Magnetic Polarity Stratigraphy and Biostratigraphy of Paleocene and Lower Eocene Continental Deposits, Clark's Fork Basin, Wyoming Robert F

University of Portland Pilot Scholars Environmental Studies Faculty Publications and Environmental Studies Presentations

5-1981 Magnetic Polarity Stratigraphy and Biostratigraphy of Paleocene and Lower Eocene Continental Deposits, Clark's Fork Basin, Wyoming Robert F. Butler University of Portland, [email protected]

Philip D. Gingerich

Everett H. Lindsay

Follow this and additional works at: http://pilotscholars.up.edu/env_facpubs Part of the Earth Sciences Commons

Citation: Pilot Scholars Version (Modified MLA Style) Butler, Robert F.; Gingerich, Philip D.; and Lindsay, Everett H., "Magnetic Polarity Stratigraphy and Biostratigraphy of Paleocene and Lower Eocene Continental Deposits, Clark's Fork Basin, Wyoming" (1981). Environmental Studies Faculty Publications and Presentations. 3. http://pilotscholars.up.edu/env_facpubs/3

This Journal Article is brought to you for free and open access by the Environmental Studies at Pilot Scholars. It has been accepted for inclusion in Environmental Studies Faculty Publications and Presentations by an authorized administrator of Pilot Scholars. For more information, please contact [email protected]. MAGNETIC POLARITY STRATIGRAPHY AND 40K-40AR DATING OF LATE MIOCENE AND EARLY PLIOCENE CONTINENTAL DEPOSITS, CATAMARCA PROVINCE, NW ARGENTINA'

ROBERT F. BUTLER, LARRY G. MARSHALL, ROBERT E. DRAKE, AND GARNISS H. CURTIS Departmentof Geosciences, The Universityof Arizona, Tucson, Arizona85721 Departmentof Geology and Geophysics, Universityof California,Berkeley, California94720 ABSTRACT Magnetostratigraphicand 40K- 40Ardata on a 2300 m thick sequenceof continentalsediments at Puerta de CorralQuemado in CatamarcaProvince, NW Argentinapermit calibration of landmammal faunas of late Tertiary(Huayquerian and Montehermosan)age. The sequence represents(from oldest to youngest) the ChiquimilA, Araucanense,and CorralQuemado Formations. Paleomagnetic samples were collected from 99 stratigraphiclevels. Strong-fieldthermomagnetic and isothermalremanent magnetization experiments indicatethat the dominantferrimagnetic mineral is magnetite.Progressive alternating-field (AF) and themal demagnetizationof the naturalremanent magnetism (NRM) demonstratesthat AF demagnetizationto 20 mT peak field is sufficient to isolate the primaryNRM which is of depositionalorigin. The resulting paleomagneticdata provide a well-definedmagnetic polarity zonation, although sampling is less dense in the upperhalf of the section. 40K-40Ar data obtainedfrom mineralseparates of four tuffs withinthe section allow reliableage determinationsfor those levels. The combinedmagnetostratigraphic and 40K- 40Ardata allow the magneticpolarity zonation to be correlatedwith the magneticpolarity time scale. This correlation indicates a nearly constant rate of sediment accumulationbetween -8.0 Ma and 3.5 Ma. The boundary between the Araucanenseand CorralQuemado Formations approximates the boundarybetween the Huay- querianand Montehermosanland mammalage faunas at this locality. The data presentedhere allow the boundarybetween the Araucanenseand CorralQuemado Formations to be datedat 6.4 Ma. Combinedwith geochronologicdata from similarage rocks and faunasfrom San Carlos, MendozaProvince, west-central Argentina,the geochronologicdata from Puerta de CorralQuemado allow the Huayquerian-Monteher- mosanland mammal age boundaryto be placedtentatively at 6.0 Ma. A specimenof the fossil land mammal Cyonasua(family Procyonidae) from unit 14 of the AraucanenseFormation is dated at 7.0 to 7.5 Ma. This specimenis the earliest known representativeof this North Americangroup in South Americaand represents the oldest dated participantin the GreatAmerican Faunal Interchange on that continent.

INTRODUCTION the appearance of the land bridge and in- The continents of North and South volved most of the interchange participants, America are today connected by the Panama- which dispersed north and/or south after 3.0 nian land bridge, a structure which has per- Ma. The other phase occurred prior to emer- mitted the reciprocal interchange of terres- gence of the land bridge and involved fewer trial biotas following its final emergence participants. The land mammals in this phase about 3.0 Ma. This biotic event is known as were waif immigrants in the Late Miocene the Great American Faunal Interchange and dispersed on rafts of vegetation and/or by (Webb 1976), and it represents the best docu- island hopping along island chains. The par- mented example in the fossil record of an in- ticipants are known to include members of terchange of two long separated continental the North American raccoon family Procy- faunas. onidae (and possibly the rodent family Two primary phases of the Great American Cricetidae), which dispersed to South Amer- Faunal Interchange are recognized, the par- ica, and members of the ground sloth families ticipants in which are identified on the basis Megalonychidae and Mylodontidae, which of their time and means of dispersal (Simpson dispersed to North America. In this paper we 1940, 1950, 1980). One phase occurred after provide geochronologic data that securely calibrates the beginning of this early phase of the Great American Faunal Interchange in 1 Manuscript received February 22, 1984; revised June 11, 1984. South America. Late Tertiary land mammal faunas which [JOURNALOF GEOLOGY, 1984, vol. 92, p. 623-636] © 1984 by The University of Chicago. All rights include fossil Procyonidae have been re- reserved. corded from a thick sedimentary sequence 0022-1376/84/9206-001$1.00 at Puerta de Corral Quemado and Chiquimil

623

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions 624 R. F. BUTLER ET AL.

FIG. 1.-Map of Catamarca Province, northwest Argentina showing Puerta de Corral Quemado and Chiquimil localities. in CatamarcaProvince, northwestArgentina corded 1,525 m of sediment and that from (fig. 1). A largecollection of fossil vertebrates Puerta de Corral Qeumado 1,913 m. Fossil from these localities was made by Elmer S. mammalsand tuff beds were indicatedto be Riggs in 1926 and is deposited in the Field especially abundantin the upper half of each Museum of Natural History, Chicago. De- section. tailed stratigraphicprofiles of each locality These published data indicated potential were made by RudolphStahlecker, geologist for calibrating the age of these mammal- on the Riggs' expedition, and recordwas kept bearingrocks using magnetostratigraphicand of the position of most fossils collected within 40K- 40Ar dating techniques. In May 1977, this stratigraphic framework. Stahlecker's Butler, Marshall, and personnel from the sections were publishedby Riggs and Patter- Museo de La Plata and Museo Municipalde son (1939) along with a preliminarystudy of Ciencias Naturales "Lorenzo Scaglia" in the faunas. The section from Chiquimilre- Argentina visited these localities to deter-

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions MAGNETIC POLARITY STRATIGRAPHY 625 mine the feasibility of such a study. The re- results of that trip are reportedhere. A 2,300 sults of that trip are reported by Marshall et m thick composite paleomagneticsection, in- al. (1979). The rocks at Puerta de Corral cluding the two sections collected in 1977, Quemado proved particularly suitable for was obtained. This section correlates with paleomagnetic analysis, and two sections, the upper half of Stahlecker's section, span- each including a securely dated tuff, were ning his units 13 through30 (see below) and sampled. The dated tuff in the upper part of includes four securely dated tuffs. We our lower 400 + m section we believed to be confirmedthe identityof the tuff in our upper Stahlecker's unit 8, and the dated tuff in the section of Marshallet al. (1979) as Stahleck- top of our upper 500 + m section we believed er's unit 29, but found that unit 8 in the lower to be his unit 29. Correlation of our two sec- section of Marshall et al. (1979) is actually tions with units in Stahlecker's section was Stahlecker's unit 15. The unit identified as based largely on the identity of these tuff unit 15 by Marshallet al. (1979) cannot be levels. securely correlated with a specific unit in Our preliminary study demonstrated the Stahlecker's section, although we believe ideal nature of these rocks for magnetostrati- that unit is correlativewith Stahlecker'sunit graphic and 40K- 40Ar analyses. We felt 20. confident that additional work at Puerta de Stratigraphicnomenclature of the late Ter- Corral Quemado would permit us to sample tiary sedimentsat Puertade CorralQuemado the sequence between our upper and lower and Chiquimilhas had a confused history as sections. The anticipated results would pro- reviewed by Marshall and Patterson (1981, vide a more secure correlation of our data pp. 14-15) and Bossi and Palma(1982, fig. 1). with Stahlecker's section and would permit We follow the terminologyused by Riggs and chronostratigraphic calibration of the fossil Patterson (1939) because it is the one which faunas, including three specimens of Procy- appears most frequently in paleontological onidae. The mammal faunas are assigned to literature. the Huayquerian and Montehermosan Land The following abbreviations are used: Mammal Ages (see below). Thus, the oppor- FMNH, Field Museum of Natural History, tunity existed to date the Huayquerian-Mon- Chicago;MACN, Museo Argentinode Cien- tehermosan boundary at this locality. cias Naturales "Bernardino Rivadavia," An unpublished manuscript by Stahlecker Buenos Aires, Argentina; Ma, millions of on the geology of Puerta de Corral Quemado years before present. and Chiquimil was discovered in the archives of Field Museum in late 1979 and published in PALEOMAGNETISM AND ROCK MAGNETISM Marshall and Patterson (1981). Subsequent Oriented samples were collected from 99 study of this manuscript revealed to us that paleomagneticsites (3 sample/site)in a 2300 the stratigraphic thickness of Stahlecker's m thick stratigraphicsection at Puerta de sections as published by Riggs and Patterson Corral Quemado (fig. 1). This section was (1939) were erroneously off by a factor of located on the west limb of a gently plunging two. Thus, his Puerta de Corral Quemado anticline.Samples for paleomagneticanalysis section is not 1,913 m thick, as published in were preferentiallycollected from claystones Riggs and Patterson (1939), but is closer to and siltstones ratherthan sandstones. In has 4000 m. In addition, no detailed descriptions generallybeen observed (e.g., Johnson et al. of the lithologies were published by Riggs and 1975) that finer-grainedcontinental deposits Patterson (1939). Additional field work was yield more reliable paleomagneticresults. needed to clarify aspects of the geology re- Magnetic separates for thermomagnetic ported in Stahlecker's manuscript, and to ex- analysis were prepared from bulk samples pand the geochronological work begun in collected at three stratigraphiclevels. An ex- 1977. ample of the results of the thermomagnetic In September 1980, Butler, Drake, and analyses is shown in figure2. All thermomag- Marshall revisited Puerta de Corral Quemado netic analyses were similarin indicatingonly (with Stahlecker's unpublished manuscript in a single Curie temperatureof approximately hand) and collected additional rock samples 580°C. This observation indicates that mag- for paleomagnetic and 40K-40Ar study. The netite (Fe304) is the dominantferrimagnetic

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions 626 R. F. BUTLER ET AL.

u 100 200 300 400 500 600 Temperature(C) FIG.2.-Strong-field thermomagnetic results on magneticseparate. Magnetizing field was 300mT. Heatingand cooling curves are indicated by arrows directionof temperaturechange. Sample showing O 100 200 300 400 500 600 chamberwas evacuated then back-filled with argon MagneticField (mT) gas. Heatingand cooling rate was 15°C/min. FIG. 3.-Acquisition of isothermal remanent magnetizationas a functionof magnetizingfield. mineral.The heatingand cooling curves were very similarin all cases, suggestingabsence of low temperatureoxidation. Single samples from 20 sites in the section sive alternating-field(AF) and thermal de- were selected for study of acquisition of magnetizationof the NRM are illustratedin isothermal remanent magnetization (IRM). figures4 and 5, respectively. In many cases, Typicalresults are illustratedin figure3. IRM no systematic directional change was ob- is acquiredrapidly in magnetizingfields up to served duringeither progressive AF or ther- 300 millitesla (mT; 1 mT = 10 oe), but only mal demagnetization,and no significantsec- minoramounts of IRM are acquiredin higher ondary components are present in these intensity magnetizingfields. These observa- samples. Where present (e.g., fig. 4b), sections are consistent with the thermomagnetic ondary components were adequately erased results indicatingmagnetite as the dominant by AF demagnetizationto 20 mT peak field. ferrimagneticmineral. The parameterS is de- No secondary components were erased by fined as (IRM(600mT)- IRM(300mT))/IRM thermaldemagnetization which were not also (300mT)where IRM(H)is the IRM acquired erased by AF demagnetizationto 20 mT peak in magnetizingfield H (Butler 1982).This pa- field. Thus, the blanket demagnetization rameteris a measureof the ratio of high coer- treatmentchosen was AF demagnetizationat civity phases (such as hematite)to low coer- 20 mT. The primary NRM component has civity phases (such as magnetite).Because it coercivity dominantly<80 mT and blocking is advisable to avoid sediments which have temperaturedistribution dominantly <5750C. suffered post-depositional oxidation with These observationsindicate that the primary possible attendant secondary components NRM is a depositionalremanent magnetiza- and remanentmagnetization, low values of 8 tion (DRM)carried by the detritalmagnetite. are desirable. The § values observed in the Following AF demagnetizationto 20 mT, Puertade CorralQuemado sediments are low the clusteringof NRM directionsfor each site for continental sedimentary rocks (Butler was excellent. All but four of the 99 sites 1982). For all but one specimen, the 8 values show within-site clustering which is signifi- are <0.1, with an average § value of 0.028. cant from randomat the 95%confidence level Thus, only minor amounts of high coercivity (Watson 1956). For N = 3, this requires R mineralssuch as hematitecould be present in >2.62. Given the uncomplicatednature of the these rocks. NRM and the well-determinedsite mean di- The natural remanent magnetism (NRM) rections, the resulting determinationsof po- intensities were strong, rangingfrom 5.2 x larityof the primaryNRM were in most cases 10-3 to 5.3 x 10-1 A/m. Results of progres- unambiguous.

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions MAGNETIC POLARITY STRATIGRAPHY 627

Up,North

(b) Up,NT Up,North

CQ101B (b)

(c)

CQ160B

a Inclination

CQ153A East Down,South -=2.0x102A/m FIG. 5.-Vector demagnetizationdiagrams of a Inclination progressivethermal demagnetization behaviors of three representative Symbols in a Declination samples. as figure 4 except numbersadjacent to data points indicate temperaturein °C. FIG. 4.-Vector demagnetization plots of pro- gressivealternating-field demagnetization behaviors of three representativesamples. Projectionof fold test indicates the anticipated result that NRM into the horizontal plane is indicated by the primary NRM is prefolding in origin. triangles, vertical versus horizontalplot is shown by squares. Numbers adjacent to the data points An interesting feature of the distribution of indicatethe peak alternating-fieldin mT. site mean directions shown in figure 6 and tabulated in table 1 is that the mean direction after structural correction is deflected in a clockwise sense from the axial geocentric di- Althoughthe paleomagneticsamples were pole field direction (declination = 0°, inclina- collected from only one limb of the plunging tion = -45.90). Although it is possible that anticline, variations in structural attitude this observation could indicate regional tec- within this limb are sufficientto allow a fold tonic rotation, we believe such a conclusion test to be performed.Results of the fold test would be premature. Sufficient structural are tabulated in table 1 and illustrated in data do not exist to allow application of a figure 6. Although the improvementin clus- plunge correction, so the "structural correc- tering of site mean vectors following struc- tion" was done only about the local strike tural correction is not visually dramatic,the line. It is certainly possible that the clockwise increasein the k value producedby the struc- deflection of the mean direction from the tural correction is significant at the 99% axial geocentric dipole direction is simply a confidence level (Watson 1956). Thus, the result of the incomplete structural correc-

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions 628 R. F. BUTLER ET AL.

TABLE 1

AVERAGES OF SITE MEAN DIRECTIONS BEFORE AND AFTER STRUCTURAL CORRECTION

Inclination Declination a95% (o) (o) Na Rb kb (o)

Before structural 38.0 7.8 99 83.0 5.81 6.4 correction After structural 46.9 26.6 99 89.4 10.22 4.7 correction

NOTE.-In computing average direction, antipodes of reversed polarity sites are averaged with normal polarity sites. a N = number of directions averaged. b Statistical parameters of Fisher (1953). tions applied to the cleaned NRM directions. components are easily cleaned. The primary However, this possibility of incomplete NRM is carried by detrital magnetite and is structural correction would not affect the fold clearly depositional in origin. Thus, the test. The section collected was of such a lim- stratigraphic sequence of cleaned NRM di- ited spacial extent with respect to the dimen- rections should provide an accurate record of sions of the anticlinal structure that a single the geomagnetic polarity sequence during de- plunge correction would have to be applied to position. all sites in the section. It is clear that these sediments provide un- Another interesting feature of the mean di- usually reliable paleomagnetic data. Unfortu- rection of cleaned NRM is the mean incli- nately, the sedimentology of these rocks has nation. The observed mean inclination of not been studied in any detail. Thus little in- -46.90 is close to the axial geocentric dipole formation is available regarding provenance, direction of -45.9°. Thus, no inclination er- sedimentary environment, or diagenesis. It is ror is evident in the NRM of these continental clear, however, that the dominant deposi- sediments. tional environment is fluviatile and that the The rock-magnetic and paleomagnetic data finer-grained sediments are overbank depos- indicate that these sedimentary rocks contain its rich in detrital magnetite. We also ob- an unusually strong and stable NRM. The served abundant carbonate cement in many NRM is almost entirely devoid of secondary of the sediments. It is possible that this components, and, where present, secondary cement decreased permeability of the sedi-

* Lowerhemisphere OUpper hemisphere * Lowerhemisphere o Upperhemisphere FIG. 6.-Equal area stereographic projections of site mean directions of NRM (a) before and (b) after structural correction. Squares indicate the average directions for the normal and reversed polarity directions. Circle of 95% confidence surrounds each mean direction.

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions MAGNETIC POLARITY STRATIGRAPHY 629 ments, preventing post-depositional oxida- trates (KA 3307, 3307R1, 3307R2, see table 2) tion of the detrital magnetite or post- from LGM 2 yielded reproducible results. depositional precipitation of other magnetic The average age of these three sanidine dates phases. These conditions could in part ac- based on current decay constants is 6.70 count for the quality of the paleomagnetic ± 0.05 Ma (fig. 7). data obtained from these continental sedi- The highest tuff sampled in the section is ments. from unit 29 of Stahlecker (fig. 7), a 1 m thick tuffaceous sand. Mineral separates of biotite, POTASSIUM-ARGON AGE DETERMINATIONS plagioclase, and sanidine were dated (Mar- Tuffs are abundant in the Corral Quemado shall et al. 1979, table 1). Recalculation of the section, especially in the upper part (above dates for two biotite (KA 3278, 3285), one unit 14) of the section measured by Stah- plagioclase (KA 3438), and one sanidine (KA lecker (in Riggs and Patterson 1939; Marshall 3343) age using improved data reduction and Patterson 1981). These tuffs apparently techniques gave an average age of 3.48 Ma originated from ash-flow eruptions in the ad- (table 2). The date of 3.53 ±0.04 Ma on the jacent Andean Cordillera, and some may sanidine concentrate (KA 3343) is technically prove to be identified with eruption events of the best (fig. 7), and the biotite and plagio- nearby calderas (see Francis and Baker 1978; clase dates are concordant within their ana- Francis et al. 1978). They vary from several lytical errors. The concordant ages obtained meter thick coarse sands rich in biotite, glass, for these three mineral separates suggests plagioclase, and sanidine to several centime- that this tuff, as opposed to those below it, ter thick water-worked sands and silts. was apparently uncontaminated by older de- The primary deposition of these tuffs was trital material. in subaerial environments. Subsequent re- In 1980, nine additional tuffs between units working by wind and water resulted in admix- 15 and 29 were sampled. Six of these (LGM ing of detrital material derived from nearby 213, 214, 216, 219, 220, 223) contained biotite Precambrian and Paleozoic crystalline base- and/or glass suitable for dating (table 2). Of ment (Ruiz Huidobro 1972; Allmendinger et these, only glass concentrates from LGM 216 al. 1983; Jordan et al. 1983). Preliminary and 220 yielded reproducible and technically 40K- 40Arage determinations on mineral sep- precise dates. Two results (KA 4099, 4099R2) arates from some of these tuffs yielded mark- from LGM 216 yielded an average age of 5.3 edly discordant dates due to this contamina- +0.2 Ma (fig. 7), and two others (KA 4368, tion. Further study demonstrated that only 4368R) from LGM 220 yield an average age of mineral concentrates carefully hand-picked 4.95 ± 0.14 Ma (fig. 7). All biotite dates to remove detrital grains gave reproducible proved unreliable due to contamination, and hence reliable, age estimates (see Marshall et the other glasses gave imprecise results due al. 1979). to high concentrations of atmospheric 40Ar Mineral separates from two tuffs sampled trapped in vesicles. LGM 216 apparently cor- in 1977 and reported by Marshall et al. (1979) relates with part of Stahlecker's unit 23, and yielded reliable age determinations permit- LGM 220 with part of his unit 24. ting calibration of the upper and lower limits of the 2300 m thick section shown in figure 7. DISCUSSION The lowest tuff sampled in our measured sec- Geochronology. -The paleomagnetic, tion is from unit 15 of Stahlecker (= unit 8 of 40K-40Ar, and paleontologic data are illus- Stahlecker in Marshall et al. 1979; Marshall trated in stratigraphic context in figure 7. The and Patterson 1981). This is the most promi- magnetic polarity zonation determined from nent tuffaceous unit in the section and con- the paleomagnetic data is shown along with sists of 3 m of coarse white crystal-rich sand its correlation to the magnetic polarity time overlain by 2.5 m of tan and 6 m of dark green scale of Ness et al. (1980). Abundant siltstone tuffaceous sand. Sample LGM 3 is from the layers are interbedded with the predominant basal eolian sand, and LGM 2 is from the sandstone in the bottom half of the section. overlying tan sand. Biotite, plagioclase, and Thus, paleomagnetic sampling in this part of sanidine concentrates were dated from both the section was quite dense, and the resulting samples, but only the three sanidine concen- magnetic polarity zonation is defined with

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions 630 R. F. BUTLER ET AL. Stratigraphic VGP Magnetic Magnetic Thickness Lithology Latitude Polarity Polarity Ft. M Zonation TimeScale

Age(Mo)

-90° 0 +90° FIG.7.-Paleomagnetic, 40K- 40Ar,and paleontologicdata as a functionof stratigraphicposition. Site average virtual geomagnetic pole (VGP) latitude, interpretedmagnetic polarity column, and lithologic column are plotted against stratigraphicthickness. Solid symbols indicatesites with within-siteclustering which is significantfrom randomat the 95%confidence level while open symbolsindicate sites with poorer within-siteclustering. Black bars in polaritycolumn indicate normal polarity, white bars indicatereversed polarity. Correlationof magneticpolarity zones with the magneticpolarity time scale of Ness, Levi, and Couch (1980)is illustratedat the rightside of the diagram.The positionof fossil specimensof Procyonidae in the section are shown.

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions MAGNETIC POLARITY STRATIGRAPHY 631

TABLE 2

ANALYTICAL DATA FOR 40K- 40AR DATES OF MINERAL SEPARATES OF TUFFS FROM CORRAL QUEMADO, CATAMARCA PROVINCE, ARGENTINA, IN STRATIGRAPHIC ORDER (YOUNGEST TO OLDEST)

Sample 40Ara Collection Sample Dated Weight K (x 10-11 40Ara Age ± 2r Number Number Material (grams) (%) mole/gram) (%) (Ma) (LGM)

KA 3278 biotite 2.7520 4.281 2.59 9.6 3.49 ± .2 4A KA 3285 biotite 1.3556 4.648 2.76 9.4 3.42 + .6 4A KA 3343 sanidine 4.6977 4.783 2.93 83.2 3.53 ± .04 4A KA 3438 plagioclase 1.9678 .416 .25 4.9 3.46 _ .5 4A KA 4392 glass 2.8786 4.323 4.61 6.2 6.1 ± 1.3 223 KA 4393 biotite 1.37 4.934 4.17 10.0 4.9 ± .4 223 KA 4368 glass 3.0822 4.445 3.77 31.6 4.88 _ .13 220 KA 4368R glass 2.9131 4.445 3.88 23.4 5.02 _ .15 220 KA 4367 glass 1.2535 4.723 5.64 9.8 6.9 ± .5 219 KA 4099 glass 5.44 5.072 4.62 58.2 5.2 _ .3 216 KA 4099Ra,b glass 3.0018 5.060 5.15 35.2 5.87 ± .3 216 KA 4099R2 glass 2.58 5.060 4.77 61.7 5.4 + .2 216 KA 4103 biotite .5498 6.642 11.98 36.0 10.4 + .2 216 KA 4095 biotite .4418 6.540 6.97 8.5 6.14 ± 1.6 214 KA 4373 biotite 1.7327 6.351 28.66 76.0 25.1 ± 1.5 214 KA 4098 biotite .3372 6.711 25.41 35.2 21.7 ± .6 213 KA 3307 sanidine 2.0359 9.477 11.07 91.2 6.73 ± .09 2 KA 3307R1 sanidine 2.0058 9.477 10.95 68.2 6.67 ± .12 2 KA 3307R2 sanidine 3.3533 9.549 11.10 71.3 6.69 ± .09 2

NoTE.-Calculations are based on the radio-active decay constants for 40KXp= 4.962 x 10 10yr ' and he + Xe' = 0.581 x 10- 0 yr ' and on the isotopic abundance 4K = 0.01167% of total K. a Radiogenic 40Ar. high fidelity. In the upper half of the section, tion, these correlations are shown in figure 7 however, the sampling density is generally by solid lines. The less secure correlations in lower with some significant gaps between the upper part of the section are shown by paleomagnetic sites. This is a reflection of the dashed lines and were strongly influenced by lower abundance of siltstone interbeds. The the 40K- 40Ar ages. resulting magnetic polarity zonation in the Two problems with the correlations shown upper part of the section is correspondingly in figure 7 are evident. Within the normal less well defined. It is fortunate that this por- polarity zone is a thin reversed polarity zone. tion of the section contains three tuff levels The normal zone is correlated with the older from which 40K- 40Ar analyses have pro- normal polarity interval of chron 5. No corre- vided reliable age determinations (see above, sponding short-duration reversed polarity in- fig. 7). terval is present in the magnetic polarity time The correlation of the magnetic polarity zo- scale. It is possible that the high sediment nation with the magnetic polarity time scale accumulation rate in this section allowed re- integrates the paleomagnetic and 40K-40Ar cording of a short duration reversed polarity data. Both the pattern of polarity zonation interval that is not included in the magnetic and 40K- 40Ar ages were used to accomplish polarity time scale. The correlation also im- this correlation. The correlation of polarity plies that the reversed polarity interval be- zones with polarity intervals of the magnetic tween the two older normal polarity events in polarity time scale is considerably more se- the Gilbert epoch is not recorded in the mag- cure for the lower half of the section than for netic polarity zonation. However, the sam- the upper half. To some degree this is a result pling interval would appear to have been of the fidelity with which the magnetic polar- sufficient to anticipate having recorded this ity zonation could be determined. Because polarity interval. Despite these problems, the the correlations of magnetic polarity zones overall match of the magnetic polarity zona- with the magnetic polarity time scale inter- tion with the magnetic polarity time scale is vals are firm in the lower portion of the sec- good. The goodness of the match can be ob-

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions 632 R. F. BUTLER ET AL.

Ness, Couch Ft. M Levi,and (1980) 8000- 0.0

1.0

-2.0

Corral Quemado -3.0

3.53Ma -4.0 2000m- 4.95Ma Age(Ma) -5.0 5.30Ma

6.0

8.0 7.0 6D 5.0 4.0 3.0 6.70Ma -7.0 Age(Mo) 0- FIG.9.-Ages impliedby correlationof magnetic polarity zonation with the magnetic polarity time -8.0 scale and 40K-40Ar age determinationsplotted againststratigraphic thickness. Ages impliedby the magnetostratigraphiccorrelation are shown by (x) -9.0 while 40K- 40Arage determinationsare shown by solid circles with attendanterror estimates. FIG.8.-Magnetic polarityzonation from Puerta de CorralQuemado section correlatedwith magnetic polaritytime scale of Ness, Levi, and Couch imply major and repeated changes in sedi- (1980). ment accumulation rate. We thus believe that the magnetostratigraphic and 4K - 40Ar data are providing accurate age determinations for served more easily in figure 8, where the mag- this stratigraphic section. netic polarity zonation is shown against a Age of Formation Boundaries.-Stah- larger portion of the magnetic polarity time lecker (in Riggs and Patterson 1939, fig. 1; scale. Marshall and Patterson 1981, fig. 6) recog- An effective evaluation of the consistency nized five major mappable units in his Puerta of the magnetostratigraphic and 40K- 40Ar de Corral Quemado profile; from oldest to data is a plot of the ages implied by the mag- youngest: the Calchaqui, Chiquimil B, Chi- netostratigraphic correlation and the 40K- quimil A, Araucanense, and Corral Quemado 40Ar age determinations against stratigraphic Formations. Our measured section (fig. 7) in- thickness (fig. 9). The internal consistency cludes the upper part of Chiquimil A, all of evident in figure 9 is to some degree forced the Araucanense, and all except the very top because the 40K- 40Ar age determinations of the Corral Quemado Formations. Correla- were used as guidance in the magnetostrati- tions of our current section with that of graphic correlations. However, the fact that Stahlecker are based on the secure the data are consistent with a simple constant identification of units 15 and 29 in both. Pro- sediment accumulation rate of 56 cm/1000 yr portional dividers were used to extrapolate during deposition of this section provides units from Stahlecker's section into ours, us- confidence that the correlation of figure 7 is, ing units 15 and 29 as reference points in this in fact, correct. Any other correlation would process. With this approach we were able to

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions MAGNETIC POLARITY STRATIGRAPHY 633 place the following ages on formational 1976). Dispersal of these groups occurred boundaries: Chiquimil A-Araucanense 7.5 prior to the emergence of the Panamanian Ma, and Araucanense-Corral Quemado 6.4 land bridge at about 3.0 Ma (Marshall et al. Ma (fig. 7). 1979, 1982). Stahlecker also made a detailed section at The earliest known Procyonidae in South Chiquimil and in it recognized the same lower America are of the genus Cyonasua from four formations that he recognized at Puerta rocks of Huayquerian age in Catamarca Prov- de Corral Quemado. His Chiquimil profile ap- ince, Argentina (Kraglievich and Olazabal parently terminated below rocks and faunas 1959, table on p. 45). Numerous specimens that would correlate with the Corral Que- were collected by early Argentine paleon- mado Formation. Based on Stahlecker's tologists from rocks of the Araucanense For- geological study and on a preliminary study mation in the Valle del Rio Santa Maria at of the fossils, Riggs and Patterson (1939, fig. and near Chiquimil (fig. 1), but the strati- 1) proposed a tentative correlation of the Chi- graphic levels from where they came were quimil and Puerta de Corral Quemado sec- not recorded, and their positions within the tions. The geochronologic data presented geochronologic time frame presented here here require modification of one important are not securely known. As summarized by aspect of that correlation. The tuff from unit Kraglievich and Reig (1954) and Kraglievich XIX at Chiquimil is confidently dated at 6.02 and Olazabal (1959) these specimens include: Ma (see Marshall et al. 1979) and that unit is (1) the type of Cyonasua argentina Ameghino placed by Riggs and Patterson (1939, fig. 1) in (1885, p. 19); (2) five specimens (MACN the middle part of their Araucanense Forma- 6687, 6688, 6689, 6692, 8211) referred to Cy- tion. This securely dated level correlates not onasua cf. C. argentina; (3) the type of Cy- with the middle part of the Araucanense For- onasua brevirostris (Moreno and Mercerat mation at Puerta de Corral Quemado as they 1891, p. 235); and (4) the types of Am- suggest, but with the middle of unit 18 in the phinasua longirostris Rovereto (1914, p. 81) lower part of the Corral Quemado Formation. and Pachynasua? robusta Rovereto (1914, p. Thus, rocks and faunas from the upper half of 82), both regarded as junior synonyms of C. the Araucanense Formation at Chiquimil cor- brevirostris (Kraglievich and Olazabal 1959). relate with those of the lower half of the Cor- The only Procyonidae from Catamarca ral Quemado Formation at Puerta de Corral with good stratigraphic control were obtained Quemado. This revised correlation clearly in 1926 by an expedition from the Field warrants a detailed reanalysis of the fossil Museum (Riggs and Patterson 1939; Marshall mammal faunas from both localities. et al. 1979; Marshall and Patterson 1981). Age of Fossil Procyonidae.-In the late Five specimens, two from Chiquimil and Miocene a limited but important interchange three from Puerta de Corral Quemado (fig. 1), of faunas between the Americas occurred were collected. The two specimens from Chi- (Simpson 1980). Members of the North quimil include FMNH P14342 from unit American family Procyonidae (raccoons and XVIIIa and FMNH P14537 from unit XIX of their allies) dispersed to South America, and the Araucanense Formation of Stahlecker in members of the South American ground sloth Riggs and Patterson (1939) and Marshall and families Megalonychidae and Mylodontidae Patterson (1981). Both specimens are re- dispersed to North America. These dispersal ferred to Cyonasua brevirostris by Tedford in events apparently resulted from members of Marshall et al. (1979). An average age of 6.02 each group being carried on rafts of vegeta- Ma was obtained on three plagioclase (KA tion that were broken away from banks of 3305, 3305R, 3390), and one sanidine (KA swollen rivers and ferried directly or by way 3386) concentrate from a prominent tuff in of island chains across the marine water bar- unit XIX (Marshall et al. 1979). This date ap- rier that then separated the two Americas proximates the age of these specimens of Cy- (Coney 1982). These animials, the "new is- onasua at this locality. land-hoppers" of G. G. Simpson (1950), rep- The three specimens from Puerta de Corral resent the first participants in the Great Quemado include: (1) FMNH P14451 from American Faunal Interchange (sensu Webb unit 14 of the Araucanense Formation (fig. 7)

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions 634 R. F. BUTLER ET AL. and referred to Cyonasua sp. by Tedford in the Huayquerias Formation, namesake and Marshall el al. (1979); (2) FMNH P14397 from type formation of the Huayquerian land unit 16 or 17 of the Araucanense Formation mammal age (Simpson 1940). At this locality (fig. 7) and identified as Cyonasua cf. C. the Huayquerias Formation is discordantly lutaria by Tedford in Marshall et al. (1979) overlain by the Tunuyan Formation (Dessanti [the preceding two specimens were incor- 1946), which contains a mammal fauna as- rectly listed by Marshall et al. (1979, p. 276) signed to the Montehermosan land mammal as coming from the Corral Quemado Forma- age. The boundary between the Huayquerias tion]; and (3) FMNH P14401 probably from and Tunuyan Formations at San Carlos is unit 21 of the Corral Quemado Formation thus 5.8 Ma or younger. The combined ages (fig. 7) and referred to Chapalmalania cf. C. from San Carlos and from Puerta de Corral altaefrontis by Kraglievich and Olazabal Quemado apparently bracket the Huayque- (1959, p. 9, 28). Chapalmalania is a bear-like rian-Montehermosan boundary between 5.8 procyonid and an apparent direct descendant and 6.4 Ma. Based on these dates we tenta- of Cyonasua (Kraglievich and Olazabal tively recognize an age of 6.0 Ma as the ap- 1959). FMNH P14401 is the specimen listed proximate boundary between these land as Cyonasua nov. sp. by Riggs and Patterson mammal ages in Argentina. (1939, p. 145). Our geochronologic data indicate the following ages for these specimens CONCLUSIONS (see fig. 7)-7.0 to 7.5 Ma, 6.4 to 6.9 Ma, and Magnetostratigraphic and 40K- 4Ar data 5.4 to 5.8 Ma, respectively. FMNH P14451 on a 2300 m thick sequence of continental from unit 14 dated between 7.0 and 7.5 Ma is sediments at Puerta de Corral Quemado in the oldest known representative of Cyonasua Catamarca Province, NW Argentina permit in South America, and FMNH P14401 from calibration of land mammal faunas of Huay- unit 21? dated between 5.4 and 5.8 Ma is the querian and Montehermosan age. The time oldest known representative of Chapal- represented includes the interval from -8.0 malania . Ma to 3.5 Ma. The Araucanense Formation Age of the Huayquerian-Montehermosan contains a mammal fauna of Huayquerian Boundary.-The geochronologic data pre- age, and the overlying Corral Quemado For- sented here permit approximate dating of the mation contains a fauna of Montehermosan boundary between Huayquerian and Mon- age. The age of the boundary between these tehermosan land mammal age faunas at this formations is dated at 6.4 Ma, and this age is locality. As summarized by Pascual and believed to approximate the boundary be- Odreman Rivas (1973, chart opposite p. 318), tween the Huayquerian and Montehermosan Marshall and Patterson (1981), and Marshall land mammal ages at this locality. A speci- et al. (1983), the Araucanense Formation at men of Cyonasua from unit 14 (dated be- Puerta de Corral Quemado contains a mam- tween 7.0 and 7.5 Ma) of the Araucanense mal fauna assigned to the Huayquerian land Formation is the oldest known representative mammal age, and the Corral Quemado For- of the North American family Procyonidae in mation contains a fauna assigned to the Mon- South America. This specimen represents the tehermosan land mammal age. The boundary earliest known participant of the Great Amer- between the Araucanense and Corral ican Faunal Interchange on that continent. Quemado Formations thus approximates but In North America the earliest undoubted does not necessarily represent the boundary participant in this early phase of the Great between Huayquerian and Montehermosan American Faunal Interchange is represented land mammal age faunas at this locality. The by a specimen of the megalonychid ground boundary between these formations is se- sloth Pliometanastes protistus Hirschfeld curely dated at 6.4 Ma (fig. 7). and Webb (1968) from the Siphon Canal lo- Rocks and faunas of Huayquerian age have cality of the Mehrten Formation, Stanislaus also been dated by the 40K- 40Armethod near County, California (Hirschfeld, 1981). This San Carlos in Mendoza Province, west-cen- specimen was collected from 4 m below a tuff tral Argentina (see Marshall et al. in press). dated at 8.19 ±0.16 Ma (Wagner 1981). An age of 5.8 Ma was obtained on biotite sep- Thus, the available geochronologic data arates from a tuff believed to be at the top of document the occurence of the earliest un-

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions MAGNETIC POLARITY STRATIGRAPHY 635 doubted participants in the Great American (LGM). The 40K- 40Ar dating was supported Faunal Interchange in rocks of similar age in by NSF Grant EAR-73-00235 A01, formerly North America and South America. These GA-40805 (GHC), and NSF Grants EAR- new data further suggest that the beginning of 7909515, EAR-8300918, and EAR-8305243 the interchange may have been a synchro- (LGM). Processing paleomagnetic samples nous event or nearly so, as proposed by was supported by NSF Grants EAR-75-13571 Webb (1976) and not markedly diachronous and EAR-8115430 (RFB), and some of the as proposed by Marshall et al. (1979). paleomagnetic equipment was provided by a Cottrell Research Grant from the Research Corporation. Kathy Flanagan's able assis- ACKNOWLEDGMENTS.-Fundsfor field work tance with the paleomagnetic analyses is were provided by grant 1698 from the Na- gratefully acknowledged. The manuscript tional Geographic Society, Washington, D.C. was improved by critical readings by John (LGM, RFB) and NSF Grant EAR-7909515 Obradovich and Steve May.

REFERENCES CITED ALLMENDINGER,R. W.; RAMOS, V. A.; JORDAN, JOHNSON,N. M.; OPDYKEN. D.; and LINDSAY, T. A.; PALMA,R. M.; and ISACKS,B. L., 1983, E. H., 1975, Magnetic polarity stratigraphy of Paleogeography and Andean structural geome- Pliocene/Pleistocene terrestrial deposits and ver- try, northwest Argentina: Tectonics, v. 2, p. 1- tebrate faunas, San Pedro Valley, Arizona: Geol. 16. Soc. America Bull., v. 86, p. 5-12. AMEGHINO,F., 1885, Nuevos restos de mamfferos JORDAN, T. E.; ISACKS, B. L.; ALLMENDINGER, f6siles oligocenos, recogidos por el professor R. W.; BREWER,J. A.; RAMOS,V. A.; and ANDO, Pedro Scalabrini y pertenecientes al Museo pro- C. J., 1983, Andean tectonics related to geome- vincial de la ciudad de Parani: Bol. Acad. Cienc. try of subducted Nazca plate: Geol. Soc. C6rdoba, v. 8, p. 3-207. America Bull., v. 94, p. 341-361. Bossi, G. E., and PALMA,R. M., 1982, Considera- KRAGLIEVICH,J. L., and OLAZABAL,A. G. DE., cion de la estratigrafia del Valle de Santa Maria, 1959, Los procionidos extinguidos del genero Provincia de Catamarca, Argentina: Quinto Chapalmalania Amegh.: Rev. Mus. Argent. Congr. Latino-Americano de Geol., Argent., Ac- Cienc. Natur., Cien. Zool., Buenos Aires, v. 6, tas 1, p. 155-172. p. 1-59. BUTLER,R. F., 1982, Magnetic mineralogy of conti- - , and REIG, O. A., 1954, Un nuevo pro- nental deposits, San Juan Basin, New Mexico, cionido del Plioceno de Las Playas (Provincia de and Clark's Fork Basin, Wyoming: Jour. C6rdoba): Rev. Assoc. Geol. Argent., v. 9, p. Geophys. Res., v. 87, p. 7843-7852. 209-231. CONEY, P. J., 1982, Tectonic constraints on MARSHALL,L. G.; BUTLER,R. F.; DRAKE,R. E.; biogeographic connections between North and CURTIS,G. H.; and TEDFORD,R. H., 1979, Cali- South America: Ann. Missouri Bot. Garden, v. bration of the Great American Interchange: Sci- 69, p. 432-443. ence, v. 204, p. 272-279. DESSANTI, R. N., 1946, Hallazgo de dep6sitos ---; BUTLER,R. F.; DRAKE,R. E.; and CURTIS, glaciales en las Huayquerias de San Carlos G. H., 1982, Geochronology of type Uquian (late (Mendoza): Rev. Assoc. Geol. Argent., v. 1, p. Cenozoic) land mammal age, Argentina: Sci- 270-284. ence, v. 216, p. 986-989. FISHER,R. A., 1953, Dispersion on a sphere: Proc. -- ; DRAKE,R. E.; and CURTIS,G. H., in press, Royal Soc. (London) Ser. A, v. 217, p. 295-305. 40K-40Ar age calibration of late Miocene- FRANCIS, P. W., and BAKER, M. C. W., 1978, Pliocene mammal-bearing Huayquerias and Sources of two large ignimbrites in the Central Tunuyan Formations, Mendoza Province, Andes: some landsat evidence: Jour. Volcan. Argentina: Jour. Paleont. Geotherm. Res., v. 4, p. 81-87. - ; HOFFSTETTER,R.; and PASCUALR., 1983, --- ; HAMMILL,M.; KRETZSCHMAR,G.; and Geochronology of the continental mammal- THORPE,R. S., 1978, The Cerro Galan Caldera, bearing Tertiary of South America: Palaeover- North-west Argentina and its tectonic setting: tebrata, Mem. Extr., 1983, p. 1-93. Nature, v. 274, p. 749-751. --- ; and PATTERSON,B., 1981, Geology and HIRSCHFELD,S. E., 1981, Pliometanastes protistus geochronology of the mammal-bearing Tertiary (Edentata, Megalonychidae) from Knight's of the Valle de Santa Maria and Rio Corral Ferry, California with discussion of Early Hem- Quemado, Catamarca Province, Argentina: Fiel- phillian megalonychids: PaleoBios, v. 36, p. 1- diana-Geology, n.s. no. 9, p. 1-80. 17. MORENO,F. P., and MERCERAT,A., 1891, Explora- - , and WEBB, S. D., 1968, Plio-Pleistocene cion arqueologica de la provincia de Catamarca: megalonychid sloths of North America: Bull. Rev. Mus. La Plata, v. 1, p. 222-236. Fla. State Mus. Biol. Sci., v. 12, p. 213-296. NESS, G.; LEVI, S.; and COUCH,R., 1980, Marine

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions 636 R. F. BUTLER ET AL.

magnetic anomaly timescale for the Cenozoic Catamarca y Tucumin: Bol. Dir. Nac. Geol. y and Late Cretaceous: a precise critique, and syn- Min. 134, p. 1-65. thesis: Rev. Geophys. Space Phys., v. 18, p. SIMPSON,G. G., 1940, Review of the mammal- 753-770. bearing Tertiary of South America: Proc. Am. PASCUAL,R., and ODREMANRIVAS, O. E., 1973, Las Philos. Soc., v. 83, p. 649-709. unidades estratigrificas del Terciario portadores - , 1950, History of the fauna of Latin de mamiferos. Su distribuci6n y sus relaciones America: Amer. Sci., v. 38, p. 261-389. con los acontecimientos diastr6ficos: Actas - , 1980, Splendid Isolation: the Curious His- del Quinto Cong. Geol. Argent., v. 3, p. 293- tory of South American Mammals: New Haven, 338. Yale University Press, 266 p. RIGGS,E. S., and PATTERSON,B., 1939, Stratig- WAGNER,H. M., 1981, Geochronology of the raphy of Late Miocene and Pliocene deposits of Mehrten Formation, Stanislaus County, Califor- the province of Catamarca (Argentina) with nia: Unpub. Ph.D. dissertation, Univ. Califor- notes on the faunae: Physis (Rev. Soc. Argent. nia, Riverside. Cien. Natur.), v. 14, p. 143-162. WATSON,G. S., 1956, Analysis of dispersion on a ROVERETO,C., 1914, Los estratos araucanos y sus sphere: Mon. Not. R. Astron. Soc. Geophys. f6siles: Anal. Mus. Nac. Hist. Nat., Buenos Suppl., v. 7, p. 153-159. Aires, v. 25, p. 1-250. WEBB,S. D., 1976, Mammalian faunal dynamics of Ruiz HUIDOBRO,O. J., 1972, Descripci6n geol6gica the Great American Interchange: Paleobiology, de la Hoja lle, Santa Maria, provincias de v. 2, p. 216-234.

This content downloaded from 64.251.254.77 on Mon, 28 Oct 2013 18:33:29 PM All use subject to JSTOR Terms and Conditions