C. Lillington-Martin. Procopius. 2013.

WAR AND WARFARE
IN LATE ANTIQUITY
EDITED BY
ALEXANDER SARANTIS
and
NEIL CHRISTIE
LEIDEN • BOSTON
2013
© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8
CONTENTS
VOLUME 8.1
Acknowledgements ........................................................................................ ix
List of Contributors ........................................................................................ xi
Foreword ........................................................................................................... xvii
Alexander Sarantis and Neil Christie
Waging War in Late Antiquity .................................................................... 1

Alexander Sarantis
Bibliographic Essays
War in Late Antiquity: Secondary Works, Literary Sources and

Material Evidence ...................................................................................... 101
Conor Whately
Military Equipment and Weaponry: A Bibliographic Essay .............. 153

Alexander Sarantis
Tactics: A Bibliographic Essay .................................................................... 177

Alexander Sarantis
Organisation and Life in the Late Roman Military:

A Bibliographic Essay ............................................................................... 209
Conor Whately
Strategy, Diplomacy and Frontiers: A Bibliographic Essay ................ 239

Conor Whately
Fortifĳications in the West: A Bibliographic Essay ................................. 255

Alexander Sarantis with Neil Christie
Fortifĳications in Africa: A Bibliography Essay ....................................... 297

Alexander Sarantis

vi contents
Fortifĳications in the East: A Bibliographic Essay .................................. 317

Alexander Sarantis
VOLUME 8.2
Strategy and Intelligence
Information and War: Some Comments on Defensive Strategy

and Information in the Middle Byzantine Period
(ca. A.D. 660–1025) .................................................................................... 373
John Haldon
Fortifications and Siege Warfare
Fortifĳications and the Late Roman East: From Urban Walls

to Long Walls .............................................................................................. 397
James Crow
Siege Warfare and Counter-Siege Tactics in Late Antiquity

(ca. 250–640) ............................................................................................... 433
Michael Whitby
Weaponry and Equipment
Late Roman Military Equipment Culture ............................................... 463

J. C. N. Coulston
Barbarian Military Equipment and its Evolution in the Late

Roman and Great Migration Periods (3rd–5th c. A.D.) ................. 493
Michel Kazanski
Recreating the Late Roman Army ............................................................. 523

John Conyard

contents vii
Literary Sources and Topography
Reporting Battles and Understanding Campaigns in Procopius and

Agathias: Classicising Historians’ Use of Archived Documents as
Sources .......................................................................................................... 571
Ian Colvin
Procopius on the Struggle for Dara in 530 and Rome in 537–38:

Reconciling Texts and Landscapes ....................................................... 599
Christopher Lillington-Martin
Ammianus Marcellinus and the Nisibene Handover of A.D. 363 .... 631
Susannah Belcher
The West
Imperial Campaigns between Diocletian and Honorius,

A.D. 284–423: the Rhine Frontier and the Western Provinces .... 655
Hugh Elton
The Archaeology of War and the 5th c. ‘Invasions’ .............................. 683

Michael Kulikowski
Controlling the Pyrenees: a Macaque’s Burial from Late Antique

Iulia Libica (Llívia, La Cerdanya, Spain) ............................................. 703
Oriol Olesti, Jordi Guàrdia, Marta Maragall, Oriol Mercadal,
Jordi Galbany and Jordi Nadal
The Balkans
The Archaeology of War: Homeland Security in the South-West

Balkans (3rd–6th c. A.D.) ........................................................................ 735
John Wilkes
Military Encounters and Diplomatic Afffairs in the North Balkans

during the Reigns of Anastasius and Justinian ................................. 759
Alexander Sarantis

viii contents
Horsemen in Forts or Peasants in Villages? Remarks on the

Archaeology of Warfare in the 6th to 7th c. Balkans ..................... 809
Florin Curta
The East
Military Infrastructure in the Roman Provinces North and South

of the Armenian Taurus in Late Antiquity ........................................ 853
James Howard-Johnston
El-Lejjūn: Logistics and Localisation on Rome’s Eastern Frontier

in the 6th c. A.D. ........................................................................................ 893
Conor Whately
Civil War
Wars within the Frontiers: Archaeologies of Rebellion, Revolt and

Civil War ....................................................................................................... 927
Neil Christie
The Justinianic Reconquest of Italy: Imperial Campaigns and

Local Responses ......................................................................................... 969
Maria Kouroumali
Abstracts in French ........................................................................................ 1001
Index ................................................................................................................... 1009
Series Information .......................................................................................... 1085

PROCOPIUS ON THE STRUGGLE FOR DARA IN 530
AND ROME IN 537–38: RECONCILING TEXTS AND LANDSCAPES
Christopher Lillington-Martin
Abstract
This paper examines Procopius’ descriptions of Roman and Persian strate-

gies to control Dara in 530 and Roman and Gothic strategies to control Rome
in 537–38 by reconciling texts with the landscapes of the areas concerned,
drawing on satellite imagery, cartography and fĳield visits. The traditional
approach to this history has been to use written sources only, but, as will
be shown, these are subject to multiple interpretations. Study of the land-
scape provides a diffferent, complementary perspective, which is in some
ways more reliable, as the physical features have not changed too signifĳi-
cantly over the centuries, and modern technology has opened up new ways
of reading them. These two case studies strongly suggest that Procopius is
reliable when he is interpreted carefully, and this has implications for stud-
ies of the many other events for which he is the main source.
Introduction
This paper re-interprets Procopius’ descriptions of opposing Roman, Per-

sian and Gothic strategies to control Dara in A.D. 530 and Rome in 537–38,
especially those relating to the location of temporary army camp sites and
the identifĳication of river bridges.1 This is achieved by reconciling texts
with the landscapes of the areas concerned, drawing on satellite imag-
ery, cartography and fĳield visits. The traditional approach to the history
of this era, which has been studied by historians such as Bury, Dewing,
Evans, Greatrex and Lieu, Kaldellis, Kouroumali and Whitby, has been to
use textual sources only, but, as will be shown, these are subject to mul-
tiple interpretations. Study of the landscape provides a diffferent, comple-
mentary perspective which is in some ways more reliable, as the physical
features have not changed too signifĳicantly over the centuries, and mod-
1 Procopius of Ceasarea was an adviser to Justinian’s general Belisarius, ca. 505–65. He

was an eye-witness to many of the events discussed, and seems to have had access to reli-
able offfĳicial written documentation (but see now Colvin, in this volume).
A. Sarantis, N. Christie (edd.) War and Warfare in Late Antiquity: Current Perspectives
(Late Antique Archaeology 8.1–8.2 – 2010–11) (Leiden 2013), pp. 599–630

600 christopher lillington-martin
ern technology has opened up new possibilities of reading them. Bring-

ing the two approaches together, therefore, provides a powerful tool for
the study of the past. Thus, by reconsidering Procopius’ account in con-
junction with visits to the landscapes he describes near Dara and Rome,
it will be shown that we can gain additional insights, which help us to
re-interpret military strategy and events.
It seems particularly appropriate to analyse landscapes and recon-
cile them with the literary description provided by Procopius as, with
Homer in mind, he himself offfers reconciliation between landscape and
literature:
Tarracina; and very near that place is Mt. Circaeum, where they say Odys-
seus met Circe, though the story seems to me untrustworthy, for Homer
declares that the habitation of Circe was on an island.2 I am able to say . . .
Mt. Circaeum, extending . . . far into the sea . . . has every appearance of
being an island . . . for this reason, Homer perhaps called the place an island.
(Procop. Goth. 5.11.2).
So, we will reconsider the strategies which culminated in three engage-
ments: one in 530 on a site 20–24 km west-north-west of Nisibis (modern
Nusaybin, Turkey), just west of the 6th c. Roman-Persian frontier in north-
ern Mesopotamia (near Dara, modern Oğuz, Mardin, Turkey); and two in
537–38, between the River Anio and the Aurelian Walls at Rome. Military
fortifĳications were related to all three engagements. Belisarius used them
as bases from which to direct fĳield armies as part of his campaign strategy.
New interpretations will be presented which are entirely compatible with
Procopius’ account, which it will be suggested has been misinterpreted by
previous Anglophone historians. We will fĳirst consider how this straight-
forward approach has been used to argue for a more precise location of
the battlefĳield of Dara, 530, with a consequent re-interpretation of camp
locations. Then we will consider the identifĳication of an anonymous bridge
which has important implications for our understanding of the Gothic
army’s strategy regarding Rome, 537–38, led by King Wittigis.
2 Hom. Od. 10.135: ‘we came to the island of Aeaea, the home of the beautiful Circe’.

procopius on the struggle for dara and rome 601
The Background to the Roman-Persian Struggle for Dara in 530
The border fortress of Dara was constructed on the orders of the emperor
Anastasius in ca. 505. Procopius described the location and founding of
Dara thus:
The Emperor Anastasius, after concluding the treaty with Kavad, built a city
in a place called Dara, exceedingly strong and of real importance, bearing
the name of the emperor himself. Now this place is distant from the city of
Nisibis 100 stades lacking two, and from the boundary line which divides the
Romans from the Persians about 28. (Procop. Pers. 1.10.14).
Fig. 1 Cities and battle sites near the Roman-Persian frontier, A.D. 530–531.
(prepared with Carlos García).

Dara was planned as a forward post to launch potential Roman attacks

eastwards towards Nisibis and to deter, or delay, potential Persian attacks
north-westwards towards Diyarbakir (ancient Amida), and westwards
towards Antakya (ancient Antioch). The existence of this fortress so close
to the border provoked diplomatic crises and intermittent warfare, and
full-scale Persian invasions in 530, 540, 544 and 573. The battlefĳield and
settlements of Dara, Ambar (2.4 km south), Amuda in north-east Syria
(ancient Ammodios), and Nisibis, are set amongst fĳield systems and
divided by the modern Turkish-Syrian border. There are sightlines from
the hilltop of Ambar to Dara, Ammodios and Nisibis, but not between
Nisibis and Dara because of the low hills to the south-east of Dara. There
is a 2–3 km geographical gap between Ambar and the low hills.
The Battle of Dara, fought in the early summer of 530, was important
because the Romans won it decisively, and it confĳirmed the importance
of the fortress as well as the relative security of that part of the frontier
(fĳig. 1). Another victory at Satala (Sadak, Gümüşhane, Turkey), achieved
by Sittas later that summer, led to Roman advantages in Armenia. Fol-
lowing these setbacks, in 531, the Persians sought an alternative route of
attack, ending in a Persian tactical, if not strategic, victory at the Battle of
Callinicum (Ar Raqqah, Syria), and further Armenian campaigns.3 These
campaigns, combined with the death of the Persian King Kavad in 531,
contributed to conditions for an ‘Endless Peace’ treaty in 532, between
the Roman Emperor Justinian and the new Persian King Khusro I. This
allowed for the redeployment of Roman troops and resources, to be led
initially by Belisarius, for the conquest of Vandal Africa in 533–34, the fĳirst
stage of the partial re-conquest of the West by Justinian’s forces. So, along
with its precursors in 527–29 (see below), and the battles of Satala and
Callinicum referred to above, the Roman victory at the Battle of Dara was
especially signifĳicant because Justinian ordered the assembly of one of the
largest armies of his reign at this frontier to force a confrontation which
was to lead to a diplomatic resolution of the war with Persia.
Locating the Battlefield of Dara, A.D. 530
The accurate identifĳication of any battle site as precisely as evidence per-

mits is signifĳicant for military history, potential battlefĳield archaeology and
3 Greatrex and Lieu (2002) 91–96.

Fig. 2 Sites and topography surrounding the Battle of Dara, A.D. 530.
the interpretation of ancient sources. Here we will argue that the Battle
of Dara occurred approximately 2–3 km south of Dara, east of Ambar
(fĳig. 2), rather than immediately outside its walls, as proposed by Bury,
Evans, Greatrex, Haldon and Whitby.4 The line of reasoning by which this
location has been determined takes into account Procopius’ description
of the topography and his record of various distances to describe sites,
including the position of this battle site. These distances can be used
alongside Procop. Pers. 1.10.14 cited above, and other internal evidence, to
tentatively convert the distances Procopius provides into kilometres and
to better locate the battlefĳield within the landscape.
However, we should bear in mind that such calculations may not be as
precise as we would like. One of his measurements of distance is at Pro-
cop. Vand. 3.17.17, where Procopius refers ‘to Decimum, 70 stades distant
from Carthage’. This suggests, considering that ad Decimum was Latin for
‘10 mile post’ (15 km), that for Procopius, one mile equated to 7 stades.5
Classical authors such as Pliny and Strabo equated one Roman mile to
8 stades.6 Therefore, if we divide 28 stades by 7, Procopius would locate
4 Bury (1923) 82–83; Evans (2000) 117; Greatrex (1998) 172; Haldon (2001) 23–35; Lillington-
Martin (2007) 302; Whitby (1986) 758–59.
5 www.unitconversion.org.
6 Feissel (2002) 383.

Dara 4 Roman miles (6 km) from the boundary with Persia, and by divid-
ing 98 by seven, we can see Procopius placed it 14 Roman miles (or 21 km)
from Nisibis. So Procopius’ stades measure ca. 211 m. Turning now to mod-
ern measurement technology in the form of the Google Earth ruler, we
fĳind that Dara is actually 27 km from Nisibis, via the geographical gap
mentioned above. This anomaly may be a coincidence, but it might indi-
cate that Procopius meant that the frontier was 98 stades (21 km) west of
Nisibis and 28 stades (6 km) east of Dara.
We can now turn to the Persian invasion of Roman territory with the
intention of capturing Dara which culminated in the battle of 530. Proco-
pius states that a trench was constructed by the Romans as a fĳield defence
across the battlefĳield near a πὺλη (interpreted here as a geographical ‘gap’
rather than Dewing’s ‘gate’), to disrupt the Persian attack:
[S]uddenly, however, someone reported to Belisarius and Hermogenes that
the Persians were expected to invade the land of the Romans, being eager
to capture the city of Dara. And when they heard this, they prepared for the
battle as follows. Not far from the gap (πὺλης) which lies opposite the city
of Nisibis, about a stone’s throw away, they dug a deep trench with many
passages across it. Now this trench was not dug in a straight line, but in the
following manner. In the middle, there was a rather short portion straight,
and at either end of this, there were dug two cross trenches at right angles
to the fĳirst; and starting from the extremities of the two cross trenches, they
continued two straight trenches in the original direction to a very great dis-
tance. Not long afterwards, the Persians came with a great army, and all of
them made camp in a place called Ammodios, at a distance of 20 stades
from the city of Dara. (Procop. Pers. 1.13.12–14).
The translation of πὺλη as ‘gate’ by Dewing seems to have contributed to
several scholars locating the trench-line and battle within a few metres of
the fortress walls. Whitby proposed that the rock-cut moat may have been
part of the fĳield fortifĳication constructed prior to the battle, and Greatrex
suggested the trenches “used in the battle may subsequently have been
integrated into the defences of the city”, or that “Belisarius may have
adapted the trenches already under construction”; both scholars implied
that the fortifĳications of Dara were not complete after 25 years of build-
ing, which seems unlikely to have been the case.7 Dara’s defences prob-
ably would have been fĳinished before 530, given the attempts at building
an outlying fort in 528. Belisarius is likely to have ensured that Dara was
7 Dewing (1914–40) 105; Greatrex (1998) 171 and n.10 and 172; Haldon (2001); Mitchell
(2007) 134; and Whitby (1986) 761.

Fig. 3 The Battle of Dara: Roman and Persian army deployments.


completed before risking the movement of the workforce elsewhere.8 Such

a fort-building workforce was destroyed at Mindouos in 528 (fĳig. 2).9
Considering the space required for the deployment of the 25,000-strong
Roman army, including cavalry, and analysing the landscape, it seems
more likely that the trenches were constructed at a greater distance from
the fortifĳication wall. This would correlate the testimony of John Malalas
(ca. 490–575), who states that the Roman army was encamped outside
Dara.10 Furthermore, by interpreting πὺλη as gap (rather than ‘gate’), we
can convincingly reconcile text and landscape because there is a geo-
graphical gap ‘opposite the city of Nisibis’, whereas there is no gate in
the eastern wall of Dara facing Nisibis.11 Procopius may even have used
πὺλη to evoke the Battle of Thermopylae, 480 B.C.,12 in his readers’ minds,
where a huge Persian army attacked another Greek-speaking one which
was defending fĳield fortifĳications in a geographical gap. Regarding dis-
tances, Procopius uses the term ‘stone’s throw’ (οσον λὶθου βολὴν) at Pers.
1.13.13, and during the siege of Rome, at Goth. 5.22.13, where the distance
concerned was approximately 85 m.13 This suggests a trench located very
near the gap 2.4 km south of Dara.
At the Battle of Dara, Procopius relates Roman deployment and all tacti-
cal manoeuvres to the trench-line and a low hill on the eastern side of the
plain (fĳig. 3). According to Procopius, the Persians were out-manoeuvred
by Belisarius’ tactics, including an ambush (by the Herulians) from behind
low hills and internal flank attacks from the centre (made possible by the
trench-line being designed with many crossing points). These attacks were
executed by well-positioned and well-led cavalry units, including Hunnic
mercenaries and Belisarius’ reserve. The Roman victory depended on a
number of factors, such as the design of, and intelligent army deployment
along, the trench-line. Therefore, the most likely position of the Roman
battle line is approximately 2–3 km south of Dara and 4–6 km north of
Ammodios, in, or very near, the geographical gap, which is 2–3 km wide
between Ambar and the low hills to the east. This is where local farmers
informed me that they fĳind arrowheads in the fĳields, which they leave in
situ. One farmer drew a sketch of an arrowhead, but I did not see any.
8 Compare with Greatrex and Lieu (2002) 89.

9 Procop. Pers. 1.13.1–8.
10 Malalas 18.50.
11 Lillington-Martin (2007) 301–303.
12 Hdt. 7.138–239.
13 Lillington-Martin (2007) 302.

The gap at Ambar was ideal for the Roman army of 25,000 men to
prevent access to Dara by the invading Persian army of 50,000, as the
restricted space within the landscape neutralised the diffference in num-
bers by preventing a Persian envelopment of the Roman army wings
(although the Persians were able to double their lines and rotate troops to
manage fatigue). The study of this battle improves our understanding of
Roman military planning by providing details of deployment, and furthers
our understanding of battle descriptions within historical sources. In addi-
tion to Procopius perhaps alluding to the Battle of Thermopylae, he may
have intended his readers to recall the Battles of Salamis, 479 B.C.,14 where
a large Persian fleet advanced into narrow straits and was ambushed
and defeated by a smaller force, or Marathon, 490 B.C.,15 where a smaller
Greek army initially defended a position before attacking and routing a
larger Persian army. Procopius’ Dara battle description brings to mind all
three of Herodotus’ famous battle descriptions.16 So, by combining the
literary evidence provided by Procopius with satellite imagery and fĳield
visit experience, it is possible to gain a better understanding of the battle
of June/July 530.17
Locating the Persian and Roman Camp Sites, A.D. 530
In this section, we will analyse Procopius’ text, and consider measure-

ments of distances and relative positions of temporary fĳield fortifĳications
linked to Dara and Ammodios, ca. 8 km to the south, to re-evaluate Roman
and Persian campaign strategies. Satellite imagery and evidence from fĳield
visits will be discussed, with particular reference to topographic features
relating to the Battle of Dara. The relationship between textual and mate-
rial evidence can be problematic, especially when there are few literary
accounts to contrast with that of Procopius, whose evidence we are using
in a relatively literal way.
Procopius explains that the ‘. . . Romans . . . dug a deep trench . . . . [the]
Persians . . . made camp in . . . Ammodios, at a distance of twenty sta-
des . . .’.18 The location of the Persian camp in Ammodios can be calculated
14 Hdt. 8.40–96.
15 Hdt. 7.207–229.
16 Procop. Pers. 1.13.9–14.
17 Lillington-Martin (2007) 310.
18 Procop. Pers. 1.13.12–14.

in a similar way to calculation of the distance between Dara and Nisibis

in the previous section. Twenty stades would equate to 2.85 Roman miles
(4.2 km). Whereas the distance from Dara to Ammodios was calculated as
7.7 km by Dillemann in 1962,19 my research using the Google Earth ruler
suggests that Ammodios is little more than 4.2 km (20 stades) from the
geographical gap discussed above, which might indicate that Procopius
meant that the Persian camp was 20 stades south of the Roman trench
and battle-line, which would be consistent with the quote cited above.
The discrepancy in the distances of 20 stades (4.2 km) and 7.7 km has
attracted attempts at reconciliation, including Whitby’s idea that Procop-
ius’ stades need to be doubled to allow for the diffference, which is unnec-
essary.20 Another, unexplained, hypothesis by Evans reads: “A battle was
fought at Dara three and a half kilometres from the fortress”.21 Perhaps,
in this latter case, Procopius’ 20 stades were similarly calculated as 4.2 km
then deducted from Dillemann’s 7.7 km, resulting in a 3.5 km diffference,
and thereby placing the Persian camp at Ammodios 4.2 km away from
a battlefĳield 3.5 km south of Dara. However, this idea was not explained
and then withdrawn completely from Evans’ second edition of The Age
of Justinian (published in 2000). To overcome this problem without seek-
ing solutions in the geography of the region, we might conclude that
Procopius’ knowledge of distances is inaccurate, unreliable or, perhaps,
rhetorical.
Alternatively, we can consider the possibility that Procopius has been
misinterpreted by modern scholars and re-examine the landscape in light
of this. ‘[A]place called Ammodios’ is likely to refer to the nearest named
area to the Persian camp, where there was a good water supply. Further-
more, it would have been fundamentally sound military strategy for the
Persians to camp at least part of their 50,000-strong army on the north
side of Ammodios, to control the east-west road, as well as to separate
this Roman outpost from Dara. Such an army would have occupied an
extensive area: at 2 m2 per man, 50,000 men would occupy 10 ha just by
lying upon the ground. Hunter-gatherer campsites have been calculated
as occupying between 5.9 and 10.2 m² per person.22 The Persian army
had myriads of horses to cater for, so its camp at Ammodios is likely to
have extended across many hectares. This interpretation could situate the
19 Greatrex (1998) 170–71, n. 9.

20 Whitby and Whitby (1986) 136 n. 17.
21 Evans (1996) 117.
22 Hassan (1981) 67 citing Wiessner (1974).

Persian camp straddling an area up to a few hundred metres north of

the current Turkish-Syrian border (perhaps to within 20 stades of Dara).
The fact that Ammodios is little more than 20 stades and ‘a stone’s throw’
from the geographical gap area, with its Roman trenches and battle-line,
suggests that Procopius’ measurement may refer, in fact, to the distance
between the Roman trenches and Persian camp defences. Whichever way
we view these possibilities, Procopius’ numerical distances should, there-
fore, be considered generally accurate and reliable.
As noted above, the Roman army was encamped, according to Malalas,
outside Dara, although part of it may have been partially accommodated
within the fortress.23 However, Dara could not have accommodated 25,000
Roman soldiers and their horses because, whether the defences had been
completed or not, it was too small.24 The wall is only about 2.7 km in cir-
cumference.25 The plans of Dara’s walls may not be entirely accurate, as
the discovery of a possible Anastasian mosaic on the floor of a structure
outside the proposed circuit may indicate that the circuit wall of Dara was
slightly larger than previously thought, if the structure had been enclosed
by it.26 However, this would not signifĳicantly change the fact that “The
internal area is fairly small . . . and the area suitable for habitation limited
by the steep, rocky ground”.27 The main Roman camp outside the city
walls would have needed several hectares for the same reasons as the Per-
sian camp at Ammodios. In any case, deploying a large army in a relatively
small fortress would not have been a sensible strategy given the potential
problems with supply were the fortress to come under attack before the
army were deployed in the fĳield (Procopius delivers a set piece description
of the opposite, more sensible, strategy of 537–38, when Belisarius garri-
soned Rome, a large fortifĳied city, with a small army (see below)).
23 Malalas 18.50.
24 Millett (1990) 182–183, citing Hassan (1981) 66–67 on urban population estimates at
between 137 and 216 people per ha.
25 Ahunbay (1991); Croke and Crow (1983); Greatrex (1998); Haldon (2001); Sinclair
(1987–90); Whitby (1986); and Zanini (1990).
26 The discovery at Dara, by Professor Metin Ahunbay, of a 2.5 × 3.5 m floor mosaic,
with a 1 × 1 m Greek, possibly Anastasian, inscription (at 37°10’37.12”N, 40°56’53.13”E,
approximately 100 m west of the probable circuit wall), was announced via www.mardin
.gov on 22 October 2007. ΤΟΠΟΤΗ at the end of line three of the mosaic inscription sug-
gests τοποτηρητὴς, which has been associated with a military rank, appointed by a dux,
associated with building projects (SEG 14, no.1931, page 563) and someone concerned with
the building and restoration of a hostel (SEG 15, no. 1505, page 511).
27 Croke and Crow (1983) 150.

Indeed, the fact that only a fraction of the Roman army could have
fĳitted into the fortress supports the view that the Roman force was never
intended as a garrison. This suggests an offfensive Roman strategy at the
outset of the campaign, something which would not have been lost on the
Persians. Indeed, Malalas informs us that the Persians invaded when they
heard that the Roman army was at Dara.28 It is likely that the Romans had
intended to invade Persian territory, but were surprised by the fact that
the Persians had already gathered an even larger army in the same area,
which is a credit to Persian strategic organisation. Procopius tells us that
the Romans’ trench-line preparation only began when: ‘someone reported
to Belisarius and Hermogenes that the Persians were expected to invade
the land of the Romans’.29
The Roman political and military strategy is likely to have been to attack
Nisibis, or perhaps to oblige the Persians to negotiate. As it transpired, the
Persians’ intelligence allowed them to pre-empt the Romans, resulting in
their strategic decision to invade Roman territory, even though their tac-
tical assault towards Dara was unsuccessful. Had the Persian command-
ers heavily fortifĳied the Ammodios area they might, with patience, have
blockaded the Roman army of 25,000 at Dara, compelling it to abandon its
defensive fĳield tactics or negotiate. They certainly lingered in the area to
negotiate even after losing the battle, as Procopius records: ‘the Persians,
though defeated by Belisarius in the Battle at Dara, refused even so to
retire from there’.30 By placing this comment immediately after Pers. 1.15
(which narrates Roman success at repelling another Persian invasion, this
time further north, at the Battle of Satala, weeks after that of Dara and
Roman advantages gained in Armenia), and just before Kavad’s supposed
diplomatic speech to Rufĳinus,31 Procopius may well have been implying
that the defeated Persian army, which remained deployed in the area
between Nisibis and Dara to pin-down Roman forces there, continued to
be a serious force to be reckoned with for several months. Therefore, Per-
sian strategy was an offfensive one, and as part of it, Kavad had assembled
two large armies: to attack the Romans in both northern Mesopotamia
and Armenia.
To conclude this section, it seems highly likely that the Roman trenches,
serving the dual purpose of protecting the troops encamped outside Dara
28 Malalas 18.50.
29 Procop. Pers. 1.13.12.

and dictating the tactical course of the battle, lay approximately 2–3 km
south of Dara, while that the northern edge of the Persian camp was about
1–2 km from Ammodios. This leaves a distance of 20 stades between them
and confĳirms the accuracy of Procopius’ account, which will have been
written up from offfĳicial records, notes and/or eye-witness testimony sev-
eral years later. This accuracy is important when interpreting his descrip-
tions of other sites, conflicts and events, such as those in Italy.
The Roman-Gothic Struggle for Rome in A.D. 537–38
Having argued for Procopius’ accuracy in recording distances regarding

relative army positions and camps near Dara in 530, this section will now
compare his text with the landscape to appreciate the probable routes
taken by the Gothic army from Ravenna to Rome in 537, and its route of
retreat in 538.
These movements led to the fĳirst and last Roman-Gothic encounters
at Rome, 537–38, which were fought near an anonymous river bridge.32
The identifĳication of this bridge has strategic implications regarding
the route of the Goths’ march, and, therefore, the strategies employed
by both Romans and Goths. Wittigis, the king of the Goths and Italians
(A.D. 536–40), approached Rome in 537 via an anonymous river bridge in
order to lay siege to the Roman army under Belisarius. There has been con-
siderable disagreement amongst scholars about the identifĳication of this
bridge, but this paper maintains that Procopius’ account needs to be more
carefully interpreted. It will compare his description of the initial stages,
and fĳinal lifting, of the siege of Rome, with topographic evidence obtained
by means of satellite imagery, fĳield-visits and cartographic materials.33 The
results challenge conventional interpretations made by 19th and 20th c.
Anglophone scholars, such as Mahon, Hodgkin and Dewing, who, follow-
ing Gibbon, have suggested that the anonymous bridge was the Milvian
crossing of the Tiber.34 Their interpretation has been repeated by 21st c.
scholars, such as Kaldellis and Kouroumali.35 My results instead support
Gregorovius’ suggestion that this bridge was in fact the Salarian, which
32 Procop. Goth. 5.17–18 and 6.10.

33 Carta Topografĳica dei Dintorni di Roma and Google Earth.
34 Mahon (1848) 193–94; Hodgkin (1896) 118–20, 194 and 250; and Dewing (1914–40) 169.
35 Kaldellis (2004) 33; Kouroumali (2006) 187.

Fig. 4 Italy, A.D. 537–38. (prepared with Carlos García).

crosses the Anio, and it will be shown that this has implications for how
we understand the strategies adopted by the Gothic army.
Procopius was not the only eye-witness account we can draw upon.
Another textual source, the Liber Pontifĳicalis, comments on some of these
events leading to the Siege of Rome in 537–38, which led directly to the
deposition of Pope Silverius. Davis notes “an incompetent chronological
join between” the two parts of the biography of Silverius, and that the
“account of that siege given in the fĳirst part suggests that the author had
not yet witnessed the two worse sieges Rome was to endure in the next
decade”.36 As the next siege occurred in 546, the author of the fĳirst part of
Silverius’ biography was a contemporary or eye-witness to the events and
“it was produced no later than the 540s”.37
Eye-witnesses’ texts such as Procopius and Liber Pontifĳicalis are obvi-
ously of immense value in historical research, and may be expected to
provide more accurate information than accounts which have acquired
their information second-hand. However, even fĳirst-hand narratives are
hard to reconcile with the landscape because they omit certain details—
in this case, the precise identifĳication of rivers and bridges, the roads
between them, and the gates (in the Aurelian Walls) of Rome. This not
only leads us to embark upon further research in order to reach a logical
reconciliation between topography and text, but raises questions about
the objectives and research methods of the primary source. It may be
the case that the details of interest to us were un-recorded, forgotten, or
intentionally left out for the sake of simplicity (in Procopius’ case, this
would aid communication to an eastern audience, far from the theatre
of military operations, who expected to be entertained with drama and
heroism rather than be bored with specifĳic geographic details).
The historical background to our discussion is as follows (fĳig. 4). In
December 536, Belisarius led his army, fresh from sacking Naples, into Rome
through the Asinarian Gate. Procopius dates this to the ‘ninth day of the last
month’, while the Lib. Pontif. by Silverius assigns it to ‘10 December’.38 At
the same time, the 4,000-strong Gothic garrison left Rome (leaving behind
their commander Leuderis) via the Flaminian Gate, and headed north up
the Via Flaminia. Belisarius sent Leuderis to Constantinople, but we hear
no more of him.39 The leaderless Gothic army will have crossed the Tiber
36 Davis (2000) xliii.

37 Davis (2000) xiii and xlvi.
38 Procop. Goth. 5.14.14; Lib. Pont. 60.4.
39 Procop. Goth. 5.14.13–15, 24.1.

at the Milvian Bridge.40 His comment that the Goths had left along the Via
Flaminia may, in part, have led some scholars to make assumptions about
their return route under King Wittigis in February or March 537.
Belisarius began preparations to withstand a siege by strengthening the
fortifĳications of the Aurelian Walls and bringing in provisions.41 Procop-
ius records that, at this time, a Goth named Pitzas came to Belisarius to
surrender ‘half of . . . Samnium, which lies on the sea, as far as the river’
(Biferno or Sangro), and was given ‘a small number of soldiers to help him
guard that territory.42 In any case, ‘the Goths who were settled on the other
side of the river were neither willing to follow Pitzas nor to be subjects of
the emperor’.43 These Goths no doubt felt safer from Roman attack since
they were settled further north and away from the coast, and there were
doubtless links between this area and the important neighbouring Gothic
settlement region ‘concentrated in Picenum’ to the north, as has been
shown by Heather.44 Belisarius was now in control of the southern half of
Italy, from Rome, around the coast, to southern Samnium, as well as the
Dalmatian coast.45 In December 536, he sent his generals Constantinus
and Bessas north to occupy numerous strategic strongholds in Tuscany:
Perusia, Spoletium, Narni ‘and certain other towns’ (fĳig. 4).46
Wittigis responded by sending an army from Ravenna to Perusia, com-
manded by Unilas and Pissas, which Constantinus defeated completely.
Both Gothic commanders were captured and sent to Belisarius in Rome,
although they are not mentioned again.47 Their capture may have been
contrived, and perhaps in light of this and previous surrenders by Gothic
commanders (Sinderith surrendered at Syracuse in 535, and Ebrimuth
at Reggio di Calabria, Leuderis at Rome and Pitzas followed in 536), no
further Gothic army was sent against the Romans until Wittigis led one
himself.48 So Procopius records that by December 536, several Gothic
leaders either changed sides, or fell into Roman hands. This suggests a less
than loyal attitude towards the Gothic kingship. When Wittigis marched
his army southwards to Rome in early 537, he doubtless collected troops
40 Procop. Goth. 5.14.14; Bury (1923) 179–80.

41 Procop. Goth. 5.14.15; Lib. Pont. 60.4.
42 Procop. Goth. 5.15.1–2; Dewing (1914–40) 148–49, n. 1.
43 Procop. Goth. 5.15.2.
44 Heather (1996) 238, Figure 8.1.
48 Jord. Get. 60.308–12.

Fig. 5 Spoletium to Rome via Sabine country (Interamna Nahars and Reate) and
Rome to Narni. (prepared with Carlos García).
from his subjects in Picenum along the northern section of the Via
Flaminia towards the Apennines, and could have been joined by others
from northern Samnium at Reate, by way of the Via Salaria, Via Caecilia
and Via Claudia Nova (fĳig. 5).49
When Wittigis marched his army southwards, Belisarius recalled the
cavalry commanded by Constantinus and Bessas from Perusia, Spoletium
and Narni, where they had established control, and posted garrisons to
obstruct one of Wittigis’ potential routes to Rome. Before leaving, how-
ever, Bessas engaged and routed Wittigis’ advance guard before being
overpowered by numbers, after which he retired fĳirst to Narni and then
to Rome (foreshadowing Belisarius’ fĳirst conflict outside Rome soon after-
wards). Wittigis had ‘made no attempt to capture Perusia and Spoletium,’
nor did he ‘attempt anything’ at Narni.50 This was because, of the ‘two
roads leading [to Narni, one] . . . is very narrow [with] precipitous rocks,
49 Talbert and Bagnall (2000) Italia, Map 42, Arretium-Asculum.

50 Procop. Goth. 5.17.1–9.

while the other cannot be reached except by way of the [Augustan] bridge
(τὴν γὲϕυραν).’51
[S]o Wittigis, not . . . to have his time wasted there [Narni area], departed
thence with all speed and went with the whole army against Rome, making
the journey through Sabine territory (διὰ Σαβὶνων τὴν πορεὶαν ποιοὺµενος).
And when he drew near to Rome, and was not more than 14 stades away
from it, he came upon a bridge over the Tiber River (Τιβὲριδος τοὺ ποταµοὺ
γεϕὺρα). (Procop. Goth. 5.17.12–13).
Procopius leads us to believe that Belisarius’ strategy, of seizing and garri-
soning Narni, Perusia and Spoletium, culminating in the conflict at Narni,
caused Wittigis to leave the Via Flaminia east of the Tiber at Interamna
Nahars (modern Terni), 12 km east of Narni, and ‘journey through Sabine
territory’, leading his army along the Via Curia until it joined the Via
Salaria at Reate (fĳig. 5). This slowed Wittigis down and gave Belisarius
more time to provision Rome and prepare its defences. Alternatively, Wit-
tigis may have planned to use this itinerary so as to collect troops from
Picenum and northern Samnium en route, and thence move south-west
to Rome.52 This route had been used by Alaric’s Goths, who laid siege
to Rome in August 410.53 Therefore, Procopius may have presented what
was a prudent Gothic strategy of advancing through Sabine country to
collect troops, and thus to arrive at the more vulnerable north-east side
of Rome, in such a way as to denigrate Wittigis’ and improve Belisarius’
reputations.
Although the Τιδὲριδος τοὺ ποταµοὺ γεϕὺρα is not named by Procop-
ius, most scholars (as mentioned above in n.34 and n.35) have taken it to
signify the Milvian Bridge (crossed by the Via Flaminia).54 This is based
largely upon the fact that the Lib. Pontif. states that Wittigis ‘collected an
enormous army of Goths and returned to attack Rome on 21 February’ 537
and ‘pitched camp by the Milvian Bridge’.55 Rubin discusses whether 21
February was the date on which Wittigis left Ravenna or arrived at Rome.
The Lib. Pontif. date and bridge-name appear to have convinced many
51 Procop. Goth. 5.17.10–11.

52 Talbert and Bagnall (2000) Italia, Map 42, Arretium-Asculum.
53 Heather (2005) 227–29; Lançon (2000) 36–39.
54 Hodgkin (1896) 118, n. 2.
55 Lib. Pont. 60.1.4, trans. Davis (2000) 53; Kouroumali (2006) 187, n. 56; Rubin (1995)
105 n. 286.

Fig. 6 Via Flaminia, Via Curia, Reate and the Via Salaria to Rome.
scholars as to the time and place of Wittigis’ arrival, including Dewing,

whose translation refers to both.56
However, there is good reason to doubt whether the Milvian Bridge
could have formed any part of Wittigis’ route, given that his march down
the Via Flaminia was blocked by Narni, and that he made his ‘journey
56 Dewing (1914–40) 169, n. 1 and margin note, but see Veh (1966) 129–33 with notes
on 1028.

through Sabine territory’ (fĳigs. 5 and 6). Hodgkin considers, but rejects,
the suggestion by Gregorovius that Wittigis did not cross the river via the
Milvian Bridge.57 He prefers “to sacrifĳice the words διὰ Σαβὶνων rather than
the words Τιβὲριδος τοὺ ποταµοὺ γεϕὺρα”, which highlights the geographi-
cal incompatibility of the two phrases. Kouroumali does not deal with this
incompatibility and states that: “Wittigis, . . . chose a route through Sabine
territory . . . Belisarius had established a guard in a tower on the Milvian
Bridge over the river Tiber as a means of checking the descending Gothic
army”.58 This is geographically impossible because the Sabine route does
not lead to the Milvian Bridge. Martindale mentions the River Anio with-
out discussing any bridge at all, and refers to Bury,59 who agreed with
Hodgkin in 1889, but had changed his mind by 1923. His note is worth
quoting in full because it explains the controversy well:
The usual view has been that the Goths advanced by the Via Flaminia
(regaining it somewhere presumably between Narnia and Ad Tiberim, now
Magliano, where there was a bridge), and that the bridge where Belisarius
placed the garrison was the Pons Milvius, now Ponte Molle, two miles from
Rome. This view was held by Gibbon and maintained by Hodgkin. But it is
certainly erroneous and inconsistent with the story. If the fĳighting had been
at the Milvian Bridge, the Roman fugitives would have returned to the Porta
Flaminia, not to the Porta Salaria. The cause of the error is that Procopius
does not name the bridge, but calls it simply Τιβὲριδος τοὺ ποταµοὺ γεϕὺρα,
‘a bridge of the Tiber’. Hence, as the Milvian was the only bridge which
spanned the Tiber north of the city, it was naturally supposed to be meant.
But Τιβὲρις is ambiguous in Procopius; it means (1) the Tiber, (2) the Anio.
That it means the Anio in this passage is shown by the statement in the
context that there are bridges over the river in other places (πολλαχὸσε τοὺ
ποταµου), meaning, of course, in the neighbourhood of Rome. This is not
true of the Tiber, which had only the one bridge outside the city; but it is
true of the Anio, which is crossed, near Rome, by the Via Nomentana and the
Via Tiburtina, as well as by the Via Salaria. In two other passages (Procop.
Goth. 7.10.23, and 24.31), Τιβὲρις clearly means the Anio. This was the view
of Gregorovius, Rome in the Middle Ages, 1.372; is accepted by Hartmann,
Gesch. Italiens, 1.295, n. 19; and has been defended in a special monograph
by L. Fink, Das Verhältnis der Aniobrücken zur Mulvischen Brücke in Prokops
Gothenkrieg, 1907. Procopius knew the localities, and the ambiguous use of
Τιβὲρις cannot be due to ignorance. The explanation may be found in the
modern name of the Anio, Teverone, and the use in Procopius be taken to
57 Hodgkin (1885) vol. 4, 134, n.1 refers to Gregorovius, Geschichte der Stadt Rom (1859),
i.349, n. 1.
58 Kouroumali (2006) 187, n. 56.
59 Martindale (1992) IIIA 197 cites Bury (1923) II, 182.

show that the old name had passed out of common speech before his time.
(Bury (1923) II, 182, n. 1.).
This idea has Procopius keep Wittigis on the eastern side of the Tiber
(from Narni) as proposed by Gregorovius (fĳig. 6).60 Hodgkin accuses
Gregorovius of suggesting that Procopius “here as elsewhere confused the
Tiber with the Anio”, but he fails to specify any other passages where Pro-
copius undoubtedly uses the name Tiber when he means the Anio. The
Tiber and Anio rivers are recorded by 19th c. cartography as the Tevere
and Teverone, so it would have been understandable if they were equated
or even confused if their names were equally similar in Late Antiquity.61
Furthermore, the Salarian Bridge crosses the Anio less than 1 km from its
confluence with the Tiber, so Procopius could easily have equated the riv-
ers or merely simplifĳied his description for his readers.
Alternatively, the confusion may have been caused by Procopius teas-
ing his readers. On 28 October 312, the Emperor Maxentius died by drown-
ing at the Milvian Bridge after losing a battle north of it to Constantine
the Great. During the skirmish at the anonymous bridge in 537, Procopius
claims that 1,000 Goths fell and ‘many of the noblest’ of Belisarius’ house-
hold were slain, but he only names one: a certain ‘Maxentius’.62 Perhaps
Procopius made this nonchalant comment for his readers to spot, alluding
to the Battle of the Milvian Bridge of 312, and perhaps this is why he does
not name the bridge.
At this stage, Belisarius’ plan seems to have been to further delay the
Goths from advancing southwards on the east of the Tiber, north of the
anonymous bridge (as his garrisoning of Tuscany had prevented Gothic
movements along the Via Flaminia on the west of the Tiber, south-west of
Narni). This would have been intended to delay the Goths or force them
to re-route their approach to Rome, perhaps from further east. Belisarius
would thereby have gained more time to bring in additional supplies to
Rome and harass the Goths from the safety of a fortifĳied camp at a river
crossing with a fortifĳied bridge. The cavalry troops which would have
been based in the intended camp would have operated to the north of the
Salarian Bridge, and later, as necessary, withdrawn to Rome just as Con-
stantinus and Bessas had withdrawn from Perusia, Spoletium and Narni.
60 Hodgkin (1896) 118, n. 2.

61 Carta Topografĳica dei Dintorni di Roma (1884).

Landscape evidence gained from satellite imagery (confĳirmed by my

fĳield visit), combined with detailed analysis of internal textual data from
Procopius, gives good reason for re-examining the situation, and suggests
the validity of Gregorovius’ proposal, which has created an opposing
convention amongst German scholars, who accept the Via Salaria route
for Wittigis’ march south.63 Belisarius had taken command of the Salar-
ian and Pincian gates and stationed Constantinus at the Flaminian and
Bessas at the Praenestine gates on their return to Rome (fĳig. 7). At the
anonymous bridge (γεϕὺρα), Procopius tells us that ‘Belisarius had built a
tower (πὺργον), furnished it with gates (πὺλαι) and a guard of soldiers, not
because it was the only point at which the “Tiber” could be crossed by the
enemy,’ but to delay the advancing Goths.64 When the Goths reached
Belisarius’ fortifĳied and garrisoned anonymous bridge, Procopius continues:
. . . the Goths bivouacked there that day, . . . supposing that they would . . .
storm the tower on the following day; but 22 deserters came to them, men
who were barbarians by race but Roman soldiers, from the cavalry troop
commanded by Innocentius. (Procop. Goth. 5.17.16–17).
From the Aurelian Wall to the Milvian Bridge, the Via Flaminia is straight
and leads northwards over the River Tiber (the landscape is very flat and
a tramline now operates along this section of the route). We cannot fully
appreciate the lines of sight northwards today because the road is tree-
lined and in a residential area. However, its straightness to the Milvian
Bridge, and the flatness of the terrain to the bridge and just beyond it,
can be appreciated to a sufffĳicient extent on the ground, on maps, and
via satellite imagery (fĳig. 7). This suggests that only Constantinus, from
his position at the Flaminian Gate, had a clear view towards the Milvian
Bridge, 2.7 km along the Via Flaminia. Belisarius’ view, towards that bridge
from the Pincian-Salarian Gate section of the wall, was obstructed by the
hilly landscape, as was his view to the Salarian Bridge (which crosses the
Anio, where it is approximately 30 m wide).
When the Goths took the anonymous bridge, Constantinus commanded
a position at the Flaminian gate, and Belisarius commanded a position at
the Salarian and Pincian Gates, prior to Belisarius’ all-day skirmish near
the anonymous bridge.65 So Belisarius could not have seen either bridge,
63 See n. 55–57.
65 Procop. Goth. 5.18.1–29.

and he could not have known of the capture of one of them by the Goths
unless a message had reached him, which it evidently had not, given his
subsequent actions.
It is apparent that the Goths had not encamped north of the Milvian
Bridge on their arrival near Rome because had they done so, Constantinus
would have seen their camp fĳires overnight from the Flaminian Gate, or
would have been able to easily receive a signal or message from a tower
had there been one on the Milvian Bridge (fĳig. 7). Furthermore, Procopius
mentions the Goths building a seventh camp there (below) as an addition.
Such information could have been conveyed to Belisarius before dawn.
However, clearly no such information reached him and:
it occurred to Belisarius to establish a camp near the Tiber River, . . . (to)
hinder . . . the crossing of the enemy . . . But all the soldiers who, as has
been stated, were keeping guard at the bridge, . . ., abandoned by night the
tower . . . (as) they could not enter Rome, they marched toward Campania.
(Procop. Goth. 5.17.18–19).
Therefore, the Goths must have been encamped north of another bridge
which is out of sight of the Aurelian Walls. This concords with the sug-
gestion that it was the Salarian Bridge, as this is completely out of sight
of the city walls, from which it is obstructed by the landscape, which rises
gradually from the Salarian Gate for about 2.5 km north-eastwards along
the Via Salaria, as far as the Cemetery of Priscilla (avoiding the Villa Ada
hilly area), then drops more steeply northwards out of sight of Rome for
1 km to the Salarian Bridge (fĳig. 7). A Roman camp would have been
advantageous here.66 This strategic point commands the view northwards.
Indeed, today there are remains of an ancient tower, the Torre Salaria, 200
m north of the Salarian Bridge. In addition, those bridge guards who did
not desert to the Goths but ‘marched toward Campania’ would, to have
escaped out of sight of Rome, have followed the Anio east of Rome before
heading south. This suggests that, in this quarter at least, the morale and
communications system of Belisarius’ army was buckling under the stress
caused by the swift march and size of Wittigis’ army, leading to Roman
desertions at a critical time and in a crucial place.67
66 See fĳig. 7; Talbert and Bagnall (2000) Italia, Map 43, Latium Vetus; and Lib. Pont.
trans. Davis (2000) Map I.
67 Liebeschuetz (1996) 230–39.

All of Procopius’ next chapter, 5.18, is devoted to the skirmish, which

lasted the whole of the following day: ‘the Goths destroyed the gates of
the tower with no trouble and made the crossing, since no one tried to
oppose them’. If this crossing had taken place at the Milvian Bridge, it
could not have proceeded unnoticed because it would have been in full
view of Constantinus’ men at the Flaminian Gate, even though it was pos-
sible that the visibility could have been reduced on a February or March
morning. Even if Procopius used this episode to glorify Belisarius in what
follows, this would not afffect the fundamental tenets of this hypoth-
esis. So, without any visual information or messages from the ‘soldiers
who . . . abandoned . . . the tower’ on the anonymous bridge,
Belisarius, who had not . . . learned what had happened to the garrison, was
bringing up 1,000 horsemen to the bridge over the river, in order to look over
the ground and decide where it would be best for his forces to make camp.
But when they had come rather close, they met the enemy already across
the river, and not at all willingly they engaged with some of them. (Procop.
Goth. 5.18–2–3).
Thus, whilst Belisarius could have known from his control of Narni that
Wittigis was not marching southwards along the Via Flaminia, he did not
know how quickly Wittigis had marched southwards east of the Tiber. He
did not believe it was yet time to prepare to receive an immediate attack
on Rome. This corresponds with the idea that the landscape did not per-
mit a clear view from Belisarius’ starting point, the Salarian or Pincian
Gate, to his destination, the anonymous towered bridge. Had Constanti-
nus seen the Goths crossing at the Milvian Bridge and an engagement
between them and Belisarius, he would surely have reacted by either lead-
ing reinforcements there or preparing to receive Belisarius’ men. There-
fore, it is reasonable to conclude that the Goths did not cross the Tiber at
the Milvian Bridge, but made their crossing further east, probably across
the Anio near the confluence with the Tiber, for which the only likely
candidate is the Salarian Bridge.
This conclusion would correlate Procopius’ comment that, after the
skirmish between Bessas and the Gothic vanguard in the Narni area, Wit-
tigis marched his army through ‘Sabine territory’. This was either because
the Roman garrison at Narni prevented his army’s progress, or because he
wanted to rendezvous with troops en route from Picenum and Samnium,
perhaps in the vicinity of Reate. This route conveyed the Goths, partly by
way of the Via Salaria, to the north-east side of Rome, where, once they
crossed the Anio, via the anonymous bridge, they:

made 6 fortifĳied camps from which they harassed the portion of the wall
containing 5 gates, from the Flaminian as far as the Praenestine Gate; all
these camps were made on the left bank of the Tiber River. (Procop. Goth.
5.19.2.).
Procopius is very clear that these 6 camps were sited east of the Tiber,
south of the Anio and north and east of the Aurelian Walls. Maps confĳirm
Fig. 7 Main gates, roads and bridges of Rome. (prepared with Carlos García).

that this area is reached from Ravenna by crossing the Anio after travers-
ing Sabine territory (fĳigs. 4 to 7 and n.55 and n.66 above).
To return to Procopius’ description of the skirmish at the anonymous
bridge: ‘And the battle was carried on by horsemen on both sides . . . the
Romans turned the enemy to flight, . . . the Gothic infantry . . . forced them
back, . . . cavalry reinforced the Goths, . . . the Romans fled at top speed
until they reached a certain hill, which they climbed, and there held their
position.’68 Such hills are discernible via satellite and by fĳield visit, and
fĳill the landscape between the rivers and walls, including along the Via
Salaria, especially where it drops north from the Cemetery of Priscilla (fĳig. 7).
Then, ‘the enemy’s horsemen were upon them directly, and a second cav-
alry battle took place . . . the Romans escaped, and arrived at the fortifĳi-
cations of Rome . . . the barbarians in pursuit . . . (to) the Salarian Gate.’69
Procopius’ statement that the Goths pursued Belisarius to the Salarian,
and not to the Flaminian, Gate favours the hypothesis presented here.
Although Hodgkin comments on the possibility that Belisarius arrived
at the nearby Pincian Gate, which is linked to the Salarian Bridge by the
Via Salaria, we can be certain, based on the evidence above, that it was
not the Flaminian Gate.70 ‘Those inside the fortifĳications . . . were without
a general . . . and the battle which had begun early in the morning did not
end until night.’71 It is reasonable to conclude that Belisarius returned to
the gate he had left that morning along the same road. Had his destina-
tion been the Milvian Bridge, he would surely have left from the Flamin-
ian Gate, which Constantinus commanded, seen the Goths crossing, and
then retreated along the Via Flaminia, rather than use either the Salarian
or Pincian Gates. Furthermore, when Procopius does specifĳically mention
the Milvian Bridge, it is two chapters after the skirmish at the anonymous
bridge, and in the context of Gothic camp building, where, he states:
the barbarians feared lest their enemy, by destroying the bridge which bears
the name of Milvius, should render inaccessible to them all the land on the
right bank of the river as far as the sea . . . so they fĳixed a seventh camp across
the Tiber in the Plain of Nero, in order that the bridge might be between
their two armies. (Procop. Goth. 5.19.3).
68 Procop. Goth. 5.18.3–17.

69 Procop. Goth. 5.18.18–19.
70 Hodgkin (1896) 122, n. 2; and Dewing (1968) 176, n. 2.
71 Procop. Goth. 5.18.25–9.

This fĳits the hypothesis that the anonymous bridge was on the Anio rather
than the Tiber because the Milvian Bridge is only mentioned after the fĳirst
six camps had already been constructed to threaten fĳive gates of Rome,
from the Flaminian (Popolo) to the Praenestine (Maggiore) Gates. In this
way, Procopius provides internal evidence that the Goths built their last
and not fĳirst, camp north of the Milvian Bridge, which means that the
anonymous bridge that they initially crossed lay to the east. Once again,
therefore, the identifĳication of the anonymous towered bridge with the
Milvian Bridge is insecure whereas the Salarian Bridge, or another on the
Anio, is a much better candidate.
We can now turn to the statement in the Lib. Pontif. that Wittigis
‘collected an enormous army of Goths and returned to attack Rome on
21 February’ 537. It is unspecifĳied whether this was the date Wittigis arrived
outside Rome, or when he left Ravenna. The statement: ‘he pitched camp
by the Milvian Bridge’72 probably denotes the last camp the Goths built,
which Procopius refers to as the ‘seventh camp across the Tiber in the
Plain of Nero’.73 It would follow that the skirmish at the anonymous bridge
occurred a few days before the seventh camp was built, because Procopius
records that the Goths returned to bury their dead ‘on the third day, when
the barbarians had made camp hard by the circuit-wall of Rome’.74 Dur-
ing the building of the fĳirst six Gothic camps, the Romans evidently still
had access to the Milvian Bridge via the west bank of the Tiber, because
the Goths feared that the Romans could destroy it and ‘render inacces-
sible to them all the land on the right bank of the river’.75 So there was
no Gothic camp there initially. This further supports the hypothesis that
the Goths approached Rome from the east to focus their attack on the
north-east section of the Aurelian Wall, where it is not protected by the
Tiber. Wittigis’ march on Rome was swifter than expected by Belisarius,
who was surprised by the Goths’ arrival at the anonymous bridge, which
may justify a slightly more positive assessment of Wittigis’ generalship.
Overall, therefore, this re-examination of textual and topographic infor-
mation provides compelling evidence that the Salarian Bridge was the
Gothic point of arrival.
We can now briefly deal with the last engagement at Rome in 538. Pro-
copius relates that ‘one year . . . and 9 days’ after the start of the siege, ‘the
72 Lib. Pont. Silverius, trans. Davis (2000) 54.


Goths, having burned all their camps, set out at daybreak’, and ‘Belisarius . . .
when he saw that more than half of the enemy had crossed the bridge,
led the army out through the small Pincian Gate, and (to) . . . battle’,
after which ‘each man’ of the Goths rushed ‘to cross the bridge fĳirst’, and
‘many . . . fell offf the bridge . . ., sank with all their arms, and perished’.76
Belisarius would have led his army from the Pincian Gate along a road
which communicated directly with the anonymous bridge (the Via Pinci-
ana connects with the Via Salaria) (fĳig. 7). Had the Goths been retreating
over the Milvian Bridge, Belisarius would surely have attacked the Goths
from the Flaminian Gate up the Via Flaminia. For Wittigis, it would have
been safer to withdraw most of the Goths from the camps over the Salar-
ian Bridge rather than over the Milvian, given that it was further from, and
out of sight of, Rome.
Therefore, it is most likely that Wittigis retreated with most of his
forces back the way he came, over a bridge to the north-east (from which
communications led to Gothic settlements in Picenum and northern
Samnium, where many of his troops had homes and families). Further-
more, Procopius states that Ildiger and Martinus were sent by Belisarius
along ‘another road’, diffferent from Wittigis’ retreat route, and Procop-
ius specifĳies that these Romans did use the Via Flaminia.77 However,
as Wittigis sent some garrisons north, it is likely that the Via Cassia
was used by part of his force, perhaps troops from the seventh camp,
north of the Milvian Bridge, who could have retreated by that route, and
perhaps some of those east of, but near, the Milvian Bridge. Even though
Procopius does not name the bridge which Wittigis used, the Salarian is
the most likely candidate when we take landscape evidence into consid-
eration alongside the textual evidence.
D’Agincourt’s (ca. 1820) detailed plan of the Salarian Bridge, which was
restored by Narses in 565, and which survived into the 19th c., includes
two inscriptions, which serve as further evidence. Line six (LIBERTATEVR-
BISROMAEACTOTIUSITALIAERESTITUTA) of the ten-line-inscription pro-
claimed ‘the restoration of the liberty of the city of Rome and the whole
of Italy’.78 Lines 9 to 10 continue: ‘the river bed has been cleansed’.79 Such
76 Procop. Goth. 6.10.13–18.

77 Procop. Goth. 6.11.4 and 8 respectively.
78 Cameron, Ward-Perkins and Whitby (2000) 534; CIL VI 1199; Zanini (1998) 182–83
and cover illustration.
79 Mitchell (2007) 385.

a cleansing would have been appropriate after signifĳicant loss of life there
in 537, and especially 538. Although this inscription dates from nearly
30 years later, it may be signifĳicant that this particular bridge was inscribed
in this way. This supports the hypothesis that the Salarian Bridge is the
most likely site of the fĳirst and last engagements in 537 and 538, confĳirm-
ing Procopius’ testimony that Wittigis did indeed march through Sabine
territory in early 537.
Conclusion
In conclusion, this paper has reconciled landscape and textual evidence

to carefully re-interpret two of Procopius’ battle descriptions. In the East,
Procopius’ accuracy is supported by topographic evidence for the precise
locations of forts, camps, trench lines and battles. In the West, this paper
has mapped out the lines of advance and retreat during the Gothic army’s
attack on Rome, 537–38, with corresponding strategic moves by Belisarius.
It has been highlighted that, to a signifĳicant extent, we can trust Proco-
pius as a sound historical source, and we can gain valuable insights by
using such textual evidence in close combination with topographic evi-
dence. Thus it has been shown that historical sources, even eye-witness
accounts, leave us records which need to be carefully interpreted, and
that their reconciliation with landscape, through the use of using satellite
imagery or fĳield visits, can lead to new interpretations of locations, events
and strategies.
Our research raises questions about the objectives and research meth-
ods of primary sources, such as Procopius, and their attitudes to readers.
Sometimes, they may have mistakenly equated, left unrecorded, forgot-
ten or purposely left ambiguous details such as river or bridge names.
Logical reconciliation of literary and landscape evidence can lead us to
recover this detailed information, and so re-interpret locations and strat-
egies. This permits a better analysis of our literary source, an improved
historical interpretation of events and, therefore, an enhanced insight into
the mind-set of both the decision makers in and commentators on the
conflicts concerned. The results of our investigation indicate that we can
usually trust the fundamental tenets of Procopius’ landscape details, from
which we can ascertain more precise information with the use of compar-
ative topographical evidence and a cautious re-appraisal of translations.
This vindicates the opinion of Peter Heather, who wisely suggests “reading

Procopius’ gripping, fĳirst-hand account of the action” as it informs more

than we might expect.80
Acknowledgements
I am delighted to acknowledge the support I have received from Dr Roger

Palmer (Aerial Archaeology Research Group, Cambridge) and the British
Institute of Archaeology at Ankara, which granted me funds to visit Tur-
key. Professor Metin Ahunbay showed me around Dara and discussed the
site. I am very grateful for the translation skills of Dr Deniz Beyazit (Met-
ropolitan Museum of Art, New York) which facilitated communication.
The commander of the Turkish garrison at Dara, Kamil Gündüz, J. Bcvş,
Oğuz J.Krk.K, expedited my visit to Dara’s hinterland. Ambar and Dara
farmers were most helpful in talking of their fĳinds. Professor Wolf Lie-
beschuetz (Nottingham), Professor Stephen Mitchell (Exeter), Professor
Enrico Zanini (Sienna) and Dr Oliver Schmitt (Martin-Luther-Universität)
have helped in various ways. I should like to thank Dr Luke Lavan (Kent)
and Dr Alexander Sarantis (Kent) for agreeing to publish this paper.
I am very grateful to Carlos García for his map-making skills. Last, but
not least, I wish to thank all of my University of Oxford tutors, especially
Dr. Bryan Ward-Perkins, and the referees who provided useful feedback
and comments.
Bibliography
Primary Sources
CIL = Reimerum G. (1862) ed. Corpus Inscriptionum Latinarum (Berlin 1862).
Hdt. = de Selincourt A. (1972) trans. Herodotus, Histories (London 1972).
Hom. Od. = Rieu E. V. (2003) trans. Homer, Odyssey (London 2003).
Lib. Pont. = Davis R. (2000) trans. Liber Pontifĳicalis. The Book of Pontifffs: The Ancient Biogra-
phies of the First Ninety Roman Bishops to AD 715 (2nd ed., Liverpool 2000).
Jord. Get. = Mierow C. C. (1915) trans. The Gothic History of Jordanes (Princeton 1915).
Strategikon = Dennis G. T. (1984) ed. Maurice’s Strategikon. Handbook of Byzantine Military
Strategy (Philadelphia 1984).
Procop. Goth., Procop. Vand., Procop. Pers. = Dewing H. B. (1914–54) ed. and trans. Procopius,
History of the Wars (London 1914–54); Veh O. (1966) trans. Gotenkriege Prokop (Munich
1966); Signes Codoñer J., García Romero F. A. and Flores Rubio J. A. vols. I–III (2000–2007)
trans. Procopio de Cesarea, Historia de las guerras (Madrid 2000–2007).
80 Heather (1996) 264.

Malalas = Jefffreys E., Jefffreys M. and Scott R. with B. Croke (1986) trans. The Chronicle of John
Malalas (Melbourne and Sydney 1986).
SEG = (1993–4) Chaniotis A., Corsten T., Stroud R. S. and Tybout R. A. (1994–95) edd. Supple-
mentum Epigraphicum Graecum, volumes 44 and 45 (Leuven and Boston 1994–95).
Theoph. Sim. = Whitby M. and Whitby M. (1986) trans. The History of Theophylact Simocatta:
an English Translation with Introduction and Notes (Oxford 1986).
Zach. Myt. Chron. = Greatrex G. Phenix R. R., Horn C. B. and Brock S. P. (2010) edd. and
trans. The Chronicle of Pseudo-Zachariah Rhetor: Church and War in Late Antiquity
(Liverpool 2010); J. F. Hamilton and E. W. Brooks (1899) trans. Historia Ecclesiastica
(London 1899).
Secondary Sources
Ahunbay M. (1991) “Dara-Anastasiopolis 1990 Yili Çalismalari.”, Kazi Sonuçlari Toplantisi 13.1
(1991) 197–203.
Bury J. B. (1923) History of the Later Roman Empire from the Death of Theodosius to the Death
of Justinian, 2 vols. (London 1923).
Cameron A. Ward-Perkins B. and Whitby M. (2000) edd. The Cambridge Ancient History Vol-
ume 14. Late Antiquity: Empire and Successors AD 425–600 (Cambridge 2000)
Croke B. and Crow J. G. (1983) “Procopius and Dara”, JRS 73 (1983) 143–59.
Evans J. A. S. (1996) and (2000) The Age of Justinian: The Circumstances of Imperial Power
(London 1996) (2nd ed., London 2000).
Feissel D. (2002) “Les itinéraires de Procope et la métrologie de l’Antiquité tardive”, Antiq-
uité tardive 10 (2002) 383–400.
Greatrex G. (1998) Rome and Persia at War, 502–532 (Leeds 1998).
Greatrex G. and Lieu S. (2002) edd. The Roman Eastern Frontier and the Persian Wars: a Nar-
rative Sourcebook (London 2002).
Hassan F. A. (1981) Demographic Archaeology (New York 1981).
Haldon J. F. (2001) The Byzantine Wars: Battles and Campaign of the Byzantine Era (Stroud
2001).
Heather P. J. (2005) The Fall of the Roman Empire: a New History (London 2005).
—— (1996) The Goths (Oxford 1996).
Hodgkin T. (1880–99) Italy and her Invaders (Oxford 1880–99).
Humphries M. (2000) “Italy, A.D. 425–605”, in The Cambridge Ancient History Volume 14.
Late Antiquity: Empire and Successors AD 425–600, edd. A. Cameron, B. Ward-Perkins and
M. Whitby (Cambridge 2000) 525–51.
Istituto geografĳico militare (1884) Carta topografĳica dei dintorni di Roma (1884).
Kaldellis A. (2004) Procopius of Caesarea: Tyranny, History, and Philosophy at the end of
Antiquity (Philadelphia 2004).
Kennedy D. L. and Riley D. (1990) Rome’s Desert Frontier from the Air (London 1990).
Kouroumali M. (2006) Procopius and the Gothic War (D.Phil. diss., Univ. of Oxford 2006).
Lançon B. (2000) Rome in Late Antiquity: Everyday Life and Urban Change, AD 312–609
(Edinburgh 2000).
Liebeschuetz J. H. W. G. (1996) “The Romans demilitarised: the evidence of Procopius”,
Scripta Classica Israelica 15 (1996) 230–39.
Lillington-Martin C. (2007) “The archaeological and literary evidence for a battle near Dara,
AD 530: topography, text and trenches”, in The Late Roman Army in the East from Dio-
cletian to the Arab Conquest: Proceedings of a Colloquium held at Potenza, Acerenza and
Matera, Italy (May 2005), edd. A. Lewin and P. Pellegrini (BAR International Series 1717)
(Oxford 2007) 299–312.
Mango M. M. (1982) “Deux eglises de Mesopotamie du Nord: Ambar et Mar Abraham de
Kashkar”, Cahiers archéologiques 30 (1982) 47–70.
Martindale J. R. (1992) The Prosopography of the Later Roman Empire. Volumes 2 and 3 (Cam-
bridge 1992).

Millett M. (1990) The Romanization of Britain: an Essay in Archaeological Interpretation

(Cambridge 1990).
Mitchell S. (2007) A History of the Later Roman Empire AD 284–641: the Transformation of the
Ancient World (Malden and Oxford 2007).
Moorhead J. (2000) “Totila the revolutionary”, Historia 49 (2000) 382–86.
Nicholson O. (1985) “Two notes on Dara”, AJA 89 (1985) 663–71.
Rubin B. (1995) Das Zeitalter Iustinians (Berlin and New York 1995).
Stanhope P. H. (1848) The Life of Belisarius (2nd ed., London 1848).
Talbert R. J. A. and Bagnall R. S. (2000) Barrington Atlas of the Greek and Roman World
(Princeton and Oxford 2000).
Whitby M. (1986) “Procopius’ description of Dara (Buildings II.1–3)”, in The Defence of the
Roman and Byzantine East: Proceedings of a Colloquium held at the University of Shefffĳield
in April 1986, edd. P. Freeman and D. Kennedy (BAR International Series 297) (Oxford
1986) 737–83.
Whitby M. and Whitby M. (1986) trans. The History of Theophylact Simocatta: an English
Translation with Introduction and Notes (Oxford 1986).
Wood G. A. (2004) “The Roman fort at Qubur al Bid, Mesopotamia”, JRA 17 (2004) 397–405.
Zanini E. (1990) “La cinta muraria di Dara. Materiali per un’analisi stratigrafĳica”, in Costanti-
nopoli e l’arte delle province orientali (Rome 1990) 229–52.
Zanini E. (1998) Le Italie bizantine. Territorio, insediamenti ed economia nella provincia
bizantina d‘Italia (VI–VIII secolo) (Bari 1998).
Websites
Fuller M. J. http://users.stlcc.edu/mfuller/area1roman.html. (accessed 10 November 2011).
Google Earth leverages Landsat satellite imagery and Shuttle Radar Topography Mission
data, http://www.google.com/.
http://www.ourpasthistory.com/roman_scotland/Imgp0696.jpg (accessed July 2007).
http://www.unitconversion.org/length/kilometers-to-miles-roman-conversion.html (accessed
July 2007).
http://www.world-gazetteer.com (accessed July 2007).
List of Figures
Fig. 1 Cities and battle sites near the Roman-Persian frontier, A.D. 530–531. (prepared with
Carlos García).
Fig. 2 Sites and topography surrounding the Battle of Dara, A.D. 530. (prepared with
Carlos García).
Fig. 3 The Battle of Dara: Roman and Persian army deployments. (prepared with Carlos
García).
Fig. 4 Italy, A.D. 537–38. (prepared with Carlos García).
Fig. 5 Spoletium to Rome via Sabine country (Interamna Nahars and Reate) and Rome to
Narni. (prepared with Carlos García).
Fig. 6 Via Flaminia, Via Curia, Reate and the Via Salaria to Rome. (prepared with Carlos
García).
Fig. 7 Main gates, roads and bridges of Rome. (prepared with Carlos García).

C. Lillington-Martin. Procopius. 2013.

Uploaded by

Copyright:

Available Formats

C. Lillington-Martin. Procopius. 2013.

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

C. Lillington-Martin. Procopius. 2013.

Uploaded by

Copyright:

Available Formats

WAR AND WARFARE

Waging War in Late Antiquity .................................................................... 1

War in Late Antiquity: Secondary Works, Literary Sources and

Military Equipment and Weaponry: A Bibliographic Essay .............. 153

Tactics: A Bibliographic Essay .................................................................... 177

Organisation and Life in the Late Roman Military:

Strategy, Diplomacy and Frontiers: A Bibliographic Essay ................ 239

Fortifĳications in the West: A Bibliographic Essay ................................. 255

Fortifĳications in Africa: A Bibliography Essay ....................................... 297

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

Fortifĳications in the East: A Bibliographic Essay .................................. 317

Information and War: Some Comments on Defensive Strategy

Fortifications and Siege Warfare

Fortifĳications and the Late Roman East: From Urban Walls

Siege Warfare and Counter-Siege Tactics in Late Antiquity

Weaponry and Equipment

Late Roman Military Equipment Culture ............................................... 463

Barbarian Military Equipment and its Evolution in the Late

Recreating the Late Roman Army ............................................................. 523

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

Literary Sources and Topography

Reporting Battles and Understanding Campaigns in Procopius and

Procopius on the Struggle for Dara in 530 and Rome in 537–38:

Imperial Campaigns between Diocletian and Honorius,

The Archaeology of War and the 5th c. ‘Invasions’ .............................. 683

Controlling the Pyrenees: a Macaque’s Burial from Late Antique

The Archaeology of War: Homeland Security in the South-West

Military Encounters and Diplomatic Afffairs in the North Balkans

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

Horsemen in Forts or Peasants in Villages? Remarks on the

Military Infrastructure in the Roman Provinces North and South

El-Lejjūn: Logistics and Localisation on Rome’s Eastern Frontier

Wars within the Frontiers: Archaeologies of Rebellion, Revolt and

The Justinianic Reconquest of Italy: Imperial Campaigns and

Abstracts in French ........................................................................................ 1001

Index ................................................................................................................... 1009

Series Information .......................................................................................... 1085

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

This paper examines Procopius’ descriptions of Roman and Persian strate-

This paper re-interprets Procopius’ descriptions of opposing Roman, Per-

1 Procopius of Ceasarea was an adviser to Justinian’s general Belisarius, ca. 505–65. He

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

ern technology has opened up new possibilities of reading them. Bring-

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

The Background to the Roman-Persian Struggle for Dara in 530

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

Dara was planned as a forward post to launch potential Roman attacks

Locating the Battlefield of Dara, A.D. 530

The accurate identifĳication of any battle site as precisely as evidence per-

3 Greatrex and Lieu (2002) 91–96.

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

Fig. 3 The Battle of Dara: Roman and Persian army deployments.

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

completed before risking the movement of the workforce elsewhere.8 Such

8 Compare with Greatrex and Lieu (2002) 89.

© 2013 Koninklijke Brill NV ISBN 978-90-04-25257-8

Locating the Persian and Roman Camp Sites, A.D. 530

In this section, we will analyse Procopius’ text, and consider measure-