8 pages ANTH 165 paper

University of Pennsylvania Museum of Archaeology and Anthropology

Chapter Title: Areas of Origin of the Crops and Domestic Animals Book Title: Origins of Agriculture in Western Central Asia

Book Subtitle: An Environmental-Archaeological Study

Book Author(s): David R. Harris

Published by: University of Pennsylvania Press; University of Pennsylvania Museum of Archaeology and Anthropology

Stable URL: https://www.jstor.org/stable/j.ctt3fj6gz.14

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

University of Pennsylvania Press and University of Pennsylvania Museum of Archaeology and Anthropology are collaborating with JSTOR to digitize, preserve and extend access to Origins of Agriculture in Western Central Asia

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

part iii

Neolithic Crop Plants and Domestic Animals

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

7

Areas of Origin of the Crops and Domestic Animals

A fundamental part of this enquiry into the transi-tion from foraging to farming in western Central Asia is to try to determine whether agriculture began there independently, or whether the new way of life was introduced—wholly or in part—from elsewhere. One approach to this question is to ask whether any of the crops and domestic animals identified in the or- ganic remains recovered at Jeitun and other early sites could have been domesticated locally from wild pro- genitors native to the region. To answer that question requires not only knowledge of the present physical environment and of environmental changes during the Pleistocene and Holocene (outlined in Chapters 1 and 2), but also familiarity with genetic studies of the species concerned and with archaeological evidence of their presence in the Neolithic.

In recent years cytological, molecular, and ar- chaeological investigation of the evolution and do- mestication of many crops and domestic animals has advanced rapidly, with the result that earlier assump- tions about their ancestry and areas of origin are being radically re-examined (Brown et al. 2009; Bru- ford, Bradley, and Luikart 2003; Burger, Chapman, and Burke 2008; Doebley, Grant, and Smith 2006; Fuller 2007; Jones and Brown 2000; Zeder et al. 2006). In particular, analyses of modern and ancient DNA of several cereals and domestic herd animals that were an integral part of Neolithic agriculture in Eurasia— barley, wheat, goats, sheep, cattle, and pigs—suggest that they may have been domesticated several times in different areas (see below).

There is also a growing realization that short- term climatic and biotic changes during the Qua- ternary were more frequent and extreme, and more

profoundly affected the ranges of wild plants and animals, than previously thought. As a result, it has become more difficult to make sound inferences, from the present geographical distributions of their wild progenitors and other closely related (congeneric) wild species, about where particular crops and domes- tic animals originated. In this chapter, the principal crops and domestic animals that were present in Turk- menistan and adjacent areas during the Neolithic are reviewed, in order to assess whether they are likely to have been domesticated in the region or introduced, as domesticates, from elsewhere.

The Crops and Their Wild Progenitors

Very little reliably identified and dated archae- obotanical evidence of domesticated plants from prehistoric sites has so far been obtained in western Central Asia. No systematic retrieval of macroscopic plant remains by flotation and/or fine sieving was undertaken in Turkmenistan until we carried it out at Jeitun between 1989 and 1997, and Naomi Miller (1999, 2003) did so at Anau North. Lisitsina (1978) had previously studied cereal grains and impressions in pottery from four Neolithic sites in southern Turk- menistan, but her identifications were insecure. Con- sequently only the crops we have definitively identified and dated from our excavations at Jeitun are discussed in this section.

The cereal remains from Jeitun comprise charred grains and chaff of hulled and naked variet- ies of (probably six-row) cultivated barley (Hordeum

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

74 origins of agriculture in western central asia

vulgare L. [syn. sativum Pers.]) and two types of culti- vated glume (hulled) wheat: einkorn (Triticum mono- coccum L., both the one-grain and the two-grain form) and another (emmer-like) type of uncertain origin (Jones, Valamoti, and Charles 2000), as well as three tentatively identified specimens of free-threshing wheat (of T. aestivum/durum type). Contrary to our earlier reports (Charles and Hillman 1992:83–94; Harris et al. 1993:332; Harris, Gosden, and Charles 1996:438), none of the remains examined can be definitely identified as cultivated emmer wheat (T. dicoccum Schübl.). Nor has any evidence of legume crops been found. These results add greatly to the earlier, less well documented reports of barley and wheat found at Jeitun, Chopan, Bami, and Chagylly (Lisitsina 1978:92; this volume, p. 62); and they raise the question of whether the wild progenitors of barley and the wheats found at Jeitun are likely to have been present in Turkmenistan in the Late Pleistocene and Early Holocene and might, therefore, have been do- mesticated locally.

Barley

Wild barley (Hordeum spontaneum L. subsp. spon- taneum C. Koch, or, according to von Bothmer et al. (1995), H. vulgare L. subsp. spontaneum C. Koch) is a predominantly self-pollinating annual diploid (chromosome number 2n=14) and the progenitor of domestic barley, with which it readily hybridizes and is fully fertile. In wild barley two rows of grain normally develop but under domestication six-row forms have been selected, as have naked-kernel forms (Harlan, de Wet, and Price 1973:317; Pourkheirandish and Ko- matsuda 2007). Wild barley is better adapted to aridity and more tolerant of poor, especially calcareous soils than wild wheats and it occurs more widely around the Southwest Asian Fertile Crescent, as Harlan and Zohary showed in 1966 when they published their well- known distribution maps of the “primary habitats” of wild wheats and barley (updated in Zohary and Hopf 2000:37, 45, 66). Subsequently Zohary (1989:30) com- mented that “more isolated populations, usually of weedy forms” extend east across Central Asia (includ- ing Turkmenistan) as far as the western Himalayas and Tibet (see also Shao 1981; Witcombe 1978), and he later elaborated that comment, stating that “in north-east Iran, Central Asia, and Afghanistan wild spontaneum barley is. . .sporadic in its distribution; it

rarely builds large stands and seems to be completely restricted in most localities to segetal habitats, ruins, or to sites which have been drastically churned by human activity” (Zohary and Hopf 2000:67).

When visiting the Badghyz Natural Reserve in southeastern Turkmenistan in 1992, two members of our research team (Hillman and Harris) observed apparently natural stands of wild barley growing in as- sociation with pistachio trees (Fig. 1.9, color), although it is possible that they are feral populations descended from domesticated barley. This possibility was also envisaged by Jan Valkoun (pers. comm. 1997) who noted that at least some of the “wild” barley popula- tions of northeastern Iran and Turkmenistan have an upright growth habit, and synchronized tillering and seed maturation, implying that they are likely to be descended from weedy forms introduced with already domesticated barley during the spread of Neolithic agriculture. However, these observations do not pre- clude the possibility that genuinely wild barley grew in the intermontane valleys and on the piedmont slopes of the region when agricultural settlements were first established there early in the Neolithic, especially as the climate was warmer and wetter then, during the mid-Holocene Climatic Optimum, than it is today (this volume, pp. 25–26). Hillman had observed wild barley growing as a component of wild almond and pista- chio woodlands in Southwest Asia, and he speculated (1996:188–89) that it might have survived the Late Pleistocene in refugia farther east and subsequently spread from them (as well as from mountain refugia in the northern Levant). Similar woodlands exist in southern Turkmenistan today (this volume, pp. 9–10), and it is quite possible that wild barley grew there in such habitats early in the Holocene.

Genetic data derived from accessions of wild barley from west and east of the Zagros mountains col- lected at a wide range of locations between the Medi- terranean and the Pamirs in Central Asia have shown that western and eastern barleys are part of an inter- breeding population that shares many haplotypes, but that they often differ in the predominant haplotype present, which implies an ancient divergence between west and east and that all the eastern barleys are very unlikely to be feral rather than genuinely wild (Mor- rell, Lundy, and Clegg 2003, Morrell et al. 2005). This conclusion does not deny the possibility that some of the weedy forms referred to by Zohary and Valkoun may have derived from domesticated barley, but it supports the palaeoenvironmental inference that

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

areas of origin of the crops and domestic animals 75

wild barley may have been present in southern Turk- menistan in the Early Holocene. If so, its local domes- tication cannot be excluded on phytogeographical grounds.

Other genetic studies of barley have concluded that it was domesticated only once or very few times. Thus Neale et al. (1988) investigated types of chloro- plast DNA in a large collection of wild and cultivated barleys from Israel and Iran and demonstrated the existence of three lineages in wild barley, only one of which was (with two exceptions) represented in the cultivated barley, implying “a single or very few events of domestication” (Zohary 1999:137). This conclusion was supported by Badr et al. (2000) who analyzed AFLP (amplified-fragment length polymorphism) data from a large sample of wild and cultivated bar- leys and concluded that barley was domesticated only once, in southeastern Turkey. However, although the authors claimed that this result “closes the long- lasting debate on the origin of barley” (ibid., p. 507), it should not be regarded as conclusive because AFLP data analyzed by cluster analysis (neighbor-joining trees) is not a genuinely phylogenetic method of in- ferring monophyletic origins for crops (or domestic animals), as has been demonstrated in simulations of cereal domestication (Allaby and Brown 2003, 2004; Allaby, Fuller, and Brown 2008; Allaby, Brown, and Fuller 2009; and see Salamini et al. 2004 and Honne and Heun 2009).

The possibility that barley was domesticated more than once received further support from Morrell and Clegg’s (2007) comparison of the haplotype com- position of wild and domesticated forms from a wide range of locations across southwestern, central, and eastern Asia. This reinforced their earlier evidence of an ancient divergence between western and eastern wild barleys, and showed that all accessions of culti- vated (landrace) barleys from Asia east of the Zagros have substantial identity with eastern wild barleys and have been subject to later admixture from imported western landraces. This conclusion is supported by a phylogeographic analysis by Saisho and Purugganan (2007) of haplotype clusters of five genes in a large sample of barley cultivars from East and South Asia (including Iran, Afghanistan, and Pakistan), as well as from the Fertile Crescent, North Africa, and Eu- rope, which revealed an area of genetic discontinu- ity between the landraces westward from the Fertile Crescent into Europe and the eastern ones from Iran to China, Korea, and Japan. Both studies imply that

barley was domesticated, independently of its domes- tication in the Fertile Crescent, somewhere east of the Zagros before western cultivars were introduced to central and eastern Asia; but the antiquity of such a putative domestication cannot be demonstrated conclusively from the genetic data alone.

Morrell and Clegg (2007) interpreted archaeo- logical finds of barley in Neolithic contexts at Jeitun, and at Mehrgarh in Baluchistan, as suggestive of areas where a second domestication of barley could have occurred. But at present there is insufficient archaeo- botanical evidence to resolve the question of whether barley was domesticated locally in Turkmenistan and/ or in Pakistan or whether it was initially introduced to those sites as an already domesticated crop. For example, at Mehrgarh there is abundant evidence of barley in the form of plant impressions in mudbrick in the earliest (aceramic Neolithic Period I) levels, which, despite inconsistency in the radiocarbon chro- nology, probably date to between c. 7000 (or perhaps earlier) and c. 6000 cal. BCE (Jarrige 2000:278–83, 2007–08:151; Jarrige, Jarrige, and Quivron 2006). The barley consists mainly of the domesticated six-row hulled and naked forms but there is also some “wild” two-row barley (Costantini 2007–08:168–69; 1983:29– 31; Fuller 2006:22; Possehl 1999:459). The presence of both forms could be interpreted as evidence of initial local domestication, but it does not necessarily support that hypothesis because, as Pourkheirandish and Komatsuda (2007) have shown, six-row spikes and naked kernels (caryopses) were selected after wild two-row barley was domesticated, and furthermore, the apparently wild barley, which typically occurs around Mehrgarh in disturbed habitats, may have reached the region from farther west as a weed of cereal cultivation.

Remains of barley recovered from Southwest Asian Neolithic sites show that its domestication was underway in the Levant and possibly also in the Zagros region by the end of the 9th millennium cal. BCE and that it spread as a domesticated crop throughout the Fertile Crescent during the Pre-Pottery Neolithic period1 (Charles 2007; Garrard 1999:77–79; Willcox 2005:535–38). As the initial domestication of barley took place more than two millennia before it was cultivated at Jeitun, there was ample time for it to be dispersed across northern Iran (together with einkorn wheat; see below) as one of the two founder crops of Neolithic cereal cultivation at the Jeitun- Culture sites—a circumstance that can be interpreted

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

76 origins of agriculture in western central asia

as evidence against an independent domestication of barley in Turkmenistan. However, lack at present of securely identified and well-dated remains of barley (and wheat) from Neolithic sites across northern Iran, from east of the Zagros mountains to the Kopetdag piedmont, means that the question of whether it was introduced to Turkmenistan as a crop from the Fertile Crescent, or locally domesticated, remains open.

There is a similar lack of archaeobotanical evi- dence of barley (and wheat) at sites across Central Asia east of Turkmenistan, which makes it also impos- sible, at present, to do more than speculate about the route(s) by which these cereals first reached China. They were present by c. 2500 cal. BCE in northwest- ern China where seeds of barley and wheat have been found at the site of Xishanping near Tianshui in Gansu and AMS radiocarbon dated to c. 2600 cal. BCE (X. Li et al. 2007). Barley has also been found at Fengtai in the upper Huang-ho valley and wheat at Donghuishan in western Gansu, and both were being cultivated in the middle and lower Huang-ho valley by c. 2000 cal. BCE in the Late Longshan period (F. Li 1989; S. Li 2002:180; Zhao Zhijun, pers. comm. 2006). Remains of barley and wheat have also been found in South Korea and dated to c. 1000 cal. BCE (Crawford and Lee 2003).

The Glume Wheats

Two types of glume wheat dominate the archaeo- botanical samples recovered at Jeitun. Initially, c. 90% of the wheat remains were believed to consist of domestic einkorn (Harris et al. 1993:332), but further investigation has shown that some of the material closely resembles a type of glume wheat, morpho- logically distinct from einkorn, that has been reported from Neolithic and Bronze Age sites in Greece, Turkey, Hungary, Poland, Austria, and Germany (Jones, Vala- moti, and Charles 2000; Köhler-Schneider 2003). Un- like einkorn, which is a diploid, the other glume wheat is probably a tetraploid and it may derive from wild Timopheev’s wheat (Triticum araraticum Jakubz. [syn. T. timopheevi Zhuk. subsp. araraticum]) or possibly from wild or cultivated emmer wheat (T. dicoccoides Körn and T. dicoccum). Wild T. dicoccoides occurs today through- out the Fertile Crescent, with a concentration in the central and southern Levant, whereas T. araraticum extends around the northern and eastern sectors of the crescent from southeastern Turkey to the southern

Zagros, with outlying populations in the Caucasus (Zo- hary and Hopf 2000:45), but neither species has been recorded east of the Caspian Sea. A domesticated form of Timopheev’s wheat (T. timopheevi Zhuk.) was culti- vated until recently in Georgia, where it may have been domesticated from T. araraticum (Zohary and Hopf 2000:58), and it is also possible that the glume wheat now recognized in the Jeitun assemblage represents an extension far to the east of a formerly much wider Neolithic distribution of T. timopheevi cultivation. How- ever, none of these possibilities implies that it might have been domesticated locally in Turkmenistan, and its presence there in the Neolithic is almost certainly the result of its introduction as a domestic cereal from somewhere west of the Caspian.

In contrast to the uncertain origin of the newly recognized glume wheat from Jeitun, the ancestry of domestic einkorn is well understood. Its progenitor is wild einkorn (Triticum monococcum L. subsp. boeoticum [Bois.] A. et D. Löve [syn. T. boeoticum Boiss. emend. Schiem.]), a self-pollinating annual diploid grass (chro- mosome number 2n=14), with which domestic einkorn is fully interfertile. A second wild diploid wheat (T. urartu Tuman.) exists in Southwest Asia. It closely re- sembles two-grain wild einkorn but is intersterile with wild and domestic einkorn and is not implicated in the latter’s domestication (Waines and Barnhart 1992; Zohary and Hopf 2000:20–22, 36–38), although Heun, Haldorsen, and Vollan (2008) have speculated that finds of “einkorn” from Neolithic contexts at sites in the middle Euphrates valley may instead be domesticated T. urartu. Like other wheats, and unlike barley, wild einkorn generally avoids calcareous soils and occurs mainly on acid (often basaltic) soils. As a result, its dis- tribution in Southwest Asia is less extensive and more patchy than wild barley. It is more cold- and drought- tolerant than other wild wheats and its range extends in a broad band from the northeastern margins of the Fertile Crescent to western Turkey, with more isolated populations in the Caucasus, the central Levant, and in Greece and the Balkans where they are weedy forms that probably spread with domestic einkorn (Nesbitt 2001:46–48; Valkoun 2001; Valkoun, Waines, and Konopka 1998:295; Zohary and Hopf 2000:37).

Heun et al. (1997, 2008) undertook AFLP analy- sis of DNA obtained from a large number of samples of wild and cultivated einkorn from sites between the Balkans and the Caucasus (mainly from within the northern Fertile Crescent), and concluded that einkorn was monophyletic and had probably been

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

areas of origin of the crops and domestic animals 77

domesticated in the Karacadağ mountains a short dis- tance east of those archaeological sites. A modification of this hypothesis was suggested by Kilian et al. (2007) who used AFPL and haplotype data to investigate einkorn domestication and proposed a “dispersed- specific model” of multiple “domestication events” in a larger area of southeastern Turkey. However, as was pointed out in the preceding section, the use of AFLP data to infer crop origins by constructing neighbor- joining trees can produce misleading results. Nor did Heun et al. (1997) distinguish taxonomically between one-grain and two-grain forms of einkorn, which may have been separately domesticated.

One- and two-grain einkorn have in the past been classified as separate subspecies of Triticum boeoti- cum: subsp. aegilopoides (Link) Schiem. (one-grain) and subsp. thaoudar Reuter ex Hausskn. (two-grain), or even as different species: Triticum aegilopoides (Link) Bal. and T. thaoudar Reuter. Whereas the one-grain form occurs mainly in the northwestern part of the range of wild einkorn in western Anatolia and around the Aegean, where it occupies what are probably sec- ondary habitats, the more robust two-grain form is found throughout the northern Fertile Crescent, with intermediate forms also present, especially in central Anatolia, the Caucasus, and northwestern Iran (Kreuz and Boenke 2002:234; Zohary and Hopf 2000:36).

The earliest records of domesticated one-grain einkorn come from sites in the northern Levant and southeastern Anatolia such as Tell el Kharkh, Nevalı Çori, and Cafer Höyük during the 9th millennium cal. BCE, in the Early PPNB (de Moulins 1997:52, 68; Pasternak 1998; Willcox 2005:536–37; 2007:24). In contrast, definite evidence of the domestic two- grain form—which could derive either from Triticum boeoticum subsp. thaudar or possibly from T. urartu (Fuller 2007:8)—is restricted to two much later sites in northern Syria: Tell Sabi Abyad I, where the grain was found in a late Neolithic/early Chalcolithic context (van Zeist 1999), and Kosak Shamali, where the grain was recovered from burnt storage structures of the Chalcolithic (Ubaid) period (Willcox 2003). There is no evidence that two-grain einkorn continued to be cultivated at Sabi Abyad in later periods, and van Zeist suggested that it may have been domesticated in southeastern Anatolia, where the wild form is abun- dant, and introduced as a crop into northern Syria, where it was only cultivated for a short period and then abandoned. It (or domesticated T. urartu rather than two-grain einkorn according to Heun, Haldorsen, and

Vollan [2008:449]) also spread west into eastern and central Europe, where there is evidence of two-grain (as well as one-grain) einkorn in the Neolithic and the Bronze Age (Köhler-Schneider 2003:109; Kreuz and Boenke 2002:238), but it did not long survive as a cultivated crop in Southwest Asia or Europe. As it has been identified at Jeitun, it evidently also spread east into western Central Asia.

These advances in understanding the begin- nings of einkorn cultivation suggest that the one- and two-grain forms may have been separately domes- ticated in Southwest Asia, but much more genetic and archaeobotanical evidence is needed before the question of where and how frequently einkorn was domesticated in the region can be resolved. It is interesting that both forms were present at early Neo- lithic Jeitun, but highly improbable that either was independently domesticated in Central Asia. Wild einkorn has never been recorded east of the Caspian Sea, and it is unlikely that its range extended into Turkmenistan even during the climatic fluctuations of the Late Pleistocene and Early Holocene. Had it done so, it would probably have survived in isolated refugia in the intermontane valleys and piedmont of the Kopetdag, as a component of woodland and/or steppe vegetation. In its main area of distribution in Southwest Asia it commonly occurs as a component of oak woodlands; the extinction of oak in the Kopetdag during the Quaternary period (this volume, p. 19), and the absence of native species of Quercus in Turk- menistan today (Nikitin and Geldykhanov 1988:140), may partly account for the absence of wild einkorn east of the Caspian.

The possibility that wild einkorn might have been present in southern Turkmenistan in the Early Holocene and have been domesticated locally can- not be totally excluded, but the weight of evidence is heavily against that hypothesis. Like the other type of glume wheat found at Jeitun, the presence there of einkorn in both the two-grain and the one-grain form is much more likely to be the result of their introduc- tion, as domestic crops, from Southwest Asia. So too is the presence of the free-threshing Triticum aestivum/ durum wheat that has been tentatively identified in three samples recovered there (this volume, p. 153); and see Zohary and Hopf 2000:42–59 for a discussion of the distribution and origin in Southwest Asia of the tetraploid durum-type and hexaploid aestivum-type wheats). Finally, it is worth noting that the case for einkorn having been introduced during the Neolithic

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

78 origins of agriculture in western central asia

from Southwest Asia as an already domesticated crop is paralleled by evidence from Mehrgarh in Baluchistan. Domestic einkorn has been reported as present there (with the barley referred to above) in the earliest levels (Costantini 1983:31; Jarrige 2007–08:142), and it is most unlikely to have been domesticated locally because the site lies far to the southeast of the known range of wild einkorn in the northern Fertile Crescent and Turkey.

The Domestic Animals and Their Wild Progenitors

The only domesticated animals whose remains have been recovered during excavations at Jeitun are the dog and two ungulate species, goat and sheep. The other domesticated ungulates associated with prehis- toric agricultural and pastoral economies in Central Asia—pigs, cattle, horses, and camels—appear later in southern Turkmenistan than the Early Jeitun phase, although all of them were present as wild animals in the Late Pleistocene/Early Holocene. Evidence of them in the archaeological record, and the question of whether any of them may have been domesticated locally, is first reviewed, before the three domestic animals recorded at Jeitun are discussed.

Pigs

Remains of pig have been recovered at Jeitun, but, on the grounds of their relative rarity and the large size of the molar teeth found, they are thought to derive from the wild boar (Sus scrofa), which prob- ably inhabited thickets of tamarisk and reeds along the Kara Su (this volume, p. 177; Kasparov 1992:57; Shevchenko 1960:466, 470–73). Farther west in the Iranian Caspian lowland, bones of wild boar have been reported from Mesolithic levels at the sites of Ali Tappeh and Ghar-i Kamarband (Uerpmann and Frey 1981:148, 151). No definite evidence of domestic pigs has been reported from any Jeitun-Culture site, but by the Early Chalcolithic period domestic pigs—much smaller than modern wild boar—were evidently being raised at Anau North (during the IB1 and IB2 phases, c. 4000–3500 cal. BCE: Moore, Ermolova, and Forsten 2003:155–56). This accords with Duerst’s conclusion (1908:355–58), from his study of over 100 pig bones excavated at Anau North in 1904, that their generally

small size and the absence of pig bones in the lowest layers pointed to their being domesticated, although not at Anau itself; he thought that pigs might have been domesticated elsewhere in “Turkestan” or imported as tamed animals “from Iran or India” (ibid., p. 357). Much farther east at the Neolithic site of Ayakagytma in the Kyzylkum desert, a few pig bones have been re- covered from both the older (c. 6000–5500 cal. BCE) and the more recent (c. 4000–3000 cal. BCE) levels (Lasota-Moskalewska et al. 2006:208, and this volume, p. 67). The bones are thought to be from domestic pigs, but in view of their poor preservation and the very small sample, this should be regarded as highly speculative. Thus, there is no conclusive evidence at present of domestic pigs in the region during the early Neolithic. Those reported from the IB and later levels at Anau and at other Late Neolithic/Chalcolithic sites in southern Turkmenistan, such as Dashlidji (Masson and Sarianidi 1972:61), are probably descendants of already domesticated pigs introduced from farther west, in and beyond Iran, although local interbreeding with wild boar may have occurred subsequently.

All domestic pigs are descended from Sus scrofa, with the possible exception of the Sulawesi warty pig which according to Groves (2007:27–29) may have been domesticated from Sus celebensis. Local domesti- cation of wild boar in western Central Asia in the Late Pleistocene/Early Holocene cannot be excluded on zoogeographical grounds, but without more definite evidence of domestic pigs at Mesolithic or early Neo- lithic sites in the region that hypothesis lacks support. This contrasts with early zooarchaeological evidence of pig domestication in the northern Levant and east- ern Anatolia. There pigs appear to have been under some degree of human management by the Late/ Final PPNB1 at the sites of Çayönü (Ervynck et al. 2001; Hongo and Meadow 2000; Hongo et al. 2002:154–57) and Hallan Çemi (Redding 2005:43–44; Rosenberg et al. 1998), and perhaps already morphometrically domesticated at Hayaz, Halula, and Gürcütepe (Peters et al. 1999:41; Peters, von den Driesch, and Helmer 2005:113–14). By early in the Pottery Neolithic period (after c. 6500 cal. BCE) domestic pigs were present at many more sites around the Fertile Crescent, at least a millennium before there is evidence of them in western Central Asia, possibly in the early Neolithic at Ayakagytma and more definitely in the Chalcolithic at Anau North.

Recent genetic research on the phylogeography of the genus Sus does not point to western Central Asia

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

areas of origin of the crops and domestic animals 79

as a probable area of pig domestication. Analyses of mi- tochondrial DNA (mt DNA) and nuclear genes of wild boar and domestic pig populations in Asia and Europe suggest that pigs were domesticated independently in eastern and western Eurasia, with the probability that several domestications from distinct wild lineages of Sus scrofa occurred in both regions (Albarella, Dob- ney, and Rowley-Conwy 2006:218–19; Giuffra et al. 2000; Kijas and Andersson 2001; Larson et al. 2005, 2007a, 2007b). But to confirm and elaborate this pat- tern, more extensive sampling of wild and domestic pig populations in Eurasia, including Central Asia, is needed, together with systematic examination and dating of existing zooarchaeological evidence (cf., for China, Flad, Yuan, and Li 2007:168–69, 192; Yuan and Flad 2002), and, where feasible, analyses of DNA from ancient specimens (Larson et al. 2007b).

Cattle

No remains of domestic (or wild) cattle have been found at Jeitun. Nor have domestic cattle been reported from other Jeitun-Culture sites, with the ex- ceptions of Chagylly in the eastern region and Chopan in the central region of the Kopetdag piedmont where they were found in Middle- and Late-Jeitun levels (Berdiev 1966:26–27). In the subsequent Early Chal- colithic period they are relatively abundant (remains of at least 22 individuals) at the small site of Dashlidji in the Geoksyur oasis, part of the ancient delta of the Tedzhen river (Masson and Sarianidi 1972:58–61), which may imply that domestic cattle played an in- creasing economic role as settlements based on irriga- tion agriculture developed in the oasis through the Chalcolithic. There is also evidence of cattle breed- ing at the large piedmont sites of Ilgynly-depe by the Middle Chalcolithic period and at Altyn-depe by the Early Bronze Age (Kasparov 1994:145–47).

Bones identified as domestic cattle, and a smaller number probably from wild cattle (aurochsen), have recently been found at the Neolithic site of Ayak- agytma in the Kyzylkum. They are more numerous in the small animal-bone assemblage so far analyzed than any of the other taxa identified (except camels), and more abundant in the earlier than the later phase of occupation (Lasota-Moskalewska et al. 2006:208). It is suggested that the wild cattle were regularly hunted only during the earliest occupation of the site (sub- unit 5c, c. 6000–5700 cal. BCE) and that some of them

may have been locally domesticated (an hypothesis that awaits further investigation through analysis of larger bone assemblages more recently excavated at the site). The presence of cattle identified as domestic diminishes rapidly from c. 5600 cal. BCE and ceases during the final phase of occupation, possibly as a result of climatic changes starting in the late Neo- lithic and continuing to the beginning of the Bronze Age (Lasota-Moskalewska et al. 2006:208–15, and see below, p. 81).

The aurochs (Bos primigenius) is the wild pro- genitor of all domestic cattle. Three subspecies are recognized: Eurasian B. p. primigenius, South Asian B. p. namadicus, and North African B. p. africanus (syn. opisthonomous). Domestic cattle are conventionally di- vided into breeds with and without humps. The former (indicine or zebu) breeds are thought to descend from the South Asian subspecies, which is often referred to as Bos indicus, and the latter (taurine) breeds from the Eurasian subspecies, often referred to as Bos taurus (Clutton-Brock 1999:27, 84–85; Grigson 1985).

Aurochsen are now extinct in the wild, but in the Late Pleistocene/Early Holocene they ranged across temperate and tropical Eurasia and North Africa where they browsed and grazed in forests, woodlands, and open-canopy shrub vegetation. Many aurochs bones were found in the Mesolithic levels at Ali Tappeh and Ghar-i Kamarband in the Iranian Caspian lowland (Coon 1951:44; McBurney 1968:396–97; Uerpmann and Frey 1981:148, 151), and although Coon speculated (ibid., p. 50) that the Bos bones he recovered from Neo- lithic levels at the latter site were from domesticated oxen, Uerpmann and Frey pointed out (ibid., p. 147) that the bones could not be distinguished from those of aurochs in the Mesolithic levels. Duerst concluded (1908:359–69), from his study of bovid bones from Anau North, that the cattle remains from the Early Chalcolithic IA stratum (now dated to c. 4500–4000 cal. BCE) derived from what he termed the Asiatic form of Bos primigenius, which he referred to as Bos namadicus (without implying that they were humped cattle of zebu type). He also showed that the smaller and more numerous bovid bones found in the upper layers of the site were from domestic cattle which, he hypothesized, might first have been domesticated lo- cally as a long-horned breed and then underwent a reduction in size, or, alternatively, that the smaller ani- mals “may have reached Anau with. . .other imported domestic animals” (ibid., p. 369). More recent exami- nation of new bone assemblages from Anau North ex-

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

80 origins of agriculture in western central asia

cavated between 1978 and 1982 produced many bones of large long-horned cattle from the IA stratum and later layers, as well as evidence of some smaller cattle, all of which are interpreted as domestic (Ermolova 1985; Moore, Ermolova, and Forsten 2003:155–56). The prevalence of large cattle at Anau is attributed by Moore et al. to diversity in the domesticated popula- tion rather than to the occurrence of wild cattle, and there is no evidence of a progressive shift through time from larger to smaller bovid metatacarpal bones, such as would suggest local domestication. It is possible that the assemblage represents a mixed population of introduced (domestic) cattle and (hunted) wild au- rochs, and, regardless of that possibility, it is probable that domestic cattle became an increasingly important component of the local agro-pastoral economy from early in the Chalcolithic period.

If we turn from the very limited evidence of domestic cattle in Neolithic/Chalcolithic southern Turkmenistan to consider the results of recent genetic analyses of modern breeds in Eurasia and Africa, the likelihood of cattle having been independently domes- ticated in western Central Asia diminishes. Analyses of mtDNA, nuclear microsatellites, and Y-chromosome markers from present-day breeds in Africa, India, and Europe have revealed a very ancient divergence (c. 200,000 years ago) between taurine and indicine cattle, and significant clustering of variance into three continental groups. This implies that distinct ancestral populations of aurochsen were independently domes- ticated in two, possibly three regions: humped cattle in South Asia, humpless cattle in Southwest Asia (the principal source of European cattle), and perhaps also taurine cattle in North Africa where, however, there is evidence of admixture of taurine and indicine compo- nents (Bradley and Magee 2006; Bradley et al. 1996; Hanotte et al. 2002; Kim et al. 2003; Loftus et al. 1994; MacHugh et al. 1997; Troy et al. 2001). The genetic data presently available do not support a hypothesis of independent cattle domestication in Central Asia, but the possibility that aurochsen were domesticated in the eastern part of their range in the arid interior of Asia should not be completely excluded.

In the Southwest Asian Fertile Crescent shifts toward increasing numbers of small cattle in bone assemblages at Neolithic sites, as well as evidence of earlier killing of the animals, suggest that the domes- tication process was underway by c. 8000 cal. BCE in the Middle PPNB and that by the Late/Final PPNB1 domestic cattle were present at many sites in the

northern Levant, including Çayönü, Hayaz, Gürcüt- epe, Halula, Tell es Sinn, Bouqras, and Ras Shamra (Helmer et al. 1998; Hongo et al. 2002:160–62; Öksüz 2000; Peters et al. 1999; Saña Seguί 2000). It is even possible that cattle were undergoing domestication by the Early PPNB at other sites in the northern Levant such as Dja’de (Helmer et al. 2005:92).

The fact that domestic cattle are present at Middle and Late/Final PPNB sites in the northern Levant more than a millennium before they appear at Neolithic and Chalcolithic sites in southern Turk- menistan, coupled with the genetic evidence that identifies Southwest rather than Central Asia as a probable center of aurochs domestication, suggests that domestic cattle spread to Turkmenistan in the latter part of the Neolithic from the Fertile Crescent. If so, their introduction probably took place across northern Iran, but at present the absence of securely identified and well-dated finds of domestic cattle at Neolithic sites in northern Iran means that this sup- position must remain tentative.

The lack of any definite evidence that aurochsen were independently domesticated in western Central Asia contrasts with the situation south of the Hindu Kush and Afghan plateau at the site of Mehrgarh in Baluchistan. Analysis by Meadow (1993:304–13) of Bos remains from the aceramic Period I and ceramic Period II levels at Mehrgarh demonstrated a gradual increase between c. 7000 and c. 5000 cal. BCE in the proportion of cattle in the bone assemblages, and a de- crease in their size. He interpreted this as evidence of a change from predominantly wild to predominantly domestic cattle of zebu type and thus of local domes- tication of “what is today called Bos indicus” (Meadow 1993:310), i.e., the South Asian subspecies Bos primige- nius namadicus. Cattle of zebu type may also have been domesticated farther east in South Asia (Fuller 2006: 30), a possibility strongly reinforced by recent analyses of mtDNA diversity in zebu cattle that have identified two distinct haplogroups in the Indian subcontinent (Baig et al. 2005; Magee, Mannen, and Bradley 2007; Chen et al. 2010).

Equids and Camels

No remains of wild or domestic equids or cam- els have been found at Jeitun. Shevchenko (1960:465) tentatively identified onager or kulan (Equus hemio- nus kulan, this volume, p. 14) from one fragment of

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

areas of origin of the crops and domestic animals 81

bone, but subsequent osteoarchaeological research at the site has not yielded any onager bones (Kasp- arov 1992:73; this volume, p. 176). Nor have remains of onager been found at any other Jeitun-Culture sites with the exception of Chagylly where they are represented by a few bones that increase in number from the Middle- to the Late-Jeitun levels (Berdiev 1966:27; Kasparov 1994:148). Onagers appear in Chalcolithic bone assemblages at the eastern pied- mont sites of Ilgynly-depe and Altyn-depe, and Kasp- arov (1994:148) suggested that their earlier scarcity in the Neolithic may have been due to more humid climatic conditions in the 6th millennium BCE to which they were less well adapted than to the more arid climate of the Chalcolithic period when they were a major hunted prey. He implied too that this may account for the scarcity of onager bones in the Mesolithic and Neolithic levels at the Jebel and Dam Dam Cheshme rockshelters in the Bolshoi Balkhan massif reported by Tsalkin (1956:220). Onager bones were also present only in small numbers at the Meso- lithic sites of Ghar-i Kamarband and Ali Tappeh in the Iranian Caspian lowland (Uerpmann and Frey 1981:148, 151).

Equid and camel bones are represented in the animal-bone assemblage from Ayakagytma in the Kyzylkum in both the older and the more recent levels. Whether the equid remains are from horse or onager (“half-ass”), or both, could not be determined because of the poor preservation of the bones. They are most abundant in the oldest stratigraphic unit (5c, c. 6000–5700 cal. BCE), diminish thereafter, and are not represented in the final phase of occupation (unit 2). This decline parallels that of cattle (referred to above) and may be related to climatic changes. Lasota- Moskalewska et al. (2006:215) infer that the horses (or half-asses) were “if not fully domesticated, then at least tamed.” However, given the very poor preserva- tion of the remains and the relatively small number of equid bones identified (179), the inference seems very speculative. An alternative hypothesis that the equids were hunted appears more plausible.

Camels comprise the largest taxon in the Ayak- agytma animal-bone assemblage (490 identified bones) and they are present in all the stratigraphic units. In contrast to the cattle and equids, camel bones increase markedly from the middle of unit 5b at c. 5600 cal. BCE to over 60% of all animal remains by the end of the early Neolithic Dariasai phase and, after the settlement hiatus from c. 5400–4000 cal. BCE, to over

75% in units 4 and 2 (4000–3000 cal. BCE). It is sug- gested that the camels were “most probably if not fully domesticated” or “at least tamed,” and that the great increase in their number by the end of the occupa- tion, when they had “completely supplanted cattle and horse,” was probably due to climatic changes (ibid., p. 215, and see below).

In his study of equid and camel bones from Anau North on the Kopetdag piedmont, Duerst (1908:384– 99, 401–31) attributed the former to the domestic horse (Equus caballus) and the latter tentatively to the domestic Bactrian camel (Camelus bactrianus). Whereas horse bones were abundant and present in all the layers of the site, only two camel bones were found. They occurred in the highest layers, and Du- erst assumed that the (probably Bactrian) camel was introduced late as a domestic animal “from Bactriana or the Iranian plateau” (ibid., p. 384). His identifica- tion of domestic horse at Anau was later disputed by several authors, most recently by Forsten (2000) who examined some of Duerst’s original specimens and concluded that the bones and teeth were from onager rather than horse (Moore et al. 2003:157–59). Since Duerst’s time, a few more camel bones have been recovered from excavations at Anau and identified by Ermolova (1985:86) as from the Bactrian rather than the dromedary, although it is unclear whether the remains represent hunted animals or isolated finds of domesticated camels (Moore et al. 2003:157).

The paucity of archaeologically recovered equid and camel bones from Neolithic sites in western Cen- tral Asia may in part be due to the general lack of zooarchaeological research. But it may also imply that wild onagers, horses, and camels were not abundant or commonly hunted at that time, perhaps as a result of the onset of wetter and warmer conditions around 6000 cal. BCE at the beginning of the Climatic Opti- mum when populations of these wild herbivores, es- pecially onagers and camels which are highly adapted to desert environments, probably declined. It is pos- sible that changes in the relative abundance of cattle, equid, and particularly camel remains between c. 6000 and 3000 cal. BCE at Ayakagytma reflects oscil- lations between relatively warm, wet conditions dur- ing the Climatic Optimum and cooler, drier phases such as the one that occurred between c. 4800 and 4400 cal. BCE (this volume, p. 26), but without more detailed local palaeoenvironmental evidence for the mid Holocene than is available at present that specu- lation cannot be tested. In Turkmenistan, bones of

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

82 origins of agriculture in western central asia

what are believed to be wild horses do not appear in archaeological contexts until the Middle Chalcolithic, for example at sites in the Geoksyur oasis (Masson and Sarianidi 1972:69); and the earliest reliable evidence of domestic horses dates to the end of the 3rd millen- nium BCE (Ermolova 1983).

West of Turkmenistan on the north Iranian cen- tral plateau, three sites on the Qazvin plain, Zaghe, Ghabristan, and Sagzabad, provide evidence of the exploitation of (mainly wild) equids from the Late Neolithic to the Iron Age (Mashkour 2002, 2003). Remains of onager dominate the equid-bone assem- blages at all three sites, but another (now extinct) wild equid, Equus hydruntinus (Clutton-Brock 1999:115), which inhabited Southwest Asia and southern Eu- rope from Palaeolithic to Neolithic times, has also been identified in the assemblages, thus extending its known prehistoric range farther east and its survival into the Iron Age. Some of the equid bones from the Neolithic and Chalcolithic levels in the Qazvin sites have “caballine” features which suggest that horses might possibly have been domesticated locally on the Iranian plateau. The domestic ass or donkey (Equus asinus) is also present by the Chalcolithic period, probably having been introduced from the Fertile Crescent. Far to the east in northern China, remains of wild horses have been found at 28 Palaeolithic sites, but there is very little evidence of horse bones at Chinese Neolithic sites, none of it incontestably of domestic horses, and it is not until c. 1400 BCE in the Late Shang dynasty that definitely domesticated horses are present in the middle and lower Huang-ho valley, probably introduced from the northwest out of Central Asia (Flad, Yuan, and Li 2007:169, 194; Yuan and Flad 2006).

The loci of the initial domestications of horses and camels remain uncertain. Analyses of mtDNA sequences from samples of living and fossil horses have revealed a high degree of genetic variation in domestic breeds which could imply that wild horses were separately domesticated many times across an extensive area, but it also possible that initial do- mestication was more restricted in time and space, and that wild horses were later incorporated into domestic herds as the practice of horse breeding spread (Jansen et al. 2002; Levine 2006; Lister et al. 1998; Vilà et al. 2001, Vilà, Leonard, and Beja-Pereira 2006). In any case, it seems certain that horses were domesticated in the Eurasian steppe zone where the wild progenitor (Equus ferus) ranged in the Late

Pleistocene/Early Holocene. There is archaeological evidence that horses were being exploited intensively for food and other purposes at such sites as Dereivka in the Ukraine and Botai in Kazakhstan by c. 3500 cal. BCE, but whether the remains represent wild or domestic horses has been uncertain, largely because it is difficult to establish robust zooarchaeological cri- teria for distinguishing between wild, captive, tamed, and domesticated horses (Clutton-Brock 1992:54–55; Levine 1999a, 1999b; Olsen 2003, 2006a, 2006b). Now a fresh approach to the problem based on three independent lines of evidence—metrical analysis of metacarpal bones, examination of damage to mouth skeletal tissues caused by the use of bridles, and organic-residue analysis of potsherds indicating that both horse meat and milk were processed—has pro- vided strong evidence that some at least of the Botai horses were closely managed, milked, and possibly ridden (Outram et al. 2009).

In the Late Pleistocene/Early Holocene the range of wild horses probably extended south of the steppe zone into the deserts of Central Asia, where they would have overlapped with the range of the onager which is better adapted to arid conditions (Uerpmann 1987:34). However, horse populations would have been smaller there, especially in the des- ert lowlands, than farther north in the center of their range. In Central Asia they may have been more numerous and more intensively hunted in montane than in lowland desert habitats: for example, remains of wild horses dominate the bone assemblage in the Upper Palaeolithic horizons at the hunting-camp site of Shugnou in eastern Tajikistan (Davis and Ranov 1999:191; Vishnyatsky 1999:93–94; this volume, p. 54), and there are (inadequately documented) reports of Equus remains in the levels assigned to the Upper Palaeolithic at the Kara Kamar and Aq Kupruk II rockshelters in northern Afghanistan (Coon 1957:230; Dupree 1972:77).

Little is definitely known about the ancestry and domestication of the two-humped and one-humped camels, the Bactrian and the dromedary. Small herds of wild camels with two humps that may represent the progenitor of the domestic Bactrian survive in the Gobi desert. Named Camelus ferus, they are often assumed to be a remnant population of the ancestral wild camel, but it has been suggested that they might be feral descendants of domestic camels (Clutton- Brock 1999:156; Mason 1984:108). C. ferus probably inhabited most of Central Asia, including the Kara-

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

areas of origin of the crops and domestic animals 83

kum and Kyzylkum deserts, in the Late Pleistocene/ Early Holocene, and Mason states that its former distribution extended into southern Russia, Iran, and Afghanistan.

Bulliet (1975:48–49) suggested that the Bactrian camel may have been domesticated on the borders of Iran and Turkmenistan several centuries prior to 2500 BCE. The earliest direct evidence of domesticated Bactrians (apart from the inconclusively identified re- mains from Anau in Turkmenistan and the as yet insuf- ficiently documented camel bones from Ayakagytma in Uzbekistan, see above) comes from the large 3rd- millennium BCE site of Shahr-i Sokhta in east-central Iran, where remains were found of five camel bones, pieces of camel dung, and fibers of what was thought to be camel hair in fragments of woven cloth (Bökönyi and Bartosiewicz 2000:117; Compagnoni and Tosi 1978; Salvatori and Tosi 2006). Also, in the 3rd-millennium BCE levels of Sialk III near Esfahan, a representation of what may be a two-humped camel was found on a pottery sherd (Zeuner 1963:359). In Turkmenistan, clay models of 4-wheeled carts pulled by camels (and less frequently by horses) have been found at Early and Middle Bronze Age sites on the eastern piedmont, for example at Altyn-depe (Masson and Sarianidi 1972:109, 120, Plate 36), indicating that by c. 2000 BCE both had been incorporated as draft animals into systems of agricultural production. Also, by the end of the 3rd millennium, representations of domesticated Bactrian camels appear in the archaeological record of Margiana (the oasis settlements of the Murghab delta) in the form of terracotta figurines and an intricately carved steatite amulet from the site of Togolok 21 (Hie- bert 1994b:378–79; Moore et al. 1994:425).

The dromedary (Camelus dromedarius) is an im- portant domestic animal in western Central Asia today, but there is no evidence that it was present in prehistoric times. Finer-limbed and faster than the Bactrian, it is well adapted to the hot deserts of Southwest Asia and North Africa, and although its area of origin remains uncertain, it was probably first domesticated in the Arabian peninsula (Bulliet 1975:42–48; Mason 1984:109; Uerpmann and Uerp- mann 2002:258; Zeuner 1963:340–44). Dromedary bones have been found at several prehistoric sites in southeastern Arabia and analyzed by Uerpmann and Uerpmann (2002), who concluded that the earlier Neolithic and Bronze Age finds came from wild cam- els that were probably hunted, whereas remains of domestic camels did not appear until the Iron Age (at

the site of Tell Abraq, c. 900–800 BCE). The domestic dromedary subsequently spread across Southwest Asia, used mainly as a pack animal and in warfare, and may only have reached Central Asia in the first millennium AD, where it, and hybrids between it and the Bactrian, gradually displaced the latter as the pre- ferred pack animal on the Silk Route to China (Bulliet 1975:168–71; Köhler-Rollefson 1996:287–88). It was also integrated into the agro-pastoral economy, and is still bred and milked in the Karakum desert.

We can now turn from the domesticated ungu- lates whose presence in southern Turkmenistan post- dates the Early Jeitun phase to the two herd animals, goat and sheep, whose remains have been recovered at Jeitun. Goat and sheep (caprine) bones are the most abundant class of animal remains at Jeitun (this volume, pp. 175–77; Kasparov 1992:51), and they are well represented at other Jeitun-Culture sites. Both played a major role in Neolithic subsistence, and the question of whether they may have been domesticated locally is critically important for our understanding of the origins of the agro-pastoral economy of the Jeitun Culture.

Goats

Many living species of wild goats have been recognized taxonomically, but only three (interfer- tile) groups in the genus Capra are implicated in the ancestry of domestic goats—the bezoar or pasang (C. aegagrus), the markhor (C. falconeri), and the ibex (C. ibex). The bezoar (including subspecies) is recog- nized as the main progenitor, and the markhor and ibex may have contributed, after the domestication of the bezoar, to the ancestry of certain domestic breeds in, respectively, northwestern South Asia and northeastern Africa (Clutton-Brock 1999:76–78; Har- ris 1962; Manceau et al. 1999; Mannen, Nagata, and Tsuji 2001; Takada et al. 1997). All species of Capra are well adapted to rocky high-mountain habitats, and the bezoar is the most widely distributed of them, with a range in recent centuries that extended from Turkey to western Central Asia (Fig. 1.16). Although only small populations of the bezoar survive in Turkmeni- stan today (this volume, pp. 12–13), it was formerly more widespread there (and in northern Iran and northern Afghanistan), so the possibility that goats were domesticated locally cannot be excluded on zoogeographical grounds.

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

84 origins of agriculture in western central asia

There is no doubt that wild bezoar goats were hunted in western Central Asia prior to and during the Neolithic, but their presence has only been docu- mented at a few sites. This may partly be due to the dif- ficulty of distinguishing between the remains of wild and domestic goats. The customary morphological criteria for recognizing domestication are reductions in the size of the limb bones and the appearance of shorter, more rounded and sometimes twisted horn cores (Clutton-Brock 1999:78–79); and analysis of age and sex profiles can, when the evidence is adequate, demonstrate human management of herds preceding full domestication (Zeder 2001). Such criteria have not been systematically applied in the examination of goat bones from sites in the region and it is therefore difficult to determine the accuracy of reports of the presence of domestic and/or wild goats. This applies particularly to their reported presence in Mesolithic and Neolithic levels at the sites of Jebel and Dam Dam Cheshme 2 (DDC 2) in the Bolshoi Balkhan massif (Tsalkin 1956:221; Markov 1966a:123), Ghar-i Kamarband in the southern Caspian lowland (Coon 1951:44–50; Uerpmann and Frey 1981:148), and Aq Kupruk I and II in northern Afghanistan (Perkins 1972). These excavations were carried out before mod- ern methods of osteoarchaeological analysis began to be applied to bone assemblages, so the interpretations of the data contained in the published reports must be treated with caution. In particular, the suggestion that goats were locally domesticated at Ghar-i Ka- marband during the Mesolithic (Coon 1951:49–50) should be regarded with great skepticism, as should Perkins’ claim that both wild bezoar and domestic goats were present at the Aq Kupruk sites (Meadow 1989a:27–28).

The goat and sheep bones excavated by Oklad- nikov at the Jebel rockshelter were examined by the zoologist V. I. Tsalkin who concluded that, because they were so poorly preserved, it was very difficult to judge whether they represented wild or domestic ani- mals, although he tentatively concluded (1956:221), on the basis of the small sizes of the bones, that some of the specimens from levels 3 and 4 were from domes- tic goats. He also studied the bones from the DDC 2 rockshelter, and according to the excavator (Markov 1966a:123), he stated that some of the caprine bones from layer 4 were “indisputably” from domestic goats (and some of those from layer 3 were “indisputably” from domestic sheep). Although no radiocarbon dates were obtained for the archaeological sequence

at the site, Markov inferred that layer 4 was contem- poraneous with the occupation of Jeitun and that it represented the beginning of the Neolithic. Assuming Tsalkin’s identifications were sound, this correlation, if correct, suggests that the presence of domestic goats in layer 4 may be the result of contact with goat (and sheep) herders of the Early Jeitun Culture rather than of local domestication of wild bezoar goats. However, Masson and Sarianidi (1972:170) disagreed with Markov’s correlation, believing instead that the (more recent) layer 3 was synchronous with Jeitun, and therefore that domestic goats were present at DDC 2 prior to the occupation of Jeitun and could have been domesticated there. But even if their correlation were correct, already domesticated goats could have reached the Bolshoi Balkhan massif from farther west earlier than the foundation of Jeitun. Without a radiocarbon chronology for the sequence at DDC 2, and more research on bone assemblages from other early sites in the region, these uncertainties cannot be resolved (and see this volume, pp. 217–18, for further discussion of the possible relations between the Jeitun- Culture and Bolshoi Balkhan sites).

A more critical analysis of goat remains from Jeitun itself was undertaken by Kasparov (1992), using Tsalkin’s bone-size criteria for discriminating between wild and domestic caprines. On this basis, Kasparov (1992:51) concluded that some of the goat bones were from the wild bezoar, although most of them were the remains of domestic goats. If correct, this implies that the inhabitants hunted wild goats, probably in the Kopetdag and/or Bolshoi Balkhan mountains, as well as pasturing domestic herds closer to Jeitun. A few caprine bones of Neolithic age have also been excavated at Ayakagytma in the Kyzylkum, but their poor preservation and small number (19 fragments) prevented discrimination between goat and sheep. All but one of the bones came from the early Neo- lithic Dariasai phase (unit 5) dated to c. 6000–5500 cal. BCE, but none could be definitively identified as Capra. They were interpreted as remains of domesti- cated animals, except for one bone (in 5b) thought to be from a wild sheep (Lasota-Moskalewska et al. 2006:208–9).

At the later site of Anau North on the Kopetdag piedmont small numbers of wild-goat bones have been identified in the Chalcolithic IB and IIA strata, and there too remains of domestic goats are much more abundant (Moore et al. 2003:155). Without more extensive and detailed osteoarchaeological evidence

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

areas of origin of the crops and domestic animals 85

from sites in western Central Asia, the question of whether goats may have been domesticated locally cannot be resolved, but consideration of wider archae- ological and genetic evidence suggests otherwise.

The earliest published zooarchaeological evi- dence for probable goat domestication comes at present (2008) from two PPNB sites in the northern and east- ern Fertile Crescent: Nevalı Çori in the foothills of the southern Taurus mountains in southeastern Turkey and Ganj Dareh in the central Zagros mountains in western Iran, both within the range of the wild progenitor (the bezoar). The later PPNB site of Ali Kosh in the western foothills of the Zagros range south of Ganj Dareh pro- vides early evidence for domestic goats at the margins of the range of the bezoar. At Nevalı Çori, an assemblage of goat (and sheep) bones, characterized by more females than males, a high proportion of immature animals, and a trend toward smaller individuals, has been recovered from levels that date to the Early PPNB,1 suggesting that goats and sheep may have been kept and bred in southeastern Turkey by that time (Peters et al. 1999:35–40). The assemblage of goat bones re- covered at Ganj Dareh in the late 1960s and early 1970s has been re-analyzed and dated to c. 7900 cal. BCE by Zeder and Hesse (2000). They constructed sex-specific age profiles which showed that young male goats were selectively slaughtered and they concluded that the animals, although morphologically wild, were man- aged rather than hunted. They suggested that the Ganj Dareh assemblage provided evidence of initial human management of goats and a transition from hunting to herding, an interpretation originally proposed by Hesse (1978), with which Hole (1996) and Legge (1996) concurred. In contrast, the goat remains from Ali Kosh, which are chronologically at least 500 years more recent than those at Ganj Dareh (Zeder and Hesse 2000:2257), represent, as Flannery originally argued, a morphologi- cally domesticated population characterized by changes in horn shape (Hole, Flannery, and Neely 1969:270–78). By the Late PPNB there is widespread evidence for domesticated goats (and sheep) at sites in the north- ern Fertile Crescent such as Çayönü, Hayaz, Gritille, and Gürcütepe, as indicated by both the diminution in size and the increased representation of female animals evident in the bone assemblages (Hongo et al. 2002:157–60; 2005; Peters, von den Driesch, and Hel- mer 2005:97–98). By the beginning of the Final PPNB domestic goats and sheep are present throughout the Fertile Crescent, integrated into the now established agro-pastoral economy (Harris 2002a:70–78).

The evidence that goats were domesticated in southeastern Turkey and western Iran and were being kept at sites around the Fertile Crescent by 7000 cal. BCE, long before they are present at Pottery Neolithic sites in southern Turkmenistan some 1000 km east of the Zagros, shows that there was ample time for domestic goats (and sheep) to have spread eastward across northern Iran during the millennium or more that separates the PPNB sites of the Fertile Crescent from the PN sites on the Kopetdag piedmont. But to reach a more definite answer to the question of whether goats are likely to have been independently domesticated in western Central Asia, we need to consider the results of recent genetic research on their origins and spread.

Several research groups have investigated mtDNA in domestic and wild goat populations across much of Eurasia and northern Africa. Luikart et al. (2001) sampled 88 breeds of domestic goat, focusing on traditional “unimproved” breeds in remote rural areas. Their sampling included 406 domestic goats (none from Turkmenistan) and 14 wild goats of which 2 bezoars and 1 markhor were from Turkmenistan. The results revealed three highly divergent maternal lineages, the largest of which (A) was widespread in Southwest Asia, Europe, and Africa, whereas the other two much smaller lineages (B and C) were de- tected respectively in southern and eastern Asia and in central and eastern Europe. The authors inferred that the three lineages had expanded at different periods, and they suggested that A corresponded to the initial domestication of goats in the Fertile Cres- cent and their outward expansion from about 10,000 years ago, and that lineages C and B expanded about 6000 and 2000 years ago and represented secondary expansions long after the initial dispersal of domestic goats by people. Subsequently, the existence of three more lineages was recognized, now labeled D, F, and G and referred to as mitochondrial haplogroups by Naderi et al. (2007:2). Further research by Luikart et al. (2006) on mtDNA, Y-chromosome, and microsat- ellite markers for domestic goats and wild species of Capra supported the established view that the bezoar was the principal progenitor of domestic goats and re- inforced the evidence for three main lineages, as well as the inference that goats were initially domesticated in the Fertile Crescent.

Since then, Naderi et al. (2008) analyzed mtDNA from 473 wild-bezoar samples collected over the whole present range of the species and compared the results

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

86 origins of agriculture in western central asia

with the known genetic diversity of domestic goats. All the mtDNA haplogroups found in domestic goats were also found in the bezoar samples, and the geographi- cal distribution of these haplogroups in the bezoars allowed the probable location of independent centers of goat domestication to be determined. Naderi et al. inferred that the most probable area of origin of the A haplogroup (to which 90% of domestic goat haplo- types belong) is eastern Anatolia, where it is common in wild bezoar populations, whereas the C haplogroup predominates in the southern Zagros and the central Iranian plateau (in Fars, Yadz, and Kerman prov- inces). Estimated bezoar population growth rates were found to be higher in the C haplogroup than in other haplogroups and Naderi et al. hypothesized that this reflected early “management” or “incipient domestica- tion” of some bezoar populations in the central Zagros prior to their “true domestication”—which accords with Zeder and Hesse’s analysis of the goat bones from Ganj Dareh.

Naderi et al. (2008) concluded that goats were probably domesticated independently in eastern Ana- tolia and the southern Zagros/central Iranian plateau, that the former region contributed more than the latter to the modern goat gene pool, and that, despite very ex- tensive sampling of bezoars across Iran, no haplotypes were found that could have been domesticated in the eastern half of the central Iranian plateau or eastward of it—a conclusion that argues against a hypothesis of independent domestication in western Central Asia. Nor did Naderi et al. find any mtDNA evidence that would support an independent domestication of bezo- ars in the Indus valley region, for example at the site of Mehrgarh where remains of very small goats presumed to be domestic have been found in the earliest aceramic (Period I) levels, and where there is also evidence that young goats were buried with humans (Lechevallier, Meadow, and Quivron 1982; Meadow 1993:310). The site is on the margins of the recent range of the bezoar and Meadow reported (ibid.) the presence of bones of “some very large animals during Period IA [which sug- gests] that wild goats were also hunted at least until the end of that period.” Wild bezoars may well have been hunted and domestic goats raised at Mehrgarh (as evi- dently happened at Neolithic Jeitun), but the mtDNA evidence suggests that the domestic goats had spread to the area from the Fertile Crescent, rather than been locally domesticated.

The recent genetic research on goat domestica- tion offers no support for a hypothesis of independent

goat domestication in western Central Asia. Further- more, the credibility of inferring past conditions from present-day data is strengthened by the observation that none of the bezoar haplogroups apparently un- derwent so much population reduction that the pres- ent genetic structures would not reflect those of the Early Holocene (Naderi et al. 2008:17,660). However, there are limits to the inferences that can plausibly be drawn from the genetic data, and what is now needed is to test and elaborate the data with analyses of ancient DNA (aDNA) recovered from goat bones in Neolithic (and Mesolithic) sites. No such investigations have as yet been undertaken in western Central Asia, but the feasibility of the technique has been demonstrated by analysis of aDNA from Neolithic and Chalcolithic/ Iron Age sites in southern France and northern Iran (Fernández et al. 2005, 2006; Mashkour, Fontugne, and Hatté 1999). In both investigations it proved pos- sible to relate the aDNA sequences to corresponding haplogroups and to infer that in both cases the goats were descended from populations that originated in the Fertile Crescent.

Sheep

All wild and domestic sheep are grouped in the genus Ovis but their classification is both confused and confusing. In recent decades taxonomists have tended to reduce the number of species and sub- species recognized, on the grounds that all extant wild sheep are capable of interbreeding and because genetic data have supplemented the morphological criteria on which many distinctions between taxa were initially based (Nadler et al. 1973). Wild sheep can be divided according to their diploid chromosome numbers into four groups and seven species, all but one of which (the American bighorn, O. canadensis) are native to western and central Asia (Clutton-Brock 1999:69–72). The Asiatic mouflon (O. orientalis, dip- loid chromosome number 2n=54) is generally believed to be the principal or only progenitor of all domestic sheep. The so-called European mouflon (2n=54), which was formerly referred to as O. musimon but is now usually classified as O. orientalis, survives in a feral state in the mountains of Corsica and Sardinia and is believed to be a descendant of domestic sheep origi- nally introduced to the islands during the Neolithic (Poplin 1979). Two other species of Asiatic wild sheep, the argali (O. ammon, 2n=56) and the urial (O. vignei,

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

areas of origin of the crops and domestic animals 87

2n=58), have been regarded as possible contributors to the ancestry of domestic sheep (Ryder 1984:65–66; Zeuner 1963:159–61), but their higher chromosome numbers make this unlikely.

Analyses of Ovis mtDNA by Hiendleder et al. (1998, 2002) also identify the Asiatic mouflon as the probable ancestor of domestic sheep, and provide no evidence of contributions from either the urial or the argali. However, two wild hybrid populations of mouf- lons and urials have been described, one in the Elburz mountains of northern Iran and the other in Iranian Baluchistan. Their existence has been attributed to the interbreeding of formerly separate populations that came into contact as a result either of the melting of montane glacial barriers early in the Holocene (Na- dler et al. 1973:118) or, more probably, of a reduction of forest barriers caused by increasing aridity during the last two millennia coupled with more intensive exploitation of vegetation (Valdez, Nadler, and Bunch 1978:68). These cases of natural hybridization in the wild show that urials could have interbred with mou- flons and contributed to the ancestry of some breeds of domestic sheep, although the available mtDNA data do not support that possibility.

Today the Asiatic mouflon survives in small numbers in mountainous and hilly habitats from south-central Turkey to Armenia, south down the Zagros range, and east to the Elburz mountains south of the Caspian Sea (Fig. 1.15). In the Late Pleisto- cene/Early Holocene its range, attested by remains found in archaeological sites, extended farther west in central Turkey and farther southwest into the southern Levant, although apparently not east of the Elburz mountains into Turkmenistan (Uerp- mann 1987:126–27). The range of the urial extends in mountain habitats from the eastern Elburz across Turkmenistan and farther east and south in western Central and South Asia (Fig. 1.15). The urial sheep of the Bolshoi Balkhan mountains in Turkmenistan have been classified as a steppe subspecies, O. v. arkal, that formerly ranged across northern Turkmenistan to the delta of the Amudarya (this volume, p. 12). Uerpmann (1987:130) argued that both the urial and the mouf- lon had more restricted ranges in the Late Pleistocene and were more separated geographically then than they are today—a view that accords with Valdez et al.’s (1978) suggestion that the urial-mouflon hybrid populations in Iran may be a recent phenomenon due to forest clearance having removed barriers to the extension and overlapping of their ranges. The

third species of wild sheep, the argali, is larger than the urial and the mouflon and lives at higher altitudes east of the range of the urial (Fig. 1.15). Its range extends eastward from beyond the upper Syrdarya valley through the Tien Shan and Altai mountains to Mongolia and south to the Tibetan plateau and the Himalayas.

The geographical pattern of adjacent but spa- tially discreet ranges of the three main Asiatic spe- cies of wild sheep, from Turkey in the west to Tibet in the east, is borne out by such osteoarchaeological evidence as there is of the remains of wild sheep. Uerpmann (1987:126–32) documented the presence of the Asiatic mouflon at some 40 Late Palaeolithic, Mesolithic, and Neolithic sites around the Fertile Cres- cent and in central Turkey, and also of the urial at four such sites farther east in northern Iran (Ali Tappeh and Ghar-i Kamarband), southern Turkmenistan (Anau), and northern Afghanistan (Aq Kupruk). At the two Iranian sites fragments of horn cores typical of the urial were found in Mesolithic contexts (Uerp- mann 1987:130; Uerpmann and Frey 1981:153–55). The small numbers of wild-sheep bones found at Anau came from the Chalcolithic IB and IIA strata and were greatly outnumbered by the bones of domestic sheep (Moore et al. 2003:155). Meadow (1989a) evaluated the sheep bone assemblages from Aq Kupruk I and II and concluded that Perkins’ (1972) identification of the remains of wild urial sheep in the Upper Palaeoli- thic (Kuprukian) levels was probably correct, but that his identification of domestic sheep in the “Neolithic” and later levels was not well founded. Remains of wild sheep (unspecified, but presumably urial) were also found by Coon in what he interpreted as Mesolithic deposits at the site of Kara Kamar in northern Af- ghanistan (Coon 1957:234).

In Turkmenistan remains of wild and domestic sheep have been reported from Mesolithic and Neo- lithic levels at the Bolshoi Balkhan rockshelters of Jebel and Dam Dam Cheshme 2. Tsalkin (1956:221) was un- able to determine, because of the poor preservation of the bones, whether any of the sheep bones excavated at Jebel were from domestic animals, but (as already mentioned) Markov (1966a:123) stated that Tsalkin reported some of the bones found in (Neolithic) layer 3 at DDC 2 to be “indisputably” from domestic sheep. Also, at Jeitun Kasparov (1992:51, 63–64) identified a few bones of wild urial sheep, all from adult animals, among the more abundant remains of domestic sheep and concluded that the urial was hunted in the moun-

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

88 origins of agriculture in western central asia

tains together with wild bezoar goats. In Uzbekistan, in the early Neolithic (Dariasai phase) levels at Ayak- agytma in the Kyzylkum, a single bone was tentatively identified as from a wild sheep, but all the other caprine bones in the small assemblage were recorded as remains of domesticated sheep (and possibly goat). Lasota-Moskalewska et al. (2006:215) found the appar- ent lack of “interest in sheep (or goat) breeding. . .very strange” and concluded that sheep did not have “major economical importance” (and see Chapter 11 for gen- eral discussion of the possible role of stockbreeding in the subsistence economy at Ayakagytma).

The genetic evidence for the ancestry of domes- tic sheep (Hiendleder et al. 1998, 2002), together with the distribution of their putative wild progenitors, identifies the Asiatic mouflon as the only, or at least the principal, ancestor, and the locus of domestication as west of Turkmenistan within the natural range of the mouflon. This general conclusion was supported by two further studies of ovine mtDNA. In an analysis of 836 samples from domestic breeds and wild sheep in Eurasia and Africa, Bruford and Townsend (2006) found no evidence of urial or argali contributions to present-day breeds, although they did emphasize the need for “further sampling of domestic populations in Central Asia” and elsewhere, as well as “a detailed study of O. orientalis” (Bruford and Townsend 2006:311, 314). This recommendation was partly addressed by an analysis by Tapio et al. (2006) that included data from Central Asia. They analyzed mt DNA from 48 breeds and local varieties of domestic sheep sampled across a wide area from northern Europe to east of the Caspian Sea, in a study that identified four divergent lineages (haplogroups A, B, C, and D) and thus confirmed the findings of Hiendleder et al. (1998, 2002), Pedrosa et al. (2005), and Guo et al. (2005).

Groups A, B, and C are indicative of population expansions and group D was found only in a single sheep in the Caucasus region. Groups A and B are interpreted as representing expansions westward and eastward out of Southwest Asia and the Caucasus ap- proximately 9000 years ago, which is broadly compat- ible with the archaeological evidence for initial sheep domestication in that region. Groups A, B, and C were all found in the two Central Asian sample areas (one south of the Aral Sea and the other in the Altai mountain region). The presence of A and B in Central Asia is attributed to a gradual spread of domesticated sheep from Southwest to Central Asia, and group C, which occurs mainly in Central Asia, is thought to

have emerged more recently there, two or three mil- lennia after the domestication of sheep in Southwest Asia, possibly as a result of introducing wild mouflon ewes or lambs into local domesticated stock (Tapio et al. 2006:1781). This interpretation of the pattern of mtDNA diversity in Central Asian breeds should be re- garded as provisional and it calls for further sampling in the region; but it does suggest that the presence of domestic sheep in Neolithic western Central Asia is attributable to their gradual spread from Southwest Asia rather than the result of local domestication of the wild mouflon, the range of which probably did not extend east of the Caspian.

The zooarchaeological evidence already dis- cussed demonstrates that wild urial sheep continued to be hunted during the Neolithic when the herding of domesticated sheep (and goats) became established as part of the agro-pastoral economy of the Jeitun Cul- ture. This, together with the possibility (already men- tioned) that sheep as well as goats were kept and bred in the Early PPNB at the site of Nevalı Çori, and defi- nite evidence of domestic sheep at many sites around the Fertile Crescent by the Late PPNB, reinforces the genetic evidence and leads to the conclusion that the domestic sheep of Neolithic Turkmenistan were probably descendants of stock domesticated from the mouflon in Southwest Asia that had spread to western Central Asia during the Pottery Neolithic period. So too, perhaps, were the sheep recorded from the earli- est aceramic and ceramic (Periods I and II) levels at Mehrgarh in Baluchistan, although, as Meadow sug- gested, it is possible that “local wild sheep. . . may have been kept and bred, only to be replaced later in the history of the region by western forms developed for their wool and fat production” (1993:311).

Dogs

The dog is the third domestic animal, in ad- dition to goats and sheep, whose remains have been found during excavations at Jeitun. Canid bones from the site were first reported, following Masson’s excava- tions in 1956–58, by Shevchenko (1960:473–75) who identified both wolf and domestic dog, but Kasparov, who analyzed the animal remains recovered dur- ing the 1989–91 excavations, identified all the (42) Canis bones he examined as domestic dog (Kasparov 1992:51, 71–73). Our excavations in 1993 and 1994 produced only two identifiable fragments of canid

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

areas of origin of the crops and domestic animals 89

remains which did not permit differentiation between wolf, jackal, and dog (this volume, p. 177). Kasparov (1992) reported that many of the goat, sheep, gazelle, and other bones from Jeitun show signs of having been gnawed by dogs and he assumed that the dogs were mainly used in hunting. During excavation in 1991, dog bones were found with a pottery vessel in a niche in a wall of one of the houses, suggesting deliberate burial of a domestic dog. Small clay figurines from Jeitun that represent animals provide further, tentative evidence of domestic dogs (Fig. 8.12, color). Kasparov (2000) analyzed similar animal figurines from Late Chalcolithic levels at the sites of Ilgynly-depe (70 figurines), Altyn-depe (13), and Kara-depe (13) on the Kopetdag piedmont east of Jeitun. Most of the identifi- able figurines are of bulls, but Kasparov also identified 7 as dogs, which in style and proportions resemble the dog-like figurines found at Jeitun.

Bones of domestic dogs were reported from Middle and Late Jeitun-Culture levels at Chagylly in the eastern piedmont by Berdiev (1966:27), and from the Chalcolithic IA and IB strata at Anau North where the remains varied in size from medium to large dogs (Moore et al. 2003:155, 157). Four bones identified as from domestic dog have also been found in the early Neolithic (Dariasai phase) levels at the site of Ayak- agytma in the Kyzylkum (Lasota-Moskalewska et al. (2006:208). Although the bone and figurine evidence is sparse, it is sufficient to demonstrate the presence of domestic dogs at Jeitun and other sites in the Neo- lithic and to suggest that they were used in hunting wild game, for herding domestic sheep and goats, and perhaps as guard dogs and pets.

Despite the existence today of over 400 distinct breeds, all domestic dogs are believed to be descen- dants of the wolf (Canis lupus) which ranged until recent times throughout Eurasia and North America. In Eurasia, both the larger subspecies of northern latitudes and the smaller wolves of Southwest, Central, and South Asia may have contributed to the ancestry of domestic dogs in a long process of taming and domestication that began at least 15,000 years ago (Clutton-Brock 1999:56–58). This view, which was originally based on anatomical and behavioral simi- larities between wolves and dogs and on archaeological evidence of dogs in Late Palaeolithic and Mesolithic contexts (see, for example, Clutton-Brock 1995), has since been borne out by genetic studies (Savolainen et al. 2002; Vilà et al. 1997; Vilà, Maldonado, and Wayne 1999; Wayne, Leonard, and Vilà 2006).

Results from an analysis of mtDNA from a sam- ple of 162 wolves from 27 populations in Eurasia and North America and of 140 dogs representing 67 breeds by Vilà et al. (1997) strongly supported the view that the wolf was the exclusive ancestor of domestic dogs and showed that more than one wolf population had contributed to their ancestry, implying either that wolves were domesticated independently in several parts of the world at different times or that there was one initial domestication followed by episodes of admixture between dogs and wolves. Subsequently, a larger-scale analysis of mtDNA from a worldwide sample of 654 domestic dogs undertaken by Savol- ainen et al. (2002) found that mtDNA diversity pres- ent in East Asian dogs was much greater than in dogs from other areas, implying a common origin of all dog populations from a single gene pool. The greater genetic variation in East Asia and the pattern of phylo- geographic variation suggested that dogs were domes- ticated approximately 15,000 years ago in East Asia, although a much earlier date of c. 40,000 years ago could also be inferred from the data. However, Wayne et al. (2006:283) subsequently argued that Savolainen et al. did not consider other factors that might have influenced the genetic diversity found in East Asian dogs and maintained that the evidence presented for the location and time of origin of domestic dogs was not conclusive— a conclusion since reinforced by an analysis of mtDNA diversity in 318 African village dogs, which was found to be similar to the diversity of the East Asian dogs sampled (Boyko et al. 2009).

An East Asian locus of domestication is not at present supported by any definite archaeological evidence of domestic dogs in the Palaeolithic record of the region, although the remains of a possibly domestic dog have been found in the lowest cultural layers of a site in northern China (Nanzhuangtou) dated to approximately 12,000 BP (c. 14,000 cal. BP; Cohen 1998:25; Underhill 1997:113–14). Domestic dogs, sometimes buried in human graves, have been found in many Chinese and other East Asian Neolithic sites (Underhill 1997:121–28; Olsen 1985:48–70) and at the Neolithic site of Burzahom in Kashmir (Allchin and Allchin 1982:113). Also, at the large Chalcolithic site of Botai in northern Kazakhstan, where (as men- tioned above) there is evidence of intensive exploi- tation of horses, dogs are the only other domestic animal found. Their bones comprise the second most abundant type of animal remains and most were found buried in pits in or close to houses. Olsen, who

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

90 origins of agriculture in western central asia

studied the bone assemblage (2000; 2003:99–100), showed that in their size and morphology the Botai dogs resemble the modern Samoyed breed, and she inferred from the placement of the burials in the foundations and at the thresholds of houses that dogs may have fulfilled a ritual role as guardians.

The—very limited—archaeological evidence of domestic dogs in Palaeolithic contexts comes from much farther west, in Europe and the Levant. For example, two skulls of adult dogs were found at the Upper Palaeolithic site of Eliseevichi I on the central Russian plain west of Bryansk and radiocarbon dated to between c. 13,000 and c. 17,000 BP (c. 15,500– 20,000 cal. BP; Sablin and Khlopachev 2002); a dog mandible was found in a Late Palaeolithic human grave at Bonn-Oberkassel in Germany dated to c. 14,000 BP (c. 17,000 cal. BP; Nobis 1979); and canid remains identified as dogs have been reported from Epipalaeolithic (Natufian) contexts at several sites in Israel dated to c. 12,000–10,000 BP (c. 14,000–11,500 cal. BP), notably at Hayonim and Ein Mallaha where the dogs were buried with humans (Davis and Valla 1978; Dayan 1994; Tchernov and Valla 1997). Remains of domestic dogs have been found at several early Neo- lithic sites around the Fertile Crescent, for example at Jericho, Mureybet, Cayönü, and Jarmo (Clutton-Brock 1979:140–41; Lawrence 1967; Lawrence and Reed 1983; Peters et al. 1999:38). However, it should not be assumed that these early Neolithic dogs were necessar- ily descended from the Natufian dogs of the Levant because, as Tchernov and Valla point out (1997:66), it is possible that the Neolithic dogs of Southwest Asia were either domesticated anew in the region or intro- duced from elsewhere.

As wolves were present in the Late Pleistocene/ Early Holocene across the whole of Asia, zoogeo- graphically it is possible that dogs could have been do- mesticated anywhere between the Levant and China, including western Central Asia. There is, however, no conclusive evidence of domestic dogs at Late Palaeo- lithic/Mesolithic sites in the region. Coon (1957:155) found the muzzle of a very large dog in the lowest Me- solithic levels at the site of Ghar-i Kamarband, which he said resembled that of a St. Bernard dog and he in- ferred that domestic dogs were used in hunting during the Mesolithic. But H-P. Uerpmann, who analyzed the bone assemblage (Uerpmann and Frey 1981), thought the specimen more likely to be from a wolf (Uerp- mann, pers. comm. 2007). No canid remains were identified at the nearby site of Ali Tappeh nor from the

Mesolithic (or Neolithic) levels at the Bolshoi Balkhan rockshelters of Jebel (Tsalkin 1956) and Dam Dam Cheshme, although according to Markov (1966a:123), Tsalkin identified several dog bones from the second (Bronze Age) layer at Dam Dam Cheshme 2.

The apparent absence of remains of domestic dogs in the Caspian Mesolithic runs counter to a hy- pothesis of independent domestication from local wolf populations, but this conclusion must remain tentative in view of the relative lack of zooarchaeological re- search in western Central Asia. On present evidence, it seems probable that the existence of domesticated dogs at Jeitun-Culture sites (and at Ayakagytma in the Kyzylkum) is due to their having been introduced into the region during the Neolithic rather than the result of their local domestication from wolves.

Conclusion

The foregoing review of archaeological, bio- geographical, and genetic evidence concerning the areas of origin of the crops and domestic animals of Neolithic Turkmenistan and adjacent areas reveals how little we definitely know about how they were in- corporated into the agricultural and pastoral systems that developed from c. 6000 cal. BCE (the beginning of the Early Jeitun phase). Without much more ar- chaeobotanical and zooarchaeological research at Late Palaeolithic/Mesolithic and Neolithic sites in the region, together with accurate identification and direct (AMS) radiocarbon dating of the plant and ani- mal remains found, it is difficult to reach well-founded conclusions about the central question of whether all the crops and domestic animals of the Jeitun Neolithic were introduced from elsewhere or whether some of them were independently domesticated in the region. The rapidly accruing genetic evidence on the ori- gins of crops and domestic animals is beginning to provide valuable insights into the question, but close correlation between it and the biogeographical and archaeological evidence is difficult to achieve. Given the present inadequacy for western Central Asia of all three types of data, we cannot answer the question definitively, but by weighing the probabilities, already discussed, of local versus external domestications, it is possible to reach general, if tentative, conclusions.

Regarding the cereal crops that were definitely present at Jeitun in the Early Neolithic, we can be con- fident that einkorn (and the tentatively identified free-

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

areas of origin of the crops and domestic animals 91

threshing wheat) were not domesticated locally, and although less is known about the ancestry of the other glume wheat, it almost certainly reached Jeitun as a domesticate from somewhere west of the Caspian. The phytogeographical and recently acquired genetic evi- dence that eastern forms of wild barley occur in Central Asia (as well as feral domestic barley) demonstrates that barley could have been domesticated independently in southern Turkmenistan, but its widespread distribu- tion as a Neolithic crop around the Fertile Crescent long before it was present at Jeitun suggests that, like the wheats, it is more likely to have reached southern Turkmenistan as an already domesticated crop.

The two domestic ungulates, sheep and goat, parallel wheat and barley in that the sheep appears to have been domesticated (from the mouflon) outside the region and subsequently dispersed eastward into Central Asia, whereas the distribution of the bezoar suggests that it could have been independently domes- ticated in the region. However, Naderi et al.’s (2008) analysis of bezoar mtDNA sampled across the wild goat’s present range found no support for possible domestication anywhere east of central Iran, so the domestic goats too are more likely to be descended from stock domesticated in Southwest Asia. The dog could have been domesticated locally from native wolves, but the lack of evidence of domestic dogs at Late Palaeolithic/Mesolithic sites in western Central Asia, compared with early evidence of them in South- west Asia, tends to suggest otherwise.

The evidence of the other domestic animals that were present in Turkmenistan and adjacent

areas in prehistoric and early historical times, but not found at Jeitun itself, reveals similar uncertain- ties. Pigs and cattle could have been domesticated locally from wild boar and aurochs, but the weight of archaeological and genetic evidence is against the possibility and in favor of both animals having been introduced as domesticates from Southwest Asia. Horses and Bactrian camels conclusively identified as domestic do not appear in the archaeological record until the Chalcolithic and the Bronze Age, but the wild progenitors of both are likely to have been pres- ent during the Late Pleistocene/Early Holocene and so might have been domesticated locally. Indeed, the discovery of horse and camel bones in early Neolithic and later levels at Ayakagytma in the Kyzylkum could be interpreted as evidence of local domestication, but the question of whether the remains derive from domesticated animals remains to be resolved. Finally, the dromedary appears only to have been introduced into Central Asia, as a domesticate, in the 1st millen- nium AD.

note 1. The Pre-Pottery Neolithic (PPN) period in the Levant

lasted approximately three millennia and is conventionally

divided into the PPNA and PPNB. The latter is subdivided into

Early, Middle, Late, and Final PPNB (sometimes referred to

as the PPNC). Approximate calibrated BCE dates for the divi-

sions (after Peters, von den Driesch, and Helmer 2005:535) are

as follows: PPNA 9500–8700, Early PPNB 8700–8200, Middle

PPNB 8200–7500, Late PPNB 7500–7000, Final PPNB (PPNC)

7000–6500, succeeded by the Pottery Neolithic.

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms

This content downloaded from ������������130.65.109.155 on Sun, 17 Jan 2021 03:52:38 UTC�������������

All use subject to https://about.jstor.org/terms