TUI CHUB TAPHONOMY AND THE IMPORTANCE OF MARSHRESOURCES IN THE WESTERN GREAT BASIN OF NORTH AMERICA

Virginia L. Butler

Debates about the importance of marsh resources to prehistoric hwnan subsistence in the western Great Basin are long-

standing. Recent questions regarding the natural vs.cultural origin offISh remains in lakeside archaeological sitesjllrtherimpede understanding of ancient subsistence patterns. Taphonamic study of a huge assemblage of tui chub (Gila bicolor)remains from an archaeological site in Stillwater Marsh, western Nevada, was undertaken to identify agents of depasition in

marsh settings. The Stillwater fISh remains showed limited surface modification-cut marks, burning, and digestive etchingand staining-and thus these attributes were not usejill indicators of origin. Fish mortality profiles, reconstructed by regres-sion analysis of body size, indicates cultural selection of young/small fish rather than natural catastrophic mass death. Thelow survivorship of vertebrae in the chub assemblage suggests differential treatment of cranial and postcranial body part'by cultural agents. The Stillwater site fish assemblage represents a vast number of small fISh; the presence of small tui chubfrom archoeological sites throughout the western Great Basin suggests that prehistoric fishers targeted relatively small chub

in the subsistence quest.

Desde hace mucho tiempo Iran habido debates acerca de /a importancia de /as recursos de /as zonas pantanosas para la sub-sistencia prehistorica humana en /a Gran Cuenca occidental de /as Estados Ullidos. Las preguntas recielltes referelltes al

origell lIatural versus cultural de los resfos de pescado en sitios arqueolOgicos ell/as oril/as de /agos obstaculizan, alin mas.la comprension de alltiguas pautas de subsiste/ICia. Se emprendio un estudio tafoltOmico de grandes colecciolles de los res/osde cachos (Gi/a bicolor) de un sitio arqueologico ell Stillwater Marsh, en eloeste de Nevada, COli el propiJsito de idelltificarlos agelltes de sedimellto ell /as sitios palltallosos. Los restos de pescado de Stillwater permitieron observar limifadas mod-ificaciones de las superftcies-cortes. quemaduras. el grabado digestivo y manChas-y, POT eso, estos atributos no fueronindicadores "tiles del origell. Las esiadisticas recollStruidas de la mortalidad de los peces utilizando el anaJisis regresivo deltamaiio del cuerpo indicall una seleccion cultural de los peces pequeiios/jovenes mas biell que UIIQ muerte llatural masivacatastrofica. EI bajo indice de supervivencia de vertebra de /os cochos (Gila bic%r) sugiere un tratamie/lto difere/ICial delas partes craneales y poscraneales del cuerpo POT los agelltes culturales. La colecciOn de /a pesqueria de Stillwater repre-sellta un nUmero ellorme de peces pequeiios; /a prese/ICia de pequeiios cochos tui de sitios arqueol/Jgicos POT todo /a GrallCuenca occidelltal sugiere que los pescadores prehistiJricos estaball mas illteresados ell/as cachos re/ativamente pequeiios

para su sustento.

whether wetland resources could have supportedsedentary or near-sedentary populations. Someresearchers have emphasized the richness of wet-land resoun:es (e.g., fIShes, plants, birds) and sug-gested that lacustrine systems could supportresidential bases for much or all of the year (e.g.Livingston 1991; Madsen 1979, 1988; but seeBinford 1983). Others challenge this view, believ-ing that marsh resources are not capable of support-ing year-round residence and that they would beused primarily as a backup resource or when uplandresources were diminished (e.g., Kelly 1985).

T he importance of marsh resources and theirrelation to hunter-gailierer mobility strate-gies in die western Great Basin have been

die subject of continuing debate for over 60 years(e.g., Aikens and Greenspan 1988; Bettinger 1993;Heizer and Kreiger 1956; Heizer and Napton 1970;Kelly 1985, 1990; Livingston 1988a, 1991; Loudand Harrington 1929; Madsen 1988; Raven andElston 1988, 1991; Raymond and Parks 1990;Rhode 1987, 1990; Thomas 1985, 1990). In itsbasic fonn, this debate centers on die relative valueof marsh versus upland resources and considers

Virginia L. Butler 8 Department of Anthropology, Portland State University, Portland, OR 97207

American Antiquity, 61(4), 1996, pp. 699-717.Copyright C> by the Society for American Archaeology

699

[Vol. e1, No.4, 19ge]

Fllure I. Map of Cano. Desert, weltera Nnad.. Ih~.llocatioD of Stillwater Manh.

Rhode 1987; Thomas 1985) have advanced ourunderstanding of prehistoric land use in marshand upland environments (Bettinger 1993).However, documenting the role of particular foodresources in prehistoric subsistence has been ham-pered by nagging taphonomic questions about theorigins of faunal materials in marsh and adjacentcave sites (e.g., Greenspan 1988; Livingston1988a; Smith 1985). Recent events in the CarsonSink of western Nevada (Figure I) illusb'ate thiscondition. The archaeological record from theCarson Sink underwent a dramatic transformationresulting from extensive flooding of the basin dur-ing 1982-1984. Subsiding floodwaters exposednumerous pithouse features, burials, and denseartifact scatters and provided, at least for someresearchers, compelling evidence for intensive

Until recently, archaeological data available toevaluate questions of hunter-gatherer mobilitystrategies and subsistence components have beenlimited. Most of the empirical record was drawnfrom cave sites and, as Thomas (1985:28) hasnoted, these were excavated long before researchfocused on ecology and prehistoric lifeways.Detailed information on subsistence, seasonalityof site use, and mobility strategies has been diffi-cuh to extract ftom site records and excavatedmaterials obtained from field techniques drivenby questions of culture history (see Livingston1988a for an important exception).

Severallarge-scale regional surveys and exca-vations in the western Great Basin over the past 15years (e.g., Aikens et aI. 1982; Beck 1984;Bettinger 1975; G. T. Jones 1984; Kelly 1985;

REPORTS 701

and residential use of the marsh for multiple sea-sons, if not the entire year (Raven and Elston1991; Raymond and Parks 1990).

Exposure of the Stillwater Marsh sites has bol-stered the view that wetlands could support resi-dential bases during some periods in the past.However, specific understanding of the nature ofmarsh resource use based on faunal remains hasbeen difficult given taphonomic concerns.Concern over the origin of the fish remains,which represent between 20 and 85 percent of thefaunal remains in marsh sites, has been acute(e.g., Greenspan 1988; Raven and Schmitt 1991).As Raven and Schmitt (1991:56-57) recentlylament "the role of fish in human subsistenceremains poorly understood as most bones arecomplete, unburned, and show no evidence of

partial digestion."Such concern over fish bone origin is cer-

tainly reasonable. Indeed, the same floods thatexposed the Stillwater archaeological sites sug-gest processes that might have naturallydeposited fish carcasses in such locales in thepast. The exceptionally high water flows into theCarson Sink in the early 1980s led to the tremen-dous population explosion of tui chub (Gilabicolor) (Rowe and Hoffman 1987), the domi-nant fish taxon in marsh archaeological sites.Beginning in 1985, lake levels began to decline;water ceased flowing into the basin while evapo-ration continued, resulting in decreased waterlevels and high salinities. In January 1987, thelake froze, which probably raised salinitiesbeneath the ice. During the winter of 1987, anestimated 7 million fish carcasses, primarilyfrom tui chub, were found along 40 miles of theCarson Sink shoreline. Apparently the fishreached maximum salt-toleration levels and per-ished (Rowe and Hoffinan 1987).

Water levels in the Carson Sink have fluctuatedin the past (Benson and Thompson 1987; Morrison1964); tui chub populations have undoubtedlyexperienced boom-bust cycles as well. It is notunlikely that bony remains resulting from carcassesstranded along lakeshores have become incorpo-rated into lacustrine archaeological sites(Greenspan 1988; Raven and Schmitt 1991).

The Stillwater Marsh archaeological recordclearly holds promise for modeling the evolution

of hunter-gatherer mobility strategies and prehis-toric land use (Kelly 1988; Raven and Elston1988; Raymond and Parks 1990). However, inorder to incorporate the faunal record into thosemodels, we must devise ways of distinguishingbetween bone deposited by natural and by culturalprocesses (e.g., Schmitt 1988a). Here, I draw onrecent analysis of fISh remains from one of theStillwater sites (26CHI062) to assess the poten-tial contribution of natural death and deposition tothe Stillwater fish bone deposit. I evaluate multi-ple lines of evidence, including fish mortality pat-terns, bone surface modification, and body partrepresentation to argue that a human agency isprimarily responsible for the deposit. I then dis-cuss the implications of these fmdings to ourunderstanding of fish resource use in the westernGreat Basin. .

The Site: 26CHIO62

26CHI062 is located in the Stillwater Marsh, oneof a few existing wetlands situated on the floor ofPleistocene Lake Lahontan in the western GreatBasin (Figure 1). Stillwater Marsh lies at 1,180 masl in the Carson Desert of western Nevada, oneof the lowest points in the region. The largestbasin in the western Great Basin, the Carson Sinkis a wide, flat valley characterized by sand dunes,alkali flats, and slightly alkaline marshes (Kelly1988). The desert receives water from the CarsonRiver, which drains the eastern side of the SierraNevada Range in California, and, periodically,water from the Humboldt River, which drainsmountain ranges in east-central and northeasternNevada. Rainfall averages 123 mm/year.

Abundant research on the Pleistocene-Holocene geomorphic history of the Carson Sinkdocuments the dynamic nature of the drainagesystems, water level, and wetland distributionover the past 13,000 years (e.g., Benson andThompson 1987; Davis 1982; Morrison 1964;Russell 1885). The earliest archaeological recordfrom Stillwater dates to between 4,950 and 3,300B.P. (Elston et al. 1988). Elston et al. (1988) sug-gest that, during the past 5,000 years, theStillwater region probably contained transitoryshallow ponds, lakes, and marshes created byblowouts and dunes, which formed dams trappingwater.

[Vol. 61, No.4, 1996JAMERICAN ANTIQUITY702

F~ of Fish Remains.Table

NISPTaxon6i~ - - 5.2S9

Gilabicob- 956Gila bicolor cr. obI;nI.J 229Gila bicolor cr. pecIbJq., 37C~id8e -~~ 51

Cyprinidae/CltoItomidie 2,270Total 9,010Notes: Slmplea Mre &un 26<:81062; 3.2-mm IIIeIh -.UIed. NISP - mDDber of M!eIItified ~

As sh(y.VD in Table I, 9,010 specimens repre-senting ~ species of fISh were identified in the3.2-mm samples from 26CHI062. Most of thespecimens that could be identified to family arecyprinid; b1i chub is the only species of this fam-ily present. The remains in theCyprinidae/Catostomidae category include frag-mentary or eroded materials that could not beassigned to family or vertebrae that are difficult todistinguish between families.

To determine whether use of 3.2-mm meshbiased taxonomic recovery, I scanned six of theflotation samples with a low-power microscope. Iwas particularly concerned to know whetherextremely small cyprinid species (e.g.,Richardsoniw egregiw and Rhinichlhys osculw)might be present only in the fme-screen sampleand thus focused analyses on pharyngeals(toothed branchial arches found in the back of themouth), which are readily identified to species.Of the 114 pharyngeals identified in the flotationsamples, only 4 (3.5 percent) represent speckleddace (Rhinichthys osculw), whereas the rest arefrom b1i chub. Thus, even if all the sediments hadbeen screened with 1.6-mm mesh, b1i chub wouldstill be the dominant fish taxon represented.Consequently, taphonomic and subsistence issuesdiscussed below focus on this taxon.

Fish Mortality Pattens

Numerous workers have used age/size class pro-files reconstructed from archaeological and pale-ontological faunas to identify exploitationstrategies and to determine whether deposits rep-resent a cultural or natural origin (e.g.. Frison1978; Hoffecker et al. 1991; Klein and Cruz-Uribe 1984; Lyman 1991; Noe-Nygaard 1983;Speth 1983; Stiner 1991). Two theoretical modelsof population structure have been defmed for fos-sil populations. A population structure that resem-bles a typical living population. whereinsuccessively older age classes contain pr0gres-sively fewer individuals. has a catastrophic ageprofile. Natural agents (i.e.. flash floods. volcan-ism. and epidemic diseases) and cultural procure-ment strategies (i.e.. bison traps) may produce apopulation structure with a catasb'ophic age pr0-file. Reconstructed fossil populations in whichprime age adults are undeiTepiesented and young

26<;H 1 062 is one of over SO archaeologicalsites in the marsh area exposed by severe floodingfrom 1982 to 1986. Intermountain Research(Silver City, Nevada) undertook test excavationsat five of the sites (Raven and Elston 1988).Robert Kelly directed excavation at a sixth site,26CHI062, which provided the fish remains Ianalyzed and describe here. Together these sitesrepresent occupations dating between 3,200 and650 B.P., based on projectile point types andradiocarbon dates (Raven and Elston 1988). All ofthe sites are characterized by variably shapeddepressions, including some probable pithouses(Raven and Elston 1988). Faunal analyses havebeen conducted by Dansie ( 1987), Greenspan(1988), Livingston (1988b), and Schmitt (1988b).

26CHI062 is composed of five to six pithousefeatures and numerous smaller features.Excavated matrix was water-screened through3.2-mm mesh. Residue from 18 flotation samples,which included all material retrieved in 1.6-mmmesh, also was available for analysis.

Preliminary sampling and analysis of the 3.2-mm screened matrix estimated over 290,000 fishremains in the excavated sample and the presenceof only ~ species, tui chub (Gila bicolor) andTahoe sucker (Catostomus tahoensis). Most of theremains were from tui chub. To obtain a representa-tive sample of this voluminous assemblage, I drewa SO percent random sample of all ~cavation unit-level bags; from each selected bag, 100 specimenswere drawn. To ensure adequate representation ofpithouse features, three unit-level bags from ~ ofthe pitbouse features ~ completely analyzed.Given the similarities in taxonomic and body partfrequencies between these ~ samples, the results

'presented here are based on the combined samples.

REPORTS

I

I

'-

/I

( /I" /

'"

and old individuals are relatively more abundanthave attritiOnal age profiles. Natura) causes forthis population structure are starvation, accidents,

Predation, and disease-factors that affect themost vulnerable members of the POPulation, theyoung and the old Cultural procurement may alsoexploit the most vuJnerable members of the POp-ulation through either capture of living individu-als or scavenging of deceased animals thatperished from other atb'itional factors.

In the context of this study, I suggest that thePOPulation structure of a DaturaJ death assem-blage of Wi chub WOuld have a catastrophic pro-file. AU age/size classes would be affected by

shrinking lake levels and rising salinities. Tuichub in three age classes died during the winter1986-1987 at Stillwater Marsh and their car- Flpre 2. Tui c"ub Opercle (left side, media' view). s.ow-casses were found together along the shoreline Inr POrtion measured (OL). (For ,cale. tllis lengt" ranged

(Rowe and Hoffman J 987). Furthermore. becaUSe behl-eeD about 4 aDd 24 mm on the modern sPecimeDs.)

the youngest/smalJest members of the population

are the mOSt abundant, members of such classes {~should be the IDOst abundant in the fossil assem- thblage. The mortality pattern should reflect that of nthe overall POPulation age/size structure. (It is be:POSSible that some Wi chub age cohorts are more plrlikely to die from elevated salinities or reduced netwater or oxygen levels than others. If this is the ~case, then the fossil population structure would swivary from the catastrophic profile described caQJ

Unfortunately, little is known about Wi chub be ephysiology or developmental changes in Pbysiol:' Togy to evaluate this POSSibility. Ideally, the theo- generetical catastrophic profile of a natural Wi chub proddeath assemblage could be compared to an empir- tribulical one, generated from a study of the recent fish modskin at Stillwater. However, information required numbto Create such a prorue [e.g., length frequencies, Thages J was not coIJected.)'""- .

704 AMERICAN ANTIQUITY [Vol. 61, No.4, 1996]

frequency20

15

10

5

0150- 60- 70- 80- 90- 100- 110- 120- 130- 140-

59 69 79 89 99 109 119 129 139 149

Standard Length (mm)

Figure 3. Reconstructed tui chub length frequencies (SL) from 26CHI062; inset Illustrates catastrophic mortality prol"lle.

Based on a modem sample of tui chub(n = 143) collected from several locations in west-ern Nevada, I performed regression analysis(Casteel 1976:95) using the variables operclelength (OL) (maximum length of anterior borderof the opercle bone) and standard length (SL)(end of snout to end ofhypural bone) (Figure 2).The exponential model was most appropriate(,2 = .985), providing the regression formula (nat-ural logarithms):

log SL = 2.51260 + .90851(1og OL:

I then measured the OL of the 74 archaeologi-cal cyprinid specimens with the anterior borderintact. Although most of these opercles probablyare from tui chub, the element can be identified tofamily only on the basis of morphological criteria.Some of the opercles may be from speckled dace,given that dace was identified in the flotationsamples and based on the extremely small recon-structed body sizes of the archaeological fish (seebelow). This analytic difficulty should not intro-duce much error into body size reconstructions,however, given the overwhelming predominanceof tui chub in the assemblage. In addition, mor-

phological similarity in opercles across familymembers suggests that the regression modelshould hold for other species in the family.

Using the regression equation obtained from themodern sample, I then calculated predicted stan-dard lengths of the prehistoric fISh. As shown inFigure 3, the reconstructed length frequencies ofthe fossil population are nonnally distributed witha mean of 89. 14 mm ($ = 18.60), ranging between52.27 and 142.69 mm SL. As compared to the insetin Figure 3, the fossil population structure bears noresemblance to a catastrophic profile. Extremelysmall fish «50 mm SL) and medium to large fISh(> 150 mm SL) are absent entirely.

How do these data correspond to gr~ andlife history characteristics oftui chub? Chub mayreach 20-40 mm SL in their first few months oflife and grow to lengths of350 mm SL (Carlander1969:394), living to be over 30 years old(Scoppettone 1988). The modem sample ofStillwater tui chub used in the regression analysisranged from about 50 to 210 mm SL.

Whereas more detailed life history informationhas not been gathered for Stillwater Marsh fish, arecent age and growth study of tui chub in EagleLake, another relatively shallow water lake in the

REPORTS 705

Table 2.1\1i Chub (Gila bicolor) mean SL (mm) by Igeclass, Eagle Lake, CaliflXnia (DaviI1986)

S.D.

:'=2.32

:'=6.48

:'= 14.3

:'=20.0

:'=22.6

.. Mea SiD0 + 22.1I + 91.52 + 141.43 + 177.64 + 212.8

10 + 341- ~ -lAp clua oS + through 9 + oot captured.

Table 3. Estimated Size of1\li Chub 0pen:Iea.

62 (14.8)218 (68.6)63 (IS.0)

6 (1.4). 1 (.2)

~ (100)aBased on IegreIIion fcxmula (natural loprithms), log SL-2.SI~+.908S1 (IogOL).

Lahontan system of northeastern California, pr0-vides some insights on the prehistoric sample fromStillwater. From seven gill net and minnow seinesets during June and July 1986, Davis (1986)recovered over 800 tui chub, ranging in size fromless than 20 mm to over 350 mm SL. He estimatedthe age of the individuals by counting annuli(growth rings) on the opercle (Scoppetone 1988)and calculated the mean body size (SL) by ageclass (Table 2). Thus, fISh of 22 mm SL wererecently hatched young of the year, those of91 mmSL had hatched the previous summer and ~ inthe 1+ age class, chub with lengths of 141 mmwere in the 2+ age class, and so forth (Table 2).

If we can assume that the chub growth patternsfor Eagle Lake and Stillwater Marsh are similar,then the body size data from the Stillwater archae-ological sample suggests that fish primarily diedtoward the end of their first and during their sec-ond year. Fish representing the 0+ and 3+ agegroups and older are absent. 1

If the 26CHI062 fish assemblage resulted pri-marily from natural mass death, we would expecta much lalger number of smaller young of theyear and more of the larger individuals. The smallmean body size and unimodal distribution of sizessuggests some selective mechanism of procure-ment. However, two 8na1ytic biases may affect thesize of the fossil open:le measured and thus thereconstructed body sizes of the fossil population:screen mesh size used and differential breakagemediated by size.

suggests that the use of 3.2-mm mesh has intro-duced slight bias to the reconstructed body sizedistribution in the direction of larger fish.

It is possible that opercles of smaller, youngertui chub are less sturdy and thus might sufferhigher attrition and fragmentation from destruc-tive processes than opercles from larger chub.3 Iflarger or smaller opercles tend to be broken withgreater than random frequency, then the sample ofmeasured opercles may not adequately reflect thefossil body size distribution. Over 700 cyprinidopercles were identified from the 26CHI062sample, whereas only 10 percent of these werecomplete enough to measure and include in thesize class reconstructions. To examine the poten-tial for bias introduced by differential breakage, Iestimated original opercle size of the brokenspecimens as follows. I selected five operclesfrom the modem tui chub sample with operclelengths measuring 5,10,15,20, and 25 mIn. Eachof the broken fossil opercles were then comparedto the opercles of known size and placed in themost appropriate size class category. For example,fossil opercles closest in size to the 5-mm operclewere placed in the OL category, 2.5-7.5 mm; fos-sil opercles closest to the 100mm opercle in size~ placed in the 7.5-mm to 12.5-mm OL cate-

Biases

Opercles of extremely small fish could have beenlost through the 3.2-mm mesh screen used duringexcavation. To examine the potential for suchloss. I retrieved and measured opercles from eightof the flotation samples that were screened with1.6-mm mesh. The size class distribution in the1.6-mm mesh sample differs somewhat from thatin the 3.2-mm sample (Figure 4). Mean fish sizein the 1.6-mm fraction (75.65 mm, S = 14.53) is

slightly, albeit significantly, smaller than themean body size in the larger mesh (89.14 mm SL,s = 18.61) (I = 2.95 > 1.0j(2).87 = 1.98). Although I

cannot exclude the possibility that opercles ofextremely small fish slipped through the 1.6-mmmesh, such loss seems unlikely? This comparison

[Vol. 81, No.4, 1886]AMERICAN ANTIQUITY706

Standard I

Figure 4. ComparilOn or reconstructed hi chub leagth fmmesh fraction 16CHIO61.

gory, and SO forth. Over 400 of the fossil opercleswere complete enough to assign to such size class

categories (Table 3).The frequency distn"bution of estimated body

size (Figure 3) is somewhat affected by bonebreakage, with larger specimens showing a greatertendency to be incomplete than smaller ones(Table 3). Over 32 pen:ent of the complete, mea-surable opercles are in the smallest size class,whereas only about 14 percent of the incompletespecimens are from this class. About 17 percent ofthe incomplete specimens are from fish larger than122 mm, whereas only about 5 percent of the com-plete, measurable opercles are from ftSh this large.Thus, the low frequencies of larger fish in Figure3 in part may reflect dle fact that only completeopercles were used to generate that graph. Size-dependent breakage, however, does not explain thelow frequencies of the smaller fish.

Screen mesh size and size-mediated breakagehave affected the body size reconstructions tosome extent. Use of 3.2-mm mesh biases bodysize in the direction of slightly larger fish, and dif-ferential breakage biases body size reconstructionin the direction of smaller fish. Neither of dlesebiases, however. challenges the basic structure of

Length (mm)j8eDda Rp~ted iD 3.2-mm mab (raedo. aDd l.6-mm

the tui chub mortality prattle. The scarcity ofyounger fishes and the relatively narrow sizerange does not suggest catastrophic mass or non-selective death but rather indicates a form of

selective mortality.

Surface Modification

Whether bone bas been burned, bears cut marks,and shows sign of digestion are other commonlyused indicators of depositional origin (e.g., Balme1980; Greenspan 1988; Lyman 1982; McGuire1980; Richter 1986; Schmitt and Sharp 1990;Shipman et aI. 1984; Smith 1985). Althoughburned bone in an archaeological site can resultfrom natural tIres (e.g., forest, grass fires), ifmul-tiple artifact classes and sediments do not showevidence ofbuming, it is reasonable to assume thatcultural agents are responsible for the burned bone.However, even if cultural agents are responsible forthe lire that burned bone, one would still need toconsider the posSloility that cultural burning post-dates the natUIal deposition of fish remains.

Widespread burning is not indicated at26CH 1 062. Thus, presumably, humans areresponsible for the IIres, which in turn burnedfaunal materials. Of the 9,010 fish remains from

REPORTS 707

Table 4. Frequency of Burned Specimensacross Body Parts, 26CHI062

Bwued Not ~Body Part ObI Bxp Obi £XVCranial 36 64.5 4.805 4.776.5Postcraniala 84 55.5 4.085 4.113.5

aElements posterior to basioccipital. including pectoral endpelvic f"m elements (chi-square - 26.59, p = .0000)

California) suggests that Smith's criteria do notapply to all coprolite bone. The Fish Slough Cavefish remains, representing the Cyprinidae andCatostomidae families, were examined by low-power magnification. Of a total of 87 fish speci-mens in the coprolite, only 40 (46 percent) arestained dark red. None of the specimens showsdigestive etching. Given the small sample size,my results obviously are teIltative, but they pointout the difficulty of using Smith's criteria todetermine whether digestive processes affectedfish remains.4

The 26CH I 062 fish specimens do not exhibitany of the attributes that Smith describes. None ofthe fish remains are stained or acid-etched norwere organic-rich materials adhering to any of thebony specimens. Although the fish remains arerelatively small, representing fish averaging about90 mm SL, this small size simply indicates thatsuch fish could have been consumed whole andnot that they were.

Raven and Schmitt (1991) suggest that the lackof evidence of staining and digestive etching onmost of the Stillwater fish remains indicates thatthe fISh were not ingested by humans, which in turncalls into question the cultural origin of the bones.As noted above, however, fISh bone that passesthrough the human digestive tract does not neces-sarily become stained or etched. I suggest that thevariables controlling staining and acid-etching aremore complex than previously conceived Thus, forthe present, the absence of staining and acid-etch-ing cannot be called upon to argue for a natural ori-gin of the Stillwater fish remains.

the site, 120 (1.3 percent) are burned. Relativelymore postcranial elements and fewer cranial ele-ments are burned than expected by chance events(Table 4; chi-square = 26.59, p = .0000), whichmay reflect differential processing of body parts.On the other hand, even if differential burningacross body parts reflects cultural processing perse, such practices were not common. given thesmall proportion of burned specimens. Burningdoes not provide compelling evidence for culturalorigins of the fish remains.

Cut marks, which would provide direct evi-dence for cultural processing, were not identifiedon any fish specimens. It is not known whetherthe absence of such marks results from limitedbutchering, limited contact of tools with bone,breakage tendencies of the bones themselves, orthe fact that human agency was not involved inbone deposition.

Evidence for human ingestion of fish car-casses including the bones also has been used toinfer cultural origin of fish remains in the westernGreat Basin (Greenspan 1988; Raven and Schmitt1991; Smith 1985). Archaeological evidence thatprehistoric inhabitants of the western Great Basinconsumed whole or partial tui chub comes frombones found in human coprolites (Follet 1967;Roust 1967) and chub remains inferred to be fromcoprolites based on digestive etching and otherattributes (Smith 1985). Based on his analysis offish remains from Hidden Cave, Smith(1985: 173) described several attributes indicatingingestion, digestion, and defecation: dark-stained,acid-etched bone; adhering matrix of organic-richsubstance that sometimes included bits of char-coal, hair, and other fibers; bones of uniformlysmall size; and overrepresentation of internalbones (e.g., vertebrae, pharyngeal teeth, andbasioccipitals). Smith does not indicate whetherthese attributes are specific to human coprolitesor whether fish remains from coprolites of non-human predators and scavengers could have simi-lar attributes.

Smith's criteria may characterize some fIShbones that have passed through the digestive tractof a human or some other vertebrate. However,my recent analysis of fish bones recovered from asingle human coprolite excavated by RobertBettinger from Fish Slough Cave (Owens Valley,

Body Part Frequency

Body part frequency is commonly used to iden-tify the agent(s) responsible for a particular fossilassemblage (e.g., Binford 1981; Binford andBertram 1977; Brain 1981; Stiner 1991). Such

AMERICAN ANTIQUITY [Vol. 81, No.4, 1998)708

studies generally assume that different agents(e.g., birds, humans. water) modify or destroyparticular classes of elements in distinctive ways.AJthough L~ (1984, 1985) and ~yson(1988) have cogently argued that the structuraldensity (g/cm3) of bones is the ultimate cause ofelement destruction in most contexts, proximatecauses may vary considerably across depositionalcontexts because of differences in the ways tapho-nomic agents procure and process carcasses.

I assume that while bone density plays a criti-cal role in structuring fossil fish assemblages,there should still be non-density-m~iated differ-ences in element representation between naturaland cultural tui chub deposits because of majordifferences in the nature of the agents generatingthe deposits (Lyman 1984). Carcasses resultingfrom natural mass death were presumablydeposited whole along the shoreline as the waterlevel declined. Such events were observed bybiologists working at Stillwater after the recent tuichub population crash (Rowe and Hoffman1987:Figure 10). AJthough scavenger birds andmammals no doubt assisted in die breakdown offiSh carcasses and modified and destroyed bones,in situ weadiering was probably the main destruc-tive agent. In these natural settings, the absence ofa dominant taphonomic agent focusing on soft tis-sue anatomy should result in higher correlationsbetween density and survivorship (Lyman 1984).

In cultural settings, human processing andconsumption of fish carcasses should signifi-cantly affect element representation (Butler1993). For example, ethnographic records of IlShprocessing in the Great Basin (Fowler 1986:88,1992:63) indicate that fish were often preservedby drying; both fresh and dried IlSh were furtherprepared by boiling or roasting in the ashes of thef1IeJ)lace (Fowler 1992:63; Kelly 1932:97). Driedfish often were pulverized and added to otherdishes (Fowler 1986:88). At Stillwater Marsh.larger fish (Tahoe sucker and large tui chub) weredried on racks; although diey were split length-wise to promote drying, they were not filleted(Fowler 1992:63). At Pyramid Lake, only the Itl-lets obtained from cui ui (Chasmistes cujus) werepreserved (Fowler 1986; Stewart 1941).

Human consumption of whole or parts of fiShmay also affect element representation. The

ethnographic record for Stillwater inhabitants,suggests that

small fish (usually dried) were boiled whole insoups. . . . Their bones softened and they couldbe eaten whole if they had Dot disintegratedduring boiling. . . . Small fresh fISh were bakedin the ashes in packets made by placing thembetween two layers of cattail leaves . . . . 11risprocess kept the fish clean and also softenedthe bones so they could be eaten along with themeat. [Fowler 1992:63]

As mentioned above, the presence of tui chubbones in prehistoric coprolites indicates that somefish were consumed along with their bones.Based on analyses of 30 human coprolites fromLovelock Cave, FoUet (1967) identified 474 tuichub Pharyngeals representing a minimum of 298IlSh. FoUet suggests that the fish represented inthese coprolites come from fish 38 to 139 mmlong, although he does not describe how such sizeestimates were made. He also argues that thewell-preserved condition of the pharyngealsreflects that at least the heads of these fish hadbeen eaten whole. Because pharyngeals alonewere identified, the Lovelock data do not provideany estimates on overall element survivorship incoprolites. FoUet's report does provide clear evi-dence that at least some parts, the head includingthe bones, of small tui chub were consumed.

With a few exceptions (Fowler 1986), ethno-graphic accounts do not specify whether heads andtrunks were treated differently during processing.Previous ~rk on salmonid taphonomy suggeststhat differences in the distribution of meat andbones ~ the head and trunk in part struc-tured the way these body parts were butchered andcooked, which in turn affected element survivor-ship of the ~ body parts (Butler 1993).

If the StiUwater IISh fauna results from naturaldeath and deposition, element survivorshipshould show a high correlation to bone densitygiven the absence of a dominant taphonomicagent (e.g., humans) focused on soft tissueanatomy. If the fish fauna was generated primar-ily by cultural agents, element survivorshipshould show a much lower correlation to bonedensity given the variety of intervening factors,including cooking, butchering, consumption, anddisposal, which may affect bones of the head andtrunk in variable ways.

REPORTS 709

~MAU

Figure 5. Element survivorship (percent MAU) of Cyprinformes remains, 26CHI062 (baslo, basioccipital; hyom,hyomandibula; opere, opercle; phary, pharyngeal; clelth, clelthrum).

Of course, chub assemblages originally pro-duced by cultural or natural agents could be mod-ified further by water sorting. Discrete elementsor articulated parts could be moved away from theoriginal site of deposition. The floodwaters of the1980s that exposed the Stillwater archaeologicalsites probably moved materials found in thosesites (Dansie 1987), and past flooding episodescould have sorted exposed bone accumulations.Such transport would affect element representa-tion and potentially obscure the distinctionsdescribed above. Controlled flume studies andfield observations of mammal bone transport bywater (Behrensmeyer 1975; Voorhies 1969) andfield observations of natural fish bone accumula-tions on lakeshores exposed to differing amountsof wave action (Stewart 1991) suggest that watereffectively sorts bones by their density, size, andshape. For example, Stewart's (1991) study ofmodern natural fish bone accumulations on theshore of Lake Turkana, Africa, found that collec-tion locales exposed to the greatest wave activityhad a disproportionately high number of denseelements. In the context of this study, a high cor-relation between bone density and element sur-vivorship could just as easily reflect natural deathand deposition (no aqueous transport) as culturaluse and deposition (followed by water sorting).

Interpretations of the bone density-element sur-vivorship relationships discussed below will bearthese issues in mind.

Element survivorships for the 26CHI062assemblage are based on identifiable remainsfrom both the cyprinid and catostomid families ofthe Cypriniformes order. Because most of the ver-tebrae could not be identified to family, the moreinclusive taxonomic category had to be used inorder for the postcranial elements to be includedin analysis. Given that over 96 percent of theassemblage is cyprinid, and virtually all of theseremains are from tui chub, body part frequenciespresented here primarily reflect that of tui chub.

Element survivorships were detennined byfIrst calculating the minimum animal unit (MAU,Binford 1978); when the entire site was treated asan aggregate, the fish renlains provided a MAU of520, based on the pharyngeal. I based MAU cal-culations on the minimum number of elements(MNE) (Bunn 1982), which selects the best rep-resented section of each element and counts thenumber of times it occurs in a given aggregate(Grayson 1988). The MNE for vertebrae werebased on the presence of 50 percent or more of thecentrum. Element survivorships (percent MAU)were calculated by comparing the number of ele-ments expected relative to the number observed

AMERICAN ANTIQUITY [Vol. 11, No.4, 1891]710

Table 6. Comparison of Rank Order Element Density(l/cm3) with Element Survivorship (Pen:ent MAU).

Table S. Frequency (MNE) and Survivorship (PercentMAU) ofCyprinifonnes Skeletal Elements, 26CHI062

EIeaaeat WNE %MAIJCranium

Articular 43 4.13Autopterotic 115 11.06Autosphenotic 37 3.55Basioccipital 309 59.42CeI'atohyaI lIS 17.79Dentary 78 7.SOEntoptery8oid 51 4.90Epihyal 110 10.58Epiotic 5 .48Exoccipital 132 12.69Frontal 175 16.83Glossohyal I .19Hyomandibula 369 35.48Maxilla 14 1.34Metaptery8oid 5 .48Opercle 927 89.13Palatine 12 1.15Parasphenoid 41 7.88Pharyngeal 1,040 100.00Premaxilla I .09Preopercle 58 S.58Prootic 55 S.29Quadrate 120 11.54Supraethmoid 15 2.88Supraoccipital 34 6.54Urohyal 54 10.38Vomer 15 2.88

FinsCoracoid 39Cleithrum 980Scapula 39Supracleithrum 22Basipteryaium 63

VertebraeF lISt vertebra 970d1er vertebra 2,522

(Binford 1984; Grayson 1988). For example, witha MAU value of 520, we would expect 1,040 ofeach paired element in the skeleton. If only 250dentaries or maxillae are present, their elementsurvivorship would be 250/1,040, or 24 percent

As shown in Figure 5 and Table 5, element sur-vivorship is extremely variable; several cranialand paired rm elements show moderate to veryhigh frequencies (basioccipital, opercle, pharyn-geal, cleithrum), whereas vertebrae and numerousother elements of the cranium and paired rIDS arescarce.

To examine the role of bone density in struc-turing body part representation in prehistoric fau-

3.7594.23

3.752.116.06

18.6514.26

nal assemblages, Lyman (1984, 1985; Lyman etal. 1992), Kreutzer (1992), ~d Butler andChatters (1994) have measured the bone mineralcontent (g) of modern skeletal elements fromartiodactyls, rodents, and saImonids by photonabsorptiometl"y. Combined with estimates of bonevolume, their values provide a valid ordinal mea-sure of volume density (g/cm3)(sensu Lyman1984), which can be compared with rank-orderelement survivorships in prehistoric faunas todetennine the degree to which bone density struc-tures body part representation.

Although bone density data are not availablefor tui chub, Butler and Chatters have recentlydetermined element densities for another speciesin the Cypriniformes order (Catostomusmacrocheilus, family Catostomidae). I assumethat density values for this taxon approximatethose for the tui chub. Volume densities of 14 ele-ments were deterDlined; representative cranial,paired 1m, and vertebral elements from liveCatostomus skeletons ~ Studied.

As shown in Table 6, the rank orders of bonedensity and element survivorship are very differ-ent; the correlation is low and not significant(Spearman's rho = -.330, .50 > p > .20). The moststriking discrepancy occurs with the opercle andcleithrum, which have close to the highest sur-vivorship and yet the lowest densities. Similarly,the very high abundance of the pharyngeal is notpredicted from its only moderately high bone den-sity. The low correlation argues against the role ofwater action in modifying the assemblage; at least

REPORTS

Table 7. Comparison of Survivorship (Percent MAU)between 3.2-mm and 1.6-mm Fraction,

Selected Elements, 26CH I 062.

water did not differentially deposit or remove ele-ments of particular density values.

As argued above, a low correlation betweendensity and survivorship is expected if culturalagents in some way modified the faunal materi-als. The relative scarcity of vertebrae, which haveclose to the highest densities and yet survivorshipof less than 20 percent, may result from differen-tial processing of trunks versus heads. For exam-ple, if trunks were consumed and heads were not,or if trunks were cooked with the heads detached,digestion and cooking may lead to greater loss ofvertebrae relative to some bones of the head.Given the small chub body size, the low correla-tion presumably is not explained by differentialtransport of carcass parts to the site (e.g., Lyman1985), although the scarcity of vertebrae mayindicate that trunks and heads were consumed anddeposited in different locations. Because we lackdata on the specific effects of cooking and diges-tion on tui chub bone preservation, other scenar-ios are equally plausible.

Additional support for the argument that morevertebrae are expected in a natural deposit ofcyprinid remains is provided by Stewart's (1991)recent analysis of natural fish bone accumulationsfrom Lake Turkana. Based on analysis of cyprini-form body part representation, Stewart suggeststhat cranial elements survive poorly, whereas ver-tebral elements survive very well. Although morespecific comparisons between Lake Turkana andStillwater are not possible because of differencesin data organization and reporting, Stewart'sresults certainly indicate that vertebrae are rela-tively abundant in natural lakeside settings, whichis contrary to that found at Stillwater.

Of course identifiability and recovery prac-tices (Lyman and O'Brien 1987) also may affectelement representation. Given the small bodysizes represented, no doubt the relative infre-quency of some elements is explained by the useof 3.2-mm mesh screens during excavation. Toexamine whether 3.2-mm recovery biased verte-brae representation in particular, I assessed ele-ment survivorships for selected elements within asingle flotation sample. I included two of the mostcommon cranial elements in the 3.2-mm fraction,the opercle and pharyngeal, and the vertebrae inanalysis. These elements provide 436 MNE; ele-

ment survivorship was based on the opercle,which produced a MAU value of 31 (Table 7).Survivorship of vertebrae in the flotation sampleis almost double that found in the 3.2-mm mesh,suggesting that a disproportionate number of ver-tebraeare lost through 3.2-mm mesh (at least rel-ative to opercles and Pharyngeals) (Table 7).However, survivorship of most vertebrae in the1.6-mm mesh is still much lower than 50 percent,suggesting that recovery bias is not a major factorin accbunting for vertebrae scarcity.

In SUfi1, tui chub element survivorship is highlyvariable; some fragile elements of relatively lowdensity (e.g., opercle, pharyngeal) are extremelyabundant, whereas many other elements of rela-tively high density are scarce. Most of this varia-tion is not explained by the intrinsic property ofbone density. Although the scarcity of some ele-ments may be explained by archaeological recov-ery practices, recovery bias does not completelyaccount for the scarcity of vertebrae. Rather, ver-tebra scarcity is best explained by the actions ofagents that differentially treat parts of the fish.Cultural agents seem to be the most likely candi-date for this intervention.

Natural vs. Cultural Origin of StillwaterFish Remains

Archaeologists working throughout the world arebeginning to recognize that before we can exam-ine evidence of human fishing strategies, we mustbe able to distinguish fish bone resulting fromnatural death and deposition from that depositedby cultural agents. Noe-Nygaard (1983), forexample, examined the contribution of naturalfish kills in 3,OOO-year-old lakeside deposits incentral Denmark. She posits that the fish remainsresult from cultural procurement based on the

712 AMERICAN ANTIQUITY [Vol. 61, No.4, 1996]

face characteristics are not necessarily usefulindicators of depositional origins. The absence ofsuch evidence at 26CHI062, then. does notundermine the argument for cultural origin of thefish bone.

A variety of additional research would greatlyincrease our understanding of the contribution ofnatural fish kills in marsh sites. For example, con-troIled coIlection and analysis of fish remainsdeposited after the recent flooding in Sti1Iwaterwould allow detailed characterization of IlShassemblages resulting from natural death anddeposition (e.g., Butler 1990, 1993; Stewart1991). Also, as suggested by Raven and Schmitt(1991), test excavation of marsh deposits free ofprehistoric cultural materials would allow us todetermine the likelihood of the deposition andpreservation of natural fish bone outside archaeo-logical contexts; if such deposits are located, theywould help characterize natural fish bone assem-blages.

narrow size range of individuals represented anda similar season of capture of the fishes and mam-mals present. On the other hand, Richter (1986)argues that the extremely small fish remains froma coastal Danish Neolithic site were naturallydeposited, based on collagen analysis, showingthat the fiSh remains were not heated. Otherinvestigators have used attributes such as bonediscoloration and fragmentation, spatial distribu-tion of faunal materials and their co-occurrencewith unambiguous artifacts, as well as hominiddietary preferences to argue whether or not peoplewere primarily responsible for a fish bone accu-mulation (Butler 1990, 1993; Stewart 1991; VanNeer and Muniz 1993).

Similarly, researchers in the Great Basin ofwestern North America have begun to share thisconcern for fish bone origin in lacustrine, fluvial,and cave deposits and have started to developapproaches to assess taphonomic agents (e.g.,Aikens and Greenspan 1988; Greenspan 1988;Smith 1985). My analysis of primarily tui chubremains from one site in the Stillwater Marsh sug~gests that humans are primarily, if not entirely,responsible for fish deposition at the site.

A relatively narrow size range of tui chub arerepresented in the bone bed, which is more sug-gestive of selective procurement than catastrophicmass death. In particular, the chub body size datasuggest that mass harvesting techniques, involv-ing the use of nets of small mesh size, were usedto catch the fish. Hamley (1975) and other fish-eries researchers (e.g., Gulland and Harding1961; Jensen 1990; McCombie and Fry 1960)have shown that gill nets of specified mesh sizeare very selective of fish of partic~ar lengths.Furthermore, a disproportionately low frequencyof vertebrae are present, which indicates thatsome intervening factors, such as cultural pro-cessing, may have affected body part representa-tion. Bone surface modification has not helpedsort out agents responsible for the fish remains.Importantly, the absence of evidence for humandigestion and processing on the bone surfaces hasbeen used to argue for a natural origin of fishremains from other Stillwater sites (e.g., Ravenand Schmitt 1991). However, given our limitedunderstanding of how fish bone is modified dur-ing digestion and cultural processing, bone sur-

Small Tui Chub-an Important MarshResource

As described in the introduction, the debate overthe importance of marsh resources to prehistoriceconomies centers, in part, on the relative value ofmarsh versus upland resources. Livingston (1991)has recently marshaled archaeological avian fau-nal data and life history characteristics of water-fowl to argue that birds were an importantsubsiStence component to peoples of the westernGreat Basin over the past 5,000 years. Drawing onexperiments with fISh procurement and process-ing and nutritional analyses, Raymond and Sobel(1990) and Lindstrom (1992) argue that fishwould have provided rates of return (calories/hr)as high or higher than terrestrial resources (e.g.,Simms 1987), depending on fish body size, meth-ods of processing, and other variables. In particu-lar, Raymond and Sobel (1990) found that smallschooling tui chub caught through mass harvest-ing (nets) had higher return rates than larger chub,because larger fish required more processingtime.s

Prehistoric evidence for the use of small tuichub is found in several western Great Basin caveassemblages, where Raymond and Sobel (1990)measured the lengths of whole desiccated tui chub

REPORTS

Basin. The results discussed here should assistwith resolving some of the taphonomic ambiguitysurrounding marsh assemblages. In addition, theregression model provided will allow reconstruc-tions of fish body size, which will assist witttinterpretations of procurement strategies as well asevaluation of resource intensification models (e.g.,Broughton 1994) and diet breadth models (e.g.,Broughton and Grayson 1993; Madsen 1993).Such models make predictions about foragingefficiencY and prey selection, factors at least inpart determined by prey body size (e.g., that preybody size is positively correlated with prey rank:the larger the animal, the higher its rank and thegreater its return rate). Importantly for tui chub,whether body size per se is an appropriate measureof prey rank may be questionable when mass har-vesting techniques are emplqyed in capture. Withincreased chronological and taphonomic controlsover fish assemblages from wetland and cave sites,we will be able to more clearly document the rela-tive importance of fiSh of various sizes to prehis-toric subsistence and how that use may have variedover time.

Acknowledgments. Analysis of 26CHl062 fish bones wassupported by National Science Foundation Grant BNS8704094 to R. Kelly. M. Sevon (Nevada State Department ofWildlife) collected the modern tui chub carcasses used inregression analyses. A. Fountain, L. Goldstein, D. Grayson,R. Lyman, D. Madsen, D. Schmitt, and two anonymousreviewers provided very useful comments on the manuscriptE. Rees translated the abstract into Spanish. L. Bergman, R.Bettinger, A. Fountain, K. Gobalet, J. Hawkins, R. Kelly, S.Lindstrom, P. Moyle, D. Nelson, C. Raven, A. Raymond, D.Rhode, N. Sharp, P. Schultz, G. Scoppetone, G. Smith, andL. Todd provided valuable assistance on various aspects ofthe work. I greatly appreciate the help of all these people.

from cultural features. Chub in Stick Cave rangedbetween 75 and 122 mm in total length (mean 99mm, n = 79), and Lovelock Cave tui chub ranged

in total length between 43 and 130 mm (mean 75mm, n = 51). Body size of the tui chub at Humbolt

Cave were larger and more variable, rangingbetween 90 and 220 mm (mean 148 mm, n = 54).

Based on the relatively high return rates of smallchub and the archaeological evidence for smallchub use, the authors suggest that prehistoric fish-ers specifically targeted smaller chub in the sub-sistence quest.

Analysis of fish remains from Stillwater cer-tainly indicates that tui chub, particularly smalltui chub, were a major food resource to inhabi-tants of the Stillwater Marsh. The fish remains inthe 26CHI062 assemblage represent over 85 per-cent of the identified vertebrate faunal remains(Livingston 1991; Schmitt and Sharp 1990). Oursampling of the faunal assemblage suggests thatover 290,000 fish remains were recovered duringexcavation; a small (but unknown) fraction of thesite was excavated. Thus, the sampling and analy-sis of the fish remains suggest that literally hun-dreds of thousands of small tui t:.hub reside in thearchaeological deposit. Such abundance supportsthe position that small tui chub were a keyresource to prehistoric marsh inhabitants.

Many questions remain. How pervasive is thistrend across environments and over time? Weresmall chub targeted by the earliest marsh andlakeside residents of the western Great Basin orhas there been an increase in the use of small chubover time? Does the use of small chub represent adecline in foraging efficiency, an example ofresource intensification, as Broughton (1994) hasdocumented for the late Holocene of centralCalifornia based on his study of mammalian andfish faunal assemblages? Did fish resource usevary across aquatic environments in the GreatBasin (e.g., deep-water lakes, perennial streams,and marshes), which support different species offishes with different life histories and growthcharacteristics? How would this variation con-strain human subsistence scheduling and mobility

patterns?Answering these questions will require addi-

tional taphonomic analysis of faunal assemblagesfrom other sites in Stillwater and the western Great

References CitedAikens, C. M., D. K. Grayson, and P. J. Mehringer

1982 Final Project Report to the National ScienceFoundation on the Steens Mountain Prehistory Project.Department of Anthropology, University of Oregon,

Eugene.AIKens, C. M., and R. L. Greenspan

1988 Ancient Lakeside Culture in the Northern GreatBasin: Malheur Lake, Oregon. Journal of Californiaand Great Basin Anthropology 10:32-61.

Balme, J.1980 An Analysis of Charred Bone from Devil's Lair,

Western Australia. Archaeology and PhysicalAnthropology in Oceania 15:81-85.

Beck, C.1984 Steens Mountain Surface ArcJIae%gy: 11Ie Sites.

AMERICAN ANTIQUITY [Vol. 61, No.4, 1996]

Ph.D. dissertation, University of Washington, Seattle.University Microfilms, Ann Arbor.

Behrensmeyer, A. K.1975 The Taphonomy and Paleoecology of Plio-

Pleistocene Vertebrate Assemblages East of LakeRudolf, Kenya. Bulletin of the Museum of ComparativeZoology 146:473-578.

Benson, L. V:, and R. S. Thompson1987 Lake-level Variation in the Lahontan Basin for the

Past 50,000 Years. Quaternary Research 28:69-85.Bettinger, R. L.

1975 The Surface Archaeology of Owens Valley, EasternCalifornia: Prehistoric Man-Land Relationships in theGreat Basin. Unpublished Ph.D. dissertation,Department of Anthropology, University of California,Riverside.

1993 Doing Great Basin Archaeology Recently: CopingWith Variability. Jounral of Archaeological Research1:43-66.

Binford, L. R.1978 Nunamiut Ethnoarchaeology. Academic Press, New

York.1981 Bones: Ancient Men and Modern Myths. Academic

Press, New York.1983 In Pumdt of the Past. Thames and Hudson, New

York.1984 Faunal Remains from Klasies River Mouth.

Academic Press, New York.Binford, L. R., and J. B. Bertram

1977 Bone Frequencies-and Attritional Processes. InFor Theory Building in Archaeology, edited by L. R.Binford, pp. 77-153. Academic Press, New York.

Brain, C. K.1981 The Hunters or the Hunted? An Introduction to

African Cave Taphonomy. University of Chicago Press,Chicago.

Broughton, J. M.1994 Resource Intensification in Central California:

Archaeological Vertebrate Evidence from theSacramento Valley. Journal of Archaeological Science21:501-514.

Broughton, J. M., and D. K. Grayson1993 Diet Breadth, Adaptive Change, and the White

Mountains Faunas. Journal of Archaeological Science20:331-336.

Bunn, H. T.1982 Meat-Eating and Human Evolution: Studies on the

Diet and Subsistence Patterns of Plio-PleistoceneHominids in East Africa. Unpublished Ph.D. disserta-tion, Department of Anthropology, University ofCalifornia, Berkeley.

Butler, v: L.1990 Distinguishing NatlU-al from CultlU-al Salmonid

Deposits in Pacific Northwest North America. Ph.D. dis-sertation, University of Washington, Seattle. UniversityMicrofilms, Ann Arbor.

1993 Natural Versus Cu\tural Salmonid Remains: Originof The Dalles Roadcut Bones, Columbia River, Oregon,U.S.A. Journal of Archaeological Science 20: 1-24.

Butler, v: L., and J. C. Chatters1994 The Role of Bone Density in Structuring Prehistoric

Salmon Bone Assemblages. Journal of ArchaeologicalScience 21:413-424.

Cailliet, G. M., M. S. Love, and A. W Ebeling1986 Fishes: A Field and Laboratory Manual on Their

StrIlcture, Identification. and Nahlral History.Wadsworth, Belmont, California.

Carlander, K. D.1969 Handbook of Freshwater Fishery Biology, vol. 1.

Iowa State University Press, Ames.Casteel, R. VI:

1976 Fish Remains in Archaeology. Academic Press, NewYork..

Dansie, A. J.1987 The Stillwater Animal Bones. In Final Report on the

Excavations in the Stillwater Marsh ArchaeologicalDistrict, Nevada, edited by D. R. Tuohy, A. J. Dansie,and M. B. Haldeman, pp. 257-311. Nevada StateMuseum Archaeological Services, Carson City.

Davis, J. O.1982 Bits and Pieces: The Last 35,000 years in the

Lahontan Area. In Man and Environment in the GreatBasin, edited by D. B. Madsen and J. F. O'Connell, pp.53-75. Papers No.2. Society for AmericanArchaeology, Washington, D.C.

Davis, L. E.1986 Some Notes on the Age, Growth, and Reproduction

of 1\ri Chubs (Gila bicolor) from Eagle Lake,California. Manuscript on' file, AnthropologyDepartment, Portland State University, Portland,

Oregon.Elston, R. G., K. L. Katzer, and D. R. Currey

1988 Chronological Swnmary. In Preliminary Investi-gations in Stillwater Marsh: Human Prehistory andGeoarchaeology, edited by C. Raven and R. G. Elston,pp. 373-384. Cultural Resource Series No. I. U.S. Fishand Wildlife Service, Portland, Oregon.

FoUet, VI: L1967 Fish Remains from Coprolites and Midden Deposits

of Lovelock Cave, Churchill County, Nevada.

University of California Archaeological Survey Report70:93-116.

Fowler, C. S.1986 Subsistence. In Great Basin, edited by W. L.

D' Azevedo, pp. 64-97. Handbook of North AmericanIndians, vol. 11, W. C. Sturtevant, general editor.Smithsonian Institution, Washington, D.C.

1992 In the Shadow of Fox Peak: An Ethnography of theCattail-Eater Northern Paiute People of StillwaterMarsh. Cultural Resource Series No.5. U.S. Fish andWildlife Service, Portland, Oregon.

Frison, G. C.1978 Prehistoric Hunters of the High Plains. Academic

Press, New York.1982 Bison Dentition Studies. In The Agate Basin Site: A

Record of Paleoindian Occupation of the NorthwesternHigh Plains, edited by G. C. Frison and D. J. Stanford,pp. 240-260. Academic Press, New York..

Grayson, D. K.1988 Danger Cave, Last Supper Cave, and Hanging Rock

Shelter: The Faunas. Anthropological Papers Vol. 66, Pt.I. American Museum of Natural History, New York..

Greenspan, R. L.1988 Fish remains. In Preliminary Investigations in

Stillwater Marsh: Human Prehistory andGeoarchaeology, edited by C. Raven and R. G. Elston,pp. 312-326. Cultural Resource Series No.1. U.S. Fishand Wildlife Service, Portland, Oregon.

Gulland, J. A., and D. Harding1961 The Selection of Clarias mossambicus (Peters) bv

REPORTS

Nylon Gill Nets. Journal du Conseil PermanentInternational pour I' ExploraJion de la Mer 26:215-222.

Hamley, J. M.1915 Review of Gillnet Selectivity. Journal of the

FISheries Research Board of Canada 32:1943-1969.Heizer, R. F., and A. D. Krieger

1956 The Archaeology of Humboldt Cave ChurchillCounty, Nevada. University of California Publicationsin American Archaeology and Ethnology, Berkeley41(1):1-190.

Heizer, R. F., and L. K. Napton1910 Archaeology and the Prehistoric Great Basin

Lacustrine Subsistence Regime as Seen from LovelockCave, Nevada. Archaeological Research FacilityContributions 10. University of California, Berkeley.

Hoffecker, J. F., G. Baryshnikov, and O. Potapova1991 Vertebrate Remains from the Mousterian Site of

II'skaya I (Northern Caucasus, U.S.S.R.): New Analysisand Interpretation. Journal of Archaeological Science18:113-141.

Jensen, W.1990 Comparing Fish Catches Taken With Gill Nets of

Different Combinations of Mesh Sizes. Journal of FIShBiology 31: 99-104.

Jones, A. K. G.1984 Some Effects of the Mammalian Digestive System

on Fish Bones. In 2nd Fish Osteoarchaeology Meeting,edited by N. Desse-Berset, pp. 61-16. Centre deRecherches Archeologiques Notes et MonographiesTechniques No. 16. Centre National de la RechercheScientifique, Paris.

1986 Fish Bone Survival in the Digestive Systems of thePig, Dog, and Man: Some Experiments. In FISh andArchaeology: Studies in Osteometry. Taphonomy,Seasonality, and FIShing Methods, edited by D. C.Brinkhuizen and A. T. Clason, pp. 53-61. BARInternational Series 294. British Archaeological Reports,Oxford.

Jones, G. T.1984 Prehistoric Land Use in the Steens Mountain Area,

Southeastern Oregon. Ph.D. dissertation, University ofWashington, Seattle. University MicroIllms, Ann Arbor.

Kelty, I. T.1932 Ethnography of the Surprise Valley Paiute.

Publications in American Archaeology and EthnologyVol. 31, No.3. University ofCatifornia, Berkeley.

Kelty, R. L.1985 Hunter-Gatherer Mobility and Sedentism: A Great

Basin Study. Unpublished Ph.D. dissertation,Department of Anthropology, University of Michigan,Ann Arbor.

1988 Archaeological Context In Preliminary Investigations inStillwater Marsh: Human Prehistory and Geoarchae-ology, edited by C. Raven and R. G. Elston, pp. 5-20.Cultural Resource Series No.1. U.S. Fish and WildlifeService, Portland, Oregon.

Klein, R. G., and K. Cruz-Un"be1984 77Ie Analysis of Animal Bones from Archaeological

Sites. University of Chicago Press, Chicago.Kreutzer, L. A.

1992 Bison and Deer Bone Mineral Densities: ~and Implications for die Interpretation of ArchaeologicalFaunas. Journal o/Archaeologicai Science 19:211-294.

Kurten, B.1953 On the Variation and Population Dynamics of Fossil and

Recent Mammal Populations. Acta Zoologica Fennica76:1-122.

Lindstrom, S.1992 Great Basin Fisherfolk: Optimal Diet Breadth

Modeling the TnICkee River Aboriginal SubsistenceFIShery. Unpublished Ph.D dissertation, Deparbnent ofAnthropology, University of California, Davis.

Livingston, S. D.1988a The Avian and Mammalian Faunas from Lovelock

Cave and the Humboldt LaMbed Site. Ph.D. disserta-tion, University of Washington, Seattle. UniversityMicrofihns, Ann Arbor.

1988b Avian Fauna. In Preliminary Investigations inStillwater Marsh: Human Prehistory andGeoarcllaeology, edited by C. Raven and R. G. Elston,pp. 292-311. Cultural Resource Series No.1. U.S. Fishand Wildlife Service, Portland, Oregon.

1991 Aboriginal Utilization of Birds in the Western GreatBasin. In Beamers, Bobwhites, and Blue-points:Tributes to the Career of Paul ~ Parmalee, edited by J.R. Purdue, W. E. Klippel, and B. W. Styles, pp.341-357. Scientific Papers XXIlI. Illinois StateMuseum, Springfield.

Loud, L. L., and M. R. Harrington1929 Lovelock Cave. Publications in American

Archaeology and Ethnology Vol. 25, No. I. Universityof California, Berkeley.

Lyman, R. L.1982 Archaeofaunas and Subsistence Studies. In

Advances in Archaeological Method and Theory, vol. 5,edited by M. B. Schiffer, pp. 331-392. Academic Press,New York.

1984 Bone Density and Differential Survivorship ofFossil Classes. Journal of Anthropological Archaeology3:259-299.

1985 Bone Frequencies: Differential Transport, In SituDestruction, and the MGUl. Journal of ArchaeologicalScience 12:221-236.

1991 Subsistence Change and Pinniped Hunting. InHuman Predators and Prey Mortality, edited by M. C.Stiner, pp. 187-199, Westview Press, Boulder,Colorado.

Lyman, R. L., L. E. Houghton, and A. L. Chambers1992 The Effect of Structural Density on Marmot Skeletal

Part Representation in Archaeological Sites. Journal ofArchaeological Science 19:1-17.

Lyman, R. L., and M. J. O'Brien1987 Plow-Zone Zooarchaeology: Fragmentation and

Identifiability. Journal of Freld Archaeology 14:493-498.McCombie, A M., and F. E. J. Fry

1960 Selectivity of Gill Nets for Lake Whitefish,Coregonus clupeafonnis. Transactions of the AmericanFisheries Society 89:176-184.

MacFadden, B. J.1992 Fossil Horses: Systematics, Paleobiology, and

Evolution of the Family Equidae. Cambridge UniversityPress, Cambridge.

McGuire, K. R.1980 Cave Sites, Faunal Analysis, and Big-Game Hunters

of the Great Basin. Quaternary Research 14:263-268.Madsen, D. B.

1979 The Fremont and Sevier: Defming PrehistoricAgriculture North of the Anasazi. American Antiquity44:711-722.

1988 The Prehistoric Use of Great Basin Marshes. In

718 AMERICAN ANTIQUITY [Vol. 61, No.4, 1996]

Middle Neolithic Site, Kainsbakke, East Jutland.Jollnlal of Danish Archaeology 5:116-124.

Roust, N. L.1967 Preliminary Examination of Prehistoric Human

Coprolites from Four Western Nevada Caves. Ullivenity

of Callfl;rnia Archaeological Survey Report. Berkeley70:49-88.

Rowe, T. G., and R. J. Hoffman1987 Wildlife Kills in tile Carson Sink, Western Nevada,

Winter 1986-1987. Water-Supply Paper 2350. U.S.

Geological Survey, Reston, Virginia.Russell, I. C.

1885 Geological History of Lake Lahontan, a QuaternaryLake of Northwestern Nevada. Monograph No. 11. U.S.Geological Survey, Washington, D.C.

Scbmitt, D. N.19888 Some Observations on Vertebrate Taphonomy and

Site Formational Processes in Stillwater Marsh. InPreliminary Investigations in Stillwater Marsh: HumanPrehistory and Geoarchaeology, edited by C. Ravenand R. G. Elston, pp. 359-371. Cultural ResourceSeries No. I. U.S. Fish and Wildlife Service, Portland,

Oregon.1988b Manunalian Fauna. In Prelimhlary Investigations

ill Stillwater Marsh: Human Prehistory andGeoarchaeology, edited by C. Raven and R. G. Elston,pp. 262-290. Cultural Resource Series No. I. U.S. Fishand Wildlife Service, Portland, Oregon.

Schmitt, D. N., and N. D. Sharp1990 Manunals in the Marsh: ZooarchaeologicalAnalysis

of Six Sites in the Stillwater Wildlife Refuge, WesternNevada. In Wetland Adaptations in the Great Basin,edited by J. C. Janetski and D. B. Madsen, pp. 75-95.Museum of Peoples and Cultures Occasional Papers No.1. Brigham Young University, Provo, Utah.

Scoppetone, G. G.1988 Growth and Longevity of the Cui-ui and Longevity

of Other Catostomids and Cyprinids in Western NorthAmerica. Transactions of the American FisheriesSociety 117:301-307.

Shipman, P., G. Foster, and M. Schoeninger1984 Bumt Bones and Teeth: An Experimental Study of

Color, Morphology, Crystal Structure, and Shrinkage.Journal of Archaeological Science 11 :307-325.

Simms, S. R.1987 BehaviolUl Ecology and Hunter-Gatherer Foraging:

An Example from the Great Basin. BAR InternationalSeries 381. British Archaeological Reports, Oxford.

Smith, G. R.1985 Paleontology of Hidden Cave: Fish. In The

Archaeology of Hidden Cave, Nevada, edited by D. H.Thomas, pp. 171-178. Anthropological Papers Vol. 61,Pt.l. American Museum of Natural History, New York.

Speth, J. D.1983 Bison Kills and Bone Counts: Decision Ma.b"ng by

Ancient Hunters. University of Chicago Press, Chicago.Stewart, K. L.

1991 Modem Fishbone Assemblages at Lake Turkana,Kenya: a Methodology to Aid in Recognition ofHominid Fish Utilization. Journal of ArchaeologicalScience 18:579-603.

Stewart, O. C.1941 Culture Element Distributions, XIll: Nevada

Shoshone. University of California AnthropologicalRecords, Berkeley 4(3).

Preliminary Investigations in Stillwater Marsh: HumanPrehistory and Geoarchaeology, edited by C. Ravenand R. G. Elston, pp. 414-418. Cultural ResourceSeries No. I. U.S. Fish and Wildlife Service, Portland,Oregon.

1993 Testing Diet Breadth Models: Examining AdaptiveChange in the Late Prehistoric Great Basin. Joumal ofArehaeological Science 20:321-329.

Morrison, R. B.1964 Lake Lahontan: Geology of Southern Carson

Desert, Nevada. Professional Paper 401. U.S.Geological Survey, Reston, Virginia.

Moyle, P. B., and J. J. Cech1982 FIShes: An -Introduction to Ichthyology. Prentice-

Hall, Englewood Cliffs, New Jersey.Noe-Nygaard, N.

1983 The Importance of Aquatic Resources to MesolithicMan at Inland Sites in Denmark. In Animals andArchaeology: Shell Middens, FIShes, and Birds, editedby C. Grigson and J. Clutton-Brock, pp. 125-142. BARInternational Series 183. British ArchaeologicalReports, Oxford.

Raven, C., and R. G. Elston (editors)1988 Preliminary Investigations in Stillwater Marsh:

Human Prehistory and Geoarchaeology. CulturalResource Series No. I. U.S. Fish and Wildlife Service,Portland, Oregon.

1991 Looking for the Marsh: Past, Present, and FutureArchaeological Research in the Carson Desert.Submitted to U.S. Fish and Wildlife Service, ContractNo. 14-16-0001-81091(TA). Copies available from theU.S. Fish and Wildlife Service, Portland, Oregon.

Raven, C., and D. N. Schmitt1991 Subsistence Technologies and Their Archaeological

Expression. In Looking for the Marsh: Past, Present,and Future Archaeological Research in the CarsonDesert, edited by C. Raven and R. G. Elston, pp. 52-64.Submitted to U.S. Fish and Wildlife Service, ContractNo. 14-16-0001-81091(TA). Copies available from U.S.Fish and Wildlife Service, Portland, Oregon.

Raymond, A. W:, and \-: M. Parb1990 Archaeological Sites Exposed by Recent Flooding

of Stillwater Marsh, Carson Desert, Churchill County,Nevada. In Wetland Adaptations in the Greal Basin,edited by J. C. Janetski and D. B. Madsen, pp. 33-61.Museum of Peoples and Cultures Occasional Papers No.I. Brigham Young University, Provo, Utah.

Raymond, A. W., and E. Sobel1990 The Use ofThi Chub as Food by the Indians of the

Western Great Basin. Journal of California and GreatBasin Anthropology 12:2-18.

Rhode, D. R.1981 The Mountains and the Lake: Prehistoric

Lacustrine-Upland Settlement Relationships in theWalker River Watershed. Western Nevada. Ph.D. disser-tation, University of Washington, Seattle. UniversityMicrofilms, Ann Arbor.

1990 Settlement Patterning and Residential Stability atWalker Lake, Nevada: The View from Above. InWeiland Adaptations in the Great Basin, edited by J. C.Janetski and D. B. Madsen, pp. 101-120. Museum ofPeoples and Cultures Occasional Papers No. I. BrighamYoung University, Provo, Utah.

Richter, J.1986 Evidence for a Natural Deposition of Fish in the

REPORTS 717

Stiner, M. C.1991 The Faunal Remains From Grona Guattari: a

Taphonomic Perspective. Current Anthropology32:103-117.

Thomas, D. H.1985 Three Modes of Inquiry into the Lacustrine Lifeway.

In 11re Archaeology of Hidden Cave. Nevada, edited byD. H. Thomas, pp. 21-38. Anthropological Papers Vol.61, Pt. I. American Museum of Natural History, NewYork.

1990 On Some Research Strategies for Understanding theWetlands. In Wetland Adaptations in the Great Basin,edited by J. C. Janetski and D. B. Madsen, pp. 277-283.Museum of Peoples and Cultures Occasional Papers No.I. Brigham Young University, Provo, Utah.

Todd, L. C.1987 Taphonomy of the Homer II Bone Bed. In 11re

Horner Site: The 7jIpe Site of the Cody CulturalComplex, edited by G. C. Frison and L. C. Todd, pp.107-198. Academic Press, New Yon.

Van Neer, W., and A. M. Muniz1993 "Fish Middens": Anthropogenic Accumulations of

Fish Remains and Their Bearing on ArchaeoichthyologicalAnalysis. Journal of Archaeological Science 19:683-695.

Voorhies, M.1969 Taphonomy and Population Dynamics of an Early

Pliocene Vertebrate Fauna, Knox County, Nebraska.Contributions to Geology Special Paper No.2.University of Wyoming, Laramie.

opercles from fish larger than 20 mm should be representedin the flotation samples if indeed they were deposited andrecovered in the sample.3. Several investigators of mammalian faunas (e.g., Frison1982; Kurten 1953; MacFadden 1992; Todd 1987) have doc-umented the underrepresentation of neonates and juvenilesin fossil assemblages and suggested that such scarcity resultsfrom differential preservation. Ontogenetic changes in bone,which would enhance preservation potential of the mature

versus young tui chub, certainly need to be considered.4. Indeed, other research involving modern feeding experi-ments suggests that bones of some fish are virtually destroyedduring vertebrate digestion. A. K. G. Jones (1984, 1986) feddogs and humans various fishes (cod, herring, and snapper);only between 9 and 17 percent of the fish remains in the feceswere identifiable to element. 1 fed a dog one complete cohosalmon (Oncorhynchus kisutch) and found only three identi-liable remains in the feces: two eye lenses and one vertebrafragment (Butler 1990). On the other hand, the fish remainsin the Fish Slough Cave coprolite were relatively well-pre-served, representing elements from all parts of the skeleton.Given that the data are archaeolOgical, the proportion of the

carcass originally ingested is not known. The data suggest,however, that bone destruction may not be as great forcyprinids/catostomids as for other fishes examined.5. Contrary to Raymond and Sobers (1990) finding thatsmaller chub had higher rates of return than larger ones,Broughton (1994) uses the Raymond and Sobel data(1990:Table 5) to show that chub length is positively corre-lated with return rate (Spearman's rho = .671, P = .027)

(Broughton 1994:Table I). Apparently, the discrepancy isdue to varying treatments of "costs", particularly "handlingtime," by the authors. Broughton (1994) uses handling timethat includes only the time needed to set and retrieve the netand remove the l1Sh. In addition to this handling time,Raymond and Sobel (1990) also include time needed toprocess (clean, gut, eviscerate, dry) the fish, which, theynote, is much greater for larger l1Sh than for smaller fish.

Notes1. This interpretation may explain my failed attempts toidentify annuli on the small opercles from the Stillwaterarchaeological deposit. Many of the chub represented in thearchaeological deposit may have expired before the annulusformed, if most of the chub died in their flTst or second yearsof life.2. Opercles are relatively large elements. OL is the largestlinear dimension on the element and would constrain themesh size through which the specimen could pass. The OLfrom a chub measuring 55 rom SL ranges between 5.2 and5.8 rom. apercles would have to be three times smaller thanthese to pass through 1.6-rom mesh. In short, opercles fromfish 15-20 mm SL may be lost through 1.6-mm mesh, but

- -RecetvedAugust 7, /995; «cc~ted December /, /995.