Identifying low-level food producers: detecting mobility from lithics.

Identifying low-level food producers: detecting mobility from lithics. Introduction Ideas regarding human development from the Pleistocene through theMiddle and Late Holocene are changing due to new data andinterpretations. Instead of a single transition from hunter-gatherer toagriculturalist, adaptation to social, cultural and naturalcircumstances is now conceived as non-linear and more variable thanproposed in the classic nineteenth- and early twentieth-century studies.Societies may occupy a middle ground practising 'low-level foodproduction' with methods of sustenance that are qualitativelydifferent to either hunting or agriculture (Smith 2001). Smithdistinguishes between low-level food producers with and withoutdomesticates. Those with domesticates have economies where thesecontribute less than 30-50 per cent of the total annual calories (Smith2001: 27), while societies without domesticates tended non-domesticatedplants or reshaped the environment to enhance the productivity ofselected species (Smith 2001: 29). The movement within the middle groundof Smith's model is expected to be conceptually nonlinear (Smith2001: Figure 7). Low-level food-producing groups differ from bothhunter-gatherers who, while also demonstrating a variable socio-economy,occupied environments characteristic of the late Pleistocene (Richersonet al. 2001); and from agricultural societies which had typically morethan a 50 per cent dependence on domesticated plants and animals. These groups developed an extremely successful set ofsocio-economic solutions that persisted for long periods of time in awide range of environmental contexts, yet they have been lessrecognised, perhaps because fewer suitable ethnographic analogies exist(Smith 2001: 33-4). Archaeologists, therefore, face a methodologicalchallenge. Botanical and faunal remains are an obvious focus of study,but a lack of preservation of organic remains is a limiting factor.Conventional approaches to the typological definition of societies donot deal with low-level food production societies at all. Tracking low-level food-producers An alternative is to move away from a typological definition andfocus on how these societies dealt with a commonly shared set ofproblems, using the assemblages that are available. If the artefacts canbe related to the degree of mobility within a society, this opens theway for a productive form of cross-cultural comparison. In this paper,we outline an artefact-based approach to determining mobility, acriterion that is thought to shift dramatically between hunter-gathererand agricultural societies but has a high degree of variability inlow-level food production societies (Andrefsky 1994; Torrence 1994;Close 1996, 1999). Mobility can be measured in a variety of ways (Wendrich &Barnard 2008), but as Close (2000) points out, many archaeologicalstudies focus on patterns of movement or mobility strategies rather thanassessing actual indications of movement. The level of generalisationthat results is an impediment to cross cultural comparison (Holdaway& Wandsnider 2006), reducing explanations to ineffective or overlysimplified statements. The alternative is to study the actual record ofindividual movements possible through the use of long distance refits(Close 2000: 52). Refits measure actual rather than potential mobilityand make it possible to determine what was moved and the route taken.Close (1996) was able to consider lithic material recovered from anentire survey area at Bir Safsaf because the conditions lent themselvesto the identification of small, single component assemblages withnumerous refits. Some artefacts were deposited by people moving throughthe area while others were moved within the 15[km.sup.2] surveyed. Despite its analytical utility, refitting has significantmethodological problems. Studies often indicate what was refitted butrarely if ever indicate what was not refitted (Close 2000). It is hardto obtain quantitative estimates of the proportion of an assemblage thatcan be refitted because the ability to 'see' refits isdependent on the observer. The method depends on selecting a suitablysized survey area that reflects the extent of the mobility pattern.Situations like that found at Bir Safsaf are tare archaeologically sincemost archaeological assemblages are derived from palimpsest deposits(Bailey 2007). What is needed is a technique applicable in situationswhere identifying refits is difficult, where record accumulation hasformed palimpsests but where the result indicates mobility over the typeof area studied by Close. Fortunately such a technique has recently been developed based onthe quantification of cortical surface area (Dibble et al. 2005). Wherecortex is present on naturally occurring cobbles, the cortical surfacearea together with their size provides the basis for determining if allthe products of core reduction are present or if some were removed.Equally, it is possible to suggest if artefacts were imported to aparticular location from elsewhere (Douglass et al. 2008). While notallowing the identification of individual movements, the cortical methoddoes provide a quantification of what was moved into, or removed awayfrom, a particular location. The approach uses all assemblage componentsrather than just the cores or retouched tools typically analysed inother studies. Results allow an interpretation of movement closer to theideal defined by Close's refitting work. Here we outline the application of the method in two case studiesfrom different geographic regions and cultural traditions, but where alack of predictablity of resources provided a set of problems which bothsocieties had to overcome. Our first case study comes from western NewSouth Wales (NSW), Australia. Aboriginal people did not developagriculture, but there is a long history of plant-tending andenvironmental modification in Australia (e.g. Cane 1989; Yen 1989).Aboriginal people in western NSW exploited small grass seeds, activitiesoften considered part of the process of domestication (Allen 1974). Theyfall into the category of low-level food producers that lackdomesticates. Artefact scatters frequently include broken seed-grindingimplements while complete grinding dishes were removed by collectors andare now found in museums, as are particular flint tools, such asbifacials. Our interest, however, is in flaked stone artefacts sincethese were much less affected by collectors. When analysed, flaked stoneartefacts may be used to provide a measure of occupation intensity andmobility. Below we outline the relevant methodology based on anassessment of the overall integrity of the assemblage, analysing whatwas abandoned in situ and what was removed and deposited at otherlocations. Our second example comes from the Fayum region of Egypt wherepeople obtained domesticated species by approximately 7000 BP. Thesespecies were introduced to Egypt from the Levant (except cattle) andformed a key component of subsistence economy during later times.However, the introductions added to, rather than replaced, existingsubsistence strategies involving fishing and the hunting of wild game(Wenke et al. 1988: 46). The Fayum example, therefore, is a low-levelfood production society with domesticates. While the Fayum is not partof the Nile Valley proper, its lake is connected to the Nile and wastherefore subject to the same fluctuations during the mid-Holocene(Hassan 1986). Archaeological remains consist largely of surfacescatters of stone artefacts, pottery, bones and hearth features, withthe exception of two mounds of stratified deposits dating to themid-Holocene (6500 BP) that accumulated over approximately 500 years. Dealing with uncertainty: Late Holocene foragers in western NSW Researchers have commented on the extent to which Aboriginal peopleadjusted their socioeconomies to deal with the fluctuating Australianenvironment and a small range of edible plants (e.g. Clarkson 2002;Hiscock 2006). Gould (1991) notes the lack of a single rich staple food.Small seed grasses were exploited; however, studies confirm that netenergy gains from this source were low given the processing costs (e.g.O'Connell & Hawkes 1984). Along the Darling River, seasonalfluctuations supported summer occupation and winter dispersal to theplains and hills, but this seasonality was very much dependent on rainsthat showed marked year-to-year variability (Allen 1974). Palaeo-environmental studies show an environment prone to episodicchange with highly localised effects. Investigations provide evidence ofpast flood events that have eroded the valley floor (Fanning &Holdaway 2001). Radiocarbon age determinations and OSL age estimatesconfirm records of palaeo-floods (Jansen & Brierley 2004; Fanning etal. 2007) and the radiocarbon determinations obtained from hearths shownumerous temporal gaps that are not due to erosion but indicate timeswhen occupation was either very much reduced or non-existent (Holdawayet al. 2002). The hearth and sediment chronologies suggest that artefactassemblages reflect occupations susceptible to climatic extremes: eitherdrought, that limited the ability of groups to inhabit some areas, orfloods, that removed the soil surfaces adjacent to drainage systems, achange that may also have affected the suitability of particularlocations for occupation (Holdaway et al. 2008a). Flakes and cores dominate stone artefact assemblages with a smallerground-stone component. Retouched tools vary in proportion to flakes(Table 1) and assemblages are dominated by scrapers and notched flakesand smaller numbers of specialised tools. Tools are generally lightlyretouched, with intensive retouching only consistently evident on tulaadzes (Moore 2004). Seed-grinding implements account for only a smallproportion of the overall assemblage count and are often broken. Twotypes of raw material, silcrete and quartz, dominate assemblages. Lithicsources occur as outcrop, streambeds and stone pavement (Gibber) withthe size, form and quality varying considerably between sources. Cortexcovers cobbles from both the stone pavement and outcrop while quartzartefacts also exhibit a rounded cortex characteristic of procurementfrom stone pavements or streambeds. Quartzite and sandstone sometimesoccur within assemblages, but in low proportions. Spatially extensive surface stone artefact distributions alongvalley floors are abundant with assemblage size declining whereversediment and vegetation obscures the surface (Fanning & Holdaway2004). Assemblage composition varies both spatially (Shiner et al. 2007)and chronologically (Holdaway et al. 2008b). Raw material access has adetectable effect, but it is not on its own sufficient to determine allaspects of assemblage composition. Instead, it can be argued thatvariation in occupation duration explains changes in assemblagecomposition largely because artefact deposition is a function of periodof use, in addition to the type of use. As time passes, retouching offlakes forms more tools and these tools acquire more retouching throughresharpening (Hiscock & Clarkson 2005). Prolonged occupation in onelocality leads to the greater use of locally abundant materials sincethere are fewer opportunities to procure stone from distant sources.Consequently, the proportions of artefacts made from local materials aregreater compared to those made from non-local raw materials and theimported materials are worked more intensively. Using these measures,analyses indicate that in some locations surface assemblages separatedby only a few hundred metres have different occupation durations whileat other locations, the situation is more complex (Holdaway et al.2004). At Fowlers Gap, for instance, a series of measures suggests acomplex occupation history where the assemblage composition representsmaterials dropped as a result of not one but many separate occupationsof different durations (Holdaway et al. 2008a). Individual locationsshow little redundancy, in the sense that places attracted the same setof activities and, while analysis of radiocarbon determinations fromhearths indicates reoccupation of places, this was intermittent. To survive in a highly variable environment, people must cope witha variety of risks and uncertainties. Bad years can lead to catastrophicpopulation losses, so activities are geared to forma type of buffer toalleviate risk. One of these is to reduce the time needed to locatestone through the curation of flakes, thus providing access to a sharpedge whenever needed. A number of authors have discussed the differencebetween provisioning individuals and provisioning places with eitherartefacts or materials (e.g. Kuhn 1994; Clarkson 2006). Provisioningpeople provided the means to exploit plant and animal resources whenencountered. Raw material in the region is widely available, thereforemoving raw material away from sources makes little sense. Equally,provisioning of places makes little sense where availability of foodresources is unpredictable. For foragers, the time it takes to find thefood source and manufacture an artefact is inhibiting (Kuhn 1995).Provisioning people with artefacts mitigates the risk implied by thistime stress. The analysis of cortex introduced above provides a means ofdetermining whether artefacts of any form or raw material weretransported irrespective of the presence of retouched artefacts or thelocality from where materials were acquired (Dibble et al. 2005). The method involves three steps. Total assemblage weight is dividedby the mass of the nodules from which cores were flaked, giving anestimate of the number of nodules reduced. Next, nodule frequency(estimated by number of cores) is multiplied by the total surface areaof the nodules. This provides the expected cortical surface area in anassemblage. Finally, this value is compared to the actual quantity ofcortex observed in the assemblage and expressed as a ratio. A slightlymodified version of the method applied to assemblages from western NSWreturns values consistently below one, indicating that cortex isunder-represented. Conversely, there is an over representation ofnon-cortical flakes and cores in assemblages probably because of theselective removal of large blanks that tend to have cortex on theirdorsal surface (Douglass et al. 2008). While transporting cores provides greater flexibility, transportingflakes provides the greatest efficiency in mass, relative to useablecutting edge (Kuhn 1994). Analyses of cortex strongly points to thetransport of flakes and a concern to have a cutting edge at hand in itsmost economical form. Curating stone in an economically efficient formin a lithic rich landscape seems incongruous; however the archaeologicalevidence makes more sense if the environment is consideredresource-impoverished. Resources were exploited as they wereencountered, hence the need to transport flakes already struck from thecore. The resulting artefacts have characteristics consistent withprovisioning people rather than places. Imported material is worked moreintensively than materials that are locally available, coresmanufactured on imported material are smaller and show more evidence ofrotation, the flake to core ratio is often higher, and tools made fromimported materials show relatively more retouching and more utilisededges (Holdaway et al. 2004; Shiner et al. 2007). In an unpredictable environment with a limited range of resources,behavioural strategies counter the uncertainty of food resourceavailability. Certainly mobility involving the transportation of stoneartefacts, the efficient use of raw materials, the concern withprovisioning people, and the need to exploit even the least efficientenergy sources would fit with the need to counter such uncertainty.Responses involved a range of technological and social solutionsalthough domestication was not among them. Given the level ofuncertainty in the environment, the low net gains from grass seedexploitation and the technological responses involving transport offlakes, Aboriginal people in western NSW occupy a particular position inthe range of low-level food production solutions. Comparison with othersocieties should attempt to account for the similarities and differencesbetween one group and the other. Dealing with uncertainty by enhancing predictability: NeolithicEgypt In the Fayum, several researchers from the late nineteenth centuryonwards have noted Neolithic remains. Diffusion of a 'NeolithicPackage', including Emmer wheat, barley, sheep, goat and pig, aswell as a very limited number of stone artefact types and ceramics, wasfrom the Levant (e.g. Caton-Thompson & Gardner 1934). However, inthe eastern Sahara, Levant-like sedentism was not viable in anenvironment with fluctuating aridity levels. Cattle herding facilitatedseasonal movement between the Nile Valley and the Western Desert(Marshall & Hildebrand 2002), while wild food resources,particularly lake resources in the Fayum, remained the most dominantcomponent (Wenke et al. 1988: 46). Pastoralism in this region was aresponse to changing climatic conditions, particularly fluctuationsbetween arid and wet phases during the early Holocene. People required amore predictable food source when other wild game and plant species werenot guaranteed (Marshall & Hildebrand 2002: 107, 111). This suggestslow-level food producers with some domestication. Neolithic stone artefacts, bone and pottery scatters occur on thesurface across the northern shore of Lake Qarun. In addition, twostratified sites (Kom W and Kom K) contain ceramic vessels, hearthfeatures and middens but no architectural features except shallow pits,a few postholes and some clay floors. Carbonised domesticated wheat andbarley have recently been found at Kom K, as well as sheep/goat dungfrom hearth contexts. Close to Kom K are the K-pit granaries excavatedby Caton-Thompson and recently re-examined (Wendrich & Cappers2005). Both Kom K and Kom W fall into a narrow period of the FayumNeolithic around 6500 BP, while the Upper K pits give slightly morerecent dates, up to 6200 BP. Owing to the topography of the lake basin, cultivation of cerealswas possible either in a narrow band around the lake, where the watertable was high enough to negate the necessity for artificial irrigation(Wenke et al. 1988: 39), or in lower wadi regions, dependent on winterrain and irregular wadi flows. The lake fluctuated during the annualNile inundation, the level of which was dependent on Ethiopian rainfall(Hassan 1986). The unpredictable nature of lake fluctuations placedlimits on the duration of occupation and may have required movement todifferent field locations each year (Wenke 1999:314). Therefore, theFayum probably required significant adaptation of the LevantineNeolithic strategies to the local environment (Wenke & Casini 1989:144). As with the Aboriginal people, those who made use of the Fayumadopted a range of responses to environmental uncertainty that includeda certain level of mobility. A direct measure of mobility is neededsince the economic evidence is intermediate between the extremes ofLevantine sedentism and African pastoralism. Caton-Thompson categorised the Fayum Neolithic stone assemblagebased on the morphology of complete tools only. She concluded that 40per cent of the tool assemblage consisted of chipped and polished flintaxes, with the remainder made up of bifacial projectile points, sickleblades, borers and other tools in low percentages (Caton-Thompson &Gardner 1934: 19-22). Typological similarities in bifacial projectilepoints, core axes and sickle blades suggested a connection with the siteof Merimde Beni-Salama (Caton-Thompson & Gardner 1934: 89-92).However, collection biases distorted the role of bifacial tools in theassemblages. Full assemblage analysis on a sample from the site of Kom Wsuggests that finished tools constitute less than five per cent of thetotal (Kozlowski & Ginter 1989). Although the locations of material sources are not well understoodfor the Fayum, flint was brought into the area as cortical pebbles.Wenke et al. (1988) suggest that during the Neolithic some localised rawmaterial sources on the lake margin were covered by the higher lakelevel. Whatever the source, cortex present on flakes suggests that flintpebbles were imported to the Fayum rather than decortified cores. Table 2 provides flake to tool ratios based on recently collectedsamples from the two Kom sites and XB11, a surface artefact scatter lessthan l km from the X Basin palaeo-shoreline (Phillipps 2006). Ali threevalues of the flake to tool ratio are high when compared to most of theAustralian sampling locations. Tool collection, particularly at Kom Wwhere the sample is based on a surface collection made in 2006, probablyhad an impact on the flake to tool ratio since earlier researchers, andcollectors removed tools but not flakes. This bias is less of an issuehowever, for the cortex ratio values (Table 2) since the calculationsinvolve measurements taken from flakes and cores which considerablyoutnumber tools (Douglass et al. 2008). The highest values for thecortex ratio are found for the two Kom assemblages with the XB11assemblage showing a lower value but still one that is high compared tothe majority of the Australian samples (Table 1). Higher values suggestlower levels of flake movement since ratios closer to one indicate thepresence of sufficient cortex to account for the number of nodulesworked. To the extent that movement of flakes equates with the movementof people, a higher value for the cortex ratio suggests lower mobilitylevels. The Fayum lithic analysis suggests two things. Firstly, as with thesituation in Australia, there is evidence that flakes manufacturedwithin an 'expedient' technology were curated and that flakeswith cortex were differentially removed from all the Fayum assemblagesstudied. This indicates that the people who occupied lakeside locationsduring the Neolithic were mobile to some degree. Secondly, a differencebetween the Kom sites and XB11 indicates variability in the degree ofmobility over time. Fayum lake-edge resources were probably available tosome degree all year round, although they probably fluctuated inabundance from year to year. Nevertheless, people selected flakes fromthe Fayum assemblages, including the stratified Kom deposits and movedthem sufficient distance away so that they did not return. A cortexratio less than one indicates not only that large flakes (with somecortex) were removed but also that large cortical flakes were notintroduced from elsewhere. Were this to happen with sufficientfrequency, the cortex ratio would of course have a value of one. That itdoes not happen at three substantial Fayum sites indicates that movementwas of sufficient distance to remove flakes from the area exploitedaround the lake. Conclusion Both of the examples discussed are low-level food producers, albeitin very different contexts. Whereas Aboriginal people reacted tounpredictability by equipping individuals with transportable, highlyefficient stone flakes and moving between resource patches as theybecame available, the people of Egypt ensured access to food by movingsome resources and establishing their own patches through the use ofdomesticates. Both societies attempted to cope with environmental uncertaintiesand challenges. Each of the environments is different, but both arelinked by the challenge provided by an inability to predict the natureof environmental variability and therefore resource availability. AsSmith (2001) suggests, one should expect a range of socio-economicsolutions from low-level food producers. In Australia, faced with a unpredictable environment and a resourcebase that was poor in comparison to that available to foragers in otherparts of the world, people reacted by extensive and frequent movementaided by an extremely efficient mobile technology, one that permitteduse of resources on an encounter basis in patches often separated bylarge distances. In the Fayum much has been written on the potential ofthe environment to sustain long terra permanent occupation with perhapsseasonal movement around the lake edge. But although results indicate alesser level of artefact movement than in Australia, it was stillsufficient to deplete the cortex below one. Although the Fayumenvironment appears more predictable than that in Australia, regularlake inundation was controlled by rains in Ethiopia, somethingunobservable to the local inhabitants. Like the Australians, part of theresponse to this unpredictability by the people who occupied the Fayumwas large scale movement. In both cases, movement was sufficient toremove flakes in a quantity and with a frequency to ensure that cortexwas never replenished. Local movement in both cases would, over time,have moved the cortex ratio back to a value of one. The utility of the cortex ratio in both the Australian and Egyptiansituations is that it measures actual mobility in the sense that Closeuses it, but does so in a quantifiable way based on whole artefactcollections. Issues with observer bias are reduced, as are the problemswhich arise from the differential removal of certain tool forms. AsClose correctly emphasised, the only test of assessments ofenvironmental potential is actual evidence of movement. Refittingcertainly indicates such movement, but so too does use of the cortexratio with the added advantages of reduced observer bias. It provides auseful mechanism for exploring the range of variability thatSmith's low-level food production designation suggests shouldexist. Acknowledgements Melinda Allen, Matthew Douglass and Harry Allen provided usefulcomments on earlier drafts. Research was supported by the NationalGeographic Society, the National Geographic Committee for Research andExploration Young Explorers Grant, University of Auckland PartnershipAppeal Award and the University of Auckland, Department of Anthropologyresearch fund. Thanks to students of the University of California, LosAngeles, Leiden University, the University of Groningen, and TheUniversity of Auckland who assisted with survey and analysis. Received: 6 October 2008; Revised: 3 July 2009; Accepted: 7 July2009 References ALLEN, H. 1974. The Bagundji of the Darling Basin: cereal gatherersin an uncertain environment. World Archaeology 5: 309-322. ANDREFSKY, W. 1994. Raw-material availability and the organizationof technology. American Antiquity 59: 21-34. BAILEY, G.N. 2007. Time perspectives, palimpsests and thearchaeology of time. Journal of Anthropological Archaeology 26: 198-223. CANE, S.B. 1989. Australian Aboriginal seed grinding and thearchaeological record, in D.R. Harris & G.G. Hillman (ed.) Foragingand farming: the evolution of plant exploitation: 99-119. London: UnwinHyman. CATON-THOMPSON, G. & E.W. GARDNER. 1934. The Desert Fayum.London: The Royal Anthropological Institute of Great Britain andIreland. CLARKSON, C. 2002. An index of invasiveness for the measurement ofunifacial and bifacial retouch: a theoretical, experimental andarchaeological verification. Journal of Archaeological Science 29: 65-75 --2006. Lithics in the land of the lightning brothers: thearchaeology of Wardaman Country, Northern Territory. Canberra: ANUE-Press. CLOSE, A.E. 1996. That weight: the use and transportation of stonetools. Current Anthropology 37(3): 545-53. --1999. Distance and decay: ah uneasy relationship. Antiquity 73:24-32. --2000. Reconstructing movement in prehistory. Journal ofArchaeological Method and Theory 7: 49-77 DIBBLE, H.L., U.A. SCHURMANS, R.P. IOVITA & M.V. MCLAUGHLIN.2005. The measurement and interpretation of cortex in lithicassemblages. American Antiquity 70: 545-60. DOUGLASS, M.J., S.J. HOLDAWAY, P.C. FANNING & J.I. SHINER.2008. AH assessment and archaeological application of cortex measurementin lithic assemblages. American Antiquity 73: 513-26. FANNING, P.C. & S.J. HOLDAWAY. 2001. Temporal limits to thearchaeological record in arid western NSW, Australia: lessons from OSLand radiocarbon dating of hearths and sediments, in M. Jones & P.Sheppard (ed.) Australasian connections and new directions: proceedingsof the 7th Australasian archaeometry conference (Research inAnthropology and Linguistics 5): 91-111. Auckland: University ofAuckland. Department of Anthropology. --2004. Artifact visibility at open sites in western New SouthWales, Australia. Journal of Field Archaeology 29: 255-71. FANNING, P.C., S.J. HOLDAWAY & E. RHODES. 2007. A geomorphicframework for understanding the surface archaeological record in aridenvironments. Geodinamica Acta 20(4): 275-86. GOULD, R.A. 1991. Arid-land foraging as seen from Australia:adaptive models and behavioral realities. Oceania 62: 12-33. HASSAN, EA. 1986. Holocene lakes and prehistoric settlements of theWestern Faiyum, Egypt. Journal of Archaeological Science 13:483-501. HISCOCK, P. 2006. Blunt and to the point: changing technologicalstrategies in Holocene Australia, in I. Lilley (ed.) Archaeology inOceania: Australia and the Pacific Islands: 69-95. Sydney: Blackwell. HISCOCK, P. & C. CLARKSON. 2005. Experimental evaluation ofKuhn's Geometric Index of Reduction and the flat-flake problem.Journal of Archaeological Science 32:1015-22. HOLDAWAY, S.J., J.I. SHINER & P.C. FANNING. 2004.Hunter-gatherers and the archaeology of discard behavior: an analysis ofsurface stone artifacts from Sturt National Park, western New SouthWales, Australia. Asian Perspectives 43: 34-74. HOLDAWAY, S.J., P.C. FANNING, M. JONES, J. SHINER, D. WITTER &G. NICHOLLS. 2002. Variability in the chronology of late HoloceneAboriginal occupation on the arid margin of southeastern Australia.Journal of Archaeological Science 29: 351-63. HOLDAWAY, S.J., P.C. FANNING & E. RHODES. 2008a. Assemblageaccumulation as a time dependent process in the arid zone of western NewSouth Wales, Australia, in S. Holdaway, S.J. & L. Wandsnider (ed.)Time in archaeology: 110-33. Sair Lake City (UT): University of UtahPress. --2008b. Challenging intensification: human--environmentinteractions in the Holocene geoarchaeological record from western NewSouth Wales, Australia. The Holocene 18: 411-20. HOLDAWAY, S.J. & L. WANDSNIDER. 2006. Temporal scales andarchaeological landscapes from the Eastern Desert of Australia andIntermontane North America, in G. Lock & B.L. Molyneaux (ed.)Confronting scale in archaeology: 183-202. New York: Springer Science& Business Media. JANSEN, J.D. & G.J. BRIERLEY. 2004. Pool-Fills: a window topalaeoflood history and response in bedrock-confined rivers.Sedimentology 51: 901-925. KOZLOWSKI, J.K. & V.B. GINTER. 1989. The Fayum Neolithic in thelight of new discoveries, in L. Krzyzaniak & M. Kobusiewicz (ed.)Late Prehistory of the Nile Basin and the Sahara: 157-79. Poznan: MuzeumArcheologiczne W. Poznaniu. KUHN, S.L. 1994. A formal approach to the design and assembly ofmobile toolkits. American Antiquity 59: 426-42. --1995. Mousterian lithic technology: ah ecological perspective.Princeton (NJ): Princeton University Press. MARSHALL, F. & E. HILDEBRAND. 2002. Cattle before crops: thebeginning of food production in Africa. Journal of World Prehistory 16:99-143. MOORE, M. 2004. The Tula Adze: manufacture and purpose. Antiquity78: 61-73. O'CONNELL, J.F. & K. HAWKES. 1984. Food choice andforaging sites among the Alyawara. Journal of Anthropological Research40: 504-535. PHILLIPPS, R.S. 2006. Neolithic surface stone artefact assemblagesfrom the Fayum, Egypt. Unpublished MA dissertation, University ofAuckland. RICHERSON, P.J., R. BOYD & R.L. BETTINGER. 2001. Wasagriculture impossible during the Pleistocene but mandatory during theHolocene? Aclimate change hypothesis. American Antiquity 66: 387-411. SHINER, J.I., S.J. HOLDAWAY, H. ALLEN & P.C. FANNING. 2007.Burkes Cave and flaked stone assemblage variability in western New SouthWales, Australia. Australian Archaeology 64: 35-45. SMITH, B. 2001. Low-level food production. Journal ofArchaealogical Research 9: 1-43. TORRENCE, R. 1994. Strategies for moving on in lithic studies, inP.J. Carr (ed.) The organization of North American prehistoric chippedstone tool technologies (International Monographs in Prehistory,Archaeological Series 7): 123-31. Ann Arbor (MI):, InternationalMonographs in Prehistory. WENDRICH, W. & R.T.J. CAPPERS. 2005. Egypt's earliestgranaries: evidence from the Fayum. Egyptian Archaeology 27: 12-15. WENDRICH, W. & H. BARNARD. 2008. The archaeology of mobility:definitions and research approaches, in H. Barnard & W. Wendrich(ed.) The archaeology of mobility: Old World and New World nomadism:1-21. Los Angeles (CA): Cotsen Institute of Archaeology. WENKE, R.J. 1999. Fayum, Neolithic and Predynastic sites, in K.A.Bard (ed.) Encyclopaedia of the archaeology of ancient Egypt: Volume 2.New York: Routledge. WENKE, R.J. & M. CASINI. 1989. The Epipaleolithic-Neolithictransition in Egypt's Fayum depression, in L. Krzyzaniak & M.Kobusiewicz (ed.) Late prehistory of the Nile Basin and the Sahara:139-55. Poznan: Poznan Archaeological Museum. WENKE, R.J., J.E. LONG & P.E. BUCK. 1988. Epipaleolithic andNeolithic subsistence and settlement in the Fayyum oasis of Egypt.Journal of Field Archaeology 15:29-51. YEN, D.E. 1989. The domestication of environment, in D. Harris& G. Hillman (ed.) Foraging and farming: the evolution of plantexploitation: 55-75. London: Unwin Hyman. Simon Holdaway (1), Willeke Wendrich (2) & Rebecca Phillipps(1) (1) Anthropology Department, University of Auckland, Private Bag92019, Auckland, New Zealand (sj.holdaway@auckland.ac. nz;rphi@auckland.ac.nz) (2) Department of Near Eastern Languages and Cultures, UCLA, 397Humanities Building, 415 Portola Plaza, Los Angeles, CA 90095, USA(wendrich@humnet.ucla.edu)Table 1. Ratio of flakes to tools and the cortex ratio for selectedwestern NSW silcrete assemblages (Douglass et al. 2008;Holdaway et al. 2008a). Burkes Nundooka Mulga DamFlake to tool ratio 8.2 5.6 5.8Cortex ratio 0.1 0.3 0.2Table 2. The ratio of flakes to tools and the cortex ratiofor selected Fayum assemblages (Phillipps 2006). Kom K Kom W XB11Flake to tool ratio 25.8 41.5 16.6Cortex ratio 0.5 0.5 0.4