Early and Middle Pleistocene Faunal and hominins dispersals through Southwestern Asia

Abstract This review summarizes the paleoecology of the Early and Middle Pleistocene of southwestern Asia, based on both flora and fauna, retrieved from a series of ‘windows’ provided by the excavated sites. The incomplete chrono-stratigraphy of this vast region does not allow to accept the direct chronological correlation between the available sites and events of faunal and hominin dispersals from Africa. It also demonstrates that hominins survived in a mixed landscape of open parkland with forested surrounding hills. In addition, the prevailing environmental conditions are not sufficient to explain successful adaptations to new ecological niches away from the African savanna of the bearers of ‘core and flake’ and the Acheulian industries, The differences in knapping and secondary shaping of stone artifacts probably reflect the learned traditions of different groups of hominins. The current distribution of lithic industries across Eurasia is undoubtedly incomplete due to lack of cultural continuities as well as paucity of field research in several sub-regions. This observation supports the contention that what we view as a constant stream of migrants was actually interrupted many times. The continuous occupation of southwestern Asia by the makers of the Acheulian is in contrast with neighboring regions such as the Iranian plateau and Eastern Europe. A more complex model is required to explain the recorded Eurasian archaeological–cultural mosaic.


Introduction 30
The effect of climate change on the tempo and mode of early hominids dispersals 31 from Africa during the Early and Middle Pleistocene is one of the main interests in 32 paleoanthropology and Paleolithic archaeology (Behrensmeyer, 2006). It has been 33 suggested that the expansion of savanna environments during the Early Pleistocene 34 allowed for the first 'Out of Africa I' dispersal (Dennell, 2004;Martínez-Navarro, 2004) 35 and that humans dispersed into Western Europe during warmer periods and that the 36 dispersal was highly influenced by climate rather than culture, which played only a minor 37 role (Agustí et al., 2009a). However, these theories assume as a working hypotheses that 38 the presence of hominins and faunal present in the paleontological and archaeological 39 record is continuous. Here we would like to focus on southwestern Asia that served as the 40 main corridor for hominin dispersal during the Early and Middle Pleistocene and address 41 the assertion that we can test hominids co-dispersed with other taxa and the extent of the 42 response to climatic forcing. We will propose that there is no correlation between 43 hominid and faunal dispersals and suggest that early Homo dispersals were not directly 44 controlled by climatic forcing. 45 Geological and archaeological investigations in southwestern Asia produced a 46 few "windows" into the geological sequence of the Early and Middle Pleistocene. However, it is most important to remember that the home ranges of the species is 196 fluid rather than static, and are constantly changing dependant on climatic and 197 environmental changes and inter and intra species competition. Therefore, the presence-198 absence of faunal elements correspond to the vegetation conditions, and disperse when 199 resources diminish during worsening climatic conditions that could also result in 200 extinction (Bennett, 1997). Thus in the following sections, we will discuss the changes in There are special localities in the Levant that were thought to provide evidence 215 for early presence of hominins, but a close examination do not stand up to the 216 geochronological criteria. The mammalian fauna of Bethlehem is the oldest Plio-217 Pleistocene assemblage known from the Levant. This was an animal bone accumulation 218 incorporated in a deposit of flint fragments and clay discovered when a local farmer was 219 digging a well (Gardner and Bate, 1937). The small collection of flints was originally 220 thought to include human artifacts but they have since been shown to be natural (Hooijer, 221 1958;Clark, 1961). 222 Another claim for an Early Pleistocene site named Yiron where core and flakes 223 have been retrieved from the gravels in a major crevice. Unfortunately the Yiron basalt 224 on the plateau did not cover this location and thus the date of the lava flow of 2.4 Ma 225 (Ronen et al., 1980;Ronen, 1991a), is unrelated to the human made lithics. 226 The 'Erq el Ahmar Formation (Horowitz, 1979) is located about 14 km south of 227 the Sea of Galilee, in the Jordan Valley and acumulated prior to the 'Ubeidiya Formation. 228 The polarity sequence is correlated with late Gauss and early Matuyuma chrons, within 229 which the Olduvai subchron was identified (Ron and Levi, 2001). Unfortunately the 230 artifacts mentioned in the different papers were found on the surface, except for a few 231 retrieved from a fluvial conglomerate of an unknown stratigraphic position (Tchernov,232 limestone. The basalt nodules occurred as pebbles, cobbles, boulders, and scree 282 components; the limestone was available as much more sparse cobbles within the beach 283 and wadi deposits; and the flint is found in the same environments as small pebbles and 284 cobbles. The 'Ubeidiya hominins employed each type of rock to shape a different type of 285 object (Bar-Yosef and Goren-Inbar, 1993). Core-choppers and light-duty tools were made 286 of flint, spheroids mainly of limestone, and the handaxe group from basalt, with a few of 287 flint and limestone. There is a direct correlation between the size of the tool-category and 288 the type of raw material. Although basalt is the most common rock and found in every 289 lithological facies at 'Ubeidiya the most abundant object is the core-chopper, which is 290 made of flint and its detached pieces (flakes). It seems that the lithic assemblages from 291 the lower most layers in the sequence (K/III-12, III-20-22, II-23-24) contain an 292 abundance of core-choppers, polyhedrons and spheroids but lack bifaces, except for one 293 trihedral. The samples are large enough to suggest that they may indicate the presence of 294 an early group of hominins that did not produce bifaces. The overlying assemblages 295 contain bifaces in varying frequencies and can be called Early Acheulian (Bar-Yosef and 296 Goren-Inbar, 1993). As the presence or absence of bifaces is taken to designate different 297 groups of hominins one may speculate that as in Dmanisi, the first hominins to arrive at 298 'Ubeidiya, were not the bearers of the Acheulian. 299 Among the so-called formal categories the frequencies of bifaces are abundant in 300 gravelly layer K-30 although the underlying layer K-29 of the same wadi fill produced 301 very few bifaces. It seems that certain activities were carried in the hilly-forested areas 302 from which the more abraded assemblage of K-30 was derived. The almost total 303 disappearance of bifaces in the later assemblages at 'Ubeidiya is noticeable and as yet 304 unexplained. 305 Estimated dates for the fossil-bearing strata of the 'Ubeidiya Formation are 306 between ca. 1.6 -1.2 Ma. Paleomagnetic analysis of the 'Ubeidiya Formation indicated a 307 reversed polarity suggesting that it predates the Brunhes -Matuyama reversal (Opdyke et 308 al., 1983;Braun et al., 1991;Verosub and Tchernov, 1991 (Rink et al., 2007). 314 The dating of these short polarity events was also corroborated by local faunal 315 turnovers (Belmaker, 2009). The 'Ubeidiya fauna (Table 1) can be assigned to a local 316 mammalian fauna biozone older than the assemblages of Bitzat Ruhama, Evron-Quarry 317 and Latamne (Belmaker, 2009

Evron-Quarry 329
The excavations at Evron-Quarry, located in the coastal plain of the Western 330 Galilee, exposed a sequence of alternating deposits of sandstone (kurkar), sometimes up 331 to three meters thick, and red-brown loams (hamra), either as isolated lenses or layers up 332 to about one to four m thick (Ronen, 1991b). The layer that contained the Acheulian 333 industry had two distinct horizons of calcareous concretions, occasional artifacts, and 334 sandy clay lenses with pebbles (two to three meter thick) separated from the dark brown-335 black clay (two m thick) that contained Late Acheulian artifacts and a few animal bones. 336 The artifacts within the archaeological horizon had a vertical distribution of 15-25 cm 337 deep, a phenomenon interpreted as the result of repeated occupations. 338 The archaeological assemblage contained small pebbles of quartz, limestone, and 339 flint, with most of the artifacts made of the latter raw material. No bifaces were found in 340 the excavated areas probably due to their small surface but earlier searches in the quarry 341 dumps recovered twenty handaxes. These are of large size (140-220 mm in length) and 342 demonstrate a relatively crude workmanship that resembles that of the Latamne site 343 (Ronen, 1991b). Large cobbles from which the bifaces were made were recovered 344 together with a group of hard calcite geodes, the heaviest of which was 580g, were 345 brought to the site by the occupants from about 5 km away. 346 Chronologically, the context of Evron-Quarry is assigned to a post-'Ubeidiya age, 347 and is perhaps contemporaneous with the Latamne site in Syria (ca 1.0-0.8 Ma). 348 Paleomagnetic and ESR studies of the archaeological bearing strata have suggested a date 349 ca. 1.0 Ma (Porat and Ronen, 2002;Ron et al., 2003).

Bizat Ruhama 352
The site of Bizat Ruhama is situated in the southern coastal plain and the 353 archaeological horizon is incorporated in a paludine deposit (Ronen et al., 1998;Zaidner, 354 2003a, b;Zaidner et al., 2003). The retrieved rich lithic industry, generally made of small 355 flint pebbles, falls within the category of "core and flake" assemblage with intensively 356 retouched flakes (small borers, notches and denticulates, etc.) No bifaces were found. 357 Paleomagnetic information and TL dates places the site at 0.99 to 0.85 Ma (Laukhin et 358 al., 2001;Ron and Gvirtzman, 2001). 359

360
Latamne 361 The site of Latamne was excavated by D. Clark (1967) and additional fieldwork 362 was conducted by Sanlaville and his associates (1993). The archaeological horizon of 363 Latamne lay in the mid-sequence of the Latamne Formation. It contained the Latamne 364 "occupation floor," a silt layer only a few centimeters thick (up to 10cm) capped by 365 sandy-silty bedding with traces of rootlets. The sequence was interrupted by erosion and 366 was overlaid by a fluvial sandy deposit, capped by a lacustrine layer (Sanlaville et al., 367 1993). occurrences (e.g., Hours, 1975Hours, , 1981Bar-Yosef, 1987;Goren-Inbar, 1995;Liubin, 2002;405 Taskiran, 1998). The nature of the deposits and the malacological assemblages, 406 dominated by Viviparus apameae, indicate that the archaeological assemblages 407 accumulated on the shores of an expanding lake that flooded the gorge (Horowitz, 1979; 408 Goren-Inbar et al., 1991, 1992aGoren-Inbar and Saragusti, 1996;Feibel, 2004). 409 The complex stratigraphic sequence, first partially exposed by M. Stekelis (1960)  The archaeological horizons of GBY are embedded in a depositional sequence 420 that accumulated above a lava flow with normal polarity. The lava flow, designated as the 421 Yarda Basalt, was first K/Ar dated to 0.68±0.12 Ma (Horowitz, 1979) and later to 422 0.9±0.15 Ma (Goren-Inbar et al., 1992a). However, the paleomagnetic sequence within 423 the excavated deposits demonstrated that the Matuyama Brunhes Boundary (MBB) was 424 located within the series of deposits and the accumulation of the entire sequence took 425 place during ca. 100,000 years (Feibel, 2004). The lithic assemblage was characterized as Late Acheulean (Garrod and Bate, 1937). 476 The TL dates for Tabun Ed to D suggested that they were deposited during MOIS 477 9 through 8 (Mercier et al., 1995), more recent dates by Mercier  Surveys along the Levantine coast located a few occurrences that appear to be of 493 Early Pleistocene age although dating, in most cases, is rather tenuous due to lack of 494 datable substances. Shorelines were dated on the basis of their elevation above sea level, 495 while the known relative ages of foraminifera and marine shell assemblages were also 496 incorporated into these figures along the western mountainous range along the Levantine 497 coast, a few artifacts were found on terraces as high as 120 m above sea level. We 498 therefore choose to summarize those sites where systematic excavations or surveys were 499 conducted and their dates depend of the geological stratigraphy and its suggested 500 correlation with sea level chronology. Here we add some brief information about the 501 main occurrences, and for a full survey see the available summaries (e.g., Hours 1975Hours , 502 1981Muhesen, 1985Sanlaville et al., 1993;Bar-Yosef, 1998a). 503 One of the distinctive sites is Kefar Menachem, situated in the interior part of the 504 coastal plain and was excavated twice (Gilead and Israel, 1975;Barzilai, 2006). The 505 lithic industry is embedded in red loam of an Early Pleistocene age (Horovitz, 1979). The 506 lithic assemblages of both excavated areas are comprised of numerous core and flake 507 products, a few flake-tools (classified as end-scrapers, side scrapers, burins, notches, and 508 denticulates). The use of direct hard hammer percussion is dominant. To date, the rare 509 bifaces have been found only on the surface and their attribution to the excavated 510 collections is doubtful. These bifaces are described as irregular ovates, picks, long 511 lanceolates, and backed bifaces (Gilead and Israel, 1975)  The general technological tendency among the Achuelian sites is toward a greater 525 use of soft hammer percussion and the sporadic appearance of the Levallois technique 526 (Copeland and Hours, 1981). Typologically, the almost total disappearance of core-527 choppers is noticeable. The cordiform and amygdaloid bifaces outnumber the ovates. The 528 length of the handaxes decreases in general, a tendency that was already noted by D. other groups with an occasional dominance of the rounded aspect over the cordiform 555 aspect (Yiron, Beith Uziel, Baqaa-Rafaim etc.). As in the EK group, the Levallois 556 technique was practiced in some sites. It is worth noting that despite the hilly distribution, 557 these assemblages are not present in the three caves where Late Acheulian layers were 558 uncovered (Tabun F, Abu Sif, Umm Qatafa D). 559 The frequency of refinement index (thickness/breadth x 100) demonstrates the 560 differences among the sites. The same is probably true when the mean length among Late 561 Acheulian sites is considered. Wherever large cobbles were available, there was a 562 tendency towards larger bifaces. However, a general tendency for decrease in biface 563 length could indicate increasing efficiency of resharpening (perhaps longer curation?) 564 during the Late Acheulian. 565 The flake industry of most of the Late Acheulian occurrences is not very well 566 known. In some places, the number of flakes cannot account for their manufacture. For 567 instance, the thousands of bifaces found in Ma'ayan Barukh may have been produced in 568 an area further north near the Litani River. The flakes collected from the same surface 569 clusters could indicate some resharpening (although small flakes and chips are not easy to 570 retrieve in the deep red soil of these hills). It seems that the concentration of bifaces near 571 the Hula Lake shores on the interfluves of freshwater creeks may represent repeated 572 butchering activities in a pristine environment. 573 A unique Late Acheulian site, embedded between a lava flow dated to 233±3 ka 574 and an older lava flow dated to 800 ka, was excavated on the edge of the crater lake 575 known as Berekhat Ram on the Golan plateau (Goren-Inbar, 1985). The rich assemblage 576 contains several thousands of artifacts, mostly in mint condition, with about four hundred 577 of retouched pieces including eight small bifaces. The makers of the industry employed 578 the Levallois centripetal (radial) technique. A special find is a human figurine  Inbar, 1986) that recently received much attention in the debate concerning the capacities 580 of Archaic Homo sapiens or late Homo erectus (e.g., Marshack, 1997;D'Errico, 2000). 581 The actual date of the site is unknown but given the current TL dates of the Acheulo-582 Yabrudian, it should be placed during the time span of 350-450 ka. 583 In general, Late Acheulian sites can be found across the southwestern Asia in 584 every environment including the coastal plain, hilly areas, inter-montane valleys, and 585 oases and in desert landscapes. The best example to date from an oasis situation is the 586 series of Late Acheulian assemblages which is characterized by high frequencies of 587 bifacial cleavers, uncovered in the Azraq basin (Copeland and Hours, 1989;Rollefson, 588 1997

The Lower Paleolithic of Iraq, Iran and the Caucasian region 626
The vast geographic area summarized here is generally poorly known. The scant 627 evidence from Turkey on one end of the region and India on the other, including a few 628 recorded find spots from Iran (Smith, 1986) and from the Arabian Peninsula (Zarins et 629 al., 1979(Zarins et 629 al., , 1980(Zarins et 629 al., , 1982Whalen et al., 1983Whalen et al., , 1984Abdul Nayeem, 1990) functioned as a selective filter for most organisms, through which only specific species 741 could spread or pass (Tchernov, 1988;Tchernov and Belmaker, 2004). 742 The onset of the Pliocene is marked by an abrupt transgression of the 743 Mediterranean and reestablishment of the barrier between Africa and Eurasia. Thus, 744 Africa became considerably isolated from the rest of the world by the Saharo-Arabian 745 arid belt. North Africa was even further isolated from both sub-Saharan Africa and the 746 eastern Mediterranean region (Thomas, 1985). During this period, the main biotic 747 influences in the region were Asiatic, primarily from the Irano-Turanic region. These are 748 recorded mainly in Anatolian sites, which include many forest dwellers such as the 749 Cervids. The Taurus Artemisia levels were over 90% indicating extreme aridity, however these conditions 782 were not sustained over a period longer than 10ka (Tzedakis et al., 2006). 783 The majority of pollen data is derived from late Pleistocene and Holocene 784 sediments and post date the period discussed in this study. A playnostraigraphy of the Levant, but unfortunately is not well dated. The sequence records alternations between 787 wet and dry Mediterranean flora. The interglacial flora was poor in arboreal pollen in 788 comparison to modern day conditions. A north -south gradient indicated 3-5% arboreal 789 pollen in the north associated with steppe elements while southern cores indicate no 790 arboreal pollen but proliferation of desert plants. In comparison, the pluvial periods or the 791 glacial flora was dominated by arboreal pollen, and of interest is the increase of pollen 792 the winter deciduous oak Quercus ithaburensis (Horowitz, 1988). There is a small pollen spectrum from Bitzat Ruhama (n=114) reflecting 811 relatively high frequencies of arboreal pollen including Quercus, Pinus, Olea and Cedrus. 812 (Zaidner, 2003b). Among the non-arboral pollen we encounter Chenopodiaceae, Palmae, 813 Poaceae and Liliaceae. The presence of cedar is indicative of a colder environment in the 814 region than the present (Zaidner, 2003b). bovids are represented only by 11 specimens (30%) .

Stephanorhinus etruscus, Mammuthus trogontherii, Equus suessenbornensis and E. 925
altidens (Vekua, 1986(Vekua, , 1987Hemmer et al., 2001;Tappen et al., 2002) and Dursunlu in 926 the generally more arid central Anatolia (1.0-0.78 Ma) is characterized by the high 927 frequency of Equus caballus mosbachensis and E. altidens (Güleç et al., 1999). Pleistocene sites of Qesem, Hayonim and Meged sites did not indicate any significant 960 change in niche utilization between the ungulates, suggesting that the change in 961 abundance may be due to human hunting preferences (Rowland, 2006). During this 962 period there is no sound evidence for an additional dispersal from Africa following the  (Table 2). 983 The evidence presented in this paper indicates that a few faunal dispersals 984 between African and Eurasia occurred throughout the Late Pliocene through the Middle 985 Pleistocene. There is a general agreement among scholars that these events were 986 generally coincident with global climate changes. The main difficulty is attributing each 987 dispersal event to the time when it took place. The reason is that the known sites and 988 assemblages, described and discussed above, do not necessarily date these events. We 989 tend to forget that the discovery of Lower Paleolithic sites is accidental and they cannot 990 be considered as a continuous chronological sequence. Not surprisingly this is the nature 991 of windows into the remote past, and one can argue that the dispersals of large mammals 992 The site of GBY is interpreted to be the archaeological remains of a group of 1019 hominids that migrated from Africa (Bar-Yosef, 1987; Goren-Inbar and Saragusti, 1996). 1020 The lithic industry at GBY bears an African stamp and is still a rare occurrence in the 1021 Human migrations occurred along the "Levantine Corridor" as defined by the 1031 paleontologists (e.g., Thomas, 1985). The Lower Paleolithic assemblages from el- that what we view as a constant stream of migrants, was actually interrupted many times 1094 (e.g., Dennell, in press). 1095 Chronologically, the earliest hominins reaching southwestern Asia after leaving 1096 Africa were the makers of 'core and flake' industries, and is probably why this kind of 1097 simple way of obtaining sharp edges spread across Asia and Eastern Europe. It is not 1098 impossible that even migrants in later time (such as the makers of the Karari industry) 1099 carried the 'core and flake' production into the same areas. This would explain why the 1100 Acheulian contexts dated to 1.5 Ma to 0.25 Ma demonstrate interstratifications with those 1101 who manufactured cores and flakes (sometimes with additional types such as spheroids). 1102 The continuous occupation of southwestern Asia by the makers of the Acheulian 1103 industry needs to be stressed. The contrast with the neighboring regions is striking. No there are suitable raw materials for making bifaces, but those who needed and knew how 1108 to make these tool-types were not present in these region. Natural boundaries such as the 1109 mountain ranges of the Caucasus limited early hominin moves as shown by the 1110 decreasing numbers of Acheulian handaxes (Liubin, 2002). 1111 In brief, we have no persuasive explanation why the Acheulian is not represented 1112 in the vast area between the Zagros and Baluchistan. A similar observation concerns 1113 Eastern Europe where only 'core and flake' industries were found and stand in marked 1114 difference with the proliferation of handaxes in Western Europe. 1115 While we agree that technological innovations and their social role are also 1116 considered as an important adaptation for dispersal (Carbonell et al., 1995;Larick and 1117 Ciochon, 1996), we wonder how exactly the technological innovations assisted in the 1118 survival or organization of those early hominin groups, and in particular given the lack of 1119 correlation between lithic technology, typology and environment. Detaching a few flakes 1120 from a nodule can be done in different ways, as it often depends of the fracture 1121 mechanics of the raw material, its size, shape and volume (Hovers and Braun, 2009 and 1122 papers therein). 1123 Therefore, if we accept the notion that adaptation to a specific environment during 1124 Lower Paleolithic times, did not influence the way stone tools were shaped, we have to 1125 adopt the position that it was due to the intrinsic behavioral capacities, unique to 1126 hominins, that facilitated the moves from Africa into new territories. It should be 1127 remembered that the initial dispersal of hominins has been attributed to both 1128 morphological and behavioral characteristics such as the capacity for long distance walking (Steudel, 1994), endurance running (Bramble and Lieberman, 2004), heat 1130 adaptation (Walker and Leakey, 1993), greater brain capacity (Aiello, 1993;Aiello and 1131 Wheeler, 1995) and social structure (Kroll, 1994). Even the release from tropical diseases 1132 that allowed for an increase in population size in higher latitudes (Bar Yosef and Belfer-1133 Cohen, 2001), does not explain what happened to the various groups in Asia. 1134 The variability selection hypothesis states that the adaptability of hominins to a 1135 wide range of habitats and specifically to a variable climate in Africa may have provided 1136 a pre-adaptation to survival in novel environments (Potts, 1998(Potts, , 2002. While we agree 1137 that the presence of hominins in a wide range of environments in Eurasia supports this 1138 hypothesis, however, this explanation is valid for the "longue durée" when a period of 1139 one million years or more is considered. In addition, hypothesizing what may constitute 1140 the pre-adaptability traits is rather intriguing. These may include an increase of resource 1141 exploitation by using of stone tools, wooden tools that did not survive, and fire. For the 1142 latter there is hardly any solid evidence prior to 0. and gatherers when compared to their ancestors, and these capacities required larger 1150 territories (Walker and Shipman, 1996). Specifically, a high proportion of meat in the diet 1151 was seen as a critical for the success of hominins and this proposition was supported, according to various authors, by the taphonomic analyses of faunal remains 1153 (Brantingham, 1998 In sum, this paper provides an overview of the environmental conditions that 1160 prevailed during the Lower and Middle Pleistocene in southwestern Asia indicating that 1161 regional topographic configuration played a primary role in shaping the effects of 1162 climatic amplitude on the biotic responses of flora and fauna. The emphasis on climatic 1163 shifts and regional ecological variability is commonly seen as the sound background for 1164 the archaeological contexts and therefore facilitating the understanding hominin 1165 population dynamics. Hominins were able to disperse into regions that were beyond the 1166 their sub-tropical and tropical African homeland (Bosinski, 2006;Dennell 2009) and in 1167 routes, times and into environments unrelated to other large African taxa both carnivore 1168 and herbivore suggesting a unique biological, behavioral and cultural suite of characters 1169 which allowed them to do so. However, without a better understanding the reasons for 1170 success and failure of survival of various hominin migrant groups, and whether it 1171 depended on their ability to keep their mating and reproductive systems, we will need to 1172 resort to oversimplifications of the "human success".