gondwana from Rodi-nia and its subsequent amalgamation with WestGondwana. See ¢gure 12 in Dalziel [4].

4. Samples, results, interpretations

4.1. La Ce¤bila

Our La Ce¤bila sample is from a large outcropof the La Ce¤bila Formation along La Ce¤bilacreek, which separates the Sierra de Ambato in

Catamarca Province to the north from the Sierrade Velasco in La Rioja Province to the south (Fig.1). Both ranges are part of the Sierras Pampeanasimmediately east of the Sierra de Famatina. Theage of the unfossiliferous La Ce¤bila Formation isuncertain, and it has been correlated to either theUpper Neoproterozoic^Lower Cambrian Punco-viscana Formation or the Middle to Upper Cam-brian Meso¤n Group [26,27]. The age distributionof detrital zircons, presented below, demonstratesthat the La Ce¤bila Formation is lowest Ordovi-cian (6 480 Ma), and it might be correlative, atleast in part, with the Middle to Upper CambrianMeso¤n Group, which also is generally unfossilif-erous. The La Ce¤bila Formation is composed ofmature siliciclastic sediments deposited in a shal-low marine platform environment. Given its ageand geographic and geologic setting, it was likelydeposited on the Gondwanan margin, and mustoverlie the Neoproterozoic to early CambrianPuncoviscana Formation, which was deformedduring the Pampean Orogeny, 530^520 Ma [6].The sediment of the La Ce¤bila Formation wasprobably derived in part, and reworked from,the Puncoviscana Formation, which is a thickmeta-sedimentary sequence also interpreted tohave a Gondwanan provenance (i.e., the Amazo-nian craton).

The detrital zircons of the La Ce¤bila sample(Fig. 2) fall into three broad groups of decliningabundance that represent the following sources:(1) 0.48^0.72 Ga grains likely derived from intru-sions of the Pampean Orogeny, reworked Punco-viscana sediments, and possibly sediment derivedfrom extensions of Brasiliano/Pan-African oro-genic belts between the older cratons of WestGondwana, (2) 1.05^1.48 Ga grains representingGrenvillian age-equivalent terranes within Gond-wana, perhaps part of the Sunsas orogenic belt ofthe Amazonian craton or the Kibaran belts of theCongo and Kalahari cratons, and (3) 1.90^2.30Ga grains likely derived from older Transamazo-nian/Birimian age terranes [28^33]. In the 0.48^0.72 Ga group, the youngest peak (480^520 Ma)overlaps in age (earliest Ordovician) with, and thedetrital zircons could have been eroded from, theearliest magmatic units (515^450 Ma [34] or 490^450 Ma [35]) of the Famatina arc. These granitic

EPSL 6476 3-1-03

S. Finney et al. / Earth and Planetary Science Letters 205 (2003) 349^359 351

to tonalitic rocks are widespread in the westernSierras Pampeanas especially to the west of thedepositional site of the La Ce¤bila Formation inclose proximity to the proto-Andean margin ofGondwana. The Grenvillian age-equivalent popu-lation, with its broad group that includes distinc-tive peaks at 1.30 and 1.45 Ga, is markedly dif-ferent from Grenvillian populations of detritalzircons from the Ouachita orogen and Appala-chian river sediments of North America, whichare concentrated in a narrow group (0.95^1.25Ga) with a large peak at 1.02^1.05 Ga (Fig. 2).These di¡erences allow for the ready di¡erentia-tion of Grenvillian-age populations of Gondwanaand Laurentia, respectively. Because of their un-

equivocal Gondwanan provenance, detrital zirconpopulations in the La Ce¤bila sample are a criticalreference for evaluating the two samples from thePrecordillera.

4.2. Cerro Totora

The Cerro Totora sample (Fig. 1) is from a wellstudied outcrop area of the Cerro Totora Forma-tion immediately east of Cerro Totora in thenorthern part of the Argentine Precordillera [36].The formation is composed of gypsum and inter-bedded stromatolitic and oolitic dolostone in itslower part and red silty mudstone, quartzosesandstone and oolitic dolostone in its upper

Fig. 1. Index map showing the location of La Ce¤bila, Cerro Totora, and Las Vacas samples in Sierras Pampeanas and Precordil-lera of northwest Argentina. Substantially modi¢ed from Astini et al. [1].

EPSL 6476 3-1-03

S. Finney et al. / Earth and Planetary Science Letters 205 (2003) 349^359352

part. It is capped by a glauconitic sandstone andan overlying green shale with Early Cambrianolenellid trilobites. The sandstone bed sampledfor detrital zircons is within the upper part ofan V30 m thick interval dominated by red siltymudstone, V75 m below the top of the forma-tion. The Cerro Totora Formation is overlain bythe Middle Cambrian Hornos Formation (possi-bly a facies of the widespread La Laja Forma-tion), the lowest unit at Cerro Totora of the thickCambrian^Lower Ordovician carbonate platformsuccession, which is characteristic of the Precor-dillera. The Cerro Totora Formation was inter-preted recently as a synrift sedimentary successiondeposited in a graben that developed on the Pre-cordillera terrane as it rifted from Laurentia[36,37]. The Cerro Totora Formation is boundedbelow by a Tertiary thrust fault and is structurallydetached from its basement, but that basement isassumed to be Precordilleran (Laurentian) base-

ment of Grenvillian age (e.g. V1.0^1.2 Ga) [22].The contact between the Cerro Totora and Hor-nos formations is interpreted as a record of thetransition in this part of the Precordillera fromactive rift to passive margin. Given this interpre-tation for the depositional setting and location ofthe Cerro Totora Formation, and the widely ac-cepted Grenvillian-age basement of Laurentian af-¢nity for the Precordillera, the detrital zircon agepopulations in our sample are surprising (Fig. 2).

Most zircons in the Cerro Totora sample cor-respond chronologically to the Brasiliano/Pan-African and the Transamazonian/Birimian/Ebur-nian orogenic cycles. The Brasiliano/Pan-Africanpopulation (0.55^0.80 Ga) is very similar to thatfrom the La Ce¤bila sample except for the peak inthe La Ce¤bila sample of young grains possiblyderived from the Famatina arc, re£ecting thefact that deposition of the Cerro Totora Forma-tion pre-dated the Famatina arc. The most abun-dant population of zircons in the Cerro TotoraFormation is Early Proterozoic, with a large num-ber of 2.00^2.25 Ga grains, an age range that isdistinctive of Transamazonian/Birimian/Eburnianprovenance and found only in limited areas ofnorthwesternmost Laurentia (Fig. 3) [38]. Thesample also includes a substantial population ofArchean grains (2.40^2.80, 2.67^2.75 Ga). Anysignature of a Grenvillian provenance is essen-tially lacking in this sample, raising serious ques-tions regarding the a⁄nity of the Precordillerabasement. The sandstone bed we sampled is avery mature quartzite, suggesting considerable re-working and recycling; yet given its age and geo-logic setting (both paleogeographic and paleo-environmental), it must include a substantialcomponent derived from the basement of the Pre-cordillera. Finally, we note that the detrital zirconage population of the Cerro Totora Formation isremarkably similar to that of exotic early Paleo-zoic sedimentary rocks in the Florida basement(Suwannee basin; Fig. 2), also interpreted asGondwanan [39]. Based on these data and region-al comparisons, we ¢nd it di⁄cult to deny anunambiguously Gondwanan provenance for theCerro Totora Formation and a Gondwanan a⁄n-ity for associated basement rocks of the Precor-dillera terrane.

Rela

tive a

ge p

robabili

ty

0 500 1000 1500 2000 2500 3000

Detrital zircon age (Ma)

La Cebila (n=95)

CerroTotora (n=87)

Las Vacas (n=87)

Appalachian rivers (n=637)

Suwannee basin (n=46)

Ouachita orogen (n=92)

Fig. 2. Relative age^probability curves showing U^Pb indi-vidual detrital zircon age spectra for La Ce¤bila Formation(Upper Cambrian? to Lower Ordovician), Cerro Totora For-mation (Lower Cambrian), and Las Vacas Formation (lowerUpper Ordovician), all discussed in this paper. Shown forcomparison are U^Pb individual detrital zircon age data forUS modern Appalachian rivers [52], Ouachita orogen Paleo-zoic sandstones [53,54], and early Paleozoic sandstone ofGondwanan a⁄nity from the Florida subsurface Suwanneebasin [39]. Number of grains analyzed in each sample isshown in parentheses. In this ¢gure, each curve incorporatesthe age and analytical uncertainty for each grain as a normalprobability distribution. Each curve is then normalized to thenumber of grains analyzed resulting in curves of equal area[55].

EPSL 6476 3-1-03

S. Finney et al. / Earth and Planetary Science Letters 205 (2003) 349^359 353

4.3. Las Vacas

The Las Vacas sample was collected from a bedof quartz arenite in the upper conglomeraticmember of the Las Vacas Formation from Que-brada de Las Plantas, a section southwest ofGuandacol at the north end of the Precordillera[40]. The Las Vacas Formation, included in theTrapiche Group [1,40], is early Late Ordovician inage (V455^458 Ma). The sandstone beds are in athick sequence of basin-¢ll conglomerate that ac-cumulated in fault-bounded basins. These basinsdeveloped atop the Precordillera carbonate plat-form as it broke apart and foundered in Mid toLate Ordovician time [7,11]. Astini and co-au-thors [1,7,11,12] interpret basin formation as aresponse to the docking of the Precordillera ter-rane with Gondwana in the early Mid Ordovician,which resulted in subsidence of a peripheral fore-land basin and drowning of the carbonate plat-form, followed by post-collisional extension andconsiderable sediment progradation from the east.

The Las Vacas sample has a large population ofGrenvillian-age zircons with an age distributionvirtually identical with that of the La Ce¤bila sam-ple and di¡erent from populations of Grenvillian-age zircons in Ordovician quartz sandstones in theOuachita orogen and Appalachian river sediments(Fig. 2). Detrital zircons from the Las Vacas sam-ple are concentrated in a single broad group from1.0 to 1.5 Ga with peaks at 1.19, 1.30, and 1.45Ga; those of the La Ce¤bila sample also compose asingle broad group from 1.0 to 1.5 Ga with peaksat 1.08, 1.19, 1.30, and 1.45 Ga (Fig. 2). Accord-ingly, we conclude that the provenance of the LasVacas sample must represent a Gondwanan prov-enance, such as that recorded in the La Ce¤bilasample. We particularly note the absence of detri-tal zircons less than 520 Ma (Fig. 2), which indi-cates that the Precordillera terrane was not at itspresent location close to the Sierra de Famatinaduring deposition of the Las Vacas Formation.

In addition to carbonate clasts of intrabasinalprovenance, the conglomerates of the Las Vacas

EAST ANTARCTICA

AMAZONIA

WEST

AFRICA

CONGO

LAURENTIA

BALTICA

AUSTRALIA

INDIA

PLA

TA

KALAHARI

N

0o

Grenvillian belts

pre-Grenvillian (Transamazonain/Birmian/Eburnian) orogenic belts

pre-Grenvillian (without Transamazonian/Birmian/Eburnian) orogenic belts

Brasiliano/Pan-African belts

Equatorial oceanic currents facilitating dispersal of larvae of benthic organisms

Present location of Famatina system (inboard) andArgentine Precordillera (outboard) relative to Precambrian cratons and orogenic belts of South America

L

LL

PrecordilleraTerrane

Fig. 3. Early Cambrian (V500 Ma) paleogeographic reconstruction, modi¢ed substantially from Ho¡man [46], showing theGondwanan model for the Precordillera terrane, and the present position (rectangle) of Famatina System and Argentine Precor-dillera relative to Precambrian cratons and orogenic belts of South America.

EPSL 6476 3-1-03

S. Finney et al. / Earth and Planetary Science Letters 205 (2003) 349^359354

Formation include igneous and sedimentaryclasts, which indicate an extrabasinal provenance.As the source of these clasts, Thomas et al. [12]proposed present-day basement rocks exposed inthe Sierra de Pie de Palo and Sierra de Valle Fe¤rtil(Fig. 1), ranges in the westernmost Sierras Pam-peanas, west of the Sierra de Famatina, and con-sidered to represent the eastern part of the Pre-cordillera terrane. The Sierra de Famatina mustalso be considered as the source of the igneousclasts because of its voluminous igneous rocksof compositions (granodiorite, tonalite, and gab-bro) similar to the clasts and because, accordingto orthodox interpretations [1,12,34], the Famati-na arc stood as a topographic high immediatelyeast of the Precordillera terrane in Mid to LateOrdovician time, having formed during Early toMid Ordovician time as the Precordillera terraneapproached and docked with the proto-Andeanmargin of Gondwana. However, these paleogeo-graphic relationships must be questioned becauseof the near absence of metamorphic clasts in theLas Vacas conglomerate [41] in contrast to thedominance of Proterozoic-age metamorphic rockin the westernmost Sierras Pampeanas. Becauseof their textural and mineralogical maturity, thequartz sandstones beds in the Las Vacas Forma-tion must represent a source on the Precordilleraterrane that is not known, or they too must havea provenance to the east. The absence of detritalzircons with ages equivalent to the igneous rocksof the Famatina system and the dominance ofGrenvillian-age zircons of Gondwanan characterin the Las Vacas sample indicate that the prove-nance of the quartz sand in the Las Vacas For-mation, as well as the igneous clasts, was an areaof Gondwana far away from the Famatina sys-tem.

5. Discussion

Most models arguing for the Laurentian a⁄nityof the Precordillera terrane begin with the as-sumption that the basement of the Precordillerais of Grenvillian age, even though the basement isnot exposed anywhere in the Precordillera. Thisassumption is based on crustal xenoliths from

Miocene lava domes in the Precordillera thathave (1) U/Pb zircon crystallization ages of1.102 and 1.118 Ga and (2) whole-rock Pb iso-topic signatures compatible with Grenvillian base-ment in the Llano uplift along the southern mar-gin of Laurentia [22]. These common Pb isotopicsignatures are widely assumed to be signi¢cantlydi¡erent from nearby areas of South America,particularly basement and volcanic rocks in theAndes [22]. In addition, the basement of the west-ernmost Sierras Pampeanas (e.g. the Sierra de Piede Palo and Sierra de Valle Fe¤rtil), considered tobe the eastern part of the Precordillera terrane,was deformed and metamorphosed during theGrenvillian orogeny as determined by a U/Pbupper intercept age of 1.060 Ga for metagneiss[42]. These data represent a very tiny sample ofthe Precordilleran basement, especially when con-trasted with our Cerro Totora sample (87 zircongrains). We believe that the Cerro Totora sample,with its clear Gondwanan a⁄nity, is much morelikely to re£ect the age of the Precordilleran base-ment. Similarity of Pb isotope ratios betweenthe xenoliths of the Precordillera and rocks ofthe extensive Grenvillian age-equivalent beltsthroughout Gondwana cannot be ruled out be-cause such comparisons were not made by Kayet al. [22], and, in fact, Wareham et al. [43] re-ported close similarity of Pb and Nd isotope char-acteristics of Grenvillian-age rocks of Laurentiaand the Precordillera to Mesoproterozoic rocksin East and West Antarctica, Natal, and the Falk-land Islands. A recent isotopic study of Grenvil-lian-age basement in the southern part of thewestern Sierras Pampeanas reveals signi¢cant iso-topic di¡erences from the Laurentian Grenville[44]. Thus, the widely held assumption of domi-nantly Grenvillian-age basement of Laurentian af-¢nity for the Precordillera may be untrue.

Our proposed early Gondwanan a⁄nity for thePrecordillera terrane is based on only threesampled formations. Although preliminary, theyare compelling, positive evidence. Most impor-tantly, the large number of grains analyzed ineach sample (87 in both the Cerro Totora andLas Vacas samples) re£ects extensive samplingof source areas. If the basement of the ArgentinePrecordillera is considered Gondwanan, then its

EPSL 6476 3-1-03

S. Finney et al. / Earth and Planetary Science Letters 205 (2003) 349^359 355

paleogeographic history must be reconsidered, al-beit with the constraints imposed by the sedimen-tological and paleontological evidence seeminglyto the contrary.

The Precordillera continental fragment mayhave always been part of Gondwana, moving toits present position adjacent to the Famatina sys-tem sometime during the latest Ordovician to De-vonian. Paleomagnetic evidence from the CerroTotora Formation has been used to argue thatthe Precordillera terrane was located at the siteof the Ouachita embayment of southern Lauren-tia in the early Cambrian, and was not close to itspresent location relative to the rest of SouthAmerica [45]. However, the only paleomagneticconstraint available is paleolatitude (V20‡S). Pa-leolongitude is only speculative. Although this pa-leolatitude is the same as that of the Ouachitaembayment, it does not exclude a position innorthern Gondwana (Cambrian coordinates). Insome Early Cambrian paleogeographic recon-structions [4,46], a northern location in easternGondwana would place the Precordillera micro-continent within the tropics with only a narrowocean separating it from Laurentia (Fig. 3). Sucha paleolatitude could account for the thick Cam-brian to Lower Ordovician carbonate platformsuccession of the Precordillera, as well as warm-water tropical faunas. Strong west-£owing equa-torial currents would have enabled rapid dispersalof larvae of endemic benthic organisms from Lau-rentia, resulting in the Laurentian paleobiogeo-graphic a⁄nity of the Precordillera. A later stepin this Gondwanan hypothesis would be the mi-gration of the Precordillera microcontinent to itspresent position adjacent to the Famatina systemafter the Late Ordovician. Mid to Late Ordovi-cian drowning of the carbonate platform and for-mation of fault-bounded basins, in one of whichthe Las Vacas sample accumulated, may recordthe initial rifting or transcurrent transport of thePrecordillera terrane from its original position innorthern Gondwana. Its post-Late Ordovician ar-rival at high paleolatitudes, adjacent to the Fama-tina system, would be consistent with the paleo-biogeography of Mid Ordovician graptolites.Although generally cosmopolitan, these plankton-ic organisms displayed marked latitudinal provin-

cialism in the Mid Ordovician [47,48], with thediverse, low-latitude, Paci¢c province fauna wellrepresented in the Precordillera stratigraphic suc-cessions in sharp contrast to the high-latitude At-lantic province fauna in Ordovician strata of theEastern Cordillera of Argentina, Bolivia, andPeru [47,49]. A low-latitude position for the Pre-cordillera in the Mid Ordovician would be consis-tent with graptolite paleobiogeography, eliminat-ing the need to invoke long-lived hypotheticaloceanic currents to explain the present close geo-graphic juxtaposition of the Atlantic and Paci¢cfaunas in South America [4,47].

Our proposed paleogeographic placement ofthe Precordillera terrane has antecedents. In1989, Baldis et al. [16] proposed a similar modelin which the Precordillera terrane was displacedalong a transcurrent fault from a Mid Ordovicianlocation in northeast Gondwana to a ¢nal posi-tion adjacent to the Famatina system in the LateSilurian. In the Baldis et al. model, thick succes-sions of extrabasinal siliciclastic sediment of LateOrdovician and Silurian age were delivered suc-cessively from Gondwana to the Precordillera ter-rane as it passed along the proto-Andean marginof Gondwana. Acen‹olaza and Toselli [50] alsoproposed a similar Gondwanan model, andmore recently Acen‹olaza et al. [51] argued that,because many of the Laurentian faunal elementsin the Precordillera also occur in Australia, Ant-arctica, and on other paleo-plates, the paleobio-geography of benthic fossils of the Precordilleradoes not preclude a northern Gondwana location.

6. Conclusions

Detrital zircon populations in the La Ce¤bilasample are typical of a Gondwanan provenance.They are characteristic of (1) Brasiliano/Pan-Afri-can orogenic belts, (2) Grenvillian age-equivalentterranes of Gondwanan, not Laurentian, charac-ter, and (3) Transamazonian/Birimian-age ter-ranes. This is consistent with a depositional set-ting inboard of, and close to, the proto-Andeanmargin of Gondwana. The abundance of 480^520Ma grains is considered as evidence that theLa Ce¤bila Formation accumulated in a sedimen-

EPSL 6476 3-1-03

S. Finney et al. / Earth and Planetary Science Letters 205 (2003) 349^359356

tary basin inboard of, and close to, the Famatinaarc.

Our sandstone sample from the Cerro TotoraFormation has a distinctly Gondwanan prove-nance with two prominent detrital zircon popula-tions regarded as Brasiliano/Pan-African andTransamazonian/Birimian/Eburnian. Given thegeologic setting of the Cerro Totora Formation,the source of zircons must either be the basementof the Precordillera terrane or Gondwana proper,but not Laurentia. If it is the basement of thePrecordillera, then the basement rocks are ofGondwanan a⁄nity. If it is Gondwana proper,then the Precordillera must have been adjacentto Gondwana in Early Cambrian time.

The detrital zircon population in the Las Vacassample is Grenvillian-age equivalent, but the pop-ulation is very similar to that of the La Ce¤bilasample and signi¢cantly di¡erent from a Lauren-tian population (Fig. 2). Thus, it is evidence of aGondwanan provenance. Because the quartzsandstone of the Las Vacas sample is within astratigraphic sequence of conglomerates com-posed in part of extrabasinal igneous and sedi-mentary clasts with a source to the east, the prov-enance of the sandstone was to the east. Thesource area was not the basement rocks presentlyexposed in the westernmost Sierras Pampeanasbecause of the scarcity of metamorphic clasts inthe Las Vacas conglomerates. Nor was it the Fa-matina arc because of the lack of age-equivalentdetrital zircons. Both the westernmost SierrasPampeanas and the Sierra de Famatina are di-rectly east of the Precordillera today. But, inLate Ordovician time, the Precordillera musthave been removed from its present position rel-ative to these mountain ranges, yet still in positionto receive sediment dispersed from Gondwananbasement rocks.

The individual zircon grains of the Cerro To-tora and Las Vacas samples represent 174 samplesof exposed basement (plutonic) rock incorporatedat various stages into the sedimentary system.More samples are needed to test and fully evalu-ate our results. Nevertheless, given their nature,our results require critical re-evaluation of widelyheld assumptions regarding the paleogeographyof the Argentine Precordillera.

Acknowledgements

Acknowledgement is made to the Donors of thePetroleum Research Fund, administered by theAmerican Chemical Society, for support of thisresearch. The Radiogenic Isotope Lab at the Uni-versity of Arizona funded sample preparation andanalysis. Universidad Nacional de San Juan pro-vided ¢eld vehicle and equipment support. PaulHo¡man, Warren Hu¡, A.M. Ce“lal Sengo«r, andEd Stump provided valuable reviews.[SK]

References

[1] R.A. Astini, J.L. Benedetto, N.E. Vaccari, The early Pa-leozoic evolution of the Argentine Precordillera, Geol.Soc. Am. Bull. 107 (1995) 253^273.

[2] I. Dalziel, L. DallaSalda, C. Cingolani, P. Palmer, TheArgentine Precordillera: a Laurentian Terrane?, GSA To-day 6 (1996) 16^18.

[3] W.A. Thomas, R.A. Astini, The Argentine Precordillera:a traveler from the Ouachita embayment of North Amer-ican Laurentia, Science 273 (1996) 752^757.

[4] I.W.D. Dalziel, Neoproterozoic-Paleozoic geography andtectonics: review, hypothesis, environmental speculation,Geol. Soc. Am. Bull. 109 (1997) 16^42.

[5] R.J. Pankhurst, C.W. Rapela, (Eds.), The Proto-AndeanMargin of Gondwana, Geol. Soc. London, Spec. Publ.142 (1998) 1^383.

[6] C.W. Rapela, R.J. Pankhurst, C. Casquet, E. Baldo, J.Saavedra, C. Galindo, Early evolution of the Proto-An-dean margin of South America, Geology 26 (1998) 707^710.

[7] R.A. Astini, Stratigraphic evidence supporting the rifting,drifting and collision of the Laurentian Precordillera ter-rane of western Argentina, in: R.J. Pankhurst, C.W. Ra-pela (Eds.), The Proto-Andean Margin of Gondwana,Geol. Soc. London, Spec. Publ. 142 (1998) 11^33.

[8] V.A. Ramos, R.D. Dallmeyer, G. Vujovich, Time con-straints on the Early Palaeozoic docking of the Precordil-lera, central Argentina, in: R.J. Pankhurst, C.W. Rapela(Eds.), The Proto-Andean Margin of Gondwana, Geol.Soc. London, Spec. Publ. 142 (1998) 143^158.

[9] V.A. Ramos, J.D. Keppie, (Eds.), Laurentia-GondwanaConnections before Pangaea, Geol. Soc. Am. Spec. Pap.336 (1999) 1^276.

[10] M. Keller, Argentine Precordillera: Sedimentary andPlate Tectonic History of a Laurentian Crustal Fragmentin South America, Geol. Soc. Am. Spec. Pap. 341 (1999)1^131.

[11] R.A. Astini, W.A. Thomas, Origin and evolution of thePrecordillera terrane of western Argentina: a drifted Lau-rentian orphan, in: V.A. Ramos, J.D. Keppie (Eds.), Lau-

EPSL 6476 3-1-03

S. Finney et al. / Earth and Planetary Science Letters 205 (2003) 349^359 357

rentia-Gondwana Connections before Pangaea, Geol. Soc.Am. Spec. Pap. 336 (1999) 1^20.

[12] W.A. Thomas, R.A. Astini, G. Bayona, Ordovician colli-sion of the Argentine Precordillera with Gondwana, inde-pendent of Laurentian Taconic orogeny, Tectonophysics345 (2002) 131^152.

[13] L.H. DallaSalda, C.A. Cingolani, R. Varela, Early Paleo-zoic orogenic belt of the Andes in southwestern SouthAmerica: results of Laurentia-Gondwana collision?,Geology 20 (1992) 617^621.

[14] L.H. DallaSalda, I.W.D. Dalziel, C.A. Cingolani, R. Var-ela, Did the Taconic Appalachians continue into southernSouth America?, Geology 20 (1992) 1059^1062.

[15] C. MacNiocaill, B.A. van der Pluijm, R. Van der Voo,Ordovician paleogeography and the evolution of the Ia-petus ocean, Geology 25 (1997) 159^162.

[16] B.A. Baldis, S.H. Peralta, R. Villegas, Esquematizacionesde una posible transcurrencia del terrane de Precordilleracomo fragmento continental procedente de areas pampea-no-bonaerenses, Univ. Nac. Tucuma¤n Ser. Correl. Geol. 5(1989) 81^100.

[17] A.V. Borrello, The Cambrian of South America, in: D.H.Holland (Ed.), Cambrian of the New World, Wiley-Inter-science, London, 1971, pp. 385^438.

[18] O.L. Bordonaro, Biogeograf|¤a de trilobites Ca¤mbricos enla Precordillera Argentina, Univ. Nac. Tucuma¤n Ser. Cor-rel. Geol. 7 (1980) 25^30.

[19] N.E. Vaccari, Early Ordovician trilobite biogeography ofPrecordillera and Famatina, western Argentina: prelimi-nary results, in: J.D. Cooper, M.L. Droser, S.C. Finney(Eds.), Ordovician Odyssey: Short Papers for the SeventhInternational Symposium on the Ordovician System, Pa-ci¢c Section, SEPM (Society for Sedimentary Geology)Publ. 77 (1995) 193^196.

[20] O. Lehnert, J.F. Miller, J.E. Repetski, Paleogeographicsigni¢cance of Clavohamulus hintzei Miller (Conodonta)and other Ibexian conodonts in an early Paleozoic car-bonate platform facies of the Argentine Precordillera,Geol. Soc. Am. Bull. 109 (1997) 429^443.

[21] J.L. Benedetto, Early Paleozoic brachiopods and associ-ated shelly faunas from western Gondwana: their bearingon the geodynamic history of the pre-Andean margin, in:R.J. Pankhurst, C.W. Rapela (Eds.), The Proto-AndeanMargin of Gondwana, Geol. Soc. London, Spec. Publ.142 (1998) 57^83.

[22] S.M. Kay, S. Orrell, J.M. Abbruzzi, Zircon and wholerock Nd-Pb isotopic evidence for a Grenville age andLaurentian origin for the basement of the Precordilleraterrane in Argentina, J. Geol. 104 (1996) 637^648.

[23] J.S. Stacey, J.D. Kramers, Approximation of terrestriallead isotope evolution by a two-stage model, Earth Plan-et. Sci. Lett. 26 (1975) 207^221.

[24] B.D. Webby, Steps toward a global standard for Ordovi-cian stratigraphy, Newsl. Stratig. 36 (1998) 1^33.

[25] R.A. Cooper, The Ordovician time scale ^ calibration ofgraptolite and conodont zones, Acta Univ. Carol.-Geol.43 (1999) 1^4.

[26] S.M. Espizu¤a, R. Caminos, Las rocas metamor¢cas de laFormacio¤n La Ce¤bila, Sierra de Ambato, Provincias deCatamarca y La Rioja, Bol. Acad. Nac. Ci. Co¤rdoba 53(1979) 125^142.

[27] M. Ho«ckenreiner, Die Formation La Ce¤bila und ihr geo-logisches Umfeld (Sierra de Ambato, NW Argentinien),Thesis, Diplomarbeit Institut fu«r Angewandte Geologieder Ludwig-Maximilians-Universita«t, Munich, 1998, 96pp.

[28] M. Boher, W. Abouchami, A. Michard, F. Albarede,N.T. Arndt, Crustal Growth in West Africa at 2.1 Ga,J. Geophys. Res. 97 (1992) 345^369.

[29] B.B. BritoNeves, U.G. Cordani, Tectonic evolution ofSouth America during the Late Proterozoic, PrecambrianRes. 53 (1991) 23^40.

[30] B.B. BritoNeves, M.C. CamposNeto, R.A. Fuck, FromRodinia to Western Gondwana An approach to the Bra-siliano-Pan African cycle and orogenic collage, Episodes22 (1999) 155^166.

[31] A. Bernasconi, The major Precambrian terranes of easternSouth America: a study of their regional and chronolog-ical evolution, Precambrian Res. 37 (1987) 107^124.

[32] J. Klerkx, J.-P. Lie¤geois, J. Lavreau, W. Claessens, Crus-tal evolution of the northern Kibaran belt, eastern andcentral Africa, in: A. Kro«ner (Ed.), Proterozoic Litho-spheric Evolution, Am. Geophys. Union and Geol. Soc.Am., Geodyn. Ser. 17 (1987) 217^233.

[33] C.C.G. Tassinari, M.J.B. Jacambira, Geochronologicalprovinces of the Amazonian Craton, Episodes 22 (1999)174^182.

[34] S.M. Quenardelle, V.A. Ramos, Ordovician western Sier-ras Pampeanas magmatic belt: record of Precordillera ac-cretion in Argentina, in: V.A. Ramos, J.D. Keppie (Eds.),Laurentia-Gondwana Connections before Pangaea, Geol.Soc. Am. Spec. Pap. 336 (1999) 63^86.

[35] R.J. Pankhurst, C.W. Rapela, J. Saavedra, E. Baldo, J.Dahlquist, I. Pascua, C.M. Fanning, The Famatinianmagmatic arc in the central Sierras Pampeanas: an Earlyto Mid-Ordovician continental arc on the Gondwanamargin, in: R.J. Pankhurst, C.W. Rapela (Eds.), The Pro-to-Andean Margin of Gondwana, Geol. Soc. London,Spec. Publ. 142 (1998) 343^367.

[36] W.A. Thomas, R.A. Astini, R.E. Denison, Strontium iso-topes, age, and tectonic setting of Cambrian salinas alongthe rift and transform margins of the Argentine Precor-dillera and Southern Laurentia, J. Geol. 109 (2001) 231^246.

[37] R.A. Astini, N.E. Vaccari, Sucesio¤n evapor|¤tica del Ca¤m-brico Inferior de la Precordillera: signi¢cado geolo¤gico,Rev. Asoc. Geol. Argent. 51 (1996) 97^106.

[38] P.F. Ho¡man, Precambrian geology and tectonic historyof North America, in: A.W. Bally, A.R. Palmer (Eds.),The Geology of North America ^ An Overview, TheGeology of North America A, Geol. Soc. Am., Boulder,CO, 1989, pp. 447^512.

[39] P.A. Mueller, A.L. Heatherington, J.L. Wooden, R.D.Shuster, A.P. Nutman, I.S. Williams, Precambrian zircons

EPSL 6476 3-1-03

S. Finney et al. / Earth and Planetary Science Letters 205 (2003) 349^359358

from the Florida basement: a Gondwanan connection,Geology 22 (1994) 119^122.

[40] R.A. Astini, El Conglomerado de Las Vacas y el GrupoTrapiche de la Precordillera: tecto¤nica distensiva en elOrdov|¤cico Tard|¤o, Rev. Asoc. Geol. Argent. 53 (1998)489^503.

[41] S.H. Peralta, Estratigraf|¤a y consideraciones paleoamb|¤en-tales de los depo¤sitos marino-cla¤sticos eopaleozoicos de laPrecordillera Oriental de San Juan, XI Congreso Geolo¤-gico Argentino Actas I (1993) 128^137.

[42] M.R. McDonough, V.A. Ramos, C.E. Isachsen, S.A.Bowring, G.I. Vujovich, Nuevas edades de circones del ba-samento de la sierra de Pie de Palo, Sierras PampeanasOccidentales de San Juan: sus implicancias para los mod-elos del supercontinente proterozoico de Rodinia, XII Con-greso Geolo¤gico Argentino y II Congreso de Exploracio¤nde Hidrocarburos, Mendoza, Actas III (1993) 340^342.

[43] C.D. Wareham, R.J. Pankhurst, R.J. Thomas, B.C. Stor-ey, G.H. Grantham, J. Jacobs, B.M. Eglington, Pb, Nd,and Sr isotope mapping of Grenville-age crustal provincesin Rodinia, J. Geol. 106 (1998) 647^659.

[44] A.M. Sato, A.H. Tickyjb, E.J. Llamb|¤asa, K. Satoc, TheLas Matras tonalitic-trondhjemitic pluton, central Argen-tina: Grenvillian-age constraints, geochemical character-istics, and regional implication, J. S. Am. Earth Sci. 13(2000) 587^610.

[45] A.E. Rapalini, R.A. Astini, Paleomagnetic con¢rmationof the Laurentian origin of the Argentine Precordillera,Earth Planet. Sci. Lett. 155 (1998) 1^14.

[46] P.F. Ho¡man, Did the breakout of Laurentia turn Gond-wanaland inside-out?, Science 252 (1991) 1409^1412.

[47] S.C. Finney, X. Chen, The relationship of Ordoviciangraptolite provincialism to palaeogeography, in: W.S.

McKerrow, C.R. Scotese (Eds.), Palaeozoic Palaeogeog-raphy and Biogeography, Geol. Soc. Mem. 12 (1990) 123^128.

[48] R.A. Cooper, R.A. Fortey, K. Lindholm, Latitudinal anddepth zonation of early Ordovician graptolites, Lethaia 24(1991) 199^218.

[49] C.E. Mitchell, E.D. Brussa, R.A. Astini, Biogeography ofMiddle and Upper Ordovician graptolites, Precordilleraterrane, Argentina: plate tectonic implications, Geol.Soc. Am. Abstr. with Progr. 29 (1997) A379.

[50] F.G. Acen‹olaza, A.J. Toselli, Argentine Precordillera: al-lochthonous or autochthonous Gondwanic?, Zbl. Geol.Pala«ontol. Teil I 7/8 (1999) 743^756.

[51] F.G. Acen‹olaza, H. Miller, A.J. Toselli, Proterozoic-EarlyPaleozoic evolution in western South America ^ A discus-sion, Tectonophysics 354 (2002) 121^137.

[52] K.A. Eriksson, I.H. Campbell, J.M. Palin, C.M. Allen,The Grenville superorogeny revealed by detrital zirconsin Appalachian rivers, Geol. Soc. Am. Abstr. with Progr.33 (2001) A29.

[53] J.D. Gleason, S.C. Finney, G.G. Gehrels, Paleotectonicimplications of a Mid- to Late-Ordovician provenanceshift, as recorded in sedimentary strata of the Ouachitaand Southern Appalachian Mountains, J. Geol. 110(2002) 291^304.

[54] J.D. Gleason, G.G. Gehrels, S.C. Finney, Tectonic recy-cling in the Paleozoic Ouachita Assemblage from U-Pbdetrital zircon studies, EOS 82 (2001) 435.

[55] J.H. Stewart, G.E. Gehrels, A.P. Barth, P.K. Link, N.Christie-Blick, C.T. Wrucke, Detrital zircon provenanceof Mesoproterozoic to Cambrian arenites in the westernUnited States and northwestern Mexico, Geol. Soc. Am.Bull. 113 (2001) 1343^1356.

EPSL 6476 3-1-03

S. Finney et al. / Earth and Planetary Science Letters 205 (2003) 349^359 359