ABSTRACT

Ahmarian assemblages in the southern Levant were situated on or directly adjacent to water sources, regardless of elevation, while flake-based assemblages in the Central Negev currently labeled ‘Levantine Aurignacian’ were consistently situated on high plateaus at greater distances from water. Beyond a basic blade/flake dichotomy, other technological differences involving cortical elements and tool/debitage ratios occur between these areas on the landscape. These trends are examined in terms of two possible explanations: (1) sites situated away from water were drawn to raw material, and (2) sites situated away from water were occupied during a period of climatic improvement.

Introduction

Researchers have long recognized spatial and temporal trends within the Levantine Upper Paleolithic. Major phytogeographic differences in this small area imposed obvious archaeologically recognizable constraints on human subsistence. A recent study by Gladfelter (1997), for example, revealed that Ahmarian sites were almost exclusively situated in montane areas of the southern Levant, and date to a wet phase of the Upper Paleolithic. Such a pattern suggests that groups exploiting a Mediterranean environment followed its growth into interior, high-altitude zones during an interval of climatic improvement (Gladfelter 1997).

In a similar manner, this study reveals co-associated site placement and technological trends during the Upper Paleolithic. Specifically, Ahmarian assemblages throughout the Levant were situated on or directly adjacent to water sources, regardless of elevation, while flake-based assemblages in the Central Negev currently labeled ‘Levantine Aurignacian’ were consistently situated on high plateaus at greater distances from water. The purpose of this analysis is to describe this trend and discuss how it relates to other factors, such as raw material exploitation and paleoclimatic settings.

Background

To describe material revealed from extensive fieldwork in the Negev and Sinai during the late 1970’s, Gilead (1981) and Marks (1981) independently proposed a model, resembling ideas of Ronen (1976), which suggested that two traditions co-existed in the Levant: (1) the Levantine Aurignacian, and (2) the Ahmarian. In general terms, the ‘two-tradition’ model described the Levantine Aurignacian as technologically dominated by flakes, and typologically dominated by scrapers and burins, particularly carinated varieties. The Ahmarian was distinguished by a blade-bladelet technology and a toolkit composed of various elongated pieces with fine retouch, specifically el-Wad points.

Integral to the two-tradition model is the idea that technological differences between these traditions are unrelated to function, and instead represent the cultural traits of two different, largely contemporaneous groups (Gilead 1991, Marks 1981). While no direct evidence is available to indicate that groups of each tradition occupied the same space at the same time, some claim that broad spatial and temporal dispersal of the Ahmarian and Levantine Aurignacian are partly conformal. In general, however, the Levantine Aurignacian prevailed in the core Mediterranean zone, whereas the Ahmarian was mostly situated in the southern, ‘marginal’ steppes and deserts. These two commonly recognized phytogeographic zones, core and marginal, are based on the work of Zohary (1962), and have been adopted as analogies for the Pleistocene (Bar-Yosef and Belfer-Cohen 1989). The core zone along the eastern Mediterranean Sea receives sufficient annual precipitation to allow the growth of Mediterranean maquis dominated by oak, pistachio, cedar, pine, cypress, olive, laurel, and juniper (Gilead 1991). The marginal zone, to the south of the core, is an open steppe/desert with shrub (Irano-Turanian) or desert (Saharo-Arabian) plant communities corresponding to elevational zones.

Proponents of contemporaneity believe the Levantine Aurignacian existed in both areas between ca. 32,000 and 20,000 B.P. (Marks 1981, Gilead 1991), while the Ahmarian appeared earlier, around 38,000 B.P. at Boker A in the Central Negev (Marks 1981), and persisted until some 20,000 years ago. New dates from Kebara, however, reveal the appearance of an Early Ahmarian at 43,000 B.P., while the Levantine Aurignacian may have emerged some 36,000 years ago (Bar-Yosef et al. 1996). Other researchers believe that Levantine Aurignacian groups only occupied the Mediterranean core area (Belfer-Cohen 1995, Belfer-Cohen and Bar-Yosef 1981, Belfer-Cohen and Goring-Morris 1986). For example, Belfer-Cohen (1995: 247) notes that the Levantine Aurignacian has been broadened to a “macro-tradition”, which lumps under its heading most flake-oriented assemblages that lack many elements found in northern cave sites, such as a bone tool industry, nosed and shouldered endscrapers, lamellar retouch, and particular blade/flake ratios for debitage and tool blanks.

While no one denies that the rather dramatic environmental differences between the northern and southern Levant should be reflected in material remains, most researchers addressing pan-Levantine variability are concerned with choosing the most relevant industrial criteria for grouping or splitting assemblages under different names. This paper does not attempt to contribute to the current debate over how to define and describe each Upper Paleolithic tradition. Rather, an attempt is made to demonstrate how the artifacts of numerous Upper Paleolithic assemblages, given their tradition names, pattern differently with environmental settings, such as elevation and distance from permanent water sources.

Analysis

This analysis uses 51 Levantine Upper Paleolithic assemblages. Various archaeological units were recorded from the available literature for sites located in both phytogeographic zones, including cave sites in the core zone of Israel and Lebanon, and marginal open-air sites in the Negev, Sinai and Jordan (Table 1, Figure 1). These assemblages have sufficient artifact samples, and details that allow inter-assemblage comparison (adequate recovery techniques, debitage and tool counts, and environmental settings). Of these assemblages, 13 were labeled ‘Levantine Aurignacian’ by the excavator, 32 fall under ‘Ahmarian’, and the remaining five were never named. As revealed in Table 1, there is a paucity of data available from the Mediterranean core zone. Only nine core zone assemblages, all of which are located in caves or large rockshelters, were published in enough detail to allow basic technotypological and environmental comparisons. This analysis, therefore, is necessarily biased toward southern, marginal zone assemblages.

FIGURE 1. Map of the southern Levant, showing the distribution of Upper Paleolithic sites mentioned in text.

Data were organized into units that allowed comparison of relevant technological and environmental variables. Technological variables include relative percentages of flakes, blade-bladelets, primary elements, and core trimming elements within the debitage. Flakes and blade-bladelets have obvious importance, due to the general tendency for Upper Paleolithic assemblages to be either flake- or blade-dominated. Artifacts with cortex (primary elements and core trimming elements) were also deemed important because they generally indicate initial stages of reduction, which often reflects factors such as distance from raw material and curation. To obtain these technological indices, the total number of artifacts in each class (flakes, blade-bladelets, primary elements, core trimming elements) was divided by the total number of artifacts in the debitage category (excluding chips and chunks) for each assemblage. In some comparisons, restricted percentages of flakes and blades (class n / total flakes, blades, and bladelets) were used to narrow the scale of variability. In addition, environmental variables such as phytogeographic zone, elevation, paleo water source, and distance from water were recorded for each assemblage.

Technological Variability

The first variables to be tested were flakes and blade-bladelets, as they relate to initial core reduction and maintenance (core-trimming elements and primary elements). Flake and blade technologies exhibit different relationships with core-trimming elements and primary elements (Figure 2). Specifically, flakes are associated with higher percentages of cortex in the debitage, while blade-bladelets exhibit the opposite trend. The markers used in Figure 2 reveal the technological dichotomy between the flake-dominated Levantine Aurignacian and the blade-oriented Ahmarian. This difference is predictable and is largely due to the definition of each tradition, which emphasizes these particular technological differences. A noticeable outlier in Figure 2 is Ein Aqev, which falls outside the expected blade and flake percentages for the Levantine Aurignacian.

Of particular importance to the trends revealed in Figure 2 is the near absence of core Mediterranean assemblages. Hayonim D was the only Mediterranean core assemblage with sufficient data to allow comparison of the given variables. While Hayonim D conforms to the pattern exhibited by the southern Levantine Aurignacian, it is impossible to generalize that all assemblages in the core zone exhibit the same relationship between flakes, blades, and decortication. Assemblages in marginal settings, however, display a demonstrable pattern: flake-dominated technologies generally exhibit high percentages of cortical debitage (> 20 % primary elements and core trimming elements).

Numerous factors could have influenced these trends (Figure 2), such as distance from raw material, curation, a technological continuum, or blade technology efficiency. Further, these patterns could be related to post-depositional agents, such as the removal of smaller artifacts during deflation and slope wash. To test the possibility that deflated assemblages have erroneously high percentages of larger artifacts (i.e., primary elements and core trimming elements) and low percentages of small artifacts (chips and bladelets), a linear regression was calculated for two variables: average thickness of the excavated layer and the relative percentage of blades/ bladelets. Blades were included because most site reports did not separate blades & bladelets in their tabulations. If bladelets were removed during deflation, then assemblages with deeper, in-situ deposits should have higher percentages of blade-bladelets. In fact, this is not the case for the assemblages used in this analysis. The linear regression revealed an almost completely random distribution (R-square = .0003).

FIGURE 2. Scatterplots showing the relationship between decortication and flake/blade technological orientation.

Hit Counter