Exploitation of Upland and Lowland Environments: Early Bronze and Iron Age Sites from Eastern Turkey

Sarah Howell-Meurs (University of Melbourne, Australia)


Eastern Anatolia is characterised by a climatic and topographic dichotomy dictated by the presence of the Anatolian plateau and highlands to the north and semi-arid lowlands to the south. Early Bronze Age settlement patterns and architectural evidence broadly support this notion of a dichotomy, with the northern provinces such as Bayburt and Erzurum being characterised by low-density occupation consisting of small villages, in contrast to a tendency toward higher-density occupation which included medium- and large-sized towns and administrative centres characteristic of the southeastern provinces of Urfa and Malatya (Yakar 1985). While obvious contrasts in the exploitation strategies employed at upland and lowland sites have been noted for early pastoral economies of south eastern Europe (Greenfield 1991), little research examines whether altitude and consequently topography and climate exerted a significant impact over the economic systems of eastern Anatolia. Given the complex interrelationship between settlement, environment and subsistence, an examination of the economic strategies employed at upland and lowland sites throughout eastern Anatolia highlights the extent to which topographical and climatic factors influenced economic patterns. While the reconstruction of animal resource exploitation patterns through the analysis of faunal remains has been relatively intense in the southeast, evidence has until lately been entirely lacking from the northeastern regions. Recent analysis of the faunal assemblages from the northeastern Anatolian site of Sos Höyük and Büyüktepe Höyük (Howell-Meurs 1998), however, permits an examination of the effects of topography and climate over animal resource use in the upland compared with the lowland sites of eastern Anatolia during the Early Bronze Age period.

Investigation and comparison of the economic and animal resource-use patterns of upland and lowland sites in eastern Anatolia is hampered by a poor representation of sites, and a lack of published analyses of faunal remains from excavated sites. Of those sites with detailed data relating to pastoral production and exploitation of wild animals, many are characterised by small assemblage size or a lack of correlation and thus comparability between the analytical methods chosen to investigate assemblage characteristics. The sites used in the current analysis were chosen on the basis of the accessibility of detailed data and use of generally comparable methods of assemblage analysis in the reports. Sos Höyük and Büyüktepe Höyük are located in the provinces of Erzurum and Gümüşhane, and lie within the Anatolian plateau at altitudes of approximately 1800m and 1500m above sea level respectively. These sites comprised small villages in the Early Bronze period. The lowland sites considered in the current study, Hassek Höyük (Stahl 1989) and Lidar Höyük (Kussinger 1988), and Gritille (Stein 1988), located in the Urfa and Adiyaman provinces respectively, are all located in the Karababa Basin at an altitude of approximately 400m above sea level. Greater diversity in site function is apparent amongst these lowland sites with Hassek Höyük functioning as a large urban centre, Lidar Höyük as an administrative district centre, and Gritille as a small-scale village settlement (Yakar 1985, Stein 1988). A further site, Korucutepe (Boessneck and von den Driesch 1975), located on the Altinova plain in the Elazig province at approximately 800m above sea level, is intermediate in altitude between those of the upland and lowland sites and comprised a large urban centre. This site was examined to investigate whether it displayed pronounced pastoral and economic affinities with the economies of either upland and lowland environments.

All of the sites considered in the current study are situated on plains bordered by mountain ranges, although the climate and surrounding vegetation are markedly different between the upland and lowland sites. The high-altitude sites are subject to a continental climate and dominant summer drought system, with low annual precipitation of approximately 400mm per year, short summers and extremely long and harsh winters with daily maximum temperatures in January consistently below -15oC (Alex 1985a, 1985b, Alex 1983). This is in contrast to the semi-arid climate of the Karababa Basin in which the annual precipitation of 400-600mm is concentrated during winter, and average temperatures reach between 0-5oC in winter and 30-35oC in summer (Stein 1988). The climate in the Altinova Plain is intermediate between the upland and lowland regions, consisting of a summer drought with an annual precipitation of 400mm, and average summer and winter temperatures of approximately 19 oC and 7 oC respectively (Alex 1985c:105). In terms of vegetation, the regions surrounding Sos Höyük and Büyüktepe Höyük are characterised by steppe. By contrast, the sites in the Karababa Basin, located at the junction between the Anatolian highlands and North Syrian plain, exhibit three vegetation types, being mixed broad-leaved and needle-leaved woodland, Irano-Turanian steppe and desert vegetation (Stein 1988). The region surrounding Korucutepe is characterised by natural vegetation of broad-leaved and needle-leaved woodland (Zeist and Bottema 1991). The vastly differing climatic and topographical environments of the upland and lowland sites might be expected to impose significant constraints over the activities of the inhabiting human populations and by extension their systems of pastoral production.

The relative representation of wild to domestic taxa within each assemblage was examined to determine if any differences were apparent between the exploitation patterns of highland and lowland sites (Figure 1).


Figure 1. The relative abundance of wild to domestic taxa in terms of NISP for Sos Hoyuk (N=2477), Buyuptepe (N=34), Korucutepe (N=866), Lidar Hoyuk (N=2989), Hassek Hoyuk (N=12962) and Gritille (N=1205).


Examination of the results reveals a consistently low-level exploitation of wild resources, with all assemblages providing clear evidence of an overwhelming emphasis on domestic taxa for subsistence needs. Further, the pattern of wild animal use appears to be consistent across assemblages, with a relatively broad suite of species each contributing relatively few specimens (Howell-Meurs 1998, Kussinger 1999, Stahl 1989, Stein 1988). Some exception to this is provided by the assemblage from Korucutepe, in which the hunting of red deer figures more prominently than is apparent elsewhere, suggesting a greater abundance and proximity of woodlands to this site (Boessneck and von den Driesch 1975 124). The likelihood of detecting rare taxa in an assemblage increases with, and is thus dependent on sample size. It is therefore possible that a number of the Early Bronze Age assemblages may reflect an under-representation of wild relative to domestic taxa. Despite this, an increase in the representation of wild taxa for the smaller-sized assemblages would not alter the fact that exploitation of wild resources between upland and lowland sites was reasonably consistent. The low level of exploitation of wild resources may have resulted from a variety of factors that were consistent throughout differing altitudes in eastern Anatolia during the Early Bronze Age. These may have included a reduction in the proximity of habitat suitable for some wild species through processes including increased agricultural intensification involving land clearance, and deforestation for wood resources for building and fuel, with the result that many wild taxa were confined to more remote and perhaps inaccessible environments. Similarly the continued emphasis upon pastoralism evidenced in these assemblages may have provided neither the labour nor the impetus for the large scale exploitation of wild resources.

Trends in the abundance of the main domesticates for each assemblage may provide insights into altitudinal differences in the focus of pastoral strategies (Figure 2).


Figure 2. The relative abundance of the main domesticates in terms of NISP for Sos Hoyuk (N=2362), Buyuptepe (N=29), Korucutepe (N=721), Lidar Hoyuk (N=2791), Hassek Hoyuk (N=12709) and Gritille (N=1133).


With the exception of Hassek Höyük, sheep and goats emerge as the most abundant taxa in each assemblage. Although this predominance varies significantly from between 52% to 69%, it does not appear to be correlated to site altitude. Sos Höyük and Büyüktepe Höyük reflect a similar abundance of ovicaprids to Korucutepe and the lowland sites. The favouring of ovicaprids as the principal herded domesticate at both upland and lowland sites may have occurred for two reasons: these taxa are able to adapt to a wide diversity of habitats and furnish multiple products. The suitability of ovicaprids to marginal environments including steep and mountainous terrain and their ability to graze very low vegetation, allows for the utilisation of the hilly lands adjacent to the plains in which each of the sites is located. As neither cattle nor pigs can effectively utilise these regions, the keeping of ovicaprids allows for the exploitation of what would be an otherwise under-utilized resource. Sheep and goats may also yield a variety of products including milk, wool, hair, meat, hides and horn. An emphasis on the herding of ovicaprids could therefore provide a wide array of returns. These features may have made ovicaprids equally desirable and versatile as stock for both upland and lowland pastoralists.

Sites from the lowland semi-arid ecosystems, however, show a consistently lower abundance of cattle, relative to sheep and goats, than is apparent at the upland sites. Korucutepe appears to indicate an intermediate level of cattle abundance, although tending more toward the lesser relative representation apparent at the lowland sites. The lower representation of cattle within the lowland environments accords well with the fact that cattle have a lower tolerance for semi-arid conditions (Spooner 1973:8), and thus will tend to occupy a less significant role in the subsistence strategies of herders occupying such areas. This is in part owing to the high water requirements of domestic cattle, which increase significantly with higher ambient temperatures (Wells 1970:122). The higher altitude and lower temperatures of the highlands surrounding both Sos Höyük and Büyüktepe Höyük may have thus favoured the herding of a greater proportion of cattle than at sites in the lowland, semi-arid environments. Cattle in addition require substantial quantities of forage and pasture. The location of both the upland and lowland sites on plains adjacent to mountain ranges would however have limited the grazing land available for cattle, as the species is best suited to flat ground or land with only low undulations. The subsequent confinement of this species principally to the plains would have resulted in it being in direct competition with arable lands. The lesser relative abundance of cattle at lowland sites may therefore indicate the lesser availability of land for cattle to graze through a more extensive use of plains for agriculture perhaps coupled with the apparently higher levels of settlement density in these regions.

With pig abundance ranging from between one to twenty percent within the majority of Early Bronze Age assemblages, a clear neglect of this species as a major food source is evident, irrespective of altitude. The relatively low representation of pigs is however a trend apparent throughout the Near East for many millennia following their initial domestication (Zeder 1996:298). The consistently low representation in the upland and lowland sites suggests significant pressures discouraging the widespread large scale herding of pigs during the Early Bronze Age period. Pigs require shelter, typically in the form of vegetation, from both sun and extremes of weather and also need a reliable water source close by, soft ground and, in harsh sun, mud wallows. With a dietary preference for acorns and beech-mast, their ideal habitat comprises moist, open woodland (Diener and Robkin 1978, Grigson 1982:300). Pigs have, in addition, much higher water requirements than do the other main domesticates, with daily watering of three parts water to one part feed necessary. Even higher levels are required by pregnant sows (Zeder 1996, 301). Their low abundance within most assemblages suggests that these sites were lacking in one or more of these characteristics. As all sites were located adjacent to perennial water sources, the low abundance of pigs may be related to a lack of settlement proximity to woodland and vegetation cover as a source of food and shelter, a conclusion also suggested by the low representation of wild resources at each site. A clear exception to this pattern is provided by the assemblage from Hassek Höyük, in which pigs comprised over 50% of the main domesticate bones identified. The proximity and environmental comparability of this site to the other lowland sites analysed suggests that cultural rather than ecological factors may be dictating the nature of pig exploitation.

The nature of pastoral strategies can most readily be investigated through analysis of the age and sex data relating to the main domesticates. Patterns of eruption and attrition on the mandibular remains in addition to the state of epiphyseal fusion of the long bones all provide indicators of mortality. Sexual dimorphism in the size and morphology of various skeletal elements including the metapodials, horn cores and pelvis may permit for differentiation between male and female specimens. This allows for the reconstruction of survivability profiles for the herd in which peaks in mortality of different age classes and between males and females can be highlighted and interpreted with reference to models of idealised herd management for specific products such as milk, wool, traction or meat (Payne 1973, Higham and Message 1970). The analysis was hampered, however, by a lack of data or poor sample size. As a results, herd management regimes could only be investigated for the ovicaprid and cattle remains from a selection of sites.

Mortality profiles for the ovicaprid assemblages from upland and lowland Early Bronze Age sites appear to be comparable in form (Figure 3).


Figure 3. Survivorship curve for the ovicaprid remains for Sos Hoyuk (N=58), Korucutepe (N=16), Lidar Hoyuk (N=53), Hassek Hoyuk (N=90) and Gritille (N=32).


In each case, juvenile mortality is poorly represented, while the highest mortality is apparent amongst the subadult age classes. Although lacking for the other assemblages, epiphyseal fusion data from Sos Höyük and Gritille support the profiles suggested by the dental remains (Table 1). With the exception of Gritille, the assemblages each indicated a predominance of adult females (Boessneck and von den Driesch 1975:68, Howell-Meurs 1998:113, Kussinger 1988:59, Stahl 1989:72, Stein 1988:188). With high subadult mortality and a predominance of adult females, these profiles conform most readily to a meat production strategy (Payne 1973), although this may have been supplemented by the exploitation of secondary products from adult breeding stock. The assemblage from Korucutepe, with a later peak in mortality at two to four years, provided the only evidence of a divergence from the mortality patterns apparent for the other Early Bronze Age assemblages, reflecting a greater emphasis on wool production (Boessneck and von den Driesch 1975 38).


os coxae134
p. radius34280
d. humerus314123
12-24p. phalanx 129910
p. phalanx 211120
d tibia3010142
d metapodials3223145
30-36p ulna4303
p femur61452
36-48d humerus0512
p tibia3503
d radius61311
d femur3611

Table 1. Epiphyseal fusion data for ovicaprid remains from Sos Huyuk and Gritelle


Evidence for the construction of survivability profiles for the cattle remains from Early Bronze Age sites is generally lacking. Only three assemblages, Sos Höyük, Lidar Höyük and Hassek Höyük, allow for the determination of survivability curves based upon mandibular eruption and attrition. The resulting profiles suggest extremely low mortality amongst juveniles and subadults prior to two and a half years (Figure 4).


Figure 4. Survivorship curve for the cattle remains for Sos Huyuk (N=25), Lidar Hoyuk (N=20) and Hassek Hoyuk (N=21).


Epiphyseal fusion data from Sos Höyük suggests a peak in mortality between two and a half and four years, although the majority of animals nevertheless survived beyond four years into full adulthood (Table 2). Females outnumber males amongst the adults within each of these assemblages (Howell-Meurs 1998:86, Kussinger 1988:21, Stahl 1989:22). These results suggest a mixed strategy in which some males were culled for meat prior to four years, with the remainder retained for traction work, while adult females provided milk resources and served as breeding stock (Stahl 1989:15, Kussinger 1988:19). Epiphyseal fusion data from Gritille contrasts somewhat with the dental data from Sos Höyük, Lidar Höyük and Hassek Höyük, with a pronounced peak in mortality apparent prior to four years, suggesting a strategy more specifically geared toward meat production (Stein 1988:222). The lack of dental data and specimens identified to gender within this assemblage precluded further investigation of this trend. Skeletal-part representation analysis within each assemblage again indicates the presence of entire carcasses at the settlements suggesting local production and consumption rather than exchange in each case.


Age (Months)Elementfusedunfused
7-10os coxae160
12-20p. phalanx 1411
p. phalanx 2341
p radius250
d humerus253
24-30d tibia153
d metapodials272
42-48p humerus41
p femur93
p tibia13
d radius72
d femur32

Table 2. Epiphyseal fusion data for cattle remains from Sos Hoyuk.


In contrast to Greenfield’s findings for the eastern Balkans, in which pronounced differences in economic strategies were apparent between upland and lowland sites, the herding strategies and animal use patterns apparent for the eastern Anatolian sites examined here appear to be reasonably consistent. The ovicaprid and cattle mortality data and relative representation of wild to domestic taxa indicate little variation in the nature of animal production strategies and use of wild resources between upland and lowland sites, suggesting that these factors were not largely dependent on site altitude or environment. Some variation was apparent however between upland and lowland sites in terms of relative species abundance of the main domestic taxa and this could be correlated to some extent with climate and topography and thus altitude. Broadly, though, trends were comparable across the assemblages analysed, irrespective of site altitude. While animal use patterns do not appear to correlate obviously with site altitude, various differences in exploitation strategies such as the apparent focus on wool resources at Korucutepe may in part be related to site function. Ethnographic analogues, for instance, imply that smaller-scale settlements such as those represented at Sos Höyük, Büyüktepe Höyük and Gritille may have been able to meet their wool and milk requirements without structuring herd productivity exclusively toward these goals (Redding 1981:48). Twentieth-century Lur nomads, who raise flocks of ovicaprids primarily for subsistence, with fibre constituting a by-product, obtain enough wool and goat hair from their stock to supply their own needs plus furnish a surplus that is sold either to itinerant dealers or in the local townships (Mortensen 1993:279). Similarly, Black-Michaud (1986:43) provides an account of the regime undertaken by Lur nomads to permit simultaneous use of sheep milk resources by offspring and humans including the restriction of suckling time and milking prior to suckling. The wool profile apparent at the large urban centre of Early Bronze Age Korucutepe may therefore indicate larger scale production for local consumption and perhaps to create a surplus for exchange. Why this pattern was apparent at Korucutepe but not at the comparably-sized settlements of Hassek Höyük and Lidar Höyük requires further investigation.

The small number of sites examined and lack of detailed data for various aspects of the analysed assemblages significantly limited the current investigation. Future analyses of assemblages, particularly from highland contexts, and the availability of detailed data from a wider geographic range of upland and lowland habitats would serve to clarify significantly the influence of altitude over economic strategies and allow for assessment of the pervasiveness of the trends detected in the current study.



Alex, M. 1983 Vorderer Orient. Mittelere Januartemperaturen. 1:8,000,000. Karte A IV 2. Wiesbaden: Dr. Ludwig Reichert Verlag.

Alex, M. 1985a Vorderer Orient. Niederschlagsverläßigkeit. 1:12,000,000. Karte A IV 5. Wiesbaden: Dr. Ludwig Reichert Verlag.

Alex, M. 1985b Vorderer Orient. Jahresgangtypen des Niederschlags 1:12,000,000. Karte A IV 6. Wiesbaden: Dr. Ludwig Reichert Verlag.

Alex, M. 1985c Klimadaten ausgewahlter Stationen des Vorderen Orients. Wiesbaden: Dr. Ludwig Reichert Verlag,.

Black-Michaud, J. 1974 An Ethnographic and Ecological Survey of Luristan, Western Persia: Modernisation in a Nomadic Pastoral Society. Middle East Studies 10:210-28.

Boessneck, J. and A. von den Driesch 1975 Tierknochenfunde vom Korucutepe bei Elazig in Ostanatolien. In van Loon, M. (ed.) Korucutepe: Final Report on the Excavations of the Universities of Chicago, California (Los Angeles) and Amsterdam in the Keban Reservoir, Eastern Anatolia, 1968-1970, Vol. 1, pp. 1-191. Amsterdam: North Holland Publishing Company.

Diener, P. and E. Robkin 1978 Ecology and Evolution and the Search for Cultural Origins: The Question of Islamic Pig Prohibition. Current Anthropology 19:493-540.

Greenfield, H. 1991 Fauna from the Late Neolithic of the Central Balkans: Issues in Subsistence and Land Use. Journal of Field Archaeology 18:161-186.

Grigson, C. 1982 Porridge and Pannage: Pig Husbandry in Neolithic England. In Limbrey, S. and M. Bell (eds) Archaeological Aspects of Woodland Ecology, pp. 297-304. International Series 146. Oxford: BAR

Higham, C. and M. Message 1970 An Assessment of a Prehistoric Technique of Bovine Husbandry. In D. and E. Higgs (eds) Science in Archaeology (rev. ed.), pp. 315-330. Praeger, New York.

Howell-Meurs, S. 1998 Early Bronze and Iron Age Animal Exploitation in Northeastern Anatolia: The Faunal Remains from Sos Höyük and Büyüktepe Höyük. Unpublished Ph.D. thesis, The University of Melbourne.

Kussinger, S. 1988 Tierknochenfunde vom Lidar Höyük in Südostanatolien (Grabungen 1979-86). Unpublished thesis, Ludwig-Maximillians-Universität, München.

Mortensen, I. 1993 Nomads of Luristan: History, Material Culture and Pastoralism in Western Iran. London: Thames and Hudson.

Payne, S. 1973 Kill-off patterns in sheep and goats; the mandibles from Aşvan Kale. Anatolian Studies 23:281-303.

Redding, R. 1981 Decision Making in Subsistence Herding of Sheep and Goats in the Middle East. Unpublished Ph.D. Dissertation, University of Michigan.

Spooner, B. 1973 The Cultural Ecology of Pastoral Nomads. Module in Anthropology 45. Reading (Mass.):Addison-Wesley

Stahl, U. 1989 Tierknochenfunde vom Hassek Höyük (Südostanatolien). Unpublished thesis, Ludwig-Maximillians-Universität, München.

Stein, G. 1988 Pastoral Production in Complex Societies: Mid-late Third Millennium B.C. and Medieval Faunal Remains from Gritille Höyük in the Karababa Basin. Unpublished Ph.D. thesis, University of Pennsylvania.

Wells, D. 1970 The water requirements of particular stock producing systems. In A. Taylor, (ed.) The Role of Water in Agriculture, pp. 121-132. Oxford: Pergamon Press.

Yakar, J. 1985 The Later Prehistory of Anatolia: The Late Chalcolithic and Early Bronze Age. International Series 268. Oxford: BAR.

Zedder, M. 1996 The Role of Pigs in the Near East. In Seger, J. (ed.) Retrieving the Past: Essays on Archaeological Research and Methodology in Honour of Gus W. van Beek, pp. 297-312. Winona Lake (Indiana): Mississippi State University.

Zeist, W. van and S. Bottema 1991 Late Quaternary Vegetation of the Near East. Beihefte zum Tübinger Atlas. Tübingen: Luwig Reichert Verlag.