MAPPING TRADE BY SHIPWRECKS: AN ANALYSIS OF AMPHORA ASSEMBLAGES FROM FOURNOI ARCHIPELAGO (AEGEAN) By Konstantinos Raptis May 2024 Director of Thesis: David Stewart, Ph.D. Department of History, Program in Maritime Studies ABSTRACT This thesis examines nineteen amphorae from fourteen different shipwrecks, documented in the Fournoi Archipelago of the east-central Aegean, dating to the Late Roman period (fourth to seventh centuries CE). Both the macroscopic observations and typological study of the raised amphorae indicate a variety of provenances. While some of the artifacts are in good condition, others are less well preserved. The shipwrecks' significance lies particularly in the cargo’s heterogeneity albeit with chronological range. Despite their distinct geographical production areas, these amphorae ended up on the same vessels’ decks, revealing complex economic activities. Furthermore, these sunken vessels in Fournoi not only highlight the importance of the island as a vital sea lane but also provide valuable research on the trade and connectivity of various scales and the economic implications of material culture. Many of the Fournoi shipwrecks, whose discovery is not more than 10 years of age, have yet to be studied more closely. Even though the amphorae forms have enjoyed significant so far research, what is missing is an overall understanding of their strong economic ties. The present study seeks not only to put light into our understanding of the newly discovered shipwrecks but also to view the Fournoi Archipelago as a microcosm for studying economic maritime activities during the Late Roman period in the eastern Mediterranean. Starting from the Fournoi archipelago, this study aims to provide information about these tradable commodities that were exported far beyond this island in Late Antiquity. At the same time, it intends to analyze, reevaluate, and interpret distributional patterns, and commercial activities, utilizing pottery as an interpretive key, Keywords: Shipwreck Archaeology, Mediterranean, Late Antiquity, Network Analysis, Trade, Connectivity, Late Roman Amphorae Mapping Trade by Shipwrecks: An Analysis of Amphora Assemblages from Fournoi Archipelago (Aegean). A Thesis Presented to the Faculty of the Department of History East Carolina University In Partial Fulfillment of the Requirements for the Degree Master Arts in Maritime Studies By Konstantinos Raptis May 2024 Director of Thesis: David Stewart, Ph.D. Thesis Committee Members: David Stewart, Ph.D. Helen Dixon, Ph.D. Nathan Richards, Ph.D. ©Konstantinos Raptis 2024 DEDICATION This thesis is dedicated to two amazing people, my uncle Vassilis and my friend Brett, who both passed a year ago. ACKNOWLEDGEMENTS This thesis was only possible because of a vast network of family, friends, colleagues, and professionals to whom I must express my heartfelt gratitude. First, I would like to thank Dr. David Stewart. Your unconditional support, patience, and infinite advice have played a pivotal role in shaping this thesis and my graduate journey. I would also like to acknowledge my committee members, Dr. Helen Dixon and Dr. Nathan Richards for the fruitful discussions and aspirations. In the face of my committee members, I found the passion of true scholars who highly inspired me to continue my academic journey. On the other hand, I want to thank my Greek colleagues. My Fournoi team with their continuous support provided all the necessary data for my current research. Within the team, I want to special thank Dr. George Koutsouflakis and Dr. Peter Campbell for their continuous support and assistance, but also the conservationist Angelos Tsopanidis for letting me conduct a direct inspection of the amphorae at the Greek Emphorate of Underwater Antiquities in Athens. Furthermore, I want to thank Fulbright for providing me with financial support to cover my expensive master’s studies. I also want to express my gratitude to my friends and loved ones, who made me feel like I found a new family in the U.S., and their love and energy played a vital role in this amazing journey. Finally, I am wholeheartedly thankful to my parents. Without their support, I would not be currently authoring this thesis. Table of Contents DEDICATION ........................................................................................................................ iii ACKNOWLEDGEMENTS .................................................................................................. iv List of Tables ............................................................................................................................ x List of Figures ........................................................................................................................ xii Chapter 1. Introduction ........................................................................................................... 1 Research Questions ................................................................................................................ 4 History and Archaeology of Fournoi ...................................................................................... 5 The Maritime Cultural Heritage of Fournoi: Navigation and the Role of the Island’s Geography .............................................................................................................................. 9 Thesis Outline ...................................................................................................................... 15 Chapter 2. The Two Late Roman Amphorae Types: LRA1 and LRA2 ............................ 17 The Late Roman Amphorae ................................................................................................. 17 The case of the Late Roman Amphora 1 (LRA1): ............................................................... 18 Scholarly Community: ...................................................................................................... 18 Typology: .......................................................................................................................... 19 Distribution: ...................................................................................................................... 24 Workshops: ....................................................................................................................... 24 The South-East Mediterranean Cluster ................................................................................ 25 The Aegean and Black Sea Cluster ...................................................................................... 27 The Case of the Late Roman Amphora 2 type (LRA2): ...................................................... 28 Scholarly Community: ...................................................................................................... 28 Typology: .......................................................................................................................... 28 Distribution ....................................................................................................................... 31 Workshops: ....................................................................................................................... 32 A Final Word for the Major Workshops of Both Types ........................................................ 33 Discussing Amphorae Contents ........................................................................................... 34 Epigraphy ............................................................................................................................. 35 Chapter 3. From Production to Distribution, and the Complexity of the Late Roman Trade on the East, Through the Study of Maritime Activities and Amphorae Containers (LRA1 and LRA2).................................................................................................................. 36 An Overview of the Reorganization of the Roman East ...................................................... 36 The Annona System .......................................................................................................... 38 Quaesture Exercitus .......................................................................................................... 40 The Two Late Roman Amphorae as Evidence for the State-lead supply system ................. 40 Standardization ................................................................................................................. 42 Shipwreck Archaeology and Amphorae ............................................................................... 44 Local, Regional, and Interregional scales ............................................................................ 46 Scales of Complexity, and the Late Roman “Mosaic” ......................................................... 51 “Marketing” ...................................................................................................................... 52 “Redistribution” ................................................................................................................ 53 Either or, or Maybe Both? ................................................................................................ 55 Chapter 4. Amphorae Data: Methods and Analysis ........................................................... 58 Introduction of the Amphorae Samples ................................................................................ 58 Methodological Approach for Analyzing the Samples from the Fournoi Archipelagos ...... 59 Fabric Analysis ................................................................................................................. 60 A Guide on How to Read the Following Pages .................................................................... 62 Catalog of the Fournoi Archipelago Amphorae Finds ......................................................... 63 Amphora 1/1 BE: 2015/5-52 ............................................................................................ 63 Amphora 1/2 BE: 2015/5-53 ............................................................................................ 65 Amphora 2/1 BE: 2015/5-2 .............................................................................................. 67 Amphora 2/2 BE:2017/10-4 ............................................................................................. 69 Amphora 2/3 BE:2017/10-5 ............................................................................................. 71 Amphora 2/4 BE:2017/10-21 ........................................................................................... 72 Amphora 5/1 BE: 2015/5-7 .............................................................................................. 74 Amphora 6/1 BE: 2015/5-8 .............................................................................................. 75 Amphora 8/1 BE: 2015/5-43 ............................................................................................ 77 Amphora 10/1 BE: 2015/5-26 .......................................................................................... 78 Amphora 10/2 BE:2015/5-27 ........................................................................................... 80 Amphora 14/1 BE:2015/5-35 ........................................................................................... 81 Amphora 19/1 BE:2015/5-55 ........................................................................................... 83 Amphora 24/1 BE: 2016/4-10 .......................................................................................... 84 Amphora 25/1 BE:2016/4-13 ........................................................................................... 86 Amphora 42/1 BE: 2016/4-85 .......................................................................................... 87 Amphora 44/1 BE: 2016/4-106 ........................................................................................ 88 Amphora 47/1 BE: 2017/10-2 .......................................................................................... 90 Amphora 53/1 BE: 2017/10-45 ........................................................................................ 92 Chapter 5. An Overall Synthesis of Shipwreck Data, Land Sites, and Network Analysis .................................................................................................................................................. 95 Results of the Sample Analysis from the Fournoi Archipelagos .......................................... 95 Results of the Various Types and Subtypes ...................................................................... 95 Results of the Chronological Analysis.............................................................................. 96 Fabric Analysis Results .................................................................................................... 97 Results of the Parallels and Provenance Hypothesis ........................................................ 99 Analyzing Data through Networks ..................................................................................... 100 Social Network Graphing of Shipwrecks and their Components, Utilizing the Gephi Platform: Data Preparation ............................................................................................. 101 Making Sense of Shipwreck Distribution and Production Centers: An Overview and Patterns ............................................................................................................................... 110 Looking Beyond Shipwrecks, Circulation, and Consumption of Large Areas and Major City Actors.......................................................................................................................... 114 Asia Minor ...................................................................................................................... 115 The Danubian Provinces ................................................................................................. 117 Conclusions ........................................................................................................................... 121 Future Research .................................................................................................................. 126 A Final Word ...................................................................................................................... 127 Works Cited .......................................................................................................................... 128 APPENDIX A: Production centers of LRA1 and LRA2 .................................................. 143 APPENDIX B: A guide for ceramic past inclusions ......................................................... 147 APPENDIX C: Fabric section of the Fournoi Shipwrecks .............................................. 148 APPENDIX D: General catalogue of shipwrecks found in the Eastern Mediterranean with LRA1 and LRA2 .......................................................................................................... 150 APPENDIX E: Comparative dates for APPENDIX C catalogue .................................... 153 List of Tables TABLE 1. Recovered Amphorae Measurements SW1/1 BE:2015/5-52, 1/2 BE:2015/5-53. Here and below, H/PH (preserved height), PMD/MD (preserved maximum diameter), ERD (rim – exterior- diameter), NH (neck height), in centimeters; empty weight (EW) in kilograms and Color Munsell (CM). ......................................................................................................... 63 TABLE 2. Recovered Amphorae Measurements SW2/1 BE:2015/5-2, SW2/2 BE:2017/10-4, SW2/3 BE:2017/10-5, SW2/4 BE:2017/10-21. ....................................................................... 67 TABLE 3. Recovered Amphorae Measurements SW5/1 BE:2015/5-7. ................................. 73 TABLE 4. Recovered Amphorae Measurements SW6/1 BE2015/5-8 ................................... 75 TABLE 5. Recovered Amphorae Measurements SW8/1 BE:2015/5-43. ............................... 77 TABLE 6. Recovered Amphorae Measurements SW10/1 BE2015/5-26, 10/2 BE2015/5-27. .................................................................................................................................................. 78 TABLE 7. Recovered Amphorae Measurements SW14/1 BE:2015/5-35. ............................. 81 TABLE 8. Recovered Amphorae Measurements SW19/1 BE:2015/5-55. ............................. 82 TABLE 9. Recovered Amphorae Measurements SW24/1 BE2016/4-10. ............................... 84 TABLE 10. Recovered Amphorae Measurements SW25/1 BE2016/4-13. ............................. 85 TABLE 11. Recovered Amphorae Measurements SW42/1 BE:2016/4-85. ............................ 87 TABLE 12. Recovered Amphorae Measurements SW44/1 BE:2016/4-106. .......................... 88 TABLE 13. Recovered Amphorae Measurements SW47/1 BE:2017/10-2. ............................ 90 TABLE 14. Recovered Amphorae Measurements SW53/1 BE:2017/10-45. .......................... 92 TABLE 15. LRA1 Fabric Groups. Each group refers to a production area and not a specific workshop. ................................................................................................................................. 98 TABLE 16. LRA2 Fabric Groups ............................................................................................ 99 TABLE 17. Example of data preparation for creating nodes ................................................. 102 TABLE 18. Example of data preparation for creating edges. ................................................ 103 TABLE 19. LRA1 Production Centers – The last two rows are not illustrated on the map in FIGURE 6. ............................................................................................................................. 142 TABLE 20. LRA2 Production Centers – the last four rows are not illustrated on the map in FIGURE 6 as no kiln sites are yet linked with the production of this type. .......................... 144 TABLE 21. Shipwrecks with LRA 1 and LRA 2 amphorae found in the Eastern Mediterranean. ....................................................................................................................... 149 List of Figures FIGURE 1. The map of the Fournoi Archipelagos and its 58 shipwrecks that have been surveyed since 2015 to date. Yellow circles denote the four villages and their harbors (contemporary), (Campbell and Koutsouflakis 2021: fig. 2). .................................................... 6 FIGURE 2. The Fournoi complex depicted on the Tabula Peutingeriana (or Peutinger Table) of the 13th century- Courtesy Peter Cambell. .......................................................................... 11 FIGURE 3. Navigational routes in the Central Aegean, produced by the author, according to Campbell et. al (2021: fig. 5,7). ............................................................................................... 12 FIGURE 4. Places mentioned in this Chapter. Produced by the author, however, the three days sailing scales, follow Leidwanger’s time zones (2020: fig. 4.20). .................................. 14 FIGURE 5. Typology of the LRA1 type and its subtypes – according to Pieri (2012 fig. 2.3) and Reynolds (2011 fig. 31-q) classifications. ......................................................................... 22 FIGURE 6. LRA1 and LRA2 Production Centers and Potential Sites – produced by the author through Geographic Information System (GIS), sites, locations, and short descriptions can be seen in Appendix A: tables 19 and 20........................................................................... 26 FIGURE 7. LRA2 forms and related workshops, after Reynolds (2017: Fig. 24). ................. 30 FIGURE 8. Maritime trade scales, After Tartaron (2018: Fig. 4-2). ........................................ 50 FIGURE 9. Left Amphora 1/1 BE:2015/5-52, right Amphora’s profile view. ........................ 64 FIGURE 10. Left: Amphora SW1/2 BE:2015/5-53, right: side view. ..................................... 66 FIGURE 11. Left: Amphora 2/1 BE2015/5-2, middle: handle detail, Right: pitch on the interior. ..................................................................................................................................... 68 FIGURE 12. Left: Amphora 2/2 BE:2017/10-4, middle: profile view, Right: top view. ......... 70 FIGURE 13. Left: top view, right: Amphora 2/3 BE:2017/10-5. ............................................ 71 FIGURE 14. Left: top view, right: Amphora 2/4 BE:2017/10-21. .......................................... 73 FIGURE 15. Left: Amphora SW5/1 BE:2015/5-7, Right: the inscription “ΠΟP”. ................. 75 file:///D:/Final.docx%23_Toc163056132 file:///D:/Final.docx%23_Toc163056133 file:///D:/Final.docx%23_Toc163056133 file:///D:/Final.docx%23_Toc163056134 file:///D:/Final.docx%23_Toc163056137 FIGURE 16. Left: Amphora SW6-1 BE:2015/5-8, right: top view. ........................................ 76 FIGURE 17. Left: profile view, right: Amphora SW8/1 BE:2015/5-43. ................................. 78 FIGURE 18. Left: Amphora SW10/1 BE:2015/5-26, right: detail on the handles, neck, and preserved body. ........................................................................................................................ 79 FIGURE 19. Left: Amphora SW10/2 BE:2015/5-27, middle: profile view, right: top view. .. 81 FIGURE 20. Left: Amphora SW14/1 BE:2015/5-35, middle: profile view, right: top view ... 82 FIGURE 21. Left: Amphora SW19/1 BE:2015/5-55, top right: detail of the handle, bottom right: detail of the base. ............................................................................................................ 84 FIGURE 22. Left: Amphora SW24/1 BE:2016/4-10, right: detail of the mouth, neck, and handle. ...................................................................................................................................... 85 FIGURE 23. Left: Amphora SW25/1 BE:2016/4-13, Middle: top view, right: detail on the rim and handle. ............................................................................................................................... 86 FIGURE 24. Left: Amphora SW42/1 BE:2016/4-85, right: top view. .................................... 88 FIGURE 25. Left: Amphora SW44/1 BE:2016/4-106, Middle: profile view, right: top view. 90 FIGURE 26. Left: Amphora SW47/1 BE:2017/10-2, middle: profile view, right: top view. .. 90 FIGURE 27. Left: Amphora SW53/1 BE:2017/10-45, middle: profile view, right: top view. 93 FIGURE 28. Late Roman amphorae types and their quantities from Fournoi, produced by the author........................................................................................................................................ 95 FIGURE 29. Late Roman shipwrecks from Fournoi and their suggested dates. ..................... 97 FIGURE 30. Two-mode network of shipwrecks and amphorae forms. Produced by the author. ................................................................................................................................................ 105 FIGURE 31. Two-mode network with shipwrecks and their dates, shortened by 100 years. Produced by the author........................................................................................................... 106 FIGURE 32. Two-mode network with shipwrecks and their ceramic fabric groups. Produced by the author. .......................................................................................................................... 107 file:///D:/Final.docx%23_Toc163056142 FIGURE 33. Two-mode network with shipwreck’s cargo amphorae and their similarities found elsewhere, in both land and underwater sites. Produced by the author. ...................... 108 FIGURE 34. Multi-mode network consisting of shipwreck sites from Fournoi, dates, fabric groups, similarities with other land and underwater sites, and cargo types utilizing Gephi’s ForceAtlas layout. Produced by the author. ........................................................................... 109 FIGURE 35. Shipwreck Distribution of LRA1 and LRA2 Amphorae, produced by the author utilizing data from various sources (APPENDIX D). ............................................................ 112 FIGURE 36. Transport or sea zones division of the Mediterranean Sea (Parker 2008: 192) 113 FIGURE 37. A guide for ceramic paste inclusion’s analysis, according to Bonnet, et al. (2019). .................................................................................................................................... 146 FIGURE 38. Comparative dates CE of shipwrecks with LRA1 amphorae cargo (produced by K. Raptis using Microsoft Word). .......................................................................................... 152 FIGURE 39. Comparative dates CE of shipwrecks with mixed LRA1 and LRA2 amphorae cargo (produced by K. Rapti). ................................................................................................ 152 FIGURE 40. Comparative dates CE of shipwrecks with LRA2 amphorae cargo (produced by K. Raptis). .............................................................................................................................. 153 file:///D:/Final.docx%23_Toc163056159 file:///D:/Final.docx%23_Toc163056159 file:///D:/Final.docx%23_Toc163056160 file:///D:/Final.docx%23_Toc163056160 file:///D:/Final.docx%23_Toc163056162 file:///D:/Final.docx%23_Toc163056162 Chapter 1. Introduction Lying in the eastern Mediterranean and known as the “capital of shipwrecks,” the waters of the Fournoi Archipelago have vast amounts of sunken cargo that can reveal information about trade and exchange of the Late Roman period.1 The material culture can provide details such as fruitful insights for the amphorae themselves, regarding typological sequencing, but also dating of the wrecks that have yet to be examined more detailly. Through these matters, pottery analysis, along with spatial distribution and environmental data, can provide evidence of the maritime routes travelled. The material culture’s economic implication provides an opportunity for expanding on broader economic topics, namely trade scales of both long and short-distance, as well as various economic behaviors of state-led and free market trade. Overall, the Fournoi case should be viewed as a microcosm of the Late Roman economy of the Eastern Mediterranean due to quantitative data, such as the coexistence of numerous forms of amphorae from a long chronological range yet are found in the same location. On the other hand, the purpose of this study is not only to discuss amphorae samples but also to look closer into their economic ties to understand the infrastructure used to produce and distribute them but also wider economic systems, therefore various interdisciplinary methodological tools will be exercised. In conclusion, Foirnoi’s scattered information should also be viewed as a chance to establish a research way of “making sense” of the limited survey data available, often common in Mediterranean archaeology of the Classical times. In more detail, the study of Late Roman amphorae has been long and continuous. Late Roman amphorae have been examined for decades since 1899 (see LRA1 and LRA2 in chapter 2), but significant gaps and questions are still open regarding our understanding of their production, distribution, and consumption. Several scholars have been involved in the tedious processes of analysis, ranging from typological examination to scientific DNA 1 In historiography the Late Roman Empire spans from Diocletian in 284 until the death of Heraclius in 641. 2 extraction, often revealing vital information and adding to the “mosaic” of knowledge. All these actions have helped to facilitate the research of material culture towards a better understanding of economic ties. There is, however, more work ahead to comprehend commonly transported commodities as well as fully understand their economic implications in more depth, in order to create a complete picture of their significant yet fragmentary typological and petrographic sequence. In addition, after personal communication with Justin Leidwanger, what is currently missing from the research of Late Roman amphorae “is to create a comprehensive map of the various networks of state-led and free-market commerce, that distributed Aegean commodities, while also highlighting the connection nodes and key redistribution points of Late Antiquity.” Or to frame it differently as Wilson (2009:245) states “trade in what goods, between what regions, at what times?” Shipwrecks equipped with bulk cargoes expected of state supply, on the other hand, have remained rather elusive (McCormick 2012; Brennan et al. 2020). Thus, towards this direction, one cannot ignore the pivotal role that the sea had in Late Antiquity, as it formed the basis of trading networks of the Mediterranean. Within these networks and landscapes, it is highly possible for these bulk cargoes to be lying still underwater. Completing the picture of these trade commodities and as a result, reconstructing their economic implications requires looking closer at them by utilizing shipwreck archaeology, especially from a mariner’s perspective, as these prevalent types would be distributed through the sea. In the last two decades, the catalog of late Roman shipwrecks has been dramatically expanded, with the Fournoi Archipelago adding recently to the list several new examples of great potential for interpreting the economic ties of complex societies. The vessels that sunk in Fournoi are only some of the dozens of mixed cargo wrecks along the coast of the central and eastern Mediterranean. Mixed and unified cargoes seem to cover a wide timeline, containing various 3 typologies from several workshops within a corpus of both unified and mixed cargoes, and thus offer a chance to study maritime commerce through a new prism. A large number of shipwrecks became overwhelmingly predominated during the middle fourth century (Leidwanger 2017:605-06).2 This is explained by the shift of the gravity center leading to Constantinople in 324 CE (Brennan et al. 2020:328). During Late Antiquity, in the Eastern Mediterranean, according to Matthew Harpster (2005:356-360), the numbers dropped again.3 Although the purpose of this thesis is not to discuss the overall shipwreck presence and absence, it is important on the other hand to pay attention to the Fournoi’s chronological distribution within the wider norm and the importance of amphorae usage in shaping our understanding about trade.4 Thus, an updated list of the Fournoi shipwrecks representing 23 percent of shipwrecks found in Greek waters will contribute to our understanding of discussing patterns of maritime activities (According to the Ephorate of Underwater Antiquities (EUA). This thesis intends to cover trade scales of short and long distances, commercial or state-driven powers, the fundamentals of all production and distribution, and exchange patterns through the archaeological record. It is understood that the data provided is fragmentary and scattered. However, looking at a few pottery assemblages representing a fraction within a dozen wrecks, of various tonnages and hundreds of sherds still lying on the seafloor, should not be viewed as an incomplete effort. Instead, it aims to generate a useful methodological tool for interpreting scattered data from survey seasons. Consequently, a 2 Parker’s (1992) shipwreck catalog in the 1990s revealed a decline in maritime traffic roughly around 300- 450CE. However, McCormick’s (2012:84-6) updated list seems to minimize the drop, adding new evidence between 300-700 CE. 3 He states also that approximately 19 sites date between the middle to late 6th century, while only 12 dates to the middle or late 7th century. Additionally, between 750 and 850, there are only four known shipwrecks in the Mediterranean. However, all these catalogs do not come without caution, and we have to pay particular attention to newer studies but also the limitations of even existing ones. 4 For a detailed discussion about quantifying shipwreck archaeology, amphorae and their connection see Wilson (2009). 4 comparative maritime study of material culture, geography, harbors, cities, forts, landscapes, historical records, and networks offers a great interdisciplinary and multi-temporal opportunity forced by the data’s nature. Finally, very little underwater archaeology has been carried out or published for the area that today covers modern Greece, an area very vivid and promising, “a key sample area” bearing so much information for the ancient past” (Wilson 2009:22). Research Questions The island’s maritime assemblages offer several research questions regarding complex economic systems within the economic transformation of Late Roman times. In particular, the questions that triggered this research are described below: 1. How are the multi-temporal corpus of amphorae and hence, shipwrecks found best interpreted? 2. Can a more precise chronology and re-evaluation of the shipwreck’s dating be provided? 3. What are the conditions within which two commonly used jars from diverse geographical origins, from the Greek mainland on one hand and the other from the southeast Mediterranean, tie together to the same maritime activities? 4. How can network analysis be utilized to offer fruitful insights for the site-centric study of shipwrecks and their counterparts- cargoes? 5. Can a methodological tool for interpreting scattered survey data into a cohesive study be offered? 6. What can be said about the Fournoi’s role in navigation? 7. Finally, what are the broader socioeconomic aspects of the Late Roman period, and how can shipwreck archaeology fill the gap on questions regarding distribution? 5 Although scholars have studied material from the eastern Mediterranean, the location of Fournoi compared to other studies is far more interesting. All amphorae production centers appear to be south of Fournoi. Previous studies have focused on both regional and interregional scale trade, yet Fournoi’s assemblages’ apparatus remain rather unclear. It might have served as an interregional scale route, facilitating the distribution of Late Roman products within more hierarchical and organized networks. If this hypothesis is correct, then Fournoi might serve as an example to re-evaluate the nature of such prevalent types within a well-organized system operating under complex administration mechanisms. Before we start discussing all these topics it is necessary to begin by presenting the site of Fournoi. Thus, its terrestrial and maritime heritage as well as information about its significant role on navigation routes will next be discussed in depth. History and Archaeology of Fournoi To begin, the Fournoi Archipelago5 (FIGURE 1) or to use the names used before the 10th century CE, Korseai, or Corassiae (e.g. see Strabo 10.5.13), comprises a complex of twenty islands and islets, southward of its neighbors, the larger eastern Aegean islands of Samos and Ikaria. In more detail, Fournoi is located in the Ikarian Sea, a section of the Aegean that has been characterized as notorious since Antiquity due to its dangers. The Fournoi Archipelago complex consists of 45,25 km² and played a significant part in the Aegean’s navigational landscape due to its proximity to major city-states and its geographical importance, lying in between those two bigger islands. This landscape forced ancient sailors to pass by Fournoi while crossing the eastern Aegean or traveling west to the Greek mainland (Cambell and Koutsouflakis 2021:279-280). 5 The earliest-known source to use this name is the Stadiodromikon of the De Ceremoniis, likely dating around the 10th century CE (Campbell and Koutsouflakis 2021:281). Other names mentioned in the bibliography are - Κορσεαί/Korseai, Κορσοί/Corsoi, Κορασσίαι/Cορασσιαι, and Kουρσοί/Coursoi (Viglaki-Sofianou et all 2019:11, 12, 27). Fick (1905:53) argues the etymology of ‘Corassiai’ to be Carian. 6 FIGURE 1. The map of the Fournoi Archipelagos and its 58 shipwrecks that have been surveyed since 2015 to date. Yellow circles denote the four villages and their harbors (contemporary), (Campbell and Koutsouflakis 2021: fig. 2). The major work by Viglaki Sofianou et al., The Korseai Islands (2019), is significant for our knowledge about Fournoi and consists of the only complete study regarding the island. As they comment, “the poor historical record albeit with limited attention from the Greek Archaeological Authority for the island of Fournoi”, has resulted in an obscure historiographic narrative of its terrestrial heritage (Sofianou et al. 2019:39). However, even from these sporadic and scattered, so to speak, efforts, one can point several things regarding both its history and archaeology. 7 To begin with, the island was likely inhabited since the Neolithic Period with a population that did not peak until Hellenistic (c. 330 BCE) and Roman times. Most current settlements and villages remain in their ancient locations. For example, the main village of the island named Fournoi served as the main village of Korseai in antiquity and dates between the third century BCE and the second or the third century CE. It also includes a Hellenistic fortification on an acropolis. Another village on the island's northern part is called Chrysomilia, where the remains of a tower are still visible today, which likely served as a warehouse and watchtower. Its long history of occupation is shown by the fact that it has a Byzantine temple built atop its ruins. Additionally, a village on the eastern side of the archipelago called Kamari, dates to the Roman period with some remains still visible. These include the remains of an ancient wall next to the coastline (Viglaki-Sofianou et al. 2019:39- 141). In more detail regarding the ancient world, both the acropolis and the watchtowers likely served as part of a wider organized network for supervising shipping routes and channels. The communication system of Phryskotiai through watchtowers was well-known in ancient times6. Evidence of the acropolis, accompanied by an inscription, testifies to the existence of the ancient city of Korseai in the modern settlement of Fournoi. Furthermore, the Acropolis, during the Hellenistic kingdoms and the period known as the Ptolemaic thalassocracy (second century BCE), was seized by Samos, a major naval station, which likely established a garrison on it to better control the area. The evidence is mainly epigraphical7, and the situation was probably similar to that of the Samians occupation in Ikaria and Amorgos (Viglaki-Sofianou et al. 2019:39-141). 6 For more details on the communication system of Phryskotiai see Viglaki-Sofianou et al. (2019:51). 7 IG XIII, 6, II, 1213 8 Fournoi was probably a Milesian colony (in the early Archaic period, c. 650-550 BCE) before the Samian occupation; for supervising and controlling the Fournoi passage, due to the island’s important location. In the year 129 BCE, Fournoi became part of the Roman province of Asia and was thus ruled by Rome.8 According to the Notitia Dignitatum, an administrative document dating to c. 395-413 CE, Fournoi continued to be part of the Diocese Asiana, along with Lycia and Caria, the so-called Asiana VII region (Campbell and Koutsouflakis 2017, 2021:281). Scattered traces of a Roman necropolis together with a later Roman settlement at Kamari and many other remains of early Christian and Byzantine temples highlight the archaeological milieu of a “mare incognitum,” as has been addressed by Voutyrea (2021:13) since no systematic scientific research has been conducted on the island. Despite its prominent geographical position, the archipelago was not rich in resources. The only export commodity according to Kienast (2007:165-176), which was mostly valued in the Asian Minor and the neighboring islands, was a type of greyish-white marble, as the largest quarry at Petrokopio was located on the west coast of the main island of Fournoi, south of the Korseai village. Thus, the island’s poor production power could not sustain its prosperity. On the other hand, Fournoi, as it appears from the historical and archaeological record, was a key waypoint, a stopover for routes transiting the eastern Aegean. According to observations on the modern harbor at Kamari, remains of ancient structures have been found. These remains include an anchor graveyard, with more than 30 anchors of different periods (e.g., from Bronze Age to the Age of Sail), lying next to each other. Although modern maritime facilities consist of a single small dock to serve local fishermen’s needs, a few corroded walls stand along the coastline, testifying to a maritime landscape likely used for more extensive shipping in the past (Viglaki-Sofianou et al. 2019). 8 A Roman statue of Augustus was found at the Acropolis at Fournoi, indicating the inclusion of the island to Rome, alongside inscriptions testifying such actions (IG XII, 6, II, 1205). 9 The Maritime Cultural Heritage of Fournoi: Navigation and the Role of the Island’s Geography While Fournoi’s terrestrial archaeological sites consist of small-scale settlements of limited duration, maritime archaeology reveals extensive connectivity in nearly every period. Since 2015, a collaborative survey team combined from the Greek Ephorate of Underwater Antiquities (EUA) and RPM Nautical Foundation, located 58 shipwrecks distributed throughout the islands from the Archaic period (650 BCE) to the 20th century (Campbell and Koutsouflakis 2017).9 Beginning with sites reported by sponge divers, the team began a systematic diver survey and a multibeam geophysical survey, utilizing a remotely operated vehicle (ROV). These methods were continued in 2021, followed by a systematic excavation of one of the Late Roman shipwrecks.10 The team to date has been focused on systematic excavation efforts. Furthermore, according to Campbell and Koutsouflakis (2021:281-4), the largest concentration of shipwrecks is located on the east side of the Agios Menas Channel between the large island of Fournoi and the small island of Agios Menas, where more than 16 of the known shipwrecks lie almost next to each other. Considering the status quo of Fournoi as was previously explained and being one of the smallest islands of the Aegean it is surprising that Fournoi is represented in most portolans and nautical charts from the 13th century onwards. Viglaki-Sofianou et al. (2019:146-150) offer a concise review of portolans and ancient navigation maps, reconstructing hence the importance of the island as a vital passage. Thus, they note that Fournoi is depicted, despite its size, in the work of the early portolans in a “definite way.” For example, on the pan-Mediterranean map of R.W Seale (1747), the island has a “proper size and location,” despite many other big islands like Kalymnos that are absent. Accordingly, the famous Tabula Peutingeriana or the Peutinger Table of the 13th century, the most complete 9 55,1% are dating in Roman times. 10 For more information on the late Roman shipwreck number 15 see Voutyrea (2021). 10 extant Roman map that consists of a copy of an earlier map from the fourth century CE, which itself is thought to have been built based on a first-century CE original. This map (FIGURE 2) includes Fournoi, appearing as Corsa or Cross-a phonetic corruption, among the 22 depicted Aegean islands. Those portolans and many others include Fournoi while surprisingly more prevalent and bigger islands are not depicted. Consequently, the significance of the “Fournoi Pass” or “Stenon Fournon” as a maritime “chokepoint,” a constriction in the age of wind-powered vessels, can be better understood through the island's pivotal references in portolans and periploi, combined with its substantial number of wrecks. Such significance can be observed through Joseph Tournefort’s account during the 18th century, highlighting the importance of Fournoi as a passage: All the Ships coming down from Constantinople into Syria and Egypt, after resting at Scio [Chios], are obliged to pass through one of these Straits. The same must they do, that go up from Egypt to Constantinople (Campbell and Koutsouflakis 2021:285 after Tournefont, 2014:306). Thus, as Tournefort comments, followed by a strong agreement regarding several portolans, historical sources, and environmental data, ancient sailors would be obliged to pass by the ‘’Big Strait’’ of the Fournoi archipelago, a constriction within the Aegean landscape, as it is formed by Ikaria and Samos dividing the eastern Aegean in two. Additionally, the archipelago, unlike its neighboring islands, offers safe anchorage. Strabo (14.1.19), Tournefort (1718:333), and the local bishop Georgirenes (1678:54-55) state that both the island of Ikaria alongside the western side of Samos are harborless. As a result, during severe weather, especially the northwestern Etesian wind, the so-called ‘Meltemi,’ a strong and consistent primary summer wind, the Fournoi Pass seems the safest place in the region for vessels to anchor. Three navigational routes (FIGURE 3) drawing from Campbell and Koutsouflakis 2021:285-290) and Avramea (2002:84) seem to provide 11 options for sailors traveling north to south in the eastern Aegean and vice versa. Simultaneously the island’s location was also prominent in the east-west and west-east axis which has also been discussed by Arnaud (2005:57). FIGURE 2. The Fournoi complex depicted on the Tabula Peutingeriana (or Peutinger Table) of the 13th century- Courtesy Peter Cambell. The first route following the north-south axis runs past the western cape of Ikaria, with the second going through the narrow Samos Strait (its narrowest point is only 1.6 km wide), separating the mainland from the island. Both options are problematic since the first one is dangerous, being affected by the strong Etesian winds, and the second one is time consuming, with a landscape perfect for piratical actions. The third and last route runs through the Fournoi Pass, and it is the most direct and safest. On the other hand, one cannot ignore, that even this passage has its dangers with fierce winds that gain speed after crossing Samos and Ikaria, creating large waves with strong local currents, as the shipwreck assemblage reveals. However, even when these dangerous conditions prevail, 12 Fournoi offers a place to anchor and wait them out. This north-south arterial network especially after transferring the new capital from Rome to Constantinople, attracted many pirates that traveled annually from as far away as France, Italy, Malta, and Sardinia to hunt ships (Ormerod 1924:19). FIGURE 3. Navigational routes in the Central Aegean, produced by the author, according to Campbell et. al (2021: fig. 5,7). As can be seen, shipwrecks that sunk in the Fournoi Archipelagos were probably not a function of trade within the villages and communities on the island, as the lack of port facilities, and settlement status of the Island of Fournoi reveals. Instead, they indicate a high volume of trade that passed through the Fournoi Strait. The discovery of shipwrecks from different periods along the same routes, often in the same positions as in Aspros Cavos (white cape), on the western part of the island, is indicative of the degree to which maritime people stuck to predetermined routes (Avramea 2002:80). The extensive number of wrecks 13 undoubtedly pinpoint the importance of the Fournoi Pass together with the settlements (towers/acropolis) that aimed to maintain and control the “Big Strait” that creates the Ikarian Sea. The Fournoi Archipelago served a dual role in the landscape of the central-eastern Mediterranean. Its proximity to several significant centers and various other islands within a regional scale is undoubtedly observable based on the Island’s geography. Thus, several cities lie from a day to a few days sail from the island, as Leidwanger (2020:145-48) noted, utilizing sailing, alongside eastern Mediterranean wind data, to calculate time. Consequently, most of the 12 city-states that formed the Ionian League in Asia Minor and continued playing a significant role during Late Roman times, are within a day away, including ones such as Ephesus, Miletus, Priene, Samos, and Teos but also Claros, Didyma, Heracleia, Lebedos and Notium. Additionally, several other major centers lay within a few days of sailing such as Chios, Smyrna, the Datca Peninsula, Kos, Rhodes, and others (FIGURE 4). Through the documented data, it is possible to state that Fournoi played a vital role on both regional and interregional scales, which will be discussed in Chapter 2. Since the foundation of Constantinople, according to Avramea (2002:83-83), the eastern part of the Mediterranean became busier, and all the sea lanes led sailors to pass by Fournoi. Consequently, the various harbors along the Propontis and the North-South axis of the eastern Aegean linked the capital with nearby provinces, especially in the eastern Mediterranean, Egypt, and the coast of Africa, ultimately creating a traffic route past the islands of Myteline, Chios, Kos, Samos, and Fournoi. A journey from Constantinople with a sail-equipped merchant ship to Rhodes would require approximately seven days, while sailing to Cyprus was around 18 days (stopping almost every two days on islands along the way). Sailing to Egypt would require approximately 12 days as well. These time and range examples, whether 14 particularly accurate or not, highlight the fact that the Aegean could be crossed in 2-6 days, depending on the route followed. FIGURE 4. Places mentioned in this Chapter. Produced by the author, however, the three days sailing scales, follow Leidwanger’s time zones (2020: fig. 4.20). Commercial activities of various scales would utilize the Fournoi passage to distribute their goods, destined for neighboring cities or further destinations and marginal places such as the Danubain Provinces. As Christy Constantakopoulou (2007:15-28, 61-89, 176-227) has noted, insularity in the Aegean should relate to an exploration of the theme of connectivity, which creates the dense cluster of numerous islands, serving as “bridges,” highly facilitating navigation. Such clusters would shape the Aegean into a world of increased mobility, including island hopping, cabotage, and longer-distance trade. Therefore, ancient sailors were not restricted to long-term journeys, with essential but also frequent stops along the way, especially when a storm or a strong wind prevailed. Of course, the maritime network of 15 communication existed for centuries in the Aegean, and it was not a “new” activity within the Late Roman Empire. The lack of self-sufficiency or on the contrary island wealth, is another key for one to understand the need for constant trade and exchange, as islands would either trade to survive or prosper by selling their surplus. Additionally, another pattern that is usually observed in the Aegean milieu, is that of a larger island controlling their smaller neighbors, a “micro-imperialism,” due to access to agricultural resources. For example, Constantakopoulou (2007:15-28, 61-89, 176-227) observes that the island of Samos exercised some control over Ikaria and Fournoi during the third and the second centuries BCE, as was earlier explained. Such control must have continued in later years, as Samos continued to expand and gain power. Regional and interregional economic activities are more blurred throughout this dense matrix of islands and will be properly discussed in later chapters. Thesis Outline This thesis is divided into 4 chapters together with an introduction and conclusion. The first section contains the introduction, discussing the overall history and archaeology of the Fournoi Archipelagos, highlighting the island’s significant geography as an important mode of passage and navigation, as it likely would have been viewed as an unintentional stop of the various sunken vessels. Chapter 2 focuses on the two late Roman types of LRA 1 and LRA 2 presented in a long and in-depth discussion regarding their production sites, providing information in typology, distribution as well as epigraphic and content insights. Chapter 3 reviews the economic complexity of the Late Roman mosaic: distribution, maritime trade, economic behaviors, and connectivity through the implications of the amphorae trade. This chapter contains the necessary theoretical and historical formulas that will be referred to in the following chapters. Chapter 4 presents the analysis of all the amphorae samples collected from Fournoi. Stylistic, fabric, and macroscopic analysis, followed by measurements, and photographs of every individual sample all together aim to better extract essential information 16 from these traded commodities. The last chapter presents and discusses the results of the Chapter 4 analysis, while also utilizing the outcome as raw data for a network analysis experiment. Furthermore, this chapter covers topics regarding other distribution sites (e.g., a long catalog of shipwrecks with similar cargoes found elsewhere), but also consumption sites in Asia Minor and the Danubian provinces, essential pieces of spatial and geographic information for comparative patterns tracking. Chapter 2. The Two Late Roman Amphorae Types: LRA1 and LRA2 The Late Roman Amphorae One of the earliest amphorae classifications derived from efforts to illustrate a corpus of amphorae inscriptions from Rome, which started with Heinrich Dressel in 1899. A particularly important conclusion in the wide field of Late Roman amphorae taxonomy after years of research, was the identification of a group of six types11 which Riley (1979:212-32) very accurately described as a “standard package with diverse origins,” and Karagiorgou (2009:42-3) as “branded” or internationaux, reflecting their popularity. Amphorae were produced before but these branded containers are the result of a massive production in more than one center. These types appeared in large numbers, counter to localized amphorae production that had been common, in numerous places all around the Mediterranean and far beyond during the Late Antiquity (fourth to eighth centuries CE). They can overall be summarized, according to Diamanti (2008), as closed-shaped jars with cylindrical, spherical, and sometimes long, narrow necks with two handles to facilitate transport. However, an issue often following the classification process based on types and several subtypes is the fact that a single amphora is known by numerous different names; often after the various scholars who studied them. Thus, to find one’s way around the bewildering array of amphora typologies demands a common classification system. Being made in small, locally produced batches, they did, however, differ slightly from one province to another. Arthur (2007:174) highlighted a gradual fading of these prevalent ceramic types’ use starting in the seventh and eighth centuries. Across the Mediterranean, ceramic production became less sophisticated and declined in numbers than it had been in former centuries. The 11 LRA1 and LRA2: see later in this chapter, LRA3: Western Asia Minor, LRA4: Gaza type, LRA5/6: Palestinian? LRA7: Egyptian (Karagiorgou 2009:41). 18 absence of any large industrial production centers, albeit with the introduction of different forms and much later types such as wooden barrels, resulted in an overall decline of these branded amphorae. Smaller containers began also to appear with a shape, notably the so- called globular types, which often imitated old forms. The overall picture that emerges in later times is that of fragmentation and simplification of the cycle of production, distribution, and demand (Cosentino 2021:24-28). The case of the Late Roman Amphora 1 (LRA1): Scholarly Community: First, we shall begin by addressing and presenting one of the most prevalent types, a type within a corpus of the most exported amphorae of the Byzantine East, classified as Late Roman 1. The Late Roman 1 (LRA1) name was coined after Riley’s excavation efforts at Sidi Khrebish, Benghazi (Berenice), and Carthage in the 1970s (Riley 1979:212-32). Besides LRA1, numerous designations have been given to this class reflecting its extensive scholarly attention over the years. Thus, one can divide these studies into three dissimilar categories following Demesticha’s taxonomy (2013:71-72). The following taxonomy is the result of various scholars designating numerous names for the same typological group for different research purposes, for example by looking at them as key for maritime networks or classifying them according to a specific site or geographic area. It is important to understand but also break down, the historiography that follows such types to comprehend better how they appear as well as the wider theoretical approach that accompanied them. It is also intended to do so for one not to get lost in the modern “tower of Babel” of classification names. To begin with, the first category includes earlier classifications created during the second half of the 20th century; however, these studies deal with the LRA1 amphorae as one 19 type, without distinguishing variants. They were, however, studied as proxies for maritime connectivity and exchange networks. Consequently, LRA1 appears in the bibliography as British Bii after Thomas (1959) who studied the amphorae of the eastern Mediterranean imported into Britain; Kuzmanov 13a and 13b (1973), who classified the Roman amphorae from the Black Sea; and Peacock and Williams 44 (1986), after a division on 66 general “Classes” based on Roman amphora types from all over the Mediterranean. The second category includes typologies, in which variants within the types are distinguished based on similarities in material deriving either from a specific site or a specific geographical area. Egloff 164 and 169 (1977:112-113), studied amphorae found in Kellia, Egypt, similarly Riley LR1 and LR1 A (1979:212-216), based on findings at Benghazi in Libya and Carthage (1981:85-124) and Keay 3 (1984:268-278) for amphorae from Catalonia Spain. Less common names include Agora M 333 after Robinson (1959), Hayes 5 (1972), and Ballana 6 after Farid (1963) can also be included in this category. The third category includes recent publications. Although most scholars today tend to use the established term “LRA 1”, they still group them in several diverse ways. Andrei Opaiţ LR1 A-E (2004b:8-10), divided the amphorae from Roman Scythia into five variants. Likewise, Pieri LR1 A and 1B and LR1 A transition (2005:70-76), classified the LR1 amphorae found in Gaul into two “generations”. Finally, Peter van Alfen (1996:203) subdivided the LR1s from the seventh-century Yassi Ada shipwreck into eleven groups (I- XI). Essential finally is the contribution of Reynold’s in understanding linear typologies (2005). Typology: In 2005 Pieri, subdivided the LRA1 type into LRA 1A-C, aiming to establish a general typological sequence. However, the issue with such taxonomy is the fact that these 20 variants he studied from the Western empire, are typically Cilician forms, and he thus not equal material from the East is representable. Thus, he did not include, as they were excavated later, other major production centers like those in the Aegean Islands. Later scholars added more variations, identifying more variants as will be discussed later. Even though the type is well established from the fourth to the seventh centuries CE, its forerunners date back to the mid-third century CE. The origins of these prototypes12 (transitional variants) were suggested to have been in eastern Cilicia, as first addressed by Reynolds. Reynolds (2005, 2011, and 2017) defined the Pompeii 5 type, as a precursor of the recent LRA 1, whose development can be traced from the mid-first to early third centuries CE. The creation of a universal and linear classification is sometimes problematic as various places of production have unique styles that do not follow the norm of their times. For example, while the LRA 1B type in Cilicia has a wider and higher neck, the same type in Cyprus has a narrower and shortened neck as well as an ovoid, almost body shape (see fig. 2 LRA IB from Kourion, Demesticha 2013:173). Having said that, it is considered preferable to use the specific classification of their places of origin. However, as has been highlighted before, a common classification system is easier for one’s research, thus this thesis will follow the general classification of Riley for naming and Pieri for variants A and B. The first examples of the LRA1 type appear as exports to the Danube and Black Sea forts in the late fourth century. Although this type has several variations dating to different periods, in almost every case the handle has a single or double groove. A groove is also observable under the rim, and the handles have a characteristic “Γ” shape. Their body is usually pear-shaped or cylindrical with heights ranging between 50-55 cm, and a capacity often of 10-15 Liters (L), while sometimes reaching 19L (Karagiorgiou 2010:49, 51). Following Pieri’s classification, 12 For more information about the forerunners of LRA 1, see Opaiţ (2010:1015-1022) and Arthur and Oren (1998). 21 but adding some more into it based on other areas, this thesis proposes a more detailed categorization of his two main variants of LRA, 1A and 1B, by utilizing various scholars’ contributions; drawing especially from Demesticha’s classification system (2013 and 2014) and expanding into typological characteristics of the seventh century (FIGURE 5). The First LRA 1A variant (middle of the fourth to fifth century CE) Regarding this early form which includes the amphorae identified as Egloff 169 (1977), Keay LIII A (1984), Williams Proto – LR1 (1995), Opaiţ LRA 1A2 (2004a) and Pieri LR1 A (2005), the type has a round or quadrangular, thickened or folded rim, and is almost square in cross-section (the external diameter of the rim is usually 5-8 cm). It has a narrow (cylindrical or conical) short neck and two angled handles that run from under the neck to the end of the shoulders (grooved and folded). The body is ovoid and covered with grooves which become more frequent on the shoulder and towards the bottom, and a base that ends in a small knob. Towards the second half of the fifth century, changes were observed in the morphology which will be explored in a subsequent section on LRA 1B. During the fifth century, LRA amphorae began transitioning from the 1A to the 1B types. The volume increased during the first half of the century, and in the second half the rim at the end bore a concave face and its diameter gradually increased, reaching almost 9 cm in diameter. This slow and late evolution sometimes makes it difficult to distinguish LRA 1A and LRA 1B forms. Additionally, the neck becomes wider and remains either cylindrical or conical. As before, the handles are decorated from a double-stepped groove: the base is narrow and bears a small nipple and lastly, the body is indented being pear-shaped (Bezeczky 2013:158-59; Demesticha 2013:72, 2014:601-603; University of Southampton 2014; Pieri 2005:71-72). Finally, Dominique Pieri (2005:76-77) addressed a particular generation which he calls the LRA 1 sous-module, which combines characteristics of both the LRA 1A and 1B variants. In short, this variant includes a 22 long neck, slender handles, and bodies of small capacity (3.5-5 liters). Additionally, several sub-variations are smaller and slimmer, which are commonly produced in south Cypriot workshops (Demesticha and Michaelides 2001). This variant can be considered mainly a Cilician variant, with possible minor Cypriot production sites (Pieri 2005 and Reynolds 2005). FIGURE 5. Typology of the LRA1 type and its subtypes – according to Pieri (2012 fig. 2.3) and Reynolds (2011 fig. 31-q) classifications. The Second LRA 1B variant (end of fifth to eighth centuries CE) As Demesticha (2013:72) states, most of the changes of this sub-type are not universal and are “typical of a new historical and economic context” (Demesticha 2014:601-603). Due to the lack of studies and excavations of kilns, addressing the beginning of such changes in 23 either Cyprus or Cilicia is impossible at this stage. On the other hand, several different forms might have coexisted during the sixth century CE. Thus, we can distinguish three different subtypes based on different workshops rather than sorting them by chronology, following Demesticha’s points (Demesticha 2013:72, 2014:601-603; University of Southampton 2014). I. The first form includes typologies from Keay LIIB (1984), Opaiţ LRA 1A3 (2004b), and Pieri LR1B and C (2005). The mouth is wider, between 10-12.5 cm, with a thickened or inverted lip followed by a ridge at mid-collar. The neck is broader and shorter than the previous ones, often deformed and encircled by a ridge. The handles are usually angled, becoming rounder towards the end of the sixth century, with various patterns on their outer face, and the body is more cylindrical, covered with grooves, and without bearing the characteristic nipple of the earlier types. II. The second form includes Keay LIII B and C (1984) and Opaiţ LRA 1B1 (2004a). The rim is simpler, slightly inverted, or thickened. The neck is rather short and cylindrical or slightly concave. The handles and body are like the earlier form. III. The third form includes Opaiţ LRA 1A4 (2004b) and Demesticha LR1.T1 (2003). The rim is folded with a flange shape ending, a short cylindrical, often deformed, neck whose diameter is equal to the height. The handles are angled from the lower rim to the end of the shoulders, with a double off-center grooving and an often ‘knee’ shaped junction at their attachment to the shoulder. The body is smaller, almost barrel-shaped, and covered with irregular stepped ridging. This fifth to sixth-century triple form-variant also stands as the peak of LRA1 production, which gradually became weaker and lost its dynamic presence in the archaeological record after the seventh century, until it completely faded during the eighth. Amphorae of the LRA 1B type were also produced in Sinope. Having their distinctive characteristics, they usually 24 appear in the historiography as Snp I and II and D Snp I similis. According to Tezgör (2020:41-43) “Snp” is used simply to distinguish them from other production sites. Distribution: Both generations of LRA A and B, from the end of the fourth to the sixth centuries CE, had a wide distribution all over the mare nostrum. Appearing in almost all corners of the Roman Empire, and of course the Fournoi Archipelago, LRA1 were the most prevalent eastern amphora export of the fifth and sixth centuries. Those exports were on a massive scale, and not just in Constantinople13 or ports cities throughout the Roman world - namely, Naples and Carthage. But they reached rural sites too. Specifically, this type is found in western Britain, in the south of France, Italy, the Adriatic region, Moesia and the eastern Alps, the Greek mainland, Egypt, Tunisia, Cyrenaica, the Levant, Cyprus, the Aegean, the Black Sea, and even as far as India. As the archaeological record from the West reveals, the first generation was probably more frequently traded there than the second, (Reynolds 2017:353-396; Bezeczky 2013:159). Workshops: So far, 18 kilns related to the production of this type have been located, with three potential areas still to be confirmed. Empereur and Picon (1989:236-243) were the first to catalog all but one of the sites utilized in this thesis. As Karagiorgiou (2009:43-5) points out, fabric analysis of such types is still in its “infancy” as very few examples of Late Roman kilns have been located and properly studied, with the known sites found in Cyprus, Kos, Paros, and few in the Cilicia region of modern Turkey. Nevertheless, the production of the LRA 1 type can be traced in a wide geographical area of the eastern Mediterranean, including 13 at Constantinople, LRA 1 was the most numerous amphorae present in seventh century deposits, 15% of the total amphorae, 17.1% in Berenike in the mid-sixth century, 42.5% in Carthage in the fifth and sixth (Decker 2001:76-7) 25 Cilicia, Caria, northern Syria, Cyprus, Sinope, the Aegean islands, and Asia Minor. Pieri (2007:301) has argued Cilicia was the birthplace of this type and that it was soon adopted by Cyprus. Regardless of the truthfulness of such a statement, all centers had significant distribution power, with Cilicia, Cyprus, and the Aegean containing the cluster of the most prevalent ones. What is also important here is that after a series of dramatic events, namely war campaigns led by the Persians Sassanids that led to the sack of Antioch, one of the most important cities of the Empire in its role as an essential administrative and economic center, in 540 CE, we observe not only a relocation of production areas from Cilicia to Cyprus and the Aegean, but also redistribution from the west to more stable markets such as Palestine and Egypt in the south and east (Pieri 2007:301-303). The decline of Cilicia, together with a new administrative unit of the Quaesture Exercitus by Justinian in 536, linked Cyprus with the Annona system, created a new market for Cypriot products destined for the Empire’s frontiers in a time of war. Both these systems will be thoroughly discussed in Chapter 3. The island’s strategic position, as a key factor in commercial activities, followed the administrative and trade links which naturally made Cypriot potters manufacture the common amphora shapes and ship their products. According to Demesticha (2013:176-177), the difference in their distribution and typological change in both Cyprus and Cilicia areas is indicative of their parallel function and rise in demand and not a matter of diffusion. The sixth-century LRA1s were associated with these new administrative formations for Cyprus, while Cilicia continued to supply older trade networks. During the seventh century, an overall decline of the type was witnessed, with new production centers such as the kiln at Kos. The South-East Mediterranean Cluster 26 LRA1s were produced across a large region (FIGURE 6) ranging from rural establishments to larger complexes, but there are several dense clusters of production sites, all located around the Eastern Mediterranean, with shared economic interests that extended far beyond a regional and provincial scope. Although the purpose of this thesis is not to analyze each of the production centers, a short presentation is considered valuable for mapping the geographic production of this type followed by a detailed table (APPENDIX A). FIGURE 6. LRA1 and LRA2 Production Centers and Potential Sites – produced by the author through Geographic Information System (GIS), sites, locations, and short descriptions can be seen in APPENDIX A: tables 19 and 20. The first cluster includes eight production sites in the Cilicia I and II, Syria I, and Isayria region in the southern part of modern Turkey. Following a west-to-east direction: the workshops were located at Antiochia and Cragum, Elaiussa Sebaste, Soli-Pompeipolis, Tarsus, Karatas, and Yumartalik. There were also kilns located at Selecia Pieria, one between 27 the two villages of Arsuz and Iskenderun, and a final significant cluster of LRA1 workshops is on the southern coast of Cyprus. All three known sites in Cyprus, lie along a 100 km stretch of coastline from Paphos to Zygi-Petrini. In more detail, petrographic analysis from Kourion indicates that even though a kiln has not yet been found there, the excavation of water wells in Amanthous (Empereur and Verlinden 1987; Touma 1989) implies the existence of a kiln at the city’s ancient harbor. Demesticha and Michaelides (2001) published the discovery of an LRA1 kiln site on the island of Paphos and another one was surveyed at Zygi-Petrini (Manning et al. 2000). All three kilns on the island date to the sixth and seventh centuries CE. Demesticha (2013:70) suggests that two more fabric groups named “X” and “ZA” exist on the island, fired at as yet unidentified kilns. The Aegean and Black Sea Cluster Four Aegean workshops appear at the islands of Kos, Paros, Rhodes, and Lipsi. As Demesticha (2013:70) comments, regarding Appolakia Rhode’s workshop, “no further archaeological work has been published for that workshop since Empereur and Picon in 1989”. One production site has been found in Halasarna, Kos (modern Kardamaina) that was used for the production of LRA1 and LRA2/LRA13 amphorae from the end of the sixth until the first half of the seventh century (Diamanti 2010a: 146-7, 2010b: 80-114, 2014:183). In Paros, a sixth and seventh centuries workshop has been found that was responsible for producing LRA1s and LRA13s at Lageri in Naousa-Z. Pighi (Diamanti 2016:691-97). At Lipsi, a workshop has been found and the research published by Papavassiliou et. al. (2014:159-168). Another LRA1 workshop exists at Sinope, on the Black Sea, and it is related to a different morphological variant, that of Snp I and II and D Snp I similis (Tezgör 2020:41- 43). Chios seems also to have produced an early LRA 1 variant (Opaiţ and Tsaravopoulos 2011:292). The Kiliseyanı/Reşadiye workshop in the Datça Peninsula and on the southwest corner of Anatolia, at İçmeler, seem also to be responsible for producing LRA 1 from the 28 fourth to the seventh centuries CE (Sakarya et al. 2019:323-324). However, no kiln has yet been identified. Finally, three more sites have been suggested to produce this type. The first one is in Turkey, the Delice kiln, but there is not enough evidence to catalog it as a production center (Rauh 2023). The second one is on the Carmel Coast, on the Levant. Although direct evidence in the form of kilns is lacking Gillet (2023) through petrographic similarities identified with this region, believes that a production center must have existed in this area. And the third, Ghaly (1992), proposed the production of the type in the monastery of Saint Jeremiah, in Egypt. That is a structure in use ca. 5th-10th centuries CE outside Cairo. The Case of the Late Roman Amphora 2 type (LRA2): Scholarly Community: The LRA2 is another important amphora, falling into the taxonomy of a standard package, with diverse origins and branded types within the corpus of the Late Roman Amphorae generation (Riley 1979:212-32). The LRA2 type, besides being the most prevalent form in the Northern Black Sea, is typically found at sites comparable with that of the LRA1 type (Badescu 2012:321). The name comes from Riley (1979) and, as with the LRA1 type, many other designations have been given this class: British B1 (1959); Benghazi and Carthage Late Roman Amphora 2 (1979); Keay 65 (1984); Peacock – Williams 43 (1986); Kuzmanov 19 (1973); Scorpan 7A (1975), and Hayes 9 (1992). Additionally, essential is the classification of types by Van Doornick adding four major working types after Yassi Ada A (1989), Kuzmanov 28 (1992), and Opaiţ (1984), after studying the type within the context of the Danube frontier and the Black Sea. Typology: Regardless of the variations, according to Karagiorgou (2001:130), the type is typically a wheel-made and broad-bellied container with wide shoulders, a tapered conic 29 neck, and a wide conical or cup-shaped mouth. Maximum diameter occurs at the shoulder, while the handles are crudely crafted, flattened, and oval in cross-section. The upper part of the body is decorated with “stylus grooving,” which may be straight or undulating, and the lower part has a small basal knob, at least in the case of early variants (Bezeczky 2013:160- 62).14 This type generally occurred from the early fourth century CE, often in Scythian contexts (Opaiţ 1996, 2004a, and 2004b), until it disappeared sometime in the early seventh century CE (Riley 1981; Peacock 1984). However, the production tradition seemed to carry on as amphorae similar to this type recently appeared in contexts of the eighth century. In conclusion, the capacity of this type was large in the fourth century CE, while in the next centuries, its volume was reduced in half, ranging between 30-64 liters according to Badescu (2012:318). The antecedence of this type is believed by Opaiţ Andrei and Dyczek Pochmarski (2008), to go back to the first and third centuries CE. The Peloponnesian or Boeotian precursors seem to have been the Dressel 24 type (or mentioned with other names such as Zeest 90, Knossos 15, Peacock, and Williams 57), similar to the LRA1 type, which appears in contexts of the lower Danube and the Black Sea (Opaiţ 2007a and 2010). According to various studies and scholars on wide area coverage, it is suggested there are two main generations of this type (FIGURE 7). The first variation of the LRA 2A (fourth and fifth centuries CE). Regarding this first variant, the mouth is broad and funnel-shaped. Both the shoulder and the lower part of the body are decorated with a deep horizontal grooving, which becomes broader on the central girth and appears only on the shoulder towards the end of the fourth century. Additionally, this earlier type bears a small basal knob/nipple. During the fifth century, the form is generally shorter, the grooving occurs only on the shoulder and the body is squatter (Reynolds 2017:327-35; Swan 2004:371-82; Opaiţ 2004:10-12). 14 According to van Alfen (2015:18), the existence of a basal knob dates them before 550 CE. 30 FIGURE 7. LRA2 forms and related workshops, after Reynolds (2017: Fig. 24). The second variation of the LRA 2B (sixth to eighth centuries CE) 31 Even though LRA2 is often labeled an “Aegean” amphora, it is not visible in Aegean production centers until the sixth century (Reynolds 2017:327-35). By the middle of the sixth century, the shape became more oval and the neck sharply conical and narrow. Towards the end of the same century, the decoration became undulated, and the basal knob disappears. However, as Sawn (2004:371-82) highlights, these neck and rim variations might not reflect chronological but rather regional differences. Within LRA2, some sub-variants appear after the seventh century which are sometimes ambiguous in the bibliography: the LRA 13 or LRA 2/13 variant or LRA 2C (after Pieri 2005:85-89). According to Bonifay and Pieri (1995:109-110), this variant has a rather short and more vertical mouth, with an oval body and base together with ridges along it. This form appeared during the mid-sixth century and goes all the way through to the eighth century, and it seems to have replaced the traditional LRA 2 and became an entirely different type rather than just a regular variant. It is present also in the workshops of Kos and Cyprus, and while distribution sites in the east and south Aegean have been found, other workshops have not yet been located (Karagiorgou 2001:77-80). Distribution The distribution of LRA 2, much like the LRA1 discussed previously, is overly broad and wide. Examples have been found in western Britain, the south of France, Italy, the Adriatic region, Moesia and the eastern Alps, the Greek mainland, Egypt, Tunisia, Cyrenaica, the Levant, Cyprus, the Aegean, Black Sea regions, and India (Bezeczky 2013:160). As in the case of LRA1, most assemblages seem to be surrounded by the Northern frontier and along the coast of the Black Sea (Moesia I, Dacia, Moesia II). The largest concentration, other than the Yassi Ada A shipwreck, was discovered at Tomis, Constanza, in modern Romania (about 120 LRA2s) (Karagiorgou 2001:132-39). 32 Workshops: Given the number of workshops and its range (see previously, FIGURE 6), the LRA2 type seems to have been equally as LRA1 was in the Aegean. This is not surprising as either the Greek mainland or the surrounding islands were the birthplace of both types. Both Poulou, Papadimitriou, Nodarou (2007), and Riley J. (1981) believe that the Greek Mainland handled the initial production of this type. Regardless of its initial birthplace, a couple of workshops are responsible for its production. However, only a few of them have been found as of today, and based on the number of LRA2 amphorae found, more workshops are expected to be related to their production. In more detail, for the early type LRA 2A, during the fourth and fifth centuries, a workshop was found at Delion/Dilesi in Boeotia. Another kiln from the late sixth and the beginning of the seventh-century was responsible for the production of LRA 2B in the Argolid (Gerousi 2014). Workshops have also been found in the islands of Kos and Paros that date to sixth and seventh centuries that produced imitations but also a smaller number of some local forms (Diamanti 2010a; 2010b; 2016; Diamanti et al. 2014). Finally, according to Demesticha (2003), Cyprus was responsible for the production of LRA 2C or LRA13 after the seventh century, while no evidence supports the production of earlier LRA 2 forms. Besides the substantial numbers of late fabrics of LRA2 and LRA2 C found in Chios, no definitive production sites of LRA2 have been found there (Reynolds 2017:327-35). However, Opaiţ Andrei and Tsaravopoulos Aris (2011:317) have addressed the production of the Dressel 24 similis, the LRA2 prototype on the island. This fact is rather a suspicious phenomenon since the prototype is not followed by the actual later commonly produced form. Future research on the island might help clear such paradox. Karagiorgou (2001:76-77) and Reynold (2017:335) also suggest potential workshops in Samos and Crete due to the considerable number of assemblages that have been found there. Another important site on 33 which the production of LRA2 is under investigation is that of Burgaz, in the Datça Peninsula (see also Leidwanger et al. 2015; Leidwanger 2019; Sakarya et al. 2019). Overall, one can observe that the production of this type has three phases, the first probably on the Greek mainland and Chios (possibly Delion, or Halieis) for the earlier types (fourth-fifth centuries). The second phase ran from the sixth to early seventh centuries on the Aegean islands and was characterized by, as Reynolds (2017:335, 338) believes, completely new workshops caused by the new administrative organization that resulted from the Quaesturae Exercitus. The third phase, mid to late seventh and eighth centuries, reflects workshops that are yet unlocated. Potential actor areas might exist in Asia Minor, Samos, and Crete. Furthermore, this third phase might also reflect production on the same old workshops yet at a different pace and numbers. Cyprus seems to belong to this category with several kilns responsible for the production of the LRA 2C/LRA13 form which was later adopted on the island. A Final Word for the Major Workshops of Both Types As Stella Demesticha (2013:176-177) and Pieri (2005, 2007:301-302) state, while Cilicia seems to have been involved in a more interregional trade market (at least until the beginning of the sixth century), Cypriot trade was so to speak, more focused on the south and east. Paphos on the other hand could participate as a redistribution center, a “promoter” if one examines the city’s administrative role and strong link between the entire island and the eastern regions (Kaldeli 2013). Cilicia was extremely popular for exporting this type, being evolved with a well-recognized wine trafficking throughout the Mediterranean and was part of a long tradition in maritime activities. Kinglsey and Decker (2001:4-15) inform us that commercial transactions of imported eastern foodstuffs continuously flow on the west, even after the divide of the Roman empire, an action that reflects how commercially successful the 34 LRA1 type was. Pieri (2012:30) states that LRA1 and 2 averaged around 30-45 percent in Marseille, Arles, and Narbonne during the fifth century CE. Such a number should not be solely viewed for the army supply, but instead as commercial, in nature. Early Byzantine Egypt was a massive market, for a wide range of East Mediterranean foodstuffs. Syrian, Cilician, and Cypriot LRA 1 amphora, were extremely popular, on at least one-way traffic. Discussing Amphorae Contents Tackling the containers’ contents is a particularly challenging task as most of these containers would be subject to repeated reuse. However, in the case of both LRA1 and LRA2, one can say at least that they would probably transport olive oil, wine, or even some nonliquid products (at least for the LRA1), based on shape and the existence of resin lining on the interior used to waterproof them (Bezeczky 2013:159; van Alfen 1996:203). John Lund (2023:48-51), noted that oil amphorae seem to have bigger shapes, and thus, larger carrying capacities. They suggest that such a feature can lead towards accepting and connecting olive oil products for smaller containers, which would perhaps be more useful for storing it for a longer period. On the other hand, wine containers seem to have smaller shapes so that the containers could be emptied faster. Since oxygen has a significant impact on the flavor and body of the wine, they likely preferred to drink the wine while it was fresh rather than leave it sitting in a container for long periods. However, it is hard to support these notions based solely on shape, because these differences may reflect regional patterns instead of functional ones. Thus, one cannot ignore the production tradition of their places of manufacture. The LRA1 type was mainly produced in areas with long wine traditions such as Cilicia, Syria, and Cyprus, while the LRA2 type was in areas with the production of olive oil, such as the Aegean islands and the Greek mainland. 35 After all, it seems that LRA1 jars would possibly carry wine (Bass and Van Dornick 1989:252, 327-331; van Alfen 1996:203-209). In the case of the LRA2 containers, despite the large oil tradition of the areas in which they produced them, there is a disagreement, as it was suggested that they would carry wine too (for an opposing view see Reynolds 2017:332). In the interior of one of the Fournoi’s LRA 2 amphora (BE2015/5-55), a liquid residue was found, which was probably the result of grapes and thus wine trafficking. Finally, olive oil is often accompanied by wine on the markets and hence, it is quite common when both are distributed together for logistical purposes (wine was cheaper and more profitable than olive oil). Both would consist of a “complete diet” (“βρώσιν και την πόσιν,” for drinking and eating) of the Roman population with also larger reuse rates, making the distinction overly complex (Demesticha 2020). Epigraphy The last thing that remains to be discussed is the epigraphic sources found on these containers, which consist mostly of stamps at the neck and shoulders. They bear marks of the different forms of economic behaviors. These inscriptions namely tituli picti, or dipinti but also stamps, according to Keay (1984) can be divided into three types: “commercial notations”, “tax notations” and “religious formulae”. Unfortunately, within the corpus that this thesis examines, only one amphora bears an inscription on the lower neck (“ΠΟΡ”), which will be discussed in Chapter 4. In more detail, these inscriptions usually refer to the weight or volume of the commodities transported, the names of people or institutions, abbreviated theological formulas, and Christian symbols (see Fournet 2015:62-68). Finally, both Top-down and commercial purposes are reflected through stamped amphorae; resulting in an unambiguity of their economic ties (Cosentino 2021:24-28). These economic ties and the highlighted complexity will be discussed in the following chapter, in more detail. Chapter 3. From Production to Distribution, and the Complexity of the Late Roman Trade on the East, Through the Study of Maritime Activities and Amphorae Containers (LRA1 and LRA2) An Overview of the Reorganization of the Roman East The task of rebuilding the status quo of a pre-industrial empire, or as Marlia Mango (2009:31) describes it, an “early globalization era”, is inevitably too great for the scope of this thesis. Indeed, reconstructing a comprehensive economic picture of the Late Roman Empire is not the intended purpose of this study. Furthermore, it is suggested here that the notion of one pan-Mediterranean market is incorrect and a problematic concept as the Roman political system encompasses a variety of diverse cultures unequal social structures and overlapping powers yet with the Mediterranean being busier than ever. However, trying to discuss key points that resulted in trade reaching its peak during this Roman “mosaic” within the height of the ancient Mediterranean, and the extent of the economic mobility and connectivity within the world of Late Antiquity, even briefly and schematically, is after all, not only a fascinating intellectual minefield but a worthwhile endeavor. What will be attempted here, therefore, is to sketch out some of the structural determinants of the region’s economic performance for our case study. As it will be discussed later in this chapter, microeconomics, or regional economy, were as significant, if not more so, because they were more frequent and common than interregional scale trade as Walter Scheidel et al. (2007:4) correctly noted it. This early overview of the reorganization of the Roman East is an effort to not only highlight the significance of economic reformations that led the way for more complex activities but also a chance to explain terms that will be discussed later. Even though this present study focuses on the eastern part of the Empire, one cannot also ignore its implications regarding the West. 37 The regions of focus for this study range from the Balkan peninsula, arcing eastward and southward, to Roman Egypt. The Mediterranean climate becomes increasingly arid as one moves to the east and south, creating a smaller micro-climate affecting the agricultural success and the concomitant need for exchange (Grove and Rackham 2001). According to Alcock (2008:671-97), uneven resource division and irregular productivity would require a frequent change of distribution patterns. In some cases, these strong agricultural economies create “geographical pockets with significant cash crops” due to their continuous growth and expansion (Syria, Cilicia). To continue, the Roman Empire emerged from the ill-attested third century, after a military and political crisis. The state's response to the crisis was a set of civil, political, military, and economic actions, reforms, and regulations. Regardless of whether these changes were fully successful or not, they revealed a move towards a more rational and stable economic system, which in turn, boosted maritime mobility as pointed out by Wilson and Bowman (2017). To be specific, during the fourth century CE, according to Gregory (2010:99-100), legislations were established to guarantee the collection of the tax owed to the state and to ensure that the food supplies were delivered to the cities as well as support the essential collegia (occupational “guilds”) of ship-owners and agricultural goods producers, to maintain for the state. Dominique Tezgör (2023:240) identified these changes in Diocletian’s reforms (284-305 CE), such as the Price Edict (301), which set up a strict system of empire- wide price controls and lumped large areas into a single entity called the Diocese of Asia. This resulted in the organization of the Diocesan population’s foodstuff supply and services for military and other official use. Simultaneously, later initiatives of various Emperors like Valentinian’s II Imperial Edict (386), Theodosius II (438), and Justinian’s (545) formations show the effort taken in defining standards of weight and capacity. All these actions should be viewed under the same lens: to facilitate an efficient and effective supply system. 38 Laiou Aggeliki and Cecile Morisson (2007:23-41) highlights the sharp contrast between the Empire’s Eastern and Western parts regarding economic prosperity. The East was more urbanized and prosperous due to its proximity to fertile regions of agricultural production (e.g., Egypt, Asia Minor, Cilicia, and Syria). However, after a series of dramatic events beginning in the mid-sixth century CE: plague (541-42), Slavs and Avar’s invasion (560s), the war with the Sassanians (572), and finally the Arab’s conquest (630s) resulting in the loss of 50 percent of imperial land, the east started showing signs of economic decay. This would have profoundly negative consequences for the Byzantium economy. Reynolds (2017:338) adds that by the end of the sixth century, most of the lower Danube was lost with its remaining coastal ports, as well as much of Greece to the Slavs, by 615-617 CE. This paralleled a shift from tied navicularii (ship owners) to private, free naukleroi (ship owners) in the seventh century, with the application of new sea laws. The Annona System An important system was set in place in c. 330 CE called the Annona. Annona was a goddess of Roman Egypt and reflected the personification of the grain supply. Although it had started unofficially by the time Septimius Severus was on the Roman Imperial throne from 193 to 211 CE, it was not codified until the period of Diocletian’s rule, where it served as an official system to both military and civilian supply networks (McGeer 1991). This system was encouraged in an official capacity, or in other words, “severely underwritten by the state” (Angelo 1942:393-444). Alcock (2008:671-97), states that subsequent land taxation was not a new phenomenon in the area but evolved from an occasional to a regular element of state and troop supply during the Hellenistic period (c. 323-32 BCE) in Late Antiquity. In Late Roman times, however, land taxation became highly regularized taking the form of tax- in-kind. 39 Both the Annona Militaris (providing for the army) and Annona Civica (for the state) were payment-in-kind. As Kingsley (2009:34) comments regarding the Annona Civica, more than 31,200 tons of foodstuff were shipped each year from Alexandria to the new capital Constantinople alone while serving as the new “magnet” of the east, with more than 80.000 beneficiaries, equivalent to 20 percent of the population between 332 and 620 CE. Consequently, a Rome-like system was later adopted for the new capital, with the East relying greatly upon this new system. The Annona Civica served to provide for the dietary needs of the Roman population in major cities, which would be dependent on the foodstuff supply from the state. Amphorae, according to Kaldeli (2013:198-202), were frequently collected and distributed from individual households as taxed products. Thus, middlemen (aristocratic intermediaries) engaged with commercial distribution, which belonged securely in the private sector, while the imperial agents (conductors) responsible for managing imperial land, also collected and distributed the produce from the tenants. The Annona Militaris, on the other hand, was a payment in kind for military purposes was preordained for the armies stationed on the Empire’s borders (Segre 1942:393-444). There are, however, still questions regarding the military Annona as Justin Leidwanger (2023) has articulated. The questions centered on whether it was a specific tax-in-kind levied on the provinces to provide for the army or whether it was taken from a non-specific taxation system (Leidwanger 2023:241). Finally, according to Opris and Ratiu (2019:139-155), the daily food ration for a soldier (through papyri records) consisted of three pounds of bread, two pounds of meat, two sextarii of wine, and 1/8 sextarii of olive oil. This new system, although somewhat ambiguous in its function, combined with the establishment Constantinople created as Reynolds (2017:61-87) and Brennan et al. (2020) point out, new dynamics and trade routes, at least not until the fifth and sixth centuries. Antioch rather than Constantinople was, so to speak, the catalyst of investments in the east 40 during the fourth century. This new era in the Mediterranean’s affairs shifted the Empire’s commercial center of gravity eastward, making the Aegean a vital maritime corridor. Finally, Reynolds (2017:338) demonstrated that not just Constantinople, but also the troops and sailors in the Cyclades received Annona goods rations which can be seen on several Aegean sites. Consequently, a private fleet under state contract was transferring goods from provinces to the capital and the Empire’s borders (Segre 1942:393-444). Quaesture Exercitus Adding to the Annona system, Opriș Carol, and Alexandru Ratiu (2019:139-155) discussed another imperial initiative during the year 536. The Quaesture Exercitus was set in place for the maintenance of an efficient supply of the Annona system for the Danuban provinces, for which Justinian created a new administrative unit, a new “praetorian prefecture.” His sophisticated system combined two Danuban provinces of Moesia Secunda (II) and Scythia, with three Aegean ones: Caria, the Cyclades, and Cy