Reports

Recent Paleolithic Surveys in Luristan1 k . ro u s t a e i , h . v a h d a t i n a s a b , f . b i g l a r i , s . h e y d a r i , g . a . c l a r k , and j . m . l i n d l y Center for Archaeological Research, No. 15, Ekbatan St., Baharestan Sq., Tehran, Iran (Roustaei)/ Department of Anthropology, Arizona State University, Tempe, AZ 85287-2042, U.S.A.([email protected], [email protected]) (Vahdati, Clark)/Center for Paleolithic Research, National Museum of Iran, Si-ye Tir St., Imam Khomeini Avenue, Tehran, Iran (Biglari, Heydari)/SWCA, Inc., Environmental Consultants, 2120 N. Central Avenue, Suite 130, Phoenix, AZ 85004, U.S.A. (Lindly). 15 iv 04 [Supplementary material appears in the electronic edition of this issue on the journal’s web page (http://www.journals.uchicago. edu/CA/home.html).]

Judging from scarce archaeological evidence, the intermontane valleys in the Zagros Mountain chain have been occupied by hunter-gatherers at least since the end of the Middle Pleistocene. Although research in the area dates from Carleton Coon’s work in the 1940s and 1950s (e.g., 1951, 1957), geopolitical instability (e.g., the revolutions in Iraq [1958] and Iran [1979], the Iraq-Iran War [1980–88]) has effectively precluded sustained research there since the early 1970s. With the exception of the pioneering efforts of Frank Hole and Kent Flannery in Luristan in the early 1960s, what was done was not conducted under problem-oriented “modern-era” research designs, and as a result even the basic culture history of the region is very little known. Although isolated Acheulean bifaces and other typologically Lower Paleolithic tools have been recovered from a few surface collections (e.g., Pal Barik in Luristan), nothing more than their presence is known. The Middle Paleolithic is somewhat better reported, 䉷 2004 by The Wenner-Gren Foundation for Anthropological Research. All rights reserved 0011-3204/2004/4505-0007$10.00 1. We are grateful to Maziyar Asrafian, director of the Paleontology and Paleoanthropology Group, Iranian Cultural Heritage Organization, and Davood Davoodi for their help and support. Jalil Golshan, deputy research director, and Masoud Azarnoush, director of the Archaeological Research Center, Iranian Cultural Heritage Organization, also provided support. This contribution first appeared in Farsi in the Iranian Journal of Archaeology and History 16(1): 45–64, and a modified and expanded version is published here with the permission of the ICHO.

692

in part because of reanalysis and publication during the 1980s and 1990s of collections stored in the United States and the United Kingdom for more than two decades (e.g., Baumler and Speth 1993, Dibble and Holdaway 1993). More important, surveys by Iranian archaeologists over the past two decades have located a number of Mousterian sites in the Kermanshah region (e.g., Roustaei 1999). Mostly caves and rockshelters, they include Do-Ashkaft (Biglari and Heydari 2001), Awi Kher (Biglari and Abdi 1999:6), Cham-e Souran (Biglari and Abdi 1999), and new sites near the famous Bisitun rockshelter (Biglari 2000, 2001). The Middle–Upper Paleolithic transition is completely undocumented, however, and our knowledge of the later phases of the Upper Paleolithic and the entire Epipaleolithic rests largely upon modern studies of old collections from Warwasi rockshelter (Olszewski 1993a, b), first tested by Bruce Howe in the 1950s (Braidwood 1960). The best-studied part of the Zagros is Iraqi Kurdistan, where Robert Braidwood’s investigations into the origins of food production during the 1950s also yielded considerable information on the Middle and Upper Paleolithic (Braidwood and Howe 1960). Later, some of Braidwood’s students, including Hole and Flannery, continued his work, although the corpus of well-published Paleolithic research in the region remains meager. Hole and Flannery’s research during two seasons of fieldwork on Paleolithic sites in the Khorramabad Valley is summarized in a long article in the Proceedings of the Prehistoric Society (Hole and Flannery 1967), and Philip Smith’s excavation in Ghar-i Khar has been published only in preliminary fashion (Young and Smith 1966; Smith 1967, 1986). Another obstacle to a better understanding of the Iranian Paleolithic is the marked disinclination of Iranian archaeologists—until the 1980s—to participate in this area of research. Only the late Enayat Amirloo conducted officially sanctioned modern-quality research in this area, at the Ghale Asgar open-air site in the Demavand Valley (Amirloo 1990). However, a number of important Paleolithic sites were discovered and, in some cases, tested as a consequence of organized surveys conducted under the aegis of the Iranian Cultural Heritage Organization (ICHO) (Biglari and Abdi 1999, Biglari, Nokandeh, and Heydari 2000, Biglari 2001, Biglari and Heydari 2001). It is our intention here to summarize these diverse efforts and to provide a modern view of Paleolithic research in Iran, insofar as it is known.

history of paleolithic research in luristan Henry Field was perhaps the earliest researcher to identify Paleolithic sites in Luristan. Although he visited the region prior to World War II, in 1950 he dug a trench in Kunji Cave, on the outskirts of the modern city of Khor-

Volume 45, Number 5, December 2004 F 693

ramabad, publishing his results a year later (Field 1951). No subsequent work was accomplished until 1963, when Hole (then at Rice University) and Flannery (University of Michigan) identified 17 Paleolithic sites in the Khorramabad Valley. Five of these yielded typical Middle Paleolithic (Mousterian) artifacts, and the rest consisted of Upper Paleolithic and Epipaleolithic materials. Over a two-year period, Hole and Flannery excavated or tested Ghamari and Kunji Caves (both Middle Paleolithic [fig. 1]), the Gar Arjeneh rockshelter (Middle and Upper Paleolithic [fig. 2]), Yafteh Cave (Upper Paleolithic [fig. 3], and the Pasangar rockshelter (Upper and Epipaleolithic). Hole and Flannery (1967:151) saw the Middle Paleolithic

Fig. 1. Plan of Ghamari Cave.

of the region as part of an uninterrupted continuum with the Upper Paleolithic Baradostian and the Epipaleolithic Zarzian. Although the nature of the relationship of the Baradostian to the Aurignacian—supposedly lacking in the Zagros—has recently been revisited by Olszewski (e.g., 1999, 2001; Olszewski and Dibble 1994), continuity has been clearly demonstrated for the late Baradostian and the early Zarzian (unit 1) at Warwasi (Olszewski 1993b:215). A monograph-length treatment of this early work has yet to appear. In 1969 Charles McBurney (University of Cambridge) excavated four newly discovered rockshelters at high elevations in the Kuh-i Dasht Mountains of western Khor-

694 F c u r r e n t a n t h ro p o l o g y

Fig. 2. Plan of the Gar Arjeneh rockshelter. ramabad (Mir Malas, Barde Spid, Humian I and II). Only Humian I yielded (Middle) Paleolithic remains; again, only preliminary notes were published (McBurney 1969a, b, 1970). A decade later one of McBurney’s students, Robert Bewley, used the results obtained from Humian I as the basis for his Ph.D. dissertation (1980), eventually publishing an account of McBurney’s excavations in the Kuh-i Dasht area (Bewley 1984). Humian I is an important Paleolithic site. At 2,000 m above sea level, it is one of the highest Middle Paleolithic sites in western Eurasia (Smith 1986:21) and, arguably, the oldest dated site in the Zagros.2 The most recent sustained Paleolithic field research 2. There is a single Th/U date on one piece of bone from layer 2a at Humian I: 148,000Ⳳ35,000 years bp (Bewley 1984:35–38, n. 53). It should be kept in mind that the chronology of the Middle Paleolithic in western Asia has changed dramatically over the past 20 years. On the basis of thermoluminescence, electron spin resonance, and uranium-series chronologies from Middle Paleolithic sites in the Levant and Central Asia, the Middle Paleolithic is considered to begin ca. 250,000 years ago and end sometime around 40,000 years ago in western Asia (Mercier and Valladas 1994, Davis and Ranov 1999). This agrees with current dates for the European Mousterian and contrasts sharply with the “traditional” Lower–Middle Paleolithic boundary at either 100,000 or 130,000 years ago (Middle Upper Pleistocene boundary) (Bar-Yosef 1994). Thus, the Iranian Middle Paleolithic, like that of its neighbors in central Asia, Turkey, and the Levant, must date to the late Middle Pleistocene. The single date from Humian I is consistent with this claim.

in Luristan was undertaken by John Speth as part of his doctoral program at the University of Michigan. During the spring and summer of 1969, he excavated a substantial surface area in Kunji Cave to study the effect of activity loci on what he originally took to be a relatively pristine “living floor,” sampled earlier by Hole and Flannery. However, he quickly discovered that most of the site had been extensively reworked by porcupines, to the point that he was forced to change the focus of his research (Speth 1971). In the mid-1980s Speth, in collaboration with Mark Baumler (then at the University of Arizona), reanalyzed the disturbed 1969 collections and those produced by Hole and Flannery’s 10 # 1-m trench in what were apparently rare intact deposits (Baumler and Speth 1993). This effort was part of a number of restudies of Luristan Paleolithic collections housed in British and American museums that were eventually more completely published during the 1980s and 1990s. The early 1960s collections from Kunji were also studied by James Skinner, who created a typology of Southwest Asian Middle Paleolithic assemblages as part of his Ph.D. dissertation research (Skinner 1965). During the 1980s, Angela Minzoni-De´roche analyzed partial collections from Kunji, Gar Arjeneh, Yafteh, and Pasangar in an effort to compare the Zagros Mousterian with that of the Taurus Mountains in Turkey (Minzoni-De´roche 1993, 1999). The most recent of these systematic reanalyses of the

Volume 45, Number 5, December 2004 F 695

Fig. 3. Plan of Yafteh Cave. Zagros Mousterian is that of John Lindly (1997), who was interested in the effects of seasonal mobility on lithic assemblages at sites at high elevations. After studying substantial portions of collections from seven high-altitude sites in the Zagros Mountains, he concluded that the generally colder environments prevailing there during the Pleistocene strongly suggested that these sites could have been occupied only during relatively warm climatic intervals and even then probably only during the summer months. The evidence from the sparse faunal remains indicated short and intensive but sporadic use of these high-altitude caves and rockshelters and a pattern of seasonal movement to the highlands during

the summer still seen among pastoralists in the region today (Vahdati Nasab 1999). Tools were repeatedly resharpened because raw-material availability and size, coupled with the time constraints of a short summer season, precluded the manufacture and transport of sufficient numbers of flakes to meet demands. Neandertal foragers responded to these constraints with innovative technological solutions designed to wring as much usable cutting edge as possible out of the very limited quantities of stone that they could bring with them (Lindly 1997, n.d.).

696 F c u r r e n t a n t h ro p o l o g y

choosing the survey areas The province of Luristan is located in the central Zagros and includes a variety of climatically and ecologically distinct microenvironments determined mainly by soils, exposure, and elevation. In the east, mountains over 4,000 m are the dominant landforms and are characterized today (at an interglacial maximum) by a mosaic of generally cold environments created by a lattice of individual valleys. Although they are in a precipitation shadow compared with the western Zagros slopes, snowfall is often heavy there, and snowmelt has important effects on the adjacent western and central lowlands, which have milder and drier climates and less of a mosaic character. Our preliminary goals in undertaking our surveys were to locate Paleolithic sites and assess their potential for further investigation. After studying the geology and geomorphology of the area and taking into consideration the present climate, we decided to focus our efforts on the central and southwestern parts of the province.3 We selected the Khorramabad Valley in central Luristan and the Baba Zeid plain, located between Malavy and Pol-i Dokhtar in the southwestern part of the province. These regions appeared to exhibit the greatest bioenvironmental potential for foragers, as they contained numerous caves and rockshelters, easy access to water, and river valleys that formed natural migration routes. The Khorramabad Valley was also the region in which Hole and Flannery identified 17 Paleolithic sites during their projects in the early 1960s, but, apart from the three sites they tested (Kunji, Gar Arjeneh, Ghamari), there is no published information about the remainder, which cannot be located (or relocated) with certainty. We started with a survey of the southern slopes of the Kuh-i Sefid cuesta, where the Upper Paleolithic cave of Yafteh is located, some 13 km to the west of Khorramabad (Km 17 on the Khorramabad to Kuh-i Dasht road). Ten sites spanning the Middle Paleolithic to the Epipaleolithic were located, although only a small part of the valley was surveyed. The potential for Paleolithic research in the region is clearly enormous. The Baba Zeid plain is geomorphologically, topographically, and microclimatically very different from the Khorramabad Valley. These two regions were selected for their contrasting topographies, microenvironments, and ecologies with an eye to assessing the possibility of seasonal transhumance between them. The Baba Zeid plain is located at the junction of three convergent valleys that offer natural routes to the north, leading to the Zagros highlands, to the west, in the direction of Islamabad and the Hulailan sites, and to the south, toward the Deh Luran plain and the lowlands of Khuzistan (Hole, Flannery, and Neely 1969). Finally, on the basis of local information, we also briefly investigated the slopes of Mapel Mountain, about 20 km from Khorramabad in the eastern Kuh-i Dasht Valley. Humian rockshelter, at 2,000 m arguably the 3. In fact, the high elevations in northeastern Luristan (e.g., Garin Mountain at 3,700 m, Oshtorankooh and Ghadikooh at 1 4,000 m) precluded any attempt at survey except during the height of summer.

highest Middle Paleolithic site in western Eurasia, is located in the Kuh-i Dasht (Bewley 1984). These three survey regions are distinct from one another environmentally and ecologically, primarily as a consequence of differences in elevation. Sites in the Khorramabad Valley occur at elevations ranging from 1,200 to 1,400 m; those on the Baba Zeid plain are much lower (750–900 m), and those on Mapel Mountain are all above 1,800 m asl (fig. 4). In each survey area, all recognizable lithics were collected, including shatter and de´bitage. At a few sites such as Gilvaran I and Sorkh-e Lizeh, where extremely dense lithic scatters occurred, we sampled the scatters systematically with the aim of obtaining a representative collection of the full range of stone artifacts. zagros and luristan paleoclimates Although Pleistocene paleoclimatic change in the Zagros has been studied for decades (e.g., de Morgan 1924, 1926, and several publications in Farsi dating back to the nineteenth century), there is no consensus on the paleoclimatic succession in the region (Brooks 1982). A good summary of what was known through the mid-1980s is provided by Philip Smith (1986:10, 11), and our account is drawn from it. Many of the data pertain to peripheral regions, and there are differences of opinion as to the significance of the single palynological sequence in the area, that of Lake Zeribar near Mirabad. There are also disagreements about whether and to what extent the evidence for mountain glaciation can be squared with continental glacial sequences first defined in Europe, although these controversies can be resolved to some extent by increasing use of marine paleotemperature core chronologies in the still-unusual situation in which sites can be dated radiometrically. Small glaciers evidently formed during cold intervals in the Zagros Mountains when the snow line descended to between 300 and 1,150 m below its modern average, but there is no evidence for ice sheets of any size or description (Wright 1980, Schweizer 1975). There were also montane glaciers in the Elburz Mountains north of Tehran, but these, too, are poorly understood. The only relatively credible evidence from the study area consists of cores sunk into the sediments of Lakes Zeribar and Nilofar (1,400 m above sea level) near Mirabad during the 1960s.4 Published by Herbert Wright (1960, 1961, 1980), the intent of this research was to reconstruct the local flora in the vicinity of Lake Zeribar from 42,000 years ago until the present. Wright also drew upon what little was known of the paleoclimatic sequence in Shanidar Cave, several hundred kilometers to the northwest of Lake Zeribar but at a significantly lower elevation (765 m).5 Wright estimated the local expression of the last glacial maximum in the Zagros at around 4. To the best of our knowledge, the sequence from Lake Nilofar has never been published. 5. The data from Shanidar indicated that the cave was first occupied during a warm interval prior to the last glacial maximum and the formation of mountain glacie5rs and ended before the pleniglacial maximum (ca. 18,000 years ago) (Trinkaus 1983).

Volume 45, Number 5, December 2004 F 697

Fig. 4. Geological structures, topography, and the distribution of Paleolithic sites in Luristan Province (based on Broude 1990; Darvish Zadeh 1981; Heydari 1999a, b; Oberlauder 1968; Sto¨cklin 1974, 1981). 20,000 years ago (early oxygen isotope stage 2) and also indicated two earlier glacial episodes, one at approximately 75,000–70,000 years ago (stage 4) and the other at around 50,000 years ago (stage 3 [59,000–24,000 years ago], regarded as a moderately warm interval [Klein 1999: 59]), separated by an interglacial at ca. 60,000 years ago (Wright 1961, Skinner 1965). This temperate episode probably falls in stage 3 and has been confirmed by studies of the Humian rockshelter deposits and floral remains (Bewley 1984). Since the appearance of oak (Quercus spp.) in this region is usually taken as a sign of a warm, postglacial climate, its occurrence in Mousterian levels suggests a relatively warm interstadial or interglacial episode. Either or both stage 3 and stage 5 (127,000–71,000 years ago) are the most likely candidates, but—given the extended Mousterian chronologies in both Europe and the Levant—stage 7 (242,000–186,000 years ago) is at least a possibility. It should be kept in mind that detailed correlation between continental and marine oxygen isotope chronologies is notoriously difficult. Old radiocarbon dates on solid samples of wood charcoal from the Lake Zeribar

cores bracket 42,000 and 10,000 years bp. Insofar as these extremely sparse data can be generalized, it would seem that steppe landscapes with occasional stands of trees in sheltered localities prevailed between roughly 125,000 and 42,000 years ago at relatively high elevations (ca. 1,400 m) over most of the Zagros chain. Then, between 42,000 and 24,000 years ago, the arboreal pollen disappears altogether, probably as a consequence of the cold, dry climates characteristic of the last glacial cycle (van Zeist and Bottema 1977) (figs. 5 and 6). To summarize, overall glacial climates in western, highland Iran were cooler than at present, although there is no consensus as to how much cooler. Estimates of temperature depression in the Zagros have varied from as little as ca. 3–4⬚C to as much as ca. 10–11⬚C (Wright 1980), while others prefer intermediate figures of ca. 5–6 ⬚C (Krinsley 1970, van Zeist and Bottema 1977). Cold intervals were probably somewhat drier than today, and warm periods probably broadly resembled the present. It is important to point out that there are contrasting views as to the significance of the absence of trees during “glacial” episodes. Workers such as van Zeist and Bottema

698 F c u r r e n t a n t h ro p o l o g y

Fig. 5. Late Pleistocene and Early Holocene climates and vegetation according to van Zeist and Bottema (1977; from Smith 1986: fig. 11, reprinted with the permission of the University of Pennsylvania Museum).

(1977) and Smith (1986) interpret this as reflecting cold, dry conditions (the view taken here), while others (e.g., Farrand 1981) point to high lake levels in Iran and eastern Turkey and montane glaciers on Iranian mountains and conclude from this that cold intervals were also relatively humid. In this view, it was the cold and not the aridity that prevented tree growth (Smith 1986:10). Given its latitude, Iran during the Pleistocene appears to have been generally semiarid, although its climatic regimen was altered during cold intervals by lower temperatures and evaporation rates. Given the disagreement about the meaning of arboreal pollen just noted, the cooler phases did not necessarily lead to increased vegetation cover and more abundant animal resources for foragers. In the mountains, at least the summers, while short, would have been relatively mild, but, to judge from Coon’s experiences trying to traverse the Zagros during the late fall and winter of 1949, the winters were probably long, severe, extremely cold, and possibly with greater snowfall than at present (Wright 1976). Lake levels were seasonally higher during cool periods because of decreased evaporation, while during warm episodes and seasons they probably declined overall. Smaller seasonal lakes probably dried up altogether.

the lithic industries We divided the lithic collections into retouched pieces and de´bitage, including cores and shatter. Any flake, blade, or bladelet with at least some regular marginal retouch was scored as a retouched piece. De´bitage consisted of all unretouched flakes, blades, microblades and cores, as well as chipping debris. In an effort to determine rough stages in reduction sequences, we also noted the presence or absence of cortex on all flakes and blades and whether or not cores appeared to be exhausted. This allowed us to identify (1) a set of Paleolithic and early Epipaleolithic industries (Mousterian, Baradostian, or Aurignacian, probably Zarzian), found predominantly in sites bordering the Khorramabad plain, (2) Late Epipaleolithic and Proto-Neolithic industries, found on the margins of the Baba Zeid plain, and (3) nondescript flake industries, probably mostly Paleolithic, that do not resemble any Zagros lithic assemblages described in the literature. Twenty-one collections ranging in size from a single artifact (Dozaleh II rockshelter, region 3) to 357 pieces (Gilvaran I, region 1) were recovered. Although in no sense representative because of obvious factors of selection (e.g., slope wash, erosion in general, etc.) and heavily

Volume 45, Number 5, December 2004 F 699

Fig. 6. A synthesis of the Iranian Paleolithic. Chronological subdivisions are not to scale and positions of individual sites within units are only approximate (from Smith 1986: fig. 15, reprinted with the permission of the University of Pennsylvania Museum).

dominated by large flakes, blades, and cores, the collections nevertheless give a preliminary idea of the richness of the Stone Age archaeological record in the Iranian Zagros and provide a much-needed update on the decades-old research that has been reported in Western literatures. The small Mousterian collection (15 pieces) from Gachi rockshelter is perhaps the most informative of those assigned to the Middle Paleolithic. It produced no fewer than six sidescrapers (single, double and convergent), a Mousterian point, and an exhausted disk core (fig. 7). Moreover, the pieces are textbook examples and heavily reduced and thus tend to support Lindly’s conclusion that the region was occupied only seasonally during the Mousterian and that hominids were bringing relatively large, easily modified flake and blade tools and cores into the highlands and subsequently reworking them as circumstances required (Lindly 1997). There are

no suggestions of primary reduction in the Gachi surface collection. Other collections are assigned to an undifferentiated Upper and/or Epipaleolithic because pieces such as Arjeneh points, keeled or carinate endscrapers, and Dufour bladelets, supposedly diagnostic of the Baradostian or Zagros Aurignacian (more generally, early Upper Paleolithic), do not occur in them. Typical Zarzian pieces are also lacking (e.g., geometric microliths, small burins). The two large collections from Sorkh-e Lizeh and Gilvaran I exhibit some generic Early Upper Paleolithic characteristics (e.g., many flakes, retouched pieces made on flakes, flake cores, denticulates and notches, sidescrapers), but they occur along with lamellar elements and even bullet cores (figs. 8 and 9). Sidescrapers, notches, denticulates, and retouched flakes predominate in early Upper Paleolithic layers AA–LL at Warwasi (Olszewski and Dibble 1994). Hammerstones and grinding

700 F c u r r e n t a n t h ro p o l o g y

Fig. 7. Gachi rockshelter, selected pieces. 1, Mousterian point; 2 and 3, convergent sidescrapers; 4, de´jete´ sidescraper; 5, notched flake; 6, convex sidescraper; 7, disk core. stones were reported by Hole and Flannery (1967) from their tests at Gar Arjeneh and also show up at Gilvaran I but are of uncertain chronological or diagnostic significance. Judging from the dates from Yafteh, the early Upper Paleolithic in the region would seem to fall between 40,000 and 30,000 years ago (Hole and Flannery 1967). Distinctive Zarzian diagnostics such as geometrics and other retouched microblades were not recovered in any of the survey collections, and the most likely candidate for a Zarzian site (probably Ab Zadeh Cave in region 1) must await testing to confirm its culture-stratigraphic affinity. All assessments of Zarzian affinity must rest upon comparisons with Olszewski’s comprehensive study of the Warwasi rockshelter, which contains a long Zarzian sequence (1993b). Three sites (Gilvaran I and II and Sorkh-e Lizeh) can plausibly be assigned to the Baradostian (Solecki 1958, 1963) or, as it is increasingly called, the Zagros Aurignacian (Olszewski 1993a, 1999, 2001; Olszewski and Dibble 1994), but the remaining five can only be reasonably claimed to be Upper and/or Epipaleolithic.6 An apparently later group of samples includes those 6. On the basis of textbook generalizations about these industries in highland western Asia.

from Takht-e Shir I and II and Vare Zard. Although one is from a rockshelter and the other from an adjacent open-air site, the two Takht-e Shir collections are very similar to one another, probably represent the same interval of time, and are almost certainly part of a single “site.” They consist mostly of small lamellar elements, blade and bladelet cores (fig. 10). To judge from their size and neatly parallel lamellar detachments, pressure flaking was used to detach the microblades. Although no bullet cores were recovered, two blades with silica gloss on them imply plant processing and perhaps signal an Epipaleolithic or Proto-Neolithic component. The Vare Zard collection is also dominated by lamellar elements and includes drills, endscrapers made on blades, and bullet cores, all of which are characteristic of the Zagros Neolithic (fig. 11 [Kozlowski 1999]). The third group of collections is dominated by nondescript chert flakes and flake cores recovered from four rockshelters and a cave on the margins of the Baba Zeid plain (region 2). These pieces generally represent simple percussion detachments from unprepared or minimally prepared cores, as indicated by the presence of cortical material on some of the platforms and on the dorsal faces of the flakes themselves. Technically, the latter would be considered primary or (mostly) secondary decortica-

Volume 45, Number 5, December 2004 F 701

Fig. 8. Sorkh-e Lizeh, selected artifacts. 1, cortical bladelet; 2, backed and truncated bladelet; 3, backed bladelet with inverse retouch; 4, flake with lateral retouch on both edges; 5, flake with inverse retouch; 6, unretouched blade; 7, denticulated blade; 8, plain blade; 9, notched flake; 10, multiple dihedral burin on a break; 11, double endscraper; 12, endscraper on a flake; 13, combination endscraper-sidescraper; 14, blade/bladelet core; 15, bullet core.

tion flakes (fig. 12). Retouched pieces are rare, and much of what could—by some stretch of the imagination—be considered retouch was likely produced simply by use of unretouched edges. The general impression is one of an expedient technology in which flakes were being detached for more or less immediate use, primarily as unmodified cutting tools. Lamellar technology is absent from these collections, although they likely represent a “structural pose” (Binford 1980) of a wider range of lithic reduction systems other aspects of which were directed toward the production of blades and bladelets. They cannot be dated, although a test at the most promising of these sites, Dar Eshkaft Cave, might allow us to do so or at least to put them into a more precise behavioral context.

conclusions With many potentially informative sites, adequate water, easily exploited plants and animals of known importance to foragers throughout the Middle and Upper Pleistocene, and a climate suitable for humans in at least some places, the province clearly was an attractive place for hunter-gatherers during the warm intervals in the Quaternary. The discovery of 21 sites, many with buried deposits, in the course of a very brief archaeological survey underscores the potential of the region for prehistoric investigations of all kinds but especially those bearing on the Middle and Upper Paleolithic and the biological and behavioral transitions between archaic and modern

702 F c u r r e n t a n t h ro p o l o g y

Fig. 9. Tang-e Barik I, selected artifacts. 1, retouched bladelet; 2, unretouched bladelet; 3, partially backed bladelet; 4, inversely retouched blade; 5, inversely retouched notched blade; 6 and 7, unretouched bladelets; 8, partially retouched flake; 9, flake endscraper; 10, multiple burin; 11 and 13, single platform bladelet cores; 12, polyhedric-carinate burin; 14 and 15, multiple platform flake cores.

Volume 45, Number 5, December 2004 F 703

Fig. 10. Takht-e Shir, selected artifacts. 1, plain flake; 2, sickle blade segment; 3, truncated sickle blade segment; 4, 7–9, bladelet cores; 5, flake struck from a bladelet core; 6, retouched blade segment. Homo sapiens. The work reported here produced five substantive results. First, we were able to test the idea that different geological zones of the Zagros were characterized by different kinds of sites. Region 1 (Khorramabad), located on the boundary between the Zagros thrust and the Zagros folded zone, is characterized by high mountain ranges with steep escarpments surrounding the Khorramabad plain, and most of its sites have formed at the base of the limestone cliffs abutting the margins of the plain in the thrust zone. Big caves such as Ghamari, Kunji, and Yafteh are more or less confined to this region. Region 2 (Baba Zeid) is located near the center of the folded zone. Its plains are not as extensive as those of region 1, and the landscape is one of lower relief, with fewer cliffs and escarpments. Most of its sites are relatively small compared with those in region 1, are found in small side

valleys that do not have direct access to the Baba Zeid plain, and date to later (i.e., post-Paleolithic) periods. For the Khorramabad sample, the ratio of Upper to Middle Paleolithic sites is similar to that proposed by Hole and Flannery (1967) for the same region (70.6%), and this could be taken to indicate denser Upper Paleolithic settlement than was typical of the Mousterian. This suggestion squares well with patterns detected in other parts of the Zagros such as the Marv Dasht region (Rosenberg 1988), the Hulailen Valley (Mortensen 1993), Izeh, in Khuzistan (Wright, 1979), and the plains of Kermanshah (Biglari and Heydari 2001). Although it clearly would be premature to ascribe a cause to this apparent increase in site frequency, more intensive land use by Upper Paleolithic foragers with improved technologies and living at higher population densities than their Middle Paleo-

704 F c u r r e n t a n t h ro p o l o g y

Fig. 11. Vare Zard, selected artifacts. 1, unretouched bladelet; 2, notched bladelet; 3, blade segment; 4, notched blade; 5, blade endscraper; 6, borer; 7, unretouched blade; 8, used blade; 9, flake with continuous retouch along two edges; 10, denticulated flake; 11, notch/sidescraper; 12 and 16, single-platform bladelet cores; 13 and 15, bullet cores; 14, mixed flake/blade core.

Volume 45, Number 5, December 2004 F 705

Fig. 12. Tool Kash I, selected artifacts. 1, burin: 2, secondary decortication flake; 3, denticulated flake; 4, inverse sidescraper; 5, disk core (exhausted). lithic predecessors could be a factor. Climatic change is unlikely to have played a major role. Second, we were able to show that there are significant differences in contexts between Middle Paleolithic sites and Upper Paleolithic and Epipaleolithic ones. All the Middle Paleolithic sites in region 1 except for Gar Arjeneh are in large rockshelters and caves, whereas all the Upper Paleolithic sites in the region except for Yafteh are in rockshelters. While we hesitate to propose an explanation for this pattern, it is probably linked to the kinds of solution cavities available before and after ca. 40,000 years ago. A third result was the recognition of previously unknown flake industries in the area surrounding the Baba Zeid plain. Although a functional explanation appears likely in the absence of any relative or absolute chronology, it is nevertheless true that no similar industries have been described for any Paleolithic period in the Zagros. A similar assemblage has, however, recently been reported from rockshelter deposits near Islamabad, some 150 km northwest of the Baba Zeid plain (Abdi 1999 and Biglari’s observations). Islamabad is located in the Zagros folded zone, and Abdi notes a remarkable consistency of raw material (brown and reddish brown chert nodules) across all these sites, so far undated. A fourth finding is the so-far unique Vare Zard site complex, a series of rockshelters cut into a cliff face with

a scatter of Epipaleolithic and Proto-Neolithic artifacts extending for more than 200 m along the slope in front of it. The sheer extent of the scatter suggests repeated use of the location over a substantial interval in the early Holocene in a setting quite distinct from that of the penecontemporaneous Belt and Hotu Caves along the shore of the Caspian Sea, several hundred kilometers to the northeast (Coon 1957). Although there are no indications that all or most of the Vare Zard shelters were occupied simultaneously, the numerous projections between the relatively shallow shelters would have facilitated the construction of tents, ramadas, or lean-tos that could have sheltered a group in excess of 100 individuals. The near certainty of buried deposits at the site augurs well for future excavations. Finally, important paleoclimatic evidence came from a pothole at the Gachi rockshelter, our best candidate for a Mousterian site with buried deposits. Two strata were exposed in a section measuring ca. 90 cm deep. The uppermost level (ca. 40 m deep) consisted of a flowstone deposit incorporating cryoclastic debris that overlies the cultural layers and may eventually provide a terminus ante quem uranium-series date for the Mousterian deposits underneath. Flowstones are common features of active karstic systems and usually form at some distance from the cave mouth, which implies that the shelter overhang once extended outward some meters from its

706 F c u r r e n t a n t h ro p o l o g y

present location. The precipitation of calcium carbonates that results in flowstone deposits also must have taken place under a hydrologic regime that was considerably wetter than that of today. An analogous situation occurs at Shanidar Cave, in Iraqi Kurdistan, some 250 km to the northwest (Solecki 1961, 1971). There a thin flowstone deposit was uncovered in Mousterian layer D at a depth of ca. 8.5 m below the modern cave floor. Long before the advent of oxygen-isotope-based marine paleotemperature chronologies (which he himself pioneered), Robert Emiliani assumed that this layer corresponded to Wu¨rm I in the European continental glacial sequence, which at the time implied an age of ca. 60,000 years ago (Solecki 1961).7 Pointing to pollen and sediment evidence from the Humian rockshelter, Bewley (1980, 1984) also maintains that the Zagros Mousterian developed under warm, humid conditions contemporary with the flowstone deposit at Shanidar. If the flowstone deposit at Gachi could be shown to be contemporaneous with the travertines at Shanidar and Humian, it would imply that the Mousterian at Gachi also dates to a warm, wet period ca. 60,000 years ago (in modern terms, early in oxygen isotope stage 3). Whatever the case, it might be possible to date the flowstone directly and thus provide the beginnings of a radiometric chronology for the region.

References Cited a b d i , k . 1999. Archaeological research on the Islamabad plain: Preliminary report, 1998 and 1999 (in Farsi). Iranian Journal of Archaeology and History 26/27:47–60. a m i r l o o , e . 1990. Archeological survey of the Demavand region (in Farsi). Iranian Journal of Archaeology and History 7: 78–79. b a r - y o s e f , o . 1994. The Lower Paleolithic of the Near East. Journal of World Prehistory 8:211–65. b a u m l e r , m . , a n d j . d . s p e t h . 1993. “A Middle Paleolithic assemblage from Kunji Cave, Iran,” in The Paleolithic prehistory of the Zagros-Taurus. Edited by D. Olszewski and H. Dibble, pp. 1–73. Philadelphia: University of Pennsylvania Museum. b e w l e y, r . h . 1980. Houmian, Iran. M.Phil. thesis, Department of Archaeology, University of Cambridge, Cambridge, U.K. ———. 1984. The Cambridge University Archaeological Expedition to Iran, 1969. Excavations in the Zagros Mountains, Houmian, Mir Malas, and Barde Spid. Iran 22:1–38. b i g l a r i , f . 2000. Lithic analysis of artifacts from the Luristan Paleolithic Survey, winter 2000. MS, Institute of Archaeology, Iranian Cultural Heritage Organization, Tehran, Iran. ———. 2001. Report on newly discovered Paleolithic sites at Bisitun, central-western Iran (in Farsi). Journal of Archaeology and History 28:50–60. b i g l a r i , f . , a n d k . a b d i . 1999. Paleolithic artifacts from Cham-e Souran, the Islamabad Plain, central-western Zagros Mountains, Iran. Archa¨ologische Mitteilungen aus Iran und Taurus 31:1–8. b i g l a r i , f . , a n d s . h e y d a r i . 2001. Do-Ashkaft: A recently discovered Mousterian cave site in the Kermanshah Plain, Iran. Antiquity 75:487–88. b i g l a r i , f . , g . n o k a n d e h , a n d s . h e y d a r i . 2000. A re7. Although probably impractical because of the unsettled political situation, it would be very useful to go back to Shanidar and try to recover a sample of this flowstone for uranium-series dating.

cent find of a Paleolithic assemblage from the foothills of the Zagros Mountains. Antiquity 74:749–50. b i n f o r d , l . 1980. Willow smoke and dogs’ tails: Hunter-gatherer settlement systems and archaeological site formation. American Antiquity 45:4–20. b r a i d w o o d , r . j . 1960. Seeking the world’s first farmers in Persian Kurdistan: A full-scale investigation of prehistoric sites near Kermanshah. Illustrated London News 237:695–97. b r a i d w o o d , r . j . , a n d b . h o w e . 1960. Prehistoric investigations in Iraqi Kurdistan. Oriental Institute, University of Chicago, Studies in Ancient Oriental Civilization 31. b r o o k s , i . a . 1982. “Geomorphological evidence for climate change in Iran during the last 20,000 years,” in Paleoclimates, paleoenvironments, and human communities in the Eastern Mediterranean region in later prehistory, pt. 1. Edited by J. L. Bintliff and W. van Zeist, pp. 191–228. British Archaeological Reports International Series 133. b r o u d e , j . 1990. Geological map of Bakhtaran (Kermanshah). Translated by A. Agha Nabati. Tehran: Iranian Geological Organization. c o o n , c . s . 1951. Cave explorations in Iran 1949. Philadelphia: University of Pennsylvania Museum. ———. 1957. The seven caves: Archaeological explorations in the Middle East. New York: Alfred Knopf. d a r v i s h z a d e h , a . 1981. Geology of Iran. Tehran: Danesh-e Emroos Publishers. d a v i s , r . , a n d v. r a n o v. 1999 Recent work on the Paleolithic of Central Asia. Evolutionary Anthropology 8:186–93. d i b b l e , h . , a n d s . h o l d a w a y. 1993. “The Middle Paleolithic industries of Warwasi,” in The Paleolithic prehistory of the Zagros-Taurus. Edited by D. Olszewski and H. Dibble, pp. 75–100. Philadelphia: University of Pennsylvania Museum. f a r r a n d , w. 1981. “Pluvial climates and frost action during the last glacial cycle in the eastern Mediterranean: Evidence from archaeological sites,” in Quaternary paleoclimate. Edited by W. C. Mahaney, pp. 393–409. Norwich: GeoAbstracts. fi e l d , h . 1951. Reconnaissance in southwestern Asia. Southwestern Journal of Anthropology 7:86–102. h e y d a r i , s . 1999a. Geoarchaeology of Paleolithic sites in the Southwestern Prave Mountains (Kermanshah Region) and a study of the paleoclimate at the end of the Quaternary. M.A. thesis, Department of Geology, Islamic Azad University at Najaf-Abad, Tehran, Iran. ———. 1999b. A study of the paleoclimate and its relationship with the Paleolithic sites in Luristan Province. MS, Institute of Archaeology, Iranian Cultural Heritage Organization, Tehran, Iran. h o l e f . , a n d k . fl a n n e r y. 1967. The prehistory of southwestern Iran: A preliminary report. Proceedings of the Prehistoric Society 33:147–206. h o l e , f . , k . fl a n n e r y, a n d j . n e e l e y. 1969. Prehistory and human ecology of the Deh Luran Plain. Memoirs of the Museum of Anthropology, University of Michigan, 1. k l e i n , r . 1999. The human career: Human biological and cultural origins. Chicago: University of Chicago Press. k o z l o w s k i , s . 1999. The eastern wing of the Fertile Crescent: Late prehistory of Greater Mesopotamian lithic industries. British Archaeological Reports International Series 760. k r i n s l e y, d . 1970. A geomorphological and paleoclimatological study of the playas of Iran. 2 pts. Washington, D.C.: Air Force Cambridge Research Laboratories. l i n d l y, j . 1997. The Zagros Mousterian: A regional perspective. Ann Arbor: University Microfilms. ———. n.d. The Zagros Mousterian: A regional perspective. Arizona State University Anthropological Research Papers 57. In press. m c b u r n e y, c . 1969a. On an examination of rock paintings in the Kuh-i Dasht area. Bastan Shenasi Va Honar-e Iran 3:7–8. ———. 1969b. Report on further excavations in the caves of the Kuh-i Dasht area during August, 1969. Bastan Shenasi Va Honar-e Iran 3:8–9. ———. 1970. Paleolithic excavations in the Zagros Mountains. Iran 8:185–86.

Volume 45, Number 5, December 2004 F 707

m e r c i e r , n . , a n d h . v a l l a d a s . 1994. “Thermoluminescence dates for the Paleolithic Levant,” in Late Quaternary chronology and paleoclimates of the Eastern Mediterranean. Edited by O. Bar-Yosef and R. Kra, pp. 13–20. Tucson: Radiocarbon, Inc. m i n z o n i - d e´ r o c h e , a . 1993. “Middle and Upper Paleolithic of the Taurus-Zagros region,” in The Paleolithic prehistory of the Zagros-Taurus. Edited by D. Olszewski and H. Dibble, pp. 147–58. Philadelphia: University of Pennsylvania Museum. ———. 1999. The Middle and Upper Paleolithic periods in the Taurus-Zagros region. Translated by F. Biglari. Journal of Archaeology Students in Tehran University 11:19–24. m o r g a n , j . d e 1924. Le monde oriental avant l’histoire: L’Asie ante´rieure et l’Egypte. L’Anthropologie 34:17–56. ———. 1926. La pre´histoire orientale. 2 vols. Paris: Masson. m o r t e n s e n , p . 1993. “Paleolithic and Epipaleolithic sites in the Holailan Valley, northern Luristan,” in The Paleolithic prehistory of the Zagros-Taurus. Edited by D. Olszewski and H. Dibble, pp. 159–86. Philadelphia: University of Pennsylvania Museum. o b e r l a n d e r , t . m . 1968. “The origin of the Zagros defiles,” in The Cambridge history of Iran, vol. 1. Edited by W. B. Fisher, pp. 195–211. Cambridge: Cambridge University Press. o l s z e w s k i , d . i . 1993a. “The late Baradostian occupation at Warwasi rockshelter, Iran,” in The Paleolithic prehistory of the Zagros-Taurus. Edited by D. Olszewski and H. Dibble, pp. 187–206. Philadelphia: University of Pennsylvania Museum. ———. 1993b. “The Zarzian occupation at Warwasi rockshelter, Iran,” in The Paleolithic prehistory of the Zagros-Taurus. Edited by D. Olszewski and H. Dibble, pp. 207–36. Philadelphia: University of Pennsylvania Museum. ———. 1999. “The early Upper Paleolithic in the Zagros Mountains,” in Dorothy Garrod and the progress of the Palaeolithic. Edited by W. Davies and R. Charles, pp. 167–80. Oxford: Oxbow Books. ———. 2001. “Ruminations on the Early Upper Paleolithic and a consideration of the Zagros Aurignacian,” in Questioning the answers: Re-solving fundamental problems of the Early Upper Paleolithic. Edited by M. Hays and P. Thacker, pp. 79–90. British Archaeological Reports International Series 1005. o l s z e w s k i , d . , a n d h . d i b b l e . 1994. The Zagros Aurignacian. current anthropology 35:68–75. r o s e n b e r g , m . 1988. Paleolithic settlement patterns in the Marv Dasht, Fars Province, Iran. Ann Arbor: University Microfilms. r o u s t a e i . k . 1999. Studying and documenting caves, rockshelters, and open Paleolithic sites in Luristan Province. MS, Institute of Archaeology, Iranian Cultural Heritage Organization, Tehran, Iran. s c h w e i z e r , g . 1975. Untersuchungen zur Physiogeographie

von Ostanatolien und Nordwestiran. Tu¨bingen: Geographischen Instituts der Universita¨t Tu¨bingen. s k i n n e r , j . 1965. The flake industries of Southwest Asia: A typological study. Ph.D. diss. Columbia University, New York, N.Y. s m i t h , p . e . l . 1967. Ghari Khar and Ganj-i Dareh. Iran 5: 138–39. ———. 1986. Paleolithic archaeology in Iran. (American Institute of Iranian Studies, Monograph 1.) Philadelphia: University of Pennsylvania Museum. s o l e c k i , r . s . 1958. The Baradostian industry and the Upper Palaeolithic of the Near East. Ph.D. diss., Columbia University. New York, N.Y. ———. 1961. Three adult Neanderthal skeletons from Shanidar Cave, northern Iraq. Sumer 17:71–94. ———. 1963. Prehistory of the Shanidar Valley. Science 139: 189–93. ———. 1971. Shanidar: The first flower people. New York: Alfred Knopf. s p e t h , j . 1971. Kunji Cave. Iran 9:172–73. s t o¨ c k l i n , j . 1974. “Possible ancient continental margins in Iran,” in The geology of continental margins. Edited by C. Burk and C. Drake, pp. 873–87. New York: Springer-Verlag. ———. 1981. “A brief report on geodynamics in Iran,” in ZagrosHindu Kush-Himalaya geodynamic evolution. Edited by H. K. Gupta and F. Delany, pp. 70–73. Boulder: Geological Society of America. t r i n k a u s , e . 1983. The Shanidar Neandertals. New York: Academic Press. v a h d a t i n a s a b , h . 1999. Paleobiology and animal distributions associated with Paleolithic sites in Luristan Province. MS, Institute of Archaeology, Iranian Cultural Heritage Organization, Tehran, Iran. v a n z e i s t , w. , a n d s . b o t t e m a . 1977. Palynological investigations in western Iran. Palaeohistoria 19:19–85. w r i g h t , h . e . , j r . 1960. “Climate and prehistoric man in the eastern Mediterranean,” in Prehistoric investigations in Iraqi Kurdistan. Edited by R. Braidwood and B. Howe, pp. 71–95. Oriental Institute, University of Chicago, Studies in Ancient Oriental Civilization 31. ———. 1961. Pleistocene glaciation in Kurdistan. Eiszeitalter und Gegenwart 12:131–64. ———. 1976. The environmental setting for plant domestication in the Near East. Science 194:385–89. ———. 1980. Climatic change and plant domestication in the Zagros Mountains. Iran 18:145–48. w r i g h t , h . t . 1979. Archaeological investigations in Northeastern Xuzestan. Museum of Anthropology, University of Michigan, Research Reports in Archaeology 10. y o u n g , t . c . , a n d p . e . l . s m i t h . 1966. Research in the prehistory of central-western Iran. Science 153:386–91.