Author: Vladimir Kondić, former director of the Institute of Archaeology of the Serbian Academy of Sciences and Arts, author and editor of numerous important books and journals in the field of archaeology
This paper was presented at the very first Scientific Conference “Danube – River of Cooperation” and published in the, now out-of-print, proceedings. Reedited for Danubius by Enisa Imamović.
The Đerdap, or Iron Gate region encompasses the banks of the Danube from Golubac (Roman Cuppae) to the mouth of the river Timok (Roman Timacus) with its immediate hinterlands. In this region for millions of years the Danube cut its way through the rocky massif of the Transylvanian Carpathians, forming not only the largest but also one of the most beautiful river gorges in Europe. Between the river and the mountains of Homolje, Miroč, and Deli Jovan on its south bank, the space available for human habitation is very limited. In some parts of the Danube gorge, there are many submerged reefs, rapids and cataracts, while in the Great and Small Gorges in the middle of the canyon, the Danube is the deepest river in Europe. In antiquity the river often froze during severe winters, making crossings very easy during those periods. From the end of the Gorge, at the Roman castellum Diana, to the mouth of the Timok river, the Danube again becomes a broad and smooth flowing stream with numerous islands and sandbars. Inland from its banks, endless plains, fields, and forests spread. In this Iron Gate region, which is some 120 Roman miles long, some of the oldest cultures connected with the origins of European civilization developed. The river also separated two worlds which, most of the time from the prehistoric period to the Middle Ages, were in opposition to one another. From the north side of the river arose problems and threats to the peoples occupying the south bank, and it also happened that it was on this stretch of the river that the Roman empire steadfastly defended the integrity of its borders for four long centuries.
Roman interest in the Balkans was heightened immediately after the victory at Pydna and the conquest and annexation of Macedonia; at that time the new territory faced many dangers from barbarian tribes in adjacent lands, including some from the Danube system. The historical sources bear witness to many expeditions which the Romans undertook to punish neighbouring barbarians, but in actual fact they did not reach the Danube before 75-73 BC, when the consul Gaius Scribonius Curio advanced along the Timok valley and arrived at its confluence with the Danube. This event, however, did not cause any lasting consequences. The Iron Gate region remained as it had been previously.
It is difficult to claim with certainty what the ethnic composition of the population was like at the time of the complete conquest and annexation of Moesia. After the very powerful tribe of the Triballi, which had occupied the middle and lower Danube valley from the sixth to the third century BC, the ethnic composition was reduced by conflicts with the Scordisci, its survivors moved eastward towards the area of the later Trajanic bridge crossing. It is known also that the Moesi inhabited the territory between the rivers Cibrica (Ciabrus) and Timok. In the same territory some smaller barbarian tribes, the Celegri and the Timaci, are also mentioned. In campaigns during 29 and 28 B.C. the proconsul of Macedonia, Marcus Licinius Crassus, the grandson of the triumvir and a person who had joined Octavian just before the battle of Actium, added to the empire a new territory which was to become the province of Moesia. Campaigning against the Bastarnae, the proconsul reached the Danube, first encountering the Moesi and, after bitter battles, conquering their territory. Augustus formally proclaimed this event in 27 BC in Rome. In the older literature it was often thought that the Iron Gate region was inhabited predominantly by the Scordisci. According to the Roman historian Appian, after their defeat in 84 BC, the surviving groups of the Scordisci withdrew to the south bank of the river and to some islands in the stream. New excavations, however, suggest that the Scordisci left behind few traces in these places and most likely settled much farther to the west in the modern region of Srem. The Iron Gate, in general terms, was the territory of the Dacians and the Getae, tribes that were united about the middle of the first century BC under the leadership of Burebista into a cohesive and strong confederation, which was a clear forerunner of the Dacian state whose strength the Romans were to experience to their disadvantage in the first century AD
After the proclamation of the annexation of the new territory, there is almost no information about this region in the texts of the Roman historians. However, we do know that in the beginning Moesia was both administratively and militarily subject to the province of Macedonia and Achaia. It is thought that Moesia received independent military status some time between the years 2 and 6 AD and an independent civilian administration at the beginning of Tiberius’ reign. Due to the uncertain and insecure situation in neighbouring Thrace and the threatening presence of aggressive population to the north of the Danube, from the very beginning two legions (the Fourth Scythica and the Fifth Macedonica) were stationed at Moesia. We still do not know exactly where these legions were stationed in the first century A.D. It is quite possible that the permanent garrisons of these units were located in one of the big provincial towns, the future locations of the legionary bases at Singidunum, Viminacium, Ratiaria, Oescus or Novae. In any case, vexillations of these units must have been stationed in auxiliary camps and smaller forts along the Danube.The salvage excavations conducted in the Iron Gate region enabled detailed investigation of these auxiliary camps. The camps were, moving from east to west, Novae, Smyrna, Taliata, Transdierna, Diana, and Pontes. Some of them, like Cuppae, Egeta and Aquae, still have not been examined. All of these camps were, undoubtedly, constructed immediately after the arrival of the legions Fourth Scythica and Fifth Macedonica, that is in the last decades of the first century B.C. At this early stage, until the time of Vespasian, camps on the frontier consisted of earthwork fortifications with wood palisades. Their existence and design are well documented at Novae, Smyrna, Diana and Pontes. Some time later, the wood and earth constructions were replaced by stone walls. The first camps had the classical plans which are well documented in the ancient sources, especially by Tacitus. They were of quadrangular shape with palisade ramparts and four gates with wood towers. Interior buildings were also made of wood. Sometimes these forts had a defensive ditch, a so-called spitzorabe outside of the wall. The castellum Diana, where excavation is still in progress, revealed the well preserved remains of earth fortifications in which three variants of fossae and the remains of a wood gate and palisade could be distinguished. There is no doubt that the oldest ditch should be dated to Trajan or even to an earlier period.
The first stone fortifications were constructed just before Domitian’s Dacian wars (for example, at Taliata and Smyrna). In the winter of 85/86 AD a group of Dacians led by Diurpaneus crossed the frozen Danube and made a totally unexpected attack on the military camps and stations on the south bank. Furthermore the Dacians defeated the army of Gaius Oppius Sabinus in a decisive battle, in which even the provincial governor was killed. During this first series of attacks some of the camps were burned to the ground (for example, Diana). The emperor moved quickly to the Danubian frontier and appointed Cornelius Fuscus commander of the army. He also reorganized the military and civilian administration of the province by dividing it into two parts: Moesia Superior and Moesia Inferior. The new governor of Moesia Superior defeated and expelled the barbarians, and restored all the Iron Gate fortifications. At this time Decebalus became the new Dacian chieftain, and the Iron Gate fortifications experienced new destruction from renewed Dacian attacks. Domitian, in a very short period of time, suffered two severe defeats on the Danubian frontier. Later, in 88 AD, he returned to the Danube and organized a punitive expedition which defeated Decebalus in the battle of Tapae, in the Romanian Iron Gate. By terms of the treaty which was signed after the battle, Dacia became a client kingdom of Rome.
Already at the beginning of Tiberius’s reign, it had become obvious that the contemporary geopolitical situation concerning the Dacians and Roman plans for Thrace and the kingdom of Rhescuporis, in conjunction with the Greek colonies there, required urgent construction of communications between Pannonia, the lower Danube and the Balkan coast of the Black Sea. We should keep in mind that the greatest concentration of military power in Moesia and that many regions depended completely on the mobility and speed of those military forces on the Danube. Already in 30 AD a road was marked out through the Iron Gate gorge, and its actual construction was begun by the two Moesian legions, the Fourth Scythica and the Fifth Macedonica. The inscriptions carved on the cliff faces above the road bear witness to all phases of this work. These are the Tiberian, Claudian and Domitianic inscriptions. As well as it is known, the road carved into the rock was two meters wide and then extended in width by six meter-long beams set into the rock, so that the wood part of the road, three or four meters wide, was directly above the water. This highly elaborate, technical undertaking was executed on the 35 Roman mile long section of the road, from Cuppae to Taliata. There the road left the river bank, turned inland over the Miroč mountains (Gerulatis), and then returned to the Danube at Egeta. The work on the lower part of the Iron Gate road was continued in connection with Trajan’s preparations for his Dacian campaigns. The road was well protected with numerous watchtowers at Gospođin Vir, Pesaca, and Lepenski Vir, and also means of auxiliary camps.
At the beginning of his reign, the emperor Trajan decided to eliminate the hundred year old Dacian threat on the northern frontier. In the winter of 98/99 AD he arrived on the Danube and most resided at Diana. He concentrated his major war preparations at the Gate. He extended the road in the gorge for 30 miles more and marked his accomplishment with the well-known inscription of 100 AD. He also, for the first time in history, made this part of the Danube navigable by constructing a canal to bypass the impassable cataracts near the castrum Diana in 101 AD. This achievement is recorded by an inscription on a marble plaque which reads
“that because of the dangerous cataracts he diverted the river and made the whole Danube navigable”: (ob periculum cataractarum, derivato flumine, tutam Danuvii navigationem facit).
Trajan restored earlier stone fortifications in the Iron Gate and rebuilt all the earthwork fortifications in stone. These were then the preparations for the beginning of the first Dacian war, and land and river communications for supplying the army were set in place. Just below the castellum Pontes a large river port and massive horrea (300 meters on a side) were built. The first Dacian war was successfully completed in 102 AD. During the peaceful interval between the first and second Dacian wars, from 103 to 105, the imperial architect Apollodorus of Damascus designed and constructed a magnificent bridge across the Danube, one of the greatest achievements in Roman architecture. The bridge between Pontes and Drobeta had a portal at each bank, twenty piers in the river, and four piers on the river banks, all of them connected with arches. The length of the bridge was 1139.90 meters.
Some basic details about bridge are supplied in the ancient literary sources, although unfortunately the most important ancient treatment of the structure, a work of Apollodorus of Damascus, is not preserved. In that text, the architect described his own work and discussed a number of technical details. Dio Cassius (Historia Romana 68. 13) also wrote about this bridge. From the chronological information he supplies, it is possible to date precisely the time of the construction of the bridge. In addition Dio notes that the bridge was built of stone, which is only partially correct because it is clear from coin representations of the bridge that its superstructure was constructed in wood. Along with the information already cited, he informs us that the bridge was 150 paces high (50 m.) and that the interval between the piers was 170 paces (56.7 m.). He goes on to state that it was built on a stretch of the Danube which was narrow and the stream was fast-flowing and deep. These details are also not completely correct because in that section the river extends to almost its greatest width with an average depth and very strong current. A survey which we completed in 1983 with sonar and the help of a diver verified all the remaining technical details and measurements which Dio transmitted to us. Procopius in a passage (De Aedificiis, 4.6) before his description of the bridge cites the lost text of Apollodorus and in another passage mentions the diversion of the river. This last has been interpreted in the literature on the bridge as one of the possible means of facilitating its construction.
The grammarian Joannes Tzetzes (Historiarum Variarum Chiliades 5. 61-73, 88-94) also supplies details completely in agreement with those provided by Dio Cassius. He, however, cites an interesting technical detail from Theophilos Patrikios (Theophilos borrowed it from Apollodorus’ text) and relates that Apollodorus, before constructing the piers in the bed of the Danube, made caissons to facilitate the construction of the piers. The designer and builder of this part of the project must have had t0 solve unusually complicated technical problems. Because of the strong current and a depth in the channel of about 20 m., it appears that contemporary technology did not allow for the employment of caissons. However, on the occasion of the underwater survey mentioned above, on the eleven piers still preserved in the river (of the original piers, seven were carried away by the current and two were destroyed with explosives by the Romanians because they obstructed navigation), we found well-preserved wood beams the caissons had been made from. There remains, however, the general question of how the caissons were placed in position.
Finally, Dio Cassius left behind one comment which creates a last point of confusion. He states that Trajan’s successor, Hadrian, rendered the bridge incapable of use by removing its superstructure. It is virtually impossible to accept that comment when it is known as a fact that Hadrian began to accelerate the romanization of the new province and that the bridge to him was the best and most secure link with Dacia.
After the completion the bridge, the final victory over the Dacians and Decebalus’ suicide, the large province of Dacia was formed north of the and the Iron Gate limes lost its primary importance. Smaller fortifications were abandoned, and in others garrisons were considerably reduced. At this point the Iron Gate gorge entered a hundred-year period of peace, the only of this kind in the history of the Roman occupation of this area.
At this time it is difficult to ascertain the reasons that made Septimius Severus pay so much attention to the Moesian limes and grant such great benefits to that province and Dacia. It is possible that he already had anticipated future danger from the barbarians in the Black Sea area, and so, along with the new limes in Dacia the Limes transalutanus) he reconstructed the limes in Moesia Superior. During the reign of Julia Dornna and Caracalla certain reconstructions of fortifications also took place. In any case clear phases of restoration can be noticed at Pontes where, as it was the case in so many other Iron Gate forts, the original Trajanic plan was preserved with certain alterations of the gates and towers. Similar is the situation at Novae and Diana where the porta principalis sinistra and dextra were completely reconstructed, while the porta praetoria remained unchanged. In addition to already existing military camps, a new castellum – Campsa/Ravna – was built on an island at the Porečka river during the Severan dynasty. It is interesting to note that this castellum is of a quadriburgium type, with a square ground plan and projecting square towers at the corners. The dating of this castellum to the beginning of the third century is absolutely certain, because of the archaeological material and the coin findings (especially the hoard of denarii in which the latest coin was an issue of Maximinus Thrax from 237 AD). After this great Severan restoration, which could easily be considered as a distinct phase in the development of the limes fortifications, no significant changes were executed in any fort. In the middle of the third century certain repairs on walls as well as rebuilding in the interiors are evident, but these did not change the original function of the fortifications. New excavations at Diana have shown that there were some destruction and burnt damage due to barbarian attacks from the reign of Philip the Arab to that of Aurelian when Dacia was partially abandoned.
After the dramatic events on the Danube frontier in the last quarter of the third century, when the province of Dacia was almost entirely abandoned and the south bank of the Danube with its few fortresses became the last line of defence, changes in the conception and organization of the defensive system on the middle and lower Danube became inevitable. Significant works of restoration and the building of new fortifications were clearly observed during the archaeological investigation of the Iron Gate region. This shows clearly that the largest new constructions, with precise plans and locations at almost uniform distances from one another, were executed under the first tetrarchy and then continued during the reign of Constantine the Great. In the second half of the fourth century (in the 370’s), during the reigns of Valentinian and Valens (especially after the death of the latter) the Iron Gate fortifications underwent more alterations, but on a smaller scale than before, with only repairs and reinforcements of structures.
During the period in question, the reigns of Diocletian and Constantine, all of the older fortifications (that is, those at Diana, Pontes, Taliata, Novae, Smyrna and Campsa) maintained their function, but with significant changes in ground plan. Exterior towers of various shapes (rectangular, square, horseshoe and fan-like) were added, and the more important forts (for example at Diana and Novae) were enlarged.
At this point it seems that one of the most important results of the twenty years of salvage excavation on the Danubian limes in the Iron Gate is that we have acquired archaeological proof of the significant work of the two great military reformers, Diocletian and Constantine. Their extensive activities of reconstruction and fortification on the limes are evident in new forms of fortifications. One form is the quadriburgium (for example at Sapaja and Sip), with a symmetrical square plan and projecting square corner towers. For the lower part of the Iron Gate, the other characteristic form was the so called burgus, which has analogies in Gaul and Pannonia, dating in the latter region to the time of Valens and Valentinian. The fortifications excavated at Hajdučka Vodenica, Donje Butorke, Glameja-Rtkovo, Mihajlovac, Mora Vagei, Bordzej are also evident at other sites, represent the most striking examples of Diocletianic and Constantinian activity to strengthen the Danube frontier. These fortifications have the form of a symmetrical square (with dimensions of 19.5 meters on one side) with a single interior tower supported on four massive piers. Certainly such strongly founded towers, raised high above the perimeter walls, were watchtowers. Keeping in mind the disposition of these buildings and their mutual visibility, it is obvious that their function was to signal impending danger. The dating of these small fortifications is not doubtful. The type of construction and the quality of workmanship, as well as the numerous archaeological finds, indicate the period of the transition from the third to the fourth century. This dating is strongly supported by an important find from Donje Butorke. It is an inscription found in situ (in the area of the tetrapylon) that is precisely dated to 292 AD. From the inscription it is obvious that the rulers of that time were responsible for the construction of the praesidium (… praesidium constituerunt)
Although it is very probable that Valentinian and Valens undertook reconstruction work on a larger scale because of the Gothic threat, the archaeologically most noticeable phase is the period of horrible destruction and fire immediately following the battle of Harianople. But, basic fortification elements like walls, gates, and lowers remained almost unchanged.
In the following period, under new conditions, when units of foederati protected the frontier, all the ruins were filled in and levelled, making a platform where the limitanei built the wattle and daub homes which are so evident at Pontes and Diana. This period in the history of the frontier lasted until the middle of the fifth century. Thanks to the fortunate discovery of five solidi (the latest an issue of 443) of Theodosius II we know the exact termination date of the Iron Gate Roman limes. The Huns’ invasion from the direction of Niš (Naissus) caused such destruction that Procopius correctly described the situation as disastrous. Fortifications were razed to the ground, and at Diana the south wall with its gate and the fourth-century porticoed building were destroyed. Thick layers of burnt rubble, building debris and ash covered most of the fortress and mark the end of the five centuries of its existence. Other fortresses suffered a similar fate. This period was the terminal phase of the restored, late Roman limes and the northern frontier of the empire. The final renaissance of the Danubian limes occurred under Justinian I.
The significant testimony of Procopius concerning the renovation and reinforcement of the Danubian frontier has been confirmed in its entirety by our recent archaeological research. Procopius paid considerable attention to construction work on the Iron Gate frontier (limes) and provided at times rather detailed information about the former Roman frontier. The sequence in which he comments on the fortifications in those sectors which have been investigated make it possible to identify the Roman and early Byzantine toponyms for some sites whose ancient names were not known previously (e.g. Kantabaza, Smyrna, Campsa). Furthermore, excavations in the Iron Gate gorge have demonstrated that Justinian’s builders in the early Byzantine period entirely retained the disposition of fortifications from the former Roman frontier. Some elements of the earlier Roman castella were altered, most likely because of the requirements of a new defensive strategy, and at locations which were in greater danger because of their topographic circumstances completely new fortifications were constructed. Now it is possible with complete certainty to reconstruct the composition of the Justinianic limes on this part of the Danube. The fortresses can be divided typographically into the following groups:
- Renovated Roman auxiliary and other minor forts.
- Renovated late Roman burgus – forts (from the Diocletian and Constantine periods).
- New early Byzantine forts built around renovated late Roman burgus-forts.
- Completely new early Byzantine forts.
After the Huns’ invasion in 443 AD damages to the forts were not repaired until the early sixth century, which for this sector of the Danubian limes is the only period devoid of any traces of activity. Then in the early Byzantine period all the auxiliary bases on the limes were renovated. The former Roman forts for the most part were renovated on the basis of their original plans. The most frequent alterations which can be observed are the closing-off of gates. These were either walled up or replaced by large rectangular or circular towers. Usually the corner towers were completely rebuilt. At the fortress Diana (early Byzantine Zanes) at the southeast
corner a new tower was built in a horseshoe-shape with an apsidal termination, and two fortification walls were joined together in a point to form a type of bastion. The southern wall and gate, which had been razed to their foundations by the Huns, were rebuilt in exactly the same plan as before and the gate remained the only one in use. In the interior of the fortress, without any type of regular disposition, buildings of wood, earth and courses of poorly joined stones were erected. At Novae (early Byzantine Nobas) the former Roman south gate was completely closed-off and new circular towers were built in place of the earlier east and west gates. All the towers in this fortress were built afresh, with circular plans. The situation is similar at other forts. Everywhere fortification walls were significantly reinforced, most often from the inside. At the former Roman quadriburgium Campsa, the alterations were somewhat more radical. The south gate was closed-off and two new U-shaped towers were added there. Additionally, all of the auxiliary bases contained solidly built, single-nave churches.
The second category of renovation was the least complicated. The Diocletian-Constantine period castella received reinforced fortification walls (cc. one meter thick), and new entrances without towers, features not previously present in these complexes, were constructed. Certainly the most interesting form of renovation consisted of the erection of completely new and characteristically early Byzantine fortification walls around the former burgi. In these situations the renovated burgi functioned as watch towers. Two outstanding fortifications of this type are Glamija and Donje Butorke. The latter has a more complex plan, with piers on two of the towers and one rectangular tower with an apsidal termination. This type of fortification recalls in a certain sense an inaccurate statement of Procopius (De Aedificiis, 4.1) in which he states that Pincum, Cuppae, and Noveae were formerly only Roman towers around which Justinian caused buildings to be erected and to which he granted municipal status after their defenses were strengthened. As mentioned above, during the Roman period civilian settlements of a type which did not exist in Justinian’s time developed around the auxiliary bases. Could it be that Procopius in his exaggeration of credit to the emperor actually had in mind the construction of new fortresses around earlier Roman towers?
Finally, the last group consists of purely Justinianic castella which were completely new constructions. Up to date six of these have been discovered on the Iron Gate section of the limes. Saldum (Kantabaza) in plan is an irregular rectangular with three circular towers and a single elongated one with an apsidal termination. The fort at Bosman is the only complex with a triangular plan on this part of the Danube and is skillfully into the restricted space between the mountain range and the river. The eastern fortification wall, located right on the river, was laid out in a convex line so that high water levels on the river would not be able to damage it seriously. The fort at Hajdučka Vodenica was constructed on the site of an earlier tower which was not renovated in the sixth century. It is situated high on the river bank, and from each and of its northwest perimeter wall extends a fortification wall with a tower at its end to protect a small river harbour. The forts at Milutinovac and at the mouth of the Slatinska river are very similar in both construction and size (55 x 55 m.). They are defended by circular towers with square foundations on defensive walls which are turned toward the river and form the foundation for an upper-level entrance.
In almost all of the fortresses of Justinian time one layer of ash and destruction debris can be observed which can be dated to 580 AD when a forceful Slavic incursion on this part of the Danube was recorded. However, the fortresses themselves did not experience such significant destruction that they could not be once again renovated after the passage of that crisis. However, even this strong system of fortifications could not withstand a disastrous attack by combined forces of Slavs and Avars in 596 AD, and it was then that the Justinianic limes was definitively destroyed.
The recent excavations of Roman military sites in the Iron Gate region, especially at Diana, have produced a wide range of significant archaeological finds. It is important to note is that the archaeological strata on these sites are accurately dated with numerous coin finds, so that the artifacts found in these so-called “closed layers” can be dated with certainty. Here, as on all archaeological sites, the most numerous type of artifact in the category of small finds is pottery. The pottery types are extremely diverse and of special interest for students of Roman ceramic production and distribution. In the first century AD, before the reign of Trajan, most of the pottery was imported into the region. The most frequently appearing types were the luxury wares from Italian workshops; Gallic terrasigillata was also frequently imported. These types of pottery are sometimes found on the same levels with Dacian cups, handmade-vessels in simple forms with a long tradition in prehistory. Beginning with the second century the situation changes somewhat. Imported pottery, from Italy, Pannonia and Germany, is still present, but in addition a large quantity of vessels for everyday use was produced in local military workshops. Two large well preserved kilns for producing this type of pottery were discovered adjacent to the castellum Diana, and similar finds appeared at the fortresses of Smyrna, Novae, and Taliata. In the fourth century the import of pottery was infrequent. Demand for luxury wares at that time was already completely satisfied by the products of local workshops. These are most often glazed vessels produced in a large variety of forms. Production of this relatively well made pottery lasted up to approximately the sixth decade of the fourth century. After that time there appears to be a significant decline in the quality of ceramic production. The presence of federated units of Goths is evident in the military bases in the products of their own material culture; also very clearly present is the typically poor quality pottery of late antiquity. It is evident that this was a period of crisis and reduced economic ability, when the needs of the military garrisons and local populations were satisfied by markedly simpler and much poorer products. The rest of the archaeological material show similar characteristics. Jewelry of gold, silver, and bronze, which was so abundantly represented from the first century to the end of Constantine’s and Constantius’ reigns, is replaced in the later period by of iron and very rarely of bronze. The large number of agricultural tools found in the military camps in the last decades of the fourth century, bears witness to the concerns of the garrisons for securing their own food supplies. The only luxury objects recovered from this period on the Iron Gate sites are beautifully designed and decorated Gothic worked in gilded silver.
Among the abundant coin finds mentioned above there were sestertii, denarii and aureii from the first and second centuries and numerous finds of silver coins (denarii and antoniniani) from the second and third centuries. Hoards of silver coins were also found, which is completely understandable when one considers that soldiers on the frontier in the third century were paid almost exclusively in silver coinage. Because of frequent wars, garrisons or detachments from garrisons were often transferred away from their bases which provided the occasion for some soldiers to conceal their savings in the ground. The abundance of bronze, especially small coins (nummi) was the result of dramatic inflation. This circumstance, however, considerably eases the job of the archaeologist in dating thin, often barely discernible layers which in the fourth century, directly thanks to that type of coinage, can be finely distinguished within a range of two to three decades. A very valuable find from that period is a gold medallion of the emperor Valentinian I (weight, 86 grams), minted in Constantinople most probably on the occasion of the beginning of his quinquennalia in 369 AD or of some later imperial victory over the barbarians. This is a unique object which is exceptionally well worked on its reverse field with a representation of the emperor in elegant parade uniform and conquered enemies in Germanic attire kneeling underneath him.
The unexpected quality of the bronze and stone sculpture recovered during the past ten years on the Danube limes was a pleasant surprise. Representative examples from that impressive group of objects include the following: a statuette of Priapus, one of the Greco-Roman gods of fertility, with a silver bracelet on his arm and meticulously defined face, hair and heard. A nude Venus with a broad, somewhat flat face and a diadem in the form of a mural crown (corona muralis) is certainly an eastern product, probably from a Syrian workshop, while a Jupiter with a lightening bolt in his hand represents the yield of some western provincial workshop. All of these statuettes were found in levels of the early decades of the third century at Diana. A bronze portrait bust, which according to its stylistic characteristics can be assigned to the transition from the first to the second century, probably represents a portrait of Trajan’s father. This was recovered right at Trajan’s bridge which suggests that it was part of a large statue (or statuary group?) which was located on the portal at the approach to the bridge from the Moesian side. An imperial portrait of rare beauty and expressiveness, found in the vicinity of Singidunum (modern Belgrade) can certainly be described as the best portrait of the emperor Macrinus (217-218 AD) recovered to date. Stylistic analysis and comparison with imperial portraits on coins from mints at Rome and at Antioch make it completely certain that this bust was produced in Antioch.
A marble bust of the usurper Albinus, found in the fortress at Diana, was unexpected on the Moesian limes because almost all of his activity was connected with Britain. This is a realistic piece, executed in the style of the Severan period, in which the actual appearance of the emperor is represented with a strong, almost brutal expression.
To conclude this paper, I would like to direct your attention to an important function of this section of the Danube which has not always been given sufficient emphasis. In addition to its traditional mission – that of maintaining the frontier and defending the empire – the limes from the end of the third until the mid-fifth century must have had significant strategic importance for the economic interests of the empire. It is now becoming increasingly clear that, immediately after the abandonment of Dacia and the rich Transylvanian mines there which produced for the empire an immense quantity of gold, silver, copper, and iron, the Romans moved their mining and metallurgical operations to the Moesian mines, the exploitation of which had begun already during the reign of Domitian. At the end of the third century and especially from the beginning of the reign of Diocletian, the extraction and processing of ores reached an unprecedented scale. Recent survey and excavation in eastern Serbia have revealed hundreds of mining sites and a large number of metallurgical centres in which ore was refined into metals. These complexes (e.g. Kraku’lu Yordan, Bukova Glava) were all fortified against both possible incursions of Goths and other tribes from the north and also the attacks of powerful robbery bands (latrones) which are frequently mentioned in the epigraphical sources from this region. The attraction for the robbery and barbarian groups was the rich plunder to be found in these centres. Slobodan Dušanić, a distinguished ancient historian who has devoted much time to the study of this period, concludes that the mining industry at Moesia was of greater importance for the economic life of the Roman empire than the ancient literary sources would lead us to believe. It can also be concluded that the economic crisis during the epoch of the “soldier emperors“ represented in a significant measure also a crisis in the mining industry, endangered by the general instability of the times and also by the exhaustion of certain large ore deposits in the Balkans and the Iberian peninsula, too. The demand for metals had priority over many other needs and led to important administrative restructuring. If the literary sources speak infrequently of or in fragmentary fashion about the mining industry, this is not to be taken as an indication that this economic activity was unimportant: involved here are men and factors which for the most part functioned beyond the focus of interest of the ancient writers but which, on the other hand, supported the infrastructure of the ancient economy. Research in the last two decades has elucidated in large measure two factors which are synchronous by and closely interrelated: the development and significance of the late antique limes and to the same extent both the importance and the influence of mining and metallurgy in the Danube hinterlands.