Dietary Variability Among Prehistoric Forager-Farmers of Eastern North America

Wright State University CORE Scholar

Sociology & Anthropology Faculty Publications Sociology & Anthropology

6-1996

Kristen J. Gremillion

Kristin D. Sobolik Wright State University - Main Campus, [email protected]

Follow this and additional works at: https://corescholar.libraries.wright.edu/socanth

Part of the Anthropology Commons, and the Sociology Commons

Repository Citation Gremillion, K. J., & Sobolik, K. D. (1996). Dietary Variability Among Prehistoric Forager-Farmers of Eastern North America. Current Anthropology, 37 (3), 529-539. https://corescholar.libraries.wright.edu/socanth/11

This Article is brought to you for free and open access by the Sociology & Anthropology at CORE Scholar. It has been accepted for inclusion in Sociology & Anthropology Faculty Publications by an authorized administrator of CORE Scholar. For more information, please contact [email protected].

Dietary Variability among Prehistoric Forager-Farmers of Eastern North America Author(s): Kristen J. Gremillion and Kristin D. Sobolik Reviewed work(s): Source: Current Anthropology, Vol. 37, No. 3 (Jun., 1996), pp. 529-539 Published by: The University of Chicago Press on behalf of Wenner-Gren Foundation for Anthropological Research Stable URL: http://www.jstor.org/stable/2744553 . Accessed: 13/02/2013 09:59

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].

The University of Chicago Press and Wenner-Gren Foundation for Anthropological Research are collaborating with JSTOR to digitize, preserve and extend access to Current Anthropology.

http://www.jstor.org

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions Volume37, Number3, JuneI996 1 529

age behavioral patternof sequential communitiesover DietaryVariability among a period of time. Although advances have been made PrehistoricForager-Farmers of in identifyingdifferences in production,processing, and consumption patterns between households (Hastorf EasternNorth America' I990) and other small communitygroups (Gumerman I994), variability,which is the basis of evolutionary change (Braun I990:63-63; Dunnell I980:38; O'Brien KRISTEN J. GREMILLION AND KRISTIN D. and Holland I990:40-4I), is obscured in the resulting SOBOLIK analyses. Departmentof Anthropology,The Ohio State In contrastto midden refuse,human paleofecesoffer University,Columbus, Ohio 43 210-13 64 directevidence of food consumption by individuals (who (gremillion.ir@osu. edu)/Department of Anthropology differin gender,age, and social status) in a varietyof and Institutefor QuaternaryStudies, Universityof environmentaland cultural contexts.Refined analytic Maine, Orono, Maine 04469, U.S.A. i2 X 95 techniquesmake it possible forus to obtain information about these and otherdimensions of prehistoricdietary Until recently,subsistence studies have tended to em- variabilityin unprecedenteddetail. Our goal is to dem- phasize one type of variabilityover othersas a focus of onstratethe potential of such an approach with refer- inquiry:long-term, directional change. However,in or- ence to recentlyobtained paleodietarydata fromSalts der to explain such trends (including the origin and Cave and Mammoth Cave, Mammoth Cave National growingimportance of agriculturalsubsistence), it is Park,Kentucky. critical to document variabilitythat occurs on a rela- tivelyrestricted temporal scale and along nontemporal PALEODIETARY RESEARCH IN THE dimensions. For example, plant food consumption is MAMMOTH CAVE AREA likely to be sensitiveto short-termseasonal changes in food availability. Analyses of agriculturalorigins in Salts Cave and Mammoth Cave are part of the Mam- many regions(e.g., Cowan I985, McCorristonand Hole moth Cave system,an extensive network(the world's i992, Wills I988) emphasizethe causal significanceof longest) running through a karstic plateau in west- seasonality and the role played by crops as storablere- central Kentucky (Crothersand Watson I993) (fig. I). sources.In addition,the costs and benefitsof producing, These caves were exploitedfor minerals by prehistoric processing,and consuming crops were probablysome- people,primarily during the mid-3dmillennium B.P. Be- what differentfor women and men, childrenand adults, cause of the constant humidity and temperaturein largehouseholds and small ones. These costs and bene- many of the cave interiorsas well as the exsiccating fits also varied accordingto the part of the plant used effectof some ofthe minerals,organic materials (includ- and the culturaland environmentalcontext of use. ing human paleofeces)are extremelywell preserved.Re- Neglect of these issues in studies of prehistoricsub- mains of this typefrom Salts Cave and Mammoth Cave sistence has been largelya productof limitationsinher- have providedkey evidence for the early development ent in the data base. Most investigatorsof archaeobo- in easternNorth America of agriculturebased on native tanical materials,for example, work with remains that crops (Watson I974a; Yarnell I969, I974). Paleofecal representcarbonized waste from refuse deposits. Oc- data were particularlycritical in demonstratinga sub- casionally they are able to recoverdesiccated or water- stantialdietary role forcultivated plant species. Overall, loggedmaterial that is well-preservedin its noncarbon- I127 paleofeceshave been analyzedfrom Salts and Mam- ized, organic state and offersa more complete record moth; ioo samples were studiedfor their macrobotani- of plant use. Analysis of such materialgives us the op- cal contentby Yarnell (I969, I974) and 27 by Stewart portunityto generalizeabout changes in the processing (I974). In addition,pollen frompaleofeces was analyzed and disposal of plant products but provides no direct by Schoenwetter(I974) and Bryant(I974). Yarnell (I969, evidence of consumption.Furthermore, refuse deposits I974) estimates that approximately74% of the Salts must usually be regardedas representinga sortof aver- Cave diet was composed of plant foodsthat were either cultivated(about 42%) or a productof habitats created by agriculturalactivity (another 3e2%). Skepticismabout i. ? I996 by The Wenner-GrenFoundation for Anthropological the representativenessof this dietarypattern diminished Research. All rights reserved OOII-3204/96/3703-0007$I.00. We in view of corroboratingevidence fromthe vestibule,or are gratefulto PattyJo Watson for the opportunityto workwith entrychamber, of Salts Cave, where numerouscarbon- the Salts Cave and MammothCave material.Washington Univer- sityin St. Louis and The Ohio StateUniversity provided financial ized seeds of cropplants were depositedduring the main supportfor our research,which was carriedout in collaboration period of cave utilization (GardnerI987, Yarnell I974). withthe Cave ResearchFoundation Archaeological Project. Spen- The presentproject was devised to expand and refine cerTomb and HeatherAlmquist-Jacobsen assisted with identifica- this database by applyingnew analytic techniques to a tionof Asteraceae pollen. Stephen Bicknell drafted figures I, 3, and 4. VaughnBryant, Paul Gardner,Bruce Smith, Patty Jo Watson, and series of i2 paleofecal samples collected fromSalts and one anonymousreferee provided valuable comments that helped us Mammoth. Each specimen providedmaterial for accel- to improvethe qualityof the manuscript. eratorradiocarbon dating and was processedfor macrore-

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions 530 1 CURRENT ANTHROPOLOGY

FG.T Mamm oth Cave s h_ \ ~~~National Park <

\L>>

~~~~~~~~~~~~~II ] ? ~ ~~~~~~~~~~2mi

FIG. I. The Mammoth Cave system. mains, parasite remains, and pollen. We do not report cave interiorand occupation of Salts Cave vestibule on faunalremains (which were poorlyrepresented in the (Gardneri987, Kennedyn.d., Kennedyand Watsonn.d.). specimens)or parasites(analysis of which is still in prog- Judgingby these determinationsand by analyses of pa- ress). Related researchalso produceda pioneeringeffort leofecal content,crops had become a significantportion to extractand measure steroidsin the paleofecesin order of the diet of the people using the caves before2,500 B.P. to determinethe sex of the individualswho deposited Withinthe time periodrepresented by the paleofeces, them(Sobolik et al. I995, WhittenI994). In thispaper there is evidence of considerabledietary continuity in we synthesizeresults fromthese differentlines of evi- that the same typesof plant foodswere consumedregu- dence to discuss the implicationsof variabilityin plant larly. In fact,diet has been judged as having remained use by the prehistoricforager/farmers of the Mammoth quite consistent during the period of major cave use Cave area along threemajor dimensions:temporal (both (Watson i969:77). With the availability of i2 directly long-termand seasonal), interindividual,and contextual. dated paleofecesthat span this period,there is an opportunityto evaluate this conclusion. Is there perhaps a trendtoward consumption of greaterquantities of culti- DIMENSIONS OF DIETARY VARIABILITY vated plant products? Long-termtemporal variability. For the Mammoth Aftercalibration (using CALIB version 3.0; Stuiver Cave area, therenow exists an extensiveseries of radio- and Reimer I993), radiocarbonages of the most recently carbon dates that providea sound chronologicalframe- collected i2 specimens fall within the 3d millennium work forearly food production (Gardner I987, Kennedy B.P. and forma continuousseries with considerableover- n.d., Kennedy and Watson n.d., Watson I974a). All di- lap at the two-sigmalevel (table i, fig.2, Kennedyand rectlydated paleofeces fall between ca. 2,700 and 2,250 Watson n.d.). The total of I4 determinations(including B.P., although there are indications of earlier (Late Ar- duplicates for2 samples that were run to test formold chaic period,ca. 4,000 to 2,700 B.P.) explorationof the contamination)are significantlydifferent at the .o5 level

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions Volume 37, Number 3, fune I996 1531

TABLE I AMS Radiocarbon Determinationsfor Salts Cave and Mammoth Cave Paleofeces

RadiocarbonAge Laboratory CalibratedAge Specimena (UncalibratedB.P.) Numberb (YearsB.P.)C

MCF-iA 2,335 ? 75 AA-Io079 2,707 (2,340) 2,I47 MCF-iB 2,485 ? 70 AA-ioo8o 2,752 (2,7IO, 2,630, 2,490) 2,346 MCF-2 2,575 ? 65 AA-ioo8i 2,780 (2,740) 2,464 MCF-3 2,365 ? 70 AA-ioo82 2,7I2 (2,350) 2,I59 MCF-4 2,485 ? 70 AA-ioo83 2,752 (2,7IO, 2,630, 2,490) 2,346 MCF-5 2,605 ? 70 AA-ioo84 2,846 (2,750) 2,478 MCF-6 2,700 ? 80 AA-ioo85 2,954 (2,780) 2,720 SCF-i 2,570 ? 70 AA-ioo86 2,782 (2,740) 2,369 SCF-2 2,4Io ? 70 AA-ioo87 2,729 (2,360) 2,322 SCF-3 2,605 ? 80 AA-ioo88 2,854 (2,750) 2,464 SCF-4A 2,590 ? 70 AA-ioo8g 2,794 (2,740) 2,468 SCF-4B 2,580 ? 70 AA-ioo0o 2,788 (2,740) 2,377 SCF-5 2,500 ? 80 AA-ioogi 2,760 (2,7IO, 2,62o, 2,580, 2,540, 2,530, 2,5I1) 2,345 SCF-6 2,703 ? 62 AA-iI738 2,935 (2,780) 2,742

NOTE: Calibratedages werecalculated using CALIB Version3.0 (Stuiverand ReimerI993). Cal- ibrationdatasets are fromStuiver and Pearson(I993) forMCF-iA, MCF-3,and SCF-2 and Pear- son and Stuiver(I993) forall others. IMCF = MammothCave feces;SCF = Salts Cave feces. bAll14C measurements were conducted at the NSF-ArizonaAMS Facility,University of Ari- zona, Tucson. cTwo-sigmamaximum (calibrated ages) two-sigma minimum.

(x2 = 22.40). However,if the two oldest samples (SCF-6 (when plant foods are in shortsupply) was emphasized and MCF-6) are removed,the remainingi2 are statisti- in earlierMammoth Cave area research(Watson I974, cally the same (x2 = I9.70), indicatinga highprobability Yarnell I974). Other researchers(e.g., Cowan I978a, that the majority of the paleofeces were deposited I985; Gremillion I993a, i995) have explored similar around the same time. The remainingpair representa causal links between food storageand food production somewhat earlier entryinto the cave. The contentsof for the Cumberland Plateau to the east of Mammoth specimens fromthese two temporalgroups do not ap- Cave, where a similar recordof early plant domestica- pear to differin any significantway; all i2 containsome tion has been found. Seasonal variations in resource remainsof crops, although their quantity and the species abundance have also been cited as causal factorsin ini- representedvary widely. However, the earlier sample tial cultivationof crops in other regions (McCorriston may be too small to reflectany differencesin composi- and Hole i992, Wills I988). tion that mightdifferentiate it fromthe later group. Previous analyses of paleofeces fromSalts Cave and Thus, Watson's (i969) originalassessment of consid- Mammoth Cave revealed seasonal variationin the con- erabledietary stability during the EarlyWoodland period sumption of crop plants. For example, Yarnell (I969) cannot be furtherevaluated using these data, given the notedthat maygrass (Phalaris caroliniana),whose grains likelihood that most of the paleofecalspecimens form a ripen in early summer,was generallyeither extremely contemporaneousgroup. However, directdating of pa- abundant or present in trace amounts. Furthermore, leofecesshould proveto be a usefultool in futurepaleo- when it was abundantmaygrass tended greatly to exceed dietarystudies, particularly if it can be used on a large other crops in quantity.Maygrass seeds and pollen are sample of specimens that can be grouped chronologi- often associated with strawberry(Fragaria virginiana) cally. This technique,especially if it provespossible to and blackberry(Rubus sp.), both of which are late- reduce the size of standarddeviations of radiocarbonde- spring-summerfruits likely to be eaten fresh(Bryant terminations,offers the prospect of tracing dietary I974, Yarnell I969). In contrast,several of the othercrop change over relativelybrief periods of time. taxa (includingsumpweed [Iva annual, chenopod [Che- Seasonality. The question of why cultivated plants nopodium berlandieri],and sunflower[Helianthus an- acquired subsistence importancedemands attentionto nuus]) were frequentlyassociated with hickory(Carya cyclical variationin diet that occurs at a restrictedtem- sp.) and seldom occurred with maygrass (Marquardt poral scale in responseto seasonal variationin resource I974). Thus, it seems likely that sumpweed,sunflower, abundance. The importanceof crops as storedfoods in- and chenopod (as well as hickory)were storedand con- tended to be consumed duringthe winter and spring sumed out ofseason, oftenin quantity.In contrast,may-

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions 532 1 CURRENT ANTHROPOLOGY

E 11_

Fl~~~~~~~~~~~~~~~~~~E LIEZ EI LIIIY II

L Z

_~~ I

I ~ I I I _ _ I I I I

3000 2900 2800 2700 2600 2500 2400 2300 2200 2100 2000 ca I BP FIG. 2. Calibrated age ranges for 12 paleofecal specimens. Solid bars, calibrated one-sigmaranges; emptybars, calibrated two-sigmaranges. The total of i4 determinationsincludes duplicates taken fromMCF-i and SCF-4. The chart was produced using CALIB Version3.0 (Stuiverand Reimer 1993). grass was apparentlyutilized when available. Although as ingredientsin medicinal teas or foods. In contrast, maygrassis considereda crop plant (Cowan I978b), be- backgroundpollen is ingestedduring respiration or as a cause ofits earlyavailability (before other crops and nuts contaminantof food and water. This type of ingestion were harvestable)it was probablynot storedin any quan- is prevalentduring the pollinationseason ofa plant (e.g., tity,if at all. springin the case of pine and oak and fall forragweed Studies of pollen (BryantI974, SchoenwetterI974) and juniper).Economic pollen fromknown crops may provided informationon seasonal utilization of the have been storedwith cropseeds and fruits(either inten- caves. Many of the feces were deposited during the tionallyor unintentionally)and is thereforea poor indi- springand summermonths, although there is evidence cator of seasonality.Analysis of economic noncroppol- of activitythroughout most of the year (Watson I974, len and backgroundpollen is moreuseful for seasonality Yarnell I974). However,in the earlierwork most of the determinationbecause the presenceof pollen in a sam- specimens studiedfor macrobotanical content were not ple usually reflectsthe season in which a plant blooms. subjectedto pollen analyis. We have analyzed both pol- Most of the pollen grainsidentified in the paleofeces len and macrobotanical content of each specimen to are probablyeconomic pollen. Of these, only one repre- maximize chances of documentingseasonal variationin sents economic noncrop pollen (the mustard family, plant food consumption. Brassicaceae). Mustard pollen made up 58% of pollen Pollen remainsfound in paleofecescan be classifiedas fromsample SCF-3, which is thereforelikely to repre- eithereconomic pollen or backgroundpollen. Economic sent springdeposition (figs. 3 and 4). Backgroundpollen pollen enters the body through the intentional con- was usually not observedin the samples in largeenough sumption of flowersor seeds or throughthe uninten- frequenciesto indicateseasonality. Three exceptionsare tional ingestionof pollen adheringto plant parts used sample SCF-3, in which i9% of the pollen was hickory;

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions Volume 37, Number 3, JuneI996 1533

CROPS

CLo CL POLLEN

. L CONCENTRATION rEU @ t 6 I SAMPLE a a 0 EU L. I. o (grains/gram) L d:o 3 zn SCF-1 ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~90.172 SCF-2 16.175

SCF-3 113.885 SCF-4 157.939

SCF*5 tra%52.310 MCF-SSCF-6 .l | - > | . [| | o 116.1612.9

MCF-1 29.224

MCF.2 *162.302

MCF-3 *56.561

MCF-4 *.34.546

MCF-5 * * .122.499

MCF-6 *811

Pollen= of pollencounted Macrobotanical=estimated volume X of material =trace I I 0 loot FIG. 3. Crop remains and selected pollen taxa fromSalts Cave and Mammoth Cave paleofeces.

sample SCF-6, in which I3% of the pollen was oak midsummeruntil firstfrost. Maygrass in MCF-2 and (Quercus sp.); and sample MCF-5, in which 4% of the MCF-4 suggestsearly to midsummer,although the trace pollen was oak. Both treesbloom in the spring,and nei- amountsrepresented might have been retainedfor some thertype of pollen is likely to be ingestedwith the nut- weeks in the intestineand been deposited later in the meats (which could have been storedand consumedout year. SCF-2, SCF-4, and MCF-3 lack small noncrop of season). seeds. However, high frequencies of grass pollen in Samples withoutsufficient quantities of noncroppol- SCF-2 probablyrepresent warm-season ingestion. SCF-4 len did contain some seeds that reflectseasonality (fig. contains cucurbit pollen, which is unlikely to have 4). Seeds of wild or weedy taxa producingfruits or small storedwell and was probablyingested in summer.Be- grainsare assumed not to have been stored.MCF-5 and cause of the lack of otherseasonal indicatorsin MCF-3, MCF-6 contain strawberryseeds fromfruits that would we suggestthat it was depositedin the winter. most likely have been eaten freshin the late spring,as Many ofthe specimensindicate that crops were stored well as maygrass,which ripens duringMay and June. and consumedout ofseason (fig.4). This patternis repre- These occurrencesindicate that MCF-5 and MCF-6 were sented by the four samples deposited duringspring or probablydeposited in the late springor early summer. early summer (SCF-3, SCF-6, MCF-5, and MCF-6). As- Summer or early fall deposition is likely for MCF-i, sumingthat maygrasswas consumed at harvest,MCF-2 MCF-2, MCF-4, SCF-i, and SCF-5, based on the occur- and MCF-4 containstored crops and/or hickory nut (pre- rence ofweed seeds such as nightshade(Solanum), ama- sumably also stored). SCF-4 representssummer con- ranth(Amaranthus), knotweed (Polygonum), poke (Phy- sumption of squash/gourd (Cucurbita) flowers along tolacca americana), wood sorrel(Oxalis), and/orpanic with stored seeds of the same taxon and other crops. grass (Panicum). These plants typicallyhave long sea- MCF-i, SCF-i, SCF-2, and SCF-5 may have been depos- sons of floweringand fruitingthat last fromearly or ited duringthe late-summer-early-fallharvest season

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions 534 CURRENT ANTHROPOLOGY

SAMPLE J F M A M J J A S 0 N D OUTOF SEASON(stored) SEASONALITYINDICATORS

MCF-1A77| ? knotweed,poke

MCF-2 L chenopod, sunflower, sumpweed amaranth,poke, panic grass, maygrass

MCF-3 chenopod [?]

MCF-4 squash, chenopod, sunflower, sumpweed knotweed,maygrass

MCF-5 I hickory, sunflower, sumpweed strawberry,maygrass, oak pollen

MCF-6 | chenopod, sunflower, hickory strawberry,maygrass

SCF-1 amaranth,poke, knotweed,panic grass

SCF-2 grass pollen

SCF-3 squash hickory, mustardpollen

SCF-4 squash, sunflower, sumpweed cucurbit pollen

SCF-5 EIZZiIZZIIIZZIII nightshade, panic grass

SCF-6 squash, sunflower, hickory wood sorrel, maygrass,oak pollen

LIZZZI Seeds

Pollen

OverlapBetween Seeds andPollen FIG. 4. Probable season of deposition of Salts Cave and Mammoth Cave paleofeces based on macrobotanical and pollen evidence. Ranges of seed and pollen productionrepresent estimates for species likely to have been foundin west-centralKentucky. Sources forseed seasonality: Asch and Asch (I977), Cowan (i985), and Radford,Ahles, and Bell (I968). and thereforedo not necessarilycontain storedfoods. If When comparingindividual specimens,the source of MCF-3 was in fact deposited in winter,the chenopod variabilitycan be difficultto determine.First of all, containedin it was stored. there is no way of telling at the presenttime whether Althoughit appears likely that crops were consumed specimens of similar age were depositedby the same or over much of the seasonal cycle, no obvious seasonal differentindividuals (although DNA analysismakes this patterningin the dietary importance of crops emer- a veryreal possibilityfor the near future;see, forexam- gences fromanalysis ofour sample. However,the poten- ple, Sutton,Malid, and Ogram I995). Even ifwe assume tial for such an investigationis high given consistent that the specimens are fromdifferent individuals, the application of analytic and quantitativemethods to a possible underlyingcauses of dietary differencesare largerdata base. many: seasonality,year-to-year variation in productivity Interindividualvariability. The issue of interindivid- of differenttaxa, individual food preferences,and indi- ual variabilityin diet is seldom addressedin prehistoric vidual physiologyand health status. One inferencethat subsistence research,primarily because of the lack of is supportedby available data, however,is that meals relevantevidence. However, consumptionof particular eaten by contemporaneousindividuals (perhapsby the kinds of foods (includingcrops) probablyvaried widely same individual)are as likelyto varywidely in cropcon- within as well as between communities.Differences of tentas theyare to be similarin thisrespect. How can we this kind are likely to be relevantto why food produc- determinewhether these differencesreflect systematic tion began and succeeded when and where it did. For dietarypatterns within a food-producingcommunity? example, the use of crop-basedgruels to wean infants One way to approachthis question is to considerfood appears to have accompanied the intensificationof food consumptionby differentsubgroups. Analysis of fecal productionin the midcontinentalUnited States (Braun steroidsoffers one method fordetermining the biologi- I987, Buikstraet al. I987). Differencesin diet and heatlh cal sex of individuals as a basis for identifyinggender betweensocial ranksin stratifiedsocieties (Powell I988, differencesin diet. Ratios of hormonesin the I2 speci- i99i) also indicate that the benefitsof food production mens fromSalts Cave and MammothCave stronglysug- cannotbe assumed to have been identicalfor all individ- gest that the cave explorersresponsible for the fecal de- uals withina community.Fortunately, paleofecal analy- posits were predominantlyor exclusivelymale (Sobolik sis gives us the opportunityto examine interindividual et al. I995, WhittenI994). Theremay not have been variabilityin food consumption.Furthermore, it holds significantdifferences in male and female diet among the potential for comparing diet between major sub- earlyforager-farmers of the Mammoth Cave area, but if groupswithin a population. cavers were of one sex the probabilityof ever finding

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions Volume 37, Number 3, JuneI9961 535

out is low. However, one veryimportant piece of infor- only the pollen intact in the sample. Pollen may thus mation has been obtained: that genderis not likely to indicatethe intentionalconsumption of flowers or seeds account for dietary variabilityrepresented in the pa- or the intentionalor unintentionalingestion of pollen leofeces. adheringto otherplant partsfor medicinal teas or foods When in the caves, adult males frequentlyate crop (Bryant I974, Sobolik and Gerick I992). Pollen, de- seeds, probably on a daily basis; however, quantities pending on size and structure,becomes caught in the consumed varied considerably.Comparison of samples intestinallumen, allowing it to be excretedin fecalsam- of similarage and season of depositionpermits elimina- ples for up to one month afteringestion (Kelso I976, tion of some potentialcauses ofvariation and allows for Williams-Dean I978). Consequently,the pollen content identificationof behavioraldifferences between individ- of paleofeces does not reflectone meal but rathermay uals withinsimilar environments. For example, samples be a productof numerousmeals containinga varietyof SCF-i and MCF-2 producednearly identical radiocarbon pollen types (Sobolik I988, Williams-Dean I978). determinations(table I) as well as macrobotanicaland Previous studies have indicated that paleofeces and pollen profilesthat are similarwith respectto seasonal- modern human feces that contain pollen frequencies ity (summer-fall)(figs. 3 and 4). Both contain high per- above ioo,ooo grains/gramcontain recently ingested centages of crop remains as well as panic grass,poke, and easily recognized economic pollen (Sobolik I988, and amaranth seeds and small amounts of hickory. Sobolik and Gerick I992). Paleofecal samples that con- However, whereas MCF-2 is dominated by chenopod, tain pollen concentrationsbetween 99,ooo and 25,000 SCF-i has a more or less equal mix of squash, chenopod, grains/gramprobably contain economic pollen types and sunflower.This analysis shows that much of the that are not as easy to recognize because pollen inges- variationin plant food consumptionobserved in these tion was not recent; and samples that have pollen con- paleofecal samples may be attributableto seasonal pat- centrationvalues less then 25,000 grains/grammay con- ternsof resourceabundance and thatin similarenviron- tain economic pollen, but this pollen will not be mental contextsdiet compositionamong caversmay in recognized because ingestion may have taken place fact have been quite consistent.On the otherhand, no many days and even weeks ago. single crop species was invariablychosen as the primary An example of an economic pollen typefound in pa- componentof meals by differentindividuals or by the leofeces fromthe caves is sumpweed, as indicated by same individualson differentoccasions. Futurecompar- the high pollen concentrationvalues of the samples in isons of this kind utilizing largersamples may help to which it occurs and its extremelyhigh frequencies in six clarifythis issue. of the samples. Two of the specimensthat containhigh Variabilityin the contextand characterof plant use. frequenciesof sumpweed pollen exhibitpollen concen- A final dimension of variabilityis the culturalcontext tration values indicating the recent ingestion of eco- ofplant use. Althoughsubsistence research tends to em- nomic pollen, and the remainingfour samples contain phasize nutritionalcontent, plants have a multitudeof pollen concentrationvalues indicatingthe probablein- otheruses that may well have some bearingon the pro- gestion of pollen, although not as recently (fig. 3). cess ofdomestication. The prehistoricseed cropsof east- Sumpweed pollen was thus probablyingested either as ern NorthAmerica, although probably primarily impor- food or along with sumpweed achenes. tant for the food value of their seeds and fruits,also Evidenceof a similarpattern is offeredby the presence producedflowers, stems, leaves, and roots and probably of significantlyhigh frequencies of cheno-am pollen had nonfood uses. Various modes of use are likely to (Chenopodiaceae and/or Amaranthus) and chenopod have had differentutility dependingon their environ- seeds, high-spinepollen (sunflowertype) and sunflower mental and culturalcontext. For example,consumption achenes, sumpweed pollen and sumpweed achenes, and of food would have been subject to varyingconstraints maygrass grains/glumesand Poaceae (grass) pollen. at a riverinebase camp, on a caving expedition,and on However, the combination of Cucurbitaceae (squash an overnighthunting trip. family)pollen and cucurbitseeds in sample SCF-4 indi- Midden refuse seldom gives us direct evidence for cates that cucurbitflowers were intentionallyingested how plantswere used (Sobolik I994) or forwhat purpose at the same time as cucurbitfruits (or seeds) because (as food,medicine, or constructionmaterial or in ritual). cucurbitflowers and fruitproduction take place in spa- Paleofeces most likely underrepresentthe ingestionof tially separate locations on the plant. Mustard pollen meat, particularlyof large animals, because of the ab- was observedin high frequencyin sample SCF-3, indi- sence of large bone remains (Sobolik I993). However, cating the intentionalingestion of mustardpollen and they offermore precise informationon exactly what flowers.Mustard seeds were not observedin this sam- partswere used. In many plants,pollen is presentat the ple, which suggestseither that mustardseeds were in- same time and foundin the same location as seeds and/ gestedand excretedmany days beforesample deposition or fruits.For example, chenopod and sunflowerfruits or that mustardflowers were ingestedindependently of both ripenin the same location as residualpollen. With the seeds, most likely as medicine. the ingestionof these seeds and/orfruits, pollen is auto- Broadeningthe geographicalscale bringssubsistence maticallyingested as well, eitherintentionally or unin- diversityaccording to environmentalcontext into sharp tentionally.If a flowertype is frequentlyingested, the focus. In the Mammoth Cave area, it is only withinthe soft flowerparts will most likely be digested,leaving caves themselves that remains of domesticatedplants

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions 536 1 CURRENT ANTHROPOLOGY

dating to the 3d millennium B.P. are abundant. Rock- seeds (GremillionI995). Domesticated chenopodwas at shelters in Mammoth Cave National Park with Early least presentin the diet ofthese populations several cen- Woodland components have yielded little evidence of turies earlier (Smith and Cowan I987). Thus, even if food productionwith the possible exception of bottle heavy consumptionof crop seeds duringthe EarlyWood- gourd (Lagenaria siceraria) (Gremillion I990, Wagner land in the Mammoth Cave area was limited to caving I978, Watson I985). However, such contemporaneous expeditions,such a specialized dietaryrole was not nec- noncave data sets are limited in number and thus are essarilytypical of early farmingpopulations in the up- unlikelyto be an adequate reflectionof subsistence vari- lands of the midcontinentalUnited States. ability.For the period priorto ca. 3,000 B.P. (the tradi- Variable preservationof organic remains is a major tional boundary between the Late Archaic and Early factorinfluencing the archaeologicalrecord of foodpro- Woodlandperiods in EasternWoodlands prehistory), the duction found in differentenvironmental situations. prevalentenvironmental context of evidence for early The exceptionalopportunities for survival of organicre- foodproduction is just the opposite,with noncave sites mains foundin drycave passages and in some rock shel- such as rocksheltersand shell middens in and near the tersaccount to some degreefor the substantialarchaeo- Green River floodplainproviding the bulk of the data logical recordof early food productionat such sites. In (Crawfordi982; Gremillion I990, I994; Marquardtand contrast, the collections of charred crop seeds that Watson i983; Wagner I978; Watson i985). amply documentthe earlydevelopment of foodproduc- These Late Archaic collections of botanical material tion at nonshelteredsites in the largerriver valleys of (which date primarilyto the mid-4thmillennium B.P.) the Midwest and Mid-South to the west of Mammoth are no match for the later cave assemblages in either Cave (e.g., Asch and Asch I985, Chapman and Shea quantityor variety.Shell middenshave producedno evi- I98I, CritesI99I, JohannessenI984) have survivedin dence of food productionsave remains of gourd/squash spite of the great potential for rapid mechanical and (which,in lightof our presentunderstanding of the evo- chemical breakdownof plant tissues in such environ- lution of Cucurbita pepo, may representa wild gourd ments. Thus, it seems unlikely that preservationalfac- ratherthan, or as well as, a garden crop [Cowan and tors are solely responsiblefor the absence of a recordof Smith I993, Decker i988, Decker-Walters et al. I993, agriculturalsubsistence of similar compositionand age Smith,Cowan, and Hoffmann1992]). Possible seed crops fromopen sites in the Mammoth Cave area. The inten- representedat nearby Peter Cave (which is actually a sity of archaeological researchis implicated as well as rock shelter rather than a true cave) include gourd/ geographicalvariability in the behavior of prehistoric squash, chenopod,and knotweed,but only chenopodis people. Systematic comparison of the formationpro- abundant and exhibits morphologicalevidence of do- cesses affectingdifferent site types between regions mestication(Crawford I982, Yarnell I993). should help to clarifythe extentto which the paleofecal Althoughlimited in quantity,archaeological evidence evidence (necessarilyrestricted to shelteredsites) is rep- of the use of crops outside the caves by theirLate Ar- resentativeof subsistence behavior over a broad range chaic predecessorschallenges the contentionthat the of environmentalcontexts. caver's mode of plant use is an aberration,a special diet of "trail mix" that was consumed only in special con- CONCLUSIONS texts (see discussions in Gardner i987 and Watson i985:I28). The rock sheltersof this region,unlike the Our studyillustrates the advantagesof applyingmulti- deep caves, were suitable for a wide range of human ple analytic techniques to prehistoricpaleofecal mate- activities and are likely to representa largersubset of rial as a method for investigatingdietary variability. subsistencepractices. Thus, the occurrenceof cultivated Many types of variability (seasonal, interindividual, plants at such sites indicates that theywere consumed functional,environmental) are relevantto the explana- in settingslacking the constraintson food choice im- tion of long-termsubsistence change and meritcompre- posed by the rigorsof cave exploration.Compelling evi- hensive documentation. Although in most cases the dence of a similar contextfor early food productionis complexityof plant use is obscuredin the archaeological provided by midden-derivedplant remains from rock record,the feasibilityof this line of researchis greatly shelters on the Cumberland Plateau, located approxi- enhanced by continuing methodological innovation mately 2oo km to the northeastof Mammoth Cave aimed at the fullestpossible exploitationof directevi- (GremillionI993b). Furthermore,the fact that remains dence forthe diet ofindividuals. Awareness of the many of crops were deposited in quantityat the entranceof dimensions of variabilitywithin and between prehis- Salts Cave as well as withinthe cave itselfsuggests that toric farmingcommunities of eastern North America they were unlikely to have been used exclusively as (FritzI990, GremillionI995, ScarryI993) continuesto "trail mix." growin the wake ofthe applicationof new analytictech- The presenceof crop remains amongstmidden refuse niques and resultingrefinement of the evidentialbasis in noncave locations does not necessarilyargue forthe forprehistoric subsistence. widespreadand generalizedadoption of an agricultural Direct assessment of the age of individualpaleofecal diet in the Mammoth Cave area. However, paleofeces specimensholds considerablepotential for tracking sub- fromrock shelterson the CumberlandPlateau thathave sistence change.The small size and apparentcontempo- been directlydated to ca. 3,000 B.P. have been foundin raneityof our sample prohibitsassessment of fluctua- some cases to contain substantial quantities of crop tion in the dietaryrole of crops duringthe major period

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions Volume37, Number3, June1996 1 537

of cave use. However,larger samples in combination rior,on the one hand,and noncaveand cave entrance withsmaller standard deviations for radiocarbon dates archaeobotanicalassemblages, on the other.These ex- would provideopportunities for delineation of subtle planationsremain incomplete, but the variationbe- changesin cropusage over relativelybrief periods of tweensite typesand preservationalcontexts in thefre- time.Such a databasewould also permitus to pinpoint quencyof crop remains illustrates the need for attention timesof initial introduction of individual crops and in- to thefull range of environmental settings used bypre- tensificationof farmingin orderto correlatethese hu- historicpopulations. There is no reason,on theoretical manbehavioral changes with environmental and social orempirical grounds, to assumethat crops were equally variables. importantin all contexts.Thus, in orderto document Combinedpollen and macroplant data provide a strong and explainthe behaviorof earlyforager-farmers accu- evidentialbasis for seasonality of crop consumption and rately,we needto considerthe constraints and opportu- have allowedus to substantiatethe claim thatstored nitiesoffered by nonagriculturalas well as agricultural cropswere an importantpart of the caver diet. The fact activities. that crops were consumedyear-round indicates that For the mostpart, our studycomplements and con- theywere not reserved for periods of shortage but inte- firmsthe results of earlierinvestigations of theprehis- gralcomponents of human diets. However, the frequent toricforager-farmers of the MammothCave area (Wat- consumptionof cropseeds out of seasonsuggests that son I969, I974). However,by utilizing newly available theirrole as storableresources may have been an impor- techniqueswe have refinedour understandingof early tantdeterminant of theirdietary importance. Whether food-producingsocieties in easternNorth America and thestorability of food crops was a keyfactor influencing charted some potentially profitable directions for future theinitiation of plant domestication and farming can be researchinto the origins and development of agriculture determinedonly through recovery and analysisof pa- worldwide. leofecalmaterial from the 4th millennium B.P. Preliminarycomparisons of samples of similar age and ReferencesCited seasonalprofile indicate that, although any conclusive statementsabout interindividualdietary variability ASCH, DAVID L., AND NANCY E. ASCH. I977. Chenopodas thisline of investiga- cultigen:A reevaluationof some prehistoriccollections from wouldat thispoint be premature, eastemNorth America. Midcontinental Journal of Archaeol- tionholds much promise for the future. Our preliminary Ogy 2:3-45. assessmentdoes suggest that consistent consumption of 1.I985. "Prehistoricplant cultivation in west-centralIlli- cropseeds had highutility for cave explorersbut that nois," in Prehistoricfood production in NorthAmerica. Edited themix ofspecies used was likelyto be quitevariable. by R. I. Ford,pp. I49-203. Museumof Anthropology, Univer- sityof Michigan, Ann Arbor,Anthropological Papers 75. Thispattern contrasts with the dietary reliance of many BRAUN, DAVID P. I987. "Coevolutionof sedentism,pottery Late Prehistoricpopulations in easternNorth America technology,and horticulturein the centralMidwest, 200 on a singlestaple carbohydratesource, maize (Smith B.C.-A.D. 6oo," in Emergenthorticultural economies of the i989). Analysis of fecal steroids (Sobolik et al. i995, EasternWoodlands. Edited by WilliamF. Keegen,pp. I53-82. shownfurther that our subjectsare Centerfor Archaeological Investigations, Southem Illinois Uni- WhittenI994) has versity,Carbondale, Occasional Paper7. males and demonstratesthat we have thepotential to . I990. "Selectionand evolutionin nonhierarchicalorgani- examinegender differences in diet. zation,"in The evolutionof political systems: Sociopolitics of Our analyseshave also allowedus to documentthe small-scalesedentary societies. Edited by SteadmanUpham, consumptionof pollen and/orflowers (items that are pp. 62-86. Cambridge:Cambridge University Press. BRYANT, VAUGHN M., JR. I974. "Pollen analysisof prehistoric usuallynot preservedin open sites)as foodand/or as humanfeces from Mammoth Cave," in Archaeologyof the medicine.Such conclusionsmay provide insights into MammothCave area. Editedby PattyJo Watson, pp. 203-49. the domestication,management, or persistentutiliza- New York:Academic Press. tionof plants whose uses in prehistoryare imperfectlyBUIKSTRA, JANE E., JILL BULLINGTON, DOUGLAS K. understood.For example,gourd/squash (C. pepo) was CHARLES, DELLA C. COOK, SUSAN R. FRANKENBERG, LYLE W. KONISGSBERG, JOSEPH B. LAMBERT, AND LIANG XUE. used forits fruits(both as containersand as food)and I987. "Diet, demography,and the developmentof horticul- its seeds (Cowann.d.). Our researchdemonstrates that ture,"in Emergenthorticultural economies of theEastern its flowerswere also consumedprehistorically in east- Woodlands.Edited by WilliamF. Keegan,pp. 67-86. Center ernNorth America. These multipleuses help explain forArchaeological Investigations, Southem Illinois University, the millennia-longecological association between pre- Carbondale,Occasional Paper7. CHAPMAN, JEFFERSON, AND ANDREA B. SHEA. I98I. The historichuman populations and this species. Pollen evi- archaeobotanicalrecord: Early Archaic period to contactin the dencefrom paleofeces also providesinformation on the lowerLittle Tennessee River Valley. Tennessee Anthropologist medicinaland/or nutritional usefulness of plants such 6:6I-84. as the mustards,whose seeds are seldomrecovered ar- COWAN, C. WESLEY. 1978a. Seasonalnutritional stress in a Late Woodlandpopulation: Suggestions from some eastemKen- chaeologically.The same is probablytrue of many me- tuckycoprolites. Tennessee Anthropologist 3: I I7-28. dicinalplants, whose preparation methods often ensured . I978b. "The prehistoricuse and distributionof maygrass the destructionor alterationof anatomicalstructures in eastemNorth America: Cultural and phytogeographicalim- thatmake identification of macroremains possible. plications,"in The natureand statusof ethnobotany. Edited by R. I. Ford,pp. 263-88. Universityof Michigan Museum of Finally,we have consideredvariability in the spatial Anthropology.Anthropological Papers 67. contextof plant use in theMammoth Cave areaby pro- . I985. Fromforaging to incipientfood production: Subsis- posingexplanations for the contrast between cave inte- tencechange and continuityon the CumberlandPlateau of

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions 538 CURRENT ANTHROPOLOGY

easternKentucky. Ph.D. diss. Departmentof Anthropology, KENNEDY, MARY c. n.d. "Radiocarbondates from Salts and Universityof Michigan,Ann Arbor,Mich. MammothCaves," in Of caves and shellmounds. Edited by . n.d. "Evolutionarychanges associated with domestication K. C. Carstensand P. J.Watson. Tuscaloosa: Universityof Ala- of Cucurbitapepo: Evidencefrom eastem Kentucky," in Peo- bama Press.In press. ple, plants,and landscapes: Studiesin paleoethnobotany. KENNEDY, MARY, AND PATTY JO WATSON. n.d. "The chronol- Editedby KristenJ. Gremillion. Tuscaloosa: Universityof ogyof earlyagriculture at Salts Cave and MammothCave, AlabamaPress. In press. MammothCave NationalPark, Kentucky," in Cave archaeol- COWAN, C. WESLEY, AND BRUCE D. SMITH. 1993. New per- ogyin NorthAmerica. Edited by J.Steele. Special Bulletin, Na- spectiveson a wild gourdin easternNorth America. Journal of tionalSpeleological Society. In press. EthnobiologyI 3: I7-54. MARQUARDT, WILLIAM H. I974. "A statisticalanalysis of con- CRAWFORD, GARY. i982. Late Archaicplant remains from west stituentsin humanpaleofecal specimens from Mammoth centralKentucky: A summary.Midcontinental Journal of Ar- Cave," in Archaeologyof theMammoth Cave area. Editedby chaeology7:205-24. PattyJo Watson, pp. I93-202. New York:Academic Press. CRITES, GARY D. I99I. Investigationsinto earlyplant domesti- MARQUARDT, WILLIAM H., AND PATTY JO WATSON. I983. cates and foodproduction in MiddleTennessee: A statusre- "The Shell MoundArchaic of Western Kentucky," in Archaic port.Tennessee Anthropologist I6:69-87. huntersand gatherersin theAmerican Midwest. Edited by J. CROTHERS, GEORGE M., AND PATTY JO WATSON. I993. "Ar- Brownand J.Phillips, pp. 323-39. New York:Academic Press. chaeologicalcontexts in deep cave sites:Examples from the MCCORRISTON, JOY, AND FRANK HOLE. I992. The ecologyof EasternWoodlands of North America," in Formationprocesses seasonal stress and the origins of agriculturein the Near East. in archaeologicalcontext. Edited by Paul Goldberg,David T. AmericanAnthropologist 93:46-69. Nash, and Michael D. Petraglia,pp. 53-60. Madison:Prehis- O'BRIEN, MICHAEL J., AND THOMAS D. HOLLAND. I990. toryPress. Variation,selection, and thearchaeological record. Archaeologi- DECKER, DEENA S. I988. Origin(s),evolution, and systematics cal Methodand Theory2:3I-79. of Cucurbitapepo (Cucurbitaceae).Economic Botany 42:4-I 5. PEARSON, G. W., AND M. STUIVER. I993. High-precisionbi- DECKER-WALTERS, DEENA S., TERRENCE W. WALTERS, C. decadal calibration of the radiocarbon time scale 500-2500 B.C. WESLEY COWAN, AND BRUCE D. SMITH. I993. Isozymic Radiocarbon35:25-33. characterizationof wild populationsof Cucurbitapepo. Journal POWELL, MARY LUCAS. I988. Statusand healthin prehistory: of EthnobiologyI3:55-74. A case studyof theMoundville chiefdom. Washington, D.C.: DUNNELL, ROBERT C. I980. Evolutionarytheory and archaeol- Smithsonian Institution Press. ogy.Advances in ArchaeologicalMethod and Theory3:35-99. . I99I. "Rankedstatus and healthin theMississippian FRITZ, GAYLE j. I990. Multiplepathways to farmingin precon- chiefdomat Moundville,"in Whatmean thesebones? Studies tacteastern North America. Journal of WorldPrehistory in Southeasternbiorchaeology. Edited by MaryLucas Powell, 4:387-435. PatriciaS. Bridges,and Ann MarieWagner Mires, pp. 22-5 I. GARDNER, PAUL S. I987. New evidenceconcerning the chronol- Tuscaloosa: Universityof Alabama Press. ogyand paleoethnobotanyof Salts Cave, Kentucky.American RADFORD, ALBERT E., HARRY E. AHLES, AND C. RITCHIE Antiquity52:358-66. BELL. I968. A manual of the vascularflora of the Carolinas. GREMILLION, KRISTEN J. I990. "Plantremains from ten sites Chapel Hill: Universityof North Carolina Press. withinMammoth Cave NationalPark," in MammothCave SCARRY, C. MARGARET. I993. "Variabilityin Mississippian National Park archaeologicalinventory project interim re- cropproduction strategies," in Foragingand farmingin the port-I989 investigations,by Guy Prentice,pp. 76-96. Talla- Eastern Woodlands. Edited by C. Margaret Scarry,pp. 78-go. hassee: NationalPark Service, Southeast Archaeological Gainesville:University Press of Florida. Center. SCHOENWETTER, JAMES. I974. "Pollen analysisof human pa- . I993a. "Paleoethnobotany,"in The developmentof leofeces fromUpper Salts Cave," in Archaeology of the Mam- Southeasternarchaeology. Edited by JayJohnson, pp. I32-59. mothCave area. Editedby PattyJo Watson, pp. 49-58. New Tuscaloosa: Universityof Alabama Press. York:Academic Press. . I993b. Planthusbandry at the Archaic/Woodlandtransi- SMITH, BRUCE D. I989. Originsof agriculturein easternNorth tion:Evidence from the Cold Oak Shelter,Kentucky. Midconti- America.Science 246:I566-7I. nentalJournal of Archaeology18:I6 I-89. SMITH, BRUCE D., AND C. WESLEY COWAN. I987. Domesti- . I994. "Evidenceof plant domestication from Kentucky cated Chenopodiumin prehistoricEastem North America: caves and rockshelters,"in Agriculturalorigins and develop- New acceleratordates from eastem Kentucky. American Antiq- mentin themidcontinent. Edited by WilliamGreen, pp. 87- uity52:355-57. I04. Officeof the StateArchaeologist, University of Iowa, Re- SMITH, BRUCE D., C. WESLEY COWAN, AND MICHAEL P. port I9. HOFFMAN. I992. "Is it an indigeneor a foreigner?"in Rivers . I995. "Botanicalcontents of paleofeces from two eastem of change:Essays on earlyagriculture in easternNorth Kentuckyrockshelters," in Currentarchaeological research in America,by BruceSmith, pp. 67-IOO. Washington,D.C.: Kentucky,vol. 3. Editedby JohnF. Doershuk,Christopher A. SmithsonianInstitution Press. Bergman,and David Pollack,pp. 52-69. Frankfort,Ky.: Ken- SOB OLIK, KRISTIN D. I988. The importanceof pollen concen- tuckyHeritage Council. trationvalues fromcoprolites: An analysisof southwest Texas GUMERMAN , GEORGE, IV. I994. "Feedingspecialists: The ef- samples.Palynology I2:20I-I4. fectof specializationon subsistencevariation," in Paleonutri- . I993. Directevidence for the importanceof small ani- tion: The diet,health, and nutritionof prehistoric Americans. mals to prehistoricdiets: A reviewof coprolite studies. North Editedby K. D. Sobolik,pp. 80-97. Centerfor Archaeological AmericanArchaeologist I4:227-44. Investigations,Southern Illinois University, Carbondale, Occa- I994. "Introduction," in Paleonutrition:The diet and sional Paper 22. healthof prehistoric Americans. Edited by K. D. Sobolik,pp. HASTORF, CHRISTINE A. I990. The effectof the Inka stateon i-i8. Centerfor Archaeological Investigations, Southern Illi- Sausa agriculturalproduction and cropconsumption. American nois University,Carbondale, Occasional Papers 22. Antiquity55:262-90. SOBOLIK, KRISTIN D., AND DEBORAH J. GERICK. I992. Pre- JOHANNESSEN, SISSEL. I984. "Paleoethnobotany,"in Ameri- historicmedicinal plant usage: A case studyfrom coprolites. can Bottomarchaeology. Edited by CharlesJ. Barreis and Journalof Ethnobiology I2:203-II. JamesW. Porter,pp. I97-2I4. Urbana:University of Illinois SOBOLIK, KRISTIN D., KRISTEN J. GREMILLION, PATRICIA Press. WHITTEN, AND PATTY JO WATSON. I995. Sex determina- KELSO, GERALD. 1976. Absolutepollen frequencies applied to tionand dietaryanalysis of prehistoric human paleofeces. MS. the interpretationof human activities in NorthernArizona. STEWART, ROBERT B. 1974. "Identificationand quantification Ph.D. diss.,University of Arizona, Tucson, Ariz. of componentsin Salts Cave paleofeces,1970-1972," in Ar-

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions Volume 37, Number 3, June1996 1539

chaeologyof theMammoth Cave area. Editedby PattyJo Wat- erally.Among these are concealed ratherthan advertised son, pp. 4I-48. New York:Academic Press. patternsof ovulationand the lack ofperiodic restriction STUIVER, M., AND G. W. PEARSON. I993. High-precisionbi- to sexual activity.In recentyears two othersexual traits decadal calibrationof the radiocarbontime scale, A.D. I950- 500 B.C. and 2500-6000 B.C. Radiocarbon35:I-23. have been given prominence.The firstis that women STUIVER, M., AND P. J. REIMER. I993. Extended14C database can have veryconspicuous blood loss duringmenstrua- and revisedCALIB radiocarboncalibration program. Radiocar- tion, and the second is a tendencyunder some circum- bon 35:2I5-30. stances forthem to synchronisetheir ovulatory cycles SUTTON, MARK Q., MINNIE MALID, AND ANDREW OGRAM. and, in particular,menstruation. These traitShave all I995. Experimentson the determinationof gender from coprolites by DNA analysis.Journal of Archaeological Science. In been ascribedevolutionary significance. press. Knight(I99I) has linked these physiologicaland other WAGNER, GAIL. I978. An archaeobotanicalanalysis of five sites anthropologicalobservations togetherto construct a in theMammoth Cave area. Master'sessay, Department of An- comprehensivemodel forthe role of female sexual be- thropology,Washington University, St. Louis,Mo. WATSON, PATTY JO. I969. The prehistoryof Salts Cave, Ken- haviourin the evolution of modernhuman culturaland tucky.Illinois State Museum Reports of Investigations i6. cognitive systems. His argumentin essence is as fol- . I974. "Prehistorichorticulturists," in Archaeologyof the lows: Human females conceal theirperiod of ovulation. MammothCave area. Editedby PattyJo Watson, pp. 233-38. In contrastto the situationwith many mammals,there New York:Academic Press. . I985. "The impactof earlyhorticulture in theupland is no overt advertisementof when ovulation occurs or drainagesof the Midwestand Midsouth,"in Prehistoricfood when the female is sexually and reproductivelyre- productionin NorthAmerica. Edited by RichardI. Ford,pp. ceptive. Chimpanzees and baboons, for example, both 99-I48. Universityof MichiganMuseum of Anthropology An- have major genital swelling around the time of ovula- thropologicalPapers 75. tion. In contrast,human females have a more marked WHITTEN, PATRICIA. I994. Sexingprehistoric human remains withfecal steroid analysis. Paper presented at the annualmeet- period of menstruation,with relativelyhigh levels of ing of the AmericanAssociation for Physical Anthropology. blood loss, which can give rise to visible signals. This WILLIAMS-DEAN, GLENNA. I978. Ethnobotanyand cultural signal,Knight argues, is the inverseof signallingovula- ecologyof prehistoric man in southwestTexas. Ph.D. diss., tion. Females will be infertile,and so in effectby the Texas A&M University,College Station,Tex. of WILLS, W. H. I988. Earlyagriculture and sedentismin theAmer- high visibility menstrualbleeding they are signalling ican Southwest:Evidence and interpretations.Journal of World to males that conception cannot occur even if thereis Prehistory2:445-88. sexual activity. This gives menstruation,although a YARNELL, RICHARD A. I969. "Contentsof human paleofeces," characteristicof all anthropoidprimates, an additional in The prehistoryof Salts Cave, Kentucky,by PattyJo Watson, significancefor humans the pp. 4I-53. IllinoisState Museum, Reports of Investigations i6. beyond purelyphysiologi- . I974. "Plantfood and cultivationof the Salts Cavers,"in cal. In Knight'smodel this significanceis enhanced by Archaeologyof theMammoth Cave area. Editedby PattyJo the fact that some women have been observedto syn- Watson,pp. II 3-22. New York:Academic Press. chronisetheir menstrual cycles. Some reportshave indi- . I993. "The importanceof native crops during the Late Ar- cated thatunder certain circumstances women livingin chaic and Woodlandperiods," in and Foraging farmingin the close proximitywill in their EasternWoodlands. Edited by C. MargaretScarry, pp. I3-26. converge ovulatorycycles Gainesville:University Press of Florida. in such a way as to end up menstruatingat approxi- matelythe same time (and, by implication,presumably ovulatingsimultaneously as well). Knighthas placed the occurrenceof visible signalling Is ReproductiveSynchrony an and menstrualsynchronisation in an evolutionarycon- text.For much oftheir evolution, human femaleswould EvolutionarilyStable Strategy have lived in small groups in close proximityto each other.Synchronisation of menstrualcycles, possibly re- forHunter-Gatherers?' lated to lunar cycles,would have occurred.The selective pressureleading to this would have been the reproduc- tive advantages R. A. FOLEY AND C. M. FITZGERALD accruingto women when theirreproduc- Human EvolutionaryBiology Research Group, tive behaviourwas coordinated.Women who synchro- Departmentof Biological Anthropology,University of nised theirreproduction within a groupwould have had Cambridge,Downing St. Cambridge CB2 3DZ, U.K. enhanced evolutionaryfitness. In Knight's model this I2 IX 95 would have arisen throughthe followingset of circumstances: These small groupsof women would have regu- The human female has a numberof strikingreproduc- larlybeen signallingto men thatthey were infertile,and tive and sexual characteristics,ones that are rareor un- because this would have been universal within the usual in the contextof mammals or primatesmore gen- groupthere would have been no directreproductive ben- efitto men arisingfrom copulation. Females' signalling infertilityin concert,Knight argues, would have encour- i. ? I996 byThe Wenner-GrenFoundation for Anthropological aged males to increase their economic effortson the Research.All rightsreserved OOII-3204/96/3703-0008$i.oo. We thankChris Knightfor his constructivecomments on an earlier groundsthat therewould have been little advantagein draftof this paper,a versionof whichwas also presentedat the guardingfemales. By reducingmale competitionfor fe- conferencehe organisedon ritualand the originsof culture at the males, the synchronisedregulation of periods of fertility School of Orientaland AfricanStudies, London, in MarchI994. and infertilitywould have acted as a stimulus to male

This content downloaded on Wed, 13 Feb 2013 09:59:12 AM All use subject to JSTOR Terms and Conditions