Our project attempts to understand how the disintegration of the Roman Empire and the formation of new early medieval states influenced diet and economic practices. A team from Germany compiled available historical, archaeological, and ecological information on the economy, while a team from Poland generated a new, comprehensive set of isotopic data from regions ranging from the Italian Peninsula in the west to the Caucasus and the Levant in the east. 
Analyses of stable isotope ratios of various elements in bone collagen and the mineral fraction of human and animal bones and teeth are a standard method for reconstructing the diet, economy, and mobility of ancient human groups. Carbon isotope ratios (δ13C) vary among plants with different photosynthetic pathways and also allow for distinguishing between marine and terrestrial dietary components. Nitrogen isotope ratios (δ15N) indicate the contribution of plant and animal proteins (especially fish) to the diet and are also sensitive to differences in fertilization levels. Sulfur isotope ratios (δ34S) vary between terrestrial and marine dietary components and also exhibit variability between different locations. Finally, strontium isotope ratios (87Sr/86Sr) are an indicator of human mobility and food imports. 

he research hypothesis assumed that the collapse of the western half of the Empire and the profound crisis in the eastern half caused the disintegration and regional fragmentation of the previously relatively homogeneous economic system. However, our results showed that such fragmentation did not actually occur, as even during the Imperial era, diet and food economy were largely local in nature, and mass food imports could only have affected the capitals of the empire and populations with high economic status. 
We also confirmed a relatively high share of millet in the diet as early as Roman times, but primarily in the Balkans. In the Early Middle Ages, this share remained locally high, including at sites in southern Austria and Slovenia. Individual regional analyses showed that dietary variability was most often greater between regions than between chronological units, although very interesting changes over time were also observed. For example, in central Anatolia, agriculture was more intensive during the Roman Empire, with a growing share of pastoralism. Later it became more extensive, with significantly lower levels of fertilization. 
At the Collegno site (northern Italy), isotopic data obtained as part of the project were compared with the results of ancient DNA studies (HISTOGENES project), which allowed us to determine the degree of dietary similarity between related individuals and to link the arrival of a new population at the site with a dietary shift involving a reduction in the use of millet. This combination of genetic and isotopic methods opens up new possibilities for understanding the history of ancient human groups, even at the scale of individual families. 

The main objective of the project was to collect available historical, archaeological and ecological information on diet and subsistence in the Central and Eastern Mediterranean during the “Fall of Rome”. The main source was to be a new comprehensive set of isotopic data from various regions, from the Apennine Peninsula in the west to the Caucasus and the Levant in the east, used to test the hypothesis that the fall of the western part of the Empire and the deep crisis in the eastern part of the Empire caused the collapse and regional fragmentation of a previously fairly homogeneous economic and subsistence systems. The results of the project showed that such fragmentation did not actually occur, as even earlier, the diet and food economy were mainly local in nature, and large-scale food imports could only have affected the capitals of the empire and groups of people with high material status. The original list of archaeological sites that we planned to include in the project has undergone significant changes. It turned out that obtaining permits to collect and export samples from Sicily and Greece was practically impossible, but instead we managed to collect samples in northern Italy, on both sides of the Karawanken Mountains, and in Turkey. Cooperation with Turkey also required an increase in financial expenditure on the analysis of carbon and nitrogen isotope ratios in collagen (δ13C, δ15N) at the expense of other analyses,

and the use of DFG funds (direct and overheads) for isotope analyses. This is due to the Turkish government's policy, which does not allow the export of samples but allows biochemical analyses to be carried out in a laboratory in Istanbul, which, as a monopoly, imposes high prices. We also completely abandoned the analysis of light isotopes (δ13C, δ18O) in biophatite; they make more sense in arid areas where collagen is poorly preserved or not preserved at all, but in regions where collagen extraction was not problematic, the traditional set of light isotope measurements in collagen (δ13C, δ15N, δ34S) offered greater interpretative possibilities. 
As planned, measurements of strontium isotope ratios (⁸⁷Sr/⁸⁶Sr) were carried out to identify mobility patterns and identify the high proportion of imported food in the diet. The data collected enabled the creation of detailed regional maps of strontium isotope ratios covering part of northern Italy (Liguria and Piedmont), the upper Drava (Carinthia) and Sava (northern Slovenia) river basins, a large part of the lower Danube valley, especially Dobruja, almost all of Lebanon (except for areas controlled by Hezbollah) and part of Georgia. These maps fill important gaps in existing knowledge about the variability of 87Sr/86Sr values in Europe and south-western Asia. Various data from Egypt were also obtained during the project. Although it was not possible to carry out isotopic studies there, a mass grave in Marea, preliminarily dated to the time of the first plague pandemic, provided samples for aDNA analyses (we are still waiting for permission to analyse them in a laboratory in Cairo), but also classical osteological indicators of diet (enamel wear patterns, caries frequency, skeletal indicators of vitamin deficiencies). A collection of non-metric dental features was also gathered in Egypt, which can be used to reconstruct mobility patterns. A marginal effect of the project is the creation of a simple application in R for statistical analysis of non-metric features. 
The project was completed with a delay resulting primarily from the Covid pandemic; this significantly delayed not only the process of obtaining samples from local institutions, but also the process of obtaining export permits. Coordinating a project involving the collection of samples from nearly 50 archaeological sites located in different countries and the analysis of these samples in several laboratories was a major logistical challenge, but we ultimately managed to obtain a data set of the planned size, and even more diverse than anticipated. Some of the project results have already been published, and further articles are at various stages of preparation (at least four will be submitted for publication in 2026). Given the pace of work to date, we expect that all project results, including a synthesis article, will be submitted for publication by the end of 2027. 
The results obtained in northern Italy (Liguria and Piedmont) have been largely published and formed the basis for Sarah Defant from the German project team to obtain her PhD. At the Collegno site, isotopic data obtained as part of the project were compared with the results of ancient DNA research (the HistoGenes project funded by the ERC), which made it possible to determine the degree of similarity in diet within individual families and to link the appearance of a new population at the site with a change in diet consisting of a reduction in the proportion of millet. At the Ventimiglia site, significant differences in diet between women and men and a certain level of regional mobility were observed. At the Frascaro site, representing the Ostrogothic population, we found a low level of mobility. 
Results from southern Austria (Carinthia) and Slovenia are currently being prepared for publication. In two regions (the upper Drava basin and the upper Sava basin on both sides of the Karawanken mountains), significant diachronic differences in diet between late antiquity and the early Middle Ages have been observed, which, however, go in opposite directions in terms of the proportion of millet and animal protein. This is probably the result of social and economic changes in the early Middle Ages, resulting from interactions between Slavic groups and the Frankish state expanding towards Carinthia. 
A publication on sites from the lower Danube (Romania and Bulgaria) has been submitted for publication. The collected data set allows for a comparison of dietary patterns in different ecosystems (mountains, river valleys, coast); both during the Roman Empire and in the early Middle Ages, there was a tendency to use primarily local resources. The use of δ34S values was very important in this context, as they proved to be clearly differentiated locally, with low variance within sites and high variance between sites. 
The results from Croatia indicate a diversified economy in the coastal zone, best seen at two late antique sites on the Adriatic Sea. Vir-Smratina is characterised by consistent use of marine resources, while in Podvrsje a significant proportion of individuals based their diet mainly on terrestrial resources, which fits into the broader pattern of irregular use of marine fish in Late Antiquity. These sites also differ in terms of mobility: in Vir, a homogeneous distribution of strontium isotope ratios is observed, indicating the local origin of the population, while in Podvrsje, two distinct isotopic groups are identified, one of which most likely represents migrants. Against this background, early medieval Dubravice shows similarities to Podvrsje in terms of both diet and mobility patterns. 
Analysis of samples from Butrint (Albania) indicates a diet based primarily on terrestrial resources, both plant and animal, with a small proportion of millet, fish and other marine resources. Strontium isotope data for the same sample also indicate relatively low mobility of the population studied. This is a surprising result, given the broad chronological horizon of the sample, covering the period from Late Antiquity to the Middle Ages, with a few individuals from the modern period, and yet characterised by a distinctly local isotopic profile, consistent use of terrestrial resources, avoidance of marine resources, and limited mobility. The migrants present in the sample come from different geological areas, suggesting rather individual and diverse cases of mobility than mass migration or settlement by a homogeneous group. This is particularly relevant in the context of the demographic changes in Albania documented by historical sources. We aim to publish these results in 2026 together with aDNA carried out on the same individuals in MPI EVA and the new palaeoecological data (high-resolution pollen profile) from Lake Butrint. 
The results from Turkey will be submitted for publication soon. In central Anatolia, we observed very interesting diachronic changes in diet/food procurement methods. Around the 4th century CE, the proportion of animal protein in the diet increased (especially among men), which may indicate a higher proportion of pastoralism in the local economy. Then, between the 6th and 8th centuries, there was a marked extensification of agriculture, which was probably the result of political changes during that period. Data from Georgia and Armenia are not as extensive, but no similar pattern of change can be seen there. 
The results of research conducted in the Levant will also be submitted for publication within a few months. The most important data comes from three sites in Lebanon (Jiyeh, Chhim, Barja), which are located close to each other and have similar dating (4th-7th century). During the Roman period, pigs were the main source of animal protein, and the consumption of sea fish left an isotopic trace only at the port site (Jiyeh). 
The most problematic region included in the project is Egypt (with comparative data from Sudan). Firstly, collagen is poorly preserved in dry climates, and δ13C, δ15N, δ34S measurements and radiocarbon dating are problematic. Secondly, Egyptian policy makes it very difficult to export samples, while the local infrastructure does not ensure the reliability of isotope results. Nevertheless, we decided to include several sites in the project, hoping that isotopic measurements could be performed during planned genetic studies (Marea) or using skeleton collections located in Europe (Wadi Qitna). In Egypt, we also managed to collect data for the analysis of non-metric dental features, which can be used in mobility analyses. In parallel to the generation of the isotopic data, the team in Berlin/Jena (PI Adam Izdebski and postdoc Sabina Fiołna) worked on the collection of large-scale datasets for ecological and economic change. For the first one, Adam Izdebski with the support of the broad network of colleagues, we collected published and unpublished high-quality pollen data from 6 regions in our study area (Central Italy, Sicily, Southern and Northern Greece, SW Turkey, Israel) and several more in the West, which we will use in the final synthesis of the project results. For the economic change, Sabina Fiolna focused on numismatic data, both those collected by herself (published in the Science Advances paper, with a follow up in the Journal of Quaternary Science) and those provided by the FLAME project at Princeton. As a result, Sabina Fiolna was able to analyse over 65,000 records from around the Mediterranean Basin from the period 4th-9th cent CE with the use of Social Network Analysis. The material collected in the FLAME database is rich and numerous, but its quality is very uneven geographically and chronologically. We identified a particularly acute gap in the good-quality numismatic data from the area of modern Turkey, the heart of the Eastern Roman Empire. Despite the number of excavated Byzantine coinage in Turkey, the  material turned out unsuitable for Social Network Analysis. On the other hand, the Social Network Analysis of the whole body data have shown unequivocally the importance of Constantinople in the analysed period. The analysis has also shown the increasing (but not complete) disintegration between Eastern, Central and Western Mediterranean monetary systems and a shift in the economic importance of different regions connected to the historical events, such as the flourishing of Visigothic Iberia. We plan to include the results of these analysis in the synthesis article of the project, together with the pollen and isotopic data. 
In summary, our findings change the existing picture of diet and economy in the Roman Empire. We show that the diet before, during and after the fall/crisis of the Empire was based mainly on local resources, although there were also some general patterns (for example, the important role of pigs as a significant source of animal protein). We also point to the low proportion of marine fish in the diet, despite the popularity of the fish sauce garum. These results entirely falsify our initial hypotheses formulated in the project proposal, based on the broad consensus of the scholarly literature as of the year 2019. 
In our research, we show how important it is to link biochemical studies of human remains not only to the archaeological context, but also to historical and ecological sources. This provides broader interpretative possibilities, including the use of Bayesian modelling, an example of which we showed in our article on Ventimiglia. We expect that the results of our research will reach not only bioarchaeologists who deal with isotopic analyses on a daily basis, but also historians debating the causes and effects of the fall of the Roman Empire. The raw data from the project (isotopic measurements, regional maps of strontium isotope ratios) will be available for use in future cross-sectional isotopic studies. We also anticipate widespread use of the application for analysing non-metric dental data, as well as the published sampling protocol for combined genetic and isotopic studies. 

•    Description of the handling of research data generated in the project: as the publication of the data progresses, they are uploaded to the standard public repository of archaeological science, that is the IsoMemo network at Pandora (https://pandoradata.earth/dataset/?groups=isomemo-group&page=1).
•    Description of any research data, methods, standards, software or infrastructures generated in the project that are re-usable and openly accessible to others: as part of the project, we generated a protocol for multi-proxy sampling of teeth for Sr and aDNA in particular (see point 4.2). 

Implementation of scientific events, science communication measures: together, Polish and German team members delivered more than 20 talks at different international seminars and conferences in the course of the project. We also organised two workshops. The first one took place online in 2020 at the Warsaw’s Department of Archaeology annual international conference (it was a project launch, with a large number of invited collaborators responsible for sites that we included in the project). The second one was organised on July 10-11, 2024, at the Free University in Berlin and was devoted to discussing project results together with a number of external collaborators (workshop title: First millennium socio-environmental change in Western Eurasia).

Category A – Articles in peer-reviewed journals, contributions to peer-reviewed conferences or to anthology volumes, and book publications

Defant, S., Derda, T., Borowski, T., & Sołtysiak, A. (2025). Human remains from 
‘Marea’/Philoxenite, Egypt, 2023. Bioarchaeology of the Near East, 19, 105-109. http://www.anthropology.uw.edu.pl/19/bne-19-06.pdf Open Access 
Defant, S., Giostra, C., Micheletto, E., Garbarino, G.B., Izdebski, A., Sołtysiak, A. ().

Isolation and Identity in a 5th/6th century Ostrogothic Rural Community in Frascaro, Italy. Prähistorische Zeitschrift. https://doi.org/10.1515/pz-2025-2034  

Defant, S., Carabia, A., Fetner, R., Craig-Atkins, E., Fernandes, R., Martino, G.P., Costa, S.,

Sołtysiak, A., Izdebski, A. (2025). Isotopic data reveal a localist Roman population in late Roman Albintimilium, Liguria. Scientific Reports, 15(1), p.12097. https://www.nature.com/articles/s41598-025-92851-7 Open Access 

Tian, Y.*, Koncz, I.*, Defant, S.*, Giostra, C., Vyas. D.N., Sołtysiak, A., Pejrani Baricco L., Fetner, R.A., Posth, C., Brandt, G., Modi, A., Lari, M., Vai, S., Francalacci, P., Fernandes, R., Steinhof, A., Pohl, W., Caramelli, D., Krause, J., Izdebski, A., Geary, P.J., Veeramah, K.R. (2024). The role of emerging elites in the formation and development of communities after the fall of the Roman Empire. Proceedings of the National Academy of Sciences, 121(36), p.e23178. https://www.pnas.org/doi/10.1073/pnas.2317868121 Open Access 

Doğan M., Fiołna S., Duman B., Eastwood W., Haldon J., Izdebski A., Şenkul Ç., Environmental change and globalization dynamics in Roman Anatolia: Stabilizing an accelerating system. Science Advances 11, eadt7107 (2025). https://doi.org/10.1126/sciadv.adt7107 Open Access 
Fiolna S., Doğan M., Eastwood W., Haldon J., Liakopoulos G.C., Şenkul D., Şenkul Ç.,

Izdebski A., Imperial systems and local landscapes of Buldan Yayla in Western Anatolia (Türkiye) during the last 4000 years: An integrated palynological, historical, and archaeological approach. Journal of Quaternary Science 40 (2025), 1285-1304. https://doi.org/10.1002/jqs.70008 Open Access 

Category B – Any other form of published results

Defant, S., Ingrová, P., Zlámalová, D., Sołtysiak, A., & Hofmanová, Z. (2025). Multi-Proxy

Sampling Protocol for Teeth from Archaeological Collections. https://www.protocols.io/view/multi-proxy-sampling-protocol-for-teeth-from-archa-bp2l69o8rlqe/v1 

Pickel, D., Fetner, R. Ailincăi, S.-C., Lemke, M., Nuţu, G., Russeva, V., Sarantis, A., Vasile, G., Izdebski, A., Soltysiak, A. (2026) Dietary change in the face of societal transformation? Isotopic analysis of Carbon, Nitrogen, and Sulfur from the Lower Danube region during the Late Antique to Early Medieval periods. http://dx.doi.org/10.2139/ssrn.6108796  (under review in the Journal of Archaeological Anthropology)

The above table provides background information to the map in the public part of the report. It contains information on the sites we studied and the results we obtained, together with the publication status for each site. 
As is clearly visible from this table, our project was massive in scale and presented a huge logistical challenge. The table does not include, in fact, some two dozen more sites in Italy, Greece, Turkey, Tunisia, Jordan, and Israel, which we also tried to obtain. Each of these sites required ongoing contacts – online calls, numerous emails, meetings – with local archaeologists over prolonged periods of time, and time-consuming permit procedures that in some cases failed at an advanced stage. All this effort – except for Northern Italy, where Sarah Defant focused on her PhD project – was the responsibility of the project PIs and kept them very busy over the course of the project. This demanding logistics posed a major challenge for our project, and slowed it down considerably, in addition to problems caused by the pandemic in the first years of the project. These challenges can also be visible in our publication outcome: while we achieved several high-impact publications in international multidisciplinary journals, all these publications could happen only because team members other than the PIs were the lead authors for each of them and these junior team members could bring these studies relatively quickly to publication. In fact, we even managed to obtain additional funding on both sides (from the Max Planck in Germany, from the Excellence Initiative in Poland) to either prolong contracts of existing team members (Sarah Defant, Sabina Fiolna) or even higher a new postdoc (David Pickel), in order to remedy this problem. Now that the data collection phase of the project is finally almost over, the publication of the remaining sites will proceed quicker, as this could become the main focus of the project PIs. In this sense, while we managed to achieve all of our goals, our initial planning was not realistic enough and we should have planned to have a larger group of 2/3-year postdocs (like David Pickel) in the final stage of the project to actually work on the interpretation and publication of the obtained data, and perhaps one more PhD student working on a coherent region (e.g., the Balkans). In fact, in our planning we counted on the local archaeological partners to lead publications of site-specific results; this, however, turned out to be unrealistic. Not being isotopic specialist and having their own busy agendas, these collaborators could (most of the times) provide the necessary contextual information, but none of them so far took a leading role in any publication (in this respect, our collaboration is most fruitful in Turkey).  
For the reasons of time, given the challenging logistics, we also gave up on the technically challenging and bearing higher risk analysis of isotopes in single amino acids, initially planned at MPI SHH/GEA. This freed up significant budget that – as we already informed the DFG - we redirected towards enriching the study of the Balkans with sedaDNA analysis of lake and terrestrial cores from Lake Dojran and its catchment (at the border of Greece/N Macedonia). This area lies just south of our study sites in the Lower Danube and understanding local biodiversity and agricultural dynamics in the Roman-Medieval period will help us contextualise the results of isotopic analysis. This way, the Balkans and Anatolia come out as the core parts of the Eastern Roman Empire where we manage to combine ecological (sedaDNA, pollen analysis) and bioarchaeological approaches (isotopes on human and animal remains).