Commentary - The Roman Empire and its Neighbours

It was always intended that the map should depict the Roman Empire at the apex of its power, prosperity and territorial reach. This restricts the time scope in question to somewhere between the Flavian dynasty and the beginning of the Crisis of the Third Century. In the end, the decision was made to select the reign of Septimius Severus, at the end of that period.

Unlike the territories conquered by Trajan during his Parthian war, the provinces of northern Mesopotamia (created by Septimius Severus) were held in the Empire for more than 160 years. Taken in conjunction with the continuous expansion into the periphery of the Sahara Desert, these conquests give the Roman Empire its largest sustained territorial extent. Simultaneously, this map shows the world of the discontinuing Principate in its final form, before the deep political, economical and religious upheavals of the Third Century.

The map shows the Mediterranean from an Earth orbit perspective. To achieve this effect, the background was composed from landclass and bathymetry data, which was then combined with a shaded relief. Most of the geodata comes from the Natural Earth Project[1], while the shaded relief was calculated using SRTM-based digital elevation data[2]. An equidistant cylindrical projection, centered on a latitude of 40°N, i.e., the center of the Mediterranean, was used. In places where the topography has changed significantly, especially along the coastlines, the geodata was corrected to reflect the ancient topography. The road network is based on the Itinerarium Antonini but was significantly extended for the map.



Most space in the map is given to the cities, which in Severan times still were the mostly autonomous base units of the Empire. Regional differences in the organization of cities are indicated by the city categories: In the Greek and Hellenized East we find a dense network of cities that were mostly organized in the tradition of the Greek Polis. Civitates and their territories in the Celtic northwestern provinces were oriented on the boundaries of conquered Celtic tribes, leading to larger units of territory. Smaller, newer units could be found in the parts of Agri Decumates to the left of the Rhine. In Egypt, the Romans, just as the Ptolemies before them, took over the existing administrative infrastructure and the nomes, which in part had already existed for thousands of years. The Latinized remainder of the Empire followed the Italian model. Cities that were either founded as Roman colonies or later elevated to that status can be found in all parts of the Empire.

The Latin term "Civitas", which literally means citizenry, is a generic term that may be used for all kinds of autonomous communities, including colonies and independent states. On the map, in contrast, the Civitas symbol is only used for the main settlements of the aforementioned mentioned larger administrative units.

The category of non-autonomous settlements on the map includes notable Vici[3] that were part of a larger administrative unit without being the main settlement in that unit, and Canabae, cities that grew around forts and were subordinates to the military. Both types of settlements could attain a respectable size and surpass proper cities in population.

Also marked on the map are notable sanctuaries which, unlike most holy sites, were located further away from any major settlements. In some cases, the accompanying settlement had been abandoned. One example of this is the island sanctuary of Delos. Once a prosperous free port, Delos lost this function during the Pax Romana and did not recover from the devastation of the Mithridatic Wars.

Sites in Southern Arabia are usually marked with the local name, as used in Antiquity, as well as the one recorded by Greco-Roman geographers. A question mark next to the name indicates that the assignment of the name to this particular site is still uncertain. Names in brackets are modern names for sites that, as of 2014, could not be identified with a historical name.



Most of the outer borders of the Roman Empire are well-defined through rivers or fortifications. However, for boundary sections where this is not the case, known outposts are connected with lines as a "de facto" border on the map. The statement of Bowersock, that „Just as the Romans did not have provincial boundaries within the waters of the Mediterranean, they were not likely to have defined clear boundaries in the wastes of the great Syrian desert“  holds true for the boundaries through deserts. In these regions, it was attempted to denote the area under direct military control, so far as it is indicated by inscription finds or similar evidence. One example case of this is the bend in the border line through the Sahara desert west of Dimmidi that was included to take into account CIL VIII 21567 at Agueneb and two possible forts at El Khadra and El Bayadh.

Both the eastern and the southern borders have so far not been as intensely researched as the borders within Europe. The emphasis of archaeological work in the southern and eastern regions has been on large urban centers. Forts often have not been investigated through excavations, making their dating and the question of precursor structures uncertain.



At its peak, the Imperial Roman Army encompassed around 400 alae and auxiliary cohorts, which were distributed over a much larger number of forts, primarily along the borders of the Empire. Of these, only the ones that are essential in understanding the border regions have been included in the map. Another goal of marking these forts was to contrast the staged defense system of the Dacian provinces with the linear nature of the rest of the northern border. In addition, it was attempted to represent the very unequal density of forts that mirrors the non-uniform distribution of troops over the Empire. For instance, while 50.000 - 60.000 troops were stationed in the north of the Danube section of Dacia alone, the entire boundary of Africa, between Sala at the Atlantic and Egypt, was secured with a mere 35.000 troops.

Just as the Legionary fortresses, the garrisons of the special Alae Milliariae are also represented by their own icon on the map. The rare Alae Milliariae were double strength elite cavalry regiments, commanded by equestrian prefects in an extraordinary fourth military position of their cursus honorum.


Mines and Quarries

The map also denotes mines of precious and non-ferrous metals. Most mines were Imperial property and thus contributed significantly to the state treasury. Others, especially iron mines, were so numerous that they produced almost exclusively to fulfill local demand. Regarding quarries, all sites that exported regionally or super-regionally were marked on the map. Again, most quarries were Imperial property, which meant that their products were used for Imperial (but also private) construction projects in the entire Empire.


Roman Territories


Arabia and the Red Sea

There is now some certainty on the southern border of the Roman province of Arabia. A Latin inscription found in Hegra in 2003 proves that this former center of the Nabataean Kingdom was firmly integrated into the Roman province. When also taking into account inscription finds of Roman soldiers in the territory of the former kingdom, it now becomes clear the Nabataean Kingdom was, in its entirety, annexed in 106 CE. The precise location of the Imperial boundary, coinciding with one of the main entry points of the Incense Route, was located approximately 20 km south of Hegra, close to the present-day oasis al-'Ula (in antiquity: Dedan), as is indicated by graffiti of Roman soldiers (including Benificiarii) found there.
Inscriptions demonstrate that the Dumata oasis, located deeply within the desert, belonged to the Nabataean realm of influence. In Severan times, roman troops based in the fort at Qasr al-Azraq patrolled the Wadi Sirhan. This Wadi connected Dumata to the provincial capital Bostra, an important trade route and attack vector to the urbanized core of the province. In the oasis itself a ceremonial altar, erected by a centurio of Legio III Cyrenaica based in Bostra, was found, which can be dated to the reign of Septimius Severus, proving the imperial army was in control of Dumata.

The status of the old oasis city of Tayma in Roman times is unclear. Excavations show that it remained an urban center and that it belonged to the Nabataean Kingdom. Taking into consideration the fate of other Nabataean properties, it is only reasonable to assume Roman dominance, as is the case on the map.

The Roman base on the Farasan Islands (lat. Portus Ferresanus) has been verified through inscriptions for the year 144 CE and for the reign of Trajan. Since the installations could so far not be excavated, there is no clarity on whether or not the base was in use in the year 211 CE, nor on how long it existed. The fact that the base was in use for at least a few decades, thus surviving various changes in leadership, and the fact that the islands had their own prefecture in 144 CE, indicate a certain permanence of the installations.

Archaeological evidence from the harbor cities of Egypt show no decrease in activity before the crisis of the third century. To the contrary, the Roman Empire continually extended its regional power in the Orient (projected through bases) after the successful Parthian wars of Lucius Verus and Septimius Severus. As an example, there is the aforementioned inscription from Severian times in Dumata that shows trade routes and military access vectors being important drivers of imperial policy. In addition, more recent excavations in Myos Hormos and Berenice repeatedly produced evidence of a continued presence of Roman Naval units on the Red Sea. Because the Empire considered southern Arabia to be in its realm of influence, with the corresponding diplomatic relations, there must also have been credible options for military intervention. A base on the Farasan islands would have been extremely useful in this regard, which fuels speculation about predecessors and additional bases. All these considerations, together with the desire to visualize the extent of Roman influence, led to the inclusion of Portus Feresanus in the map.

New archaeological finds fuel the old discussion on whether a separate classis Maris Rubri existed or if the ships were subordinate to the fleet in Alexandria. In any case, there must have been a so far unidentified naval base on the east coast of Egypt. Being the seat of the prefect in charge of the region, Berenice seems the most likely candidate. There is some hope that these questions will be answered in the near future, through the continuing investigations in Berenice and Myos Hormos, as well as the evaluation of existing written evidence.


Traveling Times in the Red Sea and Beyond

All times noted on the map are the actual travel times for the corresponding route. Given the wind system of the Red Sea and the Indian Ocean, the travel times along a multi-station route may deviate drastically, which can be misleading [4].

Ships had to leave Egypt in summer so that they could cross the Red Sea using the predominant northerly winds of this season. In the Gulf of Aden, they would then encounter the Southwest Monsoon, which allowed to continue towards India, but made a change of course on the Horn of Africa (Aromata Promontorium), so as to travel in a south-western direction along the East African coast, impossible. Doing so only became possible from around November, when the Northeast Monsoon begins. Likewise, the ships had to wait for 8 months in Raphta for the return of the Southwest Monsoon. Back in Egypt, another 6 months had to pass before the next journey could be started in the following July. With the then common rigging (just as with more modern rigging), travel against strong Monsoon winds was practically impossible.



The development of administrative structures in the new trans-Euphratic territories during the time of Septimius Severus cannot be unambiguously reconstructed based on present-day knowledge. This is especially true for the province of Osrhoene (as evidenced by inscriptions) that most likely was put in place already in 195 CE, after the first Parthian War of Septimius Severus. The map follows the conventional wisdom that it existed as an independent province.

However, it is not unlikely that Osrhoene was a sub-province adjunct to Syria Coele, subordinate to the governor there. In this case, its territory would be smaller than shown here, without Rhesaina and the remaining Habur area. After 205 CE, Osrhoene was de facto unified with Mesopotamia and was administered from Nisibis. The commonly accepted date for this is the year 212 or 213, when the client kingdom of Regnum Abgari around Edessa was annexed. Nonetheless, an earlier date, for instance during the reorganization of border defenses in the last years of the reign of Septimius Severus, remains possible.

Regarding the Regnum Abgari, a residual kingdom still attributed to the former King of Osrhoene, Abgar VIII., no precise borders, except at one place, are known. It is usually assumed that this territorial entity was from its inception limited to the city of Edessa and its immediate surroundings.

The city Amida was already raised to the rank of Colonia during Severan times. This also could already have been ordered by Septimius Severus himself, or equally likely by one of his successors.


Surrounding Territories


Aksum and Southern Arabia

At same point between the very late 2nd and the first decades of the third century Aksumite troops crossed the narrow stretch of sea between Adulis and southern Arabia to launch an invasion of Himyar. They enjoyed some initial success, apparently Aksum controlled the coastal plain for several decades and could, in concert with allied local chiefdoms, launch attacks against the inland centers from there. It wasn’t until the later decades of the 3rd century that the invaders could finally be dispelled. By the 4th century rulers of Aksum could, now anachronistically, claim to be kings of Himyar, Raydan, Saba and Salhan as part of their royal titles. However, unlike the later 6th century Aksumite intervention in southern Arabia, this earlier episode is only poorly understood and all details are still a matter of debate. No complete account has survived and evidence is mainly limited to contemporary local epigraphy. Thus other than depicted on my map it may very well be possible that Aksum already held some southern Arabian possessions in 211 AD.



Reconstructing the situation in Germania at the beginning of the third century posed a problem. Lacking literary evidence from the cultures in question, modern researchers are completely dependent on writings of Roman and Greek authors living far away. While the source base is relatively acceptable for the first century due to original authors such as the historian Tacitus and the Geographers Strabo and Pliny the elder, literary evidence becomes scarce during the time of the Adoptive Emperors and the Severans. For instance, the Geography of Claudios Ptolemaios, dating from the second century, is problematic as evidence since his sources are unknown and seem to dating from a large interval of time.

The dynamic changes in power balance, dependency relations and settlement patterns apparent from these works of writing cannot be clearly retraced for the 100 years that follow. Just as in the Roman realm, Germania experienced massive upheavals during the third century, such as the migration of the Goths to the black sea or the formation of large meta-tribes such as the Franks or the Alemanni. The ethnogenesis of the latter tribe is a controversial topic of present-day research, which also affects this map.



Currently, there are two competing models for the formation of the Alemanni. On the one hand, it is assumed that they originate in Suebian Tribes (or parts thereof), such as the Hermunduri and Semnones, which supposedly migrated to the river Main area shortly after the Marcomannic Wars. According to this theory, this new, militarily powerful union directly adjacent to Roman territory pushed Emperor Caracalla towards a proactive campaign in 213 CE. It follows that the Alemanni would later have been instrumental in the fall of the Limes Germanicus in 260 CE, and would have settled in the former Agri Decumates territory, the area between Limes, Rhine and Danube (Grid D2 to E3 of the map).

The second theory, which is preferred in contemporary research within Germany, assumes a haphazard abandonment of the Limes Germanicus over time, as key troop concentrations were withdrawn to other hot spots during the chaos of the third century. This would have allowed the Suebian (and various other Germanic groups) to settle in the now accessible former Agri Decumates territory. Under this hypothesis, the Alemanni (whose first definite evidence of existence dates to a Panegyricus from 289 CE [5]) would have formed there, within what was once the Roman realm.

A key factor for the plausibility of this scenario is the validity of references to the Alemanni in various Roman reports from the first half of the third century. Particularly relevant are the two contemporary authors Asinius Quadratus and Cassius Dio, the latter of which writes about the campaign in 213 CE, thus supplying an alternate first notification of existence. However, their works for this period are only available via later summaries and quotations in Byzantine sources. The interpretation of the people against which Caracalla campaigned as Alemanni in the writings of Cassius Dio is controversial, because their name appears in three different variants throughout the various fragments, some of which deviate strongly from the standard Latin spelling.
I used the first hypothesis for this map, since I consider the arguments for[6] the veracity of these writings to be stronger. Accordingly, the bulk of the Semnones, who settled between Elbe and Oder in the area around the Havel and the Spree during the first and second centuries, are already shifted towards the south-west and the Roman border on the map.



Another special case is the population that is described as “Cotini” or “Gotini” in Roman sources and has been given the name “Kotiner” in the German literature. These people are likely identical to the archaeologically identified Púchov culture (in the later phase), which had its center in present day Slovakia[7]. Many of the properties ascribed to the Cotini, such as the settlement area, their practice of iron mining and processing, their culturally Celtic traits, and their political dependency on the nearby Quadi, correspond to evidence in the archaeological record of the Púchov culture.

The Púchov culture settlements were abandoned in the late second or early third century. In the known fragments of the history by Cassius Dio (whose contents range until 229 CE), they are described as having been eradicated later, sometime after the Marcomannic wars[8]. Unfortunately, the segments of his work concerning this period have not been preserved. Two inscriptions from the city of Rome dating to the reigns of Severus Alexander[9] and Decius[10] speak of Cotini from the province of Pannonia Inferior[11]. It is assumed that some time at the end of the Marcomannic wars, probably around 180 CE, the remainders of the Cotini populace were resettled to this province, which had been depopulated by war and the Antonine plague. Whether a part of the populace stayed behind, or if the source area was uninhabited for the next decades is unknown. Considering the general uncertainties, I decided to also mark the Cotini at their ancestral location, since the map is supposed to give an overview of the last 135 years of the principate.



Similar conditions apply to the Saxones. For them too, the veracity of the first mention by Ptolemaios in the second century is in question, and their name is not recorded correctly in all manuscripts. However, since the Saxones do not appear in any source that is relevant for the time scope of this map, and since they make an appearance as a regional power only in at the turn of the third and fourth centuries, I left them out of this map. If inserted into the map, they would be located east of the Chaucii in the general area of the mouth of the Elbe.



Only few written sources from the Roman realm exist for the Ireland of antiquity. The depiction in this map is based largely on Ireland and the Classical World by P. Freeman, in which the author has united these sources with the archaeological evidence.

A series of Roman coins from the late first to early fourth century and various valuables indicate that the neolithic burial mound of Brú na Bóinne/Newgrange had regional religious significance.


Oasis Cultures and Trans-Saharan Trade

Since the 1990s, after decades of neglect, archaeological research fortunately resumed the investigation of Saharan cultures, which allowed an extraordinary improvement of our knowledge of the region to take place. Of special note is the work of the Italian research mission to the Tadrart Acacus mountains and its surrounding oases (near present day Ghat[12]) led by Mario Liverani and the achievements of the Fazzan project (later: Desert Migrations Project) in central-Libyan Fazzan led by David Mattingly[13]. Both regions are considered to be part of the Garamantian kingdom, whose core territory in antiquity was the Wadi al-Ajal with the capital Garama (present day Jerma). Some doubts, however, remain regarding the area surrounding Ghat.

Compared with research on the ancient Mediterranean, research on the ancient Sahara realm is overall still in its infancy. For instance, the Kufra oases[14] and their surroundings belong to the least explored sites among all presented here.

Regarding both of the two main questions relevant for this map, i.e. the existence of continuous trade in the pre-Islamic Sahara and the existence of an organized state structure for the Garamantes, the investigating archaeologists have converged on confirmation[15]. Both had been strongly doubted or even disregarded in the past[16]. While in earlier research the focus was firmly on trans-Saharan trade towards the Roman cities of Africa (not unlike the better documented medieval trade), newer results show that there were population centers within the desert capable of generating significant imports and exports even in antiquity. Thus, the trade relations between these centers also have to be taken into consideration. With regard to the importance of this trade to the Roman world, it has to be noted that, at the time, the North-African provinces of the Empire already had hundredfold the population of the Garamantian realm (with the entire Roman Empire having thousandfold the population). Exports and imports that were important for the economic development of the Fezzan were therefore never comparable in importance to the trade with the East and Africa that took place via the Red Sea and the Caravan cities of Syria and Arabia. However, they did still contribute to the blossoming of the coastal metropoles of Tripolitania.

In addition to the archaeological record of trade goods, there are some ceramics finds from the Roman era along the Abu Ballas Trail[17] and some Berber-Libyan inscriptions in the Selima Oasis[18] proving that the Sahara was also crossed in all directions away from the main routes by mobile groups.

Differences in the degree of research activity can also be noted in the used and marked paths between the oasis groups. While the routes of antiquity in the deserts west and east of the Nile in Egypt are well-documented through archaeological prospection, the pathways in the central Sahara region can often only be extrapolated indirectly via the distribution of trade goods and the natural conditions, such as water sources.




[2] SRTM30 Dataset with 1 km resolution. Source: U.S. Geological Survey,

[3] The Latin term Vicus originally designated a quarter or suburb of a larger city, in our time the meaning was expanded to include seperate settlements that legally belonged to their larger city state.

[4] See L. Casson, Rome's Trade with the East: The Sea Voyage to Africa and India, Transactions of the American Philological Association 110 (1980) 21-36.

[5] An overview over this thesis and supporting literature is given e.g. by W. Pohl: "Die Germanen"

[6] As shown in Bruno Bleckmann: "Die Alamannen im 3. Jahrhundert: althistorische Bemerkungen zur Ersterwähnung und zur Ethnogenese"

[7] K. Pieta: "Die Púchov Kultur".

[8] Dio. 72.12.3

[9] 222 to 235 CE.

[10] 249 to 251 CE.

[11] CIL 06, 32542 and CIL 06, 32557, ..ex provincia Pannonia inferiore cives Cotini..

[12] Grid D8 of the map

[13] Grid E8 and F8 of the map

[14] Grid G8 of the map

[15] The recent survey article WILSON 2012: "Saharan trade in the Roman period", for spezific regions for example MATTINGLY 2013a: "The first towns in the central Sahara", MORI 2010: "Archaeological Research in the Oasis of Fewet and the Rediscovery of the Garamantes", or for classical Siwa KUHLMANN 2007: "Das Ammoneion – ein ägyptisches Orakel in der libyschen Wüste".

[16] For example SWANSONS 1971: "Myth of Trans-Saharan Trade during the Roman Era"

[17] Grid H8 to G9 of the map, HENDRICKX 2013

[18] Grid A8, Ancilliary map Aethiopia-Kush, PICHLER 2005