Society for American Archaeology

Occasional Hunts or Mass Kills? Investigating the Origins of Archaeological Pronghorn Bonebeds in Southwest Wyoming Author(s): Jack N. Fenner Source: American Antiquity, Vol. 74, No. 2 (Apr., 2009), pp. 323-350 Published by: Society for American Archaeology Stable URL: http://www.jstor.org/stable/20622429 . Accessed: 10/02/2015 00:46

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

.

Society for American Archaeology is collaborating with JSTOR to digitize, preserve and extend access to American Antiquity.

http://www.jstor.org

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions OCCASIONAL HUNTS OR MASS KILLS? INVESTIGATING THE ORIGINS OF ARCHAEOLOGICAL PRONGHORN BONEBEDS IN SOUTHWEST WYOMING

JackN. Fenner

It is often unclear whether an archaeological animal bonebed was formed by a mass kill of individuals from a single prey population or through the accumulation of a number of separate kills which occurred over a long time period or at multi ple locations. The isotope ratio variation within animals from an archaeological bonebed can be used to investigate this question by comparing the isotope ratio variation within an archaeological assemblage to the variation within modern pop ulations. If the archaeological variation exceeds that expected within a single population, then an accumulation is indi cated. Otherwise, a single mass kill or a series of kills from a single population over a fairly short timeperiod ismore likely. This technique is demonstrated using carbon, oxygen, and strontium isotope ratios to characterize the amount of variation within pronghorn tooth enamel from six archaeological sites in southwestern Wyoming. The results indicate that theAustin Wash and Trappers Point sites are the result of an accumulation of pronghorn from multiple populations, while the Fire hole Basin, Gailiun, and Boars Tusk sites do not show evidence of accumulation and therefore may be the result of mass kills. The Eden-Farson site formation characteristics are currently indeterminate.

A menudo es dificil determinar si un dep?sito arqueol?gico de restos ?seos de animal seform? a traves de la muerte en masa de individuos de una ?nica poblaci?n de presas, o a traves de la acumulaci?n de un n?mero de muertes por separado que tuvieron lugar durante un periodo de tiempo prolongado o en multiples lugares. Es posible utilizar la variaci?n de relaciones isot?picas en los animales de un dep?sito arqueol?gico de huesos para investigar dicha cuesti?n. La variaci?n de relaciones isot?picas de un contexto arqueol?gico es comparada con la variaci?n en de poblaci?n moderna y, si la variaci?n arque ol?gica supera aquella que se espera dentro de una ?nica poblaci?n, se indica una acumulaci?n. De otra manera, es mas prob able una ?nica muerte en masa o una serie de muertes de una ?nica poblaci?n a lo largo de un periodo de tiempo relativamente corto. Dicha tecnica queda demostrada mediante el uso de relaciones isot?picas de carbono, oxigeno y estroncio para car acterizar la magnitud de la variaci?n en el esmalte de dientes de antttopes americanos proveniente de seis yacimientos arque ol?gicos del sudoeste de Wyoming. Los resultados indican que los yacimientos de Austin Wash y Trappers Point son el resultado de una acumulaci?n de antilocapras americanas de diversas poblaciones, mientras que los yacimientos de Firehole Basin, Gailiun y Boars Tusk no presentan evidencia de acumulaci?n y por lo tanto pueden ser el resultado de muertes en masa ?ni cas. Las caracteristicas de la formacion del yacimiento de Eden-Farson a?n no han sido determinadas.

Pronghorn (Antilocapra americana) roam over time.This is an importantquestion because it wild onWyoming rangelands today as they has implications for a variety of human behaviors have formany millennia. Archaeological of archaeological interest.An improved under evidence indicates that people have successfully standing of prehistoricmass kill strategies in gen hunted these fairly large and very fleet animals for eral, and pronghorn procurement in particular, at least 13,000 calendar years (Frison 1991; Hill would affectnot only the interpretationof siteoccu 2001; Lubinski 2000a), which in some instances pation histories, but also larger issues such as the has led to substantial archaeological pronghorn prevalence of communal hunting and population bonebeds. As is often the case for archaeological aggregation (Arkush 1986; Davis et al. 2000; Faw bonebeds, it is difficultto determine whether a par cett 1987; Lubinski 1999), resource intensification a ticular pronghorn bonebed resulted from single (Lubinski 1997), population size and mobility or an mass kill from accumulation ofmultiple kills (Todd 1987;Widga 2004), gender roles and meat

N. Jack Fenner Department of Archaeology and Natural History, The Australian National University, Canberra ACT 2600, Australia ([email protected])

American Antiquity 74(2), 2009, pp. 323-350 Copyright ?2009 by the Society forAmerican Archaeology

323

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 324AMERICAN ANTIQUITY [Vol. 74, No. 2,2009]

sharing (Dewar et al. 2006), and cultural or tech pie populations and a singlemass kill (or a series nological change (Lubinski 1997, 2000a; Ugan et of fairlyclosely related kills from a single popula al. 2003). tion) ismore likely. Much of our understanding of mass procure Hoppe (2004) assessed thecarbon, oxygen, and ment events derives from ethnographic and histor strontium isotope ratio variation of two paleonto ical sources. Within western North America, logical mammoth bonebeds thought to represent traditionalpeoples communallymass-hunted bison catastrophic herdmortality, and decided thatonly (De Smet 1859), rabbits (Chamberlin 1911:336; carbon isotopes showed sufficient inter-herdvari Steward 1933:253-254), deer (Steward 1933:252, ation to confidentlydistinguish between herds. She 1938:82-83), and even insects such as crickets then compared the carbon isotope variation in the (Egan 1917:230-233). Of particular interestto this paleontological mass mortality bonebeds to the study are numerous reports of historic communal variation within theBlackwater Draw, Miami and pronghorn hunts (Arkush 1986; Egan 1917: Dent mammoth archaeological assemblages. Her 238-241; Grinnell 1923:277-290; Hyde analysis suggests that all three assemblages are 1968:18-20; Irving 1986:225-226,353-354; Kelly composed of mixtures of individuals frommulti 1932:82-86; Kindig 1987; Lowie 1909:185, ple herds (Hoppe 2004:139) and therefore likely 1924:197; Lubinski 1999; Steward 1938:34-36, representmultiple kill events within each site. 105, 115, 120, 128, 175, 233; Sundstrom 2000). Widga (2004:38-41), as part of a larger analy Pronghorn herds were driven or led into brush or sis of Early Archaic bison procurement, compared timber corrals, driven into pits, or surrounded by the carbon isotope variation within bison remains people either on foot or horseback. The pronghorn recovered from several components of the Spring were often then run to exhaustion, at which point Creek and Logan Creek sites inNebraska to that they could be easily killed. There are also numer of bison remains fromPaleoindian mass kill com ous accounts of individual pronghorn hunting, ponents at theAgate Basin site and themultiple using disguises to stalk pronghorn (e.g., Lowie kill Plains Village sites ofHelb and Talking Crow. 1909, 1924) or colored flags thatflap in thewind He found thatSpring Creek and at least one Logan and attract curious pronghorn (e.g., Chittenden Creek component appear to represent single 1905; Schoolcraft 1860). While valuable, these episode procurement events.Widga (2004:39,41) accounts lack timedepth andmay reflectresponses acknowledges that the comparative samples from toEuroamerican contact ratherthan traditionalpre differingtime periods are of uncertain comparative historic behavior. For a long-termperspective, we value, and thata number of untested assumptions must turn to archaeology. are included inhis isotopic analysis. Nevertheless, Unfortunately, theuse of archaeology tounder a qualitative assessment of the range of variation stand communal mass-kill behavior is often ham suggests thatbison mass kill sitesmay have sub pered by an inabilityto reliablydistinguish between stantiallyless carbon isotope variation thando mul sites thatcontain the remains of large kills from a tiple kill sites (Widga 2004:41). single population and those containing an accu These interestingstudies introduce a technique mulation of small, presumably noncommunal kills thatpromises to improve our understanding of pre obtained frommultiple prey populations. Hoppe historic hunting practices. The study reportedhere (2004) andWidga (2004) recently and indepen furtheranalyzes and develops isotopic characteri dently introduceda technique thatuses isotope ratio zation of bonebeds along several lines.Both Hoppe analysis to investigate this issue. Simply put, a site andWidga used ancient animal remains thought to whose prey remains contain more isotopic varia result fromknown mortality events as comparative tion than is found in a singlemodern prey popula samples to characterize unknownmortality events. tion is likely to be the result of an accumulation This was necessary because the animals used in frommultiple populations and thereforeprobably those studies (mammoth and bison) no longer roam is not the result of a singlemass kill. Conversely, wild, so no modern sample set was available. if the isotopic variation within a site is less than or Although partially circumscribed by fences and equal to that found inmodern populations, then other artificial structures,pronghorn remain wild there is no evidence of accumulation frommulti withinWyoming and thereforeprovide a good com

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS

Figure 1.Wyoming pronghorn bonebeds and hunt areas. parative sample for assessing archaeological iso of these traditionalmethods and their results is in tope ratio variation.My study also expands upon order. theHoppe-Widga work by using a scaling and dis tance that approach simultaneously incorporates Previous Work isotopic informationfrom multiple elements, and a bootstrap technique that provides quantitative Archaeologists confrontedwith substantialprehis comparisons of isotopic variability between loca toric bonebeds currentlyuse a number of criteria tions. for assessing whether a particular archaeological The isotope ratio analysis technique is used to faunal assemblage is the resultof a singlemass kill investigate the formation processes of bonebeds or the accumulation of multiple kill events (e.g., from six archaeological sites in southwestern Davis et al. 2000; Frison 2000; Hill 2002; Lubin Wyoming: Austin Wash (48UT390), Boars Tusk ski 1997, 2000a; Todd et al. 1992, 2001; Widga (48SW1373), Eden-Farson (48SW304), Firehole 2004). Lubinski (1997:124-206, 2000a) provides Basin (48SW1217), Gailiun (48SW1156), and a detailed discussion of many of these criteria as Trappers Point (48SU1006) (Figure 1).The carbon, theyrelate to identificationof a pronghornmass kill. oxygen, and strontiumisotope ratios aremeasured He separates the criteria into threecategories: evi in enamel collected from pronghorn teeth found dence for human-caused mortality (e.g., associa within bonebeds at each of these sites.The amount tionwith projectile points or presence of cutmarks), of isotopic variation within each site is computed evidence for a single depositional episode (e.g., and compared withmodern pronghornpopulations stratigraphicassociation), and evidence for a sin to assess whether thepronghorn remains at the site gle mortality event. The work discussed in this were obtained frommultiple pronghorn popula paper is particularly concerned with the last cate tions (and thereforelikely representmultiple mor gory,which Lubinski evaluates using two typesof talityevents). The number of mortality events at evidence: uniform season of death and presence of each of these sites has previously been analyzed a catastrophic demographic profile.A uniform sea using traditionalmethods. Before presentingdetails son of death is evidence thatthe animals died within of the currentproject and its results,a brief review a fairlynarrow window of time,while a catastrophic

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 326AMERICAN ANTIQUITY [Vol. 74, No. 2,2009]

demographic profile suggests that an entire herd Table 6 for a summary of Lubinski's results). In may have been killed at once. However, each of addition, investigatorsat some of these sites have these lines of evidence has difficulties.The season used other techniques, such as evaluating the clus of deathmay be uncertain. This is a particular con teringof radiocarbon dates, to assess the number cern for pronghorn bonebeds because season of ofmortality events. The remainder of this section death analyses typically depend on tooth eruption describes the non-isotopic evidence formultiple and wear techniques that in pronghorn often pro prey populations within each of six archaeological duce very wide seasonality ranges (Lubinski and pronghorn bonebed sites in southwestern O'Brien 2001; Miller et al. 1999b:483). Addition Wyoming. Subsequent sections describe the iso ally, thebonebed may be the result ofmultiple kill tope ratio variation evidence developed during the episodes during the same season over differing currentproject. Finally, in theDiscussion section, years without discernable stratigraphicseparation the combined non-isotopic and isotope ratio vari (asmight be expected ifthe kill episodes are depen ation evidence is considered and the pronghorn dent on animal migration [Hill 2002:331; Lubin population characteristics of each site are assessed. ski 1997:198; Lubinski and O'Brien 2001]). Austin Wash Tooth eruption and wear patterns are used by prehistorians?especially those evaluating large The Austin Wash site (48UT390) is a largemulti prehistoric bison bonebeds?to develop prey age component locality in southwesternWyoming that profiles for use in estimating herd demographics has been the subject of several investigations (Reiss as well as the seasonality ofmortality events (Fri and Walker 1982; Schroedl 1985; SHPO son 1978, 1991; Frison and Reher 1970; Todd et 1981-1999). It is "situated on theupper of two ter al. 1992; but seeWhittaker and Enloe 2000). More races above and just west of the confluence of generally, catastrophic (or "non-selective") prey AustinWash and an unnamed, ephemeral drainage age demographic profiling is used to distinguish thatflows from thenorthwest" (Schroedl 1985:35). between bonebeds created by animal predation, Two early components of the sitehave been radio human hunting, and human scavenging (e.g.,Klein carbon dated to 3030 ? 70 B.P. and 1740 ? 60 B.P. 1982; Stiner 1990, 1991a, 1991b). Catastrophic The component of interestfor thepresent analysis age profiles are those that reflect a "normal" age is a large pronghorn bonebed in the northernpor profile for the species in question and are thought tion of the site thathas been radiocarbon dated to to indicate that an entire herd or population died 1197 ? 26 B.P. Except where specifically noted, all at once in a mass mortality event. The use of references toAustin Wash in thisreport refer to this catastrophic age profiles to identifymass kills is component. No cultural or geological stratification controversial (Caughley 1974; Lubinski 2000b; is visible within the component, and all materials Lyman 1987; Steele 2003; Whittaker and Enloe were recoveredwithin 40 cm of thepresent ground 2000). For age profiles to be acceptable mortality surface. In addition to pronghorn, the remains of indicators, the animal population in question must at least twobison, one canid, and a handful of small, have a stable population level, a consistent normal possibly intrusive,rodents and rabbitswere recov living age structure, and a clear age indicator. ered.The pronghorn,bison, and canid bones exhibit Wildlife biologists have rejected theuse of age pro cut marks. files inwildlife management and moved tomore In themost detailed report available on the complex population models that incorporatemany AustinWash site,Schroedl (1985:91) indicates that variables that cannot be determined in archaeo "analysis of the radiocarbon dates..., the distribu logical situations (e.g., animal densities on differ tionof the artifactsand the faunal remains, and the ent range types; see Kohlmann [2004] formodern cohesiveness of the artifactassemblage all indicate pronghorn population modeling discussion). that the main excavation area [the pronghorn Despite these concerns, Lubinski (1997,2000a, bonebed] represents a single, short-termantelope 2000b) has used uniform season of death and pres processing locale." His ANOVA analysis of six ence of a catastrophic demographic profile as cri radiocarbon dates from the component indicated teria to assess formation history at five of the six thatthey are not significantlydifferent. This is sup archaeological sites discussed in this report (see ported by the radiocarbon clustering method of

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 327

Lubinski (1997:226-242, 2000a, 2000b) ana lyzed the pronghorn mortality pattern at Austin Wash and concluded that the evidence is "incon sistentwith a single mass kill event" (Lubinski 1997:242). His seasonality assessment was incon clusive because itwas based on three samples that produced mortality periods thatoverlap fromOcto ber to but range from to As 700 600 800 900 1000 1100 January July April. pre viously noted, this broad range is common in Year (A.D.) pronghorn tooth-based seasonality assessment (Lubinski 2001; Lubinski and O'Brien 2001) and Figure 2. Austin Wash pronghorn bonebed radiocarbon dates. The 95 percent confidence intervals are shown for provides little insight into thenumber ofmortality the six radiocarbon dates from the bonebed component events. His assessment of age demographic struc 7271 - 1197 ? (Beta 2444, 3325, 7274). Weighted average: ture indicates thatAustin Wash to have 26 B.P.; Cal. A.D. 724 to 739 and 771 to 893. "appears too many mature animals for a living structure" (Lubinski 2000b: 124), which leads him to con clude that was a mass ormor Wilson andWard (1981), which does not find sig AustinWash not kill = nificant structurewithin the dates (T' 9.07; p > tality site. .05). Likewise, using a current radiocarbon cali Firehole Basin #11 bration curve (BCal 2006), the 95 percent confi The Firehole dence intervals for the six dates overlap formost Basin#l 1 site (48SW1217; generally to as on of the period between A.D. 770 and 876 (Figure referred herein simplyFirehole Basin) lies a 2). Note, however, that theyalso present a spread hill between theGreen River and Little Bitter Itwas of 435 years. The radiocarbon date overlap there Creek. excavated during 1976 and 1977, but fore supports a single contribution to thebonebed, has been only partially reported (Lubinski 1997, et but thedates also include sufficientuncertainty to 2000a; Middleton al. 2007). Thirty-fourtest units were also be consistentwith multiple contributions that excavated, although many of the units were accumulated over a timeperiod of up to several hun dug only five centimeters below the surface.Nev a dred years. ertheless, substantial bone, lithic, and ceramic Schroedl's reference to the distribution of the assemblage was recovered. Lubinski (1997:243 artifactsapparently relates to the fact thatthe stone 245) indicates that hafted "Shoshonean" knives, and bone artifacts are concentrated in the same ceramics, tri-notched projectile points, and two ? area. It is not clear what ismeant by the "cohe radiocarbon dates of645 135B.P. (calA.D. 1159 to ? siveness" of theartifact assemblage, unless itrefers 1469) and 625 50 B.P. (calA.D. 1283 to 1406) all to to consistency in the temporally diagnostic stone supportLate Prehistoric Protohistoric affili tool collection. Schroedl (1985:50-56) identified ation. The bonebed includes at least 26 pronghorn 16 of 18 temporally diagnostic projectile points individuals, plus five rabbit bones, three pocket found within the pronghorn bonebed as "Rose gophermandibles, and one sage grouse bone (Mid et Spring" type.The other two points were identified dleton al. 2007). Lubinski (1997:245) reports two as Avonlea (thought to be diagnostic of theLate that bison bones from Firehole Basin were were Prehistoric period) and Lovell Constricted. The examined in 1986, though they subsequently In was Lovell Constricted point is interestingas it is a late lost. any case, the faunal assemblage clearly Paleoindian point type (Frison 1991). Schroedl dominated by pronghorn. (1985:55) states without furtherdiscussion that Lubinski's (1997:249) eruption and wear analy seven "although recovered fromwithin thebone midden, sis of Firehole Basin mandibles produced it does not represent an earlier occupation in the seasonality estimates that ranged fromAugust to midden area." Such an early point does seem out April and overlapped during December and Janu of place and probably can be dismissed as intru ary. Seasonality analysis thereforeis not helpful in sive to thebonebed, but perhaps ithints at a weak assessing the number of mortality events. Lubin ness in the cohesiveness of the collection. ski also assessed age demographics using Juvenile/

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 328AMERICAN ANTIQUITY [Vol. 74, No. 2,2009]

Prime/Old (JPO) partitions and Fawn/Youth/ dismissed [as a single mass kill site] because it Mature (FYM) partitions, as well as a histogram lacks fawns" (Lubinski 2000b: 124). As he notes of annual age categories (Lubinski 1997:249-253, (Lubinski 1997:279), additional excavations may 2000b), all ofwhich were derived frompronghorn yield more samples thatcould revise his interpre mandibles. When compared tohis model distribu tation. tions, two of thesemeasures (FYM and histogram) Eden-Farson most closely resemble an entire living population and therefore suggest a mass kill, while the third Eden-Farson (48SW304) is a large site located on (JPO) ismore compatible with an attritionalmor rolling hills about sixmiles east ofEden, Wyoming talitypattern. His assessment is that,on balance, (Frison 1971,2000; Larson et al. 1969;Lytle 1969). "themass kill hypothesis ismost parsimonious and The Wyoming Archaeological Society excavated a Firehole Basin is evaluated as a singlemass kill" substantialportion of the siteduring 1969 and 1970. (Lubinski 1997:253-254). Additional material was collected from exposed areas during subsequent visits, but it appears that Gailiun substantial additional material was lostdue to loot The Gailiun site (48SU1156) was partially exca ing. Eden-Farson has been interpretedas a proto vated during 1978 and 1990 but has not been for historic Shoshonean occupation. It includes a mally reported.Thus we must rely on Lubinski's number of post holes, depressions, hearths, and (1997, 2000a) pronghorn bonebed analysis publi bone debris piles that are interpretedas evidence cations for site information.The site is located on of tenhabitations or lodges. The remains of at least an alluvial silt terrace along Dry Sandy Creek. It 226 pronghornwere recovered. Though theEden includes five shallow depressions with artifactclus Farson faunal collection is dominated by pronghorn ters thatBill Current, who excavated the site in remains, relatively small amounts of bison, canid, 1990, identified as the remains of five structures rabbit, sage grouse, and rodent remains were also (Current 1993, cited inLubinski 1997:271). Four recovered. The vertebrae of large troutwere pre 2 x 2 m units and one 1 x 1m unitwere excavated; sent andmay have been used as beads, but no other thedepth of excavation is not known butwas appar fish remainswere recovered (Frison 1971:266). ently very shallow. Artifacts recovered include Nimmo (1971) estimated the age at death of 79 ceramics, lithics, and bone. Despite a radiocarbon Eden-Farson pronghorn and created a histogram of date of 150 ? 60 B.P. (cal A.D. 1646 to 1892) on thepopulation age structure.He thenvisually com charcoal recovered from the 1978 excavation, no pared thishistogram to age distributions of a Dall Euroamerican tradegoods were recovered.Though Sheep attritionaldeath collection and a catastrophic some 5,500 bone fragmentswere collected during death assemblage ofMerycodusfurcatus (an extinct the 1978 and 1990 excavations, it is likely thatthese Antilocapridine). The Eden-Farson histogram represent only a small proportion of thebone pre clearly resembles the catastrophic death assem sent at the site (Lubinski 1997:273). blage profile, which Nimmo (1971) and Frison Lubinski's (1997:276-277) seasonality analy (2000) accept as strongevidence thatEden-Farson sis of pronghornmandibles fromGailiun did not was the result of a very large population kill rather yield useful information.Most specimens produced than a series of attritionalkills. age estimates exceeding six months in duration. Lyman (1987) critiquesNimmo's (1971) analy Only one specimen produced a reasonable range, sis on the grounds thatold pronghorn individuals fromAugust to January.Mandible-based season are absent from the Eden-Farson assemblage. ality thereforecannot informus regarding thenum Lyman raises thepossibility thatthe pronghorn had ber ofmortality events.No old adults were present "steepened their survivorship curve" (1987:140); in theGailiun sample so the JPO age analysis is that is, thatolder individuals were absent from the not applicable, and no fawns were present so the parent population due to hunting pressure. This FYM age analysis also cannot be applied. Likewise, might cause thepopulation age structurehistogram there are insufficientsamples to produce a mean to not conform to expected catastrophic or attri ingfulhistogram of age categories. Lubinski nev tional age profiles. So, while Lyman does not claim ertheless states that "Gailiun clearly should be thatEden-Farson's pronghorn remains are neces

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 329

sarily the product of an attritional mortality Fisher and Frison assessed the season of death sequence, his analysis questions the large popula using sixmandibles and twomaxillae whose teeth tionkill scenario. had not fully erupted. They estimate a season of Lubinski (1997:254-269), as usual, begins his death between October and January,with "the pre investigationof Eden-Farson pronghornmortality ponderance of evidence pointing to somewhere with a seasonality analysis.With such a large sam during November-December" (Fisher and Frison ple, he is able to disregard pronghornmandibles 2000:91). Because of thewide range of variation whose seasonality estimate ranges exceed six in a fairlysmall sample, theywere unable to resolve months. This leaves 90 mandibles with individual whether Boars Tusk was the result of a mass kill seasonality estimates of sixmonths or less,which or an accumulation frommultiple kills. when combined produce an assemblage seasonal Lubinski (1997:284) was only able to assess the ityestimate thatspans theentire year. However, 93 season of death forone pronghornmandible, which percent of the season estimates include theOcto produced an estimate ranging fromAugust to Jan ber to Januaryperiod. Lubinski thereforeindicates uary.This is obviously of littleuse in assessing the that "most of the kills at Eden-Farson occurred siteformation process. His age-based analyses were fromfall tomid-winter" (1997:263). His age struc also hindered by small sample sizes, which pre tureanalyses produced conflicting results,with the vented him from performing the JPO or age his JPO partition conforming to an accumulation of togram analyses. His FYM analysis was "largely multiple events but theFYM partition and age his consistent" (Lubinski 1997:285) with a mass kill togrammore closely resembling a single popula mortality event, but overall Lubinski (1997:287, tion.Because two of the analyses favored a single 2000:124) considers theBoars Tusk mortality event population, Lubinski considers Eden-Farson tobe analysis to be inconclusive. "more consistentwith a singlemortality event than Trappers Point multiple events" (1997:267). Trappers Point (48SU1006) is a large,multicom Boars Tusk ponent pronghorn bonebed site just west of The Boars Tusk site (48SW1373) is in theGreen Pinedale, Wyoming, thatwas partially excavated River Basin, about 2.4 km southeast of thepromi between 1990 and 1992. Excavation results have nent volcanic formationof the same name (SHPO been reported (Miller et al. 1999a; Miller and 1979). Itwas partially excavated in 1981, and sum Sanders 2000; Sanders andMiller 2004) and analy mary resultswere reported by Fisher and Frison sis continues (Clayton and Kunselman 2002; Fen (2000; see also Lubinski 1997:279-287). Boars ner and Walker 2008; Leopold and Vita-Finzi Tusk contains at least 11 concentrations of heat 2005). Pronghorn remains were recovered from altered rock, artifacts,bone, and charcoal laid out several components, but the currentanalysis only in a roughly linear,or perhaps oval, configuration. concerns bones from StratumV, an Early Archaic Five of these concentrations have been excavated. component that contains the most extensive Boars Tusk is considered a Late Prehistoric or Pro bonebed at the site.The StratumV excavation pro tohistoricsite based on diagnostic projectile points duced many stone tools, several features,and abun and a single radiocarbon date of 100 ? 80 B.P. (cal dant animal bone. About 85 percent of the identified A.D. 1653 to 1919). Not all of the artifacts from bone is from pronghorn; the remainder is distrib Boars Tusk were included in the summary publi uted in low percentages among bison, elk, deer, cation, but in addition to projectile points they grouse, and smallmammals such as voles and rab include at least six bifaces, fourunifaces, a scraper, bits. One goose bone and one fish bone were also a core, a mano, and four bone beads. No Euroamer recovered. Except as specifically noted, all refer ican trade goods are reported.Pronghorn remains ences toTrappers Point in this article refer specif dominate the faunal collection, with at least nine ically toTrappers Point StratumV. pronghorn individualspresent (Lubinski 1997:282). Miller et al. (1999a) used several avenues of The only other identifiedanimals are one rabbitand evidence to assess the number of Trappers Point one or two rodents, though threebison-sized but mortality events.Their general conclusion was that unidentifiable bones were also recovered. the bonebed likely represents a single mortality

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 330 AMERICAN ANTIQUITY [Vol. 74, No. 2, 2009]

Figure 3. Histogram of combined Trappers Point Stratum Y radiocarbon date probability distributions using 18 bins. Redrawn from Francis and Sanders 1999:Table 5.2, which is based on a 1993 version of Kintigh's C14 program. event, or possibly a series of eventswithin a single This issue can more profitably be addressed season of one year.The evidence did not, however, using a statistical approach. Using Wilson and permit firm conclusions: "Unfortunately, investi Ward's (1981) radiocarbon date clustering analy gators cannot demonstrate unequivocally whether sismethod, the 11 StratumV radiocarbon dates are = or not the bonebed represents a single cohort kill not a single cluster (T' 21.97;/? < .02; test includes or separate closely spaced springevents in a single sunspot effectuncertainty for all samples). year" (Miller et al. 1999b:283). The radiocarbon data may also be examined Eleven radiocarbon dates fromTrappers Point using calibrated dates (Figure 4). The 95 percent rangefrom 5160 ? 210 to6010 ? 130B.P. The prob confidence intervalsof some dates do not overlap. abilities associated with thenormal distributionsfor In particular, sample 79197's intervaldoes not over these dates were summed and graphed, resulting lapwith the intervalsfor 79195 or 79191. The inter in a "fairly steep unimodal curve" (Francis and val for 79191 also does not overlap with that of Sanders 1999:49) (Figure 3). This is interpretedas 79194 or 53885. Assuming thedated material was supporting a single occupational event, because a deposited simultaneously with thebonebed itwas continuous occupation over the entire time span foundwithin, the lack of overlap indicates that the would likely produce a flatterdistribution. How bonebed resulted from at least two depositional ever, one could also make a case that the distribu episodes. tion in Figure 3 is, in fact, fairly flat and bimodal Miller interpretsan age distribution histogram with modes at 5700 and 5500 B.P. As the apparent based on fetalbone and tootheruption/wear analy bimodality could be an artifactof the largebin size ses (Figure 5) as indicating "catastrophic [i.e., non used in Figure 3, the 11 Trappers Point radiocar natural] mortality of pronghorn dominated by bon date probability distributionshave been com mature, breeding age individuals" (Miller bined and graphed using a much finer-grained 1999:240). An age structuredominated bymature, histogram binning size (not shown; see Fenner breeding individuals suggests hunter selection for 2007:95). It is clear from thefine-grained histogram prime animals (and thereforean accumulation from that there are at least two truemodes, with a pos multiple kills) rather than a mass kill of all indi sible thirdmode around 6000 B.P. There are, of viduals in a population. However, thenear absence course, any number of intermittentoccupation pat of pronghorn over three years of age raises con terns that could produce such a distribution.Fur cern. As Lyman (1987:140) pointed out for the thermore,Bartlein et al. (1995) has shown that, in Eden-Farson sample, pronghornmay have had a some circumstances, modality in uncalibrated steep survivorship curve, perhaps due to predation radiocarbon date histograms can be the result of by humans. This concern, in combination with the fluctuations in atmospheric radiocarbon produc known imprecision involved in aging pronghorn tion as reflected in non-linearity in calibration based on toothwear (Lubinski 2001; Miller et al. curves, and thereforemodality is not necessarily 1999b:483) and likelyvariability in age profiles due related to thenumber of occupations. to natural events such as severe winters, prevents

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 331

79197

39687

79194

53885

79192

79190

79196

79193

79195

79189

79191

5000 4900 4800 4700 4600 4500 4400 4300 4200 4100 4000

Years (B.C.)

Figure 4. Calibrated Trappers Point Stratum V radiocarbon dates. The 95% confidence intervals are shown. The only AMS-measured date is the second from the bottom. Data from Francis and Sanders 1999:42. Beta samples 39687,53885, 79189-79197. Calibration from BCAL 2006. age profiles from being used to assess a hypothe considerable number of years. sis of multiple pronghorn accumulations at Trap Trappers Point fetal pronghorn mortality sea pers Point. sonally estimates are compatible with a single kill Miller et al. (1999a) also examined maps of the event, though thereare hints of structurewithin the vertical distributions of pronghorn bones. These date estimate distribution thatsuggest two or three maps show a "single, relatively compact, 10 to 20 kill eventsmay have occurred (Fenner 2007; Fen cm thickbone layer.There is no evidence for ver ner andWalker 2008). Of course, the fetal prong tically separate bone layers,which agrees with the horn data only addresses kills of mature pregnant field observations" (Sanders 1999:389). This is females in latewinter and spring.Other kill events strongevidence for either a single kill or accumu involvingmales or occurring during other seasons lations thatwere very closely spaced in time.The would not be detected through analysis of fetal are maps in question, however, drawn east-to-west remains. and thereforedo not address potential vertical sep To summarize the results forTrappers Point so aration along other axes. Sediment deposition rates far,radiocarbon dates indicate thatthere were mul are unknown in the Trappers Point area and the tipleoccupations separated by substantial timeperi bones may have been on or near the surface for a ods, pronghorn age distribution data is not vertical distribution but 10 -, informative, maps support do not demonstrate a single kill or multiple kills fairlyclosely spaced in time, and fetal seasonality data are compatible with a single kill event but do not rule out additional events. 2 This review demonstrates that the pronghorn population formationprocesses at these six sites is difficultto interpretusing current techniques. Sea sonality and age distribution patterns are often inconclusive or conflicting, and radiocarbon date Fetal 1 23456789 cluster analysis may not be precise enough (ormay Age Group lack sufficientsamples) to resolve the issue. Iso ratio variation offers an Figure 5. Trappers Point pronghorn age profile. Redrawn tope analysis independent means from Miller 1999:Figure 11.14. of assessing prey population characteristics

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 332AMERICAN ANTIQUITY [Vol. 74, No. 2,2009]

that, in combination with other approaches, may pronghorn enamel therefore are controlled by a permitmore reliable assessment ofwhether a spe number of factors. InWyoming, the ratios appear cific archaeological bonebed was the result of an to be more strongly influenced by aridity thanby accumulation of pronghorn frommultiple popula temperature or other factors, and show a geo tions. The following sections introduce relevant graphical pattern that is similar but not identical to aspects of pronghorn behavior and sources of iso the carbon isotope ratio geographic pattern (Fen tope variation, and then report on the analysis of ner 2007; Fenner and Frost 2009). pronghorn isotopic variation at each of the six Strontium is a heavy elementwhose isotopes do archaeological sites discussed above. not undergomeasurable biological fractionation.In mammals, thismeans that strontiumisotope ratios in tissues reflect the ratios in the diet, Pronghorn Behavior and Isotope Chemistry preserved which in turnreflect the ratios present in the sur Though often called antelopes, pronghorn are not face geology underlying the lowest component of closely related to true Eastern Hemisphere the food chain (Bentley et al. 2002; Ericson 1985; antelopes but instead are indigenous to theprairies Price et al. 1994, 2002). Pronghorn strontiumiso of North America. Aspects of theirbehavior rele tope ratios are therefore informative of the geo vant to isotope analysis have been presented else graphic region fromwhich a pronghorn's food and where (Fenner 2007,2008) andwill only be briefly water was derived. Modern Wyoming pronghorn summarized here. More detailed discussions of groups show a fairlywide range of geographic vari pronghorn biology and behavior are also available ation, reflectingthe area's diverse geology (Fenner (McCabe et al. 2004; O'Gara and Yoakum 2004). 2007). Pronghorn are ruminants that forage primarily on Pronghorn groups shift in both size and com shrubs (most of which are sagebrush species in position over the course of a year (Byers 2003; Wyoming) with forbsand occasionally grasses pro O'Gara 2004). During thefall mating season, bucks viding smaller contributions to the diet (Yoakum attempt to gather does intoharems, which must be 2004a:465-466). Most such plants utilize theC3 defended from the advances of other bucks. Does photosyntheticpathway, resulting incarbon isotope will often leave a harem to join thatof amore desir ratios thatrange fromabout -20 to-35 permil (%o) able buck. Following mating season, does and (Ehleringer and Rundel 1989). Pronghorn enamel bucks gather into large groups, sometimes includ thereforeshows a strongC3 signal (afteradjustment ing hundreds of individuals. By lateMay, these for dietary fractionation).Within the C3 range, groups will have broken up and pregnant does will pronghorn carbon isotope ratios have been shown find isolated locations to give birth to twins.After tohave significantgeographic variation both across birthing season, pronghorn often re-gather into Wyoming in general and within southwestern small groups or "bachelor herds" of only a few Wyoming inparticular, probably as a resultof con individuals. Pronghorn groups thusmerge and shift suming plants that grew in differing humidity both seasonally and reproductively,and it is diffi regimes (Fenner 2007; Fenner and Frost 2009). cult to definewhat ismeant by a "pronghorn herd" Oxygen isotope ratios inmammal tissues are over the course of a year.However, pronghorn typ derived frombody water, which includes drinking ically have relatively small home ranges, averag water,water obtained directly fromfood, andmeta ing less than 70 square kilometers (O'Gara bolic water produced by oxidation of food (Koch 2004:170). Some pronghorn migrate seasonally, 1998). The oxygen isotope ratios in rain and snow but usually return to the same ranges. For exam are determined by temperatureand the number of ple, Hoskinson and Tester (1980) found thatall 22 previous rainouts, andmay also be affectedby arid radio-collared pronghorn in Idaho returned to the ity (Clark and Fritz 1997; Higgins andMacFadden same range in consecutive summers. Itmay be that 2004; Levin et al. 2006). Prior to consumption by home range consistency is related to the need to pronghorn, oxygen isotope ratios in local water know dominance relations, which for does are may be furtheraffected by other factors, such as largely established at birth (Byers 1997:102-118). a evaporation enrichmentwithin plant leaves or trans In any case, while the composition of particular port via river channels. Oxygen isotope ratios in groupmay vary,pronghorn within a particular large

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 333

area tend to remain or return there.Hunt areas in topic variation to distinguish accumulations from Wyoming covering hundreds of square kilometers single-population kills. includemultiple entirehome ranges,but for thepur While analysis ofmodern prey populations may poses of this study are treatedas containing a sin adequately model spatial variation, temporally gle pronghorn population. This may overestimate based isotopic variation ismore difficultto predict. the isotopic variation within a truemodern "herd," Climatic models may suggest changing environ butwill minimize thechance of false positive (mul mental conditions thatwould be expected to pro tiple kill) findings. A low percentage of modern duce differingmean carbon or oxygen isotope pronghorn in southwestern Wyoming undergo ratios, but rarely provide adequate temporal reso long-range seasonal migrations that (ifperformed lution to predict the isotopic variation potential of during enamel formation) could affectenamel iso a single bonebed. Note, however, thatregardless of tope ratios.More importantlyfor thecurrent analy whether isotopic variationwithin a bonebed is ulti sis, the potential for congregation of multiple mately derived from spatial or temporal variation, populations duringmigration must be considered if it substantially exceeds the variation expected when applying the technique to archaeological within a population existing at a single place and sites. time, then thebonebed likely results from an accu Pronghorn roamwild inWyoming, but thatdoes mulation ofmultiple populations. notmean theyare unaffected by humans. They are For certain prey animal species, including managed as a game animal, and pronghorn popu pronghorn, it ispossible thata multiple-population lation size inWyoming has fluctuated with hunt accumulation was the result of a single intercep ing pressure as well as weather and nonhuman tionof populations thathad temporarilyaggregated predator pressure. Unlike deer, pronghorn avoid while migrating ratherthan being theresult of mul jumping over fences (although theyoften are able tiplekill events. Isotopic analysis alone may not be topass underwire-strand fences) and thereforeare able to distinguish between these two circum somewhat restricted in theirmovements in devel stances. This is particularly a concern for theTrap oped areas of the state (Karhu andAnderson 2006; pers Point site, which is located at a modern Yoakum 2004b). Interstatehighways, particularly pronghornmigration bottleneck. 1-80 in southernWyoming, also restricttheir move ments. Nevertheless, pronghorn are clearly wild. Samples They are not routinelyprovisioned, and isotopic dif ferences among modern hunt areas should reflect The modern comparative sample is a collection of trueenvironmental or behavioral differences rather male pronghorn firstincisors obtained fromhunters thanmanagement differences. by theWyoming Game and Fish Department Isotope ratio analysis will only produce evi (WGFD) during late September and earlyOctober dence of accumulation when themultiple prey pop of 2004 and 2005 as part of a separate study.The ulations representedwithin the accumulation have data collected byWGFD included thehunt area in substantially differentmean isotope ratios for at which each pronghornwas killed.Wyoming hunt least one element. This requires spatial or tempo areas vary in size and terrain. Some hunt area ral variation in theclimatic, geologic, or behavioral boundaries are natural formations such as moun conditions thatcontrol prey isotopic composition. tainranges orrivers while otherboundaries are arti Analysis ofmodern pronghorn shows substantial ficial constructions such as highways. Tooth variation across Wyoming for all three elements cementum-based age determination performed by under consideration (Table 2; see also maps inFen WGFD shows that the sample has an average age ner 2007; Fenner and Frost 2009). This was of 4.7 ? 1.9 years,with a range of 1.5 to 10.5 years. expected because Wyoming's varied topography Sample size varies by hunt area, ranging from one and geologic history produces a wide range of cli to 24 samples (Table 2). matic and geologic conditions across the state Six archaeological sites in southwestern (Knight 1994). Areas with more uniform environ Wyoming that contain substantial pronghorn mental and geologic conditions, such as portions bonebeds are included in the analysis (Table 1). In of theGreat Plains, may not produce sufficientiso addition, a virtual site ismodeled using random

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 334 AMERICANANTIQUITY [Vol. 74, No. 2, 2009]

Table 1.Archaeological Sites Analyzed.

SiteApproximate Pronghorn PronghornHunt Site Name_Identifier_Age_MNI_NISP_Area AustinWash 48UT390 1197BP 188094 18 Boars Tusk 48SW1373 Historic Period 71792 9 Eden-Farson 48SW304 Historic Period > 92226 FireholeBasin 48SW1217 630 BP 43359 26 Gailiun 48SU1156 150BP 91 8 to Trappers Point 48SU1006_6010 5160 BP_27_1683_8^_ The Trappers Point component of interest is Stratum V. Sources: Eckles 1999; Fisher and Frison 2000; Lubinski 1997; Middleton et al. 2007; Miller et al. 1999a; Schroedl 1985. Blanks indicate thatNISP data is not currently available for Eden-Farson or Gailiun.

samples frommodern hunt areas (seeMethods sec likelybe less thanotherwise would be thecase. This tion formore information). Finally, the Reiser favors theposition of not identifyinga site as highly Canyon natural mass die-off site is included for variable and thereforenot likely the result of an comparative purposes. In 1991, about 150 prong accumulation frommultiple populations. horn died afterfalling over a cliff inReiser Canyon, near of which is located the southern edge Hunt Methods Area 96 (Lubinski and O'Brien 2001; Ottman 1992). Pronghorn sexwas not recorded during col Prior to isotope ratio analysis, each sample under lection of theReiser Canyon sample, but presum went a pretreatmentprocedure intended to remove ably includes a mix of males and females. organics and other adheringmaterials. Each tooth Comparing the all-male hunt area samples to pre was mechanically cleaned, then enamel was col sumably mixed-sex samples fromReiser Canyon lected by using a Dremel drill to carve a groove and archaeological sites introducesuncertainty, par along the length of the buccal enamel surface of ticularly since there is some evidence that,during the tooth.Enamel development proceeds from the summer,female pronghorn range over a largerarea occlusal surface to the cervix (Hillson 1986:114), than do males (Hoskinson and Tester 1980; so sampling along the tooth lengthprovides an iso Reynolds 1984). However, males and females are topic ratio thataverages theentire period of enamel together inmixed groups most of the year, and formation (excluding any loss due to toothwear probably consume similar isotopic regimes. during life). This reduces the influence of short Tooth enamel was used in all reported isotope term isotope ratio variations due to seasonal fluc analyses due to concerns with potential diagenetic tuations or nursing by fawns.Using theprocedure contributions to isotopic variability in bone sam described by Koch et al. (1998:125; see also Koch ples. All teeth selected for isotope analysis are et al. 1997), the enamel was then soaked inbleach incisors. Incisors have tapered root structuresthat followed by .1N acetic acid. Koch et al.'s proce allow them to easily slip out of theirmandibular dure was modified slightlyby reducing soak time sockets following post-mortem tissue degenera in response toGarvie-Lok et al.'s (2004) work that tion, and therefore incisors are commonly found showed long soak timesmay cause substantial sam loose in archaeological contexts ratherthan embed ple loss and possible isotopic inaccuracy due to ded in bone. It is often difficult to assign side and recrystallization.Details of the pretreatmentpro toothnumber to loose pronghorn incisors, so there cedure are provided elsewhere (Fenner 2007). is a possibility thatmore than one tooth from an Carbon and oxygen isotope ratios inenamel car individualmay be included in archaeological sam bonate were measured using continuous-flow iso ples. Every attemptwas made to obtain teethfrom tope ratiomass spectrometryat theUniversity of separate individuals by comparing overall size, Wyoming Stable Isotope Facility and are reported wear amount and intra-sitelocation (cf.Klein and inparts permil (%o)with respect to theV-PDB stan Cruz-Uribe 1984:26-27). If two teethfrom an indi dard. For compatibilitywith archaeological analy vidual nevertheless are included, the resulting iso ses, 1.5\%c is added tomodern carbon isotope ratios topic variationwithin thearchaeological sitewould to account for themodern carbon ratiooffset, which

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 335

Table 2. Hunt Area Isotopic and Environment Data.

513C 8180 Mean Mean Mean Mean Hunt 813C 813C Std 8180 8180 Std 87Sr/86Sr 87Sr/86Sr 87Sr/86SrMay-Oct Annual May-Oct Annual Area n Mean Dev n Mean Dev n Mean Std Dev Temp. Temp. RH RH 17 2 -11.74 .66 2 -9.01 1.35 15.5 6.9 53.5 58.6 19 3 -11.98 .33 3 -9.96 1.08 15.5 7.0 54.4 59.8 23 7 -11.55 .45 6 -10.01 .47 15.1 6.8 51.4 57.0 24 8 -11.35 .60 8 -8.27 1.17 15.3 7.0 52.6 57.8 32 1 -11.06 1 -9.78 14.4 6.5 50.0 55.6 42 13 -11.07 .45 13 -5.96 1.36 13.1 5.6 50.1 53.8 43 1 -10.91 1 -8.36 12.6 5.2 49.0 51.8 46 4 -10.81 .65 4 -6.98 .44 12.9 5.6 48.5 54.8 47 16 -10.83 .40 17 -6.81 1.20 .71153 .00140 12.9 5.3 49.4 55.9 53 16 -10.99 .51 17 -7.01 1.71 13.5 5.7 52.1 59.3 57 12 -10.34 .81 12 -2.81 1.18 .71234 .00071 13.4 5.6 47.0 56.6 59 3 -11.01 .65 3 -6.05 1.12 .71200 .00153 13.4 5.5 45.5 55.7 61 1 -10.06 1 -7.04 13.6 5.6 48.3 58.0 64 1 -11.64 1 -7.36 12.8 4.8 46.7 55.9 65 2 -10.82 .49 2 -5.43 .53 10.4 3.0 48.7 54.6 66 4 -11.21 .65 4 -7.44 1.05 15.0 6.6 45.0 52.2 75 1 -11.53 1 -7.21 14.6 6.3 47.1 53.8 87 8 -10.56 .53 8 -8.34 1.78 .71200 .00090 7.2 -.4 47.7 52.8 88 9 -10.55 .87 9 -7.92 2.21 .71114 .00064 8.7 1.1 42.9 49.2 89 10 -10.28 .75 9 -6.98 1.22 9 .71086 .00041 8.9 1.2 42.1 49.3 90 10 -10.17 .90 10 -5.90 1.05 3 .71209 .00098 10.4 2.2 43.3 50.5 91 24 -10.37 .68 24 -8.31 2.46 10 .71278 .00072 10.7 2.4 45.2 53.4 92 17 -10.18 .66 17 -7.59 1.59 10 .71381 .00129 12.4 4.1 44.0 53.8 93 4 -9.47 .86 4 -4.55 3.31 3 .71041 .00017 11.1 3.0 42.0 50.4 94 1 -9.31 1 -4.98 1 .71037 12.8 4.6 42.3 51.4 95 3 -10.45 .33 3 -3.92 1.54 3 .71008 .00013 13.9 6.0 42.9 51.9 96 10 -10.09 .99 10 -6.32 2.27 4 .71121 .00013 12.8 4.2 42.5 51.3 106 2 -10.99 .61 2 -7.36 1.41 14.1 6.1 47.8 56.0 107 6 -10.68 .54 6 -7.85 .91 4 .71581 .00114 10.3 2.3 46.4 54.1 112 1 -10.12 1 -5.92 1 .71176 13.1 5.4 46.0 55.3 115 4 -11.07 .66 4 -7.29 1.12 1 .71006 13.4 5.7 49.3 55.0

Outliers excluded. 813C mean values increased by +1.51%o to account formodern carbon offset. 813C and 8180 values are per-mil VPDB. Blank indicates no measured values, or insufficient samples to compute standard deviation, n: number of samples; Temp.: temperature in C; RH: relative humidity in%. Temp, and RH means computed by averaging raster pixel values area areas within hunt after eliminating above 2800m elevation; raster data provided by Spatial Climate Analysis Service at Oregon State University and Climate Source Inc.

= is likely produced by a recent change in atmos 268) and oxygen (n 267) ratios. Results are pheric C02 composition due to theburning of fos reported as unitless 87Sr/86Srratios. Twelve NIST sil fuels (Tieszen and Fagre 1993). Strontium 987 strontiumstandards were run,with measured isotope ratios were measured using thermal ion values averaging .71026 ? .00001. This indicates mass ization spectrometry (TIMS) at theUniver that strontiummeasurement uncertainty iswithin sityof Wyoming IsotopeGeology Laboratory. Prior the fifthdecimal place. Intra-individual variation toTIMS analysis, enamel strontiumwas separated was .17 ? A4%c for carbon, .61 ? 33%c for oxy from other elements (notably rubidium, which gen, and .00003 ? .00004 for strontium. overlaps with strontium at mass 87) using a Post-depositional alteration of the chemical strontium-specific cation exchange resin. The properties of a sample is always a concern when strontium separation procedure is described in analyzing isotope ratiosfrom archaeological sites, detail inFenner (2007). Due to this lengthyprocess, and has received much research attention (for were fewer teeth processed for strontium isotope reviews of thiswork, see Hedges 2002; Hoppe et ratios = was case = (n 107) than the for carbon (n al. 2003; Kohn and Cerling 2002; Lee-Thorp and

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 336 AMERICAN ANTIQUITY [Vol. 74, No. 2, 2009]

Oxygen Isotope Ratio -16 -15 -14 -13 -12 -11 -10 -9 -8 -f? ?I?

Austin Wash Enamel AustinWash Sediment

Figure 6. Carbon and oxygen isotope ratios of Sediment (filled diamonds) and pronghorn enamel (empty squares) from Austin Wash (48UT390). Oxygen isotope ratio is 8lsO (per mil); carbon isotope ratio is 813C (per mil).

Sponheimer 2003). The current consensus is that quartile are considered outliers and removed from thenon-porous, crystalline structureof enamel pro furtherconsideration. While thiscriterion is some tects its hydroxyapatite from significantdiagene what arbitrary, it is a commonly applied rule of sis inmost situations, even over millions of years thumb for identifying suspected outliers (Moore (e.g.,Hodell et al. 2004; Knudson et al. 2004; Lee andMcCabe 1999:47). As a result,out of a total of Thorp and Sponheimer 2006). A qualitative check 667 measurements from 276 individuals, 13mea for diagenesis was performed by comparing the surements (1.9 percent) were discarded. For multi isotope ratios inpronghorn enamel from theAustin element analyses, all isotope values for these 12 Wash site to the ratios contained in the surround individuals have been excluded. Basic descriptive ing sediment carbonate. The carbon and oxygen statisticswere then computed for the remaining sam isotope ratios in carbonate from six sediment samples. ples saved from the 1981 Austin Wash excavation Standard deviation was used as the basic mea were measured and compared to the ratios in sure of spread or variability foreach group and iso enamel from the site.The ratios in the sediment are tope ratio. Each group's isotope ratio values were distinct from the ratios in the enamel (Figure 6), centered about the origin, and multiplied by an and thereforediagenetic alteration of isotopic com element-specific scaling factor so that all isotope position is not indicated. ratios use comparable scales. The three scaled iso tope ratios foreach individualwere thencombined Statistical Techniques into a three-dimensional Euclidean distance, and The statistical approach used in this study is essen thesedistances used to assess variabilitywithin and tially the same as is described in detail in Fenner among groups. Some hunt areas lack measured (2008; see also Fenner 2007:59-62). Isotope ratio strontiumratio samples but include sufficientcar values lyingmore than 1.5 times the interquartile bon and oxygen samples for comparison with range below the first quartile or above the third archaeological sites, so two-dimensional carbon

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 337

Table 3. Archaeological Site Summary Isotope Results.

813C (%oVPDB) 8180 ($oVPDB) 87Sr/86Sr Std Std Dev Site_n_Mean_Std Dev_n_Mean Dev_n_Mean Austin Wash 7 -10.15 1.20 6 -9.90 1.84 6 .71065 .00006 Boars Tusk 6 -11.51 .72 6 -9.10 .84 6 .71190 .00047 Eden-Farson 12 -8.84 1.00 12 -10.54 1.78 7 .71175 .00061 FireholeBasin 7 -8.65 .69 8 -12.60 2.37 7 .71196 .00098 Gailiun 8 -9.19 .76 8 -9.83 1.47 7 .71176 .00053

Trappers Point 19 -10.02 1.05_19 -9.04 2.06_6 .71290 .00034 = Outliers excluded, n Number of Samples.

oxygen Euclidean distance values are also reported. from the parent population. That is, it represents This study also investigates the isotopic vari the amount of geographically based variability one ability among groups throughnon-parametric boot would expect if themodeled site had been accu strapping (see Buntinas and Funk [2005:471-502] mulated from all hunt areas (spanning the entire for general discussion regarding such techniques). state, thoughmost samples are from southwestern Non-parametric bootstrapping, sometimes referred Wyoming) ratherthan as amass kill of a single hunt to as resampling, can provide an estimate of the area group. This virtual site can be thoughtof as probability that a sample of a certain size drawn representing an average random geographic accu froma population would be at least as variable as a mulation from a verywide area. As such, itwould given amount.When applied to thisproject, thepop be expected to be more variable than any single ulation data are isotopic distance values for a par modern hunt area. Because the virtual site is an ticularmodern hunt area.A subsample of a certain averaged value, an archaeological site thatwas in size is randomly selected from this group and the fact accumulated frommultiple populations may standarddeviation of distance values for thatgroup have somewhat less variability than thevirtual site. is computed. This subsampling process is repeated The relativemagnitude of virtual site and archae 1,000 times for each hunt area. The standard devi ological site variabilitywill thereforebe used as a ation of isotope distances obtained from a specific meaningful but not conclusive indicator when archaeological site of interestis then compared to assessing the significance of isotope variability, the listof 1,000 subsample standard deviation val along with comparisons tomodern hunt area and ues for each hunt area, and thenumber ofmodern Reiser Canyon sample variability. subsamples whose variability is less than thatof the site of interest is determined. This archaeological Results number,when divided by 1,000, is interpretedas theprobability that the archaeological site ismore After removing outliers, isotope ratio summary variable than themodern hunt area.An Excel-based descriptive data were calculated for each archaeo Visual Basic program for performing these com logical site (Table 3). Interestingly,there is no sig putations is available from the author. nificant correlation between isotope ratios for any Another aspect of the investigationwas model two elements (Kendall's p, all /?> .09). This is ing a virtual site for comparison to archaeological somewhat surprising,since both 813C and 8180 are sites. To create the virtual site, one instance of a driven to some degree by climate. The lack of sig virtual pronghorn group is created by randomly nificant correlationmay be the result ofmicrocli selecting a pronghorn subsample from among all mate differences among the areas inhabited by modern hunt area pronghorn,and the individual iso pronghornduring enamel formationor, more likely, tope distances computed. The standard deviation is simply an artifact of small sample size and a of thisgroup of pronghorn distances is thencalcu weak truecorrelation. Analysis of element isotope lated, and this process repeated a total of 20,000 ratio correlations among modern hunt areas sup times.The average of the group isotopic distance ports the latterhypothesis; there is no significant standard deviations is thenused as an estimate of correlationbetween 813C and 8180 using pronghorn the standarddeviation of a virtual site accumulated from six hunt areas in southwesternWyoming

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 338 AMERICAN ANTIQUITY [Vol. 74, No. 2, 2009]

Table 4. Archaeological Site and Hunt Area Means Comparisons.

o13C o180 87Sr/86Sr Hunt Arch. Site_Area_N_t_Sig._N_t_Sig._N_t_Sig. Boars Tusk 92 6,17 -3.994 .00 6,17 -2.923 .01 6,10 -4.26 .00 Eden-Farson 92 12,17 4.031 .00 12,17 -4.572 .00 7,10 -4.41 .01 Firehole Basin 59 7,3 5.159 .01 8,3 -6.202 .00 7,3 -.03 .98 Gailiun 91 8,24 3.944 .00 8,24 -2.095 .05 7,10 -3.67 .00 Trappers Point 87 17,8 1.702 .10 17,8 -.863 .40 6,8 6.67 .00 Trappers Point 88_17,9 1.368 .19 17,9 -1.259 .23 6,7 2.46 .04 The hunt area containing theAustin Wash site has only one sample and could not be included inmeans comparisons. Modern 813C values increased by 1.51%o prior tomeans comparison to account formodern carbon offset. Trappers Point is close to the border between Hunt Areas 87 and 88 so both were tested. N: Number of site samples, number of hunt area samples, t: independent samples t-test, equal variances not assumed. Sig.: Two-tailed significance probability; bold type = face indicates statistical significance at a .05. Non-parametric Mann-Whitney U tests provide essentially the same results, except theGailiun/HA 91 8180 means are not significantly different (p =.17).

= = (Kendall's p,p .53, n 6) even though the larger (More quantitative assessments of comparative state-wide sample of modern hunt areas shows a variation produced using bootstrap analysis are pro significant though not particularly strong correla vided for individual sites in the Discussion sec = = tion (Kendall's p .42,p < .01, n 22). tion.)The virtual site inFigure 7 was derived from The mean isotope ratios frommost archaeo random selection of pronghorn from all modern logical sites do not closely match the ratios of their hunt areas, and serves as an indicator of a level of modern hunt areas (Table 4). For carbon and oxy variation tobe expected in an accumulated site.The gen isotopes, this could be related to climatic Reiser Canyon site,which was formed by a mod changes; for instance, colder conditions during the ernmass pronghorn die-off, serves as an indicator Little Ice Age could have caused more depleted of the level of variation to expect in a mass kill sit mean oxygen isotope values at the Eden-Farson uation, as do modern hunt areas. site. Changes inmean environmental aridity also to likely contribute prehistoric-to-modern differ Discussion ences, particularly for carbon isotopes. Strontium differences perhaps indicate prehistoric pronghorn The distance metrics show some interestingpat migration or human prey transportbehavior. The ternsamong archaeological sites (Figure 7 and Fig implications of thesemean differenceswill be fur ure 8). Austin Wash and Trappers Point show the therdiscussed in the individual archaeological site most variation among archaeological sites in both discussion sections. the three-dimensional and two-dimensional dis As discussed in theMethods section, individ tances. The variation at each of these sites is sim ual site and hunt area isotope ratioswere scaled and ilar to or greater than thatof thevirtual site,which converted to three-and two-dimensional distances. was assembled as an example of a geographically The standard deviation of each group's carbon based accumulation site, and substantially exceeds oxygen-strontium distances was then computed thatof Reiser Canyon and most hunt areas, which (Table 5, Figure 7). The two-dimensional carbon are considered single-population assemblages. On oxygen distance standard deviations, forwhich a the other hand, theGailiun, Firehole Basin, and larger sample is available, were also calculated Boars Tusk sites show relatively little variation, (Figure 8). These standard deviations of hunt area while theEden-Farson site is intermediatebetween isotopic distances measure the amount of prong the two groups. Overall, Figure 7 and Figure 8 indi horn isotopic variability within each area. An cate thatAustin Wash and Trappers Point are likely archaeological site isotopic distance standarddevi to be accumulations frommultiple populations, ation that substantially exceeds modern hunt area Gailiun, Firehole Basin, and Boars Tusk show lit distance standard deviation values indicates that tle evidence ofmultiple-population accumulation, the sitewas formed throughaccumulation of prong and Eden-Farson shows weak support for accu horn frommultiple temporal or spatial populations. mulation frommultiple populations.

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 339

Table 5. Scaled Isotope Distance Measures.

Combined 513C, 5180 and 87Sr/86Sr Combined 813C and 8180

Average Standard Average Standard Location N Distance Deviation N Distance Deviation

Modern Hunt Area 19 3 1.03 .59 Hunt Area 23 6 1.18 .52 Hunt Area 24 8 1.72 .81 Hunt Area 42 13 1.67 .58 Hunt Area 46 4 1.48 .62 Hunt Area 47 15 1.47 .61 Hunt Area 53 16 2.02 .80 Hunt Area 57 3 1.69 .26 12 2.14 1.12 Hunt Area 59 3 2.68 1.31 3 1.60 .72 Hunt Area 66 4 1.61 .85 Hunt Area 87 7 2.62 .89 8 1.95 .95 Hunt Area 88 8 3.24 1.18 9 2.74 1.43 Hunt Area 89 8 2.07 1.11 9 2.18 1.21 Hunt Area 90 3 3.87 1.33 10 2.17 1.36 Hunt Area 91 10 2.87 1.12 24 2.54 1.13 Hunt Area 92 10 3.06 1.00 17 2.12 .99 Hunt Area 93 3 2.84 2.31 4 3.06 1.95 Hunt Area 95 3 1.39 .58 3 1.38 .57 Hunt Area 96 4 2.01 1.08 10 2.92 1.71 Hunt Area 107 4 2.29 .68 6 1.43 .71 Hunt Area 115 4 1.72 .67

Reiser Canyon 9 3.12 .76 15 2.90 1.18 Virtual Site 7 N/A 1.38 7 N/A 1.25 Archaeological Austin Wash 5 2.46 1.52 7 3.09 1.67 Boars Tusk 4 1.60 .97 5 1.82 1.11 Eden-Farson 6 2.94 1.23 12 2.88 1.23 Firehole Basin 7 2.69 .89 7 2.04 1.00 Gailiun 7 2.00 .92 8 2.06 1.22 Trappers Point 6 2.45 1.37 17 2.83 1.95

The standard deviation columns contain the key variability measures. N: Number of samples. For the virtual site, TVis the number of subsamples per iteration.

This overall assessmentmust be evaluated in full mass kill. This can be investigatedmore quantita contextbefore assigning a particular formationhis tively using bootstrap analysis. Note that both tory to a site. This includes consideration of non carbon-oxygen-strontium distances and carbon isotopic formation evidence (Table 6) and a oxygen distances atAustin Wash aremore variable quantitative assessment of relative isotope varia than at least 90 percent of subsamples from Reiser tion.The remainderof thissection reviews theover Canyon and all hunt areas (Table 7). In addition, in all site formation evidence based on isotopic data every instanceAustin Wash ismore variable than as well as traditional analyses of each of the six is the virtual site,which of course is known to be archaeological pronghorn bonebed sites included based on an accumulation from multiple popula in this study. tions. It appears therefore thatAustin Wash is an accumulation of multiple populations rather than Austin Wash theresult of a singlemass kill. This is in agreement As noted previously, thecarbon-oxygen-strontium with Lubinski's (1997) analysis, though not with and carbon-oxygen results (Figure 7 and Figure 8) Schroedl's (1985) initial assessment. seem to support thedesignation ofAustin Wash as At thispoint, itmay be informativeto consider a multiple-population accumulation rather than a AustinWash's isotope values in thecontext of sur

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 340 AMERICAN ANTIQUITY [Vol. 74, No. 2, 2009]

3 M i " 5 .t: .2 S *3? ^ '? .-: s ? tS If 9 ?* a O .? ? fa a. ? ? 3 3 ta * ? * ^ J?

Location

Figure 7. Comparison of carbon-oxygen-strontium scaled distance metric variation for modern hunt areas, the Reiser Canyon sample, a virtual site, and six archaeological sites. Minimum four samples per hunt area.

rounding hunt area isotope values. Austin Wash analysis of Firehole Basin isotope variability (Fig lies within modern Hunt Area 94, forwhich there ure 7 and Figure 8) supports Lubinski's assess is unfortunatelyonly a singlemodern sample avail ment. Firehole Basin has the leastoverall variability able. It isnot farfrom Hunt Area 95, which has three of any of the archaeological sites.More impor modern samples available. These two hunt areas tantly,it is less variable than six of seven hunt areas produced very similar strontiumisotope ratios that for the carbon-oxygen-strontium distance metric are just slightly lower than themean Austin Wash and less variable thanReiser Canyon for thecarbon strontium isotope ratio.While the small sample oxygen distance metric. Quantitative assessment of sizes involvedmake itdifficult to reach firm con these relativedifferences furtherillustrates this pat clusions, the strontium isotope ratios suggest that tern (Table 7). In every case and by wide margins, themultiple populations included within Austin Firehole Basin is less variable than thevirtual site. Wash did not come fromnonlocal geographic areas. It is less variable thanmore than half of subsam Instead, it ismore likely that thepopulations were ples from two other hunt areas, and almost forty temporally,rather thangeographically, separate. percent of subsamples fromanother. Using carbon oxygen-strontiumdistances, Firehole Basin is less Firehole Basin #7 7 variable thanmore than half of Hunt Area 91 and As was the case forAustin Wash, comparative 92 subsamples. None of thesemeasures indicate

Table 6. Assessments of Evidence forMultiple Pronghorn Populations within Bonebeds.

Lubinski Current Current Radio- Overall Isotopic Overall Assessment Site_carbon PS PD Assessment Other Types of Assessments_Assessment Austin Wash ? ? Yes Yes Cohesiveness of diagnostic stone tools: No Yes Yes FireholeBasin ? No No No No Gailiun ? ? Yes Lack of fawns:Yes No No Eden-Farson ? No No Nimmo's (1971) age demography: Yes ? No (but see Lyman's (1987) critique) Boars Tusk ? ? ? Fisher and Frison (2000) seasonality: Inconclusive No No Trappers Point Yes ? ?_Vertical bone distribution: No_Yes_Yes PS: Pronghorn seasonality, PD: Pronghorn demography. Yes: Supports multiple populations, No: Does not support multiple populations, ?: Inconclusive, Blank: Not Available. PS and PD are Lubinski's (1997, 2000a,b) assessments, except for Trappers Point which is based on Miller et al. (1999) supplemented with thework described in Fenner and Walker (2008).

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 341

Figure 8. Comparison of carbon-oxygen scaled distance metric variation formodern hunt areas, the Reiser Canyon sam ple, a virtual site, and six archaeological sites. Minimum six samples per hunt area. multiple population sources. Finally, it is less vari 2008) carbon-oxygenmeasure, Gailiun and thevir able than nearly two-thirds of Reiser Canyon tual site show essentially equal variation.Gailiun's carbon-oxygen subsamples. The exception to the carbon-oxygen-strontium variation is greater than pattern is that Firehole Basin's carbon-oxygen the variation in less than half of subsamples from strontiumdistance variation is greater than thevari Hunt Areas 91 and 92. Gailiun's variation relative ation in88 percent ofReiser Canyon's subsamples. toReiser Canyon of about 91 and 69 percent for Taken together, the isotope variation data clearly the two isotope distances is, however, suggestive do not provide evidence of an accumulation of ofmultiple populations. pronghorn frommultiple populations at Firehole The Gailiun site isotopic variation data thus is Basin. somewhat equivocal but overall does not indicate Firehole Basin is located insideHunt Area 59, an accumulation. Lubinski (1997, 2000a) consid and itsmean strontium isotope value is indistin ers Gailiun to be the result of an accumulation guishable from thatof Hunt Area 59 (Table 4). The rather than a mass kill site (Table 6; see also the pronghorn thus are likely to have been local resi PreviousWork section).My assessment, however, dents, and there is no reason not to accept a single is thatoverall Gailiun does not show sufficientiso mass kill scenario forFirehole Basin. topic variation to support a designation as a multi ple population accumulation site. Lubinski's Gailiun analysis, which I argue uses methods that in gen Isotope variation at theGailiun site is generally not eral may not be reliable (see previous discussion), consistentwith amultiple population accumulation. and which in this case is driven almost entirelyby The comparative carbon-oxygen-strontiumisotope theabsence of fawns in the sample, is not adequate reverse variation (Figure 7) indicates thatGailiun has less to the isotopic assessment. variation than do most hunt areas, and only a bit It should be admitted thatthe difference between more variation than Reiser Canyon. Gailiun's my assessment and thatby Lubinski may simply carbon-oxygen variation (Figure 8) ismore impres be a resultof thedefault conditions of the analytic sive, being larger than thevariation at 11 of 14 hunt techniques in use. The isotopic method seeks to areas (but again is only slightly larger thanReiser demonstrate an accumulation frommultiple pop Canyon's variation).Quantitatively, Gailiun shows ulations, and if itcannot then itdefaults to not indi much less variation than thevirtual sitewhen using cating accumulation. Lubinski's method seeks to carbon-oxygen-strontiumdistance variation (Table match archaeological population age characteris 7). Using the less accurate (Fenner 2007:72-73, tics tomodern population characteristics, and if

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 342 AMERICAN ANTIQUITY [Vol. 74, No. 2, 2009]

Table 7. Percentage of Hunt Area Subsamples inWhich Each Archaeological Site has Greater Distance Variation than the Subsample.

Austin Firehole Eden Boars Trappers Virtual Location Wash Basin Gailiun Farson Tusk Point Site

Combined 813C and 8180 Hunt Area 42 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Hunt Area 47 100.0 100.0 100.0 100.0 100.0 100.0 100.0 Hunt Area 53 100.0 90.8 100.0 100.0 98.2 100.0 100.0 Hunt Area 57 99.9 40.5 71.6 73.1 54.5 100.0 75.6 Hunt Area 91 91.3 39.2 55.9 56.6 48.7 98.9 57.9 Hunt Area 92 100.0 58.3 86.5 87.5 72.0 100.0 88.7 Reiser Canyon 98.2 34.7 69.2 70.2 52.1 99.8 71.4

Combined 813C, 8180, and Sr Hunt Area 42 Hunt Area 47 Hunt Area 53 Hunt Area 57 Hunt Area 91 93.9 43.6 45.0 63.9 46.1 84.7 86.5 Hunt Area 92 99.4 41.0 46.5 94.8 57.9 99.4 99.4

Reiser Canyon_100.0 88.0 91.2 100.0 96.0 100.0 100.0 The virtual site is included for comparison. Combined 813C and 8180: Subsample size 7 selected from locations with at least 12 samples. Combined 813C, 8180 and Sr: Subsample size 7 selected from locations with at least 9 samples. Virtual site subsample size: 7.

low sample size causes a characteristicnot tomatch the hunter's shoulders. In thismanner the carcass then itdefaults to indicatingan accumulation. These was carried to camp, where itwas skinned and differences in technique are most apparent when butchered. Intactcarcass transportseems especially equivocal results are produced. likely if thehunters intended toprocess thehide or Gailiun's mean strontiumratio is quite unlike bones. Numerous ethnographic reports demon that of modern pronghorn from the hunt area in stratethe utility of pronghorn hide for clothing and which it is located (Table 4). This suggests that the bones formarrow and bone grease extraction kill may have occurred while thepronghorn were (McCabe et al. 2004:77-84). away from theirhome area,which was presumably in the lower strontiumratio area to thewest. Alter Eden-Farson natively, the kill may have occurred in the prong Eden-Farson's relative isotopic variability is diffi horn's home area, and the carcasses transportedto cult to interpret (Figures 7 and 8). Its carbon theGailiun site forprocessing. In the lattercase, it oxygen-strontium distance metric is greater than might seem surprising to find teeth and other low that of Reiser Canyon or any of the seven hunt utility remains atGailiun, but ethnographic hunters areas. It is,however, less than thatof thevirtual site. sometimes transportentire fairly lightweight ani Using carbon-oxygen distance, Eden-Farson is mals from the kill location to a camp or process more variable than Reiser Canyon (though just ing area (O'Connell andMarshall 1989; O'Connell barely) and 11 of 14 hunt areas. et al. 1990). An eviscerated but otherwise intact The quantitative assessment shows thatEden pronghorn carcass weighs 33 to 42 kg (O'Gara Farson is only slightly less variable than thevirtual 2004:243), which iswithin the ability of a human site (Table 7). Eden-Farson ismore variable than to carry for a fair distance. Northern Paiute infor themedian subsample variation from every hunt mants described toWillard Park themethod they area, but in some cases (particularlyHunt Area 91), used to carry a pronghorn or deer carcass (Fowler thedifference is not large. Similarly, it ismore vari 1989:20). The animal was bled and its intestines able than is Reiser Canyon, though not by a very removed. Skin was then stripped from the legs and large amount in carbon-oxygen distance. used totie the legs togetherand as a tumplineacross These results suggest thatEden-Farson prong

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 343

Table 8. Modern and Projected Relative Humidity atMass Kill Sites.

Modern HA Projected MeanHunt Growing Season Site Site_N_813C (%c)_Area ID_RH (%)_RH(%) 6 Boars Tusk -11.51 9244.0 51 Firehole Basin 7 -8.65 5945.5 35 8 Gailiun -9.19 9145.2 38

horn are the result of an accumulation frommulti tion.The Reiser Canyon exception is of some con ple populations. The isotopic data is, however, cern sinceReiser Canyon carbon-oxygen-strontium equivocal. It does not seem strongenough to fully distance is probably the best single isotopic indi overturn theNimmo (1971) and Lubinski (1997, cator.Nevertheless, theoverall isotopic assessment 2000b) analyses, which were based on large sam is thatthe pronghorn remains at theBoars Tusk site ple sizes and agreed thatEden-Farson was theresult are less likely tohave originated inmultiple prong of a singlemortality event.My overall assessment, horn populations than in a single population. therefore,is thatEden-Farson's pronghorn forma Boars Tusk pronghornmean strontiumvalues tionhistory is currently indeterminate. are substantially differentfrom those of modern Eden-Farson's mean strontiumisotope value is pronghorn in theirhome region of Hunt Area 92 significantlydifferent from thatof its home Hunt (Table 4). The Boars Tusk values are more com Area 92 (Table 4). Its value is, however, interme patible with pronghorn fromhunt areas to thenorth diate between themeans values of Hunt Area 92 west or southeast of Hunt Area 92. This indicates and nearby Hunt Area 96. This may indicate that that the pronghorn spent substantial time outside thepronghorn spent time inHunt Area 96, or itmay ofHunt Area 92 during their tooth formationperi be an artifactof the size of thehunt areas. That is, ods, which in turn suggests that theywere killed the regionwithin Hunt Area 92 along theborder of duringmigration from one area to another.As at Hunt Area 96 may in fact produce strontium iso Gailiun, human transportof pronghorn carcasses tope ratios intermediatebetween the two hunt area is also a possibility. mean values. Trappers Point Boars Tusk Trappers Point's relative carbon-oxygen-strontium Boars Tusk's relative isotopic variation also appears isotopic distance variation (Figure 7) is only slightly to be somewhat equivocal (Figures 7 and 8). It less than thatof the virtual site, and substantially shows less carbon-oxygen-strontiumvariation than exceeds the variation of Reiser Canyon and all thevirtual site and five of seven hunt areas, butmore modern hunt areas. Its carbon-oxygen variation is thanReiser Canyon. Its carbon-oxygen variation is the largestmeasured, substantially exceeding that less than thatof thevirtual site,Reiser Canyon, and of the virtual site,Reiser Canyon and all modern six of 14 hunt areas. The quantitative assessment hunt areas (Figure 8). The quantitative analysis sup shows thatBoars Tusk is clearly quite a bit less vari ports this overall assessment (Table 7). The Trap able than is the virtual site (Table 7). Its carbon pers Point carbon-oxygen-strontium variability is oxygen-strontiumdistance ismore variable than96 about the same as thatof thevirtual site and greatly percent of Reiser Canyon subsamples, but its exceeds the variation in modern locations. Its carbon-oxygen distance is close to themedian of carbon-oxygen variation substantiallyexceeds that Reiser Canyon subsamples. Boars Tusk's carbon inboth thevirtual site and allmodern locations.The oxygen-strontium distance is also close to the isotopic data stronglyindicates thatTrappers Point median ofHunt Area 91 and 92 subsamples. pronghornwere accumulated frommultiple popu Except for the carbon-oxygen-strontium dis lations. tance relative toReiser Canyon, the isotopic vari Trappers Point, however, presents a complica ability indicators thus suggest that Boars Tusk tion.As previously noted, it is located at a natural pronghorn were collected from a single popula bottleneck between theGreen River and theWind

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 344AMERICAN ANTIQUITY [Vol. 74, No. 2,2009]

RiverMountains (Eckerle 1999; Sanders andMiller pations atTrappers Point harvested pronghorn from 2004). Modern pronghornmove through the area multiple populations. That thesemultiple popula during springmigrations north into theGreater Yel tions formed during migration through the area lowstoneArea and returnin thefall. Crucially, these remains a possibility. migrations include pronghorn that originate in a wide swathof southwestern and Wyoming (Sawyer Conclusions Lindzey 2000) and thereforewould be expected to include substantial isotopic variability. If these Isotope ratio variation is a useful source of infor migrations also occurred during theEarly Archaic mation when assessing whether an archaeological period, then isotopic analysis of Trappers Point bonebed results from a singlemass kill or an accu pronghornmight not distinguish between multiple mulation of animals frommultiple populations. accumulations frommultiple populations and a sin Archaeologists using isotope data in conjunction gle accumulation frommultiple populations that with other archaeological informationwill often gathered togetherduring a migration (although the improve their understanding of the formation radiocarbon dates would not be similarlyaffected). processes of particular sites, and thereforeof cul Multiple modern pronghorn groups congregate in tural and regional patterns. the area just south of Trappers Point during spring With respect to pronghorn bonebeds in south migrations (Sawyer and Lindzey 2000:18), western Wyoming, it appears that bonebeds at although thismay be related tomodern highway Austin Wash and Trappers Point are the result of and housing developments in the area. The spring accumulation frommultiple populations and Fire timeframe for at least some of the pronghorn at hole Basin is likely tobe derived froma single pop Trappers Point (Miller 1999; see also Fenner and ulation. Gailiun and Boars Tusk are less certain,but Walker 2008) lends support to the possibility that seem unlikely tobe frommultiple populations. The its pronghorn were intercepted during migration pronghorn population characteristics at theEden (Miller et al. 1999b). Farson site are indeterminate. Strontium isotope ratios may provide insight It is interestingthat Austin Wash and Trappers into thisproblem. The mean Trappers Point archae Point are the sitesmost likely tohave been formed ological strontiumvalues are significantlydiffer through accumulation, because they are also the ent from those ofHunt Areas 87 and 88 (Table 4). two oldest sites in the sample (Table 1). Four expla Itmay be that some of theTrappers Point prong nations for this circumstance may be suggested. horn spentmuch of their time during theperiod of First, itmay simply be a coincidence related to enamel formation in a different location but were small sample size and thereforenot indicative of moving through theTrappers Point area as part of any archaeological phenomenon. The only way to a routine, seasonal migration. If so, then hunters resolve this is to increase the sample size by find may have intercepted them during movement ing and excavating more pronghorn bonebeds in through a particularly small migration corridor. the area. Second, diagenesis may have altered the This could be further investigated through stron archaeological site isotope ratios in theoldest sites. tium isotope analysis of intra-toothisotopic varia This does not, however, seem likely to be the case tion of archaeological samples. The possibility of (see theMethods section fordiscussion). multiple population accumulation through inter The differencebetween AustinWash and Trap ception ofmigrating pronghorn is furtherexplored pers Point bonebed formationprocesses and those in the following section. of later sitesmay reflectdifferences in the cultural The isotopic data in combination with theradio groups thatkilled thepronghorn and deposited their carbon data discussed previously present a per remains. Specifically, Lubinski (1997,2000a) ten suasive case thatTrappers Point pronghorn remains tatively proposed thatmass pronghorn kills are are theresult of at least twokill events substantially indicative of the expansion of Numic cultural separate in time. It is difficultto reconcile thiswith groups from the Great Basin into southwestern thevertical distributiondata, unless sediment accu Wyoming between 500 and 700 B.P. The results of mulation in the area was very slow or intermittent. the current analysis are compatible with this Overall, it seems quite likely thatmultiple occu hypothesis: the two sites identified as accumula

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 345

tionsdate towell before 700 B .P,and the sites iden and thousands of pieces of debitage in a very spe tifiedas single pronghorn population sites consis cific location. Both also had associated ash and tentwith mass kills occur during or after theperiod. fire-cracked rock suggestive of hearths or roasting While much more analysis of a variety of cultur pits.While not forgettingthe possibility of tempo ally diagnostic archaeological materials is needed raryprey population aggregation duringmigration, to confirm thehypothesis, it is an intriguingpossi the presence of multiple prey populations within bility. Itmay be informative to perform similar each bonebed suggests thatpeople visited the area on analyses Great Basin pronghorn bonebeds to more than once over a fairly shortperiod during determine whether the pattern holds in other time which thebonebed was exposed, and each timedis periods and regions. carded theirrefuse in exactly the same location.The Finally, perhaps Trappers Point andAustin Wash presence of a number ofwhole, intact lithicsalong show evidence of accumulation because they with chipped stone debitage throughout each involved interceptionof migrating pronghorn from bonebed indicates that the bones were also multiple populations while theothers resulted from processed in some manner (presumably involving a drive, surround,or trap of local pronghorn pop defleshing) at the exact same locations. ulations. Trappers Point, at least, clearly lieswithin Along these lines, it is interestingthat Trappers a modern pronghornmigration corridor (Sawyer Point and Austin Wash are both multicomponent and Lindzey 2000). It is possible that the compar sites.At Trappers Point, StratumVII overlies and a ison of site's mean strontium isotope values to is stratigraphically separated from StratumV (the mean values in pronghorn from thehunt area sur focus of interestelsewhere in thisreport), which in rounding the sitewould distinguish between sites turnoverlies and is stratigraphicallyseparated from with local pronghorn and those with nonlocal, StratumHI (Francis and Sanders 1999).While Stra migrating pronghorn. The Trappers Point prong tumV has by far the largestbonebed, all threestrata mean horn strontium isotope ratio is significantly include substantial bone deposits and are domi differentfrom themean ofmodern local pronghorn nated by pronghorn remains.Austin Wash has been two (Table 4). However, other sites identifiedby less thoroughly investigated thanTrappers Point, isotope analysis as containing single-population and only themain component iswell understood. means are bonebeds have strontiumisotope that sta However, radiocarbon dates obtained fromnearby tistically different from their surrounding hunt features suggest that there are at least two compo so areas, nonlocal strontium signatures occur in nents spatially slightlyoutside themain bonebed both single- and multiple-population sites. (Schroedl 1985). Having distinguished between multiple- and Thus, bone was processed and discarded at both single-population archaeological bonebeds, it is Trappers Point andAustin Wash multiple times,and now to appropriate consider what anthropological spanning two temporal scales. Deposition of bones issues may best be investigatedusing each typeof frommultiple independent prey populations sug bonebed separately.Out of a range of possibilities, gestmultiple events at an intermediate scale, dur I will on focus two: the value of multiple ingwhich the bonebed formed during a previous sites reuse a accumulation in foregrounding the of deposition eventwas stillvisible on the landscape, on specific location the landscape, and the recon while the presence ofmultiple components sepa structionof specific environmentalconditions using rated by hundreds, perhaps thousands, of years single-population bonebeds. indicatemultiple events at the same exact location Both of the sites identified in this as analysis despite theprobable invisibilityof previous depo remains from containing multiple pronghorn pop sition events.What was it thatdrew people topref ulations are most (like archaeological sites) discard erentially process and discard pronghorn kills at locations.While one may suppose thatTrappers those exact locations rather than elsewhere on the were Point and Austin Wash associated in some landscape? The association of location and discard manner with kill or nearby sites camps, ultimately behavior may be reflective of cultural history or the evidence to only recovered date relates to the may be associated with specific economic strate discard of a number of a large pronghorn bones, gies (Holdaway andWandsnider 2006). Trappers few bones fromother animals, hundreds of lithics, Point and Austin Wash do not seem to share spe

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 346AMERICAN ANTIQUITY [Vol. 74, No. 2,2009]

cific environmental features thatwould lead peo process differencesmust await a larger sample of ple touse those particular locations. Trappers Point bonebeds, which may involve expanding focus is located near the crest of a low rise (one ofmany beyond southwesternWyoming into surrounding lowrises in the area) while AustinWash is on a gen regions. The currentanalysis has shown, however, tle slope near thebase of a steep cliff.Trappers Point that isotope analysis, in combination with other is a few kilometers from theGreen River (the only methods, can in some circumstances resolve the largeriver in the region) but is not directly associ prey population formationhistory of large archae ated with a modern water source,while theAustin ological pronghornbonebeds. It furthermoreseems Wash site is adjacent to the intermittentstream of likely that themethod can be adapted to use with the same name and an unnamed ephemeral bonebeds dominated by the remains of other drainage. The modern vegetation in both areas is species, such as bison. dominated by sagebrush, but thatis trueof themost Acknowledgments. I appreciate helpful discussions and com of thebasin area of southwesternWyoming. It thus ments on an earlier version of this article by Marcel is that features of the unlikely physical landscape Kornfeld, Todd Surovell, Carol Frost, David Williams, Mark come to caused people to repeatedly these partic Miller, and James Ahern. Anonymous reviewers also pro ular locations forpronghorn processing. Instead, it vided helpful comments on an earlier draft. The author would like to thank Martha Joni Stainbrook, the seems likely that long-lasting cultural or historical Rogers, University of Wyoming Archaeological Repository and features of the landscape made these attractivepro Curation Facility and Western Archaeological Services for sites for who hunted cessing people successfully providing archaeological samples. Thank you also to Shay pronghorn. Westbrook and theWyoming Game and Fish Department for Turning now to the three sites identified in this providing the modern hunt area pronghorn teeth. Patrick Lubinski, Walker, and the of analysis as resulting frommass kills of single Danny University Wyoming Comparative Osteology Museum and Zooarchaeological pronghorn populations, we can consider theuse of Laboratory kindly provided the Reiser Canyon samples. bonebeds in envi pronghorn reconstructing past Mike Meredith and Shikha Sharma provided help and good ronmental conditions. Gailiun, Firehole Basin, and advice regarding isotope analysis. My thanks to Ranel Boars Tusk each contain the remains of a statisti Capron of theU.S. Bureau of Land Management for permis sion to enamel from recovered from BLM cally meaningful sample of pronghorn from a sin sample pronghorn sites. Kirk Nordyke ofWyoming Game and Fish Department gle population and thereforea single snapshot of provided a GIS shapefile of the pronghorn hunt areas. The time. of modern Analysis pronghorn populations Spanish translation of the abstract was provided by ASTA mean are (Table 2) indicates that 813C values USA. Funding for this project was provided by the stronglycorrelated with growing season humidity University of Wyoming Department of Geology and National Science Foundation Doctoral (r2= .69,p < .001;Fenner and Frost 2009). A regres Geophysics, Dissertation Grant and a of sion of thiscorrelation can be used with mean 813C Improvement #44310, University Wyoming Summer Doctoral Augmentation Award. values from archaeological single-population bonebeds to obtain a rough idea of humidity con References Cited ditions at fairly specific times and places in thepast is (Table 8). The Boars Tusk humidity projection Arkush, Brooke S. substantiallygreater than itsarea's modern 30-year 1986 Aboriginal Exploitation of Pronghorn in the Great Basin. Journal of Ethnobiology 6:239-255. average growing season humidity, but is in line Bartlein, Patrick J.,Mary E. Edwards, Sarah L. Shafer, and areas to northeast with humidity values from the Edward D. Barker, Jr. such as Hunt Areas 64 and 75.More interestingare 1995 Calibration of Radiocarbon Ages and the Interpreta tion of Paleoenvironmental Records. Research the forGailiun and Firehole Quartenary humidity projections 44:417-424. Basin, which are well below themodern 30-year BCal average growing season humidity values for any 2006 BCal: An on-line Bayesian radiocarbon calibration tool. Calibrated using the 2004 atmospheric calibration measured hunt area inWyoming. Itmay be thatvery curve, http://bcal.shef.ac.uk. arid conditions increased themotivation or the Bentley, R. Alexander, T. Douglas Price, Jens Luning, Detlef D. opportunity forpeople to communally huntprong Gronenborn, JoachimWahl, and Paul Fullagar 2002 Prehistoric Migration in Europe: Strontium Isotope horn in southwesternWyoming. Analysis of Early Neolithic Skeletons. Current Anthro Full resolution of the of south implications pology 43:799-804. M. westernWyoming pronghorn bonebed formation Buntinas, Martin, and Gerald Funk

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 347

2005 Statistics for the Sciences. Brooks/Cole, Belmont, Ehleringer, James R., and Paul W. Rundel California. 1989 Stable Isotopes: History, Units, and Instrumentation. Byers, JohnA. In Stable Isotopes inEcological Research, edited by Paul 1997 American Pronghorn: Social Adaptations and the W. Rundel, James R. Ehleringer and Kenneth A. Nagy, pp. Ghosts of Predators Past. University of Chicago Press, 1-15. Springer-Verlag, New York. Chicago. Ericson, Jonathon E. 2003 Built for Speed: A Year in theLife of Pronghorn. Har 1985 Strontium Isotope Characterization in the Study of vard University Press, Cambridge. Prehistoric Human Ecology. Journal ofHuman Evolution Caughley, Graeme 14:503-514. 1974 Interpretation ofAge Ratios. Journal ofWildlifeMan Fawcett, William Bloys, Jr. agement 38:557-562. 1987 Communal Hunts, Human Aggregations, Social Vari Chamberlin, Ralph V. ation, and Climatic Change: Bison Utilization By Prehis 1911 The Ethno-Botany of the Gosiute Indians of Utah. toric Inhabitants of theGreat Plains. Ph.D. Dissertation, Memoirs of theAmerican Anthropological Association 2 Anthropology, University ofMassachusetts, Amherst. Pro Part 5:331^05. Quest Dissertation Number AAT 8710451. Chittenden, Hiram Martin Fenner, Jack N. 1905 Life, Letters and Travels of Father Pierre-Jean De 2007 Prehistoric Hunting On theRange Where theAnte Smet, S. J: 1801-1873, Volume 4. Francis P. Harper, New lope Play: Archaeological Pronghorn Bonebed Forma York. tion Analysis. Ph.D. Dissertation, Department of Clark, Ian D., and Peter Fritz Anthropology, University ofWyoming, Laramie. ProQuest 1997 Environmental Isotopes inHydrogeology. Lewis Pub Dissertation Number AAT 3259784. lishers, Boca Raton. 2008 The Use Of Stable Isotope Ratio Analysis To Distin Clayton, Carmen, and Raymond Kunselman guish Multiple Prey Events From Mass Kill Events. Jour 2002 Obsidian Source Utilization at the Trappers Point nal ofArchaeological Science 35:704-716. Antelope Kill Site 48SU1006. Wyoming Archaeologist Fenner, Jack N. and Carol D. Frost 46(2):83-91. 2009 Modern Wyoming Plant and Pronghorn Isoscapes and Current, Bill Their Implications for Archaeology. Journal of Geo 1993 Archaeological Data Recovery Plan for theGailiun chemical Exploration, in press. Site (48SUU56). Sublette County, Wyoming. Fenner, Jack N., and Danny N. Walker Davis, Leslie B., JohnW. Fisher, Jr.,Michael C. Wilson, Stephen 2008 Mortality Date Estimation Using Fetal Pronghorn A. Chomko, and Richard E. Morlan Remains. International Journal of Osteoarchaeology 2000 Avonlea Phase Winter Fare at Lost Terrace, Upper 18:45-60. Missouri River Valley ofMontana: The Vertebrae Fauna. Fisher, JohnW, Jr.,and George C. Frison In Pronghorn Past and Present: Archaeology, Ethnogra 2000 Site Structure and Zooarchaeology at theBoar's Tusk phy, and Biology. Plains Anthropologist Memoir 32., edited Site,Wyoming. In Pronghorn Past and Present: Archae by Jana V. Pastor and Patrick M. Lubinski, pp. 53-69. ology, Ethnography, and Biology. Plains Anthropologist Plains Anthropological Society, Lincoln, Nebraska. Memoir 32., edited by JanaV. Pastor and PatrickM. Lubin De Smet, Pierre J. ski, pp. 89-108. Plains Anthropological Society, Lincoln, 1859 Western Missions and Missionaries: A Series of Let Nebraska. ters.New York, P. J.Kennedy. Fowler, Catherine S. Dewar, Genevieve, Dave Halkett, Tim Hart, Jayson Orton, and 1989 Willard Z. Park's Ethnographic Notes on theNorth ern Judith Sealy Paiute ofWestern Nevada, 1933-1944. University of a 2006 Implications of Mass Kill Site of Springbok (Anti Utah Anthropological Papers Number 114 1. University dorcas Marsupialis) in South Africa: Hunting Practices, of Utah Press, Salt Lake City. Gender Relations, and Sharing in the Later Stone Age. Francis, Julie E., and Paul H. Sanders Journal ofArchaeological Science 33:1266-1275. 1999 Excavation Summary. In The Trappers Point Site Eckerle, William P. (48SU1006): Early Archaic Adaptations in the Upper 1999 Paleoenvironmental Reconstruction and Prehistoric Green River Basin, Wyoming, edited by Mark E. Miller, Lifeways. In The Trappers Point Site (48SU1006): Early Paul H. Sanders and Julie E. Francis, pp. 38-57. Cultural Archaic Adaptations in the Upper Green River Basin, Resource Series No. 1. vol. 1.Office of theWyoming State Wyoming, edited byMark E. Miller, Paul H. Sanders, and Archaeologist, Wyoming Department of State Parks and Julie E. Francis, pp. 452-475. Cultural Resource Series Cultural Resources, Laramie, Wyoming. No. 1. vol. 1.Office of theWyoming State Archaeologist, Frison, George Wyoming Department of State Parks and Cultural 2000 Observations on Pronghorn Behavior and Taphonomic Resources, Laramie, Wyoming. Analysis of Bonebeds: Implications forAnalysis of the Eckles, David G. Eden-Farson Pronghorn Kill. In Pronghorn Past and Pre 1999 Taxonomic Identification, Bone Element Frequen sent: Archaeology, Ethnography, and Biology. Plains cies, and Bone Artifacts. In The Trappers Point Site Anthropologist Memoir 32, edited by Jana V. Pastor and (48SU1006): Early Archaic Adaptations in the Upper Patrick M. Lubinski, pp. 29-37. Plains Anthropological Green River Basin, Wyoming, edited by Mark E. Miller, Society, Lincoln, Nebraska. Paul H. Sanders, and Julie E. Francis, pp. 193-205. Cul Frison, George C. turalResource Series No. 1. vol. 1. Office of theWyoming 1971 Shoshonean Antelope Procurement in the Upper State Archaeologist, Wyoming Department of State Parks Green River Basin, Wyoming. Plains Anthropologist and Cultural Resources, Laramie, Wyoming. 16:258-284. Egan, Howard R. 1978 Animal Population Studies and Cultural Inference. 1917 Pioneering theWest 1846 to 1878. Howard R. Egan Plains Anthropologist 23:44-52. Estate, Richmond, Utah. 1991 Prehistoric Hunters of theHigh Plains. 2nd ed. Aca

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 348AMERICAN ANTIQUITY [Vol. 74, No. 2,2009]

demic Press, San Diego. Kelly, Isabel T. Frison, George C, and Charles A. Reher 1932 [ 1986] Ethnography of the Surprise Valley Paiute. In 1970 Age Determination of Buffalo by Teeth Eruption and A Great Basin Shoshonean Source Book, edited by David Wear. Plains Anthropologist Memoir 7:46-50. Hurst Thomas. Garland Publishing, New York. Garvie-Lok, Sandra, Tamara L. Varney, and M. Anne Katzen Kindig, JeanMatthews berg 1987 An Evaluation of an Ethnohistoric Account of a Plains 2004 Preparation of Bone Carbonate for Stable Isotope Indian Communal Hunt in theBoulder Valley, 1862. South Analysis: The Effects of Treatment Time and Acid Con western Lore 53:17-27. centration. Journal ofArchaeological Science 31:763-776. Klein, Richard G. Grinnell, George Bird 1982 Age (Mortality) Profiles as a Means of Distinguish 1923 [1972] The Cheyenne Indians: Their History and Ways ing Hunted Species from Scavenged Ones in Stone Age of Life. University of Nebraska Press, Lincoln. Archeological Sites. Paleobiology 8:151-158. Hedges, R. E. M. Klein, Richard G., and Kathryn Cruz-Uribe 2002 Bone Diagenesis: An Overview of Processes. 1984 The Analysis of Animal Bones from Archeological Archaeometry 44:319-328. Sites. University of Chicago Press, Chicago. Higgins, Pennilyn, and Bruce J.MacFadden Knight, Dennis H. 2004 "Amount Effect" Recorded in Oxygen Isotopes of 1994 Mountains and Plains: The Ecology of Wyoming Late Glacial Horse (Equus) and Bison (Bison) Teeth from Landscapes. Yale University, New Haven. the Sonoran and Chihuahuan Deserts, Southwestern Knudson, K. J.,T. D. Price, J.E. Buikstra, and D. E. Blom United States. Palaeogeography, Palaeoclimatology, 2004 The Use of Strontium Isotope Analysis to Investigate Palaeoecology 206:337-353. Tiwanaku Migration and Mortuary Ritual in Bolivia and Hill, Matthew E., Jr. Peru. Archaeometry 46:5-18. 2002 The Milnesand Site: Site Formation Study of a Pale Koch, Paul L. oindian Bison Bonebed in Eastern New Mexico. Plains 1998 Isotopic Reconstruction of Past Continental Environ Anthropologist 47:323-337. ments. Annual Review of Earth and Planetary Sciences Hill, Matthew Glenn 26:573-613. 2001 Paleoindian Diet and Subsistence Behavior on the Koch, Paul L., Kathryn A. Hoppe, and S. David Webb Northwestern Great Plains ofNorth America. Ph.D. Dis 1998 The Isotopic Ecology of late Pleistocene mammals in sertation,Department ofAnthropology, University ofWis North America: Part 1. Florida. Chemical Geology consin, Madison.. ProQuest Dissertation Number AAT 152:119-138. 3012451. Koch, Paul L., Noreen Tuross, and Marilyn L. Fogel Hillson, Simon 1997 The Effects of Sample Treatment and Diagenesis on 1986 Teeth. Cambridge University Press, Cambridge. the Isotopic Integrity of Carbonate inBiogenic Hydroxy Hodell, David A., Rhonda L. Quinn, Mark Brenner, and George lapatite. Journal ofArchaeological Science 24:417^129. Kamenov Kohlmann, Stephan G. 2004 Spatial Variation of Strontium Isotopes (87Sr/86Sr) 2004 Population Dynamics and Modeling. In Pronghorn in theMaya Region: A Tool forTracking Ancient Human Ecology and Management, edited by Bart W. O'Gara and Migration. Journal ofArchaeological Science 31:585-601. Jim D. Yoakum, pp. 644-674. University Press of Col Holdaway, Simon J., and LuAnn Wandsnider orado, Boulder, Colorado. 2006 Temporal Scales and Archaeological Landscapes from Kohn, Matthew J., and Thure E. Cerling the Eastern Desert of Australia and Intermontane North 2002 Stable Isotope Compositions of Biological Apatite. In America. In Confronting Scale inArchaeology: Issues of Phosphates. Geochemical, Geobiological, and Materials Theory and Practice, edited by Gary. Lock and Brian L. Importance. Reviews inMineralogy and Geochemistry, Molyneaux, pp. 183-202. Springer, New York. edited by Matthew J.Kohn, John Rakovan, and JohnM. Hoppe, Kathryn A. Hughes, pp. 455^-88. vol. 48. Mineralogical Society of 2004 Late Pleistocene Mammoth Herd Structure, Migra America, Washington D. C. tion Patterns, and Clovis Hunting Strategies Inferred from Larson, Robert, Joe Bozovich, and Jack Krmpotich Isotopic Analyses ofMultiple Death Assemblages. Pale 1969 Preliminary Report on Farson Site #48 SW 304. obiology 30:129-145. Wyoming Archaeologist 12(4): 13-16. Hoppe, K. A., P. L. Koch, and T. T. Furutani Lee-Thorp, Julia, and Matt Sponheimer 2003 Assessing the Preservation of Biogenic Strontium in 2003 Three Case Studies Used toReassess theReliability Fossil Bones and Tooth Enamel. International Journal of of Fossil Bone and Enamel Isotope Signals for Paleodi Osteoarchaeology 13:20-28. etary Analysis. Journal of Anthropological Archaeology Hoskinson, Reed L., and JohnR. Tester 22:208-216. 1980 Migration Behavior of Pronghorn in Southeastern 2006 Contributions of Biochemistry to Understanding Idaho. Journal ofWildlife Management 44:132-144. Hominin Dietary Ecology. Yearbook of Physical Anthro Hyde, George S. pology 49:131-148. 1968 Life of George Bent: Written From His Letters. Uni Leopold, Luna B., and Claudio Vita-Finzi versity of Oklahoma Press, Norman. 2005 Archaeological Trash: Geomorphology and Early Irving,Washington Human Occupation inWyoming. Catena 62:1-13. 1986 The Adventures of Captain Bonneville U.S.A. in the Levin, Naomi E., Thomas E. Cerling, Benjamin H. Passey, John Rocky Mountains and theFar West. University of Okla M. Harris, and James R. Ehleringer homa Press, Norman. 2006 A Stable Isotope Aridity Index for Terrestrial Envi Karhu, Rory R., and Stanley H. Anderson ronments. Proceedings of theNational Academy of Sci 2006 The Effect ofHigh-Tensile Electric Fence Designs on ence 103:11201-11205. Robert H. Big-Game and Livestock Movements. Wildlife Society Lowie, Bulletin 34:293-299. 1909 [1986] Anthropological Papers of the American

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions REPORTS 349

Museum ofNatural History: The Northern Shoshone. Vol 1999a The Trappers Point Site (48SU1006): Early Archaic ume 2 Part 2. InA Great Basin Shoshonean Source Book, Adaptations in the Upper Green River Basin, Wyoming. edited by David Hurst Thomas. Garland Publishing, New Cultural Resource Series No. 1. Office of theWyoming York. State Archaeologist, Laramie, Wyoming. 1924 [1986] Notes on Shoshonean Ethnography. Anthro 1999b Intermountain Archaeology and theTrappers Point pological Papers of theAmerican Museum ofNatural His Site. In The Trappers Point Site (48SU1006): Early Archaic tory,Volume 20, Part 3. In A Great Basin Shoshonean Adaptations in the Upper Green River Basin, Wyoming, Source Book, edited by David Hurst Thomas. Garland Pub edited by Mark E. Miller, Paul H. Sanders, and Julie E. lishing, New York. Francis, pp. 476-489. Cultural Resource Series No. 1. vol. Lubinski, Patrick M. 2. Office of theWyoming State Archaeologist, Wyoming 1997 Pronghorn Intensification in theWyoming Basin: A Department of State Parks and Cultural Resources, Study ofMortality Patterns and Prehistoric Hunting Strate Laramie, Wyoming. gies. Ph.D. Dissertation, Department of Anthropology, Moore, David S., and George P. McCabe University ofWisconsin, Madison. 1999 Introduction to thePractice of Statistics. W. H. Free 1999 The Communal Pronghorn Hunt: A Review of the man, New York. Ethnographic and Archaeological Evidence. Journal of Nimmo, Barry W. a California and Great Basin Anthropology 21:158-181. 1971 Population Dynamics of Wyoming Pronghorn 2000a Prehistoric Pronghorn Hunting in Southwest Cohort from the Eden-Farson Site, 48SW304. Plains Wyoming. In Pronghorn Past and Present: Archaeology, Anthropologist 16:285-288. Ethnography, andBiology. Plains Anthropologist Memoir O'Connell, James F., K Hawkes, and N. Blurton Jones 32, edited by Jana V. Pastor and Patrick M. Lubinski, pp. 1990 Reanalysis of Large Mammal Body Part Transport 109-118. Plains Anthropological Society, Lincoln, Among the Hadza. Journal of Archaeological Science Nebraska. 17:301-316. 2000b A Comparison ofMethods forEvaluating Ungulate O'Connell, James F, and Brendan Marshall Mortality Distributions. Archaeozoologica 11:121-134. 1989 Analysis ofKangaroo Body Part Transport among the 2001 Estimating Age and Season of Death of Pronghorn Alyawara of Central Australia. Journal ofArchaeological Antelope (Antilocapra americana Ord) byMeans of Tooth Science 16:393-405. Eruption and Wear. International Journal of Osteoar O'Gara, BartW. chaeology 11:218-230. 2004 Physiology and Genetics. InPronghorn Ecology and Lubinski, Patrick M., and Christopher J.O'Brien Management, edited by Bart W. O'Gara and Jim D. 2001 Observations of Seasonality and Mortality from a Yoakum, pp. 231-273. University Press of Colorado, Boul Recent Catastrophic Death Assemblage. Journal of der, Colorado. Archaeological Science 28:833-842. O'Gara, BartW, and JimD. Yoakum Lyman, R. Lee 2004 Pronghorn Ecology and Management. University 1987 On theAnalysis ofVertebrae Mortality Profiles: Sam Press of Colorado, Boulder. ple Size, Mortality Type, and Hunting Pressure. American Ottman, Jim Antiquity 52:125-142. 1992 The Plunge: A Herd Falls to Its Death. Wyoming Lytle, John C. Wildlife 52:10-13. 1969 Farson Site 48 SW 304. Wyoming Archaeologist Price, T. D., J.H. Burton, and R. A. Bentley 12(4): 17-19. 2002 The Characterization of Biologically Available Stron McCabe, Richard E., BartW O'Gara, and Henry M. Reeves tium Isotope Ratios for the Study of Prehistoric Migration. 2004 Prairie Ghost: Pronghorn and Human Interaction in Archaeometry 44:117-135. Early America. University Press of Colorado, Boulder, Price, T. Douglas, Clark M. Johnson, Joseph A. Ezzo, Jonathon Colorado. Ericson, and James H. Burton Middleton, Jessica L., Patrick M. Lubinski, and Michael D. 1994 Residential Mobility in the Prehistoric Southwest Metcalf United States: A Preliminary Study using Strontium Iso 2007 Ceramics from the Firehole Basin Site and Firehole tope Analysis. Journal of Archaeological Science Phase in theWyoming Basin. Plains Anthropologist 21:315-330. 52:29^11. Reiss, David, and Danny N. Walker Miller, Mark E. 1982 Pronghorn Utilization at 48UT390 in Southwestern 1999 Seasonality of theEarly Archaic Trappers Point Prong Wyoming. Wyoming Contributions to Anthropology horn Bonebed. In The Trappers Point Site (48SU1006): 3:1-25. Early Archaic Adaptations in theUpper Green River Basin, Reynolds, Timothy D. Wyoming, edited byMark E. Miller, Paul H. Sanders and 1984 Daily Summer Movements, Activity Patterns, and Julie E. Francis, pp. 206-241. Cultural Resource Series Home Range of Pronghorn. Northwest Science No. 1. vol. 1.Office of theWyoming State Archaeologist, 58:300-311. Wyoming Department of State Parks and Cultural Sanders, Paul H. Resources, Laramie, Wyoming. 1999 Investigation of Bone Weathering Stages, Age Com Miller, Mark E. and Paul H. Sanders position, and Spatial Distributions of Lithic and Faunal 2000 The TrappersPoint Site (48SU1006):Early Archaic Remains. In The Trappers Point Site (48SU1006): Early Adaptations and Pronghorn Procurement in the Upper Archaic Adaptations in the Upper Green River Basin, Green River Basin, Wyoming. InPronghorn Past and Pre Wyoming, edited byMark E. Miller, Paul H. Sanders, and sent: Archaeology, Ethnography, and Biology. Plains Julie E. Francis, pp. 364-451. Cultural Resource Series Anthropologist Memoir 32., edited by Jana V. Pastor and No. 1. vol. 2. Office of theWyoming State Archaeologist, Patrick M. Lubinski, pp. 39-52. Plains Anthropological Wyoming Department of State Parks and Cultural Society, Lincoln, Nebraska. Resources, Laramie, Wyoming. Miller, Mark E., Paul H. Sanders, and Julie E. Francis Sanders, Paul H., and Mark E. Miller

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions 350AMERICAN ANTIQUITY [Vol. 74, No. 2,2009]

2004 Archaeological Evidence of Pronghorn (Antilocapra and Biology. Plains Anthropologist Memoir 32., edited by americand) Migration in theUpper Green River Basin of JanaV. Pastor and PatrickM. Lubinski, pp. 119-132. Plains Wyoming: Implications forWildlife Management. In Anthropological Society, Lincoln, Nebraska. Zooarchaeology and Conservation Biology, edited by R. Tieszen, Larry L., and Tim Fagre Lee Lyman and Kenneth P. Cannon, pp. 164-177. Uni 1993 Carbon Isotope Variability inModern and Archaeo versity of Utah Press, Salt Lake City. logical Maize. Journal of Archaeological Science Sawyer, Hall, and Fred Lindzey 20:25-40. 2000 Jackson Hole Pronghorn Study.Wyoming Coopera Todd, Lawrence C. tiveFish andWildlife Research Unit, Laramie, Wyoming. 1987 Analysis of Kill-Butchery Bonebeds and Interpreta Schoolcraft, Henry R. tion of Paleoindian Hunting. In The Evolution ofHuman 1860 History, Antiquities, Language, Ethnology, Pictogra Hunting, edited byM. Nitecki and D. Nitecki, pp. 225-266. phy, Rites, Superstitions, and Mythology of the Indian Plenum Press, New York. Tribes of the United States, Volume IV J.B. Lippincott, Todd, Lawrence C, Jack L. Hofman, and C. Bertrand Schultz Philadelphia. 1992 Faunal Analysis and Paleoindian Studies: A Reex Schroedl, Alan R. amination of theLipscomb Bison Bonebed. Plains Anthro 1985 Archaic and Late Prehistoric Adaptation in South pologist 37:137-165. western Wyoming: The Frontier Pipeline Excavations. Todd, Lawrence C, David C. Jones, Robert S.Walker, Paul C. Bureau of Land Management Wyoming Cultural Resource Burnett, and JeffreyEighmy Series Number 3. 2001 Late Archaic Bison Hunters in Northern Colorado: SHPO 1997-1999 Excavations at the Kaplan-Hoover Bison 1979 48SWI373 (Boars Tusk) Site Report. Copy on file at Bonebed (5LR3953). Plains Anthropologist 46:125-147. Wyoming State Historic Preservation Office Cultural Ugan, Andrew, Jason Bright, and Alan Rogers Records Office, Laramie, Wyoming. 2003 When Is Technology Worth theTrouble? Journal of 1981-1999 48UT390 (AustinWash) Site Reports. CRM site Archaeological Science 30:1315-1329. reports provided by various organizations. Copy on file at Whittaker, William E., and James G. Enloe Wyoming State Historic Preservation Office Cultural 2000 Bison Dentition Studies Revisited: Resolving Ambiq Records Office, Laramie, Wyoming. uity between Archaeological and Modern Control Sam Steele, Teresa E. ples. Archaeozoologica 11:113-120. 2003 Using Mortality Profiles to InferBehavior in theFos Widga, Chris sil Record. Journal ofMammalogy 84:418^430. 2004 Early Archaic Subsistence in theCentral Plains: The Steward, Julian H. Spring Creek (25FT31) Fauna. Plains Anthropologist 1933 Ethnography of theOwens Valley Paiute. University 49:25-58. of California Publications inAmerican Anthropology and Wilson, S. R., and G. K. Ward Ethnology 33(3):233-350. 1981 Evaluation and Clustering ofRadiocarbon Age Deter 1938 Basin-Plateau Aboriginal Sociopolitical Groups. minations: Procedures and Paradigms. Archaeometry Smithsonian Institution Bureau ofmerican Ethnology Bul 23:19-39. letin 120. Smithsonian Institution,Washington DC. Yoakum, JimD. Stiner,Mary C. 2004a Foraging Ecology, Diet Studies and Nutrient Values. 1990 The Use ofMortality Patterns inArchaeological Stud In Pronghorn Ecology and Management, edited by Bart ies ofHominid Predatory Adaptations. Journal ofAnthro W. O'Gara and JimD. Yoakum, pp. 447-502. University pological Archaeology 9:305-351. Press of Colorado, Boulder, Colorado. 1991 a An Interspecific Perspective on theEmergence of the 2004b Habitat Conservation. In Pronghorn Ecology and Modern Human Predatory Niche. In Human Predators Management, edited by Bart W. O'Gara and Jim D. and PreyMortality, edited byMary C. Stiner, pp. 149-185. Yoakum, pp. 571-630. University Press of Colorado, Boul Westview Press, Boulder. der, Colorado. 1991b Introduction: Actualistic and Archaeological Stud ies of Prey Mortality. InHumanPredators and Prey Mor tality,edited byMary C. Stiner, pp. 1-13. Westview Press, Boulder. Sundstrom, Linea Received Revised 2000 Cheyenne Pronghorn Procurement and Ceremony. In July 6, 2007; April 12, 2008; Accepted Pronghorn Past and Present: Archaeology, Ethnography, August 21, 2008.

This content downloaded from 130.56.5.15 on Tue, 10 Feb 2015 00:46:38 AM All use subject to JSTOR Terms and Conditions