Part 2. Near Eastern Crop Diversity and its Global Migration

Diversity of Major Cultivated Plants Domesticated in the Near East - A.B. Damania
The Spread of Neolithic Agriculture from the Levant to Western Central Asia - D.R. Harris
The Spread of Agriculture to the Eastern Arc of the Fertile Crescent: Food for the Herders - F. Hole
Early History of Sesame Cultivation in the Near East and Beyond - D. Bedigian
Grain Legumes: Evidence of these Important Ancient Food Resources from Early Pre-agrarian and Agrarian Sites in Southwest Asia - A. Butler

Diversity of Major Cultivated Plants Domesticated in the Near East - A.B. Damania

Our fathers planted gardens long ago...
Whose fruits we reap with joy today;
Their labor constitutes a debt we owe...
Which to our heirs we must repay;
For all crops sown in any land...
Are destined for a future man.
Arab Poet - Nizami

Introduction

Alexander von Humboldt was probably the first author to refer to the question of origins of crops in his work Essai sur la Géographie des Plantes in 1807. Possible reasons for humans to have abandoned their hunting/gathering ways and settled down to the sedentary pastoral life of a cultivator have been discussed by several authors, among them Ucko and Dimbleby (1969), Harlan et al. (1976a), Zeven and de Wet (1982), Smith (1995), Harris (1996) and Diamond (1997).

Alphonse de Candolle, in his 1882 book Origine de Plantes Cultivées, was among the first to indicate regions where plant domestication may have taken place: China, Southwest Asia including Egypt, and Tropical Asia. In 1926, during the Fifth International Genetics Congress in Berlin, Vavilov expounded his theory of centers of origin of crop plants for the first time. The centers recognized by Vavilov were: China, India, Indo-Malaya, Central Asia, Near East, the Mediterranean, Ethiopia, Southern Mexico and Central America, South America and Chile. Vavilov continued to work on his theory until his death in 1943.

Harlan et al. (1976b) refer to the Near East as the “center of agricultural innovation” where barley was the first crop to be domesticated followed by wheat. Later the other 'founder crops' such as pea, lentil, vetch, faba bean, flax, tree and vine fruits were domesticated and the entire system moved out of the nuclear area together with an array of agricultural techniques. The system spread, moving along the shores of the Mediterranean and up the banks of the Danube River and down the Rhine, eastward to the Indus and northern India and southward across Arabia, the Yemen and into the Ethiopian plateau. It did not proceed further down into tropical Africa. It reached China in the second half of the 2nd millennium BC (Harlan et al. 1976b).

Arab geographers, authors of garden books and other writers from the 10th century AD onward tell about a countryside in the Islamic world that had changed significantly since ancient times (Watson 1983). The 'new' crops were mostly fruit trees, grains other than wheat and barley, and vegetables and several others. These introductions greatly increased the diversity of crops that appealed to different tastes. Watson (1983) gives a few examples of the diversity available at that time: the writer Al-Jahiz in the 9th century AD stated that there were 360 varieties of dates (Phoenix dactylifera) to be found on sale in the market at Basra in Iraq, while in the following century Ibn Rusta stated there were 78 kinds of grapes (Vitis vinifera) being grown in the vicinity of San'a (Yemen). Al-Ansari, writing about a small town in North Africa, stated that in the year 1400 AD there were 65 kinds of grapes, 36 kinds of pears, 28 kinds of figs, 16 kinds of apricots, and so on. The Arab conquests during the 7th and 8th centuries AD greatly facilitated the introduction of new varieties of crops in the Near East.

This paper focuses on the crop diversity of the above region, an important area recognized by most authors as one of the major centers of origins of crop plants and genetic diversity as well as innovation of agricultural techniques. However, this Vavilovian center overlaps in places with the Mediterranean Center and hence crops of the latter with wide distribution in the Near East are also included for comprehensiveness. Much of the information given here has been adapted from Zeven and de Wet (1982) and Vavilov (1992). Brief descriptions of selected major crops which have their greatest diversity and/or were domesticated in this region are given below in alphabetical order of their botanical families.

Major crop plants domesticated in the Near East

The genus Allium (Family: Alliaceae) includes the common garlic and the onion which form the very basis of Asian and Mediterranean food preparations. Allium ampeloprasum or sweet leek (Levant garlic) is found all over Europe, Asia Minor and North Africa. Its presence in the Channel Islands and western France is thought to represent relics of former cultivation rather than recent introduction. According to Vavilov, the primary center of origin of onion is in Central Asia and that of leek in the Near East and the Mediterranean. Allium ascalonicum or shallot is closely related to A. deserti-syriaci from the deserts of Syria and Iraq. It has also been described as a variant of A. cepa, the common onion. Allium kurrat or salad leek originated in Sinai and is grown in the Nile delta, Palestine and Arabia for its leaves. Allium porrum or leek has its greatest diversity in Asia Minor and is derived from A. ampeloprasum, its wild progenitor. And finally, A. sativum or garlic has its primary center in Central Asia but a secondary center in Southwest Asia. Wild Allium species often invade cereal fields all over Southwest Asia. They can be distinguished quite simply by their deep purple or light green inflorescence.

Pistacia vera (Anacardiaceae) or pistachio is native to the Near East and Southwest Asia where it has been cultivated for over 4000 years. Kernels of naturally occurring forms are known from sites long before that. Secondary centers of diversity have developed in Iran and Syria. Relic populations of former extensive cultivation are found in Turkmenistan and Kyrgyzstan. Another species, P. terebinthus, is used as an alternative pollinator and as rootstock on which young male and female pistachio trees can be grafted.

Carthamus tinctorius (Asteraceae) or safflower has its origin in the Near East according to Ashri and Knowles (1960) since the cultivated form is similar to two closely related species: C. flavescens found in Turkey, Syria and Lebanon (usually as a weed in wheat fields), and C. palaestinus found in the deserts of Western Iraq and Palestine (see McCorriston, this volume). In these areas introgression between the wild and cultivated forms may still occur, leading to an exchange of genes. One species in the group, C. leucocaulos, is distinctly different from other species. Stems are smooth, head size is small, colored flowers are light purple or light pink instead of purple, and self-fertility is high. It is endemic to the Aegean Islands, but has been introduced to southern Europe, Australia and the USA. Carthamus divaricatus, which is endemic to Libya, is the only species which is morphologically distinct. It has horizontal branches; strongly divaricate outer involucral bracts; yellow, purple and white flowers; yellow pollen, and dark purple striped anthers. Carthamus oxyacanthus, found in continental areas from western Iraq through to northwestern India, and extending northward to southern former USSR, in all areas is a very serious weed. Two other species also found are C. gypsicolus, similar to C. oxyacanthus and restricted to the former USSR, and C. curdicus, with characteristics of both C. gypsicolus and C. flavescens and restricted to northern Iraq. All the above have 12 pairs of chromosomes. There are two species with 32 pairs of chromosomes, C. turkestanicus and C. baeticus. Carthamus turkestanicus ranges from Turkey in the west to northern Pakistan and Kashmir, India in the east, with presumably introduced populations in Ethiopia. Carthamus baeticus may be found around the entire Mediterranean Sea and in its islands. All wild species of safflower are spiny weeds, some of them very serious because they occupy fields sown to other crops. Others are more prevalent in roadsides and waste places. They are characterized by yellow flowers, although white flowers with yellow pollen grains are not unknown. At an early stage in its evolution, C. tinctorius spread to Egypt, Ethiopia, southern Europe, South Asia and the Far East where distinctive types have evolved.

Several wild Beta spp. (Chenopodiaceae), which are relatives of Beta vulgaris or beetroot, are widespread in this region. Among them are B. corolliflora found in Turkey, Azerbaijan, Georgia and in parts of Iran. It is known for its resistance to frost and is often used in crosses with cultivated species. Beta intermedia is said to have evolved as a result of a cross between B. lomatogona and B. trigyna, two other wild forms. Beta lomatogona is itself a weed found all over Asia Minor and B. trigyna is found around the Black Sea with a few samples found as far as the Caspian Sea, Ukraine and Hungary. Beta macrorrhiza is found at higher altitudes in the mountains of Iran, Turkish Armenia near Lake Van and the Caucasus. It is a cold-resistant species, sweet tasting with a white pulp. The sugar beet, B. vulgaris, has its origins in the western Mediterranean.

Spinacia oleracea (Chenopodiaceae) or spinach is a native of Southwest Asia but has become well established almost all over the world, having been introduced in Europe, the Americas and the Far East.

Cichorium intybus (Compositae) or chicory is a perennial salad plant native to the parts of southern Europe bordering on the Mediterranean. It was cultivated in Greek and Roman times and has become naturalized in parts of North America. Cichorium endivia or endive is indigenous to the eastern Mediterranean.

Cynara scolymus (Compositae) or artichoke was probably developed from its wild progenitor C. cardunculus, a native of the Mediterranean whose stalks and leaves are edible. There are several words for the plant in the Berber language, which indicates the possibility of its origin in the Maghreb. It moved in to Europe in the Middle Ages.

Lactuca sativa (Compositae) or lettuce evolved as a natural selection from its closest wild progenitor, L. serriola, which is a prickly biennial of the temperate regions (Purseglove 1968). Lactuca sativa originated in the eastern Mediterranean about 7000 BP, most likely in Egypt when the climate of that part of northeastern Africa was not as dry as it is today. Lettuce was a common oilseed crop in Upper Egypt as far back as 6500 BP as depicted by rosettes of a tall, large vegetable with subulate leaves painted on the walls of tombs at Thebes (2500 BC). The presence of latex in primitive lettuce also marked the plant as a fertility symbol. Unfortunately, the available illustrated record found in ancient monuments is not complemented by dry or carbonized remains of leaves or seed of this crop in Egypt.

Early human selection for non-shattering seed heads, late flowering, non-prickly leaves, decrease in latex content, and hearting character (the tendency of leaves to congregate in layers in a heart-shaped head) is said to have led to its domestication as a leafy salad vegetable. Lettuce reached China around the 7th century AD where a special cultivar of the asparagus type (L. Serriola f. integrifolia) was developed and reported by Fuchs in 1543 from Sinkiang. The local name of this variety is 'usum' or 'on-sen'. According to Vavilov (1992) this form of lettuce is practically unknown in Southwest Asia.

Cornus mas (Cornaceae) or Cornelian cherry is found all over Caucasia and Asia Minor as an underbrush of deciduous forests. The fruits are edible and are used to produce an alcoholic beverage, vin de Cornouille. The fruits are of great antiquity and have been reported from archaeological sites over 10,000 BP in 1991 from the PPNB site of Nevali Çori, near Urfa, in southeastern Turkey (see Pasternak, this volume).

Corylus avellana (Corylaceae) or European hazelnut has its primary center in the Caucasus where a number of other Corylus species are also found. However, it was domesticated relatively recently in 17th century Italy. Corylus colurna or Turkish hazelnut or cob nut is cultivated in Turkey for export and also used as rootstock for C. avellana since it possesses resistance to several diseases. Hybrids between the two are called 'trazels'. Corylus maxima or filbert nut tree is native to West Asia, southeast Europe and the Caucasus.

Brassica oleracea (Cruciferae) or cabbage has a secondary center in Asia Minor where convar. oleracea and capitata are found. There are wild as well as cultivated forms. Some of the wild inedible forms are also used as ornamentals. Brassica oleracea var. botrytis or cauliflower is a native of the eastern Mediterranean

Cucumis melo (Cucurbitaceae) or musk melon (cantaloupe) has its origins in Africa but has an important secondary center in the Near East where some of the sweetest fruits are found. Similarly, Cucumis sativus or cucumber, a native of India, also has a secondary center in this region and cucumbers are part of all Near Eastern cuisines from Turkey through to Iran and Afghanistan. In Iran tender cucumbers are eaten with yogurt (mást-khiyár) on a daily basis.

Ricinus communis (Euphorbiaceae) or castor was known to the ancient Egyptians as far back as 6000 BP. Its oil was used as an illuminant. Although it originated in Africa, it grows wild in the Yemen and all of the Near and Middle East. It was taken at an early date to India. Today castor oil has great commercial value in the paint and lubricant manufacture industries. The oil also has been used as a purgative (home remedy) since ancient times.

Castanea sativa (Fagaceae) or chestnut is found from southern Italy to Asia Minor and western Georgia in the former USSR. It was introduced to southern Europe by the Greeks and the Romans, and then transported to northern Europe and Britain. Nuts are roasted, the shells removed and the insides eaten or used in confectionery.

A number of Aegilops spp. (Gramineae) or goatgrass are native to West Asia and the Near East. Aegilops was known to the ancient Greeks and the name is related to its supposed healing properties of an eye disease from which goats suffer (van Slageren 1994). Ae. columnaris is spread from Turkey through Iraq, Iran and Caucasia. It is a common weed found in and on the borders of fields. Ae. crassa is also found in Turkey, Syria, Palestine, Iraq, Iran and Afghanistan. Ae. cylindrica is found in the Balkan peninsula, Crete, Turkey, Caucasia, Armenia, Azerbaijan, Iran, Iraq and Afghanistan. It is a weed in fallow fields and along slopes of hillsides. Ae. kotschyi is found from North Africa across Palestine, Iraq, Iran, Afghanistan and Caucasia. Ae. laurentii has a wider distribution, i.e. from southern Europe, former USSR, Turkey, Palestine, Iraq and Iran. Ae. mutica is restricted somewhat to Armenia and Anatolia in Turkey and Ae. ovata is just the opposite, being found all over the Mediterranean, Palestine, Syria, Lebanon, Turkey, Iraq, Iran and Afghanistan. A weed of cultivated wheat fields, it has a male-sterilization action on the nucleus of Triticum aestivum and T. turgidum. Ae. speltoides has its primary center in southern Turkey and northern Syria and Iraq. It is less common in places bordering the above. It is often found growing with wild wheat (T. boeoticum) in southern Turkey and northern Syria. Ae. triaristata is found in the Mediterranean, West Asia, Iraq, Iran and southern parts of the former USSR. Ae. triuncialis is all over the Mediterranean area, Turkey, Palestine, Syria, Lebanon, Iraq, Iran, Turkmenistan and Afghanistan. It has been suggested that it is a hybrid between female Ae. caudata × male Ae. umbellulata that originated in West Asia. Ae. umbellulata, on the other hand, occurs on moist steppe, dry slopes of hills, and is a weed in cultivated fields in the Greek islands, Turkey, northern Syria, Iraq, northwestern Iran and Transcaucasia. It is resistant to leaf rust and hence used often in wheat breeding. Some Aegilops spp. cross readily with others, e.g. Ae. sharonensis, whereas others are cross-pollinators which readily produce hybrids. Aegilops spp. that occur on the borders of cultivated Triticum fields often exchange genes although the hybrids are mostly sterile.

Hordeum vulgare (Gramineae) or barley is one of the oldest crops to be domesticated in this region, from its wild progenitor H. spontaneum. The wild progenitor of barley is found in the eastern Mediterranean, West Asia, and as far as Turkmenia and Afghanistan. Domestication of barley is said to have taken place around 9000 BP in the Fertile Crescent. However, grains of barley have been found in Egyptian tombs dating from 15,000 BP. But it would be difficult to distinguish between wild H. spontaneum and its domesticated form from looking at carbonized remains. Other finds include those discovered in Cayönü in Turkey dating from 9000 BP (Braidwood et al. 1969). The domesticated barley spread through the Mediterranean to Rome and beyond as well as eastwards through Iran and Afghanistan into India and from India to China through Tibet. There are several forms but they all have a brittle rachis and very long awns. For some time H. agriocrithon, a brittle rachis six-row form first reported from Tibet, was thought to be the progenitor of six-row barley. But later this theory was discounted as such brittle rachis forms were also found where H. spontaneum came into contact with six-row barley. These forms mature about two weeks before the cultivated six-row barley and occur as weeds in barley fields all over Tibet. They are easily distinguished from cultivated forms by their purple to black spikes. However, H. agriocrithon, unlike H. spontaneum, has not yet been found growing by itself in the wild state away from cultivated fields.

Avena sativa (Gramineae) or oats. Two hexaploid oats have been recognized: A. sativa and A. byzanthina. Avena sterilis, the wild oat, is recognized as the progenitor with its center of diversity in the Near East and the Mediterranean. Wild oats are perpetual invaders of cereal fields in Southwest Asia and elsewhere. Other wild species found in this region are: A. damascena (found in an area 60 km north of Damascus in Syria), A. clauda, A. longiglumis, A. moroccana (found in Morocco and sometimes mistaken for A. sterilis) and A. strigosa (black oat, found all over the Mediterranean).

Secale cereale (Gramineae) or rye has its primary center of origin in northeastern Turkey and northwestern Iran. A secondary center has been suggested in Afghanistan from where the crop may have migrated to central Asia and Europe. A number of genetic variants exist, one of which (S. segetale) evolved into the annual wild rye weed (see Jaaska, this volume). Hybridization between cultivated wheat and rye has resulted in the triticales which have resistance to several diseases. Cultivated rye is known from the Neolithic age in Austria, but it seems to have become widespread in Europe only after the Bronze Age.

Secale montanum (Gramineae) or mountain rye is a perennial species found in the central Atlas Mountains of Morocco, Sierra Nevada Mountains of Spain, Italy, Sicily, Yugoslavia, Greece, Lebanon, Turkey, Iran and Iraq. It is highly polymorphic and crosses readily with other Secale spp., producing several weedy forms and varieties. Other species include S. silvestre and S. vavilovii, which occurs on the slopes of Mt. Ararat and along the banks of the Araks River.

Sorghum halepense (Gramineae) or Johnson grass, a relative of the cultivated sorghum (S. bicolor), is found from the Mediterranean up to Pakistan. It is a rhizomatous perennial first identified near Aleppo (Halep) in Syria. It occurs as a weed on the borders or in cereal fields. It is a source of disease resistance for cultivated sorghum with which it crosses readily.

Triticum aestivum (Gramineae) or common wheat or bread wheat has its primary center in Transcaucasia and adjacent areas and is by far the most important crop that this region has given the world (see Dvorak, this volume). Natural cross-fertilization still takes place between wild and cultivated Triticum spp. and many authors believe that wheat is still evolving. It is said to have arisen as a result of a cross between T. dicoccum (emmer wheat) and Ae. squarrosa. This cross must have taken place after emmer evolved from its wild progenitor, T. dicoccoides, in the area south of the Caspian Sea. Several other hexaploid forms eventually evolved and they are as follows: T. vavilovii, a species peculiar to Armenia, T. spelta (spelt wheat) is still cultivated in parts of Spain, T. macha (macha wheat), T. compactum (club wheat, was grown in Syria until the last century and volunteers can still be observed in cultivated fields of T. aestivum) and T. sphaerococcum (dwarf shot wheat with globoid grains native to India, where it is still grown in small pockets). A secondary center of diversity for hexaploid wheat developed in the Hindu Kush Mountains where the crop encountered a physical barrier to further migration, i.e. the theory of peripheral diversity (Yamashita 1980).

Triticum monococcum (Gramineae) or einkorn was domesticated in the Fertile Crescent from one of its wild progenitors, T. boeoticum or T. urartu. Its earliest recorded archaeological find is from the site at Ali Kosh in Iraq dating from 8500 BP. Triticum boeoticum subsp. thaoudar, the wild progenitor with two grains and two awns, has been reported growing in wild stands from Diyarbakir in Turkey and remains have been found from the site at D'jade on the Euphrates where it was grown and harvested in the green state. Triticum monococcum, although considered obsolete, is still cultivated as a feed for poultry and swine in the mountainous villages in Italy, Spain and Turkey and elsewhere.

Triticum timopheevi (Gramineae) consists of two subspecies: araraticum and timopheevi. The former grows wild in the northeastern arc of the Fertile Crescent. The latter species, however, is part of what is termed the Zanduri wheat, cultivated in Georgia in the former USSR. This type of wheat is formed by diploid einkorn, tetraploid timopheevi and hexaploid zhukovskyii wheats. Neither subspecies crosses readily with other wheats.

Triticum dicoccoides (Gramineae) or wild emmer is found all over the Fertile Crescent. Recently wild stands have been reported from Suweida province south of Damascus in Syria. Wild emmer has been used in crosses with cultivated tetraploid wheat to improve disease resistance of the plant and protein content of the grain. Several cultivated tetraploid wheat forms may have been derived from wild emmer: T. dicoccum (emmer wheat), T. turgidum (poulard wheat), T. carthlicum (Persian wheat), T. durum (macaroni wheat), T. polonicum (Polish wheat) and includes T. ispahanicum, T. turanicum (Khurasan wheat) and T. turgidum subsp. turgidum (English wheat).

Triticum dicoccum (Gramineae) or emmer wheat is, according to some authors, the oldest cultivated wheat, being domesticated in the Fertile Crescent around 10,000 BP. Until recently, it was cultivated in large quantities in Ethiopia (the secondary center of diversity for tetraploid wheats), Iran, eastern Turkey, Transcaucasia, Volga Basin, Yugoslavia, Czechoslovakia, Italy, Spain and India. It was seen from frescoes and wall paintings at Thebes and Luxor that T. dicoccum was the favored cereal used for bread-making in Egypt. But today its cultivation is on a decline and it can be found only in small pockets of traditional farming communities in the above areas. In India it is called 'khapli' wheat and is now grown only in limited areas in the state of Rajasthan.

Triticum durum (Gramineae) or macaroni wheat is cultivated all over West Asia and North Africa (WANA) and the Mediterranean. In fact, 80% of all the durum wheat grown in the world is cultivated in this region. It is rarely grown outside the WANA region in the Old World. Its price in the world market is higher than bread wheat because of its great demand in the making of pasta dishes in Italy and elsewhere. Ancient texts speak of the possibility of storing durum wheat over long periods because of the low water content of its grains.

Triticum zhukovskyi (Gramineae) or Zanduri wheat is cultivated in western Georgia in the former USSR. It is a hexaploid with a unique genome formula. The designation 'Zanduri', however, is reserved for a mixture of T. monococcum, T. timopheevi subsp. and T. zhukovskyi.

Cicer arietinum (Leguminosae) or chickpea, also referred to as garbanzo, occurs in two forms: the Mediterranean type (plants with green stem, white flower, owl-head shaped and salmon-white seeds) and the desi type (plants with anthocyanin in the stem and leaves and small seeds with an angular shape and darker color). Cicer reticulatum, a wild annual species found in Turkey, could be the wild progenitor since it produces fertile progenies when crossed with cultivated chickpea.

Lens esculentum (Leguminosae) or lentil. The genus Lens comprises two species: (1) L. culinaris (syn. Lens esculentum), which includes three subspecies - culinaris (cultivated lentil), and the wild orientalis and odemensis, and (2) L. nigricans, with subspp. nigricans and ervoides.

The cultivated lentil has been domesticated from its wild progenitor subsp. orientalis (see Erskine, this volume). Wild lentils occur from western Europe (Spain and Portugal) to central Asia. According to Harlan (1951) there is a microcenter of diversity for lentils in Turkey.

Lathyrus sativus (Leguminosae) or grass pea was domesticated in West Asia with a center of diversity in the Mediterranean region.

Lathyrus tingitanus (Leguminosae) or Tangier pea has a microcenter of diversity in Morocco where it is cultivated as a winter annual.

Lupinus albus (Leguminosae) or white lupin is found wild in western Turkey but may have been domesticated in the Balkans. Other lupins from the Mediterranean area are: L. angustifolius, L. cosentini, L. luteus, L. pilosus and L. termis. Egyptian lupin has been found in Palestine and Egypt since ancient times. The seeds contain an alkaloid which has to be neutralized before consumption.

Medicago sativa (Leguminosae) or alfalfa evolved around the Caspian Sea where it occurs as a wild plant. It is believed that it was the first plant to be cultivated as a forage crop to provide feed for horses and cattle. The Greeks came across alfalfa being used to feed the horses of the mighty Persian cavalry of Darius III around 350 BC. The Greeks quickly adopted the crop for their own use and it spread later into southern Italy and Europe and subsequently to other parts of the world. Several wild Medicago species occur in West Asia: Medicago cancellata, which is used as a gene source for adaptation of M. sativa to poor soils, Medicago dzhawakhetica, found in Georgia and parts of Asia Minor, and M. glutinosa, found in the Caucasus and also used as a gene source for crop improvement.

Pisum sativum (Leguminosae) or pea was domesticated from the wild P. humile, the latter being common in northern Iraq, Jordan, Syria, northwestern Iran, Palestine, Turkey and Cyprus. The crop reached the Greeks through the Black Sea and went eastwards to India and China via the Himalayan trade routes passing through Tibet. Pisum elatius is another wild form which is found all over the Near East and in southern Italy and Sardinia. It can also be found in scattered places in North Africa.

Trifolium ambiguum (Leguminosae) or Caucasian clover is found in Caucasia, Crimea and Turkey. It is a very valuable fodder plant. Material taken from Crimea to Australia gave rise to cultivars which can withstand long periods of flooding.

Trifolium israeliticum (Leguminosae) is a forage species reported from northern Palestine.

Trigonella fbenum-graecum (Leguminosae) or fenugreek probably also evolved in West Asia, but is cultivated in southern Europe, North Africa and India as a fodder crop. In India the seeds are also eaten and used in traditional medicine.

Vicia sativa (Leguminosae) or common vetch has a great variation in form, chromosome number and karyotype. Zohary and Plitmann (1979) described seven subspecies as follows: sativa, macrocarpa, nigra, cordata, incisia, amphicarpa and pilosa. All these subspecies are found in the Mediterranean and the Near East, extending up to southern and central Europe. Other useful Vicia species include V. ervilia or bitter vetch, with its primary center in the Mediterranean. It was known to be cultivated for food in Turkey (7750 BP) and Greece (7500 BP). Vicia narbonensis or the narbonne vetch has its primary center of origin in eastern Georgia and a secondary center of origin in the Mediterranean. Narbonne vetch occurs as a weed in barley and wheat fields in Transcaucasia as does V. villosa or hairy vetch which is found in central Europe, the Mediterranean region and West Asia.

Vicia faba (Leguminosae) or faba bean probably originated in Southwest Asia or the Mediterranean. Its wild progenitor has not been identified with certainty although several theories have been put forward but discounted for one reason or another. A center of diversity has been established in the Mediterranean region. Several finds of the small-seeded var. minuta have been found in the archaeological remains. The oldest documented record of an archaeological find of large-seeded faba bean comes from Iraq around 1000 AD which perhaps accounts for the late development of the large-seeded types we see today.

Linum usitatissimum (Linaceae) or flax (linseed) was probably domesticated almost around the same time (ca. 8200 BP) as wheat and barley in the Near East from where it spread to other parts of the Old World along trade routes. Its wild progenitor, L. bienne, does not occur in Central Asia which Vavilov thought to be a center of origin of flax. Linum bienne has two main races: a continental winter annual found in the semi-arid foothills of Iraqi Kurdistan and Iran, and an Atlantic-Mediterranean form found along the coastlines. The latter is a perennial also sometimes referred to as L. angustifolium. It has a very high oil content. Hence, during the domestication process, according to some authors, two forms emerged, one for fibre (flax) and another for its oil (linseed).

Gossypium herbaceum (Malvaceae) or short-staple cotton, a native of southern Africa (see Holubec, this volume), was probably introduced to Ethiopia and thence to Arabia and Balochistan. It is widely cultivated in Syria and Egypt. In Iran it evolved into a special race which spread to western India where it was the first cotton to be cultivated. However, owing to its short staple it was not much in demand compared with the Egyptian cotton G. barbadense which is also later in maturity. Another cotton species, G. incanum or hillcoat cotton, became established in the Yemen, no doubt having arrived there from Ethiopia. It is extremely drought resistant as is G. areysianum found in southern Arabia. Another theory suggests that the northwestern part of the Indian subcontinent was probably the cradle of cotton cultivation, from which an 'ennobled' plant diffused to other parts of Asia Minor, the Near East, Africa and Europe. The earliest archaeological finds of cotton are from Mohenjo-Daro in the Indus Valley 2300 BC (Watson 1983).

Cannabis sativa (Moraceae) or hemp originated perhaps near the Caspian Sea. The first use of this plant was for fibre. Hemp is reckoned to be one of the oldest cultivated non-food crops. Evidence of the use of hemp in China goes back to Neolithic times. The crop reached West Asia around 2000 BC and was taken to Europe around 1500 BC. It became well established in the Mediterranean region.

Ficus carica (Moraceae) or common fig, one of the most ancient fruits of the Near East, arrived there from South East Asia, probably along the trade routes. However, it developed a secondary center in the Mediterranean and Asia Minor. By 6000 BP figs were cultivated in Egypt. It also grows wild in these regions and all over tropical Asia.

Nelumbo nucifer (Nelumbonaceae) or the Indian lotus is not native to India but to wetlands of northern Iran. It spread to China, Japan and India where it is often cultivated for its rhizomes and fruits. It was also grown for some time in the eastern Mediterranean region but is not found there today.

Olea europaea (Oleaceae) or olive is one of the longest-living fruit-bearing trees in the world (see Besnard et al., this volume). It is one of the relics of an age when the Mediterranean had a somewhat tropical climate, i.e. during the mid-Tertiary period. Excavations in Palestine and elsewhere in the Mediterranean indicate that it is closely associated with the history of civilization, being the source of edible fruits and oil for over 10,000 years. There is little doubt that the olive tree was first domesticated in the eastern Mediterranean where wild Olea europeae subsp. oleaster forms thrive (Damania 1995). The Phoenicians, from such ancient cities as Ugarit on the Syrian coast, introduced the olive around 5000 BP to the west and a secondary center of diversity developed in the areas bordering the Aegean Sea. From there the tree spread to southern Italy and as far as Spain where groves known to be 1000 years old still exist.

Phoenix dactylifera (Palmae) or date palm is another tree plant whose cultivation goes back to ancient times in the Near East. This tree and its symbol are synonymous with Arabia. Its present natural distribution includes North Africa, the Middle East and the western coast of India. At one time in the past the genus Phoenix was widespread in Europe but retreated southward as the climate changed. Its cultivation goes back to Neolithic times. The earliest archaeological record dates back to 6500 BP from Egypt. Traces of date palm stones from 5000 BP have been found at achaeological sites in western India. Its wild ancestor is not known although it crosses readily with the two other species of palms - P. dactylifera and P. reclinata. Phoenix atlantica, or false date palm, could also be an ancestor or an intermediate form. Phoenix atlantica var. maroccana also produces edible fruits.

Papaver somniferum (Papaveraceae) or opium poppy is a native of West Asia. It is the source of opium which is obtained from dried latex which drips from cuts made in the unripe capsular fruits. Papaver setigerum, the wild progenitor, is found from Persia to the Balkans. Arab traders are credited with spreading the narcotic use of opium. The plant had reached China by the 18th century and was the principal cause of the Opium Wars.

Fagopyrum esculentum (Polygonaceae) or buckwheat was introduced into China, Tibet and northern India from Central Asia. It is cultivated as a grain crop and is highly nutritious. It is a popular food in Japan.

Rheum rhaponticum (Polygonaceae) or rhubarb of Europe and North America originates from Armenia and Azerbaijan around the Caspian Sea where several other species of Rheum are grown. In China and Tibet R. palmatum and R. officinale have been used in traditional medicine as laxatives since 2700 BC.

Punica granatum (Punicaceae) or pomegranate is the traditional fruit of the central Iranian plateau. It is also one of the most ancient fruit trees to be domesticated and is known to have been grown in the Hanging Gardens of Babylon. It has some cultural significance as well and is found in the central courtyard of every home in the central Iranian plateau. The field genebank maintained by the Agricultural Research Center at Yazd in central Iran has over 700 different types, some of which go back to antiquity (Damania et al. 1993). Its only related species is P. protopunica which is endemic to the island of Socotra (Yemen) in the Indian Ocean.

Cydonia oblonga (Rosaceae) or quince is native to southern Daghestan and Azerbaijan. It is cultivated in the Near East as a garden tree. Its fruits are used to make jam.

Malus pumila (Rosaceae) or apple and Pyrus communis (Rosaceae) or pear have their primary center of origin in Asia Minor, the Caucasus and Central Asia. The secondary centers of origin are more toward the east, perhaps in western China. However, both apples and pears have existed in Europe from prehistoric times.

Prunus armeniaca (Rosaceae) or apricot was thought by Vavilov to have its secondary center of diversity in the Near East. It is believed to be native to western China.

Prunus amygdalus, syn. Amygdalus communis (Rosaceae) or almond, has its primary gene center in Central Asia as well as the Near East region and is a fruit species of great antiquity. Syria has several cultivars. Amygdalus fenzliana or fenzel almond is found in Anatolia and is grown as an ornamental. Amygdalus persica or peach has its center of origin in China but also a secondary center in Afghanistan and northeastern Iran. The landraces in the native habitats have better taste but also have some bitter forms and are not uniform in shape and size and hence not preferred by commerce.

Prunus avium (Rosaceae) or sweet cherry comes from Asia Minor and wild trees can be found in West Asia and North Africa. Domestication is said to have occurred in different places, leading to several ecotypes. It is interesting to observe that the sweet forms were domesticated whereas the bitter ones were left to grow unprotected in the wild. Another Prunus spp. native to the Near East is P. cerasifera or cherry plum which grows wild in Caucasia, Iran, Asia Minor, the Altai and Central Asia. It could be one of the parents of P. domestica or the garden plum. Prunus spinosa or blackthorn or sloe is found all over the Near East, North Africa and the southern Mediterranean.

Pyrus spp. (Rosaceae) or pears. The Near East is the main geographic center of origin for this genus. Pyrus caucasica, P. syriaca and P. takhtadzhiana are some of the other species of Pyrus found in this region. Hybridization among the wild species is common and the progenies are appreciated as ornamentals, rootstocks for cultivated species, and increasingly in recent years as donors of genes of disease resistance in fruit-breeding programs, particularly in California.

Coffea arabica (Rubiaceae) or Arabic coffee has its primary center in southwest Ethiopia, where it was chewed before being used for making a beverage. From Ethiopia it migrated to the Yemen. The discovery that coffee could be brewed into a beverage was made in Arabia around the 15th century. A single coffee plant shipped by Dutch plant explorers to Amsterdam from the East Indies in 1706 was nurtured in the botanical garden of that city. Amsterdam was one of the major world centers of maritime trade in those days. Some of the seedlings from this single plant were taken via France to the Caribbean island of Martinique and from there to South America. Hence the genetic base of the Latin American coffee is indeed very narrow and devastating attacks of diseases are common. The African form C. canephora often provides genes for disease resistance. Besides Latin America, coffee is also cultivated in Kenya, India, Java and other parts of the East Indies. Coffee is said to be the second most valuable commodity in world trade after petroleum products.

Citrus medica (Rutaceae) or citron is probably the only citrus native to West Asia, although the northeastern part of India (Assam and the Indo-Burma border area) has been mentioned as a center of origin of citrus species. Citrus medica var. ethrog is used by people of the Jewish faith at the Feast of Tabernacles. Citron was grown in Sardinia and the area around Naples after the 3rd century AD.

Cuminum cyminum (Umbelliferae) or cumin is a native of the Levant whose fruits are used as spice in cooking, for flavoring soups, and are an essential ingredient of curries. Its essential oil is also used in perfumery. It is also grown extensively in the drier parts of western India.

Daucus carota (Umbelliferae) or carrot. Wild forms occur in West and Central Asia. However, a secondary center of diversity has developed in Turkey and unique landraces can even be found in Syria. There is evidence to believe that the carrot was domesticated in Afghanistan from where it spread all over the world, introgressing with the indigenous wild forms to give rise to local cultivars. Two main groups have been recognized: the Asian group with purple and yellow roots, and the Western group with mainly orange roots. In Turkey the two groups seem to converge and give rise to hybrids. In the Near East countries uniform non-hairy hybrid varieties from the West have become popular, but the natives still prefer the hairy purple type with non-uniform roots.

Foeniculum vulgare (Umbelliferae) or fennel is native to the Mediterranean where it has been cultivated since early times. All parts of the plant are aromatic. The essential oil is used as a flavoring agent.

Pastinaca sativa (Umbelliferae) or parsnip is a native biennial of West Asia and has been consumed as a root vegetable since Greek times. Feral forms are known with tough, dry roots. In the tropics, where it has not gained much popularity, parsnip grows well only at higher altitudes.

Pimpinella anisum (Umbelliferae) or anise was domesticated in the Orient and cultivated in the Near East for its aromatic fruits which are used to flavor strong liqueurs, such as Syrian 'arak' and the Greek 'ouzo'. It was known to the ancient Egyptians, Hebrews, Greeks and Romans and was valuable for medicinal purposes during the Middle Ages.

Celtis australis (syn. C. excelsa) (Ulmaceae) or hackberry is native to the eastern Mediterranean and one of the tree species whose fruits were gathered by early farmers from ancient times. The tree is cultivated as an ornamental and for its edible fruits (Mansfield 1959). Charred remains of fruits were found at the PPNB site of Nevali Çori, near Urfa, in southeastern Turkey in 1991 (see Pasternak, this volume).

Vitis vinifera (Vitidaceae) or grapevine is perhaps as old as time itself and was one of the first fruits to be domesticated from the wild in Central Asia, the Near East and the Mediterranean region. The wild progenitor, subsp. sylvestris, is found in areas bordering the Mediterranean Sea except in Libya and Egypt. Its primary center is probably in Armenia and northern Iran. Natural hybrids between the wild and cultivated forms are often found in Tajikstan and as a result newer cultivars come into existence all the time.

Conclusion

There is little doubt that the Near East and the Mediterranean regions have played a major role in the domestication and spread of several major and many more minor crops throughout the world. But why is it that crops originating in the Americas and Africa took a long time to spread to the rest of the world whereas crops domesticated in the region in and around the Fertile Crescent quickly spread far and wide? Most of the progenitors of 'founder crops' were not domesticated again elsewhere after their initial domestication in the Fertile Crescent. One of the explanations could be that the spread of domesticated crop plants was so rapid that people soon stopped gathering the wild species and began to eat only the cultivated crops. Eurasia's west-east axis permitted the Near East crops to promote agriculture over the band of temperate latitudes from Europe to the Indus Valley.

Another question which needs an answer is why the region which gave the world agriculture and several major crops finds itself today in the grip of food deficit? Some say that despite the early lead of the Fertile Crescent in agricultural development the region lost out eventually because of the ecologically suicidal and unsustainable policies adopted by the people of that region, e.g. deforestation, overgrazing, monoculture and the practice of irrigation without crop rotation. These practices destroyed the topsoil and caused widespread erosion as well as salinization followed by a drop in soil fertility and crop yields.

References

Ashri, A. and P.F. Knowles. 1960. Cytogenetics of safflower (Carthamus L.) species and their hybrids. Agron. J. 52:11-17.

Braidwood, R.J., H. Cambel and P.J. Watson. 1969. Prehistoric investigations in southwestern Turkey. Science 164:1275-1276.

Damania, A.B. 1995. Olive, the plant of peace, reigns throughout Mediterranean. Diversity 11(1&2):131-132.

Damania, A.B., M. van Slageren, M. Nikpour, N. Khaghani, A. Soltani and N. Lessani. 1993. Collecting genetic resources of wheat and barley in Iran. Plant Genet. Resour. Newsl. 96:39-44.

Diamond, J. 1997. Guns, Germs and Steel - The Fates of Human Societies. Jonathan Cape, London, UK.

Harlan, J.R. 1951. Anatomy of gene centers. Am. Nat. 85:97-105.

Harlan, J.R., J.M.J. de Wet and A.B.L. Stemler (eds.). 1976a. Origins of African Plant Domestication. Mouton, The Hague, Netherlands.

Harlan, J.R., J.M.J. de Wet and A.B.L. Stemler. 1976b. Plant domestication and indigenous African agriculture. Pages 3-19 in Origins of African Plant Domestication (J.R. Harlan, J.M.J. de Wet and A.B.L. Stemler, eds.). Mouton, The Hague, Netherlands.

Harris, D.R. (ed.). 1996. The Origins and Spread of Agriculture and Pastoralism in Eurasia. UCL Press, London, UK.

Mansfield, R. 1959. Vorläufige Verzeichnis landwirschaftlich oder gärtenrisch kultivierter Pflanzenarten. Kulturpflanz. Suppl. 2, Berlin, Germany [in German].

Purseglove, J.W. 1968. Tropical Crops. Dicotyledons. Longman, Essex, UK.

Smith, B.D. 1995. The Emergence of Agriculture. Freeman, New York, USA.

Ucko, P.J. and G.W. Dimbleby (eds.). 1969. The Domestication and Exploitation of Plants and Animals. Duckworth, London, UK.

van Slageren, M.W. 1994. Wild wheats: a monograph of Aegilops L. and Amblyopyrum (Jaub. & Spach) Eig (Poaceae). Wageningen Agricultural University, Wageningen, Netherlands.

Vavilov, N.I. 1992. (English Translation) Origin and Geography of Cultivated Plants. Cambridge University Press, Cambridge, UK.

Watson, A.M. 1983. Agricultural Innovation in the Early Islamic World - The diffusion of crops and farming techniques, 700-1100. Cambridge University Press, Cambridge, UK.

Yamashita, K. 1980. Origin and dispersion of wheats with special reference to peripheral diversity. Z. Pflanzenzüchtg. 84:122-132.

Zeven, A.C. and J.M.J. de Wet. 1982. Dictionary of Cultivated Plants and Their Regions of Diversity. Pudoc, Wageningen, Netherlands.

Zohary, D. and U. Plitmann. 1979. Chromosome polymorphism, hybrixation and colonization in Vicia sativa group (Fabaceae). Plant Syst. Evol. 131:143-156.

The Spread of Neolithic Agriculture from the Levant to Western Central Asia - D.R. Harris

Introduction

For a symposium which brought together botanists and archaeologists to discuss recent research on the origins of agriculture in the Near East, it is fitting first to recall the profound influence on the subject of two early 20th-century pioneers, one a botanist, N.I. Vavilov, and the other an archaeologist, V. Gordon Childe.

It was in 1926 that Vavilov first portrayed in map form his concept of centers of origin of cultivated plants (Harris 1990). These he argued were to be found in “mountainous districts”, which were not only the homelands of cultivated plants but also of “primeval agriculture” (Vavilov 1926:219). Southwest Asia, which he defined as stretching from the eastern Mediterranean to northern India, was the first of the five centers he proposed at that time, and the one where many of the Old World cereal and pulse crops had been domesticated.

Two years later Childe referred to Vavilov's work when he first adumbrated his hypothesis of the Neolithic or Agricultural Revolution, which proposed that animals were domesticated and agriculture began in the Near East (including the Nile Valley in Egypt) when the climate became more arid following the end of the last glacial period (Childe 1928:42-43). He subsequently elaborated this model, first briefly (Childe 1934:23-30) and then more comprehensively (Childe 1936:66-104), and it subsequently became widely known as Childe's 'desiccation', 'oasis' or 'propinquity' theory. Childe envisaged people and animals being forced by increasing post-glacial aridity (desiccation) into closer proximity (propinquity) around diminishing sources of water (oases) where sheep, goats and cattle were domesticated, a process that, he speculated, would have been facilitated if the people were already cultivating grain and the animals were attracted to the farmers' fields. In fact, he suggested that cereal cultivation (which he thought might have begun in Palestine or Egypt) preceded livestock-raising - a view that finds support today (see below). Vavilov (1926:244) was more concerned to show how a “thorough knowledge of cultivated plants, with their multitude of varieties and differentiation into geographical groups,... makes us refer [their] origin... to the remotest past” than he was to speculate about the particular environmental and cultural contexts of the earliest shifts to cultivation and domestication. Nevertheless, both men regarded the Near East as the region in which agriculture first originated, and their very different but parallel contributions to the subject encouraged later scholars, notably the American archaeologist Robert Braidwood (1960), to undertake archaeological field research on 'agricultural origins' in the region.

Since Vavilov and Childe laid the foundations of the subject in this way, new techniques of bioarchaeological research have been developed and much new evidence has been acquired in the field and laboratory. These investigations have tended to reinforce the view that the transition from hunting and gathering to agriculture took place earlier in Southwest Asia than elsewhere in the world, and that an exceptionally large array of plants and animals was domesticated in the region in the course of the transition. In this contribution to the symposium, attention is focused, however, not on the earliest stages of the process, but on the way in which the system of agropastoral production which originated in western Southwest Asia (the Levant) appears to have spread from that 'core area' during Neolithic times. In particular, the putative spread of agropastoralism northeast into Central Asia is examined in the light of new evidence from recent archaeological investigations in southern Turkmenistan. But before that topic is addressed directly, it is necessary to outline the temporal and methodological context of the study.

Chronology and methodology

The chronological context is the Neolithic which, in the southern and central Levant at least, is conventionally divided into two phases, the second approximately twice as long as the first, i.e. the Pre-Pottery (or Aceramic) Neolithic A (PPNA) from ca. 8500 to ca. 7500 bc, and the Pre-Pottery Neolithic B (PPNB) from ca. 7500 to ca. 5500 bc. (In this paper “bc” refers to uncalibrated radiocarbon dates, “BC” to calibrated dates). The PPNB is commonly subdivided into Early (ca. 7500-7200 bc), Middle (ca. 7200-6500 bc), Late (ca. 6500-6000 bc) and Final (alternatively called PPNC or even Early Late Neolithic, ca. 6000-5500 bc). Following the widespread archaeological appearance of pottery by about 5500 bc, the PPNB is succeeded by the Pottery (or Ceramic) Neolithic (ca. 5500-4200 bc). An alternative chronological scheme (Cauvin 1987), not followed here, defines a sequence of Stages for the northern Levant, in which Stage 2 broadly corresponds to the PPNA, Stage 3 to the Early and Middle PPNB, Stage 4 to the Late PPNB and Stage 5 to the Final PPNB.

Although these chronologies, which are based on radiocarbon (¹⁴C) dates, refer strictly to the southern/central and northern Levant respectively, they correspond sufficiently closely to each other, and refer to a sufficiently large region (the western Fertile Crescent), to provide a sound temporal framework for discussion of the origins and spread of agriculture in Southwest Asia as a whole. However, not all ¹⁴C dates from archaeological sites and sequences are equally reliable because of sampling biases, unclear contextual associations, differences in materials sampled, and interlaboratory inconsistencies in sample-processing (Kuijt and Bar-Yosef 1994). We must therefore be both cautious and critical when comparing ¹⁴C dates from different sites and constructing regional chronologies.

The data and methods of analysis on which the interpretation offered in this paper is based derive from four main fields of study: (1) the biogeography of the wild progenitors of the crops and domestic animals of Neolithic agropastoralism, (2) the bioarchaeology of plant and animal remains recovered from Neolithic sites, (3) analysis of the distributional, architectural and artefactual features of Neolithic settlements, and (4) the ¹⁴C dating of Neolithic sites and sequences, including in particular the use of the accelerator mass spectrometric (AMS) ¹⁴C technique to date individual seeds and bones from the Neolithic crops and domestic animals. These four sources of evidence are invoked in the discussion that follows, which starts by examining how the agropastoral system arose in the Levant and began to spread within the Fertile Crescent during the PPNA and the Early and Middle PPNB.

Origins in the Levant

Because this paper focuses on the Neolithic spread of agriculture and pastoralism, after the crops and livestock of the agropastoral system had been domesticated, we are not directly concerned here with the earlier stages of the transition from hunting and gathering, nor with the problematic question of how to define, and recognize archaeologically, the very beginnings of the process (see Harris 1998 for discussion of that currently controversial question). For our purposes, it is sufficient to start at a point along the continuum of interaction between people, plants and animals where agriculture and pastoralism, based on crop cultivation and the raising of domesticated animals, are being practised, in contrast to techniques of wild plant and animal management (cf. Harris 1996a:444-456).

Leaving aside the question of how plants and animals were exploited by the immediately pre-Neolithic (Epipalaeolithic) people of the Levant, we can start by recognizing, from the bioarchaeological evidence, the presence in the PPNA and Early and Middle PPNB of a suite of domesticated cereals, pulses and one oil/fibre crop - barley, einkorn and emmer wheat, pea, lentil, chickpea, bitter vetch and flax - which constitute the 'founder' crops of Southwest Asian agriculture (Zohary 1989, 1996; Zohary and Hopf 1993). Table 1 gives the full nomenclature of these crops and of their wild progenitors, and Fig. 1 shows the location of the archaeological sites from which their remains have been recovered.

As Fig. 1 shows, only three PPNA sites, all in the southern and central Levant (Jericho, Iraq el-Dubb and Aswad I), have yielded remains of (cereal) crops interpreted as domestic. The crops are two-row barley, emmer and einkorn. Remains of lentils and other herbaceous legumes also have been recovered, but whether they are from wild or domestic forms cannot be determined because of their charred state. The primary evidence for the presence of domesticated cereals and other possible crops at the three sites in the PPNA is presented in Hopf (1983:591-592, 605, 609, 611), Colledge (1994:163-165, Table 5.4) and van Zeist and Bakker-Heeres (1982:171,185-190). In the southern Levant, along the middle Euphrates and in the catchment of the upper Tigris, there are six other PPNA sites where remains of morphologically wild forms of barley, einkorn, rye, lentil, pea and bitter vetch have been found (Fig. 1). As a group, all the PPNA sites are located within areas in which the wild progenitors of barley, einkorn, emmer, pea, lentil, bitter vetch and flax occur today and are believed to have done so in early Neolithic times, although the wild ancestor of the chickpea is restricted today to southeastern Turkey (Zohary 1989). From the presently available archaeobotanical evidence of the plant remains, coupled with the data on the ranges of their wild progenitors, we can infer that during the PPNA the wild cereals and pulses were widely harvested, and perhaps cultivated, by hunter-gatherer groups in the western Fertile Crescent and that the cereals (at least) may possibly have been domesticated first in the central and/or southern Levant.

Table 1. The 'founder' crops of Southwest Asian agriculture.

English name		Traditional binomial† and authority of cultivated crop	Traditional binomial and authority of wild progenitor
Barley		Hordeum vulgare L.	Hordeum spontaneum C. Koch
	Two-row	H. vulgare subsp. disticum
	Six-row	H. vulgare subsp. vulgare
Einkorn wheat		Triticum monococcum L.	Triticum boeoticum Boiss. emend. Schiem.
Emmer wheat		Triticum dicoccum Schübl.	Triticum dicoccoides (Körn.) Aarons.
Lentil		Lens culinaris Medik.	Lens orientalis (Boiss.) Hand-Mazz.
Pea		Pisum sativum L.	Pisum humile Boiss. & Noë
Chickpea		Cicer arietinum L.	Cicer reticulatum Ladiz.
Bitter vetch		Vicia ervilia (L.) Willd.	Never named as an independent species
Flax		Linum usitatissimum L.	Linum bienne Mill.

Sources: Zohary 1989:371; Zohary and Hopf 1993:24, 58, 88, 95-96, 102, 110, 120.
^† The traditional binomials are given here because they are still widely used by archaeobotanists, but, strictly, the rules of nomenclature require that the wild progenitors be named as subspecies (wild races) of the crops, e.g. wild barley should be referred to as Hordeum vulgare subsp. spontaneum.

1: Beidha; 2: Nahal Hemar; 3: Jilat 7; 4: Jericho, Jericho (Tell es-Sultan); 5: Netiv Hagdud; 6: Ain Ghazal; 7: Iraq el-Dubb; 8: Yiftah-el; 9: Aswad I, Aswad II; 10: Ghoraifé I; 11: Mureybit II, III; 12: Jerf el-Ahmar; 13: Halula; 14: Can Hassan III; 15: Haçilar; 16: Asikli Höyük; 17: Cafer Höyük; 18: Çayönü; 19: Hallan Çemi; 20: Quermez Dere; 21: M'lefaat; 22: Ganj Dareh; 23: Abdul Hosein; 24: Ali Kosh.

Although wild, and possibly domesticated, cereals and pulses are likely to have been cultivated in the Levant during the PPNA, it is probable that this was a minor subsistence activity carried out in the vicinity of the typically small but permanent settlements, probably on patches of alluvium with water-retentive soils. There is abundant bioarchaeological evidence that the inhabitants of these sites hunted, trapped, fished and gathered a wide variety of wild animals and plants (Clutton-Brock 1978; van Zeist and Bakker-Heeres 1982; Hillman et al. 1989; Kislev 1989; Bar-Yosef and Belfer-Cohen 1991; Tchernov 1995). Thus an essentially foraging economy, with a minor component of crop cultivation, characterized the PPNA. It was not until after ca. 7500 bc, during the PPNB, that agriculture and pastoralism became the dominant modes of subsistence for a majority of the population of Neolithic Southwest Asia.

The establishment of agropastoralism in Southwest Asia

Evidence for the widespread development of an agropastoral economy during the two millennia of the PPNB comes mainly from changes in the distribution, size and architecture of settlements and from increases in the diversity and relative abundance of the remains of crops and domestic animals found at the sites. Comparison with the PPNA settlement pattern shows a substantial increase in the number of sites and an extension of their overall distribution beyond the Levant and the northern margins of the Fertile Crescent northwest into south-central Anatolia and southeast along the foothills of the Zagros Mountains. The increased number of settlements was accompanied by increases in their size, particularly in the Levant and southeastern Anatolia. These changes increased dependence, at least in the larger 'villages', on local agricultural production and a corresponding decrease over time in dependence for food on wild plants and animals. Architectural changes also occur in the PPNB, notably from the curvilinear structures characteristic of earlier periods to rectilinear layouts of rectangular, mainly mud brick houses with features such as internal fireplaces or ovens, containers for storage, and plaster floors.

When these changes in settlement pattern and architectural style are viewed in conjunction with the more abundant evidence at the PPNB sites for crops and domestic animals, it becomes clear that we are witnessing the establishment and spread within Southwest Asia of an essentially new way of life based increasingly on agriculture and pastoralism, and involving the elaboration of exchange networks which connected the growing settlements and promoted the diffusion of ideas and techniques as well as products.

Comparison between the PPNA and Early and Middle PPNB shows that a major change took place over that time period. Compared with only three PPNA sites, in the southern and central Levant, which have yielded possible evidence of domesticated crops, there are, by the Middle PPNB, 17 sites with definite evidence of such crops, and they are distributed across most of the Fertile Crescent and west as far as central Anatolia (Fig. 1). The representation of the founder crops also increases to include - in addition to two-row barley, einkorn and emmer - naked six-row barley, free-threshing (tetraploid and hexaploid) bread and hard wheat, and rye; and the pulses, especially lentil and pea, are now more widely represented (Garrard 1996), although discrimination between wild and domestic forms remains problematic.

Although this bioarchaeological record of the founder crops is still quite meagre and is hampered by problems of identification and of intersite variations in recovery techniques, the evidence for domestic goats and sheep is even more inadequate (Bar-Yosef and Meadow 1995:82-90; Legge 1996). But, taken as a whole, the evidence for the appearance of domestic caprines in Southwest Asia suggests that they may have been present at a very few Levantine sites (Abu Hureyra, Aswad, Jericho) before the beginning of the Middle PPNB (ca. 7200 bc), and that size reduction (which is an indicator of domestication) began later in sheep than in goats, perhaps not until the end of the Middle PPNB (ca. 6500 bc) in the Levant (Bar-Yosef and Meadow 1995:89). There, up to a millennium apparently separated the beginnings of cereal cultivation and of caprine husbandry. However, by the end of the Final PPNB (ca. 5500 bc) domestic goats and sheep are found at Neolithic sites throughout Southwest Asia, frequently associated with evidence of cereals. It is, therefore, in the later Middle PPNB and especially the Late PPNB, i.e. during the 7th millennium, that we can expect agropastoralism to have become the dominant mode of subsistence across much of the region.

Despite recent advances in archaeological and biological research on the earliest crops and domestic animals, the bioarchaeological evidence is still too uneven and uncertain to permit well-grounded hypotheses about precisely where within Southwest Asia the cereals, pulses and caprines were first domesticated. However, regardless of these uncertainties, we can confidently hypothesize that agropastoralism, as a functionally integrated system of crop and livestock production, came into being during the PPNB, probably no earlier than the Middle PPNB, and that by the end of the PPNB villages sustained mainly by a combination of grain (cereal and pulse) cultivation and (sheep and goats) herding were established throughout the Fertile Crescent from the southern Levant to the central Zagros.

Whether the spread of agropastoralism was due more to adoption of this new way of life by pre-existing hunter-gatherer groups or to colonization by farmers remains uncertain. But it can be argued that the functional integration of grain and livestock production in a system of mixed farming - in which cereals and pulses were grown on the flatter, better-watered, mainly lowland soils and sheep and goats grazed and browsed on rougher upland terrain (whether locally or by means of seasonal transhumance) - proved to be highly effective, ecologically and nutritionally, in sustaining the growing number of sedentary villages. The grain crops alone provided a mix of carbohydrates, vegetable proteins and oils which contained most of the essential nutrients, and this was augmented by animal protein and fat from the caprines. An assured supply of new weaning foods derived from the cultivated cereals - with the addition of milk after goats and sheep began to be routinely milked, perhaps not until after the PPNB - as well as increased storage of grain in the villages and of animal products 'on the hoof, would also have tended to accentuate population growth and stimulate the foundation of new settlements near cultivable land, thus promoting agricultural colonization. Agropastoralism can therefore be seen as a uniquely successful self-sustaining subsistence system which had a 'built-in' tendency to expand spatially and was not restricted territorially by a need for continued access to wild plant and animal foods, as were the hunter-gatherer subsistence systems that it progressively replaced.

By the end of the PPNB agropastoralism had come to support most of the inhabitants of the Fertile Crescent and had spread northwest across Anatolia and southeast to the southern Zagros. Bar-Yosef and Meadow (1995:Fig. 3.6) have constructed the extent of this spread and they also indicate by arrows the directions of further expansion northwest into Europe, southwest to the Nile Valley, southeast toward the Indus Valley and northeast to Central Asia. It is to the last of these putative expansions beyond the Fertile Crescent that we now turn our attention.

The development of agropastoralism in western Central Asia

The next logical step in attempting to trace the postulated spread of agropastoralism toward Central Asia would be to examine the archaeological evidence of Neolithic settlements associated with the remains of crops and domestic livestock in the western half of the Iranian Plateau, east of the crest of the Zagros Mountains. Unfortunately, however, this vast area of mountains, upland valleys, steppe and desert is almost devoid of excavated Neolithic sites, the main exception being Tepe Sialk (Fig. 2), near the modern town of Kashan, where the Pottery Neolithic is represented in the lowest levels (Ghirshman 1938-39:74-77). Remains of domestic two-row hulled barley and of domestic goats were found in these levels (Sialk I), as well as stone hoe and sickle blades, suggesting that agropastoralism was practised at Sialk from the time the site was first occupied, which may have been ca. 5500 bc (Mellaart 1975:187-188). Excavations at another Neolithic tell - Tepe Zaghe - northwest of Sialk and about 65 km south of the town of Qazvin, have revealed early ceramic, and even some aceramic, levels, associated with small rectangular mud brick houses with plastered platforms, dated (by one ¹⁴C determination) to 6269±150 bc (Neghaban 1971; Mellaart 1975:190, 194), which suggests initial occupation there in the Late PPNB.

Farther east, along the northern margins of the Iranian Plateau and the southern shores of the Caspian Sea, there are no reported Neolithic sites until the Elburz Mountains are reached, where there is a tell site, Tepe Hissar, on the southern slopes of the mountains near Damghan, and, still farther east, three sites in the Gorgan Valley (Tureng, Shir-i Shayn and Yarim) and another cluster of small tells (at Sang-i Chakmak) near the town of Bastam (Fig. 2). It is not until one reaches the piedmont zone of southern Turkmenistan, between the northern edge of the Iranian Plateau (formed by the Kopet Dag Range) and the Kara Kum Desert to the north, that known Neolithic sites become more numerous. Several of them have been excavated since the 1950s, and the existence recognized of a local Pottery Neolithic 'Jeitun Culture' (Masson 1957, 1961, 1971; Masson and Sarianidi 1972:33-46; Mellaart 1975:209-216; Kohl 1984:45-55). Recent investigations by an international team of British, Russian and Turkmenian archaeologists have sought to determine how agriculture began in the piedmont zone and have included further excavations at the eponymous site of Jeitun (Masson 1992; Harris et al. 1993, 1996; Harris and Gosden 1996). In view of the dearth of information on Neolithic settlement on the Iranian Plateau, we can best approach the questions of whether, and if so how, when and by what route agropastoralism spread from the Fertile Crescent to western Central Asia, by examining the evidence now available from southern Turkmenistan.

Fig. 2. The northern Iranian Plateau and Kopet Dag Piedmont: Pottery Neolithic sites and present-day zone of dry-summer tropical (mediterranean) climate; site distribution after Kohl 1984, Mellaart 1975 and Sarianidi 1992, climatic zone modified from Trewartha 1960. Sites (Mesolithic (in bold), Pottery Neolithic)

1: Zaghe; 2: Sialk; 3: Hissar; 4: Ghar-i Kamarband; 5: Hotu; 6: Ali Tappeh; 7: Sang- i-Chakmak; 8: Shir-i Shayn; 9: Tureng; 10: Yarim; 11: Jebel; 12: Dam Dam Cheshme; 13: Naiza; 14: Bami; 15: Kelyat; 16: Gievdzhik; 17: Pessedjik; 18: Togolok; 19: Chopan; 20: New Nissa; 21: Jeitun; 22: Yarti Gumbez; 23: Mondjukli; 24: Gadimi; 25: Chagylly.

From 1989 to 1994 excavation at Jeitun was renewed and off-site ecological surveys were undertaken. A principal objective of these new investigations was to obtain from the site evidence of agriculture in the form of cereal or other plant remains and to date them directly by the AMS ¹⁴C method. Systematic sampling and flotation were employed throughout the excavation of parts of the site (Harris et al. 1996:426-429) and our preliminary results confirm that barley and wheat were cultivated locally and sheep and goats raised. A suite of 11 AMS dates from the site, all from charred cereal remains, also confirms that agriculture was being practised at Jeitun by slightly before 5000 bc, or ca. 6000 BC calibrated (Harris et al. 1996:436).

The range of cereals so far identified is, however, less than at most Southwest Asian Neolithic sites. Domestic einkorn is present in all the samples analyzed, but (contrary to our earlier report, in Harris et al. 1993:332) none of the cereal remains can definitely be identified as from domestic emmer; both naked and hulled barley occur at low frequencies, but there is no evidence of any of the pulse crops (Charles and Bogaard 1996). Quite a wide spectrum of wild and weedy herbaceous plants, including grasses, sedges, legumes and crucifers, has also been identified, and, when analyzed by time of flowering/fruiting can be classified into two groups: those that arrived on site as 'probably harvested' crop weeds and those that probably did so in sheep/goat dung. Analysis of the plant remains from Jeitun, coupled with confirmation from the new excavations of the abundance of sheep and goats (as well as gazelle), suggests that the inhabitants practised a form of agropastoralism essentially similar to that in evidence at Middle and Late PPNB sites in the Fertile Crescent. The apparent absence of emmer at Jeitun does, however, appear anomalous, particularly because it is usually more abundant than the other wheats, and sometimes barley, in the Neolithic crop assemblages of the Fertile Crescent. If its absence, and the dominance of einkorn, is confirmed by analysis of further samples from Jeitun, an explanation of this apparently anomalous situation may lie in selective adaptation to local conditions, which (despite average annual precipitation - today - being less than 250 mm) might have favored the cultivation of drought-tolerant forms of einkorn as a rain-fed winter cereal rather than the cultivation of typical forms of emmer, which in this environment would require irrigation. Another explanation of the anomaly could be that emmer was not introduced with einkorn as part of a Neolithic crop 'package' coming from farther west, a possibility that could best be tested by archaeobotanical work at other early Jeitun Culture sites on the piedmont.

Mention of the introduction of a Neolithic 'package' from the west brings us back to the main theme of this paper and raises the fundamental question of whether any of the founder species of Neolithic agropastoralism in Southwest Asia which have been identified at Jeitun, i.e. einkorn, ?emmer, barley, goats and sheep, could have been domesticated locally rather than introduced as domesticates. Reference has already been made to Zohary's (1989) account of the identification and distribution of the wild progenitors of the founder crops, which shows them to be largely concentrated in the Fertile Crescent, with some species extending westward and/or eastward, mainly as weedy forms. In the updated versions of his 1989 maps published with Hopf in 1993 there are only very minor changes in the distribution patterns, which still show that neither wild einkorn nor wild emmer occurs today east of the Zagros Mountains. Wild barley too occurs mainly in the Fertile Crescent, but “more isolated populations, usually of weedy forms” do extend east across Central Asia as far as Tibet (Zohary 1989:360). Valuable additional information on the present-day distribution of wild wheats and barley, incorporating new field data, is now available from the work of Jan Valkoun et al. (this volume). Although their map of the distribution of Triticum boeoticum (wild einkorn) marks a few collection sites farther east in Iran than any shown by Zohary and Hopf (1993:34), the sites are still confined to the mountains of western Iran (southern Zagros and western Elburz), and their map of the distribution of T. dicoccoides (wild emmer) does not extend its distribution eastward at all in comparison with Zohary's and Hopf's map (1993:41).

Assuming that the distributions of the wild progenitors did not differ grossly in Neolithic times from the present patterns, we can exclude the possibility of local domestication, east of the Caspian Sea in southern Turkmenistan, of both einkorn and emmer. Barley is more problematic, in that apparently wild stands of the wild progenitor occur in southeastern Turkmenistan and northeastern Iran today (Harris and Gosden 1996:386), but these populations have characteristics, such as an upright habit and synchronised tillering and seed maturation, which suggest that they are descended from weedy forms introduced with domestic barley during the spread of Neolithic agriculture from farther west (Jan Valkoun, pers. comm.). The probability is therefore that barley was not independently domesticated in western Central Asia and that the remains recovered at Jeitun and Chagylly derived from domestic barley originally introduced from Southwest Asia.

The natural ranges of the wild progenitors of domestic sheep and goats -respectively the western Asiatic mouflon (Ovis orientalis Gmelin) and the bezoar (Capra aegagrus Erxleben) - extend across Southwest Asia and into Central Asia (Harris 1962; Uerpmann 1987:113-118, 124-132, 1989), so their local domestication in southern Turkmenistan appears to be a possibility. But, although the bezoar is native to the area, the range of the Asiatic mouflon extends eastward only as far as the Elburz Mountains, where it overlaps the western limit of the urial (Ovis vignei Blyth) (Clutton-Brock 1981:53-54; Uerpmann 1987:126-130). The mouflon does not extend into Turkmenistan, and although the urial can interbreed with the mouflon - viable hybrids have been reported in north-central and southeastern Iran (Valdez et al. 1978; Meadow 1989:29) - it has a higher chromosome number than the mouflon and is unlikely to be a direct ancestor of domestic sheep (Clutton-Brock 1981:54).

Consideration of the distributions of the progenitors of the early Neolithic crops and domestic animals represented at Jeitun thus tends to support the view that they were introduced to the Kopet Dag piedmont rather than domesticated locally, although local domestication of the goat cannot be wholly excluded on these grounds. There is no evidence reported of domestic (or wild) cattle at Jeitun, but according to Masson and Sarianidi (1972:44) the “abundant osteological evidence from the upper [Late Jeitun] layers” at Chagylly “points to cattle... having been domesticated.” Small numbers of wild boar bones have been recovered at Jeitun and they have been reported from other Jeitun Culture sites, but there is no indication at present that domestic pigs were part of the Neolithic economy.

Viewed overall, the bioarchaeological evidence from Jeitun and the other Jeitun Culture sites on the piedmont suggests the presence from ca. 5000 bc of an already 'developed' Neolithic agropastoral system which strongly resembles the village-based grain-caprine economy of the PPNB in the Fertile Crescent. There are similarities too in settlement pattern, architecture and artefacts. The Jeitun Culture settlements were small, compact villages consisting of rectilinear mud brick structures, most of which were single-roomed houses with lime-plaster floors and interior ovens with adjoining low platforms (Masson and Sarianidi 1972: Fig. 9; Harris et al. 1996: Figs. 3-6). Clay figurines and 'gaming counters', pottery, small stone axes and an array of other tools are very similar to artefacts found at Neolithic settlements in the Fertile Crescent, particularly at such Zagros sites as Jarmo, Tepe Guran and Tepe Sarab (Masson and Sarianidi 1972:45).

The implication of these similarities is that the settlement-subsistence system of the Jeitun Culture is more likely to have been 'implanted' in the piedmont zone from outside than to have developed in situ from local pre-Neolithic precursors. This conclusion is reinforced by the fact that all the known Early Jeitun sites are located on the western piedmont whereas the eastern sites (insofar as they have been excavated) lack evidence of early occupation - implying that settlement spread eastward across the piedmont during the 6th millennium and that the antecedents of the Jeitun Culture are most likely to be found farther west in Turkmenistan or northeastern Iran.

This line of reasoning directs attention to the valleys of the Sumbar and Chandyr Rivers which dissect the western end of the Kopet Dag Range and give easy access to the piedmont zone from the west (Fig. 2). These long east-west trending valleys are well suited to rain-fed agriculture, having higher rainfall than the piedmont and deeper, more fertile soils, and they appear to offer likely routes for the spread of Neolithic agropastoralism into southern Turkmenistan. In 1995 this possibility was investigated by members of the international team who had worked at Jeitun. They found, and trial-excavated, eight rock shelters and one open site in the Sumbar and Chandyr Valleys but detected no signs of Neolithic (or earlier) occupation other than three pottery sherds of unknown type which may be Neolithic (Gosden et al. 1996). This surprisingly negative result does not of course mean that the valleys were not settled by Neolithic farmers - early sites may have been buried under more recent alluvium or destroyed by later agricultural activities - but it does show that it will not be easy to test the hypothesis that these upland valleys functioned as conduits for the spread of agropastoralism to the piedmont zone.

In searching for the antecedents of the Jeitun Culture, attention turns next to the few known Neolithic sites, already referred to, in the Gorgan Valley and on the southeastern side of the Elburz Mountains (Fig. 2). At Yarim and Tureng, pottery of Jeitun Culture type was recovered from the lowest occupation levels although at Tureng only Late Jeitun ceramics were found (Crawford 1963; Deshayes 1967; Sarianidi 1992:113). At Sang-i Chakmak (Sang-e Caxmaq) there are several small tells, two of which - the eastern and western tells - have been partially excavated (Masuda 1974a, 1974b, 1976). The architecture, pottery and other artefacts of the eastern tell closely resemble the material culture of the Jeitun sites as well as that found at Yarim, and although there are no ¹⁴C dates from it, it is thought to date to the Late Jeitun phase. The western tell proved to be almost devoid of pottery, but architecturally it is very similar to the Jeitun sites (Aurenche 1985:236, Fig. 3). Two ¹⁴C dates, of 5505±155 bc and 5540±130 bc, were obtained from the middle occupation levels, which, when calibrated and compared with the AMS dates from Jeitun, indicate that the western tell at Sang-i Chakmak was occupied some 200-300 years before Jeitun (Harris and Gosden 1996:382). This comparison suggests that the Sang-i Chakmak tells, and possibly those in the Gorgan Valley, may represent antecedents of the Jeitun Culture, but without more complete excavation and dating of them this possibility cannot be confirmed.

In seeking evidence for possible local antecedents of the Jeitun Culture we should consider, in addition to the Neolithic tells in northeastern Iran, the rock shelter sites with evidence of Mesolithic occupation that exist close to the southern and eastern coasts of the Caspian Sea (Fig. 2). They form two clusters: the south-Caspian sites of Ali Tappeh, Hotu and Ghar-i Kamarband (Belt Cave), which were investigated by Coon (1957) and McBurney (1968), and the east-Caspian sites of Jebel and Dam Dam Cheshme at the foot of the Bolshoi Balkhan Mountains, which were excavated by Okladnikov (1956) and Markov (1966). Gupta (1979:II:49-52) suggested, mainly on the grounds of the almost complete absence of pottery at the western Sang-i Chakmak tell and some similarities in the stone tools, that the site might represent a transition from the south-Caspian Mesolithic to the Jeitun Culture. It is quite possible that there were connections across the Elburz Mountains in the late Mesolithic/early Neolithic, but occupation had ceased at Ali Tappeh by ca. 9000 bc, and, although (?domestic) sheep and goats increase at the two other rock shelters after ca. 7500 bc and pottery appears by ca. 6100 bc shortly before these sites too are deserted (Mellaart 1975:209), the postulated cultural similarities with Sang-i Chakmak appear tenuous. It is more likely that such 'Neolithic' elements as are present in the south-Caspian sites, and the early Neolithic culture of Sang-i Chakmak, both derive from farther west (see below).

A somewhat different problem arises when we consider the east-Caspian sites. The rock shelters excavated here are thought to have been occupied from the 10th to the 5th millennia, and they yielded bones of domestic goats and sheep (as well as abundant gazelle, onager and fish bones) in levels dated to the 6th and early 5th millennia. No evidence of crops was reported, but the sites were excavated before modern techniques of recovery of very small plant (and animal) remains, e.g. by flotation, were generally applied. The potential significance of these long-occupied sites has led members of the international team that worked recently at Jeitun to initiate a program of site survey and excavation in the hope of determining more precisely the nature of the local Mesolithic-Neolithic occupation and its possible relationship with the Jeitun Culture, but results are not yet available from the first field season in 1997, when limited excavation was carried out in several rock shelters, flotation was systematically applied and samples taken for AMS dating.

The presence of the remains of domestic caprines in the rock shelter deposits could be explained in several ways: that resident hunter-gatherers were herding caprines adopted from Jeitun Culture farmers, that they were hunting feral goats and sheep, or that transhumant pastoralists from the Jeitun Culture villages used the rock shelters seasonally. We hope to obtain new evidence that will help us choose between these hypotheses, and also demonstrate whether the occupants of the rock shelters cultivated any crops (or perhaps obtained grain from elsewhere). However, even in the absence of such new evidence it is highly improbable that the Jeitun Culture developed out of the east-Caspian Mesolithic, although the interesting question of how the inhabitants of the sites in the Bolshoi Balkhan region interacted with those living on the piedmont in the 6th millennium itself deserves investigation.

When we look still farther west, beyond the southeastern Caspian and the eastern Elburz Mountains, for similarities to and antecedents of the Jeitun Culture, we encounter the problem, already mentioned, of the lack of excavated Neolithic settlements on the Iranian Plateau. Over 600 km (in a direct line) of mountain and desert terrain separate the eastern-Elburz and Gorgan Valley sites from the easternmost Neolithic sites of the Zagros; and Sialk I, which Mellaart (1975:194) interprets as culturally equivalent to the Jeitun Culture and Masson and Sarianidi (1972:45) claim has 'as its basis, local Neolithic traditions of the Djeitun type', lies over 400 km southwest of Sang-i Chakmak across the forbidding desert of Dasht-e-Kavir. Even the closest sites with excavated Neolithic levels in northern Iran - Hissar and Zaghe - are over 300 km apart (Fig. 2), and it is difficult to infer anything about possible early Neolithic connections between them from the published reports (Schmidt 1937; Neghaban 1971).

In the absence of new archaeological investigations in northern Iran, we cannot at present follow a trail of Neolithic settlement westward, in the hope of tracing more precise connections between the PPNB and/or the early Pottery Neolithic of the eastern Fertile Crescent and the Pottery Neolithic Jeitun Culture sites. But we can, in conclusion, speculate about when and by what route agropastoralism spread from west to east, and ask whether colonization is likely to have been the dominant process or whether the spread was due more to pre-exisiting Mesolithic hunter-gatherers adopting grain cultivation and caprine herding.

Conclusion: the spread of agropastoralism from the Fertile Crescent to western Central Asia

We have already established that by the end of the PPNB, ca. 5500 bc, most of the inhabitants of the Fertile Crescent, from the southern Levant to the central Zagros, were supported by an agropastoral village-based economy, which, as an integrated system of grain cultivation and caprine herding, is unlikely to predate the beginning of the Middle PPNB, ca. 7200 bc. We also know that by 5000 bc a 'developed' Neolithic grain-caprine economy was in place, as the Jeitun Culture, on the Kopet Dag piedmont in southern Turkmenistan and probably a few centuries earlier at Sang-i Chakmak and other early Neolithic sites in northeastern Iran. The maximum time available for the spread of agropastoralism from the Fertile Crescent to the region of the Jeitun Culture is therefore about 2000 years, but a more realistic estimate, which spans the period between the Late PPNB in the Zagros (including the partly aceramic site of Tepe Zaghe already referred to) and the beginning of Jeitun settlement, is about 1000 years, i.e. from ca. 6300 to ca. 5300 bc.

Any attempt to infer the route by which agropastoralism spread must also be speculative, but here the assumption made is that it would have done so in the environment most favorable to rain-fed agriculture in the area between the Zagros and the Kopet Dag piedmont. This assumption is likely to be valid even if the Neolithic farmers were familiar with techniques of irrigation and added water to their fields to ensure or enhance grain production, because both rain-fed and irrigation agriculture would have been more dependable and productive in this zone than in any of the other major biomes of northern Iran: the high mountains, the deserts to the south, and the heavily forested coastal lowlands to the north. Allowance should be made for climatic and other environmental changes that may have occurred in the area during and since the 6th millennium (about which we are presently ignorant), but the contrasts in terrain between upland and lowland and between forest, steppe and desert are so extreme in northern Iran that any such changes are unlikely greatly to have altered the west-east trending pattern of climatic and vegetational zones.

Today the zone most propitious for agriculture is that designated by the so-called mediterranean or dry-summer subtropical climate with its associated woodland/grassland vegetation containing a great diversity of woody and herbaceous species growing on rocky slopes (well suited to caprine pastoralism) and on fertile valley soils (well suited to grain cultivation). This zone extends from the northeastern Fertile Crescent in a long narrow lobe that projects west-east south of the Caspian Sea to approximately 60 E longitude (Fig. 2). It encompasses the middle and lower elevations of the Elburz Mountains and the Gorgan Valley, and follows the course of the Atrek River (including its north-bank tributaries the Sumbar and Chandyr) as far east as the modern city of Mashhad. It offers an inviting 'corridor' for the spread of agropastoralism from the northern Zagros to the Kopet Dag piedmont, and one where systematic prospecting for early Neolithic sites is likely to be rewarded.

Even if these speculations about when and by what route agropastoralism spread to western Central Asia are broadly correct, they leave open the question of whether it reached the Kopet Dag piedmont as a result mainly of colonization by farmers (primary or demic diffusion) or the adoption of domesticates and agropastoral techniques by resident foragers (secondary diffusion). These need not, of course, be mutually exclusive processes, and it is likely that both colonization and adoption occurred, but the conclusion to be drawn from the evidence and reasoning presented in this paper is that the spread of agropastoralism across northern Iran was due primarily to colonization. This conclusion is based partly on the (already discussed) assumption that the grain-livestock mixed-farming system which originated in the Fertile Crescent during the PPNB was inherently expansive and accentuated population growth; partly on the - admittedly still rather tenuous - evidence for similarities in the material culture of Neolithic sites in the eastern Fertile Crescent and the Jeitun Culture region; and partly on the fact that when the Jeitun Culture appears archaeologically at ca. 5000 bc it does so relatively suddenly and exhibits the main features of a 'developed' Neolithic economy in the Fertile Crescent: small village settlements with rectilinear mud brick architecture, pottery and domesticated barley, wheat, goats and sheep. However, this conclusion does not preclude the possibility that some of the elements of the agropastoral economy (e.g. goats, sheep, pottery) were selectively adopted through trade or by other means, such as intermarriage, by hunter-gatherer groups living in the region, e.g. at the rock shelter sites in the Bolshoi Balkhan Mountains.

If colonization from the west was responsible for the foundation of the first Jeitun Culture settlements it is likely to have taken place quite rapidly, spreading from the northern Zagros to the western Kopet Dag piedmont within a millennium. If, as is suggested, the spread took place along the zone of mediterranean climate across northern Iran the distance was of the order of 1000 km, thus suggesting an average rate of spread of approximately 1 km per year. Interestingly, this approximation is of the same order of magnitude as a comparable calculation of the rate of spread of Neolithic agriculture from the Levant across Anatolia to southeastern Europe (Harris 1996b).

In recent decades there has been much controversy among archaeologists about the relative significance in explaining culture change of processes of diffusion and of independent (autochthonous) development. The debate has tended to become polarized between proponents of these two modes of explanation, and it has been particularly lively in discussions of 'agricultural origins'. In this paper an attempt has been made to assess critically whether agropastoralism originated independently in western Central Asia or spread there from the Fertile Crescent. The evidence presently available decisively supports the hypothesis that it spread mainly by colonization during the 6th millennium bc across northern Iran within the belt of mediterranean climate that extends eastward past the Caspian Sea.

Acknowledgments

I wish to thank all members of the international team who have, since 1989, participated in the archaeological and environmental investigations at Jeitun and other sites in southern Turkmenistan, particularly Professor V.M. Masson (St Petersburg), Dr K. Kurbansakhatov (Ashgabat), Dr Chris Gosden (Oxford) and Dr Mike Charles (Sheffield). I also wish to thank Catherine Pyke of the UCL Geography Department for preparing Figures 1 and 2, and my colleagues Andrew Garrard and Gordon Hillman for valuable comments on an earlier version of this paper. The recent fieldwork in Turkmenistan was supported by the British Academy, the British Institute for Persian Studies, University College London, the University of Oxford and the Institute for the History of Material Culture in St Petersburg.

References

Aurenche, O. 1985. La tradition architecturale à l'est de la Mésopotamie (Iran-Turménistan) du 10e au 4e millénaires. Pages 235-240 in L'Archéologie de la Bactriane Ancienne. Actes du Colloque Franco-Soviétique, 27 Octobre a 3 Novembre 1982, Dushanbe. Editions du CRNS, Paris, France [in French].

Bar-Yosef, O. and A. Belfer-Cohen. 1991. From sedentary hunter-gatherers to territorial farmers in the Levant. Pages 181-202 in Between Bands and States (S.A. Gregg, ed.). Occasional Paper No. 9, Center for Archaeological Investigations, Southern Illinois University, Carbondale, USA.

Bar-Yosef, O. and R.H. Meadow. 1995. The origins of agriculture in the Near East. Pages 39-94 in Last Hunters-First Farmers: New Perspectives on the Prehistoric Transition to Agriculture (T.D. Price and A.B. Gebauer, eds.). School of American Research Press, Santa Fe, New Mexico, USA.

Braidwood, R.J. 1960. The agricultural revolution. Sci. Am. 203:130-141.

Cauvin, J. 1987. Chronologie relatif et absolue dans le Néolithique du Levant nord et d'Anatolie entre 10,000 et 8,000 BP. Pages 325-342 in Chronologies in the Near East: Relative Chronologies and Absolute Chronology 16,000-4,000 BP (O. Aurenche, F. Evin and F. Hours, eds.). British Archaeological Reports International Series, 379 (i), Oxford, UK.

Charles, M. and A. Bogaard. 1996. Jeitun charred plant remains. Report on Project 9601. Sheffield Environmental Facility, Department of Archaeology and Prehistory, University of Sheffield, UK.

Childe, V.G. 1928. The Most Ancient East: the Oriental Prelude to European Prehistory. Kegan Paul, Trench, Trubner, London, UK.

Childe, V.G. 1934. New Light on the Most Ancient East. Routledge & Kegan Paul, London, UK.

Childe, V.G. 1936. Man Makes Himself. Watts, London, UK.

Clutton-Brock, J. 1978. Early domestication and the ungulate fauna of the Levant during the Pre-Pottery Neolithic Period. Pages 29-40 in The Environmental History of the Near and Middle East since the Ice Age (W.C. Brice, ed.). Academic Press, New York, USA.

Clutton-Brock, J. 1981. Domesticated Animals from Early Times. Heinemann and British Museum (Natural History), London, UK.

Colledge, S. 1994. Plant Exploitation on Epipalaeolithic and Early Neolithic Sites in the Levant. PhD Thesis. University of Sheffield, UK.

Coon, C.S. 1957. Seven Caves: Archaeological Explorations in the Middle East. Jonathan Cape, London, UK.

Crawford, V.E. 1963. Beside the Kara Su. Bull. Metropolitan Museum of Art 22:263-273.

Deshayes, J. 1967. Céramiques peintes de Tureng Tépé. Iran 5:123-131.

Garrard, A. 1996. Charting the emergence of cereal and pulse domestication in South-West Asia. Paper delivered at the International Symposium on The Origins of Agriculture held at the International Research Center for Japanese Studies, 13-16 December 1996, Nara, Japan.

Ghirshman, R. 1938-39. Fouilles de Sialk près de Kashan 1933, 1934, 1937, Vol. 1. Série Archéologique Tome IV. Musée de Louvre, Département des Antiquités Orientales, Librairie Orientaliste Paul Geuthner, Paris, France [in French].

Gosden, C., K. Kurbansakhatov, M. Charles, S. O'Hara, N. Schlanger, C. Merrony and V. Zav'yalov. 1996. Report [to the British Academy] on archaeological fieldwork in southern Turkmenistan, March 1996.

Gupta, S.P. 1979. Archaeology of Soviet Central Asia and the Indian borderlands, 2 Vols. B.R. Publishing Corporation, Delhi, India.

Harris, D.R. 1962. The distribution and ancestry of the domestic goat. Proc. Linn. Soc. Lond. 173:79-91.

Harris, D.R. 1990. Vavilov's concept of centres of origin of cultivated plants: Its genesis and its influence on the study of agricultural origins. Biol. J. Linn. Soc. 39:7-16.

Harris, D.R. 1996a. Domesticatory relationships of people, plants and animals. Pages 437-463 in Redefining Nature: Ecology, Culture and Domestication (R. Ellen and K. Fukui, eds.). Berg, Oxford, UK.

Harris, D.R. 1996b. The origins and spread of agriculture and pastoralism in Southwest Asia. Paper given at the International Symposium on The Origins of Agriculture held at the International Research Center for Japanese Studies, 13-16 December 1996, Nara, Japan.

Harris, D.R. 1998. The origins of agriculture in Southwest Asia. The Review of Archaeology 19 (in press).

Harris, D.R. and C. Gosden. 1996. The beginnings of agriculture in western Central Asia. Pages 370-389 in The Origins and Spread of Agriculture and Pastoralism in Eurasia (D.R. Harris, ed.). UCL Press, London, UK.

Harris, D.R., C. Gosden and M.P. Charles. 1996. Jeitun: recent excavations at an early Neolithic site in southern Turkmenistan. Proc. Prehistoric Soc. 62:423-442.

Harris, D.R., V.M. Masson, Y.E. Berezkin, M.P. Charles, C. Gosden, G.C. Hillman, A.K. Kasparov, G.F. Korobkova, K. Kurbansakhatov, A.J. Legge and S. Limbrey. 1993. Investigating early agriculture in Central Asia: new research at Jeitun, Turkmenistan. Antiquity 67:324-338.

Hillman, G.C., S.M. Colledge and D.R. Harris. 1989. Plant-food economy during the epipalaeolithic period at Tell Abu Hureyra, Syria: dietary diversity, seasonality and modes of exploitation. Pages 240-268 in Foraging and Farming - The Evolution of Plant Exploitation (D.R. Harris and G.C. Hillman, eds.). Unwin Hyman, London, UK.

Hopf, M. 1983. The plants found at Jericho. Pages 580-621 in Excavations at Jericho V (K.M. Kenyon and T.A. Holland, eds.). British School of Archaeology in Jerusalem, London, UK.

Kislev, M.E. 1989. Pre-domesticated cereals in the Pre-Pottery Neolithic A period. Pages 147- 151 in People and Culture in Change (I. Hershkovitz, ed.). British Archaeological Reports International Series 508 (i), Oxford, UK.

Kohl, PL. 1984. Central Asia: palaeolithic beginnings to the Iron Age. Synthèse 14, Editions Recherche sur les Civilisations, Paris, France.

Kuijt, I. and O. Bar-Yosef. 1994. Radiocarbon chronology for the Levantine Neolithic: observations and data. Pages 227-245 in Late Quaternary Chronology and Paleoclimates of the Eastern Mediterranean (O. Bar-Yosef and R.S. Kra, eds.). Radiocarbon, Department of Geosciences, University of Arizona, Tucson, USA.

Legge, T. 1996. The beginning of caprine domestication in Southwest Asia. Pages 238-262 in The Origins and Spread of Agriculture and Pastoralism in Eurasia (D.R. Harris, ed.). UCL Press, London, UK.

Markov, G.E. 1966. Grot Dam Dam Cheshme 2 v vostochnom prikaspii. Sovetskaya Arkheologiya 2:104-125 [in Russian].

Masson, V.M. (ed.). 1992. New Research on the Jeitun Settlement: Preliminary Reports of the Work of the Soviet-British Expedition. Academy of Sciences of Turkmenistan, Ashgabat, Turkmenistan [in Russian].

Masson, V.M. 1957. Jeitun and Kara-tepe. Sovetskaya Arkheologiya 1:143-160 [in Russian].

Masson, V.M. 1961. The first farmers in Turkmenia. Antiquity 35:203-213.

Masson, V.M. 1971. The Jeitun Settlement: the Emergence of a Productive Economy. Materials and Research on the Archaeology of the USSR 180, Nauka, Moscow, USSR [in Russian].

Masson, V.M. and V.I. Sarianidi. 1972. Central Asia: Turkmenia before the Achaemenids. Thames & Hudson, London, UK.

Masuda, S. 1974a. Tepe Sang-e Caxmaq. Iran 12:222-223.

Masuda, S. 1974b. Excavations at Tappeh Sang-e Caxmaq. Pages 23-33 in Proc. IInd Annual Symposium on Archaeological Research in Iran (F. Bagherzadeh, ed.). Iranian Centre for Archaeological Research, Teheran, Iran.

Masuda, S. 1976. Report of the archaeological investigations at Sahrud, 1975. Pages 63-70 in Proc. of the IVth Annual Symposium on Archaeological Research in Iran (F. Bagherzadeh, ed.). International Centre for Archaeological Research, Teheran, Iran.

McBurney, C.B.M. 1968. The cave of Ali Tappeh and the Epi-Palaeolithic of N. E. Iran. Proc. Prehistoric Soc. 34:385-413.

Meadow, R.H. 1989. Prehistoric wild sheep and sheep domestication on the eastern margin of the Middle East. Pages 25-36 in Early Animal Domestication and its Cultural Context (P.J. Crabtree, D. Campana and K. Ryan, eds.). MASCA Research Papers in Science and Archaeology, University Museum of Archaeology and Anthropology, University of Pennsylvania, Philadelphia, USA.

Mellaart, J. 1975. The Neolithic of the Near East. Thames & Hudson, London, UK.

Neghaban, E. 1971. Excavations at Tepe Sagzabad and Zaghe. Symposium on Archaeological Research in Iran. Tehran University, Iran, June 1971.

Okladnikov, A.P. 1956. Peschera Djebel - pamyatnik drevnei kul'turi prikaspiskikh Turkmenii. Trudy IuTAKE 7:11-219 [in Russian].

Sarianidi, V.I. 1992. Food-producing and other Neolithic communities in Khorasan and Transoxania: eastern Iran, Soviet Central Asia and Afghanistan. Pages 109-126 in History of Civilizations of Central Asia, Vol. 1 (A.H. Dani and V.M. Masson, eds.). UNESCO, Paris, France.

Schmidt, E.F. 1937. Excavations at Tepe Hissar, Damghan. Publications of the Iranian Section of the University Museum, University of Pennsylvania Press, Philadelphia, USA.

Tchernov, E. 1995. Environmental and socioeconomic background to domestication in the southern Levant. MASCA Res. Papers in Science and Archaeol. 12:39-77 (Supplement).

Trewartha, G.T. 1960. Climates of the Earth (map). Pages 8-9 in Goode's World Atlas, 11th Edition (E.B. Espenhade Jr., ed.). Rand McNally, Chicago, USA.

Uerpmann, H-P. 1987. The Ancient Distribution of Ungulate Mammals in the Middle East: Fauna and Archaeological Sites in Southwest Asia and Northeast Africa Beihefte zum Tübinger Atlas des vorderen Orients, Reihe A: Naturwissenschaften, Nr. 27, Ludwig Reichert, Wiesbaden, Germany.

Uerpmann, H-P. 1989. Middle East: Ancestral Forms of Domestic Animals and Early Domestication. Tübinger Atlas of the Middle East, Map A VI 16, Ludwig Reichert, Wiesbaden, Germany.

Valdez, R., C.F. Nadler and T.D. Bunch. 1978. Evolution of wild sheep in Iran. Evolution 32:56- 72.

van Zeist, W. and J.A.H. Bakker-Heeres. 1982. Archaeobotanical studies in the Levant, 1: Neolithic sites in the Damascus Basin: Aswad, Ghoraifé and Ramad. Paleohistoria 24:165-256.

Vavilov, N.I. 1926. Studies on the Origin of Cultivated Plants. Institut Botanique Appliqué et d'Amelioration des Plantes, Leningrad, USSR.

Zohary, D. 1989. Domestication of the Southwest Asian Neolithic crop assemblage of cereals, pulses and flax: the evidence from the living plants. Pages 358-373 in Foraging and Farming - The Evolution of Plant Exploitation (D.R. Harris and G.C. Hillman, eds.). Unwin Hyman, London, UK.

Zohary, D. 1996. The mode of domestication of the founder crops of Southwest Asian agriculture. Pages 142-158 in The Origins and Spread of Agriculture and Pastoralism in Eurasia (D.R. Harris, ed.). UCL Press, London, UK.

Zohary, D. and M. Hopf. 1993. Domestication of Plants in the Old World: the Origin and Spread of Cultivated Plants in West Asia, Europe, and the Nile Valley, 2nd edition. Clarendon Press, Oxford, UK.

The Spread of Agriculture to the Eastern Arc of the Fertile Crescent: Food for the Herders - F. Hole

Introduction

The link between the Böling-Alleröd amelioration, the Younger Dryas cold period, and the advent of agriculture and animal domestication seems well established (Hole 1989, 1996; McCorriston and Hole 1991; Moore 1992; Moore and Hillman 1992; Bar-Yosef and Meadow 1995). In this paper I discuss the spread of agriculture from its apparent heartland, the southern Levant, across the Fertile Crescent, into the Zagros. The timing of this expansion was determined by the climatic/environmental changes during the Climatic Optimum which immediately succeeded the Younger Dryas. At least for the eastern side of the Near East, the spread of agriculture was closely related to the expansion of animal husbandry and the development of patterns of seasonal transhumance.

We should bear in mind that “Late Quaternary paleoclimatic data available for Southwest Asia and the Near East are outstanding neither in quality nor in quantity” (Roberts and Wright 1993). Although much remains to be discovered about the precise nature of climate at any time or any place in the Near East during the crucial millennia, the broad outline and implications can now be sketched. Paleoclimatology has advanced more rapidly than archaeological discoveries which now lag seriously behind, both because of the inherently slow pace of excavation and analysis and also because many parts of our topical region have been inaccessible to modern research for prolonged periods. Further, the field of archaeology is notoriously underfunded so that skilled research teams can seldom sustain their efforts. An examination of the pattern of discovery reveals significant gaps in geographic coverage and an embarrassingly thin network of sites that span the crucial periods. However, notwithstanding such caveats, one can perceive an outline of the way agricultural economies spread across the Near East.

More than 30 years ago, I was involved in the excavations that first recovered plant remains through the use of flotation, from an early village in Iran (Hole et al. 1969). Hans Helbaek, a pioneer in the technique of identifying plant remains and one of the first to apply his knowledge to the study of agricultural origins, examined this material from Ali Kosh (Helbaek 1969). He concluded that the site recorded some of the early attempts to bring cereals under human control. This unprecedentedly complete record, coupled with early radiocarbon dates, led us to believe that we had recovered some of the oldest evidence of domestication. We now know this to be untrue, although the local adoption of agriculture may have belatedly recapitulated similar events that took place earlier and repeatedly across the Fertile Crescent. Ali Kosh, on the Deh Luran Plain of southwestern Iran, lies near the southern edge of the Fertile Crescent, nearly as far from the southern Levant as one can get in the Near East. Despite the long geographic arc involved in travel, the Fertile Crescent probably provided the natural route for the spread of cereals and agriculture, although not by simple diffusion. Rather, the mountain zone also played an important role via transhumant pastoralism.

The spread of agriculture eastward

It is agreed that the first evidence of morphologically altered domesticated cereals is found at PPNA sites in the southern Levant: Jericho, Netiv Hagdud, Gilgal and Gesher, all clustered within a 15-km radius on alluvial fans in the Jordan Valley. The contemporary village of Aswad in the Damascus basin lies at the edge of what then was a marsh/lake. All these sites are dated to approximately 9000 BC calibrated (11,000 BP), just following the Younger Dryas cold interval which terminated ca. 9000 BC or a little later (Becker et al. 1991; Kromer and Becker 1993). Presumably there are other such sites in similar locations, most of them probably buried. Harlan notes that the Jordan Valley sites lie well below the oak woodland zone in which wild cereals occur today, as does Aswad (Harlan 1995), but he does not speculate about the distribution of trees 11,000 years ago when these sites were occupied. At Abu Hureyra on the Euphrates, Hillman reports finding both domestic and wild rye in late Epipaleolithic layers, as much as 500 years earlier than the occurrence of domestic grains in the southern Levant. Until the dating is confirmed, the next oldest site with plant remains in the Euphrates region is Jerf el-Ahmar, dated to 9000 BC. Here, wild barley, rye and einkorn were harvested but there is no evidence of the cultivated varieties (Wilcox 1995). The first actual domesticates on the Euphrates appear in Halula no earlier than 8000 BC, a full 1000 years after their occurrence in the Jordan Valley.

Much farther to the east, in the Tigris drainage of eastern Anatolia, the earliest settlement at Hallan Çemi dates to about 10,500 BC and lay in what may have been a refuge for pistachio and oak trees during the Younger Dryas (van Zeist and Bottema 1982; Rosenberg and Davis 1992; Rosenberg 1994). Despite this, no trace of cereals has turned up in the voluminous flotation samples. In other words, it appears that cereals were not part of the oak-pistachio-almond complex at this location and that this site - like contemporary Natufian sites - represents sedentism without agriculture. I have speculated that tree fruits provided much of the villagers' sustenance, in the absence of cereals (Hole 1996). By 9000 BC, after the Younger Dryas, Qermez Dere, located on the upper Mesopotamian Plain just south of the Jebel Sinjar, lay within the cereal belt, for the presence of einkorn, barley, legumes and pistachio nuts was reported (Watkins 1990). The first sites on the Jezireh Plain may have benefited from the massive increase in carrying capacity that the spread of wild cereals engendered (Hillman 1996). Whether cereals were present or not, each of these sites fits with the same late Epipaleolithic pattern of hunting and gathering in relatively rich environments which is manifest in the southern Levant among Late Natufian sites, and on the Euphrates at Abu Hureyra, Mureybit and Jerf el-Ahmar. The presence or absence of sites in any particular region before, during and after the Younger Dryas no doubt depended on the relative richness of local resources. Only after the Younger Dryas were rich resources widely spread.

In the Zagros the first known settlements were Zawi Chemi and Shanidar Cave, both of which are considered to be 'proto-Neolithic'. These sites, which were occupied during the onset of the Younger Dryas, were situated, like Hallan Çemi, in a forest refuge (Solecki 1981). In each case, occupation terminated during the Younger Dryas. It took another 2000 years before the first settlements appear in the true interior uplands. The first of these, Asiab, is a long-term but seasonal hunters' camp which carries on in the Epipaleolithic tradition (Braidwood et al. 1961; Bökönyi 1977). Significantly this is the last indication of such an adaptation. True agricultural villages also appear in the uplands of the Zagros at about the same time, as exemplified by Ganj Dareh at 8000 BC and Abdul Hosein, probably a little later. This first evidence of agriculture in the Zagros occurs 1000 years later than in the Levant. Both agriculture and stock-raising are attested at these Zagros sites which lie near the modern environmental limits for agriculture, and while neither enjoys a broad expanse of good alluvial soil, both have excellent access to still higher pastures. Ganj Dareh is a tiny site at 1400 m elevation, filled in part with small cubicles that impress one more as storage containers than as dwellings (Hole 1987a). Abdul Hosein at 1600 m asl is higher still, but it may actually be substantially later than Ganj Dareh (Hole 1987b). However, settlement of the eastern arc of the Fertile Crescent was anything but even. Jarmo and Ali Kosh, with locations that seem more favorable than those of Ganj Dareh or Abdul Hosein, were founded 1000 years after those sites.

A series of early villages with pottery, dating to around 6000 BC, is found in the intermediate valleys. One of these, Sarab in the Kermanshah Plain, gives evidence of initially having been a camp of transhumant herders who did not practise agriculture, at least in this location (Braidwood et al. 1961; Hole 1987a). The contemporary site of Guran in a nearby valley shows a similar practice of early herding followed by settled agriculture (Mortensen 1963, 1972). That transhumance was practised in the past is indicated by the occurrence of Neolithic sherds typical of Deh Luran, in mountain caves and shelters (Hole and Flannery 1967). The finding of a camp area used around 6000 BC by tent-dwellers would appear to solidify the case for systematic transhumance at that time (Hole 1974; Pires-Ferreira 1977), but we cannot say whether the practice dates back to the preceramic periods. This evidence, limited though it may be, suggests that the initial penetration of the Zagros was by herding people who carried agriculture with them seasonally to their summer pastures and then to their winter camps below the tree zone in the Fertile Crescent.

The climatic environment

There are no analogs for the climatic conditions that existed at the time agriculture began and spread. This is well illustrated empirically by the recent work on the Euphrates where, in the 11th-10th millennia (uncalibrated), the vegetation included rye and einkorn, along with arboreal species such as Pistacia, Rhannus and Quercus, which are adapted to cooler, wetter conditions than exist today. “The Neolithic vegetation resembles neither the surrounding steppe nor the Mediterranean vegetation to the west. It appears to correspond with more continental zones at higher altitudes” such as are found today above 800 m (Willcox 1996). Blumler (1996) also reminds us that “species which associated with each other in the past do not do so today and vice versa.” Further, that “present-day plant communities almost certainly were not identical 10,000 years ago, and that whole communities did not shift north and south, or up and down slope, as climate changed. Rather, each individual species shifted location according to its own requirements and adaptations, forming new species groupings” (Blumler 1996).

Pollen profiles which reflect vegetational changes are the principal source of information about past climates in the Near East. These show that “essentially modem” conditions had been established throughout the Near East by around 6000 years ago (van Zeist and Bottema 1991; Roberts and Wright 1993). This date coincides with the termination of rapid sea level rise and with the attenuation of the Climatic Optimum. Since that time there have been relatively minor changes in global climate despite perturbations that sometimes had significant local effects.

Hillman has recently provided a detailed interpretation of the pattern of vegetational changes as they relate to agricultural potential. He refers to enormous increases in carrying capacity when cereals invaded the steppe and he traces this movement from its inception in the northwest Levant ca. 13,000 BC and its spread eastward to Zeribar at ca. 9000 BC (Hillman 1996). He argues that pistachio and almond trees were followed in succession by oaks. Cereals came in with the former, he says, and stayed with the latter. He sees oaks migrating at a rate of 150-200 km/millennium and terebinth at 200-300 km/millennium (Hillman 1996). He maintains that the forest is ecologically linked with the grasses which “spread together with (or slightly behind) the first terebinths” (Hillman 1996). Although this association is clearly manifest today, it may not have been so tightly linked in the past, as the evidence from Hallan Çemi suggests. In the Anatolian Tigris drainage, there was a relict forest with pistachio-almond but there is “a relative paucity of wild grasses” although pulses are common (Rosenberg 1994).

Another version of regional vegetational change is given by El-Moslimany (1983, 1986, 1994) who sees summer rainfall as a significant factor during the Climatic Optimum. In her scenario this resulted from a shift in the ITCZ northward to near the head of the Gulf (El-Moslimany 1983). As evidence she cites Gramineae in peat deposits from Bubiyan Island at the head of the Persian Gulf, as well as Poaceae/Chenopod+Artemisia ratios in the Zagros pollen cores. The early Holocene records uniformly indicate much higher percentages of grass (Poaceae) than are found in surface samples today, even in areas protected from grazing (El-Moslimany 1994). Since “grasses play a dominant role in vegetation only where there is adequate moisture during the growing season” and “grasses adapted to summer aridity produce less pollen than those growing in regions of summer precipitation” she infers that a high P/C+A ratio implies warm-season precipitation. Using this line of reasoning she finds that summer precipitation extended to northern Mesopotamia; for example to Mureybit on the Euphrates where “Poaceae pollen reaches 60-70% of the total pollen between 8000 and 7600 BP” whereas today it “comprises only 3.8% of modern samples” (El-Moslimany 1994). El-Moslimany concludes that “even at its maximum, summer rainfall was probably only a small percentage of total annual precipitation” (El-Moslimany 1994). She sees a shift throughout the region to dominant winter rainfall after 6000