Bigger beans- arriving in Italy and beyond

 An important contribution to evolutionary history, and breeding, of broad beans (Vicia faba) appeared in Vegetaion History and Archaeobotany recently, namely work by the Lecce archaeobotany laboratory on beans from southern Italy, by Grasso, Arthur and Fiorentino (2025). This study documents the appearance of two step changes in broad bean sizes in the Medieval period, backed up by measurement data on a small set of comparative land races, but only from Italy. This paper seems to miss the opportunity to think more broadly about the timing and processes of the evolutionism of gigantism in Vicia faba, which is just one of a group of pulses that evolved larger-seeded varieties in circum-Mediterranean agriculture- a pattern that Vavilov drew attention to under the heading "Regularities of type-forming processes" (Vavilov 1926). This provides an opprotunity to ponder what processes drove evolution of, and cultural preferences for, macro-seeded forms of pulses in this region but not in others, and it relates to the broader issue of the mechanisms that drive regional varietal evolution in crops, which can indicate remarkable parallel evolution across different crops in the same region.

While broad beans (Vicia faba) are among the Neolithic domesticates of the Near East, and part of the the Neolithic dispersal in Mediterranean Europe, the beans one is likely to be familiar with today-- at least in traditional agriculture in Europe, North Africa or Turkey, are generally quite different and much larger than those of prehistory. To be sure beans would have increased over the size of their wild progenitor (although the wild progenitor appears now to be extinct: see Kosterin 2015, Caracuta et al. 2016), but they were modest-sized cultivars, classed as variety minor. In Britain we typically refer to these as "Celtic breans", appearing in the Middle Bronze Age and becoming moderately common in southern England in Late Bronze Age and Iron Age times (as per the useful review of Treasure and Church 2017). In England such archaeological finds, charred, are consistenlty less than 8mm long, putting their uncharred dimeniosn likely under 1cm (assume -20% for charring). Although poorly undertsood some small-seed faba beans are also grown in part of the Indian Himalaya, referred to in the past as subsp. paucijuga, but with no real archaeobotanical indications yet of when they got there.

Larger beans in historical Italy

Somewhat larger beans in Southern Italy and Naples, that can be classed as variety equina, the "horsebean", appear as a minoriy of finds only in the late 7th century AD at the site of Colmitella (Grasso et al 2025). Such seeds have lengths consistely >1cm and typically near 1.5cm, and maximum widtth >1cm. After this equina types become more widespread in Southern Italy. By around this time equina type beans have also been reported from Panjakent in Uzbekistan, and although Grasso et al raises some questions about this, I would see this as also representing larger faba beans of the mid First MIllennium AD.

Even bigger beans, of variety major, with lengths around 2cm beginning to appear in Southern Italy in the 13th century. Grasso et al. argue that these steps of size increase may derive from breeding for larger beans, perhap in Southern Italy-- farmers selecting first for equina and then some centuries later for even larger major varieties. 

The overall conlusion of Grasso et al. is to emphasize the importance of archaeological seed measurments, as a tool to discern variation and evolution within crops over time, even long after domestication (when measurments are more commonly used and reported). This I fully concur with. However, I would raise some alternative hypotheses and unanswered questions about the drivers of seed size increase in Vicia faba, by noting that this is one among a group of peer species that have similar geographically restricted patterns of seed gigantism in the Mediterranean.

Mediterranean macrosperma pulses

As I already noted Vavilov drew attention to the occurrence of large seeded pulses in Mediterranean Europe and Turkey as a recurrent varietal feature.  This pattern is seen not just in Vicia faba, but also lentil (Lens culinaris), pea (Pisum sativum), chickpea (Cicer arietinum) and grasspea (Lathyrus sativus). The most stark contrast is between the small seed size and seed weight of varieties of these species found in India versus the larger sizes found in Spain and Italy. Smaller sizes also typically characterize these crops in Ethiopia, Afghanistan and Iran. Vavilov's data indicates that average seed dimeters of Italian lentils are two times those of India, with seed weights about 4 times greater. In chickpea, Indian types are typically ~30% shorter than those of Italy with as much of 70% less weight. Peas are about 50% lighter in India than Italy, and grasspeas too are also substantially smaller (Vavilov compares those from France, Ethiopia and India). Thus rather than seeing this as a product of targetted breeding of bigger beans by Medieval Italians, I think there is an issue of parallel evolution that demands explanation. The map at left/above from Muratova (1931), part of Vavilov's research team, shows that in the early 20th century large-seeded (major) faba beans are European and Mediterranean. Similarly the large seeded (macrosperma) varieties of lentil overlap this: the map of another part of Vavilov's team, Helena Barulina shows this (right/below). Grasspea, pea and chickpea similarly overlap. 

So what we have is a feature of geography and history that is influencing several crops in parallel. I would therefore suggest that instead of explaining this by intentional selective breeding we need to think about evolutionary factors, environmental and cultural, that might drive this. It would be nice to have comparable seed size data across all of these to look at species and see if the timing of stepped increases were similar. (In some of the cereals we do see linked regional patterns of change... but I'll leave that for another blog...see Dal Martello et al 2025). Perhaps, they might contend, in Italy (or elsewhere in the Mediterranean) breeders were targetting these pulses to select them bigger. (And if this is the case, is there a cultural factor to do with bean cooking methods than favours this?). But such a view has the implication that farmers in other parts of the world were somehow incapable of this, which I find this an unconvincing notion of European exceptionalism! 

Instead we might frame hypotheses about factors that favoured smaller seeds, for example in Afghanistan or India. Small seeds should require less water during the period of seed filling, and in crops such as this that mature in the months after winter that are hot and dry in places like India this could be adaptive. By contrast in the more gentle, showery spring around the Mediterranean water is not in such short supply. Smaller beans can be expected to cook faster too, so they might have lower wood fuel demands in areas where fuel is at a premium (across Iran and semi-arid India).  We can also think about other factors that promote larger seeds, such as increasing field disturbance and burial depth. It is only in the Mediterranean from late Roman times onwards that true ploughs, the mouldboard plouogh that turns over the soil, comes into use, whereas in India scratch ploughs (the ard) persist (even to today). I have previously speculated (Fuller 2007) that technological differences such as this might favour size increases under the deeper ploughing regimes. On top of this some farmer choice may well have operated too as an additional amplifier-- after all seed size multi-genic trait. 

I would note that there is another set of pulses and geographical context where gigantism can be noted, in East Asia. Here soybeans and adzuki beans also have macro-seed varieties much larger than the early domesticates that are well-documented in Neolithic or Bronze Age times. It also a part of the world were true ploughs that turn the soil come into use by the early centuries AD (Han Dynasty), spreading thereafter. Is this another context in which advances in tillage technology, perhaps coupled with irrigation and reduction of water stress, set in place conditions of seed gigantism to evolve in legumes. This seems to me to call for thinking about common causal drivers of unconscious selection that cut across cultures and crop species. Clearly more time series of archaeobotanical measurements are needed to make such comparative studies feasible.

Bibliography

Barulina, Helena (1930) Lentils of the U.S.S.R. and of other Countries. Bulletin of Applied Botany, of Genetics and Plant-Breeding 40. Leningrad [in Russian, with English summary]

Grasso, A. M., Arthur, P., & Fiorentino, G. (2025). Morphometrics shed new light on the first archaeobotanical evidence for the cultivation and breeding of Vicia faba var. equina (horse bean) and var. major (broad bean) in medieval southern Italy. Vegetation History and Archaeobotany, 34(6), 813-823.

Muratova, V. S. (1931) Common Beans (Vicia faba L.). Bulletin of Applied Botany, of Genetics and Plant-Breeding 50. Leningrad [in Russian, with English summary]

Vavilov, N. I. (1926) Centres of origin of cultivated plants. Bulletin of Applied Botany, of Genetics and Plant-Breeding  16(2) [Russian original], English Translation by Doris Love in Vaviliv (1992) Origin and Geography of Cultivated Plants. Cambridge University Press.

The earliest wheats of Ukraine (5400 BC)

The eastern areas of Europe and their transition to the steppe that lead to Central Asia remains one of the less well-studied regions archaeobotanically. The sparseness of reliable evidence has meant that the region is sometime discussed in terms of an alternative eastern source of crops from Europe, in addition to the main thrust from Anatolia through Greece and the Balkans, and it is sometimes mooted as a region of some crop origins, such as spelt wheat. New data is always welcome, especially of a high empirical calibre, from systematic sampling and backed up by AMS dating.

New data from the Ratniv-2 site in Western Ukraine, near the eastern frontiers of the Linear Pottery (LBK) culture, has been published by Motuzaite Matuzeviciute and Telizhenko in Archaeologia Lituana. This is an important record of early crops, and as the authors point out, it clearly points to similarities to the West and Southwest in Neolithic  Europe and suggest a spread towards Ukraine from the West in the Neolithic, rather than the east. The assemblage consists of wheats, barley, flax, lentil and pea. Two direct AMS radiocarbon dates on emmer wheat grains place these assemblages between 5400 and 5200 cal.BC. Of particular interest is that the wheats here include not just einkorn and emmer but apparently at some of the socalled "new type glume wheat," which these and other authors sometimes equate with Triticum timopheevi, and 20th century relict wheat found north of the Caucasus (western Georgia). That the archaeological "new type" has the AAGG genome of timopheevi remains unproven-- although I agree it is likely. It is perhaps more accurate to regard T. timopheevi as the relict remnant of what was a once a much more diverse and widespread species of wheat, which in all likelihood originated in the Anatolia and spread through many part of Europe and east through northern Iran in the Neolithic. I have sometimes offered the name "striate emmeroid" as a descriptive alternative to  "new type", as it is hard to think of something that has been largely extinct since the Bronze Age as new, and this wheat type has been in discussion by archaeobotanists for around 20 years...

In any case, what is notable about this assemblage is that is corresponds to those crops that are most common in the Neolithic of southwest Europe, supporting the ceramic and settlement evidence that attributes to the origins of agriculture in western Ukraine to spread from the west.

Dating archaeobotanical treasure troves from Armenia and Peru

The arrival of a new issue of Antiquity is always a welcomed event. The issue that arrived in my mail box this week a particular trove of treasures of an archaeobotanical sort. Not so much for archaeobotanical reports as such (the only such is the study by Willcox and  Strodeur on the details of Jerf el Ahmar), but for tantalizing new information of sites that have exceptional plant assemblages, or that one expects will in the future, like the Han era Pompeii, Sanyanzhuang (blogged below). Two sites, which are well-known for exceptional preservation are the Areni-1 cave complex in Aremenia (Areshian et al in Antiquity), left, and Huaca Prieta mound in coastal peru, below right (Dillehay et al.). 

     Areni made the news a couple of years ago for its early leather footwear, from ca. 3500 BC (Pinhasi et al in PLOSone), but it also has dessicated plant remains of many sorts: emmer and free-threshing wheat and barley, of course, but also lentil, grasspea, grapes, plums, walnuts, almonds and pears. How many of these fruit and nuts represent species that were available wild in the region, were under cultivation or introduced as cultivars still needs to be clarified. The site has produced features indicating on-site wine production, so presumably grapes were cultivated, from at least, from around 4000 BC. 4000 BC is associated with the earliest material reported so far, but the site still has much to yield to excavation, presumably including earlier material. Late samples from the past 2000 years include cotton and textiles.

     Huaca Prieta also boasts exceptional archaeobotanical preservation, and with a long sequence it provides information that suggests the chronology of cultivar introductions in this region. Few plants are likely to be native here and so their introductions point earlier cultivation and domestication elsewhere. This includes Cucurbita sqaush, avocado and lima bean at 7000-5500 BC, and thereafter the appearance of chillis and bottlegourds. Cotton cultivation was established around 4800 BC, and after 4500 BC maize was added to the repertoire. This is some of the best dated and documented early maize in South America, detailed of which were published earlier this year (see previous blog). Peanut, sweet potato and quinoa also come from later levels. Full details are not yet published but some summary can be found in the on-line supplement. This site has also produced coca leaves, indicating the long traditions of chewing this drug plant. Dillehay and colleagues reported the earliest for use of this drug, back to ca. 6000 BC, from elsewhere in northern Peru about a year and half ago, also in Antiquity. Intriguingly this drug plant, plausibly from across the Andes, appeared to already have domestication features at this date.

De-centering the fertile crescent

The Near Eastern "fertile crescent" is the classic centre of origin for domesticated plants. Although when James Breasted coined the term (1906) he was thinking about the beginnings of agrarian civilizations in Egypt and Mesopotamia. The term become subsequently linked to Gordon Childe's notion (1935) of the "Neolithic Revolution" and with Vavilov's "centres of origin" idea, and the Fertile Crescent became, in archaeological argot, the centre of agricultural origins par excellence. In another classic paradigm-setting paper Jack Harlan used the Fertile Crescent as representative of a "centre" of origin, a focused area in which a package of crops was domesticated together, opposed to his notion of a "non-centre" of diffuse origins of crops that were not-packaged and spread out in domestications in space and time, of which the sub-Saharan savannah was perhaps the classic Harlanian exemplar. Within "centres" the Neolithic was meant to be a revolution, domestication rapid, and hunter-gatherers transformed to village farmers in one process. However, the evidence for this has been gradually unravelling for the Near East, with domestication-- the evolution of domestication syndrome traits in crops-- taking place quite slowly (3000-4000 years), and taking place not in a focused area, as a package of crops, but spread out and with as many dead-end proto-domesticates as paths into the crops and farming systems we know for later prehistory or history. In other words, there was no Neolithic Revolution as such but evolutionary processes, in the slow Darwinian sense, in which many incremental changes and transformations only added up to something revolutionary in retrospect, after millennia (some 150 to 200 human generations) of small steps.

You can find the case for this made in the December World Archaeology in my joint paper with George Willcox and Robin Allaby "Cultivation and domestication had multiple origins: arguments against the core area hypothesis for the origins  of agriculture in the Near East" which is a companion piece with our in press paper in the Journal of Experimental Botany  "Early agricultural pathways: moving outside the ‘core area’ hypothesis in Southwest Asia". Both are responses precipitated by publication late last year by Abbo et al. of a re-iteration of the "core area" view of  a tightly focused area where a single package of "founder crops" was domesticated rapidly. Such a view requires the assumption that archaeobotanical evidence is false fossil record, a poor reflection of actual domestication processes, i.e. that the gradual changes that appear in systematic archaeobotanical evidence, are somehow misleading or mistaken. I find this hard to accept-- perhaps because I am a practicing archaeobotanist and regard the preserved grains and rachises of prehistoric crops as our most material record of what these plants were like and how they were used in the past-- but also because recent years have seen increasing sampling and sample size and it is the increase in these data, and the detail with which they have been studied, that most points to the more gradual evolutionary processes. In addition, the "core area" view delimits a smaller number of founder crops and sets aside (or even rejects implicitly) the presence of past cultivars and domesticates now extinct, from the Abu Hureyra rye, 2-grained einkorn, the "new type" emmer ('striate emmeroid'), or the Gilgal oats-- all species which are the product of careful efforts of archaeobotanists to document the material remains of past crops and not to assume that all crops that ever were must still exist today. These constitute the "lost crops" of the Fertile Crescent just as much as Iva annua is a lost crop of the American midwest.  Coupled with the genetics of known crops that support multiple "domestication pathways" (such as in barley, emmer, pea, probably one-grained einkorn), the Fertile Crescent as a whole was host to something like 20 domestication "events," only a fraction of which can be localized in any one sub-area of the Fertile Crescent or can be expected to be present in modern germplasm collections. With this number of domestications and their diffuse nature across the Near East, the Fertile Crescent as a whole starts to look like a Harlanian "non-centre". On the level of individual crops and domestication events there may well be many centres of origins but in terms of regions it look increasingly like all were non-centres. The closer one looks for a core centre, the blurrier it becomes. 

Last year I declared in General Anthropology, a paradigm shift in agricultural origins research. Perhaps rather pretentious, but it remains the case that domestication appears to be a slower process as we gather more evidence, and there is evidence for more places of domestication around the world (North America, South India, separate North and South China, various parts of Africa, New Guinea, to name a few). This year I have put my money where my mouth is, and brought out a number of contributions looking at domestication processes in the Near East in particular and in comparison to the best documented crops from elsewhere (mainly in the Old World). This includes attempting to objectively calculate rates of domestication in terms of phenotypic units, the darwin and haldane, reported in terms of more general conclusions-- that domestication was slow and not somehow special compared to other forms of evolution-- in the journal Evolution (Purugganan and Fuller), and unpacked with more consideration of the variation across crops in the Near East and elsewhere in Vegetation History and Archaeobotany (Fuller, Asouti, Purugganan). See also the updated rachis data of Tanno and Willcox (2011)phenotypic change may differ in adjacent geographical contexts, and in particular that there appear to be in island effects on Cyprus, where grain size change was sped up (Lucas, Colledge, Simmons & Fuller). A careful consideration of the hard evidence, such as the essay assaying the southern Levant (Asouti & Fuller), shows that even for the Fertile Crescent we still lack the evidence we need to be clear about domestication processes in particular micro-regions, early cultivation or when agriculture emerged (keeping in mind that cultivation, domestication and agriculture are really different things from among the many transitions that gradually came together). We also note that there is alot more work to be done on the species that were important wild food stuffs, which were abandoned as cereal agriculture took off, the small-seeded grasses and legumes, wild nuts and nutlets-- evidence in other new archaeobotanical datasets such as that from Jordan (el-Hemmeh) of White et al. (which reports important evidence for how barley was harvested prior to domestication) or 3 sites from the northern and eastern Fertile Crescent (SE Turkey and Iran) of Riehl et al.. Of course as the role of many wild foods along side pre-domesticated cereals gains recognition, the difficulty of being clear what was a likely weed as opposed to gathered become acute-- an issue discussed in a short paper by Willcox on early weeds. Another recent paper out of London (Wollstonecroft, Hroudova, Hillman & Fuller on Bolboschoenus) illustrates an example of challenges that still confront archaeobotanical identification, and the the potential ecological and dietary implications of refining these to species level, in this case for the sedge Bolboschoenus glaucus.

 Most of these papers, now available on-line for a forthcoming Vegetation History and Archaeobotany issue on the Near East which will illustrate the revised (and more diverse) understanding of the precursors of Fertile Crescent agriculture: the tides seem to have turned on the simpler core area paradigm!

Review on lentil domestication

A recent review article by Sonnante, Hammer and Pignone (2009) " From the cradle of agriculture a handful of lentils: History of domestication" in the rather obscure Rendiconti Lincei of April reviews the archaeology and genetics of Lentil domestication. It provides a useful overview, including a wide range of neutral genetic evidence that confers with the orthodoxy (e.g. of Zohary and Hopf) of a single Lentil domestication somewhere in the Levant, although the authors note the possibility that pre-domestication cultivation began with more than population of Lentils, but in the end only one was domesticated. They also provide a tabulation of the identified domestication-related genes in Lentil, which have generally received less attention than those in peas or Phaeseolus, and discussion of the morphological domestication syndrome in Lentils. They some speculation on why/how people came to cultivate lentils, which remains something of mystery for this and other pulses, in which seed germination rates of wild types are so low as to make them unlikely candidates for domestication. They concur the early Near Eastern agriculture including that of Lentils did not come from a dump-heap origin, but they still suggest that wild lentils might have occurred as weeds in early cereals and thus been a co-domesticate. Given what we know of wild lentil habitats this hypothesis seems little stronger than a dump-heap model, and I must favour some sort of intentional interest in Lentils and other pusles (perhaps for their protein content, or taste, or storability) which lead early cultivators to persist in their efforts despite initially low germination rates.