Nicolae Gudea Mihail Zahariade

DACIA RIPENSIS

Festungen an der Nordgrenze der

Provinz und ihre Truppenkörper

Amsterdam 2016

I. The northern border of the province of Dacia Ripensis (275-380 AD)

- p. 9
1. Introduction
The history of the territories of the former Dacian provinces in the period after Aurelianus has
always preoccupied historiography in general, and especially Romanian historiography, and will
continue to do so. The more important reasons for this are: the need to become acquainted with
the historical processes that took place after the Roman withdrawal (determination of the
economic, social and spiritual phenomena that took place here), the role and limits of relations
with the Roman Empire.
In general, this issue has been treated from an internal perspective of this former province, and
attempts have been made to use all of the written and especially archaeological information.
Although the conclusions reached are or appear to be satisfactory, they have not yet been
adequately argued (Macrea 1969, 464-488; Toropu 1976, 13-77; Tudor 1978, 416-454). This
situation led us to treat the subject from a different perspective and to see things from the
outside, etc. by studying the history of the neighboring contemporary late Roman provinces.

So we investigated the northern border of the province of Dacia Ripensis in order to draw
conclusions about the relations with the areas north of the river from its way of organization, its
strength and its economic system. In this first phase of the research, we turned our attention more
to the border of the province, which to some extent is still unknown. In a second stage we will
deal in detail with the location of the territories north of the Danube.
The purpose of the present study is to use a single late Roman province (DR) to record the
changes that occurred in the perception of the organization of defense and in the arrangement of
troops after the year 275 AD up to the end of the 4th century AD (Fig. 1). In the present work we
will only deal with the fortresses and the defense system they form.
1.a. The work equipment

1
The funds at my disposal were comparatively meager and of little historical value. The most
important written source remained the NotDign (cf. Jones 1964, 347-380; Hoffman 1970; Ward
1974). To the extent that the data it offers are also confirmed by others, especially archaeological
sources, its value remains intact. However, if these sources provide other information, then the
data from the NotDign must be treated with even greater caution than before. The Late Antique
Itineraria also contain some information because they mention localities from the

- p. 10
border and sometimes also their role in relation to them (ItAnt; Tab.Peut). Isolated data can also
be found in some later authors (Jones 1964; FHDR I-II). One should also mention the imperial
decrees, which refer to the organization of the border in general (Mommsen 1910; Grosse 1920),
but can also apply to an individual province. In general, however, the data that the literary
sources offer us are insufficient to understand the picture of the border.
There are still few epigraphic sources. There are two or three inscriptions, one of which relates to
a building (IDR II-III / 1; IGLR; IMS). The brick inscriptions are more numerous, but they
contain indirect data as they relate more to the troops (IDR II-III / 1; IGLR; IMS). Many of them
have not yet been read, while a considerable part is unpublished. The brick stamps have a special
contribution to the identification of fortresses and border sections.
Although the archaeological research provides the most extensive information, it has not yet
been fully evaluated. More or less extensive systematic archaeological excavations were carried
out at the forts of Hajdučka Vodenica, Kladovo-Karataš, Kladovo-D. Butorke, Orşova, Hinova,
Kula and Celei. Other forts have only been identified on the surface with their general layout.
There are also numerous fortresses for which only a few technical data are available and nothing
is known about the archaeological material. The archaeological material from the local fortresses
of the 4th century AD has hardly been researched at all, so that it is difficult to give a general
picture of the late material culture.

1. b. Current state of research

So the research is still in an initial phase. Of the large number of settlements that may be forts,
not a single one has been thoroughly explored. There are four attempts at monographic
representations, namely for Drobeta, Hinova, Kula and Celei. The excavation reports are quite
numerous, but they are very scarce, contain no technical data and no richly illustrated
archaeological material. Apart from the excavations of Sucidava, which were published in
successive reports, there are no series of reports. Sometimes even dating to the 4th century AD is
problematic. In 1951 H. Vetters published a first monographic sketch of the province of Dacia
Ripensis (Vetters 1951, 6-7), but without devoting himself particularly to the border. The essays
on the Limes from the great encyclopedias give very vague pictures of the limits and are
generally based on the NotDign (Forni Limes, 1267-1269). The most recent monograph of the
late Roman Empire deals little with the boundaries of the provinces (Jones 1964, 347-380).

- p. 11
Other special works that dealt with the military strategy from the late Roman period did not
delve into the system, because actually no section of the late Roman provincial border is known
in detail. Everything is judged according to the NotDign. E. N. Lutwak (Lutwak 1978, 127-190)

2
attempted a strategic pattern for the time of the tetrarchy - but on a very general level. We don't
share his judgments, but we will summarize them here:
- the re-fortification of the old border, but combined with elements of the mobile army; the
newly fortified border had three functions: it secured replacement battalions for the army, tactical
liaison for troops from within, and refuge for the mobile forces;
- The re-fortification of the old border meant not only the repair work on the old fortresses, but
also the creation of new bases; these new fortresses were not necessarily designed strategically,
but for tactical domination;
- the outer defensive elements of the fortresses (berms, ditches) became wider;
- the troop units were equipped with artillery; the types of troops were fundamentally changed;
- the real legions and auxiliary troops had disappeared; even if they kept the names, they
represented drastically reduced static forces; Even if the legions (or parts of them) have been
preserved, they played no different role among the frontier units;
- there were new border units (limitanei) that lived from agriculture; the military quality was
poor; the weakening of these forces continued after the time of the Tetrarchy through the transfer
of soldiers from the border area to the mobile forces from within.
These findings confirm the statements made by R. Grosse (Grosse 1920, 67) about the troops and
fortresses on the Danube. It also has particularly valid formulations: the late Roman fortresses
were built along the two river banks, but there are also fortresses that were advanced into the
barbaricum or withdrawn into the interior of the province.
The most complete map, which contains only about half of the fortresses, was drawn up by D.
Tudor (Tudor 1960, the map). V. Velkov made another - also incomplete - map (Velkov 1987,
164-166). Few of the fortresses known today can be found on such maps. NotDign OR XLII
contains for Dacia Ripensis 9 cunei equitum, 2 legiones, 2 cohortes, 6 auxilia, 1 milites; probably
20 fortresses with a total of 19,500 soldiers (Váradi 1961, 50). The difference between this data
and the data obtained today (see below) is very large.

1. c. Working and presentation methods

- p. 12
In order to obtain as much data as possible about the organization of the border of the province
of Dacia Ripensis, I'm proceed as follows:
- we examined the relief of the province and especially that of the border strip;
- we compiled a list of the certain or presumed fortresses, which I numbered from west to east.
Each fortress was systematically represented according to the following points: a. Name of the
village; b. Position of the fortress opposite the village; c. Relief form on which the fortress is
located; d. archaeological excavations; e. Technical specifications; f. occupied military units; G.
Literature;

3
- This list certainly still has a lot of uncertainties, but it represents a first systematic basis for
work. In the illustrations, the fortresses have the number from the list;
- we collected all information about the border of the province from sources or from modern
specialist literature;
- we do not have complete data and of the same historical value for all fortresses. From this point
of view our work is incomplete, etc. as a result of defects both historical and technical in nature
(plans, dimensions, construction system). It is possible that future research will change the
picture of the border presented here. Over time, especially over the past 25 years, the current
frogs have attributed several fortresses to several eras without making clear delineations between
them. We ourselves (Gudea 2001, the repertory) have some of the fortresses, or some of their
ruins, as for the 2nd-3rd Century and the research afterwards proved (!) That the fortresses are
from the 4th or 6th century (!!). Such ambiguities persist because of the lack of research.

2. The organization of the border

The total number of fortresses recorded in the area between the confluence of the Porečka and
Vit rivers with the Danube amounts to 45 (Fig. 4). Of these, 32 are certain and 13 are suspected.
NotDign mentions 23 forts, but only 10 of them have been "identified". Not mentioned in the
NotDign or other sources, but archaeologically documented 18 fortresses. Another 11 fortresses
are held on the basis of typical plans for late Roman fortresses.
2. a. Number of fortresses
The research on the northern border of the province of Moesia Superior showed that in the 2nd-
3rd Century AD on the section between the confluence of the rivers Porečka and Lom in the
Danube 8 Roman camps functioned (Gudea 1976; Gudea 2001) (Fig. 3); only three are safe
(Kladovo-Karataš, Kladovo-Kostol and Drobeta); three are likely (Brza Palanka, Prahovo and
Arčar); the rest are completely unknown.

- p. 13
Research on the northern border of the province of Moesia Inferior showed that at the end of the
2nd century AD 11 fortresses existed between the confluence of the rivers Lom and Vit and
Danube, which were abandoned after the conquest of Dacia (Zahariade-Gudea 1977; Gudea
2005). In the 2nd-3rd In the century AD, this section did not function as a Limes because it was
behind the Dacian provinces (Dacia Malvensis and Dacia Apulensis). (Fig. 3).
It must be emphasized that the Serbian historiography has many fortresses from the 2nd-3rd
centuries. or 4th century dated for the 6th century without any archaeological, stratigraphic
arguments or relevant archaeological material as evidence. Except for the archaeologists of
Roman times Vl. Kondić (Kondić 1972), P. Petrović (Petrović 1979), M. Vasić (Vasić-Kondić
1986) also has an archaeologist of the early Byzantine period Dj. Janković (Janković 1981) dated
numerous fortresses for the 6th century. We did not take these opinions into account. We have
dated all fortresses of the quadriburgium type to the age of tetrachia and to the 4th century based
on the inscription by Donje Butorke and the analogies as well as on the brick stamps discovered
there.

4
The northern border of the province of Dacia Ripensis was organized along the Danube.
Between the mouth of the Porečka and Cerna rivers, the Danube flows through the streams and
from Orşova to Prahovo through a mountainous area; its course is narrow, with numerous
meanders and frequent changes of direction. From Prahovo, the Danube then flows almost in a
straight line from west to east. Along the southern bank there is a narrow river plain as far as
Arčar, which from here becomes wider and wider and is 40-50 km wide at the mouth of the Vit
river (Vetters 1951, 5-7); north of the stream, the river plain is consistently wide.
The latest research carried out on the border of the province of Moesia Inferior (Gudea 2005) has
proven this situation.
The studies by S. Bobčev on the late Roman building systems (Bobčev 1952) show (Fig. 15-19)
that at the end of the 3rd century and in the 4th century the system of leveling rows of 1 to 4
bricks integrated between the rows of stone was practiced.

2 B. Density of fortresses
In the 4th century AD the number of fortresses grew considerably; Depending on the terrain and
the route of the river, they are more dense or scattered (Carrié 1979, 839: “en effet le dispositif
tétrarchique visait à l'occuper tout le territoire de la province, sans tenir compte de la valeur
strategique réele des sites”); in the actual streams there are two (M. Golubinje, H. Vodenica),
from Orşova to Prahovo 23 fortresses; the greatest density is on the Orşova-Drobeta section,
where there are 9 fortresses; from Prahovo to the east, so there are few on the plain and the
existing ones are far apart.

- p. 14
The NotDign reflects a different situation: 9 fortresses called castella are mentioned here and
another 14 that function as headquarters of the troops; they are not listed in any particular order,
but rather according to the categories of troops; only Drobeta, Orşova, Tekija, Brza Palanka,
Vidin, Hărletz and Gigen have been archaeologically identified.
Despite its general title, the work of S. Johnson (Johnson 1983) stops analyzing the fortresses at
Pannonia II, so it does not include Dacia Ripensis. The work by J. Lander (Lander 1984)
contains more data - but without an analysis of the boundary. In chapter IB: Fortification most
readily dated to the period of Diocletian and the Tetrarchy (Lander 1984, 181-193) he mentions
after listing the ancient information about the building activities of Diocletianus (Eumenius,
Lactantius, Ammianus Marcellinus, Zosimus, Malalas, Procopius ) - which is why he was also
called innovator, recycler, restorer - and of the contemporary inscriptions only the fortresses of
Donje Butorke (e.g. in Dacia Ripensis not Moesia Superior, as the author writes) and Castra
Martis / Kula (Lander 1984, 185 , Fig. 172; 188, fig. 180).
In Chapter IV / V (Lander 1984, 206), in which he talks about the square corner towers, he
casually mentions the fortresses of Gornea and Ravna (Moesia I: Fig. 207) and of Porečka Reka,
Orşova and Puţinei (Dacia Ripensis, Fig. 207-209). For the forts with round towers he mentions
Hajdučka Vodenica and Kula (Lander 1984, 240, fig. 256-257) as well as those from Moesia I:
Saldum, Bosman, (Lander 1984, 240 fig. 254-255). For the horseshoe-shaped towers he only
refers to Drobeta (Lander 1984, 252 fig. 274).

5
The buildings from the time of Valentinianus are mentioned in general (after Ammianus: Lander
1984, 270). The author (J. Lander) did not record the decree of the emperor to Tautomedus dux
Dacia Ripensis.
Since the area of the Lower Danube belonged to Licinius, Constantinus could not start the
fortification program earlier than 324.
Procopius also adds (De Aed. IV 5. 1-3) that: "because he thought that it was not useless that the
river should be guarded on both sides." The remark of the Byzantine author is perfect in my
opinion valid.

2. c. Arrangement of the fortresses

In the 2nd-3rd In the 17th century AD, the Roman camps were located near crossing points,
fords or bridges. The concern for such places remained in the later age. The fortresses were built
everywhere, not only where there are easier transition options, but to control all transition
options. They worked as a linear defense system. Most of the fortresses are located on the right
bank of the Danube, the majority directly on the bank, especially at the estuaries. Some are to the
south, sometimes even at a considerable distance from the stream (Kula, Vidrovac). Ten
fortresses are on the left bank; eight of them are bridgeheads in the sense that they are opposite a

- p. 15
fortress on the right bank of the Danube, and two / three are away from the river (Mehadia ?,
Puţinei, Bistreţu) along roads. There are two fortresses at the ends of the bridges (Drobeta,
Celei). Incidentally, the establishment of advanced Roman fortresses in enemy territory was also
common in the late Roman period. Amm. Marc. mentioned these acts twice during the reign of
Valentinianus: (February 28, 1) “et utrique Rhenus celsioribus castris munivit”; (29. 6. 2) “in
ipsis terris quadorum, quasi romano iure vindicatis, aedificari castra praesidiaria mandavit”.
For Dacia Ripensis, the small, contra-named fortresses opposite the cities or fortresses fortified
on the right bank are missing. But in Pannonia-Valeria and Moesia Prima, the NotDign also
shows many such fortresses (Contra Acinco, Contra Bononia, Contra Margum, etc.). Procopius
(De Aed. IV 5. 3) claims that “the former Roman emperors built these fortresses and castella not
only on the right bank and not with the purpose of making them invincible, but rather around the
banks of the river to leave defenseless ”.

2. d. The streets and their relationship to the fortresses

The main road along the river was retained. The journeys of Diocletianus in the years 294 and
303, of Constantinus in the years 328 and 334 prove its use (Mommsen 1910, 195; Enßlin 1948,
2434, 2439-2440; Danoff 1937, 1035-1036). The following roads branch off from the main road
to the south: along the Porečka valley to Garičin Grad; in the Timoctal to Timacum Minus; in the
Lom valley to Niš through the St. Nicola bottleneck; in the Ogost, Isker and Vittal to Serdica.
This road was repaired under Constantinus and Valentinianus (CIL III 12521; Morfova 1963, 33-
34). An inner road, which ran parallel to the river, connected the cities that lay at the foot of the
Balkan Mountains (Timacum Minus - Asamus - Conbustica - Montana); it was repaired under
Constantine (CIL III 12520). Roads branch off from the main road to the north that continue on
the other side of the Danube: in the Cernatal to Tibiscus, in the Pleşovatal to the Vulcan
bottleneck; up along the old valley; repaired under Constantine (IGLR 278).

6
It is certain that bridges worked in Drobeta and Celei. The former is older; the second was built
under Constantine (Tudor 1974, 129-133; 135-160). It is believed that this bridge was the work
of the architect Theophylus Patricius. The Drobeta bridge was probably destroyed. In any case, it
was mentioned as destroyed by Procopius (De Aed. IV 6. 14). The bridge at Celei was almost
certainly destroyed by the great floods of 376 or 378 (cf. Amm. Marc.).
Celei was chosen for the location of the bridge because in this way it had a direct connection to
the road south of the Danube, which connected Oescus / Gigen with Serdica / Sofia along the

- p. 16
Iskăr river; at the same time it joined the former Roman road north of the river, which led from
Celei to the north (into the former Dacia) and was rebuilt by Constantinus. The structure or the
bridge (Tudor 1971, 161-162) was 2,437 m long; it ended in 328. This bridge was mentioned by
Sextus Aurelius Victor (De Caes. 17. 41) after the middle of the 4th century. Chronicon Pasquale
(cf. MGH, Chron. Min. I, 233) claims that it was a stone bridge; this statement was also repeated
by the later Byzantine authors (Tudor 1971: Theophanes Confessor and Georgius Cedrenus).
The archaeological excavations carried out by Cezar Bolliac around 1869-1870 identified the
stone bridge piers from the Romanian bank. The research of Gr. Tocilescu from 1888 (Tudor
1971, 166-167) have confirmed the presence of the bridge piers. In 1933 D. Tudor carried out
research with the help of divers who identified three pillars of the stone bridge (Tudor 1971, 169-
171). The later research of O. Toropu provided new elements about the shape, the working
technique and the dimensions of the northern bridgehead.
The Celei Bridge was depicted on a gold medallion minted in Constantinopolis. On the basis of
this, O. Toropu graphically restored the northern bridgehead with the portal (Tudor 1978, 420
Fig. 126).

2. e. The ancient names of the border section and fortresses

The ancient names of the border are difficult to determine. When Lactantius (De Mort., 17, 4)
spoke of Diocletian's travels, he wrote: “per circiutum ripae Istricae”; Anon. Val. (V 21) “Gothi
per neclectos limites eruperunt”; Eumenes (Pan. Lat XVIII 4. 18) “limite restituta”; the term
“limes” was also used in the early Aurelian period (SHA XXVI 10.2); CIL III 13734 = ILS 8938
“ad confirmandam limitis tutelam”; the brick stamps read “ripa”: ripa Var(iniae), ripa
Val(erianae), ripa Oes(co); the most frequently used name is ripa.
There is even less data on the fortresses. Aur. Victor (De Caes. 41.18) calls them "castra
castellaque"; Amm. Marc. used several names for the fortresses: castra (XXX 7. 6; XXVIII 2.
1), castra praesidiaria (XXVIII 3. 7; XXIX 6. 2), castella (XXVIII 2. 1; XXX 7. 6), turres
(XXVIII 2. 1). Vegetius (IV.10) writes castellum parvulum quam burgum vocant; Orosius (Adv.
Pag) uses the same term (VI.32, 12); Amm. Marc. (XXVII 4. 6) writes "castris ... et castellis".
Only the name praesidium is mentioned in inscriptions (CIL III 6151 = ILS 641; Russu 1933-
1935, 210-211); The NotDign uses the name castellum, but only for the settlements where
Roman camps used to be (Drobeta, Bononia, Dortico, Cebro, Augustae, Almo); on the
inscriptions by D. Butorke it says praesidium, meaning a quadriburgium; Procopius (De Aed. IV
5. 4) claims that

7
- p. 17
most of the fortresses consist of a tower; the order to Tautomedus (CTh, XV 1. 13) of 365 only
mentions turres, a term whose meaning we do not know. At the time of the Gothic invasions,
laws were passed that mention repairs (CTh, XI 16. 18), but do not mention the names of the
fortresses. The name of some fortresses is confirmed by late Roman brick stamps or based on the
name inherited from the earlier period (2nd-3rd centuries AD). We have stamps with the place
names Karataš / Diana, Tekija / Transdierna, Orşova / Dierna, TurnuSeverin / Drobeta, Prahovo /
Aquis, Vidin / Bononia, Cibăr / Cebrus, Hărletz / Augustae, Lom / Almus, Gigen / Oescus, Arčar
/ Ratiaria, Vit / Utus.
Numerous contemporary authors have tried to find the ancient names of some fortresses that
were archaeologically discovered based on the data provided by ItAnt or NotDign. So Vl. Kondić
- among other things - the fortress of Insula Banului Transdiana, that of Donje Butorke Zanes (cf.
Petrović 1980, 549) and that of Karataš, first Caput Bovis and then Diana. All of these and other
attempts are very likely, but completely uncertain in the absence of some epigraphic evidence.

3. Types of fortress and related technical data

There are several types of fortresses on the border of the province: rebuilt old Roman camps,
new type fortresses (quadriburgia), fortified cities, individual towers, castella (NotDign) and
ports.

3. a. Rebuilt Roman camps

Only three of the rebuilt old Roman camps are known: Drobeta, Kladovo-Kostol, Kladovo-
Karataš; perhaps one could also add Hărletz and Gigen, but there is still very little data. The
general plan was retained, the ramparts were re-used through reconstruction; the architecture of
the interior buildings changed; Trapezoidal towers appeared at the corners; the gates were
blocked or replaced with protruding square gate towers. In Hărletz and Gigen there are round
outer towers at the corners and on the sides. There is no definite information about the date. It is
said that Drobeta and Gigen date from the time of Constantine. The reconstruction works of
Karataš are dated to the end of the 3rd century AD. In the old fortresses the NotDign mentions
troops: in Vrav - Brza Palanca - Hărletz cunei and in Arčar legio.
The dating for the age of Constantine is based on the horseshoe-shaped corner towers. The claim
from the Excerpta Valensiana (27), according to which Constantinus restored the “castles” from
the right bank, is only partially confirmed. On the other hand, Zos claims. Hist. Nov (II 34) that
he had "destroyed the Limes" with the withdrawal of the troops from the border.

- p 18
Unfortunately, the internal organization of these fortresses is unknown. For the Drobeta fort we
have two proposals for a floor plan: one by D. Tudor with the restored old barracks (Tudor 1978,
436 fig. 137) and the other by O. Toropu (Toropu 1976, 21 fig. 4) with the organization of the
interior in the shape of a cross, with the soldiers' dwellings along the beams of the cross. In the
case of the two “new” fortresses of Vidin and Kula, the structure cannot be determined.

3. b. New type of fortresses

8
The new fortresses are all of the quadriburgium type (Fig. 6-9) and have the following main
features (Lander 1979, 1051-1060): - small dimensions: on average 40 x 40 m; - square plan,
mostly regular and approaching the square; - protruding square (Orşova, Sip, Ostrovul Banului,
Puţinei, Hinova) or round corner towers (Malo Golubinje, Hajdučka Vodenica, Donje Butorke,
Slatinska, Vidrovac, Miliutinovac, Vidin, Kula); - the tower entrances are transverse (Hajdučka
Vodenica, Donje Butorke, Vidrovac, Kula, Orşova) or parallel to the wall (Hinova); - sometimes
there are protruding intermediate towers; - very high walls; - the interior is uninhabited;
sometimes there is a small square building in the center (Donje Butorke, Hajdučka Vodenica,
Rtkovo, Ljubičevać, etc.); in Kula the principia is next to the wall; - you lived along the
ramparts; Usually the apartments had clay walls (Carrié 1979, 826); in Celei there are still the
substructures of the walls that supported the roof of these apartments; - the building system of
walls with rows of leveling bricks. The external defensive elements (ditches) are missing (cf.
Anthes 1917, 157); only in Hinova were three defensive trenches dug to the north.
Little can be said about the entrances to the fortresses (Fig. 10). None of the new typical
quadriburgia has an actual gate. In Hajdučka Vodenica (north wall), Orşova (south wall),
Ostrovul Banului (north wall), Kostolac (west wall) we have simple breaks in the wall. In
Vidrovac there is a break with a side wall. In Kula and Kostolac there is a gate but in a
completely new form. When entering the central building, the surrounding wall was doubled in
both cases and an entrance with two “towers” was built in. In the former, newly used auxiliary
camps, the side gates were mostly blocked with large protruding square towers (Karataš,
Drobeta). In Kostol, the porta praetoria was blocked in this way. In Drobeta and Karataš, the
south gate (praetoria) was rebuilt with horseshoe-shaped projecting towers. In Kladovo Fetislan,
entry was made possible through an intermediate tower on the north side. All of this indicates
that there has been less movement of troops; no real gates were needed, only a few small
entrances. This is another indication of the small number of late Roman units.

- p. 19
The construction and / or. The repair program was extensive. Apart from stone - as already
mentioned - brick was used. With their help, the leveling rows were made in the walls. They
were also used in large quantities at the tower entrances. The brick production must have been
very large in some centers so that the products could also be exported. Bricks with the stamp
DRP DIERNA were used in the fortresses of Golubac / Cupae, Ribnički Potoc, Gornea, Donji
Milanovac in Moesia Prima, and in those of Malo Golubinje, Dubova - Veterani cave, Hajdučka
Vodenica, Tekija, Dierna, Drobeta Aquis / Prahovo, Arčar / Ratiaria etc. found in Dacia
Ripensis. It is noteworthy that in some of the centers mentioned above, stamped bricks were
made, on which the name of the fortress appears. This indicates that construction activity is
generally continuing.
We know very little about architecture. The well-preserved tower of Kula indicates very high
ramparts and towers. On a corner tower, three floors were preserved at a height of 16 m, which
were separated from each other by wooden platforms. Next to the tower, the wall was 12 m high.
Most of the fortresses of the newer type have been dated to the time of the Tetrarchy, to the end
of the 3rd century AD (Gudea 1972, 143; Lander 1974, 1050). However, the written sources
show a longer construction period: v. Gall. (SHA XXIII 13.6) v. Aurel. (SHA XXVI 10. 2:
limitem restitueret); for the tetrarchic period: Eumenius (Pan. Lat. IV 18 (297), Eunapius (FHG

9
IV. 14); for the Constantinian period: Zos. Hist. Nov. (II 34), Vegetius (II 11), Aurelius Victor (
41, 1. 18) “pons per Danubius ductus castra castellaque pluribus locis commode posita.”
Lactantius (De mort. Pers. VII 8-9) accuses Diocletianus of the unlimited passion for buildings
and mentions the concern for the provinces He says they built basilicas, circuses, workshops for
weapons, etc. Diocletianus visited the area three times (Moesia I, Dacia Ripensis, Moesia II,
Scythia Minor) and knew the realities personally (Ensslin 1948, 2438-2440, 2447-2448). One of
the visits, the one from the year 294, began in Singidunum and on the way Cebrus, Ratiaria and
Varinia are mentioned (10.10 - 13.10.294); the journey was continued afterwards to Moesia II as
far as Durostorum. The buildings are mentioned in the speech of Eumenes ( Pan.Lat. IV 18); he
says: “nam quid ego alarum et cohortium castra percenseam t oto Rheni et Histri et Euphratis
limite restituta ”. The facts are - as already mentioned - by Zos. Hist. Nov. (II 34) and Eunapius
(FHG IV 14. 4) confirmed.
The mints of all mints that have fortresses on their lapel (RIC VI) must also be mentioned here.
The dating is based on the floor plan, building system and archaeological material. It is generally
recognized, however, that the fortresses of the quadriburgium type - regardless of whether they

- p. 20
had square or round, protruding corner towers - were predominantly built under the Tetrarchy
(Berchem 1952, 12 Fig. 14) or under Constantinus (Lander 1984, 252, 256 -257). It is believed
that the fan-shaped corner towers - the variants called "large fan" and "small fan tower" - could
also be dated for this period. The latest architectural analyzes by J. Lander show that the
horseshoe-shaped and round intermediate towers with counter pillars on the perimeter or on the
gates can also be dated to the time of the Tetrarchy.
Our panels, which show the corner towers of the fortresses in Dacia Ripensis (Fig. 11 - 12), show
great differences within a so-called “type” (square, round, fan-shaped). There are square corner
towers with different positions opposite the rampart wall: fully protruding, half protruding,
oblique. This also applies to the round corner towers. In Drobeta and Kladovo-Karataš, the fan-
shaped corner towers were added to the ramparts. In Celei there is the same type of corner tower
(Fig. 13); but it is half protruding. At the moment we do not know how the simultaneous
presence of towers of different floor plans (square, round, horseshoe-shaped, fan-shaped) should
be explained. Such “combinations” exist in Hărletz and Gigen, but they are not definitely dated.
From the point of view of the ground plan and the corner or central wall towers, the fortress of
Celei / Sucidava is very similar to that of Pannonia I (Soproni Limes Nitra 133); irregular floor
plan, horseshoe-shaped towers at the corners, square intermediate towers on the sides and living
next to the walls.
However, it could be that some of the drawings of towers that we used in the work actually
accurately depict the floor plan, the location and other technical details. Therefore, only future
research will solve this question.
Here we repeat the very important opinions of J. Lander about the dating of the tower floor plans
(Lander 1984, 255): “Twelve different categories and subdivisions of tower shape have been
distinguished, and if for the moment we divide the imperial frontiers into four sectors: Britain
and the Rhine, the Danube, the Egypt and the East and the North Africa, we discover that only

10
one of the twelve tower shape is found in all four sectors: not surprisingly, it is the traditional
square tower though now, of course, projecting outward to a varying degree. Four tower shapes
can be found simultaneously easily in only two sectors, and five tower shapes are not known
outside a particular sector. The clustering appears to be more particular still: the majority of
polygonal towers are found in a single province, Britain. Lower Moesia has a majority of the
“large fan” towers and nearly a majority of all U-shaped towers. All “simple rounded” examples
come from Pannonia; all "external circular" towers come from Middle Danube; most "splayed
fan" towers from Pannonia and Upper Moesia ". "Apparently all of the late sites in North Africa

- p. 21
had square towns only and were located in two clusters of apparently quite different dates". From
J. Lander's sentences it can be seen that the tower types were not uniformly distributed, perhaps
this was closely related to the local or regional needs of the fortification activities. One should
not forget, however, that J. Lander did not record everything correctly; Incidentally, he did not
delimit the provinces in the sense of the late Roman administration (Lower Moesia, Upper
Moesia etc. are very general formulations, they do not properly cover certain late Roman
provinces). Only Sucidava was built around AD 328. It can be assumed that the sons of
Constantine also continued the fortification work. At least that is what the inscription from
Troesmis (Scythia Minor) suggests (CIL III 12483).
But Julianus' building activities are safer. The trip from Lutetia (Paris) to Constatinopol served as
the occasion for the crossing of Dacia Ripensis. Amm. Marc. (XXII 7. 7) writes: Hister, qua
romanam cespitem lambit urbibus multis et castris contiquus et castellis “.
A construction and repair phase should have started as a result of the order from Valentinianus to
Tautomedus dux Daciae Ripensis (CTh. XV 1. 13: “in limite gravitati tuae commisso praeter eas
turres, quas refici oportet, si forte indigeant refectione, turres administrationis tempore
quotannis locis opportunis extrue ”), which, however, has not been identified archaeologically.
The repairs, which Amm.Marc. describes, (XXVI 7. 7) "nec tamen cum corrigendis civilibus ita
diligenter in et omisit castrensia rectores militibus diu explorates adponens urbes quin etiam per
Thraciam omnes cum munimentis reparens extimis curansque soliciti ne arma vel indumenta, aut
stipendium vel alosimenta de supercilia Histri dispersos, excursibusque barbarorum opposite
aggere vigilante audiebat et fortiter ”. For the activities of Valentinianus and Valens for the
defense of the borders see Vegetius (III 13; IV 10), Themistius (Or. X 136a-136b, 136d), Amm.
Marc. (XXX 7. 6) and CTh (XV 1. 13). It is believed that Valentinianus and Valens wanted to
give up the offensive policy of Constantinus and his successors. The fortification of the borders
seemed (Valentinianus) to be more successful than the war. Amm. Marc. (XXIX 4.1) expresses
these political wishes very clearly: “quod maius pretium opera foret in coercendis verius limite
barbaris quam Pellendis”. A. H. M. Jones (Jones 1964, 462) is of the opinion that, according to
Valentinianus, responsibility for the buildings on the borders was transferred from the praesides
to the duces; this took place as early as 349. He gives numerous examples in this regard. V.
Velkov mentions some remarks by Themistius, who accompanied Valens on a tour along the
Danube; he describes in detail how a fortification (a tower) is built and gives that of
Commercium (Pannonia II) as an example, which was built in 371 within 48 days. J. Lander

- p. 22

11
 suspects that the so-called “splayed fan towers”, the protruding fan-shaped corner towers, can be
dated to the Valentine period (Lander 1984, 252). The coins found do not help to date the
buildings, but help to determine their ends. The last coins are from: Valentinianus II (Orşova,
Hajdučka), Teodosius II (Hinova), Constantius II (Puţinei), Honorius (Arčar). Apparently all of
them were destroyed towards the end of the 4th century AD (between 380-390 AD) or the
beginning of the 5th century AD (kula). For some (Hajdučka, Orşova, Hinova) the destruction is
evidenced by a layer of ash. For details on the destruction of most of this fortress, see Chapter
III.

3. c. Fortified cities
The data on the fortified cities are missing. In Vidin and Kula (Figs. 104, 112) the new fortresses
appear as parts of large late walls that probably encompassed the entire settlement. In Kula, the
large rampart wall was also built with rows of leveling bricks. The building system with leveling
rows made of bricks plays or could play an important role in dating the fortresses. It is believed
that this system generalized at the end of the 3rd century and especially in the 4th century
(Bobčev 1952). The number of bricks remains low: 1 (Hinova, Prahovo), 3 (Orşova, Tekija,
Kula, Oescus), 4 (Tekija, Oescus) and 5 (Malo Golubinje, Hajdučka Vodenica, Orşova, Kladovo-
Karataš, Bordej, Celei). From the examples of this kind (Fig. 15-19), some remarks emerge: a.
the rows of bricks are small, the distance between the leveling rows is quite large: (1.25 - 1.00 -
1.67 - 1.80) Kula (Fig. 18.30.33), Celei (1.35 - 0.70 - 0.70) (Fig. 19. 41), Gigen (0.50 - 1.35 -
0.70) (Fig. 19. 40), Tekija (1.00 - 0.60) (Fig. 15. 3 ), Vidin (0.30-0.30) (Fig. 18.30); the series of
levels are very seldom repeated. In Kula - the only very well-preserved fortress (one of the
corner towers is 16 m high) - the rows of leveling are repeated 7-8 times. All these indications
regarding the building characteristics point to an earlier dating of the fortresses (end of the 3rd-
beginning of the 4th century). In Gigen, the part of the city that has been incorporated into the
new wall has retained its old organization (Fig. 19.40). Hierocles (Synecd. 655. 1-6) mentions
the following cities: Ratiaria, Aquae, Bononia, Castra Martis, Oescus.
There are still too few plans of the fortified cities to be a basis for discussion. We only know
proper late Roman city walls in Vidin and Kula. It is believed, however, that the plans of Hărletz,
Gigen and Sucidava represent fortified cities. In the case of Hărletz and Gigen, the walls look
more like a city fortification than a fortress. The plans by Vidin, Hărletz and Gigen are regular
squares. The wall is about 3 m thick, with round, protruding towers at the corners; on the sides
there are round, semicircular and horseshoe-shaped towers (Hărletz, Gigen). The plans of Kula
and Celei are irregular squares. In Kula there are round towers protruding on the corners and
sides. In Celei there are semi-protruding towers at the corners and square and semicircular

- p. 23
intermediate towers on the sides. It is noteworthy that in Vidin and Kula the small fortresses
(quadriburgia) functioned as corner towers. The surface of the fortified cities looks very large:
Vidin 135 hectares, Kula 70 hectares, Celei 180 hectares, Oescus 10 hectares, Hărletz 105
hectares.

3. d. Individual towers

12
The individual watchtowers and observation towers - apparently - played a major role in defense.
One could assume that the small towers of Mihailovac-Blato, Mihailovac-Moara Veche, Bordej
and probably also those of Pečka Bara, Dubova, with square floor plans, small dimensions
(between 7-15 m) and inside with a small building, which has four systematically arranged bases,
for which the time of Valentinianus can be dated. There is currently no archaeological evidence
for this. But the similarity with the small fortresses from Pannonia (Zeiselmauer, Pilismarot-
Malompatak, Budakalász, Gomboshatvánpustza) (Soproni 1978 No. 17, 30, 44, 72) or from other
provinces (Lander 1984, 289-293) can now be used for this dating serve as evidence. Likewise,
the very small distance between the two towers of Mihailovac indicates the possible density of
these small fortresses.
It is assumed that the individual towers played a major role in the defense policy of
Valentinianus and Valens (cf. the order to Tautomedus - already mentioned above). Amm. Marc.
mentions the towers twice: (XXVIII 2. 1) “... castra extolens altius et castella, turres assiduas
per habiles locos et opportunos”; (XXX 7. 6) “castris munivit et castellis ne latereusquam hostes
ad nostra se pro ripens posuit”. This multiplication of small fortresses shows the embodiment of
a sentence by Amm. Marc: (XXVIII 3. 7) “instaurabant urbes et praesidiaria castra (ut diximus)
limitique vigilis tuebatur et praetenturis”. By analogy with buildings from Pannonia I and II, the
towers of Mihailovac / Blato and Mihailovac / Moara Veche could be dated for the same age.
Possibly also the square buildings from the interior of some quadriburgia.

3. e. Ports
In Hajdučka Vodenica, Kladovo-Karataš, Drobeta, Korbovo (?) And Celei, the fortresses have
two walls extended against the river. Some of these double walls end with towers (Hajdučka
Vodenica, Kladovo-Karataš, Drobeta). These two walls enclose a certain space that was probably
used as a harbor (Bosković 1978, 432 note 18). The situation is very similar to the ship lands
from Germania and Pannonia.

3. f. The fortresses from the Notitia Dignitatum

With regard to the fortresses (castella) from NotDign, the technical data is missing. Only
Drobeta, Karataš and Celei have been identified. We would be inclined to believe that Drobeta is
a model of a castellum, but the dating does not correspond to the NotDign and is also not certain.

- p. 24
The fact that the list of fortresses from the NotDign begins with Transdierna and the two more
western quadriburgia (M. Golubinje, H. Vodenica) as well as most of the forts of the newer type
are missing, could be an argument for the dating of the NotDign after the year 390 . Likewise,
the NotDign only knows three forts on the left bank, while several have been discovered
archaeologically and are all quadriburgia. The list from NotDign has the following
characteristics: few fortresses; no particular arrangement; the troops mentioned are of a newer
type and differ from those covered by the stamped bricks.

4. Relations between the border and the areas north of the Danube

13
The way in which the border was organized and the troops were deployed at the moment allow a
conclusion to be drawn about a reconquest of the areas north of the Danube, or rather, do not
provide sufficient, but nonetheless certain, evidence in this regard.

4. a. the wall "Brazda lui Novac"

Brazda lui Novac (Troianul) is a huge wall; it begins somewhere in the Drobeta area (it was
assumed that Hinova was the starting point), reaches the Alt near the village of Ostrov and
continues in an easterly direction to near the city of Buzău (Tudor 1978, 244-249). It was better
preserved just before the border of the province of Dacia Ripensis. Its average width at the
bottom is 20-30 m, its height between 1.00-2.00 m; on the northern side, the trench is 10 m wide
and 1.50 m deep. It has not been explored by excavations, and no reliably dated materials have
been found that are connected with it. It is believed to be a late construction because: it was not
organized as a Limes; because somewhere in the east it pierces the transalutanus wall built in the
2nd or 3rd century (Tudor 1978, 249); because it is not accompanied by roads and fortresses.
Apart from the fact that it begins next to Hivnova and that the road to the north along the old part
of the river ends near it, there are no other connections to the border on the Danube.

4. b. the area between the Danube and "Brazda lui Novac"

The area between the Danube and the Brazda-lui-Novac-Wall was recently investigated for the
period between the years 270-450 AD (Toropu 1976, 13-35; 39-65). It is assumed that it was
reoccupied after the year 322 AD (the war of Constantine against the Goths) and that the area
north of the Brazda-lui-Novac-Wall came under Roman control (Toropu 1976, 28¸Tudor 1968,
50; Horedt, Limes Mamaia 212-213 note 10; Vulpe, Limes Mamaia 267-270). The main
evidence in support of this hypothesis was the miliarium of Constantinus of Celei (IGLR 272),

- p. 25
which assumed the existence of a road along the old town to the north. The evidence is not yet
sufficient, although this possibility is emerging: Fig. 31 - of the 78 localities (from all of Oltenia)
with late discoveries, only 6 are certain (+ Drobeta, Celei); the remaining 70 are only assumed
(Toropu 1976). All 6 are on the banks of the Danube and belonged to the border strip; - of the 78
localities, 43 are between Brazda-lui-Novac and the Danube; - Late Roman products: fibulae
(Toropu 1976, 59-60), pendants (Toropu 1976, 59-60), glazed ceramics (Toropu 1976, 51-53)
are very few and these too are unevenly distributed; - 15,825 coins (Toropu 1976, 65) were
discovered in 58 localities (Drobeta, Desa, Celei), 10,000 of them in Celei (individual finds and
in 8 hoards); the others were mainly discovered in localities on the banks of the Danube; - after
the year 380 the coin circulation decreased by 70% compared to the previous period (Toropu
1976, 18-35).
The late Roman coin circulation was weaker in the area between the Danube border and Brazda
lui Novac (Novaks Graben) very north of Brazda lui Novac and almost absent to the east of Alt.
But it is also certain - according to today's archaeological evidence - that the barbarians did not
cross the area west of the Alt until the end of the 4th century (380-390). The finds that represent
the Sarmatian, including the late Sarmatian tribes (the Alans) (Bichir 1977, 167-197), the finds
of the Sîntana-Cherniakov culture, which represents the Goths (Diaconu 1964, 211-232; Mitrea,
Preda 1966 ) (Fig. 33), keep east of the Alt. All these archaeological finds are now evidence that

14
the territory north of the border of Dacia Ripensis functioned as the foreland of the border. Some
researchers regard this situation as a real occupation (Tudor, Schmidt, Toropu). E. Démouget
writes (Démougeot 1982, 91, 97, 101): “L´ancienne Dacie resta donc exterieure à la culture
Tscherniakow-Sîntana qui l´entourait tant au nord quà l´est, elle apparait même comme une môle
de resistance sur la carte ou K. Horedt a reduit, cependant, l'aire geographique de cette culture au
territoire de la Roumanie. ”He also adds:“ Il est certain also that Constantin réoccupa et controla
outre Danube une bande des exprovinces de Dacie Inferieur et de Dacie Malvensis ”(see Moesia
Prima). About the politics of Constantinus he says: "n´aboutit sûrement pas une rèconquète des
trois exprovinces daces, mais il y eut incontestablement, comme l´attestent quelques inscriptions
et monnaies, un Dacie constantinienne d´outre Danube, face à la Dacie Ripuaire” . Thompson
(1956, 371) also explains on the basis of the text by J. Lydus (De Mag. II 10; III 31. 40) that the
area north of Dacia Ripensis between 337-35) was recaptured.
The border construction policy and the border organization together with some written sources
(Pan.Lat. IV 8; III 3: "Dacia restituta" etc.) show an earlier activity in this direction. We tend to

- p. 26
think that the retention or reconquest of southern Dakia was still under Aurelian and certainly
under the tetrarchical emperors. Although decisive evidence is not yet available, it can be
claimed that the area was under the influence of the Romans or even under Roman rule; that life
of the Roman type continued in the old provincial economic, social and spiritual forms; that
Christianity was making great strides. In any case, the situation was different from that of the
barbarian areas.

5. Conclusions
- the river formed the northern border of the province; most of the fortresses were built during
the Tetrarchy; the Tetrarchs found a solution for defense for quite a long period of time;
Constantinus relied on these fortresses when he went on the attack; the area north of the Danube
may have played the role of an attack interception zone, which was strictly controlled by the
Romans; a direct connection between the border and the Brazda-lui-Novac-Wall cannot be
established at the moment;
- in the first phase, the frontier represented a power in terms of both the number of fortresses and
the number of troops; in the second stage their power and attack power diminished; the fortresses
were transformed from attack bases into guard points. The border itself became a guard line;
- the border is no longer a fortress line, but has become a network of fortresses that were laid out
in all important points (many of them were inherited fords, bridges, harbors, crossroads,
checkpoints, but not only). The large settlements and cities south and north of the river were also
incorporated into the system (Fig. 31);
- the archaeological discoveries north of the river up to Brazda lui Novac prove the constant links
with the empire, a further development of the old provincial civilization, an economic and social
dependence of the area on the empire (Christianity), which is not enough to ensure the existence
of a Dacia to prove restituta (Pan. Lat. IV (8) III 3; Iulianus (Caes. 24), which, however, remains
a reality that has to be proven and its classification in the border strip becomes a certainty.

15
- p. 43
II. The troops on the northern border of the province of Dacia Ripensis (275-380 p. Chr.)
II.1. introduction
We intend to study the troops from the northern border of the province of Dacia Ripensis more
closely to find out what the forms of organizing the defense of a late Roman province in a
precisely defined period of time and to what extent the military reforms from the 4th century to
the older were tactical organization and the structure of the army. Such an investigation will
make it possible to assess the extent to which the limes-type defense system was retained in the
late Roman epoch. From the beginning we would like to note that, with the exception of the
treatise by D. Tudor (Tudor 1960), our work is the first to deal with this topic in detail. As a
result, the results achieved are just a beginning that needs to be deepened in the future. In my
work I will try to depict the military units in close connection with the fortresses.

II.1.a. Literary, epigraphic and archaeological sources

There are few literary sources that contain data on the troops of the Dacia Ripensis Province. M.
Zahariade (Zahariade 2006, 167-168) assumes that the reform of the army from the provinces of
Moesia I and Dacia Ripensis took place around the years 315/316; therefore the military
organization does not correspond to that of Scythia Minor, because the reforms were introduced
there later. Actually they are limited to the NotDign (NotDign XLI, Or; for dating in 390-395 see
Bury 1920, 131-114; Stein, 1928, 92-114; Lot 1936, 285-338; Polaschek, 1937, 1081-1097;
Hoffmann 1970). Other literary sources (laws, ancient authors, etc.) contain only general
information that is valid for the entire frontier of the empire (cf. Mommsen 1910, 206-282).
More important information is contained in the inscriptions and, above all, the brick stamps,
which were discovered in large numbers in the fortresses (see the list of inscriptions and brick
stamps). Unfortunately, there are very few and uncertainly dated inscriptions and the brick
stamps cannot be dated very precisely either. There is still little archaeological material that
allows conclusions to be drawn about the lives of the soldiers, for which reason the relevant
section of the work has been rather difficult to draft.

II.1.b. Current state of research

Research into the army of the province of Dacia Ripensis has lagged very far behind the
demands of historiography. D. Tudor's work mentioned above is actually a simple listing of a
few brick stamps. The general works on the history of the late Roman Empire note only or

- p. 44
mainly the data from the NotDign about the provincial army (Forni Limes, 1268-1269; Berchem
1952, 92-93). Some of the late brick stamps and inscriptions have been published in recent years
either in collections or in various articles (Morfova 1963; Tudor 1960; IDR II; IGLR). The
unpublished archaeological materials from the various fortresses are still numerous. Only single
published archaeological material is known: Comori 1978; Davidescu 1980 (Drobeta, Hinova);
SKuD (Kladovo, Hajdučka Vodenica, Malo Golubinje); Tudor 1974 (Celei). Apart from a few
metal objects from the fortresses Drobeta, Hinova and Celei and with the exception of a few

16
vessels from Drobeta, Hinova, Kula and Celei, the archaeological material from the 4th century
is still unknown.

II.1.c. Working method

This state of research compelled us to proceed in a certain way in order to get closer to our goal.
We compiled a list of the inscriptions that were found mainly in the fortresses but also in other
localities on the former border. We presented them simply, without further details that are
usually required of such a working tool (description, dimensions, etc.). In the context of each
village, the inscriptions were arranged according to the military units. All of the brick stamps
shown in the figures are reproduced on a 1: 1 scale. The brick stamp and inscription numbers on
the tables (images) correspond to their numbers in the list of military inscriptions. In addition to
the data on the troops (tactical organization, inventory, etc.), I collected several pieces of
information about the soldiers' lives (clothing, equipment, employment). The results achieved are
still quite modest.

II.2. Organization and structure of the army

On the basis of the epigraphic material we tried first of all to find out the types of units and their
stationing locations, the differences in organization compared to the previous age and at the same
time the differences compared to the situation reflected by the NotDign. Here, too, as in the case
of the province of Moesia I, we took into account the fact that the area belonged to the central
empire led by Gallienus, that the military army was formed since Aurelianus (Grosse 1920, 15)
and that the troops from Dacia were moved south (Macrea 1969, 445-456). The two source
groups I work with (Ziegelstempel and NotDign) show a different situation in the way they are
organized and deployed from the start. The brick stamps reflect a location and contain data from
a different era than the NotDign. That's why I present them separately.

- p. 45
II.2.a. The troops mentioned by brick stamps
The troops mentioned by the brick stamps (Appendix B.I.) are the following:
imagine
- legio XIII Gemina: 1st legionary division in Pontes (brick stamp with place names); 2. Legion
division in Ratiaria (brick stamp with place names) (Fig. 20. 2-3); 3rd Legion Division in

- p. 46
Orşova; 4th in Mehadia; 5. in Drobeta; 6. Legion division sent to Egypt (Mommsen 1910, 213;
Ritterling 1925, 1720; Berchem 1952, 60, 105-106; Jones 1964, 55);
- legio V Macedonica: 1st legionary division in Valeriana (brick stamp with place names); 2.
Legion division in Variana (brick stamp with place names) (Fig. 20.4-5; Fig. 21.3-4); 3. Legion
division in Oescus (brick stamp with place names); 4. Legion division in Utus (brick stamp with
place names); 5th Legion Division in Orşova; 6. in Drobeta; 7. in Hinova; 8. In Celei there are
actually still three legionary divisions mentioned: the cohorts III and IV of the legion and
probably also a cohort V (in Hărletz, Vidin, Gigen); 9. Legion division sent to Egypt (Mommsen

17
1910, 213, mentions another legion division in Thrace; Ritterling 1925, 1359-1360; Berchem
1952, 60, note 4, believes that the Romans stayed here for good);
- legio VII Claudia: 1st legionary department in Prahovo; 2. Legion division in Celei (see the
army of Moesia Prima);
- legio IIII Flavia: Legion division in Orşova (see the Moesia Prima army)
It would be particularly interesting to examine the situation in more detail, which is reflected in
the brick temples of the legio V Macedonica. These can be divided into several groups: 1. Whole
Legion LVM; 2nd Legion under praefectus PRELVM (OES, VTO, VAL); 3rd legionary division
under a praepositus LEGVMSCROM; 4. Legion division from an old subunit LVMIII,
LVMCIII, LVMPCIII or LEGVMCII, LEGVMCIII, LEGVMCIIII, LVMCORSIII; the stamp
type LEGVMCV with the reading L (egio) V M (acedonica) C (ohors) V (quinta) can also be
added to this group; if the reading is accepted then it would mean a subdivision of the Legion; 5.
Legion division from a sub-unit without the name of the legion: CORSIIII / CORSIIII; it can be
assumed that the stamp type CORSARI also belongs to this group and shows a subdivision of the
legion; 6. Legion division with the name of the station location LVMVAR, LVMOES,
PRLVMOES, PRLVMVAL, PRLVMUTO (Fig. 21.1-2); 7th legionary division under division
commanders PPRIPVAR (Fig. 21.3) or PRVAR, PROES (Fig. 21.5). Th. Mommsen (Mommsen
1910, 225) observed the factioning of the legions from Dacia Ripensis: seven detached troop
detachments for the legio XIII Gemina and six detached troop detachments for the legio V
Macedonica; an observation based solely on the NotDign. D. Tudor identified eight detached
troop divisions (praefecturae) for the legio V Macedonica on the basis of the brick stamp, which
he dated as pre-Constantinian. K. H. Dietz (Dietz 1999) has partially confirmed this
fractionation.
In addition to the brick stamps with the name of the unit, there are also brick stamps that indicate
the province and the stationing location (brickworks) DRPDIERNA, DRPAST, DRPAQVIS,

- p. 47
DARDIANA, DARPΣON, DARP (Fig. 22.1-3; 6-7) or just the stationing location: DIERTRA,
DIANA, DRVB or DRVBETA, AQVIS, BON, CEBR, ALM, VTO, OESC, VARINIA,
AVGVSTIS (Fig. 23 1-8) and also III VIMINACE. All or the majority of these brick dies were
found in fortresses. Their military character is almost certain, but it is unknown what kind of
units they represent.
There is a third category of brick stamps, the meaning of which has not yet been clarified: CAD,
GTS, ET, ERBYN, TQP, DAS, NVPPR. All of them were discovered in fortresses and are
undoubtedly related to the army. The brick stamps found in the fortress of Jatrus (Krivina;
Bulgaria) (Moesia II) are very similar to these stamp types mentioned above, also illegible, but
certainly indicate a military origin (see Table B.II).
Other units are not assigned, or more precisely, cannot be determined. It is not certain whether
the brick stamps VARIDAL, DALVARI and EQNC represent equites or cunei units, as was
previously believed. A grave inscription from Mokres (Arčar?; Vidin) mentions a numerus
Dalmatarum (Rep. 74); at the same time this unit is mentioned by other inscriptions from Mokres
(Rep. 75). The numerus has a subdivision, a centuria, similar to that in 2-3 AD. If the dating of

18
the two inscriptions is certain, then at the end of the 3rd century we have a numerus somewhere
in Arčar or Vidin.
In the period of time represented by the brick stamps, it is very difficult to determine the
mounted and infantry units. Perhaps the Legion's subdivisions are infantry units. This Arčar
numerus was certainly a mounted unit; the equites from the second above-mentioned inscription
are also depicted on the monument. If this was done correctly, then the equites (dalmati?)
Extended their activities in Dacia Ripensis during the Diocletian period. A brick from Celei
(Rep. 110) depicting two (or three) horses is another evidence of the presence of mounted units
on the late Roman border in Dacia Ripensis.
The investigation of the brick punches from the group COILING, LCOELCR, COELPRIMI,
COLPRO is very interesting. Your reading remains a mystery.
The main conclusions drawn from this list are: the old-style military units have disappeared; of
the two legions brought from Dacia, very few were moved away from the province; the legionary
divisions probably stationed temporarily; the legio XIII Gemina had 6 departments and the legio

- p. 48
V Macedonica at least 7 departments. The former legion has 3 and the other 4 departments with
place names.

II.2.b. The troops from the Notitia Dignitatum

The troops from the Notitia Dignitatum (Table B.IV) are as follows: 9 cunei equitum, 6 auxilia, 1
milites, 4 divisions of Legion V Macedonica, 5 legionary divisions of legio XIII Gemina, 2
cohorts, 2 naval units. Remarks: old type units are mentioned; the divisions of the two legions
correspond only partially to the situation reflected by the brick temples; new types of units
appear which are more numerous than the legionary divisions. Apparently the stamps reflect the
situation at the time of the Tetrarchy and the NotDign the situation after the reforms of
Constantine. A. H. M. Jones (Jones 1964, 99) thinks that the disappearance of the cavalry
divisions of the alae and cohortes is due to the reforms of Constantine. Legion XIII Gemina had
three detached troops in the western part, between Egeta and Ratiaria, and Legion V Macedonica
had four detached troops between Cebrus and Oescus. Each detached troop division was led by a
praefectus legionis, which indicates that each troop division was an independent formation (cf.
Berchem 1952, 92-93).

II.2.c. The relationship between troops and fortresses

In the case of the province of Dacia Ripensis, it is difficult to determine the relationship between
troops and fortresses. The list of fortresses (Appendix I) contains 45 points (for the Age of
Tetrarchy - Constantine in particular), but brick stamps were discovered only in 17 (Appendix
III). In the western section the DRPDIERNA (10) brick stamps predominate; in the central
section the brick stamps DARDIANA (8) and DRPAQVIS (7). Legion divisions appear only or
mainly to the east of Orşova: in the central area the legio XIII Gemina (Orşova - Arčar), in the
central and eastern area the legio V Macedonica (Arčar-Vit) (see Table B.II).

19
In the NotDign a different situation appears: there are 23 fortresses and in each there is at least
one unit; In the former Roman camps there are 2-3 units each (Appendix B.IV). The shifting of
units starts from Orşova; the number of units is greater than the number of fortresses.
In general, however, it can be stated that the number of fortresses is greater than in the case of
the province of Moesia I; the number of units is smaller only because no research has been done
and no materials have been published.

II.2.d. Number, line-up and tactical organization

If one assumes that the number of a legionary division was usually 1000 soldiers (Mommsen
1910, 260; Grosse 1920, 34; Berchem 1952, 93; Saxer 1967, 118-119), that the two legions only
lost one legionary division and that the number of the Legion until Diocletianus remained
complete (Grosse 1920, 35-36), then 4,000 + 4,000 soldiers remained in the province. However,
we must assume that under Gallienus the cavalry units were withdrawn from the legions (Parker
1928, 210-211; Grosse 1920, 36; Mommsen 1910, 17, note 3, speaks of the increase in the
number of legion riders) and that both legions are one lost a certain number of soldiers. At the
time of Gallienus and Claudius II, troops were sent from this area to Italy and Gaul (cf. Saxer
1967, 57-58 No. 108). It is assumed that under Constantine the population was greatly reduced
by the relocation of some units to comitatus. Th. Mommsen (Mommsen 1910, 213) mentions one
legion division each from Legions V Macedonica and XIII Gemina in the Orient and in Thrace.
Each legion was left with 3,000 soldiers. In this case the divisions of the two legions numbered
approximately 300/500 soldiers each. This number seems closer to reality than the estimated
stocks. Mommsen (1910, 220, notes 1-2; 260) claims that the stocks increased greatly during the
time of the tetrarchy; - a late legion is said to have had 1,000 soldiers! (?). In any case, the name
legio is a traditional tradition (Berchem 1952, 91, 93). If it should ever turn out that the brick
stamps DALVARI and EQNC represent equites units, then these two units can also be added to
the legions, as well as the units represented by the brick stamps DRP ... and DAR ...
In the NotDign, the number of units is greater. But we only have 2 legionary divisions, which
means that some of the legions have disappeared. If we assume that a legionary division has 300-
500 soldiers (Amm.Marc. XX 4. 2; XXI 11. 2; XXXI 10. 13), a cuneus 1000 (Grosse 1920, 51-
53), the auxilia 500 (Grosse 1920 , 38-42), the milites 300 (Grosse 1920, 54), the cohortes 500
(Grosse 1925, 42-45; Breeze-Dobson, 1991, 209) claim, due to the organization of the late
barracks, that a miliaria cohort 600 soldiers and a quingenaria cohort had 300 soldiers) and the
fleet units had 300 soldiers (Grosse 1920, 72-76), then the number would be 3500 + 9000 + 3000
+ 300 + 1000 + 600 soldiers, which is a much larger number than those from the previous age.
At the moment it is difficult to explain that the border troops under Constantinus were reduced
by the transfer of some of their units to the military army (Mommsen 1910, 220-221).
Th. Mommsen (Mommsen 1910, 208) and R. Grosse (Grosse 1920, 52) were of the opinion that
the cunei equitum were formations formed from barbarians. A. Alföldi (cf. Berchem 1952, 94)
claimed that the cunei were formations that arose from the former alae and cohortes.
However, we would like to mention that of the 45 fortresses, at least 35 are quadriburgia of very
small dimensions. At the same time, of the 23 fortresses from the NotDign, at least 5

- p. 50

20
quadriburgia are very small. The interior of the Orşova fortress was 32 x 32 m; but if you
subtract the 644 m² of the uninhabited central zone from the 966 m², there are only 322 m² that
were habitable; In addition, there are 196 m² of the four corner towers, which results in a total
area of 518 m². If you do a similar calculation for Hinova, you can see that the habitable area is
340 m². No more than 200-250 people could live in such an area.
The list of units is more difficult to determine. During the time of the Tetrarchy, infantry units in
particular seem to have been concentrated in the western part, etc. those named DRPDIERNA,
DRPAST, DARP, DARPΣON, DARDIANA, DRPAQVIS. Infantry units (the legionary
divisions) were safely stationed in the central area. In the eastern sector the divisions of the
legions and the presumed equites units (Appendix B.III) were under praefecti ripae. In the
NotDign (Appendix V) in the Eastern sector, mainly cunei, but also auxilia units are mentioned.
In contrast, there were predominantly legionary divisions in the western sector, i.e. infantry.
A law from the year 311 (TIR I 93 No. 24) shows that at that time all troops from Illyricum were
commanded by a duces. In the time of the Tetrarchy, the praefecti were in command of the
legionary divisions. They headed both the Legion (PRLVM) and the zonal (PRLVMOES,
PRLVMUTO) and border departments (PRRIPVAR, PRVAR). We have a praefectus ripae - if
not a praepositus ripae - in Varinia. It can be stated that the praefecti ripae or the preapositi
ripae from Legion V Macedonica appear either by border section or by fortress. At the moment it
cannot be determined what the difference between these offices was. The praepositi were at the
head of both local and legionary divisions from neighboring provinces (Appendix B.III). An
inscription (AnnÉp 1942-1943 No. 84) indicates that the praepositi limitis received orders for the
construction of fortresses from the praesides. The relationships between praesides and duces
must also be determined in this way. In our case, praepositis ripae, praepositus legionis and
praepositus (fortress) can hardly be equated. The existence of older ranks, tribuni, centurio,
circitor, praepositi is more difficult to explain (Mommsen 1910, 273-275; Grosse 1920, 145-
150). The dedicatory inscription by Drobeta dedicated to Iuppiter shows that one of the cohors
survived in Drobeta; their leader was still a tribunus (see also Appendix B.III). We do not know
whether she was a legionary cohort or an auxiliary cohort. Perhaps it can be dated for the time of
the tetrarchy, before the military reforms. The same situation is also valid for the numerus
dalmatarum. In the NotDign the praefecti were in command of the legions, milites and clasis,

- p. 51
while the tribuni were at the head of the cohorts (Mommsen 1910, 275; cf. CTh. VII 21. 1 (313);
VII 4.1 (325 ).
An inscription (AnnÉp 1942-1943 No. 84) indicates that the praepositi limitis received orders for
the construction of fortresses from the praesides. The relationships between praesides and duces
must also be determined in this way. In our case, praepositis ripae, praepositus legionis and
praepositus (fortress) can hardly be equated.
We do not have any information regarding recruitment into these units. The generally applicable
norms were probably used (Jones 1964, 614-619). If the inscription from Glava (next to Hărletz)
(Appendix B I 128) was correctly dated, then Valerius Tzitza qui et Vitalis - a veteran - had two
sons as milites. This is called a traditional family business and suggests that it will be recruited
on the spot. Like Tzitza, the father who has a Thracian or even Dacian name, the sons have Latin

21
names (Valerius, Vitalis); this certainly indicates a Romanized population. Most of the Romans
were safe (Appendix B III). Malalas claims that Diocletianus settled in the border fortresses
stratiotai limitaneous (Neumann 1968 col. 876-888). The other sources call them ripenses (SHA,
v. Aurel. XXVI 38. 4); ripariensium et castricianorum (VII 1. 18) [400]; militia ripensis (CTh.
VII 20. 4 [372]; VII 22. 8 [372]); riparienses milites (CTh. VII 4. 14 [365]); ripensis militis
(CTh. VII 13.7) [375]; riparienses castriciani (CTh. VII 1. 18 [400]); ripenses qui in ripa per
cuneos auxiliaque fuerunt constituti (CTh. VII 13. 7 [375]); castellanus miles (CTh. VII 15. 2
[423]. Apparently the NotDign also confirms this situation: dacisci, marianensi, crispitiensi are
Romanized locals.

II. 3. About food, clothing and armor

The sources relating to food were collected and treated by A. H. M. Jones (Jones 1964, 623-630),
but these apply to the frontier army in general. Most likely, the soldiers were paid in cash. The
large number of small bronze coins from the fortresses (e.g. Hinova, Kula) or even the hoards
(e.g. Dierna) indicate this. M. Vasić (Vasić 1983, 99-112) claims that the 5 solidi of Thedosius II
found in Kladovo / Pontes belonged to some federati. Their presence confirms that the limitanei
were also paid in cash. The information on remuneration is also of a general nature. In the case
of the province of Dacia Ripensis, a fabrica is documented on the border at Ratiaria (NotDign.
XLII, 38). Until new data are found, we consider this general information also valid for Dacia
Ripensis. As in the case of the province of Moesia I, the archaeological discoveries show a
certain uniformity of the ceramic material in all fortresses (bowls, mortaria, jugs, jugs, chalices,

- p. 52
sled lamps), both glazed ceramics and imported ceramics (amphorae) or ceramics for the
everyday use (pots, lids). All point to care from the same center. The coin circulation also has a
uniform character.
Few weapons were discovered. Old type lance tips from Orşova (excavations N. Gudea), D.
Butorke, Prahovo (Janković 1981, 179 fig. 72.a), Radujevac (Janković 1981, 179 fig. 72, 5),
arrowheads from Sucidava (Tudor 1937 -1938, 370 fig. 7 k, l; fig. 6 t, u) and even a gladius
(Hinova) (Davidescu 1980, 134). Glandes made of clay and stone (Hinova) were still used
(Davidescu 1980, 134). At the same time, later-type weapons appeared: the balistae (Orşova:
Gudea-Baatz, 1974, 50-72 fig. 6-7; Baatz 1975, 432-434; Drobeta: Davidescu 1980, 134) and the
late spear and arrowheads (Orşova , Drobeta, Celei) (excavations N. Gudea; Davidescu 1980,
134; Toropu 1976, 64). Numerous iron axes have been published as weapons in the fortress of
Kula / Castra Martia. It is unknown to what extent they belonged to the Roman army. But such
finds are represented as weapons of the Goths (Magomedov-Levada 1996, 307-308, 562 fig. 5).
Since the work on late Roman weapons is incomplete and very theoretical (Couissin 1926;
Feugère 1993), we have tried to compile a typology of weapons for long-range combat based on
the published finds from the fortresses of Dacia Ripensis. I have found that there are several
groups of weapons: Group I: large lances (Fig. 24); medium-sized lances and small lances with a
nozzle or with a fastening pin (Fig. 25-26); Group II: heavy arrowheads with a socket or with a
fastening spike (Fig. 27); Group III: light arrowheads (Fig. 28). Analogies with weapons from
late Roman fortresses from other Roman border areas from Raetia II (Vemania - Fig. 30) or from
Moesia I (Gornea - Fig. 29) show that this division is valid.

22
Compared to the weapons of the Roman troops from the area from the 2nd-3rd centuries.
Century AD (Vlădescu 1983; Gudea 1994) it can be stated that both the type and the shape of the
weapons for long-range combat have remained unchanged. The heavy lances and arrows
continued to be used. Apparently, however, the number of light arrows has increased and new
types of these weapons have appeared. The presence of the latter partly explains the presence of
the throwing machines (ballista) and the bows.
The data on the outfit are based only on the metal objects belonging to the same period: Onion-
head fibulae (Karataš, Orşova, Drobeta, Hinova, Puţinei, Izvoarele, Celei) (SkuD, 164, plate
XIV; excavations N. Gudea; Davidescu 1980, 178, 179, 184; Tudor 1974, Fig. 36; Toropu 1976,
59). Parts of a belt buckle (Hinova, Celei) (Benea 2002, 290 plate II Fig. 78; Tudor 1974, 89 Fig.

- p. 53
35), brooches with a folded foot (Toropu 1976, 59), strap tongues (Drobeta, Celei) (Toropu
1976, 60), various fittings (Celei) (Toropu 1976, 60). Their presence means that the outfit did not
change significantly.
The old-style heavy weapons, the lance and the spear, seem to have disappeared.

II.4. Data on the life of the soldiers

Data on the lives of the soldiers were obtained only by studying the archaeological material. The
known material is much more sparse than in the case of the province of Moesia I. Also the data
regarding the situation in the 1st-3rd centuries AD are completely absent. There is only a little bit
of information about Drobeta (Davidescu 1980, 133). Exposition Totius Mundi et Gentium, a
kind of inventory of the resources of the Roman provinces around 359-360, also provides brief
information on Dacia Ripensis. Together with the province of Moesia I, it is referred to as "sibi
sufficientes"; The peculiar climate of the two provinces is also emphasized. This means that these
provinces were based on their own economy, which could at least secure regular food for the
troops (Zahariade 2006, 13). The literary sources (decrees, ancient authors, etc.) offer us only
general data that contain the duties and benefits that the emperors offered the ripenses (Jones
1964, 630-636; Grosse 1920, 234-246; Berchem 1952, 75- 87). Storage depots for the provisions
of the army (horrea) are believed to be next to the fortress of Porečka Reka (Petrovič 1976, 259-
275). However, it is unknown what was stored in these storage facilities.
Sure the ripenses were farmers. This fact is confirmed by the fundi limitrophi (on the soils
assigned to the garrisons: CTh, V 12. 2; VII 15. 2 = CJ XI 60. 2). According to a decree of
Valentinianus I from the year 364 (CTh VII 20. 8) a veteran received two pairs of oxen and 50
modii of grain, sufficient for the cultivation of 10-12 yokes. A veteran protector primae received
two pairs of oxen and 100 modii of grain. These data are confirmed by Annonimus Valesianus
(De rebus bellicis V 4). Numerous agricultural implements were found in the fortresses:
plowshares, hoes, picks, scythes (Celei) (Tudor 1974, 76-77 fig. 81; Tudor 1978, 445 fig. 144;
Toropu 1976, 61-62), axes, hoes, Sensen, Hippen, Messer (Hinova) (Davidescu 1980, 101-106),
ploughshare, hoes (Kula) (Atanasova 1972, 170 plate 39), mills (Orşova, Hinova, Celei)
(excavations N. Gudea; Davidescu 1980, 180; Tudor 1974, 76.). Other tools indicate other
occupations: carpentry (Hajdučka Vodenica, Celei) (Toropu 1976, 62; Comori 1978, 208,
no.161), domestic activities such as spinning and weaving (Orşova, Hinova, Kula, Celei)

23
(excavations N. Gudea; Davidescu 1980, 193, 197-199; Atanasova 1972, 170-171; Toropu 1976,
62), metallurgical activities (Hinova) (Davidescu 1980, 106 - ladle, anvil, slag). Cattle were
raised; the bells and chains discovered (Celei) (Tudor 1974, 76) prove this: It was fished; Fishing

- p. 54
rods, harpoons, tools for weaving nets, clay and lead weights were found (Hinova, Celei)
(Davidescu 1980, 103, 106; Toropu 1976, 62).
People lived simple and poor in rooms with mud walls that stuck to the ramparts; its roof was
based on the wall and on the pedestals built parallel to it (Celei, Hinova, Orşova) (excavations N.
Gudea; Tudor 1974, 70; Toropu 1976, 62-63; Davidescu 1980, 206). In Hinova, two rooms were
discovered on either side of the intermediate tower. The hearths were made of bricks (Celei).
Little building material was found. Apparently no more roof tiles and hollow tiles were made,
only tiles. Hook nails, construction clips and nails were still used in building; but they seem to be
bigger. They were found in large numbers in the rubble or fire layer, from which conclusions can
be drawn about the way in which the buildings were roofed.
Few objects made of bronze and glass were discovered. Only a few fragments of goblets and
goblets discovered in Orşova and Celei can be dated with certainty as of the 4th century. On the
other hand, in addition to other bone objects (needles, tubes, plates), bone combs appeared
(Orşova, Hinova, Celei) (Toropu 1976, 61; Davidescu 1980, 193). Of the jewelry items we
mention only the glass beads (Hinova, Celei) (Tudor 1974, 77) and a necklace made of gold
(Celei) (Toropu 1976, 60-61). Earrings and rings made of bronze, earrings made of gold, a little
chain made of gold and a bracelet made of bronze were also discovered in Celei (Tudor 1974,
77; Davidescu 1980, 203). With regard to food, discoveries were made in Celei and Hinova
(Tudor 1974, 76); Charred wheat grains and numerous bones of domestic and wild animals as
well as bones were discovered in some vessels.
Apparently the soldiers-to-population ratio was pretty bad. Amm. Marc. (XXII. 4, 7) speaks
about the "tyranny of the army". Themistius (Or. X 137 e) writes that the soldiers oppressed the
provincial residents and that “they were not soldiers, but bandits”.
There is very little data on the spiritual life of the soldiers' ripenses. In Celei, monuments of the
ancient deities (Lar, Thracian knight, Mithras) were found in the apartments (Tudor 1974, 88-89;
Toropu 1976, 107-109), in Hinova fragments of the Mithraic relief, satyr statuette, bronze
statuette of Venus (Davidescu 1980 , 180-203), indicating the persistence of pagan religions. The
erection of an inscription in honor of Juppiter cohortalis, or the protector of the cohorts, at the
beginning of the 4th century in Drobeta (Rep. No. 41) can also be an argument for the continued
existence of paganism within the military society during the 4th century be considered in the
border area. V. Velkov (Festschrift Vetters 1985, 270-272) thinks that at the end of the 3rd

- p. 55
century Christianity was not very widespread among soldiers from the border of the province of
Dacia Ripensis. Nevertheless, a verse from the 28th Psalm for holy water was carved into a
bronze vessel discovered in Kladovo-Kostol / Pontes. In fact, the writers dealing with 4th century
Christianity don't particularly mention the presence of Christians in the troops on the border; for
example B. Brenk (Christianization of the late antique world. City, country, house, church and

24
monastery in early Christian times, Wiesbaden 2003) says nothing about the Christians in the
army and in the fortresses. On the other hand, the data on Christianity are very sparse and
confusing (Tudor 1974, 134), vessel opening with cross (Celei) (Davidescu 1980, 226), vessel
bottom with cross (Hinova); Davidescu 1980, 226, lamp with a cross-shaped handle (Insula
Banului); Davidescu 1980, 176, Basilica (?) (Drobeta). There are very few bricks with italic
incised inscription (Orşova, Hinova) (IDR III 1 No. 48, 49 (Orşova); Davidescu 1980, 199
(Hinova), but all are in Latin. This is further evidence of ethnicity of the soldiers from the
fortresses.

- p. 67
Anexa B. III. Names of commanders, officers and soldiers (vezi tabel p. 67)
B. IV. Table showing the distribution of military units in the Notitia Dignitatum (vezi tabel
p. 68)

- p. 70
APPENDIX B V
III. The collapse of the late Roman border on the Middle and Lower Danube after 378 with
a special view of the province of Dacia Ripensis
III.1. The Goths on the Lower Danube (240-378)
This chapter is an introductory sketch of the military events of the late 4th century that led to the
collapse of the Roman border on the Danube.
So we do not intend to provide an exhaustive history of the Goths in the period mentioned. There
are several general (Wolfram 1990; Kazanski 1991; Bierbrauer 1994), regional: Russia, Ukraine
(Magomedov 2001) or local works: Romania (Ioniţă 1972; Harhoiu 1990).
The most complete work on the Goths, which connects the written sources with the
archaeological facts, is that of B. Magomedov (Magomedov 2001). In this work the author
presents the chronological sequence of the movements of the Goths from north to south and their
material culture and correctly determines their borders with the territory of the Dacian provinces.
He established three major periods of expansion and development of their material culture: I.
The period 230-270, when the Goths moved south and reached the north coast of the Black Sea,
reaching the south of the Vltava and the Romanian Plain. This is the time of the great attacks
against the Roman Empire, from Lower Moesia to Pontus and Bithynia, followed by a return to
the “regions of origin” (Magomedov 2001, 287 Fig. 88); II. 270-330, a period of relative calm in
which attacks against the Empire are almost completely absent (Magomedov 2001, 288 fig. 89);
III. 330-378, a period of general attacks against the Roman border, followed by settlement on the
territory of the empire, especially in the provinces on the Middle and Lower Danube
(Magomedov 2001, 289, Fig. 90).
In the first period (230-270) the relations between the Goths and the empire were rather violent
(cf. Klose 1934, 147 ff.). The series of attacks against the empire on the Lower Danube (Upper
Moesia, Dacia Malvensis, Lower Moesia), against the provinces in Asia Minor (Bithynia,

25
Galatia, Pontus, etc.) was almost endless. We're going to cover just a few of the most important
attacks to give the most truthful picture possible:
- 235 - attacks against cities on the seashore; Histria and Olbia are destroyed (SHA, v. Max. Et
Balb. XX 11-12; XXI 16. 3; Gerov 1980, 671);
- 238 - joint attack by the Carps and the Goths (Vulpe-Barnea 1968, 230-231);
- 242 - Attack of the Goths and Carps against Low Moesia; they reach Thrace (Vulpe-Barnea
1968, 234; Gerov 1980, 378);

- p. 71
- 248 - the Goths, Carps and Taifals attack low moesia and reach Marcianopolis (Gerov 1980,
381-382);
- 249-250 - the Goths and Sarmatians attack low moesien; Abrittus and Novae (the legionary
camp) are destroyed (Gerov 1980, 381-383);
- 251 - the Goths, under the leadership of King Kniva, attacked Low Moesien along the entire
length of the front (Gerov 1980, 384-385);
- 253 - Attack against Low Moesia (Gerov 1980, 391);
- 258-260 - the Goths attack low moesien on land and on the sea; they reach Byzantium (Vulpe-
Barnea 1968, 254);
- 263 - the Goths, together with other barbarians, attack low moesia on the sea and the provinces
of Pontus, Asia, and Bithynia also on the land; they penetrate as far as Macedonia and Greece:
“multa gravia in solo romano fecerunt” (SHA, v.Gall. XIII 6; Eutrop. IX 8.2; Gerov 1980, 393);
- 269 - the Goths, Herulen, Gepids and Bastarnen attacked Lower Moesia and the provinces in
the north of Asia Minor on the sea and on the land; in the country they penetrated as far as
Thrace, Greece, and in Upper Moesia as far as Naissus; they were defeated by Emperor Claudius
II (Aur. Vict. Caes. 33.3; Eutrop. IX 11. 1; Gerov 1980, 393-394).
At the end of this period of time, the Greek cities on the north coast of the Black Sea, the
Bosporan Empire and its cities were destroyed, Lower Moesia was plundered and devastated, the
Dacian provinces were abandoned in order to strengthen the Danube border with troops.
Eutropia. (IX 15. 1) clearly shows “(Aurelianus) provinciam Daciam, quam Traianus ultra
Danubius fecerat, intermisit, vastato omni Illyrico et Moesia”.
After 275 and up to the middle of the 4th century there was a period of relative calm. It is the
time of the great heyday of the Sântana de Mureş-Černjachow Gothic culture. At the end of the
period, conflicts occur again.
- In 323 the Goths took advantage of the concentration of troops for the battle between
Constantinus and Licinius; since “the Limes was neglected”, the Goths attacked the empire under
Rausimond and advanced as far as Thrace (Anon.Val. V 13; Lyd. De Mag, II 10; III 31. 40);

26
- 331-332 - war of the emperor Constantinus I against the Goths in the "Sarmatian areas"; it ends
with the peace of the year 332; the Goths become foederati (Euseb. Chron. 233; Eutrop. X 7. 1);
this status is maintained until late in the 4th century when the Goths settled on the territory of the
empire. L. Váradi (Váradi 1961, 40) mentions an army body of 40,000 Goths who were available
to the Romans;
- 338 - Emperor Constantinus I divided the empire between his three sons and his nephew
Dalmatius; the latter received Illyricum and the Gothic bank! (Anon.Vales VI 35);
- before 353 the Goths reached Thrace (Thompson 1956, 379).

- p. 72
The situation changed in the following period when changes in the structures of the Goths took
place due to pressure from the Huns. The state of Ermanarich collapsed in 375. Some of the
Goths, under the leadership of Athanaric, fled west and probably reached the south of the Vltava.
- 376 - Emperor Valens signs a peace with the Goths and gives them access to the empire (Them.
Or. X 133-140) - it concerns the group of the Fritigern; consequently, the author writes: “the
bank of the river is full of fortresses, the fortresses full of soldiers, the soldiers full of weapons”;
Eusebius, Chron. 248;
- 377 - as a result of the looting caused by the Goths - they were accepted into the empire in 376
- the population of Dacia Ripensis and Moesia II was exempt from taxes (CTh VII 6. 3);
- 378 - as a result of the dissolution of the border defense after the battle of Hadrianopolis,
groups of barbarians (Goths, Alans, Huns) invaded Dacia Ripensis; they were neutralized one
after the other (Amm. Marc. XXXI 1.16);
- after 378 - the Sarmatians and Huns settled in Castra Martis in Illyricum; the Huns conquered
Utus and Oescus in Dacia Ripensis (Iord., Get. 265);
Another part of the Goths, led by Alatheus and Saphrax, tried to resist. The last two groups
began to put pressure on the Roman border and both eventually crossed the Danube after 378 and
settled in the empire. In 382 a treaty was signed by which the empire ceded the area between the
Danube and the Balkan Mountains to the Goths (i.e. the provinces of Dacia Ripensis and Moesia
Secunda); the Goths had to defend it (cf. Zos. Hist.Nov. IV 34. 4: the group of Athanaric).
III.2. The battle of Adrianople and its aftermath
The barbarian pressure on the borders on the Middle and Lower Danube began after the middle
of the 4th century AD. Ammianus Marcellinus (XXVI 4. 5) expresses this very vividly: “hoc
tempore velut per universam orbem Romanam bellicum cannentibus bucinis; excitae gentes
saevissimae limites sibi proximos persultabant ”. It is possible that the hoards in the Banat and
Little Wallachia, which usually end around 361-365, reflect exactly this fact (Chirilă et alii 1974,
84-92; M. Vasić, who notes the existence of a layer of destruction in some fortifications
mentioned in the area of the Iron Gates (Vasić 1977, 327). What is certain is that the emperors
took a number of measures to defend and strengthen the borders (see CTh, XV 1. 13
(Valentinianus an Tautomedus dux Daciae Ripensis); Them. Or. X (FHDR II 65), CTh XXIX 6.

27
2 (Valens). However, as a result of an agreement in 376, Emperor Valens allowed large groups
of Goths to settle in the provinces of Moesia II and Dacia Ripensis, on the condition that that
they participated in the defense of the empire (Zosimus IV. 20; Eunap. frg. 42; cf. Thompson
1963, 107-108; Stahlknecht 1966, 65, note 74 with all sources). Up to 370 were along the Lower
Danube several emporia (Them. Or. X 135); after the ers During the last war against the Goths,
their number was reduced to two. Some of the Goths among Fritigern received receptio through a
foedus, who allowed them to settle in Moesia II, Dacia Ripensis and Thracia. Apparently the
uncertainty of the borders grew from that point on. It should also be mentioned here that an army
sent by Emperor Gratianus and commanded by Frigeridus, magister militum per Illyricum, came
from Augusta Treverorum (Trier), at Castra Martia on a number of Alans from the Gothic group
under Alatheus and Saphrax bumped. The Roman army did not advance further east (Fig. 32).

III.2.a. The location of the provinces on the lower Danube

After the Battle of Adrianople (378) the situation in all provinces along the Danube, from
Pannonia Valeria to the sea, was desperate (on the course of the battle: Amm. Marc. XXXI 1-17;
13. 1-6; Zos. Hist .Nov. IV 50; Eunap. Frg. 60). The groups of Alatheus and Saphrax (Goths,
Alans, Huns) joined the group of Goths under Fritigern, who rose up against the Romans. This
emerges from the overview of the most important literary sources:
- the Goths occupied Moesia II, Dacia Ripensis and Moesia I (A, I, 26); they ruled the entire
northern part of the empire up to the Danube "tamquam genitales terras" (Appendix C.I, 9
Aur.Vict.); the mine officials in Moesia I had not yet returned to work in 381 because they feared
the barbarians (CTh I 32.5);
- the Gothic attacks can also be felt in the west, in Pannonia II, Valeria, up to the borders of Italy
(A, I, 10 Amm.Marc.);
- 381 - the city of Mursa (Savia) is occupied by Goths;
- 389 - Valeria and both pannonia are lost (Appendix C. I 17 PanLat; Appendix C. I 12
Claudianus); Dalmatia is lost (Appendix C.I 14-15 Hieronymus); Ambrose buys off his believers
captured in northern Italy (Appendix C.I 7);
- 392 - the Sarmatians reach Italy; the alpine passes are being fortified;
- 401 - Alaric moves without hindrance to Ravenna (Appendix C.I 23);
- the Balkan chain is crossed south by the barbarians (Appendix C.I, 19-20 Themistius);
- Thrace is gradually occupied and devastated (Appendix C.I 12-13, Claudianus; Appendix C.I
14 Jerome); 386, 388, 391 great battles are fought in Thrace, in which the emperor also takes
part; - Hellas, Macedonia, Dardania and Thessaly are sacked (Appendix C.I 18 Theophanes) -

- p. 74
Iustina, mother of Valentinianus II, is besieged in Thessalonica; Pillage under the walls of
Constantinople;
- Scythia is sacked to the east (Appendix C.I18; Theophanes; Appendix C.I 14 Hieronymus);

28
- Looting and destruction throughout the area: - we lost 25 peoples (Appendix C.I 21 Libanius); -
they sacked about 20 provinces (Appendix C.I 18 Theophanes); - the largest and most powerful
part of the empire was lost (Appendix C.I 29 Philostorgius); lost provinces (Appendix C.I 20
Themistius); Illyricum is a grave (Appendix C.I PanLat).
This overview shows the following:
1. All Roman provinces on the Middle and Lower Danube were threatened, plundered, and
devastated;
2. some of the provinces, and especially Moesia II, Dacia Ripensis, and Thracia, were occupied
by the barbarians and were considered lost (Libanius) or "wrested" (Philostorgius); the peace
made in 382 envisaged the settlement of the Goths in Moesia II and Dacia Ripensis as foederati.
Alaric was appointed magister militum for both the eastern and western empires; in 409 he
signed a treaty with Emperor Honorius (Zosimos V 50);
3. The incursions and destruction were not limited to the border provinces, but extended south of
the Haemus, into the provinces of Thracia, Dardania, Macedonia, Thessalia, Achaia and west of
the Haemus (Dalmatia, Pannonia II);
4. The difficult connections between the cities, between different parts of the empire, the
insecurity of traffic, the siege of large inner centers, including the capital, indicate the extent of
the catastrophe;
- the crossing of the Danube by the barbarians continued after 378 (Iord. Get. 129; Zos.
Hist.Nov. IV 32-34; Orosius VII 33; Amm.Marc., XXXI 9. 3 “transire flumen direpturi vaccus
defensoribus loca ”; another group of Goths under Farnobius crossed the Danube; the well-
known agreement between Emperor Theodosius I and the Visigoths did not put an end to the
looting and destruction (on the peace of October 382: Them. Or. XVII 210-214; Iord Get.131;
Zosimus II 34; Aur.Vict. Caes. 47, cf. Stahlknecht 1966, 76; Jones 1964, 157);
- this fact lasted for a long time. It can be said that Jerome's assertion, "viginti et amplius anni",
records a fact. Until Alaric withdrew to the west (401), when the first measures to rebuild the
fortifications (408), one cannot speak of Roman rule here.

III.2.b. The army

After 378/380 there is no more reliable information in connection with the border army. The
location offered by the stamped bricks from the time of the Tetrarchy is unique (see below). It is

- p. 75
assumed that Constantine I reorganized the frontier army (in 325: Berchem 1952, 83-88);
including on the Danube, the old military units (cohorts and ales) were replaced by infantry units
of the auxilia type and the cavalry units by cunei (Jones 1964, 608). He also completed the
system with the duces. In any case, it is assumed that he reduced the number of troops at the
borders in favor of the mobile military army (Berchem 1952, 101-110, 115), this in the period
312-324: “This security (which Diocletianus had enforced on the border) was destroyed by

29
Constantinus, who withdrew most of the soldiers to transfer them to the cities ”(Zos. Hist.Nov. II
34).
Under the sons of Constantine, later during the reigns of Valentinian and Valens, the situation
remained unchanged. The same author is of the opinion that towards the end of the reign of
Theodosius II the system documented by the Notitia Dignitatum was introduced (Jones 1964,
609). The frontier armies in our areas were commanded by praefecti legionum and duces (in
Dacia Ripensis, Moesia I or Moesia II and Scythia) (Jones 1964, 609). The responsibility of
these duces limitis for the fortifications, probably also the praefecti legionibus (ripae), was
established (cf. CTh XV 1.13). Presumably these commanders were also responsible for the
fundi limitrophi, areas that were intended for the maintenance of the army (cf. CTh V 12.2). This
author is of the opinion that the last phase of the dissolution of the army in the western part of the
empire took place in the time of Honorius (Jones 1964, 612), in the eastern part later. L. Váradi
believes that “the gates of the Roman army were opened to the Goths after the peace of 382”
(Váradi 1961, 41). The Federation began to be used extensively in the Illyricum as well, and the
recruitment of Roman citizens was abandoned.
So it is assumed in the historiography that is based almost exclusively on the Notitia Dignitatum
and various texts of historical literature (e.g. Oros. Adv. Pag. VII XL. 7).
None of the authors who wrote about the Roman army in the 4th century also referred to the
examination of the stamped bricks, but mainly to the Notitia Dignitatum (R. Grosse, D. van
Berchem, B. Hoffmann).
The facts that can be inferred from the stamped bricks are completely different from that which
emerge from the Notitia Dignitatum. In the time of the Tetrarchy and in the 4th century we are
not dealing with units of the type in the Notitia Dignitatum (see Chapter II.2.a.). In the course of
the 4th century up to 478 there are no clearly attested units of this new type.
- the end of the fortifications between 378-390 indicates their destruction or abandonment;

- p. 76
- The dismantling of the fortifications means, without reservation, either the destruction of the
military units or their withdrawal from the site. Given that these limitanei were actually armed
peasants, their withdrawal or transfer means the disappearance of the Roman element from the
area;
- the end of the rows of coins in the fortifications, both in the case of hoards and individual finds,
suggests the same thing;
- It is known from the sources that the defense in our area was taken over by Alaric, who was
appointed magister militum in Moesia Prima and Dacia Ripensis in 390, 397 magister militum
per Illyricum (Jones 1964, 183), after having spent many years in Thrace and had occupied
Dacia. A. Schwarcz (foederati, in RGA 9, 1995, 295) shows that in his capacity as magister
militum according to the foedus of 397 he held the jurisdiction over the exercitus Gothorum,
controlled the gunsmiths, collected taxes and also the tasks of a praeses provinciae would have.
He was later appointed magister militum for Dacia and was commissioned to occupy Epirus
(Jones 1964, 185);

30
- From this period, when Alaric was magister militum in Illyricum and Dacia, there is no
information about the border army; the troops in the Notitia Dignitatum certainly come from the
time after the Visigoths withdrew from this area (408-410).
From the representations of Priscus Panites (Priscus 281, 2; Theoph. Conf. Chron. 102, 13-26) it
was concluded that the city of Ratiaria resisted until the attacks of the Huns from the 5th century.
This statement could be confirmed with the information from Iordanes (Rom. 33), according to
which Arnigia, magister militum in Dacia Ripensis, was killed in the battle against the Huns of
Etzel (Attila) (cf. also with Marcellinus Comes ad.an. 447. 5 ).

III.2.c. The location of the border

The position of the border is derived indirectly from the information given above. Since the
fortifications were destroyed, the military units disappeared, because the Goths / Visigoths were
more concerned with the plundering of the inner provinces than with the defense of the empire, it
can be said that the border no longer worked at least in the period 379-400. This is also evident
from the fact that the sources mention multitudes of Goths, Huns and Alans who constantly
crossed the border to the south without anyone stopping them (Váradi 1961, 41).

III.3. Conclusions
III.3.a. Some questions of military history
The political and military situation described above clearly shows that there was no longer any
fortified frontier on the Danube in the province of Dacia Ripensis. This statement based on the
literary sources and the political events is supported by the archaeological sources. In the

- p. 77
investigated fortifications of the provinces of Moesia I and Dacia Ripensis, there is a level of
violent destruction, which can be dated from the end of the coin rows or from incendiary layers
for the period between 378-395.

III.3.b. The consequences of the attacks in the texts

Any information about the troops disappears. L. Váradi (Váradi 1969, 371) believes that the
duces disappeared from Pannonia at this time. Due to Amm. Marc. (XXI 10. 9) and Zos. Hist.
Nov. (IV 34) it is assumed that the destruction in Pannonia II was due to the penetration of the
group led by Alatheux and Saphrax. There are no more inscriptions, stamps or other information
to prove the presence of the troops in the last phase. It is assumed that they were withdrawn to
the northern border of Thrace (Iord. Get. 132; cf. Stahlknecht 1966, 77-78; Lot 1927, 312). In
any case, the army's arsenal was left to the Goths.
III.3.c. The occupation of the provinces
Combining the information from the three sources, it can be said that for twenty years (after 378)
the Roman Empire lost control of the border in the area of the province of Dacia Ripensis (and
the province of Moesia Prima, and west and east of it). The barbarians destroy and burn down
fortifications, destroy and burn down internal settlements; among the provinces mentioned in the
sources, Dacia Ripensis and Moesia Inferior appear steadily (Iord. Get. 132); Far-reaching

31
disruption of traffic, far-reaching disruption of the economy, abandonment of the border areas by
the state officials, by the bishops, looting, robbery and rape in different intensity depending on
the place and time (Claudianus, in Eutropium II 214-218), all of this clearly shows that Collapse
of the Roman defense.

III.3.d. Historical remarks

Which position the Notitia Dignitatum reflects (Polaschek 1937, 1081-1097) is more difficult to
determine because it is archaeologically not confirmed by anything.

- p. 78
ATTACHMENTS
C.1. Directory of literary texts on the attacks of the barbarians
Ambros. De fide II 16 (after 378) The Gothic attacks make themselves felt “per Ripensem
Daciam et Mysiam omnemque Valeriam Pannoniorum”.
Ambros. De obitu Valentiniani 4 = PL 16, 1416 “audite omnes populi et videte dolorem meum.
Virgines meae et iuvenes mei abierunt in captivitate ... Nos adhuc muros Alpium addere
parabamus. Valentiniani gratia non expectavi Alpium vallum, fluente omnium ageres nivium sed
Alpes et fluvios supergressos, moro nos sui imperii protexit ”.
Ambros. Ep. 12 (381) The bishops assembled in the Council of Aquileia complain to Theodosius
II that because of the barbarians they could not send representatives for the Council of
Constantinople.
Ambros. Ep. 15 (381) is surprised that he received the news of the death of Bishop Acholius of
Thessalonica "clauso tunc mari, occupatis terrarum barbarico infestatione regionibus".
Ambros. Ep. 15 (381): “The bishop of Mursa put on Gothic clothes at the time when the city was
occupied by Goths”.
Ambros. Ep. 40 = P. 16, 1156; “Ego (christus) cum periculum summum esset ne Alpes infide
barbarorum penetrarent consilia, intra ipsa Alpium vallum victoriam tibi contuli, ut sine damno
vinceres” (letter to Theodosius II).
Ambros. Ep. 59 (392) writes to Bishop Servius of Neapolis about a serious threat in Pannonia
because the Alpine passes were blocked in a hurry.
Amm. Marc. XXI 16. 3 (after 378) “Hunii Halanisque permixti belicosis et fortibus ... agros vero
fertiles latestentos et longe ad extremum vastavere penuriam cultoribus caesis aut captis ...
itineris lentis missentes cuncta populationibus et incendiia nullo retinentes pergebant”.
Amm. Marc. XXXI 16. 7 “exinde digressi sunt efusori per arctoas provincias, quae peragravere
licenter ad usque radices Alpium Iuliarum quae venetas appelabant antiquitas”.
Aur.Vict. Caes. 47. 3 (378) "Thraciam Daciamque tamquam genitales terras possidentibus Gothi
Taiphalisque atque omni pernicie atrocioribus Hunis et Alanis extremum periculum in nomini
romano accito from Hispania Theodosio ... imperium contulit".
32
Claud. In Eutrop. II 214-218 (after 378) “Vastastor Achaiae / Getis et Epirum nuper populatim
inultam / praesidet Illyrico iam quos obsedit amicus / ingreditus muros, illis responsa daturus /
quorum coniugibus potitur, patosque peremit / sic hostes punire solent”.

- p. 79
Claud. In Ruf.. II 45 “Nam plaga Pannoniae miserardaque moenis Thracium / arvaque Mysorum
iam nulli flebile damnum / sed cursus solemnis erat campusque furori / expositus sensumque
malis detraxerat usus”.
Claud. Paneg. de quarto cons. Hon. Aug. VIII 50-55 (after 378) “Nam cum barbaris penitus
comote gementem / inruerunt Rhodope et mixto turbine gentes / Iam deserto suas in nos
transfunderat arctoas / Danuvii totae vomerent com proelis ripae / Cum Geticae ingens
premeretur Mysia palustris / Flavaque bistoni os oper campos / omnibus adflictis et vel
latentibus ictu / vel prope casuris ”.
Consularia Constatinopolitana (CCCCXVI, 490-494) “et pugna magna fuit cum Romanis et
Gothis miliario XII from Hadrianopoli the Vid. Aug. Ex ea the Valens Augustus musquam
apparuit et toto anno per diocesim Thraciam et Scythie et Moesiae Gothi habitaverunt et eas
praedaverunt. Deinde usque ad portas urbis Constantinapolitanae venerunt.
Eunap. Frg. 50 (after 378 = during the reign of Theodosius) “The barbarians gradually
devastated Thrace; the infinite enemy tribes had crossed the river from the beginning and more
and more of them continued to cross it without anyone stopping them ”.
Greg. Naz. Or. XII (after 378) “The earth is covered with corpses and red with blood” (south of
the Danube).
Jerome. Coment.in Soph. CI = PL 25, 1340; (a. 392) “et vastatis urbibus, hominisque interfectis,
solitudinem et raritatem bestiarum quoque fiere et volatilim pescicumque testis Illyricum, testis
Thraciam, testis in quo ortus sum, ubi praeter caelum et terrae”.
Jerome. Ep. LX 16 (a. 403) “Viginti et amplius anni sunt quod inter Constantinopolim et Alpes
Iulias quotidie Romanus sanguinis effunditur. Scythiam, Thraciam, Dardaniam, Daciam,
Thesaliam, Achaiam, Dalmatiam cunctasque Pannonias Gothus, Sarmata, Quadus, Alanus,
Huni, Vandali, Marcomani vastant, trahunt, rapiunt ”.
Jerome. Ep. LXVI = PL 22, 647 (a. 400) “compulsi summus frater Paulinus ad patriam mittere,
ut semirutas vilula que barbarorum effugerunt manus et parentem communium census venderet”.
Lord. Get. 132 (a. 376) “susceptoresque in partibus Moesiase Getas, quasi murum regni sui
contra citeras stattuit gentes”; 133 “Ipsi quoque dictum est Danuvio transmeantes Ripensem
Daciam, Mysiam Thraciamque permisio principi insederunt”.
Lord. Get. 137 (after 378) “Gothi iam non ut advense et peregrini, sed ut cives et domini
possesoribus imperare totasque partes septentrionales usque ad Danubium suo iure tenet”.
Lord. Get. 147 (a. 395) “per Pannoniem et per Sirmium dextraque latere quasi viris v / acuum
intravit (Alaricus) Italiam nulloque penitus to Ravenna”.

33
Lord. Rome. 313 (after 378) “superatisque Romanis (Gothi) in congressione funduntur in
Thracias”.

- p. 80
Lib. Or. XXIV 15 (a. 380) "We lost 25 peoples who were torn from us beyond the walls (the
borders) ... so the Scythians mocked us, who always trembled when they heard of our skill in
battle" .
Oros. Adv. Pag. VII 33. 11 (after 378) “deinde propter intolerabilem avaritiam Maximi ducis
fame et iniuria adacti in arma surgentis victo Valentis exercitu per Thraciam sesse, miscentes
simul omnis caedibus incendiis rapinisque fuderunt”.
Pan. Lat. XII (2) XI 4 (390-400) "perdidi infortunata Pannonia, lugeo funus Illyrici ... quidquid
atterit Gothus, quidquid rapit Hunnus, quidquid aufert Halanus, id olim desiderabat Arcadius".
Philostorg. Hist. Eccl. IX 17 (after 378) “So Valens went under and at the same time also the
largest and strongest part of the Roman Empire. The barbarians among Fritigers fearlessly
sacked Thrace ”.
Them. Or. XIV 181 (a. 379) “We are convinced that we will stop the Scythians in their course ...
and put out the fire that devours everything and that neither the Haemus nor the borders of the
Thracians and Illyrians stopped are difficult to cross even for a hiker ”.
Them. Or. XVI 206 (383) “The Thracians were sacked, the Illyrians were sacked, whole armies
were crushed as if they were shadows. Neither the inaccessible mountains, the inaccessible
waters, nor the roughness of the places stopped the invasions of the enemy ”.
Theoph. Conf. Chron. A. 5870 (a. 378) “The Goths reunited and invaded the land of the Romans,
where they devastated many provinces, Shytia, Macedonia and all of Hellas, about 20
provinces”; (after 378) "The Goths dared themselves because of their victories and devastated
the surroundings of Constantinople in a terrible way".
Veg. Epit. I 20. 16-23 "ad tartum urbium excidiae (Alaric).

C.II. The consequences of the attacks are reflected in the texts

378 Numerous other barbarian groups cross the Danube and penetrate into the empire; among
them those led by Alatheus and Saphrax (Amm.Marc. XXXI 4. 12-13).
380 The connections between Thessalonike and Italy are resumed (Ambros. Ep. 15).
381 Arbogast and Bauto - magistri militum in the west - were sent to Moesia I against the Goths
(Ambros. De fide 16).
386 Order to bring the mine officials who ran away from the barbarians back to work (CTh. I 32.
5).
386-387 Theodosius I returns to Constantinople after fighting in Thrace (Marcellinus Comes, a.
386; Zos. Hist. Nov. IV 48-49).

34
387 Iustina, mother of Valentinianus II, besieged in Thessalonica, asks Magnus Maximus for
help.

- p. 81
390 Alaric becomes magister militum in Moesia I and Dacia Ripensis
(Claud. In Eutrop. II 214-218; Claud. Bell. Goth. 535-539; Vetters 1950, 37).
391 Promotus, magister militum, falls in the fighting in Thrace (Jones 1964, 158).
392 The iazygian Sarmatians reach Italy (Claud. In Ruf. II 26, 31).
395-396 After the death of Theodosius I, the Goths resumed their attacks on Macedonia and
Thessaly, central Greece and the Peloponnese; they could only be stopped in this way by giving
them new territories; Alaric receives the title of a magister militum per Illyricum (Claud. In Ruf.
II 101-106).
407 Stilichon orders Alaric to occupy West Illyricum (Claud. In Eutrop. II 214-216; Claud. Bell.
Goth. 535-539).
408 The first defense measures (CTh. XI 17. 4); Order to praefectus praetorio Illyrici for the
restoration of the city walls (after the devastation of the Alaric), supplies, necessary transports;
complete commitment without privileges.
409 Theodosius II sends 4,000 soldiers to Dacia Ripensis to offer Honorius help (Jones 1964,
182; cf. Zos. Hist. Nov. IV 45.1).
C. III. The end of the Roman defense due to the coin finds in the border fortresses (vezi p.
81-83)

35