A Further Assessment of Paleoindian Site-Use at Bonfire Shelter Author(S): Ryan M

Society for American Archaeology

A Further Assessment of Paleoindian Site-Use at Bonfire Shelter Author(s): Ryan M. Byerly, Judith R. Cooper, David J. Meltzer, Matthew E. Hill, Jason M. LaBelle Reviewed work(s): Source: American Antiquity, Vol. 72, No. 2 (Apr., 2007), pp. 373-381 Published by: Society for American Archaeology Stable URL: http://www.jstor.org/stable/40035821 . Accessed: 13/01/2012 10:15

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 A FURTHER ASSESSMENT OF PALEOINDIAN SITE-USE AT BONFIRE SHELTER

Ryan M. Byerly, JudithR. Cooper,David J. Meltzer,Matthew E. Hill, and Jason M. LaBelle

In Byerly et al. (2005) we explored the hypothesis that the Paleoindian component at Bonfire Shelter was the result of a jump kill. Our efforts involved extensive mapping and GIS analysis, a re-examination of the Paleoindian-age bison assemblage, and consideration of the geomorphic history of the canyon in which the site is located. Weconcluded that the pre- ponderance of evidence indicated the Paleoindian-age bison remainsat Bonfire Shelter markeda processing site as Binford (1978) suggested, rather than a primary kill locality as originally interpreted(Dibble 1968). Bement (this issue) raises sev- eral concerns about our analysis and discussion, including that we omit pertinent informationrelevant to the interpretation of the site. His comments,however, resultfrom a misreading of our discussion and a misconstrual of the data set, as we explain in this response.

En Byerly et al. (2005) exploramosla hipotesis que el componentepaleoindio de BonfireShelter fuese el resultadode una matanzade salto. Nuestros esfuerzos incluyeronla preparacionde extensivosmapas y un andlisis SGI (GIS), una reexami- nacion de la asamblea bisontede fecha paloeoindia y la evaluacionde la historiageomorfologica del canon en que se encuen- tra el sitio. Concluimosque la mayorparte de la evidenciaindica que los restosbisontes de fecha paleoindia en BonfireShelter demuestranquefue un sitio de procesamientotal como sugirio Binford(1978), en vez de ser una localidad de una matanza primariatal como originalmentesepropuso (Dibble 1968). Bement(this issue) plantea variosproblemas sobre nuestroandli- sis, entre otras cosas que omitimosla informacionpertinente que tiene que ver con la interpretaciondel sitio. Sin embargo, su comentarioes el resultadode una mala interpretacionde nuestradiscusiony los datos, tal como explicamosen esta respuesta.

et al. (2005) explored whether the how muchof the Late Glacialvalley fill was since Paleoindian-agebison bone assemblageat removed, a matter relevant to the question of BonfireShelter (41VV218) was the result whetherbison carcasssegments could have been of a jumpkill at thatspot, as originallyinterpreted hauledinto this shelter,now -18 m abovethe val- (Dibble and Lorrain1968), or alternativelywas a ley floor. localitywhere carcass parts of animalskilled else- Ourresults, in brief,indicated this would have where were transportedand processed (Binford been an ideal spot for jumping bison relativeto 1978:475).Our study involved three elements: (1) otherareas in the vicinity of the shelter.However, fieldmapping and Geographic Information System it maynot have been so used.Reanalysis of thebison (GIS) analysis of the terrain,in search of drive remainsindicated that skeletal parts present at the lanes or topographicfeatures that might indicate site were more suggestiveof a processinglocality. how animalswere driven toward the cliff edge, and Importantly,we expresslystated this did not speak the feasibilityof doing so; (2) a thoroughreanaly- to howthe bison were killed, only that these remains sis of thebison remains recovered in the 1963-1964 didnot indicate this was the probable kill locus.And excavations,to see if the skeletalelement patterns althoughwe were unable to securely determine indicateda primarykill or secondaryprocessing; where the valley floor was in Late Glacial times, and (3) a study of the geomorphichistory of the evidencesuggested it was perhapsjust a few verti- canyonin whichthe shelteris located,to determine cal metersbelow the shelterentrance, and hence

Ryan M. Byerly, Judith R. Cooper, and David J. Meltzer Departmentof Anthropology,Southern Methodist University, Dallas, TX 75275 ([email protected]) Matthew E. Hill Departmentof Anthropology,University of Arizona,Tucson, AZ 85721 Jason M. LaBelle Departmentof Anthropology,Colorado State University,Fort Collins, CO 80523

373 374 AMERICAN ANTIQUITY [Vol. 72, No. 2, 2007]

hunters who had made a kill nearby were not case? Second, if the facts are as they appear(i.e., impededby the currentlysteep slope. BB2 resultedfrom a jump kill), why does such an Bementcriticized portions of ourstudy, leading anomalouspattern exist in thearchaeological record, with the suggestionwe omitted "pertinentinfor- and how might it be explained (Byerly et al. mationthat is directlyrelevant to the interpretation 2005:626; see also Forbis 1969; Frison 1991)? of the site and to an evaluationof [our] conclu- Some obvious hypothesesexplain the latter.For sions."As we had explicitly called for more evi- example,repeated jump kills on theNorthern Plains dence to resolve the use of this site in Paleoindian (e.g., Frison1991) were probably the result of inten- times (Byerlyet al. 2005), Bement'scomments, as sificationstrategies geared towards the relief of sea- a participantin the work at BonfireShelter in the sonallyconstrained resource limitations, such that early 1980s, are welcome. That said, they do not long-termnutrient storage in the form pemmican significantlyadvance the issue, sincethey are based productionwas implementedto subsidizelean win- on a misinterpretationof our argumentsand avail- ter months (e.g., Reeves 1990; Todd 1987). The able data.Although we arecontent to let ourpaper absenceof such sites on the SouthernPlains indi- standon its merits,we takethis opportunityto clar- catesthat hunter-gatherers in theregion were never ify the natureof the argumentand evidence, and faced with the same limitations,and thereforedid expandon our results. not need to use similaracquisition strategies. Bement(this issue) assertswe committeda log- But, of course,one must resolve the firstques- ical fallacy by reasoningthat the Paleoindian-age tionbefore turning analytical attention to thesecond. bonebedat Bonfire Shelter(BB2) "didnot result Thatis whywe analyzedthe BB2 faunalassemblage, froma jump because no otherPaleoindian jump site the surfacetopography, and the geomorphichistory has everbeen found."Had we actuallyreasoned in of thecanyon, and maintained a firewallbetween the thatmanner, he wouldhave been rightto makethat issue of the largerpatterning in the archaeological claim. However,we did not do so. To be sure,we record,and the empiricalevidence for/against BB2 illustratedthe distributionof bison jump kills in being the resultof a jumpkill. time and space (Byerly et al. 2005:Figures1, 13). Inregard to thatempirical evidence, we deemed Yet,this merely reinforces the point, made long ago it usefulto reanalyzethe bison from the 1963-1964 by Dibble(Bonfire Shelter's original investigator), excavations ratherthan rely on published data, that Paleoindianbison jumps are scarce on the which would do little more than replicateother SouthernPlains (Dibble 1970:252;see also Forbis studies(e.g., Binford1978:475). We realized these 1969:91).We then went on to agree with Dibble dataare biased, most notablythat they come from thatjust becausea phenomenonis rare,that is not a relativelylimited area of the site (Byerly et al. evidence againstany particularlocality, say BB2 2005:605).Those biases, of course,equally hinder at BonfireShelter, being an expressionof thatphe- originalinterpretations of the site. We also did not nomenon (i.e., a bison jump; Byerly et al. reanalyzethe bison from the 1980s excavations 2005:597). Bement's commentabout our reason- but,as explained,that should not materiallyimpact ing is thereforeincorrect and irrelevant. our conclusions,since thatlater sample was small Moreover,if usingthe lack of Paleoindianbison (Numberof IndividualSpecimens [NISP] = 51; jumps as evidenceagainst BB2 being a jump is a Bement 1986:26)and represents a minorpercent- fallacy,then why is it logical for Bementto use the age (< 2.5 percent)of the total remainsfrom BB2 Late Archaicbison bonebed at Bonfire (BB3) as (Byerly etal. 2005:605). evidence/or a Paleoindianjump? (e.g., "Byerlyet Based on that analysis,we calculateda Mimi- al. fail to explorethe implicationsof BB3 which is mum Numberof Individuals (MNI) of 24 to 27 undeniablythe result of LateArchaic bison jumps" Bison antiquus(Byerly et al. 2005:625).Nowhere [Bementthis issue]). The logic cuts both ways, so do we suggestthis is the "totalor ultimate" (Bement too does thefallacy. For that reason, we also explic- this issue) numberof animalsin BB2. The origi- itly omittedBB3 from our discussion. nal investigatorsarrived at a similarMNI estimate, The scarcityof Paleoindianbison jumps does but then arbitrarilymultiplied it by four, on the warranttwo questions:first, is therean alternative undemonstratedassumption that only half the ani- hypothesisthat might accountfor the facts of the malspreserved and that unexcavated portions of the COMMENTS 375

site,particularly the talus cone, containedskeletons notfurther our understanding of BonfireShelter site in the sameor greaterdensity and proportion (e.g., history or Folsom and Plainview hunting strate- Lorrain1968:84). This assumption rests on thecon- gies,"it is the only way the availablearchaeologi- tention that the bison enteredthe shelter via the cal data can be evaluated.But then perhapsthe notch in the cliff face. As argued(Binford 1978; issue is not so muchthe separationof these strata, Byerly et al. 2005), thatmay not be the case at all. but their overlap.Bement (this issue) assertsthat BB2 bone andlithics (Byerlyet al. 2005:Figure1) the presenceof "bothFolsom and Plainviewpro- are highly concentratedin the southernend of the jectile points suggests at least two temporaland shelter, and certainly greater numbers are not societalgroups." That's hardly a compellingclaim expectedin the unexcavatednorthern portions of on eithertheoretical or empiricalgrounds. Projec- the shelter. Indeed, Dibble notes the bonebed tile point styles are not nicely layer-cakedin time; thinnedand element frequencies diminished away they can and do overlap(Sellet 2001). Moreover, from the excavations(Dibble 1968:29), a pattern they actuallydo at Bonfire Shelter.Granting that Bementlikewise observed(1986:19). Regardless, threedepositional episodes compose BB2, Dibble the mattermust be resolvedby testing,not through reportsthat the lowest "stratum,"Component A, unfoundedassertions about a relationshipbetween yieldedboth Folsom andPlainview points (Dibble skeletalcounts and site area. 1968:Table 2; also Cooper and Byerly 2005). Our analysisof dentalpatterns indicated BB2 Bement's (this issue) concernsabout Folsom and probablyrepresented a single summerkill (Byerly Plainvieware thus misguided,even acceptingthe et al. 2005:612). Bement (this issue) points to the tripartitestratigraphic division. difficultyof making such a seasonal assignment, Moreover,accepting this division renders mean- andasserts we failto considerthe criteria that would inglessBement's estimate of thetime spanningand indicatemore than one event took place, namely separatingthe supposed multiple kills. Leaving "thepresence of stratifieddeposits, differential fau- aside the degree of confidenceone can have in a nal taphonomies, or artifact assemblage differ- seriesof radiocarbonages with standarddeviations ences." We agree the identification of a single as largeas theones fromBB2 (Holliday2000:25 1), componentvia dentalcohort clustering is a diffi- and from a period characterizedby significant cult assessment,especially with as small a sample radiocarbonplateaus (i.e., the latterportion of the as thatfrom BB2. But it is hardlyunheard of (e.g., YoungerDryas; see Blackwell et al. 2006; Kita- Toddet al. 1992).Furthermore, we do notpreclude gawa and van der Plicht 1998), such an estimate the possibility that the dental data may indicate necessarilyassumes sediment deposition is linear "multiple,closely timed deathevents" (Byerly et over time. This reasoningis inconsistentwith the al. 2005:612), nor limit our interpretationto den- assertionthat three separate depositional episodes tal data. occurred. We consideredthe possibility,raised by Dibble Although three BB2 deposits (A, B, and C; (1968) and assertedby Bement (this issue), that Bement1986:Figure 5) wereidentified in the 1980s BB2 was subdivisibleinto severalstrata. However, excavations,bone is only sparsely present in A, Dibble (1968:29-30) observedthat BB2 is divisi- with the majorconcentration lying at the B/C con- ble into threestrata only on the southernslope of tact; no differentiationby burning(as seen in the the talus cone and not at all in the "centraland 1960s)is noted(Bement 1986). This, coupled with northernareas of its occurrence"(Dibble 1968:29). the reworkingof Zone 2a underBB2 in the central Lorrain(1968:92), in fact, cites the lack of distin- and northerndeposits (Dibble 1968:21),indicates guishablestrata across BB2 as a reasonwhy bones thata directcorrelation between Bement's (1986) werenot separatelyanalyzed, and why seasonality andDibble's (1968) stratigraphicsequences is per- was likewisenot assessedper stratum.These facts hapsunfounded, and that the BB2 separationcould unfortunatelynegate any investigationof "differ- representan isolatedpostdepositional burn feature entialfaunal taphonomies" (Bement this issue) in buriedby one or more series of talus slope wash the extantassemblage. events (Byerly et al. 2005:612). For that matter,although Bement (this issue) Bement (this issue), following Lorrain suggeststhe "collapseof threestrata into one does (1968:100), furtherargues that crania and pha- 376 AMERICAN ANTIQUITY [Vol. 72, No. 2, 2007]

Table 1. SHEI for "High-SurvivalElements" in BB2.

Element TPb SFUI MNE MAU p. PiPnCpj)] SHEI Femur 100.0 100.0 33 16~5 !l80 ^309 Tibia 58.1 62.8 26 13.0 .142 -.277 Metatarsal 15.9 37.0 14 7.0 .077 -.197 Humerus 28.4 36.8 30 15.0 .164 -.296 Radius 19.7 25.8 26 13.0 .142 -.277 Mandible 10.4 11.5 26 13.0 .142 -.277 Skull 10.4 9.1 5 5.0 .055 -.159 Metacarpal 6.0 5.2 18 9.0 .098 -.228 Total 178 91.5 1.000 -2.021 .972a aTheevenness value fits only within the range of the bulk transportstrategy modeled by Faith and Gordon(2006) at sample size (MNE total) = 150. It also does not fit within the range of values for an unconstrainedor non-transportedbone assemblage, implying selective transport(Faith and Gordon2006). Element data from Byerly et al. (2005:Table2). bPj= Element MAU/X MAU (Faithand Gordon2006); SHEI = -Z {p^ln^)] }/ln(X Element),I Element = 8 in this instance (Faithand Gordon2006); TP = Total Products(Byerly et al. 2005:Table2); SFUI = StandardizedFood Utility Index. langesoccur in equallyhigher frequency than they craniarelative to thatof otherelements, in thiscase, ought to were bison carcass parts differentially high-utilitylimb bones. As noted, comparisonof transportedto this spot for processing.There are ourderived relative element frequencies to standard two problemswith this argument:(1) the cranial bison utility indices best fit a bulk transportstrat- MNI of 18 may be inaccurate,and (2) utilitymod- egy (rs= .54,/? = .05). eling incorporatesrelative, not absolute, element However,traditional methods of utilitymodel- abundance. ing relying on statisticalcorrelation and pattern In regardto the former,Lorrain's MNI deriva- "eye-balling"(Faith and Gordon2006) are sub- tions are problematic. In her Table 9 (Lorrain jective means of evaluating transportstrategies 1968:80) cranialdata are presentedaccording to (Byerly et al. 2005:616-618; also Beaver 2004; specifiedor unspecifiedportion (maxilla and mis- Faithand Gordon 2006; Rogers2000; Rogersand cellaneous) and anatomicallandmark (auditory Broughton2001). Faithand Gordon(2006) show bulla),per unit. These dataare further summarized that rank-order(rs) correlationsfor bulk utility per element and elementportion in her Tables 18 strategies,in particular,do not reliably indicate and 19 (Lorrain1968: 100). The maximumper por- relationshipsbetween economic utility and ele- tioncranial MNI is thatfor left maxilla(MNI =11). mentfrequency at low samplesizes (i.e., Minimum The cranialMNI of 18 was apparentlyderived by Numberof Elementstotals [MNE]), such that Type addingthe individuallyderived MNIs for left and II errorsdramatically increase with low element rightmaxilla as well as thatfor auditorybulla (Lor- abundance(Faith and Gordon2006). rain 1968:80). Since auditorybulla are temporal Accordingly,and following Faith and Gordon features,not maxillary,and because it is unknown (2006), we reevaluateour originaltransport infer- if maxillawere maximallydistinguished by age or ences here (Tables1 and 2) in termsof the Shan- overlappinganatomical features (see Hill 2001) in non EvennessIndex (SHEI),using our published Lorrain'sanalysis, we consideredthe maximum MNE data(Byerly et al. 2005:Table2). These data maxillaryestimate of 11 (Byerlyet al. 2005:Table indicatea cranialMNE of five, based on left inci- 2) to be the best minimumestimate for BB2 cra- sive counts (NISP = 5, MNE = 5), left maxilla nia, and not the originallyreported 18. counts(NISP = 7, MNE= 3), andleft petrous counts Nonetheless,this wouldstill implythat 1 1 of 24 (NISP= 5, MNE = 5). Unconstrainedassemblages (~ 46 percent)expected crania are present. This is should be characterizedby an evenness of one, indeed a large numbergiven that craniaought to indicating no differentiationin relative element be rareat a processinglocality. It is not, however, abundance,with evenness declining across bulk absolutecranial abundance that is importantto our andunbiased assemblages, and gourmet strategies transportmodeling, but rather the relative frequency characterizedby highlyuneven distributions of ele- (as percent MinimumAnimal Units [MAU]) of ments (Faith and Gordon 2006; also see Todd COMMENTS 377

Table 2. Spearman's(rs) Correlationsfor "High-Survival 1968:Table14). Elements"in BB2. However,the degreeor lack of articulationmay have little to do with butcheryactivity, and may Correlations p InferredStrategy rs speak to other taphonomic variables. Lorrain %MAUvs. SFUI ^59 ~A3 Weak Bulk refers to the %MAUvs. TPMRa .74* .04 Bulk (1968:93) "mixing of the bones by waterthat the notch"to the %MAUvs. TP .83* .02 Bulk pouredthrough explain aUses the cranialTP value for mandibles(Emerson 1993). low abundanceof mandiblesin N50/W70.We con- Utility values presentedin Table 1. sideredthe possiblerole of waterin influencingthe *p < .05 distributionand representation of skeletalelements (Byerly 2005:613).Although detailed data on the 1983:Table24). force and volume of flow from the notch during At a samplesize of 178 (totalMNE) for "high- periodsof high-intensityoutput do not exist, there survivalelements" (e.g., femora, tibiae, metatarsals, areindications that water-flow occurred within the humeri,radii, metacarpals, mandibles, and crania; shelterand distributed portions of the BB2 deposit see Faithand Gordon2006), the SHEIfor BB2 is (Byerly et al. 2005:613-614). only consistentwith a bulktransport strategy where Bement (this issue) asserts otherwise,but the high-utility upper limbs were selectively trans- archaeologicalpatterns he cites are equivocalon ported over low-utility crania and metapodials this issue, includingthe "spoke-like"arrangement (Faithand Gordon 2006:Table 4). These datasup- of bonesfrom a smallportion (just ~ 9 m2)of north- portthe hypothesis that BB2 representsthe remains ern portion of the rockshelter excavated in of transportedbison carcassportions (Byerly et al. 1983-1984. We talliedthe orientationof the bone 2005;also Binford1978). Of course,such evidence elementsin thisapparent feature (data derived from is only one partof the equationof site-use;hence Bement1986:Figure 13) in 10°increments between ourexploration of skeletalelement articulation and 180° and 360° (to remove mirrorimaging; see distribution,and our acknowledgmentof fluvial Lyman1994: 178). Theirdistribution was thensta- activityas a possible taphonomicagent governing tisticallycompared to a randommodel (following those variables within BB2 (Byerly et al. Kreutzer1988; Meltzer 2006). Expected values 2005:613). were derivedby dividingthe total sample size (in In regardto the latter,Bement (this issue) cites this case 24 elements)by the numberof possible element co-association,most notablythe scapula classes (n = 18), modelingan assumedeven distri- concentrationin N60/W50, andthe skulland axial bution.The resultingstatistic (%2 = 13.50,/?= .70) elementconcentration in N60/W50 andN50/W50- showsthat the difference between the observed and N50/W60 at the base of the talus cone (Lorrain modeled expectedvalues is insignificant.In non- 1968:Table 17, Figure40), as testimonythese ani- statisticalterms, bone orientationsshow no pre- mals died in place.Yet, these concentrationswere ferredalignment or pattern. not interpretedas such by Lorrain.Rather, they In any case, the criticaltaphonomic question of weretaken as indicatingdeliberate stacking related what pre- and postdepositionalprocesses influ- to "intensive"butchery (Lorrain 1968:96). Lorrain enced bone distributionpatterns will need to be commentsthat: resolvedbased on a muchlarger area than this small of the shelter.Here, too, it must be The articulationsin the lowerbed (BoneBed portion recognized thatfluvial action was probablynot uniform 2), which were very few in number,usually acrossthe site since some areas,like those nearest consistedof onlytwo bones, whereas the upper the rearwall of the shelter,where the hearth bonebed [BB3] oftencontained articulations only featureis or far from the are of anentire limb or anentire vertebral column preserved, pour-off, better thanothers. we and, in at least one case, of an entire protected Accordingly, sug- young the "bonebedas mosaic" and bison [1968:93]. gest concept (Todd Rapson 1999), incorporatingan explorationof all Indeed, only seven unspecified articulationsare taphonomicvariables, is a far more illuminating cited (Lorrain 1968:Table 16), including two perspectivethan a simple searchfor "patterningin unfused femur/innominatearticulations (Lorrain culturalremains" (Bement this issue). It is via the 378 AMERICAN ANTIQUITY [Vol. 72, No. 2, 2007]

Table 3. SignificantSpearman's (rs) Correlationsfor Utility among SampledBison Bonebeds.

TF TF SKF MAR SFUI Assemblage21 MNEb rs P rs P rs P rs P rs P CSP 643 T64 X)9 ^43 40 ^43 40 3l !54 ^73* !()3 GN(U) 308 -.90** .00 -.94** .00 -.94** .00 -.83* .04 -.71* .05 GN(L) 840 -.83* .01 -.89* .02 -.89* .02 -.77 .07 -.67 .07 AB(AB) 367 -.50 .21 -.90* .02 -.90* .02 -.75 .08 -.28 .51 BB2 178 .74* .04 .84* .04 .84* .04 .70 .13 .59 .13 aAB(AB) = Agate Basin, Agate Basin; BB2 = Bonfire Shelter,Bonebed 2; CSP = Casper;GN(L) = Glenrock,lower; GN(U) = Glenrock,upper. Data from Byerly et al. (2005:Table2), Hill (2001:Table3.21), and Todd (1987:Table2). bTotalMNE for "high-survivalelements" listed in Table 1. cUtility data summarizedin Byerly et al. (2005:Table2); TP = Total Products,TF = Total Fat, SKF = Skeletal Fat, MAR = MarrowFat, SFUI = StandardizedFood Utility Index. */?<.05;**p<.01 formerthat a truly "systematicinvestigation fol- analyses, however, and much in line with Todd lowing the scientific method" is achieved in (1987), we canexplore skeletal element abundance bonebedanalyses. and utility across GreatPlains kill-butcheriesto Finally,Bement raises severalquestions about illustratenumerically, rather than graphically, the ourFigures 14 and15 (Byerlyet al. 2005:622-623). positionof BonfireShelter relative to othersites in These figures were intendedto display BB2 ele- termsof inferredactivity, regardless of temporalor mentpatterning and butchery activities in termsof regionalassociation. the continuumof kill andprocessing strategies evi- Here,we presenta broad-basedcomparison of dencedamong Great Plains Paleoindian bonebeds, element patterningamong 13 Great Plains kill- not to "showthe site [BB2] is not a kill" (Bement butcheries,incorporating data from 1 6 components. thisissue). We deliberately did not incorporate data Percent MAU data for "high-survivalelements" from knownjump sites because, as we discuss in (Faithand Gordon 2006) in Paleoindian,Late Pre- detail(Byerly et al. 2005:620-621), thevast major- historic,and modernbison assemblages,summa- ity of jumpkills areLate Prehistoric in age. We are rizedin Todd(1987:235, Table 2) andHill (2001), quite explicit on what we considerto be "jumps" arecorrelated (Spearman's r) to Emerson's(1993) (Byerly et al. 2005:599) and recognize the many bisonutility values and Metcalfe and Jones' (1988) variantsof this strategy,although (contra Bement StandardizedFood Utility Index (SFUI), as pre- this issue) neitherLipscomb nor Olsen-Chubbuck sented in Faith and Gordon(2006). Results indi- comply with what we would defineas a jump kill. catesignificant relationships for five of the sampled Plainview may be an exception, though data to assemblagesincluding BB2 (Table 3), implying resolve this are lacking (see Holliday 1997:107). good fits with Binford's(1978) transportmodels. Here, beyond the co-associationof bison with a These data indicate strongreverse utility pro- slope amenableto jumping,the bison assemblage files for identifiedkill sites for eithertotal products is also associatedwith a marshor pond (Holliday orfat, including the Late Prehistoric Glenrock bison 1997:107;Sellards et al. 1947), which may better jump kill and the Agate Basin componentkill at explainbison presencein the areaand at the site. AgateBasin (Table 4). A weaktotal product reverse Thus, Bement's (this issue) comments notwith- utility profile also characterizesthe Casperkill. standing,data from BB3 andother Late Holocene Bothcomponents of Glenrocklikewise show strong jumps have no bearingon our Figures 14 and 15. selectionagainst high-total utility and high-fat ele- Furthermore,butchery evidence (e.g., green- ments, implying probabletransport of these ele- break and cutmarkfrequencies) is neither well- ments from the site, althoughcarnivore-induced represented nor consistently recorded across deletion may be an agent of this patterning(see Paleoindianassemblages (Todd 1987, 1991; Hill Toddet al. 1997). The Agate Basin componentof 200 1) . OurFigure 1 5 was thuslimited by whatdata AgateBasin displays similar selection against high- was available,hence the incorporation of bothNISP fatutility elements. Bonfire Shelter BB2, however, and MNE data for green-brokenelements from exhibitsa disparatepattern for high- total utility and such a variety of sites. Following our previous high-skeletalfat limb bones, implying selective COMMENTS 379

Table4. StrategiesInferred from Spearman's(rs) Correlationsand SHEI among SampledBison Bonebeds.

Assemblage Strategy (r/ SHEI Strategy (SHEI)a CSP Reverse Bulk: TP .967 Reverse Bulk GN(U) Reverse Bulk: TP & TF; Unbiased SKF & MAR .950 Reverse Bulk / Unbiased GN(L) Reverse Bulk: TP & TF; Unbiased SKF & MAR .925 Reverse Bulk / Unbiased AB(AB) Reverse Bulk: TF; Unbiased SKF & MAR .949 Reverse Bulk / Unbiased BB2 Bulk: TP & TF; Unbiased SKF .912 Bulk aBasedon data from Table 3.

transportof these partsto the site. These data do a jump, canyontrap, etc. (Byerlyet al. 2005). not directlyidentify BB2 as a processingsite, but We also recognize the anomalies in our theystrongly suggest that based on modeledtrans- processing-sitehypothesis; most obviously, that portbehaviors, and in the absence of evidence of the BB2 lithic assemblageis dominatedby pro- significantcarnivore activity or density-mediated jectile points (Byerly et al. 2005). A processing attrition,selective transport of high-utilityelements locality ought to yield more butcheringtools and was a factorin site formation(Byerly et al. 2005). tool resharpening/productiondebris. In an effortto In this regard,we recognizethat utility model- determinewhether this scarcityis real,or possibly ing is by no means a perfectanalytical system; it a resultof fluvialreworking, or even archaeologi- assumesstrict behavioral guidelines for activities cal bias, we have since conductedadditional test- knownamong modern hunter-gathers to be logis- ing at the site and on extantcollections. We found tically highly variable(e.g., Egeland and Byerly some indicationof excavatorbias (Byerly et al. 2005; Lupo 2006). Nonetheless,it serves as a rea- 2007), but only continuedtesting and excavation sonable analyticalbaseline to apply archaeologi- will ultimatelyresolve this issue and othersmen- cal andmid-range data to understandingprehistoric tioned. Bement (this issue) indirectly offers a humanlifeways (e.g., Toddand Rapson 1999). hypothesisin this regard,contending that flakes of We furtheragree with Bement (this issue) that rockshelterlimestone could have been used by understandingBonfire Shelterdepends on know- Paleoindianhunters in processingBB2 bison.Test- ing "the relationshipof the animals to the sur- ing this hypothesiswill requiredetailed analysis of roundingterrain," of which the notch above the the limestone spall from the shelter, which cer- shelteris certainlyone majorcomponent. But that tainlypresents a formidabletask. is just one component,which makes it all the more Pendingsuch an analysis,or furtherinvestiga- puzzlingwhy Bementignores our comments on the tion of taphonomicissues (including,but not lim- geomorphichistory of the canyon.Although our ited to studies of excavation bias, rockshelter findingsas to the elevationof the valley floor in hydrology,and canyon paleohydrology),or per- LateGlacial times were inconclusive, that does not haps additionalexcavations within BB2 itself, we meana plungeinto the shelteris the most, or only, continueto see farmore compelling empirical sup- plausiblemode of entry.Nor shouldthe presence portfor the hypothesisthat the BB2 bison remains of the cliff and the notch overshadowother possi- representa processinglocality, as Binford(1978) ble use(s)of thisrockshelter. Rockshelters are well- suggested, ratherthan a kill, as originally inter- recognized as landscape features employed in preted(Dibble 1968). multiplecapacities by prehistorichumans and we cannotdiscount intothe shelter eitherPale- entry by We thank Lee Bement and Elton Prewitt oindiansor bison via the floor. Acknowledgments. canyon for theirhistorical perspective and willingness to discuss site Thereis no doubtthat Paleoindians were capa- activities at Bonfire Shelter. We also thank J. Tyler Faith ble of andemployed multiple strategies to dispatch (GWU) for the invaluablecomments. CharlesEgeland (IU) largeherds of bison (Frison2004), and at no point provided useful feedback during this writing. We are also do we contestthis. our gratefulto David Wilson (SMU) who composed the Spanish Furthermore, interpretation version of our abstract.Our work at Bonfire Shelterwas of where the animalsin BB2 were sup- says nothing ported by the Quest Archaeological Research Program killed (only thatit does not appearto have been in (SMU). As always, the Skiles family graciouslyfostered our theshelter), or what strategy was employed, whether researchat Bonfire Shelter. 380 AMERICAN ANTIQUITY [Vol. 72, No. 2, 2007]

References Cited NorthwesternGreat Plains of NorthAmerica. Unpublished Ph.D. dissertation,Department of Anthropology,Univer- Beaver,Joseph E. sity of Wisconsin,Madison. 2004 IdentifyingNecessity and SufficiencyRelationships Holliday,Vance T. in Skeletal-PartRepresentation using Fuzzy-SetTheory. 1997 PaleoindianGeoarchaeology of the SouthernHigh AmericanAntiquity 69: 13 1-140. Plains. Universityof TexasPress, Austin. Bement,Leland C. 2000 The Evolution of PaleoindianGeochronology and 1986 Excavationof the Late PleistoceneDeposits of Bon- Typology on the Great Plains. Geoarchaeology fire Shelter,41W218, ValVerde County, Texas. Archaeo- 15:227-290. logical Series 1. TexasArchaeological Survey, Austin. Kitigawa,Hiroyuki, and Johannes van der Plicht Binford,Lewis R. 1998 AtmosphericRadiocarbon Calibration to 45,000 yr 1978 NunamiutEthnoarchaeology. Academic Press. New B.P.:Late Glacial Fluctuationsand CosmogenicIsotope York. Production.Science 279:1187-1190. Blackwell,Paul G., CaitlinE. Buck, and PaulaJ. Reimer Kreutzer,Lee 2006 ImportantFeatures of the New RadiocarbonCalibra- 1988 MegafaunalButchering at LubbockLake, Texas:A tion Curves.Quaternary Science Reviews 25:408^4-13. Taphonomic Re-analysis. Quaternary Research Byerly,Ryan M., JudithR. Cooper,David J. Meltzer,Matthew 30:221-231. E. Hill, andJason M. LaBelle Lorrain,Dessamae H. 2005 On Bonfire Shelter (Texas) as a PaleoindianBison 1968 Analysisof the Bison Bones fromBonfire Shelter. In Jump:An Assessment using GIS and Zooarchaeology. Bonfire Shelter: A StratifiedBison Kill Site, Val Verde AmericanAntiquity 70:595-629. County,Texas, pp. 77-132. MiscellaneousPapers No.l. Byerly,Ryan M., JudithR. Cooper,David J. Meltzer,and Jim Texas Memorial Museum Publications, University of Theler Texas,Austin. 2007 Exploring PaleoindianSite-Use at Bonfire Shelter Lupo, KarenD. (41VV218). Bulletinof the TexasArchaeological Society, 2006 What Explains the Carcass Field Processing and in press. TransportDecisions of ContemporaryHunter-Gatherers? Cooper,Judith R., and Ryan M. Byerly Measuresof EconomicAnatomy and Zooarchaeological 2005 The Significanceof a SecondFolsom Projectile Point Skeletal PartRepresentation. Journal of Archaeological fromBonfire Shelter, Texas. Current Research in thePleis- Methodand Theory13:1 9-66. tocene 22:41-43. Lyman,R. Lee Dibble, David S. 1994 VertebrateTaphonomy. Cambridge University Press, 1968 TheArchaeology. JnBonfire Shelter: A StratifiedBison Cambridge. KillSite, Val Verde County, Texas, pp. 1-76. Miscellaneous Meltzer,David J. PapersNo. 1. TexasMemorial Museum Publications, Uni- 2006 Folsom:New Archaeological Investigations of a Clas- versityof Texas,Austin. sic PaleoindianBison Kill.University of CaliforniaPress, 1970 On the Significanceof AdditionalRadiocarbon Dates Berkeley. From Bonfire Shelter, Texas. Plains Anthropologist Metcalfe,Duncan, and KevinT. Jones 15:251-254. 1988 A Reconsideration of Animal Body-Part Utility Egeland,Charles P., and Ryan M. Byerly Indices.American Antiquity 53:486-504. 2005 Applicationof ReturnRates to LargeMammal Butch- Reeves, BrianO.K. ery andTransport among Hunter-Gatherers and its Impli- 1990 CommunalBison Huntersof the NorthernPlains. In cations for Plio-PleistoceneHominid CarcassForaging Huntersof the RecentPast, editedby Leslie B. Davis and and Site Use. JournalofTaphonomy 3:135-158. BrianO.K. Reeves, pp. 168-194. UnwinHyman, London. hmerson,Alice M. Rogers,Alan R. 1993 The Role of Body PartUtility in Small-scaleHunt- 2000 Analysis of Bone Countsby MaximumLikelihood. ing underTwo Strategiesof CarcassRecovery. In From Journalof ArchaeologicalScience 27:1 1 1-125. Bones to Behavior:Ethnoarchaeological and Experimen- Rogers,Alan R., andJack M. Broughton tal Contributionsto the InterpretationifFaunal Remains, 2001 Selective Transportof Animal Parts by Ancient editedby JeanHudson, pp. 138-155. Centerfor Archae- Hunters:A New StatisticalMethod and an Applicationto ological InvestigationsOccasional Paper No. 21. South- the EmeryvilleShellmound Fauna. Journal of Archaeo- ern Illinois University,Carbondale. logical Science 28:763-773. haith,J. Tyler,and Adam D. Gordon Sellards,Elias H., Glen L. Evans,and Grayson E. Meade 2006 Skeletal Element Abundances in Archaeofaunal 1947 Fossil Bison andAssociated Artifacts from Plainview, Assemblages:Economic Utility, Sample Size, andAssess- Texas. Bulletin of the Geological Society of America ment of CarcassTransport Strategies. Journal ofArchae- 58:927-954. ologicalScienceM press[doi:10.1016/j.jas.2006.08.007]. Sellet, Frederic Forbis,Richard G. 2001 A ChangingPerspective on PaleoindianChronology 1969 Review of Bonfire Shelter:A StratifiedBison Kill andTypology: A View fromthe NorthwesternPlains. Arc- Site, Val Verde County, Texas. American Antiquity tic Anthropology38:48-63. 34:90-91. Todd,Lawrence C. Frison,George C. 1983 The HornerSite: Taphonomyof an EarlyHolocene 199 1 PrehistoricHunters of theHigh Plains, 2nded, edited Bison Bonebed.Unpublished Ph.D. dissertation,Depart- by GeorgeC. Frison.Academic Press. New York. ment of Anthropology,University of New Mexico, Albu- ZUU4 survival by Hunting:Prehistoric Human Predators querque. andAnimal Prey. University of CaliforniaPress, Berkeley. 1987 Analysis of Kill-ButcheryBonebeds and Interpreta- Hill, MatthewG. tion of PaleoindianHunting. In TheEvolution of Human 2001 PaleoindianDiet and SubsistenceBehavior on the Hunting, edited by Matthew H. Nitecki and Doris V. COMMENTS 381

Nitecki,pp. 225-266. PlenumPress, New York. L. AdrienHannus, Lynette Rossum, and R. PeterWinham, 1991 Seasonality Studies and Paleoindian Subsistence pp. 136-157. OccasionalPublication No. 1. Archeology Strategies.In HumanPredators & PreyMortality, edited Laboratory,Augustana College, Sioux Falls. by Mary C. Stiner,pp. 217-238. WestviewPress, Boul- Todd,Lawrence C, and David J. Rapson der. 1999 FormationalAnalysis of Bison Bonebeds and Inter- Todd,Lawrence C, JackL. Hofman,and C. BertrandSchultz pretationof PaleoindianSubsistence. In Le Bison: Gibier 1992 FaunalAnalysis and PaleoindianStudies: A Reex- et Moyen de Subsitancede homes du Paleolithiqueaux aminationof theLipscomb Bison Bonebed. Plains Anthro- Paleoindiensdes GrandesPlaines, editedby Jean-Philip pologist 37:137'-165. Brugal,Francine David, James Enloe, and Jacques Jaubert, Todd, LawrenceC, MatthewG. Hill, David J. Rapson, and pp. 480-499. EditionsAPDCA, Antibes, France. GeorgeC. Frison 1997 Cutmarks,Impacts, and Carnivores at the CasperSite Bison Bonebed.In Proceedingsof the 1993 Bone Modifi- cation ConferenceHot Springs,South Dakota, editedby Received January8, 2007; Accepted January8, 2007.