Previous Page Table of Contents Next Page

Part 4. Domestication of Crop Plants

Use of Historical and Archaeological Information in Lentil Improvement Today - W. Erskine
What Can Molecular Markers Tell Us about the Process of Domestication in Common Bean? - Paul Gepts
On the Origin and In Statu Nascendi Domestication of Rye and Barley: A Review - V. Jaaska
Plant-Gathering Versus Plant Domestication: An Ethnobotanical Focus on Leafy Plants - F. Ertug
Origins and Domestication of Mediterranean Olive Determined through RAPD Marker Analyses - G. Besnard, A. Moukhli, H. Sommerlatte, H. Hosseinpour, M. Tersac, P. Villemur, F. Dosba and A. Bervillé

Use of Historical and Archaeological Information in Lentil Improvement Today - W. Erskine


Lentil (Lens culinaris Medikus) is a short-statured, annual, self-pollinating, food legume mainly grown in the Indian subcontinent, the Mediterranean region and North America. The crop is grown in dryland cereal-based rotations because of its nitrogen-fixing ability, its high-protein seeds for human consumption, and its straw, which is a valued livestock feed.

ICARDA has a global mandate for research on lentil improvement and is situated in the Near East arc, where the crop was domesticated. The temporal dimension given by the history of the spread of the lentil following domestication and the consequent selective forces give a perspective of applied evolution to plant breeding which is highly relevant to ICARDA's current role in any new 'spread' of the crop.

Lentil origins

The putative progenitor of the cultivated lentil is Lens culinaris subsp. orientalis (Boiss.) Ponert which is distributed from Greece in the west to Uzbekistan in the east, and from the Crimean Peninsula in the north to Jordan in the south (Ladizinsky 1979, Cubero 1981). The oldest carbonized remains of lentil are from Franchthi cave in Greece dated to 11,000 BC and from Tell Mureybit in Syria dated 8500-7500 BC (van Zeist in Zohary 1972; Hansen and Renfrew 1978). But as it is not possible to differentiate wild from cultivated small-seeded lentil, the state of domestication of these and other carbonized remains in the aceramic farming villages in the 7th millennium BC in the Near East arc is unknown. The finding of a large hoard of lentil (about 1.4 million seeds) at Yiftah-el dated to 6800 BC is, however, suggestive of domestication (Garfinkel et al. in Zohary 1992). The oldest find of lentil seeds that are larger than wild seeds, and therefore unequivocally domesticated (Helbaek 1969), was at Tepe Sabz, Iran; they have been dated to 5500-5000 BC. The overlap in the distribution of wild lentil and the early archaeological record indicates that lentil was domesticated in the Near East arc.


From these beginnings the crop spread to the Nile, and to Central Europe via the Danube. As lentils are repeatedly found in the early agricultural settlements of the 5th millennium BC in Europe, situated outside the distribution of L. culinaris subsp. orientalis, this is indicative of earlier domestication. Lentils were definitely associated with the start of the 'agricultural revolution' in the Old World, which was initiated by the domestication of einkorn and emmer wheats, barley, pea, flax and lentil (Zohary 1976). The crop was part of the assemblage of Near Eastern grain crops introduced to Ethiopia by the invaders of the Hamites. From the Bronze Age onward, maintained itself as an important companion of wheat and barley throughout the expanding realm of Mediterranean-type agriculture.

The dissemination eastward of the Near Eastern grain crops, including lentil, reached Georgia in the 5th and early 4th millennia BC. The crop appears in the archaeological record in India around 2500 BC as part of the Harappan crop assemblage. Alphonse de Candolle (1882) wrote that on linguistic grounds, “It may be supposed that the lentil was not in this country (India) before the invasion of the Sanskrit-speaking race.” The invasion occurred before 2000 BC. The crop probably reached its current Old World range about 3000 years ago. It was carried to the New World after Columbus.

Selective forces operating during spread

Our knowledge of the selective forces operating during the spread of the culture of lentil derives from an analysis of the variation found in current landrace populations from different geographic regions. Following widespread collecting and evaluation of landraces in the 1920s, Barulina (1930) classified the assembled variation into six groups (grex varietatum), each of which was geographically differentiated and also characterized by a complex of morphological characters, mainly qualitative, common within a group but differing in other groups. This type of geographic association is mirrored by variation in quantitative morphological traits (Erskine et al. 1989).

Such morphological characters are readily observable and consequently often the subject of human selection. Geographic differentiation between landraces also has been found for other more cryptic, ecophysiological factors, such as those resulting from selection for soil conditions and for climatic conditions, as the following examples illustrate.

Iron-deficiency symptoms are observable on some accessions of lentil grown in calcareous soil. In a germplasm collection of 3512 accessions originating from 18 countries, landraces from those Mediterranean countries where lentil originated (Syria and Turkey) exhibited Fe-deficiency symptoms only at very low frequencies (<1.5%) (Erskine et al. 1993). In these regions the soils are generally highly calcareous with pH >8.0, conditions known to reduce Fe availability. Those landraces exhibiting symptoms of Fe deficiency mostly originate from relatively warm climates, such as India (37.5% accessions showing Fe deficiency) and Ethiopia (30%), where the pH is generally 6.5-7.0 (Fig. 1). This arose from either the chance introduction to these regions of Fe-deficient founder populations, or selection pressure in favor of either Fe-deficient types or genes linked to Fe-deficiency.

In a survey of patterns of morphological variation in the world germplasm collection of lentil, phenology was found as the key to the adaptation of the crop on a macrogeographic scale (Erskine et al. 1989). To further understand phenology, the flowering responses to temperature and photoperiod of a world collection of 369 accessions from 13 major lentil-producing countries (25 randomly selected accessions per country) together with lines from the ICARDA breeding program were studied (Erskine et al. 1994). The distribution of country means for sensitivity to temperature and photoperiod illustrates the responses to selection for adaptation to new ecological environments following the spread of the crop from its origin (Fig. 2).

Fig. 1. Map indicating countries with a high frequency of winterhardiness (hatching) among lentil germplasm and countries with a high frequency of accessions showing iron deficiency (striped) (from Erskine 1997).

Fig. 2. Mean values for lentil accessions of different countries of origin for responses (x104) in flowering to temperature and photo-period. The respective means for the countries of West Asia are joined to illustrate the variation inherent in the region where the crop was domesticated. Country names are abbreviated as follows: Afg = Afghanistan, Chl = Chile, Egy = Egypt, Eth = Ethiopia, Grc = Greece, Ind = India, Irn = Iran, Jor = Jordan, Leb = Lebanon, Pak = Pakistan, Syr = Syria, Tur = Turkey, Rus = Russia (ICA represents ICARDA selections) (from Erskine et al. 1994).

Dissemination to lower latitudes such as into Egypt, Ethiopia and India was accompanied by a reduction in photoperiodic response. Obligate photoperiodic control of the onset of flowering ensures that flowering starts annually in the same calendar period, irrespective of fluctuations in temperature. Consequently, selection against photoperiodic control in a long-day plant such as lentil implies an adaptation to relatively short days, which occur at low latitudes and which would otherwise delay flowering to an unacceptable extent. Under these conditions, the crop relies rather more on temperature than photoperiod to ensure that flowering occurs at an ecologically and agronomically appropriate time. There was also evidence that flowering in the subtropical group is more temperature-sensitive than in West Asian germplasm, on average.

Movement from West Asia to higher latitudes, for example to Russia, resulted in a modest reduction in photoperiod sensitivity and an increase in temperature sensitivity, probably reflecting the change in sowing date from winter to spring.

Temperature sensitivity to processes other than flowering, such as the base temperature for rate of germination (Tb), also has been altered during dissemination from the Near East to environments with higher temperature regimes. The mean Tb of 15 randomly selected accessions from Ethiopia and India was higher than that of similar samples from Lebanon and Turkey (Ellis and Hong 1995).

The degree of winterhardiness also has been affected in the spread of the lentil from its origins as a winter-sown crop in the Near East. A world collection of 3910 accessions was screened for winterhardiness near Ankara, Turkey, in the winter of 1979/80, when temperatures dropped to a low of -26.8°C and there were 47 days of snow cover (Erskine et al. 1981). In total, 238 accessions were undamaged by the cold winter. Their origins were mostly from Chile (frequency of winterhardy accessions was 28%), Greece (33%), Syria (33%) and Turkey (15%), where selection for winterhardiness had occurred through winter sowing (Fig. 1), although lentil production is found in countries farther north, where the winter is colder, such as Russia (5%), and Hungary (4%), where sowing is exclusively done in spring. The frequency of winterhardy accessions from countries with warmer winter environments, such as Egypt, Ethiopia, India and Pakistan, was <1%.

Isozyme data of the locus aat-p from the world collection show differences among countries and regions in the frequency of alleles and the extent of polymorphism, reflecting adaptation to local conditions (Skibinski et al. 1984). There is evidence that a low frequency of Aat-pF is associated with the spread into South Asia, whereas a higher frequency of Aat-pF is characteristic of northern European material.

Such regional differences result from the spread of the crop to new physical environments with the consequent natural and artificial selection for local adaptation. To summarize with an extreme example, the spread of lentil from the Near East into India resulted in a loss of winterhardiness and the ability to extract iron from calcareous, high-pH soils, a reduction in photoperiodic sensitivity for flowering, increased intrinsic earliness in flowering, an increase in temperature sensitivity for flowering, and in the base temperature for rate of germination. This list of factors affecting adaptation is not exhaustive. Each factor individually may be of minor importance, but collectively they illustrate the complex of interacting ecophysiological factors that determine adaptation. Superimposed on these natural selective forces is the effect of human selection on observable morphological traits, particularly the seed, for culinary and agronomic preferences.

The pilosae group of the Indian subcontinent is also characterized by two endemic qualitative morphological traits: precocity in flowering and maturity, and a low biomass. Furthermore, lentil germplasm from India is among the least variable among lentil-producing countries, despite India being the largest lentil-producing country in the world. Recent evidence from isozyme (7 loci) and random amplified polymorphic DNA (RAPD) (22 loci) analysis indicates a similar striking difference between germplasm from South Asia and that from the rest of the world; additionally South Asian germplasm was found low in diversity (Ferguson et al. 1998). These data confirm the morphological and physiological distinctiveness of pilosae germplasm. The simplest explanation has the introduction around 2000 BC of a founder population adapted to the new environment that was probably both small in number and low in variability which resulted in a genetic bottleneck in South Asia (Erskine et al., unpublished). This still limits breeders' progress today. Indeed, a reconstruction of the phenological problems associated with the initial spread of the crop into the Indo-Gangetic Plain was inadvertently made with the introduction into that region of lentil selected in West Asia through ICARDA international nurseries in its early years (Ceccarelli et al. 1994). Lentil selected in West Asia, when sown in India and Pakistan, mostly came into flower as the indigenous lentils were maturing.

Use of information in breeding

Armed with an understanding of the specific adaptation of lentil, the local constraints to production and the various consumer requirements of different geographic areas for seed, the breeding program at ICARDA aims to produce genetic material suitable for national cooperators. The program has been designed as a series of separate, but finely targeted streams linked closely to national breeding programs (Ceccarelli et al. 1994) (Table 1). As we have seen, many of the selective forces important in the adaptation of the crop are environment-specific, such as temperature and photoperiodic sensitivities. Clearly, selection undertaken in conditions very different from those in the target environment will have a lower response in the target environment than selection conducted directly within the target environment. Harnessing the specific comparative research advantages of ICARDA and its national program partners, crosses (previously agreed with cooperators) are made and segregating generations advanced at ICARDA. The selection of bulk segregating populations is undertaken in the target environment by national programs. Since 1985, we have made specific crosses for national programs, for example, Algeria, Bangladesh, Egypt, India, Jordan, Morocco, Nepal, Syria and Turkey. Details may be found in ICARDA (1995).

We have made a major effort to widen the genetic base in the Indian subcontinent using three approaches, namely plant introduction, hybridization and mutation breeding. As mentioned above, plant introductions from West Asia flower as indigenous material matures in the Indian subcontinent. The asynchrony in flowering has isolated pilosae lentils reproductively. However, the introduction of ILL 4605, an early flowering, large-seeded line, has resulted in its release as 'Manserha 89' for wetter areas of Pakistan and its widespread use as a parent in breeding programs in the region. Hybridization between pilosae and exotic germplasm, primarily at ICARDA, followed by selection in the Indian subcontinent has resulted in cultivars with improved disease resistance and yield in Bangladesh and Pakistan. Mutation breeding has given new morphological markers and several promising lines.

Table 1. Target agro-ecological regions of production of lentil and key breeding aims.


Key traits for recombination

Mediterranean low to medium elevation

300-400 mm annual rainfall

Biomass (seed + straw), attributes for mechanical harvest and wilt resistance

<300 mm annual rainfall

Biomass, drought escape through earliness


Biomass, attributes for mechanical harvest and rust resistance


Seed yield, response to irrigation, earliness and wilt resistance

High elevation

Anatolian highlands

Biomass and winterhardiness

N. African highlands

Seed yield and low level of winter hardiness

South Asia and E. Africa

India, Pakistan, Nepal, Ethiopia

Seed yield, early maturity and resistance to rust, ascochyta and wilt


Seed yield, extra earliness and resistance to rust and Stemphylium

The example of the widening of the genetic base of the lentil in South Asia illustrates the value of the historical perspective in plant improvement. Although ICARDA and its national program partners are undertaking the first systematic lentil improvement program with access to a wide range of genetic diversity and modern biotechnological tools, the efficiency of this research has been increased by an awareness that we belong to the approximately 200th generation of cultivators/selectors who have grown the crop since its domestication.


Barulina, E.I. 1930. The lentils of the USSR and other countries. Tr. po Prikl. Bot. Genet. Sel. [Bull. Appl. Genet. & Select.] Suppl. 40:1-319 [in Russian].

Ceccarelli, S., W. Erskine, J. Hamblin and S. Grando. 1994. Genotype by environment interaction and international breeding programs. Exp. Agric. 30:177-188.

Cubero, J.I. 1981. Origin, taxonomy and domestication. Pages 15-38 in Lentils (C. Webb and G. Hawtin, eds.). CAB, Farnham, UK.

de Candolle, A.P. 1882. Origins of cultivated species. Reprint 1967. Hafner, London.

Ellis, R.H. and T.D. Hong. 1995. The effect of cool temperatures on the germination of lentil. Pages 95-106 in Autumn sowing of lentil in the Highlands of West Asia and North Africa (J.D.H. Keatinge and I. Küsmenoglu, eds.). Central Research Institute for Field Crops, Ankara, Turkey.

Erskine, W. 1997. Lessons for breeders from land races of lentil. Euphytica 93:107-112.

Erskine, W., A. Hussain, M. Tahir, A. Baksh, R.H. Ellis, R.J. Summerfield and E.H. Roberts. 1994. Field evaluation of a model of photothermal flowering responses in a world lentil collection. Theor. Appl. Genet. 88:423-428.

Erskine, W., K. Myveci and N. Izgin. 1981. Screening a world lentil collection for cold tolerance. LENS Newsl 8:5-8.

Erskine, W., Y. Adham and L. Holly. 1989. Geographic distribution of variation in quantitative characters in a world lentil collection. Euphytica 43:97-103.

Ferguson, M, L.D. Robertson, B. Ford-Lloyd and H.J. Newbury. 1998. Contrasting genetic variation amongst lentil landraces from different geographical origins. Euphytica (in press).

Hansen, J. and J.M. Renfrew. 1978. Paleolithic-Neolithic seed remains at Franchthi cave, Greece. Nature 71:349-352.

Helbaek, H. 1969. Plant collecting, dry-farming and irrigation agriculture in prehistoric Deh Luran. Pages 383-426 in Prehistory and Human Ecology of the Deh Luran Plain: An Early Village Sequence from Khuzistan (F. Hole, K.V. Flannery and J.A. Neely, eds.). Memoirs Museum Anthropology No. 1. Ann Arbor, University of Michigan, USA.

ICARDA. 1995. Annual Report for 1994 of the Germplasm Program: Legumes. International Center for Agricultural Research in the Dry Areas (ICARDA), Aleppo, Syria.

Ladizinsky, G. 1979. The origin of lentil and its wild genepool. Euphytica 28:179-187.

Skibinski, D.O.F., D. Rasool and W. Erskine. 1984. Aspartate aminotransferase allozyme variation in a germplasm collection of the domesticated lentil (Lens culinaris). Theor. Appl. Genet. 68:441-448.

Zohary, D. 1972. The wild progenitor and place of origin of the cultivated lentil Lens culinaris. Econ. Bat. 26:326-332.

Zohary, D. 1976. Lentil. Pages 163-164 in Evolution of Crop Plants (N.W. Simmonds, ed.). Longman, London, UK.

Zohary, D. 1992. Domestication of the Neolithic Near Eastern crop assemblage. Pages 81-86 in Préhistoire de l'agriculture: nouvelles approches expérimentales et ethnographiques. Monographie du CRA No. 6, Centre de Recherches Archéologiques (PC. Anderson, ed.). CNRS, Paris, France.

What Can Molecular Markers Tell Us about the Process of Domestication in Common Bean? - Paul Gepts


Alphonse de Candolle, the founder of the study of crop evolution, was primarily interested in biogeography, the study of distribution of plants in relation to their environment. In his opinion, among the factors influencing this distribution are 'historic' factors such as glaciation events that are not related to the intrinsic adaptation characteristics of the species. When he considered crop plants, he posited that one particular historic event - domestication - had had a paramount importance in determining their adaptation and, hence, distribution, as well as their genetic characteristics (de Candolle 1882).

Four types of evidence could be utilized to determine the origin of domestication of crop plants according to de Candolle (1882). These were archaeological (or more specifically archaeobotany), botanical (i.e. the distribution of the wild, ancestral relative), historical (or the existence of a written record documenting the existence or importance of the crop) and linguistic evidence (i.e. the existence of words designating the crop or objects or concepts related to the crop in native languages). Of these four types, the first two provide the most abundant and reliable evidence (Harlan and de Wet 1971). Since the time of de Candolle, however, additional scientific methods have increased considerably our power to determine centers of crop domestication based on archaeological and botanical arguments (Smith 1995).

Although common bean is not a species originating in the Near East, the focus of the present symposium, it presents nevertheless an interesting domestication history, which is only recently being elucidated. The first botanically accurate description of common bean in European herbals dates from 1542. Whereas barely 50 years had passed after the conquest of the Americas, the actual origin of the crop had already been lost to the writer of the herbal and later European botanists. Linnaeus assigned the origin of common bean to India; de Candolle (1882) himself expressed doubts that common bean had been domesticated in the Americas. It was only during and after the Second World War that it became generally admitted that common bean originated in the Americas. The conclusion was based on the discovery of wild common bean in Argentina (Burkart and Brücher 1953) and Guatemala (McBryde 1947) and archaeological remains in the Americas (Kaplan and Kaplan 1988).

Based on the extensive distribution of the wild relative (Fig. 1), Heiser (1965) and Harlan (1975) speculated that common bean originated through multiple domestications as did other crops originating in the Americas, such as cotton, pepper and amaranth. In the variability present in the Cambridge (UK) collection, Evans (1976) also observed that there was a Mesoamerican cultivar group with smaller seeds and an Andean group with larger seeds. None of these observations, however, provided substantial experimental evidence and hence a clearer picture of the actual process of domestication for this species. For example, one could ask whether the two cultivated groups identified by Evans (1976) resulted from distinct domestications or a single domestication event followed by divergence into a small-seeded and a large-seeded group. One could also ask where these single or multiple domestications took place.

The archaeological record of the origins of agriculture in the Americas is rather limited. In Mesoamerica, there are only five major sites. This important type of information therefore pales in comparison with what is available in the Near East (Zohary and Hopf 1993). In part, this may be due to the relatively humid environment required by many crop relatives in the Americas, which render long-term conservation of crop remains less likely. Therefore, in relative terms compared with archaeological information, there is a relatively higher emphasis on botanical evidence to decipher crop evolution in the Americas than in the Near East.

In this chapter, I would like to show how the use of molecular marker information has increased our understanding of the domestication process in common bean (Phaseolus vulgaris L.), a crop originating in the Americas. In addition, I hope to demonstrate how molecular analyses can bridge the gap that has separated botanical and archaeological evidence, the twin pillars on which rests the study of crop evolution. The power of molecular biology to change scientific paradigms is by no means unique to this field but extends to many other fields as well.

Once upon a time there was a bean...

General methodological principles

Our current understanding of the origins of domestication in common bean is based on a two-pronged approach relying on the study of wild-growing P. vulgaris beans, on one hand, and the use of biochemical or molecular markers, on the other. The use of wild beans is predicated on the fact that their dispersal occurs over much smaller distances (at most a few meters after each vegetation cycle, if at all) than their domesticated descendants. The latter are subject to long-distance exchange, trade, gift or human migration, all of which tend to obscure the original geographic distribution patterns of genetic diversity. A prerequisite for such an approach is of course the availability of wild germplasm. Recent explorations have provided a more complete representation of the actual distribution of wild P. vulgaris (Fig. 1) (Debouck et al. 1993; Freyre et al. 1996; Beebe et al. 1997). Wild beans are distributed from approximately 30°N Lat. to 30°S Lat. at mid-elevations from 1000-2500 m asl in regions with moderate rainfall. A common characteristic of the environments in which wild beans are found is the existence of a marked dry season, an important feature needed for seed dispersal.

Genetic relatedness is traditionally assessed with morphological phenotypic traits. Plants with a similar phenotype are inferred to be genetically related either through common ancestry or through geneflow (hybridization). Occasionally, however, inferences based on phenotypic similarity are erroneous. Plants can arise from different selection episodes leading to the same phenotype (convergent evolution). In addition, similar phenotypes can have a different genetic basis (genetic heterogeneity). In both cases, an inference of genetic relatedness would be unwarranted. In contrast, biochemical, and even more so, molecular markers provide information on genetic relatedness that is generally (although not always) devoid of selective influences.

Fig. 1. Distribution of the three major genepools of wild Phaseolus vulgaris L. in Latin America.

Not all biochemical and molecular markers were created equal! Differences have been observed with regard to the ease of analysis, reproducibility, level of polymorphism, number and genome distribution of loci (Rafalski and Tingey 1993). Of particular note is the molecular basis of polymorphisms. The more complex the basis, the more desirable the marker, because complexity will reduce the probability of repeat mutations, especially in conspecific materials, which are, by definition, closely related. For example, polymorphisms based on DNA nucleotide substitutions are simple changes that could presumably occur repeatedly. Analyses based on single or few RAPD or RFLP markers are therefore unreliable in determining genetic relatedness (Skroch et al. 1992). Conversely, analyses that sample a large number of changes at the molecular level, because they include either a large number of markers or complex markers such as actual DNA sequences or multigene families, are more robust with regard to repeated changes. For example, banding patterns of seed proteins such as phaseolin in common bean, when analyzed in denaturing SDS-polyacrylamide gel electrophoresis, represent a composite image resulting from changes at the DNA nucleotide level (duplication of genes, within-gene duplications, nucleotide substitutions) and at the translation level (co - and post-translational modifications). Therefore each pattern is unlikely to have arisen more than once. As a consequence, genotypes that exhibit the same protein pattern are likely to share a common ancestor (Gepts 1988).

One of the major experimental difficulties in crop evolutionary studies is the distinction one has to make between ancestry (domestication) and geneflow as a cause of genetic relatedness. This is particularly the case given that wild ancestors and domesticated descendants belong to the same biological species and can therefore freely hybridize. Hence, similarity between ancestors and descendants even at the molecular level cannot be assumed to be due to domestication without additional data. For example, DNA sequence data and the utilization of an outgroup (e.g. closely related species) can shed additional light on this problem because they can provide information on the nature of the polymorphism and the direction of the change, i.e. the ancestral vs. descendant states of the molecular polymorphism. This will be illustrated below with the case of phaseolin.

Domestication pattern

The origin of the Phaseolus vulgaris progenitor in its wild state can be traced to the Pacific slope of Ecuador and northern Peru, based on the distribution pattern of direct repeats in phaseolin seed protein genes (Kami et al. 1995). Phaseolin is coded by a small multigene family inherited as a single Mendelian unit and located on linkage group B7 (Brown et al. 1981; Nodari et al. 1993). The phaseolin locus is at most 190 Kb long (Llaca and Gepts 1996) and contains 6 to 9 genes (Talbot et al. 1984). Differences among these phaseolin sequences include nucleotide substitutions and the presence or absence of tandem direct repeats. Three repeats have been identified so far: a 21 bp repeat in the third intron, a 15 bp repeat in the fourth exon and a 27 bp repeat in the sixth exon.

The existence of these tandem repeats provides an interesting landmark to follow the evolution of phaseolin genes. Indeed, the generation of a repeat is a more likely event than the precise loss of such a repeat, which would be needed to restore a sequence without repeats. Hence, phaseolin genes without repeats represent probably a more ancestral state than phaseolin genes with repeats. This was verified by analyzing the presence of tandem repeats with a polymerase chain reaction (PCR) test of phaseolin sequences in two subspecies of a related species, P. coccineus subsp. polyanthus and P. coccineus subsp. coccineus. In these two taxa, none of the three repeats could be identified, suggesting that the repeatless state is indeed the ancestral state.

A survey was then conducted among wild P. vulgaris to determine if any phaseolin haplotypes lacked genes with tandem repeats. Only wild bean populations from Ecuador and northern Peru on the Pacific slope of the Andes appeared to be devoid of phaseolin genes with tandem repeats (Fig. 2; Kami et al. 1995). It was inferred that these wild populations therefore represent the presumed ancestor of the species. This conclusion is somewhat unexpected as most wild Phaseolus species are distributed in Mexico. One could have thought that P. vulgaris would have originated in that region as well and that other wild P. vulgaris populations distributed elsewhere were ultimately derived from Mexican ancestors. It may be, however, that some wild Phaseolus ancestors ('proto-vulgaris' and other species) were dispersed from Mexico to Central America and the Andes to give rise to locally distributed wild Phaseolus species, such as P. costaricensis (Freytag and Debouck 1996) and P. augusti (Freyre et al. 1996). From such a local distribution, a species such as P. vulgaris could have acquired a secondary, broader distribution by dispersal toward the north and south (see below). It is worth remembering that lima bean (P. lunatus) also shows an extensive distribution of its wild relative from Mexico to Argentina and that the species consists of two genepools, an Andean genepool originating on the Pacific side of the Andes in Ecuador and northern Peru (although at lower altitudes than wild P. vulgaris) and a 'Mesoamerican' genepool distributed from Mexico to the southern Andes. Although the domestication region of the 'Mesoamerican' genepool is unknown so far, it is likely to be located somewhere in Mesoamerica, given that most Mesoamerican cultivars are located there (Salgado et al. 1995).

Fig. 2. Polymerase chain reaction amplification of part of the phaseolin genes surrounding the 15 repeat. 1: T phaseolin (Andean); 2: S phaseolin (Mesoamerican); 3: I phaseolin (Ecuador and northern Peru). Adapted from Kami et al. (1995).

From the small nuclear area in Ecuador and northern Peru, wild P. vulgaris were dispersed toward both the north (Colombia and Venezuela, Central America and Mexico) and the south (southern Peru, Bolivia and Argentina) to achieve their current distribution (Fig. 1). In a subsequent step, multiple domestications took place in the Mesoamerican and Andean regions. At least one domestication may have taken place in west-central Mexico (Jalisco), a conclusion based on phaseolin data. Most Mesoamerican cultivars exhibit the same phaseolin type ('S'-type). Wild beans with the same phaseolin type and without morphological signs of past hybridizations to domesticated beans are concentrated in the Mexican state of Jalisco, suggestive of a possible domestication site in that area (Gepts 1988). Interestingly, Doebley et al. (1984) suggested a similar domestication area for maize.

A second major domestication took place in the southern Andes although a more specific region awaits confirmatory data. Phaseolin data suggest possible multiple domestications within the Andean region because at least four phaseolin types have been identified among Andean beans. Alternatively, the original populations that were domesticated were polymorphic for phaseolin. Further data are needed to distinguish between these possibilities. A third, likely minor domestication may have taken place in Colombia, although regions in Central America deserve further study because samples of wild P. vulgaris from that area were unavailable.

Issues raised by these observations

1. Molecular analyses appear to allow us to locate more precisely the centers of origin and domestication. It remains to be determined, however, if these results can be confirmed with further analyses of additional markers and plant materials.

2. Data from several crops mutually can potentially reinforce the identification of centers of domestication. When several crops appear to have been domesticated in the same well-defined area, such as maize and common bean in west-central Mexico, one can infer that several thousands of years ago, people in these areas developed the technology of domestication and agriculture and applied it to several crops simultaneously.

3. Once domestication was accomplished in a certain area for a certain crop, it may not have been attempted again within a certain distance. It is as if crops, or the people who domesticated them, had gained a competitive advantage, such that domestication was not attempted again. One can speculate as to the nature of this advantage. Perhaps the assemblage of domestication traits characterizing a crop is an unusual one and could only be reconstituted with very low probability elsewhere. Alternatively, the new agricultural societies overwhelmed other, non-agricultural societies by their increased population growth caused by sedentism and the attendant reduction in birth spacing.

4. Molecular biology provides a way of characterizing genetic diversity in archaeological remains. PCR amplification of remnant DNA sequences and sequencing of the amplification products provides a way of comparing genetic diversity in archaeological remains and contemporary plant materials. Such an analysis has been accomplished by Goloubinoff et al. (1993) for maize and Brown et al. (1994) for wheat. The probability of DNA amplification is higher with desiccated than with charred remains, although Goloubinoff et al. (1993) claim to have amplified DNA from charred remains too.

Post-domestication divergence

Common bean is perhaps best known for the high level of phenotypic diversity, particularly for seed type (color, color pattern, size and shape). This broad diversity is a consequence of the wide range of cultural environments under which common bean is grown, both in the Americas and elsewhere. This diversity has made it difficult to recognize patterns of genetic diversity in the domesticated genepool because of the possibility of convergent and reticulated evolution.

Analyses of biochemical traits such as isozymes and phaseolin seed protein have provided additional information on this problem. Koenig and Gepts (1989) and Singh et al. (1991a) analyzed isozyme diversity in wild and domesticated germplasm, respectively. From these studies, it was possible to identify clusters of cultivars within the Andean and Mesoamerican genepools that shared a common isozyme allele. Further analyses of phenotypic diversity were conducted with multivariate statistical analyses (canonical and discriminate analyses) (Singh et al. 1991b). Results showed that the isozyme clusters could also be distinguished by morphological, agronomic and ecological traits. For example, in the Mesoamerican genepool, three subdivisions or races can be distinguished (Singh et al. 1991b, 1991c). Race Mesoamerica represents cultivars originating predominantly from the warmer, more humid areas of Mexico, Central America and Colombia. Possibly, this race could be split further to account for an additional, minor domestication area in Central America or Colombia (the 'B' phaseolin cultivars, Gepts and Bliss 1986). Race Durango includes cultivars from the northern highlands of Mexico, which have a cooler, more arid environment. Race Jalisco consists of the most traditional common bean cultivars, originating predominantly in the humid central and southern highlands of Mexico. A similar subdivision in three ecogeographical races has been established for the Andean genepool (Singh et al. 1991b, 1991c).

Figure 3 summarizes the current status of the domestication scenario in common bean. The question marks refer to the racial diversification process (see below). The arrows indicate rare cases of geneflow between the two major genepools.

Fig. 3. Current domestication scenario in common bean. The question marks refer to the current lack of data regarding the actual pattern of domestication within the Andean and Mesoamerican regions. The arrows at the top of the figure indicate rare instances of geneflow between the Mesoamerican and Andean domesticated genepools.

Issues raised by these observations

1. One question arising from these observations is the cause of this racial diversification. In the Mesoamerican genepool, phaseolin data suggest a single domestication event because the domesticated genepool contains only one phaseolin electrophoretic type compared with over 15 in the corresponding wild genepool (Gepts et al. 1986). This suggests that racial diversification in the Mesoamerican genepool cannot be attributed to separate domestications but to events subsequent to domestication. One possibility is the existence of geneflow from wild to domesticated bean populations. As domesticated beans were disseminated from their presumed center of domestication in Jalisco, Mexico, they may have been subjected to geneflow from wild populations in other regions of Mexico. An AFLP diversity study of wild and domesticated common bean from Mexico does not appear to support this possibility as most domesticated beans are located in a distinct cluster from the wild beans. In turn, these observations suggest that racial diversification in the Mesoamerican genepool may have taken place as a consequence of selection of genetic diversity within geographically isolated populations after dispersal from the center of domestication. In the Andean genepool, phaseolin data (Gepts et al. 1986) suggest a multiple domestication pattern or, alternatively, domestication from a phaseolin-diverse wild population. Data are currently lacking to determine a possible role for geneflow in the Andean genepool.

2. If racial diversification in the Mesoamerican genepool involved selection after dispersal to other ecogeographic regions of Mexico and Central America, then it should be interesting to attempt to trace ancient human migrations in Mexico, which may be associated with the spread of agriculture and crops. Such a study would be very similar to the ones performed by Ammerman and Cavalli-Sforza (1984) and Sokal et al. (1991). In these studies they identify gradients of gene frequencies among contemporary European populations, the most important of which is a SE-NW gradient, which, the authors claim, stems from the introduction of agriculture from the Near East into Europe via diffusion.

What does it take to get a domesticated bean?

Wild relatives and their crop descendants show marked phenotypic differences, collectively called the domestication syndrome (Hammer 1984). As a consequence, they have been classified in different taxonomic species, which is unjustified given that often they can be freely crossed. Based on the traditional neo-Darwinian theory of evolution, one would expect that these marked phenotypic differences would be controlled by a large number of genes, each with a small phenotypic effect. This view reflects a view of evolution involving very gradual changes in environment and adaptation. This may not have been the case during domestication (nor in natural environments for that matter) where the domesticated and natural environments are quite different. Hence, adaptation to cultivation represented selection for a radically different environment over a time scale that would have been very short in an evolutionary context (even a few 100 generations should be considered a very short period in evolution). This suggested that major genes might have played a role during domestication because only such genes would have provided a fast response to the rapidly changing environment imposed by cultivation.

To test this view of evolution during domestication, an experiment was conducted to determine the inheritance of the domestication syndrome in common bean (Koinange et al. 1996). A recombinant inbred population was established from a cross between 'Midas' (a snapbean cultivar representing the 'ultimate' in bean domestication) and 'G12873' (a wild bean from Mexico). In this RI population, a linkage map was constructed with RFLP markers regularly spaced throughout the genome. In turn, this map was used to conduct a quantitative trait locus (QTL) analysis for quantitative traits and to map major genes involved in the domestication syndrome. The following traits were studied.

Pod dehiscence

The onset of the dry season coincides with the maturation phase of the life cycle of wild beans. During the progressive desiccation of the plant, fibers located in the pod walls and sutures shorten and, hence, cause the explosive dehiscence of the pods. (Without desiccation during the dry season, pods rarely open and seeds germinate in the pods, causing a marked reduction in fitness.) Pods of domesticated beans contain fewer or no fibers and therefore do not open at maturity. This difference in the capability of seed dispersal is one of the most important characteristics distinguishing wild and domesticated wild beans. It can be easily recognized in dry pod remains: dried wild pods are strongly twisted (owing to the oblique orientation of fibers in the pod walls), whereas dried domesticated pods are not or only slightly twisted. Presence or absence of fibers is controlled by the St gene.

Seed dormancy

Because the onset of the growing (wet) season is irregular, plants have adopted seed-dormancy mechanisms to insure against premature germination during the first rains. In common bean, dormancy is essentially impermeability of the seed coat, which prevents immediate imbibition of the seed. Prior to this study, no studies had ever been undertaken about the inheritance of dormancy in common bean.

Growth habit

Wild beans are viny plants that use shrubs or trees as support. During common bean evolution, there has been a selection for a bush growth habit. Growth habit consists of several subtraits, including the number of nodes and pods on the main stem, the length of the internodes, and type of stem termination (vegetative or indeterminate vs. reproductive or determinate). The latter traits are conditioned by the fin gene.

Photoperiod sensitivity

Photoperiod sensitivity plays a crucial role in scheduling of flowering, such that maturation will take place at the beginning of the dry season. Earlier maturation will result in reduced or lack of seed dispersal. Later maturation will prevent sufficient development of seeds before the dry season sets in. Common bean, having originated in the tropics, is a short-day (long-night) plant, which will flower only under daylengths of 12 hours or less. During dispersal from centers of domestication, dayleingth neutrality was selected, whereby common bean plants will flower under any daylength. Photoperiod sensitivity is controlled by the Ppd gene.

Seed size and color

Selection by farmers and consumers has led to larger pods and seeds with different colors, among them white (absence of pigmentation) conditioned by the p gene.

Four conclusions were drawn from the results. First, for many traits it was possible to identify major genes. For example, at least four genes controlled seed dormancy, one of which accounted for at least 50% of the phenotypic variation. Second, for most of the traits, more than 50% of the phenotypic variation could be accounted for in genetic terms, suggesting that heritability of most traits was high. Third, distribution of genes involved in the domestication syndrome appeared to be non-random, as there was a relatively higher concentration of genes on linkage groups 1, 2 and 7. Linkage group 1 appeared to be particularly important because it included genes for photoperiod (Ppd), determinacy (fin) and internode length.

Issues raised by these observations

1. The genetic control of the domestication syndrome in common bean, which is actually quite similar to the one observed in maize (Doebley and Steck 1991), reflects the process of domestication in that we observe major genes and high heritability for most traits. If farmers practised conscious selection, mutations with major effect would have been more easily detected than minor ones. Even in the case of unconscious selection, adaptation to the cultivated environment would have been more rapid with these major mutations. In addition, highly heritable traits would also have been more amenable to selection, especially in variable environments. Finally, the concentration of many genes for domestication in a limited number of genomic regions would have facilitated recovery of domesticated types after outcrossing between wild and incipient domesticates.

2. Given the simple genetic control of the domestication syndrome, domestication could have proceeded rapidly. There are at least two caveats for this statement:

· We have in many cases little information about the actual selection pressure that was applied during domestication. Archaeobotanical sequences encompassing wild to domesticated materials could help shed light on this problem.

· Arguments about speed of domestication are generally based on single traits (e.g. Hillman and Davies 1990). How long did it take, however, to assemble the domestication syndrome? It is unlikely that all mutations necessary for the domestication syndrome occurred in a single plant. This suggests that hybridization and recombination took place. However, many crops are predominantly self-pollinated. This reproductive mode would have slowed down the recombination among mutant genotypes compared with allogamous plants. Again, a sequence from wild to domesticated plants including different types of organs would be very useful.


Molecular biology methods can provide a bridge between the main study fields of crop evolution - archaeobotany and plant science - by providing information on putative centers of domestication, potential identification of plant (and human) migration patterns from the center of domestication, the time frame for the domestication process, and a molecular characterization of archaeological remains. Thus, by analogy with the Tigris and Euphrates Rivers of the Fertile Crescent, which after having flowed separately for most of their courses are finally united before they reach the sea, a closer association of these two fields can be achieved using molecular biology as a unifying tool.


Ammerman, A.J. and L.L. Cavalli-Sforza. 1984. The Neolithic Transition and the Genetics of Populations in Europe. Princeton Univ. Press, Princeton, New Jersey, USA.

Beebe, S., O. Toro, A. González, M. Chacón and D. Debouck. 1997. Wild-weed-crop complexes of common bean (Phaseolus vulgaris L., Fabaceae) in the Andes of Peru and Colombia, and their implications for conservation and breeding. Genet. Resour. & Crop Evol. 44:73-91.

Brown, J.W.S., F.A. Bliss and T.C. Hall. 1981. Linkage relationships between genes controlling seed proteins in French beans. Theor. Appl. Genet. 60:251-259.

Brown, T., R. Allaby, K. Brown and K. O'Donoghue. 1994. DNA in wheat seeds from European archaeological sites. Experientia 50:571-575.

Burkart, A. and H. Brücher. 1953. Phaseolus aborigineus Burkart, die mutmassliche andine Stammform der Kulturbohne. Züchter 23:65-72 [in German].

de Candolle, A. 1882. L'origine des plantes cultivées. English translation: The origin of cultivated plants. Appleton, New York, USA.

Debouck, D.G., O. Toro, O.M. Paredes, W.C. Johnson and P. Gepts. 1993. Genetic diversity and ecological distribution of Phaseolus vulgaris in northwestern South America. Econ. Bot. 47:408-423.

Doebley, J. and A. Stec. 1991. Genetic analysis of the morphological differences between maize and teosinte. Genetics 129:285-295.

Doebley, J.F., M.M. Goodman and C.W. Stuber. 1984. Isoenzymatic variation in Zea (Gramineae). System. Bot. 9:203-218.

Evans, A.M. 1976. Beans. Pages 168-172 in Evolution of Crop Plants (N.W. Simmonds, ed.). Longman, London, UK.

Freyre, R., R. Ríos, L. Guzmán, D. Debouck and P. Gepts. 1996. Ecogeographic distribution of Phaseolus spp. (Fabaceae) in Bolivia. Econ. Bot. 50:195-215.

Freytag, G.F. and D.G. Debouck. 1996. Phaseolus costaricensis, a new wild bean species (Phaseolinae, Leguminosae) from Costa Rica and Panama, Central America. Novon 6:157-163.

Gepts, P. 1988. Phaseolin as an evolutionary marker. Pages 215-241 in Genetic Resources of Phaseolus Beans (P. Gepts, ed.). Kluwer, Dordrecht, Netherlands.

Gepts, P. and F.A. Bliss. 1986. Phaseolin variability among wild and cultivated common beans (Phaseolus vulgaris) from Colombia. Econ. Bot. 40(4):469-478.

Goloubinoff, P., S. Pääbo and A. Wilson. 1993. Evolution of maize inferred from sequence diversity of an Adh2 gene segment from archaeological specimens. Proc. Nat. Acad. Sci. USA 90:1997-2001.

Hammer, K. 1984. Das Domestikationssyndrom. Kulturpflanze 32:11-34 [in German].

Harlan, J.R. 1975. Geographic patterns of variation in some cultivated plants. J. Heredity 66:184-191.

Harlan, J.R. and J.M.J. de Wet. 1971. Towards a rational classification of cultivated plants. Taxon 20:509-517.

Heiser, C.B. 1965. Cultivated plants and cultural diffusion in nuclear America. Am. Anthropol. 67:930-949.

Hillman, G.C. and M.S. Davies. 1990. Domestication rates in wild-type wheats and barley under primitive cultivation. Biol. J. Linn. Soc. 39:39-78.

Kami, J., B. Becerra Velásquez, D.G. Debouck and P. Gepts. 1995. Identification of presumed ancestral DNA sequences of phaseolin in Phaseolus vulgaris. Proc. Nat. Acad. Sci. USA 92:1101-1104.

Kaplan, L. and L.N. Kaplan. 1988. Phaseolus in archaeology. Pages 125-142 in Genetic Resources of Phaseolus Beans: their Maintenance, Domestication, Evolution and Utilization (Paul Gepts, ed.). Current Plant Science and Biotechnology in Agriculture, Vol. 5. Kluwer Academic Publishers, Dordrecht.

Koenig, R. and P. Gepts. 1989. Allozyme diversity in wild Phaseolus vulgaris: further evidence for two major centers of diversity. Theor. Appl. Genet. 78:809-817.

Koinange, E.M.K., S.P. Singh and P. Gepts. 1996. Genetic control of the domestication syndrome in common-bean. Crop Sci. 36:1037-1045.

Llaca, V. and P. Gepts. 1996. Pulsed field gel electrophoresis analysis of the phaseolin locus region in Phaseolus vulgaris. Genome 39:722-729.

McBryde, F.W. 1947. Cultural and historical geography of southwest Guatemala. Smithsonian Inst. Publ. 4:1-184.

Nodari, R.O., S.M. Tsai, R.L. Gilbertson and P. Gepts. 1993. Towards an integrated linkage map of common bean. II. Development of an RFLP-based linkage map. Theor. Appl. Genet. 85:513-520.

Rafalski, J. A. and S. V. Tingey. 1993. Genetic diagnostics in plant breeding: RAPDs, microsatellites and machines. Trends in Genetics 9:275-280.

Salgado, A., P. Gepts and D. Debouck. 1995. Evidence for two gene pools of the lima bean, Phaseolus lunatus L., in the Americas. Genet. Resour. Crop Evol. 42:15-28.

Singh, S.P., P. Gepts and D.G. Debouck. 1991a. Races of common bean (Phaseolus vulgaris L., Fabaceae). Econ. Bot. 45:379-396.

Singh, S.P., J.A. Gutirrez, A. Molina, C. Urrea and P. Gepts. 1991b. Genetic diversity in cultivated common bean: II. Marker-based analysis of morphological and agronomic traits. Crop Sci. 31:23-29.

Singh, S.P., R. Nodari and P. Gepts. 1991c. Genetic diversity in cultivated common bean. I. Allozymes. Crop Sci. 31:19-23.

Skroch, P., J. Tivang and J. Nienhuis. 1992. Analysis of genetic relationships using RAPD marker data. Pages 26-30 in Proc. Joint Plant Breeding Symposia Series, Nov. 1 1993. CSSA, Minneapolis, Minnesota, USA.

Smith, B. 1995. The Emergence of Agriculture. Scientific American Library, New York, USA.

Sokal, R.R., N.L. Oden and C. Wilson. 1991. Genetic evidence for the spread of agriculture in Europe by demic diffusion. Nature 351:143-145.

Talbot, D.R., M.J. Adang, J.L. Slightom and T.C. Hall. 1984. Size and organization of a multigene family encoding phaseolin, the major seed storage protein of Phaseolus vulgaris L. Mol. & General Genet. 198:42-49.

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.

On the Origin and In Statu Nascendi Domestication of Rye and Barley: A Review - V. Jaaska


The crucial difference between cultivated cereal crops and their wild relatives is in ear fragility at maturity. Easy shattering of spikes into spikelets upon maturity is essential for seed dispersal and survival in the wild, whereas forms with non-brittle ears survive only under cultivation. It is generally assumed that most Triticeae crops have been domesticated from their wild relatives by selection of non-shattering individuals which sporadically appear in wild populations as rare mutants (Zohary and Hopf 1993). However, there are no documented cases of the appearance of non-brittle mutants in wild populations, except observations of introduction from cultivars into weedy forms. The appearance of non-brittle mutants seems to be a rare event which may be induced under specific conditions.

Interspecific hybridization may activate transposition of mobile genetic elements which may be one potential source of increased mutation rates. A hypothesis was proposed that rye and wheat forms with varying ear fragility may have arisen as a result of interspecific hybridization processes between different wild species (Jaaska 1975). Taking into account that non-brittle mutants will persist in the wild only for a limited number of generations after their appearance, it follows that rye and barley should have been domesticated in statu nascendi from the ancient hybrid populations of their wild relatives.


Despite considerable disagreement among taxonomists about the delimitation of rye species and their intraspecific taxa (Kobyljanskij 1975; Gandilyan 1976; Hammer et al. 1987) the rye genus Secale L. may be treated as consisting of three valid biological species (Sencer and Hawkes 1980; Jaaska 1975,1983):

· The outcrossing annual Secale cereale L. s.l, including the European cultivated rye (subsp. cereale), the Transcaucasian cultivated and weedy rye with a tough rachis (subsp. segetale Zhuk.), and various weedy and wild forms of the annual self-incompatible rye with a brittle rachis such as subsp. ancestrale Zhuk. s.l, including weedy subsp, ancestrale Zhuk. s.str, weedy subsp. afghanicum (Vav.) Hammer, weedy subsp. dighoricum Vav., and wild subsp. vavilovii (Grossh.) Kobyl. s.str., non s.l. of Kobyljanskij (1975).

· The outcrossing perennial Secale strictum Presl., syn. S. montanum Guss., with several subspecies: subsp. strictum (= S. dalmaticum Vis.), subsp. anatolicum (Boiss.) Hammer, subsp. kuprijanovii (Grossh.) Hammer, subsp. ciliatoglume (Boiss.) Hammer, and subsp. africanum (Stapf) Hammer.

· Secale sylvestre Host. - a cleistogamous annual of sandy dunes and coasts.

A cleistogamous annual, originally collected by Kuckuck in Iran and attributed to S. vavilovii (Kuckuck and Kranz 1957), may deserve a specific rank as S. iranicum Kobyl., syn. S. vavilovii Grossh. sensu Khush, non sensu Grossheim (1924) or sensu lato of Gandilyan (1976), owing to its reproductive isolation as a result of its cleistogamous breeding system. Likewise, South African endemic perennial rye may be recognized as S. africanum Stapf because of its autogamous breeding system.

Grossheim (1924) described S. vavilovii as a short-statured (stems 20-35 cm) wild annual with short (4-8 cm) fragile spikes which he collected in Nakhitshevan from a dry habitat on volcanic ash soil. Populations of short wild annual rye were later found to occur in different dry regions of Armenia by Gandilyan (1976) who attributed the short stature to a phenotypic adaptation to a dry and nutrient-poor habitat. On those grounds he proposed to treat all wild and weedy annual ryes with a brittle rachis as S. vavilovii s.l. separately from the cultivated S. cereale s.str. To give these forms a specific rank is questionable because they are not reproductively isolated from non-brittle weedy S. cereale subsp. segetale. Their inclusion under S. cereale subsp. vavilovii s.l. as proposed by Kobyljanskij (1975) contradicts the botanical nomenclature rules because of the priority of subsp. ancestrale Zhuk. at the subspecies level (Zhukovsky 1928).

Vavilov (1917) expressed a view that cultivated rye evolved from a wild perennial rye through the appearance of annual weedy forms which in turn were domesticated by unconscious selection of non-brittle forms under cultivation in barley and wheat fields. Later he reported finding a weedy rye in Afghanistan with fragile ears which he attributed to S. cereale var. afghanicum and proposed that it might be an initial form from which semi-brittle and non-brittle forms of weedy rye were spontaneously selected under cultivation in wheat and barley fields in northern areas and higher altitudes (Vavilov 1926). He also further argued that S. cereale may have evolved from a perennial S. montanum with wild S. vavilovii Grossh. acting as a link between them. According to Vavilov, rye was domesticated in the Near East region as a secondary weedy crop (subsp. segetale) and only thereafter spread to European countries and was used as a distinct crop.

Alternatively, the present-day weedy ryes with brittle and semi-brittle rachis are products of introgressive hybridization of the cultivated non-brittle subsp. segetale with wild rye subsp. vavilovii s.str. Segregants from this hybridization acquired a weedy trait from subsp. segetale and became field weeds. Artificial hybrids between wild, weedy and cultivated forms of S. cereale are completely interfertile and show regular meiosis (Nürnberg-Krüger 1960a; Khush 1962, 1963), indicating that they comprise a single biological species.

It is generally accepted that all annual rye species have evolved from the perennial S. strictum (Stutz 1972; Hammer 1990). Secale cereale and S. strictum differ by two large translocations between three of the seven chromosomes, but intraspecific taxa have the same chromosome arrangement in both species (Riley 1955; Stutz 1957; Khush and Stebbins 1961). Polymorphism for reciprocal translocations has been described in populations of S. cereale (Candela et al. 1979). The evolution of S. cereale from S. strictum should thus involve fixation of two reciprocal translocations (Riley 1955; Khush and Stebbins 1961) which would be possible only by the involvement of an autogamous intermediate. Stutz (1972) suggested that cleistogamous annuals S. sylvestre and S. vavilovii sensu Khush (= S. iranicum) might represent intermediate stages in the evolution of S. cereale from S. strictum.

Secale sylvestre has the same chromosomal arrangement as S. strictum, but a low crossability with it, or differs in only a small chromosomal translocation, whereas S. iranicum was found to have the same chromosomal arrangement as S. cereale (Nürnberg-Krüger 1960b; Stutz 1972). According to these data, Stutz (1972) suggested that cleistogamous annual S. sylvestre has been derived from S. strictum and has then given rise to a similarly cleistogamous annual S. iranicum by fixation of two chromosomal translocations, while S. cereale was assumed to emerge from S. iranicum by acquisition of the outcrossing breeding system by introgression from S. strictum.

My studies of isoenzyme variability among the rye species established that S. cereale and S. strictum display partially homologous polymorphism of many isozymes with most allozymes shared between them. It may be seen from Table 1 that S. iranicum is fixed for one of the allozymes found in S. cereale and S. strictum, whereas S. sylvestre has unique allozymes of acid phosphatase ACP-B, anodal peroxidase PRX-C, and aromatic alcohol dehydrogenases AAD-A and AAD-E (Jaaska and Jaaska 1984). In the light of the isoenzyme data, it has been concluded that S. sylvestre is a lateral branch of evolutionary divergence form S. strictum and not an intermediate leading to S. iranicum and S. cereale.

Morever, S. strictum and S. sylvestre are well isolated in nature by growing in different ecological habitats and do not hybridize freely. Instead, I suggest that some unknown inbreeding form of subsp. vavilovii related to S. iranicum might be an intermediate from which both brittle- and tough-rachis forms of S. cereale evolved by introgression of self-incompatibility from S. strictum (Jaaska 1975). It was also assumed that non-brittle rye S. cereale subsp. segetale might be domesticated in statu nascendi, directly from some such hybrid populations currently extinct, while fragile subsp. vavilovii s.str. has persisted in the wild. Later, S. sylvestre was found to differ from the other rye species also in allozymes of cathodal peroxidases CPX-4 and CPX-5 (Vences et al. 1987a) and chloroplast DNA (Muray et al. 1989; Petersen and Doebly 1993), whereas S. cereale and S. strictum revealed homologous polymorphism and could not be definitely distinguished by allozymes and chloroplast DNA (Vences et al. 1987b; Petersen and Doebly 1993). These data provide further support to my conclusion that S. sylvestre has not been involved in the origin of S. cereale.

Zhukovsky (1971) pointed out that there have been no reports of spontaneous appearance of non-brittle forms for more than 100 years of cultivation of wild and weedy ryes in different botanical gardens and collections. The reason may be in the recessive nature of the non-brittle mutation combined with the outcrossing breeding system of rye (Sybenga and Prakken 1962). The rare non-brittle mutants will remain phenotypically unexpressed in outcrossing rye populations in heterozygous genotypes until two parents with recessive non-brittle alleles cross and segregate into homozygotes in a progeny. The probability for this depends on the frequency of non-brittle mutations which may be extremely low. Therefore, the involvement of an inbreeding intermediate would be an important premise for the appearance of recessive non-brittle mutants phenotypically expressed in homozygotes.

Zhukovsky (1971) also has proposed a hypothesis that weedy subsp. segetale might have arisen from interspecific hybridization between wild perennial and annual ryes with the appearance of non-brittle forms in the hybrid progeny. He assumed that interspecific hybridization might be the crucial mutagenic factor which has caused the appearance of semi- and non-brittle forms of rye. Stutz (1972) noted that “on the slopes of Mt. Ararat, S. vavilovii makes contact and hybridizes rather freely with S. montanum.” The seed obtained from crosses between S. cereale and S. strictum were found to germinate only when S. cereale was the mother parent (Riley 1955), indicating that the introgression is possible from the perennial rye. Sympatric populations of wild perennial and annual ryes in the Near East region should be examined for the fragility, chromosome rearrangements and molecular characters in order to have more information about the origin of wild, weedy and cultivated forms of S. cereale.

Table 1. Electrophoretic variants of aliphatic and aromatic alcohol dehydrogenase (ADH and AAD), aspartate aminotransferase (AAT), acid phosphatase (ACP), esterase (EST) and anodal peroxidase (APX) isoenzymes in rye species: major electrophoretic variants (allozymes) are numbered in the order of their decreasing mobilities and listed in the order of decreasing occurrence; rare allozymes are labelled by letter 'r'; unique allozymes are in italics.











S. strictum










S. cereale

subsp. cereale










subsp. segetale










subsp. ancestral










S. iranicum










S. sylvestre










Source: Adapted from Jaaska (1975, 1983) and Jaaska and Jaaska (1984).


The cultivated barley Hordeum vulgare L., incl. H. distichon L. and H. hexastichon L., and its closest wild relative H. spontaneum C. Koch are autogamous annuals which with the allogamous perennial H. bulbosum L. share basic genome I (Von Bothmer and Jacobsen 1985). They also share a number of allozymes (Jaaska and Jaaska 1986; Jaaska 1992). Their hybrids are completely interfertile and have normal chromosome pairing in meiosis (Takahashi 1955; Zohary 1960). Cultivated barley and its closest wild relative, H. spontaneum, are now commonly recognized as subspecies of H. vulgare s.l. (Bowden 1959; Von Bothmer and Jacobsen 1985). However, taking into account their autogamous breeding system, which itself provides a reproductive barrier between them, and the fact that the biological species concept is not strictly applicable to uniparentals, they could formally be accepted at a species level as separate phylogenetic lineages.

There is a general agreement among investigators that cultivated barley has originated from its closest wild relative H. spontaneum through the latter's domestication (Takahashi 1955; Harlan 1968; Zhukovsky 1971; Trofimovskaya 1972; Von Bothmer and Jacobsen 1985; Zohary and Hopf 1993). Different views, however, have been expressed with respect to the place, time and mechanisms of the origin of different forms of cultivated barley, particularly six-row forms (Vavilov 1926; Åberg 1940; Nevski 1941; Takahashi 1955; Zohary 1959, 1960; Staut 1961; Bakhteyev 1964, 1975; Trofimovskaya 1972).

Monophyletic origin of the cultivated barley as a two-row form by domestication of H. spontaneum and the secondary origin of the six-row barley from a cultivated two-row form is now generally accepted (Zohary and Hopf 1993), thanks to experimental evidence on the secondary origin of the Tibetan semi-brittle six-row barley by hybridization between the cultivated six-row barley and wild H. spontaneum (Zohary 1959, 1960; Tovia and Zohary 1962). Therefore, this aspect of barley evolution is not discussed here.

Instead, I would like to focus attention on the appearance of non-brittle mutants in pure, wild populations of H. spontaneum as a critical issue in the domestication of barley which needs to be documented. The semi-brittle forms in the present-day mixed populations of weedy and cultivated barley are evidently derived by introgression from the cultivated barley. Genetic studies of hybrid progeny between H. vulgare and H. spontaneum have shown that ear fragility is controlled by two closely linked genes (Bt and Bt2) with a recessive mutation at either loci giving rise to non-brittle ears (Takahashi 1955).

In the case of a simple in statu nascendi domestication of barley by selecting out non-brittle mutants, the cultivated barley has evidently been domesticated repeatedly from different genotypes of wild barley, i.e. it has a polytopic origin. This, however, does not explain a strikingly wider morphological variability among cultivated barley than in other cereals. Indeed, morphological variation within H. spontaneum is limited to only three (Nevski 1941) or four (Bakhteyev 1962) botanical varieties which differ in the apex shape and awn length of lateral spikelets. At the same time, the cultivated barley is remarkably more variable with over 100 botanical varieties described among primitive barley landraces (Mansfeld 1950). Although a remarkable morphometric variation was observed among a set of 77 accessions of wild barley under cultivation around Moscow (Russia) under unusual climatic conditions, it still remained relatively limited and no non-brittle cultivar-like forms were observed (Bakhteyev 1979). Characters of the wild parent dominate in hybrids with cultivated barley and segregation appears in later generations. No morphologically new types were found among the progeny of hybrid generations between various morphological types of H. spontaneum. Therefore, the question as to why cultivated barley is so morphologically variable still remains unanswered.

Comparative studies of isoenzyme variation among a set of morphologically different accessions of Hordeum vulgare s.str. and H. spontaneum showed that both-had the same allozymes of essentially monomorphic isoenzymes and displayed variation of 10 polymorphic esterase isozymes (Table 2). This shows that the isozyme loci which were monomorphic in wild barley have remained also in the cultivated barley, i.e. cultivation has not induced new variation at the isozyme loci in contrast to those controlling variable morphological characters. The same also has been demonstrated with respect to wild and cultivated wheats and rye (Jaaska 1987). The cultivated barley and its wild progenitor could not definitely be distinguished by any of the isoenzymes studied except by a frequency of allozymes of some isoesterases.

Hordeum bulbosum displays extensive intrapopulation polymorphism of several isozymes which were largely monomorphic in H. vulgare and H. spontaneum. The allozyme genepool of the two annuals was found to be a subset of most frequent allozymes of H. bulbosum (Table 2). This result is consistent with a view that H. spontaneum might have evolved from H. bulbosum by fixation of genes controlling self-compatibility and annual habit. Isoenzyme data also argue against a multiple, polytopic origin of H. spontaneum, as well as against later, repeated introgression from the bulbous barley. Moreover, crossing experiments between H. vulgare s.l. and H. bulbosum have shown that hybrids could be obtained only by the embryo rescue technique, had disturbed meiosis, were sterile and frequently haploid owing to spontaneous elimination of bulbosum chromosomes (Kasha and Sadasivaiah 1971; Lange 1971).

Table 2. Electrophoretic variants of aliphatic and aromatic alcohol dehydrogenase (ADH and AAD), aspartate aminotransferase (AAT) and esterase (EST) isoenzymes in cultivated barley (vulgare) and its wild relative species (spontaneum, bulbosum): major electrophoretic variants (allozymes) are numbered in the order of their decreasing mobilities and listed in the order of decreasing occurrence; rare variants are labelled by letter 'r'.






































































Source: adapted from Jaaska (1992) and Jaaska and Jaaska (1986).
Hordeum spontaneum and H. bulbosum are reproductively isolated by strong sterility barriers. Observations that crossability and chromosome elimination varies depending on the genotype of parental species suggested that introgression from the bulbous barley in some genotype combinations may be possible (Craig and Fedak 1985; Thomas and Pickering 1985).


Åberg, E. 1940. The taxonomy and phylogeny of Hordeum L. sect. Critesion Ands. with special reference to Tibetian barleys. Symb. Bot. Upsaliensis 2:1-156.

Bakhteyev, F.Kh. 1962. Anew link in the wild barley species. Bot. Zhurnal 47:844-847 [in Russian, with Latin diagnosis].

Bakhteyev, F.Kh. 1964. Origin and phylogeny of barley. Pages 1-18 in Barley Genetics I, Proc. 1st Int. Barley Genetics Symposium, Wageningen. Pudoc, Wageningen, Netherlands.

Bakhteyev, F.Kh. 1975. Studies on barley, Folia Quaternaria 46:73-81.

Bakhteyev, F.Kh. 1979. Near-Eastern collection of Hordeum spontaneum C. Koch emend. Bacht. under conditions of Moscow environs. II. Collection from Iran, Iraq and Turkey. Z. Pflanzenzüchtg. 83:211-221.

Bowden, W.M. 1959. The taxonomy and nomenclature of the wheats, barleys, and ryes and their wild relatives. Can. J. Bot. 37:657-684.

Candela, M., A.M. Figueiras and J.R. Lacadena. 1979. Maintenance of interchange heterozygosity in cultivated rye, Secale cereale L. Heredity 42:283-289.

Craig, I.L. and G. Fedak. 1985. Variation in crossability of diverse genotypes of Hordeum bulbosum with H. vulgare (4x) cv. Betzes. Cereal Res. Commun. 13:393-397.

Gandilyan, P. A. 1976. On the systematics of the genus Secale L. and its diversity in the Armenian SSR. I. Species delimitation. Biol. Zhurn. Armenii 29 (11):27-35 [in Russian].

Grossheim, A.A. 1924. A new race of wild rye Secale vavilovi m. from Transcaucasia. Bull. Appl. Bot. & Select. 13:461-482 [in Russian with English summary].

Hammer, K. 1990. Breeding system and phylogenetic relationships in Secale L. Biol. Zbl. 109:45-50.

Hammer, K., E. Skolimowska and H. Knüpfer. 1987. Vorarbeiten zur monographishen Darstellung von Wildpflanzensortimenten: Secale L. Kulturpflanze 35:135-177 [in German].

Harlan, J.R. 1968. On the origin of barley. Pages 9-31 in Agriculture Handbook No. 388. USDA, Beltsville, Maryland, USA.

Jaaska, V. 1975. Evolutionary variation of enzymes and phylogenetic relationships in the genus Secale L. Proc. Acad. Sci. Estonian SSR, Biology 24:179-198 [in Russian].

Jaaska, V. 1983. Secale and Triticale. Pages 79-101 in Isozymes in Plant Genetics and Breeding (S.D. Tanksley and T.J. Orton, eds.). Elsevier, Amsterdam, Netherlands.

Jaaska, V. 1987. Isoenzyme variability and speciation in the Triticeae grasses. Proc. Acad. Sci. Estonian SSR, Biology 36:281-295 [in Russian].

Jaaska, V. 1992. Isoenzyme variation in barley (Hordeum L.). 2. Aspartate aminotransferase and 6-phosphogluconate dehydrogenase. Hereditas 116:29-35.

Jaaska, V. and V. Jaaska. 1984. Isoenzymes of aromatic alcohol dehydrogenase in rye and triticale. Biochem. Physiol. Pflanzen 179:21-30.

Jaaska, V. and V. Jaaska. 1986. Isoenzyme variation in barley genus Hordeum L. I. Alcohol dehydrogenase and superoxide dismutase. Biochem. Physiol. Pflanzen 181:301-320.

Kasha, K.J. and R.S. Sadasivaiah. 1971. Genome relationships between Hordeum vulgare L. And H. bulbosum L. Chromosoma 35:264-287.

Khush, G.S. 1962. Cytogenetic and evolutionary studies in Secale. II. Interrelationships of wild species. Evolution 16:484-496.

Khush, G.S. 1963. Cytogenetic and evolutionary studies in Secale. III Cytogenetics of weedy ryes and origin of cultivated rye. Econ. Bot. 17:60-71.

Khush, G.S. and G.L. Stebbins. 1961. Cytogenetic and evolutionary studies in Secale. I. Some new data on the ancestry of S. cereale. Am. J. Bot. 48:723-730.

Kobyljanskij, V.D. 1975. On the systematics and phylogeny of the genus Secale L. Bull. Vavilov Inst. Plant Industry 48:64-71 [in Russian].

Kuckuck, H. and A.R. Kranz. 1957. A genetic analysis of rye populations from Iran. Wheat Inf. Serv. 6:20-21.

Lange, W. 1971. Crosses between Hordeum vulgare L. and H. bulbosum L. I. Production, morphology and meiosis of hybrids, haploids and dihaploids. Euphytica 20:14-19.

Mansfeld, R. 1950. Das Morphologische System der Saatgerste, Hordeum vulgare L. s. 1. Der Züchter 20:8-24 [in German].

Muray, K., X. Naiyu and K. Tsunewaki. 1989. Studies on the origin of crop species by restriction endonuclease analysis of organellar DNA. III. Chloroplast DNA variation and interspecific relationships in the genus Secale. Jpn. J. Genet. 64:36-47.

Nevski, S.A. 1941. Beiträge zur Kenntnis der wildwachsenden Gersten in Zusammenhang mit der Frage über den Ursprung von Hordeum vulgare L. und Hordeum distichon L. Pages 248-255 in Acta Inst. Bot. Nom. V. L. Komarovii Acad. Sci. USSR, Ser. 1, Fasc. 5 [in Russian with German summary].

Nürnberg-Krüger, M. 1960a. Cytogenetische Untersuchungen der Gattung Secale L. Z. Pflanzenzücht. 44:63-72.

Nürnberg-Krüger, M. 1960b. Cytogenetische Untersuchungen an Secale silvestre Host. I. Der Bastard mit Secale cereale L. Der Züchter 30:147-150 [in German].

Petersen, G. and J.F. Doebley. 1993. Chloroplast DNA variation in the genus Secale L. Plant Syst. Evol 187:115-125.

Riley, R. 1955. The cytogenetics of differences between some Secale species. J. Agric. Sci. 46:377-383.

Sencer, A. and J.G. Hawkes. 1980. On the origin of cultivated rye. Biol J. Linn. Soc. 13:377-383.

Staut, G. 1961. The origin of cultivated barleys: a discussion. Econ. Bot. 15:205-202.

Stutz, H.S. 1957. A cytogenetic analysis of the hybrid Secale cereale L. x S. montanum Guss. And its progeny. Genetics 42:199-221.

Stutz, H.S. 1972. On the origin of cultivated rye. Am. J. Bot. 59:59-70.

Sybenga, J. and R. Prakken. 1962. Gene analysis in rye. Genetica 33:59-70.

Takahashi, R. 1955. The origin of cultivated barley. Pages 227-266 in Advances in Genetics (M. Demerec, ed.). Acad. Press, New York, USA.

Thomas, H.M. and R. A. Pickering. 1985. The influence of parental genotype on the chromosome behaviour of Hordeum vulgare x H. bulbosum diploid hybrids. Theor. Appl. Genet. 71:437-442.

Tovia, T. and D. Zohary. 1962. Natural hybridization between Hordeum vulgare and Hordeum spontaneum: Corroborative evidence from progeny tests and artificial hybrids. Bull. Res. Council Isr. 11:43-45.

Trofimovskaya, A.Y. 1972. Barley (Evolution, Classification, Selection). Pages 7-78 in Problems of Evolution and Classification of Barley Species. Kolos, Leningrad, USSR [in Russian].

Vavilov, N.I. 1917. On the origin of cultivated rye. Bull. Appl. Bot. 10: 561-590 [in Russian with English summary].

Vavilov, N.I. 1926. Centres of origin of cultivated plants. Tr. po Prikl. Bot. Genet. Sel. [Bull. Appl. Bot. & Genet. Sel.] 16(2):139-248 [in Russian].

Vences, F.J., F. Vaquedro and P. de la Vega. 1987a. Phylogenetic relationships in Secale (Poaceae): an isozymatic study. Plant Syst. Evol. 157:33-47.

Vences, F.J., F. Vaquedro, P. Garcia and P. de la Vega. 1987b. Further studies on phylogenetic relationships in Secale: On the origin of its species. Plant Breed. 98:281-291.

Von Bothmer, R. and N. Jacobsen. 1985. Origin, taxonomy and related species. Pages 19-56 in Barley (D. Rasmusson, ed.). Agronomy Monograph No. 26, ASA, Madison, Wisconsin, USA.

Zhukovsky, P.M. 1928. A new wild growing form of rye in Anatolia. Bull. Appl. Bot. 19:49-56 [in Russian with English summary].

Zhukovsky, P.M. 1971. Cultivated Plants and their Relatives. Leningrad, Kolos [in Russian].

Zohary, D. 1959. Is Hordeum agriogrithon the ancestor of six-rowed cultivated barley? Evolution 13:279-280.

Zohary, D. 1960. Studies on the origin of cultivated barley. Bull. Res. Counc. Isr. 9D:21-42.

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.

Plant-Gathering Versus Plant Domestication: An Ethnobotanical Focus on Leafy Plants - F. Ertug


In this paper some information about the gathering of leafy plants in Turkey is presented. An ethnoarchaeological study on a contemporary rural village located close to a prehistoric site in Central Anatolia was carried out. It was observed that the village women gathered a large number of wild leafy plants. The number of species of plants gathered was much higher than expected considering that a sufficiently diverse variety of cultivated plants was available in their own gardens as well as in local markets.

Gathering of wild plants by contemporary villagers in Central Anatolia raised several questions about the domestication of leafy plants and our knowledge of human diets in the past. Through ethnobotanical study, answers to simple questions were attempted, such as: why do people continue to gather leafy plants thousands of years after domestication? What was the role of these plants in past diets?

Material and methods

The study area, the Melendiz Plain, is located on the Central Anatolian Plateau, southeast of Ankara, in the province of Aksaray. The modern town of Aksaray is in the southwest corner of Cappadocia, an ancient name once given to most of Central Anatolia. The Melendiz Plain is at an altitude of about 1100-2000 m asl, and is bordered by a range of volcanic mountains including Melendiz Dag (2935 m) and Hasan Dag (3268 m). The fertile plain is watered by the Melendiz Su and is suitable for cereal crops. The average annual precipitation is 339 mm. Most of the Aksaray area belongs to the Irano-Turanian floristic region, and is predominantly treeless steppe vegetation (Zohary 1973; Kürschner et al. 1995). The research area is in B5 square of the Flora of Turkey grid-system Map 1 (Davis 1965-1985). A considerable amount of information on plant-gathering and a large collection of plants were obtained during the fieldwork in the Melendiz Plain. The use of the plants for food, fuel and fodder, as well as in textiles, basket-making, drugs, glues, brooms, dyes, amulets and medicines was documented. Of the 600 plant specimens that were collected, 313 belong to 205 genera and 70 families. Most of the identifications were made by Professors Tuna Ekim, Mecit Vural and their team in the Department of Biology, Gazi University, Ankara, and the samples were deposited in the Gazi University Herbarium.

Another indigenous vegetation belt once surrounded this steppe vegetation (Davis et al. 1988). Zohary (1973) uses the term Xero-Euxinian vegetation belt for an area with lower average precipitation than the mediterranean zone, and with increasing continentality toward the Central Anatolian plateau. This belt of 'steppe-forests' starts east and south of the research area, where the altitude reaches 1400 m. The dominant tree in the deciduous forests is Quercus cerris; a mixed forest of Q. pubescens, Q. infectoria, Q. ithaburensis, in addition a few Q. trojana, occurs in the area, and Q. vulcanica is found at 2000 m, on the slopes of the Melendiz Mountains. Neither Pinus nor Juniperus is represented in the present steppe-forest, but the species that originally grew together with Pinus nigra, such as Q. vulcanica and Acer hyrcanum subsp. tauricolum, are still present. Solitary wild fruit trees of Pyrus elaeagrifolia and Crataegus species indicate the lower limits of the forest-steppe, at about 1200 m.

There are also some small patches among the volcanic areas, where Crataegus, Celtis, Prunus, Pyrus and Amygdalus trees have survived, and marshy areas nearby that contain many species used for mat- and basket-making, such as Typha and Phragmites. The rivers and small streams that provide moisture for the grazing lands are bordered by some edible species such as Berula erecta, Rorippa nasturtium-aquaticum and Veronica anagallis-aquatica. Triticum boeoticum (einkorn) and some wild species of Hordeum are also present in patches. Today, most of the area is devoted to cereal production with gardens near available water, and partly to orchards and vineyards, so the area is predominantly a man-made (anthropogenic) landscape.

The village of Kizilkaya, central to this study, is about 25 km southeast of the city of Aksaray, and 1 km north of an archaeological site Asikli Höyük. This Pre-Pottery Neolithic site, which dates from 8000 to 7450 years BP (calibrated), has been under excavation for the last eight years by a team from the University of Istanbul (Esin et al. 1991; Esin 1996). The excavations indicated that the economy of Asikli was based on crop husbandry (van Zeist and de Roller 1995), on hunting game animals and gathering wild plants. The ethnoarchaeological fieldwork was conducted from January 1994 to the end of June 1995 to provide comparative data to aid in the interpretation of the archaeological remains from Asikli. Plant-gathering was an important aspect of the economy of an early Neolithic settlement, but has not been previously studied in Anatolia.

Specimens of all plants gathered by the village folk were collected, pressed and put in a card file according to the local names of the plants. Questionnaires, one for the men concerning land-ownership and agriculture, and one for the women about plant-gathering and gardening, were completed. Thirty households were chosen randomly from three income groups, so that different attitudes of rich and poor toward gathering and/or agricultural decisions could be detected. These questionnaires helped to quantify the number of wild plants that each woman knew and gathered. However, the most satisfactory way to collect information about the plants was to accompany and question the women while they were gathering, as well as to attend all agricultural activities, such as planting, weeding, harvesting, preparing and cooking.


The modern village of Kizilkaya has a population of about 1300 occupying 300 houses, and its economy is based on agriculture, field-cropping and gardening, as well as the husbandry of sheep and cattle. The total number of known cultivars is 70, and of these 20 are trees, 10 are used primarily as animal fodder, 4 species are no longer cultivated, and some others are planted only rarely. The basic cereal crop is bread wheat, and basic legumes such as beans, lentil and chickpea are also grown. About 20 vegetables are more or less regularly planted in spring and consumed during summer, and some are dried, made into paste or pickled, and stored for winter.

Onion, potato, garlic, green beans, squash, pepper and tomato are the most important. Beet, spinach, cabbage, leek, purslane, green onion, cress, lettuce, parsley and chicory are regularly planted for the consumption of their green leaves. Whereas the number of green leafy vegetables does not exceed 10 during the summer, as many as 40 different kinds of wild greens are gathered during the winter and spring. Gardening is a recent development within the last 20 to 30 years, and as the climate is continental, and the frosts start as early as October, it is strictly limited to the summer.

Until recently, fresh products of the village gardens, such as tomato and leek, could be eaten up to October, but from November to June, a period of 7 to 8 months, no fresh vegetables would be available. This was true until about 20 to 30 years ago, but since transportation became available, fresh vegetables and fruits can be found in the local village markets all year round. Despite this easy accessibility, the local tradition of gathering wild plants for food persists in Central Anatolia, and probably goes back at least 500 years. Historical records for many villages in this area go back to the early 16th century (Ertug-Yaras 1997).

Over 100 wild species of plants in and around the Melendiz Plain are considered by the local people as edible (Ertug-Yaras 1996). These edible species belong to 36 botanical families. Among these, 42 wild greens make up the most commonly and regularly consumed group (see Appendix), other than wild fruits, roots and stems. Among the 42 greens in 18 families, species belonging to the Asteraceae and Brassicaceae families are the most numerous. Twelve of these greens (see (*) in the Appendix) were not previously recorded in the literature as edible.

Wild greens are an important component of the local diet, and are regularly gathered from October to June. During the winter, unless there is deep snow, it is possible to find green leaves of 13 to 16 different varieties of edible plants. Three different kinds of edible aquatic plants can be found, even when the ground is covered with snow, but these are frozen if the temperature drops below zero, or are destroyed when there is spring flooding. The maximum number of plants with edible leaves is gathered in April and May, 33 and 35 respectively, while in June there are only 9.

The gathering of leafy plants is exclusively carried out by women. Groups of women, mostly with their children, go to gather greens during winter and spring, when fresh greens are most needed in the diet. Most of these greens are eaten raw with salt between folds of the local flat bread (yufka), but some greens require cooking. These are chopped and cooked together with onions and cracked wheat. These dishes are called 'cacik', and are usually eaten with yogurt.

According to the questionnaires, women in middle-income and poorer groups go more often, and gather more greens than do women in more affluent households. Low-income women also gather more species than do the richer households, but they are less selective. The implication is that gathering wild greens is related to nutrition and taste more than to economic need.

Nutritional analysis for twelve of the most commonly consumed wild plants indicates that these plants are a very good source of raw protein and minerals (Ertug-Yaras 1997). Most of them have protein and mineral values as high as cultivated green vegetables, and probably make a significant nutritional addition to local diets.

Some women in the village of Kizilkaya told me that they made several attempts to cultivate the wild plants in their gardens, but have found that the desirable sharp or bitter taste of the wild plants changes when they are planted in gardens. This is probably an important reason why so few of the many leafy plants were selected for cultivation. Chicory, Crepis foetida and wild spinach were the most common experiments. These showed phenotypic changes, with larger leaves, as well as different tastes. Some villagers attribute these changes to water availability, some to the soil, and some botanists explain it as a reaction of the plants to reduced competition for water and nutrients. Wild spinach (Spinacia tetranda) and wild beet (Beta macrorrhiza) grow in fallow fields and waste places and find their place in the diet while their domesticated relatives grown from commercial seeds bought in the market are commonly planted in gardens.


Some of the reasons village women of all economic groups still gather leafy plants became apparent during the fieldwork: (1) the villagers feel the need for green leafy plants in their diet all year round, and in winter and spring when these are not available from their own gardens the women gather the wild plants; (2) these plants grow abundantly in most areas during winter and spring, are easy to gather, and most require no processing except washing; (3) their taste seems to play an important role in their dietary use, perhaps because they add variety to the otherwise monotonous diet; (4) gathering is also a social activity, as the women take pride in providing for their families, and also share the plants they gather with their friends and neighbors. The women consider plant-gathering a pleasant way to get out of the house and to socialize with other women; (5) gathered greens, as well as bulbs, roots, mushrooms and fruits, provide a buffering/risk-management technique against potential hunger.

The drive to gather wild plants is closely related also to insecurity during the times of scarcity, such as droughts, epidemics, wars or rebellions. Recent reports about strife in North Korea, Bosnia and various parts of Africa have indicated that wild food may make the difference between survival and death when crops fail repeatedly, or when harvests are stolen or destroyed. While this critical economic aspect of plant-gathering is frequently emphasized, we tend to forget or consider unimportant the role of wild plants in the daily diet.

Archaeobotanists considering past diets usually emphasize recovery and study of cereals and pulses, especially their initial domestication and early use. Although the primary role of cereals in the subsistence of both modern and past societies is extremely important, the role of vegetables, mushrooms and soft fruits tends to be ignored, and, of course leaves, stems, flowers and bulbs are very difficult to find among archaeobotanical remains. Only very recently have techniques been developed for the recovery of such plant parts. In general, as soon as the domestication of cereal crops is detected on a site, it should not be assumed that people limited themselves to these cultivated cereals, and stopped using wild plants as food.

Ethnobotanical studies in various countries indicate that while farmers restricted themselves to a limited number of crops, their wild plant kit, used as food and for other purposes, is much larger. Farmers use this knowledge and pass it on to future generations, and they do not limit themselves to the surrounding environment, but trade with other villages for plants not available to them.


Wild plant-gathering in Central Anatolia is not unique to this region, nor to Anatolia itself. It is true that the development of agriculture and the grazing of animals have decreased the availability of wild plant sources, but unless there is very dense population, many resources are still available. A good indication of this is that even after thousands of years of agriculture and animal grazing in the Melendiz area, more than 100 edible species, mostly perennials, are still available and consumed. Although not well studied, wild plant-gathering is a common tradition among farming societies in the Near East. The role played by women in early agriculture and plant-gathering and domestication should not be underestimated.

It is hoped that the role of wild leafy plants in the diet of agricultural societies and local traditions of wild plant use can be better understood through further ethnological studies, and that they will illuminate archaeological research to provide better understanding of past diets. Perhaps such studies can help improve diets to suit future environmental changes.


This project was supported by a National Science Foundation Grant for Improving Doctoral Dissertation Research (Grant No. SBR-9319164); a Wenner-Gren Foundation for Anthropological Research, Developing Countries Training Fellowship (Grant No. 4201-11); and the Department of Anthropology, Washington University, St. Louis, USA.


Davis, P.H. (ed.). 1965-1985. Flora of Turkey and the East Aegean Islands. Volumes 1-9. Edinburgh University Press, Edinburgh, UK.

Davis, P.H., R.R. Mill and K. Tan (eds.). 1988. Flora of Turkey. Vol. 10: Supplement. Edinburgh University Press, Edinburgh, UK.

Ertug-Yaras, F. 1996. Contemporary Plant Gathering in Central Anatolia: An Ethnoarchaeological and Ethnobotanical Study. Pages 945-962 in Plant Life in Southwest and Central Asia (M.A. Öztürk, Ö. Seçmen and G. Görk, eds.). Proc. 4th Plant Life of Southwest Asia Symposium. Ege University Press, Izmir, Turkey.

Ertug-Yaras, F. 1997. An Ethnoarchaeological Study of Subsistence and Plant Gathering in Central Anatolia. PhD Thesis. Washington University, St. Louis, USA.

Esin, U. 1996. Asikli, Ten Thousand Years Ago: A Habitation Model From Central Anatolia. Pages 31-42 in Housing and Settlement in Anatolia: A Historical Perspective. Türkiye Ekonomik ve Toplumsal Tarih Vakfi, Istanbul, Turkey.

Esin, U., E. Biçakçi, M. Özbasaran, N. Balkan-Atli, D. Berker, I. Yagmur and A.K. Atli. 1991. Salvage Excavations at the Pre-Pottery Site of Asikli Höyük in Central Anatolia. Anatolica 17:123-174.

Kürschner, H., T. Raus and J. Venter. 1995. Plflanzen der Türkei: Agais-Taurus-Inneranatolien. Quelle & Meyer Verlag, Wiesbaden, Germany.

van Zeist, W. and G.J. de Roller. 1995. Plant Remains from Asikli Höyük, a pre-pottery Neolithic site in Central Anatolia. Veget. History & Archaeobot. 4:179-185.

Zohary, M. 1973. Geobotanical Foundations of the Middle East. Volumes I and II. Gustav Fischer Verlag, Amsterdam, Netherlands.

Appendix: Edible greens of the Melendiz area, Aksaray, Turkey (B5).



Local names

Gathering frequency


Amaranthus retroflexus L.

Pancar otu



Berula erecta (Huds.) Coville (*)




Centaurea depressa Bieb. (*)




Chondrilla juncea L. var. Juncea




Cichorium intybus L.

Çitlik/Yabani hindiba



Crepis foetida L. subsp. rhoadifolia (Bieb.) Celak. (*)

Kohum otu



Lactuca serriola L.

Marul otu/Yazi marulu



Scariola viminea (L.) F.W. Schmidt

Kedi çitligi



Scorzonera cana (C.A. Meyer) Hoffm. var. radicosa (Boiss.)

Tekercik/Dede sakali



Sonchus asper (L.) Hill subsp. glaucescens (Jordan) Ball

Su kangali



Taraxacum microcephaloides van Soest (*)

Ebem çitligi/Karahindiba



Taraxacum serotinum (Waldst. et Kit.) Poiret (*)

Çukur çitligi



Tragopogon buphthalmoides (DC.) Boiss.




Anchusa undulata L. subsp. hybrida (Ten.) Coutinho (*)




Barbarea plantaginea DC. (*)

Götlez götü



Camelina hispida Boiss. (*)




Capsella bursa-pastoris (L.) Medik.

Kuskus ekmegi



Rorippa nasturtium-aquaticum (L.) Hayek (Syn. Nasturtium officinale R.Br.)

Aci tere



Sinapis arvensis L.

Hardal otu



Sisymbrium altissimum L. (*)

Elgelen hardali



Campanula cymbalaria Sm. (*)

Yer otu



Silene vulgaris (Moench.) Garcke var. vulgaris

Tavsan ekmegi



Stellaria media (L.) Vill. subsp. pallida (Dumort.) Aschers. et. Graebn.

Haval otu



Beta macrorrhiza Stev.

Kizil pancar



Chenopodium album L. subsp. album

Köpürgen otu



Chenopodium album L. subsp. Aellen iranicum

Sirken otu



Spinacia tetrandra Stev.

Yazi ispanagi



Convolvulus lineatus L (*).

Tavsan kulagi



Sempervivum armenum Boiss. et Huet var. armenum

Musluk otu



Ononis spinosa L. subsp. leiosperma (Boiss.) Sirj.

Sigek dikeni



Erodium cicutarium (L.) L'Hérit. subsp. cicutarium




Malva neglecta Wallr.




Malva sylvestris L.




Papaver rhoeas L.

Gül otu/Gelincik



Polygonum bellardii All. (*)

At mercimelegi



Polygonum cognatum Meissn.




Rumex acetosella L.

Eskileme/Kuzum oglagi



Rumex crispus L.

Kazan kulpu



Rumex scutatus L.




Portulaca oleracea L. subsp. oleracea

Temizlik otu/Semizotu



Veronica anagallis-aquatica L




Urtica dioica L.



(*) mark at the end of species' names indicates that these are not reported as edible in the literature.

The 'Gathering frequency' column offers an impressionistic and tentative evaluation: x-rarely gathered; xx - commonly; xxx - most commonly gathered species.

Origins and Domestication of Mediterranean Olive Determined through RAPD Marker Analyses - G. Besnard, A. Moukhli, H. Sommerlatte, H. Hosseinpour, M. Tersac, P. Villemur, F. Dosba and A. Bervillé


Olive (Olea europaea L.) is one of the most important cultivated trees in the Mediterranean basin and the spread of its cultivation was associated with that of grapevine, almond and fig. These species form the oldest group of domesticated trees (Zohary and Hopf 1993).

Olea europaea is a complex of four subspecies (Table 1, Fig. 1) (Green and Wickens 1989). Olea europaea subsp. europaea is separated into two taxa: var. sylvestris (oleasters: feral and wild forms) and var. sativa (cultivated forms). Cultivated trees could develop from oleasters (Zohary and Spiegel-Roy 1975).

The area of olive domestication appears to have been the Near East. Artifacts associated with the oil-extraction process as well as those associated with clonal propagation are the main evidence for domestication which occurred during the early Bronze Age (Liphschitz et al. 1991). Nevertheless, the exploitation of wild forms before domestication is suggested by discoveries of olive tree remains in archaeological sites since Palaeolithic times about 43,000 years BP (Zohary and Spiegel-Roy 1975). This was not specific to the Near East. On the basis of archaeobotanical evidence, Terral and Arnold-Simmard (1996) and Galili et al. (1989) have suggested a pre-domestication period in eastern Spain and Palestine. During this period, harvesting and caring of wild trees could have occurred without vegetative multiplication, thus leading to higher production. Evidence of olive remains from the Holocene, outside the natural distribution of the species, has been observed at a Syrian archaeological site (Willcox 1996) but the low frequency of olive fragments is insufficient to conclude that domestication occurred there.

The number of generations required for the improvement and selection of cultivated forms is likely to be very low in comparison with annual plants. A close genetic similarity is therefore expected between wild forms and their direct cultivated descendants (Liphschitz et al. 1991). Elant (1976) has suggested that cultivated forms were propagated in the entire Mediterranean basin and crosses between them and with local wild forms have led to the creation of new cultivars. At least, the crosses between related species of the Olea europaea complex could have occurred in different areas (northern Africa, the Middle East and the Canary Islands). In southern Morocco, O. maroccana has been considered as an intermediate form between O. europaea and O. laperrinei by Sauvage and Vindt (1952) or as a variant of O. laperrinei (Maire 1933; Green and Wickens 1989; Zohary 1994). The distribution area of this form could overlap with the cultivated form (Zohary 1994). Some examples of spontaneous interspecific crosses are known for several cultivated trees, notably apple, pear, date palm and grapevine (Zohary and Hopf 1993). Consequently, the origin of Mediterranean olive is not clear. Two places have been suggested: North Africa (Maley 1980) and the Near East (Chevalier 1948; Turrill 1951).

Table 1. The Olea europaea L. complex: subspecies and their main taxa.


Geographical origin

subsp. europaea Mansfield

- O. europaea L.

var. sylvestris D.C. (oleasters)

Mediterranean basin

var. sativa Lehr. (cultivars)

Mediterranean basin

subsp. laperrinei Ciferri


- O. laperrinei Batt. & Trab.

Saharan Mountains

- O. maroccana Gren. & Burd.

Atlas Mountains (Morocco)

subsp. cerasiformis Kunk. & Sund.

Canary and Madeira Islands

subsp. cuspidata Ciferri

- O. cuspidata Wall.


- O. chrysophylla Lam.

Arabia, Abyssinia

- O. africana Mill.

Eastern and southern Africa

A molecular analysis of different wild and cultivated accessions could clarify the origins of the domestication and of the selection of olive trees. The RAPD markers have been used to identify cultivars (Koller et al. 1993), to reveal the extent of diversity (Isabel et al. 1995) and to differentiate species for detection of interspecific hybridization (Jeandroz et al. 1996; Vigouroux et al. 1997) or in a phylogenetic approach (Wang et al. 1994). In this preliminary study we chose cultivars and wild forms with RAPDs to estimate the similarities between them.

Material and methods

Plant materials

Sixty-four cultivars from different countries around the Mediterranean basin were analyzed (Table 2). These cultivars were selected from a germplasm collection maintained at Montpellier (France) and provided by Dr Rallo (Cordoba) and Dr Baldoni (Perugia) from the collections of Cordoba (Spain) and Perugia (Italy).

Oleasters (wild olive trees) were studied from three sources: from Corsica (11 individuals: 'Corsica 1' to '11'), from Morocco (9 individuals: 'Morocco 1' to '9'), and from southern France (5 individuals: 'Hérault 1' to '5'). We have no criteria to distinguish wild from feral forms. Four individuals of O. maroccana from southern Morocco and four individuals of O. europaea subsp. cuspidata from four different areas [Kenya (O. africana), Iran, India, China] were also studied.

Molecular analysis

DNA preparation was carried out according to a modified method described by Saghai-Maroof et al. (1984). Five grams of leaves were ground and 20 ml of 2X CTAB buffer (100 mM Tris-HCl pH 8,1.4M NaCl, 20 mM EDTA, CTAB 2%) were added with 0.2 ml of bmercaptoethanol. This mixture was incubated at 65°C for 1 hour. Ten milliliters of chloroform/isoamylic alcohol (IAA) (24/1) were added and mixed at 100 rpm for 20 min. This mixture was centrifuged at 4500 rpm for 20 min. The supernatant was recovered and 10 ml of chloroform/IAA(24/l) were added and mixed at 100 rpm for 20 min. This mixture was centrifuged at 4500 rpm for 20 min. The supernatant was recovered and mixed with 20 ml of isopropanol and stored at -20°C overnight. The DNA pelot was recovered, washed with ethanol solution (ethanol 76%, 10 mM ammonium acetate), dried, and removed in 0.5 ml of 1X TE (10 mM Tris-HCl pH 8,1 mM EDTA). Then 500 µg of RNAse A (Boehringer) were added. The incubation was performed at 37°C for 1 hour; 530 µl of phenol/chloroform/IAA (25/24/1) were added and mixed. After centrifugation at 12,000 rpm for 10 min, the supernatant was recovered and mixed with 530 µl of chloroform/IAA (24/1). Another centrifugation was performed at 12,000 rpm for 10 min, and then the supernatant was mixed with 1 ml of ethanol and 188 µl of 7.5M ammonium acetate solution. This mixture was kept overnight at -20°C. A final centrifugation at 14,000 rpm for 10 min was performed to recover DNA pelot which was dried, then dissolved in 0.5 ml of 1XTE.

Fig. 1. Herbarium samples of Olea europaea taxa. A = O. africana (Reunion Island); B = O. chrysophylla (Yemen); C = O. laperrinei (Algeria); D = O. europaea var. oleaster (feral?-France); and E = O. europaea var. oleaster (Corsica).

Table 2. List of Olea cultivars studied according to their geographic origins.

Geographical origin

No. of cvs.




Barouni, Bid el Hamam, Chemlal de Kabylie, Chemlali de Sfax, Chetoui, Meski, Picholine Marocaine, Sigoise, Zarazzi

Iberian Peninsula


Arbequina, Cornicabra, Empeltre, Galega, Lechin de Sevilla, Manzanilla, Picual, Villalonga



Amellau, Belgentier, Bouteillan, Cayet Bleu, Cayet Rouge, Cayon, Colombale, Corniale, Filitosa, Sabina, Lucques, Olivière, Picholine, Picholine de Rochefort, Pigale, Rougette de Pignan, Tanche, Verdale, Verdelet, Vermillau



Ascolana Tenera, Cassanese, Cellina, Dolce Aggogia, Frantoio, Giarraffa, Leccino, Leucocarpa, Nocelara del Belice, Pendolino, San Felice






Amygdalolia, Carolia, Gaïdourolia, Kalamata, Koroneiki, Valanolia



Ayvalik, Domat, Sofralik, Uslu

Near East


Kaissy, Merhavia, Souri, Zaity






'Filitosa' corresponds to a cultivar which was identified by the authors at the Filitosa archaeological site in Corsica (France) without known denomination.
The procedure for obtaining RAPD markers has been described by Quillet et al. (1995). Thirty decamer primers (Bioprobe) were tested on eight genotypes: five cultivars ('Olivière', 'Lucques', 'Giarraffa', 'Arbequina', 'Domat'), two wild forms corresponding to the cuspidata subspecies (from Iran and China) and Ligustrum vulgaris L. as an outgroup. The choice of primers was based on the level of polymorphism with the clearness of banding profiles.

Data exploitation

The bands were noted and coded according to the primer and the size of the fragments as judged with the 1 Kb ladder from BRL as reference. Several individuals from 15 cultivars were analyzed; fragments, which were not reproducible, were not taken into account here. The probability to obtain the same pattern for two different cultivars was calculated according to the formula from Nybom and Hall (1991): (mean of Nei and Li similarities)/Mean of marker number by individual.

A Factorial Correspondence Analysis (FCA) was performed on the data with the procedure CORRESP from SAS (SAS 1992) to show the structure of the data. Distances between all individuals were computed according to Sokal and Michener's index (Sokal and Michener 1958): dij = a + d/a + b + c + d, where a is the common band between two individuals, i and j, b and c the bands present in one individual (i or j, respectively), and d the common absence of bands between the two individuals i and j. Dendrograms were constructed with the UPGMA algorithm (Benzécri 1973).


Out of 30 primers checked, 6 (A1, A2, A9, A10, C15, E15) were retained because they reveal the highest level of polymorphism. These were then applied on all the individuals. They enabled us to note 101 markers. Thirteen are conserved in Olea individuals and three are conserved between Ligustrum vulgaris and Olea genus on the basis of size homology. Twenty-three markers are unique to L. vulgaris.

Specific markers were found only for related species: four for O. maroccana and one for O. cuspidata from Asia. We have one unique marker for 'Lechin de Sevilla' (Spain) and another for 'Zaity' (Near East). The cultivar markers have a mean frequency of presence (0.57) lower than that in wild populations (superior to 0.69)(Table 3).

Analysis and comparison with oleaster cultivars

Most cultivars (56/64) are differentiated with our markers. Four couples of cultivars cannot be distinguished: 'Picholine Marocaine' and 'Sigoise', 'Cellina' and 'Frantoio', 'Meski' and 'Bid el Hamam', 'Belgentier' and 'Merhavia'. The probability to obtain the same pattern for two different individuals was 3.9 -10-4 (for Mean of Nei and Li similarities = 0.592 and Mean of marker number by individual = 14.98).

The first dimension (13.9%) of the FCA performed on the cultivar data permitted separation of 'Chemlal de Kabylie' from the other cultivars. On the other dimensions (explaining less than 9% of the variability) several cultivars are separated, namely 'Zaity' and 'Lechin de Sevilla' (data not shown). FCA of Mediterranean olive trees (wild, feral and cultivars) separated wild forms from both Corsica and Morocco, 'Sabina', 'Filitosa', 'Chemlal de Kabylie', 'Galega' and 'Koroneiki' from the rest of the individuals. We found 13 markers present in oleasters from the Western Mediterranean and absent in all cultivars from the Near East (except 'Koroneiki' from Crete). Seven of these markers are present in the five cultivars clustered with the oleasters from Morocco and Corsica.

Analysis of the wild forms

The FCA constructed with wild form data only enabled us to distinguish four groups corresponding to: (1) O. europaea from Mediterranean basin, (2) O. maroccana from southern Morocco, (3) O. africana from Kenya, and (4) O. cuspidata from Asia.

Phenogram constructed with total data

On the phenetic tree constructed on the total data (Fig. 2), a clear structure fitting a geographic origin of Mediterranean olive was not obtained but small groups corresponding to regions were found. The oleasters are related with cultivars. The oleasters from Corsica and from Morocco form homogeneous groups (except for 'Morocco 5' and '6' which are related with 'Picholine Marocaine') in contrast to wild forms from southern France. The oleasters from Corsica and from Morocco form a subgroup (subgroup 1.2) and are related with 'Chemlal de Kabylie' (Algeria), 'Filitosa' (Corsica), 'Sabina' (Corsica), 'Galega' (Portugal), 'Pigale' (France) and 'Koroneiki' (Crete). The forms from Asia (O. cuspidata) and from Africa (O. maroccana and O. africana) form two well-separated groups.

Fig. 2. Dendrogram based on Sokal and Michener (1958) distances computed between all studied individuals and constructed according to UPGMA algorithm. The index numbers correspond to the geographic origin of the individuals: 1 = eastern Mediterranean basin; 2 = Italy and Yugoslavia; 3 = France; 4 = Iberian Peninsula; 5 = Maghreb; and 6 = Asia.


Analysis of cultivars in relation to oleasters

RAPDs permit identification of most cultivars. The probability of obtaining the same pattern between two different cultivars is very low (3.9x1-5). Nevertheless, we cannot eliminate the possibility that occasional mutations can generate phenotypic variations leading to different clones. Such a phenomenon has been reported in numerous vegetatively propagated species (Zohary and Hopf 1993).

The pattern obtained does not show a clear geographic origin of the olive. This can be expected because of commercial exchanges between the different countries of the Mediterranean basin since antiquity. It seems that there exists a high variability in cultivated olive. The number of polymorphic markers is higher in the cultivar group than in the wild populations but this depends on the population size which is insufficient here to detect all polymorphic markers (Table 3). Crosses between forms of different geographic origin have occurred and this probably explains the large genetic diversity (Elant 1976). Thus, a high level of heterozygosity is therefore expected and the low mean frequency of present markers in the cultivar group reflects this fact.

Some varieties which differentiate from the rest of cultivars support the theory that selection of olive occurred in different places. Since 'Chemlal de Kabylie' is male-sterile (Ouksili 1988), we suggest this cultivar is derived from an interspecific hybrid.

In contrast to oleasters from Corsica and Morocco, the trees from southern France are very close to cultivars. These forms did not exhibit specific or unique markers and this suggests that they are feral forms. The present climate of southern France is not favorable for the continued existence of indigenous forms (Zohary and Hopf 1993) although these existed during the Holocene in this region (Solari and Vernet 1992). The occurrence of relics of former populations is, therefore, likely.

Relationships between species of O. europaea complex

Olea maroccana is closely related to O. africana from Kenya. This supports the Maley hypothesis (1980) of a tropical origin of Olea (O. laperrinei) in the mountains of North Africa. The distribution of O. laperrinei in the Atlas Mountains is unknown (except for O. maroccana). Olea maroccana is certainly a relic of an ancient population of tropical Olea. Other relic populations could exist locally and we suggest that O. laperrinei was distributed all over the mountains of Maghreb. The sympatry between cultivated forms and O. maroccana is suggested by Zohary (1994) and we can predict that they may hybridize. The existence of introgression of O. maroccana in cultivated forms is therefore possible but not detected yet. On the basis of our molecular data, O. africana appears intermediate between O. cuspidata and O. maroccana. It is differentiated from O. cuspidata from Asia, suggesting a differentiation between Asian and Eastern African forms as suggested by Zohary (1994) because of the discontinuous geographical distribution and an adaptation to different environments. We recognized the different subspecies as described by Green and Wickens (1989) and we will be able to detect easily any hybrids.

The origin of olive by hybridization between oleasters and related species suggested by Elant (1976) is not supported here. We now have to verify whether O. chrysophylla is related to the cultivated forms from the Mediterranean basin as suggested by Chevalier (1948).

Table 3. Mean frequency of present markers and percentage of polymorphic markers in studied populations and cultivar group.

Groups of individuals

No. of individuals

No. of polymorphic markers

Mean freq. of present markers

R p/c

Corsica 1 to 11





Morocco 1 to 9





Hérault 1 to 5





O. maroccana





Cultivated forms





Rp/c corresponds to the following ratio - number of polymorphic markers: number of conserved markers in the group of individuals. The number of individuals studied in O. maroccana and southern France populations is insufficient to consider this ratio.


The present-day olive cultivars may have derived from oleasters of different Mediterranean areas. Human migration and exchange of germplasm have increased variability of the cultivated and the wild forms. Nevertheless, the classification of the indigenous wild forms in the Mediterranean basin is awaited. Evidence of hybridization with related species was not found, although two origins of Olea in the Mediterranean basin are suggested (Middle East and North Africa). The present study is preliminary and analysis of several more accessions of cultivated and wild forms is needed, particularly of the related species (O. chrysophylla from Yemen, O. laperrinei from Hoggar) to verify our results.


This work was supported by a grant from the EC-FAIR No. CT 95-0689 on the 'Evaluation of olive genetic diversity and identification of molecular markers related to the aspects of food products, oil and olive, their mapping for marker assisted breeding'.


Benzécri, J.P. 1973. L'analyse des données. Tome I. La Taxonomie (Dunod, ed.). Paris, France [in French].

Chevalier, A. 1948. L'origine de l'olivier cultivé et ses variations. Rev. Int. Bota. Appl. Agric. Trop. 28:1-25 [in French].

Elant, H. 1976. Olive. Pages 219-221 in Evolution of Crop Plants (M.W. Simmonds, ed.). Longman, London, UK.

Galili, E., M. Weinstein-Evron and D. Zohary. 1989. Appearance of olives in submerged Neolithic sites along the Carmel coast. J. Isr. Prehistoric Soc. 22:95-97.

Green, P.S. and G.E. Wickens. 1989. The Olea europaea complex. Pages 287-299 in The Davis and Hedge Festschrift. Edinburgh University Press, UK.

Isabel, N., J. Beaulieu and J. Bousquet. 1995. Complete congruence between gene diversity estimates derived from genotypic data at enzyme and random amplified polymorphic DNA loci in black spruce. Proc. Nat. Acad. Sci. 92:6369-6373.

Jeandroz, S., N. Frascaria-Lacoste and J. Bousquet. 1996. Molecular recognition of the closely related Fraxinus excelsior and F. oxyphylla (Oleaceae) by RAPD markers. Forest Genet. 3:237-242.

Koller, B., A. Lehmann, J.M. McDermott and C. Gessler. 1993. Identification of apple cultivars using RAPD markers. Theor. Appl. Genet. 85:901-904.

Liphschitz, N., R. Gophna, M. Hartman and G. Biger. 1991. The beginning of olive (Olea europaea) cultivation in the Old World: A reassessment. J. Archaeol. Sci. 18:441-453.

Maire, R. 1933. Études sur la flore et la végétation du Sahara central. Pages 166-168 in Mémoires de la Société d'Histoire Naturelle de l'Afrique du Nord No. 3, Mission du Hoggar II, Hoggar, Algeria [in French].

Maley, J. 1980. Les changements climatiques de la fin du Tertiaire en Afrique: Leur conséquence sur l'apparition du Sahara et de sa végétation. Pages 63-86 in The Sahara and the Nile (M. Williams and H. Faure, eds.). AA Balkema, Rotterdam, Netherlands [in French].

Nybom, H. and H.K. Hall. 1991. Minisatellite DNA 'fingerprints' can distinguish Rubus cultivars and estimate their degree of relatedness. Euphytica 53:107-114.

Ouksili, A. 1988. Observations sur la microsporogénèse et recherche de pollinisateurs chez deux variétés d'olivier. Olivae 16:23-27 [in French].

Quillet, M.C., N. Madjidian, Y. Griveau, H. Serieys, M. Tersac, M. Lorieux and A. Bervillé. 1995. Mapping genetic factors controlling pollen viability in an interspecific cross in Helianthus sect. Helianthus. Theor. Appl. Genet. 91:1195-1202.

Saghai-Maroof, M.A., K.M. Soliman, R.A. Jorgensen and R.W Allard. 1984. Ribosomal DNA spacer length polymorphisms in barley: Mendelian inheritance, chromosomal location, and population dynamics. Proc. Nat. Acad. Sci. 81:8014-8018.

SAS Institute. 1992. SAS/STAT Language Guide for Personal Computers. Release 6.03. SAS Institute, Gary, North Carolina, USA.

Sauvage, C. and J. Vindt. 1952. Flore du Maroc analytique, descriptive et illustrée: Spermatophytes. Pages 95-96 in Travaux de l'Institut Scientifique Chérifien (Internationales, ed.). Tangers, Morocco [in French].

Sokal, R.R. and C.M. Michener. 1958. A statistical method for evaluating systematic relationships. Univ. Kansas Sci. Bull. 28:1409-1438.

Solari, M.E. and J.L. Vernet. 1992. Late glacial and Holocene vegetation of the Corbières based on charcoal analysis at the Cova de L'Espérit (Salses, Pyrénées Orientales, France). Rev. Paleobot. Palynology 71:111-120.

Terral, J.F. and G. Arnold-Simard. 1996. Beginnings of olive cultivation in eastern Spain in relation to Holocene bioclimatic changes. Quartenary Res. 46:176-185.

Turrill, W.B. 1951. Wild and cultivated olives. Kew Bull. 3:437-442.

Vigouroux, A., G. Besnard, K. Sossey-Alaoui, M. Tersac and A. Bervillé. 1997. Le statut d'hybride de Platanus acerifolia confirmé et celui de P. densicoma mis en évidence à l'aide de marqueurs génétiques moléculaires: conséquences. Acta Bot. Gallica 144(2):243-251.

Wang, W.Y., R.C. Pai, C.C. Lai and T.P. Tin. 1994. Molecular evidence for the hybrid origin of Paulownia taiwaniana based on RAPD markers and and RFLP of chloroplast DNA. Theor. Appl. Genet. 89:271-275.

Willcox, G.H. 1996. Evidence for plant exploitation and vegetation history from three early Neolithic pre-pottery sites on the Euphrates (Syria). Veget. Hist. Archaeobot. 5:143-152.

Zohary, D. 1994. The wild genetic resources of the cultivated olive. Acta Horticult. 356:62-65.

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.

Zohary, D. and O. Spiegel-Roy. 1975. Beginnings of fruit growing in the Old World. Science 187:319-327.

Previous Page Top of Page Next Page