 Okay, we'll probably get started with today's brown bag. Our special today is John Marston from the Department of Anthropology at Boston University. He's a freshly minted associate professor, so I think we should extend our congratulations to him. We're not seeing him better, but of course, he's this semester a visitor at Stanford University. But he's going to Australia, especially Queensland, on a full bright. So he's got a pretty nice year ahead of him once he finishes his talk here. John, who did his PhD in the Prozincute of Archaeology, UCLA, and his BA at Washington University in St. Louis, is, I guess he described himself as a fairly much no botanist. I would say an archaeoplatanist, but, you know, reverse. And he's pretty interested in kind of Anatolia, Levant, Iron Age, and early historical society, and in particular issues of agriculture and intersections with food production and environment, and evolution of productive systems and their impact on the environment and things of that sort, which I'm probably only partially correctly characterizing, but you get the idea. He's going to speak to us today on the topic, which is announced on the screen there, which I soon will pick up on a lot of research that was included in his recently published 2017 monograph on just these issues at the site of Gordian. So without further ado, I will turn things over to John Marston, and let's hear what he has to say. Probable that he built this for his father, Gordius, after whom the site is named. It's an entirely anthropogenic 53-meter-high earthen mound, and it contains within it a wooden chamber. Here's the model of it from the Museum of Anatolian Civilizations. It's essentially a log cabin. Now, when it was excavated in the 1950s, they found the tomb completely intact. The dead king was laid out on an imported cedar coffin with remnants of his funerary feast, elaborate furniture which held bronze-serving vessels with their contents still preserved, the dirty dishes left over from the funeral party. My colleague, Pat McGovern, an archaeochemist at Penn, reconstructed their food as a spicy roasted lamb and lentil stew, their drink as a mixture of honeymead, grape wine, and barley beer, flavored with saffron. And then he teamed up with Dogfish Head Brewery in Delaware to recreate the beverage, Midas Touch. So who's tried this? Anyone had the pleasure? What do you think? Yeah? Yeah. It's gotten better over the iterations. You'd bring the early ones to holiday parties, and we're all like, Pat, this needs a little more work. But I recommend looking it up if you haven't had a chance. It's pretty fun. All right. So getting back to our actual focus today, what data do we have from Gordian to reconstruct agricultural strategies and what were the environmental impacts of those agricultural practices? Were they sustainable over the long term? So here I define a sustainable agricultural system as one that balances cultural needs with what a given ecosystem can support. Sustainable systems are resilient to variability and designed to mitigate long-term risks. In contrast, systems designed to maximize short-term profits may not be sustainable over the long term. To conceptualize the environmental changes resulting from these practices, I apply the concept of legacy effects that comes from resilience thinking. I use this to frame the consequences of land use during one period for subsequent inhabitants of the region. Environmental legacies constrain future choices and may change the sustainability of certain agricultural practices. So my data to explore these questions come from environmental archeology, the direct result of interactions with the environment, which provide a unique perspective through which to assess the long-term sustainability of agricultural systems and to see why specific agricultural systems succeed or fail within their contemporary social and environmental context. In prior work, I argued that one of the best ways to achieve a sustainable agricultural system is to manage risk through diversification of agricultural and subsistence strategies. Diversification includes diversifying what is grown among different crops, among plants and animals, and between domesticated and wild foods. Additionally, farmers can diversify where food is grown through techniques such as field dispersion and terracing and when crops are grown over the years, such as farming multiple crops with different growing seasons. Now, how do we see these in the archeological record? Primarily through the study of plant remains. One core tool that I use with both wood charcoal and charred seeds from archeological sites are indices that correlate specific patterns of archeobotanical remains with human behavior. I'll present several of these, and for each of these I'll describe how I draw from general theory to specific expectations about human behavior or plant ecology. And through this middle-range theory building, we can use plants as proxies for human decision-making. This relies on extremely simple statistics, generally ratios, and the development of these indices requires a thorough understanding of plant ecology within the region. So developing these methodological tools that link theories with everyday archeological data in a way that can be used by other scholars is the focus of many of my prior publications. At Gordian, I developed indices for agricultural risk. Here specifically looking at the risk of crop failure due to drought and for agricultural intensification. Here specifically looking at irrigation. These are based on the relative presence of certain seeds in the paleoethnobotanical assemblage. For risk management, I compared drought-tolerant barley to more sensitive wheat. And for intensification I look at the frequency of sedge seeds, which preferentially colonize irrigation canals and wet fields among all wild plant seeds. So from a sustainability perspective, we expect that risk-managed systems should be more sustainable and intensive systems less so. So let's graph these indices against one another to assess this relationship at Gordian. Risk management above, with a high value indicating a safer system and a lower value, a less safe system. And irrigation below, a high value indicating more irrigation, a lower less. They share the same horizontal axis, which is time, from the Bronze Age to the Medieval. And when we look at the data, what do we find? We find an increase in irrigation during the Phrygian period and then it moved to less risky agricultural systems following the conquest of Phrygia by the Persian Empire. In the Roman, we find a much riskier agricultural system focused on wheat cultivation and a significant substantial increase in irrigation, perhaps to minimize the risk associated with intensive farming of wheat. So to assess the sustainability of these practices, we need to consider not only their temporal duration but also their environmental implications or the legacies of these practices. One place we can see a legacy of human impact is in the structure of woodlands. One type of forest in this region is dominated by slow-growing juniper trees, which provide excellent wood for construction. When cut down, however, this forest is typically replaced by a secondary forest of faster-growing scrub vegetation that is more diverse. So how can we see if people were responsible for these changes over time? I identified wood charcoal fragments from trash deposits at Gordian to see what types of trees were harvested for fuel and how they changed over time. We show a similar graph here to before, but now the vertical axis is the proportion of the total charcoal assemblage comprised by slow-growing junipers versus faster-growing types that replace juniper. Juniper use declines substantially into the Phrygian period and is replaced to some extent by secondary trees that are minor components of the climax juniper, pine, and oak forests in the region that increase in frequency during later stages of occupation. Another measure of sustainability is to consider the effect of animal grazing. So the same soils can produce completely different plant communities depending on grazing pressure compared here inside and outside of this fenced area. We can use this as a marker of grazing intensity and thus a measure of the sustainability of agro-pastoral practices. To understand how plant communities respond to different anthropogenic pressures, I spend much of my time in the field studying modern plant communities and the effects of human activities on their species composition. This gives me insight into the interpretation of archaeological plant remains and what they show us about human actions in the past. So the natural step vegetation in this region is quite diverse, especially among perennial grasses. These are often preferred fodder for animals as they're very nutritious. However, in contrast, severely overgrazed step is very low species diversity. The only plants that survive are those that are anti-pastoral in some way, either extremely spiny or full of phytochemicals that render them unpalatable to grazing animals. So the presence of these types of seeds, plants, in the archaeological record is a good marker of overgrazed step. How do these seeds get into the archaeobotanical assemblage at Gordian? Through the burning of dung fuel, which deposits seeds into the archaeological record and shows us what the grazer's diet is. Thus we can track change in grassland composition over time. I developed an index for overgrazing that's based on a ratio that compares the number of seeds of species typical of healthy step grasslands to the number of seeds of plants typical of those overgrazed landscapes. So this is an indicator of forage quality. A high value indicates healthy and a low value indicates overgrazing. What we see is that the value of this index drops significantly into the Phrygian period and reaches its lowest levels in the entire record in the Roman period. Note these are the two regional peaks of population at Gordian. So the combination of deforestation and overgrazing in these periods led to erosion. Along the Sicaria River, which flows next to the site, five meters of alluvium post-date settlement and are the results of upstream erosion. This profile has been dated, so we know how erosion changes over time. Here we have a horizontal axis again of time and now the vertical axis represents depth. A flat line is very slow or no accumulation and a steep line is rapid erosion of upstream soils and alluviation. So we see slow rates up to the Phrygian period when upstream erosion begins in earnest and then an acceleration following the Roman period. Most erosion post-dates the Roman period and represents a legacy of Roman land use practices. So let's consider what these data mean together for our two periods of interest beginning with the Phrygian. The Phrygian is the peak of power, wealth and urban population as Gordian became the capital of the Phrygian kingdom and wealth increased. This allowed or perhaps required a greater investment in irrigation and agricultural intensification. Maybe there was less of a focus on risk management because of the need to feed the larger urban population. However, after the Persian conquest we see a shift to risk management. The legacies of this system are changes in forest structure that indicate deforestation and the beginnings of upstream erosion. In contrast, the Roman system was decentralized with a low urban population density but the population instead spread out across the landscape. We see evidence for high levels of irrigation over grazing due to the step health declines and a focus on wheat production which historical sources indicate was likely needed to pay taxes. It's possible that the locations of farming and grazing also changed over this period. The legacies are severe over grazing perhaps now differentially distributed across the landscape that induced massive erosion on a regional scale. So to summarize, based on this research I suggest that erosion is a result of intensification and it's the most damaging aspect of agriculture at Gordian because it removed productive graze land soils and buried good quality agricultural soils. In contrast, diversification offers the best protection against agricultural risk and it's probably the most sustainable agricultural system in central Anatolia and probably we can generalize this a bit more broadly. Now the legacies of one period affected later periods. So without Phrygian deforestation and initial erosion the environmental implications of Roman practices would have been dramatically different. This case study focused on identifying which strategies were sustainable and which were unsustainable but now that we understand the processes underlying the shifts we see at Gordian especially during the Phrygian Roman periods it brings us to question how to identify imperial policies and the specific ways in which they influenced large-scale agricultural strategies. More broadly, how do empires structure environments? This was the broad subject of a recent edited special issue of the Journal of Anthropological Archaeology that I put together with Melissa Rosenzweigen who was published last year. So go there for further examples and case studies on this topic. Now one way we can study this question is through a multi-site regional comparison of agricultural practices within a single imperial economy. I'm working on a multi-regional project that includes excavation at multiple sites integrated with the analysis of both published and unpublished environmental archaeological data from other previously excavated sites. So let's stay in central Anatolia and explore the differences between two imperial periods the Phrygian Iron Age and the Roman period. Remember that rainfall in central Anatolia is conditioned on elevation. At high elevations we have more precipitation. So if we look at contemporaneous sites especially those within a single political and economic system that lie across different elevations we should be able to see how farmers adapt or don't their agricultural system to different rainfall amounts. In other words it's a measure of how adaptive states were to local conditions in their agricultural strategies. There are four Phrygian-affiliated Iron Age sites with published analyses of archaeological plant remains. Gordian and Kerkenes where I've worked for about a decade as well as comparative though preliminary published data from the sites of Kaman-Kalei-Huyuk and Charter-Huyuk. Here I present one measure of environmental adaptation comparing rainfall to the choice of the main crop, wheat or barley. These two cereal crops are staples for central Anatolia throughout the entire period. Free threshing or naked wheats, primarily bread wheat which is what you see here and whole barley. Now in any one location there's more risk associated with growing wheat which needs more water for good harvests and that's the basis of the risk management metric that I showed you earlier. However, wheat was and remains preferred for human consumption so where rainfall permits people tend to plant more wheat. We expect the choice of cereals might then correlate with water availability. In analyzing charred archaeological seed assemblages from these sites we're able to identify the relative quantities of wheat and barley. The seeds look similar at first. We see wheat above and barley below. The modern comparative specimens are on the right and the archaeological specimens on the left. But these are distinguishable. I assure you of that. You spend a month or so staring through a microscope at these things and they become readily apparent. So in studying the ratio of barley to wheat across these different sites our expectation is that higher elevation sites which have more rainfall ought to prefer wheat and the lower lying sites with less rainfall ought to prefer barley. Since it's difficult to get local rainfall data, for example the site of Karakinas sits on top of a mountain. They don't have a weather station up there. I use elevation as a proxy for rainfall and that represents the x-axis on this plot. Our dependent variable which is displayed on the vertical axis is the ratio of barley to naked wheat. And again we expect to find an inverse relationship between the two, a higher value at lower elevations. I have good data from Gordian which is a low lying dry site and Karakinas, a high elevation site, a wet site, as well as preliminary data from Chotr and Khamon which lie in the middle. This is only four data points and the middle two definitely require larger sample sizes which is part of the project I have for my Fulbright next semester. But what we see here is a linear relationship. So this provides a point of departure for future studies to confirm this pattern but if this result is real what we see in the Phrygian period is an empire that adapts to local climate conditions. In other words, they're adapting their cereal farming to these local conditions. Now does this extend beyond cereal agriculture? Was the environmental impact of this uniform across all regions? And that's where I'm going next with this work by collaborating with colleagues working in the region to bulk up these comparative data sets and also to expand it to consider herding and also changes in forest and grassland composition as well as the agricultural component. Here we seem to find an agricultural system that is focused on cereal farming and that's an adaptive to local climate within the Phrygian state. But in general, we should expect that imperial strategies have a range of relationships with local scale agricultural decision making. So some might be adaptive to local circumstances while others might be intent on enforcing specific agricultural practices regardless of local setting. This can best be illustrated by further exploration of imperial strategies at Roman Gordian contemporary Roman sites across Anatolia. So again, I adopt a multi-sided regional perspective here, one that combines environmental archaeology with historical evidence to understand Roman Gordian within Roman Anatolia. Now unfortunately, Roman Gordian is the only fully published botanical data set from Roman Anatolia. So we have to turn to faunal remains to explore imperial strategies in the region. This map represents the Roman road network, that's the solid lines that are within and beyond the province of Galatia, that's this funny shaped dotted line. Gordian lies on the Roman road from Ankara, the capital of the province of Galatia, to Pessinus, another major Roman city, which includes one of the only published faunal data sets from Roman Anatolia. Other comparative data sets I'll discuss come from the highland city of Sagalassos, as well as the Aegean city of Ephesus, which lies just a little bit off the map to the left here. So what was the nature of Roman Gordian? According to Andy Goldman, who excavated the Roman levels, it is a Roman garrison fort. He bases this interpretation on the architecture. We find that the plan of the site is laid out exactly as military barracks in other parts of the Roman world. The inscription shown here, which is an epitaph for a soldier, and especially military hardware found across the site, a javelin head, arrowheads, fragments of scale mail. The inscription, as well as historical evidence, point to Gordian as the home of an auxiliary, or a light infantry unit, from the Danube region. The inference is that the role of this military unit is to protect the road into managed local tax collection and military provisioning within the region. Now I published the first account of agricultural strategies at Roman Gordian as part of an article in 2012. I identified a significant shift in the agricultural system, which I already described to you, towards bread wheat, which is risky in Anatolia because it's dry, but it can be ameliorated with irrigation, which is also evident botanically. I suggested this was a taxable commodity. There's evidence of significant overgrazing, but by which animals? I didn't have any access to faunal data, I'll turn to that in a moment, but I hypothesized, based on the taxation system that we know to have been in place during this period, that it was based on the production of another taxable resource, which is wool, and so I suggested that sheep were primarily responsible for this overgrazing. There's evidence that this expansive rural system had significant environmental implications in terms of erosion, and I concluded this was all unsustainable. Okay, so this study was then used by Julian Bennett, a Roman archaeologist and historian with a focus on the military, especially in Anatolia, as the basis for a paper in 2013. Bennett argued based on historical evidence that the unit stationed at Gordian was primarily concerned with managing the flow of resources to its military posts. And in fact, both the garrison itself but also for transport east to the front lines, this is during the period of Hadrian's Parthian Wars. He identifies my wheat as destined for bread for soldiers, the barley was horse feed, and summarizes that the animal economy was primarily based on beef and pork, based on comparative data from military sites and texts that are related primarily to northern Europe, which is the only area where a large number of garrison sites have been well excavated, but it's much wetter in those regions and it's hard to intensively manage cattle and pigs in dry Anatolia. So, regarding the animal economy, I predicted sheep, and Bennett, cows, and pigs. Who was right? Well, we needed animal bones and they were unfortunately absent from the levels sampled in the initial publication on faunal remains done by Mindy's Eater during the 1990s. So I teamed up with our new zoo archaeologist, John Unchakrilar, to study and publish the Roman faunal data. These results are based on the analysis of 1100 bones, an ISP count, from Goldman's excavations of the Roman period. I'll present these data and then revisit the debate. So, the data are presented by bone count or number of identified specimens as a percentage of the total assemblage by period. Across all periods we find relatively similar assemblages with a majority of sheep and goats as shown here in yellow, but a large transition into the Phrygian period, an increase in cattle and pigs at the expense of sheep and goats. Presumably this may have been due to the need to feed this large urban population and it continues to some extent after the Persian conquest, but returns to a focus on sheep and goats as the site declined in size. Again, however, we find a sudden change in the Roman period with more pigs, that's the purple bar in particular, and fewer sheep and goats. Bone weight instead of bone count provides a better metric for understanding meat contribution because cattle are so much larger than sheep, goats and pigs. So let's look just at the three types of animals that provided the majority of the meat component of the diet, cattle, sheep and goats, and pigs. Comparing now only the immediately preceding Hellenistic period in yellow with the Roman period in blue, we find a substantial increase in both pigs and cattle and fewer sheep and goats. So there is more beef and pork in the diet, but there's still a major contribution of sheep and goats and in fact their meat is providing about a third of the total meat supply to the site. Age at death data can help us to distinguish animal husbandry strategies that were used to raise these animals. So first let's consider pigs. A survivorship graph such as this is based on estimations of age at death from dental eruption and where. The line will decline from 100% at the lowest age class, surviving at birth, until all animals are dead. And again, we're going to compare the Roman with the preceding Hellenistic period. So the Hellenistic data indicate a rather steady kill rate across age classes. While in the Roman period we find more animals killed at younger ages. The line declines more steeply at younger ages. In fact fewer than 30% of pigs survived more than a year. This is consistent with a strategy of intensive meat harvest as pigs begin to reach their maximum weight at around one year in age. This is a strategy that requires stying in pens rather than extensive pig herding, letting your pigs kind of go wild in the forest and then hunting them once a year. So either the pigs were kept on site by the garrison itself or close by in specialized pig farms. I won't show the cattle age data here but they were also killed young and also do not appear to have been used in field labor. So we infer a similar system where cattle were raised specifically for meat for the garrison. In contrast sheep and goats show a dramatically different survivorship pattern. We're going to see exactly the same type of graph here. And again the x-axis represents dental age, although in different units. The animals were kept to old ages during the Hellenistic. Note the relatively slow decline of the curve and even older ages during the Roman period. These animals that were eaten on site were very old. In fact more than half of them were older than four years. This is really tough mutton. So this pattern is consistent with a wool based economy. Both male and female animals were kept until old ages. And if this is a wool economy we ought to expect more sheep than goats relative to the Hellenistic in particular. So the Hellenistic ratio of sheep to goats is about 1.3 sheep for every goat and in the Roman more than 3 to 1. We conclude that the sheep and goats were raised for wool not primarily for meat for the garrison and the pastoralists raising these animals are thus functioning independently and pursuing their own economic goals which are maximizing wool production. So let's wrap up with a comparison with the other three published contemporary assemblages from Roman Anatolia and compare them by bone count. So here again is the Gordian assemblage. Cattle, sheep and goats, pigs and other which in the Roman period is mostly little chickens. Pessinus which lies also on the central Anatolian plateau and shares a very similar climate, had a similar amount of sheep and goat, much less pig, and the cattle while relatively numerous are old draft animals, used up draft animals. So this is an extremely different meat provisioning system than what we see at Gordian. The other two sites lie in wetter regions, Sagalassos which is up in the Highlands and Ephesus which is on the Aegean coast in a Mediterranean climate. These have less sheep and goat and more cattle and pigs respectively. Van Gordian. So this is evidence for a distinct set of meat procurement strategies at Gordian. It's environmentally constrained to a dependence on sheep and goats due to the relatively dry environment. But while we might anticipate a dominance of beef and pork if the site lay in a wetter region, we do see that the presumed preferences of the Roman soldiers were met to a limited extent through specialized pork and beef production. So to revisit our initial hypotheses, yes we find support for a wool economy and evidence that the sheep and goat pastoralists are operating in an independent economic system, not primarily producing for the urban meat market as are cattle and pig herders and farmers. This supports interpretations about the rural impact and the large-scale erosion visible archaeologically. It's due to extensive sheep herding. This supports interpretations also made by Bennett. Bennett's supposition about a preference for beef and pork consumption is supported, although comparison through the, yes, which we can see through comparison both with the Hellenistic and through the comparison with Pessinus in a similar environmental region. But although this preference existed, it doesn't seem to have been enough supply. The garrison was still eating lots of mutton. So we find evidence that supports both hypotheses, although evidently the garrison could not entirely stick to its preferred diet of beef and pork. We see evidence here supporting the existence of a distinct military meat economy with dedicated meat provisioning systems for pork and beef. Pastoralists pursued their own economic goals apparently related to the production of wool, presumably another taxable commodity desired by the Roman state. Earlier botanical data support intensive focus on wheat, also likely for taxes, and these two specialized extractive economies led to broad-scale environmental change, including the legacy of erosion recorded at Gordian. Of course, this set of studies is limited in both space and time, and so I'm broadening the scope of this work while also applying new methods to address questions yet unanswered in central Anatolia. I'm broadening this study to include other areas of the eastern Mediterranean and a broader time frame from the Middle Bronze Age through the Middle Islamic period. In addition to ongoing work at Gordian and Kerkenes, I serve as the Paleoethnobotanist for other active field projects. Two late Bronze Age sites in western and central Turkey, two sites in the southern Levant with continuous occupation across this entire time frame, as well as one in Egypt that's focused on several Roman period sites. This is all parallel with the central Anatolian studies I've described, but given my limited time remaining, rather than talking about work to date at these sites, I want to focus on new methods that I'm applying to expand my current work in Turkey. I'm working with these new methods to consider two major variables that haven't yet been captured in any of my prior work, the impacts of climate variation through high-resolution climate reconstruction and considering how agriculture, both farming and herding, varied spatially over time. So first, let's turn to climate. One great source of paleoclimatic data are lakes, which we can core and get various types of paleoclimatic and paleoenvironmental data. This map represents all published lake cores from the region. But notice that there are very few in central Anatolia and the two that are there are essentially completely worthless. And that is because it is an incredibly dry environment, so there are no lakes, perennial lakes that you can core. We don't know, however, if the surrounding lakes up in the mountains from which excellent paleo-ecological archives come, represent conditions on the plateau. They have very different climatic systems. So we need a local paleoclimate record. But fortunately, trees can provide that record. And we have many that live for a long time and are preserved in the unique structure at Gordian, especially that tomb structure. This is what the real thing looks like. These trees, which can still be found in the landscape today, albeit on a very limited scale, have lived for over 1,000 years in some cases. And in collaboration with Sturt Manning at Cornell and Charlotte Pearson at the University of Arizona Laboratory for Tree Ring Research, we're working to produce a detailed local climate record using measurements and isotopic carbon and oxygen signatures of individual tree rings. The goal is to produce an absolutely dated, through dendrochronology, local record within 30 kilometer catchment, roughly, of the site where we estimate these trees to have been taken, annually resolved paleoclimate record. Of course, this requires an absolutely massive amount of work, and so it's going extremely slowly. But in the end, we should have the first robust paleoclimate record from the central Anatolian Plateau itself, and one of the best paleoclimate record directly linked to an archeological site anywhere in the Near East. My second aim is to explore spatial changes in agriculture over time, including pastoral mobility through periods of increased climatic variability, to see how changes in the reliability of farming and the availability of fodder influence relationships between sedentary farmers and mobile herders in the region. So we know a lot about farmers from my earlier work, but rather less about herders and very little about the location of any of these activities. We can assess herding mobility, however, by assessing isotopic values of animal teeth from Gordian. So strontium is a trace element that appears in different forms in different geological formations. It's incorporated into teeth in place of calcium, with which it shares a chemically identical bonding pattern, and it's effectively a geological tracer of where an individual lived at the point when the tooth was growing. So since sheep, goats, and cow teeth form over about 12 months, if we take multiple samples up and down the tooth, we get a record of where the animal lived and moved over that time span. We can then map this onto a geological landscape such as the area around Gordian to find where the animals moved, and it's especially effective in a landscape like that of Gordian with diverse geologies that have different and distinct strontium signatures. We have some pilot data from a BU undergraduate honors thesis that I directed a few years back, and we recently received NSF funding to expand this project to a much larger body of samples over the next three years. The pilot data comes from three different time points of five animals, so it really is a tiny data set. They're sheep in yellow and goats in blue, but what we found, even from this tiny data set, was clear. Evidence that animals came from distinct geological areas, especially the sheep, and that animals seemed to have moved between areas over the course of their first year of life. That is, not all of these lines are perfectly horizontal. At least in one case, two animals converged on the same geology later in life as evidenced by changes in these two samples. In Adam's thesis, he also applied predictive GIS modeling to identify potential movement paths through the landscape. In our larger study, we're going to combine this approach to identify routes of mobility between the areas that we can identify isotopically and to link farming and herding patterns with geologies that show evidence for erosion. Now that we can tell where the animals moved, we question where are the fields. We do have some data already from wood charcoal that suggests that the frequency of trees that are found in open woodlands, which have to be cut to clear land for new agriculture, increased through the Persian period, and then dropped off to almost nothing in the Roman period. So by the Roman period, the Romans didn't need to clear any new land for farming, at least within the fuelwood catchment of Gordian. I have made some ecological suppositions about the location of fields based on this and related evidence as well as that erosional evidence, but how spatially diverse was farming, especially as herding location varied. It's possible to use stable isotope analysis of seeds to help with this. So within plants, carbon varies with water stress and nitrogen with fertilization practices. These signatures are present in the seeds that are produced by the plants. The combination of carbon and nitrogen analyses allows us to identify the diversity of fields that contributed to a single cache of seeds. A wide dispersion of values indicates that fields were in multiple locations and that diverse farming methods were used, while a narrow clustering of values indicates that fields were treated similarly. Our pilot data from barley grains show that during two periods of political centralization, the Phrygian and the Hittite periods, we find values relatively tightly clustered, while in the intervening periods of decentralization, the early Iron Age and the early Phrygian period, values are much more dispersed, indicating a greater diversity of farming practices. Our NSF award will enable us to include wheat, as well as a much larger dataset for both cereal grains, which will allow us to compare agricultural strategies for these two crops across all periods of settlement. So by incorporating these methods in a few years, I hope to be able to share with you a much more complete record of agricultural adaptation to imperial expansion and climate change across the eastern Mediterranean. This would not have been possible without my sponsors and partners, so I thank them and I thank you for your attention. Of course. Did you have some basal data to link these conditions that you're promoting for the carbon in their location? I mean, these are archeological. So do you have modern that are somehow allowing you to place their strategy as well as their geology? Yes. We have some general models that have been produced in prior research that was done across broader areas of the Mediterranean world to look at effects of different fertilization practices and different rainfall amounts or irrigation values that established that there seems to be a fairly consistent set of values that fall along these extremes from no fertilization to extreme fertilization. So it doesn't look like we actually need local proxies. We can actually just plot these into the general, more or less, values. What we see here is that this represents a fairly broad range of carbon values, which indicate water stress, and relatively low values of nitrogen, so at most moderate fertilization. Again, this is for the seeds. We're going to have 1,000 seeds probably by the end of this project, so hopefully we'll have a much broader data set to be able to sort these out first. Same thing about the taxation regime in the Roman period. It gets me early mid-empires, people are in Ossia, and I think the land tax is one-tenth tax in kind. Is that something that you or your colleagues have explored? Yeah. And so I gather that they're grain and lullaby than it is taxing kind, or you want to expand on that? Sure. So we don't have a ton of great data from central Anatolia. Historically speaking, textually speaking, we have some indications that the land holding system is, and the state is kind of latitude system. We have interpretations that likely would have been sort of a local tax bidding system on the provincial level as we have in other areas of the empire, and thus that different regions would be sort of targeted and potentially assigned specific crops that they would be required to pay in taxes. The inference is that that might lead to kind of monoculture for both the agricultural and the pastoral side, but we have to also overlay that with the fact that clearly Bennett is right that there is some form of military procurement that is operating independently of a taxation system here. The folks who live there are probably also enforcing the tax collection. So we definitely have two different systems that we're seeing, and so this is my best interpretation about it, but according to the historical data, we don't necessarily have anything that says clear cut is what you're describing sort of from the records in Austria. Yeah, Anatolia is not great for rolling through the data. I mean, it really is just kind of a black box because we meant a ton of local textual elements. I wanted to ask about something you talked about earlier on the talk in respect of overgrazing. So if I understood correctly, your measure for overgrazing was from data sets that you're collecting from Don and you used as fuel. So I was just wondering how the utilization of different species like cattle, all which raise very different distances and very different things versus what's represented in the Dung Fuel record would affect your assessments of overgrazing. No, it's a great question, and there's a couple of reporting clarifications all that now about how that works. So the first is that you don't use pig feces as a source of fuel. It's too wet. But you can use both cow and sheep and goat and we tend to be based on what we know ethnographically. There's separate systems for processing sheep versus caprid duck into these cakes. They certainly move different distances across the landscape, but we have to keep in mind that what we're seeing is a local Dung record is probably a much narrow catchment than the actual grazing activity of these animals. So if we have something which hopefully with the strontium isotope data we'll be able to say, transhuman pastoralism where you say, take the cattle up into the mountain pastures during the summer, none of that Dung is coming back to the site. I guess it's conceivable, but it's unlikely that they would be bringing Dung back to the site through those distances. And so we're going to, that portion of the animal diet is going to become visible in the proxy record that we have. So what we're looking at is probably a fairly narrow strain, 5, 10 kilometer, maybe boundary around the site of Orion, and that's the area that we're getting our data from during every time period. So because of this minimal catchment zone or maximal catchment zone, I think that actually the data are fairly well comfortable over time and I think the sheep, goat and cattle are not going to be too different because within that catchment zone, they're going to be eating relatively similar things. Now the goats are going to be eating potentially some plants that can't be eaten by some of the others and the limitation of that I think is overcome by the specific species that I chose to use in this metric. So some kind of wood ear plants are things that might be eaten by goats, but not by a sheep, but they don't have seeds that show things in this proxy. So I think it's fairly well controlled for that, but we can never be 100% certain. Do you have an idea of the relative representation of sheep and goat fighting versus cattle? I have absolutely no idea. I think I was just thinking maybe even your estimates of overgrazing are under-representing the amount of overgrazing. If sheep and goats will eat almost anything are not being as well represented in the dump record? Yeah, so my supposition is that you can see more sheep and goats, but that's based on certain anecdotal data, which is that when the sheep, dragons, and goats can go dumb, but that's because you don't see dragons and cattle. It's an invisible thing. So that's my inclination, but I don't actually have data that can support that. We could look at the animal bone counts and say, okay, maybe it's somewhat correlated with us. Who knows? Who knows? It's an important point. This overgrazing record is not an overgrazing proxy calculation. It's not something that I would necessarily use to compare one site with another site because they're doing such ecological differences with the areas, but I think it works very well to track sheep over time in one particular location. Yeah, thank you. That's a great question. Anything else? Mark, is barley preferred? So I think barley is probably only preferred in conditions where you can't grow enough wheat, but that being said, even in cases where we have very wet areas and people grow lots of wheat, they grow some barley, and typically in those conditions barley is reserved solely as animal feed and for deer brewing. And we know they're brewing deer based on chemical data from the site, so we know that some portion of the barley is going into that. Following us on that, why do you say that they're not eating barley? Is there some kind of documentation that says that they're not making the barley stew? Yeah, so I'm certain they are eating barley, and I'm certain they're eating quite a bit of barley at Gordian, especially during a bad year. We lack the resolution to see that in these horse-grading data sets. One reason I'm hoping for this very fun-granging, paleoclimate data set is that it might actually allow us to say that there might be certain periods of time that would be more variable or more drier than other periods of time, and that barley would probably become a greater part of the diet. A lot of it's based on ethno-archaeological parallels of the work of Jones and Paul Halstead and looking at sort of how, in this case, Greek farmers kind of managed this and they grew it together and then they would sieve them out. So in a good year they'd eat the more wheat-heavy fraction, and in a bad year they'd eat the more barley-heavy fraction and not use it for animal feed. So we can revision something like that happening, but there's no actual evidence So when you say they were brewing wheat barley and they were giving it to animals, that's a supposition based on just your ethno-archaeological work. So it would be safer to say they're eating wheat and barley and drinking barley and feeding it to animals. Yes, that's true. They are absolute, the one thing we don't have any primary evidence for is are they eating wheat and barley directly or in the clouds or something like that? That's the thing that if you can't see I've never found anything that I thought was bread. So we know they're brewing because of the chemical data and we know they're feeding barley and in some cases, chaff to animals because you find that in the dung pallets. So that's the only primary evidence we have. Yes? Have you been to the hospital at all for a very long period of time? Yeah, it's a great question. We have very few variables from this item. So do you have the big dream variables that were excavated in the 1950s when they threw out bones? They did keep the stone with King Lydus but he's in the Museum of Ethnology. King Lydus. So unfortunately, he's there. We have some Roman period graves that were excavated later. The bones are preserved and we have part of the Hellenistic period that actually counts living at the site and they were doing some really creepy, even sacrifice stuff. So we have some heads on top of animal bodies and stuff like that. But we have some of those skeletal remains as well but they're all in Turkey and none of them are currently in existence. It would be great to do it. But for the animals? Oh, for the animals. Then I'll count those in the animals. It's actually a great idea and it's something I would love to have the capacity to do at some point. I am sure we can reach so many animal bodies that it would be easy enough to find some of the head calculus. So it's a great idea. Okay, thank you.