The drug before the calorie? Some hazy thinking on Cannabis domestication

Many have been excited this week about headlines claiming marijuana (Cannabis) was domestication in China 12,000 years (making it the first crop in East Asia). As the reputable journal Nature put it "pot farming first blossomed" in China 12,000 years ago. But was it so? How clear or otherwise in the evidence? Is it really farming? The study by Ren et al in Science Advances paper is important-- it represents the largest collection of Cannabis genomes sequences, it provides some important information on subpopulations and genes that have been selected for differently in fibre hemp from drug strains. However, I finds is dicussion and conclusions riddles with both imprecision (about chronology, geography and cultures) and inaccuracies. So what should be questioned?.

A major source of inaccuracy lurks in sampling and geographical representation. This is compounded by the fact that in many countries growing or collection or transporting of Cannabis is illegal (although legalization is on the rise). Thus traditional drug varieties have not been sampled across most of Central Asian countries, Afghanistan, Russia, Iran. It should go without saying that sampling in the modern time plane will miss past diversity that has been lost to ha bitat destruction and environmental change; such a problem, for example, has plagued some genetic studies of rice-- as wild population no longer exist in the regions were it was first cultivated (such that a modern genome map is not a map of origins). But even so the limited sampling across free-growing (feral/ wild) outside of South Asia and China) is notable. Compare the map of Ren et al (above) with that from the Cannabis book by Clarke and Merlin (2013), which highlight how much more wild/feral diversity there is out there across Asia and eastern most Europe, including 'wild' populations south of ot eh Caspian sea in Iran and along the Volga River in Russia. Even the map from Clarke and Merlin is incomplete with regards to Afghanistan where Vavilov collected apparently wild drug types of Cannabis (C. indiva var. afghanica) in the 1920s.

Given how few wild/feral samples that they have can they really rule out multiple domestications. Genetic analyses often err on the side of single origins. Simulation work (Allaby et al 2008 PNAS) have shown that this will be true even for crops with multiple origins, because gene flow among crops of different origins and pruning of lost branches (not sampled or not surviving to present). Their analytical methods are more modern and more sophisticated but I am not sure they can rule out multiple origins, and they certainly can’t rule out origins from regions not sampled or where wild populations are extirpated (e.g. Japan). Also wild populations (lets assume there are some in central Asia) can be heavily inundated with gene flow from crops over time make their original dinstictiveness hard to find in modern genomes.

A major issue is imprecision in dating. Genomic dervied estimates will never rival radiocarbon dating on archaeobotanical remains. Their proposed date of origins (is this domeatication?) has 6000 year error margins. And it not clear what they date! Was a divergence  between two wild population separated due to climatic vicariance at the end of the Pleistiocene or start of the Holocene, or is it meant to be the domestication bottleneck? The equivalent genomic dates for Asian rice domestication are ca. 18,000 BP and something close to 10,000 for African rice. Both of which are way off. African rice is domestication at more like 3000 BP and Asian rice at more like 8000 BP (of course it matters whether one is talking about the beginning or end of a process, as domestication takes 3000-4000 years in terms of morphological evolution/ genetic fixation.

By contrast archaeological dates are much more precise, at worst with 100-200 year error margins, but Ren at all quote these very imprecisely. It is as if they wish archaeology was less precise, but even then it would not approach the dating imprecission attached the the genomic dates. The say "~3000 BP" for the appearance of Cannabis in India, but if we are rounding off it is closer to 4000, as it occurs in the Late Harappan horizen (3900-3500 BC). Although Indian epics are not well dated some parts of them are from oral traditions that probably also date around then and make reference to Cannabis- I think Ren et al refer to this as ~2000 BP. Cannabis comes to India in my view as part of “Chinese horizon”, which is really just piece meal adoption of various things coming in via central Asia including crops and technologoies (harvest knives) from China, peaches, apricots, millets, japoonica rice. For discussions of this see, for example Fuller & Boivin (2009); Stevens et al (2016). Whether or not it arrived in South Asia earlier, or was already utilized from wild populations in the Himalayas, does not really make much differece to the whether or not cultivation began at the start of the Holocene in China. The imprecision in genetic dating, however, makes it about as likely that domestication took place around the start of the Yangshao period (~5000 BC), the period when we generally see the ending of the millennia long domestication process for the native China millets (

They are also quite imprecise about geography: are they suggesting a NW China (Xinjiang) or NE (Chifeng) source? And if either of these then discussion of early cord-marked pottery in South China (mostly south the Yangtze) is really not relevant, and yet they discuss this as though it indicates the use of hemp cords-- for which there is simply no evidence. 

Cannabis is undoubtedly an important crop brought into cultivation early in east Asia, esepcially for its medicinal and/or social uses, but becoming important for larger oily seeds and fibres over time. But in my view its development as a crop either parallels or is even inspired by the increasing importance of cultivation of other taxa, like the China millets. And this process could easily have played out multiple times-- perhaps in very different contexts in parts of central Asia or Jomon Japan, even amongst non-farming cultures. While genomic data will contribute to this, modelling such data really requires some calibration points in time and space, which will ultimately come from archaeobotany.

Chicken origins: closing in with new genomic evidence

wild Gallus gallus spadiceus

The past week saw the publication of a landmark genomic study on chickens (Wang et al 2020, Cell Research), which clarifies much about origins, and focuses some questions for further research. For a news summary see Lawler's Science piece.

It is transformative because it includes a substantial sample of genomes from across all of the wild subspecies of Red Jungle Fowl (142 wild red jungle fowls) and other wild Gallus species. The first thing to note is that is does support the reality of these different wild taxa. They aren’t merely geographical feral populations derived from escaped chickens, but they are differentiated from each other, making it reasonable to ask which population(s) are ancestral to domesticated chickens. In addition there has, of course, been gene flow via introgressions with domesticated chickens, but this has been on a more limited scale. So the answer to big question (of origins) appears to be Gallus gallus spadiceus. G. g. spadiceus is geographically focused on Burma, Yunnan, Guangxi, northern Thailand and bits of Laos. This struck me as the most surprising—this geographical derivation. If one favours a Chinese origins then you would look to G. g. jaboulliei (of the Guangdong and Fujian and perhaps further north in the past); if one favours an Indus domestication then one looks to G. g. murghii. Previously I have accepted the likelihood of an Indus Chicken domestication and a spread through India in post-Harappan times (e.g. Fuller 2006). This now appears unlikely. Instead it probably means that wild jungle fowls attracted attention in the Harappan period as pretty birds that were captured sometimes, traded, etc.,but not really domesticated subsistence species. Presumably the first Bronze Age Mesopotamian and Ramesside Egyptian “chickens” were actually pet wild jungle fowl-- fancy exotic birds-- and not connected to chickens as we understand them now. The "multi-colored birds of Meluhha" that were imported to Mesopotamia at the end of the Third Millennium BC from the Indus region, are plausible painted ivory statuettes of murghii jungle fowl (see, e.g. During-Caspers 1990).

These new genetic data also make it clear that as chickens spread out of their northern SE Asian homeland they did pickup some genetic material through introgression with local wild jungle fowl (such as G. g. murghii in northern India) and even grey jungle fowl in South India (the source of yellow legs: G. sonneratii). This process can be called “introgressive capture” and it is widespread in most livestock and many crops. This process has sometimes confused genetic studies into inferring multiple domestications, but with more genomic data it can now be disentangled (see Larson and Fuller 2014). 

It is also quite exciting that they have some genetic loci that might be under positive selection as part of the domestication process. One of the real mysteries with animal domestication is what constitutes domestication in a genetic sense in terms of adaptations. In plant it is well known that certain genes for seed dispersal, growth habit, dormancy, grain size, etc. were selected. We can find this evidence genetically and tie it to morphological changes in the archaeobotanical record. There is so far nothing equivalent in animals that links genetic loci to the morphological adaptations we see with animal domestication. So on a more theoretical level this may be the first step to actually starting to unravel the genetics of animal domestication.

These raises fascinating questions then about the contexts in which G. g spadiceus was domesticated—what kinds of human societies and agricultural economies did it interact with in its wild form and how did it get incorporated into ecology of human settlements. Equally at what period and in what contexts did these early chickens then spread. Their phylogenetic results suggest the first wave of chickens spread through SE Asia and SW China only.  Sadly we know little about the Neolithic in Myanmar, or Guangxi or Southern Yunnan; we do have some data from northern Yunnan where Chinese millet and rice agriculture (with pigs) arrives from the North around 2600 BC. One presumes there was some further Southern diffusion towards the China/Burma borderlands. And perhaps it was in these borderland zones where early sedentary rice/millet farmers began to isolate some G. g. spadiceus populations that came feed within the human settlement niche. As hypothesized in Larson and Fuller (2014), chickens likely followed a commensal pathway to domestication. But we now need to refine the map (right). And work out when this happened.

There estimate of the age of the last common ancestor of domesticated chickens and G. g. spadiceus 9500 BP (+/- 3000). But I would regard domestication any time between 10000 BC and 4500 BC as highly unlikely. As the authors themselves not in the first paragraph of their discussion such genetic estimates of domestication age tend to be over estimates (by upto 15,000 years!), so these are not exactly reliable.  In fact I would regard the tendency genetic coallesence ages as to tell us anyting about the timing of domestication to be a highly  misleading tradition that is entrenched in genetics but has little to back it up. Take the example of rice (Oryza sativa), where the genetic estimate of last common ancestor of cultivated rice and modern wild population is ca. 18,000 (Choi et al 2017). But archaeologically even the more generous estimates are ~10,000 (and more like 7,000-6500 by more cautious approaches). I suspect a more general problem is that what is being picked up the last major cladogenetic event that structured wild populations and not domestication itself. Often this can be expected to be something climatic, so 9500 BP is telling us something about how Early Holocene climatic changes—which restructured vegetation in big ways—restructured wild jungle fowl. Then it was one of these localized population that millennia later got domesticated. In all likelihood that localized population that was actually domesticated won’t exist anymore. It is also worth noting that the reality of domestication bottlenecks is itself somewhat dubious and is in the past year or two come to be questioned. Where ancient DNA is available (e.g. maize, sorghum, barley) it is demonstrable that no such bottleneck occurred and age estimates (see Allaby, Ware and Kistler 2019) that conceive some sort of a bottleneck may not be really telling up about domestication. 

Given what we know of the archaeology of SE Asia, one would tend think the initial  domestication and spread of chicken is unlikely earlier than the grain-based Neolithic that starts around 2500 BC (in southern bits of China) and reaches southern Thailand at 2000 BC. However, as far as I know there are no archaeological chicken finds at early sites. So I wonder whether the first spread of domesticated chicken might represent a secondary later spread perhaps closer to 1000 BC (the period when Bronze working spread southwards from China); this might also be the period when new crops spread like sticky rice. It may be that at that time chickens also spread rapidly via trade routes to India. I have long argued (e.g. Fuller 2007) that in South India the Dravidian linguistics suggest arrival of chickens after the South, South-Central and Central languages had fully diverged (which is something like 3000 years ago). Not long after this there are good chicken terminologies in Sanskrits and Prakrits from the 1^st Millennium BC, so it makes sense that chickens really only became established as livestock in India at around that time, and of course it is the later Iron Age when they first turn up in the west , such as the Hellenistic era evidence from the Levant (Perry-Gal et al 2015), or as an exotic animal in western Europe (Sykes 2012).

Citrus diversity in Roman Naples: pollen evidence for tangerines

The early history of Citrus fruits in the Mediterranean has been an active area of discussion amongst archaeobotanists, historians and palynologists in recent years. It has been well-established that citrons (Citrus medica) and lemons (Citrus limon) were known. For one thing they appear distinctively in Roman art, and for another their seeds are well known from Pompeii (e.g. Celant and Fiorentino 2018). What has been less clear is whether anything that we would call an orange today was known. Despite plausible textual sources, it is always hard to translate ancient terms into botanical species, especially in Citrus fruits that are so variety-rich, prone to both hybridization and frequent somatic mutations. So I have tended to think that some of the few orange-like seeds from Pompeii and from Rome, might just be outliers in the range of variation of early lemons (which are likely to have some pomelo and orange related ancestry in their South Asian origins (see, e.g. Fuller et al 2018). But an important new morphometric investigation of pollen of several Citrus species and archaeological pollen from Oplontis (near Pompeii) seems to have cleared this up. 

Lumaga et al. (2020) in a recent Vegetation History and Archaeobotany article, demonstrate the clear distinction in exine form, especially the cell or lumen size, that differentiates oranges (C. reticulata- the more primitive species, or less hybridized, of mandarin oranges or tangerines and similar small, sweet fruits). Much larger lumens characterize lemons. So I stand corrected on the Citrus diversity of Roman Italy: at least citrons, lemons and mandarins were grown.

This raises interesting questions about how these got to Rome. While these oranges were certainly known in China prior to the Han Dynasty, I have previously deduced that early orange in India-- known from Prakrit and Pali sources of the First Millennium BC-- naranga-- were perhaps most likely bitter oranges (Citrus aurantium), with sweet fruits coming later. Long distance transport of fruits from China to Rome strikes me as unlikely, so perhaps then there was more fruit diversity in northern Indian after all included under the rubrice of naranga, or other less obvious terms. Early South Dravidian languages (precursor to Old Tamil and Kannada) do seem to have two different kinds of oranges. This highlights all the more need for archaeobotanical investigations of early citrus (hidden in flots as charred rind fragments) throughout South Asia and Middle East in the Iron Age period.

I was recently interviewed about the history of Citrus fruits, especially oranges, in Europe in relation to how we can understand their use and symbolism in the writings of William Shakespeare: find the podcast here: That Shakespeare Life.

Structured diversity in tea does not mean multiple domestications

Genetic population structure of tea.

When and where were tea domesticated? And how many times? This is a question I get asked sometimes, and a paper published a few years ago was just brought to my attention that claims to have genetic evidence for three separate domestications, by Meegahakumbura et al. (2016) in PLOSone. I disagree. The paper convinced me of quite the opposite. It remains entirely likely that tea was domesticated only once in ancient times, even if there were widespread use of wild tea leaves across their wild range by people who encountered them-- but a key problem remains determining what that wild range was. The Meegahakumbara paper is an interesting one in terms of raising the question as to how much underlying geographical structure there is in the genetic diversity of tea, and how this relates to different cultural traditions of use, and which (if any) can be regarded to evolving into domesticated forms. But in the end there are severe limitations to their study. Despite some sampling from "wild" populations in India and China, as well as cultivars, there is a lack of samples from in between (northern Vietnam, Myanmar and presumably Laos, which are all presumably within the range of wild teas. What is really at issue is whether or not the large tree forms of tea are primarily wild (and recently managed or cultivated) or represents a distinct domesticated form, and this in turn raises the issues of what is "domestication" in tea. Domestication implies morphological adaptations, underpinned by genetics, which have been favoured through the propagation over generations by people. We know full well what this entails in cereals and seed crops (see, e.g. this Annals of Botany article); we can propose what is involved in tree fruits (which are harvested for fruits and presumably early on cultivated from seed: see "long and attenuated"); and we can also propose what is common about domestication in tubers and other vegecultural crops (see this recent Annals of botany article). But what about tea? Since tea can be propagated by seed or by cuttings it might share some similarities with either tree fruit or vegecultural domestications, but its use for leaf harvesting makes it quite different. Characteristics that might fit in the vegecultural domestication syndrome, include increased proportion of edible (leaf) to inedible (trunk) parts, and more asynchronous production of those. In developmental terms there appear to be selection for dwarfism-- smaller leaves, smaller stature, and more compact growth, which together with human management give teas their very trimmed hedge look (below).

Tea (C. sinensis sinensis) cultivation in Zhejiang
(Photo: DQ Fuller, 2004)

So what did Meegahakumbara et al. find? They demonstrated 3 distinct populations, and a 4th (lasiocalyx) that looks admixed between the Indian and Chinese tree teas (Figure at top). Truly domesticated tea (C. sinensis sinesis) looks to a be a strong clade. That is (in my view) the only really domesticated form. It has distinctive morphology (dwarfism).  It could already be distinct by the Han period, or at least under selection, since small, immature leaves are among the grave goods in the Yangling tomb of a Han Dynasty queen in Shaanxi, indicating trade from distance Yunnan already by 2200 years ago (see this Scientific Report). While it could be that tea was still a wild utilized plant, it might make sense that by this period experiments with cultivation had begun with an ancestral form of C. sinensis sinensis. It is also plausible that some cultivation developed in parallel with the distinct species C. taliensis, but its cultivation never spread beyond the Myanmar-Yunnan border regions (see, this paper on taliensis tea), so it is less relevant to the main domestication story of tea.

But is there any reason to assume that the distinct populations of large tea forms of tea, which they refer to as Assam tea (C. sinensis assamica). They divide this tall tree tea (see left) into two geographical groups (Indian assamica and Yunnan assamica, which leads to the somewhat oxymoronic terminology of "Chinese assam tea"). I do not see any reason in these data, or rationale provided, to not assume that Chinese tea is a domestication bottleneck from the Yunnan tree tea (assamica sensu  lato). The Assam and Yunnan populations of tree teas are distinct as one would expect of any geographically distant populations of a wild species. Trees of this assamica tea are also wild through large parts of northern Vietnam (see Zhao et al. 2017), and I would expect them also in Laos and parts of Myanmar.  In India these tree teas were used from the wild by a few local minority tribes prior to the British introduction of tea drinking and Chinese domesticated tea to India. But these were presumably only ever gathered wild. These Assam tree teas only came into cultivation (equivalent to pre-domestication cultivation) in the colonial era.  though they have now been cultivated by ~200 years, there seems no basis to assign them to domesticated status.

The Yunnan tree teas (What they call Chinese Assam tea) are presumably also native to the forests of parts of Yunnan, where they were also used by indigenous people. In this region their cultivation and management has continued while perhaps unmanaged populations went extinct in the wild. They could still be also the source of domesticated C. sinensis Chinese tea, or related to that source.

Tree (assamica) tea, leaves and fruits
(after a photo in the Hangzhou Tea museum)

The Cambod teas (C. lasiocalyx) look to me like part of the structured variation of a wild species that included geographical populations: i.e that the tea trees of Yunnan, India and adjacent Northern SE Asia (Vietnam, etc) borderlands should probably all be C. assamica (or  C., sinensis assamica, as taxonmically described by Zhao et al. 2017).  In their K=3 structure analysis these are grouped with Indian Assam wild tea trees, so it may be that lasiocalyx is just a northerly somewhat smaller variant of wild tea trees. It is plausible that domesticated Chinese tea came from this Cambod end of the wild range, i.e. further east, like SE Yunnan and is more distance from the Assam end of the geographical variation. It is a pity their is no geographical information on the assamica and lasiocalyx populations-- in terms of where and what habitats they come from. Although with a species like this it is not really surprising if wild populations have been completely lost. One is hard pressed to identify truly wild populations of some of the world's more popular nuts, whether chestnuts or walnuts, as these trees have been so heavily anthropogenically influenced in terms of where they grow.

So the origins of tea cultivation: southwest China, some place and time still to be determined. But the use of wild tea leaves by local peoples, widespread from Northeast India through Vietnam.

On the history of tea, see especially this magnum opus by Van Driem, reviewed here.

In Memoriam Alison Weisskopf (1960-2018)

Alison and Oryza nivara in
Orissa, Sept. 2010

Alison Weisskopf (1960-2018), passed away peacefully in hospice in the presence of her immediate family on 11 January 2018. She was a beloved colleagues at the Institute of Archaeology, a fixture in the archaeobotany laboratory for many years and a leading figure in archaeological phytolith research, respected globally. Her research legacy is substantial as her work takes a distinctively ecological assemblage approach to reconstructing rice cultivation ecology as well as crop processing. This has proved innovative and has proved fruitful, and can be expected to continue to inspire further research and agricultural ecology approaches to phytoliths around the world. Despite first being diagnosed with late stage cancer in 2010, she soldiered on was at her most productive as a researcher over the past half dozen years, which is readily evident from her publications list. 

Bangladesh, Nov. 2013: ethnobotany

She has made lasting empirical contributions on archaeological research in China, Southeast Asia (Vietnam, Thailand, Cambodia), and South Asia (Bangladesh, Sri Lanka, India). Through ethnobotanical fieldwork (in India, Thailand, Laos) and archaeological projects (in China, Bangladesh, Fiji), many further collaborations she was a key colleague in many international networks and she leaves behind many friends around the world.

Alison joined UCL as a BSc Archaeology student in 2000/01, essentially a career reboot as a mid-life adult. She demonstrated a strong affinity for environmental archaeology and archaeobotany from the beginnings of her studies. She took my “Plants and Archaeology” in 2001/02, and a new course on “Origins of Agriculture” the following year. Her BSc dissertation on phytoliths (“A study of the phytoliths from the late Bronze Age site of Krasnoe Smarskoe, Samara Valley, Russia, and the information they provide on agro pastoral economies and environments”) supervised by Dr. Arlene Rosen was passed with distinction in 2003. In receipt of a AHRC scholarship, she continued her studies in the MSc Palaeoecology of Human societies, with a dissertation on “An investigation of the Neolithic ash mound and settlement at Sanganakallu in the south Deccan, India, using phytoliths and macro-archaeobotanical material”, combined analyses of plant macro-remains and phytoliths and received a distinction in 2005.

Liu River, near Huizui, Henan, China, 2006

She began her PhD in 2005, again funded through an AHRC studentship. She submitted her PhD thesis, Vegetation, agriculture and social change in Neolithic north central China, a phytolith study, in 2009 and was awarded her doctorate in 2010. Her doctoral research took her on field to China several times, such as to the sites of Huizui and Xipo, where she worked alongside colleagues including Arlene Rosen (now University of Texas at Austin), Gyoung-Ah Lee (University of Oregon) and Liu Li (Stanford University). Her PhD represents years of dedicated laboratory work. She later published a revised version of her PhD as a monograph in 2014.

Sept 2010: Sampling Oryza rufipogin in Orissa, with
Rabi Mohanty and Mukund Kajale

In 2009 she took up a post-doctoral research associate position funded as part of a NERC project  'The Identification of Rice in Prehistory' (2009-2012), which came to be dubbed the Early Rice Project, and spawned follow on research projects, including 'The Impact of Evolving of Rice Systems from China to Southeast Asia' (2013-2016), and 'The impact of intensification and de-intensification of Asian rice production: transitions between wet and dry ecologies' (2016-2019). During a intermission between the first and second NERC projects she secured funding through a British Academy small grant to explore comparisons between phytoliths and diatoms in rice paddy soils, and she received a travel grant from the Thai Ambassador to the UK for ethnobotanical fieldwork on non-rice plant use in Thailand. Her research, and her development of phytolith approaches to rice cultivation ecology was central to these projects and their success. This sent Alison into the field to study modern rice ecologies, both cultivated and wild, in far flung parts of Asia, from central China to Laos and the highlands of northern Thailand, through Bangladesh and Assam, remote parts of Odisha state in India, and the Western Ghats mountains along western coast of India. Her unique experience and expertise has meant that she attracted archaeological collaborations and samples for analysis from an even wider range of countries. She authored 29 academic papers or book chapters, in addition to 1 monograph, with many more still in the pipeline. For a list her published academic papers and chapters: see here.

While many have approached phytoliths typologically and metrically to attempt to look at morphological differentiation between domesticated and wild rice (e.g. bulliforms or double-peaks), Alison’s innovation was to focus on the plant communities that occurred with rice and were sampled in harvests, sub-sampled in crop-processing and ended up to systematically recorded, quantified and discriminated in the micro samples from archaeological sediments. In her fieldwork and analyses, her focus on plant communities and how human communities intersected these is evident. It offers a legacy for phytolith archaeology.

Ethnobotanical fieldwork in Thailand,
Nov. 2012: with Katie Manning.

Alison, herself was a key node in our community. Having worked in the archaeobotany lab as a post-graduate student and post-doctoral staff member for some 15 years, she was often the focus of discussions, both of science and of social life. She has also trained and supported numerous students, offered countless cups of tea, words of encouragement, and a warm sense of humour. She is warmly remembered.

I invite comments to be posted to this blog by those who knew and miss here. And I append below various photos of Alison in action.

Gyoung-Ah Lee and Alison on the Liu river, Henan, China (2006)

Alison collecting rice weeds in Bangladesh, Nov. 2013.

Nov 2011: Northern Thailand: Cristina Castillo (Left) and ALISON (right) with Karen rice farmers in Northern Thailand

Ellie Kingwell-Banham and ALISON WEISSKOPF in Maharashtra, India (Sept. 2010)

14 July 2004, IoA foyer on lab botanical shirt day: Phil Austin, Emma Harvey, Meriel McClatchie, Jon Digby, ALISON WEISSKOPF, Emma Jenkins. Alison was an MSc student at the time, and was apparnelty the original source of the idea for this day.  Below a full photos of the whole lab group.

Dorian, ALISON, and Deepika Tripathi at the IWGP in Thessaloniki (2014)

Indo-Pacific Prehistory Association conference, Siem Reap, Jan. 2014. Participants in session on "Foraging and Farming". Alison fifth from Left.

Early Rice and Its Weed Flora, Symposium at Peking University May 2011

Using big machines to look at the finer aspects of seeds

This year has seen three studies on high resolution x-ray computed tomography applied to archaeobotany, one using ct-scanning to recovered chaff hidden in ceramics (see Finding Rice Domestication in Clay), and two using a synchrotron to peer inside seeds, including soybeans and horsegram.

This past summer, I published with colleague Charlene Murphy, a Scientific Reports article on domestication of the Indian crop horsegram. While this article represents an important contribution on the domestication history of a major crop in India, and evidence for evolution of morphological change during that crops domestication in South India (see also our GRCE paper, reviewing all that is known about horsegram origins), this is really more significant for the methodological contribution to the archaeobotanical documentation of domestication. We were able to put our small archaeological seeds in a very large machine, the Diamond Light synchrotron (shown at left). which allowed us to non-destructively capture the the internal structure of the entire seed (not as straightfoward as it sounds as it takes a lot of computing time). And from this we could measure seed coat thickness on any of the 1000s of cross-section slices through our seeds (like that below/right)

One of the well-known domestication syndrome traits in pulses is the thinning of the seed coat, tied to loss of germination inhibition. But it has been difficult to document this archaeologically. Seed coats are often destroyed in charring, but even if preserved they study on charred seeds would require destructive breaking of seeds. And even if damaged, it might only be possible to document the seed coat thickness in one or two places with an SEM or high powered normal microscope. As a result this has been rarely documented, which has lead to a fair degree of speculation on the evolution of thin-seedcoat, readily germinating pulses, as the result of conscious selection of the readymade mutants in the wild (although none have been documented in the present day)-- the domestication before cultivation hypothesis applied to lentils-- or positing a rapid conscious selection by those who initiated cultivation-- lets call this the pea breeding before agriculture hypothesis. The truth appears to be, however, a gradual evolutionary process as seed coats thinned over time, much like the evolution of increasing seed size or the non-shattering in cereals-- at least in horsegram. This can be seen in the chart below showing the thinning seed coat along side a trend in seed size increase in horsegram. Further work is needed on additional pulses to see if this pans out as typical of the pulses domestication processes, or whether there was variation, or indeed any cases of plucking domesticated types from the wild-- of which I am doubtful. At least now we have a method for approaching this.

This is actually, quite logical: established stands of pulses could be maintained and wild-type dormant seeds would constitute an established seed, and would recurrently add new plants to the the stand over a series of years. But due to annual human harvests mutations that reduced dormancy would get selected, and would be particularly important for any new populations planted in areas without existing wild populations. In this context we can expect the gradual evolution for thinner coated, more easily germinating seeds through selection across what are presumably multiple loci, as is evident in our archaeological horsegram data (shown left).

Soybean oil content in charred seeds?
The claim for earliest use of a synchrotron to look at charred archaeological pulse seeds, however, goes to our colleagues in China, in collaboration with Prof. Gary Carwford, Shandong archaeobotanist Xuexiang Chen. They argue that soybean underwent selection for increased oil content in prehistory during domestication-- undoubtedly true-- and that this can be tracked archaeological through a change in the number and size of pores visible on the inside of charred soybeans viewed through the synchrotron and High-Resolution Computed Tomography. I remain unconvinced on this last point, and although the paper reports on examination of modern soybeans and, other oily crop seeds, and experimentally charred seeds none of these are illustrated or really described so as to support this interpretation. The authors infer that more small pore is a product of more oil whereas large pores represent burned out protein, but is this true. The differences look to me more like artefacts of carbonization processes, and not a good proxy for the internal anatomy of the original uncharred soybeans. As the few illustrated example suggest larger and irregular pore are present in seeds with more distorted external surface anatomy (e.g. c), whereas small pores are more evident in better preserved examples (e.g. f).

Unfortunately, the central claim in this paper does not really add up, or at least are not well justified and explained in the text. This makes me very nervous about accepting the main conclusion of the paper, i.e. that the authors have demonstrated an increase in oil content in soybean during domestication by measuring the quantity of bubbles (voids) of different sizes in charred archaeological soybeans. Small voids are attributed to oil content and large voids to protein—but this is never demonstrated (for example in modern and experimental charred examples) or backed up by citations on soybean anatomy, as to why these voids should differ between oil and protein. That soybeans are oily, in contrast to most pulses is clear, but this also has major implications for the nature of archaeological finds. Most carbonized archaeological soybean are poorly preserved, distorted, full of large voids and small voids and very shiny on their interior. This is contrast to pretty much every other pulse I have seen archaeobotanically, from Vigna spp. to lentils and peas to Lablab. Even in the Chinese samples, presumably subjected to similar formation processes Vigna angularis seed present typical features of carbonized pulses, including a dense charred matrix with distinct cotyledons. In Glycine cotyledons are rarely evident and their interiors are heavily distorted by voids and bubbles. The obvious deduction is that this state of things is the result of the oil content in soybeans, and of course many other oily seeds, from cotton to sesame, also tend to show similar levels of bubbling and porosity when charred. If large voids in soybean are due to protein burning up during carbonization then surely one would expect to see this in any pulse, all of which have at least 20% protein content. It is true that the oil in soybean is contained in fresh seeds in many small droplets/sacs but upon charring things are likely to end up being very different. Oils are going to burn to more readily to gas than carbohydrates or proteins and thus create more bubbles and explosions of expanding gas. As this progresses and cracks to the outside of the seed allow penetration of gas (and some oxygen) from the exterior, one would expect this to speed up. The persistence of small voids then might be predicted to be the result of less oxidation, less temperature and perhaps other variables of charring conditions of a given seed. Cracking and penetration of gases into the charring seed may indeed be affected by aspects of domestication—thinning of seed coat, increase in seed size. Indeed, larger seeds seem likey to leave larger parts of their interior cotyledons unexposed to exterior cracks and oxygen; and in this context would be expected to preserved more small oil bubbles as a side effect of seed volume increase: i.e. the difference over time would reflect preservation artefacts rather than selection for genetic change. It is hard to see how at this stage we can deduce difference in underlying phenotype and genetics from this sort of data—at least until we have much better grasp on who charring conditions affect the distribution of seed contents, and this calls for some systematic experiments.

Undoubtedly soybeans were selected for oil content, but when and how this took place in relation to other domestication traits remains sadly unclear. I find I have to reject to conclusions of Zong et al., although their paper doe illustrate the potential analytical power of using a synchrotron to peer inside archaeological seeds

Early Rice Project symposium last week

Last week we hosted in London a symposium for the Early Rice Project, 

Investigating the evolution and impact of rice cultivation through the later prehistory of monsoon Asia. We brought in colleagues and collaborators on the archaeology of India, Southeast Asia and China, from countries across several continents, and had a success full exchange, not just on the archaeobotany of the region and new data (much of it generated at UCL through our NERC and ERC projects), but also on the stories of domesaticated fauna, our current understanding of Neolithic spread processes, Mesolithic persistence, demographic growth and the emergence of complex societies and irrigation. What is clear is that there is much new to say about rice, when it first arrived in several regions of monsoon Asia, and as it was transformed into the cornerstone species in the subsistence base of large complex societies. Nevertheless the meeting highlighted also the major gaps in empirical evidence, both geographically and chronologically. We hope to be able to pull this together for publication to further broaden out our dialogue on what we know and what we need to know. There has certainly been a rapid increase in data as the chart (below) of published, or recently counted archaeological spikelet bases indicates (from my introduction presentation)..

Some recent outputs from the Early Rice Project include publication of ancient DNA from charred rice grains from sites in Thailand and India (Castillo, Tanaka et al.), which add some flesh on the skeleton of the Proto-indica hypothesis; and publication of the first of a new generation spatial modelling of the early geography of rice, this one aimed at deducing the most like region (or regions) from which rice originated and spread, in particular the originals of early japonica rice that was so important to the Neolithic developments in China and throughout Southeast Asia (Silva et al. in PLOSone).  See also, the paper on phytoliths as a reflection of weed flora (Weisskopf et al 2014), the first of several in the pipeline that will illustrate new and more robust approaches to determining past rice ecology.

Unravelling agricultural packages

Two recent studies, one for the west and and one for the east, illustrate how crop packages unravel and become less diverse as they spread. The spread of agriculture is so often presented as a processing of unfolding, like a blanket being stretched from the point of origin outwards. This is especially true of the spread of Near Eastern agriculture, a truly diversified crop package of cereals (multiple kinds of wheat and barley, pulses, flax, plus livestock). But when the spread of agriculture is examined in detail, it is clear that crop species and varieties drop out along the way, and those which do make it probably become less genetically diverse. A recent database analysis of Neolithic Ireland illustrates the extreme western edge of Neolithic dispersal from western Asia. Published by Meriel McClatchie (whose PhD hails from here at UCL) and various collaborators (including UCL colleague, Sue Colledge), has been published in Journal of Archaeological Science, "Neolithic farming in north-western Europe: archaeobotanical evidence from Ireland" . This study demonstrates the clear pattern of quantitative reduction in most crops in Neolithic Ireland compared with elsewhere in Europe. Emmer wheat, virtually no einkorn (and one has to ask how securely identified any einkorn was), naked barley and a bit of flax-- that pretty much sums up Neolithic Ireland, in contrast to the 8 "founder crops" that are meant to characterize the start of agricultural dispersal from the Near East. 

A similar barley and wheat (with a dash of flax) characterizes the UK early Neolithic, which like Ireland see the dramatic introduction of cereal farming shortly after 4000 BC. As recently suggested in the study of radiocarbon dates from the UK (Stevens and Fuller in Antiquity, Sept 2012). One looks forward to further Irish Analyses to see whether the collapse of Neolithic cereal farming that Chris Stevens and I see in the England and Scotland cereal data also held true in Ireland.

This parallels what we see in the East, in India for example, which has recently been mapped in the paper I co-wrote with Nicole Boivin and Alison Crowther, "Old World Globalization and the Colombian Exchange: comparions and contrast." In South Asia wheats (including glume and free-threshing), barley, several pulses and flax, all seem to be important on the Indus Valley, but this package becomes less frequent and less stable as one moves into "inner" India. Sure enough wheat and barley make it both eastwards to Bihar and south to Karnataka, but generally with a strong preference for barley few or no pulses. In China only select wheat, and rarely barley, makes any showing at all, and there wheat is quantitatively negligible. This highlights that in some cases the caloric and subsistence needs are not likely to be served by the introduced cereals from the Near East. Some years ago I made the case (Antiquity 2005) that wheat and barley in Southern India might also have been status crops, used perhaps for beer, rather than as staples. One can ask the question as to what extent some the westernmost spread of cereals in Europe was as much about preferred foodstuffs rather than subsistence necessity when wild sources like hazelnuts were still so readily used and available?

There are broad similarities but also differences in the outward spread of crops from the Fertile Crescent. While in India and China this spread is seen largely in terms of the adoption of crops by local populations, in western Europe there is evidence for a greater role of migration. While in India we tend to attribute this to the local importance of other crops, Brachiaria ramosa and mungbean in the south or rice in the Ganges, that was clearly not the case in Ireland. So I wonder if we are seeing both the effects of crossing ecological frontiers, perhaps quicker than some crops can adapt, or beyond which some crops just can not adapt. Northern Europe certainly presented great challenges to agriculture, highlighted in its extreme margins such as Norway, but also in Britain by the apparent abandonment of cereals in the later Neolithic, perhaps as temperature retreated somewhat (Stevens and Fuller 2012). Monsoon Asia was not the most suited to the Near Eastern crops either, which also points towards social rather than caloric drivers in crop spread. In another parallel with distant Britain the agriculture and sedentism in parts of the Deccan, most clearly in western Maharashtra, where wheat and barley were quite prominent, appear to have collapsed and possible were abandoned over a wide area (in this case around 1200-1000 BC at the end of the Jorwe period).

Both of these studies show the importance of larger regional datasets, in which broad patterns are often visible even with simple quantification. This broad patterns raise questions that in turn call for more intensive sampling and local studies to work out wheat is actually happening at the periods of intial adoption or abandonment. What is missing currently is more usable data from the middle, Central Asia, the Iranian plateau, etc., so that archaeobotanical databases can become truly continental across all of Eurasia.

South Indian aridification press release

The Woods Hole Oceanographic institute has put out a press release on the palaeoclimatic data for India aridification over the late Holocene, based on the GRL paper published a few weeks ago and blogged previously.

Bay of Bengal ardification data and South Indian agricultural adaptation

A new article, out this week in Geophysical Research Letters, "Holocene aridification of India", by, Ponton, Giosan, an others, presents important new, and quite high resolution, data on past monsoon dynamics and vegetation of peninsular India spanning the whole Holocene. This research, lead by researchers from the Woods Hole Oceanographic Institute, analyzed evidence from a Bay of Bengal sediment core, which captures discharges from the large Godavari river system. The core data comes from carbon isotopes of leaf waxes, reflecting the amount of arid-adapted/ savannah vegetation in the Godavari catchment, and oxygen isotopes from a marine microfossil that record salinity. This points to a general aridification trend over the course of the middle and late Holocene, supporting what we already would infer from pollen data in Rajasthan or monsoon proxies in the Arabian Sea, but this time providing more direct evidence from South India. My own involvement in this work came in the form of trying to think about how this might be correlated with archaeological evidence for settlement, agriculture and population in South India-- where the archaeological record suggests increasing sedentism, population and agriculture in response to, or despite, aridification, a contrast from the Indus region for example where the long-term trend of population depletion as aridification proceeded. This suggests long term cultural adapatation processes to aridification in peninsular Indian agricultural practices.

To quote from part of our conclusion: "The significant aridification recorded after ca. 4,000 years ago may have spurred the widespread adoption of sedentary agriculture in central and south India capable of providing surplus food in a less secure hydroclimate. Archaeological site numbers and the summed probability distributions of calibrated radiocarbon dates from archaeological sites, which serve as proxies of agricultural population, increase markedly after 4,000 BP in peninsular India [discussed in detail in the electronic supplementary text]...In contrast, the same process of drying elicited the opposite response in the already arid northwestern region of the subcontinent along the Indus River. From 3,900 to 3,200 years BP, the urban Harappan civilization entered a phase of protracted collapse. Late Harrapan rural settlements became instead more numerous in the rainier regions at the foothills of the Himalaya and in the Ganges watershed."  Most of the archaeological information is summarized in the electronic supplement, Section 4., and included an attempt to sum Neolithic/Chalcolithihc radiocarbon dates (as limited as they are) and to tally known site numbers through the Iron Age. 

This work complements recent sedimentary studies of the Indus river system, such as the Clift et al Geology paper, blogged earlier.

Expanding Indus fibre crops

Two recent articles in Archaeological and Anthropological Sciences, report new evidence for species used in Harappan fibre work. Rita Wright and colleagues have reported evidence for jute textiles (Cochorus capsularis), based on analysis of fibre impressions preserved in ceramics. As with work done on plant impressions in pottery, this demonstrates that quite fine detail can be preserved in impressions, recorded in casts and studied with SEM. While we have perhaps long suspected jute, which is native to South Asia, was grown in the Indus period, seed finds from sites such as Rojdi (Weber 1991) were ambiguous as whether this species was cultivated, and processed for fibres.  Processing  involves retting (rotting in water), something I encountered a few years ago in the Son valley. My photo below shows a stack of harvested jute which is about to be weighed down with stones (visible in the water behind) for a week or two, before it is pounded to remove the fibres.

Posing for a photo with recently harvested jut that is about to retted  in the side channel of the Son river behind.

 Finds of textiles from eastern Iran published a few years ago by Irene Good, in the book  Ancient Textiles: Production, Craft and Society, included a couple examples of jute, as well as many of sunn hemp (Crotalaria juncea), which were also presumed to have been imported from the Indus to the east. Unfortunately as a small-seeded legume, recognizing the presence of sunn hemp in seed assemblages, especially as this crop and not a related weedy species, is not yet really possible, and could prove intractable. Taken together with evidence for flax seeds, and cotton in the Indus Valley [see my 2008 review pdf], as well as wild silk production (from the Assam silk moth), reported by Good & al. from Harappa in Archaeometry 2009, the Harappan civilization was quite the centre of textile crop diversity in the Bronze Age (compared to apparently only flax cultivation in contemporary Egypt or Mesopotamia). This adds weight to the notion that Indus exports, including those of the textually known Meluhha merchants of the Persian Gulf, included a range of cloth types.

Wild fibre sources were also important, and spun and woven net material from Shahi Tump, has also been reported recently by Thomas, Tengberg et al. in AASc. In this case, they appear to be using the local dwarf Mazari palms (Nannarrhops ritchiana). One of the excellent components of the paper is comparative study of palm phytoliths (admittedly of the limited taxa range that might be found in Pakistan), but which shows clearly that there is significant and taxonomically-informative variation in the spikey silica balls that palms produce. The next challenge will be doing more work on this variation and it taxonomic interpretation in the palm-rich tropics.