The Initial Domestication of Goats (Capra hircus) in the Zagros Mountains 10,000 Years Ago

See allHide authors and affiliations

Science  24 Mar 2000:
Vol. 287, Issue 5461, pp. 2254-2257
DOI: 10.1126/science.287.5461.2254


Initial goat domestication is documented in the highlands of western Iran at 10,000 calibrated calendar years ago. Metrical analyses of patterns of sexual dimorphism in modern wild goat skeletons (Capra hircus aegagrus) allow sex-specific age curves to be computed for archaeofaunal assemblages. A distinct shift to selective harvesting of subadult males marks initial human management and the transition from hunting to herding of the species. Direct accelerator mass spectrometry radiocarbon dates on skeletal elements provide a tight temporal context for the transition.

The fertile crescent region of the Near East was the center of domestication for a remarkable array of today's primary agricultural crops and livestock animals. Wheat, barley, rye, lentils, sheep, goats, and pigs were all originally brought under human control in the broad arc that stretches from the southern Levant through southeastern Turkey and northern Syria, to the high Zagros mountain pastures and arid lowland plains of Iraq and Iran. For more than 50 years researchers have sought to define the sequence, temporal placement, and social and environmental context of domestication (1). Central to addressing this process is the ability to identify early domesticates in the archaeological record, and to place them within a secure temporal context. Here we describe recent research that uses a study of modern wild goats (Capra hircus aegagrus) to develop an unequivocal marker of early goat domestication, which we apply to assemblages that lie both within and outside the natural range of wild goats in the eastern fertile crescent region—a region long thought to be the initial heartland of goat domestication (2).

Two markers have been used to identify domestication in goats. It has been proposed that changes in skeletal morphology, particularly population-wide reduction in body size, quickly follow human controlled breeding (3, 4). Uncertainty over the pace and causal connection between body size reduction and human control limits the utility of this marker, however, as does difficulty in distinguishing between human-induced changes and those resulting from other biological or environmental factors (5). Alternatively, changes in age and sex profiles that resulted from controlled breeding and selective harvesting of young males have been proposed as providing a better early marker of domestication (6). Difficulty in distinguishing between various selective hunting strategies and those that reflect deliberate herd management and domestication has been the main perceived limitation of this demographic profiling approach (4). Until now a key obstacle has been an inability to construct the separate male and female age profiles necessary to detect the distinctive sex-specific harvest patterns of managed herds.

Recent metrical analysis of a skeletal collection of 37 modern wild goats, curated by the Field Museum of Natural History, from different habitats in Iran and Iraq provides an empirical basis for assessing the utility of these different markers of early goat domestication (7). In all measurements taken on nine different postcranial skeletal elements, sex was the single most significant factor influencing size (Fig. 1). Even the unfused bones of young males older than 1 year were absolutely larger than the fully fused bones of older, adult females. Environment also influenced body size, with a clear pattern of size reduction from the cooler, wetter regions in northern Iran and northwestern Iraq (Fig. 1A) to the arid, hotter, and sparser pastures of southern Iran (Fig. 1C). While the available sample of domesticated goats (C. h. hircus) from the region was too small to draw firm conclusions, domestic status does not appear to be a major factor influencing size in this modern sample, especially in female goats (Fig. 1, A and B).

Figure 1

Greatest breadth of distal metacarpal at epiphysis of modern goats older than 1 year of age (red, fused females; pink, unfused females; blue, fused males; light blue, unfused males; light blue triangle, male mean; pink triangle, female mean; D, domestic). (A) Northern Zagros: Caspian and Azerbaijan regions of Iran and Iraqi Kurdistan (individual animals = 17, specimens measured = 32). (B) Central Zagros: Kermanshah and Isfahan regions of Iran (individual animals = 12, specimens measured = 22). (C) Southern Zagros: southern Fars, Iran (individual animals = 10, specimens measured = 18).

This modern reference class of goat skeletons provides a baseline of comparison for analysis of goat bone assemblages from a number of archaeological sites, also curated by the Field Museum and by the Smithsonian's National Museum of Natural History, that bracket the transition from hunting to early herd management. This sample includes two sites widely accepted as providing the earliest evidence of domesticated goats—Ganj Dareh, an upland settlement located within the present-day geographical range of wild goats (8), and Ali Kosh, located in a lowland environment well outside the natural habitat of wild goats (9). In his initial study of the Ganj Dareh assemblage, Hesse proposed that domestic goats were present at the site on the basis of a demographic profile that indicated herd management (8). Subsequent researchers also concluded that the Ganj Dareh goats were domesticated, claiming these animals were considerably smaller than wild bench-mark populations from Late Pleistocene/Early Holocene contexts (3, 4, 10). Flannery's identification of domesticated goats at Ali Kosh was based on a high percentage of young animals, on a gradual change in horn form, and on the abundance of goats despite the site's location (9).

A reanalysis of the Ganj Dareh goats by Zeder (7) contradicts claims for a marked diminution in their size relative to wild populations, while supporting Hesse's original conclusions. In all postcranial elements examined in this analysis, there is an almost complete overlap in the size range, mean, and medians of the dimensions of the Ganj Dareh goat bones (Fig. 2C) when compared to wild goats from earlier nearby upland Zagros sites dating to the Middle and Upper Paleolithic (Fig. 2A). There is also no observed size difference between the Ganj Dareh goats and those from the nearby site of Asiab, a likely hunting camp occupied slightly earlier than Ganj Dareh (Fig. 2B) (3, 9, 10).

Figure 2

Length of second phalanx of goats from Zagros sites (shaded, fused bones; unshaded, unfused and fusing bones; triangle, mean). (A) Paleolithic sites: Palegawra, Kunji, Kobeh, and Yafteh Caves, ∼50,000 to 15,000 14C yr B.P. (n = 23). (B) Asiab, ∼10,00014C yr B.P. (n = 6). (C) Ganj Dareh, ∼8900 14C yr B.P. (n = 135). (D) Ali Kosh, ∼8000 to 8400 14C yr B.P. (n = 47).

The apparent population-wide size reduction observed in the Ganj Dareh goats (3, 4, 10) is, in fact, an artifact of a long-standing practice of excluding the unfused bones of young animals from metric analyses. Very few of the large (male) Ganj Dareh goats survived to adulthood, and the exclusion of their unfused skeletal elements from size analysis skewed the resultant size profiles toward smaller adult females. When compared to goats in a hunted population, in which adult males are likely to be well represented, the Ganj Dareh assemblage can appear to indicate an overall reduction in body size. Methods that normalize the dimensions of a number of different skeletal elements into standardized species size profiles (3,4) tend to exacerbate this bias, while combining assemblages from several different regions across the Near East to form broad temporal samples for comparison (3) further obscures actual patterns of size variation. When all bones, fused and unfused, are measured, and only direct element-by-element comparisons are made of assemblages from the highland Zagros region, the apparent differences in the size of these wild and domestic goat populations disappears.

There is a marked difference in the dimensions of postcranial elements of goats from all of these upland sites, however, when compared to the goats from the lowland settlement Ali Kosh (Fig. 2D). Yet the size difference of ancient goats from these different environmental zones is no greater than that noted between the modern wild goats from northern cooler and wetter upland areas and those from more arid southern regions (Fig. 1) (7). It is difficult to determine, then, whether the smaller size of the Ali Kosh goats is attributable to selective factors that favor smaller animals in managed herds, the response of colonizing populations of domesticated goats to harsher environmental conditions, or the original smaller size of a progenitor stock of wild goats.

At the same time that this study shows that size reduction is not a good marker of early domestication, it also confirms and extends Hesse's original conclusion that the Ganj Dareh goats were domesticated on the basis of demographic profiling (6, 8). The marked bimodality in the size distribution of the postcranial skeletal elements noted by Hesse in his initial study is clearly evident in essentially all dimensions of postcranial skeletal elements measured in this analysis (Fig. 3). With the benefit of Zeder's recent study of size variation in modern goats (7), this bimodality can be securely attributed to the marked dimorphism in the size of male and female goats. Moreover, the study of the modern collections provides solid empirical support for Hesse's conclusion that the disproportionate number of unfused elements of large animals can be attributed to the early slaughter of young males. Finally, the strong regularity of the proportional relation between the size of male and female animals in the modern sample allows the separation of the ancient goat assemblage into male and female subpopulations. Sex-specific age profiles can then be constructed for each of the different size-separate subpopulations.

Figure 3

Ganj Dareh postcranial bone dimensions (shaded, fused bones; unshaded, unfused and fusing bones). (A) Breadth of distal radius (fusion ∼36 months, n = 186). (B) Breadth of distal metacarpal (fusion ∼24 months,n = 165). (C) Length of first phalanx (fusion ∼16 months, n = 83). (D) Breadth of distal humerus (fusion ∼10 months, n = 292).

When this is done for the Ganj Dareh assemblage (Fig. 4A), there is a clear pattern of selective kill-off of subadult males and delayed slaughter of females that matches culling patterns in modern domesticated herds managed for meat (11). In contrast, the sex-specific age profiles of goats from nearby Asiab (Fig. 4B) are more consistent with a hunted population in which there is a focus on fully adult males, with females and young taken when encountered. Preliminary examination of the size and fusion patterns of the Ali Kosh goats also shows the distinctive demographic profile of a managed herd, although both subadult males and adult females seem to have been slaughtered at somewhat older ages than the Ganj Dareh goats.

Figure 4

Sex-specific survivorship curves, displayed as a percentage of animals living beyond a certain age for each fusing element (red, females; blue, males). (A) Ganj Dareh (females = 783, males = 367). (B) Asiab (females = 19, males = 24).

Previous efforts to date these sites have been problematic (12). An early date for the initial occupation of Ganj Dareh suggested that the settlement was established about 11,000 years ago, whereas four subsequent levels were dated to about 9500 years ago. At Ali Kosh, a number of laboratories, using different extraction techniques on different types of organic materials, have produced several nonconforming series of dates. On the basis of site stratigraphy, material culture, and some of the dates obtained, excavators identified three different phases of occupation, from 9500 to 7500 years ago (10).

Accelerator mass spectrometry (AMS) dating of four samples of charred barley at Ganj Dareh (13), and 12 bone collagen samples that span all five levels at the site obtained as part of this study (Table 1), clearly indicate that Ganj Dareh was occupied briefly, for perhaps no more than 100 to 200 years, at about 10,000 calendar years before the present (yr B.P.). While the seven new AMS dates obtained for Ali Kosh provide the most consistent age determinations obtained for this site, there are still some internal inconsistencies, possibly stemming from the different material dated—collagen and carbon extracted from charred bone. It is clear, however, that the occupation of Ali Kosh was much briefer than previously thought, perhaps lasting only 500 years. Moreover, whether the date of initial occupation was 8900 calendar yr B.P. (14), as indicated by the date derived from bone collagen, or 9500 calendar yr B.P., as indicated by the earliest carbonized bone date, these new dates confirm that the site is chronologically more recent than Ganj Dareh by at least 500 years, and perhaps by as much as a millennium.

Table 1

AMS 14C dates on animal bones from Ganj Dareh (GD) and Ali Kosh (AK).

View this table:

By resolving the temporal placement of these two sites, it is possible to place the initial domestication of goats in the highland Zagros region well within the natural habitat of this species. The long-term concentration on the hunting of wild goats in this upland region seen in assemblages stretching as far back as the Middle Paleolithic (2) suggests that the eventual domestication of goats grew out of a long-term process in which hunting strategies evolved into the actual management of captive animals. Ganj Dareh represents an early stage in herd management, in which the feeding, movement, and other environmental circumstances of these animals was little changed. Although the Ganj Dareh goats may not yet be morphologically, or even genetically, distinguishable from wild goats, however, the distinctive profile of young male slaughter and prolonged female survivorship documented here marks the goats of Ganj Dareh as a managed, and therefore domesticated, population.

The subsequent establishment some 500 to 1000 years later of settlements like Ali Kosh, outside this natural habitat zone, represents a break with the environmental, biological, and social context of initial domestication. This move seems to have necessitated some adjustments in management strategies from those developed in the upland zone. It also allowed the expression of morphological traits not widely seen in either wild or domesticated goats in upland habitats, such as the gradual alteration in the shape of horns identified by Flannery and, possibly, the reduction in the size of the Ali Kosh goats seen in this study.

Such morphological changes, if related to the process of domestication at all, are only delayed, and possibly indirect, artifacts of human management. Instead, it is the transformation of strategies that seek to maximize offtake from wild herds into those that seek to control the productivity of captive herds that lies at the heart of the process of animal domestication. The method for computing sex-specific age profiles developed here provides the most sensitive tool to date for monitoring this transformation and should be applicable not only to goats, but to any sexually dimorphic species. Combined with high-precision small-sample AMS dating, it promises a much finer understanding of the timing and trajectory of the origins of animal domestication across the Near East and beyond.

  • * To whom correspondence should be addressed. E-mail: zeder.melinda{at}


View Abstract

Stay Connected to Science

Navigate This Article