Investigations at Ramat Saharonim: A Desert Neolithic Sacred Precinct in the Central Negev Steven Rosen, Fanny Bocquentin, Yoav Avni, Naomi Porat

To cite this version:

Steven Rosen, Fanny Bocquentin, Yoav Avni, Naomi Porat. Investigations at Ramat Saharonim: A Desert Neolithic Sacred Precinct in the Central Negev. Bulletin of the American Schools of Oriental Research, American Schools of Oriental Research, In press, pp.1-27. ￿hal-02014702￿

HAL Id: hal-02014702 https://hal.parisnanterre.fr//hal-02014702 Submitted on 11 Feb 2019

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non, lished or not. The documents may come from émanant des établissements d’enseignement et de teaching and research institutions in France or recherche français ou étrangers, des laboratoires abroad, or from public or private research centers. publics ou privés. Investigations at Ramat Saharonirn: A Desert Neolithic Sacred Precinct in the Central Negev

STEVEN A. ROSEN F ANNY BOCQUENTIN Archaeological Division Centre Nationale Recherche Français Ben-Gurion Uni versity of the Negev MAE Uni versité Paris-X POB 653 Paris, France Beersheva. lsrael fa nny.bocquentin @libertysurf.fr [email protected]

YOAV AVNI NAOMl PORAT Geological Survey of Israel GeologicaJ Survey of Israel 30 Malkhei Israel Street 30 Malkhei Israel Street Jerusalem 95501. lsrael Jerusalem 95501, lsrael [email protected] naom [email protected]. i1

fn ve.stigations aJ t/l e open-air sltrf11e ami cairn complex at Ramat Saharonim in the Makhtesh Ramon in 1/te central Negev reveal a sai:red precinc/ or ritua/ ce111er with a focus on a morcuary cult, artributable to .tlie Late Neolirliir, ra. 5000 n.c. The four shrines are aligned wirh the set1in1: Still of the s11111mer solstice. along with other land­ scape features. The tilree /1111111/i l!Xrm•ared, mughly rontemporary with the shrines. re­ vea/ed primary and secondary /Juriai.f and inlentimia/ bnne realig11111ent. Excavations al Shrine 4 allow derailed recon'i/rttf'lirm. of siteformatio11 pmces.ves, demonstrating fo11g- 1erm development of the fen111res of the complex. ln general, the megalithic aspect of the site, 1he symbolic nspear 1~{ tlw alig11111en1.1·, and the a11riburio11 11• the Lme Neolithic suggest a close rela1io11 s/Jip berween 1he rise of the de.sert cuit and tribal sociery asso­ ciated witlz the earlieJI intmductin11 of domestic herd animais i11 to the central Negev.

LN T RODUCTLON ca. 5500- 5000 B.c.. onl y slightly postdates the ear­ liest i11fiJtration and adoption of berd animals-sheep s with the transition from hu111ing-galbering and goat- replacing hunting as a primary subsistence to farming in the Mediterranean zone. the base. A rise of pastoral nomadism in the desert per­ The presence of ancient cuit and mortuary sites iphery entailed far-reacbing transformdeserts, in the Late Neoli1hic 1990; 2002; also Yisrael and Nachlieli L998). Other difficult-to-classify features. such as Lh e " K-Line" Open-air shrinei., sometimes referred to as tem­ (e.g., Haiman 2000) may also be included in Lhe ples, have also been investigated. Yogev (1983) ex­ general category. The general c lass of c alL struc­ cavaced a courlyard s hrine with stelae in a cis t in tures ranges in date from the sixtb mHlennium a.c. the focal corner of the shrine. in the Uvda Valley. through recent lime s-a~. for example, in open-air llated to the sixlh millennium cal s.c. Eddy and Wen­ mosques- and thus has been assocint:ed with the com­ dorf ( J 998; 1999: 36, 39) documented a rectangular plete complement of cultures lmown in Lhe deser1 shrine in eastern Sinaï similar to those al Ramat from the historical and pmtohistorical periods. Saharonim. also dating it by radiocarbon to the sixth ln cerms of the earlier part of this long Lime pan, milleon.ium caJ o.c., and Rothenberg (1979: 125. several recenc investigations provide imporiant back­ lig. 28) excavated anothcr, uggesting it be dated 10 ground for the excavations at Ramat Saharonim. the Pre-Ponery Neolithic on the basis of associated The earliest of these is the systematic exploration of anifacts. Given the absence of artifacts ar vinoally the nawamis fields al Ein el Hudera (Bar-Yosef et al. ail other such sites, and 1heir consistent Late Neo­ J 977) and Geb.el Gunna (Bar- Yosef et al. 1986) in lithic attributîon, il is likcly ù1at the shrine at Ein Sinai. These studies documented the nawamis. cy­ Yarka was builL on an earlier occupation. More sig­ lindrical corbel arched buildings usually 4-8 m in nificantly, Avner's (e.g., 1984: 1990: 2002; Avner extemal diameter and 2 m high, as mortuary struc­ and Cam1i 2001) pioneering long-tenn research pro­ tures dating to roughly the early fourth millennium gram on tbe desen cuit has documented numerous B.c. They reflect a local pastoral society (e.g., Coren hrines in the Negev and Sinaï, many typologjcally 1980) organized at a tribal level. wilh cultural linki. 10 idemical 10 those of Ramat Saharooirn. He. too. has Egypt. La1er analyses (Bar- Yosef et al. 1983; Hersh­ dated the origins of these structures to the Late kowitz et al. 1985) also established a seasonal and Nenlithic and has noced solstice alignments which he cosmological aspect to the nnwarnis; doorways are has interpreted in a cosmological framework drawn aligned to the west, facing the sening sun. with de­ from lmer Mesopotamian civilization (Avner 2002: viations apparently in accordance with the season of 102-J). On chis basis. he suggests a winter sunrise as construction but wiLh modaliUes suggesüog seasonal opposed to a summer sunsel orientation. preferences. The focus on the setling sun clearly has Wilh respect to Ramat Saharonim itself, Cohen symbolic meaning and is tied Lo Egyptiao beliefs in cxplored the site in the 1970s. publishiog plans and lhe connections be1ween dealh, lhe west. and the set­ n few surface artifacts in his doctoral thesis ( 1986: ting sun. 8-9. pis. 5-6; also see Avner 2002: table 14.9-12) The tumulus fields of the central Negev are con­ and Inter in hil> synthetic study of che Negev High­ ceptualJ y si milar to the nawamis in terms of spatial land ( 1999: 2 1-24). Tn formal test excavations werc clustering of the structures and 1heir morruary asso­ also cbnducted but never publisbed. Chronologi­ ciations. However, unlike 1·he nawamis, tumu li ex­ call y. Cohen attributed the site to the Chalcolithic cavations have usuall y (bu t not aJways) shown Lhem perioù based on su.rfacc ar1ifac1s, including several to be empty of burial remains, either from poor prcs­ wbu lar scrapers !Cohen 1999: fig. 9: 1. 2, 7) and a ervation or perhaps the removal of 1hc bones (c.g., simple bifacially retouchcd knife (Cohen 1999: fig. Haiman 1992). Also contrasting wilh the nawnmis, 9: 11 ). ln fact. tabular scrapers as a class appear in buriaJ goods are rare in the tu mu li, rendering chrono­ 1he late stages of ù1e Pottery Neoli thic (second half cultural anribution problematic. Although rcciangu­ of Lhe sixtb millennium B.C.) (Ro en L997: 75) and lar tumuli have usually been auribULed ro the ln1er­ continue through the Early Bronze Age. The bifacial mediate Bronze Age (= EB l V = MB l), Haiman knifo is 1101 diagnostic. lniûal assessments based on ( 1992: l993) bas suggested that the s tandard round survey work also tended 1oward the Early Bronze tumuJi be dated to Lhe Early Bronze Age, based on a Age attribution (Rosen and Rosen 2003). especially pauern of geographic association with large Early based on the strong architectural simiJarities be­ Bronze Age campsites. Avner (2002: 154-55) has lween the Early Bronze Age tumuli in the Negev presented radiocarbon date indicaling earlier oc­ Highlands (Haiman 1992: 1993) and those of Ramat currences. and, anticipaling later discussion. the Ra­ Sahnronim. mat Saharonim excavations indicaie a deeper his1ory The investigations at Ramai Sabaronim were ini- with the tradilion of tumulus burial beginning in 1he 1ia1ed in order to build on these pioneering works. Late Neolithic. The general goal of 1he project was to survey the site (Rosen and Rosen 2003) and excavate part of il m,ation (Zak 1968) (prirnarily, Limestone wiù1 clays, in a methodologicaJly dgorous fashion in anticipa­ maris. and some sandstones) to the Middle Jl!rassic tion that the greater detail would provide answers lo lnmar Formation (sandstone) (fig. 3). some of the quesùons conceming the early desert Historicall y, the area has been primarily the reaJm cuit not yet rcsolved. Three sets of issues were LO be of pastoral nomadic societies. The Azazmeh and addressed: Saidiyyin Bedouin tribes inbabited the region in the l 9lh and early 20th centuries. During the periods of 1. The explication of the relationships. chrono­ clnssical antiquity, the Early 1slamic and Byzantine, logical and cultural, between the different com­ Ro man, and Hellenis1ic periods. pastoral sites of vari­ ponents of the site. using a suite of field and ous kinds dominate the arcbaeology of the Makhtesh laboratory methods. including assay in both Ramon. including most notably Nabataean campsiles radiocarbon and opticaJly stimulated lumines­ (e.g .. Rosen 1993). lndeed. the site lies adjacent to cence (OSL) datiJ1g. the Nabataean spice route leading from Perra Lo Gaza 2. The documentation of ù1e slratigraphy of o (e.g .. Cohen 1982). Although intensive run-off irri­ shrine in order to understand its construction, ils gation farming was practiced in desert areas some­ original fon11 . the different phases of develop­ whal farther north, in ùie lrano-Turaninn zone, ùiis menl, and the post-abandonment formation pro­ re.gion remained unexploited agriculturally. With cesses that result in the current ite features. re pect to site prior LO the classical era. survey in 3. Examinaiion of the relaùonship between the the general region has revealed campsi1es primariJy site, the landscape. and Olher natural features. from lhe Early Bronze nnd lntermediate Bronze Ages, including astronomical aspects (cf. Ti lJey 1994; although not in the immediate vicinity of the site. Cannichael et al. 1.994). Neolithic sites are known from the northern Arava (e.g.• Taute 1994). farther south, in the Uvda Valley The first phase of the project. initiated in 1999, (e.g.. Goring-Morris and Gopber 1983; Avner 1990). consisted of the intensive mapping of the site (Rosen and in lùgher area.-; to the west and north (e.g.. Noy and Rosen 2003). A large-scale map of the entire u111d Cohen 1974; Rosen 2002: Goring-Morris 1993: precinct was prepared, and 1:20 stonc-by-stone plans Si mmons 1981 ). Across the Rift Valley. the bigh­ of each shrine were completed. The solstice orient<-t· lands of southèrn Jordan saw the evolution of village lion of the shrines was also documented. and Ltrbao settlements from the Nêolithic thrnugh the Bronze Age in a less arid cnvironment more suilable lo scdentary and agricultural pursuits. No habitation THE REG JON AND T H E sites are found in the immediate vicinity of Ramat S ITE A R EA Saharonim.

The cuit complex at R amat Saharonim (Israel Grid 1434/0035) is located south of Mt. Ardon in the GENERAL Sfl'E DESC RlP TLON eastem half of 1he Makhtesh Ramon. a large ero­ sionaJ cirque (e.g.. Y. Avni 1993: Zilbem1an 2000) The Ramat Saharonim sacred precinct (fig. 2) located in the soulhem Negev Highlands (fig. 1). The con ists of three areas: ( l ) the Sbrine Area. wilh four region is a rocky desert, receiving roughly 75 mm shrines and associated instaUations: (2) Ramat Sa­ of rainfall per year, and is characlerized by sparse haronim East, consisting of L4 tumuli on 1wo parallel Saharo-Arabian vegetatfon (e.g .. Roseni\11 and Gllead cuesta cliffs east of ù1e Shrine Area: and (3) cbe l 985a: l 985b; Danin 1983: 35. 53). Surface sedi­ Southern Ridge, with 16 tumuli aligned on the cuesta ment are reg soils. and the shrines resl on a devel­ cfüf south of the Shrine Area and Ramat Saharonim oped desert pavement with a sandy substrale. The East. ln addition 10 the cuit complex. a sundstone tumuli rest on limestoae bedrock. Geomorpbologi­ quarry for the production of milling stones is locaced cally, the site is located in a shnllow valley fonned approximately 100-200 m nortb of the Shrine Area. between sets of paraUel cuesta clitfa varying in probably dating 10 the Early Bronze Age (Abadi heigbt from roughly 2 m in the west to up Lo 5 m in 2003; Abadi and Rosen in press). tbe east (fig. 2). GeologicaJJy, the site lies on the tran­ The Shrine Area con s i sL~ of four courtyard shrines sition from the Lower- Middle Jurassic Ardon For- (numbered Shlines 1- 4; fig. 4). A detailed descrip1ion •

Mediterranean Sea

Fig. 1. Location map showing the Mskhtesh Ramon and the Negev ln the Levant, and the location of Ramat Sa­ haronim with1n the Makhtesh Ramon. ,.. ,\ .. •: 1 '• 11~ ~" ~ t

Fig. 2. Aenal photograph of Ramat Saharonim wlth srte leatures tndlcated

Fig. 3. Geolog1cal sec1ion or Ramat Saharonlm. Note lhat vertical scale 1s not applicable to horizontal measurements. c c

\ 12 1 " B ., ,,,. .... c .. " .., ~...... ; .. ~ 11 ~~~ ..... , ~ c ;, çl';; • !r . ·~ ...f . ·; ~ .l; ~ c ' . ....~ B ,):·~~,,'\ D D D ...... Il Shnnr 4 11 ~ 11

10•

B A Upper conglamcrate Ramat B Lower conglomelllte wcll prclCIYl:

Fig. 4. Deta1led schemallc of shrines in relahon to geological and geomorphologlcal units. Scales are approx1mate. Fig. S. Norlhwest v1ew of unexcavaled Shrine 3 loward Lower Cre1aceous volcanlc Ml. Ga'ash and the northern wall of lhe Makhtesh Ramon Nole the orientation between fow hills on e1ther sfde of lhe shnne.

of the individual shrinc!.1 as bai-.cd on 'urvey has been The sccondury stn1cturcs, buih on the north side presented in the repon on 1hc 'itc survcy (Rosen and of 1he primtu·y shrines. ure square or ni.!ar square Roscn 2003; nlso sec Cohen 1999: 21-24; Avner structures, upprnximately 8 mon a sic.li.!. built of what 2002: table 14: 10-12). but se verni fc111urcs derived appearJ. to have been a si ngle row and single course from the '! urvcy arc imporcant for undcri.tanding the t1f roumkd wudi cobbles placcd carefully one against nature of the complex, a wcll a' it., chronology and the othcr n1ey each also show an intcrn:ù feature: a development. First. each shnne i. how' two compo­ mali '-lone pile. poorly pre.\crved and d10icuh to de­ nentl.. a largcr. primary rectangular 'tructure. and a scnbe. TI1c \Ource for the i.tone of 1hc econdary maller. more squarii.h '>Ccondttr) one. located on i.tructurc' appcar. to be conglomerate e~pol>ures lo­ the nonh side or the primary. Although Shrine 3 laèb cated in the immediatc vicinity of the 5hrinc (fig. 5). thii. structure, it 'hows rcmains of u dilTerent charac­ The con1ra ... 1 with the primary s1ructurci. is triking. ter. vaguely rcminisccnt of the .,ecundary ~trucLUres As abovc, Shnne 3 doei. not show a ·imilar second­ associ:ued with the other shrines. but con'ltructed in ary 'itructure. although smalt "fonce" slabi. suggc t a difTcrent fa.,hion and IC\'\ well prescrvcd. Avoer 1ha1 some other feature remim cent ol the 'ecoodary {2002: 120-22. 126) ha ~ refcrred lO lhese pairs or ... 1ruc1urcs walt prei.ent (contra Cohen 1999: fig. 23: i.tructure a temple pairi. or 1w1n temple<.. 1mplying Avner 2002: fig. 5:2). 1heir con1emporaneity and llUggc.,ting thcy con~ù tute ln add111on to the 'econdary '>lructure ... nonde­ male and femalc pairs, perhap <;pecihc godl> and '>Cnpl slone 'cutter in roughly linear pattern!>. appar­ go 'C!.. ln thil> light, the importance of establish­ enlly panially lhe re ults of human fü: ti v iti e~. and a ing 1hc chronological relationship j., clcar. few iima ll coni.1ruc1ed features ~uch a!. boxe!. made The priniary structures are cach on the order of of small limcl>tonc slabs. arc locatcd uround 30 m 20-22 m in leng1h. They cons1s1 of a IHrge forw ard wesl of Shrinci. 1. 2. and 4 and coincidl.' wi1h con­ wnll on the wcs1 side of the rcctungle. built or two glomcratl.! cxposures. Avner {2002: 1 16- 17, fig. 5:2) rows of large limestone blocks or slabs with a space refen. to thc!tc remuins as "circlc chains," but it is dif­ of 20-40 cm be1ween the rows. and in the east. a ficult to 'lce any patterns. courtyard, fcnced off by a si ngle row. '>ingle course, The 30 large tumuli at Ramat Saharonim East and or small Monc slab . This fencc i' now fallen but the Southern Ridge are arranged in 1wo rough lines origmally stood upright, a<; rcnected 111 a few slabs along the parallel cuc!.ta cliffo;. '!Omewhat converg­ -;till embeddcd in the ground in Shnne 1 The source ing toward the ca ... 1. The tumuli arc cach 4-8 macro s for 1he limcstone bloclo..s (sec di\CUS'>ton below) is m the ba\I.' and about 1- 2 m high. con.,tructed of locaJ. The we 1ern wall. ongm:illy slood to a height limestonc bluet....:;. ln many cases. a margin of larger of about 1.5 m. ba ed on the pre,crvcd height of the '>IOnes i' cv1dent around the ba.,ul circumference. walls and the quantity of ... 1one fall on 1hc surface and wilh the remaining -.1one~ of the cairn piled more found in excavation. haphazardly on top. MoM of the tumuli are concave on top. and excavations at 01hcr 'itci. (e.g.. Haiman 1992: 1993). U'> well ~Ls at Ramat Saharonim. indicate 1Thc teml ..1 hnnc .. 1 ~ emplo}cd hcrc 111 ordc:r11111,01d wme of the prc. encc of bunal ci t. in .. ide. often lacking skel­ the p1tlolls of u~mg londcd 1enn• -uch J\ .. temple .. or .. ,illlclll.ir)." etons. al tht: b~e of 1hc tumulu~ . A . ingle cairn wh1c:h rc,onuu.: "uh 01hcr mcan111g,. \UO:h '" hou\c or god. or pro­ opcned in the 1980s by urcbaeologi ·1s '1dmg shcher, etc Other opt1m1'. \UCh "' "euh ~tructun::· pc:rhap ~ tcchmcnll} more corrcc1. 'eem awk" urd working for the l rael Department of Antiqu iti e~ revcaled a keleton (Y. Israel. persona! communica- Shrine Aren were constructed (figs. 4 and 6). 1ion. 2000). nllowing beuer comprehensioa of 1he detail. The specific locations of lhe featurei. in lhe com­ of 1he placement of individual archaeological plex were clearly chosen for Lheir position!> amid the feaiure in the l311dscape. smaJI- cale topographie relief and larger -scale lnnd- 3. Geoarchaeological survey (on which lhe ectioa cape features. Alignmems of the shrines were derer­ and maps were based) also locatcd special gco­ mined by landscape feamres- mos1 nouibly a large. logical features, sucb as the limestone quarries/ black volcanic mouniain in the distance (fig. 5)-ancl cxposures from which lhe building material. for to accord generally with the seuing sun of the 1>um­ the primary shrines and mmuli were taken. Thi mcr olsticc. with nzimuth deviations from only 2° survey also documented the location of the con­ to 8° (Roscn and Roscn 2003). Thrce of the four glomcrate e.xpo ures from which the cobbles shrines were placed so as 10 view this sol1>1ice i.uni.c!I used in 1he consLruction of the seconclary . truc­ in Lh e shallow depre sion between two low hi lb (fig. lllrc origioated. 5). The oricntution lOward spccific geogrnphic fca­ ture. in the nonhwes1 <: trongly suppons the summt!r The gcological section was constructed using an sols1icc sunset interprclation of 1he alignmc111, ai. cxposurc east of lhc site area. in the area or the wa- oppose(.) 10 the wintcr sunrise suggcs ced by Avncr 1ershcd berween Naha! Ardon and Nahal Ramon. (2002: 102-3). for whicb no geographic pmtern s arc The sec tion tilts 7°-10° 10 the norlh. resulting in a evident. Although il is diflicuh 10 pcrceivc any clèar typical laycred eue ta. The specilics of the section pattern with respect LO the placement of 1he wmuli ure sttmmarized in tigure 3. The uppem1os1 unit of vis à vi th e shrines, the cliffs constitute folsc hori­ the i.cction. the lnmar Formation. dates to the Middle zons. wilh Lhe tumuli visible in silhouetLc from grcm Jurasi.ic. Il provide. the source materials for the !>a nd­ dis1ances. seeming to integrate visually wnh more "'onc milling <,tone quarrics mcntioncd earlier (Abadi distant ridges. 2 The shrines are locaied tll the open 2003: Abad1 and Ro en in press). The iumuli and (ea~ t ) end, between the two lines of 1umuli. They nre. shnncl. are associated with different facies of the on one hand, set off from the tumuli, and on 1he Ardon Ponna1ion, dating tO the Lower-Middle Juras­ other. merge with them, forming :l large- cale cum­ '>ÎC (fig. 3). The transitional Trias ic- Jura sic Mishor ple.>. . There can be littlc doubt as to the deliberate Fonnt11ion lies bencath the Ardon Formation. choice in thesc alignments and Lhc placement ot lhe ln nddition to the geological section. geological StlC features. and geomorphologica.I mapping defined imponant femures in the landscape (fig. 6). Survey wa,<; con­ T H E GEOLOGY ANO ductcd u-;ing a 1:5 000 color aerial photograph pm­ GEOMORl'HOLOGY OF T HE SITE duced hy Ofck Aerial Photographs Lld. in 1989, cnlargèd 10 1: 4000. Geologically, Ram at Saharonim The special nature of the cuit complex ai R<1mm cun hù dividcd into two areas. The geology of che Saharonim suggested 1ha1 rela1ively straigh1forward, consisting geological and geomorphological con1ex1s might o (. of a r-cquencc of cues1as rilted 7°-10° to the nonh­ fer important insigh1s into understanding various is­ CtruCtcd (fig. 3). providing a kcy for un­ em . the rc~u lt of a nonhwest-soutbeast fault which derstanding û1e nature of 1hc site sediments, creatcd f\ tructural. geomorphological. and topograph­ 1hcir di tribu1ion, and 1hcir ultimaie origin!.. ical anomalies relative to lhe eastern area. The ~ub - 2. Dctailed geological and geomorphologicnl idcnce of the block south of the fault line created a maps of both the general site area and 1hc long. shullo\\ vaUey in10 which ahal Ramon pcne­ Lr:ucd at leas1 twice during 1he Pleistocene. depos­

: yu, al Yd.uueli suggc~ t cd 1he pos~ibihlu~' of a rcln11on\h1p iting two conglomcrnte units along both the main bc 1w~n the di\1ant ndgcs and th.: tumuli on th.: ncarby .:l1fl\. chnnncl and itt- tributaries. The two units are 15 m 0035 1 ,1 1 1432 1 1 I I I §l 1440 I $° I ~ I ~ l <:'$ ).

11 ' 1~ ' ' . { . , . ...J.. 0035 7-10• ' . 7-10 ' B Lower cooglomerate we!J preserved ' c Lower cooglomerate eroded D Surface covaal by local rock fragments M Mjsbor ·rormanon p Playa ll No. of Sll'lltigraphic unit '<. Dip

1432 "<-z Fault Une .. - .. - .. Dike ,, ,,- Drainage line ·---,, Cliff / zoom ~ N

0030 1440

Fig. 6. Map showing geomorphological and geological features. The stratigraphie unit numbers correspond to the stratigraphie units in figure 3. and 10 m above the modern chan11 cl of Nahul RH­ the shrinc, beyond the area immediately adjacent lo mon. They show similar components. primarily large the primary wall , rendered excavation in 1his area cobble , 20-40 cm in diamett:I', of limesrone, dolo­ pointless. and it was not tested. Ali sedimenls were mite, llînt. harù sandstone. and basall, t.h e bedrock of sieved through 2- to 3-mm mesh , but in the event , Makhtesh Ramon. These terraces arc loca1ed in close no in situ artifacts were recovered. Severa! sections proximity co the hrîncs and providcd rhe source wcre intentionally left intact. bo1h for later investi­ materials for the secondary structures of the shrines. ga1io11 and for stratigraphie documentation. Thal is. The g.coarchaeologîcal su.rvey focused on the areas elle cntire lengtb of the shrine wall was not excavated in the immediate vicinity of the shrines (tig. 5) but since total exposure was considered both scientifi­ exrendcd throughout the site arca. or particular note cally unnecessary and potentially damaging to larer js the presence of small-scate quarries from which rcsearch. The interior of the primary wall comprised limestone blocks were pried out of the bedrock u ~­ smaller cobbles in an apparently intentional fill (fig. ing fi ssures already present in the bcdrock, provid­ 8). This was cleared in the squares opened. but the in ing block 40-60 cm in length of Lhe primary wall were left in place. These quarries are particularly nbtable in Uni t 11 leaving the W:\11 itself intact. of the Ardon Fonnation, 200-300 m nonh of the The excavalious revealed a massive double wail. ~ hrinc s and somewhat closer to the tumuli of Ramar with 20-40 cm between the rows filled with cobbles Saharonim East No drag marks were eviderll bc­ and smaller slabs (fig. 8), which served as the west­ tween the quarries and the shrines or the tumul i. ern wall of the primary shrine. The double wall was Ahhough lrnowledgc of modern geological sci­ built or large limestone blocks, some up to 450 kg ence was obviously not a prerequisitc for si1u a1ing. (based on linear dimensions and a specific gravity of the Ramat Saharonim shrines in ancient limes. eanh 2.5 for limestone). and many greater than 1OO kg. sciences anaJyses indicate that the locales cbosen for The wall is preserved to a beight of approxirnately the shrines are indecd unusual. For cxample. lhe va­ 0.75 m. Considering che stonc fall present on both riery of surface colors and textures, rcflected in the sides of 1he structure, the original height of the wall cliJferent geological and geomorphological units pres­ can be estimated at about 1.5 m, and a conscrvative ent in the dlrect vicin ity of the shrines. con1 rasts estimmc of the total mass of the western wall, includ­ sjgnlficantly wïth 1he eastem part or Ramat Sahar­ ing inccrnal 1111 , ii. 30 tons.3 onim. Shrines 1 and 4 arc even alig11ed wi th color Stratigraphically. four general units can be de­ contrasts such that the structures dernarcate unit dis­ fined surrounding the double wall. on either side and tinctjons, reflected in surface cotors und textures, Jt atso at each of the ends (figs. 9, 10): is difficult to be unequi vocal in rcconstructing chc Unit/: The modern land surface in the vici nity of specific motivation and decisions in 1he placement 1be sh1;nc, beyond 1hc confines of the site itself. is a of lhc shrine on one spot or ano1 hcr: however, givcn defl atcd desert pavement. Il shows a stony gravel and the solstice aJ ignment and the clear topographie dc­ cohblc surface with a substrate of red saody clay or cisions in placement, mi cro-deci~ i on s based on i; ur­ sill with calcium carbonate nodules. probably reflect­ face colon, and texture do not ~eem too far-fetchcd. .iog Pleistocene pedogenesis known io the Ncgev Of course. speculation as to meuning is beyond our Highlands (cf. Avni and Porat 2002). reacb. Unit 1/; The upper surface of the excavation con­ sists of a crust of silt, sand. and grave! 1- 2 cm thick, EXCAVATIONS AT S HR lNE 4 wi th occasional limestone blocks fallen from the double walls rcsting on and in it. This surface Following the carlier survey work resuhing in 1:2 0 stone-by-stone sit e plans ( fig. 4; aJso Rosc.:n 1Avncr (2002; 100) ~ugge$ IS a gencral helgbt of half thi ~ for and Rosen 2003), excavati ons at Shrine 4 were con­ open-air sa11ctu1lfics. bul his suggestion wa.~ made prior 10 i:xca­ ducted according to a mcter grid square (fig. 7) and valions .11 Ramm Saharonim and lhe documcn1u1ion of large quan­ arbitrary levels (spits) of 5 cm. 111 the presence of liuc~ of sione fait. He also sur;gests thauomc of the slabs wcrc se1 horilonwlly, sort or .1~ roofing ~labs . ll1i~ wa; indeed the apparent discernible surfaces (the original land surface), the plcturc hu~od on survcy, hui is not borne oui in excavatiC)n, The arbicrary spit level was abandoned in favor of the appurcnl rnofing slubs ;ire fal ic·n and not present over most of the natural surface. DeAation in the courtyard area of double wall, .... 0 0 o • a • ... D a 0 "o ae 0, • ~ a.i ~ cP ..~~ ... ~o "I 1l). .. 0 • D t • ~ ~ 'O 434 t> -~ ' Ramat -~ 433 0 ...... lllf -.!:!! "q\ ..,. . -"".. 0 • • • ,1·· ...... Saharonim 431 (. ':. o-:'· f - ..m ~ 431 c ' .~ ·, r::: ~ 430 c 1) O• • o .. Shrine 4 ....." . 4l9 .. q. ... w OO . . ~ . 418 - c • , .. • .·~ DOUBLE w~ ~ o •• 0 r. •• •••• 0 ... .. :l .JI! • "1. o<> . f' ~ .. ' J' . • & ~ - r::to :" 4l5 0.0 Cft c. 'i 1F,.. '14 0 • • o .. ,. . 413 H ... "' . ' . 1 11 "" .... • C) : ' • 0 412 ""',, 0 ••0 0, COURTVARO ... 4ll l - "' • 410 rt 1 ~ d .,...... 419 - "' .1 • . ,. 0 .. 418 .. " . 417 ~ - ' 0' ~ 416 "' .,r' "1 B 415 • " 1,,-ll, .. 414 • 413 11 -p ~ ~ 412 - ï:JJ ~ - Fig. 7. Plan of excavations of Shrlne 4: (A) photograph ol secondary structure; (B) photograph of prlmary structure betore excavation: (C) grid and plan of excavations of prlmary structure: (D) plan of unexcavated areas. Note that the darkened areas ln excavation grid are unexcavated. Fig. 8. Photographs oi primary double wall: (A) portion of wall wlth some 1nterior till duringJatter excavation; (8) portion of wall before excavation; (C} vlew of entire wall af1er excavation (note thal the foreground ls unexcavated; also note the presence of sections left in place); (0) view of section of the western face of the primary wall.

reache~ to the top or to within a few centimeters of modern land surface, but bas been disturbed and is the top of the preserved wall height and slopes awuy lacking the desert pavemenl The contact between from the wall. this layer and Unit Ill is sharp and clear. TJ1e deeper Unit Ill: This layer is comprised of ligbt gray­ parts of rhis layer are more consolidated, supporting brown silt andsand . li is 20-40 cm thick adjacent lo the idea or disturbance of the upper portion. Indeed, Lhe double wall and tapers to a feather t.ermination activities hy the excavators around the site destroyed away from the wall. le lies directly beneath the uppcr the dcscrl pavement. leaving a norizon equivalenL surface crnst anation of ac ti vi tics around the shrioe, the massive double wa ll actcd ai; n sand trap. accumulating wi nd­ blow11 sedimcnts (Slrntigraphic Unit Ll i) all around the structure. like a dune. abovc the origi nal l:rnd sur­ face (Strutigraphic Un it IV). The higher elemcnl!. of the structure also cxpo~c" unprotected areas to wi nd de­ natio11 and to local cro-.ion caused by runoff (from Stage l - Pre·occupat1011 (fig. 11 · / ): The land­ lhe increu.,.ed relief) during the infrequeot rain events. scapc is essemially n:u. with a cover of <> mail gravel This cau,es the stones of the coun ya rd fence to (2- 5 cm: desen pavement). Lhe uppermosl part of a 1opplc. and after the upright stones have fallcn. ped­ yo ung lope also on the north Thi phase conSÎ'> t'> of Lhe construction of the pri­ and 'lOuth ends of the structure. uJtimately capped mary double watt. ca. 1.5 m h1gh. Il '' al-;o po~sib le by the 'urface cru -.t (Stratigraphie Unit Il ). Round that the counya rd fcncc was constructcd al &hi -. lime. conglomeralc cobble~ of the subi.idiary tructure but there is no tirm evidcnce conccrning th1 .., . The ac­ are placed on these accurnu lati oni.. resting in and tivities in and around the shrine destroycd the desert abovc Stratigraphie Uni1 li (fi gs. 9. 11 ). Therefore, pavement (Stratigraphie Unit 1). leaving only the soft Lhcy mulil postdate the double wall construction by and porous sill and sand A horiton of the reg soil i.ome i.igniticunl span of timc, probably thou aods and underl ying palcoi.ol (Stratigraphie Unit IV). of ycars ('ee Absolute Chronology below). Section ~--OS

B E F O(; H 0(1, - ~ •o< b e ~ "0 4pl ~ 1 ~ ~ 434 160 43 3 ,If' 0 . ~, 170 432 117 12S Sec1lon 43Cl-OI ~ ' 185 ·-..::::::::;~';--'-:· ~· · ::::i 43 1 0 ~ ~ i-m 1...-: · ~ •• · ·~LI ~ ~ 430 196 r::;C (\ ~ l:M [189 Il 42 9 89 _q;t :Q Secdon 427-'f28 33 .,~ 42 8 Il 0 ' 42 7 35 r} . 190 [' • . 130 1 42 6 101 A 01 ' ';J' '"- ç 0 c 42 s .~

Section 417""'8 18 42 0 (l ~ ns 1 !l 117 41 9 '(1fa ' 1) "~ ri~() 41 8 12 31 417 il~ 0 e 41 6 ~ ... ~ j;) .. ' . 41 0 Jj . ... s ....__.' ,, _ ~ ""' ~ ;:;J:;cz;s1\. .I ~ 4 ':J 1 l 41 . \' ..J 1 8 1o; c D jj,__ IJ 41 3 ~ 41. 2 jb71J

Fig. 1 O. Ramat Saharonlm Shrine 4 section drawlngs. Ramat Saharonim Shrine 4: Schematlc of Phases of Construction, Occupation, and Deposition

stones / desert pavement s 8 • 8 8 a a 0 • 0 8 s s s a s 8 s original surface /// ,.. //-///"'//~ 4/'sand/ loess o o / '/"' o ' subsurface 0 0 0 OO 0 0 0 0 0 0 o --- Pleistocene paleosol 1 . Pre-occupation stratigraphy wlth CaC03 nodules

primery fntenUonal stone flll 2. Construction I shrine ~ls :--.roeen walls occupation phase

courtyard 0 fenœ 0 dlsturbance of desert end subsurface deposlts

,,.-~~...... _---~--~~

/ "";./·,,,,,,, -/ / / / ..... ) ~ ...; 0 0 , 0 0 0 0 smell foundation 0 " tranches 0

fallen fence post-construction 3. Post-construction I slab on fallen sedlment accumulation abandonment pedes~ bl~k't _ Lr,ef::;;J__ phase . /e/~/~/~7,-:-/,-;-. -,,.- eros1on o o / 0 "' ,... "/ , 0 0 0 0 0 0 0 0

stones from secondary square structure . (restlng on sedlment post-constructton accumulation) sedlment accumulation / ~~ ~/ "o/ / ,,.....- ,,o// /u / / ·, post-constructlon ° " sediment accumulation Post-construction/ abandonment profile

Fig. 11 . Ramat Saharonim Shrlne 4 slle formation sequence. Fig. 12. Tumulus 28 looking west. wlth Tumulus 29 ln the background

EXCAVATION OF TUMULl 28, 29, AND The excavation revealed remai_ns of seven in· 1 30 0 THE SOUTHE RN RIDGE dividual.;;; six of whom could be associated witb the pcriod of the construction of tbe Lumuü. The seventh, Tumulu 28. Tumulus 29. and Tumulus 30-ihe round an 1lw upper level of Tumulus 29, clearly rep· westernmost 1umuli on the Southern Ridge (figs. 2. re~cn t !. a Inter reuse of the tu mulus. and indeed lies 12)-were excavated. Excavations wcre conducted almmt a metcr above the very poorly pre erved re· from th e top down inlo the cist of each 1umulus s-uch m111ns of an inûividual on the lowest surface of the that the structure of each was left iniact. A 11 sedi· tumulus. ments were sieved through 2- lO 3-mm mesh. The basic llLrucLure of the 1umuli is idcntical to 1'11 Ill""' s 2 8 that describcd by previous scholars {e.g., Haiman 1992: 1993). Tumulus cons1n1ction begins with 1hc Tumulus 28 has a relatively small cist opening at excavation of a shallow pit-in th e case of Ramm 1.75 m (nbovc dnium). wilh a rectangular shape ori· Saharonim, no more 1han 20 cm deep due LO the sh•ll · en1 ed northwest- southeas t. From the top. the cist low dcpth or surface sedimcms on the limcstonc lay· narrows fmm 0.70 tu 0.60 min width. but increases ers or the Ardon Formation. The interior wull of 1hc frorn 0.85 10 1.20 m in length. A few isolated human turnului:; was Lhen constructed around the . hallow pit. boncs werc round as high as l.65 m. being probably A ring of large margin stones was placed dellmiting movcd hy later ùisturbance. The main level of bone 1he ou1 er edge of the tumulus. The interior c:îs t was wns fCJund from 1.25 10 1.1 3 m. Two di fferent areas. constructed using horizontal slabs of varying lcngth!. the !ICluthcaM and the nonhwest. were clearly recog­ 10 forrn a rough oval or polygon. At some point 1hc nizc

Tumulus 29

Tumulus 29 has the largest cist or the lhree tumuli excavated ( 1.60 x 1. 10 m). The opening is pentag­ onal in s hape. oriented norlheast-southwei.t. The highest point of the tumulus is al 2. 12 m. and the highest human remains were discovered al 1.68 m. At this level, a weU-preserved and complete skeleton of an old woman (over 60 years old) wns found. The body was placed on the left lateral side with the head Fig. 14. Tumulu s 29: (A) view of upper skeleton, a second­ lifted up (fig. L4A). The legs were nexed as well as ary use of tumulus, (B) the lower scatter of bones. the arms. which were tight to the body with the right hand under the chin and the tert under the thorax. fibers, wcre ulso found. The excellent state of pres­ Pieces of desiccated leather ·urrounded the whole ervation, the pre ervation of Lhe leather. and the high skeleton. uggesting burial in a ti ghtened sack. or location in the tumulus suggest thar this burial wa shroud. Arouad the skull, pieces of a different k.iad ··iocrusive·· to the tumulus. as reflected clearly in the of organic macerial. perhap rope made of vegetal radiocarbon and OSL dates fixing the burial to the Fig. 15. Tumulus 30: (A) ce ntral cluster of bones wlth wall affect on south sida; (B) closeup view of articulated foot. second half of the ftrst millennium B.C.E. (see larer This couId be eitber the resull of a perishable struc­ discussion). The proximi ty to the Nabataean spice ture. which prevenred the dcposit of Lhe body or route (e.g .• Cohen 1982) suggests a Nabataean cul­ bones in tbis area, or the result of a specific tapho­ tural attribution. nomic process in tbis area thaL destroyed the re­ Under tllis skeleton, che internai structure of the main s :tfterward. In fact. the alignmenl of che bones cist was organized with slabs one on top of rhe other at the limit of the empty area suggests that the exis­ fillin g the east half of the cist down 10 the bottom Lence of a hard perishable structure, now gone, is the of lhe tumulus and a fill of infiltrated sandy matrix on most likely bypothesis. Analyses of Lhe sedimem the wcst side. in which a second cl us 1er of hum an re­ samplcs taken Crom this area will help to answer this mains was found between 1.02 and 0.89 m high (fig. quei.tion. 148). Only about 40 bone fragments, in a very poor The level of bones was approximately 20 cm statc of preservation, were found scattercd. These th ick. lsolated bones were mixed with parts of artic­ inclùde pieces of long bones. vertebrac, ribs, and ulmed skelecons. Ali catcgories of bones are pres­ teeth . More detaîled study is necessary bcforc a1- ent: long bones, skulls and mandibles, hand and foot tempting to ioterpret this assemblage in terms of bone ~. girdJe. and thorax. A minimum of three adult fune rary treatment, but there is no doubt that this is individuaL<; i present in the assemblage. Al Ica t 1wo part of the initial use of the tumulus. Small slabs of them remain parrially articulated at the bollom of paved the bouom of the cist immedintely beneath the pit (one complète left foot and leg [fig. 1581, one the bones. complete right hand and forearm. and a skull articu­ Ja1ed with the mandible). but later disturbance does Tumulus 30 not allow reconstruction of lhe original position of the bt)di e~. The few additional remains suggest Lhat Tumulus 30 is 1.84 m high. The opening of the a secondary burial was also made in Tumulus 30. Ac­ cist (0.95 x 0.75 m) was pentagonal and oricnced cording to the disturbancc&, the cist was visi ted sev­ nonheast-soutbwest. The cist was fillecl by infil­ eral times for funerary purposes. and the burials were Lrated sedünent almosl ro the top. lts internai stnu.:­ probably not contemporaneous but successive ones. ture was organized with two rows of slahs one above The final stage of the grave. with two skulls lying lhe 01her on lhe north side from 1.36 to 1.19 m high. agai ns1 each other and a cluster of long bones ut lhe Slabs were absent from lhe central part of lhe cisL, center of the cist. is certainly a deliberate reorgani­ and the south edge of these areas is filled onl y by zaLion of the bones. Eighc Conus sbells, with holes sediment. A few isolaced bone were found between drilled in the Hat end to form beads, were also re­ the slabs al the northwes1 corner of the cist, bu1 the covered at the burial level. Four of lhese were round major layer of human remains was round henemh in a cluster becween foot bones in articulation and a Lhe level of the slabs in the central and north part of skull (fi g. 158 ). the cist. No bones were round on the sou1h side, The bones in Tumulus 30 were resting in a shal­ and a very clear "waJ I effect" is shown (fig. 15A). low basin filled wilh stony sediment. The bouom is T ABL~ 1. Radiocarbon Determinations

14C Age± Context lu years Calihrated Calibra1ed guc Lab code (square, depth) 1}•pe 8. P. age B.C age B.P. %o PDB

RTI466J Shrinc 4 Chorcnal Cll1SU~40 5210-5050 (68.2%) 7170-7000 (68.2%) -24.6 (422F , 1.40-1.42) 5280-4990 (95.4%) 7230-6940 (95.4%) RTT4665 Shrine 4 bearth Chun:oal 5945 ±45 -19 10-4730 (68.2%) 6860-6680 (68.2%) (-23) (432E, 1.94) 494()-47 10 (95.4%) 6890-6660 (95.4%) RTI4664 Tumulus 29 L.:ulhcr 2225 ± 35 180-200 (68.2%) .23'.l0-2150 (68.2%) - 20.4 Upper burial 390- 200 !95.4CK) 2)4()-2150 (95 .4%)

situaLed in t:he middle of Lbe cist, l .OJ m ltbove da· at leasl in Tumulus 30. This tumulus is the best pre­ Lum, at approximarely the base of the tumulus ilself. servcd and seems not to have been reopened since Al this level, the cist was much larger Lh un Lbe upper Neoli1hic limes. The state of preservation of the scat- opening ( 1.30 x 1.50 m). Under the basin, a Sandy. 1ered remains in Tumuli 28 and 29 are more difficult oft sediment was present. sîmilar LO lhe sediment to interpret as secondary inhumations because the covering t:he upper parl of tbe cii.t. poor slate of the assemblages could also be 1he result of taphonornic processcs. Summary ·a11d Discussin11 Although dst burials are also known from Ponery Neolilhic contexts .in Jordan (e.g., Ban ning 1995: The excavation of 1he westernmost tumuli of tbe 1998: 224). tJ1ey differ in numerous particulars. in­ Southern Ridge reveals well-preserved cist burials. cl uding social context (sedentary fanning vi llages), Two are pentagonal-shaped with a major axis ori­ the presence of grave offerings like pots. the exclu­ en1ed northeast-southwest Jike the shrines, and Lhe sive pracLicc of prima.ry burial. and the absence of Lhird one is rectangular with an axis perpendicular tumulus superstructures. These contrasts undoubt­ to 1he previous ones. ln each tumulus. human re­ edly reflect the fundamenlal contrasts berween the mains were discovered, clearly associated with Lhe li feways of Lbe desert and the sown. initial construction of the cist. with slab covers and/ Finall y. ir is imporcanl to emphasize that the burial or pavements. At least six adult individuals were bur­ remains from Lbe Ramat Saharonim tumuli suggest a ied in tbese three rumuli during the Laie Neoli1hk. closer behavioral relalionship with Lhe Sinai nawamis Severa! Lhousand year la1er, Tumulus 29 was re used tombs than previously assumed. Wbereas previous by Nabataeans (mos1 likely) for the inhumation of an assessments teuded 10 view the tumuli as single-epi­ old woman. The Neolithic remains are very poorl y sodc. single-burial tombs. in significanl conLrast to preserved, and the bones crumbled on louch. At the nawamis. Lhe reused and multi-burial iumuli at leas1 three burials were primary inhu mations. With Ramat Snharonim suggesl similarities not recognized the exception of the half skeleton in Tumulus 28. in earlicr studies. whicb was well pro1ec1ed by a large slab, the original position of the bodies at burial could not be recon­ ABSOLUT E C HRONOLOGY structed because the Neolithic people reorganized the boues after the decay of soft tissue. Neverthe­ fn the virtual absence of malerial culture remains less, given the arriculated anatomical clusters al the associal'ed with the Ramat Saharonim complex, two base of the cists, il is clear that the bodies were not independent methods of dating, L4c and OSL, were covered immecliarely by sedimenl. The stale of pres­ nssayed in order to place the site in historical coo­ ervation of t:he articulations shows thal the decay oc­ texl. A dctailed discussion of the methods and resullS curred in an empty space, allowing small movements has been publisbed elsewhere (PoraL et al. 2006), but of the bones during the process of decomposiûon. a review bere will allow better evaluation of the More likely. and in accord with geomorphological dates in light of lhe detailed site description. analysis (see discussion below). the bodie!l were cov­ The melhod of radiocarbon dating is well knowa ercd by scooe slabs and the sedimenl infi ltrated the (e.g .. Mook and Waterbolk 1985; Ramsey 1995), and graves Jater. Secondary burial is also likely to have the results of the de1ermina1ions are presented in been part of funerary customs at Ramat Saharonim. table 1. The two dates from Shrine 4 derive from charcoaJ fragrnenlS associated with the contact bc­ which prt:sumably accumulated soon after the con­ t ween Units Ill and fV, that is. the original land sur­ struction or abandonment of the site. This age will face associated with the construction of the shrinc. be a minimum age for rhe sire. By combining the two They clearl y place Lhe construction ca. 5000 cal B.c .• typei.. the age of the site can be constrained. the Late Pouery Neolithic in standard Levantine The need for alternative dating merhod terns chronological tcrminology. and Lo the early phase of Erom the anticipated paucity of charcoaJ on lhe site. the Timnian in Rothenbcrg's (e.g .. Rothenberg and and indeed. the OSL date. provided lhe onl y means Glass 1992) desert framcwork. of daring the earlicr phase of tumulus construction Although both dates dcrive from wcll-controllcd and use. The details of the methods as applied to Ra­ contexls, RIT-4665 from a 'lmall in siru heanh and mat Sahuronim arc published el ewhcre (Porat et al. RTT-4663 from a concentrntion of charcoal on the 2006; also see Por:u 2002). Briefly stated, 1wo tech­ original land surface, givèn only two dmes (the rotai niques were employed, one utilizing single aliquots charcoal recovered). it i, difficult to as~css the minor Con the tumuli and the shrîne) whereby a large num­ differcnce betwecn them. Although they overlup ut ber of quurt7 graini. (several thousands) are measured the 2-sigma confidence imcrval, in gcneral lhi:: con­ LOgether, and the second 11tilizing single graill analy­ struction of four shrines nnd 30 tumuli supports the ses. Results arc summarized in table 2. likelihood that the span in the dates renecrs a long With respect to the tumuli at Ramat Saharonim. period of site use, implying that the differencc 111 one can c;afely assume that ail the scdiment within dtues is meaningJul. thcm, betwcen the slabs abovc and below the burial • The 14C date for tJ1c upper buriul in Tumulus 29 and be1wccn Lhe srones of the cist walls. accumulaled ·uggest a Hellenistic/carly Nabarnean reuse of the after burial (see Porat et al. 2006 for detailed discus­ tumulus. Beyond dating the burial itself, and pos­ sion). The bodies were not covered with soil. as no sible implications for :in early use of the sou thcrn soil is nvall:lble in the area. but with Stone slabs (sce spice route. the relativcly dose agreement between aJso discussion of <;kcletal remains above). The infil- the t.ic determination und the OSL date (sec he­ 1ra11ng grains arc fine. wind-bome. and were blown low) lends grcater confidence 10 the chronology in into the tumulus Lhrough gaps in the stone,. There­ generaJ. forc. 1hcy were mosr Ukely well bleached ut the tirne Luminescence dating rnethods (Aitkin 1998) date of dcposition, of the sedimem will reduce ihe previously acquirec.J B.1> .. while a second sarnpJe fRS- 10), taken from under OSL signal 10 zero ("blcaching''), and after burial it <• tone in the wull of the cist. from the same level as will grow again. These mcthods are used ex1e11:.ivt:ly the upper. well-preserved skcleton. gave un age of for dating Late Pleistocene 10 Holocent! aeolian. allu­ 2000 ± 200 years B.P. Evidently. the two burials are vial. fluvial. and colluvial sedimcnts, nncl major ap· of differcnt ages, supporting the idea of secondury plicatioas include paJaeoseismology. palaeoclimatc. use of Tumulus 29. As these ages are from sedimems land cape evolutioa. and prehistoric sites. ln order depoi. 11ed after burial. the burial in Tumulus 29 took to date an archaeological installation, one needi. w place ahout 7500 ± 700 years ago, and the b11rial in sample sedimems (preforably aeolian) deposited cloi.e Tlumllus 28 about 2000 ± 200 years ago. ro the lime of construction that are likely Lo have ln the shrine. both pre- and pos1-construc1ion ed­ becn exposcd to unLight. Appropriarc ·edimenl con­ imcnr., were identified. At the time of construction, texts include two basic types: ( 1) sediments underly­ the uppcnnosr pan of the surface sediment on which ing stones used for constructi on whercin presumably the sroncs wcrc placcd was disturbed and mixed. the uppermost sand grains were exposed 10 the sun exposing a layer of several cenrimeters 10 sunlight prior to the placement of stones. 1hus providing a (a bimilur process happened during the excavation. maximum luminescence age for the site, and (2) sed­ when a large number of people treaded the surface). iment füling interstices betweea construction stoncs This layer could poteotialJy give the timc of con- TABLE 2. Summary of Op1ically Stimula1cd Luminescence (OSL) Dates

S111,f?le Aliquot Mt'fH1tre111et11J Axe (ka) YmmgeH age St1111ple Deptlr (111) De (Gy} Dose rate H.P. (~a) H.I'. Slmne 4: RS· I 04 268±7.9 11 22 ± 66 23.9±77 12.5 RS-2 0 -~ -Il 2 ± 11)5 l:.?TH.!4 J2 4 ± 81 19 7 RS-3 04'i :?O.ll±7 6 !071 i 2.1 20.8 ± 7 l 9 l RS--1 () ~ 16.8 ± 5.5 9110 ± 60 17.2 ± 57 q4 RS-5 0.-1 73±21 121K ± 76 60± 18 41 RS-6 0,5 l l 9±2.5 991 ± 23 12.0± 2.5 iUl RS-7 0,2 234 ± 19 15·1Jl 76 152 ± 14 (il) "/1111111/i T28 RS-8 O.H 118±1.2 197H1R 60±06 T28 RS-9 U.H 147±1.4 1960 ± 28 7 5 ±0.7 T29 RS- 10 () 7 3 :'il: 0 35 1714 ± 2'1 2.010211 T29 RS- 11 0.7 ) 0±0.27 IM0±26 1 8 ± 1117

Single Grain Mem11re11w11t,f No. 11/ De (Gy) A}le(ka) Somple grc111u Main peak Dose "''" H.I'.

RS· l 12Œl -17± 1.4 1122 ± 66 4 z ± 1.) R ·2 3!17.l 8.2± 2.6 1173 ± 24 6.4 ±2. 1 RS·J 271.16 5.4 ± 2.4 1071.123 5.U± 2.2 RS-l 30/40 5.1 ± 1.5 9ll0i60 5.11± 1.6 RS·5 9142 5.4 ± 0.7 1211! ± 76 4.8±0 8 RS·6 23/34 6 1 ±2-6 991 ± 2.1 6.1 ±26

struction. when the placement of stones ~caJ ed the largc1.>t proportion of well· blcached grains. and Lbe sedimcnLS off from furthcr exposure to sunlight. Af­ age culculaled from it could indkate the true time ter construction and abe 10 the base uf Single grain measurcmems (B01ter-Jen en et al. the site. from the layer tha1 was d"turbed during 2000) of huntlrcdi. of individual grains from the ~i t e construction. One obvious rea<;On for Lhe e large shrine howed a mixed population of young and old scaner and old ages is thai in 1.>ome or the sediment age1.>. For each <;ample. 1hcre is a distinc1 peak distri· grai ns. the OSL signal was not fully rcset during buLion in the young ages (4000-6000 years 11.P.) with transportation. and thcy carried a c;ubi.tantial resid­ a tail of older grain1.> (~ee Porat el al. 2006). This con· ual signaJ at the timc of deposition. The aJiquot with firm~ our inference from the '>ingle aJiquot measure­ 1he lowesr dose equivalcnt (De) would contain the rncnt 1.> that some of the grains in the edimenl were not welJ exposed to sunlight al the Lim e of deposi­ and has constituted one of the prime difficullies in tion, and that the OSL signal or thcse grains was not 1hcir lnterpretation (cf. Avner 2002: 114- 15). fully z.eroed. ft must be notecl 1hat each sample also Second. no domeslic sites were found witrun at contains young grajns. with ages as young as 3000 least a 2-km radius of ù1e site, from any period. A ycars B.t>., indicaling that deposition conti nued for milling stone quarry, dated roughly to the Early many thousands of years. Bronze Age, but possibly earlier5 (Abadi 2003; Abadi Ages calculated from the younger grain population and Rosen in press). is located about 200 m north of for ail samples cluster beLween 4200 and 6400 years the site, acljacen1 10 sandstone outcrops, and the Na­ B.P. (table 2). with an average of .5400 :t 800 year:. baiaean inn at Ein SaJrnronim is located about 2 km B.P. for au six samples. Thi i. age ii, intemally consis­ south. but no contemporary (Late Neolithic) sites have tent. and it confonns bener to the archaeologicaJ data. been discovered. nor any other habitation sites from This is a post-construction age and rhus is a minimum olher periods. age of the shrine. Given the radiocarbon dates from Thus, the cornplcx at Ramat Saharonim is clearl y the sbrine and U1e OSL dates from Tumu lus 28, we a cul! sile, undoubtedly linked to mortuary behavior may assume that the shrine and 1he tumulus were con­ (see also Rosen and Rosen 2003 ). Beyond lhe obvi­ structed and used at the same time; the younger age ous burials in the tu mu li, and the elaborate behaviors of the sbrine OSL dates could be attl'ibutable lo associated with the burials. the su.mmer solstice set­ the longer use of the sbrine, and/or the li.me re­ Lin g sun a1ignmen1 can also be placed in such a mor­ quired for sediment accurnulati.on. If the sbrine were tuary context. Certainly the setting sun in the west in use for a long tune. ~edimentation •rnd hence burin! signified death in ancient Egypt (e.g .. Erman 1894: would have begun only after il was abandoned. 3 J 0), and clcar parallel.s exist between the tumuJus Ahhough no absolute dates were obiained for 1he fields of rhe Negev and the Earl y Bronze Age l nawa­ norLllem secondary structure, il is clearly lacer than mis tomb fields of Sinai. already tied to Egypr. The Lhe primary structuTe. Given tbat 1he discrepancy be­ summer solstice, marking the dry season in the Near tween the OSL date& and 14C dates runs on the orcler East, i!. the seasun of death, in marked contrast to the of 1000-2000 years. that some effon wa:. made to winter and spring. the seasons of rain and rebirth. obtain OSL dates as close to thè interface between The absence of dome. tic sites and activities in the the original surface and the sand/silt accumu lation ac; vicinity of l'he complex also suggests a distinccion possible. and that the stoncs of the secondary struc­ between the sacreEin Gedi). The presence of typicul open-air mosques and wi th the majority of ancestors justifies ownership. ccmeteries. and corpo­ other torical context, Lhi s megalithic a.<; pect re­ send a clear message of terri tory. 1n the case of Ramat fl ectl> the organization of Jabor an order of magnitude Saharonim, the tumuli may be murking a tribal bor­ greater thon 1hot of the preceding desert Pre-Pottery der. since Late Neolithic sites are prcsenl in the Arava Neolithic socie Li es7 (cf. Renfrew 1984: 165-99). That Valley, south and easr of the Makhtesh Ramon. bu1 this labor cou Id be drafted is undoubtedly a function are relatively scarce in the central Negev. lt may be of 1he power or the cull ideologies also embodied in noted that the Bedouin graveyard at Ein Saharonim, the shrines. The stability or Lh e system, reflected in onl y 2 km south. marked the l 9th- to 20th-cemury 1he building of four shrines and tumulus construc­ rribal border between the Azazmeh and Saidiyyin ti on over what must have been everal generations, tribes (for map of Lribal borders. ee Bailey 1980: rnap and in the conlinued power of the place as a symbol 11 .3). ln the same sense tbat 1he cemetery ·eem Lo over 1he long tenu, is suggestive of the long-renn im­ have marked tbe tribal border in recent times, the tu­ portance of such cuits for desert nomadic societies. mulus field may bave functioned in a similar fashion. Furchermore, the seasonal aspect also plays a 71! ls Important lO distinguish hcre betwecn the large PPNB role. The orientation of the shrines lndicate11 some sites of ~o ulh o rn Jordan. Lhe megasitcs, clctlrly integrated into a kind of aggregation at the s ununer solstice. The lurgcr ~ph ere of PPNB ~e d c ntary agricuhural socicty and Jocated absence of occupational accumulations indicating in the Mcditerraucan and steppe zones enst and southeast of Ùle Dead Sea. and the smnJJer-scale hunLer-gatherer socielies localcd in the dccpcr dcsen. AltlJougb located due wcst of the mcgasite re· "Although the PPNB sue of Nnhnl Hcmnr m lhe Judaean gion, Ramat Saharonim. lllld 1he centrJI Negev in gcncral, is more De~ert hru. been intcrpreted as a euh site, it is bc•t IJ$.O.OCiated wilh and thon the Pe1ra/Ba,1a region, environmeni3lly comparable to the Judacan Hi Us and the ShephclaJ1 (B:tr· Yc»ef nnd Alon 198!!: are115 m Jordan f:u1hcr south. and es~enually comprises a different 28) and i~ lbus Linked to the core an:a of the ngricuhural PPNB. cnvirunmental (und cullurall zone. ACKNOWLEDGMENTS

The Ramai Saharonim Excavation Projcct was supponed I lar Manegcv Field Scbool, affiliated witlJ the Society for the by a grant from the National Gcographic Society Commiuee Protection of Nature. and 4 high school student volunteers for Research and Exploration. Beyond lhe aulhors. partic­ from Mi l2pc Ramon. Map and plans were prepared for ipating field excavatjon staff included Yael Abadi (Bcn­ publication by Pll trice KRminsky and Helena Sokoloskayu, Gurion University): Tali Cohen (Bar-llan University): Ser­ of the Archaeological Division of Ben-Gurion University. ena Love (University College London): Brian Pittman Alter Fogcl, al o of the Archaeological Divi~ion of Ben­ (University of Cambridge); Amon Ram (Ben-Ourion Uni­ Gurîon University, developed and scanned the photographs. versity); Dan Rabin, artist, survcyor; Ynniv Rosen (Uni­ 1 nm grateful to Daniella Bar-Yosef for idenûfying û1e sea­ versity of Califomia, Berkeley): Jacob Vardi (Bcn-Gurion shcll s. Benjamin Saidd and Tsaac Gileaù commcmed on University): rmd Rona Winter (Bcn-Gurion University). early draf'ts of this papcr. We are grnteful to two anony­ Other participants included 14 volunteers and staff from che rnoui. reviewer!. for their comments.

REFERENCES

Abadi. Y. Avni. G. 2003 Early Bronze Age Grinding Stone Production 1994 Early Mosques in the Negev Highlands: New in the Ncgev Hi ghlandi.. M.A. thcsis. Ben­ Archaeologiéal Evidence on lslamic Penetra­ Gurion University or the Negev (Hcbrew). ti on of Southern Palestine. Bulletin t>f the Abadi. Y., and Rosen, S. A. Amnica11 Schools of Oriental Researc/1 294: ln press A Chipoff theOld Mlllstonc: GrindingSronePro­ X3-100. duction and Distributi on 111 thc Early Bronze Age Avni, Y. of tht: Negcv. ln New Approaches 10 Old Swnes: 1993 The Structural and Landscape Evolution of the Rece111 S111d;es ofGro1111d Stone Artifncts. eds. Y. Western Ramon Strucrw·e. Israel Journal of M. Rowan and J. R. Ebeling. London: Equinox. Earth Sciences 42: 177- 88. Aitken. M. J. Avni. Y. , and Porat, N. 1998 An /111roduc1io11 to Op1icc1/ Dati11g: Tire Dat· 2002 E11viro11111e11tal Impact Res11/ti11g from Ems/011 ing of Quatemary Sedimems by the Use of Pho· t!( Pleis1oce11e Sedime111s in the Negev High­ ra11-Sti111 11/ated L11111 i11esce11ce. Oxford: Oxford /antls. Jerusalem: lsrael Geological Survey University. Report GSI/ 112002 (Hebrew). Amit. R.: Harrison. J. B. J.: and Erize!, Y. Bailey. C. 1995 Use of Soils and Colluvial Deposit:. in Annlyz· 1990 111e Ottomans and Lhe Bedouin Tribes of tbe ing Tectonic Events-The Southern Arava Rift, Negev. Pp. 321-32 in 0110111a11 Palestine. fsrael. Geo111orplwl<>1:1Y 12: 91 - 107. 1800-1914: Studies in Economie and Soci<1/ Avner, U. History, ed. G. G. Gi lbar. Leiden: Brill. 1984 Ancient Cult Sites in tht: Negev and Sinaî IJanning, E. B. Desens. Tel Aviv 11: 115- 31. ! 995 Herdcrs or Horncsteaders? A Neolithic Pann in 1990 Ancient AgriculturaJ Settlcment und Religio11 Wadi Ziqlab. Jordan. Near Eastern Arcliaeol­ in the Uvda Valley in Southern Israel. Biblical ogy 58: 2-13. Arc/weologist 53: 125-41. 1998 The Neolithic Period: Triumphs of Architec­ 2002 Scudies in the Material and Spiritual Culrure of ture. Agriculture and Art. Netir E1wem Ar­ the Negcv ami Sinai Populations, during thè diaeology 61 : 188-237. 6th-3rd Millennia a.c. Ph.0. dissertation, Hc­ Bar-Yosef. O.. und Alon. D. brew University or Jerusalcm. 1988 NaJ:ial l~emar Cave: The Excavations. 'Atiqor Avner. U.. and Carmi. l. J8: l- JO 2001 Seulement Patterns in tht: Saut.hem Levant Bar-Yosef. O.: Hershkovi.tz. 1. : Arbel. G.: and Goren. A. Descrts during the 6Lh-3rd Millennia llC: A 1983 The Orientation of Nawamis EnLrances in Revision llased on 14 C Oating. Ratlioc<1rbn11 Soulhcrn Sinai: Expressions of Religious Belief 43: 1203-16. and Seasonalily? Tel A viv 10: 52-60. Bar-Yosef, O.: Belfer-Cohen, A.; Goren. A.: and Srnitb. P. Erman. A. 1977 The Nawamis near Arabah, from the Chalcolithic 2006 Dating the Ramat Snhoronim LM.' Ncoli1hic Pcriod 10 che Early Bronze Age IV. l eva111 24: Desert Cult Site. Journal of Archaeo/ngical 141- 57. Science 33: 134 J- 55. Schmidt. K. Rumsey. B. 2001 Glibek li Tepe. Southeastem Turkey. A Prelim­ 1995 Radiocarbon Calihration and An;i lysil. of inary Repon on the 1995- 1999 Excavations. Stratigraphy: The. OxCal Pro!,rram. Rmfiocar­ Pc1/éwie111 26/ 1: 45- 54. bo11 37: 425- 30. Simmons. A. H. Rcnfrew. C. 19R 1 A Paleosubsistence Modcl for Earl y Neolithic 1984 Megaliths, Terrîloncs and Popul utions. Pp. 165- Occupation of the Western Negev Desert. Bul· 99 in Approacl1es w Social Arclllu·ology, by C lerin of the American Schouls uf Ori1111tal Renfrcw. Cambridge. MA: Harvard Uuivcrsity. Rese