 I'm very pleased to be here and have a chance to talk about this theme and so I will launch right into it. Okay, so I'm going to do this in six parts whether or not this number seven parts of the knowledge not all of equal length so because some of us are archaeologists and archaeologists and some of us are not. I thought I'd briefly introduced what is that I'm going to talk about progress in the study of plant domestication through our chemical evidence what I would call conventional methods, because things have advanced a lot in the last couple of decades, and so get a sense of where we are and then I'm going to talk about the some of the new lines of evidence that we can generate using these new techniques like high resolution x-ray computed tomography, including a bit of work that I was involved with on it using a synchrotron in the UK, and with colleagues in Australia using micro CT scanning so I'll talk to those case studies. So I'm not going to, I don't know why it says five minutes of insight I'm not going to talk about that. And then I'm going to talk about new developments in the study of cooking and cooked foods, which so far we've used through research, which have potential using three dimensional imaging methods, and then then I'll conclude with just some sort of random thoughts on future directions, and then hopefully we'll have time for discussion. So yeah, six parts, not seven or the next extra one. So what is archaeopathy so archaeology, as anyone who's an archaeologist will know, and most of you will also know is the escalation and study of past sites of human activity. And I'm studying all the finds and artifacts and lines of evidence that come out of that. Archaeopathy is, is particularly I would say a subset of archaeology that uses botanical methods to retrieve and study the evidence for plants and plant use on those sites. Now most often we deal with material that's preserved by being charred in the past so that differentially destroys some evidence that means we're left with woody parts and hard, hard seeds and nut shells and hard bits of chaff and things like that. So we recover that material by concentrating it because it's dispersed as small particles in the sediment, and we concentrate that through flotation so mixing it through water and using buoyancy. And so on this slide as pictures of a machine flotation system at Chetlpiaq and Neal of Excite and Turkey I worked on for a number of years, and when I was much younger doing bucket flotation and Orissa in Eastern India. So it's a kind of very low tech method of concentrating the charred material catching on fine sieves. Plant macromains tend to go down to about 250 micron so they're quite large compared to what people using X-ray imaging usually look at. But that of course provides the potential to look inside that material. And there's a nice example of some charred rice spike that basis about 200 rice spike that basis or something like that and a single rice grain. And you know what the size of a grain of rice is because you probably eaten it. That's on the right of that lower photograph and the rest of the spike that basis which attached the rice to the plant. And the spike that basis for those of us interested in domestication are kind of gold dust because they tell you whether the plant is wild, has a wild seed dispersal mechanism, or is domesticated in the sense of being dependent on humans to harvest it and plant it. So that's a key change that we're interested in studying. So some of the kinds of things we can use this to study by no means exhaustive. I think the most common things that lines of research that I've been involved in this tells us about what people ate. So what plant foods are on those archaeological sites are they fruits and nuts are they cereals are they tubers. Are they flax seeds, what combination they were. Are they farming so are they cult are these wild food they collected or they cultivating cereals and flax and legumes rice to bring out part of the world you're in. And then we can also say things about how they were farming. So we often use our simply to say well, they're irrigating or maneuvering these fields or they're not. So rain fed agriculture or irrigated agriculture these kinds of questions are very common for our give on this to them. But because we have the remains of the plants themselves we can also ask questions about the evolution of crops over the long term or the evolution of the week floor. And so I'm going to talk a lot about that last but the evolution of crops in this talk I've been quite interested in that in recent years. So plant domestication and the expanding archaeological record. When I was a PhD student I finished my PhD in my in 2000 really early 2000. We sort of thought there are a few centers where plants are domesticated domestication happened very few times it happened very rapidly. It was a non question you know you could have said the archaeological question was let's get plant remains. Do we have domesticated crops and agriculture and was a yes no tick the box, or not. You know, I would say in the last two decades we've come to realize that of course it's an evolutionary process that takes thousands of years so that's thousands of plant generation. Hundreds of human generations. And things are changing so there's a lot to unpick that box of domestication and understand it as a protractive process to compare across different crops, compared across different cultural contexts and different geographies. So essentially we can do that by looking at some of the morphological differences that are obvious between a modern domesticated plant and it's closest while relatives will be often referred to as well for dentures. And that's things that I've already referred to so things like seed dispersal so in this picture on the left, you have a series of barley spikelets in my wild barley spikelets in my hand so this is kind of freshly picked off the wild barley plants and in Iraqi Kurdistan. You can see that each of the spikelet and the single grain inside, and the ones that are sticking in the soil there in the middle are, they fall into the soil they're shaped like little projectile points little arrowheads, and they're weighted so that if you drop them, they will always fall point down, and they're covered with little microscopic hairy barbs made of silica that allow them to dig their way into the soil, and given time they literally will dig their way into the soil because the wind will blow them animals will vibrate the ground, and they will drill them down over a period of days to weeks. And so they're designed for getting shattered off the plant when they're mature and sticking themselves in the soil. On the right is domesticated barley, you can see that it's all dry, it's all held together it's all still on the plant. It is sitting there waiting for a human being to come along and harvest it, and to thresh it to break the grains apart, and then plant those grains so it's reproductive cycle has become dependent on the intervention of people. It's a hard domestication trade, which is typical of most cereals and to varying degrees, there's a certain amount of spectrum of all domesticated. So they've become increasingly dependent on humans for their reproductive cycle. That of course makes them good to harvest you get a higher yield, if you're a person. It's good from the point of view of the plant in that they can then people spread them about to places where they're while the janitors wouldn't go. And so it's a kind of code evolutionary process. Now we can study that through some of the morphological changes that result that come about as part of that process which of course are underpinned by genetic changes. Another thing that changes with domesticated plants is seed size so wild relatives and wild plants tend to have smaller seeds and domesticated plants tend to have larger seeds and the lower left of that. Along the bottom wild barley wild, just the middle and wild melon and wild pea, and at the top some domesticated varieties, they're not not necessarily bigger in all the dimensions that they're fatter. And basically we think that's because it's driven by kind of level playing field effect you clear the soil, I tell you you put seeds in it, and the competition is all amongst those seeds of the same species. Which means the ones that have bigger seeds, more quickly established larger seedlings, they capture more soil to where the nutrients are in the water is and they capture more leaf area more sunlight and they are proud out their siblings. And so there's a selection process for larger seeds. Now of course we can study by simply measuring them seeds get bigger over time also better data. Then we know from modern studies that there's also things in some plants that are, which involve physical changes related to dormancy. So many plants are taking a wild, wild pea like that one down here it's black it's also got a thick ornamented seed coat. And that seed coat is hard and doesn't allow water to penetrate the seed. And so it doesn't allow it to germinate but it doesn't get me water so it will sit wild peas will sit in the soil for five years 10 years 15 years before they germinate. The domesticated one has a thin seed coat permeable, and so it will germinate as soon as it gets wet so people were germinated beans, you know as a school experiment at home, you stick them in a dish or water they germinate within a week. And that comes about through physically through a thinning of the seed coat, which allows that permeability it also removes various chemicals usually involves a change of color. Of course archaeologically we can't see color, because the seeds are generally all charged a little black, but this just shows some SCM through a section of a soybean so a domesticated or wild to the outer seat. So I mean, it's very very thick, and the kind of all your own layer with those palisade palisade cells is very thin and then the, and that of course allows water to percolate around the seed that layer. And then in this, this one it's, it's much thinner in the domesticated and actually the percolation layer if we call that as much bigger so it's allowing that water to get in and germinate quickly. So we ought to be able to see over time, but the problem is it's on the inside of the seed kept so I'll come back to that's what that's one of the new directions that we're now now using these new technologies for to wherever we've been so if we go back to the end of the 20th century, where we knew there are things like differences in seed size, you know, the thickness, but the approach tended to be a typological one that tended to be, it is or isn't the domesticated type. So these are examples so here's North American domesticates published in science in 1989 earlier ones, you see it's not there's no sample size or standard deviation or mean it's sort of like I've looked at a few specimens from this side of this date, and they are small these are small seeds. And then these are large seeds and they're across this dashed line this kind of nominal threshold that this scholar created it says they're longer than this they're just domestic. So it's either wild or domesticate so even though he dashed that line there he doesn't have an evolutionary process that it's kind of not a statistical test it's just a kind of is it or isn't it. So this is a number of things so this is I've annual that's an extinct crops of the one that the sunflower back at the kind of podium so the North American version of quinoa. And this is actually seed coat thickness this is the only seat of thickness study that really existed back in the 20th century and that was from finding broke lots of archaeological seeds some of which were broken so you can kind of see a little bit of the seed cut and say oh it's then. So it's been around less than 20 microns, or it's thick around 50 microns. So again it was an easy. And similarly this was some work by a couple of colleagues in the college in the college in the UK so college measuring wheat grains of early sites in Jordan, saying okay the ones with thickness and breadth that are on the lower end those are the wild ones the ones on the upper end are the domesticated ones it was a kind of threshold was an either for. Things started to change in a kind of pivotal paper was very short, I think they used to call a science post ill or something was like a one page paper brief brief brevia science brevia, I don't think they have them anymore anyways. I published two images in it from a George Wilcox was based in France now retired and can eat you can I was a Japanese post off working with them and they had they worked on three sites. Well, Al Kirk and coach Dr Molly at a foresight in the voluntary. And what they did is they were looking at the rake is is of wheat of iron corn wheat and said well you can see some with the smooth wild types car shattering and the rough they were domesticated but they actually counted up the proportions and said well in the earlier sites, they took a couple existing published data sets were directly not a week there are barley so this graph is curious in a number of ways it combines two different species and so well we can just put them together and assume that they're evolving in the same way at the same time in the same space but anyways that's. It was it was really good anyways you have the indecent domesticated type of purple, the dehiscence of the shattering wild type in dark blue and then ones where they weren't sure, but we're possibly domesticated in light blue. And that's about the time series, as you can see that there was this change, it was a gradual change so they said, you can't just say type of this while the domesticated we look at a population level through time. So that was actually a kind of watershed moment. But of course they're combining data from the number of different prop with two different crops. And that kind of got the floodgates open for people to start to think about the stuff including myself when I was younger at that point I was just starting to work in China. And I just started a collaboration on this on this cycle, Tim Loshan, which is discovered in 2004 and I did field work including systematic. In 2006 or just as that paper came out. And it's a waterlog site that's part of it there in the picture, and in amongst that waterlog material there's lots of wild foods and acorn not children things but there's also lots of rice and lots of rice spike with bases and prior to this. And even tried to recover rice by the bases in China because again the question was, is it well domesticated. It's an argument it's wild, it's domesticated they just assumed it was wild. They didn't think well we should try to find the attachment scar and ask is it a smooth to his type with that round scar so that's a modern one. That's a waterlog archaeological one from the site that the SCM of a carbonized archaeological one, or the domesticated type which is ripped out because it doesn't chatter so people have to break it fresh and tears. And so these are the domesticated types, modern and archaeological. And then we get a third category which I argued was probably immature green harvested. Right so that could be either way it says harvested before it's ready to his perhaps. So we have all three on the site. So then it became interesting to do a sequence, and we had three phases of the site so we could count them up. And the sample size is something like 2600 across the bottom for the different phases in total. And so then for each, for each phase you have multiple samples across the site that are roughly the same age. So from that we can get what is the average that period, what is the standard deviation. So that's what these data are and this is the non shattering type increasing and the wild type in the immature type increasing at that time there was kind of nothing to compare to so I got a whole sample from a much later site, and said well this is where it's headed towards domesticated up here at like 70% and very low proportions of immature and lower proportions of war. This of course, then led to a kind of strategy of let's compile existing data because there was some data out there. Let's go back to all the sandwiches and count stuff up. And then we can start to look think about rates of change so it's an evolutionary process how fast is it. So this is a compilation I was involved in compiling and I think it was published 2010 so this is brain size we also said we can look at grain size right to this is brain size assemblages mean and standard deviation maximum up here minimum down here for barley insights in the Near East through time and down average they go up and then they kind of level off. Sometimes they level off sometime around let's say seven or 6000 BC. This is an important week averages going up. And as of 2009 this was the non chattering data we had we had those three data points for rice, Tim Loshan, which were larger assemblages and we have these barley and wheat assemblages from the Near East. And so we've published this, this data looking back at it now it's like very few sites very few assemblages small data sets, but we can at least start to see that there were similar slow evolutionary tendencies and two different parts of the world, involving at different crops. And this just shows the kind of, you know, the end point in the beginning point of kind of average time for wheat grain size getting fatter and thicker to time that's from the site of average radar from the lowest level We've been playing a game of using those to calculate rates of evolution what's called the Haldane rates what's a rate of phenotypic change per generation, which is basically shifted if you imagine a natural distribution of a biological trade is a normal bell curve distribution just got a mean, and then a distribution around that standard deviation that's a shift. So one how they use as a shift in a standard deviation unit in one generation which of course is very, very rapid nothing to hold up so everything's a kind of fraction one, but then not allowed you to standardize. It's kind of a scale for unit, right it's not changes in percentages or changes in millimeters of grain thickness. And so this is our first attempt to come up with kind of Haldane rates of evolution various crops. I'm not going to digress on that it did. The one thing I would point out is that it did assume a constant range of evolution, and one of the future direction to really start to recognize that evolution isn't constant it's constantly fluctuating. So this is a change, cultural techniques and technologies change, and the plants are adapting and the cultures are re adapting to those plants and environmental conditions of course we expect evolution to be directional but not stable in its rate. This is just a snapshot of sort of where we are now, by no means all the species that speaking of large data sets for seed size increase. So we've got percentage and increase in the mean of populations and you can put lots of species on one percentage graph. And this is percentage of non shatter and so we've got two species of week I current number week and barley and rice. Some sort of a permanent data from Africa here, and various population so you can see that they're starting in any different times the rates of things like seed size increase on average are not the same for all crops with the kind of directionality So this is sort of a snapshot of where we are today using what I would call conventional method which is identifying things counting up qualitative differences quantifying those quantifying metrical differences. So these are things like compare across within a crop across traits and across a region. So this is just an example of barley from the early barley from the Near East separating out different read, but what data we have from different regions of barley and you might look at it and say well, we have a lot of data from the Southern Levant still Israel, Syria, Jordan, not so much from Cyprus or central Turkey or the East in Iraq and so there's still some patchiness geographically so we can target that as our target. So I'll just try to fill that in, if it's possible to go and sample in these areas and this is the grain size increase. So from those we can then look at the rates and say there's certain there's a certain sub period for changes most rapid. So that's what I've labeled as the episode of different types of kind of period of most rapid change. So we can start this is stuff I've been working on recently trying to calculate how they rates but over subsets of that distribution so take every three or four data points and say what is the rate. Over that period of time was the next three or four data points. What you get is the kind of variable rate so the rate of evolution the rate of morphological changes changing through time. So high rates, low rates, high rates and low rates for different props, then you can so to one of the periods when it's quick evolution and barley size or critical evolution and barley makers or quick evolution and we break is pretty evolution P, and see if they overlap geographically chronologically and culturally and then start to think about what's driving. So that's a kind of rough compilation of in 250 year bins. When you have the fastest rates of evolution in cereal grain rate is so grain size other crops in size so interestingly there is a kind of poor period, but most of the crops from the Near East are evolving most quickly without them. There's a question about one of the underlying cultural correlates our cultures adaptive that are driving that so doesn't correlate with things like diversification of simple sickle types and sickle complexity which is, this is basically the example for harvesting and the old they do is that sickles drive domestication. This kind of turns it around it says well domestication actually drives sickle evolution so sickles as you move up the scale with some more complicated. More steps to producing sickle brain more specialized tools adapted culturally to harvest things that are domesticated. So, again it shows that co evolution. So that's the sort of direction of travel for where we are, I think in domestication studies using conventional method there's a lot more to do, especially gathering more data, especially expanding to more crops in mobile regions there's a lot of data from the Near East as you've seen some data from China, much less from India much less from Africa much less from the Americas so there's a lot of work to do. So far this work is focused on two traits really monitoring in cereals where we get the spike the basis preserve and seed size and sort of anything. There's other really key traits like that loss of dormancy and secret spinning, which this isn't getting at. And it's also true that some of those regions of the world like Africa have less data because there are less of these seeds recovered and preserved so are there other ways to get that information, and the beauty of these new of these. These imaging methods these x-ray methods that are allowing us to get new data from old collections or from existing collections in two different ways both in filling in gaps geographically, as we'll see with Africa. So the work that I'll come on to a little bit that I've done with colleagues at Australian National University in Canberra and working on African ceramics to get evidence for crop domestication there. And then, and then also looking inside of seeds to look for seed to thinning which is work that we did some pilot work on with the synchrotron in the UK at the Diamond Light Source. So I'll talk through both of those. So, I guess everyone here will be familiar with what the synchrotron is usually in my archaeological lectures do a few slides on that. So I presume you all understand the synchrotron is probably much better than I do. It's a, well I guess in the old language we say it's a nuclear accelerator produces a high power x-rays and allows you to look into things, and that's much more complicated than that but that's the easiest way. So here's some pictures of the one it did caught. So we wrote. I had a postdoc Charlie Murphy who appears in some of these pictures I think that's her in the distance there loading a seed in. It came to me I guess in 2014 and she said, Oh, I just saw this paper on imaging fossils inside rock without taking the rock with plant fossils. Isn't that really interesting you think we could try something like that with seeds and of course part of her postdoc. So for a scientist use a postdoc on a project I had was to work on legume domestication in India so we're getting lots of archaeological legumes and modern ones that you're measuring them and do lots of measurements. I guess you measure a lot of stuff it drives you a bit fatty. That's what she was looking for other things to read to take your mind off. So she came across that and said couldn't we couldn't we try this maybe we could look inside and do the seed job. So we wrote a, you know, one of these proposals to diamond light source for beam time. And they, they said we've never heard of this archaeological stuff before why not. We'll give you a bit of beam time. And then we went and had our, our two days of beam time in 2015 I can't remember the exact date it might say it on the slide. And we didn't know any of them so it was just by email correspondence. They determined that we should be put on this be mine over here, which isn't even in the main range I never got a full we never got a full two of the main facility we were, you know we were using this. I 13 be mine which is a very, very, very, very long to but it had been used for kind of biological imaging of things. And because it has good contrast imaging for low attenuating elements and partial coherence. 250 meter long to anyways. There's a bit of background out of already mentioned that seed coat thinning is something you know happens we've seen this picture before. That's the domesticated wild seed seed. This is just an example with modern teams here's a wild kind of podium album at hand it's a common weed all across your Asia, it's got a black seed coat. It's also got a thinner seed with pointy edges, and as a proportion a little bit hard to see this photograph that's when I just, you know section the seat the black seat that is relatively thick, compared to the overall seat thickness this is a domesticated one. And it's a cultivated domesticated prop and parts of the hills of southern China, Southwest China and you're not and also in Taiwan, the material the sample I have here is from from Taiwan. You can see the seat is bigger. That's one of our traits is kind of rounded at the edges but it also got a change in color of the secret and secret is much, much thinner. So we knew that this was a trade and of course it's been studied in the past, as I mentioned in North America by work by Drew Smith back in the 80s. Using the SCM at that time on specimens like this he'd go through archaeological sample say ah here's one that's broken so we can mount it in such a way that we can measure the seat coat because it happens to be broken. But of course means you're picking out the rare broken specimens, and you know you can see one little bit of the seat coat and estimate the measurement from that. So we thought well if we can look inside the scene on destructive you can measure the seat coat everywhere around and make it more quantitative more precise. So that's what we set out to do and we thought we'd do it with species called horse gram which is an important bean crop in India. So it's called horse gram because the British didn't really like it. So it's also considered a poor man's crop, it's under researched crops, because it was denigrated it's never had a lot of research put into it, but it turns up a lot archaeologically so in modern India it's about the fifth most commonly cultivated crop and archaea botanically it's like the fourth most commonly found crop in prehistoric sites in India, but it's the least discussed in research we thought that would be a good interesting candidate for those sort of reasons. So this one preserves archaeologically with the seat coat the seat coats on damage so. So what we did is we mounted a bunch of these. That's a modern one. We mentioned a bunch of archaeological ones on the end of little needles using. I think we use fingernail polish in the end. And then this is Charlie and loading it into the chamber and that's the needle with the whole thing on top, and then it was x-rayed with the fingerprint. This was done in 2015 and we didn't really know what we're doing I think the beeline scientists who helped us I mean they obviously knew what they were doing but they didn't really necessarily know what we wanted to get out of it. So there was a little bit of maybe we didn't quite approach it in the most efficient way but anyways. This is us doing that you know learning to both to you know run it through the machine and then also to do the computer tomography to get the slices later. So it just shows an archaeological modern and SM and then these are the kind of slice images we got which are not really great now but they were good enough that we could measure the seat to thickness on a bunch of these. So we published that in scientific reports initially. The thing to notice we didn't have a very big sample size and it was 13 seeds but from those 13 seeds we've got able to get lots of measurements on one. The thing we didn't do what you would do today segmented we didn't segment the images so we took the slices. And you know we worked out about 4000 slices proceed, so we took every 200 slice and we measured six or eight times around, and we took those as a kind of sample a kind of judgmental sample that'd be obviously better ways to do that. So we had a lot of documentation and to get many, many more measurements but anyways from those data we can see that. So these are those each of those seeds the mean and standard deviation of those stick this measurements plotted against the age of the specimen or the estimated age of the specimen. You can see this gradual thinning. And then these are some historic ones we had early century baby which are much thinner. In fact, we kind of concluded it's a kind of step change these are kind of thick. They're semi thick, and these are thin. This one's also thin. And then the data set at the pop was the other data sheet and generating which was measuring the seeds of seed. I think that seed, but they seed width, and that's in millimeters. So, but at the same time, so you can see that the seeds are getting bigger seed cuts are getting thinner. And the where the places these the sites these beans come from around itself in the red shows the approximate wild distribution. So we have a domestication sequence for this problem. We then went back and had a second session. We wrote for more times and we need to do more, more horse ground but we'd like to do some other things too. And we got really excited to do we did some Chinese canopodium. I was so, which do so he told me we have a public that we did a few lentils from the mirrors, we did a sequence of Chinese soybeans. And the next time I realized that when you have only two people because only the two of us went. And we had 20 we had 48 hours on the beam. You can't stay away for 48 hours. So there's a lot of downtime where we weren't producing. We went back to the next time I took two more postdocs and then we had somebody came part way through, help us out so we had a kind of constant churn for our all of our being time we used all of it that And then we published a kind of methods paper here with Bodhi and rower both the be my time talking about what we've done in both times. The other thing, you know, when we were there we said, well, you know, hitting with high radiation, is it going to excite the carbon in there is it going to turn carbon 13 back into carbon 14 and screw up radio carbon dates. And Christoph row is one of the physicists be my fifth. I don't know and never thought about that. Well, we can do very easy test so so this we took soybeans the ones with the stars were put in the synchrotron and blasted for 45 minutes or whatever it was. And the ones with all the stars were from the same flotation sample that had not been in the machine. And then we just said we'll just am a state all of them they can compare. And there's no significant change. Here's a series I was one secret on one and a bunch of others in the same contact they basically fall together and that was true at two different sites. So we're pretty satisfied it doesn't affect being able to synchrotron it then take it out and send it off and directly data. So that's a, that's a good thing for, because I'll just always want to be, make sure we can date our specimens if you know that can be important. And so I got some really good result, not all that the says written up yet that's on my plate because I'm still getting around to some of these postdocs and moved on to other things. I normally do nice reconstructions but Bode who was one of the be my own time to say, Oh, you can quickly put all the slices back together making good for animation. So I still need to learn to do that myself and it was that's an example of a archaeological soybean from early in the sequence or something like close to 3000 BC, and it's still thick seed coated and looks a bit more like the wild type. So with these we started to do some segmentation to that's an example of segmented soybean where you then you can kind of automate the measuring of the seat cutting you get hundreds of thousands of measurements from each seat. So the soybean data was again it was really a pilot so we have here we have T code measurements on six sorry beans two from one site but otherwise have different periods, you have the mean and standard deviation so again we see six seated early and thinner seated at the end. And what looks like what could be interpreted as a gradual process. The data below was previously published data on soybean size. So we still only grass or did it that's our different that's 5000 BP. So here that's 3000 BC here, but if you look at seed size in pieces soybean it is this period. At least some of them are getting bigger, and that's the period we're seeing Phoenix or again it fits with a domestication period for that problem. Okay, so that was one direction that was this that was the kind of sync patron adventure that I had. And that's a little bit different because we get so many images it takes time to work through until we still have. I have another former PhD students working on the kind of podium Chinese going to put your data still with other data sets to put together so that's on the way. I'm getting new data from much older data sets and this is looking inside ceramics. And this is to look at both the domestication of pearl millet a very important prop in some sort of Africa which originates in West Africa, somewhere like Molly and sorghum which of course is a very important prop in Africa but also in other parts of the world, and originates somewhere around central eastern Sudan. And this gives a sense of a distribution so the hatched area is traditional cultivation or it's an important prop so Sub-Saharan Africa, India, sorghum also goes into North Africa there's a little bit cultivated and bits of Spain and Italy and northeast China and even Japan and Korea. Of course, in the new world too it's a major cattle fodder crop in places like Mexico and Texas today. So it's an important global prop we don't really know the origin of it. And then pearl millet comes from West Africa it's it's particularly important because it's highly, highly drought tolerant. So most drought tolerance here. So the places they grow it in northern Molly or Sudan where you have 150 millimeters of rainfall a year. And it's grown in the most drought prone parts of India so it's an important prop that's also an important staple food for the first food producers in West Africa. And these are the sites traditionally where we heard the earliest finds of it. And one of the challenges is that the environment in which it comes from is one that is the Sahel so here's the modern Sahelian northern Sahelian kind of vegetation and in Sudan. And of course, we know that through the Holocene the Saharan desert expands. So the Sahel retreat does this hell retreats it becomes desert, which has massive events of windblown sand and that tends to stour your archaeological sites and removes stratigraphy. And that means you tend to become deflated, you tend to lose stratigraphy you lose preservation of plant matter. So, most of these sites have zero or very thin stratigraphy is no organic preservation no very little char material preserved so you don't get the traditional plant remain, but you do get the artifacts the stone tools and pottery. And luckily in the Neolithic very often pottery was tempered with plant materials you take your chaff or other plant material and you navigate the clay you add it to the clay to give it to bind it together, which is not uncommon with kind of handmade low fired ceramics and so when it's fired than that organic material partly burns away and leaves behind the kind of fossil of what was that tap material so we can use the ceramics to look for these crops. Now this is something that we've been doing for a while anyways, but we've only been working by old fashioned methods of looking what you could see on the surface, making silicon rubber cast of impressions on the surface, and hoping that they happen to be sort of a little bit or happen to be the right bits of the plant that we can say something interesting about it, which is still useful because it's very low tech, the budget you can do a lot of this later on the table with a simple microphone to do a lot of that material. But it's not the most efficient at recovering kind of high quality quantitative data. And this is where our collaboration with college in Australia. In Canberra came in to kind of try, try out some new techniques using a micro CT scanner to look inside the box. So these are the sites we'll we'll talk about we'll talk about some sites in Eastern sea and especially kg 23. And we'll talk about a few sites in northern Mali, especially easy 22 and in case 36. That's here at the base now both of these all these collections of ceramics are old in the sense that the archaeology was done in the early 80s. And so kg 23 was excavated by a Southern Methodist University University cartoon collaboration in the early 80s back when American student were friendly. Long time ago, and those ceramics were kicking around store rooms and Southern Methodist University, wherever, and then eventually we're donated to the British Museum for whatever reasons. So they came that way so allowed us to go back to this old material that was gathering dust and it's been studied. There's nothing more to do with it well there's more we could do with it. So the work in Mali was by a French project French team working in northern Mali, really on paleo environment and climate change and dry erudification of the Sahara in the early 80s, and they collected lots of certain material, and that was sitting in in in to lose. Anyways, in the south of France. So we've been able to go in and kind of, in a sense, reescavate re examine the surrounding ceramics to get new information out of sites that have long been finished. So I'll just walk through these two cases. Here's where kg 23 is so it's kind of between cartoon and a city called cassava. It's on a paleo channel dry paleo channel that feeds into the at bar river here. And the upper of course comes off North Indian European and joins the Nile north of cartoon. When the work was here's what the site looks like on the earth today so it's an impressive kind of natural mound that then has kind of about a meter of archaeology on top of it. And when it was excavated they didn't do archaeo botanies we do it today they didn't float, unfortunately, they didn't collect fauna, and the fauna is interesting because it's basically a hunted fauna. You know, gazelle and hairs and so forth, so forth, some word hogs in the very upper levels of a little, a few bugs or she's gotten cattle so it looks like that the transition to pastoralism, and most of it is pre pastoral. It's also curious for this part of the world because there's no fish on the site. So most sites in this period that are better known are near the Nile Valley and they eat a lot of fish. Here, it's a probably seasonal river and they're not really easy. They've got a lot of ceramics and Frank Winchel who did a PhD on the ceramics on the site back in the late 80s, they're only published that much more recently, kind of moving towards retirement he said oh I need to publish PhD and work on the I found that about 20% of the pottery was this type, this cordyce plain type so not decorated, but heavily tempered with plant material, including chaffee type material so he did some of the work that we published on chaff impressions from elsewhere, and sort of got in touch and he said well I think that a bunch of this stuff is a chapter but might have sorghum in it. Could you take a look at it and it's being donated from being shipped from the US to Britain anyways to go into the British Museums question maybe to look at it first before it gets lost in the British Museum base. So that's what we did so he, he picked out 92 shirts that he thought were promising. We got 147 shirts of those. We looked at them we found 92 shirts with some sort of plant impressions on the surface. We did a first paper on that in 2017. We said okay it's about one third indeterminate one third wild one third domestic. And then it was after that that we got in touch with colleagues in Australia, Professor Tim denim and his PhD student then at least Baron with exploring micro CT uses for archaeology there. And they suggested that they could try a few shirts to see what they could do to look inside not just on the surface but inside so we sent them 12 shirts. So this is what the kind of shirt impressions look like on the surface with conventional photography is kind of see that's bits of chaff is a bit of a spike the base there rake rakes. You know there's more bits of chaff maybe the rakes is smooth or it's broken. It's not the best way to study it the conventional ways to then make casts of that with silicon to get a positive. And then you can both coat them and SCM them so that looks like this or here's one we got to two husks of two bits of chaff and a torn Rekilla that's characteristic of domesticated Thresh sorghum. Here's one with a smooth spike the base so that's characters of wild. So that's the work we done and published in 2017. Here's comparing looking for smooth versus torn Rekilla, but a lot of times from the surface and from these kind of things. They're partial they're incomplete you don't see bits of them. You can't turn them over there are not three dimensional objects are just whatever happened to be stuck on the surface. And there's a lot of certain amount of indeterminacy in them so here's, here's some ones that are domesticated here's that's a I guess a wild one, but what it did indicate clearly. This was chaff 2017 papers and this was chaff from the de husking and winnowing of processing sorghum. So this is traditional, you know, de husking of sorghum, it breaks off the gloom so they can eat the grains then you've got this nice chat product which can be used by the potters. So, we know they're using the stuff but can we get more out of it by micro CT. And you know sometimes it's like, well type probably just get that probably, you know, a little bit hard to tell. Okay, anyway, so this is just some examples, you know, I had a postdoc christy with two went through these over and over and photographed all of them and, you know, we looked at the SCM again and we would argue about whether it was domesticated or wasn't always easy. And some had no impressions on the surface so when we published it in 2017. And we did another site nearby from another from an Italian project cost of the K one, you can see the white here is the indeterminate that's where we said yeah it's a bit of sorghum but we don't know. We can't tell, we can't even agree whether we think it might be well domesticated, and then the wild ones are the blue and the domesticated ones are gray. Now prior to this work we've done that been previous work with some impressions of sorghum mostly wild but very very very small sample size. So that was the state of plan 2017 when we sent these 12 shares off to Australia. And of course what they did is. So what they've been doing at that point in 2017, you can see Baron at all. This is the Australian group published a paper around the same time in 2017, where they took a shirt, or a few shirts from a Neolithic site in Vietnam, I think. So they took a piece of a shirt, the second and micro CT scanner, and you could separate out the or the different densities of materials or the organics, the component in green, or maybe that's the mineral component in green and the organic component in red. And this is Vietnam and in Vietnam they could go through that organic component and find the rice husks, and sort of set the segment out the rice husk so here's a rice spike that that was used to temper ceramics and Vietnam and then you'll have the period. And they were able to pull these out of shirts so this this was sort of where the idea came from so they were driving this so like we can get inside the shirt and find the rice husk temper. So a follow up study to that that again at least better mothers take a single shirt from the site in do a Sarah is Indonesia. And so here's the do a serious shirt, and you can see there's lots of organic material in there in this case organic materials in green, and you can pull out all the individual rice spike that basis so she could segment out and get high resolution models for the right right back to the basis over in the interior of the ceramics, and then we could look at those spike the basis, you know domesticated type, mostly they were domesticated. Some of them, because I've done that work on physical by the basis in China they got in touch with me to say oh can you check what we've called wild and domesticated. And then gave us gave the idea of they argued that a single shirt could be a whole archetypal assemblage because from that, you had 50 plus right spike the basis, which is good for one rotation sample but that's one shirt sitting in the museum. So that was the background so then, then it was like okay 12 shirts from Sudan so they're Tim denim and at least he's not finished a PhD that she was one who did all the topography work in terms of the machine but especially sitting at the computer like processing those machine this is some pictures of their facility in a new. So, basic question for us was can we get more information out, and obviously the answer is yes. I'm also not going to explain what a micro CT machine is here because there are people in the room who will correct me when I make lots of mistake those very similar ones that I saw yesterday. It's, you know, it's producing an x-ray you put your sample in front of the x-ray. It's detecting it there. The longer you leave it in the higher resolution you can get but of course there's different resolutions within the machine built in machine, but you get down to one micron or something and that sort of work. And then they use this software that they were working with in Australia to do that. So the samples were put in a tube I mean if they were in fact stacked in a tube, so they stacked several shirts in one tube, and then obviously scan them. And the point really was to separate out the clay of the pottery from the mineral inclusions that are harder things like port sand and the organics and or pasts that around the organic stuff is burnt out. So there's just one one animation of one of these examples that they at least produced for us. And then of course once you've segmented all that stuff out you can then remove the, let's remove the clay component and just leave behind the organic component there. And then once you've got the organic component you can focus on the individual fragments within it to say which are the ones that look like sorghum spikelets. And once you've got the sorghum spikelets you can look at those in quite high definition and unlike the surface impressions, you can turn these around and look at the spike the base of an ambiguous that's got a torn Rikila of the, of the domesticated type, or the surface impressions are so like, is that a torn pillow. So, much, much higher reliability, and it reduced our indeterminacy rate along. So this is an example of one of the students so some of the impressions on the surface versus what you could see from the interior so suddenly from a shirt that might have one ambiguous impression or two on the surface. You have several in the interior, which are unambiguous. That just shows another animation of these are some of the then resulted so you can pull out the specimens and say okay that's a wild type domesticated wild type of Thresh broken domesticated type. And suddenly also it's not a sorghum that's like a mechanical or some other grass right so. Yeah, and so what that did is allow us, and even when you have mixed so chaff of course it's mixed together so here's two fragmented bits of chaff, both a wild type but they're kind of one spike the bases here with the tornado ones like it is here kind of disentangle those you can really see a lot. And the key things in the graph of those statistics so if we look at what we published in 2017, we had around 5051 and about a third of them were indeterminate. And then in this study, kg 23 that produced 83, not only 12 shirts, and you know 5% or less were indeterminate and increase the number of identifiable sorghums for sure to about seven where before we have less than one. So much higher returns great. So that was great now that I think at that point, maybe at least was kind of winding down and finishing a PhD. And so at the same time we sent them that we send them a little bit of stuff from the French retail from Molly, not as much as I wish we had. But here we're looking at a very simple process so here's wild pro mella smooth scar down here domesticated promo has three things that change you get a torn. And instead of having one grain in here you get two grains or you have paired spikelets. And then of course the grains also get fatter. So we're kind of looking for those kind of things and previous work I've been doing these are the kind of surface impression approach or you peel bits off surface impressions or he has paired spikelets. And then this is a side of the torn or killer is often surfacing the site in more Kenya. And then this is a site in Molly, you've got that bit of the torn or killer torn or killer. Right, this one probably other. That's one from Sudan, and then occasional grain grains are really rare so like this site a character got nor we have two grains. And so we have a very small sample size from the surface of the pottery we had, I think, nine impressions that we can identify as domesticated a while so again very small sample size. And so we had this new material from the French collections in the South of France, but the 22, and these cover a range of different periods. So we had, we'd already been working on easy to do so we had surface impressions by SN these are them here. Most of them sort of look like wild pro mill or you couldn't really tell because you can see these are kind of broken off. So we needed to look inside. So these did so from a single shirt from AZ 22. It was just chocolate block full of perl millet cast inside, and they are all single spike lit narrow and all have a smooth base so they're entirely the wild type. And there were more she could have counted you because limited based on time how many she could go through, but they were 100% wild. So we had these two other sites where again, we sent them I guess 10 shirts or micro CC from this site, and I'm from this one. And I regret that but in the case these are two slightly later sites. But from the MK 36 you can, you get a mixture of wild types and domesticated types, as well as one you can't tell and someone paired spikelets which is a domestication trade. And this site is later it's sort of, let's say 3000 2500 BC. Now the dates unfortunately are all dates based on charcoal so it would be nice to date have new dates. So this yeah so this shows a parent spike that this shows a non shattering type where it's torn these show some wild type ones. And some with mixed features so ones that have to a pair of spike or shattering. So the kind of mixed features in that. So small data separate put together you know here the one that appears as a non shattering, the kind of maximum minimum estimate from various sites. These have none, and then paired spikelets, these have none and then they're starting to appear around 2500 BC and then from those we could combine that with brain measurements from charred material as well as the estimated average of the maximum within pottery, and you have a grain size 10 so that's the average and standard deviation of the maximum grain size reported on. So you can see that there's a trend in grain size increase, and then those other kind of hard domestication rates, kicking in. And a lot of that although a lot of that data was without the micro CT the micro CT stuff was was really key because it grounded us in that this is 100% wild. So here is starting to get a mixture of domesticated and wild. So able to kind of book in that domestication process. And until we get more data, that's sort of where we are. And then at least publish over this year kind of summary of all of the method they developed there that she worked on the PhD, including these various case studies, and others so she's done some exciting works imaging charge tuber fragments with the tissue of tubers and three dimensions, which opens a potential methodology for studying domestication of props like yam and sweet potato and terror and so forth, which would be an important direction. Okay, the last thing I want to talk about is something else we've been working on, not just in my lab but in my lab and in other labs across Europe is increasing interest in the charred remains of foods. Because those breadcrumbs it falls in the bottom of your toaster and get charred and accumulate right. This kind of stuff, tiny charred specks of burnt bread and so forth. They can survive and they come up with flotation, and they end up in your flotation sample. And so you have your brains and your chaff and your wood charcoal, and you often have not all the time maybe half the time maybe 30% of the time something you have a few of these amorphous charred lumps and our care botanists have tended to just set them inside and say, jump. And then the larger ones that say well maybe that's charred bread, maybe that's charred food with a big question. So we, you know, I thought it was time to start trying to system out a look at this stuff. But of course, the different ways of cooking things, porridges flatbreads, leavened breads. These are Indian idlies which are steamed. These are Indian water which are deep fried, they're in fact the same ingredients, right, but they have very different textures as a result of how they're cooked. So these are examples of charred lumps. So these are nice big ones from the site of Chattel here in Turkey, which is the site. So these are really big ones, these are millimeter scales and several millions across. Those are exceptional, but you do get those occasionally. And it's clear that they have variable texture. So I said a PhD student, a task of kind of working on this Laura Gonzalez has now started teaching at University of York, finished a PhD a few years ago just before COVID. And her PhD title was something like the or a methodological approach to the origins of bread culture. And this built on a kind of observation that if you look broadly at different parts of the world you get different early cooking traditions. So for the East Asia they have really, really pottery back in the Pleistocene 15,000 years ago. By the time you get to the Neolithic and their cooking rice, they've got really elaborate multi stackable steamers and things and they're, you know, really set up for boiling cooking. There's no side in site, no site, no architectural site in China has an oven. Most kitchens in Shanghai don't have an oven, right. They wouldn't cook in an oven. You know, what's that for. That's not part of the tradition and that goes right the way back to the beginning of agriculture and before. Whereas if you look at the Miri's, there in red we have these clay bill ovens, these little taboos ovens for baking flatbreads and naans and things. And those are earlier than cooking pots in sites in the Miri's. So we have a pre ceramic Neolithic and you have pre ceramic bacon. So I think I've written about this a number of years ago but we didn't really have a way to approach that archaeo botanically. So that's sort of what a lot of started doing by both looking at the archaeological stuff from sites like in Turkey, but then also doing experimental recipes of making bread, making batter making porridges, and then touring them, and then SCMing them. So these are some archaeological examples of different matrix types. And essentially we approach this as you might approach a ceramic sin section where they have lots of porosity, and it's a porosity kind of regular or irregular. They have lots of inclusion particles and all those particle sizes regular irregular what's their model, what's their distribution. And then doing the same with the experimental stuff. And so in fact we published a kind of semi quantitative qualitative set of charts and could use to estimate the percentages of voids. So those are your pores, as well as poor size for size across the top and then this is particle size and particle So rather than actually trying to count each one of us a kind of as as you will find in kind of soil, soil science manuals or ceramic analysis manual these kind of percentage model charts we wanted to do some that we could apply to this kind of food material because of course food is quite irregular. It's not regular throughout right so you're looking for tendencies within it. It tends to have small regular voids and large irregular voids, but it tends to be cracked or not correct. So we were kind of semi quantitative and qualitative. And that seems to work for telling apart breads. So baked breads from batters from porridges in a broad sense of some results from Taddle here to this kind of thing with SCM we zoom in and you can also then see, so you can both describe the. These are the bread like products with the lots of these small micro pores, and then you can zoom in on some of the particles in there and say well that's a bit of a charred seed and in fact it's got what looks like a kind of seed card from a pulse. And here's another food fragment that's got a bit of what looks like all your own tissue that's underneath the layer underneath the brand and wheat and barley is what's called the all your own layer. And sometimes those are there's enough material that you can say well that looks like a wheat or barley or it just kind of looks like generic cereal out your own right. So let's just look at this, you know, different level of identification. And some of these things. This methodology than allowed. So this is a pre agricultural site in Jordan culture, but the date has about 14,000 14 and a half thousand years ago. So it's pre agricultural but they're starting to build some of these interesting kind of communal stone structures with the kind of pit of them this is pre regular but a pit of them. And amongst the flirtation samples there there are lots of these food ways. So the postdoc who's doing there if you bought me a Maya around a tag we came over to London partly to use our reference collection for Near Eastern seeds of them, got to talking with me we should see some of the food so this was done in our department where they work together on identifying these pre agricultural breads. So sometimes people ask me what came first meat or bread and the answer is clear bread came first. So this stuff is made with a mixture of wild and corn wheat and sedge tuber ground up sedge tuber starch, and maybe some other seeds so they're making fat breads before they domesticated. Raise also entered questions like did they prefer to domesticate week because of the gluten content because it made better bread. Why is it that all the gluten containing sales were domesticated in the Near East or not in China or Africa right so it starts to perhaps we can ask me questions anyway that allowed us to take this back to a very early. There's a little bit of work in other places like in this site called Hammadab in central Sudan. And there is nice because we had a set of systematic flirtation samples that they collected and one of my students helped collect the numbers years ago, and out of 60 architectural samples the samples with food was just about 30 so we were able to determine that 50% of flirtation was actually a lot of this material to work on just hasn't been worked on. And so very preliminary work that Laura did here show that we have we did have some breads but most of it had porridge type matrices, and mostly when you can identify bits of plant tissue it looked like it came from the surfaces of sorghum grain so this is a sort of culture, rather than a weak bread culture there's a little bit of weak bread, but very, very little. So it's just about the first century. So there's a lot more work to be done on this and what we need is more controlled experiments of food preparations more gathering of ethnographic recipes and ethnographic samples, because a lot of ways you can cook things that are sometimes quite surprising that we may not think that people might have used in the past. And so, before all these ethnographic traditions disappear we need to record them, and think about how they would leave behind chart tracing the art and the archaeological record so that's a direction that works I don't have that one of the PhD students here is one of my PhD students from Taiwan who's done some ethnographic work in Thailand with the Aka people, and then she's reproducing some controlled experiments and cooking pots of different methods for steaming and boiling rice that they're reported where we can control temperature and cooking time and quantify the contents and all that kind of stuff. And then that's chart and it's just some other strange. It's a raised stiff porridge or sorghum flower that I photographed earlier this year. And so I think that the potential then that we've been doing SM work and this kind of semi quantitative qualitative description which I think that's quite powerful and those potential to take it a bit further perhaps with a three dimensional approach. So we look through the entire fragment of food, because then you can slice it and look in lots of, lots of different views, not that one kind of haphazard view that you have that you stick under your CNN, and you could potentially then start to quantify things in different ways. And this was a piece of food from charge food from South India that Steven Allen, and Michael Favel have put in a microscopy scanner over the down the road or whatever in who sits yesterday. Obviously the full tomography is not there but it just shows it shows some of the potential so I think that's another direction we can go both with the ethnographic and experimental. And I understand that a lot of these kind of tools are used by a specialist from the food industry so it's probably also a lot of specialist knowledge and how foods behave. And we can also collaborate with to better understand ethnographic and ancient foods. Okay, so my summary is really that we have a, you know, plant domestication studies have been going apace in recent decades for an expanding our chemical records more species more sites more region of the world, getting involved and we can continue to do that and should do kind of conventional archaeobotan simple, you know, more data quantitative time series and analysis but that will continue. And some labs like mine will, that's our kind of bread and butter that's what we do, but it's also the opportunity to get new domestication traits things like seat cut thickness through using new technologies imaging that requires collaborations with those who are who can do the x-ray tomography and also we can get new data from perhaps surprising sources like inside churns of a gathering dust and, you know, in a French or British Museum, we can say well those could be really useful for parts of the world where we don't have other lines of plant evidence, and we can we can get those through these x-ray tomography methods as well. In the early stages of developing this kind of archaeobotany of cooked foods and processed foods, and that I think can also benefit from these new imaging technologies by looking at them in three dimensional, three dimensionally and also through those interdisciplinary collaborations that will create. And that's what I wanted to say today.