/./. TRAEEEMEXTDES RÉCOLTES : l X REGARD SI R L\ 1)1 'XÉOUTHIQl E Al'PRÉMXT XXIIF rencontres internationales d'archéologie et d'histoire d'Antibes Sous la direction de P. C. Anderson, L. S. Cuminings, T. K. Schippers, B. Simonel Éditions APDCA. Amibes. 2003

Ancient threshing floors, threshing tools and plant remains in 'Uvda Valley, southern Negev désert, Israël. A preliminary report

Uzi AVNER*, Patricia C. ANDERSON'*, But THI MAI**,

Jacques CHABOT** and Linda S. CUMMINGS****

Résumé La N'allée d'Uvda, dans le Néguev du Sud, est actuellement une zone désertique extrême. Néanmoins, depuis le \ millénaire av. J-C. un vaste système agricole s'y est progressive­ment développé, qui a atteint son apogée au IIP' millénaire av. J-C. Un des phénomènes agricoles remarquables est un groupement de 29 aires de battage anciennes. Une aire de battage que nous avons fouillée a fourni une séquence stratigraphique et chronologique du développement de la technologie du battage (dépiquage) pendant les I V et III ' millé­naires av. J -C . Des analyses de laboratoire récentes de micro-traces d'utilisation sur les outils en silex, du pollen et des phytolithes, ont révélé de nouvelles précisions sur cette agriculture du désert, en particulier à propos de l'utilisation du tribulum comme technique de dépiquage, et sur le paléoenvironnement. Les résultats préliminaires de ces études sont présentés et discutés ici.

Abstract 'Uvda Vallev, in the southern Negev, is an extrême désert area todav. N'evei theless. since the 61'1 millennium B C a vast agricultural seulement System gradually developed here, reaching its climax in the S millennium B C . One of the unusual agricultural phenomena is a concentration of 29 ancient threshing floors. An excavated threshing floor yielded a stratigraphie and chronological séquence of development of the threshing technology during the 4 , h - 3 r d millennia BC. Récent laboratorv analyses, of microNvear on llint tools. of pollen and of phytoliths, supplied new évidence for the désert agriculture, for the thresh­ing technology, and for the paleo-environment. The preliminary results are presented and discussed below.

A raya Institiile. l'.ilal. Israël. Mél : u/\.\</' 012net.il ** Cépam, UMR 6130, 250, rue Albert-Einstein, F-06560 Valbonne. Mél:

andersoiitfrccpam.ciirs.fr ; buitliitfr cepama nrs.f'r *** Laboratoire de recherche sur la pierre taillée and CELAT, Université Laval, Québec,

Canada. Mél : jacqucs.chabot@hst.ulaval.ca ****Paleo Research Institute, 2675 Youngfield St., Golden, C O 80401 USA. Mél:

I.indaê PaIeoresearch.com

l ' / i A \ \ l K . Pairie ia C . A N D I R S O N . B u T l I I MAI et al.

'Uvda Valley is situated in the southern Negev, 40 km nor th of the Gulf of Aqaba. The area is hyper arid, with mid-day summer températures above 40

degrees C , an average annual précipitation of 28 m m , and a potential animal evaporalion raie of 1 <>()() mm. This restricts végétation in the région t<> the wadi-beds, and therefore. the carrying capacity is very low. Since 1978, one of the writers (Avner, 1979, 1982a, b, 1990, 1998) hasconducted an archaeological survey and excavations in the area. The eastern side of the valley was found to be extre-mely rich in sites, 400 o f which were documented, dating f rom the Early Neolithic to the Early Islamic periods. The majority of the sites were dated to the 6 - 3 ™ millennia BC, including 186 stone-built habitation sites and corrals, 40 tent camps, agricultural installations, and cuit and burial sites (fig. 1). Despite the présumée! environmental conditions, a vast farming seulement system dcveloped in the valley. The site densin in the eastern side of ihe valle\s higher tl i .m thaï found in much more moderate désert areas, such as the Negev Highlands. and reached a population of at least 3000 at its peak, in the 3 r d mi l l enn ium BC (Avner, 1998). Abundant finds pointée! to an agricultural economy of the inhabitants, inc luding hundreds of glossed and non-glossed flint blades of various types (see below), hundreds of gr inding stones. two stone arc! tips. the earliest known to date in the Near East (ca. 3000 BC), botanical remains o f domesticated wheat and barley, silos built in the habitation sites and the remains of vast cultivated fields. The inhabi­tants were able to develop an agricultural system in the valley due to a rare combi-nation of local environmental conditions. One is the unique soil, consisting of lime-sand mixed with silt, which is well ventilated, easily ti l led and highly permé­able. The other is the hydraulic régime, in which floodwater f rom a drainage area o f 400 sq. k m runs slowly over the surface, whose gradient is only 0,3 %. As a resuit, floods are non-destructive, and the water is w i l l absorbed by the soil. The ancients further utilized the natural conditions by bui lding a long séries of ' limans' (large level basins. surroimded by earth einbanktnents). which supplice! thein with 1200 hectares o f fertile cultivated field, watered by floods. Today, dense vvild cereal stands still grow in the ancient field areas, followïng the floods (fig. 2).

The threshing floors

D u r i n g the survey, thirtv-two circulai" structures were found, measuring 8-15 m in diameter. Twenty-nine were built in an area of 0,5 x 2,5 km along the eastern side o f the valley (f ig. 3) . Twenty were eut into the rock surface to a m a x i m u m depth of 80 cm, while twelve were circular structures of beaten earth. The inter­prétation of thèse installations as threshing floors initially faced objections, due to the absence o f known archaeological parallels. Over time, however, conclusive archaeological and ethnographie évidence lias been collected to support this interprétation.

A i l o f thèse structures yielded surface collections o f f l int and pottery dated to the 5 , , , -3" 1 mil lennia BC, as well as f r o m several later periods up to the 2 0 , h century.

456

Fig. 1 . Survey map of 'Uvda Valley (due to map scale, not ail sites are shown).

457

U z i AVNER, Patricia C . A N D E R S O N , BUI T H I MAI et al.

Fig. 2. Northern 'Uvda Valley, a growlh of xvild cerealfollowing aflood (photo U. Avner).

Excavations of a threshing floor

A n aerial photograph of Site 96, at the m o u t h o f Nahal Yitro (f ig. 4, 5) , showed i n t r i g u i n g features, particularly a l ight-colored strip 120 m i n length, orientée! roughly North-South, which included two circular threshing floors eut into the bedrock. This raised the possibility of an on-site development o f the floors, which motivated the excavation o f the northernmost one. The first excava­t ion was conducted i n 1982, to examine site interprétation and technological development (Avner, 1982, 1990, 1998). A n extension o f the dig took place in the summer o f 2002, in an attempt to verify stratigraphie questions and to select soil samples for micro-botanical analyses The later excavation yielded a develop­ment a séquence in the site that was slightly différent f r o m the one previously published (Avner, op. cit.). Preliminary results o f our récent laboratory analyses o f f l i n t micro-wear (by Anderson and Chabot) , of pol len (by Bui-thi-Mai and Cummings) and o f phytoliths (by Anderson and Cummings) , contr ibuted infor­mation vital to understanding the site and the ancient agriculture i n the région (see below). Following is a brief description of the principal stages identi f ied d u r i n g the 2002 excavation :

Field observations showed that in the first stage, the upper layer o f weathered rock was removed f r o m the surface, and a flat bedrock f loor was cleared over an

1. T h e first excavation was made on behalf of the Israël Depar tment of Antiquities (now the Israël Antiquities Authority) a n d the second was supported by the G o o d m a n F o u n d a t i o n in New Y o r k ) . We thank them ail for enabl ing the dig a n d research.

458

A N C I E N T T H R E S H I N G F L O O R S , T H R E S H I N G T O O L S AND PLANT REMAINS IN ' U V D A V A L L E Y , ISRAËL

Fig. 3. Eastern 'Uvda Valley, map of the main concentration of threshing floors.

459

U z i AVNER, Patricia C . A N D E R S O N , B U I T H I MAI et al.

Fig. 4. Uvda Site 96, An aerial photograph from the North, before excavation : 1, perimeter of first working area ; 2, flint worhshop ; 3, silos ; 4, circular, rock-cut threshing floors (photo U. Avner).

Fig. 5. Uvda Site 96, a gênerai plan (with numbers of Loci).

A N C I E N T T H R E S H I N G F L O O R S , T H R E S H I N G T O O L S A N D P L A N T R E M A I N S I N ' U V D A V A L L E Y , I S R A Ë L

461

U z i AVNER, Patricia C . A N D E R S O N , BUI T H I MAI et al.

Fig. 7. 'Uvda Site 96, an aerial photo of the Northern threshing floor during the first excavation : 1, perimeter of the first cleared rock surface and re-excavated rock floor ; 2, flint workshop (L. 11) ; 3, silos ; 4, rock-cut threshing floor.

Fig. 8. The Northern threshing floor during the second excavation, from theEast : 1, the earlier cleared rock floor ; 2, flint workshop ; 3, silos ; 4, rock-cut threshing floor ; 5, part of the stone dump left un-excavated.

area o f ca. 18 x 35 m . The rocks removed were used for b u i l d i n g a low wall surrounding the cleared area, whose remains were f o u n d mainly to the East (W. 7) and West (W. 8, see f ig . 6) . A vague, curved l ine Connecting the two walls marks its original contour at the N o r t h e r n end (fig. 5, 7) . The cleared area was gradually extended southward u n t i l i t reached a length o f 120 m (figs. 4, 5) . Part o f the rock f loor was later covered by a stone d u m p (see below), but when re-exposed by excavation ( in Locus 7, hereafter, a gênerai term for Loci 7, 14, 15, 16), a dark gray patination was f o u n d o n its surface, indicat ing prolonged expo-sure and use. The circular silo i n Locus 10 seems to belong to this stage, and was later eut by L o c i 5 and 11 (fig. 6) . Ethnographie documents (e. g. Dalman, 1933, p. 188-9 ; Avitzur, 1966, p. 89) suggest that the silos i n the site were used for storage of cleaned grain d u r i n g the threshing season.

I n the second stage a circular structure (Locus 11) was bui l t , partially o n top o f the cleared rock surface (fig. 6, 8) . I t contained a stone anvil and a flagstone bench, hundreds of flint debitage pièces and several dozen tools (see below). A large amount of flint was also f o u n d around this structure, suggesting that i t

A N C I E N T T H R E S H I N G F L O O R S , T H R E S H I N G T O O L S A N D P L A N T R E M A I N S I N ' U V D A V A L L E Y , I S R A Ë L

Fig. 9. The Northern threshing floor, excavation of Locus 7 : 1, the earlier cleared rock floor ; 2, occupation level ; 3, stone dump (photo U. Avner).

Fig. 10. The Northern threshing floor following the second excavation, from the North : 1, the earlier cleared rock floor ; 2, flint workshop ; 3, silos ; 4, rock-cut threshing floor ; 5, part of the stone dump left un-excavated (photo U. Avner).

most probably served as a f l i n t workshop. Then, the area of Locus 7 was covered by a t h i n occupation level 5-7 cm thick, o f fine silt and sand, mixed with ash and containing artifacts ( f ig. 9 : 2 ) . 1 4 C dates show that this level dates f r o m ca. 3000 BC (see below). The accumulation o f the occupation level indicates that this part o f the large cleared rock surface was no longer used for threshing, although threshing at the site d i d continue. This can be deduced f r o m the construction o f Silo 5 at this stage, f r o m the continuous use o f the f l i n t workshop, and f r o m the richness o f cereal phytoliths and pol len i n soil samples f r o m the occupation level (see below).

I n the t h i r d stage, Silo 6 was bui l t on top o f the occupation level and against Wall 7 (fig. 6, W. 7, 10). This indicates, again, continuation o f threshing on other parts o f the cleared rock surface.

I n the f o u r t h stage, the two circular threshing floors were eut into the large cleared rock surface ; the n o r t h e r n one (Locus 1) measures 13 x 17 m and is 0.6 m deep. Part o f the excavated rocks were used for b u i l d i n g the wall surroun-d i n g the threshing f loor (W. 1, termed « Shall » i n the Bible- 2 Sam 6 v. 7) . The

463

U/.i AVNF.R. Patricia C . AXDF.RSOX. 1U I I I II MAI et al.

rest o f the rock débris was d u m p e d on top o f the occupation level to the east (Locus 7. f ig . 9 : 3 ) , also covering the silo in Locus 6. A t h i n incl ined wall (W. 2, f ig . 6, 10) was built to support the dumped material and prevent it f rom spill ing onto the threshing floor.

Artifacts f rom the excavation included large amounts o f f l int debitage and tools, pottery sherds, two pairs and one t r io of g r i n d i n g stones (saddle querns) made o f hard sandstone, a stone ard t ip, the end of an elongated wooden imple-ment (95 x 28 x 16 m m ) , and some sea-shells, fragments o f ostrich egg-shell and caprine boue fragments.

Flint implements in the threshing floors and micro-wear analysis

The excavation of the nor thern threshing floor yielded several thousand flint items, mostly from within the f l int workshop and around it . The large majo-rity was debitage, while scraper-like implements were dominant among the tools (39 ont o f 90, 43 % ) . A n additional pile of 29 « scrapers » was found during the site's survey, on the surface o f Silo 22, slightly covered by wind-blown sand. Similar tools were collected in almost ail threshing floors in the valley, in a few habitation sites, and in several threshing floors in the Negev Highlands. Thèse tools présent a variety of forms ; some are heavy duty while others are elongated and thinner. Some spécimens o f both are made on transverse flakes, similar to « fan scrapers », and 19 of thèse 29 tools retain cortex on the dorsal face (f ig. 11). A i l o f thèse heavier tools are made f r o m local, coarse-grained flint.

The occurrence of thèse tools in the threshing floors posed the question o f whether they were used for threshing crops. In fact, f l in t sledge blades, smaller and thinner than the 'Uvda scraperlike tools, are well known as inserts in ethno­graphie material, for example f r o m Cyprus, Turkey, Greece, the Balkans and Spain -. They were also identif ied in archaeological contexts in 3 I ( 1 mi l l enn ium Mesopotamia, Syria and Israël (Adams, 1975; Anderson, 1994, 1998, 1999; Anderson, Inizan, 1994 ; Anderson, Chabot, 2001 ; Skakun, 1999 ; Ataman, 1999 ; Chabot, 2002) and even in PPNB sites, in Syria and Iraq (Anderson, 1994, 1998, 1999 and this volume) and possibly in Turkey (Ataman, 1999). I t is clear, by comparison with microwear traces obtained with experiments and ethnographie threshing sledge inserts, that thèse tools were inserted in a raft-like structure, known f r o m 3 l d mi l l en ium texts (Grégoire in Anderson, Inizan, op. cit.). Traces on blades f r o m the PPNB show that the threshing sledge, as the one described in cunei form texts, was pul led around a threshing floor over harvested plants (Anderson, op. cit. ; Ataman, op. cit.) a l though there is no évidence preserved indicating the appearance o f the PPNB instrument.

2. H o r n e l l , 1930: Crawford. 1035: Borda/ . 1900: Whal lon . 197S : C h c c t h a m . 1982: Pearlman, 1984, 1987 : Fox, Pear lman, 1987 : S k a k u n . 1999 : Ataman. 1999. In 1990. Prof. A. R o u e n , Haifa l nivci sii\ <lo< u m c n t c d on video a (Talisman in Nicosia preparing lliis type ol sledge blade loi sale. In 199."). wlien Avner visiter! Cvprus . lie was told In Pearlman that the (Talisman harl rlierl. Récently. |. VVhittaker interviewer! retirer! Ilint knappers in C v p r u s (Whi i iakei . 1991). 1999).

464

A N C I E N T T H R E S H I N G F L O O R S , T H R E S H I N G T O O L S AND PLANT REMAINS IN ' l ' Y D A \. ISRAËL

Microwear analysis o f the 'Uvda material, by Anderson and Chabot, included two principal types o f artifacts. One was a group o f 20 of the scraper-like tools f r o m the f l in t workshop and f r o m Silo 22 (fig. 11), and the other was a group of 30 glossed blades f r o m nearby habitation sites (fig. 12), inc luding one large « Canaanean » blade (fig. 13). The tools were examined under a metallographic microscope, at 100X and 200X magnification, and compared wi th récent exami-nation o f about fïfty ethnographie threshing sledge flints f r o m various countries and with flints used in reconstructed threshing sledges f r o m 1995-2002 (Anderson, Chabot, 2001).

A n earlier micro-wear analysis o f the « scrapers » (Bueller, 1988, p. 20-23, 30), had suggested their use as inserts in a threshing sledge, for non-cereal plants, and fol lowing this observation, Avner (1998, p. 167) had suggested that they could have been used for threshing chickpea. I n the présent analysis, however, the results d i d not support the earlier hypothesis. Most tools, f r o m both the flint workshop and f r o m Silo 22, d i d not show any use traces, while some bore only faint use traces o f a scraping mot ion , probably on wood.

Two elongated, tabular scraper-like spécimens (fig. 11 : second f r o m left, f o u r t h f r o m left) carried diffuse microscopic abrasion traces on the retouched edges, parallel to their long axis, which could have been produced by cutt ing of plant material i n contact wi th an abrasive such as soil. Thèse traces, however, were too poorly developed and covered too l imi ted an area of the cutt ing edge lo point to any spécifie agricultural use. Since most tools d i d not bear any use wear, it is possible that those f r o m the f l i n t workshop never reached the stage o f utilisa­t ion , while the pile in Silo 22 was possibly made for some symbolic or r i tual purpose (cf., the papers o f V i k l u n d , Whittaker, Garine et ai, this volume, for possible ethnographie parallels).

T h i r t y backed blades, mostly macrolunates (fig. 12a, d, f, and g) , had been previously assumed to be sickle blades due to their gloss (Avner, 1998, p. 160), but had not been analysed for microwear traces. Indeed, the présent micro-wear analysis yielded différent results. Some were patinated and therefore had a poorly-visible use-wear pattern which could be identi f ied only as f r o m cutt ing

465

A N C I E N T T H R E S H I N G F L O O R S , T H R E S H I N G T O O L S AND PLANT REMAINS IN ' U V D A VALLEY, ISRAËL

cereals in gênerai, wi thout further détail possible, but others bore clear, well-deve-loped wear patterns characteristic of threshing sledge blades (ex. f ig . 12b, c, e). Thèse included uni-directional , longi tudinal abrasion traces, often in the f o r m of wide, comet-shaped grooves, and large shallow pits, and randomly-oriented scrat-ches ( f ig . 12 b, c, e, and cf. Anderson, 1994, 1998, 1999 ; Anderson, Chabot, 2001) . Most blades were apparently set parallel to the plank or raft in the under-side o f the sledge, but a few had traces oriented slightly obliquely to the edge, indicating a canted set in the threshing sledge. Ethnographie threshing sledges often have parallel-set blades nearer their f ront and middle , and canted-set blades nearer their distal ends (Anderson, Inizan, 1994). No traces of b i tumen were found preserved on thèse blades, unlike many sledge blades f r o m N o r t h e r n Mesopotamia (see e. g. Anderson, Chabot, 2001, p. 266).

The analyzed Canaanean blade (f ig. 13) is one out of only five found to date in 'Uvda. Since this type is rare i n the Negev, it was most probably imported f r o m further N o r t h . Examination by Jacques Pelegrin (personal communicat ion, 2002) showed that it was removed f r o m a core by a spécifie pressure technique, using a copper-tipped lever. The technique was first recognized on blades f rom N o r t h e r n Mesopotamia and then reconstructed in experiments by Pelegrin (Annex in Chabot, 2002 ; Chabot, 2002, chap. 3). The blades f r o m this régions had been long assumed to be sickle inserts (e. g. Rosen, 1983c, 1997, p. 55-58) and indeed, some possible Canaanean blades were discovered in three wooden sickle remains at Tell A b u Al-Kharaz, Jordan (Fischer, 1994, p. 132; plate 11). However, récent micro-wear analyses on blades f r o m EB sites has revealed a diffé­rent picture. One hundred and eighty Canaanean blades f r o m Tell 'At i j , NE Syria, ail showed use-wear typical for sledge blades and not sickles (Anderson, Chabot, 2001 ; Chabot, 2002, chap. 5) . Out of several hundred blades analyzed f r o m 30 additional sites f r o m n o r t h e r n Syria and Mesopotamia (op. cit., Anderson, Inizan 1994 ; Anderson, 1999), only one blade, f rom Telul eth-Talatat V shows double use-wear (Nishiaki, Chabot, in préparation), of both harvesting and threshing while ail others showed threshing wear only (Eid, personal commu­nicat ion) . The présent Canaanean blade f r o m 'Uvda, analyzed microscopically by Anderson and Chabot, also showed the typical sledge use-wear (fig. 13b) and would have been easily inserted i n a threshing raft such as that cited in cunei form texts, which we have reconstructed and used in experiments over six seasons 3

(«g. H ) .

< Fig. 12. Tools with gloss and microwear traces showing use in a threshing sledge : a, d. /. macro-lu notes with gloss (small circles) from l'vda Site 166, Icite 3"1 milieu ht m, Early Bronze 4 with traces characteristic of their use as inscris in a threshing sledge or raft (b, traces on tool in a (arrow) : c, traces on tool d (arrow) ; e, traces on tool f(arrow) ; g, small backed tool, with traces from use in threshing sledge, from excavated threshing floor, Locus 5 silo, nearjlint workshop (photos P. C. Anderson, drawingsJ. Courbet).

3. For the distinction and comparisnn between sickle a n d sledge use-wear see A n d e r s o n , Chabot . 2001, table l , p . 264.

467

U z i AVNER, Patricia C . A N D E R S O N , BUI T H I MAI et al.

Fig. 14. Reconstruction of a threshing instrument of the tirne, according to cuneiform texts and archeological data, used in experiments (draxvingj. Courbet).

468

A N C I E N T T H R E S H I N G FLOORS, T H R E S H I N G T O O L S AND PLANT REMAINS IN LA'DA VALLEY, ISRAËL

Micro-botanical analysis

Pollen analysis

Pollen analysis was carried out on seven soil samples : one f r o m the excavated threshing floor's occupation level (Locus 7, fig. 9) , one f r o m the unexcavated, southern threshing f loor at the same site (Locus 20, fig. 5) , three samples f rom occupation levels of 4 , h - 3 r d mi l lennia habitation sites, two f r o m underneath the pavements of bui l t silos wi th in habitation sites and one f r o m another unexcavated threshing floor. A l t h o u g h detailed pollen counting is incomplète as yet, the preli­minary results greatly contribute to understanding the sites.

The first sample, f r o m the excavated threshing floor, contained many thou-sands of domesticated cereal pollen, slightly over 50 % o f the total, while others are f r o m typical agricultural weeds and typical désert végétation (Chenopodicaea, Amaratharaca, LiguUfhrae, etc.). In one sample, from an occupation level o f habi­tation Site 124/IV, cereal pol len reached 8-10 %, while in others it was even lower. The principal resuit here, therefore, is the outstanding richness o f cereal pollen in the threshing floor, in comparison to ail the other samples.

Deserl végétation was dominant in ; i l l samples excluding the first, however, some pollen indicate a richer végétation than the présent one, inc luding Tamarisk, a tree associated with wet areas which is very rare today in the 'Uvda vicinity, Plaiitago, Bocrhaavia which represent weedv plants, and Myrlareac, which is a Mediterranean type. Sweet water plants such as sedge (Cypenis and Scirpus sps.) were also identif ied, which may indicate the présence o f a spring or springs in the région that no longer exist today. It seems, therefore, that the 'Uvda environment was somewhat wetter d u r i n g the 4 , h -3 l < 1 mil lennia BC than it is today.

Six charcoal samples f r o m the excavated threshing floor were recently iden­tif ied under a microscope by N . Liplischitz (personal communicat ion. 2003) ; they represent three species used for combustion. Atlantic pistachio, Tamarisk, and White Broom. The first two were dominant , although they are very rare today in the area, whereas the YVTute Broom, which is common and excellent for combustion, was f o u n d in only one sample. Recovery of Tamarisk in both the pollen and charcoal records substantiates the présence o f thèse trees, growing either along wadis or near springs, since they require a great deal o f groundwater to survive.

Phytolith analysis

Phytoliths (silica deposited in plant cells which survives decay and shows cell patterns allowing identif ication o f the plant) were analyzed f r o m 17 soil samples f r o m excavated and unexcavated threshing floors, f r o m excavated habitation sites and one control sample o f the présent désert surface. Five samples f rom the exca­vated threshing f loor ( f rom Loci 7, 22, and 24) yielded eut cereal stem phytoliths, with smooth, regular profiles, straight and curved, and perpendieular to the stems' ( longitudinal) fibers. Based on ethnographie and expérimental ( f ig. 15a)

469

Uzi AVNER, Patricia C. ANDERSON, BUI THI MAI et al.

Fig. 15. a, a phytolith from cereal stems eut by the expérimental threshing sledge on the threshing floor then burnt, with a smooth, curved eut profile (arrow) ; b, a similar smooth-cut phytolith (arrow) from below the flagstone pavement of a silo in Locus 16, probably from burnt cereal, from stems eut by a threshing sledge (photos P. C. Anderson).

material, thèse are clear indications for the use o f a threshing sledge equipped wi th f l i n t blades (Anderson, 1998, 1999, p. 145 ; Anderson, Chabot, 2001 ; Anderson, this volume ; see also Cummings, this volume ; Khedhaier et al. ; this volume) . A l t h o u g h phytoliths of cultivated cereal are hardly distinguishable f r o m those o f grass, the sharp perpendicular cuts do indicate they are f r o m domesti-cated cereal processed under a sledge. Cereal phytoliths wi th smooth, curved and straight cuts were also f o u n d i n a sample taken below the pavement of a bui l t silo i n Site 16, and some were b u r n t ( f ig. 15b), suggesting intentional or accidentai in t roduct ion o f a winnowing residue wi th the grain (or possibly midden mate­r ia l ) . I n contrast, the control sample f r o m the désert surface, 1 k m south o f the threshing f loor area, d i d not contain any cereal phytoliths, indeed almost no phytoliths o f any k i n d .

1 4 C Dating

Following the first excavation, one 1 4 C date on charcoal f r o m Silo 3 was early 3 r d m i l l e n n i u m calibrated B C Recently two additional results were received. One, f r o m a hearth at the top o f the occupation level i n Locus 14b ( included i n the gênerai Locus 7) , was around 3000 BC ; the other f r o m Locus 15b was 600 years later and most probably represents an intrusion 4 . Addi t iona l thermolumines­cence dates f r o m the base of Wall 11 (the f l i n t workshop) is now under analysis, by N . Porat. The second date ment ioned here is presently the most significant. I t

4. The first date was analyzed by I. Carmi in Weizman Institute, Rehovot, Israël, the second by E . Boaretto and G. Mintz in the same laboratory. The third sample was prepared by the latter and dated by AMS in the University of Arizona, Tucson, by T. Jull. Following is the dating information : Samp. No. Locus Material BP cal. BC (1 sigma) 1. RT 648B 3 Charcoal 4250 ± 50 2920-2700 ; 2. RT 4408 14b « 4350 ± 40 3020-2900 ; 3. RT 4407 15b « 3870 ± 40 2460-2280. (Date n° 3 is not yet final). Calibration is based on OxCal 3.4 (Ramsey, 2000).

A N C I E N T T H R E S H I N G F L O O R S , T H R E S H I N G T O O L S A N D P L A N T R E M A I N S IN ' U V D A V A L L E Y , I S R A Ë L

Fig. 16. Northern Sinai. Arabs threshing using a sledge made from a flattened oil drum (photo U. Avner).

dates the end o f the occupation level, and therefore the first large cleared rock surface and the f l in t workshop are older than 3000 BC.

The évolution of threshing technology in 'Uvda

A l t h o u g h there are still some uncertainties, a gênerai line o f development can be hypothesized, based on the identif ication o f the various stages i n the site, the finds and the laboratory analyses :

The first stage in the site has a large cleared rock surface (120 m long) which is very différent f r o m the circular threshing floors i n the 'Uvda Valley and else-where, i n size and in ( irregular) shape.

Theoretically, threshing could have been done there by beating (cf. Judges, 6 : 13 ; Ruth 2 : 17 ; Dalman, 1933, abb. 25) or by animal t rampl ing. The labour investment i n removing the top rock layer may indicate that the hard, flat surface was desired for better efficiency of both threshing and collection o f the grains. A t the présent t ime, this stage is not directly dated, but based on the 1 4 C dates o f the fo l lowing stage, and the prolonged use of the cleared rock surface, as indicated by the patination (see above), i t probably dates f r o m the 4 t h m i l l e n n i u m BC.

I n the second stage the f l in t workshop was built , partially o n top o f the cleared rock f loor while cutt ing into Wall 7 (fig. 10) and it was followed by the accumula­t ion o f the occupation level to its n o r t h . The necessity of a f l in t workshop next to the threshing floor, and the réduction of the required threshing area, can be best explained by the in t roduct ion o f the sledge equipped wi th f l i n t blades. This is well supported now by the occurrence of abundant cereal pol len i n the soil sample, and more so by the cereal phytoliths with sharp, perpendicular cuts. Since the 1 4 C date f r o m the hearth at Locus 7 marks the end o f the occupation level's accumulation, the construction t ime o f the f l i n t workshop was probably late 4 t h m i l l e n n i u m BC, i. e. Early Bronze I . To date, the hard évidence for the use of f l i n t blades for equipping a threshing sledge in 'Uvda cornes f r o m habitation sites (mainly 124/IV and 166), but the persistent occurrence o f the scraper-like tools ( f ig. 11) i n threshing floors speaks for their being associated in some way

U z i AVNER, Patricia C . A N D E R S O N , BUI T H I MAI et al.

Fig. 17. Western Negev Highlands, Bédouin threshing lr< Irampling (photo M. Haiman, 1981).

with the threshing process, despite the fact that that microwear traces o n them do not correspond to use for threshing with the threshing sledge. I t should be noted that no wall surrounding a circular space could be identif ied i n the first two stages. I t was only in the f o u r t h stage that the circular threshing f loor was eut into the older rock surface, and some of the cut-out rocks were used for b u i l d i n g the surrounding wall (W. 1). This may mean that the circular, bui l t threshing f loor represents consolidation o f a development process, fo l lowing the invention or adoption o f the sledge.

I t is interesting to find that the désert people at 'Uvda already used the thresh­ing sledge by the laie 41'1 mi l l enn ium BC. whereas people in Egypt, despite ils cultural and technological advancement, d i d not know of the sledge d u r i n g the entire Pharaonic per iod (Murray, 2000, p. 524-5).

A n interesting philological point unités 'Uvda Valley with the threshing floors. The Bédouin-'Arabie name for the valley is Wadi ' U q f i , and ' U q f i means a wooden pôle harnessed to the animais dragging the sledge (Avitsur, 1966, p. 76) ; the w o r d originated f r o m the Semitic root « walk about ». I t is interesting to see threshing in another région carried out today with improvised threshing sledges (fig. 16). The Bédouin of the Negev and Sinai, however, thresh their yields by t rampl ing only (fig. 17), and when questioned, were unfamil iar with the sledge and w i t h the meaning o f the valley's name. Since they could not have been the source of the geographical name, a possible alternative or ig in is f r o m the Nabateans, who also used the same k i n d o f threshing floors in the valley some 2000 years ago. Their language was Aramaic, and the root « walk about » also appears in Aramaic documents (e. g. Jean, Hoftijzer, 1965, p. 220). Therefore, i t is possible that the or ig in o f the valley's name is ancient and was passed on to the Bédouin like many other geographical names. I n any case, the occurrence o f an agricultural geographical name i n such a harsh désert environment is unex-pected, yet i t accords weJl with the archaeological finds.

472

A N C I E N T T H R E S H I N G F L O O R S , T H R E S H I N G T O O L S A N D I M A M R E M A I N S IN T Y D A V A L L E Y , I S R A Ë L

Intérim conclusions

1. The prel iminary results of laboratory analyses, the micro-wear, pol len and phytoliths. contributed new, additional évidence for the interprétation of the circular installations as threshing floors, and to the agricultural nature o f the seu­lement system of the 'Uvda Valley. This intensive seulement stands out in sharp contrast to the gênerai environmental conditions of the southern Negev, and présents a unique combination o f local conditions, with the wisdom of l iving developed by the désert populat ion.

2. The concentration of threshing floors in the valley is remarkable. According to oui présent knowledge, this concentration is the largest and earliest identi f ied to date in the Near East.

3. The animal-drawn sledge can be seen as an additional innovation related to the « secondary product révolution », next to the plough (Sherratt, 1981 ;

Chapman, 1983 ; Davis, 1987, p. 168 ; Grigson, 2000). We do not suggest that the « patent » on the sledge was first « registered » in the name of the 'Uvda Valley farmers, but it is still interesting that, in this highly arid area, the sledge was in use already by the 4lh m i l l e n n i u m BC. It is also of interest that the two earliest ard tips known to date were discovered in 'Uvda Valley, one o f them next to the f l int workshop in the excavated threshing floor.

4. In the study of 'Uvda Valley, many questions remain unanswered, and much work awaits us. Excavations of additional threshing floors and habitation sites are planned, as well as further laboratory analyses. Al though the good state of préservation o f thèse désert remains is highly promising for future progress in this research, the valley's remains are severely threatened by the demands of mi l i -tary t ra in ing in the area. Al though great efforts have already been invested in developing an archaeological park in the valley, it is not yet clear today how many of the sites wil l survive in the near future, nor how many will remain accessible to visitors and researchers.

Bibliography

A D A M S R.. I(.I7,").- An Ancient l'ruk Threshing Sledge or H arrow. Su mer, 31, p. 17-20.

ANDERSON R C , 1994 - Interpreling Traces of Near Eastern Neolithic Craft Activities : An Ancestor of the Threshing Sledge for Processing Doinestic Crops }. in : A. Van Cijn (dir.), The Use oj Uthic Tools in Neolithic Crafi Activities, Hetinhm, 34/2, p. 306-321.

A N D E R S O N P. C , 1098.- I liston ol Ilarvcsting and Threshing Techniques loi Cereal s in the Prehistoric Near l'.ast. ///: A. B. Damania. ). Valkonn, C». Willcox, C. O. Qnalset (dir.). The origins of agriculture and crop domestication. The harlan symposium. Aleppo. p. 145-150.

A N D E R S O N . P. C. 1090 - Expérimental cnltivation. harvest and threshing of wild cereals : their relevance for interpredng the use of epipalaeolithic and Neolithic artifacts, in : P. C. Anderson (dir.), Préhistory <>j agriculture. \eu> expérimental and ethnographie approaches, Monograph 10. Institute ol arc haeologv. universin ol ( lalifornia. Los Angeles, p. 104-1 IS.

A N D E R S O N P. C , C H A B O T J . , 2001.- Functional analysis of glossed blades from Northern Mesopotamia in the Early Bronze âge (3000-2500 BC) : the case of Tell 'Atij, Recherches

Uzi AVNKR, Patricia C. ANDERSON, Blï THI MAI et al.

archéométriques, collection •• Cahiers d'archéologie du C E L A T ••, n" 10 et « Série archéo inétrie », n° 1, Québec, p. 257-276.

ANDERSON P. C , INIZAN M. L. , 1994 - Utilisation du tribulum au début du IIP' millénaire : des lames « cananéennes » lustrées à Kutan (Ninive V) dans la région de Mossoul, Irak, Paléorient, 20, p. 85-103.

A T A MAN k... ÎOOO-Threshing sledges and archaeology, in: P. C. Anderson (dir.). Préhistory of agriculture. Sexv expérimental and ethnographie approaches. Monograph 40. Institute of archaeo­logy, university of California, Los Angeles, p. 211-222.

AYITSl 'R S.. 1060.- Implements for harvesting and similar pu rposes used in the traditional agriculture of Eretz Israël, Tel Aviv, (Hebrew).

AVNER U. , 1979 - A Survey in the 'Uvda Valley, Hadashot Arkheologiyot, 69-71, p. 1-2, (Hebrew).

AVNER U. , 1982a.- Biq'at 'Uvda Survey, ESI, 1, p. 79-80.

AVNER U. , 1982b.- Excavation of a Threshing Floor in Biq'at 'Uvda. /..S/. 1. p. 80-81.

AVNER U. , 1990 - Ancient agricultural seulement and religion in the 'Uvda Valley in southern Israël, A4, 53, p. 125-141.

AVNER U. , 1998- Seulement, agriculture and paleoclimate in 'Uvda Valley, Southern Negev Désert. f)' h-3 n l Millennia BC. in: A. Issar. X. Brown (dir.). Water, environ ment and society in times of climate change, Dordrecht, p. 147-202.

BORDAZj., 1969.- Flint flaking in Turkey, Natural history, 78, p. 73-79.

BUELLERj . , 1988 - Handling, hafting and ochre stains, in: S. Beyries (dir.), Industries lithiques. tracéologie et technologie, vol. L, Bar International Séries 411 (i), Oxford, p. 5-32.

CHABOT J. , 2002.- Tell 'Atij-Tell Gudeda. Industrie lithique. Analyse technologique et fonctionnelle. Éditions du Centre interuniversitaire d'études sur les lettres, les arts et les traditions ( C E L A T ) , Collection « Cahiers d'archéologie du C E L A T », n° 13 et « Série Archéométrie », n° 3, Québec, 225 p.

CHAl'MAN C f . 1083- 1 lie secondary prodnets révolution and the limitation ol the Neolithic, BIAUL, 19, p. 107-122.

CHEETHAM L. , 1982- Threshing and winnowing- an ethnographie study, Antiquity, 56, p. 127-130.

CRAWFORD O. G. S., 1935-A primitive threshing machine, Antiquity. 9, p. 335-339.

DALMAN G. , 1933.-Arbeit und Sitte in Palâstina, Band III. Gutersloh.

DAVIS S. J . M., 1987.- The archaeology of animais, I.ondon.

FISCHER P. M., 1994 - Tell Abu Al-Kharaz the Swedish Jordan expédition 1992, Third season, Preliminary Excavation Report, ADAJ, 38, p. 127-143.

FOX W. A., PEARLMAN D. A., 1987-Threshing sledge production in the Paphos district, Cyprus, SIMA, 77, p. 227-231.

GRIGSON, C. 2000 - The secondary products révolution ? Changes in animal management from the fourth to the fifth Millennium, at Arjoune, Syria, in : M. Mashkour, A. M. Choyke, H. Buitenhuis, F. Poplin (dir.), Archaeozoology of the Near East, IV B, Groningen, p. 12-28.

HARLAN, J . R. 1967.- A wild wheat harvest in Turkey, Archaeology, 20, p. 197-201.

U n i MAN G . C , DAMES M. S., 1999 - Domestication rate in Wild Wheats and Barley under primitive cnltivation : preliminary results and archaeological implications of field measure-ments of sélection coefficient, in : P. C . Anderson (dir.), Préhistory of agriculture. New expéri­mental and ethnographie approaches, Monograph 40, Institute of archaeology, university of California, Los Angeles, p. 70-102.

474

A N C I E N T T H R E S H I N G F L O O R S . T H R E S H I N G T O O I . S A N D P L A N T R E M A I N S IN T \ D A V . \ L L E Y , I S R A Ë L

HORNELJ., 1 9 3 0 . - T h e Cypriot treshing sledge, Man, 30, p. 135-139.

JEAN C. F., HOFTIIZERJ., 1965.- Dictionnaire des inscriptions sémitiques de l'Ouest, I.eiden.

MURRAY A. M., 2000 - Cereal Production and Processing, in: P. T. Nicholson, I. Shaw (dir.), Ancient egyptian mutinais and technoloirs, Cambridge, p. 523-531).

NAVEH.J. 1982 . - Early history ofthe alphabet, Jérusalem.

PEARLMAN D. A., 1984.- Threshing sledges in the east mediterranean : ethnoarachaeology with Cher! knappers and dhoukanes in Cyprus, unpublished MA thesis, University of Minnesota, Minneapolis, 301 p.

PEARLMAN D. A., 1 9 8 7 . - O n the Dhoukani Trail, Cyprus Life, 1. p. 57-58.

PELEGRIN J . , 2 0 0 2 . - Chapitre annexe, La reconnaissance des techniques de taille, in : J . Chabot (dir.), Tell Atij-Tell Cudeda. Industrie lithique. Analyse technologique et fonctionnelle. Editions du CELAT, Collection « Cahiers d'archéologie du C E L A T », n° 13 et « Série Archéométrie », n° 3, Québec, 225 p.

RAMSEYC. B., 2 0 0 0 . - Oxford calibration program 3.4. http://www.rlhah.ox.ac.uk

ROSEN S., 1 9 8 3 - T h e Canaanean blade and the early Bron/e âge. II.f. 33. p. 15-20.

ROSEN S., 1997.- Lithics after the Bronze âge, A handbook of stone tools from the hvant, Walnut Creek, Ca., AltaMira Press, 184 p.

SHERRVTT A., 1 9 8 1 - Plough and pastoralism : aspects of the Secondary Products Révolution. in: I. Hodder, G. Isaac, N. Hammond (dir.), Pattern of the past : studies in honour of David Clarke, Cambridge, p. 261-305.

SkAkt'N \ \ 1000.- Evolution ol agricultural techniques in Eneolithic (Chalcolithic) Bulgaria : data from use-wear analysis, in: P. C . Anderson (dir.), Préhistory of agriculture. New expéri­mental and ethnographie approaches. Monograph 40. Institute of archaeology, university of California, Los Angeles, p. 199-210.

W H A I . I . O N R. J . , 1978.- Threshing sledge Hints. A distincove pattern of wear, Paléorient, I. p. 319-324.

WHITTAKER J . -C . , 1996.- Athkiajas : a Cypriot flintknapper and the threshing sledge industry, Lithic technology, 21, p. 108-120.

WHITTAKERJ. C , 1999 - Alonia : the ethnoarchaeologv of Cypriot threshing floors, JMA, 12. p. 7-25.

WHITTAKERJ. C. , 2 0 0 0 - T h e ethnoarchaeology of threshing in Cyprus, NEA, 63, p. 62-69.