On the ‘lost’ crops of the neolithic Near East

Abstract

The claim that the ‘classic’ eight ‘founder crop’ package (einkorn wheat, emmer wheat, barley, lentil, pea, chickpea, bitter vetch, and flax) underlying the emergence of agriculture in the Near East is a relic of a larger number of domesticated species is addressed. The ‘lost’ crops concept relies on the idea that additional taxa were at certain points in time and in certain locations genuine crops, which were later abandoned. The issue is highly relevant to the debate concerning mono- versus polyphyletic domestication, because if there were numerous ‘false starts’ that were subsequently lost, this implies that plant domestication occurred over a protracted time period, and across a wide geographic range. Different criteria were used for declaring those taxa as ‘lost’ crops, including, but not limited to (i) identification in archaeobotanical assemblages of grains from species which are not known as crops at present; (ii) identification of such grains in what is interpreted to have been Neolithic storage facilities; and (iii) recent botanical observations on populations of crop wild relatives in disturbed habitats. The evidence for four presumed ‘lost’ crops (wild oat, rambling vetch, rye, and wild black lentil) and the broad bean is evaluated, and discussed in light of data on Croatian and Israeli wild pea, and Moroccan wild lentil in disturbed habitats. Based on present knowledge, the broad bean might emerge as a founder crop (without an identified wild progenitor). The same may hold true for rye, which was never lost since its adoption in the Pre-Pottery Neolithic B period in Anatolia. In the remaining three cases, there are alternative, more likely, explanations for the archaeological finds or the recent botanical observations rather than ‘lost’ domestication episodes.

Introduction

The question of the geographic origin of plant domestication attracted much attention during generations of scholarly investigations (e.g. de Candolle, 1885; Vavilov, 1951; Helbaek, 1959; Braidwood, 1967; Harlan, 1992; Zohary and Hopf, 2000; Abbo et al., 2010a ; Zohary et al., 2012). Based on immense archaeobotanical databases, botanical observations, genetic data, and archaeological considerations on Near Eastern plant domestication, the so-called Near Eastern Fertile Crescent was proposed as the centre of Near Eastern agriculture (e.g. Zohary and Hopf, 2000; Zohary et al., 2012). Based on archaeological and environmental considerations, Braidwood (1960, 1975) proposed his nuclear zone hypothesis (see Redman, 1978) and although he referred to major parts of the Fertile Crescent, he highlighted the piedmont hills and lower intermountain valleys of the Zagros-Taurus arc as the nuclear zone. More recently, based on geo-botanical, archaeological, archaeobotanical, and genetic evidence, a hypothesis suggesting a well-defined ‘core area’ origin for all Near Eastern Neolithic grain crops was formulated (Lev-Yadun et al., 2000; Gopher et al., 2001; Abbo et al., 2010a ). The core area hypothesis maintains that the Near Eastern Neolithic crop package including both cereals and legumes (Abbo et al., 2009, 2010b ), namely diploid einkorn wheat (Triticum monococcum L.), tetraploid emmer wheat (T. turgidum L.), barley (Hordeum vulgare L.), lentil (Lens culinaris Medikus), pea (Pisum sativum L.), chickpea (Cicer arietinum L.), bitter vetch [Vicia ervilia (L.) Willd.], and flax (Linum usitatissimum L.), arose within a relatively restricted area in south-eastern Turkey (Lev-Yadun et al., 2000) within a short time frame (e.g. Ladizinsky, 1987; Hillman and Davies, 1999), and radiated from there in all directions (Gopher et al., 2001), and that each and every crop plant was adopted only once or a very limited number of times; that is, monophyletic domestication (sensu Zohary, 1996; Abbo et al., 2010a ).

An alternative hypothesis favours a geographically diffused Near Eastern plant domestication pattern implying a series of events in which certain crops were domesticated through a sequence of genetically and culturally independent (i.e. autonomous) occurrences (e.g. Willcox, 2005; Weiss et al., 2006; Brown et al., 2009; Allaby et al., 2010; Fulle et al., 2011, 2012a , b ). The debate concerning ‘core area’ versus diffused domestication is part of the discussion on Near Eastern plant domestication which, as a whole, can be characterized by a series of bipolar views partly listed below (reviewed by Abbo et al., 2011).

• Time frame. Domestication is viewed as having occurred in one major event (singular timing) versus a recurrent event repeated multiple times.

• Rate. Domestication is perceived to have unfolded as a fast, revolutionary change achieved in a single major move versus a slow and gradual (protracted) process perhaps involving numerous stages.

• Scope. The scale of domestication is considered as an inclusive act of deployment involving most package plants together versus an exclusive act by which domestication of each and every species has occurred independently.

• Intention. The process of domestication as a whole is deemed as a conscious effort to select and manipulate the various plant species that were eventually transformed versus an unconscious dynamics, and automatic selection occurring inadvertently in cultivated fields.

Interestingly, most advocates of the geographically diffused domestication process also favour a protracted and unconscious domestication process slowly unfolding through several millennia (e.g. Tanno and Willcox, 2006a ; Weiss et al., 2006; Allaby et al., 2008; Fuller et al., 2012a , b ), thereby making all the above aspects of Near Eastern plant domesticated inter-related.

In support of the multiple plant domestication foci model, some of the above-mentioned recent reviews have presented a claim regarding ‘lost’ founder crops (see Fuller et al., 2011, 2012a , b ). Moreover, Fuller et al. (2011, 2012b ) state that there were numerous ‘false starts’ and dead ends of early cultivars, and that as many as 16 or 17 (partly named) species were involved in (pre-domestication) cultivation in the Near East. This would indicate that the Neolithic Near Eastern founder crop package (listed above) is only a relic of a large number of taxa that were auditioned as candidate crops (sensu Smith, 2011) during millennia of pre-domestication cultivation, and that several such taxa were, at certain points in time (mostly not specified in details) and in certain locations, genuine crops which were later abandoned. Listed as one out of eight reasons, in favour of a protracted and geographically diffused domestication model (Fuller et al., 2012b ), the ‘lost’ crops concept might seem a minor issue. However, being fundamental to the discussion on the geography of Near Eastern plant domestication (localized versus diffuse), the ‘lost’ crops concept becomes inseparable and therefore highly relevant to all other inter-related bipolar aspects of plant domestication (Abbo et al., 2011). More specifically, the ‘lost’ crops issue is central to the debate concerning mono- versus polyphyletic domestication, because if there were numerous ‘false starts’ that were subsequently lost, it implies that plant domestication took place over a protracted time period, and possibly also across a wide geographic range.

The following 10 plant species are listed by Fuller et al. (2012b) as additional founder crops of Near Eastern agriculture: the two-grained form of Triticum monococcum or T. urartu Tuman. (einkorn wheat), Secale sp. (rye), Triticum turgidum/timopheevi (striate emmeroid tetraploid wheat), Avena sterilis L. (oat), Vicia peregrine L. (rambling vetch), Vicia faba L. (broad bean), Lathyrus sativus L. (grass pea), Lens nigricans (M. Bieb.) Godron (black wild lentil), and Ficus carica L. (common fig). Three out of the above 10 species are extant crops, namely broad bean, grass pea, and the common fig, while the other seven species apparently represent some of the presumed numerous ‘false (domestication) starts’.

Different criteria were used for declaring those taxa as ‘lost’ crops, including, but not limited to (i) the identification, in the archaeobotanical assemblages, of grains from species which are not known as crops at present; (ii) the identification of grains from such taxa in what is interpreted to have been storage facilities; and (iii) botanical observations on populations of crop wild relatives in disturbed habitats. The ‘lost’ crops list presented in table 1 of Fuller et al. (2012b) is highly important to the debate concerning the core area versus multiple origins mainly because it includes several of the so-called ‘non-core crops’ such as the wild oat of Weiss et al. (2006) and the rambling vetch of Melamed et al. (2008), both from the lower Jordan Valley, and the west Mediterranean black lentil of Ladizinsky et al. (1983). Hence, accepting the arguments in favour of these ‘lost’ non-core area crops may become a decisive consideration in viewing the Mediterranean basin as a non-centre of plant domestication (Harlan, 1971; Abbo et al., 2010a ). In this paper, we evaluate several of the additional founder crops cases based on the evidence and arguments presented in the primary literature sources cited by Fuller et al. (2012b) and consider their bearing on the study of Neolithic Near Eastern plant domestication.

Evaluating the evidence and arguments for several ‘lost’ crops cases

Broad bean

We see no reason why the broad bean cannot be added to the classical eight ‘founder’ crops list (Zohary and Hopf, 2000), as advocated earlier by Tanno and Willcox (2006b). We presume that Zohary and Hopf (2000, and in earlier editions) refrained from doing so simply because the wild ancestor of broad bean is still elusive (Zohary and Hopf, 2000). We agree that due to the fact that broad bean is a major grain crop to this very day it cannot be termed a ‘lost’ founder crop. What might have been lost is its wild ancestor (Ladizinsky, 1975a ), although careful botanical surveys may still bring it to light in the future. Indeed, early broad bean remains [10 350 calibrated years before present (BP)] were excavated (alongside chickpea remains) in Early Pre Pottery Neolithic B (EPPNB) Tell Ein el-Kerkh in northern Syria (Tanno and Willcox, 2006b ) which is only 150 km west of the core area of Lev-Yadun et al. (2000). Therefore, it is quite likely that both taxa originated well within, or very close to, the core area of Lev-Yadun et al. (2000).

Wild oat

Weiss et al. (2006) mention wild oat in the context of the proposed Pre Pottery Neolithic A (PPNA) pre-domestication cultivation. Indeed, a pile of charred grains containing large amounts of wild barley (Hordeum spontaneum C. Koch) and wild oats (260 000 and 120 000 seeds, respectively) was found in the excavation of locus #11 in the PPNA site Gilgal I. In terms of weight, however, these numbers amount to no more than 3kg of oats and 8kg of barley seeds. Such a quantity of seeds can easily be obtained from several hours of gathering in wild populations by a group of 2–3 people (see Harlan, 1967; Ladizinsky, 1975b ; Kislev et al., 2004; Abbo et al., 2008a ), in ecologies located only a few kilometres to the west of this PPNA site. Therefore, cultivation is neither the only possible nor the most likely explanation for these oat and barley finds. Also note that Weiss et al. (2006) have stated that the presumed PPNA cultivation of oats in the southern Levant ‘did not necessarily lead to its domestication, which occurred much later in a faraway region’. Another difficulty of the proposed scenario of PPNA oat cultivation in the lower Jordan Valley in the southern Levant is the mere fact that this oat species is highly adapted to the ecology of the region by its early flowering and excellent plastic tillering potential. Namely, in dry years, it will produce only a few panicle-bearing tillers and will mature in the early spring mostly prior to barley, presenting a classical example of drought escape. In more lush years with a longer rainy season, this species will produce many fertile tillers, hence exploiting the advantage of the better moisture regime. Therefore, we wonder why did the PPNA ‘cultivators’ (or their PPNB successors) abandon such a presumed crop species so well adapted to the regional conditions. After all, the simplest expectation regarding the cultivation of such an adapted species is that such practice would have persisted for millennia, if not to this very day.

Rambling vetch

A total of 313 seeds of rambling vetch were found in eight different loci bearing archaeobotanical remains (out of ~100 excavated loci) at the PPNA site of Netiv Hagdud (located <1 km west of Gilgal I) mixed together with thousands of wild barley seeds (Kislev, 1997; Melamed et al., 2008). A first reservation concerns the naming of this toxic legume as ‘edible’. Indeed, Melamed et al. (2008) devoted a considerable part of their discussion to the neurotoxins of this legume and to the possible ways by which its seeds might have been processed to render it edible. Several toxic plant species are being used as food plants (e.g. cassava), so the option that rambling vetch seeds were treated, one way or another, to reduce their toxicity may be true to life. However, if this was indeed the case, one would expect the toxic seed to have been stored separately, first to avoid contamination of truly edible grains caches, and secondly to allow an effective detoxification processes. The same would hold true if the inhabitants of Netiv Hagdud used to mix rambling vetch seeds with other edible grains (as suggested by Melamed et al., 2008) and thereby dilute their effect, because to ensure the effectiveness of such practice both types of grains (toxic versus edible) must be kept separately. Hence, the fact that the rambling vetch seeds were scattered (and mixed with wild barley grains) in eight different loci suggests, in our view, that it is unlikely to have been a foodstuff. Melamed et al. (2008) write that rambling vetch could have been collected in wild habitats, or ‘more reasonably it could have been brought from an artificial habitat—namely, from cultivated fields’. Therefore, the reasoning of this claim for cultivating rambling vetch is similar to the claim of Weiss et al. (2006) of wild barley and oat cultivation in both PPNA sites of Netiv Hagdud and Gilgal I. It is interesting to note that Weiss et al. (2006) have used two criteria to justify their idea of cultivation of morphologically wild plants in these two PPNA lower Jordan Valley sites. Their first criterion was the amount of seeds, indicating (in their view) harvest from a cultivated field rather than gathering from the wild. The second criterion was the presence of seeds of presumed weeds of cultivation, indicating human-induced vegetation changes. Interestingly, Melamed et al. (2008) reported that most of the identified plant species at Netiv Hagdud are not known today as discarded weeds. Therefore, the finds of rambling vetch do not conform to either of the criteria used by Weiss et al. (2006) for declaring cultivation following this very same research group. Moreover, Melamed et al. (2008) stress that the potential grain yield of rambling vetch is higher than the potential yield of lentils, and its seeds are three times as large. Therefore, we wonder why such an attractive, and so-called ‘edible’, legume was abandoned in favour of the flimsy low-yielding lentil. A parsimony alternative interpretation of this archaeobotanical find may be that rambling vetch was neither cultivated nor a deliberate target for collecting, but simply reached the site as an unavoidable companion of other truly edible species as described for wild grass pea (Lathyrus sp.) seeds and discussed in details by Abbo et al. (2008b).

Rye

The notion of Central European rye (Secale cereale L.) domestication was formulated by Vavilov (1917) based on the observation that rye is much more adapted to certain north European growing environments than wheat and barley. Because wild rye (Secale vavilovii Grossh.) is not native to Europe, Vavilov (1917) suggested that some form(s) of weedy rye with partly or fully non-brittle rachis was introduced to Europe as a companion weed of the Near Eastern wheat and barley (Zohary and Hopf, 2000, pp. 70–71), and later became a (secondary) crop in its own right due to its better adaptability to the edaphic conditions of Central Europe. Rye is an important crop to this very day, so, regardless of its geographical and cultural origin, we find it difficult to accept the proposed terminology for rye as a ‘lost’ (founder) crop. Fuller et al. (2012b) follow the suggestion of Hillman et al. (2001) that rye was cultivated in (Late) Epipalaeolithic Tell Abu Hureyra. The claim for Epipalaeolithic rye cultivation was based mainly on two lines of evidence. The first line of evidence concerned the presence of weeds of cultivation (therein). The second is the find of five charred, plump, and therefore so-called domestic-type grains, of which three yielded Epipalaeolithic AMS ¹⁴C dates (Hillman et al., 2001). The circular nature of the ‘weeds of cultivation’ argument was already pointed out by Abbo et al. (2012). In brief, when one assumes that there was ‘necessarily a stage of production (cultivation) that precedes morphological domestication’ (Fuller, 2007), such (presumed weeds) seeds may be interpreted as ‘weeds of cultivation’ (e.g. Willcox et al., 2008, and references therein). However, to use the presence of such (presumed weeds) seeds after making the above assumption (regarding the necessity of the pre-domestication cultivation) as a proof of pre-domestication cultivation is a circular argumentation. Also note that in a recent reassessment of the evidence for cultivation of wild plants at Tell Abu Hureyra, Colledge and Conolly (2010) stated as follows: ‘cultivation is not needed to explain the changes in proportion of plant taxa’, and they concluded that a model which does not include cultivation of cereals and legumes in the Late Epipalaeolithic is more parsimonious (p. 137, therein). Accepting the latter conclusion of Colledge and Conolly (2010) moves the focus to the five plump rye grains. One of the present authors (MH) warned already in 1998 (Hillman et al., 2001, p. 389, fourth paragraph, left column), that drought-mediated fertility problems associated with sink–source distortion of photosynthetic products can easily explain how reduced seed set success may result in a few plump grains. Despite the problematic and meagre evidence for rye cultivation in Tell Abu Hureyra, this case is kept alive by some authors apparently because it may provide a link with the previous Late Natufian culture (e.g. Hayden et al., 2012). However, if indeed tough rachis rye first appears in PPNB Can Hasan III in south-central Anatolia, as indicated by Hillman (1978), this implies that rye has developed morphological domestication traits while still in Anatolia, well before its introduction into the Central European environs. Accepting the old Hillman (1978) suggestion for an Anatolian Neolithic origin of rye domestication, one need not assume that rye was a genuine crop (e.g. in Late Epipalaeolithic Tell Abu Hureyra), which was later abandoned (i.e. became a ‘lost’ crop). To conclude the rye case we note that deciding between Vavilov’s (1917) hypothesis on rye domestication and that of Hillman (1978) is outside the scope of the present review; however, it seems that under both reconstructions rye does not deserve the status of a ‘lost’ crop.

Wild black lentil

Ladizinsky et al. (1983) sampled wild black lentil populations (L. nigricans) across the Mediterranean basin in the early 1980s and noted that L. nigricans stocks sampled from secondary habitats (i.e. human made or disturbed) have larger grain size compared with those sampled in primary (undisturbed or less disturbed) habitats. Ladizinsky et al. (1983) also found that L. nigricans from the primary habitats form a distinct cytogenetic gene pool, which is reproductively isolated from the populations sampled from secondary habitats in northern Italy, south-eastern France, and along the Adriatic coast of Croatia. In their discussion, Ladizinsky et al. (1983) wrote that a possible explanation for their observations of the L. nigricans populations native to human-made habitats is that those populations have evolved from cultivated lentil derived from a past European domestication of this wild lentil taxon (thereby assuming that those populations are in fact feral) independent of the Near Eastern L. orientalis (Boiss.) domestication which gave rise to domesticated lentil. However, as noted by Ladizinsky et al. (1983) and reiterated by Fuller et al. (2012b), no archaeological evidence for such a west Mediterranean lentil domestication centre was ever presented. Therefore, the hypothesis of Ladizinsky et al. (1983) can be considered only as a ‘possible explanation’ (among others) to account for their observations and cannot be regarded as evidence for past domestication.

Wild crop relatives in secondary habitats and the ‘lost’ crops concept

As mentioned earlier, the importance of the ‘lost’ crops claim is its role in promoting the idea of independent domestication foci and their diffused geographic pattern. The claim for a past west Mediterranean domestication of L. nigricans by Ladizinsky et al. (1983) is very important since by listing it among the Near Eastern ‘lost’ crops, Fuller et al. (2012b) considerably enlarge the geographic context of the debate regarding the core area hypothesis of Lev-Yadun et al. (2000). It is therefore of interest to consider additional recent field observations of crop wild relatives in secondary Mediterranean habitats and evaluate their value for declaring lost domestication episodes.

It is worth noting that while arguing in favour of a geographically diffused pattern of plant domestication and in favour of multiple ‘domestication events’ for some crops, Gross and Olsen (2010) suggested that the monophyletic signatures obtained from the analyses of present-day cultigens and wild relatives (e.g. Heun et al., 1997) are probably due to the fact that many of the presumed ancient (genetically and geographically independent) domesticated lineages did not survive to the present day (Gross and Olsen, 2010; Fig. 1). While some of our field observations of crop wild relatives (below) in secondary Mediterranean habitats might seem anecdotal and/or overstated, we think that such an elaboration may contribute to the discussion on Near Eastern plant domestication in two ways. First, our observations may help to weigh the level of details of the published cases cited by Fuller et al. (2012b) and therefore can help to assess their potential as possible evidence for past domestication episodes. Secondly, and no less important, if indeed one accepts the presence of crop wild relatives in secondary habitats as genuine evidence for past domestications, these should be seriously considered as a living proof for the (thus far) hypothetical ‘lost’ crops lineages (sensu Gross and Olsen, 2010) and as such merit careful description.

Fig. 1.

A map of the Bet-Guvrin site (near Kiryat-Gat, Israel), in which a population of ‘southern’ Pisum humile persists in an abandoned olive grove. A congeneric wild pea species, P. fulvum, thrives in the adjacent woodland hillside, and also invades the olive grove, while P. humile is confined to the man-made habitat.

Wild pea (Pisum elatius) in Croatia

In his field diaries, Gideon Ladizinsky (in Hebrew, unpublished) mentioned a single wild pea (P. elatius M. Bieb.) plant growing along a stony footpath in an olive grove, some 2 km south of Primosten near the Adriatic coast, where he encountered one of the above-mentioned L. nigricans populations. In June 2011, one of the present authors (SA) visited this site (43°34’53.4’’N, 15°56’10.4’’E) in search of wild legumes. Unfortunately, plants of wild black lentil were not encountered in June 2011, but several wild P. elatius plants were observed in the olive grove. Would the readers consider entertaining a hypothesis concerning an Adriatic P. elatius domestication based on these two observations (made in 1980 and 2011) of this species in a human-made habitat? Such reasoning might seem inappropriate. It could be argued that the observations of an isolated population of wild P. elatius is not equivalent to over a dozen secondary habitats harbouring wild black lentil populations reported by Ladizinsky et al. (1983); hence, granting a ‘lost’ crop status might be unjustified in the case of the wild P. elatius from Primosten, Croatia.

Israeli wild pea (Pisum humile)

Wild P. humile can be found in certain secondary habitats in Israel. These are represented in the herbarium of The Hebrew University of Jerusalem, Israel, where dozens of P. humile (Boiss. et Noe) specimens are stored, and include the herbarium specimens HUJ-HERB-VP-102653 through HUJ-HERB-VP-102668, many of which underwent taxonomic verifications by Professor Daniel Zohary in 2004.

In their work, Ben-Zeev and Zohary (1973) mention a ‘southern P. humile’ type occurring along roadsides and field edges at the Judean foothills in the Bet-Guvrin area, Israel. We have observed such plants near Khirbet Minya close to the north-western shore of the Sea of Galilee and near Kibbutz Gesher (~10 km) south of the Sea of Galilee, along paved and unpaved roads and field edges in the Bet-Shean Valley at the foot of Mt Gilboa (Fig. 1), and elsewhere. Seeds from all the populations marked in Fig. 1 were sampled and plants were grown in screen-houses in Rehovot, Israel. Without exception, these southern P. humile types have fully dehiscent pods, camouflage seed coloration, and strong (~90%) seed dormancy mediated by water-impermeable seed coats. Despite their wild-type seed dispersal mode and wild-type seed dormancy, and again without exception, these southern P. humile never invade adjacent less disturbed habitats where such habitats are available. In many of the sites, just a few metres apart, in the adjacent woodland and or rangeland, a different wild pea species (P. fulvum Sm.) is a common element, for example near Lakhish, Bet-Guvrin, Mt. Gilboa, and the eastern Galilee. Over the past 6 years we have documented the situation in one such southern P. humile site, near Bet-Guvrin, Israel (Fig. 1). The site is an abandoned olive grove, in which no husbandry operations were carried out at least since autumn 2005 (and most probably since several years earlier). The bordering rocky hills are partly under Aleppo pine reforestation and partly under degraded kermes oak–pistachio plant formation, and serve for occasional cattle grazing (Fig. 1). In the olive grove, a large population of the southern P. humile persists, among other annuals and perennials, which invaded it from the adjacent rangeland. The many years since abandonment of the olive grove can be deduced from the fact that several nicely developed perennial shrubs typical of the surrounding vegetation (e.g. Prasium majus L.) established themselves among the olive trees. Interestingly enough, among the invading annual legumes, wild P. fulvum can be encountered in small numbers, but for the time being, only along the 3–4 rows of trees adjacent to the rocky hillside. On the other hand, not a single P. humile individual was ever encountered outside the olive grove or on the adjacent rocky slope, despite the fact that it is fenced with barbed wire that does not form a barrier for seed dispersal (Fig. 1).

According to the ‘lost’ crops rationale, the Israeli southern P. humile populations may perhaps serve as evidence for a past pea domestication centre, thereby granting further support to the claims of Fuller et al. (2012b). Moreover, if accepting Gross and Olsen’s (2010) hypothesis regarding ‘lost’ domesticated lineages, and in conjunction with Harlan’s (1951) methodology of high-resolution mapping of crops’ gene centres, Israel may well be declared as a mosaic of multiple independent southern pea domestications in its own right. This situation is similar to the status of the Croatian and Italian L. nigricans populations as documented by Ladizinsky et al. (1983). In fact, based on the geography of the L. nigricans inhabiting secondary habitats and the ‘lost’ founder crops reasoning, one can easily be tempted to propose a number of independent L. nigricans domestication centres. After all, a southern Levant and an eastern Turkey lentil domestication centre were proposed by Weiss et al. (2006). Accordingly, the Croatian, Italian, and possibly the French L. nigricans populations could be portrayed as each resulting from a ‘lost’ independent domestication episode.

Our view is that without an appropriate archaeological and, more specifically, archaeobotanical support, such claims for ‘lost’ domestication episodes based on the occurrence of crop wild relatives in human-made habitats are no more than unproven hypotheses, and as such can hardly grant any support to the diffused domestication model.

Moroccan Lens lamotii

This wild lentil species, typical of the west Mediterranean basin, can be found within the Roman ruins of the site of Volubilis (34°04’22’’N, 5°33’15’’W), Morocco (observed on different occasions by both G. Ladizinsky and S. Abbo; Fig. 2). The surrounding hills are mostly overgrazed, and therefore we failed in spotting this wild lentil species outside the Volubilis compound. Again, can we seriously consider these observations as evidence for past domestication and or cultivation of L. lamotii during the classical, Roman (or earlier) periods? Also maybe this hypothetical (lost) Moroccan (domesticated) lentil should be seen as the missing (legume) link of a ‘Lost West Mediterranean’ crop package together with the Moroccan barley of Molina-Cano et al. (1987, 2005). Again, we think that without additional support based on independent lines of evidence (e.g. archaeobotany and genetics), such an idea is unacceptable. We argue that the purported case of the Croatian and Italian L. nigricans domestication is by no means stronger than a hypothetical lost domesticated L. lamotii and its possible role in a lost ancient Moroccan crop package.

Fig. 2.

A wild, pod-bearing, plant of Lens lamotii in the Roman compound of Volubilis, Morocco (May 2009, photo by S. Abbo).

Conclusions

Prior to the PPNB and to this very day, a large number of wild plant species were gathered across the Near East (e.g. Ertug, 2000), but ultimately only a few were domesticated in what we consider to have been well focused effort and knowledge-based decisions (Abbo et al., 2009, 2010b , 2011). Hence, contrary to the ‘lost’ crops concept, we argue that skilful archaeobotanical identifications of seed remains (e.g. rambling vetch) cannot possibly sustain a claim for past cultivation or domestication. Otherwise, why not consider the Cicer pinnatifidum (and likewise the dozens of taxa identified by Lev et al., 2005) remains from the Mousterian (60 000–50 000 years BP) layers of Kebara cave (therein) as evidence for past auditioning (or indeed cultivation) of such species as candidate crops à la Smith (2011). In addition, we argue that floristic observations (e.g. concerning L. nigricans, L. lamotii, or P. elatius and P. humile by the same token) cannot by themselves serve as evidence for past domestication events let alone for their (agronomic, genetic, or cultural) role in presumed independent domestication centres.

In our view, only clear identification of domesticated plant forms conforming to the classical diagnostic morphological criteria (i.e. non-brittle rachis) can attest to past domestication if such observations come from relevant archaeological contexts and if such morphotypes are not observed among the living wild relatives of the taxon under discussion. From a genetic perspective, DNA sequences obtained from ancient (charred or otherwise) plant remains which control domestication genes but which are absent in living relatives of this putative ancient crop would allow the granting of a ‘lost’ crop status. For example, study of the sequence of the brittle/non-brittle alleles of emmer wheat (A and B genomes) may allow the search for corresponding loci in the Triticum urartu genome (living wild and presumed ancient lost domesticated).

Our arguments in this paper concerning the ‘lost’ crops notion are made in the context of an ongoing debate on the geography of Near Eastern plant domestication (e.g. Abbo et al., 2010a ; Fuller et al., 2011, 2012b ; Heun et al., 2012). Therefore, we suggest that the above treatment of the presumed ‘lost domestication episodes’ will be considered in its wider context pertaining to all other inter-related aspects of Near Eastern plant domestication (e.g. Abbo et al., 2003, 2009, 2010b , 2011, 2012) and will assist students of the subject to distinguish between what could serve as mere working hypotheses or possible explanations and what can be considered as hard evidence for past domestication events.