You are on page 1of 9

Maney Publishing

New Evidence on Prehistoric Trade Routes: The Obsidian Evidence from Gilat, Israel Author(s): Joseph Yellin, Thomas E. Levy and Yorke M. Rowan Reviewed work(s): Source: Journal of Field Archaeology, Vol. 23, No. 3 (Autumn, 1996), pp. 361-368 Published by: Maney Publishing Stable URL: http://www.jstor.org/stable/530487 . Accessed: 03/03/2013 19:01
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Maney Publishing is collaborating with JSTOR to digitize, preserve and extend access to Journal of Field Archaeology.

http://www.jstor.org

This content downloaded on Sun, 3 Mar 2013 19:01:33 PM All use subject to JSTOR Terms and Conditions

361

New Evidence on Prehistoric TradeRoutes: The Obsidian Evidence from Gilat, Israel
Joseph Yellin
The Hebrew University of Jerusalem Jerusalem, Israel

Thomas E. Levy
The University of California, San Diego La Jolla, California

Yorke M. Rowan
The University of Texas at Austin Austin, Texas

Obsidian artifacts are rarefinds in prehistoricsites in Israel. Thescarcityof the material and the absenceof obsidiansourcesin Israel makessuch artifacts especiallyimportantfor sourcesof obsidianfound in Israel understanding ancient exchangepatterns. Theclosest are in the CycladicIslands of Greeceto the westand Anatolia to the north. Using neutron activation analysis (NAA), we identify the origin of sevenobsidian artifacts from

theChalcolithic northern (ca. 4500-3500 B.C.)siteof Gilat in Israel's Negevdesert. Anatolia. HotamisDa' alsoin central centralAnatolia,and, mostinterestingly,

These Dagin finds have been traced to the Nemrut Dagarea of easternAnatolia, G6illii

Introduction at the Chalcolithic(ca. 4500-3500 Recent excavations of in Israel's northernNegev desert have site Gilat B.C.) wide range of objects made of revealedan astonishingly imported minerals (Alon and Levy 1989; Levy 1992, such as violin-shaped figur1995). Some of these artifacts, ines, palettes, votive axes, fenestrated"incense burner" acstands,and standingstones, are linked to cult-related of the tivities.These importedartifacts centrality support for the center Gilatas a late prehistoric exchange regional of exotic and cultic materials (Alon and Levy 1989). Durseasonsat Gilat,six ing the recent 1990-1992 excavation In addition,two obsidian obsidianpieceswere discovered. at Gilat bladeletswere recoveredfrom earlierexcavations seasons. the 1975-1977 during The contextin whichsix of the eight objectswerefound places them securelyin the Chalcolithicperiod; all the artifactsare summarizedin Table 1. Seven stratawere defined at Gilat, all dated securely to the Chalcolithic sites in the period.The paucityof obsidianat Chalcolithic southern Levant makes these materials especially significantin helping to identify long-range exchange patlittoral. ternsalong the easternMediterranean

Site Location and Regional Context Gilatis located in the northernNegev Desert of Israel Situatedon the some 25 km Nwof the city of Beersheva. interfacebetween the Negev coastalplain and the more aridinlandfoothill zone, the site is locatedon the edge of the best grazingland in the Negev (FIG. 1). The landscape is mostlyan undulating, gentlyslopingplateaudissectedby a nearbyseasonaldrainagecalled the Nahal Patish.The climatein this areaborderson arid;wintersare cool and summers hot. The average annual rainfallranges from 250-350 mm, all of which fallsduringthe winterseason. This region has been cultivatedfor thousandsof years, beginningin the late Neolithic period, and there are few of the natural remnants vegetationleft.
Due to long-term agriculturalexploitation of the region, the main plants in the area are Leopoldia eburnea and Lolium multiflorum (weeds in cultivated land). The primary crops grown both in prehistoric and historical times were wheat, barley, and most recently, sorghum. Most of these crops were grown under dry-farming regimes with occasional supplementary irrigation, even during the Chalcolithic. Extending over an area of some 10 hectares, Gilat is one

This content downloaded on Sun, 3 Mar 2013 19:01:33 PM All use subject to JSTOR Terms and Conditions

362 PrehistoricTradeRoutes/Yellin,Levy,and Rowan

Table 1. Summary of information on obsidian artifacts from Gilat.


Text reference HUAL no. Locus Basket Stratum Context Artifact type Dimensions (mm) L x W x Th NAA source

Gilat 1 Gilat 2 Gilat 3 Gilat 4 Gilat 5 Gilat 6 Gilat 7 Gilat 8

14508 14509 41210 41211 41212 41209 41208 not analyzed

60 55 609 1124 624 854 1168 806

222 174 5040 3572 5134 2449 3729 2381

III II IIA IIC IIA IIIA IIC IIA

Floor Pit Fill Pit Pit Pit Pit Pit

Bladelet Retouched bladelet Debitage Debitage Debitage Retouched bladelet Debitage Micro-endscraper

29 23 20 17 28 23 17 34

x x x x x x x x

8 x 3 10 x3 9.5 x 3 15 x 15 15,5 x 5 8.5 x 3 18 x 9 10.5 x 4

Hotamis Dag Hotamis Dag Nemrut Dag Nemrut Dag Gollu Dag Gollu Dag Nemrut Dag Unknown

of the largest Chalcolithic sites in Israel. The site is made up of two main components: a low mound extending over an area of ca. 30,000 sq m and rising about 2 m above the surrounding plain; and the flat, loessial plain that includes an area of approximately 70,000 sq m, where Chalcolithic pottery sherds, flint tools, debitage, and other artifacts have been found. All of the excavations at the site have concentrated on the low mound located roughly in the center of the site. Gilat is a well-stratified site with seven major strataidentified to date (Strata I, IIA, IIB, IIC, IIIA, IIIB, IV). The total area excavated at the site is 1,200 sq m, a relatively small sample compared to some other Chalcolithic excavations in Israel. For example, at Shiqmim a total of ca. 5,000 sq m has been opened and at Bir esSafadi over 10,000 sq m (Levy et al. 1991; CommengePellerin 1990). The earliest occupation at Gilat (Stratum IV) is characterized mostly by pits and some small plat-

forms. Beginning in Stratum IIC, the presence of the main cult area became apparent with the discovery of a small group of structures filled with exotica (Alon and Levy 1989). For example, in Room A more than 70 objects related to cult activities were found. These include anthropomorphic and zoomorphic statuettes, fenestrated stands, violin-shaped figurines, pendants, and unusual pottery types. When this structure was destroyed, the cult complex was moved to the west where large numbers of cult-related finds came to light as well as four non-domestic circular platforms and a multiple grave in a well-built burial monument. Stratum IIA represents a more ephemeral occupation which may reflect a period of social disintegration following the destruction or abandonment of the previous Stratum IIB. Unfortunately, the last occupation phase (Stratum I) was thoroughly disturbed by deep ploughing during the early 1950s. The centrality of the site in the northern

Figure 1. Map showing the location of Gilat and obsidian sources relevant to the Gilat obsidian artifacts.

NEMRUT HOTAMIS DA .VAN GOLLU DAG

DA&

-/

O
GILAT

MAP SOURCELOCATION
OBSIDIAN SOURCE 0 500 km

This content downloaded on Sun, 3 Mar 2013 19:01:33 PM All use subject to JSTOR Terms and Conditions

23, 1996 363 Journal of Field Archaeology/Vol.

Figure 2. Obsidian microendscraper (Gilat 8).

Negev and the archaeological context suggest that Gilat functioned as a prehistoric sanctuary.The obsidian artifacts reported on here add another dimension to the study of the exchange of rare raw materials in the Southern Levant during the Chalcolithic period.

The Obsidian Pieces


Both the context and the technological elements of the obsidian artifacts suggest a Chalcolithic manufacture and use. Dimensions and contextual numbers are presented in Table 1, with other summarized data. The best example is a micro-endscraper (FIG. 2) that was not subjected to NAA because of the destructive nature of the analysis. Though data collected on micro-endscrapers from Gilat are under analysis, the obsidian bladelet clearly follows the same manufacturing techniques as the standard Chalcolithic micro-endscraperspreviously reported (Gilead 1984). This bladelet has an asymmetric cross-section that resulted from the complete dorsal retouch on one lateral margin (right) to the proximal area near the bulb of percussion. The distal, dorsal aspect was retouched creating a convex distal tip. This piece (Gilat 8; see TABLE 1) was found in a Stratum in this artifacts the all of pit were Chalcolithic. The IIA pit; a violin-shaped figurine with was found micro-endscraper fragment, at roughly 0.30 m above a complete semi-flexed burial of an adult. Gilat 6, a proximal bladelet fragment with unilateral dorsal fine retouch, was recovered from an older Stratum IIIA pit in a clear Chalcolithic context. This piece exhibits a triangular cross-section and probably represents the proximal fragment of a micro-endscraper. Gilat 1, recovered from the 1975 excavation, is a medial fragment, recovered from a Chalcolithic floor in Stratum III. The fragment is un-retouched, although there may be use-wear damage on one lateral edge. No bulb of percussion or striking platform remains. Gilat 2, also a tool fragment from Stratum II, was

recovered in 1975 from a shallow stone-lined pit. This bladelet retains the bulb of percussion and has fine dorsal, unilateralretouch. The distal end is snapped, and, based on the typical micro-endscraper retouch on the lateral margin and the standardized size, we speculate that this is the proximal fragment of another micro-endscraper. The remaining four obsidian pieces appear to be debitage fragments. Gilat 7 is probably a bladelet core trimming element, perhaps a striking-platform fragment. This piece was found in a Stratum IIC pit that was sealed by a later Stratum IIB surface. Gilat 4 dates to Stratum IIC and was also found in a pit sealed by the same Stratum IIB surface. This piece was a triangular flake fragment with two parallel dorsal flake ridges, indicating a previous bladelet probably was removed before this flakewas struck. This flake fragment was found during sieving of the pit's contents. Because the top of the pit was sealed by a Stratum IIB surface, the artifact cannot be considered intrusive. Another obsidian bladelet core trimming element (Gilat 5) was recovered from Stratum IIA. This piece was a striking platform with numerous bladelets removed from along a right-angle corner on the dorsal side of the piece. The piece was found in fill about 0.60 m below the ground surface. Because no other intrusive material was found at this depth in the fill, it seems unlikely that this piece could be intrusive, but the possibility cannot be ruled out. The final obsidian piece was a flake fragment (Gilat 3), also recovered from fill in Stratum IIA about 0.50-0.55 m below ground surface. The dorsal aspect of the fragment had two non-parallel flake ridges. Because this piece was from the top of the fill, below the disturbed topsoil horizon, it cannot be positively placed in an undisturbed Chalcolithic context.

Chemical Characterization of the Gilat Obsidian by NAA


The method employed to measure the elemental abundances of the obsidian artifacts by Neutron Activation Analysis has been previously reported (e.g., Yellin and Garfinkel 1986; Yellin and Frachtenberg 1992; Yellin and Gopher 1992). Table 1 gives the correspondence between the textual numbers, the objects analyzed, the Hebrew University Archaeometry Laboratory (HUAL) sample identification, and the provenience of each obsidian analyzed. Table 2 shows the analyticalresults for each sample. Only a few of the 24 elements measured for each obsidian are given, enough to differentiate the compositions (see Yellin and Perlman 1981). For the complete set of elements measured for the source obsidians see Yellin (1995). The elements measured are arsenic, barium, cal-

This content downloaded on Sun, 3 Mar 2013 19:01:33 PM All use subject to JSTOR Terms and Conditions

364 PrehistoricTrade Routes/Yellin,Levy,and Rowan

Table 2. Analytical results for the Gilat samples with values normalized to the Cl values, which are in parts per million. Unavailable values are indicated by "na."
Element Atomic weight Gilat 7 Gilat 3 Gilat 4 Gilat 5 Gilat 6 Gilat 1 Gilat 2 Cl

La Ce Nd Sm Eu Tb Yb Lu

139 140 144 150 152 159 173 175

396.3 321.8 192.7 128.1 10.5 135.0 81.7 77.3

389.3 382.9 210.7 na 13.2 136.9 na na

396.3 320.8 197.8 130.7 9.8 135.6 83.0 79.7

57.8 91.4 38.4 27.3 7.0 20.1 19.5 17.8

33.3 93.8 38.4 28.0 7.4 19.4 17.7 20.2

116.5 91.4 40.2 23 6.02 na 19.1 20.2

116.8 92.2 38.8 23.4 7.08 na 19.3 18.2

0.23 0.59 0.42 0.13 0.05 0.032 0.15 0.023

cium, cerium, cobalt, chromium, cesium, europium, iron, hafnium, potassium, lanthanum, lutetium, sodium, neodymium, nickel, rubidium, antimony, scandium, samarium, tantalum, terbium, thorium, uranium, and yitterbium. In the case of Gilat 3 there are a few missing values because of an instrumental problem in data recording. The elements shown, plus the many other elements measured but not shown, confirm that Gilat 3 is of the same composition as Gilat 7. Table 2 is organized so that obsidians of similar composition are adjacent. Relevant obsidian source compositions are given in Table 3.

Gilat 1 and 2
The composition of Gilat 1 and 2 was measured by NAA in the 1970s along with hundreds of source and archaeological obsidian samples. Measurements were carried out by I. Perlman and J. Yellin in the then newly established Archaeometry Laboratory in Jerusalem. It was noted then that the Gilat obsidians were enigmatic in terms of their composition (Yellin and Perlman 1979). They are the only archaeological finds linked to the Hotamis Dag source in central Anatolia. Perlman and Yellin (1980) noted that all Pre-Pottery Neolithic obsidian finds from prehistoric sites in present-day Israel could be traced to the Gollii Dag source, also in central Anatolia, but that Chalcolithic period obsidians had many origins, including Gllii Dag and sources in eastern Turkey, as well as yet undefined sources (Yellin and Perlman 1979).

No obsidian artifact, however, other than those from Gilat, was ever traced to Hotamis Dag The question as to why the overwhelming majorityof obsidian artifactsfound in Palestine originate in the Gllii Dagsource when there are several other sources in the vicinity of Gllii Dagis an intriguing one. Perhaps an examination of obsidian artifacts nearer to these sources will throw some light on this question. It was also noted that obsidian found in Palestine and attributed to the Neolithic periods came from an Anatolian source called Agigol (Renfrew, Dixon, and Cann 1966, 1968), named after the nearby village of Agigol. Agigol corresponds, we believe, to the G611tiDagvolcanic source. Figure 1 shows the location of Gilat in relation to three Anatolian sources. The Hotamis Dagsource is just several kilometers north of G6lli Dag but its composition is easily differentiated from that of Gllu Dagas well as other Anatolian sources. Figure 3 shows the differences in the above-mentioned compositions. Measures of error are not shown in this figure but may be seen in Table 3, where the root-meansquare-deviation is given for each of the sources shown (i.e., the spread in composition observed in our measurements). Figure 4 compares the composition of Gilat 1 with the sources Gllii Dag(GLD) and Hotamis Dag(HTMD). Gilat 2 is of the same composition. The rare earth pattern shown in Figure 4 and subsequent figures may not be the best way to differentiate among sources of obsidian, but it

Table 3. Selected source compositions for the Gilat samples. Values are normalized to Cl chrondrite composition.
Element Atomic weight NMRD1 N = 34 NMRDI RMSD GLD N = 83 GLD RMSD HTMD N = 61 HTMD RMSD

La Ce Nd Sm Eu Tb Yb Lu

139 140 144 150 152 159 173 175

397 348.6 200.7 136.5 10.1 99.1 87.6 82

9.1 8.4 4.8 2.1 0.2 1.9 1.7 1.7

98.7 77.6 34 21.3 3.04 13.4 15.7 14.9

3 1.9 1.9 0.5 0.36 1.2 0.6 0.6

129.6 95.8 42.8 24.1 8.12 13.6 19.3 18.6

3.1 1.9 1.9 0.5 0.36 0.8 0.5 0.8

This content downloaded on Sun, 3 Mar 2013 19:01:33 PM All use subject to JSTOR Terms and Conditions

23, 1996 365 Journal of Field Archaeology/Vol.

400

C.

200

10

La

Ce

Nd

Sm 1

Eu

Tb

Yb

Lu

Elements M GLD
NMRD1 l HTMD

Figure 3. Differences in composition between central Anatolian (Hotamis DaW=HTMDand G6llfi DaW=GLD)and eastern Turkey (Nemrut DaW=NMRD1) obsidian sources.

has special meaning to geochemists studying fractionation processes within the earth and for this reason is used.

Gilat 3-7
Gilat 3-7 fall into two compositional groups. Gilat 3, 4, and 7 have a composition that matches most closely a source we call Nemrut Dag 1 (NMRD1), after the mountain on the western side of Lake Van. There are many obsidian sources in the Lake Van region of modern Turkey; these have been summarized by Altinli (1964). Our NMRD1 is a particular source whose location is not known, except that it is in the Lake Van region. A second source composition (Pullar, Yellin, and Perlman 1986) likewise cannot be identified more precisely than "Nemrut Dag/Lake Van." Tables 2 and 3 and Figure 5 show why
Figure 4. Composition of Gilat 1 obsidian artifact and the G6llfi Dag' (GLD) and Hotamis Da'g(HTMD) sources. Gilat 2 is of the same composition as Gilat 1.
1000

Gilat 4 and 7 are assigned to NMRD1 and Figure 6 shows why Gilat 5 and 6 are assigned to G6llii Dag Gilat 3 is not shown in Figure 5 because of missing values (see TABLE 2). Comparing Figures 5 and 6 shows the great dissimilarity between the source obsidians of central Anatolia and the Nemrut DagLake Van areas. Note that there are three curves in Figure 5-Gilat 7, Gilat 4, and NMRD1. However, Gilat 7 is so close in composition to Gilat 4 that it is hardly visible (see TABLE2). Based on analytical work reported thus far there is little chance that an archaeological obsidian will be assigned to the wrong geographical region. We cannot, however, rule out the possibility that the Gilat obsidian will match better some as yet unidentified source in a different geological region, but this is extremely unlikely.

O 100
o
1

130

140

150

160 0 Gilat 1 0 HTMD

170

180

Atomic Weight
E
GLD

This content downloaded on Sun, 3 Mar 2013 19:01:33 PM All use subject to JSTOR Terms and Conditions

366 PrehistoricTradeRoutes/TYellin, Levy,and Rowan

1000

100

0
c

100

10
0 0

130 139 140 144 150 152 159 173 175

140

150

160

170

180

La, Ce, Nd, Sm, Eu, Tb, Yb, Lu E Gilat 7 0 Gilat 4 O NMRD1
Figure 5. Composition of Gilat 4 and 7, and the NMRD1 (Nemrut Dag obsidian source.

La, Ce, Nd, Sm, Eu, Tb, Yb, Lu


E Gilat 5
E Gilat6 O GLD

Figure 6. Composition of Gilat 5 and 6, and the G6llii Dag (GLD) obsidian source.

Discussion Literatureon obsidian from the southern Levantsuggests that sourcesin Chalcolithicsites vary extensivelyin comparisonto the Neolithic obsidianthat has been tested by NAA. One of the most intriguingaspectsconcerning obsidian exploitation during the Chalcolithicperiod in record. ComPalestineis its rarityin the archaeological Neolithic with the Period, and earlierPre-Pottery pared sites such as Nahal Lavan109, where some 356 obsidian were found in a remote desertsite with no archiartifacts tecture (Burianand Freedman1988), there is a realpau(Yellinand Frachtcity of obsidianduringthe Chalcolithic enberg 1992: 139). One of the reasonsfor this discrepancy may relateto technologicalchange and the gradualeffect economiesin Palestine. on Chalcolithic of earlymetallurgy With the addition of metal tools to the Chalcolithictool kit, the attractionof high qualityraw materialsfor lithic tool productionmay have waned. This is consistentwith the small sample of obsidian found at Gilat, where it appearsrelativelyearly and then goes into decline. One enigmatic aspect of this problem is that while copper artifactsare common from sites in the BeershevaValley some 16 km south of Gilat, no metal remainshave been found at Gilatitself. from Gilat, Giventhe smallnumberof obsidianartifacts however,it is difficultto generalizeaboutvery broad-scale patternsof procurementand exchange mechanisms.Future excavationsat more southern Levantinesites and better sieving methods should provide largersamplesto these patterns. clarify As noted above,the three sourcesfor the Gilatobsidian are Gllu Dag and Hotamis Dag in centralAnatoliaand Nemrut Dag 1 near Lake Van. During the Pre-Pottery Neolithic (PPNA), most sites such as Jericho,Beisamoun, Netiv Hagedud,Yiftahel, Munhatta,Beidha,andEl Khiam

received most of their obsidian from Golli Dagand central Anatolia (Bar-Yosef 1991; Dixon, Cann, and Renfrew 1979: 113). The same basic exploitation pattern continued during the following Pre-Pottery Neolithic B period (PPNB). The small sample of five pieces of PPNB obsidian from Beidha analyzed by M.-C. Cauvin reveals sources in the Cappadocia region, where Renfrew and his colleagues first noted the main Pre-Pottery Neolithic (PPN) source area, and from eastern Anatolia (B. Byrd, personal communication). Recent studies by Cauvin (1991) and Cauvin et al. (1991) show that Bingol was an additional important obsidian source for PPNB and Chalcolithic villages in the Taurus mountains and in upper Mesopotamia and also to the south at El Kowm, an oasis in the Syrian desert. Cauvin's studies are important because they suggest the route of obsidian trade, which began in the Pottery Neolithic and continued in the Chalcolithic for the eastern Anatolian obsidians. El Kowm is located on an old Roman road that no doubt followed a prehistoric trade route crossing the northern Syrian desert. Dilleman's (1962) study of the historical geography of northern Mesopotamia allows us to hypothesize a route for the movement of obsidian from northern Mesopotamia into the southern Levant. Movement would have been across the Dhyarbakir Plateau, through the Tur Abdin (Maszi Dao Mountains, and into the upper Habur Valley (probably via the traditional Roman road along the Jagh Jagh branch of the upper Habur). From there, the route would have gone through the El Kowm oasis, westward into the Orontes Valley, and into the Levantine corridor. During the Chalcolithic, we see a shift in the exploitation of obsidian not only within the Cappadocia region of central Anatolia, but also from the volcanic source at Nemrut Dagi on the western shore of Lake Van some 600 km away in eastern Anatolia. Bearing in mind the small

This content downloaded on Sun, 3 Mar 2013 19:01:33 PM All use subject to JSTOR Terms and Conditions

Journal of Field Archaeology/Vol. 23, 1996 367

sample size of NAA-sampled obsidian from Gilat, the stratigraphy of the site provides a tentative view of the development of exchange relations that ultimately reached Anatolia. In the early Strata III and IIIA, central Anatolian sources were exploited following the earlier PPNA and PPNB pattern. The later Stratum IIC at Gilat, the occupation horizon associated with the construction of the first sanctuary complex at the site, shows the first evidence that exchange relations extended to eastern Anatolia and the region around Lake Van. These indirect relationships with Anatolia continue in the following Strata II and IIa. The small quantity of obsidian found in Chalcolithic contexts suggests a "downthe-line" model of exchange as outlined by Renfrew (1975). For the northern Negev and the very southern Levant, these relationships are quite interesting and probably point to the concern, or at least curiosity, of the inhabitants of Gilat in exploring the fullest range of exchange relationships in their Levantine world. This is especially interesting when cast in the light of the following Early Bronze I period when exchange with these "remote" northern regions disappears and the focus of exchange shifts to Egypt and the Nile Valley.

sity of Texasat Austin. He has conductedfield work in Israel, Jordan, Italy, and Texas, and is currently completeing doctoral researchon late prehistoriccraft specialization and trade in the southern Levant. Mailing address: Department of Anthropology, University of Texasat Austin, Austin, iX 78712-1086.
Alon, D., and Levy,T. E.

1989 "TheArchaeology of Cultandthe Chalcolithic Sanctuary


at Gilat,"Journal of MediterraneanArchaeology 2: 163-

221.

Altinli, I.

Area. M.T.A (Maden 1964 GeologicalMap of Turkey-Van Ankara. Teknik Arama)

Bar-Yosef, O. RecentAdvances," 1991 "TheEarlyNeolithicof the Levant:


12(2): 1-18. TheReviewof Archaeology

E. F., andFriedman, Burian,


1988 "A Note on the Obsidian Finds from the Pre-Pottery Neolithic B Site 109 near Nahal Lavan,Negev," Journal 21: 95-98. Society of theIsraelPrehistoric

M. Cauvin, et au Levantprehistorique: 1991 "L'obsidienne provenance


fonction," Cahiersde l'Euphrate5-6: 163-190.

Acknowledgments
We would like to thank the Soreq Nuclear Research Center for the neutron radiations, particularlyA. Nagler and S. Diga. Thanks are also owed to R. Asia for help with electronics, and R. Borosh for technical support. The authors thank G. Algaze, D. Alon, B. Byrd, and C. Commenge for many useful discussions concerning this paper and D. Ladiray for the drawing of the obsidian blade. We are grateful for generous support from the C. Paul Johnson Family Charitable Foundation of Chicago. JosephYellin is a professorof archaeometryand physics and director of Hebrew University'sArchaeometryLaboratory. He has published extensivelyon the provenience of ancient ceramics and obsidian. Mailing address:Institute of Archaeology, The Hebrew University of Jerusalem, 91905 Jerusalem, Israel. ThomasE. Levy is a professorin the Department of Anand Judaic Studies Program at the University thropology of California, San Diego. He has directed archaeological projectsat Gilat, Shiqmim, and Nahal Tillah in Israel and conducted ethnoarchaeologicalresearchnear Lake Chad in northern Cameroon. Mailing address:Department of Anthropology, University of California at San Diego, 9500 Gilman Drive, La Jolla, CA 92903-0101. email: tlevy@weber.ucsd.edu YorkeM. Rowan is a doctoral candidate at the Univer-

andR. Montigny M. C., Y. Besnus, J. Tripier, Cauvin, du Proche-Orient: 1991 "Nouvellesanalysesd'obsidiennes


Modele de geochimie des magmasutilise pour la recherche archeologique,"Paleorient17/2: 5-20. au IVe Millinaire La PoterieD'Bir es-Safadi (Beersheva) Paris:AssociationPaleorient. Avant L'ereChretienne. Orientaleet PaysAdjacents.Paris:LiHaute Misopotamie brairieOrientalistePaul Geuthner. "Obsidianand the Origins of Trade,"in C. C. LambergAmerican:Hunted., Readingsfrom Scientific Karlovsky, ers, Farmersand Civilizations-Old WorldArchaeology.

C. Commenge-Pellerin,
1990

L. Dillemann,
1962

Dixon,J. E., J. R. Cann,and C. Renfrew


1979

108-116. andCompany, SanFrancisco: W.H. Freeman

Gilead, I. a New Tool Type of the Chal1984 "The Microendscraper: colithic Period," TelAviv 11: 3-10.

T. E. Levy, NorthernNegev Desert,Israel," 1992 "The GilatSanctuary, 8: 372-374. & Exploration NationalGeographic Research Peand RankSocieties-Chalcolithic 1995 "Cult,Metallurgy riod (ca. 4500-3500 B.C.E.),"in T. E. Levy,ed., The Facts in theHolyLand.New York: Archaeology of Society on File,Inc.,226-244. Y. Rowan, A. Holl, and ThomasE., Alon D., C. Grigson, Levy, P. Smith in the NorthandAdaptation Settlement 1991 "Subterranean ern Negev Desert,ca. 4500-3700 B.C." NationalGeo& Exploration(National Geographic Sographic Research ciety)7: 394-413.

This content downloaded on Sun, 3 Mar 2013 19:01:33 PM All use subject to JSTOR Terms and Conditions

368

Prehistoric Trade Routes/Yellin, Levy, and Rowan

Perlman,I., and J. Yellin 1980 "The Provenience of Obsidian from Neolithic Sites in Israel,"IsraelExploration Journal 30: 83-88. Pullar,J., J. Yellin, and I. Perlman 1986 "Sourcesof Obsidianfrom Tepe Abdul Hosein as Determined by Neutron ActivationAnalysis,"in J. S. Olin and J. Blackman,eds., Proceedings of the 24th International Washington:SmithsonianInstiArchaeometry Symposium, tution, 389-402. Renfrew,C. 1975 "Tradeas Action at a Distance,"in J. A. Sabloffand C. C. eds., Ancient Civilizationand Trade. Lamberg-Karlovsky, Albuquerque:Universityof New Mexico Press, 3-59. Renfrew,C., J. E. Dixon, and J. R. Cann 1966 "Obsidianand EarlyCulturalContactsin the Near East," 32: 30-72. Proceedings of thePrehistoric Society 1968 "FurtherAnalysis of Near Eastern Obsidians," Proceed34: 30-72. ings of thePrehistoric Society

Yellin, J. of CentralAnatolianOb1995 "TraceElement Characteristics sidian Flows and their Relevance to Prehistory,"Israel 35: 175-190. Journal of Chemistry Yellin, J., and F. Frachtenberg 1992 "A Re-Examination of Nahal Lavan Obsidian," in M. Waelkens, N. Herz, and L. Moens, eds., Ancient Stones:Quarrying, Trade and Provenance.Acta ArchaeologicaLovaniensia.Leuven, Belgium:Leuven University Press. Yellin, Joseph, and Y. Garfinkel 1986 "The Source of Archaeological Obsidian from a PrePottery Neolithic B Site at Yiftahel,Israel,"Paleorient12: 99-104. Yellin, J., and A. Gopher 1992 "The Origin of the ObsidianArtefactsfrom Mujahiya-A PPNB Site in the Golan Heights," TelAviv 19: 94-99. Yellin, J., and I. Perlman 1979 "Characterization of Anatolian Obsidian Flows and the Provenienceof ObsidianArtifactFinds from Israel,"HebrewUniversity Archaeometry Laboratory ReportHUAL 6 Jerusalem. 1981 "SourceObsidianRareEarthPatternsand ObsidianArtifact Provenience,"RevueD'Archeometrie, III SupVolume plement, Actes Du XX Symposium International D'Archeometrie,Paris26-29 Mars 1980, 325-330.

This content downloaded on Sun, 3 Mar 2013 19:01:33 PM All use subject to JSTOR Terms and Conditions

You might also like