Res

earc

h

Grinding flour in Upper PalaeolithicEurope (25 000 years bp)Biancamaria Aranguren1, Roberto Becattini2, Marta Mariotti Lippi2

& Anna Revedin3

The authors have identified starch grains belonging to wild plants on the surface of a stone from theGravettian hunter-gatherer campsite of Bilancino (Florence, Italy), dated to around 25 000 bp.The stone can be seen as a grindstone and the starch has been extracted from locally growingedible plants. This evidence can be claimed as implying the making of flour – and presumablysome kind of bread – some 15 millennia before the local ‘agricultural revolution’.

Keywords: Gravettian, cat’s tail, flour, grindstone, reedmace, starch, Typha

IntroductionBy virtue of the surviving evidence, the diet of Palaeolithic people has been consideredprimarily carnivorous, and there has been little evidence for the exploitation of plants. Theresearch reported here has identified the remains of starches from Typha and Graminaecf. Brachypodium found on the surface of a Gravettian period grindstone, and draws theconclusion that parts of these plants were being ground into a flour. The survival of tracesof perishable starches on such an ancient grindstone suggests an intensive and repeateduse. The routine production of a Palaeolithic flour has at least two important implications.Firstly, a food was available which was portable, and had high energy content and goodproperties of conservation, so allowing hunter-gatherers a greater mobility and autonomy.Secondly, the technical ability to obtain flour from wild plants in the Upper Palaeolithicchallenges the idea that the exploitation of wild cereals necessarily began in Pre-Neolithicwestern Asia.

Archaeological contextThe Bilancino settlement in the Mugello area (north-west of Florence, Italy) was discoveredin 1992 during the digging of an artificial lake (Aranguren & Revedin 1998), and excavated in1995-96 by the Archaeological Department of Tuscany (Figure 1). The excavation broughtto light an occupation surface of 120m2, from which 15 000 artefacts were recovered(Aranguren et al. 2003a). Among about 1600 tools, 65 per cent were classic Noailles burinswith fine flaking and notches (Figure 2). The burins were obtained from different groups

1 Archaeological Department of Tuscany, Florence, Italy (Email: arangurenb@yahoo.it)2 Department of Plant Biology, Florence University, Italy (Email: robertobecattini@yahoo.it; mariotti@unifi.it)3 Italian Institute of Prehistory & Protohistory, Florence, Italy (Email: annarevedin@iipp.it)

Received: 17 July 2006; Accepted: 6 September 2006; Revised: 22 September 2006

antiquity 81 (2007): 845–855

845

Grinding flour in Upper Palaeolithic Europe (25 000 years bp)

Figure 1. The Bilancino settlement: location in central Italy.

of raw materials, both local and exotic(Aranguren et al. 2004b), and the chaıneoperatoire, from knapping to discard, hasbeen worked out in detail (Aranguren& Revedin 2005). The high degreeof standardisation in the lithic industrysuggests the use of specialised tools for aprocess which demanded high precision.

On the basis of the lithic industry, thesettlement of Bilancino has been ascribedto the Gravettian culture (Noailles facies)(Aranguren & Revedin 2001). Among thethree Bilancino AMS dates (Aranguren &Revedin 1997), the most reliable is the

oldest one, taken from a domestic hearth, of 25 410 +− 150 bp (Beta – 106549; Arangurenet al. 2001). Palaeoecological analysis suggests that the Bilancino site was a summer camplocated in a damp situation, with abundant wetland plants, such as Cyperaceae and reedmaceor cat’s tail (Typha) (Figure 3) (Aranguren et al. 2003b). Use-wear analysis on Noailles burinshas identified the presence of organic residues located on the active edge of the tool. Chemicalanalysis performed on the residues indicates the presence of carbon and chlorine, suggestingorganic materials of vegetal origin (Aranguren et al. 2007). A series of striations and polishingtraces were also noted on the burins. These could be replicated experimentally by 40-90minutes spent cutting Typha and separating its fibres.

The grindstoneAnalysis has also been applied to the corpus of pebbles found on the site (Arangurenet al. 2006; in press). The most important of these was a piece of sandstone measuring13.6 × 9.7 × 6.2cm with one long side flat and the other dished to a maximum depth ofaround 0.8mm. It had originally been joined to a smaller piece found 1.15m away (Figure 4),the two together measuring 25 × 9.7 × 6.2cm (Figure 5). The original stone – a quartz-feldspar-micaceous sandstone of local origin – had been broken in two with a blow onthe edge. The resulting larger fragment had the concave depression on its upper surface,indicating that it had been used as a grindstone, while the smaller no doubt functioned as thegrinder.

Grindstones are sporadically attested from the Aurignacian, but it is from the Gravettianthat these artefacts become more frequent. The earliest examples have been identified fromwear traces rather than their shape (De Beaune 2000), but those from the Epipalaeolithic andthe Mesolithic have a decidedly concave surface. Wear traces revealing an axial movementback-and-forth (like that from Bilancino) were observed in some Natufian specimens fromthe Levant (12 000-10 500 bp: Wright 1991), and classified as Trough Quern type A.7(Wright 1992: Figures 4, 7). Their asymmetrical form is intentional, flour being gatheredfrom the open side in a lower container (De Beaune 2000; 2004).

846

Res

earc

h

Biancamaria Aranguren et al.

Figure 2. Noailles burins from the Bilancino settlement (Scale 1:1).

ArchaeobotanyPollen analyses (Mariotti Lippi & Mori Secci 2002; Aranguren et al. 2003b) showedthat the very scanty treecover was mainly represented by pine (Pinus), whose grains

847

Grinding flour in Upper Palaeolithic Europe (25 000 years bp)

Figure 3. Typha latifolia growing in the Bilancino area.

are probably attributable to P. sylvestris,a tree which was widespread in centralItaly during the glacial period. In the mostancient sediments, pine was accompaniedby birch (Betula) and by willow (Salix)and alder (Alnus) the latter being treeswhich grow in wet environments. Heather(Erica) was also present. Among the her-baceous plants, Gramineae, Compositae,Chenopodiaceae and Cyperaceae were wellrepresented, and there was a significant per-centage of helophytes (Typha, Sparganiumerectum type,cf. Scirpus) (Figure 6).

With the aim of verifying if plantremains were present on the grindstone,a thin layer of mud was removed froman area of less than 5cm2, using distilledwater. The residue was observed under alight microscope. To ascertain the presenceof starch grains, observation under a

polarising microscope and IKI methodology (Jensen 1962) were used. Starch was extractedaccording to Barton et al. (1998), and identified using literature (Tateoka 1962; Seidemann1996) and fresh reference material. The reference specimens were suggested by the palaeo-ecological data, especially the pollen spectra; they included grasses, herbs, and ferns whichcould be used as food supply (Kunkel 1984), particularly caryopses of Bromus secalinus L.and Brachypodium ramosum L., seeds of Scirpus lacustris L. and rhizomes of Sparganiumerectum L., Typha latifolia L. and Typha angustifolia L. Phytolith recovery was also attempted(following Pearsall et al. 2004).

Starch grains were found on both the grindstone and the grinder (Figure 7: a-g; Table 1).They were very different in size and shape, and morphology of the hilum, but only two typeswere tentatively identified at genus level. Circular or elliptical single grains with a Y-shapedhilum and ranging between 13 and 18µm in size may be attributed to Typha (Typhaceae).Grains ranging between 7 and 18µm, but with no evident hilum might belong to the sameplant. Angular irregular grains 9 and 14µm in size, with no evident hilum, may belong toBrachypodium (Gramineae), a common wild grass, or to a related plant. Phytoliths were onlyrepresented by an elongate smooth one, about 40µm long and a dubious Chloridoid oneaccording to Pearsall’s (1989) classification (Figure 7: h). The Chloridoid class morphologyoccurs in Chlorideae, Eragrosteae and Sporoboleae. The last two are currently representedin Italy by a few species, some of them growing in wet environments.

DiscussionCompared to the evidence from animal bones, the evidence for plant use in Palaeolithicsites has been scarce and indirect, such as the study of the microwear on fossil human teeth

848

Res

earc

h

Biancamaria Aranguren et al.

Figure 4. Location of the grindstone (1) and its related fragment (2) on the Bilancino occupation surface.

Figure 5. Refitting of the grindstone and fragment (grinder) (Scale 2:1).

(Lalueza et al. 1996). Increasing numbers of researchers are using residues on stones tosearch for information about food preparation and other types of behaviour (Piperno &Holst 1998; Kealhofer et al. 1999; Piperno et al. 2004; Torrence et al. 2004). At Bilancino,the ability to produce flour from plants in the Upper Palaeolithic has been demonstrated bya variety of evidence: the morphology of the grindstone, use-wear analysis (Aranguren et al.

849

Grinding

flourin

Upper

PalaeolithicEurope

(25000

yearsbp)

Figure 6. Pollen diagram from the Bilancino archaeological site.

850

Res

earc

h

Biancamaria Aranguren et al.

Figure 7. a, b, Particles of modern Typha rhizome with starch in the cells of the storage parenchyma. c-h, Starch and phytolithgrains recovered from grindstones: c, e, starch grains attributable to Typha as seen under a light microscope; d, f, the same inpolarised light; g, grain attributable to Typha colored with IKI methodology; h, phytolith belonging to the Chloridoid type(Scale bars, 10µm).

851

Grinding flour in Upper Palaeolithic Europe (25 000 years bp)

Table 1. Morphological and dimensional features of the examined starch grains.

Dimensional Main NumberShape range (µm) dimension (µm) Hilum of grains

Fresh materialBromus secalinus L. Circular-elliptic 4-10 Not evident 50Brachypodium ramosum Angular 3-16 Not evident 50

(L.) R. et S.Scirpus lacustris L. Angular 3-9 Not evident 50Sparganium erectum L. Angular 4-16 Not evident 50Typha angustifolia L. Circular-elliptic 4-18 Y shaped 50Typha latifolia L. Circular-elliptic 3-18 Y shaped 50

Archaeological findings – possible attributioncf. Brachypodium sp. Angular 9; 14 Not evident 2Indet. Angular 18; 24 Not evident 2cf. Typha sp. Circular-elliptic 7-18 Not evident 10Indet. Circular-elliptic 19-34 Not evident 13cf. Typha sp. Circular-elliptic 13-18 Y shaped 5Indet. Circular-elliptic 19; 19 ; 25 Y shaped 3Indet. Reniform 22; 23 ; 27 ; 27 Not evident 4Indet. Reniform 27 V shaped 1Indet. Elliptic 16; 19; 23 Linear 3Indet. Circular-elliptic 24 X shaped 1Indet. Angular 23 Star 1Indet. Angular 10; 18; 32 Compound 3

in press), the finding of the grinder, and especially by residues of starch from different plantsfound on the grinder and grindstone (Typha and wild Gramineae). This discovery encouragessystematic organic micro-residue recovery on Palaeolithic sites, particularly on artefacts thatare potentially involved in the treatment of organic substances or in the preparation of food,like grindstones, hearthstones and silica tools (Fullagar et al. in press).

The variety of the starch grains recovered from the Bilancino grindstone suggest thatit was used for processing more than one kind of plant, with rushes (Typha) and grasses(Gramineae) predominating. It is possible that different plants were exploited as they cameinto season. The dry season (before the rains began) corresponds to the most suitable periodfor gathering Typha leaves, although the same period is not the best for gathering rhizomes.

Ethnographical comparisons reveal the use of Typha for different purposes all over theworld particularly in non-European countries from China to Australia (Gott 1999) andNorth America. Almost every part of the plant is utilised – buds, spikes, pollen, but aboveall the raw rhizomes. They are sun-dried and ground to obtain flour: in this way the productmaintains better qualities of nourishment. The flour is similar to cereal flour, both in tasteand in nourishment value (content of Typha (mg/100g): energy 277Kj, water 69.9, protein2.8, fat 0.1, carbohydrate 14.1, fibre 12.2, ash 0.9 (Gott 1999). Other uses include theweaving of fibres and as material for construction (Aranguren & Revedin in press). In thewet regions of the Po delta, Typha plants are used for weaving cordages, mats and baskets(Aranguren et al. 2004a). The leaves are harvested and the fibres separated with a metal tool,

852

Res

earc

h

Biancamaria Aranguren et al.

Figure 8. A-B, Noailles burins from Bilancino; C, Arfinden, a metal tool used in historic times to harvest Typha in the Podelta (Scale 1:1).

the arfinden (parting tool), which bears some morphological resemblance to the Noaillesburin, the artefact most represented in the Bilancino site (Figure 8).

The earliest evidence for the use of wild plants as food currently comes from the siteof Cuddie Springs in Australia (Fullagar & Field 1997), where remains rich in starch andsilica have been found on various grindstone fragments dating to around 27 000 years BP(Fullagar et al. in press). At the site of Dolni Vestonice (26 000 BP) the remains of tap roots,probably of the Compositae family, have been identified as a potential food-plant material(Mason et al. 1994; nd). The most ancient remains of wild cereals (grasses) have been foundon grindstones in the Early Epipalaeolithic pre-Natufian, in the site of Ohalo II in Israel,dated to around 19 500 BP (Piperno et al. 2004).

ConclusionThe interdisciplinary and integrated research at Bilancino has revealed a seasonal UpperPalaeolithic site processing local reeds and grasses, most probably for food to be consumedas a product made from flour. This implies the availability of an elaborate product, a flour,with high energy content, that is rich in carbohydrates, easily storable and transportable,to make a kind of bread (biscuits) or a porridge. Together with other vegetal resources,such as seeds and fruits, such portable foodstuffs would allow travel over long distancesand greater independence from a home territory. Traditionally, agriculture has its origins inthe cultivation of wild cereals in western Asia. But the evidence for the processing of otherwild plants in 25 000 BP raises the possibility of alternative and independent trajectoriestowards the preparation of elaborate vegetable foods many millennia before their routineproduction by farmers. The Upper Palaeolithic technological and cultural innovations,like weaving, ceramic technology, stockage and sedentary life, anticipate (according toKozlowski 2002) the ‘Neolithic Revolution’ and seems to have a polycentric origin inEastern and Mediterranean Europe, now with the new evidence offered by the discovery offlour production technique in the Italian Gravettian.

853

Grinding flour in Upper Palaeolithic Europe (25 000 years bp)

Further investigations should bring new understanding to the historical study of nutrition,with important implications for modern populations. The so-called ‘diseases of civilisation’have been related to the incomplete evolutionary adaptation to some foods introduced withthe spread of agriculture (Cordain 2002).

NoteThese authors contributed equally to this work: B. Aranguren and A. Revedin for thearchaeological research, R. Becattini and M. Mariotti Lippi for starch analysis.

AcknowledgementsWe would like to thank Richard Fullagar for his suggestion and for providing important references.

ReferencesAranguren, B. & A. Revedin. 1997. Il ciottolo inciso

e utilizzato dall’insediamento gravettiano diBilancino e i “ciottoli a cuppelle” in Italia. Rivista diScienze Preistoriche 48: 187-222.

–1998. L’habitat gravettien de Bilancino (Barberino diMugello – Italie centrale), in Proceedings of XIIIInternational Congress of UISPP, Forlı (Italy) 8-14September 1996, 2: 511-16.

–2001. Interpretation fonctionnelle d’un site gravettiena burins de Noailles. L’Anthropologie 105: 533-45.

–2005. Dalla tipologia analitica alla catena operativa efunzionale: per una nuova definizione del Bulino diNoailles, in F. Martini (ed.) Askategi: Miscellanea inmemoria di Georges Laplace (Supplemento 1 dellaRivista di Scienze Preistoriche): 137-50. Firenze:Istituto Italiano di Preistoria e Protostoria.

–In press. Evidenze archeologiche di strategie disussistenza legate alle risorse vegetali presso icacciatori-raccoglitori, in 4˚ Convegno Nazionale diEtnoarcheologia, Roma 17 -19 maggio 2006. Roma:Associazione Italiana di Etnoarcheologia.

Aranguren, B., G. Giachi, P. Pallecchi &A.Revedin. 2001. Primi dati sul focolaregravettiano di Bilancino, in Atti della XXXIVRiunione Scientifica IIPP ‘Preistoria e Protostoriadella Toscana’ dedicata ad Antonio Mario Radmilli :337-48. Firenze: Istituto Italiano di Preistoria eProtostoria.

Aranguren, B., A. Revedin, S. Bartolacci & R.Morandi. 2003a. L’insediamento gravettiano diBilancino (Firenze), in C. Peretto (ed.) Analisiinformatizzata e trattamento dati delle strutture diabitato di eta preistorica e protostorica in Italia:71-81. Firenze: Istituto Italiano di Preistoria eProtostoria.

Aranguren, B., G. Giachi, M. Mariotti Lippi, M.Mori Secci, A. Revedin & G. Rodolfi. 2003b.Paleoecological data on the Gravettian settlement ofBilancino (Florence, Italy), in M. Patou-Mathis &H. Bocherens (ed.) Le role de l’environnement dansles comportements des chasseurs-cueilleursprehistoriques, Actes du XIVe Congres UISPP,Universite de Liege 2001, Colloque/Symposium C3.1(British Archaeological Reports International SeriesS1105): 171-79. Oxford: Archaeopress.

Aranguren, B., A. Revedin & M.R. Bagnari. 2004a.Interpretazione funzionale del sito di Bilancino(Firenze), in M. Barogi & F. Lugli (ed.) 2˚Convegno Nazionale di Etnoarcheologia, Mondaino7-8 giugno 2001: 126-39. Rimini: Raffaelli.

Aranguren, B., A. Revedin, M. Sozzi, M.L.Vannucci & S. Vannucci. 2004b. First results onprovisioning sources of siliceous raw materials fromthe Gravettian site of Bilancino (Florence, Italy), inSecretariat du Congres (ed.) Section 6. LePaleolithique Superieur / The Upper Palaeolithic,General sessions and posters, Acts of the XIVth UISPPCongress, University of Liege, 2001 (BritishArchaeological Reports International Series 1240):119-26. Oxford: Archaeopress.

Aranguren, B., P. Pallecchi & A. Revedin. 2006.Circolazione e utilizzo dell’ematite nell’ambito delladiffusione di conoscenze tecnologiche nelPaleolitico superiore: l’esempio di Bilancino, in D.Cocchi Genick (ed.) Materie prime e scambi nellapreistoria italiana, Atti della XXXIX RiunioneScientifica dell’ Istituto Italiano di Preistoria eProtostoria, Firenze, 2004: 253-66. Firenze: IstitutoItaliano di Preistoria e Protostoria.

854

Res

earc

h

Biancamaria Aranguren et al.

Aranguren, B., L. Longo, P. Pallecchi P & A.Revedin. 2007. The operative chain of the NoaillesBurin: tipology, technology and functionality, inBurins: formes, fonctionnements, functions, Actes de laTable Ronde d’ Aix-en-Provence, 3-5 mars 2003:87-106.

Aranguren, B., L. Longo & A. Revedin. In press.Bilancino, a specialized site for ‘latent technology’:an integrated approach, in Prehistoric Technology,Proceedings of an International Congress in Verona20-23 april 2005.

Barton, H., R. Torrence & R. Fullagar. 1998.Clues to Stone Tool Function Re-examined:Comparing Starch Grain Frequencies on Used andUnused Obsidian Artefacts. Journal ofArchaeological Science. 25: 1231-38.

Cordain, L. 2002. The Nutritional Characteristics of aContemporary Diet Based Upon Paleolitithic FoodGroups. Journal or American NeutraceuticalAssociation 5:15-24.

De Beaune, S.A. 2000. Pour une archeologie du geste.Paris: CNRS.

–2004. The Invention of Technology: Prehistory andCognition. Current Anthropology. 45(2): 139-62.

Fullagar, R. & J. Field. 1997. Pleistoceneseed-grinding implement from the Australian aridzone. Antiquity 71: 300-307.

Fullagar, R., J. Field & L. Kealhofer. In press.Grinding stones and seeds of change: starch andphytoliths as evidence of plant food processing, inY.M. Rowan & J.E. Ebeling (ed.) New Approaches toOld Stones: Recent Studies of Ground Stone ArtifactsIV. London: Equinox.

Gott, B. 1999. Cumbungi, Typha Species, a StapleAboriginal Food in Southern Australia. AustralianAboriginal Studies 1999/1: 33-49.

Jensen, W.A. 1962. Botanical histochemistry. SanFrancisco: W.H. Freeman & Co.

Kealhofer, L., R. Torrence & R. Fullagar. 1999.Integrating Phytoliths within Use-Wear/ResidueStudies of Stone Tools. Journal of ArchaeologicalScience. 26: 527-46.

Kozlovski, J.K. 2002. Les premiers hommes moderneset les premiers agriculteurs en Europe: voies dediffusion et interactions entre populations, inM. Otte & J.K. Kozlowski (ed.) Prehistoire de laGrande Plaine du nord de l’Europe, Liege, ERAUL99: 9-34.

Kunkel, G. 1984. Plants for human consumption.Koeningstein: Koeltz Scientific Books.

Lalueza, C., A. Perez-Perez & D. Turbon. 1996.Dietary inferences through buccal microwearanalysis of middle and upper Pleistocene humanfossils. American Journal of Physical Anthropology100 (3): 367-87.

Mariotti Lippi, M. & M. Mori Secci. 2002.Palynological investigation at Bilancino dam(Mugello, Central Italy). Webbia 57: 251-64.

Mason, S.L.R., J.G. Hatler & G.C. Hillman. 1994.Preliminary investigation of plant macro-remainsfrom Dolni Vestonice II, and its implications forthe role of plant foods in Palaeolithic andMesolithic Europe. Antiquity 68: 48-57.

Mason, S., G. Hillman & J. Hather. nd.Investigation of the role of Wild Plant Foods inPre-Agrarian Europe, in http://www.ucl.ac.uk/archaeology/research/profiles/smason/smresear.htm.

Pearsall, D.M. 1989. Paleoethnobotan: a handbook ofprocedures. San Diego: Academic Press.

Pearsall, D.M, K. Chandler-Ezell & J.A. Zeidler.2004. Maize in ancient Ecuador: Results of residueanalysis of stone tools from Real Alto site. Journal ofArchaeological Science 31: 423-42.

Piperno, R.D. & I. Holst. 1998. The presence ofstarch grains on prehistoric stone tools from thelowland Neotropics: Indication of early tuber useandagriculture in Panama. Journal of ArchaeologicalScience 25: 765-76.

Piperno, R.D., E. Weiss, I. Holst & D. Nadel.2004. Processing of wild cereal grains in the UpperPalaeolithic revealed by starch grain analysis. Nature430: 670-73.

Seidemann, J. 1996. Starke Atlas: Grundlagen derStarke-Mikroskopie und Beschreibung der wichtigstenStarkearten. Berlin: Paul Parey.

Tateoka, T. 1962. Starch Grains of Endosperm inGrass Systematics. Botanical Magazine, Tokyo 75:377-83.

Torrence, R., R. Wright & R. Conway. 2004.Identification of starch granules using imageanalysis and multivariate techniques. Journal ofArchaeological Science 31: 519-32.

Wright, K. 1991. The origins and development ofground stone assemblages in Late PleistoceneSouthwest Asia. Paleorient 17/1: 19-45.

– 1992. A classification system for ground stone toolsfrom the Prehistoric Levant. Paleorient 18/2: 53-81.

855